Professional Documents
Culture Documents
HEAT3 User
HEAT3 User
Copyright 2005-2007 by the Cranes Software, Inc. 1607 E. Big Beaver Road,
Suite 250, Troy, Michigan 48083, U.S.A.
All rights reserved. No part of this manual may be reproduced, stored in a retrieval
system, or transmitted in any way without the prior authorization of Cranes
Software, Inc.
1-iii
Table of Contents
Preface - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - x
Version 15.0 New Features - - - - - - - - - - - - - - - - - - - xii
1 Introduction - - - - - - - - - - - - - - - - - - - - - - - - - - 1.1
1.1 General Description - - - - - - - - - - - - - - - - - - - - 1.1
1.1.1 NISA/HEAT Capabilities - - - - - - - - - - - - - - - 1.1
1.2 NISA Suite of FEA Software - - - - - - - - - - - - - - 1.3
1.2.1 NISA Softwares - - - - - - - - - - - - - - - - - - - - - 1.3
1.2.2 NISA Documentations - - - - - - - - - - - - - - - - 1.8
2 Theoretical Overview - - - - - - - - - - - - - - - - - - - - 2.1
2.1 Introduction - - - - - - - - - - - - - - - - - - - - - - - - - - 2.1
2.2 Governing Equations - - - - - - - - - - - - - - - - - - - - 2.1
2.3 Boundary Conditions - - - - - - - - - - - - - - - - - - - 2.6
2.4 Finite Element Formulation - - - - - - - - - - - - - - - 2.8
2.4.1 Galerkin Method - - - - - - - - - - - - - - - - - - - - - 2.8
2.4.2 Element Shape Function - - - - - - - - - - - - - - - - 2.9
2.4.3 Finite Element Formulation of Energy
Equation - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 2.15
2.5 Surface Radiation Heat Transfer - - - - - - - - - - - 2.17
2.5.1 Introduction - - - - - - - - - - - - - - - - - - - - - - - 2.17
2.5.2 Governing Equations - - - - - - - - - - - - - - - - - 2.17
2.5.3 Computation of View Factors - - - - - - - - - - - 2.20
2.5.4 Solution Procedure - - - - - - - - - - - - - - - - - - 2.21
2.5.5 Using Artificial Convection Boundary
Condition to Simulate Surface Radiation - - - - - - - - 2.24
2.5.6 Special Cases - - - - - - - - - - - - - - - - - - - - - - 2.25
iv
2.6 Phase Change Effects - - - - - - - - - - - - - - - - - - - 2.30
2.7 Convection Link - - - - - - - - - - - - - - - - - - - - - - 2.32
2.8 Radiation Link - - - - - - - - - - - - - - - - - - - - - - - 2.32
2.9 Beam Element - - - - - - - - - - - - - - - - - - - - - - - 2.34
2.10 Heat Transfer in Composites - - - - - - - - - - - - - 2.35
2.11 References - - - - - - - - - - - - - - - - - - - - - - - - - 2.40
3 NISA/HEAT Solution Technique - - - - - - - - - - - - 3.1
3.1 Introduction - - - - - - - - - - - - - - - - - - - - - - - - - - 3.1
3.2 Frontal Solver - - - - - - - - - - - - - - - - - - - - - - - - - 3.1
3.3 Iterative Solver - - - - - - - - - - - - - - - - - - - - - - - - 3.7
3.4 Solution Errors - - - - - - - - - - - - - - - - - - - - - - - 3.12
3.5 NISA/HEAT Files - - - - - - - - - - - - - - - - - - - - - 3.15
4 Element Library - - - - - - - - - - - - - - - - - - - - - - - - 4.1
4.1 Introduction - - - - - - - - - - - - - - - - - - - - - - - - - - 4.1
4.1.1 Element Types - - - - - - - - - - - - - - - - - - - - - - 4.1
4.1.2 Element Input - - - - - - - - - - - - - - - - - - - - - - 4.5
4.2 2-D Element (NKTP = 2) - - - - - - - - - - - - - - - - - 4.7
4.3 Axisymmetric Element (NKTP = 3) - - - - - - - - - - 4.9
4.4 3-D Element (NKTP = 4) - - - - - - - - - - - - - - - - 4.13
4.5 3-D Thin Shell Element for Heat Transfer
(NKTP=20) - - - - - - - - - - - - - - - - - - - - - - - - - - - - 4.17
4.6 3-D Laminated Composite Shell for heat transfer
(NKTP = 32) - - - - - - - - - - - - - - - - - - - - - - - - - - - 4.21
4.7 3-D Bar Element for Heat Transfer (NKTP = 12) 4.25
4.8 3-D Convection Link Element (NKTP = 49) - - - 4.27
4.9 3-D Radiation Link Element (NKTP = 50) - - - - 4.29
5 Input Setup - - - - - - - - - - - - - - - - - - - - - - - - - - - 5.1
5.1 Introduction - - - - - - - - - - - - - - - - - - - - - - - - - - 5.1
5.2 Free Format Input Features - - - - - - - - - - - - - - - - 5.5
6 Executive Commands - - - - - - - - - - - - - - - - - - - - 6.1
v
6.1 General Description - - - - - - - - - - - - - - - - - - - - 6.1
6.2 Executive Commands - - - - - - - - - - - - - - - - - - - 6.4
7 Model Data - - - - - - - - - - - - - - - - - - - - - - - - - - - 7.1
7.1 Introduction - - - - - - - - - - - - - - - - - - - - - - - - - - 7.1
7.2 Title - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 7.4
7.2.1 *TITLE Data Group - Problem Title - - - - - - - - 7.4
7.3 Element Data - - - - - - - - - - - - - - - - - - - - - - - - - 7.5
7.3.1 *ELTYPE Data Group - Element Type Selection
Table - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 7.5
7.3.2 *ELEMENTS Data Group - Element Definition 7.6
7.3.3 *E1 Data Group - Alternate Form for Element
Definition - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 7.13
7.3.4 *RCTABLE Data Group - Real Constant
Table - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 7.16
7.3.5 *LAMANGLE Data Group - Rotation Angles
for Laminated Composite Elements - - - - - - - - - - - 7.18
7.3.6 *LAMSEQ data group - Composite lamination
sequence data - - - - - - - - - - - - - - - - - - - - - - - - - - 7.20
7.4 Nodal Data - - - - - - - - - - - - - - - - - - - - - - - - - 7.29
7.4.1 *LCSYSTEM Data Group - Definition of
Local Coordinate Systems - - - - - - - - - - - - - - - - - - 7.29
7.4.2 *NODES Data Group - Nodal Coordinates - - 7.33
7.5 Material Data - - - - - - - - - - - - - - - - - - - - - - - - 7.40
7.5.1 *MATFLUID Data Group - Material Property
Data for Fluid - - - - - - - - - - - - - - - - - - - - - - - - - - 7.40
7.5.2 *PRINTCNTL data group - Selective printout
control - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 7.43
7.6 Miscellaneous Data - - - - - - - - - - - - - - - - - - - - 7.45
7.6.1 *SETS Data Group - Definition of Sets - - - - - 7.45
7.6.2 *TEMPFN Data Group - Temperature
Dependency Curves - - - - - - - - - - - - - - - - - - - - - - 7.48
7.6.3 *TIMEAMP Data Group - Definition of
Time Amplitude Curves - - - - - - - - - - - - - - - - - - - 7.49
7.6.4 *CPVAR Data Group - Coupled Variable Data 7.53
vi
8 Analysis Data - - - - - - - - - - - - - - - - - - - - - - - - - - 8.1
8.1 Analysis Control Data - - - - - - - - - - - - - - - - - - - 8.3
8.1.1 *FLCNTL Data Group - Fluid Load Case
Control - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 8.3
8.2 Boundary Conditions - - - - - - - - - - - - - - - - - - - - 8.6
8.2.1 *BCDVAR Data Group - Variable Nodal
Boundary Condition - - - - - - - - - - - - - - - - - - - - - - 8.6
8.2.2 *BCGVAR Data Group - Variable Elemental
Boundary Condition - - - - - - - - - - - - - - - - - - - - - - 8.8
8.2.3 *ICDS Data Group - Specified Dirichlet Type
Initial Conditions - - - - - - - - - - - - - - - - - - - - - - - -8.16
8.2.4 *ICGRAD Data Group - Neumann Type
Initial Conditions - - - - - - - - - - - - - - - - - - - - - - - -8.17
8.2.5 *PCHANGE - - - - - - - - - - - - - - - - - - - - - - -8.19
8.2.6 *STEPSIZE - - - - - - - - - - - - - - - - - - - - - - - -8.19
8.3 Output Control Data - - - - - - - - - - - - - - - - - - - 8.22
8.3.1 *PRINTCNTL data group - Selective
printout control - - - - - - - - - - - - - - - - - - - - - - - - - -8.22
8.4 Data Terminator - - - - - - - - - - - - - - - - - - - - - - 8.24
8.4.1 *ENDDATA Data Group - Input Data
Terminator - - - - - - - - - - - - - - - - - - - - - - - - - - - - -8.24
9 Hints for Using NISA/HEAT - - - - - - - - - - - - - - - 9.1
9.1 Introduction - - - - - - - - - - - - - - - - - - - - - - - - - - 9.1
9.2 Running NISA/HEAT - - - - - - - - - - - - - - - - - - - 9.1
9.2.1 Monitoring Convergence - - - - - - - - - - - - - - - 9.3
9.2.2 Managing Disk Space - - - - - - - - - - - - - - - - - 9.3
9.2.3 Killing a Running NISA/HEAT Job - - - - - - - 9.4
9.3 Modeling Hints - - - - - - - - - - - - - - - - - - - - - - - - 9.5
9.4 Analysis Hints - - - - - - - - - - - - - - - - - - - - - - - - 9.9
A Appendix A: Verification Problem - - - - - - - - - A.1
B NISA/HEAT to NISA/HEAT3 Translator - - - - B.1
C FLUTL: NISA/HEATs Utility Package- - - - - - C.1
C.1 About this Appendix - - - - - - - - - - - - - - - - - - - C.1
vii
C.2 Introduction - - - - - - - - - - - - - - - - - - - - - - - - - C.1
C.3 Basic Features of FLUTL - - - - - - - - - - - - - - - - C.1
C.4 Executing FLUTL - - - - - - - - - - - - - - - - - - - - - C.4
C.5 Menus and Pertinent Notes - - - - - - - - - - - - - - - C.5
C.6 Some Examples of FLUTL Usage - - - - - - - - - C.15
C.6.1 Example #1 - - - - - - - - - - - - - - - - - - - - - - - C.15
C.6.2 Example #2 - - - - - - - - - - - - - - - - - - - - - - - C.15
C.6.3 Printouts for Example #1 and Example #2 - - C.16
D Solving Thermal Stress Problems Using
NISA/HEAT and NISA II- - - - - - - - - - - - - - - - - - - D.1
E System of Units - - - - - - - - - - - - - - - - - - - - - - - - E.1
F Keyword Index - - - - - - - - - - - - - - - - - - - - - - - - F.i
viii
ix
Preface
You have with you the latest version of NISA - A standard in Finite Element
Analysis. NISA is one of the few, commercially available, proven and robust
Finite Element Analysis software that has enjoyed a long-standing presence in
the arena of engineering analysis and design. Today it is the result of more than
three decades of innovation and dedication of highly skilled scientists, technol-
ogy architects and software engineers. As a result, generations of scientists,
engineers and researchers have come to depend on NISA to solve their most
complex engineering problems.
NISA which has a heritage of more than 30 years, changed hands from EMRC,
to Cranes Software, Inc. in July of 2005. Cranes Software, Inc is a wholly
owned subsidiary of Cranes Software International Limited - a global software
products and solutions provider. With this change comes an induction of fresh
talent and resources which is poised to take NISA to a new level in the world of
FEA. NISA Version 15.0 is the achievement of a great development team
which has worked rigorously for the past year in accordance to the best in breed
software development life cycle management practices. As manuals are very
important to us, a lot of thought has gone into the design and content of the
manuals. These have been totally revamped as per the new features and look of
the product. New additions to the various modules have been consolidated and
presented in an integrated manner.
The NISA Shell is now more appropriately called Application Launcher. The
new application launcher makes selecting modules, input/ output files, analysis
type and various CAD/FEA translators a simple task. Significant updates and
error correction have been made to individual modules of the NISA suite of
programs.
xi
Version 15.0 New Features
NISA II: Improved iterative and sparse matrix equation solvers; end release for
pipe and elbow element; General spring element (NKTP=38) upgraded to a
general spring and damper element; facility to input user element stiffness,
mass, and damping matrices; rigid link forces output for linear transient
dynamic analysis; ability to post process larger problems involving multiple
load cases.
DISPLAY III: Pre & post processing support new features of NISA modules.
In addition to this, general features are: Viewing the input function such as time
amplitude, spectra etc. as a graph; Realistic plot for 3-D General Beam, 3-D
Straight pipe, 3-D Elbow elements (NKTP=48); Automatic selection of Master
and Slave node for Rigid links; Post processing of Non-linear Beam element
(NKTP 39) stresses; Post processing of 3-D beam results such as - filtering of
the results and report generation, reporting maximum stresses and ASME ratios
at critical points across section; Option to plot the XY points in the graph with-
xii
out the connecting lines; Area under the curve with respect to X-Axis; History
plot for nonlinear results for external results; Reading/ Writing of Multiple
External results for linear and nonlinear static analyses; Viewing results for
complete model using Symmetry for external results; Variation of stress triaxi-
ality along a line of nodes; Crack mouth opening displacement graph; Crack
opening area calculation; facility to integrate fluid quantities such as pressure,
temperature on surface.
DISPLAY IV: Pre & post processing support new features of NISA modules.
Key additional features in DISPLAY IV are: Enhancement of Dialog boxes,
accelerator keys. XP style file open dialog; enhancement of Entity-Status view
in the workspace to facilitate the deletion of a single entity or the entire group
using mouse right click or by using Delete key on the keyboard; a wizard that
helps navigate analysis data.
xiii
Chapter
1
Introduction
1-1
General Description
a
- Specified temperature ( T )
c
- Convection h , Ta
b a
- Radiation , T s
T : ambient temperature
: emmissivity
Ts : temperature of surroundings
- Polynomial
- Tabular
Post-processing with DISPLAY III
1-2
Introduction
- Nodal temperature
- View factors
NISA
This is the parent program of the NISA family. It is a general purpose finite
element analysis program to which all other programs in the NISA family are
1-3
NISA Suite of FEA Software
interfaced. NISA capabilities and input specification are given in this manual.
Other NISA related documentations are listed in Section 1.2.2.
DISPLAY
A 3-D interactive color graphics finite element pre- and post-processing program.
Highlights of the main capabilities are given in Display III/IV Users Manual.
COMSYN
A free interface component mode synthesis program to compute natural
frequencies and mode shapes of a structure from NISA eigenvalue analyses of its
components. As a special feature, this program can also compute natural
frequencies and mode shapes of cyclic symmetric structures given the results of
NISA eigenvalue analysis of a single sector.
STROPT
A general purpose finite element structural optimization program employing the
state-of-the-art optimization techniques including design sensitivity computation.
The optimization process involves appropriately re-sizing the cross-sectional
parameters (design variables) while simultaneously satisfying prescribed limits
(constraints) on the structural response to minimize material volume, mass or
weight. Constraints include displacements, stresses and natural frequencies in
addition to limits on the design variables.
The design variables may be fixed (constant valued) or free (variable valued), and
they may be grouped and linked in many possible ways. This facility is especially
useful for uniformity of design in certain regions of the structure, for creating a
symmetrical design under unsymmetric loading conditions, and for satisfying
certain design specifications. Optimization may be performed for multiple loading
conditions.
SHAPE
A finite element program for shape optimum design of structures. SHAPE adopts a
novel approach for the shape optimization of 2-D and 3-D continuum structures.
The novelty lies in that the design variation is not limited to the original
boundaries. Instead, new boundaries may be created by removal of material from
1-4
Introduction
SHAPE uses NISA as the analysis module and accepts the same NISA input file.
Specific optimization information, e.g., constraints, number of iterations, frozen
regions, are read from a separate file. For each new design, a new NISA input file
and an updated optimization input file are generated. In addition, SHAPE also
provides a NISA input file with smoothed boundaries. Furthermore, a separate file
for design history is output, to enable tracing of the changes in design. Each
improvement in the design may be plotted and postprocessed in DISPLAY.
SECOPT
A program for optimizing the area of beam cross-section under a set of constraints
on section properties as well as other constraints imposed by code requirements.
Additional constraints may be imposed to control local buckling of webs or flanges
and to satisfy other practical considerations. The section properties of the optimum
section may be written to a file in NISA input format.
ENDURE
Single and multi-channel loading may be used. Both deterministic and probabilistic
description of loading are available. The rainflow method of cycle-counting is used
for the general case of sequential load history. Stress input may be specified
1-5
NISA Suite of FEA Software
Fatigue life contours may be plotted using the postprocessing module of the
DISPLAY program.
FEAP
A program for the analysis of Printed Circuit Boards (PCBs) and electronic
systems. FEAP is completely interfaced with NISA for various types of analysis
including stress and thermal analysis, heat transfer with convective heat flow and
random vibration analysis.
The finite element model of the PCB is created interactively using a large built-in
library of commercial components such as: Dual Inline Packages (DIPs), hybrid
packages, flat packs, Leadless Ceramic Clip Carriers (LCCCs), etc. Direct
interface with several CAD geometry data bases is also available.
The Integrated Circuits (ICs) are modeled according to the method by which they
are mounted on the PCBs, either the traditional poke-through attachment or the
more recent surface-mounted technology. Once all components on a PCB are
defined, a finite element mesh is automatically generated for each IC and the entire
board. Appropriate elements, including isotropic or composite elements, are
chosen from the NISA element library.
NISA-3D/FLUID
A finite element program for analyzing a wide range of 2-D, axisymmetric and 3-D
fluid flow problems. Analysis capabilities include steady state and transient,
laminar and turbulent, incompressible and compressible and Newtonian and Non-
Newtonian fluid flow. In addition conductive and convective heat transfer
problems may be analyzed.
1-6
Introduction
DYMES
A general purpose program for mechanical system dynamics to analyze
constrained, multi-body and spatial mechanical systems, in which body elements
are connected through mechanical joints such as spherical, revolute and
translational joints and force elements such as nonlinear springs, dampers and
actuators.
NISA/P-ADAPT
A program for linear, static, structural analysis based on the adaptive p-version of
finite element method. The order of the displacement shape functions can be varied
from one to eight. The model may be constructed by using different shapes of
linear or quadratic elements. The adaptive program may be controlled either
manually or automatically. Certain types of geometric curves/surfaces are
represented exactly in the analysis, thereby increasing the accuracy of the analysis.
Zero and/or non-zero boundary constraints can be imposed on the nodes, edges
and/or faces of a model.
NISA/CIVIL
A finite element program that offers CAD based solutions to a wide variety of
problems encountered in the Analysis and Design of Reinforced Concrete and Steel
Structures. Backed by powerful NISA II Analysis and DISPLAY III/IV - the
graphical Pre and Post processor of NISA family of programs, NISA/CIVIL
provides excellent tools for modeling, associating design information and carry out
design process in Limit state and working stress methodologies of design. Design
results are processed to produce structural engineering drawings in AutoCAD
environment. Equipped with an extremely user friendly GUI and graphic displays,
NISA/CIVIL, presents an elegant platform for analysis and design of different
types of structures encountered in practice.
1-7
NISA Suite of FEA Software
NISA/EMAG
NISA/HEAT
A finite element program for analyzing a wide range of 2D, axisymmetric and 3D
thermal problems. Analysis capabilities include steady and transient analysis, 2D,
3D and axisymmetric domains, phase change, orthotropic thermal conductivity,
surface radiation with view factor computation, volumetric/nodal heat generation.
In addition conductive and convective heat transfer problems may be analyzed. In
addition heat transfer in laminated composite shells can be modeled.
NISA/AERO
1-8
Introduction
Training manuals
1-9
NISA Suite of FEA Software
1-10
Chapter
2
Theoretical Overview
2.1 Introduction
In this chapter, a brief review of the theoretical aspects of NISA/HEAT is given.
Section 2.2 presents the basic equations for heat transfer, namely the energy
equation. Initial and boundary conditions for this equation is discussed in
Section 2.3. Finite element formulation of the energy equation is given in
Section 2.4. Theoretical aspects for surface radiation and phase change are given
in Section 2.5 and Section 2.6, respectively.
T
q = [ K ] ------ (2.2-1)
x
or in component form:
qx k xx k xy k xz T x
q y = k yx k yy k yz T y (2.2-2)
qz k zx k zy k zz T z
2-1
Governing Equations
where q i is the rate of heat flow conducted per unit area in the direction i, [K] is the
thermal conductivity tensor for the material (Eqn. 2.2-1 assumes anisotropic
material behavior), and T x is the temperature gradient vector in Cartesian
coordinates.
2-2
Theoretical Overview
In the material principal directions, Eqn. 2.2-2 reduces to the diagonal form
q1 k 11 0 0 T x 1
q 2 = 0 k 22 0 T x 2 (2.2-3)
q3 0 0 k 33 T x 3
or simply,
T
q n = k n ------ (2.2-4)
n
The thermal conductivity matrix in Eqn. 2.2-2 may be obtained from that in Eqn.
2.2-3 by simple second order tensor transformation as
T
K x, y, z = T K 1, 2, 3 T (2.2-5)
where T is the second order transformation tensor between the two sets of axes
x,y,z and 1,2,3. It is noted that the principle of irreversible thermodynamics shows
that the tensor K has to be symmetric. If the material is homogeneous and isotropic,
then:
k xy = k yz = k zx = 0 , and
(2.2-6)
k xx = k yy = k zz = k
T T T T
k x ------ + k y ------ + k z ------ + q = c ------
b
(2.2-7)
x x
y y
z z t
In deriving the above equation, it is assumed that the principal material directions
are along the global xyz axes. kx, ky, kz are the same as kxx, kyy, kzz, respectively, qb is
the rate of heat generated per unit volume, t is the time, is the mass density of the
material and c is the specific heat (where in the case of conduction in solids, no
2-3
Governing Equations
distinction is made between the specific heat at constant pressure, c p , and the
specific heat at constant volume, c v ).
Eqn. 2.2-7 is the general equation for heat conduction in solids and it may take
one of the following special cases:
b 1 T
2T + q k = --- ------ (2.2-8)
t
where = k c is the thermal diffusivity
1 T
2T = --- ------ (2.2-10)
t
which is the diffusion equation
2T = 0 (2.2-11)
which is Laplaces equation.
