Feasibility study of surrogate model for the application of vehicle suspension system

Abstract

High-fidelity (HF) model always provides better performance in assessing vehicle suspension system design compared to low-fidelity (LF) model. However, HF model is computationally expensive. On the contrary, LF model, which depends on a few parameters allow the simulation of ’what-if’ problem run faster and the results potentially comparable with HF model. This research attempts to conduct feasibility study on LF model using the surrogate model for the application of vehicle suspension study. The surrogate models are classified into three types which are Response-Based (RB) model, Variable-Based (VB) model, and Parameter-Based (PB) model. Through three statistical metrics and graphical interpretation, the results show that VB model gave the most superior performance compared to RB model and PB model.