Research On Abnormal Identification Of Wind Turbine State Parameters And Evaluation Method Of Operating State

In order to improve the ecological environment and alleviate the shortage of conventional energy supply,we need to develop new energy.In all kinds of new energy,wind energy has attracted the attention of all countries with the characteristics of clean,environmental,pollution-free,inexhaustible and inexhaustible.Wind energy is the most technical and exploitable new energy,the development of the industry is rapid,but the high operation and maintenance cost of wind power equipment also restricts the development of the wind power industry.The traditional maintenance mode is passive maintenance and the maintenance after the failure of the wind power equipment.This maintenance mode is unavoidable economic loss.At present,the development direction of the maintenance mode of wind power equipment is prescient.Maintenance,the predictive maintenance strategy is to find out the abnormal and deterioration trend of the equipment before the failure,and carry out targeted maintenance to avoid the failure,so as to avoid economic loss and improve the economy and security.The abnormal identification of the state parameters of the wind turbine and the evaluation of the running state are the precondition of the predictive maintenance.Therefore,this paper focuses on the abnormal identification of the state parameters of the wind turbine and the evaluation method of the running state.(1)The wind turbine is a complex nonlinear system with strong coupling,and the fault location can not be accurately located when the fault occurs.The relationship of the working conditions of each component is complex.Therefore,this paper first studies the correlation of the working conditions of the wind turbine and measures the correlation method.This paper makes a comparative analysis,and combines the characteristics of the correlation coefficient of Pearson,Kendall and Spearman,puts forward the overall correlation index,and uses the overall correlation index to measure the correlation between the parameters of the state parameters.(2)The identification of state parameters of wind turbines is studied.With the rapid progress of sensor technology,the monitoring of the state of the wind turbine will be more comprehensive,the type and quantity of the sensors will be more,and the state parameters of the wind turbine will be more.The number of the input parameters is simplified on the basis of the degree.The algorithm of Least squares support vector machines(abbreviated LS-SVM),backpropagation neural network(Back-propagation neural network,BPNN)and radial basis function neural network(Radial basis)are studied and then built.The combined forecasting model based on three kinds of algorithms,combined with the forecast information entropy residuals established parameter anomaly discriminant model.The two sub models constitute theanomaly identification model of work condition index.Finally,an example is given to analyze the model,which verifies the accuracy and effectiveness of the model.(3)Cloud model is studied.Based on cloud model,the evaluation method of overall operation state of wind turbines is proposed.First,the evaluation grade of the operating state of the unit is established,and the parameters of the unit are evaluated,and the weight is determined by the subjective and objective empowerment method,and then the membership degree cloud model is established.Finally,the fuzzy comprehensive evaluation is carried out.The evaluation model of the work condition evaluation model of a wind farm is evaluated and compared with the other two evaluation methods,which verifies the accuracy and effectiveness of the proposed model.