MSR Report

“Comparative Study of Finite Element Analysis (FEA) And Isogeometric Analysis (IGA)”
Contents
LIST OF FIGURES ................................................................................................................. 4
LIST OF TABLES ................................................................................................................... 7
Abstract……………………………………………………………………………………….8
Chapter-1 Introduction ........................................................................................................ 9
1.1 Structure ................................................................................................................................. 9

1.1.1 Types of Structure ........................................................................................................... 9
1.2 Structural Analysis................................................................................................................. 11
1.2.1 Conditions of Structural Analysis .................................................................................. 12
1.2.2 Methods of Structural Analysis ..................................................................................... 13
1.3 Stiffness Method ................................................................................................................... 13
1.4 Introduction of Finite Element Method (FEM) ..................................................................... 16
1.4.1 Definition in Finite Element Analysis Field.................................................................... 17
1.4.2 Steps Involved In FEM ................................................................................................... 18
1.4.3 Advantages of FEM ....................................................................................................... 19
1.4.4 Disadvantages of FEM ................................................................................................... 20
1.5 What is Isogeometric Analysis? ............................................................................................ 20
Chapter-2 Literature Review ............................................................................................ 22
2.1 General.................................................................................................................................. 22
2.2 Literature Survey................................................................................................................... 23
2.3 Summary of Literature .......................................................................................................... 26
2.4 Scope of Work ....................................................................................................................... 26
2.6 Objectives of Study ............................................................................................................... 26
Chapter-3 B-Spline & NURBS .......................................................................................... 27
3.1 General.................................................................................................................................. 27
3.2 B-Spline ................................................................................................................................. 27
3.2.1 Parametric Domain of B-Splines ................................................................................... 28
3.2.2 B-Spline Basis Function ................................................................................................. 28
3.2.3 B-Spline Derivatives ...................................................................................................... 30
3.2.4 B-Spline curves .............................................................................................................. 31
3.2.5 B-Spline Surfaces and Solids ......................................................................................... 32
3.3 Refinement............................................................................................................................ 32
P a g e 1 | 92
3.3.1 Knot insertion: h-refinement ........................................................................................ 32

3.3.2 Degree Elevation: p-refinement ................................................................................... 33
3.3.3 Continuity and degree elevation: k-refinement ........................................................... 36
3.4 Non-Uniform Rational B-Splines (NURBS) ............................................................................ 38
3.4.1 NURBS basis functions .................................................................................................. 38
3.4.2 NURBS derivatives......................................................................................................... 38
3.4.3 NURBS Curves ............................................................................................................... 39
3.4.4 NURBS surfaces and solids ............................................................................................ 40
Chapter-4 Numerical Study............................................................................................... 42
4.1 Validation Problem & Result ................................................................................................. 42

4.2 Problem Formulation ............................................................................................................ 49
4.2.1 Data for Hook Models ................................................................................................... 49
4.2.2 Data for Beam Curved in Plan ....................................................................................... 50
Chapter-5 Finite Element Analysis ................................................................................... 51
5.1 Introduction .......................................................................................................................... 51

5.2 Finite Element Analysis of Hook............................................................................................ 51
5.2.1 Result of Hook Model-1 ................................................................................................ 51
5.2.5 Comparison of Hook Results ......................................................................................... 59
5.3 Finite Element Analysis of Beam Curved in Plan................................................................... 60
5.3.1 Result of Curved Beam Model-1 ................................................................................... 60
5.3.2. Result of Curved Beam Model-2 ................................................................................... 62
5.3.3 Comparison of Curved Beam Result ............................................................................. 64
Chapter-6 Iso-Geometric Analysis .................................................................................... 65
6.1 Steps of IGA Using NURBS in ABAQUS 6.14 .......................................................................... 65

6.2 Isogeometric Analysis of Hook Models ................................................................................. 75
6.2.5 Comparison of Hook Results ......................................................................................... 83
6.3 Comparison of FEA and IGA for Hook Models ...................................................................... 83
6.4 Iso-Geometric Analysis of Curved Beam ............................................................................... 84
P a g e 2 | 92

6.4.3 Comparison of Curved Beam Results: ........................................................................... 88
6.5 Comparison of FEA and IGA for Curved Beam Models ......................................................... 88
Chapter-7 Summary of Findings ...................................................................................... 89
References…………………………………………………………………………………91
P a g e 3 | 92
LIST OF FIGURES
Fig.-1. 1 Truss ....................................................................................................................................... 10
Fig.-1. 2 Cable Supported Bridge & Arch ............................................................................................ 10
Fig.-1. 3 Frame Structure ...................................................................................................................... 11
Fig.-1. 4 Element................................................................................................................................... 17
Fig.-1. 5 Nodes...................................................................................................................................... 17
Fig.-3. 1 Basis Function and their type…………………………………………………………………………………………….29
Fig.-3. 2 Quadratic basis functions for the non-uniform knot vector Ξ = {0,0,0,1,2,3,4,5,5,5}
……………………………………………………………………………………………………………………………………………30
Fig.-3. 3 The creation of a curve ........................................................................................................... 31

Fig.-3. 4 Knot Intersection: h-refinement ............................................................................................. 34
Fig.-3. 5 Degree Elevation: p-refinement ............................................................................................. 35
Fig.-3. 6 k-refinement versus p-refinement .......................................................................................... 37
Fig.-3. 7 Example of the construction of a NURBS quarter circle ....................................................... 40
Fig.-3. 8 Construction of a circle using NURBS .................................................................................. 41
Fig.-4. 1 Cantilever Beam detail given in book .................................................................................... 42
Fig.-4. 2 Ansys Result of cantilever beam ............................................................................................ 44
Fig.-4. 3 STAAD Pro. Result of cantilever beam ................................................................................. 45
Fig.-4. 4 Equivalent Stresses in ABAQUS 6.14 ................................................................................... 47
Fig.-4. 5 Equivalent Stresses in ANSYS 15.0....................................................................................... 47
Fig.-4. 6 Max. Principal Stresses in ABAQUS 6.14 ............................................................................. 47
Fig.-4. 7 Maximum Principal Stresses in ANSYS 15.0 ....................................................................... 48
Fig.-4. 8 Min. Principal Stresses in ABAQUS 6.14 ............................................................................. 48
Fig.-4. 9 Minimum Principal Stresses in ANSYS 15.0 ........................................................................ 48
Fig.-5. 1 Equivalent Stress For Hook Model- 1 .................................................................................... 51
Fig.-5. 2 Maximum Principal Stress For Hook Model-1 ...................................................................... 52
Fig.-5. 3 Minimum Principal Stress For Hook Model-1 ....................................................................... 52
Fig.-5. 4 Total Deformation For Hook Model-1 ................................................................................... 53
Fig.-5. 5 Equivalent Stress For Hook Model-2 ..................................................................................... 53
Fig.-5. 6 Maximum Principal Stress For Hook Model-2 ...................................................................... 54
Fig.-5. 7 Minimum Principal Stress For Hook Model-2 ....................................................................... 54
Fig.-5. 8 Total Deformation For Hook Model-2 ................................................................................... 55
P a g e 4 | 92
Fig.-5. 9 Equivalent Stress For Hook Model-3 ..................................................................................... 55

Fig.-5. 10 Maximum Principal Stress For Hook Model-3 .................................................................... 56
Fig.-5. 11 Minimum Principal Stress For Hook Model-3 ..................................................................... 56
Fig.-5. 12 Total Deformation For Hook Model-3 ................................................................................. 57
Fig.-5. 13 Equivalent Stress For Hook Model-4 ................................................................................... 57
Fig.-5. 17 Equivalent Stress for Curved Beam Model-1 ....................................................................... 60
Fig.-5. 18 Maximum Principal Stress for Curved Beam Model-1 ........................................................ 60
Fig.-5. 19 Minimum Principal Stress for Curved Beam Model-1......................................................... 61
Fig.-5. 20 Total Deformation for Curved Beam Model-1 ..................................................................... 61
Fig.-6. 1 Standard/Explicit Model......................................................................................................... 65
Fig.-6. 2 Set working Directory ............................................................................................................ 66
Fig.-6. 3 Import Geometry .................................................................................................................... 66
Fig.-6. 4 Give Material Property ........................................................................................................... 67
Fig.-6. 5 Create Assembly .................................................................................................................... 67
Fig.-6. 6 Create Step ............................................................................................................................. 68
Fig.-6. 7 Give Loading condition .......................................................................................................... 68
Fig.-6. 8 Give Boundary Condition ...................................................................................................... 69
Fig.-6. 9 Create mesh seeds .................................................................................................................. 69
Fig.-6. 10 Set the Element type ............................................................................................................. 70
Fig.-6. 11 Create mesh part ................................................................................................................... 70
Fig.-6. 12 Create Job for analysis ......................................................................................................... 71
Fig.-6. 13 Submit the Job ...................................................................................................................... 71
Fig.-6. 14 FEM Results ......................................................................................................................... 72
Fig.-6. 15 NURBS Plugins.................................................................................................................... 72
Fig.-6. 16 IGA Analysis Process........................................................................................................... 73
Fig.-6. 17 NURBS Post processing Process ......................................................................................... 73
Fig.-6. 18 IGA Process.......................................................................................................................... 74
P a g e 5 | 92
Fig.-6. 19 Iso-Geometric Analysis Results ........................................................................................... 74

P a g e 6 | 92
LIST OF TABLES
Table 1 Comparison of deflection for Validation ................................................................................. 45
Table 2 Result of ABAQUS and ANSYS ............................................................................................. 46
Table 3 Model Information of Hook ..................................................................................................... 49
Table 4 Model Information of Beam Curved in Plan ........................................................................... 50
Table 5 FEA Result of Hook ................................................................................................................ 59
Table 6 FEA Result of Curved Beam ................................................................................................... 64
Table 7 IGA Result of Hooks ............................................................................................................... 83
Table 8 Comparative percentage difference between FEA and IGA For Hook Models ...................... 83
Table 9 IGA Result of Curved beam .................................................................................................... 88
Table 10 Comparison percentage difference between FEA and IGA for Curved Beam Models ......... 88
P a g e 7 | 92
Abstract
In the real world of designing field the different geometries are generated using the Computer
Aided Design (CAD). And then after this geometries are analysed using different analytical
methods like Finite Element Method (FEM), Finite Difference Method and other traditional
methods. Among them Finite Element Method is more popular in recent time among the
designers. But, FEM consumes more time in preparing the boundary of the element. Hence, it
consumes more time to get final results. In last decade, a new emerging technology named as
Isogeometric Analysis has been developed by Hughes in 2005. As compared to FEM, the
Isogeometric Analysis prepares the meshing 80% faster than FEM. So, from the literature, this
study is concentrated on the comparison of Isogeometric Analysis with FEM.
In this research work, the main aim is to compare Isogeometric Analysis with Finite Element
Analysis using Isogeometric tool. Currently, four models of Crane Hook and Beam Curved in
Plan have been analysed and results of Finite Element Analysis and Isogeometric Analysis
using ABAQUS 6.14 software have been compared.
P a g e 8 | 92
Chapter-1 Introduction
1.1 Structure
Structure is an assembled system of the interconnected members in a stable configuration to
resist different types of loads with maximum factor of safety.
The loads acts upon the structure can be divided in following types:
 Dead Load
 Live Load
 Wind Load
 Earthquake/Seismic Load
 Snow load etc.
The structural members are connected by different types of support as
 Pin/Hinge Support
 Roller Support
 Fix Support
The Structure can be classified according to their uses as
 Civil Structures – Buildings, Bridges, Towers etc.

