A Mesh-Free Neural Network-Based Method for the Analysis of Electromagnetic Structures Conference

Priota, NZ, Volakis, JL, Pulugurtha, MR et al. (2026). A Mesh-Free Neural Network-Based Method for the Analysis of Electromagnetic Structures . 474-475. 10.23919/NRSM68586.2026.11550895

cited authors

  • Priota, NZ; Volakis, JL; Pulugurtha, MR; Zekios, CL

authors

abstract

  • In this paper, we build on deep learning principles to develop a neural network-based framework capable of modeling both closed and open-radiating electromagnetic structures. Concurrently, we adapt computational tools that incorporate mesh refinement with appropriate basis functions. Specifically, we examine the impact of network depth (number of layers), network width (number of neurons), and activation functions, and demonstrate, for the first time, how these parameters govern solution accuracy in electromagnetic problems.

publication date

  • January 1, 2026

Digital Object Identifier (DOI)

start page

  • 474

end page

  • 475