arXiv subjects

R. Vinuesa

Publications and source records attributed to R. Vinuesa.

At least 19 recordsLinked to original sources

arXiv2025

Discovering Flow Separation Control Strategies in 3D Wings via Deep Reinforcement Learning

In this work, deep reinforcement learning (DRL) is applied to active flow control (AFC) over a threedimensional SD7003 wing at a Reynolds number of Re = 60,000 and angle of attack of AoA = 14 degrees. In the uncontrolled baseline case, the flow exhibits massive separation and a fully turbulent wake. Using a GPU-accelerated CFD solver and multi-agent training, DRL discovers control strategies that enhance lift (79%), reduce drag (65%), and improve aerodynamic efficiency (408%). Flow visualizations confirm reattachment of the separated shear layer, demonstrating the potential of DRL for complex and turbulent flows.

cs.CE↗

arXiv2025

Deep Reinforcement Learning for Active Flow Control around a Three-Dimensional Flow-Separated Wing at Re = 1,000

This study explores the use of deep reinforcement learning (DRL) for active flow control (AFC) to reduce flow separation on wings at high angles of attack. Concretely, here the DRL agent controls the flow over the three-dimensional NACA0012 wing section at the Reynolds number Re = 1,000 and angle of attack AoA = 20 degrees, autonomously identifying optimal control actions through real-time flow data and a reward function focused on improving aerodynamic performance. The framework integrates the GPU-accelerated computational fluid dynamics (CFD) solver SOD2D with the TF-Agents DRL library via a Redis in-memory database, enabling rapid training. This work builds on previous DRL flow-control studies, demonstrating DRL potential to address complex aerodynamic challenges and push the boundaries of traditional AFC methods.

cs.CE↗

arXiv2025

Assessment of non-intrusive sensing in wall-bounded turbulence through explainable deep learning

In this work we present a framework to explain the prediction of the velocity fluctuation at a certain wall-normal distance from wall measurements with a deep-learning model. For this purpose, we apply the deep-SHAP method to explain the velocity fluctuation prediction in wall-parallel planes in a turbulent open channel at a friction Reynolds number ${\rm{Re}}_\tau=180$. The explainable-deep-learning methodology comprises two stages. The first stage consists of training the estimator. In this case, the velocity fluctuation at a wall-normal distance of 15 wall units is predicted from the wall-shear stress and wall-pressure. In the second stage, the deep-SHAP algorithm is applied to estimate the impact each single grid point has on the output. This analysis calculates an importance field, and then, correlates the high-importance regions calculated through the deep-SHAP algorithm with the wall-pressure and wall-shear stress distributions. The grid points are then clustered to define structures according to their importance. We find that the high-importance clusters exhibit large pressure and shear-stress fluctuations, although generally not corresponding to the highest intensities in the input datasets. Their typical values averaged among these clusters are equal to one to two times their standard deviation and are associated with streak-like regions. These high-importance clusters present a size between 20 and 120 wall units, corresponding to approximately 100 and 600${\rm\mu m}$ for the case of a commercial aircraft.

physics.flu-dyn↗

arXiv2024

Bursting bubble in an elasto-viscoplastic medium

A gas bubble sitting at a liquid-gas interface can burst following the rupture of the thin liquid film separating it from the ambient, owing to the large surface energy of the resultant cavity. This bursting bubble forms capillary waves, a Worthington jet, and subsequent droplets for a Newtonian liquid medium. However, rheological properties of the liquid medium like elasto-viscoplasticity can greatly affect these dynamics. Using direct numerical simulations, this study exemplifies how the complex interplay between elasticity (in terms of elastic stress relaxation) and yield stress influences the transient interfacial phenomena of bursting bubbles. We investigate how bursting dynamics depends on capillary, elastic, and yield stresses by exploring the parameter space of the Deborah number $De$ (dimensionless relaxation time of elastic stresses) and the plastocapillary number $\mathcal{J}$ (dimensionless yield-stress of the medium), delineating four distinct characteristic behaviours. Overall, we observe a non-monotonic effect of elastic stress relaxation on the jet development while plasticity of the elasto-viscoplastic medium is shown to affect primarily the jet evolution only at faster relaxation times (low $De$). The role of elastic stresses on jet development is elucidated with the support of energy budgets identifying different modes of energy transfer within the elasto-viscoplastic medium. The effects of elasticity on the initial progression of capillary waves and droplet formation are also studied. In passing, we study the effects of solvent-polymer-viscosity ratio on bursting dynamics and show that polymer viscosity can increase the jet thickness apart from reducing the maximum height of the jet.

