Turbulence models have long been developed and examined for their accuracy and stability in variety of environments. While many flows work with excited turbulence intensity, models have rarely been tested to explore whether their accuracy withstands with augmented free stream turbulence intensity or decline in reasonable solutions. In present study the turbulent intensity of the air, moving parallel to a flat plate is increased from 0.4 to 6.6% for the whole flow, downstream to the screen. Three popular turbulence models are examined by investigating the turbulence penetration into flow field as well as into turbulent boundary layers over the flat plate. Results of numerical solutions for Standard k-e, Realizable k-e and finally two equations Shear Stress Transport k-w model are compared to experimental measurements and results are discussed. Results of variation of free stream turbulence intensity from flow field out of boundary layer, in addition, streamwise mean velocity, streamwise rms velocity and skin friction coefficient from boundary layer are investigated. Conclusion is made that despite restrictions of these turbulent models specially in predicting flow near a turbulent/non-turbulent interface, they have acceptable performance in both low and high intensity turbulent flows.

There are many approaches to determine the sound propagated from turbulent flows. In hybrid methods, the turbulent noise source field is computed or modeled separately from the far-field calculations. To have an initial and quick estimation of the sound propagation, less computationally intensive methods can be developed using stochastic models of the turbulent fluctuations. In this paper, turbulent mean flow of a two dimensional, compressible, cold-jet at Mach 0.56 is computed using RANS with 2 equation k-e RNG model. The above mean-flow quantities are then used in a stochastic model to generate the details of the turbulent velocity fluctuations. This method is based on the use of classical Langevin equation to model the details of fluctuating flow field superimposed on the averaged computed quantities. The resulting sound field due to the generated unsteady flow is then evaluated using Lighthill's acoustic analogy. Our results are validated by comparing the directivity and the overall sound pressure level (OASPL) magnitudes with the available experimental data. Numerical results show reasonable agreement with the experiments, both in maximum directivity and the magnitude of the OASPL.

In this paper, linear accelerated turbulent pipe flow has been simulated at various Reynolds numbers using five common turbulence models. The models considered are the Baldwin-Lomax algebraic model, the Spalart-Allmaras one-equation model, theκ-e model with wall correction of Lam and Bremhorst, theκ-ω model and the κ-ε-ν2 model. The goal is to evaluate the performance and precision of these models for prediction of the wall shear stress, Reynolds stress, turbulence viscosity, delay time in response and mean velocity. Factors such as changes in pipe diameter, fluid type, initial Reynolds number of acceleration and rate of acceleration and its effect on the above parameters has examined carefully. In order to verify the results, the experimental and numerical results (turbulence modeling and Large Eddy Simulation) of other researchers have been compared with the present results. The results show the desired accuracy of the one-dimensional modeling of accelerated turbulent pipe flow in comparison with Large Eddy Simulation results (three-dimensional). The response of delay time, simulated by the models (except BL model) shows relatively good agreement with experimental data. Comparing the distribution of mean velocity, turbulent kinetic energy and turbulent viscosity shows  k-e-ν2 model leads to a better accuracy compared with the other models.

Distillation Column is one of the eldest and important separation equipments which many new designs and various changes in its internals are still being made for increasing its performance. In this research with utilizing computational fluid dynamics, sieve tray hydraulics in experimental scale has been simulated in two phase and three dimensions in Eulerian frameworks by FLUENT 6.3 commercial software. And the simulation have been validated with Krishna et al. (1999) experimental and simulations results. Then simulating with RSM and k-e in three forms of Standard, Realizable and RNG turbulence models, the results have been discussed and compared and has been demonstrated that the RNG model predicts lower clear liquid height which are closer to the experimental results.

Recognizing of river flow pattern, deposition and erosion areas in meanders is highly important. Despite the limitations of the physical models in laboratory investigation of flow pattern in meanders, mathematical models can be helpful. In this research, RSM and LES turbulence models have been compared on sharp river bend of a laboratory flume using Fluent Software. The Laboratory flume is a sharp river bend established in hydraulic laboratory of EPFL. In order to investigate turbulence models, depth-averaged parameters were used. Moreover, to quantitatively examine those models, some longitudinal velocity profiles were selected and compared with measured ones. Results of longitudinal depth averaged velocities showed that RSM model is not able to precisely determine the important flow points. This model shows flow separation region within 85 degree and maximum width of 40 percent, though measurements indicated it was located in the region with maximum width of 60 percent of total width and approximately 75 degree. On the contrary, LES model suitably shows separation region. Besides, it better determines maximum measured depth averaged velocity situations rather than RSM model. Investigation of transversal depth-averaged velocity distinguished LES model from RSM as the better model and finally quantitative comparison of velocity profiles showed LES model is more accurate and presents more valid authentic results.

In the current study, after validating the accuracy of the current solution, mixed convection heat transfer in a square enclosure for drilling mudas a non-Newtonian fluidand wateras a Newtonian fluid—is investigated by finite volume method. The turbulence methods used in this study are the reliable methods such as RNG, standard, and RSM. The outcome of the investigation implies that under natural convection conditions, velocity boundary layer is somewhat asymmetrical on the cold wall and the fluid at the center of enclosure remains stratified and still. The existing graphs also indicate that the turbulence intensity is higher for forced convection than natural convection. Also, the maximum turbulence intensity is greater for drilling mud than water. One of the most prominent outcomes to be named is that, under similar circumstances, the Nusselt number for water is far more than the one for drilling mud, which confirms that convection heat transfer is greater in the water than the drilling mud.

Turbulence schemes have long been developed and examined for their accuracy and stability in a variety of environments. While many industrial flows are highly turbulent, models have rarely been tested to explore whether their accuracy withstands such augmented free-stream turbulence intensity or declines to an erroneous solution. In the present study, the turbulence intensity of an air flow stream, moving parallel to a flat plate is augmented by the means of locating a grid screen at a point at which Rex=2.5×105 and the effect on the flow and the near-wall boundary is studied. At this cross section, the turbulence intensity is augmented from 0.4% to 6.6% to flow downstream. Wind tunnel measurements provide reference bases to validate the numerical results for velocity fluctuations in the main stream and at the near-wall. Numerically, four of the most popular turbulence models are examined, namely the one-equation Spalart-Almaras, the two equation Standard k-e, the two equation Shear Stress Transport k-w and the anisotropy multi equation Reynolds Stress Models (RSM). The resulting solutions for the domain are compared to experimental measurements and then the results are discussed. The conclusion is made that, despite the accuracy that these turbulence models are believed to have, even for some difficult flow field, they fail to handle high intensity turbulence flows. Turbulence models provide a better approach in experiments when the turbulence intensity is at about 2% and/or when the Reynolds number is high.