Problem comparing CFD results with experiment.
 

I am using STARCCM+ to simulate tubes with internally helical grooved (the ratio of tube length to pitch size is over 100 )with Re ranged between (4000-10000). Fluid model is solved with K-epsilon turbulent model in steady state. The mesh independence study have been done and the wall y+ is around 1. I calculated the Nu and F value for different Re which there are huge errors between experimental and CFD values especially in large Re. Since the groove is too small and generate swirl, I am thinking the realizable K-Epsilon two layer model ( I used all y+ wall treatment) cannot capture swirl effect inside the tube. Do you have any suggestion?

First of all, you have to be absolutely sure that you are comparing apples with apples and not something else. The Nu number can be automatically calculated in STAR-CCM+ (be careful with the inputs). You are comparing the experimental and numerical results at the same location? The boundary conditions are correct?

I think a two equation turbulence model is not suitable for such an application. Give it a try with V2F or an eliptic blending version of the k-epsilon.

Thanks Tim,
I tried to be as accurate as possible to match the boundary condition of simulation to study. Also, I am using the exact formula of the study to calculate Nu and friction factor. In Tubes with groove that has a larger pitch sizes, the results is acceptable but for smaller pitch the results is far off from the study. Again thanks, I 'll try with these two version to see if it is going to change.