Trouble with Electric Potential Boundaries (Computational Physics)

Click For Summary
The discussion revolves around a Python implementation for simulating electric potential using the Jacobi method. The user is facing challenges with maintaining a constant voltage at the center block while updating potential values outside of it. Attempts to treat the block as a boundary have led to incorrect results, prompting the user to explore various approaches, including separating the grid into regions. The user is seeking effective strategies to handle boundary conditions without compromising the integrity of the simulation. Solutions to this boundary problem are essential for achieving accurate results in the electric potential model.
CrosisBH
Messages
27
Reaction score
4
Homework Statement
Recreate the given figure. A block (I don't know what width I'm eyeballing it) of constant potential 1 is centered at the origin, use the Jacobi method to find the potential around the block (and eventually the E-field)
Relevant Equations
$$\Delta V_{end} \leq 10^{-5}$$, meaning the program stops when this is satisfied
$$V_{n+1}(i,j) = \frac{1}{4} [V_n(i-1,j)+V_n(i+1,j)+V_n(i,j-1)+V_n(i,j+1)] $$
DOqqrBR.png


This is in python:
Python:
#ELECTRIC POTENTIAL

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

dx = 0.1
dy = 0.1

xrange=np.arange(-1,1,dx)
yrange=np.arange(-1,1,dy)

X,Y = np.meshgrid(xrange, yrange)

max_dV = 10e-5

blockRadius = 3

V = np.zeros((int(2/dx),int(2/dy)))
for i in range(-blockRadius,blockRadius):
    for j in range(-blockRadius,blockRadius):
        V[i+int(len(V)/2)][j+int(len(V[0])/2)] = 1
        
#setting up the boundries for guessing
for i in range(0,len(V)):
    for j in range(0,len(V[0])):
        if(i==0):
            V[i][j] = 1
        if(i==len(V)-1):
            V[i][j] = 1
        if(j==0):
            V[i][j] = 1
        if(j==len(V[0])-1):
            V[i][j] = 1
            
#Step function
def update_V(iV):
    dV=0
    newV=np.zeros((int(2/dx),int(2/dy)))
    
    midX = int(len(iV)/2)
    midY = int(len(iV[0])/2)
    
    for i in range(1,len(iV)-1):
        for j in range(1,len(iV[0])-1):
            newV[i][j]= (iV[i-1][j]+iV[i+1][j]+iV[i][j-1]+iV[i][j+1])*0.25
            dV += abs(iV[i][j] - newV[i][j])
    return newV, dV#initial plot so i can see what the hell is going on
fig = plt.figure(figsize=plt.figaspect(2.))

ax = fig.add_subplot(2, 1, 1)
ax.plot([-1,1],[0,0], color = 'black')
ax.plot([0,0],[-1,1], color = 'black')
ax.contour(X,Y,V)
ax.set_title("Equipotential Lines (Before Jacobi Method)")
ax.set_xlim(-1,1)
ax.set_ylim(-1,1)

dV = 1
i=0

while(dV>max_dV):
    i+=1
    V, dV = update_V(V)

        
print("Took ", i, " steps")

fig = plt.figure(figsize=plt.figaspect(2.))
ax = fig.add_subplot(2, 1, 1)
ax.plot([-1,1],[0,0], color = 'black')
ax.plot([0,0],[-1,1], color = 'black')
ax.contour(X,Y,V)
ax.set_title("Equipotential Lines")
ax.set_xlim(-1,1)
ax.set_ylim(-1,1)ax = fig.add_subplot(2, 1, 2, projection='3d')
ax.set_title("Perspective Plot")

surf = ax.plot_surface(X, Y, V, cmap=cm.coolwarm,
                       linewidth=4, antialiased=False, rstride=1, cstride=1,)

plt.xlim(-1,1)
plt.ylim(-1,1)

plt.show()

Capture.PNG


(Obviously there's a center issue I'm working on at the moment). This is the plot I generate. I know what the problem is, but I don't know how to handle it. The problem states that the block is a at a constant voltage, so it doesn't dissipate like this, so the center needs to stay the same. I tried treating the block as a boundary so it doesn't mess with the voltage in there, but only affects the voltage outside the block. My first try was checking if it would be updating in the region in the box by:

if(not i in range(int(len(iV)/2)-blockRadius, int(len(iV)/2)+blockRadius)):
if (not j in range(int(len(iV[0])/2)-blockRadius,int(len(iV[0])/2)+blockRadius)):

before doing the calculation in update_V. This produced:
Capture.PNG

Interesting but not correct

I tried another (bad) approach. Where I separated it into 4 regions around the box and used 4 for loops. This produced this beautiful hot mess:
unknown.png


This damn box is going to be the end of me. What would be a good way to approach this boundary problem?
 
Physics news on Phys.org
I can’t say what has gone wrong, but perhaps it’s more in the nature of a programming error than a physics problem. If you set the nodes that are on the surface of the block to V and then never update them, I see no reason this shouldn’t work.
 

Similar threads

  • · Replies 4 ·
Replies
4
Views
2K
Replies
19
Views
3K
  • · Replies 6 ·
Replies
6
Views
2K
Replies
4
Views
614
  • · Replies 8 ·
Replies
8
Views
2K
Replies
7
Views
2K
  • · Replies 7 ·
Replies
7
Views
4K
  • · Replies 4 ·
Replies
4
Views
968
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K