Insight最新文献

In most cases, over 80% of a system's cost is already committed during the preliminary design review (PDR) of the life cycle (Claxton 1993 and NASA 2017). Overruns often occur because requirements of certain stakeholders, such as those related to maintainability, are not involved in the system design process early enough. Consequently, a collaborative effort during the design stage should involve the requirements of key stakeholders and decision-makers. In this perspective, domain ontologies and the sharing of needs should facilitate a common agreement. Systems engineers are key stakeholders to guarantee the viability of the solution system. The integration of domain ontologies within a model-based systems engineering framework will enable the interconnection of the heterogeneous data needed to integrate maintenance issues into collaborative system design. Adding maintenance requirements earlier in the design process will help limit additional costs and delays.

Large amounts of heterogeneous data are generated, used and managed all along engineering projects. The concept of digital thread (DTH) is therefore evoked to assume digital continuity, consistency, and traceability of these data. To this purpose, this paper proposes an approach to address data, tool, and processes interoperability problems in the DTH.

The optimization of the system development process has long been a significant challenge. In the context of aircraft business class seats, the products are mechatronics systems as they incorporate diverse discipline-specific components (mechanical, electrical, and electronic). Therefore, their design, manufacturing, and V&V (verification and validation) are complex. The V&V phase typically occurs later in the development process, adhering to the V-model and ISO 15288 technical processes. However, ensuring system behavior testing early in the development process is essential to prevent extensive rework and physical prototyping iterations. This paper presents an innovative approach integrating virtual testing in the development process of mechatronics systems. This methodology is built upon the traditional V-model, enhanced by a parallel testing process using the digital twin concept. The digital twin architecture is designed following the model-based systems engineering (MBSE) approach.

Performing verification and validation (V&V) as early as possible has become a critical challenge for companies. Requirements and their associated models, elaborated during the engineering phase, often present quality issues that must be systematically verified. This paper introduces a conceptual model of requirements, together with writing patterns specifically designed for system requirements. These contributions provide a structured approach to writing system requirements and will form a foundation for future work on requirements modeling and verification.

Conversational generative artificial intelligence (CGenAI) offers new opportunities to support systems engineering practices but raises numerous challenges regarding human-AI teaming design. This ongoing project is exploring the design of human-CGenAI interaction (H-CGenAI.I) within socio-technical systems and proposes a modelling approach grounded in model-based systems engineering (MBSE). A scenario-driven methodology and SysML profile will be introduced to capture and formalise interactive patterns. The study also intends to develop an evaluation framework based on heuristics and practitioner feedback. This research aims to foster a human-centred integration of H-CGenAI.I into systems engineering analysis, enhancing traceability, collaboration, and reliability in the engineering of socio-technical systems.

The digital thread concerns the integration of engineering data across a system's lifecycle and has emerged as a cornerstone of modern digital engineering. One promising application is in engineering education, where it can expose students to authentic, connected workflows across multiple domains. The digital engineering factory (DEF), developed at the University of Arizona, is a web-based platform designed to support systems and software engineering students by providing integrated access to tools for project management, requirements, modeling, analysis, verification, and test planning. These tools are linked through Violet, a central hub that aggregates data into a comprehensive, semantically structured database. Using the ontological modeling language (OML) and the University of Arizona ontology stack (UAOS), the DEF supports reasoning, validation, and querying. These capabilities enable dynamic dashboards to guide students and assist instructors with assessment. In a classroom deployment, students worked in role-defined teams using the DEF to manage their project data and participate in model-based design and review. The DEF provided students with a clear view of the full engineering lifecycle and enabled automated grading based on traceable, semantically validated data. Ongoing development is focused on improving automated workflows and enhancing the student and instructor experience.

Digital engineering (DE) promises faster, more reliable product development—but only when its benefits are tangible to practitioners. This concise 5-page paper distills a years-long applied research effort at Dassault Systèmes into a portable showcase: a five-axis Arduino-based robotic arm modelled, simulated, manufactured, and verified through a single digital thread. Using SysML-based model-based systems engineering (MBSE) (MagicGrid method), Modelica multiphysics, robotic simulation, FMI co-simulation, and MQTT-enabled hardware-in-the-loop, we demonstrate how requirements trace directly to architecture, mechanics, electronics and code. The result is a replicable template for universities and industry teams seeking to adopt digital engineering (DE) with minimal cost and maximum pedagogical impact.