Data driven failure prediction of wind turbine components

The PhD study will investigate how machine learning can be deployed to improve the optimization of offshore wind farm, by to improving the efficiency of data analysis, processing and providing faster response, on top of minimizing high-cost repair operations and wind turbine operational down time. The study will enable effective fault prediction and mitigation system in offshore

wind turbine by implementing machine learning in the process of data analysis for wind turbine mechanical performance data and environmental data to enable accurate prediction of mechanical failure of wind turbine components. A mitigation system shall be designed to benefit the result of prediction system to minimize operational repair cost in the case of predicted damage. Challenges include developing an algorithm capable of providing accurate fault prediction of wind turbine without comprising accuracy due to excessive noise within data, as well as the development of autonomous fault prevention mechanism within wind turbine capable of mitigating potential damage without compromising wind power generation.