Session X10: Turbulence: LES Simulations (10:45am - 11:30am CST)

Large eddy simulation of flow over axisymmetric hull at high Reynolds numbers

Wall-modeled large eddy simulation is performed for flow over an axisymmetric hull geometry at Reynolds numbers (Re) of 1.1, 12 and 67 million, based on hull length and freestream velocity. The domain is chosen to minimize confinement effects and to capture the evolution of the turbulent wake. The no-slip wall boundary condition is replaced by a prescribed stress boundary condition (wall model) to alleviate stringent near-wall resolution requirements at high Re. The prescribed stress is obtained assuming that the near-wall streamwise velocity field satisfies Reichardt’s law. The boundary layer is tripped on the bow region of the hull to make it turbulent, similar to the experiment and the wall model is active throughout the hull, post-tripping. Simulations are performed using two grids, with nominal wall-normal grid resolution of 8 and 16 points respectively per boundary layer thickness in the mid-hull region. The predictive capability of the employed wall model is assessed using available data for the pressure and skin-friction coefficients on the hull, as well as the velocity profiles in stern and wake regions.   

Large-eddy simulation of appended submerged vehicles using an unstructured overset grid method

Appended axisymmetric bodies are of particular interest in marine vehicle applications, in which the wakes and horseshoe vortices of the appendages are ingested by the propulsor. We discuss the application of a novel unstructured overset method to wall-resolved large-eddy simulation (LES) of an appended hull geometry. The numerical algorithm used in the present work is based on that developed by Horne and Mahesh [J. Comput. Phys (2019) 397: 108790] to address the discrete conservation and scaling challenges of overset methods. Simulation results are compared to relevant experimental data and LES of the unappended axisymmetric hull.   

Wall-resolved and wall-modeled LES of active flow control for external aerodynamics

We investigate the aerodynamic performance of active flow control of airfoils and wings using synthetic jets with zero net-mass flow. The study is conducted via wall-resolved and wall-modeled large-eddy simulation using two independent CFD solvers: Alya, a finite-element-based solver; and charLES, a finite-volume-based solver. Our approach is first validated in a NACA4412, for which numerical and experimental results are already available in the literature. The performance of synthetic jets is evaluated for two flow configurations: a SD7003 airfoil at moderate Reynolds number with laminar separation bubble, which is representative of Micro Air Vehicles, and the high-lift configuration of the JAXA Standard Model at realistic Reynolds numbers for landing. In both cases, our predictions indicate that, at high angles of attack, the control successfully eliminates the laminar/turbulent recirculations located downstream the actuator, which increases the aerodynamic performance. Our efforts illustrate the technology-readiness of large-eddy simulation in the design of control strategies for real-world external aerodynamic applications.   

Backscattering and stability of data-driven subgrid-scale parameterization for 2D turbulence

To alleviate the stringent requirement of DNS, large eddy simulation (LES) is one of the favorite alternatives of atmospheric scientists and fluid mechanics engineers. This study investigates data-driven models for parameterization of the subgrid-scale interactions for 2D turbulence. The data-driven models are based on the framework of random forest, fully connected neural network, or convolutional neural network. The models provide a functional joining the resolved flow field in terms of vorticity and stream function to the SGS momentum flux, which acts as a closure of the system. The SGS momentum flux can be obtained from filtered DNS data and used as the target for the training process. The models accurately predict the SGS momentum flux in a priori analysis. However, when coupled back to the LES, the data-driven models may lead to instabilities. The instabilities can be attributed to the backscattering effect, which is a physical process transferring energy from SGS terms to the large scale (grid resolved) terms. Data-driven models can be stabilized by canceling the backscattering effect at post-processing. However, the backscattering cancellation leads to an inaccurate prediction of the SGS term and, eventually, an over-diffusive system. We proposed regularization methods to alleviate the backscatter cancellation while maintaining the stability of the data-driven models.   

Extension of the Integral Length-Scale Approximation (ILSA) model to passive scalar

The Integral Length-Scale Approximation (ILSA) is a versatile LES model in which the turbulence characteristics fully determine the length scale, rather than being related to the grid size. This approach decouples modeling errors from numerical errors, one of the central complications of LES. ILSA was applied successfully to momentum transport over a variety of flows, both internal and external. In this work, the ILSA model is formulated for passive scalar transport and applied to a plane channel flow over a range of Prandtl numbers. The resulting mean scalar and turbulent scalar fluxes are in good agreement with previous studies. A grid convergence study was performed, demonstrating the decoupling of the model from the numerics. In addition, a clear and rigorous procedure is presented to determine the only model coefficient without the need for \it{a-priori} information.   

Large-Eddy Simulation of a Ducted Propeller in Crashback Mode of Operation

We discuss LES of flow around a ducted, marine propeller in the off-design mode of operation known as crashback. In crashback, the propeller rotates in the reverse direction while the vessel moves in the forward direction yielding highly unsteady loads. The simulations use a novel, unstructured grid methodology, developed by Horne {\&} Mahesh [J. Comput. Phys (2019) 376:585-596]. Experiments and previous LES studies have shown that the addition of a duct exacerbates side-forces to about three times the magnitude of the case without a duct. The results are compared to available experimental data and previous LES studies. Details of the flow field and the mechanisms behind the high side-forces are discussed.   

