基于模糊微分演化算法的配电网综合规划

针对配电网综合规划问题提出一种模糊微分演化算法。设计了模糊控制器来自适应调整演化参数中的变异缩放因子和交叉概率常数，避免了迭代陷入局部最优解，同时加快了演化速度。为避免盲目搜索设计了一种保持配电网辐射约束的初始化染色体编码方法，并对演化过程产生的不可行个体的修复代价进行估计，尽可能修复成可行个体。在潮流校验时，将网络拓扑分析得到的母线-支路编号直接用于潮流计算，并尽量将结线分析限制在结构发生变化的馈线及与该馈线相关的电气岛内。对IEEE 50节点算例系统的仿真分析表明，文中算法能有效提高收敛速度并明显改善收敛性能。

Application of Fuzzy Differential Evolution Algorithm in Comprehensive Planning of Distribution Network

To solve the comprehensive planning of distribution networks, a fuzzy differential evolution algorithm is proposed.A fuzzy logic controller is designed to adaptively adjust the mutation zoom factors and crossover probability constant in evolution parameters to avoid the locally optimal solution during the iteration and quicken the evolution speed.To prevent blind search, an initial chromosome coding method that can keep the radiation constraint of distribution network is designed, and the cost to repair the non-feasible individuals into feasible ones as possible is estimated.During the power flow verification, the bus-branch numberings are directly applied to power flow calculation and the wiring analysis is confined to the feeders whose structure is varied and the electrical sub-island related to above-mentioned feeders.The results of simulation for IEEE 50-bus distribution network show that the proposed algorithm can effectively improve convergence speed and convergence performance of comprehensive planning of distribution networks.