 Military Structures – Ships, Aircraft Structures, Tanks, etc.
1.1.1 Types of Structure

As an engineer we must be able to classify a structure according to their form and functions
and must also be able to recognize various types of elements composing a structure. Each
structural system may be composed from the four basic types of structure as
a) Trusses
b) Cable & Arches
c) Frames
d) Surface Structures
P a g e 9 | 92
(a). Trusses:
Truss is an assemblage of slender members arranged in triangular/rectangular pattern.

All members are connected together by pins/hinges which are free to rotate. Truss members
are usually subjected to axial forces only like axial tensile force and axial compression force.
Trusses are more suitable when the span of the structure is large and there is no restriction on
depth in the design criteria.
Fig.-1. 1 Truss
Truss structure is classified in two types:
Plane Truss (2D): All the members are lies in the same plane. They are used to support bridges
and roofs in industrial building.
Space Truss (3D): All the members are extending in three dimensions. They are used to
connect the member of transmission line towers and microwave towers.
(b). Cables & Arches:

Cables are usually flexible and carries a tensile load. The external load is usually acting
in vertical direction (not along the axis of the cable) as a result the cable deforms with a sag
see Fig. 1.2(a). Cables are mostly used to support bridge decks and building roofs. E.g. Bandra-
Worli Sea Link, New Yamuna Bridge, etc.
Fig.-1. 2 Cable Supported Bridge & Arch

An arch has the reverse curvature of a cable and it carries compressive load. The arch
must be rigid in order to maintain its shape. Arch is primarily subjected to compression but
P a g e 10 | 92
also some shear and moment. Arches are used in bridge structures, dome roofs and opening in
masonry walls.
(c). Frames:
Frames are commonly used in building structure. Frame is an assembly of beam and
column connected together at junction. Frame is divided in two groups as Planner (2D) Frame
in which all members are lies in one plane and Space (3D) Frame in which all members are lies
in more than one plane.
Fig.-1. 3 Frame Structure

There are different types of frame structure depending upon the connection between beam and
columns as under:
Braced Frame: All Connection between beams and columns are pinned joints and there are
diagonal members.
Moment Resisting Frame: All connection between beams and columns are rigid and there is
no need of diagonal bracings.
(d). Surface Structures:

Membrane, plates and shell types structure with much less thickness as compared to its
other dimensions are fall in this category. The structure is subjected to in-plane tension or
compression forces. Surface structures may be made of rigid material like R.C.C. The
structures may be shaped as folded plates, shell, cylinders, etc.
1.2 Structural Analysis

The structural analysis is a mathematical process by which the response of a structure to
specified loads and actions is determined. This response can be measured by determining the
internal forces or stress resultants and displacements or deformations throughout the structure.
P a g e 11 | 92
The structural analysis is an application of solid mechanics to predict the response in terms of
forces and displacement in a given structure (existing or proposed) subjected to specified loads
or load combinations.
1.2.1 Conditions of Structural Analysis

It is not important which structure is considered and which method is applied for analysis, the
solution must satisfy following three conditions:
A. Equilibrium of Forces
B. Compatibility of Displacements
C. Force displacement relations as specified by elastic properties
(A). Equilibrium of Forces:
The Solution should be satisfy equilibrium condition of forces (∑H=0, ∑V=0, ∑M=0)
for any part of the structure (members, joints). So, the reaction components developed at
various supports and the external load acting on the structure should satisfy equilibrium
condition forth structure as a whole. Every member must be in equilibrium under the effect of
load acting on it and the end forces. At every joint, forces extended by members on the joint
and the external load acting on the joints must satisfy the equilibrium condition of the joint.
Similarly, internal forces developed throughout the structure should be such as to keep any part
of the structure in equilibrium.
(B). Compatibility of Displacement:
Compatibility means geometrical relationship between joint displacement and
deformation of the members. Also, it refers to requirements of continuity of slopes and
displacement throughout the structure. The compatibility condition ensure that the slopes and
the displacement at supports are consistent with the type of support.
So, in a plane truss, the member deformation (elongation of the member) should be
calculated from the displacement of the joints at the end of the member. The twist of the grid
member must be consistent with the rotation of the joints at the end of the member. In a
Continuous beam, the slopes at the sections to the left and right of the support must be identical.
A clamped support of a plane frame does not allow rotation and displacements. Hinge supports,
allows rotation but does not allow displacement.
(C). Force displacement relations as specified by elastic properties:
The member end forces are related to the deformation of the member and elastic
property of the material. The deformation of the member can be represents in terms of the
member end displacements. So, the member end forces, which are internal resisting forces
P a g e 12 | 92
developed in the end sections are related to the member end displacement throughout elastic
constants like modulus of elasticity, modulus of rigidity and geometrical properties of the
member such as length, cross sectional area, moment of inertia, etc. These relations are linear
when small displacements and linear material behaviour are considered.
1.2.2 Methods of Structural Analysis

Various methods are available for analysis of the structure. These methods are classified as
under:
(A). Traditional Methods:
 Slope Deflection Method
 Moment Distribution Method
 Kani’s Method
 Column Analogy Method
 Energy Principles
(B). Matrix Methods:
 Flexibility Method
 Stiffness Method
(C). Numerical Methods:
 Finite Difference Method (FDM)
 Finite Element Method (FEM)
Traditional methods are used to analyse a simple building members like beam, truss, frame
etc. and Matrix methods are mostly used for analysis of complex structure. Among that
methods Stiffness Method is mostly used in current time.
1.3 Stiffness Method

The given indeterminate structure is first made kinematically determinate by
introducing constraints at the nodes. The required number of constraints is equal to degrees of
freedom at the nodes that is kinematic indeterminacy (K.I.). The kinematically determinate
structure comprises of fixed ended members hence all nodal displacements are zero. These
results in stress resultant discontinuities at these nodes under the action of applied loads or in
other words the clamped joints are not in equilibrium. In order to restore the equilibrium of
stress resultants at the nodes the nodes are imparted suitable unknown displacements. The
P a g e 13 | 92
number of simultaneous equations representing joint equilibrium of forces is equal to kinematic

indeterminacy (K.I.). Solution of these equations gives unknown nodal displacements. Using
stiffness properties of members the member end forces are computed and hence the internal
forces throughout the structure.
Since nodal displacements are unknowns, the method is also known as displacement
method. Since equilibrium conditions are applied at the joints the method is also known as
equilibrium method. Since stiffness properties of members are used in these method is also
known as stiffness method.
Stiffness method is classified in two ways as (i). System Approach (ii). Member
Approach. But, now a day’s member approach is well popular method among the analysts. So,
we are discusses here general procedure/steps of stiffness member approach or direct stiffness
method.
 Direct Stiffness Method:

The direct stiffness method is a very computer friendly and most powerful analytical tool
for complex structure. To simplify the assembly process for the joint stiffness matrix [Sj], the
key is to use member stiffness matrices for action and displacement at both ends of each
member. If the displacement are referenced to structural (Global) co-ordinates, they coincide
with joint displacements. In that case all the geometric complications must be handled locally,
and the transfer of member information to structural array is straight forward. That is the
stiffness matrix and the equivalent load vector can be assembled by direct addition instead of
by matrix multiplication.
General procedure/step of stiffness method is as under:
Step-1: Generate Joint Stiffness Matrix (Sj):
The assembly of joint stiffness matrix assuming ‘m’ members may be stated as under
𝑚
𝑆𝑗 = ∑ 𝑆𝑚𝑠 𝑖
𝑖=1
th
Where, 𝑆𝑚𝑠 𝑖 = the i member stiffness matrix with end actions and displacements (for both
ends) taken in the directions of structural axes. For matrix addition, all such member stiffness
matrices should be expanded to the standard size as [Sj] by augmenting them with rows and
columns of zero’s. However, this operation can be avoided in computer programming.
Step-2: Generate Joint Load Vector (Aj):
P a g e 14 | 92
The load applied at the joints may be listed in a vector [AJ], which contains the applied
loads corresponding to all possible joint displacements including those at support restraints.
The elements in [AJ] are numbered in the same sequence as joint displacements.
Step-3: Generate Equivalent Joint Load Matrix [AE]:
Equivalent load vector represents the actions produced at each end of a member
corresponding to displacements due to externally applied member loads.
For each members, equivalent load vector [AML] is prepared, and the summation of
[AML] for different members gives equivalent joint load matrix [AE].
So, [AE] = [AML1] + [AML2] +…..
The direction of [AE] matrix is changed as reactions are opposite to the applied actions.
Step-4: Generate Combined Joint Load Matrix [AC]:
The algebraic sum of [AJ] and [AE] gives the matrix [AC].
So, [AC] = [AJ] + [AE]
The elements of [AC] are rearranged. The elements corresponding to restrained
displacement forms [AFC] which is written at top, while the element corresponding to restrained
displacement forms [ARC] which is written below the [AFC] matrix.
Step-5: Generate Rearranged [SJ] matrix:
[SJ] matrix is obtained by assembling member stiffness matrix [SM] for all members.
The rearranged [SJ] matrix has the form as under,
𝑆𝐹𝐹 𝑆𝐹𝑅
[𝑆𝐽 ] = [ ]
𝑆𝑅𝐹 𝑆𝑅𝑅
Where, [SFF] = Joint Stiffness Matrix corresponding to free displacements
[SFR]= Joint Stiffness Matrix corresponding to free displacements and Joint
actions
[SRF]= Joint Stiffness Matrix corresponding to joint action and free
displacements
[SRR]= Joint Stiffness Matrix corresponding to joint actions
Step-6: Determination of deformations [DF]:
Deformation can be calculated as under,
[DF] = [SFF]-1 [AFC]
Where, [AFC] = [AD] – [ADL] = loads corresponding to free displacements
Step-7: Determine Reaction Matrix [AR]:
Reactions are calculated as under,
[AR] = - [ARC] + [SRF] [DF]
P a g e 15 | 92
Where, [AR] = Member End Reaction

