Turbulent Flow Field Research Articles

Comparison and analysis of various deep learning models for super-resolution reconstruction of turbulent flows

Single image super-resolution (SR) has become a promising research topic, with many deep learning-based models invented to reconstruct high-fidelity high-resolution (HR) images from low-resolution (LR) images. Motivated by a large amount of turbulent flow field data collected by experimental measurements and numerical simulation, researchers begin investigating the application of these data-driven deep learning models to conduct SR reconstruction of LR flow field data. Due to the limitations of experimental equipment and computing power, sometimes researchers can only obtain LR data. However, deep learning models can quickly reconstruct HR spatial-temporal turbulent data from LR data so that researchers can easily conduct further qualitative and quantitative analyses. This article reviews the development of flow field data SR reconstruction models and the problems encountered from the two aspects of network structure and loss function definition. Finally, we propose the research direction of applying the conditional generative adversarial network (cGAN) to turbulent flow SR reconstruction since seldom study has been conducted in this field.

Atmospheric Turbulence Degraded Video Restoration with Recurrent GAN (ATVR-GAN).

Atmospheric turbulence (AT) can change the path and direction of light during video capturing of a target in space due to the random motion of the turbulent medium, a phenomenon that is most noticeable when shooting videos at long ranges, resulting in severe video dynamic distortion and blur. To mitigate geometric distortion and reduce spatially and temporally varying blur, we propose a novel Atmospheric Turbulence Video Restoration Generative Adversarial Network (ATVR-GAN) with a specialized Recurrent Neural Network (RNN) generator, which is trained to predict the scene's turbulent optical flow (OF) field and utilizes a recurrent structure to catch both spatial and temporal dependencies. The new architecture is trained using a newly combined loss function that counts for the spatiotemporal distortions, specifically tailored to the AT problem. Our network was tested on synthetic and real imaging data and compared against leading algorithms in the field of AT mitigation and image restoration. The proposed method outperformed these methods for both synthetic and real data examined.

Effect of Turbulence and Diffusional–Thermal Instability on Hydrogen‐Rich Syngas Turbulent Premixed Flame

As a kind of fuel with low‐carbon emissions, hydrogen‐rich syngas has a large resource potential and is a promising alternative energy resource. A series of experiments in a constant volume combustion bomb are carried out to investigate the effects of turbulence and diffusional–thermal (DT) instability on a hydrogen‐rich syngas turbulent premixed flame. The flow field in the constant volume combustion bomb is calibrated, and the turbulent flow field's homogeneity and isotropy were investigated. The effects of different speeds of fan (1366–4273 rpm), hydrogen fractions (55%–95%), pressures (1.0–3.0 bar), and equivalence ratios (0.6–1.0) on the hydrogen‐rich syngas turbulent premixed flame are studied. According to the analyses, the flow field in the experimental device exhibits good homogeneous and isotropic characteristics. With the increase of turbulence intensity, equivalence ratio, hydrogen fraction, or pressure, the turbulent flame propagation speed increases gradually. As the hydrogen fraction increases or the equivalence ratio decreases, the effective Lewis number decreases, and the effect of DT instability on the flame increases. The research in this article is crucial for the clean and efficient utilization of hydrogen‐rich syngas and the design of related burners.

Numerical simulation of the dynamic turbulent flow field through a channel provided with baffles: comparative study between two models of baffles: transverse plane and trapezoidal

Une contribution à l’étude numérique d’un écoulement turbulent d’air en présence des chicanes transversales a été menue. Deux différentes formes de chicanes rectangulaire plane et trapézoïdale, disposées en chevauchement dans une conduite de section rectangulaire est présentée dans cet article. Les équations gouvernantes, basées sur le modèle k - à bas nombre de Reynolds employé pour modéliser la turbulence, sont résolues par la méthode des volumes finis à l’aide de l’utilisation de l’algorithme SIMPLEC. Les profils de vitesse ont été obtenus pour toute la géométrie considérée et pour différentes sections choisis, à savoir, en amont, en aval et entre les deux chicanes, ainsi que les coefficients de frottement ont été obtenues pour différentes sections et pour différents nombres de Reynolds.

