Integrated Modeling and Machine Learning Based Fault Diagnosis of BLDC Motor Drive for EV Applications

The rapid adoption of Electric Vehicles (EVs) has strengthened the demand for reliable and fault-tolerant motor drive systems. Among various motor technologies, Brushless Direct Current (BLDC) motors are widely employed in EVs owing to their high efficiency, compact structure, and minimal maintenance requirements. Nevertheless, BLDC motors are susceptible to multiple fault modes that adversely affect performance, operational safety, and service lifetime. This paper presents a comprehensive investigation into Condition Monitoring (CM) and Fault Diagnosis (FD) of BLDC motor drives. Both internal faults, including inter-turn short circuits, phase-to-phase short circuits, and open-circuit failures, as well as external inverter–motor connection faults such as single-phase, double-phase, and ground faults, are systematically analyzed. MATLAB/Simulink-based simulation is conducted to evaluate the performance degradation under faulty conditions. Fault characterization is performed using diagnostic approaches encompassing signal-based, model-based, and data-driven techniques. Supervised machine learning algorithms – including Support Vector Machines (SVM), K-Nearest Neighbors (KNN), decision trees, and neural networks – are employed for fault classification. Experimental validation confirms the efficacy of the proposed framework, achieving high diagnostic accuracy and demonstrating its practical relevance for enhancing the reliability of BLDC motor drives in EV applications.