Influence of turbulence model parameter settings on conjugate heat transfer simulation

Corner separation is one type of the three-dimensional (3D) separated flows which is commonly observed at the junction of the blade suction surface and endwall of an axial compressor. The commonly used Reynolds-Averaged Navier-Stokes (RANS) turbulence models, namely Spalart-Allmaras (SA) and Menter’s Shear Stress Transport (SST) models, have been found to overpredict the size of corner separation. The physical reason is partly attributed to the underestimation of turbulence mixing between the mainstream flow and the endwall boundary-layer flow. This makes the endwall boundary layer unable to withstand the bulk adverse pressure gradients, and in turn leads to its premature separation from the endwall surface during its migration towards the endwall/blade suction surface corner. The endwall flow characteristics within the compressor stator cascade are then studied to facilitate understanding the physical mechanisms that drive the formation of 3D flow structures, and the physical reasons that lead to RANS modelling uncertainties. It is found that the insufficient near-wall boundary layer mixing is partly due to the failure of both SA and SST models to reasonably model the non-equilibrium turbulence behaviors inside the endwall boundary layer, which is caused by the boundary layer skewness. Based on the understanding of the skew-induced turbulence characteristics and its effect on mixing, a detailed effort is presented towards the physical-based modelling of the skew-induced non-equilibrium wall-bounded turbulence. The source terms in the SA and SST models that control mixing are identified and modified, in order to enhance mixing and strengthen the endwall boundary layer. The improved turbulence models are then validated against the compressor corner separation flows under various operating conditions to prove that the location and extent of the corner separation are more realistically predicted.

Liquid microjet impingement cooling has shown the potential to be the solution for heat removal from electronic devices such as very-large-scale integration (VLSI) chips. The post-impingement dynamics of the jet, specifically the interaction between the liquid fronts on the surface engendered by the jets is a critical criterion improving the heat transfer characteristics. While some seminally important experimental studies have investigated this attribute, the amount of accurate data and analysis is limited by the shortcomings of real-life experiments. In this article, numerical investigations into the fluid dynamics and heat transfer in microjet cooling systems are carried out. Specifically, this paper addresses the question regarding the necessary fidelity of the simulations. Different Reynolds-averaged Navier–Stokes (RANS) models are compared to the Large Eddy Simulations (LES) simulation and the potential fidelity of different eddy-viscosity-based closures is clearly shown. Recommendations are made regarding the RANS closures that should give the best performance. It is demonstrated that the transition Shear Stress Transport (SST) model and k - ω SST model both show excellent ability to predict the local or average Nu, and also local level pressure coefficient f with less than 5% difference in the range of 30 < Red < 4000, compared with the reference LES model. For the experimental measurements in the range of 130 < Red < 1400, the LES model, transition SST model and k - ω SST model all show less than 25% prediction error. Moreover, it is shown that the validity of the unit cell assumption for the temperature and flow distribution depends on the flow rate.

Turbulent flow and heat transfer past a cube (L × L × L) mounted on one wall of a channel of height 2L have served as a benchmark problem to validate computational methods and turbulence models because of the richness in the flow features induced by an unassumingly simple geometry. These include an unstable bi-modal horse-shoe vertical structure upstream of the cube, vortex shedding from the two sides of the cube that extend to the aft of the cube, and periodic separation from the top of the cube. In this study, simulations based on steady and unsteady Reynolds-Averaged Navier-Stokes (RANS and URANS) and large-eddy simulations (LES) were performed to understand the nature of the flow and how that flow affects pressure, drag, and surface heat transfer distributions. For RANS and URANS, the following turbulence models were assessed: the realizable k−ε model, the shear-stress transport (SST) model with and without corrections for curvature, the Reynolds stress model with linear pressure strain (RSM-LPS), and the stress-omega Reynolds stress model. In this study, the RANS and URANS results were generated by using the ANYSYS Fluent code, and the LES results were generated by using the PowerFlow code. Results obtained show RANS to be unacceptable in predicting the reattachment of the separation bubble downstream of the cube. URANS by being able to resolve the oscillatory shedding from the two sides of the cube was able to predict the reattachment with reasonable accuracy. URANS with the SST model that has the curvature correction gave the best results for the reattachment length downstream of the cube (relative error within 5% of the experimental data). Also, this model was able to capture all details of the flow features revealed in the experimental oil-flow study with considerable accuracy. However even with this model, the predicted pressure just downstream of the cube is incorrect by up to 50%. This is because URANS was unable to predict the unstable bi-modal, horse-shoe vortex system upstream of the cube induced by interactions between the spanwise vortices in the turbulent boundary layer and the horsehoe vortex system. This interaction was found to cause the separation bubble on top of the cube to shed periodically. By being able to capture this interaction, LES was able to better predict the pressure just downstream of the cube. This shows why LES is needed and why RANS and URANS are inadequate unless those models can be modified to account for these unsteady interactions. Steady and unsteady RANS based on the SST model were also employed to study the heat transfer about the surface of the cube. Reasonably good results were obtained for the leading and trailing faces of the cube but not for the top and the two side faces. A parametric study on the effects of the turbulent Prandtl number show the connection between the turbulent eddy viscosity and the turbulent thermal conductivity cannot be represented by a simple algebraic equation, the Prandtl number. Thus, better RANS models are needed to capture the physics embedded in the correlations between the fluctuating velocities and temperatures. A paper on this work is being prepared. The order of authorship will be X. Chi, K, Adeel Ahmad, Surya Muthukannan Chinnamani, and T. I-P. Shih. The author of this abstract thanks his students who worked on the different parts of this problem.

