Fourier transform and image processing for automatic detection of broken rotor bars in induction motors. (3rd August 2018)

• Record Type: Journal Article
• Authors: De Santiago-Perez, J Jesus; Rivera-Guillen, Jesus R; Amezquita-Sanchez, Juan P; Valtierra-Rodriguez, Martin; Romero-Troncoso, Rene J; Dominguez-Gonzalez, Aurelio
• Is Part Of: Measurement science & technology. Volume 29:Number 9(2018:Sep.)
• Publication Date: 2018-08-03
• Subjects: fault diagnosis -- discrete Fourier transform -- image processing -- induction motor; Physical measurements -- Periodicals; Scientific apparatus and instruments -- Periodicals; Equipment and Supplies -- Periodicals; Science -- instrumentation -- Periodicals; Technology -- instrumentation -- Periodicals; Mesures physiques -- Périodiques; Physical measurements; Scientific apparatus and instruments; Periodicals; 502.87
• DOI: 10.1088/1361-6501/aad3aa
• Languages: English
• ISSNs: 0957-0233

Abstract

Abstract: In the literature, many research studies have proposed the diagnosis of the induction motor condition where both the electrical and mechanical faults have been considered. Despite obtaining promising results, the diagnosis is mostly achieved qualitatively, requiring an expert user to interpret the results. This disadvantage could lead to time delays and additional costs that at an inopportune stage could prevent the diagnosis of the motor. In this study, a methodology based on signal processing and image processing is proposed to automatically diagnose a broken rotor bar (BRB) by using current signals. For the signal processing, first a decimation stage is proposed and then short-time Fourier transform is applied to the current signal. The proposed signal processing eliminates the 60 Hz component of the power line and its associated leakage. Next, the fault diagnosis is automated by applying image processing algorithms to the time-frequency plane of the current signal. From this time-frequency plane, the region of interest, in this case, the V-shaped pattern associated with the BRB condition, is automatically located mainly by using mathematical morphology-based algorithms. In addition, the area of the V-shaped pattern is also computed in order to automatically distinguish between the faulty and healthy condition, half BRB, one BRB, and two BRB. This last step of the method avoids the need of an expert user. For the area values, an analysis of variance is performed, where a 100% effectiveness is obtained for automatically determining the motor condition.