引用本文: | 王彬,郭文鑫,刘文涛,等.基于K-邻近法的电网关键断面在线分布式发现方法[J].电力系统保护与控制,2019,47(7):113-118.[点击复制] |
WANG Bin,GUO Wenxin,LIU Wentao,et al.Power system distributed key section detection online based on K nearest neighbor algorithm[J].Power System Protection and Control,2019,47(7):113-118[点击复制] |
|
摘要: |
随着可再生能源大规模接入电网,电力系统正面临着越来越复杂的运行环境,从而对电网在线操作的时间粒度提出了更高的要求。关键断面在线发现以及其极限传输容量计算是保证大电网在线安全运行的重要手段,调度员通过调控关键断面控制电网安全、稳定运行。从数据驱动的角度出发,对电网在线运行状态建立特征集合,运用层次聚类和分布式特征选择筛选出影响断面出现与否的关键特征,随后运用二分类分布式的K-邻近法(KNN)建立特征集合与断面出现与否的映射。算例分析表明,相比于传统方法,所提机器学习方法可以大大减少关键断面在线发现所需时间,且关键断面预测精度达到工程应用需求。 |
关键词: 关键断面 K-邻近法 分布式 机器学习 数据挖掘 |
DOI:10.7667/PSPC161482 |
投稿时间:2018-11-06修订日期:2019-01-15 |
基金项目:南方电网科技项目(GDKJ00000058)“面向大数据的复杂大电网安全特征选择和知识发现的关键技术与示范应用” |
|
Power system distributed key section detection online based on K nearest neighbor algorithm |
WANG Bin,GUO Wenxin,LIU Wentao,LU Jiangang,XIANG Dejun,ZHOU Zhemin,YU Zhiwen |
(Guangdong Power Grid Power Dispatching Control Center, Guangzhou 510600, China;Qing Da Gao Ke System Control Company, Beijing 100084, China) |
Abstract: |
The renewable energy such like wind energy and solar energy is rather uncertain and intermittent, thus it has raised a more strict requirement for the efficiency of operating. Key section automatic detection and total transfer capability calculation online is the key way to guarantee security of large power grid by operators, thus it has been put much more attention. From the perspective of data-driven, in this paper, feature set is built to describe the state, and then machine learning method is utilized to map the feature set to whether the section exits or not, where KNN method is to play as a classifier. Numerical tests show that the proposed machine learning method can reduce the time needed to discover key sections and have high performance in accuracy compared with traditional methods. This work is supported by the Science and Technology Program of China Southern Power Grid:the Key Technology and Demonstration Application for Security Feature Selection and Knowledge Discovery in Complex Large-scale Power System Based on Big Data (No. GDKJ00000058). |
Key words: key section K nearest neighbor algorithm (KNN) distributed machine learning data mining |