Research On Intelligent Monitoring And Maintenance Of Offshore Wind Farms Based On Data Driven

Offshore wind power is an important pillar of renewable energy development in China,but its marine environment has led to frequent failures of offshore wind turbine components,and the operation and maintenance of faulty components are severely affected by tides.In addition,the lack of dedicated offshore wind power maintenance strategies has led to high operating and maintenance costs for offshore wind farms,greatly limiting the development of the offshore wind power industry,Therefore,this article conducts in-depth and detailed research on the intelligent monitoring and maintenance of offshore wind farms based on data-driven methods,which has important engineering application value.In response to the above issues,the work done in this article is as follows:(1)As a high-risk component of offshore wind turbines,bearings are the research object.Based on data from four operating states of bearings,a data-driven method is proposed.The variational modal decomposition is used to extract features to form the initial feature data set,and a new feature set is formed according to the feature importance score as the input of the fault diagnosis model.Semi supervised learning is used to solve the problem of insufficient labeled data,and the bearing fault diagnosis is completed by combining with random forest.(2)Tide is the environmental factor that has the greatest impact on the offshore wind power when it is out of the sea for maintenance.Taking tide as the research object,using the measured data of a tide station in Shandong,and combining with the time span of operation and maintenance work,an advance multi-step tide prediction method based on random forest is proposed.And compared with traditional methods,it indicates that the method proposed in this thesis can provide more accurate tidal information for offshore operation and maintenance work in the case of small sample data.(3)When the tide meets the requirements for sea operation and maintenance,repair the faulty components of the unit.In order to avoid "over maintenance" and "under maintenance",this article introduces the concept of "opportunity" based on the operational data of offshore wind farms,preventive maintenance and post maintenance.Based on the economic correlation of unit components,the goal is to minimize the number of offshore repairs and maintenance costs,and proposes a component level opportunity maintenance strategy for offshore wind turbines.Compared with other maintenance strategies,the method proposed in this article requires lower maintenance costs.