Beruflich Dokumente
Kultur Dokumente
Farzana Hussain
Registration No: 6012009, Session: 2009-10,
Department of Mathematics
Shahjalal University of Science and Technology, Sylhet,
Bangladesh.
A thesis submitted for the degree of
Doctor of Philosophy
January 2014
Submitted by
FARZANA HUSSAIN
Registration No: 6012009, Session: 2009-10
A Dissertation Submitted for the Degree of
DOCTOR OF PHILOSOPHY
(Signature)
External member
(Signature)
External member
(Signature)
Dedicated
to
my loving parents
and
my kids
Declaration
I hereby declare that the matter embodied in this thesis is the result
of investigation carried out by me in the Department of Mathematics,
Shahjalal University of Science and Technology, Sylhet, Bangladesh
under the supervision of Dr. Md. Shajedul Karim, Professor, Department of Mathematics, Shahjalal University of Science and Technology, Sylhet, Bangladesh. This thesis has not been submitted for
the Award of any type of Degree, Diploma, Associateship, Fellowship
etc. of any other University or Institute.
Farzana Hussain
Certified
Acknowledgements
It gives me immense pleasure to express my deep sense of respect and
gratitude to my guide Dr. Md. Shajedul Karim, Professor of
Mathematics, Shahjalal University of Science and Technology, Sylhet
for his enthusiastic guidance, encouragement and all kind of technical
support. He has always given me suggestions, instructions, inspirations in my academic progress as well as during the course of my
Ph.D. work.
I am also thankful to the SPM, Entity of the sub-project Postgraduate
Research Enhancement Schemes in Mathematics and Statistics under HEQEP, Higher Educational Quality Enhancement Project, UGC,
Bangladesh for the much needed financial support during the research
work.
I am grateful to Dr. Md. Rashed Talukder, Professor of Department of Mathematics, Shahjalal University of Science and Technology,
Sylhet for helping me in preparation of this thesis in LATEX. I am
deeply thankful to all of my colleagues for their help, kind cooperation and encouragement in my academic progress. Thanks to all the
office staffs of the Department of Mathematics, SUST. I would like to
thank The University Authority also for their collaboration and
cooperation.
My loving mother Mrs. Farida Parvin, My husband Mr. Icramul
Jalil, son Shadwan Sabah Jalil and my daughter Nujhat Tanjim
Jalil had always been my source of strength and inspiration to accomplish my Ph.D. work. I express my heartiest thanks to all of my family
members and friends for their great contribution during my research
period.
Farzana Hussain
Abstract
This thesis is concerned with accurate numerical schemes to evaluate
the integrals of rational functions of a n-th degree bivariate or trivariate polynomial numerator with linear or nonlinear denominator over a
triangular or tetrahedral domain in the local parametric space in two
dimensions or three dimensions respectively. These integrals encountered in finite element application to higher order partial differential
equations governing diverse physical phenomena employing the most
popular isoparametric, subparametric and superparametric finite elements. To derive a suitable, highly accurate technique to evaluate
such integrals numerically, this thesis studied on different approaches
to establish the new formulae.
Many realistic problems involving multiple integration frequently occur in a large number of engineering applications in Computer aided
design (CAD), Computer aided Engineering (CAE), Computer aided
Manufacturing (CAM), in geometric modelling as well as in robotics.
Particularly for irregular, complicated or distorted type of domain
such integrals cannot be evaluated analytically or explicitly and numerical integration schemes are the only alternative. Finite element
method (FEM) gained more importance due to its flexibility and versatility to obtain the solution of many complicated problems which
cannot be evaluated accurately by any other methods. In practical finite element analysis, complex geometrical domains are discretized by
different types of curvilinear isoparametric finite elements, which are
advantageously mapped into n-simplex (a square or a unit triangle)
or n-hyper (a cube or an orthogonal tetrahedron) in a local parametric space. If the domain consist a polygonal boundary, elements with
straight sides, usually triangle, quadrilateral, tetrahedron, cube are
It is astounding to note here that the Thesis is fully concerned with the
systematic gradual development of quadrature formulae to evaluate
the domain integrals and its contribution is directed in the area of
FEM and other numerical methods. Thus:
Chapter 1 describes the objectives and scope of the Thesis, important
applications and relevant literature review.
Chapter 2 contains some basic concepts and a brief historical background of FEM, elements in FEM, Interpolation functions. Some problems and some mathematical preliminaries are also discussed.
Chapter 3 is concerned with an accurate numerical integration scheme
to compute the triangular domain integrals. At first, the two dimensional domain of integration is divided into finite number of triangular
elements. Using mathematical transformation based on shape functions, each of the arbitrary triangle is transformed into a unit triangle
{(u, v)| 0 u, v 1, 0 u + v 1}, then the unit triangle is
subdivided into three quadrilaterals. Through isoparametric transformation each quadrilateral is again transformed into its equivalent
2-square given by {(, )| 1 , 1}. Then Gaussian quadrature
formula is derived for accurate evaluation of the triangular domain integrals employing the Gaussian quadrature formula for square domain
integrals.
In Chapter 4, each triangular element of the original domain is transformed into a standard triangular element {(s, t)| 1 s, t 1, s +
t 0} which is then subdivided into three quadrilaterals. Through
isoparametric transformation each quadrilateral is transformed into
its equivalent 2-square {(, )| 1 1, 1 1}. Then
quadrature formulae are obtained for accurate evaluation of the triangular domain integrals employing the Gaussian quadrature formulae
for square domain integrals. These developed techniques of Chapter
3 and 4 will be the best testing tools for the accuracy of higher order
(Gaussian) quadrature formulae developed in Chapter 5, 6, 7.
Chapter 5, 6, 7 concerned with different approaches for accurate nu-
Contents
Contents
xi
List of Figures
1 Introduction
1.1 Objectives and Scope .
1.2 Important Applications
1.3 Literature Review . . .
1.4 Plan of The Thesis . .
xvi
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2 Some Preliminaries
2.1 Finite Element Method Preliminaries . . . . . . . . . . . . .
2.1.1 Some Basics of the Finite Element Method . . . . . .
2.1.2 History of the Finite Element Method . . . . . . . .
2.1.3 Weighted Residual Method . . . . . . . . . . . . . . .
2.1.4 Finite Element Formulation of Problems . . . . . . .
2.1.5 Elements in FEM . . . . . . . . . . . . . . . . . . . .
2.1.6 Shape Functions (Interpolation Functions) . . . . . .
2.1.7 Isoparametric Elements . . . . . . . . . . . . . . . . .
2.1.8 Some Practical Problems . . . . . . . . . . . . . . . .
2.2 Mathematical Preliminaries . . . . . . . . . . . . . . . . . .
2.2.1 Properties of Multivariate Polynomial of n-th Degree
2.2.2 Integration of Monomials . . . . . . . . . . . . . . . .
2.2.3 Gauss Quadrature Formula . . . . . . . . . . . . . .
2.3 Finite Element Formulation of Physical Problems . . . . . .
xi
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2
3
7
8
12
.
.
.
.
.
.
.
.
.
.
.
.
.
.
16
16
16
19
20
21
22
23
24
25
28
28
29
30
30
CONTENTS
2.3.1
2.3.2
2.3.3
Torsion Problem . . . . . . . . . . . . . . . . . . . . . . .
Elasticity Problem . . . . . . . . . . . . . . . . . . . . . .
Heat Conduction Problem . . . . . . . . . . . . . . . . . .
31
33
37
of Gaussian
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
41
41
43
46
46
46
51
52
Domain Inte.
.
.
.
.
55
55
57
59
61
62
.
.
.
.
.
.
.
.
65
66
68
70
71
73
76
80
83
99
99
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
xii
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
CONTENTS
6.2
6.3
6.4
6.5
Problem Statement . . . . . . . . . . . . . . . . .
Numerical Evaluation Procedures . . . . . . . . .
6.3.1 Gauss Quadrature using Standard Triangle
6.3.2 Gauss Quadrature using Standard Triangle
Application Examples . . . . . . . . . . . . . . .
6.4.1 Two Dimensional Elasticity Problem . . .
Conclusions . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . .
(GQSTS) .
(GQSTM)
. . . . . . .
. . . . . . .
. . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
102
103
104
106
110
112
115
7 A Symmetrical Extended Gaussian Quadrature Formula for Evaluation of Triangular Domain Integrals
128
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.2 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7.3 Integration Over Arbitrary Triangle (IOAT) . . . . . . . . . . . . 131
7.4 Transformation of Triangular Domain . . . . . . . . . . . . . . . . 134
7.4.1 Symmetric Gauss Quadrature for Triangle (SGQTS) . . . 138
7.4.2 Symmetric Gauss Quadrature for Triangle (SGQTM) . . . 140
7.5 Application Examples . . . . . . . . . . . . . . . . . . . . . . . . 143
7.6 Error Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
7.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
8 Efficient Gauss Quadrature Rules for Tetrahedral Domain Integrals
170
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
8.2 Problem Statement over Tetrahedral Domain . . . . . . . . . . . . 172
8.3 Integration Over Arbitrary Tetrahedron . . . . . . . . . . . . . . . 173
8.4 Integration over Orthogonal Tetrahedron and Quadrature Formulae
(GQTTS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
8.5 Efficient Gaussian Quadrature formula for Tetrahedron (GQTTM) 181
8.6 Application Examples . . . . . . . . . . . . . . . . . . . . . . . . 186
8.7 Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . 191
8.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
9 Accurate Gaussian Quadrature Rules for Numerical Integration
over Tetrahedron
202
xiii
CONTENTS
9.1
9.2
9.3
9.4
9.5
9.6
9.7
9.8
Introduction . . . . . . . . . . . . . . . . . . . . . . . . .
Problem Statement over Tetrahedral Domain . . . . . . .
Integration over Arbitrary Tetrahedron . . . . . . . . . .
Integration Over Standard Tetrahedron . . . . . . . . . .
Numerical Evaluation Procedures . . . . . . . . . . . . .
9.5.1 Gaussian Quadrature for Tetrahedron (GQTSS) .
9.5.2 Gaussian Quadrature for Tetrahedron (GQTSM)
Application Examples, Results and Discussions . . . . .
Error Analysis . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
203
204
205
207
211
211
213
217
221
224
266
Appendix A
269
Appendix B
283
xiv
CONTENTS
References
332
List of Publications
341
xv
List of Figures
1.1
1.2
2.1
2.2
2.3
2.4
2.5
2.6
2.7
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
17
22
22
23
23
34
34
3.1
3.2
3.3
3.4
3.5
3.6
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
44
45
47
48
51
52
4.1
4.2
4.3
58
60
62
5.1
5.2
5.3
5.4
69
73
74
77
xvi
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4
5
LIST OF FIGURES
5.5
5.6
83
6.1
6.2
6.3
6.4
6.5
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
7.10
7.11
7.12
7.13
7.14
7.15
7.16
7.17
7.18
7.19
7.20
8.1
8.2
8.3
xvii
.
.
.
.
.
78
102
107
109
110
113
LIST OF FIGURES
8.4
8.5
8.6
8.7
.
.
.
.
185
186
187
190
9.1
9.2
205
9.3
9.4
9.5
9.6
9.7
9.8
9.9
9.10
9.11
9.12
9.13
9.14
9.15
9.16
206
214
217
234
234
235
235
236
236
237
237
238
238
239
239
10.1
10.2
10.3
10.4
10.5
10.6
248
249
252
253
256
257
xviii
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Chapter 1
Introduction
Chapter 1
Introduction
For numerical solution of continuum mechanics problem using finite element method
(FEM) an extra and important task is involved in the calculation of element matrices. Among all the numerical rules, Gaussian quadrature rule occupies a central
role for such calculations. Complication arise from two main sources, firstly the
large number of integrations that need to be performed and secondly in methods
which use isoparametric or subparametric elements, the presence of the determinant of Jacobian matrix in the denominator of the stiffness matrix for which the
integrands are rational functions. Among all the recently developed numerical
integration schemes, the technique based on symbolic manipulation can reduce
the computation time of the usual numerical integration and give adequate integration accuracy by increasing integration points. It is highly expected that the
expressions for the exact values of the integrals have to be evaluated with care
and hence the higher order numerical integration techniques are the best choice.
Explicit formulae are available by which a rational integral of monomial numerators with linear denominators can be evaluated. But when the denominator is
bilinear or expression of higher order then Gaussian quadrature formula is the
only alternative for the evaluation of these integrals.
Although, such integrals of rational functions occurs in the calculation of stiffness matrix of higher order partial differential equation employing curved triangular or tetrahedral elements, very few explicit formulae are developed so far.
Many researchers had also shown that the most accurate 7 - points and 13 points Gaussian quadrature formulae for the triangular domain integrals are not
2
1. Introduction
sufficient for satisfactory degree of accuracy and for some element geometry these
formulae are not at all reliable. Similar complication arises for the high order
transformation method as in this case the denominators are linear, bilinear or of
higher order. This thesis originated with the development of the integration formulae to compute the corresponding domain integrals encountered in the arena
of computational engineering science and to present fast algorithm to compute
element matrices in order to solve physical problems.
This chapter includes the objectives and scope of this thesis, important applications, relevant literature survey and plan of the Thesis.
1.1
1. Introduction
1. Introduction
and Zienkiewicz and Morgan [1983].
1. Introduction
the integrals have to be evaluated with care and hence the accurate integration
techniques are the best choice. For this reason most emphasis has been placed on
the four node isoparametric planar quadrilateral elements and exact integration
formulae are given by Babu and Pinder [1984]; Griffths [1994]; Griffths and
Mustoe [1995]; Hacker and Schreyer [1989]; Okabe [1981]; Rathod [1988];
Rathod and Islam [1998]; Rathod and Karim [2002]; Videla [1996].
Although integrals of rational functions occurs in the calculation of stiffness
matrix for the second or higher order partial differential equations employing
curved elements as well as in axisymmetric case with linear triangular or tetrahedral elements, such development has not been made so far. On the other hand
for the numerical integration method of integrating an arbitrary function of two
variables over the surface of a triangle very few integrating coefficients have been
published. According to the Gaussian quadrature formulae given by Cowper
[1973] the 13-point rule is the most accurate method. But Lague and Baldur
[1977] practically investigated that the method is not sufficient for a satisfactory
degree of accuracy and for some element geometry it is not at all reliable. Karim
[2001]; Rathod and Karim [2002] presented integration formulae to compute
the integrals of rational functions of a n-th degree bivariate polynomial numerator
with linear denominator over a unit triangle in a local two dimensional parametric
space. They also presented methods based on analytical integration schemes for
the calculation of stiffness matrix for the linear quadrilateral elements.
Triple integral formulae have always been of massive curiosity in all sides of
applied science and engineering applications. An idea about methods for evaluating volume integrals has given by Lee and Requicha [1982]. A hypothetical
advancement to the evaluation of volume integrals by transforming the volume
integral to a surface integral over the boundary of the domain of integration
was discussed by Timmer and Stern [1980]. Bernardini [1991]; Cattani and
Paoluzzi [1990a,b]; Lien and Kajiya [1984]; Zienkiewicz and Taylor [2000]
have presented formulae and algorithms for computing integrals of polynomials
over tetrahedron. Rathod et al. [2004, 2005] proposed formulae based on GaussLegendre quadrature rule for the numerical integration of an arbitrary function
over the standard tetrahedron. It is clearly noticed that they used Gaussian
quadrature schemes applicable for square or cubes to evaluate triangular or tetra6
1. Introduction
hedral domain integrals respectively and that exhibits, enforces severe crowd of
Gaussian points involved in the calculation process. These higher order (more
Gauss points and weights) formulae become erroneous due to the crowding of
Gaussian points and that effects the accuracy and efficiency of the calculations.
The main objective of the present Thesis is to develop efficient numerical integration schemes using less number of Gaussian points and weights for the desired
accuracy of both triangular and tetrahedral domain integrals of different types of
functions, e.g. (i) polynomial functions, (ii) non-polynomial functions and (iii)
rational functions of bivariate or trivariate polynomial numerator and a linear
or nonlinear denominator. These integrals encountered in many areas of science
and engineering. Finally, by algebraic manipulation and geometric consideration
the crowding of Gaussian points is totally removed from the scheme. Hence with
considerable savings in computational effort, numerical evaluation of the said domain integrals as well as the element matrices will be possible without resorting
to uncertain and expensive integration schemes. Further the Thesis enhances its
objectives by presenting computer codes (FORTRAN) for such higher order elements compatible with the element formulations. The Thesis will be (hopefully)
directed to contribute on this important topic and related basic developments in
the area of Finite Element Method and other numerical methods employed in
computational engineering.
1.2
Important Applications
Since the advent of freely available powerful computers in the mid-1960s, the finite element method has been a standard technique for the numerical analysis of
a wide range of continuum mechanics problems. In FEM solution procedure, the
high degree of accuracy which can be achieved with the higher order elements
means that a small number of such elements can often be used to obtain practically adequate solutions. Unfortunately, the simple shapes of elements restricts
severely their application in the analysis of practical problems, where often quite
complex geometrical boundaries have to be modeled. This restriction would be
removed if we could map a simple element in the local coordinate system, into
a more complex shape in the global coordinate system. Once a particular form
7
1. Introduction
of mapping is adopted and the coordinates are chosen for every element so that
these map into contiguous spaces, then shape functions written in the local space
can be used to represent the function variation over the element in the global
space without upsetting the inter element continuity requirements. This has led
to the development of elements of more arbitrary shape and are called isoparametric elements. The isoparametric elements are widely used for the analysis of
continuum mechanics problems of complex shapes involving curved boundaries.
Complication that arise for these elements is the presence of the determinant of the
Jacobian matrix in the denominator of the stiffness matrix, i.e., each component
of the stiffness matrix is an integral of rational function of bivariate or trivariate
polynomial numerator with a denominator, which is an expression of more than
one local variables. In the evaluation of these integrals normally a high order
and relatively too expensive numerical integration scheme is required. In axisymmetric case some integrands are rational functions of product of shape functions
in the numerator with nonlinear denominator in local variables. Such integrals
could not evaluated analytically and efficient numerical integration schemes are
demanded.
This Thesis mainly concentrates on the development of efficient numerical
integration formulae to compute the domain integrals encountered in many areas
computational engineering. Presented computer codes in this Thesis are straight
forward to evaluate such integrals.
1.3
Literature Review
A brief discussion to the literature concerned with this work is very necessary and
essential to build up a proper frame work for the research undertaken in this study.
Particularly in the finite element applications to problems involving an enclosed
region in R2 or R3 , if the original domain has a polygonal boundary, elements with
straight sides usually triangles, quadrilaterals, tetrahedrons, cubes are perfectly
satisfactory. If some part of the boundary of the region is curved, however elements
with at least one curved side are desirable, Mcleod and Mitchel [1972]. The two
principal methods of dealing with curved boundaries in the finite element method
are:
8
1. Introduction
The direct method: In this case basis functions are constructed to match
the curved boundaries and the integrations carried out directly in the original plane. This method originally developed with some success by Mcleod
and Mitchel [1975]; Wachspress [1972, 1973] for triangular elements. It
is shown that the basis functions are no longer polynomials and so the numerical work is necessarily more involved.
The use of isoparametric transformations: This is the most popular
method where each element involving part of the curved boundary in the
global (x, y, z) space is mapped into a standard straight sided element in
local (u, v, w) space by a formula of the same form as that used for the
interpolation on the standard element. Isoparametric elements were first introduced by Ergatoudis et al. [1968] and later mathematically analyzed by
Ciarlet and Raviart [1972]; Strang and Fix [1973]; Zlamal [1974]. According to Mcleod and Mitchel [1972, 1975] the only element, which can
deal adequately with most plane problems involving curved boundaries and
interface is the triangular element with two straight sides and one curved
side. For this element the isoparametric transformation enables the integration to be carried out in the normalized triangles.
An alternative method to rational basis function method and the isoparametric
method is given by Mcleod [1978]. This method, called a high order transformation method, allows any degree of approximation like the rational basis function
method, while retaining the simplicity of the isoparametric method. High order
basis functions, as well as Isoparametric functions, have been in the search for exact analytical expressions of integrals of rational function of bivariate polynomial
numerator with linear denominator over a unit triangle in local space.
Lague and Baldur [1977]; Rathod [1988] have shown that such integrals
are also encountered in axisymmetric finite element solution employing triangular
elements in a local coordinate system. In their study Lague and Baldur [1977]
clearly shown that the Gauss quadrature formulae developed by Cowper [1973]
are not satisfactory in respect of integration accuracy. Consequently they have
developed a technique based on a transformation of the triangular surface into a
square for which the Gaussian quadrature formulae have been extensively ana9
1. Introduction
lyzed and are readily available. It is also demonstrated that for some element the
proposed technique is very inefficient to obtain the sufficient accuracy and little
assurance is guaranteed if any divergence is evident. These integrals were considered by Andersen and Mcleod [1979]; Baart and Mcleod [1983]; Mcleod
[1978] extending the work of Mcleod and Mitchel [1972]. They employed recursive and reduction algorithms to derive analytical, closed-form formulae. They
also considered in detail the propagation of errors in the associated recursive formulae. The expressions derived by them ultimately depend upon the evaluation
of several integrals, which in turn require a recursive procedure in obtaining the
final result. Therefore the evaluation procedure is very lengthy even for one final
integral. In their study only one type of integrals with rational integrands had
been discussed.
The approach of Rathod and Karim [2000] differs from that of Andersen
and Mcleod [1979]; Baart and Mcleod [1983]; Mcleod [1978]; Mcleod and
Mitchel [1972] in the sense that they have identified four distinct types of rational integrands depending on the position of intermediate node(s) along the
curved side. Accordingly integration formulae they have derived and their application have shown in the calculation of the components of element stiffness
matrices for the second order partial differential equation employing the curved
triangular elements and in axisymmetric case with linear triangular elements. A
method for exact integration of the polynomials in the convex and concave circular sectors is provided in Silva and Mote [1988]. Because of the circular sector
these composite boundary elements have got limited applications in modelling
arbitrary curved domains. Several Authors considered a convex linear quadrilateral element and presented integration formulae. Babu and Pinder [1984];
Griffths [1994]; Griffths and Mustoe [1995]; Hacker and Schreyer [1989];
Okabe [1981]; Rathod [1988]; Videla [1996] have successfully attempted to derive the explicit finite element relations for general convex linear quadrilateral
finite elements. A good overview of numerical integration schemes which employ
Gaussian quadrature is given by Yagawa and Yashimara [1990]. They have
developed a numerical integration scheme based on symbolic manipulation, which
can dramatically reduce the computing time of the usual numerical integration
and give adequate integration accuracy by increasing integration points. They
10
1. Introduction
have shown that the evaluation of the coefficient of the bivariate polynomial in
numerator with linear denominator requires relatively large amount of time in
the total estimation of the stiffness matrix. The computation of these coefficient has to be improved in order to reduce the total computation time. Karim
[2001]; Rathod and Karim [2002] have presented methods for computing the
components of element stiffness matrix in an efficient way employing the exact
integration formulae encountered for planar quadrilateral and curved triangular
elements. Karim [2001] mainly contributed with special attention on the fournode, six-node and ten-node triangular elements with two straight sides and one
curved side.
Triple integral formulae have always been of great interest in all sides of applied science and engineering applications. Lee and Requicha [1982] have given a
good inkling of existing methods for evaluating volume integrals. A hypothetical
advancement to the evaluation of volume integrals by transforming the volume
integral to a surface integral over the boundary of the integration domain was discussed by Timmer and Stern [1980]. Bernardini [1991]; Cattani and Paoluzzi
[1990a,b]; Lien and Kajiya [1984] has presented formulae and algorithms for computing integrals of polynomials over tetrahedron. Rathod et al. [2004] derived
explicit integration formulae for computing volume integrals of trivariate polynomials over an arbitrary tetrahedron in Euclidean space. They evaluated this
volume integral as a sum of four integrals over the unit triangle. In Zienkiewicz
and Taylor [2000] a numerical integration formulae for tetrahedral regions with
a degree of precision up to 3 are listed. Rathod et al. [2005] proposed product
formulae based on Gauss-Legendre quadrature rule for the numerical integration
of an arbitrary function over the standard tetrahedron.
The serious drawbacks of such higher order (more Gauss points and weights)
formulae for two and three dimensional problems developed so far is that, they
become erroneous due to crowding of Gaussian points and weights in the calculation process that effects the accuracy and efficiency of the calculations and also
require much computing time and memory. The existing formulae for triangular
domain integrals are not sufficient for satisfactory degree of accuracy and for some
element geometry these formulae are not at all reliable.
11
1. Introduction
At first, in this thesis Gaussian quadrature formula is derived for accurate evaluation of the triangular domain integrals using the Gaussian quadrature formula
for square domain integrals as they are widely developed (from lower order to
higher order). Secondly, the developed technique is used as standard testing tools
for the accuracy of higher order (Gaussian) quadrature formulae so developed in
this thesis. Curved boundary can be accurately modeled by curved finite elements.
Because fewer elements are required, the effort needed to obtain a solution is usually reduced. This Thesis also presents methods for computing the components of
element stiffness matrix in an efficient way employing the numerical integration
scheme using triangular elements for two dimensional and tetrahedral elements
for three dimensional problems. This thesis considered general types of elements:
1. triangle with three straight sides or triangular element with two straight
sides and one curved side,
2. tetrahedron with four plane surfaces or tetrahedral element with one curved
and three plane surfaces.
The versatility of these elements enhanced in this thesis by presenting efficient
numerical integration schemes.
1.4
To achieve the objectives of this Thesis the plan of work is as follows. The work
in the thesis is spread over the following chapters:
Chapter 1 gives the objectives and scope of the Thesis, important applications
and relevant literature review.
Chapter 2 contains some basic concepts and a brief historical background of
FEM., elements in FEM, Interpolation functions, some problems are also discussed
and some mathematical preliminaries are described.
Chapter 3 concerned with accurate numerical integration schemes to compute
triangular domain integrals employing the Gaussian quadrature for square domain
integrals. At first, the arbitrary domain of integration is divided into finite number
of triangular elements. The arbitrary triangle is transformed into a unit triangle
12
1. Introduction
{(u, v)| 0 u, v 1, 0 u + v 1}, then the unit triangle is subdivided into
three quadrilaterals. Through isoparametric transformation each quadrilateral is
again transformed into its equivalent 2-square given by {(, )| 1 , 1}.
Then Gaussian quadrature formula for the triangular domain integrals is derived
employing the Gaussian quadrature formula for square domain integrals.
In Chapter 4, each triangular element of the original domain is transformed
into a standard triangular element {(s, t)| 1 s, t 1, s + t 0}, which
is then subdivided into three quadrilaterals. Each quadrilateral is transformed
into its equivalent 2-square {(, )| 1 1, 1 1}. Then quadrature
formula is obtained for accurate evaluation of the triangular domain integrals using
the Gaussian quadrature formula for square domain integrals. These developed
techniques of Chapter 3 and 4 are used as standard testing tools for the accuracy
of higher order (Gaussian) quadrature formulae developed later in this thesis.
Chapter 5, 6, 7 concerned with accurate numerical integration schemes to
compute two dimensional domain integrals of arbitrary functions. In these chapters numerical schemes are presented to evaluate the integrals with higher order
precision with lesser number of Gauss points in the calculation process. Different
types of elements are used for the search of an accurate method. Chapter 5 deals
with unit triangle, Chapter 6 concerned with standard triangle and Chapter 7
deals with a triangle which is symmetric about the line of symmetry. All these
studies are done to obtain a highly accurate, efficient numerical scheme to evaluate the triangular domain integrals. Comparison between the results of Chapter
5, 6, 7 and by the present methods are also shown.
Chapter 8 is devoted with development of efficient methods to evaluate the
tetrahedral domain integrals in three dimension with reduced number of Gaussian
points. At first, the three dimensional domain of integration is divided in to
finite number of tetrahedron. Using mathematical transformation depending on
shape functions, each of the tetrahedron is transformed into its equivalent unit
tetrahedron {(u, v, w)|0 u, v, w 1, 0 u + v + w 1} whose base triangle
is an unit triangle in w = 0 plane. Using mathematical transformations the
unit tetrahedral element is then transferred into its equivalent 2-cubic element
{(, , )| 1 , , 1}. Finally, two sets of Gaussian quadrature formulae
are derived for the evaluation of tetrahedral domain Integrals. The first set of
13
1. Introduction
quadrature formulae involves a huge crowding of Gaussian points and weights
within the domain of integration whereas the second set of formulae are completely
free of such crowding.
Chapter 9 devoted with development of efficient method with tetrahedral
elements {(s, t, r)| 1 s, t, r 1, s + t + r 1}, whose base triangle is a
standard triangle in r = 1 plane. Here also two sets of quadrature formulae are
developed of which the first set of quadrature formulae involves a huge crowding
of Gaussian points and weights within the domain of integration and the second
set of formulae are completely free of such crowding. All these studies are under
taken for obtaining a highly accurate, efficient numerical scheme to evaluate the
said triple integrals. Comparison between the obtained results of Chapter 8, 9
are also shown to verify the accuracy and efficiency of the presented schemes.
Chapter 10 is devoted to focus on some applications of the derived formulae
on some real world problem of Applied Science and Engineering.
Finally, Chapter 11 focus on the main conclusions of the present research
work. Computer codes based on the presented integration schemes are also appended.
The development of the numerical integration rules in this thesis is believed
to provide the way of accurate evaluation of integrals encountered in solving continuum mechanics problems with less computational effort.
14
Chapter 2
Some
Preliminaries
Chapter 2
Some Preliminaries
The primary intension of this chapter is to introduce the essentials of Finite Element Method, its adaptability, popularity and related information. The second
step is to bring in some well known theorems or formulae from a range of Mathematical origins - particularly from Algebra, Advanced Calculus and Numerical
Analysis those are essential to develop new and powerful techniques pertaining to
the problems investigated in this Thesis for ready reference.
2.1
In this section, brief discussion of the Finite Element Method, its history and
relevant definitions are narrated in a concise form.
2.1.1
The basic idea in the finite element method (FEM) is to find the solution of a
complicated problem by replacing it by a simpler one. Since the actual problem
is replaced by a simpler one in finding the solution, we will be able to find an
approximate solution rather than the exact solution. The majority of practical
design problems fall outside the reach of closed form solutions due to the complex
and irregular form of structures, complexity of loading conditions, non-linearity
and in homogeneity in properties of materials. For this reason, there is a growing
interest in numerical methods for the solution of continuum mechanics problems.
16
2. Some Preliminaries
In such instances Finite difference (FD) Method were used extensively in the
1960s. In the 1970s Finite Element began to replace Finite Differences as the
dominant technique. Now a day the FEM is a powerful tool for the approximate
solution of differential equations governing diverse physical phenomena. Its use
in industry and research is extensive, and indeed without it many problems in
science and engineering would be incapable of solution.
2. Some Preliminaries
the entire problem domain. As a consequence of these operations, a set of finite
linear equations is obtained in terms of a set of unknown parameters over each
element. Solution of these equations is achieved using the techniques of linear
algebra.
The advantages of finite elements are:
1. It is easier to conform the physical geometry where as with finite differences
the geometry is usually adjusted to fit the grid spacing,
2. The boundary conditions are easier to apply for geometrical reasons,
3. It is easier to use variable grid spacing by using small elements in regions of
interest and large elements in regions of little interest,
4. It is easier to write universal programs which can be applied to any geometry and a large number of physically different situations without having to
change of the code for each individual case.
The primary disadvantages of finite element calculations are:
a. The time consuming data input of the grid although the grid generation programs may substantially shorten this process;
b. Long computing times, specially for complex problems and three-dimensional
problems;
c. The requirement, that the analyst must find an element size such that the
governing differential equation is satisfactorily approximated everywhere.
With the advent of computers and its use in FEM such disadvantages are
considerably reduced and the method with above mentioned advantages become
a powerful numerical procedure for solving physical problems governed by a differential equation or an energy theorems in the realm of science and engineering
analysis. This method has principal characteristics that distinguish it from other
numerical procedures:
1. The method utilizes an integral formulation to generate a system of algebraic
equations;
18
2. Some Preliminaries
2. The method uses continuous piece wise smooth functions for approximating
the unknown quantity or quantities.
Actually, the second characteristic distinguishes the finite element method
from other numerical procedures that utilize also an integral formulation. The
implementation of the FEM can be subdivided into three broad steps:
1. Establishing the element interpolation properties;
2. Evaluating the element matrices; and
3. Solving an actual problem.
2.1.2
2. Some Preliminaries
model heat transfer problems with complex boundaries are discussed by Huebner
[1975]; a comprehensive three-dimensional finite element model for heat conduction is described by Heuser [1975].
Recent and ongoing research in finite element method indicates a profuse literature describing the versatility and reach mathematical abstraction of the subject.
Excellent reviews and description of the method given in the texts by Bickford
[1990]; Desai [1979]; Finlayson [1972]; Fletcher [1984]; Hughes [1987]; Reddy
[1984]; Segerlind [1984]; Zienkiewicz and Taylor [1989]. A short monograph
of the development of the FEM is given by Owen and Hinton [1980]. Programming of the finite element method is described by Smith [1982] and a rigorous
mathematical discussions is given in text book by Johnson [1987].
2.1.3
The underlying mathematical basis of the finite element method first lies with the
classical Rayleigh-Ritz method and variational calculus procedures introduced by
Rayleigh [1877] and Ritz [1909]. These theories provided the reasons why the
finite element method worked well for the class of problems in which variational
statements could be obtained. However, as interest expanded in applying the finite
element method to more areas, especially in mechanics, classical theory could no
longer be used to describe such problems.
Extension of the mathematical basis to nonlinear and nonstructural problems
was achieved through the method of weighted residuals, originally conceived by
Galerkin [1915] in the early 20th century. The method of weighted residuals was
found to provide the ideal theoretical basis for a wide set of problems including
those to which the Rayleigh-Ritz method can be applied. Basically, the method
requires the governing differential equation to be multiplied by a set of predetermined weights and the resulting product to be integrated over the problem domain; this integral is then required to vanish. The number of weighting functions
equals the number of unknown coefficients in the approximate solution. There are
several choices for the weighting functions, and some of the more popular choices
are: (1) Collocation method, (2) Subdomain method, (3) Least square method
and (4) Galerkins method. A rigorous description of the method of weighted
20
2. Some Preliminaries
residuals can be found in Finlayson [1972]. Technically, Galerkins method is a
particular case of the general weighted residual procedure where various types of
weights can be utilized; in the case of Galekins method, the weights are chosen
to be the same as functions used to define the unknown variables.
The underlying principle of the finite element method resides in the method
of weighted residuals. The two most commonly used procedures are the RayleighRitz and Galerkins methods. The Rayleigh-Ritz method is based on calculus of
variation; however the method is difficult to use on complicated equations. Since
the variational method has a major disadvantage: it is not applicable to any
differential equation containing a first derivative term. The Galerkins method
is simple to use and is guaranteed to yield a compatible approximation to the
governing differential equation. For the linear and elliptic equations, this method
yields results identical to the Rayleigh-Ritz approximations. Hence, Galerkins
method is the most preferable method to establish the approximations to the
governing equations.
2.1.4
In the realm of science and engineering many problems are amenable to solution
using two and three dimensional analysis. The best way to solve any such physical
problem governed by a differential equation is to obtain the analytical solution.
There are many situations, however, where the analytical solution is not at all
possible to obtain.
A numerical method can be used to obtain an approximate solution when an
analytic solution cannot be developed. All numerical solutions produce values at
discrete points for one set of the independent parameters. The calculated values
provide important information about the physical process even though they are
at discrete points for one set of the independent parameters. Hence, numerical
solutions, however, are now desirable with great importance than no solution at
all.
A primary advantage of the finite element method is the ease with which it
can be generalized to solve problems composed of several different materials and
having irregular boundaries. In fact it is in the problem domains that the strength
21
2. Some Preliminaries
of the finite element method becomes evident. The concept of a mesh and the
choices for available element discretizations become significantly important.
For the solutions of the problems, the physical domain is subdivided into sub
regions or elements. Different type of shapes can be used to define the elements.
Specifically, various types of element shapes and node numbering are the special
feature of the finite element method among all other numerical methods. Since
the FEM is based on the use of unstructured grids i.e. calculations are performed
on element-by-element basis, we are free to place our elements where we wish to
connect them to other elements without regard for orthogonally or sequential
node numbering.
2.1.5
Elements in FEM
Finite elements can normally be classified into three groups according to their
polynomial order: simplex, complex and multiplex, Oden [1972]. In a simplex
element, the number of coefficients in the polynomial is equal to the problems
dimensional space, plus one, the two dimensional three node triangular element
is the simplex element utilize a polynomial that contains constant and linear
terms. Similarly, the three dimensional four node tetrahedral element is the simplex element utilize a polynomial that contains constant and linear terms. A
complex element utilizes a polynomial that contains constant and linear terms,
plus higher-order terms. While the shape may be identical to a simplex element,
the complex element has additional boundary nodes and can have internal nodes.
The multiplex elements also use polynomials with higher order terms; however,
22
2. Some Preliminaries
the boundaries of the elements must be parallel to the co-ordinate axes. An example of a multiplex element is a rectangular or quadrilateral element, cubic or
cuboid element.
Figure 2.4: Discretized domain in 3D. Figure 2.5: A single complex 3D element.
Historically, the finite element method first employed triangular elements to
model structural problems in two dimension. The generation of the simple mesh
using linear triangular elements is relatively easy for simple geometries, and in
some instances may not require sophisticated mesh-generation techniques. Use
of quadratic elements would more accurately approximate the curved boundaries.
The quadrilateral element, a four-sided polygon is just as versatile as the triangle
in discretization a region. For three dimensional problems tetrahedron, cubic elements are very widely used. Many commercially available finite element codes for
structural analysis use a mixture of triangles and quadrilaterals; cubes and tetrahedrons for two and three dimensional problems respectively. A more extensive
discussion on accuracy and convergence properties of various elements are given
by Oden and Reddy [1976] and Carey and Oden [1983].
2.1.6
In finite element analysis using the displacement model, one assumes the variation
of displacement within an element since the true variation of displacement is not
known. In general, in higher mathematics, it is necessary in many situations to
deal with functions whose analytical form is either totally unknown or else is of
23
2. Some Preliminaries
such a nature that the function can not easily be subjected to such operations as
may be required. In either case, it is desirable to replace the given function by
another function, which can be more easily handled. This operation of replacing
or representing a given function by simpler one is known as interpolation in a
board sense. In finite element literature interpolating functions are referred to
as Shape Functions.
There are two types of interpolation functions: (1) Lagrange interpolation and
(2) Hermite interpolation. In the Lagrange interpolation, which is widely used in
practice, the assumed function takes on the same values as the given function
at specified points. In the Hermite type of function, the slopes of the function
also take the same value as the given function at specified points. Using the
Lagranges interpolation formula the Lagrange type shape function can be derived
more directly. An easy and systematic method of generating shape functions can
be achieved by simple products of appropriate polynomials in two or three coordinates. Now a days lower and higher shape functions for all finite elements are
available.
For any specified element the values of shape function Ni (say) gives a unit
value at the node i and zero value at the other nodes of the element. The variation
of field variable u (say) in the element can be described as the sum of all shape
functions each multiplied by the corresponding model displacement i.e.
u=
n
X
ui Ni
i=1
Hence, shape function plays the most important role in the finite element solution
procedure.
2.1.7
Isoparametric Elements
For the analysis of physical problems of complex shapes involving curved boundaries, simple elements are no longer sufficient. This has led to the development of
elements of more arbitrary shapes are called isoparametric elements. The concept
of isoparametric elements is based on the transformation of the parent element
into local or natural coordinate system to an arbitrary shape in the global coordinate system. A convenient way of expressing the transformation is to make use of
24
2. Some Preliminaries
the shape functions of the rectilinear elements in their natural coordinate system
and the nodal values of the coordinates. Thus the global x, y, z coordinates of a
point in an element may be expressed as
x=
n
X
xi Ni0 ,
y=
i=1
n
X
yi Ni0 ,
z=
i=1
n
X
zi Ni0
i=1
Ni0
Where,
are the shape functions of the parent rectilinear element and xi , yi , zi
are nodal coordinate of the element. Thus, the shape functions Ni0 used in the
above transformation help us to define the geometry of the element in the global
coordinate system. If this shape functions Ni0 are the same as the shape functions
Ni used to represent the variation of displacement i.e.
u=
n
X
ui Ni
i=1
2.1.8
2
2
+
D
G + Q = 0
y
x2
y 2
(2.1)
has many important applications in the physical sciences, which are known as
two dimensional problems. A few of these are discussed in this section.
1. The first application area is the torsion of non-circular sections. With
Dx = Dy = g1 , G = 0, Q = 2, the field equation reduces to the torsion
problem
25
2. Some Preliminaries
1 2 2
(
+
+ 2g) = 0
g x2 y 2
(2.2)
Where g is the shear modules of the material and is the angle of twist.
2. Several fluid mechanics problems are embedded within the field equation.
The streamline and potential formulations for an ideal irrotational fluid are
respectively governed by
2 2
+
=0
x2 y 2
(2.3)
2 2
+ 2 =0
x2
y
(2.4)
3. The seepage of water under a dram or retaining wall and within a confined
aquifer is governed by
2
2
Dx 2 + Dy 2 = 0
x
y
(2.5)
Dx
2
2
+
D
+Q=0
y
x2
y 2
(2.6)
Where Q is the point sink term, Dx , Dy are the permeability and represents
the piezometric head and aquifer is assumed to be confined.
5. The heat transfer from a two-dimensional fin to the surrounding fluid by
convection is governed by
Dx
2T
2T
2h
2hTf
+
D
T+
=0
y
2
2
x
y
t
t
26
(2.7)
2. Some Preliminaries
the coefficients Dx , Dy represents thermal conductivities in the x and y directions respectively, h is the convection coefficients, t is the thickness of
the fin, Tf is the ambient temperature of the surrounding fluid and T is the
temperature of the fin. If the fin is thin and the heat loss from the edges is
neglected.
When the body is very long in the z-direction and the temperature is a
function of only the x- and y- co-ordinate directions the heat transfer is
governed by
Dx
2T
2T
+
D
=0
y
x2
y 2
(2.8)
6. Seiche motion, which describes the standing waves on the bounded shallow
body of water, is governed by
h
2 w 4 2
2w
+
h
+
w=0
x2
y 2
gT 2
(2.9)
Where h is the water depth at the quiescent state, w is the wave height above
the quiescent level, g is the gravitational constant, and T is the period of
oscillation.
7. A fluid vibrating within closed volume is governed by
2 p 2 p w2
+
+ 2p=0
x2 y 2
c
(2.10)
Where p is the pressure excess above ambient pressure, w is the wave frequency, and C is the wave velocity in the medium.
Above different physical problems are contained within the general differential equation. Finite element formulation of these equation is given in detail by
Segerlind [1984].
27
2. Some Preliminaries
2.2
Mathematical Preliminaries
This section briefly introduces some of the mathematical results, well known theorems and formulae needed to develop the new and efficient techniques, which
highlight the lemmas and theorems pertaining to the problems investigated in the
thesis for ready reference.
2.2.1
n X
m
X
fmk, k xmk y k
m=0 k=0
k
n X
m X
X
m
X
fmk,k xmk y k
k=0
(n+1)(n+2)
2
the proof of the results follows Karim [2001]; Rathod and Karim [2000]. Similar
properties exists for trivariate polynomials also.
28
2. Some Preliminaries
2.2.2
Integration of Monomials
(a) The integral of a monomial of the type f (p, q) = pm q n (where m, n are nonnegative integers) over a unit triangle T = {(0, 0), (1, 0), (0, 1)} in the local
p, q coordinate system is denoted by ITm,n and given by
ITm,n
1p
pm q n dq dp =
=
p=0
q=0
m! n!
(m + n + 2)!
(b) The integral of a monomial of the type f (p, q) = pm q n (where m, n are nonnegative integers) over a 2-square S = {(1, 1), (1, 1), (1, 1), (1, 1)} in
the local p, q coordinate system is denoted by ISm,n and given by
ISm,n
p=1
pm q n dq dp =
=
q=1
IVm,n,r
=
p=1
q=1
pm q n rl dr dq dp
r=1
1p
1pq
=
p=0
q=0
pm q n rl dr dq dp =
r=0
29
m! n! l!
(m + n + l + 3)!
2. Some Preliminaries
2.2.3
m
X
wim f (im )
i=1
I
(m)
(m)
where i is the i-th sampling point and wi is the weight of the i-th sampling
point of m-th order.
The error in the m-th order one dimensional Gaussian quadrature formula
is given by
(22m+1 [m!]4 )
f 2m ()
=
(2m + 1)[(2m)!]3
Where f 2m () is the (2m)-th partial derivative of the function with respect to
x and is a point somewhere in the interval [1, 1].
The Gauss Quadrature formula for numerical integration of a function f over
domain D is of the form
ZZZ
f dD =
D
N
X
wi f (i , i , i )
i=1
where (, , ) are the coordinate of the i-th sampling point and wi is the weight
of the i-th sampling point.
2.3
This section presents Finite Element formulations for some two dimensional field
problems. The underlying mathematical basis of the finite element method first
lies with the classical theories introduced by Rayleigh [1877] and Ritz [1909].
Extension of the ideas, originally conceived by Galerkin [1915] in the early 20th
century and the method was found to provide the idea theoretical basis for wide
set of problems encountered in the realm of engineering science. Basically this
30
2. Some Preliminaries
method requires the governing differential equations to be multiplied by a set
of predetermined weights and the resulting product to be integrated over the
problem domain, this integral is then required to vanish. The number of weighting
functions equal to the number of unknown coefficients in the approximate solution.
A rigorous description of the method can be found in Finlayson [1972]. Here we
wish to use Galerkins techniques to present finite element problem equations of
some physical problems in a general manner.
2.3.1
Torsion Problem
and
in A
(2.11)
= 0 on C
(2.12)
2 2
+
+h
x2 y 2
(2.13)
where
R=
W
=
{W }
2
x
x
x
x x
2
W
W 2 =
{W }
y
y
y
y y
W
31
2. Some Preliminaries
Hence from Eq.(2.12) and (2.13) we obtain
ZZ
[ (W ) +
(W )]d
x
x
y
y
ZZ
W W
(
+
)d +
x x
y y
ZZ
W hd = 0
(2.14)
We known from the Greens theorem in plane
ZZ
F
G
(
+
)d =
x
y
Z
(F dy Gdx)
Which gives
ZZ
[ (W ) +
(W )]d =
x
x
y
y
Z
(W
dy W dx)
x
y
(2.15)
Thus we have
ZZ
(W )]d = 0
[ (W ) +
x
x
y
y
W W
+
)d +
(
x x
y y
ZZ
W hd = 0
(2.16)
M
X
Nj j
(2.17)
j=1
ZZ
M ZZ
X
W W
{
(
+
)d}j =
hNi d = 0 1 i m
x x
y y
j=1
32
(2.18)
2. Some Preliminaries
It is clear that Eq.(2.18) can be written as
[K]{} = {F }
(2.19)
Kij =
Ni Nj Ni Nj
+
)dxdy
x x
y y
(2.20)
ZZ
Fi =
hNj dxdy
(2.21)
Matrix [K] usually known as the stiffness matrix which is symmetric and {F } is
called load vector. Similarly, employing weighted residual Galerkins formulation
for other problems the stiffness matrix can be obtained as in the same form.
From Eq.(2.20), it is clear that the evaluation of the stiffness matrix [K] requires
to integrate the product of the global derivatives of shape functions.
2.3.2
Elasticity Problem
Two dimensional elasticity is generally categorized into two modes: plane strain
and plane stress. When the thickness of a solid object is large, a state of plane
strain is considered to exist. If this thickness is small compared to its overall
dimensions (x, y), the condition of plane stress is assumed. Both cases are subsets
of general three dimensional elasticity problems. In this instance, body forces (or
loads) cannot have components in the z-direction, nor vary in the direction of the
body thickness.
The governing equations which described two dimensional elastic stress are
defined as Reddy [1984]
x xy
+
+ fx = 0
x
y
xy y
+
+ fy = 0
x
y
(2.22)
where x and y are the normal stress components in the x and y directions,
33
2. Some Preliminaries
respectively; xy is the shear stress which acts in the xy plane; and fx and fy are
the body force terms. The strain displacement relations are defined from Eq.(2.22)
as
u
v
u v
(C11
+ C12 ) C33 (
+
) = fx
x
x
y
y y x
u v
u
v
C33 (
+
)
(C12
+ C22 ) = fy
x y x
y
x
y
(2.23)
(2.24)
tx = (C11
(2.25)
ty
(2.26)
u =
n
X
Ni ui
v =
i=1
n
X
Ni vi
i=1
34
(2.27)
2. Some Preliminaries
ZZ
[
u
v
u v
(C11
+ C12 ) + C33 (
+
) + fx ]Ni dxdy = 0
x
x
y
y y x
ZAZ
[C33
(2.28)
u v
u
v
(
+
)+
(C12
+ C22 ) + fy ]Ni dxdy = 0
x y x
y
x
y
G F
(
)dxdy =
x
y
Z
(F dx + Gdy)
If we choose
F = C33 (
u v
+
)Ni
x y
G = (C11
u
v
+ C12 )Ni
x
y
ZZ
[
u
v
u v
(C11
+ C12 ) + C33 (
+
)]Ni dxdy
x
x
y
y y x
(2.29)
ZZ
=
Ni
u
v
Ni u v
(C11
+ C12 ) + C33
(
+
)]dxdy
x
x
y
y y x
Z
Ni {C33 (
u v
u
v
+
)dx + (C11
+ C12 )dy}
y x
x
y
u v
u
v
Ni {C33 (
+
)dx + (C11
+ C12 )dy} =
y x
x
y
Z
Ni tx ds
S1
35
(2.30)
2. Some Preliminaries
ZZ
[
u
v
u v
(C11
+ C12 ) + C33 (
+
)]Ni dxdy
x
x
y
y y x
(2.31)
ZZ
=
Ni
u
v
Ni u v
(C11
+ C12 ) + C33
(
+
)]dxdy
x
x
y
y y x
Z
Ni tx ds
+
S1
ZZ
Ni
u
v
Ni u v
[
(C11
+ C12 ) + C33
(
+
) Ni fx ]dxdy
x
x
y
y y x
Z
Ni tx ds = 0
S1
(2.32)
Similarly, we have,
ZZ
Ni
u
v
Ni u v
( + )+
(C12 +C22 )Ni fy ]dxdy
[C33
x y x
y
x
y
Z
Ni ty ds = 0 (2.33)
S2
ue X Nj
=
uj
y
y
j
v e X Nj
=
vj
y
y
j
A
Z
n
X Z
j=1
uj (C11
Ni Nj
Ni Nj
+ C33
)dxdy
x x
y y
Ni Nj
Ni Nj
vj (C12
+ C33
)dxdy =
x y
y x
ZZ
Z
Ni fx dxdy +
Ni tx ds
S1
(2.34)
36
2. Some Preliminaries
for i = 1, 2, 3, ..., n. Similarly,
n ZZ
X
uj (C33
j=1
A
Z
n
X Z
j=1
Ni Nj
Ni Nj
+ C12
)dxdy
x y
y x
Ni Nj
Ni Nj
vj (C33
+ C22
)dxdy =
x x
y y
ZZ
Ni ty ds
Ni fy dxdy +
A
S2
(2.35)
Eq.(2.34) and (2.35) can be written in matrix form as
K11 u + K12 v = F1
(2.36)
K21 u + K22 v = F2
(2.37)
where
ZZ
K11 =
Ni Nj
Ni Nj
+ C33
)dxdy
x x
y y
ZZ
Ni Nj
Ni Nj
=
(C12
+ C33
)dxdy
x y
y x
(C11
A
K12 = K21
(2.38)
(2.39)
ZZ
Ni Nj
Ni Nj
+ C22
)dxdy
x x
y y
ZAZ
Z
=
Ni fx dxdy + Ni tx ds
K22 =
F1
(C33
Ni fy dxdy +
A
2.3.3
(2.41)
Z1
ZZ
F2 =
(2.40)
Ni ty ds
(2.42)
S2
2
2
+
k
+ Q = 0 in
y
x2
y 2
(2.43)
where kx and ky are the thermal conductivities and Q is an internal heat source
37
2. Some Preliminaries
or sink. The differential Eq.(2.43) is embedded within Eq.(2.1) in page 25. The
parameters for Eq.(2.1) Dx = kx , Dy = ky , G = 0, Q = Q with standard boundary
conditions
= on
= q
on q
(2.44)
where + q = . Now the weighted residual statement, using the trial functions themselves as weighting functions, can be deduced as the similar procedure
in the previous cases. We obtain,
Z
Z
Z
Ni b Ni b
kx
+
ky )dxdy = QNi dxdy qNi d
(
x x
y y
q
(2.45)
(2.46)
Kije
Z
=
(2.47)
(2.48)
eq
Here e is the surface of the element e and eq is the position of the element
boundary which lies on, or approximates to, a portion of q . The integral over
eq in Eq.(2.47) will thus only appear for elements that are adjacent to q . This
finite element formulation, is completely general. We wish to note that one can
use only the general triangular or the convex quadrilateral elements as well as
both type of elements if necessary for discretization of the problem domain. At
a glance, it can be said that in Finite Element equation of the physical problems
following integrals are needed to evaluate to form the element matrices:
38
2. Some Preliminaries
ZZ
A
ZZ
ZZ
Ni Ni
Ni Nj
Ni Ni
dxdy,
dxdy,
dxdy,
x x
y y
x y
A
Z
Z A
ZZ
ZZ
Ni Ni
Ni ds
Ni dxdy, Ni ds,
dxdy,
y x
A
S1
S2
for i, j = 1, 2, 3,..., N, and N is the number of nodes specified over the element
and s1 , s2 part of the boundary of the domain A.
39
CHAPTER 3
Evaluation of Triangular
Domain Integrals by use
of Gaussian Quadrature
for Square Domain
Integrals
Chapter 3
Evaluation of Triangular Domain
Integrals by use of Gaussian
Quadrature for Square Domain
Integrals
This chapter presents a technique to evaluate the integrals over the triangular surfaces using readily available Gaussian quadrature for the square domain integrals.
