Aplicação da norma ieee 1459 - 2010 e parâmetros de hjorth para a identificação de falhas de estator em máquinas de indução trifásicas

Abstract

Three-phase induction motors (TIMs) are widely employed in industrial applications due to their robustness, simple construction, and reliability. However, stator inter-turn short-circuit faults represent a significant portion of failures in three-phase electrical machines, accounting for up to 36% of the cases reported in the literature. Although several diagnostic methodologies based on current signals have been proposed, they may be susceptible to voltage unbalances and asymmetries arising from the manufacturing process or from the use of frequency inverters to drive the motor. This limitation highlights the need for alternative approaches capable of robustly extracting discriminative features from signals acquired from the machine while maintaining diagnostic effectiveness under adverse operating conditions. In this context, this work investigates the features defined by the IEEE 1459-2010 standard, obtained from line voltage and current signals, and Hjorth parameters extracted from the motor current signal to identify inter-turn short-circuit faults with different severity levels. These features are used independently and jointly, considering both single-phase and three-phase observation scenarios, as inputs to classification models aimed at detecting stator short-circuit faults under different fault severities, supply unbalance conditions, and motor drive methods. The results demonstrate that combining the features defined by IEEE 1459-2010 with Hjorth parameters achieves accuracy rates above 97% for the single-phase monitoring scenario, whereas, in the three-phase observation scenario, the models achieve accuracies above 98.5%, regardless of the motor drive method employed. Furthermore, reactive power was found to be one of the features carrying the most information regarding the motor operating condition