(2.2-12)
= e + e + e
z x y y z z
2-4
Theoretical Overview
where ex, ey, and ez are the unit vectors in x,y and z directions, respectively, and
2 T 2 2
-
2
- + -------
= = -------2- + ------- 2
(2.2-13)
2
x y z
2-5
Boundary Conditions
T ( x, t ) = T s ( x, t ) on S 1, S 1 CS (2.3-1)
T s
k n ------ =q , S 2 CS (2.3-2)
n s = s 2
T
------ =0 , S 3 CS (2.3-3)
n s = s
3
which is the natural boundary condition and is a special case of the Neu-
mann of Cauchy B.C.
k T
------ = h ( T s T f ) (2.3-4)
n
where T s is the body surface temperature, T f is the fluid temperature, n
is the normal to the surface S c and h is the convective heat transfer coef-
ficient. The convective heat transfer coefficient, h, may be a function of
temperature or time and the fluid temperature, T f , may be a function of
time.
2-6
Theoretical Overview
where is the absorbtivity of the surface. Owing to Kirchoffs law, = and the
heat flux out of the surface is given by q:
4 4 T (2.3-7)
q = [ T s T e ] = k n ------
n
Note that may be function of temperature, surface finish and view factor between
the two bodies.
It is convienent to put equation (2.3.7) into the following form which is similar to
the convective boundary condition form:
q = h ( Ts Te ) (2.3-8)
where
2 2
h = ( T s + T e ) ( T s + T e ) (2.3-9)
2-7
Finite Element Formulation
T
T ( x, t ) = T ( t ) (2.4-1)
where the unknown T is column vector of nodal points and is the column vectors
of the shape functions. Substituting these into the governing equation yields an
equation of the form
f ( , T ) = R (2.4-2)
where R is the residual resulting from the use of the approximations of Eqn. 2.4-1.
The Galerkin form of the method of weighted residuals seeks to reduce the error
(residual) R to zero. This is done by achieving orthogonality between the residual
and weighting functions of the element which is expressed as
E ( f w ) d E ( R w ) d = 0 (2.4-3)
2-8
Theoretical Overview
For each of the element types mentioned above, the shape functions used are given
in terms of the natural coordinates of the element, i.e., and . For a 4 noded
isoparametric quadrilateral, the shape function is defined as
2-9
Finite Element Formulation
1 4(1 )(1 )
1 4(1 + )(1 )
= (2.4-4)
1 4(1 + )(1 + )
1 4(1 )(1 + )
1 4(1 ) ( 1 ) ( 1 )
2
1 2(1 )(1 )
1 4(1 + )(1 )( 1)
1 2(1 + )(1 )
2
= (2.4-5)
1 4(1 + )(1 + )( + 1 )
2
1 2(1 )(1 + )
1 4(1 )(1 + )( + 1 )
2
1 2(1 )(1 )
L1
= L2 (2.4-6)
L3
2-10
Theoretical Overview
L 1 ( 2L 1 1 )
4L 1 L 2
L 2 ( 2L 2 1 )
= (2.4-7)
4L 2 L 3
L 3 ( 2L 3 1 )
4L 3 L 1
1 8(1 )( 1 )( 1 )
1 8(1 + )(1 )(1 )
1 8(1 + )(1 + )(1 )
1 8(1 )(1 + )(1 )
= (2.4-8)
1 8(1 )(1 )(1 + )
1 8(1 + )(1 )(1 + )
1 8(1 + )(1 + )(1 + )
1 8(1 )(1 + )(1 + )
1
= 2 (2.4-9)
3
where
2-11
Finite Element Formulation
1 8(1 ) ( 1 ) ( 1 ) ( 2.0 )
2
1 4(1 )(1 )(1 )
1 8 ( 1 + ) ( 1 ) ( 1 ) ( 2.0 + )
2
1 4(1 + )(1 )(1 )
1 =
1 8 ( 1 + ) ( 1 + ) ( 1 ) ( 2.0 + + )
2
1 4(1 )(1 + )(1 )
1 8 ( 1 ) ( 1 + ) ( 1 ) ( 2.0 + )
2
1 4(1 )(1 )(1 )
2
1 4(1 )(1 )(1 )
1 4( 1 + )( 1 )( 1 )
2
2 =
1 4(1 + )(1 + )(1 )
2
1 4(1 )(1 + )(1 )
2
1 8(1 ) ( 1 ) ( 1 + ) ( 2.0 + )
2
1 4(1 )(1 )(1 + )
1 8 ( 1 + ) ( 1 ) ( 1 + ) ( 2.0 + + )
2
1 4(1 + )(1 )(1 + )
3 =
1 8 ( 1 + ) ( 1 + ) ( 1 + ) ( 2.0 + + + )
2
1 4(1 )(1 + )(1 + )
1 8 ( 1 ) ( 1 + ) ( 1 + ) ( 2.0 + + )
2
1 4(1 )(1 )(1 + )
2-12
Theoretical Overview
1 2 L1 ( 1 Z)
1 2 L2 ( 1 Z)
1 2 L3 ( 1 Z)
= (2.4-10)
1 2 L1 ( 1 + Z)
1 2 L2 ( 1 + Z)
1 2 L3 ( 1 + Z)
where (L1, L2, L3, Z) are the natural coordinates for the element and L1 + L2 + L3 = 1.
Also Z lies between -1 and 1.
2-13
Finite Element Formulation
2
1 2 L 1 ( 2L 1 1 ) ( 1 Z ) 1 2 L1 ( 1 Z )
2L 1 L 2 ( 1 Z )
2
1 2 L 2 ( 2L 2 1 ) ( 1 Z ) 1 2 L 2 ( 1 Z )
2L 2 L 3 ( 1 Z )
2
1 2 L 3 ( 2L 3 1 ) ( 1 Z ) 1 2 L 3 ( 1 Z )
2L 3 L 1 ( 1 Z )
2
L1 ( 1 Z )
= L2 ( 1 Z )
2
(2.4-11)
L3 ( 1 Z )
2
2
1 2 L 1 ( 2L 1 1 ) ( 1 + Z ) 1 2 L 1 ( 1 Z )
2L 1 L 2 ( 1 + Z )
1 2 L 2 ( 2L 2 1 ) ( 1 + Z ) 1 2 L 2 ( 1 Z )
2
2L 2 L 3 ( 1 + Z )
2
1 2 L 3 ( 2L 3 1 ) ( 1 + Z ) 1 2 L 3 ( 1 Z )
2L 3 L 1 ( 1 + Z )
L1
L2
= (2.4-12)
L3
L4
where L1, L2, L3, and L4 are the natural coordinates related by
L1 + L2 + L3 + L4 = 1 .
2-14
Theoretical Overview
L 1 ( 2L 1 1 )
4L 1 L 2
L 2 ( 2L 2 1 )
4L 2 L 3
L 3 ( 2L 3 1 )
= (2.4-13)
4L 3 L 1
4L 1 L 4
4L 2 L 4
4L 3 L 4
L 4 ( 2L 4 1 )
For shell elements shape functions are same as those for corresponding 2D
elements.
T- T T T
Cp ----- = ----- kx------ + k y ------ + k z------ + q (2.4-14)
t x x y y z z
n 1
TT - T T T
Cp --------------------- = ----- kx------ + k y ------ + k z------ + q
t x x y y z z
2-15
Finite Element Formulation
n1
T ----- T- T k T
C p ------ -----
- k k ------
t x x x y y y z z z
------ d = T
C p ------------- + q d
t
T T T T n1
C p ------- -----
- k ---------- k ----------
t x x x y y y z z z
k ---------- Tj d = C p ------------
T -
t
+ q d
C p T T T
---------- T + kx ------ -
- --------
+ k y ------- --------- + k z ------- --------- T j d
t x x y y z z
(2.4-15)
C
= T
kx ------ n x + ky T ------ z d + ----------p- T T jn 1 d + q d
------ y + k T
x y z z t
The boundary integral term appearing on the right hand side is evaluated based on
the type of boundary condition (i.e., heat flux, convection, radiation, surface
radiation) applied to the boundary surfaces.
[Kreyzig, 1972] has been used to convert volume integral WA into surface
integral appearing on the right hand side. The above can be written as
In the case of steady state problems, the terms containing t are set to zero.
2-16
Theoretical Overview
2.5.1 Introduction
Heat transfer by surface radiation is usually neglected or approximations are made
to include its effect on real life problems. Computed temperature distributions may,
therefore, be different than as would be expected. With the availability of more
(and increasing) computing resources, it is becoming possible to account for the
heat transfer via the surface radiation. Among a long list of applications involving
surface radiation, a few important ones relate to cooling of automobile engines, air
conditioning of passenger compartments, heat transfer in furnaces, etc.
Motivated by these applications, NISA/HEAT now allows its users to account for
heat transfer in computational domains by surface radiation in addition to
conduction and convection.
It must be pointed out here that the radiative heat transfer phenomenon is different
from its other two counterparts, i.e., conduction and convection. In the latter two
modes, the heat transfer rate is a function of local temperatures and/or their
gradients. However, in surface radiation heat transfer, all surfaces are connected
with each other in radiation heat transfer sense and heat transfer rate is proportional
to the difference in the fourth power of temperatures.
2-17
Surface Radiation Heat Transfer
N N
ij 1 j
-----
( ij Fij )Tj
4
---------- F ij q r = (2.5-1)
j j
j
j=1 j=1
where
1 j
R ij = -----ij ------------- F ij (2.5-3)
j
j
and
3
S ij = ( ij F ij )T j (2.5-4)
where ij is the Kronecker Delta, j refers to the surface emissivity of the jth
surface, Fij is a view factor between an ith and jth surface, q rj refers to radiative
flux leaving (+ve) the jth surface, is the Stefan-Boltzman constant, Tj is the
absolute temperature of the jth surface and N is the total number of surfaces with
radiative heat exchange. It is the radiative flux, qr, that couples the radiative heat
2-18
Theoretical Overview
balance with the overall energy conservation expressed by Eqn. . This will be
explained in Section 2.5.4.
2-19
Surface Radiation Heat Transfer
In the above defined Eqn. 2.5-1, Fij, the view factor between the surfaces i and j
must be computed. For a better understanding, i,j in the above equation can be
replaced with 1 and 2. Thus, view factor, F1-2, between two arbitrary surfaces (see
Figure 2.5.1), 1 and 2 is defined as a fraction of diffuse radiant energy leaving
surface 1 that arrives at surface 2. Mathematically,
1
F 1 2 = ------
A1 Cos 1 Cos 2
-------------------------------- dA 1 dA 2
2
(2.5-5)
A1 A2 r 12
where A1 and A2 are the areas of surfaces 1 and 2, respectively, r12 is the distance
between the two elemental areas dA1 and dA2, 1 is the angle between the position
dependent normal vector n and the line connecting dA1 and dA2. Angle 2 is
defined in a similar way. It must be noted that in defining the above expression for
the view factor, it is assumed that the two surfaces are gray, diffuse and isothermal.
Another important relation connecting the view factors F1-2 and F2-1 is as follows:
2-20
Theoretical Overview
In order to calculate view factors internally, the user must specify the radiation
surfaces in terms of the finite element faces of a discretized domain. The user must
also specify which of the two sides is a radiatively active side. These pieces of
information can be supplied very easily via the already existing *BCGVAR card in
the NISA file of NISA/HEAT. Each radiating face is taken as one radiation surface.
View factors between the radiating surfaces are automatically generated by NISA/
HEAT taking into account self-obstruction and obstructions due to a third surface.
However, in the absence of any third surface obstruction, setting IOBST=0 in the
Executive command,SRAD = ON, IOBST, NAXI, NDIV, , will result in a
saving of substantial computer time.
As can be assessed from the preceding discussion, computing view factors can
result in usage of excessive computer time. To economize this computation,
different techniques are used depending on whether the geometry being analyzed is
2D, 3D or axisymmetric. For example, double area integration method [Siegel and
Howell] is employed in comparison with contour integration method [Siegel and
Howell] when a 3D geometry, with radiation surfaces, is being analyzed. No
special directives are required when computing view factors for axisymmetric
geometries. NISA/HEAT internally generates a complete 3D model (with the axis
of symmetry as the X-axis [NISA/HEAT]) to calculate the required view factors.
Furthermore, for 2D problems, a completely different approach, called Hottels
crossed-string method [Siegel and Howell], is employed for its computational
efficiency and accuracy. Reference [Siegel and Howell] provides more details for
evaluating view factors for interested readers.
q d (2.5-7)
where
2-21
Surface Radiation Heat Transfer
q = qa + qc + qr
In the above equality qa, qc and qr refer to the applied heat flux, effect due to
convection boundary condition (*BCGVAR card) and effect due to radiation on
this surface, S respectively. Radiative effects can be specified via *BCGVAR card
with labels TG, and/or RAD. In the latter case, gray-body radiative effects will be
considered and q r is evaluated by solving Eqn. 2.5-2. It must be pointed out here
that a temperature field (see the right hand side of Eqn. 2.5-2) must be available, if
q r s, due to gray-body radiation effects are to be included. On the other hand, q r s
affect the distribution of temperature in the domain. This leads to a coupling
between the energy and radiative energy transfer systems. An effective solution
strategy must be established to solve this system of equations.
It has been observed that q r (and hence T) solution may not converge or may do so
slowly. An under relaxation of q r leads to its stabilization. This is achieved as
follows:
2-22
Theoretical Overview
i+1
q ri + 1 = q r + ( 1 )q ri (2.5-8)
* *
4
T h L
NR1 = -------------------------
-
k ( Th Tc )
where is the Stefan-Boltzman constant, Th and Tc are upper and lower absolute
temperatures, k is the material conductivity and L is characteristic length. If
radiation-conduction parameter is much larger than one, low values of relaxation
parameters must be used in order to get convergence. For problems involving
convection boundary conditions there is an important dimensionless grouping,
NR2, which is defined as
4 4
( T h Tc )
NR2 = ------------------------------
-
h ( Th Tc )
small value of NR2 helps convergence. The relaxation factor, should be picked
on the basis of the minimum of these dimensionless parameters (whichever ones
are relevant for a particular problem).
Together with checking for convergence of the velocity and temperature fields,
convergence is also checked for q r . L2 Norm (See Section 3.4) is evaluated and a
check is made against a user-defined tolerance (in the *FLCNTL card). The
solution procedure is continued until L2 Norm falls below the specified tolerance.
2-23
Surface Radiation Heat Transfer
2-24
Theoretical Overview
boundary condition, the effect of surface radiation is brought into the energy
equation. Contribution of radiative heat flux is not directly added. Note that with
this technique, temperature and radiative flux are still solved in a segregated
manner.
q = 0 = qr = qc
where q c is the local heat transfer flux (via conduction) and q r is the radiative flux
evaluated using Eqn. 2.5-1. This is automatically satisfied on the surfaces, where
no flux or temperature is prescribed (as is the case for a plane of symmetry).
Use of symmetry plane must be made very carefully. Figure a and Figure b show
two conduction problems with external boundaries as radiation surfaces. It is
appropriate to use symmetry plane for the problem shown in Figure a. However,
use of symmetry plane is not appropriate for the problem shown in Figure b . This
2-25
Surface Radiation Heat Transfer
is due to the fact that the use of symmetry plane will lead to, effectively, neglecting
radiation exchange between certain surfaces, for example, between surfaces A and
B. Therefore, if a problem such as the one shown in Figure b is to be solved, it is
recommended to model the complete geometry.
In some cases, radiation surfaces do not form a complete enclosure. In such cases,
radiant energy is received from the surroundings beyond the opening. In these
cases, it is still possible to complete the enclosure by considering a pseudo surface
at the opening. The pseudo surface is black with a temperature that yields the
equivalence of the black body emissive power and the radiant energy from the
surroundings. The total radiation heat leaving the pseudo surface is given by
j Aj Fjk T j
4 4
Q k = A k T s
j =1
j k, other openings
j Ak Fkj T j
4 4
= A k T s (2.5-9)
N
Q
j Aj Fkj T j
4 4
q k = -----k- = T s
Ak
j =1
j k , other openings
This value of heat flux is applied as a boundary condition at the pseudo surface
(opening), surface k, for solving radiation exchange Eqn. 2.5-1. In the above
equation, Ts is the temperature of the surroundings beyond the opening. The user
just specifies the temperature, Ts. Heat flux, q k , is calculated internally in the
program. However, heat flux values at radiation outlets are not used as boundary
conditions for the energy Eqn. 2.2-7.
2-26
Theoretical Overview
Fij Ts
4
1 must be included on the right hand side of Eqn. 2.5-9. Thus
j=1
this equation is modified as
N
1 j N N
-----ij ------------
j
- Fij q rj =
( ij
4
F ij )T j
1
Fij Ts4
(2.5-10)
j=1
j
j=1 j=1
Note that the last term will have contribution only for external surfaces. For an
internal surface,
Fij = 1 , implying that the last term in the above equation becomes zero. If Fij
j=1
= 0, for all j, Eqn. 2.5-9 reduces to radiation boundary condition
4 4
q rj = j ( T j Ts ) (2.5-11)
as desired.
2-27
Surface Radiation Heat Transfer
2-28
Theoretical Overview
2-29
Phase Change Effects
T = Tf (2.6-1)
dV
QA = L ------- (2.6-2)
dt
where Tf is the phase change temperature, Q is the energy that must be balanced in
converting the volume V with interface area A evolving or absorbing the latent
heat L.
There are several ways to handle the above mentioned conditions (Eqn. 2.6-1) and
(Eqn. 2.6-2) Reference [Tamma and Namburu], a review paper, evaluates some of
the techniques employed to satisfy Eqn. 2.6-1 and Eqn. 2.6-2. NISA/HEAT
employs the following methods for handling the phase change effects:
2-30
Theoretical Overview
dH dt
C = ----------------
dT dt
n n1
H(T ) H(T )
or, C = ------------------------------------------- (2.6-3)
n n1
T T
C ( T i T f )V i
Q i = ----------------------------------
- (2.6-4)
t
L Vi (2.6-5)
At that instant Eqn. 2.6-1 and (Eqn. 2.6-2) will be satisfied and the
phase change at node i is said to be completed. Note that for nodes in
liquid or solid regions, Qi will be 0.
2-31
Convection Link
hA 1 1
1 1
Note that the element consists of two nodes and the temperature need not be known
at either of nodes. The heat transfer coefficient and area for heat convection must
be specified and is given under *RCTABLE card in NISA/HEAT.
4 4
Q r = A i i G ij ( T i T j )
A i i G ij = A j j G ji
The effect of radiation heat transfer is modeled using the following element
stiffness matrix.
2-32
Theoretical Overview
2 2
A i i G ij ( T i + T j ) ( T i + T j ) 1 1
1 1
Note that the element stiffness matrix is linearized using the most recent available
node temperatures dependent and an iterative procedure is necessary to get the
final solution. This data needed to evaluate the element stiffness matrix (i.e.,
Stefan-Boltzman constant, emissivity, gray-body radiation factor Gij) is specified
under *RCTABLE card in NISA/HEAT.
The Gij values depend upon view factor matrix, Fij and emissivity values.
Calculation of matrix Gij is computationally very expensive. Each row of this
matrix obtained by solving NxN matrix, the coefficients of which depend on
emissivity values and view factors. Thus a different system of NxN equations must
be solved for each node. The user is referred ro Siegel and Howell (1992) for
further details.
2-33
Beam Element
2-34
Theoretical Overview
Theory:
In this first release, the first order lamination theory for composites is adapted to
understand the heat transfer analysis. Based on this, a linear temperature
distribution is assumed through the thickness of the laminated composites.
However, in-plane variation of the first order or higher orders can be used with a
choice of higher order elements using NKTP=32 and appropriate NORDR value.
In the laminated composites, the layers are stacked up three-dimensionally as
shown in Figure 2.10.1. Each layer can have principal directions for the thermal
conductivity and must possess a finite thickness. The local thermal conductivity
principal directions must be specified with respect to the global coordinate system.
Based on the first order lamination theory, the temperature field in three
dimensions can be written as [Noor, Argyris]:
where
1
T 0 = --- ( T t + T b ) (2.10-2)
2
and
1
T 1 = --- ( T t T b ) (2.10-3)
h
where Tt and Tb are the temperatures at the top and bottom surfaces of a laminate
2-35
Heat Transfer in Composites
2-36
Theoretical Overview
2-37
Heat Transfer in Composites
The energy Eqn. 2.2-7 is solved for obtaining the temperature distribution in the
laminated plates (or shells). In this equation K, the thermal conductivity can be
written as a tensor Kij where i, j both vary from 1 through 3, ie,
K xx K xy K xz
K yx K yy K yz (2.10-4)
K zx K zy K zz
Eqn. 2.2-7 is discretized using the Galerkin Procedure and after some algebraic
manipulations, one obtains algebraic equations with two degrees of freedom at
each node. (This is the main difference between the thin shell formulation with one
degree of freedom per node with NKTP=20 implemented in earlier versions of 3D-
FLUID and HEAT III).
The two degrees of freedom here are the average of the top and bottom temperature
gradient through the thickness and the temperature.
The boundary conditions of convection, radiation, heat flux and heat generation
can be aplied and have been described under *BCGVAR and *BCDVAR cards.
Whenever (average laminate) temperatures are known, these must be applied using
*BCDVAR card. The second degree of freedom will automatically be set to 0 at
those nodes when the boundary conditions are applied using DISPLAY III.
Another important point to remember is the order in which laminates must be
specified. These are further elaborated in Model Data under groups *LAMSEQ and
*LAMANGLE.
2-38
Theoretical Overview
For users convience, the thermal conductivitys variation with temperature can be
described by
(a) a polynomial
(b) a table
Both of the above capabilities are supported in 3D-FLUID and HEAT 3. The user
must input either the coefficient of the polynomial (up to 3 rd order, i.e., 4
coefficients) or provide a table specifying the amplification values of thermal
conductivity as described in *MATFLUID card.
2-39
References
2.11 References
Kreyszig, E., Advanced Engineering Mathematics, John Wiley, 1972.
Rohsenow, W.M., and Harnett, J.P., Handbook of Heat Transfer, McGraw Hill,
1973.
Tamma, K.K. and Namburu, R.R., Recent Advances, Trends and New
Perspectives via Enthalpy Based Finite Element Formulations for Applications to
Solidification Problems, International Journal for Numerical Methods in
Engineering, Vol. 30, pp. 803-820, 1990.
2-40
Chapter
3
NISA/HEAT Solution Technique
3.1 Introduction
NISA/HEAT supports both direct and iterative solution methods. The direct solver
used by NISA/HEAT in frontal solution technique. Two iterative solvers are
available: (1) conjugate gradient method and (2) Generalized minimal residual
(GMRES). Each of these methods are described in this chapter. A description of
different types of error reporting is also included.
3-1
Frontal Solver
In the frontal technique, the ordering of the elements determines the wavefront
size. The node numbering sequence does not have any effect on the wavefront size.