In order to perform the integration we transformed the arbitrary triangular domain in (x, y)-space to a unit triangular domain {(u, v)| 0 u, v 1, u + v 1},
then the unit triangle is divided into three quadrilaterals. Each of the quadrilaterals is transformed into its equivalent 2-square {(, )| 1 1, 1 1}.
As the technique suitably can employ higher order Gaussian quadrature for square
domain integrals, higher degree of accuracy of the integrals is possible to achieve
without resorting to inefficient quadrature formulae for triangles.
3.1
Introduction
Most of the domain integrals encountered in several areas of science and engineering are not amenable to evaluate analytically or tedious in calculation. In
recent and ongoing researches, we have been witnessing finite element method
41
3.2
Consider the integral over the arbitrary two dimensional domain as:
ZZ
I=
f d;
: original domain
(3.1)
To solve a practical problem using FEM, we divide the irregular domain into
finite number of elements. Observe that I can be calculated as a sum of integrals
evaluated over simplex divisions i :
=
i ;
i : completely covers
(3.2)
f d =
X ZZ
i
f di
(3.3)
f (x, y) dx dy;
: triangular domain
(3.4)
!
=
x1 x2 x3
y1 y2 y3
44
N1
N2
N3
(3.5)
N2 (u, v) = v,
N3 (u, v) = 1 u v
(3.6)
The original and the transformed triangles are shown in Figure 3.2. From
Eq.(3.5) and Eq.(3.6), we obtain
(3.7)
(3.8)
1u
II = J
u=0
(3.9)
(3.10)
v=0
1v
II = J
v=0
u=0
45
3.3
3.3.1
It is well known that the Gaussian quadrature formula for triangle to evaluate the
integral in Eq.(3.9) or in Eq.(3.10) is
II = J
N
GP N
GP
X
X
i=1
Wi Wj f (x(ui , vj ), y(ui , vj ))
(3.11)
j=1
where (ui , vj ) are the ij-th sampling points; Wi , Wj are corresponding weights
and NGP denotes the number of Gauss points in the formula and NGP 13.
It is thoroughly investigated by several authors that in some cases e.g., specially
for the nonpolynomial functions (integrand) available Gaussian quadrature for
triangle can not evaluate the integral II accurately and hence either higher order
formulae are required or a different method is time demanded.
3.3.2
Integration over the normalized unit triangle can be calculated as a sum of integrals evaluated over three quadrilaterals (Figure 3.3). Now, Eq.(3.9) can be
written as
46
II = J
Zu=0
Z
= J{
1u
e1
e2
ZZ
+
= J
3 ZZ
X
i =1
(3.12)
ei
47
u=
4
X
u i ni ,
v=
i=1
4
X
vi ni
(3.13)
i=1
where the shape functions to transform a quadrilateral to its equivalent 2square is given by
(1 + )(1 )
,
4
(1 + )(1 + )
n3 (, ) =
,
4
n1 (, ) =
(1 )(1 + )
,
4
(1 )(1 )
n4 (, ) =
4
n2 (, ) =
(3.14)
(3.15)
Jei =
For element e1 using the coordinates of A0 (1, 0), B 0 (1/2, 1/2), C 0 (1/3, 1/3), D0 (1/2, 0)
we have
u = (14 + 4 4 2)/24,
v = (5 + + 5 + )/24
1
f (x(u, v), y(u, v)) dv du =
96
e1
=1
1
f (x(u, v), y(u, v)) dv du =
96
e2
=1
1
f (x(u, v), y(u, v)) dv du =
96
=1
Combining the results for all the three quadrilateral elements we get
Z 1Z 1
J
II =
[f (X1 , Y1 )(4 + ) + f (X2 , Y2 )(4 + + )
96 1 1
+ f (X3 , Y3 )(4 )]d d
(3.16)
49
X1 =
1
[a
24 11
1
[b
24 11
X2 =
1
[a
24 21
Y2 =
1
[b
24 21
X3 =
1
[a
24 31
Y3 =
1
[b
24 31
and
a11 = 14x1 + 5x2 + 5x3
a14 = 2x1 + x2 + x3
b14 = 2y1 + y2 + y3
a24 = x1 + 2x2 x3
b24 = y1 + 2y2 y3
a34 = x1 x2 + 2x3
b34 = y1 y2 + 2y3
Now right hand side of Eq.(3.16) with Eq.(3.17) can be evaluated by use of
available higher order Gaussian quadrature for square domain. Since the unit
triangle is divided into three quadrilaterals and each quadrilateral is transformed
into equivalent 2-square, for m n points Gaussian quadrature formula, total
number of points within the triangle is 3 m n. Figure 3.5 and 3.6 shows
the distribution of Gauss points within the Unit triangle for m = n = 7 and
50
3.4
Application Examples
To show the accuracy and efficiency of the derived formulae, following examples
with known results are considered:
1y
I1 =
(x + y) 2 dx dy = 0.4
y=0
1
x=0
1y
I2 =
(y) 2 dx dy = 0.266
y=0
x=0
51
3.5
Conclusions
This chapter presented a new technique to use readily available Gaussian quadrature formulae for square domain integrals to evaluate the integral over the triangular surfaces. To do so, the original triangle is first transformed into a unit
triangle and then the unit triangle subdivided into three quadrilaterals. Finally,
all the quadrilaterals individually transformed into 2-square {(, )| 1
52
Gauss Points
77
GQT
13 13
355
Proposed Method
366
IOUT
377
388
399
3 10 10
I1
0.4001498818
0.4000451564
0.4000032328
0.4000013845
0.4000006727
0.4000003534
0.4000002024
0.4000001234
I2
0.2682188451
0.2674752175
0.2668388764
0.2667701218
0.2667336604
0.2667125320
0.2666994391
0.2666908966
53
Chapter 4
An Accurate Evaluation
Procedure of
Triangular Domain
Integrals
Chapter 4
An Accurate Evaluation
Procedure of Triangular Domain
Integrals
This chapter mainly concentrates to present quadrature schemes to evaluate accurately the triangular domain integrals with satisfactory degree of accuracy. In
this chapter, the original triangular domain in (x, y) space is first transformed to
Standard triangular domain {(s, t)| 1 s, t 1, s + t 0}. Then the standard triangle is divided into three quadrilaterals. Each of the quadrilaterals are
transformed into a square domain in {(, )| 1 1, 1 1}. The
technique successfully accommodates the readily available Gaussian quadrature
formulae for square and therefore it guaranties the high accuracy in evaluating
such integrals. Hence, the technique instead of the usual Gaussian quadrature
formula for triangle will find better place in application when high accuracy is
desired.
4.1
Introduction
Most of the domain integrals encountered in real problems defies our analytical
skills and we are resort to numerical integration methods, Bathe [1996]; Hughes
[1987]; Irons and Razzaque [1972]; Rogers and Adams [1990]; Strang and
Fix [1973]; Wachspress [1975]; Zienkiewicz and Taylor [1989]. Among vari55
4.2
In many practical problems, the complications arises with the numerical evaluation
of the integral over the irregular surface as:
ZZ
f (x, y) dx dy
I=
: irregular domain
(4.1)
the irregular domain can be divided over simplex divisions i where i are
triangles in two dimensional domain (see Figure 3.1 in page 44). Then I can be
calculated as a sum of integrals over 4i :
=
4i ;
4i : completely covers
(4.2)
To evaluate the integral I in Eq.(4.1), it is now required to evaluate the triangular domain integral
ZZ
II =
4 : arbitrary triangle
f (x, y) dx dy;
(4.3)
In FEM, integration over triangular domains is usually carried out in normalized co-ordinates. In this chapter, we are transforming each arbitrary triangles
into a normalized Standard triangle {(s, t)| 1 s, t 1, s + t 0}. To perform
this we first map one vertex (vertex 1) to the (1, -1), the second vertex (vertex 2)
to point (-1, 1) and the third vertex (vertex 3) to point (-1, -1) (as seen in Figure
4.1).
This transformation by using shape functions can be easily accomplished as:
57
x
y
!
=
x1 x2 x3
y1 y2 y3
N1
N2
N3
(4.4)
1
N3 (s, t) = (s + t)
2
(4.5)
where
1
N1 (s, t) = (s + 1),
2
1
N2 (s, t) = (t + 1),
2
1
1
1
x(s, t) = (x1 + x2 ) + (x1 x3 )s + (x2 x3 )t
2
2
2
1
1
1
y(s, t) = (y1 + y2 ) + (y1 y3 )s + (y2 y3 )t
2
2
2
(4.6)
(4.7)
The original and the transformed triangles are shown in Figure 4.1. Eq.(4.3)
58
s = 1
(4.8)
(4.9)
t = 1
t = 1
s = 1
4.3
The main interest of this chapter is to obtain a suitable, highly accurate, stable
technique to evaluate the triangular domain integrals of type given in Eq.(4.8) or
Eq.(4.9). In this section, an alternative procedures is proposed to evaluate the
integral II in order to use extensively the Gauss quadrature for square domain
integrals.
Integration over the normalized Standard triangle can be calculated as a sum
of integrals evaluated over three quadrilaterals (Figure 4.2). Following similar
procedure of calculation described in section 3.3.2 in page 46 we have
II =
=
J
4
i =1
ei
1
J
[f (X1 , Y1 )(4 + ) + f (X2 , Y2 )
96 1 1
(4 + + ) + f (X3 , Y3 )(4 )]d d
(x, y)
dt ds
(s, t)
59
(4.10)
1
[a
24 11
1
[b
24 11
X2 =
1
[a
24 21
Y2 =
1
[b
24 21
X3 =
1
[a
24 31
Y3 =
1
[b
24 31
and
a11 = 14x1 + 5x3 + 5x2
a14 = 2x1 + x2 + x3
b14 = 2y1 + y2 + y3
60
a24 = x1 + 2x2 x3
b24 = y1 + 2y2 y3
a34 = x1 x2 + 2x3
b34 = y1 y2 + 2y3
4.4
Application Examples
To show the accuracy and efficiency of the derived formulae, following examples
with known results are considered:
1y
I1 =
y=0
1
x=0
y
I2 =
y=0
(x + y) 2 dx dy = 0.6666667
1
(x2 + y 2 ) 2 dx dy = 0.881373587
x=0
4.5
Conclusions
An alternative technique to use readily available Gaussian points and corresponding weights for square domain integrals to evaluate the integral over the triangular surfaces is derived in this chapter. To do so, each triangle is first transformed into a standard triangle then the standard triangle is subdivided into three
quadrilaterals. Finally, all the quadrilaterals individually transformed into square
{(, ) | 1 1, 1 1} and consequently the original integral over the
62
Gauss Points
77
13 13
355
366
377
388
399
3 10 10
Test example
I1
I2
0.6606860757 0.8315681219
0.6637058258 0.8501738309
0.6660540844 0.8704939556
0.6662960549 0.8735856484
0.6664256193 0.8755247201
0.6665011716 0.8768203350
0.6665481694 0.8777286724
0.6665789279 0.8783900003
arbitrary domain reduces as resulted integral over the square domain. Hence, it is
possible to use higher order Gaussian quadrature formulae for the square domain
integrals to evaluate the triangular domain integrals with satisfactory degree of
accuracy. The technique is demonstrated through application examples and it is
recommended that the technique will be suitable alternative method to evaluate
triangular domain integrals instead of using the usual Gaussian quadratures for
triangles. Thus, the versatility of the triangular finite elements is now enhanced
with the proposed method.
It can be be seen that the final form of the techniques IOST in Eq.(4.10) for
standard triangular element and IOUT in the Eq.(3.16) for unit triangular element
are identical. Thus any choice of triangular domain will produce the same result by
these two methods. It is expected now that these developed techniques will be the
better alternatives to obtain the desired accuracy of triangular domain integrals
using higher order Gauss quadrature formulae. Also, the developed technique can
be used as the best alternative to test the accuracy of higher order (Gaussian)
quadrature formulae developed later in this thesis.
63
Chapter 5
Appropriate Gaussian
Quadrature Formulae
for Triangles
Chapter 5
Appropriate Gaussian
Quadrature Formulae for
Triangles
This chapter mainly presents higher order Gaussian quadrature formulae for numerical integration over the triangular surfaces. The arbitrary triangular domain
in (x, y)-space is transformed to a normalized Unit triangle in (u, v)-space, then
the Unit triangle is transformed into its equivalent 2-square. In order to show
the exactness and efficiency of derived quadrature formulae, it shows the effective
use of available Gaussian quadrature for square domain integrals to evaluate the
triangular domain integrals. Finally, it presents n n points (for n > 1) and
n(n+1)
1 points (for n > 2) Gaussian quadrature formulae for triangle utilizing
2
n-point one dimensional Gaussian quadratures. By use of simple but straightforward algorithms, Gaussian points and corresponding weights are calculated and
presented for clarity and reference. The proposed n(n+1)
1 points formulae com2
pletely avoids the crowding of Gaussian points and overcomes all the drawbacks
in view of accuracy and efficiency for the numerical evaluation of the triangular
domain integrals of any arbitrary functions encountered in the realm of science
and engineering. To verify the efficiency and accuracy, the method is applied to
evaluation of triangular domain integral of different types of functions and to the
evaluation of integrals for element stiffness matrices in two dimensional elasticity
problem.
65
5.1
Introduction
The integration theory extends from real line to the plane and three dimensional
spaces by the introduction of multiple integrals. Integration procedures on finite
domains underlie physically acceptable averaging process in engineering. In probabilistic estimations and in spatially discretized approximations, e.g., FEM and
boundary element methods (BEM), evaluation of integrals over arbitrary shaped
domain are the pivotal task. In practice, most of the encountered integrals
either cannot be evaluated analytically or the evaluations are very lengthy and
tedious. Thus, for simplicity numerical integration methods are preferred and the
methods extensively employ the Gaussian quadrature technique that was originally designed for one dimensional cases and the procedure naturally extends to
two and three dimensional rectangular domains according to the notion of the
Cartesian product. Gaussian quadratures are considered as the best method of
integrating polynomials because they guarantee that they are exact for polynomials less than a specified degree. In order to obtain the result with the desired
accuracy, Gaussian integration points and weights necessarily increase and there
is no computational difficulty except time in evaluating any domain integral when
the two and three dimensional regions are bounded respectively, by systems of
parallel lines and parallel planes.
Analysts cannot ignore at all the randomness in material properties and uncertainty in geometry that are frequently encountered in complex engineering systems. This entails the evaluation of an integral of the function over the domain
of the component. In general, the -shape-class is very irregular in two and
three dimensional geometry. For non-parallelogram quadrilateral, very frequent
in finite element modeling, there is no consistent procedure to select the sampling
point to implement a Gaussian quadrature on the entire element. Special integration schemes, e.g., reduced integration over quadrilaterals have been successfully
developed in Hughes [1987] and are widely used in commercial programs. There
is no methodical way to design such approximate integration schemes for polygons with more than four sides. An attempt to distribute the sampling points
according to the governing perspective transformation fails to assure the error
order germane to the quadrature formula. The reason can be traced to the crowd-
66
5.2
One of the main task of this thesis is to evaluate the integral of an arbitrary
function f for two dimensional problems:
ZZ
II =
f (x, y) dx dy;
: triangle (arbitrary)
(5.1)
These type of integrals arise in axisymmetric finite element method with linear
triangular element as well as in finite element formulations of second order linear
differential equations by use of quadratic triangular element with two straight
sides and one curved side. In FEM, the original domain of integration is usually
discretized into finite number of elements. Unfortunately, the simple shapes of the
elements restrict severely their applications in the analysis of practical problems,
where often quite complex geometrical boundaries have to be modeled. This
restriction can be removed by mapping a simple element in the local coordinates
into a more complex shape in the global coordinate system. Once a particular
form of mapping is adopted and the coordinates are chosen for every element so
that these map into contiguous space, then shape functions written in the local
element space can be used to represent the function variation over the element
68
!
=
x1 x2 x3
y1 y2 y3
N1
N2
N3
(5.2)
where
N1 (u, v) = u,
N2 (u, v) = v,
N3 (u, v) = 1 u v
(5.3)
69
(5.4)
(5.5)
1u
II = 4
u=0
(5.6)
(5.7)
v=0
1v
II = 4
v=0
u=0
The evaluation of integrals II in Eq.(5.6) and in Eq.(5.7) will yield the same
results. Any one of these two can be evaluated numerically. Influences of these
integrals will be investigated later to present new quadrature formulae for triangles.
5.3
70
(5.9)
= (1 )
(,
8
du dv = dv du =
(u, v)
1
d d = (1 ) d d
(,
8
(5.10)
(5.11)
1 + (1 )(1 + )
1 + (1 )(1 + ) 1
,
), y(
,
)}
dd
2
4
2
4
8
1 1
Z 1Z 1
1 + (1 )(1 + ) 1
,
}
dd
(5.12)
= 4
f{
2
4
8
1 1
Z
II = 4
f {x(
In order to evaluate the square domain integral II in Eq.(5.12) efficient Gaussian quadrature coefficient (points and weights) are readily available so that any
desired accuracy can be obtained.
5.3.1
II = 4
m X
n
X
i=1
mn
X
= 4{
(m)
(m)
1 + i
f{
2
j=1
(1 i
f (ur , vr )Gr }
(n)
)(1 + j ) 1 i(m)
(m)
(n)
}(
)Wi Wj
4
8
(5.13)
r=1
Here, (ur , vr ) are the new Gaussian points, Gr is the corresponding weights
(m)
(n)
for triangles where i , i = 1, m; j , j = 1, n are the m-th and n-th order
(m)
(n)
Gaussian quadrature points respectively. Wi , Wj are the corresponding m-th
and n-th order weights respectively.
71
u = (1
1+ 1+
)(
),
2
2
v=
1+
2
Then one can obtain (on the same lines of Eq.(5.12) and Eq.(5.13))
II = 4
=
f(
(1 + )(1 ) 1 + 1
,
)
dd
4
2
8
1 1
mn
X
4{
G0r f (u0r , vr0 )}
r=1
(5.14)
where G0r and (u0r , vr0 ) are respectively weights and Gaussian points for triangle.
All the Gaussian points and corresponding weights can be calculated simply using
the following algorithm:
step 1. r 1
step 2. i = 1, m
step 3. j = 1, n
(m)
(1 i
Gr =
8
(m)
(m)
(n)
(m)
Wi Wj ,
1 + i
ur =
2
(m)
(n)
(1 j ) (m) (n)
Wi Wj ,
=
8
step 4. compute step 3
G0r
u0r
(1 + i
vr =
(1 i
(n)
)(1 j )
,
4
(n)
)(1 + j )
4
(n)
vr0
1 + j
=
2
For clarity and reference, computed Gauss points and weights for different
values of m = n based on the above algorithm are listed in Table 5.1. Figure 5.2
and Figure 5.3 shows the distribution of Gaussian points for m = n = 10 and
72
5.3.2
n(n + 1) (n m + 1)(n m + 2)
;
2
2
nm
The new Gaussian quadrature formulae for Eq.(5.6) can be written as:
74
II = 4
m1
X n+1i
X
i=1
j=1
(m1)
(m1)
1 + i
f{
2
(1 i
(m1)
1 i
8
(m1)
)Wi
(n+1i)
Wj
= 4{
(n+1i)
)(1 + j
4
TP
X
f (pr , qr )Lr }
(5.15)
r=1
where (pr , qr ) are the new Gaussian points, Lr is the corresponding weights
for triangles. Similarly, the new Gaussian quadrature formulae for Eq.(5.7) can
be written as:
f{
II = 4
1 1
T
P
X
(1 + )(1 ) 1 + 1
,
}
d d
4
2
8
= 4{
(5.16)
r=1
where (p0r , qr0 ) and L0r are respectively the Gaussian quadrature points and
weights for triangle. All the Gaussian points and corresponding weights can be
calculated simply using the following algorithm:
step 1. r 1
step 2. i = 1, m 1
step 3. j = 1, n + 1 i
(m1)
(n+1i)
(m1)
(1 i
)(1 + j
1 + i
, qr =
pr =
2
4
(m1)
(1 i
) (m1) (n+1i)
Lr =
Wi
Wj
8
75
(m1)
(m1)
)(1 j
)
1 + j
=
, qr0 =
4
2
(m1)
(1 j
) (n+1j) (m1)
L0r =
Wi
Wj
8
r =r+1
p0r
(1 + i
5.4
Application Examples
To show the accuracy and efficiency of the derived formulae, following examples
with known results are considered:
1y
I1 =
(x + y) 2 dx dy
y=0
x=0
1y
y=0
(x + y) 2 dx dy
I2 =
x=0
76
Figure 5.4: Gauss Points for m =n =10 (54 points) using GQUTM.
Z
(x2 + y 2 ) 2 dx dy
I3 =
y=0
x=0
exp|x + y 1|dx dy
I4 =
y=0
x=0
The computed results are summarized in Table 5.3 for m m points GQUTS
and m(m+1)
1 points GQUTM formulae. In order to compare the results obtained
2
by methods GQT and IOUT are also included for all examples.
Some important remarks from the Table 5.3 are:
77
Figure 5.5: Gauss Points for m =n =15 (119 points) using GQUTM.
Usual Gauss quadrature (GQT) for triangles e.g. 7-point and 13point rules cannot evaluate the integral of non-polynomial functions
accurately.
Splitting unit triangle into quadrilaterals (IOUT) provides a way of
using Gaussian quadrature for square. But the convergence rate is
slow (as it needs the function to be calculated at more points) but
satisfactory in view of accuracy.
New Gaussian quadrature formulae for triangle (GQUTS and GQUTM)
are exact in view of accuracy and efficiency. GQUTM scheme needs
78
p,q
y=0
Example-1: I
r,0
Example-2: I
x=0
Example-3: I
1y
=
y=0
0,0
x=0
1y
y=0
0,r
1y
x=0
1y
=
y=0
x=0
1y
xp y q
dx dy
+ x + y
xr
dx dy
0.375 0.375 x
6= = 0
yr
dx dy
0.375 0.375 y
6= = 0
1
dx dy
12 + 21.53679831x 8.821067231y
6= 0, 6= 0
Example-4: I
0,0
=
y=0
x=0
1
dx dy
12 + 9.941125498(x + y)
= 6= 0
Results are summarized in Tables 5.4, 5.5, 5.6 and 5.7 for mm points GQUTS
and m(m+1)
1 points GQUTM formulae.
2
In Tables (5.4 - 5.7) for method GQUTS, Formula 1 is for Eq.(5.13) and
Formula 2 is for Eq.(5.14); for method GQUTM, Formula 1 is for Eq.(5.15) and
Formula 2 is for Eq.(5.16). These tables substantiated the influences of numerical
79
5.4.1
From the principle of virtual work, the stiffness matrix of an arbitrary element is
given by
80
[K]e =
(5.17)
A0
Where [D] is a material property matrix, [B] is the strain displacement matrix
and the integration is performed over the element A0 in global coordinates. The
symmetric material property matrix [D] for anisotropic materials can be written
as
D11
D12
D13
D14
D12
D22
D23
D24
D13
D23
D33
D34
D14
D24
D34
D44
A0
For a two dimensional element, the matrix [Bi ] has the form
Ni
x0
Ni
y 0
Ni
y 0
Ni
x0
In which Ni are the appropriate element shape functions. The element stiffness
relating nodes i and j for an anisotropic material under two dimensional loading
is given by:
[Kij ]e =
Kij11 Kij12
Kij21 Kij22
81
Ni Nj
dx0 dy 0
t s
t, s = x0 , y 0
A0
We shall evaluate these double integrals, where A0 is the triangle element with
two straight and one curved side. It can be shown that the evaluation of integrals
now amounts to the evaluation of integrals of the type
II
pq
xp y q
dx dy
J(x, y)
A
Z 1 Z 1x p q
xy
=
dx dy
x=0 y=0 J(x, y)
ZZ
(5.18)
Where J is the Jacobian of transformation and A is the unit triangle: {(x, y)|0
x, y 1, x + y 1}. The transformation which maps a one side curved triangle
A0 in to an orthogonal unit triangle A is given by
t=
m
X
Ni (x, y)ti ,
t = x0 , y 0 and
ti = (x0i , yi0 )
i=1
Where Ni are the corresponding shape functions for each nodes. Then for
quadratic curved element (m = 6, Fig 5.6) we have
x0 = x03 + (x01 x03 )x + (x02 x03 )y + (4x04 2x01 2x02 )xy
y 0 = y30 + (y10 y30 )x + (y20 y30 )y + (4y40 2y10 2y20 )xy
The shape functions Ni which transforms a quadratic curved triangular element to a orthogonal unit triangle are given in Appendix A. If the coordinates
are given 1(1, 0), 2(0, 1), 3(0, 0) and 5, 6 are at mid points of two straight sides,
then we have
82
5.5
Conclusions
Though Gaussian quadrature formulae for the triangular domain integrals are
widely used in FEM and BEM, they are not so developed comparing the square
domain integrals. To achieve the desired accuracy of the triangular domain integral it was necessary to increase the number of points and corresponding weights.
Therefore, it is an important task to make a proper balance between accuracy and
efficiency of the calculations.
83
84
Table 5.1: Computed weights G and corresponding Gauss points (u, v) for n n
point method (GQUTS).
n
G
0.684643776713535D-01
0.109543004274166D+00
0.684643776713535D-01
0.617283950617284D-01
0.987654320987654D-01
0.617283950617284D-01
0.869611615580696D-02
0.139137858492911D-01
0.869611615580696D-02
0.408492975558259D-02
0.882401185479183D-02
0.120457917319127D-01
0.131855763377203D-01
0.120457917319127D-01
0.882401185479183D-02
0.408492975558259D-02
0.788426937775230D-02
0.170311096195916D-01
0.232494247307213D-01
0.254493080585855D-01
0.232494247307213D-01
0.170311096195916D-01
0.788426937775230D-02
0.868834290475827D-02
0.187680193729915D-01
0.256205064439922D-01
0.280447438447181D-01
0.256205064439922D-01
0.187680193729915D-01
0.868834290475827D-02
0.676492884474095D-02
0.146131796369908D-01
0.199486720263899D-01
0.218362349021241D-01
0.199486720263899D-01
0.146131796369908D-01
0.676492884474095D-02
0.367196988822002D-02
0.793196156673469D-02
0.108280404231932D-01
0.118526002080616D-01
0.108280404231932D-01
0.793196156673469D-02
0.367196988822002D-02
0.117014141080756D-02
0.252766688745083D-02
0.345055620900489D-02
0.377705121539608D-02
0.345055620900489D-02
0.252766688745083D-02
0.117014141080756D-02
0.106659360355737D-03
0.230398933768031D-03
0.314521061065587D-03
0.344281351761558D-03
0.314521061065587D-03
0.230398933768031D-03
0.106659360355737D-03
u
0.112701665379258D+00
0.112701665379258D+00
0.112701665379258D+00
0.500000000000000D+00
0.500000000000000D+00
0.500000000000000D+00
0.887298334620742D+00
0.887298334620742D+00
0.887298334620742D+00
0.254460438286208D-01
0.254460438286208D-01
0.254460438286208D-01
0.254460438286208D-01
0.254460438286208D-01
0.254460438286208D-01
0.254460438286208D-01
0.129234407200303D+00
0.129234407200303D+00
0.129234407200303D+00
0.129234407200303D+00
0.129234407200303D+00
0.129234407200303D+00
0.129234407200303D+00
0.297077424311301D+00
0.297077424311301D+00
0.297077424311301D+00
0.297077424311301D+00
0.297077424311301D+00
0.297077424311301D+00
0.297077424311301D+00
0.500000000000000D+00
0.500000000000000D+00
0.500000000000000D+00
0.500000000000000D+00
0.500000000000000D+00
0.500000000000000D+00
0.500000000000000D+00
0.702922575688699D+00
0.702922575688699D+00
0.702922575688699D+00
0.702922575688699D+00
0.702922575688699D+00
0.702922575688699D+00
0.702922575688699D+00
0.870765592799697D+00
0.870765592799697D+00
0.870765592799697D+00
0.870765592799697D+00
0.870765592799697D+00
0.870765592799697D+00
0.870765592799697D+00
0.974553956171379D+00
0.974553956171379D+00
0.974553956171379D+00
0.974553956171379D+00
0.974553956171379D+00
0.974553956171379D+00
0.974553956171379D+00
85
v
0.100000000000000D+00
0.443649167310371D+00
0.787298334620742D+00
0.563508326896291D-01
0.250000000000000D+00
0.443649167310371D+00
0.127016653792583D-01
0.563508326896291D-01
0.100000000000000D+00
0.247985426820927D-01
0.125945902810518D+00
0.289517979151782D+00
0.487276978085690D+00
0.685035977019597D+00
0.848608053360861D+00
0.949755413489286D+00
0.221575394388360D-01
0.112532875195889D+00
0.258684799487838D+00
0.435382796399849D+00
0.612080793311860D+00
0.758232717603808D+00
0.848608053360861D+00
0.178865986691016D-01
0.908417823768389D-01
0.208822428275864D+00
0.351461287844349D+00
0.494100147412834D+00
0.612080793311860D+00
0.685035977019597D+00
0.127230219143104D-01
0.646172036001514D-01
0.148538712155651D+00
0.250000000000000D+00
0.351461287844349D+00
0.435382796399849D+00
0.487276978085690D+00
0.755944515951914D-02
0.383926248234639D-01
0.882549960354370D-01
0.148538712155651D+00
0.208822428275864D+00
0.258684799487838D+00
0.289517979151782D+00
0.328850438978473D-02
0.167015320044137D-01
0.383926248234639D-01
0.646172036001514D-01
0.908417823768389D-01
0.112532875195889D+00
0.125945902810518D+00
0.647501146528088D-03
0.328850438978473D-02
0.755944515951914D-02
0.127230219143104D-01
0.178865986691016D-01
0.221575394388360D-01
0.247985426820927D-01
Table 5.2: Computed Gauss points (p, q) and corresponding weights L for
{ n(n+1)
1} point method GQUTM.
2
n
n=5
n=9
p
0.694318442029737D-01
0.694318442029737D-01
0.694318442029737D-01
0.694318442029737D-01
0.694318442029737D-01
0.330009478207572D+00
0.330009478207572D+00
0.330009478207572D+00
0.330009478207572D+00
0.669990521792428D+00
0.669990521792428D+00
0.669990521792428D+00
0.930568155797026D+00
0.930568155797026D+00
0.198550717512319D-01
0.198550717512319D-01
0.198550717512319D-01
0.198550717512319D-01
0.198550717512319D-01
0.198550717512319D-01
0.198550717512319D-01
0.198550717512319D-01
0.198550717512319D-01
0.101666761293187D+00
0.101666761293187D+00
0.101666761293187D+00
0.101666761293187D+00
0.101666761293187D+00
0.101666761293187D+00
0.101666761293187D+00
0.101666761293187D+00
0.237233795041836D+00
0.237233795041836D+00
0.237233795041836D+00
0.237233795041836D+00
0.237233795041836D+00
0.237233795041836D+00
0.237233795041836D+00
0.408282678752175D+00
0.408282678752175D+00
0.408282678752175D+00
0.408282678752175D+00
0.408282678752175D+00
0.408282678752175D+00
0.591717321247825D+00
0.591717321247825D+00
0.591717321247825D+00
0.591717321247825D+00
0.591717321247825D+00
0.762766204958164D+00
0.762766204958164D+00
0.762766204958164D+00
0.762766204958164D+00
0.898333238706813D+00
0.898333238706813D+00
0.898333238706813D+00
0.980144928248768D+00
0.980144928248768D+00
q
0.436530238707252D-01
0.214742881469342D+00
0.465284077898513D+00
0.715825274327684D+00
0.886915131926301D+00
0.465186775265609D-01
0.221103222500738D+00
0.448887299291690D+00
0.623471844265867D+00
0.371926177849334D-01
0.165004739103786D+00
0.292816860422638D+00
0.146726751310273D-01
0.547591690719464D-01
0.156037898816279D-01
0.803566392721822D-01
0.189476014677302D+00
0.331164789916112D+00
0.490072464124384D+00
0.648980138332656D+00
0.790668913571466D+00
0.899788288976586D+00
0.964541138367140D+00
0.178364709110403D-01
0.913306309413408D-01
0.213115003430640D+00
0.366773901111335D+00
0.531559337595478D+00
0.685218235276173D+00
0.807002607765473D+00
0.880496767795773D+00
0.194093822823562D-01
0.985756383301930D-01
0.226600619520678D+00
0.381383102479082D+00
0.536165585437487D+00
0.664190566627971D+00
0.743356822675808D+00
0.199794790791376D-01
0.100234137152044D+00
0.225261107830170D+00
0.366456213417655D+00
0.491483184095780D+00
0.571737842168687D+00
0.191525719105520D-01
0.942174931981956D-01
0.204141339376088D+00
0.314065185553979D+00
0.389130106841623D+00
0.164715798970249D-01
0.782894009149582D-01
0.158944394126877D+00
0.220762215144811D+00
0.114580133114576D-01
0.508333806465933D-01
0.902087479817289D-01
0.419587036543941D-02
0.156592013857925D-01
86
L
0.191734646470685D-01
0.387333412614470D-01
0.460377090452794D-01
0.387333412614470D-01
0.191734646470685D-01
0.379971476479502D-01
0.712356204997401D-01
0.712356204997401D-01
0.379971476479502D-01
0.298908447599701D-01
0.478253516159522D-01
0.298908447599701D-01
0.603805085320831D-02
0.603805085320831D-02
0.201598349766349D-02
0.448091604484195D-02
0.646435948463322D-02
0.774766276990871D-02
0.819147462543485D-02
0.774766276990871D-02
0.646435948463322D-02
0.448091604484195D-02
0.201598349766349D-02
0.505566374507052D-02
0.111063912872568D-01
0.156674725751545D-01
0.181135411193860D-01
0.181135411193860D-01
0.156674725751545D-01
0.111063912872568D-01
0.505566374507052D-02
0.774594695636761D-02
0.167323141055649D-01
0.228415344658821D-01
0.250028228175863D-01
0.228415344658821D-01
0.167323141055649D-01
0.774594695636761D-02
0.919182785685142D-02
0.193554244975459D-01
0.251043168357703D-01
0.251043168357703D-01
0.193554244975459D-01
0.919182785685142D-02
0.877088559745420D-02
0.177185350308217D-01
0.210599120522939D-01
0.177185350308217D-01
0.877088559745420D-02
0.647199750524138D-02
0.121334570276227D-01
0.121334570276227D-01
0.647199750524138D-02
0.314010549248654D-02
0.502416878797847D-02
0.314010549248654D-02
0.502474962829402D-03
0.502474962829402D-03
87
Exact Value
GQUTM
GQUTS
IOUT
GQT
Method
77
13 13
377
388
399
3 10 10
77
88
99
10 10
44
54
77
90
Points
I1
0.4001498818
0.4000451564
0.4000006727
0.4000003534
0.4000002024
0.4000001234
0.4000037510
0.4000020035
0.4000011460
0.4000006929
0.4000016184
0.4000009417
0.4000003700
0.4000002468
0.4
I2
0.6606860757
0.6637058258
0.6664256193
0.6665011716
0.6665481694
0.6665789279
0.6659893927
0.6662009982
0.6663329015
0.6664193644
0.6662596411
0.6663718426
0.6664974532
0.6665339400
0.6666667
I3
0.8315681219
0.8501738309
0.8755247201
0.8768203350
0.8777286724
0.8783900003
0.8696444210
0.8722479925
0.8740714951
0.8753981854
0.8725338709
0.8742865042
0.8765237986
0.8772635781
0.881373587
Test example
Points
r=2
77
0.7288889289
GQT
13 13
0.7883351445
355
0.8536515855
366
0.8636423810
377
0.8699174296
IOUT
388
0.8741141888
399
0.8770583628
3 10 10
0.8792029273
Formula 1
Formula 2
55
0.8888888936 0.8189709473
66
0.8888888912 0.8386859037
77
0.8888888942 0.8511113215
GQUTS
88
0.8888888822 0.8594404172
99
0.8888888894 0.8652928433
10 10
0.8888888916 0.8695606956
14
0.8888888591 0.7979759423
44
0.8888888822 0.8620172475
GQUTM
77
0.8888888823 0.8738937177
90
0.8888889030 0.876131041
Exact
Value
0.8888888
Method
r=6
0.2209523767
0.2803986370
0.3457150818
0.3557058569
0.3619809270
0.3661776763
0.3691218499
0.3712664246
Formula 1
Formula 2
0.3809523857 0.3110344248
0.3809523670 0.3307493958
0.3809523811 0.3431748166
0.3809523803 0.3515039099
0.3809523861 0.3573563424
0.3809523859 0.3616241943
0.3809523785 0.2900394411
0.3809523803 0.3540807426
0.3809523815 0.365957215
0.3809523782 0.368194523
0.3809523
88
Points
r=2
77
0.7288889289
GQT
13 13
0.7883350849
355
0.8536515855
366
0.8636423810
377
0.8699174296
IOUT
388
0.8741141888
399
0.8770583628
3 10 10
0.8792029273
Formula 1
Formula 2
55
0.8189709473 0.8888888936
66
0.8386859037 0.8888888912
77
0.8511113215 0.8888888942
GQUTS
88
0.8594404172 0.8888888822
99
0.8652928433 0.8888888894
10 10
0.8695606956 0.8888888916
14
0.7979759423 0.8888888591
44
0.8620172475 0.8888888822
GQUTM
77
0.8738937177 0.8888888823
90
0.8761310414 0.8888889030
Exact
Value
0.8888888
Method
r=6
0.2209523618
0.2803986370
0.3457150818
0.3557058569
0.3619809270
0.3661776763
0.3691218499
0.3712664246
Formula 1
Formula 2
0.3110344248 0.3809523857
0.3307493958 0.3809523670
0.3431748166 0.3809523811
0.3515039099 0.3809523803
0.3573563424 0.3809523861
0.3616241943 0.3809523859
0.2900394411 0.3809523785
0.3540807425 0.3809523803
0.3659572152 0.3809523814
0.3681945230 0.3809523782
0.3809523
89
GQUTM
GQUTS
IOUT
GQT
Method
Points
77
13 13
355
366
377
388
399
3 10 10
55
66
77
88
99
10 10
14
44
77
90
Exact
Value
0.03694801040
90
GQUTM
GQUTS
IOUT
GQT
Method
Points
77
13 13
355
366
377
388
399
3 10 10
55
66
77
88
99
10 10
14
44
77
90
Exact
Value
0.02731723349
91
Table 5.8: The absolute error over corresponding monomials integrals for GQUTS.
N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
17
19
20
21
22
23
24
25
26
27
28
i
1
0
1
2
3
3
4
3
3
7
6
7
8
10
3
11
4
6
6
7
8
15
14
11
11
6
j
0
2
2
2
2
3
3
5
6
3
5
5
5
4
12
6
15
14
15
15
15
9
11
15
16
22
m
2
2
3
3
5
5
6
6
7
7
7
8
8
9
9
10
11
11
12
12
13
13
14
14
15
15
Formula 1
0.11102230246251565404D-15
0.69388939039072283776D-16
0.93675067702747583098D-16
0.25153490401663702869D-16
0.24286128663675299322D-16
0.11926223897340548774D-16
0.69931040125315035993D-17
0.45265440701269810120D-17
0.26834003769016234742D-17
0.76300727888667374543D-17
0.13145951341386741262D-17
0.18821072087990553534D-17
0.10693790959085541781D-17
0.34711410178481227895D-17
0.97019153778507560837D-17
0.22703129964375418776D-18
0.14525556254227182940D-17
0.72970564916977889366D-19
0.62256921623191074922D-19
0.16136639389684463881D-19
0.67353181377130278426D-20
0.26850282683471278326D-20
0.86559281224845889752D-21
0.43010806913343210882D-21
0.28096998962953544725D-21
0.87461942068294381203D-20
92
Formula 2
0.13877787807814456755D-15
0.55511151231257827021D-16
0.62450045135165055399D-16
0.25153490401663702869D-16
0.32959746043559334794D-16
0.11817803680091998331D-16
0.80773061850170080334D-17
0.35507621148900270214D-17
0.19244588561617703704D-17
0.75487576259303246218D-17
0.14196272195982073683D-17
0.22107559923337238850D-17
0.12730905197232134096D-17
0.40056188075655863035D-17
0.99340024053984343766D-17
0.21292290711604974773D-18
0.12598556299223649668D-17
0.82638651916497676830D-19
0.69430231895250930918D-19
0.19764653556342043263D-19
0.77750841676921836300D-20
0.23076271138698089560D-20
0.82382019131453099994D-21
0.46376398030668342234D-21
0.29547149974335571368D-21
0.87056623568143397642D-20
Table 5.9:
GQUTM.
N
1
2
3
4
5
6
7
9
10
11
12
13
14
15
17
19
20
21
22
23
24
24
25
25
26
26
27
27
28
28
i
0
0
1
2
3
3
4
4
7
6
7
8
10
3
11
4
6
6
15
9
15
15
17
17
11
11
11
11
6
6
j
1
2
2
2
2
3
3
5
3
5
5
5
4
12
6
15
14
15
7
14
9
9
8
8
15
15
16
16
22
22
m
3
3
4
6
7
6
8
8
8
8
10
9
10
11
13
16
23
17
17
23
15
23
17
25
15
21
16
23
16
23
Formula 1
0.52735593669694935670D-15
0.31918911957973250537D-15
0.57176485768195561832D-14
0.23210600108569678923D-14
0.81141690588815151841D-15
0.35247412627503749150D-15
0.84947240214239272404D-16
0.22020029571447052863D-12
0.24570731733952744236D-16
0.20913072536459919454D-12
0.82127035546026606883D-14
0.36685927411276306534D-13
0.19663415267925790988D-12
0.13295543797322249835D-12
0.64774877142554824301D-14
0.96761867520779067109D-17
0.17882321354416373486D-18
0.89397710137790976567D-18
0.88817939246452559470D-17
0.54887569545956151366D-20
0.37408630675467081079D-16
0.80600478887167016639D-20
0.97180131397516505553D-17
0.95833009867331021272D-20
0.52781280261133280729D-17
0.99993211360589904835D-21
0.81391527216587962374D-18
0.25275537592323451890D-21
0.45687021572263586205D-17
0.10446463955932186506D-19
93
Formula 2
0.33306690738754696213D-15
0.97144514654701197287D-16
0.13357370765021414627D-14
0.23210600108569678923D-14
0.32916377956659914616D-15
0.35247412627503749150D-15
0.35517455946806886224D-11
0.33732556136388536228D-11
0.89810173187210295964D-11
0.15063337289479821618D-10
0.15025672712483265428D-12
0.23860034588321524327D-11
0.49008535029954492651D-12
0.85906081510531140388D-17
0.13019009425257495457D-14
0.60896692589019075365D-15
0.87985547396040447721D-18
0.91583551450074603259D-16
0.65983039410497443386D-18
0.96866985633022305866D-20
0.99060573491742491661D-17
0.32247636180379783905D-20
0.33752484509766879502D-18
0.46148406374333413981D-20
0.11767352280171164067D-16
0.74633663237304036071D-20
0.26984091943001830982D-17
0.93141829443178030296D-21
0.18615208877889429484D-15
0.73087735050746025737D-18
M
2
3
4
5
6
7
8
9
10
11
12
13
14
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
Formula 1
0.13877787807814456755D-15
0.61062266354383609723D-15
0.56996074526693973894D-13
0.81185058675714572018D-15
0.70429773124658368033D-15
0.10373646386341306425D-14
0.23835100559921329477D-14
0.43966566498632175808D-14
0.65156213757688874466D-14
0.83275400464266624567D-14
0.49274820335121205517D-14
0.56465249143045070923D-14
0.34867941867133822598D-15
0.38172243144174444751D-12
0.22305768343500176343D-12
0.14502982148556498032D-12
0.10154724283673033369D-12
0.86659845965897375208D-13
0.64941108046667750386D-13
0.50034629217599047024D-13
0.39316640221276344391D-13
0.31355126828280788231D-13
0.25294870364955812647D-13
0.20589432936368723404D-13
0.18630062770252919790D-13
0.15331919761552015302D-13
0.12692104311984309106D-13
0.10559261798270824784D-13
0.96363889090511634095D-14
94
Formula2
0.13877787807814456755D-15
0.61062266354383609723D-15
0.56996074526693973894D-13
0.81532003370909933437D-15
0.70082828429463006614D-15
0.10338951916821770283D-14
0.23835100559921329477D-14
0.43957892881252291772D-14
0.65164887375068758502D-14
0.83284074081646508603D-14
0.49274820335121205517D-14
0.56473922760424954959D-14
0.34954678040932662952D-15
0.38172243144174444751D-12
0.22300217228377050560D-12
0.14503676037946888755D-12
0.10154377338977838008D-12
0.86659845965897375208D-13
0.64941108046667750386D-13
0.50029425047171116603D-13
0.39316640221276344391D-13
0.31356861551756765039D-13
0.25291400918003859033D-13
0.20588999255499729202D-13
0.18630062770252919790D-13
0.15332353442421009504D-13
0.12692971673722297510D-13
0.10559261798270824784D-13
0.96350878664441808041D-14
M
3
4
6
8
9
10
11
13
14
15
16
17
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
Formula 1
0.52735593669694935670D-15
0.34833247397614286456D-13
0.33834046675451645569D-13
0.40010356139319469548D-11
0.64681315858905463756D-11
0.12284062655965044542D-11
0.90869672897397890665D-12
0.87030209428018423523D-13
0.88775341244851091460D-13
0.96405956134804071667D-13
0.60945692936797968287D-12
0.28793634143653434876D-12
0.17497808757482857800D-12
0.11922407505693399798D-12
0.86691070988464957736D-13
0.75148220979315283330D-13
0.57768026473503653051D-13
0.45427203665404647381D-13
0.36325976948692328961D-13
0.29435655302112451182D-13
0.24149952870811119965D-13
0.21992824228433960343D-13
0.18524244638218334558D-13
0.17214962094724839403D-13
0.15642001582882869570D-13
0.15574347367319774094D-13
0.16233108607321966588D-13
0.17945714358980069392D-13
0.21206127132078478326D-13
95
Formula2
0.52735593669694935670D-15
0.34833247397614286456D-13
0.33834046675451645569D-13
0.40010356139319469548D-11
0.64681315858905463756D-11
0.12284062655965044542D-11
0.90869846369745488346D-12
0.87030209428018423523D-13
0.88775341244851091460D-13
0.96405956134804071667D-13
0.60945692936797968287D-12
0.28793634143653434876D-12
0.17497114868092467077D-12
0.11922407505693399798D-12
0.86691070988464957736D-13
0.75148220979315283330D-13
0.57768026473503653051D-13
0.45428938388880624188D-13
0.36325976948692328961D-13
0.29434787940374462778D-13
0.24149952870811119965D-13
0.21992824228433960343D-13
0.18524244638218334558D-13
0.17214962094724839403D-13
0.15642001582882869570D-13
0.15574347367319774094D-13
0.16233108607321966588D-13
0.17945714358980069392D-13
0.21206127132078478326D-13
M
2
3
4
5
6
7
8
9
10
11
12
13
14
15
POINTS
4
9
16
25
36
49
64
81
100
121
144
169
196
225
I22 2
0.274994667299803D-02
0.325156766835848D-02
0.326147362431669D-02
0.326151351705471D-02
0.326151369867892D-02
0.326151374374570D-02
0.326151362742850D-02
0.326151380595353D-02
0.326151374714880D-02
0.326151374399900D-02
0.326151376422602D-02
0.326151371351188D-02
0.326151391384561D-02
0.326151371024150D-02
I32 4
0.177094483171524D-03
0.223692365249628D-03
0.230577406022251D-03
0.227783881738088D-03
0.227781427793912D-03
0.227781433987425D-03
0.227781420418245D-03
0.227781441200401D-03
0.227781439860638D-03
0.227781433283100D-03
0.227781436303126D-03
0.227781431639375D-03
0.227781448332771D-03
0.227781430014434D-03
I52 3
0.546336186031229D-04
0.128392351031521D-03
0.111603259040227D-03
0.113382200902182D-03
0.113363432955252D-03
0.113363393267416D-03
0.113363383742831D-03
0.113363402109816D-03
0.113363382009822D-03
0.113363396969313D-03
0.113363395315162D-03
0.113363385540465D-03
0.113363408869501D-03
0.113363387592995D-03
2
Table 5.12: The values of integrals Ipq
using GQUTS
I82 8
0.130239293785839D-07
0.206906550387037D-06
0.116184871832071D-06
0.149778923092289D-06
0.140874043849123D-06
0.142527331787087D-06
0.142336737470806D-06
0.142346694835904D-06
0.142346752435195D-06
0.142346764992089D-06
0.142346772794248D-06
0.142346743090658D-06
0.142346775552142D-06
0.142346747743832D-06
96
M
3
4
5
6
7
8
9
10
11
12
13
14
15
16
POINTS
5
9
14
20
27
35
44
54
65
77
90
104
119
135
I22 2
0.275485757590948D-02
0.325158889113563D-02
0.326147298020953D-02
0.326151343707490D-02
0.326151371810622D-02
0.326151376345027D-02
0.326151359160585D-02
0.326151374400892D-02
0.326151364663877D-02
0.326151364910581D-02
0.326151370828131D-02
0.326151374783817D-02
0.326151391876362D-02
0.326151371009277D-02
I32 4
0.179182452053709D-03
0.223687133177605D-03
0.230575241486151D-03
0.227783627665424D-03
0.227781398791806D-03
0.227781436381915D-03
0.227781415052045D-03
0.227781429756894D-03
0.227781415126699D-03
0.227781414262975D-03
0.227781424787914D-03
0.227781435253389D-03
0.227781450366441D-03
0.227781424592687D-03
I52 3
0.545678818403744D-04
0.128396027303604D-03
0.111603887708712D-03
0.113382277228293D-03
0.113363444919348D-03
0.113363396118669D-03
0.113363381437413D-03
0.113363399280595D-03
0.113363379195043D-03
0.113363391444485D-03
0.113363391961321D-03
0.113363387385158D-03
0.113363411069091D-03
0.113363391435240D-03
2
Table 5.13: The values of integrals Ipq
using GQUTM
I82 8
0.171783947167217D-07
0.206903489038316D-06
0.115385158234197D-06
0.149694890639606D-06
0.140862213129016D-06
0.142525165812834D-06
0.142336270369235D-06
0.142346588190201D-06
0.142346683853969D-06
0.142346730172438D-06
0.142346767351738D-06
0.142346747351104D-06
0.142346795007845D-06
0.142346742313464D-06
97
Chapter 6
Accurate Evaluation
Schemes for Triangular
Domain Integrals
Chapter 6
Accurate Evaluation Schemes for
Triangular Domain Integrals
In this chapter an alternative higher order Gaussian quadrature formula for numerical integration over the triangular surfaces is derived using Standard Triangle
in the geometrical transformation. It presents n n points (for n > 1) and
n(n+1)
1 points (for n > 2) Gaussian quadrature formulae for triangle utilizing
2
1
n-point one dimensional Gaussian quadrature formulae. The proposed n(n+1)
2
points formulae reduces the total number of Gaussian points in the calculation
process and completely avoids the crowding of Gaussian points as well. To verify
the efficiency and accuracy, the method is applied to evaluate triangular domain
integrals of different types of functions and to the evaluation of integrals for element stiffness matrices in two dimensional elasticity problem. Thus, it overcomes
all the drawbacks in view of accuracy and efficiency for the numerical evaluation
of the triangular domain integrals of any arbitrary functions encountered in the
realm of science and engineering.
6.1
Introduction
101
6.2
Problem Statement
The solution of any physical phenomena using finite element method requires the
accurate evaluation of components of element matrices. This thesis intents to
derive accurate methods to evaluate the associated domain integrals for the accurate evaluation of the element matrices. Consider the two dimensional triangular
domain integral:
ZZ
f (x, y) dx dy;
I=
: triangle (arbitrary)
(6.1)
x
y
!
=
x1 x2 x3
y1 y2 y3
102
N1
N2
N3
(6.2)
1
N1 (s, t) = (s + 1),
2
1
N3 (s, t) = (s + t)
2
(6.3)
From Eq.(6.2) and Eq.(6.3), we obtain the transformation equation from the
arbitrary triangle into a standard triangle as:
1
1
1
x(s, t) = (x1 + x2 ) + (x1 x3 )s + (x2 x3 )t
2
2
2
1
1
1
y(s, t) = (y1 + y2 ) + (y1 y3 )s + (y2 y3 )t
2
2
2
(6.4)
(6.5)
s = 1
(6.6)
(6.7)
t = 1
t = 1
s = 1
The evaluation of integrals I in Eq.(6.6) and in Eq.(6.7) by the existing Gaussian quadrature (i.e. 7-point and 13-point) will yield the same results. Thus, any
one of these two can be evaluated numerically. Influences of these integrals will
be studied later to present new quadrature formulae for triangles.
6.3
t=
1
1
1
(x1 + x2 ) + (x1 x3 ) + (x2 x3 )( 1)
2
2
4
1
1
1
y=
(y1 + y2 ) + (y1 y3 ) + (y2 y3 )( 1)
2
2
4
x=
(6.9)
Now the determinant of the Jacobian and the differential area are:
s t
s t
1
(s, t)
=
= (1 )
(,
2
ds dt = dt ds =
(s, t)
1
d d = (1 ) d d
(,
2
(6.10)
(6.11)
(1 )(1 + )
(1 )(1 + )
1
1), y(,
1)}
dd
2
2
8
1 1
Z 1Z 1
(1 )(1 + )
1
1}
dd
(6.12)
= J
f {,
2
8
1 1
Z
I = J
f {x(,
6.3.1
I = J
m X
n
X
(m)
(m)
f {i ,
(1 i
i=1 j=1
= J{
mn
X
(n)
(m)
)(1 + j )
1 i
(m)
(n)
1}(
)Wi Wj
2
8
f (sr , tr )Gr }
(6.13)
r=1
where (sr , tr ) are the new Gaussian points, Gr is the corresponding weights for
(m)
(n)
standard triangle where i , i = 1, m; j , j = 1, n are the m-th and n-th order
(m)
(n)
Gaussian quadrature points respectively. And Wi , Wj are the corresponding
m-th and n-th order weights respectively.
