HI everyone, I have a small understanding of LES simulations and have a very basic question. What happens if you use LES with a coarse mesh. In an ideal world when using a Ke turbulence model you would want to model down to y+ values of approx 11 to 50. But often y+ values (at least ina comercial field) very rarely get down to these levels. So what would happen if LES were used at mesh sizes above the Kolmogarov length scales (assuming that it is correct to try and model down to that level?)? Dave

Hi,

The simple answer is the larger the cell you use the more important the Subgrid model becomes. In LES the scales above the filter size (related to your cell size) are directly simulated and the scales below the filter size are approximated by the subgrid model. When the filter size = the kolmogorov length scale you do not require a subgrid model (This is DNS). The use of a large cell also means that you do not resolve important phenomena at the wall. In general I would advise running a simulation on a coasre mesh to check if everything is o.k. If you do this you will get an idea of the accuracy. If you need to improve the accuracy put more cells in (or change the subgrid model - or both)

Good luck.

Tim.

Cheers Tim, that has cleared up a few questions, no doubt there will be more as we progress, Dave