Automatic Modal Parameter Identification for Offshore Wind Turbines Using Modified Clustering-Based Methodology.

Area of Science:

• Engineering
• Renewable Energy Systems
• Data Science

Background:

• Offshore wind turbines (OWTs) are crucial for carbon neutrality goals.
• Monitoring OWTs' dynamic response is vital for safe operation and modal parameter assessment.
• Accurate modal parameter identification is essential due to limited natural frequency ranges under dynamic loads.

Purpose of the Study:

• To introduce a novel machine learning method for automated modal parameter identification of OWTs.
• To combine Stochastic Subspace Identification (SSI-data) with DBSCAN and K-means clustering.
• To enable automatic determination of K-means clusters for enhanced analysis.

Main Methods:

• Utilized SSI-data for modal parameter identification.
• Integrated DBSCAN and K-means clustering algorithms.
• Validated the method using a four-degree-of-freedom model and Opensees numerical simulation.

Main Results:

• The proposed method demonstrated high accuracy for automated modal parameter identification.
• Frequency identification differences were 0.0%, 0.30%, and 0.18% for the first three orders compared to benchmark results.
• The method successfully validated through theoretical analysis and numerical simulations.

Conclusions:

• The developed machine learning approach provides accurate automated modal parameter identification for OWTs.
• Offers valuable insights for dynamic monitoring professionals in the offshore wind sector.
• Provides technical support for diagnosing abnormal states in OWTs using dynamic response data.