Hard Constraint Projection in a Physics Informed Neural Network (Burgers)

Abstract

Physics informed machine learning provides a framework to study and simulate fluid dynamics, potentially at a fraction of the computational cost of traditional numerical methods, with the goal of improving the explainability, accuracy and robustness of purely data driven machine learning methods. In this work, we extend the hard constraint projection (HCP) method developed by Chen et al. [Journal of Computational Physics, 445, 110624, 2021] from a linear partial differential equation (PDE) to a nonlinear PDE: the 1D Burgers' equation. To strictly impose a discrete approximation of the governing equation, the Burgers' equation is first locally discretised (using a finite difference stencil) then decomposed into a constraint matrix and a prediction matrix. Using tools from linear algebra, the constraint matrix is used to construct a projection matrix. A physics informed neural network (PINN) [Raissi et al. Journal of Computational Physics, 378, 686, 2019] is used to predict the velocity, and a final unlearnable layer projects the velocity prediction to the hyperplane of exact solutions to a discretised form of the Burgers' equation. Our initial results indicate that the HCP-PINN initially trains similarly to an equivalent PINN, and as the PINN loss plateaus, the HCP-PINN loss continues to decline. When tested on a previously unseen uniform grid of points covering the domain, the HCP-PINN exhibits higher accuracy, converging to a data error approximately an order of magnitude lower than the comparable PINN. We also observe that the inclusion of the HCP layer results in better observance of the governing PDE. Our findings suggest that the HCP-PINN can be more robust to the selection of physics collocation points, and less dependent on collocation point quantity.

Projection-based segregated, iterative, and multi-stage physics-informed neural network for incompressible Navier-stokes equations
Zhenyu Zhang, Zhixiang Ling, Yinghao Huo +2
arXiv5 Nov 2026PreprintarXiv licenceFull textComputational physics Fluid dynamicsarXiv
On the Assessment of Domain Decomposition Approaches in Extended Physics-Informed Neural Networks for the Shock-Dominated Burgers Equation
Andrés F. Bolaños-Acosta, Julián C. Restrepo
arXiv1 Oct 2026PreprintarXiv licenceFull textComputational physics Fluid dynamicsarXiv
Energy-Guided Transport for Projection-Free Physics-Informed Flow Matching
Giuliano Galadini, Aurelio Uncini, Danilo Comminiello
arXiv4 Nov 2026PreprintCC BY 4.0Full textComputational physicsarXiv
PINNIES: An efficient physics-informed neural network framework for integral operator problems
Alireza Afzal Aghaei, Mahdi Movahedian Moghaddam, Kourosh Parand
Computer Physics Communications6 Jun 2026Peer-reviewedLicence not statedFull text10 citedStatistical & nonlinear physicsOpenAlex
Comparison of a Parametric Physics-Informed Neural Network and a Tensorial Reduced-Order Model for the Shallow-Water Dam-Break Problem
Anton Myshak, Md Rezwan Bin Mizan, Ilya Timofeyev
arXiv5 Oct 2026PreprintarXiv licenceFull textNumerical analysis & scientific computing Fluid dynamicsarXiv

Cite

BibTeX RIS

BibTeX

@misc{horne2026hard,
  title = {{Hard Constraint Projection in a Physics Informed Neural Network (Burgers)}},
  author = {Miranda J S Horne and Peter K Jimack and Amirul Khan and He Wang},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.42106},
  eprint = {2610.42106},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.42106},
}

RIS

TY  - UNPB
TI  - Hard Constraint Projection in a Physics Informed Neural Network (Burgers)
AU  - Miranda J S Horne
AU  - Peter K Jimack
AU  - Amirul Khan
AU  - He Wang
PY  - 2026
DA  - 2026/10/30
PB  - arXiv
DO  - 10.48550/arxiv.2610.42106
UR  - https://arxiv.org/abs/2610.42106
AB  - Physics informed machine learning provides a framework to study and simulate fluid dynamics, potentially at a fraction of the computational cost of traditional numerical methods, with the goal of improving the explainability, accuracy and robustness of purely data driven machine learning methods. In this work, we extend the hard constraint projection (HCP) method developed by Chen et al. [Journal of Computational Physics, 445, 110624, 2021] from a linear partial differential equation (PDE) to a nonlinear PDE: the 1D Burgers' equation. To strictly impose a discrete approximation of the governing equation, the Burgers' equation is first locally discretised (using a finite difference stencil) then decomposed into a constraint matrix and a prediction matrix. Using tools from linear algebra, the constraint matrix is used to construct a projection matrix. A physics informed neural network (PINN) [Raissi et al. Journal of Computational Physics, 378, 686, 2019] is used to predict the velocity, and a final unlearnable layer projects the velocity prediction to the hyperplane of exact solutions to a discretised form of the Burgers' equation. Our initial results indicate that the HCP-PINN initially trains similarly to an equivalent PINN, and as the PINN loss plateaus, the HCP-PINN loss continues to decline. When tested on a previously unseen uniform grid of points covering the domain, the HCP-PINN exhibits higher accuracy, converging to a data error approximately an order of magnitude lower than the comparable PINN. We also observe that the inclusion of the HCP layer results in better observance of the governing PDE. Our findings suggest that the HCP-PINN can be more robust to the selection of physics collocation points, and less dependent on collocation point quantity.
ER  -

Abstract

Related

Cite