Wind power has the disadvantages of intermittence and randomicity, which will bring challenge to the safety and stabilization of power grid and power quality, with the increasing proportion of wind connected to grid. Short-term wind power prediction is an effective approach to the above problem. This paper studies short-term wind and power prediction. The statistical rules of parameters from the wind farm in North China are acquired after pretreatment. The paper analyses wind speed's Chaotic characteristic, and then establishes the six-hours-advance prediction model base on neural network, which adopts the gradient descent algorithm with the amount factor, effectively improves the convergence speed and the shortcoming of easy fall into local minimum value. The wind power is predicted with two approaches. One is based on wind-power curve, the other is wind power directly prediction using RBF network. The paper presents a integrate wind power prediction method. The simulation results that the proposed methods have a better global forecasting performance.
|