physics.flu-dyn↗

arXiv2024

Flow control of three-dimensional cylinders transitioning to turbulence via multi-agent reinforcement learning

Designing active-flow-control (AFC) strategies for three-dimensional (3D) bluff bodies is a challenging task with critical industrial implications. In this study we explore the potential of discovering novel control strategies for drag reduction using deep reinforcement learning. We introduce a high-dimensional AFC setup on a 3D cylinder, considering Reynolds numbers ($Re_D$) from $100$ to $400$, which is a range including the transition to 3D wake instabilities. The setup involves multiple zero-net-mass-flux jets positioned on the top and bottom surfaces, aligned into two slots. The method relies on coupling the computational-fluid-dynamics solver with a multi-agent reinforcement-learning (MARL) framework based on the proximal-policy-optimization algorithm. MARL offers several advantages: it exploits local invariance, adaptable control across geometries, facilitates transfer learning and cross-application of agents, and results in a significant training speedup. \rev{For instance, our results demonstrate $16\%$ drag reduction for $Re_D=400$, outperforming classical periodic control, which yields up to $6\%$ reduction.} A proper-orthogonal-decomposition (POD) analysis at $Re_D=400$ reveals that the DRL control results in a stable wake structure with longer recirculation bubble. To the authors' knowledge, the present MARL-based framework represents the first time where training is conducted in 3D cylinders. This breakthrough paves the way for conducting AFC on progressively more complex turbulent-flow configurations.

physics.flu-dyn↗

arXiv2024

Active flow control for drag reduction through multi-agent reinforcement learning on a turbulent cylinder at $Re_D=3900$

This study presents novel drag reduction active-flow-control (AFC) strategies} for a three-dimensional cylinder immersed in a flow at a Reynolds number based on freestream velocity and cylinder diameter of $Re_D=3900$. The cylinder in this subcritical flow regime has been extensively studied in the literature and is considered a classic case of turbulent flow arising from a bluff body. The strategies presented are explored through the use of deep reinforcement learning. The cylinder is equipped with 10 independent zero-net-mass-flux jet pairs, distributed on the top and bottom surfaces, which define the AFC setup. The method is based on the coupling between a computational-fluid-dynamics solver and a multi-agent reinforcement-learning (MARL) framework using the proximal-policy-optimization algorithm. This work introduces a multi-stage training approach to expand the exploration space and enhance drag reduction stabilization. By accelerating training through the exploitation of local invariants with MARL, a drag reduction of approximately 9% is achieved. The cooperative closed-loop strategy developed by the agents is sophisticated, as it utilizes a wide bandwidth of mass-flow-rate frequencies, which classical control methods are unable to match. Notably, the mass cost efficiency is demonstrated to be two orders of magnitude lower than that of classical control methods reported in the literature. These developments represent a significant advancement in active flow control in turbulent regimes, critical for industrial applications.

physics.flu-dyn↗

arXiv2023

Drag-reduction strategies in wall-bounded turbulent flows using deep reinforcement learning

In this work we compare different drag-reduction strategies that compute their actuation based on the fluctuations at a given wall-normal location in turbulent open channel flow. In order to perform this study, we implement and describe in detail the reinforcement-learning interface to a computationally-efficient, parallelized, high-fidelity solver for fluid-flow simulations. We consider opposition control (Choi, Moin, and Kim, Journal of Fluid Mechanics 262, 1994) and the policies learnt using deep reinforcement learning (DRL) based on the state of the flow at two inner-scaled locations ($y^+ = 10$ and $y^+ = 15$). By using deep deterministic policy gradient (DDPG) algorithm, we are able to discover control strategies that outperform existing control methods. This represents a first step in the exploration of the capability of DRL algorithm to discover effective drag-reduction policies using information from different locations in the flow.

physics.flu-dyn↗

arXiv2023

Active flow control for three-dimensional cylinders through deep reinforcement learning

