Large Eddy Simulations for Next-Generation Digital Twin Engine Analysis
Large eddy simulations (LES) have emerged as indispensable tools in augmenting the capabilities
of next-generation digital twin frameworks for integrated engine analysis. The importance of LES
lies in its ability to provide comprehensive insights into the unsteady flow physics in multicomponent engine systems. By resolving large-scale turbulent structures and capturing turbulent
interactions between components such as compressors, combustors, and turbines, LES-driven
digital twins offer deeper understandings of system behavior. This abstract highlights the critical
role of LES and GPUs in enabling multi-component engine analysis, offering insights into complex unsteady flow phenomena. Results from our HPCMP Frontier project will be presented in terms of performance, turbulence statistics, flow dynamics, and heat transfer characteristics within a realistic engine configuration. Additionally, the scalability and efficiency of LES on GPU-accelerated platforms will be discussed, demonstrating the potential for widespread adoption in the aerospace industry. Furthermore, the integration of LES with advanced machine learning techniques will be explored, showcasing the synergy between high-fidelity simulations and data-driven modeling for enhanced predictive capabilities. Overall, this abstract underscores the transformative impact of LES and GPU technology on the development of next-generation digital twin frameworks, paving the way for more efficient and reliable engine designs. This work is conducted in support of the DoD High Performance Computing Modernization Program
(HPCMP) Frontier program at ARL and NAVAIR in collaboration with the University of Cincinnati,
Cadence Design Systems, and Pratt & Whitney.
PRESENTER
Bravo, Luis
luis.g.bravorobles.civ@army.mil
914-843-4278
DEVCOM Army Research Laboratory
CO-AUTHOR
Russell Powers
russell.w.powers2.civ@us.navy.mil
CATEGORY
Computational Fluid Dynamics
SYSTEMS USED
Nautilus, Narwhal, Warhawk
SECRET
No