New sliding mode control of building structure using RBF neural networks

Abstract

The undue chattering effect is the major disadvantage of conventional sliding mode controllers. In this study, based on the advantage of RBF neural network control method, a new adaptive sliding mode control method, which is one of the active control algorithms, has been applied for seismically-excited building structures. The undue chattering effect, the major disadvantage of conventional sliding mode controller, has been avoided by introducing the new control method. First, we build the motion equation and design the switching surfaces. Next, based on the RBF neural network control algorithm, we adjust the control gain parameter and then design the neurocontroller. For numerical applications, a three-storey shear building model subjected to ground excitations has been considered. The ground accelerations recorded in two different earthquake events have been used to evaluate the effectiveness of the control algorithm for varied disturbances. The simulation results show preliminarily that our new adaptive sliding mode control method is quite effective: not only can it reduce the peak-response of the ground motion, but also it can keep the chattering effect sufficiently low.