Due to the non-linear dynamics of the human body and the complex interaction between vehicle’s subsystems, it is still challenging to predict the vibration discomfort of a vehicle occupant accurately. In the automotive industry, Finite Element (FE) method has widely used to predict the vibration transmission through a large structure (e.g., a vehicle body). However, it is still challenging and costly to accurately model a complex structure including subsystems with non-linear dynamics, such as the human body and seat foam. A vehicle plus occupant’s body FE model often requires high computational time and resources to run and calibrate. This paper presents a data-driven hybrid method is presented to predict the vehicle body vibration transmission to the occupant’s body by combining the numerical transfer matrices of subsystems. The transfer matrix of a subsystem can be obtained from either vibration experiments or the FE models. This method allows the direct combination of the physical experiment data and the FE models of the subsystems. This study further demonstrates a modular numerical method in predicting the vibration discomfort of the vehicle occupant.

RMIT University: Jianchun Yao, Mohammad Fard; NHK Spring Co Ltd: Kazuhito Kato