Again, this is unlike the banded solvers in which the node numbering is what
affects the bandwidth. The instantaneous wavefront size is defined as the number
of active equations that are in core when a typical element is assembled to the
coefficient matrix in the front. When an equation is fully summed, meaning that the
corresponding unknown has made its last appearance and will not appear again in
any subsequent element, it is eliminated from the current set of equations, thus
making room for a new unknown to occupy the same place. As an element is
brought into the front, unknowns making their first appearance occupy vacant or
new positions in the coefficient matrix, and unknowns which appeared earlier are
assembled to the existing coefficients. The assembled coefficient matrix, therefore,
keeps expanding and contracting as elements are assembled and unknowns making
their last appearance are eliminated.
Three quantities are defined to reflect the varying size of the frontwidth. The
maximum wavefront (Wmax) is the maximum size that the instantaneous wavefront
can attain:
W i = W i 1 + F i L i 1 , i = 1, 2,..., M (3.2-2)
where Fi is the number of variables making their first appearances as element (i) is
assembled to the front, and Li-1 is the number of variables that make their last
appearances when element number (i-1) was assembled to the front. Clearly, WM +
1 must be zero. The average wavefront size (Wav) is defined as
M
1-
W av = ----
M Wi (3.2-3)
i=1
3-2
NISA/HEAT Solution Technique
12
M
1
Wi
W rms = ---- 2 (3.2-4)
-
M
i= 1
The largest problem that can be solved using NISA/HEAT depends on the memory
available at a particular installation. On most main frames and mini computers, the
problem size is limited to a maximum wavefront of 2000. The maximum wavefront
should not be confused with the total number of degrees of freedom in the problem.
The wavefront limit is sufficiently large for most practical problems. It is not
unusual to have a problem with 30,000 DOF and a maximum wavefront of 500.
The three measures of the wavefront size are printed in the output file. In addition,
the wavefront history (the instantaneous frontwidth as the front advances through
the elements) may also be printed, see the executive command EXECUTION in
Section 6.2.
As mentioned earlier, the wavefront size depends on the element sequence. A nodal
degree of freedom becomes active as the element in which it appears first is
assembled into the front, and it becomes eligible for elimination when it makes its
last appearance. In order to reduce the wavefront size, the elements should be
ordered such that the first and last appearances of a typical node, as implied by the
element connectivity, are as close as possible. This is evident from Eqn. 3.2-2,
where the objective is to make the quantity (Fi - Li-1) minimum.
In other words, the element should be sequenced such that the wavefront advances
through the model continuously from one end to the other in the direction which
has the largest number of nodal points.
3-3
Frontal Solver
Wavefront Optimizer
One may easily optimize the wavefront size for a simple mesh such as the
rectangular mesh shown in Figure 3.2.1. But it becomes exceedingly difficult and
3-4
NISA/HEAT Solution Technique
more time consuming to minimize the wavefront for a more complicated mesh. In
addition, the elements are generally numbered in an order which is convenient to
the user for easier identification of the regions in the model. In many cases, the
user element sequence, defined by the ascending order of the element ID numbers,
does not yield a cost-effective wavefront size. As mentioned earlier, the solution
time is proportional to the square of the frontwidth.
NISA/HEAT has a built-in wavefront optimizer which does not require additional
input from the user, and can be controlled using a single command (see the
executive command RESEQUENCE in Section 6.2). The internal element
sequence is invisible to the user. All communications between the user and the
program, as far as the input specification and the output request are concerned, are
in terms of the users element identification numbers.
The specified boundary conditions are not considered in the resequencing process,
since these conditions are load case dependent, and can only reduce the frontwidth.
The calculated frontwidth is hence an upper bound. The optimizer is invoked only
once prior to the solution of all iterations.
The wavefront optimizer is cost-effective. The CPU time spent in the resequencing
process is only a small fraction of the total CPU time. The latter is, of course,
reduced due to reduction of the CPU time in the solution phase. For medium size
problems, say less than 5000 DOF, the wavefront optimization takes about 5 to
10% of the CPU time taken in a single pass through solution phase, and about 1 to
10% for larger size problems, depending on the computer installation.
The algorithm performs five independent trials, each with a different starting
element, for the minimization of the frontwidth. The starting elements are chosen
such that they are the least connected to the model or the most distant from each
other (in the topological sense). A typical trial may be terminated before
completion if it becomes apparent that the previous trial has better wavefront
statistics. The final internal element sequence is chosen from the completed trial
3-5
Frontal Solver
with the least mean square frontwidth (this is the default option, see executive
command RESEQUENCE).
1. The optimizer works with nodes on the front rather than nodal degrees of
freedom on the front. During the solution, some nodal degrees of free-
dom may be eliminated before all the elements connected to the node are
assembled.
2. The specified boundary conditions, which reduce the number of active
DOF, consequently the frontwidth, are not considered in the
resequencing process.
3-6
NISA/HEAT Solution Technique
where A is the coefficient matrix, x is the solution of the system of length N and b
is the right hand side.
In this type of solution procedure, the coefficient matrix (A) has to be symmetric
and positive definite, which is true for pure heat transfer problems. The basic
algorithm is as follows [2]:
3-7
Iterative Solver
3-8
NISA/HEAT Solution Technique
(1)
3. Set = r0 r0 2
4. For i = 1, 2,......, k do the following till the residual norm W(2) = ||b - A xi||2
is less than or equal to the tolerance limit or i = k (k is preset integer - no.
of orthogonal vectors to be formed before restart).
(i)
(i) W = A
(m)
hm,i = ( w, )
(m)
W = W h m, i
end
(iii) hi+1,i = || W ||2
(i + 1)
(iv) = W/hi+1,i
3-9
Iterative Solver
7. s(i+1) = b Ax 2
else x(0) = x
9. Restart the algorithm (go to step 2).
GMRES is guaranteed to converge to the exact arithmetic for any N x N matrix A.
3-10
NISA/HEAT Solution Technique
References:
1. W.E. Arnold, The Principle of Minimized Iterations in the solution of the
Matrix Eigenvalue Problem, Quart. Appl. Math., V. 9 (1951), pp. 17-29.
2. R. Chandra, Conjugate Gradient Methods for Partial Differential Equations,
Ph.D. Thesis, Department of Computer Science, Yale University, Connecticut,
1978.
3-11
Solution Errors
where A is the coefficient matrix and x and b are the vectors of unknown variable
and known right-hand side, respectively. It must be pointed out here that x refers to
the solution at each iteration in the case of steady state analysis. For transient
analysis, the solution is obtained at every inner iteration in each time step. In
problems dealing with nonlinear heat transfer, the matrix A is a function of the
solution from the previous iteration(s). The difference in solutions obtained at
subsequent iterations can be termed as Error at each node. For transient analysis,
it is possible to have one inner iteration per time step. In such cases, the above-
mentioned difference in solutions at subsequent iterations refer to that between two
consecutive time steps.
There are two methods of reporting the change in nodal solutions at two
consecutive iterations. These are termed as maximum difference and L2 Norm
errors and are reported in the non-dimensionalized form of temperatures.
Temperature is non-dimensionalized as either of the two methods can be selected
via an executive command ERROR (Chapter 6).
Maximum Difference
In this method of reporting the errors, a maximum value of the difference in
solutions at nodes between two consecutive iterations is reported. This can be
written as follows:
k k k 1
D error = Max [ T i T i ] , i=1,...., N (3.4-2)
where Derror is the error in the variable T, the superscript k represents the iteration
number, the subscript i indicates the node number and N is the total number of
nodes in the model.
3-12
NISA/HEAT Solution Technique
When reporting this error in the output file (and log file, if any), also given are the
corresponding variable name, iteration number and the node number where this
maximum difference is found. A typical message for the variable temperature is as
follows:
T T ref1
T = ------------------------------
- , where Tref1 and Tref2 are the minimum and maximum
T ref2 T ref1
allowable temperatures in the domain, and inputted through *FLCNTL Card (see
Chapter 8 for details).
L2 Norm
N
k 2
( xi
k+1
xi )
i=1
L 2 Norm = -------------------------------------------------
-
N
k+1 2
( xi )
i=1
where xi is any nodal quantity at the in node, k is a step no. or an iteration no., and
N is the total no. of nodes.
3-13
Solution Errors
Monitoring Convergence
Error at each iteration or time-step is also written to file 56. A plot of error vs.
iteration can be obtained by typing the following command:
monitor [- options...]
Note that we must save file 56. File 56 can be saved by including unit 56 in the
executive command:
3-14
NISA/HEAT Solution Technique
3-15
NISA/HEAT Files
+
file codes designated by dash () are system dependent
3-16
Chapter
4
Element Library
4.1 Introduction
The available NORDR values for each element type are given in the detailed
description of each element. NORDR controls both the element shape functions
(linear through cubic) and the geometry of the element (quadrilateral or triangular,
with straight or curved sides). Considering all available combinations of NKTP
and NORDR, more than twenty-five different element types are available in the
NISA/HEAT element library.
4-1
Introduction
4-2
Element Library
- Phase Change
4-3
Introduction
All the elements used in NISA/HEAT are isoparametric which means that the same
shape functions are used to represent both the element geometry and the
temperature variation within an element. Elements with mid-side nodes can have
curved sides. For example, NORDR = 2 elements use three nodes to define each
side. Since a parabolic curve can exactly fit three points, the sides of these elements
can exactly match a parabolically curved boundary. For shell elements, coordinates
of a generic point, with M nodal points are given by
M M
tk k Vni
k k
xi = k x i + ( 2)
k=1 k=1
4-4
Element Library
where
( , , ) natural coordinates
4-5
Introduction
4-6
Element Library
4-7
2-D Element (NKTP = 2)
4-8
Element Library
Note:
1. The element connectivity must be given in the order shown in Figure 4.2.1, in
which the node numbering sequence starts at a corner node and proceeds along
the perimeter of the element in a counterclockwise direction.
4-9
Axisymmetric Element (NKTP = 3)
4-10
Element Library
4-11
Axisymmetric Element (NKTP = 3)
Notes:
1. The element connectivity must be given in the order shown in Figure 4.3.1, in
which the node numbering sequence starts at a corner node and proceeds along
the perimeter of the element in a counterclockwise direction.
2. Once a reference is made to NKTP = 3, the first coordinate entry in the
description of nodes is interpreted as that in the direction of the axis of
symmetry (i.e., Z-axis of the Z-R- system) while the second entry refers to
the radial direction (i.e., R-axis). If the modelling plane is represented by X-Y
plane (as in the case of NISA pre-processor DISPLAY III), this can be easily
achieved by making Z-R of the axisymmetric system coincident to the X-Y
system of the modelling plane. Also, the model must be contained in the
positive Z - positive R quadrant (or if using DISPLAY III, then in the positive
X - positive Y of the DISPLAY III modelling plane).
3. Axis of symmetry must be X or Y-axis. Default is X-axis. Axis of symmetry is
specified in the executive card DIME (Section 6.2).
4-12
Element Library
The element can be shaped as an 8-, 20- or 27- noded hexahedron (brick) element,
or as a 6-, 15- or 18-noded wedge element depending on the selected NORDR
value. The element configuration, node locations and face numbering convention
are shown in Figure 4.4.1. A general description of the input and output data
pertinent to this elemet is given in table Table .
4-13
3-D Element (NKTP = 4)
4-14
Element Library
4-15
3-D Element (NKTP = 4)
Notes:
1. The element connectivity must be given in the order shown in Figure 4.4.1, in
which the node numbering sequence starts at a corner node and proceeds along
the perimeter of a face on the element. This face will be considered face-1 of
the element. Following the right hand rule, the midside nodes of the edges are
then defined with the first edge midside node above the first node of face-1.
Face-2 of the element follows similarly. For wedge elements, face-1 must be a
triangular face.
2. The linear quadrilateral element (NORDR = 1) uses trilinear shape functions
to represent the element geometry.
4-16
Element Library
4-17
3-D Thin Shell Element for Heat Transfer (NKTP=20)
4-18
Element Library
4-19
3-D Thin Shell Element for Heat Transfer (NKTP=20)
Notes:
1. The element connectivity must be given in the order shown in Figure 4.5.1, in
which the node numbering sequence starts at a corner node and proceeds along
the perimeter of the element in a counterclockwise direction.
2. The element is suitable for modeling curved thin shell structures. Heat flux,
convection and radiation, if any, may be applied to the top and bottom faces.
Top and bottom are determined as follows:
(a) Choose a view point so that the element connectivities are defiend coun-
terclockwise.
(b) The top face of the element is now closer to the observer. The bottom
and top faces are numbered 1 and 2, respectively. The inplane faces of
quadrilateral elements are numbered 3 to 6 (3 to 5 for triangles), as
shown in Figure 4.5.1.
4-20
Element Library
4.6 3-D Laminated Composite Shell for heat transfer (NKTP = 32)
This is a curved thin 3D Shell element oriented arbitrarily in space and is suitable
for modeling heat transfer in moderately thick to thin laminated composite shells.
The element consists of a number of layers of perfectly bonded orthogonal
materials. The element has two degrees of freedom per node (T and T n ). The
theoretical basis of the element is discussed in section 2.10. An element reference
guide is briefed in Table 4.6.1 with the available loading listed in Table 4.6.2.
Figure 4.6-1 3-D laminated composite shell (NKTP = 32), (a) element configuration,
(b) layer setup, (c) face numbering convention for top and bottom surfaces
4-21
3-D Laminated Composite Shell for heat transfer (NKTP = 32)
node locations and face numbering convention are shown in Table 4.6-1. The
lamination sequence is between the top and bottom faces of the element with the
layer setup starting from the top face. Table 4.6-1 gives the output pertinent to this
element.
4-22
Element Library
Notes:
For pressure loading, the pressure (force/area) is integrated over the area of the
loaded face. For curved shells, this area will be somewhat different than the
area of the shell midsurface.
4-23
3-D Laminated Composite Shell for heat transfer (NKTP = 32)
4. For most NISA elements, the variable MATID in the *ELEMENTS data group
refers to a material property table defined in *MATERIAL data group.
However, this is inadequate for composite shells, since each layer may be of a
different material. Therefore, for this element, MATID is a pointer to an entry
in the *LAMSEQ (lamination sequence) data group. The *LAMSEQ data
group in turn contains pointers to the layer thicknesses, rotation angles, and
materials for all layers (*RCTABLE, *LAMANGLE and *MATERIAL data
groups, respectively). In this way, we retain the ability to model composite
shells which have tapered thickness or variable rotation angles. There is no
limitation on the number of layers or the leap sequence. Note that the number
of real constant tables, lamination angle tables, or material property tables may
not necessarily be equal to the number of layers, since more than one layer
may be of the same material type, thickness type, or lamination angle
(orientation) type.
5. There is no need to enter a non-zero value for the variable IDRC (real constant
ID number) in the *ELEMENTS data group for this element. Any value
entered for IDRC will be reset to zero.
6. The element consists of a number of layers of perfectly bonded orthotropic
materials. For each different material, for transient or steady state heat transfer
analysis DENS, COND, SPEC or DENS, COND respectively are to be
specified. Several options are available for defination of material principal
axes, see *LAMSEQ for details.
7. The following special plots are available for this element. These plots may be
obtained using the postprocessing module of the DISPLAY program:
(a) Contour plots of temperature on the top and bottom faces of the elements.
4-24
Element Library
The element configuration and face number convention are shown in Figure 4.7.1.
Table gives the output pertinent to this element.
4-25
3-D Bar Element for Heat Transfer (NKTP = 12)
Notes:
1. For steady state heat transfer analysis, the values of DENS and C are not
required.
4-26
Element Library
2. The element has three faces. Area and perimeter are specified at the two end
nodes (input in *RCTABLE data group).
4-27
3-D Convection Link Element (NKTP = 49)
Notes:
1. The convection surface area, A, in *RCTABLE data group, for this element is
the surface area over which heat is convected. In an axisymmetric analysis, the
convection area must be the total circumferential area.
2. Nodal points may or may not be coincident.
4-28
Element Library
4-29
3-D Radiation Link Element (NKTP = 50)
Notes:
1. If the emmissivity is temperature dependent, then the curve identification
number is given in *RCTABLE data group and the corresponding curve is
input in *TEMPFN data group.
4-30
Element Library
2. The input quantity, A in *RCTABLE data group, for this element is the surface
area over which heat is radiating. In an axisymmetric analysis, the radiation
area must be the total circumferential area.
3. Nodal points may or may not be coincident.
4-31
3-D Radiation Link Element (NKTP = 50)
4-32
Chapter
5
Input Setup
5.1 Introduction
The input data setup for a typical NISA/HEAT analysis run, is described in this
chapter.
This chapter also explains the rules of the free format input. A detailed description
of the executive commands, which constitute the first data block in a typical
NISA/HEAT deck, and their applicable analysis types is given in Chapter 6.
Detailed description of the other two data blocks, namely, the model data block
and the analysis data block are given in Chapter 7 and Chapter 8, respectively. A
list of all available cards in NISA/HEAT is given in Table .
The input data for a typical NISA/HEAT analysis type consists of three data
blocks and a data terminator as shown in Figure 5.1.1.
The three data blocks and the data terminator constituting a NISA/HEAT input
deck should appear in the sequence shown in Figure 5.1.1, namely:
- Data terminator
A brief description of each data block and its general format are as follows:
5-1
Introduction
where the left hand side (Command name) is a character string, and the right
hand side (option or data) is an applicable option(s) or data for the specific com-
mand. Detailed description of these commands is given in Chapter 6.
5-2
Input Setup
5-3
Introduction
The analysis data block consists, in general, of a delimiter data group followed by
distinct data groups that can be arranged in arbitrary form in the block. The
delimiter data group is *FLCNTL. Each data group, including the delimiter group,
consists of a group identification card that has to be the first card in the group,
followed by the pertinent free format data of the group.
Detailed descriptions of all the groups of the analysis data block are given in
Chapter 8.
4. Data Terminator:
The input data terminator signals the end of the data deck. The *ENDDATA group
identification card represents the data deck terminator which must be the last card
in the input data deck.
5-4
Input Setup
Comment card(s) may be inserted anywhere in a NISA deck. Any card starting
with two asterisks (**) is a comment card, e.g.:
** boundary conditions
Comment cards are for the users reference only and they are ignored by the
program.
where the group name is a character string constituting a valid group name. The
parameter is a pertinent group feature and the data is a value assigned to the
parameter.
5-5
Free Format Input Features
3. Data Cards:
The data for each group follows the appropriate group identification card. The
precise content of each data group is detailed in Chapter 7 and Chapter 8. In what
follows, we provide some general rules and/or restrictions applicable to all data
cards:
- Special tab character ($) is marked on many data cards. Inserting the tab char-
acter on a data card has the effect of supplying null entries for all variables up
to the variable following the tab character. For example, in the data for the
*NODES data group (see Chapter 7 ), there is a tab character ($) between the
4th and 5th entry on the card. Thus the following entries are equivalent:
5-6
Input Setup
*ELTYPE
1, 2, 1
*READ, FILE2
...
...
...
5-7
Free Format Input Features
Table 5.2-1 Alphabetical list of all valid command and input data groups
Section
Group ID(1) Description
No.
ANALYSIS 6.2 Specify analysis type
ANISO Indication of presence of direction-dependent
6.2
properties
*BCDVAR 8.2.1 Variable Dirichlet type boundary conditions
*BCGVAR 8.2.1 Variable Neumann type boundary conditions
BFACE 6.2 Generate a list of boundary faces
CNVBC Lump matrix due to convection boundary
6.2
condition
*CPVAR 7.6.4 Coupled variable data
DIMENSION 6.2 Specify dimension of problem
DTAUTO 6.2 Automatic time step size control
*E1 7.3.3 Alternate element definition
ELEMENTECHO 6.2 Echo element input
*ELEMENTS 7.3.2 Element definition
*ELTYPE 7.3.1 Element type specification
*ENDDATA 8.4.1 Input data terminator
ERROR 6.2 Specify convergence criteria
EXECUTION 6.2 Select execution/checking run
FILENAME 6.2 Specify NISA file prefix
*FLCNTL 8.1.1 Fluid load case control
*ICDS 8.2.3 Dirichlet type initial condition
*ICGRAD 8.2.5 Neumann type initial condition
INITIALC 6.2 Specify a uniform initial condition
ITERATION 6.2 Iteration strategy
*LCSYSTEM 7.4.1 Local coordinate system definiton
MASS_LUMPING 6.2 Lump mass matrix
*MATFLUID 7.5.1 Material property data
MEMORY 6.2 Initial captured memory
5-8
Input Setup
Table 5.2-1 Alphabetical list of all valid command and input data groups
*NODES 7.4.1 Nodal coordinates
NODEECHO 6.2 Echo node input
PCHANGE 6.2 Turn on phase change computation
*PCHANGE 8.2.6 Specification of phase change parameter
*PRINTCNTL 8.3.1 Selective printout control
Table 5.2-2 Alphabetical list of all valid command and input data groups (cont.)
Section
Group ID(1) Description
No.
RESEQUENCE 6.2 Element resequencing
RESTART 6.2 Restart options
SAVEFILE 6.2 Save specific NISA file
*SETS 7.6.1 Specification of a set of numbers
SRAD 6.2 Analysis with surface radiation
SOLUTION 6.2 Solution method
STDS 6.2 Steady state analysis
STEP 6.2 Time step size control
TCUTOFF 6.2 Temperature cutoff
TEMP_OFFSET 6.2 Temperature offset
*TEMPFN 7.6.2 Temperature dependency curves
*TIMEAMP 7.6.3 Time amplitude curves
*TITLE 7.2.1 Problem title
UFIJ 6.2 View factor and/or iteration control
WARNING 6.2 Set warning flag
(1)
Acceptable minimum abbreviations are in bold face
5-9
Free Format Input Features
5-10
Chapter
6
Executive Commands
The option or data is a character string or a data value that will be assigned to the
specific command. Section 6.2 gives detailed description of executive commands
that are applicable to specific analysis types. The listing of the commands within
each section is arranged alphabetically and an overall alphabetical listing is given
in Table . The following general notes apply to all executive commands:
6-1
General Description
6-2
Executive Commands
(1)
Acceptable minimum abbreviations are in bold face.
6-3
Executive Commands
ANALYSIS = [HEAT]
where
HEAT : indicates heat transfer analysis
ANISO = ON
{ OFF }
where,
BFACE = ON
{ OFF }
6-4
Executive Commands
CNVBC = LUMP
{ OFF }
This feature is useful when there is a large heat gain or loss due to convection
boundary condition. This option is useful in avoiding any unphysical over-shoot or
under-shoot in the temperature solution.
where
Notes:
1. There are no defaults for TEND, DTMIN and DTMAX. The user must supply
these values.
2. Currently this capability is available for ANAL = HEAT only.
6-5
Executive Commands
DIMENSION = {2D}
AX, AXIS
3D
where
2D : Two-dimensional problem
AX : Axisymmetric problem
AXIS : Axis of symmetry. It must be either x or y-. The default is x, i.e.,
x-axis.
