Detecting industrial motor faults with current signatures

Area of Science:

• Electrical Engineering
• Industrial Automation
• Predictive Maintenance

Background:

• Induction motors are crucial in industry but prone to wear and tear, necessitating frequent maintenance.
• Current monitoring methods for three-phase motors are often complex, requiring multiple sensors and leading to significant downtime.
• Early detection of motor faults is vital to prevent minor issues from escalating into major failures, reducing costs and production interruptions.

Purpose of the Study:

• To propose a novel, cost-effective method for monitoring faults in three-phase industrial motors.
• To reduce the number of sensors required for accurate fault detection.
• To develop a universal fault detection system applicable to various motor configurations (delta/star) and operating conditions.

Main Methods:

• Utilized the Hilbert Transform (HT) to analyze the instantaneous current signature of three-phase motors.
• Developed a system incorporating normalized HT curves into a fault analysis database for accurate signature recognition.
• Reduced sensor requirements by focusing solely on the current signature, eliminating the need for voltage sensors.

Main Results:

• The proposed HT instantaneous current signature method demonstrated accurate fault detection capabilities.
• Experimental results showed close resemblance to the benchmark V and I Lissajous trajectory curve method.
• The HT method achieved a 50% reduction in sensor costs compared to existing techniques.

Conclusions:

• The Hilbert Transform instantaneous current signature offers a universal and cost-effective solution for three-phase motor fault monitoring.
• The system's accuracy and reduced sensor count present a significant advantage over traditional methods.
• This approach can enhance predictive maintenance strategies, minimizing industrial downtime and repair expenses.