Hello, I modelled a cilinder in a wind tunnel with Spalart Allmaras turbulence model because my Reynolds was around 2E5 (It should be turbulent) but when I compared my results to the experimental data, it showed me that I should of used the laminar model (looking at the pressure distribution over the surface)

After that,I used the laminar model and got good results but how can I explain the fluid separation in a laminar model ¿Is fluent capable of doing this or was it just luck that I got good results?

PD: I always thought that the laminar model is for flows that remain attached to the surface.

Hi

Fluent made a validation case of flow around bodies in laminar flow.

You can find it under the name "Steady and unsteady laminar flow behind a circular cylinder"

Hopefully, it help you to visualize more your conclusions!

However, I thihk you should use another models such as k-e model or LES. There are some papers that talk about simmulation of flow around solids. In there the authors claim that they obtain accurate results.

Regards

Alex Munoz

Hello I think it is quite strange with a such Reynolds. try to use k-e RNG model, I think it will be a good model for your case. It will discribe in a good way the boundary layer. but I remain think that Spalart Allmaras is a good model for the interaction between fluid and structure. May be your grid is not suitable around the cylinder.

I hope you will succeed.

until know I study such a case with laminar model to simulate stroual instabilities. you can take a look on the following web site: www.enseeiht.fr/travaux you will find student works, it is in french but everything concerning CFD is in english.

Mehdi icho

Hi, I have one question about your model: are you simulating the flow around cylinder (so the flow is normal to the cylinder axis) or along the cylinder (flow is parallel to the axis)? If it is along the cylinder, which length scale are you using?

I am trying to simulate the flow along the cylinder (the ratio of radius to length is very small), actually cylinder is moving in a free space and gas is rest. If I use radius as length scale, Re is very small (less than 100), but if cylinder length is used, Re is large (about 1e5). I first used laminar model, the heat transfer between cylinder and gas is smaller than experimental data; then I used Spalart Allmaras model, the agreement is better. But I am fully convinced the length scale used for Re. Does anyone have any insight on this?

Also, when I used turbulent model, I find the far boundary is farther than that in laminar flow. anyone have any idea about this too?

Thanks in advance for any help

The flow I'm studying is just a common cilinder with the flow normal to the cil. axis so the diameter is the char. length. My Reynolds is about 2E5.

I have never seen a case like yours and I am sorry but I cant help you.