OVERVIEW OF MACHINE LEARNING AND DEEP LEARNING APPLICATIONS IN FINITE ELEMENT ANALYSIS FOR STRUCTURAL MODELING AND SIMULATION

Abstract

Abstract: The integration of Machine Learning (ML) and Deep
Learning (DL) into Finite Element Analysis (FEA) has emerged as
a prominent trend, addressing computational challenges while
opening new research directions in computational mechanics.
This article presents a comprehensive review of the role of ML
and DL in enhancing FEA across preprocessing, solving, and postprocessing stages. Surrogate modeling, data-driven material
modeling, and Physics-Informed Neural Networks (PINNs)
are identified as breakthroughs with capability of reducing
computational costs, managing uncertainty, and enabling
inverse analysis. A comparative discussion between ML/DL
approaches and traditional FEM highlights both advantages and
limitations. Future perspectives include the development of hybrid
models, robust uncertainty quantification frameworks, and high
performance data-driven solvers. The findings suggest that ML
and DL will play a pivotal role in the evolution of next-generation
computational mechanics.