Reynolds-averaged Navier-Stokes Model Research Articles

Generalizability evaluation of k-ε models calibrated by using ensemble Kalman filtering for urban airflow and airborne contaminant dispersion

The low accuracy of Reynolds-averaged Navier-Stokes (RANS) modeling of urban airflow and airborne pollutant dispersion is attributed to model flaws and uncertainty contributed by closure coefficients. Previous studies have attempted to improve the performance of RANS modeling by ad hoc calibration of the turbulence model coefficients specifically for urban problems. However, these models failed to accurately reproduce the key features like the reattachment lengths. In addition, there was a lack of generalizability evaluations of the calibrated models. To optimize the accuracy and generalizability of the turbulence model, this study considered the effects of optimization objectives and model formulations. Six k-ε based turbulence models calibrated using the ensemble Kalman filtering (EnKF) approach were compared. A wind tunnel experiment consisting of key features of the airflow around a single-building model was conducted to provide the training data. The proposed optimization objective prioritizing the reproduction of the reattachment lengths in the calibrations enabled the calibrated models to capture the key features of the airflow with better accuracy. The six turbulence models before and after the calibrations were compared for a single-building test case and a building-array test case with respect to the reattachment lengths, velocity, turbulence kinetic energy, and airborne contaminant concentration. The calibrated models with different formulations exhibit distinctive generalizability. The calibrated Murakami-Mochida-Kondo k-ε model (MMK) exhibited strong potential for generalizability to single-building problems. However, the generalization from the single-building isolated roughness flow regime to the street canyon skimming flow regime is limited.

Estimates of turbulence modeling uncertainties in NACA65 cascade flow predictions by Bayesian model-scenario averaging

PurposeThe Reynolds-averaged Navier–Stokes (RANS) equations represent the computational workhorse for engineering design, despite their numerous flaws. Improving and quantifying the uncertainties associated with RANS models is particularly critical in view of the analysis and optimization of complex turbomachinery flows.Design/methodology/approachFirst, an efficient strategy is introduced for calibrating turbulence model coefficients from high-fidelity data. The results are highly sensitive to the flow configuration (called a calibration scenario) used to inform the coefficients. Second, the bias introduced by the choice of a specific turbulence model is reduced by constructing a mixture model by means of Bayesian model-scenario averaging (BMSA). The BMSA model makes predictions of flows not included in the calibration scenarios as a probability-weighted average of a set of competing turbulence models, each supplemented with multiple sets of closure coefficients inferred from alternative calibration scenarios.FindingsDifferent choices for the scenario probabilities are assessed for the prediction of the NACA65 V103 cascade at off-design conditions. In all cases, BMSA improves the solution accuracy with respect to the baseline turbulence models, and the estimated uncertainty intervals encompass reasonably well the reference data. The BMSA results were found to be little sensitive to the user-defined scenario-weighting criterion, both in terms of average prediction and of estimated confidence intervals.Originality/valueA delicate step in the BMSA is the selection of suitable scenario-weighting criteria, i.e. suitable prior probability mass functions (PMFs) for the calibration scenarios. The role of such PMFs is to assign higher probability to calibration scenarios more likely to provide an accurate estimate of model coefficients for the new flow. In this paper, three mixture models are constructed, based on alternative choices of the scenario probabilities. The authors then compare the capabilities of three different criteria.

Dynamic calibration of differential equations using machine learning, with application to turbulence models