Again, if we consider the integral I of Eq.(6.7) and substitute
s=
(1 + )(1 )
1,
2
t=
I = J
= J{
f(
1 1
mn
X
(1 + )(1 )
1
1, )
d d
2
8
(6.14)
r=1
where G0r and (s0r , t0r ) are respectively weights and Gaussian points for triangle.
All the Gaussian points and corresponding weights can be calculated simply using
the following algorithm:
step 1. r 1
step 2. i = 1, m
step 3. j = 1, n
105
(n)
(m)
(1 i )(1 + j )
(1 i ) (m) (n)
(m)
Gr =
Wi Wj , sr = i ,
tr =
1
8
2
(n)
(m)
(n)
(1 j ) (m) (n)
(1 + i )(1 j )
(n)
0
0
Gr =
Wi Wj , sr =
1, t0r = j
8
2
r =r+1
step 4. compute step 3
step 5. compute step 2
Computed Gauss points and weights for different values of n( = m) based on
above algorithm listed in Table 6.1 and the distribution of Gaussian points for
m = n = 10 is plotted in Figure 6.2. It is seen that there are crowding of Gauss
points at the terminal points in the s-direction within the triangle and that is
one of the major causes of error germane in the calculation. We know, the Gauss
quadrature formulae enforces the evaluation of the function values at each Gauss
points within the transformed domain. Thus, increasing number of Gauss points
causes a severe problem due to crowding of Gaussian points. To avoid such type
of crowding further modification is needed and that will be described in the next
section.
6.3.2
It is clearly noticed in the Eq.(6.13) that for each i (i = 1, 2,..., m) j varies from
1 to n and hence at the terminal value of i = m-1, m(say) there are respectively
n-1, n crowding points as shown in Table 6.1 and Figure 6.2. To overcome this
situation, algebraic manipulation is done in this section. We can use the advantage
of Eq.(6.12) by making j dependent on i for the calculation of new Gauss points
and corresponding weights. To do so, we wish to calculate Gauss points and
weights for i = 1, m-1 and j = 1, n + 1 - i. The number of total Gaussian
quadrature points within the standard triangle is then given by
n(n + 1) (n m + 1)(n m + 2)
;
nm
2
2
The new Gaussian quadrature formulae for Eq.(6.12) is then
TP =
106
I = J
m1
X n+1i
X
i=1
(m1)
(m1)
f {i
,
(1 i
j=1
(m1)
1 i
(
8
(m1)
(n+1i)
)Wi
Wj
(n+1i)
)(1 + j
2
= J{
TP
X
1}
f (pr , qr )Lr }
(6.15)
r=1
where (pr , qr ) are the new Gaussian points, Lr is the corresponding weights
for triangles.
Similarly, for Eq.(6.7) the new Gaussian quadrature formulae can be written
as
107
I = J
m1
X n+1j
X
j=1
(n+1j)
f{
(m1)
)(1 j
2
(1 + i
i=1
(m1)
1 j
{
8
(m1)
(n+1j)
Wj
Wi
= J{
(m1)
1, j
TP
X
(6.16)
r=1
where L0r and (p0r , qr0 ) are respectively weights and Gaussian points for triangle.
All the Gaussian points and corresponding weights can be calculated simply using
the following algorithm:
step 1. r 1
step 2. i = 1, m 1
step 3. j = 1, n + 1 i
(m1)
pr =
(m1)
,
i
qr =
(1 i
(n+1i)
)(1 + j
2
(m1)
) (m1) (n+1i)
(1 i
Wi
Wj
8
step 4. j = 1, m 1
Lr =
step 5. i = 1, n + 1 j
(n+1j)
(m1)
)(1 j
)
=
1,
2
(n+1j)
) (n+1j) (m1)
(1 i
L0r =
Wi
Wj
8
r =r+1
p0r
(1 + i
(m1)
qr0 = j
Figure 6.3: Distribution of Gauss Points for m=6, n=8 (30 points) GQSTM.
109
Figure 6.4: Distribution of Gauss Points for m=n=10 (54 points) GQSTM.
6.4
Application Examples
To show the accuracy, efficiency of the derived method and to compare the results
between the derived formulae of this thesis we are following the same examples
(of previous Chapter) with known results:
1y
I1 =
(x + y) 2 dx dy
y=0
1
x=0
1y
y=0
(x + y) 2 dx dy
I2 =
x=0
110
(x2 + y 2 ) 2 dx dy
I3 =
y=0
1
x=0
y
exp|x+y1| dx dy
I4 =
y=0
x=0
Computed values are summarized in Table 6.3 for m m points GQSTS and
1 points GQSTM formulae.
Some important remarks from the Table 6.3 are:
m(m+1)
2
Usual Gauss quadrature (GQT) for triangles e.g. 7-point and 13point rules cannot evaluate the integral of non-polynomial functions
accurately.
Splitting standard triangle into quadrilaterals (IOST) provides a way
of using Gaussian quadrature for square. But it need comparatively
a large number of Gauss points whereas the convergence rate is slow
but satisfactory in view of accuracy.
New Gaussian quadrature formulae for triangle (GQSTS and GQSTM)
are exact in view of accuracy and efficiency. Also (GQSTM) is faster
and needs significantly lesser number of Gauss point.
Again, we consider the following integrals of rational functions to test the
influences of formulae in Eq.(6.13), Eq.(6.14), Eq.(6.15) and Eq.(6.16). Consider
I
p,q
1y
y=0
Example-1: I
r,0
x=0
Example-2: I
0,r
Example-3: I
1y
=
y=0
x=0
1y
=
y=0
1y
=
y=0
0,0
xp y q
dx dy
+ x + y
x=0
x=0
xr
dx dy
0.375 0.375 x
yr
dx dy
0.375 0.375 y
1
dx dy
12 + 21.53679831x 8.821067231y
111
Example-4: I
0,0
1y
=
y=0
x=0
1
dx dy
12 + 9.941125498(x + y)
Results are summarized in Tables 6.4, 6.5, 6.6, 6.7 for m m points GQSTS
and m(m+1)
1 points GQSTM formulae.
2
In Tables 6.4, 6.5, 6.6, 6.7 for method GQSTS, Formula 1 is for Eq.(6.13) and
Formula 2 is for Eq.(6.14), for method GQSTM, Formula 1 is for Eq(6.15) and
Formula 2 is for Eq.(6.16). These tables substantiated the influences of numerical
evaluation of the integrals as described before. Some important comments may
be drawn from the Tables 6.4, 6.5, 6.6, 6.7.
xr
with 6= = 0 first formulae described in
For the integrand +x+y
Eq.(6.13) and Eq.(6.15) are more accurate and rate of convergence is higher. But
the new formula in Eq.(6.15) requires very less computational effort.
yr
Similarly for the integrand +x+y
with 6= = 0 second formula described
in Eq.(6.14) and Eq.(6.16) are more accurate and convergence is higher. Here also
the new formula in Eq.(6.16) requires very less computational effort.
Similar influences of these formulae in procedure-3 may be observed for
different conditions on , .
General Gaussian quadrature e.g. 7-point and 13-point rules and IOST
cannot evaluate all these integrals of rational functions accurately.
It is evident that the new formulae e.g. Eq.(6.15) and Eq.(6.16) are very fast
and accurate in view of accuracy and equally applicable for any geometry that is
for different values of , and . We recommend this is appropriate quadrature
scheme for triangular domain integrals encountered in science and engineering.
6.4.1
112
II
pq
xp y q
dx dy
J(x, y)
U
Z 1 Z 1x p q
xy
=
dx dy
x=0 y=0 J(x, y)
ZZ
(6.17)
Where J is the Jacobean of transformation and U is the unit triangle: {(x, y)|0
x, y 1, x + y 1}. For cubic curved triangular element (m=10, Fig 6.5) the
transformation which maps a one side curved triangle A0 in to an orthogonal unit
triangle U is given by
t=
10
X
Ni (x, y)ti ,
t = x0 , y 0 and
ti = (x0i , yi0 )
i=1
10
X
Ni (x, y)ti ,
t = x0 , y 0 and
i=1
113
ti = (x0i , yi0 )
9
x0 = x03 + (x01 x03 )x + (x02 x03 )y + (x01 x02 + x04 + x05 )xy
4
9
y 0 = y30 + (y10 y30 )x + (y20 y30 )y + (y10 y20 + y40 + y50 )xy
4
Using Appropriate transformation we have the relation between A0 and the
standard triangle S with vertices 1(1, 1), 2(1, 1), 3(1, 1) given by
x+1
y+1
(1 + R(
))
2
2
y+1
x+1
y0 =
(1 + R(
))
2
2
9
R =
(t1 t2 + t4 + t5 )
4
1
R
J(x, y) =
(1 + (x + y + 2))
4
2
x0 =
By choosing x04 = 0.8660254, y40 = 0.5 and x05 = 0.5, y50 = 0.8660254, we have
computed the values of the integrals of the Eq.(6.17) for different values of p, q
using m m points GQSTS and m(m+1)
1 points GQSTM formulae and the
2
results are summarized in Table 6.8 and 6.9.
We know the m-th order Gaussian quadrature formula can evaluate the integral
of polynomial of degree up to 2m-1 accurately. The presented methods of this
114
6.5
Conclusions
This Chapter presented new techniques to derive quadrature formulae utilizing the
one dimensional Gaussian quadrature points and weights and that overcomes all
the difficulties pertinent in the derivation of the higher order formulae. The first
technique (GQSTS) presents m n point quadrature formulae for the triangles
utilizing the one dimensional quadrature points. Finally, in the second technique
(nm+1)(nm+2)
; n m points quadrature formulae are de(GQSTM), n(n+1)
2
2
rived utilizing the one dimensional Gaussian quadrature formulae. For m = n,
1 quadrature points are calculated by the formulae GQSTS
m m and m(m+1)
2
and GQSTM respectively. It is observed that the scheme GQSTM is appropriate
for the triangular domain integrals as it requires less computational effort for desired accuracy. Through practical application examples, it is demonstrated that
the new appropriate Gaussian quadrature formulae for triangles are accurate in
view of accuracy and efficiency and hence a proper balance is observed. Computer
codes in FORTRAN for these two formulae are appended (Program 4: GQSTS.for
and Program 5: GQSTM.for).
Thus, we believe that the newly derived appropriate quadrature formulae for
triangles will ensure the accurate evaluation of the integrals in an efficient manner
and enhance the further utilization of triangular elements for numerical solution
of field problems in science and engineering.
115
Table 6.1: Computed Gauss points (u, v) and corresponding weights G for n n
point method (GQSTS).
n
u
-0.77459666924148340428D+00
-0.77459666924148340428D+00
-0.77459666924148340428D+00
0.00000000000000000000D+00
0.00000000000000000000D+00
0.00000000000000000000D+00
0.77459666924148340428D+00
0.77459666924148340428D+00
0.77459666924148340428D+00
-0.94910791234275848627D+00
-0.94910791234275848627D+00
-0.94910791234275848627D+00
-0.94910791234275848627D+00
-0.94910791234275848627D+00
-0.94910791234275848627D+00
-0.94910791234275848627D+00
-0.74153118559939446008D+00
-0.74153118559939446008D+00
-0.74153118559939446008D+00
-0.74153118559939446008D+00
-0.74153118559939446008D+00
-0.74153118559939446008D+00
-0.74153118559939446008D+00
-0.40584515137739718416D+00
-0.40584515137739718416D+00
-0.40584515137739718416D+00
-0.40584515137739718416D+00
-0.40584515137739718416D+00
-0.40584515137739718416D+00
-0.40584515137739718416D+00
0.00000000000000000000D+00
0.00000000000000000000D+00
0.00000000000000000000D+00
0.00000000000000000000D+00
0.00000000000000000000D+00
0.00000000000000000000D+00
0.00000000000000000000D+00
0.40584515137739718416D+00
0.40584515137739718416D+00
0.40584515137739718416D+00
0.40584515137739718416D+00
0.40584515137739718416D+00
0.40584515137739718416D+00
0.40584515137739718416D+00
0.74153118559939446008D+00
0.74153118559939446008D+00
0.74153118559939446008D+00
0.74153118559939446008D+00
0.74153118559939446008D+00
0.74153118559939446008D+00
0.74153118559939446008D+00
0.94910791234275848627D+00
0.94910791234275848627D+00
0.94910791234275848627D+00
0.94910791234275848627D+00
0.94910791234275848627D+00
0.94910791234275848627D+00
0.94910791234275848627D+00
v
-0.80000000000000004441D+00
-0.11270166537925829786D+00
0.57459666924148344869D+00
-0.88729833462074170214D+00
-0.50000000000000000000D+00
-0.11270166537925829786D+00
-0.97459666924148335987D+00
-0.88729833462074170214D+00
-0.80000000000000004441D+00
-0.95040291463581461073D+00
-0.74810819437896391992D+00
-0.42096404169643553139D+00
-0.25446043828620812377D-01
0.37007195403919368459D+00
0.69721610672172240619D+00
0.89951082697857298598D+00
-0.95568492112232794611D+00
-0.77493424960822177017D+00
-0.48263040102432486211D+00
-0.12923440720030276996D+00
0.22416158662371921118D+00
0.51646543520761634127D+00
0.69721610672172240619D+00
-0.96422680266179672248D+00
-0.81831643524632213804D+00
-0.58235514344827121747D+00
-0.29707742431130146343D+00
-0.11799705174331709401D-01
0.22416158662371921118D+00
0.37007195403919368459D+00
-0.97455395617137918762D+00
-0.87076559279969723004D+00
-0.70292257568869853657D+00
-0.50000000000000000000D+00
-0.29707742431130146343D+00
-0.12923440720030276996D+00
-0.25446043828620812377D-01
-0.98488110968096176379D+00
-0.92321475035307232204D+00
-0.82349000792912596669D+00
-0.70292257568869853657D+00
-0.58235514344827121747D+00
-0.48263040102432486211D+00
-0.42096404169643553139D+00
-0.99342299122043054016D+00
-0.96659693599117268992D+00
-0.92321475035307232204D+00
-0.87076559279969723004D+00
-0.81831643524632213804D+00
-0.77493424960822177017D+00
-0.74810819437896391992D+00
-0.99870499770694387554D+00
-0.99342299122043054016D+00
-0.98488110968096176379D+00
-0.97455395617137918762D+00
-0.96422680266179672248D+00
-0.95568492112232794611D+00
-0.95040291463581461073D+00
116
G
0.27385751068541380748D+00
0.43817201709666231402D+00
0.27385751068541380748D+00
0.24691358024691342776D+00
0.39506172839506170646D+00
0.24691358024691342776D+00
0.34784464623227831503D-01
0.55655143397164555386D-01
0.34784464623227831503D-01
0.16339719022330379677D-01
0.35296047419167315973D-01
0.48183166927650789013D-01
0.52742305350881341086D-01
0.48183166927650789013D-01
0.35296047419167315973D-01
0.16339719022330379677D-01
0.31537077511009203867D-01
0.68124438478366217131D-01
0.92997698922885035921D-01
0.10179723223434195001D+00
0.92997698922885035921D-01
0.68124438478366217131D-01
0.31537077511009203867D-01
0.34753371619033066042D-01
0.75072077491965857887D-01
0.10248202577596891849D+00
0.11217897537887257242D+00
0.10248202577596891849D+00
0.75072077491965857887D-01
0.34753371619033066042D-01
0.27059715378963785509D-01
0.58452718547963133933D-01
0.79794688105559577473D-01
0.87344939608496471872D-01
0.79794688105559577473D-01
0.58452718547963133933D-01
0.27059715378963785509D-01
0.14687879552880067288D-01
0.31727846266938745934D-01
0.43312161692772820076D-01
0.47410400832246554770D-01
0.43312161692772820076D-01
0.31727846266938745934D-01
0.14687879552880067288D-01
0.46805656432302384165D-02
0.10110667549803324661D-01
0.13802224836019571369D-01
0.15108204861584317857D-01
0.13802224836019571369D-01
0.10110667549803324661D-01
0.46805656432302384165D-02
0.42663744142294604891D-03
0.92159573507212338376D-03
0.12580842442623490877D-02
0.13771254070462312334D-02
0.12580842442623490877D-02
0.92159573507212338376D-03
0.42663744142294604891D-03
Table 6.2: Computed Gauss points (p, q) and corresponding weights L for
1 point method GQSTM.
n
n=5
n=9
L
0.76693858588273891574D-01
0.15493336504578786306D+00
0.18415083618111743879D+00
0.15493336504578786306D+00
0.76693858588273891574D-01
0.15198859059180083730D+00
0.28494248199896055063D+00
0.28494248199896055063D+00
0.15198859059180083730D+00
0.11956337903988054072D+00
0.19130140646380897063D+00
0.11956337903988054072D+00
0.24152203412833226498D-01
0.24152203412833226498D-01
0.80639339906539760777D-02
0.17923664179367811145D-01
0.25857437938532890770D-01
0.30990651079634820575D-01
0.32765898501739394721D-01
0.30990651079634820575D-01
0.25857437938532890770D-01
0.17923664179367811145D-01
0.80639339906539760777D-02
0.20222654980282091242D-01
0.44425565149027380907D-01
0.62669890300618077972D-01
0.72454164477543922529D-01
0.72454164477543922529D-01
0.62669890300618077972D-01
0.44425565149027380907D-01
0.20222654980282091242D-01
0.30983787825470423399D-01
0.66929256422259622572D-01
0.91366137863528443241D-01
0.10001129127034502631D+00
0.91366137863528443241D-01
0.66929256422259622572D-01
0.30983787825470423399D-01
0.36767311427405684199D-01
0.77421697990183765969D-01
0.10041726734308135949D+00
0.10041726734308135949D+00
0.77421697990183765969D-01
0.36767311427405684199D-01
0.35083542389816799179D-01
0.70874140123286688553D-01
0.84239648209175455040D-01
0.70874140123286688553D-01
0.35083542389816799179D-01
0.25887990020965513660D-01
0.48533828110490902230D-01
0.48533828110490902230D-01
0.25887990020965513660D-01
0.12560421969946170689D-01
0.20096675151913882124D-01
0.12560421969946170689D-01
0.20098998513176078752D-02
0.20098998513176078752D-02
p
-0.86113631159405257254D+00
-0.86113631159405257254D+00
-0.86113631159405257254D+00
-0.86113631159405257254D+00
-0.86113631159405257254D+00
-0.33998104358485631282D+00
-0.33998104358485631282D+00
-0.33998104358485631282D+00
-0.33998104358485631282D+00
0.33998104358485631282D+00
0.33998104358485631282D+00
0.33998104358485631282D+00
0.86113631159405257254D+00
0.86113631159405257254D+00
-0.96028985649753628717D+00
-0.96028985649753628717D+00
-0.96028985649753628717D+00
-0.96028985649753628717D+00
-0.96028985649753628717D+00
-0.96028985649753628717D+00
-0.96028985649753628717D+00
-0.96028985649753628717D+00
-0.96028985649753628717D+00
-0.79666647741362683899D+00
-0.79666647741362683899D+00
-0.79666647741362683899D+00
-0.79666647741362683899D+00
-0.79666647741362683899D+00
-0.79666647741362683899D+00
-0.79666647741362683899D+00
-0.79666647741362683899D+00
-0.52553240991632899082D+00
-0.52553240991632899082D+00
-0.52553240991632899082D+00
-0.52553240991632899082D+00
-0.52553240991632899082D+00
-0.52553240991632899082D+00
-0.52553240991632899082D+00
-0.18343464249564978008D+00
-0.18343464249564978008D+00
-0.18343464249564978008D+00
-0.18343464249564978008D+00
-0.18343464249564978008D+00
-0.18343464249564978008D+00
0.18343464249564978008D+00
0.18343464249564978008D+00
0.18343464249564978008D+00
0.18343464249564978008D+00
0.18343464249564978008D+00
0.52553240991632899082D+00
0.52553240991632899082D+00
0.52553240991632899082D+00
0.52553240991632899082D+00
0.79666647741362683899D+00
0.79666647741362683899D+00
0.79666647741362683899D+00
0.96028985649753628717D+00
0.96028985649753628717D+00
117
n(n+1)
q
-0.91269395225854965670D+00
-0.57051423706131632052D+00
-0.69431844202973769242D-01
0.43165054865536878204D+00
0.77383026385260200719D+00
-0.90696264494687817148D+00
-0.55779355499852401046D+00
-0.10222540141661973223D+00
0.24694368853173420675D+00
-0.92561476443013324111D+00
-0.66999052179242812866D+00
-0.41436627915472301620D+00
-0.97065464973794535197D+00
-0.89048166185610722057D+00
-0.96879242023674416728D+00
-0.83928672145563554263D+00
-0.62104797064539618567D+00
-0.33767042016777559787D+00
-0.19855071751231800903D-01
0.29796027666531199607D+00
0.58133782714293258387D+00
0.79957657795317182980D+00
0.92908227673428056548D+00
-0.96432705817791941616D+00
-0.81733873811731838632D+00
-0.57376999313871901442D+00
-0.26645219777733009092D+00
0.63118675190956929910D-01
0.37043647055234596444D+00
0.61400521553094544736D+00
0.76099353559154647719D+00
-0.96118123543528766195D+00
-0.80284872333961398994D+00
-0.54679876095864454744D+00
-0.23723379504183550459D+00
0.72331170874973427232D-01
0.32838113325594298075D+00
0.48671364535161654175D+00
-0.96004104184172489589D+00
-0.79953172569591102192D+00
-0.54947778433966076861D+00
-0.26708757316468956233D+00
-0.17033631808439197997D-01
0.14347568433737456495D+00
-0.96169485617889594042D+00
-0.81156501360360888686D+00
-0.59171732124782483453D+00
-0.37186962889204100424D+00
-0.22173978631675383966D+00
-0.96705684020595017447D+00
-0.84342119817008365246D+00
-0.68211121174624533836D+00
-0.55847556971037892737D+00
-0.97708397337708474684D+00
-0.89833323870681347501D+00
-0.81958250403654209215D+00
-0.99160825926912121719D+00
-0.96868159722841506998D+00
118
Exact Value
GQSTM
GQSTS
IOST
GQT
Method
Test example
I1
I2
I3
I4
0.4001498818 0.6606860757 0.8315681219 0.6938790083
0.4000451564 0.66370582580 0.85017383098 0.72387170791
0.4000006727 0.6664256193 0.8755247201 0.7178753416
0.4000003534 0.6665011716 0.8768203350 0.7179654297
0.4000002024 0.6665481694 0.8777286724 0.7180285699
0.4000001234 0.6665789279 0.8783900003 0.7180745324
0.4000037510 0.6659893927 0.8696444210 0.7184323903
0.4000020035 0.6662009982 0.8722479925 0.7185688310
0.4000011460 0.6663329015 0.8740714951 0.7181265336
0.4000006929 0.6664193644 0.8753981854 0.7182531970
0.4000016184 0.6662596411 0.8725338709 0.7180036731
0.4000009417 0.6663718426 0.8742865042 0.7175459725
0.4000003700 0.6664974532 0.8765237986 0.7179128710
0.4000002468 0.6665339400 0.8772635781 0.7180958213
0.4
0.6666667
0.881373587
0.71828183
Points
77
13 13
377
388
399
3 10 10
77
88
99
10 10
44
54
77
90
Points
r=2
77
0.7288889289
GQT
13 13
0.7883351445
355
0.8536515855
366
0.8636423810
377
0.8699174296
IOST
388
0.8741141888
399
0.8770583628
3 10 10
0.8792029273
Formula 1
Formula 2
55
0.8888888936 0.8189709473
66
0.8888888912 0.8386859037
77
0.8888888942 0.8511113215
GQSTS
88
0.8888888822 0.8594404172
99
0.8888888894 0.8652928433
10 10
0.8888888916 0.8695606956
14
0.8888888591 0.7979759423
44
0.8888888822 0.8620172475
GQSTM
77
0.8888888823 0.8738937177
90
0.8888889030 0.876131041
Exact
Value
0.8888888
Method
r=6
0.2209523767
0.2803986370
0.3457150818
0.3557058569
0.3619809270
0.3661776763
0.3691218499
0.3712664246
Formula 1
Formula 2
0.3809523857 0.3110344248
0.3809523670 0.3307493958
0.3809523811 0.3431748166
0.3809523803 0.3515039099
0.3809523861 0.3573563424
0.3809523859 0.3616241943
0.3809523785 0.2900394411
0.3809523803 0.3540807426
0.3809523815 0.3659572152
0.3809523782 0.3681945230
0.3809523
119
Points
r=2
77
0.7288889289
GQT
13 13
0.7883350849
355
0.8536515855
366
0.8636423810
377
0.8699174296
IOST
388
0.8741141888
399
0.8770583628
3 10 10
0.8792029273
Formula 1
Formula 2
55
0.8189709473 0.8888888936
66
0.8386859037 0.8888888912
77
0.8511113215 0.8888888942
GQSTS
88
0.8594404172 0.8888888822
99
0.8652928433 0.8888888894
10 10
0.8695606956 0.8888888916
14
0.7979759423 0.8888888591
44
0.8620172475 0.8888888822
GQSTM
77
0.8738937177 0.8888888823
90
0.8761310414 0.8888889030
Exact
Value
0.8888888
Method
r=6
0.2209523618
0.2803986370
0.3457150818
0.3557058569
0.3619809270
0.3661776763
0.3691218499
0.3712664246
Formula 1
Formula 2
0.3110344248 0.3809523857
0.3307493958 0.3809523670
0.3431748166 0.3809523811
0.3515039099 0.3809523803
0.3573563424 0.3809523861
0.3616241943 0.3809523859
0.2900394411 0.3809523785
0.3540807425 0.3809523803
0.3659572152 0.3809523814
0.3681945230 0.3809523782
0.3809523
120
GQSTM
GQSTS
IOST
GQT
Method
Points
77
13 13
355
366
377
388
399
3 10 10
55
66
77
88
99
10 10
14
44
77
90
Exact
Value
0.03694801040
121
GQSTM
GQSTS
IOST
GQT
Method
Points
77
13 13
355
366
377
388
399
3 10 10
55
66
77
88
99
10 10
14
44
77
90
Exact
Value
0.02731723349
122
M POINTS
I22 2
I32 4
I52 3
I82 8
3
9
0.325947151477D-02 0.224256520359D-03 0.128714495542D-03 0.207440393085D-06
4
16
0.326936162780D-02 0.231153940689D-03 0.111884992164D-03 0.116484848575D-06
5
25
0.326940113312D-02 0.228354076715D-03 0.113667897705D-03 0.150166858783D-06
6
36
0.326940130900D-02 0.228351635765D-03 0.113649141988D-03 0.141238492500D-06
7
49
0.326940135410D-02 0.228351642100D-03 0.113649102677D-03 0.142896199680D-06
8
64
0.326940123749D-02 0.228351628498D-03 0.113649093133D-03 0.142705081674D-06
9
81
0.326940141646D-02 0.228351649332D-03 0.113649111547D-03 0.142715068831D-06
10
100
0.326940135750D-02 0.228351647989D-03 0.113649091396D-03 0.142715126336D-06
11
121
0.326940135435D-02 0.228351641395D-03 0.113649106393D-03 0.142715138925D-06
12
144
0.326940137462D-02 0.228351644422D-03 0.113649104734D-03 0.142715146747D-06
13
169
0.326940132378D-02 0.228351639748D-03 0.113649094935D-03 0.142715116967D-06
14
196
0.326940152460D-02 0.228351656481D-03 0.113649118323D-03 0.142715149511D-06
15
225
0.326940132050D-02 0.228351638118D-03 0.113649096993D-03 0.142715121632D-06
2
Table 6.8: The values of integrals Ipq
using GQSTS
123
M POINTS
I22 2
I32 4
I52 3
I82 8
4
9
0.325949245527D-02 0.224251251839D-03 0.128718170385D-03 0.207437326110D-06
5
14
0.326936098818D-02 0.231151779323D-03 0.111885618075D-03 0.115682730958D-06
6
20
0.326940105419D-02 0.228353823362D-03 0.113667973621D-03 0.150082546398D-06
7
27
0.326940132861D-02 0.228351606865D-03 0.113649153897D-03 0.141226615819D-06
8
35
0.326940137387D-02 0.228351644526D-03 0.113649105522D-03 0.142894025164D-06
9
44
0.326940120159D-02 0.228351623122D-03 0.113649090819D-03 0.142704612880D-06
10
54
0.326940135436D-02 0.228351637861D-03 0.113649108710D-03 0.142714961834D-06
11
65
0.326940125673D-02 0.228351623193D-03 0.113649088574D-03 0.142715057565D-06
12
77
0.326940125921D-02 0.228351622326D-03 0.113649100854D-03 0.142715104012D-06
13
90
0.326940131853D-02 0.228351632878D-03 0.113649101372D-03 0.142715141290D-06
14
104
0.326940135820D-02 0.228351643370D-03 0.113649096785D-03 0.142715121238D-06
15
119
0.326940152954D-02 0.228351658522D-03 0.113649120528D-03 0.142715169019D-06
16
135
0.326940132035D-02 0.228351632683D-03 0.113649100845D-03 0.142715116188D-06
2
Table 6.9: The values of integrals Ipq
using GQSTM
124
Table 6.10: Absolute error in M point formula for Polynomial of order N (GQSTS).
N
2
4
6
8
10
12
14
16
18
20
23
26
30
3
6
8
10
11
14
17
19
21
23
25
27
29
30
M
2
3
4
5
6
7
8
9
10
11
12
13
14
15
15
15
15
15
15
15
15
15
15
15
15
15
15
Formula 1
0.13877787807814456755D-15
0.61062266354383609723D-15
0.56996074526693973894D-13
0.81185058675714572018D-15
0.70429773124658368033D-15
0.10356299151581538354D-14
0.23835100559921329477D-14
0.43966566498632175808D-14
0.65156213757688874466D-14
0.83275400464266624567D-14
0.49274820335121205517D-14
0.56465249143045070923D-14
0.34867941867133822598D-15
0.38172243144174444751D-12
0.17785078965104617055D-12
0.12044532038402167018D-12
0.86659845965897375208D-13
0.74709335939893151135D-13
0.50034629217599047024D-13
0.35053557279063340957D-13
0.28127673801225938632D-13
0.22797735921287198835D-13
0.18630062770252919790D-13
0.15331919761552015302D-13
0.12692104311984309106D-13
0.10559261798270824784D-13
0.96363889090511634095D-14
125
Formula 2
0.13877787807814456755D-15
0.61062266354383609723D-15
0.56996074526693973894D-13
0.81532003370909933437D-15
0.70082828429463006614D-15
0.10338951916821770283D-14
0.23835100559921329477D-14
0.43957892881252291772D-14
0.65164887375068758502D-14
0.83284074081646508603D-14
0.49274820335121205517D-14
0.56473922760424954959D-14
0.34954678040932662952D-15
0.38172243144174444751D-12
0.17785078965104617055D-12
0.12045919817182948464D-12
0.86659845965897375208D-13
0.74712805386845104749D-13
0.50029425047171116603D-13
0.35053557279063340957D-13
0.28127673801225938632D-13
0.22797735921287198835D-13
0.18630062770252919790D-13
0.15332353442421009504D-13
0.12692971673722297510D-13
0.10559261798270824784D-13
0.96350878664441808041D-14
M
3
4
6
8
9
10
11
11
13
14
15
15
15
15
15
15
15
15
15
15
15
17
18
21
22
22
22
Formula 1
0.52735593669694935670D-15
0.34833247397614286456D-13
0.33834046675451645569D-13
0.40010356139319469548D-11
0.92428148468215454159D-12
0.64681315858905463756D-11
0.31802338540387609100D-12
0.90869672897397890665D-12
0.87030209428018423523D-13
0.36290415117434804415D-13
0.25661764380124907348D-13
0.23849845709467132338D-13
0.22624263573689518125D-13
0.22124663212608197682D-13
0.22604314253715784844D-13
0.24489958672102574155D-13
0.28469414325993369630D-13
0.35621679217445745280D-13
0.47618159415563354742D-13
0.66877926807595855507D-13
0.96405956134804071667D-13
0.23984286778855334887D-13
0.19719469113166354646D-13
0.19230277092940895045D-13
0.18625292280693983571D-13
0.17502926191736989381D-13
0.16464260510495876133D-13
126
Formula 2
0.52735593669694935670D-15
0.34833247397614286456D-13
0.33834046675451645569D-13
0.40010356139319469548D-11
0.92428842357605844882D-12
0.64681315858905463756D-11
0.31802338540387609100D-12
0.90869846369745488346D-12
0.87030209428018423523D-13
0.36290415117434804415D-13
0.25661764380124907348D-13
0.23849845709467132338D-13
0.22624263573689518125D-13
0.22124663212608197682D-13
0.22604314253715784844D-13
0.24489958672102574155D-13
0.28469414325993369630D-13
0.35621679217445745280D-13
0.47618159415563354742D-13
0.66877926807595855507D-13
0.96405956134804071667D-13
0.23984286778855334887D-13
0.19719469113166354646D-13
0.19230277092940895045D-13
0.18625292280693983571D-13
0.17502926191736989381D-13
0.16464260510495876133D-13
Chapter 7
A Symmetrical
Extended Gaussian
Quadrature Formula
for Evaluation of
Triangular Domain
Integrals
Chapter 7
A Symmetrical Extended
Gaussian Quadrature Formula for
Evaluation of Triangular Domain
Integrals
This chapter is devoted to present symmetrical quadrature formulae for triangular
1
domain integrals. As a result, it presents n n points (for n > 1) and n(n+1)
2
points (for n > 2) quadrature formulae in which the last one is totally free of
crowding of Gaussian quadrature points and weights. By the suitable transformation of a triangle in global space into its contiguous space, Gauss points and
weights are computed which are symmetric about the line of symmetry. For clarity
and reference, Gaussian integration points and weights for different values of n are
presented in tabular form. The efficiency and accuracy of the schemes are tested
through application examples. Finally, an error formula also presented to evaluate
the error in monomial/polynomial integration using m n points method successfully. The error calculated by the new error formula and the error in calculation
of integrals by the proposed methods are found in good agreement.
128
7.1
Introduction
The basic problems of integrating an arbitrary function of two variables over the
surface of the triangle were first described by Hammer and Stroud [1956, 1958];
Hammer et al. [1956]. In FEM solution of real word problems, the integrands
in practical situations are either polynomials or rational expressions. To get the
desired accuracy for the polynomial of higher degree it is necessary to increase
the Gaussian quadrature points with corresponding weights. If the integrands are
rational functions of bivariate or trivariate polynomial numerator with linear or
nonlinear denominator then more Gaussian points are needed. It is well known
that the first step of the Gaussian quadrature rule is to evaluate the integrands
at all Gaussian points which require more computing time before the application
of the weighted sum. In addition, specially for two and three dimensional cases
the higher order (more Gauss points and weights) formulae become erroneous due
to crowding of Gauss points and that effects the accuracy and efficiency of the
calculations.
A table of Gaussian quadrature formulae with symmetrically placed integration points is provided by Cowper [1973]. A detailed study of symmetric quadrature rules by formulating the problem in polar coordinates is made by Lyness and
Jespersen [1975]. Some numerical integration formulae for triangles with precision limited up to degree ten are derived in Dunavant [1985]; Hillion [1977]; Lannoy [1977]; Laurie [1977]; Laursen and Gellert [1978]; Lether [1976]; Reddy
[1978]; Reddy and Shippy [1981] and it is not likely that these techniques can be
extended much further to give a greater accuracy which may be demanded in future. Lague and Baldur [1977] proposed the product formulae based only on the
sampling points and weight coefficients of Gauss-Legendre quadrature rules. So,
numerical integration rules of very high degree of precision can be obtained as the
derivation rely on standard Gauss-Legendre quadrature rules. However, Lague
and Baldur [1977] have not worked out explicitly the weight coefficients and sampling points for the higher order quadrature rules for triangles solely. Bernardini
[1991]; Cools [2003]; Cools and Rabinowitz [1993]; Lyness and Cools [1994];
Rathod and Rao [1996, 1998]; Rathod et al. [2004] provided these informations
in a systematic manner. But, they have not considered the symmetry as well as
129
7.2
Problem Statement
It is well known that the domain integrals frequently encountered in many areas of
continuum mechanics. In two dimensional case, any arbitrary domain can be subdivided using triangles as shown in Figure 7.1. Thus, an arbitrary domain integral
renders to triangular domain integrals which may be evaluated numerically.
130
can be expressed as
ZZ
f (x, y) dx dy =
n ZZ
X
i=1
f (x, y) dx dy
where i are the triangles used in the mesh. Therefore, derivation of numerical
integration scheme is now required to evaluate the integral
ZZ
f (x, y) dx dy;
I=
: triangle (arbitrary)
(7.1)
7.3
By use of the property of definite integrals, the triangular domain integrals can be
expressed as the sum of three quadrilateral domain integrals as shown in Figure
7.2. In this chapter we are dividing the original shape of the triangle without any
131
3 ZZ
X
i=1
f (x, y) dx dy = I1 + I2 + I3
(7.2)
ei
where
ZZ
f (x, y) dx dy
I1 =
e
Z Z1
f (x, y) dx dy
I2 =
e
Z Z2
I3 =
f (x, y) dx dy
e3
4
X
xi i (, )
y=
i=1
4
X
yi i (, )
i=1
where i (, ) are the standard shape functions for the quadrilateral elements.
Now, the integrals I1 , I2 , I3 reduces to
I1
I2
I3
Z Z
J 1 1
=
f (X1 , Y1 )(4 + ) dx dy
96 1 1
Z Z
J 1 1
f (X2 , Y2 )(4 ) dx dy
=
96 1 1
Z Z
J 1 1
=
f (X3 , Y3 )(4 + ) dx dy
96 1 1
(Above derivation are on the same line of Eq.(3.12) to Eq.(3.16) in page number
47 to 49.)
Each quadrilaterals in Figure 7.2 is now transformed into its equivalent 2square in {(, ) | 1 , 1} in (, ) space and that results the equivalent
integral I in Eq.(7.3).
132
I =
3 ZZ
X
i =1
f (x, y)dxdy
ei
1
Z 1
J
=
[f (X1 , Y1 )(4 + ) + f (X2 , Y2 )(4 + + )
96 1 1
+ f (X3 , Y3 )(4 )]d d
Z
(7.3)
X1 =
1
[a
24 11
1
[b
24 11
X2 =
1
[a
24 21
Y2 =
1
[b
24 21
X3 =
1
[a
24 31
Y3 =
1
[b
24 31
133
a14 = 2x1 + x2 + x3
b14 = 2y1 + y2 + y3
a24 = x1 + 2x2 x3
b24 = y1 + 2y2 y3
a34 = x1 x2 + 2x3
b34 = y1 y2 + 2y3
Now, the right hand side of Eq.(7.3) with Eq.(7.4) can be evaluated by use of
available higher order Gaussian quadrature for square. Since the triangle is divided
into three quadrilaterals and each quadrilateral is transformed into equivalent 2square, for m n points Gaussian quadrature formula, total number of points
within the triangle is 3 m n. Figure 7.3 shows the distribution of Gauss points
within the triangle for m = n = 9. This scheme will ensure the desired accuracy
of the evaluation of the triangular domain integral derived later in this chapter.
7.4
134
!
=
x1 x2 x3
y1 y2 y3
N1
N2
N3
(7.5)
where
1
N1 (u, v) = (1 + u),
2
1
N2 (u, v) = (1 u + 2v),
4
1
N3 (u, v) = (1 u 2v) (7.6)
4
The original and the transformed triangles are shown in Figure 7.4.
From Eq.(7.5) using Eq.(7.6), we obtain the transformation equations
135
1
{(2x1 + x2 + x3 ) + (2x1 x2 x3 )u + (2x2 2x3 )v}
4
1
y(u, v) =
{(2y1 + y2 + y3 ) + (2y1 y2 y3 )u + (2y2 2y3 )v}
4
x(u, v) =
(7.7)
(7.8)
(1u)/2
(7.9)
v = (u1)/2
Here, we wish to present a new technique to calculate Gaussian points and corresponding weights for triangle. Consequently, new Gaussian quadrature formulae
will be presented as the final result of this study.
Using mathematical transformation equations:
136
(1 )
(7.10)
2
the integral I of Eq.(7.9) is transformed into an integral over the surface of
the standard square {(, )| 1 , 1} and the Eq.(7.7) reduces to
u = ,
v=
1
{(2x1 + x2 + x3 ) + (2x1 x2 x3 ) + (x2 x3 )(1 )}
4
1
y=
{(2y1 + y2 + y3 ) + (2y1 y2 y3 ) + (y2 y3 )(1 )} (7.11)
4
x=
= (1 )
(,
2
du dv = dv du =
(u, v)
1
d d = (1 ) d d
(,
2
(7.12)
(7.13)
(1 )
(1 ) 1
), y(,
)}
dd
2
2
8
1 1
Z 1Z 1
(1 ) 1
}
dd
= J
f {,
2
8
1 1
Z
I = J
f {x(,
(7.14)
1
N2 (u, v) = (1 + v),
2
1
N3 (u, v) = (1 2u v)
4
x(u, v) =
The original triangle transforms to a triangle with vertices 1(1, -1), 2(0, 1),
137
v=
(1 )
(1 )
1
, ), y(
, )}
dd
2
2
8
1 1
Z 1Z 1
1
(1 )
, )}
dd
f {(
= J
2
8
1 1
Z
I = J
(7.16)
f {x(
(7.17)
7.4.1
Here, our strive is to compute Gaussian points and corresponding weights for
the symmetric quadrature formulae for triangles. Conventionally, the numerical
schemes for the integral in Eq.(7.14) is expressible as
I = J
= J
m X
n
X
(1 im )jn
1 im
(
)Wim Wjn f (im ,
)
8
2
i=1 j=1
mn
X
Gr f (ur , vr )
(7.18)
r=1
where (ur , vr ) are the new Gaussian points, Gr is the corresponding weights for
triangles. Here im , i = 1, m; jn , j = 1, n are the m-th and n-th order Gaussian
quadrature points respectively; Wim , Wjn are the corresponding m-th and n-th
order weights respectively. For Eq. (7.17) we will get
138
I = J
= J
n X
m
X
1 jn
(1 jn )im n
(
)Wim Wjn f (
, j )
8
2
j=1 i=1
mn
X
(7.19)
r=1
where (u0r , vr0 ) are the new Gaussian points, G0r is the corresponding weights for
triangles. In order to calculate such Gaussian points and corresponding weights,
we present the following simple algorithm:
step 1. r 1
step 2. i = 1, m
step 3. j = 1, n
(1 im ) m n
Wi Wj ,
Gr =
8
(1 jn ) m n
Wi Wj ,
8
r =r+1
G0r =
(1 im )jn
vr =
2
ur =
im ,
u0r =
(1 jn )im
,
2
vr0 = jn
139
Figure 7.5: Gauss Points for m = n = 10 using SGQTS (100 points, formula 2).
7.4.2
Figure 7.6: Gauss Points for m = n = 15 using SGQTS (225 points, formula 1).
that is
m(m+1)
2
I2 = J
m1
X m+1i
X
i=1
j=1
m1
m+1i
(1 i )j
1 im1
)Wim1 Wjm+1i f (im1 ,
8
2
m(m+1)
1
2
= J
Lr f (pr , qr )}
(7.20)
r=1
where (pr , qr ) are the new Gaussian points, Lr is the corresponding weights for
the quadrature rules for triangles.
For Eq.(7.17) we can write
141
I2 = J
m1
X m+1j
X
j=1
i=1
(1 jm1 )im+1j m1
1 jm1
m+1j
m1
)Wi
Wj f (
, j )
(
8
2
m(m+1)
1
2
= J
(7.21)
r=1
where (p0r , qr0 ) are the new Gaussian points, L0r is the corresponding weights
for the quadrature rules for triangles. The Gaussian points and corresponding
weights can be calculated using the following algorithm:
step 1. r 1
step 2. i = 1, m 1
step 3. j = 1, m i + 1
pr =
im1 ,
(1 im1 )jm+1i
,
qr =
2
Lr =
(1 im1 ) m1 m+1i
Wi Wj
8
r =r+1
step 4. compute step 3
step 5. compute step 2
step 6. k 1
step 7. j = 1, m 1
step 8. i = 1, m j + 1
(1 jm1 )im+1j
=
,
2
k =k+1
p0r
qr0
jm1 ,
L0r
(1 jm1 ) m+1j m1
=
Wi
Wj
8
7.5
Application Examples
To show the accuracy and efficiency of the derived formulae some examples with
known results are considered. To compare the results we have taken the results
obtained by available Gaussian 7-points and 13-points rules, the quadrature rule
143
1y
(x + y) 2 dx dy = 0.4
I1 =
y=0
1
x=0
1y
I2 =
y=0
(x + y) 2 dx dy = 0.6666667
x=0
144
(x2 + y 2 ) 2 dx dy = 0.881373587
I3 =
y=0
1
x=0
y
I4 =
y=0
x=0
Computed values are summarized in Table 7.3. Some important remarks from
the Table 7.3 are:
Splitting the original triangle into quadrilaterals without any transformation (IOAT) provides the way of using Gaussian quadrature for
square domain integrals and the convergence rate is slow (as it needs
large number of points) but satisfactory in view of accuracy.
New Gaussian quadrature formulae for triangle (SGQTS and SGQTM)
are exact in view of accuracy and efficiency and the formula (SGQTM)
is faster.
Usual Gauss quadrature (GQT) for triangles e.g. 7-point and 13point rules cannot evaluate the integral of nonpolynomial functions
accurately.
The developed integration formulae of this chapter are always consistent, reliable and provides considerable savings in computational effort, in comparison with existing numerical quadrature rule of Dunavant [1985]; Taylor et al. [2000]; Wandzura and Xiao [2003].
The Fekete quadrature rule for triangle have some quadrature
points within the domain of integration where the integrands becomes
undefined and the models fail in that case, approximations are done
by simply excluding that points.
Again, we consider the following integrals of rational functions due to Rathod
and Karim [2002] to test the influences of formulae in Eq.(7.18), Eq.(7.19),
Eq.(7.20) and Eq.(7.21);
p,q
1y
=
y=0
x=0
xp y q
dx dy
+ x + y
145
Example-1: I
Example-2:I
r,0
r,0
Z
=
y=0
x=0
1y
=
y=0
Example-3: I
0,0
0,0
x=0
1y
x=0
1y
=
y=0
x=0
xr
dx dy, 6= = 0
0.375 0.375 x
xr
dx dy, = 0, 6= 0
1y
=
y=0
Example-4: I
1y
1
dx dy, , 6= 6= 0
12 + 21.53679831x 8.821067231y
1
dx dy = 6= 0
12 + 9.941125498(x + y)
Results are summarized in Tables 7.4, 7.5 and 7.6. In these tables formula
1 stands for Eq.(7.18) and Eq.(7.20) whereas formula 2 stands for Eq.(7.19) and
Eq.(7.21).
These tables substantiated the influences of numerical evaluation of the integrals as described in equations Eq.(7.18), Eq. (7.19), Eq.(7.20) and Eq. (7.21).
r
x
For the integrand +x+y
with 6= = 0 first formulae described in
Eq.(7.18) (SGQTS) and Eq.(7.20) (SGQTM) are more accurate and
rate of convergence is higher. But the formula in Eq.(7.20) requires
very less computational effort.
r
y
with 6= = 0 second formulae
Similarly for the integrand +x+y
described in Eq.(7.19) (SGQTS) and Eq.(7.21) (SGQTM) are more
accurate and convergence is higher. Here also the formula in Eq.(7.21)
requires very less computational effort.
7.6
Error Analysis
The n-point Gauss quadrature formula can evaluate exactly the integral of polynomial of order up to 2n-1. The total error in n-point Gauss quadrature formula
to evaluate the integral of polynomial of high order is given by
=
22n+1 [n!]4
f 2n (x1 )
(2n + 1)[(2n)!]3
I=
f (x, y)dydx
x=1
y=1
I=
y=1
m
X
i=1
147
22m+1 [m!]4
fx2m
3
(2m + 1)[(2m)!]
I =
m X
n
X
i=1 j=1
m
22m+1 [m!]4
f 2m
(2m + 1)[(2m)!]3 x
XX
22n+1 [n!]4
2n
+
f
=
f (xi , yj )W (m, i)W (n, j) + E
(2n + 1)[(2n)!]3 y
i=1 j=1
The total error in the 2-square domain integral is found to be given by the
error formula
2(2n+2) [n!]4
22m+2 [m!]4
2m
fx (x1 , y1 ) +
fy2n (x2 , y2 )
E=
3
3
(2m + 1)[(2m)!]
(2n + 1)[(2n)!]
(7.22)
Where fx2m (x1 , y1 ) is the 2m-th partial derivative of the function with respect
to x, fy2n (x2 , y2 ) is the 2n-th partial derivative of the function with respect to y
and (x1 , y1 ) and (x2 , y2 ) are points somewhere in the domain [1, 1] [1, 1].
Consider the square domain polynomial integral
Z
J=
1
N
X
(
xi y N i )dydx
1 i=0
148
(1x)/2
xp y q dy dx
JJ =
x = 1 y = (x1)/2
Z 1
1
p (1
Z
=
= 1
=
=
1
2q+1
1
2q+1
= 1
1
= 1
= 1
) q 1
}
(1 )d d
2
2
p (1 )q+1 q d d
= 1
p+q+1
Al d
l=p
q d
= 1
Let p+q = 2N-1. Then N N -point rule can calculate the monomial over the
triangular domain, if p+q is odd. If p+q is even, then (N + 1) (N + 1)-point rule
can calculate the monomials over the triangular domain. The methods SGQTS
and SGQTM are verified for the integral of all monomials xi y j where i , j are
non-negative integers such that i + j M = 30 over unit triangle. In Table 7.8
and 7.9 , we present the absolute error over corresponding monomial integrals. In
Table 7.10 and 7.11 we present the absolute error over corresponding polynomial
integrals. All these Tables are in good agreement with the above statement. It
is observed that the new methods are always accurate for monomials/polynomial
functions in view of both accuracy and efficiency and hence a proper balance is
observed.
Figure 7.9, 7.10, 7.11 and 7.12 shows the absolute error in the integral of
polynomial of order 15, 20, 25 and 30 using the methods SGQTS (m m points)
1 points). It is seen from the figures that, the error
and SGQTM ( m(m+1)
2
decreases with increase in number of total Gaussian integration points. These
figures verifies that, if SGQTS gives accurate results for a value N, SGQTM gives
the results of same accuracy for N+1 to evaluate the polynomials of degree at
most 2N 1 accurately. Table 7.8, 7.9, 7.10, 7.11 and Figure 7.9, 7.10, 7.11, 7.12
proves the efficiency of the method SGQTM to calculate the polynomial integral
over triangle with significantly less number of Gaussian integration points and
weights. It is seen from the figures that, the method SGQTM which is completely
149
7.7
Conclusions
Three numerical schemes IOAT, SGQTS and SGQTM to evaluate the triangular
domain integrals are presented in this chapter. The scheme IOAT is derived based
on the subdivision of an arbitrary triangular domain by three quadrilaterals. This
scheme is different from IOUT, IOST of previous chapters. The main advantage of
this scheme is that the readily available Gauss quadrature formulae for the square
domain is employed and that grantees the highest accuracy of the evaluation.
The scheme further is used as the testing tools for the accuracy of the developed
new quadrature formulae SGQTS and SGQTM. Both the quadrature schemes so
derived are symmetric and that substantially reduces the computational effort.
Here, the quadrature scheme SGQTS and SGQTM are respectively m m points
1 points formulae (for n=m). Figures shown that, in SGQTM there
and m(m+1)
2
are only two points at one terminal side of the triangle, more points can be taken
by choosing n > m. It is clearly shown that SGQTS suffers due to crowding of
Gauss points whereas the SGQTM is totally free from such crowding. As far as
the accuracy is concern, both the schemes are so accurate but the last one i.e.
