Power Network Partitioning Based on Clustering Analysis in Mvar Control Space

Qing-lai GUO,Hong-bin SUN,Bo-ming ZHANG,Wen-chuan WU
DOI: https://doi.org/10.3321/j.issn:1000-1026.2005.10.008
2005-01-01
Abstract:The shortcoming of traditional definition of electrical distance recently used in power network partitioning for secondary voltage control is discussed. Based on a quasi-steady-state sensitivity of voltage w.r.t. Mvar output, a concept of Mvar control space is presented. In this control space, the sensitivity of a load node voltage to a Mvar source reactive power is defined as the component of the coordinate for the load node, which denotes the control ability of certain reactive power source, and by this way the problem of power network partitioning can be transformed into a clustering analysis problem in a high-dimension space. Electrical distance on this space is defined and a hierarchical clustering method is adopted to solve the problem of power network decomposition. Finally, some digital test results for an IEEE 39-bus system and a real-life power network are presented. This method has been applied to a secondary voltage control system in a provincial control center in China.
What problem does this paper attempt to address?