神经网络技术在智能BIT故障诊断系统中的应用  

Application of Intelligent BIT Fault Diagnosis System of Neural Network Research

作　　者：祁涛[1] 张彦斌[1] 姚人前 

机构地区：[1]军械工程学院 [2]解放军66476部队

出　　处：《火力与指挥控制》2016年第6期125-128,共4页Fire Control & Command Control

基　　金：军队重点课题基金资助项目(装通XXXX号)

摘　　要：论述了智能BIT的智能设计、智能检测、智能诊断和智能决策,构建了基于神经网络的某高炮装备随动系统的智能BIT故障诊断系统。用Multisim进行电路仿真,提取输出信号的均值、峭度、偏斜度构成三维向量,以它作为特征向量利用神经网络进行模拟电路的故障诊断。通过比较BP神经网络、SOM神经网络和小波神经网络的诊断结果,得知利用均值、峭度和偏斜度作为特征,BP神经网络和SOM神经网络能够有效识别故障状态模式。This paper discusses the intelligent BIT of intelligent design,intelligent detection, intelligent diagnosis and intelligent decision,and builds the certain anti-aircraft gun equipment servo system which based on the network of intelligent BIT fault diagnosis system. The circuit is simulated by Multisim,and then it is used to extract the output signal of the mean,kurtosis,and skewness,so to constitute a three-dimensional vector to analog circuit fault diagnosis. Comparing the BP neural network,SOM neural network and wavelet neural network diagnostic results,that using the mean, kurtosis and skewness as characteristics,can easily realize fault diagnosis by BP and SOM.

关 键 词：智能BIT 软故障 神经网络 MULTISIM仿真 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]