4/8/16
You are invited to attend a lecture
by
Balanced Electric-Magnetic Absorber Green’s Function Method for MoM Matrix Thinning |
Naor Shay
MS.c. candidate of
Professor Raphael Kastner of Electrical Engineering, Physical Electronics Department
A new numerical approach to solving the classic problem of electromagnetic (EM) scattering off a perfect electric object is studied with the objective of substantially reducing computation times. The method considered here is the frequency domain Method of Moments (MoM) formulation involving the use of a dyadic Green’s function (GF). Traditionally, this GF is formulated in free space, as afforded by the equivalence principle. However, since the resultant equivalent sources generate a null filed inside the scatterer volume, the door is open for the inclusion of arbitrary fillers therein.
We suggest the usage of balanced absorbers as fillers and using their Green’s function instead of the free space one. To this end, the solution of the essentially volumetric problem of the absorber is required as a preprocessing stage. Balanced absorbers have both electric and magnetic Weston-like or Perfect Matched Layers (PML) loss mechanisms. Many interactions between pairs of base functions are then virtually eliminated. As a result, the MoM matrix, representing the GF, is significantly thinned.
The cost of calibrating this modified GF using the volumetric representation of the absorber is investigated. The effort incurred in the pre-processing stage can be alleviated by choosing absorber configurations that apply to many problems with high degrees of symmetry, thin absorbing shells rather than volumetric scatterers or homogeneous absorbers that lend themselves to surface formulations.
It is shown that this form of thinning has little effect on the accuracy. Moreover, most of the thinned elements need not be computed at all.
Finally, we demonstrate how this technique can solve some oscillation phenomena related to the EFIE in EM scattering problems.
Thursday, August 4th, 2016, at 15:00
Room 011, Kitot Building
You are invited to attend a lecture
by
Balanced Electric-Magnetic Absorber Green’s Function Method for MoM Matrix Thinning |
Naor Shay
MS.c. candidate of
Professor Raphael Kastner of Electrical Engineering, Physical Electronics Department
A new numerical approach to solving the classic problem of electromagnetic (EM) scattering off a perfect electric object is studied with the objective of substantially reducing computation times. The method considered here is the frequency domain Method of Moments (MoM) formulation involving the use of a dyadic Green’s function (GF). Traditionally, this GF is formulated in free space, as afforded by the equivalence principle. However, since the resultant equivalent sources generate a null filed inside the scatterer volume, the door is open for the inclusion of arbitrary fillers therein.
We suggest the usage of balanced absorbers as fillers and using their Green’s function instead of the free space one. To this end, the solution of the essentially volumetric problem of the absorber is required as a preprocessing stage. Balanced absorbers have both electric and magnetic Weston-like or Perfect Matched Layers (PML) loss mechanisms. Many interactions between pairs of base functions are then virtually eliminated. As a result, the MoM matrix, representing the GF, is significantly thinned.
The cost of calibrating this modified GF using the volumetric representation of the absorber is investigated. The effort incurred in the pre-processing stage can be alleviated by choosing absorber configurations that apply to many problems with high degrees of symmetry, thin absorbing shells rather than volumetric scatterers or homogeneous absorbers that lend themselves to surface formulations.
It is shown that this form of thinning has little effect on the accuracy. Moreover, most of the thinned elements need not be computed at all.
Finally, we demonstrate how this technique can solve some oscillation phenomena related to the EFIE in EM scattering problems.
Thursday, August 4th, 2016, at 15:00
Room 011, Kitot Building