This paper presents for the first time successful results of active flow control with multiple independently controlled zero-net-mass-flux synthetic jets. The jets are placed on a three-dimensional cylinder along its span with the aim of reducing the drag coefficient. The method is based on a deep-reinforcement-learning framework that couples a computational-fluid-dynamics solver with an agent using the proximal-policy-optimization algorithm. We implement a multi-agent reinforcement-learning framework which offers numerous advantages: it exploits local invariants, makes the control adaptable to different geometries, facilitates transfer learning and cross-application of agents and results in significant training speedup. In this contribution we report significant drag reduction after applying the DRL-based control in three different configurations of the problem.

physics.flu-dyn↗

arXiv2023

Deep reinforcement learning for turbulent drag reduction in channel flows

We introduce a reinforcement learning (RL) environment to design and benchmark control strategies aimed at reducing drag in turbulent fluid flows enclosed in a channel. The environment provides a framework for computationally-efficient, parallelized, high-fidelity fluid simulations, ready to interface with established RL agent programming interfaces. This allows for both testing existing deep reinforcement learning (DRL) algorithms against a challenging task, and advancing our knowledge of a complex, turbulent physical system that has been a major topic of research for over two centuries, and remains, even today, the subject of many unanswered questions. The control is applied in the form of blowing and suction at the wall, while the observable state is configurable, allowing to choose different variables such as velocity and pressure, in different locations of the domain. Given the complex nonlinear nature of turbulent flows, the control strategies proposed so far in the literature are physically grounded, but too simple. DRL, by contrast, enables leveraging the high-dimensional data that can be sampled from flow simulations to design advanced control strategies. In an effort to establish a benchmark for testing data-driven control strategies, we compare opposition control, a state-of-the-art turbulence-control strategy from the literature, and a commonly-used DRL algorithm, deep deterministic policy gradient. Our results show that DRL leads to 43% and 46% drag reduction in a minimal and a larger channel (at a friction Reynolds number of 180), respectively, outperforming the classical opposition control by around 20 percentage points.

physics.flu-dyn↗

arXiv2022

Fully convolutional networks for velocity-field predictions based on the wall heat flux in turbulent boundary layers

Fully-convolutional neural networks (FCN) were proven to be effective for predicting the instantaneous state of a fully-developed turbulent flow at different wall-normal locations using quantities measured at the wall. In Guastoni et al. [J. Fluid Mech. 928, A27 (2021)], we focused on wall-shear-stress distributions as input, which are difficult to measure in experiments. In order to overcome this limitation, we introduce a model that can take as input the heat-flux field at the wall from a passive scalar. Four different Prandtl numbers $Pr = \nu/\alpha = (1,2,4,6)$ are considered (where $\nu$ is the kinematic viscosity and $\alpha$ is the thermal diffusivity of the scalar quantity). A turbulent boundary layer is simulated since accurate heat-flux measurements can be performed in experimental settings: first we train the network on aptly-modified DNS data and then we fine-tune it on the experimental data. Finally, we test our network on experimental data sampled in a water tunnel. These predictions represent the first application of transfer learning on experimental data of neural networks trained on simulations. This paves the way for the implementation of a non-intrusive sensing approach for the flow in practical applications.

physics.flu-dyn↗

arXiv2022

The potential of artificial intelligence for achieving healthy and sustainable societies

In this chapter we extend earlier work (Vinuesa et al., Nature Communications 11, 2020) on the potential of artificial intelligence (AI) to achieve the 17 Sustainable Development Goals (SDGs) proposed by the United Nations (UN) for the 2030 Agenda. The present contribution focuses on three SDGs related to healthy and sustainable societies, i.e. SDG 3 (on good health), SDG 11 (on sustainable cities) and SDG 13 (on climate action). This chapter extends the previous study within those three goals, and goes beyond the 2030 targets. These SDGs are selected because they are closely related to the coronavirus disease 19 (COVID-19) pandemic, and also to crises like climate change, which constitute important challenges to our society.

cs.CY↗

arXiv2021

Applying Bayesian Optimization with Gaussian Process Regression to Computational Fluid Dynamics Problems

