A New Acoustic Emission-Based Approach for Supply Disturbances Evaluation in Three-Phase Induction Motors

• Autor(a) principal: Lucas, Guilherme Beraldi [UNESP]
• Data de Publicação: 2021
• Outros Autores: De Castro, Bruno Albuquerque [UNESP], Rocha, Marco Aurelio [UNESP], Andreoli, Andre Luiz [UNESP]
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1109/TIM.2020.3047492; http://hdl.handle.net/11449/205679
• topic: Acoustic emission (AE); fault diagnosis; induction motor; piezoelectric sensors; wavelet transform (WT); zero-crossing-weighted energy

The interruption of a three-phase induction motor (TIM) on production lines represents a high financial and operational cost. However, these machines are often exposed to mechanical and electrical failures that can cause unexpected stoppages. Among these failures, the voltage unbalance (VU) is a supply fault that can lead to winding wear, torque losses, overheating, and other side effects. In this context, the acoustic emission (AE) analysis stands out as a promising nondestructive technique (NDT) in TIM monitoring. However, the AE method was not previously completely validated for VU diagnosis, and several research gaps need to be filled. Therefore, this work proposes a novel AE approach for detection, phase identification, and magnitude classification of VU. For this purpose, an electrical machine monitored by piezoelectric sensors was subjected to different levels of unbalanced voltages. The AE signals were processed using the novel zero-cross-weighted energy (ZE) index. This metric was based on the energy of the wavelet transform (WT) coefficients weighted by zero-crossing rate values. Experimental results revealed that the proposed index proved to be effective for detecting the VU occurrence. Besides, ZE-based data separation was proposed and achieved VU phase identification. Finally, the magnitude of the unbalanced voltages was classified by linear regression. The accuracy parameters for detection, phase identification, and magnitude classification stated the reliability of the new approach. Finally, the efforts of this work provide new functionalities to traditional AE systems.