Turbulent Flow Field around a Cylindrical Pier on a Gravel Bed

Turbulent Flow Field around a Cylindrical Pier on a Gravel Bed

Theoretical prediction of unsteady wall pressure during turbulence-airfoil-interaction

The turbulence-airfoil-interaction generates unsteady surface pressure, which is an unwanted source of noise in many aerospace and marine applications. Consequently, an accurate methodology is desired to calculate the wall-pressure-spectrum along the chord to create new technology such as quiet wind turbine and drone propeller blades. Theoretical approaches are ideal as they not only provide an efficient means of predicting wall pressure spectra, but also give improved insights into the physical mechanisms controlling noise generation. This paper will explore the accuracy of current theoretical approaches to unsteady wall pressure prediction. It will do this by comparing experimental measurements of wall pressure spectra in a variety of turbulent flow fields with theory. The results show the limitations of common theoretical approaches and the importance of including the effects of turbulence distortion in the theoretical model.

Research on the Influence of Turbulent Flow Induced by Dunes on AUVs

The demand for oceanic resource exploration and development is increasing, and autonomous underwater vehicles (AUVs) have emerged as potential tools for ocean exploration. To meet specific operational requirements, AUVs are required to operate in close proximity to the seafloor. However, complex and changeable seafloor terrains have complex spatiotemporal characteristics, which produce unsteady hydrodynamic interference on AUVs, affecting their safety and stability. In order to investigate the hydrodynamic characteristics of AUVs in turbulent flow fields near seafloor terrains, this study selects a typical dune terrain. Based on the computational fluid dynamics (CFD) method, the flow field around an AUV near the seafloor is simulated, and the flow field is solved using the large eddy simulation (LES) method, calculating the AUV’s hydrodynamic instantaneous and average characteristics, as well as flow field characteristics at different inflow velocities and distances from the seafloor. The results indicate that when the AUV is situated within the turbulent flow field caused by the terrain, its drag, lift, and pitch moment performance present significant fluctuations. Variations in total resistance are primarily caused by variations in pressure resistance. Variations in lift and pitch moment are more sensitive to variations in the flow field structure. The AUV also affects the development of the turbulent flow field near the terrain. When the AUV is in close proximity to the seafloor, it hampers the development of vortices below it. These results will offer guidance for the maneuverability and control of AUVs in practical engineering applications.

Numerical study on sealing performance after wear of plug valve during opening–closing process

The failure of the plug valve seal may result in dangerous accidents in the production and transportation process of petroleum and chemical products. In the paper at hand, the operating life of the plug valve is studied in the opening–closing process. The adhesion wear between the sealing surface of the plug valve is numerically studied in micro-scale, and the contact force between the sealing surface is balanced with the internal flow field of the plug valve and the deformation force between the set screw and the spool. The internal flow field of the plug valve in opening–closing process is calculated by finite element method. The deformation coupling model between the set screw and the spool is built to analyze the structure and the mechanics of the plug valve. Then, a leakage identification algorithm based on auto-correlation function is implemented for analyzing contact stress of the plug valve's sealing surfaces after wear. Finally, taking plug valve F-2 ISO-STANDARD as an example, the force of turbulent flow field on the spool is calculated during the closing process. Results show, that while the maximum contact stress increases, the pressure decreases, and the displacement trajectory of spool increases. The operating life decreases by increasing the roughness of the spool sealing surface. The model proposed provides a new practical method to evaluate the operating life of the plug valve, which is a good guidance for the design of the valve.

Characterization of refill region and mixing state immediately ahead of a hydrogen-air rotating detonation using LES

The extraordinary complexity of real Rotating Detonation Combustors (RDC) demands a deep knowledge of each phenomenon involved in the wave development and propagation. Since the reactants are typically injected separately, a key element for the combustor design optimization is understanding the refilling process and the reactants mixing. However, due to the harsh environment and high working frequencies of RDCs, the experimental diagnostics is usually limited, so high-fidelity simulations represent an essential tool to complement the measurements with detailed insights into the flow. In the present work, the non-premixed RDC installed at TU Berlin is simulated with the AVBP code by solving the fully-compressible, spatially-filtered reactive Navier-Stokes equations. The complete hydrogen and air injection system is included in the numerical model to accurately describe both the reactants mixing and the complex turbulent flow field in the resulting refill region. This study shows how both the injection system configuration and its transient interaction with the wave are fundamental for the reactants mixing, as they directly influence the refilled gas properties. Limiting the imbalances between the blockage dynamics of the fuel and oxidizer ducts and optimizing the fuel injectors can improve considerably the homogeneity of the fresh mixture, and consequently the leading shock strength and reasonably the pressure gain. In fact, the detailed analysis of the detonation front speed shown a higher instability near the chamber base for the periodic presence of unmixed reactants. Nevertheless, the unstable root of the front does not affect the whole wave speed, and remaining part propagates steadily thanks to the tangential mixture uniformity. Moreover, the local speed distribution does not appear directly related to the small-scale mixture properties, indicating a higher sensibility to the annulus curvature rather than to the local gas state.