[ARC] = Load corresponding to restrained displacement
[SRF] = Matrix relating joint actions [R] to the free displacement [DF]
Step-8: Member End Action [AM]:
Member End Actions are calculated as under,
[AM] = [AML] + [SM] [DM]
Where, [AM] = Member End Action
[AML] = Member End Action Due to loads
[SM] = Member Stiffness Matrix
[DM] = Deformation Matrix
1.4 Introduction of Finite Element Method (FEM)

In 1909 Ritz developed an effective method [5] for the approximate solution of
problems in the mechanics of deformable solids. It includes an approximation of energy
functional by the known functions with unknown coefficients. Minimisation of functional in
relation to each unknown leads to the system of equations from which the unknown coefficients
may be determined. One from the main restrictions in the Ritz method is that functions used
should satisfy to the boundary conditions of the problem.
In 1943 Courant considerably increased possibilities of the Ritz method by introduction
of the special linear functions defined over triangular regions and applied the method for the
solution of torsion problems [6]. As unknowns, the values of functions in the node points of
triangular regions were chosen. Thus, the main restriction of the Ritz functions – a satisfaction
to the boundary conditions was eliminated. The Ritz method together with the Courant
modification is similar with FEM proposed independently by Clough many years later
introducing for the first time in 1960 the term “finite element” in the paper “The finite element
method in plane stress analysis” [7]. The main reason of wide spreading of FEM in 1960 is the
possibility to use computers for the big volume of computations required by FEM. However,
Courant did not have such possibility in 1943.
An important contribution was brought into FEM development by the papers of Argyris
[8], Turner [9], Martin [9], Hrennikov [10] and many others. The first book on FEM, which
can be examined as textbook, was published in 1967 by Zienkiewicz and Cheung [11] and
called “The finite element method in structural and continuum mechanics”. This book presents
the broad interpretation of the method and its applicability to any general field problems.
P a g e 16 | 92
Although the method has been extensively used previously in the field of structural mechanics,
it has been successfully applied now for the solution of several other types of engineering
problems like heat conduction, fluid dynamics, electric and magnetic fields, and others.
1.4.1 Definition in Finite Element Analysis Field

Finite Element: The basic concept of FEM is the body of the structure may be divided into
smaller elements of finite dimensions called Finite Element.
Fig.-1. 4 Element
Nodes/Nodal Point: The Original shape of the structure is considered as an assemblage of
finite element connected at finite number of joints called Nodes.
Fig.-1. 5 Nodes
Nodal Degree of Freedom: The number of unknowns at a node is called Nodal Degree of
freedom.
Mesh: FEA uses a complex system of points called nodes which make a grid called a Mesh.
Finite Element Method: A numerical procedure for solving differential equations associated
with field problems with an accuracy acceptable to engineers is called Finite Element Method.
P a g e 17 | 92
Discretization: A body of the structure is divided in to equivalent system of smaller body or

units. This assembly of such unit then represents the original body. This process is known as
Discretization.
1.4.2 Steps Involved In FEM

Even though FEM procedure may vary depending upon problem and approach used for
solution, general steps in FEM remains more or less the same. The first step to the solution of
the considered problem is to define the problem in detail with the available data. Based on the
problem statement and available data, a mathematical model is prepared defining the geometry
of the problem, material properties, governing equations, initial and boundary conditions. The
general steps used in the solution using the FEM are discussed as below:
Step-1: Discretization and select element:
1D, 2D, 3D and multimode elements may be selected for various problems of structural
engineering like beam, plate bending, plane truss, plane stress/strain, 3D problems,
axisymmetric model, etc.
For discretization, divide the body into equivalent system of finite elements with nodes
and appropriate element type. The nodel degree of freedom is one of the important criteria to
start the problem.
Step-2: Select Displacement Function:
The distribution of unknown quantity (u) within each element can be chosen from
appropriate function within nodel values (u1, u2) of the element, linear, quadratic and cubic
polynomials are frequently used function as they are simple to operate with however
trigonometric series could also be used for N node element. The appropriate or interpolating
function can be written as
u(x) = N1U1 + N2U2 + …. + NnUn
Where, U1, U2, U3,… = Value of unknown at nodel point 1,2,3,… respectively.
N1, N2, N3,… = Shape Functions
So, 𝑢 = [𝑁]{𝛿}
The accuracy of the solution depends greatly on selection of these displacement
function.
Step-3: Define Gradient of unknown quantity and constitutive relationship (Stress-Strain
Relation & Displacement-Strain Relation):
These relations are basic requirement of each finite element problem. In stress analysis
problem, gradient of unknown gives strains and constitutive relationship is the relationship
P a g e 18 | 92
between stress and strain. We know the first derivative of displacement along X-direction gives
strain in particular direction.
𝑑𝑢 𝑑 1
𝜀𝑥 = = [[𝑁]{𝛿}] = [−1 1]{𝛿} = [𝐵]{𝛿}
𝑑𝑥 𝑑𝑥 𝐿
𝑑[𝑁] 1
Where, [𝐵] = = [−1 1]= Strain Displacement Matrix.
𝑑𝑥 𝐿
Now, from Hook’s Law, 𝜎𝑥 = 𝐸 𝜀𝑥 = [𝐷][𝜀𝑥 ] = [𝐷][𝐵]{𝛿}.

Where, [D] matrix is called Constitutive Matrix which includes material constants namely,
modulus of elasticity (E) and Poisson’s ratio (μ).
Step-4: Derive Element Stiffness Matrix and Nodel Forces:
To derive element stiffness matrix for an elastic body, the principle of minimizing
potential energy can be used which is called Energy Approach. It is based on the idea of finding
equilibrium state of body associated with stationary values of a scalar quantity. (πp= potential
energy).
[K]{δ}= {F}
Where, [K] = Element Stiffness Matrix
{δ} = Displacement Matrix
{F} = Nodel Force Matrix
Step-5: Assemble the element equations and introducing the boundary conditions:
The element stiffness matrix is assemble to the unknown nodel values and write in
above equation form. Where, [K], {δ}, {F} values are assembled values.
Step-6: Solve for the unknown degree of freedom (Primary Unknowns):
The assembled equations are solved for the δ by using Gauss Elimination or iterative
methods. Primary unknowns in most of the cases are displacements at each nodes.
Step-7: Solve for the secondary unknowns:
In most problems, it is necessary to compute secondary unknowns from the primary
quantities. In the case of stress analysis problems secondary unknowns are stresses, strains,
moment and shear forces.
1.4.3 Advantages of FEM

Compared to all other numerical methods some of the advantages of FEM are listed below:
1. Modelling of complex geometries and irregular shapes are easier as varieties of finite
elements are available for discretization of domain.
P a g e 19 | 92
2. FEM comparatively considered fewer nodes to get good results.

3. Different types of material properties can be easily accommodated in modelling from
element to element or even within an element.
4. Problem with heterogeneity, anisotropy, non-linearity and time dependency can be easily
dealt with it.
5. Whenever complexities are faced, other methods make the assumptions (i.e. shape,
boundary condition, loading), while in FEM can be applied without any made assumptions,
the problem is treated as it is.
6. FEM makes piecewise approximation and analysis which considered continuity of each and
every node of element.
7. FEM can consider the sloping boundaries exactly.
8. FEM can be easily coupled with the Computer Aided Design – CAD programmes in various
streams of engineering.
9. In FEM, it is relatively easy to control the accuracy by refining the mesh or using higher
order element.
10. FEM is simple, compact, and result oriented and hence widely popular among engineering
community.
1.4.4 Disadvantages of FEM

With many advantages, FEM has also some of the disadvantages discussed as under:
1. In FEM, exact equations are formed but approximate solutions are obtained.
2. There are many types of problems where some methods of analysis may prove efficient than
the FEM (for all regular problem).
3. There are other trouble spot such as “Aspect Ratio” which may affect the final results.
4. Stress value may vary by 25% from fine mesh analysis to coarse mesh analysis.
1.5 What is Isogeometric Analysis?

