Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
problem with parametric sweep (Comsol4.2 64bit)
Posted 2011年9月23日 GMT-4 02:46 Version 4.2 2 Replies
Please login with a confirmed email address before reporting spam
Hello!
When I use the parametric sweep function, it seems the calculation does not just take the input values but much more. For instance, when I put 0.1 0.2 0.5 1 for one parameter. From the log, I saw the programme takes 0.1 0.2 and then 0.25 0.2528 etc. so the calculation time is much longer than I expect. I saw other posts on the forum concerning this problem also.
I am wondering is this a bug of COMSOL or some settings are wrong with my COMSOL 4.2 64bit.
Anybody can help me?
When I use the parametric sweep function, it seems the calculation does not just take the input values but much more. For instance, when I put 0.1 0.2 0.5 1 for one parameter. From the log, I saw the programme takes 0.1 0.2 and then 0.25 0.2528 etc. so the calculation time is much longer than I expect. I saw other posts on the forum concerning this problem also.
I am wondering is this a bug of COMSOL or some settings are wrong with my COMSOL 4.2 64bit.
Anybody can help me?
2 Replies Last Post 2011年9月23日 GMT-4 10:13
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
It's not a bug. What probably happened in your case is that the solution didn't converge at the 0.5 value, so the software tried taking a smaller interval. You can probably find some more information in the documentation, and you can adjust the settings by expanding the solver tab.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
this is a feature of the solver nad the "stepping algorythm, it assumes by default a continuous parameter and when the convergence is bad and it has poour values for the derivatives (jacobian) then COMSOl tends to add extra points in between, otherwise it cannot converge. But normally it will only "save" the values you have specified (or values interpolated to your values from COMSOL results
By tweaking the solver parameters you might change somewhat this behaviour, but probably not to only use your points. It could mean too that your initial conditions are far to far from the final results.
If you use a "parametric" sweep, COMSOl restarts at the original initial conditions each time, if you use a solver "continuation" parametric sweep then COMSOL will use last Parameter values as starting point to the next simulation (with next parameter) this might be your issue too ;)
There are many subtilities on the buttons under the hood of comosol ;)
--
Good luck
Ivar
this is a feature of the solver nad the "stepping algorythm, it assumes by default a continuous parameter and when the convergence is bad and it has poour values for the derivatives (jacobian) then COMSOl tends to add extra points in between, otherwise it cannot converge. But normally it will only "save" the values you have specified (or values interpolated to your values from COMSOL results
By tweaking the solver parameters you might change somewhat this behaviour, but probably not to only use your points. It could mean too that your initial conditions are far to far from the final results.
If you use a "parametric" sweep, COMSOl restarts at the original initial conditions each time, if you use a solver "continuation" parametric sweep then COMSOL will use last Parameter values as starting point to the next simulation (with next parameter) this might be your issue too ;)
There are many subtilities on the buttons under the hood of comosol ;)
--
Good luck
Ivar