150
151
Table 7.1: Computed Gauss points (u, v) and corresponding weights G for n n
point method (SGQTS).
n=8, 64 points
n=3, 9 points
u
-0.7745966692D+00
-0.7745966692D+00
0.0000000000D+00
0.0000000000D+00
0.7745966692D+00
0.7745966692D+00
-0.9602898565D+00
-0.9602898565D+00
-0.9602898565D+00
-0.9602898565D+00
-0.7966664774D+00
-0.7966664774D+00
-0.7966664774D+00
-0.7966664774D+00
-0.5255324099D+00
-0.5255324099D+00
-0.5255324099D+00
-0.5255324099D+00
-0.1834346425D+00
-0.1834346425D+00
-0.1834346425D+00
-0.1834346425D+00
0.1834346425D+00
0.1834346425D+00
0.1834346425D+00
0.1834346425D+00
0.5255324099D+00
0.5255324099D+00
0.5255324099D+00
0.5255324099D+00
0.7966664774D+00
0.7966664774D+00
0.7966664774D+00
0.7966664774D+00
0.9602898565D+00
0.9602898565D+00
0.9602898565D+00
0.9602898565D+00
v
G
0.6872983346D+00 0.6846437767D-01
0.0000000000D+00 0.1095430043D+00
0.3872983346D+00 0.6172839506D-01
0.0000000000D+00
0.9876543210D-01
0.8729833462D-01 0.8696116156D-02
0.0000000000D+00
0.1391378585D-01
0.9412232325D+00 0.2510939335D-02
0.7808486073D+00 0.5516085752D-02
0.5150979262D+00 0.7781386411D-02
0.1797925345D+00 0.8996247611D-02
0.8626602969D+00 0.5055663745D-02
0.7156719768D+00 0.1110639129D-01
0.4721032318D+00 0.1566747258D-01
0.1647854365D+00 0.1811354112D-01
0.7324766495D+00 0.6055613217D-02
0.6076702656D+00 0.1330310188D-01
0.4008583619D+00 0.1876631018D-01
0.1399177461D+00 0.2169618166D-01
0.5682201415D+00 0.5431069819D-02
0.4714013539D+00 0.1193109146D-01
0.3109666298D+00 0.1683085382D-01
0.1085414553D+00 0.1945855410D-01
0.3920697150D+00 0.3747417313D-02
0.3252651235D+00 0.8232407275D-02
0.2145657801D+00 0.1161322448D-01
0.7489318721D-01 0.1342632758D-01
0.2278132070D+00 0.1883402929D-02
0.1889962118D+00 0.4137500225D-02
0.1246740480D+00 0.5836654737D-02
0.4351689638D-01 0.6747896641D-02
0.9762955961D-01 0.5721629090D-03
0.8099450059D-01 0.1256939834D-02
0.5342917807D-01 0.1773129531D-02
0.1864920601D-01 0.2049957612D-02
0.1906662400D-01 0.5086480501D-04
0.1581787007D-01 0.1117409020D-03
0.1043448371D-01 0.1576297352D-03
0.3642107988D-02 0.1822395206D-03
152
Table 7.2: Computed Gauss points (p, q) and corresponding weights L for
1 point method (SGQTM).
n
m=3, 5 points
m=4, 9 points
m=5, 14 points
m=8, 35 points
p
-0.5773502692D+00
-0.5773502692D+00
0.5773502692D+00
-0.7745966692D+00
-0.7745966692D+00
0.0000000000D+00
0.0000000000D+00
0.7745966692D+00
-0.8611363116D+00
-0.8611363116D+00
-0.8611363116D+00
-0.3399810436D+00
-0.3399810436D+00
0.3399810436D+00
0.3399810436D+00
0.8611363116D+00
-0.9491079123D+00
-0.9491079123D+00
-0.9491079123D+00
-0.9491079123D+00
-0.7415311856D+00
-0.7415311856D+00
-0.7415311856D+00
-0.7415311856D+00
-0.4058451514D+00
-0.4058451514D+00
-0.4058451514D+00
0.0000000000D+00
0.0000000000D+00
0.0000000000D+00
0.4058451514D+00
0.4058451514D+00
0.7415311856D+00
0.7415311856D+00
0.9491079123D+00
m(m+1)
q
L
0.6109051324D+00 0.1095382131D+00
0.0000000000D+00 0.1752611410D+00
0.1220084679D+00 0.5283121635D-01
0.7640848152D+00 0.4286819789D-01
0.3016646138D+00 0.8036768192D-01
0.3872983346D+00 0.6172839506D-01
0.0000000000D+00
0.9876543210D-01
0.6506833684D-01 0.1565300908D-01
0.8432621081D+00 0.1917346465D-01
0.5010823929D+00 0.3873334126D-01
0.0000000000D+00
0.4603770905D-01
0.5769531667D+00 0.3799714765D-01
0.2277840768D+00 0.7123562050D-01
0.2556242426D+00 0.2989084476D-01
0.0000000000D+00
0.4782535162D-01
0.4008649394D-01 0.6038050853D-02
0.9358542787D+00 0.3193509426D-02
0.7763944673D+00 0.7015570468D-02
0.5121596892D+00 0.9896667159D-02
0.1787669565D+00 0.1144177446D-01
0.8264505139D+00 0.7884269378D-02
0.6456998424D+00 0.1703110962D-01
0.3533959938D+00 0.2324942473D-01
0.0000000000D+00
0.2544930806D-01
0.6554538727D+00 0.1149574334D-01
0.4647790050D+00 0.2420682761D-01
0.1677308129D+00 0.3139666970D-01
0.4530899230D+00 0.1237822093D-01
0.2692346551D+00 0.2500590605D-01
0.0000000000D+00
0.2972154195D-01
0.2558241574D+00 0.9864562309D-02
0.1010006927D+00 0.1849370967D-01
0.1001045414D+00 0.5020494508D-02
0.0000000000D+00
0.8032791213D-02
0.1469128025D-01 0.8237200311D-03
153
154
I4
0.6938790083
0.7238717079
0.7165792295
0.7162337687
0.7173750469
0.7164511237
0.7171579302
0.7173863082
0.7153629197
0.7166547182
0.7152268938
0.7178753416
0.7179654297
0.7180285699
0.7180745324
0.7184323903
0.7185688310
0.7181265336
0.7182531970
0.7175459725
0.7179128710
0.7180958213
0.7182831374
0.71828183
N.B.: For the highest accuracy, the scheme SGQTM require less number of points.
N.B.: * - Method fails, approximation done excluding points of discontinuity.
Method
Fekete
Dunavant
IOAT
SGQTS
SGQTM
Exact
Value
54
0.8095840242
0.4558851351
0.3053608511
91
0.8460916001
0.4911054664
0.3392939554
136
0.8651135795 0.5097291161 0.3575192530
25
0.8314373633
0.4758817940
0.3235008513
48
0.8608934610
0.5053379103
0.3529569503
79
0.8716316464
0.5160761035
0.3636951444
355
0.8536515855
0.4980960396
0.3457150818
366
0.8636423810
0.5080868226
0.3557058569
377
0.8699174296
0.5143618757
0.3619809270
388
0.8741141888
0.5185586402
0.3661776763
399
0.8770583628
0.5215028018
0.3691218499
3 10 10 0.8792029273
0.5236473748
0.3712664246
Results obtained using formula 1
55
0.8888888936
0.5333333316
0.3809523857
66
0.8888888912
0.5333333289
0.3809523670
77
0.8888888942
0.5333333215
0.3809523811
88
0.8888888822
0.5333333368
0.3809523803
99
0.8888888894
0.5333333261
0.3809523861
10 10
0.8888888916
0.5333333260
0.3809523859
44
0.8888888822
0.5333333368
0.3809523803
77
0.8888888823
0.5333333366
0.3809523814
90
0.8888889030
0.5333333378
0.3809523782
104
0.8888889011
0.5333333300
0.3809523797
Points
0.8888888
0.5333333
0.3809523
N.B.: For the highest accuracy, the scheme SGQTM require less number of points.
N.B.: * - Method fails, approximation done excluding points of discontinuity.
155
Fekete
Dunavant
IOAT
SGQTS
SGQTM
Exact
Value
55
0.1111115101
0.0400000244
0.0204081649
91
0.1111109809
0.0400000033
0.0204081632
136
0.1111110973
0.0400000004
0.0204081631
25
0.1111191026
0.0400000359
0.0204081637
48
0.1111094171
0.0399999999
0.0204081633
79
0.1111107485
0.0399999998
0.0204081633
355
0.1111078131
0.03999999048
0.02040816293
366
0.1111099001
0.03999999791
0.02040816328
377
0.1111105972
0.03999999965
0.02040816323
388
0.1111108684
0.03999999955
0.02040816329
399
0.1111109861
0.03999999964
0.02040816318
3 10 10
0.1111110426
0.03999999981
0.02040816318
Results obtained using formula 2
55
0.1111111124
0.04000000014
0.0204081636
66
0.1111111103
0.04000000016
0.0204081631
77
0.1111111106
0.04000000038
0.0204081631
88
0.1111111121
0.04000000059
0.0204081634
99
0.1111111110
0.03999999975
0.0204081633
10 10
0.1111111107
0.03999999995
0.0204081630
44
0.1111111119
0.04000000020
0.0204081633
77
0.1111111112
0.04000000003
0.0204081634
90
0.1111111111
0.03999999995
0.0204081632
104
0.1111111115
0.04000000005
0.0204081633
Points
0.1111111111111
0.04
0.0204081632653
N.B.: For the highest accuracy, the scheme SGQTM require less number of points.
N.B.: * - Method fails, approximation done excluding points of discontinuity.
156
Fekete
Dunavant
IOAT
SGQTS
SGQTM
Exact
Value
Points
77
13 13
54
85
126
55
91
136
25
48
79
355
366
377
388
399
3 10 10
55
66
77
88
99
10 10
44
77
90
104
Example -3
0.03669412062
0.03688941523
0.036947991050
0.036948008095
0.036948009595
0.03694928293
0.036948026227
0.036948013918
0.036948615464
0.036947887278
0.036948007107
0.03694724295
0.03694789912
0.03694799350
0.03694800728
0.03694800931
0.03694800972
formula 2
0.036956844135
0.036950091822
0.036948511561
0.036948132905
0.036948040297
0.036948017307
0.036948121605
0.036948011096
0.036948010012
0.036948009928
Example -4
0.02731705643
0.02731722965
0.027317233499
0.027317233545
0.027317233566
0.027317233923
0.027317233402
0.027317233345
0.027317233460
0.027317233409
0.027317233437
0.02731723353
0.02731723339
0.02731723359
0.02731723343
0.02731723344
0.02731723331
formula 1 / 2
0.027317233290
0.027317233662
0.027317233227
0.027317233349
0.027317233489
0.027317233318
0.027317233552
0.027317233459
0.027317233575
0.027317233635
0.03694801040
0.02731723349
N.B.: For the highest accuracy, the scheme SGQTM require less number of points.
157
Table 7.7: Error calculated using the new error formula and the m m point
method
Points
22
33
44
55
66
77
88
99
10 10
158
Table 7.8: The absolute error over corresponding monomials integrals for SGQTS.
N
1
2
3
3
4
5
5
6
7
7
8
9
10
11
12
13
14
16
17
19
20
21
22
23
24
25
26
27
28
29
30
30
30
30
30
i
1
0
1
2
2
3
3
3
4
4
3
3
7
6
7
8
10
3
11
4
6
6
7
8
15
14
11
11
6
15
15
30
30
30
30
m
2
2
3
3
3
4
5
5
5
6
6
7
7
7
8
8
9
10
10
11
11
12
12
13
13
14
14
15
15
15
15
16
18
19
20
Formula 1
0.11102230246251565404D-15
0.69388939039072283776D-16
0.93675067702747583098D-16
0.62450045135165055399D-16
0.25153490401663702869D-16
0.51451898297472098420D-14
0.24286128663675299322D-16
0.12034644114589099217D-16
0.51499603193061460615D-17
0.70473141211557788210D-17
0.45536491244391186228D-17
0.12834243216797158738D-16
0.76300727888667374543D-17
0.13145951341386741262D-17
0.18812601758518010531D-17
0.10693790959085541781D-17
0.34711410178481227895D-17
0.12001609829646181504D-16
0.22705776942335588464D-18
0.14525556254227182940D-17
0.72970564916977889366D-19
0.62256921623191074922D-19
0.16136639389684463881D-19
0.67353181377130278426D-20
0.26850282683471278326D-20
0.86559281224845889752D-21
0.43010806913343210882D-21
0.28096998962953544725D-21
0.87461942068294381203D-20
0.89334258922459623774D-19
0.41875729320400059259D-19
0.63853002746361298136D-14
0.72730450134672608442D-14
0.74376269032505604173D-14
0.74504204888858893696D-14
159
Formula 2
0.13877787807814456755D-15
0.55511151231257827021D-16
0.62450045135165055399D-16
0.97144514654701197287D-16
0.25153490401663702869D-16
0.46187012547882488889D-14
0.33393426912553536567D-16
0.11817803680091998331D-16
0.63425827090402009389D-17
0.80773061850170080334D-17
0.35507621148900270214D-17
0.12387009820646888159D-16
0.75487576259303246218D-17
0.14196272195982073683D-17
0.22107559923337238850D-17
0.12730905197232134096D-17
0.40056188075655863035D-17
0.10882679306323250756D-16
0.21292290711604974773D-18
0.12598556299223649668D-17
0.82638651916497676830D-19
0.69436849340151355139D-19
0.19766307917567149319D-19
0.77750841676921836300D-20
0.23076271138698089560D-20
0.82382019131453099994D-21
0.46373813091254114022D-21
0.29547149974335571368D-21
0.87056623568143397642D-20
0.83832869692064102146D-19
0.41875731743780760098D-19
0.46824523425303965496D-14
0.69096204452501197579D-14
0.79038338374193273239D-14
0.87811702353945975119D-14
Table 7.9: The absolute error over corresponding monomials integrals for SGQTM.
N
1
2
3
4
5
5
6
7
7
9
11
12
14
15
17
19
19
20
21
22
23
24
25
26
27
28
29
30
i
0
0
1
2
3
3
3
4
4
4
6
7
11
12
11
4
4
6
6
15
9
15
14
15
6
7
13
11
m
3
3
4
4
5
7
7
7
8
8
8
10
11
12
13
14
16
16
17
17
17
17
17
17
17
17
17
17
Formula 1
0.52735593669694935670D-15
0.31918911957973250537D-15
0.57176485768195561832D-14
0.65832755913319829233D-15
0.42149443657546470376D-14
0.81141690588815151841D-15
0.19320482713691689014D-15
0.11703962451981020365D-15
0.84947240214239272404D-16
0.22020029571447052863D-12
0.20913072536459919454D-12
0.82127035546026606883D-14
0.10281285913772697516D-16
0.77063057541196244848D-17
0.64774877142554824301D-14
0.34968188216441369043D-15
0.96761867520779067109D-17
0.53413037386019124233D-17
0.89397710137790976567D-18
0.88817939246452559470D-17
0.52169371335045647124D-18
0.32294409108116594855D-17
0.96615005738923339810D-18
0.85244149353887245280D-18
0.63794981254172174580D-18
0.38053348066190405421D-18
0.95259153317372358323D-19
0.78654016277078905607D-19
160
Formula 2
0.33306690738754696213D-15
0.97144514654701197287D-16
0.13357370765021414627D-14
0.65832755913319829233D-15
0.42500725161431773813D-14
0.32916377956659914616D-15
0.19320482713691689014D-15
0.18841499974803221917D-10
0.35517455946806886224D-11
0.33732556136388536228D-11
0.15063337289479821618D-10
0.15025672712483265428D-12
0.11435428782559115685D-12
0.17804363194808166260D-13
0.13019009425257495457D-14
0.30886005041982282464D-14
0.60896703176930916043D-15
0.24019216008850158479D-15
0.91583551450074603259D-16
0.65983039410497443386D-18
0.34613012908714982488D-17
0.38533092141955847647D-18
0.33492863534759454854D-18
0.23097206563013923845D-18
0.77370757148537395401D-16
0.75016567210806363964D-17
0.20251223810865608639D-18
0.51832691377395518820D-18
Table 7.10: Absolute error in M point formula for Polynomial of order N for
SGQTS.
N
2
4
6
8
10
12
14
16
18
20
23
26
30
3
6
8
10
11
14
17
19
21
23
25
27
29
30
M
2
3
4
5
6
7
8
9
10
11
12
13
14
15
15
15
15
15
15
15
15
15
15
15
15
15
15
Formula 1
0.13877787807814456755D-15
0.61062266354383609723D-15
0.56996074526693973894D-13
0.81185058675714572018D-15
0.70429773124658368033D-15
0.10356299151581538354D-14
0.23835100559921329477D-14
0.43966566498632175808D-14
0.65156213757688874466D-14
0.83275400464266624567D-14
0.49274820335121205517D-14
0.56465249143045070923D-14
0.34867941867133822598D-15
0.38172243144174444751D-12
0.17785078965104617055D-12
0.12044532038402167018D-12
0.86659845965897375208D-13
0.74709335939893151135D-13
0.50034629217599047024D-13
0.35053557279063340957D-13
0.28127673801225938632D-13
0.22797735921287198835D-13
0.18630062770252919790D-13
0.15331919761552015302D-13
0.12692104311984309106D-13
0.10559261798270824784D-13
0.96363889090511634095D-14
161
Formula2
0.13877787807814456755D-15
0.61062266354383609723D-15
0.56996074526693973894D-13
0.81532003370909933437D-15
0.70082828429463006614D-15
0.10338951916821770283D-14
0.23835100559921329477D-14
0.43957892881252291772D-14
0.65164887375068758502D-14
0.83284074081646508603D-14
0.49274820335121205517D-14
0.56473922760424954959D-14
0.34954678040932662952D-15
0.38172243144174444751D-12
0.17785078965104617055D-12
0.12045919817182948464D-12
0.86659845965897375208D-13
0.74712805386845104749D-13
0.50029425047171116603D-13
0.35053557279063340957D-13
0.28127673801225938632D-13
0.22797735921287198835D-13
0.18630062770252919790D-13
0.15332353442421009504D-13
0.12692971673722297510D-13
0.10559261798270824784D-13
0.96350878664441808041D-14
Table 7.11: Absolute error in M point formula for Polynomial of order N for
SGQTM.
N
2
3
3
5
5
7
7
11
16
18
19
20
21
22
23
24
25
26
27
28
29
21
26
27
28
29
30
M
3
4
6
8
9
10
11
11
13
14
15
15
15
15
15
15
15
15
15
15
15
17
18
21
22
22
22
Formula 1
0.52735593669694935670D-15
0.34833247397614286456D-13
0.33834046675451645569D-13
0.40010356139319469548D-11
0.92428148468215454159D-12
0.64681315858905463756D-11
0.31802338540387609100D-12
0.90869672897397890665D-12
0.87030209428018423523D-13
0.36290415117434804415D-13
0.25661764380124907348D-13
0.23849845709467132338D-13
0.22624263573689518125D-13
0.22124663212608197682D-13
0.22604314253715784844D-13
0.24489958672102574155D-13
0.28469414325993369630D-13
0.35621679217445745280D-13
0.47618159415563354742D-13
0.66877926807595855507D-13
0.96405956134804071667D-13
0.23984286778855334887D-13
0.19719469113166354646D-13
0.19230277092940895045D-13
0.18625292280693983571D-13
0.17502926191736989381D-13
0.16464260510495876133D-13
162
Formula2
0.52735593669694935670D-15
0.34833247397614286456D-13
0.33834046675451645569D-13
0.40010356139319469548D-11
0.92428842357605844882D-12
0.64681315858905463756D-11
0.31802338540387609100D-12
0.90869846369745488346D-12
0.87030209428018423523D-13
0.36290415117434804415D-13
0.25661764380124907348D-13
0.23849845709467132338D-13
0.22624263573689518125D-13
0.22124663212608197682D-13
0.22604314253715784844D-13
0.24489958672102574155D-13
0.28469414325993369630D-13
0.35621679217445745280D-13
0.47618159415563354742D-13
0.66877926807595855507D-13
0.96405956134804071667D-13
0.23984286778855334887D-13
0.19719469113166354646D-13
0.19230277092940895045D-13
0.18625292280693983571D-13
0.17502926191736989381D-13
0.16464260510495876133D-13
Chapter 8
Efficient Gauss
Quadrature Rules for
Tetrahedral Domain
Integrals
Chapter 8
Efficient Gauss Quadrature Rules
for Tetrahedral Domain Integrals
This chapter concentrates to present crowding free Gaussian Quadrature formulae
to evaluate tetrahedral domain integrals in Euclidean three dimensional space
(x, y, z). Through mathematical transformations, the tetrahedron T in (x, y, z)
space is transformed to a unit tetrahedron V : {(u, v, w)| 0 u, v, w 1, u + v +
w 1} whose base triangle is an unit triangle in w = 0 plane. It presents two
set of quadrature formulae and each set contains 6 formulae, for which Gaussian
points and corresponding weights are different. Among them one such set of
formulae are crowding free but the other one is not. Then it demonstrates the
efficiency and accuracy evaluating the integrals of polynomial, non-polynomial and
rational functions. In all the test cases the efficiency and accuracy of the derived
formulae are found satisfactory. Finally, it correctly evaluates all the integrals to
form element matrices for tetrahedral finite elements employing less number of
Gaussian points and weights. Thus, it ensures the balance between accuracy and
efficiency of evaluation of tetrahedral domain integrals encountered in science and
engineering.
8.1
Introduction
8.2
The finite element solution for three dimensional problems with tetrahedron element requires the numerical integration of terms containing product of shape
functions and the product of their global derivatives over an arbitrary tetrahedron T with coordinates (x1 , y1 , z1 ), (x2 , y2 , z2 ), (x3 , y3 , z3 ) and (x4 , y4 , z4 ) in the
natural coordinate space (x, y, z). The numerical integration of an arbitrary function f , over the tetrahedron T is given by
ZZZ
III =
f (x, y, z) dx dy dz
(8.1)
8.3
In this section, we first obtain the volume integral of a function f over an arbitrary
tetrahedron T in the Cartesian coordinate system (x, y, z) by transforming it to a
normalized tetrahedron V . The transformation from Cartesian coordinate system
(x, y, z) to tetrahedron V = {(1, 0, 0), (0, 1, 0), (0, 0, 1), (0, 0, 0)} is comfortably
accomplished by using shape functions:
Figure 8.1: The Arbitrary Tetrahedron and the Transformed Unit Tetrahedron.
x
x1 x2 x3 x4
y = y1 y2 y3 y4
z
z1 z2 z3 z4
N1
N2
N3
N4
(8.2)
where
N1 = u,
N2 = v,
N3 = w,
N4 = 1 u v w
where (xi , yi , zi ) refer to the Cartesian coordinates of vertex i of the tetrahedron. We can rewrite the relations (8.2) as:
173
(8.3)
If we consider the mapping (see Fig. 8.1) between the three dimensional spaces
(x, y, z) and (u, v, w) by the Eq.(8.3), we have the Jacobian of transformation and
the volume element
(x1 x4 ) (x2 x4 ) (x3 x4 )
Jacobian = (y1 y4 ) (y2 y4 ) (y3 y4 )
(z1 z4 ) (z2 z4 ) (z3 z4 )
= J (say)
(8.4)
ZZZ
III =
f (x, y, z) dx dy dz
Z ZTZ
|J| f {x(u, v, w), y(u, v, w), z(u, v, w)} du dv dw
=
V
ZZZ
= |J|
f (u, v, w) du dv dw
(8.5)
8.4
1u
1uv
III = |J|
u=0
or
v=0
1u
f (u, v, w) dw dv du
(8.6)
f (u, v, w) dv dw du
(8.7)
f (u, v, w) dw du dv
(8.8)
f (u, v, w) du dw dv
(8.9)
f (u, v, w) dv du dw
(8.10)
f (u, v, w) du dv dw
(8.11)
w=0
1uw
III = |J|
u=0
or
Z
w=0
v=0
1v
1uv
III = |J|
v=0
or
Z
u=0
w=0
1v
1vw
III = |J|
v=0
or
w=0
u=0
1w
1uw
III = |J|
w=0
or
Z
u=0
v=0
1w
1vw
III = |J|
w=0
v=0
u=0
It is now clear that either all or at least one form of the integral III described in
Eq.(8.6) to Eq.(8.11) will ensure the effective use of Gauss points of the quadrature
formulae. Thus, considering the presence or absence of the independent variables
of the integrands we need six quadrature formulae for the integrals mentioned
in Eq.(8.6) to Eq.(8.11). Using isoparametric transformation, the domain of the
integrals can be transformed into equivalent 2-cube given by {(, , )| 1
, , 1} in (, , ). Now, the task is to present quadrature formulae one by
one.
Formula - 1: For the integral III
Here, we are required to evaluate the integral of Eq.(8.6) given as
Z
1u
1uv
III = |J|
f (u, v, w) dw dv du
u=0
v=0
w=0
175
1+
2
(1 )(1 + )
4
v=
w=
(1 )(1 )(1 + )
8
}
{
}+
{
}
(1 )2 (1 )
=
64
and
(u, v, w)
(1 )2 (1 )
d d d =
d d d
(, , )
64
dw dv du =
III = |J|
f (, , )
=1
=1
=1
(1 )2 (1 )
d d d
64
1u
1uw
III = |J|
f (u, v, w) dv dw du
u=0
w=0
v=0
Using transformation
u=
1+
2
v=
(1 )(1 + )(1 )
8
w=
(1 )(1 + )
4
176
(8.12)
(u, v, w)
(1 )2 (1 )
d d d =
d d d
(, , )
64
dv dw du =
III = |J|
f (, , )
=1
=1
=1
(1 )2 (1 )
d d d
64
(8.13)
(1 + )(1 )
4
1+
2
v=
w=
(1 )(1 )(1 + )
8
III = |J|
f (, , )
=1
=1
=1
(1 )(1 )2
d d d
64
(8.14)
(1 + )(1 )(1 )
8
v=
1+
2
w=
(1 )(1 + )
4
III = |J|
f (, , )
=1
=1
=1
(1 )2 (1 )
d d d
64
u=
(1 + )(1 )(1 )
8
v=
177
(1 + )(1 )
4
w=
1+
2
(8.15)
III = |J|
f (, , )
=1
=1
=1
(1 )(1 )2
d d d
64
(8.16)
(1 + )(1 )
4
(1 )(1 + )(1 )
8
v=
w=
1+
2
III = |J|
f (, , )
=1
=1
=1
(1 )(1 )2
d d d
64
(8.17)
III = |J|
p
m X
n X
X
mnp
X
(u, v, w)
= |J|{
f (ur , vr , wr )Gr }
(, , )
r=1
(8.18)
Here (ur , vr , wr ) are the new Gauss points and Gr are the corresponding
weights for the integral over unit tetrahedron where im , i = 1, m; jn , j =
1, n; kp , k = 1, p are the m-th, n-th and k-th order one dimensional Gaussian
quadrature points respectively and Wim , Wjn , Wkp are the corresponding weights
respectively. The new Gauss points (ur , vr , wr ) and the corresponding weights
Gr can be computed by the following algorithm:
178
step 1. r 1
step 2. i = 1, m
step 3. j = 1, n
step 4. k = 1, p
For Formula 1.
(1 im )(1 + jn )
1 + im
, vr =
,
ur =
2
4
(1 im )2 (1 jn ) m n p
Gr =
Wi Wj Wk
64
(1 im )(1 jn )(1 + kp )
wr =
,
8
For Formula 2.
(1 im )(1 + jn )(1 kp )
1 + im
ur =
, vr =
,
2
8
(1 im )2 (1 kp ) m n p
Wi Wj Wk
Gr =
64
wr =
(1 im )(1 + kp )
,
4
For Formula 3.
(1 + im )(1 jn )
1 + jn
, vr =
,
4
2
(1 im )(1 jn )2 m n p
Gr =
Wi Wj Wk
64
ur =
wr =
(1 im )(1 jn )(1 + kp )
,
8
For Formula 4.
(1 + im )(1 jn )(1 kp )
1 + jn
, vr =
,
ur =
8
2
(1 jn )2 (1 kp ) m n p
Gr =
Wi Wj Wk
64
179
(1 jn )(1 + kp )
wr =
,
4
1 + kp
wr =
,
2
For Formula 6.
(1 im )(1 + jn )(1 kp )
(1 + im )(1 kp )
ur =
, vr =
,
4
8
(1 im )(1 kp )2 m n p
Wi Wj Wk
Gr =
64
r=r+1
wr =
1 + kp
,
2
Figure 8.2: Distribution of Gauss Points for GQTTS (m =5, 125 points).
8.5
181
Figure 8.3: Distribution of Gauss Points for GQTTS (m =10, 1000 points).
III = |J|
m (n=m+2i)
X
X (p=n+2j)
X
i=1
j=1
k=1
TP
X
(u, v, w)
= |J|{
f (u0r , vr0 , wr0 )G0r }
Wi Wj Wk
(, , )
r=1
(8.19)
Where (u0r , vr0 , wr0 ) are the new Gaussian points, G0r is the corresponding weights.
For a given value of m, total Gaussian points TP can be expressed as
m
X
(r + 2)(r + 3)
m3 + 9m2 + 20m
m(m + 4)(m + 5)
TP =
{
1} =
=
2
6
6
r=1
182
step 1. r 1
step 2.
For Formula 1.
i = 1, m; j = 1, (n = m i + 2); k = 1, (p = n j + 2)
(1 i0 )(1 + j0 )
1 + i0
, vr0 =
,
2
4
(1 i0 )2 (1 j0 )
G0r =
Wi Wj Wk
64
u0r =
wr0 =
(1 i0 )(1 j0 )(1 + k0 )
,
8
For Formula 2.
i = 1, m; k = 1, (n = m i + 2); j = 1, (p = n k + 2)
(1 i0 )(1 + j0 )(1 k0 )
1 + i0
, vr0 =
,
2
8
(1 i0 )2 (1 k0 )
G0r =
Wi Wj Wk
64
u0r =
wr0 =
(1 i0 )(1 + k0 )
4
For Formula 3.
j = 1, m; i = 1, (n = m j + 2); k = 1, (p = n i + 2)
1 + j0
(1 + i0 )(1 j0 )
; vr0 =
,
4
2
(1 i0 )(1 j0 )2
G0r =
Wi Wj Wk
64
u0r =
183
wr0 =
(1 i0 )(1 j0 )(1 + k0 )
8
wr0
(1 j0 )(1 + k0 )
=
4
For Formula 5.
k = 1, m; j = 1, (n = m k + 2); i = 1, (p = n j + 2)
(1 + j0 )(1 k0 )
(1 + i0 )(1 j0 )(1 k0 )
, vr0 =
,
8
4
(1 j0 )(1 k0 )2
G0r =
Wi Wj Wk
64
u0r =
wr0 =
1 + k0
2
wr0 =
1 + k0
2
For Formula 6.
k = 1, m; i = 1, (n = m k + 2); j = 1, (p = n i + 2)
(1 i0 )(1 + j0 )(1 k0 )
(1 + i0 )(1 k0 )
, vr0 =
,
4
8
(1 i0 )(1 k0 )2
0
Gr =
Wi Wj Wk
64
r=r+1
u0r =
185
Figure 8.5: Distribution of Gauss Points for GQTTM,(m =10, 350 points).
8.6
Application Examples
To show the accuracy and efficiency of all the quadrature formula so presented,
we consider some typical integrals thoroughly studied by Rathod and Rao [1996,
1998]; Rathod et al. [2004, 2005]; Reddy and Shippy [1981].
Consider
Z
1x
1xy
p
(x + y + z) dz dy dx
I1 =
x=0
1
x=0
1
y=0
1x
z=0
1xy
y=0
1x
I2 =
Z
z=0
1xy
I3 =
dz dy dx
p
(x + y + z)
sin(x + 2y + 4z) dz dy dx
x=0
y=0
z=0
186
1x
1xy
(1 + x + y + z)4 dz dy dx
I4 =
x=0
y=0
z=0
ZZZ
=
x y z dz dy dx
1x
1xy
=
x=0
y=0
z=0
x y z
p
(ax + by + cz)
dz dy dx
187
[K]e =
V
(8.20)
Where [D] is a material property matrix, [B] is the strain displacement matrix and
the integration is performed over the volume element V 0 in global coordinates. The
symmetric material property matrix [D] for anisotropic materials can be written
as
D11
D12
D13
D14
D15
D16
D12
D22
D23
D24
D25
D26
D13
D23
D33
D34
D35
D36
D14
D24
D34
D44
D45
D46
D15
D25
D35
D45
D55
D56
D16
D26
D36
D46
D56
D66
V0
For a three dimensional element, the matrix [Bi ] has the form
Ni
x0
Ni
z0
0
0
Ni
y 0
0
Ni
z 0
0
Ni
y 0
188
Ni
0
z
Ni
y 0
Ni
x0
Ni
x0
Where Kij11 , Kij12 , ...Kij33 are volume integrals over domain V 0 and they are triple
integrals of the product of global derivatives of shape functions. Hence we need
to evaluate
ZZZ
Ni Nj
dx0 dy 0 dz 0
t s
t, s = x0 , y 0 , z 0
V0
IJpqr
xp y q z r
dx dy dz
J(x, y, z)
V
Z 1 Z 1x Z 1xy p q r
xy z
dx0 dy 0 dz 0
=
J(x,
y,
z)
x=0 y=0
z=0
ZZZ
(8.21)
20
X
Ni (x, y, z)ti ,
t = x0 , y 0 , z 0 and
i=1
189
x0
= x + Rx(y + z) + Sxyz,
z0
= z + Rz(x + y) + Sxyz,
y 0 = y + Ry(x + z) + Sxyz,
190
8.7
For clarity and reference, Gaussian points and weights for the schemes GQTTS
and GQTTM are presented in Tables 8.1 and 8.3 respectively. The scheme
GQTTM is so faster as it require function values at less number of Gauss points,
e.g. for m = 10, the scheme GQTTS require function values at 101010 = 1000
points where as the scheme GQTTM require function values at 10(10+4)(10+5)
= 350
6
Gauss points.
Computed results of Tables 8.5, 8.6 and 8.7 clearly establishes the best accuracy of the schemes so presented. At the same time the scheme GQTTM is always
in good agreement with the result of the scheme GQTTS. For example, in table
8.5 for the same accuracy GQTTS utilizes 1000 points and GQTTM utilizes 350
points. Further it is observed that the absence of any independent variable in the
explicit form of the integrand enhances faster convergence of one among the six
quadrature formulae for both GQTTS and GQTTM. For all case studies in Table
8.8 for IJpqr in Eq.(8.21) the method GQTTS require more Gaussian points than
the method GQTTM. Results are compared with the results obtained by Rathod
and Rao [1996, 1998]; Rathod et al. [2004, 2005]; Reddy and Shippy [1981].
It is found that the developed integration formulae are always consistent, reliable
and provides considerable savings in computational effort, in comparison with the
existing numerical methods that employs Gaussian quadrature procedures.
8.8
Conclusions
This chapter presented suitable Gaussian Quadrature formulae to evaluate tetrahedral domain integrals in Euclidean three dimensional space. It is also demonstrated that there are two set of quadrature formulae and each set contains six
quadrature formulae for which Gaussian points and corresponding weights are
different. Derived formulae substantially reduced the number of Gaussian in191
192
Table 8.1: Computed Gauss points (u, v, w ) and corresponding weights G for
n n n point method (GQTTS).
u
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.500000000000000
0.500000000000000
0.500000000000000
0.500000000000000
0.500000000000000
0.500000000000000
0.500000000000000
0.500000000000000
0.500000000000000
0.500000000000000
v
0.044709521703645
0.044709521703645
0.044709521703645
0.044709521703645
0.044709521703645
0.219940124839679
0.219940124839679
0.219940124839679
0.219940124839679
0.219940124839679
0.476544961484666
0.476544961484666
0.476544961484666
0.476544961484666
0.476544961484666
0.733149798129653
0.733149798129653
0.733149798129653
0.733149798129653
0.733149798129653
0.908380401265687
0.908380401265687
0.908380401265687
0.908380401265687
0.908380401265687
0.036084856923188
0.036084856923188
0.036084856923188
0.036084856923188
0.036084856923188
0.177512700518577
0.177512700518577
0.177512700518577
0.177512700518577
0.177512700518577
0.384617327526421
0.384617327526421
0.384617327526421
0.384617327526421
0.384617327526421
0.591721954534264
0.591721954534264
0.591721954534264
0.591721954534264
0.591721954534264
0.733149798129653
0.733149798129653
0.733149798129653
0.733149798129653
0.733149798129653
0.023455038515334
0.023455038515334
0.023455038515334
0.023455038515334
0.023455038515334
0.115382672473579
0.115382672473579
0.115382672473579
0.115382672473579
0.115382672473579
w
0.042612194596523
0.209622716641314
0.454190200632844
0.698757684624373
0.865768206669165
0.034392113505281
0.169185566063329
0.366574899064827
0.563964232066324
0.698757684624372
0.022354760851822
0.109970062419839
0.238272480742333
0.366574899064827
0.454190200632844
0.010317408198364
0.050754558776350
0.109970062419839
0.169185566063329
0.209622716641314
0.002097327107122
0.010317408198364
0.022354760851822
0.034392113505281
0.042612194596523
0.034392113505281
0.169185566063329
0.366574899064827
0.563964232066324
0.698757684624372
0.027757722467940
0.136548920950906
0.295860977267132
0.455173033583358
0.563964232066324
0.018042428461594
0.088756350259289
0.192308663763210
0.295860977267132
0.366574899064827
0.008327134455248
0.040963779567671
0.088756350259289
0.136548920950906
0.169185566063329
0.001692743417907
0.008327134455248
0.018042428461594
0.027757722467940
0.034392113505281
0.022354760851822
0.109970062419839
0.238272480742333
0.366574899064827
0.454190200632844
0.018042428461594
0.088756350259289
0.192308663763210
0.295860977267132
0.366574899064827
193
G
0.001439311118758
0.002907629360183
0.003455952678910
0.002907629360183
0.001439311118758
0.002346734777065
0.004740764418080
0.005634781968811
0.004740764418080
0.002346734777065
0.001813025505580
0.003662589777903
0.004353284200533
0.003662589777903
0.001813025505580
0.000704005022098
0.001422198192795
0.001690397586996
0.001422198192795
0.000704005022098
0.000070841369555
0.000143110438979
0.000170098332251
0.000143110438979
0.000070841369555
0.001894039243550
0.003826249962299
0.004547807567383
0.003826249962299
0.001894039243550
0.003088149395941
0.006238535737859
0.007415004329977
0.006238535737859
0.003088149395941
0.002385822920682
0.004819728467296
0.005728637128330
0.004819728467296
0.002385822920682
0.000926424538886
0.001871519752855
0.002224452604614
0.001871519752855
0.000926424538886
0.000093222606464
0.000188323972528
0.000223838273981
0.000188323972528
0.000093222606464
0.000951130350813
0.001921429284706
0.002283774120164
0.001921429284706
0.000951130350813
0.001550777064587
0.003132807678149
0.003723595323417
0.003132807678149
0.001550777064587
v
0.250000000000000
0.250000000000000
0.250000000000000
0.250000000000000
0.250000000000000
0.384617327526421
0.384617327526421
0.384617327526421
0.384617327526421
0.384617327526421
0.476544961484666
0.476544961484666
0.476544961484666
0.476544961484666
0.476544961484666
0.010825220107480
0.010825220107480
0.010825220107480
0.010825220107480
0.010825220107480
0.053252644428581
0.053252644428581
0.053252644428581
0.053252644428581
0.053252644428581
0.115382672473579
0.115382672473579
0.115382672473579
0.115382672473579
0.115382672473579
0.177512700518577
0.177512700518577
0.177512700518577
0.177512700518577
0.177512700518577
0.219940124839679
0.219940124839679
0.219940124839679
0.219940124839679
0.219940124839679
0.002200555327023
0.002200555327023
0.002200555327023
0.002200555327023
0.002200555327023
0.010825220107480
0.010825220107480
0.010825220107480
0.010825220107480
0.010825220107480
0.023455038515334
0.023455038515334
0.023455038515334
0.023455038515334
0.023455038515334
0.036084856923188
0.036084856923188
0.036084856923188
0.036084856923188
0.036084856923188
0.044709521703645
0.044709521703645
0.044709521703645
0.044709521703645
0.044709521703645
w
0.011727519257667
0.057691336236790
0.125000000000000
0.192308663763210
0.238272480742333
0.005412610053740
0.026626322214291
0.057691336236790
0.088756350259289
0.109970062419839
0.001100277663512
0.005412610053740
0.011727519257667
0.018042428461594
0.022354760851822
0.010317408198364
0.050754558776350
0.109970062419839
0.169185566063329
0.209622716641314
0.008327134455248
0.040963779567671
0.088756350259289
0.136548920950906
0.169185566063329
0.005412610053740
0.026626322214291
0.057691336236790
0.088756350259289
0.109970062419839
0.002498085652232
0.012288864860910
0.026626322214291
0.040963779567671
0.050754558776350
0.000507811909116
0.002498085652232
0.005412610053740
0.008327134455248
0.010317408198364
0.002097327107122
0.010317408198364
0.022354760851822
0.034392113505281
0.042612194596523
0.001692743417907
0.008327134455248
0.018042428461594
0.027757722467940
0.034392113505281
0.001100277663512
0.005412610053740
0.011727519257667
0.018042428461594
0.022354760851822
0.000507811909116
0.002498085652232
0.005412610053740
0.008327134455248
0.010317408198364
0.000103228219901
0.000507811909116
0.001100277663512
0.001692743417907
0.002097327107122
194
G
0.001198089532334
0.002420324733784
0.002876751714678
0.002420324733784
0.001198089532334
0.000465222935413
0.000939821730277
0.001117054144152
0.000939821730277
0.000465222935413
0.000046813628964
0.000094570715294
0.000112404944503
0.000094570715294
0.000046813628964
0.000170456069101
0.000344347420568
0.000409284761972
0.000344347420568
0.000170456069101
0.000277921277830
0.000561443635575
0.000667321173390
0.000561443635575
0.000277921277830
0.000214714662336
0.000433756571553
0.000515554769800
0.000433756571553
0.000214714662336
0.000083374558238
0.000168429403669
0.000200191969715
0.000168429403669
0.000083374558238
0.000008389667270
0.000016948415501
0.000020144562700
0.000016948415501
0.000008389667270
0.000003486737214
0.000007043744304
0.000008372059841
0.000007043744304
0.000003486737214
0.000005684974827
0.000011484521661
0.000013650282922
0.000011484521661
0.000005684974827
0.000004392061882
0.000008872639079
0.000010545849211
0.000008872639079
0.000004392061882
0.000001705455115
0.000003445281079
0.000004094995235
0.000003445281079
0.000001705455115
0.000000171613514
0.000000346685638
0.000000412063922
0.000000346685638
0.000000171613514
Table 8.3: Computed Gauss points (u, v, w) and corresponding weights G for
method GQTTM.
u
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.046910077030668
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
0.230765344947158
v
0.032181312753100
0.032181312753100
0.032181312753100
0.032181312753100
0.032181312753100
0.032181312753100
0.032181312753100
0.161448959877800
0.161448959877800
0.161448959877800
0.161448959877800
0.161448959877800
0.161448959877800
0.362832190643147
0.362832190643147
0.362832190643147
0.362832190643147
0.362832190643147
0.590257732326185
0.590257732326185
0.590257732326185
0.590257732326185
0.791640963091532
0.791640963091532
0.791640963091532
0.920908610216232
0.920908610216232
0.036084856923188
0.036084856923188
0.036084856923188
0.036084856923188
0.036084856923188
0.036084856923188
0.177512700518577
0.177512700518577
0.177512700518577
0.177512700518577
w
0.023433480857716
0.119013078326949
0.273581157949139
0.460454305108116
0.647327452267094
0.801895531889283
0.897475129358516
0.026729949407128
0.134100263792109
0.301370120404256
0.490270842687276
0.657540699299423
0.764911013684404
0.027689035691369
0.136211029207980
0.295128866163093
0.454046703118206
0.562568696634817
0.025192108132559
0.119738061911055
0.243094128732092
0.337640082510588
0.018195566651977
0.080724479938900
0.143253393225823
0.006800711586111
0.025380601166989
0.024754981014778
0.124192134960240
0.279103095011448
0.454046703118206
0.608957663169414
0.708394817114875
0.027757722467940
0.136548920950906
0.295860977267132
0.455173033583358
195
G
0.000576651691559
0.001245647212280
0.001700451805538
0.001861350220022
0.001700451805538
0.001245647212280
0.000576651691559
0.001381103480046
0.002908218533639
0.003772009211444
0.003772009211444
0.002908218533639
0.001381103480046
0.001847056770230
0.003731338155492
0.004434997208027
0.003731338155492
0.001847056770230
0.001666971143698
0.003125174680918
0.003125174680918
0.001666971143698
0.000913358455696
0.001461373529114
0.000913358455696
0.000155625745624
0.000155625745624
0.001369601055914
0.002883997638157
0.003740594295469
0.003740594295469
0.002883997638157
0.001369601055914
0.003088149395941
0.006238535737859
0.007415004329977
0.006238535737859
v
0.177512700518577
0.384617327526421
0.384617327526421
0.384617327526421
0.384617327526421
0.591721954534264
0.591721954534264
0.591721954534264
0.733149798129653
0.733149798129653
0.034715922101487
0.034715922101487
0.034715922101487
0.034715922101487
0.034715922101487
0.165004739103786
0.165004739103786
0.165004739103786
0.165004739103786
0.334995260896214
0.334995260896214
0.334995260896214
0.465284077898513
0.465284077898513
0.026007638687364
0.026007638687364
0.026007638687364
0.026007638687364
0.115382672473579
0.115382672473579
0.115382672473579
0.204757706259795
0.204757706259795
0.009913265714653
0.009913265714653
0.009913265714653
0.036996811316015
0.036996811316015
w
0.563964232066324
0.026704690362579
0.126927363566585
0.257689963959836
0.357912637163842
0.020005976974413
0.088756350259289
0.157506723544164
0.007625627532458
0.028459229390730
0.021826511935363
0.107371440734671
0.232642038949257
0.357912637163842
0.443457565963151
0.023259338763280
0.110551611250369
0.224443649645845
0.311735922132934
0.018596308892467
0.082502369551893
0.146408430211319
0.007336337565514
0.027379584535973
0.014216705160388
0.067571983801774
0.137185722458021
0.190541001099406
0.013003819343682
0.057691336236790
0.102378853129897
0.005496060745114
0.020511577942250
0.004169602249037
0.018498405658008
0.032827209066978
0.002094919542875
0.007818346171778
196
G
0.003088149395941
0.003502853052757
0.006567016899386
0.006567016899386
0.003502853052757
0.002172316996694
0.003475707194710
0.002172316996694
0.000393465715983
0.000393465715983
0.001363446374903
0.002754370934147
0.003273792643220
0.002754370934147
0.001363446374903
0.002702019388299
0.005065644124426
0.005065644124426
0.002702019388299
0.002125571182931
0.003400913892690
0.002125571182931
0.000429372505117
0.000429372505117
0.000546317509560
0.001024215479118
0.001024215479118
0.000546317509560
0.000786674148254
0.001258678637206
0.000786674148254
0.000199483844893
0.000199483844893
0.000028555008472
0.000045688013556
0.000028555008472
0.000013772324628
0.000013772324628
197
Exact
Value
GQTTM
GQTTS
Method
8
27
64
125
216
343
512
729
1000
14
28
48
75
110
154
208
273
350
Points
0.14285714285714
I1
0.143127410953799
0.142875312759849
0.142860037924269
0.142857834882225
0.142857355360141
0.142857220464426
0.142857175072351
0.142857157618562
0.142857150174568
0.143058065243440
0.142868320929028
0.142858821339166
0.142857541720436
0.142857267261819
0.142857189383667
0.142857162683365
0.142857152181787
0.142857147595743
0.2
I2
0.197660776240555
0.199583323221216
0.199881018522955
0.199956079429135
0.199980842555110
0.199990593978398
0.199994953918813
0.199997099779897
0.199998238575602
0.198325920889004
0.199730676421398
0.199924597609756
0.199971858116752
0.199987455481037
0.199993687508278
0.199996530092608
0.199997959524722
0.199998734389269
I4
0.020103982733156
0.020798626362385
0.020831943224904
0.020833282126123
0.020833331531397
0.020833333271718
0.020833333331268
0.020833333333265
0.020833333333331
0.020334811873381
0.020811918661132
0.020832494969141
0.020833301701749
0.020833332126155
0.020833333282340
0.020833333330381
0.020833333333016
0.020833333333269
0.131902326890181 0.020833333333333
I3
0.131949528497795
0.131902664864685
0.131902289803505
0.131902327570123
0.131902326884088
0.131902326890216
0.131902326890182
0.131902326890182
0.131902326890182
0.133142199123991
0.131877204546620
0.131902589866991
0.131902325527761
0.131902326955921
0.131902326902415
0.131902326892685
0.131902326890580
0.131902326890160
Value of I
198
Exact
Value
GQTTM
GQTTS
Method
8
27
64
125
216
343
512
729
1000
14
28
48
75
110
154
208
273
350
Points
15721.6666666667
I52,1,0
15550.9773662551233
15721.6666666665187
15721.6666666666715
15721.6666666666587
15721.6666666666897
15721.6666666666751
15721.6666666667060
15721.6666666666787
15721.6666666667078
15550.1862778507402
15721.6666928552531
15721.6666681218612
15721.6666668061934
15721.6666666861729
15721.6666666702421
15721.6666666674991
15721.6666666669025
15721.6666666667388
109662.063492063
426917356.623377
Value of I
I52,2,0
I54,4,0
107484.1792409692279 387905448.6299038529396
109657.4916666656208 425756672.2764838933944
109662.0634920635493 426910389.2180011272430
109662.0634920634620 426917352.2770764231681
109662.0634920636221 426917356.6233774423599
109662.0634920635347 426917356.6233763694763
109662.0634920638694 426917356.6233769655227
109662.0634920634765 426917356.6233772635459
109662.0634920637239 426917356.6233774423599
107434.4429454763012 384070078.9281456470489
109657.4927840148156 425742213.6549844145774
109662.0635520677169 426910391.3103920221328
109662.0634976892179 426917352.7276153564453
109662.0634928344079 426917356.6618708372116
109662.0634922022436 426917356.6281496286392
109662.0634920942829 426917356.6241724491119
109662.0634920715237 426917356.6235436201095
109662.0634920659504 426917356.6234186291694
330565.079365081
I51,2,3
281238.2292672184994
329105.6422399957082
330562.9232264332240
330565.0793650789419
330565.0793650801642
330565.0793650793493
330565.0793650807463
330565.0793650799896
330565.0793650812120
278927.4392087566666
329272.2495105185662
330562.7560054035275
330565.0694366566604
330565.0784725357079
330565.0792490325402
330565.0793439878034
330565.0793599335593
330565.0793634889996
Table 8.6: Calculated values of the integrals I52,1,0 , I52,2,0 , I54,4,0 and I51,2,3
GQTTM
Formula
1-6
GQTTS
199
Formula 6
Formula 5
Formula 4
Formula 3
Formula 2
Formula 1
Case
Method
Points
8
27
64
125
216
343
512
729
1000
110
208
350
544
110
208
350
544
110
208
350
544
110
208
350
544
110
208
350
544
110
208
350
544
222
I111
0.0000269860345093
0.0000202862454377
0.0000231837802848
0.0000233422437574
0.0000233426594248
0.0000233426697419
0.0000233426703367
0.0000233426703927
0.0000233426704000
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
022
I111
0.0006015651986381
0.0008488381955233
0.0008546314116172
0.0008546969203429
0.0008547004410159
0.0008547007916240
0.0008547008422886
0.0008547008517505
0.0008547008538885
0.0008547006761305
0.0008547008477305
0.0008547008541789
0.0008547008546419
0.0008547006761305
0.0008547008477305
0.0008547008541789
0.0008547008546419
0.0008547003467509
0.0008547008418419
0.0008547008538989
0.0008547008546175
0.0008547001755466
0.0008547008358275
0.0008547008535749
0.0008547008545902
0.0008547001755466
0.0008547008358275
0.0008547008535749
0.0008547008545902
0.0008547003467509
0.0008547008418419
0.0008547008538989
0.0008547008546175
222
I101
0.0000333182575959
0.0000258267656879
0.0000288727421023
0.0000292327112870
0.0000292495089719
0.0000292513882514
0.0000292517007873
0.0000292517681974
0.0000292517857518
0.0000292495203673
0.0000292516825132
0.0000292517802690
0.0000292517912085
0.0000292484738197
0.0000292516351558
0.0000292517791266
0.0000292517918411
0.0000292511271472
0.0000292517169499
0.0000292517800403
0.0000292517906566
0.0000292511271472
0.0000292517169499
0.0000292517800403
0.0000292517906566
0.0000292495203673
0.0000292516825132
0.0000292517802690
0.0000292517912085
0.0000292484738197
0.0000292516351558
0.0000292517791266
0.0000292517918411
GQTTM
GQTTS
Method
Points
27
64
125
216
343
512
729
1000
14
28
48
75
110
154
208
273
350
M
3
4
5
6
7
8
9
10
2
3
4
5
6
7
8
9
10
IJ0,0,0
0.057883923504
0.057917371950
0.057918451577
0.057918486772
0.057918487951
0.057918487991
0.057918487993
0.057918487993
0.057106321695
0.057890655765
0.057917552866
0.057918455127
0.057918486659
0.057918487915
0.057918487985
0.057918487992
0.057918487993
200
Chapter 9
Accurate Gaussian
Quadrature Rules for
Numerical Integration
over Tetrahedron
Chapter 9
Accurate Gaussian Quadrature
Rules for Numerical Integration
over Tetrahedron
This chapter presents an accurate Gaussian Quadrature formulae to evaluate
tetrahedral domain integrals in Euclidean three dimensional space (x, y, z). The
tetrahedral element T in (x, y, z) space is transformed into a standard tetrahedron
V : {(u, v, w)| 1 u, v, w 1, u+v +w 1} whose base triangle is a standard
triangle in w = 1 plane. The derivation is based on the transformation of the arbitrary tetrahedron in global space into its contiguous standard tetrahedron in the
local space. As a result, it presents two quadrature schemes GQTSS and GQTSM
and each scheme contains six Gaussian quadrature formulae. It is shown that the
scheme GQTSS suffers from huge crowding of Gaussian points whereas GQTSM
is free of such crowdings. The efficiency and accuracy of the schemes are tested by
evaluating the integral of some polynomial and nonpolynomial functions. Finally,
element matrices for tetrahedral finite elements are evaluated through the newly
derived formulae. In all the test cases the results are found satisfactory and the
scheme GQTSM is faster. Thus, it ensures the balance between accuracy and
efficiency of evaluation of tetrahedral domain integrals encountered in practice.
202
9.1
Introduction
203
9.2
Most computational studies of triple integrals deal with problems in which the
domain of integration is very simple, like a cube or a sphere, but the integrand
is complicated. However, in real applications, we confront the mixed problems:
both of the integrating function f (x, y, z) and the domain are very complicated.
In finite element method we divide a complex domain into finite number of simple
element. The finite element method for three dimensional problems with tetrahedron element requires the numerical integration of expressions containing product
of shape functions and their global derivatives over an arbitrary tetrahedron T
with vertices (x1 , y1 , z1 ), (x2 , y2 , z2 ), (x3 , y3 , z3 ) and (x4 , y4 , z4 ) in the natural coordinate space (x, y, z).