We present a methodology for calibration of parametric ordinary and partial differential equation models, using off-the-shelf software for back-propagation in Neural Networks (NN). As a prototypical example, we consider calibration of a Reynolds-averaged Navier-Stokes (RANS) turbulence closure model, against ground truth data from direct numerical simulations (DNS) of two different turbulent flows. Numerical time integration is represented as a custom NN, where only the RANS model parameters are trainable. A loss function is defined to quantify the mismatch between the NN prediction and the ground truth over a predefined, finite time integration window. This loss function is then minimized using a gradient descent method utilizing the back-propagation algorithm. This dynamic approach to training is to be contrasted with a static approach, wherein a least square regression estimate for parameters is obtained in the limit of an infinitesimal time integration window. In a first test of static and dynamic approaches against ground truth data generated by the model, the former proves to be significantly faster and more accurate than the latter at recovering the parameters. When both calibration approaches are tested against DNS data, for which it is known that the model cannot achieve a perfect fit, the static approach yields a good prediction only for short times, while the dynamic approach results in physical and stable predictions over the entire integration window. After optimization of the dynamic approach for time step, spatial resolution, stability, and physics-based constraints, we obtain a 50% improvement of outcomes over those obtained from the existing, manually calibrated set of parameters, demonstrating the merits of this systematic and automated procedure.

Topology optimisation of turbulent flow using data-driven modelling

Fluid topology optimisation has become a popular approach for optimisation of geometries in aero-thermal applications. However, one of the main limitations of current approaches considering turbulent flow is the fidelity of the Reynolds Averaged Navier–Stokes models employed. In response, this paper shows the development of the first data-driven fluid topology optimisation technique based on the continuous adjoint method. The technique first extracts data from a high fidelity simulation of a standard topology-optimised geometry. These data are fed through a symbolic regression-based machine learning algorithm called gene expression programming, to learn an explicit model for the anisotropy tensor. The novel aspect of the work is the derivation of the adjoint form of the generalised explicit algebraic stress model such that the developed turbulence model can be inserted directly into the primal and adjoint system of equations. This allows a second, data-driven optimisation to be performed. Finally, a high fidelity simulation of the resulting geometry is also conducted to allow comparison against the standard geometry. The method is first applied to the back-facing step to verify the approach and then to a 2D u-bend configuration. The data-driven optimisation was able to find geometries exhibiting significant reductions in pressure loss when compared with results from the standard optimisation.

How Turbulence Models Perform in Simulating Pipeflow Response to Targeted Wall-Shapes?

Abstract This study evaluates how Reynolds-averaged Navier–Stokes (RANS) models perform in simulating the characteristics of mean three-dimensional perturbed flows in pipes with targeted wall-shapes. The principal objective of this investigation is to evaluate which of the well-established RANS models can best predict the flow response and recovery characteristics in perturbed pipes at moderate and high Reynolds numbers (1×104−1.58×105). First, the flow profiles at various axial locations are compared between simulations and experiments. This is followed by assessing the well-known mean pipeflow scaling relation in the far downstream region, where the flow obtains a fully-developed state. The consistency of computationally predicted results and their similarities with experiments suggested that the Standard k−ε model can accurately capture the pipeflow characteristics in response to introduced perturbation with smooth sinusoidal axial variations.

Assessment of low- and high-fidelity turbulence models for heat transfer predictions in low-prandlt number flows

Linear eddy viscosity-based Reynolds-averaged Navier-Stokes (RANS), hybrid RANS/Large Eddy Simulation (LES), and LES models are applied for simulations of plane channel flow at Reynolds number Reτ = 150 and 640. Results are obtained for fluids with Prandtl number (Pr) of 0.025 and 0.71, for cases without buoyancy (Gr = 0) and for a vertical channel flow with buoyancy (Ri = 0.1–0.13). The objective is to evaluate the predictive capability of each class of model for low-Pr flow typical of liquid metal reactor cooling systems. For this purpose, RANS, partially-averaged Navier-Stokes (PANS) and LES computations are performed on grids which contain 0.67%, 2%, and 5% of the grid points used in the Direct Numerical Simulation (DNS), respectively, and the predictions of the mean and turbulent flow and thermal quantities are validated against DNS results. RANS results obtained using Kays variable turbulent Prandtl number (PrT) formulation were more accurate than those using constant PrT for low-Pr forced convective flows, but did not play a significant role for mixed convective flows. The variable PrT formulation likewise did not have significant effect for hybrid RANS/LES or LES computations, wherein > 90% turbulence was resolved. The RANS model performed reasonably well for flows without buoyancy, but for mixed convective flows, significantly under predicted the sharp pressure gradient around the mid-channel, and the limitations were identified due to over prediction of heat transfer by wall-normal fluctuations on the aiding side. Both PANS and dynamic Smagorinsky (DSM) showed similar predictions, and under predicted the mid-channel temperature gradient for the low-Pr vertical channel case and over predicted flow turbulence on the opposing flow side and thermal fluctuations on the aiding flow side. The Wall-Adapting Local Eddy-viscosity (WALE) model performed best among the LES subgrid stress models. The improved predictions by WALE over PANS and DSM were identified because of higher modeled dissipation predictions away from the wall. Future work will focus on investigation of non-linear RANS models and extend the validation for higher Re flows.

