An Explainable AI-Based Fault Diagnosis Model for Bearings

Area of Science:

• Mechanical Engineering
• Artificial Intelligence
• Signal Processing

Background:

• Bearing faults are critical in rotating machinery, leading to equipment failure.
• Accurate fault diagnosis is essential for predictive maintenance and operational safety.
• Existing AI models often lack transparency in their decision-making processes.

Purpose of the Study:

• To propose a novel explainable AI-based fault diagnosis model for bearings.
• To enhance the interpretability of AI models in bearing fault diagnosis.
• To improve the accuracy and reliability of fault detection under variable operating conditions.

Main Methods:

• Data preprocessing using Stockwell Transformation Coefficient (STC) for vibration signal analysis.
• Statistical feature extraction to identify significant patterns from STC data.
• Explainable feature selection using Boruta and multicollinearity avoidance.
• Fault diagnosis and explanation using k-Nearest Neighbors (k-NN) with additive Shapley values.

Main Results:

• Demonstrated effectiveness of the proposed model on two distinct bearing datasets.
• Successfully incorporated explainability into both feature selection and classifier interpretation.
• The model accurately diagnoses bearing faults and provides explanations for individual predictions.

Conclusions:

• The developed explainable AI model offers a significant advancement in bearing fault diagnosis.
• Explainability in feature selection and classification enhances model transparency and debugging capabilities.
• The approach provides a robust framework for reliable fault diagnosis in rotating machinery.