基于遗传算法的配电网故障定位系统

【作者】 陈碧云；

【机构】 广西大学电气学院；

【摘要】 结合北海市城网自动化设计方案,提出了一种可行的基于开关电流越限信号的配电网故障定位遗传算法。该算法采用树型数据结构自动识别网络拓扑、单一故障完全初始化、最佳保留机制、两断点交叉、可变的交叉及变异概率。在PⅡ—200的微机上的仿真测试,表明本算法能进行全局寻优求解,能自动识别网络拓扑的变化,具有很好的容错性能,远优于传统的故障电流判别法。此外,对于每回配线包含30个以下元件的一般配电系统,计算时间不超过15s,具有在线应用的潜力。更多还原

【Abstract】 This paper presents a feasible method of fault section diagnosis in distribution networks based on Genetic Algorithm(GA),using overflow signals of breakers,combining the project of Beihai City-Distribution-Network’s Automatization.Automatic recognition of Topo structure, Tree-type data structure,completely initialization of single fault,best preserved mechanism, two-point crossover,alterable probabilities of crossover and mutation have been introduced in this paper.The procedure is worked out accordingly,and the results of simulation show that this algorithm can fred global optimum solution and recognize the change of Topo structure of the networks automatically.The method has high fault-tolerance performance,which is superior to traditional fault current discriminance method.Furthermore,the computing time for common system, with no more than 30 elements each distributing line,doesn’t overrun 15s,so this method can be implemented on-line.更多还原