A Framework for Cyber-physical System Tool-chain Development : A Service-oriented and Model-based Systems Engineering Approach

Sammanfattning: The development of complex Cyber-physical Systems (CPS) requires tight interactions of projects, system components, stakeholders, data and models. These models and data support component design, which are implemented by various engineering tools used by stakeholders. Effective tool integration thus relies on development of tool-chains. Five common challenges have been identified as part of this work concerning the development of tool-chains to support Model-based System Engineering (MBSE) for complex CPS:Models from different domains are represented by different syntaxes and seemingly similar syntax may have completely distinct semantics. Moreover, such models and their views have completely different purposes. A unified formalism, therefore, does not exist to accommodate all system artifacts from all constituent models.Modeling tools are developed by different tool suppliers and hence the interoperability of these tools may be limited, because the interfaces may not be fully open.The interoperability limitation is especially prohibitive for co-simulations across multiple simulation tools.Tool-chains must support design automation of product development workflows as adopted by different business units.Users and developers of the MBSE tool-chain must have evaluation criteria to judge the effectiveness of different tool-chains and workflows.To alleviate these challenges, this thesis proposes a novel Model-based System Engineering (MBSE) framework, called SPIRIT, to support model and data integration and tool-chain development. The framework contributes to four main phases for CPS tool-chain development, namely concept, design, early evaluation and operation phases. For the concept phase, the framework exploits the systems thinking approach to develop novel concepts of MBSE tool-chains. For the design phase, the framework adopts a service-oriented approach to construct tool-chains from the perspectives of social networks, process, information-service-infrastructure, and technology. For the early evaluation phase, quantitative metrics are defined to measure, (i) the MBSE capabilities of tools within a tool-chain, and (ii) the interoperability of the tool-chain. For the operation phase, several MBSE tool-chain prototypes are developed to support product development. An advantage of the new framework is to support tool-chain development using systems thinking and considering integrations of several open standards, including: 1) a domain-specific modeling (DSM) approach based on the Graph-Object-Property-Point-Role-Relationship meta-meta model; 2) ontology design based on the Web Ontology Language; 3) co-simulation using the High-Level Architecture and the Functional Mock-up Interface; 4) model-driven process management using BPMN; 5) tool-integration based on Open Services for Lifecycle Collaboration; and 6) value selections of design parameters based on an automated decision-making algorithm.The effectiveness of the novel MBSE framework is investigated and verified by three case studies. Three main contributions are concluded from this thesis:Benefits and challenges of MBSE tool-chains in industry are identified through a questionnaire survey and literature review. The results and the use of a systems thinking approach led to the development of a conceptual architectural model aiming to support MBSE tool-chain formalisms.The SPIRIT framework is defined to provide support for MBSE tool-chain development. The framework has the following properties: architecting by DSM, integrating tools and models via open standards, enabling automatic co-simulations, and supporting design automation.A DSM approach supporting visualization and Bayesian network analysis is presented to support MBSE tool-chain assessment. Quantitative metrics are defined to evaluate the effectiveness of MBSE tool-chains.