Gaussian process hydrodynamics

DOI10.1007/S10483-023-2990-9zbMATH Open1516.35310arXiv2209.10707OpenAlexW4366449378MaRDI QIDQ6132295FDOQ6132295

Authors:

Publication date: 14 July 2023

Published in: AMM. Applied Mathematics and Mechanics. (English Edition) (Search for Journal in Brave)

Abstract: We present a Gaussian Process (GP) approach (Gaussian Process Hydrodynamics, GPH) for approximating the solution of the Euler and Navier-Stokes equations. As in Smoothed Particle Hydrodynamics (SPH), GPH is a Lagrangian particle-based approach involving the tracking of a finite number of particles transported by the flow. However, these particles do not represent mollified particles of matter but carry discrete/partial information about the continuous flow. Closure is achieved by placing a divergence-free GP prior $xi$ on the velocity field and conditioning on vorticity at particle locations. Known physics (e.g., the Richardson cascade and velocity-increments power laws) is incorporated into the GP prior through physics-informed additive kernels. This approach allows us to coarse-grain turbulence in a statistical manner rather than a deterministic one. By enforcing incompressibility and fluid/structure boundary conditions through the selection of the kernel, GPH requires much fewer particles than SPH. Since GPH has a natural probabilistic interpretation, numerical results come with uncertainty estimates enabling their incorporation into a UQ pipeline and the adding/removing of particles (quantas of information) in an adapted manner. The proposed approach is amenable to analysis, it inherits the complexity of state-of-the-art solvers for dense kernel matrices, and it leads to a natural definition of turbulence as information loss. Numerical experiments support the importance of selecting physics-informed kernels and illustrate the major impact of such kernels on accuracy and stability. Since the proposed approach has a Bayesian interpretation, it naturally enables data assimilation and making predictions and estimations based on mixing simulation data with experimental data.

Full work available at URL: https://arxiv.org/abs/2209.10707

Recommendations

• Gaussian processes for computer experiments
• Gaussian processes and Gaussian random fields
• Numerical Gaussian processes for time-dependent and nonlinear partial differential equations
• A new class of high-order methods for fluid dynamics simulations using Gaussian process modeling: one-dimensional case
• Gaussian processes and mixed volumes
• Applying Bayesian optimization with Gaussian process regression to computational fluid dynamics problems
• Solving and learning nonlinear PDEs with Gaussian processes
• scientific article; zbMATH DE number 3850197
• Enhanced Gaussian processes and applications

zbMATH Keywords

EulervorticityLagrangianGaussian process (GP)Navier-Stokes (NS) equationphysics-informed kernel

Mathematics Subject Classification ID

Gaussian processes (60G15) Probabilistic methods, particle methods, etc. for boundary value problems involving PDEs (65N75) Navier-Stokes equations (35Q30) Navier-Stokes equations for incompressible viscous fluids (76D05) Nonlinear boundary value problems for ordinary differential equations (34B15) Spline approximation (41A15) Probabilistic methods, particle methods, etc. for initial value and initial-boundary value problems involving PDEs (65M75) Spectral, collocation and related methods for boundary value problems involving PDEs (65N35) Kernel operators (47B34)

Cites Work

• Kernels for vector-valued functions: a review
• Multigrid with Rough Coefficients and Multiresolution Operator Decomposition from Hierarchical Information Games
• Operator-adapted wavelets, fast solvers, and numerical homogenization. From a game theoretic approach to numerical approximation and algorithm design
• Compression, inversion, and approximate PCA of dense kernel matrices at near-linear computational complexity
• Scattered Data Approximation
• Two-dimensional turbulence
• Can a finite element method perform arbitrarily badly?
• Bayesian numerical homogenization
• Title not available (Why is that?)
• Smoothed particle hydrodynamics: theory and application to non-spherical stars
• Smoothed particle hydrodynamics (SPH): an overview and recent developments
• Title not available (Why is that?)
• Vortex Methods
• Vortex methods for flow simulation
• The anisotropic Lagrangian averaged Euler and Navier-Stokes equations
• Sixth problem of the millennium: Navier-Stokes equations, existence and smoothness
• Stability of two-dimensional Navier-Stokes motions in the periodic case
• Kernel mode decomposition and the programming of kernels
• Reproducing kernel Hilbert spaces associated with analytic translation-invariant Mercer kernels
• Gamblets for opening the complexity-bottleneck of implicit schemes for hyperbolic and parabolic ODEs/PDEs with rough coefficients
• Can the atmospheric kinetic energy spectrum be explained by two-dimensional turbulence?
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
• Kernel flows: from learning kernels from data into the abyss
• Learning dynamical systems from data: a simple cross-validation perspective. I: Parametric kernel flows
• Learning dynamical systems from data: a simple cross-validation perspective. III: Irregularly-sampled time series
• Statistical Numerical Approximation
• Learning ``best kernels from data in Gaussian process regression. With application to aerodynamics
• Solving and learning nonlinear PDEs with Gaussian processes
• Consistency of empirical Bayes and kernel flow for hierarchical parameter estimation
• Sparse Cholesky Factorization by Kullback--Leibler Minimization
• One-shot learning of stochastic differential equations with data adapted kernels
• Do ideas have shape? Idea registration as the continuous limit of artificial neural networks
• Computational graph completion
• De-noising by thresholding operator adapted wavelets

Cited In (3)

• Surrogate Modeling with Gaussian Processes for an Inverse Problem in Polymer Dynamics
• A new class of high-order methods for fluid dynamics simulations using Gaussian process modeling: one-dimensional case
• Preface: machine-learning approaches for computational mechanics

This page was built for publication: Gaussian process hydrodynamics

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6132295)