[SMM]

Hi all,

I am trying to model a simple laminar flow in star-cd inside a pipe of dia = 50mm and L=5m.Theory indicates that the flow will become fully developed at a length of 3.75m at Re= 1500. But I have several issues.

1) First of all for laminar flows usually the grid near the wall can be arbitrary refined to resolve viscous boundary layer, this may be achieved by using exponential meshing criteria (fill ratio in VFILL command), but in case of star-cd I have observed that if I use an arbitrary refined grid near the wall the error in computed friction factor increases, however if I follow the R/38 criteria of designing the grid near the wall (first wall adjacent cell height >= R/38 where R is pipe radius) the results get better (1.2% error in computed friction factor) but again if I follow the R/38 criteria and try to insert more cells in radial direction the error again starts to increase. This suggests that in star-cd for laminar flows in pipes we must find the optimal no of radial cells and should follow the R/38 criteria as well for first wall adjacent cell height. I would certainly like to know if some one has ever experienced the same thing. Any suggestions/opinions can be of great help.

2) secondly it is very clearly mentioned in the manuals that to model a fully developed flows in pipes/ducts one must impose a uniform pressure boundary condition at the pipe inlet and must specify the mass flow rate at the pipe exit. The mass flow rate can be explicitly defined by imposing OUTLET type of boundary condition at pipe exit (because I cannot find any other B.C type in star-cd that allows an explicit definition of mass flow rate). I simulated first the developing flow (mentioned above) and read the value of the pressure at the point where the flow becomes fully developed (0.4817Pa at 3.774m) also calculated the mass flow rate from the simulation of developing flow, I than used these values as B.C at pipe inlet and exit respectively but the code inspite of the clear expalnation in the manuals was not able to predict any laminar fully developed flow???. Any suggestions on this issue.

Thanking you guys in advance for your time and consideration