System integration is a key enabler to maximize information value in an engineering context. The valuable information is normally represented by information models which play a decisive role in the implementation of system integration. The information models are designed to efficiently and effectively capture, process and communicate information among different functional and business units. However, use of the information models in implementing system integration is challenged by insufficient support from current settings of modeling architectures. This situation calls for new strategies to ease the use of information models.

To address this challenge, this study presents a new twofold solution: Model driven system integration. It includes 1) a modeling architecture to guide the development of information models and 2) an integrated implementation process to guide the use of information models. Thus, this work improves practical applicability of an information model in its entire modeling lifecycle.

The results contribute not only to the performance of modeling practices but also to improved understanding of information modeling in system integration. Implementation contexts and implementation models are introduced to develop an implementation-oriented modeling architecture. Further, the potential of information models as a knowledge base to sup-port implementation practices is identified.

To concretely discuss behaviors and structures of information models, this study adopts ISO 10303 and the related standards as major references of existing information models.

Case studies on model driven system integration validate this research in scenarios concerning kinematic modeling, kinematic error modeling, cutting tools classification and product catalogue modeling. Model driven system integration exhibits high efficiency in implementation, enhanced interoperability and increased value of information models.