Opened 4 years ago
Last modified 3 years ago
#6422 new enhancement
Performance problems with Power Electronics
| Reported by: | Owned by: | Andreas Heuermann | |
|---|---|---|---|
| Priority: | high | Milestone: | |
| Component: | Run-time | Version: | 1.16.5 |
| Keywords: | Cc: | Christian Kral |
Description
OM 1.16.5 on Windows 10 64bit
Please consider the enclosed model:
It's a bidirectioanl DC/DC converter for coupling a 13 V battery with an 48 V battery in automotive systems.
As a reference, in Dymola simulation takes about 2 minutes (I know, simulation of switching power electronics can't be really performant). OM reports "The initialization finished successfully without homotopy method.", then it seems to get stuck at 3% for a long time without any feedback.
You may add that model to your test cases.
Attachments (1)
Change History (5)
by , 4 years ago
| Attachment: | BiChopper.mo added |
|---|
comment:1 by , 4 years ago
| Component: | Backend → Run-time |
|---|---|
| Milestone: | Future → 1.18.0 |
| Owner: | changed from to |
comment:2 by , 4 years ago
comment:3 by , 4 years ago
You may check PR https://github.com/modelica/ModelicaStandardLibrary/pull/3774.
Although simulating a shorter time span, OM hangs.
Note:
See TracTickets
for help on using tickets.
I managed to get through the simulation using euler instead of dassl. Profiling reveals that most of the time is spent solving system 191, and some logging reveals the nonlinear solver does a large number of iterations on it. I also tried kinsol, same result.
I guess you should try to figure out what is going wrong with the solver in that case.