EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Physics-informed differentiable solvers for learning parametric solution manifolds in heterogeneous physical systems

Milad Panahi iD, Giovanni Michele Porta iD, Monica Riva iD, Alberto Guadagnini iD

PNAS Nexus · 30.05.2026 · 5 (6)

DOI: 10.1093/pnasnexus/pgag195 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PNAS Nexus, online erschienen am 01.06.2026. Verfasst von Milad Panahi und 3 weiteren Autor:innen. Beteiligte Einrichtung: Dipartimento di Ingegneria Civile e Ambientale, Politecnico di Milano, Piazza L. da Vinci 32, Milano 20133, Italy. Die Arbeit nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: European Union’s Horizon 2020 research and innovation programme.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Quantifying parametric uncertainty in partial differential equations is a central challenge to our ability to model the behavior of heterogeneous systems. This challenge is relevant to a variety of fundamental and application-oriented implications where system properties exhibit significant (and often uncertain) spatial heterogeneity. We address this by reformulating a physics-informed neural network as a differentiable solver that learns the continuous solution manifold for steady-state Darcy flow. Our framework requires only a single training run, circumventing the need for costly retraining for each new parameter instance. The approach is demonstrated through two representations of spatially heterogeneous hydraulic conductivity fields: a direct analytical form and a novel data-driven formulation resting on an autoencoder to create a low-dimensional latent encoding. A key innovation is the integration of the differentiable decoder into the physics-informed loss function, enabling on-the-fly reconstruction of complex conductivity fields. The approach yields accurate, mass-conserving flow solutions and supports efficient uncertainty quantification, providing a general methodology for physics-constrained data-driven modeling of heterogeneous systems.

Schlagwörter

Stichwörter der Autor:innen

Darcy flowdifferentiable physicsparameterized PDEsphysics-informed neural networksuncertainty quantification

Autor:innen und Institutionen

  • Dipartimento di Ingegneria Civile e Ambientale, Politecnico di Milano, Piazza L. da Vinci 32, Milano 20133, Italy.

ORCID-Profile: Milad Panahi, Giovanni Michele Porta, Monica Riva, Alberto Guadagnini

Förderung

  • European Union’s Horizon 2020 research and innovation programme

Zitieren

APA

Panahi, M., Porta, G. M., Riva, M., & Guadagnini, A. (2026). Physics-informed differentiable solvers for learning parametric solution manifolds in heterogeneous physical systems. PNAS Nexus, 5(6). https://doi.org/10.1093/pnasnexus/pgag195

Vancouver

Panahi M, Porta GM, Riva M, Guadagnini A. Physics-informed differentiable solvers for learning parametric solution manifolds in heterogeneous physical systems. PNAS Nexus. 2026;5(6). doi:10.1093/pnasnexus/pgag195

BibTeX
@article{Panahi2026,
  author = {Panahi, Milad and Porta, Giovanni Michele and Riva, Monica and Guadagnini, Alberto},
  title = {Physics-informed differentiable solvers for learning parametric solution manifolds in heterogeneous physical systems},
  journal = {PNAS Nexus},
  year = {2026},
  volume = {5},
  number = {6},
  doi = {10.1093/pnasnexus/pgag195}
}
Kostenlos starten →