3D : Three-dimensional problem
ELEMENT_ECHO = { ON }
OFF
where
ERROR = { MAXDIF }
L2NORM
6-6
Executive Commands
where
EXECUTION = {GO}
CHECK
CGO
where
FILE_NAME = fname
{ no default }
where
6-7
Executive Commands
This command must be used in conjunction with the SAVE FILE command which
specifies the NISA file numbers to be saved (e.g., 26, 27,...).
INITIAL_CONDITION = VALUE ,T
{ T lower , T }
where
If temperature has been specified on some particular nodes in the *ICDS data
group, then the initialization values given here will be superceded by those in the
*ICDS card. If this card is not present, then temperature is initialized to TLower as
specified in the *FLCNTL card (ref. Analysis data). Further, *BCDVAR will take
precedence over initial condition.
where
METHOD : Name iteration type adopted in each time step.
= FIXED Fixed point iteration
= MNR Modified Newton-Raphson iteration
(currently unavailable)
6-8
Executive Commands
n
n i
K u i u in = f u
where, u is a typical variable (say velocity), and
n n n1
ui = u i 1 + ( 1 ) u
6-9
Executive Commands
where
Notes:
1. Usually lumping of mass matrix for the energy equation can help in avoiding
unphysical overshoot or undershoot in temperature solution.
2. This option is useful for transient problems only.
MEMORY - Initial captured memory
MEMORY = { 1000000 }
N
where
6-10
Executive Commands
NODEECHO = { ON }
OFF
where
Note:
Users are advised to use the more general *ECHO data group (Section 5.2) since
the NODE ECHO command will not be supported in future versions.
PATH = Pname
{ no default }
where
6-11
Executive Commands
Note:
NISA temporary file will be opened in users current working directory if this
command is not used.
PAT1
PAT2
PAT3
PAT4 - Directory for NISA files
where
NISA file numbers which should be open in pname directory. These four
commands allow the user to open files in four different directories other than the
current working directory during a NISA run. Saved files, however, will be opened
in the current working directory.
6-12
Executive Commands
PCHANGE = { OFF }
ON, METHOD
where
If 1 is chosen for METHOD, it implies that enthalpy formulation for the phase
change is desired while METHOD = 2 invokes source method formulation. The
default value of 1 is chosen in case MEHTOD is not indicated.
where
6-13
Executive Commands
The element resequencing is transparent to the user. The element IDs are not
changed. All communications between the user and the program as far as input and
output are concerned are in terms of the users element identification numbers.
The command RESE = ON (default option) resequences the elements and selects
the trial with the least RMS wavefront, but does not provide a table of
correspondence for element numbering. The options may appear in any order.
RESTART = { 0 }
n, ISTEP
where
6-14
Executive Commands
Whenever ISTEP is applicable and is not supplied, the last step is used. Every time
a RESTART is made, the post-processing files, i.e., files 26 and 27, are
overwritten. This means that if the results prior to any RESTART are needed later
on, the files 26 and 27 should be copied with a different set of file names before
starting the RESTART run. Currently ISTEP is fixed at the last time step. Option 2
is currently unavailable.
where
This command must be used in conjunction with the FILE NAME command which
specifies a character prefix (fname) for files to be saved. Refer to Chapter 3 for
NISA file numbers.
Note:
For certain computer systems, the NISA files to be saved are actually specified by
the Job Control Language (JCL) and this command is not needed. Refer to
pertinent system manual for details.
{ FRONT }
SOLU = ITERATIVE, MAXIT, TOLER, METHOD, IOCORE
SPARSE
6-15
Executive Commands
where
FRONT : Direct, frontal solver used for the solution of the system of
linearized algebraic equations.
ITERATIVE : Iterative solver used for the solution of the system of
linearized algebraic equations
MAXIT : Maximum number of iteration for the iterative solver
TOLER : Required level of convergence or tolerance for the iterative
solver
METHOD : Iterative solution method
1 : Conjugate gradient (CG)
2 : Generalized minimization of residual (GMRES)
3 : Pre-conditioned Conjugate Gradient Method (PCG)
4 : Pre-conditioned Conjugate Gradient Method (PCKSS)
IOCORE : The out-of-core scheme activated instead of the in-core
scheme for the iteration process to speed up the
performance of the solver
= 0 in-core scheme for iteration is used.
= 1 out-of-core scheme for iteration is used.
Unless specified by the user, the iterative solver uses the following default values;
MATRIX = 400, TOLER = 10-6 and METHOD = 2.
SPARSE : Direct, sparse solver used for the solution of the system of
linearized algebraic equations.
6-16
Executive Commands
where
Notes:
1. Whenever possible, the user should use IOBST = 0. This leads to substantial
savings of CPU time for view factors calculation.
2. If NAXI is specified to be less than 8, NISA/HEAT will set this value equal to
8. If NAXI is specified to be a very large value (say 9999), it will be reset to the
maximum allowable value based on available memory.
3. For 2D-problems having third surface obstructions, the user can instruct
NISA/3D-FLUID to subdivide a radiation surface into NDIV number of
divisions for view factor computation. This leads to more accurate values of
view-factors. However, CPU time will also be higher. Maximum allowable
value for NDIV is equal to 10. There is no provision to subdivide a surface for
3D-problems.
4. Relaxation parameter for radiative flux should lie between 0 and 1
0 < ( ) < 1 . A small value of can slow down the rate of convergence and a
6-17
Executive Commands
high value of may cause solution to diverge. Values between 0.1 and 0.3 are
recommended.
5. Radiation view factors are stored in file 54 (binary), only if obstruction checks
are on. Therefore, view factors do not have to be recomputed at restarts. How-
ever, the user can force NISA/3D-FLUID to recompute the view factors at
restarts using the executive command SRAD=ON, IOBST, NAXI, NDIV,
RELAXQ, IRVIEW by setting IRVIEW to 1 will trigger recomputation of
view factors on restarts. Users may want to recompute view factors if they
wish to
Note that view factors will not be stored in file 54 if obstructions checks are not
performed. This is done to save the disk space required to store the view factors. If
view factors are not stored in file 54, they will automatically get recomputed at
restarts.
6. Unless specified by the user, the following default values will be used; IOBST
= 1, NAXI = 32, NDIV = 1, =0.2 and IRVIEW = 0.
STDS = ON, , ,
{ OFF }
6-18
Executive Commands
where
Notes:
1. The nonlinear finite element system of equations at nth iteration can be written
as
n 1
n 1 T n = f T
K T
n1
where T is a typical variable (say temperature), T is the under-relaxed value
of T after n -1 iteration. Solution after n iterations is taken to be
n n n1
T = T + ( 1 )T
when Tn is obtained after solving the above given system of equations, is called
the relaxation parameter.
2. Relaxation factors , and are set to 1.0 by default. For nonlinear energy
equation, a value of less than 1.0 may be used for temperature relaxation
parameter, .
6-19
Executive Commands
The users can subdivide the time step given in *FLCNTL card into still smaller
time steps using the executive command.
STEP = N
{1}
where, N is the number of subdivisions of the time step with default value of N
being 1.
In DISPLAY, the user will be able to look at the results at time steps given in the
*FLCNTL or *STEPSIZE cards. However, NISA/HEAT will use smaller time
steps size (DT/N) for time accurate solution.
TCUTOFF = ON
{ OFF }
If TCUTOFF = ON, REFT1 and REFT2 will be the upper and lower cutoff
temperatures. Otherwise, these are just used to non-dimensionalize the energy
equation.
TEMP_OFFSET = TVAL
{ 0, 0 }
This card can be used by users who wish to input data in degree Celsius in a
radiation problem. In such a case TVAL should be set to 273.15. The final results
will be output in degrees Celsius.
6-20
Executive Commands
where
NSTMIN : Minimum number of time steps for which solution must be run.
EPSUVW : Velocity tolerance for checking steady state
EPSTMP : Temperature tolerance for checking steady state.
Notes:
1. This card is used for transient analysis only.
2. Tolerance giving in *FLCNTL card is used to check convergence within the
time step. Tolerances given in this card are used to check if steady state has
been reached. If change in solution between two successive time steps is
smaller than these tolerances, solutions will be terminated.
6-21
Executive Commands
User defined view factors can also be read by NISA/HEAT. In addition, view
factors computed by NISA/3D-FLUID can be printed out in the output file. This is
done using an executive command.
where
IUSRVF : 0 = Do not read user defined view factors. Compute view factors
internally.
This is a default.
1 = Read user defined view factors, do not compute view factors
internally.
IPRVF : 0 = Do not print view factor matrix in the output file. This is
default.
1 = Print the view factor matrix in the output file.
IRADH : 0 = Do not use artificial convection boundary condition method
X for solving surface radiation problems with heat flux.
element heat generation etc. type of b.c.
1 = Use artificial convection boundary condition method.
HRAD : Typical value of radiative heat transfer coefficient in the domain
2 2
hr = ( T + Tr ) ( T + T r )
6-22
Executive Commands
Notes:
1. If user defined view-factors are to be read, they should be given the file vfac-
nisa.dat. Format for this file is given under Analysis Hints (Section 9.4 ).
2. Solution is independent of the value of HRAD. However, use of this option
can help stabalize the solution.
WARNING - Set warning flag
WARNING = { GO }
STOP
where
6-23
Executive Commands
6-24
Chapter
7
Model Data
7.1 Introduction
The NISA data deck consists of distinct data blocks as has been mentioned in
Chapter 5. The data blocks must be ordered in the sequence shown in Figure5.1.1.
The model data block represents the majority of the input data and is detailed in
this chapter. It describes the physical characteristics of the finite element model,
but without any boundary condition. The latter is described in the next chapter.
The description of the model primarily consists of:
1. Element Data
(a) Element types selection: A table of the element types used in the model.
2. Nodal Data
(a) Definition of local coordinate systems
3. Material Data
This includes specification of material properties.
7-1
Introduction
4. Miscellaneous Data
This includes the definition of sets (groups) of nodes or elements, definition of
temperature and time amplitude curves, vectors, etc., which may be referenced
in the analysis data.
The above input items of the model data are arranged in distinct data groups. Each
data group consists of a group ID card followed by its free format data, which may
consist of one or more card sets. Each card set may consist of one or more cards.
The free format rules are given in Section 5.2. Each group has a descriptive
identification name indicating the function of the data group. For example, the
element definition is given in the *ELEMENTS data group, the nodal coordinates
in *NODES data group.
The detailed description of the data groups is presented in this chapter in the
sequence shown in Table 7.1.1. This sequence is adopted according to the function
of each data group as shown above. For easy reference, an alphabetical list of all
data groups available in NISA/HEAT is given in Table . It should be noted that the
data groups may be arranged in the model data quite arbitrarily, with the following
exceptions:
1. The *ELTYPE data group must precede the *ELEMENTS data group.
2. It is recommended that the *TITLE data group be the first data group in
the model data set. This will allow the title of the problem being ana-
lyzed to be printed on every page of the output file.
It should also be noted that not all of the data groups given in this chapter need to
be present in the model data for the heat transfer analysis.
7-2
Model Data
7-3
Title
7.2 Title
Entry No: 1
Variable: Problem title
Max char: 80
Note:
1. Up to 6 cards (80 characters per card) may be used for the title. The content of
all cards will be printed when this data group is read in. The content of the first
card will be printed as a heading on all subsequent pages of the output file.
7-4
Model Data
Entry No: 1 2 3
Variable: NSRL NKTP NORDR
Max char: 2 3 2
Note:
1. The same element NKTP and NORDR values may be specified more than
once, with a different NSRL in this data group. The user may wish to do this so
7-5
Element Data
that, for example, changes to a model are easily recognized. It may also be
useful in plotting, since an option is available in the DISPLAY program to sort
elements by the NSRL values before plotting.
For individual element definition, the first card set (one card) will include only
element parameters. For first level element generation, the first card set will
include both element and generation parameters. For second level element
generation, the first card set will include only generation parameters and will be
followed by the appropriate sets of cards defining the base element(s) either
individually or via first level generation. Examples for various methods of element
definitions are given below.
7-6
Model Data
Entry No: 1 2 3 4 5 6 7
Variable: NELID MATID NSRL IDRC KISO NSET NODINC
Max char: 6 6 2 6 1 6 6
Entry No: 8 9
Variable: NELSET NELINC
Max char: 6 6
7-7
Element Data
Card set 2: more than one card if the element has more than ten nodes
7-8
Model Data
Notes:
1. If the element has more than 10 nodes (e.g., the 20-noded solid element), con-
tinue on additional cards in card set 2 to complete the element node list.
2. For first level element generation: (see example below)
- If there is only one element in the base set (NELSET = 1), then only two
sets of cards are required to generate NSET number of elements.
- If there is more than one element in the base set (NELSET > 1), then the
first base element along with the generation parameters will be defined
by the first two card sets. Definition for additional elements in the base
set should follow immediately. The total number of elements generated
will be equal to NELSET*NSET.
3. For second level element generation, card set 1 is used to define the generation
parameters, and should include only entries 6-9, i.e., start the card with a tab
($). Immediately after the definition of the second level generation parameters,
the NELSET number of base elements are defined on subsequent cards using
individual definition or first level generation. See examples below.
The following cards define one 15-noded wedge element whose element ID is 10,
material ID is 1, identification index is 1, and node list is 1 to 15. Comment cards
are included for explanation.
7-9
Element Data
**
**Card set 1: defines element parameters, no generation parameters needed
**
10, 1, 1, 0, 0
**
**Card set 2: element node list; note that two cards are required since the element has more
than 10
**nodes
1, 2, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 13, 14, 15
**
**
**Card set 1: defines element parameters, for the first element in the base set, along with first
level
**generation parameters.
**Five sets are generated including the base set which contains 3 elements
**
1, 1, 1, 1, 0, 5, 1, 3, 1
**
**Card set 2: defines node list for the first base element
**
1, 2, 8, 7
**
**Second element in the base set. Two sets of cards are needed. No generation parameters
are specified.
**
6, 1, 1, 0
7-10
Model Data
7, 8, 14, 13
**
**Third element in the base set.
**
11, 1, 1, 0
13, 14, 20, 19
**
7-11
Element Data
**
**Card set 1: defines second level generation parameters only 3 sets are to be generated from
a base set
**which contains 5 elements
**
$-3, 6, 5, 5
**
**Card set 1: defines the base element parameters and generation parameters for first level
generation
**
1, 1, 1, 0, 0, 5, 1, 1, 1
**
**Card set 2: defines node list for the base element of the first level generation
**
1, 2, 8, 7
**
7-12
Model Data
Users are advised to use the standard form of element definition, the *ELEMENTS
data group, since this data group (*E1) will not be supported in future versions.
For first level element generation, the first card set will define the first base
element node list and element parameters as well as the generation parameters. The
remaining elements in the base set, if any, should follow immediately with no
generation parameters. For second level generation, the first card set will define
only generation parameters. All elements in the base set should follow
immediately.
Card set 1: one card (except if the element has more than 8 nodes)
$
Entry No: 1 2 3 4-8 9 10
Variable: NELID NODE1 NODE2 ... NODE8 MATID
Max char: 6 6 6 ... 6 4
7-13
Element Data
Entry No: 11 12 13 14 15 16
Variable: NSRL IDRC NSET NODINC NELSET KISO
Max char: 2 4 4 6 3 1
7-14
Model Data
Notes:
1. If the element has more than 8 nodes, (e.g., the 20-noded solid element), com-
plete the node list on additional cards (8 nodes per card).
2. For first level element generation:
- If there is only one element in the base set (NELSET = 1), then one set of
cards (one card except if the element has more than 8 nodes) will be
required to generate the number of elements equal to NSET.
- If there is more than one element in the base set (NELSET > 1), then only
the first element definition will contain the generation parameters, i.e.,
entries 13-15. Definition for additional elements in the base set, if any,
should follow immediately. The total number of elements generated will
be equal to NELSET * NSET.
7-15
Element Data
3. For second level element generation, card set 1 is used to define the generation
parameters and should include only entries 13-15, i.e. start the card with two
tabs ($$). Immediately after the definition of the second level generation
parameters, start the definition of NELSET base elements using individual or
first level element generation.
Entry No: 1 2 3 4
Variable: IDRC NUMRC IFORM IDCRV
Max char: 6 2 1 6
entry variable description
1 IDRC real constant table identification number. It is the number
referred to in the *ELEMENTS data groups to define the
element real constants.
2 NUMRC number of real constants in this table to be input in the next
card(s).
3 IFORM key for number of entries per card in card set 2
= 0 (or blank)-8 entries per card
> 0-4 entries per card
7-16
Model Data
IFORM = 0
Entry No: 1 2 3 4 5 6 7 8
Variable: RC1 RC2 RC3 RC4 RC5 RC6 RC7 RC8
Max char:
OR
IFORM >0
Entry No: 1 2 3 4
Variable: RC1 RC2 RC3 RC4
Max char: 20 20 20 20
7-17
Element Data
Notes:
1. Use as many cards as needed in card set 2 to enter NUMRC number of real
constants in this table. Data should be entered according to the IFORM param-
eter (8 entries per card if IFORM = 0 OR 4 entries per card if IFORM > 0).
2. If there is more than one real constant table in this data group repeat the same
card set sequence (card set 1 and 2) to define the additional tables.Group ID
card:
IFORM = 0 OR 4 entries per card if IFORM > 0).
7.3.5 *LAMANGLE Data Group - Rotation Angles for Laminated Composite Ele-
ments
This data group is required if the model includes any composite element (NKTP =
32). Two sets of cards are required for a typical rotation angles table.
Entry No: 1 2
variable: IDRT NUMRT
Max char:
7-18
Model Data
Entry No: 1 2 3 4 5 6 7 8
Variable: RA1 RA2 RA3 RA4 RA5 RA6 RA7 RA8
Max char: 10 10 10 10 10 10 10 10
Notes:
1. If the element has more than eight nodes, use additional cards as required.
2. Rotation angles are measured with respect to the coordinate system defined
with variable IDRA in *LAMSEQ data group.
3. If there is more than one rotation angles table in this group repeat the same
card set sequence (card set 1 and 2) to define additional tables.
7-19
Element Data
It should be noted that the input for this data group does not include a table
identification number. However, each data set entered in this group will be given an
identification number by the program, in ascending order starting with one. These
identification numbers must match with the entry MATID in the *ELEMENT data
group.
7-20
Model Data
7-21
Element Data
7-22
Model Data
Notes:
1. Each card can have a maximum of 20 integer entries. Use additional cards if
required.
2. The layer numbering sequence always starts at the top surface of an element.
Example 7.3.6 (a)
Consider a flat plate lying in the global XY plane and modeled using 8 node
laminated composite shell elements (NKTP = 32, NORDR = 2), as shown in
Figure7.3.3. The plate has 4 layers. The lamination angles, thicknesses and
material specification are shown in the figure. The lamination angles (45/0/90/-45)
are about the global Z-axis and are measured from the global X-axis (i.e., IDRA =
1). The lamination data for this problem are as follows:
7-23
Element Data
**
*LAMSEQ
**
*RCTABLE
**
1, 8
0.0055, ///////
**
2, 8
0.0075, ///////
**
7-24
Model Data
*LAMANGLE
**
1, 8
45.0, ///////
**
2, 8
0.0, ///////
**
3, 8
90.0, ///////
**
7-25
Element Data
4, 8
-45.0, ///////
**
*MATERIAL
**
EX, 1, 0, 25.0E6
**
**
EX, 2, 9.0E6
**
To understand the need for more general ways to define fiber angles, consider the
complication that arises in the problem of the previous example when the flat plate
is changed to a cylindrical shell, as shown in Figure7.3.4. The rotation angle of the
top layer is still 45o. However, it is no longer a rotation of 45o about the global Z-
axis measured from the global X-axis. It is a rotation of 45o about the normal to the
shell surface measured from the projection of the global X-axis onto the shell
surface (i.e., the projection of the global X-axis on the plane tangent to the shell
7-26
Model Data
surface at each node). Thus, IDRA may be set to 4 in the previous example to
define the lamination angles for this problem.
These examples of using IDRA = 0 in Figure7.3.5(a) and Figure7.3.5(a) are for flat
plates; the generalization to shells consists in realizing that the rotation angle is
7-27
Element Data
measured from the projection of the reference line defined above onto the tangent
plane to the shell surface at each node.
7-28
Model Data
Defining a local coordinate system which may be referenced in the *NODES data
group to specify nodal coordinates and boundary conditions in local coordinate
systems.
The free format data for a typical local coordinate system consists of one card set
which may contain one or two cards depending on the method of definition.
$
Entry No: 1 2 3 4-12
IDSYS ITYPE METHOD Variable, depends on the value
Variable:
of METHOD
Max char: 6 1 1 max. 10 characters each
7-29
Nodal Data
For METHOD = 1 (requires entries 4-9 on the first card, entries 10-12 on the
second card)
7-30
Model Data
For METHOD = 3 (requires entries 4-9 on the first card, entries 10-12 on the
second card)
7-31
Nodal Data
Notes:
1. Refer to Section 7.4.2 for an example of defining a local coordinate system.
2. If the local coordinate system is cylindrical (ITYPE = 1 on first card) and is
referenced in the *NODES data group to define nodal coordinates, then the
corresponding nodal coordinates must be cylindrical (R, ,Z). The same holds
for a local spherical coordinate system.
3. A local cylindrical (or spherical) coordinate system may be referenced in the
*NODES data group to define both coordinates and local velocity coordinate
systems at more than one node. The local velocity coordinate system directions
will be dependent on the coordinates of each node.
4. The user is not allowed to specify only a particular component of velocity in
local coordinate system. All velocity components (2 for 2-D and 3 for 3-D)
must be specified whenever a local coordinate system is used to specify bound-
ary conditions at nodes.
7-32
Model Data
ndividual definition, first level generation and second level generation. Second
level generation is activated whenever a base set, defined using individual nodes
definition and/or first level generation, is preceded by the second level generation
card as shown below. Examples for first and second level node generations are also
given below
$ $
Entry No: 1 2 3 4 5 6 7 8
Variable: NODE IDCSYS INDEX NSET X Y Z IDDSYS
Max 1
6 6 6 2 12 12 6
char: 2
7-33
Nodal Data
7-34
Model Data
7-35
Nodal Data
Notes:
1. For second level generation, the nodes in the base set may be defined individu-
ally or by using first level generation. They may not be defined using first level
generation.
2. All nodes defined using first or second level generation must have the same
local displacement coordinate system ID (IDDSYS).
Example 7.4.1: First level node generation
The six nodes shown in Figure7.4.2 are generated using first level generation.
Nodes 1 and 6 are defined in the global cartesian coordinate system. The
coordinates for the generated nodes (2, 3, 4 and 5) are linearly interpolated between
nodes 1 and 6 in the same coordinate system.
