Research On Wind Power Prediction Based On Atmospheric Disturbance Model

Green wind energy is one of the effective renewable energy sources to replace traditional coal and fossil fuels,as well as to alleviate carbon emissions and environmental impact.However,the wind's random volatility makes wind power facing enormous challenges.Accurate wind power prediction can enhance the stability of wind power grid and reduce the waste of wind energy to the maximum extent.Most traditional prediction algorithms are focused on the improvement of the algorithms,but lack of researching on the essential of wind power.In this paper,the Lorenz system,which characterizes the atmospheric power system,is used as an atmospheric disturbance model to explore the significant improvement effect of Lorenz disturbance on the prediction of wind power.The main work includes the following points:(1)In view of the volatility of wind power,we explore the prediction error sequence of wind power at different time,and estab lish the error correction model;(2)At first,it's going into detail on the four Lorenz chaos attractors of Lorenz equation.Then,it's defining the Lorenz comprehensive disturbance flow by Chebyshev distance and establishing the correspondiong Lorenz disturbance model.Finally,the Lorenz disturbance corrects the preliminary wind speed prediction sequence obtained by neural network(RBF,BP and Elman);(3)Due to the non-stationary characteristics of the original wind speed series,the original wind speed series are divided into trend and detrend of piecewise prediction,namely linear prediction of the trend,the detrend to establish the time series model of wavelet decomposition.When the parameters are same and the initial values of the Lorenz equation are h=(0.01,1,1)and h=(0.001,1,1),the improvement effect of wind speed prediction is explored;(4)Considering the multiple influencing factors of wind speed,the Lorenz disturbance model based on principal component analysis is established.And RBF neural network and least squares support vector machine model are used as contrast models to verify its validity.The establishment of principal component analysis model improves the prediction results of single factor wind speed.Lorenz disturbance model greatly enhances the overall wind speed prediction results.Comparing with traditional wind power prediction,the atmospheric disturbance model makes an obvious improvement in the errors and prediction accuracy,which meets what power grid dispatching needed.The research opens up a new perspective in the wind power forecasting field and promotes the new development of wind power...