Hi,

I have used the below written ccl inorder to use Rough wall treatment along with SST. Its a very simple pipe problem. Smooth wall with SST gives a certain value of Head loss, later when the roughness values are increased, there is no change in headloss, i.e. it remains constant for varying roughness apart from a smooth wall.

Re Number is of the order of 10e6. The average yplus is around 1.5 when viewed in POST along the wall, could this low y+ number be the reason for these strange results. Please let me know.

BOUNDARY:PIPEWALL

Boundary Type = WALL

Frame Type = Stationary

Location = PIPEWALL

BOUNDARY CONDITIONS:

WALL INFLUENCE ON FLOW:

Option = No Slip

TURBULENT WALL FUNCTIONS :

Option = Scalable

END

END

WALL ROUGHNESS:

Option = Rough Wall

Roughness Height = 0.00635 [mm]

END

END

END

FLOW: EXPERT PARAMETERS:

effective viscosity type = 2 END END

If your y+ wall distance is really 1.5, then this means you resolve the viscous sublayer. In theory, there is no effect of wall roughness if it is within the viscous sublayer. You need to use a grid with larger wall distance (e.g. y+>70). Your roughness of 0.00645mm seems really small. What is this in dimensionless form (k+)? What is your pipe diameter?

Yes the Y+ is very small. The pipe dia is around 2in. But if I increase the y+ to over 70 and use a SST model would the results be any better than a k-e model? I have results using a k-e model and would like to compare that with an SST model. My basic question would be which resolves roughness better SST or k-e?