一种基于改进遗传算法的变电站故障识别方法  

A Fault Recognition Method for Substation Based on Improved Genetic Algorithm

作　　者：许伟东 陈义森 Xu Weidong;Chen Yisen(School of Electric Power Engineering,South China University of Technology,Guangzhou Guangdong 510640,China)

机构地区：[1]华南理工大学电力学院,广东广州510640

出　　处：《电气自动化》2022年第4期94-96,共3页Electrical Automation

摘　　要：针对传统优化算法求解变电站故障诊断解析模型时存在易于早熟、求解准确度不高以及收敛速度慢等问题,设计了一种基于改进的遗传算法的变电站故障识别新方法。首先,针对传统遗传算法由于采用固定的交叉与变异概率导致收敛慢、全局搜索能力不足的缺陷,设计了与进化代数相关的交叉概率、与个体适应值相关的变异概率,并对现有的自适应遗传算法进行改进,提高算法的全局搜索能力与收敛速度;其次,依据电网拓扑及其保护配置情况构建电网故障识别解析模型;最后,应用改进的遗传算法对上述故障诊断模型进行求解。通过包含断路器正常动作、据动和误动场景的故障算例分析,发现所提方法较常规优化算法在多次重复性试验中仍能保持接近1.0的故障识别准确率,且收敛代数保持在10左右,能够实现故障的快速准确定位。Aiming at the problems of easy maturity,low solution accuracy and slow convergence speed when solving the analytical model of substation fault diagnosis with traditional optimization algorithm,a new method of substation fault identification based on improved genetic algorithm was designed.The whole idea is as follows:firstly,in view of the slow convergence and insufficient global search ability of traditional genetic algorithms due to the fixed crossover and mutation probability,the crossover probability related to evolutionary algebra and the mutation probability related to individual fitness were designed to improve the existing adaptive genetic algorithm,so as to raise the global search ability and convergence speed of the algorithm;secondly,an analytical model for power grid fault identification was constructed based on the power grid topology and its protection configuration.Finally,an improved genetic algorithm was used to solve the above-mentioned fault diagnosis model.Through the analysis of fault examples including the normal operation,data movement,and maloperation scenarios of the circuit breaker,it is found that the proposed method can still maintain a fault recognition accuracy close to 1.0 in multiple repetitive tests compared with the conventional optimization algorithm,the convergence algebra remains at about the 10th generation,and the fault can be located quickly and accurately.

关 键 词：电网故障识别 数学解析模型 改进遗传算法 收敛性 全局搜索能力 

分 类 号：TM64[电气工程—电力系统及自动化]