Hi, everyone..

I'm making the LES code by modified laminar NS code.

First, I have been defined the outlet bc as du/dn=0. I think it was wrong.. So, I found the outlet bc in the paper. They used the convective bc. du/dt+U*du/dx=0, U is the free stream velocity.

does anyone know how to coding this convective bc? Do I treat this bc as 1-d for each direction?

If you know how to coding this one, please send or let me know the code of bc.

Thank you everyone

Hi Jay,

The LES treatment of outlet bc aims to carry waves away from the computational domain. If you are solving compressible equations, waves are acoustic and convective; if you are doing a low-Mach number approximation, you need only deal with convective waves.

If you are solving compressible equations, have a look for paper dealing with non-reflecting boundary conditions such as Poinsot (for reactives flows). The principle is to decompose you system into incoming and outgoing waves and treat them depending on the bc type. Note: this treatment applies also to wall/inlet/outlet...

If you are in the latter case, the convective bc you mentionned is enough. How to code is very simple you determine (1) U at your boundary (U is local, not averaged along the boundary) and (2) du/dx at your boundary where x is the direction normal to the boundary (local as well). (3) This is your source term to integrate. I did it for a finite difference code where you solve the velocity at the boundary, but I am not familiar on how to implement it in a finite volume code.

The validation is usually performed by initialising waves in your domain (acoustic, vortex, enthalpy...) and see how they propagate through the boundary condition. The important factor is the ratio initial energy/reflected energy.

Hoppe this help, Julien

Julien, can you quote some reference papers?

Hi.. jdc

every time I thank you for ur help.. In my case, I solve the incompressible flow case. if you don't mind to send a part of your code, Would you send it to me by e-mail?

I want to look at the code of outlet bc of yours..

Thanks

Hi Centaur,

Here are some reference to papers (on the subject of non-reflective boundary conditions for compressible flow):

(1) Thompson, K. W. (1987). Time dependent boundary conditions for hyperbolic systems. J. Comput. Phys., 68:1-24

(2) Thompson, K. W. (1990). Time dependent boundary conditions for hyperbolic systems II. J. Comput. Phys., 89:439-461.

(3) Poinsot, T. and Lele, S. K (1992). Boundary conditions for direct simulations of compressible viscous flows. J. Comput. Phys., 101:104-129

(4) Pakdee, W. and Mahalingam, S. (2003), An accurate method to implement boundary conditions for reacting flows based on characteristic wave analysis. Combustion Theory and Modelling, 7:705:729

I don't have any of them on electronic form. For papers (1-3), I will have problems to get a paper version since they are lost in the middle of my documentation.

For the incompressible approximation, I am affraid it is so trivial that no-one bothered to write a paper about it. In the worst case (and if you read french), I could send a copy of my PhD dissertation.

Good reading. Julien

Hi Jay,

The treatment in my code is so simple that I would be ashamed to send it (as well as I don't have the right to do so). There are two possibilities:

1) I can send you the part of my phd related to the boundary condition. Note: it is in french.

2) you can send me your code (or the part that is computing the convection fluxes) and I could send you some ideas on how to implement.

Sincerely, julien

I can send a part of my code to you. Can I know your e-mail address? send me your email address to me.

I will send a subroutine of bc condition for each velocity components. you may figure it out without any problem.. the code was written by fortran and used by FVM

Thanks Julien, i have #3 which is a really good paper, i think i can get my hands on to #1 and #2.