That was my workflow initially too. The entire ecosystem around computational simulation tools is lacking. Partly I also think it's because of the lack of attention to languages like Fortran which is mostly viewed as old by current generation of engineers or software developers, getting lesser attention compared to other languages like say Python. But then again, even the current python fem frameworks still run into some of the issues I outlined.

▲

elderring 9 hours ago | parent [-]

Your comment hits the nail on its head! These languages are still used in deep engineering space, nuclear & defence companies and institutions like NASA and the US Airforce. I like what your project is doing. How does the analytical testing works? can I create write my own tests?

	▲	white_tiger 8 hours ago \| parent [-]
		Each test is responsible for a part of the UMAT subroutine, checks it's correctness, compares the subroutine (FEA) with the analytical. Plot their results and show whether the tests passes or fails. AI is built into the process too, but in a manner where they are deterministic and with strong grounding on the physics. You can also write your own tests in Fortran and run it. You can also share it to the community for others to use. You can read the docs https://jacobee.netlify.app/docs