Isogeometric analysis was introduced in 2005 by Hughes et al. to get exact engineering
geometry to Finite Element Analysis (FEA) and to mitigate the inconvenient process of
meshing altogether.
The key concept of IGA outlined by Hughes et al. in 2005 is, “To employ Non-Uniform
Rational B-Splines (NURBS) not only as a geometry discretization technology but also a
P a g e 20 | 92
discretization tool for analysis”. The IGA concept merge the two fields of CAD and FEA by
expanding the solution space using the same basis at that of the geometry description from
CAD. Since, its introduction, IGA has successfully applied to a wide variety of problems in
structural analysis, electromagnetics, turbulence, fluid structure interaction and higher order
partial differential equations.
There are several candidate technologies available to the IGA framework, of which
NURBS is most commonly used tool since it is standard method employed in CAD programs.
NURBS generalizes B-Splines and consequently inherit all of their favourable properties for
free from design. NURBS are commonly used in Computer Aided Design (CAD), Computer
Aided Manufacturing (CAM) and Computer Aided Engineering (CAE).
Isogeometric Analysis is based on NURBS, has refinement procedure related to h-

refinement, p-refinement which are respectively known as Knot Intersection and Degree
Elevation. Hence, Isogeometric Analysis is an emerging technology capable of:
 Directly interacting with the CAD systems
 Greatly simplifying the refinement processes
 Improving the solution accuracy
 Reducing the computational costs and time
P a g e 21 | 92
Chapter-2 Literature Review

2.1 General
In real word of designing the engineering problems includes analysis on products like
aircraft, automobiles, boats, wind turbines and components of these products. The geometry of
all these products are described using Computer Aided Design (CAD).
In the field of engineering a lot of time is waste on approximating this geometry for
analysis purposes. Most of the analysis methods are uses piecewise linear or piecewise
quadratic approximations of the boundary. Since such an approximation is not unique,
engineers waste time going back and forth between these two definitions of the geometry.
Traditionally, geometry has been represented differentially in the field of CAD and FEA. This
means that the CAD geometry, which is exact, must be converted to an Analysis Suitable
Geometry (ASG) for input into a FEA programme. In order to obtain an ASG, features like
insert holes and another difficult boundaries are often omitted to avoid for making analysis
easy and simple during analysis process. This process takes up to 80% analysis time and this
is also known as ‘Meshing”.
The difference between the fields of FEA and CAD on the subject of geometry
representation is remarkable. This has mainly to do with the fact that they are seen as separate
fields, which are interfaced using complicated and expensive mesh generation schemes. In
order to avoid this problem, it is preferable to use an integrated approach where the CAD
geometry is directly used in the FEA. Some attempts at this have been made in the recent years
but among that the Isogeometric Analysis Method is very good and gives user interfaced
results.
To overcome this critical issues in FEA, an idea has been developed by Hughes et al.
in 2005 that, In the analysis framework, it is employed the same function used to describe the
geometry of the computational domain i.e. typically use B-Splines and/or Non–Uniform
Rational B-Splines (NURBS).
P a g e 22 | 92
2.2 Literature Survey

This section is discussed on the literature available on the Isogeometric Analysis.
“Fundamental aspect of shape optimization in the context of Isogeometric Analysis

(IGA)” by Fubeder D., Simeon B. et al., Computer Methods in Applied Mechanics and
Engineering, Vol.-286(2015), pg.-313 to 331.
A mathematical model has been developed for shape optimization problem under static
equation constraints where the discretization is done by the B-splines and/or Non-Uniform
Rational B –splines (NURBS).
IGA has been used for solving the Partial Differential Equations (PDEs) and nodal approach
to change domain where control points takes the place of nodes.
The optimization problem has been solved by a gradient descent method where shape gradient
is defined in Isogeometric term.
There are two techniques used for solving the shape optimization problem:
(1) Optimize first and discretize later
(2) Discretize first and optimize later
It has been concluded that optimize first and discretize later is more suitable technique than
the other one since they obtained same results.
“GeoPDEs: A research tool for Isogeometric Analysis of Partial Differential Equations

(PDEs)”, by Falco C. et al., Advances in Engineering Software, Vol.-42(2011), pg. 1020-
1034.
GeoPDEs is a free software tool for the application of IGA. Its main objective is to provide a
common framework for the implementation of the IGA methods for discretization of PDEs
which are based on B-splines and NURBS. The main feature of IGA is to maintain the same
exact description of the computational domain geometry throughout the analysis process
including refinement. The design process using GeoPDEs is done in 3 steps as under:
(1) Define the parameterization: the geometry structure
(2) Domain Partition: Mesh Structure

P a g e 23 | 92
(3) The discrete space: The space structure
GeoPDEs is also able to solve more complex structure as under:
- GeoPDEs is able to solve the non-isoparametric elements where the solution space
does not coincide with the geometry space.
- GeoPDEs is also able to solve liner elasticity, non-homogeneous boundary conditions
and 3D Liner elasticity example problems.
“NURBS based IGA of Beams & Plates using Higher Order Shear Deformation Theory
(HSDT)” by Xinkang Li., Hindawi Publication, volume 2013.
IGA is based on the NURBS developed for static analysis of beam and plates using Third Order
Shear Deformation Theory (TSDT) with C1 continuity. NURBS based on IGA utilizes the
concept of isoparametric which refers to the use of same basis functions for both geometry and
unknown field of discretization. To apply TSDT to plates five degree of freedoms w, ф x, фy,
u0, v0 has been considered.
In this paper some examples are and their results are discussed as under:
(1). Beams under a Uniform Pressure: from the result it has been concluded that IGA based
on TSDT gets more accurate results than FEA based on FSDT. FEA needs more nodes to
represent the geometry than the control points needed in IGA.
(2). Beams at extreme thickness situation: from the result it has been concluded that NURBS
element modelled with lower order basis functions suffers from shear locking. As the order of
basis function increases the shear locking problem is reduced.
(3). Square plate under a uniform pressure: From the results it has been concluded the shear
locking phenomenon is more obvious in clamped plates than simply supported beam and this
problem is improved by using the h-refinement technique.
“From the finite element analysis to the Isogeometric analysis in an object oriented
computing environment”, by Daniel Rypl, Patzak Borek, Advances in Engineering
Software,vol.44 (2012), pg. 116-125.
In Recent time the Isogeometric analysis has been introduced as a viable alternative to the
standard, polynomial-based finite element analysis. Moreover, it has been shown that IGA is
P a g e 24 | 92
superior to the classical finite element method in many aspects. This paper presents how the
Isogeometric analysis can be integrated within an object oriented finite element environment.
The class hierarchy and corresponding methods are designed in such a way, that most of the
existing functionality of the finite element code is reused. The missing data and algorithms are
developed and implemented in such a way that the object oriented features, such as modularity,
extensibility, maintainability and robustness, are fully retained. The results shows that the
amount of the modified and/or added code is rather limited (mostly related to handling B-spline
or NURBS based shape functions). On the other hand, the less transparent physical meaning
of primary variables may complicate the debugging of the code. The functionality of the
implementation and the performance of the IGA code has been verified on a few numerical
examples. The results reveal that the IGA is a viable methodology which can be profitable for
the whole engineering community.
“Isogeometric analysis: An overview and computer implementation aspects”, by Nguyen

Vinh Phu et al., Mathematics and Computers in Simulation
Isogeometric analysis (IGA) represents a recently developed technology in computational
mechanics that offers the possibility of integrating methods for analysis and Computer Aided
Design (CAD) into a single, unified process. In this paper a self-contained Matlab
implementation applied to IGA applied analysis problems and the related computer
implementation aspects. Furthermore, implementation of the extended IGA which are applied
for both two-dimensional and three-dimensional bodies. The MATLAB code which
accompanies the present paper can be applied to one, two and three-dimensional problems for
linear elasticity, linear elastic fracture mechanics, structural mechanics (beams/ plates/ shells
including large displacements and rotations).
“ISOGAT: A 2D tutorial MATLAB code for Isogeometric Analysis (IGA)”, by Vuong A.

V. et al., Computer Aided Geometric Design, Vol-27(2011), pg. 644-655.
In this Paper it is enlightening that the main feature of IGA to use one common geometry
representing for creating CAD model, for meshing, for numerical solution, which makes a
seamless integration of all computational tool within a single design loop. This paper resents a
tutorial of 2D MATLAB code for solving diffusion type problems on single patch geometries.
This ISOGAT code is used for the solving elliptic diffusion type problems including Poisson’s
P a g e 25 | 92
equation on single patch geometries. ISOGAT code includes an example of L-shape, circular,
square type elements. Knot Refinement procedure in ISOGAT is h-refinement.
2.3 Summary of Literature

From the above literature survey the main summary of the literature is listed as under:
An Isogeometric analysis is a new emerging technology in analysis field which is faster in

meshing in curved surfaces. Hence, the time required to get solution is very less as compared
to other analytical methods. From the literature it has been noted that most of the work has
been done on two dimensional bodies like plates sections like circular, square, etc. Using the
Isogeometric Analysis tool like GeoPDEs, ISOGATE and IGAFEM the analysis procedure
become more speedy and easy.
2.4 Scope of Work

To study behaviour of structural curvilinear element with Isogeometric analysis using various
IGA tools using NURBS or B-Spline Functions.
2.6 Objectives of Study

The main objectives of the study are:
A. To study FEA for curvilinear element by mesh refinement technique to achieve optimal
results using ANSYS and ABAQUS.
B. To study Isogeometric Analysis (IGA) as an emerging method which may prove to be
efficient method of analysis for curvilinear elements (with NURBS function) using IGA
tool.
C. To validate results obtained from ANSYS & IGA tool with theoretical results.
D. To compare the results of curvilinear element obtained from ANSYS & IGA tool with
theoretical results.
P a g e 26 | 92
Chapter-3 B-Spline & NURBS

3.1 General
In the recent time Computer Aided Geometric Design (CAGD) is concern with the
generation of smooth curves and surfaces, which generally have to satisfy a large number of
constraints. While using CAGD process requires higher degree approximations because degree
‘p’ polynomials can satisfy p+1 constraints. Higher degree polynomials are inefficient to
process and become unstable and have the disadvantage that changes are made global, while
local control is needed. In addition, continuity should be maintained when local changes are
made.
These issues were overcome by the definition of a Spline in the mathematical sense. A
function constructed from polynomial elements pieced together with a certain level of
continuity between the elements. The required continuity is directly built into the basis which
made Basis Splines or B-Splines. The B-Spline basis allows an arbitrary choice of continuity
between the elements from C0 continuity to a maximum C p-1 continuity. Curvature continuity
is an important requirement in design process and hence Cubic Splines are most commonly
used in CAGD.
The B-Splines are convenient for free form modelling, but B-Spline has lack the ability to exact
represent some simple engineering shapes like circle and ellipsoids. That’s why today, the
standard technology in CAD is a generalization of B-Splines called Non Uniform Rational B-
Splines (NURBS). The NURBS are rational function of B-Splines and inherit all the properties
of B-Splines.
3.2 B-Spline
Knot Vector: A knot vector in one dimension is a non-decreasing set of coordinates in the
parameter space, written = {ξ1, ξ2, . . . , ξn+p+1}.
Where, ξi ∈ R is the ith knot,
i= the knot index, i = 1, 2, . . . , n + p + 1,
p = the polynomial order, and
n = the number of basis functions used to construct the B-spline curve.
P a g e 27 | 92
- Knot vectors may be uniform if the knots are equally spaced in the parameter space.
- Knot vector may be non-uniform if they are unequally spaced in the parameter space.
- A knot vector is said to be open if its first and last knot values appear p + 1 times.
3.2.1 Parametric Domain of B-Splines

