Application of Discretization Method with a Phase Compensation Technique for Adaptive-Notch-Filter Based Resonance Suppression in Industrial Servo Systems

Abstract

This paper presents an application of a discretization method with a new phase compensation technique for adaptive-notch-filter (ANF) based resonance suppression in an actual industrial servo system. The ANF includes a frequency estimator which estimates the frequency of the vibration in the input signal in real time and automatically filters it out. Because the ANF is designed in the continuous-time domain, a discretization process is required for digital implementation. However, the discretization error may restrain the performance of the ANF, especially when the resonance frequency is near the Nyquist frequency. This paper develops the discretization method for the ANF that combines a modified bilinear transform method with a compensation method for the frequency and damping ratio warping, and a phase compensation technique. The proposed method reduces the magnitude and phase error near Nyquist frequency, and thus the discretized ANF can suppress the resonance in relatively high-frequency. Experiments using an industrial servo system are performed to show the improved frequency estimation performance of the developed ANF.