10 X October 2022

https://doi.org/10.22214/ijraset.2022.47247
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue X Oct 2022- Available at www.ijraset.com

Numerical Approximation of Poisson Equation

Using the Finite Difference Method
Arunachalam Sundaram
Department of Mathematics, Rizvi College of Arts, Science and Commerce, University of Mumbai, Mumbai, Maharashtra, India.

Abstract: In this paper, we implement the Finite Difference Method to approximate the homogeneous form of the Poisson
equation. The Poisson equation is discretized using the central difference approximation of the second derivative and the grid
function is determined by the five point method approximates the exact solution of the Poisson equation. The finite difference
approximation is consistent and convergent. The method of solving the numerical approximation of Poisson equation is
implemented using Python Programming.
Keywords: Poisson equation, Finite Difference method, Discrete Grid, Numerical Approximation, Mesh Points, Matrix form,
Python.

I. INTRODUCTION
Partial Differential Equations arise in the study of many branches of Applied Mathematics, for instance, in the field of fluid
dynamics, heat transfer, boundary layer flow, elasticity, quantum mechanics and electromagnetic theory. The analytical method of
these equations is a rather involved process and requires applications of advanced mathematical methods. In most of the cases, it is
easier to develop approximate solutions by numerical methods. Several numerical methods have been proposed for the solution of
partial differential equations, the method of finite difference is most commonly used. In this method, the derivatives appearing in the
equation and the boundary conditions are replaced by their finite difference approximations.
H.Bennour and M.S.Said have investigated the solution of Poisson equation in n-dimensional domain with Dirichlet boundary
conditions to establish the existence, uniqueness and regularity of the solution [1]. Mohammad Mehdi Mazarei and Azim Aminataei
have introduced a new way for numerical solution of Poisson’s partial differential equation by a special combination between
logarithmic and multi-quadric radial basis function [2]. Peng Guo studied the two-dimensional Poisson equation with Dirichlet
boundary conditions. Five point difference method and Chebyshev spectral method is used to solve the corresponding two-
dimensional Poisson equation [3]. James R. Nagel has proposed the practical application of multiple dielectrics, conductive
materials and magnetostatics using the finite difference method from Poisson equation [4]. Iman Shojaei et al., have developed the
solution of a governing equation on an arbitrary domain is sought through a geometrical transformation from the rectangular domain
into the original domain using conformal mapping. They have proved that conformal mapping preserves the Laplace and Poisson
equation which are used in engineering problem [5]. Mohammad Mehdi Mazarei and Azim Aminataei have proposed that the
transformation of Poisson’s equation into the polar coordinate can achieve a better accuracy than the direct radial basis function
network method and the indirect radial basis function network method on the Cartesian coordinates [6].
Benyam Mebrate and Purnachandra Rao Koya have introduced Microsoft Office Excel worksheet implementations of numerical
methods for solving Poisson’s equation in two dimensions with Dirichlet’s boundary conditions. They have used finite difference
method and finite element method and the numerical solutions obtained by these two methods are also compared with each other
graphically in two and three dimensions [7]. Mohammad Asif Zaman has presented a comprehensive discussion on how to build a
finite difference matrix solver that can solve the Poisson equation for arbitrary geometry and boundary conditions [8]. D.J. Evans
has introduced a rhombic region to solve the Poisson equation using skew rectangular coordinates by the successive block over-
relaxation method [9]. Genet Mekonnen Assefa and Lemi Guta have showed that finite difference method for two-dimensional
Poisson equation with non-uniform mesh is not sufficiently accurate than finite difference method for two-dimensional Poisson
equation with uniform mesh size [10]. Mohammad Aslefallah and David Rostamy have presented a numerical scheme for solving
fractional Poisson equation. The method is used to find the numerical solutions of these equations based on the Grunwald estimates
for Riemann-Liouville fractional derivative [11]. This paper proposes the numerical approximation of the Poisson equation using
finite difference method. The paper is organized as follows: Section II presents the Finite Difference Method, Section III discusses
the Poisson Equation, Section IV focuses on Implementation and Results, Section V analyses on Consistency and Convergence of
the Poisson equation and finally the Conclusion is presented in Section VI.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 1513
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue X Oct 2022- Available at www.ijraset.com

