
[Sponsors] 
July 8, 2001, 05:40 
One/Twoequturbulencemodel

#1 
Guest
Posts: n/a

Dear all,
I am trying to implement the kepsmodel (LamBremhorst) into my 2D  FD  code. (it is a basic code, not very sophisticated, but everybody starts with small steps !!). My Question is: Knows somebody an Oneequmodel, where I can at first test the kequ. before I implement the epsequ. The geometry I am using is a canal flow and the backwardfacing step. Further questions concerning the keqsequ (LamBremhorstmodel):  do I have to use the walldamping function only in the near of the wall or also at the inlet or outlet.  I am using noslip conditions at the wall, should I implement other bcd at the wall, eg. for consideration of the loglaw layer. Thaks for the help Matthias 

July 8, 2001, 12:24 
Re: One/Twoequturbulencemodel

#2 
Guest
Posts: n/a

(1). My suggestion is: try the fullydeveloped channel flow case first, which is just 1D problem. In this case you are not going to have inlet or exit. Just the wall point and the centerline point. (2). In the fullydeveloped channel flow case, you can also try out two turbulence models: one is the high Reynolds number model and the other is the Low Reynolds number model. (3). These two models require different sets of equations and boundary conditions. For the high Reynolds number turbulence model, normally some kind of wall function treatment is required, and you have to be very careful in the implementation of the method because most published methods did not handle it correctly.


July 8, 2001, 15:15 
Re: One/Twoequturbulencemodel

#3 
Guest
Posts: n/a

Matthias,
There is a simple trick for implementing a oneequation model for the kind of testing you want to do. Note that the keps model assumes the turbulent length scale may be estimated by L = D k^1.5 / eps, where D is a constant usually assumed to be unity. All you need to do is implement the k equation, use D/L as an input number to the calculation, and use this relation to compute eps instead of getting it from the eps transport equation. 

July 9, 2001, 03:05 
Re: One/Twoequturbulencemodel

#4 
Guest
Posts: n/a

Dear John,
thanks for the suggestion but I have already an 2D code, so I think it is to much effort to first try the implementation for the 1D case. For my problem I am mainly interested in lowReflow (Re <=50000). As I deduce from you post  there I don 't need an special wall treatment (wall fct.). Matthias 

July 9, 2001, 03:07 
Re: One/Twoequturbulencemodel

#5 
Guest
Posts: n/a

Dear Dean,
I have done this test (found in Lecture notes on turbulence, Launder, 1972 ?) but I thougth there is a more sofisticated model for my lowReproblem. thanks again Matthias 

July 9, 2001, 03:22 
Re: One/Twoequturbulencemodel

#6 
Guest
Posts: n/a

(1). You can easily run 1D problem using a 2D code, if you update inlet and exit solutions by using the computed interior station. (that is, you can use 3 rows in the axial direction) (2). No, you don't need to use the wall function treatment, and a low Reynolds number model is required.


July 11, 2001, 03:39 
Re: One/Twoequturbulencemodel

#7 
Guest
Posts: n/a

(1). I think, in 80's Dr. W. Rodi had published a review article in AIAA Journal about various low Reynolds number twoequation turbulence models. I think, the model you mentioned is an old and better one. (2). Since then, there has been a lot of development and new related models published, with some from Japan. (3). There was also a near wall turbulence modeling conference held at Arizona State University in 90's (?) (4). So, if you do some paper search, you should be able to find many different new models, hopefully with improved results.


July 11, 2001, 10:41 
Re: One/Twoequturbulencemodel

#8 
Guest
Posts: n/a

Older one, but models presented here are still in use in Fluent and co :
Turbulence Model for NearWall and Low Reynolds number Flows : A Review Patel, Rodi and Scheuerer AIAA Journal vol 23 no 9  1984 and in 1993 : Low Reynolds number kE modelling with the aid of direct simulation data Rodi and Mansour JFM vol 250 pp 509529  1993 And you may take a look at jobs done by Durbin and co, the keyword should be elliptic relaxation. 

July 11, 2001, 11:23 
Re: One/Twoequturbulencemodel

#9 
Guest
Posts: n/a

Thanks for help
Matthias 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
SimpleFoam case with SpalartAllmaras turbulence model implemented  nedved  OpenFOAM Running, Solving & CFD  2  November 30, 2014 23:43 
Wrong calculation of nut in the kOmegaSST turbulence model  FelixL  OpenFOAM Bugs  27  March 27, 2012 09:02 
Low Reynolds kepsilon model  YJZ  ANSYS  1  August 20, 2010 13:57 
KOmega Turbulence model from wwwopenFOAMWikinet  philippose  OpenFOAM Running, Solving & CFD  30  August 4, 2010 10:26 
SSG Reynolds Turbulence Model  Georges  CFX  1  February 28, 2007 17:15 