对已有产品的模块化改造是实现复杂产品系统由集成化设计向模块化设计转变的前提。但目前的模块化方法主要是面向新产品和结构相似的变型产品的。对于结构差异较大的异质类复杂产品系统的模块化，还缺少有效的理论和方法。针对这一现状，本项目拟开展异质复杂产品系统模块化方法研究，包括数据建模、功能-结构同质化处理、模块聚类划分以及模块参数规划等方面。并以典型产品为例，开发原型系统，进行实证研究。本项目创新点在于:(1)提出了基于矩阵的异质多产品功能-结构同质化方法，减少了功能冗余，提高了零部件共享率；(2)采用定量化方法描述零部件之间的耦合关系强度，应用改进的GN社区发现算法进行模块划分，提高了模块聚类效率，使聚类方法适用于复杂产品系统零部件数量规模巨大和耦合关系复杂的情况；(3)提出了基于顾客参数需求分布的复杂产品系统模块参数优化取值模型，采用启发式算法获取模块参数谱系，为后续模块化配置设计奠定基础。

The modularization of existing complex products and systems (CoPS) is the precondition, which can facilitate the transformation of product design from integrated design to modular design. However, the available modular design approaches are mainly focus on new products and variant products with similar physical structures. Research on the effective theory and approach for the modularization of heterogeneous CoPS with various physical structures is insufficiency. Motivated by this issue, this project intends to study the modularization of heterogeneous CoPS, including data modeling, homogenization of function and physical structure, module clustering and module parameter planning, then with typical products, develop a prototype system to accomplish the empirical research. The main innovation points of this project are: (1) developing the homogenization approach of product functions and physical structures based on matrices theory to reduce the function redundancy and increase the components share ratio; (2) representing the coupling relation intensity with quantitative method, establishing module partition with modified GN community detection algorithm, and the proposed approach could cope with the huge components amount and complex coupling relations of CoPS to improve the module clustering efficiency; (3) proposing the module parameter evaluation model for CoPS based on the customer requirements distribution and obtaining the module parameter portfolio with heuristic algorithm, which could lay the foundation for subsequent modular configuration.