double refinement with diferent refinement factors

Report or discuss software problems and other woes

Moderators: arango, robertson

Post Reply
Message
Author
mma
Posts: 4
Joined: Fri Jun 06, 2003 4:55 pm
Location: TAMU

double refinement with diferent refinement factors

#1 Unread post by mma »

Hello

I have 3 nested domains with refinement factors 3,3 and all looks fine (3 from grid 0 to grid 1; and 3 from grid 1 to grid 2). When I change the refinement factors to 3,5, I get problems with the free surface of the intermediate domain.

Here is the zeta initial (left) and after one day (right) of simulation using 3,3:

Image

The same using 3,5:

Image

Looking at zeta of the intermediate domain, we can see the problems at the boundary (left plot). Note that in both cases, the refinement of the intermediate domain is 3, only the refinement of the 3rd (higher resolution) domain changes from 3 to 5.

Image

Continuing the 3,5 simulation (about 70 days), the free surface of the intermediate domain continues increasing to insane values. The next image plots some point in the middle of the intermediate domain vs some point inside the coarser domain:

Image

The 1 day simulations (ie, the first three figures) have no tides, no river and no atmospheric forcing. The dt was uses in accordance with the refinement factors.

Is this a bug or is this a feature? Can't we use different refinement factors in multiple refinement nesting configurations?

Thank you
mma

User avatar
wilkin
Posts: 875
Joined: Mon Apr 28, 2003 5:44 pm
Location: Rutgers University
Contact:

Re: double refinement with diferent refinement factors

#2 Unread post by wilkin »

In principle you can use different ratios in the different grids, but I will confess I don't know that we ever tested that.

Quick question - is this 2-way or 1-way nesting? If it's not 2-way, please give that a try. We did notice some odd volume conservation issues in earlier versions of the code, but I thought they were solved and there was no suggestion the nesting ratio made a difference anyway,

I have to ask - have you made quite sure your grid metrics pm,pn and the depth are sensible (smooth and continuous) at the perimeter transitions for the factor of 5 grid?
John Wilkin: DMCS Rutgers University
71 Dudley Rd, New Brunswick, NJ 08901-8521, USA. ph: 609-630-0559 jwilkin@rutgers.edu

mma
Posts: 4
Joined: Fri Jun 06, 2003 4:55 pm
Location: TAMU

Re: double refinement with diferent refinement factors

#3 Unread post by mma »

I used 2-way nesting as I only activated the NESTING option.

The second (intermediate) grid has a bathymetry smoothly converging to the bathymetry of the first grid (coarser). The third grid for both factors 3 and 5 was extracted from the intermediate one using coarse2fine.m without any modification (of h, mask, ...).

Anyway, I get problems at the intermediate grid boundary when I add a third grid with a refinement factor 5. So, there seems to be no problems with the grids and with the transition from the second to the third one. The problem is in the results of the second domain which uses the same grid in both experiments.

This really looks like a bug in the model or in the creation of the contact file.

Thanks for your reply
mma

Post Reply