摘要
为了解决现代化武器装备因集成化和大型化引起的故障诊断难题,简要介绍了几种典型的深度学习模型,对其在旋转机械、齿轮箱、滚动轴承和电子电路系统中故障的特征提取与诊断方法进行了分析,总结了当前深度学习在故障诊断领域的现状及其存在的问题,提出了在模型、信息与故障特征提取、故障预测等方面的发展方向与改进建议,以更好地适应装备发展要求,促进装备保障能力的有效提升。
In order to solve the problem of fauh diagnosis caused by modernization of weaponry due to integration and large-scale, and further improve the support capabilities of equipment, several typical deep learning models were briefly introduced, and feature extraction and diagnostic methods for equipment fail- ures in different systems (i. e. rotating machinery, gearbox, rolling bearing and electronic circuit system) were performed. Current status of deep learning in fault diagnosis and its existing problems are summarized, and development directions and suggestions for improvement are proposed to better meet the requirements of equipment development which includes model, information and fault feature extration, and fault prediction, and promote the effective improvement of equipment support capabilities.
作者
王应晨
段修生
Wang Yingchen;Duan Xiusheng(Department of Electronic and Optical Engineering,Army Engineering University of PLA,Shijiazhuang 050003,China)
出处
《战术导弹技术》
北大核心
2018年第5期25-30,共6页
Tactical Missile Technology
基金
装备状态监控与维修总装重点实验室基金项目(ECMM2016007)
陆军装备军内科研项目基金(装综(2018)11号)
关键词
现代化装备
深度学习
故障诊断
modern equipment
deep learning
fault diagnosis