I have a geometry where the flow is transitional. i.e. a part of it has laminar flow and rest part has turbulent flow.
Can I use k-epsilon model for this problem? Does it resolve the laminar region well?
Dear, Hardik M.
Certainly, you can use k-epsilon model for transit flow.
No, RANS formulation does not generally recover the solution in a laminar flow region. You need of specifical formulations
No. Certainly the standard k-epsilon model is not able to predict transitional flow regime. You need a special formulation of a transition model. E.g. ANSYS CFX and ANSYS Fluent are providing several transition model formulations. And in addition to using the right model, you need to prepare a suitable mesh with very good near wall mesh-resolution. As far as I know, most of the time a mesh with a near-wall y+
The transit regime flow is near to the turbulent regime, thus you can use k-epsilon model.
Saad Najeeb : Sorry, but I disagree or we speak about two different things. I speak about prediction of complete laminar to turbulent transition, i.e. a flow which goes through all the intermediate flow regimes. K-Epsilon model is clearly not capable of predicting this.
I agree, without ad hoc adjustements in no way RANS formulation is suitable for transitional flows. What is more, it will not provide a laminar solution if you have a laminar flow problem.
Thanks Filippo Maria Denaro , Saad Najeeb and Thomas Frank
So, as you said k-epsilon model is incapable of resolving laminar and transitional regimes, what should be the strategy for solving such problems?
I am using Fluent and my geometry has a pipe like volume first where the flow is laminar and then suddenly expands in a large diameter chamber which also has some obstruction. Flow is turbulent here.
Thomas Frank mesh with y+ < 0.1..isnt it too fine?
Hi Harkik,
I understand your question re transition of one state of motion to a (more or less) other state of motion. This might be laminar-turbulent, or eben low-Re to fully developed turbulence. The latter case is a turbulent-turbulent transition and its laws of motion take into account the INERTIA of motion. I.e., although turbulence is GENERALLY kinda STEADY STATE (actually the german word FLIESSGLEICHGEWICHT, formed and used by Ludwig Bertalannffy, fits better as it indicates already verbally the FLUX of energy continuously flowing through the spectral energy cascade) I will use it her. So, if you go from one flow state to anothr you need to either accelerate or de-celerate. While accelerating flow is comparatively simple to describe, the opposite is REALLA hard and almost not understood. It means deceleration from high-Re flow to low flow or even to laminar state.
For a quite minute, start with the attached part 1, then part 2, and only then Baumert-Wessling who deal with the non-Newtonian case. For all further questions please ask my agent ...
Hardik Mistry : The strategy should be to apply an appropriate transition model - see attachment. The Fluent documentation will provide you with the details about the model formulations. And "no", a y+
Thanks Helmut Ziegfeld Baumert for the detailed explanation.
Thanks Thomas Frank for the help.
Good day!
In my opinion k-e model of turbulence was produced and recommended for jet flows. I think in more common case, the turbulent Reunold's number may help to filter turbulent, laminar, domains in flow pattern. Then molecular viscosity and turbulent viscosity are of same values in first approximation.
W.b.r,
Serge.
Transition flow is a different flow regime from turbulence. To predict a complete laminar to turbulent transition, there are some models to do so. For example, k-kI-Omega….. Choosing a good model also depends on the type of flow where some models work well for external wall-bounded flow lie boundary layer
- Badges
- Science topic