The numerical integration of an arbitrary function f , over the tetrahedron T
204
ZZZ
f (x, y, z) dx dy dz
I =
(9.1)
9.3
In this section, we first obtain the volume integral of a function f over an arbitrary tetrahedron T in the cartesian coordinate system (x, y, z) by transforming
it to a normalized orthogonal tetrahedron S. The transformation from arbitrary
tetrahedron in (x, y, z)-space into a standard tetrahedral domain S with vertices
{(1, -1, -1), (-1, 1, -1), (-1, -1, 1), (-1, -1, -1)} is effortlessly accomplished by using
shape functions:
x1 x2 x3 x4
x
y = y1 y2 y3 y4
z1 z2 z3 z4
z
205
N1
N2
N3
N4
(9.2)
Figure 9.2: The Arbitrary Tetrahedron and The Transformed Standard Tetrahedron.
where (xi , yi , zi ) refer to the Cartesian coordinates of vertex i of the tetrahedron
and
N4
u+1
,
2
w+1
,
2
1 u v w
= 1 N1 N2 N3 =
2
N1 =
N2 =
v+1
,
2
N3 =
206
(x1 x4 ) (x2 x4 ) (x3 x4 )
1
Jacobian = (y1 y4 ) (y2 y4 ) (y3 y4 )
8
(z1 z4 ) (z2 z4 ) (z3 z4 )
vol
=
8
(9.4)
and
vol
|du dv dw
8
So, using Eq.(9.3) and Eq.(9.4), Eq.(9.1) becomes
dx dy dz = |
ZZZ
=
f (x, y, z) dx dy dz
Z ZTZ
|
=
Z ZSZ
=
vol
| f {x(u, v, w), y(u, v, w), z(u, v, w)} du dv dw
8
vol
f (u, v, w) du dv dw
8
(9.5)
9.4
Through several inspections and rigorous study of the nature of the integrand,
it is observed that due to ineffective contribution of Gauss points the scheme
become inefficient in view of accuracy. The main objective of this chapter is to
present suitable alternative Gauss quadrature schemes in the sense that the scheme
ensures the effective use of all the Gauss points. By use of calculus and geometrical
concept the integral of Eq.(9.5) can be calculated in six ways (as shown in previous
chapter) by rearranging the order of variables in the integration process. Using
isoparametric transformations the domain of these integrals can be transformed
to a 2-cube given by {(, , )| 1 , , 1} in (, , ) space. These six cases
are as follows:
Case - 1: For the integral I 3
207
vol
I =
8
3
1uv
f (u, v, w) dw dv du
u=1
v=1
w=1
Using transformation
u=
v=
(1 )(1 + )
(1 )(1 )(1 + )
1 w =
1
2
4
Now the determinant of the Jacobian and the differential area are:
(u, v, w)
(, , )
u v w v w
u
v
w
v
w
u
v
w
v
w
}
{
}+
{
}
= {
2
(1 ) (1 )
=
8
and
(u, v, w)
(1 )2 (1 )
dw dv du =
d d d =
d d d
(, , )
8
I = vol
f (, , )
=1
=1
=1
(1 )2 (1 )
d d d
64
1uw
f (u, v, w) dv dw du
u=1
w=1
v=1
Using transformation
u=
v=
(1 )(1 + )(1 )
(1 )(1 + )
1 w =
1
4
2
208
(9.6)
(u, v, w)
(1 )2 (1 )
d d d =
d d d
(, , )
8
dv dw du =
I = vol
f (, , )
=1
=1
=1
(1 )2 (1 )
d d d
64
(9.7)
vol
I =
8
3
1uv
f (u, v, w) dw du dv
v=1
u=1
w=1
(1 + )(1 )
1 v =
2
w=
(1 )(1 )(1 + )
1
4
I = vol
f (, , )
=1
=1
=1
(1 )(1 )2
d d d
64
(1 + )(1 )(1 )
1 v =
4
209
w=
(1 )(1 + )
1
2
(9.8)
vol
I =
8
3
1vw
f (u, v, w) du dw dv
can be evaluated as
Z 1
Z
3
I = vol
=1
v=1
w=1
u=1
f (, , )
=1
=1
(1 )2 (1 )
d d d
64
(9.9)
vol
I =
8
3
1vw
f (u, v, w) du dv dw
w=1
v=1
u=1
With transformation
u=
(1 + )(1 )(1 )
(1 + )(1 )
1 v =
1 w =
4
2
we have
Z
I = vol
f (, , )
=1
=1
=1
(1 )(1 )2
d d d
64
1uw
f (u, v, w) dv du dw
w=1
u=1
v=1
(1 + )(1 )
(1 )(1 + )(1 )
1 v =
1 w =
2
4
210
(9.10)
I = vol
f (, , )
=1
=1
=1
(1 )(1 )2
d d d
64
(9.11)
In order to evaluate the integral I 3 in Eq. 9.6 - 9.11 available Gaussian quadrature coefficients (points and weights) are computed in this study so that preferred
precision can be readily obtained.
9.5
9.5.1
vol X X X
f (u(i , j , j ), v(i , j , j ), w(i , j , j ))
=
8 i=1 j=1 k=1
mnp
X
(u, v, w)
Wi Wj Wk
=
vol{
f (ur , vr , wr )Gr }
(, , )
r=1
(9.12)
Where (ur , vr , wr ) are the new Gaussian points, Gr is the corresponding weights.
All the Gaussian points and analogous weight can be calculated simply by using
the following algorithm:
211
step 1. r 1
step 2. i = 1, m
step 3. j = 1, n
step 4. k = 1, p
For Case 1:
(1 i )(1 + j )
1,
2
(1 i )2 (1 j )
Gr =
Wi Wj Wk
64
ur = i ,
wr =
vr =
(1 i )(1 j )(1 + k )
1,
4
For Case 2:
(1 i )(1 + j )(1 k )
1,
4
(1 i )2 (1 k )
Gr =
Wi Wj Wk
64
ur = i ,
vr =
wr =
(1 i )(1 + k )
1,
2
For Case 3:
(1 + i )(1 j )
1, vr = j ,
2
(1 i )(1 j )2
Gr =
Wi Wj Wk
64
ur =
wr =
(1 i )(1 j )(1 + k )
1,
4
For Case 4:
(1 + i )(1 j )(1 k )
1,
4
(1 j )2 (1 k )
Gr =
Wi Wj Wk
64
ur =
vr = j ,
212
wr =
(1 j )(1 + k )
1,
2
ur =
vr =
(1 + j )(1 k )
1,
2
wr = k ,
For Case 6:
(1 + i )(1 k )
(1 i )(1 + j )(1 k )
1, vr =
1,
2
4
(1 i )(1 k )2
Wi Wj Wk
Gr =
64
r=r+1
ur =
wr = k ,
9.5.2
In Eq.(9.6) to Eq.(9.11) it can be seen that for each i (i = 1, 2, ..., m); j varies
from 1 to n and k varies from 1 to p hence at the terminal value i = m there
are n k crowding points as shown in Table 9.1 and Fig. 9.3. To avoid such
crowding, keeping the generality, we can select the terminal values of the inner
dummy indices depending on the present and terminal value of its outer indices.
Therefore, the new Gaussian quadrature formulae now can be written as:
213
Figure 9.3: Distribution of Gauss Points for GQTSS, m =15 (3375 points).
n(i) p(j)
vol X X X
=
Wi Wj Wk f (u(i0 , j0 , j0 ), v(i0 , j0 , j0 ), w(i0 , j0 , j0 ))
8 i=1 j=1 k=1
tp
X
(u, v, w)
(9.13)
= vol{
G0r f (u0r , vr0 , wr0 )}
(, , )
r=1
Where (u0r , vr0 , wr0 ) are the new Gaussian points, G0r is the corresponding weights.
Now six cases arise for the evaluation of Eq.(9.13). All the Gaussian points and
corresponding weight for the six cases can be calculated simply by using the following algorithm:
214
step 1. r 1
step 2.
For Case 1.
i = 1, m; j = 1, (m i + 2 = n); k = 1, (n j + 2 = p)
(1 i0 )(1 + j0 )
1,
2
(1 i0 )2 (1 j0 )
G0r =
Wi Wj Wk
64
u0r = i0 ,
wr0 =
vr0 =
(1 i0 )(1 j0 )(1 + k0 )
1,
4
For Case 2.
i = 1, m; k = 1, (m i + 2 = n); j = 1, (n k + 2 = p)
(1 i0 )(1 + j0 )(1 k0 )
1,
4
(1 i0 )2 (1 k0 )
Wi Wj Wk
G0r =
64
u0r = i0 ,
vr0 =
wr0 =
(1 i0 )(1 + k0 )
1
2
For Case 3
j = 1, m; i = 1, (m j + 2 = n); k = 1, (n i + 2 = p)
(1 + i0 )(1 j0 )
1; vr0 = j0 ,
2
(1 i0 )(1 j0 )2
G0r =
Wi Wj Wk
64
u0r =
wr0 =
(1 i0 )(1 j0 )(1 + k0 )
1
4
For Case 4.
j = 1, m; k = 1, (m j + 2 = n); i = 1, (n k + 2 = p)
(1 + i0 )(1 j0 )(1 k0 )
1,
4
(1 j0 )2 (1 k0 )
0
Gr =
Wi Wj Wk
64
u0r =
vr0 = j0 ,
215
wr0 =
(1 j0 )(1 + k0 )
1
2
vr0
(1 + j0 )(1 k0 )
=
1,
2
wr0 = k0
For Case 6.
k = 1, m; i = 1, (m k + 2 = n); j = 1, (n i + 2 = p)
(1 i0 )(1 + j0 )(1 k0 )
(1 + i0 )(1 k0 )
, vr0 =
,
4
8
(1 i0 )(1 k0 )2
Wi Wj Wk
G0r =
64
r=r+1
u0r =
wr0 =
1 + k0
2
Figure 9.4: Distribution of Gauss Points for GQTSM, m =15 (950 points).
9.6
To verify the accuracy and efficiency of the derived quadrature formulae and to
compare the results of the projected techniques, we reflect on the same typical
integrals with known correct values following Rathod and Rao [1996, 1998];
Rathod et al. [2004, 2005]; Reddy and Shippy [1981].
217
1x
1xy
p
(x + y + z) dz dy dx
I1 =
x=0
1
y=0
1x
z=0
1xy
I2 =
y=0
x=0
1x
z=0
dz dy dx
p
(x + y + z)
1xy
sin(x + 2y + 4z) dz dy dx
I3 =
x=0
Z 1
y=0
Z 1x
z=0
Z 1xy
I4 =
y=0
x=0
(1 + x + y + z)4 dz dy dx
z=0
The results are summarized in Table 9.5. It is seen from the Table that both
of the methods GQTSS and GQTSM can successfully evaluate the results of the
considered integrals but the method GQTSM calculate the integrals with significantly less number of points, the results of 350 points GQTSM is approximately
equal or better than 1000 points GQTSS.
Consider the monomial integrals
I5,,
ZZZ
=
x y z dz dy dx
1x
1xy
=
x=0
y=0
z=0
218
x y z
p
(ax + by + cz)
dz dy dx
Ni Nj
dx0 dy 0 dz 0
t s
t, s = x0 , y 0 , z 0
V0
We shall evaluate these triple integrals, where V 0 is the tetrahedral element with
three plane surface and one curved surface. It can be shown that the evaluation
of integrals now amounts to the evaluation of integrals of the type
IJpqr
xp y q z r
dx0 dy 0 dz 0
J(x, y, z)
V
Z 1 Z x Z 1xy p q r
xy z
=
dx dy dz
J(x, y, z)
x=1 y=1 z=1
ZZZ
(9.14)
20
X
Ni (x, y, z)ti ,
t = x0 , y 0 , z 0 and
i=1
219
(x + 1)(y + 1)(z + 1)
For the 20 nodal data of Fig 8.7 in page 190 we have
x0 = u + Ru(v + w) + Sxyz,
y 0 = v + Rv(u + w) + Suvw,
z 0 = w + Rw(u + v) + Suvw
x+1
y+1
z+1
where u =
, v=
, w=
2
2
2
1
J(x, y, z) =
[1 + 2R(u + v + w)
8
+ R2 (u2 + v 2 + w2 ) + (2R2 + s)(uv + vw + wu)
+ RS(u2 v + uv 2 + vw2 + wv 2 + u2 w + uw2 )
+ 4R3 uvw + R2 S(u2 vw + v 2 uw + w2 uv)]
Where R = 0.842329219213245 and S = 1.534481952840430. We have computed the values of the integrals of the Eq.9.14 for different values of p, q, r. Results
are summarized in Table 9.8. For all case studies in Table 9.8 for IJpqr in Eq.(9.14)
the method GQTSS require more Gaussian points than the method GQTSM. Further, it is observed that the absence of any independent variable in the explicit
form of the integrand enhances faster convergence of one among the six quadrature
formulae for both GQTSS and GQTSM.
Comparison between the results of Chapter 8, 9 and existing methods of
Rathod et al. [2004, 2005] are also done. It is found that all the methods
220
9.7
Error Analysis
22n+1 [n!]4
f 2n (x1 )
(2n + 1)[(2n)!]3
I=
f (x, y, z)dzdydx
z=1
y=1
x=1
I=
z=1
y=1
m
X
{
f (xi , y, z)W (m, i) +
i=1
22m+1 [m!]4
f 2m (x1 , y, z)}dydz
(2m + 1)[(2m)!]3 x
Where fx2m (x1 , y, z) is the 2m-th partial derivative of the function with respect
to x, m points are taken along x direction, W(m,i) is the corresponding weight.
Now integrating with respect to y and z successively, we have
221
I =
m X
n
X
z=1
i=1 j=1
2n+1
4
22m+1 [m!]4
f 2m (x1 , y1 , z)
(2m + 1)[(2m)!]3 x
2
[n!]
f 2n (x2 , y2 , z)}dz
(2n + 1)[(2n)!]3 y
p
m X
n X
X
=
f (xi , yj , zk )W (m, i)W (n, j)W (p, k) +
2(2n+3) [n!]4
22m+3 [m!]4
2m
f
(x
,
y
,
z
)
+
fy2n (x2 , y2 , z2 )
1
1
1
x
3
3
(2m + 1)[(2m)!]
(2n + 1)[(2n)!]
(2p+3)
4
2
[p!]
f 2p (x3 , y3 , z3 )
+
(9.15)
(2p + 1)[(2p)!]3 z
Where fx2m (x1 , y1 , z1 ), fy2n (x2 , y2 , z2 ) and fz2p (x3 , y3 , z3 ) are the 2m-th, 2n-th
and 2p-th partial derivative of the function with respect to x, y and z respectively,
(x1 , y1 , z1 ), (x2 , y2 , z2 ) and (x3 , y3 , z3 ) are points somewhere in the cubic domain
{(x, y, z)| 1 x, y, z 1}.
Consider the cubic domain monomial integrals
Z
JJ =
z=1
y=1
N
X
(
xi y ji z N j )dzdydx
x=1 i=0
The computed errors for the integral JJ using mmm Gaussian quadrature
formula and the error formula given by Eq.(9.15) are given in Table 9.9. The table
9.9 shows that the error formula gives exactly same results for the error calculated
using the m m m Gaussian quadrature method. Thus, this error formula can
be used to evaluate the error in the integral over cubic domain.
We know, the n-point Gaussian Quadrature rule gives exact results for polynomials of degree at most 2n-1. Thus with n = 2, we have a rule with 8 nodes
which is exact for any polynomial of degrees at most 3 in x, y and z separately, so
222
1xy
xm y n z p dz dy dx
It =
x=1
y=1
z=1
The absolute error of the integral of all monomials xi y j z k where i , j and k are
non-negative integers are calculated with the methods GQTSS and GQTSM. In
Table 9.10 we present the absolute error over corresponding polynomial integral
of order up to m + n + p = 20. It is observed that the method GQTSM is
always accurate for the integration of monomial /polynomial functions and hence
a proper balance is observed.
Figure 9.5, 9.6, 9.7 and 9.8 shows the percentage error in the calculated results
of I1 , I2 , I3 , I4 . The figures proves the efficiency of the method GQTSM with
very less number of Gauss points to obtain an accurate results. It is seen from
these figures that, the error decreases with increase in number of total Gaussian
integration points. These figures also show that, if SGQTS gives accurate results
for a value N, SGQTM gives the results of same accuracy for N or N+1.
Figure 9.9, 9.10, 9.11 and 9.12 shows the percentage error in the calculated
results of the monomial integrals I5,, . Same observation is found in these figures.
Figure 9.13, 9.14, 9.15 and 9.16 shows the error precisions calculated for the
,,
. These figures show the sucresults of the integrals of rational functions Ia,b,c
cess of the newly derived method GQTSM to calculate the integral of rational
functions.
223
9.8
Conclusions
224
Table 9.1: Computed Gauss points (u, v, w ) and corresponding weights G for
n n n point method (GQTSS).
u
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
v
-0.910580956592710D+00
-0.910580956592710D+00
-0.910580956592710D+00
-0.910580956592710D+00
-0.910580956592710D+00
-0.560119750320643D+00
-0.560119750320643D+00
-0.560119750320643D+00
-0.560119750320643D+00
-0.560119750320643D+00
-0.469100770306681D-01
-0.469100770306681D-01
-0.469100770306681D-01
-0.469100770306681D-01
-0.469100770306681D-01
0.466299596259307D+00
0.466299596259307D+00
0.466299596259307D+00
0.466299596259307D+00
0.466299596259307D+00
0.816760802531374D+00
0.816760802531374D+00
0.816760802531374D+00
0.816760802531374D+00
0.816760802531374D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.230765344947159D+00
-0.230765344947159D+00
-0.230765344947159D+00
-0.230765344947159D+00
-0.230765344947159D+00
0.183443909068528D+00
0.183443909068528D+00
0.183443909068528D+00
0.183443909068528D+00
0.183443909068528D+00
0.466299596259307D+00
0.466299596259307D+00
0.466299596259307D+00
0.466299596259307D+00
0.466299596259307D+00
-0.953089922969332D+00
-0.953089922969332D+00
-0.953089922969332D+00
-0.953089922969332D+00
-0.953089922969332D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.500000000000000D+00
-0.500000000000000D+00
-0.500000000000000D+00
-0.500000000000000D+00
-0.500000000000000D+00
w
-0.914775610806955D+00
-0.580754566717371D+00
-0.916195987343129D-01
0.397515369248745D+00
0.731536413338329D+00
-0.931215772989438D+00
-0.661628867873342D+00
-0.266850201870347D+00
0.127928464132649D+00
0.397515369248745D+00
-0.955290478296355D+00
-0.780059875160321D+00
-0.523455038515334D+00
-0.266850201870347D+00
-0.916195987343129D-01
-0.979365183603272D+00
-0.898490882447301D+00
-0.780059875160321D+00
-0.661628867873342D+00
-0.580754566717371D+00
-0.995805345785755D+00
-0.979365183603272D+00
-0.955290478296355D+00
-0.931215772989438D+00
-0.914775610806955D+00
-0.931215772989438D+00
-0.661628867873342D+00
-0.266850201870347D+00
0.127928464132649D+00
0.397515369248745D+00
-0.944484555064120D+00
-0.726902158098188D+00
-0.408278045465736D+00
-0.896539328332845D-01
0.127928464132648D+00
-0.963915143076812D+00
-0.822487299481423D+00
-0.615382672473579D+00
-0.408278045465736D+00
-0.266850201870347D+00
-0.983345731089503D+00
-0.918072440864658D+00
-0.822487299481423D+00
-0.726902158098188D+00
-0.661628867873342D+00
-0.996614513164185D+00
-0.983345731089503D+00
-0.963915143076812D+00
-0.944484555064120D+00
-0.931215772989438D+00
-0.955290478296355D+00
-0.780059875160321D+00
-0.523455038515334D+00
-0.266850201870347D+00
-0.916195987343129D-01
-0.963915143076812D+00
-0.822487299481423D+00
-0.615382672473579D+00
-0.408278045465736D+00
-0.266850201870347D+00
-0.976544961484666D+00
-0.884617327526421D+00
-0.750000000000000D+00
-0.615382672473579D+00
-0.523455038515334D+00
225
G
0.143931111875796D-02
0.290762936018300D-02
0.345595267890959D-02
0.290762936018300D-02
0.143931111875796D-02
0.234673477706453D-02
0.474076441808048D-02
0.563478196881094D-02
0.474076441808048D-02
0.234673477706453D-02
0.181302550558012D-02
0.366258977790325D-02
0.435328420053336D-02
0.366258977790325D-02
0.181302550558012D-02
0.704005022097696D-03
0.142219819279516D-02
0.169039758699556D-02
0.142219819279516D-02
0.704005022097696D-03
0.708413695548069D-04
0.143110438979225D-03
0.170098332250659D-03
0.143110438979225D-03
0.708413695548069D-04
0.189403924354977D-02
0.382624996229895D-02
0.454780756738284D-02
0.382624996229895D-02
0.189403924354977D-02
0.308814939594078D-02
0.623853573785862D-02
0.741500432997716D-02
0.623853573785862D-02
0.308814939594078D-02
0.238582292068215D-02
0.481972846729571D-02
0.572863712833023D-02
0.481972846729571D-02
0.238582292068215D-02
0.926424538886224D-03
0.187151975285516D-02
0.222445260461427D-02
0.187151975285516D-02
0.926424538886224D-03
0.932226064642663D-04
0.188323972528047D-03
0.223838273981466D-03
0.188323972528047D-03
0.932226064642663D-04
0.951130350812901D-03
0.192142928470617D-02
0.228377412016423D-02
0.192142928470617D-02
0.951130350812901D-03
0.155077706458653D-02
0.313280767814902D-02
0.372359532341708D-02
0.313280767814902D-02
0.155077706458653D-02
0.119808953233352D-02
0.242032473378445D-02
0.287675171467764D-02
0.242032473378445D-02
0.119808953233352D-02
v
-0.230765344947159D+00
-0.230765344947159D+00
-0.230765344947159D+00
-0.230765344947159D+00
-0.230765344947159D+00
-0.469100770306681D-01
-0.469100770306681D-01
-0.469100770306681D-01
-0.469100770306681D-01
-0.469100770306681D-01
-0.978349559785040D+00
-0.978349559785040D+00
-0.978349559785040D+00
-0.978349559785040D+00
-0.978349559785040D+00
-0.893494711142838D+00
-0.893494711142838D+00
-0.893494711142838D+00
-0.893494711142838D+00
-0.893494711142838D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.560119750320643D+00
-0.560119750320643D+00
-0.560119750320643D+00
-0.560119750320643D+00
-0.560119750320643D+00
-0.995598889345954D+00
-0.995598889345954D+00
-0.995598889345954D+00
-0.995598889345954D+00
-0.995598889345954D+00
-0.978349559785040D+00
-0.978349559785040D+00
-0.978349559785040D+00
-0.978349559785040D+00
-0.978349559785040D+00
-0.953089922969332D+00
-0.953089922969332D+00
-0.953089922969332D+00
-0.953089922969332D+00
-0.953089922969332D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.910580956592710D+00
-0.910580956592710D+00
-0.910580956592710D+00
-0.910580956592710D+00
-0.910580956592710D+00
w
-0.989174779892520D+00
-0.946747355571419D+00
-0.884617327526421D+00
-0.822487299481423D+00
-0.780059875160321D+00
-0.997799444672977D+00
-0.989174779892520D+00
-0.976544961484666D+00
-0.963915143076812D+00
-0.955290478296355D+00
-0.979365183603272D+00
-0.898490882447301D+00
-0.780059875160321D+00
-0.661628867873342D+00
-0.580754566717371D+00
-0.983345731089503D+00
-0.918072440864658D+00
-0.822487299481423D+00
-0.726902158098188D+00
-0.661628867873342D+00
-0.989174779892520D+00
-0.946747355571419D+00
-0.884617327526421D+00
-0.822487299481423D+00
-0.780059875160321D+00
-0.995003828695537D+00
-0.975422270278180D+00
-0.946747355571419D+00
-0.918072440864658D+00
-0.898490882447301D+00
-0.998984376181768D+00
-0.995003828695537D+00
-0.989174779892520D+00
-0.983345731089503D+00
-0.979365183603272D+00
-0.995805345785755D+00
-0.979365183603272D+00
-0.955290478296355D+00
-0.931215772989438D+00
-0.914775610806955D+00
-0.996614513164185D+00
-0.983345731089503D+00
-0.963915143076812D+00
-0.944484555064120D+00
-0.931215772989438D+00
-0.997799444672977D+00
-0.989174779892520D+00
-0.976544961484666D+00
-0.963915143076812D+00
-0.955290478296355D+00
-0.998984376181768D+00
-0.995003828695537D+00
-0.989174779892520D+00
-0.983345731089503D+00
-0.979365183603272D+00
-0.999793543560198D+00
-0.998984376181768D+00
-0.997799444672977D+00
-0.996614513164185D+00
-0.995805345785755D+00
226
G
0.465222935413471D-03
0.939821730277485D-03
0.111705414415182D-02
0.939821730277485D-03
0.465222935413471D-03
0.468136289636074D-04
0.945707152938267D-04
0.112404944502805D-03
0.945707152938267D-04
0.468136289636074D-04
0.170456069100769D-03
0.344347420567744D-03
0.409284761972482D-03
0.344347420567744D-03
0.170456069100769D-03
0.277921277830243D-03
0.561443635574862D-03
0.667321173390394D-03
0.561443635574862D-03
0.277921277830243D-03
0.214714662336039D-03
0.433756571552638D-03
0.515554769799693D-03
0.433756571552638D-03
0.214714662336039D-03
0.833745582383381D-04
0.168429403668663D-03
0.200191969714883D-03
0.168429403668663D-03
0.833745582383381D-04
0.838966726974746D-05
0.169484155008367D-04
0.201445626995949D-04
0.169484155008367D-04
0.838966726974746D-05
0.348673721407182D-05
0.704374430430755D-05
0.837205984069912D-05
0.704374430430755D-05
0.348673721407182D-05
0.568497482726903D-05
0.114845216605656D-04
0.136502829220052D-04
0.114845216605656D-04
0.568497482726903D-05
0.439206188153587D-05
0.887263907855757D-05
0.105458492105091D-04
0.887263907855757D-05
0.439206188153587D-05
0.170545511491618D-05
0.344528107924556D-05
0.409499523510995D-05
0.344528107924556D-05
0.170545511491618D-05
0.171613514481643D-06
0.346685637877664D-06
0.412063922372617D-06
0.346685637877664D-06
0.171613514481643D-06
Table 9.3: Computed Gauss points (p, q, r) and corresponding weights L for
method GQTSM.
u
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.906179845938664D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
-0.538469310105683D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
0.000000000000000D+00
v
-0.935637374493801D+00
-0.935637374493801D+00
-0.935637374493801D+00
-0.935637374493801D+00
-0.935637374493801D+00
-0.935637374493801D+00
-0.935637374493801D+00
-0.677102080244399D+00
-0.677102080244399D+00
-0.677102080244399D+00
-0.677102080244399D+00
-0.677102080244399D+00
-0.677102080244399D+00
-0.274335618713707D+00
-0.274335618713707D+00
-0.274335618713707D+00
-0.274335618713707D+00
-0.274335618713707D+00
0.180515464652371D+00
0.180515464652371D+00
0.180515464652371D+00
0.180515464652371D+00
0.583281926183063D+00
0.583281926183063D+00
0.583281926183063D+00
0.841817220432465D+00
0.841817220432465D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.927830286153624D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.644974598962845D+00
-0.230765344947159D+00
-0.230765344947159D+00
-0.230765344947159D+00
-0.230765344947159D+00
0.183443909068528D+00
0.183443909068528D+00
0.183443909068528D+00
0.466299596259307D+00
0.466299596259307D+00
-0.930568155797026D+00
-0.930568155797026D+00
-0.930568155797026D+00
-0.930568155797026D+00
-0.930568155797026D+00
-0.669990521792428D+00
-0.669990521792428D+00
-0.669990521792428D+00
-0.669990521792428D+00
-0.330009478207572D+00
-0.330009478207572D+00
-0.330009478207572D+00
227
-0.694318442029738D-01
w
-0.953133038284567D+00
-0.761973843346101D+00
-0.452837684101723D+00
-0.790913897837677D-01
0.294654904534188D+00
0.603791063778566D+00
0.794950258717031D+00
-0.946540101185744D+00
-0.731799472415783D+00
-0.397259759191488D+00
-0.194583146254489D-01
0.315081398598846D+00
0.529822027368808D+00
-0.944621928617262D+00
-0.727577941584041D+00
-0.409742267673815D+00
-0.919065937635887D-01
0.125137393269633D+00
-0.949615783734883D+00
-0.760523876177890D+00
-0.513811742535817D+00
-0.324719834978824D+00
-0.963608866696046D+00
-0.838551040122200D+00
-0.713493213548354D+00
-0.986398576827778D+00
-0.949238797666023D+00
-0.950490037970444D+00
-0.751615730079521D+00
-0.441793809977105D+00
-0.919065937635887D-01
0.217915326338827D+00
0.416789634229750D+00
-0.944484555064120D+00
-0.726902158098188D+00
-0.408278045465736D+00
-0.896539328332845D-01
0.127928464132648D+00
-0.946590619274843D+00
-0.746145272866830D+00
-0.484620072080328D+00
-0.284174725672316D+00
-0.959988046051174D+00
-0.822487299481423D+00
-0.684986552911672D+00
-0.984748744935084D+00
-0.943081541218540D+00
-0.956346976129275D+00
-0.785257118530658D+00
-0.534715922101487D+00
-0.284174725672316D+00
-0.113084868073699D+00
-0.953481322473439D+00
-0.778896777499262D+00
-0.551112700708310D+00
-0.376528155734133D+00
-0.962807382215067D+00
-0.834995260896214D+00
-0.707183139577362D+00
-0.985327324868973D+00
G
0.576651691559072D-03
0.124564721228047D-02
0.170045180553786D-02
0.186135022002191D-02
0.170045180553786D-02
0.124564721228047D-02
0.576651691559072D-03
0.138110348004552D-02
0.290821853363949D-02
0.377200921144364D-02
0.377200921144364D-02
0.290821853363949D-02
0.138110348004552D-02
0.184705677022988D-02
0.373133815549181D-02
0.443499720802718D-02
0.373133815549181D-02
0.184705677022988D-02
0.166697114369810D-02
0.312517468091849D-02
0.312517468091849D-02
0.166697114369810D-02
0.913358455696345D-03
0.146137352911416D-02
0.913358455696345D-03
0.155625745624003D-03
0.155625745624003D-03
0.136960105591409D-02
0.288399763815690D-02
0.374059429546915D-02
0.374059429546915D-02
0.288399763815690D-02
0.136960105591409D-02
0.308814939594078D-02
0.623853573785862D-02
0.741500432997716D-02
0.623853573785862D-02
0.308814939594078D-02
0.350285305275700D-02
0.656701689938599D-02
0.656701689938599D-02
0.350285305275700D-02
0.217231699669361D-02
0.347570719470980D-02
0.217231699669361D-02
0.393465715983046D-03
0.393465715983046D-03
0.136344637490265D-02
0.275437093414734D-02
0.327379264321987D-02
0.275437093414734D-02
0.136344637490265D-02
0.270201938829868D-02
0.506564412442597D-02
0.506564412442597D-02
0.270201938829868D-02
0.212557118293120D-02
0.340091389268994D-02
0.212557118293120D-02
0.429372505117035D-03
v
-0.930568155797026D+00
-0.930568155797026D+00
-0.930568155797026D+00
-0.930568155797026D+00
-0.930568155797026D+00
-0.669990521792428D+00
-0.669990521792428D+00
-0.669990521792428D+00
-0.669990521792428D+00
-0.330009478207572D+00
-0.330009478207572D+00
-0.330009478207572D+00
-0.694318442029738D-01
-0.694318442029738D-01
-0.947984722625272D+00
-0.947984722625272D+00
-0.947984722625272D+00
-0.947984722625272D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.769234655052841D+00
-0.590484587480411D+00
-0.590484587480411D+00
-0.980173468570694D+00
-0.980173468570694D+00
-0.980173468570694D+00
-0.926006377367970D+00
-0.926006377367970D+00
228
w
-0.956346976129275D+00
-0.785257118530658D+00
-0.534715922101487D+00
-0.284174725672316D+00
-0.113084868073699D+00
-0.953481322473439D+00
-0.778896777499262D+00
-0.551112700708310D+00
-0.376528155734133D+00
-0.962807382215067D+00
-0.834995260896214D+00
-0.707183139577362D+00
-0.985327324868973D+00
-0.945240830928054D+00
-0.971566589679223D+00
-0.864856032396452D+00
-0.725628555083959D+00
-0.618917997801187D+00
-0.973992361312636D+00
-0.884617327526421D+00
-0.795242293740205D+00
-0.989007878509772D+00
-0.958976844115500D+00
-0.991660795501926D+00
-0.963003188683985D+00
-0.934345581866044D+00
-0.995810160914250D+00
-0.984363307656444D+00
G
0.136344637490265D-02
0.275437093414734D-02
0.327379264321987D-02
0.275437093414734D-02
0.136344637490265D-02
0.270201938829868D-02
0.506564412442597D-02
0.506564412442597D-02
0.270201938829868D-02
0.212557118293120D-02
0.340091389268994D-02
0.212557118293120D-02
0.429372505117035D-03
0.429372505117035D-03
0.546317509560028D-03
0.102421547911849D-02
0.102421547911849D-02
0.546317509560028D-03
0.786674148253923D-03
0.125867863720628D-02
0.786674148253923D-03
0.199483844892810D-03
0.199483844892810D-03
0.285550084723027D-04
0.456880135556845D-04
0.285550084723027D-04
0.137723246280313D-04
0.137723246280313D-04
229
Exact
Value
GQTSM
GQTSS
Method
8
27
64
125
216
343
512
729
1000
14
28
48
75
110
154
208
273
350
Points
0.14285714285714
I1
0.143127410953799
0.142875312759849
0.142860037924269
0.142857834882225
0.142857355360141
0.142857220464426
0.142857175072351
0.142857157618562
0.142857150174568
0.143058065243440
0.142868320929028
0.142858821339166
0.142857541720436
0.142857267261819
0.142857189383667
0.142857162683365
0.142857152181787
0.142857147595743
0.2
I2
0.197660776240555
0.199583323221216
0.199881018522955
0.199956079429135
0.199980842555110
0.199990593978398
0.199994953918813
0.199997099779897
0.199998238575602
0.198325920889004
0.199730676421398
0.199924597609756
0.199971858116752
0.199987455481037
0.199993687508278
0.199996530092608
0.199997959524722
0.199998734389269
I4
0.020103982733156
0.020798626362385
0.020831943224904
0.020833282126123
0.020833331531397
0.020833333271718
0.020833333331268
0.020833333333265
0.020833333333331
0.020334811873381
0.020811918661132
0.020832494969141
0.020833301701749
0.020833332126155
0.020833333282340
0.020833333330381
0.020833333333016
0.020833333333269
0.131902326890181 0.020833333333333
I3
0.131949528497795
0.131902664864685
0.131902289803505
0.131902327570123
0.131902326884088
0.131902326890216
0.131902326890182
0.131902326890182
0.131902326890182
0.133142199123991
0.131877204546620
0.131902589866991
0.131902325527761
0.131902326955921
0.131902326902415
0.131902326892685
0.131902326890580
0.131902326890160
Value of I
230
Exact
Value
GQTSM
GQTSS
Method
8
27
64
125
216
343
512
729
1000
14
28
48
75
110
154
208
273
350
440
544
663
798
950
Points
15721.6666666667
15613.2458847736452
15721.6666666665296
15721.6666666666770
15721.6666666666642
15721.6666666666770
15721.6666666666661
15721.6666666666879
15721.6666666666661
15721.6666666667024
15612.4547963692658
15721.6666928552586
15721.6666681218594
15721.6666668061935
15721.6666666861638
15721.6666666702495
15721.6666666674864
15721.6666666668807
15721.6666666667461
15721.6666666666024
15721.6666666665897
15721.6666666665933
15721.6666666665951
15721.6666666665969
I52,1,0
109662.063492063
17357405.82972
I54,0,4
19959073.9239162914455
17447147.0802131444216
17356864.7296515144408
17357404.7323902957141
17357405.8297258540988
17357405.8297257982194
17357405.8297258764505
17357405.8297258913517
17357405.8297258727252
15916008.4708074014634
17395025.3200442716479
17356769.8780943527818
17357410.5770655572414
17357406.3276728168130
17357405.8860965371132
17357405.8378492966294
17357405.8311539664865
17357405.8300211951137
17357405.8297953382134
17357405.8297439366579
17357405.8297308571637
17357405.8297272846103
17357405.8297261968255
Value of I
107754.6467764059344
109657.4916666657664
109662.0634920635057
109662.0634920633602
109662.0634920636367
109662.0634920634475
109662.0634920636076
109662.0634920634766
109662.0634920636221
107739.9768428653770
109657.4923773875344
109662.0635323561437
109662.0634959668678
109662.0634926121420
109662.0634921646124
109662.0634920864541
109662.0634920696466
109662.0634920655139
109662.0634920636658
109662.0634920631564
109662.0634920630109
109662.0634920630400
109662.0634920631128
I52,2,0
Table 9.6: Calculated values of the integrals I52,1,0 , I52,2,0 , I54,4,0 and I51,2,3
330565.079365081
I51,2,3
380177.2290809325059
331027.3599999966100
330568.5208940719021
330565.0793650789419
330565.0793650799314
330565.0793650795240
330565.0793650801061
330565.0793650801061
330565.0793650796986
329118.0069088268210
330824.6705146196764
330568.9540617458988
330565.1115077321883
330565.0829795532045
330565.0799310719012
330565.0794796128175
330565.0793933367240
330565.0793732203892
330565.0793677304173
330565.0793660327327
330565.0793654504814
330565.0793652340653
330565.0793651472777
GQTSS
GQTSM
Case
1-6
231
Case 6
Case 5
Case 4
Case 3
Case 2
Case 1
Case
Method
Points
8
27
64
125
216
343
512
729
1000
110
208
350
544
110
208
350
544
110
208
350
544
110
208
350
544
110
208
350
544
110
208
350
544
222
I111
0.0000269860345093
0.0000202862454377
0.0000231837802848
0.0000233422437574
0.0000233426594248
0.0000233426697419
0.0000233426703367
0.0000233426703927
0.0000233426704000
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
0.0000233426460487
0.0000233426700306
0.0000233426703855
0.0000233426704001
022
I111
0.0006015651986381
0.0008488381955233
0.0008546314116172
0.0008546969203429
0.0008547004410159
0.0008547007916240
0.0008547008422886
0.0008547008517505
0.0008547008538885
0.0008547006761305
0.0008547008477305
0.0008547008541789
0.0008547008546419
0.0008547006761305
0.0008547008477305
0.0008547008541789
0.0008547008546419
0.0008547003467509
0.0008547008418419
0.0008547008538989
0.0008547008546175
0.0008547001755466
0.0008547008358275
0.0008547008535749
0.0008547008545902
0.0008547001755466
0.0008547008358275
0.0008547008535749
0.0008547008545902
0.0008547003467509
0.0008547008418419
0.0008547008538989
0.0008547008546175
222
I101
0.0000333182575959
0.0000258267656879
0.0000288727421023
0.0000292327112870
0.0000292495089719
0.0000292513882514
0.0000292517007873
0.0000292517681974
0.0000292517857518
0.0000292495203673
0.0000292516825132
0.0000292517802690
0.0000292517912085
0.0000292484738197
0.0000292516351558
0.0000292517791266
0.0000292517918411
0.0000292511271472
0.0000292517169499
0.0000292517800403
0.0000292517906566
0.0000292511271472
0.0000292517169499
0.0000292517800403
0.0000292517906566
0.0000292495203673
0.0000292516825132
0.0000292517802690
0.0000292517912085
0.0000292484738197
0.0000292516351558
0.0000292517791266
0.0000292517918411
GQTSM
GQTSS
Method
Points
27
64
125
216
343
512
729
1000
14
28
48
75
110
154
208
273
350
440
544
663
798
950
3
4
5
6
7
8
9
10
2
3
4
5
6
7
8
9
10
11
12
13
14
15
IJ0,0,0
0.057883923504
0.057917371950
0.057918451577
0.057918486772
0.057918487951
0.057918487991
0.057918487993
0.057918487993
0.057106321695
0.057890655765
0.057917552866
0.057918455127
0.057918486659
0.057918487915
0.057918487985
0.057918487992
0.057918487993
0.057918487993
0.057918487993
0.057918487993
0.057918487993
0.057918487993
Value of I
IJ2,1,2
IJ1,2,1
0.002450805315 0.003530383143
0.002403948085 0.003416615059
0.002414474857 0.003423823551
0.002413862879 0.003424492584
0.002413804476 0.003424516105
0.002413802396 0.003424516819
0.002413802334 0.003424516841
0.002413802332 0.003424516841
0.001951020091 0.002545145635
0.002450792220 0.003530376620
0.002403953091 0.003416615436
0.002414475280 0.003423823630
0.002413862920 0.003424492594
0.002413804481 0.003424516106
0.002413802396 0.003424516819
0.002413802334 0.003424516841
0.002413802332 0.003424516841
0.002413802332 0.003424516842
0.002413802332 0.003424516842
0.002413802332 0.003424516842
0.002413802332 0.003424516842
0.002413802332 0.003424516842
232
Table 9.9: Error calculated using the new error formula and the method GQTSS
Points N Error calculated using error formula Error using m m m points
22 4
2.133333333
2.133333333
33 6
0.548571428
0.548571428
44 8
0.139319727
0.139319727
5 5 10
0.035181749
0.035181749
R 1 R 1x R 1xy
Table 9.10: Calculated Errors in the integral x=0 y=0 z=0 xi y j z k dz dy dx for
GQTSS
i, j, k
3, 4, 5
5, 7, 8
7, 4, 9
8, 7, 5
6,10, 4
6, 6, 6
TP
73
103
93
103
103
103
GQTTS
Absolute error
0.407419326993121D-12
0.693087437300628D-12
0.693691154321356D-16
0.295766425612876D-16
0.634844425226502D-16
0.818942693206072D-17
233
TP
208
350
440
350
273
544
GQTTM
Absolute error
0.465243339007592D-12
0.693061853399468D-12
0.668925855145236D-16
0.941456653102749D-16
0.588036204043632D-16
0.854762493675243D-17
2,2,2
Figure 9.13: Error Precision in I1,1,1
.
0,2,2
Figure 9.14: Error Precision in I1,1,1
.
238
2,2,2
Figure 9.15: Error Precision in I1,0,1
.
2,2,0
Figure 9.16: Error Precision in I1,1,0
.
239
Chapter 10
Appropriate Gauss
Quadrature and
Recursive Schemes to
Obtain Element Matrices
for General Triangular
Finite Elements
Chapter 10
Appropriate Gauss Quadrature
and Recursive Schemes to Obtain
Element Matrices for General
Triangular Finite Elements
This chapter concentrates to show the applications of the derived appropriate
Gaussian quadrature formulae along with the recursive scheme to form the element matrices for the general (straight sided or curved sided) triangular finite
elements. Also it presents a recursive formula to evaluate the integrals when the
Jacobean is a nonlinear function of special type. The formulation of element matrices is one of the most important, time consuming and sensitive task in the
FEM solution procedure. To demonstrates the application, this chapter presents
the Prandtle stresses and torsional constant k for different types of cross sections. Also, the curved triangular finite elements are applied to integrate some
complicated functions over the curved domain in the shape of a lunar model. Obtained results are compared with the results obtained by others and found in good
agreement.
241
10.1
Introduction
The finite element method has proved superior to other numerical methods due to
its better adaptability to any complex geometry. If the domain of a problem consist a polygonal boundary, element with straight sides, usually triangle (in 2D)
or tetrahedron (in 3D) are perfectly satisfactory to represent the domain. But
the domain of real problems often contains curved boundaries, however, elements
with at least one curved side or face are desirable to represent the domain accurately. In classical finite element applications curved boundaries are discretized
by extremely refined meshes because simplifying the curved domains by polygonal domains may cause global changes in the physical solution of the problem.
When dealing with curved or irregular boundaries, the side of the element should
closely approximate the boundary which requires many elements in order to reduce the error in the solution. Many researchers strongly suggested that a large
number of lower order elements should be used in the portion of the domain where
the function varies rapidly to get the desired accuracy, which generally requires
much more computational effort, computing memory and time. Alternatively, a
fewer number of higher order elements are suitable to achieve better accuracy with
less computational effort. For irregular, complicated or distorted type of domain
the integrals cannot be evaluated analytically or explicitly and numerical integration scheme are the only alternative. Complication arise from two main sources:
firstly the large number of integrations that need to be performed and secondly
in methods which use isoparametric or subparametric elements, the presence of
the determinant of Jacobean matrix in the denominator of the stiffness matrix for
which the integrands are rational functions.
Among all the numerical rules, Gaussian quadrature rule occupies a central
role for such calculations. To get the desired accuracy for the polynomial of higher
degree it is necessary to increase the Gauss points with corresponding weights.
In addition specially for two and three dimensional cases the higher order (more
Gauss points and weights) formulae become erroneous due to crowding of Gaussian
points and that effects the accuracy and efficiency of the calculations. Hence, it
is an important task to make a proper balance between accuracy and efficiency of
calculations. The previous chapters of this Thesis successfully derived numerical
242
10.2
Problem Statement
In the FEM solution procedure of physical problems, formulation of element matrices is one of the most important, time consuming and sensitive task. By using
the Galerkins weighted residual FE procedure (Discussed in Section 2.3 in page
30) for any problem we achieve the following FE equations:
[K]{u} = {F }
243
ZZ
{
Ni Nj Ni Nj
ij
ij
+
}dxdy = kxx
+ kyy
x x
y y
(10.1)
Fie
ZZ
f (x, y)Ni dxdy
(10.2)
The Lagrange interpolation function for the Prandtl stress function u(x, y) is
u=
NP
X
ui Ni (, )
i=1
10.3
ij
ij
For curved elements to calculate kxx
and kyy
we have to calculate the integrals of
rational function of a monomial numerator with a linear or nonlinear denominator.
Rathod and Karim [2002] showed that for a rational integral of n-degree bivariate polynomial numerator with a linear denominator there are (n+1)(n+2)
rational
2
integral of monomial numerator with the same linear denominator. Rathod and
Karim [2001] presented a recursive scheme to compute n(n+1)
integrals whenever
2
(n + 1) integrals of order 0 to n in one of the variables are known. In this section we wish to present numerical integration schemes to compute the integrals
in Eq.10.1 and Eq.10.2 in an efficient manner. Here, two cases depending on the
Jacobian of transformations will be discussed.
10.3.1
To compute kxx and kyy for a curved element when the Jacobian is a linear function
of the type J(, ) = 0 + 1 + 2 we need to calculate
244
ij
ktz
ZZ
=
Z
Ni Nj
dtdz
t z
Ni Nj
|J|dd
t z
=0 =0
Z 1 Z 1
n X
m
X
mk k
=
dd
Cij
0
1
2
=0
=0
m=0 k=0
n X
m
X
Cij Imk,k
(10.3)
m=0 k=0
where
Z
Imk,k =
=0
=0
mk k
d d
(0 + 1 + 2 )
1 (m k)! k!
{
0 Imk,k 2 Imk,k+1 }
1 (m + 2)!
1 (m k)! k!
= {
0 Imk,k 1 Imk+1,k }
2 (m + 2)!
Imk+1,k =
if 1 2
Imk,k+2
if 2 > 1
(10.4)
10.3.2
ij
ktz
ZZ
=
Ni Nj
dtdz
t z
Ni Nj
|J|dd
t z
Z 1 Z 1
=
=0
=0
n X
m
X
m=0 k=0
n X
m
X
Cij
=0
=0
mk k
dd
0 + 1 + 2 + 3
Cij IImk,k
(10.5)
m=0 k=0
where
Z
IImk,k =
=0
=0
mk k
d d
(0 + 1 + 2 + 3 )
m = 0, n; k = 0, m.
(10.6)
Now, we know
Z
=0
p q d d =
=0
p! q!
(p + q + 2)!
(10.7)
then
=0
mk
d d =
=0
=0
=0
mk k (0 + 1 + 2 + 3 )
dd
(0 + 1 + 2 + 3 )
(m k)! k!
(m + 2)!
(10.8)
1 (m k)! k!
{
0 IImk,k 1 IImk+1,k 2 IImk,k+1 } (10.9)
3 (m + 2)!
Now if IIr,0 and II0,r for r = 0 to n ie. (2n+1) integrals are known or calculated
by the derived quadrature formulae (so presented in previous chapters) then the
246
10.4
To show the application of the derived quadrature formula so presented in previous chapters we consider the following two dimensional boundary value torsion
problem following Lewis and Ward [1991]; Segerlind [1984].
2u
x2
2u
y 2
+2=0
u = 0 on C1 ,
and
within A
u
n
= 0 on C2
2
(10.10)
where A is the interior of the quarter ellipse x62 + y22 = 1 and the boundary of
A = C1 + C2 for which the finite element equation is of the form
[k]{{u} = {F }.
and the components of element matrices are given by Eq.10.1 and Eq.10.2.
The domain of the problem is shown in Figure 10.1 and FE models using the
quadratic and the cubic curved triangular elements are shown in Fig 10.2a and
10.2b respectively.
10.4.1
If nodes 5, 6 are at the mid points of two straight sides (Fig 10.2a), then the
transformation formulae from quadratic curved element to unit triangular element
is
247
(10.11)
(10.12)
10.4.2
If nodes 6, 7, 8 and 9 are at the trisection points of two straight sides (Fig 10.2b)
and following Karim [2001] if chosen
x5 = x4 31 (x1 x2 ),
x10 =
1
(x1
12
1
y5 = y4 (y1 y2 )
3
1
y10 = (y1 + y2 + 4y3 + 3y4 + 3y5 )
12
then the transformation formulae from cubic curved element to unit triangular
element is
248
(10.13)
10.4.3
Calculation Procedure
249
10.4.4
k
kexact
Computed Results
The geometry and FE model for this problem are shown in Figures 10.1, 10.2a,
10.2b. To show the application of the quadratic and cubic curved triangular
element we shall illustrate the solution by use of a single element.
We have four cases on account of the numerical values of 1 and 2 in the
Jacobian, which are
Case 1. 1 6= 2 6= 0,
Case 2. 1 = 2 6= 0,
Case 3. 1 = 0, 2 6= 0,
Case 4. 1 6= 0, 2 = 0
The cases can be attributed to a choice of point along the curved side of the
triangle. For six node element considering a = 6, b = 2.
x10
b
a
x4 = (1 + 17), y4 = (1 + 17)
6
6
a
b
x5 = (1 + 17), y5 = (1 + 17)
6
6
1
1
= (a + 3x4 + 3x5 ), y10 = (b + 3y4 + 3y5 )
12
12
Computed results of element of stiffness matrix and load vectors are given in
Table 10.4. It is found that the calculated results with SGQTS and SGQTM
completely matched with results of the analytical integration formula given by
Karim [2001] for M = 4 (16 points) and M = 5 (14 points) respectively.
10.5
Finite element models with straight sided quadratic and cubic triangular elements
are shown in the Figure 10.3. The cross section is modelled by eight six node
element and four ten node element as shown in Fig 10.3. Results are tabulated
in Table 10.5. Nguyen [1992] has shown that the approximate result for k is
erroneous, it should be greater than 0.1406 and finally the computed value of k
251
10.6
(10.14)
u = 0 on the boundary of .
For this problem we have to determine
ZZ
Fi =
252
10.7
The curved triangular finite elements are applied to integrate some complicated
functions which are very recently studied by Rathod et al. [2013] over the curved
domain in the shape of a lunar model. The theoretical developments of the preRR
vious chapters used to compute the integral f (x, y)dxdy over the domain in
the shape of a lunar model. The boundary of the domain is composed of two
circular arcs. The outer circular arcs satisfies the equations
(x 0.5)2 + (y 0.5)2 = 0.25
253
f1 (x, y) = (x + y)19
p
(x 0.5)2 + (y 0.5)2
f2 (x, y) =
2 +(y0.5)2
f3 (x, y) = e(x0.5)
f4 (x, y) = e100(x0.5)
2 +(y0.5)2
ZZ
f1 (x, y)dxdy =
ZZ
ZZ p
(x 0.5)2 + (y 0.5)2 dxdy = 0.20646770293563
f2 (x, y)dxdy =
ZZ
ZZ
f3 (x, y)dxdy =
2 +(y0.5)2
e(x0.5)
254
dxdy = 0.57263720432530
ZZ
f4 (x, y)dxdy =
ZZ
2 +(y0.5)2
dxdy = 0.03137185199242
ZZ
cos[20(x + y)]dxdy = 0.0062895812195655
f5 (x, y)dxdy =
ZZ
ZZ
1dxdy = 0.6426990816987241
f6 (x, y)dxdy =
e100(x0.5)
Using the numerical scheme SGQTS, SGQTM results are summarized in Table
10.7. The distribution of Gaussian points (for m =10) in the Lunar domain for
these two methods SGQTS, SGQTM are shown in Fig 10.5 and 10.6. Fig 10.5
shows that there are a huge crowding of Gaussian points with in the lunar model
domain, whereas Fig 10.6 is free of such crowding. The present schemes are
shown efficient to evaluate the integrals of these complicated test functions using
significantly less number of Gaussian points and weights with desired accuracy.
10.8
Conclusions
257
Table 10.1: Computed Prandtl stress values ui , torsional constant k, error Ek for
Example -1.
SGQTM
SGQTS
Analytic
13 13
77
Method
u3
u5
u6
k
u3
u5
u6
k
u3
u5
u6
k
u3
u5
u6
k
Ek
M
u3
u5
u6
k
Ek
M
258
259
SGQTM
SGQTS
Method
condition
condition
condition
condition
condition
condition
condition
condition
1
2
3
4
1
2
3
4
Condition
k33
1.408954947029
1.340335082143
1.398016972353
1.345959227348
1.408954687936
1.340336710677
1.398009859097
1.345946297006
k66
3.473000140410
2.512902376386
2.768895479951
2.801709347663
3.473000094361
2.512902878264
2.768845175747
2.801697361016
Table 10.3: Computed Values for six node FE model for example -1 (same for
both SGQTS and SGQTM)
F3
F5
F6
Condition 1
Condition 2
Condition 3
Condition 4
-0.211928851253 -0.331370866666 -0.292820323027 -0.292820323027
4.259644256269 5.988225173333 6.342562584220 5.171281292110
6.283501958777 5.988225226666 5.171281292110 6.342562584220
260
K
3
6
7
8
9
10
261
6
1.0336408565
6.6949714454
-4.6536505668
-0.3055757768
0.1409903483
0.0887848905
7
-1.8618026534
-4.6536505668
5.6678980122
-0.4448290735
0.0233451488
0.0786123665
8
0.1834227902
-0.3055757768
-0.4448290735
3.4970038177
-0.2907481188
-3.6294305476
9
0.0014101027
0.1409903483
0.0233451488
-0.2907481188
2.7984437166
-1.4636074881
10
-0.2181688406
0.0887848905
0.0786123665
-3.6294305476
-1.4636074881
1.0176103061
3
1.2104154852
1.0336408565
-1.8618026534
0.1834227902
0.0014101027
-0.2181688406
F
0.5684658437
1.65809628
0.8999999999
0.8999999999
1.6580963137
8.4323851875
Table 10.4: Computed elements of Stiffness matrix of TEN node cubic FE model using SGQTM.
