Short-circuit diagnosis in induction motor based on Frequency Response Analysis (FRA) and statistical indices

Frequency Response Analysis (FRA) is widely recognized as an effective diagnostic technique for detecting faults in power transformers. However, its application to fault diagnosis in rotating electrical machines, such as induction motors (IMs), remains relatively unexplored. This study experimentally investigates the feasibility of applying FRA to identify electrical faults in induction motors. Specifically, the frequency responses of two short-circuit fault types—turn-to-turn (TT) and turn-to-ground (TG)—were examined. These faults were deliberately introduced into the IM windings under controlled conditions, with varying resistance levels to simulate different fault severities. The main contribution of this work is a comparative experimental evaluation of five curve-based and two resonance-based statistical indices across different frequency regions, with respect to fault detection, discrimination between TT and TG faults, and assessment of fault severity. The experimental results reveal that these indices offer a reliable means for detecting and characterizing electrical faults in rotating machines. Additionally, a notable correlation was observed between the type of fault and the number of resonance points in the frequency response, providing further diagnostic insight. These complementary indicators are subsequently integrated into a preliminary machine-specific diagnostic framework for fault detection, type discrimination, and severity assessment.