基于模糊神经网络的方向元件

根据电力系统正常运行，发生各种类型正方向故障和反方向故障时，母线处获得的信号的特点，提出了用模糊神经网络来识别电力系统故障方向的模型和算法。经EMTP仿真表明，该方法能够正确地区分故障发生的方向，而且计算和响应速度快。另外，系统正在振荡时又发生故障，模糊神经网络的模型及算法也能正确区分出故障的方向。缺点是需要经过大量的训练：但是由于是离线训练，不影响此方法的实时应用。

DIRECTIONAL PART BASED ON FUZZY NEURAL NETWORK

When power system runs normally,and when positive directional fault and inverse directional fault of various fault types occur in power system,the characteristic of the signal of the bus is very different.According to this information,this paper poposed the model and algorithm of the fuzzy neural network to be used to identify the fault direction.EMTP simulation calculation indicated that this directional part could identify the fault direction,not influenced by the fault type,transition resistance and operation mode.And the speed of the directional part is high.In addition,if fault occurs when the system is in oscillation,the method proposed in this paper could identify the fault direction correctly.This method needs a large number of training,however,it doesn't influence on-line application because the training is off line.