Page 1 of 1

Fixed boundary conditions - Repost from old forum

PostPosted: Sun Nov 30, 2008 1:32 am
by malonso
Hi All,

I wish to impliment a idealised simulation with the values on one boundary held constant. Eventually, I am hoping to impliment a new boundary condition option to do this.

Has anybody had any experience working with the boundary conditions in the idealised simulations, who has done this already?

Thanks,
Sarah

Re: Fixed boundary conditions - Repost from old forum

PostPosted: Sun Nov 30, 2008 1:32 am
by malonso
Hi Sarah,

I would also like to set up a constant inflow condition in an idealised case.

At the moment, I'm achieving this with a pretty horrible hack -- I've added code to set the various *tend arrays to zero near the end of the subroutine "rk_tendency" in dyn_em/module_em.F

Have you made any progress towards implementing your own constant BC, or have otherwise discovered a more elegant way of doing this yourself?

Cheers, Alan.

Re: Fixed boundary conditions - Repost from old forum

PostPosted: Sun Nov 30, 2008 1:33 am
by malonso
Hi Alan,

Your email has excellent timing! I have recently returned to the problem, as a number of people have expressed interest in fixing the inflow at the boundary.

I ended up modifying my experiments so that the constant inflow was 25m/s. At this speed, the open boundary conditions permit no waves to radiate upstream (this is detailed in Chapter 6 of the tech notes - Lateral Boundary Conditions, and implemented in module_advection).

I did attempt a similar hack to you when I first posted this question, but found that it quickly generated instabilities. I cannot however be sure this was not just user error - I was quite green with the code then. I was planning on returning to the problem later this week. I'll keep you posted!

Cheers, Sarah

Re: Fixed boundary conditions - Repost from old forum

PostPosted: Sun Nov 30, 2008 1:33 am
by malonso
Hello,
I'm also trying to implement fixed boundary conditions in idealized cases.
I've modified the module advect_em.F to have a fixed inflow wind speed on one or more boundaries. The problem that I've experienced is a mass drift after some integtration time, so that pressure and temperature are decreasing more or less linearly with time and eventually the simulation crashes getting CFL > 1.
Does some of you have the same problem and/or hints on how to deal with that?
Any suggestion will be appreciated,
francesco

Re: Fixed boundary conditions - Repost from old forum

PostPosted: Thu Dec 15, 2011 2:33 am
by tomk
Has anyone found a good way to do this?

Perhaps there is a way to use WRF's own "specified" option for the inflow, and an "open" boundary for the outflow?

Re: Fixed boundary conditions - Repost from old forum

PostPosted: Mon May 22, 2017 4:52 am
by engrossment
could you offer some ideas about implementing inflow boundary condition. really appreciate it :cry:
why can't just modify the module_bc.F to make a constant boundary condition?