8 six node
4 ten node
262
SGQTM
Analytical M.
SGQTM
Analytical M.
FE Model
i
7
8
9
7
8
9
5
6
7
9
10
5
6
7
9
10
Computed Stress
ui
i
0.0960802472
12
0.1236618019
13
0.0903051441
14
0.09608024972341 12
0.12366180644874 13
0.09030514551905 14
0.055549792
12
0.089792221
13
0.0550452768
14
0.1241786087
16
0.1238791629
17
0.059023000487
12
0.096295652764
13
0.058562968131
14
0.129507227925
16
0.129106499063
17
function values ui
ui
0.123643778
0.1621862722
0.1209024325
0.123643783156
0.162186275133
0.120902433437
0.094800726
0.1620487414
0.0923284681
0.1318099076
0.1343235984
0.094725716001
0.163109710489
0.092548242198
0.136907410754
0.134420821735
at node i
i
ui
17
0.0972297317
18
0.1233425114
19
0.100282778
17 0.097297319135
18 0.123342514484
19 0.100282779360
19
0.056065628
20
0.101048889
21
0.06697125
19 0.056777163927
20 0.101949643739
21 0.066973467265
0.16780884
0.167526745
0.1605573640
0.1605573601
Torsional
Constant k
Table 10.5: Computed Stress function values ui , the torsional constant k for example-2
Table 10.6: Computed Values for six node FE model for example-3, m=10
Method
Gauss 7-point
Gauss 13-point
Analytical M.
SGQTS
SGQTM
K10,10
7.811317529
8.262921381
8.274075346
8.274075342
8.274075300
263
F10
u10
5.368507707 0.687273009
5.385059531 0.651713756
5.385091615 0.6508390835
5.372745751 0.6493469698
5.302350861 0.6408390870
264
f1
f2
f3
f4
f5
f6
f1
f2
f3
f4
f5
f6
f1
f2
f3
f4
f5
f6
f
f1
f2
f3
f4
f5
f6
SGQTS
Ea
SGQTM
Ea
0.6385586779312916D+03 0.1245184271624566D-02 0.6386421161691412D+03 0.8468342212120206D-01
0.2064676940601683D+00 0.8875461671165041D-08 0.2064676832778503D+00 0.1965777970958094D-07
0.5726372020008313D+00 0.1645301561659096D-08 0.5726372042263089D+00 0.3870779141657010D-08
0.3115537787347361D-01 0.2164741189463852D-03 0.3165191572921538D-01 0.2800637367953787D-03
0.5290947148691215D-02 0.9986340708742856D-03 0.1046493985035890D-01 0.4175358630793395D-02
0.6426990807542486D+00 0.9444754978105152D-09 0.6426990807542489D+00 0.9444752757659103D-09
Number of triangle = 200, (m=5) SGQTS with 25 points, SGQTM with 14 points
0.6385586866716534D+03 0.1253924633374481D-02 0.6386421544879713D+03 0.8472174095129503D-01
0.2064676999357412D+00 0.2999888781829441D-08 0.2064676970158540D+00 0.5919775947216266D-08
0.5726372015421841D+00 0.1186654330354031D-08 0.5726372029562166D+00 0.2600686777043393D-08
0.3115544111954633D-01 0.2164108728736729D-03 0.3165178537247875D-01 0.2799333800587514D-03
0.5290956980767337D-02 0.9986242387981636D-03 0.1046457069760205D-01 0.4174989478036545D-02
0.6426990816396912D+00 0.5903288968767129D-10 0.6426990816396914D+00 0.5903277866536882D-10
Number of triangle = 400, (m=5) SGQTS with 25 points, SGQTM with 14 points
0.6385574325131282D+03 0.2338917965971632D-06 0.6385574316997154D+03 0.1047304635903856D-05
0.2064677035093190D+00 0.5736890129615091D-09 0.2064677023774669D+00 0.5581630713180630D-09
0.5726372045185323D+00 0.4163002498280832D-08 0.5726372035271892D+00 0.3171659379574976D-08
0.3137186096244889D-01 0.8970028893273696D-08 0.3137186784505314D-01 0.1585263313674812D-07
0.6289581650472803D-02 0.4309073028951893D-09 0.6289566892772273D-02 0.1432679322746488D-07
0.6426990816396914D+00 0.5903277866536882D-10 0.6426990816396915D+00 0.5903266764306636D-10
Number of triangle = 400, (m=10) SGQTS with 100 points, SGQTM with 54 points
0.6385574323024241D+03 0.4445959120857879D-06 0.6385574325546615D+03 0.1923584704854875D-06
0.2064677030899757D+00 0.1543457028851947D-09 0.2064677035359275D+00 0.6002975061480953D-09
0.5726372033211775D+00 0.2965647727393161D-08 0.5726372045429263D+00 0.4187396540622501D-08
0.3137185186267821D-01 0.1297417936974021D-09 0.3137185184658459D-01 0.1458354062511269D-09
0.6289581219460250D-02 0.1052500100962028D-12 0.6289581346177302D-02 0.1266118017720363D-09
0.6426990816396916D+00 0.5903255662076390D-10 0.6426990816396912D+00 0.5903288968767129D-10
Number of triangle = 400, (m=20) SGQTS with 400 points, SGQTM with 209 points
Table 10.7: Computed Integrals and Absolute Error for ten node Lunar model
Chapter 11
Conclusions
Chapter 11
Conclusions
The main concern of this thesis was to develop suitable quadrature schemes to
evaluate the triangular/tetrahedral domain integrals encountered in many areas
of science and engineering.
The employment of general triangular/ tetrahedral finite elements in FEM
solution procedure give rise to huge number of triangular or tetrahedral domain
integrals of rational functions of n-th degree bivariate or trivariate polynomial
numerators with linear or nonlinear denominators. Generally, such integrals as
well as domain integrals of nonpolynomial, irrational functions are evaluated by
numerical integration schemes that employ Gaussian quadrature.
In some recent studies, many researchers have also shown that the most accurate 7 - points and 13 - points Gaussian quadrature rules for the triangle are not
satisfactory in view of accuracy and for some element geometry these formulae are
not at all reliable. Hence, among the others Nguyen [1992] first proposed a technique to evaluate the triangular domain integrals using the quadrature rules for
square domain integrals to achieve the desired accuracy. This technique seriously
affected by the huge crowding of quadrature points and that leads to computational inefficiency. Similar technique for tetrahedral domain integrals lead to
severe computational inefficiency. Thus, the higher order quadrature rules for triangles and tetrahedrons which ensures the efficiency and accuracy of calculations
are highly demanded.
Therefore, this thesis was fully devoted to develop higher order quadrature
schemes for triangular and tetrahedral domain integrals ensuring the accuracy and
efficiency of evaluations. Furthermore, it is concentrated to present the recursive
266
11. Conclusions
schemes to compute the components of element matrices efficiently. The gradual
development of the research work, relevant topics and demonstrations are shown
through out the thesis.
In Chapter 1 the objectives and scope of the Thesis, important applications
and relevant literature review are described. In Chapter 2 some basic concepts
and a brief historical background of FEM, elements, Interpolation functions are
discussed. Some problems and some mathematical preliminaries are also included.
The numerical integration scheme IOUT that can evaluate the triangular domain integrals using the quadrature schemes for square domain integrals is presented in Chapter 3. This technique is almost similar to the extended Gaussian
quadrature for triangles. It is demonstrated that the scheme IOUT so developed
is always reliable and appropriate in view of accuracy.
Chapter 4 presented the numerical integration schemes IOST by transforming
the triangular domain integrals into square domain integrals using the transformation of arbitrary triangle first to standard triangle, then standard triangle is
subdivided into three quadrilaterals. The reliability and accuracy of the scheme
is demonstrated through application examples.
The developed quadrature schemes are shown sufficient to obtain the desired
accuracy of evaluation of the triangular domain integrals. As it is so, such schemes
so developed in Chapter 3 and 4 are used to test the accuracy of the higher
order quadrature schemes developed in other chapters.
The higher order Gaussian quadrature schemes GQUTS, GQUTM; GQSTS,
GQSTM; SGQTS and SGQTM are respectively presented in Chapter 5, 6 and
7. It is thoroughly investigated that the developed quadrature schemes based
on different technique/approaches are also different but their accuracy is almost
same. Any one of the higher order quadrature schemes so presented may be used
for obtaining the exact value of the triangular domain integrals efficiently and
hence its direct implication enhanced the versatility of triangular finite elements.
Chapter 8 developed two sets of quadrature formulae to evaluate the tetrahedral domain integrals. It is shown that the quadrature formulae of one set GQTTS
exhibits crowding of Gauss quadrature points but the quadrature schemes of the
other set GQTTM are completely free of such crowding. The accuracy and efficiency of the schemes are investigated through practical application examples.
267
11. Conclusions
Chapter 9 also derived quadrature schemes for the tetrahedral domain integrals using a standard tetrahedral element but this concept is different from the
concept of Chapter 8. Here also, two sets of quadrature formulae GQTSS and
GQTSM are presented among which the scheme GQTSM is totally free of crowding of quadrature points. The efficiency and accuracy of the presented quadrature
schemes of Chapter 8 and 9 are thoroughly investigated. It is observed that all
the schemes are accurate in view of accuracy but the schemes free of crowdings
of quadrature points are so efficient as they employ very less number of Gauss
points.
It is our firm belief that the thesis may be considered as the complete reference
for the higher order quadrature schemes to evaluate the triangular or tetrahedral
domain integrals encountered in many areas of continuum mechanics. The computer codes in FORTRAN for such schemes enhances the contribution of the
thesis. So, researchers may use such computer codes whenever needed.
268
Appendix A
The shape functions Ni which transforms a quadratic curved triangular element to a orthogonal unit triangle are listed below:
N1 (, ) = + 2 2 ,
N2 (, ) = + 2 2 ,
N3 (, ) = 1 3 3 + 2 2 + 2 2 + 4,
N4 (, ) = 4,
N5 (, ) = 4 4 4 2 ,
N6 (, ) = 4 4 4 2
269
Appendix A
Shape functions for a cubic curved triangular element:
N1 (, ) =
N2 (, ) =
N3 (, ) =
N4 (, ) =
N5 (, ) =
N6 (, ) =
N7 (, ) =
N8 (, ) =
N9 (, ) =
N10 (, ) =
1
(2 9 2 + 9 3 ),
2
1
(2 9 2 + 9 3 ),
2
1
(2 11 11 + 18 2 + 18 2 + 36 9 3 9 3 27 2 27 2 ),
2
9
( + 3 2 ),
2
9
( + 3 2 ),
2
9
( + + 4 2 3 2 9 3 ),
2
9
(2 5 5 2 + 3 2 + 6 2 + 3 3 ),
2
9
(2 5 5 2 + 3 2 + 6 2 + 3 3 ),
2
9
( + 4 2 + 3 3 3 2 ),
2
27( 2 2 )
270
Appendix A
m=6, 20 points
m=7, 27 points
m=8, 35 points
p
0.888727297463133
0.630192003213731
0.227425541683039
0.697064941206465
0.414209254015687
0.430568155797026
0.169990521792428
0.178750067572431
2.708354560136226E-002
0.000000000000000E+000
0.000000000000000E+000
0.917061053145689
0.716493205000875
0.392141691261992
0.774513554793957
0.549203619608679
0.198198215856174
0.561205871610772
0.333479209306941
0.327826332907397
0.129427521840461
0.131213040406755
1.949437207665818E-002
0.000000000000000E+000
0.000000000000000E+000
0.000000000000000E+000
0.935854278720920
0.776394467312567
0.512159689180237
0.178766956543018
0.826450513922025
0.645699842407919
0.353395993824022
0.655453872674869
0.464779005004411
0.167730812890342
0.453089922969332
0.269234655052842
0.255824157429295
0.101000692742857
0.100104541368752
1.469128025426521E-002
0.000000000000000E+000
0.000000000000000E+000
0.000000000000000E+000
q
-0.906179845938664
-0.906179845938664
-0.906179845938664
-0.538469310105683
-0.538469310105683
0.000000000000000E+000
0.000000000000000E+000
0.538469310105683
0.906179845938664
-0.538469310105683
0.538469310105683
-0.932469514203152
-0.932469514203152
-0.932469514203152
-0.661209386466265
-0.661209386466265
-0.661209386466265
-0.238619186083197
-0.238619186083197
0.238619186083197
0.238619186083197
0.661209386466265
0.932469514203152
-0.932469514203152
-0.238619186083197
0.661209386466265
-0.949107912342758
-0.949107912342758
-0.949107912342758
-0.949107912342758
-0.741531185599394
-0.741531185599394
-0.741531185599394
-0.405845151377397
-0.405845151377397
-0.405845151377397
0.000000000000000E+000
0.000000000000000E+000
0.405845151377397
0.405845151377397
0.741531185599394
0.949107912342758
0.000000000000000E+000
-0.741531185599394
0.741531185599394
271
L
9.671808407443765E-003
2.036612959907326E-002
2.641521865037473E-002
2.180780247074805E-002
4.405510797397064E-002
2.473634454310782E-002
4.637476656800330E-002
1.534040420130362E-002
2.778564607155512E-003
5.236305923555275E-002
2.454464672208581E-002
5.358724942779149E-003
1.157558519997316E-002
1.580200602497686E-002
1.283435698778468E-002
2.702557440373014E-002
3.505263253647306E-002
1.716438073492758E-002
3.467468341195214E-002
1.549087599565812E-002
2.904171054783806E-002
8.487682963359328E-003
1.446203274907969E-003
1.729720730431097E-002
4.121366590559383E-002
1.358029274137493E-002
3.193509426162615E-003
7.015570468089016E-003
9.896667159022517E-003
1.144177445787316E-002
7.884269377752301E-003
1.703110961959155E-002
2.324942473072126E-002
1.149574334237841E-002
2.420682760510148E-002
3.139666969662116E-002
1.237822093354783E-002
2.500590605057915E-002
9.864562308726572E-003
1.849370967345216E-002
5.020494508311843E-003
8.237200310701331E-004
2.972154195011338E-002
2.544930805858548E-002
8.032791213298952E-003
Appendix A
m=9, 44 points
m=10, 54 points
p
0.948937348485512
0.819431649704404
0.601192898894164
0.317815348416544
0.862660296884733
0.715671976824132
0.472103231845533
0.164785436484144
0.723947440393452
0.565614928297778
0.309564965916809
0.551758363089550
0.391249046943736
0.141195105587486
0.369977534931071
0.219847692355784
0.204290635247786
8.065499321191914E-002
7.875073467027130E-002
1.146333102035305E-002
0.0
0.0
0.0
0.0
0.958402053212199
0.851291661485933
0.668593449333590
0.426495791355461
0.146504277333313
0.888786158306404
0.767489559872218
0.563084515391895
0.297669686013274
0.774652110792637
0.642659468024678
0.423939488558651
0.147974105985064
0.628429701049770
0.490987605545342
0.268720915541678
0.466234757101576
0.330604693233132
0.119309593041598
0.306173964336758
0.181934396453014
0.166469949200553
6.572319189508555E-002
6.350487906200106E-002
9.191347145602645E-003
0.0
0.0
0.0
0.0
q
-0.960289856497536
-0.960289856497536
-0.960289856497536
-0.960289856497536
-0.796666477413627
-0.796666477413627
-0.796666477413627
-0.796666477413627
-0.525532409916329
-0.525532409916329
-0.525532409916329
-0.183434642495650
-0.183434642495650
-0.183434642495650
0.183434642495650
0.183434642495650
0.525532409916329
0.525532409916329
0.796666477413627
0.960289856497536
0.183434642495650
-0.960289856497536
-0.525532409916329
0.796666477413627
-0.968160239507626
-0.968160239507626
-0.968160239507626
-0.968160239507626
-0.968160239507626
-0.836031107326636
-0.836031107326636
-0.836031107326636
-0.836031107326636
-0.613371432700590
-0.613371432700590
-0.613371432700590
-0.613371432700590
-0.324253423403809
-0.324253423403809
-0.324253423403809
0.0
0.0
0.0
0.324253423403809
0.324253423403809
0.613371432700590
0.613371432700590
0.836031107326636
0.968160239507626
0.836031107326636
0.324253423403809
-0.324253423403809
-0.836031107326636
272
L
2.015983497663494E-003
4.480916044841954E-003
6.464359484633224E-003
7.747662769908705E-003
5.055663745070523E-003
1.110639128725685E-002
1.566747257515452E-002
1.811354111938598E-002
7.745946956367606E-003
1.673231410556490E-002
2.284153446588211E-002
9.191827856851423E-003
1.935542449754594E-002
2.510431683577034E-002
8.770885597454200E-003
1.771853503082168E-002
6.471997505241379E-003
1.213345702762273E-002
3.140105492486543E-003
5.024749628294020E-004
2.105991205229387E-002
8.191474625434849E-003
2.500282281758626E-002
5.024168787978471E-003
1.333102027244705E-003
2.988298775027559E-003
4.380659742804402E-003
5.384022372785824E-003
5.909044540081589E-003
3.369591872501191E-003
7.489574345999887E-003
1.080477752203405E-002
1.294973968319760E-002
5.320342238493575E-003
1.168784271708034E-002
1.648770968862596E-002
1.906183693488799E-002
6.694804681384016E-003
1.446170176936692E-002
1.974188730358828E-002
7.072261232401939E-003
1.489220864908214E-002
1.931544949367340E-002
6.250944716822827E-003
1.262786770048660E-002
4.381211780382315E-003
8.213730741263730E-003
2.056991588376092E-003
3.234696324495897E-004
3.291186541401748E-003
1.500924217028327E-002
2.160988400642072E-002
1.369154378180941E-002
Appendix A
q
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.010743648455019
0.055731967907878
0.055731967907878
0.055731967907878
0.055731967907878
0.055731967907878
0.055731967907878
0.055731967907878
0.055731967907878
0.055731967907878
0.055731967907878
0.055731967907878
0.133163679476162
0.133163679476162
0.133163679476162
0.133163679476162
0.133163679476162
0.133163679476162
0.133163679476162
0.133163679476162
0.133163679476162
0.133163679476162
0.237314822536963
0.237314822536963
0.237314822536963
0.237314822536963
0.237314822536963
0.237314822536963
0.237314822536963
0.237314822536963
0.237314822536963
0.360463383316429
0.360463383316429
0.360463383316429
0.360463383316429
0.360463383316429
0.360463383316429
0.360463383316429
0.360463383316429
0.493476632129293
0.493476632129293
0.493476632129293
0.493476632129293
0.493476632129293
0.493476632129293
0.493476632129293
0.626489880942157
0.626489880942157
0.626489880942157
0.626489880942157
0.626489880942157
0.626489880942157
0.749638441721623
0.749638441721623
0.749638441721623
0.749638441721623
0.749638441721623
0.853789584782424
0.853789584782424
0.853789584782424
0.853789584782424
0.931221296350708
0.931221296350708
0.931221296350708
0.976209615803566
0.976209615803566
r
0.009000343078836
0.046800828160385
0.112311611011103
0.201432090647445
0.308564484743052
0.426977779097407
0.549231836706159
0.667645131060515
0.774777525156121
0.863898004792463
0.929408787643181
0.967209272724731
0.010136968592262
0.052584856125216
0.125644099593486
0.223913962990053
0.340108484621152
0.465610648175354
0.591112811729556
0.707307333360655
0.805577196757222
0.878636440225491
0.921084327758446
0.011139167091428
0.057603746063275
0.136858385783597
0.241880555591099
0.363341112359269
0.490448472423154
0.611909029191324
0.716931198998827
0.796185838719148
0.842650417690996
0.011934154220146
0.061458692597240
0.144915818357696
0.253282805339145
0.374819220860811
0.496355636382478
0.604722623363926
0.688179749124382
0.737704287501476
0.012439001537527
0.063693197178343
0.148624572011216
0.255784966802195
0.370704914139962
0.477865308930941
0.562796683763814
0.614050879404630
0.012557028009562
0.063774160020431
0.146600766830786
0.246738316064646
0.346875865298507
0.429702472108862
0.480919604119731
0.012171133693667
0.061060805395110
0.137224952088334
0.223238431228095
0.299402577921320
0.348292249622762
0.011132456605728
0.054764036883816
0.118657411268482
0.182550785653147
0.226182365931235
0.009245799846884
0.043945276380129
0.089218403096034
0.123917879629279
0.006281085598081
0.027865983953939
0.049450882309797
0.002270400063718
0.008473248391302
273
L
0.000021086958459
0.000047800933658
0.000071553598188
0.000090814044281
0.000104369100374
0.000111366527676
0.000111366527676
0.000104369100374
0.000090814044281
0.000071553598188
0.000047800933658
0.000021086958459
0.000053546428804
0.000120793127271
0.000179188652286
0.000224304198222
0.000252786187266
0.000262520924912
0.000252786187266
0.000224304198222
0.000179188652286
0.000120793127271
0.000053546428804
0.000087221974090
0.000195517907109
0.000286617065143
0.000352264905689
0.000386615967301
0.000386615967301
0.000352264905689
0.000286617065143
0.000195517907109
0.000087221974090
0.000116860652101
0.000259745564195
0.000374719964555
0.000449109293105
0.000474835765027
0.000449109293105
0.000374719964555
0.000259745564195
0.000116860652101
0.000137086819800
0.000301155285992
0.000424831258156
0.000491157615021
0.000491157615021
0.000424831258156
0.000301155285992
0.000137086819800
0.000143441629306
0.000309853709416
0.000422985974226
0.000463009321186
0.000422985974226
0.000309853709416
0.000143441629306
0.000133493245590
0.000281099524076
0.000364590893663
0.000364590893663
0.000281099524076
0.000133493245590
0.000107845482009
0.000217864425396
0.000258949491601
0.000217864425396
0.000107845482009
0.000070977335331
0.000133065633519
0.000133065633519
0.000070977335331
0.000031981545160
0.000051170472256
0.000031981545160
0.000004919349119
0.000004919349119
Appendix A
q
0.012166494443092
0.012166494443092
0.012166494443092
0.012166494443092
0.012166494443092
0.012166494443092
0.012166494443092
0.012166494443092
0.012166494443092
0.012166494443092
0.012166494443092
0.062916342903180
0.062916342903180
0.062916342903180
0.062916342903180
0.062916342903180
0.062916342903180
0.062916342903180
0.062916342903180
0.062916342903180
0.062916342903180
0.149480367469124
0.149480367469124
0.149480367469124
0.149480367469124
0.149480367469124
0.149480367469124
0.149480367469124
0.149480367469124
0.149480367469124
0.264188373451698
0.264188373451698
0.264188373451698
0.264188373451698
0.264188373451698
0.264188373451698
0.264188373451698
0.264188373451698
0.396850822703577
0.396850822703577
0.396850822703577
0.396850822703577
0.396850822703577
0.396850822703577
0.396850822703577
0.535680860640916
0.535680860640916
0.535680860640916
0.535680860640916
0.535680860640916
0.535680860640916
0.668343309892794
0.668343309892794
0.668343309892794
0.668343309892794
0.668343309892794
0.783051315875368
0.783051315875368
0.783051315875368
0.783051315875368
0.869615340441313
0.869615340441313
0.869615340441313
0.920365188901401
0.920365188901401
0.013367999605811
0.013367999605811
0.013367999605811
0.013367999605811
0.013367999605811
r
0.010018792579021
0.051971825849235
0.124179350182252
0.221303590942994
0.336143525628120
0.460182594450700
0.584221663273280
0.699061597958407
0.796185838719148
0.868393363052166
0.910346396322380
0.011345641543418
0.058671483157382
0.139395178702936
0.246364028614955
0.370075965732413
0.499539374708900
0.623251311826357
0.730220161738377
0.810943857283931
0.858269698897895
0.012466083175355
0.064198028585252
0.151375004189405
0.264572123000961
0.391525657937684
0.518479192874407
0.631676311685963
0.718853287290116
0.770585232700013
0.013270004372377
0.067948299748769
0.158553619796689
0.272872996889125
0.395470313003669
0.509789690096105
0.600395010144026
0.655073305520417
0.013630958658022
0.069228398473477
0.159138690332064
0.267840430320458
0.376542170308851
0.466452462167439
0.522049901982894
0.013399764423026
0.067224666852556
0.151077301196801
0.245773521506776
0.329626155851020
0.383451058280551
0.012393096949226
0.060965521130610
0.132094186725849
0.203222852321088
0.251795276502472
0.010378697585519
0.049329938070762
0.100150429398362
0.139101669883605
0.007090776624761
0.031458171451590
0.055825566278419
0.002571082800639
0.009595411642452
0.010780997661352
0.055751551843689
0.132457833243631
0.234103114061445
0.351658220984276
274
L
0.000059509032276
0.000134243912621
0.000199142006864
0.000249281344617
0.000280934914111
0.000291753652722
0.000280934914111
0.000249281344617
0.000199142006864
0.000134243912621
0.000059509032276
0.000150952285656
0.000338376599161
0.000496039003275
0.000609653624843
0.000669103910383
0.000669103910383
0.000609653624843
0.000496039003275
0.000338376599161
0.000150952285656
0.000242902915463
0.000539899048033
0.000778881259316
0.000933504603098
0.000986978847182
0.000933504603098
0.000778881259316
0.000539899048033
0.000242902915463
0.000317364752095
0.000697193740583
0.000983511523017
0.001137061279546
0.001137061279546
0.000983511523017
0.000697193740583
0.000317364752095
0.000357101778493
0.000771389109565
0.001053034913256
0.001152674107609
0.001053034913256
0.000771389109565
0.000357101778493
0.000350036453042
0.000737079092838
0.000956004198305
0.000956004198305
0.000737079092838
0.000350036453042
0.000293619012570
0.000593155468982
0.000705013252414
0.000593155468982
0.000293619012570
0.000198459095465
0.000372063633308
0.000372063633308
0.000198459095465
0.000091004790589
0.000145607664942
0.000091004790589
0.000014131170069
0.000014131170069
0.000102967932626
0.000230814251757
0.000338359306367
0.000415858382636
0.000456410753006
Appendix A
q
0.013367999605811
0.013367999605811
0.013367999605811
0.013367999605811
0.013367999605811
0.068842731814761
0.068842731814761
0.068842731814761
0.068842731814761
0.068842731814761
0.068842731814761
0.068842731814761
0.068842731814761
0.068842731814761
0.162326928825093
0.162326928825093
0.162326928825093
0.162326928825093
0.162326928825093
0.162326928825093
0.162326928825093
0.162326928825093
0.283713816620238
0.283713816620238
0.283713816620238
0.283713816620238
0.283713816620238
0.283713816620238
0.283713816620238
0.419852392074756
0.419852392074756
0.419852392074756
0.419852392074756
0.419852392074756
0.419852392074756
0.555990967529274
0.555990967529274
0.555990967529274
0.555990967529274
0.555990967529274
0.677377855324419
0.677377855324419
0.677377855324419
0.677377855324419
0.770862052334751
0.770862052334751
0.770862052334751
0.826336784543702
0.826336784543702
0.014230084199161
0.014230084199161
0.014230084199161
0.014230084199161
0.014230084199161
0.014230084199161
0.014230084199161
0.014230084199161
0.014230084199161
0.072864333686846
0.072864333686846
0.072864333686846
0.072864333686846
0.072864333686846
0.072864333686846
0.072864333686846
0.072864333686846
0.170024914572384
0.170024914572384
0.170024914572384
0.170024914572384
r
0.474678563559426
0.592233670482257
0.693878951300071
0.770585232700013
0.815555786882350
0.012272031559499
0.063198698562623
0.149018645439834
0.260453696446561
0.385431026167376
0.510408355888190
0.621843406894918
0.707663353772128
0.758590020775252
0.013449385920162
0.068866812722558
0.160696919295911
0.276561645299257
0.400816210025162
0.516680936028508
0.608511042601861
0.663928469404257
0.014147770528067
0.071853163097369
0.165172364573945
0.277995483764637
0.390818602955329
0.484137804431906
0.541843197001207
0.014176417999888
0.071121024752307
0.159833778001397
0.260018614073359
0.348731367322450
0.405675974074868
0.013309036992320
0.065471316758644
0.141856908310119
0.218242499861594
0.270404779627918
0.011270658032131
0.053569425080607
0.108757503744486
0.151056270792962
0.007758690524781
0.034421365907380
0.061084041289980
0.002824990717434
0.010543008888376
0.011183200273643
0.057591418310226
0.135797023368205
0.237345042281768
0.351233806432731
0.465122570583695
0.566670589497258
0.644876194555237
0.691284412591820
0.012783357625703
0.065456452867118
0.152739032168659
0.262866010269902
0.380967353107876
0.491094331209119
0.578376910510660
0.631050005752075
0.013910659583212
0.070648932977925
0.162404142163887
0.273336391246120
275
L
0.000456410753006
0.000415858382636
0.000338359306367
0.000230814251757
0.000102967932626
0.000260265082353
0.000578489846161
0.000834553981011
0.001000229461794
0.001057525927395
0.001000229461794
0.000834553981011
0.000578489846161
0.000260265082353
0.000410939770525
0.000902761361728
0.001273499961506
0.001472323874040
0.001472323874040
0.001273499961506
0.000902761361728
0.000410939770525
0.000517102354726
0.001117012428919
0.001524850521750
0.001669133370859
0.001524850521750
0.001117012428919
0.000517102354726
0.000546256852972
0.001150264500013
0.001491912742958
0.001491912742958
0.001150264500013
0.000546256852972
0.000482818900053
0.000975368279414
0.001159304092992
0.000975368279414
0.000482818900053
0.000338401945391
0.000634423214647
0.000634423214647
0.000338401945391
0.000158880691017
0.000254209105627
0.000158880691017
0.000024984583805
0.000024984583805
0.000139415696920
0.000309878545114
0.000447043928551
0.000535790995223
0.000566482882934
0.000535790995223
0.000447043928551
0.000309878545114
0.000139415696920
0.000349625324427
0.000768064462517
0.001083486849256
0.001252645153975
0.001252645153975
0.001083486849256
0.000768064462517
0.000349625324427
0.000535672337832
0.001157126154460
0.001579610373778
0.001729074653693
Appendix A
q
0.170024914572384
0.170024914572384
0.170024914572384
0.292615255613059
0.292615255613059
0.292615255613059
0.292615255613059
0.292615255613059
0.292615255613059
0.424082441451564
0.424082441451564
0.424082441451564
0.424082441451564
0.424082441451564
0.546672782492239
0.546672782492239
0.546672782492239
0.546672782492239
0.643833363377778
0.643833363377778
0.643833363377778
0.702467612865463
0.702467612865463
0.014617153391652
0.014617153391652
0.014617153391652
0.014617153391652
0.014617153391652
0.014617153391652
0.014617153391652
0.014617153391652
0.074237047072965
0.074237047072965
0.074237047072965
0.074237047072965
0.074237047072965
0.074237047072965
0.074237047072965
0.170652314741006
0.170652314741006
0.170652314741006
0.170652314741006
0.170652314741006
0.170652314741006
0.287218584745408
0.287218584745408
0.287218584745408
0.287218584745408
0.287218584745408
0.403784854749810
0.403784854749810
0.403784854749810
0.403784854749810
0.500200122417850
0.500200122417850
0.500200122417850
0.559820016099163
0.559820016099163
0.014369232340683
0.014369232340683
0.014369232340683
0.014369232340683
0.014369232340683
0.014369232340683
0.014369232340683
0.072088346222680
0.072088346222680
0.072088346222680
0.072088346222680
0.072088346222680
r
0.384268640328352
0.476023849514315
0.532762122909028
0.014319246644569
0.071837575264130
0.161444117220108
0.262638324231456
0.352244866187434
0.409763194806995
0.013726604181157
0.067525460398349
0.146307627806530
0.225089795214711
0.278888651431902
0.011805143379214
0.056109833340320
0.113915081232065
0.158219771193171
0.008211931753257
0.036432166843423
0.064652401933588
0.003007170628092
0.011222913571069
0.011115266587425
0.056915087943902
0.132808226959586
0.228564815792052
0.331255200307111
0.427011789139577
0.502904928155262
0.548704749511739
0.012728114238126
0.064643066302190
0.148598164008093
0.250100061208925
0.351601958409758
0.435557056115660
0.487472008179724
0.013633893699332
0.068399259338159
0.153717020678344
0.250067834071466
0.335385595411650
0.390150961050477
0.013473445935047
0.066280095784009
0.143609292372704
0.220938488961399
0.273745138810361
0.011848704929974
0.056316881342594
0.114335433398412
0.158803609811032
0.008366638837962
0.037118523536483
0.065870408235004
0.003088967973098
0.011528185418554
0.010463250321044
0.053140360903866
0.122156335037195
0.205596799084250
0.289037263131306
0.358053237264635
0.400730347847457
0.011935151820329
0.059876918699973
0.134564345272441
0.218910139014064
0.293597565586532
276
L
0.001579610373778
0.001157126154460
0.000535672337832
0.000635662488365
0.001338527819612
0.001736093490551
0.001736093490551
0.001338527819612
0.000635662488365
0.000606552151637
0.001225328437748
0.001456401959262
0.001225328437748
0.000606552151637
0.000447572137223
0.000839091376250
0.000839091376250
0.000447572137223
0.000217154553172
0.000347447285074
0.000217154553172
0.000034748745319
0.000034748745319
0.000155709948310
0.000342066976851
0.000482544225219
0.000557880961558
0.000557880961558
0.000482544225219
0.000342066976851
0.000155709948310
0.000384423220178
0.000830407193042
0.001133601389558
0.001240863862768
0.001133601389558
0.000830407193042
0.000384423220178
0.000560512390214
0.001180282683454
0.001530846840634
0.001530846840634
0.001180282683454
0.000560512390214
0.000603540458013
0.001219244354405
0.001449170534097
0.001219244354405
0.000603540458013
0.000480978848727
0.000901721020056
0.000901721020056
0.000480978848727
0.000244790553608
0.000391664885772
0.000244790553608
0.000040163151675
0.000040163151675
0.000143400918949
0.000309765769426
0.000422865926019
0.000462877913901
0.000422865926019
0.000309765769426
0.000143400918949
0.000343450840604
0.000723211630750
0.000938017862639
0.000938017862639
0.000723211630750
Appendix A
q
0.072088346222680
0.162007687132911
0.162007687132911
0.162007687132911
0.162007687132911
0.162007687132911
0.263555143376274
0.263555143376274
0.263555143376274
0.263555143376274
0.353474484286505
0.353474484286505
0.353474484286505
0.411193598168501
0.411193598168501
0.013289732853664
0.013289732853664
0.013289732853664
0.013289732853664
0.013289732853664
0.013289732853664
0.065376353661207
0.065376353661207
0.065376353661207
0.065376353661207
0.065376353661207
0.141651151467688
0.141651151467688
0.141651151467688
0.141651151467688
0.217925949274170
0.217925949274170
0.217925949274170
0.270012570081712
0.270012570081712
0.011129592453413
0.011129592453413
0.011129592453413
0.011129592453413
0.011129592453413
0.052898940541990
0.052898940541990
0.052898940541990
0.052898940541990
0.107396275308498
0.107396275308498
0.107396275308498
0.149165623397075
0.149165623397075
0.007603791647411
0.007603791647411
0.007603791647411
0.007603791647411
0.033734158327754
0.033734158327754
0.033734158327754
0.059864525008097
0.059864525008097
0.002757099674531
0.002757099674531
0.002757099674531
0.010289636066883
0.010289636066883
r
0.341539332466176
0.012363392077610
0.060819393573824
0.131777571688137
0.202735749802450
0.251191751298664
0.011248492492696
0.053464072296347
0.108543614836563
0.150759194640214
0.008124476673733
0.036044173111340
0.063963869548947
0.003036576090371
0.011332656250313
0.009117040014437
0.045738862139902
0.102791195188291
0.167221374893421
0.224273707941810
0.260895530067275
0.010222923067433
0.050289756857191
0.108962974637085
0.167636192416979
0.207703026206737
0.009835100679876
0.046746222583354
0.094904928884335
0.131816050787812
0.007368023934041
0.032688176830603
0.058008329727165
0.002808451006571
0.010481281847093
0.006997370883884
0.034422256537484
0.074582811698537
0.114743366859591
0.142168252513190
0.007456721455199
0.035441788775994
0.071954486532504
0.099939553853299
0.005961798695881
0.026449470270995
0.046937141846109
0.002351959627232
0.008777632826181
0.004156504373647
0.019755860661066
0.040108664347031
0.055708020634450
0.003801895823705
0.016867079163877
0.029932262504048
0.001606870246458
0.005996921400953
0.001159659120884
0.005144818033441
0.009129976945999
0.000582643717629
0.002174455956902
277
L
0.000343450840604
0.000459323439224
0.000927903842541
0.001102888770595
0.000927903842541
0.000459323439224
0.000414540002858
0.000777163975548
0.000777163975548
0.000414540002858
0.000227132676091
0.000363412281746
0.000227132676091
0.000038700788121
0.000038700788121
0.000104510657004
0.000220070280072
0.000285434919692
0.000285434919692
0.000220070280072
0.000104510657004
0.000235648564158
0.000476046266093
0.000565819492374
0.000476046266093
0.000235648564158
0.000267293510290
0.000501111800218
0.000501111800218
0.000267293510290
0.000165763800754
0.000265222081206
0.000165763800754
0.000030024334684
0.000030024334684
0.000053966849766
0.000109021319164
0.000129580656044
0.000109021319164
0.000053966849766
0.000106949181923
0.000200504295923
0.000200504295923
0.000106949181923
0.000084132667633
0.000134612268213
0.000084132667633
0.000016995080924
0.000016995080924
0.000014581587902
0.000027337011495
0.000027337011495
0.000014581587902
0.000020996870944
0.000033594993510
0.000020996870944
0.000005324360227
0.000005324360227
0.000000621553620
0.000000994485792
0.000000621553620
0.000000299780623
0.000000299780623
Appendix A
v
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.9785127030899614
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.8885360641842432
-0.7336726410476753
-0.7336726410476753
-0.7336726410476753
-0.7336726410476753
-0.7336726410476753
-0.7336726410476753
-0.7336726410476753
-0.7336726410476753
-0.7336726410476753
-0.7336726410476753
-0.5253703549260738
-0.5253703549260738
-0.5253703549260738
-0.5253703549260738
-0.5253703549260738
-0.5253703549260738
-0.5253703549260738
-0.5253703549260738
-0.5253703549260738
-0.2790732333671420
-0.2790732333671420
-0.2790732333671420
-0.2790732333671420
-0.2790732333671420
-0.2790732333671420
-0.2790732333671420
-0.2790732333671420
-0.0130467357414141
-0.0130467357414141
-0.0130467357414141
-0.0130467357414141
-0.0130467357414141
-0.0130467357414141
-0.0130467357414141
0.2529797618843137
0.2529797618843137
0.2529797618843137
0.2529797618843137
0.2529797618843137
0.2529797618843137
0.4992768834432453
0.4992768834432453
0.4992768834432453
0.4992768834432453
0.4992768834432453
0.7075791695648470
0.7075791695648470
0.7075791695648470
0.7075791695648470
0.8624425927014152
0.8624425927014152
0.8624425927014152
278
w
-0.9819993138423283
-0.9063983436792293
-0.7753767779777941
-0.5971358187051090
-0.3828710305138969
-0.1460444418051857
0.0984636734123185
0.3352902621210299
0.5495550503122417
0.7277960095849270
0.8588175752863625
0.9344185454494611
-0.9797260628154768
-0.8948302877495671
-0.7487118008130278
-0.5521720740198941
-0.3197830307576960
-0.0687787036492924
0.1822256234591111
0.4146146667213093
0.6111543935144430
0.7572728804509823
0.8421686555168919
-0.9777216658171440
-0.8847925078734500
-0.7262832284328062
-0.5162388888178013
-0.2733177752814612
-0.0191030551536918
0.2238180583826483
0.4338623979976532
0.5923716774382970
0.6853008353819912
-0.9761316915597071
-0.8770826148055193
-0.7101683632846072
-0.4934343893217102
-0.2503615582783774
-0.0072887272350445
0.2094452467278523
0.3763594982487646
0.4754085750029524
-0.9751219969249455
-0.8726136056433137
-0.7027508559775688
-0.4884300663956099
-0.2585901717200765
-0.0442693821381174
0.1255933675276275
0.2281017588092593
-0.9748859439808757
-0.8724516799591379
-0.7067984663384281
-0.5065233678707071
-0.3062482694029861
-0.1405950557822763
-0.0381607917605384
-0.9756577326126661
-0.8778783892097810
-0.7255500958233327
-0.5535231375438092
-0.4011948441573610
-0.3034155007544759
-0.9777350867885435
-0.8904719262323680
-0.7626851774630369
-0.6348984286937057
-0.5476352681375303
-0.9815084003062328
-0.9121094472397426
-0.8215631938079326
-0.7521642407414426
-0.9874378288038375
-0.9442680320921216
-0.9010982353804058
G
0.0000210869584592
0.0000478009336582
0.0000715535981885
0.0000908140442808
0.0001043691003742
0.0001113665276756
0.0001113665276756
0.0001043691003742
0.0000908140442808
0.0000715535981885
0.0000478009336582
0.0000210869584592
0.0000535464288036
0.0001207931272715
0.0001791886522862
0.0002243041982222
0.0002527861872664
0.0002625209249121
0.0002527861872664
0.0002243041982222
0.0001791886522862
0.0001207931272715
0.0000535464288036
0.0000872219740904
0.0001955179071091
0.0002866170651435
0.0003522649056890
0.0003866159673009
0.0003866159673009
0.0003522649056890
0.0002866170651435
0.0001955179071091
0.0000872219740904
0.0001168606521007
0.0002597455641951
0.0003747199645547
0.0004491092931053
0.0004748357650267
0.0004491092931053
0.0003747199645547
0.0002597455641951
0.0001168606521007
0.0001370868198001
0.0003011552859919
0.0004248312581563
0.0004911576150206
0.0004911576150206
0.0004248312581563
0.0003011552859919
0.0001370868198001
0.0001434416293063
0.0003098537094155
0.0004229859742259
0.0004630093211860
0.0004229859742259
0.0003098537094155
0.0001434416293063
0.0001334932455900
0.0002810995240760
0.0003645908936630
0.0003645908936630
0.0002810995240760
0.0001334932455900
0.0001078454820094
0.0002178644253955
0.0002589494916014
0.0002178644253955
0.0001078454820094
0.0000709773353307
0.0001330656335193
0.0001330656335193
0.0000709773353307
0.0000319815451597
0.0000511704722556
0.0000319815451597
Appendix A
v
0.9524192316071329
0.9524192316071329
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.9756670111138167
-0.8741673141936402
-0.8741673141936402
-0.8741673141936402
-0.8741673141936402
-0.8741673141936402
-0.8741673141936402
-0.8741673141936402
-0.8741673141936402
-0.8741673141936402
-0.8741673141936402
-0.7010392650617517
-0.7010392650617517
-0.7010392650617517
-0.7010392650617517
-0.7010392650617517
-0.7010392650617517
-0.7010392650617517
-0.7010392650617517
-0.7010392650617517
-0.4716232530966042
-0.4716232530966042
-0.4716232530966042
-0.4716232530966042
-0.4716232530966042
-0.4716232530966042
-0.4716232530966042
-0.4716232530966042
-0.2062983545928467
-0.2062983545928467
-0.2062983545928467
-0.2062983545928467
-0.2062983545928467
-0.2062983545928467
-0.2062983545928467
0.0713617212818314
0.0713617212818314
0.0713617212818314
0.0713617212818314
0.0713617212818314
0.0713617212818314
0.3366866197855889
0.3366866197855889
0.3366866197855889
0.3366866197855889
0.3366866197855889
0.5661026317507365
0.5661026317507365
0.5661026317507365
0.5661026317507365
0.7392306808826250
0.7392306808826250
0.7392306808826250
0.8407303778028012
0.8407303778028012
-0.9732640007883790
-0.9732640007883790
-0.9732640007883790
-0.9732640007883790
-0.9732640007883790
-0.9732640007883790
-0.9732640007883790
-0.9732640007883790
279
w
-0.9954591998725648
-0.9830535032173966
-0.9799624148419588
-0.8960563483015299
-0.7516412996354959
-0.5573928181140126
-0.3277129487437590
-0.0796348110985994
0.1684433265465601
0.3981231959168137
0.5923716774382970
0.7367867261043313
0.8206927926447598
-0.9773087169131647
-0.8826570336852367
-0.7212096425941288
-0.5072719427700896
-0.2598480685351741
-0.0009212505822009
0.2465026236527146
0.4604403234767538
0.6218877145678618
0.7165393977957897
-0.9750678336492901
-0.8716039428294952
-0.6972499916211889
-0.4708557539980772
-0.2169486841246318
0.0369583857488138
0.2633526233719254
0.4377065745802315
0.5411704654000264
-0.9734599912552455
-0.8641034005024621
-0.6828927604066217
-0.4542540062217496
-0.2090593739926616
0.0195793801922106
0.2007900202880510
0.3101466110408346
-0.9727380826839559
-0.8615432030530465
-0.6817226193358712
-0.4643191393590843
-0.2469156593822975
-0.0670950756651221
0.0440998039657872
-0.9732004711539487
-0.8655506662948876
-0.6978453976063979
-0.5084529569864489
-0.3407476882979591
-0.2330978834388980
-0.9752138061015480
-0.8780689577387805
-0.7358116265483021
-0.5935542953578239
-0.4964094469950564
-0.9792426048289611
-0.9013401238584765
-0.7996991412032752
-0.7217966602327908
-0.9858184467504782
-0.9370836570968202
-0.8883488674431620
-0.9948578343987214
-0.9808091767150953
-0.9784380046772969
-0.8884968963126230
-0.7350843335127386
-0.5317937718771106
-0.2966835580314484
-0.0506428728811481
0.1844673409645141
0.3877579026001421
G
0.0000049193491194
0.0000049193491194
0.0000595090322757
0.0001342439126211
0.0001991420068639
0.0002492813446167
0.0002809349141111
0.0002917536527217
0.0002809349141111
0.0002492813446167
0.0001991420068639
0.0001342439126211
0.0000595090322757
0.0001509522856561
0.0003383765991611
0.0004960390032748
0.0006096536248433
0.0006691039103832
0.0006691039103832
0.0006096536248433
0.0004960390032748
0.0003383765991611
0.0001509522856561
0.0002429029154634
0.0005398990480330
0.0007788812593161
0.0009335046030977
0.0009869788471822
0.0009335046030977
0.0007788812593161
0.0005398990480330
0.0002429029154634
0.0003173647520954
0.0006971937405828
0.0009835115230168
0.0011370612795458
0.0011370612795458
0.0009835115230168
0.0006971937405828
0.0003173647520954
0.0003571017784935
0.0007713891095648
0.0010530349132563
0.0011526741076089
0.0010530349132563
0.0007713891095648
0.0003571017784935
0.0003500364530424
0.0007370790928380
0.0009560041983049
0.0009560041983049
0.0007370790928380
0.0003500364530424
0.0002936190125703
0.0005931554689816
0.0007050132524139
0.0005931554689816
0.0002936190125703
0.0001984590954645
0.0003720636333079
0.0003720636333079
0.0001984590954645
0.0000910047905887
0.0001456076649420
0.0000910047905887
0.0000141311700691
0.0000141311700691
0.0001029679326262
0.0002308142517569
0.0003383593063674
0.0004158583826363
0.0004564107530060
0.0004564107530060
0.0004158583826363
0.0003383593063674
Appendix A
v
-0.9732640007883790
-0.9732640007883790
-0.8623145363704783
-0.8623145363704783
-0.8623145363704783
-0.8623145363704783
-0.8623145363704783
-0.8623145363704783
-0.8623145363704783
-0.8623145363704783
-0.8623145363704783
-0.6753461423498142
-0.6753461423498142
-0.6753461423498142
-0.6753461423498142
-0.6753461423498142
-0.6753461423498142
-0.6753461423498142
-0.6753461423498142
-0.4325723667595235
-0.4325723667595235
-0.4325723667595235
-0.4325723667595235
-0.4325723667595235
-0.4325723667595235
-0.4325723667595235
-0.1602952158504878
-0.1602952158504878
-0.1602952158504878
-0.1602952158504878
-0.1602952158504878
-0.1602952158504878
0.1119819350585480
0.1119819350585480
0.1119819350585480
0.1119819350585480
0.1119819350585480
0.3547557106488386
0.3547557106488386
0.3547557106488386
0.3547557106488386
0.5417241046695025
0.5417241046695025
0.5417241046695025
0.6526735690874035
0.6526735690874035
-0.9715398316016786
-0.9715398316016786
-0.9715398316016786
-0.9715398316016786
-0.9715398316016786
-0.9715398316016786
-0.9715398316016786
-0.9715398316016786
-0.9715398316016786
-0.8542713326263087
-0.8542713326263087
-0.8542713326263087
-0.8542713326263087
-0.8542713326263087
-0.8542713326263087
-0.8542713326263087
-0.8542713326263087
-0.6599501708552311
-0.6599501708552311
-0.6599501708552311
-0.6599501708552311
-0.6599501708552311
-0.6599501708552311
-0.6599501708552311
-0.4147694887738811
-0.4147694887738811
-0.4147694887738811
-0.4147694887738811
-0.4147694887738811
280
w
0.5411704654000264
0.6311115737647004
-0.9754559368810017
-0.8736026028747540
-0.7019627091203327
-0.4790926071068780
-0.2291379476652488
0.0208167117763804
0.2436868137898351
0.4153267075442564
0.5171800415505041
-0.9731012281596763
-0.8622663745548832
-0.6786061614081772
-0.4468767094014855
-0.1983675799496759
0.0333618720570159
0.2170220852037219
0.3278569388085149
-0.9717044589438657
-0.8562936738052630
-0.6696552708521097
-0.4440090324707260
-0.2183627940893423
-0.0317243911361890
0.0836863940024137
-0.9716471640002230
-0.8577579504953869
-0.6803324439972055
-0.4799627718532823
-0.3025372653551008
-0.1886480518502648
-0.9733819260153596
-0.8690573664827117
-0.7162861833797618
-0.5635150002768118
-0.4591904407441639
-0.9774586839357379
-0.8928611498387868
-0.7824849925110273
-0.6978874584140763
-0.9844826189504376
-0.9311572681852391
-0.8778319174200406
-0.9943500185651311
-0.9789139822232480
-0.9776335994527140
-0.8848171633795486
-0.7284059532635898
-0.5253099154364650
-0.2975323871345371
-0.0697548588326091
0.1333411789945156
0.2897523891104743
0.3825688251836399
-0.9744332847485946
-0.8690870942657640
-0.6945219356626813
-0.4742679794601965
-0.2380652937842477
-0.0178113375817628
0.1567538210213200
0.2621000115041505
-0.9721786808335768
-0.8587021340441509
-0.6751917156722266
-0.4533272175077608
-0.2314627193432952
-0.0479523009713708
0.0655242458180552
-0.9713615067108625
-0.8563248494717401
-0.6771117655597831
-0.4747233515370886
-0.2955102676251314
G
0.0002308142517569
0.0001029679326262
0.0002602650823529
0.0005784898461613
0.0008345539810105
0.0010002294617935
0.0010575259273953
0.0010002294617935
0.0008345539810105
0.0005784898461613
0.0002602650823529
0.0004109397705245
0.0009027613617285
0.0012734999615058
0.0014723238740403
0.0014723238740403
0.0012734999615058
0.0009027613617285
0.0004109397705245
0.0005171023547262
0.0011170124289186
0.0015248505217504
0.0016691333708590
0.0015248505217504
0.0011170124289186
0.0005171023547262
0.0005462568529719
0.0011502645000127
0.0014919127429584
0.0014919127429584
0.0011502645000127
0.0005462568529719
0.0004828189000526
0.0009753682794138
0.0011593040929919
0.0009753682794138
0.0004828189000526
0.0003384019453905
0.0006344232146466
0.0006344232146466
0.0003384019453905
0.0001588806910169
0.0002542091056270
0.0001588806910169
0.0000249845838053
0.0000249845838053
0.0001394156969197
0.0003098785451142
0.0004470439285509
0.0005357909952225
0.0005664828829345
0.0005357909952225
0.0004470439285509
0.0003098785451142
0.0001394156969197
0.0003496253244272
0.0007680644625167
0.0010834868492557
0.0012526451539752
0.0012526451539752
0.0010834868492557
0.0007680644625167
0.0003496253244272
0.0005356723378323
0.0011571261544599
0.0015796103737784
0.0017290746536926
0.0015796103737784
0.0011571261544599
0.0005356723378323
0.0006356624883655
0.0013385278196120
0.0017360934905505
0.0017360934905505
0.0013385278196120
Appendix A
v
-0.4147694887738811
-0.1518351170968717
-0.1518351170968717
-0.1518351170968717
-0.1518351170968717
-0.1518351170968717
0.0933455649844783
0.0933455649844783
0.0933455649844783
0.0933455649844783
0.2876667267555559
0.2876667267555559
0.2876667267555559
0.4049352257309258
0.4049352257309258
-0.9707656932166957
-0.9707656932166957
-0.9707656932166957
-0.9707656932166957
-0.9707656932166957
-0.9707656932166957
-0.9707656932166957
-0.9707656932166957
-0.8515259058540692
-0.8515259058540692
-0.8515259058540692
-0.8515259058540692
-0.8515259058540692
-0.8515259058540692
-0.8515259058540692
-0.6586953705179881
-0.6586953705179881
-0.6586953705179881
-0.6586953705179881
-0.6586953705179881
-0.6586953705179881
-0.4255628305091844
-0.4255628305091844
-0.4255628305091844
-0.4255628305091844
-0.4255628305091844
-0.1924302905003809
-0.1924302905003809
-0.1924302905003809
-0.1924302905003809
0.0004002448357003
0.0004002448357003
0.0004002448357003
0.1196400321983269
0.1196400321983269
-0.9712615353186331
-0.9712615353186331
-0.9712615353186331
-0.9712615353186331
-0.9712615353186331
-0.9712615353186331
-0.9712615353186331
-0.8558233075546404
-0.8558233075546404
-0.8558233075546404
-0.8558233075546404
-0.8558233075546404
-0.8558233075546404
-0.6759846257341786
-0.6759846257341786
-0.6759846257341786
-0.6759846257341786
-0.6759846257341786
-0.4728897132474524
-0.4728897132474524
-0.4728897132474524
-0.4728897132474524
-0.2930510314269907
-0.2930510314269907
-0.2930510314269907
281
w
-0.1804736103860092
-0.9725467916376855
-0.8649490792033028
-0.7073847443869405
-0.5498204095705783
-0.4422226971361956
-0.9763897132415725
-0.8877803333193609
-0.7721698375358703
-0.6835604576136586
-0.9835761364934850
-0.9271356663131544
-0.8706951961328238
-0.9939856587438157
-0.9775541728578627
-0.9777694668251506
-0.8861698241121969
-0.7343835460808280
-0.5428703684158955
-0.3374895993857777
-0.1459764217208451
0.0058098563105240
0.0974094990234775
-0.9745437715237478
-0.8707138673956205
-0.7028036719838149
-0.4997998775821498
-0.2967960831804849
-0.1288858877686792
-0.0250559836405518
-0.9727322126013357
-0.8632014813236818
-0.6925659586433119
-0.4998643318570689
-0.3292288091766991
-0.2196980778990452
-0.9730531081299070
-0.8674398084319827
-0.7127814152545922
-0.5581230220772018
-0.4525097223792776
-0.9763025901400513
-0.8873662373148126
-0.7713291332031754
-0.6823927803779368
-0.9832667223240769
-0.9257629529270346
-0.8682591835299924
-0.9938220640538042
-0.9769436291628915
-0.9790734993579121
-0.8937192781922685
-0.7556873299256109
-0.5888064018314990
-0.4219254737373872
-0.2838935254707294
-0.1985393043050859
-0.9761296963593420
-0.8802461626000544
-0.7308713094551189
-0.5621797219718718
-0.4128048688269363
-0.3169213350676487
-0.9752732158447804
-0.8783612128523528
-0.7364448566237263
-0.5945285003950997
-0.4976164974026720
-0.9775030150146072
-0.8930718554073052
-0.7829127703268735
-0.6984816107195715
-0.9837510466525349
-0.9279116537773202
-0.8720722609021055
G
0.0006356624883655
0.0006065521516366
0.0012253284377478
0.0014564019592622
0.0012253284377478
0.0006065521516366
0.0004475721372230
0.0008390913762495
0.0008390913762495
0.0004475721372230
0.0002171545531715
0.0003474472850744
0.0002171545531715
0.0000347487453190
0.0000347487453190
0.0001557099483103
0.0003420669768513
0.0004825442252185
0.0005578809615585
0.0005578809615585
0.0004825442252185
0.0003420669768513
0.0001557099483103
0.0003844232201780
0.0008304071930423
0.0011336013895582
0.0012408638627684
0.0011336013895582
0.0008304071930423
0.0003844232201780
0.0005605123902143
0.0011802826834540
0.0015308468406342
0.0015308468406342
0.0011802826834540
0.0005605123902143
0.0006035404580134
0.0012192443544051
0.0014491705340967
0.0012192443544051
0.0006035404580134
0.0004809788487273
0.0009017210200564
0.0009017210200564
0.0004809788487273
0.0002447905536076
0.0003916648857722
0.0002447905536076
0.0000401631516755
0.0000401631516755
0.0001434009189485
0.0003097657694262
0.0004228659260195
0.0004628779139008
0.0004228659260195
0.0003097657694262
0.0001434009189485
0.0003434508406038
0.0007232116307498
0.0009380178626390
0.0009380178626390
0.0007232116307498
0.0003434508406038
0.0004593234392237
0.0009279038425407
0.0011028887705954
0.0009279038425407
0.0004593234392237
0.0004145400028581
0.0007771639755478
0.0007771639755478
0.0004145400028581
0.0002271326760911
0.0003634122817458
0.0002271326760911
Appendix A
v
-0.1776128036629980
-0.1776128036629980
-0.9734205342926717
-0.9734205342926717
-0.9734205342926717
-0.9734205342926717
-0.9734205342926717
-0.9734205342926717
-0.8692472926775869
-0.8692472926775869
-0.8692472926775869
-0.8692472926775869
-0.8692472926775869
-0.7166976970646236
-0.7166976970646236
-0.7166976970646236
-0.7166976970646236
-0.5641481014516603
-0.5641481014516603
-0.5641481014516603
-0.4599748598365756
-0.4599748598365756
-0.9777408150931738
-0.9777408150931738
-0.9777408150931738
-0.9777408150931738
-0.9777408150931738
-0.8942021189160209
-0.8942021189160209
-0.8942021189160209
-0.8942021189160209
-0.7852074493830036
-0.7852074493830036
-0.7852074493830036
-0.7016687532058506
-0.7016687532058506
-0.9847924167051783
-0.9847924167051783
-0.9847924167051783
-0.9847924167051783
-0.9325316833444923
-0.9325316833444923
-0.9325316833444923
-0.8802709499838063
-0.8802709499838063
-0.9944858006509373
-0.9944858006509373
-0.9944858006509373
-0.9794207278662345
-0.9794207278662345
282
w
-0.9939268478192584
-0.9773346874993747
-0.9817659199711265
-0.9085222757201959
-0.7944176096234181
-0.6655572502131575
-0.5514525841163795
-0.4782089398654491
-0.9795541538651344
-0.8994204862856184
-0.7820740507258301
-0.6647276151660418
-0.5845939475865258
-0.9803297986402473
-0.9065075548332930
-0.8101901422313307
-0.7363678984243763
-0.9852639521319173
-0.9346236463387935
-0.8839833405456696
-0.9943830979868571
-0.9790374363058146
-0.9860052582322312
-0.9311554869250322
-0.8508343766029254
-0.7705132662808184
-0.7156634949736194
-0.9850865570896014
-0.9291164224480115
-0.8560910269349921
-0.8001208922934022
-0.9880764026082388
-0.9471010594580104
-0.9061257163077820
-0.9952960807455358
-0.9824447343476380
-0.9916869912527055
-0.9604882786778677
-0.9197826713059386
-0.8885839587311007
-0.9923962083525891
-0.9662658416722462
-0.9401354749919031
-0.9967862595070837
-0.9880061571980946
-0.9976806817582317
-0.9897103639331173
-0.9817400461080028
-0.9988347125647420
-0.9956510880861953
G
0.0000387007881208
0.0000387007881208
0.0001045106570038
0.0002200702800715
0.0002854349196915
0.0002854349196915
0.0002200702800715
0.0001045106570038
0.0002356485641582
0.0004760462660934
0.0005658194923739
0.0004760462660934
0.0002356485641582
0.0002672935102895
0.0005011118002183
0.0005011118002183
0.0002672935102895
0.0001657638007540
0.0002652220812064
0.0001657638007540
0.0000300243346837
0.0000300243346837
0.0000539668497656
0.0001090213191645
0.0001295806560438
0.0001090213191645
0.0000539668497656
0.0001069491819233
0.0002005042959233
0.0002005042959233
0.0001069491819233
0.0000841326676332
0.0001346122682131
0.0000841326676332
0.0000169950809241
0.0000169950809241
0.0000145815879016
0.0000273370114952
0.0000273370114952
0.0000145815879016
0.0000209968709439
0.0000335949935103
0.0000209968709439
0.0000053243602271
0.0000053243602271
0.0000006215536202
0.0000009944857924
0.0000006215536202
0.0000002997806231
0.0000002997806231
Appendix B
This section includes the complete computer codes for the derived formulae of
the thesis using FORTRAN for readily use and reference. Thus, it enhance the
use of the derive formulae of the thesis in the solution of practical problems in
real world.
