DESIGN OF ROBUST FUZZY SLIDING MODE CONTROLLER FOR MAGNETIC LEVITATION SYSTEM

Abstract

Magnetic levitation ball system is a nonlinear open loop unstable complex system. But approximate model can be obtained through linearization and can be controlled by nonlinear control theory. Uncertainty comes mainly due to modelling error and other outside disturbances. Due to these uncertainties and unstable nature, controlling the position of the levitated object becomes very difficult. So, there is a need of controllers with high performance for adjustment of levitated object position. Firstly, a sliding mode controller is developed by using a control law that drags the state trajectory onto a predefined surface, which is called switching surface. But imperfect control switchings in control law lead to the problem of chattering. To avoid chattering a boundary layer is introduced in the surface with help of fuzzy controller. The fuzzy controller which we designed is different from the conventional fuzzy controller as in this we have not fuzzified error and error change rather we - fuzzified the sliding surface. Sliding surface is a composition of error and error change, so the number of rules also gets reduced. Continuous approximation and smoothening of discontinuous - control law is been done in this boundary layer which ultimately removes chattering completely. Robustness is achieved after the trajectory strikes the sliding surface since the system becomes insensitive to external disturbances and parameter variations. To ensure stability we use Lyapunov's theory. Simulations are performed on magnetic levitation ball system through MATLAB and results are presented and discussed in chapter 5.