I did in the past several numerical simulation using the k-ε standard model with reasonable results. However, The k-ω model was claimed to perform better in adverse pressure gradient and to reproduce better flow separation than the k-ε model. You could even try the SST model, which is aimed at using the favourable properties of both models.

I did in the past several numerical simulation using the k-ε standard model with reasonable results. However, The k-ω model was claimed to perform better in adverse pressure gradient and to reproduce better flow separation than the k-ε model. You could even try the SST model, which is aimed at using the favourable properties of both models.

Please, note that the turbulent length scale should be related to the radius/width of your channel.

Yes I am using the formula "turbulent length scale"= 0.07 x hydraulic diameter.  

Any suggestion abt the outlet Boundary Condition. The flow is pump driven having a variable speed drive and close recirculating water tunnel.

you can use SST. The model combines the k-omega turbulence model and K-epsilon turbulence model such that the k-omega is used in the inner region of the boundary layer and switches to the k-epsilon in the free shear flow.

I recommend using the turbulent model of modified gama-Re. This model exists in the StarCCM+ software.

Two-equation models are good enough for this simulation, and the k-ε model has been applied to this type of physics with quite reasonable results. However, you may need to have the step at a sufficient distance from the inlet boundary to make sure that the turbulence is fully developed at the step. The other important issue (somehow more crucial than the turbulence model itself) is the hydrodynamic model which should be accurate enough for simulating the pressure variation over a step.

The SST model would be suitable for this problem as it involves adverse pressure gradient and separation. The realizable k-epsilon model may also work. 

One thing to make sure with these turbulence models is the near wall mesh. Check the references and documentation regarding sensitivity of these models to different y+ values. 

eg. in Fluent, with y+ ~ 1 for a plane channel flow, the enhanced wall treatment option in k-epsilon gives similar results to SST model for Re~590 compared with DNS results from Moser et al, 1999 ( http://dx.doi.org/10.1063/1.869966). 

Dear Hassan

standard K Epsilon and standard K Omega models could help you to get good results. although Reynolds stress model (RSM) is a perfect one but it would increase your solution time. I attached you an ANSYS tutorial for simulating turbulence flow over an step. I hope it would help you.

Good Luck

I would suggest k-epsilon model for said turbulence flow.

two equations models are nice for this simulation, for this only thing is that you should put sufficient distance from this inlet boundry to make sure that turbulence is fully developed at the step

Hi, I also suggest the standard k-epsilon model and it can simulates the wall turbulent problems well. 

k-w model

before giving you any idea, what about the Re number? What about the wall effects in the experiment?

My suggestions are: 

- use the SST model since from both litterature and personal experiences it performs better

- If the reattachment point is too far from model test try to adjust the turbulence intensity instead of mixing length, however first of all you should be shure that your solution is mesh indipendent if not it will be useless.

best regards

Again, if the experimental setting causes relevant 3D effects, the problem is not the best turbulence model but the bc.s settings int the numerical model.... many years ago we worked about that, in order to match experimental results

https://www.researchgate.net/publication/230392664_Analysis_of_3-D_backward-facing_step_incompressible_flows_via_local_average-based_numerical_procedure

Article Analysis of 3-D backward-facing step incompressible flows vi...

Hi, I am working on Forward facing step in Adverse pressure gradients. Reynold number : 4800. Step height : 6mm,  inlet height = 45mm. Smooth wall  with no surface roughness was used in experiments.

You may use any relevant model but make sure that it captures all the eddies formed on the step, since there is singularity type of flow and some turbulence model unable to predict it. 

The reattachment length (Xr) in the recirculation region was determined where the dividing streamline ψ = 0 . i.e., from leading step edge to the point where the very first streamline attaches to the top of the step at some distance Xr

The Re number is low such that this flow case was studied in DNS more than 20 years ago, e.g. see

http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19940028784.pdf

Dear Hassan,

I am agree with Mohsen J., however, to capture all eddies I suggest LES model. If you are in trouble with computational cost, you might use a hybrid model such as SST-SAS. The RANS models are weak as compared LES and SST-SAS hybrid model.

Please refer to this article (page 132):

The Scale-Adaptive Simulation Method for Unsteady Turbulent Flow Predictions. Part 1: Theory and Model Description By F. R. Menter · Y. Egorov, 2010.

For your experiments, you can use 0 pressure gauge as outlet boundary condition if the length of test section is long enough. With a long test section, It doesn't matter the water channel is close or open.

Dear Hassan

SST k Omega I think its good model to predict your case and you have to calculate boundary layer this you give you more accurate results  

1/ LES (SGS) models may be, in general,  the more useful and that I know in many case that K-w , k-eps and their hybrid models < LES < DNS models: '

however, please provide some details (figures) about the discrepancy between experimental and numerical results.... the reattachment lenght is ony one of the flow parameters of interest

All three models k-epsilon, k-omega and SST model of Menter

can be used for a range of flow problems with good accuracy as discussed in the following references

[1] Wilcox, D. C : Turbulence Modeling for CFD. DCW Industries, Inc., 5354 Palm Drive,

La Canada, Calif., 1998.

