基于遗传算法的接地网腐蚀故障诊断研究

Research of Fault Diagnosis of Grounding Grid Corrosion Based on Genetic Algorithm

为了对接地网的腐蚀故障进行准确定位和判断,将遗传算法应用于接地网腐蚀故障诊断中。实际的接地网腐蚀故障诊断有电压法和电阻法两种思路,笔者分别根据这两种思路建立了适应度评价函数,并编制了相应的MATLAB程序。采用多层减维技术对程序进行优化,以解决接地网腐蚀故障诊断中支路变量维数多、算法后期收敛速度慢的问题。最后以一个60条支路、36个节点的接地网为例检验和对比了电压法和电阻法的诊断效果。仿真实验结果表明,在经过5层遗传算法迭代后,两种方法腐蚀诊断的误诊率和漏诊率均为0。电阻法的诊断效果略优于电压法,但其运算时间要高于电压法。

In order to position and judge the corrosion condition of grounding gird accurately,genetic algorithm was applied to fault diagnosis of grounding gird corrosion.There are two kinds of train of thought in real fault diagnosis of grounding gird corrosion,voltage method and resistance method.This paper established two fitness functions according to these two kinds of train of thought respectively and corresponding MATLAB programs were compiled.Technology of multilayer reducing dimension was used to optimize programs in order to solve the problems that branch variable dimension number is large and convergence rate is slow in late iterative procedure.Finally we took a grounding gird with 60 branches,36 nodes as an example to test and comprise the effect on fault diagnose of voltage method and resistance method.Simulation experimental results showed that after five layer genetic algorithm iteration,both misdiagnosis rate and missed diagnosis rate of these two methods were 0.Diagnosis result of resistance method was slightly better than voltage method,but its computation time was higher than voltage method.