B-splines are defined on a parameter space Ω΄. The B-Spline parameter space is local
to “patches” instead of elements, where the patch can be seen as a “macro-element”. The
parameter domain itself is defined by the knot vector(s) Ξ. The knot vector is defined as under,
Ξ = {(𝜉1 , … , 𝜉𝑝+1 = 𝑎), 𝜉𝑝+2 , … , 𝜉𝑛, (𝜉𝑛+1 , … , 𝜉𝑛+𝑝+1 = 𝑏)}
Where, ξi ∈ R is the ith knot,
i is the knot index, i = 1, 2, . . . , n + p + 1,
p is the polynomial order, and
n is the number of basis functions used to construct the B-spline curve.
Higher Dimensional parameter space are constructed using a tensor product of 1D knot
vectors. Hence the parametric domains are defined by the set [a, b]d ∈ Rd with d is the
dimension of the space. Using the knot vector we can construct B-Spline basis function
of order p+1 which are piecewise polynomials of degree p. Repeated knots are allowed,
hence 𝜉1 ≤ 𝜉 ≤ ⋯ ≤ 𝜉𝑛+𝑝+1 . A knot which is repeated k times in knot vector is said to
have a multiplicity k.
3.2.2 B-Spline Basis Function

The B-Spline basis function are defined recursively starting with piecewise constants (p = 0):
1 𝑖𝑓 𝜉𝑖 ≤ 𝜉 ≤ 𝜉𝑖+1
𝐵𝑖,𝑝 = { } (3.1)
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
For polynomial order p=1,2,3,…… they are defined by,
𝜉−𝜉𝑖 𝜉𝑖+𝑝+1 −𝜉
𝐵𝑖,0 (𝜉) = 𝜉 𝑩𝑖,𝑝−1 (𝜉) + 𝐵𝑖+1 , 𝑝−1 (𝜉) (3.2)
𝑖+𝑝 −𝜉𝑖 𝜉𝑖+𝑝+1 −𝜉𝑖+1
So given a knot vector and a polynomial degree the B-Spline function space B is uniquely
defined as
P a g e 28 | 92
By using the recursive algorithm. Higher-dimensional B-Spline function spaces are constructed
using tensor products of univariate B-Spline basis functions namely,
The result of (3.1) and (3.2) is shown in Fig.-3.1 for the knot vector Ξ = {0,0,0,0,1,2,3,4,4,4,4}
Fig.-3. 1 Basis Function and their type

Fig. 3.1 shows Recursive generation of a cubic basis for the uniform knot vector 𝛯 =
{0,0,0,0,1,2,3,4,4,4,4}. An example of a quadratic basis for an open, non-uniform knot vector
is shown in Fig.-3.2. Here the implications of the repeated knots at the ends of the interval and
also at 𝜉 = 4 are shown, where the continuity is lowered to C0. The other basis functions are
C1 continuous. Degree p basis functions have up to p−1 continuous derivatives. A repeated
knot will reduce the number of continuous derivatives by 1. When the multiplicity equals p,
the basis function is nodal. `
P a g e 29 | 92
Fig.-3. 2 Quadratic basis functions for the non-uniform knot vector 𝜩 =

{𝟎, 𝟎, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 𝟓, 𝟓, 𝟓}
The basis functions possess the following important properties:
a) Non-negativity: Bi,p (ξ) ≥ 0 for ∀i, p and a ≤ ξ ≤ b.

b) On a knot span [ξi , ξi+1 ) there are p+1 non-zero functions.
c) Partition of Unity. ∑ni=1 Bi,p (ξ) = 1.
d) The basis function from a linear independent basis which makes them suitable for
analysis.
e) B0,p (0) = Bn,p (1) = 1.
f) For compact support[ξi , ξi+p+1 ). Higher order functions have support across larger
portions of the domain. This increase in support has no implications on the bandwidth
of the resulting linear system in numerical applications. The total number of functions
that any function shares support with (including the function itself) is 2p+1 which is
equal to that for Lagrange polynomials.
3.2.3 B-Spline Derivatives

Derivatives of B-Spline basis functions are generated using the B-Spline lower order bases
When the denominator becomes zero due to repeated knots, the coefficient is defined to be
zero.
P a g e 30 | 92
3.2.4 B-Spline curves

B-Spline curves are defined by the coefficients of the basis functions, the control points Pi. The
curve is constructed in Rd by taking linear combinations of a set of n basis functions Bi,p where
i=1,2,…,n with their corresponding control points Pi ∈ Rd, i=1,2,…,n. The piecewise
polynomial B-Spline curve is given by,
(3.3)
Hence given a degree p, a knot vector Ξ and set of control points Pi the curve is defined. The
curve C (ξ) is a vector-valued function of one parameter. It maps a line segment into
C: Ω΄ Ω, this is show by graphically in Fig.-3.3.
Fig.-3. 3 The creation of a curve

B-Spline Curve has following properties:
1. A B-Spline curve of degree p has p-1 continuous derivatives in the absence of repeated
knots or control points. Another important property is that the compact support of the basis
gets passed on the curve, hence moving a single control point does not affect more than
p+1 elements of the curve.
2. Repeating a knot or control point k times, reduces the number of continuous derivatives
by k.
3. Non-negativity of the basis follows to the convex hull property, if ξ ∈ [ξi , ξi+1 ) then C (ξ)
lies within the convex hull of the control points Pi-p,…,Pi.
4. Variation decreases property with increasing degree. That means higher degree curve is
flatter and smoother than the lower degree curve.
P a g e 31 | 92
5. Affine invariance property. Affine transformation of a B-Spline curve are applied to the
control points directly.
3.2.5 B-Spline Surfaces and Solids

The B-Spline surface is defined by a control net Pij, i=1,2,…,n, j=1,2,…,m and the knot vectors
Ξ = {ξ1 , ξ2 , … , ξ𝑛+𝑝+1 }, Ή = {𝜂1 , 𝜂2 , … , 𝜂𝑚+𝑞+1 }. Taking the tensor product of the univariate
basis functions Bi,p (ξ) and Bj,q(𝜂) with the control net results in a B-Spline surface defined as
S: Ω΄ Ω with the map defined as,
(3.4)
It is important to note that the univariate function are defined on their own knot vector, hence
they can have a different parameterization. Furthermore a different degree can be chosen for
each co-ordinate direction.
Analogous to B-Spline surfaces, B-Spline solids are defined as the tensor product of three
univariate basis functions. Given a control net Pi,j,k, i=1,2,…,n, j=1,2,…,m, k=1,2,…,l and knot
vectors Ξ = {ξ1 , ξ2 , … , ξ𝑛+𝑝+1 }, Ή = {𝜂1 , 𝜂2 , … , 𝜂𝑚+𝑞+1 }, ɀ = {𝜁1 , 𝜁2 , … , 𝜁𝑙+𝑟+1 }, the B-
Spline solid is defined as V : Ω΄ Ω with the map defined as
(3.5)
3.3 Refinement
The B-Spline basis can be enriched by three types of refinement of which two have an analogue
in standard FEM bases. These are knot insertion, degree elevation and degree and continuity
elevation. The first two are equivalent to h- and p-refinement respectively, the last one is
dubbed k-refinement and has no equivalent in standard FEM. In this section these three
enrichments are discussed and examples are shown.
3.3.1 Knot insertion: h-refinement

Knot intersection or h-refinement in classical FEM improves the basis by increasing the
resolution of the parameter space. Given a knot vector,Ξ = {ξ1 , ξ2 , … , ξ𝑛+𝑝+1 }, and introducing
an extended knot vector Ξ′ = {ξ′1 , ξ′2 , … , ξ′𝑛+𝑝+1 }, such that Ξ ⊂ Ξ′. The new n + m basis
functions are formed by (3.1) and (3.2) by applying them to Ξ′. The new n + m control points,
P a g e 32 | 92
P′ = {P′1, P′2,…, P′n+m}T are formed from linear combinations of the original control points,
P={P1, P2,…, Pn}T by𝑃′ = 𝛼𝑖 𝑃𝑖 + (1 − 𝛼)𝑃𝑖−1 . (3.6)
where, (3.7)
Note that choosing the control points as in (3.6) and (3.7) the continuity of the curve is
preserved. Fig.-3.4 gives an example of knot intersection. The initial knot vector is Ξ =
{0,0,0,0,1,2,3,4,4,4,4}, a new knot is inserted at ξ′ =2.5. The curve with its new and old control
polygon is shown in Fig.-3.4(a). The new curve is geometrically identical to the original curve,
but the basis function and control points are changed. The new knot added one basis function,
compare the Fig.-3.4(b) with Fig.-3.4(c) and one control point.
3.3.2 Degree Elevation: p-refinement

Degree elevation is the second method of enriching the basis and is the equivalent of p-
refinement in regular FEM. This procedure starts by effectively subdividing the curve into
Bezier elements by knot insertion, (3.6) and (3.7) to raise the multiplicity to the polynomial
degree. Then the order of the polynomial is raised on each individual segment. Finally excess
knots are removed to create the new B-spline. We know that the basis has p − k continuous
derivatives, so increasing p also implies increasing k. This ensures that the basis keeps its
original continuity.
As with h-refinement the solution space spanned by the degree elevated basis functions
contains the space spanned by the original functions. So it is possible to degree elevate without
changing the geometry of the B-Spline curve. An example of degree elevation is given in Fig.-
3.5. Again the starting point is a cubic curve with the control net P and basis functions shown
in Fig.-3.5(a) and Fig.-3.5(b) respectively. After degree elevation the multiplicity of the knots
is increased by one and consequently the locations of the control points, P′ change. Though,
the elevated curve is still geometrically and parametrically the same. Fig.-3.5(c) shows the
enriched basis, there are now 11 quartic basis functions.
P a g e 33 | 92
Fig.-3. 4 Knot Intersection: h-refinement

