Hi,
I am trying to simulate the steady flow on the wake of a circular cylinder to find the optimum Turbulence model for different Reynolds numbers. However, in most of the research papers I found they just mentioned the inlet velocity and the ratio of domain dimensions. Whenever, I apply small diameter (i.e. .151 m) of cylinder with inlet velocity 1 m/s and use the default density and viscosity, the drag and lift coefficient is much less than the other research papers for same Reynolds number. However, for larger diameter of the cylinder (i.e. 1 m) the drag and lift coefficient becomes more for a higher Reynolds number which is not accurate as Drag and Lift coefficient decrease with the increase of Reynolds number.
My question is which parameter should I change to vary the Reynolds number. Inlet velocity / Domain size including the diameter of cylinder / density / viscosity ? Should I run hybrid or standard initialization?
Thanks.
Theoretically, you can change velocity, lenght or fluid viscosity in the Reynolds number. No matter what you select, flow problems having the same Re number must show the same solution (I assume that Re is the only rlevant non-dimensional parameter of the problem).
Therefore, I suggest to provide more details about your setting in Fluent.
It is known that RANS can fail to produce correct solution for certain Re numbers. The grid size close to the wall is also a relevant parameter.
Theoretically, you can change velocity, lenght or fluid viscosity in the Reynolds number. No matter what you select, flow problems having the same Re number must show the same solution (I assume that Re is the only rlevant non-dimensional parameter of the problem).
Therefore, I suggest to provide more details about your setting in Fluent.
It is known that RANS can fail to produce correct solution for certain Re numbers. The grid size close to the wall is also a relevant parameter.
Hi Filippo,
Thanks for your suggestion. I did mesh independence study and found better results. For my case, I changed the inlet velocity to variate the Reynolds number and observed almost similar results (Drag and lift coefficients) with other literatures.
The variable that you use for changing the Reynolds number is immaterial. Any variable should do.
Also, initialization should not be a major problem. You can go for any of the two and the final results won't differ much if you've achieved proper convergence.
My guess is there's something wrong with your reference area selection. Drag or lift coefficients are associated with a reference area. Are you choosing the correct reference area for calculation? Do that first, keep the geometry fixed and use parameter changes in density, velocity or viscosity to change Reynolds number. If you change geometry, you've to change the reference area as well.
Hi Navin,
Yup I tried to monitor the drag and lift coefficients for cylinder only, not for the entire domain. Now I have got my desired results. The problem was the turbulence intensity and turbulence length, which should be the same as the article that I have chosen as a reference. However, thanks for your suggestion.
- Badges
- Science topic