II. FINITE DIFFERENCE METHOD

This is a numerical technique to solve a partial differential equation. Here we approximate second order partial derivatives using
finite differences. Consider a two-dimensional region where the function ( , ) is defined. The domain is split into regular
rectangular grids. The points of intersection of these points are called mesh points, grid points or nodal points. The approximation is
made discrete values of the independent variables and approximation is implemented through a computer program. The finite
difference method replaces all partial derivatives and other terms in the partial derivatives by means of their finite difference
approximations. After some modification, a finite difference scheme is created from which the approximate solution is obtained
[12].
To find the solution of the function ( , ) on the region, we divide the solution region into equal rectangles or meshes. We
now construct a finite difference equation to represent the given equation and boundary conditions [13]. This will help us to
calculate the values of ( , ) at the nodes of the rectangles in the region. Here we directly consider the difference equation as

1
, = , + , + , + , − ℎ , … … … (1)
4
Equation (1) shows that the value of at any point is the mean of its values at the four neighbouring points. This is called the
standard five point formula. Solution of elliptic partial differential equations is over closed regions on which boundary values
are given. The boundary values determine the solution of the partial differential equation in the interior of the region. The two
most widely used elliptic partial differential equations are Laplace equation and Poisson equation [14].

III. POISSON EQUATION

The general two dimensional Poisson’s equation is of the form

+ = ( , ), ( , ) ∈ Ω = (0,1) × (0,1) … … … (2)

with boundary conditions ( , ) = ( , ) ℎ ( , ) ∈ Ω-boundary

The region Ω = (0,1) × (0,1) is discretized into a uniform mesh Ω . In the directions into N steps giving step size
of ℎ = where

[ ] = 0 + ℎ, = 0,1,2, … … … ,
} … … … (3)
[ ] = 0 + ℎ, = 0,1,2, … … … ,
The Poisson equation is discretized using is the central difference approximation of the second derivative in the direction
1
= , −2 , + , … … … (4)
ℎ
and is the central difference approximation of the second derivative in the direction
1
= , −2 , + , … … … (5)
ℎ
This gives the Poisson difference equation,
− , + , = , , ( , ) ∈ Ω … … … (6)
, = , ,( , ) ∈ ∂Ω
where , is the numerical approximation of , .
Expanding the Poisson difference equation gives the five point method,
−[ , + , −4 , + , + , ]=ℎ , … … … (7)
for = 1,2, … … … , − 1 = 1,2, … … … , − 1.
Equation (7) can be written as ∇ , = , … … … (8)

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 1514
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue X Oct 2022- Available at www.ijraset.com

IV. IMPLEMENTATION AND RESULTS

We will implement a finite difference method to approximate the Poisson equation by using Python programming. Let us
consider the homogeneous second order form of the Poisson’s equation where ( , ) = 0.
∴ The homogeneous second order Poisson equation is + = 0 … … … (9) with boundary conditions
( , 0) = sin(2 ) , 0 ≤ ≤ 1, ,
( , 1) = sin(2 ) , 0 ≤ ≤ 1, ,
(0, ) = 2sin(2 ) , 0 ≤ ≤ 1, ,
(1, ) = 2sin(2 ) , 0 ≤ ≤ 1, ℎ.

The region Ω = (0,1) × (0,1) is discretized into a uniform mesh Ω . In the directions into = 10 steps giving step
size of
ℎ= = 0.1 where
[ ] = 0 + ℎ, = 0,1,2, … … … ,10
[ ] = 0 + ℎ, = 0,1,2, … … … ,10
The Figure 1.below shows the discrete grid points for = 10, the known boundary conditions (green colour) and the unknown
values (red colour) of the Poisson equation.

Figure 1.Discrete grid points for = 10 of the Poisson equation.

