Robust adaptive controller for semi-active control of uncertain structures using a magnetorheological elastomer-based isolator

Publication Type:
Journal Article
Citation:
Journal of Sound and Vibration, 2018, 434 pp. 192 - 212
Issue Date:
2018-11-10
Filename Description Size
1-s2.0-S0022460X18304917-main.pdfPublished Version6.57 MB
Adobe PDF
Full metadata record
© 2018 Elsevier Ltd The isolator based on magnetorheological elastomers (MREs) that is used with semi-active controllers has emerged as a kind of smart devices that could potentially improve vibration control in traditional systems. The nonlinear damping characteristics of the isolator make the controller design more difficult. Generally, the command current/voltage is determined according to the relative responses in conventional semi-active controllers. However, these controllers may exhibit unsatisfactory isolation performance and even cause instability owing to significant excitations or nonlinear effects. In this study, the design of the new semi-active controller for an MRE-based isolator was investigated to overcome the drawbacks of traditional controllers from two perspectives. Firstly, an inverse model is designed for the isolator so that it can be used to predict an appropriate electric current supplied to the electromagnet based on the desired control force. Secondly, a robust adaptive controller for semi-active control is proposed for a nonlinear system with unknown dynamic parameters. The control scheme consists of three parts: a standard adaptive linearizing controller, an adaptive sliding mode controller, and a single robust controller. The proposed method guarantees zero convergence of the displacement response and provides robust stability. In addition, the singularity problem that usually appears in standard adaptive control is eliminated. Simulations demonstrate that the proposed controller exceeds the performance of the passive system as assessed in the protection of a two-story shear building during seismic events.
Please use this identifier to cite or link to this item: