Validation of surface roughness model approaching a hydraulically smooth surface
 

We have used the surface roughness model in CFX previously on transitionally rough surfaces with fairly good results.

Now we are studying a case where the surface is just on the limit between hydraulically smooth and transitionally rough. Has anyone done any validations of the surface roughness model in CFX for this type of almost smooth surfaces?

I'm responding to myself here. We have digged out some measurement data for a case with two different surface roughness values - one is almost hydraulically smooth and one is a bit rougher and transitionally rough most of the time.

Although the model does not predict the correct absolute losses the loss-trend when changing surface roughness seems to be fairly well captured with this model. We are running it with resolved boundary layers. The loss trend is what we were most interested in. We want to assess if it is worth the extra cost to polish theses surfaces to mirrors or if we should devote our effort to other problems instead (welds, tolerances, ...).

The fact that the absolute numbers are not matching is probably due to problems of comparing apples to apples - measuring loss in a boundary layer is difficult and comparing measured loss with CFD results in a consistent manner is also difficult.