Structural Mechanics PDF

Structural Mechanics
Bao Shihua
Gong Yaoqing
WUPT
Wuhan University of Technology Press
本书简介
本书是根据中国高等学校土木工程专业的教学计划、结构力学课程的教学大纲和课程基本要求，
为中国高等学校土木工程专业结构力学课程‘双语教学’编写的英文结构力学教材。全书共分 14 章，
包括：绪论，结构的几何组成分析，静定梁，静定刚架，三铰拱，静定桁架和组合结构，静定结构
总论，影响线，虚功原理和结构的位移，力法，位移法，渐进法和超静定结构的影响线，矩阵位移
法和超静定结构总论。除第一章外，每章均有提要、小结、思考题和习题，书后附有答案。
本书与第一作者主编的中文结构力学（参考文献 1）在章节次序和内容上基本上是对应的，便
于中文与英文教学中交互使用。
本书可作为土木工程专业，即‘大土木’的房建、路桥、水利等各类专门化方向的‘双语’教
材及参考用书，也可供工程技术人员学习专业外语之用。
Brief Introduction of the Book
The textbook, written in terms of the national education program, syllabus and requirement of the
course—structural Mechanics—for the undergraduates majoring civil engineering in China, is intend to be
a sort of bilingually teaching material. It includes 14 chapters, which are: introduction, geometric
construction analysis of structures, statically determinate beams, plane statically determinate rigid frames,
three hinged arches, plane statically determinate trusses and composite structures, general remarks on
statically determinate structures, influence lines, principle of virtual work and displacement of structures,
force method, displacement method, method of successive approximations, matrix displacement method,
and general remarks on statically indeterminate structures. All chapters except for chapter 1 are arranged by
abstract, text, summary, problems for reflecting, and problems for solution; the answers to selected
problems are attached to the back as well.
It should be pointed out that the arrangement of the content and the order of the chapters for the book
is almost the same as its Chinese counterpart (reference book 1), so as to give an alternant usage for
Chinese readers and teachers.
The book can be not only used as a textbook and/or reference book for the students majoring civil
engineering structures, but also employed as a specialized English book for the engineer and technicians
interested in civil engineering and English.
Preface
Background
When time reaches to the 21st century, strong pressures developed from many ranks to write a
textbook of Structural Mechanics in English for Chinese undergraduates majoring in civil engineering
specialty. The naissance of the book is the balance of the pressures and the request of WUTP, the Wuhan
University of Technology Press, a press in china.
In recent years, the manner of teaching technical courses in English or partially in English is strongly
promoted by Chinese Ministry of Education. In the circumstance, the course of Structural Mechanics has
been taught in English or partially in English in Civil Engineering Department of Tsinghua University since
middle 80th of last century. Here, ‘taught in English’ means that English language is employed in all
teaching procedure, such as in class, textbooks and the like; whereas ‘taught partially in English indicates
that only Chinese language is used in class and English language are adopted in other teaching process,
such as textbook, students’ homework and so on. The two teaching ways are strongly dependent on the
English oral level of the teachers. Nowadays, undergraduates learning Structural Mechanics in Civil
Engineering Department of Tsinghua University were grouped into two classes, an English class and a
Chinese class. In the English class, class language is English, that is, the teacher must use English to prelect
and the students must read English textbooks and use English do their home works; while in the Chinese
class, class language is Chinese. The two teaching ways have been normalized into teaching plan of
Tsinghua University.
The first author of the book, Professor Bao Shihua, has been taught Structural Mechanics partially in
English in Tsinghua University since middle 80th of last century. The second author of the book, professor
Gong Yaoqing, one of PHD students of the first author, has been taught Structural Mechanics partially in
English in Ningxia University and Henan Polytechnic University for many years. The first draft of the book
has also been used in Henan Polytechnic University. Apparently, the didactical experience has provided a
bed for the appearance of the book.
In China, every specialty in a university has its own teaching plan and every course has its own
teaching program and demand depend on different specialties. So does Structural Mechanics. The course
has formed its own teaching system after nearly 50 years’ didactical experience. Nowadays, the teaching
system is quite different from that of America or the Great British because the content and system of the
textbooks of Structural Mechanics coming from the two countries do not have unified syllabus and
requirement. So it is hard for the original English textbook of Structural Mechanics to meet Chinese
syllabus and requirement. In addition, in the universities of China the teaching plan and program of
Structural Mechanics must be unchangeable no matter how the course is taught, in English or in Chinese.
The tenet has been obeyed for many years by the teachers who teach Structural Mechanics in Tsinghua
University. Based on the tenet, the book is written by means of teaching plan and program of Structural
Mechanics used in civil engineering specialty of Chinese universities. In other words, the content and
system of the Structural Mechanics written in English are identical to those of the Structural Mechanics
written in Chinese. The original intention of the consistence is to facilitate Chinese readers and teachers.
Organization and approach
Since there are too many contents pertinent to Structural Mechanics, the contents are divided, by
Chinese teaching program and demand, into two portions, fundamental portion that will be a required
course for the common students majoring civil engineering specialty and advanced portion that is planed to
be a selected course for the undergraduates or graduate students who have some advanced requirement for
structural analysis. The contents of the fundamental portion are organized by the book, named Structural
Mechanics including the analysis of statically determinate and indeterminate structures, matrix structural
analysis and so on; whereas the contents of the advanced portion are composed by another book, known as
Advanced Topics of Structural Mechanics or Advanced Structural Mechanics comprising the structural
dynamics, stabilities of structures, plastic analysis and the like. It is actually the sister book of Structural
Mechanics.
The compiling outline, content and requirement are designed by the first author, Professor Bao Shihua
of Tsinghua University. The initial draft of the book including English composition, figure design and so on
is completed by the second author, professor Gong Yaoqing of Henan Polytechnic University. The final
manuscript of the book is also revised and checked by the first author.
Each chapter of the book beginning with an abstract introducing its objective, proceeding with text
presenting its contents, ending with a summary outlining its salient features, and providing with problems
for both reflecting and solution.
The nature of the book is consistent with that of the book written in Chinese.
Contents
This book consists of 14 chapters. Chapter 1 provides a brief introduction of the various types of
structural forms and loads. The structurology (or geometric construction) of framed structures is discussed
in chapter 2. The analysis of statically determinate structures is covered in the next 5 chapters. Chapter 3
through 6 discuss the analysis of statically determinate beams, rigid frames, three hinged arches and plane
trusses and composite structures, respectively. Chapter 7 presents general remarks on statically determinate
structures so as to enhance and deepen the comprehension about the types of structures and their analytical
methods. In chapter 8 influence lines for beams, girders and trusses are discussed by static and mechanical
methods, respectively. Chapter 9 covers the principle of virtual work and evaluation of the displacements of
statically determinate structures.
In chapter 10 through 12, the analysis of statically indeterminate structures is discussed by the force
method (chapter 10), displacement method (chapter 11) and successive approximation method (chapter 12),
respectively.
Chapter 13 presents an analytical method by computer program, matrix displacement method. In the
chapter, a computer program for the analysis of plane framed structures is also attached.
Chapter 14 makes a general discussion about the analytical methods, behavioral characteristics and
computing models of statically indeterminate structures.
Acknowledgements
The development of this book was strongly influenced by the authors’ colleagues, their students and
numerous books published here and abroad. We would like to hereby acknowledge all of their valuable
suggestions, comments, help and English composition. In particular we would like to thank the second
author’s students, Gong Yun and He Shufeng, who prepared the most of the figures on the book.
The authors will be greatly indebted to the readers who dig out the errors slipped the authors’ notice in
the book because the book is written in a relatively short time, and not in the authors’ mother language.
Bao Shihua Tsinghua University
Gong Yaoqing (gongyq@263.net) Henan Polytechnic University
June, 2005
作者简介
清华大学土木工程系教授，中国力学学会《工程力
学》编委，中国建筑学会高层建筑结构委员会委员。
包世华，1985~1986 年美国伊利诺大学（ University of
Illinois）土木工程系访问学者，1991~1993 年为香
港理工大学土木与结构系研究员。长期从事高层建筑结构、结构力学、弹性
力学、能量原理及有限元、板壳结构和薄壁杆件结构等领域的教学和研究工
作。
出版教材和专著 25 本。教材有《高层建筑结构设计》、
《结构力学》、
《结
构力学教程》等，分别于 1987 年获建设部优秀教材二等奖，1988 年、1992 年获国家教委国家优秀
教材奖，1988 年度获教育部科学技术进步奖一等奖，1999 年度获国家科学技术进步奖二等奖，2002
年获全国普通高等学校优秀教材一等奖。专著有《薄壁杆件结构力学》
、《高层建筑结构计算》、《新
编高层建筑结构》等。
在国内外发表学术论文 110 多篇。其中，壳体研究成果被收入国家行业标准《钢筋混凝土薄壳
结构设计规程》。提出和创建高层建筑结构解析和半解析常微分方程求解器法系列。科研成果 1993
年获北京市科委技术成果奖，1986 年、1992 年、1994 年分别获国家教委科学技术进步奖一、二、
三等奖。
Bao Shihua, professor in Civil Engineering Department of Tsinghua University, is one of consulting
editors of Engineering Mechanics, an academic periodical superintended by the China Science Association
(CSA) and the Chinese Society of Theoretical and Applied Mechanics (CSTAM); one of commissioners in
the Tall Building Committee of the Chinese Society of Architecture. He was a visiting Scholar in Civil
Engineering Department of University of Illinois during 1985~1986, research fellow in Civil Engineering
and Structure Department of Hong Kong Polytechnic University during 1991~1993. He is engaged in the
education and academic research for the fields of tall building structures, structural mechanics, elastic
theory, energy theorem and finite element method, plate and shell structures, thin walled member structures
and so on.
He has been written totally 25 books involving textbooks and monographs. His textbooks written in
Chinese are: Design of Tall Building Structures, Structural Mechanics and so on, which were born off the
palms of second-class prize awarded by China Construction Ministry in 1987, advanced textbook prize
awarded by China Education Committee in 1988 and 1992, first-class prize of technology advancement
awarded by China Education Ministry in 1988, second-class prize of national technology advancement
awarded by China Science And Technology Ministry in 1999, and first-class advanced textbook awarded
本书简介
本书是根据中国高等学校土木工程专业的教学计划、结构力学课程的教学大纲和课程基本要求，
为中国高等学校土木工程专业结构力学课程‘双语教学’编写的英文结构力学教材。全书共分 14 章，
包括：绪论，结构的几何组成分析，静定梁，静定刚架，三铰拱，静定桁架和组合结构，静定结构
总论，影响线，虚功原理和结构的位移，力法，位移法，渐进法和超静定结构的影响线，矩阵位移
法和超静定结构总论。除第一章外，每章均有提要、小结、思考题和习题，书后附有答案。
本书与第一作者主编的中文结构力学（参考文献 1）在章节次序和内容上基本上是对应的，便
于中文与英文教学中交互使用。
本书可作为土木工程专业，即‘大土木’的房建、路桥、水利等各类专门化方向的‘双语’教
材及参考用书，也可供工程技术人员学习专业外语之用。
Brief Introduction of the Book
The textbook, written in terms of the national education program, syllabus and requirement of the
course—structural Mechanics—for the undergraduates majoring civil engineering in China, is intend to be
a sort of bilingually teaching material. It includes 14 chapters, which are: introduction, geometric
construction analysis of structures, statically determinate beams, plane statically determinate rigid frames,
three hinged arches, plane statically determinate trusses and composite structures, general remarks on
statically determinate structures, influence lines, principle of virtual work and displacement of structures,
force method, displacement method, method of successive approximations, matrix displacement method,
and general remarks on statically indeterminate structures. All chapters except for chapter 1 are arranged by
abstract, text, summary, problems for reflecting, and problems for solution; the answers to selected
problems are attached to the back as well.
It should be pointed out that the arrangement of the content and the order of the chapters for the book
is almost the same as its Chinese counterpart (reference book 1), so as to give an alternant usage for
Chinese readers and teachers.
The book can be not only used as a textbook and/or reference book for the students majoring civil
engineering structures, but also employed as a specialized English book for the engineer and technicians
interested in civil engineering and English.
by China Education Ministry in 2002, respectively. His monographs written in Chinese are: Structural
Mechanics for Thin Walled Member Structures, Analysis of Tall Building Structures, Newly Edited Tall
Building Structures and so forth.
His published papers both in his mother language and foreign language exceed 110 pieces, of which
papers pertinent to shell structures are adopted by national profession code—Design Code of Reinforce
Concrete Thin Shell Structures. His pioneering work is analytical and semi-analytical methods by Ordinary
Differential Equation Solver used for analyzing tall building structures. His research findings were brought
down the persimmon, and awarded for technology advancement prize by Beijing government in 1993, for
technology advancement fist-class prize, second-class prize and third-class prize by China Education
Committee in 1986, 1992 and 1994, respectively.
作者简介
1999 年清华大学土木工程系毕业并获结构工程
龚耀清专业工学博士学位，师从我国著名土木工程专家
，龙驭球院士和我国著名高层建筑专家包世华教
授。现为河南理工大学特聘教授，研究生导师;
国家一级学会中国力学学会主办的中文核心期刊《工程力学》编委会委员;
国家自然科学基金委员会工程与材料科学部项目评审专家。1994 年被中国
力学学会评为优秀力学教师。长期从事《结构力学》的教学工作，近年来
主要从事超高层建筑结构与大型桥跨结构的分析与研究，发表学术论文 30
多篇，其中部分已被 EI 收录；出版了首部用半解析方法分析超高层建筑结构的专著《弹性地基上高
层建筑结构及半解析法研究》；先后参加和主持国家级、省部级科研基金项 9 项，研究成果已得到国
内外同行的认可和关注。目前正在英国爱丁堡大学 (Edinburgh University) 作高级研究者。
Gong Yaoqing, graduated and obtained doctor’s degree of structural engineering in Civil Engineering
Department of Tsinghua University, is a student of Academican Long Yuqiu who a well known expert in
civil engineering and Professor Bao Shihua who is a famous expert in tall building structures, a specially
engaged professor and supervisor of postgraduate students in Henan Polytechnic University, one of editors
of Engineering Mechanics, an academic periodical superintended by the China Science Association (CSA)
and the Chinese Society of Theoretical and Applied Mechanics (CSTAM), one of experts examining and
commenting proposal projects in Department of Engineering & Materials Science of National Natural
Science Foundation of China. He got an honor of excellent teacher from Chinese Society of Theoretical and
Applied Mechanics in 1994.
He has been engaged in the education and academic research for structural mechanics, super tall
building and very long-span bridge structures for many years. His published papers have been more than 30
pieces, some of which are accepted by EI Index. In 2003, he published a monograph—Tall Building
Structures on Elastic Subgrade and Research of Semi-Analytical Method (in Chinese), which is a first
monograph pertinent to the semi-analytical analysis about tall and super tall building structures. He has
finished 9 researching projects involving with those helped by National Natural Science Foundation of
China and local government. His research findings have been paid attention to and interested in by
pertinent fellows. He has been doing some research work in Edinburgh University of United Kingdom.
Contents
Chapter 1 Introduction 1
1.1 Structures and Their Classification 1

1.2 Objective and Learning Method of Structural Mechanics 3
1.3 Analytical Models of Structures 5
1.4 Classification of Framed Structures 13
1.5 The Classification of Loads 17
Summary 18
Problems for Reflecting 18
Chapter 2 Geometric Construction Analysis of Structures 19
2.1 Purpose of Analyzing Geometric Construction of Structures, Stable and Unstable Structural
Systems 19
2.2 The Concept of Degrees of Freedom and Restraints 20
2.3 Geometric Construction Rules of Planar Stable Framed Systems Without Redundant
Restraints 25
2.4 Illustration of Geometric Construction Analysis 31
2.5 The Relationship between Static Determinacy and Geometric Construction of Structures
32
Summary 33
Problems for Solution 34
Chapter 3 Statically Determinate Beams 38
3.1 The Analysis of Single Span Beams 38

3.2 Construction of Bending Moment Diagram By Principle of Superposition for Straight
Members 52
3.3 The Analysis of Simply Supported Inclined Beams 55
3.4 The Restraint Force Calculation and Geometric Construction of Statically Determinate
Multispan Beams 59
i
ii Contents
3.5 The Construction of Internal Force Diagrams of Statically Determinate Multispan Beams
62
Summary 66
Chapter 4 Plane Statically Determinate Rigid Frames 73
4.1 Geometric Construction and Characteristics of Plane Statically Determinate Rigid Frames
74
4.2 The Analyzing of Reactions for Statically Determinate Rigid Frames 76
4.3 Determination of Internal Forces of Member Ends by Using The Method of Sections
79
4.4 The Construction of Internal Force Diagrams of Statically Determinate Frames 84
4.5 Internal Force Diagrams of Statically Determinate Three-hinged Frames, Multispan and
Multistory Frames 87
Summary 92
Chapter 5 Three Hinged Arches 101
5.1 The Constitution and Type of Three Hinged Arches 101

5.2 The Reactions of Three Hinged Arches Under The Action of Vertical Loads 102
5.3 The Formula for Calculating Internal Forces of Three Hinged Arches Under Action of
Vertical Loads 105
5.4 Stressing Performance of Three Hinged Arches 110
5.5 Rational Axial Lines of Three Hinged Arches 111
Summary 116
Chapter 6 Plane Statically Determinate Trusses and Composite Structures 120
6.1 Characteristics and Classification of Trusses 120

Contents iii
6.2 The Method of Joints 126

6.3 The Method of Sections 132
6.4 The Combination of The Method of Joints and The Method of Sections 136
6.5 Form and Stressing Characteristics of Girder Trusses 138
6.6 Composite Structures 142
Summary 147
Chapter 7 General Remarks on Statically Determinate Structures 155
7.1 Analytical Methods For Statically Determinate Structures 155

7.2 General Property of Statically Determinate Structures 160
7.3 Stressing Characteristics of Various Types of Structures 164
Chapter 8 Influence Lines 169
8.1 Concept of influence lines 169

8.2 Influence Lines for Statically Determinate Single Span Beams by Static Method 172
8.3 Influence Lines for Girders with Floor Systems 180
8.4 Influence Lines for Trusses by Using Static Method 184
8.5 Virtual Displacement Method for Constructing Influence Lines for Statically Determinate
Beams 187
8.6 Application Of Influence Lines 197
Summary 212
Chapter 9 Principle of Virtual Work and Displacement of Structures 220
9.1 Introduction for Calculation of Structural Displacement 220

9.2 Virtual Work and Principle of Virtual Work 222
9.3 General Equation and Unit Load Method for Computing Displacements 235
iv Contents
9.4 Calculation of Displacements Caused by Loads 236

9.5 Graph-Multiplication Method 245
9.6 Calculation of Displacements Caused by Temperature Changes 260
9.7 Calculation of Displacements Caused by Support Movement 263
9.8 Reciprocal Laws for Linear Elastic System 266
Summary 269
Chapter 10 Force Method 280
10.1 Statically Indeterminate Structures and Degrees of Indeterminacy 280

10.2 Basic Concept of Force Method 283
10.3 Canonical Equations of Force Method 287
10.4 Statically Indeterminate Beams, Rigid Frames and Bent Frames 292
10.5 Statically Indeterminate Trusses and Composite Structures 304
10.6 Analysis of Symmetric Structures 310
10.7 Statically Indeterminate Arches 326
10.8 Internal Forces due to Support Settlements and Temperature Changes 340
10.9 Computation of Displacements of Statically Indeterminate Structures 347
10.10 Verification of Calculation of Statically Indeterminate Structures 352
Summary 356
Chapter 11 Displacement Method 370
11.1 Fundamental Concepts of Displacement Method 370

11.2 Member End Forces of Various Single-Span Indeterminate Prismatic Beams Due to Their
End Displacements and External Loadings 374
11.3 Primary Unknowns and Primary Systems in Displacement Method 382
11.4 Displacement-Method Equations 387
11.5 Analysis of Statically Indeterminate Beams and Rigid Frames with No Sidesway 393
11.6 Analysis of Statically Indeterminate Rigid and Bent Frames with Sidesway 401
11.7 Analysis of Statically Indeterminate Symmetric Structures 411
Contents v
11.8 Development of Displacement-Method Equations by Direct Stiffness Method 416

Summary 421
Chapter 12 Method of Successive Approximations

and Influence Lines for Indeterminate structures 431
12.1 General Remarks for Method of Successive Approximations 431

12.2 Concepts and Terminology in Moment Distribution Method 432
12.3 Moment Distribution at A Single Joint—Moment-Distribution Process 439
12.4 Moment Distribution at Multi-Joints—Successive Approximations 446
12.5 No-Shear Moment Distribution Method 461
12.6 Shear Distribution Method 469
12.7 Influence Lines for Forces of Statically Indeterminate Structures 476
12.8 Most Unfavorable Arrangement of Live Loads and Envelops for Internal Forces of
Continuous Beams 483
Summary 488
Chapter 13 Matrix Displacement Method 495
13.1 Introduction 495

13.2 Member Stiffness Matrix in Local Coordinate System 500
13.3 Member Stiffness Matrix in Global Coordinate System 507
13.4 Global Stiffness Matrixes of Continuous Beams 513
13.5 Global Stiffness Matrixes of Rigid Frames 525
13.6 Equivalent Nodal Loads 531
13.7 Procedure for Analysis and Examples 538
13.8 Global Analysis of Rigid Frames without Considering Axial Deformation 549
13.9 Block Diagram and Computer Programs for Plane Framed Structures 556
Summary 583
vi Contents
Chapter 14 General Remarks on Statically Indeterminate Structures 588
14.1 Classification and Comparison of Fundamental Method used to Analyze Statically

Indeterminate Structures 588
14.2 Combined Method—Simultaneously Applying Moment Distribution and Displacement
Methods to Frames with Sidesway 591
14.3 Approximate Analysis of Statically Indeterminate Structures 596
14.4 Properties of Statically Indeterminate Structures 604
14.5 Complementary Discussion about Computing Models of Structures 611
Summary 618
Answers to selected problems Ⅰ
Bibliography
CHAPTER 1
INTRODUCTION
The abstract of the chapter

The chapter will introduce four questions such as the objective and learning method of structural
mechanics, the analytical models of structures, the classification of framed structures and their loads.
Among the four questions, the analytical models are the most significant since they will lay the
foundations of the other chapters of the book.
1.1 Structures and Their Classification
In civil engineering project, the generic term an engineering structure or briefly named a structure is
referred to as a frame or skeleton used to carry loads applied on it and composed by members of buildings
or other constructions made of construction materials. The following figures show some photographs of
engineering structures. Fig.1.1 is a tall building suspended structure, Fig.1.2 is a bridge structure, Fig.1.3 is
Fig. 1.2 Bridge structure
Fig. 1.3 Hydraulic power station
Fig.1.1 Tall building suspended structure

Fig. 1.4 Industrial premises
the structure of a hydraulic power station, Fig.1.4 is the structure of an industrial premises. Speak in detail,
the roof panel, the roof truss, the beams, the columns, the foundation and their combination of the
1
2 Chapter 1 Introduction
one-storey workshop of a plant shown in Figure 1.18 are all structures.

Structures can be classified into three categories by their geometric characteristics.
(1) Framed structure
A framed structure is comprised of members whose
h
cross-sectional dimension (e.g. the width b and depth h of
b a rectangular cross section, the radius of a circular cross
l
section, etc.) is much smaller than the length l as shown
Fig.1.5 Bar in Fig. 1.5. The most commonly used types of structures
in the structural engineering are framed structures which
h will be the main attention the book focuses on.
b (2) Plate, Slab and Shell (or thin-walled structure)
l When the thickness of the structure is very small in
comparison with its other two dimensions (length and
Fig.1.6 Plate
width), the structure is referred to as a thin-walled
structure. The geometric characteristic of a thin-walled
structure is that its thickness h is much smaller than its length l and width b. The plate shown in Fig. 1.6 is
one of instances of thin-walled structures. The combination of finite number of plates would develop a
floded plate shown as in Fig. 1.7 (a). Figure 1.7 (b) shows a building with a roof structure composed of
floded plates. If a structure has a curved middle surface, it is called a shell, as shown in Fig. 1.8. In that
context, a plate or a slab can be considered as a thin-walled structure with a plane middle surface.
(a) floded plate (b) floded plate roof
Fig.1.7 Floded plate structure
(3) Massive Structure

The three dimensions (length l, width b and depth h) of a massive structure have the same order of
magnitude. Consider, for example, the retaining wall shown in Fig. 1.9 and the dam shown in Fig. 1.10 are
two of projects applying massive structures.
1.2 Objective and Learning Method of Structural Mechanics 3
(a) shell (b) shell roof

Fig.1.8 Shell structure
Fig.1.9 Retaining wall Fig.1.10 Dam
1.2 Objective and Learning Method of Structural Mechanics
1.2.1 The relationship between structural mechanics and other curricula
In structural mechanics, the primary focus will be on the analysis of. In this context, you can name the
course as Mechanics of Framed Structure. For simplicity, Structural Mechanics is adopted.
Structural Mechanics belongs to one of technically fundamental courses, and plays a very important
role and a connecting link between the preceding courses and the following (or subsequent) courses learned
by the undergraduates majoring in the specialty pertinent to civil engineering.
Structural Mechanics is the following course of Theoretical Mechanics and Strength of Materials. The
objective of Theoretical Mechanics is the investigation of essential rules and analysis of mechanical motion
(including static state and equilibrium) of rigid bodies. The attention paid by Strength of Materials is the
strength, stiffness and stability of a single member (or a bar). While the contents treated of in Structural
Mechanics are the strength, stiffness and stability of framed structures, which are composed of many
members. Therefore, Theoretical Mechanics and Strength of Materials would provide primary principles
and basement of mechanical analysis for the studying of Structural Mechanics.
Structural Mechanics is meanwhile the preceding course of Theory of Elasticity (focusing on the
strength, stiffness and stability of plate, shell and massive structures), Reinforced Concrete Design,
Masonry Structure, Steel Structure and other specialized curricula associated with building construction,
structural engineering, highway engineering, bridge engineering, water conservancy engineering and
underground engineering. By this token, Structural Mechanics will provide basic mechanical knowledge for
the studying of the subsequent courses and play a very important role in the specialty pertinent to civil
engineering.
1.2.2 Objective and learning method of Structural Mechanics
The objective of Structural Mechanics comprises following aspects:

（1） Discuss principles for constructing structures, rational configurations of structures and
selections of computing models for analyzing structures;
（2） Investigate the methods for analyzing internal forces and displacements of structures so as to
check their strength and stiffness;
（3） Study structural stability and structural response induced by dynamic loading.
The analyzing problems involved in Structural Mechanics can be classified into two categories: the
first category is statically determinate problems, which could be solved by means of force equilibrium
conditions, the first fundamental condition listed in next paragraph; the second category is statically
indeterminate problems, which could be figured out only by satisfying all of the following three types of
fundamental conditions. The three types of fundamental conditions are:
（1） Force equilibrium conditions The entire structure or part of it must be balanced under the
action of a system of forces.
（2） Compatibility conditions or geometrical conditions of displacements The continuity of a
structure must be maintained after the structure has deformed under the action of the loadings applied on it.
That is, there are no overlap and gap existing in the materials composing the structure, and meanwhile the
deformation and displacement of the structure should satisfy the restraint conditions provided by the
supports and the connecting joints.
（3） Physical conditions This is the physical performance condition linking stress and strain or
forces and displacements yielding in a structure, i.e. physical or constitutive equation.
The all analyzing methods formulated in Structural Mechanics are permeated by aforementioned three
types of fundamental conditions, i.e. every analyzing method discussed in the course has to utilize the three
conditions in some kind of degree and order.
During studying the course, we should focus our attention on the relationship between Structural
Mechanics and other curricula. A necessary review about the knowledge associated with Theoretical
Mechanics and Strength of Materials should be taken and the knowledge also should be consolidated and
advanced during the study.
In the process of studying the course, we have to pay attention to the leaning method and clue of
solving problems. All of the analyzing methods discussed in the book are the concrete application of above
mentioned three types of fundamental conditions. How the three types of conditions are applied to the
computing process of every analyzing method is the point we should focus on. During studying the course,
the clues of solving problems must be mastered; especially the general method of analyzing problems
should be holden. For instances, the method of how passing through unknown field to known field; the
method of how partitioning the entire structure into elements then integrating them into the very beginning
structure, the method of how contrasting pertinent problems between which some sort of relationship exist
in some extent, etc.
Study should associate with practice. Doing exercises is a very important approach for learning
Structural Mechanics. We will not predominate the primary concept, principle and method of Structural
Mechanics without doing some amount of exercises. However, doing exercises should avoid silliness. The
number one silliness is only doing exercises without reading and review the book. The number two silliness
is doing exercises fast and more without understanding them completely. The number three silliness is
doing exercises with referring to the answers but being not able to verify the exercises by self. The number
four silliness is doing exercises without rectifying fault happened in doing exercises and extracting lesson
from them.
1.3 Analytical Models of Structures
1.3.1 The rules of developing analytical model
Real structures are usually too complex to perform a rational analysis in their real practical states.
They have to be simplified or idealized, by throwing away some unimportant details, as analyzing models
prior to computation. An analytical model is a simplified representation of sketch, or an ideal sketch, of a
real structure for the purpose of analysis. The simplified representation of sketch of a structure is referred to
as its analytical model. All the analysis of structures is performed in their analytical models. Therefore, the
development of analytical model is the basement of structural analysis. Establishment of an analytical
model is one of the most important steps of the analysis process; it requires experience and knowledge of
design practices in addition to a thorough understanding of the behavior of structures. Remember that the
structural response predicted from the analysis of the model is valid only to the extent that the model
represents the actual structure.
Development of the analytical model should generally obey the following rules:
（1） The analytical model should reflect, as accurately as practically possible, the main stressed and
strained characteristics of the structure of interest to the analyst;
（2） In order to facilitate the analysis, the analytical model should maintain principal factors and
discard much of the detail about the members, connections, and so on, that is expected to have little effect
on the desired characteristics.
It should be point out that the development of analytical model should meet the site and time
conditions under the above rules. It should have alternative ways. For instance, a more precise analytical
model should be developed for an important structure; while a less precise analytical model should be used
for an unimportant structure. Furthermore, in schematic design stage we can develop a rough analytical
model for a structure; while in technical design stage we can select a more precise analytical model for the
same structure. For hand-oriented method, the simplest analytical models of structures should be used;
while for computer-oriented method, complex models of structures might be selected.
1.3.2 Simplifying points of analytical model
(1) The simplification of structural system

Generally, the actual structures are space, or three-dimensional, structures, whose members are
connected as a space frame to undergo the loadings likely acting on it in various sense. Fortunately, many
actual three-dimensional structures can be subdivided, by discarding some subsidiary space restraints, into
plane structures for simplifying analysis. If all the members of a structure as well as the applied loads lie in
a single plane, the structure is called a plane structure. The book will mainly discuss the calculation
problems of plane structures.
(2) The simplification of members
The main attention the book focuses on is the framed structures. A framed structure is comprised of
members (or bars), whose cross-sectional dimensions (e.g. the width b and depth h of a rectangular cross
section, the radius of a circular cross section, etc.) is much smaller than the length l. Also cross-sectional
deformation of the members satisfies plane-section assumption, and stress acting on the cross sections of
members can be determined by the stress resultants (i.e. internal forces, bending moment, shear force and
axial force). The cross-sectional deformation can be also evaluated by the strain components of the
centroidal axes of the members. In the circumstances, the analytical model of a two- or three-dimensional
structure selected for analysis is represented by a line diagram. On this diagram, each member of the
structure is represented by a line coinciding with its centroidal axis; each connection between the members
is represented by a kind of joint; the length of each member is represented by the distance between joints to
which the member be attached; the position of the loads acting upon members is also transmitted to their
centroidal axes. However, the simplifying approach is only suitable for the members whose ratios between
their longitudinal and lateral dimensions are great than 4.
(3) The simplification of connections
The connections between members of a structure are commonly simplified into joints. Two types of
joints are commonly used to join members of structures:
c Flexible, or hinged, joint

A hinged joint prevents only relative translations of member ends connected to it; that is, all member
ends connected to a hinged joint have the same translation but may have different rotations. Such joints are
thus capable of transmitting forces but not moments between the connected members. Fig.1.11 (a) shows a
roof structure, the relative translations of each member are restrained by the gussets, but slight rotation
between members will exist. So the gussets or connections of the roof structure are idealized into hinges.
Hinged joints are usually depicted by small circles at the intersections of members on the line diagram of
the structure, as shown in Fig. 1.11(b).
d Rigid joint
A rigid joint prevents relative translations and rotations of the member ends connected to it, that is, all
member ends connected to a rigid joint have the same translations and rotation. In other words, the original
angles between the members intersecting at a rigid joint are maintained after the structure has deformed
(a) (b)
Fig.1.11 Roof structure and its computing model
(a) (b)
Fig.1.12 Connection of reinforced concrete bars and its computing model
under the action of loads. Such joints are, therefore, capable of transmitting forces as well as moments
between the connected members. Fig.1.12 (a) shows a connection of a reinforced concrete side column and
beam, the relative translations and rotation of the column and beam are restricted by the arrangement of the
reinforcements which are cast into the whole body by the concrete. So the connection is simplified into a
rigid joint. Rigid joints are usually represented by monolithic point at the intersections of members on the
line diagram of the structure, as shown in Fig. 1.12 (b).

(4) The simplification of connections between structure and its foundation
Supports are used to attach structures to their foundations or other bodies, thereby restricting their
movements under the action of applied loads. The loads tend to move the structures; but supports prevent
the movements by exerting opposing forces, or reactions, to neutralize the effects of loads, thusly keeping
the structures in equilibrium. The type of reaction depends on the type of supporting device used and the
type of movement it prevents. A support that prevents translation of the structure in a particular direction
exerts a reaction force on the structure in that direction. Similarly, a support that prevents rotation of the
structure about a particular axis exerts a reaction couple on the structure about that axis.
The types of supports commonly used for plane structures are grouped into four categories, depending
on the number of reactions (1, 2, or 3) they exert on the structures.
(c) the rocker support of a bridge
(a) Photo of the roller support of a bridge A A
RA (d) computing models of roller support
A A
(b) the roller support of a bridge (e) computing model of roller support
Fig.1.13 Roller support
c Roller support
Fig.1.13 (a) shows a photo of the roller support of a bridge structure; Fig.1.13 (b) and (c) depict the
roller and rocker supports used in bridge structures. The supports only prevent translation perpendicular to
the supporting surface. So the reaction force RA acts perpendicular to the supporting surface and may be
directed either into or away from the structure. The magnitude of RA is the unknown. The support is
thusly idealized, in according with its behavioral characteristics, as a roller-liked symbol shown as in
Fig.1.13 (d) or a link shown as in Fig.1.13 (e).
d Hinged support
The conformations of this kind supports are depicted in Fig.1.14 (a) and (b), they are simply referred
to as hinge. The supports are able to prevent translations in any direction. So the reaction force RA may
YA RA
A XA A
(a) (b)
A A
A
(c) (d) (e)
Fig.1.14 Hinged support

(a) and (b) conformations of hinged support;
(c), (d) and (e) computing models of hinged support
lattice beam
asphaltum with jute fiber A
panel support
support arm
main truss
vertical beam
(a) (b)
Fig.1.15 Practical hinged supports

(a) precast reinforced concrete column embedded in cup-liked foundation;
(b) arc-shaped sluice with hinged support
act in any direction. It is usually convenient to represent R A by its rectangular components, X A and Y A .
The magnitudes of X A and Y A are the two unknowns. The support is thusly idealized, in according with
its behavioral characteristics, as a hinge-liked symbol shown as in Fig.1.14 (c) or two concurrent links
shown as in Fig.1.14 (d) and (e).

Fig.1.15 (a) shows a precast reinforced concrete column embedded in cup-liked foundation and the
interspace around the column bottom end is filled by asphaltum with jute fiber. In practice, there is slight
rotation between the column and the foundation but no translations. So the attachment can be simplified as
a hinge. Fig.1.15 (b) shows another example of hinged support, an arc-shaped sluice, there is a rotation
about pin A when the sluice is being opened.
e Fixed support
The fixed support can prevent both relative translations and rotation between structure and its
foundation. So the reactions consist of two force components X A , Y A and a couple of moment M A .
The magnitude of X A , YA and M A are the three unknowns. The support is thusly idealized, in
according with its behavioral characteristics, as a symbol shown as in Fig.1.16 (b).
Fig.1.16 (c) shows a precast reinforced concrete column embedded in cup-liked foundation and the
interspace around the column bottom end is filled by crushed stone concrete. In practice, there is no
movement between the column and the foundation when the depth of embedment reaches some kind of
degree. So the attachment generally is simplified as a fixed support.
crushed stone concrete

MA
A
XA
YA
(a) (b) (c)
Fig.1.16 Fixed support

(a) conformation of fixed support; (b) computing model of fixed support;
(c) precast reinforced concrete cup-liked foundation
f Directional support
The directional support (or double-link support) restricts all relative movement between structure and
its foundation but slides along its supporting surface. So the reactions consist of a force YA perpendicular
to the supporting surface and a couple of moment M A . The magnitudes of Y A and M A are the two
unknowns. The support is thusly idealized, by its behavioral characteristics, as a symbol shown as in
Fig.1.17 (b) or two parallel links shown as in Fig.1.17 (c).
(5) The simplification of property of material
Generally, the structures constructed in civil engineering are made of materials such as steel, concrete,
bricks, stone, timber, and so on. However, in order to simplify the structural analysis, all materials
composing structures are assumed to be continuous, homogeneous, isotropic, perfectly elastic or plastic.
A
A MA
YA
(a) (b) (c)
Fig.1.17 Directional support

(a) conformation of directional support; (b) schematic symbol of directional support;
(c) computing model of directional support
Above assumption is suitable for metal within some stressing extent, but for concrete, reinforced
concrete, bricks, stone and the like, the assumption will have some degree of approximation. As far as
timber, because the property along the timber grain is quite different from that cross the timber grain, the
attention should be paid when applying the assumption.
(6) The simplification of load
The loads applied on structures may be divided into two categories such as body force and surface
force. The body force indicates the gravities and inertial forces of structures and the like; the surface force
means the action upon structures transmitted by other bodies attached to them, for instance, the
compression of soil and vehicular wheels and so on. In framed structures, the members are represented by
lines coinciding with their centroidal axes, so both body force and surface force can be simplified as the
forces acting on the lines. The loads can be simplified into concentrated loads and distributed loads
according to their distribution. The simplification and determination of loads are of complexity. The special
discussion about them will be given in a latter section.
1.3.3 Illustration of analytical models of structures
Fig.1.18 (a) shows an illustrative diagram of a workshop of a factory. Now its analytical models will
be discussed.
(1) The simplification of structural system of the workshop
The structure of the workshop is, like clockwork along its longitudinal direction, collocated by a series
of planar elements, composed by roof structure, columns and foundation shown as dashed area or in
Fig.1.18 (b), and then connected by roof panels and other longitudinal members as a space structure. The
loads acting upon the workshop are usually uniformly distributed along its longitudinal sense. Therefore,
the portion between adjacent centroidal axes of two columns could be extracted as the analytical element
from the space structure, while the loads acting upon the structure could be distributed to each analytical
element by the longitudinal members. Because in each analytical element, its members as well as its loads
lie in a single plane, the space structure would be thusly divided into plane structures shown as in Fig.1.18
(b).
For the plane structure shown in Fig.1.18 (b), the analytical models of its roof structure and columns
will be discussed in the following.
roof structure
foundation
roof panel
column
(a) (b)
(c) (d)
Fig.1.18 Analytical model of the structure of one-storey workshop

(a) the structure of one-storey workshop; (b) planar analytical element;
(c) analytical model of roof structure under vertical load;
(d) analytical model of framed bent under horizontal load
(2) The analytical model of roof structure under vertical loads

The analytical model of roof structure under the action of vertical loads is shown in Fig.1.18 (c), here
the following simplification has made:
c Each member of the roof structure is represented by a line coinciding with its centroidal axis;
d Each connection between the members is idealized as a hinged joint;
e The two ends of the roof structure welded together with the columns below it by steel gussets are
simplified as one roller and one hinged support.
f The loads are recognized as concentrated forces located at four corners of the roof slab and acting
on the top chords of the roof truss.
(3) The analytical model of the columns under action of horizontal loads
The analytical model of the columns of the workshop structure under the action of horizontal loads
(for example, the lateral wind load) is shown in Fig.1.18 (d), here the following simplification has made:
c Each column of the workshop structure is represented by a line coinciding with its centroidal
axis;
d The connections between two ends of the roof structure and the support heads of the columns are
simplified as hinged joints. Here the function of the roof structure is realized as a link which connect the
two columns;
e Because the crushed stone concrete would be used to fill the interspace between the bottom ends
of the columns and their foundations, the supports are simplified as fixed supports.
The structure shown as in Fig.1.18 (b) or (d) are referred as framed bent, it is one of common types of
structures to be used in one-storey workshops in industrial factories.
Establishment of the analytical model is the most important step of the analysis process; it is complex
too. It requires experience and knowledge of design practices in addition to a thorough understanding of the
behavior of structures. For a new type of structures, a rational analytical model would be developed by
many tests and practices sometimes. However, for commonly used types of structures, their analytical
models will be straight developed according to the experience of the predecessors devoted to structural
analysis. Regrettably it is nearly impossible to give the training of ability of developing analytical models
of structures only in school except in a most indirect sense. A rational analytical model comes from
intuition based on our experience, judgment, knowledge, and our mastery of information concerning the
engineering structures and civil engineering projects.
It should be note that the analytical model of the two- or three-dimensional structure selected for
analysis is represented by a line diagram. On this diagram, two lines ( ) are often used in this text to
represent members on the line diagrams. This is done, when necessary, for clarity of presentation; in such
cases, the distance between the lines does not represent the member depth.
1.4 Classification of Framed Structures
In practice, perhaps the most important decision made by a structural engineer in implementing an
engineering project is the selection of the type of the structure used for supporting or transmitting the
loadings. Commonly used structures can be classified according to their composition and stressing
characteristics or their method being analyzed or the arrangement of their members as well as the locations
of the applied loads acting on. Essentially, the classification of framed structures is the classification of
their analytical models.

Commonly used structures can be classified into following categories, depending on their composition
and stressing characteristics.
(a) (a)
(b) (b)
Fig.1.20 Multispan beam

Fig.1.19 Single span beam
(a) statically determinate multispan beam;
(a) statically determinate beam;
(b) statically indeterminate multispan beam
(b) statically indeterminate beam
(1) Beams
The centroidal axis of a beam is generally a straight line. For a horizontal beam, there is no horizontal
reaction under transverse loads, its internal forces are bending moment and shear force. There are single
span beam [Fig.1.19 (a) and (b)] and multispan beam [Fig.1.20 (a) and (b)].
H H
H H
(b)
(a)
Fig.1.21 Arch
(a) three-hinged arch; (b) hingeless arch
(2) Arches
The centroidal axis of an arch is a curved line. Horizontal thrusts H [Fig.1.21 (a) and (b)] will be
developed at the supports of the arch under the action of vertical loads. The horizontal thrusts would
improve, out and away, the stressing property of the arch.
(3) Rigid frames
A rigid frame is a system of several members, hinged or rigidly connected at their ends as shown in
Fig.1-22. Generally, there exist bending moment, shear and axial force in rigid frame members, of which
bending moment is the dominated internal force.
(a) (b)
Fig.1.22 Rigid frame

(a) statically determinate rigid frame; (b) statically indeterminate rigid frame
(4) Trusses
A truss consists of a group of members which are assumed to be connected at their joints with
frictionless hinges as shown in Fig. 1-23. When a truss is subjected to the loads applied at its joints, it
transfers the loads by developing axial forces in its members.
(a) (b)
Fig.1.23 Truss
(a) statically determinate truss; (b) statically indeterminate truss
(5) Composite structures

A composite structure is a combination of beam-typed members (mainly resist bending) and links or
two-force members as shown in Fig.1.24.
Framed structures can be also classified into two categories, statically determinate structures and
statically indeterminate structures, according to analyzing method applied to them.
(1) Statically determinate structures
A stable structure is considered to be statically determinate if all its support reactions and internal
forces can be determined by solving the equations of equilibrium. The structures shown in figure (a) from
Fig.1.19 to Fig.1.24 are all statically determinate structures.
(2) Statically indeterminate structure
Conversely, a stable structure is considered to be statically indeterminate if all its support reactions
and internal forces can not be determined by solving the equations of equilibrium. The structures shown in
figure (b) from Fig.1.19 (b) to Fig.1.24 (b) are all statically indeterminate structures.
B
(a)
(b)
Fig.1.24 Composite structure

(a) statically determinate composite structure; (b) statically indeterminate composite structure
Furthermore, framed structures can be classified into another two categories, plane structures and
space structures, according to the arrangement of their members as well as the locations of the applied
loads acting on.
(1) Plane structures
If all the members of a structure as well as the applied loads lie in a single plane, the structure is called
a plane structure. The foregoing mentioned structures are the primary structures concerned with in the
book.
y P
z
P
x
(a) (b)
Fig.1.25 Space structure

(a) space structure; (b) plane structure under the action of nonplanar load
(2) Space structures

Some structures, such as latticed domes, aerospace structures, and transmission towers, cannot, due to
their shape, arrangement of members, or applied loadings, be subdivided into planar components. Such
1.5 The Classification of Loads 17
structures, called space structures, are analyzed as three-dimensional bodies subjected to three-dimensional
force systems.
Fig.1.25 (a) is a space structure whose members do not lie in a single plane; although the members of
the structure shown in Fig.1.25 (b) lie in a single plane, the applied load lie in another plane, the structure
belongs to space structure.
1.5 The Classification of Loads
The objective of a structural engineer is to design a structure that will be able to withstand all the loads
to which it is subjected while serving its intended purpose throughout its intended life span. In designing a
structure an engineer must, therefore, consider all the loads that can realistically be expected to act on the
structure during its planned life span. Well then, what is load? In fact, the loads are the external active
forces imposed on structures. The deadweights (or gravities) of structures, the crane loads acting on the
structures of workshops, the loads due to vehicular traffic on bridges, hydraulic pressure and soil pressure
acting on the hydraulic structures are the examples of loads.
The loads may be divided into the following categories according to their duration and the manner of
application.
(1) According to their duration, the loads can be divided as:
c Dead loads Dead loads are loads of constant magnitudes and fixed positions that act
permanently on the structure. For instance, the weights of the structural system itself and of all other
materials and equipment permanently attached to the structural system and the like.
d Live loads Live loads are loads of varying magnitudes and/or positions caused by the use of
the structure. Sometimes, the term live loads are used to refer to all loads on the structure that are not dead
loads, including environmental loads, such as crowd loads on floors, crane loads, snow loads or wind loads,
etc. Live loads can be subdivided into following two categories according to the situation of their varying
positions:
a. Movable loads Movable loads are those which may be located at one position or another on a
structure, such as furniture and goods on a building floor and snow on a roof, etc.
b. Moving loads Live loads that move under their own power are said to be moving loads, such as a
series of trucks or railroad trains, etc.
(2) According to their manner of application, the loads can be divided as:
c Static loads The magnitudes and positions of static loads are not varying or varying very
slowly with time, so that the inertial forces could be neglected due to no effect of structural vibration.
d Dynamic loads Dynamic loads are loads of varying magnitudes and positions with time.
Dynamic loads give rise to vibration of structures. So the inertial forces have to be taken into consideration.
18 CHAPTER 1 INTRODUCTION
Blast loads caused by explosion, impact loads, centrifugal forces and the like are examples of dynamic
loads.
Some external effects such as support settlement, temperature change, manufacture errors, shrinkage
and creep of materials may cause internal forces and deformations in structures. These factors may be
called loads in generalized sense.
SUMMARY
In this chapter we have discussed four questions: (1) objective and learning method of structural
mechanics, (2) the analytical models of structures, (3) classification of framed structures and (4) the
classification of loads. They are all important questions throughout the book, but a primary understand will
satisfy the requirement of the introduction. The profound comprehension needs our gradually leaning the
other content of the book.
It is worthy of emphasize that the analytical models of the structures are the key point of the chapter
and also the starting point of computing structures in the latter study. Some special attention about the
principle of developing analytical models and the key points of simplification (especially, the simplification
of connecting joints and supports) ought to be paid to during the studying, so as to lay a solid foundation of
knowledge for analyzing the internal forces and deformations of structures.
Problems for Reflecting
1.1 What is the analytical or computing model of a structure? What is the relationship and difference
between the model and its realistic structure? Why we have to simplify actual structures into
analytical models?
1.2 Which two kinds of joints are the connections of a plane structure commonly simplified? What are
the conformation, the characteristics of restricting movement and restraint forces of the joints?
1.3 What kinds of supports for plane structures are usually simplified into? What are the conformation,
the characteristics of restricting movement and restraint reactions of the supports?
1.4 How many categories are there of commonly used framed structures?
CHAPTER 2
GEOMETRIC CONSTRUCTION ANALYSIS
OF STRUCTURES

The chapter is the essential requisition of structural analysis. Firstly, degrees of freedom are
introduced; then, geometric construction rules of plane stable structures with no redundant restraint
are discussed in the chapter. For plane structures, there are three construction rules recognized as the
key point of the chapter.
2.1 Purpose of Analyzing Geometric Construction of Structures, Stable and Unstable

Structural Systems
In order to withstand and transmit loads the geometric shape of a structure must not be changed under
loads. If the shape of a structural system is variable under loads the structural system cannot be used as a
structure.
It should be realized that all physical bodies deform when C D
subjected to loads; the deformations in most engineering structures
under service conditions are so small that their effect on the geometric
construction analysis of the structures can be neglected. This means A B
(a)
that all members composing of the structures are recognized as rigid
bodies. For plane structures, the rigid bodies are referred to as rigid
sheets. In the following discussion of the chapter, the all members of A B
structures are considered to be rigid bodies or rigid sheets; and for (b)
simplicity we call a structure as a system. For instance, framed
B
structures are named framed systems.
According to the shape and location of their members, framed
systems can be classified into two categories: A C
(c)
(1) Geometrically stable system Under the action of the
loads, the system still maintains its shape and remains its location if Fig.2.1 Stable system
the small deformations of the members are neglected as shown in
Fig.2.1 (a) and (b).
(2) Geometrically unstable system Under the action of the loads, the system will change its
19
20 Chapter 2 Geometric Construction Analysis of Structures
shape and its location if the small deformations of the members are neglected as shown in Fig.2.2 (a) and
(b).
Corresponding to geometrically stable and unstable systems, there are internally stable and unstable
systems as well. A structure is considered to be internally stable, or rigid, if it maintains its shape and
remains a rigid body when detached from the supports. Conversely, a structure is termed internally unstable
(or nonrigid) if it cannot maintain its shape and may undergo large displacements under small disturbances
when not supported externally. Fig.2.1 (c) shows a triangular structure conformed by three members and
detached from the supports. Because the triangle can maintain its shape and remains a rigid body although
it can only move in its own plane as an integer, it is referred to as an internally stable system, simply the
internally stable. However, the structure shown as in Fig2.2 (c), conformed to a parallelogram by four bars
and connected by pins each other, cannot maintain its shape and may undergo large displacements under
small disturbances when not supported externally. This kind of structures is thusly named as internally
unstable system, simply the internally unstable.
Purpose of analyzing geometric construction of structures is as following:
(1) To estimate whether or not a system is geometrically stable so as to determine whether or not the
system can be used as a structure;
(2) To discuss geometric construction rules of stable systems.
In addition above two points, analyzing geometric construction of structures are helpful to make their
static analysis due to the close relationship between their geometric construction analysis and mechanical
analysis.
C D C D
A B
A B B′ A B
(a) (b) (c)
Fig.2.2 Unstable system
2.2 The Concept of Degrees of Freedom and Restraints
In the analyzing geometric construction of structures, it is very feasible to consider one part of the
members (i.e., rigid bodies) or joints of a system as an object which possesses degrees of freedom, whereas
other part of the members or joints of the system as restraints which restricts the movement of the object.
The relationship of these tow parts are then analyzed and whether or not the system is stable or unstable
will be determined. Accordingly, the concept of degrees of freedom and restraints of a system is discussed
first of all.
2.2.1 The degrees of freedom
The degrees of freedom of a system are the numbers of independent movements (or coordinates)
which are required to locate the system fully. Obviously, a rigid body has three degrees of freedom in a
planar coordinate system (six degrees of freedom in a three dimensional coordinate system), e. g., the
position of member AB may be determined by three parameters xA , y A and θ A [Fig. 2-3 (b)].
The degrees of freedom of a joint
The movement of a point in a planar coordinate system could be decomposed to two translations in
any different directions, i.e., a point possesses two independent moving styles or two independent
coordinates are needed to locate its position in a planar coordinate system. So a joint (or a point) has two
degrees of freedom in a planar coordinate system (three degrees of freedom in a three dimensional
coordinate system). In Fig. 2-3 (a) the parameters xA , y A will locate joint A.
The degrees of freedom of a rigid body
The movement of a rigid body (or rigid sheet) in a planar coordinate system could be decomposed to
two translations in any different directions
and a rotation about some point in the y y
system, i.e., a rigid body possesses three
independent moving styles or three B
independent coordinates are needed to locate A A θA
its position in a planar coordinate system. yA yA
x x
Therefore, a rigid body has three degrees of xA
xA
freedom in a planar coordinate system (six (b)
(a)
degrees of freedom in a three dimensional
Fig.2.3 The degree of freedom
coordinate system), e.g., the position of
(a) the degree of freedom for a joint;
member AB may be determined by three (b) the degree of freedom for a rigid body
parameters xA , y A and θ A [Fig. 2-3 (b)].
2.2.2 Restraints
The devices or connections which can reduce the degrees of freedom of a system are defined as
restraints. The number of the degrees of freedom of a system reduced by the device or connection is named
the number of its restraints. There are two kinds of restraints, support restraints and connecting restraints
between rigid bodies.
1. Support restraints
(1) The roller support [Fig.2.4 (a)] can restrict the translation of joint A in the direction perpendicular
its moving surface but cannot prevent its translation along its moving surface and rotation about joint A, i.e.,
one roller support reduces one degree of freedom and is equivalent to one restraint.
(2) The hinged support [Fig.2.4 (b)] can restrict the translations of joint A in vertical and horizontal
directions but cannot prevent the rotation about joint A, i.e., one hinged support reduces two degrees of
freedom and is equivalent to two restraints.
(3) The fixed support [Fig.2.4 (c)] can restrict the translations of joint A in vertical and horizontal
directions and the rotation about joint A, i.e., one fixed support reduces three degrees of freedom and is
equivalent to three restraints.
A A A
(b) (c)
(a)
Fig.2.4 Support restraint

(a) roller support; (b) hinged support; (c) fixed support
2. Connecting restraints between rigid bodies

We will pay more attention to connecting restraints between two rigid bodies. One rigid body has
three degrees of freedom and two independent rigid bodies have six degrees of freedom in a planar
coordinate system. When connecting them together their degrees of freedom would be reduced. Now we
will discuss the equivalent restraints of a few kinds of commonly used connections.
A C A
A C
E F
B D B C
(a) (b) (c)
B
Fig.2.5 Connecting restraint

(a) restraint of a link; (b) restraints of a simple hinge; (c) restraints of a simple rigid joint
(1) Link [Fig.2.5 (a)] The two independent rigid bodies AB and DC have three relative degrees of
freedom with respect to each other, when they are connected by link EF which is a bar with hinges at its
ends, the relative translation between the two rigid bodies in the direction of the link is restricted, i.e., one
of the degrees of freedom between them is removed. Therefore, one link is equivalent to one restraint.
(2) Simple hinge (the hinge which connects two rigid bodies) [Fig.2.5 (b)] The two independent
rigid bodies AB and AC have three relative degrees of freedom with respect to each other, when they are
connected by hinge A only one relative degree of freedom between them is existed, i.e., the relative rotation
about A, and two translation degrees of freedom between them are removed. Therefore, one simple hinge is
equivalent to two restraints.
(3) Simple rigid joint (the joint which connects two rigid bodies rigidly) [Fig.2.5 (c)] The two
independent rigid bodies AB and AC have three relative degrees of freedom with respect to each other,
when they are connected by a rigid joint A there exist no relative degree of freedom between them in the
plane, i.e., three of the degrees of freedom between them are removed. Therefore, one simple rigid joint is
equivalent to three restraints.
By similar analysis, a hinge connecting three rigid bodies [Fig.2.6 (a)] makes the system remove four
degrees of freedom; the hinge is thusly equivalent to two simple hinges. While a rigid joint connecting
three rigid bodies [Fig.2.6 (b)] set the system remove six degrees of freedom; the rigid joint is then
equivalent to two simple rigid joints. Therefore, if a hinge or a rigid joint is used to connect more than two
rigid bodies, it is called a multiple hinge or multiple rigid joint. A multiple hinge connecting n bars is
equivalent to n-1 simple hinges or 2(n − 1) restraints, while a multiple rigid joint connecting n bars is
equivalent to n-1 simple rigid joints or 3( n − 1) restraints.
2.2.3 Restraint substitution and virtual hinges
Foregoing discussion tells us

that one hinge is equivalent to two A
B
restraints and two links are also D
equivalent to two restraints;
(b)
whereas restraints can be substitute. C (a)
In order to extend the concept of Fig.2.6 Multiple joint
restraint substation the concept of (a) multiple hinge; (b) multiple rigid joint
virtual or instantaneous hinge are
introduced.
Two noncollinear links which connects two rigid bodies are equivalent to one hinge. If two
noncollinear links AB and AC intersect at point A which lies on the rigid bodyⅠ[Fig.2.7 (a)], the point A
will be a actual simple hinge. If the intersectional point A does not lie on the two rigid bodies, i.e., the
intersecting point lies outside of the two rigid bodies, the intersection is referred to as a virtual or
instantaneous hinge, which has a variable position.
Two links AB and AC connecting two rigid bodiesⅠandⅡshown as in Fig.2.7 (b) intersect at point E
which lies outside of the two rigid bodies. Because the two rigid bodies are connected by two links the two
of their relative degrees of freedom are removed, but their relative rotation still exist. Where is their
instantaneous rotational center? Obviously it is the intersectional point E about which the two rigid bodies
are able to rotate relatively and instantaneously and soon the point E will move to a new position E′ .
E
E′
Ⅰ Ⅰ
A B D
B C A
Ⅱ C
Ⅱ
(a) (b)
Fig.2.7 Actual hinge and virtual hinge

(a) actual hinge; (b) virtual hinge
2.2.4 Necessary restraints and redundant restraints
In a framed system, the restraints which can restrict the degrees of freedom of the system are named
as necessary restraints; whereas the restraints which are not able to restrict the degrees of freedom of the
system are referred to as redundant restraints.
A A
1 2
1 2
A 3
(a) (b) (c)
Fig.2.8 Connection of a joint

(a) a free joint; (b) two links attached to a joint; (c) three links attached to a joint
A free joint A shown as in Fig.2.8 (a) has two degrees of freedom; if it were attached to the ground
with two noncollinear links 1 and 2 shown as in Fig.2.8 (b), its position would be determined and its two
degrees of freedom would be restricted. Therefore, the link 1 and 2 are all necessary restraints in order to
locate the position of the joint. If a link 3 were added to attach the joint with the ground shown as in Fig.2.8
(c) the link 3 would be redundant (you can also recognize any one of the three links as a redundant
restraint).
2.3 Geometric Construction Rules of Planar Stable Framed Systems without Redundant Restrains 25
A free member AB shown as in Fig.2.9 (a) has three degrees of freedom; if it were attached to the
ground with three noncollinear links 1 and 2 and 3 shown as in Fig.2.9 (b), its position would be
1A B
A B
2 3
4
(a)
(c)
1A B 1 A B
2 3 2 3
(b)
(d)
Fig.2.9 Connection of a member

(a) a free member; (b) three links attached to a member;
(c) four links attached to a member; (d) three links attached to a member
determined and its three degrees of freedom would be restricted. Actually, it is a simply supported stable
beam. Therefore, the link 1, 2 and 3 are all necessary restraints in order to locate the position of the member
AB. If a link 4 were added to attach the member with the ground as shown in Fig.2.9 (c) the link 4 would be
redundant (you can recognize any one of the three vertical links as a redundant restraint as well). It should
be noted that in Fig.2.9 (c) you cannot recognize the horizontal link 1 as a redundant restraint, and it is a
real necessary restraint. If the three links connecting the member AB intersected to one point as shown in
Fig.2.9 (d) only one link of link 1 and 3 would be necessary, the other one would be redundant, while the
vertical link 2 would be necessary restraint. It seems unclear for a system which restraints are necessary or
redundant restraints. In fact, the stability of a system depends on not only the number of its restraints but
also the arrangement of its restraints.
2.3 Geometric Construction Rules of Planar Stable Framed Systems without

Redundant Restraints
There are three elementary geometric construction rules for planar stable framed systems without
redundant restraints. Now we will discuss them through examples. It is worth of mentioning in anticipation
that a hinged triangle [Fig.2.1 (c)] is the simplest stable form of framed system. It is the basic stable shape
from which the general rules for stability may be developed.
2.3.1 The connection between a point and a rigid body
There are two degrees of freedom for a free point A [Fig.2.10 (a)] in planar coordinate system; when
the point is connected with a rigid body by a single link AB only the motion along the direction of the link
is restricted. So the point still has a degree of freedom in planar coordinate system. If another link AC,
which is noncollinear with link AB, is used to connect the point [Fig.2.10 (b)], all of the degrees of freedom
A A A
B C C B C
B
(a) (b) (c)
Fig.2.10 Connection between a point and a rigid body

(a) a link connecting a point and a rigid body;
(b) two links connecting a point and a rigid body;
(c) a point attached to the ground by two links
of the point is restricted; an internally stable system without redundant restraints is then composed by the
point A, the two links AB and AC and the rigid body.
Thereby, the following rule will be obtained:
Rule 1: A point and a rigid body connected by two noncollinear links form an internally stable system
without redundant restraint.
The rule can be reworded by the construction of binary system:
If a binary system is attached to a rigid body to form a new system the new system is still a stable
system without redundant restraints.
Here, the binary system indicates the construction of a new joint (connection) which is formed by two
noncollinear links (or members) connected each other. The attachment of binary systems will not change
the initial degrees of freedom of a system.
Lastly, it should be mentioned that in the geometric A C E
construction analysis the foundation of a structure or the
ground can be also regarded as a rigid body [Fig.2.10 (c)],
thus rule 1 will be a standard model to fix a point to the F
B D
ground or foundation. And it is helpful to note that a hinged
support is equivalent to two noncollinear links and a roller Fig.2.11 A framed system
support is equivalent to one link.
Example 2-1
Analyze the geometric construction of the framed system shown as in Fig.2.11.
Solution
According to rule 1, (1) starting from stationary points A and B, use two noncollinear links AC and BC
to fix point C; (2) beginning at stationary points B and C, add a binary system to fix point D; reincrease a
binary system to fix point E; again starting from stationary points D and E, reincrease a binary system to fix
point F.
Therefore, the integral system is a stable system without redundant restraint.
2.3.2 The connection of two rigid bodies in the same plane
In planar coordinate system，if rigid bodies AB and BC are connected by one hinge the relative
rotation between them will still exist [Fig.2.12 (a)]; If a link AC, which will not intersect with hinge B, is
used to connect rigid bodies AB and BC, the degrees of freedom of the new system are restricted and the
system will become an internally stable system without redundant restraints [Fig.2.12 (b)].
A A Ⅰ
C
1 2 3
B C B C Ⅱ
(a) (b) (c) (d)
Fig.2.12 Connection between two rigid bodies

(a) one hinge connecting two rigid bodies; (b) one hinge and one link connecting two rigid bodies;
(c) and (d) three links connecting two rigid bodies
Thereby, the following rule will be achieved:

Rule 2:
Two rigid bodies, connected by one hinge and one link that do not cross the hinge, form an internally
stable system with no redundant restraint.
By the concept that one hinge is equivalent to two links and the concept of virtual or instantaneous
hinge, rule 2 can be reworded in another way:
Two rigid bodies connected by three links, which are nonparallel and nonconcurrent, will form an
internally stable system with no redundant restraint [Fig.2.12 (d)].
When the ground is regarded as a rigid body [Fig.2.12 (c)], rule 2 will be a standard model to fix a
rigid body to the ground or foundation.
Example 2-2
Analyze the geometric construction of the system shown as in Fig.2.13.
Solution
28 Chapter 2 Geometric Construction Analysis Of Structures
Firstly, assume rigid body AB to be a moveable
object, it is connected with the foundation by a fixed
support; so AB and the foundation are fixed together A B C D
and they form a new rigid body with no redundant
restraint.
Fig.2.13 The system of example 2-2
Secondly, regard rigid body BC as a moveable
object, AB and the foundation as a new rigid body,
by rule 2, hinge B and roller C (a roller is equivalent to a link) are used to connect the two rigid bodies and
the roller C does not cross hinge B. Therefore, the system is stable with no redundant restraint.
2.3.3 The connection of three rigid bodies in the same plane
In nature, a hinged triangle is the simplest internally stable system with no redundant restraints
[Fig.2.14 (b)]. Based on this system rule 3 will be acquired [Fig.2.14 (a)]:
(Ⅰ,Ⅱ) (Ⅰ,Ⅱ)
A A Ⅱ
Ⅰ Ⅱ Ⅰ
(Ⅰ,Ⅲ) B C (Ⅱ,Ⅲ) (Ⅰ,Ⅲ) B C (Ⅱ,Ⅲ)

Ⅲ Ⅲ
(a) (b)
Fig.2.14 Connection of three rigid bodies
(a) connection of three rigid bodies; (b) two rigid bodies attached to the foundation
Rule 3:
Three rigid bodies (I, II and III) joined pairwise by hinges, provided that the three hinges (A, B and C)
do not lie on the same straight line, form an internally stable system with no redundant restraint.
When the ground is regarded as a rigid body [Fig.2.14 (b)], rule 3 will be a standard model to fix two
rigid bodies to the ground or foundation.
It is worthwhile to emphasize that all the three rules are only different explanations of the geometric
construction of the identical basic hinged triangle [Fig. 2-1 (c)] from which any one of the rules developed
formerly may be derived.
Example 2-3
Analyze the geometric construction of the system shown as in Fig.2.15 (a), (b) and (c).
Solution
Fig.2.15 (a): According to rule 3, rigid bodies ADC, BEC and the foundation which is regarded as the
third rigid body are connected pairwise by hinges A, B and C to form a stable system with no redundant
restraint.
Fig.2.15 (b): AFD is a rigid body, then a binary system is added to AFD to fix point C and to form a
larger rigid bodyⅠ. Similarly, BGEC can be composed to another larger rigid body II. The foundation is
regarded as rigid body Ⅲ. Upon that, three rigid bodiesⅠ, II and Ⅲ are connected pairwise by hinges A,
B and C to form a stable system with no redundant restraint.
Fig.2.15 (c): The system can be recognized that link FG is attached to the system shown in Fig.2.15
(b). Accordingly, the system shown as in Fig.2.15 (c) is a stable system with one redundant restraint.
(Ⅰ,Ⅱ) (Ⅰ,Ⅱ)
D E D E C E
C D
Ⅰ C Ⅱ
Ⅰ Ⅱ F G
F G
(Ⅰ,Ⅲ) A B (Ⅱ,Ⅲ) (Ⅰ,Ⅲ) A B (Ⅱ,Ⅲ) A B

Ⅲ Ⅲ
(a) (b) (c)
Fig.2.15 The systems of example 2-3
2.3.4 Instantaneously unstable system
In the construction of framed system, although the number of restraints is sufficient or even more than
sufficient, it is still unclear whether the system is stable or unstable, because the stability of a system
depends on not only the number but also the arrangement of restraints. If the arrangement of restraints of a
system does not satisfy the requirement of geometric construction rules as discussed in previous
subsections, the system will not remain stable.
In Fig.2.16 (a), the ground and joint A are connected by two links lying on the same straight line. The
system is unstable, because link AB and AC can not prevent the infinitesimal normal displacement of joint
A. However, when the system undergoes an infinitesimal displacement, A′ is the new position of joint A,
the two links will not lie on the same line any more. In this case, the system is termed an instantaneously
unstable system.
In Fig.2.16 (b), three links are intersected at point O, an infinitesimal rotation of rigid bodyⅠaround
rigid bodyⅡwill exist instantaneously. After the infinitesimal rotation happened the three links will not
intersect at one point and the system will not change again.
A Ⅰ
B C
A′
(a) Ⅱ
(b)
Fig.2.16 Instantaneously unstable system
The system existing infinitesimal instantaneous displacement is referred to as an instantaneously

unstable system. Although the displacement of an instantaneously unstable system is an infinitesimal value,
the internal forces of its members are infinitely large under the action of the loads which have finite values.
Therefore, instantaneously unstable system can not be used in practice.
Example 2-4
Analyze the geometric
construction of the system shown in
B C D E
Fig.2.17 (a), and (b). A
Solution (a)
Fig.2.17 (a): Rigid body AB is
fixed with the foundation to form a C D E
larger rigid body; then the new rigid B
A
body is connected with rigid body (b)
CDE by a hinged support D and a
Fig.2.17 The systems of example 2-4
link BC, but BC cross the hinge D.
Thereby, the system does not meet
rule 2. It is an instantaneously
unstable one.
Fig.2.17 (b): Analyzing similarly as Fig.2.17 (a), the stable portion composed by AB and the
foundation is connected with rigid body CDE by a hinged support D and a link BC, but BC does not cross
the hinge D. According to rule 2 the system is stable one with no redundant restraint.
2.4 Illustration of Geometric Construction Analysis 31
2.4 Illustration of Geometric Construction Analysis
Example 2-5
Analyze the geometric construction of the systems shown as in Fig.2.18 (a), and (b).
Solution
C E
A D D
G
C E F G
H A B
B F
(a) (b)
Fig.2.18 Systems of example 2-5
Fig.2.18 (a): The hinged triangle CGH is the basic stable portion (the hinged triangle GDE, or GHE,
HEF can be also regarded as a basic stable portion). By using rule 1, adding binary system from triangle
CGH, joints E, D, F, B and A are connected in the mentioned order. Consequently the portion becomes
internally stable with no redundant restraint. Afterward, connect the stable portion with the ground with one
hinged support and one roller support (equivalent one link, by rule 2) to form a stable system with no
redundant restraint.
Fig.2.18 (b): The hinged triangle AFD can be regarded as a basic rigid body. By using rule 1, then add
a binary system to AFD to form a larger rigid body AFCD. Similarly, another larger rigid body BGCE can
be obtained. And then, connect the two rigid bodies AFCD and BGCE with one hinge and one link (rule 2)
to form an internally stable system with no redundant restraint. Finally, attach the system to the ground with
a hinged support and one roller support which does not cross the hinged support. Therefore, the integral
system is stable with no redundant restraint.
Example 2-6
Analyze the geometric construction of the systems shown as in Fig.2.19 (a), and (b).
Solution
1. Fig.2.19 (a): Rigid bodies AC, BC and the foundation which is regarded as the third rigid body
are connected pairwise by hinge C and virtual hinges D and E, and the three hinges do not lie on the same
straight line. By rule 3, the system is a stable one with no redundant restraint.
2. Fig.2.19 (b): Although the rigid bodies AD and BE are angle bars, their ends are connected by
hinges. So AD and BE can be recognized as two links. The two links and roller support C (the roller support
is equivalent to one link) are used to connect rigid body CDE and the ground. However, the three links are
concurrent. Thereupon, the system is an instantaneously unstable system.
O
D E
Ⅰ C Ⅱ D E
A B
Ⅰ
2 3
2
1 4 3
A C B
Ⅲ Ⅱ
(a) (b)
Fig.2.19 Systems of example 2-6
2.5 The Relationship between Static Determinacy and Geometric Construction of

Structures
Another important function of geometric construction analysis is that the analysis can be used to
determinate the static determinacy or indeterminacy of a structure by analyzing whether or not a system has
redundant restraints.
P
P A C B
A B
P
P XA
XA
YA YB YC
YA YB
(a) (b)
Fig.2. 20 Structures
(a) statically determinate structure; (b) statically indeterminate structure
A stable structure is considered to be statically determinate if all its support reactions and internal
forces can be determined by solving the equations of equilibrium. For example, the simply supported beam
shown in Fig.2.20 (a) has one hinged support and one roller support, so it is a stable system with no
redundant restraint; obviously, the two supports will yield three reactions to the beam. Because the action
lines of the three reactions are not concurrent at one same point, the three equations of equilibrium
Summary 33
(∑ X = 0, ∑ Y = 0 and ∑ M = 0 ) of a coplanar force system can be used to solve the three reactions.
Once the reactions are solved all internal forces of the beam will be determined.
Conversely, if all support reactions and internal forces of a stable structure cannot be determined by
solving the equations of equilibrium the structure is recognized to be statically indeterminate. For the
structure shown in Fig.2.20 (b), the beam is attached to the foundation by one hinged support and two roller
supports, and the supports will develop four reactions. It is clear that the number of reactions is in excess of
those of equilibrium equations of a coplanar force system.
Thereby, statically determinate structures are stable with no redundant restraint systems whose
mechanical characteristics are that all of their support reactions and internal forces can be uniquely
determined by their static equilibrium equations; whereas statically indeterminate structures are stable
systems but with redundant restraints whose mechanical characteristics are that all of their support
reactions and internal forces cannot be uniquely determined by their static equilibrium equations.
SUMMARY
1. The main purpose of geometric construction analysis of a system is to determinate whether or not
the system is stable or unstable so as to make a decision that the system whether or not can be utilized as a
structure. Analyzing the geometric construction rules of a stable system with no redundant restraint will
help us to select correct method and procedure of static analysis, the idea will be frequently referred to in
the consequent chapters.
2. There are three geometric construction rules: (1) one point and one rigid body are connected with
two noncollinear links; (2) two rigid bodies are connected with one hinge and one link which does not cross
the hinge or with three nonparallel and nonconcurrent links; (3) three rigid bodies are connected pairwise
with hinges.
The essential of the three rules is a triangle rule, i.e., when the three side length of a triangle is
determined its geometric configuration is uniquely definite. Understanding the three rules is not difficult.
The main attention of the chapter should be paid to the sound application of the rules to analyze a variety of
framed structures. The difficulty of a novice is how to begin with analyzing a structure. Obviously, doing
some amounts of exercise will overcome the difficulty.
While using the three rules to analyze a system, their strictness should be pay attention to: distinguish
restricted objects and their restraints which play a role to restrict them; the number of the restraints and
their arrangement whether or not meet the requirement of geometric construction rules. On the other hand,
flexibility should be also paid when analyzing a system, such as the relation of substitution between
restricted objects and their restraints, the concept of virtual or instantaneous hinge and the like, and not
confused by the variation of configuration of the system.
3. Make a sound understanding of relationship between geometric construction and static
characteristics:
A stable with no redundant restraint system means that it is a statically determinate structure.
A stable with redundant restraints system means that it is a statically indeterminate structure.
An unstable (include instantaneously unstable) system means that it cannot be used as a structure.
2.1 Whether the instability of an unstable system can be changed by adding or discarding binary system
in turn from it? Contrariwise, whether an unstable system can be changed to a stable one by adding or
discarding binary system in turn from it?
2.2 What kind of relationship is there among the three rules mentioned in the chapter? Why it can be said
that their essential is the same?
2.3 What is the instantaneous unstable system? Why instantaneous unstable system cannot be used to
engineering structure?
2.4 What kind of characteristic of geometric construction is there for a statically determinate structure?
What kind of characteristic of static analysis is there for a statically determinate structure?
2.5 What is the difference between the geometric construction and characteristics of static analysis of a
statically indeterminate structure and those of statically determinate one?
Problems for Solution
Analyze the geometric construction of following systems.
problem 2-1
problem 2-2
problem 2-3
problem 2-4 problem 2-5 problem 2-6
problem 2-7 problem 2-8 problem 2-9
problem 2-10
problem 2-11
problem 2-12 problem 2-13



CHAPTER 3
STATICALLY DETERMINATE BEAMS

The objective of this chapter is to present the analysis of support reactions and internal forces
that may develop in beams under the action of coplanar systems of external forces and the
construction of internal force diagrams. The analysis of single span beams is the foundation of the
analysis of multispan beams; the content pertinent to single span beams should be soundly mastered.
Pay attention to the characteristics of constructing bending moment diagrams of straight members,
which are plotted by using the method of superposition segment by segment.
3.1 The Analysis of Single Span Beams
In engineering practice, single span beams have a variety of application. The analysis pertaining to
single span beams is the foundation of statically determinate multispan beams, frames and so forth. The
commonly used single span beams are the following three kinds: (1) simply supported beams or simple
beams [Fig.3.1 (a)], (2) cantilever beams [Fig.3.1 (b)] and (3) overhanging beams [Fig.3.1 (c)].
(a) (b) (c)
Fig.3.1 Single span beams

(a) simple beam; (b) cantilever beam; (c) overhanging beam
The chapter will review and discuss the calculation of internal forces and construction of their
diagrams of statically determinate single span beams.
3.1.1 Internal forces and their sign convention
For a planar member, it is generally subjected to shear forces and bending moments as well as axial
forces under the action of external loads.
The internal force component oriented in the direction of the centroidal axis of a beam at the section
under consideration is referred to as the axial force. Positive axial forces make the beam in tensile state
while negative in compressing state [Fig.3.2 (a)]
The internal force component oriented in the direction perpendicular to the centroidal axis of a beam
38
P
B B 10kN
10kN
3m
a
q = 1kN/m
C
Pa
C 0.6m
a
6m
A
A
(e)
(f)
Problem 3-1
q
B M
A A B
C C D
M diagram M diagram
Q diagram Q diagram
(a) (b)
A B
C D A B C
M diagram M diagram
Q diagram
Q diagram
(c) (d)
problem 3-2 (contd)
70 Chapter 3 Statically Determinate Beams
Q diagram M diagram Q diagram M diagram

B B
A A
(e) (f)
Problem 3-2
q q
ql2 ql2 ql2 ql2
8 8 8
A C B A C B 8
l l l l
2 2 2 2
(a) (b)
P
q
ql2 ql2
Pl Pl
8 C B 8 4 B 4
A A C
l l l l
2 2 2 2
(c) (d)
P P
Pl Pl Pl Pl
4 B 4 4 4
A C A C B
l l l l
2 2 2 2
(e) (f)
problem 3-3 (contd)

10kN 3kN
3kN/m 2kN/m
A B
A B
D E
D C E C
2m 2m 2m 2m 2m 2m 2m 2m
(g) (h)
4kN
2kN/m 2kN
A 2kN/m
D E
C B A D C B
2m 4m 4m 2m 2m 2m 2m
(i) (j)
Problem 3-3
q q
B B
90o
3m
3m
C C
D D
A A
1m 1m 2m 1m 1m 2m
(a) (b)
Problem 3-4
20kN 10kN 2kN/m

A B C
(a) F
D E
3m 3m 1.5m 2m 2.5m 2m 4m
10kN 5kN/m 20kN ⋅ m 10kN

(b) E
A B C D
2m 2m 2m 2m 2m 2m 2m
Problem 3-5
P
B C D E F
A
l l l l l l
2 2 2 2
Problem 3-6
P
A B C D E F
a a a a a a
2 2 2 2 2
Problem 3-7
at the section under consideration is called the shear force (or, simply, shears). As shown in Fig.3.2 (b), the
shears are considered to be positive when they tend to make the portion of the member on the left of the
section rotate clockwise and vice versa.
The internal couple which is the moment about (the centroid of the cross section of a beam at) the
section under consideration of all the internal forces is termed the bending moment. The positive bending
moment is shown in Fig.3.2 (c). The bending moments are considered to be positive when they tend to
bend a horizontal beam concave upward, causing compression in the upper fibers and tension in the lower
fibers of the beam at the section and vice versa.
+ +
N N +
Q Q M M
N − N − −
Q Q M M
(a) (b) (c)

Fig.3.2 Sign convention of internal forces
(a) sign convention of axial force; (b) sign convention of shear force; (c) sign convention of bending
3.1.2 The method of computing internal forces— method of sections
The method of sections is an elementary method to determinate the internal force components in a
member. The method means to pass an imaginary section through the member and cut the member into two
parts and then consider the free body of the either side of the section. We may obtain the three internal force
components at the section by equilibrium equations of the free body.
Now we will discuss the method of sections by concrete examples.
Example 3-1
Fig.3.3 (a) shows a simple beam, determine the internal forces at sections C , D1 and D2 .
Solution
1. Before determining the internal forces the reactions can be obtained by considering the
equilibrium conditions of the free body of the entire beam [see Fig.3.3 (a)].
∑ X = 0 ， X − 10 = 0 ， X = 10kN(→)
A A
∑ M = 0 ， 2 × 2 ×1 + 10 × 0.2 − Y × 8 = 0 ，
A B YB = 0.25kN(↑)
∑ M = 0 ， −2 × 2 × 7 − 10 × 0.2 + Y × 8 = 0 ，
B A YA = 3.75 kN(↑)
Check: ∑ Y = 0 ， 2 × 2 − 3.75 − 0.25 = 0
2. The internal forces at section C

An imaginary section is passed through section C, cutting the beam into two portions, AC and CB. The
portion AC shown in Fig.3.3 (c), which is to the left side of the section, is used here to compute the internal
forces. By utilizing the three equilibrium conditions, the internal forces can be determined.
2kN/m 10kN
e = 0.2m
A B
C D1 D2
2m 4m 2m
(a)
10kN 2kN/m
2kN/m MC
A e = 0.2m 10kN
B NC
XA A C
C D1 D2 QC
YB 3.75kN
YA 2m 4m 2m
(c)
(b)
2kN/m QD1 10kN

A M D1
N D1 N D1 B
10kN C D1 M D1 D1 D2
QD1
3.75kN 0.25kN
(d)
(f)
2kN/m 10kN QD 2
A D1 D2 B
ND2 ND2
10kN C M D2 M D 2D2
QD 2
3.75kN
(e) (g) 0.25kN
Fig.3.3 Diagrams of example 3-1
∑ X = 0, 10 + NC = 0 , NC = −10kN
∑Y = 0 , 3.75 − 2 × 2 − QC = 0 , QC = −0.25kN
∑M = 0 , C 3.75 × 2 − 2 × 2 × 1 − M C = 0 , M C = 3.5kN ⋅ m (tension in the lower fibers)
The negative N C and QC indicate that their real sense are opposite to those shown on the free
body, i.e., axial force is a pressure. The shear is really negative. The sense of M C is positive means that
its actual direction is the same as that shown on the free body, that is, M C causes tension in the lower
fibers of the beam at section C.
3. The internal forces at section D1
An imaginary section is passed through section D1 , cutting the beam into two portions, AD1 and
D1 B . The portion AD1 shown in Fig.3.3 (d), which is to the left side of the section, is used here to
compute the internal forces. By utilizing the three equilibrium conditions, the internal forces can be
determined.
∑ X = 0, 10 + N D1 = 0 ， N D1 = −10kN
∑Y = 0 , −3.75 + 2 × 2 + QD1 = 0 ， QD1 = −0.25kN
∑M D = 0 , 3.75 × 6 − 2 × 2 × 5 − M D1 = 0 ， M D1 = 2.5kN ⋅ m ( tension in the lower fibers)
The meanings of the internal forces are the similar as the above.
An imaginary section is passed through section D2 , cutting the beam into two portions, AD2 and
D2 B . The portion AD2 shown in Fig.3.3 (e), which is to the left side of the section, is used here to
compute the internal forces. By utilizing the three equilibrium conditions, the internal forces can be
determined.
∑ X = 0， 10 − 10 + N D 2 = 0 ， N D 2 = 0
∑Y = 0 , −3.75 + 2 × 2 + QD 2 = 0 ， QD 2 = −0.25kN
∑M D2 = 0 , 3.75 × 6 − 2 × 2 × 5 − M D 2 = 0 ， M D 2 = 0.5kN ⋅ m ( tension in the lower fibers)
The meanings of the internal forces are similar as the foregoing.
The selection of the free body can be the either side of the section under consideration when use the
method of sections. For instance, in above example we can select the left side of the section under
consideration or right side of the section as well. Although either of the two sides of the beam can be
employed for computing internal forces, we should select the side that will require the least amount of
computational effort, such as the side that does not have any reactions acting on it or that has the least
number of external loads and reactions applied to it. Based on this argument, select segment D1 B [Fig.3.3
(f)] and segment D2 B [Fig.3.3 (g)] can more easily find internal forces at sections Dl and D2
respectively. It is recommended for readers to check them.
The procedure for determining internal forces at a specified location on a beam can be summarized as
follows:
(1) Pass an imaginary section perpendicular to the centroidal axis of the beam at the section where
the internal forces are desired, thereby cutting the beam into two portions. Then select any one of the two
portions (commonly the simplest portion) as the free body.

(2) Draw the free-body diagram upon which reactions and applied loads and desired internal forces
should be actually imposed.
(3) Apply the equilibrium conditions of the free body and determinate the desired three internal
forces.
Obviously, in order to establish the correct equilibrium equations so as to obtain correct internal forces,
the key step is presenting the free-body diagram correctly. Following consideration has to take when
present free-body diagrams:
(1) The free body should be isolated completely with its restraints and all the restraints should be
substituted by their corresponding constraint forces.
(2) The constraint forces should meet the characteristics of restraints. For example, while remove one
roller support, one hinged support and one fixed support, one support reaction, two support reactions and
three support reactions should be imposed on the free body respectively; while cut one link, one hinge and
one rigid joint, a axial force, a axial force and a shear, and a axial force and a shear and a couple should be
exerted on the free body respectively.
(3) All the forces acting on the free body should be indicated actually. Do not leave out the forces
directly exerting on it and do not superimpose the forces not directly exerting on it. The forces exerting on
a free body can be classed as two groups. One is the externally applied loads and the other is internal forces
exerted on by the other portion corresponding to the free body.
(4) The presentation of a free body should be clear and accurate. Generally, given forces ought to
indicate according to their actual locations and magnitudes and directions, whereas unknown forces should
be indicated at their locations in their positive directions. Based on this regulation, if consequent unknowns
are positive they are actually positive; if some of the unknowns are negative they are actually negative,
their direction indication are just in the opposite direction. Therefore, you can avoid the ambiguity caused
by directions of unknown forces.
Above mentioned procedure of method of sections is a normal progress to calculate internal forces. Its
counterpart method, which does not need free-body diagram and may be developed by the definition of
internal forces, can be stated as following:
Axial force: The internal axial force on any section of a beam is equal in magnitude but opposite in
direction to the algebraic sum (resultant) of the components in the direction parallel to the axis of the beam
of all the external loads and support reactions acting on either side of the section under consideration.
According to the sign convention adopted in the preceding paragraphs, if the portion of the beam to
the left side of the section is being used for computing the axial force, then the external forces acting to the
left are considered positive, whereas the external forces acting to the right are considered to be negative. If
the right portion is being used for analysis, then the external forces acting to the right are considered to be
positive and vice versa.
Shear: The shear on any section of a beam is equal in magnitude but opposite in direction to the
algebraic sum (resultant) of the components in the direction perpendicular to the axis of the beam of all the
external loads and support reactions acting on either side of the section under consideration.
If the left portion of the beam is being used for analysis, then the external forces acting upward are
considered positive, whereas the external forces acting downward are considered to be negative. If the right
portion has been selected for analysis, then the downward external forces are considered positive and vice
versa.
Bending moment: The bending moment on any section of a beam is equal in magnitude but opposite
in direction to the algebraic sum of the moments about the centroid of the cross section under consideration
of all the external loads and support reactions acting on either side of the section.
If the left portion is being used for analysis, then the clockwise moments are considered to be positive,
and the counterclockwise moments are considered negative. If the right portion has been selected for
analysis, then the counterclockwise moments are considered positive and vice versa.
Example 3-2
When the reactions of example 3-1 [Fig.3.3 (a)] are calculated try to determine the internal forces at
sections C , D1 and D2 again by employing the counterpart of method of sections.
Solution
1. The internal forces at section C

The portion AC (from A to C ), which is to the left side of the section, is used here to compute the
internal forces.
Axial force: Considering the external forces acting to the left to be positive, we write
NC = −10kN
Shear: Considering the external forces acting upward to be positive, we write
QC = 3.75 − 2 × 2 = −0.25kN
Bending moment: Considering the clockwise moment of external forces about C to be positive, we
write
M C = 3.75 × 2 − 2 × 2 ×1 = 3.5kN ⋅ m
The portion AD1 (from A to D1 ), which is to the left side of the section, is used here to compute the
internal forces.
N D1 = −10kN
QD1 = 3.75 − 2 × 2 = −0.25kN
Bending moment: Considering the clockwise moment of external forces about D1 to be positive, we
write
M D1 = 3.75 × 6 − 2 × 2 × 5 = 2.5kN ⋅ m
The portion AD2 (from A to D2 ), which is to the left side of the section, is used here to compute
the internal forces.
N D 2 = −10 + 10 = 0
QD 2 = 3.75 − 2 × 2 = −0.25kN
Bending moment: Considering the clockwise moment of external forces about D2 to be positive, we
write
M D 2 = 3.75 × 6 − 2 × 2 × 5 − 10 × 0.2 = 0.5kN ⋅ m
It is clear that in the above computations if we select right segment BD1 (from B to D1 ) and segment
BD2 (from B to D2 ) to compute the internal forces at section D1 and D2 respectively, the analysis
will be more easily. It is recommended for readers to check them.
3.1.3 Relationships between loads, shears and bending moments
The construction of internal force diagrams can be considerably expedited by using the basic
relationships that exist between the loads, the shears, and the bending moments.
1. Differential relationships
To derive these relationships, consider a beam subjected to an arbitrary loading, as shown in Fig.3.4
(a). All the external loads shown in this figure are assumed to be acting in their positive directions. As
indicated in this figure, the external distributed and concentrated loads acting downward (in the positive y
direction) are considered positive; the external couples acting clockwise are also considered to be positive
and vice versa. Next, we consider the equilibrium of a differential element of length dx , isolated from the
beam by passing imaginary sections at distances x and x + dx from the origin A, as shown in Fig.3.4
(a). The free-body diagram of the element is shown in Fig.3.4 (b), in which Q and M represent the shear
and bending moment, respectively, acting on the left face of the element (that is at distance x from the
origin A), dQ and dM denote the changes in shear and bending moment, respectively, over the distance
dx . As the distance dx is infinitesimally small, the distributed load q acting on the element can be
considered to be uniform of magnitude q (x ) . In order for the element to be in equilibrium, the forces and
couples acting on it must satisfy the two equations of equilibrium, ∑ Y = 0 and ∑ M 0 = 0 . The third
equilibrium equation, ∑ X = 0 , is automatically satisfied, since no horizontal forces are acting on the
element. Applying the equilibrium equations the following formulae can be obtained.
q P
m
A B
C D E F
(a)
q
P m
M +dM Q EL Q ER Q FL M FR
Q
x x x
M Q + dQ M EL E M ER M FL F Q FR
dx
y y y
(b) (c) (d)
(1) Distributed loads

Considering the equilibrium of an arbitrary free body of a differential element of length dx isolated
from the segment subjected to distributed loads of the beam, we write
dQ ⎫
∑ Y = 0, dx
= −q (a) ⎪
⎪
⎬ (3-1)
dM
∑M 0 = 0,
dx
=Q (b) ⎪
⎪⎭
Combine Eqs. (3-1) (a) and (b), we obtain
d 2M
= −q (3-2)
dx 2
Above differential relationships represent the following geometric meanings:
Eq. (3-1) (a) can be expressed as: Slope of shear diagram at a point equals to the intensity of
distributed load at that point but opposite sign.
Eq. (3-1) (b) can be stated as: Slope of bending moment diagram at a point equals to the shear at that
point.
Eq. (3-2) can be presented as: Curvature of bending moment diagram at a point equals to the intensity
of distributed load at that point but opposite sign.
Based of above differential relationships, the internal force diagrams located in the segments imposed
by distributed loads possess the following geometric characteristics:
c At the segment imposed by uniformly distributed loads (q=constant)
Q diagram is a linear function of x, the curve of which is an inclined straight line. If consider upward
direction as positive ordinate of Q and q acting downward in the segment, the Q curve will be straightly
inclined downward from the left to the right of the segment.
While M diagram is a quadratic function of x, the moment curve is a parabolic curve, which is concave
upward for a downward load q.
d At no loading segment (q=0)
Q is a constant and its diagram is a horizontal line.
While M diagram is a linear function of x, the moment curve is an inclined straight line, the inclined
sense of which is the same as the sign of Q.
(2) Concentrated loads
The relationships between the loads and shears derived thus far [Eqs. (3-1) through (3-2)] are not valid
at the points of application of concentrated loads, because at such a point the shear changes abruptly by an
amount equal to the magnitude of the concentrated load. To verify this relationship, we consider the
equilibrium of a differential element that is isolated from the beam of Fig.3.4 (a) by passing imaginary
sections at infinitesimal distances to the left and to the right of the point of application E of the
concentrated load P. The free-body diagram of this element is shown in Fig.3.4(c). Applying the
equilibrium equations we write
∑ Y = 0, QER = QEL − P (a ) ⎪⎫
⎬ (3-3)
∑ M = 0,
E M EL = M ER (b) ⎪⎭
Eq. (3-3) (a) can be expressed as: The magnitudes of shears at the two sides of the point of application
of concentrated load are unequal. The shear difference at the two sides is equal to the magnitude of P.
Therefore, the shear diagram will have abrupt change at the point of application of concentrated load P by
an amount equal to the magnitude of the concentrated load P.
Eq. (3-1) (b) can be stated as: The bending moments at the two sides of the point of application of
concentrated load are identical. Note that because of the abrupt change in the shear diagram at such a point,
there will be an abrupt change in the slope of the bending moment diagram at that point. That is, there are
cusps (angle points) on the moment diagram at the locations where concentrated loads are applied on.
When concentrated loads are downward the cusps are downward as well.
(3) Couples or concentrated moments
Although the relationships between the loads and shears derived thus far [Eq. (3-1) (a), Eq. (3-2) and
Eq. (3-3) (a)] are valid at the points of application of couples or concentrated moments, the relationships
between the shears and bending moments as given by Eqs. (3-1) (b) and (3-3) (b) are not valid at such
points, because at such a point the bending moment changes abruptly by an amount equal to the magnitude
of the moment or the couple. To derive this relationship, we consider the equilibrium of a differential
element that is isolated from the beam of Fig.3.4 (a) by passing imaginary sections at infinitesimal distance
to the left and to the right of the point F of application of the couple m. The free-body diagram of this
element is shown in Fig.3.4 (d). Applying the equilibrium equations we write
∑ Y = 0, QEL = QER (a ) ⎪⎫
⎬ (3-4)
∑ M = 0,
F M ER = M EL + m (b) ⎪⎭
Eq. (3-4) (a) can be expressed as: The magnitudes of shears at the two sides of the point of application
of concentrated moment m are equal; thusly the shear diagram will not change at the two sides.
Eq. (3-4) (b) can be stated as: The magnitudes of bending moments at the two sides of the point of
application of concentrated moment m are unequal. Their difference at the two sides is equal to the
magnitude of m. Thereby, the moment diagram will have abrupt change at the point of application of
concentrated moment m by an amount equal to the magnitude of the concentrated moment m, whereas the
slope of the moment diagram will not change since the constant value of shears at the two sides of the point,
i.e., the slopes at the two sides remain parallel.
2. Integral relationships
To determine the change in shear and bending moment between points A and B along the axis of the
member (see Fig.3.5), we integrate Eq. (3-1) from A to B to obtain the following integral relations between
loads and internal forces [Eq. (3-5)].
(a ) ⎫⎪
B
QB = QA − ∫ q ( x)dx
A
B ⎬ (3-5)
q ( x)
M B = M A + ∫ Qdx (b) ⎪
A ⎭ MB
These equations imply that: B
MA A
(1) The shear at end B is equal to the difference QB
QA
between the shear at end A and the resultant of load q
between A and B, or the change in shear between Fig.3.5 Segment AB of a member
sections A and B is equal to the area of the distributed
load diagram between sections A and B.
(2) The bending moment at end B is equal to the sum of that at end A and the area of the shear diagram
between A and B, or the change in bending moment between sections A and B is equal to the area of the
shear diagram between sections A and B.
If not only the distributed but also concentrated loads are exerted on segment AB, Eq. (3-5) may be
written as
QB = QA − ∫ q ( x)dx − ∑ P (a ) ⎪⎫
B
A
⎬ (3-6)
M B = M A + ∫ Qdx + ∑ m
B
(b) ⎪
A ⎭
Where ∑ P represents the algebraic sum of the vertical concentrated loads between points A and B,
and ∑ m is the sum of the concentrated moments exerted on segment AB. The downward P and clockwise
m are assumed to be positive in the equations.
3.1.4 The construction of internal force diagrams
An internal force diagram is the diagram on which the variation law along the location of member
sections of the internal force is clearly represented. On such diagrams, the abscissas represent the axial
lines of members, whereas the ordinates denote the internal force values of member sections (i.e., plot the
internal force values in the direction perpendicular to the axial lines of members). The regulations of
constructing internal force diagrams are as followings:
The ordinates of bending moment diagrams have to be plotted on the sides where the fibers of the
members are under tensile action, and no sign indication is necessary on bending moment diagrams. Shear
and axial force diagrams can be depicted on any sides of the members, but sign indication must be made
(generally, for horizontal members positive shears and axial forces are plotted on the top sides of the
members).
The basic differential relationships that exist between the loads and internal forces tell us that there is
direct relationship between the shape of internal forces and load distributions. In the segments imposed by
distributed loads or no load, the curves of internal forces are continuous; while at the points where loads
vary abruptly, internal force curves vary abruptly as well. Utilizing these features, we could depict internal
force diagrams very soon. The key points to construct them are stated as followings:
1. Select control sections
The control sections (or points) indicate the sections at which loads vary discontinuously, such as the
starting and ending points of distributed loads, the locations of concentrated forces and couples. The control
sections are the sections where internal force curves vary discontinuously as well.
2. Determinate the internal force values of control sections
The primary method to determinate internal forces is method of sections. See subsection 3.1.2.
3. Construct internal force diagrams segment by segment
Divide a member into several segments by control sections, and then consider the internal force values
of control sections as ordinates and lay off them based on corresponding abscissas, finally plot the internal
force diagrams by means of their curve properties dominated by the basic differential relationships that
exist between the loads and internal forces.
Now illustrate above statement through the following example.

Example 3-3
Depict the internal force diagrams for the simple beam shown in Fig.3.6 (a).
Solution
1. Find reactions
Applying the equilibrium conditions of entire beam, we obtain
∑ X = 0， X =0 A
∑ M = 0 ， 16 ×1 + 4 × 4 × 4 − Y × 8 = 0 ，
A B YB = 10kN(↑)
∑ M = 0 ， Y × 8 − 16 × 7 − 4 × 4 × 4 = 0 ，
B A YA = 22kN(↑)
∑ Y = 0 ， 22 − 16 − 4 × 4 + 10 = 0
2. Construct shear diagram
(1) Determinate the internal force values of control sections. The control sections are sections A, C,
D, E and B, at which loads vary discontinuously. The five sections divide the beam into four segments, AC,
CD, DE and EB. As no load is applied to the segments AC, CD and EB, the shear curves in these segments
are horizontal lines, which could be plotted just by one numerical value of their shears. As the uniformly
distributed loads are applied in segment DE the shear curve in this segment is an inclined straight line,
which needs two numerical values of the shears to be plotted.
QA = QCL = YA = 22kN
QCR = QD = YA − P = 22 − 16 = 6kN
QE = QB = −YB = −10kN
(2) Construct Q diagram
Take axial line AB as the abscissa axis; indicate the locations of control sections (A, C, D, E and B);
lay off the ordinates at points A and Cleft (just to the left of section C) by the magnitude of 22kN
upwardly, we obtain points A1 and C1 ; do so at points Cright (just to the right of section C) and D by
the magnitude of 6kN upwardly, we get points C2 and D1 ; do so at points E and B by the magnitude of
10kN downwardly, we obtain points E1 and B1 . Then connect points A1 and C1 by a horizontal line; do
so between points C2 and D1 , E1 and B1 respectively; and connect points D1 and E1 by a straight,
sloping line (downward to the right). Finally, indicate positive sign on upward side and negative sign on
downward side of the abscissa. The final shear diagram will be completed as shown in Fig.3.6 (b)
3. Construct moment diagram
(1) Determinate the magnitude of bending moments of control sections. The control sections are
section A, B, C, D and E as well. The magnitude of the bending moments is:
MA = 0
M C = 22 ×1 = 22kN ⋅ m (Tension in the lower fibers)
M D = 22 × 2 − 16 ×1 = 28kN ⋅ m (Tension in the lower fibers)
M E = 10 × 2 = 20kN ⋅ m (Tension in the lower fibers)
MB = 0
16kN 4kN/m
(a) XA = 0 A B
C D F E
YA = 22kN YB = 10kN
1m 1m 2m 2m 2m
(b)
22 22
A1 C1
Q (kN) C2 D1
6 6
G E B
A C D 10 10
1.5m
E1 B1
(c) A C D G F E B
A1 B1
M (kN ⋅ m)
20
22 E1
C1 28
D1 32.5 32
G1 F1
16kN
4kN/m
(d)
A
M max
G
1m 1m x
22kN Q=0

(2) Construct M diagram

On downward side of the abscissa axis, lay off the values of the bending moments of each control
sections(A, C, D, E and B) by their corresponding magnitude (0, 22, 28, 20, 0), whose corresponding points
are ( A1 , C1 , D1 , E1 and B1 ) shown as in Fig.3.6 (c).
In the segments to which no load is applied (i.e., segments AC, CD and EB), connecting A1C1 , C1 D1
and E1 B1 by straight sloping lines will finish the construction of bending moment diagram about these
segments of the beam.
In the segment DE to which uniformly distributed loads is applied, the bending moment diagram
should be a quadratic parabola. Three points are needed to determinate the parabola. Since we only have
two points ( D1 and E1 ), another additional point is necessary. We may take the value of bending moment
at the middle point F of segment DE or the maximum value M max of the bending moments at the point G
where the shear is equal to zero.
The two bending moment quantities are calculated as following:
M F = 22 × 4 − 16 × 3 − 4 × 2 × 1 = 32kN ⋅ m (Tension in the lower fibers)
Consider AG as a free body shown in Fig.3.6 (d); assume the maximum value M max of the bending
moments occurs at section G at a distance x from the starting point of distributed loads, applying method of
sections, we obtain
QG = 22 − 16 − qx = 0
22 − 16
x= = 1.5m
4
1.52
M max = 22 × 3.5 − 16 × 2.5 − 4 × = 32.5kN.m
2
Now we can lay off the ordinate at point F downwardly by the magnitude of 32kN ⋅ m , obtain point F1 ;
or do so at point G downwardly by the magnitude of 32.5kN ⋅ m , obtain point G1 .Then connecting the
three points D1 , F1 and E1 or points D1 , G1 and E1 by a parabola will accomplish the construction of
the bending moment diagram of the beam as shown in Fig.3.6 (c).
4. The check of the shape characteristics of internal force diagrams
By analyzing the load diagram, shear diagram and bending moment diagram shown in Fig.3.6 (a), (b)
and (c) respectively, shear and bending moment curves in segments AC, CD and EB, to which no load is
applied, are horizontal and inclined lines respectively. Since a downward concentrated force
of 16kN magnitude acts at section C, the shear diagram decreases abruptly from 22kN to 6kN at this section;
while bending moment curve is not derivable at the location of this section, that is, there are two slopes or
there is a downward cusp at the location of the section for the bending moment curve. In segment DE, the
beam is subjected to a downward uniformly distributed load of 4kN / m . Because the load intensity is
constant and positive in segment DE, the shear diagram in this segment is a straight line with negative slope;
whereas the bending moment diagram in this segment is a second-order upward-concave parabola with a
minus constant curvature of 4kN / m . Note that the bending moment curve should be derivable at points
D1 and E1 , i.e., there is only one slope or tangent line at each of the two points for the curve.
3.2 Construction of Bending Moment Diagram by Principle of Superposition for

Straight Members
The principle of superposition is extremely convenient in structural analysis. If a structure is subjected

to a variety of loads, the displacements in the structure vary linearly with the applied loads, that is, any
increment in displacement is proportional to the loads causing it, and if all deformations of a structure are
small enough so that the resulting displacement of the structure does not significantly affect the geometry
of the structure and hence do not change the initially acting property of the loads in the members. Under
such conditions, reactions, internal forces and displacements due to the loads can be obtained by utilizing
the principle of superposition.
The principle can be expressed as, if a structure is linearly elastic, the forces acting on the structure
may be separated or divided into some expediently individual forms and then the structure may be
separately analyzed for each individual forms. The final results can then be obtained by adding up the
individual results.
Now we will discuss how the principle of superposition be used to facilitate the construction of
bending moment diagrams for straight members.
3.2.1 Superposition method of bending moments for simple beams
Fig.3.7 (a) shows a simply supported beam subjected to not only the uniformly distributed loads in the
span but also the end couples. By principle of superposition, divide the loads applied on the beam into two
groups, one is the two end couples [Fig.3.7 (b)] and the other is uniformly distributed loads [Fig.3.7 (c)].
If the beam undergoes the end couples only, the moment diagram is simply a straight line as shown in
Fig.3.7 (b), it is denoted by M diagram; and if the beam is subjected to the uniformly distributed loads in
0
the span only, the bending moment diagram M ( x ) is a second-order parabola and may be plotted as
shown in Fig.3 7 (c). Then the final bending moment diagram of the beam may be obtained by the principle
0
of superposition. That is, depicting the M diagram first, then the ordinates of M ( x ) diagram being
superimposed on the M diagram will complete the final bending moment diagram of the beam as shown
in Fig.3.7 (d).
Note that it is the ordinates not the figures of the two diagrams which are superimposed together. The
ordinate M ( x) at any arbitrary section shown in Fig.3.7 (d) should meet the following formula.
3.2 Construction of Bending Moment Diagram by Principle of Superposition for Straight Members 53
M ( x) = M ( x) + M 0 ( x)
Where the ordinates of M ( x ) are perpendicular to the beam axis AB not the dashed line A′B′ .
0
3.2.2 Superposition method segment by segment
Now extend the superposition method discussed in above subsection to the construction of bending
moment for any arbitrary segment of a straight member. Fig.3.8 (a) shows a simply supported beam
subjected to the uniformly distributed loads only in segment AB. The bending moments M A at section A
and M B at section B are determined by the method of sections. Next we will discuss how to construct the
bending moment diagram of segment AB by utilizing principle of superposition.
(a) q
(a) MA q MB
A B
A B
MA MB MA q MB
A B (b)
(b)
A B
M
QA QB
q (c) MA q MB
(c)
A B
A B
M0
YA0 YB0
A B (d) A B
(d)
MA MA M M MB
M M MB
A′ B′
A′ B′
M0 M0
Fig.3.7 Construction of moment diagram Fig.3.8 Construction of moment diagram for arbitrary
by superposition method segment by superposition method
Take segment AB of the beam shown in Fig.3.8 (a) as an example and compare the free body diagram
of the segment [Fig.3.8 (b)] with the simply supported beam in Fig.3.8 (c). We may find that in these two
cases both the load q and the end couples are the same. By employing the equilibrium equations, we also
find YA0 = QA and YB0 = −QB . Since the stressed state of segment AB [Fig.3.8 (b)] is the same as that of
the simple beam [Fig.3.8 (c)], the bending moment diagrams of these two cases must be identical. Thereby,
the superposition method used for constructing bending moment diagram of the simple beam AB [Fig.3.8
(b)] can be employed to construct the bending moment diagram of the segment AB shown in Fig.3.8 (a).
Therefore, for a segment of a member, the moment diagram may be constructed as follows.
Firstly, determine the end moments M A at section A and M B at section B of the segment and take
the end couples M A and M B as ordinates (indicate on tension side of the segment), draw a straight
dashed line which represent the M diagram [Fig.3.8 (d)].
Secondly, take the dashed line as the base line, superimpose the ordinates (in the direction
0
perpendicular to AB) of the M diagram, we accomplish the final M diagram of segment AB [Fig.3.8 (d)].
Example 3-4
Construct the bending moment diagram for the overhanging beam shown in Fig.3.9 (a).
Solution
It is no necessary for constructing bending moment diagram to find reactions.

1. Determination of bending moments of control sections
The sections A, B and D are control sections at which loads vary discontinuously, which divide the
beam into two segments AB and BD. By the method of sections we obtain the bending moments
M A = −3kN ⋅ m (Tension in the upper fibers)
MD = 0
M B = −1× 2 × 1 = −2kN ⋅ m (Tension in the upper fibers)
2. Constructing bending moments diagram segment by segment
Take axial line AD as the abscissa axis; indicate the locations of control sections (A, B, and D); lay off
the ordinates at points A, B, and D on tension side of the beam by the magnitudes of M A , M B and
M D respectively, we obtain points A1 , B1 and D1 [Fig.3.8 (b)].
Connect points A1 , B1 and D1 with dashed lines, then take the dashed lines as base lines and
0
impose the ordinates of the corresponding bending moment diagram M ( x ) . The final bending moment
diagram is shown as in Fig.3.9 (b).
Note that since a concentrated load of 4kN is applied at the middle of the segment AB, the bending
0
moment diagram M ( x ) in segment AB should be that of a simple beam with the same span of AB and
subjected to the same external load of segment AB (i.e., a simple beam with a span of 5m and subjected to a
concentrated downward force of 4kN at the middle of the beam). While the bending moment diagram
M 0 ( x) in segment BD should be that of a simple beam with the same span of BD and subjected to the
same uniformly distributed load of segment BD (i.e., a simple beam with a span of 2m and subjected to
uniformly distributed load of 1kN/m in the whole span of the beam).
0
Knowing from strength of materials, the bending moment value M at the middle of segment AB
Pl 4 × 5
will be = = 5kN ⋅ m , the total value of the bending moment at section C will be
4 4
(3 + 2)
MC = − + 5 = 2.5kN ⋅ m (Tension in the lower fibers)
2
ql 2 1
0
The bending moment value M at the middle of segment BD will be = ×1× 22 = 0.5kN ⋅ m ,
8 8
the total value of the bending moment at the middle section E will be
(2 + 0)
MC = − + 0.5 = −0.5kN ⋅ m (Tension in the upper fibers)
2
The bending moment diagram in segment BD can be plotted by considering the segment as a
cantilever beam and directly drawing the moment diagram with no need of method of superposition as well.
4kN
3kN ⋅ m 1kN/m
(a) B
A C E D
2.5m 2.5m 2m
(b) A1
3
2 B1
M diagram (0.5)
(KN ⋅ m) C 5 D1
A B (0.5) D
E1
C1 2.5
Fig.3.9 Diagram of example
3.3 The Analysis of Simply Supported Inclined Beams
In the civil engineering construction, inclined beams shown as in Fig.3.10 (b) and (c), are commonly
used as the analyzing models of stairs [Fig.3.10 (a)].
The axial line of an inclined beam is straight slope line; its slope angle is denoted by θ . The loads
imposed to the inclined beams can be classified into two groups: (1) vertically distributed loads along the
axial lines of inclined beams, such as deadweight loads; (2) vertically distributed loads along the horizontal
lines (the projections of axial lines of inclined beams), such as live loads. In order to facilitate the analysis,
vertically distributed loads along the axial lines of inclined beams [Fig.3.10 (b)] are usually altered to
(b) (c) q
(a)
q′
l′
θ θ θ
l l
Fig.3.10 Simple inclined beam
vertically distributed loads along the horizontal projections of the beams [Fig.3.10 (c)]. The final analyzing
results under the action of vertically distributed loads along the axial lines of inclined beams are the same
as those under the action of vertically distributed loads along the horizontal projections of the beams. If the
distributed loads are uniformly distributed the alternative relationship of this two kinds loads will be
l q′
ql = q ' l ′ = q′ ∴ q=
cos θ cos θ
In which, q′ is vertically distributed loads per unit length along the axial lines of inclined beams, while q is
vertically distributed loads per unit length along the horizontal projections of the beams.
Next we will discuss the analytical nature, by an example, of statically determinate inclined beams
subjected to vertically uniformly distributed loads along their horizontal projections.
Example 3-5
Construct the internal force diagrams for the simply supported inclined beam shown in Fig.3.11 (a).
Solution
1. By employing equilibrium conditions of entire beam, the reactions will be

∑ X = 0, XA = 0
1
∑M = 0 A YB = ql (↑)
2
1
∑M B = 0, YA = ql (↑)
2
ql ql
Check: ∑Y = 0 , ql − − =0
2 2
(a) q
B B
(c) M diagram
1
YB = ql
2
C
XA = 0 A 1 2
θ A ql
8
1
YA = ql x
2
l B
(d) Q diagram ql
cos θ
2
qx sin θ ql
(b)
qx cos θ
2 ql
q sin θ
s A 2
qx cos θ
B
MC NC
C (e) N diagram
QC
A θ
ql r
sin θ 1 A
2 θ ql
2
ql x ql
cos θ sin θ
2 2
Fig.3.11 Diagram of example 3-5
2. Determinate the internal forces on an arbitrary section C by employing method of sections

An imaginary arbitrary section at a horizontal distance x from the origin A is passed through section C,
cutting the beam into two portions, AC and CB. The portion AC shown in Fig.3.11 (b), which is to the left
side of the section, is used here to compute the internal force components. Note that since the axial line of
the beam is slope straight line the action directions of axial force and shear are all inclined. When utilizing
the equilibrium conditions to evaluate internal force components, in order to avoid solving simultaneous
equations in the computations two projective equations, one of which is along the direction s tangent to the
axial line and the other of which is along the direction r normal to the axial line of the inclined beam,
should be employed to directly obtain the axial force N c and shearing force Qc . Therefore, both reaction
YA and resultant of uniformly distributed load qx ought to be decomposed along the direction s and r.
Thusly, we obtain
1 x 1
∑M C = 0,
2
qlx − qx − M C = 0,
2
M C = q(lx − x 2 )
2
(a)
1 1
∑ r = 0, − ql cos θ + qx cos θ + QC = 0,
2
QC = q(l − 2 x) cos θ
2
(b)
1 1
∑ S = 0,
2
ql sin θ − qx sin θ + NC = 0, NC = − q(l − 2 x) sin θ
2
(c)
3. Construct M, Q and N diagrams

(1) M diagram
Since all the forces are vertical, only the horizontal distances are involved in the bending moment
calculation, hence the bending moment at any arbitrary section C, denoted by M C , of the inclined beam is
the same as that of a simply supported horizontal beam with the same span and subjected to the same load.
The bending moment diagram is depicted as shown in Fig.3.11 (c).
(2) Q diagram
Recalling from equation (b), the shape of the shear of the inclined beam is a linear function with
respect to the horizontal distance x from the origin A, which will be plotted by two numerical values of its
ordinates. The two numerical values are
1
x =0, QA = ql cos θ
2
1
x=l, QB = − ql cos θ
2
When comparing the two shear values with those of its corresponding horizontal beam (the same span
subjected to the same load), their difference is only a constant cosθ , i.e., the magnitude of shears of inclined
beam is the projection of that of its corresponding horizontal beam along the normal direction of the axial
line of the inclined beam. The Q diagram is shown in Fig.3.11 (d).
(3) N diagram
Knowing from equation (c), the shape of the axial force of the inclined beam is a linear function with
respect to the horizontal distance x from the origin A, which will be depicted by two numerical values of its
ordinates. The two numerical values are
1
x =0, N A = − ql sin θ
2
1
x=l, N B = ql sin θ
2
When comparing the two axial force values with those of its corresponding horizontal beam, their
3.4 The Restraint Force Calculation and Geometric Construction
of Statically Determinate Multispan Beams 59
difference is only a constant sin θ , i.e., the magnitude of axial forces of inclined beam is the projection of
shears of its corresponding horizontal beam along the tangent direction of the axial line of the inclined
beam. The N diagram is shown in Fig.3.11 (e).
The example shows that:
(1) For simply supported inclined beam, the method to calculate its reactions and internal forces is
still the method of sections and application of equilibrium conditions of free bodies.
(2) Under the action of vertical loads, the reactions of a simply supported inclined beam are the
same as those of a simply supported horizontal beam with the same span and subjected to the same load.
(3) Under the action of uniformly distributed loads, the bending moment diagram of a simply
supported inclined beam are the same as those of a simply supported horizontal beam with the same span
and subjected to the same load.
(4) Under the action of vertical loads, a simply supported inclined beam will be imposed by axial
force; the magnitude of shears and axial forces of the inclined beam is the two projections of shears of its
corresponding horizontal beam along the normal and tangent directions of the axial line of the inclined
beam respectively.
Finally, the additional explanation is needed to mention, which indicate that although the axial line of
an inclined member is a straight slope line, the superposition method can be applied to plot bending
moment diagrams for all segments with straight axial lines of the member.
3.4 The Restraint Force Calculation and Geometric Construction of Statically

Determinate Multispan Beams
A statically determinate multispan beam is constituted by several single span beams connected by
hinged joints each other based on the geometric construction rules of a stable system with no redundant
restraint. This type of structure is a commonly used structural form in road bridges [Fig.3.12 (a)] and purlines
of buildings [Fig.3.13 (a)], whose corresponding computing model is shown in Fig.3.12 (b) and Fig.3.13 (b)
separately.
Generally, the reactions of a statically determinate multispan beam are more than three. The statically
determinate multispan beam shown in Fig.3.14 (a) have five reactions, which is greater than the number of
equilibrium equations provided by the free body of the entire beam. Obviously, the three equilibrium
equations are not sufficient to determine the five unknown reactions at the supports for this multispan beam.
However, the presence of the internal hinges at C and E yields two additional equations that can be used
with the three equilibrium equations to determine the five unknowns. The additional equations are based on
the conditions that internal hinges cannot transmit moments; that is, the moments at the internal hinged
joints must be equal to zero. Therefore, when an internal hinge is used to connect two portions of a
structure, the algebraic sum of the moments about the hinge due to loads and reactions acting on the
structure from either side of the hinge must be zero. Such equation is commonly referred to as the equation
(a) A B C D
640cm 1680cm 490cm 700cm 490cm 1680cm 640cm
(b) A B C D
Fig.3.12 Multispan road bridge
P u rlin
A B C D E F G H I
J
(a)
Top chord of roof
(b) A B C D E F G H I J
Fig.3.13 Multispan purlin
of condition or construction. For the multispan beam of Fig.3.14 (a), there are two internal hinges at C and
E which yield two equations of condition. For simplicity, we use the symbols M C = 0 , M E = 0 express
the equations. Therefore, all five unknown reactions for the beam of Fig. 3.14(a) can be determined by
solving the three equations of equilibrium plus two equations of condition ( M C = 0 and M E = 0 ).
Although the total number of equilibrium equations and condition equations of a statically determinate
multispan beam is sufficient to determine its all reactions, the difficulty of solving simultaneous equations
will sometimes occur. In order to avoid solving simultaneous equations the best procedure for determining
reactions of a multispan beam can be stated as: first, analyze the geometric construction of the beam, then,
find reactions in the order which is the reverse of that of its geometric construction.
The constituent characteristics of statically determinate multispan beams are that the ends of some
constituent members serve as supports for the other constituent members. Therefore, the former is
considered as the main portion of a multispan beam, and the latter as a subsidiary one. For example, in the
3.4 The Restraint Force Calculation and Geometric Construction
of Statically Determinate Multispan Beams 61
statically determinate multispan beam shown in Fig.3.14 (a), the overhanging beam ABC (the portion is
attached to the foundation by one hinged support and one roller support to form a stable system) is regarded
as the main one with respect to the subsidiary member CDE which is supported at end C by beam ABC; the
beam CDE is the main one with respect to the subsidiary member EFG at the same time. The geometric
construction relationship between members of the statically determinate multispan beam is presented by the
figure shown in Fig.3.14 (b).
q
A B C D E
(a) G
F
q
E
C D F G
(b) A B
q
XE = 0
G
E F
YE
YF
XC = 0 C D
(c) E XE = 0
YC YD
XA A B
C XC = 0
YA YB
Fig.3.14 Statically determinate multispan
The figures shown in Fig.3.14 (b) through (c) give a better idea about the transmitting behavior of the
forces of the entire beam: The load applied on the main beam produces no reactions and internal forces in
the subsidiary portions, whereas the load applied on the subsidiary portion always gives rise to reactions
and internal forces in the main portion of the beam; a multispan beam remains stable if only a subsidiary
portion is removed, but becomes unstable when a main portion is removed.
When analyzing the multispan beam shown in Fig.3.14 (a), take analyzing sequence as the order
which is opposite to that of the geometric construction, that is, first, calculate subsidiary portion EFG, then,
compute portion CDE, finally, analyze the main portion ABC [Fig.3.14 (c)]. We can find the restraint forces
X E , YE and reaction YF by applying three equilibrium equations of the free body EFG [Fig.3.14(c)],
then X E and YE are considered as the loads which are applied on member CDE as shown in Fig.3.14(c).
Similarly, the restraint forces X C and Y C and reaction YD may be obtained through the free body
CDE. Finally, the reactions X A , Y A and YB may be obtained by applying three equilibrium equations
of the free body ABC where X C and Y C are regarded as the external loads applied on.
In conclusion, it may be worth of mentioning that dividing a multispan beam into main and subsidiary
portions will facilitate its analysis. The analysis may be implemented in an order which is the reverse of
that of its geometric construction. The loads exerted on a subsidiary portion would be transmitted to its
main portion through the connecting device between them. By the analysis of the subsidiary portion we
may determine the restraint forces provided by the main portion, and these restraint forces (in opposite
direction) turn to the load acting on the main portion. Ultimately, all restraint forces and reactions may be
obtained successfully without solving simultaneous equations.
3.5 The Construction of Internal Force Diagrams of Statically Determinate Multispan

Beams
1. The sign convention of the internal forces and construction convention of the internal forces
diagrams of statically determinate multispan beams are the same as those of single span beams
2. The internal force calculation of a statically determinate multispan beam is, in fact, the internal
force computation of its each individual single span beam. Connecting internal force diagrams of all of its
individual single span beams together will accomplish the construction of the internal force diagrams of the
entire multispan beam.
3. The relationships between loads, shears and bending moments derived from straight members and
characteristics of internal force diagrams occurring on straight members are still suitable for constructing
internal force diagrams of statically determinate multispan beams, which can be employed to check the
correctness of the internal force diagrams of multispan beams and to depict the diagrams quickly. It should
be realized that at the locations of middle hinges of multispan beams bending moments must be equal to
zero ( M = 0 ).
Example 3-6
Construct the internal force diagrams for the statically determinate multispan beam shown in Fig.3.15
(a).
Solution
1. Geometric construction analysis, draw construction relationship diagrams of members

The geometric construction relationship between members of the statically determinate multispan
beam is expressed by the figure shown in Fig.3.15 (b). The beam ABC is fixed to the foundation by a fixed
3.5 The Construction of Internal Force Diagrams of Statically Determinate Multispan Beams 63
support to form the main portion of the statically determinate multispan beam. At the right end of member
CE, there is a hinge E which will provide horizontal restraint for member EFG, but under the action of
6kN 6kN
2kN 2kN/m
(a) A B C D E F G
2m 2m 1m 1m 2m 6m
(b) D G
A
B C E F
(c) D G
A
B C E F
6kN
2kN
D
(d) 5kN 3kN
6kN
32kN ⋅ m 5kN 3kN 2kN/m
G
A B C E F
11kN
10kN 5kN
(e) 32
M diagram
(kN ⋅ m)
10 6 (9)
D
A C G
B E F
3 6
(f )
11 11
Q diagram 7
5 5 3 3
(kN) E
G
A B C D F
3 3
5
Fig.3.15 Statically determinate multispan beam of example 3-6

vertical load no horizontal restraint force will exist at hinge E. Therefore, removing the horizontal restraint
at hinge E to location G would not change the stressing state of the statically determinate multispan beam.
So doing yields a construction relationship diagram shown in Fig.3.15 (c), in which cantilever ABC and
overhanging EFG are the main portions of the statically determinate multispan beam and simple beam CDE
is a subsidiary portion supported on the two main portions.
2. Compute restraint forces and reactions
In order to avoid solving simultaneous equations we should find restraint forces and reactions in the
order which is the reverse of that of the geometric construction of the beam. On this point of view, the
restraint forces of beam CDE is first calculated, which are considered as the loads applied on beam ABC
and EFG at the positions C and E (in opposite direction) respectively; then reactions of ABC or EFG
separately. The final results are shown in Fig.3.15 (d).
Check: applying the equilibrium condition in vertical direction Y of the free body of the entire beam,
we obtain
∑Y = 0 , 6 + 2 + 6 + 2 × 6 − 11 − 10 − 5 = 0
3. Construction of internal force diagrams
As soon as the restraint forces and reactions are determined, the bending moment and shear diagrams of
these individual members may be plotted by method of superposition segment by segment; put these
diagrams together, we obtain the bending moment and shear diagrams of the multispan beam. They are
shown in Fig.3.15 (e) and (f) separately.
4. characteristics of internal force diagrams
In Fig.3.15 (e) and (f), it is indicated that at hinges C and E the ordinates of bending moment diagram
are equal to zero; in the segments where no load is applied to, such as segments AB, BC, CD, DE and EF,
the shape of moment diagrams are straight slope lines while that of shear diagrams are horizontal lines. All
of these features of the diagrams coincide with the relationship existing between loads and shears and
bending moments of straight members.
Example 3-7
Construct the bending moment diagram for the statically determinate two-span beam shown in
Fig.3.16 (a) and then compare it with that of two neighboring simply supported beams.
Solution
1. Geometric construction analysis

The geometric construction relationship existing between members of the statically determinate
two-span beam is expressed by the figure shown in Fig.3.16 (b). The beam BCD is the main portion of the
beam. Whereas the beam AB is a subsidiary portion of the two-span beam.
3.5 The Construction of Internal Force Diagrams of Statically Determinate Multispan Beams 65
2. Compute reactions
In order to avoid solving simultaneous equations, the reactions of subsidiary beam AB is first
calculated, one (at position B) of which is considered as the load (in opposite direction) applied on beam
BCD; then reactions of BCD can be found easily. The final results are shown in Fig.3.16 (c).
1kN/m
(a) A D
B C
5.1m 0.9m 6m
(b) A B
C D
1kN/m
A B
2.55kN 2.55kN
2.55kN 1kN/m
(c )
B D
C
6.9kN 2.55kN
(d)
2.7
M diagram
A D
(kN ⋅ m)
C
3.25 3.15
(e)
M diagram A C D
(kN ⋅ m)
4.5 4.5
Fig.3.16 Statically determinate two-span beam of example 3-7
Check: applying the equilibrium condition in vertical direction Y to the free body of the entire beam,
we obtain
∑Y = 0 , 1× 12 − 2.55 − 6.9 − 2.55 = 0
3. Construction of bending moment diagram
The bending moment diagrams of individual beam AB and BCD may be plotted separately; put these
diagrams together, we obtain the bending moment diagram of the two-span beam. It is shown in Fig.3.16 (d).
4. Comparison of bending moment diagrams
Compare the bending moment diagram for the statically determinate two-span beam shown in Fig.3.16
(d) with that for two neighboring simply supported beams shown in Fig. 3.16 (e), we can find the difference
between the two cases. The maximum bending moment in each span of the two-span beam is smaller (in a
percentage of 28%) than that in the simple beams. The bending moment distribution of the two-span beam
is more reasonable than that of the simple beams, because the negative bending moment due to the load
acting on the overhang BC decreases the positive moment distributed in the span. The maximum and
minimum moments for the two-span beam can be controlled by adjusting the length of the overhang
SUMMARY
The primary content of the chapter is the calculation of reactions and construction of internal force
diagrams for both statically determinate single span beams and multispan beams.
Although the analysis of statically determinate single span beams has been learned in the course of
strength of materials, it is yet the foundation of analysis of statically determinate multispan beams and rigid
frames; the requirement now is that internal force diagrams of statically determinate single span beams
must be constructed as soon as correctly possible. It should be realized that the more practice would
enhance your sound proficiency about the construction of internal force diagrams. The key points of the
chapter are stated as followings:
1. Calculating step: firstly, find reactions (except cantilever beam); then, determinate internal forces;
finally, construct internal force diagrams.
The attention that the order to find the reactions is the reverse of that of their geometric construction
should be paid for multispan beams
2. The method of sections is the elementary method to determinate the internal force components in
a member. But the other form of method of sections should be applied skillfully. That is: The internal axial
force at any section of a member is equal in magnitude, but opposite in direction, to the algebraic sum of
the components in the direction parallel to the axis of the member of all the external loads and reactions
acting on either side of the section. We consider it to be positive when the external forces tend to produce
tension. The shear at any section of a member is equal in magnitude, but opposite in direction, to the
algebraic sum of the components in the direction perpendicular to the axis of the member of all the external
loads and reactions acting on either side of the section. We consider it to be positive when the external
forces tend to push the portion of the member on the left of the section upward with respect to the portion
on the right of the section. The bending moment at any section of a member is equal in magnitude, but
opposite in direction, to the algebraic sum of the moments about the section of all the external loads and
reactions acting on either side of the section. We consider it to be positive when the external forces and
couples tend to bend the member concave upward, causing compression in the upper fibers and tension in
the lower fibers at the section.
3. Shear, bending moment, and axial force diagrams depict the variations of these quantities along
the length of the member. Such diagrams can be constructed by determining and plotting the equations
expressing these stress resultants in terms of the distance of the section from an end of the member. The
construction of shear and bending moment diagrams can be considerably expedited by applying the
differential relationships that exist between the loads, shears, and bending moments.
4. The practical method for constructing bending moment diagram is the method of superposition.
The steps are: (1) dividing a member into several segments by control sections; (2) determining bending
moments of the control sections; (3) in segment where there is no load applied on, directly connecting the
ordinates with a straight line between two adjacent sections; (4) in segments where there are loads applied
on, connecting the ordinates with a straight dashed line between two adjacent sections and imposing the
ordinates of the corresponding bending moment diagram of a corresponding simple beam on the dashed
line, which is considered as a base line. Shear and axial force diagrams can be depicted in the similar way.
5. Depict the ordinates of internal forces in the direction perpendicular to the axial line of the
member. The ordinates of bending moment diagrams have to be plotted on the sides where the fibers of the
members are under tensile state, and no sign indication is necessary on bending moment diagrams. Shear
and axial force diagrams can be depicted on any sides of the members, but sign indication must be made
(for horizontal members, positive shears and axial forces are generally plotted on the top sides of the
members).
3-1 Why the moment diagram for a certain straight segment of a member may be constructed by the
method of superposition of a simple beam with the same span and subjected to the same loads?
3-2 Why the bending moment diagram of an inclined beam subjected to vertical loads is the same as that
of a horizontal beam with the same span and subjected to the same loads?
3-3 Would internal forces yield on the subsidiary portion when loads were applied on the main portion
for a statically determinate multispan beam? Why?
3-4 Why, generally the maximum bending moment in each span of a multispan beam is smaller than that
in a series of simple beams with the same spans?
3-1 Construct the bending moment and shear diagrams of following statically determinate single span
beams or columns (please follow the regulation required by the chapter).
3-2 Check and revise the following internal diagrams by the differential relationship existing between the
loads and shears and bending moments.
3-3 Construct bending moment diagrams of the following beams by using the method of superposition
segment by segment and calculate the bending moment at section C.
3-4 Construct internal force diagrams of the following inclined beams and calculate the internal forces at
sections C and D.
3-5 Calculate the reactions of the following statically determinate multispan beams and depict their
internal force diagrams.
3-6 A concentrated load P is applied on the multispan beam, please construct the bending moment
diagrams in the following three cases and compare them with each other: (1) P is only applied at the
middle of segment CD; (2) P is only applied at the middle of segment AB; (3) P is only applied at the
middle of segment EF.
3-7 Depict the bending moment diagram straightforwardly according the stressing characteristics and the
differential relationship existing between the loads and shears and bending moments of the multispan
beam.
ql 2
4kN
2 q 3kN/m
A A
C B C D B
l l 2m 2m 2m
2 2
(b)
(a)
6kN 8kN
4kN⋅ m 2kN/m 10kNim
1kN/m
A B
A D C D
C B 4m 2m 2m
2m 4m 2m
(c) (d)
Problem 3-1 (contd)

CHAPTER 4
PLANE STATICALLY DETERMINATE RIGID FRAMES
The chapter will discuss the analysis of plane statically determinate rigid frames. Firstly, present
the analysis of the support reactions; then, the internal forces and the construction of their diagrams.
When finding the reactions of three-hinged frames the condition of moment at middle hinge, i.e.,
M = 0 , must be used; when calculating the reactions of multispan and multistory rigid frames the
geometric construction analysis must be made in order to proceed the analysis in the order which is the
reverse of that of their geometric constructions; while computing internal forces the attention should be
paid on the equilibrium of rigid joints where several members are connected together; while
constructing bending moment diagrams the method of superposition segment by segment is still
powerful.
In practical engineering, there is a variety of employment for plane statically determinate rigid frames,
such as awnings in train or bus stations, car sheds [Fig.4.1 (a) and (b)], one-storey industrial workshops,
warehouses [Fig.4.2 (a) and (b)] and the like. The analyzing for statically determinate rigid frames is the
foundation of calculation of displacements of rigid frames and the base of analysis of statically
indeterminate rigid frames as well. Therefore, the content of this chapter must be really mastered.
(a) (b)
C20 crushed stone concrete
Fig.4.1 Y-shaped rigid frame
The procedure to analyze a plane statically determinate rigid frame has the following steps: (1) find
reactions; (2) calculate internal forces of control sections by using the method of sections; (3) construct
internal force diagrams by employing the method of superposition.
73
74 Chapter 4 Plane Statically Determinate Rigid Frames
Ridge joint
C
C
D E
D E
Asphaltum with jute fiber

A B B
A
(a) (b)
Fig.4.2 Three hinged rigid frame
4.1 Geometric Construction and Characteristics of Plane Statically Determinate Rigid

Frames
Rigid frames, usually referred to simply as frames, are composed of straight members (columns and
beams) connected either by rigid (moment-resisting) connections or by hinged connections to form stable
configurations. The frame shown in Fig.4.3 (a) is a portal frame, whose joints C and D are rigid joints. If all
of the joints of the frame shown in Fig.4.3 (a) were changed into hinged joints, the system would be
unstable truss; the dashed lines shown in Fig.4.3 (b) represent its movement. In order to make the system
[shown in Fig.4.3 (b)] be stable a restraint must be added, for instance, a slope member is attached to it as
shown in Fig.4.3 (c) to form a stable truss.
P C D P C D P C D
0
90
0
900
900 90
A B A B A B
(a) (b) (c)
Fig.4.3 Frame and truss

(a) a gate-shaped frame; (b) an unstable system; (c) a truss
Comparing the frame shown in Fig.4.3 (a) with the truss shown in Fig.4.3 (c), it can be seen that the
members of the frame are connected by rigid joints which are able to transmit bending moments and make
the system be stable; while the members of the truss are connected by hinges which are only able to
transmit forces but not moments, and in order to prevent the relative rotations of the members connected to
4.1 Geometric Construction and Characteristics of Plane Statically Determinate Rigid Frames 75
hinges inclined member has to be employed.

