Interim major project report on

SOLUTION OF POTENTIAL PROBLEMS

USING BOUNDARY ELEMENT METHOD
FACULTY OF ENGINEERING AND TECHNOLOGY
In partial fulfillment of the requirement for the degree of
BACHELOR IN CIVIL ENGINEERING

BY
Souvik Pramanik, roll – 001910401037
Manash Dey, roll- 001910401045
Sayan Banerjee, roll- 001910401059

Under the guidance of Dr. Arup Guha Niyogi

Professor, Department of Civil Engineering
Jadavpur University
Kolkata-700032
 CONTENTS

• Introduction
• Literature Review And Brief History
• Analytical Formulation Of Two-dimensional Potential Problems
• The Basic Integral Equation And Fundamental Solution
• The Boundary Element Method
• Conclusions
• References
 INTRODUCTION
What is Boundary Element Method(BEM)?

• The Boundary Element Method (BEM) is a fully equipped numerical

techniques to solve linear partial differential equations widely applied
in many engineering areas

• The Boundary Element Method tries to use the given boundary

conditions to fit the boundary values into the integral equation instead
of using the entire space defined by the partial differential equation.

• Once calculated by this way in the post processing phase the integral
equation can be used to directly compute the numerical solution
again at any desired point in the internal solution domain.
 INTRODUCTION

Advantages Disadvantages
• Unfamiliar mathematics
• Less data preparation time
• The modelling for inferior in non-linear
• High resolution of stress problems

• Less computer time and • Fully populated and unsymmetrical

solution matrix
storage
• Poor for thin structures in three
• Less unwanted information dimensional analysis
 INTRODUCTION

Finite Element Method Boundary Element Method

• Discretization of whole domain • Discretization of boundary
• Solvable in most linear second order • Maybe unsolvable for homogeneous or
PDEs linear problems
• More precise in finite domains • More precise in infinite or semi-infinite
LITERATURE REVIEW AND BRIEF HISTORY

• C. F. Gauss (1813) : Developed the Divergence Theorem

• Green (1828) : Wrote a famous essay on the application of mathematical analysis to the theories of electricity and
magnetism.
• E. I. Fredholm (1903) : Proved the existence and uniqueness of solution of the linear integral equation
• M. A. Jaswon and A. R. Ponter (1963) : First formulated 2D potential problem in terms of a direct Boundary Integral
Equation (BIE) and solved it numerically.
• F. J. Rizzo (1967) : Extended the work into the 2D elastostatic case.
• T. A. Cruse and F. J. Rizzo (1968) : Extended the work into 2D elastodynamics case.
• P. K. Banerjee and R. Butterfield (1975) : Coined the term “Boundary Element Method” in an attempt to make an
analogy with Finite Element Method (FEM).
• C. A. Brebbia (1978) : Published the first textbook on BEM, ‘The boundary Element Method for Engineers’.
ANALYTICAL FORMULATION OF
TWO-DIMENSIONAL POTENTIAL
PROBLEMS
 ANALYTICAL FORMULATION

• The differential equation representing Laplace’s equation in two dimensions:

where is the Laplacian operator, is the potential function (temperature in heat

conduction) and x and y are the Cartesian coordinate axes.
• A complementary function that satisfies the above differential equation is the
"fundamental solution".
• Now consider the problem's physical domain, where the solution is sought is depicted
arbitrarily–

A two-dimensional physical domain.

• Consider any point on the boundary Q (typically referred to as the "field" point) with coordinates xQ
and yQ. Assume that there is an interior point p (typically referred to as the "load" point) of
coordinates Xp and Yp.
• The three-dimensional solution of a concentrated potential or source at point p, which is
the basis for the fundamental solution of Laplace's equation, can be conveniently
validated as follows:
λ(p,Q) =

where r(p, Q) is the distance between p and Q, given by

r(p,Q) =

• The fundamental solution is 'singular', i.e., it becomes infinite as the distance r

approaches zero (p coincides with Q).
• A boundary element formulation's main goal is to reduce the problem's dimensionality
by one, or to change the variables from area variables to boundary ones. Use of
Green's second identity is the most straightforward strategy.
• For the purposes of this problem we assume two functions-
• = unknown potential at any given point
• = known fundamental solution of Laplace's equation
• Applying Green's second identity to these two functions results in the following
transformation from an area integral dA to a line integral dΓ:

where n is the unit outward normal and ∂/∂n is the derivative in the direction of the
outward normal.
• Now, = 0 everywhere in the solution domain, but = 0 everywhere except at the point p
itself where it is singular.
• Hence, , we can surround the point p by a very small circle of radius ε and examine the
solution in the limit at ε0. By excluding this small circle, the new area is (A - Aε) and the
new surface is (Γ+ Γε), and the previous equation becomes-

• Within the area (A - Aε), = 0, = 0 everywhere which makes LHS=0, and the RHS can
be split into two integrals as follows-
• To evaluate the , it is convenient to use the angle α measured anticlockwise from the x
axis at point p and substitute = ε dα. The following substitution can also be used:

