Identifying regions of importance in wall-bounded turbulence through explainable deep learning.

Despite its great scientific and technological importance, wall-bounded turbulence is an unresolved problem in classical physics that requires new perspectives to be tackled. One of the key strategies has been to study interactions among the energy-containing coherent structures in the flow. Such interactions are explored in this study using an explainable deep-learning method. The instantaneous velocity field obtained from a turbulent channel flow simulation is used to predict the velocity field in time through a U-net architecture. Based on the predicted flow, we assess the importance of each structure for this prediction using the game-theoretic algorithm of SHapley Additive exPlanations (SHAP). This work provides results in agreement with previous observations in the literature and extends them by revealing that the most important structures in the flow are not necessarily the ones with the highest contribution to the Reynolds shear stress. We also apply the method to an experimental database, where we can identify structures based on their importance score. This framework has the potential to shed light on numerous fundamental phenomena of wall-bounded turbulence, including novel strategies for flow control.

Visualizing Academic Contributions to Achieving the Sustainable Development Goals through AI: The Case of Universitat Politècnica de València.

How are you contributing to SDGs and measuring sustainable improvements? AI solutions can help you to quantify it. This pilot experience shows the case of the university's scientific contributions.

ß-Variational autoencoders and transformers for reduced-order modelling of fluid flows.

Variational autoencoder architectures have the potential to develop reduced-order models for chaotic fluid flows. We propose a method for learning compact and near-orthogonal reduced-order models using a combination of a ß-variational autoencoder and a transformer, tested on numerical data from a two-dimensional viscous flow in both periodic and chaotic regimes. The ß-variational autoencoder is trained to learn a compact latent representation of the flow velocity, and the transformer is trained to predict the temporal dynamics in latent-space. Using the ß-variational autoencoder to learn disentangled representations in latent-space, we obtain a more interpretable flow model with features that resemble those observed in the proper orthogonal decomposition, but with a more efficient representation. Using Poincaré maps, the results show that our method can capture the underlying dynamics of the flow outperforming other prediction models. The proposed method has potential applications in other fields such as weather forecasting, structural dynamics or biomedical engineering.

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 30% 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 and 10 percentage points, respectively.

A deep-learning approach for reconstructing 3D turbulent flows from 2D observation data.

Turbulence is a complex phenomenon that has a chaotic nature with multiple spatio-temporal scales, making predictions of turbulent flows a challenging topic. Nowadays, an abundance of high-fidelity databases can be generated by experimental measurements and numerical simulations, but obtaining such accurate data in full-scale applications is currently not possible. This motivates utilising deep learning on subsets of the available data to reduce the required cost of reconstructing the full flow in such full-scale applications. Here, we develop a generative-adversarial-network (GAN)-based model to reconstruct the three-dimensional velocity fields from flow data represented by a cross-plane of unpaired two-dimensional velocity observations. The model could successfully reconstruct the flow fields with accurate flow structures, statistics and spectra. The results indicate that our model can be successfully utilised for reconstructing three-dimensional flows from two-dimensional experimental measurements. Consequently, a remarkable reduction in the complexity of the experimental setup and the storage cost can be achieved.

Application and Advances in Radiographic and Novel Technologies Used for Non-Intrusive Object Inspection.

Increase in trading and travelling flows has resulted in the need for non-intrusive object inspection and identification methods. Traditional techniques proved to be effective for decades; however, with the latest advances in technology, the intruder can implement more sophisticated methods to bypass inspection points control techniques. The present study provides an overview of the existing and developing techniques for non-intrusive inspection control, current research trends, and future challenges in the field. Both traditional and developing methods, techniques, and technologies were analyzed with the use of traditional and novel sensor types. Finally, it was concluded that the improvement of non-intrusive inspection experience could be gained with the additional use of novel types of sensors (such as biosensors) combined with traditional techniques (X-ray inspection).

In situ visualization of large-scale turbulence simulations in Nek5000 with ParaView Catalyst.

