Methodology for fault detection in induction motors via sound and vibration signals

• Published in: Mechanical systems and signal processing 2017-01, Vol.83, p.568-589
• Main Authors: Delgado-Arredondo, Paulo Antonio, Morinigo-Sotelo, Daniel, Osornio-Rios, Roque Alfredo, Avina-Cervantes, Juan Gabriel, Rostro-Gonzalez, Horacio, Romero-Troncoso, Rene de Jesus
• Format: Article
• Language: English
• Subjects: Acoustic sound; CEEMD; Decomposition; Fault detection; Fault diagnosis; Frequency domains; Induction motors; Methodology; Motors; Sound; Spectral analysis; Vibration; Vibration analysis
• ISSN: 0888-3270; 1096-1216
• DOI: 10.1016/j.ymssp.2016.06.032

Nowadays, timely maintenance of electric motors is vital to keep up the complex processes of industrial production. There are currently a variety of methodologies for fault diagnosis. Usually, the diagnosis is performed by analyzing current signals at a steady-state motor operation or during a start-up transient. This method is known as motor current signature analysis, which identifies frequencies associated with faults in the frequency domain or by the time–frequency decomposition of the current signals. Fault identification may also be possible by analyzing acoustic sound and vibration signals, which is useful because sometimes this information is the only available. The contribution of this work is a methodology for detecting faults in induction motors in steady-state operation based on the analysis of acoustic sound and vibration signals. This proposed approach uses the Complete Ensemble Empirical Mode Decomposition for decomposing the signal into several intrinsic mode functions. Subsequently, the frequency marginal of the Gabor representation is calculated to obtain the spectral content of the IMF in the frequency domain. This proposal provides good fault detectability results compared to other published works in addition to the identification of more frequencies associated with the faults. The faults diagnosed in this work are two broken rotor bars, mechanical unbalance and bearing defects. •Fault detection in induction motors with vibrations and acoustic signals is shown.•Broken rotor bars, bearing defects and unbalance in induction motors are analyzed.•The Complete Ensemble Empirical Mode Decomposition method is used.•The frequency marginal of the Gabor is used for spectral estimation.