Hi. With NeuralPDE i have this conditions.
L=10
bcs0 = [G(x,y,0)~0, G(-L,y,z)~G(L,y,z), G(x,-L,z)~G(x,L,z)]
I get loss like
0 loss: 67.44680398772921 pde: [0.8533529399741632, 1.3552322555600897, 20.41989735134126] bcs: [44.8183214408537, 0.0, 0.0]
20 loss: 3.9507670556189387 pde: [0.10522889646263646, 0.3102654926375333, 0.6664227288652805] bcs: [2.8688499376534886, 0.0, 0.0]
40 loss: 1.2851344326840106 pde: [0.09905081249455622, 0.10090247334471997, 0.12011270767184003] bcs: [0.9650684391728944, 0.0, 0.0]
60 loss: 0.20369897412886434 pde: [0.009124741276919066, 0.009213327251510667, 0.05609956005273404] bcs: [0.12926134554770058, 0.0, 0.0]
80 loss: 0.06461636009430785 pde: [0.0026174909670680026, 0.003940402902969963, 0.0404453620141796] bcs: [0.017613104210090283, 0.0, 0.0]
100 loss: 0.03800169007950788 pde: [0.0022196631450851086, 0.003278956546181802, 0.024756007897911575] bcs: [0.007747062490329393, 0.0, 0.0]
120 loss: 0.03319021162891614 pde: [0.0012682617417848698, 0.002768210254526406, 0.01837854348364282] bcs: [0.010775196148962038, 0.0, 0.0]
140 loss: 0.18376030447113537 pde: [0.01034530027230838, 0.028421757495622733, 0.038018971224707536] bcs: [0.10697427547849672, 0.0, 0.0]
160 loss: 0.07395316386033662 pde: [0.0031282077964498375, 0.0059535830476358355, 0.016372415561220908] bcs: [0.04849895745503005, 0.0, 0.0]
180 loss: 0.034562800263553906 pde: [0.0013684440550966322, 0.0023915343421130626, 0.01522692624185329] bcs: [0.015575895624490927, 0.0, 0.0]
200 loss: 0.026810932794939613 pde: [0.0010110818569102188, 0.0017320878888202173, 0.013766375266484308] bcs: [0.01030138778272487, 0.0, 0.0]
The first bcs condition works fine, but the 2 periodic ones return 0 loss, and the result don’t look right.