Data Management in a Collaborative Design Architecture, Part C: Simulation Topology

Abstract

In this three-part paper series, a central data management system (DMS) in a collaborative design architecture for aero engines is described. This paper aims at closing the gap between pre-design and high-fidelity simulation by extending an existing data model of the engine for including complex simulation configurations. The model links computational domains with the geometric features found in pre-design and allows to integrate high-fidelity simulations into the overall design considerations. To integrate high-fidelity simulations, i.e. CFD or CSM, into the holistic engine perspective, the parametric description of the engine components must be translated into computational domains. Based on an ontology, a data model is introduced which describes those domains and its meta information. This data model is called domain-based simulation topology and provides the structure to build a knowledge graph. It links the multiple disciplines required to investigate complex physical phenomena. Based on the data model, the necessary information and data to run computations on a HPC cluster can be deployed to these infrastructures. Data transfer is only required once and the individual simulation goals, i.e. boundary conditions or solver settings, become part of the knowledge graph. The results of the simulations are written back to the data management system, from which the collaborative design platform can extract individual pieces of information.