• Taking the integral to the limit as ε0, within the limits of α from 0 to 2π, in the integral
becomes-

• Substituting this result and rearranging the terms results in the following equation:

• The above equation is called the Boundary Integral Equation (BIE) and forms
the starting point for the subsequent numerical formulation.
• K1 and K2 are called the kernels and are defined as –

• To make the BIE equation a truly 'boundary-only' one, we must move the interior load
point p to the boundary (and refer to it as P), which results in the following equation:

• The function C(P) can be calculated by surrounding the boundary point P by a small
circle of radius ε and taking each term in the above equation in the limit as ε0.
• To calculate the first kernel K1(P, Q), the function (p, Q) is differentiated with respect to
the unit normal at the point Q:

• Therefore-

K1(p,Q) =

• Here, ; and ;
• The potential kernels are functions of geometry only.
 PHYSICAL SIGNIFICANCE OF BIE

• The BIE is an integral equation that connects the boundary variables of the actual
problem that we are attempting to solve (here the kernels K1 and K2).

• The fundamental solution (which produces the kernels) only determines the "influence"
of a concentrated potential at a particular location p on any other point Q, with the
exception of the case where p coincides with Q, in which case the solution is singular.
THE BASIC INTEGRAL EQUATION AND
FUNDAMENTAL SOLUTION
Basic Integral Solution
Let’s start with the solution of a Laplace equation in Ω domain , Ω
with the following conditions on the Γ boundary
on
In principle, the error introduced into the above equations
when substituting approximate solutions for exact values ​of
u and q can be minimized by orthogonalizing them with
respect to weighted function and derivatives in the

boundary.
If R is the residuals, then,
 BASIC INTEGRAL SOLUTION
Therefore we can write ,

Let’s do Integrating by parts two times on the above equation we got,

 FUNDAMENTAL SOLUTION
The fundamental solution satisfies Laplace's equation and it represents the electrical field
generated by a concentrated unit charge acting at a point. The effect of this charge is propagated
from i to infinity without any consideration of boundary conditions. Because of this the
solution can be written

Therefore,

Let’s put this value in the equation( )

 FUNDAMENTAL SOLUTION

For an isotropic 3-D medium the solution of the above partial differential equation is

And for an isotropic 2-D medium the solution of the partial differential equation is
 BOUNDARY INTEGRAL EQUATION
Consider the equation ( ) before any boundary conditions have been applied

Let’s consider a hemisphere of very small radius (for 3-D) or a semicircle

of very small radius (for 2-D)
Let’s do the limits,

Therefore from above two equations we can write,

THE BOUNDARY ELEMENT METHOD
 THE BOUNDARY ELEMENT METHOD
• The boundary integral equation obtained by last chapter is

• If u and q are known on Γ we can compute u at any point in the solution domain Ω by
evaluating the line integral over Γ.

• However, the boundary conditions of the boundary value problems give us either u or q (not
both) at each and every point on Γ. At a given point on the boundary Γ if u is specified by the
boundary conditions, then q is unknown or vice versa.

• We will now show how the boundary integral equation for any points on a smooth part of Γ
can be used to determine approximately unknown u or q on Γ whichever is not given by the
boundary conditions.
 DISCRETIZATION OF BOUNDARY

• For the sake of simplicity now we will

assume the body is 2 dimensional and
the border is split into N segments
apart.
• Physically not realistic but we are
assuming the simplest case as each
element is considered as a constant
element which have a constant shape
function.
BOUNDARY ELEMENT SOLUTION WITH CONSTANT ELEMENT
• The BIE maybe discretized at a given point i before adding any boundary conditions since the
value of u and q are assumed to be constant across each element and equal to value at the
mid element node

• As, u and q are constant over each element so they can be taken out of the integrals. For a
certain element ‘j’, they are called uj and qj respectively.

• Points to be mentioned that for this type of constant element the boundary is always smooth
due to the position of the node at the centre of element therefore the multiplier of u is ½. We
get
 BOUNDARY ELEMENT SOLUTION WITH CONSTANT ELEMENT

• The purpose of this two integrals of previous equation are to set relation
between node ‘i’ where the fundamental solution is acting and node ‘j’. Assume,
;

• For a particular ‘i’ point one can write,

• We assume that the fundamental solution is applied at each node progressively,

one would get a system of equations by applying the equation to each boundary
point in succession.
 BOUNDARY ELEMENT SOLUTION WITH CONSTANT ELEMENT

• Now it can be written that,

• Hence the previous equation can be written as

• This equation can be expressed as a matrix as

• HU=GQ
BOUNDARY ELEMENT SOLUTION WITH CONSTANT ELEMENT

• Here H and G are two N × N matrices and U, Q are vectors of length N. Now as we
know N solutions of u and q then we have only other N unknowns in the system of
equations. One must restructure the system by moving columns of H and G from one
side to the other in order to incorporate these boundary conditions into the matrix.
When all unknowns are passed to the left-hand side one can write,

