another question is on NRBC: my comp. domain is meshed also outside the pipe = suitable far-field boundary is to be set at the far left, far-north and on a piece of upper west boundary. First, I simply put there outlet condition - that lead to nice wave reflections. Then I modified it to more proper, setting the state in external ghost cell to be the same as in the last cells of the domain, that would, imho, cause zero flux accross the boundary (at least for planar wave). result are better, but still reflecting the waves. Now I'm going through maaaaaany papers on nrbc, but: I want something really simple (no buffer zone with increased dose of artificial viscosity or modified equations). How good do you think is simple extrapolation technique? I am intending to extrapolate the state in external ghost cells from the state values in last two interior cells, enough? more? linear extrapolation? Another idea that come to my mind is to set such a state in the ghost cells that would cause a flux of the same magnitude but opposite direction (I mean ghost cell would "suck" just as the last domain cell "spits") actually, any advice is welcome

Hi queram,

just as a hint and just in case you aren't aware: There are two good references on non-reflecting boundary conditions to start with: Chapter 19 in "Numerical Computation of Internal and External Flows" Vol. 2 by C. Hirsch

And: "Numerical and Theoretical Combustion" by Poinsot and Veynante

There is also a famous paper on BC written by Lele and Poinsot: Poinsot, T. & Lele, S. 1992 "Boundary conditions for direct simulations of compressible viscous flows." J. Comp.Phys. 101, 104-129

(You probably know all these.)

I am currently involved in implementing inlfow BC for incoming boundary layers and I was looking for someone to discuss with (see my post in this forum below). Maybe we can solve our both problems togther? Are you interested?

Alan