Re: NEC-LIST:nec-2 arbitrary shape patches

Arie,

The integration of current over a patch in NEC is evaluated as just the current
at the patch center times the patch area, so all that matters to the
code is the center point, area and normal vector. You can get away
with that representation for the Magnetic Field Integral Equation
(MFIE), but not for the EFIE. The EFIE, which can model open plates,
is usually solved in terms of potentials. Also, continuity of
current is usually enforced, which requires keeping track of the
shape and neighboring patches. A good way of solving the EFIE was
not published (by Rao, Wilton and Glisson) until after the basic NEC
code had been developed.

 From the center point, normal and area in NEC you could draw square
patches with an arbitrary orientation, but it might not make a nice
picture. The NECPLOT program that we include with the codes plots
patches if they are entered as a shape: triangle, rectangle,
quadrilateral. It gets the shapes from reading either the input file
or output file. From an output file it will also plot current
vectors as arrows at the centers of the patches and oriented in the
plane of the patches. The cluster of current arrows at least gives
an idea of the surface shape when the patches cannot be drawn.

I almost never use the MFIE patch model in NEC. The people at the
NAVY (SPAWAR)
compared NEC patch and wire grid models of ships a long time ago. I
think they decided that the MFIE patch model often was not as
accurate as wire grid models, but was faster in filling the matrix
and was accurate enough for their needs in determining the
approximate number and depth of nulls in radiation patterns. They
would go on to make scale model measurements for more accuracy.

Jerry Burke
LLNL

-- 
The NEC-List mailing list <nec-list_at_gweep.ca>
http://www.gweep.ca/mailman/listinfo.cgi/nec-list