Machine Learning-Based Stator Current Data-Driven PMSM Stator Winding Fault Diagnosis

Area of Science:

• Electrical Engineering
• Machine Learning
• Condition Monitoring

Background:

• Permanent magnet synchronous motors (PMSMs) are critical in modern drive systems.
• Effective fault diagnosis and condition monitoring are essential for PMSM reliability.
• Stator winding faults are a common failure mode in PMSMs.

Purpose of the Study:

• To develop an intelligent method for detecting and classifying stator winding faults in PMSMs.
• To utilize stator current data for automated fault diagnosis.
• To demonstrate the feasibility of an online fault diagnosis system for industrial deployment.

Main Methods:

• Feature extraction from stator phase current symmetrical components using Short-Time Fourier Transform (STFT).
• Fault detection and classification using Support Vector Machine (SVM), Naïve Bayes, and Multilayer Perceptron (MLP) algorithms.
• Online verification and experimental evaluation of the proposed intelligent fault diagnosis system.

Main Results:

• The proposed method effectively detects and classifies stator winding faults in PMSMs.
• Machine learning algorithms accurately identify different fault types based on extracted current features.
• The system demonstrated reliable performance in online verification, showing industrial applicability.

Conclusions:

• The developed intelligent system offers a robust solution for PMSM stator winding fault diagnosis.
• The integration of STFT and machine learning provides an effective approach for condition monitoring.
• The methodology shows significant potential for real-world industrial deployment, enhancing system reliability.