Stator Winding Faults Diagnosis of Three-Phase Induction Motor using Adaptive Centroid Defuzzification Based Fuzzy Inference System

Objective: To develop a reliable diagnostic approach for induction motors capable of identifying incipient stator winding faults under variable operating conditions, thereby enhancing predictive maintenance and minimizing downtime. Method: A laboratory setup was established to simulate both healthy and faulty stator conditions in a three-phase induction motor. Voltage and current signals were acquired and preprocessed to extract significant features. A Mamdani fuzzy inference system integrated with an adaptive centroid defuzzification approach was implemented in a MATLAB-based GUI for incipient fault monitoring. Findings: The proposed diagnostic framework achieved an accuracy of 99.2%, precision of 98.7%, and fault tolerance of 95.6% in detecting stator winding defects. These results demonstrate the method’s robustness and reliability under variable and noisy operating conditions. Novelty: This manuscript introduces an optimized fuzzy-based diagnostic model combining adaptive centroid defuzzification with enhanced feature preprocessing. This integration significantly improves classification accuracy, decision sensitivity, and fault tolerance, establishing a powerful tool for intelligent predictive maintenance of induction motors. Keywords: Induction motor, Mamdani fuzzy inference system, Adaptive centroid defuzzification, MATLAB GUI, Stator winding fault diagnosis, Predictive maintenance