283
C (PROGRAM-1) IOUT.FOR
C COMPLETE PROGRAM TO DETERMINE THE INTEGRAL OVER THE UNIT
C TRIANGULAR, STANDARD OR THE ORIGINAL TRIANGULAR SURFACES
C USING THE NEW TECHNIQUE. THAT IS THE TRIANGLE IN (X,Y) SPACE IS
C EITHER IN ORIGINAL SHAPE OR TRANSFORMED INTO STANDARD OR
C UNIT TRIANGLE WHICH IS AGAIN SUB-DIVIDED INTO THREE
QUADRILATERALS.
==============================================================
PARAMETER (NP=20)
CHARACTER*8,OUTFILE, OUTF1
INTEGER TYPE
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),INT(2:NP,2:NP),F1,F2,F3
1 ,X1,Y1,X2,Y2,X3,Y3,AREA,U(3),V(3),X,Y,A11,A12,A13,A14,
2 B11,B12,B13,B14,A21,A22,A23,A24,B21,B22,B23,B24,A31,A32,
3 A33,A34,B31,B32,B33,B34
C HERE THE INTEGRAND SHOULD BE GIVEN AS STATEMENT FUNCTION
F(U,V)=(U+V)**0.5D0
PRINT*,'TYPE THE FILE NAME FOR OUTPUT OF THE INTEGRAL'
READ*,OUTFILE
OPEN( UNIT =2, FILE= OUTFILE)
PRINT*,'TYPE THE FILE NAME FOR OUTPUT OF THE GAUSSPOINTS AND WTS'
READ*,OUTF1
OPEN( UNIT =3, FILE= OUTF1)
C
GAUSS OR SAMPLE POINTS AND CORRESPONDING WEIGHT POINTS
CALL GQPW(NP,S,W)
C
DO 2 JJ=1,3
WRITE(*,3) JJ
3 FORMAT('TPYE X AND Y COORDINATES OF NODE:-',I2)
READ*,U(JJ),V(JJ)
2
CONTINUE
AREA=(U(1)-U(3))*(V(2)-V(3))-(U(2)-U(3))*(V(1)-V(3))
C
A11=14.D0*U(1)+5.D0*U(2)+5.D0*U(3)
A12=4.D0*U(1)+U(2)-5.D0*U(3)
A13=-4.D0*U(1)+5.D0*U(2)-U(3)
A14=-2.D0*U(1)+U(2)+U(3)
B11=14.D0*V(1)+5.D0*V(2)+5.D0*V(3)
B12=4.D0*V(1)+V(2)-5.D0*V(3)
B13=-4.D0*V(1)+5.D0*V(2)-V(3)
B14=-2.D0*V(1)+V(2)+V(3)
A21=5.D0*U(1)+14.D0*U(2)+5.D0*U(3)
A22=U(1)+4.D0*U(2)-5.D0*U(3)
A23=-5.D0*U(1)+4.D0*U(2)+U(3)
284
A24=-U(1)+2.D0*U(2)-U(3)
B21=5.D0*V(1)+14.D0*V(2)+5.D0*V(3)
B22=V(1)+4.D0*V(2)-5.D0*V(3)
B23=-5.D0*V(1)+4.D0*V(2)+V(3)
B24=-V(1)+2.D0*V(2)-V(3)
A31=5.D0*U(1)+5.D0*U(2)+14.D0*U(3)
A32=5.D0*U(1)-U(2)-4.D0*U(3)
A33=-U(1)+5.D0*U(2)-4.D0*U(3)
A34=-U(1)-U(2)+2.D0*U(3)
B31=5.D0*V(1)+5.D0*V(2)+14.D0*V(3)
B32=5.D0*V(1)-V(2)-4.D0*V(3)
B33=-V(1)+5.D0*V(2)-4.D0*V(3)
B34=-V(1)-V(2)+2.D0*V(3)
C THIS PART FOR THE SUM IN THE GAUSS INTEGRATION FORMULA
DO 7 M=2,NP
N=M
INT(M,N)=0.D0
DO 9 I=1,M
DO 10 J=1,N
X1=(A11+A12*S(M,I)+A13*S(N,J)+A14*S(M,I)*S(N,J))/24.D0
Y1=(B11+B12*S(M,I)+B13*S(N,J)+B14*S(M,I)*S(N,J))/24.D0
F1=F(X1,Y1)*(4.D0+S(M,I)-S(N,J))
X2=(A21+A22*S(M,I)+A23*S(N,J)+A24*S(M,I)*S(N,J))/24.D0
Y2=(B21+B22*S(M,I)+B23*S(N,J)+B24*S(M,I)*S(N,J))/24.D0
F2=F(X2,Y2)*(4.D0+S(M,I)+S(N,J))
X3=(A31+A32*S(M,I)+A33*S(N,J)+A34*S(M,I)*S(N,J))/24.D0
Y3=(B31+B32*S(M,I)+B33*S(N,J)+B34*S(M,I)*S(N,J))/24.D0
F3=F(X3,Y3)*(4.D0-S(M,I)-S(N,J))
INT(M,N)=INT(M,N)+(F1+F2+F3)*W(M,I)*W(N,J)
10 CONTINUE
9
CONTINUE
WRITE(2,11) M,N,INT(M,N)*AREA/96.D0
11 FORMAT(2X,'I(',I2,',',I2,')=',D18.10)
7
CONTINUE
15 PRINT*, 'YOUR DATA FILE NAME ', OUTFILE
STOP
END
C================================****===========================
285
SUBROUTINE GQPW(NP,S,W)
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),Z1,Z,XM,XL,PP,P3,P2,P1,
EPS,X1,X2
EPS=0.0000000000000003D0
X1=-1.D0
X2=1.D0
C
DO 70 N=2,NP
M=(N+1)/2
XM=0.5D0*(X2+X1)
XL=0.5D0*(X2-X1)
DO 80 I=1,M
Z=DCOS(3.141592654D0*(I-0.25D0)/(N+0.5D0))
82
CALL NEWTON (N,P1,P2,P3,PP,Z1,Z)
IF(DABS(Z-Z1).GT.EPS) GO TO 82
S(N,I)= XM-XL*Z
S(N,N+1-I)=XM+XL*Z
W(N,I)=(2.D0*XL)/((1.D0-Z*Z)*PP*PP)
W(N,N+1-I)=W(N,I)
80
CONTINUE
70
CONTINUE
RETURN
END
C==============================================================
SUBROUTINE NEWTON (N,P1,P2,P3,PP,Z1,Z)
DOUBLE PRECISION Z1,Z,PP,P3,P2,P1
P1=1.D0
P2=0.D0
DO 90 J= 1,N
P3=P2
P2=P1
P1=((2.D0*J-1.D0)*Z*P2-(J-1.D0)*P3)/J
90 CONTINUE
PP=N*(Z*P1-P2)/(Z*Z-1.D0)
Z1=Z
Z=Z1-P1/PP
RETURN
END
C==============================================================
286
C
PROGRAM-2 GQUTS.FOR
C COMPLETE PROGRAM TO EVALUTE THE INTEGRAL OVER THE UNIT
C TRIANGULAR SURFACES USING THE NEW WEIGHTS AND GAUSSIAN
C POINTS. THAT IS THE TRIANGLE IN (X,Y) SPACE IS TRANSFORMED INTO
C A UNIT TRIANGLE WHICH IS AGAIN TRANSFORMED INTO A 2-SQUARE.
C NP= NUMBER OF GAUSSIAN POINTS, ITP= TOTAL NEW GAUSSIAN POINTS
FOR TRIANGLES.
===============================================================
PARAMETER(NP=11, ITP=NP*NP)
INTEGER R
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),XC(3),YC(3),X,Y,
1 G(ITP),U(ITP),V(ITP),U1(ITP),V1(ITP),G1(ITP),XR,YR,XR1,YR1,
2 INT1,INT2,AREA
C HERE THE INTEGRAND SHOULD BE GIVEN AS STATEMENT FUNCTION
F(U,V)=1.D0/(12.D0-8.821067231*U+21.53679831D0*V)
OPEN( UNIT =1, FILE= '11.DAT')
C
DO 2 JJ=1,3
WRITE(*,3) JJ
3 FORMAT('TPYE X AND Y COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ)
2 CONTINUE
AREA=DABS((XC(1)-XC(3))*(YC(2)-YC(3))-(XC(2)-XC(3))*(YC(1)-YC(3)))
C
CALL GQPW(NP,S,W)
C THIS PART FOR THE SUM IN THE GAUSS INTEGRATION FORMULA
DO 7 M=2,NP
C FOR MN POINTS N=2,NP; FOR MM POINTS N=M
C DO 7 N=2,NP
N=M
INT1=0.D0
INT2=0.D0
R=1
DO 9 J=1,N
DO 10 I=1,M
G(R)=(1.D0-S(M,I))*W(M,I)*W(N,J)/8.D0
U(R)=(1.D0+S(M,I))/2.D0
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))/4.D0
XR=XC(3)+(XC(1)-XC(3))*U(R)+(XC(2)-XC(3))*V(R)
YR=YC(3)+(YC(1)-YC(3))*U(R)+(YC(2)-YC(3))*V(R)
INT1=INT1+F(XR,YR)*G(R)
G1(R)=(1.D0-S(N,J))*W(M,I)*W(N,J)/8.D0
U1(R)=(1.D0+S(M,I))*(1.D0-S(N,J))/4.D0
V1(R)=(1.D0+S(N,J))/2.D0
287
10
9
XR1=XC(3)+(XC(1)-XC(3))*U1(R)+(XC(2)-XC(3))*V1(R)
YR1=YC(3)+(YC(1)-YC(3))*U1(R)+(YC(2)-YC(3))*V1(R)
INT2=INT2+F(XR1,YR1)*G1(R)
R=R+1
CONTINUE
CONTINUE
12
7
INT1=INT1*AREA
INT2=INT2*AREA
WRITE(1,12) M,N,INT1,INT2
FORMAT(2X,2I4,4D25.15)
CONTINUE
INT1=0.D0
INT2=0.D0
R=1
DO 19 I=1,M
N=M-I+2
DO 20 J=1,N
G(R)=(1.D0-S(M,I))*W(M,I)*W(N,J)/8.D0
U(R)=(1.D0+S(M,I))/2.D0
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))/4.D0
C
XR=XC(3)+(XC(1)-XC(3))*U(R)+(XC(2)-XC(3))*V(R)
YR=YC(3)+(YC(1)-YC(3))*U(R)+(YC(2)-YC(3))*V(R)
INT1=INT1+F(XR,YR)*G(R)
C
G1(R)=(1.D0-S(M,I))*W(N,J)*W(M,I)/8.D0
U1(R)=(1.D0+S(N,J))*(1.D0-S(M,I))/4.D0
V1(R)=(1.D0+S(M,I))/2.D0
C
XR1=XC(3)+(XC(1)-XC(3))*U1(R)+(XC(2)-XC(3))*V1(R)
YR1=YC(3)+(YC(1)-YC(3))*U1(R)+(YC(2)-YC(3))*V1(R)
INT2=INT2+F(XR1,YR1)*G1(R)
R=R+1
20
CONTINUE
19
CONTINUE
INT1=INT1*AREA
INT2=INT2*AREA
WRITE(1,11) M+1,-1+(M+2)*(M+1)/2,INT1,INT2
11 FORMAT(2X,2I5,4D25.15)
17 CONTINUE
PRINT*, 'COMPUTATION OVER'
STOP
END
C================================****===========================
C
PROGRAM-4 GQSTS.FOR
C COMPLETE PROGRAM TO EVALUTE THE INTEGRAL OVER THE
C STANDARD TRIANGULAR SURFACES USING THE NEW WEIGHTS AND
C GAUSSIAN POINTS. THAT IS THE TRIANGLE IN (X,Y) SPACE IS
C TRANSFORMED INTO A STANDARD TRIANGLE WHICH IS AGAIN
C TRANSFORMED INTO A 2-SQUARE. NP= NUMBER OF GAUSSIAN POINTS,
C ITP= TOTAL NEW GAUSSIAN POINTS FOR TRIANGLES.
==============================================================
PARAMETER (NP=15,ITP=NP*NP)
INTEGER R
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),AREA,
1 XC(3),YC(3),X,Y,G(ITP),U(ITP),V(ITP),U1(ITP),V1(ITP),
2 G1(ITP), XR,YR,XR1,YR1,INT,INT1
289
C
C
3
2
DO 2 JJ=1,3
WRITE(*,3) JJ
FORMAT(' TPYE X AND Y COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ)
CONTINUE
AREA=DABS((XC(1)-XC(3))*(YC(2)-YC(3))-(XC(2)-XC(3))*(YC(1)-YC(3)))
AREA=AREA/4.D0
CALL GQPW(NP,S,W)
C
C THIS PART FOR THE SUM IN THE GAUSS INTEGRATION FORMULA
DO 7 M=2,NP
C
DO 7 N=2,NP
N=M
INT=0.D0
INT1=0.D0
R=1
DO 9 I=1,M
DO 10 J=1,N
G(R)=(1.D0-S(M,I))*W(M,I)*W(N,J)/2.D0
U(R)=S(M,I)
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))/2.D0-1.D0
XR=(XC(1)+XC(2)+(XC(1)-XC(3))*U(R)+(XC(2)-XC(3))*V(R))/2.D0
YR=(YC(1)+YC(2)+(YC(1)-YC(3))*U(R)+(YC(2)-YC(3))*V(R))/2.D0
INT=INT+F(XR,YR)*G(R)
C
G1(R)=(1.D0-S(N,J))*W(M,I)*W(N,J)/2.D0
U1(R)=(1.d0-S(N,J))*(1.d0+S(M,I))/2.D0-1.D0
V1(R)=S(N,J)
XR1=(XC(1)+XC(2)+(XC(1)-XC(3))*U1(R)+(XC(2)-XC(3))*V1(R))/2.D0
YR1=(YC(1)+YC(2)+(YC(1)-YC(3))*U1(R)+(YC(2)-YC(3))*V1(R))/2.D0
INT1=INT1+F(XR1,YR1)*G1(R)
R=R+1
10 CONTINUE
9 CONTINUE
INT=INT*AREA
INT1= INT1*AREA
WRITE(1,11) M, N, INT, INT1
7
CONTINUE
STOP
END
C================================****===========================
290
C
PROGRAM-5 GQSTM.FOR
C
COMPLETE PROGRAM TO EVALUTE THE INTEGRAL OVER THE
STANDARD TRIANGULAR SURFACES USING THE NEW WEIGHTS AND
GAUSSIAN POINTS. THAT IS THE TRIANGLE IN (X,Y) SPACE IS
TRANSFORMED INTO A STANDARD TRIANGLE WHICH IS AGAIN
TRANSFORMED INTO A 2-SQUARE. NP= NUMBER OF GAUSSIAN POINTS, TP=
TOTAL NEW GAUSSIAN POINTS FOR TRIANGLES WHICH IS COMPLETELY
CROWDING FREE WITH LESS NUMBER OF POINTS.
===============================================================
PARAMETER (NP=15, TP=NP*(NP+1)/2)
INTEGER R
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),AREA,
1 XC(3),YC(3),X,Y,L(ITP),P(ITP),Q(ITP),P1(ITP),Q1(ITP),
2 L1(ITP),XR,YR,XR1,YR1,INT,INT1
C
C
HERE THE INTEGRAND SHOULD BE GIVEN AS STATEMENT FUNCTION
F(X,Y)=1.D0/(12.D0+21.53679831D0*Y-8.821067231*X)
OPEN( UNIT =1, FILE= 'P41.DAT')
C
DO 2 JJ=1,3
WRITE(*,3) JJ
3 FORMAT(' TPYE X AND Y COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ)
2 CONTINUE
AREA=DABS((XC(1)-XC(3))*(YC(2)-YC(3))-(XC(2)-XC(3))*(YC(1)-YC(3)))
AREA=AREA/4.D0
CALL GQPW(NP,S,W)
C
C THIS PART FOR THE SUM IN THE GAUSS INTEGRATION FORMULA
DO M=2,NP-1
INT=0.D0
INT1=0.D0
R=1
DO I=1,M
N=M-I+2
DO J=1,N
L(R)=(1.d0-S(M,I))*W(M,I)*W(N,J)/2.D0
P(R)=S(M,I)
Q(R)=(1.D0-S(M,I))*(1.D0+S(N,J))/2.D0-1.D0
XR=(XC(1)+XC(2)+(XC(1)-XC(3))*p(R)+(XC(2)-XC(3))*q(R))/2.D0
YR=(YC(1)+YC(2)+(YC(1)-YC(3))*p(R)+(YC(2)-YC(3))*q(R))/2.D0
INT=INT+F(XR,YR)*L(R)
C
L1(R)=(1.D0-S(M,I))*W(N,J)*W(M,I)/2.D0
291
P1(R)=(1.D0-S(M,I))*(1.D0+S(N,J))/2.D0-1.D0
Q1(R)=S(M,I)
XR1=(XC(1)+XC(2)+(XC(1)-XC(3))*P1(R)+(XC(2)-XC(3))*Q1(R))/2.D0
YR1=(YC(1)+YC(2)+(YC(1)-YC(3))*P1(R)+(YC(2)-YC(3))*Q1(R))/2.D0
INT1=INT1+F(XR1,YR1)*L1(R)
R=R+1
ENDDO
ENDDO
INT=INT*AREA
INT1=INT1*AREA
WRITE(1,11) M+1,-1+(M+2)*(M+1)/2,INT,INT1
11 FORMAT(2X,2I5,5D25.15)
ENDDO
STOP
END
C================================****===========================
C
PROGRAM-6 SGQTS.FOR
C COMPLETE PROGRAM TO EVALUTE THE INTEGRAL OVER THE
TRIANGULAR SURFACES USING THE NEW WEIGHTS AND GAUSSIAN
POINTS. THAT IS THE TRIANGLE IN (X,Y) SPACE IS TRANSFORMED INTO A
SYMMETRIC TRIANGLE WHICH IS AGAIN TRANSFORMED INTO A 2SQUARE. NP= NUMBER OF GAUSSIAN POINTS, ITP= TOTAL NEW GAUSSIAN
POINTS FOR TRIANGLES
===============================================================
PARAMETER(NP=17, ITP=NP*NP)
INTEGER R,K
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),XC(3),YC(3),X,Y,AREA,
1 G(ITP),U(ITP),V(ITP),XR,YR,XR1,YR1,INT1,INT2,L(TP),P(TP),Q(TP)
C HERE THE INTEGRAND SHOULD BE GIVEN AS STATEMENT FUNCTION
F(U,V)=U**6/(1.D0-1.D0*V)
OPEN( UNIT =1, FILE= 'NFA11.DAT')
OPEN( UNIT =2, FILE= 'NFA22.DAT')
C
DO 2 JJ=1,3
WRITE(*,3) JJ
3 FORMAT('TPYE X AND Y COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ)
2 CONTINUE
AREA=DABS((XC(1)-XC(3))*(YC(2)-YC(3))-(XC(2)-XC(3))*(YC(1)-YC(3)))
C
CALL GQPW(NP,S,W)
DO 7 M=2,NP
N=M
INT1=0.D0
INT2=0.D0
R=1
292
DO 9 I=1,M
DO 10 J=1,N
G(R)=(1.D0-S(M,I))*W(M,I)*W(N,J)/8.D0
U(R)=S(M,I)
V(R)=(1.D0-S(M,I))*S(N,J)/2.D0
XR= ((2.D0*XC(1)+XC(2)+XC(3))+((2.D0*XC(1)-XC(2)-XC(3))*U(R))+
1
((2.D0*XC(2)-2.D0*XC(3))*V(R)))/4.D0
YR= ((2.D0*YC(1)+YC(2)+YC(3))+((2.D0*YC(1)-YC(2)-YC(3))*U(R))+
1
((2.D0*YC(2)-2.D0*YC(3))*V(R)))/4.D0
INT1=INT1+F(XR,YR)*G(R)
L(R)=(1.D0-S(N,J))*W(M,I)*W(N,J)/8.D0
P(R)=(1.D0-S(N,J))*S(M,I)/2.D0
Q(R)=S(N,J)
XR1= ((XC(1)+2.D0*XC(2)+XC(3))+2.D0*(XC(1)-XC(3))*P(R)+
1
(-XC(1)+2.D0*XC(2)-XC(3))*Q(R))/4.D0
YR1= ((YC(1)+2.D0*YC(2)+YC(3))+2.D0*(YC(1)-YC(3))*P(R)+
1
(-YC(1)+2.d0*YC(2)-YC(3))*Q(R))/4.d0
INT2=INT2+F(XR1,YR1)*L(R)
R=R+1
10 CONTINUE
9 CONTINUE
INT1=INT1*AREA
INT2=INT2*AREA
WRITE(1,12) M, INT1, INT2
12 FORMAT(2X,I4,2F25.12)
7
CONTINUE
PRINT*, 'COMPUTATION OVER'
STOP
END
C=============================****==============================
C
PROGRAM-7 SGQTM.FOR
C COMPLETE PROGRAM TO EVALUTE THE INTEGRAL OVER THE
TRIANGULAR SURFACES USING THE NEW WEIGHTS AND GAUSSIAN
POINTS. THAT IS THE TRIANGLE IN (X,Y) SPACE IS TRANSFORMED INTO A
SYMMETRIC TRIANGLE WHICH IS AGAIN TRANSFORMED INTO A 2SQUARE. NP= NUMBER OF GAUSSIAN POINTS, TP= TOTAL NEW GAUSSIAN
POINTS FOR TRIANGLES. CROWDING FREE FORMULA.
===============================================================
PARAMETER(NP=17, TP=NP*(NP+1)/2)
INTEGER R,K
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),XC(3),YC(3),X,Y,AREA,
1 G(ITP),U(ITP),V(ITP),XR,YR,XR1,YR1,INT1,INT2,L(TP),P(TP),Q(TP)
C HERE THE INTEGRAND SHOULD BE GIVEN AS STATEMENT FUNCTION
F(U,V)=U**6/(1.D0-1.D0*V)
OPEN( UNIT =1, FILE= 'NFA11.DAT')
OPEN( UNIT =2, FILE= 'NFA22.DAT')
293
3
2
DO 2 JJ=1,3
WRITE(*,3) JJ
FORMAT('TPYE X AND Y COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ)
CONTINUE
AREA=DABS((XC(1)-XC(3))*(YC(2)-YC(3))-(XC(2)-XC(3))*(YC(1)-YC(3)))
C
CALL GQPW(NP,S,W)
DO 17 M=2, NP-2
INT1=0.D0
INT2=0.D0
R=1
K=1
DO 19 I=1,M
N=M-I+2
DO 20 J=1,N
G(R)=(1.D0-S(M,I))*W(M,I)*W(N,J)/8.D0
U(R)=S(M,I)
V(R)=(1.D0-S(M,I))*S(N,J)/2.D0
XR= ((2.D0*XC(1)+XC(2)+XC(3))+((2.D0*XC(1)-XC(2)-XC(3))*U(R))+
1
((2.D0*XC(2)-2.D0*XC(3))*V(R)))/4.D0
YR= ((2.D0*YC(1)+YC(2)+YC(3))+((2.D0*YC(1)-YC(2)-YC(3))*U(R))+
1
((2.D0*YC(2)-2.D0*YC(3))*V(R)))/4.D0
INT1=INT1+F(XR,YR)*G(R)
R=R+1
20 CONTINUE
19 CONTINUE
C
N=M
DO 21 J=1,N
NN=N-J+2
DO 22 I=1,NN
L(K)=(1.D0-S(N,J))*W(NN,I)*W(N,J)/8.D0
P(K)=(1.D0-S(N,J))*S(NN,I)/2.D0
Q(K)=S(N,J)
XR1= ((XC(1)+2.D0*XC(2)+XC(3)) +2.D0*(XC(1)-XC(3))*P(K)+
1
(-XC(1)+2.D0*XC(2)-XC(3))*Q(K))/4.D0
YR1= ((YC(1)+2.D0*YC(2)+YC(3)) +2.D0*(YC(1)-YC(3))*P(K)+
1
(-YC(1)+2.D0*YC(2)-YC(3))*Q(K))/4.D0
INT2=INT2+F(XR1,YR1)*L(K)
K=K+1
22 CONTINUE
21 CONTINUE
INT1=INT1*AREA
294
INT2=INT2*AREA
WRITE(1,12) -1+(m+1)*(m+2)/2,INT1,INT2
17 CONTINUE
PRINT*, 'COMPUTATION OVER'
STOP
END
C=============================****==============================
C
PROGRAM 8 GQTTS.FOR
C ARBITRATY TETRAHEDRON IS TRANSFORMED TO UNIT TETRAHEDRON
C THEN TO 2-CUBE M*N*P POINTS
PARAMETER(NP=10, ITP=NP*NP*NP)
INTEGER R,M,N,P
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),X,Y,G(ITP),U(ITP),V(ITP),
1T(ITP),XR,YR,ZR,INT1,XC(4),YC(4),ZC(4)
C
OPEN( UNIT =11, FILE= 'N_1.DAT')
OPEN( UNIT =12, FILE= 'N_2.DAT')
C
DO 2 JJ=1,4
WRITE(*,3) JJ
3 FORMAT('TPYE X,Y AND Z COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ),ZC(JJ)
2 CONTINUE
VOL=DABS((XC(1)-XC(4))*((YC(2)-YC(4))*(ZC(3)-ZC(4))-(YC(3)-YC(4))*
1(ZC(2)-ZC(4)))-(XC(2)-XC(4))*((YC(1)-YC(4))*(ZC(3)-ZC(4))2(YC(3)-YC(4))*(ZC(1)-ZC(4)))+(XC(3)-XC(4))*((YC(1)-YC(4))*
3(ZC(2)-ZC(4))-(YC(2)-YC(4))*(ZC(1)-ZC(4))))
CALL GQPW(NP,S,W)
C
C THIS PART FOR THE SUM IN THE GAUSS INTEGRATION FORMULA
C CASE-1, XYZ
DO 14 M=2,NP
N=M
P=M
INT1=0.D0
C
R=1
DO 11 I=1,M
DO 12 J=1,N
DO 13 K=1,P
G(R)=((1.D0-S(M,I))**2)*(1.D0-S(N,J))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))/2.D0
295
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))/4.D0
T(R)=(1.D0-S(M,I))*(1.D0-S(N,J))*(1.D0+S(P,K))/8.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
13 CONTINUE
12 CONTINUE
11 CONTINUE
WRITE(11,*) M*M*M,INT1
14 CONTINUE
C CASE-2, XZY
DO 24 M=2,NP
N=M
P=M
INT1=0.D0
c
R=1
DO 21 I=1,M
DO 23 K=1,P
DO 22 J=1,N
G(R)=((1.D0-S(M,I))**2)*(1.D0-S(P,K))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))/2.D0
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))*(1.D0-S(P,K))/8.D0
T(R)=(1.D0-S(M,I))*(1.D0+S(P,K))/4.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
22 CONTINUE
23 CONTINUE
21 CONTINUE
WRITE(11,*) M*M*M,INT1
24 CONTINUE
C CASE-3, YXZ
DO 34 M=2,NP
N=M
296
P=M
INT1=0.D0
C
R=1
DO 32 J=1,N
DO 31 I=1,M
DO 33 K=1,P
G(R)=((1.D0-S(N,J))**2)*(1.D0-S(M,I))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))*(1.D0-S(N,J))/4.D0
V(R)=(1.D0+S(N,J))/2.D0
T(R)=(1.D0-S(M,I))*(1.D0-S(N,J))*(1.D0+S(P,K))/8.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
33 CONTINUE
31 CONTINUE
32 CONTINUE
WRITE(11,*) M*M*M,INT1
34 CONTINUE
c CASE-4, YZX
DO 44 M=2,NP
N=M
P=M
INT1=0.D0
c
R=1
DO 42 J=1,N
DO 43 K=1,P
DO 41 I=1,M
G(R)=((1.D0-S(N,J))**2)*(1.D0-S(P,K))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))*(1.D0-S(N,J))*(1.D0-S(P,K))/8.D0
V(R)=(1.D0+S(N,J))/2.D0
T(R)=(1.D0-S(N,J))*(1.D0+S(P,K))/4.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
297
41
43
42
CONTINUE
CONTINUE
CONTINUE
WRITE(11,*) M*M*M,INT1
44 CONTINUE
c CASE-5, ZYX
DO 54 M=2,NP
P=M
N=M
INT1=0.D0
c
R=1
DO 53 K=1,P
DO 52 J=1,N
DO 51 I=1,M
G(R)=((1.D0-S(P,K))**2)*(1.D0-S(N,J))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))*(1.D0-S(N,J))*(1.D0-S(P,K))/8.D0
V(R)=(1.D0+S(N,J))*(1.D0-S(P,K))/4.D0
T(R)=(1.D0+S(P,K))/2.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
51 CONTINUE
52 CONTINUE
53 CONTINUE
WRITE(11,*) M*M*M,INT1
54 CONTINUE
C CASE-6, ZXY
DO 64 M=2,NP
P=M
N=M
INT1=0.D0
C
R=1
DO 63 K=1,P
DO 61 I=1,M
DO 62 J=1,N
G(R)=((1.D0-S(P,K))**2)*(1.D0-S(M,I))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))*(1.D0-S(P,K))/4.D0
298
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))*(1.D0-S(P,K))/8.D0
T(R)=(1.D0+S(P,K))/2.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
62 CONTINUE
61 CONTINUE
63 CONTINUE
WRITE(11,*) M*M*M,INT1
64 CONTINUE
STOP
END
C============================****===============================
FUNCTION F1(X,Y,Z)
DOUBLE PRECISION X,Y,Z,F1
F1=(X**2.D0)*Y
C F1=DSQRT(X+Y+Z)
RETURN
END
C============================****===============================
C
PROGRAM-9 GQTTM.FOR
C COMPLETE PROGRAM TO EVALUATE INTEGRALS OVER THE ARBITRARY
C TETRAHEDRON WHERE THE ARBITRARY TETRAHEDRON IS FIRST
C TRANSFORMED TO UNIT TETRAHEDRON THEN THE UNIT TETRAHEDRON
IS AGAIN TRANSFORMED TO A 2-CUBE. TOTAL NUMBER OF POINTS
ITP = M*(M+4)*(M+5)/6 (GQTTM)
===============================================================
PARAMETER(NP=10, ITP=NP*NP*NP)
INTEGER R,M,N,P
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),X,Y,G(ITP),U(ITP),V(ITP),
1 T(ITP),XR,YR,ZR,INT1,INT2,INT3,INT4
C
OPEN( UNIT =1, FILE= '1.DAT')
OPEN( UNIT =2, FILE= '2.DAT')
OPEN( UNIT =11, FILE= 'N_1.DAT')
DO 2 JJ=1,4
WRITE(*,3) JJ
FORMAT('TPYE X,Y AND Z COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ),ZC(JJ)
299
CONTINUE
VOL=DABS((XC(1)-XC(4))*((YC(2)-YC(4))*(ZC(3)-ZC(4))-(YC(3)-XC(4))*
1 (ZC(2)-ZC(4)))-(XC(2)-XC(4))*((YC(1)-YC(4))*(ZC(3)-ZC(4))2 (YC(3)-XC(4))*(ZC(1)-ZC(4)))+(XC(3)-XC(4))*((YC(1)-YC(4))*
3 (ZC(2)-ZC(4))-(YC(2)-XC(4))*(ZC(1)-ZC(4))))
PRINT*,VOL
CALL GQPW(NP,S,W)
C
C THIS PART FOR THE SUM IN THE GAUSS INTEGRATION FORMULA
c CASE-1, GQTTM
DO 14 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
M=L
DO 11 I=1,M
N=M-I+2
DO 12 J=1,N
P=N-J+2
DO 13 K=1,P
G(R)=((1.D0-S(M,I))**2)*(1.D0-S(N,J))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))/2.D0
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))/4.D0
T(R)=(1.D0-S(M,I))*(1.D0-S(N,J))*(1.D0+S(P,K))/8.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
ENDIF
13
12
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(I+3)*(I+2)/2
300
11
9
14
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
FORMAT(2X,I5,3D25.15)
CONTINUE
c CASE-2, GQTTM
DO 24 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
M=L
DO 21 I=1,M
P=M-I+2
DO 23 K=1,P
N=P-K+2
DO 22 J=1,N
G(R)=((1.D0-S(M,I))**2)*(1.D0-S(P,K))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))/2.D0
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))*(1.D0-S(P,K))/8.D0
T(R)=(1.D0-S(M,I))*(1.D0+S(P,K))/4.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
22
23
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(I+3)*(I+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
CONTINUE
21
24
c CASE-3, GQTTM
DO 31 L=2,NP-2
TT=0
301
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
N=L
DO 32 J=1,N
M=N-J+2
DO 33 I=1,M
P=M-I+2
DO 34 K=1,P
G(R)=((1.D0-S(N,J))**2)*(1.D0-S(M,I))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))*(1.D0-S(N,J))/4.D0
V(R)=(1.D0+S(N,J))/2.D0
T(R)=(1.D0-S(M,I))*(1.D0-S(N,J))*(1.D0+S(P,K))/8.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
34
33
32
31
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(J+3)*(J+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
CONTINUE
c CASE-4, GQTTM
DO 44 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
N=L
DO 42 J=1,N
P=N-J+2
302
DO 43 K=1,P
M=P-K+2
DO 41 I=1,M
G(R)=((1.D0-S(N,J))**2)*(1.D0-S(P,K))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))*(1.D0-S(N,J))*(1.D0-S(P,K))/8.D0
V(R)=(1.D0+S(N,J))/2.D0
T(R)=(1.D0-S(N,J))*(1.D0+S(P,K))/4.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
41
43
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(J+3)*(J+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
CONTINUE
42
44
c CASE-5, GQTTM
DO 54 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
P=L
DO 53 K=1,P
N=P-K+2
DO 52 J=1,N
M=N-J+2
DO 51 I=1,M
G(R)=((1.D0-S(P,K))**2)*(1.D0-S(N,J))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))*(1.D0-S(N,J))*(1.D0-S(P,K))/8.D0
V(R)=(1.D0+S(N,J))*(1.D0-S(P,K))/4.D0
T(R)=(1.D0+S(P,K))/2.D0
303
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
51
52
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(K+3)*(K+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
CONTINUE
53
54
c CASE-6 GQTTM
DO 64 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
P=L
DO 63 K=1,P
M=P-K+2
DO 61 I=1,M
N=M-I+2
DO 62 J=1,N
G(R)=((1.D0-S(P,K))**2)*(1.D0-S(M,I))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=(1.D0+S(M,I))*(1.D0-S(P,K))/4.D0
V(R)=(1.D0-S(M,I))*(1.D0+S(N,J))*(1.D0-S(P,K))/8.D0
T(R)=(1.D0+S(P,K))/2.D0
XR=XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+(XC(3)-XC(4))*T(R)
YR=YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+(YC(3)-YC(4))*T(R)
ZR=ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+(ZC(3)-ZC(4))*T(R)
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
304
62
61
63
64
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(K+3)*(K+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
CONTINUE
STOP
END
C=============================****==============================
PROGRAM-10 GQTSS.FOR
C
PARAMETER(NP=15, ITP=NP*NP*NP)
INTEGER R,M,N,P
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),X,Y,G(ITP),U(ITP),V(ITP),
1T(ITP),XR,YR,ZR,INT1,VOL,XC(4),YC(4),ZC(4)
c
OPEN( UNIT =11, FILE= 'tT3d.dat')
OPEN( UNIT =12, FILE= 'sq3d.dat')
C
DO 2 JJ=1,4
WRITE(*,3) JJ
3 FORMAT('TPYE X,Y AND Z COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ),ZC(JJ)
2 CONTINUE
VOL=(XC(1)-XC(4))*((YC(2)-YC(4))*(ZC(3)-ZC(4))-(YC(3)-yC(4))*
1 (ZC(2)-ZC(4)))-(XC(2)-XC(4))*((YC(1)-YC(4))*(ZC(3)-ZC(4))2(YC(3)-yC(4))*(ZC(1)-ZC(4)))+(XC(3)-XC(4))*((YC(1)-YC(4))*
3(ZC(2)-ZC(4))-(YC(2)-yC(4))*(ZC(1)-ZC(4)))
CALL GQPW(NP,S,W)
C
C THIS PART FOR THE SUM IN THE GAUSS INTEGRATION FORMULA
c CASE-1, xyz
DO 14 M=2,NP
N=M
P=M
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
DO 11 I=1,M
DO 12 J=1,N
305
DO 13 K=1,P
G(R)=((1.d0-S(M,I))**2)*(1.d0-S(N,J))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=S(M,I)
V(R)=((1.d0-S(M,i))*(1.d0+S(N,J))/2.D0)-1.D0
T(R)=((1.d0-S(M,i))*(1.d0-S(N,J))*(1.d0+S(P,K))/4.D0)-1.D0
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.d0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.d0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.d0
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
13 CONTINUE
12 CONTINUE
11 CONTINUE
WRITE(11,*) M*M*M,INT1
14 CONTINUE
c CASE-2, XZY
DO 24 M=2,NP
N=M
P=M
INT1=0.D0
c
R=1
DO 21 I=1,M
DO 23 K=1,P
DO 22 J=1,N
G(R)=((1.d0-S(M,I))**2)*(1.d0-S(P,K))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=S(M,I)
V(R)=((1.d0-S(M,i))*(1.d0+S(N,J))*(1.d0-S(P,K))/4.D0)-1.D0
T(R)=((1.d0-S(M,i))*(1.d0+S(P,K))/2.D0)-1.D0
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.d0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.d0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.d0
306
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
22 CONTINUE
23 CONTINUE
21 CONTINUE
WRITE(11,*) M*M*M,int1
24 CONTINUE
c CASE-3, YXZ
DO 34 M=2,NP
N=M
P=M
INT1=0.D0
c
R=1
DO 32 J=1,N
DO 31 I=1,M
DO 33 K=1,P
G(R)=((1.d0-S(N,J))**2)*(1.d0-S(M,I))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=((1.d0+S(M,i))*(1.d0-S(N,J))/2.D0)-1.D0
V(R)=S(N,J)
T(R)=((1.d0-S(M,i))*(1.d0-S(N,J))*(1.d0+S(P,K))/4.D0)-1.D0
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.d0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.d0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.d0
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
33 CONTINUE
31 CONTINUE
32 CONTINUE
WRITE(11,*) M*M*M,int1
34 CONTINUE
c CASE-4, YZX
DO 44 M=2,NP
N=M
P=M
INT1=0.D0
c
R=1
307
DO 42 J=1,N
DO 43 K=1,P
DO 41 I=1,M
T(R)=(1.d0-S(N,J))*(1.d0+S(P,K))/4.D0
G(R)=((1.d0-S(N,J))**2)*(1.d0-S(P,K))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=((1.d0+S(M,i))*(1.d0-S(N,J))*(1.d0-S(P,K))/4.D0)-1.D0
V(R)=S(N,J)
T(R)=((1.d0-S(N,J))*(1.d0+S(P,K))/2.D0)-1.D0
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.d0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.d0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.d0
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
41 CONTINUE
43 CONTINUE
42 CONTINUE
WRITE(11,*) M*M*M,int1
44 CONTINUE
c CASE-5, ZYX
DO 54 M=2,NP
P=M
N=M
INT1=0.D0
c
R=1
DO 53 K=1,P
DO 52 J=1,N
DO 51 I=1,M
G(R)=((1.d0-S(P,K))**2)*(1.d0-S(N,J))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=((1.d0+S(M,i))*(1.d0-S(N,J))*(1.d0-S(P,K))/4.D0)-1.D0
V(R)=((1.d0+S(N,J))*(1.d0-S(P,K))/2.D0)-1.D0
T(R)=S(P,K)
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.d0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.d0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
308
(ZC(3)-ZC(4))*T(R))/2.d0
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
51 CONTINUE
52 CONTINUE
53 CONTINUE
WRITE(11,*) M*M*M,int1
54 CONTINUE
c CASE-6, ZXY
DO 64 M=2,NP
P=M
N=M
INT1=0.D0
c
R=1
DO 63 K=1,P
DO 61 I=1,M
DO 62 J=1,N
G(R)=((1.D0-S(P,K))**2)*(1.D0-S(M,I))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=((1.D0+S(M,I))*(1.D0-S(P,K))/2.D0)-1.D0
V(R)=((1.D0-S(M,I))*(1.D0+S(N,J))*(1.D0-S(P,K))/4.D0)-1.D0
T(R)=S(P,K)
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.d0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.d0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.d0
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
R=R+1
62 CONTINUE
61 CONTINUE
63 CONTINUE
WRITE(11,*) M*M*M,int1
64 CONTINUE
STOP
END
309
C================================****===========================
PROGRAM -11 GQTSM.FOR
C COMPLETE PROGRAM TO EVALUATE INTEGRALS OVER THE ARBITRARY
C TETRAHEDRON WHERE THE ARBITRARY TETRAHEDRON IS FIRST
C TRANSFORMED TO STANDARD TETRAHEDRON THEN THE STANDARD
C TETRAHEDRON IS AGAIN TRANSFORMED TO A 2-CUBE. TOTAL NUMBER
OF POINTS ITP = M*(M+4)*(M+5)/6(GQTSM)
PARAMETER(NP=17, ITP=NP*NP*NP)
INTEGER R,M,N,P,TT
DOUBLE PRECISION S(2:NP,NP),W(2:NP,NP),X,Y,G(ITP),U(ITP),V(ITP),
1T(ITP),XR,YR,ZR,INT1,INT2,INT3,INT4,VOL,XC(4),YC(4),ZC(4)
C
OPEN( UNIT =1, FILE= 'TT3D.DAT')
OPEN( UNIT =2, FILE= 'SQ3D.DAT')
OPEN( UNIT =11, FILE= 'TT1.DAT')
C
DO 2 JJ=1,4
WRITE(*,3) JJ
3 FORMAT('TPYE X,Y AND Z COORDINATES OF NODE:-',I2)
READ*,XC(JJ),YC(JJ),ZC(JJ)
2
CONTINUE
VOL=DABS((XC(1)-XC(4))*((YC(2)-YC(4))*(ZC(3)-ZC(4))-(YC(3)-YC(4))
1 *(ZC(2)-ZC(4)))-(XC(2)-XC(4))*((YC(1)-YC(4))*(ZC(3)-ZC(4))2 (YC(3)-YC(4))*(ZC(1)-ZC(4)))+(XC(3)-XC(4))*((YC(1)-YC(4))*
3 (ZC(2)-ZC(4))-(YC(2)-YC(4))*(ZC(1)-ZC(4))))
PRINT*,VOL
CALL GQPW(NP,S,W)
C
C THIS PART FOR THE SUM IN THE GAUSS INTEGRATION FORMULA
c CASE-1, xyz
DO 14 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
M=L
DO 11 I=1,M
N=M-I+2
DO 12 J=1,N
P=N-J+2
DO 13 K=1,P
310
G(R)=((1.D0-S(M,I))**2)*(1.D0-S(N,J))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=S(M,I)
V(R)=((1.D0-S(M,I))*(1.D0+S(N,J))/2.D0)-1.D0
T(R)=((1.D0-S(M,I))*(1.D0-S(N,J))*(1.D0+S(P,K))/4.D0)-1.D0
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.D0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.D0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.D0
C
13
12
11
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(I+3)*(I+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
FORMAT(I5,4F25.16)
CONTINUE
9
14
C
c CASE-2, XZY
DO 24 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
M=L
DO 21 I=1,M
P=M-I+2
DO 23 K=1,P
N=P-K+2
DO 22 J=1,N
C
G(R)=((1.D0-S(M,I))**2)*(1.D0-S(P,K))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=S(M,I)
V(R)=((1.D0-S(M,I))*(1.D0+S(N,J))*(1.D0-S(P,K))/4.D0)-1.D0
T(R)=((1.D0-S(M,I))*(1.D0+S(P,K))/2.D0)-1.D0
311
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.D0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.D0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.D0
22
23
21
24
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(I+3)*(I+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
CONTINUE
c CASE-3, YXZ
DO 34 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
N=L
DO 32 J=1,N
M=N-J+2
DO 31 I=1,M
P=M-I+2
DO 33 K=1,P
C
G(R)=((1.D0-S(N,J))**2)*(1.D0-S(M,I))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=((1.D0+S(M,I))*(1.D0-S(N,J))/2.D0)-1.D0
V(R)=S(N,J)
T(R)=((1.D0-S(M,I))*(1.D0-S(N,J))*(1.D0+S(P,K))/4.D0)-1.D0
C
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.D0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.D0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
312
33
31
32
(ZC(3)-ZC(4))*T(R))/2.D0
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(J+3)*(J+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
CONTINUE
34
C
c CASE-4, YZX
DO 44 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
N=L
DO 42 J=1,N
P=N-J+2
DO 43 K=1,P
M=P-K+2
DO 41 I=1,M
G(R)=((1.D0-S(N,J))**2)*(1.D0-S(P,K))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=((1.D0+S(M,I))*(1.D0-S(N,J))*(1.D0-S(P,K))/4.D0)-1.D0
V(R)=S(N,J)
T(R)=((1.D0-S(N,J))*(1.D0+S(P,K))/2.D0)-1.D0
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.d0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.d0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.d0
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
313
41
43
42
CONTINUE
CONTINUE
TT=TT-1+(J+3)*(J+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
CONTINUE
44
c CASE-5, ZYX
DO 54 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
P=L
DO 53 K=1,P
N=P-K+2
DO 52 J=1,N
M=N-J+2
DO 51 I=1,M
G(R)=((1.D0-S(P,K))**2)*(1.D0-S(N,J))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=((1.D0+S(M,I))*(1.D0-S(N,J))*(1.D0-S(P,K))/4.D0)-1.D0
V(R)=((1.D0+S(N,J))*(1.D0-S(P,K))/2.D0)-1.D0
T(R)=S(P,K)
C
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.d0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.d0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.d0
C
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
C
51
52
53
R=R+1
CONTINUE
CONTINUE
TT=TT-1+(K+3)*(K+2)/2
CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
314
54
CONTINUE
c CASE-6, ZXY
DO 64 L=2,NP-2
TT=0
INT1=0.D0
INT2=0.D0
INT3=0.D0
INT4=0.D0
c
R=1
P=L
DO 63 K=1,P
M=P-K+2
DO 61 I=1,M
N=M-I+2
DO 62 J=1,N
C
G(R)=((1.D0-S(P,K))**2)*(1.D0-S(M,I))*W(M,I)*W(N,J)*W(P,K)/64.D0
U(R)=((1.D0+S(M,I))*(1.D0-S(P,K))/2.D0)-1.D0
V(R)=((1.D0-S(M,I))*(1.D0+S(N,J))*(1.D0-S(P,K))/4.D0)-1.D0
T(R)=S(P,K)
C
XR=(XC(1)+XC(2)+XC(3)-XC(4)+(XC(1)-XC(4))*U(R)+(XC(2)-XC(4))*V(R)+
1 (XC(3)-XC(4))*T(R))/2.D0
YR=(YC(1)+YC(2)+YC(3)-YC(4)+(YC(1)-YC(4))*U(R)+(YC(2)-YC(4))*V(R)+
1 (YC(3)-YC(4))*T(R))/2.D0
ZR=(ZC(1)+ZC(2)+ZC(3)-ZC(4)+(ZC(1)-ZC(4))*U(R)+(ZC(2)-ZC(4))*V(R)+
1 (ZC(3)-ZC(4))*T(R))/2.D0
C
INT1=INT1+F1(XR,YR,ZR)*G(R)*VOL
INT2=INT2+F2(XR,YR,ZR)*G(R)*VOL
INT3=INT3+F3(XR,YR,ZR)*G(R)*VOL
INT4=INT4+F4(XR,YR,ZR)*G(R)*VOL
R=R+1
62 CONTINUE
61 CONTINUE
TT=TT-1+(K+3)*(K+2)/2
63 CONTINUE
WRITE(11,9) TT, INT1,INT2,INT3,INT4
64 CONTINUE
STOP
END
C================================****===========================
PROGRAM 12 NIPTE.FOR
315
READ(1,*)X(J1),Y(J1)
PRINT*, X(J1),Y(J1)
ENDDO
IF(TE.NE.1) THEN
IF(NN.EQ.2) THEN
YY = 2.D0*Y(4) - Y(1) - Y(2)
XX = 2.D0*X(4) - X(1) - X(2)
B3 =-2.D0*YY
C3 = 2.D0*XX
ELSE
YY = Y(4) + Y(5) - Y(1) - Y(2)
XX = X(4) + X(5) - X(1) - X(2)
B3 =-9.D0*YY/4.D0
C3 = 9.D0*XX/4.D0
END IF
END IF
B1 = Y(2) - Y(3)
B2 = Y(3) - Y(1)
C1 = X(3) - X(2)
C2 = X(1) - X(3)
AL0 = B1*C2 - B2*C1
AL1 = B2*C3 - B3*C2
AL2 = B1*C3 - B3*C1
C==============================================================
IF(NN.EQ.2) THEN
DO 15 J=1,NP
E(1,J)= -1.D0*(D1(1,J)+3.D0*D1(3,J)-4.D0*D1(6,J))
E(2,J)= -1.D0*(D1(2,J)+3.D0*D1(3,J)-4.D0*D1(5,J))
E(3,J)= 4.D0*(D1(1,J)+D1(3,J)-2.D0*D1(6,J))
E(4,J)= 4.D0*(D1(3,J)+D1(4,J)-D1(5,J)-D1(6,J))
E(5,J)= 4.D0*(D1(2,J)+D1(3,J)-2.D0*D1(5,J))
15
CONTINUE
F(1) = (2.D0*AL1- AL2)/60.D0
F(2) =(2.D0*AL2- AL1)/60.D0
F(3) =(AL1 + AL2)/(-60.D0)
F(4) =(5.D0*AL0 + 2.D0*AL1 + 2.D0*AL2)/15.D0
F(5) =(5.D0*AL0 + AL1 + 2.D0*AL2)/15.D0
F(6) =(5.D0*AL0 + 2.D0*AL1 + AL2)/15.D0
C****FOR THE GLOBAL DERIVATIVES W.R.TO X AS IN EQUATION
(*)**********
CALL GDQT(NP,B1,B2,B3,E,AX)
