Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Solving problem
Posted Jan 12, 2011, 2:30 p.m. EST 1 Reply
Please login with a confirmed email address before reporting spam
Hi,
Comsol can slove my model for time 5 seconds, but can not solve for 10 s. At onether time, Comsol solves the same model for time 10s, but does not solve for 15s or 20s. I can not understand why comsol does this? I have tried different mesh denses and mesh types. However, the best mesh type is default type (triangular advancing front). I see that the convergence is always increasing when I increase the time period to solve. It (convergence) starts to decrease first, everything goes good. But after some time, the convergence increases continuously and this implies that soln. will not be reached and an error message will come. The error may sometimes becomes "out of memory LU factorization", sometimes singularity or sometimes "last step size is not converged". Some BCs are function of dependent variables of physics. I checked my BCs and ICs again and again but I could not find the problem.
Comsol can slove my model for time 5 seconds, but can not solve for 10 s. At onether time, Comsol solves the same model for time 10s, but does not solve for 15s or 20s. I can not understand why comsol does this? I have tried different mesh denses and mesh types. However, the best mesh type is default type (triangular advancing front). I see that the convergence is always increasing when I increase the time period to solve. It (convergence) starts to decrease first, everything goes good. But after some time, the convergence increases continuously and this implies that soln. will not be reached and an error message will come. The error may sometimes becomes "out of memory LU factorization", sometimes singularity or sometimes "last step size is not converged". Some BCs are function of dependent variables of physics. I checked my BCs and ICs again and again but I could not find the problem.
1 Reply Last Post Jan 13, 2011, 4:31 a.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
from my experience that often is related to poorly defined initial values.If you have a reasonable reply once, look at the dependetn variables and see if there is not a simple way to define them, rather than using a "0" by default.
i.e. in CFD you define a first flow velocity , as a parabola for a pipe, 0 at the no-slipwalls, with perhaps a first pressure drop. In ACDC you can setup the voltage drop as a Vin*y or V*(1-(y-Y0)/Length) or whatwever
--
Good luck
Ivar
from my experience that often is related to poorly defined initial values.If you have a reasonable reply once, look at the dependetn variables and see if there is not a simple way to define them, rather than using a "0" by default.
i.e. in CFD you define a first flow velocity , as a parabola for a pipe, 0 at the no-slipwalls, with perhaps a first pressure drop. In ACDC you can setup the voltage drop as a Vin*y or V*(1-(y-Y0)/Length) or whatwever
--
Good luck
Ivar