摘要
针对电子信息系统结构复杂、难于建立精确数学模型等特点,以某型电子信息系统为例,采用故障字典和BP神经网络相结合的方法,利用Multisim软件进行电路仿真。再由实测样本数据对BP神经网络进行训练,完成对网络各参数的设置。仿真结果表明,该方法可以较好地将故障定位到元器件,同时可为其他电子信息系统模块级故障诊断技术研究提供借鉴。
Since it is difficult to establish precise mathematical model for electronic information system due to its complex structure,taking a certain type of electronic information system as an example,a method to combine a fault dictionary with BP neural network is adopted. The circuitry emulation was performed by means of Multisim software. The BP neural network was trained with the measured data. The setup of network parameters was accomplished. The emulation result indicates that this meth-od can preferably localize a fault to a certain component,and can be used as a reference for research on block-level fault diagno-sis technology for other electronic information systems.
出处
《现代电子技术》
2013年第20期166-170,共5页
Modern Electronics Technique
关键词
电子信息系统
故障诊断
神经网络
故障字典
electronic information system
fault diagnosis
neural network
fault dictionary
