[nishant_hull]

Hi forum...

I was trying to get the postchannel plots for the channel395 case in OF-1.5. Unfortunately postChannel utility has not been able to generate the Uf.xy, u.xy etc etc plots in OF-1.5.

I noticed that instead of Umean, now OF-1.5 writes UMean (capital M) in every time directory. Also R file is not present in time directory. Is this error in postChannel utility because of that? I go through both the codes and find no difference. I am still looking into it but It will be helpful if someone can tell me fast.

regards,
Nishant

[nishant_hull]

can anyone take out some time to comment on the above problem. I changed the postchannel utility with UMean instead of Umean.. and now the utility is complaining about R file not found.
Now the question is, why channeloodle utility is not writting R field??

[leonardo.morita]

Hello,

I have the same problem...after running the postChannel utility, no files are created...has somebody figured out what the cause of the problem is?

Thanks,

Leonardo

[lakeat]

1. As I remember, R is the same as UPrime2Mean
2. Many of us, would rewrite postChannel utility a bit to get what we need, just search the forum and spend some time to work it out.

[fs82]

UPrime2Mean is not R I guess. Nevertheless you have to use UPrime2Mean to calculate the reynolds stresses.

[lakeat]

Originally Posted by fs82

UPrime2Mean is not R I guess. Nevertheless you have to use UPrime2Mean to calculate the reynolds stresses.

Then, their differences are???

[fs82]

Allright I think the following, but this is just an assumption, not proved up to now:

R = average(u'_i u'_j)
UprimeToMean = average(u_i u_j)

An by this R becomes:

R = average ( u'_i u'_j) = average(u_i u_j) - average(u_i) * average (u_j) (1)

Edit: I found a working postChannel in the GIT repository. But I am still 99% sure, that R is not the same als Uprime2Mean, if R is the reynolds stress tensor. But I am not able to figure out where the calculation (see equation 1) is done so far. But I am not so familiar with C++, so it is a bit difficult for me to understand the source code :-/ Please correct me if I am wrong.

[fs82]

Mea culpa :-/ I went into the fieldAverage routines and checked the calculation of the reynolds stresses and I have to take back everything and state the opposite. OpenFoam calculates average(u_i u_j) but writes out everytime step average(u'_i u'_j). This is done by equation (1). But before calculation the new reynolds stress tensor equation (1) is used again to get average(u_i u_j) from average(u'_i u'_j). So everything is fine and you can use UPrimeToMean instead of R from former OpenFoam versions.

Sry,
fs82

[lakeat]

[sam1364]

Hey guys,

I am trying to get Reynolds Stress Components for Lid driven cavity flow. By using fieldaveraging, I get 6 components for uprime2mean as written below

UPrime2Mean:0
UPrime2Mean:1
UPrime2Mean:2
UPrime2Mean:3
UPrime2Mean:4
UPrime2Mean:5

My questions are that, are these components the Reynolds Stress tensor?!
Which of these representing <uu>, <uv>, <vv> and etc?!


Thanks

[fs82]

Hey,

have a look for example at this page http://www.foamcfd.org/Nabla/guides/...sGuidese4.html.

[sam1364]

Thanks. I went through this website, but I could not get any help from that. Do you think these six components are Reynolds Stress tensor?!

I guess Uprime2Mean (0) is <uu> and Uprime2Mean (1) is <vv>

[fs82]

Alright, the site I suggested wasn't the best one to read. The best way is to try it out. I guess you have performed a test run and already have the field "UPrime2Mean". So do the following: "foamCalc components UPrime2Mean" in your foam case directory. This will provide you each component uu, vv, ww, uv, uw, vw in a separate file. Compare the values of your original symmTensor field to the ones in the files and you have the answer.
For my understanding UPrime2Mean represents more or less the Reynolds stresses. If you perform an LES you have:
Mean Value + Fluctuations + Subgrid stresses = Instantaneous value
and UPrime2Mean include for my understanding only the Fluctuations. In case of RANS I am not sure, but I think UPrime2Mean does not exist there (never tried a RANS), but you got an additional field named R, which represents the Reynolds stresses. May be one of the more advanced user could give you more information.

kind regards,
Fabian

[sam1364]

Thanks for your complete response. I do appreciate your help.
I will try those suggested ways.


Best Wishes,

[sam1364]

Fabian,

I tried to get access to the source code to check the calculation of UPrime2Mean. I went through fieldAverage.C and fieldAverage.H, but unfortunately I could not see the calculation of Uprime2Mean in those files. Where should I check that out?

Thanks,

Originally Posted by fs82

Alright, the site I suggested wasn't the best one to read. The best way is to try it out. I guess you have performed a test run and already have the field "UPrime2Mean". So do the following: "foamCalc components UPrime2Mean" in your foam case directory. This will provide you each component uu, vv, ww, uv, uw, vw in a separate file. Compare the values of your original symmTensor field to the ones in the files and you have the answer.
For my understanding UPrime2Mean represents more or less the Reynolds stresses. If you perform an LES you have:
Mean Value + Fluctuations + Subgrid stresses = Instantaneous value
and UPrime2Mean include for my understanding only the Fluctuations. In case of RANS I am not sure, but I think UPrime2Mean does not exist there (never tried a RANS), but you got an additional field named R, which represents the Reynolds stresses. May be one of the more advanced user could give you more information.

kind regards,
Fabian

[fs82]

You have to go to the follwing directory: <foampath>/OpenFOAM-1.6.x/src/postProcessing/functionObjects/field/fieldAverage/fieldAverage and everything is there. The calculation of the mean and Prime2Mean values could be found in fieldAverageTemplates.C after a quick look.

hope this helps,
Fabian