[tony256]

Hii,
I am still stuck on this problem. Please help!

I am new to OpenFoam and trying to simulate flow past a circular cylinder at Re = 3900. But my drag coefficient is coming out to be very low. I have attached the required file. Please help where i am getting wrong?

I am using rhoPimpleFoam solver.
Please ask if you need to know any more data but please help.
Thanks in advance.
https://drive.google.com/drive/folde...Vh?usp=sharing

[Ardali]

Are you using any turbulence model? If no, you are right stay with laminar.

Is it 2D case with empty boundaries in sides? If yes, that is a problem in underestimating the drag.
The only reason with low drag is coarse mesh. Meh the resolution finer in azimuthal and wall normal directions.

[tony256]

Originally Posted by Ardali

Are you using any turbulence model? If no, you are right stay with laminar.

Is it 2D case with empty boundaries in sides? If yes, that is a problem in underestimating the drag.
The only reason with low drag is coarse mesh. Meh the resolution finer in azimuthal and wall normal directions.

yes i amusing kEqn turbulent model.

Ok, then can you suggest what BC can be used at front and back?
Thank you so much the reply.

[GasHighSpeedMixer]

Hi and welcome to OpenFOAM!

How do you calculate the drag? I suggest you to use functions (forceCoeff) in your controlDict, if you did not already do that. Are you sure, that Aref and lref are correct? If your case is 2D, front and back are supposed to be empty.

Kind regards

[tony256]

Originally Posted by GasHighSpeedMixer

Hi and welcome to OpenFOAM!

How do you calculate the drag? I suggest you to use functions (forceCoeff) in your controlDict, if you did not already do that. Are you sure, that Aref and lref are correct? If your case is 2D, front and back are supposed to be empty.

Kind regards

Hii, Thanks for your reply.
The diameter of cylinder in my case is 1.
Hence, i have set the value lref = 1 and Aref = 0.041315 (lref*pi*D/76).
This is a 3D case, still i am not getting correct results. Please see the drive link i have attached and help!.

[GasHighSpeedMixer]

Originally Posted by tony256

The diameter of cylinder in my case is 1.
Hence, i have set the value lref = 1 and Aref = 0.041315 (lref*pi*D/76).
This is a 3D case, still i am not getting correct results. Please see the drive link i have attached and help!.

I did a lot of airfoil simulation in the past few months to calculate lift and drag and your case is quite similar to an airfoil calculation. On the first glance it seems that you have calculated Aref wrong. In an airfoil case, lref = chord length, while Aref = lref*thickness perpendicular to lref. This means in your case Aref = 1*thickness (thickness is here the length of the cylinder). This might solve your problem.

[Ardali]

Originally Posted by tony256

yes i amusing kEqn turbulent model.

Ok, then can you suggest what BC can be used at front and back?
Thank you so much the reply.

Is it LES k equation model?
If you are using 3D case with spanwise extension, that should be fine enough. You can find reasonable resolution in the literature.

You should know that your viscous sublayer in fully laminar. It makes you simulation tricky. conventional URANS models do not work for your case. Keep your domain large enough and resolution fine enough.
It will work.
Good luck
Ardalan

[tony256]

I am still not getting correct results. Can you send me your email ID so that i can send you my case file? I am really stuck on this problem.

[tony256]

Originally Posted by GasHighSpeedMixer

I did a lot of airfoil simulation in the past few months to calculate lift and drag and your case is quite similar to an airfoil calculation. On the first glance it seems that you have calculated Aref wrong. In an airfoil case, lref = chord length, while Aref = lref*thickness perpendicular to lref. This means in your case Aref = 1*thickness (thickness is here the length of the cylinder). This might solve your problem.

I am still not getting correct results. Can you send me your email ID so that i can send you my case file? I am really stuck on this problem.

[GasHighSpeedMixer]

I advise you to watch this tutorial. This should describe your case very well. So the learning effect is definitely higher:
https://youtu.be/YOT62t2IiDE

[tony256]

Originally Posted by GasHighSpeedMixer

I advise you to watch this tutorial. This should describe your case very well. So the learning effect is definitely higher:
https://youtu.be/YOT62t2IiDE

Actually Re = 3900 is not complete turbulent case. It is in transition region. I am using rhoPimpleFoam. Is the selection of Solver correct or it depends on other things?

[GasHighSpeedMixer]

Have a look on the OpenFoam solver list:
https://www.openfoam.com/documentati...rd-solvers.php

pimpleFoam solves transient simulations and stands for the property of a parameter to follow a constant value over time.

rhoPimpleFoam takes compressible and heat transfer effects into account.

simpleFoam solves steady-state simulations.

I would recommend you to start easy and to use simpleFoam. If that works and you want to simulate something similar like the Karman Vortex Street, switch over to pimpleFoam. But I don't think that you should take compressibility (rhoPimpleFoam) into account at these low velocities.

[tony256]

Originally Posted by GasHighSpeedMixer

Have a look on the OpenFoam solver list:
https://www.openfoam.com/documentati...rd-solvers.php

pimpleFoam solves transient simulations and stands for the property of a parameter to follow a constant value over time.

rhoPimpleFoam takes compressible and heat transfer effects into account.

simpleFoam solves steady-state simulations.

I would recommend you to start easy and to use simpleFoam. If that works and you want to simulate something similar like the Karman Vortex Street, switch over to pimpleFoam. But I don't think that you should take compressibility (rhoPimpleFoam) into account at these low velocities.

Hey! Actually my case is a 3D problem. In this i don't think Aref = length * Diameter ?