Let's suppose:

nu = eddy viscosity, and nu,x = d(nu)/dx, where x could be in any of the three directions.

The quantity I'm interested in is the maximum value of nu,x/sqrt(nu)

If you have LES (or DNS) data of various flow topologies I'd really appreciate some numbers for the above quantity.

Alternatively, if you know how/where I can access LES data (so that I can evaluate this quantity myself), I'd appreciate names/addresses

Thanks

Adrin Gharakhani

hi there,

Just by checking the ressources available on cfd-online.com, i'm sure that you can find a link to some LES database. Cheeck: http://www.cfd-online.com/Resources/refs.html#data

Sincerely,

Frederic Felten

(1). Looks like that you are looking for the eddy viscosity distribution. (2). In Hinze's book of "Turbulence", there are eddy viscosity distributions in the boundary layer. From the charts, you can compute the number you are looking for. (3). Eddy viscosity disappears at the wall. It can also go to zero at the edge of the boundary layer, depending on the model.

Take a look at the ERCOFTAC databases. They have plenty of DNS and LES data for various types of flows.

Srinivasan

Yes I am looking for the eddy viscosity (EV) distribution. Specifically I'd like to know the order of magnitude of the ratio of the gradient of EV to Sqrt(EV). If we multiply this term to sqrt(timestep) the new term looks something like a local grid Reynolds/Peclet number in an explicit time integration scheme (if we think of the gradient term as the velocity)

So the former pretty much determines how large a timestep one can take and still get a positive solution! I need this info to determine if it makes at all sense to takes this route.

BTW, I've checked the database recommended by others. I've found only three sets of data (for the same problem) with the required info, and the maximum value varied from order 0.1 to 20, depending on the assumptions.

Other results do not have the eddy viscosity info and I hesitate to make any assumptions using just the published velocity fields.

Thanks

Adrin Gharakhani