Figure 3.4(a) shows an example of knot insertion performed on a cubic B-Spline curve defined
by the initial knot vector 𝛯 = {0, 0, 0, 0, 1, 2, 3, 4, 4, 4, 4} and control net P. The basis functions
defined by 𝛯 are shown in Figure 3.4(b). Inserting the knot 𝜉 ′ = 2.5 results in an additional
control point and basis function as shown in Figure 3.4(a) and Figure 3.4(c) respectively.
P a g e 34 | 92
Fig.-3. 5 Degree Elevation: p-refinement

Figure 3.5(a) shows a degree elevation of a cubic B-Spline curve defined by the initial knot
vector Ξ = {0, 0, 0, 0, 1, 2, 3, 4, 4, 4, 4} and control net P. The basis functions defined by Ξ are
shown in Figure 3.5(b). The degree elevation results in a quartic curve with the enriched knot
vector Ξ′ = {0, 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 4, 4, 4} and control net P′. The enriched basis is
shown in Figure 3.5(c).
P a g e 35 | 92
3.3.3 Continuity and degree elevation: k-refinement

A potentially more powerful type of refinement which is unique to the B-Spline basis is k
refinement. Basically k-refinement is a different order elevation strategy taking advantage of
the fact that knot insertion and degree elevation do not commute. Inserting a unique knot value
ξ′ between two distinct knots in a degree p curve lowers the number of continuous derivatives
at ξ′ to p-1. Elevating the degree to q, using the process of Section 3.3.2, increases the
multiplicity of each knot so that discontinuities in the pth derivative of the basis are preserved.
Hence the basis still has p-1 continuous derivatives at ξ′ . The above process can be turned
around by first elevating the curve degree to q and then inserting the unique knot ξ′ . Now the
basis has q-1 continuous derivatives at ξ′ . This process is called k-refinement and has, as said
before, no analogue in classical FEM.
By enriching the basis using k-refinement there is also the potential to save a significant amount
of degrees of freedom. Consider again the previous example but this time keeping track of the
number of basis functions, see also Figure 3.6. Let’s start with a single element2 of degree p,
this element has p+1 basis functions. Then perform knot insertion to arrive at n-p elements
with a total of n basis functions. Like before elevate the degree keeping the continuity p-1 by
increasing the multiplicity of each knot by one. This adds a basis function per element, hence
the total number of basis functions is now 2n-p. After r degree elevations the number of basis
functions has become (r + 1)n-rp. This is a large number because in practice the number of
elements surpasses the polynomial degree by multiple orders of magnitude.
Repeating this process using k-refinement starting again with one element and p+1 basis
functions. First elevate the degree r times adding one basis function each time and then insert
knots until the number of elements is n-p. This results in n+r basis functions all having r+p−1
continuous derivatives. Note that n + r<<(r + 1)n − rp especially when going to higher
dimensions, then these get raised to the dth power namely (n + r)d << ((r + 1)n − rp)d.
Furthermore k-refinement adds the benefit of smoother derivatives leading to more accurate
representation of physical quantities. Some examples are: the representation of the structure of
bones for a smooth transition from the solid outer part to the softer inner part, the representation
of structural vibration modes resulting in converging spectra, enabling the use of higher order
gradient damage models.
P a g e 36 | 92
Fig.-3. 6 k-refinement versus p-refinement

Fig.-3.6(a) Starting with one linear element. Fig.-3.6(b) Classic p-refinement approach, knot
insertion followed by degree elevation results in seven piecewise quadratic basis functions that
are C0 at the internal knots. Fig.-3.6(c) k-refinement approach, degree elevation followed by
knot insertion results in five piecewise quadratic basis functions that are C1 at the internal knots.
P a g e 37 | 92
3.4 Non-Uniform Rational B-Splines (NURBS)

B-splines have their rational counterparts giving the ability to exactly represent objects that
cannot be represented by polynomials. For example in CAD circular and conic shapes are
often used, which can be exactly represented by NURBS.
3.4.1 NURBS basis functions

The NURBS basis is defined by associating the B-spline basis functions with a strictly positive
weight, wi as
(3.8)
Where,
Spanning the NURBS functions space uniquely defined as Ɲ≡Ɲ(Ξ; p; w) := span {Ni,p}ni=1.
Analogous to B-Splines higher dimensional function spaces are constructed using tensor
products of univariate basis functions Ɲ≡Ɲ (Ξ, Ή, …; p, q, …; w) := span{ Ni,p ⨂ Nj,q ⨂ …
}𝑛,𝑚,…
𝑖,𝑗,…=1
.
The NURBS basis has the following properties:
1. The NURBS basis constitutes a partition of unity ∑𝑛𝑖=1 𝑁i,p(ξ) = 1 ∀ξ.

2. NURBS inherit their properties from the B-Spline basis functions like continuity across
knots, local support and non-negativity.
3. The NURBS basis functions are not polynomial but rational functions.
4. If the weights are equal the basis is again polynomial. Hence, B-Splines are a special
case of NURBS.
3.4.2 NURBS derivatives

NURBS derivatives are found by using the quotient rule on (2.8), expressing the derivative in
the B-Spline basis, namely
P a g e 38 | 92
And let
Then higher order derivatives are expressed in terms of lower order derivatives as
Where,
3.4.3 NURBS Curves

Using (2.8) the NURBS curve can be defined in the same way as the B-Spline curve namely
Fig.-3.7 shows the Example of the construction of a NURBS quarter circle based on the knot
vector Ξ = {0,0,0,1,1,1}. The dashed line indicates the unweight curve. Note that one weight
1
has the value to allow exact representation of the circle. The dashed line shows the curve
√2
when all weights are equal to one and hence this curve is polynomial. Comparing the two
curves it is clear that due to the weight the middle control point pulls the curve less strong.
P a g e 39 | 92
Fig.-3. 7 Example of the construction of a NURBS quarter circle
3.4.4 NURBS surfaces and solids

In an analogous way the NURBS surface and NURBS solid are defined as,
Fig.-3.8 shows the construction of a circular surface by a mapping from parameter space to
physical space using the control points and weights of the circular arc, Figure 2.10, for each
side of the parameter space forming a circular surface in physical space. The surface is
constructed using the knot vector Ξ= Ή= {0,0,0,1,1,1} and control points and weights as shown
in the figure.
P a g e 40 | 92
Fig.-3. 8 Construction of a circle using NURBS
P a g e 41 | 92
Chapter-4 Numerical Study

In this chapter, the validation problem, problem formulations, model and results are discussed.
4.1 Validation Problem & Result

PROBLEM-1: The Book: Finite Element Analysis – Theory and application with ANSYS,
Third edition, by Moaveni, Saeed Published by Pearson Education.
This is having a chapter on Finite element formulation of beams.

Example no. 4.3 shows deformation of beam at mid span and free end of beam.
A cantilever beam having span 10ft with uniformly distributed load 1000 lb/ft is analysed to
get the value of deflection using ANSYS 15 and the same problem is also analysed for
deflection manually. And value of deflection for both analysis is compared.
Fig.-4. 1 Cantilever Beam detail given in book

Beam: W18x35
A= 10.3 in2
L= 10 ft
I= 510 in4
W= 1000 lb/ft
E= 29x106lb/in2
According to the book deflection at free end and mid-point is given as

U21= -0.146 in
U31= -0.048 in
Deformation at free end is given below:
P a g e 42 | 92
 12 6 L  12 6 L 
 6 L 4 L2  6 L 2 L2 
K e   K G   EI3  
L  12  6 L 12  6 L 
 2 
 6L 2L  6L 4L 
2
  WL 
 2 
  WL2 
 
F e   F G    12
WL 

 
 2 
 WL2 
 12 
  WL 
 2 
 12 6 L  12 6 L  U11    WL2 
 2    
EI  6 L 4 L  6 L 2 L  U12   12 
2
  
L3  12  6 L 12  6 L  U 21    WL 
 2 
 6 L 2 L  6 L 4 L  U 22  
2 2 
 WL2 
 12 
Applying boundary condition U11=0, U12=0 at node 1.
 0 
 1 0 0 0  U 11   
 0  U   0 
EI  1 0 0   12     WL 
L  12  6 L 12  6 L  U 21   2 
3
 2  WL2 
 6 L 2 L  6 L 4 L  U 22  
2
 12 
Simplifying,
  WL 
 12  6 L  U 21  L3  2 
 6 L 4 L2  U   EI  WL2 
   22   
 12 
  100010
 12  610 ft  U 21  10 ft 3  2 
 610 ft  410 ft 2  U   2  2 
 1 ft   100010 
   22 
 
29x106 lb / in 2    
 12in   12
U21=-0.01217 ft = -0.146 in, U22= -0.00163 rad
Deflection at end point C is given as,

P a g e 43 | 92
2 3
3x 2 2 x 3 3  L  2 L 1
S 21   3  2   3  
L2
L L 2 L 2 2
2 3
L L
   
x 2 3
 2
   2  
x 2 L
S 22   2 
L L L L 8
1   120in 
VB    0.146in     0.00163rad   0.048in
2  8 
According to Ansys Deflection at free end and mid-point is

U21= -0.155 in
U31= -0.055 in
Result of Ansys At free end and Mid-point is:
Fig.-4. 2 Ansys Result of cantilever beam
P a g e 44 | 92
According to STAAD Pro. Deflection at free end is

