Multi-directional design control of plastic crash components by means of domain-predictive feed-forward neural networks

The robust crash design of plastic components is of great interest in automotive engineering. For an efficient development, a prototype-free development is aimed at, whereby at the same time the test programme can be strongly reduced and in the future possibly completely omitted, if the virtual desi...

Full description

Saved in:
Bibliographic Details
Published inStructural and multidisciplinary optimization Vol. 64; no. 6; pp. 4115 - 4128
Main Authors Schrader, Kai, Schauer, Marco
Format Journal Article
LanguageEnglish
Published Berlin/Heidelberg Springer Berlin Heidelberg 01.12.2021
Springer Nature B.V
Subjects
Artificial neural networks
Automotive engineering
Computational Mathematics and Numerical Analysis
Engineering
Engineering Design
Industrial Application Paper
Mathematical analysis
Neural networks
Robustness
Theoretical and Applied Mechanics
Crash-simulation
Finite-element-method
Design of plastic components
Artificial neural networks
Online AccessGet full text

Cover

More Information
Summary:The robust crash design of plastic components is of great interest in automotive engineering. For an efficient development, a prototype-free development is aimed at, whereby at the same time the test programme can be strongly reduced and in the future possibly completely omitted, if the virtual design delivers robust results. Based on the plastic component design of a head box, a possible procedure consisting of finite element calculation and artificial neural networks is described exemplarily. The head box is the load-bearing part of the head restraint of a car seat. It can be shown that, if these methods are used correctly, components can be designed very quickly to fulfil the specified objective functions.
ISSN:1615-147X
1615-1488
DOI:10.1007/s00158-021-03031-w