hi folks, just wonedring whether it is a wise idea to model the flow over a flat plate as turbulent from the very beginning. laminar solution does not agree with experiment, i suspect there is some degree of turbulence in the wind tunnel free stream flow. i don't really thing Reynolds Number is reliable enough to predict turbulence.there are way too many sources of pertubations to be accounted for by a simple comparison of inertial and viscous forces. will be grateful for some insight on this. maybe somebody has had this problem be4. how does prescribing a small turb intensity to a laminar flow affect the solution.

Hi,

Have you considered a transitional turbulence model? That is designed to do the sort of modelling you are talking about so is likely to be useful.

Note that you will need to know the free-stream turbulence intensity AND dissipation. If you can't match the experimental conditions exactly then you are likely to have problems with accuracy.

You cannot specify a turbulence intensity (as a k value) in a laminar flow, as k has no meaning in laminar flow. You are talking about LES/DNS simulations and I suspect you don't want to go there. Use the transitional turbulence model instead.

Glenn Horrocks

hi Glenn, thanks for your suggestion. i have thought about that. it seems the problem with the transition models is this requirement to prescribe the free turb. intensity and the dissipation. i guess once the turbulence intensity value is known, using an integral length scale the dissipation can be approximated. problem is, in most cases the transitional point has to be prescribed as well, at least this is what i get from literature on current transition modelling. so it is as if we force the numerical approach to confirm experimental results by prescribing all these details. Glenn, suppose the transition to turbulence does not occur along the plate, could the use of a transitional model still provide a solution different from the laminar case? in my model i expect transition to set in much further down the plate than the length currently modelled. will be grateful for some ideas. if possible, some tips on the free stream turbulence values (guesses)as well. thanks ben akih

Hi,

1) You have to prescribe free stream turbulence intensity and length scale/dissipation for transitional or fully turbulent simulations. Isotropic turbulence is generally specified by two variables, the first is almost always turbulence intensity, the second is either a length scale, time scale or dissipation rate. It is easy to convert between these parameters so any one of them is enough, however to fully define the free stream turbulence you need to fix two turbulence parameters.

2) If you approximate the experimental conditions you are likely to get just approximate simulations! Having said that, some simulations are fairly insensitive to free stream conditions - but turbulence transition point usually is very sensitive to both free stream turbulence intensity AND length scale/time scale/dissipation (any one is enough to define the turbulence).

3) You can prescribe the transition point in the CFX transition point if you like.

4) The transitional model will give laminar flow if no transition occurs.

Glenn Horrocks

hi sir, i want a source code which is written in c language for a turbulent flows over a flat plate.