Hello everybody
Im working on plasma, whit helium gas and time-dependent study.
its make this error:Failed to find consistent initial values.
Last time step is not converged.
can anybody help me, how can I solve this problem?
thank alot
It may be that your initial guess for a particular value may be too far from the correct value, so as COMSOL iterates, the error becomes boundless. Sometimes it is worth doing some hand calculations in order to identify a good initial guess
This is discussed in this blog: https://www.comsol.com/support/knowledgebase/1172/
I recommend contacting the techinal support team @COMSOL Multiphysics to help answer any COMSOL related questions http://www.comsol.com/support. Their team of experts will guide you through the modeling process.
Salam
Moshkele shoma dorost shod?
Man ham be eror zir barkhord kardam:
The following feature has encountered a problem:
Failed to find consistent initial values.
Out of memory during assembly.
- Feature: Automatic Remeshing (sol2/t1/arDef)
In my model coordinate of center PML was different to center of sphere of domain with PLM.
I am working drift-diffusion PDE time-dependent study. getting an error Failed to find consistent initial values.Last time step is not converged.
Well if you don't want best results at t=0s or nothing important is happening at initial time apart from initial conditions (for example like flow isn't turning to turbulent or fracture isn't initiating right after 0s). You can try turning off the "consistent initialization". This also helps if solver takes too much time to go beyond 0s when you know at t=0 everything should be zero or initial defined values (BCs).
Other way is to solve everything first using stationary study and use those as the initial conditions for time dependent study (find knowledgeable of COMSOL regarding this issue, there are fine examples too).
Also make sure that your absolute tolerance is turned from factor to manual and use some value (default is 0.001). This happened to me many times, not sure why factor don't work.
- Badges
- Science topic