Assessment of RANS-based turbulence model for forced plume dynamics in a linearly stratified environment

We present a computational study on the dynamics of a forced plume released in a linearly stratified medium for a range of buoyancy frequencies N∞=0 - 0.7 s−1. Three dimensional unsteady Reynolds-Averaged Navier–Stokes (RANS) simulations are performed using the standard k−ε turbulence model. The mean flow parameters such as the maximum height Zm, mean centerline velocity 〈wc〉, mean axial velocity 〈w〉, and turbulence parameters such as shear production (P), buoyancy production (B), and dissipation rate (ε) are compared with the experimental results of Mirajkar et al. (2020) and a good agreement is found. The validated model is then used to study the forced plume characteristics and energetics at various values of N∞. The passive scalar contours reveal that the plume reaches a maximum height in a stratified medium where the momentum becomes zero, then falls back to the neutral buoyancy height before spreading in the lateral direction. We also found that the maximum height reached by the plume decreases with increasing value of N∞, consistent with the previous studies. Quantification of turbulence parameters reveal that the production flux, P, and viscous dissipation, ε, from the RANS computation are in reasonable agreement with the experimental values at N∞ = 0.4 s−1. Further, the residual, given by 〈〈P〉〉 + 〈〈B〉〉 - 〈〈ε〉〉, is found to be larger for higher N∞, indicating more non-homogeneity in the flow at higher stratification. Overall, the RANS modeling is found to accurately predict the mean flow quantities, such as mean centerline velocity and maximum height reached by the plume. The comparison of the modeled turbulence quantities with experimental data seems satisfactory, but indicates a need for some improvement at high values of N∞.

Application of Linear and Nonlinear Two-Equation Turbulence Models in Hypersonic Flows

A nonlinear two-equation model has been implemented in the OpenFOAM framework and has been assessed for the computation of hypersonic flows with shock-wave boundary-layer interactions (SWBLIs), together with two other linear, low-Reynolds number models, namely the Launder–Sharma model and the shear stress transport model. Supersonic and hypersonic flow computations resulting from the use of these three models have been compared with experimental benchmark cases over a range of conditions. The three original models tested do generally return reasonable predictions of wall pressure in most cases, with the nonlinear model resulting in the best capability among the three in predicting flow separation. The wall heat flux in the interaction region, however, is overpredicted, in most cases by all three models (sometimes showing quite a dramatic overprediction). While the overall wall heat flux predictions of the nonlinear model are closer to the measured values, it is nevertheless evident that there is a need for further improvement. To address this need, the inclusion of a new source term in the dissipation rate equation is proposed, which aims to restrict the turbulent length scales in the shock-wave boundary-layer interaction region. This is inactive in incompressible flows and exerts only a minor influence in supersonic SWBLI cases. Computations of a range of supersonic and hypersonic flows with SWBLI show that this inclusion of the proposed source term in the dissipation rate equation, of both the nonlinear and the linear models, has significant effects only in hypersonic flows. These effects are mainly confined to the thermal predictions of these models. In the nonlinear model predictions, the overestimation of the wall heat flux in the SWBLI region is largely eliminated, while, in the corresponding predictions of the linear model, the overestimation is substantially reduced. The cubic nonlinear model tested, with the proposed new source term to the dissipation rate equation, is thus shown to be a very reliable and cost-effective tool for the Reynolds-averaged Navier–Stokes modeling of supersonic and hypersonic flows.

