Hello CFD community,
I am trying to simulate turbulent flow through a straight pipe and get a fully developed hydrodynamic profile
the details of my simulation are as follows
diameter of pipe: 5 mm
L/D ratio: 50 (to ensure fully developed flow)
Re: 10000
Fluid: Water
friction factor: 0.031 (by defining roughness height on wall as 0.0015 mm => copper pipe)
Model: SST k-w
Turbulent Intensity: 5.1% (from formula)
My mesh Yplus value: 0.89
so I am comparing my velocity profile to the power law formula (n=5.7)
Since this is a emperical simulation and nothing fancy in terms of vortexes or whatever, i am not able to get a model which closely matches the power law profile.
I would highly appreciate any insights on the same
Thank you
I would judge this a pretty perfect agreement. Why changing a turbulence model like SST, which is much more universal and most likely even more accurate in comparison to experimental data, just to be in agreement with a formula, which by itself is only a coarse representation of the time or ensemble averaged flow field in a pipe? I would very discourage to do that. There is almost no justification. The power profile is not representing the real world. And models should only be sharpened with reference to real-world data, not in reference to whatever correlations which might be out there in literature.
Can you show your comparison in a diagram? In my honest opinion SST turbulence model should result in a fully developed turbulent velocity profile. Which of cause is not identical (!) with a power law formula, since the latter is a very simplified empiral correlation and in particularly in the center of the pipe it will lead to deviation from the CFD calculated profile.
Thomas Frank Hello Thomas,
This is the profile I am getting but is it possible to get the exact profile by changing mesh parameters or model selection?
I would judge this a pretty perfect agreement. Why changing a turbulence model like SST, which is much more universal and most likely even more accurate in comparison to experimental data, just to be in agreement with a formula, which by itself is only a coarse representation of the time or ensemble averaged flow field in a pipe? I would very discourage to do that. There is almost no justification. The power profile is not representing the real world. And models should only be sharpened with reference to real-world data, not in reference to whatever correlations which might be out there in literature.
Thomas Frank Dear sir, would you suggest i go ahead with these results? there is another formula which has been mentioned in literature (which uses friction factor), and on using that formula , i get exactly the same results. MY Nusselt number correlations also show an error of only 8% (Dittus Bolter and Gnilienski) which is acceptable as they are known to vary by 20%
Yes, depending on what you finally would like to achieve I would recommend to go ahead with this SST result.
- Badges
- Science topic