Hi,

I am doing a LES for a surface mounted cube inside a channel and would like to know how I determine the starting point of averaging my results. Right now, I take a look at the drag coefficient and if it looks pretty 'turbulent' I would start averaging; but shouldn't be there a more mathematical way to determine the start point?

Regards! fab

Hi,

One way is to low pass the unsteady drag coefficient signal. If the filtered data are (almost) steady or periodic, you probably can initiate the averaging process.

Another possibility is to start immediately the process but then split your averaged data into several files of equal duration. At the end of the runs, you will be able to compare the first order statistics coming from each files to see if the earlier files differ from the latter ones. The averaging over all the files whose statistics closely match the last ones should result in converged first and second order statistical data.

Hope this helps,

Lionel

Hi Lionel,

thanks!

Regards fab