Therefore, the members of rigid frames are able to transmit not only forces but also bending moments,
whereas those of trusses are only able to transmit forces; a frame structure can develop larger internal space
with fewer members than a truss with the same dimension.
The members of frames are usually connected by rigid joints, although hinged connections are
sometimes used. A rigid joint prevents relative translations and rotation of the member ends connected to it,
so the joint is capable of transmitting two rectangular force components and a couple between the
connected members. Under the action of external loads, the member ends connected to one rigid joint of a
frame have the same displacements, that is, their ends have the same translations and rotate the same angle,
as dashed lines shown in Fig.4.3 (a) (a rectangular angle still remains rectangular under the deformation of
the frame); whereas a hinged joint only prevents relative translations but not the rotations of the member
ends connected to it, so a hinged joint is only capable of transmitting two rectangular force components
between the connected members. Under the action of external loads, the member ends connected to one
hinged joint of a structure only have the same translations, that is, their ends rotate different angles or they
have relative rotations, as dashed lines shown in Fig.4.3 (c).
Due to the hinged connections between the beam and the columns shown in Fig.4.4 (a) and their
disability of transmitting moments, under the action of the load distributed on beam CD, only the beam CD
yields bending moments and bending deformation, while the columns AC and BD are only subjected axial
q
q
C D
Δ
C D
2
Δ
ql
8 ql2
A 8
A B B
l l
(a) (b)
Fig.4.4 Frames
(a) a frame with hinged joints; (b) a gate-shaped frame
pressures and with no bending deformation. However, because of the rigid joints between the beam and the
columns shown in Fig.4.4 (b) and their ability of transmitting moments, under the action of the load
distributed on beam CD, not only the beam CD but also the columns AC and BD develop bending moments
and bending deformation. Contrasting the two situations, the maximum bending moment and bending
deformation of beam CD shown in Fig.4.4 (b) have reduced a lot in comparison with those of beam CD
shown in Fig.4.4 (a), whose dashed curves represent the deformation configurations.
The geometric construction rules for statically determinate rigid frames must obey the rules of stable
systems. Commonly utilized plane rigid frames have the following types:
(1) simply supported rigid frames and overhanging rigid frames [Fig.4.5 (a) and (b)];
(2) three hinged rigid frames [Fig.4.5 (c)];
(3) multispan and multistory rigid frames [Fig.4.5 (d) and (e)].
(b)
(a) (c)
(e)
(d)
Fig.4.5 Frames
(a) simply supported frame; (b) Y-shaped frame; (c) three hinged frame;
(d) multispan frame; (e) multistory frame
4.2 The Analyzing of Reactions for Statically Determinate Rigid Frames
Generally, in the analysis of a statically determinate rigid frame the reactions may be determined first,
and then the internal forces on control sections are calculated; finally the internal force diagrams may be
depicted. Due to the simplicity of simply supported and overhanging rigid frames, the following discussion
will only involve the analysis of reactions of statically determinate three hinged and multispan rigid frames.
It is worth of emphasizing that the order of calculating reactions of a frame should be the reverse of
4.2 The Analyzing of Reactions for Statically Determinate Rigid Frames 77
that of its geometric construction.

Example 4-1
Calculate the reactions of the three hinged frame shown in Fig.4.6 (a).
Solution
C C
2
l
P P
2
l
B XA B XB
A A
l l YA YB
2 2 (b)
(a)
Fig.4.6 A three hinged frame of example 4-1
There are totally four unknown support reactions on the frame, i.e., X A and YA , X B and YB .
From the viewpoint of geometric construction, three rigid bodies (AC, BC and the foundation) joined
pairwise by hinges A, B and C, which do not lie on the same straight line, form an internally stable system
with no redundant restraint. Thus the steps in the determination of reactions of the frame are:
(1) to evaluate the reactions YA and YB by taking the entire frame as a free body and then
applying the equations of moment equilibrium about hinges B and A ( ∑ M B = 0 and ∑ M A = 0 ), i.e.,
1 P
∑M B P × + YA × l = 0 ,
= 0,
2
YA = − (↓)
2
1 P
∑ M A = 0 , P × 2 − YB × l = 0 , YB = (↑)
2
P P
check: ∑Y = 0 , 2 2
− =0
(2) to determine the reaction X A or X B by considering either the right or the left portion as a free
body and then employing the equations of moment equilibrium about hinge C, if taking the right portion as
a free body we write
1 YB P
MC = 0 , X B × l − YB × = 0, XB = = (←)
2 2 4
(3) to evaluate reaction X A by taking the entire frame as a free body and then applying the equations
of force equilibrium in horizontal direction, i.e.,
3
∑ X = 0, P + XA − XB = 0 , X A = X B − P = − P (←)
4
Example 4-2
Calculate the reactions of the two-span frame shown in Fig.4.7 (a).
Solution
2kN/m (a) 2kN/m (b)
D D
E F G E F G
4m
A B C XA A C XC
B
YB YC
2m 2m 4m
4kN
2kN/m YE
(c) (d)
XE
D E 1kN E F G
4m
4m
XA
A B C XC
2m
2m 4m
YB YC
Fig.4.7 A two-span frame of example 4-2
There are totally four unknown support reactions [Fig.4.7 (b)] on the frame, i.e., X A , YB , X C and YC .
The geometric construction order can be stated as: in the beginning of geometric construction, a rigid frame
BCEFG is fixed on the foundation by a hinged support C and a roller support B to form a stable system with
no redundant restraint, and then the left portion is constructed. Thus the steps in the determination of
reactions of the frame are:
(1) To evaluate the reactions and restraint forces acting on the subsidiary portion ADE [Fig.4.7 (c)] by
4.3 Determination of Internal Forces of Member Ends by Using The Method of Sections 79
considering the left portion as a free body, i.e.，

∑M = 0, E 2 × 2 ×1 + X A × 4 = 0 , X A = −1kN(←)
∑Y = 0 , 2 × 2 − YE = 0 YE = 4kN(↑)
∑M = 0, A 2 × 2 ×1 − 4 × 2 + X E × 4 = 0 , X E = 1kN(→)
(2) To determinate reactions on the main portion BCEFG [Fig.4.7 (d)] by taking the portion as a free body
and then applying the equations of equilibrium, i.e.,
∑ X = 0, XC +1 = 0 , X C = −1kN(→)
∑ M = 0,
B −4 × 2 − 1× 4 − YC × 4 = 0 , YC = −3kN(↓)
∑ M = 0,
C −4 × 6 − 1× 4 + YB × 4 = 0 , YB = 7kN(↑)
check: ∑Y = 0 , 4+3−7 = 0
The foregoing steps for calculating reactions are the commonly used procedure for determinate
unknown forces, that is, the calculating order should be the reverse of that of its geometric construction of
the frame. However, the reactions of the example can be calculated in another way, that is, (1) taking the
subsidiary portion ADE as a free-body and utilizing the condition that the bending moment at hinge E must
be equal to zero, M E = 0 , to find reaction X A
ME = 0, 2 × 2 ×1 + X A × 4 = 0 , X A = −1kN(←)
(2) then considering the entire frame [Fig.4.7 (b)] as a free body and employing the equations of
equilibrium to obtain the other all reactions
∑M = 0,B −2 × 2 × 3 − YC × 4 = 0 , YC = −3kN (↓)
∑M = 0,C −2 × 2 × 7 + YB × 4 = 0 , YB = 7kN(↑)
∑ X = 0, X A − XC = 0 , X C = −1kN(→)
4.3 Determination of Internal Forces of Member Ends by Using The Method of

Sections
1. Internal forces of rigid frames and their sign convention

For a rigid frame, it is generally subjected to shear forces and bending moments as well as axial forces
under the action of external loads. It is no need for bending moments to regulate sign convention if
depicting the ordinate of the bending moments on the tension side of a member. The sign convention for
shears and axial forces are the same as those of beams discussed previously. That is, a positive shear makes
the section rotate clockwise and positive axial forces make the section in tensile state.
2. The member-end sections and the representation of internal forces of member ends at a rigid joint
A rigid frame is composed of straight members (columns and beams) with different axial-line
directions connected by rigid joints. There are thusly different member-end sections at a rigid joint. For
example, there are two member-end sections C1 and C2 at rigid joint C of the frame shown in Fig.4.8
(a). We will represent the internal forces of a member ends by using two subscripts, of which the first
represents the label of the nearend of the member whereas the second represents that of the farend of the
member. For instance, the bending moments of member-end sections C1 and C2 are represented by
M CA (C is the nearend of member CA while A is the farend of the member) and M CB (C is the nearend
of member CB while B is the farend of the member) respectively, and the shears and axial forces are
indicated by QCA and QCB , N CA and N CB separately as shown in Fig.4.8 (b).
N CA
8kN C C2 B M CA
M CB
C QCA
QCA N CB
C1 C
M CA QCB
6kN
3m
3m
N CA (c)
(a) (b)
8kN
A 8kN A
4m 6kN
6kN
QCB (e)
M CB 6kN
C B 8kN C
N CB 0
8kN 24kN ⋅ m
6kN
4m 24kN ⋅ m
6kN
(d)
Fig.4.8 A frame and free bodies of its members
3. The calculation of internal forces of member-end sections

The member-end sections here indicate the cross sections of members which are concurrent at rigid
joints, and we take this kind of sections as control sections. Then, the internal forces of the control sections
are computed by using the method of sections. We state the method as following again.
The bending moment on the end section of a member is equal in magnitude, but opposite in direction,
to the algebraic sum of the moments about the section of all the external loads and restraint forces acting on
either side of the section. We consider it to be positive when the external and restraint forces and couples
make the member rotate in the opposite sense as you assumed. The shear on the end section of a member is
equal in magnitude, but opposite in direction, to the algebraic sum of the components in the direction
perpendicular to the axis of the member of all the external loads and restraint forces acting on either side of
the section. We consider it to be positive when the external forces and restraint forces tend to make the
section rotate clockwise. The axial force on the end section of a member is equal in magnitude, but opposite
in direction, to the algebraic sum of the components in the direction parallel to the axis of the member of all
the external loads and restraint forces acting on either side of the section. We consider it to be positive
when the external forces and restraint forces tend to produce tension.
We can also use free-body diagrams of each members of a frame to calculate the internal forces of the
control end sections, and in some cases the calculation will be accomplished efficiently by taking the
free-body diagrams of rigid joints.
The control sections for the frame shown in Fig.4.8 (a) are sections C1 and C2 at rigid joint C. in
order to find the internal forces acting on the two sections we use method of sections with no free-body
diagram first. For section C1 (consider its underside), we obtain
M CA = 8 × 3 = 24kN ⋅ m (Tension in the right fiber)
QCA = 8kN
N CA = 6kN
For section C2 (consider its right side), we obtain
M CB = 6 × 4 = 24kN ⋅ m (Tension in the lower fiber)
QCB = −6kN
N CB = 0
Check: take rigid joint C as a free body [Fig.4.8 (b)] and we find that the three equilibrium
conditions, ∑ X = 0 , ∑ Y = 0 and ∑ M = 0 are satisfied.
You can use each free-body diagram shown in Fig.4.8 (c) to (e) to check the correctness of the
calculation.
Example 4-3
Calculate member end internal forces at rigid joints of the frame shown in Fig.4.9 (a).
Solution
The control sections for the frame shown in Fig.4.9 (a) are sections C1 , C2 and C3 at rigid joint C.
In order to find the internal forces acting on the three sections we use method of sections with no free-body
diagram first. For section C1 (consider its left side), we obtain
(a) (b) (c)

1kN / m 3kN 1kN / m 1kN / m 3kN
2kN C 2kN
2kN N CB C
B C1 C2 D QCA
C3 B C B D
M CB C3
2m M CA
QCB
N CA
4m
(d) (e)
QCD 3kN 2 3
M CD 2 6
C
A 2 0
N CD C2 D
2m 2m QCA
2m M CA
N CA
Fig.4.9 The frame and the free bodies of its members of example 4-3
N CB = −2kN
QCB = −1× 2 = −2kN
M CB = 1× 2 × 1 = 2kN ⋅ m (Tension in the upper fiber)
For section C2 (consider its right side), we obtain
N CD = 0
QCD = 3kN
M CD = −3 × 2 = −6kN ⋅ m (Tension in the upper fiber)
For finding N CA , QCA and M CA , we take rigid joint C as a free body [Fig.4.9 (e)] and obtain
N CA = −2 − 3 = −5kN
QCA = 2kN
M CA = −2 + 6 = 4kN ⋅ m (Tension in the left fiber)
The readers can check the results by using free-body diagrams shown in Fig.4.9 (b) to (d).
Example 4-4
Find member end internal forces at rigid joints C and D of the frame shown in Fig.4.10 (a).
Solution
D2
C C2 D
C1 D1 QDC
3kN / m
M DC
4m
N DC D
12kN QDB = 0 M DB = 0
A B
N DB = −4
6m (b)
4kN 4kN
(a)
Fig.4.10 The frame and the free body of rigid joint D of example 4-4
1. Reactions considering entire frame as a free body we will find the reactions as shown in
Fig.4.10 (a).
2. Find member-end internal forces at rigid joint C
For section C1 (consider its underside), we obtain
N CA = 4kN
QCA = −3 × 4 + 12 = 0
M CA = −12 × 4 + 3 × 4 × 2 = −24kN ⋅ m (Tension in the right fiber)
For section C2 (consider its left side), we obtain
N CD = 3 × 4 − 12 = 0
QCD = −4kN
M CD = −12 × 4 + 3 × 4 × 2 = −24kN ⋅ m (Tension in the lower fiber)
3. Find member-end internal forces at rigid joint D
For section D1 (consider its underside), we obtain
N DB = −4kN
QDB = 0
M DB = 0
For section D2 (consider its right side), we obtain
N DC = 0
QDC = −4kN
M DC = 0
We can also use the free-body diagram [Fig.4.10 (b)] of rigid joint D to find the end forces N DC ,
QDC and M DC after the end forces N DB , QDB and M DB are obtained. The results are the same as
the above calculation.
4.4 Construction of Internal Force Diagrams of Statically Determinate Frames
4.4.1 Illustration of construction of internal force diagrams of statically determinate rigid frames
Internal force diagrams of a statically determinate rigid frame are bending moment, shear and axial
force diagrams. Each kind of internal force diagram is composed of the diagrams for all the members of the
frame. In general, by following the steps: (1) evaluating the internal forces at the both ends of each member,
(2) determining the shape of the internal force diagram
by the differential relationships existing between the 4kN
internal forces and external loads, (3) employing the C
1m
B
principle of superposition when the member is subjected
D
to external loads as well as the end forces, the diagram
may be conveniently depicted. Now we utilize an YB
1kN/m
4m
example to illustrate the procedure for constructing
internal force diagrams.
Example 4-5
A
Construct internal force diagrams of the frame XA
shown in Fig.4.11. 4m
YA
Solution Fig.4.11 The frame of example 4-5
1. Calculation of reactions [Fig.4.11]
∑ M = 0 ， 4 × 5 + 1× 4 × 2 − Y × 4 = 0 ，
A B YB = 7kN(↑)
∑ X = 0 , 4 + 1× 4 − X = 0 ， A X A = 8kN(←)
∑ M = 0 , 4 × 1 − 1× 4 × 2 + 8 × 4 + Y × 4 = 0 ,
B A YA = −7kN(↓)
2. Construct M diagram
(1) Compute the bending moments on the end sections of each member
By the method of sections, the bending moment at the end section of a member is equal in magnitude,
4.4 Construction of Internal Force Diagrams of Statically Determinate Frames 85
but opposite in direction, to the algebraic sum of the moments about the section of all the external loads and
restraint forces acting on either side of the section. We consider it to be positive when the external and
restraint forces and couples make the member rotate in the opposite sense you assumed.
Member CD, from C to D:
M CD = 0
M DC = 4 × 1 = 4kN ⋅ m (Tension in the left fiber)
Member BD, from B to D:
M BD = 0
M DB = 7 × 4 = 28kN ⋅ m (Tension in the lower fiber)
Member AD, from A to D:
M AD = 0
M DA = 8 × 4 − 1× 4 × 2 = 24kN ⋅ m (Tension in the right fiber)
(2) Construct M diagram of each member of the frame
For a rigid frame，the M diagram is composed of the M diagrams of all the members of the frame.
After the bending moments at the both ends of each member are evaluated, the shape of its M diagram can
be determined by the differential relation existing between M and external loads, and then the diagram may
be plotted by using the principle of superposition when the member is subjected to an external load as well
as the end moments.
There are totally three members in the frame. We may draw the bending moment diagrams (whose
ordinates should be laid off on the tensile side of a member) of these members one by one as follows:
Member CD and BD: Since the two members carry no external load but the end moments, each of
their bending moment curves is a slope straight line. We can directly connect the ordinates of its two end
moments of each member with a straight line.
Member AD: since the member carries uniformly distributed loads, the bending moment curve is a
quadratic parabola. The parabola can be accomplished as followings: firstly, connect the ordinates of the
two end moments ( M AD and M DA ) of the member with a straight dashed line; then, superpose the
ordinates of the corresponding bending moment diagram of a corresponding simple beam(which has the
same span and is subjected to the same loads)on the dashed line. The final M diagram shown in Fig.4.12 (a)
and the final bending moment at the section E (which is located at the middle length of the member AD) is
1 1
ME = ( 0 + 24 ) + ×1× 42 = 14kN ⋅ m (Tension in the right fiber)
2 8
3. Construct Q diagram
(1) Compute the shears of each member end

Member CD, from C to D: QCD = QDC = 4kN
Member BD, from B to D: QDB = QBD = −7kN
Member AD, from A to D: QAD = 8kN
QDA = 8 − 1 × 4 = 4kN
C
C 4 C
D 7 B
D 24 B D B
4 4
28 7 7
E 14
Q diagram N diagram
M diagram
(kN)
(kN ⋅ m) (kN)
A
8 A 7
A (c)
(a) (b)
4kNim 4kN
(d) 7kN
(e)
D
D 28kNim
4kN
24kNim 7kN
Fig.4.12 Internal force diagram of example 4-5
(2) Construct Q diagram of each member

The ordinates of a shear diagram of a member can be plotted on either side of the member, but the sign
has to be indicated on the diagram. For the frame shown in Fig.4.11, directly connecting the ordinates of
the shears of each end of the members of the frame will finish the drawing of the shear diagram as shown in
Fig.4.12 (b).
4. Construct N diagram of each member of the frame
(1) Evaluate the axial forces of each member end
Member CD, from C to D: N CD = N DC = 0
Member BD, from B to D: N DB = N BD = 0
Member AD, from A to D: N AD = N DA = 7kN
4.5 Internal Force Diagrams of Statically Determinate Three-Hinged Rigid Frames,
Multispan and Multistory Rigid Frames 87
(2) Depict N diagram of each member
The ordinates of an axial force diagram of a member can be depicted on either side of the member, but
the sign has to be indicated on the diagram. For the frame shown in Fig.4.11, directly connecting the
ordinates of the axial forces of each end of the members of the frame will complete the drawing of the axial
force diagram as shown in Fig.4.12 (c).
5. Check
Taking the rigid joint D of the frame as a free body, all of the equilibrium conditions are satisfied by
the following equations:
∑M = 0,D 4 + 24 − 28 = 0
∑ X = 0, 4−4 = 0
∑Y = 0 , 7−7 = 0
4.4.2 The key points of constructing the internal force diagrams for statically determinate rigid
frames
From foregoing examples, we observe that the fundamental method of constructing internal force
diagrams of rigid frames is the method of decomposition and composition. It can be stated as following:
(1) Calculate the internal forces on the member-end sections of each member of a rigid frame;
construct the internal force diagrams of each of the members respectively by employing the method of
sections; compose each kind of internal force diagrams of all members together to form the internal force
diagram of the entire fame.
(2) The regulation of constructing internal force diagrams: The ordinates of internal force diagrams
of each member of a rigid frame must be depicted in the sense perpendicular to the axial line of the member.
The ordinates of bending moment diagram of a member ought to be laid off on the tensile side of the
member, whereas the ordinates of a shear and an axial force diagram of a member can be plotted on either
side of the member, but the sign has to be indicated on the diagram. Magnitudes of the internal forces on
some special sections must be indicated on the diagrams.
(3) Since the all members of a rigid frame are straight bars, the method of superposition segment by
segment presented previously can be implemented for each straight member.

Multispan and Multistory Rigid Frames
4.5.1 Construction of internal force diagrams of statically determinate three-hinged rigid frames,
multispan and multistory rigid frames
The steps of constructing internal force diagrams for a statically determinate three-hinged rigid frame,
or a multispan or multistory rigid frame remain the same as those discussed previously. They are: (1)
computing reactions; (2) evaluating the internal forces of member ends for each member; (3) constructing
internal force diagrams. It should be noted again that the order to determinate reactions ought to be the
reverse of that of geometric construction of a structure. The following examples explain the steps of
constructing internal force diagrams for statically determinate three-hinged rigid frames, multispan and
multistory rigid frames.
Example 4-6
Construct bending moment diagram of the three-hinged gable frame shown in Fig.4.13 (a) and
compare it with that of the frame shown in Fig.4.13 (b).
Solution
1kN/m 1kN/m
D E
C C
2m
4m
D E
2m
XA XB
A B XA XB
A B
4m 4m
YA YB 4m 4m
YA YB
(a)
(b)
Fig.4.13 Frames of example 4-6
1. Calculation of reactions [Fig.4.13 (a)]

There are totally four unknown support reactions on the frame, i.e., X A and YA , X B and YB .
From the viewpoint of geometric construction, three rigid bodies (ADC, BEC and the foundation) joined
pairwise by hinges A, B and C, which do not lie on the same straight line, form a stable system with no
redundant restraint. Thus the steps in the determination of reactions of the frame are: (1) to evaluate the
reactions YA and YB by taking the entire frame as a free body and then applying the equations of moment
equilibrium about hinges B and A ( ∑ M B = 0 and ∑ M A = 0 ), i.e.,
∑M = 0, A 1× 8 × 4 − YB × 8 = 0 ， YB = 4kN(↑)
∑M = 0, B YA × 8 − 1× 4 × 8 = 0 ， YA = 4kN(↑)
Check: ∑Y = 0 , −4 + 1 × 8 − 4 = 0
(2) to determine the reaction X A (or X B ) by using the construction condition of hinge C, i.e., M C = 0 .
We can employ the algebraic sum of the moments about hinge C of all the external loads and reactions
from the either side (left or right) of C to establish it. If taking the right side of C, we write
MC = 0 , 1× 4 × 2 − 4 × 4 + X B × 4 = 0 , X B = 2kN(←)
(3) to evaluate reaction X A by taking the entire frame as a free body and then applying the equations of
force equilibrium in horizontal direction, i.e.,
∑ X = 0, XA = XB , X A = 2kN(→)
(1) Compute the bending moments at the end sections of each member
Member AD: M AD = 0 ， M DA = 2 × 2 = 4kN ⋅ m (tension in the left fiber)
Member DC: M CD = 0 ， M DC = 2 × 2 = 4kN ⋅ m (tension in the upper fiber)
Member BE: M BE = 0 ， M EB = 2 × 2 = 4kN ⋅ m (tension in the right fiber)
Member CE: M CE = 0 ， M EC = 2 × 2 = 4kN ⋅ m (tension in the upper fiber)
(2) (2)
8 8
(2) (2) 8 D E 8
C C
4 4
4 M diagram
D M diagram E 4
(kN ⋅ m)
A (kN ⋅ m)
B
A B
(a) (b)
Fig.4.14 Bending moment diagrams of example 4-6
(2) Construct M diagram of each member of the frame

For the three-hinged gable rigid frame, the M diagram is composed of the M diagrams of all above
listed members of the frame. After the bending moments at the both ends of each member are evaluated, the
shape of its M diagram can be determined by the differential relation existing between M and external loads,
and then the diagram may be plotted by using the principle of superposition when the member is subjected
to an external load as well as the end moments.
There are totally four members in the frame. We may draw the bending moment diagrams (whose
ordinates should be laid off on the tensile side of a member) of these members one by one as follows:
Member AD and BE: Since the two members carry no external load but the end moments, each of
their bending moment curves is a slope straight line. We can directly connect the ordinates of its two end
moments of each member with a straight line.
Member DC and CE: since the two members carry uniformly distributed loads, each of the two
bending moment curves is a quadratic parabola based on its inclined axial line. The parabola can be
accomplished as followings: firstly, connect the ordinates of the two end moments of the member with a
straight dashed line; then, superpose the ordinates of the corresponding bending moment diagram of a
corresponding simple beam (which has the same span, the length of horizontal projection of member DC or
CE, and is subjected the same loads) on the dashed line. The final M diagram shown in Fig.4.14 (a) and the
final bending moment at the middle section of the member DC or CE is
1 1
M =− ( 0 + 4 ) + × 1× 4 2 = 0
2 8
3. Discussion
Fig.4.14 (b) presents the bending moment diagram of the three-hinged rigid frame shown in Fig.4.13
(b), of which the span and height of middle hinge and acting loads are the same as the frame shown in
Fig.4.13 (a). Comparing the two bending moment diagrams, we observe that the bending moment
magnitude of beams and columns of the frame shown in Fig.4.13 (a) is smaller than that of beams and
columns of the frame shown in Fig.4.13 (b)， because of the reason that the beams of the frame shown in
Fig.4.13 (a) have inclined axes and the columns have smaller heights. Therefore, the stressing state of the
frame shown in Fig.4.13 (a) is more rational than the frame shown in Fig.4.13 (b).
Example 4-7
Construct bending moment diagram of the two-span frame shown in Fig.4.15 (a) [the same as shown
in Fig.4.7 (a)]
Solution
1. Calculation of reactions and restraint forces

The reactions and restraint forces have been determined in example 4-2, and now they are shown
again in Fig.4.15 (b).
The followings are the calculation of bending moments of member ends of the frame, whose
subsidiary and main portion are ADE and CBEFG, respectively.
Member AD: M AD = 0 ， M DA = 1× 4 = 4kN ⋅ m (tension in the right fiber)
Member DE: M ED = 0 ， M DE = 1× 4 = 4kN ⋅ m (tension in the lower fiber)
Member EF: M EF = 0 ， M FE = 2 × 2 = 4kN ⋅ m (tension in the upper fiber)
Member FG: M FG = 4 × 2 = 8kN ⋅ m (tension in the upper fiber), M GF = 1× 4 = 4kN ⋅ m (tension
in the lower fiber)
Member CG: M CG = 0 , M GC = 1× 4 = 4kN ⋅ m (tension in the left fiber)
Member BF: M BF = M FB = 0
For the frame, the M diagram is composed of the M diagrams of all above listed members of the frame.
After the bending moments at the both ends of each member are evaluated, the shape of its M diagram can
be determined by the differential relationships between M and external loads, and then the diagram may be
plotted by using the principle of superposition when the member is subjected to an external load as well as
the end moments.
The final M diagram shown in Fig.4.15(c), on which the moment at hinge E is equal to zero,
i.e., M E = 0 and the bending moment curve at location E is derivable, i.e., there is only one tangent line at
E.
2kN/m (a) (c) 8

4 4
D D G
E F G E F
4
4
4m
M diagram
(kN ⋅ m)
A B C A C
B
2m 2m 4m
(b)
2kN/m
4kN 1kN E
D F G
E 1kN
4kN
4m
4m
1kN A B C 1kN
2m 2m 4m
7kN 3kN
Fig.4.15 A frame and its free-body and bending moment diagrams of example 4-7
4.5.2 The check of internal force diagrams of rigid frames
1. The check of equilibrium conditions

since any portion isolated from a rigid frame, such as a rigid joint, a member or the entire frame, must keep in
equilibrium, the equilibrium conditions that may not be used in calculation of internal forces ought to be
employed as possible as you can when check the correctness of internal force diagrams.
It should be emphasized that the three equilibrium conditions ( ∑ M = 0 , ∑ X = 0 , ∑ Y = 0 )for
each rigid joint of a frame must be employed. The three equilibrium conditions can be used to calculate the
internal forces on member end sections as well.
It should be noted that the characteristics of the bending
moment diagram at a rigid joint of a rigid frame must satisfy (a) (b)
4 2
6
the followings: when a rigid joint is composed by two D C
member ends, the magnitude of the two bending moments of
4
the two member ends must be identical and the tensile fiber 4
of the rigid joint must be yielded on the same side; while a
Fig.4.16 Free-body diagram of rigid joint
rigid joint is consisted of more than two member ends the
bending moments of the member ends must keep in
equilibrium. Such as the bending moments of member ends at the rigid joint D [see Fig.4.16 (a)] of the
frame shown in Fig.4.15 (a), the two bending moments are equal in magnitude, but opposite in direction of
rotation; they keep the joint in equilibrium and put the internal side of the joint in tension. And the bending
moments of member ends at the rigid joint C [see Fig.4.16 (b)] of the frame shown in Fig.4.9 (a), the
algebraic sum of the three bending moments are equal to zero, the clockwise is 6kN · m and the
counterclockwise is 2 kN·m +4 kN·m, the bending moments keep in equilibrium.
2. The check of characteristics of internal force diagrams and relationship existing between the
loads, shears and bending moments
The relationship existing between the loads, shears and bending moments for straight members is still
suitable for the beams and columns of rigid frames. For the rigid frames on which middle hinges exist, the
bending moments at hinges must be equal to zero, i.e., M = 0 must be satisfied at hinges.
SUMMARY
The chapter has discussed the analysis of plane statically determinate rigid frames and its content
includes:
1. Steps to calculate statically determinate rigid frames
(1) Find reactions; (2) calculate internal forces of control sections (the sections of member ends) by
using the method of sections; (3) construct internal force diagrams by employing the method of
Summary 93
superposition.
2. Calculation of reactions
In order to avoid the trouble of solving simultaneous equations, it is preferable that the order of
calculating reactions of a rigid frame should be the reverse of that of its geometric construction. Therefore,
for a rigid frame whose geometric construction is quite complex, we must analyze its geometric
construction or stability first, and then calculate reactions in the order opposite to that of its geometric
construction.
3. The internal forces of a rigid frame include bending moment, shear and axial force, the attention
to their sign convention ought to be paid.
4. The method to determine the internal forces on the member-end sections
A rigid joint may include several member ends. Generally, the method for calculating the internal
forces of the member ends at a rigid joint is still the method of sections. For emphasis, it is stated again as
the followings:
The bending moment on the end sections of a member is equal in magnitude, but opposite in direction,
to the algebraic sum of the moments about the section of all the external loads and restraint forces acting on
either side of the section. We consider it to be positive when the external and restraint forces and couples
make the member rotate in the opposite sense as you assumed. The shear on end sections of a member is
equal in magnitude, but opposite in direction, to the algebraic sum of the components in the direction
perpendicular to the axis of the member of all the external loads and restraint forces acting on either side of
the section. We consider it to be positive when the external forces and restraint forces tend to make the
section rotate clockwise. The axial force at end sections of a member is equal in magnitude, but opposite in
direction, to the algebraic sum of the components in the direction parallel to the axis of the member of all
the external loads and restraint forces acting on either side of the section. We consider it to be positive
when the external forces and restraint forces tend to produce tension.
5. Construction of internal force diagrams
The ordinates of a shear or axial force diagram must be depicted in the direction perpendicular to the
axial line of the member and can be plotted on the either side of the member, but the sign has to be
indicated on the diagrams.
The ordinate of a bending moment diagram must be depicted on the side where the fiber of the
member is in tensile state. No sign indication is necessary for bending moment diagrams. The powerful
way for constructing bending moment diagrams is the method of superposition member by member. That is,
for members imposed by no external load but the end moments, directly connect the ordinates of its two
end moments of each member with a straight line; for members imposed by loads, firstly, connect the
ordinates of the two end moments of the member with a straight dashed line, then, superpose the ordinates
of the corresponding bending moment diagram of a corresponding simple beam (which has the same span
and is subjected to the same loads) on the dashed line.
6. The check of internal force diagrams
The shapes of internal force diagrams must satisfy the relationships existing between the loads, shears
and bending moments. Each rigid joint of a frame must keep in equilibrium, i.e., ∑ M = 0 , ∑Y = 0 and
∑X =0. The bending moment curves must get across each middle hinge of a rigid frame, and the
bending moment diagrams must be located in the same side of the rigid joints composed of two member
ends.
4-1 Simply reword the characteristics of geometric construction of rigid frames.

4-2 Simply reword the difference between a rigid joint and a hinged joint based on their features of
transmitting forces and deformation as well as free-body diagrams.
4-3 Simply reword the features of calculating reactions for a statically determinate rigid frame.
4-4 For a statically determinate rigid frame, how conveniently determine its reactions when they are
more than three?
4-5 What is the sign convention of internal forces in the members of a rigid frame? Why we have to use
two subscripts to represent the internal forces of member end sections of a rigid frame?
4-6 How orient unknown bending moment directions when you isolate a free-body diagram from a rigid
frame? How determine the tensile side by the calculating result of bending moments?
4-7 Simply reword the calculating points of internal forces of statically determinate rigid frames.
Comparing the points with those of statically determinate beams, what are their differences?
4-8 What are the steps for constructing internal force diagrams of statically determinate rigid frames?
4-9 What are the regulations for constructing internal force diagrams of statically determinate rigid
frames?
4-10 How use the equilibrium conditions to check the correctness of internal force diagrams? How employ
the conditions that the bending moments at hinged joints must be equal to zero to check the bending
moment diagram of a rigid frame?
4-11 What are the characteristics of internal force diagrams at rigid joint C of the rigid frame shown in the
problem for reflecting [Fig.4.11 (a) and (b)]? Try to list the equations that the internal forces of
member ends at rigid joint C must be satisfied?
(a) A
(b) A
b
m D
P D
C C
a
a
B B
a a
Figs. of problem for reflecting 4.11
4-1 Calculate the reactions of following plane statically determinate rigid frames.
q = 2kN/m
C
C D 1.5kN
0.85m
4m
5kN/m
A B B
A
6m 1m
(a) (b)
Problem 4-1 (contd)

C D 10kN
A B
C
q a
2kN/m
4m
A B
D
4m 4m
a
(d)
(c)
A Pa
2kN/m
B
E
a
D C
2p C D
5m
a
A B
E
a a 5m 5m
(e) (f)
6kN
C E
D
D M E
C
6m
a
A B A B
a a 3m 3m 2m
(g) (h)
Problem 4-1
4-2 Calculate the internal forces on the designated section C for the frame shown in the figure of
problem 4-1(a).
problem 4-1(c).
problem 4-1(d).
4-5 Calculate the internal forces on the designated sections B and C for the frame shown in the figure of
problem 4-1(e).
8kN 8kN
2kN
D E F
C E
3kN/m
D
3kN/m
3m
3m
1m
4kN A B C
A B
6m 2m
2m 2m 2m 2m
Problem 4-6 Problem 4-7
4-6 Calculate the internal forces on the designated section D for the frame shown in the figure of
problem 4-6.
4-7 Calculate the internal forces on the designated section E for the frame shown in the figure of problem
4-7.
4-8 Construct the internal force diagrams for the following frames.
2kN/m
C D P
C B
2
h
P acts at points D,
E and B respectively
E
4m
2
h
A A B
4m l
(a) (b)
Problem 4-8
4-9 Construct the internal force diagrams for the frame shown in the figure of problem 4-1 (b).
4-10 Construct the internal force diagrams for the frame shown in the figure of problem 4-1 (d).
4-11 Construct the bending moment diagrams for the following frames.
Pa
2kN/m
C D
a C D
4m
(a) (b)
A B A B
a 2m 3m 2m
(c) 4kN
C
3m
2kN
B (d)
2kN/m
5m
B C
6m
A
A
2m 2m 2m 4m 4m 6m 4m 4m
Problem 4-11
4-12 Construct the internal force diagrams for the three-hinged frame shown in the figure of problem 4-1
(f).
4-13 Construct the internal force diagrams for the three-hinged frame shown in the figure of problem 4-1
(g).
4-14 Construct the bending moment diagrams for the three-hinged frames shown in the figures of problem
4-14.
4-15 Construct the bending moment diagram for the two-span frame shown in the figure of problem 4-15.
4-16 Construct the bending moment diagram for the two-storey frame shown in the figure of problem
4-16.
4-17 Check the mistakes for the following bending moment diagrams. If there are mistakes in the
diagrams please revise them.
C
D
4kN/m
a
P P
D C E
E
10m
a
A B A B
2.5m 5m 5m 2.5m a a a a
2 2
(a) (b)
Problem 4-14
E D 8kN
4kN/m 4kN
3m
8kN
D E F G H
F C
4m
3m
A B C A B
2m 2m 2m 2m 3m 3m

(a) (b)
Problem 4-17
CHAPTER 5
THREE HINGED ARCHES

The chapter will discuss the analysis of three hinged arches. There are two important
characteristics for a three hinged arch. They are: (1) curved members; (2) horizontal thrusts yielded
under vertical loads. The two characteristics make the analysis and stressing performance of three
hinged arches quite different from those of other type of structures. For instance, when finding the
horizontal thrusts of a three hinged arch the condition of moment at middle hinge, i.e., M = 0 , must
be used; when constructing the internal force diagrams of a three hinged arch the values of the
internal forces at a number of sections must be evaluated in order to finish the diagrams; the
dominating internal force in a three hinged arch is axial compression, its bending moments and
shears are much smaller in comparison with its axial compression; there exists a rational or optimal
axial line for a three hinged arch.
5.1 The Constitution and Type of Three Hinged Arches
We may find reactions in the way as we do in the previous analysis. It is worth of mentioned that the
analyzing order ought to be the reverse of that of its geometric construction of a three hinged arch to avoid
solving simultaneous equations. Three hinged arches composed of curved members and hinges are
C C
P P
f
HA HB
A B A B
l
VA VB VA (b) VB
(a)
Fig.5.1 Two types of three hinged arches
statically determinate arch-typed structures. They have a variety of application in bridge structure, water
power engineering, underground engineering and building roof structures. Fig.5.1 (a) and (b) show the two
101
102 Chapter 5 Three Hinged Arches
primary types of three hinged arches.

The arch shown in Fig.5.1 (a) is a three hinged arch with no tie. Two curved members AC and BC are
connected each other by hinge C and then connected on their foundation by two hinges A and B, the three
hinges do not lie on the same straight line, to form a stable system with no redundant restraint. Under the
action of vertical loads, a horizontal thrust H will develop on two foot supports. The thrust impose very
important influence on the internal forces in the arch. The arch shown in Fig.5.1 (b) is a three hinged arch
with a tie. Hinged support B as shown in Fig.5.1 (a) is now replaced by a roller support. Under the action of
vertical loads, the tensile force in the tie substitute the function of the horizontal thrust developed in the
foot supports of the arch with no tie.
Some of terminology pertinent to three hinged arches will be explained by employing the arch shown
in Fig.5.1 (a) as an example. The curve ACB is referred to as centre or axial line of the arch, the axial line
of a commonly used arch is a parabola or circular arc. The hinge C is named top or central hinge, whose
height, the distance between C and the line connecting the hinges of supports A and B, is called the rise of
the arch and denoted by f. The horizontal distance between the hinges of supports A and B is referred to as
the span and denoted by l. The value of f / l is termed as the ratio of rise to span. The ratio f / l is one
of the basic parameters of an arch, which varies in the extent of 1 ~ 1/10 .
5.2 The Reactions of Three Hinged a1 b1

Arches under Action of Vertical (a)
a2 b2
Loads
d1 P2
P1
D C
There are four reaction components acting
on the arch shown in Fig.5.2 (a), VA and H A ,
yD
HA A B HB
VB and H B . We must need four equilibrium l1 l2

VA VB
equations to determinate the four unknown l
reactions. Apparently, the equilibrium condition
of the entire arch can contribute three equilibrium P1 P2
equations; in addition, another construction (b) A B
condition at hinge C can be obtained that the D C
l1 l2 VB0
bending moment at hinge C must be equal to zero, VA0
i.e., M C = 0 . Since the number of unknown
l
reactions is just equal to that of the equilibrium

Fig.5.2 A three hinged arch and its
and construction equations, it is evident that a corresponding simple beam
three hinged arch is a statically determinate
structure.
5.2 The Reactions of Three Hinged Arches under Action of Vertical Loads 103
Usually, the two support hinges of a commonly used three hinged arch are mounted on the same level,
the directions of action lines of its loads are vertical and there is a vertical axis of symmetry in the arch. The
calculation of reactions for this kind of commonly utilized three hinged arches will be discussed as below.
In order to make a clear comprehension and compare the difference of the reactions between an arch
and a beam, a corresponding simply supported beam covering the same span and carrying the same vertical
loads is depicted as shown in Fig.5.2 (b). Under the action of vertical loads, the horizontal reaction of the
simple beam is equal to zero, i.e., H = 0 ; vertical reactions ( VA and VB ) of the simple beam can be
0 0
determined by the equilibrium conditions of the entire beam, ∑ M B = 0 and ∑ M A = 0 , respectively.

1 1
VA0 = ( Pb
1 1 + P2b2 ) ， VB = ( Pa
0
1 1 + P2 a2 ) (5-1)
l l
Considering the entire arch as a free body and applying equilibrium conditions
( ∑ M B = 0 and ∑ M A = 0 ) to it, we obtain
1 1
VA = ( Pb
1 1 + P2 b2 ) ， VB = ( Pa
1 1 + P2 a2 ) (5-2)
l l
Comparing equation (5-1) with equation (5-2), we find
VA = VA0 , VB = VB0 (5-3)
Obviously, the magnitudes of vertical reactions of the arch are the same as those of its corresponding
simple beam.
Considering the entire arch as a free body and applying the horizontal force equilibrium condition
( ∑ X = 0 ) to it, we obtain
HA = HB = H
That is, the horizontal reactions at supports A and B of the arch are equal in magnitude but opposite in
direction, and we use alphabet H indicate the thrust. The direction of H shown in Fig.5.2 (a) is recognized
to be positive.
In order to determine the thrust H, we employ the additional condition provided by hinge C, i.e., the
bending moment at hinge C must be equal to zero ( M C = 0 ), and if calculating M C from A to C we
obtain
[VAl1 − Pd
1 1 ] − Hf =0
in which, the first two terms (in square brackets) are the algebraic sum of the moments about hinge C of all
the external loads and reactions acting on left side of the hinge C, which is equal to the magnitude of the
bending moment on the corresponding section C of the corresponding simple beam. Indicating the bending
0
moment by M C , we write
M C0 = VAl1 − Pd
1 1
Upon that, the moment about hinge C of the three hinged arch may be expressed as
M C0 − Hf = 0
M C0
H= (5-4)
f
It will be seen from this that under the action of vertical loads the thrust has no relationship with the
form of the axial line of the arch; in fact, it is related to the locations of the three hinges of the arch; the
magnitude of thrusts H is in direct proportion to 1/ f , i.e., the flatter the arch becomes the lager the thrust
will be. When the direction of a load is downward the thrust, imposed to the arch by the supports, is
positive as shown in Fig.5.2 (a), and its direction is inward. In the limit case if the rise f vanishes,
i.e., f → 0 the thrusts H approach to infinite
value. As a matter of fact, if the three hinges lie on the same line the special system will be unstable
and it can not keep in equilibrium.
For the three hinged arch with a tie shown in Fig.5.3 (a), there are only three reactions, which can be
determined by the three equilibrium equations ( ∑ M B = 0 , ∑ M A = 0 and ∑ X = 0 ) of the entire arch.
The relationships H A = 0 , VA = VA and VB = VB remain usable.
0 0
d1 P1 d1
P1
C P2 C HC
VC
HA = 0 A B A
NAB
VA l1 l2 VB VA l1
(a) (b)
Fig.5.3 A three hinged arch with a tie
The tensile force of the tie can be obtained by cutting the tie and the hinge, considering the left portion
of the section C as a free body [Fig.5.3 (b)] and applying the additional condition that the algebraic sum of
the moments about hinge C of all the external load, reaction and the tensile force acting on the side must be
5.3 Formula for Calculating Internal Forces of Three Hinged Arches
under Action of Vertical Loads 105
equal to zero, i.e., M C = 0 . Thusly doing, we write
[VA l1 − Pd
1 1 ] − N AB f =0
In which the first two terms (in square brackets) remain equal to the magnitude of the bending moment on
the corresponding section C of the corresponding simple beam. Indicate the bending moment with symbol
M C0 yet.
It can be seen from the analytical results that the magnitude of the tensile force of the tie of a three
hinged arch with a tie is the same as that of the thrust H of a three hinged arch with no tie. The equation
(5-4) is still usable for computing the tensile force of the tie. It can be rewritten as
M C0
N AB = (5-4＇)
f
5.3 Formula for Calculating Internal Forces of Three Hinged Arches under Action of
Vertical Loads
A three hinged arch is composed of curved members. The cross-sectional dip ϕ , which is the angle
formed by the tangent to the centre line of the arch and the axis of abscissa, varies with the position of the
cross section of the arch. When the centre line of an arch is given (i.e., the centre line function y = f ( x)
is known), the slope of a section of the arch may be determined by the first derivative, i.e., tan ϕ = dy / dx ,
of the centre line function of the arch, sin ϕ and cos ϕ of the angle ϕ may be determined as well.
Internal force components over an arbitrary cross section of an arch may be the bending moment,
shearing force (or simply shear) and axial force.
The sign convention of internal force components are regulated as following: a positive bending
moment tends to cause tension in the inner fibers and compression in the outer fibers of the arch; a positive
shear makes the section rotate clockwise and positive axial force makes the section in tensile state.
The method to calculate the internal forces of an arch is still the method of sections. The formulae for
calculating the internal forces acting over an arbitrary cross section D [Fig.5.2 (a)] of an arch under the
action of vertical loads will be derived now.
Pass an imaginary section perpendicular to the centre line of the arch at the point D where the internal
forces M D , QD and N D [see Fig.5.4 (a)] are desired and take the left portion of section D as a free
body. In order to facilitate the analysis the internal force components on section D are replaced by its
horizontal and vertical force components X D and YD shown as in Fig.5.4 (b); the corresponding
bending moment and shear on the corresponding section of a corresponding simple beam, indicated by
M D0 and QD0 [see Fig.5.4 (c)], are herein borrowed.
By applying the equilibrium conditions in the vertical and horizontal directions to the free bodies
shown in Fig.5.4 (b) and (c) respectively, we obtain

∑Y = 0 ， Y = V − P = Q
D A 1
0
D
∑ X = 0， X = H D
(a) (b) YD = QD0

a1 P1 P1
MD
ND ϕD
ϕD XD = H
D yD D MD
H A QD H A
VA xD
VA
(c) (d) (e)

P1 H
M D0 QD0 cos ϕ D
A ϕD QD0
ϕD
H sinϕD
D H cosϕD
QD0
VA0 QD0 sin ϕ D
Fig.5.4 Internal force calculation of a three hinged arch
The internal forces on section D will be expressed by the corresponding bending moment M D0 ，shear
QD0 of a corresponding simple beam and the thrust H of the arch in the following.
1. Formula for calculating bending moments
Considering the free body shown in Fig.5.4 (a) or (b) and applying the equilibrium equation of
moment about the centroid of section D, we write
∑M D = 0 ， M D = [VA xD − P1 ( xD − a1 )] − HyD
Since VA = VA , the first two terms in square brackets on the right hand side of the equation are the
0
bending moment on the section D of the corresponding simple beam. The above equation can be thusly
expressed as
M D = M D0 − HyD
Therefore, formula for calculating the bending moment of an arbitrary section will be
M = M 0 − Hy (5-5)
0
In which M denotes the bending moment on the corresponding section in the corresponding simple
beam; y represents the distance between the centroid of the section and the line connecting the two hinges
of the arch supports.
2. Formula for calculating shears and axial forces
Obviously, the shear QD and axial force N D on section D must be the algebraic sum of the
components in the directions perpendicular and parallel to the tangent to the centre line of the arch of forces
0
H and QD acting on section D respectively. That is
QD = QD0 cos ϕ D − H sin ϕ D
N D = −QD0 sin ϕ D − H cos ϕ D
Formulae for calculating the shears and axial forces of an arbitrary section thusly be
Q = Q 0 cos ϕ − H sin ϕ (5-6)
N = −Q 0 sin ϕ − H cos ϕ (5-7)
Where, ϕ denotes the cross-sectional dip, in a rectangular coordinate system if axis of abscissa is positive
dy
to right and axis of ordinate is positive upward, > 0 , ϕ > 0 in the left half portion of an arch and
dx
dy
< 0 , ϕ < 0 in the right half portion of an arch. Q 0 denotes the shear on the corresponding section
dx
in the corresponding simple beam.
It can be seen from above formulae that if the centre line of an arch is given the location and slope of a
section (i.e., x, y and ϕ ) of the arch may be determined; and then by employing the internal forces M
0
0
and Q on the corresponding section of a corresponding simple beam of the arch, the internal forces on
each section of the arch may be calculated directly by the above formulae (5-5) through (5-7).
3. Construct internal force diagrams
Since an arch is composed of curved members the curves representing the internal force variation are
related with the ordinate y of the centre line and the cross-sectional dip ϕ of the arch. So the method of
constructing the internal force diagrams for straight members is not available for arches. When the internal
force diagrams are required, sufficient number of cross sections must be considered. Based on the obtained
numeric values of internal forces for these sections, ordinates may be laid off from the curved centre line of
the arch or from the horizontal abscissa, and then the diagrams may be plotted.
Example 5-1
A three hinged arch and the loads acting on it is shown in Fig.5.5. The centre line of the arch follows
a conic parabola given by the equation
4f
y= x(l − x)
l2
Determine the reactions and construct internal force diagrams.
Solution
1. Calculation of reactions 3kN/m
10kN
Employing equations (5-3) and (5-4), we write
10 × 3 + 3 × 6 × 9 y
C
VA = VA0 = = 16kN(↑) D
12
x
3 × 6 × 3 + 10 × 9 A B
VB = VB0 = = 12kN(↑)
12 6m 3m 3m
12m
M C0 16 × 6 − 3 × 6 × 3
H= = = 10.5kN
f 4 Fig.5.5 A three hinged arch of example 5-1
2. Calculation of internal forces
Once the reactions have been obtained the
internal forces on any section will be determined by using the formulae (5-5) through (5-7). In order to
construct the internal force diagrams eight sections, located at the ends of segments (divided along the span
by using identical horizontal projective length), are considered and the internal forces on these sections
have to be calculated. As an example to explain the computing procedure we will evaluate the internal
forces on the section D whose abscissa is 9m from the origin of the rectangular system shown in Fig.5.5.
(1) Geometric parameter of section D
By the function of the axial line of the arch when x = 9m , we determine
4f 4× 4
y= 2
x(l − x) = 2 × 9(12 − 9) = 3m
l 12
dy 4 f 4× 4
tan ϕ D = = 2 (l − 2 x) = 2 (12 − 2 × 9) = −0.0667
dx l 12
Therefore, we obtain
sin ϕ D = −0.55 ， cos ϕ D = 0.832
(2) Internal forces of section D
Based on formula (5-5), we write
M D = M D0 − HyD = 12 × 3 − 10.5 × 3 = 4.5kN ⋅ m
We will employ formulae (5-6) and (5-7) to find the shear and axial force on section D. since there is a
0
concentrated force at point D the corresponding shear QD in the corresponding simple beam will vary
abruptly from the left to the right of the section, and we must compute the shear value on the left and the
right sections of point D. We write
Table 5-1 internal force calculation of a three hinged arch
Bending moment
Geometric parameters of sections Shear computation Axial force computation
Section
computation
No.
Q0
x y tan ϕ sin ϕ cos ϕ M0 − Hy M Q 0 cos ϕ − H sin ϕ Q −Q 0 sin ϕ − H cos ϕ N
0 0 0 1.333 0.800 0.600 0.00 0.00 0.00 16.00 9.58 -8.40 1.18 -12.8 -6.29 -19.09
1 1.5 1.75 1.000 0.707 0.707 20.63 18.38 2.25 11.50 8.13 -7.42 0.71 -8.13 -7.42 -15.55
2 3.0 3.0 0.667 0.555 0.832 34.50 31.50 3.00 7.00 5.82 -5.83 -0.01 -3.89 -8.74 -12.63
3 4.5 3.75 0.333 0.316 0.949 41.63 39.38 2.25 2.50 2.37 -3.32 -0.95 -0.79 -9.96 -10.75
4 6.0 4.0 0.000 0.000 1.000 42.00 42.00 0.00 -2.00 -2.0 0.00 -2.00 0.00 -10.50 -10.50
5 7.5 3.75 -0.333 -0.316 0.949 39.00 39.38 -0.38 -2.00 -1.90 3.32 1.42 -0.63 -9.96 -10.59
-2.00 -1.66 5.83 4.16 -1.11 -8.74 -9.85

6 9.0 3.0 -0.667 -0.555 0.832 36.00 31.50 4.50
-12.00 -9.98 5.83 -4.16 -6.66 -8.74 -15.40
7 10.5 1.75 -1.000 -0.707 0.707 18.00 18.38 -0.38 -12.00 -8.48 7.42 -1.06 -8.48 -7.42 -15.90
8 12.0 0 -1.333 -0.800 0.600 0.00 0.00 0.00 -12.00 -7.19 8.40 1.21 -9.60 -6.29 -15.89
⎧⎪QDL = QDL 0
cos ϕ D − H sin ϕ D = (−2) × 0.832 − 10.5 × (−0.555) = 4.16kN
⎨ 0
⎪⎩ N DL = −QDL sin ϕ D − H cos ϕ D = −(−2) × (−0.555) − 10.5 × 0.832 = −9.85kN
0
QDR = QDR
0
cos ϕ D − H sin ϕ D＝（－12）× 0.832 − 10.5 ×（− 0.555）＝ − 4.16kN
N DR = −QDR
0
sin ϕ D − H cos ϕ D = −(−12) × (−0.555) − 10.5 × 0.832 = −15.4kN
When implementing the concrete calculation tabulation (see table 5-1) may be utilized for clarity and
simplicity. By the numeric values listed in table 5-1 of the internal forces of the arch, bending moment M
and shear Q and axial force N diagrams are depicted in Fig.5.6 (a), (b) and (c) (which ordinates from the
curve centre of the arch), or in Fig.5.6 (e), (f) and (g) (which ordinates from the horizontal abscissa).
In the interest of comparing an arch and its corresponding beam, the bending moment of a
corresponding simple beam (with the same span and carrying the same loads) is plotted in Fig.5.6 (d).
Comparing the bending moment diagram shown in Fig.5.6 (a) with that shown in Fig.5.6 (d), it will be seen
that the bending moments in the arch (with the maximum bending moment of 4.5kN·m) are much smaller
than those in a corresponding simple beam (with the maximum bending moment of 42.67kN·m). Obviously,
the reason that makes the bending moments in an arch decrease is the thrust of the arch. One of primary
characteristics of arches is that the thrust will be developed at its supports under the action of vertical loads;
consequently, the arch-typed structures are termed as compression structure.
An arch tends to flatten out under vertical loads and it is the horizontal thrust or the axial tension in the
tie who works against the deviation between the supports. The primary internal force in an arch is axial
compression, while the primary force in a beam is the bending moment.
C 0.38
(e) M diagram (kN ⋅ m) 0.38 0.38
0
0.38 A B
2.25 2.25 3.0 2.25
3.0 4.5 4.5
A 2.25 B
(a) M diagram (kN ⋅ m)
1.5m 1.5m 1.5m 1.5m1.5m 1.5m 1.5m 1.5m
4.16
(f) Q diagram (kN)
1.42 4.16 1.21
C 1.18 0.71 1.42
A B
0.71 −0.95 0.01 0.95
−0.01 2 2
1.06
1.21
−4.16 −1.06
1.18 A B 4.16
(b) Q diagram (kN)
−10.50 −10
−10.75 .59 −1
2.63 5−9. .40
.5 5 −1 85 (g) N diagram (kN)
−15
−1
C 19.09 15.55
5
.9
12.63 15.90 15.89

10.75 10.50 10.59 15.40
0
−19 .89 A
9.85
.0 9 5 B
A B −1
(c) N diagram (kN)
A (d) M 0 diagram (kN ⋅ m) B

20.63
C
18
36
34.5
41.63
39
42
Fig.5.6 Internal force diagrams of example
42.67
5.4 Stressing Performance of Three Hinged Arches
Foregoing discussion reveals that:

1. Under the action of vertical loads, unlike a beam, which develops only vertical reactions at its
supports, the arch gives rise to both horizontal thrust and vertical reactions at its supports.
2. It can be seen from the formula (5-5) that the bending moments in an arch are much smaller than
those in a corresponding simple beam.
3. Under the action of vertical loads, unlike a beam, in which there exist no axial force, an arch
yields axial compression, which is the dominated internal force in the arch.
4. Since the stress distribution on the sections of arches is more uniform than that of beams the
capabilities of constructional materials can be effectively utilized in arch-typed structures. Moreover,
because arches develop mainly compressive stresses when subjected to external loads the constructional
materials which are stronger in compression but weaker in tension such as bricks, dimension stone and
concrete, and the like, are suitable to build arch-typed structures.
However, any thing bearing two sides, when an inward thrust is provided by the supports for an arch
which makes the bending moments relatively small in the arch, an outward thrust is as well imposed on the
supports by the arch. Therefore, the three hinged arches with ties may be applied to the roof structures so as
to reduce the action of the thrusts imposed to side walls or columns by the arches.
5.5 Rational Axial Lines of Three Hinged Arches
From the previous analysis, it is realized that although the dominant normal stresses acting on the
sections of a three hinged arch are caused by axial compression, the magnitude of another kind of normal
stresses induced by bending moments and distributed unevenly on the sections are considerable as well. In
the interest of utilizing the capability of constructional materials sufficiently, the unevenly distributed
normal stresses on the sections of an arch must be decreased as less as possible. If there exists only axial
compression on the sections and the bending moments induced in an arch vanish, the normal stresses on the
sections will be uniformly distributed and the dimension of the arch will be expected to have a minimum
theoretical volume. Under the action of stationary loads the axial line of an arch that will induce the
bending moments on the sections to vanish (i.e., the arch remains no bending-moment state) is referred to
as rational or optimal axial or centre line of the arch.
5.5.1 Optimal centre line of three hinged arches under vertical loads
According to formula (5-5), it is recognized that under the action of vertical loads the bending moment
on an arbitrary section of a three hinged arch M ( x) is equal to the bending moment of its corresponding
simple beam M ( x) superposing the term − Hy ( x) , that is,
0
M ( x) = M 0 ( x) − Hy ( x)
When the centre line of an arch becomes optimal the bending moments of sections of the arch are
equal to zero, that is,

M ( x) ≡ 0 ， M 0 ( x) − Hy ( x) ≡ 0
Therefore, the function of a rational axial line ought to be
M 0 ( x)
y ( x) = (5-8)
H
0
In which M ( x) is the expression of the bending moments of the corresponding simple beam (with the
same span and carrying the same loads as those of the arch), and if depicting it as a figure the figure will be
the bending moment diagram of the simple beam.
Obviously, under vertical loads the difference between the expression of the optimal centre of a three
hinged arch and the expression of its corresponding simple beam is only a proportional constant H, that is,
the ordinates of optimal centre line of the arch is directly proportional to those of the bending moment
diagram of the simple beam.
With mastery of the concept of the optimal centre line of an arch, the rational form of an axial line of
an arch will be selected when design the arch; the capability of its constructional materials will be thusly
utilized sufficiently.
Example 5-2
In Fig.5.7 (a) a three hinged arch is under the action of uniformly distributed vertical load q, find the
optimal centre line of the arch.
Solution
Locate the origin of the rectangular system at point A shown in Fig.5.7 (a), and the corresponding
simple beam of the arch is shown in Fig.5.7 (b).
By formula (5-8), we write
M 0 ( x)
y ( x) =
H
The bending moment equation of the corresponding simple beam is
ql 1 qx
M 0 ( x) = x − qx 2 = (l − x)
2 2 2
The thrust determined by equation (5-4) is
M C0 ql 2
H= =
f 8f
After some simple operation we obtain the function of the optimal centre line of the arch as follows
qx
(l − x)
4f
y ( x) = 2 2 = 2 x(l − x) (5-9)
ql l
8f
It can be seen from equation (5-9) that under the action of uniformly distributed vertical load the
function of the optimal centre line of a three hinged arch is a quadratic parabola. Accordingly, the arches
with parabola centre lines are commonly employed in the civil engineering structures.
C
q
y
f
x ql ql
B
A 2
2 x
l l l
2 2
(a) (b)
Fig.5.7 Arch and its corresponding beam of example 5-2
5.5.2 Optimal centre lines of three hinged arches under uniformly distributed hydraulic pressure
The uniformly distributed hydraulic pressure is one sort of load which encounters frequently in water
power engineering, pipework and the like. Following example will show the optimal centre line of a three
hinged arch under uniformly distributed hydraulic pressure.
Example 5-3
A three hinged arch shown in Fig. 5.8 (a) is under the action of uniformly distributed hydraulic
pressure q, prove that the optimal centre line of the arch is a curve of circular arc.
Solution
First, assume the arch is under zero bending moment state under the action of uniformly distributed
hydraulic pressure.
(1) Take an infinitesimal element DE of the arch as a free body [Fig.5.8 (b)]. The angle between the
two end sections is denoted by dϕ ; the curvature radius at point D and E are denoted by r and r + dr
respectively; the length of the infinitesimal axial line will be ds = rdϕ .
(2) Since the arch is under zero bending moment state there exists only axial force but no bending
moment and shear on the sections D and E. The moment equilibrium equation may be set up by equating to
zero the algebraic sum of moments about the centre of curvature of the arch, i.e., ∑ M 0 = 0 , we write
∑M 0 = 0， N D r − N E (r + dr ) = 0
Neglecting the infinitesimal term N E dr , we obtain
ND = NE
This means that under the action of uniformly distributed hydraulic pressure if the arch is under zero
bending moment state the axial force is a constant throughout the sections of the arch. It may be denoted by
N.
(3) The force equation may be set up by equating to zero the algebraic sum of the projections of the
forces in the direction of the bisector of the angle dϕ . We write
dϕ dϕ
N D sin + N E sin + qrdϕ = 0
2 2
dϕ dϕ
Since N D = N E = N and dϕ is infinitesimal quantity we may replace sin by , and
2 2
then we obtain
Ndϕ + qrdϕ = 0
dϕ
(a) (b) (c) 2
O'
q q NE
C DN
E E
D ds r+dr
D NE
ND r dϕ
A dϕ
B
o 2
Fig.5.8 Arch and its infinitesimal free body of example 5-3
N
N = −qr ， r = − = constant
q
Because both the axial force N throughout the arch and the external load q remain constant, the
curvature r of each section throughout the arch remains constant as well. Recalling from geometry, the
curve with a constant curvature is a circular arc. This means that the optimal centre line of a three hinged
arch under the action of uniformly distributed hydraulic pressure is a circular arc.
In conclusion, under the action of uniformly distributed hydraulic pressure the rational or optimal
centre line of a three hinged arch is an arc of circle. This is the reason why the cross sections of structures
under uniformly distributed hydraulic pressure, such as waterpipes, high-pressure tunnels and arch dams,
are designed in annular shapes.
5.5.3 Optimal centre lines of three hinged arches under earth pressure
The earth pressure is one common load which encounters frequently in bridge engineering and
tunneling. Following example will show the optimal centre line of a three hinged arch under earth pressure.
Example 5-4
A three hinged arch shown in Fig. 5.9 is under qc q(x)

earth pressure. The upper surface of earth is a
x
horizontal plane; the arch is thusly under the load of
C
earth weight which may be expressed by
q( x) = qc + γ y A B
Where qc is the intensity of the distributed load at the y

crown hinge C, γ is the weight of the earth per unit
Fig.5.9 A three hinged arch subjected to earth pressure
volume. Find the optimal centre line of the arch under
the earth pressure.
Solution
Since earth pressure q( x) varies with respect to the ordinate y of the arch and y is an unknown
function now, we differentiate Eq. (5-8) with respect to x twice times and obtain the following differential
equation
d 2 y 1 d 2 M 0 ( x)
= (a)
dx 2 H dx 2
Because the positive y axis is in downward direction, here Eq. (a) should be expressed as
d2y 1 d 2 M 0 ( x)
= − (b)
dx 2 H dx 2
By using the differential relationship existing between the bending moment M 0 ( x) and vertical load
q( x) , we have
d 2 M 0 ( x)
= −q( x) = −(qc + γ y ) (c)
dx 2
By a simple operation, Eq. (b) may be expressed as

d2y 1
= (qc + γ y )
dx 2 H
or
d2y 1
2
− k 2 y = qc (d)
dx H
γ
where k =
2
H
0
It can be seen from Eq. (c) that both M ( x) and q( x) are related to the arch configuration. In this
situation, finding the solution of the governing differential equation (d), we obtain the optimal curve of the
arch which is expressed by a hyperbolic function
qc
y = C1 cosh(k x) + C2 sinh(k x) −
γ
Where C1 and C2 are the undetermined constants of integration which may be determined by the
boundary conditions of the arch
for x = 0, dy dx = 0, hence C2 = 0
qc
for x = 0, y = 0, thusly C1 =
γ
The function of the optimal centre line now becomes
qc
y= (cosh k x − 1)
γ
In mathematics, the curve described by such equation is referred to as catenary line.
In conclusion, under the action of earth pressure the rational or optimal centre line of a three hinged
arch is a catenary. This is the reason why the cross sections of structures under earth pressure such as
highway and bridge tunnels are designed in catenoid shapes.
SUMMARY
The key points of the chapter are as followings:

1. Composing of curved members and yielding a horizontal thrust under the action of vertical loads
are the two pronounced characteristics of a three hinged arch, which are the difference between arches and
beams as well. The two characteristics make the analysis and stressing performance of three hinged arches
quite different from those of other type of structures.
Summary 117
2. The formulae for calculating the reactions and internal forces of a three hinged arch under the
action of vertical loads are summarized as followings:
VA = VA0 ⎫
⎪
VB = VB0 ⎬
⎪
H = MC / f ⎭
0
M = M 0 − Hy ⎫
⎪
Q = Q 0 cos ϕ − H sin ϕ ⎬
N = −Q 0 sin ϕ − H cos ϕ ⎭⎪
The above formulae also express the relationship between internal forces and reactions of three hinged
arches and those of their corresponding simple beams
3. It is the first time that the method of constructing the internal force diagrams of curved members
is presented in the chapter. That is, when the internal force diagrams are required, sufficient number of
cross sections must be considered; according to the obtained numeric values of internal forces for these
sections, ordinates may be laid off from the curved centre line of the arch or from the horizontal abscissa,
and then the diagrams may be plotted.
4. Since the horizontal thrust makes the bending moments greatly small (comparing with those of
beams) and the dominant internal force in an arch is axial compression, the stress distribution on the
sections of arches is more uniform than that of beams. So the arch-typed structures may overpass longer
span and carry larger loads comparing with those of beams. Moreover, the capability of their constructional
materials can be effectively utilized and the brittle materials such as bricks, dimension stone and concrete,
and the like, which are stronger in compression but weaker in tension, are suitable to build arch-typed
structures.
5. For determined stationary loads, the centre line which makes the bending moments in an arch
vanish is referred to as the optimal centre line of the arch. Under the action of uniformly distributed vertical
load the function of the optimal centre line of a three hinged arch is a quadratic parabola; under the action
of uniformly distributed hydraulic pressure the optimal centre line of a three hinged arch is an arc of circle;
under the action of earth pressure the optimal centre line of a three hinged arch is a catenary.
5-1 How many types are there of three hinged arches?

5-2 Are ties usually used to the three hinged arches applying in roof structures? Why?
5-3 What is the difference between the reactions of three hinged arches and those of beams under the
action of vertical loads?

5-4 How does the thrust H of an arch vary when its rise f changes (increase or decrease) under the
condition that the span and loads of the arch remain unchangeable?
5-5 What is the difference between the internal forces of three hinged arches and those of beams under
the action of vertical loads?
5-6 How calculate ϕ , sin ϕ and c os ϕ of formulae (5-6) and (5-7)?
5-7 Whether or not the formulae which are used to calculate the reactions and internal forces of three
hinged arches can be utilized to compute those of three hinged rigid frames?
5-8 What are the main stressing characteristics for three hinged arches?
5-9 Why the capability of constructional materials constructed in arches can be utilized more efficiently
than it can be done in beams?
5-10 What is the optimal centre line of a three hinged arch?
P P P
l l l l l
2 2 3 3 3
(a) (b)
Figs. of problem for reflecting 5-11
5-11 What are the characteristics of the forms of the optimal centre lines of three hinged arches under the
action of vertical loads? How determine the optimal centre line of a three hinged arch under the
action of the loads shown in the figure [Fig.5.11 (a) and (b)]? Whether or not the rise of a three
hinged arch influences its optimal centre line?
5-1 The figure in problem 5-1 shows a three hinged arch with a centre line of a quadratic parabola, of
which the function is
4f
y= x(l − x), l = 16m, f = 4m
l2
(1) Find reactions
(2) Calculate the bending moment M, shear Q and axial force N on sections D and E.
5-2 The figure in problem 5-2 shows a three hinged arch with a centre line of a circle arc, Calculate the
reactions and the bending moment M, shear Q and axial force N on sections D.
5-3 The figure in problem 5-3 shows a three hinged arch with a centre line of a quadratic parabola, of
which the function is
4f
y= x(l − x)
l2
(1) Find reactions and internal forces in all two-force members.
(2) Calculate the bending moment M, shear Q and axial force N on sections K.
5-4 For the three hinged arch shown in the figure of problem 5-1,
(1) If change the height of the rise (set f = 8m ), how the reactions vary?
(2) If change the height of the rise and the length of span (set l = 32m , f = 8m ) but the ratio f / l
remains unchangeable, how the reactions and bending moments vary?
20kN/m
1kN/m
10kN C
D
y 5m
C
5m
D
E
f = 4m 300
B x A B
A
4m 4m 4m 4m 5m 3m
10kN/m
C
3m
K
1m
A B
D E
3m 3m 3m 3m 6m
Problem 5-3
CHAPTER 6
PLANE STATICALLY DETERMINATE TRUSSES AND
COMPOSITE STRUCTURES
A truss is an assemblage of straight members connected at their ends by flexible connections
(hinged joints) to form a rigid configuration. The members are subjected only to axial forces under the
action of joint loads. There are two sorts of methods used for analyzing plane trusses. They are the
method of joints, considering joints as free bodies and applying two equilibrium equations at each joint,
and the method of sections, taking a desired portion of the truss as a free body and applying three
equilibrium equations to the portion. The analyzing sequence for internal forces should be the order
which is the reverse of that of the geometric construction of the truss to avoid solving simultaneous
equations. The composite structures are composed by flexural and two-force members, which should be
distinguished clearly. Generally, the axial forces in two-force members ought to be analyzed first, then
the bending moments and shears and axial forces in flexural members will be obtained favorably.
6.1 Characteristics and Classification of Trusses
Unlike beams and frames, of which the dominating internal forces are bending moments that make the
stress distribution on the section are uneven [Fig.6.1 (a)] under the action of loads, a truss is an assemblage
of straight members connected at their ends by frictionless hinges to form a stable system, under the action
of joint loads whose members are subjected only to axial forces which develop uniformly distributed stress
[Fig.6-1 (b)]on the sections and make the efficient utilization of capacity of the constructional materials of
the truss. Because of their light weight and high strength, trusses are widely used, and their applications
range from roofs of buildings (Fig.6.2) and supporting structures of bridges (Fig.6.3) to being support
structures in space stations. Modern trusses are constructed by connecting members, which usually consist
of structural steel or aluminum shapes or wooden struts, to gusset plates by bolted or welded connections
(Fig.6.4 through Fig.6.5).
6.1.1 Assumptions for analysis of trusses and their internal force characteristics
For practical trusses, the connections between their members are almost never completely frictionless
hinges. In fact, real trusses are constructed by connecting members to gusset plates by welded or bolted
connections. Some members of the truss may even be continuous at the joints. The centroidal axes of
120
6-1 Analyze the geometric constructions and distinguish their types (simple, compound or complex) of
the following trusses.
(a) a a a P (b)
300
(c) P P
P
(d)
P 2P
(e)
(f)
problem 6-1 (contd)

150 Chapter 6 Plane Statically Determinate Trusses and Composite Structures
(g) (h) P
problem 6-1
20kN
(a) C (b)
20kN
B
D
3m
a
F E B A
4 × 3=12m P 4a
8kN
G H 4kN
(d) F
4m
P1
(c) P2 D 4kN
E D C
B E
4m
a
A A B
C F
P2 P1 3m 3m
4a
problem 6-3 (contd)

P P
(e)
G H
(d) B C D
4m
3m
P
E F
F
3m
E
C D
6m
4m
A A
B
3m 4m
2m 4m 2m
problem 6-3
6-2 Solve the following questions in the following problems: (1) point out zero-force members; (2)
point out the characters of axial forces in the members (compression or tension) by considering the
equilibrium of joints and depict out the clue of the transmission of forces.
（a） The truss shown in the figure of problem 6-1 (c);
（b） The truss shown in the figure of problem 6-1 (d);
（c） The truss shown in the figure of problem 6-1 (e);
（d） The truss shown in the figure of problem 6-1 (f).
6-3 Determine the member forces of the following trusses by using the method of joints.
6-4 Determine the axial forces in the indicated members of the following trusses by using the method of
sections.
6-5 Determine the axial forces in the indicated members of the following trusses by selecting either of
the two methods, which should be an optimal one for computation.
6-6 Determine the axial-force sign of the members and the maximum-force member of each of the
following two trusses.
6-7 Compute the axial forces in the two-force members and construct bending moment M, shear Q and
axial force N diagrams in the beam members of the composite structures shown in the following
figures.
10kN
(a)
3m (b) 10kN
10kN
10kN 10kN
5kN 1
3m
2 2 3
1 3
3m
10kN
6× 2=12m
3m
A B
1
4m P
(d)
P
2
4a
50kN 30kN
(c) 4
3
2
4m
3
6 × 3=18m 3a
1
(e)
2
4m
3
P P P P
6 × 3=18m
problem 6-4 (contd)

10kN 10kN 10kN 10kN 20kN

1
(f)
3 2
4m
6 × 4=24m
problem 6-4
1
(b)
(a) 60kN
4m
3 1
4m
2
40kN
4m
20kN 20kN 3
3m × 6=18m
3m 3m 3m 3m
1kN 3
(c)
1.5m 1.5m
(d)
1kN
2 4kN 1
3m
2kN 2kN 4kN

1 4
4 2
4m
3m
3
1kN 2kN 2kN 2kN 1kN
3m
6 × 3=18m
2m 2m 3m 3m
problem 6-5
P (b) P P P P P
(a)
A C D a c d e f g b
E
d
d
A B
F G H C D E F G
5×d 6× d
problem 6-6
20kN 30kN 40kN

(b)
(a)
1kN/m
C
3m
D
F C G E
3m
4m
A B
D E A B
3m 3m 3m 3m
4×3m =12m
problem 6-7
members of a truss may not be perfectly straight lines, and the axes of members meeting at a joint may not all
intersect at the common point. So it is almost impossible to make analysis for a truss according to its
practical stressing state. In this circumstance, trusses have to be simplified or idealized, by throwing away
some unimportant details and maintaining the very important features which is desired for analysis, as
analyzing models prior to computation. Therefore, the following assumptions are made in order to simplify
the analysis of trusses:
−σ
(a)
+σ
b
(b)
Fig.6.1 Simply supported beam and truss; (a) beam; (b) truss
600 600
3
1
120
2450
20 60
60 90 × 8 60
(a) 60
110
120 × 10 90 × 8
150 3 × 2450=7350
15000
Unit: (mm)
(b)
Fig.6.2 A reinforced concrete truss and its computing model
(1) Truss members are connected together at their ends by frictionless hinges;
(2) The centroidal axes of the members are all straight lines, those meeting at a hinge all intersect at
the centre of the hinge;
(3) External loads and reactions are all applied on the truss only at its joints
(a)
16m
128m 64m
(b)
16m
128m 64m
Fig. 6.3 The great bridge over Yangtse River in Wuhan (of China) and its computing model
Fig.6.4 A modern space truss structure
The reason for making these assumptions is to obtain an ideal truss, whose members are subjected
only to axial forces. If all the members of a truss and the applied loads lie in a single plane, the truss is
called a plane truss. The truss shown in Fig.6.6 (a) is a plane truss, Fig.6.6 (b) is its computing model which
is idealized according to above assumptions. Each of the members is represented by its centroidal axis; the
connections between members are all hinges; concentrated loads P1 , P2 and reactions YA , YB are all
applied on the joints. Fig.6.6 (c) shows an arbitrary isolated member from the truss. Since member CD is
connected at its ends by frictionless hinges [assumption (1)] with no loads applied between its ends
[assumption (3)], the member would be subjected to only two forces at its ends. Since the member is in
equilibrium, the two forces must be equal in magnitude but opposite in directions. The line of action of the
two forces must coincide with the centroidal axis of the member CD. Thereby, the arbitrary member CD is
subjected only to axial force. Because the members of a truss are subjected only to two forces they are
named two-force members. As long as two-force members with straight axial lines are subjected only to
axial forces, the internal forces in the members of a truss which satisfies above assumption are only axial
forces.
Fig.6.5 Truss members made of aluminum shapes
P2
top chord
(a) P1 vertical
C
h
A B (c)
D d
bottom chord diagonal N
C
YA YB
P2 D
(b) P1 N
C
h
A B
D d
Y YB
A
Fig.6.6 A truss and its computing model
The members of a truss are classified into two categories named as chords and web members
according to their arrangements and orientations [see Fig.6.6 (a)]. Chords indicate the members which are
located along the top and bottom periphery of the truss, including top chords and bottom chords. Web
members are arranged between top chords and bottom chords, including verticals and diagonals. The joint
at which a web member is connected to chord is called a panel point, and the length d between two panel
points on the same chord is called the panel length or simply the panel. The maximum distance between the
top chords and the bottom chords is termed as the height of the truss, denoted by alphabet h.
6.1.2 Classification of trusses by their geometric constructions
A truss is an assemblage of straight members connected at their ends by frictionless hinges to form a
stable system. The arrangement of the members of a statically determinate truss must obey the rules of
geometric construction of a stable system with no redundant restraint. According to the geometric
construction characteristics, trusses can be grouped into the following three classes.
1. Simple trusses
Beginning with the foundation or a basic triangle which is composed by three members (bars)
connected by hinges pairwise at their ends, enlarge a simple truss by attaching two new members (or links)
which do not lie on the same straight line and are pined together, or by adding a binary system, to the
foundation or the basic triangle to form a new joint. The truss can be further enlarged by repeating the same
procedure as many times as desired. Trusses constructed by this procedure are called simple trusses.
(a) (b)
E 11 1 D 7 G 13 J
F A
5 14
9 2 8
10 11
12 C I
F
7 3 9
D 6 12 15
C
B K
4 E 10 H 16
8 5 6
3
B
4 1
2
A Fig.6.7 Simple trusses
As shown in Fig.6.7 (a), beginning with the foundation, enlarge a simple truss by attaching two new
members (1, 2) to the foundation to form joint A, further enlarge the new truss by attaching in turn two new
members (3, 4), (5, 6), (7, 8) … (11, 12) to form joints B, C, D, E and F so as to construct a simple truss.
In Fig.6.7 (b), beginning with the basic triangle ABC, enlarge the triangle by attaching in turn two new
members (1, 2), (3, 4), (5, 6), (7, 8) … (15, 16) to form joints D, E, F , …, J and K so as to construct a
simple truss, which is internally stable with no redundant restraint. Finally, the truss is attached to the
foundation by one hinged support and one roller support to form a simply supported truss.
2. Compound trusses
Compound trusses are constructed by connecting two or more simple trusses to form a new truss
according to the geometric construction rule 2 or rule 3 discussed in the Chapter 2.
In Fig. 6.8, two simple trusses ABE and CDE are connected by hinge E and link BC, which does not
cross hinge E, to form an internally stable truss with no redundant restraint, and then the truss is attached to
the foundation by one hinged support and one roller support.
A D
B C
Fig.6.8 A compound truss Fig.6.9 A complex truss
N (b)
(a)
y B
N
l
x ly Y
α
α lx X
A
N
Fig.6.10 Similar relation between the dimension and member force of an inclined member
3. Complex trusses
Trusses that cannot be classified as either simple or compound are referred to as complex trusses. An
example of complex truss is shown in Fig.6.9. Sign convention of axial forces and indication of internal
forces for inclined members
It is customary to designate the tension as the positive axial force and compression as the negative. In
calculating process, it is recommended that unknown axial forces should be indicated at their locations in
their positive directions. Based on this regulation, if consequent results of the unknowns are positive they
are actually positive, this means the corresponding members are in tension; if some of the unknowns are
negative they are actually negative, their direction indication are just in the opposite direction, it means the
corresponding members are in compression. Therefore, you can avoid the ambiguity caused by directions
of unknown axial forces.
In order to facilitate the analysis of inclined members, the axial force N of an inclined member will be
resolved into its horizontal and vertical components, X and Y, in the calculation. The axial force N [see Fig.
6.10 (a)] of an inclined member AB and its horizontal and vertical force components X and Y form a
triangle shown in Fig.6.10 (b), while the length l of the inclined member AB and its horizontal and vertical
projections lx and l y forms another triangle [see Fig. 6.10 (a)]. Since the directions of axial forces in the
members of a truss coincide with their centroidal axes, the two triangles are similar to each other. There
exists the following relation:
N X Y
= = (6-1)
l lx l y
Generally, l x and l y are given, so the length l will be determined. By using Eq. (6-1), once either of
the force projections X or Y are determined another force projection and axial force N may be obtained
easily.
6.2 The Method of Joints
In the method of joints, the axial forces in the members of a plane statically determinate truss are
determined by considering the equilibrium of its joints. Since the entire truss is in equilibrium, each of its
joints must also be in equilibrium. At each joint of the truss, the member forces and any applied loads and
reactions form a coplanar concurrent force system, which must satisfy two equilibrium equations,
∑X =0 and ∑ Y = 0 , in order for the joint to be in equilibrium. These two equilibrium equations must
be satisfied at each joint of the truss. There are only two equations of equilibrium at a joint, so they cannot
be used to determine more than two unknown forces.
In its practical application, the method of joints is suitable for analysis of simple trusses. Based on the
geometric constructions of a simple truss, if we consider the joints successively in the order reverse through
that it were set up, we could proceed from one joint to another, in such a way that there are never more than
two unknown axial forces acting on the joint under consideration, to accomplish the calculation of all
unknown axial forces of the simple truss.
As the truss shown in Fig.6.11, when constructing the truss, starting from stationary joint A and B, in
turn add a binary system to fix a new joint in the order of C, D, E, F, G and H. when analyzing the truss,
the order of isolating joints should be the reverse of

above constructing order. That is, we start with joint
A 1 C 5 E 9
H at the right bottom end and proceed in succession G
3
to joints G, F and E, then to joints D and C. 2 6 10 11
Apparently, if we calculate unknown axial forces in B 7
H
this sequence, there are never more than two 4 D 8 F 12
unknown axial forces acting on the joint under P P
consideration. For each joint, we can apply two
equilibrium equations, ∑X =0 and ∑Y = 0 to Fig.6.11 A simple truss
solve two unknown axial forces, so all of the

unknown axial forces can be obtained successfully.
Example 6-1
Calculate the axial forces of every member of the truss shown in Fig.6.12 (a) by using the method of
joints.
Solution
Considering the entire truss as a free body, we write
∑ X = 0， XA = 0
∑M = 0， A YB = 12kN(↑)
∑M = 0,B YA = 12kN(↑)
2. Calculation of axial forces
The truss shown in Fig.6.12 (a) is a simple truss. When constructing the truss, starting from a triangle
BEG, in turn add a binary system to fix a new joint in the order of D, F, C and A. when analyzing the truss,
the order of isolating joints should be the reverse of the above constructing order. That is, we start with
joint A at the left end and proceed in succession to joints C, F, D and E, then to joint G. Apparently, there
are only two unknown axial forces acting on the each desired joint. For each joint, we can apply two
equilibrium equations, ∑X =0 and ∑Y = 0 to solve the two unknowns, so all of the unknown axial
forces can be obtained successfully.
(1) joint A
Taking joint A as a free body as shown in Fig.6.12 (c), depict reaction YA in its real magnitude and
direction, assume the unknown axial forces N AC and N AF being tensile forces, resolve the unknown
axial force N AC in the inclined member, whose length can be seen in Fig.6.12 (b), into components
X AC and YAC along the horizontal and vertical directions, respectively. By writing the equation of vertical
force equilibrium, we can immediately solve for YAC
9kN 6kN 9kN

C D E
D
1.5m
(a) (b) C
B 2 .5 2 .1 2
1.5
A
1.5
F G
A F
2m 1.5m 1.5m 2m 2 1 .5
YAC
(c) NAC (e) 3kN
(d) YFD
9kN NFD
XAC XFD
16kN NCD
C
A NAF 16kN F NFG
20kN NCF
12kN
12kN
9kN 6kN 9kN
(g) 6
(f) −16 −16
16 D 16
-20 -4.24 -4.24 -20
3 3 3 3
+16 +19 +16 4.24 3 3 4.24

12 unit: kN 12
Fig.6.12 The simple truss and its free bodies of example 6-1
∑Y = 0 ， Y AC + 12 = 0
YAC = −12kN
By using proportional relation, Eq. (6-1), we find
12
X AC = − × 2 = −16kN
1.5
2.5
N AC = −12 × = −20kN (compression)
1.5
By writing the equation of horizontal force equilibrium, we obtain
∑ X = 0, N AF + X AC = 0
N AF = − X AC = 16kN (tension)
(2) joint C
The free body diagram of joint C is shown in Fig.6.12 (e). In which given forces (the load and axial
force of member AC) are depicted in its real magnitude and direction, unknown axial forces N CD and
N CF are assumed to be tensile forces. Applying the equilibrium equations to the free body, we obtain
∑ X = 0, N CD + 16 = 0
N CD = −16kN (compression)
∑Y = 0 , 12 − 9 − N CF = 0
N CF = 3kN (tension)
(3) joint F
The free body diagram of joint F is shown in Fig.6.12 (d); the unknown axial force N FD in inclined
member FD, whose length can be seen in Fig.6.12 (b), is indicated by its horizontal and vertical
components X FD and YFD . By writing the equation of vertical force equilibrium, we can immediately
solve for YFD
∑Y = 0 ， YFD + 3 = 0
YFD = −3kN
By using Eq. (6-1), we find
2.12
N FD = −3 × = −4.24kN (compression)
1.5
1.5
X FD = −3 × = −3kN
1.5
By applying the equation of horizontal force equilibrium, we obtain
∑ X = 0, X FD + N FG − 16 = 0
N FG = 19kN (tension)
(4) Utilization of symmetry
Since the truss and its loads are all symmetrical the distribution of the internal forces of the truss must
be symmetrical too. So the axial forces of two members which lie on the symmetrical positions must be the
same. Therefore, the axial forces of the other portion of the truss may be determined by symmetrical
condition. The finally calculating results are shown in Fig.6.12 (f).
(5) Checking
In order to check the correctness of above calculation, apply the equilibrium conditions of joint D,
which has not been utilized so far and whose free body diagram is shown in Fig.6.12 (g). Since the
symmetry is considered the equilibrium condition ∑X =0 is automatically satisfied and the other
equilibrium condition ∑Y = 0 is needed to check as follows
∑Y = 0 , −6 + 3 + 3 = 0
The constructing sequence of the truss shown in Fig.6.12 (a) may be also considered as: starting from
a triangle ACF, in turn add a binary system to fix a new joint in the sequence of D, G, E and B. So the
sequence of isolating joints should be: starting with joint B at the right end and proceeding in succession to
joints E, G, D and C, finally to joint F.
Special cases of the method of joints—Identification of zero-force members
Members with zero forces in them are named zero-force members. Because trusses are usually
designed to support several different loading conditions, it is not uncommon to find members with zero
forces in them when a truss is being analyzed for a particular loading condition. Zero-force members are
also added to trusses to brace compression members against buckling and slender tension members against
vibrating. The analysis of trusses can be expedited if we can identify the zero-force members by inspection
instead of calculation. Three common types of member arrangements that result in zero-force members are
the following:
N1 (c)
(a) A (b) N2 N1 N2
N1 N2 N3
P
N1 = N2 = 0 N1 = N2 , N3 = 0
N 2 = 0, N1 = − P
Fig.6.13 Identification of zero-force members
1. If only two noncollinear members are connected to a joint that has no external loads or reactions
applied to it [Fig.613 (a)], then the force in both members are equal to zero. For instance, if taking the
centroidal line of member 1 as x axis of a rectangular coordinate system we can immediately obtain
N 2 = 0 by applying the equilibrium equation ∑ Y = 0 and vice versa.

2. If two noncollinear members are connected to a joint [Fig.613 (b)] and the external loads of the
joint are applied in the direction which coincides with the centroidal line of one of the two members, then
the axial force in the member whose centroidal line does not coincide with the acting-lines of the loads is
equal to zero. For example, if we take the line of action of load P and axial force N1 as x axis of a
rectangular coordinate system we can immediately find N 2 = 0 by applying the equilibrium
equation ∑ Y = 0 .
3. If three members, two of which are collinear[Fig.613 (c)], are connected to a joint that has no
external loads or reactions applied to it, then the force in the member that is not collinear is equal to zero.
For instance, if taking the connecting line of two collinear members as x axis of a rectangular coordinate
system we can immediately find N 3 = 0 by applying the equilibrium equation ∑ Y = 0 .
Example 6-2
Identify the zero-force members P P
of the truss shown in Fig.6.14. C 1 D E F 3

G
14 15
Solution 7 11
K 10 L M N 4
2 5 8
1. Starting with joint C and G, 6 9
12 13
we identify (arrangement 1) A B
H I J
N1 = N 2 = N 3 = N 4 = 0
2. Identifying joints K, L, E, M Fig.6.14 The truss of example 6-2
and N, we find (arrangement 3)

N 5 = N 6 = N 7 = N8 = N 9 = 0
3. Identifying joints H and J (arrangement 3)
N10 = N11 = 0
4. Identifying joint I
Since joint I lies on the symmetrical axis of the truss and members 12 and 13 are located at the
symmetrical position, their axial forces N12 and N13 must be equal to each other, i.e., N12 = N13 ;
again since N 7 = 0 , by the equilibrium equation ∑ Y = 0 , we find N12 + N13 = 0 , i.e., N12 = − N13 .
This conflicts with the fact of N12 = N13 , so the two axial forces must be equal to zero,
i.e., N12 = N13 = 0 .
5. Focusing our attention on joints L and M, by considering the equilibrium of the two joints, we
obtain
N14 = N15 = 0
It can be seen finally that under the action of the two concentrated loads only the members depicted
with zebralike lines are not equal to zero (see Fig.6.14).
6.3 The Method of Sections
The method of sections involves cutting the truss into two portions by passing an imaginary section
through the members whose forces are desired. The desired member forces are then determined by
considering the equilibrium of one of the two portions of the truss. Each portion of the truss is treated as a
rigid body in equilibrium, under the action of any applied loads, reactions and the forces in the members
that have been cut by the section. The unknown member forces are determined by applying the three
equations of equilibrium to one of the two portions of the truss. There are only three equilibrium equations
available, so they cannot be used to determine more than three unknown forces. Thus, in general, sections
should be chosen that do not pass through more than three nonconcurrent and nonparallel members with
unknown forces. The truss with parallel chords shown in Fig.6.15, in each panel of which there are only
three members (one top chord, one bottom chord and one diagonal). If passing an imaginary section
through any one of the panels three members will be cut and the truss will be divided into two portions, the
three unknown forces can be determined by considering the equilibrium of any one of the two portions of
the truss. For instance, if passing section Ⅰ-Ⅰ through panel BC, the unknown forces of members 1, 2 and
3 will be determined by applying equations of equilibrium ∑ M B = 0 , ∑ M C = 0 and ∑ Y = 0 to
any one of the two portions of the truss.
0.5kN 1kN 1kN 1kN 0.5kN

a b 2 Ⅰ c e
d
d
A E
B 1 Ⅰ C D
4d
Fig.6.15 A truss with parallel chords
In the process of calculation, in order to avoid solving simultaneous equations it is preferable to select
the types of equilibrium equations. The progress for applying the method of sections is now explained
through the following example.
Example 6-3
Calculate the axial forces of members 1 through 3 of the truss shown in Fig.6.16 (a) by using the
method of sections.
Solution
Applying equilibrium equations to the entire truss, we find
1
∑M A = 0 ， YB =
18
(2 × 3 + 2 × 6 + 1× 9) = 1.5kN(↑)
1
∑M B = 0, YA = (1× 9 + 2 ×12 + 2 × 15 + 1× 18) = 4.5kN(↑)
18
2kN 2kN m 1kN

(a) 1kN I
G
E J
2
C
3m
1.5m
3
A B
D F 1 mH
6 × 3m=18m
(c) 3.04 I
0.5
G
3
1kN G
(b) X2
G I 3.91
N2 2.5
Y2 J
N3 H
Y3 F 3
X3
N1 B
H
1.5kN
Fig.6.16 The truss of example 6-2
Checking: ∑Y = 0 , 1 + 2 + 2 + 1 − 4.5 − 1.5 = 0

2. Find axial forces of member 1 through 3
As shown in Fig.6.16 (b), a section m-m is passed through members 1, 2 and 3, cutting the truss into
two portions, ACGF and BJIH. For simplifying the calculation, the right-hand portion, BJIH, will be used
here to analyze the axial forces. The free-body diagram of the portion BJIH of the truss is shown in
Fig.6.16 (b). All three unknown forces, N1 , N 2 and N 3 are assumed to be tensile and indicated by
arrows pulling away from the corresponding joints on the diagram. The unknown member forces can be
determined by applying the equations of equilibrium to the free body. However, in order to avoid solving
simultaneous equations we should select the types of equilibrium equations. It is helpful to select the
intersecting point of two forces as the centre of moments to find the third unknown force, then to determine
the other two unknown forces by applying equilibrium equations of force projections.
(1) Because N 2 and N 3 pass through point G their moments about point G are equal to zero. So
we can use the equilibrium of the moments about G to calculate member force N1 as follows:
∑M G =0， N1 × 2.5 + 1× 3 − 1.5 × 12 = 0

N1 = 6kN (tension)
(2) We calculate N 2 by summing moments about point H, which is the point of intersection of the
lines of action of N1 and N 3 . In order to facilitate the calculation we resolve axial force N 2 at point I
into its horizontal and vertical components, X 2 and Y2 , along the horizontal and vertical directions; the
length of member 2 is shown in Fig.6.16 (c). Equilibrium of moments about H yields
∑M H =0, − X 2 × 3 − 1.5 × 9 = 0
X 2 = −4.5kN
By proportion relation, Eq. (6-1), we find
3.04
N 2 = −4.5 × = −4.56kN (compression)
3
0.5
Y2 = −4.5 × = −0.75kN
3
(3) After the axial forces N1 and N 2 have been determined, we may utilize the equilibrium
equation of force projection in the horizontal direction to find N 3 . By summing the force components in
horizontal direction of all forces imposed to the free body, we write
∑ X = 0, − X 2 − N1 − X 3 = 0
X 3 = − N − X 2 = −6 + 4.5 = −1.5kN
By proportion relation, Eq. (6-1), we obtain
3.91
N 3 = −1.5 × = −1.96kN (compression)
3
2.5
Y3 = −1.5 × = −1.25kN
3
(4) Checking computations
To check our computation, we apply the equilibrium equation ∑Y = 0 which has not been utilized
so far to the free-body diagram. Thus
∑Y = 0 , 1.5 − 1 + 0.75 − 1.25 = 0 (satisfied)
Note that since the truss is a simple truss the axial forces of members 1, 2 and 3 can be determined by
the method of joints. For doing so, however, the member forces connected to joints A, C, D and E have to
be involved in the calculation before joints F and G are reached that can be analyzed for members 1, 2 and
3. Therefore, the method of sections enables us to determine forces in the specific members of trusses
directly, without first calculating many unnecessary member forces, as may be required by the method of
joints.
It is instructive to emphasize that if we consider the order of calculating axial forces in the members of
a truss as the reverse order through that it were set up, we could accomplish the calculation of all unknown
axial forces without encountering any trouble. Since compound and complex trusses are constructed by
several simple trusses the method of sections is indispensable to analyzing the member forces of these
types of trusses. For instance, the truss shown in Fig.6.17 (a) is a compound truss formed by connecting
two simple trusses, AEFC and BGFD, by a common hinge F and a member CD. Thus, the order of
calculating axial forces in the members of the truss ought to be: first calculating the member force of
member CD; then the members FG and FI (or FE and FH) by using the method of sections [section aa or
bb as shown in Fig.6.17 (a)]. Once the connecting member forces of CD, FG and FI (or FE and FH) are
b a
(a) E F G (b)
2 3
H I
A B
C D
a b
1
Fig.6.17 Compound trusses
known, we can focus our attention on the simple trusses and to find corresponding member forces by the
method of joints. Apparently, following above mentioned analyzing sequence we will determine the
unknown member forces of the truss successfully. The truss shown in Fig.6.17 (b) is a compound truss as
well, after the connecting member forces of 1, 2 and 3 are determined by using the method of sections, the
whole member forces in the truss can be obtained without any difficulty.
Generally, when sections are chosen that do not pass through more than three members with unknown
forces we may directly use the equations of equilibrium for a coplanar force system to determine the
unknown forces. However, in some trusses, the arrangement of members may be such that by using
sections that pass through more than three members with unknown forces, we can determine one or, at most,
m
(a) 1 (b) m
A B
C A B
m m 1
Fig.6.18 Special trusses
two unknown forces. Such sections are employed in the analysis of only certain types of trusses. For
instance, the truss shown in Fig.6.18 (a), the section m-m cuts five members, but four of which are
concurrent to joint C. So by considering either portion of the truss (divided by section m-m) and summing
the moments about C we can find the axial force in the member (numbered 1) not crossing the joint C.
Another example is the truss shown in Fig.6.18 (b). The section m-m cuts four members, but three of which
are parallel each other. So by considering either portion of the truss (divided by section m-m) and summing
the force components in the direction perpendicular to the axial lines of the three parallel members we can
obtain the axial force in the member (numbered 1) being not parallel to the other three members.
6.4 The Combination of the Method of Joints and the Method of Sections
Although the method of joints and the method of sections described in the preceding sections can be
used individually for the analysis of trusses, the analysis of compound or some complex trusses can
sometimes be expedited by using a combination of the two methods. For some types of compound or
complex trusses, the sequential analysis of joints breaks down when a joint with more than two unknown
forces be found. In such a case, the method of sections is then employed to calculate some of the member
forces, thereby yielding a joint with two or fewer unknowns, from which the method of joints may be
continued. This approach is illustrated by the following example.
Example 6-4
Find the axial forces of members 1 through 3 of the truss shown in Fig.6.19.
6.4 The Combination of the Method of Joints and the Method of Sections 137
Solution
1. Finding reactions
Considering the whole truss as a free body, we obtain
YA = 12kN(↑), YB = 4kN(↑), XA = 0
2. Finding the axial forces of members 1, 2 and 3
(1) By using section m-m as shown in Fig.6.19 (a), the truss is cut into two portions. Taking the right
portion on the section m-m of the truss as a free body (the free-body diagram is omitted) and by summing
the moments about point C of all forces applied to the portion, we find
∑M C = 0 ， N1 × 4 − 4 × 8 = 0 , N1 = 8kN (tension)
(2) By using section n-n as shown in Fig.6.19 (a), the truss is cut into two portions. For simplifying
the analyzing, the axial force N 2 of member 2 [whose length is shown in Fig.6.19 (c)] is resolved into a
horizontal component X 2 and a vertical component Y2 . By taking the left portion on the section n-n of the
truss as a free body [see Fig. (b)] and by summing the moments about point F of all forces imposed to the
portion, we find
∑M F = 0 ， (12 − 2) × 4 − 4 × 2 − Y2 ×1 − X 2 × 2 − 8 × 2 = 0
By using the proportion relation of Eq. (6-1), we write
Y2
X 2 = 1.5 ×
2
Solving above two equations simultaneously, we obtain
Y2 = 6.4kN
Y
N 2 = 2.5 × 2 = 8kN (tension)
2
(3) With N1 and N 2 now known, there are only two unknowns, N 3 and N DH , at joint D. These
forces can be determined by applying the two equations of equilibrium to the free body of joint D, as shown
in Fig.6.19 (d). For simplifying the analyzing, the axial force N 3 of member 3 [whose length is shown in
Fig.6.19 (c)] is resolved into its horizontal and vertical components, X 3 and Y3 . By summing the force
components in the vertical direction of all forces applied to joint D, we write
∑Y = 0 ， Y 3 + 6.4 = 0
Y3 = −6.4kN
6.4
N3 = − 5 = −7.16kN (compression)
2
It can be seen from the above example that when utilizing a combination of the method of joints and
the method of sections an optimal method of the two methods may be selected according to the location of
the member whose axial force is desired.
2kN
(a) (c) G
4kN
4kN C m
n 2.5
4kN 2
F
2kN D
4m
G 1.5
3
2
A B F
H D n 1 m E
5m 3m
2 5
2m 2m 2m 2m 8m D
1
(b) 4kN (d)

4kN
6.4kN
F N2 N 3 Y3 8kN
2kN
Y2
X2 X3 4.8kN
A 8kN N DH 8kN
D D
12kN
Fig.6.19 The truss and some its free body diagrams of example 6-4
6.5 Form and Characteristics of Member Forces of Girder Trusses
Since girder trusses are widely used in practical engineering structures it is interest to discuss some of
their form and member force characteristics. In fact, there is relation between the configuration and the
member force characteristic of a truss, i.e., different configuration has different member force characteristic.
In order to make a clear understand about the argument the member force performance of three commonly
used girder trusses, which are parallel chord trusses, triangular trusses and parabolic trusses, are analyzed
under the action of top-chord-joint loads (which are the simplification of uniformly distributed loads). The
bending moment M 0 and shear Q 0 of a corresponding simply supported beam (with the same span and
subjected to the same loads) [see Fig.6.20 (b)] are utilized here to represent axial forces in a truss.
1. Parallel chord trusses [Fig.6.20 (a)]
(1) Axial forces in chords
When calculating the axial forces in chords of a parallel chord truss the method of sections may be
6.5 Form and Characteristics of Member Forces of Girder Trusses 139
used. That is, passing a section through a panel cutting the truss into two portions, by applying moment
equilibrium equations to either of the two portions and by selecting the moment centres at the intersection
points of the chords and the web
members, the axial forces of chords
will be obtained successfully. For P/2 P/2
P P P P P
example, if we are interested in the c d e 2 f g b
a
axial force in bottom chord 1 as
h
(a)
shown in Fig.6.20 (a) the joint f, the A B
C D E 1 F G
point of intersection of top chord 2
and web member Ef, should be chosen 6× d
as the moment centre of the desired

moment equation; if we desire to find P/2 P P P/2
P P P
the axial force in top chord 2 as (b) A B
shown in Fig.6.20 (a) the joint E, the C D E F G
point of intersection of bottom chord

6× d
1 and web member Ef, should be 3P 3P
chosen as the moment centre of the

M 0 diagram
desired moment equation. The arms of (c)
2.5Pd
the axial forces in chords are equal to 4.0 Pd 4.5Pd
the height h of the truss.
The formula to calculate the axial 2.5P
forces in chords of the truss may be (d) 1.5 P
0.5 P Q 0 diagram
expressed as follows
M0 0.5 P
N' = ± (6-2) 1.5 P
h 2.5P
in which, M 0 is the bending Fig.6.20 A parallel chord truss and its corresponding simple beam
moment in the corresponding simple

beam on the section corresponding to
the desired location of moment centre of the truss; h is the arm of the axial force; the positive sign is for the
axial forces in the bottom chords under the action of tensile forces; the negative sign is for those in the top
chords under the action of compressive forces. Since the height h of the parallel chord truss is a constant,
the variation of the axial forces in chords is directly proportional to that of the bending moment diagram of
the corresponding simple beam. i.e., the axial forces in the middle chords are larger than those in the end
chords of the truss, or the axial forces in the chords would decrease from the middle to the ends of the truss.
(2) Axial forces in web members

When calculating the axial forces in the diagonals and the verticals of a parallel chord truss the method
of sections may be used as well. That is, passing a section through a panel cutting the truss into two
portions, by applying vertical force projection equations of equilibrium to either of the two portions, the
axial forces may be determined successfully. The formula to calculate the vertical components in the
diagonals and the axial forces in the verticals of the truss may be written as follows
Y = ±Q 0 (6-3)
in which, Q 0 is the shear in the corresponding simple beam on the section corresponding to the location
of the desired panel of the truss; the positive sign is for the vertical components of axial forces in the
diagonals under the action of tensile forces and the negative sign is for the axial forces in the verticals for
the arrangement of the truss shown in Fig.6.20 (a). If exchanging the right-portion’s diagonal direction with
the left-portion’s diagonal direction as
shown in Fig.6.20 (a) (the vertical line
1 1 of symmetry of the truss partitions the
1 1 1 1 1
2 2 truss into two portions), the sign for
−2.5 −4.0 −4.5
axial forces in the diagonals and the
−1.0
(a)
−1.5
−2.5
−3.0
verticals must be accordingly changed

3.5
2.2
0.7
4
as well. Apparently, the variation of the

0 2.5 4.0
axial forces in web members is directly
6× d
proportional to that of the shearing
1
1 1 force diagram of the corresponding
1 .7 4
−4 1 simple beam. i.e., the axial forces in
1 2 −1
−6.3
+2.0
.8 1 the end-portion web members are

2−7.91 −1.5
h
+0.5
(b) 8 2
0 larger than those in the middle-portion
7.5 7.5 6.0
web members of the truss, or the axial
6× d forces in the web members would
increase from the middle to the ends of
1 1
1 the truss.
(c) 1 −4.53 1
1 −4.7
5 2. Triangular trusses [see
1
2 5.15 0 0 Fig.6.21 (b)]
h
− 0
0 0 2
(1) Axial forces in chords
4.5 4.5 4.5
The formula to calculate the axial
6× d
forces in chords of a triangular truss
may be still expressed as follows
Fig.6.21 Three commonly used trusses
6.5 Form and Characteristics of Member Forces of Girder Trusses 141
M0
N' = ±
r
in which, M 0 is still the bending moment in the corresponding simple beam on the section corresponding
to the desired location of the moment centre of the truss; r is the arm of the axial force in the desired chord,
which is the distance between the desired chord and the centre of the moment, the variation of r is linear
from the middle to the ends of the truss; Since the magnitude of the arm r reduces more quickly than that of
the bending moment M 0 , the variation of the axial forces in chords increase from the middle to the ends
of the truss, or the axial forces in the middle-portion chords are smaller than those in the end-portion chords
of the truss.
(2) Axial forces in web members
It is convenient to determine the axial forces in the diagonals and the verticals of a triangular truss by
means of method of sections, i.e., passing a section through a panel cutting the truss into two portions, by
applying moment (about joint at support) equations of equilibrium to either of the two portions, the axial
forces may be determined successfully. Apparently, the axial forces in the end-portion web members are
smaller than those in the middle-portion web members of the truss, and the diagonals are compressional
member and the verticals are tensile members.
3. Parabolic trusses
Since the panel points along the top chords locate on a parabolic curve, the variations of the length r of
the verticals and the coordinates of the bending moment M 0 of the corresponding simple beam obey the
same parabolic function (the formula for calculating the horizontal components of axial forces in top chords
M0
and the axial forces in bottom chords are still N = ± ) , the horizontal component of axial forces in
r
the top chords and the axial forces in the bottom chords of each panel of the truss would be identical.
Because the variations of the inclinations of the top chords are small enough the axial forces of top chords
may be considered as being nearly identical. All of the web members of the truss are the zero-force
members for the reason that the magnitudes of the horizontal component of axial force in the top chord and
the axial force in the bottom chord of each panel of the truss are equal to each other but in opposite sense
(one is tension another is compression).
In the interest of comparing the member force characteristics of the above mentioned commonly used
three types of girder trusses, the distributions of their axial forces are shown in Fig.6.21 that are obtained in
the case h = d . By analyzing the distribution, some arguments are stated as follows:
(1) Parallel chord trusses: The axial forces in the members of a parallel chord truss are nonuniform.
The axial forces in the middle chords are larger than those in the end chords of the truss. From mechanical
point of view, when constructing this type of truss the cross-sectional areas of the chords in each panel
could be different or nonuniform, which would yield difficulty to connect the members together; if we
adopt the uniform cross-sectional area for each chord member, their constructing materials would be wasted.
However, since the advantages in manufacture of parallel chord trusses, such as uniform joint structure, the
identical length of chord members and normalizing web members and so forth, they are widely applied to
engineering structures, especially to light-weighted structures for the reason that the same cross-sectional
area of chord members will not cause too much waste of constructing materials. So the parallel chord
trusses may be often appeared in long-span crane girders, supporting-structure systems and railroad bridge
structures and the like.
(2) Triangular trusses: The axial forces in the members of a triangular truss are nonuniform as well.
The axial forces in the end chords are larger than those in the middle chords of the truss, and the
inclinations of end chords are too small to manufacture. However, since the inclined configuration satisfies
the requirement of roof drainage triangular trusses are often applied to roof structures with a bit small span
and large inclination, such as wooden roof structures, light-weighted-steel roof structures and composite
(consisting of reinforced concrete inclined beams and steel tensile members) roof structures.
(3) Parabolic trusses: The axial forces in the members of a parabolic truss are by and large uniform.
So their constructing materials would be utilized more economically. However, since the inclinations of top
chord in each panel have different value the joint conformations of the truss are complicated so that it is
quite difficult to manufacture this type of truss. In practical engineering structures, parabolic trusses may be
often occurred in long-span (18m~30m) roof structures and bridge structures.
6.6 Composite Structures
6.6.1 Constitutes and types of composite structures
A composite structure is composed by two types of members which have different stressing
performance. One type of members is the members which are only subjected to axial forces, thus called
two-force members. The other type of members is the members whose dominating internal force is bending
moment and additional internal forces are shearing forces and axial forces only in some cases, which are
thusly named beam members or flexural members. Composite structures often come into your sight when
you focus your attention on the roofs of buildings in industrial and civilian architectures, crane girders and
supporting systems of bridges.
The computing model of a pentagonal below-supported roof structure is shown in Fig.6.22 (a), whose
top chords are flexural members that are made of reinforced concrete and bottom chords and web members
are all two-force members which are made of steel shapes.
Fig.6.22 (b) shows a composite structure named three-hinged rigid-frame-like structure, of which
columns are flexural members and others are two-force members.
Fig.6.22 (c) shows an analyzing model of an arched bridge. It is a composite structure composed by
stiffened beam and linked arch constructed by two-force members.
Stiffened beam
(a) (b) (c)
Fig.6.22 Three composite structures
6.6.2 Analysis of composite structures
The most fundamental point for analyzing a composite structure is to distinguish the different stressing
characters of different members. It is essential that there exists only axial force in a two-force member, but
three types of internal force components (bending moment, shear and axial force) in a flexural member. So
when using the method of sections the
unknown internal forces acting on the
desired imaginary section should be
B
coincided with the stressing characteristic (a) B C (b)
NBC
of the cut member. That is, if a beam QBC
D
member is cut by the desired imaginary NBD
section the unknown internal forces on the

(c) NDB
section must be bending moment, shear MDB NDC
and axial force in some cases; while a QDB
two-force member is cut by the desired MDA D

A
imaginary section the unknown internal QDA
forces on the section ought to be only axial NDA
force. Fig.6.23 A composite structure and some of its free body diagrams
Similarly, when taking a joint as a
free body it is essential to differentiate
whether the joint is a perfect hinged joint, the intersection point of two-force members, or a nonperfect
hinged joint, the intersection point of two-force members and flexural members. As shown in Fig.6.23 (a),
joint C is a perfect hinged joint and joint D is a nonperfect hinged joint.
It should be noted that the conclusion about the method of joints discussed in the previous section of
the chapter is based on the fact that the joint is a perfect hinged joint. For a nonperfect hinged joint formed
by flexural members and two-force members, the conclusion about the method of joints only being suitable
to trusses must not be applied to it because there exist three internal force components in flexural members.
As shown in Fig.6.23 (a), the member BDA is a flexural member, the member BD and BC, which are
connected by joint B [Fig.6.23 (b)], are not zero-force members although there is no load applied to joint B;
whereas the member DC, which is connected to joint D [Fig.6.23 (c)], is not zero-force member as well
since there exist axial forces N DC , N DB and N DA , shears QDB and QDA , the bending moments
M DB and M DA on the free-body diagram of joint D.
It is worth of emphasizing again that the order to calculate internal forces in the members of a
composite structure should be the reverse order through that it was set up. Because the flexural members
have at least two internal force components the first step to find the internal forces in the members of a
composite structure is to determine the internal forces in two-force members, and then the internal forces in
flexural members. When the desired internal forces are determined their diagrams may be constructed. The
procedure is illustrated by the following example.
Example 6-5
Construct the internal force diagrams of the composite roof structure shown in Fig.6.24.
Solution
(a) 1kN/m Ⅰ (b) YC

A B A
F C G F C XC
3m
D E
6kN
Ⅰ D NDE
3m 3m 3m 3m
(c) 1kN / m QAF

(d)
A C NCF A NAF
F 6kN
6kN 6kN
QCF
6kN 6kN 6kN
6kN
Fig.6.24 The composite structure and some of its free-body diagrams of example 6-5
(1) Find reactions

By considering the equilibrium of the whole composite structure and utilizing the symmetry of the
structure, we write
∑ X = 0， X A =0
1 1
YA = YB = ql = 1× 12 = 6kN(↑)
2 2
(2) Calculate the axial forces in two-force members
Geometric construction analysis: The structure may be considered as a system formed by two rigid
bodies ADC and BCE by using one hinge, C, and one link, DE, which does not cross the hinge. So the
structure is a stable system with no redundant restraint. Based on the geometric construction sequence of
the structure, its analyzing steps should be: first calculating the internal force in the link, member DE, and
restraint forces of hinge C; then restraint forces at joints D, E, F and G; finally the internal forces in the
beam member AFC or CGB. The approach is implemented as follows:
By using the method of sections (the imaginary section is I-I) and by applying moment equilibrium
equation about hinge C to the left portion of the section [Fig.6.24 (b)], we obtain
∑M C = 0 ， 6 × 6 − 1× 3 × 6 − N DE × 3 = 0
N DE = 6kN (tension)
By considering the equilibrium of joint D, we write
∑ X = 0， X DA = 6kN
YDA = 6kN
N DA = 6 × 2 = 8.48kN (tension)
∑Y = 0 , N DF = −6kN (compression)
(3) Internal forces of flexural members
Take beam member AFC as a free body whose diagram is shown in Fig.6.24 (c) and select sections A,
F and C as the control sections.
When focusing our attention on the equilibrium of joint A [Fig.6.24 (d)], we find
∑ X = 0， N AF +6=0
N AF = −6kN (compression)
∑ Y = 0 , QAF + 6 − 6 = 0
QAF = 0
By considering the equilibrium of flexural member AFC [Fig.6.24 (c)], we write
∑ X = 0， N CF +6=0
N CF = −6kN (compression)
∑Y = 0 , 1× 6 + 6 − 6 − 6 + QCF = 0
QCF = 0
By using the method of sections and calculating from A to F, we find
M FA = −1× 3 ×1.5 = −4.5kN ⋅ m (tension in upper fibers)
QFA = −6 + 6 − 1× 3 = −3kN ⋅ m
N FA = −6kN ⋅ m (compression)
Similarly, by using the method of sections and calculating from C to F, we find
M FC = −4.5kN ⋅ m (tension in upper fibers)
QFC = 1× 3 = 3kN ⋅ m
N FC = −6kN ⋅ m (compression)
Because the structure and its loads are all symmetrical the distribution of its internal forces must be
symmetrical too. After the internal forces of the left portion AFC of the structure are determined, those in
the right portion CGB of the structure should be determined according to the symmetry of the structure.
(4) Constructing bending moment M, shear Q and axial force N diagrams
Since the beam members AFC and CGB are subjected to uniformly distributed loads, when
constructing their internal force diagrams the superposition method may be used. The final bending
moment M, shear Q and axial force N diagrams are shown in Fig.6.25 (a) through (c).
(a) (b)
Q diagram (kN)
4.5kN ⋅ m 4.5kN ⋅ m
3 3
A B A B
3 F C 3G
−6kN
F C G
−6kN
kN
8.
6kN
48
48
(c)
kN
8.
N diagram (kN)
M diagram (kN ⋅ m) and N diagram (kN) A F C G
B
6 6
Fig.6.25 The composite structure and its internal force diagrams of example 6-5
Summary 147
SUMMARY
The chapter was mainly discussed the analysis of plane statically determinate trusses. Its key points
are as followings:
1. The member forces of a truss are only axial forces if the truss satisfies the assumptions made for
a perfect truss. So readers should understand the conditions to simplify a practical structure to a perfect
truss.
2. There exists relationship between the analyzing sequence and the order of the geometric
construction for a truss. It is instructive to emphasize that if we consider the sequence of calculating axial
forces in the members of a truss as the reverse sequence through that the truss were set up, we could
accomplish the calculation of all unknown axial forces without encountering any trouble. The trusses may
be classified into three kinds of trusses, simple trusses, compound trusses and complex trusses according to
their geometric constructions. Generally speaking, the method of joints is suitable to analyzing the member
forces of simple trusses while the method of sections will be adapted to compound and some complex
trusses.
Since the method of joints and the method of sections are two fundamental methods to analyze trusses,
they must be soundly mastered. The method of joints is most efficient when forces in all or most of the
members of a truss are desired. The method of sections usually proves to be more convenient when forces
in only a few specific members of the truss are desired.
In the interest of simplifying the analysis, the inactive members or zero-force members may be found
before starting calculation, and proportional relation, Eq. (6-1), between the member force components and
projections of member length should be neatly used.
3. The configurations will impose an influence on the axial forces in a truss. By comparing the
stressing performance of three commonly used girder trusses with their corresponding simple beams, the
distribution of member forces and structural characteristics of the trusses ought to be understood clearly.
4. A composite structure is consisting of flexural and two-force members, which must be
distinguished accurately. It is essential to differentiate whether the joint is a perfect or nonperfect hinged
joint. The order to find the internal forces in the members of a composite structure is: first the axial forces
in two-force members, then the internal forces in flexural members.
6-1 What are the fundamental assumptions for a perfect truss? Why there exist only axial forces in the
members of a truss under the three assumptions?
6-2 How many types of geometric constructions of the trusses? How classify them?
6-3 Try to deduct equation (6-1) based on the trigonometric-function relation between axial force N and
its horizontal and vertical components X and Y in an inclined member.

