Hi, Guys,

Wall distance of each node is needed in some turbulent models. Who knows if the accuracy of wall distance is very important? How to compute the wall distances of the nodes in the wake of a profile or blade?

Thanks for your answers!

Stein

There will undoubtably be lots of suggestions. Here is just one of many.

If the model parameter is based on y+ (the non-dimensional wall distance) this could be replaced by the 'turbulent Reynolds number' Ret=k^2/(nu e) where k is the turbulent kinetic energy, e the dissipation and nu the viscosity.

Ret behaves differently from y+ so you will probably need to add a constant in front of it. Ret has the nice characteristic in that it depends on the local turbulence and not the geometric.

A good reference if you can get hold of it is 'One point closures applied to transition', Transition and Turbulence modelling, ERCOFTAC book, kluwer, 1996 (but there are lots more studies around). regards, Richard

(1). Each turbulence model will contain the complete information in regard to the wall boundary condition treatment. (2). In general, one must achieve mesh independent solution. Thus, there is no general answer to the mesh size requirement. (3). Approximate wall treatment (such as the wall functions) requires special guideline, which can only be obtained through experience with the particular class of problem being considered. (4). Specific formula sometimes are used in the in-house codes, but rarely available to general codes.

For many lower reynolds number turbulence models, the distance from a cell to a wall is needed (there are some models which use local turbulent Re as suggested by Richard). The general method is to calculate the distance of a cell to all walls and then take the minimum value of these distance - in this way, there is no problem if you have multi-walls.

[lehoanganh07]

Could we use wall damping function or turbulent reynold number for K-epsilon model with high Reynold number? I saw that this method is used for low Reynold number?