OK, so it's aluminum, and I'm talking about antenna models, and, well,
I could make the claim that the roof is "hot" with RF because of
induced currents... but... the question:

I've got a mast that can get my VHF/UHF antennas up above the roof of
my apartment building about four feet, above ground about 28 feet. My
balcony has a ceiling made of aluminum, so I've got this big metal
sheet (about 16 feet x 8 feet) close by under the antennas. I want to
use EZNEC to model the effect of the whole system, antenna above ground
and at the corner of this big aluminum sheet.

I figure a grid is OK if the mesh is fairly fine (what, maybe .1
wavelength?) but the problem is that I'm going to run out of segments.
Even if I upgrade to EZNEC+ and get 1000 more segments, I may not have
enough to do a 16 foot by 8 foot mesh. I've tried a close spaced set
of parallel wires and ignored the segmentation warnings and I see some
reasonably sensible pattern distortion, but I wonder if anyone could
point me to some tricks to make a relatively coarse grid of wires seem
like a plane. Will fat wires help?

73,
Dan
N3OX

Hello Dan, if you don't mind some additional studying, depending on
on-board RAM memory or if you are very patience, you could try 4nec2.
This will enable you to run up to 11000 segments without further costs.
To automatically create a wire grid (using the equal-area rule) or even
using surface-patches, use the geometry-builder.