This paper develops a new hybrid method based on an improved ant colony optimization algorithm that incorporates pattern search (IACO-PS) for determining the location of faults in a distribution network. The performance of the conventional ant colony optimization (ACO) algorithm is improved using the opposite-based learning strategy to generate the initial population and adding a weight coefficient into the pheromone update mechanism to dynamically adjust the pheromone volatilization factor. The hybrid IACO-PS algorithm combines the individual strengths of ACO and PS. In addition, the fitness function is constructed by counting the false and missing fault information into the fault variable. In optimizing benchmark function experiments, the proposed hybrid IACO-PS presents a superior performance when compared to other improved versions of ACO. The effectiveness of the proposed approach is corroborated by tests performed on an IEEE 134-bus network. Simulation results show that the proposed hybrid IACO-PS method can determine the location of a fault even in the presence of fault distortion. In addition, it is immune to noise and data loss errors. Finally, the method proposed in this paper significantly outperforms other published fault location methods, and it can accurately locate faults and identify the type of distortion.

本文基于改进的蚁群优化算法开发了一种新的混合方法，该算法结合了模式搜索 (IACO-PS)，用于确定配电网络中的故障位置。改进了传统蚁群优化算法（ACO）的性能，采用基于对立的学习策略生成初始种群，并在信息素更新机制中加入权重系数，动态调整信息素挥发因子。混合 IACO-PS 算法结合了 ACO 和 PS 的各自优势。此外，适应度函数是通过将错误和缺失的故障信息计入故障变量来构建的。在优化基准函数实验中，与其他改进版本的 ACO 相比，所提出的混合 IACO-PS 表现出优越的性能。在 IEEE 134 总线网络上进行的测试证实了所提出方法的有效性。仿真结果表明，即使存在故障失真，所提出的混合IACO-PS方法也可以确定故障的位置。此外，它不受噪声和数据丢失错误的影响。最后，本文提出的方法明显优于其他已发表的故障定位方法，可以准确定位故障并识别失真类型。