The Research Of Product Family Design Technology For Reconfigurable Manufacturing Systems

Reconfigurable Manufacturing System (RMS), taking advantage of highly customized product with high quantity, will be one of the most popular paradigms. Design of modularized product family is a key support technology for an effective working of RMSs. The method of classification of product families is based on experience of engineers and designers and simple statistical analysis of parts used, which is a lack of quantitative analysis and scientific basis and inadequate intuition. Therefore, based on the design of modularity, a methodology of formation of the set of product families was presented; a proper set of families is selected, in RMS, products are grouped into families, each of which required a system configuration and became the problem to be solved.Product family plays an important role in the RMS, current situation and development trend of RMS and product family model was depicted, the problem of current design of product family was analyzed; the theory, definition and character were depicted, and the definition and attributes of product family was analyzed, the feasibility of the methodology based on these attributes of product family was developed, the fundamental theory and method of modularity design was depicted; a study of method used in CMSs, such as Descriptive procedures, Mathematical programming, Group analysis and Artificial intelligence approaches, was carried out. In the context of cell and product families'formation, Average Linkage Clustering Algorithm (ALC) was considered as the most appropriate to apply.A methodology was presented in detail, which took into account the parameters and index of product families and calculated, for each parameters and index, the matrixes that summarized the similarity between pairs of products. Then through the use of the Analytic Hierarchy Process (AHP), a unique matrix that comprised the similarity values between products was obtained. The Average Linkage Clustering algorithm was applied to this matrix in order to obtain a dendogram that showed the diverse sets of product families that might be formed and was based on the similarity between pairs of products. At last, two cases about motors and keyboard in some company was introduced to illustrate the science, the feasibility and the effectiveness of this method.