Abstract
In multi-dimension problems, huge sparse matrices must be calculated according to Crank-Nicolson (CN) procedure which results in degeneration of efficiency and accuracy. Based on the iterated CN procedure, domain decomposition method, an alternative scheme is proposed for the termination of unbounded uniform finite-difference time-domain domains. Meanwhile, absorbing boundary condition is proposed in the iterated CN procedure which is incorporated with the higher order concept. The proposed scheme employs the explicit scheme during the calculation rather than the implicit one. Thus, the calculation of matrices can be avoided. It shows the advantages especially in efficiency and accuracy compared with implicit schemes. Such conclusion can be further demonstrated through the numerical example. From results, it can be concluded that the proposed scheme shows efficiency improvement and accurate maintenance. Compared with other procedures, it also holds its considerable effectiveness in nonuniform domains. For comparison, it can maintain considerable performance compared with the others.
| Original language | English |
|---|---|
| Pages (from-to) | 61-68 |
| Number of pages | 8 |
| Journal | IEEE Journal on Multiscale and Multiphysics Computational Techniques |
| Volume | 7 |
| DOIs | |
| State | Published - 2022 |
Keywords
- Domain decomposition
- finite-difference time-domain (fdtd)
- iterated crank-nicolson
- perfectly matched layer (pml)
- uniform fdtd domains
Fingerprint
Dive into the research topics of 'Iterated Crank-Nicolson Procedure With Enhanced Absorption for Nonuniform Domains'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver