Boyang Chen's Group

GitHub - BoyangChenFEM/Summer2022: Training ANNs with FEM data on Open Hole Composite plate - for summer school 2022 in DelftTraining ANNs with FEM data on Open Hole Composite plate - for summer school 2022 in Delft - BoyangChenFEM/Summer2022

Training Machine Learning model (Artificial Neural Networks) with FEM data on Open Hole Composite plate - for the AI summer school 2022 in TU Delft Based on the former Master theses work of Tom Gulikers and Karthik Ventakesan

Non-linear Modeling - Finite Element Method | TU Delft OnlineThis course will provide you with a theoretical understanding of non-linear Finite Element Modeling and its application to practical engineering problems.

Currently, I am teaching a Master course called "Non-Linear Modeling", which covers finite element theories and practical knowledge on Newton-Raphson method, Arc-length method, Geometrical Nonlinearity, Contact, Plasticity, Cohesive Zone, Damage, and Fracture. This course is offering both on campus and online. If you are not a TU Delft Aerospace Master student and are interested in taking this course, you can register as an online student on the course website.

GitHub - BoyangChenFEM/APPfemLinear: Linear FEM programme in Python, using Abaqus/Salome input file, Python solver and Paraview postprocessingLinear FEM programme in Python, using Abaqus/Salome input file, Python solver and Paraview postprocessing - BoyangChenFEM/APPfemLinear

Linear Finite Element Program in Python

It reads Abaqus input file and outputs data in VTK format for postprocessing in Paraview.

Sparse matrix from Scipy.sparse and sparse matrix solver from Pypardiso are used to efficiently store and solve the system equation, respectively.

Two direct solvers are implemented, one following the book of Tom Hughes (solving on reduced system matrix on the unknown degrees of freedom only) and another in the style of Zienkiewicz (solving on the full system matrix w.r.t. all degrees of freedom, but set the diagonal terms w.r.t. imposed degrees of freedom to a dummy value of 1). For the curious ones, you can try out both solvers to see which one is faster:).