配电网络重构算法研究和展望.pdf

配电网络重构算法研究与展望 邱荣福，吕学勤，陈树果， (上海电力学院电力系统及其自动化学院，上海 200090) 摘 要：配电网络重构是通过改变分段开关的不同位置来实现的，目的有三：减少网损、平衡负荷和改善电压质量。 配电网 络重构是复杂的组合优化问题，早期应用在配网重构的经典算法有模糊控制、神经网络和遗传等算法，但是，模糊控制算法 依赖其它技术，具有一定的局限性；而神经网络算法精确度依赖于样本的训练程度，计算复杂；遗传算法计算速度慢，全局 寻优能力差。本文介绍了粒子群寻优禁忌搜索算法（PSO-TS ），它能有效的解决配网重构算法简单、计算高效、执行性强和 全局优化能力强的现实要求。最后，提出配电网重构不仅要继续研究搜索方法，还应重视负荷预测、潮流计算等相关问题的 研究，充分利用人工智能等理论加快配电网重构的理论研究和实际应用。 关键词：配电网络重构；人工智能算法；组合优化；PSO-TS ；潮流计算 A Survey and Prospect on Artificial Intelligence Algorithm for Distribution Network Reconfiguration QIU Rong-fu, LU Xue-qin , CHEN Shu-guo, （Electric power and automation engineering, Shanghai University of Electric Power ShangHai 200090,China ） Abstract: Distribution network reconfiguration (DNRC) in distribution systems has to solve the complex combinatorial optimization problem with the purpose of reducing loss, balancing load and improving the quality of voltage in the system. DNRC is realized by changing the status of sectionalizing switches. The achievements and prospect of artificial intelligence algorithm for DNRC are introduced. The early classic methods for DNRC are fuzzy control algorithm, artificial neural network algorithm and genetic algorithm,. But these methods have shortcomings in DNRC to some extent. Fuzzy control algorithm will act upon combination with other technologies; The accuracy of neural network algorithm depends on the level of training samples and its computation is complex ,resulting in poor control of real-time; The genetic algorithm costs time and its global optimization ability is poor. In order to simplify DNRC algorithms, improved particle swarm optimization is well presented in this paper. It is indicated that the research should be br