# Pressure Spectrum

 Register Blogs Members List Search Today's Posts Mark Forums Read

 March 22, 2017, 04:06 Pressure Spectrum #1 New Member   Amar Join Date: Dec 2016 Posts: 7 Rep Power: 9 Hello everyone, I've a simple question , and hope to check with you. I calculate the one dimensional spectra of the pressure in streamwise in channel flow from DNS. My procedure is as follows: 1) get the fluctuations p_prime(i,j,it)=p_ins(i,j,it)-p_m(k) " For p_ins the instantenous field, p_m is the mean value at certain height from the bottom wall. (i,j,k,it) stand for (streamwsie, spanwise ,wall-normal,time), respectively. 2) get fft of p_rime ==> p_r+i p_i 3) get the spectra of E_pp(k_x)=(p_r^2+p_i^2)*(L_x/2pai) "L_x box length in x-direction" 4) Averge over spanwsie and time domains 5) check with the r.m.s of the pressure p_rms^2=Integration(-inf to +inf E_pp(k_x) d k_x ) "k_x is streamwise wave number" However, the resulting spectrum seems to be not smooth enough. In addition the curve is little bit lower than the expected value from other studies eve if the check is correct (step # 5). Sorry for this long explnation, but any suggestions will be appreciated.

 March 22, 2017, 04:10 #2 Senior Member   Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,793 Rep Power: 71 I have never worked on pressure spectrum, could you post your results? Then, do you get good results for the velocity spectra compared to those in the database?

March 22, 2017, 04:33
#3
New Member

Amar
Join Date: Dec 2016
Posts: 7
Rep Power: 9

the wall pressure spectra is attached.

Actually, I follow the same procedure in the velocity spectra and I obtained reasonable results. I guess the same steps can be applied for the pressure.

I'm sorry for this unintended mistake, Re=1000, based on the friction velocity.
Attached Images
 wall_p_cfd_online.png (5.1 KB, 16 views)

 March 22, 2017, 04:39 #4 Senior Member   Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,793 Rep Power: 71 I don't see the plots... what about the Re number? could you post also the comparisons of the velocity spectra?

March 22, 2017, 04:57
#5
New Member

Amar
Join Date: Dec 2016
Posts: 7
Rep Power: 9
I'm sorry for the mistake. I reloaded the pressure spectrum.

the velocity spectrum at the y^+=15 is attached in loglog scale.
Attached Images
 vel_spectrum_cfd.png (3.0 KB, 13 views)

 March 22, 2017, 05:34 #6 Senior Member   Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,793 Rep Power: 71 I am no sure if the DNS database contains the pressure spectra, have you already checked there? I have no experience in the pressure wavenumbers distribution for the channel flow but it seems reasonable, there is no theoretical reason to have an inertial slope like that in the velocity

 March 22, 2017, 05:53 #7 New Member   Amar Join Date: Dec 2016 Posts: 7 Rep Power: 9 Some DNS data base provide the pressure spectra, like those of John Hopkins http://turbulence.pha.jhu.edu/ I compared my analysis with them at the same Reynolds number and obtained a difference between the results. This makes me doubts about the methodology if it is like the velocity or not. Pressure is expected theoretically to introduce a k-1 slope in the intertial subrange at the high Reynolds number. However, this slope was just approched in the experiments introduced by Tsuji et al. (2007), and up to now, it is not achieved in DNS.

 March 22, 2017, 06:42 #8 Senior Member   Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,793 Rep Power: 71 Well, could you sumperimpose your pressure spectrum and that in the database? Start with y+ corresponding to half-height channel (you wrote 1000), I do not believe you have a theoretical inertial slope close to the wall.

 Tags dns, spectra