Research On Prediction Method For Ultra-short-Term Wind Speed In Wind Farm

With the development of wind power,China proposes to support the development of clean energy.In this case,wind power has become a hot spot for researchers.Because of the intermittency and randomness of wind speed,wind power will be instability and volatility.So it will have a negative impact to the power grid.This case will influence the safety of power system.Accurate wind speed forecast can effectively improve the efficiency of wind power grid and it's benefiting for renewable energy's development.So it is very important for us to study the ultra-short-term wind speed forecasting method.First,analysis the parameters of the wind farm.Research the characteristics of wind direction and temperature,and analysis the correlation between air pressure,wind direction,temperature and wind speed.Accurately analysis the wind speed parameter characteristics of the wind farm,and provide data for the wind speed prediction of the KNN model.Second,research on wind speed prediction,by improvement the KNN prediction mode,an ultra-short-term wind speed model based on weighted KNN is proposed.The advantage of the original KNN prediction model in running time is preserved,and improves the prediction accuracy of it.On this basis,the weighted KNN prediction model is further improved,and a KNN prediction model based on similarity weighting is proposed,and the prediction accuracy of wind speed is improved again.Then,research the advantages of a single predictive model.Based on the single prediction model,combined with the advantages of the extreme learning machine and the time series model,an ultra-short-term wind speed prediction model based on error correction is proposed.The model has the advantages of both the time series model and the extreme learning machine model.The model is compared with other prediction models.The results show that the model improves the prediction accuracy of wind speed.Last,according to different evaluation criteria,different forecasting methods are evaluated.In this paper,the wind speed historical data(a wind farm in Northeast China)is used to verify and predict the ultra-short-term wind speed forecasting method used in this paper.The experimental platform is Matlab2016 b.The results show that compared with the traditional wind speed forecasting model,the method proposed in this paper is different from the forecast results.All forecast results are better than before.