Modeling the Response of ER Damper: Phenomenology and Emulation

Abstract

In this paper physically motivated and nonparametric models are investigated that predict within some tolerance the response of a semiactive electrorheological (ER) damper that was designed, constructed, and tested. The electrorheological damper is a hydraulic device that was designed for applications in vibration control of civil structures. The simplest possible physically motivated phenomenological models are first considered to predict the damper response without and with the presence of electric field. Subsequently, the performance of a multilayer neural network constructed and trained by an efficient algorithm known as the dependence identification algorithm is examined to predict the response of the ER damper. The performance of the neural network is compared to that of the phenomenological models and some conclusions are provided.