An Experimental Study on Condition Diagnosis for Thrust Bearings in Oscillating Water Column Type Wave Power Systems

Area of Science:

• Renewable Energy Engineering
• Mechanical Systems Monitoring
• Machine Learning Applications

Background:

• Wave energy systems require high operational efficiency and reliability.
• Effective fault diagnosis is crucial for managing costs and maximizing facility utilization.
• Existing fault diagnosis methods need enhancement for wave power systems.

Purpose of the Study:

• To develop and validate a fault diagnosis and monitoring method for wave power systems.
• To analyze failure modes and reproduce specific component faults.
• To apply machine learning for classifying operational states and identifying faults.

Main Methods:

• Failure Mode and Effect Analysis (FMEA) to identify system failure modes.
• Construction of a test bench for fault reproduction, focusing on thrust bearings.
• Vibration spectrum analysis of test data to extract features.
• Classification of data using Naïve Bayes (NB), k-Nearest Neighbor (k-NN), and Multi-Layer Perceptron (MLP) algorithms.

Main Results:

• Identification of critical failure modes in wave power systems.
• Successful reproduction of thrust bearing faults using dedicated test equipment.
• Extraction of distinct vibration features for different operating statuses.
• Classification accuracy achieved using NB, k-NN, and MLP algorithms.

Conclusions:

• The developed method provides a standard for fault monitoring in wave power systems.
• Machine learning-based analysis of vibration data enables efficient fault diagnosis.
• This approach enhances the reliability and operational efficiency of wave energy converters.