Thank you very much Rafal. I’m stuck in your Portable version from 2020:Mfront_portable.7z
This is where I’m now in case it helps.
I have successfully install the new package by means of pacman -S mingw-w64-x86_64-tfel.
Add paths to the system variable.
When I run ccx 2.21 inside a folder that contains your portable files Mfront runs fine but just for some material behaviors.
example:
SaintVenantKirchhoffElasticity
DruckerPrager
That means a lot because it means the libraries are found.
Some other material behaviours throw exceptions like “tangent operator not yet available” but I guess that is related with the actual development state of the material behavior or wrong material parameters set up from my side not ccx. That’s the reason why I would like to update and see now if fixed.
But when I run ccx 2.21 using the full set of dll inside the new updated folder C:\msys64\mingw64\bin something doesn’t work.
I have compared one by one and I think this two dll are the big difference between your portable version and the full recent new one installed on my computer:
libCALCULIXBEHAVIOUR.dll
libTFELMaterial.dll
There is not too much feedback abaout the failure. I have completeed the dll collection with yours when the system asked for a missing lib ( libCALCULIXBEHAVIOUR.dll ) but doesn’t work.
It is possible that due to the lack of these: librarieslibgcc_s_seh-1.dll, libwinpthread-1.dll, libstdc++-6.dll, MFront with the Mecway or PrePoMax could not work (bconverged has these libraries in the bin folder).
below list of material model are not yet available also. I’m not currently used, but maybe someone else is interesting. Only to notify for next or future release and distribution.
Thank you very much Rafal. I need to do some more test. Actually, if I use the new libCALCULIXBEHAVIOUR.dll I get :
The entry point of the procedure is not found.
Just replacing libCALCULIXBEHAVIOUR.dll by the old one it works again but I guess it is then calling the old files. I have try to run a new material model and it failed.
unfortunately, not all MFront material models can be supported in CalculiX, and probably it’s related to hydro-mechanical problems. Previously, i’m not yet to further study the implementation and limitation, thanks for mentioning.
many thanks for updates, btw some material is missing. I did not find MohrCoulombAbboSloan in latest distribution. Also, example input files of DruckerPragerCap is different from previous.
thanks for mentioned the author of TFEL/MFront and some example. At least an addition for MohrCoulombAbboSloan could be useful since it has tension cut-off, interesting to know how numerically stable is it.
These are my results for Drucker-Prager CAP.
Time step had to be reduced for the transition area between CAP and Linear.
Three of the reference points agree with input data.
Slope (34.4º), Minimum hidrostatic pressure (-75 MPa) and Intersection with ordinate (150 MPa)
Transition point between the two plastic surfaces is slightly ofsset.