Chaotic time series prediction model of wind power

In order to reveal the interval laws of wind power time series, the phase space reconstruction which is on the basis of chaotic time series theory is used to identify the chaotic feature of wind power time series. Considering the different effects of different coordinate of phase points on predicted point, this pa per improves the distance criterion and the evolutional trend criterion by weighting. In addition, this paper proposes an improved local Volterra adaptive filter to predict wind power by proposing a comprehensive criterion to select the neighbor points as the training set. The simulation of the measured data of a certain wind farm shows the proposed model is accurate and fast.