摘要
提出了一种基于数据挖掘和同调分群的相量测量单元的优化配置方法,在对各种运行方式和故障场景进行穷尽式仿真后,通过免疫聚类算法对各种动态场景进行同调性分析,由于故障场景数目庞大,直接计算适合所有场景的分群结果就变得十分复杂,此时再对所有场景下的同调性分析结果先利用粗糙集在保持分类能力不变的条件下删除冗余信息进行场景压缩,最后对剩余场景进行简单组合就可以方便、简单地得到适合各种场景的同调分群方案。仿真分析结果验证了该方法的有效性。
A data mining and coherency analysis based optimal phase measurement unit （PMU） placement is proposed. After the exhaustive simulation on various operating modes and fault scenes, by means of immune clustering algorithm the coherence analysis on contingencies is conducted. Because of the huge number of contingencies it is very complex to calculate the final coherency result suitable for all contingencies. For this reason, firstly the superfluous information in the result of coherence analysis on contingencies is reduced to compress the contingencies by use of rough set under the condition of preserving the classification capacity unchanged, then the residual contingencies are simply combined, at last the coherency identification scheme suitable to various contingencies can be easily and simply deduced. Simulation results verify the effectiveness of the proposed method.
出处
《电网技术》
EI
CSCD
北大核心
2006年第5期49-55,共7页
Power System Technology
关键词
相量测量单元
同调分群
免疫聚类
克隆选择
粗糙集
约简
电力系统稳定与控制
Phase measurement unit
Coherency identification
Immune clustering
Clonal selection
Rough set
Reduction
Stability and control of power system
