I'm currently solving the problem of an impulse pressure wave being introduced into a pipe (exhaust) and measuring the reflections, expansions and contractions of said pulse. The pulse is being introduced with a field function. It is being solved with an explicit unsteady, coupled flow and coupled energy solver. When ran, the model residuals diverge immediatly, the energy equation diverges to ~1E09, the Z-momentum to roughly 1E06 and the rest to ~1E03. I've change the Courant number, the residual soothing values but... nothing.
The pulse and reflections are visually correct. I can see the compression wave moving down the pipe and the rarefaction expansion waves generated at changes, etc. Yet it still diverges.
It sounds like it would keep on running. I mean, the residuals only high but it doesn't give an error message, is that right?
I wish, I would get one Euro for every time somebody complains about divergence in CCM+. I could stop working, buy a boat and my own island and go fishing for the rest of my live...
The residuals are normalized, therefore you shouldn't give too much on the level of residuals as it strongly depends on your initialization. Search the forum, this topic has been discussed at least 500 Million times.
PS. When my assumption in the first line should be incorrect, I apologize for that crude answer. In that case let me know and I will look if I've got a better answer.
|All times are GMT -4. The time now is 22:16.|