Adaptive Forced Balancing for Magnetic Bearing Systems

This paper presents an autobalancing algorithm with several unique features that differentiate it from traditional (notch filter) and model based autobalancing algorithms. The algorithm, called Adaptive Forced Balancing with Frequency Tracking (AFB/FT), has the following features: 1) there is no effect on control system bandwidth or stability, 2) a dynamic model of the magnetic bearing system is not needed, and 3) regulation of the entire shaft-actuator-control system in a minimum energy state is achieved. This algorithm includes a frequency tracking section which automatically estimates the shaft rotational rate. The combined algorithm simultaneously tracks the amplitude, phase and frequency of a synchronous disturbance in realtime. AFB/FT was implemented on a TI Digital Signal Processor (DSP) as part of a complete control system designed to support a single end suspension test rig. This paper presents the AFB/FT algorithm, describes the test rig and control system components and presents test data which demonstrates the effectiveness of the algorithm. Preliminary results show that AFB/FT provides at least 30 dB reduction in actuator control current at 3600 rpm.