Fuzzy control of magnetorheological dampers for vibration reduction of seismically excited structures

Since civil structures have little damping capability, extensive damage and even failure often occur when they are subjected to seismic excitations. Control devices have therefore been developed and implemented to dissipate energy from earthquakes and reduce structural vibrations. Magnetorheological (MR) dampers are examples of such devices and consist of a hydraulic cylinder containing a solution that, in the presence of a magnetic field, can reversibly change from viscous fluid to semi-solid. The objective of this research is to develop fuzzy controllers to regulate the damping properties of MR dampers and reduce structural responses of single degree-of-freedom seismically excited structures. Three fuzzy controllers were therefore designed and their effectiveness evaluated through series of numerical simulations. Since fuzzy control uses expert knowledge instead of differential equations, it allows for the development of simple and robust algorithms that do not require information on plant's structural and vibration characteristics. They are therefore an attractive alternative for controlling systems that are complex, nonlinear, or that contain ambiguity or vagueness. The first algorithm proposed is a fuzzy control system with two inputs: structural displacement and velocity. The second is referred to as gain-scheduled fuzzy control and varies the velocity input scaling factor according to incoming ground acceleration. The last one, self-tuning fuzzy control, uses a fuzzy inference mechanism based on ground acceleration intensity and building displacement to adjust the velocity input scaling factor. Robustness of these controllers to changes in seismic motions and structural characteristics were evaluated by subjecting two different buildings controlled by each of these strategies to a wide range of earthquake records. Results show that the algorithms proposed effectively reduced responses of both structures to a wide range of seismic motions. They were also found to be robust to changes in ground excitations and structural characteristics. In addition, adjusting the velocity input scaling factor according to ground acceleration intensity considerably improved the controller's ability to reduce structural vibrations, since both the gain-scheduled and the self-tuning fuzzy algorithms reduced structural responses more effectively than the fuzzy controller with constant scaling factors. Finally, the self-tuning controller outperformed all other strategies for most earthquakes considered.