机构:
Beijing Univ Posts & Telecommun, State Key Lab Informat Photon & Opt Commun, Beijing, Peoples R ChinaBeijing Univ Posts & Telecommun, State Key Lab Informat Photon & Opt Commun, Beijing, Peoples R China
Liu, Yuying
[1
]
Zhang, Wei
论文数: 0引用数: 0
h-index: 0
机构:
Beijing Univ Posts & Telecommun, State Key Lab Informat Photon & Opt Commun, Beijing, Peoples R ChinaBeijing Univ Posts & Telecommun, State Key Lab Informat Photon & Opt Commun, Beijing, Peoples R China
Zhang, Wei
[1
]
Zhang, Zhiguo
论文数: 0引用数: 0
h-index: 0
机构:
Beijing Univ Posts & Telecommun, State Key Lab Informat Photon & Opt Commun, Beijing, Peoples R ChinaBeijing Univ Posts & Telecommun, State Key Lab Informat Photon & Opt Commun, Beijing, Peoples R China
Zhang, Zhiguo
[1
]
Feng, Feng
论文数: 0引用数: 0
h-index: 0
机构:
Tianjin Univ, Tianjin, Peoples R ChinaBeijing Univ Posts & Telecommun, State Key Lab Informat Photon & Opt Commun, Beijing, Peoples R China
Feng, Feng
[2
]
机构:
[1] Beijing Univ Posts & Telecommun, State Key Lab Informat Photon & Opt Commun, Beijing, Peoples R China
[2] Tianjin Univ, Tianjin, Peoples R China
来源:
2024 IEEE INTERNATIONAL WORKSHOP ON RADIO FREQUENCY AND ANTENNA TECHNOLOGIES, IWRF&AT 2024
|
2024年
关键词:
Electromagnetic (EM) fast frequency sweep;
finite element method (FEM);
genetic algorithm (GA);
matrix Pade via Lanczos (MPVL);
topology optimization;
D O I:
10.1109/iWRFAT61200.2024.10594639
中图分类号:
TM [电工技术];
TN [电子技术、通信技术];
学科分类号:
0808 ;
0809 ;
摘要:
Electromagnetic (EM) topology optimization aims to achieve the optimal shape and topology of EM structures by adjusting the distribution of materials within the design space. Conventional proxy-based optimization approaches encounter difficulties in addressing this objective due to the inability to correlate various shapes and topologies through adjustments in geometric parameters. In response to this challenge, this paper puts forward an efficient electromagnetic topology optimization method tailored for microwave device design. A novel integration approach is introduced, which combines the merits of Lanczos (MPVL) and Householder formulas. This integration technique streamlines the computational burden associated with solving extensive finite element matrix equations across multiple frequencies. Furthermore, a genetic algorithm(GA) with an inheritance model is incorporated to generate new computational formulas in each generation. These formulas are derived from common factors between the new and parent EM structures, leading to additional computational cost reductions. To illustrate the efficacy of the proposed technique, a microwave example is presented to highlight its practical applicability.