Dear All,
How can I eliminate the triangular holes/defects/leaks around the intersection parts from the generated mesh as shown below and circled by yellow highlights? And it is hard to find them if not zooming in the single cell size scale.
Whether they will impact the simulation or not. And whether I can select different meshing methods from the cut cell. Thank you.
And considering the edges of the generated mesh matter and the some single generated cells missing as defects, I was trying to simulate at first, but after about 13mins, DT<DTMIN type error appears, which stopped the simulation. I want to know whether such kind of defects make the simulation not converge and how can I fix that. Thank you.
I hereby attach the .mfx and .stl files to see whether you can help me to find out why the project would be not convergent after 13 mins without defects on the generated meshes from the raw eyes, which would broaden the scope of challenges the MFIX could tackle with and used as tutorial examples later. Thank you.
chamber_delicate_structure2.mfx (14.7 KB)
chamber_delicate_structure2.STL (766.9 KB)
And if I set the dot product tol small, the following bug reported. Thank you.
chamber_delicate_structure1_2021-10-05T145209.031333.zip (51.6 MB)
Hi Ju -
Thank you for the bug report - these are always appreciated, and help us improve MFiX. I can’t answer all your questions about meshing, but I can look at these tracebacks and fix any bugs that you encounter in the MFiX GUI.
The traceback in the most recent error report (chamber_delicate_structure1_2021-10-05T145209.031333.zip) is not related to holes in the mesh - you just happened to discover an unrelated GUI bug.
Error: 'QMoveEvent' object has no attribute 'modifiers'
File "C:\Users\juu\.conda\envs\mfix-21.3\lib\site-packages\mfixgui\vtk_widgets\base.py", line 125, in moveEvent
return qrw._mouseMoveEvent(ev)
File "C:\Users\juu\.conda\envs\mfix-21.3\lib\site-packages\vtkmodules\qt\QVTKRenderWindowInteractor.py", line 501, in mouseMoveEvent
self.__saveModifiers = ev.modifiers()
The bug is in code that is handling mouse events in the VTK window, it’s not related to the solver or the mesher. I think setting the dot product tolerance small probably helped, since you got further along this time.
This failure is occuring in new code I added recently to make the color-bar in the plots clickable. I will fix this, and we will probably have a 21.3.1 release containing this and a few other small bugfixes. However, this error should not be fatal - it should be safe to dismiss the error dialog and continue, maybe creating a new VTK tab will help as a workaround.
– Charles