6-4 How using the characteristics of geometric constructions of simple trusses to find their member
forces?
6-5 How to distinguish zero-force
members?
P/2 P P P P P P/2
6-6 How to selection equilibrium a c e f g b
(a) d
equations so as to avoid solving
h
simultaneous equations when
A
using the method of sections to C D E F G B
find member forces of a truss.
l = 6× d
6-7 How using the characteristics of
geometric constructions of (b)
compound trusses to find their
h
member forces?
6-8 Comparing the differences of
member forces between the three
trusses shown in the problem (c)
figure. If the joint loads are
h
removed to the corresponding
position of the bottom chords,
Problem for reflecting Fig.6-8
what are the variations of their
member forces?
6-9 How to distinguish the two-force and non-two-force members in a composite structure? What are
their stressing differences? Which attentions should be paid when analyzing them? Whether the
members a and b in the truss shown in the problem figure are zero-force members?
6-10 What is the procedure to analyze a composite structure?
C E
D
P
a
A B
F
Problem for reflecting Fig.6-9

CHAPTER 7
GENERAL REMARKS ON
STATICALLY DETERMINATE STRUCTURES

The chapter is the summaries of the five preceding chapters treated of from the analysis of the
geometric constructions of structures to that of internal forces in statically determinate structures.
Unlike the five preceding chapters, which are classified on the types of structures, the chapter,
combining all of statically determinate structures together, will discuss the structural analysis in
following three aspects, the methods for analyzing statically determinate structures, the general
performance of statically determinate structures and the stressing characteristics of various types of
statically determinate structures, so as to enhance and deepen the comprehension about statically
determinate structures and their analytical methods.
7.1 Analytical Methods For Statically Determinate Structures
The analysis of statically determinate structures mainly treats of the determination of reactions and
internal forces existing in statically determinate structures by means of equilibrium equations, and the
constructions of the internal force diagrams. The basement of the analysis is the isolation of free-body
diagrams, on which internal forces and reactions may be emerged and to which the equilibrium equations
can be applied so as to determine the unknown forces.
Fig.7.1 (a) shows a statically determinate multispan beam, which has 9 unknown restraint forces
[Fig.7.1 (b)], and three members to which 9 available equilibrium equations are able to apply. Obviously,
the number of its restraint forces is just equal to that of available equilibrium equations. In theoretical mean,
the unknown forces may be determined as a whole. However, in order to expedite the computation and
avoid solving simultaneous equations, we should write the equilibrium equations so that each equation
involves only one unknown. Based on this purpose, in the concrete analyzing process, the sequence of
isolating the free bodies of the beam must be in the reverse order by which the beam was set up. Then apply
equilibrium equations to each free-body diagram to determine the desired unknown restraint forces. In
order to implement efficient analysis, some of the analytical key points are given as follows:
(1) Types of free bodies and unknown forces
A free body isolated from a structure may be a joint, a member, a portion or the whole of the structure.
It has been observed from previous chapters that the method of joints used for analyzing trusses isolates
155
156 Chapter 7 General Remarks on Statically Determinate Structures
joints as free-body diagrams; in analyzing statically determinate multispan beams the members are isolated
as free-body diagrams; in analyzing statically determinate rigid frames the members are often isolated as
free bodies to determine the member-end moments and shears, and rigid joints are selected to calculate
member-end axial forces; the method of sections used for analyzing trusses isolates portions of trusses as
free bodies.
On a free-body diagram there exist given forces and unknown forces. The number of unknown forces
is dominated by the kinds of restraints.
P
(a) C D
B E F
A
P
C D XD
(b) XC
YC YD
YC YD
A B E F
XA XD
C XC D
YA YB YE YF
Fig.7.1 A multispan beam and some of its free-body diagrams
Fig.7.2 shows a simple example. Since the members AC and BC in Fig.7.2 (a) are two-force members
subjected only to axial force, by isolating joint C as a free body and using the method of joints the axial
forces N1 and N 2 can be conveniently determined.
In Fig.7.2 (b), since member AC is a beam member and BC is a two-force member, by isolating
member AC as a free body and considering its equilibrium, unknown forces X A , YA and N may be
obtained successfully. In Fig.7.2 (c), because members AC and BC are all beam members, by using the
method similar to that used in three hinged arches, i.e., detaching the two members from their supports and
considering the entire system as a free body and applying its equations of equilibrium and condition
( M C = 0 ), unknown forces X A , YA and X A , YB may be found successfully.
(2) Number of equations
There exists relation between the number of equations and geometric construction of a system. As
shown in Fig.7.2 (a), there are two degrees of freedom for joint C; on the other hand, two available
equilibrium equations can be applied to the joint. In Fig.7.2 (b), member AC has three degrees of freedom,
and three available equilibrium equations are able to apply to it. While in Fig.7.2 (c), the unstable system
7.1 Analytical Methods for Statically Determinate Structures 157
ACB has four degrees of freedom, and the number of its available equilibrium equations
( ∑ X = 0, ∑ Y = 0, ∑ M = 0, M C = 0 ) is also four. Generally, for statically determinate structures, the
number of degrees of freedom of a free body is just equal to that of its available equilibrium equations.
(a) (b) (c) q

q
C P C
C
A B
A B A B
q q
C C
P
C
N
XA XA XB
N1 N2 A A B
YA YA YB
Fig.7.2 Types of free-body diagrams

(a) a joint as a free body; (b) a member as a free body; (c) a portion as a free body
However, this argument is not always true for any free body. For instance, the free-body diagram of
member CD shown in Fig.7.1 (b) indicates four unknown forces but only three available equilibrium
equations can be applied to it. The three equilibrium equations may determine the unknown forces YC ,
YD and the relation of X C = X D , but not the magnitude of X C and X D . Their magnitude will be
determined after considering the equilibrium of member DF.
(3) Simplification of analysis and sequence of isolating free bodies
When considering the equilibrium of a free body to determine its unknown forces several equilibrium
equations may be used. However, in order to expedite the computation and avoid solving simultaneous
equations, we should write the equilibrium equations so that each equation involves only one unknown. For
instance, when analyzing a truss by using the method of joints the commonly considered equilibrium
equations are force projection equilibrium equations, but moment equilibrium equations may be also
utilized to simplify the calculation; when analyzing a truss by using the method of sections the selection of
the centre of moments selected for moment equilibrium equations and the axes of force projections used to
sum force components in the directions of the axes impose very important effect on the simplification of
the calculation.
The capability of identifying the stressing characters in the members of a structure may help us to
simplify the calculation. As the structure shown in Fig.7.2 (a), if you can identify out that members AC and
BC are all two-force members before implementing the calculation you may isolate joint C as free body and
find the member forces by considering the equilibrium of joint C. If the two members are recognized as
beam members the method similar to that used to analyze three hinged arches has to be utilized to find the
internal forces in the members.
The utilization of symmetry of a structure may simplify the computation of reactions and internal
forces. Here the symmetry of a structure means that the geometry of the structure (the member arrangement
and the types of supports) is symmetrical with respect to a line, which is named the line of symmetry. It is a
fact that if a symmetrical structure is under the action of symmetrical loads its reactions and internal forces
are symmetrical too with respect to the line of symmetry. Thus, for a symmetrical structure only a half of its
unknown forces are needed to determine. A symmetrical structure is shown in Fig.7.3 (a), since the
q q
XC
C
C
f
H H H
A B A
V
l l V
V
2 2 (b)
(a)
Fig.7.3 Analysis of a symmetrical three hinged structure
geometry of the structure and the applied loading are symmetrical about its vertical center line of the
structure, its reactions will also be symmetrical with respect to the line, so only two reaction components at
one support are needed to determine. Based on the symmetry, if taking member AC as a free-body diagram
shown in Fig.7.3 (b), there is only horizontal restraint force X C but no vertical force at joint C (since
point C lies on the line of symmetry and YC is a pair of antisymmetric forces).
The most important way for simplifying the calculation of a statically determinate structure is the
sequence of isolating the free bodies, which must be the reverse order by which the structure was built up.
For a multispan beam, the sequence follows thusly in this order: first the subsidiary portions, then the
7.1 Analytical Methods for Statically Determinate Structures 159
main portion.
(a)
E F G
P1 P2
Ⅱ Ⅰ Ⅱ
A B C D
(b) YE YE
F G XG XG G
E E
XE XE
P1 YG YG P2
A B C D
YA YD
Fig.7.4 The analysis of a statically determinate multispan frame
For the analysis of a truss, generally, the zero-force members must identified by inspection before the
calculation. On above argument pertinent to isolating free bodies, for a simple truss, the sequence of
isolating the joints must be the reverse order by which the truss was constructed; for a compound truss, the
axial force of the final two-force member which is used to connect simple trusses has to be determined first
by using the method of sections, then other member forces of the truss may be obtained. It is observe that in
order to select rational analytical sequence the geometric construction analysis has to be made. In the
analysis of a structure, the essential problem is the isolation of free bodies, i.e., the problem how to
disassemble a structure; whereas in the analysis of geometric construction of a structure, the main problem
is the composition of the structure, i.e., the problem how to build a structure up. From the point of
structural analytical view, there is relation between the disassembly and the build of a structure. If the
sequence of disassembly of a structure is in the reverse order by which the structure was built up, i.e., the
sequence of isolating members as free bodies is in the reverse order in which the structure was composed
by adding the members, the analysis of the structure will be simple and successful.
As the statically determinate rigid frame shown in Fig.7.4 (a), the sequence of constructing the frame
is from the main portionⅠto the subsidiary portionⅡ, but the analytical sequence, i.e., the order of
isolating free bodies must be the reverse order of the construction as shown in Fig.7.4 (b).
7.2 General Property of Statically Determinate Structures
The statically determinate structures and statically indeterminate structures are all stable systems, the
difference between them are:
From the point of geometric-construction-analysis view, a statically determinate structure is a stable
system with no redundant restraint, while a statically indeterminate structure is a stable system with at least
more than one redundant restraint.
From the point of static-equilibrium view, the internal forces of a statically determinate structure can
be determined uniquely by their equilibrium equations, whereas those of a statically indeterminate structure
are not able to be determined by their equilibrium equations, the unique solution of the internal forces may
be determined by considering both its some deformation conditions and equilibrium conditions.
One of essential characters of a statically determinate structure is that if the solution of its unknown
internal forces of the structure is obtained only by the equilibrium equations the solution is a unique
solution. Some of the undermentioned characters will be all deducted from the character.
(1) The effect on internal forces due to temperature changes, support settlements and manufacture
errors
The temperature changes, support settlements and manufacture errors and the like will not induce
internal forces in a statically determinate structure.
As the beam shown in Fig.7.5 (a), the settlement of support B may only make the beam rotate around
hinged support A but no internal forces will be developed. The conclusion may be further explained by an
imaginary experimentation: imagine removing the support B so as for the beam to be a mechanical system
and let the beam rotate around support A an arbitrary angle until the support B reaches to a new position B ′ ,
then attach the beam to the foundation with a roller support at the new position. Apparently, in the entire
process, no internal force is yielded in the beam.
There exist construction errors in the three hinged arch shown in Fig.7.5 (b), but after attached the
wrong members to the foundation, there is no internal force caused by the errors in the arch represented by
the dashed lines.
In Fig.7.5 (c), the simple beam is assumed to undergo the action of a temperature change from the
amount of −t o C at the top and the amount of +t o C at the bottom of the beam. Since the beam can be
bent freely there is no internal force induced by the temperature change in the beam as shown in dashed
lines.
A B C′
A −t 0C B
A B
B′ + t 0C
(a) (b) (c)
Fig.7.5 Support settlement, manufacture error and temperature change
(2) Characters of local equilibrium of a statically determinate structure

Under the action of a system of external loads, if a local portion of the statically determinate structure
can maintain the equilibrium with the loads by itself, the internal forces in the other portion of the structure
will be equal to zero.
A statically determinate multispan beam is shown in Fig.7.6 (a). Since the beam AB is a stable portion
it can maintain equilibrium by itself under the action of the loads imposed on it, so there is no internal force
induced by the loads in subsidiary portion BC. As shown in Fig.7.6 (b), since the member AB of a statically
determinate truss is subjected to an arbitrary balanced force system the internal forces in all other members
except in member AB of the truss will be equal to zero.
P P
P 2
(a) (b) 2 B
P A
A B C
Fig.7.6 Character of local equilibrium of statically determinate structures
In fact, since the above mentioned internal force states have satisfied all of the equilibrium conditions
of each portion of the statically determinate structures, the solution of the unknown internal forces is its
unique solution.
It should be pointed out that there is no special requirement about a local equilibrium system, which
may be a stable or unstable system, but must keep in equilibrium under some kinds of loads. As the
statically determinate truss shown in Fig.7.7 (a), which is subjected to a pair of compressive forces equal in
magnitude but opposite in direction at the ends of the bottom chords, it is able to remain in equilibrium
only by its bottom chords (which consists of a local equilibrium system) as shown in Fig.7.7 (b), so the
internal forces in other members must be equal to zero.
(a) (b)
P P P P
Fig.7.7 Bottom chords keep in equilibrium
(3) Character of equivalence of loads on statically determinate structures

When a statically equivalent transformation of loads acting on an internally stable portion of a
statically determinate structure takes place, the difference between the internal force solutions for the two
loading cases exists only in the internally stable portion, and
the internal forces in other portions of the structure for the two
P
B loading cases are keeping identical. Here, so called equivalent
(a) loads mean that the loads which distribute differently but have
A the identical resultant.
Load P1 Internal force S1 The concentrated load P as shown in Fig.7.8 (a) is
statically equivalent to the two concentrated loads in
P magnitude of P / 2 acting on the joints A and B. If we
P 2
2 change the two acting types of loads, i.e., replace the load P
(b) B
A acting at the middle point of member AB as shown in Fig.7.8
(a) with the two loads P / 2 acting at the ends of member AB
as shown in Fig.7.8 (b), only the internal force in the member
Load P2 Internal force S 2
AB will be changed and the internal forces in the other
members will remain identical.
P
P P
2 The character of statically determinate structures under
(c) 2 B the action of equivalent loads may be further explained by the
A
character of their local equilibrium. Assume that a certain
stable portion of a statically determinate structure is separately
subjected to two types of loads P1 and P2 , which are
Fig.7.8 Character of equivalent loads equivalent to each other; the internal forces corresponding the
two types of loads are S1 and S 2 respectively [see Fig.7.8 (a)
and (b), the stable portion is member AB]. By the principle of superposition, the internal force state in the
structure ought to be S1 − S2 under the simultaneous action of P1 and − P2 . Since P1 and − P2

compose a balanced force system which is imposing only on the stable portion, the internal forces (in
magnitude of S1 − S2 ) in all other members except in the stable portion (member AB) of the structure must
be equal to zero, i.e., S1 − S 2 = 0 . Therefore, when the two types of loads P1 and P2 are imposed to the
stable portion (member AB in Fig.7.8) of the structure separately, the corresponding internal forces S1 and
S2 in all other members except in the stable portion (member AB) of the structure must be identical.
(4) Character of variable conformations of statically determinate structures
The internal forces in the other portion of a statically determinate structure will remain identical when
a stable portion of the structure changes its conformation.
P
P
(a) (b)
A B A B
P
(c) P (d)
N AB A B N AB
A B
N AB N AB P
P P P
2 2
2 2
P P P P
2 2 2 2
A N AB N AB A N AB N AB B
B
Fig.7.9 Character of variable conformations
The truss shown in Fig.7.9 (a) is assumed to change its top chord AB into a small truss as shown in
Fig.7.9 (b). Apparently, the internal forces in all other members except in member AB of the truss must
remain unchangeable. The argument may be further explained by portioning the truss into two portions,
member AB and the others, which respectively keep in equilibrium under the action of their own loads and
restraint forces and whose free-body diagrams are as shown in Fig.7.9 (c). Now assume changing the
member AB into a small truss. If the internal forces in other members and the restraint forces between the
small truss and the other portion of the truss remained unchangeable, the original equilibrium conditions
existing in the other portion would be still unchangeable; while the small truss could unaffectedly keep in
equilibrium under the action of original load and restraint forces. Therefore, this sort of internal force state
existing in the members of the truss is the actual internal force state after the variation of conformation of
the truss.
7.3 Stressing Characteristics of Various Types of Structures
From the preceding chapters, we have learned five types of structures. They are: beams, rigid frames,
arches, trusses and composite structures. However, the types of structures may be classified in different
ways. The commonly used two kinds of methods of classifying structures are adduced as follows.
First classifying method: The method classifies structures into two categories, structures with thrusts
and structures with no thrust. Beams and girder trusses belong to the former, while three hinged arches and
frames, arched trusses and some of composite structures belong to the latter.
Second classifying method: The method classifies the members of a structure into two-force
members and flexural (or beam) members. The members of a truss are all two-force members; the members
of a statically determinate multispan beam and the members of a statically determinate multispan rigid
frame are beam members; for a composite structure, some of the members belong to two-force members
and some of them are beam members.
In a two-force member, the axial force is its only internal force; there is no bending moment. Under no
bending moment state, the distribution of normal stress on the cross-sectional area of a member is uniform,
so the strength of the constructional materials of such a member is utilized in the most efficient manner.
A beam member is under the action of its bending moments. Recall from previous courses on statics
and strength of materials that the bending (normal) stress varies linearly over the depth of a beam from the
maximum compressive stress at the fiber farthest from the neutral axis on the concave side of the bent
beam to the maximum tensile stress at the outermost fiber on the convex side. The material near neutral
face of a beam member is thusly not utilized in efficient way. In order to utilize the material of a beam
member most efficiently under the bending moments, the best way is to reduce the value of the bending
moments or even eliminate the bending moments so as to make the member be under no bending moment
state. Based on the argument, we will discuss the characteristics of some types of structures as follows.
(1) In a multispan beam or an overhanging beam, the value of maximum bending moment on the
section of the middle span may be reduced by utilizing the negative bending moment on the sections of
overhanging ends as shown in Fig.7.10.
(2) In a structure with thrusts, the value of maximum bending moment may be reduced by utilizing
7.3 Stressing Characteristics of Various Types of Structures 165
the actions of the thrusts. The conclusion may be further explained through the three hinged rigid frame
shown in example 4-6, the three hinged arch shown in example 5-1 and the composite structure shown in
example 6-5.
(a) q q
(e)
C
A B
f
H = M C0 / f
M C0 = ql 2 / 8 A B
l l
2 2 l l
2 2
(b) MC0 / 6 MC0 / 6
A B
(f) C
MC0 / 6
h
0.207l 0.586l 0.207l
A B
N = M /h 0
C
C
(c)
f
M C0 / 4 M C0 / 4
A B
H =M / f 0
C
(g) M C0 / 24 f1 = 5 f /12
M C0 / 24 C
(d) C A B
M C0 / 24
f
M C0 / 6 M C0 / 6
f
M C0 / 6
f /6 f 2 = 7 f /12
A B l/4 l/4 l/4 l/4
H = M C0 / f
l l
2 2
Fig.7.10 Comparison of stressing states of different structures

(a) a simple beam; (b) an overhanging beam; (c) a three hinged triangular arch with tie;
(d) a three hinged triangular arch with eccentric top chords; (e) a three hinged arch;
(f) a parallel chord truss; (g) a composite structure with lower bracing members
(3) In a truss, the members of the truss may be subjected only to the axial forces with no bending
moment by connecting the members with frictionless hinges and proper arrangement of them and by the
load-transmission manner by joints. For three hinged arches, the arches with optimal centre lines are
subjected only to axial forces with no bending moment too. From the point of mechanical view, the state of
no bending moment is a rational stressing state for a structure. The above mentioned structures (trusses and
three hinged arches with optimal centre lines) are all rational types of structures. In addition, in a composite
structure, some of the members are stayed in the state of no bending moment as well.
In order to make a comprehensive comparison about the characteristics of various types of structures,
the values of the dominated internal forces in some types of structures, with the same span and subjected to
the same uniformly distributed loads, are shown in Fig.7.10.
The beam shown in Fig.7.10 (a) is a simple beam, of which the value of the bending moment on the
2
section C, at the middle span, is M C0 = ql . If the section of the beam is rectangular (depth indicated by
8
h), the distribution of the bending (normal) stress over the sections of the beam looks like a triangularity.
The magnitude of the resultant of the compressive stress and that of the tensile stress on the section C are
0
all equal to M C .
2h 3
The beam shown in Fig.7.10 (b) is an overhanging beam. In order to reduce the value of the positive
maximum bending moment on the section at the middle of the span, the length of the overhanging is
controlled in such a length that the value of the positive maximum bending moment and those of the
negative on the sections at the two supports are equal to each other. Based on the condition, the length of
1 0
the overhanging is 0.207l and the value of the maximum bending moment becomes MC .
6
The arch shown in Fig.7.10 (c) is a triangular three hinged arch with a tie, whose thrust is
1 0
H= M C . Since the action of the thrust the maximum value to the bending moments in top chord
f
1 0
members is decreased to M C . If the top chords become the chords as shown in Fig.7.10 (d), with an
4
1 0
eccentricity of e = f , the magnitude of the negative maximum bending moments, M C , becomes just
6 6
equal to that of the positive maximum one in the top chord members.
M C0
The arch shown in Fig.7.10 (e) is a parabolic arch, of which the value of its thrust is still . Since
f
the centre line of the arch is an optimal one, it remains in the state of no bending moment.
The truss shown in Fig.7.10 (f) is a girder truss, of which the values of the axial forces in the bottom
M C0
chords are . Since the joint loads are imposed to it the truss remains in the state of no bending
h
moment.
The structure shown in Fig.7.10 (g) is a composite structure. In order to set the value of the negative
maximum bending moments and those of the positive in the top chords are equal to each other, the rises
5 7
must be: f1 = f and f 2 = f , and now the value of the maximum bending moment in the top
12 12
1 0 M C0
chords decreased to M C and the axial force of bottom chord becomes to .
24 f
It is observed from above comparison that for the structures with the same span and subjected to the
same loads, the maximum value of the bending moments occurs in a simple beam; the bending moments in
overhanging beams, statically determinate multispan beams, three hinged rigid frames and composite
structures are quite small comparing with those in a simple beam; whereas trusses and the arches with
optimal centre lines are in the state with no bending moment. In practical engineering, thereby, simple
beams are used to short span structures; overhanging beams, statically determinate multispan beams, three
hinged rigid frames and composite structures may be applied to the structures with quite long spans. When
the span of a structure is very long trusses or arches with optimal centre lines may be selected. It should be
thusly realized that when selecting a type of structure different types of structures posses different best-span
ranges.
On the other hand, every type of structure has its own advantages and disadvantages. Although simple
beams have foregoing mentioned disadvantages some of their advantages such as being constructed easily
and facilitating application make the type of structure be widely used in practical engineering projects.
While other types of structures have some advantages they have disadvantages too. For instance, a truss has
too many members and its connections (joints) are difficult to manufacture; three hinged arches have a
strict requirement on its foundation because of existence of its thrust or need a tie to undergo the thrust, and
its curved shape increases the difficulty of construction as well.
It is observed from above statement that when selecting a type of structure one must, in all aspects not
just their stressing states, analyze and compare the characteristics of a structure before make a decision.
7-1 Discuss the composition and stressing character of the truss shown in the problem figure and find
the internal forces in members 1 and 2.
7-2 The problem figure shows two different trusses subjected to the same loads. What is the difference
between their internal forces?
P 4P P
A C B
a
1
a
2
E D F
8a
problem 7-1
1 1 1 1 1 1 1
(a)
2 2
(b) 1 1 1 1 1 1 1 1 1 1 1 1 1
4 2 2 2 2 2 2 2 2 2 2 2 4
problem 7-2
CHAPTER 8
INFLUENCE LINES

In previous chapters, we discussed the analysis of statically determinate structures subjected to
stationary loadings. In this chapter, we study the analysis of statically determinate structures
subjected to variable loads. The influence lines are the important tools and means used in the
analysis of structures subjected to variable loads.
We begin this chapter by introducing the concept of influence lines. We next discuss the two
kinds of methods, named static method and mechanismic (or virtual displacement) method, to
construct influence lines. Finally, we consider the application of influence lines to determination of
the maximum values of response functions for structures due to variable loads.
8.1 Concept of Influence Lines
In the previous chapters, we considered the analysis of structures subjected to loads whose positions
were fixed on the structures. An example of such stationary loading is the dead load due to the weight of
the structure itself and of other material and equipment permanently attached to the structure. However,
structures generally are also subjected to loads (such as moving loads and movable loads) whose positions
may vary on the structure. The hoist loads shown in Fig.8.1 (a) and (b), the carload shown in Fig.8.2 and
pedestrian loads are the examples of moving loads; while the wind and snow loads imposed on buildings
are the examples of movable loads. In this chapter, we study the analysis of statically determinate structures
subjected to moving loads and movable loads, simply called variable loads.
Under the action of variable loads, the internal forces and reactions of a structure will vary with the
variation of the loads. Consider, as an example, the bridge truss shown in Fig. 8.2. As the car moves across
the bridge, the forces in the members of the truss will vary with the position x of the car. It should be
realized that the forces in different truss members will become maximum at different positions of the car.
For example, if the force in member AB reaches its maximum value when the car is at a certain position
x = x1 , then the force in another member, for example, member FG, may reach maximum value when the
car is at a different position x = x2 . The design of each member of the truss must be based on the
maximum force that develops in that member as the car moves across the bridge. Therefore, the analysis of
the truss would involve, for each member, determining the position of the car at which the force in the
169
170 Chapter 8 Influence Lines
member becomes maximum and then computing the value of the maximum member force.
(b) P P
(a)
Beam of hoist
Beam of hoist
P P
A B
Fig.8.1 Hoist loads

(a) transverse section of the workshop; (b) longitudinal section l
of the workshop; (c) the computing model of the beam
From the foregoing discussion, we can see that the analysis of structures for variable loads consists of
two steps: (1) determining the position(s) of the load(s) at which the response function of interest (e.g., a
reaction, shear or bending moment on a section of a beam, or force in a truss member) becomes maximum
and (2) computing the
maximum value of the
response function.
x
However, there are a
variety of kinds of moving Beam Deck
loads and a great deal of A E

ways of distributions of B D
C
movable loads. It is
impossible to determine the
each most unfavorable
F G H
loading position of all Truss
variable loads which may Fig.8.2 Car loads
impose on a structure. An
8.1 Concept of Influence Lines 171
efficient way to analyze structures under the action of variable loads is the utilization of influence lines.
An influence line is a graph of a response function of a structure as a function of the position of a
downward unit load moving across the structure.
When the influence line for a desired quantity (such as a reaction, a shear, an axial force and a bending
moment) is determined, the maximum value of the quantity due to the actual variable loads may be
determined by the principle of superposition. In fact, an influence line for a desired quantity of a structure
is the graphic expression of the quantity with respect to the position of a downward unit load moving across
its possible range on the structure, which shows graphically how the movement of a unit load on the
structure influences the magnitude of the quantity; whereas the actual magnitude of the desired quantity of
the structure due to variable loads could be considered as that of the combination of a series of unit loads.
The concept of influence lines will now be further explained by the following example.
Consider the simply supported beam shown in Fig. 8.3(a). The beam is subjected to a downward
concentrated load of unit magnitude, which moves from the left end A of the beam to the right end B. The
position of the unit load is defined by the coordinate x measured from the left end A of the beam, as shown
in the figure. Suppose that we wish to draw the influence line for the vertical reaction at support B.
To develop the influence line for the
vertical reaction RB of the beam, we
x P =1
determine the expression for RB in terms
(a) A B x
of the variable position of the unit load, x,
by applying the equilibrium equation of
the moment about point A, we obtain l RB
∑M A = 0, 1× x − RB × l = 0 y = RB
1.0
(b)
x
RB = (0 ≤ x ≤ l ) (8-1) y
l x
From equation (8-1), it is obvious

that RB is a linear function of x, with Fig.8.3 Determination of influence line for RB
RB = 0 at x=0 and RB = 1 (a) a simple beam; (b) influence line for RB
at x = l .
Equation (8.1) represents the equation of the influence line for RB , which is constructed by plotting
this equation with RB as ordinate against the position of the unit load, x, as abscissa, as shown in Fig.8.3
(b). Note that this influence line [Fig.8.3 (b)] shows graphically how the movement of a unit load across the
length of the beam influences the magnitude of the reaction RB . As this influence line indicates, RB = 0
when the unit load is located at the left support A of the beam (i.e., when x = 0 ). As the unit load moves
from A to B, the magnitude of RB increases linearly until it becomes 1.0 when the unit load reaches the
right support B (i.e., when x = l ). It is important to realize that the ordinate of the influence line at any
position x is equal to the magnitude of RB due to a unit load acting at the position x on the beam. For
example, from the influence line for RB [Fig.8.3 (b)], we can determine that when a unit load is applied at
a distance of 0.25l from the end A of the beam, the magnitude of the reaction RB will be 0.25 .
Similarly, when the unit load is acting atx = 0.6l the magnitude of RB will be 0.6 , and so on.
The foregoing discussion about the influence line for reaction RB has introduced the concept of an
influence line. Generally, a graph indicating the variation of a response function such as one of reactions or
internal forces of a structure, across which a downward unit load is moving, is termed as an influence line.
For the graph, the position of the unit load, x, is abscissa, the value of the response function is ordinate.
When constructing an influence line, the base line (as abscissa) represents the possible moving range of the
unit concentrated load, and the positive value of the influence line is depicted on the upper side of the base
line and vice versa. The dimension of the ordinate is equal to the dimension of the response function
divided by the dimension of the unit load.
8.2 Influence Lines for Statically Determinate Single Span Beams by Static Method
There are two kinds of elementary methods, static method and mechanismic (or virtual displacement)
method, to construct influence lines for the reactions or internal forces of a structure. The static method is
illustrated by determining the influence lines for the reactions and internal forces of a single span beam as
follows.
8.2.1 Influence lines for simple beams
(1) Influence lines for reactions

The influence line for vertical reaction at right-hand support B of a simple beam is determined in the
preceding section [see Fig.8.3]. Now we determine and draw the influence line for vertical reaction at
left-hand support A of the simple beam.
To develop the influence line for the vertical reaction RA of the beam, we determine the expression
for RA in terms of the variable position of the unit load, x [see Fig.8.4 (a)], by applying the equilibrium
equation of the moment about point B, we write
∑M B = 0 ， RA × l − 1× (l − x) = 0
l−x
RA = (0 ≤ x ≤ l ) (8-2)
l
From equation (8-2), it can be seen that RA is a linear function of x, with RA = 1 at x = 0 and
RA = 0 at x = l . Actually, the linear function is just the influence line for the vertical reaction RA and
can be constructed by plotting this equation with RA as ordinate, being dimensionless, against the
position of the unit load, x, as abscissa, as shown in Fig.8.4 (c).
x P =1
A C B x
(a)
a b
RA RB
l
x 1
l
(b)
A B
(c) 1− x
1
l
A B
MC
A C MC B
C
(d) (e)
QC
QC
a b RB
RA
1 b
l
(f)
A B
C
a 1
a
l
ab b
(g)
l
A B
C
Fig.8.4 Influence lines for a simple beam
(a) a simple beam; (b) influence line for RB ; (c) influence line for RA ;
(d) and (e) free-body diagrams; (f) influence line for QC ; (g) influence line for MC
(2) Influence lines for shears

The influence lines for shears can be developed by employing a procedure similar to that used for
constructing the influence lines for reactions. To develop the influence line for the shear at section C of the
beam [Fig.8.4 (a)], we determine the expressions for QC . It can be seen from Fig.8.4 (a) that when the unit
load is located to the left of point C—that is, on segment AC of the beam ( 0 ≤ x ≤ a )—the shear at C can
be conveniently obtained by using the free body of the portion CB of the beam that is to the right of C
[Fig.8.4 (e)]. Considering the downward external forces and reactions acting on the portion CB inducing
positive shear on section C in accordance with the beam sign convention (Section 3.1), we write
QC = −RB (0 ≤ x ≤ a )
When the unit load P = 1 is located to the right of point C—that is, on segment CB of the beam
(a ≤ x ≤ l ) —it is simpler to determine QC by using the free body of the portion AC, which is to the left
of C [Fig.8.4 (d)]. Considering the upward external forces and reactions acting on the portion AC inducing
positive shear on section C, we determine the shear as to be
QC = RA (a ≤ x ≤ l )
Thus the equations of the influence line for QC can be written as
⎧− RB (0 ≤ x ≤ a)
QC = ⎨ (8-3)
⎩ RA (a ≤ x ≤ l )
Note that Eq. (8.3) expresses the influence line for QC in terms of the influence lines for the
reactions RA and RB . This equation indicates that the segment of the influence line for QC between
points A and C ( 0 ≤ x ≤ a ) can be obtained by multiplying the ordinates of the segment of the influence
line for RB between A and C by -1. Also, according to this equation, the segment of the influence line
for QC between points C and B ( a ≤ x ≤ l ) is the same as the segment of the influence line for RA
between the same two points. The influence line for QC thus constructed from the influence lines for RA
and RB is shown in Fig.8.4 (f). It is usually more convenient to construct the influence lines for shears and
bending moments (to be discussed subsequently) from the influence lines for reactions instead of from the
equations expressing the shear or bending moment explicitly in terms of the position of the unit load, x. If
desired, such equations for the influence line for QC in terms of x can be obtained by simply substituting
Eqs. (8.1) and (8.2) into Eq. (8.3); that is,
⎧ x
⎪⎪ − RB = − l (0 ≤ x ≤ a )
QC = ⎨ (8-4)
⎪ RA = 1 − x (a ≤ x ≤ l )
⎪⎩ l
The influence line for QC [Fig.8.4 (f)] shows that the shear at C is zero when the unit load is located
at the left support A of the beam. As the unit load moves from A to C, the shear at C decreases linearly until
it becomes − a / l when the unit load reaches just to the left of point C. As the unit load crosses point C,
the shear at C increases abruptly to 1 − a / l . It then decreases linearly as the unit load moves toward B
until it becomes zero when the unit load reaches the right support B. The ordinate of QC is dimensionless.
(3) Influence lines for bending moments
When the unit load P = 1 is located to the left of point C [Fig.8.4 (a)], the expression for the
bending moment at C can be conveniently obtained by using the free body of the portion CB of the beam to
the right of C. Considering the counterclockwise moments of the external forces and reactions acting on the
portion CB yielding positive bending moment on section C in accordance with the beam sign convention
(Section 3.1), we write
M C = RB × b (0 ≤ x ≤ a)
When the unit load P = 1 is located to the right of point C, we use the free body of the portion AC to
the left of C to determine M C . Considering the clockwise moments of the external forces and reactions
acting on the portion AC yielding positive bending moment on section C, we obtain
M C = RA × a (a ≤ x ≤ l )
The equations of the influence line for M C can be thusly written as
⎧ RB × b = RB (l − a) (0 ≤ x ≤ a)
MC = ⎨ (8-5)
⎩ RA × a = RA ( a ) (a ≤ x ≤ l )
It can be seen from Eq. (8.5) that the segment of the influence line for M C between points A and C
( 0 ≤ x ≤ a ) can be obtained by multiplying the ordinates of the segment of the influence line for RB
between A and C by ( (l − a ) ). Similarly, according to this equation the segment of the influence line for
M C between points C and B ( a ≤ x ≤ l ) can be obtained by multiplying the ordinates of the segment of
the influence line for RA between C and B by a. The influence line for M C thus constructed from the
influence lines for RA and RB is shown in Fig.8.4 (g). The equations of this influence line in terms of
the position of the unit load, x, can be obtained by substituting Eqs. (8.1) and (8.2) into Eq. (8.5); that is,
⎧ x
⎪⎪ RB × b = l (l − a ) (0 ≤ x ≤ a )
MC = ⎨ (8-6)
⎪ RA × a = (1 − x )a (a ≤ x ≤ l )
⎪⎩ l
The ordinate of M C is the dimension of length.
Although the influence line for M C [Fig. 8.4(g)] resembles, in shape, the bending moment diagram
of the beam for a concentrated load applied at point C, the influence line for bending moment has an
entirely different meaning than the bending moment diagram, and it is essential that we clearly understand
the difference between the two. A bending moment diagram denotes how the bending moment varies on all
sections along the length of a beam for a loading condition whose position is fixed on the beam, whereas an
influence line for bending moment denotes how the bending moment varies on one particular section as a
unit load P = 1 moves across the length of the beam.
It should be pointed out that the influence lines [see Fig. 8.4] for the reactions, shear, and bending
moment of the simply supported beam consist of straight-line segments. We will show in the following
sections that this argument is true for the influence lines for all response functions involving forces and
moments (e.g., reactions, shears, bending moments, and forces in truss members) for all statically
determinate structures.
8.2.2 Influence lines for overhanging beams

Now we determine and draw the influence lines for vertical reactions, RA and RB , at supports A and B
of the overhanging beam [Fig. 8.5(a)]. To develop the influence lines for the vertical reactions RA
and RB of the beam, we determine the expressions for RA and RB in terms of the variable position of the
unit load, x measured from the left support A and considered to be positive in the right direction, by
applying the equilibrium equations of the moment about supports A and B, we write
l−x
∑M B = 0, RA =
l
(−l1 ≤ x ≤ l + l2 ) (8-7a)
x
∑M A = 0, RB =
l
(−l1 ≤ x ≤ l + l2 ) (8-7b)
The two equations of influence lines for vertical reactions, RA and RB , at the supports A and B of
the overhanging beam is just the same as those for the reactions of a corresponding simple beam with the
same span [Eqs.(8-1) and (8-2)], but the moving range of the unit load P = 1 being extended. For the
simple beam, the moving range of P = 1 is 0 ≤ x ≤ l , and for the overhanging beam the range becomes
−l1 ≤ x ≤ l + l2 . The influence lines for RA and RB [shown in Fig.8.5 (b) and (c)] of the overhanging
beam, which is obtained by plotting the equations, are the same as those of the simple beam in the segment
AB; and in the overhanging portions, the influence lines may be just extended to the ends of the beam. It
should be noted that when P = 1 is located to the left of A the coordinate x becomes minus.
(2) Influence lines for shears and bending moments on segment AB
The influence lines for the shear and bending moment on section C of segment AB of the overhanging
beam [Fig.8.5 (a)] is now determined as follows.
First, we place the unit load P = 1 at a variable position x to the left of point C—that is, on segment
AC of the beam—and determine the shear and bending moment at C by using the free body of the portion
CB of the beam, which is to the right of C:
QC = − RB ⎫
⎬ (0 ≤ x ≤ a)
M C = RB ⋅ b ⎭
x P =1
(a) F A C B D E
d
a b
l1 l l2
(b) l1
l+
l 1
E
F l2
A C B
l
l2
l+
1 l
(c) F A
l1 C B E
l
b
l1 1 l
(d) E l2
l A C B
F l
1
a
a l
ab
a
l
(e)
F A E
l1 l2
b C B a
l l
Fig.8.5 Influence lines for an overhanging beam
(a) an overhanging beam; (b) influence line for RA ;
(c) influence line for RB ; (d) influence line for QC ;
(e) influence line for M C
Next, the unit load is located to the right of C—that is, on segment CB of the beam—and we use the
free body of the portion AC, which is to the left of C, to determine the shear and bending moment at C:
QC = RA ⎫
⎬ (a ≤ x ≤ l )
M C = RA ⋅ a ⎭
It can be observed from above equations that the equations of the influence lines for QC and M C
between two supports of an overhanging beam are the same as those for the shear and bending moment of a
corresponding simple beam with the same span; and if we extend the varying range of x to segment FA and
BE, the above two expressions will remain unchangeable. Therefore, the influence lines for QC and M C
on the overhanging portions of an overhanging beam may be obtained by just extending the corresponding
influence lines for QC and M C of the corresponding simple beam to the range of the overhanging
portions, as shown in Fig.8.5 (d) and (e).
(3) Influence lines for shears and bending moments on overhanging portions
Now determine the influence lines for the shear and bending moment on an arbitrary section D, QD
and M D , of the right overhanging portion of the overhanging beam [Fig.8.5 (a)].
First, we place the unit load P = 1 at a variable position x to the left of point D—that is, on segment
FD of the beam—and determine the shear and bending moment at D by using the free body of the portion
DE of the beam, which is to the right of D; since there is no load imposed on the portion we find:
QD = 0 ⎫
(−l1 ≤ x ≤ l + l2 − d )
M D = 0⎭⎬
Where d is the distance between points D and E [see Fig.8.6 (a)].
Next, the unit load is located to the right of D—that is, on segment DE of the beam—and we use Eq.
(8-7) and the free body of the portion FD, which is to the left of D, to determine the shear and bending
moment at D [see Fig.8.6(a)]:
QD = RA + RB = 1 ⎫
⎬ (l + l2 − d ≤ x ≤ l + l2 )
M D = − RA (l + l2 − d ) − RB (l2 − d ) = −[ x − (l + l2 − d )]⎭
The influence lines for QD and M D of the overhanging beam are shown in Fig.8.6 (b) and (c). As all
the ordinates of the influence lines on the beam except for the segment DE are equal to zero, this means
that the shears and bending moments may be developed when the loads only applied on the segment DE of
the beam.
Through constructing the influence lines for simple and overhanging beams we may summarize the
following procedure for determining the influence lines for statically determinate structures by using the
static method.
1. Select an origin from which the position of a moving downward concentrated unit load P = 1
will be measured. It is usually convenient to assume that the unit load moves from the left end of the
structure to the right end, with its position defined by a coordinate x measured from the left end of the
structure.
2. Determine the expressions for the reactions in terms of x by applying equations of equilibrium or
condition and construct the influence lines by plotting the expressions with the magnitudes of the reactions
as ordinate against the position x of the unit load as abscissa.
3. It is generally convenient to
x P =1
construct the influence lines for shears F
(a) D E
and bending moments by using the
influence lines for support reactions.
d
Thus, before proceeding with the
construction of an influence line for
1
shear or bending moment at a section (b)
D E
of the structure, make sure that the
D E
influence lines for all the reactions, on (c)
d
either the left or right side of the
section under consideration, are
Fig.8.6 Other influence lines for an overhanging beam
available. Otherwise, draw the required (a) an overhanging beam; (b) influence line for QD ;
(c) influence line for M D
influence lines for reactions by using
the procedure described in the previous
step. An influence line for the shear (or bending moment) on a section of the structure can be constructed as
follows:
(1) Locate the unit load P = 1 on the structure at a variable position x to the left of the section
under consideration, and determine the expression for the shear (or bending moment). If the influence lines
for all the reactions are known, it is usually convenient to use the portion of the structure to the right of the
section for determining the expression for shear (or bending moment), which will contain terms involving
only reactions. The shear (or bending moment) is recognized to be positive or negative in accordance with
the beam sign convention established in Section 3.1.
(2) Next, place the unit load P = 1 to the right of the section under consideration, and determine
the expression for the shear (or bending moment). If the influence lines for all the reactions are known, it is
usually convenient to use the portion of the structure to the left of the section for determining the desired
expression, which will contain terms involving only reactions.
(3) If the expressions for the shear (or bending moment) contain terms involving only reactions, it is
generally simpler to construct the influence line for shear (or bending moment) by combining the segments
of the reaction influence lines in accordance with these expressions. Otherwise, substitute the expressions
for the reactions into the expressions for the shear (or bending moment), and depict the resulting
expressions, which will now be in terms only of a variable position x, to obtain the influence line.
(4) Repeat step (3) until all the desired influence lines for shears and bending moments have been
determined.
8.3 Influence Lines for Girders with Floor Systems
In the previous sections, we discussed the influence lines for beams that were subjected to a moving
unit load P = 1 applied directly to the beams. In most bridges and buildings, there are some structural
members that are not subjected to live loads directly but to which the live loads are transmitted via floor
framing systems. The framing system of a bridge is shown in Fig. 8.7(a). The loads caused by vehicle and
the like are applied on beams called stringers, which are supported by floor beams, which, in turn, are
supported by the girders. Thus, any live loads (e.g., the weight of the traffic), regardless of where they are
located on the stringers and whether they are concentrated or distributed loads, are always transmitted to
the girders as concentrated loads applied at the points where the girders support the floor beams.
To illustrate the procedure for constructing influence lines for shears and bending moments in the
girders supporting bridge or building floor systems, consider the simply supported girder shown in Fig. 8.7
(a). As shown, a unit load P = 1 moves from left to right on the stringers, which are assumed to be simply
supported on the floor beams. The effect of the unit load is transmitted to the girder at points A through B,
at which the girder supports the floor beams. The points A through B are commonly referred to as panel
points, and the portions of the girder between the panel points (e.g., AC or CE) are called panels. Fig. 8.7(a)
shows the stringers resting on top of the floor beams, which rest on top of the girder. Although such
sketches are used herein to show the manner in which the load is transmitted from one structural member to
the others, in actual floor systems, members are seldom supported on top of each other, as depicted in Fig.
8.7(a). Instead, the stringers and the floor beams are usually positioned so that their top edges are even with
each other and are either lower than or at the same level as the top edges of the girders.
The characteristics and constructing method of influence lines for girders under panel-point loads are
herein planning to be explained by the influence line for bending moment M D on section D of the girder
shown in Fig.8.7 (a).
In order to determine the influence line for the bending moment M D on the section D, which is
located in the panel CE, of the girder due to panel-loads, which means that a unit load will move from the
left end A to the right end B on the stringers, we will consider three sorts of situation, the unit load P = 1 is
located to the left of point C (within stringer AC), to the right of point E (on stringers EF and FB) and
within the stringer CE.
It should be noted that regardless of where the unit load is located, the influence lines for vertical
reactions RA and RB at the supports A and B of the girder is identical as those for the reactions of a
corresponding simple beam with the same span. Rewrite them as follows:
P =1
Floor beam Stringer
(a)
A B
C D E F
d d Girder
2 2
l = 4d
x P =1
C E
2d − x x−d
YC = d YE =
d d
(b)
A B
C E
15
d 3
5 16 d
d 4
8
(c)
A B
D
5 3
yC = d yE = d
8 4
(d)
A C E B
1 2.5 1
(e)
4 2
C
A B
E
1 1.5
4 4
1
Fig.8.7 Influence lines under panel-loads
(a) framing system of a bridge; (b) manner of load transmission;
(c) influence line for M D under direct loading; (d) influence line for
M D under panel-loading; (e) influence line for QD
l − x⎫
RA =
l ⎪⎪ (0 ≤ x ≤ l )
⎬
x ⎪
RB =
l ⎪⎭
Now let us consider the first sort of situation, that is, when a moving unit load is located to the left of
point C (within stringer AC). In this situation, the equation of the influence line for bending moment M D
on section D of the girder may be expressed as [reference to Eq. (8-6)]
3d x 3d
M D = RB (l − ) = (l − ) (0 ≤ x ≤ d )
2 l 2
Next, consider the second sort of situation, that is, a moving unit load is located to the right of the
point E (on stringer EF and FB). In the situation, the bending moment at D is given by
3d x 3d
M D = RA = (1 − ) (2d ≤ x ≤ l )
2 l 2
Then, consider the third sort of situation, that is, a moving unit load is located within stringer CE, as
shown in Fig. 8.7 (a). In this situation, the restraint forces YC and YE , which are exerted on the girder by
the floor beam at C and E will be
2d − x
YC =
d
x−d
YE =
d
The force YC must be included in the expression for bending moment at D, so M D is given by
3d d d 1
M D = RA − YC = + x (d ≤ x ≤ 2d )
2 2 2 8
Thus, the equations of the influence line for M D can be written as
⎧ 5d x 3d
⎪ RB × = (l − ) (0 ≤ x ≤ d )
2 l 2
⎪
⎪ 3d d d 1
M D = ⎨ RA − YC = + x (d ≤ x ≤ 2d ) (8-8)
⎪ 2 2 2 8
⎪ 3d x 3d
⎪ RA = (1 − ) (2d ≤ x ≤ l )
⎩ 2 l 2
The influence line, which is obtained by plotting the equation, is shown in Fig. 8.7(d). For comparison,
Fig.8.7 (c) shows the influence line for the bending moment M D , which may be constructed by using the
procedure discussed in subsection 8.2.1, of the girder to which a moving unit load P = 1 is applied
directly. It can be seen from Fig.8.7 (d) that the influence line for M D consists of three straight-line
segments; the two side straight-line segments, on segments AC and EB, are identical as those of the
influence line for M D shown in Fig.8.7 (c). The only difference between the two figures [Fig.8.7 (c) and
Fig.8.7 (d)] occurs within the panel CE containing the location of the response function M D under
consideration; and the influence line for the response function within the panel is a straight line as well.
Founded on the analysis, we summarize the procedure for constructing influence lines for girders due to
panel-point loads as follows:
(1) Construct the influence line (plotted by dashed lines) for desired quantity of a girder to which a
moving unit load P = 1 is applied directly.
(2) Complete the influence line by connecting the all of previously computed ordinates of the panel
points between two adjacent panel points by straight lines (plotted by solid lines) and by determining any
remaining ordinates by using the geometry of the influence line.
By the procedure, the influence line for shear OD on section D of the girder shown in Fig.8.7 (a) is
constructed as follows:
First, construct the influence line (plotted by dashed lines) for OD of the girder to which a moving
unit load P = 1 is applied directly, as shown in Fig.8.7 (e).
Then, complete the influence line OD by connecting the all of previously computed ordinates of the
panel points between two adjacent panel points by straight lines (plotted by solid lines) as shown in Fig.8.7
(e).
In order to explain the characteristics of influence lines for shears of the girders due to panel-point
loads. We write the equation of influence line for shear OD as follows (which may be obtained by using
the similar procedure utilized in determining the equation of M D ):
⎧ x
⎪ -RB = − (0 ≤ x ≤ d )
l
⎪
⎪ 3
QD = ⎨ RA − YC = −1 + x ( d ≤ x ≤ 2d ) (8-9)
⎪ 4d
⎪ x
⎪ RA = 1 − (2d ≤ x ≤ l )
⎩ l
It is observed that the expressions for shear OD do not depend on the exact location of a point within
the panel; that is, these expressions remain the same for all points located within the panel CE. The
expressions do not change because the loads are transmitted to the girder at the panel points only; therefore,
the shear in any panel of the girder remains constant throughout the length of that panel. Thus for girders
with floor systems, the influence lines for shears are usually constructed for panels rather than for specific
points along the girders.
8.4 Influence Lines for Trusses by Using Static Method
The floor framing systems commonly used to transmit live loads to trusses are similar to those used
for the girders discussed in the preceding section. Fig. 8.8 shows a typical floor system of a truss bridge.
The deck of the bridge rests on stringers that are supported by floor beams, which, in turn, are connected at
their ends to the joints on the bottom chords of the two longitudinal parallel chord trusses. Thus, any live
loads (e.g., the weight of the traffic), regardless of where they are located on the deck and whether they are
concentrated or distributed loads, are always transmitted to the trusses as concentrated loads applied at the
joints. Live loads are transmitted to the roof trusses in a similar manner. As in the case of the girder floor
systems [Fig.8.8 (c)], the stringers of the floor systems of the trusses are assumed to be simply supported at
their ends on the adjacent floor beams [Fig.8.8 (b)]. Thus, the influence lines for trusses also contain
straight-line segments between panel points.
To illustrate the construction of influence lines for trusses by using static method based on method of
joints and method of sections, consider the parallel chord bridge truss shown in Fig. 8.8(a). A unit load
P = 1 moves from left to right on the stringers [see Fig.8.8 (b)] of a floor system attached to the bottom
chord AB of the truss. The effect of the unit load is transmitted to the truss at joints (or panel points) A
through B, where the floor beams are connected to the truss. Suppose that we wish to draw the influence
lines for the vertical reactions at supports A and B and for the axial forces in members de, DE, dE, cC and
eE of the truss.
The equations of the influence lines for the vertical reactions, RA and RB , can be determined by applying
the equilibrium equations:
l − x⎫
RA =
l ⎪⎪ (0 ≤ x ≤ l )
⎬
x ⎪
RB =
l ⎪⎭
The influence lines are not plotted out. But note that these influence lines are identical to those for the
reactions of a corresponding simple beam to which the unit load is applied directly.
(2) Influence line for axial force in top chord member de
The expressions for the axial force N de can be determined by passing an imaginary sectionⅠ-Ⅰ
through the members de, dE, and DE, as shown in Fig. 8.8(a), and by placing the unit load to the left of
joint E and by applying the moment equilibrium equation ∑ ME = 0 for the right portion EB, we obtain
the following expressions for N de .
∑M E = 0 ， RB × 3d + N de × h = 0
8.4 Influence Lines for Trusses by Using Static Method 185
3d 3d x
N de = − RB = − (0 ≤ x ≤ 3d )
h h l
Ⅱ Truss Floor beams

d Ⅰ e f g b
(a) a c
(b)
h
A B
C Ⅱ D E F G
Ⅰ
P=1 Stringer
l = 6d Truss
RA RB
C D E F G
(c) E
B (d) A B
A
3d 3d
h 3d h
2h
4d 1 1
3h 2
(e) D
(f) A B
A B E
D
1
3 1
1
1
6
(g) A D C
C B (h) A B
2 5
3 6
(i) D E F
A B (j) A B
1
Fig. 8.8 Influence lines for truss
(a) a parallel chord truss; (b) floor system of the truss bridge; (c) a substitute girder; (d) influence line for N de ;
(e) influence line for N DE ;(f) influence line for YdE ; (g) influence line for N cC (through truss); (h) influence
line for N cC (deck truss); (i) influence line for N eE (through truss); (j) influence line for N eE (deck truss)
This indicates that the segment of the influence line for N de between A and E can be obtained by
multiplying the corresponding the influence line for RB of the segment by −3d / h . When the unit load is
located to the right of joint E, it is convenient to determine N de by using the free body of the left portion AD
∑M E = 0 ， RA × 3d + N de × h = 0
3d 3d x
N de = − RA = − (1 − ) (3d ≤ x ≤ l = 6d )
h h l
This indicates that the segment of the influence line for N de between E and B can be obtained by
multiplying the corresponding influence line for RA of the segment by −3d / h . The influence line for
N de thusly constructed from the influence lines for RA and RB is shown in Fig.8.8 (d).
(3) Influence line for axial force in bottom chord member DE
By considering the same sectionⅠ-Ⅰused for N de but by applying the moment equilibrium
equation ∑ M d = 0 , and by placing the unit load first to the left and then to the right of joint D, we obtain
the following expressions for N DE :
∑M d = 0 ， RB × 4d − N DE × h = 0
4d 4d x
N DE = RB = (0 ≤ x ≤ 2d )
h h l
∑M d = 0 ， − RA × 2d + N DE × h = 0
2d 2d x
N DE = RA = (1 − ) (2d ≤ x ≤ l = 6d )
h h l
The influence line for N DE thusly obtained is shown in Fig.8.8 (e).
(4) Influence line for force in diagonal member dE
The expressions for N dE can be determined by considering again sectionⅠ-Ⅰand by applying the
equilibrium equation ∑Y = 0 to one of the two portions of the truss. When the unit load is located to the
left of joint D, application of the equilibrium equation ∑Y = 0 to the right portion EB of the truss yields
h
∑Y = 0 ， N dE
ldE
= YdE = − RB (0 ≤ x ≤ 2d )
When the load P = 1 is located to the right of joint E, we write

h
∑Y = 0 ， N dE
ldE
= YdE = RA (3d ≤ x ≤ l = 6d )
In which, ldE and YdE represent the length of the diagonal dE and the vertical component of N dE
respectively. So the segments of the influence line for the vertical component of N dE between A and D
and between E and B thus constructed from the influence lines for RB and RA , respectively, are shown in
Fig. 8.8(f). The ordinates at D and E are then connected by a straight line to complete the influence line for
YdE (reference to the procedure for constructing influence lines for girders due to panel-point loads in
section 8.3), as shown in the figure.
8.5 Virtual Displacement Method for Constructing Influence Lines
for Statically Determinate Beams 187
(5) Influence line for force in vertical member cC
By considering the sectionⅡ-Ⅱand by applying the equilibrium equation ∑ Y = 0 to one of the two
portions of the truss, and by placing the unit load first to the left of joint C and considering the right portion
DB of the truss, we obtain the following expressions for N cC :
∑Y = 0 ， N cC = RB (0 ≤ x ≤ d )
When the load P = 1 is located to the right of joint D, application of the equilibrium equation ∑ Y = 0
to the left portion AC of the truss yields
∑Y = 0 ， N cC = − RA (2d ≤ x ≤ l = 6d )
The segments of the influence line for N cC between A and C and between D and B thusly constructed
from the influence lines for RB and RA , respectively, are shown in Fig. 8.8(g). The ordinates at C and D
are then connected by a straight line to complete the influence line for N cC as shown in the figure.
(6) Influence line for force in vertical member eE
The influence line for N eE can be constructed by considering the equilibrium of joint e. By applying
the equilibrium equation ∑Y = 0 to the free body of joint e, we determine that N eE is zero when the
load P = 1 is located at joints A through B. The influence line for N eE coincides with the base line as
shown in Fig.8.8 (i).
If the load P = 1 is placed to top panel points, the influence lines for chord and diagonal members
will remain unchangeable but those for vertical members will have variation.
For example, when the load P = 1 is located to the top panel points, the ordinates of the segment cb
of the influence line for N cC , obtained by using sectionⅡ-Ⅱand placing the unit load to the right of joint
c and applying the equilibrium equation ∑Y = 0 to the left portion, will become − RA and the
influence line is shown in Fig.8.8 (h). If the load P = 1 moves on the top panel points and when the unit
load reaches to the joint e, the ordinate of influence line for N eE is equal to -1, whereas the ordinates at a
through d and b through f are zero. The influence line for N eE , obtained by connecting these ordinates by
straight lines, is shown in Fig. 8.8(j). As this influence line indicates, the force in member N eE will be
nonzero only when the unit load is located in the panels de and ef of the truss.
It is observed from the above that there is difference between the influence lines for deck trusses, of
which the moving unit load is applied on the top panel points, and those for through trusses, of which the
moving unit load is applied on the bottom panel points.
8.5 Virtual Displacement Method for Constructing Influence Lines for Statically
Determinate Beams
In the preceding section, we have discussed the method for constructing influence lines for statically
determinate structures. In fact, the construction of influence lines for the response functions involving
forces and moments can be considerably expedited by applying virtual displacement method, a procedure
developed by Heinrich Müller-Breslau in 1886. The procedure, which is commonly known as
Müller-Breslau's principle, is based on the principle of virtual displacements for rigid bodies.
8.5.1 Principle of virtual displacements for rigid bodies
The principle of virtual displacements for rigid bodies (be further discussed in the next chapter) can be
stated as follows:
If a rigid body is in equilibrium under a system of forces and if it is subjected to any infinitesimal
virtual rigid-body displacement compatible with its restraints, the virtual work done by the external forces
is zero.
The term virtual simply indicates that the
infinitesimal displacement is not caused by the
P1 P2
forces but by other causation. Or you may
(a)
consider that the virtual displacement is an
P1 P2 imaginary displacement, not a real one.
(b) Consider the simple beam shown in Fig.
C1 R1 Δ1 8.9(a). The free-body diagram of the beam is
Δ2 R2
(c) C2 shown in Fig.8.9 (b), in which forces P1 ,
P2 , R1 and R2 form a balanced force
system. Now, suppose that the beam is given
Fig. 8.9 Principle of virtual work an arbitrary infinitesimal virtual rigid-body
(a) a simple beam; (b) initial equilibrium position; displacement (support settlement) from its
(c) virtual displacement position
initial equilibrium position [Fig.8.9 (b)] to
another position [Fig.8.9 (c)], as shown in Fig.
8.9(c). As the beam undergoes the virtual displacement from initial equilibrium position to virtual
displacement position, the forces acting on it perform work, which is called virtual work. The total virtual
work, We , performed by the external forces acting on the beam can be expressed as the sum of the
virtual work done by each force; that is,
We = P1Δ1 + P2 Δ 2 − R1c1 − R2 c2 = 0 (8-10)
To prove the validity of the principle of virtual displacement, consider the condition that in Fig.8.9, set
P2 = 0 , c2 = 0 and place P1 at the middle span of the beam. The equation (8-10) will become:
c P
P1 1 − R1c1 = 0, R1 = 1
2 2
In this case, equation (8-10) becomes one of equilibrium equations of the simple beam.
For general cases, the equation (8-10) may be expressed as:
∑ PΔ + ∑ R
i i c =0
K K (8-11)
in which, Pi represents the loads applied on the system, Δi represents the virtual displacements
corresponding to Pi (when its direction coincides with that of Pi the product Pi Δi is positive and vice
versa), while cK represents the virtual displacements that is consistent with the given restraint condition,
RK represents the restraint forces corresponding to cK (when its sense coincide with that of cK the
product of RK cK is positive and vice versa). In Fig.8.9 (b) and (c), the virtual restraint displacements cK
are support settlements, the corresponding restraint forces RK are support reactions.
8.5.2 Mechanismic (or virtual displacement) method
The construction of influence lines for the response functions involving forces and moments can be
considerably expedited by applying mechanismic (or virtual displacement) method. The procedure of the
method can be stated as follows:
The influence line for a force (or moment) response function is given by the deflected shape of the
released structure obtained by removing the restraint corresponding to the response function from the
original structure and by giving the released structure a unit displacement (or rotation) at the location and
in the positive direction of the response function.
This procedure is valid only for the influence lines for response functions involving forces and
moments (e.g., reactions, shears, bending moments, or forces in truss members), and it does not apply to
the influence lines for deflections.
To prove the validity of the method of virtual displacement, consider the simply supported beam
subjected to a moving unit load, as shown in Fig. 8.10(a). The influence lines for the vertical reactions at
supports A and B of this beam were developed in the previous section by applying the equations of
equilibrium (see Fig. 8.4). We now draw the influence lines for the same two response functions by using
method of virtual displacement.
To construct the influence line for the vertical reaction RA , we remove the restraint corresponding to
RA by replacing the hinged support at A by a roller support, which can exert only a horizontal restraint to
A, as shown in Fig. 8.10(b). Note that point A of the beam is now free to displace in the direction of RA .
Although the restraint corresponding to RA has been removed, the reaction RA still acts on the beam,
which remains in equilibrium in the horizontal position (shown by the solid lines in the figure) under the
action of the unit load and the reactions RA and RB . Next, a virtual unit displacement δ = 1 in the
positive direction of RA is given to point A of the released beam as shown by the dashed lines in Fig.
8.10(b). Note that the pattern of virtual displacement applied is consistent with the support conditions of the
released beam; that is, points A and B cannot move in the horizontal and vertical directions, respectively.
Also, since the original beam is statically determinate, removal of one restraint from it reduces it to a
statically unstable beam. Thus, the released beam remains straight (i.e., it does not bend) during the virtual
displacement. Since the beam is in equilibrium, according to the principle of virtual displacements for rigid
bodies, the virtual work done by the real external forces acting through the virtual displacements must be
x P =1
(a) A B
P =1
(b) δ =1
y
A B
x
RA
(c)
1
A B
P =1
x
δ =1
(d) y
A B
RB
(e) 1
A B
Fig.8.10 Influence lines for a simple beam

(a) a simple beam; (b) deflecte d shape diagram for influence line R A ;
(c) influence line for R A ; (d) d eflected shape diagram for influenc e
line R B ; (e) influence line for R B
zero; that is,

RA ×1 − 1× y = 0
from which
RA = y (8-12)
where y represents the displacement of the point of application of the unit load, as shown in Fig. 8.10(b).
Equation (8-12) indicates that the displacement y of the beam at any position x is equal to the magnitude
of RA due to a unit load acting at the position x of the beam. Thus, the displacement y at any position x is
equal to the ordinate of the influence line for RA at that position. Since the unit load P = 1 is moving from
A to B, equation (8-12) expresses the deflected shape diagram of the beam, it is the influence line diagram
as well. For simplicity, draw it in Fig.8.10 (c), which is the same as that shown in Fig.8.4 (c) obtained by
static method.
The influence line for the vertical reaction RB may be determined in a similar manner, as shown in
Fig. 8.10(d) and (e).
Through above analysis, we summarize the procedure for determining the reaction and internal force
influence lines for statically determinate structures by using mechanismic (or virtual displacement) method
as follows:
1. From the given structure, remove the restraint corresponding to the response function whose
influence line is desired, and impose a corresponding force(s) to obtain the released structure.
2. Apply a unit infinitesimal displacement to the released structure at the location and in the positive
direction of the response function. Draw a deflected shape diagram of the released structure that is
consistent with the support and continuity conditions of the released structure to obtain the general shape of
the influence line. (Remember that the influence lines for statically determinate structures consist only of
straight-line segments.)
3. By using the geometric relationship in deflected shape diagram and the given value of δ = 1 , we
can directly determine the numerical values of all the remaining ordinates of the influence line.
4. The ordinates located above the base line are positive and those below are negative.
8.5.3 Influence lines for simple beams by using mechanismic method
The following example will further explain the construction of influence lines for the response
functions involving shears and bending moments by using the mechanismic (or virtual displacement)
method.
Example 8-1
Construct the influence lines for the bending moment and shear at section C of the simple beam shown
in Fig.8.11 (a).
Solution
(1) Influence line for bending moment M C

To construct the influence line for the bending moment M C , we remove the restraint corresponding
to M C by inserting a hinge at C and impose a pair of couples M C at C, as shown in Fig. 8.11(b). The
bending moment is assumed to be positive in accordance with the beam sign convention. The portions AC
and CB of the released beam are now free to rotate relative to each other. Next, a virtual unit rotation, θ = 1 ,
(a) x P =1
A B
C
a b
l
B1
yC
θ =1 b
(b) θ1 y θ2 B
A
C
MC MC
ab
l
1 b
(c)
1 C2
QC
(d) Δ2
C B
A Δ =1
y
QC Δ1
C1 1
(e) 1
b
l
a
l 1
Fig.8.11 Influence lines for simple beam by mechanismic method

(a) a simple beam; (b) deflected shape diagram for influence line for M C ; (c) influence line
for M C ; (d) deflected shape diagram for influence line for QC ; (e) influence line for QC
is introduced at C [Fig. 8.11(b)] by rotating portion AC by θ1 counterclockwise and portion CB by
θ 2 clockwise, so that θ1 + θ 2 = θ = 1 . Applying the principle of virtual displacements, we write
We = M C × θ1 + M C × θ 2 − 1× y = M C (θ1 + θ 2 ) − 1× y = M C (1) − 1( y ) = 0
From which
MC = y
This indicates that the deflected shape of the beam [Fig. 8.11(b)] is the influence line for M C . The value
of the ordinate yC can be directly determined by using the geometric relation in the deflected shape
diagram. By applying the geometric relation θ1 + θ 2 = 1 , we obtain
ab
yC =
l
The influence line diagram is shown in Fig. 8.11(c).
(2) Influence line for shear QC
To construct the influence line for the shear QC at section C of the beam, we remove the restraint
corresponding to QC by replacing the rigid connection at C by a two-vertical-roller connection, as shown
in Fig. 8.11(d). Note that sections C of the portions AC and CB of the released beam are now free to
displace vertically relative to each other. Next, a virtual unit relative displacement, Δ = 1 , in the positive
direction of QC is given at C of the released beam [Fig. 8.11(d)] by moving the end C of the portion AC
downward by Δ1 and the end C of the portion CB upward by Δ 2 , so that Δ1 + Δ 2 = Δ = 1 . The values
of Δ1 and Δ 2 depend on the requirement that the segments AC1 and BC2 in the displaced position must
be parallel to each other. Applying the principle of virtual displacements, we write
We = QC × Δ1 + QC × Δ 2 ± 1× y
= QC (Δ1 + Δ 2 ) ± 1× y
= QC (1) ± 1( y ) = 0
From which
QC = ∓ y
The sign minus or plus before y are, respectively, applied to portions AC or CB. The above equation
indicates that the deflected shape diagram of the beam [Fig. 8.11(d)] is the influence line for QC . The
values of the ordinates Δ1 and Δ 2 can be directly determined by using the geometric relation existing in
the deflected shape diagram. Thusly, we obtain
a b
Δ1 = , Δ2 =
l l
The influence line diagram for QC is shown in Fig.8.11 (e).
8.5.4 Influence lines for statically determinate multispan beams by using mechanismic method
It is very convenient to construct influence lines for statically determinate multispan beams by using
mechanismic (or virtual displacement) method. The procedure of constructing the influence lines for the
multispan beam is the same as that of constructing the influence lines for simple beams introduced in the
preceding section. The only notation is the deflection shapes of the released structures from which the
restraints corresponding to the desired response functions were removed. Generally, a statically determinate
multispan beam is composed of a stable portion and subsidiary portion(s). It should be realized that which
portion(s) would develop virtual displacement(s) and which one(s) would not yield virtual displacement(s)
after the restraint(s) of the multispan beam was removed and a virtual displacement was given. If the
influence line of a desired response function is for the force (reactions and internal forces) in the subsidiary
portion(s) of a multispan beam a virtual displacement corresponding to the desired function will not make
the stable portion of the beam develop displacement and the virtual displacement will occur only on the
subsidiary portion(s). However, the virtual displacement corresponding to the force(s), for which the
influence line(s) is desired, of the stable portion of a multispan beam will make the entire multispan beam
develop virtual movement. The following example will explain the argument in detail.
Example 8-2
Construct the influence lines for the desired forces M B , QF , RB , QC and M G of the multispan
beam shown in Fig.8.12 (a).
Solution
(1) Influence line for M B

To determine the general shape of the influence line for M B , we remove the restraint corresponding
to M B by replacing the rigid joint at B by a hinged joint that allows relative rotation between the left
portion of B and the right portion of B, and impose a pair of couple M B as shown in Fig.8.12 (b) to
obtain a released structure (a mechanism).
Next, a unit relative virtual rotation θ = 1 in the positive sense of M B is given to hinged joint B of
the released structure, and a deflected shape diagram of the beam is drawn as shown in Fig.8.12 (b). Note
that the deflected shape is consistent with the support and continuity conditions of the released structure;
the portion AB, which is a stable portion, remains its original position; also, points B and D are attached to
the roller supports, therefore, they cannot displace in the vertical direction; thus, the portion BC rotates
about B and portion CDE rotates about D as shown in the figure.
Then, by using the geometric relation in the deflected shape diagram of the released structure, the
numerical values of the ordinate of the influence line can be determined. Since the virtual displacement is
infinitesimal, θ = 1 is small enough to let the length of CC ′ be 1× 2m and the length of EE ′ 1m .
Hence, the values of the ordinate of the influence line at C and E are 2m and 1m respectively. The influence
line diagram is shown in Fig.8.12 (c).
(a) P =1
F B C G D E
A
3m 3m 2m 2m 2m 2m
(b)
MB E′
C D
A
θ =1 E
B
C′
(c) E′
C D 1m
A
B E
2m
(d) C′
1
QF
1
E′
A B
Δ =1
2 F C D 1
1 1 E6
QF C′
3m 2 3m 2m 3 2m 2m 2m
(e) F ′′
1
C E′ 1
A 2
1 B D E 6
2 1C ′
F′
3
Fig.8.12 (contd.)
(a) a statically determinate multispan beam; (b) deflected shape of M B ; (c) influence line
for M B ; (d) deflected shape of QF ; (e) influence line for QF
(2) Influence line for QF

To remove the restraint corresponding to the shear at section F, we cut the given multispan beam at F
to obtain the released structure shown in Fig.8.12 (d). Next, a infinitesimal virtual relative displacement,
Δ = 1 , in the positive direction of the shear is given to the released structure to obtain the deflected shape
diagram or the shape of the influence line for QF shown in the figure. Note that the deflected shape is
consistent with the support and continuity conditions of the released structure. Since point A is attached to
the hinged support and points B and D are attached to the roller supports, the portions AF, FBC and CDE
must rotate about A, B and D respectively. Also, the rigid portions AF and FBC must remain parallel to
each other in the displaced configuration. Then, by using the geometric relation in the deflected shape
diagram shown in Fig.8.12 (d), the numerical values of the influence-line ordinates adjacent to F are firstly
determined, and then the ordinate values at C and E can be fined successively. The influence line diagram
for QC is shown in Fig.8.12 (e).
(3) Influence line for RB
The roller support at B is removed from the given beam, and an infinitesimal virtual displacement,
Δ = 1 in the positive direction of RB , is applied at B to obtain the shape of the influence line shown in
Fig 8.12(f). Because of the hinged support at A and the roller support at D, the rigid portions ABC and CDE
of the released beam must rotate about A and D respectively. By realizing that the influence-line ordinate at
B is equal to 1, the ordinates at C and E are then determined from the geometric relation of the influence
line. The influence line thusly obtained is shown in Fig. 8.12(f).
C′
B′
(f) 4
A
1 3 E′
2
B C D
3
E
C′
(g)
1 E′
A 1
B C D
E 2
D′
G′
(h) θ =1 2m
1m
E′
A
B C G D 1m
Fig.8.12 Constructing influence lines for a statically determinate multispan beam

(f) influence line for RB ; (g) influence line for QC ; (h) influence line for M G
8.6 Application of Influence Lines 197
(4) Influence line for QC

The restraint at C is removed from the given multispan beam, and an infinitesimal virtual
displacement, Δ = 1 in the positive direction of QC , is applied at C to obtain the shape of the influence
line shown in Fig 8.12(g). Because of the portion ABC of the released structure forms a stable overhanging
beam, it neither translates nor rotates as a rigid system; and the rigid portion CDE of the released beam can
only rotate about D. By realizing that the influence-line ordinate at C is equal to 1, the ordinate at E is then
determined from the geometric relation of the influence line. The influence line thusly obtained is shown in
Fig. 8.12(g).
(5) Influence line for M G
An internal hinge is inserted in the given multispan beam at section G, and an infinitesimal relative
rotation θ = 1 is applied at G to obtain the shape of the influence line shown in Fig. 8.12(h). It can be
seen from the figure that the length of DD′ is 1× 2m=2m , the ordinates at G and E are then determined
from the geometric relation of the influence line. The influence line is shown in Fig. 8.12(h).
8.6 Application of Influence Lines
In the preceding sections, we have learned how to construct influence lines for various response
functions of structures. In this section, we consider the application of influence lines in determining the
maximum values of the response functions at particular locations in the structures due to variable loads.
8.6.1 Response at a particular location due to moving concentrated loads
1. Response at a particular location due to a single moving concentrated load

As discussed in the preceding sections, each ordinate of an influence line gives the value of the
response function due to a single concentrated load of unit magnitude placed on the structure at the location
of that ordinate. Thus we can state the following:
(1) The value of a response function due to any single concentrated load can be obtained by
multiplying the magnitude of the load by the ordinate of the response function influence line at the position
of the load.
(2) To determine the maximum positive value of a response function due to a single moving
concentrated load, the load must be placed at the location of the maximum positive ordinate of the response
function influence line, while to determine the maximum negative value of the response function, the load
must be placed at the location of the maximum negative ordinate of the influence line.
Consider, for example, an overhanging beam subjected to a moving concentrated load of magnitude P,
as shown in Fig. 8.13(a). Suppose that we wish to determine the bending moment at B when the load P is
located at a distance x from the left support A. The influence line for M B , given in Fig. 8.13(b), has an
ordinate y at the position of the load P, indicating that a unit load P = 1 placed at that position of P causes
a bending moment M B = y . By the principle of superposition, the load of magnitude P must cause a
bending moment at B, whose value is P times as large as that caused by the load of unit magnitude. Thus,
the bending moment at B due to the load P is M B = Py .
(a) A B C D
yB
(b) y
D
A B x yD
C
P
(c) D
A C
B
P
(d)
A B C
D
Fig.8.13 Response due to a single moving load

(a) an overhanging beam; (b) influence line for M B ; (c) position of load for
maximum positive M B ; (d) position of load for maximum negative M B
Next, suppose that our objective is to determine the maximum positive and the maximum negative
bending moments at B due to the load P. From the influence line for M B [Fig. 8.13 (b)], we observe that
the maximum positive and the maximum negative influence-line ordinates occur at points B and D,
respectively. Therefore, to obtain the maximum positive bending moment at B, we place the load P at point
B, as shown in Fig. 8.13 (c), and compute the magnitude of the maximum positive bending moment as
M B = PyB , where yB is the influence-line ordinate at B [Fig. 8.13(b)]. Similarly, to obtain the maximum
negative bending moment at B, we place the load P at point D, as shown in Fig. 8.13 (d), and compute the
magnitude of the maximum negative bending moment as M B = − PyD .
2. Response at a particular location due to a series of moving concentrated loads

As mentioned in Section 1.5, live loads due to vehicular traffic on highway and railway bridges are
represented by a series of moving concentrated loads with specified spacing between the loads. Influence
lines provide a convenient means of analyzing structures subjected to such moving loads. In the following,
we will discuss how the influence line for a
response function can be used to determine the
value of the response function for a given P1 P2 P3
(a)
position of a series of concentrated loads.
A B
Consider, for example, a simple beam C
shown in Fig. 8.14(a). Suppose that we wish to a b
determine the shear at section C of the beam due l
to the three concentrated loads P1 , P2 and P3 , as
shown in the figure. The influence line for the b
shear at C is also shown in Fig. 8.14(b). The (b) 1 l y2
y3
influence line for QC , given in Fig. 8.14(b), has
the ordinates y1 , y2 and y3 at the positions of y1 1
a
the loads P1 , P2 and P3 respectively, indicating l
that a unit load P = 1 placed at those positions Fig.8.14 Response due to a series of moving loads
of P1 , P2 and P3 cause shears ( QC )1 = y1 ,
(a) a simple beam; (b) influence line for QC
( QC )2 = y2 and ( QC )3 = y3 respectively.
Because the principle of superposition is valid, the load of magnitude P1 , P2 and P3 must cause shears at C,
whose values are P1 , P2 and P3 times as large as those caused by the load of unit magnitude respectively,
i.e., P1 y1 , P2 y2 and P3 y3 . Thus, the shear at C due to the three concentrated loads P1 , P2 and P3
are
QC = P1 y1 + P2 y2 + P3 y3
Generally, suppose that we wish to determine the value of a desired response function (reactions or
internal forces), Z, at a particular location due to a given position of a series of concentrated loads P1 ,
P2 ,…, Pn . The ordinates of the influence line for Z at those positions of P1 , P2 ,…, Pn are y1 , y2 ,…, yn
respectively. Thus, the value of the response function Z is the algebraic sum of a series of product
P1 y1 , P2 y2 ,…, Pn yn . That is,
n
Z = P1 y1 + P2 y2 + " + Pn yn = ∑ Pi yi (8-13)
i =1
3. Response due to a uniformly distributed live load at a particular location

Influence lines can also be employed to determine the values of response functions of structures due to
distributed loads. Consider, for example, a simple beam subjected to a uniformly distributed live load of
intensity q , as shown in Fig. 8.15(a). Suppose that we wish to determine the shear at C when the load is
placed on the beam, from x = d to x = e , as shown in the figure. The influence line for QC is also
given in the figure. By treating the distributed load applied over a differential length dx of the beam as a
concentrated load of magnitude dP = qdx , as shown in the figure, we can express the shear at C due to
the load dP as
dQC = (dP) y = (qdx) y (8-14)
Where y is the influence-line ordinate at x, where is the point of application of dP , as shown in the figure.
To determine the total shear at C due to the distributed load from x = d to x = e , we integrate equation
(8-14) between these limits to obtain
e e e
QC = ∫ (qdx) ⋅ y = q ∫ ydx = q ∫ da = qA (8-15)
d d d
e
in which A = ∫d
ydx represents the area under
e
the segment of the influence line, which
d
(a) q corresponds to the loaded portion of the beam.
A B This area is shown as a shaded area on the
C D E influence-line for QC in Fig. 8.15(b).
x dx Generally, the equation for calculating the
a b value of a desired response function (reactions or
l
internal forces) Z due to a uniformly distributed
load at a particular location should be
(b)
b y A Z = qA (8-16)
l
Equation (8-16) indicates that under the action of
a dA a uniformly distributed load the value of the
l
desired function Z at a particular location of a
Fig.8.15 Response due to a uniformly distributed load
(a) a simple beam; (b) influence line for QC structure is equal to the load-intensity q times
the area A of the influence line for the response
function corresponding to the loaded range of the
structure. Note that the sign of A must be taken into consideration when applying Eq. (8-16).
Example 8-3
For the simple beam shown in Fig.8.16 (a), determine the magnitude of shear QC at section C due to
a uniformly distributed load with the intensity of 10kN/m applied on the whole span of the beam by
using the influence line for QC , which has

constructed as shown in Fig.8.16 (b). 10kN/m
(a) A B
Solution
C
It is observed from Fig.8.16 (b) that the 4m 6m
positive area A1 and negative area A2 of

3
the influence line QC may be easily
5
determined as (b) A B
1 3 C
A1 = × 6 × = 1.8m 2
2 5 5
1 2 Fig.8.16 due to a uniformly distributed load
A2 = × 4 × (− ) = −0.8m (a) a simple beam; (b) influence line for QC
2 5
By using equation (8-16), we obtain
QC = q ( A1 + A2 ) = 10 × (1.8 − 0.8) = 10kN
8.6.2 The most unfavorable arrangement of uniformly distributed moveable loads
The most unfavorable arrangement of a uniformly distributed moveable load, which may be
distributed arbitrarily on a structure, means the arrangement of the uniformly distributed moveable load, by
which the value of a desired response function of the structure will reach to the positive (or negative)
maximum.
(a)
A B C D
q q
(b)
A B C D
q
(c)
A B C D
Fig.8.17 The most unfavorable arrangement of uniformly distributed moveable loads

(a) influence line for Z ; (b) unfavorable position for Z max ; (c) unfavorable position for Z min
For an arbitrarily uniformly distributed moveable load, equation (8-16) also indicates that the desired
response function of a structure will be maximum positive if the uniformly distributed load is placed over
all those portions of the structure where the influence-line ordinates are positive and vice versa. From Fig.
8.17(a), we can see that the ordinates of the influence line for Z are positive between the points A and B and
between the points C and D, and negative between points B and C. Therefore, to obtain the maximum
positive Z max , we place the uniformly distributed load q over the portions AB and CD of the beam, as
shown in Fig. 8.17(b). Similarly, to obtain the maximum negative Z min , we place the load over the portion
BC of the beam, as shown in Fig. 8.17(c).
Example 8-4
Determining the maximum positive and negative shears on section C of the simple beam shown in
Fig.8.16 (a) due to a uniformly distributed moveable load with the intensity of10kN/m , which may be
arranged arbitrarily over the span of the beam by using the influence line for QC , which has constructed as
shown in Fig.8.16 (b)
Solution
From Fig.8.16 (b), we observe that the ordinates of the influence line for QC are positive between
the points C and B. Therefore, to obtain the maximum positive QC max due to the uniformly distributed
moveable load 10 kN/m , we place the uniformly distributed load over the portions CB of the beam and
compute the maximum positive value of QC max due to this load by multiplying the load intensity by the
area under the portion CB of the influence line. Thus
Qc max = qA+ = 10 × 1.8 = 18kN
Similarly, since the ordinates of the influence line for QC are negative between the points A and C, the
load must arranged over the portion AC to obtain the maximum negative shear at C. By multiplying the
load intensity by the area under the portion AC of the influence line, we obtain
Qc min = qA− = 10 × (−0.8) = −8kN
8.6.3 The most unfavorable arrangement of a series of moving concentrated loads
The most unfavorable arrangement of a series of moving concentrated loads herein means the
arrangement of the loads, by which the value of the desired response function for a structure will reach to
the maximum positive (or negative). Influence lines can also be used for determining the maximum values
of response functions at particular locations of structures due to a series of moving concentrated loads.
Consider, for example, the beam shown in Fig. 8.18(a). Suppose that our objective is to determine the
maximum positive shear at point B due to the series of four concentrated loads shown in the figure. The
influence line for QB is shown in Fig. 8.18(b). Assuming that the load series moves from right to left on
the beam, we can observe from these figures that as the series moves from the end C of the beam toward
point B, the shear at B increases continuously as the ordinates of the influence line under the loads increase.
The shear at B reaches a relative maximum when the first load of the series, the load 8kN, reaches just to
the right of B, where the maximum positive ordinate of the influence line is located. As the load 8kN
crosses point B, the shear at B decreases abruptly by an amount equal to −8(0.6 + 0.4) = −8kN . With the
series of loads continuing to move toward the left, QB increases again and it reaches another relative
maximum when the second load of the series, the load 10kN, reaches just to the right of B, and so on.
Because QB becomes a relative maximum whenever one of the loads of the series reaches the maximum
8kN 10kN 15kN

5kN
B C
A
1.5m 1m 1.5m
4m 6m (a)
0.6
1 10
A (b) Influence line for QB
10 B C
1 0.4
8kN 10kN15kN 5kN
A B C (c) Loading position 1

1.5m 1m 1.5m
4m 6m
10kN 15kN
8kN 5kN
A B C (d) Loading position 2
1.5m 1m 1.5m
4m 6m
Fig.8.18 The most unfavorable arrangement of a series moving loads
positive influence-line ordinate, we can conclude that during the movement of the series of loads across the
entire length of the beam, the (absolute) maximum shear at B occurs when one of the loads of the series is
staying at the location of the maximum positive ordinate of the influence line for QB . Since it is not
possible to identify by inspection which load placed at the maximum influence-line ordinate will cause the
maximum positive QB , we use a trial-and- error procedure to determine the value of the maximum
positive shear at B. As shown in Fig. 8.18(c), the series of loads is initially positioned on the beam, with its
first load, the load 8kN, placed just to the right of B, where the maximum positive ordinate of the influence
line is located. Noting that the slope of the influence-line segment for the portion BC is 1:10 [Fig. 8.18(b)],
we compute the value QB for this loading position as
1 1 1 1
( QB )1 = 8 × 6 × + 10 × 4.5 × + 15 × 3.5 × + 5 × 2 × = 15.55kN
10 10 10 10
Next, the entire series of loads is moved to the left by 1.5m to place the second load of the series, the
load 10kN, at the location of the maximum positive ordinate of the influence line, as shown in Fig. 8.18(d).
The shear at B for this loading position is given by
1 1 1 1
( QB )2 = −8 × 2.5 × + 10 × 6 × + 15 × 5 × + 5 × 3.5 × = 13.25kN
10 10 10 10
By inspection, no other loading positions could cause larger positive values of QB than that obtained
from the above two loading positions.
By comparing the values of QB determined from the above two loading positions, we conclude that
the maximum positive shear at B occurs for the first loading position—that is, when the load 8kN is placed
just to the right of B [Fig. 8.18(c)].
Maximum positive shear at B:
( QB )max = 15.55kN
From forgoing discussion, we may summarize that the determination of the most unfavorable
arrangement of a series of moving concentrated loads or the determination of the most unfavorable position
of a series of moving concentrated loads will follow the following procedure.
1. Determination of the critical loading positions on which the values of a desired response function,
Z, of a structure will reach to extreme values (positive or negative).
2. By comparing the magnitudes of the extreme values determined for all the critical loading
positions considered, obtain the maximum positive (or negative) value(s) of the response function so as to
determine the most unfavorable position of a series of moving concentrated loads.
The characteristics and the principle of determination of the critical loading positions will be further
explained by using one of polygonal influence lines as follows.
Consider, for example, a series of moving loads is shown in Fig.8.19 (a). The influence line for a
desired response function Z of a structure is shown in Fig.8.19 (b), of which the shape is polygonal.
Although the influence-line ordinates corresponding to the loads can be obtained by using the properties of
the similar triangles formed by the influence line, it is usually convenient to evaluate such an ordinate by
multiplying the slope of the segment of the influence line where the load is located by the distance of the
load from the point at which the influence line changes its slope. The sign (plus or minus) of the ordinate is
obtained by inspection. Therefore, the angles with respect to the horizontal coordinate axis x of each of the
straight lines of the polygon [Fig.8.19 (b)] are represented by α1 , α 2 and α 3 ( α1 and α 2 are
positive, α 3 is negative). The resultants of the loads applied over each segment corresponding to each
straight line of the influence line are represented by R1 , R2 and R3 [Fig.8.19 (a)]. y1 , y2 and y3
[Fig.8.19 (b)] are the ordinates of the influence line at positions corresponding to P1 , P2 and P3 , while
y1 , y2 and y3 [Fig.8.19 (b)] are the ordinates of the influence line at positions corresponding to
R1 , R2 and R3 .
For any arrangement of the loads, we may evaluate the magnitude of Z by adding the values of the
response function contributed by the individual loads together. Thus
Z = P1 y1 + P2 y2 + " + P6 y6
Since the influence line for quantity Z is composed of individual straight lines, corresponding to each
straight line the magnitude of Z may be expressed by the contribution of each resultant of the loads applied
on that segment. For example, corresponding to firs segment, we find
P1 y1 + P2 y2 = R1 y1
R1 R2 R3
(a)
P1 P2 P3 P4 P5 P6
α3 < 0
(b) α2 > 0
y2
y1
y2 y3
y1
α1 > 0 x
Δx Δx Δx
Fig.8.19 Determination of critical loading position for polygonal influence line

(a) a given position of a series of moving concentrated loads; (b) influence line for Z
By the principle of superposition, we may express the magnitude of Z as the sum of all contributions
of each resultant of the loads as follows
3
Z = R1 y1 + R2 y2 + R3 y3 = ∑ Ri yi (a)
i =1
It is evident that since the shape of the influence line for quantity Z is a polygon the response function
Z must be a linear piecewise function with respect to x. With the knowledge of mathematics, we know that
if the quantity Z reaches to one of the extreme values at a loading position of the series, the increment of Z,
ΔZ , must less and equal to zero with respect to the infinitesimal increment Δx (right or left), around the
position of the loads [see Fig.8.20]. That is, the ΔZ must satisfy
ΔZ ≤ 0 (b)
from equation (a), we write
ΔZ = ∑ Ri Δyi (c)
Suppose that all the concentrated loads are moved an infinitesimal distance Δx (positive to right),
the line of action of Ri is moved the same distance Δx as well. From the geometry, the increment of the
ordinate of yi must be
Δyi = Δx tan α i
Thus, the increment of Z is
ΔZ = Δx∑ Ri tan αi
ΔZ
Z Z =0
ΔZ ΔZ ΔZ Δx ΔZ
>0 <0 >0 <0
Δx Δx Δx Δx
Z = f ( x) Z = f ( x)
Extreme Z Extreme Z
O x O
(a) (b) x
Fig.8.20 The character of critical position of loads

(a) extreme point; (b) extreme domain
Therefore, when the quantity Z reaches to one of the positive extreme values the critical position of the
loads must satisfy the following condition. That is,
Δx∑ Ri tan α i ≤ 0 (d)
The condition (d) has two situations:

when the loads move slightly to the right, Δx > 0, ∑ R tan α
i i ≤ 0 ⎪⎫
⎬ (8-17a)
when the loads move slightly to the left, Δx < 0, ∑ R tan α
i i ≥ 0 ⎪⎭
Similarly, when the quantity Z reaches to one of the negative extreme values the critical position of the
loads must satisfy the condition as follows:

when the loads move slightly to the right, Δx > 0, ∑ R tan α
i i ≥ 0 ⎪⎫
⎬ (8-17b)
when the loads move slightly to the left, Δx < 0, ∑ R tan α
i i ≤ 0 ⎪⎭
Equation (8-17) indicates that when Z reaches an extreme value the sum of ∑ Ri tan αi must change
its sign with respect to Δx , a given infinitesimal increment of the position of the loads. In other words, the
response function Z reaches extreme values when a series of moving loads moves to their critical positions.
Now we will analyze when the sum of ∑ Ri tan αi changes its sign. It is observed from forgoing
sections that the shape of the influence line for a desired response function Z of a statically determinate
structure is composed of straight lines. The slope, tan α i of any straight line, must be a constant. Only
when the value of one of the resultant Ri changes its magnitude that has the possibility to make the sum
∑ Ri tan α i change its sign. Next, in order to make the resultant Ri of the loads change its magnitude,
one of the series of moving loads must move through a position where the shape of the influence line for Z
occurs a vertex. When the series of moving loads moves to the left, the load moves to the left of the
position corresponding to the vertex as well, so that the resultant Ri of the loads and the sum ∑ Ri tan α i
maybe change its magnitude and sign respectively; and vice versa. The load, which makes the sum
∑ Ri tan αi change its sign when it moves to the left or to the right of the position of a vertex of the
influence line for Z, is termed as a critical load, indicated by Pcr ; while the loading position (i.e., putting
Pcr at the vertex) is called as critical position of the loads.
When the influence line for a desired response function, Z, is composed of two straight lines as shown
in Fig.8.21 (b), the critical loading position may be expressed more simply. The case in which the desired
quantity Z reaches to its peak value is now discussed as follows.
By using equation (8-17a), now i = 1, 2; α1 = α , α 2 = − β , R1 and R2 represent the resultants
of loads applied on the left portion and the right portion (segments a and segment b) respectively. In order
to make the sum ∑ Ri tan α i change its sign one of the series of moving loads, Pcr , must be placed at the
position where the ordinate of the influence line reaches to its peak value. Rleft and Rright represent the
resultants of loads to the left of Pcr and to the right of Pcr respectively.
From Eq. (8-17a), when the loads move to the right, we write
R1 tan α − R2 tan β ≤ 0
Rleft tan α − ( Pcr + Rright ) tan β ≤ 0
c c
Rleft − ( Pcr + Rright ) ≤ 0
a b
thus, we obtain
Rleft Pcr + Rright
≤
a b
Rleft Rright
(a)
Pcr
P1 P2 Pn
(b) c
α β x
a b
Fig.8.21 Critical position of loads for influence line with a triangular shape
(a) critical position of loads; (b) influence line with a triangular shape
When the loads move to the left, the equation (8-17a) will be written as
R1 tan α − R2 tan β ≥ 0
( Rleft + Pcr ) tan α − Rright tan β ≥ 0
c c
( Rleft + Pcr ) − Rright ≥ 0
a b
thus, we obtain
Rleft + Pcr Rright
≥
a b
Therefore, for the influence line with a triangular shape, the necessary condition for the critical
loading position of a series moving loads may be expressed as
Rleft Pcr + Rright ⎫
≤ ⎪⎪
a b
⎬ (8-18)
Rleft + Pcr Rright ⎪
≥
a b ⎭⎪
Equation (8-18) indicates that the character of the critical loading position for the influence line with a
triangular shape is that one of the series moving loads, Pcr , is located at the position where the ordinate of
the influence line reaches to its peak value, and when Pcr is added to the left portion the load intensity of
the left portion will be greater than that of the right portion and vice versa.
Example 8-5
For the crane beam and its loadings shown in Fig.8.22 (a), find the most unfavorable loading position
of the loads and the maximum bending moment M C at section C. P1 = P2 = P3 = P4 = 82kN
Solution
P1 P2 P3 P4
3.5m 1.5m 3.5m
(a)
P1 = P2 = P3 = P4 = 82kN
A B
C
3.6m 8.4m
12m
2.52m
2.07m
(b) 0.07m 1.02m
A B
P1 P2 P3 P4
(c) C
P1 P2 P3 P4
Fig.8.22 Determination of M C max due to the pressure of crane wheels

(a) crane beam and its wheel pressures; (b) influence line for M C ;
(c) the most unfavorable loading position
(1) Construct the influence line for M C as shown in Fig.8.22 (b).

(2) The possible critical load Pcr may be P2 or P3 . First, consider P2 as a critical load and put
P2 at the position where the ordinate of the influence line reaches to its peak value. From equation (8-18),
we write
82 + 82 82 + 82
The load move slightly to the left, >
3.6 8.4
82 82 + 82 + 82
The load move slightly to the right, <
3.6 8.4
It is observed that when P2 is added to the left portion the load intensity of the left portion will be greater
than that of the right portion and vice versa. So P2 is one of critical loads.
Next, consider P3 as a critical load and put P3 at the location where the ordinate of the influence line
reaches to its peak value. Note that the P1 load has moved off the span of the beam. By using equation
(8-18), we obtain
82 + 82 82
3.6 8.4
82 82 + 82
The load move slightly to the right, >
3.6 8.4
Apparently, when P3 is added to no matter what portion the load intensity do not satisfy equation
(8-18). So P3 is not a critical load.
(3) From above analysis, when concentrated load P2 is placed at point C, which is the most
unfavorable loading position for M C [Fig.8.22 (c)], the bending moment M C reaches to its maximum
value. By principle of superposition, the magnitude of M C is given by
( M C )max = 82 × (0.07 + 2.52 + 2.07 + 1.02) = 465.76kN ⋅ m
Example 8-6
Determine the maximum axial force N1 of member 1 of the girder truss shown in Fig.8.23 (a) due to
the wheel loads of a group of trucks and find the most unfavorable loading position of the loads.
Solution
(1) Construct influence line for N1 as shown in Fig. 8.23(b).

(2) First, suppose the group of trucks move from right to left as shown in Fig.8.23 (a), now there are
only five concentrated loads applied on the truss. Consider the third load (numbered from the
left), P = 120kN , as Pcr and put it at joint C. From equation (8-18), we write
60 + 120 + 120 70 + 130
18 27
60 + 120 120 + 70 + 130
The load move slightly to the right, <
18 27
It is observed that when the third load is added to the right portion the load intensity of the right
portion will be greater than that of the left portion and vice versa. So the third load (numbered from the left)
is one of critical loads and the axial force of member 1 corresponding to the arrangement of loads is given
by:
( N1 )max1 = 60 ×1.26 + 120 ×1.66 + 120 ×1.80 + 70 ×1.13 + 130 × 0.87 = 682.8kN
60kN 120kN 120kN 70kN 130kN

10m
4m 1.4m
(a) 4m
h = 6m
A B
1
4d = 18m 6d = 27m
45m
(b) 1.8
1.66
1.26 1.13
0.87
A B
C
Fig.8.23 (contd.)
(a) girder truss and trucks moving towards left; (b) influence line for N1 ;
(3) Next, suppose the group of trucks move from left to right as shown in Fig.8.23 (c), now there
are seven concentrated loads applied on the truss. Consider the fourth load [numbered from the right, see
Fig.8.23 (c)], P = 120kN , as Pcr and put it at joint C. The equation (8-18) is written by
130 + 70 + 120 + 120 60 + 130 + 70
18 27
130 + 70 + 120 120 + 60 + 130 + 70
The load move slightly to the right, >
18 27
It is evident that the arrangement of the loads is not a critical loading position.
(4) By comparing the results of above calculation for the two loading positions, we conclude that
the magnitude of the maximum axial force that develops in member 1 is ( N1 )max = 682.8kN tension.
(5) This maximum force occurs when the third load (numbered from the left) of the series (move
from right to left) is placed at joint C of the truss, as shown in Fig. 8.23(b).
130kN 70kN 120kN 120kN 60kN 130kN 70kN

10m 15m
4m 1.4m 4m 4m
(c)
C
6m
A B
1
18m 27m
(d)
1.8
1.66 1.53
0.66 0.53
A 0.26 0.27 B
C
Fig.8.23 N1 due to the wheel loads of moving trucks

(c) girder truss and trucks moving towards right; (d) influence line for N1
SUMMARY
The influence lines and their applications for reactions and internal forces of plane statically
determinate beams and trusses are discussed in this chapter. The key points are as followings:
1. It is essential that we clearly understand the difference between an internal force diagram and an
influence line for an internal force at a given position. An internal force diagram denotes how the internal
force varies on all sections along the length of a member for a loading condition whose position is fixed on
the member, whereas an influence line for the internal force at the given position denotes how the internal
force varies as a unit load P = 1 moves across the length of the member
2. There are two kinds of methods for constructing influence lines, static (or equilibrium) method
and mechanismic (or virtual displacement) method. If you use static method to construct influence lines for
structures, it is generally convenient to employ the influence lines for their support reactions. Thus, before
proceeding with the construction of an influence line for shear or bending moment at a point on the
structure, make sure that the influence lines for all the reactions, on either the left or right side of the point
under consideration, are available. Otherwise, draw the required influence lines for reactions by using
equilibrium method.
3. Note that all influence lines for a statically determinate structure are composed of straight lines.
4. One of the characteristics of the influence lines for some structural members that are not
subjected to live loads directly but to which the live loads are transmitted via panel points is that the
influence line between adjacent panel points is a straight line. By using this feature, the influence lines for
statically determinate beams and trusses may be constructed conveniently.
5. The construction of influence lines for the response functions involving forces and moments can
be considerably expedited by applying mechanismic (or virtual displacement) method. The key steps of the
method is: (1) from the given structure remove the restraint and impose corresponding force, corresponding
to the response function whose influence line is desired, to obtain the released structure; (2) apply a unit
infinitesimal displacement to the released structure at the location and in the positive direction of the
response function, and draw a deflected shape of the released structure to obtain the general shape of the
influence line; (3) by using the geometric relations of the deflected shape diagram and the given unit
infinitesimal displacement δ = 1 , the numerical value of the influence-line ordinate cam be determined; (4)
the ordinates above the base line are positive and those below are negative.
6. We discussed two aspects of the application of the influence lines. One is the determination of the
magnitude of a desired response function by using its influence line. Another utilization of the influence
line is the determination of the most unfavorable loading position of a series of moving loads, at which the
desired response function will reach its maximum value.
8-1 What is the definition of an influence line? Why take a downward concentrated load of unit
magnitude as the foundation of the construction of an influence line? What mechanical meanings the
abscissa x and ordinate y of an influence line represent? What is the difference between an internal
force diagram and an influence line?
8-2 If the beam shown in Fig.8.3 (a) is subjected to a concentrated load P , whether or not be able to use
the influence line for RB shown in Fig.8.3 (b) to determine the reactions due to the concentrated
load?
8-3 Try to compare the difference between two static methods (first draw free-body diagram and then
use equilibrium equation; directly write out the equations of the desired quantity by the method of
sections) used to determine the influence line for a desired quantity Z?
8-4 When write the function of an influence line for the force of a beam why sometime the function of
an influence line is defined in the entire span of a beam and why sometimes the function of an
influence line must be defined piecewise along the length of a beam?

8-5 Try to explain the meaning of the ordinate of the influence line at the point C for the bending
moment at section C of a simple beam and the difference between the influence line and the bending
moment diagram of the beam due to the action of the unit load P = 1 applied at the point C.
8-6 Try to explain the reason why the influence line for shear at section C is parallel between the left
portion and the right one of section C, and why the influence line for shear at section C of a simple
beam has an abrupt change and what is the meanings of the ordinates corresponding to the left and
the right of the abrupt change.
8-7 Please simply state the procedure for constructing the influence line for a desired quantity Z due to
the live loads transmitted via panel points.
8-8 What is the character of an influence line due to a unit load transmitted via panel points?
8-9 What is the character of constructing the influence line for trusses by using static method?
8-10 When constructing the influence lines for trusses why we have to make clear that the truss is a deck
truss or a through truss? Under what condition the influence lines for the two trusses are identical
and under what condition the influence lines for the two trusses are different?
8-11 What is the procedure to construct influence lines by mechanismic method?
8-12 By using the character of the geometric construction of a statically determinate multispan beam,
explain why the shape of the influence line for a quantity in the subsidiary portion is coincided with
the initial base line in the stable portion of the multispan beam when using the mechanismic method
to construct the influence line?
8-13 How to determinate the most unfavorable loading position and a critical loading position? What is
their difference?
8-14 Simply explain the meanings of sign change of the discriminant, ∑ Ri tan α i , of a critical loading
position ?
8-1 Construct the influence lines for reactions and internal forces on designated sections by using static (or
equilibrium) method.
(a) Construct the influence lines for reactions RA and M A on support A and internal forces M C
and QC on section C.
(b) Construct the influence lines for reactions RA at support A and RB at support B and internal
forces M C and QC on section C and M A , ( Q A )left and ( QA ) right on section A.
(c) Construct the influence lines for reactions RA at support A and RB at support B and internal
forces M C , QC and N C on section C.

(a) Construct the influence lines for reactions RAx and RAy at support A and internal forces M C ,
QC and N C on section C.
x P =1 x P =1
D B
A B
C A C
a b
2m 3m 2m
l
Problem 8-1 (b)
Problem 8-1 (a)
B
x P =1 B P =1
C
C α
A
α
A x
a b a b
l l
Problem 8-1 (c) Problem 8-1 (d)
8-2 Construct the influence lines for statically determinate beams due to unit load transmitted via panel
points by using static method.
(a) Construct the influence lines for internal forces M C and QC on section C.
(b) Construct the influence lines for reactions RA at support A and internal forces M C and QC
on section C, ( QD )left and ( QD )right on section D.

P =1 P =1
D E F G
H
A B B
C A C D
1m 1m 2m 2m 3m 1m 2m 1m 2m 3m 1m 2m
Problem 8-2 (a) Problem 8-2 (b)
8-3 Construct the influence lines for axial forces of the designated members of the statically determinate
trusses by using static method.
(a) Construct the influence lines for axial forces of the designated members 1, 2 and 3 of the deck
truss shown in the figure.
(b) Construct the influence lines for axial forces of the designated members 1, 2, 3 and 4 of the
deck truss shown in the figure.
P =1
D 1
A B
2
2m
3
C E
4 × 2m
Problem 8-3 (a)
P =1
C D 1 E 2 F
3
a
4
A B
2a 4a 2a
Problem 8-3 (b)
(c) Construct the influence lines for axial forces of the designated members 1, 2 and 3 of the deck
truss shown in the figure.
(d) Construct the influence lines for axial forces of the designated members 1, 2, 3 and 4 of the
through truss shown in the figure.
P =1
C D 1 E A A
1 2
4m
3
2m 1m
3
2 B
D E 4 F C
B P =1
3 × 3m 3 × 4m 2 × 4m
8-4 Construct the influence lines for the designated quantities by using mechanismic (or virtual
displacement) method.
(a) The beam and its designated quantities have been pointed out in problem 8-1 (c).
P =1
E C
A D
B
1m 1m 2m 2m 2m
Problem 8-4 (b)
P =1
B C D E
A F
G H
1m 1m 1m 1m 2m 1m 2m
Problem 8-4 (c)
(b) Construct the influence lines for internal forces M E on section E, ( QB )left and ( QB )right on
section B.
(c) Construct the influence lines for internal forces YC on section C, M H and QH on section H
and QG on section G.
(d) Construct the influence lines for internal forces YA on section A, YB on section B, M E ,
( QE )left and ( QE )right on section E.
P =1
B C D E F
A
4m 1.5m 2.5m 1m 3m
Problem 8-4 (d)
8-5 Compute the magnitude of the designated quantities by using their corresponding influence lines.
(a) Compute the magnitudes of shear ( QD )left on the left section of D and bending moment M E
on section E.
(b) Compute the magnitudes of shear QD on section D and bending moment M E on section E.
(c) Compute the magnitudes of shear QC and bending moment M C on section C.
(d) Compute the magnitudes of reaction YB at support B, bending moment M E on section E and
shear ( QB )left on the left section of B.
20kN 20kN 5kN

2kN/m
B
A A
D E C
D E B C
2m 2m 4m 2m 2m 2m 4m 2m
40kN 40kN
3kN 2kN/m 20kN/m
A B A D
C E B C
2.5m 2m 8m 2m 2m 2m 2m 4m

8-6 Suppose the beams are subjected to arbitrarily uniformly distributed loads, determine the most
unfavorable arrangement of the loads by using the influence line for designated quantity Z; and
compute the maximum Z max and the minimum Z min if the load intensity q = 20kN/m .
(a) Compute bending moments ( M C ) max , ( M C )min , shears ( QC ) max and ( QC )min of the beam
shown in the figure of problem 8-1 (b).
(b) Compute bending moments ( M H ) max , ( M H ) min , shears ( QH )max and ( QH )min of the beam
shown in the figure of problem 8-4 (c).
8-7 Determine the most unfavorable arrangement of the crane loads and compute the maximum and
minimum values of the designated quantities by using their corresponding influence lines.
(a) Compute bending moment ( M C ) max , shears ( QC )max and ( QC )min of the beam shown in the
figure.
(b) Compute reaction ( RB )max of the beam shown in the figure, the loading is the same as shown in
problem 8-7(a).
(c) Compute bending moment ( M C ) max , shears ( QC )max and ( QC )min of the beam shown in the
figure.
82kN 82kN 82kN 82kN

3.5m 1.5m 3.5m
B A B C
A
C
3m 6m 9m 9m
108.5kN 108.5kN 76kN 76kN

4.4m 3.5m
1.15m
A B
C
5m 7m
Problem 8-7 (c)

CHAPTER 9
PRINCIPLE OF VIRTUAL WORK
AND DISPLACEMENTS OF STRUCTURES

The attention of the chapter is mainly focused on the calculation of the displacement of statically
determinate structures by using the principle of virtual work.
In the analysis of statically determinate and indeterminate structures, the chapter plays a role of
the link between the preceding and the following of the text: it is not only the end of the analysis of
statically determinate structures but also the beginning of the analysis of statically indeterminate
structures.
First, we will discuss the principle of virtual work for rigid bodies and its two kinds of
applications: determining the restraint forces and determining the displacement of structures.
Then, we mainly discuss the principle of virtual work for deformable bodies, the general
equations for determining the displacement of structures, the calculation of the displacement of
structures due to loadings, temperature changes and support settlements, and the method of graph
multiplication and so on.
Finally, introduce the reciprocal laws for linearly-elastic deformable systems.
9.1 Introduction for Calculation of Structural Displacement
The calculation of the displacement of statically determinate structures is a very important portion of
the content of Structural Mechanics, and it is also the foundation of the analysis for statically indeterminate
structures.
The evaluations of the displacement for a structure have two sorts of purpose. The first purpose is for
checking the stiffness requirement of the structure. In design stage of a structure, it is necessary that a
structure has to satisfy not only the requirement of strength but the requirement of stiffness as well. That is,
the deflection of the structure must not exceed the allowable value specified by design code (For example,
1 1
the allowable value of the beam of a building roof is equal to ~ of its span, while the allowable
200 400
1
value of a crane beam is equal to of its span). The second purpose is for the preparation of analysis
600
for statically indeterminate structures. Because in the analysis of statically indeterminate structures, we
220
9.1 Introduction for Calculation of Structural Displacement 221
have to consider not only static equilibrium conditions but the compatible conditions of the displacements
of a structure as well.
There are three factors that cause a structure to yield displacements. They are: (1) loadings; (2)
temperature changes and material expansion; (3) support settlements.
Under the action of above three factors, the members of a structure will be deformed, which is named
the deformation of the structure. When a structure develops deformation, the translation of a point, the
translation or rotation of a section of a member of the structure are referred to as the displacement
corresponding to the point and the section of the structure, respectively.
There are two kinds of displacements for a structure. One is linear displacement or translation, which
indicates the distance of movement of a point of the structure along a line. Another is angular displacement
or rotation, which indicate the sectional rotation of a member of the structure.
As shown in Fig. 9.1(a), a simple beam is under the action of uniformly distributed loads; the
deflection shape is shown in dashed line in the figure. The point C (at the middle of the span) in unloaded
state moves to point C ′ in the deformed state, CC ′ is the vertical displacement of C; and the rotation
θB at support B is the angular displacement of section B.
Δ Cx
(b) P
(a) B C
q ΔC y
A B θC
C C′
C′ θB
ΔC y ΔC
A
ΔC x
Fig.9.1 Displacement of structures

(a) displacement of a beam; (b) displacement of a rigid frame
Fig. 9.1(b) shows a cantilever rigid-frame. Under the action of a concentrated load P at the cantilever
end, the deformed shape of the rigid frame is shown in dashed line in the figure. As shown in the figure, the
beam end, C, in unloaded position moves to C ′ in deformed position and the displacement Δ C is the
translation of point C in the direction of CC ′ . The horizontal and vertical components of Δ C are
indicated by ΔCx and Δ Cy , respectively, and the geometric relationship between them is shown in Fig.9.1
(b). θC represents the rotation of section C and is named the angular displacement of the section.
222 Chapter 9 Principle of Virtual Work and Displacements of Structures
The chapter will pay attention to the calculation of the displacements of linearly deformable systems
(or structures). The theoretical foundation of the calculation is the principle of virtual work. The method of
the calculation is the method termed unit load (or unit force) method.
The so called linearly deformable systems implies that the displacements are proportional to the loads
applying on them. That is, when the loads are removed the displacements are disappeared. The conditions
which a linearly deformable system must be satisfied are:
(1) The materials of the system are linearly elastic, i.e., the relationship between stress and strain is
linear.
(2) The deformation of the system is small enough to neglect the changes of originally acting effect
of its loadings.
A linearly deformable system is also termed linearly elastic system. The principle of superposition can
be applied to a linearly elastic system. Therefore, the calculation of displacements of a linearly deformable
system can be implemented by using the principle of superposition.
9.2 Virtual Work and Principle of Virtual Work
9.2.1 Virtual work
As we have already known, the magnitude of the work done by a unchangeable force P is the product
of the force P and the displacement Δ of its point of application undergone in the direction of P. That is,
W = PΔ (9-1)
Now, let us extend the definition. Let W represents virtual work, P indicates a generalized force, and Δ
identifies the generalized displacement corresponding to P. Here, the definition of work is still the product
of force P and its corresponding displacement Δ . The dimension or unit of work W is Newton multiplying
meter (i.e. N ⋅ m ).
(1) Generalized force and its corresponding displacement
If P is a force, its corresponding displacement Δ will be a translation along the direction of P. as
shown in Fig.9.2 (a), a vertical force P is exerted at point C on the simple beam; when the force does work
through the displacements shown in Fig.9.2 (c), its corresponding displacement is the vertical translation
Δ at the same point C.
If P is a moment, its corresponding displacement Δ will be an angular displacement about the same
direction of the moment P. As shown in Fig.9.2 (b), a moment M is exerted at right-hand end B of the
simple beam; when the moment does work through the displacements shown in Fig.9.2 (c), its
corresponding displacement is the rotation θ about the same direction of the moment M acting at the end
section B of the beam.
(a) (b) (c)

P P M
A B A B A C B
C C Δ
θ
Fig.9.2 Forces and their corresponding displacements

(a) a vertical force P; (b) a moment M ;
(c ) displacements corresponding to P and M
If a system of forces does work through their corresponding displacements, their work can be
expressed in the form of equation (9-1) as well. The system of forces may be represented by symbol P and
their corresponding displacements may be represented by symbol Δ . Here, P is termed a generalized force
and Δ is named a generalized displacement corresponding to P. For example, a pair of forces P1 and P2
shown in Fig.9.3 (a), equal in magnitude ( P1 = P2 = P ) but opposite in direction, can be recognized as a
generalized force P. Its corresponding displacement Δ ( Δ = Δ1 + Δ 2 ) is the sum of the horizontal
components of the displacements at point A and B in the sense of the connecting line of the two forces, that
is, the relative horizontal displacement between point A and B of the rigid frame, as shown in Fig.9.3 (b). In
this case, PΔ = P(Δ1 + Δ 2 ) = P1Δ1 + P2 Δ 2 is just the work done by the two forces shown in Fig.9.3 (a).
Again for example shown in Fig.9.3 (c), the two parallel force P1 and P2 are equal in magnitude
( P1 = P2 = P ) but opposite in direction, can be considered as a generalized force P as well. Its
corresponding displacement Δ ( Δ = Δ1 + Δ 2 ) is the sum of the vertical components of the displacements
at points A and B in the sense of the two forces, that is, the relative vertical displacement between points A
and B of the rigid frame, as shown in Fig.9.3 (d). In this situation, PΔ = P(Δ1 + Δ 2 ) = P1Δ1 + P2 Δ 2 is
just the work done by the two forces shown in Fig.9.3 (c). Similarly, if a generalized force P is a couple of
moments, then its corresponding displacement Δ will be the sum of the angular displacements about the
same sense of the couple P. As shown in Fig.9.3 (e), a couple of moments M 1 and M 2 are equal in
magnitude ( M1 = M 2 = M ) but opposite in direction, acting at the left and right side of hinge C of the
rigid frame, can be realized as a generalized force M. Its corresponding displacement θ ( θ = θ1 + θ 2 ) is
the sum of the slopes of the left and right sections adjacent to hinge C in the sense of the couple, the
relative rotation between the left and right sections adjacent to hinge C, as shown in Fig.9.3 (f). In this
condition, M θ = M (θ1 + θ 2 ) = M1θ1 + M 2θ 2 is just the work done by the couple shown in Fig.9.3 (e).
(2) Virtual work
The term virtual is used here for the purpose to distinguish the traditional concept of real work and the
new concept of virtual work. The so called virtual work simply means imaginary (not real) work, of which
the force (or forces) has no relation with the displacement (or displacements) undergone by the force (or
forces) during performing the work. That is, the displacement (or displacements) is not caused by the force
(or forces) doing the work; or when performing the work through the displacement (or displacements), the
force (or forces) is a constant value (or values). As shown in Figs.9.2 (a) and (b), the generalized force in
the figure (a) does not depend on the generalized displacement in the figure (c).
(a) (b) Δ1 Δ2
P1 P2
A B A B
P1 P2
(c) (d) B
Δ2
A B Δ1
A
M1 M2 θ1 θ2
(e) (f)
C C
Fig.9.3 Generalized forces and displacements

(a) two horizontal forces P1 and P2 ; (b) corresponding displacements of the two horizontal forces P1 and P2 ;
(c) two vertical forces P1 and P2 ; (d) corresponding displacements of the two vertical forces P1 and P2 ;
(e) a couple of moments M 1 and M 2 ; (f) corresponding rotations of the couple M 1 and M 2
9.2.2 Two kinds of applications of principle of virtual work for rigid body systems
If the strain of the materials composing a system does not take into consideration, only the movements
of the members of the system are concerned during its moving process. The system is recognized as a rigid
body system. Since the principle of virtual work for rigid body systems was introduced in the previous
section 8.5.1, now two kinds of its applications will be discussed.
For work, there exist two states, one is force state and another is displacement state. Since the force
state and the displacement state are independent to each other for the virtual work performed by the force
state acting through the displacement state, an unknown force of a balance force system acting on a
structure will be determined by imagining a likely displacement state for the structure; whereas an
unknown displacement of a given displacement state of a structure will be obtained by imagining a likely
balanced force state for the structure. Now separately discuss them as follows.
1. Determination of unknown restraint forces for statically determinate structures
Fig.9.4 (a) shows an overhanging beam. The reaction X at roller support A is desired when the action
of a concentrated force P at the right end of the beam is considered.
In order to make the beam develop a rigid body displacement, the restraint against the movement in
the direction of the unknown reaction X is removed and substituted by the reaction force X (now the force X
has been an active force). The original system has thusly become an unstable system (or mechanism) with
one degree of freedom. The rigid body ABC can rotate about the hinge B freely; make point A moves to A1
and point C to C1 ; then a virtual (imaginary) displacement state has been obtained, as shown in Fig.9.4
(b).
The force system shown in Fig.9.4 (a) is in equilibrium under the action of external force P and the
reaction forces X , YB and X B , that is, the Fig.9.4 (a) shows a balanced-force-system state.
Let the balanced-force-system state shown in Fig.9.4 (a) undergo the virtual displacement state shown
in the Fig.9.4 (b) to perform virtual work, the expression for the virtual work can be written as
PΔ X + PΔ P = 0 (9-2)
In which, Δ P and Δ X are the virtual displacements along the directions of P and X, respectively. By the
geometric relation of the virtual displacement state, we obtain
Δ X = a ϕ , Δ P = −bϕ
Since the direction of Δ X is the same as that of X, it has positive sign; the direction of Δ P is opposite to
that of P, it has negative sign. Then they relation will be
ΔP b
=− (a)
ΔX a
Substitute equation (a) into equation (9-2), we obtain
b
X ΔX − P ΔX = 0
a
That is,
b
X −P =0 (b)
a
So
b
X =P (c)
a
P
(a) B
A
C
XB
X a YB b
C1
ΔP
A B ϕ
(b)
C
ϕ
ΔX
A1
b
(c) B δP = −
A a
C
δX =1
Fig.9.4 Determination of support reaction by principle of virtual work

(a) a balanced force system; (b) virtual displacement state;
(c) simple indication of virtual displacement state
ΔP
It is observed from equation (a) that the ratio does not vary with respect to the magnitude of
ΔX
Δ X , then for the convenience of calculation, the virtual displacement in the direction of X may be
imagined as a unit value [shown as Fig.9.4 (c)]. The virtual displacement in the direction of P will thusly be
δ X . Let δ X = 1 , by the geometric relation, we write
b
δP = −
a
Now the virtual-work equation and unknown reaction X become as
b
X ⋅1 + P ⋅ δ P = 0 , X = − Pδ P = P
a
The eventually positive result of X tells us that the assumed direction of X (downward) is the same as it
really is.
When the shear force X ( X = QD ) at section D of the overhanging beam shown in Fig.9.5 (a) is
desired to determine, remove the restraint against the movement in the direction of X, that is, replace the
rigid connection at D by two infinitesimal-length links parallel to the beam axis, then replace the restraint
for shear by a pair equal in magnitude but opposite in direction shear force X = QD . Here, X is a
generalized force, and the original beam has become a mechanism with one degree of freedom shown as
Fig.9.5 (c). The rigid body DBC can rotate about the hinge B freely; make point D moves to D1 and point
C to C1 ; since the two portions AD and DBC are connected by two infinitesimal-length links parallel to
the beam axis, the portion AD2 remains parallel to D1 BC1 during any virtual displacements; the
deflection of A is equal to zero because of the roller support at A. A virtual (imaginary) displacement state
has been thusly obtained, as shown in Fig.9.5 (c). Let the balanced-force-system state shown in Fig.9.5 (b)
undergo the virtual displacement state shown in Fig.9.5 (c) to perform virtual work, the equation of the
virtual work can be expressed as
X Δ X + PΔ P = 0
In which, Δ X is the virtual displacements along the directions of X, i.e. the relative deflection between
the immediate left and right sections of D. It is a generalized displacement. Let δ X = 1 , as shown in
Fig.9.5 (d), with the geometric relation of the virtual displacement state, we write
b
δP =
a
At now, the virtual-work equation will be
X ⋅1 + Pδ P = 0
So
b
X = QD = − P
a
The eventually negative result of X = QD indicates that the assumed direction of X is opposite to its real
direction, i.e. the shear force QD has negative sign.
The forgoing method used to determine reactions and internal forces of a structure is termed as unit
displacement method.
Now we conclude the procedure for determination of a restraint force X (reaction or internal force) of
a statically determinate structure by using unit displacement method as follows.
P
(a) A D B
C
a a b
2 2
P
D B
(b) A
C
D1
A Δx ϕ B
(c) C
ϕ ΔP
ϕ
D2 C1
0.5
(d) A δx = 1 B C
a
D δP =
0.5 b
Fig.9.5 Determination of internal force by principle of virtual work

(a) an overhanging beam; (b) mechanism for determining X ; (c) virtual displacement;
(d) simple indication of virtual displacement state
(1) Real system Remove the restraint against the movement in the direction of the desired force and
substitute for it a restraint force X so as to make the structure become a mechanism with one degree of
freedom, and the restraint force X becomes an active force. So the original force system of the structure and
the force X compose a new balanced force system to make the mechanism remain in equilibrium.
(2) Virtual system Apply a unit infinitesimal displacement consistent with the restraint conditions to
the mechanism along the action line in the positive direction of the restraint force X ( δ X = 1 ) so as to
obtain a virtual displacement state, in which the virtual displacement corresponding to a force P is δP .
(3) Determine δP By using the geometric relation between δ X and δ P , the value of each
virtual displacement corresponding to each force P will be determined.
(4) Determine X Let the new balanced force system undergoes the virtual displacement state to
perform virtual work, we obtain a virtual-work equation as follows.

