Detection and self-tuning compensation of periodic disturbances by the control of DC motor

Abstract

The accuracy with which a DC motor and drive system track a reference velocity profile is limited by periodic torque disturbances. The cogging and ripples torques, and constructive imperfections in the mechanical assembly of the drive cause disturbance pulsation harmonics with a base frequency dictated by the rotational velocity. This paper describes a novel compensation technique based on disturbance observation and a self-tuning feed-forward compensation algorithm. The DC motor is modeled as a linear system augmented by the nonlinear Coulomb friction and is experimentally identified from a set of the system responses. The disturbance harmonics are detected by means of the fast Fourier transformation (FFT) and analytically described by a spatial Fourier transform with respect to the angular position of the rotor. The online algorithm tunes the parameters of the feed-forward compensator using the recursive estimation technique. The proposed self-tuning compensator is experimentally verified as part of an open control loop at different level of the rotational velocity and system load.