"Sustainable Wind Turbine Predictive Maintenance: Detecting Mexican Hats in Wind Speed Data for Efficient Operations"
Master thesis
Published version
Permanent lenke
https://hdl.handle.net/11250/3100865Utgivelsesdato
2023Metadata
Vis full innførselSamlinger
Sammendrag
Wind turbines are critical for renewable energy generation, but unexpected maintenance issues can hinder their efficient operation. Predictive maintenance can mitigate these issues by detecting early signs of turbine degradation. Wind gusts, sudden changes in the wind speed in turbulent wind fields, can cause significant fatigue loads on wind turbines, leading to a shorter lifespan of their components. In addition, power generation oscillations or ramping can result in rapid grid voltage changes, creating further challenges for the stability of the electric grid. In this study, we present an innovative and novel approach for sustainable wind turbine predictive maintenance that leverages a time-series analysis-based AI model to accurately detect specific types of wind gusts, called Mexican Hats, and their characteristics in wind speed data, which may indicate potential turbine performance anomalies. Our approach has the potential to optimize maintenance schedules, reduce downtime, and extend turbine lifespan for more sustainable and efficient operations. The results of this study can contribute to the development of more robust control systems and enhance the reliability of wind turbine operations in the renewable energy sector. The research uses a multivariate time series dataset collected at different heights from offshore platforms in the North Sea (FINO1 platform), which has been made available to OsloMet for research purposes.