Hybrid time-frequency domain analysis for inverter-fed induction motor fault detection

• Main Authors: CHUA, T. W., TAN, W. W., WANG, Zhaoxia, CHANG, C. S.
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2010
• Subjects: Hybrid time-frequency method; Inverter-driven induction motor; Real-time fault diagnosis; Robust algorithm; Numerical Analysis and Scientific Computing; Operations Research, Systems Engineering and Industrial Engineering
• Online Access: https://ink.library.smu.edu.sg/sis_research/6730; https://ink.library.smu.edu.sg/context/sis_research/article/7733/viewcontent/2010_Hybrid_Time_Frequency_Domain_Analysis_for.pdf
• Institution: Singapore Management University

The detection of faults in an induction motor is important as a part of preventive maintenance. Stator current is one of the most popular signals used for utility-supplied induction motor fault detection as a current sensor can be installed nonintrusively. In variable speeds operation, the use of an inverter to drive the induction motor introduces noise into the stator current so stator current based fault detection techniques become less reliable. This paper presents a hybrid algorithm, which combines time and frequency domain analysis, for broken rotor bar and bearing fault detection. Cluster information obtained by using Independent Component Analysis (ICA) to extract features from time domain current signals is combined with information extracted from fast Fourier transformed signal to reveal any underlying faults. To minimise the effect of the noise in the raw signal and intra-class variance in the extracted feature, a novel noise reduction approach- Ensemble and Individual Noise Reduction is employed. An advantage of the proposed scheme is that time domain analysis module can provide an early fault detection with minimal computation complexity. Experimental results obtained on the three-phase inverter-fed squirrel-cage induction motors demonstrated that the proposed method provides excellent classification results.