Research On Health Assessment Of Key Components Of Doubly-Fed Wind Power Generation System

As the source of wind power generation,generator is one of the most important key components of wind turbines,and its state of health directly affects the productivity of wind turbine equipment,if the health assessment of the key components can not be carried out in time and the deterioration degree of the equipment can be judged effectively,it will greatly increase the failure rate and failure time of equipment,and will directly affect the effective output of electric energy and the benefit of power generation enterprises.The object of this study is the doubly-fed induction generator,which is widely used in the wind power industry.The DFIG study is based on the actual situation of the wind resources in the Jinzhou region of Liaoning province,using the real data collected by the SCADA wind turbine operation data monitoring platform,this paper studies the fault analysis and diagnosis,data acquisition and analysis,the selection of the key components and the deterioration trend of the key components of the wind turbine,so as to achieve the health evaluation of the key components of the wind turbine,the main contents of the paper are as follows:(1)The EMD method and PCA method are used to process the characteristic parameters of the generator of wind turbine collected by SCADA system.Through empirical mode decomposition,the local characteristics of the key data affecting wind turbine generator at different time scales can be obtained,so that more detailed but more influential factors can be obtained.Then,the principle component analysis is used to reduce the dimension,which reduces the feature dimension as the input of the short-term and long-term memory neural network without loss of information,in order to improve the efficiency and accuracy of the deep neural network.(2)The WT-VMD-NARX model is proposed.Based on the representative key historical data of the generator parts in the wind power generation SCADA system,the environment wind speed data is divided by Bin method.The deterioration degree model is established according to the characteristic quantity of the key data in the interval.The relatively stationary degree of deterioration signal is extracted by wavelet denoising,and the IMF component is decomposed by VMD,and the value is obtained by central observation method.NARX neural network is used to predict the corresponding IMF component,and the prediction value of the deterioration of wind turbine generator is obtained after superposition.The accuracy of the proposed prediction method is verified by comparison and error analysis.(3)Based on AHP,the weight of target layer,project layer and index layer is calculated by analytic hierarchy process(AHP).Get the weight of each evaluation index,and use the membership function to determine the subjectivity.The degradation degree is used to calculate the membership degree to realize the health evaluation of the generator,which is the key component of the wind turbine.