A Multi-source Data Fusion Method to Establish Product Family Architecture

Because of the distributed and heterogeneous characteristics of customized product data, establishing product family architecture based on existing customized product data has became a huge challenge for manufacturer. In order to tackle this issue, a multi-source data fusion method was proposed in this paper. Customized product data and its characteristics were analysized. And then the process of establishing product family architecture based on customized product data was presented which mainly consists of data preprocess, product structure eliciting, and data fusion modules. Based on existing research results, depth of bill of material (BOM) tree node and strength of edge in BOM tree graph were defined. And similar BOM trees were unified into a general product structure based on the conception of depth and strength and minimum weighted symmetric difference between BOM trees. Using the same way, these general product structures were unified into a GBOM. Based on it, component in each node of GBOM was classified into required component and choice component according to decision condition based on the usage status of component in customized product. Product family architecture was established in the end. Finally, a case study was given out to verify it.