Analytical and Numerical Solution of Heat

Conduction in 2 Dimensional Slab

May 8, 2025

Figure 1: Problem geometry and boundary conditions

Solving I
∂ 2T ∂ 2T
+ =0
∂x2 ∂y 2

1
 Using separation of variables, let T = f (x)g(y):

1 d2 f 1 d2 g
= − = −λ2 (due to periodicity in x)
f dx2 g dy 2

f (x) = c1 sin λx + c2 cos λx

g(x) = c3 e−λy + c4 eλy
Boundary conditions:

1. x = 0, y, T = 0
f (0) = 0 =⇒ c2 = 0

2. x = L, y, T = 0
nπ
f (L) = 0 =⇒ c1 sin λL = 0 =⇒ λn = , n = 1, 2, . . .
L

3. y = 0, x, T = 0

g(0) = 0 =⇒ c3 + c4 = 0 =⇒ c4 = −c3

g(y) = c3 (e−λy − eλy ) = 2c3 sinh(λy)

nπx nπy
T = c1 sin · 2c3 sinh
L L
∞
X  nπx   nπy 
T = An sin sinh
n=1
L L
Boundary: y = w, θ = h(x):
∞
X  nπw   nπx 
h(x) = An sinh sin
n=1
L L

Now using orthogonality to solve

for An :
mπx
Multiply both sides by sin L and integrate:
Z L  mπx  ∞
X  nπw  Z L  nπx   mπx 
h(x) sin dx = An sinh sin sin dx
0 L n=1
L 0 L L

2
Let Z L  mπx 
Im = h(x) sin dx
0 L
∞
X  nπw  L
Im = An sinh δmn
n=1
L 2
∞
X  nπw 
2Im /L = An sinh δmn
n=1
L
 mπw 
2Im /L = Am sinh
L
2Im
Am =
L sinh mπw


L
Final solution for I:
∞
2X In  nπx   nπy 
T1 = sin sinh
L n=1 sinh nπw


L
L L

3
Solving II
∂ 2T ∂ 2T
+ = Sc
∂x2 ∂y 2

• all boundaries have homogeneous boundary condition

• only non-homogeneity in governing equation

Taking a guess for T :

XX  nπx   mπy 
T = Anm sin sin
n m
L w

 
X nπ 2  mπ 2  nπx   mπy 
⇒ − + Anm sin sin = Sc
n,m
L w L w

jπx kπy



Multiply both sides by sin L
sin w
and integrate:

X  nπ 2  mπ 2  Z L  nπx    Z w  
jπx  mπy  kπy
− + Anm sin sin dx sin sin dy
n,m
L w 0 L L 0 w w
Z L   Z w  
jπx kπy
= Sc sin dx sin dy
0 L 0 w

X  nπ 2  mπ 2  Lw (1 − (−1)j ) (1 − (−1)k )
− + Anm δnj δmk = ScLw
n,m
L w 4 jπ kπ

" 2 2 #
(1 − (−1)j )(1 − (−1)k )

jπ kπ Lw
+ Ajk = −ScLw
L w 4 jkπ 2

4Sc (1 − (−1)j )(1 − (−1)k )

Ajk = − h
2
2 i
π 2 jk jπ
+ kπ L w

4
 X 4Sc (1 − (−1)m )(1 − (−1)n )  nπx   mπy 
T2 = − h i sin sin
n,m
π 2 mn nπ 2


+ mπ

2 L w
L w

Final solution:
T = T1 + T2

∞
2X In  nπx   nπy 
T =
sin sinh
L n=1 sinh nπw L
L L
X 4Sc (1 − (−1)m )(1 − (−1)n )  nπx   mπy 
− h
i sin sin (1)
n,m
π 2 mn nπ 2


+ mπ 2 L w
L w

5
Numerical Solution Strategy
The code numerically solves the 2D Poisson equation:

∂2T ∂2T
+ = Sc
∂x2 ∂y 2
on a rectangular domain with Dirichlet boundary conditions.

Discretization
The domain is divided into a uniform grid with spacing h in both x and y
directions. The Laplacian is approximated using central differences:
Ti+1,j − 2Ti,j + Ti−1,j Ti,j+1 − 2Ti,j + Ti,j−1
+ = Sc
h2 h2
Rearranged for Ti,j :

1
Ti+1,j + Ti−1,j + Ti,j+1 + Ti,j−1 − h2 Sc


Ti,j =
4

Iterative Jacobi Method

An initial guess for T is made (usually zeros except for boundaries). The update
rule is applied to all interior points:

(k+1) 1  (k) (k) (k) (k)


Ti,j = Ti+1,j + Ti−1,j + Ti,j+1 + Ti,j−1 − h2 Sc
4
Boundary values are held fixed throughout the iterations.

Convergence
Iterations continue until the change between successive iterations is below a set
tolerance ϵ:
∥T (k+1) − T (k) ∥2 < ϵ

1
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

# Domain parameters
L, W = 1.0, 1.0
Sc = 1.0
Nx, Ny = 50, 50
dx, dy = L/(Nx-1), W/(Ny-1)

# Initialize solution grid

T = np.zeros((Nx, Ny))

# Boundary conditions
x = np.linspace(0, L, Nx)
T[:, -1] = np.sin(np.pi * x/L) # Top boundary (y=W)

# Jacobi iteration parameters

max_iter = 10000
tolerance = 1e-5
residuals = []

for k in range(max_iter):
T_new = T.copy()
for i in range(1, Nx-1):
for j in range(1, Ny-1):
T_new[i,j] = (T[i+1,j] + T[i-1,j] + T[i,j+1] + T[i,j-1] +
Sc*dx**2)/4

residual = np.linalg.norm(T_new - T)
residuals.append(residual)
T = T_new

if residual < tolerance:

break

# Plotting
X, Y = np.meshgrid(np.linspace(0, L, Nx), np.linspace(0, W, Ny))

plt.figure(figsize=(12, 5))

# 3D Surface Plot
ax1 = plt.subplot(121, projection='3d')
ax1.plot_surface(X, Y, T.T, cmap='viridis')
ax1.set_title('Temperature Distribution')
ax1.set_xlabel('X')
ax1.set_ylabel('Y')

# 2D Heatmap
plt.subplot(122)
plt.imshow(T.T, extent=[0, L, 0, W], origin='lower', cmap='hot')
plt.colorbar(label='Temperature')
plt.title('2D Heatmap')
plt.xlabel('X')
plt.ylabel('Y')

plt.tight_layout()

# Convergence Plot
plt.figure()
plt.semilogy(residuals)
plt.title('Convergence History')
plt.xlabel('Iteration')
plt.ylabel('Residual')
plt.grid(True)

plt.show()