The Prediction Of Wind Power And Pitch Control Based On Support Vector Machine

There are some influences that wind power meking, but the intermittent and random-inherent shortcomings of wind power, with the increasing of wind power involved in the grid is bound to affect the safe operation of the power system and reduce power quality.Generally, pitch-adjustable blades are used in the large wind turbine unit at present, therefore, researching the pitch control technology in depth is of great significance for guaranteeing the safe and optimal running.In the ground of it, the algorithm of wind power short-term power prediction and pitch predictive control is studied in this paper, the following aspects is included:1. The basic principles of support vector machine were analyzed; support vector machine regression algorithm was studied depth and kernel function theory was introduced. In addition, several commonly used support vector machine algorithms were illustrated. Then through the comparison for support vector machine and traditional neural network, the superiority of support vector machine was verified.2. The principle and process of the empirical mode decomposition method was studied. The modeling theory of support vector machine forecasting model was presented. Next forecast models of wind speed and power based on support vector machine were built. Another forecast model based on support vector machine, in which the sequence of wind speed and power is decomposed on the basis of the empirical mode decomposition, was rebuilt. Through simulation and verification to forecast model, it turned out that the prediction accuracy is improved by applying empirical mode decomposition method.3. The GM(1,1) model was improved by analysis the principle of grey model and a new model for wind prediction is built. The new model is compared with the previously model and combined them finally. The last model which combined all single models improved the accuracy of prediction, that the model composite forecast model with different prediction principle can improve effect is showed by the result of simulation.4. The Principle of operation of the wind turbine, wind turbine pitch control strategy and model predictive control theory was studied, the control model in the grid stage was built by the neural network identification, and the fan model of the constant power operation was built by simulation, a predictive controller was designed by genetic algorithms and the support vector machine, the simulation results confirm its effectiveness.