Research On Condition Monitoring Of Key Equipment Of Wind Turbine Based On SCADA

Wind turbine operating conditions are easily affected by natural environment and wind stochastic volatility,unit operating conditions directly affect the relevant operating parameters of wind turbines,and high reliability is the basis of stable and economic operation of wind turbine units.The key parts such as gearbox,generator and main bearing belong to the high-risk part of the fault,and effective monitoring can reduce the downtime.Supervisory control and data acquisition(SCADA)is used to monitor and record the operating status of wind turbines in real time.The use of SCADA data to monitor the high frequency of fault parts,without additional sensors,can effectively predict the operating status of parameters related wind turbine.In this paper,it focuses on the sample data under normal operating conditions related to the target state parameters for the high-frequency fault location of the wind turbine.Before modeling,the data are preprocessed,a temperature prediction model is established by using the nonlinear state estimation technique(NSET)to predict the temperature and identify the state of the key parts of the unit.Based on the similarity principle,comparing with the NSET model based on fixed interval and the improved NSET model based on Mahalanobis distance,which shows that the improved NSET model is more accurate and the prediction effect is better.The effect of training data with different sample data types on the prediction accuracy of the model is analyzed.Comparing with the BP neural network prediction model,some evaluation indexes are established to highlight the superiority of the improved NSET model.Combining with the NSET model and the BPNN model,the state parameter combination forecasting model is proposed.Each evaluation index shows that the combined forecasting model has higher accuracy.For the faulty generator set of the wind farm,the validity of the model built in this paper is verified.According to the overall operation status of wind turbine,the evaluation index system of wind turbine is established by AHP,and different evaluation levels are established.Introducing the theory of variable weight,and based on the fuzzy comprehensive evaluation method,the operating states of gearbox,generator and main bearing of wind turbine are evaluated to realize the overall judgment of abnormalities.According to the Markov process,the short-term reliability prediction model of wind turbines is established.Based on the evaluation results of the running status of wind turbines,the probability of outage of the turbines is predicted.