Workshop to share knowledge about system engineering experiments and how to deal with their validity.
Monday 09 October 2023, 9:30 - 17:00.
University of Antwerp, Campus Middelheim, Building G, CDF Room (M.G.028).
Abstract: A Validity Frame captures the set of contexts in which a model (and its analysis, often by means of simulation) of a system is able to replace that system with respect to questions about a set of salient properties of interest. Even though the utility of validity frames has been reported in current literature, there does not exist any precise and general definition of the concept. In this presentation, I will describe our ongoing development of a framework for designing and using validity frames. This framework both uses and supports model management. We have developed an ontology in order to precisely define the concepts of the model validity domain. The framework currently consists of ontological definitions integrated into a workflow model that describes a general experiment, validation experiments, and the construction of validity frames. A simple resistor model validation case study is used as a running example to describe the concepts. The validity frames of different resistor models are computed. How to use the framework in different scenarios is sketched.
Abstract: In the realm of model-driven engineering, the effective management and representation of complex workflows and their associated artefacts stand as pivotal challenges. This work aims to address them by introducing a comprehensive framework designed to model workflows, including their activities and artefacts. At the core of this framework, we have ontological constructs, organizing information and forging meaningful connections between conceptual entities. In this presentation, we illustrate the application of this framework in the context of maintaining a digital model or twin of the design of a drivetrain monitoring system. By encapsulating the components, processes, and dependencies, the framework offers a structured approach to minimise the inherent complexity of such systems. Leveraging the ontologies crafted for this purpose, the framework concretises the semantic relationships between diverse elements, enabling a clearer understanding of the drivetrain system's behaviour, evolution, and performance.
Abstract: A simulation can be a complex architecture of simulation models, simulation tools, and computing hardware assembled to meet particular needs. Thus, a simulation can be a complex product in its own right with its own development cycle, to be distinguished from the development cycle of the system it represents. However, simulation development is sometimes subject to informal procedures and can begin without a clear, complete, and formal definition of the simulation needs. Simulation traceability is then compromised, which prevents from easily validating whether a simulation meets the needs, or understanding the purpose of a simulation model that can be reused. This presentation highlights how systems engineering can help formalize simulation, with a particular focus on the definition and validation of simulation needs.
| Maintained by Hans Vangheluwe. | Last Modified: 2023/10/03 17:38:35. |