Investigation of Lagrangian Areas of Minimal Stretching (LAMS) in a turbulent boundary layer

In this paper the three-dimensional finite-time Lyapunov exponent (FTLE) field of a direct numerical simulation of a flat-plate turbulent boundary layer is analysed in several wall-parallel sections. The data consider a case at a low subsonic Mach number with a moderate positive pressure gradient in the streamwise direction. In contrast to other studies mainly focusing on the maxima of the FTLE field, particular emphasis is placed on the regions of minimal stretching between the vortices and shear layers of the three-dimensional turbulent flow field. These visually appear as contiguous islands or ‘valleys’ between the ‘ridges’ of the FTLE maxima, both at forward and backward integration of the flow field in time. To clearly distinguish the structures investigated from their more common counterparts (e.g. Lagrangian coherent structures, LCS), the acronym LAMS (Lagrangian areas of minimal stretching) is proposed to denote the associated cohesive fluid regions. Consistent with intuition, the largest LAMS occur near the boundary-layer edge, where large regions of homogeneous laminar external flow coexist with upwelling turbulent structures. Compensating for turbulent regions pushing upward, they sink from there down toward the wall, becoming smaller and longer. This process is associated with an increased relative velocity of the LAMS compared with the mean flow, which is observed over the whole boundary layer in the range $y^+ \gtrsim 10$ . Furthermore, it is observed that the Q4 (sweep) events contained in the LAMS clearly dominate over Q2 (ejection) events above $y^+ \approx 10$ . Thereby, local maxima occur at $y^+ \approx 20$ and near the boundary-layer edge. Below $y^+ \approx 10$ , the relationship reverses. Sweeping LAMS from above $y^+ \approx 10$ and ejecting LAMS from below meet in the layer where the maximal vortical activity occurs. The latter is caused by mostly streamwise oriented vortices with maximal vortex stretching in the streamwise direction. Overall, LAMS are associated with cohesive fluid regions between the surrounding vortices and shear layers that both drop down from the boundary-layer edge toward the wall in the outer region of the boundary layer and lift from the wall in the near-wall region.

2D-supervised fast neural fluid reconstruction technique for time-resolved volumetric flame reconstruction

The volumetric tomography reconstruction technique (VTRT) can calculate and restore the three-dimensional physical field based on the multi-view projection information of the flame and similar flow fields, which plays an important role in the research of volumetric targets. However, the reconstruction of three-dimensional scalar/vector fields representing the physical information of the flow field (temperature, density, chemical composition, etc.) requires a significant amount of computational resources and time, especially in the turbulent flame flow field, which requires high spatial and temporal resolutions. Based on the said background, in this work, we proposed a Fast Neural Fluid Reconstruction Technique (Fast-NFRT) based on deep learning. In this investigation, we first test the reconstruction accuracy, speed, and noise robustness of Fast-NFRT using numerically simulated flames. Then Fast-NFRT is used to reconstruct the experimental turbulent jet flame under two different conditions. Finally, with the camera settings preserved, the Fast-NFRT model was tested for transfer learning between the numerical simulation flame and the two experimental jet flames to examine the generalization performance of the reconstruction. It is found that the proposed Fast-NFRT model can achieve a temporal resolution of 50-500 fps with a similar reconstruction fidelity to traditional algebraic reconstruction methods, which demonstrates the capacity of the Fast-NFRT model and its potential in real-time reconstruction applications and dynamic analysis for complex flow dynamics.

Modeling Unsteady Surface Pressure Autospectra for Turbulent Boundary-Layer Flow over Small Dense Roughness

This work concentrates on analyzing and modeling the unsteady surface pressure autospectra beneath zero-pressure gradient, turbulent boundary-layer flowfields with variably rough flow surfaces. These surface conditions are qualitatively described as being for small, dense, rigid, spatially homogeneous roughness topologies, along with corresponding quantitative parameters. Turbulent boundary-layer velocity and unsteady surface pressure data from a new wind tunnel test with multiple surface conditions are analyzed in conjunction with the compiled results from five, distinct, additional references. These compiled results are utilized to develop and propose a new empirical model for the unsteady surface pressure autospectra for a range of variably rough surfaces. This new modeling framework is approached as an extension of the Goody model, which pertains to smooth-wall flows. The variable character of the newly proposed model is controlled by parameters based on viscous scaling of the effective roughness height and friction velocity such that the surface pressure model inputs are based on readily available mean properties of the turbulent boundary-layer flowfield. For the data utilized in this analysis the approximate values of viscous scaled effective roughness height range between approximately 100 and 2000. Finally, this newly proposed model is used to compare to the unsteady surface pressure data of two additional and separate experimental references for rough-wall flows, yielding further positive comparisons as well as isolated examples of the sensitivity of analysis and modeling to the determined friction velocity for a flow configuration.

