In some papers I found that they randomly sample fluctuating components of velocities from the Gaussian distribution based on the r.m.s value of turbulent intensities. But how exactly they do it?

i could not understand what u've tried to convey.can u be more clear and specific.

senthil k

Okay, from numerical simulation we get only mean components of velocities. If we want to approximate the fluctuating velocities, we need to do it based on the auto correlation (i.e., turbulence intensities) of the turbulent flow. As a matter of fact these square root of turbulence intensities are the RMS values of the fluctuating components. So some authors sample these fluctuating velocities from a Gaussian distribution with mean = zero and standard deviation = square root of turbulence intensities. I came to knew about the sampling method later after posting the message. I was wondering why so far nobody answered to this message. However if I'm wrong or you have anything more to say please do continue.