Detection and reduction of middle frequency resonance for industrial servo

A novel method for middle frequency resonance detection and reduction is proposed for industrial servo system. Defects of traditional resonance reduction method based on adaptive notch filter in middle frequency range are analyzed and summarized as the difference between the resonance frequency and the oscillation frequency. A self-tuning low-pass filter is introduced in the speed feedback path, whose corner frequency is determined by FFT results and several self-tuning rules. With the proposed method the effective range of the adaptive filter is extended across the middle frequency range. Simulation and Experiment results show that the frequency detection is accurate and resonances during the speed steady states and dynamics are successfully reduced.