Condition monitoring of BLDC motor drive systems by Hilbert Huang Transform

Abstract

This paper presents a novel on-line condition monitoring method that is based on the input voltage and measured line current of a Brushless DC motor drive system. The time-frequency signal processing method of Hilbert Huang Transform is applied to extract the characteristic component of line current and then obtain the required equivalent electrical impedance, which is utilized for health monitoring. Gauge Repeatability & Reproducibility is employed to evaluate the reliability of the computed impedance during the training strategy. A health reference index defined from the characteristic intrinsic mode function is proposed for decision making during on-line monitoring. Theoretical analysis and experiments are conducted to evaluate the effectiveness of the proposed health monitoring approach.