跳到主要导航 跳到搜索 跳到主要内容

Physics-Informed DeepONet Solution of Electromagnetic Scattering From 3-D PEC Scatterers With Wide Angular Sweeping

  • Ji Yuan Wang
  • , Liang Zhang
  • , Yun Chuan Wang
  • , Xin Yue Lou
  • , Cheng Tao Zhao
  • , Yan Nan Liu
  • , Xiao Min Pan*
  • *此作品的通讯作者
  • Beijing Institute of Technology
  • China Academy of Engineering Physics
  • IAPCM

科研成果: 期刊稿件 › 文章 › 同行评审

摘要

The physics-informed deep operator network (PI-DeepONet) is employed to solve electromagnetic (EM) scattering from 3-D PEC scatterers with wide angular sweeping, where the combined field integral equation (CFIE) with multiple right-hand sides (RHSs) is embedded into its loss function. The performance of the proposed PI-DeepONet-based EM solver is validated against the method of moments (MoM) with GPU acceleration (MoM-GPU). Numerical results show that the PI-DeepONet-based EM solver outperforms the MoM-GPU in terms of solution time including the training and inference, while maintaining comparable accuracy.

源语言英语
期刊IEEE Antennas and Wireless Propagation Letters
DOI
出版状态已接受/待刊 - 2026
已对外发布是

学术指纹

探究 'Physics-Informed DeepONet Solution of Electromagnetic Scattering From 3-D PEC Scatterers With Wide Angular Sweeping' 的科研主题。它们共同构成独一无二的学术指纹。

引用此