• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2021, Vol. 57 ›› Issue (9): 191-202.doi: 10.3901/JME.2021.09.191

• 数字化设计与制造 • 上一篇    下一篇

扫码分享

考虑变更风险的复杂产品模块划分方法

李玉鹏, 谢卫星, 王晓琳, 任海兵   

  1. 中国矿业大学工业工程系 徐州 221116
  • 收稿日期:2020-06-05 修回日期:2021-02-03 出版日期:2021-05-05 发布日期:2021-06-15
  • 通讯作者: 王晓琳(通信作者),女,1977年出生,博士,副教授,硕士研究生导师,主要研究方向为复杂系统设计、企业管理等。E-mail:graciouswxl@126.com
  • 作者简介:李玉鹏,男,1985年出生,博士,副教授,硕士研究生导师。主要研究方向为复杂产品系统设计理论及群决策。E-mail:ypeng_li@163.com;谢卫星,男,1994年出生,硕士研究生。主要研究方向为复杂产品模块化及设计变更影响分析。E-mail:1105413061@qq.com;任海兵,男,1978年出生,博士,副教授,硕士研究生导师。主要研究方向为复杂系统设计、生产计划与控制等。E-mail:cumt_rhb@163.com
  • 基金资助:
    国家自然科学基金(51505480,51475290)和江苏省自然科学基金(BK20150197)资助项目。

Module Partition Method for Complex Products Considering the Change Risk

LI Yupeng, XIE Weixing, WANG Xiaolin, REN Haibing   

  1. Department of Industrial Engineering, China University of Mining and Technology, Xuzhou 221116
  • Received:2020-06-05 Revised:2021-02-03 Online:2021-05-05 Published:2021-06-15

摘要: 模块化方法可有效控制复杂产品的研制周期与成本,模块划分是实现模块化的基础。受顾客需求、关键技术及政策、市场等内外部动态因素的影响,持续演化是复杂产品的必然趋势。设计变更是企业实现产品演化的基本手段,为了降低演化的影响,在对复杂产品进行模块划分时,同时考虑功能、结构、变更风险三个因素。首先将复杂网络理论应用于复杂产品结构建模,以零部件为节点,零部件的关联关系为边,建立复杂产品零部件网络模型。然后基于LinkRank的社区发现算法求解出产品模块初始划分方案。其次再利用基于网络相似度优化的社区发现算法得到最终划分方案。最后以某型号摩托车发动机为例验证了所提方法的可行性和有效性。

关键词: 复杂产品, 产品演化, 模块划分, 变更风险, 网络相似度

Abstract: Modularization method can effectively control the cost of research and development for complex products, and module partition is the foundation of modularization. Impacted by the dynamic internal and external factors including customer requirements, key technologies, policy and market, continuous evolution is unavoidable for complex products. Design change is the basic means for an enterprise to establish product evolution. In order to reduce the impact of evolution, three factors of function, structure and change risk are considered in the module partition of complex products. Firstly, the complex network theory is applied to the structure modelling of complex products, with components as nodes and component relationships as edges, the network model of the components for complex products is established. Then the LinkRank community discovery algorithm is used to obtain the initial module partition scheme of the product. Afterwards, the network similarity-based community discovery algorithm is used to obtain the final partitioning scheme. Finally, a case study of a motorcycle engine is presented to verify the feasibility and validity of the proposed approach.

Key words: complex products, product evolution, module partition, change risk, network similarity

中图分类号: