基于BP神经网络的电路板故障征兆获取研究  

Study of Circuit Board Fault Symptoms Based on BP Neural Network

作　　者：谢永成[1] 刘磊峰[1] 李光升[1] 魏宁[1] 

机构地区：[1]装甲兵工程学院控制工程系,北京100072

出　　处：《四川兵工学报》2015年第10期95-98,共4页Journal of Sichuan Ordnance

摘　　要：针对电路板故障征兆获取困难的问题,提出了基于BP神经网络的故障征兆获取方法;BP神经网络获取故障征兆是指利用隶属函数将症状空间的向量模糊化,然后再通过BP神经网络映射到故障空间的一种方法;实验结果表明:BP神经网络获取电路板故障征兆不仅较为准确,而且精度较高,适合复杂电路故障的诊断。In order to overcome the difficult problem of circuit board fault diagnosis, this paper proposed a method of circuit board fault based on the BP neural network. This method used membership function to fuzz the vector of fauh space and maps the vector through BP neural network. Experimental results show that the BP neural network is accurate as well as precise and it can be used to diagnose complicated circuit board.

关 键 词：BP神经网络 隶属度函数 故障征兆 

分 类 号：TP34[自动化与计算机技术—计算机系统结构]