The discrete boundary conditions of Equation (9) are

= ( , 0) = sin(2 [ ]) , for i = 0,1, … … … 10, ,
= ( , N) = sin(2 [ ]) , for i = 0,1, … … … 10 , ,
= (0, j) = 2sin(2 [ ]) , for i = 0,1, … … … 10 , ,
= ( )
, = 2 sin(2 [ ]) , for i = 0,1, … … … 10 , ℎ.
The following Figure plots the boundary values of ( , )

Figure 2.Plots of the boundary values of ( , ).

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 1515
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue X Oct 2022- Available at www.ijraset.com

Equation (8) can be written as a system of ( − 1) × ( − 1) can be arranged in matrix form

= … … … (10)
where A is an ( − 1) × ( − 1) matrix made up of the following block tridiagonal form
0 0 . . .
⎡ 0 0 . .⎤
⎢ ⎥
⎢. . . . . . .⎥
⎢. . . 0 ⎥
⎣. . . . 0 ⎦
where I denote an ( − 1) × ( − 1) identity matrix and T is the tridiagonal matrix of the form
−4 1 0 0 . . .
⎡ 1 −4 1 0 0 . . ⎤
⎢ ⎥
=⎢ . . . . . . . ⎥
⎢ . . . 0 1 −4 1 ⎥
⎣ . . . . 0 1 −4⎦
The plot below shows the matrix and its inverse a colour plot.

Figure 3.Matrix and Matrix

The vector is of length ( − 1) × ( − 1) which is made up of ( − 1) subvectors of length ( − 1) of the form

⎡ ⎤
⎢ . ⎥
=⎢ . ⎥
⎢ ⎥
⎢ ⎥
⎣ ⎦
The vector is of length ( − 1) × ( − 1) which is made up of ( − 1) subvectors of the form = −ℎ − − , where
is the vector of left and right boundary conditions for = 1,2, … … … , − 1.

⎡ 0 ⎤
⎢ . ⎥
=⎢ .
⎥
⎢ ⎥
⎢ 0 ⎥
⎣ ⎦
where is the vector of the lower boundary condition for = 1,

⎡ ⎤
⎢ . ⎥
=⎢ .
⎥
⎢ ⎥
⎢ ⎥
⎣ ⎦

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 1516
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue X Oct 2022- Available at www.ijraset.com

Upper boundary condition for = −1

⎡ ⎤
⎢ . ⎥
=⎢ .
⎥
⎢ ⎥
⎢ ⎥
⎣ ⎦
0
⎡0⎤
⎢ ⎥
.
= 2,3, … … … , − 2, = 0, = −⎢ ⎥ = 1,2, … … … , − 1.
⎢.⎥
⎢0⎥
⎣0⎦
To solve the system for , invert the matrix from equation (10), = such that
= .
Since is a vector it has to be reshaped into grid form to plot. The following figure shows the numerical approximation of the
homogeneous Poisson equation.

Figure 4.Numerical approximation of the homogeneous Poisson equation.

V. CONSISTENCY AND CONVERGENCE

Consistency and Convergence of the grid function determined by the five point method approximates the exact solution of the
homogeneous Poisson equation.
Consistency: Let ∇ ( ) = −[ , + , −4 , + , + , ] … … … (11)
denote the finite difference approximation associated with the grid Ω having the mesh size ℎ to a partial differential operator

∇ = + … … … (12)
defined on a simply connected open set Ω ∈ . For a given function ∈ (Ω), the truncation error of ∇ is
( ) = [∇ − ∇ ] ( ) … … … (13)
The approximation ∇ is consistent with ∇ if lim → [ ( )] = 0, for all ∈
∈ (Ω). The approximation is consistent to order if ( ) = (ℎ ).
In other words, the method is consistent with the approximation of the Poisson equation.
Convergence: Let ∇ = ( ) be a finite difference approximation defined on a grid mesh size ℎ to a partial differential
equation ∇ ( ) = ( ) on a simply connected set
⊂ . Assume that ( , ) = ( , ) at all points ( , ) on the boundary Ω. The finite difference scheme converges, if
max − → 0 ℎ → 0.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 1517
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue X Oct 2022- Available at www.ijraset.com

