A Reinforcement Ensemble Learning Method for Rolling Bearing Fault Diagnosis under Variable Work Conditions

Area of Science:

• Mechanical Engineering
• Artificial Intelligence
• Machine Learning

Background:

• Precise fault diagnosis is crucial for rolling bearing operation.
• Identifying fault types under varying working conditions is a significant engineering challenge.

Purpose of the Study:

• To propose a reinforcement ensemble method for diagnosing rolling bearing faults under varying working conditions.

Main Methods:

• A reinforcement model was designed to select optimal base learners.
• Stratified random sampling extracted four datasets for training.
• A sparse Artificial Neural Network (ANN) was designed as the ensemble model.

Main Results:

• Four optimal base learners were obtained through reinforcement learning.
• The constructed reinforcement learning model successfully identified fault types under variable conditions.
• Experiments demonstrated the method's superiority over other intelligent approaches.

Conclusions:

• The proposed reinforcement ensemble method offers significant practical engineering benefits for rolling bearing fault diagnosis.
• The method achieves high accuracy in identifying fault types under diverse working conditions.