摘要
针对卫星等复杂产品装配工时定额主要依靠人工经验确定导致的准确性低、制定速度慢、管理不规范等问题,提出一种基于文本挖掘与神经网络模型的复杂产品装配工时估算方法。以卫星为例分析了装配工艺数据的特点,总结了装配工时的影响因素,并根据工艺特点划分了工艺类别。借助文本挖掘技术对工艺文本特征进行了提取并分类;在此基础上构建了工时预测神经网络模型,实现了面向复杂产品装配的定额工时准确估算。设计开发了复杂产品装配工时定额与管理系统,并在航天某所上线试运行,应用效果良好,验证了所提方法的可行性和实用性。
Aiming at the problems such as low accuracy,slow formulation speed and nonstandard management caused by artificial experience,an assembly man hour estimation method for complex products based on text mining and neural network model is proposed.Taking satellite as an example,the characteristics of assembly process data are analyzed,and the influencing factors of assembly man hour are summarized,and the process categories are classified according to the process characteristics.Text mining technology is used to extract and classify the process text features;on this basis,the neural network model of man hour prediction is constructed to realize the accurate estimation of quota man hour for complex product assembly.Finally,an assembly man hour quota and management system for complex products is designed and developed,and the system is put into trial operation in an Aerospace Institute.The application effect is good,and the feasibility and practicability of the proposed method are verified.
作者
刘子文
刘检华
程益
庄存波
LIUZiwen;LIU Jianhua;CHENG Yi;ZHUANG Cunbo(School of Mechanical Engineering,Beijing Institute of Technology,Beijing 100081;Aerospace System Engineering Shanghai,Shanghai 201109)
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2021年第15期199-210,共12页
Journal of Mechanical Engineering
基金
国家自然科学基金(52005042&51935003)
国防基础科研(JCKY2016204A502)
北京理工大学青年教师学术启动计划资助项目。
关键词
复杂产品
装配
工时定额估算
文本挖掘
神经网络
工时管理
complex product
assembly
man-hour quota estimation
text mining
neural network
man-hour management
