Creating ever-more-sophisticated simulations
By Umesh Paliath, Jeff Slotnick and Andrew Wissink|December 2024
The CFD Vision 2030 Integration Committee advocates for, inspires and enables community activities recommended by the vision study for revolutionary advances in the state-of-the-art of computational technologies needed for analysis, design and certification of future aerospace systems.
In August, the fifth High Lift Prediction Workshop was attended by a large gathering of computational fluid dynamics practitioners from government, industry, academia and commercial interests. With the continued goal of enhancing CFD prediction capability for practical high-lift aerodynamic design, attendees worked on test cases focused on solution verification, the prediction of configuration build-up and Reynolds number effects, all using the CRM-HL, or High Lift Common Research Model, reference geometry. To anchor the CFD results, researchers applied experimental data from testing of CRM-HL to ONERA’s 5.1% full-span model in the French space agency’s F1 facility. Among the outcomes: When properly verified and fully iteratively converged, Reynolds-Averaged Navier-Stokes simulations on families of best-practice meshes for the simpler CRM-HL configurations generated mesh-converged solutions that are consistent with RANS solutions obtained with mesh adaptation. Additionally, CFD efforts continued to focus on the maturation of scale-resolving simulation methods, like Wall-Modeled Large Eddy Simulation and Hybrid RANS-LES technology. The scaled resolved simulations showed more consistent improvement in the prediction of subsonic maximum lift over other methods using meshes of roughly a billion degrees of freedom. However, results indicated that these methods are not appropriately modeling leading-edge slat boundary layer transition and low angle-of-attack flap flow separation. Researchers anticipate that predictive capabilities will improve as lessons from future CRM-HL testing are incorporated to refine modeling best practices.
A key milestone of the CFD Vision 2030 study was to demonstrate efficiently scaled CFD simulations on an exascale system. As these platforms have come online at U.S. Department of Energy laboratories over the past two years, multiple CFD applications have leveraged these resources to showcase CFD predictive capabilities. This year, two breakthroughs occurred on the Frontier exascale machine at Oak Ridge National Laboratory in Tennessee. First, the FUN3D development team from NASA’s Langley Research Center in Virginia worked with NASA’s Ames Research Center in California, Analytical Mechanics Associates of Virginia and Old Dominion University in Virginia to perform full-system simulations. Applications included real-gas simulations of retro propulsion for a human-scale Mars lander concept and simulations of the CRM-HL. In March, GE Aerospace simulated a full-scale Open Fan Blade at real-world flight conditions as part of CFM International’s RISE, or Revolutionary Innovation for Sustainable Engines, program. This simulation gave engineers an enhanced view into the complex turbulent flow at a microscopic level to guide aerodynamic and aeroacoustic design. These high-fidelity simulations yielded information that would otherwise take scientists years to gather. The problem size of these simulations is of the order of hundreds of billion grid points and have been demonstrated to run efficiently on greater than 90% of the Frontier machine.
Another achievement involved the development of high-fidelity co-simulation technology for free-flight full-aircraft maneuvers, using closed-loop tightly coupled CFD and computational structural dynamics, CSD, with full-production aircraft flight control systems. Sikorsky created a virtual prototype of its S-97 Raider X rotorcraft by integrating its proprietary flight controls software with the Department of Defense’s Helios high-fidelity rotary-wing simulation code. Helios was developed under the High Performance Computing Modernization Program’s Computational Research and Engineering Acquisition Tools and Environments program. With its virtual S-97 prototype, Sikorsky validated the model against flight test data from a real-world vehicle, demonstrating that the high-fidelity CFD model was able to achieve comparable accuracy to flight. In May, the Vertical Flight Society presented the team with the 2024 Howard Hughes award for this work.
Contributors: Reynaldo J. Gomez III and Eric Nielsen