Study on Dynamic Characteristics of Mars Entry Module in Transonic and Supersonic Speeds

The aerodynamic configuration of the Tianwen-1 Mars entry module that adopts a blunt-nosed and short body shape has obvious dynamic instability from transonic to supersonic speeds, which may bring risk to parachute deployment. The unsteady detached eddy of the entry module cannot be accurately simulated by the Reynolds-Averaged Navier-Stokes (RANS) model, while the computational cost for direct numerical simulation (DNS) and large eddy simulation (LES) is huge. It is difficult to implement these methods in the coupled engineering calculation of unsteady flow and motion. This paper proposes the integrated numerical simulation method of computational fluid dynamics and rigid body dynamics (CFD/RBD) based on detached eddy simulation (DES) and calculates and studies the dynamic characteristics of attitude oscillation of the Mars entry module in free flight from transonic to supersonic speeds with one degree of freedom (1-DOF) at small releasing angle of attack. In addition, the unstable range of Mach number and angle of attack are determined, and the effect of different afterbody shapes on dynamic stability is analyzed.

2D/3D RANS and LES Calculations of Natural Convection in a Laterally-Heated Cylindrical Enclosure Using Boussinesq and Temperature-Dependent Formulations

This article presents an investigation of fluid flow and natural convection heat transfer in a cylindrical enclosure heated laterally. Two-dimensional (2D) Reynolds-averaged Navier–Stokes (RANS) equations and three-dimensional (3D) Large Eddy Simulation (LES) calculations are conducted using commercial computational fluid dynamics (CFD) software, ANSYS FLUENT. The Rayleigh number (Ra) = 2E+7 is constant in all of the simulations and is based on a length scale that is equal to the ratio of volume to the lateral area of the reactor, i.e., R/2, where R is the radius of the reactor. The validity of the Boussinesq approximation is analyzed by comparing calculations using both the Boussinesq approximation and temperature-dependent properties (non-Boussinesq approach) using 2D RANS and 3D LES (Dynamic Smagorinsky) formulations. Moreover, 2D axisymmetric k-ω SST RANS model will be implemented to investigate whether the 2D axisymmetric model can give results that are comparable to those of the 3D LES (Dynamic Smagorinsky) model when the corresponding longitudinal or azimuthal cross section are compared. In other words, the validity of using a 2D model instead of a 3D model for the current geometry, flow regime and thermal boundary conditions will be discussed. The flow and temperature contours of these two types of simulations are analyzed, compared to determine the various aspects of each case and discussed for deeper physical insight.

A New Anisotropic Four-Parameter Turbulence Model for Low Prandtl Number Fluids

Due to their interesting thermal properties, liquid metals are widely studied for heat transfer applications where large heat fluxes occur. In the framework of the Reynolds-Averaged Navier–Stokes (RANS) approach, the Simple Gradient Diffusion Hypothesis (SGDH) and the Reynolds Analogy are almost universally invoked for the closure of the turbulent heat flux. Even though these assumptions can represent a reasonable compromise in a wide range of applications, they are not reliable when considering low Prandtl number fluids and/or buoyant flows. More advanced closure models for the turbulent heat flux are required to improve the accuracy of the RANS models dealing with low Prandtl number fluids. In this work, we propose an anisotropic four-parameter turbulence model. The closure of the Reynolds stress tensor and turbulent heat flux is gained through nonlinear models. Particular attention is given to the modeling of dynamical and thermal time scales. Numerical simulations of low Prandtl number fluids have been performed over the plane channel and backward-facing step configurations.

