Assessment of Sliding-Mesh Method for Predicting Multi-Rotor UAM Aircraft Performance on Graphics Processing Units

This paper presents assessments of rotorcraft simulations using a sliding-mesh method on high-performance computing graphics processing unit (GPU) architectures. The assessment simulations are conducted for NASA’s quadrotor air taxi concept vehicle for urban air mobility. A sliding-mesh method has been implemented in FUN3D, an unstructured-grid computational fluid dynamics (CFD) flow solver developed and supported at NASA’s Langley Research Center. Hover-flight simulations are conducted on unstructured composite grid systems for the quadrotor with four rotating rotors, fuselage, wings, and rotor hubs. The superior computational efficiency of GPU-based solutions enables simulations corresponding to 120 rotor revolutions with a high-resolution hybrid SA-neg-R/DDES (Delayed Detached Eddy Simulation) turbulence model. The SA-neg-R/DDES solution is compared with an unsteady Reynolds-averaged Navier-Stokes solution using the SA-neg-R turbulence model.