Development of Modeling Method for the Self Switchable Hydromount by Machine Learning

Abstract

We present a modeling method for self switchable hydromount from the specification of spring and damping by using machine learning. This is because this component has a non-linear characteristic depending on amplitude and frequency, so it is difficult to substitute to the equivalent dynamics system which is generally performed in the time series calculations at the model-based development. We confirmed the model’s prediction accuracy and also inspected the calculation stability when it is incorporated in 1D-CAE.