Dear Comsol users,
I want to use for Comsol for time-depended simulation. And the time-dynamics of the systems, that I expect, is, that there should be first an initial phase with a slow increase of the variables I solve for. Then there should be a fast exponential increase of some variables and afterwards there should be again a slow convergence to a steady state. Because of this I want to use a big time step in the first and in the last stage of the simulation and a smaller one in the middle phase. In order to do this, I defined 3 time-dependent study-steps. For all steps I use the generalized-alpha solver and define different maximal time step sizes in the three solvers.
My problem now is, that the first solver finishes normally with the expect results. But when the seconds solver starts, I get some warnings (increase of the MUMPS allocation factor). But even tough the second time dependent solver also starts to run, it doesn't use the results of the first one as initial conditions. (The fields plotted at the end-time of the first solver, which is the start time of the second, look completely different when plot with the output from solver 1 or from solver 2, respectively).
So I would be very grateful for any hint about how to connect the two solvers properly.
Thanks a lot in advance,
Felix
I want to use for Comsol for time-depended simulation. And the time-dynamics of the systems, that I expect, is, that there should be first an initial phase with a slow increase of the variables I solve for. Then there should be a fast exponential increase of some variables and afterwards there should be again a slow convergence to a steady state. Because of this I want to use a big time step in the first and in the last stage of the simulation and a smaller one in the middle phase. In order to do this, I defined 3 time-dependent study-steps. For all steps I use the generalized-alpha solver and define different maximal time step sizes in the three solvers.
My problem now is, that the first solver finishes normally with the expect results. But when the seconds solver starts, I get some warnings (increase of the MUMPS allocation factor). But even tough the second time dependent solver also starts to run, it doesn't use the results of the first one as initial conditions. (The fields plotted at the end-time of the first solver, which is the start time of the second, look completely different when plot with the output from solver 1 or from solver 2, respectively).
So I would be very grateful for any hint about how to connect the two solvers properly.
Thanks a lot in advance,
Felix