A comparison of different testing procedures for subdomain moment method formulations for arbitrarily shaped bodies, including attached wires, is presented. Conventional evaluation of the testing procedure both for linear path and Galerkin formulations is shown to be worse than an accurate calculation of the Galerkin type integral, via a Gaussian quadrature scheme. Otherwise, the relevance of an adaptively generated mesh-more refined nearby the attachment-is shown to reduce the number of required unknowns for any of the above approaches.
Published in:
Antennas and Propagation Society International Symposium, 2000. IEEE
(Volume:4
)
Date of Conference: 16-21 July 2000