The aim of the present study is to investigate how structural models with inelastic material behaviour can be enhanced by means of artificial intelligence. To this end, artificial neural networks are proposed to replace continuum mechanical models in structural dynamics. This concerns predictions of shock wave‐loaded plates and finite element simulations with neural network enhanced material modelling. By comparison to experiments, the accuracy and the computational effort of the new approach is investigated.
Paper
Full text
Artificial neural networks in structural dynamics
Semantic Scholar · Engineering · 2019
Abstract
The aim of the present study is to investigate how structural models with inelastic material behaviour can be enhanced by means of artificial intelligence. To this end, artificial neural networks are proposed to replace continuum mechanical models in structural dynamics. This concerns predictions of shock wave‐loaded plates and finite element simulations with neural network enhanced material modelling. By comparison to experiments, the accuracy and the computational effort of the new approach is investigated.