X ⋅1 + ∑ Pδ P = 0
then
X = −∑ Pδ P
The key steps in the procedure are removal of the restraint against the movement in the direction of a
desired restraint force, the assumption of a unit virtual displacement along the action line in the positive
direction of the desired restraint force, the construction of correct virtual displacement configuration and
determination of δ P by using the geometric relation.
Example 9-1
Determine the reaction RC at support C and the bending moment M G on section G for the
statically determinate multispan beam shown in Fig.9.6 (a) by using the principle of virtual displacement.
Solution
(1) Determine the reaction RC
① Real system Remove the roller support C and substitute for it a vertical reaction X ( RC = X ), a
mechanism as shown in Fig. 9.6 (b) is thusly obtained.
② Virtual system Apply a unit infinitesimal displacement consistent with the restraint conditions
to the mechanism along the action line in the positive direction of X, i.e. δ X = 1 , so as to obtain a virtual
displacement state of the system as shown in Fig.9.6 (c).
③ Determine δ P By the geometric relation of the virtual displacement state, we find
δ P1 = −3 , δ P 2 = 1.5 (a)
The negative value of δ P1 here means that the direction of δ P1 is opposite to that of P1 .
④ Determine X Substitution of each corresponding virtual displacement will obtain the virtual
work equation as
X ⋅1 + P1δ P1 + P2δ P 2 = 0 (b)
Substitute equation (a) into equation (b), we obtain
X = −2 P(−3) − P ×1.5 = 4.5 P
(2) Determine the bending moment M G on section G
① Real system Remove the restraint against the rotation in the direction of bending moment M G ,
i.e. replace the rigid-joint section G by a hinged joint, and substitute for it a couple of moments X equal in
magnitude but opposite in direction. Then the couple of moments X = M G have been changed into a pair
of active moments from its original form of internal force, and a mechanism shown as Fig. 9.6 (d) is
consequently obtained.
② Virtual system Apply a unit infinitesimal relative rotation consistent with the restraint
conditions in the positive direction of the couple X to the mechanism, i.e., δ X = 1 , so as to obtain a
virtual displacement state of the system as shown in Fig.9.6 (e).
P1 = 2 P P2 = P
(a)
A G B C E
D F
a 2a a 2a 2a a
(b) P1 P2
A B C E
D F
X
(c)
A δX =1 δ P1 E
B
C D F
δP2
(d) P1 P2
A G B C E
X D F
δ P1
(e)
A G B C E
F
δX =1 D
δP2
δB
Fig.9.6 Determination of reactions and internal forces by principle of virtual work

(a) a statically determinate multispan beam; (b) mechanism for determining RC ;
(c) virtual displacement for determining RC ; (d) mechanism for determining M G ;
(e) virtual displacement for determining M G
③ Determine δ P By the geometric relation of the virtual displacement state, we write

δ P1 = −4a , δ P 2 = 2a (c)
④ Determine X By Substituting each corresponding virtual displacement into the virtual work
equation, we will consequently obtain
X ⋅1 + P1δ P1 + P2δ P 2 = 0 (d)
Substitute equation (c) into equation (d), we find
X ⋅1 + 2 P × (−4a) + P × 2a = 0 , X = 6 Pa
2. Determination of displacements for statically determinate structures
Fig. 9.7 (a) shows an overhanging beam, its roller support A has a vertical settlement in a distance of
c1 . Now the vertical displacement Δ on section C is desired to be determined.
Since the displacement state shown in Fig. 9.7 (a) has been given, in order to use principle of virtual
work to determine Δ , a balance force system has to be assumed. Then, a concentrated force P along the
direction of the displacement Δ is assumed to be imposed at point C so as to let it perform virtual work
through the displacement Δ . Consequently, the force P and its corresponding reactions ( RA = Pb / a )
compose a balanced force system as shown in Fig. 7(b), which is a virtual-force-system state.
Let the balanced force system shown in Fig. 9.7(b) perform virtual work through the rigid-body
displacement shown in Fig. 9.7(a), we find
PΔ + RAc1 = 0 (a)
b b
PΔ + P c1 = 0 , Δ = − c1 (b)
a a
It can be observed from the equation (b) that Δ has no relation with the magnitude of the assumed
force P. In the circumstances, in order to simplify calculation we may assume the force P to be a unit
magnitude, i.e., P = 1 as shown in Fig. 9.7(c). Then let the balanced force system formed by the unit
force P = 1 to perform virtual work through the rigid-body displacement shown in Fig. 9.7 (a), we
immediately obtain
b
Δ + c1 = 0
a
b
Δ = − c1
a
The minus sign of Δ means that the direction of displacement Δ should be in the opposite direction of
the assumed force P, i.e., upward direction as shown in the dashed line in Fig. 9.7 (a).
We may find implication from above that if we apply principle of virtual work between assumed
force-system state [Fig.9.7 (b) or (c)] and a given displacement state [Fig.9.7 (a)] the real rigid-body
displacement of a structure can be determined. The main point here is the imposition of a unit load P = 1 ,
which should be imposed in the direction of the desired displacement. The desired displacement will be
determined by establishing virtual-work equation between the force system formed by the unit load and the
displacement state of the structure. Consequently, the method is also termed unit load method.
(a)
A B Δ
C
c1
a b
P
(b)
A B
C
b
RA = P
a 1
(c) A B
C
b
a
Fig.9.7 Determination of displacement of a statically determinate structure by principle of virtual work
(a) a given displacement state; (b) a virtual force system; (c) a virtual unit force system
9.2.3 Principle of virtual work for deformable bodies
When a system undergoes a deforming process not only each members of the system develop
rigid-body deformation, the materials making of the system yield deformation as well. The system belongs
to the category of deformable systems. One of applications of principle of virtual work to deformable
bodies is the principle of virtual forces and it can be stated as follows:
If a deformable structure is in equilibrium under the action of a virtual system of forces (and couples)
and if it is subjected to any small real deformation consistent with the support and continuity conditions of
the structure, then the virtual external work done by the virtual external forces (and couples) acting
through the real external displacements (and rotations) is equal to the virtual internal work done by the
virtual internal forces (and couples) acting through the real internal displacements (and rotations).
In this statement, the term virtual is associated with the forces to indicate that the force system is
arbitrary and has no relation with the action causing the real deformation. The statement may be expressed
mathematically as
We = Wi (9-3)
In which, We represents the virtual work done by external force system; Wi represents the virtual work
done by internal force system.
It should be realized that the principle of virtual forces as described here is applicable regardless of the
cause of real deformations; that is, deformations due to loads, temperature changes, or any other effect can
be determined by the application of the principle. However, the deformations must be small enough so that
the virtual forces remain unchangeable in magnitude and direction while performing the virtual work. Also,
although the application of this principle in this text is limited to elastic structures, the principle is valid
regardless of whether the structure is elastic or not.
Now we discuss how to obtain the expression of Wi , virtual work done by internal forces.
Consider a simple beam shown in Fig.9.8 (a). The simple beam is in equilibrium under the action of a
system of forces as shown. Fig.9.8 (b) shows a displacement and deformed state of the simple beam caused
by other reason other than the system of the forces.
Take a differential segment ds of the beam as a discussing object as shown in Fig.9.8 (c). The
internal forces acting on ds caused by the force system in Fig.9.8 (a) is shown in Fig.9.8 (c), and the
relative deformation of ds due to the deformation of the state shown in Fig.9.8 (b) is shown in Fig.9.8
(d).
The virtual work done by the internal forces shown in Fig.9.8 (c) acting through the deformation
shown in Fig.9.8 (d) can be defined as
dWi = Nd λ + Qdη + Mdθ
In which, N , Q and M are generalized forces; d λ , dη and d θ are generalized displacement
corresponding to N , Q and M .
Consequently, the expression for the virtual work done by the internal forces of the simple beam will
be expressed as
B
Wi = ∫ (Nd λ + Qdη + Mdθ )
A
For a framed structure, the expression will become

Wi = ∑ ∫ (Nd λ + Qdη + Mdθ )
Introducing the expression for external virtual work We (see subsection 8.5.1), the Eq. (9-3) can be
rewritten as
∑ PΔ + ∑ R
i i c = ∑ ∫ (Nd λ + Qdη + Mdθ )
K K (9-4)
Since
d λ = ε ds, dη = γ 0 ds, dθ = κ ds (9-5)
The Eq. (9-4) will become as
∑ PΔ + ∑ R
i i c = ∑ ∫ (N ε + Qγ 0 + M κ ) ds
K K (9-6)
P
(a)
A B
a ds
(b)
A B
(c)
N N
Q Q M M
ds ds ds
dθ
(d) 1
R=
k
dη
ds dλ
ds ds
Fig.9.8 Stress and deformation state of a simple beam

(a) a given stress state; (b) a given displacement and deformed state;
(c) internal forces acting on caused by the force system in figure (a)
(d) relative deformation of due to the deformation of figure (b)
In which, N , Q and M are internal forces acting on the differential segment ds , i.e., axial force,
shear force and bending moment; d λ , dη and d θ are relative deformation corresponding to N , Q
and M , i.e., relative axial deformation, relative shearing deformation and relative rotation of the
differential segment ds ; ε , γ 0 and κ are the relative axial, tangent and bending strains of the
9.3 General Equation and Unit Load Method for Computing Displacements 235
differential segment ds . The symbol ∫ means integral over the axial length of a member, while symbol
∑ means summation over all the members.

Equation (9-6) is the mathematical expression of principle of virtual work, which is commonly termed
virtual work equation for deformable bodies. There is a variety of application of principle of virtual work.
However, the chapter will only discuss one of its applications to the calculation of displacements for
various deformable structures.
9.3 General Equation and Unit Load Method for Computing Displacements
The unit load method for deformable bodies can be very easily extended by that for rigid bodies.
Now assume a real deformed state of a structure has been given. Our purpose is to determine one of its
displacements Δ of the structure by using Eq. (9-6). In order to let one equation to solve one unknown
displacement, only one unit load P = 1 must be imposed on the position, where the displacement Δ is
desired, in the direction corresponding to the desired displacement Δ so as to let the virtual work
equation just involve in Δ and no other unknown displacement. Then take the unit load P = 1 and
reactions and internal forces caused by the unit load as a balanced force system, the real deformed state of
the structure as a virtual displacement state; let the virtual balanced force system acting through the real
displacement state perform virtual work. Consequently, the virtual work equation (9-6) will be established
and the desired displacement Δ will be determined by solving the equation.
By substituting the unit load P = 1 and its corresponding reactions RK and internal forces ( N , Q
and M ) into Eq. (9-6), the general expression for computing a desired displacement Δ by using unit load
method can be expressed as follow
1× Δ = ∑ ∫ (N ε + Qγ 0 + M κ ) ds − ∑ RK cK (9-7)
It is worth of mentioning again that the displacement and the deformed state ( Δ , cK , ε ds , γ 0 ds , κ ds )
here are given by a real structure, while the reactions ( RK ) and internal forces ( N , Q , M ) of the structure,
which compose a virtual balanced force system with P = 1 , are caused by an assumed unit load P = 1 .
If the strains ( ε , γ 0 , κ ) and support settlements ( cK ) of each differential segments of a structure are
prescribed, the procedure to find a desired displacement Δ of the structure may be stated as follows:
(1) Impose a unit load P = 1 at the location and in the direction of the desired displacement Δ .
(2) Determine the reactions ( RK ) and internal forces ( N , Q , M ) of the structure caused by the
unit load P = 1 according to the static equilibrium conditions.
(3) Calculate the desired displacement Δ by using Eq. (9-7).
In equation (9-7), the sign convention is: when the directions of internal forces ( N , Q , M ) and
reactions ( RK ) in the virtual force-system state are the same as those of the strains ( ε , γ 0 , κ ) and support
settlements ( cK ) of the real structure, the products between force and deformation are positive (for instance,
when M and κ make the fiber of the same side of a differential segment be in tension, the product of
M κ will be positive) and vice versa.
If a desired displacement Δ determined by Eq. (9-7) has positive sign, it means that the direction of
Δ is the same as that of P = 1 and vice versa.
It should be realized that the equation (9-7) is a general formula to calculate a displacement Δ . It can
be employed to determine a displacement of a structure regardless of the properties of its materials, the type
of the deformation, the factors causing the deformation, and the type of the structure.
9.4 Calculation of Displacements Caused by Loads
9.4.1 Computing formula and procedure for determination of displacements
The subsection will discuss the calculation of displacements for statically determinate elastic
structures caused by loads.
γ 0 and κ ) in Eq. (9-7) can be determined by using
For elastic framed structures, the strains ( ε ,
Hook’s law learned in strength of materials. The axial strain ε due to axial force N P , average shearing
strain γ 0 due to shear force QP and bending strain κ due to bending moment M P may thusly be
expressed as follows
NP
ε= (9-8a)
EA
Q
γ0 = k P (9-8b)
GA
M
κ= P (9-8c)
EI
In which, E and G are the elastic tension (or compression) and shearing modulus of the materials,
respectively; A and I are the area and inertial moment of a cross section, respectively; EA, EG, and EI
represent the axial, shearing and flexural rigidities of a cross section, respectively; k is a modified factor
because of non uniformity of shearing stress over a cross section, it is associated with the shape of the cross
section, and for rectangular shape, k = 1.2 , circular shape k = 10 / 9 , I shape k = A / A1 ( A1 represent
the area of the web of the I shaped cross section).
Substituting Eq. (9-8) into Eq. (9-7), the general formula for calculating elastic displacements due to
the action of loads will be written as
NN P kQQP MM P
Δ = ∑∫ ds + ∑ ∫ ds + ∑ ∫ ds (9-9)
EA GA EI
It should be noted that in Eq. (9-9) there are two sorts of internal forces. N , Q and M are internal
forces caused by a assumed unit load P = 1 , while N P , QP and M P are the internal forces due to
loads acting on a structure. The sign convention about N and N P , Q and QP , M and M P are
the same as those in the previous chapters, that is, N and N P are positive when they are tensile forces;
Q and QP are positive when they make a differential segment rotate clockwise. The sign convention of
M and M P are regulated by their product, i.e., the product of M and M P are positive when both of
them make the fibers in the same side of a member be in tension, otherwise, it is negative.
The following step-by-step procedure can be used to determine the displacements of a structure due to
action of external loads by unit load method.
(1) Real system Write out the expressions expressing the internal forces N P , QP and M P of
the structure due to its real external loads by static equilibrium equations, and then construct their diagrams.
(2) Virtual system Remove all the given loads from the structure; then apply a generalized unit
load at the location and in the direction of the desired displacement Δ . Find out the expressions expressing
the virtual internal forces N , Q and M of the structure due to the generalized unit load by static
equilibrium equations, and then construct their diagrams.
(3) The desired displacement Δ of the structure now can be determined by applying Eq. (9-9) by
substituting the results from step (1) through (2) into Eq. (9-9).
9.4.2 Formulae for calculating displacements of various types of structures
The expression (9-9) is a general equation for computing elastic deflections of a statically determinate
structure due to external loads. There are three terms on the right-hand side of the equation. The first term
represents the effect of axial deformation; the second term indicates the effect of shearing deformation and
the third term involves in the effect of bending deformation of the all members of the structure. Obviously,
the ratio of the three aspects of the effects in the equation will vary with the characteristics of deformation
and type of the structure. The different simplified equations of expression (9-9) will thusly be obtained by
consideration, on which the main effect of deformation is taken and negligible effect of deformation is
discarded, of characteristics of deformation of a structure.
(1) Beam and rigid frames
For beams and rigid frames, their deformation is mainly caused by bending moments. The deformation
due to axial force and shear forces are small enough to be neglected (see examples 9-2 and 9-8). Therefore,
the expression (9-9) can be simplified as
MM P
Δ = ∑∫ ds (9-10)
EI
(2) Trusses
For trusses, each of the members is subjected to only axial force. In addition this, the cross-sectional
area, axial forces N and N P , and the elastic modulus remain unchangeable over the length of each of the
members, generally. The expression (9-9) is thusly simplified as
NN P NN P
Δ = ∑∫ ds = ∑ l (9-11)
EA EA
(3) Composite structures
For composite structures, beam-typed (or flexural) members are mainly subjected to bending moments,
and two-force members are mainly undergone axial forces. Therefore, the expression (9-9) may be
modified as
MM P NN P
Δ = ∑∫ ds + ∑ l (9-12)
EI EA
The first summation is applied to flexural members and the second for two-force members.
(4) Arches
For a general arch with massive body, the effect of its curvature is commonly neglected, and the
deformation due to bending moment is just considered. In the circumstance, equation (9-10) may be used to
calculate the deflection of this sort of arches. However, the deformations due to bending moments and axial
forces have to be taken into consideration for arches with small rise and long span. Thusly, the expression
will be simplified as
MM P NN P
Δ = ∑∫ ds + ∑ ∫ ds (9-13)
EI EA
9.4.3 Examples for calculating deflections due to external loads
Example 9-2
Determine the deflection Δ on the middle-span section C of the simple beam shown in Fig.9.9 (a),
and compare the effect on the deflection due to the bending deformation with that due to shearing
deformation. The shape of the cross section of the beam is rectangular and its area is b × h .
Solution
(1) Real system The internal force diagrams of the beam due to real external loads are shown in
Fig. 9.9 (a), which are depicted out according to the following equations expressing (in terms of a position
coordinate x) the variations of the internal forces.
q
MP = (lx − x 2 )
2
q
QP = (l − 2 x)
2
NP = 0
(a) q (b) P =1
A B A B
C C
ql ql 1 1
x x
2 2 2 2
l l l l
2 2 2 2
M P diagram M diagram
1
4
ql 2
ql 8 1
2 QP diagram 2 Q diagram
1
ql
2
2
Fig.9.9 Determination of deflection at middle point C

(a) a simple beam and its internal force diagrams;
(b) virtual force system and its internal force diagrams
(2) Virtual system Remove all the given (real) loads from the beam; then apply a unit load P = 1
at point C, where the deflection Δ is desired, and in the direction of the desired deflection to form the
virtual force system as shown in Fig. 9.9(b). The internal force diagrams of the beam due to the virtual load
are also shown in Fig. 9.9 (b), whose equations are written, in terms of the same position coordinate x, out
as follows
1
M= x
2
1
Q=
2
N =0
(3) Calculation of deflection Δ
Substituting equations coming from steps (1) through (2) into every term of expression (9-9), we
obtain deformation caused by bending moments and shear forces, respectively, as follows
1 q
( x ) (lx − x 2 )
MM P l 2 5ql 4
ΔM = ∫ ds = 2 ∫ 2 2 dx = (↓)
EI 0 EI 384 EI
1 q
( ) (l − 2 x )
QQP l 2 ql 2
ΔQ = k ∫ ds = 2 × 1.2∫ 2 2 dx = 0.15 (↓)
GA 0 GA GA
Note: for rectangular cross section, k = 1.2 .
The positive answers for ΔM and Δ Q indicate that the directions of the two deflections coincide
with that of the unit load P = 1 .
Since the axial forces are equal to zero, there is no axial deformation. Consequently, the total
deflection is written as
5ql 4 ql 2
Δ = Δ M + ΔQ = + 0.15 (↓)
384 EI GA
(4) Compare the effects on the deflection due to the bending deformation with that due to shearing
deformation.
The ratio between bending and shearing deformation is
ql 2
ΔQ 0.15
= GA = 11.52 EI
ΔM 5ql 4 GAl 2
384 EI
I h2 1 E 8
For a rectangular cross section, = , assume Poisson ratio μ = , then = 2(1 + μ ) = .
A 12 3 G 3
Substituting them into above expression, we obtain
ΔQ ⎛h⎞
2
= 2.56 ⎜ ⎟
ΔM ⎝l⎠
h 1
Obviously, when the ratio between depth and span of the beam ( ) is equal to , the ratio between
l 10
ΔQ
shearing and bending deformation will be = 2.56% . This implies that the shearing deformation is less
ΔM
than 3 percent of bending deformation. Accordingly, the shearing deformation can be neglected for a
long-span beam and its displacements may be calculated by Eq. (9-10) straightly. However, when the ratio
1
between depth and span of a beam is more than , the ratio between shearing and bending deformation
5
becomes to 10% . Therefore, the shearing deformation cannot be neglected for a beam with quite high ratio
between the depth and span when calculating the displacements of the beam.
P = 45kN P = 45kN
(a)
EA = constant
D E
1.5m
E = 2.1 × 108 kPa A = 12cm 2

A B
C
2m 2m 2m 2m
45 45
(b)
Example 9-3
Determine the vertical displacement of joint C located at the middle point of the bottom chords of the
truss shown in Fig.9.10 (a).
Solution
(1) Real system The real system and its member axial forces N P due to real loads obtained by
using the method of joints are shown in Fig. 9.10 (b).
(2) Virtual system The virtual system consists of a unit load P = 1 applied in the vertical
direction at joint C, as shown in Fig. 9.10(c). The member axial forces N due to the virtual unit load are
determined by applying the method of joints, as depicted in Fig.9.10 (c).
Table 9-1 calculation of ΔC
NN P
member N N P (kN) l (cm) A (cm 2 ) E (kN/cm 2 ) l (cm)
EA
AC 2/3 60 400 12 2.1 × 10 4 0.063
AD -5/6 -75 250 12 2.1 × 10 4
0.062
1
DE -4/3 -60 × 400 12 2.1 × 10 4 0.063
2
DC 5/6 0 250 12 2.1 × 10 4
0
∑ =0.188 (cm)
(3) Vertical displacement at C ( Δ C )

The formula to determine the displacement of a truss is Eq. (9-11). That is,
NN P
Δ=∑ l
EA
To facilitate the computation of the desired deflection, the real and virtual member forces are tabulated
along with the length (l), the cross-sectional area (A) and the elastic modulus (E) of the members, as shown
in Table 9-1. Note that because of the symmetry of geometry and loads acting on the truss, only half of the
truss is computed. The final value of Δ C must be 2 times of the sum tabulated in the table. That is,
Δ C = 2 × 0.188 = 0.376cm ( ↓ )
Example 9-4
Determine the vertical displacement at point B for the curved beam shown in Fig. 9.11 (a), and
compare the effects of displacement components due to shearing deformation and axial deformation. The
centre line of the beam is a segment of arc; the shape of its cross section is rectangular; the central angle
and radius of the arc is equal to α and R, respectively.
Solution
(1) Real system Determine internal forces due to the real loads. The equations expressing the
variation of the internal forces of the real beam can be expressed in terms of a position coordinate x
(located at a distance from right end B) as
1
M P = − qx 2
2
N P = −qx sin θ
QP = qx cos θ
(a) q (b) x P =1
B B
y
C x C
θ θ
A dθ A
α
Fig.9.11 Determination of displacement of a curved beam

(a) real system and its load; (b) virtual system and its unit load
(2) Virtual system The virtual system consists of a unit load P = 1 applied in the vertical
direction at end B, as shown in Fig. 9.11(b). The equations expressing the variation of the internal forces of
the virtual beam can be expressed in terms of the same x position coordinate as that in step (1) as
M = −x
N = − sin θ
Q = cos θ
(3) Vertical displacement Δ at B
If neglecting the effects of the curvature of the beam, the equation determining the displacement Δ
will be
MM P NN P kQQP
Δ = ∑∫ ds + ∑ ∫ ds + ∑ ∫ ds
EI EA GA
For comparison, the displacement components due to the bending moments, axial forces and shear
forces represented separately by ΔM , Δ N and Δ Q , are determined by following equations.
MM P
A q A 3
ΔM = ∫
2 EI ∫B
ds = x ds
BEI
A NN q A
ΔN = ∫
EA ∫B
P
ds = x sin 2 θ ds
B EA
A kQQP kq A
ΔQ = ∫ ds = ∫ x cos 2 θ ds
B GA GA B
In order to identify the variables we select variable θ as unique variable. Then,

x = R sin θ , y = R(1 − cos θ ) , ds = Rdθ
Substituting them into above equations, we obtain
qR 4 α
ΔM =
2 EI ∫0
sin 3 θ dθ
qR 2 α
ΔN =
EA ∫0
sin 3 θ dθ
kqR 2 α
ΔQ =
GA ∫0
cos2 θ sin θ dθ
Since
α
α α ⎡ 1 ⎤ 2 1
∫ sin θ dθ = ∫ (1 − cos θ )sin θ dθ = ⎢ − cos θ + cos3 θ ⎥ = − cos α + cos3 α
3 2
0 0
⎣ 3 ⎦0 3 3
α
α ⎛ 1 ⎞ 1
∫ cos θ sin θ dθ = ⎜ − cos3 θ ⎟ = (1 − cos3 α )
2
0
⎝ 3 ⎠0 3
We write
qR 4 2 1
ΔM = ( − cos α + cos3 α )
2 EI 3 3
qR 2 2 1
ΔN = ( − cos α + cos3 α )
EA 3 3
kqR 2
ΔQ = (1 − cos3 α )
3GA
If α = 900 , then
qR 4
ΔM =
3EI
2qR 2
ΔN =
3EA
kqR 2
ΔQ =
3GA
(4) Comparison of effects of displacement components due to bending, shearing and axial
deformations
If we set α = 900 , h / R = 1/10 , E / G = 8 / 3 , and select a rectangular cross section
with I / A = h / 12 (h indicates the depth of the cross section) and k = 1.2 , the ratios between the
2
displacement components will be

ΔN
2
2I 1⎛ h ⎞ 1
= 2 = ⎜ ⎟ =
Δ M R A 6 ⎝ R ⎠ 600
ΔQ kEI k E⎛h⎞ 1
2
= 2 = ⎜ ⎟ =
ΔM R GA 12 G ⎝ R ⎠ 375
The results of the ratios indicate that the displacement components caused by shearing and axial
deformations may be negligible in some conditions.
9.5 Graph-Multiplication Method
Recall from previous section that when calculating the displacements of beams and frames, the
following integral have to be carried out.
MM P M ( x) M k ( x)
∫ EI
ds = ∫ i
EI
dx
The expression involves with the integral of a product of two internal force functions. Directly
performing the integral is often tedious and time consuming. However, for a straight member or a segment
of a straight member with a constant flexural rigidity EI, the integral can be conveniently performed in
terms of so called graph-multiplication method that can take advantage of the fact that the diagram of one
of the two functions is composed by straight lines to facilitate the evaluation of the integral.
9.5.1 Calculating equations of graph-multiplication method
Consider two bending moment diagrams of a straight segment AB of a member, one of which (say M i )
is a straight-line diagram as shown in Fig.9.12. If the flexural rigidity EI is constant over the length of the
segment, EI can be put outside the integral. Then we write
B MiM k 1 B
∫ A EI
dx =
EI ∫ A
M i M k dx (a)
Since the diagram of M i is a straight line with a slope of α , if orienting the intersecting point
between the line representing M i and the base line x as the origin, the ordinate of M i at an arbitrary x
position will be
M i = x tan α (b)
Substitute equation (b) into the integral of equation (a) and take the consideration of the fact that the
slope α is a constant, we write
B B B
∫ A
M i M k dx = ∫ x tan α M k dx = tan α ∫ xM k dx
A A
(c)
y Mk
Mk
× A
A C B
dx
x0
M i y0 Mi
0
α
A B x
y
Fig.9.12 Graph-multiplication method
The graphic meaning of the right-hand side of equation (c) is that M k dx is the differential area of
the diagram M k at an arbitrary position x, i.e., the dashed area shown in Fig.9.12; xM k dx is the
B
moment of the differential area with respect to coordinate axis y; the integral ∫ xM k dx means the
A
summation of the moments of the all differential areas over the segment AB with respect to axis y.
Recalling from the theorem of moment of an area, the integral equals to the moment of the area of the
bending moment M k diagram (indicated by A) with respect to axis y. If let x0 represent the distance
between the centroid of the diagram M k and the axis y, we write
B
∫ A
xM k dx = Ax0 (d)
Substituting equation (d) into equation (c) and considering the geometric relation of the bending moment
diagram M i , we obtain
B
∫ A
M i M k dx = tan α ( Ax0 ) = Ay0 (e)
Substituting equation (e) into equation (a), we find

B MiM k 1
∫ A EI
dx =
EI
Ay0 (f)
in which, y0 represents the ordinate of M i diagram at the location corresponding to the centroid of the
area A of M k diagram. Consequently, the expression for determining displacements in terms of
graph-multiplication method will be written as
1 1
Δ = ∑∫ MM P ds = ∑ Ay0 (9-14)
EI EI
a b
(b)
(a)
h
h
×C ×C
2l 3 l3
(l + a) 3 (l + b) 3
l
l
lh
lh triagle A=
triagle A= 2
2
(c) (d) vertex

vertex
h
×C ×C
h
l2 l2 5l 8 3l 8
l
2
2d-degree parabola A = lh 2
3 2d-degree parabola A = lh
3
(e) (f)
h
h
×C ×C
vertex vertex
3l 4 l4 4l 5 l5
l l
1 1
2d-degree parabola A = lh 3d-degree parabola A = lh
3 3
Fig.9.13 Graphs and their areas and centroids
Expression (9-14) is the equation of calculating displacements in terms of graph-multiplication

method. It transforms the calculation of an integral with the product of two functions, one of which is a
linear function, into the calculation of their graphic areas, locations of the centroids of the areas and their
corresponding ordinates ( y0 ). When applying the method, following points must be noted.
(1) Implementing conditions The members must be straight; their flexural rigidity EI remain
constant; one of bending moment function under the integral must be a linear function; the value of ordinate
y0 must be measured on the bending moment diagram depicted by the linear bending moment function.
(2) Sign convention If the area A of a bending moment diagram (or a portion area of a bending
moment diagram) lies on the same side with its corresponding ordinate y0 , the product of Ay0 is positive
and vice versa.
Fig.9.13 shows some of commonly used graphs and their area centroids to facilitate the calculation
when applying graph-multiplication method. It should be noted that when employing equations to
determine the areas of the parabolic graphs, the slopes at the vertexes of the graphs must be parallel to the
base lines of the graphs, i.e. shear forces at the locations corresponding to the vertexes are equal to zero
( Q = 0 ), the graphs are called standard parabolas.
9.5.2 Division of segments and superposition of graphs when using graph-multiplication method
When implementing graph-multiplication method, if one of the two graphs multiplied each other is a
straight line graph and another has a curved line, the ordinate y0 must be measured on the graph of
straight line. If the two graphs multiplied each other are all straight lines, the ordinate y0 may be
M k diagram
×A1 ×A3 ×A1
×A2 ×A2
I1 I2
M k diagram
y1 y3
y2
y1 y2
M k diagram M k diagram
Fig.9.14 Graph multiplication on Fig.9.15 Graph multiplication on
segments of polygonal diagram segments with varying rigidity
determined on any one, whose area is not concerned, of the two diagrams. If the following cases are
encountered, the product of the two graphs must be carried out by dividing them into some simple portions,
on each of which the graph-multiplication method can be applied, then superpose them together.
(1) Division
① If one of the two diagrams is a polygon and another is a curved trapezoid, the calculation of the
integral ∫ M i M k dx must be performed over several segments. For the diagram shown in Fig.9.14, the
calculation of the integral will be
∫M M
i k dx = A1 y1 + A2 y2 + A3 y3
② If the flexural rigidity EI specified over the extent of the integral varies in different segments,
the integral must be carried out over the segments divided by the variation of EI. For example, the diagram
shown in Fig.9.15, the integral becomes
MiMk 1 1
∫ EI
dx =
EI1
A1 y1 +
EI 2
A2 y2
D
M k diagram
a
×A2 A1 ×
b B
a ×A1 A A2 × b
M k diagram l C
c y1 y2 d
d
y2
M i diagram c y1 M i diagram
Fig.9.16 Graph multiplication Fig.9.17 Graph multiplication

on the same side on different side
(2) Superposition
When the diagram of M k is quite difficult to determine its area and centroid, the diagram may be
divided into several simple diagrams and then use superposition method to perform the integral.
① If the two diagrams are all trapezoid [Fig.9.16], the two trapezoids may be separated into two
triangles (or one rectangle and one triangle) to simplify the application of graph-multiplication method.
This means
∫M Mi k dx = A1 y1 + A2 y2 (a)
In which, ordinates y1 and y2 may be expressed as

2 1 ⎫
y1 = c + d ⎪
3 3 ⎪
⎬
1 2 ⎪
y2 = c + d
3 3 ⎪⎭
If the two diagrams are all straight lines and the two diagrams have positive and negative portions as
shown in Fig. 9.17. The diagram of M k may be treated as two triangles, one of which has positive area
(ADB) and another has negative area (ABC) as shown in Fig. 9.17. The equation (a) can still be used to
calculate integral ∫ M i M k dx , but ordinates y1 and y2 must be determined by following formulae.
2 1 la
y1 = − c + d (corresponding area A1 = )
3 3 2
1 2 lb
y2 = c − d (corresponding area A2 = )
3 3 2
② If the diagram of M k is the diagram of bending moment M P of a straight segment of a member

due to uniformly distributed loads and member end couples, recalling from the method of superposition
used to construct bending moment diagrams, the M P diagram is composed by simple graphs. As shown
in Fig.9.18 (a), the bending moment diagram of M P over segment AB due to uniformly distributed load q
x dx
(a)
C
MA
M P diagram
M0
A B
MB
D
(b) C
MA
M ′ diagram
A1
A M′
B
A2 MB
D
(c) 0
M diagram
A B
0 ql 2
M
8
Fig.9.18 Decomposition, superposition and graph

multiplication of diagram
is superposed by a graph with straight side (caused by member-end couples M A and M B ) and a graph
with standard parabolic-curve side (caused by q).
Therefore, the diagram may be decomposed into two simple graphs ( M ′ and M ) as shown in Fig.
0
9.18 (b) and (c), respectively. The graph of M ′ may be decomposed again into two triangles. Then, the
graph-multiplication method will be easily used to the three simple graphs.
It should be noted that the superposition of bending moment diagram is actually the superposition of
0
their ordinates. So at the same x position, the ordinates of M [Fig. 9.18 (a)] depicted on an inclined base
0
line CD, and that of M [Fig.9.18 (c)] depicted on a horizontal base line, are identical. Consequently, the
differential areas as shown in dashed lines in Fig. 9.18(a) and (c) are identical, and then the areas and their
centroids are as well identical, respectively.
9.5.3 Examples for determining displacements by graph-multiplication method
Example 9-5
Calculate the deflection at point C for the beam shown in Fig.9.19 (a) by graph-multiplication method.
Solution
(1) Real system The bending moment M P diagram due to real load is shown in Fig.9.19 (a).
(2) Virtual system The virtual system with a unit force P = 1 applied at C and its moment
diagram M are shown in Fig.9.19 (b).
(3) Deflection Δ at point C
Due to the symmetry of the two diagrams, only half of them are computed. Since the left half of
M P diagram is a curved-line graph and M is a straight-line graph, the calculation of the area (A) and
ordinate ( y0 ) used in graph-multiplication method must be performed in M P and M diagrams separately
as
2 l ql 2 ql 3
A= ⋅ ⋅ =
3 2 8 24
5 l 5l
y0 = ⋅ =
8 4 32
Finally,
1 1 1 ql 3 5l 5ql 4
Δ = ∑∫ MM P dx = 2 Ay0 = 2 ⋅ ⋅ ⋅ = (↓)
EI EI EI 24 32 384 EI
The final result is the same as that in example 9-2.
(a) q
A B
× A C
ql 2
8
l l
2 2
EI = constant
P =1
(b)
C B
A
l
5 l 4
×
8 4
Fig.9.19 Determination of the deflection at the middle

span of a simple beam
(a) real loads and bending moment M P diagram;
(b) virtual unit load and bending moment M diagram
Example 9-6
Calculate the deflection at point B of the cantilever beam shown in Fig.9.20 (a) by
graph-multiplication method.
Solution
(1) Real system The bending moment M P diagram due to real loads is shown in Fig.9.20 (b).
(2) Virtual system The virtual system with a unit force P = 1 applied at B and its moment
diagram M are shown in Fig.9.20 (c).
(3) Deflection Δ at point B
Since the slope at point B of M P diagram is not parallel to base line AB, the equation (e) used for
calculating area and centroid expressed in Fig.9.13 can not be used directly. The M P diagram is thusly
discomposed into two simple graphs A1 and A2 , in which A2 is a standard parabola, as shown in
Fig.9.20 (b). Their areas ( A1 and A2 ) and corresponding ordinates measured on M diagram are
expressed as
1 1 2
A1 = ⋅ l ⋅ ql 2 = ql 3 ， y1 = l ( y1 and A1 lie on the same side)
2 2 3
2 1 1 1
A2 = l ⋅ ql 2 = ql 3 ， y2 = l ( y2 and A2 lie on different sides)
3 8 12 2
Consequently, we obtain
1 1 1 ql 3 2l ql 3 l 7ql 4
ΔB = ∑ ∫ MM P dx = ( A1 y1 − A2 y2 ) = ( × − × )= (↓)
EI EI EI 2 3 12 2 24 EI
q ql
(a)
2
A B
l
EI = constant
(b)
ql2 A2
A1 × ×
A B
(c) P =1
l
y1 y2
A B
Fig.9.20 Determination of the deflection at the end

of a cantilever beam
(a) real loads and bending moment M P diagram;
(b) virtual unit load and bending moment M diagram
Example 9-7
Calculate the slope of section C for the beam shown in Fig.9.21 (a) by graph-multiplication method.
Solution
(1) Real system The bending moment M P diagram due to real loads is shown in Fig.9.21 (b).
(2) Virtual system The virtual system with a unit couple (corresponding to the desired slope)
M = 1 applied at C and the bending moment (due to M = 1 ) diagram M are shown in Fig.9.21 (c).
(3) Slope Δ at section C
Since the M diagram is composed by two straight lines (AB and BC), the graph-multiplication
method has to be applied to the two segment separately. The M P diagram on segment AB decomposed
into a triangle and a standard parabola as shown in Fig.9.21 (b). The areas and their corresponding
ordinates are calculated as follows.
1 2
A1 = × 4 × 2 = 4 ， y1 = ( y1 and A1 lie on the same side)
2 3
2 1
A2 = × 4 × 6 = 16 ， y2 = ( y2 and A2 lie on different sides)
3 2
1
A3 = ×1× 2 = 1 ， y3 = 1 ( y3 and A3 lie on the same side)
2
2kN
3kN/m
(a) B
A C
4m 1m
EI = 45kN ⋅ m 2
2
(b) １
×A 1 ×A 3
A C
B
６
× A2
５
l l
(c)
y1 y3 M =1
y2
A
B C
Fig.9.21 Determination of the slope at the end of an overhanging beam

(a) an overhanging beam and its loads; (b) bending moment M P diagram;
(c) virtual bending moment M diagram
Finally, we obtain
1 1
Δ=
EI
∑ Ay 0 =
EI
( A1 y1 − A2 y2 + A3 y3 )
1 2 1 13 13
= (4 × − 16 × + 1) = − =− ( )
EI 3 2 3EI 3 × 45
= −0.096 (rad)
The minus Δ means that the slope of section C is counterclockwise (opposite direction of M = 1 ).
Example 9-8
Calculate the deflection at point D (the middle section of the cantilever) due to the bending moments
for the cantilevered rigid frame shown in Fig.9.22 (a) by graph-multiplication method, and Compare the
effect on the deflection between the deflection component due to the axial forces and that due to the
bending moments. The cross-sectional area A and flexural rigidity EI are indicated in the figure as well.
Solution
(b) (c)
(a)
P P l
P =1
D Pl y1 2 A1
B l
C Pl B C 2 D C
B
EI = constant y2 real system and virtual system and

A = b×h M P diagram A2
l
M diagram
l
A Pl A A
2
l l
2 2
(d) P (e)
P =1
P B C 1 D C
B
real system and virtual system and

N P diagram N diagram
P A 1
A
Fig.9.22 Determination of the deflection at point D of the rigid frame

(a) a rigid frame and its load; (b) bending moment M P diagram; (c) virtual system and its
bending moment M diagram; (d) real axial force N P diagram; (e) virtual axial force N diagram
(1) Real system The bending moment M P diagram due to the real load is shown in Fig.9.22 (b).
(2) Virtual system The virtual system with a unit force (corresponding to the desired deflection)
P = 1 applied at D and the bending moment (due to P = 1 ) diagram M are shown in Fig.9.22 (c).
(3) Calculate deflection Δ M at section D due to bending moments
The area of M diagram, which is divided into two portions indicated by A1 and A2 , respectively,
and the ordinates y1 and y2 , which are measured from M P diagram at the positions corresponding to
the centroids of A1 and A2 , respectively, are calculated as follows.

1 l l l2 5
A1 = ⋅ ⋅ = ， y1 = Pl ( y1 and A1 lie on the same side)
2 2 2 8 6
l l2
A2 = ⋅ l = ， y2 = Pl ( y2 and A2 lie on the same side)
2 2
Therefore, we find
1 1 1 ⎡l2 5 l2 ⎤ 29
ΔM = ∑ ∫ MM P ds = ∑ Ay0 = ⎢ ⋅ Pl + ⋅ Pl ⎥ = Pl 3 (↓)
EI EI EI ⎣ 8 6 2 ⎦ 48EI
(4) Compare the effect on the deflection between the deflection component due to the axial forces
and that due to the bending moments
The axial force N P diagram due to real loads and the virtual axial force N (due to P = 1 ) diagram
are shown in Fig.9.22 (d) and (e), respectively.
The deflection component due to axial forces may be expressed as
NN P NN P Pl
ΔN = ∑ ∫ ds = ∑ l=
EI EA EA
The ratio between the deflection component due to axial forces and that due to the bending moments is
thusly expressed as
pl
ΔN 48 I
= EA 3 = (a)
Δ M 29 Pl 29 Al 2
48EI
Substituting the cross-sectional area A = b × h and inertial moment I = bh /12 , we obtain
3
ΔN 4 h2
= (b)
Δ M 29 l 2
h 1 h2 1 Δ 1
It can be observed from equation (b) that if = , 2 = and N = . This implies that the
l 10 l 100 Δ M 725
effect on the deflection due to axial deformation is much smaller than that due to bending deformation. So
for a beam (or flexural) member, the effect to a displacement due to axial forces is generally neglected.
Example 9-9
Calculate the relative horizontal displacement between points C and D due to the hydrostatic pressures
for the rigid frame shown in Fig.9.23 (a) by graph-multiplication method.
Solution
(1) Real system The bending moment M P diagram due to the real pressures is shown in
Fig.9.23 (b).
(a) (b) (c)

C D C D C 1 1 D
EI = constant
qa 2
a
qa 2 A1 A4 qa 2
q 2
6 × A2 × × 6
a
y1 y2 y3 y4
a
A B A B A B
A3 × a a
2a
Fig.9.23 Calculate relative displacement between C and D

(a) a rigid frame and its loads; (b) bending moment M P diagram;
(c) virtual system and its bending moment M diagram
(2) Virtual system The virtual system with a pair of unit forces opposite in direction
(corresponding to the desired displacement) applied at C and D, and the bending moment diagram M (due
to the two unit forces) are shown in Fig.9.23 (c).
(3) Calculate relative displacement Δ between C and D
It is noted that because the members AC and BD are subjected to the hydrostatic pressure distributed
triangularly along the axes of the two members, the functions of the M P diagram on the two members are
standard cubical parabolas. The vertexes of the two parabolas lie on point C and D, respectively. The areas
of M P diagram on the two members can be expressed as follows.
1 qa 2 qa 3 4 4
A1 = × a × = ， y1 = × a = a ( y1 and A1 lie on the same side)
4 6 24 5 5
The areas of M P diagram on the member AB can be written as
qa 2 qa 3
A2 = 2a × = ， y2 = a ( y2 and A2 lie on the same side)
6 3
2 qa 2 2 3
A3 = × 2a × = qa ， y3 = a ( y3 and A3 lie on different sides)
3 2 3
Finally, we obtain
1
Δ = ∑∫ MM P dx
EI
1
= (2 A1 y1 + A2 y2 − A3 y3 )
EI
1 ⎡ qa 3 4a qa 3 2qa 3 ⎤
= ⎢ 2 × × + × a − × a⎥
EI ⎣ 24 5 3 3 ⎦
4
1 4qa
=− × (→←)
EI 15
The minus Δ indicates that the relative horizontal displacement between points C and D is in the opposite
direction of the pair of assumed unit forces, i.e. points C and D are closed to but not deviate from each
other.
Example 9-10
Find the relative rotation between the left section and the right one of the hinge C of the three-hinged
frame shown in Fig.9.24 (a) by graph-multiplication method.
Solution
(a) (b) ql 2 A ql 2 (c)

q
8 ×
1
ql 2 8 ql 2 1 1 C 1 1
D 8 A2 D 8 D E
C E × C E
1 1
2
EI = constant
l
ql ql
A B A B 4 A B
4
l l ql ql
2 2 2 2
Fig.9.24 Find the relative rotation between the left and the right of C
(a) a three-hinged frame and its load; (b) bending moment M P diagram;
(c) virtual system and its bending moment M diagram
(1) Real system The bending moment M P diagram due to the real load is shown in Fig.9.24 (b).
(2) Virtual system The virtual system with a pair of unit couples opposite in direction
(corresponding to the desired relative rotation) applied at the left and the right of C, and the bending
moment (due to the pair of unit couples) diagram M are shown in Fig.9.24 (c).
(3) Find relative rotation Δ
Because of the symmetry of the structure and the load, the bending moment diagram of the structure
must be symmetric. So the shear force at hinge C must be equal to zero, and then the bending moment
diagrams on portion DC and CE are standard second-degree parabolas, whose vertexes at point C.
Therefore, the areas of M P diagram can be expressed as follows (only half of them are needed).
1 l 1 1 3
A1 = ⋅ ⋅ ql 2 = ql ， y1 = 1 ( y1 and A1 lie on different sides)
3 2 8 48
l l 1 1 2
A2 = ⋅ ⋅ ql 2 = ql 3 ， y2 = ( y2 and A2 lie on different sides)
2 2 8 32 3
Consequently, we find
1 2 2 ⎡ ql 3 ql 3 2 ⎤ ql 3
Δ= ∑ Ay 0 = [ − A1 y1 − A2 y2 ] = ⎢− ×1 − × ⎥ = − ( )
EI EI EI ⎣ 48 32 3 ⎦ 12 EI
The minus Δ implies that the relative rotation between the left section and the right one of the hinge C of
the three-hinged frame is in the opposite direction of the pair of assumed unit couples.
(a) 6kN (b) (c)

6kN
2 4 P =1
B EA C 4kN B 8kN B 3
3
C C
−10kN
3m
5
EA −
y 3
EI D 12kN ⋅ m 2m
D D
EA = 3738 × 102 kN
3m
EI = 3484kN ⋅ m 2 2
4kN 3
A A A
4m
1
6kN
Fig.9.25 Determine the vertical displacement at point C of the composite structure

(a) a composite structure and its load; (b) bending moment M P diagram and N P ;
(c) virtual system and its bending moment M diagram and N
Example 9-11
Determine the vertical displacement at point C of the composite structure shown in Fig.9.25 (a). The
two-force members BC and CD are I shaped steel of number 12 and the flexural member ADB is I shaped
steel of number 18, their axial rigidity EA and flexural rigidity EI are shown in the figure as well.
Solution
When calculating the displacement, the axial deformation of two-force members and bending
deformation of the flexural members are only taken into consideration.
(1) Real system The bending moment M P diagram and the axial force N P due to the real load
is shown in Fig.9.25 (b).
(2) Virtual system The virtual system with a unit force (corresponding to the desired displacement)
applied at point C, and the bending moment diagram M and the axial force N (due to the unit force)
are shown in Fig.9.24 (c).
(3) Determine the displacement Δ C
By equation (9-12) used for calculating the displacement for composite structures, the desired
displacement may be calculated as follows (the graph-multiplication method is only used to the flexural
member).
MM P NN P 1 NN P
Δ = ∑∫ dx + ∑ ∫ l=∑ Ay0 + ∑ l
EI EA EI EA
1 ⎡ 1 2 ⎤ 1 ⎡ 4 5 ⎤ 48 128 + 250
= ⎢ ( × 3 × 12) × ( × 2) ⎥ + ⎢8 × × 4 + (−10) × (− ) × 5⎥ +
EI ⎣ 2 ⎦ EA ⎣ 3 ⎦ 3484 3738 × 10 × 3
2
3 3
= 14.11×10−3 m(↓)
9.6 Calculation of Displacements Caused by Temperature Changes
For a statically determinate structure, the change of its temperature will not cause internal forces, but
the materials of the structure will have free expansion or contraction. The deformation and displacement of
the structure will be thusly induced by the temperature change. The calculation of displacements of a
structure caused by temperature changes is sometimes very meaningful for the design of the structure.
The structural displacements caused by their temperature changes are still determined by employing
the expression of the unit load method as given by Eq. (9-7). When applying the equation to calculate the
displacement induced by temperature changes, the main task will be the determination of the axial strain ε ,
the shearing strain γ 0 and the curvature κ elicited by the temperature changes.
The analysis of the deformation elicited by the temperature changes may be illustrated by reference to
the beam subjected to a change in temperature as shown in Fig. 9.26 (a). Assume that the temperature rises
0 0
are equal to t1 C and t2 C at the top and bottom side of the beam, respectively; and the rise of
temperature varies linearly along the depth h of the cross section [Fig.9.26 (b)], i.e., after the thermal
deformation, the cross sections of the beam still remain planar. Considering an element of length ds, the
deformation of a cross section may be decomposed as an axial deforming component du and a bending
deforming component dθ ; there is no shearing deformation. Then, the average rise of temperature t0 at
9.6 Calculation of Displacements Caused by Temperature Changes 261
the central line of the element is

1
t0 = (t1 + t2 ) (when the cross section is symmetric to its central line, i.e., h1 = h2 ) (9-15a)
2
or (when the cross section is asymmetric to its central line, i.e., h1 ≠ h2 )
t2 − t1 t h +t h
t0 = t1 +h1 = 1 2 2 1 (9-15b)
h h
Where h is the depth of the cross section; h1 and h2 are the distances between the central line and the
top fiber and the bottom fiber of the cross section, respectively; the difference of the temperature rises
between the top and bottom of the element is denoted by Δt
Δ t = t2 − t1 (9-16)
(b) α t1ds
(a)
t1
A +t1 °C
h1
B
+t2 °C
h
ds α t0 ds = du
h2
t2 dθ
ds α t2 ds
Fig.9.26 Deformation caused by temperature changes

(a) a simple beam and its temperature changes; (b) deformation of an element ds
If the coefficient of thermal expansion of the material of the beam is equal to α , the axial deforming
component du and the bending deforming component dθ will be determined by
du = ε ds = α t0 ds ， ε = α t0 (a)
α (t2 − t1 ) α (t2 − t1 ) αΔt
dθ = κ ds = ds ， κ = = (b)
h h h
Substituting equations (a) and (b) into expression (9-7) and setting γ 0 = 0, cK = 0 , we will find the
equation for calculating the displacements elicited by temperature changes as follows.
αΔt
Δ = ∫M ds + ∫ Nα t0 ds
h
For a framed structure, the equation will become
αΔt
Δ = ∑∫M ds + ∑ ∫ Nα t0 ds (9-17a)
h
If t0 and Δt are constants along the members, the equation may be rewritten as
αΔt
Δ=∑ ∫ M ds + ∑ α t ∫ Nds
0 (9-17b)
h
In which, ∫ Mds and ∫ Nds represent the areas of the internal force diagrams M and N of each
member, respectively; the summation symbol means to find the algebraic sum of all members.
The sign convention used for the virtual forces must be the same as that adopted for the real system;
that is, if temperature increases was considered as positive in the real system, then the virtual tensile forces
N must also be considered to be positive and vice versa; if the direction of the bending deformation
elicited by the difference of the temperature changes was the same as that caused by the virtual bending
moments M , the virtual bending moments M must be recognized to be positive and vice versa.
Example 9-12
Find the vertical displacement at point C for the rigid frame shown in Fig.9.27 (a) due to a rise in
temperature by 15 degrees inside of the frame, 0 degree outside of the frame. The coefficient of thermal
expansion α = 0.00001 per degree; the each of the cross sections of the rigid frame is a rectangle with a
depth of 40cm.
Solution
(a) (b) P =1 (c)

P =1
0°C C a
B +15°C a B C 1 B C
a = 4m
0°C +15°C
M diagram N diagram
A a A 1 A
a = 4m
Fig.9.27 displacement of rigid frame caused by temperature changes

(a) a rigid frame and its temperature changes; (b) virtual system and its bending moment M diagram;
(c) virtual system and its axial force N diagram
(1) Real system The real system consists of the temperature changes given in the problem, as
(2) Virtual system The virtual system consists of a unit force applied in the vertical direction at
point C, and the bending moment diagram M and the axial force N due to the unit force are shown in
Fig.9.27 (b) and (c), respectively.
(3) Calculation of the average rise t0 and the difference of the temperature rise Δt
t1 + t2 0 + 15
t0 = = = 7.50 C
2 2
Δ t = t2 − t1 = 15 − 0 = 150 C
(4) Determine the displacement Δ C
Substituting the above values into equation (9-17b), the desired displacement is obtained as follow
αΔt
ΔC = ∑ M ds + ∑ α t0 ∫ Nds
h ∫
15α 1
=− ( × a × a + a × a ) + 7.5α (− a )
h 2
3a
= −7.5α a ( + 1)
h
Since the direction of the bending deformation elicited by Δt was opposite to that caused by the virtual
bending moments M , the first term must be negative; and because the temperature increases was
considered as positive in the real system but the virtual forces N is compression, the second term must be
negative too. Substituting the numerical values into above expression, we obtain
Δ C = −0.93cm(↑)
The minus answer means that the vertical displacement elicited by the temperature change is actually
upward, opposite to the direction of the unit force.
9.7 Calculation of Displacements Caused by Support Movement
Since a statically determinate structure is a stable system with no redundant restraint, only rigid body
motion of the members would occur when the supports of the structure were subjected to movements. As
the multispan beam shown in Fig.9.28 (a), when the support B has a settlement cB the member AC will
rotate about hinge A; the member CD will rotate about support D; only rigid body motion occur to all the
members as shown in dashed lines in the figure.
Consequently, the calculation of displacements due to support movements belongs to the displacement
determination of rigid-body system. It may be calculated by principle of virtual work for rigid bodies. So
when using the unit load method, derived from principle of virtual work, to determine the displacements,
the calculating expression will be obtained from equation (9-7) as follow
Δ = −∑ RK cK (9-18)
Where cK represents the real support movement in the direction of K; RK indicates the reaction in
direction of K due to unit load.
RK cK means the virtual work done by RK on movement cK . When RK and cK have the same
direction their product is positive and vice versa.
In Fig.9.28 (a), if the deflection at hinge C is desired, a virtual system, which does not consider the
support settlement, with a unit load applied at C in the direction corresponding to the deflection, must be
assumed and the reactions corresponding the unit load have to be determined, as shown in Fig.9.28 (b). By
equation (9-18), we find
5 5
Δ = −(− × cB ) = cB (↓)
4 4
(a) A B C D
CB
l l 3l
4 4
(b) A P =1
B D
C
1 5
4 4 0
Fig.9.28 displacement calculation caused by support movement

(a) support B has a settlement; (b) virtual system and its balanced force system
Here, since only support B has a movement and other supports have no movements, just the virtual
reaction at support B performs virtual work and the other virtual reactions perform no work. In addition,
because the reaction RB and settlement cB are opposite in direction, the virtual work or their product must
have negative sign.
Consequently, the procedure to determine the displacements of a structure due to its support
movements may be stated as:
(1) Assume a virtual system by applying a unit load at the location and in the direction of a desired
displacement of the structure without considering its support movements.
(2) Find each reaction RK corresponding to each support movement cK due to the unit load.
(3) Determine the desired displacement Δ by means of equation (9-18).
Example 9-13
Find the horizontal displacement at support B and the slope of section B for the rigid frame shown in
Fig.9.29 (a) due to the settlement of support A.
Solution
ΔB
(a) (c)
(b) M =1
B B P =1 B
h 1
l l
h
A 1 A 0 A
a
h 1
l l
l
Fig.9.29 Determination of displacement of rigid frame caused by support movement

(a) a rigid frame and its support settlement; (b) virtual system for the horizontal displacement at B and
its balanced force system; (c) virtual system for the slope of section B and its balanced force system
(1) Find the horizontal displacement at support B

① The virtual system with a unit load applied at support B in horizontal direction is shown in Fig.
9.29 (b).
② Find each reaction RK corresponding to each support movement cK due to the horizontal unit
force [see Fig.9.29 (b)].
③ Determine the horizontal displacement Δ at support B. By equation (9-18), we find
h ha
Δ = −∑ RK cK = −( × a ) = − (←)
l l
Here, the direction of reaction RA is the same as that of settlement c A , so their product is positive.
The minus Δ indicates that the horizontal displacement at support B is opposite to the direction of
the unit force, i.e., toward left.
(2) Find the slope of section B
① The virtual system with a unit clockwise couple applied at section B is shown in Fig. 9.29 (c).
② Find each reaction RK corresponding to each support movement cK due to the unit
clockwise couple [see Fig.9.29 (c)].
③ Determine the slope Δ of section B. By equation (9-18), we write
1
Δ = −∑ RK cK = −( × a ) ( )
l
Here, the direction of reaction RA is the same as that of settlement c A , so their product is positive.
The negative answer for Δ implies that the slope of section B is opposite to the direction of the unit
couple, i.e., counterclockwise.
9.8 Reciprocal Laws for Linear Elastic System
The section will introduce three reciprocal laws that play very important roles in the analysis of
statically indeterminate linear elastic structures. They are: law of reciprocal work (E. Betti’s law in 1872),
law of reciprocal displacements (James C. Maxwell’s law in 1864) and law of reciprocal reactions.
9.8.1 Law of reciprocal work
The law, initially developed by E. Betti in 1872, can be stated as follows:

For a linearly elastic structure, the virtual work done by numberⅠsystem of forces and couples acting
through the deformation caused by numberⅡ system of forces and couples is equal to the virtual work of
theⅡsystem acting through the deformation due to theⅠsystem.
To show the validity of this law, consider the beam shown in Fig.9.30. The beam is subjected to two
different systems of forces, ⅠandⅡsystems, as shown in Figs.9.30 (a) and (b), respectively. In system
Ⅰ,let us assume that P1 , N1 , Q1 and M 1 represent the system of forces and Δ1 , ε1 , γ 01 and κ1
represent its displacements and strains. In systemⅡ,we use P2 , N 2 , Q2 and M 2 to denote the system
of forces and Δ 2 , ε 2 , γ 02 and κ 2 to denote its displacements and strains. Now, let us assume that we
subject the beam that has theⅠforces already acting on it [Fig.9.30 (a)] to the displacements caused by the
Ⅱsystem of forces [Fig.9.30 (b)]. The virtual work equation can be written as
W12 = ∑ P1Δ 2 = ∑ ∫ N1ε 2 ds + ∑ ∫ Q1γ 02 ds + ∑ ∫ M 1κ 2 ds
N1 N 2 kQ Q MM
= ∑∫ ds + ∑ ∫ 1 2 ds + ∑ ∫ 1 2 ds
EA GA EI
Next, we assume that the beam with theⅡforces acting on it [Fig.9.30 (b)] is subjected to the
displacements caused by theⅠforces [Fig.9.30 (a)]. By equating the virtual external work to the virtual
internal work, we obtain
W21 = ∑ P2 Δ1 = ∑ ∫ N 2ε1ds + ∑ ∫ Q2γ 01ds + ∑ ∫ M 2κ1ds
N N kQ Q M M
= ∑ ∫ 2 1 ds + ∑ ∫ 2 1 ds + ∑ ∫ 2 1 ds
EA GA EI
Noting that the right-hand sides of above two equations are identical, we equate the left-hand sides to
9.8 Reciprocal Laws for Linear Elastic System 267
obtain
∑ PΔ = ∑ P Δ
1 2 2 1 (9-19a)
W12 = W21 (9-19b)
Equation (9-19) represents the mathematical statement of reciprocal work (Betti's law).
Pa1 Pb1
(a)
c d
a b
Δ c1
Δd1
Pc 2 Pd 2
(b)
a b
c d
Δa2
Δb2
Fig.9.30 Law of reciprocal work

(a) systemⅠ; (b) systemsⅡ
9.8.2 Law of reciprocal displacements
The law, initially developed by James C. Maxwell in 1864, will be derived here as a special case of the
more general form of the law of reciprocal work. It can be stated as follows:
For a linearly elastic structure, the displacement at a point i due to a unit load applied at a point j is
equal to the displacement at j due to a unit load at i.
To prove the law of reciprocal
displacements, consider the beam shown in (a) P1 = 1
Fig.9.31. The beam is separately subjected to
theⅠandⅡsystems, consisting of the unit
δ 21
loads at points i and j, respectively, as shown
in Figs.9.31 (a) and (b). As the figure (b) P2 = 1
indicates, δ ij represents the displacement

δ12
at i due to the unit load at j, whereas δ ji
denotes the displacement at j due to the unit
load at i. These displacements caused by per Fig.9.31 Law of reciprocal displacements
unit load are referred to as flexibility (a) systemⅠ; (b) systemⅡ
coefficients. By employing the law of

reciprocal work, we obtain
P1δ12 = P2δ 21
Since P1 = 1 and P2 = 1 , then the equation will become
δ12 = δ 21 (9-20)
This is the mathematical statement of reciprocal-displacement law (Maxwell’s law).

The reciprocal relationship remains valid between the rotations caused by two unit couples as well as
between the deflection and the rotation caused by a unit couple and a unit force, respectively.
9.8.3 Law of reciprocal reactions
The law of reciprocal reactions is also a special case of the more general form of the law of reciprocal
work. It can be stated as
For a linear elastic structure, the reaction corresponding to displacement c2 due to a unit
displacement c1 = 1 is equal to the reaction corresponding to displacement c1 due to a unit
displacement c2 = 1 .
(a) c1 = 1
2
1
k21
k11
(b)
c2 = 1
1 2
k12 k22
Fig.9.32 Law of reciprocal reactions
(a) systemⅠ; (b) systemⅡ
To prove the law of reciprocal reactions, consider two different deformation systems of the multispan
beam shown in Fig.9.32. In the first system [Fig.9.32 (a)], the reactions at supports 1 and 2 due to the
support movement c1 = 1 at support 1 are denoted by k11 and k21 , respectively. In the second system
[Fig.9.32 (b)], the reactions at supports 1 and 2 due to the support movement c2 = 1 at support 2 are
represented by k12 and k22 , respectively. Other reactions that do not perform virtual work are not
sketched. By applying the law of reciprocal work, we write
k12c1 = k21c2
Summary 269
Since c1 = 1 and c2 = 1 , then the equation will be written as

k12 = k21 (9-21)
This is the mathematical statement of reciprocal-reaction law.

The reciprocal relationship remains valid between the couples caused by two unit rotations as well as
between the force and the couple caused by a unit rotation and a unit deflection, respectively.
SUMMARY
The chapter has mainly discussed the determination of displacements for statically determinate
structures by using the principle of virtual work. It is not only the requirement of the determination of
displacements of structures, but the preparation of analysis of statically indeterminate structures. And it
plays a role of the link between the preceding and the following of the text. Therefore, it has a very
important significance for students to have a sound mastery for the chapter.
(1) The principle of virtual work is one of the fundamental principles in mechanics. It is the
characteristics of virtual work and equation of virtual work that the forces and displacements through which
the forces perform virtual work are independent each other. By means of the characteristics, the unknown
restraint forces may be determined for a given structure under the action of a system of forces by assuming
a virtual displacement system of the structure; in a similar way, the unknown displacements may be
determined for a structure subjected to a given deformation by assuming a virtual force system for the
structure. The method of determining the displacements of a structure discussed in the chapter is actually
the method of assuming a virtual force system.
(2) The method for calculating the displacements of a structure is the method of unit load. The
expression for calculating the displacements is equation (9-7), that is
Δ = ∑ ∫ (N ε + Qγ 0 + M κ )ds − ∑ RK cK
The unit load method employs two separate systems: ① a real system of loads (or other effects)
causing the deformation (strains ε , γ 0 , κ and support movement cK ) to be determined and ② a virtual
system consisting of a unit load applied at the point and in the direction of the desired displacement，
whose internal forces and reactions are N , Q, M and RK .
(3) For a linear elastic structure under action of loads, the strain expression can be written as
NP Q M
ε= , γ0 = k P , κ = P
EA GA EI
The explicit expression for unit load method to be used to determine the displacements of the structures
under action of loads is as follow:
NN P kQQP MM P
Δ = ∑∫ ds + ∑ ∫ ds + ∑ ∫ ds
EA GA EI
Here, there is two separate systems: ① a real system of loads causing the deformation to be determined,
whose internal forces are N P , QP , M P , and ② a virtual system consisting of a unit load applied at the
point and in the direction of the desired displacement, whose internal forces are N , Q, M .
When applying equation (9-9), the explicit expressions to determine the displacements of trusses,
beams, and frames are as follows:
MM P
Beams and rigid frames Δ = ∑∫ ds
EI
NN P NN P
Trusses Δ = ∑∫ ds = ∑ l
EA EA
MM P NN P
Arches Δ = ∑∫ ds + ∑ ∫ ds
EI EA
MM P NN P
Composite structures Δ = ∑∫ ds + ∑ l
EI EA
(4) When calculating the displacements for a structure under action of loads, the
graph-multiplication method is expedient to simplify the calculation of the integrals with the mastery of the
skill of dividing segments and superposition.
(5) When calculating the displacements of a structure under action of temperature changes, the
expressions for determining strains can be expressed as:
αΔt
ε = α t0 ， κ =
h
Substituting above two expressions into equation (9-7), the expression used to determine the displacements
due to temperature changes can be obtained as equation (9-17).
(6) The calculation of displacements due to support movements belongs to the displacement
determination of rigid-body system. It may be calculated by equation (9-18). The key point of the
calculation is the determination of reactions RK .
(7) There are many sign conventions in the calculation of displacements of a structure. The main
sign convention is the sign of work. That is, the sign convention of the product of a force and a
displacement. When the force and the displacement have the same directions, their product is positive and
vice versa. In graph-multiplication method, when the area A of a bending moment diagram and the ordinate
y0 of another bending moment diagram lie on the same side of a member, their product is positive and vice
versa.
(8) Reciprocal law is another one of fundamental principle in mechanical analysis. The significance
and the application of reciprocal displacements and reciprocal reactions in different deforming conditions
must be understood. The reciprocal displacements may be applied to either the analysis of statically
determinate structures or that of statically indeterminate structures. But the reciprocal reactions is only used
to the analysis of statically indeterminate structures.
9-1 Why are the displacements of a structure desired?

9-2 What are the main reasons which caused the displacements of a structure?
9-3 How many categories of displacements are there for a structure?
9-4 What is the characteristic of virtual work?
9-5 How to understand that the force and displacement used in principle of virtual work are
corresponding to each other? Please depict out the corresponding displacements or corresponding
forces in the figure of the problem.
(a) (b)
A A
A (d) A
(c)
Δ
θ
Problem for reflecting 9-5
9-6 What are the requirements for the force and displacement used in the principle of virtual work?
9-7 What are the two kinds of applications of principle of virtual work for rigid bodies to statically
determinate structures?
9-8 For the beam shown in the figure of problems for reflecting 9-8 (a) and (b), how to assume virtual
system to determine the desired bending moment?
272 Chapter 9 Principle of Virtual Work And Displacement of Structures
(a) P (b) P
A B A B
C C
a a b
b
l l
find M C find M A
9-9 Please explain the common ground and distinction between the principle of virtual work for rigid
bodies and that for deformable bodies.
9-10 Please explain how equation (9-6) of the chapter satisfies compatible conditions of the deformation.
9-11 Please write out the general expression for calculating a displacement of a structure by unit load
method.
9-12 Why are there general universality for the application of equation (9-7) of the chapter?
9-13 Please find the difference between the equation (9-7) and equation (9-9) of the chapter.
9-14 Please write out the explicit expressions to determine the displacements of trusses, beams, and
frames due to loads, and explain the meanings and sign convention of each term.
9-15 Please explain how to determine the real direction of a desired displacement Δ ?
9-16 What are the conditions of application of graph-multiplication method? Whether or not the method
can be employed when determining the displacements of beams with variable cross sections, or
curved beams, or arches? Whether or not graph-multiplication method may be applied between the
shear forces QP and Q or between the axial forces N P and N ?
9-17 How is the sign convention regulated for the equation (9-14) used for graph-multiplication method?
9-18 Whether or not the following calculations are correct? Why?
(1) ∫M M i K dx = A1 y1 + A2 y2 [see the figure (a)]
2 1 l
(2) ∫ MM dx = ( l × ql 2 ) ×
P
3 8 4
[see the figure (b)]
1 3
∫ MM P dx = ( 3 l ×ql ) × 4 l [see the figure (c)]
2
(3)
1 1 l
(4) ∫ MM P dx = ( l ×Pl ) × ×
2 6 2
[see the figure (d)]
M P diagram
(a) (b) ql 2
× A1
8
× A2 M i diagram
second-degree parabola
l
y1 M k diagram
y2 4 M diagram
ql
2 q P
(c) (d) C
A B A B
l l l
2 2
ql 2 M P diagram Pl
M P diagram P
× A × A
A B A B
3l l
4 3
M diagram P =1 l
P =1 M diagram 2
l
A C B
l
A 3l B 12
4
Problems for reflecting 9-18
9-19 How to determine strains ε , γ 0 , κ when calculating the displacements caused by temperature
changes?
9-20 Under what condition the integral can be substituted by the calculation of the areas when applying
the equation (9-17)? How is the sign convention of the terms in the equation (9-17) prescribed?
9-21 What are the significances of RK , cK and symbol ∑ in the equation (9-18)? How to determine
the sign of the virtual work done by the reactions?
9-22 Why always is there a minus sign in front of the symbol ∑ in the equation (9-18)?
9-23 Please use the following figure to explain the law of reciprocal displacements, and to depict out the
displacements δ12 and δ 21 , and account for the dimension of δ12 and δ 21 .
(a) P1 = 1 (b) M2 =1
A EI = constant A EI = constant B
B
l l
9-24 Whether or not the law of reciprocal reactions can be applied to statically determinate structures?
What results will be obtained when use it?
9-1 Determine the designated restraint forces of the following multispan beams by means of the principle
of virtual work.
20kN 10kN
(a)
C F G H
A
B D E
2m 2m 1m 1.5m 1.5m 1m 3m
find YC , M B , QC left , QC right
(b) 2kN/m
A C
D B
1.5m 1.5m 3m
find M A , M D , QA , YC
Problem 9-1
9-2 Determine the deflections and slopes of section A for following cantilevers (neglect the effect of
shear forces).
(a) (b) P
q
EI
A A B
EI B
l l l
2 2
Problem 9-2
9-3 Determine the deflections at the middle spans for the following simple beams shown in the figure
(neglect the effect of shear forces) by integral method.
(a) (b) P
q
A B A B
EI EI
l l l
2 2
Problem 9-3
9-4 Determine the slope of section A for the following simple beam.
M EI
A B
Problem 9-4
Determine the deflections at joint C for the following trusses (elastic modulus E = 2.1× 10 kPa ;
8
9-5
in the figure (b), the area of every member is indicated in the parentheses on the member, its
2
dimension is cm ).
9-6 Determine the horizontal displacement at joint C for the following trusses. The axial rigidity of every
member is the same, i.e., EA = constant .
9-7 Determine the horizontal and vertical displacements Δ AH and Δ AV of the free end for the circular
cantilever beam shown in the figure. EI is constant.
9-8 Determine the horizontal displacement at point B Δ B for the curved beam. The centre line of the
4f
beam is a parabola y = x (l − x) . EI is constant. Both the axial and shear deformations may be
l2
neglected. Assume that the beam is flat, we may consider ds dx .
20kN
(a) (b) (50)
C (20)
5m
(50) (20)
4m (20)
A B A B
(35) C (35)
10kN 10kN
3m 5m 5m
5m 3m 5m
A = 30cm 2
Problem 9-5
P
(b)
(a) a
P C D
a
a
a
A B
a
a A B
Problem 9-6 a
P q
A
y
f
R
A B
x
B l/2 l/2

9-9 Solve the problems 9-2 to 9-4 by means of graph-multiplication method.

9-10 Find the deflection at point C and E.
9-11 Find the deflection and slope of section B for the cantilever beam.
30kN 30kN P
I1 I2 I3 EI1 EI 2
A B
C D A C B
E
l−a a
2m 2m 2m 2m
l
E = 2.06 × 10 kPa 8
I1 = 6560cm 4
I 2 = 12430cm 4
Problem 9-11
Problem 9-10
9-12 Find the deflection at point D. EI is constant.

9-13 Find the deflection at point C for the beam shown in the figure. The loads P = 9kN , q = 15kN/m ;
and the moment of inertia of the cross section I = 1660cm 4 ; the elastic
modulus E = 2.1× 10 kPa .
8
P P q
A D A B
B
C C
l 0.9m 1.5m 1.5m
3 l l
2 2
9-14 Determine the vertical displacements at points A and D for the rigid frame shown in the figure.
9-15 Determine the horizontal displacement at point C for the rigid frame shown in the figure.
9-16 Determine the horizontal displacement at point E and the slope of section B for the three hinged rigid
frame shown in the figure.
9-17 Determine the relative horizontal and vertical displacements and the relative rotation between section
E and F , termed Δ v , Δ h and Δ r , respectively, for the pair of cantilevered rigid frames shown in
the figure under the following conditions:
(1) Neglect the effect of axial deformation.
(2) Consider the effect of axial deformation.
9-18 Find the vertical displacement at point C due to the temperature changes for the three hinged rigid
frame with rectangular cross sections shown in the figure. The coefficient of thermal expansion
is α = 0.00001 .
9-19 Find the horizontal displacement at point D due to the temperature changes for the rigid frame with I
shaped steel sections shown in the figure. The coefficient of thermal expansion is α = 0.00001 .
q
C I2 = ∞ D
A D 2EI B
q I1
a
I1
1.5a
EI
A B
C
4a
a a
3
q
D C E
C D
E F
q
a
h
EI = constant EI = constant
A B A B
l l
a a
2 2
+10°C
0°C C
C
D E +20°C D
+20°C 0°C
0°C 0°C
4m
+20°C +20°C +20°C

6m
+20°C
A EI = constant B 0°C
A B
6m 6m
6m
depth of cross section h = 60cm
depth of cross section h = 18cm
Problem 9-18
Problem 9-19
9-20 Find the horizontal displacement for the rigid frame shown in the figure under the following
conditions:
(1) Support A move 1cm toward left, (2) support A has 1cm settlement, (3) support B has 1cm
settlement.
9-21 Find the vertical and horizontal displacements and rotation at point K ( v , u and θ ) for the rigid frame
shown in the figure under the conditions: support A has horizontal, vertical and rotating
displacements Δ x , Δ y and Δϕ , separately.
9-22 Determine the vertical and horizontal displacements and the rotation at point D, termed Δ v , Δ h
and Δ r , separately, due to the movement of support A for the beam shown in the figure.
9-23 Determine the vertical and horizontal displacements at point C and the relative rotation of hinge C,
termed Δ v , Δ h and Δ r , separately, due to the movement of support B for the three hinged rigid
frame shown in the figure.
a
K
D
a
Δϕ
1m
B C A
A
Δy
A
4m 1m Δx 3a
C
3m
B C
A D
2cm
A B
3m 3m 1m
0.01rad
4m 4m 1cm
1cm
Problem 9-23
CHAPTER 10
FORCE METHOD

The chapter will discuss the analysis of statically indeterminate structures by using force
method (the method of consistent deformations). Its main attention will be focused on the
fundamental concepts and principles of force method. The selection of basic unknowns, primary
systems and development of compatibility equations of the method will be discussed in detail. As
the application of force method, the analysis of statically indeterminate beams, rigid joint frames,
bent frames, trusses and composite structures will be discussed, respectively; in addition, the
analysis of internal forces of statically indeterminate structures due to temperature changes and
support movements will also be evaluated, and the simplification of analysis for statically
indeterminate symmetric structures will be discussed as well.
Finally, introduce the problems pertinent to the calculation of displacements and verification of
analyzing results of statically indeterminate structures.
10.1 Statically Indeterminate Structures and Degrees of Indeterminacy
10.1.1 Statically indeterminate structures
In the previous chapters of this text, we have discussed the analysis of statically determinate structures.
In this chapter, we begin focusing our attention on the analysis of statically indeterminate structures.
For a structure, if its support reactions and internal forces can be uniquely determined from the
equations of equilibrium (including equations of condition, if any), the structure is referred to as a statically
determinate structure. The simple beam shown in Fig.10.1 (a) is one of the examples of a statically
determinate structure. However, if the support reactions and internal forces can not be completely and
uniquely determined from the equations of equilibrium, the structure will be named a statically
indeterminate structure. The two-span continuous beam shown in Fig.10.1 (b) is one of the examples of a
statically indeterminate structure.
It can be seen, by analyzing the geometric constructions of the two structures, that the two beams are
all stable. If the roller support B is removed from the simple beam shown in Fig.10.1 (a), the simple beam
will become an unstable system. Note however, if the roller support C is removed from the two-span
continuous beam shown in Fig.10.1 (b), the two-span beam will be converted into a simple beam, and it is
280
10.1 Statically Indeterminate Structures and Degrees of Indeterminacy 281
still stable. Therefore, from the viewpoint of geometric construction, the support C is excessive (or
redundant) restraint. With extending the meaning of geometric construction of the two-span continuous
beam, we will introduce the conclusion that a statically determinate structure is a stable system with no
redundant restraint, while a statically indeterminate structure is also a stable system but with redundant
restraint(s).
(a) (b)
XA A XA A C B
B
YA YA YB
YB YC
Fig.10.1 Statically determinate and indeterminate structures

(a) a statically determinate beam; (b) a statically indeterminate beam
Generally speak, the essential characteristics of a statically indeterminate structure is that it has more
support reactions and/or members than required for static stability, the equilibrium equations alone are not
sufficient for determining the reactions and internal forces of such a structure.
10.1.2 Determination of indeterminacy
From static stability point of view, the degrees of indeterminacy indicate the number of redundants
involved in a statically indeterminate structure. If after n restraints were removed from a given statically
indeterminate structure the structure would become statically determinate, the given structure is nth degrees
of indeterminacy. Thusly, we express
Degrees of indeterminacy = number of redundants = number of released restraints (which are
required to convert a statically indeterminate structure into a determinate one) (a)
Since the given structure is stable, recalling from section 2.2.4 of chapter 2, the following expression
will thusly be obtained.
Degrees of indeterminacy = -W (b)
Where W is the degrees of freedom of a system.
From static analysis point of view, the degrees of indeterminacy are equal to the number of required
supplementary equations which are needed to determine the reactions and internal forces of the structure.
Consequently, we find
Degrees of indeterminacy = number of redundant unknown forces = number of unknown forces
– number of equilibrium equations (c)
Figs.10.2 (a) through (d) are shown four kinds of statically indeterminate structures. Fig.10.3 is shown
282 Chapter 10 Force Method
the statically determinate structures which are converted by removing sufficient excessive restraints from
the indeterminate ones. They degrees of indeterminacy are 2, 4, 6 and 3, respectively.
When using equation (a) to determine the degrees of indeterminacy, the main point is that the mastery
of how to release a statically indeterminate structure into a determinate one. The common ways to release a
statically indeterminate structure are as follows.
(a) (a)
X1 X2
(b) (b)
X1 X2 X3 X4
X1 X3 X1
(c)
(c)
X2
X4 X6 X4
X5
X2
(d) (d)
X1 X3
Fig.10.2 Original statically indeterminate structures Fig.10.3 Corresponding statically determinate structures
(a) a beam; (b) a truss; (c) a rigid joint frame; (d) an arch (a) a beam; (b) a truss; (c) a rigid joint frame; (d) an arch
(1) Removing a roller support or cutting a link (or two-force member) is equivalent to release one
restraint [Fig.10.3 (a) and (b)].
(2) Removing a hinged support or a single hinge is equivalent to release two restraints.
(3) Removing a fixed support or cutting out a flexural member is equivalent to release three
restraints [Fig.10.3 (c)].
(4) Changing a rigid joint of two member ends into a hinged joint is equivalent to release one
restraint [Fig.10.3 (d)].
The following two points must be paid attention to:
(1) Do not release a statically indeterminate structure into an unstable one. For instance, if replace
the left hinged support by a roller support for the three span beam shown in Fig.10.2 (a), the beam will
become an unstable system.
(2) The all excessive restraints must be removed completely. As the structure shown in Fig.10.4 (a),
if only one roller support is removed as shown in Fig.10.4 (b), the closed frame is still internally
indeterminate and has three excessive restraints. Therefore, only when the closed frame is cut into the frame
as shown in Fig.10.4 (c), the original structure will be converted into a determinate one. So the original
structure has four redundant restraints.
(a) (b) (c)
Fig.10.4 Statically indeterminate structures with a closed frame

(a) a statically external and internal indeterminate structure; (b) a statically
internal indeterminate structure; (d) a statically determinate structure
10.2 Basic Concept of Force Method
The force method (or the method of consistent deformations) is an essential method used to analyze
statically indeterminate structures. The method, which was introduced by James C. Maxwell in 1864,
essentially involves removing enough restraints from the indeterminate structure to render it statically
determinate so as to convert a statically indeterminate analyzing problem into a statically determinate one,
about which the analytical methods have been familiar to readers since they have already been studied in
forgoing chapters.
Since the independent variables or unknowns in the method of consistent deformations are the
redundant forces (and/or moments), which must be determined before the other response characteristics
(e.g., displacements and internal forces) can be evaluated, the method is termed force method.
10.2.1 Primary unknowns and systems of force method
When analyzing a statically indeterminate structure by means of force method, the enough restraints
must be removed from the indeterminate structure to render it statically determinate. This determinate
structure, which must be statically stable, is referred to as the primary structure. The excess restraints
removed from the given indeterminate structure to convert it into the determinate primary structure are
called redundant restraints, and the reactions or internal forces associated with these restraints are termed
primary unknowns (variables) or redundants. The primary unknowns are then applied as unknown loads on
the primary structure, and their values are determined by solving the compatibility equations based on the
condition that the deformations of the primary structure due to the combined effect of the primary
unknowns and the given external loadings must be the same as the deformations of the original
indeterminate structure. When the primary structure is under combined action of primary unknown(s) and
the given external loadings, it is referred to as a primary system.
(a)
(b) (c)
MA q q
B
B B
XA A EI = constant A A
X1
YA YB
Fig.10.5 Primary system and primary structure of force method

(a) original structure; (b) primary system; (c) primary structure
To illustrate the concept of primary unknowns and primary systems of force method, consider the
propped cantilever beam subjected to uniformly distributed load shown in Fig.10.5 (a). Since the beam is
supported by four support reactions ( X A , YA , M A and YB ), the three equations of equilibrium
( ∑ X = 0 , ∑ Y = 0 , and ∑ M = 0 ) are not sufficient for determining all the reactions. Therefore, the
beam is statically indeterminate. However, if removing roller support B and substituting for it a redundant
unknown force X 1 , which is now actively applied on the cantilever as shown in Fig.10.5 (b), the propped
cantilever will become a statically determinate cantilever that is subjected to original external load and the
unknown force X 1 . Thus, if we can determine the unknown force X 1 , then the remaining three reactions
can be obtained from the three equations of equilibrium. Here, the unknown force X 1 is named the
primary unknown (or variable) and the statically determinate cantilever shown in Fig.10.5 (b) is termed
primary system and the cantilever shown in Fig.10.5 (c) is called the primary structure of the propped
cantilever shown in Fig.10.5 (a).
10.2.2 Compatibility equations of force method
How to determine the unknown force X 1 shown in Fig.10.5? It is obvious that three equilibrium
equations can not solve it. We have to find a new supplementary condition to determine it.
To establish the new supplementary condition, we must consider the compatible condition that the
vertical displacement at the free end B of the primary system [Fig.10.6 (a)] due to the combined effect of
the unknown force X 1 and the given uniformly distributed external load must be the same as the vertical
displacement at the support B of the original indeterminate structure. That is, only if the vertical
displacement at end B of the beam shown in Fig.10.6 (a) is equal to zero, the primary system will be the
same as the original structure. The supplementary condition is thusly developed by using a compatibility
equation based on the geometry of the deformation at support B of the beam. That is
Δ1 = 0 (a)
(a) (b)
q q (c)
Δ11
= B
+
A EI = constant A
Δ1 P
B A B X
1
X1
l
Fig.10.6 Linear superposition of primary system

(a) primary system; (b) primary structure under action of external loads;
(c) primary structure under action of unknown force
If the beam is linear elastic, by using principle of superposition, the total vertical displacement Δ1 at
end B of the primary beam due to the combined effect of q and X 1 can be conveniently expressed by
superimposing (algebraically adding) the vertical displacements due to the external load q and the
redundant X 1 acting individually on the beam; that is,
Δ1 = Δ11 + Δ1P = 0 (b)
In which, Δ1P and Δ11 represent, respectively, the vertical displacements at the end B of the primary
beam due to the external load q and the redundant X 1 , each acting alone on the beam [see Figs.10.6 (b)
and (c)]. Note that two subscripts are used to denote the vertical displacements Δ1P and Δ11 of the
primary beam. The first subscript, 1, indicates the location of these deflections; the second subscript, P, is
used to indicate that Δ1P is caused by the given external loading, whereas the second subscript, 1, of Δ11
implies that it is due to the redundant X 1 . Both of these deflections are considered to be positive if they
occur in the direction of the redundant X 1 , which is assumed to be upward, as shown in Fig.10.6.
For a linear elastic system, the displacement Δ11 may be expressed by means of principle of
superposition. Since the redundant X 1 is unknown, it is convenient to determine Δ11 by first evaluating
the deflection at B due to a unit value of the redundant X 1 , as shown in Fig.10.7 (b), and then multiplying
the deflection thus obtained by the unknown magnitude of the redundant. Thus
Δ11 = δ11 X 1 (c)
δ11
(a) (b)
B B
A
Δ1P A EI = constant
EI = constant X1 = 1
l l
ql 2 A B
2
l
B
A
Fig.10.7 Bending moment M P and M 1 diagrams

(a) M P diagram; (b) M 1 diagram;
In which, δ11 denotes the vertical displacement at point B of the primary beam due to the unit value of the
unknown force X 1 . It has units of displacement per unit force, is referred to as a flexibility coefficient. By
substituting equation (c) into equation (b), we obtain the compatibility equation
δ11 X 1 + Δ1P = 0 (10-1)
Eq. (10-1) is the basic equation in the force method which is referred to as the canonical equation for a
structure indeterminate to the first degree.
The coefficient δ11 and the free term Δ1P in equation (10-3) are actually the displacements of a
statically determinate structure, so they can be calculated by unit load method. When δ11 and Δ1P are
determined the unknown X 1 can be solved by the equation.
In order to calculate coefficient δ11 and the free term Δ1P , first construct the bending moment
diagrams of M P and M for the primary structure due to the external load and the unknown X 1 , each
acting alone on the structure, respectively. Then by graph-multiplication method, we write
M 1M 1 1 l ⋅ l 2l l3
δ11 = ∫ dx = ( ⋅ )=
EI EI 2 3 3EI
M 1M P 1 1 ql 2 3l ql 4
Δ1P = ∫ dx = − ( ⋅ ⋅ l) × = −
EI EI 3 2 4 8 EI
Substituting δ11 and Δ1P into equation (10-1), we find
l3 ql 4
X1 − =0
3EI 8EI
Finally, we obtain
3
X 1 = ql
8
The positive answer indicates that the assumed direction of the unknown X 1 is the same as its real
direction.
Once the unknown X 1 has been computed, the reactions and all other internal forces of the beam can
now be determined by applying the three equilibrium equations to the free body of the indeterminate beam.
The computed results are shown in Fig.10.8.
(b) (c)
(a) ql 2
ql 2 5
q 8 ql
2 8 8
ql B
B B
8 A EI = constant A 2 A 3
ql ql
8
3 l 16 l
5 l
ql ql 2 2
8 8
Fig.10.8 Reaction and internal force diagrams

(a) reactions; (b) M diagram; (c) Q diagram;
The bending moment M on an arbitrary cross section of the structure can also be determined by
employing superposition relationships similar in form to the displacement superposition relationship
expressed in Eq. (10-1). Thus, the bending moment M on an arbitrary cross section can be expressed as
M = M1 X 1 + M P (10-2)
In which, M 1 is the bending moment on an arbitrary cross section of the primary structure due to the unit
load X 1 = 1 ; whereas M P is the bending moment on the arbitrary cross section of the primary structure
due to original given external loads.
10.3 Canonical Equations of Force Method
As mentioned previously, the analysis of a statically indeterminate structure by means of force method
is, in fact, the method by which the primary system is used as an analytical tool and the redundants of the
structure are selected as the primary unknowns. The primary unknowns are then applied as unknown active
forces on the primary structure, and their values are determined by solving the compatibility equations
based on the condition that the deformations of the primary structure due to the combined effect of the
unknown forces and the given external loadings must be the same as the deformations of the original
indeterminate structure. After the unknowns have been computed, the reactions and all other internal forces
of the structure can be determined by applying the methods used to analyze statically determinate structures.
Therefore, the key step of analyzing statically indeterminate structures is establishment of compatibility
equation(s) or force-method equation(s) and determination of the unknowns existing in the equation(s).
10.3.1 Canonical equations for structures with two degrees of indeterminacy
The force-method equations for a structure with second degrees of indeterminacy can be easily
developed by extending the method of consistent deformations developed in the preceding section for
analyzing structures with a single degree of indeterminacy. Consider, for example, the rigid frame with two
degrees of indeterminacy, as shown in Fig.10.9 (a). To analyze the frame, we must remove two excess
restraints. Suppose that we select the support reactions at support B as redundants ( X 1 and X 2 ). The hinged
support at B is then removed from the given indeterminate rigid frame to obtain the statically determinate
and stable primary rigid frame, as shown in Fig.10.9 (b). The two redundants are now treated as unknown
active forces on the primary frame, and their magnitudes can be determined from the compatibility
conditions that the deflections of the primary frame at the locations B of the redundants due to the
combined effect of the known external load and the unknown redundants ( X 1 and X 2 ) must be equal to
zero. This is because the deflections of the given indeterminate frame at the hinged support B are zero.
Therefore, the compatibility conditions may be written as
Δ1 = 0 ⎫
⎬ (a)
Δ 2 = 0⎭
In which, Δ1 represents the displacement along the direction of X 1 (i.e., the vertical displacement at
point B) due to the combined effect of the known external load and the unknown redundants, X 1 and X 2 ;
whereas Δ 2 represents the displacement along the direction of X 2 (i.e., the horizontal displacement at
point B) due to the combined effect of the known external load and the unknown redundants, X 1 and X 2 .
For a linear elastic system, in order to determine the displacements Δ1 and Δ 2 by unit load method,
we subject the primary frame separately to a unit value of each of the redundants X 1 and X 2 [Figs.10.9
(c) and (d), respectively] and the external load [Fig.10.9 (e)]. The displacements Δ1 and Δ 2 in equation
(a) will be thusly, by the principle of superposition, rewritten as follows:
Δ1 = δ11 X 1 + δ12 X 2 + Δ1P ⎫
⎬ (b)
Δ 2 = δ 21 X 1 + δ 22 X 2 + Δ 2 P ⎭
Substituting Eq. (a) into Eq. (b), we obtain
δ11 X 1 + δ12 X 2 + Δ1P = 0 ⎫

⎬ (10-3)
δ 21 X 1 + δ 22 X 2 + Δ 2 P = 0 ⎭
(a) (b)
P1 P1
C D C D
P2 P2
A A X2
B B
X1
(d) (e)
(c)
P1
C D C D C D
P2
δ 21
δ11 X2 =1
A B A A B
δ12 B
X1 = 1 δ 22 Δ1 P
Δ2P
Fig.10.9 A structure with two degrees of indeterminacy
(a) original structure; (b) primary system; (c) primary structure under action of X 1 = 1;
(d) primary structure under action of X 2 = 1; (e) primary structure under action of external loading
This equation is the force-method equation of a structure with second degrees of indeterminacy. The
meanings of the coefficients and free terms of equation (10-3) are explained as follows (note that the first
subscript denotes the location of the displacement, and the second subscript indicates the location of the
unit load causing the displacement):
δ11 and δ 21 represent, respectively, the displacements of the primary frame at the action points and
in the directions of X 1 and X 2 due to unit value of the redundant X 1 alone, as shown in Fig.10.9 (c).
δ12 and δ 22 identify, respectively, the displacements of the primary frame at the action points and in
the directions of X 1 and X 2 due to unit value of the redundant X 2 alone, as shown in Fig.10.9 (d).
Δ1P and Δ 2 P denote, respectively, the displacements of the primary frame at the action points and in
the directions of X 1 and X 2 due to the external loads alone, as shown in Fig.10.9 (e).
Since the coefficient δ and the free term Δ existing in the force-method equations are the
displacements of a primary structure that is statically determinate, they can be evaluated by using the
methods discussed previously in this text, for example, by unit load method.
Once the redundants X 1 and X 2 have been determined by solving equation (10-3), the other
response characteristics (reactions and internal forces) of the structure can be evaluated either by
equilibrium equations or by method of superposition. For example, the bending moment on an arbitrary
cross section of the rigid frame can be expressed as
M = M1 X 1 + M 2 X 2 + M P (10-4)
Where M P denotes the bending moment on the arbitrary cross section of the primary rigid frame
due to the known external loading alone; M 1 and M 2 represent, respectively, the bending moments on
the arbitrary cross section of the primary rigid frame due to X 1 = 1 and X 2 = 1 alone.
Different primary structures and unknowns can be selected for the analysis of an indeterminate
structure. The structure shown in Fig10.9 (a) can select the primary structures shown in Fig.10.10 (a) or (b)
as its primary structure. Although the form of the canonical equation (10-3) is the same, the physical
meanings of redundants ( X 1 and X 2 ) and compatibility equation (10-3) are different. As shown in
Fig.10.10 (a), X 2 represents the moment provided by support A; while Δ 2 = 0 identifies that the slope
at support A of the original structure must be equal to zero. However, in Fig.10.10 (b), X 2 denotes the
bending moment of the left and right sections adjacent to middle section E; whereas Δ 2 = 0 indicates that
(a) (b) X2 (c)

C D C X1
D C
E D
X2
X2 X1 X1
B
A A B A B
Fig.10.10 Multiple choice of a primary structure
the relative rotation of the left and right sections adjacent to middle point of the beam must be equal to zero.
However, additional consideration has to be taken into that an instantaneously unstable system cannot be
used as a primary structure. For example, the system shown in Fig.10.10 (c) cannot be employed as a
primary structure because it is an instantaneously unstable system.
10.3.2 Canonical equations for structures with n degrees of indeterminacy
For a general case in which a structure is statically indeterminate to n degree, when employing force
method to analyze the structure, the primary unknowns are n redundant forces, X 1 , X 2 ," , X n ; the
primary structure is a determinate structure that is obtained by removing n excess restraints from the
original indeterminate structure to form a stable system; the compatibility equations are the consistent
deformation conditions at the locations and in the directions corresponding to the n redundant forces of the
original structure, that is, the displacements of the primary structure at the locations and in the directions of
the n redundants due to the combined effect of the known external loading and the unknown redundants
( X 1 , X 2 ," , X n ) must be identical to the actual displacements of the original indeterminate structure. For
a linear elastic system, by principle of superposition, the n compatibility conditions may be expressed as
δ11 X 1 + δ12 X 2 + " + δ1n X n + Δ1P = 0 ⎫
δ 21 X 1 + δ 22 X 2 + " + δ 2 n X n + Δ 2 P = 0 ⎪⎪
⎬ (10-5)
"" ⎪
δ n1 X 1 + δ n 2 X 2 + " + δ nn X n + Δ nP = 0 ⎪⎭
The equation (10-5) is the general form of force-method formulation for a structure with n degrees of
indeterminacy, referred to as the canonical equation of force method.
In equation (10-5), the flexibility coefficient δ ij and free term Δ iP represent, respectively, the
displacements of a primary structure due to a unit load and the external loading. Note that the first subscript
of δ ij or Δ iP indicate the location and direction of the displacement, whereas the second subscript is
used to indicate that the displacement is due to the unit value of jth redundant or the external loading. That
is, δ ij represents the displacement of the primary structure at the action point and in the directions of X i
due to unit value of the redundant X j alone; while Δ iP denote the displacement of the primary structure
at the action point and in the directions of X i due to the external loading alone.
The sign convention is that when the direction of δ ij or Δ iP is the same as that of its corresponding
redundant X i , δ ij or Δ iP will have positive sign and vice versa.
In equation (10-5), the flexibility coefficients located on the main diagonal such asδ11 , δ 22 ,"δ nn
are termed main coefficients, which are always greater than zero. The other coefficients, δ ij (i ≠ j) , are
named secondary coefficients, which may be greater than, smaller than or equal to zero.
Recall from the law of reciprocal displacements, the secondary coefficients δ ij and δ ji are identical.
That is,
δ ij = δ ji
By using the definition of matrix multiplication, Eq. (10-5) can be expressed in matrix form as
⎡δ11 δ12 " δ 1n ⎤ X
⎧ 1 ⎫ ⎧Δ1P ⎫ ⎧0 ⎫
⎢δ ⎥
⎢ 21 δ 22 " δ 2 n ⎥ ⎪ X 2 ⎪ ⎪ Δ 2 P ⎪ ⎪0 ⎪
⎨ ⎬+⎨ ⎬=⎨ ⎬ (10-6)
⎢" " " " ⎥ ⎪" ⎪ ⎪" ⎪ ⎪"⎪
⎢ ⎥
⎣δ n1 δ n 2 " δ nn ⎦ ⎩ X n ⎭ ⎩Δ nP ⎭ ⎩0 ⎭
In which, the matrix composing of flexibility coefficients δ ij is called the flexibility matrix. Based on this
reason, the canonical equation (10-5) is also referred to as the flexibility equation, and force method as
flexibility method.
After the flexibility coefficients and free terms have been determined, the redundants
X 1 , X 2 ," , X n can be evaluated by solving equation (10-5), then the internal forces and their diagrams of
the structure can be obtained either by equilibrium conditions or by principle of superposition. The
equations used to determine, by principle of superposition, the internal forces can be expressed as
M = M1 X1 + M 2 X 2 + " + M n X n + M P ⎫
⎪
Q = Q1 X 1 + Q2 X 2 + " + Qn X n + QP ⎬ (10-7)
N = N1 X 1 + N 2 X 2 + " + N n X n + N P ⎪⎭
Where, M i , Qi and Ni identify, respectively, the bending moment, shear and axial forces of the
primary structure due to X i = 1 alone; M P , QP and N P denote, respectively, the bending moment,
shear and axial forces of the primary structure due to only the external loading.
An alternative way to construct the diagrams of shear force Q and axial force N of an indeterminate
structure when employing equation (10-7) is that first draw the bending moment diagram and then
determine the shear and axial forces by using the equilibrium conditions of the free bodies of each of the
members or joints of the structure.
10.4 Statically Indeterminate Beams, Rigid Frames and Bent Frames
10.4.1 Statically indeterminate beams and rigid frames
A common means used in analysis of statically indeterminate beams and rigid frames by force method
is that the effect on displacements due to the axial and shearing deformations is neglected and only that due
to the bending deformation is under consideration. In the circumstance, the flexibility coefficients and free
terms existing in force-method equations can be expressed as
M i2 ⎫
δ ii = ∑ ∫ dx ⎪
EI ⎪
MiM j ⎪
δ ij = ∑ ∫ dx ⎬ (10-8)
EI ⎪
MiM P ⎪
Δ iP = ∑ ∫ dx ⎪
EI ⎭
Example 10-1
Draw the bending moment M and shear force Q diagrams for the beam with two fixed supports and
subjected to uniformly distributed load of intensity q as shown in Fig.10.11 (a) by force method.
Solution
(a) (b) (c)
q M P diagram
X1 q X2 q
A B
A EI C B A B X3
D E
ql 2
l l
8
(d) (e) l (f)

M 2 diagram
X1 = 1 X2 =1 M 3 diagram
M 1 diagram
A B A A B X3 = 1
B
M 3 = 0 N 3 = −1
l
Fig.10.11 figures of example 10-1
(a) a beam with two fixed supports; (b) primary system; (c) primary structure under action of external loading;
(d) primary structure under action of X 1 = 1; (e) primary structure under action of X 2 = 1;
(f) primary structure under action of X 3 = 1
(1) Select primary system

This is an indeterminate beam with three degrees of indeterminacy. The restraints against rotation at
ends A and B and horizontal translation at end B of the fixed beam are then removed to obtain the simply
supported primary beam shown in Fig.10.11 (b).
(2) Develop compatibility equations
Noting that the slopes of the actual indeterminate beam at the fixed supports A and B and the
horizontal translation at the support B are zero, we write the compatibility equations as
δ11 X 1 + δ12 X 2 + δ13 X 3 + Δ1P = 0 (a)
δ 21 X 1 + δ 22 X 2 + δ 23 X 3 + Δ 2 P = 0 (b)
δ 31 X 1 + δ 32 X 2 + δ 33 X 3 + Δ3 P = 0 (c)
(3) Calculation flexibility coefficients and free terms

Since the flexibility coefficients and free terms are all displacements of the primary beam (a simple
beam), when just bending deformation is considered, only the bending moment diagrams, M P due to the
external loading, M 1 due to X 1 = 1 , M 2 due to X 2 = 1 and M 3 due to X 3 = 1 of the primary
beam are needed to construct, as shown in Fig.10.11 (d) through (f).
By graph-multiplication method, we write
M 12 1 1 2 l
δ11 = ∫ dx = ( × l ×1) × =
EI EI 2 3 3EI
M 22 1 1 2 l
δ 22 = ∫ dx = ( × l ×1) × =
EI EI 2 3 3EI
MM 1 1 1 l
δ12 = δ 21 = ∫ 1 2 dx = − ( × l ×1) × = −
EI EI 2 3 6 EI
When calculating δ 33 , the effect on the displacement due to the axial deformation must be considered
because of M 3 = 0 . Thus
M 32 N2 l l
δ 33 = ∫ dx + ∫ 3 dx = 0 + =
EI EA EA EA
MM
δ13 = δ 31 = ∫ 1 3 dx = 0
EI
M M
δ 23 = δ 32 = ∫ 2 3 dx = 0
EI
M 1M P 1 2 1 1 ql 3
Δ1P = ∫ dx = ( × l × ql 2 ) × =
EI EI 3 8 2 24 EI
M 2M P ql 3
Δ2P = ∫ dx = −
EI 24 EI
M M NN
Δ3P = ∫ 3 P dx + ∫ 3 P dx = 0
EI EA
Noting that the minus sign of Δ 2 P is the reason that the diagrams of M 2 and M P are located in
different sides of the beam.

(4) Determine the unknown redundants
By substituting the expressions for the flexibility coefficients and free terms into the compatibility
equations [equations (a) through (c)], we obtain
⎧ ql 2
⎪ 1
2 X − X 2 + =0
⎪ 4
⎪ ql 2
⎨− X 1 + 2 X 2 − =0
⎪ 4
⎪ l X =0
⎪⎩ EA 3
Solving the equations, we find

1 2 1 2
X1 = − ql , X2 = ql , X3 = 0
12 12
X 3 = 0 implies that the horizontal reactions of an indeterminate beam with two fixed supports subjected
to vertical loading must be zero. So to analyze this beam, we need to select only two of the remaining four
reactions as the redundants. The compatibility conditions may become
δ11 X 1 + δ12 X 2 + Δ1P = 0
δ 21 X 1 + δ 22 X 2 + Δ 2 P = 0
(5) Construct internal force diagrams
Once the redundants are determined the internal forces at the end sections of the beam can now be
determined in terms of superposition of the internal forces of the primary beam due to the external loading
and each of the redundants. And then the internal force diagrams will be drawn.
① Bending moment diagram
By the equation of superposition for bending moments, we write
M = M1 X1 + M 2 X 2 + M 3 X 3 + M P
By taking the values of M AB and M BA as the ordinates and laying them off by connecting them in
dashed lines on the tension side of the beam AB, and then superposing the bending moment diagram of a
corresponding simple beam subjected to the uniformly distributed load, the final bending moment diagram
can be obtained as shown in Fig.10.12 (a).
② Shear force diagram
The shear forces at the ends of the beam can be fined by considering the equilibrium conditions of the
free body of the beam [Fig.10.12 (b)] because the end bending moments are now known, and shear force
diagram can be thusly depicted. Considering the equilibrium conditions of the free body of the beam, we
write
l 1 1 1
∑M A = 0,
ql × − ql 2 + ql 2 + QBA × l = 0 ,
2 12 12
QBA = − ql
2
1
∑ M B = 0 , QAB = 2 ql
The shear force diagram is shown in Fig.10.12 (c).
(a) (b) (c)

ql 2 ql 2
12 q 12 ql
ql 2 ql 2
2 B
12 12
A D E B A B A ql
ql
2
l
2
24 QAB QBA
Fig.10.12 Bending moment and shear force diagrams

(a) M diagram; (b) free body diagram of beam AB; (c) Q diagram
When the internal force diagrams have been finished the elastic curve of the deformed beam may be
sketched by referring the bending moment diagram as shown in dashed lines in Fig.10.11 (a). The zero
points on the bending moment diagram are the inflection points of the elastic curve of the deformed beam.
In the segment DE, because the lower fibers are in tension its elastic curve is concave upward; whereas in
the segments AD and EB, since the upper fibers are in tension their elastic curves are concave downward.
Example 10-2
Draw the internal force diagrams for the indeterminate rigid frame shown in Fig.10.13 (a).
Solution

This is a single indeterminate rigid frame. The restraint against horizontal translation at hinged support
B is removed, which is done by replacing the hinged support by a roller support, to obtain the primary rigid
frame shown in Fig.10.13 (b).
(2) Develop compatibility equation
Noting that the horizontal translation at support B of the primary rigid frame due to the combined
effect of the known external loading and the unknown redundant X 1 must be zero, we write the
compatibility equation as
δ11 X 1 + Δ1P = 0
(3) Calculate flexibility coefficient and free term
The bending moment diagrams M P due to the external loading and M 1 due to X 1 = 1 of the
primary rigid frame are depicted out as shown in Figs.10.13 (c) and (d).
q = 10 kN m q
C D C D
I1
I1 = 2I2
6m
I2 I2
(a) (b)
X1
A B A B
6m
q 6 6
6 6
C D
C D
45
M P diagram
M1
(unit: kN ⋅ m)
diagram
(c) (d) X1
A B A B
Fig.10.13 Figures of example 10-2

(a) a redundant rigid frame; (b) primary system;
(c) primary structure under action of external loading;
(d) primary structure under action of X 1 = 1
By using the graph-multiplication method, we find

M 1M P 1 2 1080
Δ1P = ∑ ∫ dx = − ( × 6 × 45) × 6 = −
EI EI1 3 EI1
M12 1 2 1 2 216 144
δ11 = ∑ ∫ dx = (6 × 6 × 6) + ( × 6 × 6) × ( × 6) = +
EI EI1 EI 2 2 3 EI1 EI 2
Since I1 = 2 I 2 , δ11 is simplified as
504
δ11 =
EI1
(4) Determine the unknown redundant
Substituting the expressions for δ11 and Δ1P into the compatibility equation, we determine the
redundant to be
504 1080
X1 − =0
EI1 EI1
X 1 = 2.14kN
It is instructive to note from above calculations that although the flexibility coefficient and free term in
compatibility equation are related to the absolute value of flexural rigidity EI of each of the members, the
final value of the unknown is only related to the relative value of flexural rigidity EI of each of the
members. Therefore, only the relative value of flexural rigidity EI of each of the members of a statically
indeterminate structure is necessary when analyzing the structure.
(5) Construct internal force diagrams
① Bending moment diagram
The equation of superposition for bending moments now is
M = M1 X 1 + M P
By multiplying the values of M 1 by X 1 = 2.14kN and then superposing them with M P
diagram, the final bending moment diagram can be obtained as shown in Fig.10.14 (a).
② Shear force diagram
In a rigid frame, the shear forces at the ends of each of the members can be fined by considering the
equilibrium conditions of the free body of each of the members because the end bending moments of each
of the members are now known, and shear force diagram can thusly be depicted.
Considering the free body of member CD as an example shown in Fig.10.14 (b), by the equilibrium
conditions (when determining shear forces the axial forces do not need to take into account, so the axial
forces do not draw out), we write
∑M D = 0, QCD × 6 − 10 × 6 × 3 − 12.86 + 12.86 = 0 , QCD = 30kN
∑M C =0, QDC = −30kN

Using the similar approach, we can find the shear forces of members AC and BD. When the shear
forces of all the members have been determined the shear force diagram can be finished, as shown in
Fig.10.14 (c).
③ Axial force diagram

By using the known shear forces, the axial forces of each of the members can be calculated by the
equilibrium conditions of the free bodies of the associated rigid joints. After the axial forces are determined
the axial force diagram of the rigid frame can be drawn out conveniently.
12.86 45 12.86 (b)
12.86 12.86 10 kN m
12.86 12.86
C D
C D
6m
32.14
QCD QDC
(unit: kN ⋅ m) (d)
C N CD
2.14kN 2.14kN 2.14kN
A (a) B 30kN
N CA
30kN 30kN
30
2.14 2.14
D 2.14 30 30
C C
2.14 D
30
(unit: kN) (unit: kN)
(c) (e)
A B A B
2.14 2.14 30 30
Fig.10.14 Internal force diagrams of example 10-2

(a) M diagram; (b) free body diagram of member CD; (c) Q diagram;
(d) free body diagram of joint C ; (e) N diagram
Taking rigid joint C as an example, its free body diagram is shown in Fig.10.14 (d) (when determining
axial forces the bending moments do not need to take into account, so the bending moments do not draw
out). By the equilibrium conditions, we write
∑ X = 0, NCD + 2.14 = 0 , NCD = −2.14kN

∑Y = 0 , N CA + 30 = 0 , NCA = −30kN
By isolating the free body diagrams of rigid joints in an appropriate order, all the axial forces will be
determined by using two projection equations of equilibrium of each of the joints. The axial force diagram
is shown in Fig.10.14 (e).
Noting that in shear and axial force diagrams signs must be identified.
By referring to the bending moment diagram, the elastic curve of the deformed rigid frame may be
sketched as shown in dashed lines in Fig.10.13 (a). Since the member CD has two zero points on its
bending moment diagram, there are two inflection points lying on the elastic curve of the deformed member.
Because the bending moments of members AC and BD do not change their sign there is no inflection point
on the elastic curves corresponding to the portions of the two members. The cross-sectional rotations of
member ends meeting in the same joint (joints C and D) must be identical. In addition, because this rigid
frame is a symmetric structure subjected to a symmetric loading its elastic curve must be symmetric with
respect to the axis of symmetry of the frame as well.
Procedure for Analysis
Based on the foregoing discussion, we can develop the following step-by-step procedure for the
analysis of indeterminate structures by using force method.
Select excess restraint force(s) as the primary unknown(s) (redundants). The choice of redundant(s) is
merely a matter of convenience, and any reaction or internal force(s) can be selected as the redundant(s),
provided that the removal of the corresponding restraint(s) from the given indeterminate structure results in
a primary structure that is statically determinate and stable. The sense(s) of the redundant(s) is not known
and can be arbitrarily assumed. The actual sense(s) of the redundant(s) will be known after its magnitude(s)
has been determined by solving the compatibility equation(s). A positive magnitude for the redundant(s)
will imply that the sense initially assumed was correct, whereas a negative value of the magnitude will
indicate that the actual sense is opposite to the one assumed initially.
(2) Develop compatibility equation(s)
Write the compatibility equation(s) by setting the algebraic sum of the displacements of the primary
structure at the location(s) and in the direction(s) of the redundant(s) due to the external loading and the
redundant(s) equal to the given displacement(s) of the restraint(s) of the actual indeterminate structure.
(3) Calculate flexibility coefficients and free terms
a. First，draw internal force diagram(s) of the primary structure with only the external loading applied
to it.
b. Next, draw, respectively, internal force diagrams of the primary structure with only the unit value of
each of the redundants applied to it alone. The unit force (or moment) must be applied in the positive
direction of the redundant.
c. Finally, calculate the flexibility coefficients and free terms in the compatibility equation(s) by using
graph-multiplication method
(4) Determine the redundant(s)
Substitute the values of the flexibility coefficients and free terms computed in step 3 into the
compatibility equation(s), and solve for the unknown redundant(s).
(5) Draw internal force diagram(s)
First, draw the bending moment diagram of the structure by superposition of the bending moments of
the primary structure due to the external loading and due to the redundant(s). Then, draw shear and axial
force diagrams by using appropriate equilibrium equations.
10.4.2 Bent frames
A bent frame is composed of its roof (or roof beam) and columns. Fig.10.15 (a) shows a sketch of a
cut plane of the structure commonly used in one-storey precast workshops in industrial factories. The roof
(or roof beam) is usually simplified as a link with infinite axial rigidity, which is hinged at the tops of the
columns. The columns are usually with variable cross sections and fixed to their foundations. The
analytical model of a bent frame is shown in Fig.10.15 (b). The degree(s) of indeterminacy of a bent frame
is equal to its number of span(s). The primary system may be obtained by cutting the link(s) connecting the
columns, and the redundant(s) is a pair (or pairs) of forces equal in magnitude but opposite in direction,
which is actually the substitution of restraint(s) of the link(s), as shown in Fig.10.15 (c).
(a) (b) (c)

X1
P
P C EA = ∞ D P C D
C D
A B A B
A B
Fig.10.15 Computing model and primary system of a bent frame

(a) sketch of cross-sectional structure of a workshop; (b) computing model; (c) primary system
Since the axial rigidity of the link(s) is infinite, i.e., EA = ∞ , the effect to the displacements of axial
deformation of the link(s) will be neglected when calculating the flexibility coefficients and free terms in
force-method equation(s). Thusly, the expressions for the flexibility coefficients and free terms are still
expressed as form of equation (10-8).
Example 10-3
Analyze the two-span bent frame with different heights of columns as shown in Fig.10.16 (a) and draw
its bending moment diagram due to the action of a horizontal braking force of a crane.
Solution

The bent frame is indeterminate to the second degree. The primary system is obtained by cutting the
two links and by substituting them by two pairs of corresponding redundants, as shown in Fig.10.16 (b).
(2) Develop force-method equations (compatibility equations)
Noting that the relative horizontal displacements at the two cuts of the primary system due to the
combined effects of the given external loading and the unknown redundants X 1 and X 2 must be zero,
i.e., the axial displacements of the each of the two pairs of sections adjacent to the two cuts should remain
consistent, so we write the force-method equations as
δ11 X 1 + δ12 X 2 + Δ1P = 0
δ 21 X 1 + δ 22 X 2 + Δ 2 P = 0
(3) Calculate flexibility coefficients and free terms
Draw the bending moment diagrams, M P due to the external loading, M 1 due to X 1 = 1 and
M 2 due to X 2 = 1 of the primary system as shown in Figs.10.16 (c) through (e).
By employing the graph-multiplication method, we write
1 1 2 1 1 2 504
δ11 = ( × 6 × 6 × × 6) + ( × 6 × 6 × × 6) =
EI1 2 3 EI 2 2 3 EI 2
2 1 2
δ 22 = ( × 3 × 3 × × 3) +
EI1 2 3
2 ⎡1 2 1 1 1 2 ⎤
⎢ × 7 × 10 × ( × 10 + × 3) + × 7 × 3 × ( × 10 + × 3) ⎥
EI 2 ⎣ 2 3 3 2 3 3 ⎦
2270
=
3EI 2
1 ⎡1 2 1 ⎤ 144
δ12 = δ 21 = − ⎢⎣ 2 × 6 × 6 × ( 3 × 10 + 3 × 4) ⎦⎥ = − EI
EI 2 2
Δ1P = 0
(a) (b) X2
F EA = ∞ G
2m
I1 I1
3m
X1
P = 20kN P = 20kN
D EA = ∞ E
6m
7m
I1 I2 I2
I 2 = 6 I1
A B C
X2 =1
(c) (d)
X1 = 1
3 3
M 1 diagram M 2 diagram
(m) (m)
6 6
10 10
(e) (f)
20.238
P = 20kN 13.492
20
26.984
0.238
M P diagram M diagram
(kN ⋅ m) (kN ⋅ m)
160 11.562 55.898 92.540

(a) a bent frame with different heights of columns; (b) primary system;
(c) primary structure under action of X 1 = 1; (d) primary structure under action of X 2 = 1;
(e) primary structure under action of external loading; (f) bending moment diagram
1 ⎡1 2 1 ⎤
Δ2P = − ⎢ × 1× 20 × ( × 3 + × 2) ⎥
EI1 ⎣ 2 3 3 ⎦
1 ⎡ 1 2 1 1 1 2 ⎤
− ⎢ ( × 7 × 160 × ( × 10 + × 3) + × 7 × 20 × ( × 10 + × 3) ⎥
EI 2 ⎣ 2 3 3 2 3 3 ⎦
14480
=−
3EI 2
(4) Determine the unknown redundants
By substituting the expressions for the coefficients and free terms into the force-method equations, we
determine the redundants to be
504 X 1 − 144 X 2 = 0
2270 14480
−144 X 1 + X2 − =0
3 3
X 1 = 1.927kN ， X 2 = 6.746kN
(5) Construct bending moment diagram

The superimposing equation for bending moments now is
M = M1 X 1 + M 2 X 2 + M P
By multiplying the values of M 1 by X 1 = 1.927kN , M 2 by X 1 = 6.746kN and then
superposing them with M P diagram, the final bending moment diagram can be obtained as shown in
Fig.10.16 (f).
10.5 Statically Indeterminate Trusses and Composite Structures
10.5.1 Statically indeterminate trusses
A truss is assemblage of straight members connected at their ends by flexible connections, which are
modeled as frictionless hinges. The members of a truss are only subjected to axial forces when there is no
load applied between the ends of every member. In the circumstance, the flexibility coefficients and free
terms in force-method equations can be expressed as follows.
Ni Ni ⎫
δ ii = ∑
EA ⎪⎪
l
NN ⎪
δ ij = ∑ i j l ⎬ (10-9)
EA ⎪
NN ⎪
Δ iP = ∑ i P l ⎪
EA ⎭
The superimposing equation of axial forces for each of the members will be
N = N1 X 1 + N 2 X 2 + " + N P (10-10)
Example 10-4
Determine the axial forces in the members of the truss shown in Fig.10.17 (a). The elastic moduli of
the materials of each of the members are identical; the area of every member is tabulated in table 10-1.
Solution
(a) (5) (b)

(6) X1
(4)
3m
(10)
(7) (9) (8)
(1) (2) (3)

P = 30kN P
3 × 3 = 9m
(c) −10 (d) −0.7

X1 = 1
−14 0 −28 0 0
10 −14 30 −0.7 −0.7
1 1
10 20 20 0 −0.7 0
30kN
20 N1
10 N P (kN)

(a) an indeterminate truss; (b) primary system; (c) primary structure under
action of external loading; (d) primary structure under action of X 1 = 1

The truss is a single degree internally indeterminate truss. The primary system is obtained by cutting
the number 10 member and by substituting it by a pair of corresponding redundants, as shown in Fig.10.17
(b).
(2) Develop force-method equation (compatibility equation)
By the consistence deformation condition of member 10, i.e., the relative displacement in the direction
of centroidal axis at the cut through the member must be equal to zero, we write the force-method equation
as
δ11 X 1 + Δ1P = 0
Draw the axial force diagram N P due to the external loading and N1 due to X 1 = 1 of the primary
system as shown in Figs.10.17 (c) to (d). The coefficient and free term can be evaluated by equations
N12l
δ11 == ∑
EA
NN
Δ1P = ∑ 1 P l
EA
As shown in table 10-1
89.5
δ11 =
E
1082
Δ1P = −
E
Table 10-1 calculation of δ11 , Δ1P and axial forces N
A (cm 2 ) N P (kN) N12 l N1 N P l N = N1 X 1 + N P (k

member l (cm) N1
A A
(1) 300 15 10 0 0 0 10.0
(2) 300 20 20 -0.7 7.5 -210 11.5
(3) 300 15 20 0 0 0 -20.0
(4) 424 20 -14 0 0 0 -14.0
(5) 300 25 -10 -0.7 6 84 -18.5
(6) 424 20 -28 0 0 0 -28.0
(7) 300 15 10 -0.7 10 -140 1.5
(8) 300 15 30 -0.7 10 -420 21.5
(9) 424 15 -14 1 28 -396 -1.9
(10) 424 15 0 1 28 0 12.1
∑ 89.5 -1082

By substituting the expressions for the coefficient and free term into the force-method equation, we
determine the redundant to be
Δ1P 1082 E
X1 = − =− = 12.1kN
δ11 89.5E
(5) Calculate axial forces of each member
The superimposing equation of axial forces for each member now is
N = N1 X 1 + N P
The calculating results are tabulated in table 10-1.
10.5.2 Statically indeterminate composite structures
A statically indeterminate composite structure is a composition of flexural members and two-force

members. In the structure, the two-force members are subjected to only axial forces, whereas the flexural
members are undergone bending moments, shear forces and axial forces. However, the effects on
displacements due to the shear and axial forces are, compared to that due to bending moments, negligible in
the flexural members. Thusly, when evaluating the flexibility coefficients and free terms in force-method
equation(s) of a composite structure, only axial forces in two-force members and bending moments in
flexural members are taken into account. In the circumstance, the expressions calculating the coefficients
and free terms can be written as
M i2 Ni2 ⎫
δ ii = ∑ ∫ dx + ∑ l ⎪
EI EA ⎪
MiM j Ni N j ⎪
δ ij = ∑ ∫ dx + ∑ l ⎬ (10-11)
EI EA ⎪
MM NN ⎪
Δ iP = ∑ ∫ i P dx + ∑ i P l ⎪
EI EA ⎭
The superposing equations of bending moment for flexural members and axial force for two-force
members will be written as
M = M 1 X 1 + M 2 X 2 + " + M P ⎪⎫
⎬ (10-12)
N = N1 X 1 + N 2 X 2 + " + N P ⎪⎭
Example 10-5
Analyze the internally indeterminate trussed beam shown in Fig.10.18 (a). The inertial moment of the
−4
cross section of the beam is I = 1× 10 m ; the cross sectional area of each of the two-force members is
4
A = 1×10−3 m 2 ; the elastic modulus of the materials of the structure is E = constant .