Experiment and RANS modeling of solitary wave impact on a vertical wall mounted on a reef flat

A series of two-dimensional flume experiments were carried out to study turbulent bore impact on the vertical wall mounted on a reef flat. Turbulent bores were generated by solitary waves propagating on typical fringing reef profiles with and without reef crest. The idealized reef model has a 1:4 face slope and a long reef flat, with an impermeable vertical wall at the end. The solitary wave height varies from 14.1% to 30.0% of the water depth to cover a range of nonlinear wave conditions. Datasets of free surface elevation and flow velocity along fringing reef profiles and pressure on the vertical wall are acquired. Corresponding RANS (Reynolds-averaged Navier–Stokes)-type simulations are performed utilizing the open-source CFD (Computational Fluid Dynamics) OpenFOAM (Open Field Operation and Manipulation) package with the SST k−ω turbulence closure solver and validated using experimental data. The characteristics of the main hydrodynamic and impact processes are examined in detail. The predictive capability of OpenFOAM and three existing formulas for calculating the maximum lateral force are evaluated using measured data. A new predictive equation for peak pressure along the vertical wall is also proposed to achieve better agreement with the experimental data.

Exact Eddy-Viscosity Equation for Turbulent Wall Flows—Implications for Computational Fluid Dynamics Models

For turbulent channel flow, pipe flow, and zero-pressure gradient boundary layer, Heinz yielded recently analytical formulas for the eddy viscosity as a product of a function of (the wall-normal distance scaled in inner units) and a function of (the same scaled in outer units). By calculating the eddy-viscosity turbulent diffusion term, an exact high-Reynolds-number equation with one production and two dissipation terms is constructed for those flows. One dissipation term is universal, peaks near the wall, and scales mainly with . The second, smaller one, is flow dependent, peaks in the wake, and scales mainly with . The production term is flow dependent, peaks in between, and scales similarly. The universal dissipation term implies a length scale analogous to the von Karman length scale used in the scale-adaptive simulation models of Menter. This length scale also appears in the production term. This confirms the relevance of these length scales. An asymptotic analysis of all terms in the budget in the limit of infinite Reynolds numbers is provided. This yields a test bench of existing Reynolds-averaged Navier–Stokes models with a similar equation. It is shown that some models, e.g., the one of Spalart and Allmaras, do not respect the flow physics: they display a production peak in the near-wall region. The most promising model, a scale-adaptive simulation model, is modified. As a step forward toward a solution to the wall damping problem, the equation of our model behaves much more correctly in the near-wall region.

Proper orthogonal and dynamic mode decomposition analyses of nonlinear combustion instabilities in a solid-fuel ramjet combustor

In ramjet propulsion systems, syelf-excited combustion instability is highly undesirable. It may lead to violent structural vibration and even mission failure. For this, self-excited combustion instabilities in such SFRJ are investigated numerically by 2D unsteady Reynolds-Average Navier-Stokes (URANS) model and varying the inlet mass flow rate ṁair. When combustion instability occurs, large-amplitude temperature, velocity, and reacting species wavy motions are observed along the axial flow direction. To shed light on the flow features and the energy distribution among the unstable eigenmodes, proper orthogonal (POD) and dynamic mode decomposition (DMD) analyses of nonlinear combustion instabilities in a solid-fuel Ramjet (SFRJ) combustor are conducted. The POD studies reveal that the fluctuation energy of the first six modes contributes to 99% of the total fluctuation energy of all modes. It means that the main features of the SFRJ combustors could be captured by using the first six modes. Unlike POD studies, DMD analysis captures the frequency information and spatial structures in the entire flow field. Examining the DMD growth rate reveals that all dominant DMD eigenmodes are marginally stable (growth rate = 0) or unstable (positive growth rate). And the frequency of each dominant mode could be enhanced by 3–10 Hz as theṁair is increased by 0.2 kg/s. The mode energy decreases with the increase of the mode order, with the decrease up to around 90%. The DMD analysis on velocity field indicated that the flow separation could be observed near the backward facing step. And the DMD mode structure of the temperature field can significantly affect the mode structures of Mach number and YCO2 field. In general, the present work provides detailed analyses of the SFRJ flow field in the presence of self-excited combustion instability by using POD and DMD approaches.

