Research On Improved Wind Forecasting Method Based On Lorenz Disturbance Model

High precision wind speed and power forecasting method is one of the most critical taches to achieve large-scale integration of wind power into electric system, to improve wind power quality, and to develop the wind power industry. Many existing numerical prediction models, such as statistical models, artificial neural network, gray model etc. have devoted to optimizing the forecasting algorithms. Based on a further study on the influence of nonlinear disturbance on atmospheric dynamical system, this paper proposed a novel short-term wind speed and power prediction method. The primary contents are as follows:(1) The Lorenz system was regarded as an atmospheric disturbance model. The disturbance sequences were obtained by numerically resolving the Lorenz equation. Then define the Lorenz comprehensive disturbance flow with two forms.(2) WNN, BP, and SVM models were taken as the basic forecasting models. A feasibility validation was made for the application of Lorenz disturbance method into short-term wind speed and power predictions.(3) Deal the wind data and the relevant disturbance series with grey generating method and the normalization constant. Then calculate the Lorenz disturbance when the Rayleigh number is equal to 45 and found polynomial disturbance models with less than fifth-degree. Various error criteria were introduced to analyze the performance of forecasting models and prediction results. In order to verify the improvements, other three forms of Lorenz disturbance, when Rayleigh number was separately 0.7, 12, and 16, were employed to the same forecasting procedure.(4) Introduce the disturbance models into grey model GM(1, N) to replace the forecasting process above. Here let N be 1.(5) Considering the great significance of Weibull distribution on wind resource assessment and site selection, the Lorenz wind disturbance model was also used to improve the parameter estimation and achieved better results.