ABSTRACT
In Chapter 6, we have shown how the SIBC concept can be coupled with boundary integral equations to form an efficient method of solution for low-frequency applications. The concepts developed are now extended to include high-frequency applications. We start, by reintroducing the displacement currents in Maxwell’s equations. This means that the basic derivations in Chapter 3, including the surface impedance functions must be generalized to take into account the displacement currents in Maxwell’s equations. Using the vector Green’s function we first develop the electric field integral equation (EFIE) and the magnetic field integral equation (MFIE) representation of the fields based on the incident electric and magnetic fields. These equations are then coupled with the SIBCs previously developed to implement formulations of various orders of approximations in the time and frequency domains in a manner analogous to that in Chapter 6. The shift from low-frequency to high-frequency carries both advantages
and penalties. Of course, in conductors, the skin depths tend to be thinner, resulting in more accurate representation of fields. In that sense, one of the main differences is that at high frequencies, it is also possible to treat lossy dielectrics-that is, the surface impedance may now be applied at the interfaces between lossy dielectrics and dielectrics (such as free space). On the other hand, the introduction of displacement currents in Maxwell’s equations results in negation of one of the most important of the advantages we found in the low-frequency transient representation-that of separation of variables. We saw in Chapter 6 that the variables in the surface impedance function given in Equations 3.82 through 3.84 admitted separation into time and space variables only by neglecting displacement currents and treating each term of Equations 3.82 through 3.84 independently. In the high-frequency case we cannot do that and the time convolution integrals cannot be precomputed and tabulated in advance. Yet, direct computation of the convolution integral in the time-domain integral equations is impractical due to the large computational time and storage requirements. To mitigate this difficulty we also include a discussion of methods for efficient
implementation of the time convolution integrals in the formulations. More specifically, we show in Appendix 7.A.1 how Prony’s method may be incorporated to approximate the impedance by decaying exponentials and by doing so we arrive at a simple recursive evaluation of the convolution integrals. As in previous chapters we use an example to illustrate the theory developed. In this case, the problem of scattering from a cylinder is used as an example.
