Premium
An intelligent meta‐element for linear elastic continua
Author(s) -
Koeppe Arnd,
Bamer Franz,
Markert Bernd
Publication year - 2018
Publication title -
pamm
Language(s) - English
Resource type - Journals
ISSN - 1617-7061
DOI - 10.1002/pamm.201800283
Subject(s) - finite element method , benchmark (surveying) , element (criminal law) , computer science , linear elasticity , artificial neural network , boundary value problem , boundary (topology) , mathematical optimization , mixed finite element method , mathematics , algorithm , structural engineering , artificial intelligence , mathematical analysis , engineering , geology , geodesy , political science , law
Large numbers of nodes and nonlinearities make the numerical solution of finite element models computationally expensive. We introduce “intelligent” elements for linear elastic continua, which use neural networks to reduce the number of nodes, and demonstrate their efficiency on two benchmark examples. It is shown, that models with “intelligent” meta‐elements accurately solve the boundary value problem and require a smaller number of nodes than conventional finite element models.