Modeling the behaviors of magnetorheological elastomer isolator in shear-compression mixed mode utilizing artificial neural network optimized by fuzzy algorithm (ANNOFA)

The main purpose of this paper is to develop numerical models for the prediction and analysis of the highly nonlinear performances of magnetorheological elastomer (MRE) isolator in shear-compression mixed mode. Dynamic behaviors of MRE isolator is experimentally tested, and the unique strain-dependent, frequency-dependent-stiffening and field-induced performances are observed and analyzed. An artificial neutral network approach optimized by fuzzy algorithm (ANNOFA) system is proposed for approximately capturing the nonlinear functional relationship between inputs (displacement, frequency and current) and output (force) of MRE isolator. Comparisons of the trained ANNOFA models with experimental results demonstrate the proposed ANNOFA modeling framework is an effective way to describe the complex behavior of the MRE isolator. In addition, the proposed ANNOFA model has a better forecasting accuracy than the conventional models (e.g. viscoelastic model, Bouc-Wen model and nonparametric model with back propagation neutral network) in the nonlinear system identification of MRE isolator.