Roughness effect
 

Hello Glen Gorrock,
I 'm following your answers regularly basis and believe that you are providing valuable information with logical analysis to CFD users. I have some problems and struggling with no results. I need your advice please.
I'm using Ansys 15.0 Fluent for the simulation of V-cone flow meter to estimate discharge coefficient. For validation of Fluent code,I have experimental data with rough pipe. In V-cone flow meter, cone element is installed co-axially at the center of a circular pipe with surface roughness height (k)1 mm (as the pipe was very old with rust). The length of the domain is 85Diameter(4.337m)and cone element is located at 60 Diameter(3.126m) length from inlet. I have produced boundary layer mesh on the pipe-wall and cone element surface and the first layer height is 0.06 mm with growth factor 1.2. In Fluent wall boundary,I have set 1 mm roughness height and roughness constant 0.5. After well converged solution (SST k-omega and RNG k-epsilon), I calculated discharge coefficient. But these values are very high(Cd=0.84 i.e. similar to smooth pipe Cd value) as compared to experiments data of rough pipe(Cd=0.71). It seems that surface roughness height is not effective in simulation. Differential pressure of upstream and downstream the cone is also low(about 65000Pa) as compared to experiment(about 92000Pa). Ansys 13.0 Fluent manual says the condition to capture the effect of roughness, first layer height of BL mesh should be double the the roughness height(k). But Ansys 15.0 Fluent says that it has newly defined y+ as y+(new)=y+(old) +(k/2) and all the previous conditions are superseded.But still I'm not getting the effect of roughness on co-efficient of discharge on pipe flow. I tried with 1.5 mm roughness height also but changes in results are insignificant. Can you please share your views related to this problem?
With regards.
Nasiruddin