CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   LES energy spectrum (http://www.cfd-online.com/Forums/cfx/25956-les-energy-spectrum.html)

 Hummi June 16, 2008 16:18

LES energy spectrum

Hi Guys

I want to plot energy spectrum (temporal spectrum) for my LES simulation. Can you guys please explain how can i do that. I have matlab and mathcad. If some one has a code for it and can pass it on to me, it will very helpful. Thank you so much.

 Glenn Horrocks June 16, 2008 21:41

Re: LES energy spectrum

Hi,

Put a monitor point in your domain and record the U, V, W velocities (or any other turbulence parameter which you wish to get the spectrum from, maybe pressure, temperature, species etc). Do an FFT on it. Draw spectrum.

Glenn Horrocks

 Hummi June 17, 2008 05:15

Re: LES energy spectrum

Thank you Glenn. Which velocities should I monitor, instantaneous or mean? How do I do FFT, can you please expand on that. Thanks

Hummi

 Hummi June 18, 2008 16:10

Re: LES energy spectrum

I know how to do FFT on it velocities. How do I do last bit i.e. draw spectrum.

Hummi

 eng1182 February 12, 2010 08:21

hi Hummi

energy spectrum is a type of transient results representation which only needs any variable as a function of time and transform it to the strength of this variable as a function of frequency. so it dose not depend on a specified variable the variable can be any variable like velocity, pressure or temperature

 MadsR February 15, 2010 04:53

Often, it would be interesting to show a separate energy spectrum from samplings of u, v, and w. But, you could show an energy spectrum for the resulting velocity (i.e. the speed) only. It depends on your case really.
If you want to show/look for turbulent coherent structures in the flow, velocities would be a good choice. Pressure would also reflect this, but obviously you loose the directional information seen in separate spectra of u,v,w.
One could also make a spectrum of the lift/drag forces on an airfoil to try to investigate phenomena related to the surface flow. Possibilities are infinite :-)

Hope it helps you a bit.

Oh, and remember to sample at at least twice the rate as you want to resolve in your spectrum and also to get "sufficiently" long time-series. If you can see some periodicity in the velocities, for instance, be sure to get, say, at least 20 of those cycles. Again, this depends heavily on the flow at hand, and needs your engineering judgement :-) all kinds of weird stuff can happen when sampling a signal, consult text books or Google on this :-)

/Mads

 All times are GMT -4. The time now is 21:43.