TurbulentInlet BC

chris1980 · September 5, 2006, 05:09

I want to use the 'turbulentInlet' boundary condition. As far as I can see the use only inputs a fluctuation scale but what about the integral length scale or other appropiate scales because the flucation scale alone do not define a turublent inlet.

Ok I know it is anyway no real turbulent inlet because there is no turbulence strcuture etc.

Additionally, I was wondering what I have to set for k and epsilon at this turbulent boundary (using high Re k-epsilon turb model)?

chris1980 · September 5, 2006, 11:15

Is there nobody who is able or want to answer my question?

pierre · September 5, 2006, 12:21

Hi,

I think that that turbulentInlet is to be used for LES/DNS, not RANS. It just produces a random noise at the inlet of given magnitude, which anyway is most lileky to be dissipated quickly.

For RANS, if you use K-Eps, just use fixed value boundary conditions and set k and Eps to what you like.
You can use something like
K_inlet = 3/2 (Uref*T)^2 and Eps = Cmu*k^(3/2)/L
where T is tubulence intensity (10% for example) and L is a turbulent lengthscale.

Hope this helps

daniel_mills · September 20, 2006, 13:50

Excuse me for the stupid question. I did not know you could actually set BC's like this. What would an example of the actual inlet BC be for k and epsilon?

sihaqqi · September 9, 2013, 00:32

chris

if you have found answer to your query, can you please let me know. I want to use turbulentInlet BC.

regards

vut · March 27, 2014, 08:35

Hi,

I am searching what does "fluctuationScale" in turbulentInlet BC mean?

If I set it 0.01, for example? What is this value signify?

Thanks in advance,

vut

Garusen · October 21, 2015, 04:32

To get information, OpenFoam source code is the best place.

If you read comment in the .H file, you have in comments:
fluctuationScale
RMS fluctuation, provided as the fraction of the mean field

So 0.001 means the fluctuation will be 0.001 time the mean flow.

To be used in RANS, you need to create fields for k and epsilon as explained above, for example, see implementation of AtmBoundaryLayer.

beatlejuice · July 4, 2017, 10:02

Hello everybody,

I also want to use the turbulentInlet BC for an LES simulation. The case of interest is a flow over a cylinder - longitudinal orientated to the flow.

I'm not quite sure how to set the fluctuationSclae value. My idea was the following:

Since i know the turbulent intensity of the flow field (freestream), I want to use this to set the fluctuationScale.

Def. of turb. Intensity: $I = \frac{u'}{\overline{U}}$

with $u' = \sqrt{\frac{1}{3}((u'_x)^2+(u'_y)^2+(u'_z)^2)}$

where $\overline{U}$ is the magnitude of the mean velocity @ inlet and

is the root-mean-square of the turbulent velocity fluctuations.

By assuming that

and $\overline{U} = 1$ i can rewrite the first equ. to

.

Therefore, I want to set the fluctuationScale components equal to turbulence intensity.

Is this a way to set the fluctuationScale value or do i make a fundamental mistake by following this idea? How can I estimate the fluctuationScale values for a flow case like mine?

Any help or ideas are greatly appreciated =) thanks!

hconel · January 18, 2020, 09:16

Quote:

Originally Posted by beatlejuice

Hello everybody,

I also want to use the turbulentInlet BC for an LES simulation. The case of interest is a flow over a cylinder - longitudinal orientated to the flow.

I'm not quite sure how to set the fluctuationSclae value. My idea was the following:

Since i know the turbulent intensity of the flow field (freestream), I want to use this to set the fluctuationScale.

Def. of turb. Intensity: $I = \frac{u'}{\overline{U}}$

with $u' = \sqrt{\frac{1}{3}((u'_x)^2+(u'_y)^2+(u'_z)^2)}$

where $\overline{U}$ is the magnitude of the mean velocity @ inlet and

is the root-mean-square of the turbulent velocity fluctuations.

By assuming that

and $\overline{U} = 1$ i can rewrite the first equ. to

.

Therefore, I want to set the fluctuationScale components equal to turbulence intensity.

Is this a way to set the fluctuationScale value or do i make a fundamental mistake by following this idea? How can I estimate the fluctuationScale values for a flow case like mine?

Any help or ideas are greatly appreciated =) thanks!

Hello,
I'm also in a similar situation. The RMS velocity depends on both the fluctuationScale and the mean velocity. Hence, I thought it would be a good idea to write a simple script which would generate an array of randomized velocity data and then do the calculations.