Solution

The trussed beam is a single degree internally indeterminate structure. Its primary system is obtained
by cutting the vertical two-force member and by substituting it by a pair of corresponding redundants, as
shown in Fig.10.18 (b).
(2) Develop force-method equation (compatibility equation)
By the consistence deformation condition of the vertical two-force member, i.e., the relative axial
displacement of the up and down sections adjacent to the cut must be equal to zero, we write the
force-method equation as
δ11 X 1 + Δ1P = 0
(a) (b)
10kN/m 10kN/m
I X1
2m
2A I
A A
4m 4m
4m
(c) (d) 10kN/m NP = 0
N1 = 1
2 X1 = 1 0 80 0
0
− 5/2 − 5/2
M P (kN ⋅ m), N P
M 1 (m), N1
(e) (f)
1.76m 9.8 N = −44.9 1.5m 20 N = −50

−44.9
15.4
−50
11.3
+50.2 +50.2
+55.9 +55.9
M (kN ⋅ m), N (kN) when A → ∞, M (kN ⋅ m), N (kN)

(a) an indeterminate trussed beam; (b) primary system; (c) primary structure under action of X 1 = 1;
(d) primary structure under action of external loading; (e) bending moment diagram and axial force;
(f) bending moment diagram and axial force when A → ∞

Draw the bending moment diagram and axial force, M 1 and N1 due to X 1 = 1 , M P and N P
due to the external loading of the primary structure as shown in Figs.10.18 (c) to (d). The coefficient and
free term can be evaluated as follows
M 12 N 2l 1 4× 2 2× 2
δ11 = ∑ ∫ dx + ∑ 1 = −4
(2 × × )+
EI EA E × 1× 10 2 3
1 ⎡12 × 2 5 2 ⎤
−3 ⎢
+ 2 × (− ) × 2 5⎥
E × 1× 10 ⎣ 2 2 ⎦
1 1
= (1.067 ×10 + 0.122 × 10 ) = (1.189 × 105 )
5 5
E E
MM NN
Δ1P = ∑ ∫ 1 P dx + ∑ 1 P l
EI EA
1 4 × 2 × 80 5 × 2 1
= −4
(2 × × ) + 0 = (5.333 ×106 )
E × 1×10 3 8 E
Δ1P 5.333 ×106
X1 = − =− = −44.9kN
δ11 1.189 ×105
(5) Calculate internal forces of each member
The superimposing equations of internal forces now should be
M = M1 X 1 + M P
N = N1 X 1 + N P
The bending moment diagram and axial forces of the composite structure are depicted out as shown in
Fig.10.18 (e).
(6) Discussion
It is noted from comparing the bending moment diagrams M [Fig.10.18 (e)] and M P [Fig.10.18 (d)],
which is actually the bending moment diagram of a simple beam, that because of the elastic restraint
provided by two-force members the bending moment M in the beam of the composite structure has been
reduced by a percentage of 80.75% , from 80kN ⋅ m to 15.4kN ⋅ m .
If we change the magnitude of the cross-sectional area A of the two-force members, the distribution of
internal forces in the composite structure will be changed as well. When the area A is reduced, the positive
bending moment of the beam will become greater, whereas its negative bending moment becomes smaller.
When the area A vanish the bending moment diagram M is the same as M P . Contrariwise, when the area
A is increased, the positive bending moment of the beam will become smaller, whereas its negative bending
moment becomes greater. When the area A goes infinitely great the elastic restraint in the middle of the
beam will become a rigid support and the bending moment diagram M of the beam will become the same as
that of a two-span continuous beam as shown in Fig.10.18 (f).
10.6 Analysis of Symmetric Structures
Many structures, because of aesthetic and/or functional requirements, are arranged in symmetric forms.
If a symmetric structure is linearly elastic, the response (i.e., member forces and deformations) of the entire
structure under symmetric loading can be obtained from the response of one of its portions separated by the
axes of symmetry. Thus only a portion (usually half) of the symmetric structure needs to be analyzed. In the
section, we discuss how to recognize structural symmetry and how to utilize it to simplify the
computational effort required in the analysis of symmetric structures.
When employing force method to analyze a statically indeterminate structure, the higher the degrees
of its indeterminacy, the more the required computational effort. The main calculation work is the solution
of compatibility equations, i.e., the evaluation of flexibility coefficients and free terms in compatibility
conditions and the solution of the simultaneous linear equations. In the circumstance, the key step to
simplify the calculation work in force method is the simplification of force-method equations. In other
words, if the approach making some or a part of the flexibility coefficients and free terms in canonical
equations to be zero is possible, the solution of the equations will be simplified. As we know, since the
coefficients located in the main diagonal in flexibility matrix [Eq. 10-6] are always positive and greater
than zero, obviously the principle of simplification of force-method equations will come down to the way
by which the secondary coefficients and free terms in flexibility matrix can be made to be zero as many as
possible. In fact, there are a few of simplifying methods such as the utilization of symmetry of structures,
the method of elastic centre and etc. The key points of the methods are the selections of reasonable primary
systems and primary unknowns. The objective of the section is the discussion of the utilization of
symmetry of structures.
In practical engineering, a lot of structures actually can be considered to be symmetric. The utilization
of symmetry will help us to simplify the analysis of the structures.
10.6.1 Symmetric structures and symmetric loadings
(1) Symmetric structures

A plane structure is considered to be symmetric with respect to an axis of symmetry in its plane if the
reflection (or mirror image) of the structure about the axis is identical in geometry (dimension and shape),
supports, and material properties (axial rigidity EA, flexural rigidity EI and shear rigidity GA etc.) to the
structure itself. Some examples of symmetric structures are shown in Fig. 10.19. For each structure, the
vertical axis of symmetry is identified as y-y axis and horizontal axis of symmetry as x-x axis. Note that the
reflection of each structure about its axis of symmetry is identical in geometry, supports and material
properties to the structure itself.
(2) Symmetry of loadings

Symmetric loadings: A loading is considered to be symmetric with respect to an axis in its plane if
the reflection (or mirror image) of the loading about the axis is identical to the loading itself. An example of
symmetric loading is shown in Fig.10.20 (b).
(a) y axis of symmetry (b) y

axis of symmetry
EI1
EI1
2
b
h EI 2 EI 2
EI 2 EI 2 x axis of symmetry x
2
b
EI1
a a
l l
2 y 2
2 y 2
Fig.10.19 Symmetric structure

(a) a symmetric structure with one axis of symmetry;
(b) a symmetric structure with two axes of symmetry
(a) (b) (c)

P P P P
P a a a a a
2 2 2 2
C C C
Fig.10.20 symmetric and antisymmetric loadings

(a) any general loading; (b) symmetric loading; (c) antisymmetric loading
Antisymmetric loadings: A loading is considered to be antisymmetric with respect to an axis in its

plane if the negative of the reflection (or mirror image) of the loading about the axis is identical to the
loading itself. An example of antisymmetric loading is shown in Fig.10.20 (c).
Decomposition of a general loading into symmetric and antisymmetric components: Any general
loading [Fig.10.20 (a)] can be decomposed into symmetric [Fig.10.20 (b)] and antisymmetric [Fig.10.20 (c)]
components with respect to an axis by implementing the following procedure:
① Divide the magnitudes of the forces and/or moments of the given loading by 2 to obtain the half
loading.
② Draw a reflection (or mirror image) of the half loading about the specified axis.
③ Determine the symmetric component of the given loading by adding the half loading to its
reflection.
④ Determine the antisymmetric component of the given loading by subtracting the symmetric
loading component from the given loading.
10.6.2 Analysis by using symmetric system
When analyzing a statically indeterminate symmetric structure, the utilization of a symmetric system
can considerably expedite the analysis of the structure. Considering the portal rigid frame shown in Fig.20
(a) P (b)
X1 X1 = 1
X2 X2
X3
M 1 diagram
(c) X2 =1 (d)
X3 = 1
M 2 diagram M 3 diagram
Fig.10.21 Bending moment diagrams and deformed shapes of symmetric primary structure
(a) primary system; (b) primary structure under action of X 1 = 1;
(c) primary structure under action of X 2 = 1; (d) primary structure under action of X 3 = 1
(a), a symmetric system is obtained by cutting the frame through its axis of symmetry and by adding three
pairs of redundants as substitution of the restraints at the cut, as shown in Fig.10.21 (a). Apparently, by the
definitions of symmetric and antisymmetric loadings, both of the pair of bending moments X 1 and the
pair of axial forces X 2 is symmetric loadings, whereas the pair of shear force X 3 is antisymmetric
loading.
By setting the algebraic sum of the relative rotation, relative horizontal displacement and vertical
relative deflection at the location of the cut of the symmetric primary structure due to each of the
redundants and the external loading equal to zero, we can write compatibility conditions as
δ11 X 1 + δ12 X 2 + δ13 X 3 + Δ1P = 0 ⎫
⎪
δ 21 X 1 + δ 22 X 2 + δ 23 X 3 + Δ 2 P = 0 ⎬ (a)
δ 31 X 1 + δ 32 X 2 + δ 33 X 3 + Δ 3 P = 0 ⎭⎪
Figs.10.21 (b) through (d) show, separately, the bending moment diagrams and deformed
configurations due to X 1 = 1 , X 2 = 1 and X 3 = 1 . It can be observed that the bending moment
diagrams of M 1 , M 2 and their corresponding deforming shapes are symmetric, while that of M 3 and
its corresponding deforming shape are antisymmetric. Therefore, the secondary flexibility coefficients δ13 ,
δ 31 , δ 23 and δ 32 will be
M 1M 3
δ13 = δ 31 = ∑ ∫ ds = 0
EI
M M
δ 23 = δ 32 = ∑ ∫ 2 3 ds = 0
EI
Thusly, the force-method equations can be simplified as
δ11 X 1 + δ12 X 2 + Δ1P = 0 ⎫
⎪
δ 21 X 1 + δ 22 X 2 + Δ 2 P = 0 ⎬ (b)
δ 33 X 3 + Δ 3 P = 0 ⎪⎭
Obviously, the force-method equations are now decomposed into two sets. The first set only involves
symmetric unknowns X 1 and X 2 , whereas the second set only antisymmetric unknown X 3 .
The free terms in equation (b) can be also simplified according to the symmetry of loadings.
Symmetric structures subjected to symmetric loadings: when a symmetric structure is subjected to a
symmetric loading the bending moment diagram of the symmetric primary structure due to the external
loading is symmetric as well. As shown in Fig.10.22 (a), the bending moment diagram of M P is
symmetric, while the bending moment diagram of M 3 is antisymmetric [Fig.10.21 (d)]. In the
circumstance, the free term Δ 3P will be
M 3M P
Δ3P = ∑ ∫ ds = 0
EI
Recalling from equation (b), the antisymmetric unknown X 3 must be equal to zero, i.e., X 3 = 0 .
Only the symmetric unknowns X 1 and X 2 exerting on the symmetric section [Fig.10.22 (b)] are needed to
determine.
Symmetric structures subjected to antisymmetric loadings: For the case that a symmetric structure is
subjected to an antisymmetric loading, the bending moment diagram of the symmetric primary structure
due to the external loading is also antisymmetric. As shown in Fig.10.22 (c), the bending moment diagram
of M P is antisymmetric, while the bending moment diagrams of M 1 and M 2 are symmetric [Fig.10.21
(b) and (c)]. Therefore, the free terms Δ1P and Δ 2 P will vanish
M 1M P
Δ1P = ∑ ∫ ds = 0
EI
M M
Δ2P = ∑ ∫ 2 P ds = 0
EI
(a) (b) P P
P P
2 X1 2
2 2
X2 A X2
X3 = 0
M P diagram
P P P P
(c) (d)
2 2 2 2
X3
X1 = X 2 = 0
M P diagram
Fig.10.22 Symmetric primary structure under action of symmetric and antisymmetric loadings
(a) symmetric primary structure under action of symmetric loading; (b) only symmetric unknowns X 1 and X 2 ;
(c) symmetric primary structure under action of antisymmetric loading;
(d) only antisymmetric unknown X 3
It is noted from the first two equations of expression (b) that the symmetric unknowns X 1 and X 2 must
be equal to zero, i.e., X 1 = 0 and X 2 = 0 . Only the antisymmetric unknown X 3 exerting on the symmetric
section [Fig.10.22 (d)] is needed to evaluate.
Generally, when a symmetric structure is subjected to a loading that is symmetric with respect to the
structure's axis of symmetry, the response of the structure (i.e., reactions, internal forces and displacements)
is also symmetric, with the point(s) of the structure at the axis of symmetry neither rotating (unless there is
a hinge at such a point) nor deflecting perpendicular to the axis of symmetry [Fig.10.20 (b)]; thus, in the
symmetric system, there is no antisymmetric unknown(s) on the symmetric axis and only symmetric
unknowns are existed on the symmetric axis and they are needed to determine.
When a symmetric structure is subjected to a loading that is antisymmetric with respect to the
structure's axis of symmetry, the response of the structure (i.e., reactions, internal forces and displacements)
is also antisymmetric, with the point(s) of the structure at the axis of symmetry not deflecting in the
direction of the axis of symmetry [Fig.10.20 (c)]; thusly, in this system, there is no symmetric unknown(s)
on the symmetric axis and only antisymmetric unknown(s) is existed on the symmetric axis and it is needed
to determine.
10.6.3 Select half a structure to analyze
(1) Frames with odd-number spans

① Symmetric structures subjected to symmetric loadings
For emphasis, we rewrite, in italic form, the performance of a symmetric structure under action of a
symmetric loading. When a symmetric structure is subjected to a loading that is symmetric with respect to
the structure's axis of symmetry, the response of the structure is also symmetric, with the points of the
structure at the axis of symmetry neither rotating (unless there is a hinge at such a point) nor deflecting
perpendicular to the axis of symmetry. Thus, to determine the response (i.e., member forces and
deformations) of the entire structure, we need to analyze only half the structure, on either side of the axis of
symmetry, with symmetric boundary conditions (i.e., slopes must be either symmetric or zero, and
deflections perpendicular to the axis of symmetry must be zero) at the axis. The response of the remaining
half of the structure can then be obtained by reflection.
Consider a symmetric portal rigid frame subjected to a loading that is symmetric with respect to the
frame's axis of symmetry as an example, shown in Fig.10.23 (a). The deflected shape (elastic curve) of the
frame is also shown in the figure. It can be seen that, like the loading, the deflected shape is symmetric with
respect to the axis of symmetry of the frame. Note that the slope and the horizontal deflection are zero at
point C, where the axis of symmetry intersects the frame, whereas the vertical deflection at C is not zero.
The response of the entire frame can be determined by analyzing only half the frame, on either side of the
axis of symmetry. The left half of the frame cut by the axis of symmetry is shown in Fig.10.23 (b). Note
that the symmetric boundary conditions are imposed on this substructure by supporting it at the end C by a
double link support, which prevents the rotation and the horizontal deflection at the axis of symmetry but
cannot prevent the vertical deflection along the axis. Once the response of the left half of the frame has
been determined by analysis, the response of the right half can be obtained from that of the left half by
reflection (or mirror image).
② Symmetric structures subjected to antisymmetric loadings
(a) (b) q
q
D E
D
C C
C′
A B A
Fig.10.23 Odd-number-span rigid frame under action of symmetric loading

(a) a portal rigid frame subjected to symmetric loading;
(b) equivalence of half the frame;
P D C P P D C
C′ E
A (a) B A (b)
Fig.10.24 Odd-number-span rigid frame under action of antisymmetric loading

(a) a portal rigid frame subjected to antisymmetric loading;
(b) equivalence of half the frame;
For the same reason of emphasis, we rewrite, in italic form, the performance of a symmetric structure
under action of an antisymmetric loading. When a symmetric structure is subjected to a loading that is
antisymmetric with respect to the structure's axis of symmetry, the response of the structure is also
antisymmetric, with the points of the structure at the axis of symmetry not deflecting in the direction of the
axis of symmetry. Thus to determine the response of the entire structure, we need to analyze only half the
structure, on either side of the axis of symmetry, with antisymmetric boundary conditions (i.e., deflections
in the direction of the axis of symmetry must be zero) at the axis. The response of the remaining half is
given by the negative of the reflection (or mirror image) of the response of the half structure that is
analyzed.
Take a symmetric frame subjected to a loading that is antisymmetric with respect to the frame's axis of
symmetry as an example, shown in Fig.10.24 (a). It can be seen that, like the loading, the deflected shape of
the frame is antisymmetric with respect to the frame's axis of symmetry. Note that the vertical deflection is
zero at point C, where the axis of symmetry intersects the frame, whereas the horizontal deflection and
slope at C are not zero. The response of the entire frame can be determined by analyzing only half the
frame, on either side of the axis of symmetry. The left half of the frame cut by the axis of symmetry is
shown in Fig.10.24 (b). Note that the antisymmetric boundary conditions are imposed on this substructure
by supporting it at end C by a roller support, which prevents the vertical deflection at the axis of symmetry
but cannot prevent the horizontal deflection and rotation at C. Once the response of the left half of the
frame has been determined by analysis, the response of the right half is given by the negative of the
reflection of the response of the left half.
(2) Frames with even-number spans
① Symmetric structures subjected to symmetric loadings
Consider an even-number-span symmetric rigid frame subjected to symmetric loading, as shown in
Fig. 10.25 (a). The left half of the frame with symmetric boundary conditions is shown in Fig. 10.25 (b). As
this figure indicates, if we neglect the axial deformation of column CD (in fact, the members located along
the axis of symmetry will undergo only axial deformations without bending), all of the displacements at
joint C have been restrained. Thus, the symmetric boundary conditions are imposed on this substructure by
(a) q (b) q
E F
E
C C
A D B A
Fig.10.25 Even-number-span rigid frame under action of symmetric loading

(a) a rigid frame subjected to symmetric loading; (b) equivalence of half the frame;
supporting it at the end C by a fixed support. Once the response of the left half of the frame has been
determined by analysis, the response of the right half is obtained by reflection (or mirror image). However,
it should be noted that there is an axial force in member CD (that is located along the axis of symmetry)
and it can be determined by appropriate equilibrium equations.
② Symmetric Structures Subjected to Antisymmetric Loadings
Take a six-degree-indeterminate symmetric rigid frame subjected to an antisymmetric loading as an
example, as shown in Fig.10.26 (a). The structure contains a member along the axis of symmetry. If we
halve the properties of the member, cross-sectional inertia moment I and area A (if there are any
concentrated loads and couples acting on the structure’s axis of symmetry, their magnitudes should be
halved as well), as shown in Fig.10.26 (b), only a pair of shear forces equal in magnitude but opposite in
direction should be imposed on the cut (at points C1 and C2 ). The half structure selected for analysis will
be simplified as the structure shown in Fig.10.26 (c).The antisymmetric boundary conditions are thusly
imposed on this substructure by supporting it at end C1 by a roller support, which prevents the vertical
deflection at the axis of symmetry but cannot prevent the horizontal deflection and rotation at C1 .
Note that the roller support just affects the magnitude of axial force in member CD that is along the
axis of symmetry and cannot undergo any axial deformations and has no any axial forces. Therefore, the
half structure selected for analysis may be further simplified as portal rigid frame shown in Fig.10.26 (d).
Once the response of the left half of the frame has been determined by analysis, the response of the right
half is given by the negative of the reflection (or mirror image) of the response of the left half. However, it
should be mentioned that there exist bending moments and shear forces in member CD that is located along
the axis of symmetry and the internal forces can be obtained by doubling the bending moments and shear
force in member C1 D1 [Fig.10.26 (d)].
Now we will summarize the analysis of simplification for symmetric structures as follows:
(1) Select symmetric primary system, and then the primary unknowns will be decomposed into two
sets; the first set only involves symmetric unknowns, whereas the second set only antisymmetric
unknown(s). Then, the secondary flexibility coefficients relating to the above two sets are equal to zero, i.e.,
δ ij = 0 ; the subscript i denotes the location and direction of a symmetric unknown, while the subscript j
represents that of an antisymmetric unknown. Consequently, a set of multi-unknown linear equations
(force-method equations) will become two sets of few-unknown linear equations. For different loading
conditions, there exist following three cases:
① When loading is symmetric as well, only the symmetric unknowns are needed to determine
(antisymmetric unknown(s) are equal to zero).
② When loading is antisymmetric, only the antisymmetric unknown(s) is needed to determine
(symmetric unknowns are equal to zero).
③ When loading is asymmetric, the loading may be decomposed into superposition of a symmetric
loading and an antisymmetric loading. The response (reactions, internal forces and displacements) of the
structure can be obtained by using principle of superposition. Or you may just employ symmetric primary
system and symmetric primary unknowns, which will simplify the force-method equations, to analyze the
structure directly.
(2) Select half a structure to analyze. It should be mentioned again that the internal forces and
displacements in the member that is located along the axis of symmetry of a symmetric structure with odd
number of span(s) are different from those of a symmetric structure with even number of spans under action
of symmetric or antisymmetric loadings. If the half a structure is selected to analyze, the loadings must be
symmetric or antisymmetric. Any general loading or asymmetric loading can be decomposed into
symmetric and antisymmetric components with respect to the structure’s axis of symmetry. Once the
response of half the structure has been determined by analysis, the response of the entire structure can be
given by using the principle of superposition.
(a) (b)
P C′ P E C1 C 2 F
E C F
l l
2 2
A D B A B
D1 D 2
(c) (d)
P E C1 P E C1
l l
2 2
A D1 A D1
Fig.10.26 Even-number-span rigid frame under action of antisymmetric loading

(a) a rigid frame subjected to antisymmetric loading; (b) halved portions of the frame;
(c) equivalence of the left half of the frame; (d) final equivalence of the left half of the frame
Example 10-6
Analyze the portal rigid frame shown in Fig.10.27 (a) and draw the bending moment diagram for the
frame and then discuss the relation of variation with the change of ratio k between the flexural rigidity of
the beam and that of the columns.
Solution
(1) Analysis of symmetry

The frame is symmetric with respect to the vertical axis passing through the middle of the beam and
the loading is asymmetric, which can be decomposed into symmetric and antisymmetric components as
shown in Figs.10.27 (b) and (c).
P P P P P
EI 2 2 2 2 2
h
EI1 EI1
(a) (b) (c)
l
P P P P 2
2 2 2 2 l
2
X1 M P diagram X1 = 1
M 1 diagram
Ph Ph l l
(d) (e) (f)
2 2 2 2

(a) a rigid frame subjected to asymmetric loading; (b) the rigid frame subjected to symmetric loading;
(c) the rigid frame subjected to antisymmetric loading; (d) symmetric primary system;
(e) primary structure under action of external loading; (f) primary structure under action of X 1 = 1
For the case shown in Fig.10.27 (b), if neglecting the axial deformation of the beam the two columns
are inactive (zero force member), only the beam is subjected to an axial force N = − P / 2 . The stressing
and deforming states not only satisfy the equilibrium conditions but also the compatibility conditions of the
displacements, so the internal force state is the real stressing state. In the circumstance, the bending
moment diagram of the entire frame is the same as that of the frame only subjected to the antisymmetric
loading as shown in Fig.10.27 (c), which can be obtained by subsequent steps.
(2) Primary system
In this example, the symmetric primary system is planed to use. Under action of antisymmetric
loading, the symmetric primary system is shown in Fig.10.27 (d). Since bending moment and axial force on
the cut (passing through the intersection point of the axis of symmetry and the beam) are symmetric
unknowns, only antisymmetric shear force X 1 is imposed on the cut.
(3) Force-method equation (compatibility condition)
Noting that the relative vertical displacements at the two ends of the cut of the primary system due to
the combined effects of the given external loading and the unknown redundant X 1 must be zero, we write
the force-method equation as
δ11 X1 + Δ1P = 0
(4) Flexibility coefficient and free term

Draw, respectively, the bending moment diagrams, M P due to the external loading and M 1 due to
X 1 = 1 of the primary structure as shown in Figs.10.27 (e) and (f). By employing the graph-multiplication
method, we write
⎡ l l 1 l l l 1 ⎤ l 2h l3
δ11 = 2 × ⎢( ⋅ h) × + ( × × )× × ⎥ = +
⎣ 2 2 EI1 2 2 2 3 EI 2 ⎦ 2 EI1 12 EI 2
1 Ph l Ph 2l
Δ1P = 2( × × h) × =
2 2 2 EI1 4 EI1
(5) Solve the force-method equation

Δ1P 6k Ph
X1 = − =−
δ11 6k + 1 2l
In which,
I2h
k=
I1l
By using the superimposing equation for bending moments M = M 1 X 1 + M P , we draw the bending
moment diagram as shown in Fig.10.28 (a).
(7) Discussion
① When the flexural rigidity of the beam is much smaller than that of the columns, i.e.,
EI 2 EI1 , their ratio k vanish, i.e., k → 0 . It implies that the beam is so weak that it can not resist the
rotation at the joints caused by the deflection of the columns. Consequently, the top ends of the columns
behave like hinged ends. The inflection points, where the bending moment is equal to zero, are at the top of
the columns. In the circumstances, the bending moment diagram is shown in Fig.10.28 (b).
② When the flexural rigidity of the beam is much greater than that of the columns, i.e., EI 2 EI1 ,
their ratio k become infinite great, i.e., k → ∞ . It means that the beam is strong enough to hold the
rotation at the joints. The restraint conditions at the two ends of the column are identical, thus the inflection
point locates at the mid-height of the columns. In this situation, the bending moment diagram is shown in
Fig.10.28 (d).
③ In general cases, as the bending moment diagram shown in Fig.10.28 (c) indicates, the inflection
point will locate somewhere between the top end and the mid-height of the column. When k = 3 , we can
approximately realize that the inflection point is at the mid point of the column. At the inflection point, the
bending moment is equal to zero and the shearing force is equal to P / 2 .
6k Ph
⋅
6k + 1 4
P P
k vanish
(b) Ph
(a)
2
6k + 2 Ph
⋅
6k + 1 4
18 Ph Ph
⋅
19 4 4
P P
k =3
k becomes infinite great
(c) (d)
Ph
20 Ph 4
i
19 4
Fig.10.28 Bending moment diagrams of example 10-6
(a) bending moment diagram in a general case; (b) bending moment diagram in the case of k → 0;
(c) bending moment diagram in the case of k = 3; (d) bending moment diagram in the case of k → ∞
Example 10-7
Analyze the statically indeterminate rigid frame shown in Fig.10.29 (a) and draw the bending moment
diagram for the frame.
Solution

In this example, we are planed to show how selecting the half of a symmetric structure for analysis. As
shown in Fig.10.29 (a), the frame is symmetric with respect to the vertical axis that coincides with the
centroidal line of member BE, and its loading is also symmetric. The left half of the frame is selected for
analysis, as shown in Fig.10.29 (b). Note that the symmetric boundary conditions are imposed on this
substructure by supporting it at the end E by a fixed support for the reason that the axial deformation of
member BE is small enough to be ignored.
(a) (b) (c)

E D E
D E F D
2I 2I 2I
X1 X2
2kN/m
2kN/m
2kN/m
2kN/m
4m
I I I I
B C
A A A
6m 6m
(d) (e) (f) (g) 2.56 2.56

2.56
D 1 E D X2 =1 E D E 2.56 E 2.56
D F
1.28 1.28
2kN/m
1 X1 = 1 1
4 2.72 2.72 4
4
M 1 diagram M 2 diagram M P diagram
(kN ⋅ m) A B C
A A A
M diagram
(kN ⋅ m)

(a) original rigid frame; (b) equivalence of half the frame; (c) primary system;
(d) primary structure under action of X 1 = 1; (e) primary structure under action of X 2 = 1;
(f) primary structure under action of external loading; (g) bending moment diagram
(2) Primary system

As shown in Fig.10.29 (b), the substructure is indeterminate to the second degree. The primary system
may be selected as the three hinged frame shown in Fig.10.29 (c). The primary unknowns, X 1 and X 2 ,
are end bending moments of member DE as shown in the figure.
(3) Force-method equations (compatibility conditions)
Noting that the relative rotation at the two positions of the primary system due to the combined effects
of the given external loading and the unknown redundants X 1 and X 2 must be zero, the force-method
equations will be written as
δ11 X1 + δ12 X 2 + Δ1P = 0
δ 21 X1 + δ 22 X 2 + Δ 2 P = 0
(4) Flexibility coefficients and free terms

Draw, respectively, the bending moment diagrams, M P due to the external loading, M 1 due to
X 1 = 1 and M 2 due to X 2 = 1 of the primary structure as shown in Figs.10.29 (d) through (f). By
using the graph-multiplication method, we write
7 1 1
δ11 = , δ12 = , δ 22 =
3EI 2 EI EI
16
Δ1P = , Δ2P = 0
3EI
determine the redundants to be
X1 = −2.56kN ⋅ m
X 2 = 1.28kN ⋅ m
By using the superimposing equation for bending moments M = M 1 X 1 + M 2 X 2 + M P , we draw
the bending moment diagram as shown in Fig.10.29 (g).
Example 10-8
Analyze the statically indeterminate rigid frame shown in Fig.10.30 (a) and draw the bending moment
diagram for the frame.
Solution

The purpose of this example is further to show how selecting the half of a symmetric structure for
analysis. As shown in Fig.10.30 (a), the frame and its loading have two axes of symmetry, x-x and y-y. The
one quarter of the frame is selected for analysis, as shown in Fig.10.30 (b). Note that the symmetric
boundary conditions under the action of the symmetric loading are imposed on this substructure by
supporting it at the ends B and D by a double-link support, respectively, for the reason that only the
translations of points B and D along the axes of symmetry will occur.
(2) Primary system
As shown in Fig.10.30 (b), the substructure is indeterminate to the first degree. The primary system
may be selected as the frame shown in Fig.10.30 (c). The primary unknown, X 1 , is end bending moment
of member CD as shown in the figure.
(a) (b) (c)

y
C D E C D C D
EI 2 EI 2 X1
2
2
l1
l1
EI1 P 2
x P EI1 EI1 P x P 2 B
B B
2
l1
EI 2 l2
A 2
l2 Pl12
EI1
y (f) y
⎛ l l ⎞
8⎜ 1 + 2 ⎟
C ⎝ 1
EI EI 2 ⎠
(d) 1 (e)
D
1
x B x
1
X1 = 1 Pl1
Pl12 2 Pl1l2
+
1 P 2
EI1 EI 2
4
M 1 diagram M P diagram A ⎛ l1 l ⎞
8⎜ + 2 ⎟
⎝ 1
EI EI 2 ⎠
y
(a) a rigid frame with two axes of symmetry; (b) one quarter structure of the frame;
(c) primary system; (d) primary structure under action of X 1 = 1; (e) primary structure
under action of external loading; (f) bending moment diagram of original frame

Noting that the slope at support D of the primary system due to the combined effects of the given
external loading and the unknown redundant X must be zero, the force-method equation is thusly written as
1
δ11 X1 + Δ1P = 0
(4) Flexibility coefficient and free term

Draw, respectively, the bending moment diagrams, M 1 due to X 1 = 1 and M P due to the
external loading of the primary structure as shown in Figs.10.30 (d) and (e). By using the
graph-multiplication method, we find
l1 l
δ11 = + 2
2 EI1 2 EI 2
Pl12
Δ1P =
16 EI1
(5) Solve the force-method equation (compatibility condition)

obtain
Pl12
EI1
X1 = −
l1 l
8( + 2 )
EI1 EI 2
First, draw the bending moment diagram of one quarter of the frame, as shown in the top left corner in
Fig.10.30 (f) by using the superimposing equation for bending moments M = M 1 X 1 + M P ; then, by
employing the property of symmetry of the frame, the final bending moment diagram will be finished as
shown in Fig.10.30 (f).
10.7 Statically Indeterminate Arches
Statically indeterminate arches are the types of structures which are widely adopted in engineering
structures. Such as the Chinese Zhaozhou Bridge [Fig.10.31] that was made of squared stone and designed
by Chinese craftsman, Lichun, in AD 605, it is a classical example
of utilization of arches in bridges. Nowadays, double arch bridges
are often selected in bridge structures. Arches are also commonly
used in roof structures of buildings. Fig.10.32 (a) shows a roof
structure that is reinforced concrete arch with a steel tie. Its
computing model is shown in Fig.10.32 (b). The lining of tunnels
Fig.10.31 Zhaozhou Bridge used underground engineering and water conservancy engineering
as shown in Figs.10.32 (c) and (d) are sorts of arches as well.
In practical engineering, most of statically indeterminate arches are two-hinged arches or hingeless
arches. Their analytical models are shown in Fig.10.33. Figs.10.33 (a) and (b) show, respectively, a
two-hinged arch with no tie and a two-hinge arch with a tie; whereas Figs.10.33 (c) and (d) depict,
respectively, a hingeless arch and a closed-loop arch that may be realized as a kind of a special hingeless
arch.
50cm
25cm
2.60m
(a) (b)
φ 12
2L90 × 60 × 10
15.6m
arch ring
side wall
(d)
(c)
inverted arch
Fig.10.32 arch structures

(a) an arched roof with a tie; (b) computing model of arches with ties;
(c) arched roof of tunnels; (d) double U shaped lining of a tunnel
(a) (b)
(d)
(c)
Fig.10.33 two-hinged arches and hingeless arches

(a) a two-hinged arch; (b) a two-hinged arch with a tie;
(c) a hingeless arch; (d) a ring shaped structure
10.7.1 Analysis of two-hinged arches
(1) Two-hinged arches

① Primary systems of two-hinged arches
A two-hinged arch [Fig.10.34 (a)] is a single degree indeterminate structure, so if the horizontal
reaction at hinged support B is selected to be the redundant X 1 , the primary system will be obtained by
replacing the hinged support of the given two-hinged arch by a roller support as shown in Fig.10.34 (b),
which now is a curved beam.
(a) (b)
X1
A B A B
(c) (d)
y
N
ϕ M
ϕ
C
C
y A Q
1 A x 1
B
X1 = 1
0
x
0 0
Fig.10.34 a two-hinged arch

(a) a two-hinged arch; (b) primary system; (c) primary structure
under action of X 1 = 1; (d) free body diagram of segment AC
② Force-method equation (compatibility condition)

Because the horizontal displacement at support B of the actual indeterminate arch is zero, the algebraic
sum of the displacements of the primary structure (a curved beam) at B due to the redundant X 1 and
external loading must also be zero. Thus the compatibility equation can be written as
δ11 X1 + Δ1P = 0 (a)
③ Determination of flexibility coefficient and free term

Since the simply supported arch is a curved member, so we cannot use graph-multiplication method
given previously to calculate the flexibility coefficient δ11 and free terms Δ1P in the compatibility
equation. Therefore, we will use integrate method, discussed in section 9.4, for determining the coefficient
and free term. Noting that the ratio between the depth (indicated by h) and the rise (represented by f) of an
arch is generally less than 0.5, i.e., h < 1 , so the effect on the coefficient and free term due to axial
f 2
deformation is negligible (when h > 1 , axial deformation must be considered). So in general, only
f 2
bending deformation is considered when calculating the free term Δ1P , and bending deformation and
sometimes axial deformation are taken into account when determining δ11 . And the shear deformation is
ignored. Thus, we rewrite the following expressions:
M 12 N2 ⎫
δ11 = ∫ ds + ∫ 1 ds ⎪
EI EA ⎪
⎬ (b)
ds = 0 ⎪
M 1M P
Δ1P = ∑ ∫
EI ⎪⎭
Assume that the origin of the rectangular coordinate system used here is oriented at point A, the
abscissa and ordinate of an arbitrary cross section C of the arch are, respectively, denoted by x and y as
shown in the figure. ϕ is the angle formed by the tangent to the centre line of the arch and the axis of
abscissa, which is positive in the left portion and negative in the right portion of the arch. The bending
moment M has positive sign when it makes the inner fibers of the arch in tension, and axial force N has
positive sign when it is a tensile force.
The expressions for the bending moment M 1 and axial force N1 due to X 1 = 1 [Fig.10.34 (c)]
on an arbitrary cross section C [Fig.10.34 (d)] of the primary structure (a simply supported curved beam)
can be written, by equilibrium conditions, as follows
M1 = − y ⎫⎪
⎬ (c)
N1 = − cos ϕ ⎪⎭
Under the action of vertical loading, the bending moment M P of the simply supported curved beam
0
(the primary structure) is identical to that, denoted by M , of a corresponding simple beam covering the
same span and carrying the same loading. That is,
MP = M 0 (d)
By substituting the expressions (c) and (d) into expression (b), we obtain
y2 cos 2 ϕ ⎫
δ11 = ∫ ds + ∫ ds ⎪
EI EA ⎪
0 ⎬ (10-13)
Δ1P = − ∫
M y
ds ⎪
EI ⎪⎭
④ Solve the force-method equation (compatibility condition)

By substituting the expressions for the coefficient and free term into the force-method equation, the
horizontal thrust H or the redundant will be obtained. That is,
M0y
Δ1P ∫ EI ds
X1 = H = − = (10-14)
δ11 y2 cos 2 ϕ
∫ EI ∫ EA ds
ds +
⑤ Calculation of internal forces

After the thrust H has been obtained the internal forces of a two-hinged arch can be determined by the
formulation used for determining the internal forces of a three hinged arch. Under the action of vertical
loading, the equations for calculating internal forces are rewritten as follows
M = M 0 − Hy ⎫
⎪
Q = Q cos ϕ − H sin ϕ ⎬
0
(10-15)
N = −Q 0 sin ϕ − H cos ϕ ⎭⎪
If an internal force diagram is desired, it may be obtained by first dividing the curved axial centre line
of the arch into several segments, then calculating the internal forces of each ends of the segments, finally
taking the values of the internal forces as ordinates of the curved axial line and connecting them into a
smooth curve.
⑥ Summary of analyzing formulation and the characteristics of internal forces of two-hinged arches
a. From the viewpoint of force method, the analyzing procedure for a two-hinged arch is the same as
that for a two-hinged rigid frame except that the method used to calculate flexibility coefficient δ 11 and free
terms Δ1P in the compatibility equation is different. In other words, for a two-hinged rigid frame,
graph-multiplication method can be employed to determine δ11 and Δ1P , while for a two-hinged arch,
integrate method must be used to determine Δ1P and δ11 ; whereas in computing δ11 the axial
deformation has to be taken into account sometime.
b. From the viewpoint of characteristics of internal forces, the internal forces of a two-hinged arch are
almost same as that of a three-hinged arch except that the horizontal thrust H is different. For a three hinged
arch, the thrust H is determined by an equilibrium condition, whereas for a two hinged arch, the thrust H
must be determined by a compatibility condition.
(2) Two-hinged arches with ties
Two hinged arches with ties are often applied to roof structures of buildings as shown in Fig.10.35 (a).
The function of the tie, which must posses enough stiffness, of a two hinged arch is to resist the horizontal
reaction induced by the arch so as to relieve the thrust acting on the wall or column(s) of a building, and to
reduce the bending moment in the arch.

① Primary systems of two-hinged arches
As shown in Fig.10.35 (a), a two-hinged arch with a tie is a single degree indeterminate structure too,
so if the axial force of the tie is selected to be the redundant X 1 , the primary system will be obtained by
cutting the tie of the given arch as shown in Fig.10.35 (b), which now is a curved beam.
(a) (b)
EA
EI
E1 A1
X1
Fig.10.35 a two-hinged arch with a tie

(a) a two-hinged arch with a tie; (b) primary system

Since the horizontal relative displacement at the cut of the actual tie is zero, the algebraic sum of the
relative displacements at the left and right sections adjacent to the cut of the primary structure (a curved
beam) due to the redundant X 1 and external loading must also be zero. Thus the compatibility equation can
be written as
δ11 X1 + Δ1P = 0 (e)
③ Determination of flexibility coefficient and free term

As shown in Fig.10.35 (b), the primary curved beam is a curved member, so we cannot use
graph-multiplication method to calculate the flexibility coefficient δ11 and free terms Δ1P in the
compatibility equation. Therefore, we will use integrate method for determining the coefficient and free
term. Note that only bending deformation is considered when calculating the free term Δ1P , bending
deformation and sometimes axial deformation are taken into account when determining δ11 . Thus, we
write the following expressions:
M 12 N2 N2
δ11 = ∫ ds + ∫ 1 ds + ∫ 1 dx (f)
EI EA E1 A1
In which, the first two items are the integer with respect to the arch, while the last item is the integer with
the tie. E1 A1 is the axial rigidity of the tie.
Since the axial force in the tie due to X 1 = 1 is equal to a unit magnitude, for a tie with a length of
l , the third integer of equation (f) will be equal to
l N 12 l 1
2
l
∫0 E1 A1
dx = ∫
0 E A
1 1
dx =
E1 A1
(g)
Because expressions of the bending moment and axial force for a two hinged arch with a tie is
identical to that for a two hinged arch with no tie, the expression for determining the coefficient δ11 can
be written, by substituting equations (c) and (g) into equation (f), as
y2 cos 2 ϕ l
δ11 = ∫ ds + ∫ ds + (h)
EI EA E1 A1
As long as there is no axial force in the tie of the primary curved beam due to external loading, the
expression for the free term Δ1P can be written as
M 1M P M0y
Δ1P = ∫ ds = − ∫ ds (i)
EI EI
④ Solve the force-method equation (compatibility condition)
By substituting the expressions for the coefficient and free term into the force-method equation, the
redundant or the tensile force in the tie will be expressed as
M0y
Δ1P ∫ EI ds
X1 = − = (10-16)
δ11 y2 cos 2 ϕ l
∫ EI ∫ EA ds + E1 A1
ds +
⑤ Calculation of internal forces

The expressions [Eq. (10-15)] used to determine the internal forces of a two hinged arch with a tie are
the same as that used to determine the internal forces of a two hinged arch with no tie if changing thrust H
into the tensile force X 1 . That is
M = M 0 − X1 y ⎫
⎪
Q = Q 0 cos ϕ − X 1 sin ϕ ⎬ (10-17)
⎪
N = −Q 0 sin ϕ − X 1 cos ϕ ⎭
It is observed, from comparing the expression used to determine the tensile force of the tie, X 1 [Eq.
l
(10-16)], with that used to determine the thrust H [Eq. (10-14)], that the item is added in the
E1 A1
denominator of Eq. (10-16). Eq. (10-16) will approach to Eq. (10-14) provided that the axial rigidity of the
tie goes to infinitely great, i.e., E1 A1 → ∞ . In the circumstance, the internal forces in the two kinds of
arches mentioned above are almost the same. In the case that the axial rigidity of the tie vanishes, i.e.,
E1 A1 → 0 , the tensile force in the tie approaches to infinitely small, i.e. , X 1 → 0 . In this situation, the two
hinged arch is actually a simple curved beam.
Example 10-9
Determine the horizontal thrust of the two hinged arch due to the action of two concentrated loads as
shown in Fig.10.36 (a), which possesses a constant cross section and a conic parabola centre line that may
be expressed as
4f
y= x (l − x )
l2
Solution
(a) (b)
y P P
A D C E B
D C E
x B Pl Pl
A
l l l l 4 4
4 4 4 4
(c) 0.028 (d)

0.011 C
D D E
C 0.0156
A
0.042 B A B
0.0034 (× Pl ) 0.0156 0.0625
(× Pl )

(a) a two-hinged arch; (b) M 0 diagram of a corresponding simple beam;
(c) M diagram of the two hinged arch; (d) M diagram of a corresponding three-hinged arch
(1) Assumptions for simplification

The example adopts two assumptions: ① Only bending deformation is taken into account whereas
the axial deformation is ignored. ② The curvature of the arch can be neglected for flat arch (in the case of
f 1 ). That is, the differential length of a curved line ds is equal to that of a straight line dx i.e., ds = dx .
<
l 5
The expressions for determining the coefficient and free term in the compatibility equation can be thusly
written as follows
1 l
δ 11 =
EI ∫
0
y 2 dx
1 l
Δ 1P = − ∫M
0
ydx
EI 0
(2) Primary system

The primary system of the two-hinged arch is shown in Fig.10.34 (b), which actually is a curved
beam.
(3) Calculation of flexibility coefficient and free term
The bending moment expression, by employing the symmetric property of the arch, for the
corresponding curved beam can be expressed, separately in segment, as
l
Segment AD: M 0 = Px, 0 < x ≤
4
Pl l l
Segment DC: M0 = , <x≤
4 4 2
After above preparation that is needed to determine δ11 and Δ1P , we finally obtain
1 l 1 l 4f 16 f 2 l 2 2 8 f 2l
δ 11 = ∫0 y dx = EI ∫0 [ l 2 x(l − x)] dx = EIl 4 ∫0 (l x − 2lx + x ) dx = 15EI
2 2 3 4 2
EI
1 l 0 2 ⎡ l4 l 2 Pl ⎤
Δ 1P = −
EI ∫ 0
M ydx = − ⎢
EI ⎣ ∫0
Pxydx + ∫
l 4 4
ydx ⎥
⎦
2 ⎡ l44f l 24f Pl ⎤
EI ⎢⎣ ∫0 l 2
=− x(l − x) Pxdx + ∫ x(l − x) dx ⎥
l 4 l2 4 ⎦
2
19 Pfl
=−
128 EI
(4) Solve the force-method equation (compatibility condition)
obtain
Δ1 P Pl
H =− = 0.278
δ11 f
(5) Calculate internal forces
Taking the bending moment as an example, by superposing equation
M = M 0 − Hy
We draw the bending moment diagram as shown in Fig.10.36 (c).
(6) Discussion
It can be observed from the computing result that the thrust ( 0.278Pl ) of a two hinged arch approaches
f
to that ( 0.25Pl ) of a corresponding three hinged arch, only 10% of relative difference, in the condition that
f
the axial deformation and curvature of a two hinged arch are ignored. Since the thrust of a two hinged arch
is greater than that of a three hinged arch, the bending moment of the two hinged arch is smaller than that
of the three hinged arch [Fig.10.36 (d)].
10.7.2 Symmetric hingeless arches
Symmetric hingeless arches are the structures which are commonly adopted in bridge structures. The
computing model of this sort of structure is shown in Fig.10.37 (a), indeterminate to third degrees. If we
select a symmetric primary system as shown in Fig.10.37 (b) the force-method equations may be expressed
as
X3
X1 X1
X2 X2 x
D
X3
y
(a) (b)
Fig.10.37 symmetric hingeless arch

(a) original structure; (b) primary system
δ11 X 1 + δ12 X 2 + δ13 X 3 + Δ1P = 0 ⎫

⎪
δ 21 X 1 + δ 22 X 2 + δ 23 X 3 + Δ 2 P = 0 ⎬ (a)
δ 31 X 1 + δ 32 X 2 + δ 33 X 3 + Δ 3 P = 0 ⎪⎭
The primary unknown X 1 is a pair of bending moments, X 2 a pair of axial forces and X 3 a pair of
shear forces, so the following secondary coefficients will be zero in terms of the unknowns’ symmetric
properties. That is,
δ13 = δ 31 = 0 ⎫
⎬ (b)
δ 23 = δ 32 = 0⎭
Substituting equation (b) into equation (a), we obtain the following two sets of equations
δ11 X 1 + δ12 X 2 + Δ1P = 0 ⎫
⎪
δ 21 X 1 + δ 22 X 2 + Δ 2 P = 0 ⎬ (10-18)
δ 33 X 3 + Δ 3 P = 0 ⎭⎪
In order to evaluate the coefficients in equation (10-18), the expressions of internal forces due to
X 1 = 1 , X 2 = 1 and X 3 = 1 must be first determined.
Assume that the origin of the rectangular coordinate system used here is oriented at the top of the arch,
the abscissa and ordinate of an arbitrary cross section of the arch are, respectively, denoted by x and y as
shown in Fig.10.37 (b). ϕ is the angle formed by the tangent to the centre line of the arch and the axis of
abscissa, which is positive in the right portion and negative in the left portion of the arch. The sign
convention for internal forces is identical to that used previously. The expressions of internal forces due to
X 1 = 1 , X 2 = 1 and X 3 = 1 are thusly expressed as follows [see Figs.10.38 (a) through (c)].
M 1 = 1⎫ M2 = y ⎫ M3 = x ⎫
⎪ ⎪ ⎪
N1 = 0 ⎬ N 2 = − cos ϕ ⎬ N 3 = sin ϕ ⎬
Q1 = 0 ⎭⎪ Q2 = sin ϕ ⎪⎭ Q3 = cos ϕ ⎭
⎪
When evaluate coefficients δii , δ ij and Δ iP , only bending deformation is taken into account
whereas the axial deformation is considered in the case of f < 1 (noting that f indicating the rise and l the
l 5
span of the arch). The expressions for determining the coefficients and free terms in the compatibility
equations can be thusly written as follows
M 12 1 ⎫
δ11 = ∫ ds = ∫ ds ⎪
EI EI ⎪
M 1M 2 y ⎪
δ12 = ∫ ds = ∫ ds ⎪
EI EI
⎪
M 22 N 22 y2 cos 2 ϕ ⎪
δ 22 = ∫ ds + ∫ ds = ∫ ds + ∫ ds ⎪
EI EA EI EA
⎪
M 32 x2 ⎪
δ 33 = ∫ ds = ∫ ds ⎬ (10-19)
EI EI ⎪
MM M ⎪
Δ1P = ∫ 1 P ds = ∫ P ds ⎪
EI EI ⎪
M 2M P MPy ⎪
Δ2P = ∫ ds = ∫ ds ⎪
EI EI
⎪
Δ3P = ∫
M 3M P
ds = ∫
MPx
ds ⎪
⎪
EI EI ⎭
(a) (b) (c)

X2 =1 X3 =1
O M1 O
X1 =1 O M2 M3
N1 N3
N2
Q1 Q3
Q2
Fig.10.38 internal forces due to unit loading

(a) under action of X 1 = 1; (b) under action of X 2 = 1; (c) under action of X 3 = 1
Example 10-10
Determine the horizontal thrust and bending moments on the sections of the top and skewback of the
hingeless arc arch due to the action of uniformly distributed load as shown in Fig.10.39 (a).
Solution
20kN/m
20kN/m
X1 X1
X2 X2 x
(b)
y
(a) C D
4m
E
A B A ϕ B
ϕ0 R ϕ0 R
EI = constant
O′ O′
8m 8m x
y
(a) original structure and loading; (b) primary system
(1) Primary system

The symmetric primary system is selected as shown in Fig.10.39 (b), noting that the antisymmetric
unknown (shear force) X 3 = 0 due to the symmetric loading.
(2) Relation between the radius R and central angle ϕ0 and coordinates
From the rectangular triangle O ′AE [see Fig.10.39 (a)], we obtain
l
R 2 = ( )2 + ( R − f )2
2
l2 + 4 f 2
R= = 10m
8f
l
AE 2 O′E R − f
sin ϕ0 = = = 0.8 ， cos ϕ0 = = = 0.6
O′A R O′A R
Thusly
ϕ0 = 0.9273 radian
The relation between the radius R and angle ϕ and coordinates will be written as
x = R sin ϕ ， y = R − R cos ϕ
(3) Force-method equations
Based on the primary system, we write the force-method equations as
δ11 X1 + δ12 X 2 + Δ1P = 0
δ 21 X1 + δ 22 X 2 + Δ 2 P = 0
(4) Calculation of flexibility coefficients
Due to f = 1 > 1 , only the bending deformation is concerned when calculating the flexibility
l 4 5
coefficients. So by equation (10-19), we write
ϕ0
EI δ11 = ∫ M 12 ds = 2∫ Rdϕ = 2 Rϕ0
0
ϕ0
EI δ12 = ∫ M 1M 2 ds = ∫ yds = 2 ∫ ( R − R cos ϕ ) Rdϕ
0
ϕ0
= 2 ∫ ( R − R cos ϕ ) dϕ = 2 R 2 [ϕ − sin ϕ ]0 0
2 2 ϕ
0
= 2 R 2 (ϕ0 − sin ϕ0 )
ϕ0 ϕ0
EI δ 22 = ∫ M 22 ds = 2 ∫ y 2 ds = 2 ∫ ( R − R cos ϕ ) 2 Rdϕ = 2 R 3 ∫ (1 − 2 cos ϕ + cos 2 ϕ )dϕ
0 0
ϕ0
⎡ ϕ 1 ⎤ 3 1
= 2 R 3 ⎢ϕ − 2 sin ϕ + ( + sin 2ϕ ) ⎥ = 2 R 2 ( ϕ0 − 2 sin ϕ0 + sin 2ϕ0 )
⎣ 2 4 ⎦0 2 4
Substituting ϕ0 = 0.9273 radian into above expressions, the coefficients can be evaluated as
EIδ11 = 1.855R ， EI δ12 = 0.2546 R 2 ， EI δ 22 = 0.0619 R 3
(5) Calculation of free terms
The bending moment expression of the primary structure due to the external loading can be expressed,
by equilibrium equation, as
q 2 q
MP = − x = − R 2 sin 2 ϕ
2 2
By using equation (10-19), we find
ϕ0 q
EI Δ1P = ∫ M 1M P ds = 2∫ 1× (− R 2 sin 2 ϕ ) Rdϕ
0 2
ϕ0
⎡ϕ 1 ⎤
= −qR 3 ⎢ − sin 2ϕ ⎥
⎣2 4 ⎦0
ϕ 1
= −qR 3 ( 0 − sin 2ϕ0 )
2 4
ϕ0 q
EI Δ 2 P = ∫ M 2 M P ds = ∫ M P yds = 2∫ ( R − R cos ϕ )(− R 2 sin 2 ϕ ) Rdϕ
0 2
ϕ0
= −qR ∫ (sin ϕ − sin ϕ cos ϕ )dϕ
4 2 2
0
ϕ
⎡ϕ 1 ⎤
0
1
= −qR 4 ⎢ − sin 2ϕ − sin 3 ϕ ⎥
⎣2 4 3 ⎦0
ϕ 1 1
= −qR 4 ( 0 − sin 2ϕ0 − sin 3 ϕ0 )
2 4 3
Substituting ϕ0 = 0.9273 radian into above expressions, the free terms can be written as
EI Δ1P = −0.2237 qR 3 , EI Δ 2 P = −0.0530 qR 4

(6) Solve the force-method equations
write
1.855 X 1 + 0.2546 RX 2 − 0.2237qR 2 = 0 ⎫⎪
⎬
0.2546 X 1 + 0.0619 X 2 − 0.0530qR 2 = 0 ⎪⎭
Finally, we obtain
X 1 = 0.0071qR 2 = 14.2kN ⋅ m
X 2 = 0.827qR = 165.4kN
(7) Calculation of bending moments on the sections of the top and skewback of the arch
By equilibrium conditions, the bending moments on the sections of the top and skewback of the arch
will be evaluated as
M C = X1 = 14.2kN ⋅ m
q l 20 16
M A = M B = X 1 + X 2 f − ( ) 2 = 14.2 + 165.4 × 4 − ( )2 = 35.8kN ⋅ m
2 2 2 2
If other internal forces are desired, they can be determined by equilibrium conditions as well.
(8) Discussion
Comparing the thrust of a three hinged arch with the same span and carrying the same loading, i.e.,
M 0 ql 2 20 × (16) 2
H′ = C = = = 160kN , with that of the hingeless arch, i.e., H = X 1 = 165.4kN , it can be
f 8f 8× 4
observed that the two thrusts only have a relative difference of 3.4%, and they are very closed.
10.8 Internal Forces due to Temperature Changes and Support Settlements
Thus far, we have considered the analysis of structures with unyielding supports. However, support
movements due to weak foundations and the like may induce significant stresses in externally
indeterminate structures and must be considered in their designs. Unlike support settlements, which affect
only externally indeterminate structures, temperature changes may affect the stress conditions of externally
and/or internally indeterminate structures. The force method, as developed in the preceding sections, can be
easily modified to include the effects of support settlements and temperature changes in the analysis. And
the procedure for the analysis of structures subjected to support settlements and temperature changes is the
same as used previously for the case of external loads. The only difference is that the primary structure is
now subjected to the prescribed support settlements and temperature changes (instead of external loads) to
evaluate its displacements at the locations of redundants due to these effects. The redundants are then
determined by applying the usual compatibility conditions that the displacements of the primary structure at
the locations of the redundants due to the combined effects of support settlements and temperature changes
and the redundants must equal the known displacements at the corresponding locations on the actual
indeterminate structure.
10.8.1 Temperature changes
As mentioned above, the procedure for the analysis of structures subjected to temperature changes is the
same as used previously for the case of external loads. The procedure is illustrated by the following
example.
Example 10-11
Determine the internal forces of the rigid frame shown in Fig.10.40 (a) due to a temperature change as
indicated in the figure.
Solution
10.8 Internal Forces due to Temperature Changes and Support Settlements 341
(1) Primary system

The horizontal reaction at support B is selected to be the redundant. The primary rigid frame obtained
by replacing the hinged support by a roller support. Next, the primary rigid frame is subjected to the
prescribed temperature changes and an unknown concentrated force so as to form a primary system as
shown in Fig.10.40 (b).
(a) C −500 C B (b) −500 C

C B X1
00 C 00 C
−500 C 00 C
6m
−500 C 00 C
EI = 144000kN ⋅ m 2
α = 0.00001
A A
8m
(c)
(d)
6 B X1 = 1 1 1 X1 = 1
6 C
C 3 B
4
M 1 diagram N1 diagram
(unit: m) (unit: kN)
A A 3
4

(a) original structure and temperature changes; (b) primary system;
(c) bending mement diagram due to X 1 = 1; (d) axial force diagram due to X 1 = 1

The compatibility equation can be expressed as
δ11 X1 + Δ1t = 0 (a)
in which, δ 11 denotes the horizontal displacement at support B due to a unit value of the redundant X 1
and Δ1t denotes the horizontal displacement at support B of the primary rigid frame due to the
temperature changes.
(3) Calculation of the flexibility coefficient and free term
The approach for determining coefficient δ11 is the same as used previously for the case of external
loads, so bending moment and axial force diagrams due to X 1 = 1 are drawn in Fig.10.40 (c) and (d). By
graph-multiplication method, δ11 is determined to be
M 12 1 ⎡ 1 2 1 2 ⎤ 168
δ11 = ∑ ∫ ds = ( × 6 × 6) × × 6 + ( × 8 × 6) × ×6 =
EI EI ⎢⎣ 2 3 2 3 ⎥⎦ EI
The approach for determining free term Δ1t is different from that used previously for the case of
external loads, but we can use Eq. (9-17b) in section 9.6 to determine it. Thusly, we rewrite
α Δt
Δ1t = ∑ ∫ M ds + ∑ α t ∫ N ds
1 0 1
(b)
h
In which, Δt = t2 − t1 represents the difference of the temperature changes between the inner and outer
t1 + t2
fibers of the members; t0 = denotes the average changes of the temperature at the central line of
2
the members. They are evaluated to be
Δt = t2 − t1 = 00 C − ( −500 C ) = 50 0 C
1 1
t0 = (t1 + t2 ) = (00 C − 500 C ) = −250 C
2 2
Finally, we find
αΔt
Δ1t = ∑ M 1ds + ∑ α t0 ∫ N1ds
h ∫
50 1 1 3
= −α ( × 6 × 6 + × 8 × 6) + α ⋅ 25 × (1× 8 + × 6)
0.6 2 2 4
= −3187.5α
Note that the two terms in equation (b) represent virtual work done by the virtual force system, so the sign
of each term will be determined by the sign of virtual work. Consequently, the first term in Δ1t must be
negative because the virtual bending moment perform negative virtual work; whereas the virtual axial force
performs positive work, so the second term is positive.
find
168
X 1 − 3187.5α = 0
EI
X1 = 18.97 EIα (c)
(5) Construction of internal force diagrams

Since the primary structure is a statically determinate structure, the temperature changes will not
induce internal forces in the primary structure. All of the internal forces of the indeterminate rigid frame are
thusly caused by its redundant. The bending moment and axial force equations can be written as
M = M 1 X 1 ⎫⎪
⎬ (d)
N = N1 X 1 ⎪⎭
The bending moment and axial force diagrams are depicted as shown in Fig.10.41 (a) and (b).
The characteristics of analyzing the internal forces of a statically indeterminate structure due to
temperature changes may be summarized, by considering the analytical procedure of above example, as
follows
(1) The force-method equation(s) [see Eq. (a)] is symbolically the same as that formulated
previously for the case of external loads, but the meaning of free term(s) is different, which now is the
displacement(s), evaluated by Eq. (b), at the location(s) of redundant(s) due to the effect of temperature
changes (instead of external loads).
(2) All of the internal forces of an indeterminate structure are caused by its redundant(s) [see Eq.
(d)].
(3) It is observed from the computing result [see Eq. (c)] and internal force diagrams [see Fig.10.41]
that the magnitude of the internal forces, caused by temperature changes, is not only in self-equilibrium but
also in direct proportion to the absolute value of the flexural rigidity EI of the member. Therefore, selecting
a large cross section for the member will not modify its stress condition caused by its temperature changes.
(4) It can be observed from Fig.10.41 (a) that the tensile stress due to temperature changes occurs
on the side of the member where the temperature rise is lower. It is instructive to notice that the cracks may
be occurred in the concrete member due to decrease of its temperature.
163.93
163.93 B C 20.49 B
C 27.32
27.32
M diagram N diagram
(unit: kN ⋅ m) (unit: kN)
A 20.49
A
(a) (b)
Fig.10.41 internal force diagrams of example 10-11
(a) bending moment diagram; (b) axial force diagram
10.8.2 Support settlements
Support settlements do not cause any stresses in determinate structures; they may, however, induce
significant stresses in externally indeterminate structures, which should be taken into account when
designing externally indeterminate structures. Consider the determinate and indeterminate beams shown in
Fig.10.42. It can be seen from Fig. 10.42 (a) that when the support B of the determinate beam undergoes a
small settlement Δ B , the beam moves as rigid body without bending—that is, they remain straight. Thus,
no stresses develop in the determinate beam. However, when the indeterminate beam of Fig.10.42 (b) is
subjected to a similar support settlement, it bends, as shown in the figure; then, bending moments develop
in the indeterminate beam. Therefore, support movements due to weak foundations and the like may induce
significant stresses in externally indeterminate structures and must be considered in their designs.
Support settlements, however, do not have any effect on the stress conditions of structures that are
internally indeterminate but externally determinate. This lack of effect is due to the fact that the settlements
cause such structures to displace and/or rotate as rigid bodies without changing their shapes. The force
method, as developed in the preceding sections, can be easily modified to include the effect of support
settlements in the analysis. The analyzing procedure is illustrated by the following example.
(a) A B (b) A B
ΔB
ΔB
B1 B1
Fig.10.42 effect due to support settlements

(a) a determinate beam; (b) an indeterminate beam
Example 10-12
Determine the internal forces and draw their diagrams for the indeterminate beam shown in Fig.10.43
(a) due to a rotation at support A as shown in the figure. EI = constant .
Solution
(a) (b) (c)

θ θ
EI B EI B M 1 diagram
A A B
θ θ A
l X1 1 X1 = 1
(d) (e) (f)

θ M diagram Q diagram
A B A B
EI B
A
θ
3EIθ 3EIθ
lθ
l l2

(a) rotation at support A; (b) first primary system; (c) primary beam under action of X 1 = 1;
(d) primary structure subjected to a support settlement; (e) bending moment diagram;
(f) shear force diagram
(1) First primary system

① Primary system 1
The vertical reaction at support B is selected to be the redundant. The primary beam obtained by
removing the support forms a determinate beam. Next, the primary beam is subjected to the prescribed
support rotation θ and an unknown concentrated force so as to form a primary system as shown in
Fig.10.43 (b).
By realizing that the deflection of the actual indeterminate beam at support B is equal to zero due to
the combining actions of the redundant and support settlement, we write the compatibility equation as
δ11 X1 + Δ1c = 0 (a)
in which, δ11 denotes the vertical displacement at support B due to a unit value of the redundant X 1 and
Δ1c denotes the vertical displacement at support B of the primary beam due to the support rotation θ .
③ Calculation of the flexibility coefficient and free term
The approach for determining coefficient δ11 is the same as used previously for the case of external
loads, so bending moment diagram due to X 1 = 1 is drawn in Fig.10.43 (c). By graph-multiplication
method, δ11 is determined to be
M12 1 1 2 l3
δ11 = ∫ dx = ( l × l) × l =
EI EI 2 3 3EI
The approach for determining free term Δ1c is different from that used previously for the case of
external loads, but we can use Eq. (9-18) in section 9.7 to determine it. Thusly, we rewrite
Δ1c = −∑ Rc (b)
Here, R = l , c = θ , thusly, equation (b) will be changed to be

Δ1c = −lθ
The free term Δ may be determined by the displacement shape formed by the given support rotation as
1c
shown in Fig.10.43 (d). The direction of Δ1c is opposite to X 1 , so it has a minus sign.
④ Solve the force-method equation
find
Δ1c lθ 3EIθ
X1 = − = = 2 (c)
δ11 3
l / 3EI l
⑤ Construction of bending moment diagram
Since the primary structure is a statically determinate structure, the support movement will not induce
internal forces in the primary structure. All of the internal forces of the indeterminate structure are thusly
caused by its redundant. The bending moment equation can be written as
M = M1 X1 (d)
The bending moment diagram is constructed as shown in Fig.10.43 (e).
⑥ Shear force diagram
With X 1 determined, the shear force of the beam can be found by equilibrium conditions. The shear
force diagram is shown in Fig.10.43 (f).
(2) Second primary system
① Primary system 2
The moment at support A is selected to be the redundant. The primary beam obtained by changing the
fixed support of the indeterminate beam into a hinged support. Next, the primary beam is subjected to the
prescribed support rotation θ and an unknown concentrated moment at support A so as to form a primary
system as shown in Fig.10.44 (a).
X1 X1 = 1
A B A M 1 diagram B
(a) (b)
θ
1
Fig.10.44 other figures of example 10-12
(a) second primary system; (b) primary beam under action of X 1 = 1

By realizing that the slope of the actual indeterminate beam at support A is equal to the rotation θ
due to the redundant, we write the force-method equation as
δ11 X 1 = θ (e)
③ Calculation of the flexibility coefficient

The approach for determining coefficientδ11 is the same as used above, so bending moment diagram
due to X 1 = 1 is drawn in Fig.10.44 (b). By graph-multiplication method, δ11 is calculated to be
M12 1 1 2 l
δ11 = ∫ dx = ( l ×1) × =
EI EI 2 3 3EI
④ Solve the force-method equation
By substituting the expression for the coefficient into the force-method equation, we obtain
l 3EIθ
X1 = θ , X1 = (f)
3EI l
⑤ Construction of internal force diagrams

Although the primary system 2 is different from primary system 1 their final bending moment and
shear force diagrams of the indeterminate beam are identical as those shown in Fig.10.43 (e) and (f).
The characteristics of analyzing the internal forces of a statically indeterminate structure due to
support settlements may be summarized, by comparing the analytical procedure due to support movements
with that due to external loads, as follows
(1) The form of force-method equation(s) [see Eqs. (a) and (e)] may be different for different
primary system. That is, the right hand side of the equality may not be equal to zero.
(2) There are two approaches to determine the free term in Eq. (a). It can be determined by Eq. (b)
or by the geometric relation (if convenient) of the displacement configuration of the primary structure,
which is actually a determinate structure.
(3) There are no external loads exerting on the determinate primary structure, so all of the internal
forces are caused by the redundant(s) [see Eq. (d)].
(4) It is observed from computing result [see Eqs. (c) and (f)] and internal force diagrams [see
Fig.10.43 (e) and (f)] that the magnitude of the internal forces, caused by support settlements, is also in
self-equilibrium and in direct proportion to the absolute value of the flexural rigidity EI of the member as
well.
10.9 Computation of Displacements of Statically Indeterminate Structures
The analytical elements of displacements of statically indeterminate structures are the same as that of
statically determinate structures. In other words, the unit load method, which is formulated in terms of
principle of virtual work discussed in Chapter 9, can be applied to the calculation of displacements not only
for statically determinate structures but also for statically indeterminate structures. Recalling for unit load
method, the equilibrium condition is the only condition that has to be satisfied for the virtual force system,
so the unit force may be applied on any one of the possible primary structures of a desired indeterminate
structure. Consequently, as soon as the internal forces of an indeterminate structure are obtained by the
force method, any one of its displacements can be conveniently determined by using the unit load method
to any one of the possible primary structures of the analyzed indeterminate structure. The internal force
components due to the unit force, which are needed when applying unit load method, may be thusly
determined by equilibrium equations alone.
In fact, the basic consideration of force method used to analyze a statically indeterminate structure is
selecting a proper statically determinate structure, named primary structure, obtained by removing enough
redundant restraints from the given structure, and then determining internal forces of the original
indeterminate structure by the primary system gained by applying external loadings and the unknown
redundants to the primary structure. In addition the consideration, the consistence of deformation is also
taken into account. That is, in order to make the deforming conditions of the primary system to meet the
deforming conditions of the original structure, the conditions of consistence of deformation or
compatibility conditions have been satisfied in force method. These two aspects of consideration ensure
that the stressing and deforming conditions of the determinate primary system are identical to those of its
original indeterminate one. However, the two states are represented in different forms. For the original
indeterminate structure, the redundants are passive forces that are existed in corresponding restraints, but
for the primary determinate system, the redundants are active forces that are applied to the primary
structures.
(a) q (b) q
A B A B
Δ C Δ C
3ql
YB =
l l 8
3ql
2 2 YB =
8
(c) (d)
ql 2 M diagram l M diagram
P =1
8 2
B A B
A C
Fig.10.45 displacement calculation of indeterminate structure

(a) an indeterminate beam; (b) a primary system;
(c) bending moment diagram; (d) primary structure and its unit load
Consider, for example, the beam shown in Fig.10.45. Reaction YB is a passive force for the
indeterminate beam shown in Fig.10.45 (a); while force YB is an active force, now denoted by X 1 , for
the determinate primary system shown in Fig.10.45 (b). However, the stressing and deforming conditions
of the two beams shown in the figures (a) and (b) are identical. As the bending moment diagram shown in
Fig.10.45 (c) is not only the bending moment diagram of the indeterminate beam shown in Fig.10.45 (a)
but also the bending moment diagram of the determinate beam shown in Fig.10.45 (b). Similarly, the
deflection at the middle span of the beam shown in Fig.10.45 (a) has the same magnitude and direction of
the deflection at the middle span of the beam shown inFig.10.45 (b).
By above discussions, we conclude that the calculation of the displacements for a statically
indeterminate structure can transformed into the calculation of displacements for its statically determinate
primary system due to the combining actions of external loadings and the unknown redundants. For
instance, if the deflection of the middle span of the beam shown in Fig.10.45 (a) is desired, the unit load
may be applied to the middle point C of the determinate primary beam shown in Fig.10.45 (d), on which
the bending moment diagram M due to P = 1 is drawn. By graph-multiplication method, we obtain
MM
ΔC = ∫ ds
EI
1 ⎡ 1 l l 2 1 2 1 1 2 2 l 1 2 3 l⎤
= ( × × )( × ql + × ql ) − ( × × ql ) × × ⎥
EI ⎢⎣ 2 2 2 3 8 3 16 3 2 8 8 2⎦
ql 4
= (↓)
192 EI
By unit load method, the general expression for determining any one of the displacements of a
statically indeterminate structure can be expressed as
MM NN QQ αΔt
Δ = ∑∫ ds + ∑ ∫ ds + ∑ ∫ k ds + ∑ ∫ M ds + ∑ ∫ Nα t0 ds − ∑ Rc (10-20)
EI EA GA h
Equation (10-20) is symbolically the same as that used for determining any one of the displacements
of a statically determinate structure. However, it should be noted that M, Q and N are herein the bending
moment, shear force and axial force of a statically indeterminate structure due to external loadings;
while M , Q , N and R are the bending moment, shear force, axial force and reaction of one of its
statically determinate primary structures due to a unit load, named virtual system in Chapter 9.
The first three items are the displacements caused by external loads; the fourth and fifth items are the
displacements induced by temperature changes; whereas the last or sixth item is the displacement resulted
from the support settlements.
If only the external loads are applied, the expression for the calculation of the displacements of a
statically indeterminate structure will be
MM NN QQ
Δ = ∑∫ ds + ∑ ∫ ds + ∑ ∫ k ds (10-21)
EI EA GA
Example 10-13
Determine the deflection at the middle point C for the indeterminate beam shown in Fig.10.45 (a) by
unit load method under the condition that the primary beam shown in Fig.10.45 (b) will not be used.
EI = constant .
Solution
(1) Real system

The real system is the given indeterminate beam shown in Fig.10.45 (a) and its bending moment
diagram is shown in Fig.10.46 (a) again.

(2) Virtual system
The virtual system selected here is a simple beam with a unit load applied at the point C, which is one
of the primary beams, obtained by replacing the fixed support A by a hinged support and then applying a
unit load at the location and in the direction of the desired displacement, as shown in Fig.10.46 (b), on
which the bending moment diagram M is also depicted.
(a) 1 M diagram (b)

ql 2 P =1
8 C
C A B
B
A
l l l
2 2 4
(a) an indeterminate beam and its bending moment diagram;
(b) virtual system and bending moment due to P = 1
(3) Calculation of desired displacement Δ C

If only bending deformation is concerned, the expression for determining the displacement can be
expressed as
MM
ΔC = ∑ ∫ ds
EI
By graph-multiplication method, we obtain
1 ⎡ 1 l 1 2 2 l 1 2 5 l⎤ ql 4
ΔC = − ( l ×
⎢⎣ 2 4 16 ) × ql + 2 × ( × × ql ) × × = (↓)
EI 3 2 8 8 4 ⎥⎦ 192 EI
The positive answer for Δ C indicates that the direction of the deflection coincide with that of the
unit load P = 1 .
It is observed from the example that when determining the displacements of an indeterminate structure
by unit load method the unit load can be applied to any one of the primary structures of the given
indeterminate structure. Since a primary structure is a determinate structure the determination of the
displacement will be simplified.
Example 10-14
Determine the horizontal displacement at the top point D of column BD for the indeterminate bent
frame shown in Fig.10.47 (a) by unit load method.
Solution
(1) Real system

The real system is the given indeterminate bent frame shown in Fig.10.47 (a) and its bending moment
diagram is also shown in the figure.
(a) C D (b) C D P =1
2m
I I
3m
20kN I I
11.8 17.7
2.3 3
6I 6I 6I 6I
M diagram M diagram
(unit: kN ⋅ m) 6m (unit: m)
86.9 A 53.1 B A 9 B

(a) an indeterminate bent frame and its bending moment diagram;
(b) virtual system and bending moment due to P = 1
(2) Virtual system

The virtual system is shown in Fig.10.47 (b), which is one of the primary structures, obtained by
cutting the two-force member CD and then applying a unit load at the location and in the direction of the
desired displacement, as shown in Fig.10.47 (b), on which the bending moment diagram M is also
depicted.
(3) Calculation of desired displacement Δ D
If only bending deformation is considered, the displacement can be determined, by
graph-multiplication method, as
MM 1 1
ΔD = ∑ ∫ ds = ( × 3 × 17.7) × 2 +
EI EI 2
1 ⎡ 1 2 1 1 1 2 ⎤
⎢ ( × 6 × 17.7)( × 3 + × 9）+ ( × 6 × 53.1)
（ × 3 + × 9) ⎥
6 EI ⎣ 2 3 3 2 3 3 ⎦
53.1 1380.9 283.2
= + = (→)
EI 6 EI EI
The positive answer for Δ D indicates that the sense of the displacement is the same as that of the
unit load P = 1 .
By the way, if we apply the unit load P = 1 at the point C, the same result will be obtained. It may
be checked by readers.
10.10 Verification of Calculation of Statically Indeterminate Structures
In the process of analyzing a statically indeterminate structure, many steps and numeric operations
must be carried out, so mistakes are prone to occur. Therefore, in order to ensure the rightness of the final
analyzing results, it is instructive to do some checking work during the analyzing process.
The checking work involves the following three aspects.
(1) Verification of analyzing process
The analyzing process involves the following checking steps and each of the steps will not allow
making any mistakes.
① Whether or not the determination of the degrees of indeterminacy of an indeterminate structure
is correct? Whether or not the primary structure is a stable system?
② Whether or not the internal force diagrams or expressions due to the external loading and due to
a unit load are correct?
③ Whether or not the calculation of flexibility coefficients and free terms in compatibility
equations has no mistakes? Whether or not the areas of internal force diagrams, their corresponding
ordinates at the position of their centroids, measured from another bending moment diagram with straight
lines, and flexural rigidity EI are correctly determined?
④ Whether or not the solution of the force-method equations is correct? Whether or not the
answers of redundants satisfy the force-method equations?
⑤ The final internal force diagrams must be checked by the equilibrium and deformation
conditions of the structure.
(2) Verification of consistent deformations
When analyzing a statically indeterminate structure, in addition that the equilibrium conditions must
be used, the conditions of consistent deformations or compatibility conditions must be considered.
Especially in force method, the unknown redundants are determined by the compatibility equations, so the
keynote checking work is the verification of deformation consistence of the structure.
The common way for checking the deformation consistence of a statically indeterminate structure can
be stated as: ① select any one of the primary structures as a desired primary structure and any one of the
unknowns as a desired redundant, say X i ; ② calculate the displacement at the location and in the
direction of X i based on the final internal force diagrams or expressions determined by force method,
named Δi ; ③ check the displacement Δi whether or not equal to the corresponding prescribed
displacement of the original indeterminate structure. That is
Δi = corresponding prescribed displacement (10-22)
For more general case, i.e., the displacements are determined by Eq. (10-20), the equation (10-22) will be
written as
MM NN QQ αΔt
∑∫ EI
ds + ∑ ∫
EA
ds + ∑ ∫ k
GA
ds + ∑ ∫ M
h
ds + ∑ ∫ Nα t0 ds − ∑ Rc (10-23)
= corresponding prescribed displacement
In which, M, Q and N are the bending moment, shear and axial force of the original indeterminate structure;
while M , Q , N and R are the bending moment, shear force, axial force and reaction of the primary
structure due to X i = 1 .
If the corresponding prescribed displacement is equal to zero and only the external loads are applied
on the indeterminate structure, the equation (10-23) will be
MM NN QQ
∑∫ EI
ds + ∑ ∫
EA
ds + ∑ ∫ k
GA
ds = 0 (10-24)
If the indeterminate structure is a beam or a rigid frame with the external loadings only applied on it
and merely the bending deformation is concerned, the Eq. (10-24) will further become
MM
∑∫ EI
ds = 0 (10-25)
We can use Eq. (10-25) to check the rightness of the final bending moment diagram. For an
indeterminate structure, if the value of one of its displacements is given, we may employ the bending
moment diagram obtained by the force method to calculate the displacement again. The coincidence of the
solution for the displacement with the given value may verify the correctness of the bending moment
diagram.
P
C D
C D 1 1
1 1
X1 = 1
M diagram M diagram
A B A 1 1 B
(a) (b)
Fig.10.48 verification of displacement of a closed rigid frame

(a) a rigid frame and its bending moment diagram;
(b) virtual system and bending moment due to X 1 = 1
Consider the closed rigid frame shown in Fig.10.48 (a). By consistent deformation, we have already
known that the relative rotation of any one of the cross sections is equal to zero. However, we can use the
bending moment diagram obtained by force method as shown in Fig.10.48 (a) again to calculate the relative
rotation of the sections adjacent to the middle point of the beam CD. The virtual force system and its
bending moment diagram due to X 1 = 1 are shown in Fig.10.48 (b). By Eq. (10-25), we write
MM M
∑∫ EI
ds = v∫ ds = 0
EI
(10-26)
The solution for the displacement shows that the final bending moment diagram [Fig. 10.48 (a)]
obtained by force method can satisfy the known displacement condition.
(a)
4kN/m
14.4 14.4 12 (b) (c)

14.4 C 18 D 14.4 D
C 12 12 D
C
I =3 5.4 5.4
3.6 5.4 5.4
12
I =1
4m
M diagram I =1 Q diagram N diagram

(unit: kN ⋅ m) (unit: kN) (unit: kN)
A B A 5.4 B 5.4 12
7.2 7.2 A 12 B
6m
(f) 4kN/m
(e)
(d) C D
14.4kN ⋅ m 12kN
C C 5.4kN
5.4kN
14.4kN ⋅ m 12kN
7.2kN ⋅ m A B 7.2kN ⋅ m
5.4kN 5.4kN
12kN 12kN
Fig.10.49 verification of internal force diagrams of a closed rigid frame
(a) a rigid frame and its bending moment diagram; (b) shear force diagram;
(c) axial force diagram; (d) balance of moment of joint C ; (e) balance of forces of joint C ;
(f) balance of load and reactions of entire structure
M
It can be seen from the discussion that the sum of the area of diagram for members of a closed
EI
frame that is only subjected to external loads should be equal to zero.
(3) Verification of equilibrium conditions
The final internal force diagrams obtained by force method must satisfy equilibrium conditions. In
other words, any one of the free body diagrams such as the entire structure, a rigid joint, a member, or a
portion of the structure must meet their corresponding equilibrium conditions
Example 10-15
Verify the correctness for the diagram of the bending moment M as shown in Fig.10.49.
Solution
The given indeterminate rigid frame shown in Fig.10.49 (a) is only subjected to external load, so its
bending moment diagram must satisfy Eq. (10-26). That is,
MM M
∑∫ EI
ds = v∫ ds = 0
EI
However,
M 1 2 2 1 1
v∫ EI ds = 3 (−14.4 × 6 + 3 × 6 × 18) + 1 (− 2 × 4 × 14.4 + 2 × 4 × 7.2) = −86.4 + 52.8 = −33.6 ≠ 0
It implies that the bending moment diagram do not meet the conditions of deformation of the structure
though the free body diagrams shown in Figs.10.49 (d) through (f) satisfy their equilibrium equations, so
the computing result of the example is wrong.
(b )
(a) a
3 B
Pa
88 a
C I1
2
a
P
M diagram M diagram
2
a
2I1
A X1 = 1
15
Pa a
88
(a) a rigid frame and its bending moment diagram;
(b) virtual system and bending moment due to X 1 = 1
Example 10-16
Verify whether or not the bending moment diagram shown in Fig.10.50 (a) satisfying the conditions of
consistent deformation.
Solution
It is known by observation that the horizontal displacement at support A of the frame shown in
Fig.10.50 (a) is zero. The given displacement condition can be used to check the correctness of the bending
moment diagram, so a virtual force system with a unit load corresponding to the given displacement
applied on one of the primary structures of the given indeterminate frame, and the bending moment
diagram due to the unit load are all shown in Fig.10.50 (b). By Eq (10-25) and graph-multiplication method,
we find
MM 1 a2 2 3
Δ AH = ∑ ∫ ds = ( ) × × Pa +
EI EI 2 3 88
1 ⎡ 1 3Pa 2a 1 15Pa a 1 Pa a⎤
⎢ ( × × a) + ( × × a） − ( × × a) ⎥
2 EI ⎣ 2 88 3 2 88 3 2 4 2⎦
=0
It implies that the bending moment diagram meets the conditions of consistent deformation of the
indeterminate frame. The equilibrium conditions may be checked by readers.
SUMMARY
(1) Principle of force method

The force (flexibility) method or the method of consistent deformations is one of the fundamental
methods used to analyze statically indeterminate structures. In the chapter we have discussed the
formulation of the force method. The method of consistent deformations involves removing enough
restraints from an indeterminate structure to render it statically determinate. The determinate structure is
called the primary structure, and the reactions or internal forces associated with the excess restraints
removed from the indeterminate structure are termed primary unknowns or redundants. The primary
unknowns are now treated as unknown loads together with the external loadings applied to the primary
structure to form a primary system, and their magnitudes are determined by solving the compatibility
equations based on the condition that the displacements of the primary structure at the locations (and in the
directions) of the primary unknowns, due to the combined effect of the prescribed external loading and the
unknown redundants, must be equal to the known displacements at the corresponding locations on the
original indeterminate structure. Once the redundants have been determined, the analysis of the
indeterminate structure will be transformed to the analysis of its primary determinate structure, and the
other response characteristics of the indeterminate structure can be evaluated either through equilibrium
conditions or by superposition of the responses of the primary structure due to the external loadings and
due to each of the redundants.
Therefore, the key points of force method are that: ① determine primary unknowns; ② select
primary structure; ③ develop force-method equations (compatibility equations).
Summary 357
(2) Determine primary unknowns and select primary structures

The primary unknowns must be equal to the degrees of indeterminacy of the structure. Remove the
restraints corresponding to the redundants from the given indeterminate structure to obtain the primary
(determinate) structure. It is possible to choose several primary structures for one indeterminate structure,
but the best one will result in simplest analysis. Also, the redundants must be chosen so that the removal of
the corresponding restraints from the given indeterminate structure results in a primary structure that is
statically determinate and stable. Applying the unknowns and the external loadings (external loads, support
settlements or temperature changes etc.) to the primary structure will form a primary system.
The senses of the redundants are not known and can be arbitrarily assumed. A positive answer for a
redundant will imply that the sense initially assumed for the redundant was correct.
(3) Develop force-method (compatibility) equations
Write a compatibility equation for the location of each redundant by setting the algebraic sum of the
displacements of the primary structure due to each of the redundants and the external loading equal to the
known displacement at the corresponding location on the actual indeterminate structure. The total number
of compatibility equations thus obtained must be equal to the number of redundants.
(4) Calculation of flexibility coefficients and free terms in force-method equations
a. First, draw internal force diagrams, especially bending moment diagram, of the primary structure
with only the external loading applied to it when the graph-multiplication is desired, or develop the
expressions of internal forces of the primary structure with only the external loading applied to it when the
integral method is employed.
b. Next, for each redundant, internal force diagrams, especially bending moment diagram, of the
primary structure with only the unit value of the redundant applied on it when the graph-multiplication is
employed, or develop the expressions of internal forces of the primary structure with only the unit value of
the redundant applied on it when the integral method is employed.
c. Finally, compute the flexibility coefficients and free terms involved in the compatibility equations
by using the graph-multiplication method or by the integral method and by the application of Maxwell's
law of reciprocal deflections. A displacement (or flexibility coefficient) at the location of a redundant is
considered to be positive if it has the same sense as that assumed for the redundant.
(5) Calculation of internal forces and the construction of their diagrams
Once the redundants have been determined, the internal forces of the indeterminate structure can be
evaluated either through equilibrium conditions or by superposition of the internal forces of the primary
structure due to the external loadings and that due to each of the redundants.
For beams and rigid frames, the member end moment should be determined first, then draw the
bending moment diagram by the method discussed in the text, and then draw shear and axial force
diagrams by any convenient method.

(6) Utilization of symmetric property and simplification of symmetric structures
When a symmetric structure is subjected to a loading that is symmetric with respect to the structure's
axis of symmetry, the response of the structure is also symmetric. Thus we can obtain the response of the
entire structure by analyzing a half of the structure, on either side of the axis of symmetry, with symmetric
boundary conditions; and by reflecting the computed response about the axis of symmetry.
When a symmetric structure is subjected to a loading that is antisymmetric with respect to the
structure's axis of symmetry, the response of the structure is also antisymmetric. Thus, the response of the
entire structure can be obtained by analyzing a half of the structure, on either side of the axis of symmetry,
with antisymmetric boundary conditions; and by reflecting the negative of the computed response about the
axis of symmetry.
The response of a symmetric structure due to a general asymmetrical loading can be obtained by
determining the responses of the structure due to the symmetric and antisymmetric components of the
asymmetrical loading, and by superimposing the two responses.
(7) The calculation of displacements of indeterminate structures and verification of consistent
deformations
The displacements for statically indeterminate structures are the same as that for the corresponding
statically determinate primary structures, so the unit load can be applied on any one of the possible primary
structures of the indeterminate structure to form the virtual system.
For an indeterminate frame or beam, if the value of one of its displacements is given, we may employ
the bending moment diagram obtained by the force method to check the consistent deformation conditions
of the structure by calculating the displacement again. The coincidence of the solution for the displacement
with the given value can verify the satisfaction of the consistent deformation conditions of the structure.
10-1 What is a statically indeterminate structure? What is the difference between a statically
indeterminate structure and a statically determinate structure? What are the degrees of
indeterminacy of an indeterminate structure? How to determine the degrees of indeterminacy of a
structure? How many methods are there for removing excess restraints?
10-2 What is the principle of force method? What are the primary system and primary unknowns? Why
do we have to determine the primary unknowns first when analyzing a statically indeterminate
structure? What is the difference between a primary system and its original structure?
10-3 What is the physical meaning of a force-method equation? The figures (b) and (c) of the problem
show two kinds of the primary structures of the single degree indeterminate rigid frame shown in
the figure (a), please lay off their corresponding primary systems and explain the physical
meanings of two kinds of force-method equations corresponding to the two primary systems, and
sketch the displacements of δ11 and Δ1P on the two corresponding primary structures,
respectively.
P
B B
B
primary structure 1 primary structure 2
A A
A
(a) (b) (c)
Reflecting problem 10-3

(a) an indeterminate rigid frame; (b) primary structure 1; (c) primary structure 2
10-4 Please explain, by physical meanings, why must the main flexibility coefficients be garter than
zero while the secondary coefficients may be greater, or less, or equal to zero?
10-5 Which deformation effects must be considered when calculating the coefficients and free terms in
force-method equations derived from indeterminate beams, bent frames and rigid frames?
10-6 Why are the internal forces of members of a statically determinate structure not related to their
flexural rigidity EI, but those of members of a statically indeterminate structure is related to their
flexural rigidity EI? Why are the internal forces of the members of a statically indeterminate
structure due to external loads related to the
relative value of their flexural rigidity EI and P
not related to the absolute value of EI? B D E
10-7 Which deformation effects must be considered I I
when calculating the coefficients and free
terms in force-method equations derived from
l
A
indeterminate trusses?
C
10-8 Please lay off the bending moment diagrams of
the structure shown in the figure of the
l l
problem under following two conditions: (1)
the area of member CD vanishes, i.e., A → 0 ;
reflecting problem 10-8
(2) the area of member CD goes infinitely great, i.e., A → ∞ .

10-9 What is a symmetric structure? Whether or not the structures shown in the figure of the problem
are symmetric structures? Why?
(a) (b)
D E F D E F
EI EI EI EI
h
EI 2 EI 2 EI EI EI EI
h
B C A B C
A
l l
l l
10-10 Why are the antisymmetric unknown forces on the section lying on axis of symmetry of a
symmetric structure subjected to symmetric loadings equal to zero? Contrariwise, why are the
symmetric unknown forces on the section lying on axis of symmetry of a symmetric structure
subjected to antisymmetric loadings equal to zero?
10-11 Why can only half a structure be analyzed when the structure is a symmetric and subjected to
symmetric or antisymmetric loadings?
10-12 How select the symmetric primary system to analyze for the structures that have two axes of
symmetry, as shown in Fig.10.30 (a) of example 10-8?
EI1 EI
10-13 In example 10-8, if set i1 = , i2 = 2 , i2 = ki1 , how the bending moment diagram of the
l1 l2
example will vary under following two conditions: (1) k → 0 ; (2) k → ∞ ?
10-14 Under what condition the shear and axial deformations may be neglected when calculating the
displacements of an arch?
10-15 Whether or not affirm that the optimal centre line of two hinged arch subjected to vertical
4f
uniformly distributed loads is still a second-degree parabola y = x (l − x ) ? Whether or not the
l2
stressing condition of a three hinged arch with a centre line determined by function
4f
y = 2 x(l − x) subjected to vertical distributed loads is the same as that of a two hinged arch with
l
the same centre line and subjected the same loads?
10-16 Please explain the function of the tie and the variation of the bending moment in a two hinged arch
when the axial rigidity of the tie varies?
10-17 When is the statement that if a structure is subjected to no external loads there is no internal forces
in the structure true?
10-18 What are the difference between the force-method equations used to analyze a statically
indeterminate structure subjected to external loads and those subjected to support movements and
that subjected to temperature changes?
10-19 In which situation is only the relative value of flexural rigidity EI of members of an indeterminate
structure required when analyzing the structure? In which situation is the absolute value of flexural
rigidity EI of members of an indeterminate structure needed when analyzing the structure?
10-20 Consider the beams had support movements as shown in the figures (a) and (b). Whether or not the
force-method equations are identical when selecting the figure (c) and figure (d) as a primary
structure, respectively? What is the meaning of each force-method equation?
(a) (b)
Δ
A EI = constant B A
B
Δ
(d)
(c)
X1
A B B
A
X1
Reflecting problem 10-20
10-21 What is the difference between the determination of displacements for a statically determinate
structure and that for a statically indeterminate structure?
10-22 Why can the unit load be applied on any one of the primary structures of a statically indeterminate
structure when use unit load method to calculate the displacements of the structure?
10-23 Please explain the physical meanings of the condition of consistent deformations of Eq. (10-25),
MM
i.e., ∑ ∫ ds = 0 .
EI
10-1 Determine the degrees of indeterminacy for the statically indeterminate structures shown in the
figure. And remove sufficient restraints to change each of the indeterminate structures into a stable
and statically determinate one.
(a) (b)
(c) (d)
(e) (f)
(g) (h)
Problem 10-1 (contd)

(i)
(j)
Problem 10-1
(a) P (b) P
A B A B
EI EI
l l l l
2 2 2 2
q (d)
(c) D
C
C I1 A I1
A I1 B I2
I1
l l EA = ∞ =k
I2
E I2 B I2 F
I1 = kI 2
①k = 10, ②k = 0.1 l l
2 2
in (d), How is the variation of the bending moment diagram when k increase from small to large
Problem 10-2
10-2 Analyze the structures shown in the figure by force method and draw their bending moment and
shear force diagrams.
10-3 Analyze the rigid frames shown in the figure by force method and draw their bending moment,
shear and axial force diagrams.
10-4 Analyze the rigid frames shown in the figure by force method and draw their bending moment
diagrams.
10-5 Analyze the bent frames shown in the figure by force method and draw their bending moment
diagrams.
q
(a) (b)
1kN/m 1kN/m B
C I 2 = 4 I1 D C
6m
I1 I1
l
B EI = constant
A
A
12m
l
Problem 10-3
10kN/m
D C E
D C E
3m
60kN
10m
3m
A B
3m 3m A B
(a) 2.5m 5m 5m 2.5m
(b)
Problem 10-4
20kN/m EA = ∞ A EA = ∞ B
2m
C D
I 4kN I
1m
6m
I I
(a) 6I (b)
6m
6I
A B
C D
12m 12m
Problem 10-5 (contd)

F EA = ∞ G 4kN/m
2.4m
1.5 1.5
D EA = ∞ E
4.5m 2.1m
① (c)
6.6m
7.5 7.5
③
A B C
numbers enclosed within circles represent the relative values of EI
Problem 10-5
(a) (b)
B EA = constant C
P P P P
D EA = constant E
a
a
A C
B
A D
2a 2a
a
Problem 10-6
3.5m
P = 7.6kN P = 7.6kN
1.2m
I20a 2φ 20 2φ 20
E EI = 1400kN ⋅ m
2
A F B
A B C
2m
C D I 20a
EA = 2.56 × 10 kN 5
P P
2m 2m 2m 4m 2m 2m 4m
10-6 Analyze the trusses shown in the figure by force method.

10-7 The figure shows a composite crane beam. Please analyze the beam and calculate the tensile force
and draw the bending moment diagram for the beam.
10-8 The figure shows a two-span suspending crane beam. Please analyze the beam and calculate the
tensile force and elongation of suspension rods by the consideration of the axial deformation, and
draw the bending moment, shear force diagrams for the beam. Assume that P = 4.5kN , the area
of the rods A = 3.14cm , the moment of inertia for the shaped steel I20a is I = 2370cm .
2 4
10-9 Analyze the structures shown in the figure by using their property of symmetry and draw their
bending moment diagrams. For problems (a) and (f), please draw shear and axial force diagrams as
well.
10kN/m q
(a) (b)
D
D E F
3m
l
E
3m
A B
A B
4m 4m l l
(c) (d) q
P E F G H A C
2I D 2I
I
l
I I I I
h
B
A B D
l l
6kN E F
(e) q (f)
2I
6m
A B I I
l
C 2I D
6m
q I I
l l A B
EI = constant 9m
Problem 10-9
10-10 Analyze the internally indeterminate structure shown in the figure and draw its bending moment
diagram.
10-11 The members of the frame shown in the figure are of a rectangular cross section of b × h . Please
analyze the frame and draw its bending moment diagram by following two conditions: (1) neglect
the axial deformation; (2) take the axial deformation into account. Finally, make a discussion by
comparing the two solutions.
10-12 Please derive the expressions for calculating the internal force of the tie of the two hinged arch
shown in the figure by neglecting the axial deformation and effect of curvature of the arch, i.e.,
ds = dx . The centre line of the arch is a parabola of y=
4f
x(l − x) .
l2
10-13 Determine the internal forces of the arc arch shown in the figure.
10-14 Determine the internal forces at sections A and B of the circular tube shown in the figure.
P P
C
l
l
A B
A B
EI = constant l l
l l
Problem 10-11
Problem 10-10
q
P
EI = constant
y EI = constant C
f =R
f
B x
A A O B
E1 A1
l R R

A
R
B
EI = constant
q
Problem 10-14
(a) (b)
t2 > t1 t2 > t1
A +t1 B A +t1
B
+t2 +t2
l l
Problem 10-15
(a) α (b) α
EI = constant EI = constant B
A α B A α
B1
l
l
(c) (d)
EI = constant B EI = constant B
A A
Δ
B1
B1
l l
Problem 10-16
10-15 Draw the bending moment diagrams for the beams shown in the figure due to temperature changes.
The coefficient of thermal expansion is α ; the depths of the beams are of h.
10-16 Draw bending moment and shear force diagrams of the beams due to support movements as shown
in the figure.
10-17 Draw bending moment diagram of the frame due to support movements as shown in the figure,
a = 30mm , b = 40mm , ϕ = 0.01 radian .
10-18 Calculate the horizontal displacement at point C of the frame shown in the figure of problem
10-3(a).
10-19 Calculate the horizontal displacement at point C of the bent frame shown in the figure of problem
10-5(a).
10-20 Calculate the rotation at end B of the beam shown in the figure of problem 10-15(a).
10-21 Please check the correctness of following bending moment diagrams by the designated conditions:
(1) the relative rotation of an arbitrary section of the beam AB shown in the figure of problem 10-2
(b) is zero; (2) the vertical displacement at support B of the frame shown in the figure of problem
10-3 (b) is zero.
C D E
2m
F
2m
A B
EI = 13440kN ⋅ m
b
ϕ
2m 2m a
Problem 10-17
CHAPTER 11
DISPLACEMENT METHOD

In this chapter, we will study a classical formulation of the displacement (stiffness) method
used for the analysis of statically indeterminate structures. We will first present the fundamental
concepts and principles of displacement method, such as the selection of basic unknowns, primary
structures and primary systems and development of fundamental (equilibrium) equations of
displacement method. We then consider the application of the method to the analysis of
continuous beams, rigid frames, bent frames and other indeterminate structures. Finally, we
introduce the development of displacement-method equations by direct equilibrium method or
slope-deflection method.
11.1 Fundamental Concepts and Analyzing Procedures of Displacement Method
The slope-deflection method for the analysis of indeterminate beams and frames was introduced by
George A. Maney in 1915. The method takes into account only the bending deformations of structures.
Nowadays, the method has been progressed a lot and can be applied to all of statically indeterminate
structures; for generalization, the method is termed displacement method. Although displacement method is
itself considered to be a useful tool for analyzing indeterminate beams and frames, an understanding of the
fundamentals of this method provides a valuable introduction to the matrix displacement method, which
forms the basis of most computer software currently used for structural analysis.
In Chapter 10, we considered the force (flexibility) method of analysis of statically indeterminate
structures. Recall that in the force method, the unknown redundant forces are determined first by solving
the structure’s compatibility equations of consistent deformations; then the other response characteristics of
the structure are evaluated by equilibrium equations or superposition. Unlike the force method, in the
displacement method the unknown displacements are determined first by solving the structure's equilibrium
equations; then the other response characteristics are evaluated through member force-deformation
relations.
The fundamental idea and analyzing procedures of displacement method may be presented by the
following example, a rigid frame shown in Fig. 11.1 (a).
1. Determination of primary unknowns
When the frame is subjected to the external loading, members AC and CB deform, as shown in the
figure using an exaggerated scale. If we ignore the axial deformations of the two members (i.e. elongation
370
11.1 Fundamental Concepts and Analyzing Procedures of Displacement Method 371
θC , the common rotation of member ends of CA and CB. If we

or contraction), only joint C has a rotation
take the rotation as a primary unknown, then develop a equation involving θC and external loads, and
then manage to resolve the θC , the other response characteristics will be obtained through member
force-deformation relations.
P P
MC = 0
C EI1 B C B B
C
θC
θC θC θC
EI 2
l2
(a) (b) (c)

A A
l1 l1
2 2
3
Pl1
16 P kCC
4EI 2 θC = 1
C B
M CP
B l2
C
5
Pl1 θC = 1 3EI1
32 l1
(d) (e)
A A 2EI 2
l2
Fig.11.1 An indeterminate rigid frame analyzed by displacement method

(a) original structure; (b) primary structure; (c) primary system;
(d) primary structure under external load; (e) primary structure under θ C = 1
2. Selection of primary system

Now let us discuss how to determine the rotation θC . The procedure is as follows.
Artificially add a restraint at joint C against the rotation of θC as shown in Fig.11.1 (b) to form a
primary structure, i.e. lock the joint so as to make it have no rotation. Thusly, the frame composed of two
members now becomes an assemblage of two individual single-span indeterminate beams. Next, let the
assemblage of single-span indeterminate beams subject the given external load and the unknown
displacement θC as shown in Fig.11.1 (c) to form a primary system. Note that the primary structure in
372 Chapter 11Displacement Method
displacement method is an assemblage of various single-span indeterminate beams and the primary system
is the assemblage of various single-span indeterminate beams subjected to given external loading and
unknown displacement.
3. Development of displacement-method (equilibrium) equation
To establish the displacement-method equation, let us consider the equilibrium condition of restraint
moments at the artificial restraint at joint C of the primary system due to the combined effect of the external
loading and the unknown rotation θC . Since there is no external moment exerted at joint C of the original
structure, the restraint moments yielded on the artificial restraint at joint C of the primary system due to the
combined effects of the external loading and the unknown rotation θC must be also equal to zero [see
Figs.11.1 (a) and (c)]. That is,
MC = 0 (a)
4. Determine primary unknown

To determine the unknown θC by using the equilibrium condition as mentioned above, the total
restraint moment at the artificial restraint of the primary structure due to the combined effects of the
external load andθC can be conveniently expressed by superimposing (algebraically adding) the restraint
moments due to the external load and the rotation θC acting individually on the primary structure; that is,
M C = M CP + M CC (b)
In which M CP and M CC represent, respectively, the restraint moments on the artificial restraint of the
primary structure due to the external load and the unknown rotation θC , each acting alone on the structure.
Since the unknown rotation θC is unknown, it is convenient to determine M CC by first evaluating
the restraint moment on the artificial restraint due to a unit value of the unknown rotation θC , as shown in
Fig.11.1 (e), and then multiplying the restraint moment thus obtained by the unknown magnitude of the
rotation θC . Thus
M CC = kCCθC (c)
In which kCC denotes the restraint moment at the artificial restraint of the primary structure due to the
unit value of the rotation θC . By substituting equations (b) and (c) into equation (a), we obtain the
equilibrium equation
M C = M CP + kCCθC = 0
The equation can be solved to express the unknown rotation θC in terms of the restraint moment M CP
and the stiffness coefficient kCC of the primary structure. That is,
M CP
θC = −
kCC
11.1 Fundamental Concepts and Analyzing Procedures of Displacement Method 373
Since the primary system is composed by two individual single-span beams, their end moments can be
easily determined by force method previously studied in last chapter. Thusly, we find
kCC = 3( EI1 / l1 ) + 4( EI 2 / l2 ) and M CP = −3Pl1 /16 by force method, we write
3Pl1
θ= 16
3EI1 4 El2
+
l1 l2
After the unknown rotation θC has been computed, all other response characteristics of the original
structure can be also conveniently determined by employing principle of superposition. For instance, the
bending moment diagram for the original structure can be obtained by superimposing the bending moment
diagram of the primary structure due to the external loading and the bending moment diagram of the
primary structure due to a unit value of the unknown rotation θ C
multiplied by the value of θC .
Based on the foregoing discussion, we can summarize the following step-by-step procedure for the
analysis of a statically indeterminate structure by displacement method.
(1) Select primary unknowns. The primary unknowns (or variables) used in displacement method are
joint displacements (joint rotations and joint translations) of the structure.
(2) Select primary system. Add artificial restraints against the primary unknown displacements to
discretize the original structure as an assemblage of individual single-span indeterminate beams. Next, let
the assemblage of individual single-span indeterminate beams subject their given external loads and
primary unknown displacements, which must satisfy the compatibility conditions of consistent
deformations, to form a primary system.
(3) Develop displacement-method (or equilibrium) equations. The displacement-method equations
are the equilibrium equations of restraint forces associated with the primary unknowns. The equations can
be developed by the consideration that the restraint forces induced at the artificial restraints of the primary
system must be equal to the joint forces acting on the original structure at the locations and in the directions
associated with the primary unknown displacements. The stiffness coefficients and free terms in the
equations can be conveniently obtained by referring the tabulated results (see tables 11-1 and 11-2) of
various single-span indeterminate beams calculated by force method learned in Chapter 10.
(4) Determine the primary unknowns. The primary unknown displacements can be determined by
solving the displacement-method equations.
(5) Determine the response characteristics of original structure. Once the unknown displacements
have been computed, the reactions and all other response characteristics of the original structure can be also
conveniently determined by employing superposition relationships. For instance, the bending moment
diagram for the original structure can be obtained by superimposing together each individual bending
moment diagram of the primary structure due to the external loading and due to a unit value of each
unknown displacement multiplied by its value individually.
11.2 Member End Forces of Various Single-Span Indeterminate Prismatic Beams

Due to Their End Displacements and External Loadings
It is known from last section that the primary structure used in displacement method is an assemblage
of individual single-span indeterminate beams. So the relationships between the member end forces and
end displacements and external loadings of various single-span indeterminate prismatic beams are the basis
of displacement method and they must be discussed before the application of displacement method.
11.2.1 Member end forces of various single-span indeterminate prismatic beams due to their end
displacements
To derive the relationship between the member end forces and displacements, let us focus our attention
on an arbitrary prismatic member AB of an indeterminate structure shown in Fig.11.2. When the structure is
subjected to external loads and support settlements, member AB deforms, as shown in the figure, and
internal moments are induced at its ends. The free-body diagram and the elastic curve for member AB are
shown using an exaggerated scale in the figure. As indicated in this figure, double-subscript notation is
used for member end moments, with the first
subscript identifying the member end at which the
moment acts and the second subscript indicating
M AB
the other end of the member. Thus, M AB denotes EI = constant
A ϕ B
the moment at end A of member AB, whereas
θA
Δ
M BA represents the moment at end B of member
QAB B′
AB. Also, as shown in Fig.11.2, θ A and θB M BA
l θB
denote, respectively, the rotations of ends A and B
QBA
of the member with respect to the undeformed
Fig.11.2 Member end forces and displacements
(horizontal) position of the member; Δ denotes
the relative translation between the two ends of the
member in the direction perpendicular to the
undeformed axis of the member; and the angle ϕ denotes the rotation of the member's chord (i.e., the
straight line connecting the deformed positions of the member ends) due to the relative translation Δ .
Since the deformations are assumed to be small, the chord rotation can be expressed as
Δ
ϕ=
l
The sign convention used in this chapter is as follows:
Due to Their End Displacements and External Loadings 375
The member end moments, end rotations, and chord rotation are positive when clockwise;
the sign convention for shear forces is the same as that previously used in the text.
Note that all the member end forces and displacements are shown in the positive sense in Fig. 11.2.
As mentioned previously, the relationship between the member end forces and displacements is the
basement of displacement method. Here, we especially interest in the member end forces expressed by end
displacements. The member end forces due to a unit value of a displacement are named stiffness
coefficients. Since the stiffness coefficients are only related to the material’s property and the size and
shape of the member, they are termed shape constants. The shape constants may be determined
conveniently by force method. Now we will give them directly as follows:
1. Member end forces (shape constants) of member AB due to a unit rotation of end A, i.e., θA = 1
(1) When end B is fixed, i.e., θB = Δ = 0
M AB M AB M AB M BA
EI = constant EI = constant B EI = constant B
A M BA A
B A
θA θA θA
QAB QBA QAB QBA QAB
QBA
l l l
2iABθ A
A B A B
A B
4iABθ A 3iABθ A
iABθ A iABθ A
(c)
(a) (b)
Fig.11.3 Member end forces due to θ A

(a) a single-span indeterminate beam with two fixed supports;
(b) a single-span indeterminate beam with one fixed support and one roller support ;
(c) a single-span indeterminate beam with one fixed support and one double-link support
Consider a single-span indeterminate beam AB with two fixed supports as shown in Fig.11.3 (a).
When support A is subjected to a rotation θ A provided that the other end displacements are prevented, its
deformed shape is denoted in dashed lines using an exaggerated scale in the figure and its corresponding
end forces determined by force method are given by
⎫
M AB = 4iABθ A ⎪
⎪
M BA = 2iABθ A ⎬ (11-1)
6i ⎪
QAB = QBA = − AB θ A ⎪
l ⎭
In which, iAB = EI is named the linear stiffness of member AB. When θ A = 1 , the bending-moment shape
l
constants at ends A and B are, respectively, 4iAB and 2iAB , the shear-force shape constants at the two
6iAB
ends are − . They are actually the restraint forces at the member ends due to the unit rotation θA = 1,
l
provided that the other end displacements are prevented.
(2) When end B is hinged, i.e., M B = 0
Consider a single-span indeterminate beam AB with fixed end A and hinged (or rolled) end B as shown
in Fig.11.3 (b). When support A is subjected to a rotation θ A , its deformed shape is denoted in dashed lines
using an exaggerated scale in the figure and its corresponding end forces determined by force method are
given by
⎫
M AB = 3iABθ A ⎪
⎪
M BA = 0 ⎬ (11-2)
3i ⎪
QAB = QBA = − AB θ A ⎪
l ⎭
It is observed that when θ A = 1 , the bending-moment shape constant at end A is 3iAB , and the
3iAB
shear-force shape constants at ends A and B are − .
l
(3) When end B is restrained by a double-link support
Consider a single-span indeterminate beam AB with fixed support A and double-link support B as
shown in Fig.11.3 (c). When support A is subjected to a rotation θ A , its deformed shape is denoted in
dashed lines using an exaggerated scale in the figure and its corresponding end forces determined by force
method are given by
M AB = iABθ A ⎫
⎪
M BA = −iABθ A ⎬ (11-3)
QAB = QBA = 0 ⎭⎪
It is obvious that when θ A = 1 , the bending-moment shape constants at ends A and B are iAB and
−iAB , respectively; and the shear-force shape constants are equal to zero because the restraint property of
the double-link support at end B.
2. Member end forces (shape constants) of member AB due to a unit relative displacement between
ends A and B, i.e., Δ = 1
(1) When end B is fixed
Consider a single-span indeterminate beam AB with two fixed supports as shown in Fig.11.4 (a).
When support A is fixed and support B is subjected to a translation Δ provided that the other end
displacements are prevented, its deformed shape is indicated by dashed lines using an exaggerated scale in
the figure and its corresponding end forces determined by force method are given by
M AB M BA
M AB
A B
B A
Δ
Δ
QAB
l QAB
QBA l
QBA
6i
Δ M diagram
l 3i M diagram
Δ
A B l
6i A B
Δ
(a) l (b)
Fig.11.4 Member end forces due to Δ
(a) a single-span indeterminate beam with two fixed supports;
(b) a single-span indeterminate beam with one fixed support and one roller support ;
6iAB ⎫
M AB = M BA = − Δ⎪
l ⎪
⎬ (11-4)
12iAB
QAB = QBA = 2 Δ ⎪
l ⎪⎭
6iAB
Thusly, when Δ = 1 , the bending-moment shape constants at ends A and B are − , and the
l
12iAB
shear-force shape constants are .
l2
(2) When end B is hinged
Consider a single-span indeterminate beam AB with fixed support A and hinged (or rolled) support B
as shown in Fig.11.4 (b). When support A is fixed and support B is subjected to a translation Δ , its
deformed shape is denoted in dashed lines using an exaggerated scale in the figure and its corresponding
end forces determined by force method are given by
3iAB ⎫
M AB = − Δ ⎪
l
⎪
M BA = 0 ⎬ (11-5)
3iAB ⎪
QAB = QBA = Δ⎪
l2 ⎭
Table 11-1 shape constants of single-span indeterminate prismatic beams
Bending moments Shear forces

Number Single-span indeterminate beam
M AB M BA QAB QBA
θA = 1
EI B
A 6i 6i
1 4i 2i − −
θA = 1 l l
l
Two fixed
end support
EI B
A
Δ =1
6i 6i 12i 12i
2 − −
l l l2 l2
l
θA = 1
A EI B
3i 3i
3 3i 0 − −
θA =1 l l
One end is l
fixed and one
end is hinged EI B
A
Δ =1
or rolled
3i 3i 3i
4 − 0
l l2 l2
l
One end is θA = 1 EI B
fixed and one A
end is 5 θA = 1 i −i 0 0
supported by
l
double links
3i
It is observed that when Δ = 1 , the bending-moment shape constant at end A is − AB , and the
l
3i
shear-force shape constants at ends A and B are AB .
l2
To facilitate the application of displacement method, table 11-1 gives the shape constants (stiffness
coefficients), denoted by M AB , M BA , QAB and QBA , for various single-span indeterminate prismatic
members with different restraint conditions.
11.2.2 Member end forces of various single-span indeterminate prismatic beams due to typical
loadings
As mentioned previously, the relationships between the member end forces and end displacements and
external loadings of various single-span indeterminate prismatic beams are the basement of displacement
method. Now we will discuss the member end forces of various single-span indeterminate prismatic beams
due to typical loadings. For easy distinguish, the member end forces caused by external loadings are termed
fixed-end forces. Since the fixed-end forces are related with the types of external loadings they are referred
to as loading constants, which can be determined by force method discussed in last chapter.
To facilitate the application of displacement method, table 11-2 gives the loading constants,
F F F F
represented by M AB , M BA , QAB and QBA , for various single-span indeterminate prismatic members with
different restraint conditions.
Table 11-2 loading constants of single-span indeterminate prismatic beams

MAB
F M BA
F QAB
F QBA
F
P
Pl Pl P P
1 A B − + + −
l 8 8 2 2
l
2 2
Two P
ends B Pa2 ⎛ 2b ⎞
A Pab 2 Pab 2 Pb 2 ⎛ 2a ⎞
are 2 − + ⎜1 + ⎟ − ⎜1+ ⎟
a b l2 l2 l2 ⎝ l ⎠ l2 ⎝ l ⎠
fixed
l
q
ql 2 ql 2 ql ql
3 A B − + + −
12 12 2 2
l
The continue of table 11-2

MAB
F M BA
F QAB
F QBA
F
q
ql 2 ql 2 3ql 7 ql
4 A B − + + −
30 20 20 20
Two l
ends
are t1
fixed A B
t2 EI αΔt EI αΔt
5 − 0 0
l h h
Δt = t1 − t2
P
3Pl 11P 5P
6 A
B − 0 + −
16 16 16
l l
2 2
A B Pb( l 2 − b2 ) Pb( 3l2 −b2 ) Pa2 ( 3l −a)

7 − 0 + −
a b 2l 2 2l3 2l3
One
l
end is
fixed
q
and
B ql 2 5ql 3ql
another 8 A − 0 + −
end is 8 8 8
l
hinged
or
rolled q
ql 2 2ql ql
9 A B − 0 + −
15 5 10
l
q
7ql 2 9ql 11ql
10 A B − 0 + −
120 40 40
l

M F
AB M F
BA Q F
AB QBA
F
One end is t1
fixed and A B
t2 3EI αΔt 3EI αΔt 3EI αΔt
another end 11 0 − −
l 2h 2hl 2hl
is hinged or
rolled Δt = t1 − t2
P
Bleft + P
Pl Pl
12 A
B − − +P
2 2 Bright 0
l
P
B
A Pa ( 2l − a ) Pa 2
13 − − +P 0
a b 2l 2l
One end l
is fixed
and
another q
end is B
ql 2 ql 2
supported 14
A − − + ql 0
by double l 3 6
links
q
B ql 2 ql 2 ql
15 A − − + 0
l
8 24 2
5ql 2 ql 2 ql
16 − − + 0
A B 24 8 2
l

MAB
F M BA
F QAB
F QBA
F
One end is
fixed and t1 B
A
t2
another EIαΔt EIαΔt
end is 17 l − 0 0
supported Δt = t1 − t2 h h
by double
links
11.3 Primary Unknowns and Primary Systems in Displacement Method
Recalling from section 11.1, when apply displacement method to analyze a statically indeterminate
structure the primary unknowns and primary system must be first selected.
The primary unknowns are the variables by which the main displacements, used to determine member
end forces, of a structure can be described completely. They are actually the joint translations and rotations
of the structure. The primary system is obtained by first adding the artificial restraints against the primary
unknowns to discretize a structure into an assemblage of single-span indeterminate members (or beams) so
as to form a primary structure, and then subjecting the primary structure to the prescribed external loadings
and the primary unknown displacements. In the following, the choices of primary unknowns and primary
systems will be discussed in details.
11.3.1 Primary unknowns in displacement method
First, let us discuss how to select joint rotations as primary unknowns. Consider the continuous beam
shown in Fig.11.5 (a). As shown in the figure, all joints (joints A, B, C and D) have no translation; the end A
is fixed, so the rotation of end A is zero; the joints B and C are the rigid joint of two members, which will
rotate when the beam is subjected to the external loading, thus we denote, respectively, the rotations as θB
and θC ; the end D is hinged, so there is rotation, represented by θ D , at the joint under the external
loading. If we assume that the rotation at end B of member AB is θ BA , the rotation at end B of member BC
is θ BC , the rotation at end C of member CD is θCD , and the rotation at end D of member CD is θ DC ,
there are totally 5 rotations for the continuous beam. Note however, that if we consider the conditions of
consistent deformations, only 3 of them are independent. That is,
θ BA = θ BC = θ B ⎫
⎬ (a)
θCB = θCD = θC ⎭
(a) P
θ BC θ CB
B
θ BA θ CD θ DC
(b) P
A B C D
θB θC
(c)
A B C D
Fig.11.5 Joint rotations as primary unknowns

(a) original structure; (b) primary system ; (c) primary structure
Therefore, if we consider the consistent deformation conditions and manage to determine the joint
rotations θ B and θC , the member end rotations θ BA , θ BC , θCD and θCB can be determined as well.
Also, the member end forces of the beam can be determined by the relationship between end force and
displacement. The joint rotations θ B and θC , again denoted by unified symbols Δ1 and Δ 2 , can be
taken as the primary unknowns of the beam.
Since end D is hinged the member end moment is zero, i.e., M DC = 0 . The rotation θ D is useless
to determine member end forces and it may thusly not be employed to be a primary unknown.
It is observed from above discussion that there are two rigid joints for the structure and also there are
two primary unknowns. It implies that the number of rotating primary unknowns of a structure is equal to
the number of rigid joints of the structure.
Next, let us discuss how to select joint translations as primary unknowns. Consider the rigid frame
shown in Fig.11.6 (a). As shown in the figure, there are two rigid joints (joints C and D) for the frame.
Recalling from chapter 2 of the text, there are two degrees of freedom for a point, one horizontal
displacement and one vertical displacement, in a plane rectangular coordinate system. There are thusly 4
translating displacements for the rigid frame. Based on the discussion, the bent frame shown in Fig.11.6 (b)
has 6 translations of the three joints (joints D, E and F).
P Δ Δ
Δ Δ Δ
C′ θD
C D′ P D D′ E E′ F F′
θC D
θC θD
A B C
A B
(a) (b)
Fig.11.6 Translating primary unknowns in displacement method
However, the computing task is directly proportional to the number of primary unknowns when use
displacement method to analyze a statically indeterminate structure. In order to reduce analyzing working,
the simplification of primary unknowns is necessary in some cases, especially for hand-oriented calculation.
Thusly, the following assumption is presented:
(1) The axial deformation is ignored due to the axial forces of the members.
(2) The member end rotation θ and the relative translation Δ between member ends are small
enough so that the chord rotation ϕ of the member satisfies the equality ϕ = Δ .
l
By the assumption (1), the straight length of a deformed member is equal to its original undeformed
straight length of the member. According to assumption (2), the length of the chord of a deformed member
is the same as that of the deformed member. By summarizing the above, we conclude that the distance
between two joints of a deformed structure is the same as that of the two joints of the original undeformed
structure. That is, the lengths of the members of a structure remain unchangeable under external loadings.
Based of above assumption, let us discuss the number of independent translations of a structure.
Consider the rigid frame shown in Fig.6 (a). Since there is no vertical translation for the joints C and D in
according with the small deformation assumption, there are only horizontal translations for the joints C and
D. However, because the length of member CD must remain unchangeable the joints C and D only have the
same horizontal translation, herein denoted by Δ . Consequently, only one horizontal translation of the four
translations of joints C and D is independent, which can be selected as one of the primary unknowns of the
structure. Finally, only three joint displacements may be chosen as the primary unknowns. They are joint
rotations at joints C and D ( θ C and θ D ) and the independent horizontal translation Δ of joints C and D.
With similar analysis, the bent frame shown in Fig.11.6 (b) has one primary unknown, which is the
independent translation of joints D, E and F, as shown in the figure.
Since we ignore the distance change between the two ends of a member in the analysis of rigid frames,
the number of independent joint translations of a rigid frame may be determined by the method for
analyzing geometric construction of a structure. In fact, the neglect of the distance change of a member will
make the number of the independent joint translations of a structure unchangeable no matter whether the
members are rigidly or pin connected at the joints. Therefore, if we change the rigid joints (include fixed
supports) of a rigid frame into pin joints to form a pin-connected system, the number of degrees of freedom
of the pin-connected system is the same as the independent translations of the original rigid frame. In other
words, the number of the restraints or links, needed to add to make the pin-connected system to be a stable
system with no redundant restraint, is equal to the number of independent translations of the original rigid
frame.
(a) (b)
Fig.11.7 Determination of translating primary unknowns

(a) an original rigid frame; (b) corresponding pin-connected system
Thusly, we may determine the independent joint translations of a rigid frame by analyzing the degrees
of freedom of a corresponding pin-connected system. Consider the rigid frame shown in Fig.11.7. In order
to determine the independent translations of the structure, change all of the rigid joints (include three fixed
supports) into hinged joints to form a pin-connected system as shown in Fig.11.7 (b) by solid lines. The
pin-connected system shown in Fig.11.7 (b) by solid lines is a mechanism, so we must add two restraints or
links, shown in the figure by dashed lines, to make the pin-connected system be a stable system with no
redundant restraint. The original rigid frame thusly has two independent translations.
In general, the primary unknowns in displacement method for a statically indeterminate rigid frame
include joint rotations and independent translations of the frame. The number of joint rotations is equal to
the number of rigid joints; the number of joint independent translations is the same as that of the degrees of
freedom of a corresponding pin-connected system obtained by changing the rigid joints of the original
frame into hinged joints.
Noted from above discussion, that the conditions of consistent deformations (or compatibility
conditions of displacements) of a structure have already been satisfied because the same rotations and the
same translations of the member ends connected to a joint are guaranteed when determining the primary
unknowns of the structure.
11.3.2 Primary systems in displacement method
As mentioned in section 11.1, the primary structure of a structure is an assemblage of individual

single-span indeterminate beams, which is obtained by adding corresponding artificial restraints against
each of the primary unknowns of the structure. The primary system is obtained by subjecting the
assemblage of individual single-span indeterminate beams to their given external loads and primary
unknown displacements, which must satisfy the compatibility conditions of consistent deformations.
P1 P2 P1 P2
Δ1 C A D
A D A D C C
Δ1 Δ1 (θ D ) Δ1
B B B
(a) (b) (c)

Fig.11.8 Primary system in displacement method
(a) original structure; (b) primary structure; (c) primary system
For example, Fig.11.8 (a) shows a rigid frame with only one rigid joint D. Since there is no joint
translation the structure only has one primary unknown, which is the joint rotation of joint D, θ D . The
primary structure can be obtained by adding artificially a restraint against the rotation θ D as shown in
Fig.11.8 (b). The primary system can be gained by subjecting the primary structure to the external loading
and the primary unknown as shown in Fig.11.8 (c), which actually is the composition of individual
single-span indeterminate beams under the combined actions of the external loading and the support
movements.
Similarly, the primary system and primary structure of the structure shown in Fig.11.5 (a) are,
respectively, the system shown in Fig.11.5 (b), which is the assemblage of three single-span indeterminate
beams under the combined actions of a concentrated force and two support rotations, and the structure
shown in Fig.11.5 (c), which is the assemblage of three single-span indeterminate beams.
Again consider the rigid frame shown in Fig.11.9 (a). Since the frame has one rigid-joint rotation and
one independent joint translation, the primary unknown for the frame is two, the rotation of joint C and the
translation of joints C and D. Therefore, we add a rotative restraint against the rotation of joint C and a
translating restraint against the translation of joints C and D as shown in Fig.11.9 (c) to form the primary
structure; and let the primary structure undergo the combined actions of the external loading and the
primary unknown to form the primary system as shown in Fig.11.9 (b).
Δ2 Δ2 Δ2 Δ2
C D C D C D
P Δ1 (θ D ) P Δ1
Δ1 (θ D )
A B A B A B
(a) (b) (c)
Fig.11.9 Primary system in displacement method

(a) original structure; (b) primary system ; (c) primary structure
It is observed from foregoing discussion that the primary system of a structure can be obtained by
adding the artificial restraints against the primary unknowns on the original structure. If a primary unknown
is a rigid-joint rotation, a restraint resisting the rotation, herein named rotative restraint, will be added;
similarly, if a primary unknown is a joint translation, a restraint resisting the translation (or a link), termed
translating restraint, will be affixed. Note that the two kinds of artificial restraints are independent each
other, the rotative restraint only resists the joint rotation, whereas translating restraint just prevent the joint
translation. The distinguish between the primary system and the original structure is that the artificial
restraints against the primary unknowns have been affixed to the original structure so as to make the
original structure to be an assemblage of single-span indeterminate beams. The following section will
discuss how to develop the fundamental formulations in displacement method by using the primary system
of a structure.
11.4 Displacement-Method Equations
How to let the response characteristics (reactions, internal forces and displacements) of the primary
system of a structure due to the combined action of the external loadings and the primary unknowns be
identical to those of the original structure? The transforming conditions are just the equilibrium conditions
of restraint forces acting on the artificial restraints; they are actually the displacement-method equations.
The equations can be developed by the consideration that the restraint forces induced at the artificial
restraints of the primary system must be equal to the joint forces acting on the original structure at the
locations and in the directions associated with the primary unknown displacements. In order to make the
expressions of the displacement-method equations have generality, all of the primary unknowns are
denoted by unified symbol Δ in the following discussion.
11.4.1 Development of displacement-method equations
To illustrate the development of displacement-method equations, consider the rigid frame shown in
Fig.11.10 (a). Since the rigid frame has only one rigid joint and has no joint translation, the primary
unknown of the frame must be the rotation of joint C, denoted by Δ1 . The primary system will be obtained
by adding artificial restraint against the rotation to the original structure as shown in Fig.11.10 (b).
The displacement-method equation can be developed by the consideration that the restraint forces
induced at the artificial restraint of the primary system must be equal to the joint forces acting on the
original structure at the locations and in the directions associated with the primary unknown displacement.
Since no joint moment is exerted at joint C, the restraint moment at the artificial restraint due to the
combined effects of the external loading and the unknown rotation Δ1 must be also equal to zero. That is,
F1 = 0 (a)
q F1 q F1P q F11
C B B B B
C C C
Δ1 Δ1 Δ1
Δ1 2I
Δ1 Δ1
I = +
A (a) A (b) A A (d)

(c)
Fig.11.10 Figures to show the development of displacement-method equation

(a) original structure; (b) primary system ; (c) primary structure under external loading;
(d) primary structure under support rotation
Apparently, the equation (a) is an equilibrium equation. It is can be also established by considering the
reason that if the value of the restraint moment at the artificial restraint yielded by the external load is the
same in magnitude but opposite in direction as the moment necessary to rotate the joint C a angle in the
value of Δ1 , the restraint moment at the artificial restraint on the primary system must be equal to zero.
Then the moment equilibrium condition at joint C of the primary system due to the combined effects of the
external load and the unknown rotation Δ1 is the same as that of the original structure at joint C.
The restraint force F1 can be determined by principle of superposition. That is, the total restraint
moment at the artificial restraint of the primary structure due to the combined actions of the external load
and the joint rotation Δ1 can be conveniently expressed by superimposing (algebraically adding) the
restraint moment due to the external load and the rotation Δ1 subjected individually to the primary
structure; that is,
F1 = F1P + F11 (b)
In which F1P and F11 represent, respectively, the restraint moments at the artificial restraint of the
primary structure due to the external load and the unknown rotation Δ1 , each acting alone on the structure.
Note that two subscripts are used to denote the restraint moments, F1P and F11 , of the primary structure.
The first subscript 1 indicates the location of these restraint moments; the second subscript P is used to
indicate that F1P is caused by the given external loading, whereas the second subscript 1 of F11 implies
that it is due to the unknown rotation Δ1 . Both of these restraint moments are considered to be positive if
they occur in the direction of the unknown rotation Δ1 , which is assumed to be clockwise.
Since the unknown rotation Δ1 is unknown, it is convenient to determine F11 by first evaluating the
restraint moment at the artificial restraint due to a unit value of the unknown rotation Δ1 , as shown in
Fig.11.10 (d), and then multiplying the restraint moment thus obtained by the unknown magnitude of the
rotation Δ1 . Thus
F11 = k11Δ1 (c)
In which k11 denotes the restraint moment at the artificial restraint of the primary structure due to the unit
value of the rotation Δ1 , which has the unit of moment per unit radian, is referred to as a stiffness
coefficient. By substituting equations (b) and (c) into equation (a), we obtain the equilibrium equation
F1 = k11Δ1 + F1P = 0 (11-6)
It is observed that when use displacement method to analyze a statically indeterminate structure, one
primary unknown matches one corresponding artificial restraint, and one artificial restraint will yield a
equilibrium equation. Therefore, the total number of primary unknowns must be equal to the total number
of equilibrium equations.
11.4.2 Canonical equations of displacement method
For a structure which has more than one primary unknown, the displacement-method equations may
be also developed by the idea used in above subsection. First, let us consider a structure with two primary
unknowns.
1. Displacement method of two primary unknowns
Now let us focus our attention on the rigid frame shown in Fig.11.11 (a) to present how
displacement-method equations for multiple primary unknowns are developed.

