The repo has the refactored code of our paper published at Computer Methods and Programs in Biomedicine, titled "PyTorch-FEA: Autograd-enabled Finite Element Analysis Methods with Applications for Biomechanical Analysis of Human Aorta" at https://doi.org/10.1016/j.cmpb.2023.107616
I am working to make it useful for more applications.
The orignal code of the paper is available at https://github.com/liangbright/pytorch_fea_paper
The preprint of our paper is available at https://www.biorxiv.org/content/10.1101/2023.03.27.533816v1
PyTorch-FEA needs the mesh library at https://github.com/liangbright/mesh
Example data: https://drive.google.com/file/d/1ByOjc9RVFEexLXB-u6Qd1SMAS-BKvW3g/view?usp=sharing
Try those examples:
forward analysis:
aorta_FEA_QN_forward_inflation.py to obtain pressurized geometry given material parameters and unpressurized geometry.
inverse analysis:
(1) aorta_FEA_inverse_mat_ex_vivo.py to obtain material parameters given pressurized and unpressurized geometries.
(2) aorta_FEA_QN_inverse_p0.py to obtain unpressurized geometry given material parameters and pressurized geometry.
(3) aorta_FEA_QN_GPA_prestress.py to obtain stress and strain of pressurized geometry given material parameters.
note: residual stress/strain is not considered
Dependency: PyTorch, PyTorch Geometric, and PyPardiso