Geometric Analysis of Signals for Inference of Multiple Faults in Induction Motors.

Area of Science:

• Electrical Engineering
• Mechanical Engineering
• Signal Processing

Background:

• Induction motors are critical in industrial processes, and unexpected failures incur high costs.
• Existing fault detection systems struggle with motors exhibiting multiple simultaneous faults.
• Accurate identification of multiple faults is crucial for preventing costly downtime and ensuring operational continuity.

Purpose of the Study:

• To present a novel methodology for detecting multiple faults in induction motors.
• To enhance the accuracy and robustness of fault diagnosis systems.
• To address the limitations of current methods in handling complex fault scenarios.

Main Methods:

• Utilizing quaternion signal analysis (QSA) for fault detection.
• Coupling motor current and triaxial accelerometer signals with quaternion coefficients.
• Applying statistical features (mean, variance, kurtosis, etc.) to analyze quaternion rotation.
• Employing four distinct classification algorithms for motor state prediction.

Main Results:

• The QSA method successfully validated ten fault classes, including single (healthy, unbalanced pulley, bearing, half-broken bar) and combined faults.
• The proposed methodology demonstrated high accuracy and superior performance compared to existing state-of-the-art techniques.
• The integration of current and vibration data via QSA proved effective in distinguishing complex fault conditions.

Conclusions:

• Quaternion signal analysis offers a powerful and accurate approach for multiple induction motor fault identification.
• The developed method significantly improves upon current fault diagnosis capabilities, especially for combined faults.
• This technique holds promise for enhancing the reliability and predictive maintenance of industrial induction motors.