Research On Operation Condition Assessment Method Of Direct-Drive Wind Turbine Generator System Based On SCADA Data

Wind power generation is a mature and widely used power generation method.Compared with doubly-fed wind turbine generator system,direct-driven wind turbine generator system accounts for more and more of the newly installed wind power capacity in China due to its advantages such as low failure rate,high power generation efficiency and good grid-connection performance.The implementation of condition monitoring and health status evaluation of wind turbine generator system can comprehensively evaluate the advantages and disadvantages of wind turbine generator system operation status,timely,scientific and reasonable maintenance,which is of great significance to ensure the stable and safe operation of wind turbine generator system,increase the operation life and improve the economic benefits of wind farms.In this paper,the evaluation methods of wind turbine generator system component level and complete machine operation state are studied respectively,which mainly contains three parts:(1)Supervisory Control and Data Acquisition(SCADA)data preprocessing.In this paper,the data preprocessing method of SCADA system is firstly explored.Considering that wind power curve is the most direct embodiment of wind turbine generator system performance,k-means clustering and Mahalanovic distance are combined to conduct cluster analysis on wind power scatter chart,and data filtering criteria are formulated to eliminate abnormal points in SCADA data.(2)Component-level health evaluation of direct-driven wind turbine generator system.Taking wind speed,generator speed and active power as input parameters,fuzzy C-means clustering is used to realize the classification and identification of wind turbine generator system operating conditions.The quartile method is used to select the health data of wind turbine operating subspace state parameters.The maximum mean difference is used to measure the probability distribution difference between the actual operating data of wind turbine generator system and the sub-space health data of working conditions,and the health index of the indicators is constructed to quantificat the deterioration degree of state parameters and realize the evaluation of wind turbine generator system component operating status.Finally,the SCADA data of a 2.5MW direct-driven wind turbine generator system is used to verify the method.(3)Assessment of the operation status of direct-driven wind turbines generator system.The SCADA system monitoring indicators of wind turbine generator system is analyzed to build the multi-level evaluation indicators system of the whole operation state.The analytic hierarchy process and entropy weight method are used to calculate the subjective weight and objective weight of the indicators.The multiplication integration method is used to calculate the combined constant weight of the indicators,and the improved variable weight formula is used to calculate the variable weight of the indicators.The operating state of wind turbine generator system is divided into good,mild deterioration,severe deterioration,major deterioration and failure,and the whole operating state of wind turbines generator system is evaluated by set pair analysis theory.When evaluating the state grade of wind turbine generator system,the support matrix is constructed by partial connection degree,and the operating state grade of wind turbine generator system is determined according to the maximum support criterion.The SCADA data of a 2.5MW direct-driven wind turbine generator system is used to verify the effectiveness of the proposed method.