Turbulent Statistical Terms

Senbonzakura · December 22, 2016, 04:55

This might be a very silly question.

I am using Matlab to calculate various statistical terms like correlation coefficient, autocorrelation etc. from a set of PIV data for a round jet (I have U and V instantaneous at different x and r, with data over 1000 images taken at 50Hz). Now the fluctuations vary with time, x (axial distance from origin), and r (radial distance).

If I am finding out, let's say the correlation coefficient between the fluctuating terms u and v (axial and radial), I am confused as to how to do it (I know matlab has a built in function for corr coef.)

I want to plot corr coef vs r/D . So, I need to provide Matlab with vectors of u and v. But these vectors should vary with x or t?

In S.B.Pope's book, there is a part about one point one time characteristics. But, if that is so, then corr coef in this situation is not going to be possible?

FMDenaro · December 22, 2016, 07:19

The flow is 3D, therefore you measured somehow U and V that are depending on the azimuthal direction, too. They were averaged?
I suggest following the Pope book and implement the computation of correlation (or auto-correlation) by yourself.

Senbonzakura · December 22, 2016, 08:36

We are taking it to be planar ofc. Neglecting the dependency on azimuthal dir., my question is when we are giving the vectors for u and v for the corr coeff., is it wrt time or space? I am not asking for the code, that is not a problem if I can understand the basic idea. I think my concept of correlation coefficient has faults.

FMDenaro · December 22, 2016, 08:49

I think you should focus first on what you want to compute...you can calculate both spatial and temporal correlations (or autocorrelation) as well as the coefficient.
Matlab function does not distinguish If the vector has spatial or temporal entries so that you can use the function by organizing properly the vectors.

December 22, 2016, 04:55	Turbulent Statistical Terms	#1
Senbonzakura New Member Arnab Mukherjee Join Date: Oct 2016 Posts: 8 Rep Power: 9	This might be a very silly question. I am using Matlab to calculate various statistical terms like correlation coefficient, autocorrelation etc. from a set of PIV data for a round jet (I have U and V instantaneous at different x and r, with data over 1000 images taken at 50Hz). Now the fluctuations vary with time, x (axial distance from origin), and r (radial distance). If I am finding out, let's say the correlation coefficient between the fluctuating terms u and v (axial and radial), I am confused as to how to do it (I know matlab has a built in function for corr coef.) I want to plot corr coef vs r/D . So, I need to provide Matlab with vectors of u and v. But these vectors should vary with x or t? In S.B.Pope's book, there is a part about one point one time characteristics. But, if that is so, then corr coef in this situation is not going to be possible?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Why can i use a laminar solver for a turbulent flow? [Heat transfer problem]	blackbow	CFX	1	November 22, 2016 04:42
turbulent kinetic energy budget terms with openfoam	Thangam	OpenFOAM Programming & Development	6	July 15, 2016 09:21
Turbulent interaction terms for multiphase/Eulerian Model	openfoammaofnepo	FLUENT	0	July 9, 2015 06:16
setting value of turbulent intensity and turbulent viscosity ratio in wind tunnel	nuimlabib	Main CFD Forum	0	August 4, 2009 00:05
Turbulent terms in the MRF equations	Francisco Saldarriaga	Main CFD Forum	1	July 22, 1999 16:49

December 22, 2016, 07:19		#2
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,768 Rep Power: 71	The flow is 3D, therefore you measured somehow U and V that are depending on the azimuthal direction, too. They were averaged? I suggest following the Pope book and implement the computation of correlation (or auto-correlation) by yourself.

December 22, 2016, 08:36		#3
Senbonzakura New Member Arnab Mukherjee Join Date: Oct 2016 Posts: 8 Rep Power: 9	We are taking it to be planar ofc. Neglecting the dependency on azimuthal dir., my question is when we are giving the vectors for u and v for the corr coeff., is it wrt time or space? I am not asking for the code, that is not a problem if I can understand the basic idea. I think my concept of correlation coefficient has faults.

December 22, 2016, 08:49		#4
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,768 Rep Power: 71	I think you should focus first on what you want to compute...you can calculate both spatial and temporal correlations (or autocorrelation) as well as the coefficient. Matlab function does not distinguish If the vector has spatial or temporal entries so that you can use the function by organizing properly the vectors.