This study investigates the accuracy of Computational Fluid Dynamics (CFD) models to predict heat transfer in turbulent separated flows at low Reynolds numbers. A novel improvement of a Scale Adaptive technique is also presented. A spectrum of turbulence models is used to simulate flow and heat transfer of two geometries; fully developed flow through a staggered tube bank and a square prism in cross flow. Experimental data for both local heat transfer and velocity data are available in the literature for these cases and have been used extensively evaluate various CFD methods. Six unsteady models were used and the results show that the unsteady Shear Stress Transport (SST) model provided good overall accuracy relative to the mean Nusselt number for both cases. However, the SST model failed to accurately predict local variations. The Partially Averaged Navier-Stokes variant of the SST model showed a marked improvement for both cases. The Dynamic Smagorinsky Large Eddy Simulation (LES) showed a much-improved fidelity to the local Nusselt but under predicted the actual values. The computational cost for the LES model was significant. In general, it was found that the computationally expensive models with higher degrees of resolved turbulence did not necessarily return more accurate results. A Scale Adaptive Simulation (SAS) modification of the SST model (SAS-SST) is also used in this study. The SAS approach for the SST model adjusts the production term of the specific dissipation transport equation based on the second velocity derivative. This modification is intended to improve the SST model where local flow accelerations/ xx decelerations are detected, as occurs in separated flows. However, the local Nusselt number for the two cases considered were found to be generally less accurate than the baseline SST model. In this study a novel modification to this model was made to reduce the SAS contribution near stagnation points in the simulation. This was done through the Kato-Launder and production limiter modification in the SAS production term. The results showed only a slight improvement of the accuracy of the Nusselt number predictions. It is possible that further adjustments to the SAS terms and constants can be made to properly support and complement the stagnation point modification in this study and yield an overall better turbulence model.

This article is focused on the choice of a suitable turbulence model for simulations of an industrial pump's intake, from the perspective of accuracy and, partially, also the CPU time. Twelve steady-state and transient simulations were made on a fine computational mesh, using turbulence models such as: the shear stress transport (SST), the scale-adaptive simulation (SAS), the Reynolds stress model, the explicit algebraic Reynolds-stress model, the detached eddy simulation and the large eddy simulation (LES). The curvature-correction (CC) option was assessed for the SST and SAS turbulence models. The results were compared with the LES and with published experimental results. Although all the models could predict the main floor vortex, there were still some substantial differences. It can be able to conclude that it is better to use either the SST-CC turbulence model, due to its low-computational resources and far better results than the SST model, or the SAS-CC turbulence model, since its predictions are quite similar to the LES results. In the final step, good agreement with experimental results was shown for a longer simulation with the SAS-CC turbulence model.

