Research On Fault Analysis And Diagnosis Method Of Pitch Control System Of A Wind Farm In Yichun

At present,the development of wind power in China has entered a reform stage,and the industrial pattern has different development trends.More and more wind farms pay more attention to the later operation and maintenance of the wind turbine.Especially for the wind farm which has been running for a long time,the mode of operation and maintenance is also changing to lean.How to improve the utilization rate of wind turbine,reduce the failure rate,and ensure that major safety accidents do not occur is a problem that many wind power practitioners are considering and are solving through various methods.Currently,the best choice to solve this problem is the customized change of maintenance mode,the rapid response of fault handling and the diversity of fault handling methods.In this paper,aiming at the scientific problems existing in the fault treatment of a wind farm: the principle of the permanent magnet direct drive wind turbine is difficult to understand,the fault handling mode of the pitch control system is single,the fault identification rate is not high,and it is difficult to deal with it.it is studied by the combination of theory and practice.Firstly,the whole structure and operation principle of GW-1500 k W permanent magnet direct drive wind turbines are analyzed,and some principle contents which are difficult to understand but play an important role in fault analysis and maintenance are analyzed comprehensively.The theoretical basic research on the fault analysis and diagnosis method of the unit is carried out.Secondly,according to the structure and fault characteristics of the pitch control system of the wind turbine,a fault analysis method of the pitch control system based on improved FMEA is proposed.The FMEA method is optimized and improved comprehensively in accordance with the fault characteristics of the variable pitch system,and the corresponding countermeasures are given according to the fault mode of each device,and the RPN value of each fault is calculated,and the fault response measures are carried out according to the value of RPN.Through the use of this method in the wind farm,the practical application effect is better.Thirdly,in order to solve the problem that the fault identification accuracy of pitch control system is not high,the fault identification method of pitch control system based on the combination of LLE and SVM is proposed.In this method,the low-dimensional feature parameter matrix obtained by LLE dimensionality reduction is input into the SVM pattern classifier for working condition recognition,and the recognition accuracy is 93.75%.The superiority and effectiveness of the algorithm are verified.Finally,a fault identification method of SVM pitch control system based on genetic algorithm optimization is proposed.After LLE dimension reduction,the SVM model optimized by genetic algorithm is used to identify and classify the data.After experimental comparison and analysis,using LLE-GA-SVM for fault identification and classification,fault data and normal data have reached 100% recognition accuracy.The results show that the SVM model optimized by genetic algorithm is suitable for fault diagnosis of fan variable pitch system,and the recognition accuracy is improved.It is verified that the establishment of the model is accurate and the algorithm is feasible and effective.