基于响应轨迹和核心向量机的电力系统在线暂态稳定评估  

Power System On-line Transient Stability Assessment Based on Response Trajectory and Core Vector Machine

作　　者：王亚俊[1] 王波[1] 唐飞[1] 陈得治[2] 王静[1] 王乙斐[1] 周雨田[1] 

机构地区：[1]武汉大学电气工程学院,湖北省武汉市430072 [2]中国电力科学研究院,北京市海淀区100192

出　　处：《中国电机工程学报》2014年第19期3178-3186,共9页Proceedings of the CSEE

基　　金：国家863高技术基金项目(2011AA05A119);国家电网公司大电网重大专项资助项目(SGCC-MPLG001-029-2012)~~

摘　　要：针对现行调度需求提出基于响应轨迹和核心向量机的电力系统在线暂态稳定评估方法。首先将响应数据构建的原始特征集映射至高维特征空间,然后将暂态稳定评估问题定义为核心向量机中的最小闭包球问题,通过最优近似求解进行故障筛选和快速暂态稳定判别,且离线的训练和在线的匹配保证了暂态稳定评估过程能够满足在线计算的要求。10机39节点和某实际省级电网算例的计算结果表明,所提方法具有更低的时间和空间复杂度,并具有更高的评估精度。On the issue of operation and scheduling, this paper proposed the assessment method for the power system on-line transient stability, based on response trajectory and core vector machine（CVM）. In the assessment, original feature sets structured by response trajectory was mapped to a higher dimensional space, so that the assessment was reformed as minimum enclosing ball problem, whose solution could be achieved by approximation algorithm. Meanwhile, a CVM was used to the off-line training and on-line test process to accomplish contingency screening and classify the system transient status. Experiment results on IEEE New England 39-bus system and a real world system demonstrated the proposed CVM based assessment algorithm has the best precision and the least time and space complexities.

关 键 词：暂态稳定评估 响应轨迹 核心向量机 人工智能 故障筛选 

分 类 号：TM74[电气工程—电力系统及自动化]