用于混合式断路器的神经网络故障电流检测方法  

Fault Current Detection Based on Neural Network in Hybrid Circuit Breaker

作　　者：赵阳[1] 魏新劳[1] 冯越[1] 

机构地区：[1]哈尔滨理工大学自动化学院,黑龙江哈尔滨150080

出　　处：《哈尔滨理工大学学报》2011年第1期53-56,63,共5页Journal of Harbin University of Science and Technology

摘　　要：为了有效的实现用于混合式电力电子断路器的故障电流检测,设计了一种基于神经网络理论的短路电流检测方法,其主旨是将动态神经网络应用于故障电流的检测,利用反馈神经网络的历史记忆效应,对信号进行预测比较,可实现一种有效的短路电流故障检测。使用Matlab神经网络工具箱进行仿真,通过仿真产生模拟训练样本,以单相工频基波叠加多次谐波分量,简化时可用类正弦函数代替,仿真结果表明了该方法的有效性和快速性。In order to effectively achieve fault current detection for a hybrid circuit breaker,we design a short-circuit current detection methods based on dynamic neural network,appling the dynamic neural network to the fault current detection.Uses historical memory effect of feedback neural networks to predict and compare the signals can realize an effective short-circuit current fault detection.Using the Matlab neural network toolbos for simulation,simulation training samples generated by simulation,we superpose multiphase harmonic components by single-phase fundamental frequency and similar sine function can be used instead when predigestion.The simulation results show that the method is effective and fast.

关 键 词：混合式断路器 神经网络 电流检测 

分 类 号：TM743[电气工程—电力系统及自动化]