量子计算和免疫优化算法相结合的有源配电网故障定位

Fault Location for Active Distribution Network Based on Quantum Computing and Immune Optimization Algorithm

针对现有配电网区段定位算法在有源配电网故障定位时存在的定位耗时长和易陷入局部最优等问题,提出了一种基于量子免疫优化算法的故障定位方法。首先,根据馈线终端单元(feeder terminal unit,FTU)上传至主控制器的故障信息畸变问题,将畸变故障信息的误报、漏报计入故障变量,构建具有容错能力的适应度函数;其次,结合量子计算与免疫算法,减小种群规模,使算法加快收敛;随后,优化克隆扩增和变异操作过程,让陷入或即将陷入局部极值的算子跳出局部最优进行全局搜索;最后,对所提算法在有源配电网分层模型中进行对比实验。结果表明:量子免疫优化算法在单一、多重故障以及含信息畸变情况下的故障定位全局寻优能力较强,故障定位耗时短、准确率高。所提故障定位方法可为解决高性能计算系统在故障定位时存在的复杂度高和容错性差的问题提供参考。

Aiming at the problems of the time-consuming and local optimum of existing distribution network segment location algorithm in the fault location with active distribution network,a fault location method based on quantum immune optimization algorithm is proposed.Firstly,the distortion of the fault information uploaded to the main controller by the feeder terminal unit(FTU)is taken into account,and the fitness function with fault tolerance is constructed by counting the false and missing fault information into the fault variable.Secondly,quantum computing and immune algorithm are combined to reduce the population size and accelerate the convergence of the algorithm.Then,the process of clonal amplification and mutation operation is optimized,so that the operator that has been trapped or is about to fall into the local extremum can jump out of the local optimum for global search.Finally,the proposed algorithm is compared and verified in the active distribution network hierarchical model.The results show that the quantum immune optimization algorithm has significant global optimization ability,short time with fault location and high accuracy of fault location in single,multiple faults and information distortion.The proposed fault location method can provide a reference for solving the problems of high complexity and poor fault tolerance of high performance computing systems during fault location.