基于DenseNet算法的一维桁架数字化装配控制系统仿真及优化  

Simulation and Optimization of One Dimensional Truss Digital Assembly Control System Based on DenseNet Algorithm

作　　者：白岩 ■ 王月[1] BAI Yan;YOU Jeongbong;WANG Yue(College of Mechanical Engineering,Beihua University,Jilin Jilin 132021,China;Department of Electrical,Electronic and Control Engineering,Kongju National University,Cheonan 03187,Korea)

机构地区：[1]北华大学机械工程学院,吉林吉林132021 [2]公州大学电气电子控制工程学部,韩国天安03187

出　　处：《机床与液压》2022年第8期133-137,143,共6页Machine Tool & Hydraulics

基　　金：吉林省科技发展计划项目(20200301044RQ);吉林省高等教育改革研究重点课题(JLB5530520190719153732);吉林省2019年教育科学“十三五”规划重点课题(zd19011)。

摘　　要：为解决一维桁架在轨建造过程中存在的精确定位难及桁架工位拓展转换控制效率低的技术难题,提出运用数字孪生和人工智能相结合的方式,通过计算机建模技术,构建并完成集自动装配、搬运、检测、信息追溯为一体的数字化装配系统,并与传统工艺对比。结果表明:与传统工艺相比,基于DenseNet算法的数字化装配系统在提高桁架素材故障检测率、空间利用率,减少配套失误,实现全生命周期实时追溯,加快动作节拍等方面效果显著。研究结果为我国宇航制造企业以及装备制造业开展面向智能制造的产品数字化生产线建设提供参考。In order to solve the technical problems of accurate positioning and low efficiency of truss station expansion and conversion control in the process of on orbit construction of one-dimensional truss, by using the combination of digital twinning and artificial intelligence and the computer modeling technology, a digital assembly system integrating automatic assembly, handling, detection and information tracing was constructed and completed.The results show that compared with the traditional process, the digital assembly system based on DenseNet algorithm has remarkable effects in improving the fault detection rate of truss materials and space utilization, reducing matching errors, realizing real-time tracing in the whole life cycle, speeding up the action beat and so on.The research results provide reference for China’s aerospace manufacturing enterprises and equipment manufacturing industry to carry out the construction of product digital production line for intelligent manufacturing.

关 键 词：桁架 数字化装配控制系统 DenseNet算法 机器学习 

分 类 号：TH165[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]