I am trying to add a subroutine to a program which is a NS code using CV scheme. This sunroutine applies a specific boundary condition; as an example, it is supposed to change the velocities, pressure, and energy at a crossection at a position in the middle of a simple pipe, ie x=L/2, so for this example, 2 blocks is defined for the code. Problem is this sunroutine makes the code totally unstable, no convergence. I know based on this general info it is hard to say anything, but if there something that might be any help, please let me know.

Steve

That's interesting that nobody made any comments on this!

Are you solving for a steady solution? Or are you solving for a transient impulse response to these boundary conditions? Is there an analytic solution to such an impulse-type problem (e.g. exponential decay, similarity)?

Thanks for the response. I am solving for a steady solution, and at the moment I am running the code for a very simple case which is a cylindar with 2 blocks (the code is Multi-block). I am applying this boundary condition right at the common patch between the blocks.

What do you think?

Steve