Modeling of Shock Absorber Behavior using Artifical Neural Networks

There is an increasing need for simulation of vehicle performance in the development process in order to achieve quality and reliability targets within (World Class) vehicle program timing. Simulation of vehicle dynamics, handling behavior and component loads has been achieved with multibody dynamic software. Prediction of service loads for component durability assessment requires accurate modeling of each component in the system over the full range of operation. In particular, the automotive shock absorber has a significant influence on the load transfer between suspension and body. Traditional models for shock absorber behavior are dependent on the internal velocity of the component. Previous studies and experimental data show that the behavior of the shock absorber is much more complex. In this study, neural network modeling techniques have been used to represent this complex behavior and provide a simple but accurate model that is computationally efficient and that can be integrated into a full vehicle simulation.