**
**Individual definition for node number 1
1, 0, 0, 0, 0.0, 0.0, 0.0
**
**Definition of node number 6. First level generation is activated between nodes 1 and 6
with node
**increment of 1
6, 0, 1, 0, 5.0, 0.0, 0.0
**
7-36
Model Data
The 18 nodes in Figure7.4.3 are generated using second level generation. The base
set is generated via first level generation (individual definition may also be used).
Three sets including the base set are generated.
**
**Define the second level generation parameters as well as the extrapolation increments for
nodal
**coordinates
-10, 0, 6, 3, 0.0, 1.0, 0.0
**
**Define nodes of the base set (nodes 1 to 6). First level generation is used
1, 0, 0, 0, 0.0, 0.0, 0.0
6, 0, 1, 0, 5.0, 0.0, 0.0
**
7-37
Nodal Data
7-38
Model Data
** nodes 7-10 defined in local system no.3, using 1st and 2nd level generation
-3, 3, 3, 2, 0.0, 2.0, 0.0, 0
7 3, 0, 0, 0.0, 0.0, 0.0, 0
9 3, 1, 0, 1.0, 0.0, 0.0, 0
**nodes 13-17 defined in local system no.4, using first level generation,
**a local displacement coordinate system, IDDSYS = 4 is specified
13, 4, 0, 0, 0.5, 0.0 0.0, 4
17 4, 1, 0, 0.5, 180.0, 0.0, 4
**
7-39
Material Data
$
Entry No: 1 2 3 4 5 6
Variable: LABEL MATID KTEMP COEF0 COEF1 COEF2
Max
4 6 1 12 12 12
char:
Entry No: 7 8
Variable: COEF3 COEF4
Max
12 12
char:
7-40
Model Data
7-41
Material Data
Notes:
1. NISA is entirely independent of any system of units. The only requirement is
that the units used must be self consistent.
2. If P = P(T) is any property given as a function of temperature T, then P is
found as follows:
4
i
P ( T ) = 0 +
i T
i=1
3. If any of the five coefficients (i.e., COEF0 through COEF4) is not specified,
then it is taken as zero. Thus, with a KTEMP value of 1, if only COEF0 is
specified, it amounts to declaring the material property to be constant.
4. If KTEMP = 0, then even if COEF 1 through COEF4 are specified, these val-
ues are ignored (only COEF0 is taken as constant property)
5. For anisotropic thermal conductivity (i.e. if executive card ANISO = ON,
TEMP is used), Kxx must be specified. If Kyy or Kzz (or both) is not specified,
the value of Kxx is picked up for the unspecified directional property. If COND
is specified for solids, Kxx, Kyy and Kzz is set equal to COND.
6. Anisotropic conductivity can be specified only for solids. For fluids if COND
along with Kxx, Kyy, and Kzz is specified, only COND is taken as conductivity
value. While, if Kxx is specified, COND is set equal to Kxx. Only Kyy or Kzz
(or Kyy and Kzz) can not be specified for fluid conductivity.
8. For transient heat transfer problems (ANALYSIS = HEAT and STDS = OFF or
no STDS card), density (DENS) and specific heat (SPEC) must be specified.
7-42
Model Data
Entry No: 1 2 3 4 5 6 7 8 9 10
Variable: Label 11 12 13 14 15 16 17 18 19
Max char: 12 6 6 6 6 6 6 6 6 6
7-43
Material Data
Notes:
1. If more than 9 sets are required to define the subset for a typical output quan-
tity,continue on additional cards. Additional cards must start with a tab ($)
character. Zero set ID numbers are ignored.
7-44
Model Data
$ $
Entry
1 2 3 4 5 6 7
No:
Variable: IDSET LABEL I1 J1 K1 I2 J2
Max
6 1 6 6 6 6 6
char:
Entry
8 9 10 11
No:
Variable: K2 I3 J3 K3
Max
6 6 6 6
char:
7-45
Miscellaneous Data
Notes:
1. LABEL must not be blank for the first card used to define the set.
2. If LABEL is left blank for any subsequent card (defining the same set), the
LABEL of the previous card is used.
3. Define as many sets as desired. Each set must start with a unique non-zero
IDSET.
4. If any set requires more than one card, continue with additional cards. Addi-
tional cards must start with a tab character $. If any additional card starts with
two consecutive tab characters $$, then the LABEL of the previous card is
used.
5. Once a set is defined, a non-zero value of IDSET on any subsequent card will
begin the definition of a new set, if any.
7-46
Model Data
Example 7.6.1:
7-47
Miscellaneous Data
Entry No: 1 2
Variable: IDCURV NPOINT
Max char: 6 2
Entry No: 1 2 3 4 5 6 7 8
Variable: T1 P1 T2 P2 T3 P3 T4 P4
Max char: 10 10 10 10 10 10 10 10
7-48
Model Data
Notes:
1. Use as many cards as necessary to define the NPOINT number of data points
(4 points per card). The data points must be entered in ascending order of tem-
perature.
2. A typical curve should cover the entire range of possible temperature values
encountered in the analysis, i.e. REFT1 T REFT2, where REFT1and
REFT2 are the reference temperatures specified in the *FLCNTL card (section
8.1.1). For temperature values outside the range of *TEMPFN table, a multi-
plier corresponding to the nearest temperature value will be used.
3. Linear interpolation is used to compute the scaling factor for a given
temperature. The exact value of a quantity of interest (e.g., conductivity) at a
given temperature is equal to the scaling factor times the base value given in
the corresponding data group (e.g., *MATFLUID).
4. The temperature axis corresponds to dimensional temperature.
5. Density cannot be specified as a function of temperature.
Two sets of cards are required for each time amplitude curve to be defined. The
first set consists of one card which defines the curve ID and the number of points
on the curve. The second set defines time and amplitude values and may consist of
more than one card if more than four points are specified on the curve.
7-49
Miscellaneous Data
Entry No: 1 2
Variable: IDCURV NPOINT
Max char: 6 2
Card set 2: more than one card if more than four points are specified
Entry No: 1 2 3 4 5 6 7 8
Variable: T1 A1 T2 A2 T3 A3 T4 A4
Max char: 10 10 10 10 10 10 10 10
Notes:
1. A typical curve should cover all possible range of time values encountered in
the analysis. For time values outside the specified range, an amplitude corre-
7-50
Model Data
sponding to the nearest time values will be used. Amplitudes within the speci-
fied time range will be linearly interpolated.
2. The total value of a quantity will be equal to the instantaneous amplitude value
times the base value of the quantity specified on its corresponding data group.
See example 1 for for details.
3. Default time amplitude curves are internally defined by two points as follows:
(0.0,0.0),(t,1.0), where t = max (1.0, max. time specified by user).
The following cards define the time amplitude curve shown in Figure7.6.1. The
curve identification number is 111 and 7 data points are specified. Analysis data
groups may reference the curve identification number to specify changes in a
specific quantity. Details of the program interpretation for incremental and total
values of the pertinent quantity are shown after the card input.
**
**Card set 1, defines curve ID and number of points.
111, 7
**Card set 2, defines the time-amplitude pairs. Two cards are required since 7 points are
specified
**(4 points per card).
0.0, 0.0, 2.0, 0.0, 5.0, 1.0, 6.0, 1.0
7.0, 0.0, 8.0, 0.0, 10.0, 0.5
**
To explain the interpretation of the program for the above curve, we assume the
following case:
7-51
Miscellaneous Data
Then, the values of the variable interpreted by the program at each step will be as
follows (see note 1 above):
Step
1 2 3 4 5 6 7 8 9 10 11
no.
Value 0.0 0.5 5.0 10.0 15.0 15.0 0.0 0.0 3.75 7.5 15.0
7-52
Model Data
Coupled variables sets may be defined using two methods: individual definition
and first level generation as shown below
$
Entry No: 1 2 3 4 5 6
Variable: LABEL NSET NODINC NODSET NODE1 NODE2
Max char: 10 10 10 10 10 10
T - temperature
2 NSET - not used in individual definition, enter zero
7-53
Miscellaneous Data
Notes:
1. Up to 10 coupled nodes may be entered on one card. The specification of the
number of coupled nodes (NODSET, entry 4) is optional since the number of
coupled nodes is counted internally. Additional cards needed to define
additional coupled nodes must start with a tab character ($). For example, if
7-54
Model Data
the U velocity is the same at nodes 1 through 11, the input may be entered as
follows:
U $ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
$ 11
Or alternative as:
U $ 1, 2, 3, 4
$ 5, 6, 7, 8
$ 9, 10, 11
2. The variable of the first node (NODE1) is the independent (master) degree of
freedom; the others are dependent degrees of freedom.
3. A dependent degree of freedom must not be referenced again (as independent
or dependent) in any subsequent coupled variable definition.
4. Dependent degrees of freedom which appear in this data group can not be con-
strained in the *BCDVAR data group.
7-55
Miscellaneous Data
7-56
Chapter
8
Analysis Data
The analysis data for fluid flow is presented in this chapter. The analysis data
generally defines the boundary conditions on the model and selects the output
options. It must follow the model data block in a typical NISA/HEAT data deck as
shown in Figure 5.1.1.
The analysis data block consists of distinct data groups. Each data group consists
of a group identification card followed by its free format data, which may consist
of one or more card sets, each card set may consist of one or more cards. The free
format rules are given in Section 5.2. Each group has a descriptive identification
name indicating the function of the data group. For example, the specified
Dirichlet boundary condition is given in the *BCDVAR data group, and the
Neumann boundary condition is given in the *BCGVAR data group.
The completion of whole analysis data block must be followed by the data deck
terminator, the *ENDDATA group ID card.
8-1
Table 8.0.1 List of analysis data groups
(1) Acceptable
minimum abbreviations are in bold face
8-2
Analysis Data
Heat transfer control card set: one card (if not provided, default values are used)
Entry No 1 2 3 4 5 6 7
Variable: NLSTP ITMAX NTPN DT EPS PENLT REFU
Max char: 4 4 4 8 8 12 8
Entry No 8 9 10
Variable: REFL REFT1 REFT2
Max char: 8 8 8
8-3
Analysis Control Data
2. For steady state problems, steady state analysis option may be used (ref.
EXECUTIVE command STDS). Total number of time steps and the parameter
NTPN should be set to 1. Maximum number of iterations to be used in
obtaining the steady state solution is specified through ITMAX. If the
convergence is not achieved in the given maximum number of iterations,
RESTART(s) may be necessary (ref. EXECUTIVE command RESTART).
Value of time step is meaningless.
3. Steady state problems may also be solved using a pseudo-transient approach.
One may set NLSTP to its maximum value (i.e., 9800) and once the solution
converges, the program will automatically stop.
4. The value of DT is fixed for all time steps.
5. The value of EPS is compared with the solution errors (ref. Section 3.4
between two time steps to check for convergence. Once the solution error goes
below the prescribed EPS, program stops automatically.
6. In case ITMAX > 1, EPS is used to check for solution error between two itera-
tions. If the error is below EPS, the computation is forwarded to the next time
step.
7. REFL is the representative length scale for the problems being analyzed. This
is used to non-dimensionalize the governing equations.
8. If TCUTOFF = ON, REFT1 and REFT2 are lower and upper cutoff tempera-
tures. The program uses them to non-dimensionalize the governing equa-
tions. If the problem is such that exact values of these limits are unknown prior
8-4
Analysis Data
8-5
Boundary Conditions
Entry No: 1 2 3 4 5 6
Variable: NODE LABEL P0 LASTND NODINC IDCRV1
Max char: 6 4 12 6 6 6
$
Entry No: 7 8 9 10-11 12
Variable: IDCRV2 LABEL LABEL ----- LABEL
Max char: 6 4 4 -- 4
8-6
Analysis Data
(i) A concentrated nodal heat source can be applied through this card, in
order to simulate internal heat generation at a node.
(ii) For a 3D model, NHS has the same unit as heat generation rate, e.g.
Watt. However, for a 2D model, it is implied that the NHS is applied
per unit width of the fluid dimension, i.e. for a 2D geometry the unit of
NHS is Watt/m.
Miscellaneous:
8-7
Boundary Conditions
Entry No: 1 2 3 4 5 6
Variable: LABEL NELID LASTEL INC IDFACE NCARDS
Max
4 6 6 6 1 2
char:
Entry No: 7 8 9 10 11
IDCRV
Variable: LFN IDCRV2 P0 IDCRV3
1
Max
3 6 6 13
char:
8-8
Analysis Data
8-9
Boundary Conditions
8-10
Analysis Data
Notes:
Heat Flux Boundary Condition:
(i) Heat flux can be specified normal to the element face (LABEL = TG,
LFN = -1) or along the global axes (LABEL = TG, LFN = 0). If LFN =
0, P1-P3 have values q x , q y and q z . If LFN = -1, P0 has a value of q n .
Heat flux, q n , is taken to be positive along the outward normal to the
element face. Therefore, if heat is to be added to the domain, q n should
be given a negative value.
(ii) Specified heat flux normal to an element face or along the global axes
become time and/or temperature dependent whenever a IDCRV1 > 0
and/or IDCRV2 > 1 is entered, respectively. The value of the heat flux
on an element face or along the global axes at any instant is the heat
flux value specified in this group times the scale factor given on the
temperature function curve times the scale factor given on the time
amplitude curve.
(i) For convective heat transfer boundary condition (LABEL = TG, LFN =
1), two values are required. P1 is the convective heat transfer coeffi-
cient ( h c ) and P2 is the ambient temperature (Tc). These values are
used to input heat flux given by q = hc (T Tc).
8-11
Boundary Conditions
8-12
Analysis Data
Heat Generation:
(i) For heat generation (LABEL = TG, LFN = 3), P0 is heat generation rate
(heat/volume) and IDFACE is ignored.
Surface Radiation:
(ii) For radiation symmetry plane (LABEL = RAD, NCARDS = -1), LFN
is set internally in the HEAT code. Value of P0 is meaningless. Radia-
tive heat flux is internally set to zero.
(iii) For radiation outlets (LABEL = RAD, NCARDS = -2), LFN is set to 0
by default in DISPLAY III. P0 is the temperature of surroundings.
Radiative heat flux at the surface is calculated internally in the pro-
gram.
(iv) If both radiation boundary condition and surface radiation cards are
present on an element face, radiation boundary condition will be
retained and the element face will be deleted from the list of radiation
surfaces.
8-13
Boundary Conditions
(v) If a surface does not interact with other surfaces in surface radiation
sense, the user is advised to make use of radiation boundary condition
(LABEL=TG, LFN=2). This will lead to reduction in disk space
requirement as well as cpu time required to solve the problem.
(vi) If view factors with obstructions are to be evaluated, then all radiation
surfaces are taken as potential obstructors. Therefore, all possible
obstructors must also be specified as radiation surfaces.
(c) If obstruction are on, use of finer mesh can enhance the accuracy
of view factor computation.
8-14
Analysis Data
(xiv) Radiation heat flux, area of radiation surfaces, radiation heat loss/gain
is written to the output file.
Miscellaneous:
(i) In the absence of a valid IDCRV1 and/or IDCRV2, this data group
degenerates to *BCGRAD.
8-15
Boundary Conditions
$
Entry No: 1 2 3 4 5 6
Variable: NODE LABEL P0 LASTND NODINC LABEL
Max char: 6 4 12 6 6 4
8-16
Analysis Data
8-17
Boundary Conditions
Entry No: 1 2 3 4 5 6 7 8 9
Variable: P1 P2 P3 P4 P5 P6 P7 P8 P9
Max char: 8 8 8 8 8 8 8 8 8
Notes:
1. For temperature gradient (TG) in 3-D problem we need three values. If LFN =
0, P1 - P3 have values TX, TY, TZ. If LFN = -1, P1 - P3 have values Tn, 0, 0.
In 2-D problem, if LFN = 0, P1-P2 have values TX, TY, and if LFN = -1, P1
- P2 have values Tn, 0.
2. In case of duplication, the last one is considered.
8-18
Analysis Data
8.2.5 *PCHANGE
This data group should be used when the phase change effects are to be considered
in the heat transfer analysis with/without fluid flow. The details of this card follow:
Entry No.: 1 2 3 4
Variable: MATID LHEAT TSOLDS TLQDS
Max Char.: 6 13 13 13
Notes:
1. All valid material IDs specified in *MATFLUID data group can be given.
2. The MATID provided under this group will be considered to be undergoing
phase change.
3. The units for this data must be consistent with the rest of the data.
4. LHEAT should be specified per unit mass.
8.2.6 *STEPSIZE
This group is used for transient problems to specify variable time step size.
8-19
Boundary Conditions
Entry No.: 1 2
Variable: NSTEP DT
Max. Char.: 6 13
Notes:
1. More than one entry may be given to get variable time step size.
2. Results are saved in 26 and 27 files for the time steps given in this card for a
minimum of 50 steps per run.
3. The program will run for the number of steps specified in the *FLCNTL card.
4. For a step number that is more than the total number of steps specified in the
*STEP card, the last given time step size will be used.
5. For a new run (Restart = 0), the user may specify more number of steps in
*STEP card than the number of steps in the *FLCNTL card. The extra infor-
mation given in *STEP card will be used for restart runs. The users may add
information to *STEP card at restart. For restarts, step number is taken to be
the sum of the step number for the current run and the total number of steps in
the previous runs.
6. Also see executive card STEP.
Example:
*STEPSIZE
10, 0.1
5, 0.2
8-20
Analysis Data
5, 0.5
10, 1.0
20, 2.0
100, 5.0
In the above example, the first 10 steps will be run using a time step size of 0.1.
The next five steps will use a step size of 0.2 and the next 5 will use a step size of
0.5 and so on.0
8-21
Output Control Data
Entry No: 1 2 3 4 5 6 7 8 9 10
Variable: LABEL I1 I2 I3 I4 I5 I6 I7 I8 I9
Max. Char: 12 6 6 6 6 6 6 6 6 6
8-22
Analysis Data
8-23
Data Terminator
The group ID card must be the last card in the data deck.
8-24
Chapter
9
Hints for Using NISA/HEAT
9.1 Introduction
This chapter gives useful hints for running NISA/HEAT. Both modeling and
analysis hints are discussed in this chapter. Modeling hints give information that a
HEAT user should know about DISPLAY III. The analysis hints cover the NISA/
HEAT code itself. In addition, it is also shown how users can run jobs in the
background, save screen messages, monitor convergence etc.
The code prompts the user to type in input nisa file name and the output file name.
When the job is done, the user gets back the UNIX prompt. Users can run their
jobs in the background. For example, to run a nisa file named test.nis, user should
type in the following commands in a file, say nisa.go.
heat3>>run.log<<!
test.nis
test.out
y
!
9-1
Running NISA/HEAT
chmod +x nisa.go
nisa.go &
All the screen messages get directed to the file run.log. One can use an editor to
view this file, use page command or use tail command. The tail command is
particularly very useful. Type
tail -f run.log
This command will update the contents of the file run.log on screen. One can use
CTRL-C to kill the above tail command. This, however, will not kill the fluids run.
One can also submit a job to be run at a later time using at command. For example,
if one wants to submit batch file nisa.go at 22:00 hrs, type the following at the
UNIX prompt.
at 22:00
nisa.go
CTRL-D
The job will start at 10 p.m. the same day. Use
at -l
to see different jobs that are queued to run at a later time. Use the following
command to remove a job from the queue.
at -r jobid
where jobid corresponds to job one wants to remove from queue and the id can be
obtained by using at -l command.
9-2
Hints for Using NISA/HEAT
Type
mklog
to grep the errors in run.log file and direct them to a file named err. The user can
edit this file to view the iteration errors.
EXECUTION = CHECK
The user may not have all the required disk space available on one disk. However,
the user may have two or more disks available to him and total free space on those
disks may be more than the required disk space to run the problem. Under such
situations, the users are encouraged to make use of executive commands PATH or
PAT1/PAT2/PAT3/PAT4 (see Section 6.2) in their nisa file. These executive
commands can be used to direct temporary files to different directories. Users
should note that files 30 and 31 take up most of the disk space.
9-3
Running NISA/HEAT
% kill -9 pid
Note that if a NISA/HEAT job is killed using kill command or using CTRL-C, one
may not be able to post-process the results. If one wishes to post-process results
and at the same time kill the job, one should create a file stopnisa.dat in the
working directory. If the file is present, NISA/HEAT will exit normally save all the
results for post-processing and restarts.
9-4
Hints for Using NISA/HEAT
In the following some guidelines for solving a heat transfer problem using finite
element are given. These are very general in nature and should not be interpreted as
strict rules. However, they should help the user in making a good model the very
first time.
9-5
Modeling Hints
that only one set of boundary conditions can be applied at one time to a given
node. Hence suitable choices must be made. Section 9.3.2 shows such an
example.
In NISA/HEAT if similar boundary conditions are applied more than once
then only the last one will be retained.
For axisymmetric models the X-axis of the modelling plane is to be treated as
the axis of symmetry and the model should be contained in the first quadrant
of the X-Y plane. (also refer to the description of NKTP = 3 in Chapter 4).
The users may also use Y-axis as the axis of symmetry using an executive
command; DIME = AX, Y.
In a typical heat transfer problem, appropriate boundary conditions are to be
prescribed on all the boundaries, either using the Dirichlet type (i.e. nodal
temperature etc.) or Neumann type (i.e. heat transfer etc.) of boundary condi-
tions. If no boundary condition is specified, normal gradient of temperature is
assumed zero; this is implicit in the formulation (ref. Chapter 2).
9-6
Hints for Using NISA/HEAT
Figure 9.3.1 Discretization of Triangular and circular domains using only Quadri-
laterals
9-7
Modeling Hints
9-8
Hints for Using NISA/HEAT
9-9
Analysis Hints
REFT2 specified by the user in the *FLCNTL card. In such cases, REFT2
should be set to a higher value.
4. If in the executive deck, TCUTOFF = ON, the lower cut off temperature,
REFT1, should also be chosen carefully. It should be the minimum of the pre-
scribed temperature boundary condition, ambient temperature (in convective
boundary condition) and the temperature of surroundings (in radiation bound-
ary condition). Note that minimum temperature in the domain may still be
lower if some heat sink is present in the domain.
Steady State Problems
Surface Radiation
1. The user is advised to read the surface radiation theory (Section 2.5). Surface
radiation analysis can be controlled through executive cards SRAD and
UFIJ. These cards are described in Section 6.2. The radiation surfaces are spec-
ified in *BCGVAR card (LABEL = RAD). The user is strongly urged to read
the notes given in Section 8.2.2 for surface radiation.
2. NISA/HEAT can internally compute the view factor matrix for the radiation
surfaces defined in the *BCGVAR card. The user can, however, also supply
externally computed view factor matrix to be read NISA/HEAT from an exter-
nal file by setting IUSRVF = 1 in the executive card (Section 6.2).
UFIJ = IUSRVF, IPRVF
Numbering radiation surfaces (for user defined view factors):
Surfaces are assigned numbers 1 through N in the order they appear in the nisa
file. Relevant data such as Stefan-Boltzman constant, surface emissivity and
T are given in the nisa file along with radiation surfaces.
