Neural Spectral Element Methods for stiff multiphysics PDEs with electrochemical transport benchmarks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Feugmo, Conrard Giresse Tetsassi, Pankaczy, David
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916074898849792
author Feugmo, Conrard Giresse Tetsassi
Pankaczy, David
author_facet Feugmo, Conrard Giresse Tetsassi
Pankaczy, David
contents The Neural Spectral Element Method (NSEM) evaluates each network only at fixed Legendre-Gauss-Lobatto quadrature nodes and replaces all derivative calls with precomputed spectral differentiation matrices. The resulting deterministic loss enables limited-memory BFGS (L-BFGS) to reach residuals of 10^-9 to 10^-10. A Kosloff-Tal-Ezer coordinate map resolves electrochemical boundary layers, while a mesh-free neural mortar framework couples multi-element domains. On the four-example Poisson-Nernst-Planck (PNP) benchmark of Huang and co-workers, NSEM attains 10^-4 to 10^-7 relative pointwise error with two orders of magnitude fewer collocation points than the adaptive-resampling PINN baseline. Both a tanh multilayer perceptron (MLP) and a basis-aligned Legendre Kolmogorov-Arnold Network (KAN) backbone attain spectral accuracy within the same NSEM infrastructure, with the KAN requiring roughly half the Adam steps to enter the L-BFGS basin of attraction on the 1D PNP benchmark.
format Preprint
id arxiv_https___arxiv_org_abs_2606_02335
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Neural Spectral Element Methods for stiff multiphysics PDEs with electrochemical transport benchmarks
Feugmo, Conrard Giresse Tetsassi
Pankaczy, David
Materials Science
Mathematical Physics
Applied Physics
Computational Physics
The Neural Spectral Element Method (NSEM) evaluates each network only at fixed Legendre-Gauss-Lobatto quadrature nodes and replaces all derivative calls with precomputed spectral differentiation matrices. The resulting deterministic loss enables limited-memory BFGS (L-BFGS) to reach residuals of 10^-9 to 10^-10. A Kosloff-Tal-Ezer coordinate map resolves electrochemical boundary layers, while a mesh-free neural mortar framework couples multi-element domains. On the four-example Poisson-Nernst-Planck (PNP) benchmark of Huang and co-workers, NSEM attains 10^-4 to 10^-7 relative pointwise error with two orders of magnitude fewer collocation points than the adaptive-resampling PINN baseline. Both a tanh multilayer perceptron (MLP) and a basis-aligned Legendre Kolmogorov-Arnold Network (KAN) backbone attain spectral accuracy within the same NSEM infrastructure, with the KAN requiring roughly half the Adam steps to enter the L-BFGS basin of attraction on the 1D PNP benchmark.
title Neural Spectral Element Methods for stiff multiphysics PDEs with electrochemical transport benchmarks
topic Materials Science
Mathematical Physics
Applied Physics
Computational Physics
url https://arxiv.org/abs/2606.02335