-
Notifications
You must be signed in to change notification settings - Fork 20
Closed as duplicate of#222
Closed as duplicate of#222
Copy link
Labels
bugSomething isn't workingSomething isn't working
Description
Description
A handful of NaNs were produced using the following mesh with large numbers of computational ions when ions initially at the top of the mesh hit the bottom of the first trench:
simulation_boundary_points = np.array([
[-0.1, -0.1],
[4.1, -0.1],
[4.1, 10.1],
[-0.1, 10.1],
[-0.1, -0.1]
])
vertices = np.array([
[0.0, 0.0],
[1.0, 0.0],
[2.0, 0.0],
[3.0, 0.0],
[4.0, 0.0],
[4.0, 1.0],
[3.0, 1.0],
[3.0, 10.0],
[2.0, 10.0],
[2.0, 1.0],
[1.0, 1.0],
[1.0, 10.0],
[0.0, 10.0],
[0.0, 1.0]
])
triangles = [
[13, 12, 11],
[13, 11, 10],
[0, 13, 10],
[0, 10, 1],
[1, 10, 9],
[1, 9, 2],
[9, 8, 7],
[9, 7, 6],
[2, 9, 6],
[2, 6, 3],
[3, 6, 5],
[3, 5, 4],
]
boundary = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
I suspect there is a divide by zero error in a trig function produced by an edge case of recoil sampling. The error is not present for an isotropic surface binding potential.
Metadata
Metadata
Assignees
Labels
bugSomething isn't workingSomething isn't working