Teoria de potência conservativa aplicada à identificação de falhas de curto-circuito em estator de motores de indução trifásicos

Abstract

The three-phase induction motor (TIM) is widely used in the industrial sector for electromechanical energy conversion due to its well-established characteristics, such as low acquisition and maintenance costs, robustness, and simplicity. However, over time, these machines can develop various types of faults. For this reason, several studies have been conducted to find alternatives that reduce the number of unexpected shutdowns, detect faults in their incipient stages, and decrease maintenance costs. With the advancement of new technologies, alternative methods based on artificial intelligence (AI), such as machine learning techniques, have been applied to the analysis of the TIM's voltage and current signals, with the aim of developing non- invasive and effective methods. In this work, combined with these Al techniques, the conservative power theory (CPT) is used to decompose the power components related to the motor and use them in the fault identification process. Despite being a relatively new approach, CPT has already demonstrated great effectiveness in various applications. The results obtained by the models in detecting faults in the TIM, using features extracted from the CPT's resulting power signals as input attributes, are promising, achieving high values in performance metrics. The combination of these techniques shows potential as a tool for accurately identifying faults. Therefore, the integration of conservative power theory with Al based methods offers an innovative and effective approach for the diagnosis and monitoring of faults in induction motors, with significant benefits for the industry, such as reduced maintenance costs and improved system reliability.