Identificação de falhas em motores de indução trifásicos utilizando a transformada de hilbert-huang e multissensoriamento

Abstract

Three-phase induction motors (TIMs) are considered robust and reliable electrical machines, which explains their wide application in various industries and production systems. However, like any other equipment, these TIMs are susceptible to adverse operating conditions and component failures, making it crucial to monitor their operation to avoid unexpected shutdowns and maintenance, which can result in losses within the industrial process. Due to the complexity of these processes and their impact on TIMs, one field of study in fault diagnosis that has received attention is the analysis of signals in the time-frequency domain, which deals with the non-stationary and nonlinear characteristics of signals. This work proposes a multi-sensor investigation using HHT in the processing of vibration, electrical current, and acoustic signals for diagnosing bearing, stator, and rotor faults in TIMs, highlighting techniques and methodologies for applying this tool to develop an alternative approach. To demonstrate the robustness of the proposed method, two TIMs with different power ratings will be used under various sinusoidal power supply conditions and oscillations in the motor shaft torque, in addition to diagnosing three types of faults with varying severities using a single methodology. Features of HHT were calculated for fault diagnosis, with their classification demonstrating separability and the potential to differentiate faults. Binary diagnostic results for faulty and healthy TIMs exceeded 99% accuracy, and multi-class faulty and healthy TIMs achieved an accuracy of over 96%. The results and experiments demonstrate the effectiveness of the HHT-based methodology for fault diagnosis in TIMs.