Umean and U

xiyuqiu · May 13, 2013, 19:49

Hello All,

Does anyone know what is the difference between Umean and U from the flow field when running turbulent model? I first thought Umean is the mean velocity and U is the instantaneous velocity, but I do not see any random fluctuations with U. Can someone help? Thanks.

ngj · May 15, 2013, 15:22

Hi Yu,

You would need to tell, which solver you are using, and which OF version you are using in order to except answers to your questions.

You would also need to tell, whether you are using RANS, LES or laminar turbulence model in your simulations.

Kind regards

Niels

xiyuqiu · May 16, 2013, 13:09

Hello Niels,

Thanks for the reply.

I am using k-epsilon and k-omega model, simpleFoam solver, OpenFoam 2.0.1. Sorry about not being clear about it.

best,

yu

ngj · May 17, 2013, 03:34

Hi Yu,

It sounds a bit weird, because simpleFoam is a steady state solver, so I do not know, where you get the Umean field from.
I looked in simpleFoam and the incompressible RAS turbulence models, and I could not find the definition of the Umean field.

Secondly, a steady state solver does not have information on the turbulent fluctuations.

Kind regards

Niels

xiyuqiu · May 17, 2013, 13:05

Hello Niels,

Thank you for your input. When I finish my openFoam run, in the result folder 3000 (I let it run for 3000 iterations), there are Umean and U files in it. When I plot them in paraView, they are clearly different for some case, close for some other case. Also, when I run buoyantBoussinesqSimpleFoam, which suppose to be a steady solver as well, I see my Umean and U velocity fields clearly been different. I will try to post pictures in the next two messages

xiyuqiu · May 17, 2013, 13:16

Umean_pic.png

U_pic.png

ngj · May 17, 2013, 19:19

Do you perform any runTime processing in your controlDict? To me it sounds like an averaging procedure is activated.

Are your simulation converged?

Kind regards

Niels

xiyuqiu · May 18, 2013, 00:49

yes. i think there is a runTime processing in my controlDict. I thought it is some kind of sudo-steady simulation i.e. it is a steady simulation, but the run time is just a way to define the number of time steps needs to run. You are also correct about the residuals are not converging for the pipe case I showed here. However, what is shown here is the buoyantBoussinesqSimpleFoam heat transfer case. When I run simpleFoam solver, the case is converging and I got both Umean and U and they are different as well.

ngj · May 18, 2013, 04:11

Of course the two fields will be different, since you are averaging a lot of non-equal fields and comparing with the final time step.

In my perspective Umean does not make sense for a steady state simulation.

Kind regards

Niels

xiyuqiu · May 19, 2013, 18:55

So you think the Umean is the averaging between the time step? And the U velocity in the results is the mean turbulent flow velocity?

May 13, 2013, 19:49	Umean and U	#1
xiyuqiu Member yu Join Date: Nov 2010 Posts: 39 Rep Power: 15	Hello All, Does anyone know what is the difference between Umean and U from the flow field when running turbulent model? I first thought Umean is the mean velocity and U is the instantaneous velocity, but I do not see any random fluctuations with U. Can someone help? Thanks.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Access UMean at runtime	deji	OpenFOAM Programming & Development	3	April 25, 2016 19:20
Whatbs the meaning of variables in the computed results	yfjma	OpenFOAM Post-Processing	8	November 12, 2012 01:52
Error in channelOodles tutorial	alberto	OpenFOAM Bugs	62	June 8, 2010 07:13
postChannel bug	zebu83	OpenFOAM Bugs	9	November 23, 2009 02:46
Umean	balakrishnan	OpenFOAM Running, Solving & CFD	4	March 11, 2005 11:31

May 15, 2013, 15:22		#2
ngj Senior Member Niels Gjoel Jacobsen Join Date: Mar 2009 Location: Copenhagen, Denmark Posts: 1,900 Rep Power: 37	Hi Yu, You would need to tell, which solver you are using, and which OF version you are using in order to except answers to your questions. You would also need to tell, whether you are using RANS, LES or laminar turbulence model in your simulations. Kind regards Niels

May 16, 2013, 13:09		#3
xiyuqiu Member yu Join Date: Nov 2010 Posts: 39 Rep Power: 15	Hello Niels, Thanks for the reply. I am using k-epsilon and k-omega model, simpleFoam solver, OpenFoam 2.0.1. Sorry about not being clear about it. best, yu

May 17, 2013, 03:34		#4
ngj Senior Member Niels Gjoel Jacobsen Join Date: Mar 2009 Location: Copenhagen, Denmark Posts: 1,900 Rep Power: 37	Hi Yu, It sounds a bit weird, because simpleFoam is a steady state solver, so I do not know, where you get the Umean field from. I looked in simpleFoam and the incompressible RAS turbulence models, and I could not find the definition of the Umean field. Secondly, a steady state solver does not have information on the turbulent fluctuations. Kind regards Niels

May 17, 2013, 13:05		#5
xiyuqiu Member yu Join Date: Nov 2010 Posts: 39 Rep Power: 15	Hello Niels, Thank you for your input. When I finish my openFoam run, in the result folder 3000 (I let it run for 3000 iterations), there are Umean and U files in it. When I plot them in paraView, they are clearly different for some case, close for some other case. Also, when I run buoyantBoussinesqSimpleFoam, which suppose to be a steady solver as well, I see my Umean and U velocity fields clearly been different. I will try to post pictures in the next two messages

May 17, 2013, 13:16		#6
xiyuqiu Member yu Join Date: Nov 2010 Posts: 39 Rep Power: 15	Umean_pic.png U_pic.png

May 17, 2013, 19:19		#7
ngj Senior Member Niels Gjoel Jacobsen Join Date: Mar 2009 Location: Copenhagen, Denmark Posts: 1,900 Rep Power: 37	Do you perform any runTime processing in your controlDict? To me it sounds like an averaging procedure is activated. Are your simulation converged? Kind regards Niels

May 18, 2013, 00:49		#8
xiyuqiu Member yu Join Date: Nov 2010 Posts: 39 Rep Power: 15	yes. i think there is a runTime processing in my controlDict. I thought it is some kind of sudo-steady simulation i.e. it is a steady simulation, but the run time is just a way to define the number of time steps needs to run. You are also correct about the residuals are not converging for the pipe case I showed here. However, what is shown here is the buoyantBoussinesqSimpleFoam heat transfer case. When I run simpleFoam solver, the case is converging and I got both Umean and U and they are different as well.

May 18, 2013, 04:11		#9
ngj Senior Member Niels Gjoel Jacobsen Join Date: Mar 2009 Location: Copenhagen, Denmark Posts: 1,900 Rep Power: 37	Of course the two fields will be different, since you are averaging a lot of non-equal fields and comparing with the final time step. In my perspective Umean does not make sense for a steady state simulation. Kind regards Niels

May 19, 2013, 18:55		#10
xiyuqiu Member yu Join Date: Nov 2010 Posts: 39 Rep Power: 15	So you think the Umean is the averaging between the time step? And the U velocity in the results is the mean turbulent flow velocity?