Bayesian optimization (BO) based on Gaussian process regression (GPR) is applied to different CFD (computational fluid dynamics) problems which can be of practical relevance. The problems are i) shape optimization in a lid-driven cavity to minimize or maximize the energy dissipation, ii) shape optimization of the wall of a channel flow in order to obtain a desired pressure-gradient distribution along the edge of the turbulent boundary layer formed on the other wall, and finally, iii) optimization of the controlling parameters of a spoiler-ice model to attain the aerodynamic characteristics of the airfoil with an actual surface ice. The diversity of the optimization problems, independence of the optimization approach from any adjoint information, the ease of employing different CFD solvers in the optimization loop, and more importantly, the relatively small number of the required flow simulations reveal the flexibility, efficiency, and versatility of the BO-GPR approach in CFD applications. It is shown that to ensure finding the global optimum of the design parameters of the size up to 8, less than 90 executions of the CFD solvers are needed. Furthermore, it is observed that the number of flow simulations does not significantly increase with the number of design parameters. The associated computational cost of these simulations can be affordable for many optimization cases with practical relevance.

physics.flu-dyn↗

arXiv2021

Effects of sweeps and ejections on amplitude modulation in a turbulent channel flow

Conditional averages are used to evaluate the effect of sweeps and ejections on amplitude modulation. This is done numerically with a direct numerical simulation (DNS) of a channel flow at friction Reynolds number $Re_τ = 1000$ in a minimal stream-wise unit (MSU). The amplitude-modulation map of such DNS is also compared to the one of a regular channel flow in a longer streamwise domain (LSD), in order to assess its validity for this study. The cheaper MSU is found to provide a good representation of the modulation phenomena in the LSD. As for conditional averages, the amplitude-modulation coefficient is conditioned on the sign of the large-scale fluctuations. Care must be exerted in defining such a coefficient, as the conditioned large-scale fluctuation has non-zero average, indeed as a consequence of conditioning. Both sweeps and ejections (positive and negative large-scale fluctuation events) are found to have a positive contribution to amplitude modulation in the buffer layer, and a negative one in the outer layer. The negative-modulation region is found to shrink in case of ejections, so that the positive-modulation region extends farther away from the wall. Two more conditional statistics are used to provide an alternative representation of amplitude modulation and insights into the characteristics of the large-scale structures.

physics.flu-dyn↗

arXiv2021

Towards adaptive simulations of turbulent wings at high Reynolds numbers

Adaptive mesh refinement (AMR) in the high-order spectral-element method code Nek5000 is demonstrated and validated with well-resolved large-eddy simulations (LES) of the flow past a wing profile. In the present work, the flow around a NACA 4412 profile at a chord-based Reynolds number $Re_c=200,000$ is studied at two different angles of attack: 5 and 11 degrees. The mesh is evolved from a very coarse initial mesh by means of volume-weighted spectral error indicators, until a sufficient level of resolution is achieved at the boundary and wake regions. The non-conformal implementation of AMR allows the use of a large domain avoiding the need of a precursor RANS simulation to obtain the boundary conditions (BCs). This eliminates the effect of the steady Dirichlet BCs on the flow, which becomes a relevant source of error at higher angles of attack (specially near the trailing edge and wake regions). Furthermore, over-refinement in the far field and the associated high-aspect ratio elements are avoided, meaning less pressure-iterations of the solver and a reduced number of elements, which leads to a considerable computational cost reduction. Mean flow statistics are validated using experimental data obtained for the same profile in the Minimum Turbulence Level (MTL) wind tunnel at KTH, as well as with a previous DNS simulation, showing excellent agreement. This work constitutes an important step in the direction of studying stronger pressure gradients and higher Reynolds complex flows with the high fidelity that high-order simulations allow to achieve. Eventually, this database can be used for the development and improvement of turbulence models, in particular wall models.

physics.flu-dyn↗

arXiv2021

Predicting the near-wall region of turbulence through convolutional neural networks

Modelling the near-wall region of wall-bounded turbulent flows is a widespread practice to reduce the computational cost of large-eddy simulations (LESs) at high Reynolds number. As a first step towards a data-driven wall-model, a neural-network-based approach to predict the near-wall behaviour in a turbulent open channel flow is investigated. The fully-convolutional network (FCN) proposed by Guastoni et al. [preprint, arXiv:2006.12483] is trained to predict the two-dimensional velocity-fluctuation fields at $y^{+}_{\rm target}$, using the sampled fluctuations in wall-parallel planes located farther from the wall, at $y^{+}_{\rm input}$. The data for training and testing is obtained from a direct numerical simulation (DNS) at friction Reynolds numbers $Re_τ = 180$ and $550$. The turbulent velocity-fluctuation fields are sampled at various wall-normal locations, i.e. $y^{+} = \{15, 30, 50, 80, 100, 120, 150\}$. At $Re_τ=550$, the FCN can take advantage of the self-similarity in the logarithmic region of the flow and predict the velocity-fluctuation fields at $y^{+} = 50$ using the velocity-fluctuation fields at $y^{+} = 100$ as input with less than 20% error in prediction of streamwise-fluctuations intensity. These results are an encouraging starting point to develop a neural-network based approach for modelling turbulence at the wall in numerical simulations.

