Waveguides

Structures consisting of an interconnected set of waveguides are easily designed and modeled with Analyst.  Microwave feeds are defined in terms of one or more ports, and a user-defined number of propagating modes at each port can be included in the analysis (over 150 modes have been used in the analysis of highly over-moded waveguides).

Both discrete frequency and fast-frequency sweeps can be used to analyze the broadband performance of a structure.  The discrete frequency sweep explicitly solves the finite-element matrix equation for each frequency in the band, and can optionally be used to obtain the field distributions in the structure for each port/mode at each frequency.

When only S-matrices are required, the Galerkin Asymptotic Wave Expansion method can be used to obtain many hundreds or thousands of frequency points in a small fraction of the time required for an equivalent discrete frequency sweep.

A variety of boundary conditions are available for use with driven frequency problems, including perfect electric conductor (PEC), perfect magnetic conductor (PMC), imperfect conductor (impedance), and “open” radiation.  There are also three types of symmetry planes, including tangential electric, tangential magnetic, and a special type called “even-odd” which is used when the geometry is symmetric but the fields are not.

Various types of media can be modeled, including simple scalar and general tensor media.  Scalar media can be characterized by either complex constituitive relations or real dielectric/permeabilities and, optionally, corresponding loss tangents.  Complex matrix elements are allowed in the specification of tensor media.

As with other Analyst solvers, the driven-frequency solver (called RF3P) can make efficient use of multiprocessor computers, networked computers, computer clusters, and even some massively parallel computers.  Highly specialized parallel algorithms make optimal use of multiple processors.  Domain decomposition, i.e., mesh partitioning, is used to distribute the computational work for problems that are too big to fit on a single processor.  The solver will also subdivide the spectrum spread across multiple processors when it is advantageous.  Both domain and spectrum decomposition are done automatically by the solver, with the appropriate combination of the two approaches chosen based upon the available computing resource and the problem characteristics.