神经网络用于同杆并架双回线故障诊断的方法

A FAULT DETECTIVE METHOD OF DOUBLE-CIRCUIT LINES WITH NEURAL NETWORK

同杆并架双回线路发生故障时，由于影响故障的因素复杂，且故障种类繁多，使得故障识别比较困难。目前在单回线上广泛采用的各种类型保护及重合闸装置都难以满足同杆并架双回线的特殊要求。本文介绍了一种应用神经网络进行同杆并架双回线故障诊断的方法。根据从黄浦到苏村之间２２０ ｋｖ 等级同杆并架双回线的实际参数，对各种故障进行了大量的仿真，并采用两个串联的神经网络来判别故障情况及选相，经检验，该方法能够正确地诊断各种故障。

With the existance of complicated fault conditions and much more fault varieties,many protective devices and reclosing devices widely applied to single circuit lines can't fully meet the special need of Double Circuit lines.This paper presents a fault detective method of Double Circuit lines using neural network.According to the actual data of 220kv Double Circuit lines form HuangPu to SuCun,two serial neural networks are used to detect and classify different faults through a lot of simulations for to all kinds of fault.As a result,this method has been proved to be a correct and effective way to detect faults.