Ultra-short Term Wind Power Forecast Model Based On Improved WD-ASD

With the massive growth of wind power installed capacity in the globe,the intermittent,volatility and randomness of wind power bring great challenges to the power grid economic dispatch.In order to meet the dispatch department's demand of high-precision wind power forecasting information,this paper proposes an ultra-short term wind power forecast model based on an improved wavelet decomposition-atom sparse decomposition algorithm.Based on the wavelet decomposition,the historical wind power data are divided into high and low frequency components,and it is further decomposed into a group of atomic components with specific expressions by the atomic sparse decomposition method.Then the self-prediction and the least squares support vector machine method are used to predict the accurate point prediction.For improving the convergence performance and calculation speed of atomic sparse decomposition,a bacterial colony chemotaxis algorithm combined with orthogonal matching pursuit algorithm is proposed to optimize the real-time prediction performance.At the same time,for the linear and non-stationary characteristics of wind power sequence in different regions,the attenuated linear atomic and Gabor atomic are constructed to achieve the effect of adaptive decomposition.Finally,the historical wind power data are divided into multiple fluctuation intervals by the swinging door and the two-dimensional kernel density estimation model considering wind power fluctuation is constructed based on point forecasts.It can scroll to obtain the confidence interval of the predicted value and reduce the impact of environmental changing.The results show that the proposed model based on the improved wavelet decomposition-atom sparse decomposition algorithm prediction model and the two-dimensional kernel density estimation model considering wind power fluctuation are more accurate than other models.It can provide more certain and uncertain information for power grid dispatch,which can reducing system operating cost and risk cost.