Research On Wind Power Prediction Based On Pattern Recognition

In this paper, the methods of wind speed prediction have been summarized and categorized, and its advantages and disadvantages also be analyzed. On this basis, using a novel method based on pattern recognition, named Mycielski algorithm to predict the wind speed of wind farms, and achieve above functions by using string matching algorithms. The wind speed prediction model is established. The model is used to predict short-term wind speed, and the results of the prediction are analyzed, which found that the prediction model has high accuracy in the smooth data interval. The parameters estimation of prediction results are analyzed , which found that the shape parameters and dimensions parameters of Mycielski results have a large similarity with the parameters of measured data, and in line with the similar parameters of the Weibull distribution. The results indicate that the prediction results have the same data composition with the measured data. The analysis of prediction error showed that the Mycielski algorithm prediction error is concentrated in both sides of zero, and submit the normal distribution. On this basis, the six hour wind speed prediction is made, which use the direct prediction and rolling prediction to forecast the six hours wind speed in future. The comparison of two methods showed that the rolling Mycielski have a higher prediction accuracy, and use this prediction results to calculate the wind power.