基于集对分析聚类法的超短期风电功率区间预测

Ultra short term wind power interval prediction with set pair analysis in cluster analysis

高精度的风电功率预测是保证含高渗透率风电电力系统安全经济运行的重要手段。文章在传统ARIMA算法的基础上,引入集对分析理论对风电功率进行超短期区间滚动预测。首先采用改进的K-means算法,建立风电功率与风速、风向之间的集对关系;在点预测结果的基础上,估计区间上下限,经过误差调整,最后得到区间预测结果。文章引入3个模型评价指标对不同方法进行比较。算例表明,所提出的基于集对分析聚类算法的超短期风电功率区间预测能够得到更精确的预测区间。

High-precision wind power prediction is an important means to ensure the safe and economical operation of wind power system with high permeability.Now the point prediction method is various,prediction accuracy is difficult to improve,so in this paper we presents a method for wind power interval prediction based on the set pair analysis theory with the ARIMA algorithm.Firstly,combined with K-means clustering algorithm,the clustering evaluation function is established and get wind power clustering results;Establish the relationship between the wind power and influence factors;For a new wind power data,calculate the distance with each class and find the cluster＇s upper limit and lower limit.According to the range of the error distribution,adjust the wind power interval and can get the final interval prediction result.Compare the method with confidence interval method,and introduce three evaluation index model,the effectiveness of the wind power interval prediction based on the set pair analysis theory is verified.