Crash problem about frequency response function

PLECS
1

danxiang.plecs (1023.4 KB)

when I start to run this for minutes, the PLECS tends to crash, I want to know the reason

image
image1140×768 192 KB

2

Thank you for posting your model. Running the model as-is mostly worked for me, noting there are a few underlying issues in the model that jump out:

  1. The simulations at several frequencies did not complete due to error messages generated in the model. In particular several simulations failed due to requiring a stiff solver while the model you shared has DOPRI configured. Solver type changed to auto for automatic detection of the appropriate solver.
  2. Some simulation runs failed due to zero-crossing errors. The components reporting the error were observation instruments (single-phase PLLs) and so I disabled them.
  3. The initialization time span is 4 seconds, but at 4 seconds you also have a change in your model (breaker closing) so it’s not in a quasi-linear operating state. I changed the Initialization time span to 5 seconds, assuming that’s the last “scheduled” event in your model.
  4. Your model includes C-Scripts with internal states such as counters, which aren’t saved as part of the . When the system state is restored at the start of each frequency injection, these states will not be correct. I increased the settling time span of the frequency analysis to 0.1 seconds from 1e-3 seconds as one approach to address this.

Even with these above changes the results aren’t exactly meaningful, but that likely has to do with the model implementation and test-case construction.

Regarding the actual crash, What version of PLECS are you using including the minor version (click File + About PLECS and report the three digit string)? I would suggest using the latest bugfix release.

Your model has a huge number of scopes in it which will consume a large amount of memory and may contribute to the issue. Try disabling unused and unecessary scopes. Since the frequency response doesn’t require any scopes at all, you can actually disable all scopes in your model (but it might be nice to have a few just to see the key waveforms). Another issue could be with multi-threading. If you go into File + PLECS Preferences + General there is a Thread Limit setting. Does PLECS still have issues if that is set to 1 or a smaller number than the default. This will cause the simulation to complete slower as you will not run as many simulations in parallel.

3

Thank you expert， the version is 5.0.2, I have changed to auto, display PLL, change to 5 seconds and 0.1s, while the crash still happens, would u send me your revised version that work? Or I wonder if I need to upload my updated model?

4

The model I have been using is attached. Please also upgrade to the latest PLECS 5.0.3 bugfix release.

danxiang_BL.plecs (992.8 KB)

You’re welcome to upload an updated model. Your model is quite busy, so perhaps there is a changed setting somewhere that might induce the issue.

When PLECS Crashes does it generate a crash response window? Does the crash happen under other conditions? Does the crash occur if you change the number of points for the analysis to something smaller (e.g. from 100 to 10)?

5

thank u，though the results of your model still seems not have the right range in phase angle，since my model is quite complex，I need sometime to simplify it.