考虑零部件间间接连接关系的模块划分方法

Module Partition Method Considering Influence of Indirect Connection between Components

模块化的产品架构可以缩短研发周期、提高研发效率和产品可靠性,有效解决大规模生产和多样化客户需求之间的矛盾。针对现有模块划分方法常忽略零部件间间接连接关系对模块划分的影响,造成重要设计依赖信息遗失的问题,建立了基于电容类比法的节点相似度模型,量化了零部件间直接和间接连接关系共同作用的效果,实现了模块的精确划分。首先,采用区间直觉模糊集分析了零部件之间的综合关联关系,将综合关联自相关矩阵映射为复杂网络;然后,建立了基于电容类比法的节点相似度模型,并以该相似度模型为中心度测度,提出了一种网络层次聚类(Network hierarchical clustering,NHC)算法,实现了复杂农机装备的模块精确划分;最后,对联合收获机脱粒装置进行模块划分,验证了NHC算法的有效性。采用NHC算法获得的划分方案的平均模块度比改进的GN算法提高了24.6%,获得的产品结构树比改进的GN算法更合理。

Agricultural machinery equipment research and development usually has problems such as long research and development cycle,low efficiency and poor product reliability.The modular product architecture can shorten the research and development cycle,improve research and development efficiency and product reliability,and effectively solve the contradiction between large-scale production and diverse customer needs.To solve the problem that the existing module division methods ignored the influence of indirect connection between components on the module division,which resulted in the loss of important design dependency information,a node similarity model was established based on capacitance analogy method,the joint effect of direct and indirect connection between components was quantified,and the accurate module division was realized.Firstly,the interval valued intuitionistic fuzzy set was used to analyze the comprehensive correlation between components,and the comprehensive correlation autocorrelation matrix was mapped to a complex network.Then,a node similarity model based on the capacitance analogy method was established,and the similarity model was used as a central measure to propose a network hierarchical clustering(NHC)algorithm,which realized the discovery of complex network communities.Finally,the effectiveness of the NHC algorithm was verified by dividing the threshing device module of the combine harvester.The average modularity of the division scheme obtained by the NHC algorithm was 24.6%higher than that of the improved GN algorithm,and it was more reasonable to obtain the product structure tree than the improved GN algorithm.