Research On Offshore Wind Power Prediction And Operation And Maintenance Strategy Based On Wind Turbine Condition Assessment

China's offshore wind power(OWP)is close to the load center,which has the advantage of local consumption.Thus,OWP has become an important part of China's energy strategy and marine strategy,and has broad development and application prospects.With the large-scale development of domestic offshore wind farms(OWF),the first batch of offshore wind turbines in China has been in operation for more than ten years.As an important power conversion equipment of OWF,the operation conditions of offshore wind turbine(OWT)has a significant impact on the daily operation and maintenance of OWF.On the one hand,the deterioration of operation condition will significantly affect the generation efficiency of OWT,and the research on the influence of external environmental factors on wind power prediction(WPP)without considering the deterioration of internal operation condition will significantly reduce the accuracy of ultra-short-term WPP.On the other hand,condition maintenance is considered to be the most suitable maintenance method for OWT,It can significantly reduce the maintenance cost of OWF.Among them,the accuracy of condition assessment of OWT is the key problem,and the random fault shutdown of OWT will significantly increase the operation and maintenance costs of OWT.In this regard,this paper focuses on the impact of the marine operating environment on the operation condition of OWT,constructs the model of ultra-short-term offshore WPP based on condition assessment,and establishes the short-term maintenance strategy of OWF based on the conditions of OWT.1)Condition assessment model of OWTAs a multi-component system,OWTs operate in the harsh marine environment.On the one hand,the operation of OWTs is affected by the external environment.On the other hand,there must be interaction among various components of OWTs,which makes the traditional condition assessment model cannot to effectively identify the deterioration degree of OWTs and affects the accuracy of the condition assessment model.In this paper,a dynamic deterioration index is proposed,and a prediction model of OWT monitoring indexes based on LSTM(Long Short-Term Memory)neural network is established.Then,the weight of each component index is obtained according to the combined weight.Finally,the operation conditions of OWT is evaluated with fuzzy comprehensive evaluation method.The results show that the proposed method can effectively eliminate the influence of marine environmental factors and the interaction between various components of OWT on the condition assessment results,and has high sensitivity in the process of reflecting the change of the operation state of OWT.2)Modelling of Ultra-short-term offshore WPP based on condition assessment of OWTThe output power of OWT not only depends on wind speed,air pressure and other weather factors,but also is affected by the internal factors of OWT operation condition.The conventional refined processing of input data of WPP model and the improvement of related forecasting algorithm,has some limitations in improving the accuracy of WPP.Considering the operation condition of OWTs,this paper establishes an ultra-short-term offshore WPP model based on CNN-LSTM,and mainly includes two parts: ?1 A multi-state WPP model based on the aggregation and classification of OWF historical operation data is proposed,which effectively reduces the impact of OWT operation conditions in the process of data-based WPP.?2 The prediction algorithm of CNN-LSTM can reduce the influence of weather factors on the accuracy of WPP.The analysis results of a domestic OWF show that the proposed method can effectively improve the accuracy of offshore ultra-shortterm WPP about 2%.3)Modelling of maintenance strategy of OWF based on condition assessment of OWTIn the process of OWF maintenance,the generation loss caused by OWT shutdown and wake effect change is one of the important factors affecting OWF maintenance.Due to the poor accessibility of OWF,random failures of OWTs may lead to significant power loss due to untimely maintenance.Therefore,it is necessary to make full use of the real-time condition information of OWTs to guide the operation and maintenance decision-making and reduce the occurrence of random faults.Considering the conditions information of OWTs,this paper proposes a short-term preventive maintenance strategy for OWF.The main contents are as follows: ?1 A short-term preventive maintenance decision-making model of OWF based on the OWT condition assessment is proposed.In the short-term maintenance process of OWF,the condition information of OWT is effectively combined.?2 an EEMD-Res Net-LSTM based prediction model of average daily wind speed is proposed to obtain the average daily wind speed within 15 working days for short-term preventive maintenance decision-making and reduce the loss of power generation during short-term maintenance of OWF.The example results show that the proposed method can effectively combine the influence of two key factors,condition information and wind speed,and reasonably arrange the preventive maintenance plan for relevant OWTs within a given time.Compared with the conventional maintenance method,the proposed method can reduce the generation loss about 15.3%.