VI. CONCLUSION
We have implemented a finite difference method to approximate numerically to the second order homogeneous Poisson equation.
The Poisson equation is discretized using finite difference method. By expanding the Poisson difference equation, we obtain the five
point method equation. This equation is represented in matrix form to find the numerical solution of Poisson equation. Python
programming is implemented to obtain the numerical solution of the Poisson equation with boundary conditions. Figure 1 shows the
discrete grid points for = 10 of the Poisson equation, Figure 2 shows the boundary values of the Poisson problem, Figure 3 shows
the matrix form and its inverse matrix form and Figure 4 gives the numerical approximation of the Poisson equation. The grid
function determined by the five point method approximates the exact solution of Poisson equation ensures the consistency and
convergence of the solution.

REFERENCES
[1] H. Bennour and M.S.Said, “Numerical Solution of Poisson Equation with Dirichlet Boundary Conditions”, International Journal of Open problems Compt.
Math., Volume 5, No. 4, December 2012, pp. 171 – 195.
[2] M.M.Mazarei and A.Aminataei, “Numerical Solution of Poisson Equation using a Combination of Logarithmic and Multiquadric Radial Basis Function”,
Hindawi Publishing Corporation, Journal of Applied Mathematics, Volume 2012, December 2011, pp. 01 – 13.
[3] Peng Guo, “The Numerical Solution of Poisson Equation with Dirichlet Boundary Conditions”,Journal Applied mathematics and Physics, Vol. 9, No. 12,
December 2021, pp. 3007 – 3018.
[4] James R. Nagel, “Numerical Solution of Poisson Equation using Finite Difference Method”, IEEE Antennas and Propagation Magazine, Vol. 56, No. 4, August
2014, pp. 209 – 224.
[5] Iman Shojaei et al., “ A Numerical Solution for Laplace and Poisson Equations using Geometrical Transformations and graph Products”, Applied Mathematical
Modeling, Volume 40, Issues 17-18, September 2016, pp. 7768 – 7783.
[6] A. Aminataei and M.M.Mazarei, “Numerical Solution of Poisson Equation Using Radial Basis Function Networks on the Polar Coordinates”, Computers and
Mathematics with Applications, Vol. 56, Issue 11, December 2008, pp. 2887 – 2895.
[7] Benyam Mebrate and Purnachandra Rao Koya, “Numerical Solution of a Two Dimensional Poisson Equation with Dirichlet Boundary Conditions”, American
Journal of Applied Mathematics, Volume 3, Issue 6, December 2015, pp. 297 – 304.
[8] Mohammad Asif Zaman, “Numerical Solution of Poisson Equation Using Finite Difference Matrix Operators”, Electronics, 11, 2365, July 2022, pp. 01 – 20.
[9] D.J.Evans, “The Numerical Solution of Poisson Equation in a Rhombus”,International Journal of Computer Mathematics, Vol. 42, Issue 3-4, March 2007, pp.
193 – 211.
[10] Genet Mekonnen Assefa and Lemi Guta, “Solution of Two Dimensional Poisson Equation Using Finite Difference Methods with Uniform and Non-Uniform
Mesh Size”, Advances in Physics Theories and Applications, Vol. 79, 2019, pp. 01 – 10.
[11] Mohammad Aslefallah and David Rostamy, “Numerical Solution for Poisson Fractional Equation via Finite Differences Theta Method”, Journal of
Mathematics and Computer Science, Vol. 12, 2014, pp. 132 – 142.
[12] S. S. Sastry, “Introductory Methods of Numerical Analysis” Third Edition, Prentice Hall of India, New Delhi, January 2002.
[13] M. K. Jain et al., “Numerical Methods for Scientific and Engineering computation”, New Age International Publishers, Sixth Edition, 2014.
[14] Kendall E. Atkinson, An Introduction to Numerical Analysis”, John Wiley & Sons.(1989)

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 1518