Affiliations: [a] School of Mathematics and Information Technology, Jiangsu Second Normal University, Nanjing, Jiangsu, China | [b] Jiangsu Regulatory Bureau of Nuclear and Radiation Safety, Nanjing, Jiangsu, China | [c] National Key Laboratory on Electromagnetic Environmental Effects and Electro-optical Engineering, PLA University of Science and Technology, Nanjing, Jiangsu, China
Correspondence:
[*]
Corresponding author: Feng Lu, National Key Laboratory on Electromagnetic Environmental Effects and Electro-optical Engineering, PLA University of Science and Technology, Nanjing 210007, Jiangsu, China. E-mail:fengluday@njmu.edu.cn
Abstract: A VIE-based domain decomposition method (DDM) is proposed
for analyzing EM scattering from inhomogeneous electrically large dielectric
objects. The volume integral equation (VIE) still uses tetrahedra to model
the entire body and uses the SWG basis functions to expand the equivalent
electric flux density. This new DDM is established by dividing the unknowns
on the whole electrically large body into groups, serving as subdomains.
Through necessary symmetry treatment of standard MoM impedance matrix, the
DDM using subdomain-decoupling technology can be combined with the VIE model
to reduce memory requirement. Actually, this is an algebraic DDM, not a
geometric DDM. In other words, it has no requirement of physical location of
basis functions belonging to the same subdomain. This decoupling procedure
is completely eliminating the coupling impact of the primary subdomain with
the rest of the dielectric body, until every subdomain is independent with
each other. In this work, when solving ultimate decoupled impedance
subdomain matrix, the LU decomposition process for solving interpolating
coefficients of multiple right sides is accelerated by GPU parallel
technology to significantly decrease CPU time. In brief, this paper first
time combines the algebraic DDM with the conventional VIE model (including
both isotropy and anisotropy VIE model) to significantly decrease the
requirement of memory. At last, a few representative numerical examples are
provided to demonstrate validity, efficiency and stability of the new
method.
Keywords: MoM, domain decomposition, volume integral equation
