An imbalance fault detection method based on data normalization and EMD for marine current turbines

Area of Science:

• Renewable Energy Systems
• Mechanical Engineering
• Signal Processing

Background:

• Marine Current Turbines (MCTs) are crucial for renewable energy, but their reliability is threatened by faults like blade imbalance.
• Environmental factors such as wave and turbulence introduce noise, complicating fault detection in variable speed MCTs.
• Existing methods struggle to isolate fault signatures from background noise and operational frequencies.

Purpose of the Study:

• To develop a robust method for detecting blade imbalance faults in variable speed direct-drive MCTs.
• To enhance the extraction of fault features concealed by supply frequency and environmental noise.
• To validate the method's effectiveness under varying fault severities and turbulence intensities.

Main Methods:

• Utilized Generalized Likelihood Ratio Test (GLRT) for monitoring variable selection.
• Applied data normalization to convert imbalance fault frequencies into a constant.
• Employed Empirical Mode Decomposition (EMD) for filtering noise and isolating fault signatures.
• Analyzed MCT stator current under wave and turbulence conditions.

Main Results:

• The proposed method successfully extracts concealed blade imbalance fault features.
• Data normalization and EMD effectively mitigate the impact of supply frequency and environmental noise.
• Experimental results demonstrate robustness against different turbulence intensities and fault severities.
• The method shows superior feasibility and efficacy compared to existing techniques.

Conclusions:

• The developed fault detection method is effective for variable speed direct-drive MCTs.
• The combination of GLRT, data normalization, and EMD provides a robust solution for imbalance fault detection.
• This approach enhances the operational reliability and maintenance strategies for marine current energy systems.