Hello Everyone,

I am running two very fundamental validation studies: (i) laminar boundary layer flow development and (ii) turbulent boundary layer flow development. Both simulations start with a uniform profile and simulate the development of a boundary layer along a no-slip wall at the bottom of the domain.

All simulations converge easily and produce expected results on a qualitative level. However, when comparing the velocity profiles of analytical approximations (Karman for laminar and Prandtl's 1/7 law for turbulent) the CFX solutions differ very significantly from the analytical solutions. I am convinced that this is due to an error on my part and am having trouble identifying the source of my error. Thus, I would greatly appreciate a critical review of my modeling approach and any tips that could be shared.

Here are more of the simulation settings: (a) Domain and Mesh - Domain is a rectangle with wall at bottom surface so a structured, hexahedral mesh is used with more elements near the wall than in the top region (high element density near wall) (b) Boundary conditions - right and upper boundaries are openings with zero ref. pressure (also tried top boundary being a free-slip wall) (c) Convergence Criteria - Max residuals below 1e-4 (d) Fluid - air w/ constant density, w/ ref. press of 1 atm (e) Turbulence model - SST / k-omega , low intensity at inlet and zero gradient at rightmost boundary

I would also like to know if any fellow CFX users have had similar experiences with this problem. Thank-you in advance for your time.

Have you checked the first cell height that you used in the mesh? This value is very important for capturing turbulent boundary layer. Make sure you using an appropriate Yplus value. Hope this help

Cheers