9-10
Hints for Using NISA/HEAT
FORTRAN subroutine can be used to create view factor input file vfac-
nisa.vfc for NISA/HEAT.
SUBROUTINE VFNISA (NSIZE, NSURF, FIJ)
IMPLICIT NONE
INTEGER*4 NSIZE, NSURF, I, J
REAL*8 FIJ
DIMENSION FIJ (NSIZE, NSIZE)
OPEN (UNIT = 21, FILE = vfacnisa.dat, STATUS = UNKNOWN)
DO 1000 I = 1, NSURF
WRITE (21,2000) (FIJ (I, J), J = 1, NSURF)
2000 FORMAT (5 (2X, E13.6))
1000 CONTINUE
CLOSE 21
RETURN
END
Any other format than the one given in the above subroutine will cause reading
error in NISA/HEAT.
Special Note:
All of the user defined surfaces must be valid surfaces. See Section 8.2.2 for
the definition of invalid surfaces. If any invalid surface is found in the nisa file,
the program will stop. the list of the invalid surfaces can be found in the output
file. The user must remove the invalid surfaces from the list of radiation sur-
faces and supply new view factor matrix in a data file vfacnisa.dat.
Transient Problems
1. For transient problems, if the time step errors oscillate a lot or diverge,
the time step should be reduced. In addition, number of iterations within
each time step may have to be increased.
9-11
Analysis Hints
9-12
Appendix
A
Appendix A: Verification Problem
Title:
One dimensional heat conduction in a bar
Element Type:
3-D eight node solid element (NKTP = 4, NORDR = 1)
Analysis Highlights:
Problem:
A 1.0 m long bar with a square cross section (0.2 m x 0.2 m) is subjected to a flux
of 1000 W/m2 at one end. The other end of the bar is fixed at 0C. All other
surfaces are insulated.
A-1
Appendix A: Verification Problem
Properties:
Material:
The bar is modeled using five equal size 8 node solid elements (Figure A.1.1). The
flux load is specified as a surface flux on the element face at X = 1.0 m. The nodes
at X = 0 are fixed at 0C temperature.
The nodal temperatures along the bar axis are listed in Table A.1.1. The
temperature varies linearly from 5C (at X = 1m) to 0C (at X = 0) along the length
of the bar. Using one dimensional steady state heat conduction equation [1], the
theoretical solution is T = 5X. Results obtained by NISA compare exactly.
Reference:
1. J. P. Holman, Heat Transfer, McGraw-Hill Book Company, Fifth edition
(1981).
A-2
Appendix A: Verification Problem
A-3
Appendix A: Verification Problem
A-4
Appendix A: Verification Problem
4, 1, 1, 0, 0
4, 5, 35, 34, 14, 15, 25, 24,
5, 1, 1, 0, 0
5, 6, 36, 35, 15, 16, 26, 25,
*MATFLUID
DENS, 1,0, 1.00000E+00,
COND, 1,0, 2.00000E+02,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID = 1
** last two entries are lower and upper
** cut-off temperatures
1,1,1,1.0,0.001,0.1E+09,1.0,1.0,0.0,10.0
*BCDVAR
** BCDVAR SET = 1
** specified temperature boundary condition
1,T , 0.00000E+00,,, 0, 0
11,T , 0.00000E+00,,, 0, 0
21,T , 0.00000E+00,,, 0, 0
31,T , 0.00000E+00,,, 0, 0
*BCGVAR
** BCGVAR SET = 1
** heat flux normal to element face
** negative value indicates heat being added to the
** system
TG , 5,,,4, 0, -1, 0, 0, -0.10000E+04
*ENDDATA
A-5
Appendix A: Verification Problem
Title:
Heat conduction in a square plate with prescribed temperature
Element Type:
2-D four node planar element (NKTP = 2, NORDR = 1)
Analysis Highlights:
Linear steady state heat transfer analysis with prescribed temperature conditions.
Problem:
A square plate (1.0 in x 1.0 in) with unit thickness has one side at 100F
temperature. Other three sides are at 0F temperature. All four sides and the two
faces of the plate are insulated.
Properties:
Material:
A-6
Appendix A: Verification Problem
( 2n + 1 )x ( b y ) ( 2n + 1 )
sin --------------------------- sinh -----------------------------------------
4T b 1 a a
T ( x, y ) = ---------
-------------------- ----------------------------------------------------------------------------------------
( 2n + 1 ) ( 2n + 1 )b
sinh ---------------------------
n=0 a
where Tb (= 100F) is the temperature at Y = 0.0 boundary and 'a' (= b = 1.0 inch)
is the size of the square plate. Isothermal contour lines are shown in Figure A.2.2
and the temperature for the nodes at X = 0.5 inch are compared in Figure A.2.3.
Excellent agreement between NISA and theoretical results is evident.
Reference:
1. H. S. Carslaw, J. C. Jaeger, "Conduction of Heat in Solids," Second Edi-
tion, Oxford University Press (1959).
A-7
Appendix A: Verification Problem
A-8
Appendix A: Verification Problem
A-9
Appendix A: Verification Problem
A-10
Appendix A: Verification Problem
A-11
Appendix A: Verification Problem
64, 1, 1, 0, 0
71, 72, 81, 80,
*MATFLUID
** material properties
DENS, 1,0, 1.00000E+00,
COND, 1,0, 1.00000E+00,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID = 1
** heat control parameters
1,1,1,1.0,0.001,0.1E+09,1.0,1.0,0.0,100.0
*BCDVAR
** BCDVAR SET = 1
** specified temperature boundary condition
l,T , 0.00000E+00,,,0, 0
9,T , 0.00000E+00,,,0, 0
10,T , 0.00000E+00,,,0 ,0
18,T , 0.00000E+00,,,0, 0
.. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. ..
79,T , 0.00000E+00,,,0, 0
80,T , 0.00000E+00,,,0, 0
81,T , 0.00000E+00,,,0, 0
** BCDVAR SET = 2
** specified temperature boundary condition
2,T , 1.00000E+02,,, 0, 0
3,T , 1.00000E+02,,, 0, 0
4,T , 1.00000E+02,,, 0, 0
5,T , 1.00000E+02,,, 0, 0
6,T , 1.00000E+02,,, 0, 0
7,T , 1.00000E+02,,, 0, 0
8,T , 1.00000E+02,,, 0, 0
*ENDDATA
A-12
Appendix A: Verification Problem
Title:
Heat conduction in a long hollow cylinder with prescribed temperature
Element Type:
Axisymmetric four node solid elements (NKTP = 3, NORDR = 1)
Analysis Highlights:
Linear steady state heat transfer analysis with prescribed temperatures.
Problem:
A long hollow cylinder is subjected to an inside temperature of 100F and an
outside surface temperature of OF. The cylinder is insulated to prevent axial heat
flow. The inner and outer radii are 1 and 2 inches, respectively, see Figure A.3.1.
The steady state temperature distribution along the radial direction is calculated
and compared with theoretical solution.
Properties:
Material:
A-13
Appendix A: Verification Problem
Ti T0 r
T ( r ) = T i ----------------- ln ---
ro r i
ln ----
ri
where Ti (= 100F) and To (= 0F) are temperatures at the inner and the outer radii,
respectively.
The calculated temperature distribution along the radius is compared with the
theoretical values in Table A.3.1. An excellent agreement with the theoretical
values is achieved.
Reference:
1. J. P. Holman, "Heat Transfer," McGraw-Hill Book Company,Fifth edition
(1 963).
A-14
Appendix A: Verification Problem
A-15
Appendix A: Verification Problem
20,,,,1.90000E+00,5.00000E-01, 0.00000E+00,0
21,,,,2.00000E+00,0.00000E+00, 0.00000E+00,0
22,,,,2.00000E+00,5.00000E-01, 0.00000E+00,0
*ELEMENTS
1, 1, 1, 0, 0
1, 3, 4, 2,
2, 1, 1, 0, 0
3, 5, 6, 4,
.. .. .. ..
.. .. .. ..
A-16
Appendix A: Verification Problem
9, 1, 1, 0, 0
17, 19, 20, 18,
10, 1, 1, 0, 0
19, 21, 22, 20,
*MATFLUID
DENS, 1,0, 1.00000E+00,
COND, 1,0, 1.00000E+02,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID = 1
1,1,1,1.0,0.001,0.1E+09,1.0,1.0,0.0,100.0
*BCDVAR
** BCDVAR SET = 1
**specified temperature boundary condition
l,T , 1.00000E+02,,, 0, 0
2,T , 1.00000E+02,,, 0, 0
** BCDVAR SET = 2
**specified temperature boundary condition
21,T , 0.00000E+00,,, 0, 0
22,T , 0.00000E+00,,, 0, 0
*ENDDATA
A-17
Appendix A: Verification Problem
Title:
Column with varying convective boundary condition on its faces
Element Type:
2-D four node planar elements (NKTP = 2, NORDR = 1)
Analysis Highlights:
Linear steady state heat transfer analysis with convection boundary condition.
Problem:
Properties:
Material:
Only one half the column cross section is modelled. The area in which the
convective coefficient is not constant has more elements. Totally 96 four node
elements are used in the mesh (Figure A.4.2). The ambient temperature and the
convective film coefficients are specified at the appropriate element faces (edges).
A-18
Appendix A: Verification Problem
The center line temperature over the width of the column (from X = 0 to X = 36 in)
is plotted in Figure A.4.3. The results are compared with the results of a Fourier
analysis solution by Peavy [I] where good agreement is shown.
Reference:
A-19
Appendix A: Verification Problem
Figure A.4.2 Finite Element Mesh with Node and Element Numbers
A-20
Appendix A: Verification Problem
A-21
Appendix A: Verification Problem
.. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
A-22
Appendix A: Verification Problem
95, 1, 1, 0, 0
74, 75, 100, 99,
96, 1, 1, 0, 0
99, 100, 125, 124,
*MATFLUID
DENS, 1,0, 1.00000E+00,
COND, l,O, 8.333338-02,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID- 1
1,1,1,1.0,0.001,0.1E+09,1.0,1.0,0.0,100.0
*BCGVAR
** convection boundary conditions
TG, 1, 4, 1, 4,1,1,0,0
3.47220E-03 1.00000E+02 0.00000E+00
TG, 1, 21, 4, 1,1,1,0,0
3.47220E-03 1.00000E+02 0.00000E+00
TG, 93, 93, 1, 1,1,1,0,0
4.16667E-02 0.00000E+00 0.00000E+00
TG, 93, 96, 1, 2,1,1,0,0
4.16667E-02 0.00000E+00 0.00000E+00
*ENDDATA
A-23
Appendix A: Verification Problem
Title:
Heat conduction analysis in an infinitely long slab
Element Type:
2-D four node planar element (NKTP = 2, NORDR = 1)
Analysis Highlights:
Transient heat transfer analysis with specified initial temperature.
Problem:
Properties:
Material:
Since the geometry is infinite in Z-direction, any cross-section of the structure will
have the same temperature distribution.
A-24
Appendix A: Verification Problem
Twenty 4-noded plane heat transfer elements (NKTP = 2, NORDR = 1) are used to
model one such cross-section, see Figure A.5.1. The initial temperature of the
structure is specified as 100F and the end nodes (1, 31, 21 and 51) have a
prescribed temperature value of 0F.
Solution Procedure:
Reference:
A-25
Appendix A: Verification Problem
Figure A.5.1 Slab Cross Selection Model with Node and Element Numbers
A-26
Appendix A: Verification Problem
A-27
Appendix A: Verification Problem
.. .. .. .. ..
19, 1, 1, 0, 0
19, 20, 50, 49,
20, 1, 1, 0, 0
20, 21, 51, 50,
*MATFLUID
DENS, 1,0, 1.00000E+00,
COND, 1,0, 1.00000E+00,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID= 1
20,1,20,1.0,0.001,0.1E+09,1.0,1.0,0.0,100.0
*BCDVAR
** BCDVAR SET = 1
l,T , 0.00000E+00,,, 0,0
21,T , 0.00000E+00,,, 0,0
31,T , 0.00000E+00,,, 0,0
51,T , 0.00000E+00,,, 0,0
*ENDDATA
A-28
Appendix A: Verification Problem
Title:
Transient heat conduction through a wall
Element Type:
2-D four node planar element (NKTP = 2, NORDR = 1)
Analysis Highlights:
Nonlinear transient heat transfer analysis with temperature dependent material
properties and time dependent specified temperatures
Problem:
An infinitely long slab with 20 cm x 1.0 cm cross section is at an initial
temperature of 100C. The left and right end faces have time dependent prescribed
temperatures. The thermal conductivity is a polynomial function of temperature
whereas the specific heat and the density are constant.
Properties:
Material:
A-29
Appendix A: Verification Problem
Since the geometry is infinite in Z-direction, any cross-section of the structure will
have the same temperature distribution.
Twenty 4-node plane heat transfer elements (NKTP = 1, NORDR = 1) are used to
model one such cross-section (Figure A.6.1). The end nodes (1, 2, 41, and 42) have
a prescribed temperature values which are time dependent (Figure A.6.1). The
initial temperature of the structure is specified to be 100C.
Solution Procedure:
Reference:
A-30
Appendix A: Verification Problem
A-31
Appendix A: Verification Problem
A-32
Appendix A: Verification Problem
A-33
Appendix A: Verification Problem
A-34
Appendix A: Verification Problem
Title:
Transient heat conduction in a stainless steel cylinder
Element Type:
Axisymmetric eight node solid elements (NKTP = 3, NORDR = 2)
Analysis Highlights:
Transient heat transfer analysis with prescribed initial temperature.
Problem:
Properties:
Material
KXX = 17.4 W/m.OK (Thermal conductivity)
C = 526 J/Kg.OK (Specific heat)
DENS = 7900 Kg/m3 (Density)
Finite Element Model:
The axisymmetric model of the cylinder is shown in Figure A.7.2, and is composed
of twenty four (4 x 6) eight node axisymmetric solid elements, (NKTP = 103 and
A-35
Appendix A: Verification Problem
NORDR = 2).
The initial temperature of the structure is specified as 600K and the ambient
temperature is maintained at 300K. Convection boundary conditions are
prescribed on the faces X = 40 mm, Y = 0, and Y = 60 mm.
Solution Procedure:
Reference:
A-36
Appendix A: Verification Problem
A-37
Appendix A: Verification Problem
Figure A.7.2 Finite Element Mesh with Element and Node Numbers
A-38
Appendix A: Verification Problem
A-39
Appendix A: Verification Problem
A-40
Appendix A: Verification Problem
TG , 24,,,3, 1, 1, 0, 0,
0.50000E+03, 0.30000E+03, 0.00000E+00
TG , 24,,,3, 1, 1, 0, 0,
0.50000E+03, 0.30000E+03, 0.00000E+00
*ENDDATA
A-41
Appendix A: Verification Problem
Title:
Heat conduction in a convectively cooled orthotropic bar
Element Type:
2-D eight node planar elements (NKTP = 2, NORDR = 2)
Analysis Highlights:
Transient heat transfer analysis in an orthotropic material
Problem:
Properties:
Material:
KXX = 20 BTU/hr.ft.0F (Thermal conductivity in X-direction)
KYY = 3.6036 BTU/hr.ft.0F (Thermal conductivity in Y-direction)
C = 0.00909 BTU/lbm.0F (Specific heat)
DENS = 400 lbm/ft3 (Density)
A-42
Appendix A: Verification Problem
Since the bar has infinite length in Z-direction, any cross section will have the same
temperature distribution. Also the geometry and the boundary conditions of the
cross section are symmetric. Hence only one quarter of the bar cross section is
modelled using 36 (6 x 6) 8-node planar elements, see Figure A.8.1. The initial
temperature is specified as 500F. Convective boundary conditions are imposed on
the element faces (edges) which are in contact with the liquid, i.e. at X = 0.1667 ft.
(2 in) and at Y = 0.0833 ft. (1 in).
Solution Procedure:
Thirty equal time steps of 2.77778 x 10-5 hr. (0.1 sec.) are used with Crank-
Nicholson time integration scheme.
Temperature at time = 8.333 x 10-4hr. (3 sec.) for the four nodes (1, 13, 157 and
169) are tabulated in Table A.8.1. These temperatures are compared with the
analytical solution [1] and results are in very good agreement.
Reference:
A-43
Appendix A: Verification Problem
A-44
Appendix A: Verification Problem
A-45
Appendix A: Verification Problem
A-46
Appendix A: Verification Problem
Title:
Electric potential distribution inside a void (Analogous field problem)
Element Type
Analysis Highlights:
Linear steady state potential flow analysis with prescribed potentials
Problem:
A rectangular void of dimensions (2m x lm) in a grounded conductor is biased to a
sinusoidal voltage of 10 sin y as shown in Figure A.9.1. The potential
distribution inside the void is computed.
Properties:
Material = Isotropic
Kxx = (Dielectric permittivity of air) = 8.854E-12 (F/m)
Finite Element Model:
Due to the symmetry of the problem, the upper half of the void is considered to be
the analysis domain. The calculation domain is modeled with 2D-linear quadratic
elements (NKTP=102, NORDR = l). Nodes on the line x = 2m are fixed at voltages
given by 10 sin y and the nodes on lines x = 0 m, y = 0.5 and y = 1.0 m are kept
at 0 volts because of the conductor being grounded. Specifying voltages here is
analogous to specifying temperatures in a heat transfer analysis run.
A-47
Appendix A: Verification Problem
References:
A-48
Appendix A: Verification Problem
A-49
Appendix A: Verification Problem
A-50
Appendix A: Verification Problem
A-51
Appendix A: Verification Problem
Title:
Magnetostatic Potential Distribution inside a Coaxial Cable (Analogous Field
Problem)
Element Type:
2-D Linear quadrilateral elements (NKTP=2, NORDR= 1)
Analysis Highlights:
Linear steady state potential flow analysis with prescribed flux conditions.
A-52
Appendix A: Verification Problem
Problem:
A coaxial cable carrying 10A, has a hollow cylindrical region with an inner radius
of 1 m and an outer radius of 2 m, between the conductors. Magnetic scalar
potential distribution is computed in this hollow region between the inner and the
outer conductors (Figure A.11.1)
Properties:
Material: Isotropic
Due to the symmetry of the problem a quarter of the model is considered in global
cylindrical coordinates. Along the line = 0 the magnetic scalar potential is set
o I
B = ----------
- (A.11)
2R
Where o is the permeability of air, I is the current through the cable, and R is the
radial location of any point in the calculation domain. The results obtained by
A-53
Appendix A: Verification Problem
NISA, given in the Table A.11.1, are in excellent agreement with those obtained
using the analytical expressions (Figure. 10.12.2 ).
References:
1. J. D. Jackson, Classical electrodynamics, John Wiley and sons, Inc., (1967)
2. M. A. Plonus, Applied electromagnetics, McGraw-Hill Book Company
(1978
A-54
Appendix A: Verification Problem
A-55
Appendix A: Verification Problem
A-56
Appendix A: Verification Problem
A-57
Appendix A: Verification Problem
A-58
Appendix A: Verification Problem
Title:
Combined surface radiation and conduction heat transfer
Element Type:
2-D four noded linear quadrilateral element (NKTP=2, NORDR=1)
Analysis Highlights:
Surface radiation analysis with view factor computation.
Problem:
A simple example of conduction heat transfer with surface radiation effects is
solved to illustrate surface radiation heat transfer capability of NISA/HEAT. Figure
A.11.1 shows schematic of the problem.Surfaces AB and CD are held at
temperatures Th (= 500K) and TC (= 300K), respectively. Surface BC exchanges
radiation energy with itself and with the surroundings that are at a temperature T
(= 10K). Emissivity of surface BC is 1.0. Surface AD is adiabatic and is not a
radiating surface. Temperature distribution in the domain shown in Figure A.12.1
is calculated.
Properties:
Material: Isotropic
A-59
Appendix A: Verification Problem
Solution Procedure:
View factors between the radiating surfaces are computed first. Radiative heat flux
which takes into account radiation energy exchange among the surfaces and with
the surroundings is calculated using Equation (2.6.39). Equation (2.6.39) is nothing
but conservation of radiative energy. Note that a temperature field must be
available, if radiative flux is to be evaluated. On the other hand, radiative flux
affects temperature distribution. This leads to coupling between solution of
temperature and radiative fluxes. NISA/HEAT solves them in a sequential manner
in an iterative loop. Solution is continued until norm for radiative flux falls below a
user specified tolerance for a tolerance of 0.001, 22 iterations were required for
convergence.
Figure A.12.3 shows the temperature distribution for the problem described above.
For the sake of comparison, the same problem was run using radiation boundary
condition (*RADBC car) instead of surface radiation (*RADSURFACE card).
Note that the use of radiation boundary conditions in this case will amount to
neglecting surface radiation exchange among the surfaces. Only the radiation
exchange with the surroundings is taken into account. Temperature distribution for
this case is shown in Figure A.12.4.
Note that the two temperature distributions are different from each other. One can
conclude that one must use surface radiation instead of radiative boundary
condition in situations where surface to surface radiation exchange is significant.
A-60
Appendix A: Verification Problem
A-61
Appendix A: Verification Problem
A-62
Appendix A: Verification Problem
A-63
Appendix A: Verification Problem
.. .. .. .. ..
63, 1, 1, 1, 0
93, 98, 99, 94,
64, 1, 1, 1, 0
94, 99, 100, 95,
*MATFLUID
DENS, 1,0, 1.00000E+00,
KXX , 1,0, 5.00000E+01,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID= 1
1,50,1,1.0,0.001,0.1E+09,1.0,1.0,10.0,500.0
*BCDVAR
** BCDVAR SET = 1
1,T , 3.00000E+02,,, 0, 0
2,T , 3.00000E+02,,, 0, 0
3,T , 3.00000E+02,,, 0, 0
4,T , 3.00000E+02,,, 0, 0
5,T , 3.00000E+02,,, 0, 0
** BCDVAR SET = 2
96,T , 5.00000E+02,,, 0, 0
97,T , 5.00000E+02,,, 0, 0
98,T , 5.00000E+02,,, 0, 0
99,T , 5.00000E+02,,, 0, 0
100,T , 5.00000E+02,,, 0, 0
*BCGVAR
** BCGVAR SET = 1
** surface radiation
RAD , 1,,,1, 1, 1, 0, 0,
0.56690E-07, 0.10000E+02, 0.10000E+01
RAD , 5,,,1, 1, 1, 0, 0,
0.56690E-07, 0.10000E+02, 0.10000E+01
.. .. .. .. .. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. .. .. .. ..
