基于CNN的某型火箭炮电气系统故障诊断研究  

The Fault Diagnosis Research on a Certain Type of Rocket Launcher Electrical System Based on CNN

作　　者：李永保 张建新 张震 王小召 李锋 谢立中 LI Yongbao;ZHANG Jianxin;ZHANG Zhen;WANG Xiaozhao;LI Feng;XIE Lizhong(Hubei Jiangshan Heavy Industry Co.,Ltd.,Hubei Xiangyang 441005,China)

机构地区：[1]湖北江山重工有限责任公司,湖北襄阳441005

出　　处：《弹箭与制导学报》2021年第2期82-86,共5页Journal of Projectiles,Rockets,Missiles and Guidance

基　　金：装备预研兵器工业联合基金(6141B012102)资助。

摘　　要：火箭炮向着自动化、信息化的方向发展,传统故障诊断方法的局限性日益突出,现代火箭炮需要快速准确地进行故障诊断,以适应现代战争的要求。针对火箭炮电气系统故障的特点,提出一种基于一维卷积神经网络的火箭炮电气系统故障诊断方法。基于某型火箭炮电气系统的故障数据,建立了故障诊断模型,经验证该故障诊断方法的准确率可达98.61%,证实了卷积神经网络应用于火箭炮电气系统故障诊断的可行性。Rocket launcher is developing in the direction of automation and information.The limitations of traditional fault diagnosis methods are becoming increasingly prominent.Modern rocket launcher requires fast and accurate diagnosis method to meet the requirements of modern warfare.Aiming at the characteristics of the faults in the electrical system of rocket launchers,this paper proposes a fault diagnosis method based on one-dimensional convolutional neural network.Based on the fault data of the electrical system of a certain rocket launcher,a fault diagnosis model was established.It has been verified that the accuracy of the fault diagnosis method can reach 98.61%,which confirms the feasibility of the convolutional neural network for the fault diagnosis of the rocket launcher electrical system.

关 键 词：火箭炮 电气系统 故障诊断 卷积神经网络 神经网络 

分 类 号：TJ713[兵器科学与技术—武器系统与运用工程] E924.93[军事—军事装备学]