When i plot my velocity profiles i get almost perfect results for my LES simulation. However, my rms values are very high compared to DNS results even though the shape of my graphs is almost similar to DNS graphs. Any idea what is going on? Thanks in advance.

Usman

you did not explain about your problem..

Hi Pankaj

I am simulating a flow on duct with four ribs inside. I have run my case using periodic boundary conditions, everything went fine. When i start collecting results at the locations specified in paper. My mean velocity profiles look very similar to DNS and experimental profiles, however when i got my rms values they were a bit too high compared to DNS values. I am thinking it might be because my rms values have not yet reached steady state. This is because when i plotted my rms values as function of time they were still changing. Any idea what can be cause of high rms values.

Thanks Usman

Which paper you are matching?

"LES and RANS for turbulent flow over arrays of wall-mounted obstacles" by Zhengtong Xie. Ian P. Castro

Usman

One more thing Pankaj, if you look at the paper can you tell how would you normalise mean velocity and rms values. In other words what would be your u* value. Thanks

Usman

i have already sent you a personal mail to your mail id. please read it. so, we can discuss more on this topic. anyway.... 1. are you running dimensional code? 2.then how you are tackling the mass flow rate or desired Re, for a given pressure gradient.?

Hi pankaj

I e-mailed you back providing you details regarding my problem.

Usman

> Any idea what is going on?

These are LES simulations? If so, somewhat counter intuitively, one can often get too large and too anistropic turbulence when using coarse grids. Try increasing the resolution and seeing what happens. Of course, without being familiar with the details of what you are doing, I would suggest a pretty low confidence level in this sort of diagnosis.

Hi Andy

I am using same grid resolution as they have in paper so it should not be a problem. My supervisor thinks i have to run it out longer. So i am letting it run for now. Thanks for reply.

Usman

Is this an inflow/outflow problem. If so, how do you prescribe velocity at the inlet. Is there turbulence at inflow?

I used periodic boundary conditions for inlet and outlet to develop turbulence.

Usman

Well, hope that running your simulation for long time would bring down the rms values! Keep us posted.