Steady-blowing jet control of a finite-length square cylinder at high reynolds number: Turbulence characteristics and DRL-based adaptive drag reduction

This study systematically investigates the turbulence characteristics and drag-reduction mechanisms of steady wall-normal blowing control on a three-dimensional square cylinder at Re = 22000 using a large-eddy simulation (LES) with the Smagorinsky subgrid-scale model. Four jet positions at the cylinder corners—the windward face (Jet 1), the two lateral faces (Jets 2 and 3), and the leeward face (Jet 4)—are first compared under a fixed mass-flow-rate ratio of Q i = 5%. The results reveal that the leeward jet (Jet 4) yields the best performance: the time-averaged recirculation bubble contracts markedly, the leeward-face pressure recovers, and the coherence of the Kármán vortex street is disrupted, achieving a 25.4% drag reduction. In contrast, windward and lateral jets produce limited benefits or even increase drag. Further analysis using Q-criterion vortex identification, wake topology and oscillation patterns, and spectral proper orthogonal decomposition (SPOD) clarifies the underlying mechanism: Jet 4 injects momentum directly into the vortex-formation region, thereby shrinking the reversed-flow zone, suppressing free-end downwash, and converting large-scale coherent vortex shedding into broadband small-scale fluctuations. Based on these findings, a TD3-based deep-reinforcement-learning framework is developed at the Jet 4 location. The adaptive policy is trained on a reduced LES mesh and then frozen and tested on the high-resolution mesh, where it dynamically adjusts the jet velocity under the constraint Q i ≤5%. After 2000 training episodes, the learned policy achieves a 29.5% drag reduction in deterministic high-resolution testing. These results demonstrate that leeward “direct-action” steady-blowing jets combined with DRL-based adaptive control constitute an effective strategy for drag reduction of bluff bodies at high Reynolds numbers, offering a practical reference for flow control of analogous engineering structures.

Authors

Institutions

Publication Details

Journal
Applied Ocean Research
Published
2026-09-18
DOI
https://doi.org/10.1016/j.apor.2026.105272
Primary Topic
Fluid Dynamics and Vibration Analysis
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Steady-blowing jet control of a finite-length square cylinder at high reynolds number: Turbulence characteristics and DRL-based adaptive drag reduction

Xin Guan, Yuheng Wu, Liang Zhong, Jinyang Liu
Applied Ocean Research
Fluid Dynamics and Vibration Analysis
article

Steady-blowing jet control of a finite-length square cylinder at high reynolds number: Turbulence characteristics and DRL-based adaptive drag reduction

Xin Guan, Yuheng Wu, Liang Zhong, Jinyang Liu
article en

Abstract

This study systematically investigates the turbulence characteristics and drag-reduction mechanisms of steady wall-normal blowing control on a three-dimensional square cylinder at Re = 22000 using a large-eddy simulation (LES) with the Smagorinsky subgrid-scale model. Four jet positions at the cylinder corners—the windward face (Jet 1), the two lateral faces (Jets 2 and 3), and the leeward face (Jet 4)—are first compared under a fixed mass-flow-rate ratio of Q i = 5%. The results reveal that the leeward jet (Jet 4) yields the best performance: the time-averaged recirculation bubble contracts markedly, the leeward-face pressure recovers, and the coherence of the Kármán vortex street is disrupted, achieving a 25.4% drag reduction. In contrast, windward and lateral jets produce limited benefits or even increase drag. Further analysis using Q-criterion vortex identification, wake topology and oscillation patterns, and spectral proper orthogonal decomposition (SPOD) clarifies the underlying mechanism: Jet 4 injects momentum directly into the vortex-formation region, thereby shrinking the reversed-flow zone, suppressing free-end downwash, and converting large-scale coherent vortex shedding into broadband small-scale fluctuations. Based on these findings, a TD3-based deep-reinforcement-learning framework is developed at the Jet 4 location. The adaptive policy is trained on a reduced LES mesh and then frozen and tested on the high-resolution mesh, where it dynamically adjusts the jet velocity under the constraint Q i ≤5%. After 2000 training episodes, the learned policy achieves a 29.5% drag reduction in deterministic high-resolution testing. These results demonstrate that leeward “direct-action” steady-blowing jets combined with DRL-based adaptive control constitute an effective strategy for drag reduction of bluff bodies at high Reynolds numbers, offering a practical reference for flow control of analogous engineering structures.

Applied Ocean ResearchVol. 176
Chongqing Jiaotong University (CN)
Chongqing Municipal Education Commission, Scientific Research and Technology Development Program of Guangxi, Chongqing Jiaotong University, Chongqing Municipal Human Resources and Social Security Bureau
Sustainable cities and communities
Openalex Percentile: Top 14%
Fluid Dynamics and Vibration Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.