Abstract: As the application of the Distributed Generators（DG） and the expansion of the power distribution network structure, the establishment of switch function in locating the fault points becomes extremely complex, and the search speed of traditional algorithms needs to be improved eagerly. In order to satisfy the current situation of DG access and deal with multiple distributed generators with multiple faults, the model is based on improved switching function in this paper. A network structure dimension reduction method for DG distribution network is raised, which can increase the calculation speed in fault points locate. The paper introduces the step of bat algorithm in the distribution network location. The example simulation comparison shows that the bat algorithm has a quicker speed and a more accurate global search capability than other artificial intelligence algorithms.

Key words: distribution network, fault location, distributed generator, bat algorithm

摘要： 针对分布式电源的大量应用与配电网结构的扩大，使故障定位开关函数的建立愈加复杂，传统算法的搜索速度与准确性亟待提高。通过建立改进的开关函数模型，能够满足目前分布式电源广泛接入的现状，处理多DG、多故障的定位要求。通过配电网结构划分降维策略进一步增加了算法的计算速度，建立改进的评价函数，防止故障定位误判。介绍了蝙蝠算法求解配电网故障定位的具体步骤。通过算例进行仿真对比表明，蝙蝠算法相比其他人工智能算法在配电网单故障与多故障定位方面具有更加快速、准确的全局寻优能力。

关键词: 配电网, 故障定位, 分布式电源, 蝙蝠算法