Comprehensive experimental study on the stabilization of active magnetic bearing system

This paper investigates performance of various controllers on the stabilization of active magnetic bearing system (AMBs). Firstly, since AMBs are inherently open-loop unstable, closed-loop system identification is adopted in order to obtain the model of the system under study. Next, based on the obtained model, Linear Quadratic Gaussian (LQG), PD-like Fuzzy Controller (PD-FLC), PID-like Fuzzy Controller (PID-FLC) and H_∞, Controller are designed. These controllers are then coded in C and implemented in real time on both x and y channels at one side of the AMBs via a Digital Signal Processor (DSP). The performance of the designed controllers is compared while the system is stationary and subject to step disturbances and while the system is rotating at high speed. As minimizing energy consumption is one of the obvious desirable objectives to achieve, these different stabilizing controllers are also compared in terms of the resulting control energy. The ultimate aim of this paper is to develop an intuition on how to design high performance stabilizing controllers of open-loop unstable systems like AMBs.