A multitasking DSP implementation of adaptive magnetic bearing control

Abstract

Several adaptive on-line synchronous vibration control algorithms are demonstrated on a laboratory rotor supported in magnetic bearings with a high-speed digital controller. Central to the controller hardware developed for this application is a floating-point digital signal processor. The algorithms are executed under a multitasking real-time operating system specifically written for the hardware platform. The multitasking operating system allows the feedback control algorithm and the adaptive vibration control algorithm to be written and executed as separate tasks. This greatly simplifies programming, code verification, and algorithm development. Tests conducted on the laboratory rotor system indicate that greater than 98% attenuation in unbalance response can be achieved with the adaptive control algorithms.