Message Boards Message Boards

[✓] Solve a NDSolveValue-problem?

GROUPS:

Goodmorning everyone, I have a small problem. tomorrow morning I have the exam at the university to which I have to present the exercise and I just can not figure out where the problem is. I'm trying to use NDSolveValue only that does not recognize the domain whwre to "work". Can anyone help me...I can not figure out where the problem may be I attach the notebook with the exercise. thank you all

Attachments:
POSTED BY: BARBU DENISIA
Answer
15 days ago

It seems likely that mesh was not defined when you executed NDSolveValue. That is what the error message shows. Since mesh is defined in another cell, it somehow became undefined. If it was not an action on your part (even an accidental one), then perhaps the kernel crashed. It possibly could be from running out of memory. NDSolve is taking up several gigabytes right now, and I might not let it run to completion.

POSTED BY: Michael Rogers
Answer
15 days ago

Thank Michael for your reply...i executed one cell at a time or evaluated the whole notebook but the same eror or nothing happening...at this point I think the problem is my computer...I hope that tomorrow turning it on the teacher's computer will work...i really hope.. thank you again

POSTED BY: BARBU DENISIA
Answer
15 days ago

Did this for tfin/50 after having kicked out some typos (missing semicolons, op == 0 twice written, double time derivation showed somehow up as Transpose[]), it finally run as seen in the picture, giving a sea of zeroes ...

enter image description here

each second it took around 10 MB of RAM, here there are only 8 GB RAM, so I had to shorten tfin. Possibly you find now an error in my editing which caused the solution staying trivial ... good luck!

Attachments:
POSTED BY: Dent de Lion
Answer
15 days ago

I did similar to @DentdeLion. A computation using {t, 0, .001*tfin} took a few seconds and brought me to MaxMemoryUsed[] of 434 Mb. I then upped it to .01*tfin.

In[690]:= AbsoluteTiming[
 uifwave = 
  NDSolveValue[{op == 0, dirc, ic, dic}, \[Eta][t, x, y], {t, 
    0, .01*tfin}, {x, y} \[Element] mesh]]

Out[690]= {65.8217, 
 InterpolatingFunction[{{0., 53.5719}, {0., 70.}, {0., 94.5}}, <>][t, 
  x, y]}

In[691]:= MaxMemoryUsed[]

Out[691]= 1084091256 

Extrapolating from that suggests the full computation really takes more memory than most Mathematica kernels will have available.

POSTED BY: Daniel Lichtblau
Answer
15 days ago

Group Abstract Group Abstract