Turbulent Length Scale for Multilayer RANS Model of Urban Canopy and Its Evaluation Based on Large-Eddy Simulations

Large-Eddy Simulation (LES) numerical experiments of neutrally-stratified turbulent flow over an urban-type surface and passive scalar transport by this flow are performed. A simple parameterization of the turbulent length scale containing only one empirical constant is proposed. Multilayer Reynolds-Averaged Navier-Stokes (RANS) model of turbulent flow and turbulent scalar diffusion is constructed. The results of the RANS model are compared with the LES experiments. It is shown that the proposed approach allows predicting the average flow velocity and the scalar concentration inside and above the urban canopy.

CFD simulations of wind flow and pollutant dispersion in a street canyon with traffic flow: Comparison between RANS and LES

• CFD simulations for a street canyon with moving traffic are systematically validated. • Among the five RANS models, SKE has the best performance against wind tunnel data. • All RANS models overestimate pollutant concentration near moving traffics. • LES outperforms the RANS simulation in predicting mean velocity and concentration. • Pollutant concentrations are strongly affected by turbulent diffusion fluxes. The dispersion of traffic pollutants in street canyons takes place under the joint influence of the urban morphology and traffic-induced air motions. Computational fluid dynamics (CFD) are increasingly used to explore air quality problems in cities, however, numerical setup guidance for studies considering moving traffics remains missing. This study presents a systematic evaluation and comparison of steady Reynolds-averaged Navier-Stokes (RANS) and large-eddy simulation (LES) in reproducing the wind flow and pollutant dispersion in a street canyon under traffic flow using the quasi-steady method. The RANS simulations are performed with five turbulence models, and the standard k-ε (SKE) turbulence model is found to yield the best agreement with the wind-tunnel data. The LES simulation with the wall-adapting local eddy-viscosity subgrid-scale model outperforms all RANS models in simulating the mean velocity and mean pollutant concentration, and can accurately predict the turbulent velocity. The counter-gradient turbulent mass flux captured by LES reveals that the gradient-diffusion hypothesis used in steady RANS fails in the regions near moving traffics, which strongly affects the predicted mean concentration. The findings of this study can support future CFD studies on pollutant dispersion in urban environments.

On the performance of RANS turbulence models in predicting static stall over airfoils at high Reynolds numbers

PurposeThis paper aims to conduct, a detailed investigation of various Reynolds averaged Navier–Stokes (RANS) models to study their performance in attached and separated flows. The turbulent flow over two airfoils, namely, National Advisory Committee for Aeronautics (NACA)-0012 and National Aeronautics and Space Administration (NASA) MS(1)-0317 with a static stall setup at a Reynolds number of 6 million, is chosen to investigate these models. The pre-stall and post-stall regions, which are in the range of angles of attack 0°–20°, are simulated.Design/methodology/approachRANS turbulence models with the Boussinesq approximation are the most commonly used cost-effective models for engineering flows. Four RANS models are considered to predict the static stall of two airfoils: Spalart–Allmaras (SA), Menter’s k – ω shear stress transport (SST), k – kL and SA-Bas Cakmakcioglu modified (BCM) transition model. All the simulations are performed on an in-house unstructured-grid compressible flow solver.FindingsAll the turbulence models considered predicted the lift and drag coefficients in good agreement with experimental data for both airfoils in the attached pre-stall region. For the NACA-0012 airfoil, all models except the SA-BCM over-predicted the stall angle by 2°, whereas SA-BCM failed to predict stall. For the NASA MS(1)-0317 airfoil, all models predicted the lift and drag coefficients accurately for attached flow. But the first three models showed even further delayed stall, whereas SA-BCM again did not predict stall.Originality/valueThe numerical results at high Re obtained from this work, especially that of the NASA MS(1)-0317, are new to the literature in the knowledge of the authors. This paper highlights the inability of RANS models to predict the stall phenomenon and suggests a need for improvement in modeling flow physics in near- and post-stall flows.

