Comprehensive Fault Diagnosis of Three-Phase Induction Motors Using Synchronized Multi-Sensor Data Collection

Area of Science:

• Electrical Engineering
• Mechanical Engineering
• Data Science

Background:

• Induction motors are vital for industry but susceptible to mechanical and electrical faults.
• Effective fault diagnosis is crucial for operational reliability and predictive maintenance.
• Existing datasets often lack synchronized multi-sensor data, limiting comprehensive analysis.

Purpose of the Study:

• To introduce a novel, synchronized multi-sensor dataset for three-phase induction motor fault diagnosis.
• To provide a rich resource for developing and validating advanced fault detection algorithms.
• To facilitate research in machine learning for motor health monitoring.

Main Methods:

• Collected synchronized real-time data of vibration, voltage, and current from a 0.2 kW squirrel cage induction motor.
• Utilized high-resolution sensors with a sampling rate of 50 kHz.
• Simulated fault scenarios including phase removal and mechanical misalignments.
• Organized data into ten CSV files representing diverse operational states.

Main Results:

• A Random Forest classifier trained on the dataset achieved 99.82% accuracy in fault diagnosis.
• The synchronized electrical and mechanical data enabled advanced cross-sensor fault analysis.
• Demonstrated the dataset's suitability for real-time fault diagnosis and predictive maintenance.

Conclusions:

• The new dataset is a valuable resource for advancing induction motor fault diagnosis.
• Synchronized multi-sensor data significantly enhances the capability of machine learning models.
• The dataset supports the development of robust predictive maintenance strategies for industrial motors.