Discussion Forum

Time dependent and non linear model takes (too?) long to solve...

Thread index  |  Previous thread  |  Next thread  |  Start a new discussion

RSS FeedRSS feed   |   Email notificationsTurn on email notifications   |   2 Replies   Last post: January 30, 2012 6:30am UTC
Norbert Bartscher

Norbert Bartscher

January 29, 2012 10:52pm UTC

Time dependent and non linear model takes (too?) long to solve...

Hello everyone,

I am simulating electrical machines within the rotating machinery interface in the ACDC module.

The model I calculate right now solves. But the way it solves is not satisfying in my oppinion. As soon as I choose the iron to be nonlinear, the solving takes days to calculate just 50ms with a time step of 100us.

I'm not sure if I'm allowed to upload the model in this forum, since the Geometry is not meant to be public.


The Geometry is a fourth of the motor and periodic boundary conditions are used for this. The current, stamped into the slots, is calculated by time dependent equations which involve the derivation of the flux field linkage. So comsol has to solve the equation for the current and vector potential simultaneously.


I already tried to improve the model this way:
http://www.comsol.com/support/knowledgebase/1127/
But it didn't really increase the speed. Actually with it's help the model could first be solved nonlinear...



Does anybody have general advices to get nonlinear time dependent models solved faster?


Thanks in advance.






Reply  |  Reply with Quote  |  Send private message  |  Report Abuse

Norbert Bartscher

Norbert Bartscher

January 29, 2012 10:55pm UTC in response to Norbert Bartscher

Re: Time dependent and non linear model takes (too?) long to solve...

By the way:

The mesh is only within the airgap very fine. All other domains use standard element size.

Reply  |  Reply with Quote  |  Send private message  |  Report Abuse

Ivar Kjelberg

Ivar Kjelberg
Moderator

January 30, 2012 6:30am UTC in response to Norbert Bartscher

Re: Time dependent and non linear model takes (too?) long to solve...

Hi

have you also checked
http://www.comsol.eu/support/knowledgebase/1118/
http://www.comsol.eu/support/knowledgebase/1062/
http://www.comsol.eu/support/knowledgebase/905/

--
Good luck
Ivar

Reply  |  Reply with Quote  |  Send private message  |  Report Abuse


Rules and guidelines