U21= -0.155 in
Result of STAAD Pro. At free end is:
Fig.-4. 3 STAAD Pro. Result of cantilever beam
Result Comparison:
Table 1 Comparison of deflection for Validation
Total deflection according to Total deflection according to
Book 0.148 in STAAD Pro. 0.156 in
ANSYS 0.156 in ANSYS 0.156 in
Difference 5.12% Difference 0%
P a g e 45 | 92
Problem:2
A beam of rectangular cross section with 100 mm width and 250 mm depth is subjected to
uniform pressure of 10 N/mm2. The beam is curved to a radius of 400 mm along the centroidal
axis and bending moment increases the curvature. Find out the stresses in curved beam.
Solution:
To solve this problem in software Steel material is used with following property:
 Modulus of Elasticity, E= 2X105 N/mm2

 Poisson’s Ratio= 0.3
This Problem is solved using ABAQUS 6.14-1 and ANSYS 15.0 FEM tools. And the results
are as under:
Table 2 Result of ABAQUS and ANSYS
(A). Equivalent Stresses:

ABAQUS 6.14 87.11
ANSYS 15.0 90.539
Error (%) 3.78 %
(B). Maximum Principal Stress:

ABAQUS 6.14 110.1
ANSYS 15.0 114.68
Error (%) 3.993 %
(C). Minimum Principal Stress:

ABAQUS 6.14 87.57
ANSYS 15.0 90.641
Error (%) 3.388 %
(D). Directional Deformation (Y-AXIS):

ABAQUS 6.14 0.0084
ANSYS 15.0 0.0084923
Error (%) 1.086 %
P a g e 46 | 92
Fig.-4. 4 Equivalent Stresses in ABAQUS 6.14
Fig.-4. 5 Equivalent Stresses in ANSYS 15.0
Fig.-4. 6 Max. Principal Stresses in ABAQUS 6.14

P a g e 47 | 92
Fig.-4. 7 Maximum Principal Stresses in ANSYS 15.0
Fig.-4. 8 Min. Principal Stresses in ABAQUS 6.14
Fig.-4. 9 Minimum Principal Stresses in ANSYS 15.0

P a g e 48 | 92
4.2 Problem Formulation

Isogeometric Analysis gets effective in curvilinear element where meshing is complex
problem. We are currently take a two types of problems as under:
1. Hook (4 Models of Hook with different cross-sections and same area)

2. Beam Curved in Plan (2 Models – Quarter Circle And Semi Circular Models)
First of all we are doing Finite Element Analysis of these models and their results are compared
with Iso-geometric Analysis results.
4.2.1 Data for Hook Models

Table 3 Model Information of Hook
Model-1 Model-2 Model-3 Model-4

(Circular) (Rectangular) (Rectangular) (Trapezoidal)
Cross Section
Material Steel Steel Steel Steel
C/S Area 1256.63 mm2 1256.63 mm2 1256.63 mm2 1256.63 mm2
Radius 20 mm - - -
Side Width - 25×50.24 mm 50.24×25 mm 15 mm , 35 mm

50.24 mm height
Maximum 21500 N 26500 N 32250 N 41000 N
Principal Load
Nodes 293 372 480 375
Element 972 1248 1713 1195
Element Type 4 Noded Linear 4 Noded Linear 4 Noded Linear 4 Noded Linear
Tetrahedron Tetrahedron Tetrahedron Tetrahedron
P a g e 49 | 92
4.2.2 Data for Beam Curved in Plan

Table 4 Model Information of Beam Curved in Plan
Model-1 Model-2
Type Quarter Circlular Beam Semi Circular Beam
Cross Section
C/S Size 230 X 300 mm 230 X 300 mm

Centroidal Radius 4000 mm 4000 mm
Length 6283.185 mm 12566.370 mm
Applied Failure Pressure 0.901 N/mm2 0.192 N/mm2
Nodes 621 344
Elements 64 28
Element Types 20 Noded hexahedron 20 Noded hexahedron
Summary:
In this chapter, we defines a different models for Finite Element Analysis and Iso-
geometric Analysis. In the following chapters we come across the results of these models by
FEA and IGA respectively.
P a g e 50 | 92
Chapter-5 Finite Element Analysis
5.1 Introduction
In this chapter, the finite element analysis of Hook (4 Models) and Beam Curved in
Plan (2 Models) are carried out and their different results like Equivalent Stress (Von Mises),
Maximum Principal Stress, Minimum Principal Stress and Total Deformation are listed out
using the FEM software tool Simulia Abaqus 6.14 version.
5.2 Finite Element Analysis of Hook
Result of different models of Hook are shown in following figures:
5.2.1 Result of Hook Model-1
Fig.-5. 1 Equivalent Stress For Hook Model- 1
P a g e 51 | 92
Fig.-5. 2 Maximum Principal Stress For Hook Model-1
Fig.-5. 3 Minimum Principal Stress For Hook Model-1
P a g e 52 | 92
Fig.-5. 4 Total Deformation For Hook Model-1
Fig.-5. 5 Equivalent Stress For Hook Model-2
P a g e 53 | 92
P a g e 54 | 92
P a g e 55 | 92
P a g e 56 | 92
P a g e 57 | 92
P a g e 58 | 92
5.2.5 Comparison of Hook Results
Table 5 FEA Result of Hook
Equivalent
213.39 193.5 206.5 213.4
Stress (N/mm2)
Maximum
Principal Stress 246.7 245.3 246.7 246.4
(N/mm2)
Minimum
(N/mm2)
Total
Deformation 0.62 0.6938 0.3258 0.3779
(mm)
P a g e 59 | 92
5.3 Finite Element Analysis of Beam Curved in Plan

Result of different models of Curved Beam are shown in following figures:
5.3.1 Result of Curved Beam Model-1
Fig.-5. 17 Equivalent Stress for Curved Beam Model-1
Fig.-5. 18 Maximum Principal Stress for Curved Beam Model-1

P a g e 60 | 92
Fig.-5. 19 Minimum Principal Stress for Curved Beam Model-1
Fig.-5. 20 Total Deformation for Curved Beam Model-1

P a g e 61 | 92
5.3.2. Result of Curved Beam Model-2

P a g e 62 | 92

P a g e 63 | 92
5.3.3 Comparison of Curved Beam Result
Table 6 FEA Result of Curved Beam
Model-1 Model-2
Equivalent 213.1 216.1

Stress N/mm2 N/mm2
Maximum 243.8 248.00
Principal Stress N/mm2 N/mm2
Minimum 243.6 248.1
Principal Stress N/mm2 N/mm2
Total 10.53 45.86
Deformation mm mm
P a g e 64 | 92
Chapter-6 Iso-Geometric Analysis
6.1 Steps of IGA Using NURBS in ABAQUS 6.14
Non-Uniform Rational B-Spline (NURBS) plugins should be installed in Abaqus 6.14

for performing an Iso-Geometric Analysis. General Steps are described here to perform an IGA
as under:
Step-1: Open Abaqus 6.14 and start with Standard/Explicit Model as shown in Fig.-
Fig.-6. 1 Standard/Explicit Model
P a g e 65 | 92
Step-2: Set the Abaqus working Directory for saving all the files in one folder as
C:\AbaqusWorking Directory\REPORT STEP PROBLEM and click ok button.
Fig.-6. 2 Set working Directory

Step-3: Import the model geometry in “.igs” or “.stp” format as shown in fig.
Fig.-6. 3 Import Geometry
P a g e 66 | 92
Step-4: Give the material property by clicking on models--property and define Section as
shown in fig.
Fig.-6. 4 Give Material Property

Step-5: Create assembly and select the part and click on ok as shown in fig.
Fig.-6. 5 Create Assembly
P a g e 67 | 92
Step-6: Create step for analysis as you want and click on continue as shown in fig.
Fig.-6. 6 Create Step

Step-7: Give the loading condition and boundary conditions as shown in fig.
Fig.-6. 7 Give Loading condition
P a g e 68 | 92
Fig.-6. 8 Give Boundary Condition

Step-8: Create meshed part and set the element types as shown in following fig.
Fig.-6. 9 Create mesh seeds
P a g e 69 | 92
Fig.-6. 10 Set the Element type
Fig.-6. 11 Create mesh part
P a g e 70 | 92
Step-9: Create Job for analysis and submit the job from job manager and get the FEM results
as shown in fig.
Fig.-6. 12 Create Job for analysis
Fig.-6. 13 Submit the Job

P a g e 71 | 92
After job status has shown as completed click on result and get the analysis result as shown in
fig.
Fig.-6. 14 FEM Results

Step-10: After FEM results are getting, The Iso-geometric analysis step has been started using
NURBS Plugins. For this process click on Plugins – NURBS – Analysis
Fig.-6. 15 NURBS Plugins
P a g e 72 | 92
Enter the path for input file (.inp) and mathematical notebook file (.NB) file from your set
abaqus working directory and click on submit and generate “.fil” file as shown in following
fig.
Fig.-6. 16 IGA Analysis Process

Step-11: Do NURBS Post Processing using above step generated .fil file and .NB file and
submit the process and generate Output Data Base File (.odb) as shown in following fig.
Fig.-6. 17 NURBS Post processing Process
P a g e 73 | 92
Enter the path for “.fil”file and mathematical notebook file (.NB) file from your set abaqus
working directory and click on conform and generate “.odb” file as shown in following fig.
Fig.-6. 18 IGA Process

Step-12: Open “.odb” file generated from above step in Abaqus 6.14 Viewer to get final Iso-
Geometric Analysis Results. As shown in following fig.
Fig.-6. 19 Iso-Geometric Analysis Results

P a g e 74 | 92
6.2 Isogeometric Analysis of Hook Models
The result of Iso-Geometric analysis is given in following clauses:
P a g e 75 | 92
P a g e 76 | 92

P a g e 77 | 92
P a g e 78 | 92

P a g e 79 | 92

P a g e 80 | 92

P a g e 81 | 92

P a g e 82 | 92
6.2.5 Comparison of Hook Results

The IGA Results of hook models are list out in following table:
Table 7 IGA Result of Hooks
Equivalent
197.3 183.8 192.2 210.9
Stress (N/mm2)
Maximum
(N/mm2)
Minimum Principal
187.5 158.9 122.6 166.8
Stress (N/mm2)
Total Deformation
0.580 0.6844 0.3182 0.3603
(mm)
6.3 Comparison of FEA and IGA for Hook Models