A gas-kinetic scheme coupled with SST model for turbulent flows

Flow separation from a continuous surface under adverse pressure gradient remains one of the most challenging problems for turbulence modeling within the Reynolds-Averaged Navier–Stokes (RANS) framework. Although the Shear Stress Transport (SST) model performs well for a wide range of flows, its predictive capability deteriorates in wake regions dominated by complex turbulent structures. To enhance its accuracy, a modified SST model incorporating the Renormalization Group (RNG) theory of turbulence is proposed. The RNG correction is introduced as an additional source term in the dissipation equation and is limited to the reattachment region through a blending function to ensure computational efficiency and numerical stability. The model is validated using several canonical problems, including the curved backward-facing step, vertical and inclined backward-facing steps, and the periodic hills case at various Reynolds numbers. The effects of the inflow boundary condition are carefully examined to isolate the intrinsic performance of the modified model. For this purpose, an experimental campaign is performed. Comparisons with benchmark Large Eddy Simulation studies and experimental data show that the RNG-enhanced SST model substantially improves the prediction of mean velocity, turbulence kinetic energy, and wall quantities such as pressure coefficient, wall shear stress, and boundary-layer parameters. In particular, the prediction of the reattachment position is improved, yielding reduced relative errors that are nearly half of the SST model. These findings demonstrate that integrating RNG theory provides a computationally cost-effective, numerically robust, and physically consistent improvement to the SST model for separated flows from a continuous surface.

Conjugate heat transfer (CHT) simulations were conducted for five film-cooled flat plates designed to model the pressure side of the High Impact Technologies Research Turbine First Vane (HIT RT1V). The numerical results of the CHT analysis were compared against experimental data. The five test cases consist of one baseline geometry and four different cooling hole geometries applied to a film-cooling hole arrangement that was optimized to achieve a more uniform cooling effectiveness. This optimized film-cooling hole configuration was designed by coupling a genetic algorithm with a Navier-Stokes fluid solver, using source terms to model film holes, starting from a baseline cooling configuration. All five plates were manufactured, and surface temperature measurements were taken using infrared thermography while the plates were exposed to flow conditions similar to the pressure side of the HIT RT1V. CHT simulations were carried out using unstructured meshes for both fluid and solid with all film holes fully resolved. Comparison of experimental data and simulations shows a consistent trend between the optimized configurations as well as correct predictions of the flow characteristics of each hole geometry although the absolute temperatures are underpredicted by the CHT. Both experimental measurements and CHT predictions show the optimized geometry with mini-trenched-shaped holes to give the best cooling effectiveness.

Laminar-to-turbulent transition phenomena are ubiquitous in natural and industrial flows. As to the Reynolds-averaged Navier–Stokes (RANS) simulation method, the workhorse for accurate prediction of such flow regime boils down to the consideration of the transition effect in turbulence modeling. In this paper, an industrial–practical transition–turbulence model with excellent accuracy, robustness, and efficiency is established by the fully connected artificial neural network (ANN), which maps the relation between the RANS mean flow variables and an intermittency factor. A one-equation local correlation-based transition model coupled with Menter's shear stress transport (SST) model is taken as the benchmark. The present two-way coupling ANN model is trained with two National Advisory Committee for Aeronautics (NACA) airfoils, that is, NACA0012 and NACA2418, at various angles of attack and Mach numbers, while tested with the A-airfoil, NACA0015, and RAE 2822 supercritical airfoils in different flow states. The a posteriori test results manifest that the mean pressure coefficient, skin friction coefficient, size of laminar separation bubble, mean streamwise velocity, Reynolds shear stress, and lift/drag/moment coefficient predicted by the ANN model are all in good agreement with those given by the benchmark transition-based SST model. Furthermore, the ANN model exhibits higher calculation efficiency and convergence speed than the traditional transition-predictive SST model. The present work may pave a new way for machine learning methods to be used in integrated transition–turbulence modeling toward industrial applications.

This paper presents results from numerical simulations of a 3-bladed horizontal axis tidal stream turbine (TST). A two-equation transient Reynolds Averaged Navier Stokes (RANS) k-! Shear Stress Transport (SST)model was used for code validation and testing of a newly implemented sliding mesh technique for Code Saturne. Wall- and blade-resolved large-eddy simulations (LES) were then performed to study the complete geometry at various tip speed ratios (TSR). Thrust and power coefficients were compared to experimental measurements in a towing tank for a range of TSR (4, 5, 6, 7 & 8) at a fixed hub pitch angle. A strong flow meandering was observed with the wake recovery and velocity deficit showing a high sensitivity to upstream turbulence intensities. The mean thrust and power coefficients by RANS k-! SST model were found to be more sensitive to the upstream turbulence than the LES. A comparison of all models are also presented for the mean sectional blade pressures and mean wake velocity profiles. The paper also presents an overview of modelling and numerical issues relating to fine LES for such rotating geometries