• AX=F

• Here, X is a vector of unknown u’s and q’s boundary values. F is found by multiplying
the corresponding columns by the known values of u’s or q’s.
 EVALUATION OF INTEGRALS
Integrals like Gij and Ĥij in the above
expressions can be calculated using
numerical integration formulae such as
Gauss quadrature rules for the case i ≠ j.
for the element i = j however the presence
Element Coordinate System
on that element of singularity due to the
fundamental solution requires a more
accurate integration. It is advised to utilise
higher order integration rules or a
particular formula, such as logarithmic and
other transformations, for these integrals.
 EVALUATION OF INTEGRALS
• The Ĥii and Gii integrals may be calculated analytically for the specific situation of constant
elements. Where Ĥii terms, for example, are identical since the normal n and the element co-
ordinate are always perpendicular to one another, i.e.,

• The integrals in Gii require special handling. For a two-dimensional element for instance, they
are
 CONCLUSION

• To start with BEM is one of the most important techniques to solve real life
complicated engineering problems, as it takes less time is spent on data
preparation, it provides high resolution of stress, it provides less unwanted data
etc. In this project we have briefly discussed about The Boundary Element
Method, to determine the method we also derived Analytical Formulation of
Two-Dimensional Potential Problems and The Basic Integral Equation and
Fundamental Solution.
 REFERENCES

1. Green, G., 1828, An Essay on the Application of Mathematical Analysis to the Theories of Electricity 13. Mukherjee, S., 2003, “Boundary Element Methods in Solid Mechanics - A Tribute to Frank Rizzo,” Electron. J.

and Magnetism, T. Wheelhouse, Nottingham. Boundary Elem.

2. Jaswon, M. A., 1963, “Integral Equation Methods in Potential Theory. I,” Proc. R. Soc. London. 14. Telles, J. C. F., 2003, “A Report on some Boundary Element Adventures,” Electron. J. Boundary Elem.

15. Yao, Z., and Du, Q., 2003, “Some Aspects of the BEM Research in China,” Electron. J. Boundary Elem., 1,
3. Jaswon, M. A., and Ponter, A. R., 1963, “An Integral Equation Solution of the Torsion Problem,” Proc.
pp. 61–67
R. Soc. London.
16. Rudolphi, T. J., and Liu, Y. J., 2003, “BIE=BEM - The Past, Present and Future, a Special Issue in Honor of
4. Rizzo, F. J., 1967, “An Integral Equation Approach to Boundary Value Problems of Classical
Professor Frank J. Rizzo”.
Elastostatics,” Quart. Appl. Math.
17. Cheng, A. H.-D., and Cheng, D. T., 2005, “Heritage and Early History of the Boundary Element Method,”
5. Rizzo, F. J., and Shippy, D. J., 1968, “A Formulation and Solution Procedure for the General Non-
Engineering Analysis with Eng. Anal. Boundary Elem.
Homogeneous Elastic Inclusion Problem,” Int. J. Solids Struct.
18. Tanaka, M., 1983, “Some Recent Advances in Boundary Element Methods,” Appl. Mech. Rev.
6. Cruse, T. A., and Rizzo, F. J., 1968, “A Direct Formulation and Numerical Solution of the General
19. Tanaka, M., Sladek, V., and Sladek, J., 1994, “Regularization Techniques Applied to Boundary Element
Transient Elastodynamic Problem - I,” J. Math. Anal. Appl.
Methods”.
7. Banerjee, P. K., and Butterfield, R., 1976, “Boundary Element Methods in Geomechanics,” Finite
20. Mukherjee, S., 1982, Boundary Element Methods in Creep and Fracture, Applied Science Publishers, New
Elements in Geomechanics, G. Gudehus, ed., John Wiley and Sons (U.K.), London.
York.

8. Brebbia, C. A., 1978, The Boundary Element Method for Engineers, Pentech Press, London.
21. Becker, A. A., 1992, “The Boundary Element Method in Engineering: A complete course”, McGraw Hill Book

9. Rizzo, F. J., 2003, “Springs, Formulas and Flatland: A Path to Boundary Integral Methods in Elasticity,” Company, New York.

Electron. J. Boundary Elem. 22. Brebbia, C. A. and Dominguez, J.,1992, “Boundary Elements: An Introductory Course”, WIT Press, New
Forest, England.
10. Cruse, T. A., 2003, “Boundary Integral Equations – A Personal View,” Electron. J. Boundary Elem.
23. Ang, W. T., 2007, “A Beginner’s Course in Boundary Element Methods”, Universal Publishers Boca Raton,
11. Watson, J. O., 2003, “Boundary Elements from 1960 to the Present Day,” Electron. J. Boundary Elem.
Florida

12. Shippy, D. J., 2003, “Early Development of the BEM at the University of Kentucky,” Electron. J.
24. Antes, H., 2010, “A Short Course on Boundary Element Methods”, Engineering school, Technische Universität
Boundary Elem.
Braunschweig, Germany.
THANK YOU❕