Development of a high-fidelity and optimal model for magneto-rheological damper by parametric sensitivity analysis

Abstract

This paper develops a high-fidelity and optimal mathematical model for the magneto-rheological (MR) damper, a popular semiactive control device, that can overcome the generic shortcomings of the existing parametric models, such as complex identification and simulation processes. A rigorous sensitivity analysis was performed using tornado diagrams and Monte Carlo simulations to reduce the number of parameters in the modified Bouc–Wen model with 21 amplitude and current parameters. Before the sensitivity analysis, a detailed experimental study with a laboratory-scale MR damper was performed to obtain its nonlinear hysteretic behavior through a series of displacement-controlled tests. From the results of the sensitivity analyses and exhaustive search, it was observed that the optimal Bouc–Wen model perfectly captured the nonlinear force–displacement and force–velocity behaviors of the damper. Moreover, the parameter identification results obtained using a genetic algorithm confirmed that the accuracy of the optimal model was the same as that of the original modified Bouc–Wen model, and the average simulation time of 1000 iterations was reduced considerably by up to 56% in the identified optimal model case. Finally, a numerical study was conducted on a multi-degree-of-freedom base-isolated building to evaluate the ability of the optimal model to simulate accurately the experimentally obtained nonlinear behavior of an MR damper. Additionally, the analysis results reinforce the broader use of mathematical models in the semiactive vibration control of structural systems.