(1) Determine primary unknowns
The frame has one rigid-joint rotation (of joint C) and one independent joint translation (of joint D)
based on the assumption in subsection 11.3.1, so the frame has two primary unknowns, denoted by Δ1
and Δ 2 as shown in Fig.11.11 (a).
Δ2 Δ2 F1 Δ2
D D D F2
C C
C
Δ1 Δ1
P Δ1 P Δ1
A B A B A (c) B
(a) (b)
Δ2 = 1 Δ2 = 1
F1P F12
F2P F11
D D F21 D F22
C C C
Δ1 = 1
P
Δ1 = 1
A (d) B A B A B
(e) (f)
Fig.11.11 A rigid frame with two primary unknowns

(a) original structure; (b) primary structure; (c) primary system ; (d) primary structure under external load;
(e) primary structure under Δ1 =1; (f) primary structure under Δ 2 = 1
(2) Develop primary structure and primary system

Remove the external loading and add artificial restraints against each of the primary unknowns, thusly
one rotative restraint against joint rotation Δ1 is affixed at joint C and one translating restraint (or link)
against joint translation Δ 2 is added at joint D, to form a primary structure, which is an assemblage of 3
single-span indeterminate beams, as shown in Fig.11.11 (b). The primary system can be obtained by
subjecting the external loading and the primary unknown displacements, which now are the support
movements of the single-span indeterminate beams, to the primary structure to form a primary system as
shown in Fig.11.11 (c).
(3) Develop displacement-method equations
The displacement-method equations are actually equilibrium conditions developed at the artificial
restraints. They can be established by using the principle of superposition as follows:

① Determine the restraint forces corresponding to the primary unknowns at each artificial restraint
of the primary structure with only external loading applied on it.
② For each primary unknown, determine the restraint forces corresponding to all the primary
unknowns at each artificial restraint of the primary structure with only the unit value of the primary
unknown subjected to it, as shown in Figs.11.11 (e) and (f). By the principle of superposition, the real
restraint force at the artificial restraint against primary unknown Δ i due to the unit value of the primary
unknown Δ j , termed Fij , must be
Fij = kij Δ j (a)
Where kij represents the restraint force against primary unknown Δ i due to the unit value of the
primary unknown Δ j . Note that two subscripts are used to denote the restraint forces Fij and kij of the
primary structure. The first subscript i indicates the location and direction of these restraint forces; the
second subscript j is used to indicate the restraint forces caused by the unknown rotation Δ j .
③ Write a displacement-method (equilibrium) equation for the location of each artificial restraint
by setting the algebraic sum of the restraint forces (or moments) of the primary structure due to the external
loading and each of the primary unknowns equal to the known joint force (or moment) at the corresponding
location on the original indeterminate structure. The total number of equilibrium equations thus obtained
must be equal to the number of the primary unknowns. Since the example has two primary unknowns and
no joint forces exerted on it, there are two equilibrium equations which can be expressed, by principle of
superposition, as
F1 = F1P + F11 + F12 = 0 ⎫
⎬ (b)
F2 = F2 P + F21 + F22 = 0 ⎭
In which F1 and F2 denote, respectively, the total restraint moment at artificial restraint 1 (rotative
restraint at joint C) and restraint force at artificial restraint 2 (translating restraint at joint D). F1P , F11 and
F12 represent, respectively, the restraint moments at the artificial restraint 1 of the primary structure due to
the external load, the unknown rotation Δ1 and unknown translation Δ 2 , each acting alone on the
structure. Similarly, F2P , F21 and F22 denote, respectively, the restraint forces at the artificial restraint
2 of the primary structure due to the external load, the unknown rotation Δ1 and unknown translation Δ 2 ,
each acting alone on the structure.
By equation (a), we write equation (b) in detail as
k11Δ1 + k12 Δ 2 + F1P = 0 ⎫
⎬ (11-7)
k21Δ1 + k22 Δ 2 + F2 P = 0 ⎭
The equation (11-7) is just the displacement-method equation for the indeterminate structure with two
primary unknowns, which can be solved by the equation.
2. Displacement method of n primary unknowns
For a structure with n primary unknowns, the canonical equations in displacement method may be
written as
k11Δ1 + k12 Δ 2 + " + k1n Δ n + F1P = 0 ⎫
k21Δ1 + k22 Δ 2 + " + k2 n Δ n + F2 P = 0 ⎪⎪
⎬ (11-8)
"" ⎪
kn1Δ1 + kn 2 Δ 2 + " + knn Δ n + FnP = 0 ⎪⎭
In which, the coefficient kii , termed main stiffness coefficient, represents the restraint force (or moment) at
the location and in the direction of ith artificial restraint of the primary structure due to the unit value of
primary unknown Δi ; the coefficient kij , termed secondary stiffness coefficient, represents the restraint
force (or moment) at the location and in the direction of ith artificial restraint of the primary structure due
to the unit value of primary unknown Δ j ; the free term FiP denotes the restraint force (or moment) at
the location and in the direction of ith artificial restraint of the primary structure due to the external loading.
The meaning of each of the equations in expression (11-8) is that all the restraint forces exerted at the
associated artificial restraint of the primary structure due to the combined actions of the external loading
and the primary unknown displacements remain in equilibrium. So the expression (11-8) is the equilibrium
conditions of the primary system. The total number of equilibrium equations is equal to the number of the
primary unknowns. Consequently, the n primary unknowns can be solved by the equations.
It should be explained that when develop the displacement-method equations all primary unknowns,
Δ1 , Δ 2 , ", Δ n are assumed to be positive. That is, positive joint rotation is clockwise; positive joint
translation makes the chord rotation to be clockwise too. The positive answer of a primary unknown
implies that the direction of the primary unknown is coincident with the assumed one and vice versa.
Obviously, the stiffness coefficient kij is the restraint force (or moment) at the location and in the
direction of ith artificial restraint of the primary structure due to the unit value of jth support movement
associated with Δ j of the primary structure. Recalling from previous discussion, it should be the shape
constant and can be obtained from table 11-1 listed in section 11.2. The free term FiP is the restraint force
(or moment) at the location and in the direction of ith artificial restraint of the primary structure due to the
external loading, it should be loading constant defined in section 11.2 and may be determined by using
table 11-2 listed in section 11.2.
Since kii is a main coefficient it is always great than zero; kij is a secondary coefficient, so it may
11.5 Analysis of Statically Indeterminate Beams and Rigid Frames without Sidesway 393
be great than or less than or equal to zero. Recalling from the law of reciprocal reactions, we write
kij = k ji (11-9)
11.5 Analysis of Statically Indeterminate Beams and Rigid Frames without Sidesway
The following examples will illustrate the analyzing procedure for statically indeterminate beams and
rigid frames without sidesway by using displacement method.
Example 11-1
Determine the internal forces for the continuous beam shown in Fig.11.12 (a) by displacement method.
The flexural rigidity of the beam is constant, i.e., EI = constant .
Solution
(1) Primary unknowns

Since the two-span continuous beam is connected by rigid joint B, the joint rotation Δ1 will be chosen
as the primary unknown of the beam.
(2) Primary structure and primary system
Remove all external loads and add an artificial restraint against the primary unknown (the rotation of
joint B) at joint B to form a primary structure. The primary system will be obtained by subjecting the
external loading and the joint rotation Δ1 to the primary structure as shown in Fig.11.12 (b).
(3) Develop displacement-method (or equilibrium) equation
The equilibrium equation is based on the requirement that the restraint moment of the artificial
restraint induced by the combined actions of the external loading and the joint rotation Δ1 must be equal to
zero because there is no joint moment acting at the joint of the original structure. By the principle of
superposition, we write
k11Δ1 + F1P = 0
(4) Calculation of stiffness coefficient
The coefficient k11 actually is the restraint moment induced at the artificial restraint due to the unit
value of the joint rotation Δ1 . Referring to the table 11-1, we find
M BC = 3i M BA = 4i M AB = 2i
The bending moment diagram of the primary beam due to Δ1 = 1 is shown in Fig.11.12 (c). By the
moment equilibrium condition due to Δ1 = 1 as shown in Fig.11.12 (d), the restraint moment k11 at the
artificial restraint due to Δ1 = 1 must satisfy the following equilibrium condition. That is,
∑M B =0 k11 = 4i + 3i = 7i
EI
In which, i = , liner flexural rigidity of the beam.
6
2kN/m 16kN (b) 2kN/m 16kN

(a)
B C C
A A
B Δ1
6m 3m 3m
2kN/m 18
4i
(e)
(c) 16kN
B Δ1 = 1 C 6 6
A
Δ1 = 1 A C
B 15
2i
M 1 diagram 3i M P diagram
(unit: kN ⋅ m)
k11 F1P
(d) (f)
4i 3i 6 18

(a) original structure; (b) primary system; (c) bending moment diagram due to Δ1 = 1;
(d) equilibrium condition of k11 ;(e) bending moment diagram due to external loading;
(f) equilibrium condition of F1P
(5) Calculation of free term

The free term F1P actually is the restraint moment induced at the artificial restraint due to the
external loading, termed fixed-end moment. Referring to the table 11-2, we find
ql 2 2 × 62
− M AB
F
= M BA
F
= = = 6kN ⋅ m
12 12
3Pl 3 × 16 × 6
F
M BC =− =− = −18kN ⋅ m
16 16
The bending moment diagram of the primary beam due to the external loading is shown in Fig.11.12 (e).
By the moment equilibrium condition due to the external load as shown in Fig.11.12 (f), the restraint
moment, F1P , at the artificial restraint due to the external loading must satisfy the following equilibrium
condition. That is,
∑M B =0 F1P + 18 − 6 = 0 F1P = −12kN ⋅ m

(6) Solve primary unknown
Substituting the numeric values of stiffness coefficient k11 and free term F1P , we solve the primary
unknown Δ1 to be
F1P 12 1
Δ1 = − = = 1.714
k11 7i i
(7) Draw bending moment diagram
By the superposing equation for bending moment, i.e., M = M 1Δ1 + M P , the member end moments
are calculated to be
⎛ 1.714 ⎞
M AB = 2iΔ1 + M AB
F
= 2i ⎜ ⎟ − 6 = −2.57kN ⋅ m
⎝ i ⎠
⎛ 1.714 ⎞
M BA = 4iΔ1 + M BA
F
= 4i ⎜ ⎟ + 6 = 12.86kN ⋅ m
⎝ i ⎠
⎛ 1.714 ⎞
M BC = 3iΔ1 + M BC
F
= 3i ⎜ ⎟ − 18 = −12.86kN ⋅ m
⎝ i ⎠
Once the member end moments have been determined, the bending moment diagram can be drawn by
superimposing method used previously in the text, as shown in Fig.11.13 (a). Mention again that the
bending moment diagram must be drawn on the tensile side of the beam.
(8) Draw shear force diagram
Once the member end moments are determined the member end shear forces can be determined by
consideration of the free body diagram of each member. By the free body diagrams shown in Fig.11.13 (b),
we write
−12.86 + 2 × 6 × 3 + 2.57
∑M B =0 QAB =
6
= 4.29kN
−12.86 − 2 × 6 × 3 + 2.57
∑M A =0 QBA =
6
= −7.72kN
−16 × 3 + 12.86
∑M B =0 QCB =
6
= −5.86kN
16 × 3 + 12.86
∑M C =0 QBC =
6
= 10.14kN
After the member end shear forces have been determined, the shear force diagram can be drawn by the
method used previously in the text, as shown in Fig.11.13 (c).
(9) Verification
By considering the moment equilibrium condition of joint B, we write
∑M B = +12.86 − 12.86 = 0
Obviously, the moment equilibrium condition is indeed satisfied.
By consideration of the equilibrium condition of vertical component projections of all the forces
acting on the entire beam, we write
∑Y = 0 4.29 + 17.86 + 5.86 − 2 × 6 − 16 ≈ 0

Apparently, the equilibrium condition is satisfied as well.
The above verification implies that the calculation for the example is correct.
(a)
2.57 (9) 12.86 (b) 2kN/m 16kN
(24) 12.86
2.57
A C A B C
B
1.29
QAB QBA QBC QCB
(unit: kN ⋅ m) 17.57
(c)
4.29 10.14 10.14
A B C
5.86 5.86
(unit: kN) 7.72
(a) bending moment diagram; (b) free body diagrams ; (c) shear force diagram
Example 11-2
Determine the bending moment diagram for the rigid frame shown in Fig.11.14 (a) by displacement
method.
Solution

Because the rigid frame consists of 5 members connected by rigid joints D and E as shown in the
figure, the rotation of joint D, Δ1 , and the rotation of joint E, Δ 2 , will be chosen as the primary
unknowns of the frame.
Remove all external loads and add two artificial restraints against the primary unknowns, the rotations
of joints D and E, at the two joints to form a primary structure. The primary system will be obtained by
subjecting the external loading and the joint rotations Δ1 and Δ 2 to the primary structure as shown in
Fig.11.14 (b).
(3) Develop displacement-method (or equilibrium) equations

Each of the equilibrium equations is based on the requirement that the restraint moments yielded at
each of the artificial restraints due to the combined effects of the external loading and the joint rotations
Δ1 and Δ2 must be equal to zero, because there is no joint moment acting at the joints for the original
structure. By superposing the combined effects of the external loading and the joint rotations Δ1 and Δ 2 ,
we write the displacement-method equations to be
k11Δ1 + k12 Δ 2 + F1P = 0
k21Δ1 + k22 Δ 2 + F2 P = 0
(a) 20kN/m
C F
4I 0 D 6I 0 E 4I 0
4m
2I 0 3I 0
A B
4m 6m 4m
(b) 20kN/m
C F
D E Δ2
Δ1
4m
A B
4m 6m 4m

(4) Calculation of stiffness coefficients

In order to calculate the stiffness coefficients, we first determine the linear flexural rigidity of each
member of the frame. As discussed in last chapter, the internal forces of a statically indeterminate structure
are only related to the relative flexural rigidity of each member of the structure. Thusly, we assume
4 EI 0 6 EI 0
iCD = iEF = =1 iDE = =1
4 6
2 EI 0 1 3EI 0 3
iDA = = iEB = =
4 2 4 4
(a) 3iDC (b) (c)

Δ1 = 1 2iDE
k11
k21
C D 3iDC 4iDE 2iDE
4iDA E F D E
Δ1 = 1
Δ1 = 1 4iDE
4iDA
M 1 diagram
A 2iDA B
Fig.11.15 M 1 diagram of example 11-2

(a) M 1 diagram; (b) equilibrium condition of k11 ; (c) equilibrium condition of k21
(a) 4iDE (b) (c)

Δ2 = 1 k22
C D E Δ2 = 1 F k12
4iEB 2iDE 4iDE E 3iEF
2iDE
Δ2 = 1 3iEF D
4iEB
M 2 diagram
A B 2iEB
Fig.11.16 M 2 diagram of example 11-2

(a) M 1 diagram; (b) equilibrium condition of k12 ; (c) equilibrium condition of k22
Actually, the coefficient kij is the restraint moment induced at ith artificial restraint due to the unit
value of the joint rotation Δ j . Referring to the table 11-1, we find the shape constants due to Δ1 = 1 as
follows
1
M DC = 3iCD = 3 M DA = 4iDA = 4 × =2
2
M DE = 4iDE = 4 M ED = 2iDE = 2
And the shape constants due to Δ 2 = 1 must be
M ED = 4iED = 4 M EF = 3iEF = 3 M EB = 4iEB = 3

The bending moment diagrams of the primary frame due to Δ1 = 1 and due to Δ 2 = 1 are shown in
Fig.11.15 (a) and Fig.11.16 (a), respectively.
By the moment equilibrium conditions due to Δ1 = 1 as shown in Figs.11.15 (b) and (c), the restraint
moments, k11 and k21 , at the artificial restraints 1 and 2 must satisfy the following equilibrium
conditions. That is,
∑M D =0 k11 = 3iCD + 4iDA + 4iDE = 3 + 2 + 4 = 9
∑M E =0 k21 = 2iDE = 2
By the moment equilibrium conditions due to Δ 2 = 1 as shown in Figs.11.16 (b) and (c), the restraint
moments, k12 and k22 , at the artificial restraints 1 and 2 must satisfy the following equilibrium
∑M D =0 k12 = 2iDE = 2 = k21
∑M E =0 k22 = 4iDE + 4iEB + 3iEF = 3 + 3 + 4 = 10

In fact, the free term FiP is the restraint moment induced at ith artificial restraint due to the external
loading, termed fixed-end moment or loading constant. Referring to the table 11-2, we find
ql 2 20 × 62
M F
DE = −M F
ED =− =− = −60kN ⋅ m
12 12
1 2 1× 20 × 42
M F
EF = − ql = − = −40kN ⋅ m
8 8
F2 P (b) (c)
(a) F1P
60 60 F1P F2 P
40
C F
D E D 60 60 40
E
M P diagram
(unit: kN ⋅ m)
A B
Fig.11.17 M P diagram of example 11-2

(a) M P diagram; (b) equilibrium condition of F1P ; (c) equilibrium condition of F2 P
The bending moment diagram of the primary frame due to the external loading is shown in Fig.11.17
(a). By the moment equilibrium conditions due to the external load as shown in Figs.11.17 (b) and (c), the
restraint moments, F1P and F2P , at the artificial restraint 1 and 2 must satisfy the following equilibrium
∑M D =0 F1P + 60 = 0 F1P = −60kN ⋅ m
∑M E =0 F2 P + 40 − 60 = 0 F2 P = 20kN ⋅ m
Substituting the numeric values of stiffness coefficients and free terms into displacement-method
equations, we write
⎧9Δ1 + 2Δ 2 − 60 = 0
⎨
⎩2Δ1 + 10Δ 2 + 20 = 0
Thusly, the primary unknown Δ1 and Δ 2 are solved to be
Δ1 = 7.442 Δ 2 = −3.488
By the superposing equation for bending moment, i.e., M = M 1Δ1 + M 2 Δ 2 + M P , the member end
moments are calculated to be
1
M AD = 2iDA Δ1 = 2 × × 7.442 = 7.44kN ⋅ m
2
1
M DA = 4iDA Δ1 = 4 × × 7.442 = 14.88kN ⋅ m
2
3
M BE = 2iBE Δ 2 = 2 × × (−3.488) = −5.23kN ⋅ m
4
3
M EB = 4iBE Δ 2 = 4 × × (−3.488) = −10.46kN ⋅ m
4
M DC = 3iDC Δ1 = 3 ×1× 7.442 = 22.33kN ⋅ m
M DE = 4iDE Δ1 + 2iDE Δ 2 + M DE F = 4 × 1× 7.442 + 2 × 1× (−3.488) − 60 = 37.21kN ⋅ m
M ED = 2iDE Δ1 + 4iDE Δ 2 + M ED
F
= 2 × 1× 7.442 + 4 × 1× (−3.488) + 60 = 60.93kN ⋅ m
M EF = 3iEF Δ 2 + M EF
F
= 4 × 1× ( −3.488) − 40 = −50.46kN ⋅ m
superimposition method used previously in the text, as shown in Fig.11.18.
37.21 ( 90 ) 60.93
22.33 50.46 ( 40 )
C
F
14.88 D E 10.46
40.93 14.77
M diagram
(unit: kN ⋅ m)
A B
7.44 5.23
Fig.11.18 Bending moment diagram of example 11-2
11.6 Analysis of Statically Indeterminate Rigid and Bent Frames with Sidesway
It is realized from last section that for the structures whose joint rotations are chosen as the primary
unknowns, the physical meanings of the corresponding displacement-method equations are the moment
equilibrium equations of the joints. Similarly, for the structures whose joint translations are chosen as the
primary unknowns, the physical meanings of the corresponding displacement-method equations are the
force component projection equilibrium equations of the forces acting on the entire or a portion of the
primary structure, which is
isolated by the cut along the associated directions of the translations. Following examples will further
illustrate the application of displacement method to the analysis of statically indeterminate rigid and bent
frames with sidesway.
Example 11-3
Draw the internal force diagrams for the rigid frame shown in Fig.11.19 (a) by using displacement
method.
Solution

By the similar analysis with that of the examples solved in last section, the rotation of joint C, denoted
by Δ1 , and the translation of joint D, represented by Δ 2 , will be chosen as the primary unknowns of the
frame.
Remove all external loads and add two artificial restraints against the primary unknowns (the rotations
of joint C and the translation of joint D) at joints C and D to form a primary structure. The primary system
will be obtained by subjecting the external loading and the joint displacements Δ1 and Δ 2 to the
primary structure as shown in Fig.11.19 (b).
(a) (b) Δ2
C D C D
3EI Δ1
10kN/m
4m
EI EI
A B A B
6m

(3) Develop displacement-method (or equilibrium) equations

The first equilibrium equation is based on the requirement that the restraint moments yielded at the
artificial restraint added at joint C due to the combined effects of the external loading and the rotation of
joint C and the translation of joint D, Δ1 and Δ 2 , must be equal to zero, because there is no joint moment
acting at the joint C for the original structure. The second equilibrium equation is based on the condition
that the restraint forces yielded at the artificial restraint added at joint D due to the combined effects of the
external loading and the rotation of joint C and the translation of joint D, Δ1 and Δ 2 , must be equal to
zero, because there is no joint force acting at the joint D for the original structure. By superposing the
combined effects of the external loading and the joint displacements Δ1 and Δ 2 , we write the
displacement-method equations to be
k11Δ1 + k12 Δ 2 + F1P = 0
k21Δ1 + k22 Δ 2 + F2 P = 0
(4) Calculation of stiffness coefficients
In order to calculate the stiffness coefficients, we first calculate the linear flexural rigidity of each
member of the frame. Thusly, we assume
EI 3EI
iAC = iBD = =i iCD =
= 2i
4 6
Referring to the table 11-1, we find the shape constants due to Δ1 = 1 as follows
M CA = 4iCA = 4i M AC = 2iCA = 2i M CD = 3iCD = 6i
And the shape constants due to Δ 2 = 1 are
6iAC 6i
M AC = M CA = − = − = −1.5i
lCA 4
3iBD 3
M BD = − =− i
lBD 4
The bending moment diagrams of the primary frame due to Δ1 = 1 and due to Δ 2 = 1 are shown in
Fig.11.20 (a) and Fig.11.21 (a), respectively.
By the equilibrium conditions of the free body diagrams shown in Figs.11.20 (b) through (e), the
restraint moment k11 and restraint force k21 , at the artificial restraints 1 and 2 must satisfy the following
equilibrium conditions. That is,
∑M = 0
C k11 = 3iCD + 4iCA = 6i + 4i = 10i [see Fig.11.20 (b)]
∑X =0 QCA + QDB = k21 [see Fig.11.20 (c)]
∑M = 0
A QCA × 4 + M AC + M CA = 0 [see Fig.11.20 (d)]
( M AC + M CA ) (2i + 4i )
QCA = − =− = −1.5i
4 4
∑ M B = 0 QDB × 4 + M BD = 0 QDB = 0 [see Fig.11.20 (e)]
Thusly, we find
k21 = −1.5i
Note that the coefficient k21 is obtained by considering the force component projection equilibrium
equation of the forces acting on a portion of the primary structure, which is isolated by the cut through the
top of the columns and along the associated direction of the translation of joint D, as shown in Fig.11.20
(c).
By the equilibrium conditions of the free body diagrams shown in Figs.11.21 (b) through (e), the
restraint moment k12 and restraint force k22 , at the artificial restraints 1 and 2 must satisfy the following
equilibrium conditions. Thusly,
6iAC
∑M C =0 k12 +
4
=0 k12 = −1.5i = k21 [see Fig.11.21 (b)]
∑X =0 QCA + QDB = k22 [see Fig.11.21 (c)]
( M AC + M CA ) (−1.5i − 1.5i ) 3
QCA = − =− = i [see Fig.11.21 (d)]
4 4 4
M (−0.75i ) 0.75
QDB = − BD = − = i [see Fig.11.21 (e)]
4 4 4
Consequently, we obtain
3 0.75 3.75
i+ i = k22 k22 = i
4 4 4
(a) (b) (c)

D k21
k11
k11 C
4iCA C Δ1 = 1 D QCA QDB
C
3iCD M CA QCA QDB
C D
Δ1 = 1 3iCD 4iCA
(d) (e)
4m
4m
M 1 diagram
A 2iCA B
QAC A B
QBD
M AC M BD
Fig.11.20 and free body diagrams of example 11-3

(a) M 1 diagram; (b) equilibrium condition of k11 ; (c) equilibrium condition of k21 ;
(d) free body diagram of member CA; (e) free body diagram of member DB
(a) (b) (c)

Δ2 = 1 Δ2 = 1 D k22
k12
k12 C
C D k 22 C 0 QCA QDB
6iAC MCA
QCA QDB
4 6iAC C D
4
M 2 diagram (d) (e)
4m
4m
6iAC A 3iBD B
4 4
QAC A B
QBD
M AC M BD
Fig.11.21 M 2 and free body diagrams of example 11-3

(a) M 2 diagram; (b) equilibrium condition of k12 ; (c) equilibrium condition of k22 ;

The fixed-end forces or loading constants can be determined by referring to the table 11-2. Thusly, we
write
1 1×10 × 42
F
M BD = − ql 2 = − = −20kN ⋅ m
8 8
The bending moment diagram of the primary frame due to the external loading is shown in Fig.11.22 (a).
By the equilibrium conditions of the free body diagrams shown in Figs.11.22 (b) through (e), the restraint
moment F1P and restraint force F2P at the artificial restraints 1 and 2 must satisfy the following
equilibrium conditions. That is,
∑M = 0 C F1P = 0 [see Fig.11.22 (b)]
∑X =0 F
QCA + QDB
F
= F2 P [see Fig.11.22 (c)]
F
QCA =0 [see Fig.11.22 (d)]
F
M BD + 10 × 4 × 2
F
QDB =− = −15kN [see Fig.11.22 (e)]
4
Finally, we find
F2 P = −15
(a) (b) (c)

C D F2 P
F1P
F2 P F1P
D QCA QDB
C F F
C
0
10kN/m
0 QCA
F
QDB
F
C D
M P diagram 0
10kN/m
(unit: kN ⋅ m) (d) (e)

4m
4m
A B
20
QAC
F
A QBD
F B
0 QBD
F
Fig.11.22 M P and free body diagrams of example 11-3

(a) M P diagram; (b) equilibrium condition of F1P ; (c) equilibrium condition of F2 P ;
Note that the restraint force F2P is obtained by considering the force component projection
equilibrium equation of the forces acting on a portion of the primary structure, which is isolated by the cut
through the top of the columns and along the associated direction of the translation of joint D, as shown in
Fig.11.22 (c).
Substituting the numeric values of stiffness coefficients and free terms into displacement-method
equations, we write
⎧10iΔ1 − 1.5iΔ 2 = 0
⎪
⎨ 3.75
⎪⎩−1.5iΔ1 + 4 iΔ 2 − 15 = 0
Thusly, the primary unknown Δ1 and Δ 2 are solved to be

1 1
Δ1 = 3.158 Δ 2 = 21.05
i i
The positive answer implies that the real directions of joint displacements Δ1 and Δ 2 coincide with
the assumed ones.
By the superposing equation for bending moment, i.e., M = M 1Δ1 + M 2 Δ 2 + M P , the member end
moments are calculated to be
6iAC 4
M AC = 2iAC Δ1 − Δ 2 = 2 × 3.158 − × 21.05 = −25.26kN ⋅ m
4 6
6i 4
M CA = 4iAC Δ1 − AC Δ 2 = 4 × 3.158 − × 21.05 = −18.94kN ⋅ m
4 6
M CD = 3iCD Δ1 = 6 × 3.158 = 18.95kN ⋅ m
3iBD 3
M BD = − Δ 2 + M BD
F
= − × 21.05 − 20 = −35.79kN ⋅ m
4 4
superimposing method used previously in the text, as shown in Fig.11.23 (a).
18.94 D 11.05 C D 11.05 C 11.05 11.05

C D
3.16
3.16 3.16 3.16
18.95
( 20 ) 2.48
(unit: kN ⋅ m) (unit: kN) (unit: kN)
A B A B A B
25.26 (a) 35.79 (b) 28.95 (c)
11.05 3.16
3.16
(a) M diagram; (b) Q diagram; (c) N diagram
(8) Draw shear and axial force diagrams

After the member end moments are determined the member end shear forces can be obtained by
consideration of equilibrium conditions of the free body diagram of each member; the member end axial
forces can be obtained by consideration of equilibrium conditions of each joint. By doing thusly, we draw
the shear and axial force diagrams as shown in Figs.11.23 (b) and (c).
Example 11-4
Draw the bending moment and shear force diagrams for the bent frame shown in Fig.11.24 (a) by
using displacement method.
Solution
Δ1 Δ1 Δ1
P D E F P D E F
h1 I1
h2 I 2
A h3 I 3
A
B B
C C
(a) (b)


The bent frame consists of 3 columns hinged by two beams as shown in the figure. The translation will
occur at the top of the three columns under the action of the external load. However, if neglect the changes
of member lengths the only independent joint displacement is the horizontal translation of joints D, E and F,
denoted by Δ1 . So the translation will be chosen as the primary unknown of the frame.
Remove the external load and add one artificial restraint against the primary unknown at joint F to
form a primary structure. The primary system will be obtained by subjecting the external loading and the
joint displacement Δ1 to the primary structure as shown in Fig.11.24 (b).
The equilibrium equation is based on the condition that the restraint forces yielded at the artificial
restraint added at joint F due to the combined effects of the external loading and the translation Δ1 must be
equal to zero, because there is no joint force acting at the joint F for the original structure. By superposing
the combined effects of the external loading and the joint displacement Δ1 , we write the
displacement-method equation to be
k11Δ1 + F1P = 0
(4) Calculation of stiffness coefficient
In order to calculate the stiffness coefficient, we assume the linear flexural rigidity of each member of
the frame to be
EI1 EI 2 EI 3
iAD = = i1 iBE = = i2 iCF = = i3
h1 h2 h3
Referring to the table 11-1, we find the shape constants due to Δ1 = 1 as follows
3i1 3i2 3i3
M AD = − M BE = − M CF = −
h1 h2 h3
Δ1 = 1 Δ1 = 1 (b)
(a) Δ1 = 1
D E F k11 D E F k11
QDA QEB QFC
D E F
QDA QEB QFC
A
(c)
h1
3i1 3i2 B (d) (e)
h2
h3
h1 h2 3i3 C
h3 A QAD
M AD B QBE
M 1 diagram QCF
M BE C
M CF
Fig.11.25 M 1 and free body diagrams of example 11-4
(a) M 1 diagram; (b) equilibrium condition of k11 ; (c) free body diagram of member DA;
(d) free body diagram of member EB; (e) free body diagram of member FC
The bending moment diagram of the primary frame due to Δ1 = 1 is shown in Fig.11.25 (a).
By the equilibrium condition of the free body diagram shown in Fig.11.25 (b), the restraint force k11
at the artificial restraint must satisfy the following equilibrium condition. That is,
∑X =0 QDA + QEB + QFC = k11 [see Fig.11.25 (b)]
M AD 3i1
QDA = − = 2 [see Fig.11.25 (c)]
h1 h1
M BE 3i2
QEB = − = 2 [see Fig.11.25 (d)]
h2 h2
M CF 3i3
QFC = − = 2 [see Fig.11.25 (e)]
h3 h3
Thusly, we find k11 to be
i1 i2 i3
k11 = 3( + + )
h12 h22 h32
(5) Calculation of free term
Note that the concentrated force P cannot yield any internal forces in the columns of the primary
structure. So by the equilibrium condition of the free body diagram shown in Fig.11.26 (b), the restraint
force F1P at the artificial restraint must satisfy the following equilibrium condition. That is,
F
QDA + QEB
F
+ QFC
F
− P = F1P
Because of
F
QDA =0 [see Fig.11.26 (c)]
F
QEB =0 [see Fig.11.26 (d)]
F
QFC =0 [see Fig.11.26 (e)]
Thusly, we find
F1P = − P
(a) (b)
P D F F1P P D F F1P
E QDA
F E QEB
F
QFC
F
D QDA
F
E QEB
F F QFC
F
A (c) (d) (e)

h1
B
h2
h3
C QAD
F
A
B QBE
F
0 C QCF
F
0
0
Fig.11.26 M P and free body diagrams of example 11-4
(a) M P diagram; (b) equilibrium condition of F1P ; (c) free body diagram of member DA;
(d) free body diagram of member EB; (e) free body diagram of member FC

Substituting the numeric values of stiffness coefficient and free term into the displacement-method
equation, we write
i1 i2 i3
3( + + ) Δ1 − P = 0
h12 h22 h32
Thusly, the primary unknown Δ1 is solved to be
P P
Δ1 = = 3
i i i i
3( 12 + 22 + 32 ) 3∑ m2
h1 h2 h3 m =1 hm

By the superposing equation for bending moment, i.e., M = M 1Δ1 + M P , the member end moments
are calculated to be
i1 i2 i3
P P P
3i h 3i h 3i h
M AD = − 1 Δ1 = − 3 1 M BE = − 2 Δ1 = − 3 2 M CF = − 3 Δ1 = − 3 3
im im im
∑ ∑ ∑
h1 h2 h3
2 2 2
m =1 hm m =1 hm m =1 hm
superimposing method as shown in Fig.11.27 (a).
(8) Draw shear force diagram
After the member end moments are determined the member end shear forces can be obtained by
consideration of equilibrium conditions of the free body diagram of each member. Thusly, we calculate
i1 i1 i1
P P P
h2 h2 h2
QAD = QDA =− 3 1 QBE = QEB =− 3 2 QCF = QFC =− 3 3
i i i
3∑ m2 3∑ m2 3∑ m2
m =1 hm m =1 hm m =1 hm
The shape of the shear force diagram is shown in Fig.11.27 (b).
P D E F
D E F
(a)
(b)
A A
M AD B
M BE B
C C
M CF
(a) M diagram; (b) Q shape of diagram
11.7 Analysis of Statically Indeterminate Symmetric Structures
It is observed from last section that when employing displacement method to analyze a statically
indeterminate structure, the higher the degrees of its indeterminacy, the more the required computational
effort. The main calculation work is the solution of simultaneous linear equations, i.e., the evaluation of
stiffness coefficients and free terms in equilibrium conditions and the solution of the simultaneous linear
equations. However, many structures, because of aesthetic and/or functional requirements, are arranged in
symmetric forms. If a symmetric structure is linearly elastic, the response (i.e., member forces and
deformations) of the entire structure under any general loading can be obtained from the response of one of
its portions separated by the axes of symmetry. Thus only a portion (usually half) of the symmetric
structure needs to be analyzed. In the section, we discuss how to utilize symmetric property of a structure to
simplify the computational effort required in the analysis of the structure.
As discussed in chapter 10, any one of asymmetric loadings may be decomposed into the
superposition of one symmetric component and one antisymmetric component. When a symmetric structure
is subjected to a symmetric loading the response characteristics of the symmetric primary structure due to
the external loading is symmetric as well. Similarly, when a symmetric structure is subjected to an
antisymmetric loading, the response characteristics of the symmetric primary structure due to the external
loading are also antisymmetric. Thus, to determine the response (i.e., member forces and deformations) of
the entire structure, we need to analyze only half the structure, on either side of the axis of symmetry, with
symmetric or antisymmetric boundary conditions at the axis. The response of the remaining half of the
structure can then be obtained by reflection. Following examples will illustrate the procedure of the
analysis.
Example 11-5
Draw the bending moment diagram for the symmetric rigid frame shown in Fig.11.28 (a) by using
Solution

Although the frame has 2 rigid-joint rotations and 1 joint translation, it is a symmetric structure under
symmetric loading. So only half of the structure shown in Fig.11.28 (b) may be selected to analyze. With
the similar analysis discussed in last section, the rotation of joint C of the half structure, denoted by Δ1 ,
will be chosen as the primary unknown of the frame.
Remove the uniformly distributed external load and add one artificial restraint against the primary
unknown (the rotation of joint C) at joint C of the half structure to form a primary structure. The primary
system will be obtained by subjecting the external loading and the joint displacement Δ1 to the primary
structure as shown in Fig.11.28 (c).
The displacement-method equation is based on the requirement that the restraint moment yielded at
the artificial restraint added at joint C due to the combined effect of the external loading and the rotation of
joint C must be equal to zero, because there is no joint moment acting at the joint C for the original
structure. Thusly, by superposing the combined effects of the external loading and the joint displacement
Δ1 , we write the displacement-method equation to be
k11Δ1 + F1P = 0
6kN/m 6kN/m 6kN/m
C C
C 3EI D 3EI E Δ1 3EI
E
EI EI
4m
4m
EI EI
4m
A B A A
6m 3m 3m
(a) (b) (c)

(a) original structure; (b) half of the original structure; (c) primary system
(4) Calculation of stiffness coefficient k11

By referring to the results listed in table 11-1, the bending moment diagram of the primary frame due
to Δ1 = 1 is shown in Fig.11.29 (a). Then, the stiffness coefficient k11 can be determined by the moment
equilibrium condition of the free body diagram shown in Fig.11.29 (b) to be
3EI 4 EI
k11 = iCE + 4iCA = + = 2i
3 4
In which, assume EI = i .
(5) Calculation of free term F1P
write
1 1× 6 × 32
F
M CE = − ql 2 = − = −18kN ⋅ m
3 3
1 1× 6 × 32
F
M EC = − ql 2 = − = −9kN ⋅ m
6 3
(c) 6kN/m
(a) (b) (d)
k11 18
k11 F1P
4iCA C F1P
E C E
iCE C C
iCE
iCE Δ1 = 1 4iCA
9 18
Δ1 = 1
(unit: kN ⋅ m)
A
2iCA
A
Fig.11.29 M 1 and M P of half structure
(a) M 1 diagram; (b) equilibrium condition for k11 ; (c) M P diagram; (d) equilibrium condition for F1P
The bending moment diagram of the primary frame due to the external loading is shown in Fig.11.29
(c). By the moment equilibrium condition of the free body diagram shown in Fig.11.29 (d), we find
F1P = −18kN ⋅ m
equation, we obtain
9
2iΔ1 − 18 = 0 Δ1 =
i 9 ( 27 ) 9
D
(7) Draw bending moment diagram 9 C 9
By the superposing equation for bending
moment, i.e., M = M 1Δ1 + M P , the bending 18
moment diagram for the half of the structure can (unit: kN ⋅ m)
A 4.5 4.5 B
be obtained and another half of the bending
moment diagram can be constructed by reflection, Fig.11.30 M diagram of example 11-5
as shown in Fig.11.30.
Example 11-6
Draw the bending moment diagram for the symmetric rigid frame shown in Fig.11.31 (a) by
Solution
y
(a) q (b) (c)
q q
D
A B A A
D EI D Δ1 EI
EI
EI
a
EI
a
C EI EI C C
x x
a a
a
EI y

q (a) original structure; (b) one quarter of the structure;
a a (c) primary system

Although the closed rigid frame has 4 rigid-joint rotations, it is a symmetric structure with two axes of
symmetry, indicated by x-x and y-y, under symmetric loading. Therefore, only one quarter of the structure
shown in Fig.11.31 (b) may be selected to analyze. The rotation of joint A of the quarter of the structure,
denoted by Δ1 , will be chosen as the primary unknown of the closed frame.
Remove the uniformly distributed external load and add one artificial restraint against the primary
unknown at joint A of the quarter structure to form a primary structure. The primary system will be
obtained by subjecting the external loading and the joint displacement Δ1 to the primary structure as
shown in Fig.11.31 (c).
The displacement-method equation is based on the requirement that the restraint moment yielded at
the artificial restraint added at joint A due to the combined effects of the external loading and the rotation of
joint A must be equal to zero, because there is no joint moment acting at joint A of the original structure.
Thusly, by superposing the combined effects of the external loading and the joint displacement Δ1 , we
write the displacement-method equations to be

k11Δ1 + F1P = 0
(4) Calculation of stiffness coefficient k11
By referring to the computing results listed in table 11-1, the bending moment diagram of the primary
frame due to Δ1 = 1 can be determined as shown in Fig.11.32 (a). Thus, the stiffness coefficient k11 can be
determined by the moment equilibrium condition of the free body diagram shown in Fig.11.32 (b) to be
k11 = 2i
Where i = EI / a .
(a) (b) (c) q (d)

qa 2
k11 k11 3 F1P
D
A D
A i A qa 2
i F1P A 3
i i qa 2
Δ1 = 1 i
6
i C C
Fig.11.32 M 1 and M P of one quarter structure

(a) M 1 diagram; (b) equilibrium condition for k11 ; (c) M P diagram; (d) equilibrium condition for F1 P
(5) Calculation of free term F1P

write
1 1
F
M AD = − qa 2 F
M DA = − qa 2
3 6
The bending moment diagram of the primary structure due to the external loading is shown in Fig.11.32 (c).
By the moment equilibrium condition of the free body diagram shown in Fig.11.32 (d), we find
1
F1P = − qa 2
3
equation, we obtain
1 qa 2
2iΔ1 − qa 2 = 0 Δ1 =
3 6i
By the superposing equation for bending moment, i.e., M = M 1Δ1 + M P , the bending moment
diagram for the quarter of the structure can be obtained and other portion of the bending moment diagram
can be constructed by reflections, as shown in Fig.11.33.
qa 2 qa 2
6 6
qa 2
3
qa 2
3
qa 2 qa 2
6 6
Fig.11.30 M diagram of example 11-6
11.8 Development of Displacement-Method Equations by Direct Stiffness Method
In this section, we present an alternative approach to develop the displacement-method equations of a

statically indeterminate structure by direct stiffness method. So called direct stiffness method means that
when establish the displacement-method equations of a structure, the primary system used in previous
section is not needed. The displacement-method equations may be directly developed by considering the
equilibrium conditions of joints or a portion of the structure in terms of slope-deflection equations of
associated members. The slope-deflection equations relate the moments at the ends of a member to the
rotations and displacement of its ends and the external loading applied to the member.
11.8.1 Slope-deflection equations of prismatic members
In section 11.2, we have already discussed the shape constants and loading constants for various single
prismatic members. That is, the member end forces of various single-span indeterminate prismatic beams
separately due to their end displacements and external loadings. If the end forces of a member are caused
by the combined effects of external loads and its end displacements, they may be determined by using
principle of superposition, i.e., by superimposing corresponding loading and shape constants together,
which have been tabulated in table11-1 and table 11-2 in section 11.2. The relationship between the end
moments and the end displacements and the external loading applied on the member is named
slope-deflection equation of the member, which is the basis of the section. Thusly doing, we can obtain the
slope-deflection equations for three kinds of prismatic members as follows.
(1) Member with two ends fixed (or rigidly connected)
Focus our attention on an arbitrary prismatic member AB with the two ends fixed as shown in
Fig.11.31. When the member is subjected to external load and support settlements it deforms, as shown in
the figure, and internal moments are induced at its ends.
M AB P M AB P
θA EI θA EI
A ϕ B
θA θA
Δ
Δ
QAB QAB
θB M BA
l l
QBA
QBA
Fig.11.32 A member with one end
Fig.11.31 A two fixed-end member fixed and another end hinged
Note that all the moments and rotations are shown in the positive sense in the figure.
The slope-deflection equations can be derived by relating the member end moments to the end
rotations and chord rotation by applying the principle of superposition to the computing results tabulated in
table 11-1 and table 11-2. Thusly, we write
Δ F ⎫
M AB = 4iABθ A + 2iABθ B − 6iAB + M AB ⎪⎪
l
⎬ (11-10)
Δ F ⎪
M BA = 2iABθ A + 4iABθ B − 6iAB + M BA
l ⎪⎭
(2) Member with one end fixed (or rigidly connected) and another end hinged or rolled
Consider an arbitrary prismatic member AB with one end fixed (or rigidly connected) and another end
hinged or rolled as shown in Fig.11.32. When the member is subjected to external load and support
settlements it deforms as shown in the figure, and internal moments are induced at its fixed end. Its
slope-deflection equations can be derived by relating the member end moments to the end rotation θ A and
chord rotation Δ / l by applying the principle of superposition to the computing results tabulated in table
11-1 and table 11-2. Therefore, we write
Δ F ⎫
M AB = 3iABθ A − 3iAB + M AB ⎪
l ⎬ (11-11)
M BA = 0 ⎪⎭
(3) Member with one end fixed (or rigidly connected) and another end fixed by double links
Imagine an arbitrary prismatic member AB with one end fixed (or rigidly connected) and another end
fixed by double links as shown in Fig.11.33. When the member is subjected to external load and support
settlements it deforms as shown in the figure, and internal moments are induced at its ends. Its
slope-deflection equations can be derived by relating the member end moments to the end rotation θ A by
applying the principle of superposition to the computing results tabulated in table 11-1 and table 11-2.
Consequently, we write
M AB = iABθ A + M AB
F
⎪⎫
F ⎬
(11-12)
M BA = −iABθ A + M BA ⎪⎭
M AB P
θA EI B
A
θA
QAB
l
M BA
QBA
Fig.11.33 A member with one end fixed
and another end fixed by two links
11.8.2 Analysis of statically indeterminate structures by direct stiffness method
Actually, the displacement-method equations of a statically indeterminate structure are the equilibrium
equations. A moment equilibrium equation corresponding to a joint rotation can be developed by the
consideration that the algebraic sum of the end moments of the members connected to the joint must be
equal to the joint moment acting on the original structure. A force equilibrium equation in the direction of
the joint translation can be established by the consideration that the algebraic sum of the end force
components of the members associated with the joint translation in the direction must be equal to the
algebraic sum of the force components acting on the portion, associated with the joint translation, of the
original structure in the same direction. When the primary system is used to establish the
displacement-equilibrium equations, the stiffness coefficients and free terms in the equations can be
conveniently obtained by the tabulated results of various single-span indeterminate beams, see tables 11-1
and 11-2.
Alternatively, the displacement-method equilibrium equations may be developed directly by using the
slope-deflection equations of the members of the structure. The following examples will illustrate the
employment of slope-deflection equations when analyzing the structure by displacement method or
slope-deflection method.
Example 11-7
Draw the bending moment diagram for the rigid frame shown in Fig.11.34 (a) by direct stiffness
method.
Solution
(a) (b) (c)

C D
C C M CD
3EI D
QCA QDB
10kN/m
M CA M CA
4m
EI EI QDB
C QCA D
10kN/m
A B
(d) (e)
6m
A QAC B
M BD QBD
M AC
(a) original structure; (b) free body diagram of joint C ; (c) free body diagram of the portion
around beam CD; (d) free body diagram of column AC ; (e) free body diagram of column BD

From Fig.34 (a), we can see the rigid-joint rotation is the rotation of joint C and independent joint
translation is the horizontal translation of joint C or D of the frame if the changes of the member lengths are
neglected. So the rotation of joint C of the structure, denoted by Δ1 , and horizontal translation of joint C
or D, denoted by Δ 2 , will be chosen as the primary unknowns of the frame.
(2) Fixed-end moments
By using the fixed-end moment expressions given in table 11-2, we evaluate the fixed-end moments
due to the external loading for each member:
ql 2 10 × 42
F
M BD =− =− = −20 (kN ⋅ m) F
M AC = M CA
F
= M CD
F
=0
8 8
Note that in accordance with the slope-deflection sign convention, the clockwise fixed-end moments are
considered as positive, whereas the counterclockwise fixed-end moments are considered to be negative.
(3) Chord rotations
Since the rotation of joint C and the translation of joints C and D are selected to be the primary
unknowns, the chord rotation for each member to be:
Δ2
ϕ AC = ϕ BD = = 0.25Δ 2 ϕCD = 0
4
(4) Slope-deflection equations
To relate the member end moments to the primary unknowns, we write the slope-deflection equations
for the three members of the frame by applying Eqs. (11-10) and (11-11). Thus the slope-deflection
equations for the members of the frame can be expressed as
Δ2 3iΔ 2 ⎫
M CA = 4iCAθ1 − 6iCA + M CA
F
= 4iΔ1 −
lCA 2 ⎪
⎪
Δ2 3iΔ 2 ⎪
= 2iCAθ1 − 6iCA + M AC = 2iΔ1 −
F
2 ⎪⎬
M AC
lCA (a)
M CD = 3iCD Δ1 + M CDF
= 3(2i )Δ1 = 6iΔ1 ⎪
⎪
3iBD 3i ⎪
M BD =− Δ 2 + M BD = − Δ 2 − 20
F
⎪
lBD 4 ⎭
EI 3EI
In which, we assume iCA = iBD = = i , iCD = = 2i .
4 6
(5) Equilibrium equations
The free-body diagram of joint C is shown in Fig.11.34 (b). Note that the member end moments,
which are assumed to be in a clockwise direction on the ends of the members, must be applied in the
(opposite) counterclockwise direction on the free body of the joint, in accordance with Newton’s third law.
By applying the moment equilibrium equation ∑ M C = 0 to the free body of joint C, we obtain the
equilibrium equation
M CD + M CA = 0 (b)
To establish the second equilibrium equation, we apply the force equilibrium equation ∑X =0 to
the free body of the portion, associated with Δ 2 , of the frame shown in Fig.11.34 (c), to obtain
QCA + QDB = 0 (b)
In which, QCA and QDB represent the shears at the upper ends of columns AC and BD, respectively, as
shown in the Fig.11.34 (c). To express the column end shears in terms of column end moments, we draw
the free-body diagrams of the two columns [Figs.11.34 (d) and (e)] and sum the moments about the bottom
Summary 421
of each column:
M AC + M CA ⎫
∑M A =0 QCA = −
l AC ⎪
⎪
⎬ (c)
M BD 1
∑M =0 Q DB =− − qlBD ⎪
B
lBD 2 ⎪⎭
By substituting Eq. (c) into Eq. (b), we obtain
M AC + M CA + M BD + 80 = 0 (e)
By substituting the end moment expressions in Eq. (a) into Eq. (b) and Eq. (e), we write
3 ⎫
10iΔ1 − iΔ 2 = 0 ⎪
2 ⎬ (f)
6iΔ1 − 3.75iΔ 2 +60 = 0 ⎪⎭
(6) Solution of primary unknowns
Solving simultaneous Eq. (f) yields
1 1
Δ1 = 3.16 Δ 2 = 21.05
i i
(7) Member end moments
By substituting the expressions of Δ 1 and Δ 2 into the slope-deflection equations expressed in Eq. (a),
we obtain
M CA = −18.94kN ⋅ m M AC = −25.26kN ⋅ m
M CD = 18.95kN ⋅ m M BD = −35.79kN ⋅ m
(8) Internal force diagrams
The bending moment and shear and axial force diagrams can now be constructed by using the methods
used previously in the text. These diagrams have already been shown in Fig.11.23.
It is observed from above progress that the principle for developing the equilibrium equations by
slope-deflection equations is the same as that for establishing displacement-method equations by primary
system. The only distinguish is the presenting form of the equations. In fact, the slope-deflection equations
have already reflected the consideration of the combined effects of external loadings and primary unknown
displacements, the same consideration as applied to the primary system.
SUMMARY
(1) The displacement method is another approach that can be used for analyzing indeterminate
structures (although it may be also used for analyzing determinate structures). The method is especially
suitable for analyzing continuous beams and rigid frames with high degrees of indeterminacy. Also, an
understanding of the fundamentals of this method provides a valuable introduction to the moment
distribution method (including no-shear distribution method) and the matrix stiffness method, which forms
the basis of most computer software currently used for structural analysis.
(2) The primary unknowns (or variables) used in displacement method are joint displacements of a
structure, i.e., rigid joint rotations and independent joint translations of the structure. A clear understanding
of the mechanical meanings of shape and loading constants of various single-span indeterminate beams
helps readers to realize why rigid joint rotations and independent joint translations rather than any
displacements (such as rotations of hinged joints) of a structure are selected to be the primary unknowns.
Note that the sign convention for member end forces and displacements will also facilitate the analysis of a
structure by displacement method.
(3) In displacement method, the displacement-method equations used to solve primary unknowns
are the equilibrium equations. For a rigid-joint rotation, a corresponding moment equilibrium equation can
be written; whereas a force equilibrium equation can be obtained for an independent joint translation. The
number of equilibrium equations is just the same as that of primary unknowns.
(4) The development of displacement-method equations by employing primary system, which will
relate the analyzing procedure of the displacement method to that of the force method discussed in last
chapter, is helpful to further understand the two fundamental methods. The utilization of primary systems
not only makes each item in displacement-method equations have a definite meaning but also echoes the
matrix stiffness method, which is plan to discuss in Chapter 13.
(5) The analysis of symmetric structures by displacement method is the analysis of corresponding
half structures. The keynote of the analysis is the choice of the half of the structures. The understanding of
response characteristics of a symmetric structure under symmetric and antisymmetric loadings will help the
choice of a half of the structure.
(6) The direct stiffness method is based on the slope-deflection equation:
2 EI
M nf = (2θ n + θ f − 3ϕ ) + M nfF (11-13)
l
The equation relates the moments at the ends of a member to the rotations and displacements of its
ends and the external loads applied to the member.
The procedure of the method essentially involves (1) identifying the unknown joint displacements
(degrees of freedom for independent joint displacements) of the structure; (2) for each member, writing
slope-deflection equations relating member end moments to the unknown joint displacements; (3)
establishing the equations of equilibrium of the structure in terms of member end moments; (4)substituting
the slope-deflection equations into the equilibrium equations and solving the resulting system of equations
to determine the unknown joint displacements; and (5) computing member end moments by substituting the
values of joint displacements back into the slope-deflection equations. Once member end moments have
been evaluated, member end shears and axial forces, and support reactions, can be determined through
equilibrium considerations.
11-1 What is the sign convention about the internal forces and displacements of a member ends in
displacement method?
11-2 What are the shape and loading constants for a single-span indeterminate beam?
11-3 Please derive equations (11-1) through (11-5) and keep the form of equation (11-1) in your mind.
11-4 Please calculate the loading constants of the single-span indeterminate beams numbered by 1, 3, 6,
8, 12 and 14 in table 11-2 and keep the constants in your mind.
11-5 What are the magnitudes of end moments M AB and M BA due to θ A = 1 for the beams shown
in the figure of the problem?
11-6 What are the magnitudes of end moments M AB and M BA due to Δ = 1 for the beams shown
in the figure of the problem?
MBA
MAB MAB
EI B
A EI
A B
θA =1 θA =1
l l
(a) one end hinged and one end fixed (b) two ends hinged
MBA MAB MBA

MAB
B A EI B
EI
A
Δ =1
Δ =1
l l
(a) one end hinged and one end fixed (b) two ends hinged

11-7 How many kinds of primary unknowns are there for displacement method?
11-8 How to determine joint rotation? What is the basis of the determination of joint rotations?
11-9 How to determine the number of independent joint translations? What is the basic assumption for
determining the independent translations? Why the determination of the number of degrees of
freedom of a hinged system can be applied to the determination of the number of the independent
joint translations of a structure?
11-10 How to select primary structure and primary system when using displacement method to analyze a
statically indeterminate structure? What is the difference between the consideration of selecting
primary structure and primary system used in force method and that used in displacement method?
As we have known in force method, there are several kinds of primary systems for a structure.
Whether or not more than one primary system for a structure can be used in displacement method?
11-11 What is the individual function of the artificial rotative restraints and artificial translating
restraints?
11-12 What are the meanings of kii , kij and FiP occurring in displacement-method equations?
11-13 Why can we say that the displacement-method equations are actually the equations for
equilibrium?
11-14 Whether or not displacement method can be used to analyze a statically determinate structure?
11-15 Why can the relative stiffness values of the members of a structure determine the internal forces?
Whether or not the solutions of primary unknowns Δ1 , Δ1 , " Δ n are their real values if the
relative stiffness values of the members of a structure are used?
11-16 Why can the rotations of a hinged joint not be used as primary unknowns?
11-17 Whether or not the structure shown in the figure of
the problem can be analyzed as the approach used in C D
example 11-4?
11-18 Why can only a half of the structure be selected to
P I1h1
analyze for a symmetric structure under symmetric I2h2
or antisymmetric loadings? Whether or not just a
half of the structure can be selected to analyze for a A B
symmetric structure under asymmetric loadings?

11-19 Whether or not the symmetric property of a structure
can be used to simplify the analysis of the structure
if the original structure instead of a half of the structure is selected as the primary structure?
11-20 Why is the direct stiffness method actually the same as the method in which the primary system is
employed?
11-21 Please compare the direct stiffness method with primary-system method.
11-22 Please compare the displacement method that uses primary system with force method.
11-1 Determine the number of primary unknowns used in displacement method for the statically
indeterminate structures shown in the figures, and lay off the corresponding primary structures.
F
(a) (b) F G H
E
D
D E
A B C A B C
C
(c) J
(d)
I α α K
B α
E F G H
A A B C D
(1) α ≠ 0 (2) α = 0 (1) α ≠ 0 (2) α = 0
(e) E C F (f)
E C F
EA
A B A B
D D
(1) EA ≠ ∞ (2) EA = ∞
problem 10-1 (contd)

(g) (h)
E
D E
C D
EA F G
A B
A B C
(1) EA ≠ ∞ (2) EA = ∞
(i)
D EI = ∞ E
F G
A B C
Problem 11-1
11-2 Lay off the corresponding primary system for each of the following structures shown in the figures
and also construct the bending moment diagrams due to the unit value of each primary unknown
and due to the external loadings.
ql q 2.5kN/m P = 10kN
B A
i C D EI B EI D
i
4m
i EI
EI = constant
A C
l l l 4m 4m
2 2
(a) (b)
2
l
E ql
l =4 q
2
l
2
B
l
MO A B
l =1 C l =2 D C
2
l
ql
l =1
l EI = constant
2
l
A D E
l l l l
(c) (d)
Problem 11-2
11-3 Draw the bending moment diagrams for the following continuous beams and rigid frames by
displacement method. Note that EI = constant .
(a) (b)
40kN 10kN/m 10kN/m 50kN 10kN
A B C A B C
D
D
3m 3m 6m 6m 6m 3m 3m 2m
(c) (d)
P 2.4kN/m
A B C D E
B
C
15m
l
l l E F A D
2 2 l l 20m 15m
problem 11-3
11-4 Lay off the corresponding primary system for each of the following two structures shown in the
figures and also construct the bending moment diagrams due to the unit value of each primary
unknown and due to the external loadings.
q q
E
F D
E
l
l
ql
B D
B C
C
l
l
A l l
A
l 2 2
l
(a) (b)
problem 11-4
11-5 Draw bending moment diagrams for each of the following structures shown in the figures by
11-6 Analyze the following bent frames shown in the figures by displacement method and draw their
bending moment, shear and axial force diagrams.
11-7 Draw bending moment diagrams for each of the following structures shown in the figures by using
the properties of the structures and by displacement method
11-8 Draw bending moment diagrams for each of the following structures shown in the figures by using
the properties of the structures and by displacement method.
11-9 Discuss how to use the property of the rigid frame to analyze it.
11-10 Analyze the structures shown in the figures of problem 11-3 (c) and (d) and in the figures of
problem 11-5 (c) and (d) by direct stiffness method.
F
D E F E
50kN D
2I 2I
q
7m
I I I
l
EI = constant
5m
A B C A B C
l l 6m 6m
(a) (b)
G H
E EI b = ∞
4m
EI EI
4m
EI = constant
10kN / m
D E F
8kN
EI b = ∞ EI b = ∞
C D
4m
2 EI
4m
2 EI 2 EI
A B C
A B
6m 6m
6m
(c) (d)
problem 11-5
(a) (b) 40kN/m

C EA → ∞ D
q = 20kN/m
C i=6 D
6m
2m
EI EI
A B P = 20kN
i=4 i=3
12m
2m
A B
4m
problem 11-6
q
20kN/m
D E F
D E C F G
l
4m
EI = constant
A B C A B
l l 6m 6m 6m 6m
problem 11-7 (a) problem 11-7 (b)
40kN/m
80kN/m
40kN/m 40kN/m E F
3I
C E D
4m
I I
EI = constant
3m
C
3I D
A F B
4m
I I
40kN/m 40kN/m
A B
80kN/m 6m
3m 3m
problem 11-8 (a)
problem 11-7 (c)
50kN C D
P
EI E C F
I I
12m
EI EI I 2I I
A I I B
A B D
10m l l
problem 11-8 (b) problem 11-9

CHAPTER 12
METHOD OF SUCCESSIVE APPROXIMATIONS AND
INFLUENCE LINES FOR INDETERMINATE STRUCTURES

In this chapter, we will study two methods of successive approximations, which are all based
on the principle of the general displacement (stiffness) method discussed in last chapter, called
moment distribution method and no-shear distribution method, respectively.
As it will be seen in the later content of the chapter, that when analyzing a structure by the
methods of moment distribution or no-shear distribution, a basic process of calculation, in which
each joint of the structure is first locked and then unlocked, is repeated in succession. The process
includes three steps: (1) evaluate the fixed-end moments at the ends of each member and restraint
moments at each joint in terms of external loadings; (2) determine distribution moments by
distribution factor; (3) calculate carryover moments by carryover factor.
The moment distribution method is the keynote of the chapter. In addition, shear distribution
method, the construction of influence lines and envelop of internal forces for statically
indeterminate continuous beams will be presented as well.
12.1 General Remarks on Methods of Successive Approximations
In this chapter we will learn another classical formulation of the displacement method, the moment
distribution method. In fact, the moment distribution method is classified as a displacement method, and from
a theoretical point of view, it is very similar to the displacement method.
Like the displacement method, the moment distribution method can be employed only for the analysis of
continuous beams and rigid frames, taking into account their bending deformations only.
Recalling from previous two chapters, force and displacement methods, a sort of simultaneous
equations associated with the equilibrium and compatibility conditions as well as member force and
displacement relations must be developed and solved when analyze an indeterminate structure. However, in
the moment distribution method the moment equilibrium equations of the joints are solved iteratively by
successively considering the moment equilibrium at one joint at a time, while the remaining joints of the
structure are assumed to be restrained against displacement. As it will be explained in detail later, that
accuracy of the successive approximations may be carried out to any desired degree.
The moment distribution method, which was initially developed by Hardy Cross in 1924, was the most
431
12.4 Moment Distributions at Multi-Joints—Successive Approximations 447
Example 12-4
Determine the member-end moments by using the moment distribution method and draw the bending
moment diagram for the continuous beam shown in Fig.12.8 (a). The relative flexural rigidity is shown in
the figure too. Assume EI = 1 .
Solution
It is usually convenient to carry out the moment distribution analysis in a tabular form, as shown in
Fig.12.8 (b). Note that the table, which is sometimes known as a moment-distribution table, consists of six
columns, one for each member end of the beam. All the computations for a particular member end moment
are recorded in the column for that member end.
(1) Bending stiffnesses and distribution factors at each rotatable joint
The first step in the analysis is to calculate the distribution factors at those joints of the structure that
are free to rotate. As discussed in Section 12.2 [Eq. (12-6)], the distribution factor for an end of a member
is equal to the bending stiffness of the member divided by the sum of the bending stiffnesses of all the
members connected to the joint. From Fig. 12.8(a), we can see that only joints B and C of the continuous
beam are free to rotate. Their bending stiffnesses and distribution factors can be calculated as follows.
Bending stiffnesses and distribution factors at joint B:
1
S BA = 4iBA = 4 × = 0.667
6
1.5
S BC = 4iBC = 4 × =1
6
Thus,
S BA 0.667 ⎫
μ BA = = = 0.4 ⎪
∑j SBj 0.667 + 1 ⎪⎪
S 1 ⎬ ∑μ Bj = 0.4 + 0.6 = 1
μ BC = BC = + 0.6 ⎪ j
∑j SBj 0.667 + 1 ⎪⎪
⎭
Bending stiffnesses and distribution factors at joint C:
1.5
SCB = 4iCB = 4 × =1
6
2
SCD = 3iCD = 3 × =1
6
Consequently,
SCB 1 ⎫
μCB = = = 0.5⎪
∑j SCj 1 + 1 ⎪⎪
S 1
⎬ ∑μ Cj = 0.5 + 0.5 = 1
μCD = CD = + 0.5 ⎪ j
∑j SCj 1 + 1 ⎪⎪
⎭
The distribution factors are written above their corresponding ends within a square as shown in Fig.12.8
(b).
(2) Fixed end moments
Next, by assuming that joints B and C are locked, we calculate the fixed end moments that develop at
the ends of each member. By using table 11-2, we obtain
1 1× 80 × 6
F
M BC = − Pl = − = −60kN ⋅ m
8 8
1
F
M CB = + Pl = +60kN ⋅ m
8
1 1× 20 × 62
F
M CD = − ql 2 = − = −90kN ⋅ m
8 8
The fixed end moments are recorded on the first line of the moment-distribution table as shown in
Fig.12.8 (b).
(3) Unlocking joint B
Since joints B and C are actually not clamped, we release them, one at a time. We can release either
joint B or joint C; let us begin at joint B because it has the maximum unbalanced moment. From Fig. 12.8
(b), we can see that there is a −60kN ⋅ m (counterclockwise) fixed-end moment at end B of member BC,
whereas no moment exists at end B of member AB. As long as joint B is restrained against rotation by the
clamp, the −60kN ⋅ m unbalanced moment is yielded at the clamp. However, when the imaginary clamp
is removed to release the joint, the negative of the unbalanced moment, 60kN ⋅ m , will equivalently act at
the joint, causing it to rotate in the clockwise direction until it is in equilibrium. The rotation of joint B
causes the distributed moments to develop at ends B of members BC and AB, DM BC and DM BA which
can be evaluated by multiplying the negative of the unbalanced moment(i.e., 60kN ⋅ m ) by the
distribution factors, respectively. Thus
DM BC = 0.6 × 60 = 36kN ⋅ m
DM BA = 0.4 × 60 = 24kN ⋅ m
These distributed moments are recorded on line 2 of the moment-distribution table [Fig. 12.8 (b)], and
they are underlined to indicate that joint B is now balanced. Note that the sum of the three moments above
the line at joint B is equal to zero (i.e., 36 + 24 − 60 = 0 ).
The distributed moment at end B of member BC induces a carryover moment at the farend C, which
can be determined by multiplying the distributed moment by the carryover factor of the member. Since
joint C remains clamped, the carryover factor of member BC is 1/2. Thus, the carryover moment at the end
C of member BC is
1
× 36 = 18kN ⋅ m
2
Similarly, the carryover moment at the end A of member AB is computed as
1
× 24 = 12kN ⋅ m
2
These carryover moments are recorded on the same line of the moment distribution table as the
distributed moments, with a horizontal arrow from each distributed moment to its carryover moment, as
shown in Fig. 12.8 (b).
It can be seen from this table that joint B is now in equilibrium, because it is subjected to two equal,
but opposite, moments. Joint C, however, is not in equilibrium, and it needs to be balanced. Before we
release joint C, an imaginary clamp must be applied to joint B in its rotated position.
(4) Unlocking Joint C
Joint C is now released. The unbalanced moment at this joint is obtained by summing all the moments
acting at the ends C of members BC and CD. From the moment-distribution table (lines 1 and 2), we can
see that there is a +60kN ⋅ m fixed-end moment and a +18kN ⋅ m carryover moment at end C of
member BC, whereas the end C of member CD is subjected to a −90kN ⋅ m fixed-end moment. Thus the
unbalanced moment at joint C is
60 + 18 − 90 = −12kN ⋅ m
When the imaginary clamp is removed to release the joint, the negative of the unbalanced moment,
12kN ⋅ m , will equivalently act at the joint, causing it to rotate in the clockwise direction until it is in
equilibrium. The rotation of joint C induces the distributed moments to develop at ends C of members BC
and CD, DM CB and DM CD , which can, as discussed previously, be evaluated by multiplying the
negative of the unbalanced moment(i.e., 12kN ⋅ m ) by the distribution factors, respectively. Thus
DM CB = 0.5 ×12 = 6kN ⋅ m

DM CD = 0.5 ×12 = 6kN ⋅ m
(a)
80kN 20kN/m
A EI B C D
1.5EI 2EI
6m 3m 3m 6m
(b)
BA BC CB CD
distribution factors A 0.4 0.6 0.5 0.5
D
B C
fixed end moments -60 +60 -90 0
⎧ unlocking B 12 ← 24 36 → 18
1 ⎨
⎩ unlocking C 3 ← 6 6
⎧ unlocking B -0.6 ← -1.2 − 1.8 → -0.9
2 ⎨
⎩ unlocking C 0.23 ← 0.45 0.45
⎧ unlocking B -0.05 ← -0.09 −0.14 → -0.07
3 ⎨
⎩ unlocking C 0.04 0.03
member end moments 11.35 22.71 -22.71 83.52 -83.52 0
83.52
(120 ) ( 90 )
(c) 2.304m
22.71
A D
B C
11.35 48.24 53.08
66.89
(unit: kN ⋅ m)
Fig.12.8 Figure of example 12-4

(a) original structure and its loads; (b) moment distribution and carryover;
(c) bending moment diagram
These distributed moments are recorded on line 3 of the moment-distribution table, and a line is drawn
beneath them to indicate that joint C is now balanced. One-half of the distributed moment are then carried
over to the farend B of member CB, as indicated by a horizontal arrow on line 3 of the table. Joint C is then
reclamped in its rotated position.
(5) Again unlocking joint B
With joint C now balanced, we can see from the moment-distribution table (line 3) that, due to the
carryover effect, there is a 3kN ⋅ m unbalanced moment at joint B. Recall that the moments above the
horizontal line at joint B were balanced previously. Thus we release joint B again and distribute the
unbalanced moment to ends B of members BC and AB as
DM BC = −0.6 × 3 = −1.8kN ⋅ m
DM BA = −0.4 × 3 = −1.2kN ⋅ m
These distributed moments are recorded on line 4 of the moment-distribution table, and one-half of these
moments are carried over to the ends C and A of members BC and AB, respectively, as indicated on the table.
Joint B is then reclamped.
(6) Again unlocking joint C
The −0.9kN ⋅ m unbalanced moment at joint C (line 4 of the moment-distribution table) is balanced
in a similar manner. The distributed and the carryover moments thus computed are shown on line 5 of the
table. Joint C is then reclamped.
(7) Thirdly unlocking joint B and C
It can be seen from line 5 of the moment-distribution table that the unbalanced moment at joint B has
now been reduced to only 0.23kN ⋅ m . Another balancing of joint B produces an even smaller unbalanced
moment of −0.07kN ⋅ m at joint C, as shown on line 6 of the moment-distribution table. Since the
distributed moments induced by this unbalanced moment are negligibly small, we end the
moment-distribution process. The final member end moments are obtained by algebraically summing the
entries in each column of the moment-distribution table. The final moments thus obtained are recorded on
line 8 of the table. Note that the final moments satisfy the equations of moment equilibrium at joints B and
C.
(8) Reactions, internal forces and their diagrams
With the member end moments known, member end shears and support reactions can now be
determined by considering the equilibrium of the free bodies of the members and joints of the continuous
beam, as discussed in previously. The bending moment diagram can then be constructed in the usual manner
by using the beam sign convention as shown in Fig.12.8 (c).
Example 12-5
Determine the member-end moments by using the moment distribution method and draw the internal
force diagrams and reactions for the rigid frame shown in Fig.12.9 (a). Assume EI = 1 .
Solution
Application of moment distribution method for rigid frames without sidesway follows the same
procedure as that given for continuous beams. However, unlike the continuous beams, more than two
members may be connected to a joint of a rigid frame. In such cases, care must be taken to record the
computations in such a manner that mistakes are avoided. Whereas some engineers like to record the
moment-distribution computations directly on a sketch of the frame, others prefer to use a tabular format
for such purposes. We will use the former for calculations, as illustrated in Fig.12.9 (b).
(1) Bending stiffnesses
2 EI 1
iDC = = S DC = 3iDC = 1
6 3
2 EI 1
iDA = = S DA = 4iDA = 2
4 2
3EI 1
iDE = = S DE = 4iDE = 2
6 2
3EI 1
iED = = S ED = 4iED = 2
6 2
4 EI 4
iEF = = S EF = 3iEF = 4
3 3
2 EI 1
iEB = = S EB = 4iEB = 2
4 2
(2) Distribution factors
At joint D:
S DC S DC 1
μ DC = = = = 0.2
∑ S Dj S DC + S DA + S DE 1 + 2 + 2
j
S DA 2
μ DA = = = 0.4
∑ S Dj 5
j
S DE 2
μ DE = = = 0.4
∑ S Dj 5
j
∑μ
j
Dj =1
10kN/m
C F
D E
4m
A B
6m 6m 3m
(a)
DC DA DE ED EB EF
0.2 0.4 0.4 0.25 0.25 0.5
C F
D E
45.00 −30.00 +30.00
−3.75 −7.50 −7.50 −15.00
−2.25 −4.50 −4.50 −2.25
+0.28 +0.56 +0.56 +1.13
−0.06 −0.11 −0.11 −0.06
+0.02 +0.02 +0.03
+42.69 −4.61 −38.08 +20.77 −6.92 −13.84
AD BE
−2.25 −3.75
+0.28
−0.06
+0.11
−2.31 −3.46
A B
(b)

(a) original structure and its loads; (b) processes of moment distribution and carryover
At joint E:
S ED S ED 2
μ ED = = = = 0.25
∑ S Ej S ED + S EB + S EF 2 + 2 + 4
j
S EB 2
μ EB = = = 0.25
∑ S Ej 8
j
S EF 4
μ EF = = = 0.5
∑ S Ej 8
j
∑μ j
Ej =1
42.69
38.08( 45 ) 32.89
( 45 ) 20.77 22.89
13.84 4.61 4.61
D
C F C F
D E 6.92 E
4.61 15.58
23.66
(unit: kN ⋅ m) (unit: kN) 27.11
37.11
2.31 A 3.46 B 1.73 A 2.60 B
(a) (b)
C 70.00 D 2.60 31.72 E 4.33 C D E

F F
2.60
4.33 4.33
⎛ unit: kN, ⎞ 4.61
22.89 ⎜ ⎟
(unit: kN) ⎝ kN ⋅ m ⎠
A B A 1.73 B 2.60
70.00 31.72
2.31 3.46
70 31.72
(c) (d)

(a) bending moment diagram; (b) shear force diagram; (c) axial force diagram; (d) reaction diagram
1 1×10 × 62
F
M DC = ql 2 = = 45kN ⋅ m
8 8
1 2 1× 10 × 62
F
M DE =− ql = − = −30kN ⋅ m
12 12
1 2 1× 10 × 62
F
M ED =+ ql = + = 30kN ⋅ m
12 12
(4) Moment distribution and carryover
As long as joints D and E are actually free to rotate, we release them, one at a time. We can release
either joint D or joint E; let us begin at joint E because it has the maximum unbalanced moment so as to
shorten the calculation process. Note that any unbalanced moment at joint D or E must be distributed to the
ends D or E of the three members connected to it in accordance with their distribution factors. The
processes of moment distribution and carryover are shown in Fig.12.9 (b)
(5) The final moments are recorded on the last line of the moment-distribution table and the moment
diagram of the frame is depicted in Fig.12.10 (a) by superposition method as used in the previous chapters.
(6) Reactions, shear and axial forces and their diagrams
With the member end moments known, member end shear and axial forces and support reactions can
now be determined by considering the equilibrium of the free bodies of the members and joints of the frame,
as discussed in previously. The shear and axial force diagrams can then be constructed in the usual manner as
shown in Fig.12.10 (b) and (c). Fig.12.10 (d) is also shown the reactions.
Example 12-6
Determine the member-end moments by using the moment distribution method and draw the bending
moment diagram for the rigid frame shown in Fig.12.11 (a).
Solution
(1) Simplification of calculation

Since the rigid frame shown in Fig. 12.11 (a) and the loading applied on it are all symmetric we can,
recalling from section 10.6.3, take one-half of the structure to analyze as shown in Fig.12.11 (b). In the
figure, supports G and H are known as double link support that prevent the translation along the links and
rotation but not translation perpendicular to the links.
(2) Bending stiffnesses and distribution factors
Bending stiffnesses:
2 EI 4 EI 4 EI
S AG = = S AC = SCA =
3 3 3
2
2 EI 4 EI
SCH = SCE =
3 3
Distribution factors at joint A:
μ AG = 0.5
μ AC = 0.5
Distribution factors at joint C:
μCH = 0.2
μCA = 0.4
μCE = 0.4
The distribution factors are recorded around their corresponding joint as shown in Fig.12.11 (b).
The fixed end moment of member AB can be determined in terms of original or its half structure by
using table 11-2.
20kN/m
(a) (b)
0.5
A G
A B
2I 0.5 −15
7.5 7.5
3m
I I −0.75
0.38 0.37
I 3.75 −0.04
C D −1.50 0.02 0.02
0.19 7.11 −7.11
3m
I I
−0.08
2.36
E F
0.4
3m H
0.2
C
(c) 0.4
7.11 7.11 −1.50 −0.75
A B −0.08 −0.03
15.37 −1.58 −0.78
I I
2.36 2.36
0.78 0.78 −0.75
C 1.58 1.58 D −0.04
−0.79
unit: kN ⋅ m E
0.79 E F 0.79 Fig.12.11 (contd)

(d)
Joint G A C H E
Member GA AG AC CA CE CH HC EC
DF 0.5 0.5 0.4 0.4 0.2
FEM -7.5 -15
Dist 7.5 7.5
0 2
CO -7.5 3.75
Dist -1.5 -1.5 -0.75
2
CO -0.75 0.75 -0.75
Dist 00.37 0.38
2
CO -0.37 0.19
Dist -0.08 -0.08 -0.032
0
CO -0.04 0.03 -0.04
Dist 0.02 0.02
∑ M -15.37 -7.11 7.11 2.36 -1.58 -0.78 0.78 -0.79

(a) original structure and its loading; (b) one-half of the frame and process of
the calculation;(c) bending moment diagram; (d) moment distribution table
1
F
In terms of original structure: M AB =− × 20 × 32 = −15kN ⋅ m
12
1
F
In terms of one-half of the structure: M AG = − × 20 × 1.52 = −15kN ⋅ m
3
1
F
M GA = − × 20 × 1.52 = −7.5kN ⋅ m
6
(4) Unlocking joints in the sequence of A, C, A, C and A will carry out the analysis as shown in
Fig.12.11 (b) or (d). Note that when columns for two ends of a member cannot be located adjacent to each
other, then an overhead arrow connecting the columns for the member ends may serve as a reminder to
carry over moments from one end of the member to the other. In Fig.12.11 (d), such an arrow is used
between the columns for the end of member CE. This arrow indicates that a distributed moment at end C of
member CE induces a carryover moment at the far end E. Also, the carryover factors for end G and H of
members AG and CH are equal to -1.
(5) Bending moment diagram
By the method of superposition, the bending moment diagram will be drawn as shown in Fig.12.11 (c)
432 Chapter 12 Method of Successive approximations and Influence Lines for Indeterminate Structures
widely used method for analysis of structures from 1930, when it was first published, through the 1960s.
Since the early 1970s, with the increasing availability of computers, the use of the moment distribution
method has declined in favor of the computer-oriented matrix methods of structural analysis. Nonetheless,
the moment distribution method is still preferred by many engineers for analyzing smaller structures, since it
provides a better insight into the behavior of structures. Furthermore, this method may also be used for
preliminary designs as well as for checking the results of computerized analyses.
12.2 Concepts and Terminology in Moment Distribution Method
12.2.1 Sign convention
We will establish the same sign convention as that adopted for the general displacement method and the
slope deflection equations: clockwise moment that act on the member are considered positive, whereas
counterclockwise moments are negative.
Since a clockwise moment at an end of a member must act in a counterclockwise direction on the
adjacent joint, the foregoing sign convention implies that counterclockwise moments on joints are considered
positive.
12.2.2 Distribution and carryover of joint moments
Consider the rigid frame with only one freely rotating joint A, on which a concentrated moment M is
applied, as shown in Fig.12.1 (a). If we use displacement method to analyze the structure, primary unknown
is the rotation of joint A, indicated by Δ1 ; and the primary equation will be written as
k11Δ1 + F1P = 0 (a)
The bending moment diagrams due to a unit rotation and due to the external load at joint A have been
drawn in Fig.12.1 (b) and (c). From the figure (b), we can find that
M AB = 4iAB ⎫
⎪
M AC = 3iAC ⎬ (b)
M AD = iAD ⎭⎪
Consider the moment equilibrium condition of the free body of the restrained joint A, we write
k11 = 4i AB + 3i AC + i AD (c)
From the figure (c), by considering the moment equilibrium condition of the restrained joint, we can
obtain
F1P + M = 0 F1P = − M (d)
Substituting Eqs. (c) and (d) into Eq. (a), solve out
F1P M
Δ1 = − = (e)
k11 4i AB + 3i AC + i AD
(a) (b) k11

M
C A D C A D
iAC iAD
Δ1 = 1
iAB
B
B
(c)
F1P
M
C D
A
Fig.12.1 Distribution of joint moment

(a) a rigid frame subjected to a concentrated moment;
(b) M 1 diagram due to Δ1 = 1; (c) M P diagram due external load
The substitution of Δ1 in the superposition equation M = M 1Δ1 + M P will yield the bending
moments of member ends as follows
4iAB ⎫
M AB = M AB Δ1 = 4iAB Δ1 = M⎪
4iAB + 3iAC + iAD ⎪
3iAC ⎪
M AC = M AC Δ1 = 3iAC Δ1 = M⎬ (f)
iAD ⎪
M AD = M AD Δ1 = iAD Δ1 = M ⎪
4iAB + 3iAC + iAD ⎭
2iAB ⎫
M BA = M BA Δ1 = 2iAB Δ1 = M⎪
−iDA ⎬ (g)
M DA = M DA Δ1 = −iDA Δ1 = M⎪
4iAB + 3iAC + iAD ⎪⎭
The final bending moment diagram can be obtained by multiplying the M 1 diagram shown in
Fig.12.1 (b) by Δ1 .
Forgoing discussion is the analyzing process for the rigid frame with a single free rotating joint, on
which a concentrated moment is applied, by means of the primary system of displacement method. Now we
like to provide the quantities appearing in the process with some new explanations and define some new
terminology as follows.
(1) Bending stiffness S AB
If we write the first equation of Eq. (b) as
M AB = 4iAB = S AB
S AB is defined as the bending stiffness of member AB at end A, which is the moment that must be
applied at end A of the member to cause a unit rotation of the end. Actually, the bending stiffness reflects
the resistance ability of the member end against its corresponding end rotation.
Recalling from the shape constants for various single-span indeterminate prismatic members with
different restraint conditions in table 11-1 of Chapter 11, the magnitude of the bending stiffness of a
member end is equal to the magnitude of the corresponding shape constant. From table 11-1, we obtain the
bending stiffnesses of near end A of member AB for three kinds of prismatic members listed in numbers 1,
3 and 5 in the table, as shown in Fig.12.2.
Note that: ① the first subscript A of S AB means that end A is the end at which a moment must be
applied to cause a unit rotation at the end; for simplicity, end A is named nearend and end B is named
EI
farend, respectively. ② the magnitude of S AB is related to both the linear bending stiffness i = ,
L
bending stiffness per length, or relative bending stiffness of the member and the support conditions of the
farend. The different support conditions yield different S AB . For example, the Fig.12.2 shows four kinds
of support conditions at the farend of member AB. Their corresponding bending stiffness at the nearend
S AB will be written as follows.
The farend is fixed, S AB = 4i (12-1)
The farend is hinged, S AB = 3i (12-2)
The farend is rollered, S AB = i (12-3)

The farend is free, S AB = 0 (12-4)
(a) (c)
S AB = 4 EI l S AB = EI l
EI B
A A EI
1 B
1
l
(b)
(d)
S Ab = 3EI l
S Ab = 0
EI B EI
A A B
1
1
l
Fig.12.2 Bending stiffness

(a) far end is fixed; (b) far end is hinged;
(c) far end is fixed by double links; (d) far end is free
(2) Distribution factor μ A j and rotational stiffness of a joint

Recalling from Eq. (f), when a moment M is applied to a rigid joint as shown in Fig.12.1, the
connecting members will each supply a portion of the resisting moment proportional to its bending stiffness
and necessary to satisfy moment equilibrium at the joint. The fraction of the total resisting moment
supplied by the member is called the distribution factor. To determine what fraction of the applied moment
M is resisted by each of the three members connected to the joint, we rewrite Eq. (f) again as
4iAB ⎫
M AB = M = μ AB M ⎪
3iAC ⎪
M AC = M = μ AC M ⎬
iAD ⎪
M AD = M = μ AD M ⎪
4iAB + 3iAC + iAD ⎭
Generally,
M A j = μA jM (12-5)
In which,
SA j
μA j = (12-6)
∑S
j
Aj
μ A j is termed distribution factor at end A of member Aj. Obviously, the second subscript j can be B, C
and D. ∑ SA j represents the sum of the bending stiffnesses of all the members connected to joint A.
j
The rotational stiffness of a joint is defined as the moment required causing a unit rotation of the joint.
From Eq. (12-6), we can see that the rotational stiffness of a joint is equal to the sum of the bending
stiffnesses of all the members rigidly connected to the joint.
Note that the distribution factors at the identical joint posses following relationship.
S AB + S AC + S AD
∑μ = μ + μ AC + μ AD = =1 (12-7)
∑ SA j
AB
A
j
(3) Carryover factor C A j

The carryover factor represents the ratio of the farend moment to the nearend moment when the
nearend rotates.
Consider again the rigid frame shown in Fig.12.1. When the concentrated moment applies at joint A,
the nearend of each connecting member supply a resisting moment, M A j , in the meanwhile, the farend of
each connecting member supply a resisting moment as well, whose magnitude is related to the support
conditions of the farend. Thus, we can express carryover factor as
MjA
CA j = (12-8)
MA j
From Eqs. (f) and (g), we can obtain

M BA 1 M CA M DA
C AB = = C AC = =0 C AD = = −1
M AB 2 M AC M AD
Therefore, we can summarize the magnitude of the carryover factor of a prismatic member for
different farend support conditions as follows.
1
The farend is fixed, C= (12-9)
2
The farend is hinged, C=0 (12-10)
The farend is fixed by double links, C = −1 (12-11)
From Eq. (12-8), we can express the farend moment by carryover factor as follows
M j A = CA j M A j (12-12)
Now reexplain the performance of the member-end moments for the rigid frame shown in Fig.12.1 as
follows:
When a moment M applies at a joint, the moment will be distributed among the nearends of the
various members connected to the joint and carryovered to the farends of the members. The distributed
moment of each nearend is named nearend bending moment or distributed bending moment whose
magnitude is equal to the multiplication of distribution factor and the moment M; whereas the carryover
moment of each farend is known as farend bending moment or carryover bending moment equaling to the
product of the carryover factor and corresponding nearend bending moment.
Forgoing discussion has shown how to use the concepts of moment distribution and carryover to
analyze a structure subjected to concentrated joint moment. The following example will give a detail
explanation.
Example 12-1
Determine the member-end moments for the rigid frame with a concentrated couple M = 100kN ⋅ m
applied at joint A as shown in Fig.12.3 by using the moment distribution method.
Solution
50
i
M = 100kN ⋅ m
37.5
B i A i D
25 12.5 12.5
i
(unit: kN ⋅ m)
E
(a) (b)
Fig.12.3 Distribution and carryover of a moment

(a ) a rigid frame subjected to a concentrated moment; (b) bending moment diagram

In terms of Eqs. (12-1) through (12-4), the bending stiffnesses of the members connected at joint A
will be evaluated as follows
S AB = 4i AB = 4i
S AC = 3i AC = 3i
S AD = i AD = i
S AE = 0
By using Eq. (12-6), the distribution factors will be calculated to be

S AB 4i
μ AB = = = 0.5
∑ S A j 4i + 3i + i
j
S AC 3i
μ AC = = = 0.375
∑ S A j 8i
j
S AD i
μ AD = = = 0.125
∑ Aj i
S 8
j
S AE 0
μ AE = = =0
∑ S A j 8i
j
(2) Member end moments

By employing Eq. (12-5), the member end moments can be obtained to be
M AB = μ AB M = 0.5 ×100 = 50kN ⋅ m

M AC = μ AC M = 0.375 ×100 = 37.5kN ⋅ m
M AD = μ AD M = 0.125 ×100 = 12.5kN ⋅ m

M AE = μ AE M = 0 × 100 = 0
In terms of Eqs. (12-9) through (12-11), the carryover moments of the members can be calculated as
1
M BA = C AB M AB = × 50 = 25kN ⋅ m
2
M CA = C AC M AC = 0 × 37.5 = 0
M DA = C AD M AD = −1×12.5 = −12.5kN ⋅ m
Under the action of a joint couple, the distribution and carryover moments of the members are their
final end moments. By the end moments, the bending moment diagram of the frame can be drawn as shown
in Fig.13.3 (b).
12.3 Moment Distribution at A Single Joint—Moment-Distribution Process
Moment distribution is based on the principle of successively locking and unlocking the joints of a
structure in order to allow the moments at the joints to be distributed and balanced. The best way to explain
the concept of the moment distribution at a single joint may be by the following practical model.
(a)
θB
A θB B C
M BA M BC
P
(b) lock joint

MB
MB
A B C
M F
AB M BA
F M BC
F
=0 M BA
F
B M BC
F
=0
P
(c) unlock joint

A C
B
M ′AB ′
M BA ′
M BC
Fig.12.4 Concept of moment distribution method for a single joint

(a) original structure; (b) locking joint B; (c) unlocking joint B
Consider the continuous beam shown in Fig.12.4 (a), which is made of a thin steel strap. As shown in
the figure, when the bobweight reaches to a magnitude of P the beam will be deformed as shown in
dashed line and the ends of the members will yield moments that we are desired to determine. The
procedure to determine the member end moments due to a general load can be taken as following
processes:
(1) Initially assume that joint B of the beam that is free to rotate is temporarily restrained (or locked)
against rotation by a clamp applied to it so as to make the joint as a hypothetical fixed support ( θ B = 0 ).
As we can see, the hypothetical clamp has made the original continuous beam becomes two independent
single span indeterminate beams. Then let the external load P apply to this hypothetical fixed structure.
As shown in dashed lines of Fig.12.4 (b), only span AB is deformed and only the fixed end moments at the
ends of member AB are yielded. Since member BC is subjected to no loading there is no deformation and
no fixed end moment at the ends of the member. Obviously, the restraint moment at the clamp, indicated by
F
M B , is temporarily balanced by the fixed end moment at the end B of member AB, M BA . The restraint
moment M B can be thusly determined by the moment equilibrium condition of joint B.
We can see from Fig.12.4 (b) that since the fixed end moment at the end B of member BC is equal to
zero, i.e., M BC = 0 , the restrain moment at the clamp can be determined to be
F
M B = M BA
F
+ M BC
F
= M BA
F
Apparently, the restraint moment at the clamp is equal to the algebraic sum of the fixed end moments of the
member ends connected to the locked joint. The sign convention of the restraint moment is that the
clockwise is positive.
(2) Since the joint B of the original continuous beam is free to rotate and there is no moment applied
to it [Fig.12.4 (a)], i.e.,∑M B = 0 , we must release joint B by removing the clamp, thereby allowing it to
rotate an original angle θ B under the action of the minus unbalanced moment, − M B , same in magnitude
but opposite in sense to the restraint moment. The rotation of joint B due to − M B will induce the
′ and M BC
distribution moments at the ends of the members connected to the joint, indicated by M BA ′ ,
respectively. The bending of member AB due to the distributed moment causes carryover moments to yield
′ , as well.
at the farend of the member, M AB
(3) If we superimpose the cases shown in Fig.12.4 (a) and (b) together, the superimposing results for
both the deformations and the member end moments are identical to the original case shown in Fig.12.4 (a).
For instance, the moment at the end B of member AB will be
M BA = M BA
F
′
+ M BA
Now we will briefly summarize the concepts and processes of moment distribution method due to the
action of general loads as follows:
(1) Lock joint Lock the joint that connects the members of a continuous beam by a clamp to
decompose the continuous beam into independent single span indeterminate beams. External loads are
applied to these individual single span indeterminate beams, and fixed-end moments at the ends of the
beams are computed, and unbalanced restraint moment is evaluated by taking the algebraic sum of the
fixed-end moments at the ends connected to the locking joint.
(2) Unlock joint Release the locking by removing the clamp, allowing it to rotate under the action
of the minus of the unbalanced moment, i.e., same in magnitude but opposite in sense to the restraint
moment. The rotation of the joint will induce the distribution moments at the nearends and carryover
moments at the farends of the members connected to the joint.
(3) Superimpose results Superposing the results of above two steps will obtain the real
member-end moments of the continuous beam.
Example 12-2
Determine the member-end moments for the two-span continuous beam shown in Fig.12.5 (a) by
using the moment distribution method. The flexural rigidity is a constant, i.e., EI = constant .
Solution
(1) Lock joint B. A clamp is applied to joint B to restrain its rotation as shown in Fig.12.5 (b). The
fixed-end moments, the clockwise is positive, due to the external loads can be evaluated by using table 11-2
as follows.
1 2 20 × 62
M F
AB = − ql = − = −60kN ⋅ m
12 12
1 2 20 × 62
M F
BA = + ql = + = +60kN ⋅ m
12 12
3 3 × 32 × 6
F
M BC =− Pl = − = −36kN ⋅ m
16 16
The fixed end moments are written below their corresponding ends. The restrain moment at joint B,
M B , must be the algebraic sum of the fixed-end moments connected to the joint. Thusly, M B can be
written as
M B = 60 − 36 = 24kN ⋅ m
(2) Unlock joint B. Release joint B by removing the clamp and apply a couple, equal in magnitude
but opposite in sense, − M B = −24kN ⋅ m , at the joint as shown in Fig.12.5 (c). The distribution moments
at the nearends and the carryover moments at the farends will be calculated as follows.
The linear stiffness (or relative stiffness) of the two members AB and BC can be defined as
EI
i=
l
The bending stiffnesses at end B for the two members will be written as
S BA = 4iAB = 4i
S BC = 3iBC = 3i
(a)
20kN/m 32kN
B C
A
EI EI
6m 3m 3m
MB
B
60 36
(b)
24kN ⋅ m
20kN/m 32kN
B C
A
−60 +48 −36
(c) 24kN ⋅ m
0.571 0.429
A C
B
−6.85 −13.70 −10.30
Fig.12.5 Process of moment distribution for a single joint

(a) original structure and its loads; (b) locking joint B and
evaluating fixed end moments; (c) unlocking joint B and
determine the distributed and carryover moments
Then, the distribution factors will be

S AB 4i
μ AB = = = 0.571
S BA + S BC 7i
S BC 3i
μ BC = = = 0.429
S BA + S BC 7i
Checking: ∑μ = μ BA +μ BC = 1
The distribution factors written above their corresponding ends within a square are all right.
The distributed moments can be evaluated by multiplying the negative of the unbalanced joint moment
by the corresponding distribution factors as follows.

'
M BA = ( −24) × 0.571 = −13.70kN ⋅ m
'
M BA = ( −24) × 0.429 = −10.30kN ⋅ m
The distributed moments are written below their corresponding ends underlined to indicate that the joint is
release and balanced as shown in Fig.12.5 (c).
The carryover moment at a farend can be computed by multiplying the distributed moment by the
member carryover factor (fixed farend, 1/2; hinged farend, 0). Thus,
1 ' 1
'
M AB = M BA = × (−13.70) = −6.85kN ⋅ m
2 2
'
M CB =0
The carryover direction of a carryover moment is indicated by an arrow as shown in Fig.12.5 (c).
(3) Superpose member-end moments. Algebraically summing the member-end moments obtained in
steps (1) and (2) can yield final member-end moments.
(a)
BA BC
A 0.571 0.429 C
distribution factors
B
fixed end moments -60 +60 -36 0
distributed moments ⎫
⎪
and ⎬ -6.85 ← -13.70 -10.30 →0
carryover moments ⎪⎭
member end moments -66.85 46.30 -46.30 0
(b)
66.85
( 90 ) 46.30 ( 48 )
A C
B
33.43 24.85
(unit: kN ⋅ m)
Fig.12.6 Calculating form for moment distribution method and M diagram

(a) calculating form for moment distribution method; (b) bending moment diagram
In order to facilitate the calculation, the above first two steps can be combined together in terms of the
form shown in Fig.12.6 (a). Note that the member end moment doubly underlined is its final result. A
positive answer for an end moment indicates that its sense is clockwise, whereas a negative answer for an
end moment implies a counterclockwise sense. These final moments must satisfy the equation of moment
equilibrium at joint B. That is,
∑M j
Bj = 46.30 − 46.30 = 0
The bending moment diagram of the beam is constructed in Fig.12.6 (b).

Example 12-3
Determine the member-end moments by using the moment distribution method and construct the
bending moment diagram for the rigid frame shown in Fig.12.7 (a). The relative flexural rigidity is shown
in the figure.
Solution
(1) Bending stiffnesses and distribution factors at joint A

Bending stiffnesses:
S AB = 3i = 3 × 2 = 6
S AC = 4i = 4 ×1.5 = 6
S AD = 4i = 4 × 2 = 8
∑S
j
Aj = 6 + 6 + 8 = 20
Distribution factors:
S AB 6
μ AB = = = 0.3
∑ S A j 20
j
S AC 6
μ AC = = = 0.3
∑ S A j 20
j
S AD 8
μ AD = = = 0.4
∑ S A j 20
j
∑μ
j
Aj = μ AB + μ AC + μ AD = 0.3 + 0.3 + 0.4 = 1
1 2 15 × 42
M F
AB = ql = − = 30kN ⋅ m
8 8
Pab 2 50 × 3 × 22
M F
AD =− 2 =− = −24kN ⋅ m
l 52
Pa 2b 50 × 32 × 2
F
M DA = = − = 36kN ⋅ m
l2 52
50kN
15kN/m AB AC AD
B 0.3 0.3 0.4 D
B D
i=2 A i=2 30 A −24 36
−1.8 −1.8 −2.4 −1.2
4m
i = 1.5 28.2 1.8 −26.4 34.8
C C −0.9
4m 3m 2m
(a) (b)
28.2 34.8
26.4
30 60
B
D
1.8
15.9
(unit: kN ⋅ m) 29.4
0.9
(c)
(a) original structure and its loads; (b) moment distribution and carryover;
(3) Distribution and carryover

The processes for moment distribution and carryover are carried out in the form as shown in Fig.12.7
(b). Note that the columns for all member ends, which are connected to the same joint, are grouped together,
so that any unbalanced moment at the joint can be conveniently distributed among the members connected
to it.
(4) The bending moment diagram
The final member end moments are obtained by summing all the moments in each column of the
moment-distribution table shown in Fig.12.7 (b). Note that the final moments satisfy the equation of
moment equilibrium at joint A. The bending moment diagram is shown in Fig.12.7 (c), which can be
constructed by superposition method discussed previously.
12.4 Moment Distributions at Multi-Joints—Successive Approximations
The moment distribution method used to analyze the continuous beams with only one rotatable joint
can also be employed to analyze both of continuous beams and rigid frames without sidesway which have
more then one free rotatable joints by an iterative procedure, in which it is initially assumed that all the
joints of the structure that are free to rotate are temporarily restrained against rotation by imaginary clamps
applied to them. External loads are applied to this hypothetical fixed structure, and fixed-end moments at
the ends of its members are computed. These fixed-end moments generally are not in equilibrium at those
joints of the structure that are actually free to rotate. The conditions of equilibrium at such joints are then
satisfied iteratively by releasing one joint at a time, with the remaining joints assumed to remain clamped.
A joint at which the maximum moments are not in balance is selected. The joint is then released by
removing the clamp, thereby allowing it to rotate under the negative of the unbalanced moment until the
equilibrium state is reached. The rotation of the joint induces distributed moments at the ends of the
members connected to it, whose values are determined by multiplying the negative of the unbalanced joint
moment by the distribution factors for the member ends connected to the joint. The bending of these
members due to the distributed moments causes carryover moments to develop at the farends of the
members, which can easily be evaluated by using the member carryover factors. The joint, which is now in
equilibrium, is reclamped in its rotated position. Next, another joint with the maximum unbalanced moment
is selected and is released, balanced, and reclamped in the same manner. The procedure is repeated until the
unbalanced moments at all the joints of the structure are negligibly small. The final member end moments
are obtained by algebraically summing the fixed-end moment and all the distributed and carryover
moments at each member end. This iterative process of determining member end moments by successively
distributing the unbalanced moment at each joint is called the moment-distribution process.
With member end moments known, member end shears, member axial forces, and support reactions
can be determined through equilibrium considerations, as discussed in the previous chapters.
The following examples will illustrate the application of moment distribution method to the analysis
of the continuous beams and rigid frames without sidesway and with more then one free rotatable joints
458
Example 12-7
Analyze the symmetric rigid frame shown in Fig.12.12 (a) by using the moment distribution method and
draw its bending moment diagram. EI = constant .
Solution
y
(a) 24kN/m (b) 24kN/m
C D
G C
G
2m
2m
x x E
E F
2m
3m
A B
3m 3m
Fig.12.12 (contd)
y
(1) Simplification of calculation

Because the rigid frame shown in Fig. 12.12 (a) and the loading applied on it are both symmetric about the axes
x and y , we can, recalling from section 10.6.3, take a quarter of the structure to analyze as shown in Fig.12.12
(b). In the figure, supports E and G are termed double link supports that prevent the translation along the links and
rotation but not translation perpendicular to the links.
EI
Bending stiffnesses (assume = 6i ):
1
EI EI
SCG = iCG = = 2i SCE = iCE = = 3i
3 2
SCG 2i SCE 3i
μCG = = = 0.4 μCE = = = 0.6
SCG + SCE 2i + 3i SCG + SCE 2i + 3i
By using table 11-2, we can find
ql 2 1
M F
CG =− = − × 24 × 32 = −72kN ⋅ m
3 3
ql 2 1
F
M GC =− = − × 24 × 32 = −36kN ⋅ m
6 6
(4) Unlocking joint C once will carry out the process of the analysis as shown in Fig.12.12 (c). Note
that the carryover factors for end E and G of members CE and CG are equal to -1.
By using the symmetric property of the frame and the method of superposition, the bending moment
diagram of the frame will be constructed as shown in Fig.12.12 (d).
(c) (d)
CE CG 43.2 (108 ) 43.2

0.6 0.4 G 43.2 43.2
C
−72 −36
64.8
43.2 28.8 −28.8
43.2 −43.2 −64.8
unit: kN ⋅ m
−43.2 E 64.8
43.2 43.2
43.2 (108 ) 43.2
(a) original structure and its loading; (b) a quarter of the frame and process of the calculation;
(c) moment distribution process; (d) bending moment diagram
Example 12-8
Analyze the continuous beam with a hangover as shown in Fig.12.13 (a) by using the moment
distribution method and draw its bending moment diagram. Assume EI = 1 .
Solution

At joint B:
1 1
S BC = 4iBC = 4 × =1 S BA = 3iBA = 3 × = 0.5
4 6
1 0.5
μ BC = = 0.667 μ BA = = 0.333
1 + 0.5 1 + 0.5
460
At joint C:
SCD = 0 SCB = 4iCB
0 4iCB
μCD = =0 μCB = =1
0 + 4iCB 0 + 4iCB
(a) 60kN 10kN

A B C
D
6m 2m 2m 2m
(b)
A 0.333 0.667 0 1
D
B C
−30 30 −20
10 20 → 10
−10 ← −20 0
3.33 6.67 → 3.33
−1.67 ← −3.33 0
0.56 1.11 → 0.56
−0.28 ← −0.56 0
0.09 0.19 → 0.10
−0.05 ← −0.10 0
0.02 0.03
14.0 −14.0 20 −20
(c)
14.0
( 60 ) 20
A
D
B C
(unit: kN ⋅ m)
43

(a) original structure; (b) process of moment distribution
and carryover; (c) bending moment diagram
Note that since end D of member CD is free no moment is needed to rotate a unit value of joint C, i.e.,
SCD = 0 . This implies that the moment at the joint is not needed to distribute at the end C of member CD
and to carry over to end D.

It can be seen that the fixed end moment at the end C of member CD can be directly obtained by the
method of cross section. It is
F
M CD = −10 × 2 = −20kN ⋅ m
Recalling from table 11-2, we find

Pl 60 × 4
F
M BC =− =− = −30kN ⋅ m
8 8
Pl 60 × 4
F
M CB = = = 30kN ⋅ m
8 8
(3) Process of moment distribution and carryover
The process of moment distribution and carryover is shown in Fig.12.13 (b).
The bending moment diagram is shown in Fig.12.13 (c).
(5) Discussion
As we can see from above process that the moments are distributed at two joints, could you distribute
them only at joint B?
12.5 No-Shear Moment Distribution Method
Thus far, we have applied moment distribution method to the analysis of structures in which the
translations of the joints were zero. In this section, we apply the moment distribution method to analyze
rigid frames whose joints may undergo both rotations and translations but whose columns’ shears can be
statically determined by equilibrium considerations. Since there is no shear yielded in the process of
moment distribution and carryover, the method is referred to as no-shear moment distribution method.
12.5.1 Applying condition of no-shear moment distribution method
As mentioned before, we will apply the moment distribution method to some rigid frames with
sidesway. The applying condition is that the shear forces of all the members can be statically determined by
equilibrium conditions except those without relative end translation.
Consider, for example, the rigid frame shown in Fig.12.14 (a), which is a commonly utilized primary
system of a symmetric structure subjected to antisymmetric loads as discussed in Chapter 10. Both ends of
each beam of the frame do not have any relative translation during the sidesway of the frame, i.e., there is
no relative translation in the direction perpendicular to its axis, so the frame’s sidesway will not induce
internal forces in the beams. These sorts of members are named members without relative end translation.
462 Chapter 12 Method of Successive Approximations and Influence Lines for Indeterminate Structures
Further more, the shear force of each column of the frame can be determined statically although the ends of
the columns have sidesway. The shear force diagram of the columns is shown in Fig.12.14 (b). These kinds
of members are known as members with statically determinate shears.
The columns AB and CD of the rigid frame shown in Fig.12.14 (c), however, are neither the members
without relative end translation and nor members with statically determinate shears, so no-shear moment
distribution method cannot be applied to the frame.
The detailed discussion about no-shear moment distribution will take place in the following
subsections.
(a)
(b) (c)
P D
B D E
D
P
P C
C
2P
P B
B A C
3P
A
A
Fig.12.14 Rigid frame being able to use no-shear moment distribution method
(a) half a portal frame; (b) shear diagram of columns; (c) rigid frame not being
able to use no-shear moment distribution method
12.5.2 Fixed end moments for the members with statically determinate shears
Actually, the analyzing process of no-shear moment distribution method is the same as that of the
moment distribution method discussed in the previous sections. The only difference between them is that
the structures to be analyzed have different characteristics in internal forces and deformations.
For instance, when use no-shear moment distribution method to analyze the rigid frame shown in
Fig.12.15 (a), the process includes following three steps: (1) Lock the rotatable displacement of joint A
[Fig.12.15 (b)], which connects the two members, and determine the fixed end moments for each member.
(2) Unlock the joint and evaluate the distributed and carryover moments of the member ends [Fig.12.15 (c)].
(3) Superpose the member end moments obtained from the first two steps to determine the final results of
the member end moments.
Now let us concern the fixed end moments for the members of the frame shown in Fig.15 (a). Since
the sidesway of the frame will not induce bending moments in beam AC, the fixed end moments of the
member can be determined as before. This means that for the members without relative end displacement,
the fixed end moments can be evaluated as previous. However, the determination of the fixed end moments
for the column will have some difference as we do previously. The difference is that when we lock the
rotation of joint A, the joint still has horizontal translation. As long as the shear QAB at the end A of
member AB is determinate ( QAB = 0 ), if we substitute the shear QAB for its corresponding restraint the
column AB will be equivalent to the single span indeterminate member as shown in Fig.12.15 (d), whose
fixed end moments can be conveniently determined by use the results in line 14 of table 11-2.
A C C A QAB = 0
A
A
B B B B
(a) (b) (c) (d)
Fig.12.15 Analytical process of no-shear moment distribution method

(a) half a portal frame; (b) locking joint A; (c) unlocking joint A;
(d) equivalent computing model of member AB
P2 P1
A A
P P+P
1
B 1 2
B B
(b)
C C
(a) (c)
Fig.12.16 Fixed end moment determination for a two-story frame
(a) locking joints A and B; (b) equivalent computing model of member AB;
(d) equivalent computing model of member BC
For the multistory rigid frame as shown in Fig.12.16 (a), which is composed of members without
relative end displacement and members with statically determined shears, the method to evaluate the fixed
end moments is the same as that mentioned above (used for one-story similar frame). If we substitute the
shear of each top end of the two columns for their corresponding restraint the determination of the fixed
end moments for the two columns will be equivalent to those of the two single span indeterminate members
as shown in Fig.12.16 (b) and (c).

It is observed from above that the steps to determine the fixed end moments of a member with
statically determinate shears can be concluded as: (1) determine the member end shear by a proper
equilibrium condition; (2) evaluate fixed end moments by substituting the shear for its corresponding
restraint of the member, which is a single span indeterminate member with one end fixed and another end
doubly linked.
12.5.3 Distribution and carryover factors for the members with statically determinate shears
Next, let us focus our attention on the distributed and carryover moments for the members of the
frame shown in Fig.12.15 (c) due to the unlocking of joint A. Unlocking joint A means releasing the joint
by removing the clamp and applying a couple, equal in magnitude but opposite in sense, at the joint as
shown in Fig.12.17 (a). Under the action of the couple, the internal force and displacement of column AB
will be equivalent to a cantilever member shown in Fig.12.17 (b). It can be seen from the figure that there is
no shear in the member, so bending moment is a constant. Thusly, this sort of member is named no-shear
member or zero-shear member. Recalling from table 11-1 (in line 5), the distribution moments at the near
ends and the carryover moments at the far ends will be calculated as follows.
M AB A
A
θA
θA
θA
Q=0 B Q=0
B
(a) (b)
Fig.12.17 Unlocking joint A
(a) performance of internal force and displacement of member AB;
(d) equivalent computing model of member AB
M AB = iABθ A M BA = − M AB
Obviously, the bending stiffness of member AB is

S AB = iAB
(12-13)
The carryover factor of the member is

C AB = −1
(12-14)
Similarly, the bending stiffnesses and carryover factors for the columns shown in Fig.12.16 can be
determined as
S BC = iBC CBC = −1
S BA = iBA CBA = −1
It can be observed from forgoing discussion that the members with statically determinate shears are
zero-shear members under the action of the joint moments. Their distributed and carryover moments are
obtained under the condition of no-shear forces. The method is thusly referred to as no-shear moment
distribution method.
Example 12-9
Determine the member-end moments by using no-shear moment distribution method and draw the
bending moment diagram for the rigid frame shown in Fig.12.18 (a).
Solution
5kN BA BC
B C 1/ 5 4 / 5 C B 1.39 C
B
i =4
1
−2.67 −3.75 1.39
1.28 5.14
1kN/m
−1.39 1.39 5.70

i =3
4m
−5.33
−1.28
unit: kN ⋅ m
A A −6.61 A
2m 2m 6.61
(a) (b) (c)

(a) original structure and its loads; (b) computing process;
It can be seen from the figure that since the member BC is a member without relative end translation
during the frame’s sidesway and the member AB is a member whose shear force can be determined
statically, no-shear moment distribution method can be utilized to the frame. Recalling table 11-2, the fixed
end moment will be determined as
3
F
M BC =− × 5 × 4 = −3.75kN ⋅ m
16
1 1× 42
F
M BA = − ql 2 = − = −2.67kN ⋅ m
6 6
1
F
M AB = − ql 2 = −5.33kN ⋅ m
3
The bending stiffnesses and distribution factors of the members can be evaluated to be
⎧ 4
⎧ S BC = 3i1 = 12 ⎪⎪ μ BC = 5
⎨S = i = 3 ⎨
⎩ BA 2 ⎪ μ BA = 1
⎪⎩ 5
The carryover factor of member BA is equal to -1.
The process of moment distribution and carryover is shown in Fig.12.18 (b). Fig.12.18 (c) shows the
bending moment diagram of the frame.
Example 12-10
Determine the member-end moments by using no-shear moment distribution method and draw the
bending moment diagram for the rigid frame shown in Fig.12.19 (a).
Solution
The rigid frame shown in Fig.12.19 (a) is a symmetric structure subjected to an asymmetric loading. If
we decompose the loading into a symmetric one and an antisymmetric one as shown in Fig.12.19 (b) and
(c), the bending moments of the structure will be induced by the antisymmetric loading provided that we
ignore the axial deformations of the two beams. If we use the symmetric property of the structure shown in
Fig.12.19 (c), half its structure as shown in Fig.12.19 (d) can be analyzed, in which the flexural rigidity of
the beams will be doubled as shown in the figure.
Since columns AB and BC are the members with statically determinate shears, their shear forces can
be determined to be
QAB = 4kN QBC = 12.5kN
From the equivalent computing models of the two columns shown in Fig.12.19 (e), their fixed end
moments will be determined, by using table 11-2, to be
4 × 3.3
F
M AB = M BA
F
=− = −6.6kN ⋅ m
2
12.5 × 3.6
F
M BC = M CB
F
=− = −22.5kN ⋅ m
2
(2) Distribution factors
The bending stiffnesses of the members connecting to joint B can be written to be
S BA = iBA = 3.5
S BC = iBC = 5
S BE = 3iBE = 3 × 54 = 162
∑S j
Bj = 170.5
8kN 4kN 4kN

A i = 2.7 A
3.3m
3.5 3.5
17kN B 8.5kN B 8.5kN

3.6m
5 5
C C
(a) (b)
4kN
4kN A 4kN 4kN A 5.4
A
3.5
B B 12.5kN B
8.5kN B 8.5kN 8.5kN 5.4
5
C C
(c) C (d) (e)

(a) original structure and its loads; (b) symmetric structure subjected to symmetric loading;
(c) symmetric structure subjected to antisymmetric loading; (d) half the structure;
(e) equivalent computing models of columns
Thusly, the distribution factors at joint B can be obtained as

3.5
μ BA = = 0.0206
170.5
5
μBC = = 0.0293
170.5
162
μ BE = = 0.9501
170.5
7.05 7.05
0.9789
(a) (b)
A D
A
0.0211 27.79
−6.60 B 21.64 E 21.64

6.15 6.15
−0.60
0.15 7.05 27.79 unit: kN ⋅ m
−7.05 7.05
C F
0.0206 23.36 23.36
0.9501
B
0.0293
−6.60 −22.5
0.60 0.85 27.65
−0.15
0.00 0.01 0.14
−6.15 −21.64 27.79
−22.5
−0.85
−0.01 Fig.12.20 Computing process and result of example 12-10
C
−23.36 (a) computing process; (b) bending moment diagram
Similarly, we can determine the distribution factors at joint A, which is written in the square box
(3) Moment distribution and carryover
The computing process is shown in Fig.12.20 (a), which adopted the order of B, A, B, A. Note that the
carryover factors of the columns are -1. The final bending moment diagram is drawn in Fig.12.20 (b).
12.6 Shear Distribution Method
Up to now, we have discussed the application of moment distribution method to the analysis of
structures without sidesway and no-shear moment distribution method to the analysis of some special
structures with sidesway. In this section, we will discuss the shear distribution method and its application to
the analysis of bent frames with sidesway.
12.6.1 Bent frames subjected to horizontal loads at their top hinged joints
Now let us reconsider the bent frame subjected to a concentrated horizontal force shown in Fig.12.21
(a), which was analyzed by displacement method in Chapter 11 (example 11-4). When one of the columns
of the bent frame is subjected to a horizontal translation of a unit magnitude at its top, Δ = 1 [Fig.12.21
3i
(b)], the shear force yielded in the column will be Q = , which is actually the stiffness coefficient due
h2
to the unit sidesway of the column. Since the each column has the same sidesway Δ , the shear forces of
the columns induced by the sidesway will be written as
3i1 ⎫
Q1 = 2
Δ = d1Δ ⎪
h1 ⎪
3i2 ⎪
Q2 = 2 Δ = d 2 Δ ⎬
h2 (a)
⎪
3i3 ⎪
Q3 = 2 Δ = d3Δ ⎪
h3 ⎭
3i j ( EI ) j
In which, d j = 2
, ij = , h j represent the stiffness coefficient due to sidesway, relative flexural
h j lj
rigidity and height of the jth column, respectively.
Recalling from the horizontal force equilibrium condition of the free body of the top portion of the
bent frame [Fig.12.21 (c)], we write
Q1 + Q2 + Q3 = P
(b)
Substituting Eq. (a) into Eq. (b), we find
dj
Qj = 3
= μjP
(c)
∑d
j =1
j
It can be observed from Eq. (c) that the shear forces of the columns are direct proportion to their stiffness
dj
coefficient due to sidesway, μ j = 3
, which is now defined as shear distribution factor.
∑dj
j =1
Δ Δ =1
P
P 3i
d=
h2
EI EI EI
i= 1
i = 2
i = i
h
3
1
h 1
2
h2
3
h 3
(a) (b)
P
dΔ
1 dΔ2
dΔ
3
(c)
Fig.12.21 Analysis of a bent frame subjected to a horizontal load at its top
(a) original structure and its loads; (b) stiffness coefficient due to sidesway;
(c) free body diagram of the top portion of the bent frame
P P
i
1
i Q =μP Q =μP Q =μP
2 i 3
1 1 2 2 3 3
h 1
h 2
h 3
Qh 1 1
(a) Qh
2 2
(b)
Qh
3 3
Fig.12.22 Shear distribution method of a bent frame

(a) original structure and its loads; (b) shear distribution
Now we will explain the problem in the shear distribution point of view as follows. When the bent
frame shown in Fig.12.22 (a) is subjected to a horizontal concentrated force, the force is distributed to the
columns in the proportion of their shear distribution factor, i.e., the column’s shear is evaluated by
multiplying the force by its shear distribution factor ( Q j = μ j P ). Since the bending moments at the hinges
are equal to zero, the bending moments of the columns can be determined by their distributed shear forces.
Thusly, the method to analyze bent frames subjected to horizontal loads at their top is referred to as shear
distribution method.
12.6.2 Shear distribution of rigid frames with beams of infinite flexural rigidity
Now let us discuss the application of shear distribution method to the rigid frame with a beam of
infinite flexural rigidity subjected to a concentrated horizontal force as shown in Fig.12.23 (a).
Δ =1
Δ
ib = ∞ θ =0 P
P 12i
d=
h2
i1
h
i2 i
i3
(a) (b)
P
d1Δ d2Δ d3 Δ
(c)
Fig.12.23 Analysis of a bent frame with beams of infinite flexural rigidities
(a) a bent frame with beams of infinite flexural rigidities;
(b) stiffness coefficient due to sidesway; (c) free body diagram of the top portion
Since the flexural rigidities of the beams are infinite, there is neither deflection nor rotation for the
beams. Therefore, only the common horizontal translation of the top joints can be considered to be the
unknown displacement of the frame. While the shear force induced in a column having no end rotation due
12i
to a unit horizontal displacement as shown in Fig.12.23 (b) will be equal to Q = 2 , the stiffness
h
12i 12i
coefficient due to the unit sidesway must be 2 , denoted as d = 2 . As long as each column has the
h h
same sidesway Δ , their shears can be expressed as
Qj = d jΔ
(a)
12i j
In which, d j = represents the stiffness coefficient of the column having no end rotation due to
h 2j
sidesway.
Recalling from the horizontal force equilibrium condition of the free body composed by the top
portion of the rigid frame [Fig.12.23 (c)], we write

Q1 + Q2 + Q3 = P
(b)
Substituting Eq. (a) into Eq. (b), we obtain
dj
Qj = 3
= μjP
(c)
∑d
j =1
j
Q1h1 Q2 h2 Q3h3
ib = ∞ P 2 2 2
P
i1
i2 i3 Q 1
h1
Q
h2 2
h3 Q 3
Q1h1
2 Q2 h2
2
Q3h3
(a) (b) 2
Fig.12.24 shear distribution of a rigid frame with beams of infinite flexural rigidities
(a) original structure; (b) shear distribution and bending moment diagram
It can be observed from Eq. (c) that the shear forces of the columns of the rigid frame with beams of
dj
infinite flexural rigidity are distributed by the distributed factor μ j = . In other words, when a rigid
∑dj
frame with beams of infinite flexural rigidity as shown in Fig.12.24 (a) is subjected to a horizontal
concentrated force at its top, the force is distributed to the columns by their shear distribution factor, i.e.,
the column’s shear is evaluated by multiplying the force by its shear distribution factor ( Q j = μ j P ). Note
that the inflection point (the location of zero bending moment, referring example 10-6) is located at the
middle of the column when subjecting the column having no end rotation to a sidesway Δ . Therefore, the
magnitude of the bending moments at the ends of the column can be determined by multiplying its
hj
distributed shear by half the column’s height, i.e., M j = ( μ j P) × . Once the bending moments of the
2
columns are known, the bending moments of the beams can be determined by equilibrium as well, and the
bending moment diagram of the frame can be conveniently constructed as shown in Fig.12.24 (b).
12.6.3 Shear distribution of bent frames or rigid frames of infinite-rigidity beams subjected to
no-joint loads
The forgoing discussed shear distribution method may also be applied to the analysis of the bent
frames or rigid frames of infinite-rigidity beams subjected to no-joint loads. Consider, for example, the bent
frame shown in Fig.12.25 (a), whose columns are subjected to no-joint loads. The following three steps will
help us to apply shear distribution method to these kinds of structure.
Δ F1P
P P F1P
= +
ib = ∞ Δ
F1P
q q
= +
(a) (b) (c)
Fig.12.25 Analysis due to no-joint loads

(a) original structure; (b) locking sidesway; (c) unlocking sidesway and shear distribution
(1) Lock the sidesway of the frame by adding an artificial corresponding restraint as shown in
Fig.12.25 (b). Then, determine the shear forces of the ends (fixed end shears) of the columns subjected to
external loads by using the results of table 11-2. By the fixed end shears, the restraint force corresponding
to the artificial restraint, denoted by F1P , will be determined.
(2) Unlock the sidesway by releasing the artificial restraint and apply the negative of the restraint
force, − F1P on the frame at the location and in the direction corresponding to the artificial restraint as
shown in Fig.12.25 (c). And then determine the shear forces distributed in the columns by shear distribution
method.
(3) Superposing the computing results obtained from steps (1) and (2) together will obtain the final
shear forces of the columns of the frame.
Example 12-11
Determine the member-end moments by using shear distribution method and draw the bending
moment diagram for the rigid frame shown in Fig.12.26 (a).
Solution
dj
(1) Shear distribution factors of the columns μ j =
∑d j
1
μ1 = μ3 = = 0.25
1+ 2 +1
2
μ2 = = 0.5
1+ 2 +1
5
Ib = ∞ 10kN 5 10 5
A B C
P = 10kN 5
i =1 Q1 = 2.5kN Q2 = 5kN Q3 = 2.5kN
4m
1 i =2
2
i =1
3
D E F
10 5
(a) 5kN ⋅ m (b)
M AB M BA
M BC
A (d)
(c)
B
5 10
(a) original structure and its loads; (b) shear distribution and bending moment diagram;
(c) moment equilibrium condition of joint A; (d) moment equilibrium condition of joint B
(2) Shears of the columns
Q1 = Q3 = 0.25 × 10 = 2.5kN
Q2 = 0.5 × 10 = 5kN
(3) The bending moments of the ends of the columns

For the rigid frame of infinite-rigidity beams, the inflection points are located at the middles of the
columns, so the end moments of the columns will be determined to be
h1 4
M 1 = M 3 = −Q1 × = −2.5 × = −5kN
2 2
h2 4
M 2 = −Q2 × = −5 × = −10kN
2 2
By the equilibrium conditions of joint moments [Fig.12.26 (c) and (d)], the end moments of the beams can
be evaluated to be
M AB = 5kN ⋅ m
1
M BA = M BC = × 10 = 5kN ⋅ m
2
Once the end moments of the members are known, the bending moment diagram will be constructed
as shown in Fig.12.26 (b).
Example 12-12
Analyze by using shear distribution method and draw the bending moment diagram for the rigid frame
Solution
Ib = ∞ Δ 6.66kN ⋅ m 10kN
A B C
A B C
(10)
5kN/m
3.34
i =1
4m
1 i =22
i =1
3
D E F
D E F 6.66
(b)
(a)
5 5kN ⋅ m
5 5 1.66 10
5kN ⋅ m 10 10kN
5
5 5
(10) 3.34
11.66 10 5
5 10 5 (d)
(c)

(a) original structure and its loads; (b) primary system and fixed end moments and shears;
(c) shear distribution and bending moment diagram; (d) bending moment diagram
(1) Lock the sidesway of the frame by adding an artificial corresponding restraint as shown in
Fig.12.27 (b). Then, the fixed end shears of the columns and the restraint force can be determined. By using
table 11-2, we write
qL 5× 4
F
QAD =− =− = −10kN F
QBE = QCF
F
=0
2 2
From the horizontal-force equilibrium condition of the beams, ∑ X = 0 , the restraint force yielded at the
artificial restraint will be
F1P = −10kN
(2) Unlock the sidesway by releasing the artificial restraint and apply the negative of the restraint
force, − F1P on the frame at the location and in the direction corresponding to the artificial restraint as
shown in Fig.12.27 (c). And then determine the shear forces distributed in the columns by shear distribution
method as shown in the figure.
(3) Superposing the computing results obtained from steps (1) and (2) together will obtain the final
shear forces of the columns. By employing the shear forces and the location of inflection points (at the
middle of the column), the bending moments of the members and the bending moment diagram of the
structure will be drawn as shown in Fig.12.27 (d).
12.7 Influence Lines for Forces of Statically Indeterminate Structures
In this section, we discuss the procedures for constructing influence lines for forces of a statically
indeterminate structure. It may be recalled from Chapter 8 that an influence line is a graph of a response
function of a structure as a function of the position of a downward unit load moving across the structure.
The basic procedure for constructing influence lines for forces of an indeterminate structure is the
same as that for determinate structures considered in Chapter 8. The procedure essentially involves
computing the values of the response function of interest for various positions of a unit load on the
structure and plotting the response function values as ordinates against the position of the unit load as
abscissa to obtain the influence line. Since the influence lines for forces and moments of determinate
structures consist of straight-line segments, such influence lines were constructed in Chapter 8 by
evaluating the ordinates for only a few positions of the unit load and by connecting them with straight lines.
The influence lines for forces of an indeterminate structure, however, are generally curved lines. Thus the
construction of influence lines for forces of an indeterminate structure requires computation of many more
ordinates than necessary in the case of determinate ones.
Although any of the methods of analysis of indeterminate structures presented in the book can be used
for computing the ordinates of influence lines, we will use the method of force method, discussed in
Chapter 10, for such purposes. Once the influence lines for forces of an indeterminate structure have been
constructed, they can be applied to determine the unfavorable loading position and absolute maximum
response and the like of the structure.
12.7.1 Principle of constructing the influence lines for forces of statically indeterminate beams by
static method
Consider the continuous beam shown in Fig.12.28 (a). The beam is subjected to a downward-moving
concentrated load of unit magnitude, the position of which is defined by the coordinate x measured from
the left end A of the beam, as shown in the figure. Suppose that we wish to draw the influence line for the
vertical reaction Z1 at right-end support of the beam by using force method. Thusly, the support C is
(a) (b)
x P =1
B B C
A C A
(c) (d)
x P =1 x P =1
B B δ1 P
A C A
C
Z1
(e)
x
δ11
B
A C
δ P1 Z1 = 1
Fig.12.28 Influence line for forces of an indeterminate structure

(a) original indeterminate structure; (b) primary structure;
(c) primary structure subjected to P = 1; (d) primary structure subjected to Z1 = 1
removed and it is replaced by a vertical force Z1 . Now we will determine Z1 by force method. Recalling
from force method, the consideration of that the deflection of the primary system [Fig.12.28 (c)] at C due to
the combined effect of the external unit load and the unknown redundant Z1 must be equal to zero, will
yield the compatibility condition as follows
δ11Z1 + δ1P = 0
Consequently,
δ1P
Z1 = −
δ11 (12-15)
Note that the primary structure is still an indeterminate beam [Fig.12.28 (b)], whose degrees of
indeterminacy are equal to that the degrees of indeterminacy of the original structure subtract one (i.e.,
n − 1 ).
δ1P denotes the deflection of the primary beam at C due to the

In which the flexibility coefficient
unit load P = 1 at position x [Fig.12.28 (d)], whereas the flexibility coefficient δ11 denotes the
deflection at C due to the unit value of the redundant Z1 = 1 [Fig.12.28 (e)].
We can use Eq. (12-15) for constructing the influence line for Z1 by placing the unit load
successively at a number of positions along the beam, evaluating δ1P for each position of the unit load,
and plotting the values of the ratio −δ1P / δ11 . However, a more efficient procedure can be devised by
applying law of reciprocal displacements (Section 9.8), according to which the deflection at C due to a
unit load P = 1 at x must be equal to the deflection at x due to a unit load of Z1 = 1 at C; that is,
δ1P = δ P1 . Thus equation (12-15) can be rewritten as
δ P1
Z1 = − (12-16)
δ11
In which δ P1 denotes the deflection of the primary structure at position x due to the unit reaction
Z1 = 1 at C [see Fig.12.28 (e)].
It can be observed that since the flexibility coefficient δ11 is a constant without respect to the unit
load position x and the deflection δ P1 varies with respect to x , δ P1 is a function of x . Thusly, the
expression for reaction Z1 in terms of the variable position x of the unit load P = 1 can be expressed
clearly as
δ P1 ( x)
Z1 ( x) = − (12-17)
δ11
Equation (12-17) implies that when the position x of the unit load varies, the shape of the function
Z1 ( x) is just the influence line for Z1 , whereas the variable shape of Z1 ( x) is proportional to that of
δ P1 ( x) and the variable shape of δ P1 ( x) is just about the deflection shape of the primary beam due to the
unit reaction Z1 = 1 at C as shown in Fig.12.28 (e).
Therefore, the deflection shape of the primary structure due to Z1 = 1 can represent the shape of the
influence line for Z1 . Eq. (12-17) also indicate that the influence line for Z1 can be obtained by
multiplying the deflected shape of the primary structure due to the unit load Z1 = 1 by the scaling factor
−1/ δ11 .
Note that the deflections δ P1 ( x) and δ11 are considered to be positive when their senses coincide
with their corresponding unit loads. However, since the negative ratio δ P1 / δ11 represents the function of
influence line for Z1 , the ordinate of the influence line for Z1 is positive in the upward direction (i.e.,
above abscissa x ).
Example 12-13
Construct the influence line for vertical reaction YB of the indeterminate beam shown in Fig.12.29 (a)
by static method.
Solution
(a)
x = αl P = 1 (b) x = αl P = 1
B
A A B
Z1 = YB
l
(c) (d) α l P =1
A B
Z1 = 1 x = αl
l (1 − α ) l
(e)
1
11 40 81 1
128 128 128
(a) original indeterminate structure; (b) primary system;
(c) M 1 diagram of primary structure;
(d) M P diagram of primary structure; (e) influence line for YB
(1) Selection of primary system

The primary system will be obtained by substituting a vertical force at B, Z1 = YB , for its
corresponding support as shown in Fig.12.29 (b).
(2) Bending moment diagrams
The bending moment diagrams due to the unit value of the reaction and the unit load P = 1 are
shown in Fig.12.29 (c) and (d), respectively.
(3) Flexibility coefficient δ11 and deflection δ P1
By graph-multiplication method, the flexibility coefficient δ11 and deflection δ P1 will be evaluated
to be
1 1 2 l3
δ11 = × l × l × l =
EI 2 3 3EI
1 ⎡1 1 1 2 ⎤ l3 ⎛ 1 2 1 3 ⎞
δ P1 = − ⎢⎣ 2 α l × (1 − α )l × α l + α l × l × α l = − ⎜ α − α ⎟
EI 3 2 3 ⎥⎦ EI ⎝ 2 6 ⎠
(4) Solution of Z1
Substituting the expressions for δ11 and δ P1 into Eq. (12-17), the solution of Z1 will be obtained
as
δ P1 ( x) 3 2 1 3 1
Z1 ( x) = − = α − α = (3α 2 − α 3 ) (a)
δ11 2 2 2
(5) Construction of influence line for Z1

The ordinates of the influence line for Z1 can now be computed by applying Eq. (a) successively for
1
each position of the unit load P = 1 . For example, when the unit load is located at x = 0 , x = l,
4
1 3 1 1 3
x = l , x = l and x = l (i.e., α = 0, α = , α = , α = , α = 1 ), the values of Z1 are given by
2 4 4 2 4
11 40 81
Z1 = 0 , Z1 = , Z1 = , Z1 = and Z1 = 1 , respectively. Connecting the points of the
128 128 128
influence line smoothly will obtain the influence line as shown in Fig.12.29 (e).
It can be seen from the shape of the influence line that the main difference between the influence lines
for statically indeterminate and those for statically determinate is that the influence lines for forces of
determinate structures consist of straight-line segments, while the influence lines for forces of
indeterminate structures are generally curved lines.
12.7.2 Qualitative influence lines for continuous beams by principle of virtual displacement
In many practical design stages, such as when designing continuous beams or building beams
subjected to uniformly distributed live loads, it is usually sufficient to draw only the qualitative influence
lines to decide where to place the live loads to maximize the response functions of interest. As in the case
of statically determinate structures (Section 8.5), mechanismic (or virtual displacement) method provides a
convenient means of establishing qualitative influence lines for indeterminate structures.
Recall from Section 8.5.2 that virtual displacement method can be stated as follows:
The influence line for a force (or moment) response function is given by the deflected shape of the
released structure obtained by removing the restraint corresponding to the response function from the
original structure and by giving the released structure a unit displacement (or rotation) at the location and
in the direction of the response function.
The procedure for constructing qualitative influence lines for indeterminate structures is the same as
that for determinate structures discussed in Section 8.5. The procedure essentially involves: (1) removing
from the given structure the restraint corresponding to the response function of interest to obtain the
released structure; (2) applying a infinitesimal displacement (or rotation) to the released structure at the
location and in the positive direction of the response function; and (3) drawing a deflected shape of the
released structure consistent with its support and continuity conditions. The influence lines for
indeterminate structures are generally curved lines since the released structures are, generally, still an
Once a qualitative influence line for a structural response function has been constructed, it can be used
to decide where to place the live loads to maximize the value of the response function. As discussed in
Section 8.6, the value of a response function due to a uniformly distributed live load is maximum positive
(or negative) when the load is placed over those portions of the structure where the ordinates of the
response function influence line are positive (or negative). Because the influence-line ordinates tend to
diminish rapidly with distance from the point of application of the response function, live loads placed
more than three span lengths away from the location of the response function generally have a negligible
effect on the value of the response function. With the live-load pattern known, an indeterminate analysis of
the structure can be performed to determine the maximum value of the response function.
Example 12-14
Draw qualitative influence lines for the bending moments on support section C and section G, and the
shear on section G of the five-span continuous beam shown in Fig.12.30 (a).
Solution
(1) Influence line for bending moment at support section C.

