Answers are grounded in this paper and cite their sources.
For a few particular waveguide problems, standard variational expressions have previously been shown to be upper or lower bounds on the quantities of interest. However, bounds have not previously been obtained for any truly three-dimensional problem; that is, where the fields cannot be derived from a single scalar potential. An example is a three-dimensional obstacle which contacts only one waveguide surface. As one consequence, no straightforward procedure exists for improving the approximations. Recently, Kato devised a rather general method for bounding the cotangent of the phase shift for a given angular momentum in a quantum mechanical central potential scattering problem. This method should be applicable whenever a system can be analyzed in terms of uncoupled standing waves each characterizable by one real phase shift. The method is here adapted to waveguides, including truly three-dimensional problems. The obstacle must be symmetric about a plane perpendicular to the waveguide axis, with certain exceptions only one mode should propagate, and the system should be lossless. Analysis is then possible in terms of ηe and η0, the real uncoupled phase shifts associated with the even and odd standing waves, respectively. The bounds obtained on cotηe and cotη0 determine bounds on the equivalent network elements.
No takes yet. Share an insight, caveat, or question.
Spruch et al. (1960) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: