Detection and phase identification of inter-turn short-circuit faults in three-phase induction motors using MEMS accelerometer and Hilbert transform

• Autor(a) principal: Rocha, Marco [UNESP]
• Data de Publicação: 2021
• Outros Autores: Lucas, Guilherme [UNESP], Souza, Wallace [UNESP], de Castro, Bruno Albuquerque [UNESP], Andreoli, André [UNESP]
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1109/INDUSCON51756.2021.9529634; http://hdl.handle.net/11449/222500
• topic: Diagnosis; Electrical machines and drives; Industry applications; Prognosis and system identification

Three-phase induction motors (TIMs) play a key role in the industrial scenario. Due to their robustness, low cost, and efficiency, TIMs have been increasingly used in industrial applications, making them the main source of electromechanical power to drive many types of loads. However, induction motors are eventually subjected to mechanical and electrical failures that can cause unexpected shutdowns. Among them, inter-turn short circuit faults (ITSC) in the stator windings correspond to the highest incidence of electrical faults in TIMs. These interruptions represent a high financial and operational cost. Therefore, the importance of detecting and identifying ITSC faults in the operation of TIMs increases. An analysis tool that shows a good result for a non-invasive technique (NIT) of damage diagnosis in induction motors is the study of mechanical vibration signals in the machine. In this paper, these signals were processed to identify and classify the ITSC faults present in the motor. For this purpose, a MEMS accelerometer was coupled to a TIM. Then, vibration signals were acquired for an ITSC fault present in each phase of the induction motor. The data was processed using the Hilbert transform energy and cross correlation. Thus, it was possible to detect the occurrence of an inter-turn short-circuit fault and identify in which phase the fault occurs by clustering the results.