Online detection and suppression of mechanical resonance for servo system

Abstract

Mechanical resonance can often deteriorate the performance of servo system and even cause instable. For the ubiquitous phenomenon in servo system, this paper proposes a strategy for online detection and suppression of mechanical resonance. The Shifted Discrete Fourier Translations (SDFT) is employed to detect the existence of mechanical resonance and its information in real time, which is used to determine the parameters of the notch filter. To eliminate the jump of control input while the notch filter is put in the system dynamically, a new method based on gradually changing parameters and repeatedly discretizing filter is proposed. When the mechanical resonant varies with the conditions of load change, component ageing and mechanical coupling looseness, the real-time SDFT can detect the new vibration information automatically and adjust the parameter of the notch filter to suppress the vibration consequently. The experimental results on the turntable servo system show sufficiently that this proposed strategy is feasible and effective.