Self-updated four-node finite element using deep learning

Abstract This paper introduces a new concept called self-updated finite element (SUFE). The (FE) is activated through an iterative procedure to improve the solution accuracy without mesh refinement. A mode-based formulation devised for four-node and assumed modal strain employed bending modes. search optimal directions implemented deep learning given deformation minimize shear locking. proposed element, which developed. passes patch zero-energy mode tests. As number of iterations increases, solutions become more accurate, resulting in significantly accurate with few iterations. SUFE very effective, especially when meshes are coarse severely distorted. Its excellent performance demonstrated various numerical examples.