To determine the qualitative influence line for the bending moment at C, we insert a hinge at C in the
actual beam and give the released beam an infinitesimal relative rotation in the positive direction of M C
by rotating the portion to the left of C counterclockwise and the portion to the right of C clockwise, as
shown in Fig.12.30 (b). The deflected shape of the released beam thus obtained represents the qualitative
influence line for M C [Fig.12.30 (c)].
(2) Influence line for bending moment and shear at section G.
The qualitative influence line for bending moment M G at section G is determined in a similar
manner and is shown in Fig.12.30 (d) and (e).
To determine the qualitative influence line for shear at section G, we insert two links at G in the actual
beam and give the released beam an infinitesimal relative displacement in the positive direction of the shear
QG by moving end G of the left portion of the beam downward and end G of the right portion upward, as
shown in Fig.12.30 (f). The influence line for QG is shown in Fig.12.30 (g).
Note that to obtain a quantitative influence line for a force of an indeterminate structure, the restraint
corresponding to the force must be removed and a corresponding unit displacement must be applied to the
released structure, and the deflection shape may be obtained by determining the deflection equations of the
members of the structure. The further discussion about the quantitative influence lines can be fined in other
special papers or books.
(a)
x P =1
A B C G D E F
(b)
θC
B C D E F
A
δ P1 MC
(c)
B C D E F
A
θG
(d)
A B C G D E F
MG
(e)
A B C D E F
(f)
A B C D E F
G
λG QG
(g)
A B C D E F
G
Fig.12.30 Influence line for continuous beam

(a) original indeterminate structure; (b) deflection shape due to M C ;
(c) influence line for M C ; (d) deflection shape due to M G ; (e) influence line for M G ;
(f) deflection shape due to QG ; (g) influence line for QG
12.8 Most Unfavorable Arrangement of Live Loads and Envelops
for Internal Forces of Continuous Beams 483
12.8 Most Unfavorable Arrangement of Live Loads and Envelops for Internal Forces
of Continuous Beams
12.8.1 Most unfavorable arrangement of live loads on continuous beams
Continuous beams are commonly adopted structures in engineering projects. Generally, the loads
applying on a continuous beam involve dead loads distributed in whole span and maintained in life, and
live loads distributed arbitrarily and maintained from time to time. Thusly, the bending moment on some
section of the beam due to the dead loads is unchangeable, while that due to the live loads is changeable. In
order to guarantee sufficient strength of a continuous beam, the maximum bending moment of the beam
due to both dead and live loads must be determined before design the beam. Actually, the determination of
the maximum bending moment is the evaluation of the effect of the live loads, because once the maximum
bending moment (positive or negative) due to the live loads is determined, the final maximum bending
moment can be obtained by adding both the maximum bending moment due to live loads and that due to
dead loads.
However, the determination of the maximum bending moment due to the live loads needs to know the
unfavorable arrangement of the live loads on the continuous beam, whereas the estimation for the
unfavorable arrangement of the live loads must resort to the help of influence line for the bending moment.
Consider, for example, the five-span continuous beam shown in Fig.12.31 (a). The qualitative
influence lines for bending moments at support section B and section 2 are also shown in Fig.12.31 (b) and
(e). It can be seen from the figure that to cause maximum positive bending moment at B, we place the live
load over spans CD and EF, where the ordinates of the influence line for M B are positive [Fig.12.31 (c)];
contrarily, the placement of the live load over other spans will cause maximum negative bending moment at
B, as shown in Fig.12.31 (d). Similarly, the arrangement of the live load to cause maximum positive and
negative bending moments at 2 are shown in Fig.12.31 (f) and (g).
It can be observed from above analysis that (1) the live-load arrangement for maximum negative
bending moment on the section near a support is that the load is placed over the adjacent and every second
spans to the support; (2) the live-load arrangement for maximum positive bending moment on the section
of a middle span is that the load must be applied on the span and every second spans to the section.
12.8.2 Envelops for internal forces
An envelop of the bending moments (or shears) of a continuous beam is a curved line whose ordinates
are composed of the maximum bending moments (or shears) against their corresponding sections of the
beam. The envelop of bending moment of the beam may be obtained by (1) evaluating the bending
moments of the beam due to dead loads; (2) estimating, by considering the unfavorable live-load
arrangements, the maximum (positive and negative) bending moments due to the live load; (3) superposing
the maximum (positive and negative) bending moments from the first two cases; (4) laying off the
maximum bending moments over or below the axis of the beam; (5) connecting the maximum bending
moments smoothly. Similarly, the envelop of shears can be also obtained by above five-step’s process. The
following example will illustrate the construction of bending-moment envelop in detail.
(a)
B 2 C D E F
A
influence line for M B

(b)
A B C D E F
live-load arrangement for positive maximum M B

(c)
A B C D E F
(d) live-load arrangement for negative maximum M B
A B C D E F
(e) influence line for M 2

A B C D E F
live-load arrangement for positive maximum M 2

(f)
A B C D E F
(g) live-load arrangement for negative maximum M 2
A B C D E F
Fig.12.31 Live-load arrangement for maximum bending moment of a continuous beam

Example 12-14
Draw the bending-moment envelop for the girder of the floor system shown in its cross section,
Fig.12.32 (a).
Solution
(a)
slab beam
girder column
37 30
730 750 2
19 18
(b)
A B
748 750 2

(a) cross section of a floor system ; (b) computing model of the girder
(1) Analytical model of the girder

The girder is the member of the floor system of an industrial workshop. Since the rigidity of the girder
is much stiffer than that of the columns and there is local deformations under ends of the girder, the
bending moments of the girder will almost not transferred to the columns, and thusly the girder is
considered to be supported by hinges and rollers. The analytical model of the girder will be simplified as
shown in Fig.12.32 (b). In order to simplify analysis, the weight of the girder is considered as a portion of
concentrated dead load which is measured by G as shown in Fig.12.33 (a).
(2) Bending moment due to dead loads
The entire girder is finally simplified as a three-span continuous beam. The bending moment diagram
due to the dead loads can be obtained by using anyone of the method discussed in Chapters 11 through 12
for analyzing an indeterminate structure. Thusly obtained bending moment diagram of the beam is shown
in Fig.12.33 (a).
(3) Bending moment due to the live load
The maximum (positive and negative) bending moment of each section might be obtained by method
of superposition, that is: (1) draw the bending moment diagram due to the live loads arranged over each
span of the beam individually, as shown in Fig.12.33 (b) through (d); (2) superpose the maximum (positive
and negative) bending moments from the figures (b) to (d) for each section together.
(a)
G 131.6kN ⋅ m G G G G
G = 66kN
A 3 4
D
1 2 B C 5 6
33.3 33.3
77.2
121.2
P = 109.8kN
(b) P 145.8kN ⋅ m
85.1
24.5
A C 5 6 D
1 2 B 3 4
12.1
36.2 24.1
177.3
225.9
P P
109.5kN ⋅ m
109.8kN
(c) 73.0
36.5
A B C D
1 2 3 4 5 6
165.0
Fig.12.33 Determination of bending-moment envelop (contd)
(a) bending moment diagram due to dead loads;
(b) bending moment diagram due to live loads arranged over first span;
(c) bending moment diagram due to live loads arranged over second span
(4) Envelop of bending moment

The envelop of the bending moments of the beam can be determined by superposing the bending
moment evaluated in cases (2) and (3). In other words, the ordinates of the envelop of the bending moments
can be obtained by adding the maximum (or minimum) bending moments due to the dead loads and those
due to the live loads. For instance, the maximum and minimum bending moments at sections 1, 2 and B
will be calculated as
M 1max = 121.2 + 225.9 + 12.1 = 359.2kN ⋅ m

M 2max = 77.2 + 177.3 + 24.1 = 278.6kN ⋅ m
M 1min = 121.2 − 36.5 = 84.7kN ⋅ m
M 2min = 77.2 − 73.0 = 4.2kN ⋅ m
M Bmax = −131.6 + 36.2 = −95.4kN ⋅ m
M B min = −131.6 − 145.8 − 109.5 = −386.9kN ⋅ m
Thusly obtained bending-moment envelop is shown in Fig.12.33 (e).
145.8kN ⋅ m P P = 109.8kN
(d) 85.1
24.5
A 1 2 B C D
3 4 5 6
12.1 24.1 36.2
177.3
225.9
386.9kN ⋅ m
(e)
95.4 76.3
A 1 2 5 6 D
3 4
4.2
B C
84.7
278.6 198.3
359.2
Fig.12.33 Determination of bending-moment envelop

(d) bending moment diagram due to live loads arranged over third span;
(e) bending-moment envelop
SUMMARY
In this chapter, we have studied two methods of successive approximations, which are two classical
formulations of the general displacement (stiffness) method, called moment distribution method and
no-shear distribution method. In addition, shear distribution method, the method of constructing influence
lines for the forces of continuous beams, the arrangement to cause maximum bending moment and envelop
of bending moments (or shears) of continuous beams are studied respectively.
1. The moment distribution method is suitable for the analysis of continuous beams and rigid frames
without sidesway. The concept and procedure of the method are clear and simple as long as there is no need
to develop the simultaneous equations of the unknown displacements, and the member end moments can be
determined directly by simple algebraic operation. So the method is a widely used hand-oriented method.
The analytical procedure of the method contains the following two steps:
(1) Lock the rigid joints. The step involves computing fixed-end moments due to the external loads
by assuming that all the free joints of the structure are temporarily restrained against the rotations.
(2) Unlock the rigid joints. The step consists of balancing the moments at free joints by the
moment-distribution process. In the moment-distribution process, at each free joint of the structure, the
unbalanced moment is evaluated and distributed to the member ends connected to it. Carryover moments
induced at the far ends of the members are then computed, and the process of balancing the joints and
carrying over moments is repeated until the unbalanced moments are negligibly small. The final member
end moments are obtained by algebraically summing the fixed-end moment and all the distributed and
carryover moments at each member end.
2. No-shear distribution method is a special case of moment distribution method.
Since there is no shear induced in the members whose shears are statically determinate and there is no
bending moment yielded in the members having no relative translation during the joint sidesway, the
moment distribution can be applied to this kind of structure only consisting of members with statically
determinate shears and those without relative translation due to the joint sidesway. Their distributed and
carryover moments are obtained under the condition of no-shear forces.
3. Shear distribution method is suitable for the analysis of the structures without joint rotations, such
as bent frames with hinged joints at their tops and rigid frames with beams of infinite rigidities. The
horizontal forces will be distributed in the columns by their corresponding shear distribution factors.
4. There are two methods, static and mechanical (or virtual work) method, to construct influence
lines for statically indeterminate structures. The formulation of static method is the principle of force
method; the principle of mechanical method is actually the application of unit displacement method applied
to determinate the deflection shapes of indeterminate structures.
The concept of envelop of bending moments (or shears) is very helpful in the design of bridges, beams
of building frames and the like.
12-1 What are the sign conventions for the fixed end moments and moments at the ends of the members
of a structure?
12-2 What is the bending stiffness? How much are the bending stiffnesses for a member with both
nearend and farend fixed and for a member with nearend fixed and farend hinged, respectively?
12-3 What is distribution factor? What is the difference between the distribution factors and bending
stiffnesses? Why is the summation of the distribution factors at a joint equal to 1?
12-4 What is the carryover factor? How to determine a carryover factor?
12-5 How to determine the distributed and carry over moments due to a joint couple?
12-6 How many steps is the computing operation of the moment distribution method due to a general
loading? What is the physical meaning of each step?
12-7 What is a fixed end moment? How to determine the restraint moment at an artificial restraint? Why
must the restraint moment be added a minus sign when it is distributed?
12-8 How are the computing steps when analyze continuous beams and rigid frames without sidesway by
moment distribution method? How to determine the end moments of a member?
12-9 Why will the restraint moments at the artificial restraints vanish when implement moment
distribution method to the continuous beams and rigid frames? i.e., why is the analysis convergent?
12-10 Can the moment distribution method be used to analyze the internal forces due to the temperature
changes and support settlements? In what circumstance it can be used and in what cases it cannot be
used?
12-11 What is no-shear distribution method? What are its application conditions?
12-12 Why can no-shear distribution method be only applied to single span symmetric rigid frames?
12-13 What problems will occur when no-shear distribution method implements to single span asymmetric
rigid frame?
12-14 What is shear distribution method? What are its application conditions? How to use shear
distribution method when the columns of a structure are subjected to no-joint loads?
12-15 What is the difference between the influence lines for statically indeterminate structures and those
for statically determinate structure?
12-16 Why must the primary structure be the beam from which a restraint corresponding to the force
whose influence line is desired is removed when determines the influence line for the force by static
method?
12-17 Please explain the principle and summarize the steps of constructing the influence lines for
indeterminate structures by virtual displacement method.
12-1 Determine the bending moments at the ends of the members for the following continuous beams and
rigid frames shown in the figure by moment distribution method.
12-2 Analyze the continuous beams shown in the figure by moment distribution method and draw their
bending moment diagrams.
12-3 Analyze the rigid frame shown in the figure by moment distribution method and draw their bending
moment diagram.
M 60kN ⋅ m
A B C A C
EI EI 2EI B 2EI
6m 6m
3m
EI
problem 12-1 (a)
D
4m 4m
problem 12-1 (b)
(a) (b)
60kN ⋅ m
40kN q = 20kN/m
A B C
EI =constant A EI = 1 B EI = 3 C
6m 2m 2m 2m 4m
problem 12-2 (a) problem 12-2 (b)
q = 20kN/m
B EI = 3 C
4m
EI = 1
A
6m
problem 12-3
12-4 Analyze the continuous beams shown in the figure by moment distribution method and draw their
bending moment diagrams and determine their reactions.
(a) 20kN/m 40kN 40kN (b) 400kN 40kN/m

A B C D
A B C D EI = 1 EI = 3 EI = 4
6m 6m 2m 2m 2m 6m 3m 3m 6m
EI =constant
(c) 40kN
20kN/m 40kN
B C D E F
A
EI =constant
4m 4m 2m 2m 4m 1.2m
problem 12-4
12-5 Analyze the rigid frames shown in the figure by moment distribution method and draw their
bending moment, shear and axial force diagrams.
(a) 80kN 15kN/m q = 10kN/m

(b) P = 20kN
A B C D A 4I B 4I C D
4m
6m
EI =constant 2I 2I
E F E F
3m 3m 6m 6m 5m 5m 1m

80kN A
(c) 20kN/m
(d)
D E F G
6m
EI EI EI 60kN
C
6m
B
2EI EI EI
20kN/m
B C
6m
2m
EI =constant
A
6m 6m 3m 3m D E
4m 4m
problem 12-5
q = 20kN/m q = 10kN/m EI = constant
B C
A B C D E
6m
4m
EI =constant
A D
12m 4m 4m 4m 4m
(a)
(b)
12kN 12kN
4kN 0.6m 4kN
1.2m 1.2m
I J K L
A 1.5 EI B
EI EI EI EI
4m
60kN 60kN 60kN 60kN

3m
EI EI
20kN/m
1kN/m
E F G H
C D 4EI 4EI
4EI
4m
EI EI EI EI
EI EI
4m
A B C D
2m 2m 2m 2m 2m 2m 2m
E F
0.5m 2m 0.5m (d)
(c) problem 12-6

C 20kN 10kN/m
B
EI = 2 C
B i1 = 4
6m
EI = 1 EI = 1
4m
i2 = 3
A D
12m
A
(a)
4m
(b)
19kN C i = 0.9 D
3.8m
0.395 0.395
38kN B i = 0.9 E
3.8m
0.395 0.395
A F problem 12-7
8m
(c)
12-8 Analyze the structures shown in the figure by shear distribution method and draw their bending
moment diagrams. EI = constant
E F G H C EI b = ∞ D
20kN/m
EA = ∞ EA = ∞ EA = ∞
20kN/m
6m
8m
I I I I I
I
A B C D A 12m B
(a) (b)
problem 12-8
12-9 Draw the influence line for the bending moment M A of the beam shown in the figure and answer
the question when the unit load reaches to where the magnitude of M A will get the maximum
value.
12-10 Draw the influence lines for the reaction RB , bending moment M D and sear force QD of the
beam shown in the figure. Assume EI = constant .
x P =1
x P =1
A D B C
A B
l 3m 3m 6m

CHAPTER 13
MATRIX DISPLACEMENT METHOD

In this chapter, we begin with an introduction, discussing the process of preparing a computing
model of the structure to be analyzed. We also define global and local coordinate systems and
explain the concept of unknown displacements of a structure. Next we derive member
force-displacement relations in local coordinates. We consider the transformation of member end
forces and end displacements from local to global coordinates and vice versa and develop the
member stiffness relations in global coordinates. We formulate the stiffness relations for the entire
structure by combining the member stiffness relations and, finally, develop a step-by-step procedure
for the analysis of continuous beams and frames by the matrix displacement (stiffness) method.
13.1 Introduction
In the previous of the text, we have focused our attention on the classical methods of structural analysis.
Although a study of classical methods is essential for developing an understanding of structural behavior and
the principles of structural analysis, the analysis of large structures by using these hand-oriented methods can
be quite time consuming. With the availability of inexpensive, yet powerful, microcomputers, the analysis of
structures in most design offices is routinely performed today on computers using software based on matrix
methods of structural analysis.
The matrix structural analysis is the method which uses the principle of classical structural mechanics
to formulate the analytical procedure of a structure by matrix algebra and compute the algebraic equations
representing the structure’s response by a computer. In matrix analysis, all the calculations are carried out
with matrix algebra. The use of matrix notations will uniformly express the formulations of structural
analysis in a simple and elegant form.
Although both the force (flexibility) and the displacement (stiffness) methods can be expressed in
matrix form, the displacement method is more systematic and can be more easily implemented on computers.
Thus, most of the commercially available computer programs for structural analysis are based on the
displacement method. In this chapter, we will consider only the matrix displacement (stiffness) method of
structural analysis.
13.1.1 The fundamental principle of matrix displacement method
In fact, the fundamental principle of matrix displacement method is identical to that of the
495
496 Chapter 12 Matrix Displacement Method
displacement method discussed in chapter 11. Matrix methods do not involve any new fundamental
principles, but the fundamental relationships of equilibrium, compatibility, and member force-displacement
relations are now expressed in the form of matrix equations, so that the numerical computations can be
efficiently performed on a computer. Like the displacement method, a structure is also considered to be an
assemblage of straight members connected at their ends to joints in the matrix displacement method. A
member is defined as a part of the structure for which the member force-displacement relations to be used
in the analysis are valid in matrix algebraic form. By assembling the force-displacement relations of the
members under the equilibrium and compatibility conditions, the force-displacement relation, expressed in
matrix form, of the entire structure can be obtained.
Therefore, the analysis of matrix displacement method comprises of the same stages as those of
displacement method presented in chapter 11. They are: (1) the discretization of a structure; (2) the
establishment of force-displacement relations of members; (3) the development of force-displacement
relation of the structure. All of these relations must be expressed in matrix algebraic form.
13.1.2 Computing model
(1) Discretization of a structure

In matrix displacement method of analysis, a member is defined as a part of the structure for which the
member force-displacement relations to be used in the analysis are valid in matrix algebraic form. In other
words, given the displacements of the ends of a member, one should be able to determine the forces and
moments at its ends by using the force-displacement relations expressed in matrix algebra. A joint is
defined as a structural part of infinitesimal size to which the member ends are connected. The members and
joints of structures are also referred to as elements and nodes, respectively.
Before proceeding with the analysis, the structure must be discretized, or a computing model of the
structure must be prepared. The model is represented by a line diagram of the structure, on which all the
joints, the joints’ unknown displacements and members are identified by numbers. Consider, for example,
the rigid frame shown in Fig. 13.1(a). The computing model of the frame is shown in Fig. 13.1 (b), in
which the element numbers are enclosed within circles to distinguish them from the joint numbers. As
shown in this figure, the frame is considered to be composed of four members and six joints for the purpose
of analysis. Note that, since the member 2 and member 4 have, generally, different rotations at their joint,
the joint are numbered by two different numbers.
(2) Global and local coordinate systems
In the matrix displacement method, the overall geometry and behavior of the structure are described
with reference to a rectangular global (or structural) coordinate system. The global coordinate system used
in this chapter is a right-handed x y z coordinate system, with the plane structure lying in the x y plane,
as shown in Fig. 13.1(b).
Because it is usually convenient to derive the basic force-displacement relations in terms of the forces
and displacements in the directions along and perpendicular to members, a local (or member) coordinate
system is defined for each member of the structure. The origin of the local x y z coordinate system for a
member may be arbitrarily located at one of the ends of the member, with the x axis directed along the
centroidal axis of the member. The positive direction of the y axis is chosen so that the coordinate system
is right-handed, with the local z axis pointing in the positive direction of the global z axis. In Fig. 13.1
(b), the positive direction of the x axis for each member is indicated by drawing an arrow along each
member on the line diagram of the structure. For example, this figure indicates that the origin of the local
coordinate system for member 2 is located at its end connected to joint 2, with the x1 axis directed from
joint 2 to joint 4. The joint to which the member end with the origin of the local coordinate system is
connected is referred to as the beginning joint for the member, whereas the joint adjacent to the opposite
end of the member is termed the end joint. For example, in Fig. 13.1 (b), member 1 begins at joint 1 and
ends at joint 2, whereas member 2 begins at joint 2 and ends at joint 4, and so on. Once the local x axis is
defined for a member, the corresponding y axis can be established by applying the right-hand rule. That
is, for each member, if we curl the fingers of our right hand from the direction of the x axis toward the
direction of the corresponding y axis, then our extended thumb points through the plane of the page,
which is the positive direction of the global z axis.
x 2 (Δ1 , Δ 2 , Δ 3 ) 4 (Δ 4 , Δ 5 , Δ 6 )
1
① ② 5 (Δ 4 , Δ 5 , Δ 7 )
③ ④
y
3 6
(a) (b)
Fig.13.1 Discretization of a structure
(a) actual structure; (b) analytical model and unknown displacements
(3) Unknown displacements or nodal displacement vector

The unknown displacements of a structure are the independent joint displacements (translations and
rotations) that are necessary to specify the deformed shape of the structure when subjected to an arbitrary
loading. Unlike in the case of the classical methods of analysis considered previously, it is usually not
necessary to neglect member axial deformations when analyzing frames by the matrix displacement method.
Consider again the plane frame of Fig. 13.1 (a). From Fig.13.1 (b), we can see that joint 1, which is
attached to the fixed support, can neither translate nor rotate; therefore, it does not have any unknown
displacement. Since joint or node 2 of the frame is not attached to a support, three displacements—the
translations Δ1 and Δ 2 in the x and y directions, respectively, and the rotation Δ 3 about the z
axis—are needed to completely specify its deformed position. Thus node 2 has three unknown
displacements. Similarly, nodes 4 and 5, which are connected to a free hinged joint, have three unknown
displacements each. However, since node 4 and node 5 have the same translations but different rotations,
their translations are numbered in the same symbols, Δ 4 and Δ 5 . Finally, joint 3 and 6, which are
attached to the fixed supports, can neither translate nor rotate; therefore, they do not have any displacement.
Thus, the entire frame has a total of 7 unknown displacements. As shown in Fig. 13.1(b), the basic
unknown displacements are defined relative to the global coordinate system, with joint translations
considered as positive when in the positive directions of the x and y axes and joint rotations
considered as positive when clockwise. The unknown displacements of the frame can be collectively
written in matrix form as
⎧ Δ1 ⎫
⎪Δ ⎪
{Δ} = ⎨ #2 ⎬ (13-1)
⎪ ⎪
⎩Δ 7 ⎭
In which, {Δ} is termed the nodal displacement vector of the structure.
As indicated in Fig.13.1 (b), the unknown displacements are numbered by starting at the lowest node
number and proceeding sequentially to the highest node number. In the case of more than one unknown
displacements at a node, the translation in the x direction is numbered first, followed by the translation in
the y direction, and then the rotation.
(a)
1 2 (Δ ) 3 (Δ , Δ ) 5(Δ )
(b) 1 2 3
5
① ② 4 (Δ , Δ ) ③
2 4
Fig.13.2 Discretization of a continuous beam

(a) actual continuous beam;
(b) analytical model and unknown displacements
In multispan or continuous beams subjected to lateral loads, the axial deformations of members are
zero. Therefore, it is not necessary to consider the joint displacements in the direction of the beam's
centroidal axis in the analysis. Thus a joint of a plane multispan or continuous beam can have up to two
unknown displacements, namely, a translation perpendicular to the beam's centroidal axis and a rotation.
For example, the two-span beam of Fig. 13.2(a) has 5 unknown displacements, the translation of node 3
and rotations of nodes 2，3 and 4, as shown in Fig. 12.2(b).
Since the joints of trusses are assumed to be frictionless hinges, they are not subjected to moments;
therefore, their rotations are zero. Thus, when analyzing plane trusses, only 2 unknown displacements,
namely, translations in the global x and y directions need to be considered for each joint or node. For
example, the truss of Fig. 13.3 (a) has 12 unknown displacements, as shown in Fig.13.3 (b).
5(Δ 7 , Δ 8 )
(a) (b)
⑥ ⑩
3(Δ 3 , Δ 4 ) 7(Δ11 , Δ12 )
⑦ ⑨
② ⑤ ⑬
1 ③ ④ ⑪ 8
① 2 (Δ1 , Δ 2 ) 4(Δ 5 , Δ 6 ) ⑧ 6(Δ9 , Δ10 ) ⑫
Fig.13.3 Discretization of a truss

(a) actual truss; (b) analytical model and unknown displacements
(4) Sign convention of member end displacements and forces

Fig.13.4 shows an arbitrary prismatic flexural member (or element) ⓔ of a structure. When the
structure is subjected to external loads, the member deforms and internal forces are induced at its ends. As
indicated in this figure, three displacements—translations in the x and y directions and rotation about
the z axis—are needed to completely specify the deformed position of each end of the member. Thus the
member has a total of six end displacements or unknown displacements. As shown in the figure, the
member end displacements are denoted by u1 , v1 , θ1 and u2 , v2 , θ 2 [ Fig.13.4 (a)], and the
θ θ 2
u 1 1
ⓔ 2 u 2
1
x
v
1 v
(a)
2
M
X 1 M 1 ⓔ
2
2 X
1 2
x
Y
1 Y 2
(b)
y
Fig.13.4 Member end displacements and forces

(a) member end displacements; (b) member end forces
corresponding member end forces are denoted by X 1 , Y1 , M 1 and X 2 , Y2 , M 2 . Note that these end
displacements and forces are defined relative to the local coordinate system of the member, with
translations and forces considered as positive when in the positive directions of the local x and y axes,
and rotations and moments considered as positive when clockwise. As indicated in Fig.13.4, the member
end displacements and forces are numbered by beginning at the member end 1, where the origin of the local
coordinate system is located, with the translation and force in the x direction numbered first, followed by
the translation and force in the y direction, and then the rotation and moment. The displacements and
forces at the opposite end 2 of the member are then numbered in the same sequential order.
In matrix displacement method, the six member end displacements and six member end forces are
arranged, in the order starting from the beginning end and ending in the ending end, in matrix form named
member end displacement vector and member end force vector, respectively. They can be written as
ⓔ ⓔ ⓔ
⎧Δ1 ⎫ ⎧ F1 ⎫
ⓔ
⎧u1 ⎫ ⎧ X1 ⎫
⎪v ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎪ ⎪
1 ⎪Δ 2 ⎪ ⎪Y1 ⎪ ⎪ F2 ⎪
⎪θ1 ⎪ ⎪Δ ⎪ ⎪M ⎪ ⎪F ⎪
⎪ ⎪ ⎪⎪ 3 ⎪⎪ ⎪⎪ 1 ⎪⎪ ⎪⎪ ⎪⎪
3
{Δ} {F }
ⓔ ⓔ
= ⎨" ⎬ = ⎨" ⎬ = ⎨" ⎬ = ⎨" ⎬ (13-2)
⎪u ⎪ ⎪Δ ⎪ ⎪X ⎪ ⎪F ⎪
⎪ 2⎪ ⎪ 4⎪ ⎪ 2⎪ ⎪ 4⎪
⎪v2 ⎪ ⎪Δ 5 ⎪ ⎪Y2 ⎪ ⎪ F5 ⎪
⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎩θ 2 ⎭ ⎩⎪Δ 6 ⎭⎪ ⎩⎪ M 2 ⎭⎪ ⎪⎩ F6 ⎪⎭
The bars over the displacement and force symbols identify that the associated physical quantities are
defined in local coordinates.
13.2 Member Stiffness Matrix in Local Coordinate System
In the matrix displacement method of analysis, the unknown displacements of a structure are
determined by solving a system of simultaneous equations, which are expressed in the matrix form
{P} = [ K ]{Δ} (13-3)
In which {Δ} denotes the unknown displacement vector, as discussed previously; {P} represents
the effects of external loads at the joints of the structure; and {K } is named the structure stiffness matrix.
As discussed in the latter sections, the stiffness matrix for the entire structure, {K } , is obtained by
assembling the stiffness matrices for the individual members of the structure. The stiffness matrix for a
member can be expressed as
{F } = [k ] {Δ}
ⓔ ⓔ ⓔ
ⓔ
In which [k ] represents the stiffness matrix for a member in global coordinates. In local coordinates, the
stiffness matrix for a member can be expressed as
{F } {Δ}
ⓔ ⓔ ⓔ
= ⎡⎣ k ⎤⎦ (13-4)
The bars over symbols in Eq. (13-4) denote that the corresponding quantities are in local coordinates.
The member stiffness matrix is used to express the forces at the ends of the member as functions of the
displacements of the member's ends. Note that the terms forces and displacements are used here in the
general sense to include moments and rotations, respectively. In this section, we derive stiffness matrices
for the members of plane rigid frames, continuous beams, and plane trusses in the local coordinate systems
of the members.
13.2.1 Stiffness matrix for a general flexural member
To establish the stiffness relationships for a general flexural member of a structure, let us focus our
attention on an arbitrary prismatic member with the length l, flexural rigidity EI and axial rigidity EA,
respectively, as shown in Fig.13.5. Our objective here is to determine the relationships between the member
end forces and end displacements. Such relationships can be conveniently established by subjecting the
member, separately, to each of the six end displacements and by expressing the total member end forces as
the algebraic sums of the end forces required to cause the individual end displacements. If we neglect the
coupling influence between axial deformation and bending deformation, the stiffness relationships pertinent
to axial deformation and those pertinent to bending deformation can be derived separately.
1 ⓔ 2 x
E , I , A, l
M
1
y
ν
θ
2
M
ν
1 2
X 1
θ X 2
2
u
1 u
Y1
2
Y
2
Fig.13.5 A general flexural prismatic member
First, let us determine the relationships between the member end axial displacements and member end
axial forces. From Fig.13.5, if neglecting the coupling influence between axial deformation and bending
deformation and recalling from strength of materials, we can see that
EA ⎫
X1 = (u1 − u2 ) ⎪
l ⎪
⎬ (13-5)
EA
X2 = − (u1 − u2 ) ⎪
l ⎪⎭
Note that the member end axial force X 2 is obtained by applying the equation of equilibrium in the x
direction.
Next, let us determine the relationships between the bending deformations and forces. In fact, the
equations between member end displacements and member end shears and bending moments have already
determined in chapter 11 (displacement method). By using the shape constants of single-span indeterminate
prismatic members shown in table 11-1 and principle of superposition, we can conveniently obtain
12 EI 6 EI 12 EI 6 EI ⎫
Y1 = 3
ν 1 + 2 θ1 − 3 ν 2 + 2 θ 2 ⎪
l l l l
⎪
6 EI 4 EI 6 EI 2 EI
M1 = 2 ν1 + θ1 − 2 ν 2 + θ2 ⎪
l l l l ⎪
⎬ (13-6)
12 EI 6 EI 12 EI 6 EI
Y2 = − 3 ν 1 − 2 θ1 + 3 ν 2 − 2 θ 2 ⎪
l l l l ⎪
6 EI 2 EI 6 EI 4 EI ⎪
M 2 = 2 ν1 + θ1 − 2 ν 2 + θ2 ⎪
l l l l ⎭
Note that the signs of bending moments M 1 and M 2 have been adopted the sign convention of this
chapter, that is, they are positive when clockwise.
Combining equation (13-5) and (13-6), we can write them in matrix form as
⎧ X1 ⎫
ⓔ
⎡ EA EA ⎤ ⓔ
⎢ 0 0 − 0 0 ⎥ ⎧u1 ⎫
⎪ ⎪
⎥⎪ ⎪
l l
⎪ ⎪ ⎢
⎢ 0 12 EI 6 EI 12 EI 6 EI ⎥ ⎪ ⎪
⎪Y1 ⎪ 0 − 3 ⎪v1 ⎪
⎪ ⎪ ⎢ l3 l2 l l2 ⎥ ⎪ ⎪
⎪ ⎪ ⎢ ⎥
⎢ 0 6 EI 4 EI 6 EI 2 EI ⎥ ⎪ ⎪
⎪M1 ⎪
⎢
0 − 2 ⎪θ ⎪
⎪ ⎪ l2 l l l ⎥⎪ 1 ⎪
⎨" ⎬ = ⎢ EA EA ⎥ ⎨" ⎬ (13-7)
⎪X ⎪ ⎢− 0 0 0 0 ⎥ ⎪u2 ⎪
⎪ 2⎪ ⎢ l l ⎥⎪ ⎪
⎪ ⎪ ⎢ 12 EI 6 EI 12 EI 6 EI ⎥ ⎪ ⎪
⎪Y ⎪ ⎢ 0 − 3 − 2 0 − 2 ⎥ ⎪v2 ⎪
⎪ ⎪2
⎢ l l l3 l ⎥⎪ ⎪
⎪ ⎪ ⎢ 6 EI 2 EI 6 EI 4 EI ⎥ ⎪ ⎪
⎪M ⎪ ⎢ 0 0 − 2 ⎥ ⎪θ ⎪
⎩ 2⎭ ⎣ l 2
l l l ⎦⎩ 2⎭
The equation can be symbolically written as

{F } {Δ}
ⓔ ⓔ ⓔ
= ⎣⎡ k ⎦⎤
The above equation is the stiffness equations (or relationships) for a general flexural prismatic member. In
is referred to as the element (or member) stiffness matrix, which is a 6 × 6 square matrix.
ⓔ
which, [ k ]
It is written again as follows.
⎡ EA EA ⎤
⎢ 0 0 − 0 0 ⎥
l l
⎢ ⎥
⎢ 12 EI 6 EI 12 EI 6 EI ⎥
⎡ k11 k16 ⎤ ⎢ 0 0 − 3
l2 ⎥
k12 k13 k14 k15
⎢ ⎥ ⎢ l3 l2 l
⎥
⎢ k21 k22 k23 k24 k25 k26 ⎥
⎢ 0 6 EI 4 EI 6 EI 2 EI ⎥
⎢ k31 0 − 2
ⓔ k32 k33 k34 k35 k36 ⎥ ⎢ l2 l l l ⎥
⎡⎣ k ⎤⎦ = ⎢ ⎥=⎢ ⎥ (13-8)
⎢ k41 k42 k43 k44 k45 k46 ⎥ ⎢
−
EA
0 0
EA
0 0 ⎥
⎢k k52 k53 k54 k55 k56 ⎥⎥ ⎢ l l ⎥
⎢ 51 ⎢
⎢ k61 k66 ⎥ ⎢ 0 12 EI 6 EI 12 EI 6 EI ⎥
⎣
k62 k63 k64 k65
⎦ − 3 − 2 0 − 2 ⎥
⎢ l l l3 l ⎥
⎢ 6 EI 2 EI 6 EI 4 EI ⎥
⎢ 0 0 − 2 ⎥
⎣ l2 l l l ⎦
13.2.2 Property of element stiffness matrix
(1) Meanings of stiffness coefficients

ⓔ
The element kij in matrix [ k ] represents the force at the location and in the direction of Fi
required, along with other end forces, to cause a unit value of the displacement Δ j while all other end
ⓔ
displacements are zero. These forces per unit displacement in matrix [ k ] are referred to as stiffness
coefficients. Note that a double-subscript notation is used for stiffness coefficients, with the first subscript
identifying the force and the second subscript identifying the displacement. For instance, k63 represents
the moment at the location and in the direction of F6 ( M 2 ) required, along with other end forces, to
cause a unit value of the displacement Δ 3 = 1 ( θ1 = 1 ) while all other end displacements are zero.
ⓔ
The ith column in matrix [ k ] denotes the end forces required to cause a unit value of the
displacement Δ i while all other displacements are zero. For example, the second column of [ k ]
ⓔ
denotes the six end forces required to cause the displacement Δ 2 = 1 ( v1 = 1 ).

(2) Symmetry of element stiffness matrix
ⓔ
The element stiffness matrix [ k ] is symmetric with respect to the main diagonal, that is
kij = k ji i≠ j (13-9)
In fact, this property is the reflection of the law of reciprocal reactions (see subsection 9.8.3).
(3) Singularity of element stiffness matrix

ⓔ
The element stiffness matrix [ k ] is singular because the determinant of the matrix is equal to zero.
That is,
[k ]ⓔ = 0 (13-10)
ⓔ ⓔ
Since the matrix [ k ] is singular, there exists no reciprocal matrix of [ k ] . In other words, by
equation (13-4), the end forces of the member can be uniquely determined by the given end displacements
[Δ ]ⓔ . But under the action of a given end force [ F ]ⓔ , the end displacements [ Δ]ⓔ cannot be uniquely
determined.
13.2.3 Special elements
Equation (13-7) is the stiffness equation for a general flexural member of a structure, in which the six
end displacements of the member are unknowns. However, in some cases, one or some of the end
displacement of a member is designated or small enough to be neglected. For instance, the axial
deformation effect may be neglected in the analysis of rigid frames and continuous beams. For this kind of
special member, the stiffness relationships may be obtained by modifying the stiffness relations of a general
element, i.e., equation (13-7).
(1) Stiffness equation for an element without considering axial deformation
In the analysis of a rigid frame, the axial deformations of the beams and columns are often be
neglected because they are to small comparing with their bending deformations. So the unknown end
displacements for one of the beams or columns remain only four, i.e., v1 , θ1 and v2 , θ 2 ; the axial
displacements vanish as shown in Fig.13.6 (a). That is,
u1 = u2 = 0 (a)
(a) (b)
ⓔ EI M 1
1 2 θ 2 M
l 1 ⓔ EI 2
2
1
ν
M u =0 θ u =0
ν
1
u =0 1 1 l 2
2
v =0 v =0
1
u =0 θ M
ν −ν
2
2
1
θ
1
2
1
Y 1
2
Y
2
Fig.13.6 Special elements

(a) beam element with no axial deformation; (b) continuous beam element
Substituting equation (a) into equation (13-7), we automatically obtain
⎧Y1 ⎫
ⓔ
⎡ 12 EI 6 EI 12 EI 6 EI ⎤ ⎧ v ⎫ⓔ
⎢ l3 −
l2 ⎥ ⎪ ⎪
1
⎪ ⎪ l2 l3
⎪ ⎪ ⎢ ⎥
⎪ M1 ⎪ ⎢ 6 EI 4 EI 6 EI 2 EI ⎥ ⎪ ⎪
− 2 ⎪θ ⎪
⎪ ⎪ ⎢ l2 l l l ⎥⎪ ⎪
⎨ " ⎬ = ⎢ 12 EI 6 EI 12 EI
⎥ "
12 EI ⎥ ⎨⎪ ⎬⎪
(13-11)
⎪Y ⎪ ⎢− − 2 − 2 v
⎪ 2 ⎪ ⎢ l3 l l3 l ⎥⎪ 2⎪
⎪ ⎪ ⎢ 6 EI 2 EI 6 EI 4 EI ⎥ ⎪ ⎪
⎪ ⎪ ⎢ 2 − 2 ⎥⎪ ⎪
⎩M 2 ⎭ ⎢⎣ l l l l ⎥⎦ ⎩θ ⎭
Equation (13-11) is the stiffness equation of an element without considering axial deformation, whose
stiffness matrix can be written as
⎡ 12 EI 6 EI 12 EI 6 EI ⎤
⎢ l3 −
l2 l3 l2 ⎥
⎢ ⎥
⎢ 6 EI 4 EI 6 EI 2 EI ⎥
− 2
ⓔ ⎢ l2 l l l ⎥
⎡⎣ k ⎦⎤ = ⎢ ⎥ (13-12)
⎢ − 12 EI 6 EI
− 2
12 EI
−
12 EI ⎥
⎢ l3 l l3 l2 ⎥
⎢ 6 EI 2 EI 6 EI 4 EI ⎥
⎢ 2 − 2 ⎥
⎣⎢ l l l l ⎦⎥
As a matter of fact, the special stiffness matrix can be obtained by deleting the 1st rows and columns and 4th
rows and columns from the stiffness matrix (13-8), respectively.
(2) Stiffness equation for an element of continuous beams
Since the axial deformations of the members of continuous beams subjected to lateral loads are
considered to be equal to zero and the beams are supported on the rigidity supports, we do not need to
consider the end displacements in the axial and lateral directions of the members. Thusly, only two
unknown end rotations θ1 and θ 2 of a member as shown in Fig.13.6 (b) are needed to take into
consideration; the other four end displacements can be recognized to be zero. That is,
u1 = v1 = u2 = v2 = 0 (b)
Substituting equation (b) into equation (13-7), we will obtain stiffness equation for an element of
continuous beams as follows.
⎧ M1 ⎫
ⓔ
⎡ 4 EI 2 EI ⎤ ⎧θ ⎫ⓔ
⎪ ⎪ ⎢ l l ⎥⎪ ⎪
1
⎨ ⎬ =⎢ ⎥⎨ ⎬ (13-13)
⎪M ⎪ ⎢ 2 EI 4 EI ⎥ ⎪ ⎪
⎩ 2⎭ θ
⎣⎢ l l ⎦⎥ ⎩ 2 ⎭
Here, the stiffness matrix for an element of continuous beams will be
⎡ 4 EI 2 EI ⎤
ⓔ⎢ l ⎥
⎡⎣ k ⎤⎦ = ⎢ l ⎥ (13-14)
⎢ 2 EI 4 EI ⎥
⎢⎣ l l ⎥⎦
It can be seen from equation (13-14) that since we have introduced the support conditions in formulating
the stiffness matrix of an element of continuous beams, the stiffness matrix is reversible. Therefore, the end
forces [ F ] of the member can be uniquely determined by the given end displacements [ Δ ] ; on the
ⓔ ⓔ
contrary, the end displacements [ Δ ] can be uniquely determined by the given end forces [ F ]
ⓔ ⓔ
as well.
(3) Stiffness equation for an element of truss members
A member of a truss is subjected to only axial forces, which can be determined from its end
displacements in the direction of the centroidal axis of the member. Thus only two unknown end
displacements u1 and u2 of a member are needed to take into consideration; the other four end
displacements can be recognized to be zero as shown in Fig.13.7. However, in order to facilitate the global
X 1
1 ⓔ EA 2 X 2
u1
u 2
l
v =0 Y v =0
1
1
Y2
2
Fig.13.7 Truss element
analysis, the stiffness equation for a truss member can be written as

ⓔ
⎧ X1 ⎫ ⎡ EA ⎤ u
ⓔ
0⎥ ⎧ 1 ⎫
EA
⎪ ⎪ ⎢ l 0 −
l ⎪ ⎪
⎪Y ⎪ ⎢ ⎥ ⎪v ⎪
⎪ 1 ⎪ ⎢ 0 0 0 0⎥ ⎪ 1 ⎪
⎨ ⎬ =⎢ ⎥⎨ ⎬
(13-15)
⎪X2 ⎪ EA EA
⎢− 0 0 ⎥ ⎪u2 ⎪
⎪ ⎪ ⎢ l l ⎥⎪ ⎪
⎪Y ⎪ ⎢ 0 ⎥⎦ ⎪⎩v2 ⎪⎭
⎩ 2 ⎭ ⎣ 0 0 0
Thus, the stiffness matrix for an element of a truss will be
⎡ EA EA ⎤
⎢ l 0 − 0⎥
l
⎢ ⎥
0 0 0 0⎥
⎡⎣ k ⎤⎦ = ⎢
ⓔ
(13-16)
⎢ EA EA ⎥
⎢− 0 0⎥
⎢ l l ⎥
⎢⎣ 0 0 0 0 ⎥⎦
The forgoing stiffness matrix for truss members can be also obtained by deleting rows 3 and 6 and columns
3 and 6 and by letting the elements in rows 2 and 5 and columns 2 and 5 to be zero, from the general
stiffness matrix (13-8), respectively.
13.3 Member Stiffness Matrix in Global Coordinate System
In last section, we formulate the stiffness matrix of an element in its local coordinate system for the
purpose of simplicity. However, for a bit complex structure, the members of the structure are oriented in
different directions. Thusly, it becomes necessary to transform the stiffness relations for each member from
the member’s local coordinate system (denoted by x y ) to a common global coordinate system (denoted
by x y ). The member stiffness relations in global coordinates thus obtained are then assembled to establish
the stiffness relations for the entire structure.
ⓔ
In order to derive the stiffness matrix in global system [k ] of an element, we use the method of
coordinate transformation. First, we discuss the transformation of member end forces from local to global
coordinates or vice versa so as to obtain the member transformation matrix; then we implement the
transformation of member stiffness matrix from local to global coordinate system, and vice versa.
13.3.1 Element transformation matrix of coordinates
Consider an arbitrary general member ⓔ as shown in Fig. l3.8 (a). The orientation of the member
with respect to the global x y coordinate system is defined by an angle α measured clockwise from the
positive direction of the global x axis to the positive direction of the local x axis, as shown in the figure.
ⓔ
The stiffness relations derived in the preceding section are valid only for member end forces [ F ] and
end displacements [ Δ ]
ⓔ
described with reference to the local x y coordinate system of the member, as
shown in the figure.
Now, suppose that the member end forces and end displacements are specified relative to the
global x y coordinate system [Fig.13.8 (b)] and we wish to determine the relations between the end forces
in the two coordinate systems. As shown in Fig.13.8 (b), the member end forces in global coordinates are
denoted by X 1 , Y1 , M 1 and X 2 , Y2 , M 2 . These global member end forces are numbered by

beginning at the member end 1, where the origin of the local coordinate system is located, with the force in
the x direction numbered first, followed by the force in the y direction and then the moment. The forces at
the opposite end 2 of the member are then numbered in the same sequential order.
(a) (b)
Y 1
X 1
M
1 O 1
X α x
M O 1 α x 1 1
y ⓔ X
Y ⓔ α
2
1
2 X 2
M
2
M
α
2
2
y y Y
y Y x 2 x
2
Fig.13.8 Transformation relations of member end forces

(a) end forces in local coordinate; (b) end forces in global coordinate
By considering the projection relations of forces shown in Fig.13.8 (b) and (a), we will find the
relation between the end forces in the global coordinate system and those in the local coordinate system as
follows
X 1 = X 1 cos α + Y1 sin α ⎫
⎪
Y1 = − X 1 sin α + Y1 cos α ⎪
M1 = M1 ⎪
⎪
⎬ (13-17)
X 2 = X 2 cos α + Y2 sin α ⎪
Y2 = − X 2 sin α + Y2 cosα ⎪
⎪
M2 = M2 ⎪⎭
Equation (13-17) can be written in matrix form as

ⓔ
⎧ X1 ⎫ ⎡ cos α sin α 0 0 0 0⎤ ⎧ X1 ⎫
ⓔ
⎪ ⎪ ⎢ ⎪ ⎪
⎪Y1 ⎪ ⎢ − sin α cos α 0 0 0 0 ⎥⎥ ⎪Y1 ⎪
⎪M ⎪ ⎢ 0 0 1 0 0 0 ⎥ ⎪⎪ M 1 ⎪⎪
⎪ 1⎪
⎨ ⎬ =⎢ 0 0 0 cos α sin α
⎥
0 ⎥ ⎨⎪ X 2 ⎬⎪
(13-18)
⎪X2 ⎪ ⎢
⎪Y ⎪ ⎢ 0 0 0 − sin α cos α 0 ⎥ ⎪Y2 ⎪
⎪ 2 ⎪ ⎢ ⎥⎪ ⎪
⎪M 2 ⎪ ⎢ 0
⎣ 0 0 0 0 1 ⎦⎥ ⎪ M 2 ⎪
⎩ ⎭
⎩ ⎭
or symbolically as
{F } = [T ]{ F }
ⓔ ⓔ
(13-19)
in which
⎡ cos α sin α 0 0 0 0⎤
⎢ − sin α cos α 0 0 0 0 ⎥⎥
⎢
⎢ 0 0 1 0 0 0⎥
[T ] = ⎢ 0 0 0 cos α sin α 0 ⎥
⎥ (13-20)
⎢
⎢ 0 0 0 − sin α cos α 0 ⎥
⎢ ⎥
⎢⎣ 0 0 0 0 0 1 ⎥⎦
is referred to as the transformation matrix of coordinates. Equation (13-19) is the transformation of end
forces from global to local coordinate systems.
Like end forces, the member end displacements are vectors as well, which are defined in the same
directions as the corresponding forces. Therefore, the transformation matrix [T ] developed for the case of
end forces [Eq. (13.20)] can also be used to transform member end displacements from global to local
coordinate systems:
{Δ} = [T ]{Δ}
ⓔ ⓔ
(13-21)
Next, we determine the transformation of member end forces and end displacements from local to
global coordinates. From the projection relations of forces shown in Fig.13.8 (a) and (b), we will observe
the relation between the end forces in the local coordinate system and those in the global coordinate system
as follows
ⓔ
0 ⎤ ⎧ X1 ⎫
ⓔ
⎧ X1 ⎫ ⎡cos α − sin α 0 0 0
⎪Y ⎪ ⎢ sin α ⎪ ⎪
⎪ 1 ⎪ ⎢ cos α 0 0 0 0 ⎥⎥ ⎪Y1 ⎪
⎪⎪ M 1 ⎪⎪ ⎢ 0 0 1 0 0 0 ⎥ ⎪⎪ M 1 ⎪⎪
⎨X ⎬ = ⎢ 0 0 0 cos α
⎥
− sin α 0 ⎥ ⎨⎪ X 2 ⎬⎪
(13-22)
⎪ 2⎪ ⎢
⎪Y2 ⎪ ⎢ 0 0 0 sin α cos α 0 ⎥ ⎪Y2 ⎪
⎪ ⎪ ⎢ ⎥⎪ ⎪
⎩⎪ 2 ⎭⎪
M ⎢⎣ 0 0 0 0 0 1 ⎥⎦ ⎪ M ⎪
⎩ 2⎭
A comparison between equations (13-18) and (13-22) indicates that the transformation matrix in
equation (13-22), which transforms the end forces from local to global coordinates, is the transpose of the
transformation matrix [T ] in equation (13-18), which transforms the end forces from global to local
coordinates. Thus equation (13-22) can be written as
{F} = [T ] {F}
ⓔ T ⓔ
(13-23)
The matrix [T ]
T
can also determine the transformation of member end displacements from local to
global coordinates; that is,
{Δ} = [T ] {Δ}
ⓔ T ⓔ
(13-24)
Recall from matrix algebra that the matrix [T ] is of orthogonality. That is,
[T ][T ] = [T ] [T ] = [ I ]
T T
(13-25)
In which, [ I ] represents an identity (or unit) matrix with the same order as that of [T ] .
13.3.2 Element stiffness matrix in global coordinates
In global coordinates, the relations between the member end forces and end displacements can be also
expressed as
{F } = [k ] {Δ}
ⓔ ⓔ ⓔ
(13-26)
ⓔ
In which, [k ] is referred to as the member (or element) stiffness matrix in global coordinates.
By using the member stiffness relations in local coordinates and the transformation relations, we can
now develop the stiffness relations for members in global coordinates.
The stiffness relation of an element ⓔ in local coordinates is
{F } {Δ}
ⓔ ⓔ ⓔ
= ⎡⎣ k ⎤⎦ (a)
Substituting equations (13-19) and (13-21) into equation (a), we will obtain
[T ]{F }ⓔ = ⎣⎡ k ⎦⎤ [T ]{Δ}ⓔ
ⓔ
(b)
Premultiplying [T ] to the both sides of equation (b) and in the meanwhile introducing equation (13-25),
T
we will write
{F } = [T ] ⎣⎡ k ⎦⎤ [T ]{Δ}ⓔ
ⓔ T ⓔ
(c)
A comparison between equations (c) and (13-26), we find

[k ] = [T ] ⎡⎣ k ⎤⎦ [T ]
ⓔ T ⓔ
(13-27)
The forgoing equation is just the transformation relations of element stiffness matrices in the two
ⓔ
coordinates. By the equation, if the stiffness matrix [ k ] in local coordinates and the transformation
matrix [T ] between the local and global coordinates of a member are determined, the stiffness matrix
[k ]ⓔ in global coordinates will be conveniently obtained.
ⓔ
It can be seen from above that the element stiffness matrix [k ] referred to global coordinates not
ⓔ
only has the same order as the stiffness matrix [ k ] referred to local coordinates, but also has the similar
ⓔ
properties with [ k ] . The properties are:
ⓔ
(1) The element kij in the matrix [k ] is the force corresponding to coordinate i due to a unit
displacement of coordinate j , which is also termed stiffness influence coefficient.
(2) [k ]ⓔ is a symmetric matrix.
ⓔ
(3) The stiffness matrix [k ] for a general member is singular.
Example 13-1
ⓔ
Determine the stiffness matrix [k ] for the members of the rigid frame shown in Fig.13.9. Assume
both the length and the size of the cross section for each member are the same. They are: length l = 5m ,
axial rigidity EA = 1.5 ×10 kN and flexural rigidity EI = 1.25 ×10 kN ⋅ m .
7 6 2
Solution
ⓔ
(1) Stiffness matrix in local coordinates [ k ]
The positive direction of the axis of abscissas x for each member is indicated by an arrow along the
axis. Since member 1 and member 2 have the same size, their stiffness matrices in local coordinates are
same too, i.e., [k ]
①
= [k ]② . By substituting the length l = 5m , axial rigidity EA = 1.5 ×107 kN and
flexural rigidity EI = 1.25 ×10 kN ⋅ m into equation (13-8), we obtain
6 2
⎡ 300 0 0 −300 0 0⎤
⎢ 0 12 30 0 −12 30 ⎥⎥
⎢
① ② ⎢ 0 30 100 0 −30 50 ⎥
⎡k ⎤ = ⎡k ⎤ = ⎢ ⎥ ×10
4
⎣ ⎦ ⎣ ⎦
⎢ −300 0 0 300 0 0 ⎥
⎢ 0 −12 −30 0 12 −30 ⎥
⎢ ⎥
⎢⎣ 0 30 50 0 −30 100 ⎥⎦
ⓔ
(2) Stiffness matrix in global coordinates [k ]
Element ① : since α = 0 , cos α = 1 , sin α = 0 . Therefore, [T ] = [ I ] , i.e., no coordinate
transformation is needed. Thusly,
[k ]
① ①
= ⎡⎣ k ⎤⎦
①
E , I , A, l x
② E , I , A, l
l = 5m
l = 5m
y
Fig.13.9 The figure of example 13-1
Element ② : since α = 900 , cos α = 0 , sin α = 1 . Thusly, the transformation matrix of element
② will be
⎡0 1 0 0 0 0⎤
⎢ −1 0 0 0 0 0 ⎥⎥
⎢
⎢0 0 1 0 0 0⎥
[T ] = ⎢ 0 0 0 0 1 0⎥
⎥
⎢
⎢0 0 0 −1 0 0⎥
⎢ ⎥
⎢⎣ 0 0 0 0 0 1 ⎥⎦
By applying equation (13-27), we can determine the stiffness matrix in global coordinates for member
② as
⎡ 12 0 −30 −12 0 −30 ⎤
⎢ 0 300 0 0 −300 0 ⎥⎥
⎢
⎢ −30 0 100 30 0 50 ⎥
[k ] = [T ] ⎡⎣ k ⎤⎦ [T ] = ⎢
② T ②
⎥ × 10
4
⎢ −12 0 30 12 0 30 ⎥
⎢ 0 −300 0 0 300 0 ⎥
⎢ ⎥
⎢⎣ −30 0 50 30 0 100 ⎥⎦
13.4 Global Stiffness Matrices of Continuous Beams 513
13.4 Global Stiffness Matrices of Continuous Beams
In previous sections, we have discussed the member analysis, involving with the establishment of the
stiffness relationships and development of stiffness matrices for the members of a structure. Beginning
from this section, we will discuss the global (or structure) analysis, comprising the establishment of the
stiffness relationships and development of stiffness matrix for an entire structure. The section will discuss
the development of global stiffness matrix for continuous beams.
The stiffness relations of a structure, which are used to express the joint forces as functions of the joint
displacements of the structure by the structure (or global) stiffness matrix, can be established by
displacement method. Generally, there are two kinds of methods to develop the stiffness relations of a
structure; one is the classical displacement method discussed in chapter 11, another is the method so called
element assembly method (or stiffness assembly method or direct stiffness method). One of the advantages
of element assembly method is that it can be easily implemented on computers.
(a) ① ②
EI1 EI 2
i1 = i2 =
l1 l2
(b)
F1 F2 Δ3 F3
① ②
Δ1 Δ2
Fig.13.10 Global analysis of a continuous beam

(a) original structure; (b) primary system in displacement method
In order to make a comparison between the two methods, let us simply review the procedure of the
development of stiffness relations of traditional displacement method through the following example.
For the continuous beam shown in Fig.13.10 (a), its primary system used in displacement method is
shown in Fig.13.10 (b). As discussed in chapter 11, the primary unknowns of the beam are the joint
rotations, Δ1 , Δ 2 and Δ 3 , which consist of the joint displacement vector {Δ} of the beam, that is,
{Δ} = [ Δ1 Δ 2 Δ3 ]
T
The joint forces, corresponding to the joint displacements Δ1 , Δ 2 and Δ 3 , are the restraint moments,
F1 , F2 and F3 , which compose of the joint force vector {F } of the beam. It can be written as
{F } = [ F1 F2 F3 ]
T
In traditional displacement method, the restraint moments F1 , F2 and F3 can be obtained by

superposing each of the restraint forces due to the joint displacements Δ1 , Δ 2 and Δ 3 individually as
shown in Fig.13.11 (a) through (c). The results of superposition can be written as following matrix form.
⎧ F1 ⎫ ⎡ 4i1 2i1 0 ⎤ ⎧ Δ1 ⎫
⎪ ⎪ ⎢ ⎪ ⎪
⎨ F2 ⎬ = ⎢ 2i1 4i1 + 4i2 2i2 ⎥⎥ ⎨Δ 2 ⎬ (13-28)
⎪F ⎪ ⎢ 0 2i2 4i2 ⎥⎦ ⎩⎪ Δ 3 ⎭⎪
⎩ 3⎭ ⎣
Or symbolically as
{F } = [ K ]{Δ} (13-29)
4i1Δ1 2i1Δ1 0
(a) ① ②
Δ1
2i1Δ 2 ( 4i
1
+ 4i2 ) Δ 2 2i2 Δ 2
(b) ②
①
Δ 2
0 2i2 Δ 3 4i2 Δ 3
Δ 3
(c)
① ②
Fig.13.11 Joint moments due to joint rotations

(a) joint moments due to Δ1 ; (b) joint moments due to Δ 2 ;
(c) joint moments due to Δ 3
Equation (13-28) or equation (13-29) is referred to as the structure (or global) stiffness relations (or
functions) of the continuous beam, which represents the transformation relationship between the joint
displacements and forces. The relationship is transformed by the matrix [ K ] , termed as structure (or
global) stiffness matrix of the continuous beam. It can be expressed as
⎡ 4i1 2i1 0⎤
[ K ] = ⎢⎢ 2i1 4i1 + 4i2 2i2 ⎥⎥ (13-30)
⎢⎣ 0 2i2 4i2 ⎥⎦
The following subsection will introduce element stiffness assembly method so as to form the global
stiffness matrix of a structure.
13.4.1 Concept of element assembly method
It can be seen from above discussion that when determining the global stiffness equation (13-29) of a
structure by traditional displacement method, the contribution to the joint force vector {F } of each joint
displacement of the structure was considered individually as shown in Fig.13.11; then the individual
contributions were algebraically added so as to form the joint force vector.
If the global stiffness equation (13-29) of a structure is formed by using element assembly method, the
contribution to the joint force vector {F } of each element (or member) of the structure was considered
individually; then the individual contributions will be assembled in the sense of matrix algebra so as to
form the joint force vector, which is the stiffness relations (or equations) between joint displacements and
forces of the structure. One of the characteristics of the method is that the global stiffness equations of a
structure can be directly assembled by the global stiffness equations of its members.
In fact, once the member stiffness equations in global coordinates have been determined, the stiffness
equations for the entire structure can be established by writing equilibrium equations for the joints of the
structure and by applying the compatibility conditions that the displacements of the member ends rigidly
connected to joints must be the same as the corresponding joint displacements.
To illustrate this procedure, again consider the two-span continuous beam shown in Fig.13.12 (a),
which indicates that the beam has three joint displacements Δ1 , Δ 2 , Δ 3 and three joint forces F1 , F2 ,
F3 . The continuous beam is divided into two elements, element ① and element ② as shown in
Fig.13.12 (b), whose numbers are enclosed within circles. There are two kinds of numbers in the figure:
One is global number such as joint numbers 1, 2 and 3 shown in Fig.13.12 (a); another is local number
which is enclosed within a bracket such as (1) and (2) shown in Fig.13.12 (b).
(1) Element stiffness equations
For the elements shown in Fig.13.12 (b), the element stiffness equations for elements ① and
element ② can be written as follows.
Element ① :
① ①
⎧ F(1) ⎫ ⎡ 4i1 2i1 ⎤ ⎧ Δ (1) ⎫
⎨ ⎬ =⎢ ⎨ ⎬ (a)
⎩ F(2) ⎭ ⎣ 2i1 4i1 ⎥⎦ ⎩Δ (2) ⎭
Element ② :
② ②
⎧ F(1) ⎫ ⎡ 4i2 2i2 ⎤ ⎧ Δ (1) ⎫
⎨ ⎬ =⎢ ⎨ ⎬ (b)
⎩ F(2) ⎭ ⎣ 2i2 4i2 ⎥⎦ ⎩Δ (2) ⎭
Here, F(1) , F(2) indicate the member end forces of element ⓘ ; Δ (1) , Δ (2) represent the member end
ⓘ ⓘ ⓘ ⓘ
displacements (or rotations) of element ⓘ

Note that for the elements of a continuous beam, their local and global coordinates are coincident, so
the element stiffness equations in their local and global coordinates remain unchangeable.
1
F1 F2 Δ 3
F3
(a) 2 3
Δ 1 Δ 2
F(1①)
F(1②)
Δ①
(2) ① Δ②
(2) F(2②)
(b) ①
F(2)
②
Δ① Δ②
(1)
(1)
F1 F3
(c) F2
① 2 3
1 F(1) ①
F(2) F(1②) F(2②)
Fig.13.12 Global analysis of a two-span continuous beam

(a) original structure; (b) elements ① and ② ; (c) equilibrium conditions of joints
(2) Equilibrium equations

By applying the equation of equilibrium, ∑ M = 0 , to the free body of each of the joints as shown in
Fig.13.12 (c), we obtain the equilibrium equations for the joints
F1 = F(1)
①
⎫
⎪
⎪
F2 = F(2)
①
+ F(1)② ⎬ (c)
⎪
F3 = F(2)
②
⎪⎭
If we express the equilibrium relation by the joint force vector of the structure, it will be the
superposition of the member end force vectors, which is extended to have the same elements as the joint
force vector. That is,
① ②
⎧ F1 ⎫ ⎧ F(1) ⎫ ⎧ 0 ⎫
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨ F2 ⎬ = ⎨ F(2) ⎬ + ⎨ F(1) ⎬ (d)
⎪F ⎪ ⎪ 0 ⎪ ⎪F ⎪
⎩ 3⎭ ⎩ ⎭ ⎩ (2) ⎭
Or symbolically as
{F } = {F } + {F }
① ②
(13-31)
(3) Compatibility equations

As mentioned previously, when developing the stiffness equations for the entire structure not only the
equilibrium equations for the joints of the structure have to be considered but also the compatibility
conditions that the displacements of the member ends rigidly connected to joints must be the same as the
corresponding joint displacements must be taken into account. By comparing Fig.13.12 (a) and (b), we can
observe that since the left end of element ① is hinged to joint 1, the displacement of end (1) of this
element, Δ (1) , must be the same as the displacement of joint 1, Δ1 . Similarly, the displacement of end (2)
①
①
of element ① is rigidly connected to joint 2, the displacement of end (2), Δ (2) , must be the same as the
displacement of joint 2, Δ 2 . Thus, the compatibility equations for element 1 are
(1) = Δ1
Δ① (2) = Δ 2
Δ① (e)
In a similar manner, the compatibility equations for element 2 are
Δ②
(1) = Δ 2 (2) = Δ 3
Δ② (f)
By substituting the compatibility equations for element ① [Eq. (e)] into the element’s
force-displacement relation [Eq. (a)] and by extending the element force and displacement vectors and
stiffness matrix to the same orders as those of the entire structure, we express the element force vector
{F }① in terms of the joint displacement vector {Δ} as
①
⎧ F(1) ⎫ ⎡ 4i1 2i1 0 ⎤ ⎧ Δ1 ⎫
⎪ ⎪ ⎢ ⎥⎪ ⎪
⎨ F(2) ⎬ = ⎢ 2i1 4i1 0 ⎥ ⎨Δ 2 ⎬
⎪ ⎪ ⎪ ⎪
⎩ 0 ⎭ ⎣⎢ 0 0 0 ⎦⎥ ⎩ Δ 3 ⎭
Or symbolically as
{F } = [K ] {Δ}
① ①
(g)
In which,
⎡ 4i1 2i1 0 ⎤
[K ] = ⎢⎢ 2i1 4i1 0 ⎥⎥
①
⎢⎣ 0 0 0 ⎥⎦
is termed element-contribution matrix, which may be recognized as the contribution to global stiffness
matrix of the structure provided by element ① .
Similarly, for element 2, substitution of Eq. (f) into Eq. (b) yields
②
⎧ 0 ⎫ ⎡0 0 0 ⎤ ⎧ Δ1 ⎫
⎪ ⎪ ⎢ ⎪ ⎪
⎨ F(1) ⎬ = ⎢0 4i2 2i2 ⎥⎥ ⎨Δ 2 ⎬
⎪F ⎪ ⎢⎣0 2i2 4i2 ⎥⎦ ⎩⎪ Δ 3 ⎪⎭
⎩ (2) ⎭
Or symbolically as
{F } = [K ] {Δ}
② ②
(h)
In which,
⎡0 0 0⎤
[K ] = ⎢⎢ 0 4i2 2i2 ⎥⎥
②
⎢⎣ 0 2i2 4i2 ⎥⎦
is named element-contribution matrix, which may be considered as the contribution to global stiffness
matrix of the structure provided by the element ② .
(4) Global stiffness equation of a structure
Finally, by substituting equations (g) and (h) into the equilibrium equation (13-31), we obtain the
desired relationship (or equation) between the joint force vector {F } and the joint displacement vector
{Δ} of the structure as
{F } = ([ K ] + [ K ] ) {Δ}
① ②
Or
{F } = [ K ]{Δ} (13-32)
Therefore, the global stiffness matrix of the structure [ K ] will be
[K ] = [K ] + [K ] = ∑[K ]
① ② ⓔ
(13-33)
ⓔ
It can be seen that the global stiffness matrix of a structure is equal to the sum of element-contribution
matrices of all the elements of the structure.
The expanding form of equation (13-31) will be
⎧ F1 ⎫ ⎡ 4i1 2i1 0 ⎤ ⎧ Δ1 ⎫
⎪ ⎪ ⎢ ⎪ ⎪
⎨ F2 ⎬ = ⎢ 2i1 4i1 + 4i2 2i2 ⎥⎥ ⎨Δ 2 ⎬
⎪F ⎪ ⎢ 0 2i2 4i2 ⎦⎥ ⎩⎪ Δ 3 ⎭⎪
⎩ 3⎭ ⎣
Note that the above equation, derived by stiffness assembly method, is identical to equation (13-28)
that is the result derived by traditional displacement method.
It can be observed from forging discussion that the steps for determination of global stiffness matrix of
a structure by so called stiffness assembly method are two: (1) determine element-contribution matrix
[ K ]ⓔ by element stiffness matrix [k ]ⓔ ; (2) superpose all of the element-contribution matrices to form
global stiffness matrix of the structure.
We will further discuss the first step as follows.
13.4.2 Element orientation vector
In the preceding subsection, we determined the element-contribution matrix by substituting the

element compatibility equations into the element global stiffness equations. In the process, the element
ⓔ
global stiffness matrix [k ] was extended by newly arranging the elements at the new positions and
ⓔ ⓔ
complementing enough zero elements in [k ] to form element-contribution matrix [ K ] . The key point
ⓔ ⓔ
for developing [ K ] is the new positions of the elements in [k ] , i.e., where the elements in
[k ]ⓔ must be placed in [ K ]ⓔ ?
ⓔ
It can be seen from the discussion of last subsection that if we extend the matrix [k ] to the same
order as that of the global stiffness matrix [ K ] of the structure, change the element local end
displacement numbers into its corresponding joint global displacement numbers by its compatibility
equations, newly place the elements in the rows and columns corresponding to its joint global displacement
ⓔ
numbers, add the zero elements at other positions, the element-contribution matrix [ K ] will be formed.
To illustrate the technique, consider the two-span continuous beam shown in Fig.13.13. The
corresponding relations between the end local displacement numbers and its joint global displacement
numbers for each element are tabulated in table 13-1.
In order to facilitate the development of element-contribution matrix, we define the vector whose
components are composed by the joint global displacement numbers of a member, termed orientation
vector and indicated by {λ} . Then, the relationship between end local displacement numbers and its
ⓔ
joint global displacement numbers of the member will be determined by the vector. So the orientation
vector is also known as number exchange vector. The orientation vectors of the continuous beam shown in
Fig.13.13 are shown in the fourth columns of table 13-1.
1 2 ② 3
(a) ①
(1) ( 2) (1) ( 2)
(b) ① ②
Fig.13.13 Relationships between local and global numbers

(a) joint global displacement numbers; (b) end local displacement numbers
Table 13-1 relationship between end local displacement numbers and its joint global displacement numbers
End local displacement Corresponding joint global
Element Orientation vectors
numbers displacement numbers
(1) 1 ⎧1 ⎫
① {λ}① = ⎨ ⎬
(2) 2 ⎩2⎭
(1) 2 ⎧2⎫
② {λ}② = ⎨ ⎬
(2) 3 ⎩3 ⎭
The procedure for developing an element-contribution matrix may be summarized as follows:

ⓔ
(1) Set up a square matrix [ K ] that has the same order as the numbers of the joint displacements
ⓔ
of a structure and let the initial values of all the elements in [ K ] to be zero;
(2) exchange numbers between the end local displacements and joint global displacements of an
ⓔ ⓔ
element; that is, exchange the row number (i) and column number (j) of the element k( i )( j ) in [k ] for
the new row number λi and column number λ j to form a new element K λⓔλ for each of the elements
i j
ⓔ
in [k ] ;
(3)
ⓔ
Add each of the elements K λi λ j to the position where is in row λi and column λ j in
[ K ]ⓔ .
ⓔ ⓔ
By the procedure, the corresponding positions between the elements in [k ] and those in [ K ]
for the beam shown in Fig.13.13 will be as follows.
Element ① : orientation vector {λ} = [1, 2]
① T
1 2 3
(1) (2)
⎡ 4i1 2i1 0 ⎤ 1
⎡ 4i1 2i1 ⎤ (1) ⎢
[k ]
①
=⎢ ⎥ ⇒ [ K ] = ⎢ 2i1
①
4i1 0 ⎥⎥ 2
⎣ 2i1 4i1 ⎦ (2)
⎢⎣ 0 0 0 ⎥⎦ 3
Element ② : orientation vector {λ} = [2, 3]

② T
1 2 3
(1) (2)
⎡0 0 ⎤1
⎡ 4i2 2i2 ⎤ (1) ⎢
[k ]
②
=⎢ ⇒ [K ]
②
= ⎢ 0 4i2 2i2 ⎥⎥ 2
⎣ 2i2 4i2 ⎦⎥ (2)
⎣⎢ 0 2i2 4i2 ⎦⎥ 3
13.4.3 Implementation of element assembly method on computers
In subsection 13.4.1, we used two steps to determine the global (or structure) stiffness matrix [ K ] .
The two steps are: (1) substitute the element compatibility equations into the element global stiffness
ⓔ ⓔ
relations, i.e., orient the elements’ positions of [k ] in the element-contribution matrix [ K ] by
orientation vector {λ} ; (2) substitute the resulting relationships derived from step (1) into the joint
ⓔ
ⓔ
equilibrium equations, i.e., accumulate the elements in each [ K ] to form [ K ] . This process of
developing three types of matrices can be quite tedious and time consuming for large structures.
From the expansion of Eq. (13-31), we observe that the stiffness of a joint in a direction equals the
sum of the stiffnesses in that direction of all the members meeting at the joint. This fact indicates that the
global stiffness matrix [ K ] can be formulated directly by accumulating the elements of the member
global stiffness matrices into their proper positions in the global matrix, thereby avoiding the necessity of
developing element-contribution matrices. In other words, we can develop the global matrix [K ] directly
ⓔ
by the elements in [k ] . The technique of directly forming a global stiffness matrix by assembling the
elements of the member global stiffness matrices is known as element assembly method.
Put concretely, the process of directly forming a global stiffness matrix [ K ] by assembling the
ⓔ ⓔ
elements in [k ] is the process that firstly calculate the numerical value of one of the elements in [k ] ;
secondly determine the position of the element in [ K ] by orientation vector {λ} ; thirdly accumulate
ⓔ
the value to its newly determined position in [ K ] ; the same three steps are repeated for next element until
ⓔ
all of the elements in all [k ] is assembled. The process can be implemented on computers by following
steps.
(1) Develop an array representing [ K ] and let the initial value of its elements to be zero.
in [ K ] by {λ}
① ①
(2) Determine the element’s positions from [k ] and accumulate them to the
newly determined positions. By now [ K ] = [ K ] .
①
in [ K ] by {λ}
② ②
(3) Determine the element’s positions from [k ] and algebraically add them
to the elements located in the newly determined positions. By now [ K ] = [ K ] + [ K ] . The same step is
① ②
then repeated for all of the next element stiffness matrices. Finally, we obtain the global stiffness matrix
[ K ] = ∑ [ K ]ⓔ .
ⓔ
It should be mentioned again that when two or more member stiffness coefficients are located in the
same position in [ K ] , then the coefficients must be algebraically added.
Example 13-2
Determine the global stiffness matrix [ K ] for the continuous beam shown in Fig.13.14.
Solution
(a)
1 ① 2 ② 3 ③ 0
i1 i2 i3
(b)
Δ1 Δ2
1 2 3 Δ3 0 Δ0 = 0
① ② ③

(a) original structure ; (b) computing model
(1) Numbers of global joint displacements

From the computing model shown in Fig.13.14 (b), we observe that joints 1, 2 and 3 are free to rotate.
Thus the beam has 3 joint global displacements, Δ1 , Δ 2 and Δ 3 , which are unknown rotations of the
three joints.
Note that at a fixed joint, the joint displacements are zero. So we let the value of a component of joint
global displacements corresponding to one of the restraints of the joint be equal to zero.
Orientation vector {λ}
ⓔ
(2)
The vector’s components are composed by the joint global displacement numbers of the element. Then,
the three elements’ orientation vectors can be written as follows.
⎧1 ⎫ ⎧2 ⎫ ⎧3 ⎫
{λ} {λ} {λ}
① ② ③
=⎨ ⎬ =⎨ ⎬ =⎨ ⎬
⎩2 ⎭ ⎩3 ⎭ ⎩0 ⎭
(3) Development of global stiffness matrix [ K ]
Stiffness matrix of element ①
(1) (2)
⎡ 4i1 2i1 ⎤ (1)
[k ]
①
=⎢
⎣ 2i1 4i1 ⎥⎦ (2)
According to {λ} , local number (1) must be changed into global number 1; local number (2) must
①
①
be changed into global number 2. Then directly place the elements in [k ] to their corresponding positions
in [ K ] . By now, [ K ] will have the form as
1 2 3
⎡ 4i1 2i1 0 ⎤ 1
[ K ] = ⎢⎢ 2i1 4i1 0 ⎥⎥ 2
⎣⎢ 0 0 0 ⎦⎥ 3
Stiffness matrix of element ②
(1) (2)
⎡ 4i2 2i2 ⎤ (1)
[k ]
②
=⎢
⎣ 2i2 4i2 ⎥⎦ (2)
By {λ} , local number (1) must be changed into global number 2; local number (2) must be changed
②
②
into global number 3. Then directly algebraically add the elements in [k ] to elements located in their
newly determined corresponding positions in [ K ] . Then, [ K ] will have the form as
1 2 3
⎡ 4i1 2i1 0 ⎤1
[ K ] = ⎢⎢ 2i1 4i1 + 4i2 2i2 ⎥⎥ 2
⎢⎣ 0 2i2 4i2 ⎥⎦ 3
Stiffness matrix of element ③

(1) (2)
⎡ 4i3 2i3 ⎤ (1)
[k ]
③
=⎢
⎣ 2i3 4i3 ⎥⎦ (2)
In accordance with {λ} , local number (1) must be changed into global number 3; local number (2)
③
③
must be changed into global number 0. Then directly algebraically add the elements in [k ] to elements
located in their newly determined corresponding positions in [ K ] . By now, [ K ] will be obtained as
1 2 3
⎡ 4i1 2i1 0 ⎤1
[ K ] = ⎢⎢ 2i1 4i1 + 4i2 2i2 ⎥⎥ 2
⎢⎣ 0 2i2 4i2 + 4i3 ⎥⎦ 3
The matrix [ K ] is just the global stiffness matrix of the continuous beam.
13.4.4 Property of global stiffness matrix
(1) Meanings of stiffness coefficients in global stiffness matrix

The element kij in global stiffness matrix [ K ] is termed stiffness coefficient, which identifies the
force (per unit displacement) at the location and in the direction of Fi required, along with other end
forces, to cause a unit value of the displacement Δ j while all other end displacements are zero. The
element kii in the main diagonal of [ K ] is the stiffness of joint i in i direction and its magnitude
equals the sum of the stiffnesses in the direction of all the members meeting at the joint.
(2) Global stiffness matrix [ K ] is symmetric. The property may be proved by reciprocal law of
reactions presented in section 9.8.
F1 , Δ1 F2 , Δ 2 F3 , Δ3 Fn , Δ n Fn+1 , Δ n+1
1 2 3 n n +1
① ② ③ ⓝ
Fig.13.15 A continuous beam with n spans
(3) If the support conditions of a structure are introduced when determine the global stiffness matrix
[ K ] , the [ K ] will be reversible.
(4) [ K ] is a sparse and banded matrix because there are many zero elements in the matrix and the
nonzero elements in it are only clustered in a rather narrow band near the main diagonal.
The Fig.13.15 shows a continuous beam with n spans. If we use the approach discussed previously, the
stiffness equation of the structure might be obtained as follows.
It can be observed that there are many zero elements in the global stiffness matrix [ K ] , so [ K ] is
known as a sparse matrix. Since the nonzero elements in [ K ] are just distributed near the main diagonal of
[ K ] , [ K ] is also referred to as a banded matrix.
13.5 Global Stiffness Matrices of Rigid Frames 525
⎧ F1 ⎫ ⎡ 4i1 2i1 0 0 ⎤ ⎧ Δ1 ⎫
⎪ ⎪ ⎢ 4(i1 + i2 ) ⎥⎪ ⎪
⎪ F2 ⎪ ⎢ 2i1 2i2 0 ⎥ ⎪ Δ2 ⎪
⎪ F3 ⎪ ⎢ 0 2i2 4(i2 + i3 ) 2i3 0 ⎥ ⎪ Δ3 ⎪
⎪ ⎪ ⎢ ⎥⎪ ⎪
⎪ # ⎪ ⎢ % ⎥⎪ # ⎪
⎪
⎨ # ⎬=⎢
⎪
% % % ⎥ ⎪⎨ # ⎪⎬ (13-34)
⎪ # ⎪ ⎢ % %
⎥⎪ ⎪
⎪ ⎪ ⎢ ⎥⎪ # ⎪
⎪ # ⎪ ⎢ 0 % % % ⎥⎪ # ⎪
⎪F ⎪ ⎢ ⎥⎪
4(in −1 + in ) 2in ⎥ ⎪ Δ n ⎪⎪
⎪ n ⎪ ⎢ 2in −1
⎢ 4in ⎦⎥ ⎩⎪ Δ n +1 ⎪⎭
⎩⎪ Fn +1 ⎭⎪ ⎣ 0 2in
13.5 Global Stiffness Matrices of Rigid Frames
In this section, we will discuss the development of global stiffness matrix for rigid frames. Comparing
with the last section that discussed the establishment of global stiffness matrix for continuous beams, the
fundamental principle and procedure to develop the global matrix are identical. However, the technique to
develop the global stiffness matrix for a rigid frame is more complex than that for a continuous beam. The
complexity occurs in following aspects:
(1) The joint displacement components for each joint of a rigid frame increase to three, the translations
in x and y directions and the rotation of the joint, respectively.
(2) The members of a rigid frame are oriented in different directions, so it becomes necessary to
transform the stiffness relations for each member from the member’s local coordinate system to a common
global coordinate system.
(3) If hinged joints occur on a rigid frame, the relative rotations between the member ends meeting at
the same hinged joint must be taken into consideration.
Unlike in the case of the classical methods of analysis considered previously, it is usually not necessary
to neglect member axial deformations when analyzing rigid frames by the matrix displacement method. In
fact, the case of neglecting member axial deformations will be discussed as a special example individually.
13.5.1 Element orientation vectors and assembly of element stiffness matrices
(1) Nodal global displacement numbers

As we have known in last section, there is only one nodal displacement component of a node for a
continuous beam, the rotation of the node. However, there are three nodal displacement components of a
node for a rigid frame, the translations in the x and y directions, respectively, and the rotation about the z axis.
Therefore, when numbering the displacements of the nodes of a rigid frame, all of the nodal displacement
components must be uniformly numbered in accordance with the connecting manner of each node.
Consider the rigid frame shown in Fig.13.16 (a). We can see that node A has three displacement
components—the translations u A and vA in the x and y directions, respectively, and the rotation θ A about
the z axis, so the nodal global displacement numbers must be [1 2 3]. Since joint B is attached to the fixed
support, it can neither translate nor rotate; it has three known displacement components, the translations
uB = 0 , vB = 0 and the rotation θ B = 0 , so the nodal global displacement numbers must be [0 0 0].
Similarly, node C, which is connected to a hinged support, has two given displacement components, the
translations uC = 0 , vC = 0 , and a unknown rotation θC , thus the nodal global displacement numbers
must be [0 0 4].
2 0
( 2) ( 5)
3 4 ( 3) ( 6)
x
1 A ① C 0 (1) ① ( 4) x
( 2)
( 3)
② (1)
B
②
0
0 (a) (b)
0
y
( 4) ( 6)
( 5)
x
Fig.13.16 Global displacement number for planar rigid frame
(a) nodal global displacement numbers; (b) element local end displacement numbers
Note that the unknown displacement components are numbered by starting at the lowest node number
and proceeding sequentially to the highest node number. In the case of more than one unknown displacement
components at a node, the translation in the x direction is numbered first, followed by the translation in the y
direction, and then the rotation. The known nodal displacement components are numbered by zero.
Since the rigid frame has four unknown nodal displacement components, its nodal (or joint)
displacement vector {Δ} will be written as follows.
{Δ} = [ Δ1 Δ2 Δ3 Δ 4 ] = [u A vA θ A θC ]
T T
Its corresponding nodal (or joint) force vector will be written as:
{F } = [ F1 F4 ]
T
F2 F3
(2) Element orientation vectors

The rigid frame shown in Fig.13.16 (a) has two elements, ① and ② . The arrows on the members
indicate the positive directions of the abscissas x of their local coordinate systems. The six numbers within
the brackets, (1), (2), (3), (4), (5) and (6) identify the element local end displacement components, the
translation in the x direction numbered first, followed by the translation in the y direction, and then the
rotation [Fig.13.16 (b)].
As defined in subsection 13.4.2, the element orientation vector is composed by its nodal global
displacement component numbers of the element.
Table 13-2 shows the corresponding relations between the end local displacement component numbers
and its nodal global displacement component numbers for each element of the rigid frame.
Table 13-2 relationship between end local displacement numbers

and its nodal global displacement numbers
Element ① Element ②
local numbers → global numbers Orientation vector local numbers → global numbers Orientation vector
⎧1 ⎫ ⎧1 ⎫
(1) →1 ⎪2 ⎪ (1) →1 ⎪2 ⎪
(2) →2 ⎪ ⎪ (2) →2 ⎪ ⎪
(3) →3 ⎪⎪3 ⎪⎪ (3) →3 ⎪⎪3 ⎪⎪
{λ} = ⎨ ⎬
①
{λ} = ⎨ ⎬
②
(4) →0 (4) →0
⎪0 ⎪ ⎪0 ⎪
(5) →0 ⎪0 ⎪ (5) →0 ⎪0 ⎪
(6) →4 ⎪ ⎪ (6) →0 ⎪ ⎪
⎩⎪4 ⎭⎪ ⎩⎪0 ⎭⎪
(3) Assembling process of element stiffness matrices

The global stiffness matrix of element ① has been calculated in example 13-1. We write it again as
follows
(1) (2) (3) (4) (5) (6)

(1) ⎡ 300 0 0 −300 0 0 ⎤
⎢
(2) ⎢ 0 12 30 0 −12 30 ⎥⎥
(3) ⎢ 0 30 100 0 −30 50 ⎥ (a)
[k ] ⎥ ×10
①
= ⎢
4
(4) ⎢ −300 0 0 300 0 0 ⎥

(5) ⎢ 0 −12 −30 0 12 −30 ⎥
⎢ ⎥
(6) ⎣⎢ 0 30 50 0 −30 100 ⎦⎥
Based on the orientation vector {λ}

①
tabulated in table 13-2 and relationships between the end local
displacement numbers and nodal global displacement numbers, place sequentially the element in row (i) and
column (j) of [k ]
①
to the position in row λi and column λ j in matrix [ K ] to form the first-stage result
of global stiffness matrix of the frame as follow
(1) (2) (3) (6)
1 2 3 4
(1) 1 ⎡300 0 0 0 ⎤
⎢
(2) 2 ⎢ 0 12 30 30 ⎥⎥
[K ] = ⎢ ×104
(3) 3 0 30 100 50 ⎥
⎢ ⎥
(6) 4 ⎣ 0 30 50 100 ⎦
Note that the nodal global displacement component numbers corresponding to the element’s local
numbers (4) and (5) are zero, so there are no positions in global stiffness matrix [ K ] for the elements in row
①
(4) and column (4), in row (5) and column (5) of [k ] .
The global stiffness matrix of element ② has been also obtained in example 13-1. We write it again as
follow
(1) (2) (3) (4) (5) (6)
(1) ⎡ 12 0 −30 −12 0 −30 ⎤
⎢
(2) ⎢ 0 300 0 0 −300 0 ⎥⎥
(3) ⎢ −30 0 100 30 0 50 ⎥ (b)
[k ] ⎥ ×10
②
= ⎢
4
(4) ⎢ −12 0 30 12 0 30 ⎥
⎢
(5) 0 −300 0 0 300 0 ⎥
⎢ ⎥
(6) ⎢⎣ −30 0 50 30 0 100 ⎥⎦
Based on the orientation vector {λ}

②
tabulated in table 13-2 and relationships between the end local
displacement numbers and nodal global displacement numbers, place sequentially the element in row (i) and
column (j) of [k ]
②
to the position in row λi and column λ j in matrix [ K ] , which, at this time, has
already added the stiffness coefficients of element ① , to form the final result of global stiffness matrix of
the frame as follows
(1) (2) (3)
1 2 3 4
(1) 1 ⎡300 + (12) 0 + (0) 0 + (−30) 0 ⎤
⎢
(2) 2 ⎢ 0 + (0) 12 + (300) 30 + (0) 30 ⎥⎥
[K ] = ⎢ ⎥
×104
(3) 3 0 + (−30) 30 + (0) 100 + (100) 50
⎢ ⎥
4⎣ 0 30 50 100 ⎦
numbers (4), (5) and (6) are zero, so there are no positions in global stiffness matrix [ K ] for the elements in
②
row (4) and column (4), in row (5) and column (5), in row (6) and column (6) of [k ] .
13.5.2 Treatment of hinged joints
In this subsection, we will discuss how to deal with the rigid frames with hinged joint. Consider the
frame shown as in Fig.13.17, on which there is a hinged joint.
First, let us consider the joint global displacement numbers.
As shown in the figure, joints B and D are attached to the fixed supports, i.e., the two joint displacements
are known to be equal to zero, so the three global nodal displacement numbers for the two joints are [0 0 0]
each. Since node A is a rigid joint, it has three unknown displacement components. Thusly, its global nodal
displacement numbers are [1 2 3]. Because joint C is a hinged joint, it has relative rotation between the two
member ends. Therefore, it is necessary to consider the joint as two nodes, nodes C1 and C2 . The two
nodes have the same translations but different rotation. In this circumstance, we use global nodal
displacement numbers [4 5 6] to identify the displacement components of node C1 , [4 5 7] of node
C2 .
Next, let us determine orientation vectors of the elements.
From Fig.13.17, the orientation vectors of the elements ① , ② and ③ , whose positive axial
directions of abscissas x in the local coordinates are indicated by the arrows shown in the figure, can be
written as follows
{λ} = [1 2 3 4 5 6]
① T
{λ} = [1 2 3 0 0 0]
② T
{λ} = [ 4 5 7 0 0 0]
③ T
Finally, let us develop global stiffness matrix of the frame by use element assembly method in terms of
the order proceeding from the element with the smallest number to the largest number.
Assume that the size and material of each member of the frame shown in Fig.13.17 are identical, which
are the same as those of the members of the frame in example 13-1.
2 5
6
3
4 C1
1 4 x
A ① C2
7
5
② ③
B D
0 0
0
0 0
0
y
Fig.13.17 Global displacement number for the
rigid frame with hinged joint
Assembly of element ① :
①
The global element stiffness matrix [k ] has been determined as equation (a) in subsection 13.5.1.
Based on the orientation vector {λ} , place the element in [k ]
① ①
to their proper positions in matrix [ K ] ,
which, at this time, has only zero elements, to form the first-stage result of global stiffness matrix of the frame
as follows
(1) (2) (3) (4) (5) (6)
1 2 3 4 5 6 7
(1) 1 ⎡ 300 0 0 −300 0 0 0⎤
⎢
(2) 2 ⎢ 0 12 30 0 −12 30 0 ⎥⎥
(3) 3 ⎢ 0 30 100 0 −30 50 0⎥
⎢ ⎥
[ K ] = (4) 4 ⎢ −300 0 0 300 0 0 0 ⎥ ×104
(5) 5 ⎢ 0 −12 −30 0 12 −30 0⎥
⎢ ⎥
(6) 6 ⎢ 0 30 50 0 −30 100 0⎥
⎢
7⎣ 0 0 0 0 0 0 0 ⎥⎦
Assembly of element ② :
②
The global element stiffness matrix [k ] has been obtained as equation (b) in subsection 13.5.1. By
the orientation vector {λ} , accumulate the element in [k ]
② ②
to their proper positions in above matrix
[ K ] to form the second-stage result of global stiffness matrix of the frame as follow
(1) (2) (3)
1 2 3 4 5 6 7
(1) 1 ⎡300 + 12 0+0 0 − 30 −300 0 0 0⎤
(2) 2 ⎢⎢ 0 + 0 12 + 300 30 + 0 0 −12 30 0 ⎥⎥
(3) 3 ⎢ 0 − 30 30 + 0 100 + 100 0 −30 50 0⎥
⎢ ⎥
[ K ] = 4 ⎢ −300 0 0 300 0 0 0 ⎥ ×104
5⎢ 0 −12 −30 0 12 −30 0⎥
⎢ ⎥
6⎢ 0 30 50 0 −30 100 0⎥
7 ⎢⎣ 0 0 0 0 0 0 0 ⎥⎦
Assembly of element ③ :
is equal to [k ] . By the orientation vector {λ} , add the
③ ② ③
The global element stiffness matrix [k ]
③
element of [k ] to their proper positions in matrix [ K ] , which, by now, has already added the stiffness
coefficients of two elements ① and ② , to form the final result of global stiffness matrix of the frame as
follow
(1) (2) (3)
1 2 3 4 5 6 7
1 ⎡ 312 0 −30 −300 0 0 0 ⎤
⎢
2⎢ 0 312 30 0 −12 30 0 ⎥⎥
3 ⎢ −30 30 200 0 −30 50 0 ⎥
⎢ ⎥
[ K ] = (1) 4 ⎢−300 0 0 300 + 12 0 + 0 0 0 − 30 ⎥ ×104
(2) 5 ⎢ 0 −12 −30 0+0 12 + 300 −30 0 + 0 ⎥
⎢ ⎥
6⎢ 0 30 50 0 −30 100 0 ⎥
(3) 7 ⎢⎣ 0 0 0 0 − 30 0+0 0 0 + 100 ⎥⎦
13.6 Equivalent Nodal Loads
In the previous two sections, we have discussed the global stiffness matrix [ K ] and developed global
stiffness equations of a structure in matrix form as follow
{F } = [ K ]{Δ} (13-35)
The global stiffness matrix [ K ] , which represents the relationships between the global joint
displacements {Δ} and global joint forces {F } (that is, the forces yielded on their corresponding artificial
restraints), is yielded in terms of displacement-method’s primary system of the original structure. The
relationships only reflect the rigidity properties but not involve the actual loads applying on the structure, so
they are not the primary equations used to analyze the original structure by displacement method.
Recalling from chapter 11, when establishing the primary equations used to analyze a structure by
displacement method, we have considered two states of the primary system:
① The primary structure was subjected only to external loads ( let the joint displacements {Δ} = {0} ).
In this situation, we used {FP } to denote the restraint forces yielded on the artificial restraints on the
primary structure.
② The primary structure was subjected only to joint displacements {Δ} ( let the external loads equal
to zero). In this case, the restraint forces yielded on the artificial restraints on the primary structure were
identified by {F } = [ K ]{Δ} .
Superimposing above two cases, we obtained the primary equations as follows
{F } + {FP } = 0
or
[ K ]{Δ} + {FP } = {0} (13-36)
(1) Concept of equivalent joint loads and equivalent nodal load vector
The loads applying on a structure may be distributed loads (named non-joint loads afterwards), or joint
loads, or the combination of joint and non-joint loads. In order to facilitate the matrix analysis, we have to
transform non-joint loads into equivalent joint loads, indicated by {PE } . The rule for transforming non-joint
loads into equivalent joint loads is that the original non-joint loads and their equivalent joint loads yield
identical restraint forces on the artificial restraints at the joints of the primary system of the structure. That is,
if the restraint forces on the primary system caused by the original loads are {FP } , the restraint forces on the
primary system induced by their equivalent joint loads {PE } are {FP } as well. Therefore, we have
following conclusion in function form as
{PE } = − {FP } (13-37)
As a matter of fact, if there are joint loads applying on the original structure, representing them by {PJ } ,
equation (13-36) must be written as
[ K ]{Δ} + {FP } = {PJ } (13-38)
Substituting equation (13-37) into equation (13-38), we obtain
[ K ]{Δ} = {PJ } − {FP } = {PJ } + {PE } = {P}

Or simply as
[ K ]{Δ} = {P} (13-39)
In which,
{P} = {PJ } − {FP } = {PJ } + {PE } (13-40)
{P} represents the effects of all external loads on the nodes of the structure. It is termed equivalent
nodal load vector.
It is observed from equation (13-35) and equation (13-39) that if we change the nodal restraint force
vector {F } in Eq. (13-35) into the equivalent nodal load vector {P} , we will obtain the
displacement-method’s primary equations of a structure.
ⓔ
(2) Equivalent joint loads {PE } of an element
First, let us consider the case in local coordinates. If we impose restraint to all likely moving directions
of the two ends of an element, i.e. 6 restraints, the two ends of the element will be fixed. Under action of
external loads, the two ends will yield 6 fixed-end forces. The matrix form of the six fixed-end forces,
{FP }ⓔ , will be written as
{F }
ⓔ T
P = ⎡⎣ X P1 YP1 M P1 X P 2 YP 2 M P 2 ⎤⎦ (13-41)
Table 13-3 shows the fixed-end forces of seven typical loading forms for a prismatic member. If add
ⓔ
minus signs in the fixed-end forces, we will obtain the element’s equivalent joint loads {PE } in its local
coordinates as follows.
{P } = − { FP }
ⓔ ⓔ
E (13-42)
Next, let us consider the situation in global coordinates. By the transformation equations of coordinates,
Eq. (13-22), we obtain
{PE } = [T ] { PE }
ⓔ T ⓔ
(13-43)
The equation (13-43) is nothing else but the equivalent joint loads of element ⓔ in global coordinates.
ⓔ
Table 13-3 fixed-end forces {FP } due to non-joint loads in local coordinates
End
Types of load Beginning end（1） Ending end（2）
Number
forces
q 0 0
XP
x
⎛ a 2 a3 ⎞ qa 3 ⎛ a⎞
1
a b YP − qa ⎜1 − 2 + 3 ⎟ − ⎜1 − ⎟
⎝ l 2l ⎠ l 2 ⎝ 2l ⎠
l qa ⎛
2
a a2 ⎞ qa 3 ⎛ a⎞
MP − ⎜6 −8 + 3 2 ⎟ ⎜4 −3 ⎟
y 12 ⎝ l l ⎠ 12l ⎝ l⎠
q
0 0
XP
x qb 2 ⎛ 2a ⎞ qa 2 ⎛ 2b ⎞
2 a b YP − ⎜1 + ⎟ − ⎜1 + ⎟
l2 ⎝ l ⎠ l2 ⎝ l ⎠
l qab 2 qa 2b
MP −
y l l2
q
0 0
XP
x
a b 6qab 6qab
3 YP −
l3 l3
l qb ⎛ 3b ⎞ qa ⎛ 3a ⎞
y MP ⎜2 − ⎟ ⎜2 − ⎟
l ⎝ l ⎠ l ⎝ l ⎠
q
0 0
XP
x qa ⎛ 3a 2 1.6a 3 ⎞ qa 3 ⎛ a⎞
4 a b YP − ⎜2 − 2 + 3 ⎟ − ⎜ 3 − 1.6 ⎟
4 ⎝ l l ⎠ 4l 2 ⎝ l⎠
l qa 2 ⎛ a a2 ⎞ qa 3 ⎛ a⎞
MP − ⎜ 2 − 3 + 1.2 2 ⎟ ⎜ 1 − 0.8 ⎟
y 6 ⎝ l l ⎠ 4l ⎝ l⎠
q
⎛ a⎞ qa 2
XP − qa ⎜1 − 0.5 ⎟ −0.5
x ⎝ l⎠ l
a b
5 YP
0 0
l
MP
y 0 0
q
qb qa
x
XP − −
l l
a b
6 YP
0 0
l
MP
y 0 0
q
0 0
XP
x
a b qa 2 ⎛ a 3b ⎞ qa 2 ⎛ a 3b ⎞
7 YP ⎜ + ⎟ − ⎜ + ⎟
l2 ⎝ l l ⎠ l2 ⎝ l l ⎠
l qab 2 qa 2b
MP − −
y l2 l2
(3) Equivalent joint loads {PE } of a structure

As mentioned in previous, the number of components of the equivalent joint load vector {PE } of a
structure is the same as that of the joint displacement vector of the structure. The components of vector {PE }
ⓔ
can be obtained by algebraically adding each of the elements in all {PE } to their proper positions oriented
by the proper {λ}
ⓔ
in {PE } .
(4) Equivalent nodal load vector {P} of a structure
In general sense, a structure might be subjected to both joint loads and non-joint loads. As known from
previous discussion, we use vector {PJ } represents the effect of joint loads and {PE } indicates the effect
of non-joint loads. Therefore, the total effects of all external loads, known as equivalent nodal load vector
{P} , on the nodes of the structure will be expressed as follows.
{P} = {PJ } + {PE } (13-44)
Example 13-3
Determine the equivalent nodal load vector {P} for the rigid frame shown in Fig.13.18.
Solution
ⓔ
(1) Fixed-end force {FP } in local coordinates
Element ① : Referring to the row 1 in table 13-3 and substituting q = 4.8kN/m , a = l = 5m into
the equations tabulated in the row, we write the fixed-end forces of the element as follows
⎧ X P1 = 0 ⎧ X P2 = 0
⎪ ⎪
⎨YP1 = −12kN ⎨YP 2 = −12kN
⎪ M = −10kN ⋅ m ⎪ M = 10kN ⋅ m
⎩ P1 ⎩ P2
Element ② : Referring to the row 2 in table 13-3 and substituting q = −8kN , a = b = 2.5m into
the equations tabulated in the row, we find that the fixed-end forces of the element will be
⎧ X P1 = 0 ⎧ X P2 = 0
⎪ ⎪
⎨YP1 = 4kN ⎨YP 2 = 4kN
⎪ M = 5kN ⋅ m ⎪ M = −5kN ⋅ m
⎩ P1 ⎩ P2
Consequently,
⎧ 0 ⎫ ⎧0⎫
⎪−12 ⎪ ⎪4⎪
⎪ ⎪ ⎪ ⎪
⎪−10 ⎪ ⎪5⎪
⎪ ⎪ ⎪ ⎪
{F } {F }
① ②
P =⎨" ⎬ P = ⎨" ⎬
⎪ 0 ⎪ ⎪0⎪
⎪ ⎪ ⎪ ⎪
⎪−12 ⎪ ⎪4⎪
⎪ 10 ⎪ ⎪−5⎪
⎩ ⎭ ⎩ ⎭
20kN
Δ2
Δ3 4.8kN/m
Δ4
Δ1 A ① C x
2.5m
8kN ②
2.5m
B
5m
y
ⓔ
(2) Element’s equivalent joint loads {PE } in global coordinates due to non-joint loads
Element ① : As known from the figure, since α1 = 0° , [T ]① = [ I ] . Referring to equations (13-42)
①
and (13-43), we write the equivalent joint loads {PE } in global system as follows
⎧ 0 ⎫
⎪ 12 ⎪
⎪ ⎪
⎪ 10 ⎪
⎪ ⎪
{PE } = − [T ] {F } = − [ I ]{ FP } = − { FP }
① ①T ① ① ①
P =⎨" ⎬
⎪ 0 ⎪
⎪ ⎪
⎪ 12 ⎪
⎪ ⎪
⎩−10 ⎭
Element ② : It can be observed from the figure, that since α 2 = 90° , by equations (13-42) and
②
(13-43), we write the equivalent joint loads {PE } in global system as follows
⎡0 −1 0 0 0 0 ⎤⎧ 0 ⎫ ⎧ 4 ⎫
⎢1 0 0 0 0 0 ⎥⎥ ⎪ 4 ⎪ ⎪ 0 ⎪
⎢ ⎪ ⎪ ⎪ ⎪
⎢0 0 1 0 0 0 ⎥ ⎪⎪ 5 ⎪⎪ ⎪⎪−5⎪⎪
{PE } = − [T ] {F }
② ②T ②
= −⎢ ⎥ =
P
⎢0 0 0 0 −1 0 ⎥ ⎨⎪ 0 ⎬⎪ ⎨⎪ 4 ⎬⎪
⎢0 0 0 1 0 0⎥ ⎪ 4 ⎪ ⎪ 0 ⎪
⎢ ⎥⎪ ⎪ ⎪ ⎪
⎢⎣ 0 0 0 0 0 1 ⎥⎦ ⎪⎩−5⎪⎭ ⎪⎩ 5 ⎪⎭
(3) Joint loads {PJ } of the structure

It can be seen from the figure that only second component of the joint load vector {PJ } is non-zero.
Thusly, the vector will be
⎧0⎫
⎪20 ⎪
{PJ } = ⎪⎨ ⎪⎬
⎪0⎪
⎪⎩ 0 ⎪⎭
(4) Equivalent nodal load vector {P} of the structure

As shown in the figure, the orientation vectors of the two elements will be written as
⎧1⎫ ⎧1⎫
⎪2⎪ ⎪2⎪
⎪ ⎪ ⎪ ⎪
⎪3⎪ ⎪3⎪
⎪ ⎪ ⎪ ⎪
{λ} {λ}
① ②
= ⎨"⎬ = ⎨"⎬
⎪0⎪ ⎪0⎪
⎪ ⎪ ⎪ ⎪
⎪0⎪ ⎪0⎪
⎪ ⎪ ⎪ ⎪
⎩4⎭ ⎩0⎭
As known from equation (13-44), the total effects of all external loads, called equivalent nodal load
vector {P} , on the nodes of the structure will be obtained by algebraically adding each of the elements in
{PE } to that in {PJ } . For the rigid frame, the equivalent joint load vector {PE } due to the non-joint loads
will be obtained by accumulating the elements at the proper positions oriented by {λ} and {λ} in the
① ②
① ②
two {PE } and {PE } together. Therefore, {PE } will be written as
(1) 1 ⎧ 0 ⎫ (1) 1 ⎧ 4 ⎫ ⎧ 4 ⎫
(2) 2 ⎪⎪ 12 ⎪⎪ (2) 2 ⎪⎪ 0 ⎪⎪ ⎪⎪ 12 ⎪⎪
{PE } = ⎨ ⎬+ ⎨ ⎬=⎨ ⎬
(3) 3 ⎪ 10 ⎪ (3) 3 ⎪−5⎪ ⎪ 5 ⎪
(6) 4 ⎪⎩−10 ⎪⎭ 4 ⎪⎩ 0 ⎪⎭ ⎪⎩−10 ⎪⎭
Finally, by equation (13-44) the equivalent nodal load vector {P} of the structure will be obtained to
be
⎧0⎫ ⎧ 4 ⎫ ⎧ 4 ⎫
⎪20⎪ ⎪ 12 ⎪ ⎪ 32 ⎪
{P} = {PJ } + { PE } = ⎪⎨ ⎪⎬ + ⎪⎨ ⎪⎬ = ⎪⎨ ⎪⎬
⎪0⎪ ⎪ 5 ⎪ ⎪ 5 ⎪
⎪⎩ 0 ⎪⎭ ⎪⎩−10 ⎪⎭ ⎪⎩−10 ⎪⎭
13.7 Procedure for Analysis and Examples
Based on the discussion presented in the previous sections, we can develop the following step-by-step
procedure for the analysis of structures by the matrix displacement method.
1. Prepare a computing model of a structure as follows:
(1) Draw a line diagram of the structure, on which each node (or joint) and element (or member) must
be identified by a number.
(2) Select a global x y coordinate system, with the x and y axes oriented in the horizontal (positive to
the right) and vertical (positive downward) directions, respectively.
(3) For each element, establish a local x y coordinate system by selecting one of the joints at its ends
as the beginning joint and the other as the end joint. On the line diagram of the structure, for each element
indicates the positive direction of the local x axis by drawing an arrow along the member pointing toward
the end joint. For horizontal members, the coordinate transformations can be avoided by selecting the joint at
the left end of the member as the beginning joint.
(4) Identify the unknown joint (or nodal) displacements, {Δ} , of the structure. The nodal
displacements are specified on the structure's line diagram by drawing arrows at the nodes and are numbered
by starting at the lowest nodal number and proceeding sequentially to the highest nodal number. In the case of
more than one nodal displacement at a node, the x translation is numbered first, followed by the y translation,
and then the rotation. Recall that a joint of a plane frame can have up to three unknown displacements (two
translations and a rotation); a joint of a continuous beam can have up to two unknown displacements (a
translation perpendicular to the beam's centroidal axis and a rotation); and a joint of a plane truss can have up
to two unknown displacements (two translations). Note that joint translations are considered as positive when
in the positive directions of the x and y axes; joint rotations are considered as positive when clockwise.
2. Evaluate the structure stiffness matrix [ K ]
ⓔ ⓔ
(1) Compute the member stiffness matrix in local coordinates, [ k ] . Expressions of [ k ] for the
members of rigid frames, continuous beams and trusses are given in equations (13-8), (13-14) and (13-16),
respectively.
(2) Compute the member's transformation matrix [T ] by using Eq. (13-20). For horizontal members
with the local x axis positive to the right (i.e., in the same direction as the global x axis), the element
stiffness equations in the local and global coordinates are the same (i.e., [k ] = [k ] ); go to step (4).
ⓔ ⓔ
(3) Determine the element stiffness matrix in global coordinates, [k ] = [T ] [k ] [T ] (Eq. (13-27)).
ⓔ T ⓔ
ⓔ
The matrix [k ] must be symmetric.
(4) Identify the member's orientation vector {λ}
ⓔ ⓔ
and store the pertinent elements of [k ] in their
proper positions oriented by {λ}
ⓔ
in the structure stiffness matrix [ K ] . The complete structure stiffness
matrix [ K ] obtained by assembling the stiffness coefficients of all the elements of the structure must be
symmetric.
3. Form equivalent nodal load vector {P} of the structure
(1) If the member is subjected to external loads, then evaluate its fixed-end force vector in local
ⓔ
coordinates, {FP } , by using the expressions for fixed-end forces given in table 13-3.
(2) Determine member’s equivalent joint loads {PE }ⓔ in local coordinates by using Eq. (13-42).
(3) Identify member’s equivalent joint loads {PE }ⓔ in global coordinates by using Eq. (13-43).
(4) Calculate equivalent joint loads {PE } of the structure by storing the pertinent components of
{PE }ⓔ in their proper positions oriented by {λ}ⓔ in {PE } .
(5) Find joint load vector {PJ } of the structure whose component numbers are associated with those
of the joint displacements {Δ} .
(6) By Eq. (13-44), add the corresponding components of {PJ } and {PE } together to form
equivalent nodal load vector {P} of the structure.
4. Determine the unknown joint displacements {Δ} . Substitute {P} , [ K ] and {Δ} into the
structure stiffness equations, {P} = [ K ]{Δ} (Eq. (13-39)), and solve the resulting system of simultaneous
equations for the unknown joint displacements {Δ} .
5. Compute member end displacements and end forces. For each member, do the following:
(1) Obtain member end displacements in global coordinates, {Δ} , from the joint displacements, by
ⓔ
using orientation vector {λ} .

ⓔ
(2) Determine member end displacements in local coordinates by using the relationship
{Δ} = [T ]{Δ}ⓔ (Eq. (13-21)). For horizontal members with the local x axis positive to the right,
ⓔ
{Δ}ⓔ = {Δ}ⓔ .
(3) Compute member end forces in local coordinates by using the relationship
{F }ⓔ = [ k ]ⓔ {Δ}ⓔ + {FP }ⓔ (13-45)
For trusses, {FP }ⓔ = {0} .

