电力系统及其自动化学报2017，Vol.29Issue(8)：22-28,7.DOI:10.3969/j.issn.1003-8930.2017.08.004
基于改进高斯过程回归模型的短期负荷区间预测
Short-term Load Interval Prediction Based on Improved Gaussian Process Regression Model
摘要
Abstract
Considering that the accuracy of short-term load prediction directly affect the economical efficiency and secu-rity of power grid operation,while the traditional point prediction method cannot take many uncertain factors into consid-eration,an interval prediction model of short-term load based on improved Gaussian process regression is proposed in this paper. Fuzzy C-means clustering algorithm is used to search for similar days from the historical data to constitute a more reasonable sample set,and multiple kernel covariance functions are used to improve the conventional Gaussian process regression algorithm. Finally,the interval prediction results can be obtained at a certain confidence level. Re-sults of a real numerical example indicate that compared with the conventional methods,the proposed method has rela-tively higher prediction accuracy and higher interval coverage probability,which is suitable for engineering applications.关键词
区间预测/高斯过程回归/电力系统短期负荷/多核协方差函数/聚类分析Key words
interval prediction/Gaussian process regression/short-term load of power system/multiple kernel covari⁃ance function/clustering analysis分类
信息技术与安全科学引用本文复制引用
宗文婷,卫志农,孙国强,李慧杰,CHEUNGKwokW,孙永辉..基于改进高斯过程回归模型的短期负荷区间预测[J].电力系统及其自动化学报,2017,29(8):22-28,7.基金项目
国家自然科学基金资助项目(51107032、61104045、51277052) （51107032、61104045、51277052）
