| Spherical shells in PV-Wave [message #1692] |
Thu, 10 February 1994 15:02 |
todd
Messages: 25
Registered: February 1994
|
Junior Member |
|
|
My research involves solving fluid flow problems in spherical shells where
I calculate temperature, pressure and the three velocity components all in
spherical co-ordinates. I use a regular (but not necessarily uniform) grid such
that any field variable, F(i,j,k) has the co-ordinates r(i), theta(j), phi(k)
{radius, colatitude and longitude, respectively}.
My problem is that the GRID_4D and FAST_GRID4 procedures don't seem
to do a good job. They "bleed" information across the boundaries of the
spherical shell such that the resulting uniformly gridded array contains
data values far outside of the actual data domain. Part of the problem,
I think, is that the resulting array, since it is cartesian, has to have spatial
locations that simply do not exist in the original data set and which should
in reality have a zero value.
Is there some aspect of the regridding procedure that I can modify
to get better results? I have tried tried reducing the neighborhood and
number of smoothing iterations, but to no avail.
Any suggestions that you might have would be greatly appreciated.
Thanks,
Todd Ratcliff
todd@artemis.ess.ucla.edu
UCLA Dept. of
Earth & Space Sciences
(310)825-3118
Keywords:
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