For example, for a mean inlet velocity of (8,0,0) m/s and fluctuationScale = (0.1,0.1,0.1), the inlet velocity ranges between (7.2,0,0) m/s and (8.8,0,0) m/s. Generating a 100000 point sample data and doing the calculations gives me the following results:

u_rms = (0.46271,0,0) (root mean square of velocity)
u_prime = 0.26714
I = 0.033 (turbulence intensity)
k = 0.10705 (turbulent kinetic energy)

Calculations can be found here: http://web.mit.edu/fluids-modules/ww...vel.record.pdf

Of course, this can be extended to a case with non-zero y and z velocity components. However, I do not know a way to directly find a value for turbulentFluctuations given the turbulence intensity.

Although, I have two questions:
1) As far as I know, there is no way to add non-zero fluctuations to a velocity component if the mean value is 0 using turbulentInlet patch type. Anybody knows a workaround?
2) In the calculations, time scale is not used anywhere. i.e. a 25 Hz data and 250 Hz data would yield the same results. This sounds not reasonable to me...

I would appreciate any ideas.
Best regards,
Huseyin

September 5, 2006, 05:09	I want to use the 'turbulentIn	#1
chris1980 Member chris book Join Date: Mar 2009 Posts: 85 Rep Power: 17	I want to use the 'turbulentInlet' boundary condition. As far as I can see the use only inputs a fluctuation scale but what about the integral length scale or other appropiate scales because the flucation scale alone do not define a turublent inlet. Ok I know it is anyway no real turbulent inlet because there is no turbulence strcuture etc. Additionally, I was wondering what I have to set for k and epsilon at this turbulent boundary (using high Re k-epsilon turb model)?

September 5, 2006, 11:15	Is there nobody who is able or	#2
chris1980 Member chris book Join Date: Mar 2009 Posts: 85 Rep Power: 17	Is there nobody who is able or want to answer my question?

September 5, 2006, 12:21	Hi, I think that that turbu	#3
pierre Member Pierre Le Fur Join Date: Mar 2009 Location: UK Posts: 60 Rep Power: 17	Hi, I think that that turbulentInlet is to be used for LES/DNS, not RANS. It just produces a random noise at the inlet of given magnitude, which anyway is most lileky to be dissipated quickly. For RANS, if you use K-Eps, just use fixed value boundary conditions and set k and Eps to what you like. You can use something like K_inlet = 3/2 (UrefT)^2 and Eps = Cmuk^(3/2)/L where T is tubulence intensity (10% for example) and L is a turbulent lengthscale. Hope this helps

September 20, 2006, 13:50	Excuse me for the stupid quest	#4
daniel_mills New Member Daniel Mills Join Date: Mar 2009 Posts: 13 Rep Power: 17	Excuse me for the stupid question. I did not know you could actually set BC's like this. What would an example of the actual inlet BC be for k and epsilon?

October 21, 2015, 04:32		#7
Garusen New Member Benoit Join Date: Jul 2015 Location: France Posts: 4 Rep Power: 10	To get information, OpenFoam source code is the best place. If you read comment in the .H file, you have in comments: fluctuationScale RMS fluctuation, provided as the fraction of the mean field So 0.001 means the fluctuation will be 0.001 time the mean flow. To be used in RANS, you need to create fields for k and epsilon as explained above, for example, see implementation of AtmBoundaryLayer. songwukong, beatlejuice, amolrajan and 2 others like this.

September 9, 2013, 00:32		#5
sihaqqi Senior Member starter Join Date: Sep 2012 Posts: 125 Rep Power: 15	chris if you have found answer to your query, can you please let me know. I want to use turbulentInlet BC. regards

March 27, 2014, 08:35		#6
vut Member Join Date: Feb 2014 Posts: 57 Rep Power: 12	Hi, I am searching what does "fluctuationScale" in turbulentInlet BC mean? If I set it 0.01, for example? What is this value signify? Thanks in advance, vut

July 4, 2017, 10:02		#8
beatlejuice Member Join Date: Nov 2013 Posts: 35 Rep Power: 12	Hello everybody, I also want to use the turbulentInlet BC for an LES simulation. The case of interest is a flow over a cylinder - longitudinal orientated to the flow. I'm not quite sure how to set the fluctuationSclae value. My idea was the following: Since i know the turbulent intensity of the flow field (freestream), I want to use this to set the fluctuationScale. Def. of turb. Intensity: $I = \frac{u'}{\overline{U}}$ with $u' = \sqrt{\frac{1}{3}((u'_x)^2+(u'_y)^2+(u'_z)^2)}$ where $\overline{U}$ is the magnitude of the mean velocity @ inlet and is the root-mean-square of the turbulent velocity fluctuations. By assuming that and $\overline{U} = 1$ i can rewrite the first equ. to . Therefore, I want to set the fluctuationScale components equal to turbulence intensity. Is this a way to set the fluctuationScale value or do i make a fundamental mistake by following this idea? How can I estimate the fluctuationScale values for a flow case like mine? Any help or ideas are greatly appreciated =) thanks! amolrajan and Abk005 like this.