Finite Difference Solution: Two-Dimensional Laplace and Poisson Equations

Definition

The Laplace and Poisson equations are fundamental partial differential equations (PDEs) in physics and engineering. The Laplace equation ($\nabla^2\phi = 0$) describes steady-state potentials (like heat distribution or electrostatic potential) in regions without sources. The Poisson equation ($\nabla^2\phi = f(x,y)$) extends this by including a source term $f(x,y)$. The finite difference method (FDM) is a numerical technique that approximates these continuous PDEs by replacing derivatives with algebraic difference equations at discrete grid points.

Main Content

1. Discretization of the Domain

• The physical domain is divided into a rectangular grid with spacing $\Delta x = h$ and $\Delta y = k$.
• Any point in the grid is denoted as $(x_i, y_j)$, where the function value is $u_{i,j}$.

2. The Five-Point Stencil Approximation

• Using Taylor series expansion, the second-order partial derivatives are approximated: $\frac{\partial^2 u}{\partial x^2} \approx \frac{u_{i+1,j} - 2u_{i,j} + u_{i-1,j}}{h^2}$ $\frac{\partial^2 u}{\partial y^2} \approx \frac{u_{i,j+1} - 2u_{i,j} + u_{i,j-1}}{k^2}$
• For a square grid where $h=k$, the Laplace equation simplifies to the "Average Property": $u_{i,j} = \frac{1}{4}(u_{i+1,j} + u_{i-1,j} + u_{i,j+1} + u_{i,j-1})$.

3. Poisson Equation Modification

• While Laplace assumes the sum of neighbors is 4 times the center, the Poisson equation accounts for the source term $f(x,y)$.
• The stencil becomes: $u_{i+1,j} + u_{i-1,j} + u_{i,j+1} + u_{i,j-1} - 4u_{i,j} = h^2 f(x_i, y_j)$.

Working / Process

1. Grid Generation and Boundary Conditions

• Define the region and divide it into a mesh of nodes.
• Assign known values to the boundaries (Dirichlet conditions) or specify flux (Neumann conditions).

  y |
    |   o---o---o (Boundary)
    |   |   |   |
    |   o---o---o (Interior node)
    |   |   |   |
    |   o---o---o (Boundary)
    +------------------ x

Visual representation of a 3x3 grid stencil for calculation.

2. Setting Up Algebraic Equations

• For every interior grid point, write the finite difference approximation.
• If you have an $N \times N$ interior grid, you will generate a system of $N^2$ linear equations.

3. Solving the Linear System

• Represent the system in matrix form $Ax = B$.
• Use iterative methods like Jacobi, Gauss-Seidel, or Successive Over-Relaxation (SOR) to find the values of $u_{i,j}$ until the change between iterations is below a predefined tolerance.

Advantages / Applications

• Computational Efficiency: FDM is relatively easy to implement on structured rectangular grids for complex engineering problems.
• Heat Transfer: Used to predict steady-state temperature distribution in metal plates or engine blocks.
• Electrostatics: Used to calculate the electric potential distribution in regions with specified charges.
• Fluid Dynamics: Helps in solving for pressure fields in incompressible flow problems.

Summary

The finite difference method transforms continuous Laplace and Poisson PDEs into a discrete system of algebraic equations by approximating derivatives with neighbor-point averages. By applying boundary conditions to a structured grid and solving the resulting matrix equation, we obtain an accurate approximation of the potential field. Key terms to remember are: Discretization, Stencil, Dirichlet Boundary Conditions, and Iterative Solvers.