C****FOR THE GLOBAL DERIVATIVES W.R.TO Y AS IN EQUATION
(*)**********
CALL GDQT(NP,C1,C2,C3,E,AY)
ELSE
DO 16 J=1,NP
317
E(1,J)=0.5D0*(2.D0*D1(1,J)-11.D0*D1(3,J)+18.D0*D1(8,J)-9.D0
1 *D1(9,J))
E(2,J)=0.5D0*(2.D0*D1(2,J)-11.*D1(3,J)-9.D0*D1(6,J)+18.D0*D1(7,J))
E(3,J)= -9.D0*(D1(1,J)-2.D0*D1(3,J)+5.D0*D1(8,J)-4.D0*D1(9,J))
E(4,J)= 4.5D0*(4.D0*D1(3,J)-D1(4,J)-D1(5,J)+D1(6,J)-5.D0*D1(7,J)
1
- 5.D0*D1(8,J)+D1(9,J)+6.D0*D1(10,J))
E(5,J)= -9.D0*(D1(2,J)-2.D0*D1(3,J)-4.D0*D1(6,J)+5.D0*D1(7,J))
E(6,J)= 13.5D0*(D1(1,J)-D1(3,J)+3.*D1(8,J)-3.D0*D1(9,J))
E(7,J)= 13.5D0*(D1(3,J)-D1(4,J)-D1(7,J)-2.D0*D1(8,J)+D1(9,J)
1
+2.D0*D1(10,J))
E(8,J)= -13.5D0*(D1(3,J)-D1(5,J)+D1(6,J)-2.D0*D1(7,J)-D1(8,J)
1
+2.D0*D1(10,J))
E(9,J)= 13.5D0*(D1(2,J)-D1(3,J)-3.D0*D1(6,J)+3.D0*D1(7,J))
16
CONTINUE
F(1) = (4.D0*AL0 + 2.D0*AL1 + AL2)/120.D0
F(2) = (4.D0*AL0 + AL1 + 2.D0*AL2)/120.D0
F(3) = (4.D0*AL0 + AL1 + AL2)/120.D0
F(4) = (AL0 + AL1)*(3.D0/40.D0)
F(5) = (AL0 + AL2)*(3.D0/40.D0)
F(6) = (AL0 + AL2)*(3.D0/40.D0)
F(7) = (3.D0/40.D0)*AL0
F(8) = (3.D0/40.D0)*AL0
F(9) = (AL0 + AL1)*(3.D0/40.D0)
F(10)= (9.D0*AL0 + 3.D0*AL1 + 3.D0*AL2)/20.D0
C*********FOR THE GLOBAL DERIVATIVES W.R.TO X AS IN EQUATION
(*)*******
CALL GDCT(NP,B1,B2,B3,E,AX)
C*********FOR THE GLOBAL DERIVATIVES W.R.TO Y AS IN EQUATION
(*)*******
CALL GDCT(NP,C1,C2,C3,E,AY)
END IF
C*****OUTPUT FOR AX AND AY WILL BE PRINTED BY THE FOLLOWING
FRAGMENTS****
WRITE( 2, 20) NP,N1
WRITE( 3, 20) NP,N1
20 FORMAT(/,8X,'FOR ',I2,'-NODED ELEMENT NO :- ',I2,/)
WRITE(2,21)((I,J,AX(I,J),I,J,AY(I,J),J=1,NP),I=1,NP)
21 FORMAT(2X,'AX(',I2,',',I2,')=',D18.10,2X,'AY(',I2,',',I2,')=',
1 D18.10)
C*******THIS PART FOR THE OUTPUT OF AL0,AL1,AL2 AND
F(I),I=1,2,3,...,NP***
WRITE(2,30) AL0,AL1,AL2
WRITE(3,30) AL0,AL1,AL2
30 FORMAT(2X,'AL0 =',D18.10,2X,'AL1 =',D18.10,2X,'AL2 =',D18.10)
WRITE(3,31) (II,F(II),II=1,NP)
31 FORMAT(3X,'F(',I2,')=',D18.10,2X,'F(',I2,')=',D18.10)
318
WRITE(3,32)
32 FORMAT(3X,'====================******================== ')
C******SUBROUTINE
'FACT'
CALLED
HERE
TO
OBTAIN
FACTORIAL(NP)*******
CALL FACT(NP,NFACT)
IF(TE.EQ.2) THEN
C******FOR ALL INTEGRALS SUBROUTINES ARE CALLED HERE****
RP=10
ITP=RP*RP
C FOR MM POINTS UNIT TRIANGLE ELEMENT
CALL NIUTS(N,RP,ITP,AL0,AL1,AL2,NFACT,AII)
C FOR M(M+1)/2-1 POINTS UNIT TRIANGLE ELEMENT
C
CALL NIUTM(N,RP,ITP,AL0,AL1,AL2,NFACT,AII)
C FOR MM POINTS STANDARD TRIANGLE ELEMENT
C
CALL NISTS(N,RP,ITP,AL0,AL1,AL2,NFACT,AII)
C FOR M(M+1)/2-1 POINTS STANDARD TRIANGLE ELEMENT
C
CALL NISTM(N,RP,ITP,AL0,AL1,AL2,NFACT,AII)
C FOR MM POINTS SYMMETRIC TRIANGLE ELEMENT
C
CALL NISYS(N,RP,ITP,AL0,AL1,AL2,NFACT,AII)
C FOR M(M+1)/2-1 POINTS SYMMETRIC TRIANGLE ELEMENT
C
CALL NISYM(N,RP,ITP,AL0,AL1,AL2,NFACT,AII)
WRITE(2,33) N1
33
FORMAT(3X,'FOR THE CURVED ELEMENT :-',I2,/,3X,'
1 CALCULATED INTEGRAL VALUES ARE',/)
DO 34 LM=1,N+1
DO 34 LK=1,LM
WRITE(2,35) LM-LK,LK-1,AII(LM-LK,LK-1)
35 FORMAT(3X,'AII(',I2,',',I2,')=',D18.10)
34 CONTINUE
ENDIF
CALL BMT(TE,NN,NP,N,AL0,NFACT,AII,B)
C*****THIS PART TO CALCULATE AND PRINTS THE COMPONENTS OF
STIFFNESS MATRICES*******
DO 36 I = 1,NP
DO 37 J = I,NP
KXX(I,J) = 0.0
KYY(I,J) = 0.0
IF(I.EQ.1) THEN
DO 40 II = 1,NP
CX(II,J) = 0.0
CY(II,J) = 0.0
DO 45 K1 = 1,NP
CX(II,J) = CX(II,J) +B(II,K1)*AX(K1,J)
CY(II,J) = CY(II,J) +B(II,K1)*AY(K1,J)
45
CONTINUE
319
40
CONTINUE
END IF
DO 50 L = 1,NP
KXX(I,J) = KXX(I,J) +AX(L,I)*CX(L,J)
KYY(I,J) = KYY(I,J) +AY(L,I)*CY(L,J)
50
CONTINUE
K(I,J) = KXX(I,J) +KYY(I,J)
C********OUTPUT OF KXX(I,J),KYY(I,J) AND K(I,J)************************
WRITE(3,51) I,J
51 FORMAT(3X,'FOR I =',I2,2X,'AND J =',I2)
WRITE(3,52) I,J,KXX(I,J),I,J,KYY(I,J),I,J, K(I,J)
52 FORMAT(3X,'CALCULATED',/,3X,'KXX(',I2,',',I2,')=',
1 D18.10,3X,'KYY(',I2,',',I2,')=',D18.10,/,3X,'AND HENCE',/
2 ,3X,'K(',I2,','I2,')=',D18.10)
WRITE(3,53)
53 FORMAT(3X,'=====================****====================',/)
37 CONTINUE
36 CONTINUE
WRITE (3, 60) N1
60 FORMAT(/,3X, 'GIVEN COORDONATES FOR ELEMENT NO:-',I2,/)
DO 61 IN = 1, JNODE
WRITE( 3,62) IN,X(IN),IN,Y(IN)
62 FORMAT(3X,'X(',I2,')=',D18.10,3X,'Y(',I2,')=',D18.10)
61 CONTINUE
WRITE( 3,63) B1,C1,B2,C2,B3,C3
63 FORMAT(/,3X,'CALCULATED',/,3X,'B1 =',D18.10,3X,'C1 =',D18.10,/,
1 3X,'B2 =',D18.10,3X,'C2 =',D18.10,/,3X,'B3 =',D18.10,3X,
1 'C3=',D18.10,/)
14 CONTINUE
PRINT*,'COMPUTATION OVER'
STOP
END
C============================================================
SUBROUTINE FACT(NP,NFACT)
DOUBLE PRECISION NFACT(0:NP)
DO 100 NFF = 1, NP+1
NF = NFF-1
IF(NF.EQ.0) THEN
NFACT(NF) = 1.D0
ELSE
NFACT(NF) = NF* NFACT(NF-1)
ENDIF
100
CONTINUE
RETURN
END
C==============================================================
320
SUBROUTINE GDQT(NP,S1,S2,S3,E,AT)
DOUBLE PRECISION S1,S2,S3,AT(NP,NP),E(NP-1,NP)
DO 100 I = 1,NP
AT(1,I) = S1 * E(1,I) + S2 * E(2,I)
AT(2,I) = S1 * E(3,I) + S2 * E(4,I) - S3 * E(1,I)
AT(3,I) = S1 * E(4,I) + S2 * E(5,I) + S3 * E(2,I)
AT(4,I) = -S3 * E(3,I)
AT(5,I) = 0.D0
AT(6,I) = S3 * E(5,I)
100 CONTINUE
RETURN
END
C==============================================================
SUBROUTINE GDCT(NP,S1,S2,S3,E,AT)
DOUBLE PRECISION S1,S2,S3,AT(NP,NP),E(NP-1,NP)
DO 105 I = 1,NP
AT(1,I) = S1*E(1,I)+S2*E(2,I)
AT(2,I) = S1*E(3,I)+S2*E(4,I)-S3*E(1,I)
AT(3,I) = S1*E(4,I)+S2*E(5,I)+S3*E(2,I)
AT(4,I) = S1*E(6,I)-S2*E(7,I)-S3*E(3,I)
AT(5,I) = -2.*(S1*E(7,I)-S2*E(8,I))
AT(6,I) = S1*E(8,I)+S2*E(9,I)+S3*E(5,I)
AT(7,I) = -S3*E(6,I)
AT(8,I) = S3*E(7,I)
AT(9,I) = S3*E(8,I)
AT(10,I)= S3*E(9,I)
105 CONTINUE
RETURN
END
C==============================================================
SUBROUTINE BMT(TE,NN,NP,N,AL0,NFACT,AII,B)
DOUBLE PRECISION B(NP,NP),NFACT(0:NP),AII(0:N,0:N),AL0
INTEGER TE
I=0
DO 102 IRR=1,NN+1
IR=IRR-1
DO 103 ISS=1,IRR
IS=ISS-1
I=I+1
J=0
DO 104 LL=1,NN+1
L=LL-1
DO 105 KK=1,LL
K=KK-1
J=J+1
IF(TE.EQ.1) THEN
321
B(I,J)=NFACT(L+IR-IS-K)*NFACT(K+IS)/(NFACT(L+IR+2)*AL0)
ELSE
B(I,J)=AII(L+IR-IS-K,K+IS)
END IF
105
CONTINUE
104
CONTINUE
103 CONTINUE
102 CONTINUE
RETURN
END
C==============================================================
SUBROUTINE NIUTS(N,NP,ITP,AL0,AL1,AL2,NFACT,AII)
DOUBLE PRECISION AII(0:N, 0:N),NFACT(0:NP),
1 AL0,AL1,AL2,U(ITP),V(ITP),G(ITP),XR,YR,Z,F,SUM(0:N,0:N),XR1,YR1,
2S(2:NP,NP),W(2:NP,NP),TEMP(N+1),PT,ETOL,U1(ITP),V1(ITP),G1(ITP)
REAL P
ETOL = 0.0000001D0
CALL GQPW(NP,S,W)
DO 7 M=2,NP
NN=M
PT=0.D0
DO P =0,N
AII(P,0)=0.D0
AII(0,P)=0.D0
R=1
DO 9 J=1,NN
DO 10 I=1,M
G(R)=(1.D0-S(M,I))*W(M,I)*W(NN,J)/8.D0
U(R)=(1.D0+S(M,I))/2.D0
V(R)=(1.D0-S(M,I))*(1.D0+S(NN,J))/4.D0
XR=U(R)
YR=V(R)
G1(R)=(1.D0-S(NN,J))*W(M,I)*W(NN,J)/8.D0
U1(R)=(1.D0+S(M,I))*(1.D0-S(NN,J))/4.D0
V1(R)=(1.D0+S(NN,J))/2.D0
XR1=U1(R)
YR1=V1(R)
IF((AL2.NE.0.).AND.(AL1.EQ.0.)) THEN
Z=F(P,0,AL0,AL1,AL2,XR1,YR1)
AII(P,0)=AII(P,0)+Z*G1(R)
ELSEIF((AL1.NE.0.).AND.(AL2.EQ.0.)) THEN
Z=F(0,P,AL0,AL1,AL2,XR,YR)
AII(0,P)=AII(0,P)+Z*G(R)
322
ELSEIF((AL1.NE.0.).AND.(DABS(AL2).GE.DABS(AL1))) THEN
Z=F(P,0,AL0,AL1,AL2,XR,YR)
AII(P,0)=AII(P,0)+Z*G(R)
ELSEIF((AL2.NE.0.).AND.(DABS(AL1).GT.DABS(AL2))) THEN
Z=F(0,P,AL0,AL1,AL2,XR1,YR1)
AII(0,P)=AII(0,P)+Z*G1(R)
ENDIF
R=R+1
10 CONTINUE
9 CONTINUE
ENDDO
7 CONTINUE
IF(DABS(AL1).GE.DABS(AL2))THEN
DO 178 A=0,N-1
DO 180 B=0,N-A-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A+1,B)=(1./AL1)*(SUM(A,B)-AL0*AII(A,B)
1 -AL2*AII(A,B+1))
180
CONTINUE
178
CONTINUE
ELSE
DO 188 B=0,N-1
DO 190 A=0,N-B-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A,B+1)=(1./AL2)*(SUM(A,B)-AL0*AII(A,B)
1 -AL1*AII(A+1,B))
190
CONTINUE
188
CONTINUE
ENDIF
RETURN
END
C================================****===========================
SUBROUTINE NIUTM(N,NP,ITP,AL0,AL1,AL2,NFACT,AII)
DOUBLE PRECISION AII(0:N, 0:N),NFACT(0:NP),
1 AL0,AL1,AL2,U(ITP),V(ITP),G(ITP),XR,YR,Z,F,SUM(0:N,0:N),XR1,YR1,
2S(2:NP,NP),W(2:NP,NP),TEMP(N+1),PT,ETOL,U1(ITP),V1(ITP),G1(ITP)
REAL P
ETOL = 0.0000001D0
CALL GQPW(NP,S,W)
DO 7 M=2,NP
PT=0.D0
DO P =0,N
AII(P,0)=0.D0
AII(0,P)=0.D0
R=1
323
DO 9 I=1,M
NN=M-I+2
DO 10 J=1,NN
G(R)=(1.D0-S(M,I))*W(M,I)*W(NN,J)/8.D0
U(R)=(1.D0+S(M,I))/2.D0
V(R)=(1.D0-S(M,I))*(1.D0+S(NN,J))/4.D0
XR=U(R)
YR=V(R)
G1(R)=(1.D0-S(M,I))*W(M,I)*W(NN,J)/8.D0
U1(R)=(1.D0+S(NN,J))*(1.D0-S(M,I))/4.D0
V1(R)=(1.D0+S(M,I))/2.D0
XR1=U1(R)
YR1=V1(R)
10
9
7
IF((AL2.NE.0.).AND.(AL1.EQ.0.)) THEN
Z=F(P,0,AL0,AL1,AL2,XR1,YR1)
AII(P,0)=AII(P,0)+Z*G1(R)
ELSEIF((AL1.NE.0.).AND.(AL2.EQ.0.)) THEN
Z=F(0,P,AL0,AL1,AL2,XR,YR)
AII(0,P)=AII(0,P)+Z*G(R)
ELSEIF((AL1.NE.0.).AND.(DABS(AL2).GE.DABS(AL1))) THEN
Z=F(P,0,AL0,AL1,AL2,XR,YR)
AII(P,0)=AII(P,0)+Z*G(R)
ELSEIF((AL2.NE.0.).AND.(DABS(AL1).GT.DABS(AL2))) THEN
Z=F(0,P,AL0,AL1,AL2,XR1,YR1)
AII(0,P)=AII(0,P)+Z*G1(R)
ENDIF
R=R+1
CONTINUE
CONTINUE
ENDDO
CONTINUE
IF(DABS(AL1).GE.DABS(AL2))THEN
DO 178 A=0,N-1
DO 180 B=0,N-A-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A+1,B)=(1./AL1)*(SUM(A,B)-AL0*AII(A,B)
1 -AL2*AII(A,B+1))
180
CONTINUE
178
CONTINUE
ELSE
DO 188 B=0,N-1
DO 190 A=0,N-B-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
324
AII(A,B+1)=(1./AL2)*(SUM(A,B)-AL0*AII(A,B)
-AL1*AII(A+1,B))
CONTINUE
CONTINUE
ENDIF
RETURN
END
C================================****===========================
SUBROUTINE NISTS(N,NP,ITP,AL0,AL1,AL2,NFACT,AII)
DOUBLE PRECISION AII(0:N, 0:N),NFACT(0:NP),
1 AL0,AL1,AL2,U(ITP),V(ITP),G(ITP),XR,YR,Z,F,SUM(0:N,0:N),
2S(2:NP,NP),W(2:NP,NP),TEMP(N+1),PT,ETOL,XC(3),YC(3),
3U1(ITP),V1(ITP),G1(ITP),XR1,YR1
REAL P
DATA XC(1)/1.D0/, XC(2)/0.D0/,XC(3)/0.D0/,YC(1)/0.D0/, YC(2)/1.D0/
1 ,YC(3)/0.D0/
ETOL = 0.0000001D0
CALL GQPW(NP,S,W)
1
190
188
DO 7 M=2,NP
NN=M
PT=0.D0
DO P =0,N
AII(P,0)=0.D0
AII(0,P)=0.D0
R=1
DO 9 J=1,NN
DO 10 I=1,M
G(R)=(1.D0-S(M,I))*W(M,I)*W(NN,J)/8.D0
U(R)=S(M,I)
V(R)=(1.D0-S(M,I))*(1.D0+S(NN,J))/2.D0-1.D0
XR=(XC(1)+XC(2)+(XC(1)-XC(3))*U(R)+(XC(2)-XC(3))*V(R))/2.D0
YR=(YC(1)+YC(2)+(YC(1)-YC(3))*U(R)+(YC(2)-YC(3))*V(R))/2.D0
G1(R)=(1.D0-S(NN,J))*W(M,I)*W(NN,J)/8.D0
U1(R)=(1.D0-S(NN,J))*(1.D0+S(M,I))/2.D0-1.D0
V1(R)=S(NN,J)
XR1=(XC(1)+XC(2)+(XC(1)-XC(3))*U1(R)+(XC(2)-XC(3))*V1(R))/2.D0
YR1=(YC(1)+YC(2)+(YC(1)-YC(3))*U1(R)+(YC(2)-YC(3))*V1(R))/2.D0
IF((AL2.NE.0.).AND.(AL1.EQ.0.)) THEN
Z=F(P,0,AL0,AL1,AL2,XR1,YR1)
AII(P,0)=AII(P,0)+Z*G1(R)
ELSEIF((AL1.NE.0.).AND.(AL2.EQ.0.)) THEN
Z=F(0,P,AL0,AL1,AL2,XR,YR)
325
AII(0,P)=AII(0,P)+Z*G(R)
ELSEIF((AL1.NE.0.).AND.(DABS(AL2).GE.DABS(AL1))) THEN
Z=F(P,0,AL0,AL1,AL2,XR,YR)
AII(P,0)=AII(P,0)+Z*G(R)
ELSEIF((AL2.NE.0.).AND.(DABS(AL1).GT.DABS(AL2))) THEN
Z=F(0,P,AL0,AL1,AL2,XR1,YR1)
AII(0,P)=AII(0,P)+Z*G1(R)
ENDIF
R=R+1
10 CONTINUE
9 CONTINUE
ENDDO
7 CONTINUE
IF(DABS(AL1).GE.DABS(AL2))THEN
DO 178 A=0,N-1
DO 180 B=0,N-A-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A+1,B)=(1./AL1)*(SUM(A,B)-AL0*AII(A,B)
1 -AL2*AII(A,B+1))
180
CONTINUE
178
CONTINUE
ELSE
DO 188 B=0,N-1
DO 190 A=0,N-B-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A,B+1)=(1./AL2)*(SUM(A,B)-AL0*AII(A,B)
1 -AL1*AII(A+1,B))
190
CONTINUE
188
CONTINUE
ENDIF
RETURN
END
C==============================================================
SUBROUTINE NISTM(N,NP,ITP,AL0,AL1,AL2,NFACT,AII)
DOUBLE PRECISION AII(0:N, 0:N),NFACT(0:NP),
1 AL0,AL1,AL2,U(ITP),V(ITP),G(ITP),XR,YR,Z,F,SUM(0:N,0:N),
2S(2:NP,NP),W(2:NP,NP),TEMP(N+1),PT,ETOL,XC(3),YC(3),
3U1(ITP),V1(ITP),G1(ITP),XR1,YR1
REAL P
DATA XC(1)/1.D0/, XC(2)/0.D0/,XC(3)/0.D0/,YC(1)/0.D0/, YC(2)/1.D0/
1,YC(3)/0.D0/
ETOL = 0.0000001D0
CALL GQPW(NP,S,W)
326
DO 7 M=2,NP
PT=0.D0
DO P =0,N
AII(P,0)=0.D0
AII(0,P)=0.D0
R=1
DO 9 I=1,M
NN=M-I+2
DO 10 J=1,NN
G(R)=(1.D0-S(M,I))*W(M,I)*W(NN,J)/8.D0
U(R)=S(M,I)
V(R)=(1.D0-S(M,I))*(1.D0+S(NN,J))/2.D0-1.D0
XR=(XC(1)+XC(2)+(XC(1)-XC(3))*U(R)+(XC(2)-XC(3))*V(R))/2.D0
YR=(YC(1)+YC(2)+(YC(1)-YC(3))*U(R)+(YC(2)-YC(3))*V(R))/2.D0
G1(R)=(1.D0-S(M,I))*W(M,I)*W(NN,J)/8.D0
U1(R)=(1.D0+S(NN,J))*(1.D0-S(M,I))/2.D0-1.D0
V1(R)=S(M,I)
XR1=(XC(1)+XC(2)+(XC(1)-XC(3))*U1(R)+(XC(2)-XC(3))*V1(R))/2.D0
YR1=(YC(1)+YC(2)+(YC(1)-YC(3))*U1(R)+(YC(2)-YC(3))*V1(R))/2.D0
IF((AL2.NE.0.).AND.(AL1.EQ.0.)) THEN
Z=F(P,0,AL0,AL1,AL2,XR1,YR1)
AII(P,0)=AII(P,0)+Z*G1(R)
ELSEIF((AL1.NE.0.).AND.(AL2.EQ.0.)) THEN
Z=F(0,P,AL0,AL1,AL2,XR,YR)
AII(0,P)=AII(0,P)+Z*G(R)
ELSEIF((AL1.NE.0.).AND.(DABS(AL2).GE.DABS(AL1))) THEN
Z=F(P,0,AL0,AL1,AL2,XR,YR)
AII(P,0)=AII(P,0)+Z*G(R)
ELSEIF((AL2.NE.0.).AND.(DABS(AL1).GT.DABS(AL2))) THEN
Z=F(0,P,AL0,AL1,AL2,XR1,YR1)
AII(0,P)=AII(0,P)+Z*G1(R)
ENDIF
R=R+1
10 CONTINUE
9 CONTINUE
ENDDO
7 CONTINUE
IF(DABS(AL1).GE.DABS(AL2))THEN
DO 178 A=0,N-1
DO 180 B=0,N-A-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
327
AII(A+1,B)=(1./AL1)*(SUM(A,B)-AL0*AII(A,B)
-AL2*AII(A,B+1))
CONTINUE
CONTINUE
ELSE
DO 188 B=0,N-1
DO 190 A=0,N-B-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A,B+1)=(1./AL2)*(SUM(A,B)-AL0*AII(A,B)
1 -AL1*AII(A+1,B))
190
CONTINUE
188
CONTINUE
ENDIF
RETURN
END
C================================****===========================
SUBROUTINE NISYS(N,NP,ITP,AL0,AL1,AL2,NFACT,BJJ,AII)
DOUBLE PRECISION AII(0:N, 0:N),NFACT(0:NP),BJJ(0:N, 0:N),
1 AL0,AL1,AL2,U(ITP),V(ITP),G(ITP),XR,YR,Z,F,SUM(0:N,0:N),
2S(2:NP,NP),W(2:NP,NP),TEMP(N+1),PT,ETOL,XC(3),YC(3),
3U1(ITP),V1(ITP),G1(ITP),XR1,YR1
REAL P
DATA XC(1)/1.D0/, XC(2)/0.D0/,XC(3)/0.D0/,YC(1)/0.D0/, YC(2)/1.D0/
1 ,YC(3)/0.D0/
ETOL = 0.0000001D0
CALL GQPW(NP,S,W)
1
180
178
DO 7 M=2,NP
NN=M
PT=0.D0
DO P =0,N
AII(P,0)=0.D0
AII(0,P)=0.D0
R=1
DO 9 J=1,NN
DO 10 I=1,M
G(R)=(1.D0-S(M,I))*W(M,I)*W(NN,J)/8.D0
U(R)=S(M,I)
V(R)=(1.D0-S(M,I))*S(NN,J)/2.D0
XR=((2.D0*XC(1)+XC(2)+XC(3))+((2.D0*XC(1)-XC(2)-XC(3))*U(R))+
1
((2.D0*XC(2)-2.D0*XC(3))*V(R)))/4.D0
YR= ((2.D0*YC(1)+YC(2)+YC(3))+((2.D0*YC(1)-YC(2)-YC(3))*U(R))+
1
((2.D0*YC(2)-2.D0*YC(3))*V(R)))/4.D0
G1(R)=(1.D0-S(NN,J))*W(M,I)*W(NN,J)/8.D0
328
U1(R)=(1.D0-S(NN,J))*S(M,I)/2.D0
V1(R)=S(NN,J)
XR1= ((XC(1)+2.D0*XC(2)+XC(3))+2.D0*(XC(1)-XC(3))*U1(R)+
1
(-XC(1)+2.D0*XC(2)-XC(3))*V1(R))/4.D0
YR1= ((YC(1)+2.D0*YC(2)+YC(3))+2.D0*(YC(1)-YC(3))*U1(R)+
1
(-YC(1)+2.D0*YC(2)-YC(3))*V1(R))/4.D0
IF((AL2.NE.0.).AND.(AL1.EQ.0.)) THEN
Z=F(P,0,AL0,AL1,AL2,XR1,YR1)
AII(P,0)=AII(P,0)+Z*G1(R)
ELSEIF((AL1.NE.0.).AND.(AL2.EQ.0.)) THEN
Z=F(0,P,AL0,AL1,AL2,XR,YR)
AII(0,P)=AII(0,P)+Z*G(R)
ELSEIF((AL1.NE.0.).AND.(DABS(AL2).GE.DABS(AL1))) THEN
Z=F(P,0,AL0,AL1,AL2,XR,YR)
AII(P,0)=AII(P,0)+Z*G(R)
ELSEIF((AL2.NE.0.).AND.(DABS(AL1).GT.DABS(AL2))) THEN
Z=F(0,P,AL0,AL1,AL2,XR1,YR1)
AII(0,P)=AII(0,P)+Z*G1(R)
ENDIF
R=R+1
10 CONTINUE
9 CONTINUE
ENDDO
7
CONTINUE
IF(DABS(AL1).GE.DABS(AL2))THEN
DO 178 A=0,N-1
DO 180 B=0,N-A-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A+1,B)=(1./AL1)*(SUM(A,B)-AL0*AII(A,B)
1 -AL2*AII(A,B+1))
180
CONTINUE
178
CONTINUE
ELSE
DO 188 B=0,N-1
DO 190 A=0,N-B-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A,B+1)=(1./AL2)*(SUM(A,B)-AL0*AII(A,B)
1 -AL1*AII(A+1,B))
190
CONTINUE
188
CONTINUE
ENDIF
RETURN
END
C================================****===========================
329
SUBROUTINE NISYM(N,NP,ITP,AL0,AL1,AL2,NFACT,AII)
DOUBLE PRECISION AII(0:N, 0:N),NFACT(0:NP),
1 AL0,AL1,AL2,U(ITP),V(ITP),G(ITP),XR,YR,Z,F,SUM(0:N,0:N),
2S(2:NP,NP),W(2:NP,NP),TEMP(N+1),PT,ETOL,XC(3),YC(3),
3U1(ITP),V1(ITP),G1(ITP),XR1,YR1
REAL P
DATA XC(1)/1.D0/, XC(2)/0.D0/,XC(3)/0.D0/,YC(1)/0.D0/, YC(2)/1.D0/
1,YC(3)/0.D0/
ETOL = 0.0000001D0
CALL GQPW(NP,S,W)
DO 7 M=2,NP
PT=0.D0
DO P =0,N
AII(P,0)=0.D0
AII(0,P)=0.D0
R=1
DO 9 I=1,M
NN=M-I+2
DO 10 J=1,NN
G(R)=(1.D0-S(M,I))*W(M,I)*W(NN,J)/8.D0
U(R)=S(M,I)
V(R)=(1.D0-S(M,I))*S(NN,J)/2.D0
XR= ((2.D0*XC(1)+XC(2)+XC(3))+((2.D0*XC(1)-XC(2)-XC(3))*U(R))+
1
((2.D0*XC(2)-2.D0*XC(3))*V(R)))/4.D0
YR= ((2.D0*YC(1)+YC(2)+YC(3))+((2.D0*YC(1)-YC(2)-YC(3))*U(R))+
1
((2.D0*YC(2)-2.D0*YC(3))*V(R)))/4.D0
G1(R)=(1.D0-S(M,I))*W(NN,J)*W(M,I)/8.D0
U1(R)=(1.D0-S(M,I))*S(NN,J)/2.D0
V1(R)=S(M,I)
XR1= ((XC(1)+2.D0*XC(2)+XC(3)) +2.D0*(XC(1)-XC(3))*U1(R)+
1
(-XC(1)+2.D0*XC(2)-XC(3))*V1(R))/4.D0
YR1= ((YC(1)+2.D0*YC(2)+YC(3)) +2.D0*(YC(1)-YC(3))*U1(R)+
1
(-YC(1)+2.D0*YC(2)-YC(3))*V1(R))/4.D0
IF((AL2.NE.0.).AND.(AL1.EQ.0.)) THEN
Z=F(P,0,AL0,AL1,AL2,XR1,YR1)
AII(P,0)=AII(P,0)+Z*G1(R)
ELSEIF((AL1.NE.0.).AND.(AL2.EQ.0.)) THEN
Z=F(0,P,AL0,AL1,AL2,XR,YR)
AII(0,P)=AII(0,P)+Z*G(R)
ELSEIF((AL1.NE.0.).AND.(DABS(AL2).GE.DABS(AL1))) THEN
Z=F(P,0,AL0,AL1,AL2,XR,YR)
330
AII(P,0)=AII(P,0)+Z*G(R)
ELSEIF((AL2.NE.0.).AND.(DABS(AL1).GT.DABS(AL2))) THEN
Z=F(0,P,AL0,AL1,AL2,XR1,YR1)
AII(0,P)=AII(0,P)+Z*G1(R)
ENDIF
R=R+1
10 CONTINUE
9 CONTINUE
ENDDO
7 CONTINUE
IF(DABS(AL1).GE.DABS(AL2))THEN
DO 178 A=0,N-1
DO 180 B=0,N-A-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A+1,B)=(1./AL1)*(SUM(A,B)-AL0*AII(A,B)
1 -AL2*AII(A,B+1))
180
CONTINUE
178
CONTINUE
ELSE
DO 188 B=0,N-1
DO 190 A=0,N-B-1
SUM(A,B)= (NFACT(A)*NFACT(B))/(NFACT(A+B+2))
AII(A,B+1)=(1./AL2)*(SUM(A,B)-AL0*AII(A,B)
1 -AL1*AII(A+1,B))
190
CONTINUE
188
CONTINUE
ENDIF
RETURN
END
C====================================================
DOUBLE PRECISION FUNCTION F(P,Q,AL0,AL1,AL2,X,Y)
DOUBLE PRECISION AL0,AL1,AL2,X,Y
F=(X**P)*(Y**Q)/(AL0+AL1*X+AL2*Y)
RETURN
END
C==============================================================
331
References
M. Abramowitz and Stegun. Handbook of mathematical functions. Dover, 1974.
42, 56
C. M. Andersen and R. J. Mcleod. Integration technique for isoparametric
higher order bases on finite elements with a curved side. Comp. Math. Appl.,
5:285295, 1979. 10
J. H. Argyris. Recent Advances in Matrix Methods of Structural Analysis. Pergamon Press, Almsford, N.Y., 1954. 19
M. L. Baart and R. J. Mcleod. Recursive calculation of curved finite elements
stiffness matrices. Report: CSIR, Pretoria, South Africa, 1983. 10
D. Babu and G.F. Pinder. Analytical integration formulae for linear isoparametric. International journal on Numerical Meth. Engneering, 20:11531166,
1984. 6, 10
K. J. Bathe. Finite Element Procedures, 4th Ed. Prentice-Hall, Englewood Cliffs,
N.J., 1996. 42, 55, 67, 101
F. Bernardini. Integrating of polynomials over n-dimensional polyhedra. Comput. Aid. Des., 23:5158, 1991. 6, 11, 100, 129, 171, 203
W. B. Bickford. A First Course in Finite Element Method. Richard D. Irwin
Inc., 1990. 20
G. F. Carey and T. J. Oden. Finite Elements: A Second Course. Prentice Hall,
Englewood Cliffs, N. J., 1983. 23
332
REFERENCES
C. Cattani and A. Paoluzzi. Boundary integration over linear polyhedra.
Comput-Aid. Des., 22:130135, 1990a. 6, 11, 171, 203
C. Cattani and A. Paoluzzi. Symbolic analysis of linear polyhedra. Enrng
Comput., 6:1729, 1990b. 6, 11, 171, 203
S. C. Chapra and R. P. Canale. The Numerical Methods for Engineers, 5th ed.
McGraw-Hill Publishing Company, 2006. 147, 221
P. G. Ciarlet and R. A. Raviart. General lagrange and hermite interpolation
in rn application to finite element method. Arc. Ret. Mech. Anal, 4:177199,
1972. 9
R. W. Clough. The finite element method in plane stress analysis. Proc., 2nd
Conf. Electronic Computations, ASCE, pages 345378, 1960. 19
R. W. Clough and Peterson. Application of finite element method. The mathematics of Finite elements and Application, 2:1520, 1976. 4
Ronald Cools. An encyclopaedia of cubature formulas. Journal of Complexity,
19(3):445453, 2003. 129
Ronald Cools and Philip Rabinowitz. Monomial cubature rules since Stroud
- a compilation. Journal of Computational and Applied Mathematics, 48(3):
309326, 1993. 129
G. R. Cowper. Gaussian quadrature formulas for triangles. International journal
on numerical methods and engineering, 7:405408, 1973. 6, 9, 42, 56, 67, 100,
129
C. S. Desai. Elementary Finite Element Method. Prentice Hall Inc., Englewood
Cliffs, N. J., 1979. 20
D. A. Dunavant. High degree efficient symmetrical gaussian quadrature rules for
the triangle. International Journal for Numerical Methods in Engineering, 21:
11291148, 1985. 42, 56, 100, 129, 130, 144, 145, 147
333
REFERENCES
J. G. Ergatoudis, B. M. Irons, and O. C. Zienkiewicz. Curved isoparametric
quadrilateral element for finite analysis. Int. J. Solids and Struct., 4:3142,
1968. 9
B. A. Finlayson. The Method of Weighthed Residuals and Variation Principles.
Academic Press, N. Y., 1972. 20, 21, 31
C. A. J. Fletcher. Computational Galerkins Method. Springer-Verlag, Berlin,
1984. 20
B. G. Galerkin. Series occuring in some problems of elasticstability of rods and
plates. Engineering Bulletine, 19:897908, 1915. 3, 20, 30
C. F. Gauss. Methodus nova integratiam valores per approximationem inveniendi
worke. Int. J., 3:163196, 1814. 171, 203
D. V. Griffths. Stiffness matrix of the four -node quadrilateral element in closed
form. Int. J. Num. Methods and Eng., 165:10271038, 1994. 6, 10
D. V. Griffths and G. G. W. Mustoe. Selective reduced integration of four-node
plane element in closed form. J.A.Mech, pages 725729, 1995. 6, 10
W. L. Hacker and H. L. Schreyer. Eigen value analysis of compatible and
incompatible finite element. Int. J. Num. Methods and Engineers, 28:687703,
1989. 6, 10
P. C. Hammer and A. H. Stroud. Numerical integration over simplexes. Math
Tables Other Aids Computation, 10:137139, 1956. 42, 56, 67, 100, 129, 203
P. C. Hammer and A. H. Stroud. Numerical evaluation of multiple integrals.
Math Tables Other Aids Computation, 12:272280, 1958. 42, 56, 67, 100, 129,
203
P. C. Hammer, A. H. Stroud, and O. J. Marlowe. Numerical integration over
simplex and cones. Math Tables Other Aids Computation, 10:130136, 1956.
42, 56, 67, 100, 101, 129, 203
J. Heuser. Finite Element Method for Thermal Analysis. NASA Technical Note
TN D-7274, Goddard Space Flight Center, Greenbelt, MD, 1975. 20
334
REFERENCES
P. Hillion. Numerical integration on a triangle. International journal on numerical methods and engineering, 11:797815, 1977. 42, 67, 100, 129
K. H. Huebner. The Finite Element Method for Engineers. John Wiley and
Sons, N.Y., 1975. 20
T. J. R. Hughes. The Finite Element Method. Prentice-Hall, Englewood Cliffs,
N.J., 1987. 20, 42, 55, 66, 100, 171, 203
B. M. Irons and A. Razzaque. Experience with the patch-test for convergence of
finite element method. Academic press, New York, 1972. 42, 55, 67, 101
C. Johnson. Numerical Solution of Partial Differential Equation by Finite Element Method. Cambridge University Press, Cambridge, U. K., 1987. 20
M. S. Karim. Integration of Some Bivariate Polymials with Rational Denumerators -An Application to Finite Element Method. A Thesis Submitted for the
Degree of DOCTOR OF PHYLOSOPHY, Department of Mathematics, Bangalore University, Bangalore, India, 2001. 6, 11, 28, 243, 248, 251, 255
J. B. Keller. Stochastic equations and wave propagation in random media. American Mathematical Society, Providence, R.I., 16, 1964. Proc., Symposium on
Applied Mathematics. 42, 56, 67
J. B. Keller and H. P. McKean. siam-ams proceedings - stochastic differential
equations. American Mathematical Society, Providence, R.I., VI, 1973. 42, 56,
67
G. Lague and R. Baldur. Extended numerical integration method for triangular
surfaces. International journal on numerical methods and engineering, 11:388
392, 1977. 6, 9, 42, 56, 67, 100, 101, 104, 129
F. G. Lannoy. Triangular finite element and numerical integration. Computers
Struct., 7:613, 1977. 42, 67, 100, 129
D. P. Laurie. Automatic numerical integration over a triangle. National Institute
for Mathematical Science, Pretoria, CSIR Spec. Rep. WISK 273, 1977. 42, 67,
100, 129
335
REFERENCES
M. E. Laursen and M. Gellert. Some criteria for numerically integrated matrices and quadrature formulas for triangles. International journal for numerical
methods in engineering, 12:6776, 1978. 42, 67, 100, 129
Y. T. Lee and A. A. G. Requicha. Algorithms for computing the volume and
other integral properties of solids i: known methods and open issues. Commun.
ACM, 25:635641, 1982. 6, 11, 171, 203
F. G. Lether. Computation of double integrals over a triangle. Journal comp.
applic. Math., 2:219224, 1976. 42, 56, 67, 100, 129
P. E. Lewis and J. P. Ward. The Finite element Method. Addison-Wesley
Publishing Company, 1991. 247
S. Lien and J.T. Kajiya. A symbolic method for calculating the integral properties of arbitrary nonconvex polyhedra. IEEE Comput. Graph. Applic., 4:3541,
1984. 6, 11, 171, 203
J. N. Lyness and Ronald Cools. A survey of numerical cubature over triangles. Proceedings of Symposia in Applied Mathematics, American Mathematical
Society, pages 127150, 1994. 129
J. N. Lyness and D. Jespersen. Moderate degree symmetric quadrature rules
for triangle. Journal of the Institute of Mathematics and its Applications, 15(1):
1932, 1975. 42, 56, 67, 100, 129
H. C. Martin. Finite element analysis of fluid flow. Proc., 2nd Conf. Matrix
Method in Structural Mechanics, Wright Patterson Air Force Bare, 1968. 19
R. Mcleod. High order transformation method for curved finite elements. J. Int.
Math. Appl., 21:419428, 1978. 9, 10
R. Mcleod and A. R. Mitchel. The construction of basic function for curved
element in the finite element method. J. Int. Math. Application, 10:382393,
1972. 8, 9, 10
R. Mcleod and A. R. Mitchel. Use of parabolic arc in matching curved boundaries in the finite element method. J. Int. Math. Appl., 16:239246, 1975. 9
336
REFERENCES
J. E. Molyneux. Analysis of dishonest methods in the theory of wave propagation in a random medium. J. Opt. Soc. Am., 58:951957, 1968. 42, 56,
67
S. H. Nguyen. An accurate finite element formulation of linear elastic torsion
calculations. Computers and structures, 42(5):111121, 1992. 251, 252, 266
J. T. Oden. Finite Elements of Nonlinear continua. McGraw Hill Book Co, N.Y.,
1972. 19, 22
T. J. Oden and C. T. Reddy. An Introduction to Mathematical Theory of Finite
Elements. John Wiley and Sons, N. J., 1976. 23
M. Okabe. Analytic integration formula related to convex quadrilateral finite
elements. Computer Methods Appl. Mechan. Engng, 29:201218, 1981. 6, 10
D. R. J. Owen and E. Hinton. A Simple Guide for Finite Elements. Pineridge
Press, U. K., 1980. 20
H. T. Rathod. Some analytical integration formulae for a four-node isoparametric
element. Computers and Structures, 30:11011109, 1988. 6, 9, 10
H. T. Rathod and M. S. Islam. Integration of rational functions of bivariate
polymial numerators with linear denumerators over a square in local parametric
space. Computers Methods in Applied Mechanics and Engineering, 161:195213,
1998. 6
H. T. Rathod and M. S. Karim. Synthetic division based integration of rational
functions of bivariate polymial numerators with linear denumerators over a unit
triangle. Computers Methods in Applied Mechanics and Engineering, 181:191
235, 2000. 10, 28
H. T. Rathod and M. S. Karim. An explicit integration scheme based on recursion for the curved triangular finite elements. Computer and Structures, 80(1):
7393, 2001. 243, 244, 245
H. T. Rathod and M. S. Karim. An explicit integration scheme based on recursion for curved triangular finite elements. Computer structure, 80:4376, 2002.
6, 11, 42, 43, 56, 57, 67, 79, 101, 104, 145, 243, 244
337
REFERENCES
H. T. Rathod and H. S. G. Rao. Integration of polynomials over an arbitrary
tetrahedron in euclidean three dimensional space. Computers and Structures,
59(1):5565, 1996. 101, 129, 186, 191, 217
H. T. Rathod and H. S. G. Rao. Integration of trivariate polynomials over linear
polyhedra in euclidean three dimensional space. J. Austral. Math. Soc. Ser B,
39:355385, 1998. 101, 129, 186, 191, 217
H. T. Rathod, K. V. Nagaraja, B. Venkatesudu, and N. L. Ramesh. Gauss
legendre quadrature over a triangle. Journal of Indian Institute of Science, 84:
183188, 2004. 6, 11, 43, 57, 101, 129, 171, 186, 191, 203, 217, 220
H. T. Rathod, B. Venkatesudu, and K. V. Nagaraja. Gauss legendre quadrature
formulas over tetrahedron. Numerical Methods for Partial Differential Equations, 22(1):197219, 2005. 6, 11, 171, 186, 191, 203, 217, 220
H.T. Rathod, A. S. Hariprasad, K.V. Vijayakumar, Bharath Rathod, and C.S.
Nagabhushana. Numerical integration over curved domains using convex quadrangulations and gauss legendre quadrature rules. International Journal Of
Engineering And Computer Science, 2(11):32903332, 2013. 243, 253
J. W. S. Rayleigh. Theory of Sound. Dover Publisher, N. Y., 1877. 20, 30
C. T. Reddy. Improved three point integration schemes for triangular finite
elements. International journal on numerical methods and engineering, 12:1890
1896, 1978. 42, 56, 67, 100, 129
C. T. Reddy and D. J. Shippy. Alternative integration formulae for triangular
finite elements. International journal for numerical methods in engineering, 17:
133139, 1981. 42, 56, 67, 100, 129, 186, 191, 217
J. N. Reddy. An Introduction to finite Element Method. McGraw Hill, N. Y.,
1984. 20, 33
W. Ritz. Uber eine neue method zur losung gewisses variations - problem der
mathematischen physik. Journal Reine Angew. Math, 135:161, 1909. 20, 30
338
REFERENCES
P. J. Roach. Computational Fluid Mechanics. Hermosa Publishers, Albuquerque,
N.M., 1972. 19
D. F. Rogers and J. A. Adams. Mathematical Elements of Computer Graphics.
McGraw-Hill, New York, 1990. 42, 55, 67, 101
L. J. Segerlind. Applied Finite Element Analysis, 2nd Ed. John Wiley and Sons,
N. Y., 1984. 20, 27, 247
P. J. Silva and C. D. Mote. Curved composite boundary elements for second
order two dimensional problems. International Journal on Numerical Methods
for Eng., 26:12651279, 1988. 10
I. M. Smith. Programming the Finite Element Method. John Wiley and Sons, N.
Y., 1982. 20
R. V. Southwell. Relaxation Methods in Theoretical Physics. Oxford University
Press, Clarendon, 1946. 19
G. Strang and G. J. Fix. An analysis of the finite element method. Prentice-Hall,
Englewood Cliffs, N.J., 1973. 9, 42, 55, 56, 67
M. A. Taylor, B. A. Wingate, and R. E.Vincent. An algorithm for computing
fekete points in the triangle. SIAM Journal on Numerical Analysis, 38(5):1707
1720, 2000. 42, 56, 130, 144, 145, 147
H. G. Timmer and J. M. Stern. Computation of global geometric properties of
solid objects. Comput. -Aid. Des., 12:301304, 1980. 6, 11, 171, 203
L. Videla. Explicit integration of the stiffness matrix of a four node plane elastic
finite element. AIAAJ, 6:11741176, 1996. 6, 10
E. L. Wachspress. A rational finite element basis. J. Int. Math. Appl., 8:5781,
1972. 9
E. L. Wachspress. A rational finite element basis. J. Int. Math. Appl., 11:83104,
1973. 9
339
REFERENCES
E. L. Wachspress. A rational finite element basis. Academic Press, San Diego,
1975. 42, 55, 67, 101
S. Wandzura and Xiao. Symmetric quadrature rules on a triangle. Computers
and mathematics with applications, 45:18291840, 2003. 42, 56, 130, 144, 145,
147
G. Ye. Yagawa and S. Yashimara. A numerical integration schemefor finite
element method based on symbolic manipulation. Int J. Numer. Method Engng,
29:153159, 1990. 5, 10
O. C. Zienkiewicz and Y. K. Cheung. Finite elements in the solution of field
problems. The Engineers, 220:507510, 1965. 19
O. C. Zienkiewicz and R. L. Taylor. The finite element method. Maidenhead,
McGraw Hill, U. K., 1989. 20, 42, 55, 67, 101
O. C. Zienkiewicz and R. L. Taylor. The finite element method, volume Vol. 1,
5th Ed. Butterworth, Heinemann. Elsevier Science, Burlington, Mass., 2000. 6,
11, 171, 203
O.C. Zienkiewicz and K. Morgan. The finite element and Approximation. John
wiley and sons, New york, 1983. 5
M. Zlamal. Curved element in finite element method. SIAM journal of Numerical
Analysis, 11(2):347362, 1974. 9
340
List of Publications
1. Hussain Farzana and Karim, M. S. (2010), An Accurate Evaluation Procedure of Triangular Domain Integrals, SUST Studies, Vol. 12(1), 15-20.
(Included in Chapter 3)
2. Hussain Farzana and Karim, M. S. (2010), Evaluation of Triangular Domain
Integrals by Use of Gaussian Quadrature for Square domain Integrals, SUST
Studies, Vol. 12(1), 52-57. (Included in Chapter 4)
3. Hussain Farzana, Karim, M. S. and Rezwan Ahmed (2012), Appropriate
Gaussian Quadrature formulae for Triangles, International Journal of Applied Mathematics and Computation, 2012, Vol. 4(1), 24-38. (Included in
Chapter 5)
4. Hussain Farzana and Karim, M. S. (2012), Accurate Evaluation Schemes
for Triangular Domain Integrals, IOSR Journal of Mechanical and Civil
Engineering, Vol. 2(6), 38-51. (Included in Chapter 6)
5. Hussain Farzana and Karim, M. S. (2013),A Symmetrical Extended Gaussian Quadrature Formula for Evaluation of Triangular Domain Integrals,
Communicated. (Included in Chapter 7)
6. Hussain Farzana and Karim, M. S. (2013), Efficient Gauss Quadrature Rules
for Tetrahedral Domain Integrals., Communicated. (Included in Chapter 8)
7. Hussain Farzana and Karim, M. S. (2013),Extended Gauss Quadrature Rules
for Numerical Integration over a Tetrahedron, Communicated. (Included in
Chapter 9)
341