Nonlinear reduced-order modeling for three-dimensional turbulent flow by large-scale machine learning

A large-scale machine learning-based nonlinear reduced-order modeling method was developed for a three-dimensional turbulent flow field (Re=1000) using a neural-network with unsupervised learning. First, a mode decomposition method was applied to three-dimensional flow field data using a convolutional-autoencoder-like neural network. Then, a reduced-order model (ROM) was constructed using long short-term memory neural networks (LSTMs). Consequently, it was demonstrated that the time evolution of the turbulent three-dimensional flow field can be simulated at a significantly lower cost (approximately three orders of magnitude) without a major loss in accuracy. However, neural-network-based mode decomposition for the three-dimensional flow field requires a huge computational cost in terms of calculation and memory usage. Therefore, a distributed machine-learning method was implemented using a hybrid parallelism scheme tailored to the network structure. Thus, it was possible to decompose 1.7 million cells of the three-dimensional flow field data into 64 modes and reproduce those with sufficient accuracy. In this study, a uniform flow around a circular cylinder model was used as a test case. To validate the method, the reduction performance of the proposed mode decomposition method was compared with the proper orthogonal decomposition (POD) method. Furthermore, the target flow field was reproduced using ROM, and the reconstruction accuracy was evaluated in terms of various criteria compared with the accuracy based on POD in conjunction with Galerkin projection method.

Visualization of Turbulent Flow Fields Around a Circular Cylinder at Reynolds Number 1.4×105 Using PIV

This study investigates the experimental parameters of particle image velocimetry (PIV) to enhance the measurement technique for turbulent flow fields around a circular cylinder at a Reynolds number (Re) of 1.4×10<sup>5</sup>. At the Korea Research Institute of Ships & Ocean Engineering (KRISO), we utilized the cavitation tunnel and PIV system to capture the instantaneous flow fields and statistically obtained the mean flow fields. An aspect ratio and blockage ratio of 16.7% and 6.0%, respectively, were considered to minimize the tunnel wall effect on the cylinder wakes. The optimal values of the pulse time and the number of flow fields were determined by comparing the contours of mean streamlines, velocities, Reynolds shear stresses, and turbulent kinetic energy under their different values to ensure accurate and converged results. Based on the findings, we recommend a pulse time of 45 μs corresponding to a particle moving time of 3–4 pixels, and at least 3,000 instantaneous flow fields to accurately obtain the mean flow fields. The results of the present study agree well with those of previous studies that examined the end of the subcritical flow regime.

Linear global stability analysis of a transient swirling flow leading to the formation of a helical precessing vortex breakdown in a straight diffuser

Vortex breakdown occurs in highly swirling flows in various engineering applications, such as hydraulic turbines working in either high-load or part-load conditions, leading to pressure pulsations and performance degradation. In the present study, the formation of a helical vortex breakdown having rotational and plunging components (axial oscillation) in a straight diffuser is investigated with transient simulation from its columnar state to a helical vortex breakdown state. It is similar to the helical precessing vortex rope occurring in hydraulic turbines operating at part-load condition. In addition, considering vortex breakdown as a global instability, a linear global stability analysis of the time-averaged turbulent flow field is conducted for different flow states. The main objective is to understand the role of plunging mode in vortex rope formation in straight draft tubes and identify its differences with the one observed in hydraulic turbines with elbow draft tubes operating at part-load condition. Results showed that in contrast to the elbow draft tubes where the plunging mode appears earlier than the rotating mode, the transient simulation showed that the plunging mode appears after several vortex rope rotations in the straight draft tube. This is confirmed by the stability analysis showing that the flow is unstable to the asymmetrical disturbances with a frequency close to the rotating frequency not the plunging ones. Finally, the stability analysis showed a higher critical flow rate than the one obtained from the contour of local swirl number and time history of pressure and velocity.

