An optimized hybrid Convolutional Perfectly Matched Layer for efficient absorption of electromagnetic waves

A hybrid technique is studied in order to improve the performance of Convolutional Perfectly Matched Layer (CPML) in the Finite Difference Time Domain (FDTD) medium. This technique combines the first order of Higdon's annihilation equation as Absorbing Boundary Condition (ABC) with CPML to vani...

Full description

Saved in:
Bibliographic Details
Published in:Journal of computational physics 2018-03, Vol.356, p.31-45
Main Authors: Darvish, Amirashkan, Zakeri, Bijan, Radkani, Nafiseh
Format: Article
Language:English
Subjects:
Absorbers
Absorbing Boundary Condition
Absorption
Boundary conditions
Computational physics
Conductors
Electric conductors
Electromagnetic radiation
Finite difference method
Finite difference time domain method
Numerical analysis
Optimization
Optimization methods
Parameters
Particle swarm optimization
Perfectly Matched Layer
Perfectly matched layers
Performance enhancement
Time domain analysis
Two dimensional analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A hybrid technique is studied in order to improve the performance of Convolutional Perfectly Matched Layer (CPML) in the Finite Difference Time Domain (FDTD) medium. This technique combines the first order of Higdon's annihilation equation as Absorbing Boundary Condition (ABC) with CPML to vanish the Perfect Electric Conductor (PEC) effects at the end of the CPML region. An optimization algorithm is required to find optimum parameters of the proposed absorber. In this investigation, the Particle Swarm Optimization (PSO) is utilized with two separate objective functions in order to minimize the average and peak value of relative error. Using a standard test, the overall performance of the proposed absorber is compared to the original CPML. The results clearly illustrate this method provides approximately 10 dB enhancements in CPML absorption error. The performance, memory and time requirement of the novel absorber, hybrid CPML (H-CPML), was analyzed during 2D and 3D tests and compared to most reported PMLs. The H-CPML requirement of computer resources is similar to CPML and can simply be implemented to truncate FDTD domains. Furthermore, an optimized set of parameters are provided to generalize the hybrid method.
ISSN:0021-9991
1090-2716
DOI:10.1016/j.jcp.2017.11.030