RAD , 57,,,1, 1, 1, 0, 0,
0.56690E-07, 0.10000E+02, 0.10000E+01
RAD , 61,,,1, 1, 1, 0, 0,
A-64
Appendix A: Verification Problem
A-65
Appendix A: Verification Problem
63, 1, 1, 1, 0
93, 98, 99, 94,
64, 1, 1, 1, 0
94, 99, 100, 95,
*MATFLUID
DENS, 1,0, 1.00000E+00,
KXX , 1,0, 5.00000E+01,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID= 1
1,50,1,1.0,0.001,0.1E+09,1.0,1.0,10.0,500.0
*BCDVAR
** BCDVAR SET = 1
1,T , 3.00000E+02,,, 0, 0
2,T , 3.00000E+02,,, 0, 0
3,T , 3.00000E+02,,, 0, 0
4,T , 3.00000E+02,,, 0, 0
5,T , 3.00000E+02,,, 0, 0
** BCDVAR SET = 2
96,T , 5.00000E+02,,, 0, 0
97,T , 5.00000E+02,,, 0, 0
98,T , 5.00000E+02,,, 0, 0
99,T , 5.00000E+02,,, 0, 0
100,T , 5.00000E+02,,, 0, 0
*BCGVAR
** BCGVAR SET = 1
TG , 1,,,1, 1, 2, 0, 0,
0.56690E-07, 0.10000E+02, 0.10000E+01
TG , 5,,,1, 1, 2, 0, 0,
0.56690E-07, 0.10000E+02, 0.10000E+01
.. .. .. .. .. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. .. .. .. ..
TG , 57,,,1, 1, 2, 0, 0,
0.56690E-07, 0.10000E+02, 0.10000E+01
TG , 61,,,1, 1, 2, 0, 0,
0.56690E-07, 0.10000E+02, 0.10000E+01
*ENDDATA
A-66
Appendix A: Verification Problem
Title:
Heat conduction in a 3-D slender bar with convective and specified heat flux
boundary condition at the ends.
Element Type:
3-D four node tetrahedral elements (NKTP = 4, NORDR = 20) and 2nd order ten
node tetrahedral elements (NKTP = 4, NORDR = 21).
Analysis Highlights:
Tetrahedral elements
Problem:
A slender bar of square cross-section (1m x 1m) is subjected to a convective
boundary condition at one end and specified heat flux boundary condition at the
other end. The sides of the slender bar is insulated. The specified heat flux value is
100 W/m2, while the convective heat transfer coefficient is 10W/m2-k and the
ambient temperature is 30C.
Properties:
KXX = 1.0 W/m-C
A-67
Appendix A: Verification Problem
Select hyperpatch no. and face no. and apply the boundary condition
[element face is immaterial]
z 1
T = q --- + --- + T
k h
A-68
Appendix A: Verification Problem
A-69
Appendix A: Verification Problem
A-70
Appendix A: Verification Problem
A-71
Appendix A: Verification Problem
A-72
Appendix A: Verification Problem
A-73
Appendix A: Verification Problem
A-74
Appendix A: Verification Problem
A-75
Appendix A: Verification Problem
799, 1, 1, 1, 0
249, 269, 245, 275,
800, 1, 1, 1, 0
269, 270, 245, 275,
*MATFLUID
DENS, 1,0, 1.00000E+00,
KXX , 1,0, 1.00000E+00,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID= 1
1,1,1,1.0,0.001,0.1E+09,1.0,1.0,0.0,1500.0
*BCGVAR
** BCGVAR SET-ID = 1
TG, 723, 723, 1, 3,0,-1,0,0 -1.00000E+02
TG, 725, 725, 1, 2,0,-1,0,0 -1.00000E+02
TG, 728, 728, 1, 3,0,-1,0,0 -1.00000E+02
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
TG, 795, 795, 1, 3,0,-1,0,0 -1.00000E+02
TG, 798, 798, 1, 3,0,-1,0,0 -1.00000E+02
TG, 800, 800, 1, 2,0,-1,0,0 -1.00000E+02
** BCGVAR SET-ID = 2
TG, 1, 1, 1, 1,1,1,0,0
1.00000E+01 3.00000E+01 0.00000E+00
TG, 2, 2, 1, 1,1,1,0,0
1.00000E+01 3.00000E+01 0.00000E+00
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
TG, 76, 76, 1, 1,1,1,0,0
1.00000E+01 3.00000E+01 0.00000E+00
TG, 77, 77, 1, 1,1,1,0,0
1.00000E+01 3.00000E+01 0.00000E+00
*ENDDATA
**EXECUTIVE data deck for HEAT
ANALYSIS = HEAT
DIMENSION = 3D
ELEMENT_ECHO = OFF
A-76
Appendix A: Verification Problem
FILE = ht12b
SAVE=26,27,54
NODEECHO = OFF
STDS = ON,1.0,1.0,1.0
**SOLU=ITERATIVE,600,0.1E-03,1
*TITLE
Tetrahedral Element (Pattern No. = 2, Order = 1).
Convective Boundary Condition at One End, Dflux at the Other.
*ELTYPE
1, 4, 20
*NODES
1,,,, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0
2,,,, 2.50000E-01, 0.00000E+00, 0.00000E+00, 0
3,,,, 5.00000E-01, 0.00000E+00, 0.00000E+00, 0
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
273,,,, 5.00000E-01, 1.00000E+00, 1.00000E+01, 0
274,,,, 7.50000E-01, 1.00000E+00, 1.00000E+01, 0
275,,,, 1.00000E+00, 1.00000E+00, 1.00000E+01, 0
*ELEMENTS
1, 1, 1, 1, 0
6, 1, 2, 26,
2, 1, 1, 1, 0
6, 2, 7, 32,
.. .. .. .. ..
.. .. .. .. ..
799, 1, 1, 1, 0
244, 250, 274, 270,
800, 1, 1, 1, 0
250, 274, 270, 275,
*MATFLUID
DENS, 1,0, 1.00000E+00,
KXX , 1,0, 1.00000E+00,
A-77
Appendix A: Verification Problem
A-78
Appendix A: Verification Problem
1, 4, 21
*NODES
1,,,, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0
2,,,, 2.50000E-01, 0.00000E+00, 0.00000E+00, 0
3,,,, 5.00000E-01, 0.00000E+00, 0.00000E+00, 0
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
1539,,,, 3.75000E-01, 1.00000E+00, 9.50000E+00, 0
1540,,,, 6.25000E-01, 1.00000E+00, 9.50000E+00, 0
1541,,,, 8.75000E-01, 1.00000E+00, 9.50000E+00, 0
*ELEMENTS
1, 1, 1, 1, 0
1, 276, 2, 500, 7, 966, 1342, 717, 1143, 27,
2, 1, 1, 1, 0
134
1, 966, 7, 280, 6, 496, 1142, 721, 31,
6,
.. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. ..
799, 1, 1, 1, 0
112 114
249, 1340, 269, 1537, 245, 1541, 1341, 275,
5, 1,
800, 1, 1, 1, 0
153
269, 491, 270, 960, 245, 1141, 715, 1341, 275,
7,
*MATFLUID
DENS, 1,0, 1.00000E+00,
KXX , 1,0, 1.00000E+00,
SPEC, 1,0, 1.00000E+00,
*FLCNTL, ID= 1
1,1,1,1.0,0.001,0.1E+09,1.0,1.0,0.0,1500.0
*BCGVAR
** BCGVAR SET-ID = 1
TG, 723, 723, 1, 3,0,-1,0,0 -1.00000E+02
A-79
Appendix A: Verification Problem
A-80
Appendix A: Verification Problem
Title:
Solidification of molten metal in a cylinder
Element Type:
Axisymmetric quadrilateral element (NKTP = 3, NORDR = 1)
Analysis Highlights:
Phase change capability using temperature dependent material property.
Problem:
The Solidification of molten alloy in a cylinder of diameter D and height H1
followed by an another material of diameter D and height H2 is considered (Figure
A.14.1). The bottom molten metal is kept at an initial temperature of TI whose
material property is temperature dependent. The top metal is kept at an initial
temperature of T2. The top surface of the top metal is exposed to atmosphere.
Properties:
D = 2.0 H1 = 5.0
H2 = 5.0 1 = 2.7
A-81
Appendix A: Verification Problem
Solution Procedure:
Transient analysis is performed using Crank-Nicholson time integration scheme.
Equal time step sizes of 3 seconds are used until a time of 2 minutes 30 seconds (1
50 seconds) is reached.
Results:
Temperature contours shows that after 150 seconds the hot molten metal reaches to
maximum temperature of 41 1.70 C. Graphs are also plotted at selected node nos. to
show how temperature at different location is varying with time.
A-82
Appendix A: Verification Problem
A-83
Appendix A: Verification Problem
A-84
Appendix A: Verification Problem
A-85
Appendix A: Verification Problem
A-86
Appendix A: Verification Problem
A-87
Appendix A: Verification Problem
A-88
Appendix
B
NISA/HEAT to NISA/HEAT3 Translator
Appendix B
NISA/HEAT NISA/HEAT3
BLANK This card is not required in HEAT3
EXECUTION EXECUTION
FILE FILE
NODE NODE
PATH PATH
PAT1 - PAT4 PAT1 - PAT4
RESEQUENCE RESEQUENCE
RESTART RESTART
SAVE SAVE
WARNING WARNING
INITIAL INITIAL
MASS MASS
ORTH ANISO
B-1
NISA/HEAT to NISA/HEAT3 Translator
B-2
NISA/HEAT to NISA/HEAT3 Translator
*NDHEAT *BCDVAR
*PRINT *PRINT
*TEMPOUT *FLCNTL
*ELHEAT *BCGVAR
Notes:
B-3
NISA/HEAT to NISA/HEAT3 Translator
B-4
Appendix
C
FLUTL: NISA/HEATs Utility Package
Appendix C
C.2 Introduction
Ordinarily when a heat transfer problem is solved, temperature field is obtained. It
is usually calculated at all the computation points. This means that there is often
more data than the analyst needs as he may be interested in only a specific portion
of the computational domain. Also, instead of seeking values of temperature at
discrete points in the domain, the analyst may want them in an overall (integrated)
sense or one may be integrated in heat flux computation or heat transfer
coefficient calculation within (or across) a portion of the domain. The main
purpose of FLUTL is to provide such facilities to the analyst.
C-1
Basic Features of FLUTL
directory and creates a file with the next higher (in alphanumeric order) file name.
The highest filename FLUTL allowed to create is flutl99.ses. The entire session,
complete with questions and remarks made by FLUTL (i.e. practically everything
that the user sees on the screen), can also be logged simultaneously in a log file
(with a user-supplied name). Please note that the session files and the log files are
in ASCII format, and may therefore occupy substantial disk space; it is advised that
these files be purged from time to time.
In broad categories, the main capabilities of FLUTL can be divided under the
following headings:
(ii) To make a temperature profile along a line in space at a given time step.
Options (i) and (ii) will produce X-Y type of plot data that could be appended to
the NISA/HEAT binary result files (26 and 27), or dumped into a user supplied
ASCII file, or both. The plots may then be viewed using the NISA post-processor
DISPLAY III POST (ref. command GNI and GPH for plotting of binary data and
ASCII data, respectively).
Please note that if a job is RESTARTed, like most other old results, the appended
results will also be lost(ref. EXECUTIVE command RESTART).
Option (iii) will usually produce ASCII formatted output, which will be dumped in
the hard copy output file (the name of this file could be supplied by the user).
Default file names for the log file, the ASCII X-Y type plot file and the hard copy
output file are flutl.log, flutl.xyz, and flutl.out, respectively. Files with default
names are always opened as new files; this means that if a previous copy of any
of these files exists in the current directory, it is lost. Checks for duplication in user
supplied file names are performed.
In addition to the above functions, this package is equipped with handy capabilities
for user convenience. For example, the user may check all the nodal boundary
conditions using FLUTL.
C-2
FLUTL: NISA/HEATs Utility Package
FLUTL is largely menu driven and most questions require no explanation. Often
the answers are either YES or NO with the default demarcated within double
quotes (for example (Y/N) means that the default is NO). Answers are to be
given on-screen in most cases; in a few instances, FLUTL can be redirected to
read from an external file containing, for example, a list of nodes or elements. The
external file should contain data laid out in the same fashion as they would appear
if entered from the screen, and should preferably end with a data terminating line
containing only *END. It should be noted here that whenever more than one
number is expected, the numbers should be separated by either a comma or a blank,
and each number should not occupy more than a 15 character field.
FLUTL execution is sequential and the user may choose to climb up or down the
menu ladders by the appropriate menu option number. If either !Q or !E is typed
as a response to a FLUTL prompt, the program execution terminates.
Data (e.g., values of temperature) that has been retrieved could be modified as
follows:
Factors and offsets will be asked for by FLUTL at appropriate times. This feature is
helpful in scaling or in non-dimensionalizing the data.
FLUTL can also be run in session file mode. In this mode, the user puts the
responses (to the questions which would be asked by FLUTL) in an ASCII file and
supplies its name to FLUTL instead of typing the responses on the screen. The user
should look at any flutlnn.ses files to get a better idea of the format.
Care must be taken to ensure that the answers given in this session file are correct;
if FLUTL encounters a wrong answer, it may continue to read the session file, but
could interpret the rest of the answers wrongly. A typical example of such an error
is as follows: in the first pass of FLUTL, one or more of the files could be newly
opened. In the next pass, FLUTL views these files as existing files and may ask if
the user wishes to overwrite or append to them. These additional answers must be
inserted into the sequence of answers given during the first pass of FLUTL for the
rest of the answers to be interpreted correctly. Please note that an empty carriage
return is indicated by CR in the session file. Lines starting with ** are comment
C-3
Executing FLUTL
lines. While reading a session or an external data file, comment lines and blank
lines are ignored.
2.1) Default Log File Name is: flutl.log. Enter a log file name or <CR>
for default >
Maximum number of trial answers allowed for questions 4 and 5 are 3. The default
file27 name, whenever possible, is constructed from the file26 name. Depending
on whether the answer to question 3 is no or yes, either question 7a or 7b is
asked. If any of the hard copy output file or the ASCII X-Y plot file already exists,
C-4
FLUTL: NISA/HEATs Utility Package
then FLUTL asks the user what has to be done with the existing files (i.e.,
overwrite or append).
Menu.1
=======================================================
===========
MAIN MENU
=======================================================
===========
(0) = Quit execution of FLUTL
(1) = Change File26 and File27
(2) = Change Hard Copy Output File
(3) = Change X-Y Plot File
(4) = Process NISA/3D-FLUID Results
(5) = Check Nodal B.C. or Wall Function Nodes
Enter a Number or <CR> for Default (def = 4) >
Option 0 will terminate execution of FLUTL. Options 1 through 3 allow the user to
change some of the filenames during execution of FLUTL.
Option 5 lets the user verify the nodal boundary conditions and the wall function
nodes visually. In this case, the user should make only a CHECK run (ref.
EXECUTIVE command EXEC = CHECK) and supply the resulting file26
name to FLUTL. If the user selects to check nodal boundary conditions for a
C-5
Menus and Pertinent Notes
particular variable (say, Temperature T), FLUTL will ask the user to supply an
initialization value. FLUTL will first initialize the selected variable (in this case T)
with this initialization value, and then superimpose the applied nodal boundary
conditions of T on the initialized values. These nodal values will be written out as
results of a fictitious HEAT run (Load Case = 1, Time Step = 1) in the
corresponding file27. The user may then plot contours of T using DISPLAY III
POST. As all the free nodes would have the user specified initialization value, there
would be spots of this value among the legitimate boundary condition values. In
order to distinguish these spots, the initialization value should be any number (the
order of the number being comparable to the applied boundary condition) that has
not been used in specifying any of the boundary conditions for the chosen variable.
An important note that concerns the viewing of boundary conditions in this way is
that to clearly view the contours the user may have to judiciously zoom into
various corners of the model, and set the range (ref. DISPLAY POST command
RNG) to current view etc., to clearly view the contours, since in the case of small
elements and bad graphic resolution, the diagnostic spots may be missed.
Menu.2
========================================================
==========
MENU FOR POST PROCESSING OF 3D-FLUID RESULTS
========================================================
==========
(0) = Go Back to Main Menu
(1) = Produce Time History Plots
(2) = Create Profiles of Fluid Qtys
(3) = Miscellaneous Capabilities
C-6
FLUTL: NISA/HEATs Utility Package
Option 1 produces time history data for any of the fluid variables. By choosing
option 2, spatial variation of a fluid variable at a given time step can be obtained.
The menus for obtaining time histories and profiles are give in the following menus
(e.g., Menu.3 and Menu.4).
C-7
Menus and Pertinent Notes
Menu.3
========================================================
==========
TIME HISTORY PLOTS CAN BE OBTAINED FOR ANY OF THE FOL-
LOWING:
========================================================
==========
(0) = Previous Menu
(1) = U Velocity [U]
(2) = V Velocity [V]
(3) = W Velocity [W]
(4) = Kinetic Energy [K]
(5) = Dissipation Rate [D
(6) = Eddy Viscosity [EDD]
(7) = Stream Lines [SL]
(8) = Temperature [T]
(9) = Pressure [P]
(10) = X- Normal Stresses [Sxx]
(11) = Y- Normal Stresses [Syy]
(12) = Z- Normal Stresses [Szz]
(13) = XY- Shear Stresses [Sxy]
(14) = XZ- Shear Stresses [Sxz]
(15) = YZ- Shear Stresses [Syz]
Enter a number >
C-8
FLUTL: NISA/HEATs Utility Package
Menu.4
========================================================
==========
PROFILES CAN BE MADE FOR ANY OF THE FOLLOWING:
CURRENT TIME STEP = 3
========================================================
==========
(0) = Previous Menu
(1) = U Velocity [U]
(2) = V Velocity [V]
(3) = W Velocity [W]
(4) = Kinetic Energy [K]
(5) = Dissipation Rate [D
(6) = Eddy Viscosity [EDD]
(7) = Stream Lines [SL]
(8) = Temperature [T]
(9) = Pressure [P]
(10) = X- Normal Stresses [Sxx]
(11) = Y- Normal Stresses [Syy]
(12) = Z- Normal Stresses [Szz]
(13) = XY- Shear Stresses [Sxy]
(14) = XZ- Shear Stresses [Sxz]
(15) = YZ- Shear Stresses [Syz]
Enter a number >
Menu.5
========================================================
==========
C-9
Menus and Pertinent Notes
The fluid quantities that could be integrated are given in the following menu:
C-10
FLUTL: NISA/HEATs Utility Package
Menu.6
========================================================
==========
ANY OF THE FOLLOWING QUANTITY CAN BE INTEGRATED OVER A
SURFACE CURRENT TIME STEP = 3
========================================================
==========
(0) = Previous Menu
(1) = U Velocity [U] (2) = V Velocity [V]
(3) = W Velocity [W] (4) = Volume Flow Rate [Q]
(5) = Temperature [T] (6) = Pressure [P]
(7) = Stream Function [S] (8) = Kinetic Energy [K]
(9) = Dissipation [D] (10 = Eddy Viscosity [EDD]
(1 = X-Normal Stress [SXX] ) = Y-Normal Stress [SYY]
1) = Z-Normal Stress [SZZ] (12 = XY-Shear Stress [SXY]
(1 = XZ-Shear Stress [SXZ] ) = YZ-Shear Stress [SYZ]
3) (14
(1 )
5) (16
)
(1 = Surface Stresses (Normal and Shear)
7) = Film Heat Transfer Coefficient [h]
(1
8)
Enter a number >
It should be pointed out here that the TOTAL AREA reported in the hard copy
output is a scalar sum of all the elemental areas while the various components of
area are vector sums. These components could be viewed as projections of the
chosen surface onto planes represented by X, Y or Z = constant, respectively. The
surface normals are always computed to be positive when pointing outwards. Also
the words face and area in the output mean edge and length in a 2D case.
C-11
Menus and Pertinent Notes
In the following, some menus are presented that indicate choices currently
available in FLUTL to define node numbers for time history or profiles, to define
planes and lines, etc.
Menu.7
========================================================
==========
...SELECTING NODE NUMBER FOR TIME HISTORY PLOTS:
========================================================
==========
(0) = Back to Previous Menu
(1) = By Giving Node Number
(2) = By Giving Approx. X, Y and Z coords.
Enter a Number or <CR> for Default (def. = 1) >
Menu.8
========================================================
==========
...COLLECTING USER DEFINED NODES:
========================================================
==========
How Do You Want to Pick the Nodes ?
(0) = Quit
(1) = By Specifying the Nodes ON-SCREEN
(2) = By Specifying a Plane
(3) = By Specifying a Line
(4) = By Reading in a File with Node Numbers
Enter a number >
C-12
FLUTL: NISA/HEATs Utility Package
Menu.9
========================================================
==========
OPTIONS FOR DEFINING A PLANE:
========================================================
==========
How Would You Like to Define the Plane?
(0) = Quit Defining the Plane
(1) = By Specifying a Constant X, Y or Z Value
(2) = By Specifying 3 Nodes
(3) = By Specifying 3 Points
(4) = By Specifying 1 Node and 1 Normal
(5) = By Specifying 1 Point and 1 Normal
Enter a number >
C-13
Menus and Pertinent Notes
Menu.10
========================================================
==========
OPTIONS FOR DEFINING A LINE:
========================================================
==========
How Would You Like to Define the Line?
(0) = Quit Defining Line
(1) = By Intersection of Two Planes
(2) = By Specifying Two Nodes
(3) = By Specifying Two Points
(4) = By One Node and One Direction Vector
(5) = By One Point and One Direction Vector
Enter a number >
When nodes are picked up using a plane or a line (in case of creating profiles), all
the nodes that are within the user specified tolerance are picked. Sorting options are
available in FLUTL; the user is advised to use them since the sequence of node
numbers may not be the desired ordering of the nodes. The user also has an option
of restricting FLUTL to a subdomain of the original computational domain by
providing minimum and maximum extents of the subdomain.
C-14
FLUTL: NISA/HEATs Utility Package
C.6.1 Example #1
In the first example, the result files for flow over a cylinder (using 2D elements) are
taken. This is the same as Problem No. 8 in the NISA/3D-FLUID Demo Problems.
The following tasks are to be performed:
(ii) make a U-velocity profile at X = 15.0 for Y between 0.0 and 4.0
The names of the binary results files are CYL26.DAT and CYL27.DAT. The time
history plot data are put in the file CYL1.XYZ and the profile data are put in the
file CYL2.XYZ. The log of this particular session is given in the file CYL.LOG.
Printouts of these files are given in Section C.6.3.
C.6.2 Example #2
In this example result files for a Couette flow problem is considered. The problem
is solved using 3D elements. This is the first problem in the NISA/3D-FLUID
Demo Problems. The following tasks are to be performed:
(i) make a U-velocity profiles along the line defined by the intersection of the
planes X = 0 and Z = 0.
C-15
Some Examples of FLUTL Usage
processor DISPLAY PRE). For the case of the entry plane, this data is entered
interactively, while for the case of the exit plane, the data are compiled in a file
named ELEM.DATA and FLUTL is made to read this file. These two approaches
are used just to exemplify the two different modes of data entry to FLUTL. The
output detailing the evaluation of the flow rates is in the file COUETTE.OUT. The
log of this particular session is saved in the file COUETTE.LOG. Printouts of these
files are given in Section C.6.3.
