Machine Learning Assisted Development of Eight Node Hexahedral Finite Element

The finite element method (FEM) is a popular numerical technique for solving partial differential equations (PDEs) arising in computational modeling. Several engineering problems, such as those involving fluid flows, electromagnetics, heat transfer, and structural analysis, have been effectively solved using FEM. It is the most powerful tool currently in structural analysis. Recently, machine learning (ML) approaches have been used to enhance the performance of FEM. The main aim of this study is to build an artificial neural network (ANN) model using deep learning, a popular and powerful ML algorithm that can estimate the elemental stiffness matrices of 3D 8-noded hexahedral elements with very high accuracy. The performance of the model is evaluated by comparing it with Cook’s beam problem. The model predicts the displacement of the beam with an error of 0.74%.