CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Post-Processing (https://www.cfd-online.com/Forums/openfoam-post-processing/)
-   -   A problem on the Energy spectrum plot by using FFT (https://www.cfd-online.com/Forums/openfoam-post-processing/122502-problem-energy-spectrum-plot-using-fft.html)

Martin duan August 21, 2013 05:55

A problem on the Energy spectrum plot by using FFT
 
1 Attachment(s)
Hi, guys

I am using the FFT to plot the energy spectrum of the turbulent velocity at the axial direction. but the figure shows that the sidelobes decrease very fast at the high frequencies, just as shown in black rectangular in the attached figure. I spend a lot of time to find out the reason. If you come across such problem before, please give me some clues.

P.S. The precision of the original singular is double precision.

Thanks in advance.

rkc.cfd October 23, 2013 13:13

I have used the Spectrum Function for LES data from the following link and it worked out fine.

Python Scripts for Turbulent Statistics

toolpost February 23, 2014 05:23

1 Attachment(s)
Hello guys!

I am simulating flow through a constricted pipe at Re=1000 using LES. DNS studies available in the literature predicts it as fully turbulent, and I am also getting similar results. I probed velocity at a suitable location and wrote a code in octave to create the spectrum. But I am getting the same trend similar to Martin's case. Towards the higher frequencies, the oscillations in energy are gone, and the curve looks like a straight line. Is this correct? Is there any problem with the simulation? or the fft code is the one to blame? Somebody please give some advice.. :confused:

Thanks a lot for your time and consideration.

huangxianbei April 20, 2014 10:24

Quote:

Originally Posted by toolpost (Post 476283)
Hello guys!

I am simulating flow through a constricted pipe at Re=1000 using LES. DNS studies available in the literature predicts it as fully turbulent, and I am also getting similar results. I probed velocity at a suitable location and wrote a code in octave to create the spectrum. But I am getting the same trend similar to Martin's case. Towards the higher frequencies, the oscillations in energy are gone, and the curve looks like a straight line. Is this correct? Is there any problem with the simulation? or the fft code is the one to blame? Somebody please give some advice.. :confused:

Thanks a lot for your time and consideration.

I think this comes from the temporal spectrum, in fact, the smooth curve appears in books and articles is averaged in time. BTW, how is the spectrum obtain? I'm disoriented now:(

lx882211 February 22, 2016 11:11

Quote:

Originally Posted by toolpost (Post 476283)
Hello guys!

I am simulating flow through a constricted pipe at Re=1000 using LES. DNS studies available in the literature predicts it as fully turbulent, and I am also getting similar results. I probed velocity at a suitable location and wrote a code in octave to create the spectrum. But I am getting the same trend similar to Martin's case. Towards the higher frequencies, the oscillations in energy are gone, and the curve looks like a straight line. Is this correct? Is there any problem with the simulation? or the fft code is the one to blame? Somebody please give some advice.. :confused:

Thanks a lot for your time and consideration.

Hi, could you tell me how to get the TKE, which parameter should i monitor? Thanks.

lx882211 February 22, 2016 11:12

Quote:

Originally Posted by Martin duan (Post 447089)
Hi, guys

I am using the FFT to plot the energy spectrum of the turbulent velocity at the axial direction. but the figure shows that the sidelobes decrease very fast at the high frequencies, just as shown in black rectangular in the attached figure. I spend a lot of time to find out the reason. If you come across such problem before, please give me some clues.

P.S. The precision of the original singular is double precision.

Thanks in advance.

Hi, could you tell me how to get the TKE, which parameter should i monitor? Thanks.


All times are GMT -4. The time now is 00:30.