遗传算法在接地网腐蚀故障诊断中的应用研究

Research on the Application of Genetic Algorithm in Fault Diagnosis of Grounding Grid Corrosion

为了准确诊断接地网的腐蚀情况,将遗传算法(GA)应用于接地网腐蚀故障诊断中,编制了相应的MATLAB程序。利用多层减维技术解决接地网支路个数多、遗传算法收敛速度慢的问题。以重庆110 kV典发变电站接地网为例进行了仿真实验,实验结果显示多层减维技术的应用不仅加速了遗传算法的收敛过程,也优化了算法的诊断效果。遗传算法的计算结果具有一定的波动性,应在多次程序运行结果统计分析的基础上进行腐蚀评判。对典发变电站接地网的10次程序运行结果进行综合分析后,假设的7条故障支路被全部准确定位,无误诊及漏诊支路,各腐蚀支路腐蚀程度的诊断误差最大不超过10%。

In order to diagnose the corrosion condition of grounding gird accurately, genetic algorithm is applied in fault diagnosis of grounding gird corrosion and corresponding-MATLAB program was compiled. Technology of multilayer reducing dimension solves the problems that branch number of grounding gird is large and convergence rate is slow. The grounding gird of Chongqing 110 kV Dianfa substation is taken as an example to do simulation experiment. Experimental results show that the application of muhilayer reducing dimension technology not only speeds up the convergence process of genetic algorithm, but also optimizes the diagnosis effect of algorithm. The results of genetic algorithm have certain fluctuation, the corrosion evaluation should carried out on the basis of statistical analysis of many times＇ running results. After comprehensive analysis of 10 times＇ program running results to grounding grid of Dianfa substation, all the hypothesis 7 fault branches get accurate positioning, no missed diagnosis and misdiagnosis branch exists, corrosion diagnosis error of each corrosion branch is less than 10%.