A deep learning approach for electric motor fault diagnosis based on modified InceptionV3

Area of Science:

• Electrical Engineering
• Computer Science
• Artificial Intelligence

Background:

• Electric motors are critical in many industries but susceptible to failures from operational stress and poor maintenance.
• Early and accurate fault detection is crucial for preventing costly downtime and ensuring operational safety.

Purpose of the Study:

• To develop and evaluate a novel thermography-based method for detecting electric motor faults.
• To improve the accuracy and efficiency of motor fault diagnosis using deep learning.

Main Methods:

• A thermography-based approach utilizing the InceptionV3 deep learning model.
• Application of Contrast Limited Adaptive Histogram Equalization (CLAHE) for image enhancement.
• Integration of a Squeeze-and-Excitation (SE) channel attention mechanism to boost InceptionV3 performance.
• Utilizing a dataset of 369 thermal images covering 11 fault types, augmented for increased data size.
• Employing fivefold cross-validation for robust evaluation.
• An alternative approach using InceptionV3 for feature extraction combined with Support Vector Machines (SVM) for classification.

Main Results:

• The proposed InceptionV3 with SE mechanism achieved high performance metrics: 98.82% accuracy, 98.93% precision, 98.82% recall, and 98.87% F1 score.
• The hybrid InceptionV3-SVM model demonstrated a perfect 100% detection rate across all evaluation metrics.
• The methods proved effective in classifying various electric motor faults from thermal images.

Conclusions:

• The developed thermography-based deep learning method, particularly the InceptionV3 with SE attention, significantly enhances electric motor fault detection.
• Combining deep learning feature extraction with traditional classifiers like SVM offers a highly accurate and robust solution for industrial motor fault diagnosis.
• This research provides a valuable tool for predictive maintenance and improving the reliability of electric motor systems.