The variation in FEA and IGA results are listed in following table:
Table 8 Comparative Percentage difference between FEA and IGA for Hook Models
Equivalent
7.50 5.012 6.92 1.171
Stress (%)
Maximum
Principal 1.62 2.119 2.31 1.01
Stress (%)
Minimum
Principal 4.23 2.997 6.98 3.695
Stress (%)
Total
Deformation 6.45 1.354 2.33 4.65
(%)
P a g e 83 | 92
6.4 Iso-Geometric Analysis of Curved Beam
The result of Iso-Geometric analysis is given in following clauses:

P a g e 84 | 92
P a g e 85 | 92
P a g e 86 | 92
P a g e 87 | 92
6.4.3 Comparison of Curved Beam Results:

The IGA Results of curved beam models are list out in following table:
Table 9 IGA Result of Curved beam
Model-1 Model-2
Equivalent Stress
203.5 204.5
(N/mm2)
Maximum Principal Stress
235.7 236.7
(N/mm2)
Minimum Principal Stress
235.6 236.7
(N/mm2)
Total Deformation
9.938 43.07
(mm)
6.5 Comparison of FEA and IGA for Curved Beam Models

The variation in FEA and IGA results are listed in following table:
Table 10 Comparative Percentage difference between FEA and IGA for Curved Beam
Models
Model-1 Model-2
Equivalent
4.50 5.36
Stress (%)
Maximum Principal Stress
3.32 4.556
(%)
Minimum Principal Stress
3.28 4.594
(%)
Total Deformation (%)
5.62 6.08
P a g e 88 | 92
Chapter-7 Summary of Findings

The main focus of this study is to understand the behaviour of curvilinear elements in terms of
stresses and deformation. It is known to us that the FEA gives a good satisfactory results for
any curved or complex geometry. But during literature review it has been found that IGA may
give more perfect and satisfactory results than FEA.
In this study, the behaviour of hook with their different cross section models and beam curved
in plan has been studied and comparative result between FEA and IGA has been presented.
From the result it has been concluded that the IGA gives less value for stress and deformation
than the FEA rather keeping same loading condition.
Comparative result for Hook models are as under:
 For Circular Cross Section of hook models, IGA values for Equivalent Stress (Von-
Mises), Maximum Principal Stress, Minimum Principal Stress and Total Deformation
gives 7.50%, 1.62%, 4.23%, 6.45% respectively less than FEA values for same loads.
 For Rectangular Cross Section (Width is blend) of hook models, IGA values for
Equivalent Stress (Von-Mises), Maximum Principal Stress, Minimum Principal Stress
and Total Deformation gives 5.012%, 2.119%, 2.997%, 1.354% respectively less than
FEA values for same loads.
 For Rectangular Cross Section (Depth is blend) of hook models, IGA values for
and Total Deformation gives 6.92%, 2.31%, 6.98%, 2.33% respectively less than FEA
values for same loads.
 For Trapezoidal Cross Section (Depth is blend) of hook models, IGA values for
and Total Deformation gives 1.171%, 1.01%, 3.695%, 4.65% respectively less than
FEA values for same loads.
Comparative results for beam curved in plan are as under:
 For Quarter Circle Curve Beam, IGA values for Equivalent Stress (Von-Mises),
Maximum Principal Stress, Minimum Principal Stress and Total Deformation gives
4.50%, 3.32%, 3.28%, 5.62% respectively less than FEA values for same loads.
P a g e 89 | 92
 For Semi Circular Curve Beam, IGA values for Equivalent Stress (Von-Mises),
Maximum Principal Stress, Minimum Principal Stress and Total Deformation gives
5.36%, 4.50%, 4.59%, 6.08% respectively less than FEA values for same loads.
As shown in advantage of IGA, IGA improves the solution accuracy and directly integrate with
CAD model. From this conclusion it has been satisfied over number of curvilinear models.
P a g e 90 | 92
References
Research Papers:
 Francisco Javier Sayas, “A gentle introduction to the Finite Element Method”.
 A. Buffa, G. Sangalli, R. V. Azquez, “AN Introduction of Isogeometric Analysis”,

Santiago de Compostela, 2010.
 T.J.R. Hughes, J.A. Cottrell, Y. Bazilevs, “Isogeometric analysis: CAD, finite

elements, NURBS, exact geometry and mesh refinement”, Computer Methods Applied
Mechanics, Vol. 194 (2005), pg. 4135–4195
 Vinh Phu Nguyena, Cosmin Anitescu et al., “Iso-geometric analysis: An overview

and computer implementation aspects”, Mathematics and Computers in Simulation.
 Fubeder D., Simeon B. et al., “Fundamental aspects of shape optimization in the

context of Iso-geometric Analysis (IGA)”, Computer Methods in Applied Mechanics
and Engineering, Vol.-286(2015), pg.-313 to 331.
 Falco C., et al., “GeoPDEs: A research tool for Iso-Geometric Analysis of Partial
Differential Equations (PDEs)”, Advances in Engineering Software, Vol.-42(2011), pg.
1020-1034.
 Vuong A. V. et al., “ISOGAT: A 2D tutorial MATLAB code for Iso-geometric

Analysis (IGA)”, Computer Aided Geometric Design, Vol-27(2011), pg. 644-655.
 Xinkang Li et al., “NURBS based IGA of Beams & Plates using Higher Order Sher
Deformation Theory (HSDT)” Hindawi Publication, volume 2013.
 Daniel Rypl, Patzak Borek, “From the finite element analysis to the Iso-geometric
analysis in an object oriented computing environment”, Advances in Engineering
Software, vol.44 (2012), pg. 116-125.
 T.J.R. Hughes a, A. Reali b,d,e, G. Sangalli, “Efficient quadrature for NURBS-based

isogeometric analysis”, Comput. Methods Appl. Mech. Engrg. 199 (2010) 301–313
P a g e 91 | 92
 Dennis Ernens, “Finite Element Methods with exact geometry representation

including IsoGeometric Analysis, NURBS Enhanced Finite Element Method and Aniso
Geometric Analysis” Delft University of Technology.
Books:
 S. S. Bhavikatti, “Finite Element Analysis”, New Age International Publisher.
 J. A. Cottrell, T. J. R. Hughes, Y. Bazilevs, “Iso-geometric Analysis: Toward

Integration of CAD and FEA”, John Wiley & Sons, Ltd.
Manuals:
 “ABAQUS 6.14 Offline Help”, Simulia Dassault Systemes.

 “ABAQUS 6.14 Example Problem Manual Volume 1: Static and Dynamic Analysis”,
Simulia Dassault Systemes.
 “ABAQUS 6.14 Example Problem Manual Volume 2: Other Application and
Analysis”, Simulia Dassault Systemes.
P a g e 92 | 92

MSR Report

Uploaded by

Copyright:

Available Formats

You might also like

MSR Report

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

MSR Report

Uploaded by

Copyright:

Available Formats

“Comparative Study of Finite Element Analysis (FEA) And Isogeometric Analysis (IGA)”

LIST OF TABLES ................................................................................................................... 7

Chapter-1 Introduction ........................................................................................................ 9

1.1 Structure ................................................................................................................................. 9

3.3.1 Knot insertion: h-refinement ........................................................................................ 32

4.1 Validation Problem & Result ................................................................................................. 42

5.1 Introduction .......................................................................................................................... 51

6.1 Steps of IGA Using NURBS in ABAQUS 6.14 .......................................................................... 65

6.4.1 Result of Curved Beam Model-1 ................................................................................... 84

Fig.-3. 3 The creation of a curve ........................................................................................................... 31

Fig.-5. 9 Equivalent Stress For Hook Model-3 ..................................................................................... 55

Fig.-6. 19 Iso-Geometric Analysis Results ........................................................................................... 74

The structural members are connected by different types of support as

The Structure can be classified according to their uses as

 Civil Structures – Buildings, Bridges, Towers etc.

1.1.1 Types of Structure

Truss is an assemblage of slender members arranged in triangular/rectangular pattern.

(b). Cables & Arches:

Fig.-1. 2 Cable Supported Bridge & Arch

Fig.-1. 3 Frame Structure

(d). Surface Structures:

1.2 Structural Analysis

1.2.1 Conditions of Structural Analysis

1.2.2 Methods of Structural Analysis

1.3 Stiffness Method

number of simultaneous equations representing joint equilibrium of forces is equal to kinematic

 Direct Stiffness Method:

Where, [AR] = Member End Reaction

1.4 Introduction of Finite Element Method (FEM)

1.4.1 Definition in Finite Element Analysis Field

Discretization: A body of the structure is divided in to equivalent system of smaller body or

1.4.2 Steps Involved In FEM

Now, from Hook’s Law, 𝜎𝑥 = 𝐸 𝜀𝑥 = [𝐷][𝜀𝑥 ] = [𝐷][𝐵]{𝛿}.

1.4.3 Advantages of FEM

2. FEM comparatively considered fewer nodes to get good results.

1.4.4 Disadvantages of FEM

1.5 What is Isogeometric Analysis?

Isogeometric Analysis is based on NURBS, has refinement procedure related to h-

 Directly interacting with the CAD systems

 Greatly simplifying the refinement processes

 Improving the solution accuracy

 Reducing the computational costs and time

Chapter-2 Literature Review

2.2 Literature Survey

“Fundamental aspect of shape optimization in the context of Isogeometric Analysis

(1) Optimize first and discretize later

(2) Discretize first and optimize later

“GeoPDEs: A research tool for Isogeometric Analysis of Partial Differential Equations

(1) Define the parameterization: the geometry structure

(2) Domain Partition: Mesh Structure

(3) The discrete space: The space structure

GeoPDEs is also able to solve more complex structure as under:

“Isogeometric analysis: An overview and computer implementation aspects”, by Nguyen

“ISOGAT: A 2D tutorial MATLAB code for Isogeometric Analysis (IGA)”, by Vuong A.

2.3 Summary of Literature

An Isogeometric analysis is a new emerging technology in analysis field which is faster in

2.4 Scope of Work