Rotary kilns are large, cylindrical, rotating ovens with a burner in one end that are used in various industrial processes to heat up materials to high temperatures. The kiln burners are characterized by long diffusion flames where the combustion process is largely controlled by the turbulent diffusion mixing between the burner fuel jet and the surrounding combustion air. The combustion air flow patterns have a significant effect on the mixing and hence the combustion efficiency, motivating a systematic study of the kiln aerodynamics. The objective of this work is to compare turbulence models when modeling the kiln aerodynamics of an iron ore pelletizing rotary kiln. Simulations of the non-reacting isothermal flow using three different ω-based turbulence models are performed on a simplified, down-scaled model of the kiln. Some of the results are validated against particle image velocimetry (PIV) experiments. The turbulence models used are the two-equation shear stress transport (SST) model, the Reynolds stress baseline (RSM-BSL) model and the delayed detached eddy simulation (DDES) turbulence model based on the SST formulation. It is found that the turbulence models produce quite different results yielding various predictions of the flow field. The SST model fails to capture the unsteady behavior of the flow field and the DDES model performs poorly on the grid applied. The Reynolds stress model agrees best when compared with the experimental data and provides a good trade-off between details captured and computational effort.

This study presents a numerical investigation of supersonic flows through a convergent conical rocket engine nozzle, focusing on the influence of turbulence models on shock structures and nozzle performance. Simulations were performed in ANSYS Fluent for nozzle pressure ratios (NPRs) from 1.4 to 7.0 under steady-state conditions. Four turbulence models [Realizable k–ε, Standard k–ω, Shear Stress Transport (SST) transition model, and Reynolds Stress Model (RSM)] were evaluated for their ability to predict Mach distributions, shock-cell structures, and pressure-based performance metrics. Results indicated that the SST k–ω model provided sharper, more accurate shock-cell structures and captured shock–boundary-layer interactions more effectively, especially at higher NPRs. At NPR = 5.0, the SST model predicteda maximum exit Mach number of 1.85, closely matching the experimental 1.88 ± 0.03, whereas the Realizable k–ε model underestimated it at 1.75. The k–ε model offered greater numerical stability and lower computational cost but tended to oversimplify flow features and underpredict shock intensity. RSM captured anisotropic turbulence effects but required more computational resources and exhibited higher sensitivity. Aerodynamic performance evaluation showed that the SST model predicted thrust and discharge coefficients (CV, CD) more accurately across all NPRs. At NPR = 4.0, it produced a CV of 0.92 and a CD within 3% of experimental values, compared to 0.87 and up to 7% deviation for the k–ε model. Therefore, the SST model was recommended for high-fidelity analyses requiring accurate prediction of shock behavior and nozzle performance, while the realizable k–ε model was suitable for preliminary analyses prioritizing faster, stable convergence with reduced detail.

Cooling of modern gas turbine blades and vanes is crucial, as they are exposed to extremely high temperatures. The present investigation makes a three-dimensional conjugate heat transfer analysis of internal convection of a ribbed gas turbine vane–blade system under rotational conditions. Simulations have been performed for a single turbine stage, including stationary vanes and rotating blades. The performance of rib-roughened coolant channels is compared to that of smooth channels for different conditions. Flow turbulence is resolved using the Shear Stress Transport (SST k–ω) model with automatic wall function. The compressibility effects in the external hot-gas flow and the rotational effect-induced complex flow patterns inside the rib-roughened coolant channels have been analyzed. The Coriolis force, caused by the rotation of blades, affects the internal flow structures within the smooth and ribbed coolant channels. Ribbed channels significantly enhance local heat transfer by disrupting Coriolis-induced vortex structures and promoting secondary flow. The flow and heat transfer characteristics within coolant channels vary at different blade positions. A transition from supersonic flow at the leading edge to the transonic flow at the trailing edge of the suction side of rotor blades is observed. Compared to the smooth channels, the ribbed channels achieve a blade temperature reduction of up to 191.8 K and a 49.9% increase in the Nusselt number, with a moderate pressure penalty. The study highlights the critical role of rib geometry and rotational effects in optimizing the internal cooling performance of turbine blades for high-temperature applications.

Conjugate heat transfer investigation of cooled turbine using the preconditioned density-based algorithm

Direct simulation of conjugate heat transfer of jet in channel crossflow