Data-driven RANS closures for wind turbine wakes under neutral conditions

The state-of-the-art in wind-farm flow-physics modeling is Large Eddy Simulation (LES) which makes accurate predictions of most relevant physics, but requires extensive computational resources. The next-fidelity model types are Reynolds-Averaged Navier–Stokes (RANS) which are two orders of magnitude cheaper, but resolve only mean quantities and model the effect of turbulence. They often fail to accurately predict key effects, such as the wake recovery rate. Custom RANS closures designed for wind-farm wakes exist, but so far do not generalize well: there is substantial room for improvement. In this article we present the first steps towards a systematic data-driven approach to deriving new RANS models in the wind-energy setting. Time-averaged LES data is used as ground-truth, and we first derive optimal corrective fields for the turbulence anisotropy tensor and turbulence kinetic energy (t.k.e.) production. These fields, when injected into the RANS equations (with a baseline k–ɛ model) reproduce the LES mean-quantities. Next we build a custom RANS closure from these corrective fields, using a deterministic symbolic regression method to infer algebraic correction as a function of the (resolved) mean-flow. The result is a new RANS closure, customized to the training data. The potential of the approach is demonstrated under neutral atmospheric conditions for multi-turbine constellations at wind-tunnel scale. The results show significantly improved predictions compared to the baseline closure, for both mean velocity and the t.k.e. fields.

Linear Stability-Based Smooth Reynolds-Averaged Navier–Stokes Transition Model for Aerodynamic Flows

The inclusion of transition to turbulence effects in computational fluid dynamics simulations makes it possible to design laminar flow airframes. It also increases the level of physical representations for simulations of standard airframes because laminar flow regions may be present in parts of the aircraft in multiple flight conditions. Modified Reynolds-averaged Navier–Stokes (RANS) models that consider transition effects became popular in the last decade and indicated favorable agreement with experimental data. However, these models present more difficult convergence behavior when compared with fully turbulent RANS approaches. To address this issue, an approximate Newton–Krylov solver is leveraged to solve the transitional flow over aeronautical configurations using a flow-stability-based, smooth RANS transition model. The amplification factor transport (AFT) model is modified to create a smooth variant, referred to as AFT-S. Strategies have been developed to obtain representative physical solutions that exhibit good agreement with experimental data. The convergence behavior for the transition RANS model is also addressed, and its impact on the numerical results is assessed. These results constitute the first in-depth investigation on the numerical behavior of an AFT-type transition model.

Comparison of low Reynolds number turbulence and conjugate heat transfer modelling for pin-fin roughness elements

This study contrasts Reynolds-Averaged Navier-Stokes (RANS) and (Numerical) Large-Eddy Simulation (NLES) for their use in predicting conjugate heat transfer for a low Reynolds number flow over a surface roughness element. The (N)LES predictions are in good agreement with experimental data, the heat transfer estimate is within 7.7% error. The linear RANS model shows larger errors up to 40%, especially in modelling the turbulent stresses. Using a one-equation LES turbulence model slightly increases the heat transfer prediction compared to a numerical LES. This indicates that more advanced turbulence models might not give more accurate heat transfer predictions.Additionally, this study investigates the time dependant development of the flow and temperature fields and how long data needs to be collected for statistically stationary results. The flow needs to develop for approximately 1800 through-flow times, TL, and 150 TL is required for collection of statistics. A significant range in turbulence length scale prediction between (N)LES and RANS was found. (N)LES scales appeared physically reasonable and should inform mesh resolution in future studies. Two additional cases were run with different cube heights showing changes in the time dependant development of the temperature field and greater sensitivity to turbulence modelling.