In situ visualization on high-performance computing systems allows us to analyze simulation results that would otherwise be impossible, given the size of the simulation data sets and offline post-processing execution time. We develop an in situ adaptor for Paraview Catalyst and Nek5000, a massively parallel Fortran and C code for computational fluid dynamics. We perform a strong scalability test up to 2048 cores on KTH's Beskow Cray XC40 supercomputer and assess in situ visualization's impact on the Nek5000 performance. In our study case, a high-fidelity simulation of turbulent flow, we observe that in situ operations significantly limit the strong scalability of the code, reducing the relative parallel efficiency to only ≈ 21 % on 2048 cores (the relative efficiency of Nek5000 without in situ operations is ≈ 99 % ). Through profiling with Arm MAP, we identified a bottleneck in the image composition step (that uses the Radix-kr algorithm) where a majority of the time is spent on MPI communication. We also identified an imbalance of in situ processing time between rank 0 and all other ranks. In our case, better scaling and load-balancing in the parallel image composition would considerably improve the performance of Nek5000 with in situ capabilities. In general, the result of this study highlights the technical challenges posed by the integration of high-performance simulation codes and data-analysis libraries and their practical use in complex cases, even when efficient algorithms already exist for a certain application scenario.

Enhancing computational fluid dynamics with machine learning.

Machine learning is rapidly becoming a core technology for scientific computing, with numerous opportunities to advance the field of computational fluid dynamics. Here we highlight some of the areas of highest potential impact, including to accelerate direct numerical simulations, to improve turbulence closure modeling and to develop enhanced reduced-order models. We also discuss emerging areas of machine learning that are promising for computational fluid dynamics, as well as some potential limitations that should be taken into account.

International Scientific Collaboration Is Needed to Bridge Science to Society: USERN2020 Consensus Statement.

Scientific collaboration has been a critical aspect of the development of all fields of science, particularly clinical medicine. It is well understood that myriads of benefits can be yielded by interdisciplinary and international collaboration. For instance, our rapidly growing knowledge on COVID-19 and vaccine development could not be attained without expanded collaborative activities. However, achieving fruitful results requires mastering specific tactics in collaborative efforts. These activities can enhance our knowledge, which ultimately benefits society. In addition to tackling the issue of the invisible border between different countries, institutes, and disciplines, the border between the scientific community and society needs to be addressed as well. International and transdisciplinary approaches can potentially be the best solution for bridging science and society. The Universal Scientific Education and Research Network (USERN) is a non-governmental, non-profit organization and network to promote professional, scientific research and education worldwide. The fifth annual congress of USERN was held in Tehran, Iran, in a hybrid manner on November 7-10, 2020, with key aims of bridging science to society and facilitating borderless science. Among speakers of the congress, a group of top scientists unanimously agreed on The USERN 2020 consensus, which is drafted with the goal of connecting society with scientific scholars and facilitating international and interdisciplinary scientific activities in all fields, including clinical medicine.

Regulating artificial-intelligence applications to achieve the sustainable development goals.

Artificial intelligence is producing a revolution with increasing impacts on the people, planet, and prosperity. This perspective illustrates some of the AI applications that can accelerate the achievement of the United Nations Sustainable Development Goals (SDGs) and highlights some of the considerations that could hinder the efforts towards them. In this context, we strongly support the development of an 18thSDG on digital technologies. This emphasizes the importance of establishing standard AI guidelines and regulations for the beneficial applications of AI. Such regulations should focus on concrete applications of AI, rather than generally on AI technology, to facilitate both AI development and enforceability of legal implications.

The role of artificial intelligence in achieving the Sustainable Development Goals.

The emergence of artificial intelligence (AI) and its progressively wider impact on many sectors requires an assessment of its effect on the achievement of the Sustainable Development Goals. Using a consensus-based expert elicitation process, we find that AI can enable the accomplishment of 134 targets across all the goals, but it may also inhibit 59 targets. However, current research foci overlook important aspects. The fast development of AI needs to be supported by the necessary regulatory insight and oversight for AI-based technologies to enable sustainable development. Failure to do so could result in gaps in transparency, safety, and ethical standards.