A numerical investigation of the aerodynamic and aeroacoustic interactions between components of a multi-rotor vehicle for urban air mobility

In this work, we conduct a numerical investigation of a hexacopter for urban air mobility, emphasizing the interactional effect between different components of the vehicle on the aerodynamic and aeroacoustic performance. The nearfield turbulent flow fields are resolved by delayed detached eddy simulations, and the acoustic waves at far-field observers are computed using an integral solution of the Ffowcs-Williams and Hawkings equation. The results show that the close placement of rotors can lead to remarkable thrust fluctuations and slightly reduce the average thrust. Consequently, the sound pressure level for the tonal components is considerably increased over a wide frequency range, especially at the observers in the upstream and downstream directions. The rotor–rotor interaction also results in a contracted wake geometry towards the vehicle centre and intensifies the dissipation of the vortices in the downstream wake. The presence of connecting arms can cause significant thrust fluctuations on the surfaces of both rotor blades and arms, which produces additional tonal noise at high-order harmonics of the blade passing frequency. In contrast, for current configuration, the existence of the fuselage only slightly reduces the total thrust and contributes little to noise generation.

Scour Development Around an Oblong Bridge Pier: A Numerical and Experimental Study

The complex flow structure around bridge piers is challenging for both experimental and numerical studies. Therefore, investigating the capabilities of Computational Fluid Dynamics (CFD) tools in resolving the flow structure and the mechanism of sediment entrainment into and out of the scour hole remains a challenging task. In this study, the scour depth around an oblong bridge pier and the bed shear stress distributions in time and space were numerically investigated using the Computational Fluid Dynamics (CFD) tool Sediment Simulation In Intakes with Multiblock option (SSIIM). Clear water scour conditions and sand of known granulometric composition were considered in accordance with the experimental study carried out. Laboratory data and the results of a scour characterization around a 0.11 m wide oblong bridge pier were considered to calibrate and validate the numerical model. The averaged form of the Navier–Stokes equations was considered to simulate the turbulent flow fields in anticipation of long time scales. The results show that calibrated numerical models can reproduce measured scour depths in the laboratory environment with considerable accuracy, with an average relative error of less than 3%, especially around oblong bridge piers.

Analysis of Available Yawing Moment of an Autonomous Underwater Vehicle Model in Simulation Frame

Research endeavors on the design and control techniques of Autonomous Underwater Vehicles (AUVs) have been going on for a long time. In the present study, the yaw motion of a small submerged underwater vehicle is investigated and visualized as a direct result of changes in the rudder tilt angle and forward velocity. The numerical analysis is performed in ANSYS-Fluent software. The turbulent flow field has been modeled using Shear Stress Transport (SST) k-ω model. A grid independence test has been conducted to ensure the validity of the findings. The forces on the rudder and the available yaw moment have been obtained for different combinations of the AUV’s rudder tilt angle and forward velocity. The trend has intuitively been consistent and agreed with the basic concept of hydrodynamics

Experimental investigation on two-dimensional aerodynamic admittances of rectangular cylinders in various turbulent flows

This paper conducts an experimental investigation of two-dimensional aerodynamic admittances (2D AAFs) for rectangular 5:1 cylinders within various turbulent flow fields. The determination of 2D AAF is achieved by removing the influence of the three-dimensional effect (3D effect) from the traditional AAF, wherein the traditional AAF can be straightly derived based on the ratio of the one-dimensional fluctuating force spectrum to the one-dimensional turbulent wind spectrum. In line with prior research, the values of traditional AAFs show variations in response to the 3D effect in different turbulent flow fields. Moreover, within homo-turbulence conditions, or more specifically, under similar turbulence intensity, the traditional AAFs differ depending on the turbulent integral scale to the model characteristic width ratios (dimensionless turbulent integral scale). Compared to the traditional AAF, the 2D AAF effectively cuts down on the discrepancies arising from the dimensionless turbulent integral scale. In cases where different dimensionless turbulent integral scales are present but the turbulence intensity remains constant, the 2D AAFs are almost consistent. Nevertheless, it can be observed that the 2D AAF of the rectangular 5:1 cylinder is still affected by turbulence intensity. For different turbulence intensities, the 2D AAFs have certain changes. When there is less turbulence intensity, it frequently approaches the quasi-steady value, and as the turbulence intensity increases, it gradually approaches the Sears function.

Numerical Predictions of Soot Formation in Kerosene/Air Jet Diffusion Flame

Soot volume fraction in a turbulent diffusion flame burning kerosene/air is simulated using steady flamelet approach. The soot inception model based on the formation rate of three and two ring aromatics is used, which further accounts for inception, coagulation, surface growth, and oxidation processes. The Favre-averaged governing equations of mass, momentum, and energy are solved in conjunction with k-? turbulence model in ANSYS-FLUENT. The fairly detailed reaction mechanism for kerosene/air is coupled to the turbulent flow field by steady laminar flamelet approach, while the soot is calculated using Moss-Brookes model. The predictions are found to match well with the measurements where no significant effects are observed with the inclusion of higher order hydrocarbon as precursors and surface growth.