DARWIN
"Digitaler Thread"-based turbomachine design with embedded artificial intelligence using high-fidelity simulation methods
The LuFo project DARWIN aims at the development, application and validation of innovative multi-disciplinary simulation and design methods as well as methods of artificial intelligence (AI) for an improved understanding of airplane turbomachines.
Four work packages are focused on the following topics:
- Improvement of the simulation framework for high-fidelity simulation of turbomachines on high performance clusters
- Robust and automated generation of digital twins of engine components and integration of data from the development, validation and operational phases of the product life cycle into the concept "digital thread"
- Application and Improvement of AI and machine learning for accelerating design processes as well as improvinmg quality and accuracy of simulation results
- Highly accurate simulation of multi-disciplinary behaviour of turbomachines with hybrid and cross-scale methods involving real geometry.
ZIH is involved in work package 1 in the improvement of the load balancing of coupled models on contemporary and novel computer architectures of the CFD code Hydra and in work package 3 in AI- and ML-based knowledge extraction from simulation and measurement data and the associated data and metadata management.
In work package 1, ZIH is significantly involved in improving the load balancing of coupled models on standard and novel computer architectures of the CFD code Hydra. In work package 3, questions of AI- and ML-based knowledge extraction from simulation and measurement data and the associated data and metadata management are researched at ZIH.
Partners
- Rolls-Royce Deutschland Ltd & Co KG (Koordinator)
- Institut für Strömungsmechanik der TUD
- DLR Braunschweig, Köln, Dresden
- Universität Darmstadt
- University of Surrey (UK)
- CFD Consultants
- Intelligent Light
Duration
01.09.2020 - 31.05.2024
Contact at ZIH
Funding
BMWi (FKZ 20D1911C)