Electrical current analysis for non-invasive early inter-turn short-circuit detection in FOC driven induction motors

Abstract With the evolution of device condition monitoring methods, it is possible to make accurate evidence-based decisions and improve predictive maintenance and asset management strategies, thus increasing equipment availability and industrial process profitability. In this way, this paper aims to develop a methodology for the non-invasive detection of incipient short circuit between stator turns of field oriented control (FOC) driven induction motor, as well as enabling the monitoring of the evolution of fault severity. Among the contributions of this work, there is the proposition of a method based on the concept of discrete wavelet transform energy, and a method for selecting their respective decomposition levels that will be used in the composition of the fault detection indexes. The results obtained show that the developed technique is also effective for feedback driven motors (FOC) and are robust in relation to noise and control reaction mechanisms.