I occasionally I have encountered the following error or one similar to it:
I was finally able to get a reproducer (at least is has repeated a couple of times for me). I was able to capture it in a GUI script:
When run, it creates a model (two spheres in a block), sets up grid, equation, boundaries, and solves. It then splits the block and repeats grid and equation, but when boundaries is selected it gets the error:
Then going back (manually) to the Equation mode and selecting subdomain 4 gets a similar error:
Then if the fea struct is exported to MATLAB there seems to be an inconsistency between the number of subdomains (4) and the number of array positions for the coefficient data (3):
This does not always happen. I have been able to create this model successfully at times. Perhaps there is some detail that is different so that it IS repeatable, except I do something else sometimes that precludes the error from appearing.
Thanks for taking a look. If it does not repeat for you, perhaps you can give me some instructions as to what else to capture or examine.
Thank you for the issue report. Somewhere along the line the internal boundary designation has been switched to external one for a subset of the boundaries. I will investigate and most likely have a fix in the next release.
The workaround is to export your geometry if you have changed it (specifically to having a model with mixed internal/external boundaries to having more external boundaries), start a new model, and re-import the geometry before starting modeling again.