Analysis And Research On The Operational Health Of Wind Turbines Based On Monitoring Data

With the rapid development of wind power industry,the single unit capacity and total installed capacity of wind turbines are increasing year by year.Most of the wind farms are located in the plateau or coastal areas with rich wind resources.The working environment is complex and the operation and maintenance of wind turbines are difficult.Once the wind turbine fails,it will not only affect the stability of power supply,but also increase the cost of operation and maintenance,resulting in serious property losses.At present,most wind farms still adopt the traditional operation and maintenance strategy of planned maintenance and post maintenance,and the operation and maintenance cost is high.If we can judge the healthy state of the wind turbine operation according to the analysis of the monitoring data of the SCADA(Supervisory Control and Data Acquisition)system of the wind turbine,and find the potential faults of the wind turbine operation,And make an accurate evaluation of the health of the wind turbine,so as to reasonably arrange the maintenance and repair plan,improve the stability of wind turbine operation,reduce the failure rate.According to the monitoring data of wind turbine,this paper carries out the analysis and Research on the operation health of wind turbine,aiming to realize the effective monitoring and evaluation of the health status of the key components and the whole machine of wind turbine.The main work of this paper is divided into the following three aspects according to the research of single-component multivariable,multi-component univariate and multi-component multivariable inside the wind turbine:(1)In order to quantify the operational health of key components of wind turbine,the assessment of operational health of key components of wind turbine is carried out,and the early warning information is generated in time after the deterioration trend of component operation,so as to avoid the secondary failure of component evolution.This part is based on the prediction residual analysis method.Firstly,the monitoring data of the wind turbine is preprocessed.Secondly,the Bayesian optimized e-lightgbm(Ensemble Light Gradient Boosting Machine)method is used to train the multi-point temperature setting prediction model of the key components,and the prediction residual of each measuring point is calculated.And a GMM(Gaussian Mixture Model)based operational health evaluation benchmark model of key components is established.Finally,the simulation case of gearbox temperature data set shows that the proposed model can reflect the health degree of components.generate early warning information in advance,and verify the reliability of the evaluation model.(2)The purpose of this work is to quantify the operation health of the nacelle composed of multiple subsystems,identify the abnormal temperature rise trend of nacelle temperature,and reduce the operation time of nacelle components under extreme temperature and unhealthy operation conditions.In this part,based on the characteristic data set of nacelle temperature,a prediction model of nacelle temperature based on MIC(Maximum Information Coefficient)and LSTM(Long Short-Term Memory)network is proposed.On this basis,a temperature uncertainty interval prediction method based on CKDE(Conditional Kernel Density Estimation)is proposed.Combined with the results of fixed value prediction and interval prediction,the health quantitative index calculation method considering the wind speed interval division is used to establish the nacelle health evaluation model.The final simulation results show the effectiveness of the model.(3)To carry out the health evaluation of wind turbine operation,the purpose is to make an overall evaluation of the current operation status of the wind turbine by integrating the multi system and multi monitoring characteristic information of the wind turbine operation.In this part,firstly,the monitoring characteristics of the three subsystems of the wind turbine are selected,and the weight matrix in the fuzzy comprehensive evaluation is determined by factor analysis(FA).Then,the membership matrix is determined according to the deterioration index of the operation parameters,and the comprehensive evaluation results of the wind turbine operation Health degree are obtained by fuzzy operation.Finally,according to the operation data of SCADA system of grid connected wind turbine.the proposed model is applied to simulate,which shows the feasibility of this method.