Fine-Pitch Interconnect Modeling Using Physics-Informed Neural Networks

Tosun R. A., Kuzucu D., DURGUN A. C., Baydogan M. G.

29th IEEE Workshop on Signal and Power Integrity, SPI 2025, Gaeta, Italy, 11 - 14 May 2025, (Full Text) identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/spi64682.2025.11014453
  • City: Gaeta
  • Country: Italy
  • Keywords: fine-pitch interconnect, hard-constrained boundary conditions,quasi-static, physics-informed neural networks, superposition
  • Middle East Technical University Affiliated: Yes

Abstract

Physics-informed neural networks (PINNs) have been demonstrated to solve partial differential equations (PDEs) effectively and show promise in electromagnetic (EM) analysis. In this study, we compare several PINN implementations for quasi-static modeling of fine-pitch interconnects. Our results indicate that imposing hard-constrained boundary conditions and implementing a superposition-based solution significantly improves the accuracy of multi-conductor interconnect modeling.