Hi,

I am trying to model a turbulent jet using STAR CD. Initially I have used the high Re number turbulence model and this has given reasonably good results, but not good enough. As one might expect, the spreading rate is too high, and I want to implement the Launder, Morse, Rodi and Spalding model which varies the constants C_{mu} and C_{epsilon 3} as:

C_{mu} = 0.09 - 0.04f

C_{epsilon 3} = 1.92 - 0.0667f

where f = | (R_{mu}/1.6u_{0} |du_{cl}/dx - |du_{cl}/dx|||

R_{mu} is the radius of the mixing region u_{cl} is the centreline velocity u_{0} is the jet velocity.

More details are in

Launder, B.E. and Spalding, D.B., " The numerical computation of turbulent flows", Computer Methods in Applied Mechanics and Engineering 3 (1974) pp 269-289

available online.

I need to use this model rather than a more up to date one to compare my results with those of my supervisor. How can I do this using STAR CD? I have seen that you can make the Prandtl number variable, and also you can write a user routine to supply turbulent viscosity, but is there any way so that one can write and use one's own turbulence model, or simply let the constants vary?

Thanks in advance for your responses,

Ant

It's possible ...

For C_mu, you could use the vistur user subroutine. The value of vist which is passed into this subroutine will be the one using the standard C_mu, so all you would need to do would be to multiply this by C_mu_new/C_mu.

For C_eps3 you would need to re-implement the source term through sorkep. The sample coding in that subroutine should give you an idea how to do it.

If you haven't thought of this already: make sure you test your subroutines by implementing the standard k-epsilon model through them first.

Thanks for your help. I think I'll be able to do it now. I'll presumably have to use POSDAT in order to get the velocity data required to calculate f, and then use VISTUR to deal with the C_mu term and SORKEP to deal with the rest.

Seems like a lot of work for something which is conceptually straightforward, do you think this method is overcomplicated?

Thanks again for your help and suggestions,

Ant

One thing which you could do to simplify the coding would be to contact star support and ask them for the common block which contains the velocity field. Then you could calculate f locally (and separately) in vistur and sorkep, without relying on posdat. That's probably about as simples as it could be.