EXCELLERATP2 D2.15 Updated Report on the STREAmS Application Use Case

This document summarises the progress made since M12 in the STREAmS use case (UC-6) on active flow control for drag reduction (DR) of transonic airfoils. During this period, the previous FLEW solver was refactored and merged into the modular, object-oriented STREAmS-2 code leading to STREAmS-2.1 version. The porting to multi-GPU architectures has been completed to enable large-scale Direct Numerical Simulation (DNS) on heterogeneous architectures. We considered and evaluated both NVIDIA and AMD Graphics Processing Unit (GPUs) from EuroHPC clusters.

In addition, code suitable for Intel GPUs was also prepared and tested. Early runs exposed the need to improve grid generation, DNS initialisation, and far-field boundary treatments to suppress spurious acoustic reflections; these enhancements were implemented and validated. Building on Quadrio et al., we completed a comprehensive parametric study of the proposed control technique at low Reynolds number (Re = 3×10⁵) comprising 26 controlled flow cases, as reported in a recently published paper. Preliminary uncontrolled simulations have been conducted at higher Reynolds number (Re = 6×10⁵), and controlled cases are also planned.

Finally, we are preparing an ambitious DNS at Re = 1.2×10⁶, which would result - to our knowledge - in the highest Reynolds number simulated by a DNS of a controlled airfoil. We aim to target at least one uncontrolled and one controlled run, with the possibility of additional exploratory cases. Over 1.1 M node-hours have been obtained through EuroHPC and other calls, so that UC-6 is positioned to deliver its high-Reynolds objectives.