摘要
为了提升不确定环境下的再制造装配质量,在分析再制造装配与原始制造装配差异的基础上,提出一种"因件制宜"再制造装配过程优化控制方法;以再制造零部件的不确定性特征为输入,再制造装配工位操作补偿值为输出,构建了基于反向传播神经网络的再制造装配过程耦合模型,实现对再制造装配工位进行"因件制宜"的操作指导,保障再制造产品质量。将本方法集成在再制造发动机装配车间制造执行系统中,验证了所提方法的有效性和可行性,为再制造装配过程控制的精准化和实时化提供了理论和方法支持,也为再制造产品可靠性和服役安全性提供了技术支持。
To improve the quality of remanufactured products under uncertainty,based on analyzing the difference between reassembly(remanufacturing assembly)and original assembly,a control method of suiting operation to different conditions of remanufactured parts for reassembly process was proposed.The reassembly process coupling model based on Back Propagation neural network was constructed by taking the feature of remanufacturing parts as input,and the operation compensation value of reassembly station as output,which achieved on the guidelines of suiting operation to different conditions of remanufactured parts in the reassembly station to improve the quality of remanufactured products.The proposed method was incorporated into Manufacturing Execution System(MES)of remanufacturing car engine assembly shop,and the validity and feasibility were verified.The research provided the theory and method support for real-time and precise control of reassembly process,and also provided technical support for reliability and service safety of remanufactured products.
作者
刘从虎
李文艺
蔡维
韩君
何康
温海骏
LIUConghu;LI Wenyi;CAI Wei;HAN Jun;HE Kang;WEN Haijun(School of mechanical and electronic engineering, Suzhou University, Suzhou234000;State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing 400030;Antai College of Economics and Management, Shanghai Jiao Tong University, Shanghai 200030;College of Mechanical Engineering ~ Automatization, North University of China, Taiyuan 030051)
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2018年第6期1357-1366,共10页
Computer Integrated Manufacturing Systems
基金
中国博士后科学基金面上资助项目(2017M611574)
教育部人文社会科学研究资助项目(17YJC630082)
安徽省高校自然科学研究重点资助项目(KJ2017A438,KJ2017A439)
安徽省高校优秀人才支持计划资助项目(gxyqZD2018082).
关键词
再制造
装配
优化控制
反向传播神经网络
remanufacturing
assembly
optimization control
back propagation neural network