Discussion Forum

Hi

the time field in the transient solver accepts a set "{}" that can be a space or comma separated list of any type, a single range(,,) or multiple chained "range(,,) range(,,) ..." or a mixture, or even 10^{range(,,)}.

Only conditions, the values should be monotonously increasing (or decreasing), none two equal values, nor any inversion

--
Good luck
Ivar

Hi

the time field in the transient solver accepts a set "{}" that can be a space or comma separated list of any type, a single range(,,) or multiple chained "range(,,) range(,,) ..." or a mixture, or even 10^{range(,,)}.

Only conditions, the values should be monotonously increasing (or decreasing), none two equal values, nor any inversion

--
Good luck
Ivar

Thank you, Ivar, but I mean the time step of computation (which is a similar parameter of grid spacing), not the time edit field. I think these two time steps are different.

Hi

that depends on your Solver Time Stepping settings, "Automatic" (the default) means COMSOL governs and might decide to take different time steps than those you specify, and will generate the results at your specified values via an interpolation of its solver step values.

If you use "Strict" it will only calculate your define time steps, with "Intermediate" it chooses a point in-between your time steps. The two latter are typically used for periodic responses, while "Automatic" is well suited for asymptomatic solutions, such as those from diffusion equation based analysis

--
Good luck
Ivar

Hi

that depends on your Solver Time Stepping settings, "Automatic" (the default) means COMSOL governs and might decide to take different time steps than those you specify, and will generate the results at your specified values via an interpolation of its solver step values.

If you use "Strict" it will only calculate your define time steps, with "Intermediate" it chooses a point in-between your time steps. The two latter are typically used for periodic responses, while "Automatic" is well suited for asymptomatic solutions, such as those from diffusion equation based analysis

--
Good luck
Ivar

I think if I make an example, my question will be much clear. If I input several time ranges in the Time edit field, like range( , , ), range( , , ), --- and in the time stepping settings, the Method I chose BDF, the Steps taken by solver I chose Free, the Initial step box was unchecked, the Maximum step box was checked and need to input a value. I want the different time ranges in the Time edit field have different Maximum steps. Can comsol do this?

Hi

Indeed it's somewhat confusing, with all these settings that also changes with the physics chosen ;)

But I believe I'm not to far off:

If you leave BDF "Free", COMSOL decides, + you might restrict, globally, the initial step and max step. (but once, not per region, without chaining several times series solver sequences). And there is another setting to decide if you store only the values you have given or all values used by the Solver

If you decide for BDF "Strict", you have full control with your range values (here you might chain ranges), which are the ones solved and stored at output

Do we agree ?

So I suppose it's rather no for your question of multiple max steps in "free mode".

But as I said too, if you chain several time solvers in sequence, using the last step of the previous to start the first step of the next solver sequence you might get to your point. But you must store and restart the initial conditions, and there something is not working for me, I'm expecting a reply from support, one of these days. If you do not store, you loose the values of all but last time series

--
Good luck
Ivar

Hello Ivar,

You mentioned that the result solved at the 'Time Steps Taken by Solver' can be recorded. However, when I do a parametric sweep, I can only see the result of 'Time Steps Taken by Solver' on the last parametric value.

Is there a way to save all the results of the 'Time Steps Taken by Solver' on all the parametric value?

Thank you in advance.
Yu Hui