Adaptive Neuro-Fuzzy Based Incipient Fault Detection and Diagnosis for Three Phase Induction Motor

Three-phase induction motors are considered as highly promising types of motors in industrial applications. Nevertheless, these motors are prone to various faults that can lead to significant financial losses. The short circuit faults and other stator winding problems can seriously affect the performance of a motor, increasing heat production, decreasing efficiency, and sometimes even damaging the motor. This paper proposed an Adaptive Neuro-Fuzzy Inference System (ANFIS) model to detect stator winding turn fault and their severity. The input to the ANFIS is the features derived from the steady-state three-phase stator current signal using frequency-based analysis. The three-phase stator line current serves as a potent indicator for detecting and diagnosing stator winding turn faults under various load conditions. Numerous experiments were conducted at different load conditions to identify the signatures of the stator turn fault in the stator current. The experimental test outcomes demonstrated a comprehensive accuracy of 98.2% in identifying the location and severity of inter-turn issues. The proposed approach in this paper advances the overall goal of improving motor reliability and maintenance methods.