Reinforcement-learning-based control of turbulent channel flows at high Reynolds numbers

TL;DR

This paper demonstrates the use of deep reinforcement learning to develop effective drag reduction strategies in turbulent channel flows at high Reynolds numbers, outperforming traditional methods and providing insights into turbulence control mechanisms.

Contribution

The study introduces a DRL-based control approach for turbulent flows at high Reynolds numbers, revealing its effectiveness and underlying nonlinear control mechanisms.

Findings

DRL achieves up to 35.6% drag reduction at Re_{τ}=180

Expanded wall actions improve drag reduction but decrease at higher Re

DRL reduces skin friction by inhibiting turbulent kinetic energy redistribution

Abstract

Deep reinforcement learning (DRL) is employed to develop control strategies for drag reduction in direct numerical simulations (DNS) of turbulent channel flows at high Reynolds numbers. The DRL agent uses near-wall streamwise velocity fluctuations as input to modulate wall blowing and suction velocities. These DRL-based strategies achieve significant drag reduction, with maximum rates of 35.6% at Re_{\tau}=180, 30.4% at Re_{\tau}=550, and 27.7% at Re_{\tau}=1000, outperforming traditional opposition control methods. Expanded range of wall actions further enhances drag reduction, although effectiveness decreases at higher Reynolds numbers. The DRL models elevate the virtual wall through blowing and suction, aiding in drag reduction. However, at higher Reynolds numbers, the amplitude modulation of large-scale structures significantly increases the residual Reynolds stress on the virtual wall, diminishing the drag reduction. Analysis of budget equations provides a systematic understanding of the drag reduction dynamics behind. DRL models reduce skin friction by inhibiting the redistribution of wall-normal turbulent kinetic energy. This further suppresses the wall-normal velocity fluctuations, reducing the production of Reynolds stress, thereby decreasing skin friction. This study showcases the successful application of DRL in turbulence control at high Reynolds numbers and elucidates the nonlinear control mechanisms underlying the observed drag reduction.