spanwise averaging for LES backward facing step
 

I completed my LES calculation for a backward facing step and I'm processing the data.

I want to average the data in the spanwise direction but don't know how to do it.

Can anyone let me know how to do it?

Sungho

Re: spanwise averaging for LES backward facing ste
 

I´m not sure, but you can make this on CFX-Pre, "Output Control", "Trn Stats" tab, for a transient simulation. In "Transient Statistics 1", choose "Arithmetic Average". In "Output Variables List", choose "Velocity".

Do u wanna calculate the recirculation point´in the average time? I did it and my result is about 6.

Re: spanwise averaging for LES backward facing ste
 

Thank you. I will try it. My recirculation point is about 7 based on mid span data. (experimental data indicates 6 as you predicted) I may need to refine the grid.

By the way, do you think it is better to average in spanwise direction rather than just process the data at mid span? Le, Moin and Kim(1997) averaged in spanwise direction in their DNS calculation.

Re: spanwise averaging for LES backward facing ste
 

I thoguth about it and realised that it is averaging in time rather than spanwise averaging.

Thanks anyway.

Re: spanwise averaging for LES backward facing ste
 

So, my results were about 7.2, using a FORTRAN code in 2D dimension. I used the finite element method, using triangle elements, Galerking scheme with LES and structure function sub-grid scale model. I did an average in time for each x location. You said: "...to average in spanwise direction rather than just process the data at mid span...", I didn´t understand you, sorry, my english is bad. What does "spanwise" mean?

Re: spanwise averaging for LES backward facing ste
 

I did a three-dimensional calculation with u(y) inlet profile.

To explain in more detail, my backward facing step is generated in (x,y) coordinate but it was extended in z (spanwise) direction with periodic condition in both z boundaries.

I wanted to average in z direction to get a data similar to two-dimensional calculation.

Since I don't know how to do the spanwise average in a simple way, I currently process data at mid z plane only.

Regards, Sungho

Re: spanwise averaging for LES backward facing ste
 

I understood now your problem! u have to average the componentes of your velocities in 3 directions, by the projection of the area of the element in each plane. The researches in Federal University of Uberlandia (www.ufu.br), in Brazil, probably did this type of simulation. Look for Professor Dr. Ellie Padilla or Professor Dr. Aristeu da Silveira Neto.