sci.physics.computational.fluid-dynamics

On Jan 25, 12:07=A0am, Olin Perry Norton
wrote:
> SK wrote:
> > Hi everybody,
> > Can anybody provide me information to write a program for k-eps model
> > I am trying to write one code for simple pipe flow
> > I am using SIMPLE scheme .....i am finding the values of u, v and pr
> > iteratively and then implementing the results in k eps
> > equations .....I am not sure but i m getting some garbage values after
> > 1 time step
> > my perception is bcoz of boundary condition and inlet conditions can
> > anybody help me out in it....
> > Also I am using finite differencing schemes
> > Thanks in advance
> > Sachin
>
> Dear Sanchin,
>
> The implementation of boundary conditions depends
> on your grid. If you have grid points very close to
> the wall, then you can perhaps apply boundary conditions
> naively. Often, it is not possible to have such a fine
> grid, so boundary conditions at wall are computed using
> "wall functions", based on logarithmic law of the wall.
> Note you need wall boundary conditions for k and epsilon
> in addition to velocity.
>
> Another thing you must do is specify k and epsilon at the
> inflow boundary. This is usually done by assuming a certain
> percentage turbulence level to get k, and then deriving
> epsilon from k and a length scale, which is assumed to be some
> fraction of the pipe diameter.
>
> One possible source of your problem may be due to the smoothness
> of the initial values and inflow values you specify. For example,
> the source term for turbulent kinetic energy depends on the velocity
> gradient, and if the velocity profile you specify at the start is
> not smooth, this term would be very large, ans that might cause problems
> on the next iteration.
>
> I'm just guessing about your problem, if you give more
> information you may get more useful responses.
>
> Olin Perry Norton

Thanks Olin
Sorry for the delayed response but had not connected to net for this
loong
anywayz i m putting initial conditions as u=3D1.0 v=3Dw=3D0 and pressure as
1.0 k as 0.01 u**2 eps related to k ^1.5
Then about my inlet and outlet boundary condiotns I have kept those
values floating that is I have linearly extrapolated those values from
inner values except for inlet of u for which i have given modulus of
sine wave
For wall boundary conditions I have specified modulus o wall functions
by calculating uplus value and then finding u* and then evaluating k
and eps from it for 'u' i have aasumed no slip condtion
let me know if i need to clarify something more

Sachin Kanetkar