Implementation of Reduced-Order Model to Design of Electric Powertrain Rubber Mount

Abstract

This paper deals with possibilities of utilizing neural networks in reduced order modelling. First section of the paper contains theory on reduced order models and their implementations. Following this, subsequent chapter contains methodology concept for obtaining sufficient amount of data and creating reduced order models as well as their evaluation process. Next section describes our findings and summarizes them. Conclusion describes further steps we plan to take in this research.

Keywords

• reduced order model
• neural network
• multilayer perceptron
• proper orthogonal decomposition
• rubber mount