Re200 flow around a square cylinder question.

fhjkfhjk · January 16, 2017, 03:38

Hello all.

I am new to openFoam and I am trying to simulate a flow over a square cylinder.

I have been researching forum and papers for a few days to choose proper solver to my simulation.

In the meantime, I have run the case and I have got the result looking fine with Vortex sheddings which is occuring as time goes by.

I have used pimpleFoam with RANS turbulence model k-epsilon model.

As far as I know, a flow of Re200 around a cylinder, character of the flow is laminar, but why am I getting vortex shedding? Is this because I used turbulence model? or my physical knowledge is wrong?

I have run with pimpleFoam turning the turbulence off (laminar) and got the same result.

Am I doing it right? Any comment is welcome. If you need more information of my question or my case, feel free to ask me.

Thank you in advance and hope you all are fine.

wildfire230 · January 16, 2017, 15:52

Hi fhjkfhjk, for Reynolds numbers probably greater than about 50 I would expect vortex shedding behind a square cylinder. This isn't coming from the turbulence model, it just comes from the nature of the Navier-Stokes equations. For Reynolds number 200 the symmetric steady solution that I think you were expecting is actually unsteady. If you decrease the Reynolds number to 10 you will probably see the steady solution you were expecting.

fhjkfhjk · January 17, 2017, 07:06

I appreciate your reply wildfire230. I should change Re number and see what hppens as you mentioned.

Furthermore, Your answer sounds like any simulationType in pimpleFoam/turbuleceModel can be applied to the laminar case among laminar, or RANS because the sheddings are occured by NS equation not turbulenceModel. I guess so far, both cases would give me the same result.

Am I following properly?

Thank you again.

wildfire230 · January 17, 2017, 10:31

Honestly I'm not that familiar with turbulence modeling. I'm not sure at what Reynolds number this flow transitions to turbulence, but I guess for Reynolds numbers less than 300-500 you should just use turbulenceModel laminar. Also, you can actually solve turbulent flows using turbulenceModel laminar, just so long as you use a fine enough grid (i.e. DNS).

fhjkfhjk · January 24, 2017, 09:22

Thank you for your reply, wildfire230.

I have found this laminarModel while studying slover's equations.

Basically, I thought there is no such laminarModel. But, there are some models.

Take a look at these links below,

https://develop.openfoam.com/Develop...Models/laminar

https://develop.openfoam.com/Develop...enceProperties

These make me confusing. and also, compare to pisoFoam/laminar, the pisoFoam does not use laminarModel whereas pimpleFoam does. is there any difference whether using it or not?

Feel free to comment to this thread, anyone!.

Thanks

January 16, 2017, 03:38	Re200 flow around a square cylinder question.	#1
fhjkfhjk New Member Dae-young(Dan) Join Date: Jan 2017 Location: South Korea Posts: 6 Rep Power: 9	Hello all. I am new to openFoam and I am trying to simulate a flow over a square cylinder. I have been researching forum and papers for a few days to choose proper solver to my simulation. In the meantime, I have run the case and I have got the result looking fine with Vortex sheddings which is occuring as time goes by. I have used pimpleFoam with RANS turbulence model k-epsilon model. As far as I know, a flow of Re200 around a cylinder, character of the flow is laminar, but why am I getting vortex shedding? Is this because I used turbulence model? or my physical knowledge is wrong? I have run with pimpleFoam turning the turbulence off (laminar) and got the same result. Am I doing it right? Any comment is welcome. If you need more information of my question or my case, feel free to ask me. Thank you in advance and hope you all are fine.

January 16, 2017, 15:52		#2
wildfire230 Senior Member Join Date: Jul 2013 Posts: 124 Rep Power: 12	Hi fhjkfhjk, for Reynolds numbers probably greater than about 50 I would expect vortex shedding behind a square cylinder. This isn't coming from the turbulence model, it just comes from the nature of the Navier-Stokes equations. For Reynolds number 200 the symmetric steady solution that I think you were expecting is actually unsteady. If you decrease the Reynolds number to 10 you will probably see the steady solution you were expecting. fhjkfhjk likes this.

January 17, 2017, 10:31		#4
wildfire230 Senior Member Join Date: Jul 2013 Posts: 124 Rep Power: 12	Honestly I'm not that familiar with turbulence modeling. I'm not sure at what Reynolds number this flow transitions to turbulence, but I guess for Reynolds numbers less than 300-500 you should just use turbulenceModel laminar. Also, you can actually solve turbulent flows using turbulenceModel laminar, just so long as you use a fine enough grid (i.e. DNS). fhjkfhjk likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Flow past a cylinder at Re 1e05 using LES, drag force coefficient is to low	Scabbard	Main CFD Forum	21	June 19, 2018 13:58
[ICEM] Flow over 2-d square cylinder	GAURAV KUMAR	ANSYS Meshing & Geometry	1	March 2, 2015 12:26
benchmark: flow over a circular cylinder	goodegg	Main CFD Forum	12	January 22, 2013 11:47
flow around a cylinder	pXYZ	Main CFD Forum	14	July 25, 2011 10:05
Simulation flow past square cylinder 3d	Gabba82	FLUENT	0	November 22, 2009 07:05

January 17, 2017, 07:06		#3
fhjkfhjk New Member Dae-young(Dan) Join Date: Jan 2017 Location: South Korea Posts: 6 Rep Power: 9	I appreciate your reply wildfire230. I should change Re number and see what hppens as you mentioned. Furthermore, Your answer sounds like any simulationType in pimpleFoam/turbuleceModel can be applied to the laminar case among laminar, or RANS because the sheddings are occured by NS equation not turbulenceModel. I guess so far, both cases would give me the same result. Am I following properly? Thank you again.

January 24, 2017, 09:22		#5
fhjkfhjk New Member Dae-young(Dan) Join Date: Jan 2017 Location: South Korea Posts: 6 Rep Power: 9	Thank you for your reply, wildfire230. I have found this laminarModel while studying slover's equations. Basically, I thought there is no such laminarModel. But, there are some models. Take a look at these links below, https://develop.openfoam.com/Develop...Models/laminar https://develop.openfoam.com/Develop...enceProperties These make me confusing. and also, compare to pisoFoam/laminar, the pisoFoam does not use laminarModel whereas pimpleFoam does. is there any difference whether using it or not? Feel free to comment to this thread, anyone!. Thanks