用于分布式故障诊断系统的新型网络分割法

大规模电力网络的故障诊断问题可由分布式人工智能技术有效地解决。文中提出了一种用于分布式故障诊断系统的有效的网络分割方法，能够将大规模电力网络分割为给定数目的连通子网络，各子网络的故障诊断负担基本相等，同时每个子网络边界元件的数目最小。该网络分割法主要由3部分组成：形成给定电力网络的深度优先搜索树；将网络分割为连通的且计算负担基本平衡的子网络；最小化子网络的边界元件数目以尽可能减小相邻子网络故障诊断的相互重叠，从而节省诊断时间。提出的网络分割法已使用稀疏存储技术编程实现，并在IEEE 14母线、30母线和118母线系统中进行了仿真研究。结果表明该网络分割法可以十分有效地分割大规模电力网络，进而由分布式人工智能技术解决故障诊断问题。

GRAPH PARTITIONING METHOD FOR DISTRIBUTED FAULT SECTION ESTIMATION SYSTEM IN POWER NETWORKS

Fault section estimation (FSE) of large--scale power networks can be implemented effectively by distributedartificial intelligence (Al) techniques. In this paper. an efficient multi--way graph partitioning method is proposed to partitionthe large--scale power networks into a desired number of connected sub--networks with balanced working burdens inperforming FSE. The number of elements at the frontier of each sub--network is minimized in the method as well. Theproposed method consists of three basic steps: forming the weighted depth--first--search tree of the studied power network:partitioning the network into connected, balanced ones and minimizing the number of the frontier nodes of the sub-networksso as to reduce the iteration of FSE among adjacent sub--networks. The method has been implemented with sparse storagetechnique and tested in the IEEE 14--bus. 30--bus and 118-bus systems respectively. Computer simulation results show thatthe proposed multi--way graph partitioning method is very fast and effective for large--scale power system FSE using adistributed Al technique.