physics.flu-dyn↗

arXiv2020

Convolutional-network models to predict wall-bounded turbulence from wall quantities

Two models based on convolutional neural networks are trained to predict the two-dimensional velocity-fluctuation fields at different wall-normal locations in a turbulent open channel flow, using the wall-shear-stress components and the wall pressure as inputs. The first model is a fully-convolutional neural network (FCN) which directly predicts the fluctuations, while the second one reconstructs the flow fields using a linear combination of orthonormal basis functions, obtained through proper orthogonal decomposition (POD), hence named FCN-POD. Both models are trained using data from two direct numerical simulations (DNS) at friction Reynolds numbers $Re_τ = 180$ and $550$. Thanks to their ability to predict the nonlinear interactions in the flow, both models show a better prediction performance than the extended proper orthogonal decomposition (EPOD), which establishes a linear relation between input and output fields. The performance of the various models is compared based on predictions of the instantaneous fluctuation fields, turbulence statistics and power-spectral densities. The FCN exhibits the best predictions closer to the wall, whereas the FCN-POD model provides better predictions at larger wall-normal distances. We also assessed the feasibility of performing transfer learning for the FCN model, using the weights from $Re_τ=180$ to initialize those of the $Re_τ=550$ case. Our results indicate that it is possible to obtain a performance similar to that of the reference model up to $y^{+}=50$, with $50\%$ and $25\%$ of the original training data. These non-intrusive sensing models will play an important role in applications related to closed-loop control of wall-bounded turbulence.

physics.flu-dyn↗

arXiv2019

Prediction of wall-bounded turbulence from wall quantities using convolutional neural networks

A fully-convolutional neural-network model is used to predict the streamwise velocity fields at several wall-normal locations by taking as input the streamwise and spanwise wall-shear-stress planes in a turbulent open channel flow. The training data are generated by performing a direct numerical simulation (DNS) at a friction Reynolds number of $Re_τ=180$. Various networks are trained for predictions at three inner-scaled locations ($y^+ = 15,~30,~50$) and for different time steps between input samples $Δt^{+}_{s}$. The inherent non-linearity of the neural-network model enables a better prediction capability than linear methods, with a lower error in both the instantaneous flow fields and turbulent statistics. Using a dataset with higher $Δt^+_{s}$ improves the generalization at all the considered wall-normal locations, as long as the network capacity is sufficient to generalize over the dataset. The use of a multiple-output network, with parallel dedicated branches for two wall-normal locations, does not provide any improvement over two separated single-output networks, other than a moderate saving in training time. Training time can be effectively reduced, by a factor of 4, via a transfer learning method that initializes the network parameters using the optimized parameters of a previously-trained network.

physics.flu-dyn↗

arXiv2019

Predictions of turbulent shear flows using deep neural networks

In the present work we assess the capabilities of neural networks to predict temporally evolving turbulent flows. In particular, we use the nine-equation shear flow model by Moehlis et al. [New J. Phys. 6, 56 (2004)] to generate training data for two types of neural networks: the multilayer perceptron (MLP) and the long short-term memory (LSTM) network. We tested a number of neural network architectures by varying the number of layers, number of units per layer, dimension of the input, weight initialization and activation functions in order to obtain the best configurations for flow prediction. Due to its ability to exploit the sequential nature of the data, the LSTM network outperformed the MLP. The LSTM led to excellent predictions of turbulence statistics (with relative errors of 0.45% and 2.49% in mean and fluctuating quantities, respectively) and of the dynamical behavior of the system (characterized by Poincaré maps and Lyapunov exponents). This is an exploratory study where we consider a low-order representation of near-wall turbulence. Based on the present results, the proposed machine-learning framework may underpin future applications aimed at developing accurate and efficient data-driven subgrid-scale models for large-eddy simulations of more complex wall-bounded turbulent flows, including channels and developing boundary layers.

physics.flu-dyn↗