LES Scale Enrichment and its Effect on the Pressure Field

High Reynolds number flows, common in engineering applications, limit the bandwidth of scales available to modern computers when solving the full nonlinear governing equations. This unavailable information is often crucial to design of aerospace systems or wind turbine blades for example. Ghate \& Lele (JFM, v. 819, 2017) have developed a method to enrich turbulence scales below the implied filter width of LES in such a way that local information of subgrid-scale dynamics is accurately represented through the use of spatially- and spectrally-localized Gabor modes. The method relies on a quasi-homogeneous assumption and expands the sub-grid velocity field as a local sum over Gabor modes which evolves dynamically with the large-scale field. We have extended the investigation of the method’s ability to enrich subgrid scales, in the context of stationary homogeneous isotropic turbulence, to its effect on the pressure field where a $32^3$ LES is enriched and compared to an independent $256^3$ LES benchmark case in the infinite Reynolds number limit. This study shows close agreement of the enriched field and its high-resolution counterpart in terms of pressure spectrum, pressure variance, pressure gradient variance, and the space-time correlation of both pressure and pressure gradient.   

A Flow-Based Coordinate Frame Representation for Invariant Data-Driven Subgrid Stress Closure

Large-Eddy Simulations are often used to get accurate statistics of turbulent flows. These simulations rely on an accurate model for Sub-Grid Scale Stress (SGS) tensor to close the system of filtered Navier-Stokes equations. We propose a data-driven approach to model SGS tensor. The model form is designed to ensure Galilean, time, frame, and dimensional invariance, thereby ensuring physical consistency with the filtered Navier-Stokes equations. The proposed model form further allows for a simplified low-cost neural network representation that is trained using a small amount of data obtained from an open-source turbulence database. We show that the data-driven model has superior performance than the classical SGS models for both a priori and a posteriori simulations. We demonstrate that the data-driven model seems to have a generalizable nature as it gives accurate results even for cases that are at a different Reynolds number or involve flow physics that is different than the training data.   

Tempered Fractional LES modeling of Turbulent Flows: A Priori Analysis

Nonlocal behavior of small scale motions and their cumulative effects on the large scale dynamics in isotropic turbulent flows shape the underlying coherent structures and the associated spatial intermittency. Filtering the Navier--Stokes equations in the large eddy simulation of turbulent flows would further enhance the existing nonlocality, emerging in the corresponding subgrid-scale (SGS) fluid motions.~Such long-range effects and anomalous behavior urge the development of nonlocal SGS models. Following [\underline {Samiee, M., et al., 2020. }\underline {\textit{Physics of Fluids}}], we outline a framework to model small scale motions at the kinetic level using a tempered \textit{L\'{e}vy} stable distribution and derive the model at the continuum level. Within this framework, the divergence of SGS motions emerges as a tempered fractional Laplacian of the resolved field. Attention in this study is focused on capturing \textit{two-point} high-order structure functions, governed by the Karman-Howarth equation in isotropic turbulent flows. Through a \textit{statistical a priori} analysis, we study the predictability of the proposed model and the role of tempering fractional model in capturing loss of energy at the dissipation range.   

Fractional LES Subgrid-Scale Modeling for Scalar Turbulence

Filtering the passive scalar transport equation in the large-eddy simulation (LES) of turbulent transport gives rise to closure term corresponding to the unresolved scalar flux. Respecting the statistical features of subgrid-scale (SGS) flux is a vital point in robustness and predictability of LES. We investigate the intrinsic nonlocal behavior of the SGS flux through its two-point statistics obtained from filtered direct numerical simulation (DNS) data in homogeneous isotropic turbulence. Presence of long-range correlations in true SGS flux urges to go beyond the conventional local closure modeling approaches that fail to predict the statistical features of turbulent transport. We propose an appropriate statistical model for microscopic SGS motions in the filtered Boltzmann transport equation (FBTE) for passive scalar by approximating the filtered equilibrium distribution with an $\alpha$-stable Levy distribution. This incorporates a power-law behavior to resemble the observed nonlocal statistics of SGS flux. Ensemble-averaging of such FBTE lets us formulate a continuum level model for the SGS scalar flux appearing in terms of fractional operators that are inherently nonlocal. In an $a-priori$ testing, our model yields a great two-point statistical behavior for SGS scalar flux.   

Modeling and simulation of non-homogeneous turbulence subjected to sudden distortion with finite upstream mean flow shear

The study of the linear evolution of turbulence with small spatial scales began with Prandtl (1933) and Taylor (1935) who analysed the distortion of unsteady flow through a contracting stream. This work was primarily motivated by ensuring optimum wind tunnel performance; i.e. that the flow in the working section downstream of a contraction has very low-intensity turbulence. It is based on the idea that the magnitude of the mean rate of strain (S) multiplied by the eddy turnover time, Teddy=Dinlet/Uinlet, is large inasmuch as S*Teddy << 1. In this talk we perform Large-Eddy Simulations to assess the distortion of turbulence through a sudden area expansion with uniform and non-uniform inflow conditions at two different subsonic Mach numbers based on the bulk flow velocity through the contraction of 0.1 and 0.5. We find that for essentially incompressible flows (Bulk Mach number of 0.1), the initial de-correlation of the two-point velocity correlation function remains the same through the contraction. However, the subsequent temporal decay occurs more slowly for the transverse correlation function components than the streamwise component. In the talk, we discuss these properties using classical Rapid-distortion theory by solving the appropriate inhomogeneous Poisson equation.