基于神经网络算法的继电保护二次回路故障检测方法

Fault Detection Method of Secondary Circuit of Relay Protection Based on Neural Network Algorithm

针对当前对继电保护二次回路故障检测时极易受到单极性影响造成检测结果不准确的问题,引入神经网络算法,开展对其故障检测方法的设计研究。结合双端口稳压控制方法,采集故障检测数据,从海量故障检测数据中提取故障特征参数,利用神经网络算法对检测模型迭代训练,从而得到检测结果,实现对故障点的定位。通过对比实验证明,新的故障检测方法在实际应用中的抗干扰能力更强,得到的检测结果更准确,可以实现从故障发生到结束整个过程的检测。

Aiming at the problem that the fault detection of relay protection secondary circuit is easily influenced by unipolar,the neural network algorithm is introduced to carry out the design and research of its fault detection method.Combined with two-port voltage-stabilizing control method,fault detection data are collected,fault characteristic parameters are extracted from massive fault detection data,and the detection model is iteratively trained by using neural network algorithm,so as to obtain detection results and realize fault location.Through comparative experiments,it is proved that the new fault detection method has stronger anti-interference ability and more accurate detection results in practical application,and can realize the detection of the whole process from the occurrence to the end of the fault.