From active learning to deep reinforcement learning: Intelligent active flow control in suppressing vortex-induced vibration

2021 ◽  
Vol 33 (6) ◽  
pp. 063607
Author(s):  
Changdong Zheng ◽  
Tingwei Ji ◽  
Fangfang Xie ◽  
Xinshuai Zhang ◽  
Hongyu Zheng ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Active Learning ◽  
Flow Control ◽  
Active Flow Control ◽  
Vortex Induced Vibration ◽  
Active Flow

scholarly journals Artificial neural networks trained through deep reinforcement learning discover control strategies for active flow control

2019 ◽  
Vol 865 ◽  
pp. 281-302 ◽  
Author(s):  
Jean Rabault ◽  
Miroslav Kuchta ◽  
Atle Jensen ◽  
Ulysse Réglade ◽  
Nicolas Cerardi
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Reinforcement Learning ◽  
Flow Control ◽  
Mass Flow ◽  
Active Flow Control ◽  
Flow Rates ◽  
Mass Flow Rates ◽  
Artificial Neural ◽  
Active Flow

We present the first application of an artificial neural network trained through a deep reinforcement learning agent to perform active flow control. It is shown that, in a two-dimensional simulation of the Kármán vortex street at moderate Reynolds number ($Re=100$), our artificial neural network is able to learn an active control strategy from experimenting with the mass flow rates of two jets on the sides of a cylinder. By interacting with the unsteady wake, the artificial neural network successfully stabilizes the vortex alley and reduces drag by approximately 8 %. This is performed while using small mass flow rates for the actuation, of the order of 0.5 % of the mass flow rate intersecting the cylinder cross-section once a new pseudo-periodic shedding regime is found. This opens the way to a new class of methods for performing active flow control.

scholarly journals Reinforcement learning for bluff body active flow control in experiments and simulations

2020 ◽  
Vol 117 (42) ◽  
pp. 26091-26098
Author(s):  
Dixia Fan ◽  
Liu Yang ◽  
Zhicheng Wang ◽  
Michael S. Triantafyllou ◽  
George Em Karniadakis
Keyword(s):  
Reinforcement Learning ◽  
Fluid Mechanics ◽  
Drag Reduction ◽  
Flow Control ◽  
Bluff Body ◽  
Control Strategies ◽  
Active Flow Control ◽  
Reduction Process ◽  
Automatic Sequence ◽  
Active Flow

We have demonstrated the effectiveness of reinforcement learning (RL) in bluff body flow control problems both in experiments and simulations by automatically discovering active control strategies for drag reduction in turbulent flow. Specifically, we aimed to maximize the power gain efficiency by properly selecting the rotational speed of two small cylinders, located parallel to and downstream of the main cylinder. By properly defining rewards and designing noise reduction techniques, and after an automatic sequence of tens of towing experiments, the RL agent was shown to discover a control strategy that is comparable to the optimal strategy found through lengthy systematically planned control experiments. Subsequently, these results were verified by simulations that enabled us to gain insight into the physical mechanisms of the drag reduction process. While RL has been used effectively previously in idealized computer flow simulation studies, this study demonstrates its effectiveness in experimental fluid mechanics and verifies it by simulations, potentially paving the way for efficient exploration of additional active flow control strategies in other complex fluid mechanics applications.

scholarly journals Deep Reinforcement Learning for Active Flow Control around a Circular Cylinder Using Unsteady-mode Plasma Actuators

2020 ◽  
Author(s):  
Mohamed Elhawary
Keyword(s):  
Fluid Dynamics ◽  
Reinforcement Learning ◽  
Drag Reduction ◽  
Flow Control ◽  
Control Strategies ◽  
Active Flow Control ◽  
Plasma Actuators ◽  
Burst Frequency ◽  
Computational Fluid Dynamics Cfd ◽  
Active Flow

Deep reinforcement learning (DRL) algorithms are rapidly making inroads into fluid mechanics, following the remarkable achievements of these techniques in a wide range of science and engineering applications. In this paper, a deep reinforcement learning (DRL) agent has been employed to train an artificial neural network (ANN) using computational fluid dynamics (CFD) data to perform active flow control (AFC) around a 2-D circular cylinder. Flow control strategies are investigated at a diameter-based Reynolds number Re_D = 100 using advantage actor-critic (A2C) algorithm by means of two symmetric plasma actuators located on the surface of the cylinder near the separation point. The DRL agent interacts with the computational fluid dynamics (CFD) environment through manipulating the non-dimensional burst frequency (f+) of the two plasma actuators, and the time-averaged surface pressure is used as a feedback observation to the deep neural networks (DNNs). The results show that a regular actuation using a constant non-dimensional burst frequency gives a maximum drag reduction of 21.8 %, while the DRL agent is able to learn a control strategy that achieves a drag reduction of 22.6%. By analyzing the flow-field, it is shown that the drag reduction is accompanied with a strong flow reattachment and a significant reduction in the mean velocity magnitude and velocity fluctuations at the wake region. These outcomes prove the great capabilities of the deep reinforcement learning (DRL) paradigm in performing active flow control (AFC), and pave the way toward developing robust flow control strategies for real-life applications.

Sign in / Sign up

Export Citation Format

Share Document