Supporting semantic PLM by using a lightweight engineering metadata mapping engine

Abstract In order to handle a high variety of interdisciplinary processes and complex smart products, integration platforms are a useful approach to view and access data all along the product lifecycle, which is stored in different data management solutions like Product Lifecycle Management (PLM), Enterprise Resource Planning (ERP) and authoring systems. Here, the Metadata Repository for Semantic Product Lifecycle Management (SP²IDER) could serve as a supporting integration platform. An additional information layer on top of data source systems like PLM and ERP provides additional information, links data objects from different source systems, and provides access to these data. The SP²IDER platform consists of three basic parts: Connector units to fetch data from the source systems, a core unit with a Service Directory and a Mapping Engine, and a Metadata Store, where information about data objects is stored. This paper focuses on the view inside the Mapping Engine and the Metadata Store. The mapping engine has a three-way approach for the mapping of data objects and data types: The initial manual mapping at the data type level, second a rule-based, and third a machine-learning mapping at the data object level. This paper describes the manual mapping process, how the mapped data objects are stored inside the Metadata Store, and how this leads to a newly formed lightweight data model, that does not need heavyweight ontologies.