6. Determine support reactions by considering the equilibrium of the joints located at the supports of
the structure.
Example 13-4
Determine the internal forces for the continuous beam shown in Fig.13.19 (a) by using matrix
displacement method. The flexural rigidity of each members is unchangeable, i.e., EI = constant .
Solution
(a) (b)
100kN q = 12kN/m 60kN
Δ1 Δ2 Δ3
0 1 2 3
① ② ③ x
4m 4m 8m 4m 4m y

(a) original structure; (b) computing model
(1) Unknown joint displacements

The computing model is shown in Fig.13.19 (b). From the model, we observe that joint 0 is fixed to the
support and joints 1, 2 and 3 are free to rotate. Thus the beam has 3 unknown joint displacements, Δ1 , Δ 2
and Δ 3 , which are the unknown rotations of joints 1, 2 and 3, respectively.
ⓔ
(2) Element stiffness matrix [k ]
Since the local x axis is identical to global x axis, the element stiffness matrices in local and global
coordinates are the same, i.e., [k ] = [k ] . In addition, the flexural rigidity per unit length of each element,
ⓔ ⓔ
EI
i= , is unchangeable, by using Eq. (13-14) we write
l
⎡ 4i 2i ⎤
[k ] = [k ] = [k ] = ⎢
① ② ③
⎥
⎣ 2i 4i ⎦
(3) Global stiffness matrix [ K ]
From Fig.13.19 (b), we can see that the orientation vectors of the elements are as follows.
⎧0 ⎫ ⎧1 ⎫ ⎧2 ⎫
{λ} {λ} {λ}
① ② ③
=⎨ ⎬ =⎨ ⎬ =⎨ ⎬
⎩1 ⎭ ⎩2 ⎭ ⎩3 ⎭
ⓔ
Based on the orientation vectors, the pertinent elements of [k ] can be stored in their proper positions
in the global stiffness matrix [ K ] . The final [ K ] will be
⎡8i 2i 0 ⎤
[ K ] = ⎢⎢ 2i 8i 2i ⎥⎥
⎢⎣ 0 2i 4i ⎥⎦
ⓔ
(4) Element’s equivalent joint loads {PE }
Referring to table 13-3, the fixed-end forces {FP }ⓔ for each element will be obtained as follows.
⎧ ql ⎫ ⎧ ql 2 ⎫ ⎧ ql ⎫
⎪⎪ − 8 ⎪⎪ ⎧−100 ⎫ − ⎪⎪ − 8 ⎪⎪ ⎧−60 ⎫
⎪⎪ 12 ⎪⎪ ⎧−64 ⎫
{F } {F } {F }
① ② ③
=⎨ ⎬=⎨ ⎬ =⎨ 2⎬
=⎨ ⎬ =⎨ ⎬=⎨ ⎬
⎪+ ql ⎪ ⎩+100 ⎭ ⎪+ ql ⎪ ⎩+64 ⎭ ⎪+ ql ⎪ ⎩+60 ⎭
P P P
⎩⎪ 8 ⎭⎪ ⎪⎩ 12 ⎪⎭ ⎩⎪ 8 ⎭⎪
Since {PE }ⓔ = {PE }ⓔ , by using Eq. (13-37) we write element’s equivalent joint loads as
⎧+100 ⎫ ⎧+64 ⎫ ⎧+60 ⎫

{PE } {PE } {PE }
① ② ③
=⎨ ⎬ =⎨ ⎬ =⎨ ⎬
⎩−100 ⎭ ⎩−64 ⎭ ⎩−60 ⎭
From the Fig.13.19, we can see that there is no joint load, i.e., {PJ } = {0} . The equivalent joint loads
{PE } of the structure can be obtained by storing the pertinent components of {PE }ⓔ in their proper
positions oriented by {λ} in {PE } . By Eq. (13-44), add the corresponding components of {PJ } and
ⓔ
{PE } together to form equivalent nodal load vector {P} of the structure. The final {P} becomes as
⎧−36 ⎫
{P} = ⎪⎨ −4 ⎪⎬
⎪ ⎪
⎩−60 ⎭
(6) Solve equation {P} = [ K ]{Δ} to determine {Δ}

Substituting {P} and [ K ] into the equation, we write the equation in expanded form as
⎡ 8i 2i 0 ⎤ ⎧ Δ1 ⎫ ⎧−36 ⎫
⎢ 2i 8i 2i ⎥ ⎪Δ ⎪ = ⎪ −4 ⎪
⎢ ⎥⎨ 2⎬ ⎨ ⎬
⎪ ⎪ ⎪ ⎪
⎣⎢ 0 2i 4i ⎥⎦ ⎩ Δ 3 ⎭ ⎩−60 ⎭
By solving these equations simultaneously, we determine the joint displacements to be
⎧ Δ1 ⎫ ⎧ −5.848 ⎫
⎪ ⎪ ⎪ ⎪1
⎨ Δ 2 ⎬ = ⎨ 5.385 ⎬
⎪ Δ ⎪ ⎪−17.693⎪ i
⎩ 3⎭ ⎩ ⎭
(7) Member end displacements {Δ}

ⓔ ⓔ
and forces {F }
The member end displacements in global coordinates, {Δ} , can be obtained by simply placing the
ⓔ
corresponding components, whose positions are oriented by {λ} , of {Δ} to the proper positions in
ⓔ
{Δ}ⓔ . Then, the end displacement vectors are as follows.
⎧ 0 ⎫1 ⎧−5.848⎫ 1 ⎧ 5.385 ⎫ 1
{Δ} {Δ} {Δ}
① ② ③
=⎨ ⎬ =⎨ ⎬ =⎨ ⎬
⎩−5.848⎭ i ⎩ 5.385 ⎭ i ⎩−17.693⎭ i
Because of {F } = {F } , by applying equation (13-45), the member end forces for the elements of
ⓔ ⓔ
the continuous beam can be determined as follows:
⎡ 4i 2i ⎤ ⎧ 0 ⎫ 1 ⎧−100⎫ ⎧−111.71⎫
{F }
①
=⎢ ⎥⎨ ⎬ +⎨ ⎬=⎨ ⎬
⎣ 2i 4i ⎦ ⎩−5.848⎭ i ⎩+100 ⎭ ⎩ 76.62 ⎭
⎡ 4i 2i ⎤ ⎧−5.848⎫ 1 ⎧−64 ⎫ ⎧−76.62 ⎫
{F}
②
=⎢ ⎥⎨ ⎬ +⎨ ⎬=⎨ ⎬
⎣ 2i 4i ⎦ ⎩ 5.385 ⎭ i ⎩+64 ⎭ ⎩ 73.84 ⎭
⎡ 4i 2i ⎤ ⎧ 5.385 ⎫ 1 ⎧−60⎫ ⎧73.84 ⎫
{F }
③
=⎢ ⎥⎨ ⎬ +⎨ ⎬=⎨ ⎬
⎣ 2i 4i ⎦ ⎩−17.693⎭ i ⎩+60⎭ ⎩ 0 ⎭
Based on the member end forces, the bending moment and shear force diagrams are drawn as shown in
Fig.13.20.
111.70 76.62 73.84 54.38 48.35 39.23
20.77 83.08
105.84 45.62 47.65 20.77
(a) M diagram (kN ⋅ m) (b) Q diagram (kN)
Example 13-5
Determine the internal forces for the rigid frame shown in Fig.13.21 (a) by using matrix displacement
method. Assume the cross-section of each member is rectangular; the elastic modulus E = 1 . Their areas
and inertial moments are also shown in the figure.
Solution
Δ2 Δ5
1 Δ6
I 2 = m4 Δ3
A A2 = 0.63m 2 12 B
Δ1 Δ4
A B
q = 1kN/m
②
A1 = 0.5m 2 A1 = 0.5m 2
6m
1 1 ① ③
I1 = m 4 I1 = m 4
24 24 x
C y D
C 12m D
{0} {0}
(b)
(a)
(a) original structure; (b) computing model
(1) Unknown joint displacements

From the computing model shown in Fig.13.21 (b), we observe that while joints C and D of the structure
can neither translate nor rotate, joints A and B are free to translate as well as rotate. Thus each of the two joints
has three unknown displacements, translations in x and y directions, respectively, and rotation about z. The
total unknown joint displacements of the rigid frame are six; the displacement numbers of joint A are
indicated by [1 2 3], joint B by [4 5 6], and joints C and D by {0}, as shown in Fig.13.21 (b).
ⓔ
(2) Element stiffness matrix in local coordinates [ k ]
The pertinent quantities are calculated as follows:
EA1 EI1 2 EI1
Column: = 83.3 × 10−3 = 6.94 × 10 −3 = 13.9 × 10−3
l1 l1 l1
4 EI1 6 EI1 12 EI1
= 27.8 × 10−3 = 6.94 × 10 −3 = 2.31 × 10 −3
l1 l12 l13
EA2 EI 2 2 EI 2
Beam: = 52.5 × 10 −3 = 6.94 × 10−3 = 13.9 × 10 −3
l2 l2 l2
4 EI 2 6 EI 2 12 EI 2
= 27.8 × 10 −3 = 3.47 × 10−3 = 0.58 × 10−3
l2 l22 l23
Elements ① and ③ :
⎡ 83.3 0 0 −83.3 0 0 ⎤
⎢ 0 2.31 6.94 0 −2.31 6.94 ⎥⎥
⎢
① ③ ⎢ 0 6.94 27.8 0 −6.94 13.9 ⎥
⎡k ⎤ = ⎡k ⎤ = ⎢ ⎥ × 10
−3
⎣ ⎦ ⎣ ⎦
⎢ −83.3 0 0 83.3 0 0 ⎥
⎢ 0 −2.31 −6.94 0 2.31 −6.94 ⎥
⎢ ⎥
⎢⎣ 0 6.94 13.9 0 −6.94 27.8 ⎥⎦
Elements ② :
⎡ 52.5 0 0 −52.5 0 0 ⎤
⎢ 0 0.58 3.47 0 −0.58 3.47 ⎥⎥
⎢
② ⎢ 0 3.47 27.8 0 −3.47 13.9 ⎥
⎡k ⎤ = ⎢ ⎥ × 10
−3
⎣ ⎦
⎢ −52.5 0 0 52.5 0 0 ⎥
⎢ 0 −0.58 −3.47 0 0.58 −3.47 ⎥
⎢ ⎥
⎢⎣ 0 3.47 13.9 0 −3.47 27.8 ⎥⎦
ⓔ
(3) Element stiffness matrix in global coordinates [k ]
Since the angle between x axis of element ① (or ③ ) and x axis is α = 90° , the direction
cosines of the two elements are as follows
cos α = 0 sin α = 1
Substitution of these values into equation (13-20) yields the transformation matrix [T ] for the
element:
⎡0 1 0 ⎤
⎢ −1 0 0
⎢ [ 0] ⎥⎥
⎢0 0 1 ⎥
[T ] = ⎢ 0 1 0⎥
⎥
⎢
⎢ [ 0] −1 0 0 ⎥
⎢ ⎥
⎢⎣ 0 0 1 ⎥⎦
ⓔ
To determine the two element stiffness matrices in global coordinates, [k ] , we substitute the matrix
[k ]ⓔ and the [T ] into the relationship [k ]ⓔ = [T ]T [k ]ⓔ [T ] and carry out the necessary matrix
multiplications to obtain
⎡ 2.31 0 −6.94 −2.31 0 −6.94 ⎤

⎢ 0 83.3 0 0 −83.3 0 ⎥
① ⎢ −6.94 0 27.8 6.94 0 13.9 ⎥
[k ] = [ k ] = [T ] ⎡ k ⎤ [T ] = ⎢
① ③
⎥ × 10−3
T
⎣ ⎦ ⎢ −2.31 0 6.94 2.31 0 6.94 ⎥
⎢ 0 −83.3 0 0 83.3 0 ⎥
⎢⎣ −6.94 0 13.9 6.94 0 27.8 ⎥⎦
Because the local x y coordinate for element ② coincides with the global x y coordinate, no
coordinate transformations are needed, or [T ] = [ I ] ; i.e., the element stiffness relations in the local and
global coordinates are the same. That is,
[k ]
② ②
= ⎡⎣ k ⎤⎦
(4) Assemble global stiffness matrix [ K ] by element assembly method

From the computing model shown in Fig.13.21 (b), the orientation vector, reflecting the relations
between the local and global numbers of each element, can be written as follows:
⎧1 ⎫ ⎧1 ⎫ ⎧4 ⎫
⎪2 ⎪ ⎪2 ⎪ ⎪5 ⎪
⎪⎪3 ⎪⎪ ⎪⎪3 ⎪⎪ ⎪⎪6 ⎪⎪
{λ} {λ} {λ}
① ② ③
=⎨ ⎬ =⎨ ⎬ =⎨ ⎬
0 4 0
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎪0 ⎪ ⎪ ⎪
5 ⎪0 ⎪
⎩⎪0 ⎭⎪ ⎩⎪6 ⎭⎪ ⎩⎪0 ⎭⎪
By using the orientation vectors, {λ} , the relevant elements of [k ]

ⓔ ⓔ
can be added into their positions
in the global stiffness matrix [ K ] . Thusly assembled [ K ] will be as follows
⎡ 54.81 0 −6.94 −52.5 0 0 ⎤

⎢ 0 83.88 3.47 0 −0.58 3.47 ⎥
⎢ −6.94 3.47 55.6 0 −3.47 13.9 ⎥
[ K ] = ⎢ −52.5 0 ⎥ ×10−3
−6.94 ⎥
⎢ 0 54.81 0
⎢ 0 −0.58 −3.47 0 83.88 −3.47 ⎥
⎢⎣ 0 3.47 13.9 −6.94 −3.47 55.6 ⎥⎦
(5) Element’s equivalent joint loads {PE }ⓔ

Referring to table 13-3, the fixed-end forces {FP }① for element ① , only the element subjected to
load, will be obtained as follows.
{P } = − { FP } = [ 0 −3 −3 0 −3 3]
① ① T
E
Since the angle between x axis of element ① and x axis is α = 90° , by using Eqs. (13-23) and
①
(13-37) we write the element’s equivalent joint loads {PE } as
⎡0 1 0 ⎤⎧ 0 ⎫ ⎧ 3 ⎫
⎢ −1 0 0 [0] ⎥ ⎪−3⎪ ⎪ 0 ⎪
⎢0 0 1 ⎥ ⎪⎪−3⎪⎪ ⎪⎪−3⎪⎪
{PE } = [T ] {P }
① ①
= −⎢ ⎥ =
T
E
⎢ 0 1 0⎥ ⎨ 0 ⎬ ⎨ 3 ⎬
⎪ ⎪ ⎪ ⎪
⎢ [ 0] −1 0 0 ⎥ ⎪−3⎪ ⎪ 0 ⎪
⎢⎣ 0 0 1 ⎥⎦ ⎩⎪ 3 ⎭⎪ ⎩⎪ 3 ⎭⎪

From the Fig.13.21, we can see that there is no joint load, i.e., {PJ } = {0} . The equivalent joint loads
{PE } of the structure can be obtained by adding the relevant elements of {PE }① into their proper
positions determined by {λ} in {PE } . By Eq. (13-44), accumulate the corresponding components of
①
{PJ } and {PE } together to form equivalent nodal load vector {P} of the structure. The final {P}
will be
{P} = [3 0 −3 0 0 0 ]
T

⎡ 54.81 0 −6.94 −52.5 0 0 ⎤ ⎧ Δ1 ⎫ ⎧ 3 ⎫

⎢ 0 83.88 3.47 0 −0.58 3.47 ⎥ ⎪Δ 2 ⎪ ⎪ 0 ⎪
⎢ −6.94 3.47 55.6 0 −3.47 13.9 ⎥ ⎪⎪ Δ 3 ⎪⎪ ⎪⎪−3⎪⎪
10−3 × ⎢ ⎥ =
⎢ −52.5 0 0 54.81 0 −6.94 ⎥ ⎨Δ 4 ⎬ ⎨ 0 ⎬
⎪ ⎪ ⎪ ⎪
⎢ 0 −0.58 −3.47 0 83.88 −3.47 ⎥ ⎪ Δ 5 ⎪ ⎪ 0 ⎪
⎢⎣ 0 3.47 13.9 −6.94 −3.47 55.6 ⎥⎦ ⎪⎩Δ 6 ⎪⎭ ⎪⎩ 0 ⎭⎪
⎧ Δ1 ⎫ ⎧u A ⎫ ⎧ 847 ⎫
⎪ Δ ⎪ ⎪ v ⎪ ⎪−5.13⎪
⎪ 2⎪ ⎪ A⎪ ⎪ ⎪
⎪ Δ 3 ⎪ ⎪θ A ⎪ ⎪ 28.4 ⎪
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨ " ⎬ = ⎨" ⎬ = ⎨ " ⎬
⎪ Δ ⎪ ⎪u ⎪ ⎪ 824 ⎪
⎪ 4⎪ ⎪ B⎪ ⎪ ⎪
⎪ Δ 5 ⎪ ⎪ vB ⎪ ⎪ 5.13 ⎪
⎪ Δ ⎪ ⎪θ ⎪ ⎪ 96.5 ⎪
⎩ 6⎭ ⎩ B⎭ ⎩ ⎭
Member end displacements {Δ}
ⓔ ⓔ
(8) and forces {F } in local coordinates
The member end displacements in global coordinate, {Δ} , can be obtained by simply placing the
ⓔ
ⓔ

⎧ 847 ⎫ ⎧ 847 ⎫ ⎧ 824 ⎫
⎪−5.13⎪ ⎪−5.13⎪ ⎪5.13⎪
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎪ 28.4 ⎪ ⎪ 28.4 ⎪ ⎪96.5⎪
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
{Δ} {Δ} {Δ}
① ② ③
=⎨ " ⎬ =⎨ " ⎬ =⎨ " ⎬
⎪ 0 ⎪ ⎪ 824 ⎪ ⎪ 0 ⎪
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎪ 0 ⎪ ⎪ 5.13 ⎪ ⎪ 0 ⎪
⎪ 0 ⎪ ⎪ ⎪ ⎪ ⎪
⎩ ⎭ ⎩ 96.5 ⎭ ⎩ 0 ⎭
Because of {F } = [ k ] {Δ} + {FP } and {F } = [T ]{F } , the member end forces for the
ⓔ ⓔ ⓔ ⓔ ⓔ ⓔ
elements of the portal frame can be determined as follows:

Element ① :
Member end force in global coordinate system:
{F } = [k ] {Δ} + { FP }
① ① ① ①
⎡ 2.31 0 −6.94 −2.31 0 −6.94 ⎤ ⎧ 847 ⎫ ⎧ −3⎫ ⎧ −1.24 ⎫

⎢ 0 83.3 0 0 −83.3 0 ⎥ ⎪ −5.13⎪ ⎪ 0 ⎪ ⎪ −0.43⎪
⎢ −6.94 0 27.8 6.94 0 13.9 ⎥ ⎪⎪ 28.4 ⎪⎪ ⎪⎪ 3 ⎪⎪ ⎪⎪ −2.09 ⎪⎪
= 10−3 × ⎢ ⎥ + =
⎢ − 2.31 0 6.94 2.31 0 6.94 ⎥ ⎨ 0 ⎬ ⎨ −3⎬ ⎨ −4.76 ⎬
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎢ 0 −83.3 0 0 83.3 0 ⎥ ⎪ 0 ⎪ ⎪ 0 ⎪ ⎪ 0.43 ⎪
⎢ −6.94 0 13.9 6.94 0 27.8 ⎥⎦ ⎪⎩ 0 ⎪⎭ ⎩⎪ −3⎪⎭ ⎪⎩ −8.49 ⎪⎭
⎣
Member end force in local coordinate system:
{F} = [T ]{ F } = [ −0.43 1.24 −2.09 0.43 4.76 −8.49]

① ① T
Element ② :
{F } = [k ] {Δ}
+ { FP }
② ② ② ②
⎡ 52.5 0 0 −52.5 0 0 ⎤ ⎧ 847 ⎫ ⎧0 ⎫ ⎧ 1.24 ⎫

⎢ 0 0.58 3.47 0 −0.58 3.47 ⎥ ⎪−5.13⎪ ⎪0 ⎪ ⎪ 0.43 ⎪
⎢ 0 3.47 27.8 0 −3.47 13.9 ⎥ ⎪⎪ 28.4 ⎪⎪ ⎪⎪0 ⎪⎪ ⎪⎪ 2.09 ⎪⎪
= 10 × ⎢
−3
⎥ + =
⎢ −52.5 0 0 52.5 0 0 ⎥ ⎨ 847 ⎬ ⎨0 ⎬ ⎨−1.24 ⎬
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎢ 0 −0.58 −3.47 0 0.58 −3.47 ⎥ ⎪ 5.13 ⎪ ⎪0 ⎪ ⎪ −0.43⎪
⎢ 0 3.47 13.9 0 − 3.47 27.8 ⎥⎦ ⎩⎪ 96.5 ⎭⎪ ⎩⎪0 ⎭⎪ ⎪⎩ 3.04 ⎪⎭
⎣
Since α = 0° , {F } = {F } . The member end force in local coordinate system will be

② ②
{F } = [1.24 0.43 2.09 −1.24 −0.43 3.04]

② T
Element ③ :
{F} = [k ] {Δ}+ { FP }
③ ③ ③ ③
⎡ 2.31 0 −6.94 −2.31 0 −6.94 ⎤ ⎧ 824 ⎫ ⎧0 ⎫ ⎧ 1.24 ⎫

⎢ 0 83.3 0 0 −83.3 0 ⎥ ⎪5.13⎪ ⎪0 ⎪ ⎪ 0.43 ⎪
⎢ −6.94 0 27.8 6.94 0 13.9 ⎥ ⎪⎪96.5⎪⎪ ⎪⎪0 ⎪⎪ ⎪⎪−3.04 ⎪⎪
= 10−3 × ⎢ ⎥⎨ ⎬+⎨ ⎬= ⎨ ⎬
⎢ −2.31 0 6.94 2.31 0 6.94 ⎥ ⎪ 0 ⎪ ⎪0 ⎪ ⎪−1.24 ⎪
⎢ 0 −83.3 0 0 83.3 0 ⎥ ⎪ 0 ⎪ ⎪0 ⎪ ⎪ −0.43⎪
⎢ ⎥ ⎩⎪ 0 ⎭⎪ ⎩⎪0 ⎭⎪ ⎩⎪ −4.38⎭⎪
⎣ −6.94 0 13.9 6.94 0 27.8 ⎦
{F } = [T ]{ F } = [ 0.43 −1.24 −3.04 −0.43 1.24 −4.38]

③ ③ T

Based on the member end forces, the bending moment, and shear and axial force diagrams are drawn,
respectively, as shown in Fig.13.22.
(a) (b)
0.43
2.09 3.04 1.24
A B
C A
M diagram Q diagram
B (kN ⋅ m) D C
(kN)
1.24 D
8.49 4.38 4.76
(c)
1.24
A B
N diagram
0.43 (kN) 0.43
C D
13.8 Global Analysis of Rigid Frames without Considering Axial Deformation
In practical engineering, the rectangular shaped rigid frames composed of columns and beams have a
variety of application. The axial deformations of the members of this sort of rigid frames are quite small in
comparison with other deformations. In order to simplify the analysis, the member’s axial deformations of
the rigid frames are often neglected. The section will discuss the matrix analysis for the plane rigid frames
without considering the axial deformations of their members.
Consider the portal rigid frame shown in Fig.13.23. Following aspects must be concerned when
analyze the frame by neglecting its member’s axial deformations.
(1) Nodal global displacement numbers
From the frame shown in Fig.13.23, we can see that joints B and D are both attached to the fixed
supports, they can neither translate nor rotate; each of them has three known displacement components, the
translations in x and y directions, and the rotation about z axis. They are all equal to zero, so the nodal
global displacement numbers for each of the two joints must be [0 0 0]. Since the member’s axial
deformations are neglected, the vertical displacements at nodes A , C1 and C2 must be equal to zero,
then the corresponding global displacement numbers must be zero as well; in addition, the horizontal
displacements at nodes A , C1 and C2 must be identical, thusly the corresponding global displacement
numbers must be the same. Consequently, the global displacement numbers for node A are represented by
[1 0 2]; the numbers for node C1 by [1 0 3] and for C2 by [1 0 4], as shown in the figure.
(2) Element orientation vectors
The rigid frame shown in Fig.13.23 has three elements, numbered by ① , ② and ③ . The arrows
on the members indicate the positive directions of the abscissas x of their local coordinate systems. For
each element, six numbers are used to identify the element’s global end displacement component numbers,
which consist of so called orientation vector. The orientation vectors for the three elements are thusly
0 0
Δ2 Δ3
Δ1 C1
Δ1 x
A ① C2 Δ1
Δ4
0
② ③
B D
0 0
0 0
0 0
y
Fig.13.23 The computing model of a rectangular shaped rigid frame
written as follows.
{λ} = [1 0 2 1 0 3]
① T
{λ} = [1 0 2 0 0 0]
② T
{λ} = [1 0 4 0 0 0]
③ T
(3) Assembly of element stiffness matrices and development of global stiffness matrix [ K ]
Assume the member’s size and materials of the frame are the same as those of the frame in example
13-1. Thusly, the global stiffness matrix of element ① has been calculated in example 13-1. We write it
again as follows
(1) (2) (3) (4) (5) (6)

(1) ⎡ 300 0 0 −300 0 0 ⎤
(2) ⎢⎢ 0 12 30 0 −12 30 ⎥⎥
(3) ⎢ 0 30 100 0 −30 50 ⎥ × 104
[k ]
①
= ⎢ ⎥
(4) ⎢ −300 0 0 300 0 0 ⎥
(5) ⎢ 0 −12 −30 0 12 −30 ⎥
⎢ ⎥
(6) ⎣ 0 30 50 0 −30 100 ⎦
Based on the orientation vector {λ} = [1 0

①
2 1 0 3]T and relationships between the end
local displacement numbers and nodal global displacement numbers, place sequentially the elements of
[k ]① to the relevant positions in matrix [ K ] to form the first-stage result of global stiffness matrix of the
frame as follows
(1)(4) (3) (6)
1 2 3 4
(1)(4)1 ⎡300 − 300 − 300 + 300 0 + 0 0 + 0 0⎤
(3) 2 ⎢⎢ 0+0 100 50 0 ⎥⎥
[K ] = ×104
(6) 3 ⎢ 0+0 50 100 0 ⎥
⎢ ⎥
4⎣ 0 0 0 0⎦
numbers (2) and (5) are zero, so there are no positions in global stiffness matrix [ K ] for the elements in
①
row (2) and column (2), in row (5) and column (5) of [k ] .
The global stiffness matrix of element ② has been also obtained in example 13-1. We write it again
as follows
(1) (2) (3) (4) (5) (6)
(1) ⎡ 12 0 −30 −12 0 −30 ⎤
⎢
(2) ⎢ 0 300 0 0 −300 0 ⎥⎥
(3) ⎢ −30 0 100 30 0 50 ⎥ × 104
[k ]
②
= ⎢ ⎥
(4) ⎢ −12 0 30 12 0 30 ⎥
(5) ⎢ 0 −300 0 0 300 0 ⎥
⎢ ⎥
(6) ⎣ −30 0 50 30 0 100 ⎦
Based on the orientation vector {λ} = [1 0

②
2 0 0 0]T and relationships between the end
local displacement numbers and nodal global displacement numbers, add sequentially the elements of
[k ]② to the pertinent positions in matrix [ K ] , which, at this time, has already added the stiffness
coefficients of element ① , to form the second-stage result of global stiffness matrix of the frame as
follows
(1) (3)
1 2 3 4
(1) 1 ⎡ 0 + 12 0 − 30 0 0⎤
(3) 2 0 − 30 100 + 100 50 0 ⎥⎥
⎢
[ K ] = ⎢⎢ ⎥
×104
3 0 50 100 0
⎢ ⎥
4⎣ 0 0 0 0⎦
numbers (2), (4), (5) and (6) are zero, so there are no positions in global stiffness matrix [ K ] for the
elements in row (2) and column (2), in row (4) and column (4), in row (5) and column (5), in row (6) and
②
column (6) of [k ] .
Since the global stiffness matrix of element ③ is equal to that of element ② , i.e., [k ]
③
= [k ]② ,
by orientation vector {λ} = [1 0
③
4 0 0 0]T and relationships between the end local
③
displacement numbers and nodal global displacement numbers, add sequentially the elements of [k ] to
the proper positions in matrix [ K ] , which, at this time, has already added the stiffness coefficients of
elements ① and ② , to form the final result of global stiffness matrix of the frame as follows
(1) (3)
1 2 3 4
(1) 1 ⎡12 + 12 −30 0 0 − 30 ⎤
⎢
2 ⎢ −30 200 50 0 ⎥⎥
[ ]
K =
⎢ ⎥
×104
3 0 50 100 0
⎢ ⎥
(3) 4 ⎣ 0 − 30 0 0 0 + 100 ⎦
numbers (2), (4), (5) and (6) are zero, so there are no positions in global stiffness matrix [ K ] for the
elements in row (2) and column (2), in row (4) and column (4), in row (5) and column (5), in row (6) and
③
column (6) of [k ] .
Example 13-6
Determine the internal forces for the rigid frame shown in Fig.13.21 by using matrix displacement
method. Neglect the member’s axial deformations and assume the member’s size and materials are the
same as those used in example 13-15.
Solution
x
(1) Unknown joint displacements y
From the computing model shown in Δ2 Δ3
Fig.13.24, if we neglect the member’s axial Δ1 A B Δ1
②
deformations the total unknown joint
① ③
displacements of the rigid frame will become
three, the translation of node A or B in x C D
direction, and the rotations of nodes A and B {0} {0}
about z axis, respectively; thus the Fig.13.24 The computing model of example 13-6
displacement numbers of joint A are
indicated by [1 0 2], joint B by [1 0 3],
and joints C and D by {0}, as shown in Fig.13.24.
ⓔ
(2) Element stiffness matrix in local coordinates [k ]
[k ]ⓔ is the same as that of example 13-5.
ⓔ
(3) Element stiffness matrix in global coordinates [k ]
[k ]ⓔ is the same as that of example 13-5.
(4) Assemble global stiffness matrix [ K ] by element assembly method
From the computing model shown in Fig.13.24, the orientation vector of each element can be written
as follows:
{λ} = [1 0 2 0 0 0]
① T
{λ} = [1 0 2 1 0 3]
② T
{λ} = [1 0 3 0 0 0]
③ T
By using the orientation vectors, {λ} , the relevant elements of [k ]

ⓔ ⓔ
can be sequentially added into
their proper positions in the global stiffness matrix [ K ] . Thusly assembled [ K ] will be as follows
⎡ 4.62 −6.94 −6.94 ⎤

[ K ] = ⎢−6.94 55.6 13.9 ⎥ ×10−3
⎢ −6.94 13.9 55.6 ⎥⎦
⎣
(5) Element’s equivalent joint loads {PE }ⓔ
{PE }ⓔ is the same as that of example 13-5.

The equivalent joint loads {PE } of the structure can be obtained by adding the relevant elements of
{PE }① (since only element ① subjected to load) into their proper positions determined by
{λ}① = [1 0 2 0 0 0]T in {PE } . Because of {PJ } = {0} , by Eq. (13-44), the final {P} will
be
{P} = [3 −3 0]
T

⎡ 4.62 −6.94 −6.94 ⎤ ⎧ Δ1 ⎫ ⎧ 3 ⎫

⎪ ⎪ ⎪ ⎪
10 × −6.94 55.6 13.9 ⎥ ⎨Δ 2 ⎬ = ⎨−3⎬
−3
⎢
⎢ −6.94 13.9 55.6 ⎥⎦ ⎩⎪ Δ 3 ⎭⎪ ⎩⎪ 0 ⎭⎪
⎣
⎧ Δ1 ⎫ ⎧u A ⎫ ⎧ 0.838 ⎫
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨Δ 2 ⎬ = ⎨θ A ⎬ = ⎨ 0.0261⎬ ×10
3
⎪⎩ Δ 3 ⎪⎭ ⎪⎩θ B ⎪⎭ ⎪⎩0.0979 ⎪⎭
Member end displacements {Δ}

ⓔ ⓔ
(8) and forces {F } in local coordinates
The member end displacements in global coordinates, {Δ} , can be obtained by simply placing the
ⓔ
ⓔ

⎧ 0.838 ⎫ ⎧ 0.838 ⎫ ⎧ 0.838 ⎫
⎪ 0 ⎪ ⎪ 0 ⎪ ⎪ 0 ⎪
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎪0.0261⎪ ⎪ 0.0261⎪ ⎪0.0979 ⎪
{Δ} {Δ} {Δ}
① ② ③
=⎨ ⎬ =⎨ ⎬ =⎨ ⎬
⎪ 0 ⎪ ⎪ 0.838 ⎪ ⎪ 0 ⎪
⎪ 0 ⎪ ⎪ 0 ⎪ ⎪ 0 ⎪
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎩ 0 ⎭ ⎩0.0979 ⎭ ⎩ 0 ⎭
Because of {F } = [ k ] {Δ} + {FP } and {F } = [T ]{F } , the member end forces for the
ⓔ ⓔ ⓔ ⓔ ⓔ ⓔ
elements of the rigid frame can be determined as follows:

Element ① :
{F } = [ k ] {Δ} + { FP }
① ① ① ①
⎡ 2.31 0 −6.94 −2.31 0 −6.94 ⎤ ⎧ 0.838 ⎫ ⎧ −3⎫ ⎧ −1.25 ⎫

⎢ 0 83.3 0 0 −83.3 0 ⎥⎪ 0 ⎪ ⎪ 0 ⎪ ⎪ 0 ⎪
⎢ −6.94 0 27.8 6.94 0 13.9 ⎥ ⎪⎪0.0261⎪⎪ ⎪⎪ 3 ⎪⎪ ⎪⎪ −2.09 ⎪⎪
=⎢ ⎥ + =
⎢ −2.31 0 6.94 2.31 0 6.94 ⎥ ⎨ 0 ⎬ ⎨ −3⎬ ⎨ −4.75 ⎬
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎢ 0 −83.3 0 0 83.3 0 ⎥⎪ 0 ⎪ ⎪ 0 ⎪ ⎪ 0 ⎪
⎢⎣ −6.94 0 13.9 6.94 0 27.8 ⎥⎦ ⎪⎩ 0 ⎪⎭ ⎪⎩ −3⎪⎭ ⎪⎩ −8.41⎪⎭
{F} = [T ]{ F } = [ 0 1.25 −2.09 0 4.75 −8.41]

① ① T
Element ② :
{F } = [ k ] {Δ} + { FP }
② ② ② ②
⎡ 52.5 0 0 −52.5 0 0 ⎤ ⎧ 0.838 ⎫ ⎧0 ⎫ ⎧ 0 ⎫

⎢ 0 0.58 3.47 0 −0.58 3.47 ⎥ ⎪ 0 ⎪ ⎪0 ⎪ ⎪ 0.43 ⎪
⎢ 0 3.47 27.8 0 −3.47 13.9 ⎥ ⎪⎪ 0.0261⎪⎪ ⎪⎪0 ⎪⎪ ⎪⎪ 2.09 ⎪⎪
=⎢ ⎥ + =
⎢ −52.5 0 0 52.5 0 0 ⎥ ⎨ 0.838 ⎬ ⎨0 ⎬ ⎨ 0 ⎬
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎢ 0 −0.58 −3.47 0 0.58 − 3.47 ⎥ ⎪ 0 ⎪ ⎪0 ⎪ ⎪−0.43⎪
⎢⎣ 0 3.47 13.9 0 −3.47 27.8 ⎥⎦ ⎩⎪0.0979 ⎭⎪ ⎩⎪0 ⎭⎪ ⎩⎪ 3.09 ⎭⎪
Since α = 0° , {F } = {F } . The member end force in local coordinate system will be

② ②
{F } = [ 0 0.43 2.09 0 −0.43 3.09]

② T
Element ③ :
{F } = [ k ] {Δ} + { FP }
③ ③ ③ ③
⎡ 2.31 0 −6.94 −2.31 0 −6.94 ⎤ ⎧ 0.838 ⎫ ⎧0 ⎫ ⎧ 1.25 ⎫

⎢ 0 83.3 0 0 −83.3 0 ⎥ ⎪ 0 ⎪ ⎪0 ⎪ ⎪ 5 ⎪
⎢ −6.94 0 27.8 6.94 0 13.9 ⎥ ⎪⎪0.0979 ⎪⎪ ⎪⎪0 ⎪⎪ ⎪⎪ −3.09 ⎪⎪
=⎢ ⎥ + =
⎢ −2.31 0 6.94 2.31 0 6.94 ⎥ ⎨ 0 ⎬ ⎨0 ⎬ ⎨ −1.25 ⎬
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎢ 0 −83.3 0 0 83.3 0 ⎥ ⎪ 0 ⎪ ⎪0 ⎪ ⎪ 0 ⎪
⎣⎢ −6.94 0 13.9 6.94 0 27.8 ⎥⎦ ⎪⎩ 0 ⎪⎭ ⎪⎩0 ⎪⎭ ⎪⎩−4.47 ⎪⎭

{F } = [T ]{ F } = [ 0 −1.25 −3.09 0 1.25 −4.47 ]

③ ③ T

Based on the member end forces, the bending moment, shear and axial force diagrams are drawn,
respectively, as shown in Fig.13.25. in comparison with the diagrams shown in Fig.13.22, it is observed
that the effect on the internal forces due to the member’s axial deformations are small enough to be
neglected. In addition, since the member’s axial deformations are assumed to be zero, the axial force
determined by the stiffness equations are zero as well. Then, the member’s axial forces must be identified
by the equilibrium equations derived from shear force diagram. Consequently, the internal force diagrams
for the rigid frame are shown in Fig.13.25.
0.43
2.09 3.09 1.25
A B
C A
(a) (b)
M diagram Q diagram
B (kN ⋅ m) D C (kN)
1.25 D
8.41 4.47 4.75
1.25
A (c) B
N diagram 0.43
(kN)
0.43
C D
13.9 Block Diagram and Computer Programs for Plane Framed Structures
Based on forgoing discussion about the basic principle and procedure of matrix displacement method,
a block diagram summarizing the various steps involved in the method and the computer programs, edited
in the Fortran 90 language, for the analysis of plane framed structures, such as rigid frames, continuous
beams, trusses, bent frames and composite structures are presented as follows.
13.9.1 Design of the block diagram
Name of the program
1
Develop a module to define the types of variables to digitalize a structure [“module variable_type_module”
and “Subroutine input_data”]
2
For each element:
(1) evaluate direction cosines [Subroutine SCL]; (2) determine element’s orientation vector {λ )ⓔ [Subroutine
EJC]; (3) Set up element’s coordinate transformation matrix [Subroutine CTM]; (4) yield element stiffness
matrix in local coordinates [k ]ⓔ [Subroutine ESM]; (5) calculate element stiffness matrix in global
coordinates, [k ]ⓔ = [T ]T [k ]ⓔ [T ] , and store [ k ]ⓔ into global stiffness matrix [K] by {λ )ⓔ
3
Form global nodal load vector {P} = {PJ } + {PE } [Subroutine nlv]
4
Solve structural stiffness equation {P}=[K]{Δ} [Subroutine BGS] to determine {Δ}
5
For each member:
(1) obtain {Δ} ⓔ
from {Δ} ; (2) calculate {Δ}ⓔ = [T ]{Δ}ⓔ ; (3) evaluate {F }ⓔ = [k ]ⓔ{Δ}ⓔ + {FP }ⓔ
[“Subroutine MQN” and “Subroutine EFX(w)”
13.9.2 Computer program for plane framed structure
The program consists of two portions, the main program and sub program composed by modules. The
main program shows how to inform computer to read and store data, and how to call relevant subroutine
programs. The further explanation for each module of the sub program will be taken place in latter
subsections.
13.9.2.1 Main program for plane framed frames
!======================================================================!
PROGRAM plane_frame_analysis_programme
use Frame_analysis_sub
implicit none
character*60,dimension(11)::FORM
print *,'input file name '
read *, input_file_name
OPEN (5,file=input_file_name)
print *,'output file name '
read *, output_file_name
OPEN (6,FILE=output_file_name)
CALL input_data
CALL nlv
CALL ssm
CALL BGS
CALL MQN
CLOSE (5)
CLOSE (6)
STOP
END PROGRAM plane_frame_analysis_programme

!======================================================================!
13.9.2.2 Subroutine analytical programs for plane framed structures
!===================================================================!
module variable_type_module
implicit none
integer(kind(1)),parameter::Ikind=kind(1),Rkind=kind(0.d0)
integer(Ikind)::NE,NJ,N,NW,NPJ,NPF,J1,J2,I,J,I1,M,K,L,IND,JJ
real(Rkind)::BL,SI,CO,S,C,G,G1,G2,G3,DX,DY,A,B,Q
integer(Ikind),allocatable::JE(:,:),JN(:,:)
integer(Ikind),dimension(:)::JC(6)
real(Rkind),allocatable::EA(:),EI(:),X(:),Y(:),P(:)
real(Rkind),allocatable::PJ(:,:),PF(:,:),KB(:,:)
real(Rkind),dimension(:)::KE(6,6),KD(6,6),T(6,6),F(6),F0(6),D(6)
character*20::input_file_name,output_file_name
end module variable_type_module

!------------------------------------------------------------------------------------------------------------------!
module Frame_analysis_sub
use variable_type_module
contains
!-------------------------- input original data ---------------------------------------------------!
Subroutine input_data
read (5,*) NE,NJ,N,NW,NPJ,NPF

allocate (JE(2,NE),JN(3,NJ),P(N),PF(4,NPF))
allocate (EA(NE),EI(NE),X(NJ),Y(NJ),PJ(2,NPJ),KB(N,NW))
read (5,*) (X(J),Y(J),(JN(I,J),I=1,3),J=1,NJ)
read (5,*) ((JE(I,J),I=1,2),EA(J),EI(J),J=1,NE)
IF (NPJ.NE.0) READ (5,*) ((PJ(I,J),I=1,2),J=1,NPJ)

IF (NPF.NE.0) READ (5,*) ((PF(I,J),I=1,4),J=1,NPF)
WRITE (6,'(/6X,"NE=",I5,2X,"NJ=",I5,2X,"N=",I5,2X,"NW=",I5,2X,&
"NPJ=",I5,2X,"NPF=",I5)') NE,NJ,N,NW,NPJ,NPF
WRITE (6,'(/7X,"NODE",7X,"X",11X,"Y",12X,"XX",8X,"YY",8X,"ZZ"/)')
WRITE (6,'(1X,I10,2F12.4,3I10)') (J,X(J),Y(J),(JN(I,J),I=1,3),J=1,NJ)
WRITE (6,'(/4X,"ELEMENT",4X,"NODE-I",4X,"NODE-J",11X,"EA",13X,"EI"/)')
WRITE (6,'(1X,3I10,2D15.6)') (J,(JE(I,J),I=1,2),EA(J),EI(J),J=1,NE)
IF (NPJ.NE.0) WRITE (6,'(/7X,"CODE",7X,"PX-PY-PM"/(1X,F10.0,F15.4))')&

((PJ(I,J),I=1,2),J=1,NPJ)
IF (NPF.NE.0) WRITE (6,'(/4X,"ELEMENT",7X,"IND",10X,"A",14X,"Q"/&

(1X,2F10.0,2F15.4))') ((PF(I,J),I=1,4),J=1,NPF)
end subroutine input_data

!---------------------------------------------------------------------------------------------------!
!------------- Calculation of direction cosines of elements -----------------------------!
Subroutine SCL
J1=JE(1,M)
J2=JE(2,M)
DX=X(J2)-X(J1)
DY=Y(J2)-Y(J1)
BL=SQRT(DX*DX+DY*DY)
SI=DY/BL
CO=DX/BL
end Subroutine SCL

!----------------------------------------------------------------------------------------------------!
!------------- Set up element’s orientation vectors ----------------------------------------!
Subroutine EJC
J1=JE(1,M)
J2=JE(2,M)
DO I=1,3
JC(I)=JN(I,J1)
JC(I+3)=JN(I,J2)
ENDDO
END Subroutine EJC

!----------------------------------------------------------------------------------------------------!
!------------- Set up element’s coordinate transformation matrix [T]-------------------!
Subroutine CTM
T=0.D0
T(1,1)=CO
T(1,2)=SI
T(2,1)=-SI
T(2,2)=CO
T(3,3)=1.D0
DO I=1,3
DO J=1,3
T(I+3,J+3)=T(I,J)
ENDDO
ENDDO
END Subroutine CTM
!----------------------------------------------------------------------------------------------------!
!----------------------- Set up element stiffness matrix --------------------------------------!
Subroutine ESM
G=EA(M)/BL
G1=2.D0*EI(M)/BL
G2=3.D0*G1/BL
G3=2.D0*G2/BL
KD=0.D0
KD(1,1)=G
KD(1,4)=-G
KD(4,4)=G
KD(2,2)=G3
KD(2,3)=-G2
KD(5,5)=G3
KD(2,5)=-G3
KD(2,6)=-G2
KD(3,3)=2.D0*G1
KD(3,5)=G2
KD(3,6)=G1
KD(5,6)=G2
KD(6,6)=2.D0*G1
DO I=1,5
I1=I+1
DO J=I1,6
KD(J,I)=KD(I,J)
END DO
END DO
END subroutine ESM

!----------------------------------------------------------------------------------------------------!
!---------------------- Assemble structural stiffness matrix ---------------------------------!
Subroutine ssm
KB=0.D0
DO M=1,NE
call SCL
call CTM
call ESM
call EJC
KE=matmul(transpose(T),matmul(KD,T))
DO L=1,6
I=JC(L)
IF (I.NE.0) THEN
DO K=1,6
J=JC(K)
IF (J.NE.0.AND.J.GE.I) THEN
JJ=J-I+1
KB(I,JJ)=KB(I,JJ)+KE(L,K)
END IF
END DO
END IF
END DO
END DO
END subroutine ssm

!----------------------------------------------------------------------------------------------------!
!-------------------------- Set up global nodal load vector ------------------------------------!
Subroutine nlv
P=0.D0
!---------------------calculate joint load vector
IF (NPJ.NE.0) THEN
DO I=1,NPJ
L=INT(PJ(1,I))
P(L)=PJ(2,I)
ENDDO
ENDIF
!------------calculate non-joint node load vector

IF (NPF.NE.0) THEN
DO JJ=1,NPF
M=INT(PF(1,JJ))
CALL SCL
CALL EFX(JJ)
CALL CTM
CALL EJC
DO L=1,6
S=0.D0
DO K=1,6
S=S-T(K,L)*F0(K)
END DO
F(L)=S
END DO
DO J=1,6
L=JC(J)
IF (L.NE.0) P(L)=P(L)+F(J)
END DO
END DO
END IF
END Subroutine nlv

!----------------------------------------------------------------------------------------------------!
!-------------------------------Solve structural stiffness equation {P}=[K]{Δ}-----------!
Subroutine BGS
N1=N-1
!------------------------------------------------loop of elimination
DO K=1,N1
IM=K+NW-1
IF(N.LT.IM) IM=N ! the maximum number of eliminating row, IM
I1=K+1
!---------------------------------------- loop of eliminating row
DO I=I1,IM
L=I-K+1
C=KB(K,L)/KB(K,1)
JM=NW-L+1 ! find the maximum column number of eliminating row JM
!----------------------------- column’s loop of eliminating row
DO J=1,JM
JJ=J+I-K
KB(I,J)=KB(I,J)-C*KB(K,JJ)
END DO
!----------------------------- column’s loop of eliminating row
P(I)=P(I)-C*P(K)
END DO
!---------------------------------------- loop of eliminating row
END DO
!------------------------------------------------loop of elimination
!----------------------------------------------------------------------------------------------------!
P(N)=P(N)/KB(N,1) ! find nth displacement
!------------------------------------------------back substitution
!---------------------------------------------loop of back-substitution row
DO K=1,N1
I=N-K
JM=K+1
IF (NW.LT.JM) JM=NW ! find the maximum column number of back-substitution row JM
!----------------------------------------- column’s loop of back-substitution row
DO J=2,JM
L=J+I-1
P(I)=P(I)-KB(I,J)*P(L)
END DO
!----------------------------------------- column’s loop of back-substitution row
P(I)=P(I)/KB(I,1) ! find ith displacement
END DO
!---------------------------------------------loop of back-substitution row
!------------------------------------------------back substitution
WRITE (6,'(/7X,"NODE",10X,"U",14X,"V",11X,"CETA")')
DO I=1,NJ
DO J=1,3
D(J)=0.D0
L=JN(J,I)
IF (L.NE.0) THEN
D(J)=P(L)
END IF
END DO
WRITE (6,'(1X,I10,3D15.6)') I,D(1),D(2),D(3) ! output nodal displacements
END DO
END subroutine BGS

!----------------------------------------------------------------------------------------------------!
!------------------ Calculate element’s end forces due to its displacements-------------!
Subroutine MQN
WRITE (6,'(/4X,"ELEMENT",13X,"N",17X,"Q",17X,"M")')
DO M=1,NE
call SCL
call CTM
call ESM
call EJC
DO I=1,6
L=JC(I)
D(I)=0.D0
IF (L.NE.0) THEN
D(I)=P(L)
END IF
END DO
DO I=1,6
F(I)=0.D0
DO J=1,6
DO K=1,6
F(I)=F(I)+KD(I,J)*T(J,K)*D(K)
END DO
END DO
END DO
IF (NPF.NE.0) THEN
DO I=1,NPF
L=PF(1,I)
IF (M.EQ.L) THEN
CALL EFX(I)
DO J=1,6
F(J)=F(J)+F0(J)
END DO
END IF
END DO
WRITE (6,'(/1X,I10,3X,"N1=",F12.4,3X,"Q1=",F12.4,3X,"M1=",F12.4,&
/14X,"N2=",F12.4,3X,"Q2=",F12.4,3X,"M2=",F12.4)')&
M,(F(I),I=1,6)
END IF
END DO
END subroutine MQN

!----------------------------------------------------------------------------------------------------!
!-------------------------------Calculate element’s fixed-end forces-------------------------!
Subroutine EFX(w)
implicit none
integer(Ikind)::w
IND=INT(PF(2,w))
A=PF(3,w)
Q=PF(4,w)
C=A/BL
G=C*C
B=BL-A
F0=0.D0
!-----------------------------------due to transverse distributed loads
select case(IND)
case(1)
S=Q*A*0.5D0
F0(2)=-S*(2.D0-2.D0*G+C*G)
F0(5)=-S*G*(2.D0-C)
S=S*A/6.D0
F0(3)=S*(6.D0-8.D0*C+3.D0*G)
F0(6)=-S*C*(4.D0-3.D0*C)
!--------------------------------------- - due to transverse concentrated force

case(2)
S=B/BL
F0(2)=-Q*S*S*(1.D0+2.D0*C)
F0(5)=-Q*G*(1.D0+2.D0*S)
F0(3)=Q*S*S*A
F0(6)=-Q*B*G
!--------------------------------------- - due to transverse concentrated couple

case(3)
S=B/BL
F0(2)=-6.D0*Q*C*S/BL
F0(5)=-F0(2)
F0(3)=Q*S*(2.D0-3.D0*S)
F0(6)=Q*C*(2.D0-3.D0*C)
!--------------------------------------- - due to transverse triangularly distributed loads
case(4)
S=Q*A*0.25D0
F0(2)=-S*(2.D0-3.D0*G+1.60*C*G)
F0(5)=-S*G*(3.D0-1.6D0*C)
S=S*A
F0(3)=S*(2.D0-3.D0*C+1.2D0*G)/1.5D0
F0(6)=-S*C*(1.D0-0.8D0*C)
!--------------------------------------- - due to axially distributed loads
case(5)
F0(1)=-Q*A*(1.D0-0.5D0*C)
F0(4)=-0.5D0*Q*C*A
!--------------------------------------- - due to axial concentrated force
case(6)
F0(1)=-Q*B/BL
F0(4)=-Q*C
!--------------------------------------- due to distributed couples
case(7)
S=B/BL
F0(2)=-Q*G*(3.D0*S+C)
F0(5)=-F0(2)
S=S*B/BL
F0(3)=-Q*S*A
F0(6)=Q*G*B
!---------------------------------------- due to member end axial displacements
case(8)
L=INT(A)
S=Q*EA(M)/BL
F0(L)=S
IF (L.EQ.1) F0(4)=-S
IF (L.EQ.4) F0(1)=-S
!---------------------------------------- due to member end transverse displacements
case(9)
L=INT(A)
F0(L)=12.D0*EI(M)*Q/(BL*BL*BL)
IF (L.EQ.2) F0(5)=-F0(2)
IF (L.EQ.5) F0(2)=-F0(5)
F0(3)=0.5D0*BL*F0(5)
F0(6)=F0(3)
!---------------------------------------- due to member end rotations
case(10)
L=INT(A)
S=2.D0*EI(M)*Q/BL
F0(L)=2.D0*S
IF (L.EQ.3) F0(6)=S
IF (L.EQ.6) F0(3)=S
F0(5)=3.D0*S/BL
F0(2)=-F0(5)
end select
END subroutine EFX
end module Frame_analysis_sub
!--------------------------------------------------------------------------------------------------------- ----!
!===================================================================!
13.9.2.3 Explanation of variable_type_module
In the module, some variable identifiers are presented, which identifies variables and arrays of double
precision. They will be explained as follows:
(1) Integer variable identifiers
NE—number of elements; NJ—number of joints; N—number of unknown nodal displacements
NW—number of maximum band; NPJ—number of joint loads; NPF—number of non-joint loads
IND—number of types of non-joint loads; M—serial number of elements
J1, J2, I, J, I1, K, L, JJ are arbitrary integer variables.
(2) Real variable identifiers
BL—length of element; SI—direction sine of element; CO—direction cosine of element
SI, S, C, G, G1, G2, G3, DX, DY, A, B, Q are arbitrary real variables.
(3) Integer array identifiers
JE(:,:)—nodal number of element ends (allocatable); JN(:,:)—nodal number of element ends
(allocatable)
JC(6) —array used to store the components of element orientation vector.
(4) real array identifiers
EA(:)—axial rigidity of element; EI(:)—flexural rigidity of element
X(:),Y(:)—coordinates of a joint (allocatable)
P(:)—global nodal load vector (finally, it will be nodal displacement vector) (allocatable)
PJ(:,:)—array used to store the loads applying at joints (allocatable)
PF(:,:)—array used to store the loads applying on elements (allocatable)
KB(:,:)—array used to store elements in global stiffness matrix (allocatable)
KD(6,6) —array used to store elements in [k ]ⓔ
KE(6,6) —array used to store elements in [k ]ⓔ
T(6,6) —array used to store elements in coordinate transformation matrix [T ]
F0(6) —array used to store components of fixed-end forces in local coordinates
F(6) —array used to store components of element equivalent joint loads in global coordinates (finally,
store the member end forces in local coordinates)

D(6) —array used to store components of member end displacements in global coordinates
13.9.2.4 Explanation of module Frame_analysis_sub
The module consists of 10 subroutines. Each of them will be interpreted as follows.

1. Subroutine input_data
The subroutine is used to let a computer to read the data of a desired structure. The following block
diagram will explain the design of the subroutine.
Develop a data file
Let computer read the data file
From the data file read following data:

Control parameter: NE, NJ, N, NW, NPJ, NPF
Nodal coordinates: X(NJ),Y(NJ)
Nodal number of element end: JE(2,NE)
Axial rigidity of element: EA(NE)
Flexural rigidity of element: EI(NE)
loads applying at joints: J(2,NPJ)
loads applying on elements : PF(2,NPF)
Output the data for checking
2. Subroutine SCL
The subroutine is used to calculate the direction cosine of elements of a desired structure. The
following block diagram will give a further explanation of the subroutine.
J1 ⇐ JE（1，M） number of beginning node of element M

J2 ⇐ JE（2，M） number of ending node of element M
DX ⇐ X（J2）− X（J1） Δx
DY ⇐ Y（J2）− Y（J1） Δy
BL ⇐ SQRT（DX 2 + DY 2） l
SI ⇐ DY/BL sinα
CO ⇐ DX/BL conα
3. Subroutine EJC
The subroutine is employed to determine orientation vectors of the elements. The block diagram of
the subroutine will be drawn as follows.
J1 ⇐ JE（1，M）
J2 ⇐ JE（2，M）
I=1，
3
JC（I）⇐ JN（I，J1）
JC（I+3）⇐ JN（I，J2）
4. Subroutine CTM
The subroutine is utilized to form the transformation matrixes of the elements. The design of the
subroutine is based on the following block diagram.
T(I ，J) ⇐ 0.0
T（1，1）⇐ C0，T（1，2）⇐ SI
T（2，1）⇐ −SI ，T（2，2）⇐ C0
T（3，3）⇐ 1.0
I=1, 3
J=1, 3
T(I+3 ，J+3) ⇐ T（I ，J）
5. Subroutine ESM
The subroutine is used to yield the element stiffness matrix in global coordinates. Following block
diagram is the base of the design of the subroutine.
G ⇐ EA(M)/BL Basic
G1 ⇐ 2.0 × EI(M)/BL Stiffness
G2 ⇐ 3.0 × G1/BL Coefficient
G3 ⇐ 2.0 × G2/BL
⎡⎣ k ⎤⎦ = [ 0]
ⓔ
KD（I，J）⇐ 0.0
1 ）⇐ G ， KD （2 ，）
KD(1 ， 3 ⇐ G2
4 ⇐ − G ， KD （2 ，）
KD(1 ，） 6 ⇐ G2
Form
4 ⇐ G ， KD （3 ，）
KD(4 ，） 5 ⇐ − G2 upper
2 ⇐ G3 ， KD （3 ，）
KD(2 ，） 6 ⇐ G1 triangular
5 ⇐ G3 ， KD （5 ，）
KD(5 ，） 6 ⇐ − G2 submatrix
5 ⇐ − G3 ， KD （3 ，）
KD(2 ，） 3 ⇐ 2.0 × G1
6 ⇐ 2.0 × G1
KD(6 ，）
I=1，5
I1 = I+1 Form
below
triangular
submatrix
J=I1，6
KD(J,I) ⇐ KD(I,J)
6. Subroutine ssm
The subroutine is employed to form the global stiffness matrix. Its design is based on following block
diagram.
KB(I，J) ⇐ 0.0 [ K ] = [ 0]
M=1，NE
CALL SCL
CALL CTM
CALL ESM
CALL EJC
Calculate
KE ⇐ [T]T [k ]ⓔ [T]
[ k ]ⓔ
L=1，6
I ⇐ JC（L）
yes (dischage zero elements)

I=0？
no
K=1，6
store elements of [k ]ⓔ in [K]
J ⇐ JC（K）
yes
J=0 or J < I？
no
JJ ⇐ J-I+1
KB（I，JJ）⇐ KB（I，JJ）+KE（L，K
7. Subroutine nlv
The subroutine is utilized to form the equivalent nodal load vector {P}. Its design is based on
following block diagram.
P(I) ⇐ 0.0
yes (no joint load)

NPJ=0？
store joint loads in {P}

no
I=1，NPJ
L ⇐ PJ（1，I）
P（L）⇐ PJ（2，I）
yes (no element loads)

NPF=0？
no
I=1，NPF
M ⇐ PF（1，I）
CALL SCL
CALL EFX
store element loads into {P}ⓔ
CALL CTM
CALL EJC
L=1，6
S ⇐ 0.0
S ⇐ S − T（K，L）× F0（K）
F(L) ⇐ S
contd
J=1，
6
assemble {P}ⓔ into {P}

L ⇐ JC（C）
yes
L=0？
no
P（L）⇐ P（L）+F（L）
8. Subroutine BGS
The subroutine is used to solve the structure stiffness equations {P} = [ K ]{Δ} , by Gauss
elimination method, so as to determine the unknown displacements {Δ} . Its program design is referenced
to following block diagram.
N1 ⇐ N-1
K=1，N
IM ⇐ K+NW-1
yes
elimination
N﹤IM？
no IM ⇐ N
I1 ⇐ K+1
contd
I=I1，IM
elimination
L ⇐ I − K+1
C ⇐ KB(K ， L） /KB （ K ， I）
JM ⇐ NW-L+1
J=1，JM
JJ ⇐ J+I-K
KB（I，J）⇐ KB（I，J）− C × KB（K，JJ）
P（I）⇐ P（I）− C × P（K）
P（N）⇐ P（N）/KB（N，
1）
K=1，N1
I ⇐ N-K
back substitution
JM ⇐ K+1
yes
NW﹤JM？
IM ⇐ N
no
J=2，JM
L ⇐ J+I-1
P （ I）⇐ P （ I）-KB （ I， J）× P （ L ）
P（I）⇐ P（I）/KB（I，
1）
contd
write title
I=1，NJ
output nodal displacements {Δ}

J=1，
3
D（J）⇐ 0.0
L ⇐ JN（J，I）
yes
L=0？
no
D（J）⇐ P（L）
output nodal
displacement
9. Subroutine MQN
The subroutine is employed to member end forces. Its program design is based on following block
diagram.
write title
get proper data of elements
M=1，N
CALL SCL
CALL ESM
CALL CT M
CALL EJC
contd
I=1，6
obtain member end displacements

L ⇐ JC （ I）
D （ I）⇐ 0.0
yes (discarge zero component in orientation vector)

L=0？
no
D（I）⇐ P（L）
I=1，6
F（I）⇐ 0.0
find member end forces

J=1，6 due to its displacements
K=1，6
F(I)⇐ F(I)+ KD(I，J)× T(J，K)× D(K)
yes (no element load)

NPF=0
det ermine total member end forces
no
I=1，NPF
L ⇐ PF(1，I)
yes
M ≠ L？
no
CALL EFX
contd
J=1，6
F(J) ⇐ F(J )+F0(J )
Output member end
stop
10. Subroutine EFX(w)

The subroutine is utilized to calculate member fixed-end forces. The following block diagram is its
design procedure.
IND ⇐ PF（2，I）
A ⇐ PF（3，I)
Q ⇐ PF（4，I)
C ⇐ A/BL
G ⇐ C×C
B ⇐ BL-A
I=1，6
F0(I) ⇐ 0.0
SELECT CASE（1， 2， 3， 4， 5， 6， 7），IND

IND=1 2 3 4 5 6 7
concentrated moment
axially concentrated load

transverse concentrated
transverse distributed
axially distributed load

transverse triangularly
distributed moment
distributed load
load
load
13.9.3 Example and its computing results
As previous discussion mentioned, when analyze a structure on computer, the structure must be
digitalized. So a data file, commonly known as input data file, about the structure have to be make
according to the relevant read statements before running the program. In addition, the computing results of
the structure are also needed to output in a data file. Following example will illustrate the development of
input and output data files.
Example 13-7
Determine the internal forces for the rigid frame shown in Fig.13.26 by using the computer program
listed in the section. Assume the axial and flexural rigidities are EA = 4.0 ×106 kN and
EI = 1.6 ×104 kN ⋅ m2 , respectively.
Solution
25kN/m 15kN ⋅ m
3 (1, 2, 4 )
4 ( 5,6,7 )
2 (1, 2,3) ②
2m
18kN ①
③
2m
1( 0,0,0 ) 5 ( 0,0,8 )
4m x
(1) Input data file

The form of the file is text on which all of the data are written in free format, and the name of the file
is defined by users. The data and their corresponding read statements of the file are written as follows.
3,5,8,7,1,2 read (5,*) NE,NJ,N,NW,NPJ,NPF
0.D0,0.D0,0,0,0 read (5,*) (X(J),Y(J),(JN(I,J),I=1,3),J=1,NJ)

0.D0,4.D0,1,2,3
0.D0,4.D0,1,2,4
4.D0,4.D0,5,6,7
4.D0,0.D0,0,0,8
1,2,4.0D+06,1.6D+04 read (5,*) ((JE(I,J),I=1,2),EA(J),EI(J),J=1,NE)

3,4, 4.0D+06,1.6D+04
5,4, 4.0D+06,1.6D+04
7.0,-15.0 READ (5,*) ((PJ(I,J),I=1,2),J=1,NPJ)
1.0,2.0,2.0, 18.0 READ (5,*) ((PF(I,J),I=1,4),J=1,NPF)

2.0,1.0,4.0, 25.0
(2) output data file

The form of the file is text on which all of the data are written in free format by computer, and the
name of the file is defined by users. The computing results of the example are written as follows.
NE= 3 NJ= 5 N= 8 NW= 7 NPJ= 1 NPF= 2
NODE X Y XX YY ZZ
1 0.0000 0.0000 0 0 0
2 0.0000 4.0000 1 2 3
3 0.0000 4.0000 1 2 4
4 4.0000 4.0000 5 6 7
5 4.0000 0.0000 0 0 8
ELEMENT NODE-I NODE-J EA EI

1 1 2 0.400000D+07 0.160000D+05
2 3 4 0.400000D+07 0.160000D+05
3 5 4 0.400000D+07 0.160000D+05
CODE PX PY PM
7. -15.0000
ELEMENT IND A Q
1. 2. 2.0000 -18.0000
Summary 583
2. 1. 4.0000 -25.0000
NODE U V CETA
1 0.000000D+00 0.000000D+00 0.000000D+00
2 -0.221743D-02 -0.464619D-04 -0.139404D-02
3 -0.221743D-02 -0.464619D-04 0.357876D-02
4 -0.222472D-02 -0.535381D-04 -0.298554D-02
5 0.000000D+00 0.000000D+00 0.658499D-03
ELEMENT N Q M
1 N1= 46.4619 Q1= 10.7119 M1= -6.8477
N2= -46.4619 Q2= 7.2881 M2= 0.0000
2 N1= 7.2881 Q1= 46.4619 M1= 0.0000

N2= -7.2881 Q2= 53.5381 M2= 14.1523
3 N1= 53.5381 Q1= 7.2881 M1= 0.0000

N2= -53.5381 Q2= -7.2881 M2= -29.1523
SUMMARY
In this chapter we have studied the basic concepts of the matrix displacement method for the analysis
of plane framed structures. The matrix displacement method is another kind of structural analytical method
by which a structure can be analyzed by a computer in terms of the principle of traditional displacement
method and the knowledge of matrix algebra.
The matrix displacement analysis of a structure consists of three pivotal processes. They are:
(1) Discretization of the structure
The Discretization means a computing model of the structure must be prepared. The model is
represented by a line diagram of the structure, on which all the joints, the joints’ unknown displacements
and members are identified by numbers.
(2) Element analysis
Element analysis involves with the development of relations between the forces and displacements of
the member's ends, known as element stiffness equations. The pivot of the equations is the element stiffness
matrix. In this chapter, we mainly derived stiffness matrices for a general flexural member, while the
stiffness matrices for other members such as the elements of continuous beams, plane trusses and those
without considering axial deformations can be recognized as special elements of the general flexural
member. The advantage of this operation can facilitate the standardization and routinization of the analysis.
Therefore, in the computer program given in the chapter, only general flexural members are programmed.
However, the program can be used to analyze any plane framed structures.
It must be kept in mind that the comprehension of meanings of stiffness coefficients in stiffness matrix
will expedite the element analysis.
(3) Global analysis
The objective of global analysis of a structure is to express the joint forces as functions of the joint
displacements of the structure by the structure (or global) stiffness matrix. The relations can be expressed
as following matrix form
[ K ]{Δ} = {P}
In the process of establishing the stiffness relations of a structure, the pivotal work is the developments
of stiffness matrix [ K ] and equivalent nodal load vector {P} .
When develop global stiffness matrix [ K ] of the structure, the method so called element assembly
method (or stiffness assembly method or direct stiffness method) is used. By the method, the elements in
are algebraically added to their proper positions, oriented by corresponding {λ} , in the
ⓔ ⓔ
each [k ]
global stiffness matrix [ K ] . The process has two steps, orienting the element’s position and adding the
value of the element to the value that has all ready stored in its proper position of [ K ] , but the two steps
are simultaneously implemented on computers.
While determine the equivalent nodal load vector {P} of the structure, the external loads applying
on the members are equivalently transformed into so called equivalent joint loads {PE }ⓔ firstly; then
algebraically add the components of each {PE }ⓔ by corresponding orientation vector {λ}
ⓔ
to the
components of {PJ } , the external loads applying on the joints of the structure.
In actually, there are two sorts of method to obtain the structure stiffness matrix for a structure. When
the support conditions are considered before the determination of the global stiffness matrix, the method is
termed preprocessing method, which has all ready presented in previous sections of the chapter; otherwise,
the support conditions are concerned after the development of the global stiffness matrix, the method is
known as postprocessing method, which can be referenced to in other pertinent books.
13-1 How to determine the positive sense of element’s local coordinates? What is the sign convention
about the end forces and displacements of a member in matrix displacement method? What is the
difference between the chapter’s sign convention and that of the traditional displacement method?
ⓔ
13-2 What is the physical meanings of each coefficient kij in the element stiffness matrix [k ] ?
Please explain the meanings of the stiffness coefficients k34 and k 42 in the element stiffness
ⓔ
matrix [ k ] (which is a stiffness matrix of a general flexural member in its local coordinate
system) ?
13-3 Why the stiffness matrix of a general flexural member [see Eq. (13-8)] is singular? Whereas the
stiffness matrix of the members of a continuous beam [see Eq. (13-14)] is nonsingular?
13-4 How to determine the angle α used in the transformation matrix [T ] [see Eq. (13-20)]?
13-5 How much property is there for an element stiffness matrix?
13-6 What is the orientation vector of an element? How determine the orientation vector? In the global
analysis, where is the orientation vector used?
13-7 What is the element assembly method? When assemble the global stiffness matrix [ K ] , how are
the equilibrium and compatibility conditions of the structure satisfied?
13-8 What is the physical meanings of each coefficient kij in the global stiffness matrix [ K ] ?
13-9 Why do the elements located on the main diagonal of element and structure stiffness matrices
always possess positive signs? Whether or not the signs for the elements located on
non-main-diagonal positions in element and structure stiffness matrices are always positive? Why?
13-10 How to deal with the hinged joints of a rigid frame when analyze the frame by matrix displacement
method? Why are the rotations of the hinged joints of the frame taken as unknown displacements in
matrix displacement method? Why do the rotations not used as unknown displacements in
traditional displacement method?
13-11 Why must the external loads applying on the elements of a structure be equivalently transformed
into joint loads in matrix displacement method?
13-12 How to evaluate element equivalent joint loads of a structure? Whether or not the equivalent joint
loads are the same as the restraint forces of the primary system of the structure used in traditional
displacement method?
13-13 Please try to derive the stiffness equations of a general flexural member without considering its
axial deformation [see Eq. (13-12)].
13-14 Whether or not a statically determinate structure can be analyzed by the matrix displacement
method?
13-1 Determine the joint rotations and member end moments of the continuous beams shown in the
following figures by matrix displacement method.
13-2 Try to analyze the continuous beam shown in the figure by using matrix displacement method, and
draw its bending moment diagram.
(a) M 1 = 50kN ⋅ m (b) 10kN/m

i1 i2 = i1 i1
① ② ① ② i2 = i1
6m 6m 6m 6m
problem 13-1
5kN/m 5kN/m
A B C D
l l l
EI = constant l = 4m
problem 13-2
13-3 Determine the global stiffness matrix [ K ] by element assembly method and write out the primary
equations for the structure shown in the figure.
13-4 Try to find the global stiffness matrix [ K ] by element assembly method for the continuous beam
shown in the figure.
4 5 6
EI
2EI
l
1 2 3
2EI x
EI
y
l
EI 2EI
1 ① 2 ② 3 ③ 4
EI 2EI EI
l l l l l

13-5 Try to determine, neglecting the axial deformation, the global stiffness matrix [ K ] by element
assembly method for the rigid frame shown in the figure. Assume that each member has the same
size and elastic modulus. That is, the element’s length l = 5m , cross sectional area A = 0.5m ,
2
1 4
cross sectional inertial moment I = m and elastic modulus E = 3 ×10 MPa .
4
24
13-6 In problem for solution 13-5, if assume that a downward uniformly distributed load q = 4.8kN/m
is applied on element ① , please determine the internal forces and draw the internal force
diagrams of the rigid frame.
13-7 Try to write out the primary equations of the displacement method with considering the axial
deformations of the members for the rigid frame shown in the figure. Assume that the elastic
modulus and size of the members, i.e., E , A and I , are all constant.
13-8 Try to analyze the rigid frame shown in the figure by matrix displacement method without
considering the axial deformations of the members.
x
1 y 3P 2P 2P
2 ① ② 3
2 ① 1 ② 3
x
③ ③
2l
3P
y
4
4
problem 13-5 l l l l l l
2 2
problem 13-7
10kN 5 2 EI 6
4m
EI EI
20kN 3 2 EI 4
6m
1.5EI 1.5EI
1 2
8m
problem 13-8
CHAPTER 14
GENERAL REMARKS ON
STATICALLY INDETERMINATE STRUCTRES

In this chapter, we will give a general discussion on statically indeterminate structures, starting
with the classification and comparison of the methods used to analyze the structures, then
presenting a combined method by simultaneously applying moment distribution and displacement
methods for the analysis of indeterminate structures with sidesway and the approximate analysis of
statically indeterminate structures, also summing up the mechanical characteristics of the
structures, finally ending with the further discussion about the computing models of the structures.
14.1 Classification and Comparison of Fundamental Method used to Analyze Statically

Indeterminate Structures
In previous chapters, we have discussed the fundamental methods for analyzing statically indeterminate
structures. In order to have an overall understanding about the methods, we now summarize them in the
following table.
Table 14-1 Classification of methods for analyzing statically indeterminate structures
Type Generalized force method Generalized displacement method
Fundamental type Force method Displacement method
Matrix type (matrix force method) Matrix displacement method
(Successively approximate force Moment distribution method,

Successively approximate type
method) no-shear distribution method
Energy type (complementary energy method) (Potential energy method)
The contents in the table can be explained as follows:

(1) The methods for analyzing statically indeterminate structures can be classified into two
categories in terms of the characteristics of the primary unknowns the methods adopted. They are: ①
generalized force method, ② generalized displacement method.
588
14.1 Classification and Comparison of Fundamental Method used to
Analyze Statically Indeterminate Structures 589
So called generalized force method means that if the method adopts force(s) as primary unknown(s), it
will be termed force method or generalized force method, such as classical force method, complementary
energy method and the like; whereas generalized displacement method indicates that the primary
unknown(s) adopted by the method are displacement(s), such as classical displacement method, potential
energy method, matrix displacement method, moment distribution method and no-shear distribution
method presented in the book.
It should be pointed out that the methods enclosed in the brackets shown in table 14-1 were not
discussed in the book; they are the subjects in some special books.
(2) The methods used to analyze statically indeterminate structures can be also categorized into two
groups by means of the formulating way utilized by the methods. They are: ①
equilibrium-geometry-physics method, ② energy method.
Recalling from the process formulating the classical force and displacement methods, three types of
equations were employed, they are equilibrium equation(s), compatibility (or geometric) equation(s) and
force-displacement (or physical) relation(s). So equilibrium-geometry-physics method means that the
equilibrium equation(s), geometric condition(s) and physical properties of the material must be used to
analyze an indeterminate structure. In static analysis, the method is called static method.
If we use energy method to formulating methods for analyzing statically indeterminate structures, the
analytical method will be referred to as energy method.
Actually, in despite of equilibrium-geometry-physics method or energy method, their principles are
identical. The only difference is their formulating way. In this context, force method is equivalent to
complementary energy method; displacement method is equivalent to potential energy method. For the
precise solution of a simple problem, it is hard to find the advantages of energy method over
equilibrium-geometry-physics method. However, for the approximate solution of a complex problem, the
energy method has remarkable advantages over equilibrium-geometry-physics method, because energy
method formulate a complex problem as a mathematical problem for minimum value or extreme value,
which is easy to solve in the viewpoint of mathematics. Energy method is widely applied in the dynamic
and buckling analyses of structures.
(3) The methods for analyzing statically indeterminate structures can be as well grouped into two
categories based on the calculating way the methods used. They are: ① hand-orientated methods, ②
computer-oriented methods.
In hand-oriented methods, simplification of calculation is the objective of the methods. So much of the
details about the members, connections and those have little effect on the desired characteristics of a
structure must be discarded so as to simplify the calculations by hand. Thusly, much of the unknowns
which have little effect on principal behavioral characteristics of the structure are neglected when use
590 Chapter 14 General Remarks on Statically Indeterminate Structures
hand-oriented method(s) to analyze the structure. For instance, when analyze a rigid frame the axial
deformations of the members are ignored to reduce the number of unknowns so that the calculating work
by hand is decreased. In that context, displacement method is advanced over force method; potential energy
method is more powerful than complementary energy method, and vice versa.
In computer-oriented methods, routinization and automatization of the computing process for a
structure are the objective of the methods. This is the reason why matrix algebra has been applied to
structural analysis. Since the displacement method is more systematic and can be more easily implemented
on computers, most of the commercially available computer programs for structural analysis are based on the
method instead of force method.
(4) The methods for analyzing statically indeterminate structures can be also divided into two groups
according to the way for solving the canonical equations (it is actually simultaneous linear equations) the
methods adopted. They are: ① direct methods, ② successively approximate methods.
Classical force and displacement methods belong to direct methods; whereas moment distribution and
no-shear distribution methods are successively approximate methods, which actually use the joint
displacements as primary unknowns though the end moments are participated in the concrete process of
operation. Since the convergent speed of the successively approximate methods based on force method are
very slow, they are not applied to structural analysis.
(5) In the viewpoint of hand-oriented methods, there is a most favorable method for an indeterminate
structure to be analyzed. Table 14-2 shows framed structures and their corresponding most favorable
methods by which the structures can be conveniently analyzed.
Table 14-2 statically indeterminate structures and their corresponding most favorable methods
Types of structures Most favorable method
Indeterminate trusses and arches Force method
Continuous beams and rigid frames without sidesway Moment distribution method
Displacement method or no-shear distribution method

Rigid frames with sidesway
or combined method
(6) In the viewpoint of computer-oriented methods, matrix displacement method is suitable for all
kinds of structures. An all purpose computer program might be used to analyze many kinds of structures as
the program shown in section 13.9.
14.2 Combined Method—Simultaneously Applying Moment Distribution and
Displacement Methods to Frames with Sidesway 591
Displacement Methods to Frames with Sidesway
In practice, when solving a complex structural analytical problem, the complex problem might be
subdivided into several simple sub-problems. For each of the sub-problems, a most favorable analytical
method can be applied to solve it. Then, combining each of the solution will obtain the solution of the
complex problem. By this combined way, a complex problem sometimes can be solved conveniently
because the advantage of each analytical method is utilized properly.
There are many combined ways, such as the combination of force and displacement methods, the
combination of moment distribution and displacement methods, and moment distribution and no-shear
distribution methods. For various different problems, different combined methods can be applied. For an
introduction, only combination of moment distribution and displacement methods are discussed in the
following.
As we known in previous chapters, moment distribution method or no-shear distribution method alone
cannot be applied to analyzing the rigid frames with sidesway. However, if we apply both moment
distribution method and displacement method to the frames, they will be analyzed conveniently, since the
effect of joint rotation(s) can be estimated by moment distribution method and the response of the sidesway
(translation) can be valued by displacement method individually. Now we will illustrate the analytical
procedure of combination of moment distribution and displacement methods by the consideration of the
rigid frame with sidesway shown in Fig.14.1 (a).
A B C A B C F1P A B C
k11Δ1
q
= q +
Fig.14.1 Sketch of combination of moment distribution and displacement methods

(a) original structure; (b) primary structure under the action of external load;
(c) primary structure under the action of sidesway
(1) Evaluating the effect due to joint translation (sidesway)

The response due to joint translation can be evaluated by displacement method, but the primary
structure used here is different from that used by traditional way. As shown in Fig.14.1 (b), on the primary
structure only translation is restrained; in other words, only the horizontal translation is chosen as the
primary unknown whereas the joint rotations do not chosen as the primary unknowns in this step.
Consequently, by considering the restraint forces at the artificial restraint caused by the external load and
the unknown translation, we can write displacement-method equation as follow
k11Δ1 + F1P = 0
(14-1)
The expression for evaluating bending moments can be expressed as
M = M 1Δ1 + M P
(14-2)
In Eqs. (14-1) and (14-2), F1P and M P are the restraint force yielded at the artificial restraint and
bending moment in the primary structure due to the external load; k11 and M 1 are the restraint force
yielded at the artificial restraint and bending moment in the primary due to the unit displacement Δ1 = 1 .
(2) Estimating the effect due to joint rotations
For the primary structure shown in Fig.14.1 (b) or (c), the moment distribution method can be
expediently applied to determine stiffness coefficient k11 , restraint force F1P , bending moments M 1
and M P . Once the quantities are obtained the translation (or sidesway) can be determined through Eq.
(14-1) as to be
F1P
Δ1 = −
k11
After Δ1 is determined the bending moment in the frame will be evaluated by Eq. (14-2).
Example 14-1
Draw the bending moment diagram for the rigid frame shown in Fig.14.2 (a) by applying both
displacement method and moment distribution method.
Solution
(1) Start with applying displacement method. Thusly, the roller support at D is replaced by a hinged
support to prevent the frame’s sidesway as shown in Fig.14.2 (b). Since the primary structure shown in
Fig.14.2 (b) has no joint translation, moment distribution method can be applied to it conveniently, and the
final bending moment diagram obtained by the method is shown in Fig.14.3.
In order to find horizontal restraint force at D, the shears at the bottom sections of the two columns
due to the external load can be determined as to be
3.45 + 1.7
QEBP = − = −1.29kN
4
9.8 + 4.9
QFCP = = 2.45kN
6
By considering the horizontal force equilibrium condition ∑X =0 of the free body diagram for the
entire frame, the horizontal restraint force at D can be determined to be
F1P = 2.45 − 1.29 = 1.16kN(→)
q = 20kN/m
D
A
4I 0
B 5I 0
C 4I 0
4m
3I 0 3I 0
6m
E
(a) original structure

F
4m 5m 4m
q = 20kN/m
A D
4I 0
B 5I 0
C 4I 0
4m
3I 0 3I 0
6m
(b) primary structure under action of external load

F
4m 5m 4m
(2) Draw the bending moment diagram M 1 for the primary structure shown in Fig.14.2 (c) due to
the unit horizontal displacement at D. Since now the support’s displacement is given, Δ1 = 1 , we can
employ moment distribution method to the structure to determine its bending moments. Thusly obtained
bending moment diagram is shown in Fig.14.4.
In order to find stiffness coefficient k11 , the shears at the bottom sections of the two columns due to
Δ1 = 1 can be evaluated as
0.965 + 0.806
QEB = = 0.443
4
0.467 + 0.436
QFC = = 0.151
6
D Δ1 = 1
A
4I 0
B 5I 0
C 4I 0
4m
3I 0 3I 0
6m
E
assume EI 0 = 1
F
4m 5m 4m
Fig.14.2 (c) primary structure under action of Δ1 = 1
46.9
43.4 24.4
14.6 F = 1.16kN
1P
A
B C 9.8 D
18.2
3.45
(unit: kN ⋅ m)
E
1.29kN 1.7
F
4.9
2.45kN
Fig.14.3 M P diagram
By considering the horizontal force equilibrium condition ∑X =0 of the free body diagram for the
entire frame, the stiffness coefficient can be written to be
k11 = 0.443 + 0.151 = 0.594kN(→)
0.340
0.318
A 0.806 C 0.436 D
B 0.096 K = 0.594
11
0.488 Δ =11
E
0.965 Q = 0.443
EB
F
0.467 Q = 0.154
EB
Fig.14.4 M 1 diagram
(3) Determine the horizontal displacement Δ1 .

By considering the displacement equation k11Δ1 + F1P = 0 , we can determine the horizontal
displacement to be
F1P 1.16
Δ1 = − =− = −1.95
k11 0.594
(4) Draw bending moment diagram.

By using the superposition equation for bending moment, M = M 1Δ1 + M P , we will draw the final
bending moment as shown in Fig.14.5.
In the forgoing example, we illustrate the analytical procedure of combined method for the rigid frame
with only one joint translation. As to the rigid frame with more than one joint translation, similar procedure
can be considered. For instance, the analytical procedure of the combined method for a rigid frame with
two joint translations, Δ1 and Δ 2 , can be written as follows:
(1) Attach two artificial restraints associated with Δ1 and Δ 2 to the original structure to prevent
the two translations so as to obtain the primary structure for displacement method.
(2) By using moment distribution method, determine the bending moment diagram M P , the
restraint forces, F1P and F2 P , yielded at the artificial restraints.
(3) Evaluate the bending moment M 1 , stiffness coefficients k11 and k 21 for the primary
structure due to the unit displacement Δ1 = 1 .

(4) Calculate the bending moment M 2 , stiffness coefficients k12 and k 22 for the primary
structure due to the unit displacement Δ 2 = 1 .
(5) Determine Δ1 and Δ 2 by using displacement-method equation
k11Δ1 + k12 Δ 2 + F1P = 0 ⎫

⎬
k21Δ1 + k22 Δ 2 + F2 P = 0 ⎭
(6) Draw final bending moment diagram by employing the superposition equation
M = M 1 Δ1 + M 2 Δ 2 + M P .
47.8
42.8 43.7
14.8
A D
5.0 B C 8.9
18.5
26.7
47.8
E 3.6
(unit: kN ⋅ m)
4.0 F
14.3 Approximate Analysis of Statically Indeterminate Structures
The analysis of statically indeterminate structures using the force and displacement methods
introduced in the preceding chapters can be considered as exact in the sense that the compatibility and
equilibrium conditions of the structure are exactly satisfied in such an analysis. However, exact analysis of
indeterminate structures involves computation of displacements and solution of simultaneous equations, so
it can be quite time consuming without assistance of computers. If we discard some of the unimportant
details of an indeterminate complex structure to simplify the analysis of the structure, approximate
method(s) will be developed. In fact, approximate analysis proves to be quite convenient to use in the
planning phase of projects, when several alternative designs of the structure are usually evaluated for
relative economy. The results of approximate analysis can also be used to estimate the sizes of various
structural members needed to initiate the exact analysis. The preliminary designs of members are then
revised iteratively, using the results of successive exact analyses, to arrive at their final designs.
Furthermore, approximate analysis is sometimes used to roughly check the results of exact analysis, which
due to its complexity can be prone to errors. Numerous methods have been developed for approximate
analysis of indeterminate structures. In the section, two approximate methods, single-storey method and
inflection-point method, will be introduced as follows.
14.3.1 Approximate Analysis of Multistory and/or Multibay Rigid Frames Subjected to Vertical
Loads
Unlike the exact methods, which are general in the sense that they can be applied to various types of
structures subjected to various loading conditions, a specific method is usually required for the approximate
analysis of a particular type of structure for a particular loading. Now we will introduce so called
single-storey method used to approximately analyze multistory and/or multibay rigid frames or rectangular
building frames subjected to vertical loads. Based on the behavioral characteristics of the multistory and/or
multibay rigid frame and the experience gained from the exact analyses of the types of the frames subjected
to vertical loads only, two simplifying assumptions about the frames can be made for the method.
(1) Neglect the effect due to the sidesway of a frame so as to moment distribution method can be
applied to the frame.
(2) Discard the effect of the loading acting on the girder(s) lying in the same storey on the other
girder(s) lying in the different stories. The assumption can make the girders of a multistory frame to be
analyzed storey by storey.
Consider, for example, the four-storey and three-bay rigid frame shown in Fig.14.6 (a). If we apply the
two assumptions to the frame, its analytical model will become four single-storey rigid frames as shown in
Fig.14.6 (b). Note that each of the columns except for the bottom ones belong to two adjacent stories, so the
bending moments obtained from the two stories of the columns must be added together.
The rationality of the first assumption can be stood for by the small sidesway of a multistory and/or
multibay rigid frame under vertical loading; whereas the rationality of the second assumption will be
explained as that: ① Unbalance moments will be yielded at the joints of some storey under the action of
its vertical loads. The unbalanced moments will be distributed to the nearends and carry over to the farends
of the columns of the storey. ② The carryover moments will further influence the columns adjacent to the
stories. However, the carryover moments have already gone a long way and become too small to affect
their neighbor columns. So the carryover moments can be neglected. Obviously, this argument is the basis
of the second assumption.
It must be pointed out that the farends of all single-storey frames except for bottom one are assumed
fixedly (or rigidly) supported. In practice, the farends are not rigidly supported; they are elastically
EI
supported. In order to reflect the characteristics of elastic supports, the relative rigidity, i.e. , will be
l
1 1
reduced by timing the relative rigidity by 0.9, and the carryover factor is changed from to .
2 3
Ⅳ Ⅲ
Ⅰ Ⅱ
(a)
Ⅰ
(b)
Fig.14.6 Sketch of single-storey method
(a) original structure; (b) single-storey frames
It is observed that the analyzed results of single-storey method will generally yield unbalanced
moments at the rigid joints of the frame. However, the experiences of analyzers have proved that the
unbalanced errors are generally very small. If necessary, the unbalanced moments can be distributed once
again.
14.3.2 Inflection-Point Method for Analyzing Multistory and/or multibay Rigid Frames Subjected
to Horizontal Loads
The behavior of multistory and/or multibay frames is different under lateral (horizontal) loads than
under vertical loads, so different assumptions must be used in the approximate analysis for lateral loads
than were used in the case of vertical loads considered previously. There are several methods used for
approximate analysis of multistory and/or multibay frames subjected to lateral loads. This subsection only
presents inflection-point method which is a commonly used method for analysis of multistory and/or
multibay frames with strong girders and weak columns subjected to lateral loads.
The fundamental assumption of inflection-point method is that the stiffnesses of the girders of the
frames are infinitely great such that the shear distribution method can be applied to the frames.
In order to illustrate inflection-point method, consider the portal frame shown in Fig.14.7 (a) whose
bending moment diagram has been drawn in Fig.14.7 (b). Because the ratio between the relative flexural
Ph
Δ 0.947
4
P P
ib Ph
θ 0.947
4
h
ic ic
ib / ic = 3
Ph Ph
1.053 1.053
4 4
Fig.14.7 A portal frame exactly analyzed

(a) original structure; (b) M diagram
Ph
Δ Ph
P θ =0 4 4
P
ib = ∞
2
h
ic ic P P P
Q= Q= Q=
h
ib Q=
P
=∞ 2 2 2
2
2
h
ic
Ph Ph
4 4
Fig.14.8 A portal frame analyzed by inflection-point method
(a) original structure; (b) M diagram
ib
rigidity of the beam and that of columns is 3, i.e. = 3 , the frame belongs to the frames with strong
ic
beams and weak columns. So we can use inflection-point method to analyze the frame. Thusly, the flexural
rigidity of the beam is assumed to be infinitely great, i.e. ib = ∞ . The frame will become into an ideal
frame as shown in Fig.14.8 (a), whose deflecting behavioral characteristic is that its rigid joints have
sidesway but no rotation, and bending-moment characteristic is that the bending moments at the midpoints
of the two columns are equal to zero as shown in Fig.14.8 (b), which has drawn by using shear distribution
method. Based of the assumption and simplification, we can obtain that:
P
(1) The shear of each column is equal to by consideration of the symmetric property of the
2
frame.
h Ph
(2) The bending moments at the ends of the columns are equal to M = Q × = by consideration
2 4
of the inflection points lying at the midpoints of the columns. Thusly, the bending moment diagram can be
drawn out as shown in Fig.14.8 (b).
(3) The bending moments of the beam can be determined by using the equilibrium conditions of the
rigid joints, and then the bending moment diagram can be drawn as shown in Fig.14.8 (b).
It can be observed from comparison between Fig.14.7 (b) and Fig.14.8 (b) that only 5% error occurs
when uses approximate method to analyze the frame. It implies that if the ratio between the relative rigidity
i
of beams and that of columns is great than 3, i.e. b ≥ 3 , the inflection-point method will obtain a quite
ic
exact analysis for the frames with strong beams and weak columns under lateral loads.
Inflection-point method can be also applied to the frames with different columns. Consider, for
example, the frame shown in 14.9 (a), whose beam has infinitely great rigidity, the relative rigidity and
height of left column are i1 and h1 , and those for right ones are i2 and h2 , respectively. Since the
sidesways of the two columns are identical the shears of the two columns will be calculated to be [see
Fig.14.9 (c)]
12i1 ⎫
Q1 = 2
Δ = d1Δ ⎪
h1 ⎪
⎬ (a)
12i2
Q2 = 2 Δ = d 2 Δ ⎪
h2 ⎪⎭
12i
In which, d = is referred to as stiffness coefficient of the column due to its sidesway, i.e. the shear
h2
induced by a unit sidesway at the top of the frame.
By consideration of the equilibrium condition of horizontal forces of the frame, we write
Q1 + Q2 = P
(b)
Substituting Eq. (a) into Eq. (b), the two shears will be determined to be
d1 ⎫
Q1 = 2
P = μ1 P ⎪
∑i =1
di ⎪
⎪
⎬ (c)
d2
Q2 = 2 P = μ2 P ⎪
⎪
∑ i =1
di ⎪
⎭
dj
Where, μ j = , known as shear-distribution factor.
∑ di
(a) (b)
Qh Qh
Δ Δ
1 1 2 2
P θ =0 P 2 2
ib = ∞
h1
2
h2
2
1
h
i1 i2
2
h
Q 1
Q
h1
2
2
h2
2
Qh 1 1
Q
1
Qh
2 2
2 2 Q 2
Δ (c)
12i
Q= Δ = dΔ
h
2
h
Fig.14.9 A frame with different columns analyzed by inflection-point method

(a) original structure; (b) bending moment diagram; (c) shear of a column
Q
It is observed from above that the shear distributed to a column is directly proportional to its
sidesway-stiffness coefficient d . Thusly, the horizontal load P is distributed to each column by means
of its shear-distribution factor. Once the shears in columns are determined the bending moments will be
determined by using the condition that the bending moments on the middle sections of the columns are
equal to zero, and thus obtained bending moment is drawn in Fig.14.9 (b).
Based on forgoing discussion, we can summarize some points used for approximate analysis of
multistory and/or multibay frames by inflection-point method as follows:
(1) When the ratio between the relative flexural rigidity of beam(s) and that of column(s) is great
i
than 3, i.e. b ≥ 3 , inflection-point method might be used to analyze multistory and/or multibay frames
ic
subjected to lateral loads.
(2) The fundamental assumption of inflection-point method is that the flexural rigidity of the beam(s)
of a frame is infinitely great so that the rigid joints of the frames only translate but not rotate.
(3) When the columns in the same storey have the same sidesway, the shears in the columns induced
by the sidesway are directly proportional to their sidesway-stiffness coefficients. All the columns locating
within some storey undertake all the lateral loads above the storey, which are then distributed to each
column of the storey by means of its shear-distribution factor.
(4) The bending moments of the columns are induced by their sidesways, so the inflection points are
located at the midpoints of the columns. Note that for multistory rigid frames, the inflection point of a
column lying on bottom storey is assumed to be located at a height of 2/3 length of the column.
(5) The bending moments of the columns are determined by their shears and the locations of their
inflection points. While the bending moments at the ends of the beams connecting to the side columns are
determined by equilibrium of the corresponding rigid joints; the bending moments at the ends of beams
connecting to the middle columns are the distribution of the unbalanced moments which are distributed by
means of the bending stiffness of the beams connected to the joints.
It must be pointed out that when shear distribution method is applied to the frames with infinitely great
beam(s), it is a sort of exactly analytical method; inflection-point method is a special kind of shear
distribution method, and it is an approximately analytical method when it applied to the rigid frames with
strong beams and weak columns.
Example 14-2
Analyze and draw the bending moment diagram for the rigid frame shown in Fig.14.10 (a) by
inflection-point method. Note that the digits enclosed in the parentheses represent the corresponding
relative flexural rigidities of the members.
Solution
Simplifying the frame In order to simplify the frame, we assume that all the inflection points are
located at the midpoints of the columns. Note that although the inflection points of bottom columns should
be located at the positions of 2/3 heights of the columns by the forgoing argument, the relative rigidities of
the bottom-storey beams are more than three times of those of the bottom columns. Thusly, the assumption
is rational for approximate analysis. If we pass imaginary sections through the inflection points the free
body diagrams of the frame will be shown as in Fig.14.10 (b). Some calculations will be made as follows.
d
(1) Determination of each shear distribution factor μ j = j for the columns
∑ di
i
The shear distribution factors for the columns on top storey:

2
μGD = μ IF = = 0.286
2× 2 + 3
3
μ HE = = 0.428
2× 2 + 3
The shear distribution factors for the columns on bottom storey:
3
μ AD = μCF = = 0.3
3× 2 + 4
4
μ EB = = 0.4
3× 2 + 4
(a) (b)
8kN 8kN
G (12) H (15) I
3.3m
G H I
(2)
(2)
QHE
(3)
QGD QIF
17kN D (12) E (15) F 17kN D E F

3.6m
(3)
(3)
(4)
QAD QBE QCF

B
A C B C
A
3.6m 4.5m

(a) original structure; (b) free body diagram
(2) Evaluation of shears for the columns

By using Q j = μ j P , the shears can be evaluated as
QGD = QIF = 0.286 × 8 = 2.29kN

QHE = 0.428 × 8 = 3.42kN
QAD = QCF = 0.3 × (8 + 17) = 7.5kN
QEB = 0.4 × (8 + 17) = 10kN
(3) Evaluation of bending moments

The following calculation will take joint E as an example.
Bending moments at the ends of the columns
h2 3.3
M EH = −QHE × = −3.42 × = −5.64kN ⋅ m
2 2
h1 3.6
M EB = −QEB × = −10 × = −18kN ⋅ m
2 2
Bending moments at the ends of the beams
The summation of the bending moments of the columns connected to joint E will be written as
m = M EH + M EB = −5.64 − 18 = −23.64kN ⋅ m
The distributed bending moments at the ends of the beams connected to joint E will be evaluated as
12
M ED = × 23.64 = 10.51kN ⋅ m
27
15
M EF = × 23.64 = 13.13kN ⋅ m
27
( 3.07 )
( 3.63) 2.51 ( 6.61) ( 3.90 )
8kN 3.78 H 5.64 3.78
G 3.13 I
3.78 ( 3.53)
( 3.63)
10.51 17.28
(11.52 ) (15.91)
3.78 13.5 18
( 2.85) D (12.11) E (18.41) F 13.5
17kN 5.64 3.78 (12.65)
( 6.14 ) ( 3.26 )
13.13
(13.42 )
17.28
(14.95) (unit: kN ⋅ m)
A B C
13.5 18 13.5
(13.63) (19.31) (13.90 )
Fig.14.11 shows the bending moment diagram of the frame. Note that the digits enclosed in the
parentheses are the values of the bending moments evaluated by an exact method.
14.4 Properties of Statically Indeterminate Structures
In this section, we will focus our attention on the properties of statically indeterminate structures.
During the discussion, some principle of force method and comparison between the indeterminate and their
comparable determine will be employed.
14.4.1 Redundant Restraints and Their Influence
As discussed previously, the support reactions and internal forces of statically determinate structures
can be determined from the equations of equilibrium (including equations of condition, if any). However,
(a) P (b) P
Fig.14.12 Bending moment distributions and comparison of maximum bending moments

(a) bending moment diagram of an indeterminate beam;
(b) bending moment diagram of the comparable determinate beam
since the indeterminate structures have more support reactions and/or members than required for static
stability, the equilibrium equations alone are not sufficient for determining the reactions and internal forces
of such structures and must be supplemented by additional relationships based on the geometry of
deformation of structures. These additional relationships, which are termed the compatibility conditions,
ensure that the continuity of the displacements is maintained throughout the structure and that the
structure's various parts fit together. The existence of redundant restraints, i.e., the excess support(s) and/or
member(s), will change the stressing properties of a structure in the following aspects.
(1) Stronger defense abilities. Unlike a statically determinate structure of which if one of supports or
members is collapsed the entire structure will be failed as well, a statically indeterminate structure, if
properly designed, has the capacity for redistributing loads when certain structural portions become
overstressed or collapse in cases of overloads due to earthquakes, tornadoes, impact (e.g., gas explosions or
vehicle impacts), and other such events. So if a part (or member or support) of such a structure fails, the
entire structure will not necessarily collapse, and the loads will be redistributed to the adjacent portions of
the structure because indeterminate structures have more members and/or support reactions than required
for static stability. (2) Uniformly internal force distributions and smaller stresses. The influence extent of
local loading in a statically indeterminate structure is generally great than that in its comparable
determinate structure, but the maximum stress in a statically indeterminate structure is generally lower than
that in its comparable determinate structure. Consider, for example, the statically determinate and
indeterminate beams shown in Figs.14.12 (a) and (b), respectively. The bending moment diagrams and their
elastic deforming curves for the beams due to a concentrate load P are also shown in the figure. It can be
seen from the figures that the bending moment distribution extends in the whole indeterminate beam but in
local portion of the determinate beam, and the maximum bending moment in the indeterminate beam is
significantly lower than in the determinate beam. This implies that the internal force distribution of an
indeterminate structure is more uniform and the maximum bending stress in the structure is lower than
those of its comparable determinate structure.
q q q
EI EI EI
l l l
f = 0.0026ql 4 / EI f = 0.0054ql 4 / EI f = 0.013ql 4 / EI

(a) (b) (c)
Fig.14.13 Comparison of maximum deflections

(a) a two fixed-support beam;
(b) one fixed-support and one roller-support beam; (c) simply supported beam
(3) Greater stiffness and stabilities.

P1 P2 Statically indeterminate structures generally have
higher stiffnesses than those of comparable
determinate structures. From Fig.14.13, we
observe that the maximum deflection of the
indeterminate beam with one fixed support and
l
one roller support is only 0.42 that of the simply

supported determinate beam, and that the
maximum deflection of the two-fixed-support
P1 indeterminate beam is only 0.2 that of the simply
P2
supported determinate beam.
Fig.14.14 Comparison of critical loads
(a) a simply supported column; Statically indeterminate structures generally
(b) an indeterminate column have higher stabilities than those of comparable
determinate structures. From Fig.14.14, we can see
that the critical force of the column with one fixed support and one axially glidingly fixed support is only
0.25 that of the simply supported column.
14.4.2 Influence due to Changes of Member Stiffness
Unlike a statically determinate structure whose stiffness changes of the members will not influence the
internal force distribution of the structure, the internal force distribution of a statically indeterminate
structure is strongly depended on its member’s relative stiffness. The more rigid is a member the more load
will be distributed to it. This can be further explained by referencing to canonical equations of force method
[see Eq. (10-5)]. Recalling from the equation that the flexibility coefficients and free terms vary with the
member’s stiffnesses, if the member’s stiffnesses are changed the flexibility coefficients and free terms in
the force-method equation are changed as well, subsequently the internal force distribution in the structure
will be also changed. However, if only change the magnitudes of the member’s stiffnesses of an
indeterminate structure but remain the proportions between the stiffnesses unchangeable, the proportions
between the flexibility coefficients and free terms—subsequently the distribution of internal forces in the
structure—will also unchangeable. Accordingly, the internal force distribution of an indeterminate structure
due to external loads is only related to the proportions between the member’s stiffnesses but not the
magnitudes of the stiffnesses. That is, the internal force distribution changes with the relative stiffnesses of
the members but not the absolute stiffnesses of the members. This implies that when we analyze the
internal forces of an indeterminate structure only the relative rigidities of the structure’s members are
necessary.
Since the analysis of an indeterminate structure involves, in addition to the dimensions and
arrangement of members of the structure, its cross-sectional and material properties (such as cross-sectional
areas, moments of inertia, moduli of elasticity, etc.), which in turn, depend on the internal forces of the
structure. The design of an indeterminate structure is, therefore, carried out in an iterative manner, whereby
the (relative) sizes of the structural members are initially assumed and used to analyze the structure, and the
internal forces thus obtained are used to revise the member sizes; if the revised member sizes are not close
to those initially assumed, then the structure is reanalyzed using the latest member sizes. The iteration
continues until the member sizes based on the results of an analysis are close to those assumed for that
analysis. However, the design of a determinate structure will not need the iterative manner.
It is obviously that we can utilize the characteristic, i.e., the more rigid is a member the more load will
be distributed to it, to adjust the relative stiffnesses of the members of an indeterminate structure so as to
modulate the internal force distribution of the structure and to implement a rational structural design.
Consider, for example, the bending moment diagram of a portal frame shown in Fig.14.15 (a). If we
increase the cross-sectional size of the girder and reduce the cross-sectional sizes of the columns—and
subsequently the member’s relative stiffnesses will be changed—the bending moment diagram will
approach the case shown in Fig.14.15 (b), where the bending moment diagram in the girder is close to that
of a simply supported comparable girder, and the maximum bending moment occurred on the middle
section is too great to be properly adopted. Contrarily, if we decrease the cross-sectional size of the girder
q = 20kN/m
53.3 I1 53.3
C D
160
6m
I2 106.7 I2
(unit: kN ⋅ m)
A B
8m
(a)
q = 20kN/m
q = 20kN/m
I1
≈ 106.7 160 ≈ 106.7
≈0 ≈0
I1
160 ≈ 53.3
I2 I2
I2 ≈ 160 I2
(unit: kN ⋅ m) (unit: kN ⋅ m)
A B A B
(b) (c)
Fig.14.15 Changes of bending moment diagrams with member's stiffnesses

(a) I1 = 2 I 2 ; (b) I1 I 2 ; (c) I1 I 2
and increase the cross-sectional sizes of the columns—and subsequently the member’s relative stiffnesses
will be changed—the bending moment diagram will approach the case shown in Fig.14.15 (c), where the
bending moment diagram in the girder is close to that of a comparable girder with two fixed supports, and
the maximum bending moment occurred at the top ends of the columns is much greater than that on the
girder’s middle-section. In practice, this kind of internal force distribution is as well not properly adopted.
Obviously, it is wise to modulate the cross-sectional sizes of the columns and the girder to make the
maximum bending moments occurred in both the girder and the columns are closed to each other.
14.4.3 Influence due to Temperature Changes and Support Settlements
The argument that if there is no load applying on a structure there will be no internal force yielded in
the structure, is only true for a determinate structure but not for an indeterminate one. Actually, temperature
changes and support settlements do not cause any stresses in determinate structures; they may, however,
induce significant stresses in indeterminate structures, which should be taken into account when designing
indeterminate structures. The reason is that the temperature changes and/or support settlements will cause
deformations of members which are restrained by the redundant restraints of an indeterminate structure;
consequently the restraint deformations will induce internal forces in corresponding members. The internal
forces induced by the temperature changes and/or support settlements and/or the like other than external
loads are referred to as self-induced internal forces in Structural Mechanics. Obviously, self-induced
internal forces can be existed in an indeterminate structure but not in a determinate structure.
Furthermore, the self-induced internal forces are directly proportional to absolute values of the
member’s stiffnesses of an indeterminate structure, i.e., the larger is the magnitude of a member’s rigidity
the greater is the self-induced internal force in the member. The argument can be further explained by
follows.
Consider the canonical equations of force method used for analysis of an indeterminate structure
subjected to temperature changes and support settlements, which can be rewritten as
δ11 X 1 + δ12 X 2 + " + δ1n X n + Δ1t + Δ1c = 0 ⎫

δ 21 X 1 + δ 22 X 2 + " + δ 2 n X n + Δ 2t + Δ 2c = 0 ⎪⎪
⎬
"" ⎪
δ n1 X 1 + δ n 2 X 2 + " + δ nn X n + Δ nt + Δ nc = 0 ⎭⎪
The corresponding superposition equations used for calculate internal forces will be written as
M = M1 X1 + M 2 X 2 + " + M n X n ⎫
⎪
N = N1 X 1 + N 2 X 2 + " + N n X n ⎬
⎪
Q = Q1 X 1 + Q2 X 2 + " + Qn X n ⎭
Above equations are the general expressions used to express the self-induced forces in an nth
It can be seen in the canonical equations that the flexibility coefficients δ ij are inversely proportional
to the member’s rigidities ( EI , EA and GA ), and that the free terms, Δ it and Δ ic , have no relationship
with the rigidities because they are determined by the equations
αΔt
Δit = ∑ ∫ M ds + ∑ α t ∫ N ds
i 0 i
h
Δ ic = ∑ RK cK
In which, there is no quantity involved with the rigidities.

It can be seen from above discussion that the self-induced internal forces (induced by temperature
changes and/or support settlements and/or etc.) are directly proportional to the absolute values of the
member’s rigidities of an indeterminate structure, i.e., if increase the lateral dimensions of the
members—actually increase the rigidities of the members—the self-induced internal force in the members
will be increased as well. Obviously, it is not a wise measure to increase the lateral sizes of the members of
an indeterminate structure in order to improve the ability to resist their temperature changes and/or support
settlements and/or the like.
Based on above discussion, following two aspects must be paid attention to when designing an
(1) Avoid the occurrence of the self-induced internal forces so as to relieve or eliminate the
disadvantageous influence caused by the self-induced internal forces.
Consider, for instance, the indeterminate rigid frame shown in Fig.14.16. If the entire structure settled
a distance Δ , i.e., the four columns A, B, C and D have an identical settlement Δ as shown in Fig.14.16
(a), the settlement will not induce internal forces in the rigid frame. However, if the four columns have
different settlements as shown in Fig.14.16 (b), the settlements will really yield internal forces in the frame.
A B C D A B C D
A′ B′ C′ D′ C′ D′
Δ
(b)
(a)
Fig.14.16 Sketch of support settlements
(a) even settlements; (b) uneven settlements
In this context, uneven support settlements of indeterminate structures are the cause to induce internal
forces in the structures, to which designer(s) must pay attention in practical design.
Another consideration to avoid the occurrence of self-induced internal forces in a long building due to
temperature changes is that temperature joint(s) is arranged in the building. In practice, the foundation of a
building is buried under the ground; while its superstructure is exposed in the atmosphere. So the
deformation of the foundation due to temperature changes is smaller than that of the superstructure, which
will subject to larger temperature changes during its lifetime. Furthermore, the deformation—consequently
temperature-induced stress—varies proportionally with the length of the building. Therefore, if a so-called
temperature joint is arranged in the superstructure of a long building, which will separate the superstructure
along its length into several portions, the temperature-induced stress will be greatly reduced in the building.
(2) Employ prestressed structures to modulate the internal forces of the indeterminate structures
during their lifetime by means of the forces deliberately provided by the prestresses.
14.5 Complementary Discussion about Computing Models of Structures
In Chapter 1, we have already given some discussion about the development of the computing model
of a structure. The section will present further complementary discussion about the establishment of
computing models.
14.5.1 Simplification of Structural System
Although all practical structures are generally three-dimensional (or space) structures based on the
viewpoint of member’s arrangement and their connections, many of them can be idealized as plane
structures for simplifying the analysis by means of their constructions and stressing characteristics. The
consideration made for simplifying analysis generally has following aspects.
(1) Take a plane element as the analytical model.
For a long prismatic structure with constant cross
sections or a structure consisting of a series of identical
plane structures, a plane element of them can be utilized as
the analytical model.
Consider, for example, a tunnel whose cross sections
remain constant along the direction of its length, are
generally subjected to unchangeable loads along the axis
of the tunnel. Thusly, an annular ring, obtained by two
Fig.14.17 Cross section of a circular tunnel
imaginary adjacent cuts with a unit distance and
perpendicular to the axis of the tunnel, is developed as the
computing model of the tunnel, as shown in Fig.14.17 for a circular tunnel.
(2) Plane structures arranged in lateral and longitudinal directions are adopted as the computing
models.
Now let us focus our attention on a multistory and multibay building frame shown in Fig.14.18. The
frame is composed by the beams and the columns to form a space structure. However, the structural
analysis of the three-dimensional frame adopts two kinds of simple plane rigid frames as its analytical
models, as shown in Figs.14.18 (b) and (c). In practice, the loads applying on the frame are its weight
(vertical load), likely wind load and seismic load (horizontal loads). Obviously, the lateral stiffness or
ability of resisting sidesway due to horizontal loads is weaker than its longitudinal one. So the response of
(a) column
7m 3m 7m
lateral direction
longitudinal direction
4m 4m 4m 4m 4m 4m 4m 4m 4m 4m 4m
(b) (c)
6m 5m 5m
Fig.14.18 Plane sections of a space rigid frame

(a) arrangement of columns; (b) longitudinal rigid frame; (c) lateral rigid frame
the lateral plane rigid frame due to the external loads is more considerable than that of the longitudinal
one. Thusly, in practice, the response of the lateral plane rigid frame shown in Fig.14.18 (c) due to
horizontal and vertical loads is analyzed, and the response of the longitudinal plane rigid frame only due
to the seismic loading is concerned.
(3) Take a presentative plane structure as a computing model of a structure.
Some of the space structures are composed by several kinds of plane structures. Under some
conditions they can be simplified as a presentative plane structures to analyze. The establishment of a
presentative plane structure needs following considerations.
① Summarize all kinds of substructures (or plane structures) composing a space structure as a
presentative plane structure, and determine the loads distributed to the substructures.
② Analyze the substructures subjected to their distributed loads.
Consider, for example, a tall building structure composed by two kinds of substructures—plane shear
wall and plane rigid frame—whose column-arrangement map is shown in Fig.14.19 (a). Now we intend to
analyze the response of the building due to its horizontal loading (wind and/or seismic load). It can be seen
from Fig.14.19 (a) that the horizontal loading is undertaken by 12 sets of plane substructures—two sets of
plane shear wall structures and ten sets of plane rigid frame structures. For the purpose of simplifying
analysis, two assumptions are made as follows.
ⓐ The in-plane rigidity of a floor is infinitely great. The assumption ensures that the cross sections
of the shear walls and floor slabs lying on the same storey have an identical sidesway.
ⓑ The plane shear walls and the plane rigid frames are arranged symmetrically. The assumption
assures that there is no torsion deformation happened in the structure under the action of symmetric
horizontal loadings.
shear wall rigid frame column
(a)
shear wall rigid bar shear wall rigid frame

rigid frame
(b) (c)
Fig.14.19 Computing model for frame and shear-wall structure

(a) arrangement map for columns and shear walls;
(b) a presentative plane structure for frame and shear-wall structure
Based on above two assumptions, the analysis of the structure under the action of lateral loads may be
simplified as the analysis of a presentative plane structure shown in Fig.14.19 (b), in which two sets of
shear wall structures are composed as one set of the structure and ten sets of plane rigid frames are
synthesized as one set of the structure as well; whereas the connections between the two presentative plane
structures are rigid bars as shown in Fig.14.19 (b).
Furthermore, if we neglect the rigidity of connecting beams between the two shear walls shown in
Fig.14.19 (b) they will be again synthesized as one shear wall as shown in Fig.14.19 (c).
When the computing model of the structure is idealized into the models shown in Fig.14.19 (b) or (c),
the analysis of the structure will become the analysis for a plane structure, which will be conveniently
analyzed by routine computer program for plane structures. Once the internal forces of the links between
the shear wall and the rigid frame are determined, the internal forces in shear walls and in the members of
the rigid frames can be obtained easily.
14.5.2 Simplification of Bars
In Chapter 1, some of the general rules are presented for the purpose of simplifying various bars and
their connections. Followings are some complementary fundamentals about the simplification of general
sense bars.
(a) (b)
Fig.14.20 Computing model of a bent frame applied in industrial buildings

(a) a bent frame applied in industrial buildings; (b) computing model of the bent frame
(1) Use straight bars to represent slightly curved or folded bars.

Consider, for example, the bent frame applied to the structure of an industrial building shown in
Fig.14.20 (a). Since the shape of the roof will not effect the internal forces of the columns the roof can be
idealized as a straight link, and although the central lines of the cross sections of the columns of the frame
are not straight lines they can be idealized as the straight lines as shown in Fig.14.20 (b) for the purpose of
simplifying the analysis, in which each column is composed by two different straight lines. The height of
the column is the distance between the top of the foundation and bottom chord of the roof; the span of the
bent frame is the distance between the central lines of the two bottom columns since the span will have no
effect on the internal forces of the columns when the roof is hinged and the loads apply on the columns. If
meeting the case that the roof is rigidly connected with the columns the distance between the two top
columns will be fitted to be the span of the frame.
Another example is the portal frame shown in Fig.14.21 (a). We can see from the figure that the
central lines of the cross sections of the two inclined beams are not straight lines and those of the two
columns are not straight vertical lines. For simplifying the analysis, the straight line that cross the centroid
of the cross section at the middle section of the beam and be parallel to the outer surface of the beam is
adopted as the axial line of the beam; while the vertical line crossing the centroid of the bottom end of the
column is used as the axial line of the column. Thusly obtained line diagram will form the computing
model of the portal frame as shown in Fig.14.21 (b).
It must be noted that the internal forces obtained from the computing model are the internal forces on
the cross sections lying on the axis of the computing model.
(a) (b)
Fig.14.21 Computing model of a portal frame

(a) a portal frame; (b) computing model of the frame
(2) Substitute solid bar for latticed member.

In the analysis for a complicated structure, a latticed member is often substituted by a solid bar.
Again let us pay our attention on the bent frame shown in Fig.14.20 (a). As mentioned above that since
the shape of the roof has no effect on the internal forces in the columns due to the loads applying on the
columns, when analyze the internal forces in the columns the latticed roof will be idealized as a solid bar
(or link) as shown in Fig.14.20 (b). However, when the internal forces in the members of the latticed roof
are interested, the computing model of the roof must be simplified as a simply supported truss subjected not
only the roof joint loads but also the longitudinal forces acting along the bottom chords, which are obtained
from the analysis of the bent frame shown in Fig.14.20 (b).
(3) Simplification of member’s stiffness
The simplification of member’s stiffness has already utilized indirectly in the previous analysis. For
example, the axial deformations of the members of a rigid frame are neglected when evaluate the
displacement(s) of the frame. Obviously, the simplification implies that the axial rigidities of the members
of the rigid frame are assumed to be infinitely great. Another instance is that when we meet the case that the
girder’s rigidity is much greater than that of the columns of a rigid frame, the girder’s rigidity will be
assumed to be infinitely great for the analysis due to horizontal loads. The utilization of the simplification
can reference to subsection 14.3.2.
14.5.3 Simplification of Joints
Now let us have some complementary discussions about the simplification of the connections between
the members of a structure.
(a) (b)
Fig.14.22 Simplification of connections

(a) analyze as a truss; (b) analyze as a rigid frame
The first consideration about the simplification of the connections is that except for the construction of
the connections, the geometric construction of the structure must be taken into account. In practice, it is
likely happened that in the viewpoint of geometric construction, the connections of the members of a
structure are rigid, but the bending moments transferred through the connections are very small. In this
context, the connections can be still idealized as hinged joints. Consider, for example, the structure shown
in Fig.14.22 (a). Although the members are almost rigidly connected they can be simplified as hinged joints
to form the structure as a simply supported truss because the bending moments yielded at the connections
are negligible. However, the connections of the structure shown in Fig.14.22 (b) must be simplified as rigid
joints to form the structure to be a rigid frame, insomuch as the bending moments are the principal internal
force components on the cross sections of the ends of the members meeting at the joints.
As to the viewpoint of geometric construction of a structure, the difference between a truss and a rigid
frame is that although the joints of the truss are hinged together the arrangement of the members ensure the
truss still to be a stable structure; whereas the geometric stability of the rigid frame is guaranteed by the
rigid joints, i.e., if the rigid joints are changed into hinged ones the system will be unstable. In practical
engineering, although the constructions of the connections between steel trusses and reinforced concrete
trusses are almost closed the internal forces of the trusses are quite different from those of rigid frames.
Axial force is the dominated internal force component, whereas bending moment is only subsidiary internal
force component in the members of the trusses. This is the reason why the connections of a truss are
generally idealized into hinged joints.

Another consideration about the simplification of the connections is the effect on the member forces
caused by a connecting region through which the connection between members is implemented. In practice,
every connection between members needs a connecting region. If the connecting region is small enough to
have little effect on the forces in the members connected to it, the connection can be simplified as a joint,
whose dimension is never concerned. However, if the connecting region is large enough (say 1/5 of a
member’s length) to have considerable effect on the forces in the members connected to it, the effect must
take into account. A roughly consideration about the effect of a connecting region is that the region is
regarded as a rigid area and the portions of the members located within the rigid region are recognized as
being rigid, i.e., they have infinitely great rigidities.
14.5.4 Simplification of Supports
In Chapter 1, we discussed the simplification about complete rigid supports. In the subsection, we
focus out attention on the simplification of elastic supports, which not only produce reactions but also
displacements. Generally, the ratios between the reactions and their corresponding displacements remain
constant. The constant ratios are known as the stiffnesses of the elastic supports. In practice, when the
rigidity of a structure approaches the rigidity of its supports the supports are generally fitted to simplify as
elastic ones.
The concept of supports can be generalized. For instance, if the response of a portion of a structure is
interested, the other portion(s) of the structure can be simplified (if conveniently) as the elastic support(s)
of the desired portion, whose stiffness is related to the rigidity of the portion(s).
Consider, for example, the rigid frame shown in Fig.14.23 (a). Assume that only the internal forces in
beam AB are desired. The beam AB can be idealized as the beam shown in Fig.14.23 (b), whose supports
are elastic. The reaction moments M A and M B yielded at the two elastic supports A and B are
proportional to the end rotations θ A and θB , that is,
M A = S A*θ A M B = S B*θ B
Where S A* and S B* are the rotational stiffnesses of the two elastic supports. Actually, they are the
bending stiffnesses of the members meeting joints A and B except for member AB, i.e., the moments needed
to let members AC and AD, BE and BF in Fig.14.23 (a) rotate a unit angle, respectively. They can be
expressed as
S A* = S AD + S AC S B* = S BE + S BF
Note that when the rigidity of the supports is greatly larger or greatly less than that of their supporting
structure, the supports may be idealized as those discussed in Chapter 1.
q q
C F A B
A B
θ A θ B
M =Sθ
A
∗
A A
M =Sθ
B
∗
B B
(b)
D E
(a)
Fig.14.23 Simplification of elastic supports

(a) an original structure; (b) computing model of beam AB
SUMMARY
In this chapter, an all-around discussion has made about the fundamental analytical methods, the
characteristics of statically indeterminate structures and the selection of computing models used for the
analysis of the structures.
(1) The attention for the discussion about the fundamental analytical methods is mainly paid to the
hand-oriented methods. The selection of the methods must be aimed at a concrete problem. This implies
that the type of an analytical method is related to the type of loads applying on a structure as well as the
type of the structures analyzed by the method in order to simplify the analysis. For the structures having too
much unknowns, the matrix displacement method should be employed so as to utilize the assistance of
computers.
(2) Because of the redundant restraint(s), a statically indeterminate structure has some special
characteristics which its comparable statically determinate one does not have. The advantages of the special
characteristics might be utilized and their disadvantages could be avoided if we have a complete cognition
about them.
(3) The general rule for the establishment of computing models can be concluded into the two
implementing principles: ① starting with practicality, ② distinguish the dominating and the subsidiary.
So-called ‘starting with practicality’ means that before the computing model of a structure is set up,
the arrangement of the members, the geometric construction and the likely behavior of the structure must
be soundly understood. While so-called ‘distinguish the dominating and the subsidiary’ implies that what
are the main factors that strongly influence the results of the analysis, and what are the subsidiary factors
which may have little effect on the results of the analysis, must be clearly known when establish the
computing model.
14-1 What is the difference between the primary system used in displacement method and that used in the
combined method by moment distribution and displacement method?
14-2 What is the stiffness coefficient of sidesway? What is the physical meaning of the coefficient?
14-3 Please use the results given in figure (a) to derive the results shown in figure (b).
Δ
Δ
Q Q
(b)
h
(a) EI
h
EI
Qh3 Qh3
Δ= Δ=
3EI 12 EI

(a) a cantilever column;
(b) a column with bottom end fixed and top end fixed freely
14-4 Why can a multistory and/or multibay rigid frame subjected to vertical loads be analyzed by
single-storey method? Why?
14-5 In what condition(s) inflection-point method can be used? In which case(s) the results obtained by
inflection-point method are exact solutions?
14-6 How many portions is a computing model composed? What are the principles to develop a
computing model?
14-1 Chose the proper method to analyze the structures shown in the figures and draw their bending
moment diagrams as well.
(a) (b) P
P C C E
D
h
hexagon
D E B EI = constant F
EI = constant
h
H
A B A G
P
l a a
2 2
(c) (d)
8kN
20kN 20kN
10m
C
m
15
4kN B D B C D E
15m
30m
5m 5m
A E A F
20m 20m 1kN/m
40m
problem 14-1
14-2 Draw the bending moment and shear force diagrams and determine the reaction at support C for the
structure shown in the figure.
14-3 Analyze the structure shown in the figure by using combined method of moment distribution and
displacement method, and draw the bending moment diagram as well. EI = constant
14-4 Draw the bending moment diagram for the structure shown in the figure by inflection-point method.
14-5 Analyze the rigid frame by shear distribution method; assume that the rigidity of the girder is
infinitely great.
q = 20kN/m
B C
A
2 EI 2 EI
6m
EI EI EI
D E F
6m 6m
problem 14-2
E
A B C D
l l l l
2 2
problem 14-3
50kN G H I
i =4 5
5m
1 1 1
A D
100kN 7 I =1 2 2 2 1
l
F B E
D E P
5m
1 2 2 2 1
2 2 2
l
C F
A B C
4m 6m problem 14-5
problem 14-4
14-6 If neglect the axial deformations due to the axial forces in the members of the structures shown in
the figures, try to compare the internal forces between following three sorts of computing models.
P P P
(a) (b) (c)
problem 14-6
14-7 Discuss the stressing characteristics of member AB shown in the figures when the ratio of the
i i
relative stiffnesses k = 2 → 0 or k = 2 → ∞ . If the computing model only for member AB is
i1 i1
desired, please draw it.
q
B
A B
i2 D i2
q q
h
h
i1 i1 i1 i1
A
l l
(a) (b)
problem 14-7
l
14-8 The middle portion structure shown in the figure is solid and the other portions of the structure
4
are truss-liked form. Please establish a proper computing model and analyzing method for the
structure.
8× a = l
problem 14-8
14-9 Try to discuss the distribution of the joint load P about the structures shown in the figure as to the
questions as follows. Assume that the cross sections of the members are identical and square, whose
area is h × h .
(a) How much load is respectively distributed to the cantilever beam AB and the simple beam CD?
(b) How much load is respectively distributed to the beam AB and the column CD?
(c) How much load is respectively distributed to the beam AB and the truss CDEF?
l l
P
l P 2 2
D
B A B
A C
l
2
C l
l
2
(a) (b) D
l l
P
2 2
A B
C E
45° D
(c) F
problem 14-9
14-10 The figure shows a rectangular board subjected to the uniformly downward distributed vertical load,
whose a pair of sides AB and CD are simply supported; the third side AD is fixed and the fourth side
BC is free as shown in the figure. Please discuss the distributing way of the load and the computing
model of the board according to the following two cases.
(a) when l1 l2 ; (b) when l1 l2 .
14-11 Try to discuss the stressing characteristics and computing model of the rectangular cistern due to the
pressure of the water for the following situations.
(a) shallow cistern ( h / a < 1/ 2 ); (b) deep cistern ( h / a > 2 ).
A B
l2
problem 14-10
D C
l1
a
a
a
a
h
(a) (b)
problem 14-11
Bibliography
1. 包世华主编，结构力学，上、下册，第二版，武汉理工大学出版社，2003
Bao Shihua (editor in chief), Structural Mechanics (volumes 1 and 2), Wuhan University of
Technology Press, 2003
2. 包世华主编，结构力学，中央广播电视大学出版社，1996
Bao Shihua (editor in chief), Structural Mechanics, Central Broadcast Television University Press,
1996
3. 龙驭球、包世华主编，结构力学教程（Ⅰ、Ⅱ）
，高等教育出版社， 2000，2001
Long Yuqiu, Bao Shihua (editors in chief), Structural Mechanics (volumes 1 and 2), High Education
Press, 2000, 2001
4. 龚耀清，胡必武，求超静定结构精确影响线的简捷方法，力学与实践 1990，12（1）
，26~30
Gong Yaoqing, Hu biwu, A simple and direct method to solve precise functions of influence lines for
statically indeterminate structures, Mechanics and Practice, 1990, 12(1), 26~30
5. Aslam Kassimali, Structural Analysis, PWS Publishing Company, 1995
6. Bingchen Zhi, Structural Mechanics, Tsinghua University Press, 1996
7. Edwin C. Rossow, Analysis and Behavior of Structures, Prentice Hall Inc. 1998
8. Harry, H. West, Louis F. Geschwindner, Fundamentals of Structural Analysis, Second Edition, John
Wiley & Sons, Inc. 2002
9. James K. Nelson, Jack C. McComac, Structural Analysis Using Classical and Matrix Method, Third
Edition, John Wiley & Sons, Inc. 2003
10. Russell C. Hibbeler, Structural Analysis, Third Edition, Prentice Hall, Inc. 1997
书后语
本书是作者应武汉理工大学出版社之邀，为高等学校土木工程专业（即大‘土木’
）编写的英文
结构力学教材。
国家教育部要求高等学校在部分课程中施行‘双语’教学，清华大学土木工程系从上世纪 80 年
代中期开始就在结构力学课程中实行了‘双语’教学。当时是按以下两种方式进行的：一种是凡书
面作业（包括课堂板书及同学作业等）全用英文，课堂讲解以汉语为主；另一种则是连课堂讲授也
以英文为主，这样进行了很长时间，因教师的不同，分别采用不同的教学方式。近年则将修结构力
学的学生分为中文班、英文班，分开授课。中文班用中文教学，用自己编写的统一教材。英文班用
英文讲授，曾用过自编的英文讲义；现由任课教师选用相应的英文材料。此做法已正式订入教学计
划要求。
本书第一作者自 80 年代中期开始，一直在清华大学土木系和水利系以上述第一种方式进行结构
力学‘双语’教学。本书第二作者近年在宁夏大学和河南理工大学也多次用这种方式讲授过结构力
学。本书的第一稿曾在河南理工大学经过了试用。本书是在积累了上述‘双语’教学经验的基础上
完成的。国内正缺合适的结构力学英文教材，所以我们乐意奉献此书给广大读者。
我国土木工程专业有自己的教学计划，结构力学课程有自己的教学大纲和教学基本要求，多年
来也形成了自己的教学体系。这些与美、英等国家的差别是比较大的。原版英文教材，因无统一的
教学大纲与课程要求，各版本的体系与内容的差别极大。直接采用原版英文教材并不太适合我国高
校的教学状况。此外，无论是用中文教学，还是用英文教学，教学计划、教学大纲应当是一样的，
结构力学的体系与内容也应当是一样的。清华大学结构力学课程多年来‘双语’教学实践也是遵循
这一原则的。基于这样的认识，本书（英文版）在编写时是以我国自己的结构力学教学大纲和课程
基本要求为依据，用英文编写。也就是说，它与国内现通用的土木工程专业结构力学（中文版）在
体系与内容上是相同的。这样，使本书不仅完全适合于国内结构力学课程的基本要求，也方便国内
教学使用。
另外，结构力学的内容较多，根据我国的教学大纲和教学基本要求，将内容分成为基本部分和
专题部分。基本部分为必修部分；专题部分为选修部分，且各校根据自己的情况常有特殊的安排。
据此，我们也将本书分册出版。基本部分（即本书）以 Structural Mechanics 为名，包含静定结构、
超静定结构和结构矩阵分析等内容。专题部分以 Advanced Topics of Structural Mechanics or Advanced
Structural Mechanics 为名，包含能量原理、动力、稳定和极限荷载等内容，另册出版。各校根据自
己的安排，可以全用，也可以分别选用。顺便说一下，多数美英原版英文结构力学是不包含动力、
稳定和极限荷载等内容的，它们是放在后续的高等结构力学中的。在这一点上，我们的分册与它们
是一致的。
本书是一本用英文编写的结构力学教材，内容包括了正文、思考题、习题及习题答案等，这些
均与国内中文教材是对应的。
‘双语’教学的方式、方法是多种多样的，本书为‘双语’教学提供了
各种英文的书面材料。各校在结构力学‘双语’教学的实践中如何使用本书，可由各校教师根据自
己的情况决定。
本书由包世华提出编写大纲和全部内容要求；龚耀清写出全书英文初稿，并设计了书中的图式
和版式，以及文字的录入工作；全书由包世华修改定稿。
本书插图的绘制和公式的录入工作由龚云和何书峰等完成，特致谢意。
由于英文不是作者的母语，难免有不恰当及错误之处，欢迎读者批评指正。
清华大学河南理工大学
包世华龚耀清
gongyq@263.net

Structural Mechanics PDF

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Structural Mechanics PDF

Uploaded by

Copyright:

Available Formats

Structural Mechanics

Brief Introduction of the Book

Organization and approach

Bao Shihua Tsinghua University

Gong Yaoqing (gongyq@263.net) Henan Polytechnic University

Brief Introduction of the Book

1.1 Structures and Their Classification 1

Chapter 2 Geometric Construction Analysis of Structures 19

Chapter 3 Statically Determinate Beams 38

3.1 The Analysis of Single Span Beams 38

Chapter 4 Plane Statically Determinate Rigid Frames 73

Chapter 5 Three Hinged Arches 101

5.1 The Constitution and Type of Three Hinged Arches 101

Chapter 6 Plane Statically Determinate Trusses and Composite Structures 120

6.1 Characteristics and Classification of Trusses 120

6.2 The Method of Joints 126

Chapter 7 General Remarks on Statically Determinate Structures 155

7.1 Analytical Methods For Statically Determinate Structures 155

Chapter 8 Influence Lines 169

8.1 Concept of influence lines 169

Chapter 9 Principle of Virtual Work and Displacement of Structures 220

9.1 Introduction for Calculation of Structural Displacement 220

9.4 Calculation of Displacements Caused by Loads 236

Chapter 10 Force Method 280

10.1 Statically Indeterminate Structures and Degrees of Indeterminacy 280

Chapter 11 Displacement Method 370

11.1 Fundamental Concepts of Displacement Method 370

11.8 Development of Displacement-Method Equations by Direct Stiffness Method 416

Chapter 12 Method of Successive Approximations

12.1 General Remarks for Method of Successive Approximations 431

Chapter 13 Matrix Displacement Method 495

13.1 Introduction 495

Chapter 14 General Remarks on Statically Indeterminate Structures 588

14.1 Classification and Comparison of Fundamental Method used to Analyze Statically

Answers to selected problems Ⅰ

The abstract of the chapter

1.1 Structures and Their Classification

Fig. 1.2 Bridge structure

Fig. 1.3 Hydraulic power station

Fig.1.1 Tall building suspended structure

one-storey workshop of a plant shown in Figure 1.18 are all structures.

(a) floded plate (b) floded plate roof

Fig.1.7 Floded plate structure

(3) Massive Structure

(a) shell (b) shell roof

Fig.1.9 Retaining wall Fig.1.10 Dam

1.2 Objective and Learning Method of Structural Mechanics

1.2.1 The relationship between structural mechanics and other curricula

The objective of Structural Mechanics comprises following aspects:

1.3 Analytical Models of Structures

1.3.1 The rules of developing analytical model

(1) The simplification of structural system

c Flexible, or hinged, joint

Fig.1.11 Roof structure and its computing model

Fig.1.12 Connection of reinforced concrete bars and its computing model

line diagram of the structure, as shown in Fig. 1.12 (b).

(c) the rocker support of a bridge

(a) Photo of the roller support of a bridge A A

RA (d) computing models of roller support

Fig.1.13 Roller support

(c) (d) (e)

Fig.1.14 Hinged support