摘要
综述了基于人工智能技术实现电网稳定评估的步骤、关键环节。总结了国内外在候选特征集的组成方式、关键特征的形成方法、智能稳定评估技术的选择等方面的研究进展。指出了动态输入特征难于提供电网智能决策所需信息、多数智能型稳定评估技术的可解释性及适应性差等不足,并提出原始输入特征应以电网实时状态信息为主、输入空间的裁减技术应以嵌入式特征方法为主要研究方向,及研究重点应为基于知识发现技术的稳定评估算法等建议。
A survey on procedures and key links for power network stability assessment based on artificial intelligence is given. The advances in the research on composition modes of initial feature set, formation approaches of key features and selection of intelligent stability assessment technology home and abroad are summarized. The defects such as difficult to offer necessary intelligent decision-making information required for power network by dynamic input features, weak interpretability and adaptability of most intelligent stability assessment technologies are pointed out. Some suggestions are proposed, they are: for initial input features the first place should be given to real-time state information of power network, for the reduction of input space the embedded feature searching should be taken as principal research direction and the major research emphasis should be paid to the stability assessment algorithm based on knowledge discovery technology and so on.
出处
《电网技术》
EI
CSCD
北大核心
2009年第12期60-65,71,共7页
Power System Technology
基金
国家自然科学基金资助项目(50407014)~~
关键词
电力系统
稳定评估
人工智能
特征选择
知识发现
power system
intelligence
feature selection
security assessment
artificial knowledge discovery