C-16
FLUTL: NISA/HEATs Utility Package
C-17
Some Examples of FLUTL Usage
========================================================
==========
MAIN MENU
========================================================
==========
(0) = Quit Execution of FLUTL
(1) = Change File26 and File27
(2) = Change Hard Copy Output File
(3) = Change X-Y Plot File
(4) = Process NISA/3D-FLUID Results
(5) = Check Nodal B.C. or Wall Function Nodes
Enter a Number or <CR> for Default (def = 4) >4
========================================================
==========
MENU FOR POST PROCESSING OF 3D-FLUID RESULTS
========================================================
==========
(0) = Go Back to Main Menu
(1) = Produce Time History Plots
(2) = Create Profiles of Fluid Qtys
(3) = Miscellaneous Capabilities
Enter a Number >1
C-18
FLUTL: NISA/HEATs Utility Package
C-19
Some Examples of FLUTL Usage
C-20
FLUTL: NISA/HEATs Utility Package
C-21
Some Examples of FLUTL Usage
========================================================
==========
MENU FOR POST PROCESSING OF 3D-FLUID RESULTS
========================================================
==========
(0) = Go Back to Main Menu
(1) = Produce Time History Plots
(2) = Create Profiles of Fluid Qtys
(3) = Miscellaneous Capabilities
Enter a Number >0
C-22
FLUTL: NISA/HEATs Utility Package
========================================================
==========
MAIN MENU
========================================================
==========
How Would You Like to Define the Line?
(0) = Quit Execution of FLUTL
(1) = Change File26 and File27
(2) = Change Hard Copy Output File
(3) = Change X-Y Plot File
(4) = Process NISA/3D-FLUID Results
(5) = Check Nodal B.C. or Wall Function Nodes
Enter a Number or <CR> for Default (def = 4) >3
========================================================
==========
MAIN MENU
========================================================
==========
(0) = Quit Execution of FLUTL
(1) = Change File26 and File27
(2) = Change Hard Copy Output File
(3) = Change X-Y Plot File
(4) = Process NISA/3D-FLUID Results
(5) = Check Nodal B.C. or Wall Function Nodes
Enter a Number >4
C-23
Some Examples of FLUTL Usage
========================================================
==========
MENU FOR POST PROCESSING OF 3D-FLUID RESULTS
How Would You Like to Define the Line?
(0) = Go Back to Main Menu
(1) = Produce Time History Plots
(2) = Create Profiles of Fluid Qtys
(3) = Miscellaneous Capabilities
Enter a Number >2
C-24
FLUTL: NISA/HEATs Utility Package
C-25
Some Examples of FLUTL Usage
C-26
FLUTL: NISA/HEATs Utility Package
C-27
Some Examples of FLUTL Usage
C-28
FLUTL: NISA/HEATs Utility Package
========================================================
==========
MENU FOR POST PROCESSING OF 3D-FLUID RESULTS
========================================================
==========
(0) = Go Back to Main Menu
(1) = Produce Time History Plots
(2) = Create Profiles of Fluid Qtys
(3) = Miscellaneous Capabilities
Enter a Number >0
========================================================
==========
MAIN MENU
========================================================
==========
(0) = Quit Execution of FLUTL
(1) = Change File26 and File27
(2) = Change Hard Copy Output File
(3) = Change X-Y Plot File
(4) = Process NISA/3D-FLUID Results
(5) = Check Nodal B.C. or Wall Function Nodes
Enter a Number or <CR> for Default (def = 4) >0
NORMAL COMPLETION OF FLUTL
End of file: CYL.LOG
C-29
Some Examples of FLUTL Usage
C-30
FLUTL: NISA/HEATs Utility Package
C-31
Some Examples of FLUTL Usage
..PLEASE WAIT..
..SEARCHING FOR END-OF-FILE FOR FILE26..
..PLEASE WAIT..
..SEARCHING FOR END-OF-FILE FOR FILE27..
========================================================
==========
MAIN MENU
========================================================
==========
(0) = Quit Execution of FLUTL
(1) = Change File26 and File27
(2) = Change Hard Copy Output File
(3) = Change X-Y Plot File
(4) = Process NISA/3D-FLUID Results
(5) = Check Nodal B.C. or Wall Function Nodes
Enter a Number or <CR> for Default (def = 4) >4
========================================================
==========
MENU FOR POST PROCESSING OF 3D-FLUID RESULTS
========================================================
==========
(0) = Go Back to Main Menu
(1) = Produce Time History Plots
(2) = Create Profiles of Fluid Qtys
(3) = Miscellaneous Capabilities
C-32
FLUTL: NISA/HEATs Utility Package
C-33
Some Examples of FLUTL Usage
C-34
FLUTL: NISA/HEATs Utility Package
(0) = Quit
(1) = By Specifying the Nodes ON-SCREEN
(2) = By Specifying a Plane
(3) = By Specifying a Line
(4) = By Reading in a File with Node Numbers
Enter a Number >3
Do You Want to Give Min. and Max. XYZ Cutoffs (Y/N)? >!CR
How Do You Want to Define the Line?
(0) = Quit Defining Line
(1) = By Intersection of Two Planes
(2) = By Specifying Two Nodes
(3) = By Specifying Two Points
(4) = By One Node and One Direction Vector
(5) = By One Point and One Direction Vector
Enter a Number >1
..CURRENT TOLERANCE = .10000E-02
Enter New Tolerance or <CR>for Curr. Value >!CR
..DEFINING A LINE BY TWO PLANES..
..DEFINING THE FIRST PLANE..
How Would You Like to Define the Plane?
(0) = Quit Defining a Plane
(1) = By Specifying a Constant X, Y or Z Value
(2) = By Specifying 3 Nodes
(3) = By Specifying 3 Points
(4) = By Specifying 1 Node and 1 Normal
(5) = By Specifying 1 Point and 1 Normal
Enter a Number >1
..SPECIFYING A PLANE AT A CONST. X OR Y OR Z..
..FOR EXAMPLE, TO SPECIFY A PLANE AT X=2.2 ENTER X2.2..
C-35
Some Examples of FLUTL Usage
C-36
FLUTL: NISA/HEATs Utility Package
(1) = X-coordinates
(2) = Y-coordinates
(3) = Z-coordinates
(4) = Node Numbers
(5) = Calculated Fluid Qty
Enter a Number >5
What Would You Like to Plot on Y-axis?
(1) = X-coordinates
(2) = Y-coordinates
(3) = Z-coordinates
(4) = Node Numbers
(5) = Calculated Fluid Qty
Enter a Number >2
..X-AXIS: FLUID QTY ; Y-AXIS: Y-COORD ..
..PUTTING 8-CHARACTER PLOT CODE..
Enter Plot Code (max. 8-char) >U @X0,Z0
Do you wish to label your plot (Y/N) >!CR
C-37
Some Examples of FLUTL Usage
C-38
FLUTL: NISA/HEATs Utility Package
C-39
Some Examples of FLUTL Usage
========================================================
==========
MENU FOR POST PROCESSING OF 3D-FLUID RESULTS
========================================================
==========
(0) = Go Back to Main Menu
(1) = Produce Time History Plots
(2) = Create Profiles of Fluid Qtys
(3) = Miscellaneous Capabilities
Enter a Number >3
========================================================
==========
MENU FOR MISCELLANEOUS CAPABILITIES
========================================================
==========
(0) = Go Back to Previous Menu
(1) = Integrate a Fluid Quantity
Enter a Number >1
..TOTAL NUMBER OF TIME STEPS = 16
..DEFAULT TIME STEP = 16
C-40
FLUTL: NISA/HEATs Utility Package
C-41
Some Examples of FLUTL Usage
(0) = Quit
(1) = By Specifying the Elements ON-SCREEN
(2) = By Reading in a File with Element Numbers
Enter a Number >1
C-42
FLUTL: NISA/HEATs Utility Package
C-43
Some Examples of FLUTL Usage
ANY OF THE FOLLOWING QUANTITIES MAY BE OVER A SURFACE
CURRENT TIME STEP = 16
XV
(0) = Previous Menu
(1) = U Velocity [U] (2) = V Velocity [V]
(3) = W Velocity [W] (4) = Volume Flow Rate [Q]
(5) = Temperature [T] (6) = Pressure [P]
(7) = Stream Function [S] (8) = Kinetic Energy [K]
(9) = Dissipation [D] (10 = Eddy Viscosity [EDD]
(1 = X-Normal Stress [SXX] ) = Y-Normal Stress [SYY]
1) = Z-Normal Stress [SZZ] (12 = XY-Shear Stress [SXY]
(1 = XZ-Shear Stress [SXZ] ) = YZ-Shear Stress [SYZ]
3) (14
(1 )
5) (16
)
(1 = Surface Stresses (Normal and Shear)
7) = Film Heat Transfer Coefficient [h]
(1
8)
Enter a Number >4
C-44
FLUTL: NISA/HEATs Utility Package
**
** This data file has the element and corresponding face numbers
** at the outflow plane for the Couette flow problem
**
3,9,3,4
12,,,4
15,,,4
18,,,4
21,27,3,4
*END
C-45
Some Examples of FLUTL Usage
C-46
FLUTL: NISA/HEATs Utility Package
C-47
Some Examples of FLUTL Usage
** CASE TITLE: Volume Flow Rate at the Entry Plane for the Couette Flow
Problem **
C-48
FLUTL: NISA/HEATs Utility Package
CASE TITLE: Volume Flow Rate at the Entry Plane for the Couette Flow Prob-
lem
RESULT INTEGRATED = -.44170E+00
OVER THE TOTAL AREA = .10000E+01
COMP# 1 OF AREA = -.10000E+01
COMP# 2 OF AREA = .00000E+00
COMP# 3 OF AREA = .00000E+00
EXTENTS OF THE CHOSEN ELEMENT FACES ARE:
MINIMUM X-COORD. = .00000E+00
MINIMUM Y-COORD. = .00000E+00
MINIMUM Z-COORD. = .00000E+00
MAXIMUM X-COORD. = .00000E+00
MAXIMUM Y-COORD. = .10000E+01
MAXIMUM Z-COORD. = .10000E+01
** CASE TITLE: Volume Rate Flow through the Exit Plane for the Couette
Flow problem **
Factor = .10000E+01, Offset = .00000E+00
C-49
Some Examples of FLUTL Usage
CASE TITLE: Volume Rate Flow through the Exit Plane for the Couette Flow
problem
RESULT INTEGRATED = .44170E+00
OVER THE TOTAL AREA = .10000E+01
COMP# 1 OF AREA = .10000E+01
COMP# 2 OF AREA = .00000E+00
COMP# 3 OF AREA = -.16653E-15
EXTENTS OF THE CHOSEN ELEMENT FACES ARE:
MINIMUM X-COORD. = .20000E+01
MINIMUM Y-COORD. = .00000E+00
MINIMUM Z-COORD. = .00000E+00
MAXIMUM X-COORD. = .20000E+01
MAXIMUM Y-COORD. = .10000E+01
MAXIMUM Z-COORD. = .10000E+01
End of file: COUTTE.OUT
C-50
Appendix
D
Solving Thermal Stress Problems Using NISA/
Appendix D
Many of the engineering problems require coupled analysis of heat transfer and
statics. Many of the engineering problems require coupled analysis of heat
transfer and statics. Consider the design of an exhaust manifold of an automobile.
Exhaust gases are emitted through the exhaust manifold. The hot gases flowing
through the exhaust manifold heat up the manifold. This causes the structure to
deform and induces unwarranted stresses. Therefore, one should be able to
perform thermal stress analysis using temperature distribution obtained by a
thermal analysis package.
A method to solve thermal stress problems using NISA/HEAT and NISA II has
been outlined. The temperature distribution predicted by the NISA/HEAT package
is available for thermal stresses calculation in the solid using NISA II.
D-1
Solving Thermal Stress Problems Using NISA/HEAT and NISA II
The user creates a model in DISPLAY III to run heat transfer analysis
(ANAL=HEAT). All the relevant boundary conditions are applied to the model.
For any NISA/HEAT analysis, files 26 and 27 must be saved which are used for
post processing of final results. For thermal stress analysis, the user must also save
file 55. This can be done via the executive card SAVE = 26,27,55. File 55
contains the temperature distribution in the solid domain, that serves as the
temperature boundary condition for the static analysis (*NDTEMPER card).
The finite element mesh created for heat transfer analysis can also be used for
doing static analysis. The static analysis model is generated in the following
manner: first all the boundary condition applied for heat transfer analysis are
removed using the command HBC,DEL,BC_DATA or by the following sequence
of operations in menu mode
All the relevant boundary conditions for the static analysis, other than temperature,
are applied to this model. Since a reference temperature is needed at which thermal
stress is zero, this information must be inputted for the static model.
Since the statics model considers multiple boundary conditions from heat transfer
model as different load cases, the analysis procedure is somewhat different for
steady state and transient analysis. For steady state analysis, there is only one load
case, and the boundary condition file 55 created by the fluids module can be
included in the static nisa file by using *READ, FILE_NAME.DAT command. The
statics code will read the boundary conditions from the file FILE_NAME.DAT. On
the other hand, for transient analysis, there are multiple load cases, and if the
thermal stress analysis is performed after certain number of time steps, the data
files created by the fluid flow analysis are read in the following fashion:
*NDTEMPER, TCRV = n, FN = FILE_NAME, Step = m, where TCRV = n is
only required for non-linear static analysis, while step = m indicates the time step
identification number on the file, FILE_NAME that would be accessed by the
static analysis for the present load case.
The user first creates two models for performing heat and static analysis. In the
statics model he/she includes the appropriate *READ card, and the Heat-Static
D-2
Solving Thermal Stress Problems Using NISA/HEAT and NISA II
Fig D.1 Fluid-structure interaction between fluid flow and solid structure
D-3
Solving Thermal Stress Problems Using NISA/HEAT and NISA II
Fig D.2 Fluid-sructure interaction between fluid flow and solid shell
D-4
Appendix
E
System of Units
Appendix E
In these tables, three systems of units are cited, the International System (SI), the
metric absolute system (CGS) and the British Gravitational System (BGS).
The SI system is based on the fundamental dimensions of length (L), mass (M)
and time (t). Length is measured in meters (m), mass in kilograms (kg) and time in
seconds (sec). The program expects, however, the mass and the mass density to be
specified in force units (i.e., F.t2/L and F.t2/L4, respectively). Therefore, for the
users convenience, Table E.1 lists force as a fundamental unit, i.e., mass is
derived from force through Newtons law F = Ma, where a is the gravitational
acceleration.
The CGS system (Centimeter, Gram and Seconds) is based on the fundamental
dimensions of length (L), mass (M) and time (t). Length is measured in
centimeters (cm), mass in grams (gm) and time in seconds (sec). As in the SI
units, Table E.1 lists force as a fundamental unit and mass and mass density are
specified in force units (i.e., F.t2/L and F.t2/L4, respectively).
E-1
System of Units
The BGS system is based on the fundamental dimensions of length (L), force (F)
and time (t). Length is measured in feet (ft), force in pounds (lbf) and time in
seconds (sec).
It should be noted that in Table E.2, the dimension of heat energy is given the
symbol (E) for convenience. The units may be in Joules (J) or British Thermal
Units (Btu) which are equivalent to force times length (F.L), see Table E.7 for
conversion factors. Also, time is kept explicit in the units given in Table E.2 (e.g.,
J/sec or Btu/sec). Note however, that Watt (W) is defined as the power that
produces energy at the rate of one Joule per second, i.e., W = J/sec.
To use Tables Table E.3 through Table E.7, the units to be converted should be
located in the left-hand column whereas the required units should be located in the
top row of the table. The conversion factors common to both units should be
multiplied by the quantity to be converted. As an example, in Table E.3 one meter
(1m) is equal to 39.37008 inches.
The conversion factors have been reported to seven significant digits. However,
this does not mean that the factors are known to this accuracy. Factors given with
fewer than seven digits are exact.
To derive conversion factors for quantities not listed in Table E.3 through Table
E.9, it is always convenient to use unity factors. The following two examples
explain the use of such factors:
Example 1:
4
To change the Stefan-Boltzmann constant from W/m2.K4 to Btu/hr.ft2. R , use the
following
factors:
E-2
System of Units
8 W 8 J hr ft 2 R4 Btu
= 5.67033 10 ------------------- = 5.67033 10 -------------------------------- ------------------------------- -------------------------------
2 4 2 4 Btu 2 4
m K sec m K hr ft R
2 4
8 J hr ft R Btu
= 5.67033 10 --------- ------- ------- --------- -------------------------------
Btu sec 2 4 2 4
m K hr ft R
8 4 3 2 5 4 Btu see Table E.3,
= 5.67033 10 ( 9.479735 10 ) ( 3.6 10 ) ( 0.3048 ) ( --- ) -------------------------------
9 2 4 E.7 and E.9
hr ft R
8 2 4
= 0.171257 10 Btu hr ft R
Note:
When changing from one system of units to another, all the relevant
nondimensional parameters (Biot number, Numlet number, etc., as applicable),
must be held constant.
E-3
System of Units
Fundamental
Quantities
Length L m cm ft
Temparature T K K R
Derived Quanti-
ties
Coefficient of
thermal expan- 1/T 1/K 1/K 1/R
sion
Mass density 2 4 3 2 4 3 2 4 3 2 4 3
F t L M LN sec m kg m dyne sec cm gm m lb
cm f sec ft slug ft
++
applicable notes are given following Table E.9
E-4
System of Units
Heat flux
E/t.L2 J/sec.m2 J/sec.cm2 Btu/sec.ft2
Heat generation
E/t.L3 J/sec.m3 J/sec.cm3 Btu/sec.ft3
rate
E/M J/kg J/gmm Btu/lbm
Latent heat(2)
Thermal
conductivity
E/t.L.T J/sec.m.K J/sec.cm.K Btu/sec.ft. R
++
applicable notes are given following Table E.9
E-5
System of Units
Centi-
Meter Millimeter Inch Foot
LENGTH meter
m mm in ft
cm
Meter
1.0 1.0 x 10+2 1.0 x 10+3 3.937008 x 10+2 3.280840
m =
Centimeter
1.0 x 10-2 1.0 1.0 x 10+1 3.937008 x 10-1 3.280840 x 10-2
cm =
Millimeter
1.0 x 10-3 1.0 x 10-1 1.0 3.937008 x 10-2 3.280840 x 10-3
mm =
Inch
2.54 x 10-2 2.54 2.54 x 10+1 1.0 8.333333 x 10-2
in =
Foot 3.048 x
3.048 x 10-1 3.048 x 10-2 1.2 x 10+1 1.0
ft = 10+1
Newton
1.0 1.0 x 10+5 1.019716 x 10-1 2.248089 x 10-1 7.233015
N =
Dyne 7.233015 x 10-
1.0 x 10-5 1.0 1.019716 x 10-6 2.248089 x 10-6
dyne = 5
Kilogramf 7.093165 x
9.80665 9.80665 x 10+5 1.0 2.204623
kgf = 101
Poundf 3.217405 x
4.448221 4.44822 x 10+5 4.535924 x 10-1 1.0
lbf = 10+1
E-6
System of Units
dyne.sec2/
N.sec2/m lbf.sec2/ft
MASS kg f.sec2/m cm pdl.sec2/in lbf.sec2/in
(kg) (gmm) (slug)
kg f.sec2/ 9.80665 x
9.80665 1.0 6.719691 x 10-1 2.161997 x 10+1 5.599743 x 10-2
m = 10+3
N-sec2/ dyne-sec2/
kg f-sec2/m4 lbf-sec2/ft4 pdl-sec2/ft4
DENSITY m4 cm4 lbf-sec2/in4
(slug/ft3) (lbm/ft3)
(kg/m3) (gmm/cm3)
kg f-sec2/m4 9.80665 9.80665 x 10-3 1.0 1.902804 x 10-2 6.122091 x 10-1 9.176331 x 10-
= 5
E-7
System of Units
++
1J = 0.999835 int J
+++
(cal)IT indicates international steam table calorie, (cal)IT = 1.000669 (cal), where (cal) indicates thermochemical
calorie.
E-8
System of Units
5--- 5---
T ( C ) = T ( C ) T(K) - 273.16
9
( T ( F ) 32 ) ( T ( R ) 491.68 )
9
T(K) 5--- 5
T ( C ) + 273.16 T(K) ( T ( F ) + 459.68 ) --- ( T ( R ) )
= 9 9
9--- 9---
T ( F ) =
( T ( C ) + 32 )
5
( T ( K ) 255.38 ) T ( F ) T ( C ) 459.68
5
9--- 9---
T ( R ) =
( T ( C ) + 491.68 ) T(K) ( T ( F ) + 459.68 ) T ( R )
5 5
++
Note that conversions of the actual temperature values are governed by the
equations given in the table, but the relations between the intervals of the
temperature scales are given by:
1 degree Celsius = 1 degree Kelvin
= (9/5) degree Fahrenheit
= (9/5) degree Rankine
For example, to change a temperature of 100 C to F, we use the following
equation from the table:
T ( F ) = (9/5) T ( C ) + 32
= ( 9 5 ) ( 100 ) + 32 = 212 F
6
However, to change a quantity of 1.0 10 ( 1 C ) to( 1 F ) we use
6 1 6 1 F 6 F 1
1.0 10 ------ = 1.0 10 ------ ( ----- ) = 1.0 10 ( ------ ) -----
C C F C F
6 F 1
- ----- = 5--- 10 6 ( 1 F )
= 1.0 10 -------------------
( 9 5 ) F F 9
E-9
System of Units
4 Radiation Conduction
T h L ( K ( T h T c ) ) Nrc
Parameter
c Specific Heat
d Characteristic length (usually transverse)
h Heat transfer coefficient
k Thermal conductivity
L Characteristic length
r0 Characteristic length
t Time
Thermal Diffusivity k c
Volumetric coefficient of thermal expansion
Mean force path
Viscosity
Density
Surface Tension
Notes:
E-10
System of Units
E-11
System of Units
E-12
Keyword Index
Keyword Index
Appendix F
A
Anisotropic properties
input cards - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 6.2-4, 7.5.1-40
Axisymmetric problems - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - A.3-13
C
Conduction heat transfer
example problems - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - A.1-1
conduction heat transfer
Example problems - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - A.12-59
Convection heat transfer coefficient
computation using FLUTL - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - C.3-1
Convergence - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 3.4-2-14
convergence - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 3.3-7
Covection boundary condition
example problems - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - A.4-18
input cards - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 8.2.2-8
D
Disk space management - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 9.2.2-3
E
Element library - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -1
Error indicators/monitoring - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 3.4-12
Executive Commands - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -1
F
Finite element formulation - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 2.4-8
F-i
Keyword Index
F-ii
Keyword Index
F-iii