[2] Cebeci, T.: Turbulence Models and Their Application. Horizons Pub., Long Beach,

Calif, and Springer, Heidelberg, 2003.

[3] Cebeci, T.: Analysis of Turbulent Flows. Elsevier, London, 2003.

Dear Hassan

for the turbulence flows, i think that RSM model is the best. you can also used other model such us SST k-w..

Hi Dr Filippo, please have a look at attached file.

APG-0 , APG-2, APG-4 stands for adverse pressure gradient with 0 degree, 2degree and 4 degree diverging angles.

Thank you all for adding comments.

Hello,

let me better see the problem:

1) flow is from right to left?

2) h/H is quite low?

3) APG-0 would be the standard backward facing step in a straight channel?

What let me some doubts is the results for APG-0 that showed large discrepancies between the simulations.

What about your computational grid? how many nodes describes the step ?

Hi,

1) Flow is from left to right. It is a forward facing step in APG.

2) blocakge ratio : h/H is 11.76 % for 6 mm step height.

3) APG-0 would be a FFS.

Computational grid:  0.3 mm tet mesh starting from -2h of the step to 4h and a height of 2.5 h. 120 nodes approximately

I am using enhanced wall treatment y+=1

Well, that means you have a flow impacting frontally with the step, not a classical backward facing step....I strongly suggest to use structured grid instead of tet elements.

@ Hassan Iftekhar

Dear friend

Please find the attachment. Hope it helps you.

With best wishes

Reza

On the principle of independence of the boundary conditions of structure direction of turbulence

Lu Panming

Abstract

This is an introduction to the turbulence modeling theory of the present author, which can take the boundary conditions of turbulence structure into consideration ( the main contents have either not been published in normal journals or only published in chinese papers).   It is included that： (I)。Difficulty Problems Encountered in “Second-Moment-Closure” turbulence modelling ，i.e.  “Gao-Ge Anomaly” ;  (II)。The improvement  to the ”Second-Moment-Closure” turbulence modelling in order to allows the boundary conditions of turbulence structure direction could be prescribed;  (III)。An explanation to the principle of  independence  of  the boundary conditions of

turbulence structure direction ;  (IV)。Six suggestions to the possible future work directions;  (V)。 An explanation to the ”Gao-Ge Anomaly”.

------------------

4. The principle of the independence of the boundary conditions is not only to the turbulence modelling, but also suitable to the other subject, because it is frequently meet the equations are not easy to solve, so there is often a need of modelling or predigestion.  For example, both the D‘Alembert anomaly (~1752) and the Stokes anomaly (~1856), are due to breach the  principle of independence  of the boundary conditions, according to today`s point of view .This is because, first, by using the inviscid potential flow`s dynamic equations it is not possible to affiliate the slip-less condition of the real fluid at the wall, and  second, using the Stokes equation it is not possible to affiliate the boundary conditions both at the wall of cylinder and at the infinite to the cylinder.  The first anomaly lead to the discovery of the Prandlt`s boundary layer theory（1904）, and the  second anomaly  was solved by Oseen（1910）with introducing a modification. All these reflect the scientific worthiness and broad latency of the present principle   of the independence of the boundary conditions.

-----------------

above is an example paragraph,for more details or need the full text, please go to the end of this letter, where will be an web address existed.

http://staff.ustc.edu.cn/~pmlu/ 

Hassan and Carlo,

I fully agree that at present k-omega is the only reasonable choice among the many semi-empirical choices, well yes, according to the attachment. Regarding the zero-pressure condition at the downstream boundary - it might be a problem in a recirculating flow IF the channel is short or, when circular, IF with a small radius of the flow. In these cases the pressure distribution would matter because pressure at the inflow boundary has to compensate the ouflow pressure AND the inner turbulent friction AND the momentum redirection against this friction, against mass inertia, and against the pressure "felt" by the step upwards-facing step.

I did some work related with this topic some years ago. I used an experimental database for a backward facing step from Driver&Seegmiller (1985). It is a bit older database but quite useful for checking the behavior of turbunece models.These authors have measured mean velocities and its fluctuations, and there is data for the the reattachment length too. I carried out some work related to this topic when I developed my PhD project. You can see this reference (and others), and some turbulence model comparisons in my PhD work. I have uploaded my PhD thesis in Research Gate, and then, I think that you can download it. The name of the work is Theoretical - numerical study of flows with strong streamlines curvature, PhD Thesis ETSEIB, UPC Barcelona Spain.