Active fault detection using time and frequency diagnostic features for electrical machine

Accurate fault detection in electrical motors is essential for ensuring system reliability and safety. This paper presents an effective diagnosis method for the fault detection of permanent magnet synchronous motors (PMSMs) operating at three different faults over a wide speed and load range. The proposed fault-detection method is based on the extracted features of stator currents from the time and frequency domains. The extracted features are then fed into an ensemble subspace discriminant tree machine-learning algorithm to classify the different types of faults. To validate the efficiency of the proposed approach, experimental tests were conducted on PMSMs operating under different speeds, loads, and fault conditions. The proposed method achieved highly accurate prediction results of 99.6% and could classify five different motor states, including two interturn short-circuit fault states.