A socio-technical framework for digital contact tracing.

In their efforts to tackle the COVID-19 crisis, decision makers are considering the development and use of smartphone applications for contact tracing. Even though these applications differ in technology and methods, there is an increasing concern about their implications for privacy and human rights. Here we propose a framework to evaluate their suitability in terms of impact on the users, employed technology and governance methods. We illustrate its usage with three applications, and with the European Data Protection Board (EDPB) guidelines, highlighting their limitations.

Revisiting History Effects in Adverse-Pressure-Gradient Turbulent Boundary Layers.

The goal of this study is to present a first step towards establishing criteria aimed at assessing whether a particular adverse-pressure-gradient (APG) turbulent boundary layer (TBL) can be considered well-behaved, i.e., whether it is independent of the inflow conditions and is exempt of numerical or experimental artifacts. To this end, we analyzed several high-quality datasets, including in-house numerical databases of APG TBLs developing over flat-plates and the suction side of a wing section, and five studies available in the literature. Due to the impact of the flow history on the particular state of the boundary layer, we developed three criteria of convergence to well-behaved conditions, to be used depending on the particular case under study. (i) In the first criterion, we develop empirical correlations defining the Re𝜃 -evolution of the skin-friction coefficient and the shape factor in APG TBLs with constant values of the Clauser pressure-gradient parameter ß = 1 and 2 (note that ß = δ∗/τw dPe /dx, where δ∗ is the displacement thickness, τw the wall-shear stress and dPe /dx the streamwise pressure gradient). (ii) In the second one, we propose a predictive method to obtain the skin-friction curve corresponding to an APG TBL subjected to any streamwise evolution of ß, based only on data from zero-pressure-gradient TBLs. (iii) The third method relies on the diagnostic-plot concept modified with the shape factor, which scales APG TBLs subjected to a wide range of pressure-gradient conditions. These three criteria allow to ensure the correct flow development of a particular TBL, and thus to separate history and pressure-gradient effects in the analysis.

Adverse-Pressure-Gradient Effects on Turbulent Boundary Layers: Statistics and Flow-Field Organization.

This manuscripts presents a study on adverse-pressure-gradient turbulent boundary layers under different Reynolds-number and pressure-gradient conditions. In this work we performed Particle Image Velocimetry (PIV) measurements supplemented with Large-Eddy Simulations in order to have a dataset covering a range of displacement-thickness-based Reynolds-number 2300

Pressure-Gradient Turbulent Boundary Layers Developing Around a Wing Section.

A direct numerical simulation database of the flow around a NACA4412 wing section at Rec = 400,000 and 5∘ angle of attack (Hosseini et al. Int. J. Heat Fluid Flow 61, 117-128, 2016), obtained with the spectral-element code Nek5000, is analyzed. The Clauser pressure-gradient parameter ß ranges from ≃ 0 and 85 on the suction side, and from 0 to - 0.25 on the pressure side of the wing. The maximum Re𝜃 and Reτ values are around 2,800 and 373 on the suction side, respectively, whereas on the pressure side these values are 818 and 346. Comparisons between the suction side with zero-pressure-gradient turbulent boundary layer data show larger values of the shape factor and a lower skin friction, both connected with the fact that the adverse pressure gradient present on the suction side of the wing increases the wall-normal convection. The adverse-pressure-gradient boundary layer also exhibits a more prominent wake region, the development of an outer peak in the Reynolds-stress tensor components, and increased production and dissipation across the boundary layer. All these effects are connected with the fact that the large-scale motions of the flow become relatively more intense due to the adverse pressure gradient, as apparent from spanwise premultiplied power-spectral density maps. The emergence of an outer spectral peak is observed at ß values of around 4 for λz ≃ 0.65δ99, closer to the wall than the spectral outer peak observed in zero-pressure-gradient turbulent boundary layers at higher Re𝜃 . The effect of the slight favorable pressure gradient present on the pressure side of the wing is opposite the one of the adverse pressure gradient, leading to less energetic outer-layer structures.