基于混合算法的单端行波故障测距方法研究

A New Single Terminal Method of Traveling Wave Fault Location Based on Hybrid Intelligent Algorithm

针对单端行波故障测距中存在故障点反射波和对端母线反射波波头性质辨识的问题,提出一种基于小波变换、神经网络、遗传算法相结合的优化故障测距方法。利用小波变换采集前3个故障行波波头信息,将后2个波头与首波头的时间差以及对应波头极性作为神经网络模型的样本属性,利用遗传算法优化神经网络结构。当实际线路发生故障时,提取故障行波信息进行归一化处理后代入故障测距网络模型得到初步测距结果。根据初测结果和小波模极大值极性判别法的关系,对波头性质进行正确辨识,得到精确的测距结果。当极性判别法失效时,将初测结果作为最终测距结果。经仿真验证,该方法有较高的可靠性和精确性。

Aiming at the problem of the correct recognition of the reflected waves coming from fault points and buses at the opposite terminal of the faulty line in the single terminal traveling wave fault location, a fault location method based on wavelet transform, neural network and genetic algorithm is put forward. Using the wavelet transform method to collect the information of the first three fault traveling wave heads, the polarity and time-lag of the wave head information are selected as the characteristic of ANN （artificial neutral network） to establish the fault diagnosis model, then the genetic algorithm is used to optimize the structure of the neural network. When the actual transmission line failure occurs, the information of the fault traveling wave is extracted, and normalized then substituted into the fault location network model to get the initial location result. According to the relationship between the initial result and wavelet modulus maximum method, the properties of traveling wave heads are correctly identified, and then the accurate location result comes out. When the polarity criterion based on the wavelet modulus maximum method fails, the initial result is set as the final location result. Simulation reveals that this method has relatively high reliability and accuracy.