Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Gaussian process hydrodynamics

Journal Article · · Applied Mathematics and Mechanics

Abstract

We present a Gaussian process (GP) approach, called Gaussian process hydrodynamics (GPH) for approximating the solution to the Euler and Navier-Stokes (NS) equations. Similar to smoothed particle hydrodynamics (SPH), GPH is a Lagrangian particle-based approach that involves the tracking of a finite number of particles transported by a 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 priorξon the velocity field and conditioning it on the vorticity at the particle locations. Known physics (e.g., the Richardson cascade and velocity increment power laws) is incorporated into the GP prior by using physics-informed additive kernels. This is equivalent to expressingξas a sum of independent GPsξl, which we call modes, acting at different scales (each modeξlself-activates to represent the formation of eddies at the corresponding scales). This approach enables a quantitative analysis of the Richardson cascade through the analysis of the activation of these modes, and enables us to analyze coarse-grain turbulence statistically rather than deterministically. Because GPH is formulated by using the vorticity equations, it does not require solving a pressure equation. By enforcing incompressibility and fluid-structure boundary conditions through the selection of a kernel, GPH requires significantly fewer particles than SPH. Because GPH has a natural probabilistic interpretation, the numerical results come with uncertainty estimates, enabling their incorporation into an uncertainty quantification (UQ) pipeline and adding/removing particles (quanta of information) in an adapted manner. The proposed approach is suitable for analysis because it inherits the complexity of state-of-the-art solvers for dense kernel matrices and results in 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 the accuracy and stability. Because the proposed approach uses a Bayesian interpretation, it naturally enables data assimilation and predictions and estimations by mixing simulation data and experimental data.

Research Organization:
Stanford Univ., CA (United States)
Sponsoring Organization:
USDOE Office of Science (SC)
DOE Contract Number:
SC0023163
OSTI ID:
2422205
Journal Information:
Applied Mathematics and Mechanics, Journal Name: Applied Mathematics and Mechanics Journal Issue: 7 Vol. 44; ISSN 0253-4827
Publisher:
Springer
Country of Publication:
United States
Language:
English

References (39)

Do ideas have shape? Idea registration as the continuous limit of artificial neural networks preprint January 2020
Multigrid with Rough Coefficients and Multiresolution Operator Decomposition from Hierarchical Information Games journal January 2017
Learning dynamical systems from data: A simple cross-validation perspective, part I: Parametric kernel flows journal July 2021
Physics-informed machine learning journal May 2021
Consistency of empirical Bayes and kernel flow for hierarchical parameter estimation journal June 2021
Learning dynamical systems from data: a simple cross-validation perspective, part II: nonparametric kernel flows text January 2021
Experimental study of the two-dimensional inverse energy cascade in a square box journal September 1986
Sparse Cholesky Factorization by Kullback--Leibler Minimization journal January 2021
A numerical approach to the testing of the fission hypothesis journal December 1977
De-noising by thresholding operator adapted wavelets journal September 2019
Gamblets for opening the complexity-bottleneck of implicit schemes for hyperbolic and parabolic ODEs/PDEs with rough coefficients journal October 2017
Vortex methods for flow simulation journal October 1980
Bayesian Numerical Homogenization journal January 2015
Grand challenges for Smoothed Particle Hydrodynamics numerical schemes journal September 2020
Smoothed Particle Hydrodynamics journal September 1992
Statistical Numerical Approximation journal November 2019
Learning “best” kernels from data in Gaussian process regression. With application to aerodynamics journal December 2022
The Anisotropic Lagrangian Averaged Euler and Navier-Stokes Equations journal January 2003
One-shot learning of stochastic differential equations with data adapted kernels journal February 2023
Computational graph completion journal April 2022
Can the atmospheric kinetic energy spectrum be explained by two-dimensional turbulence? journal June 1999
Vortex Methods: Theory and Practice book September 2009
Smoothed Particle Hydrodynamics (SPH): an Overview and Recent Developments journal February 2010
Operator-Adapted Wavelets, Fast Solvers, and Numerical Homogenization journal January 2019
Learning dynamical systems from data: A simple cross-validation perspective, part III: Irregularly-Sampled Time Series preprint January 2021
Can a finite element method perform arbitrarily badly? journal February 1999
Reproducing Kernel Hilbert Spaces Associated with Analytic Translation-Invariant Mercer Kernels journal January 2008
Simple, low-cost and accurate data-driven geophysical forecasting with learned kernels journal August 2021
Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations journal January 2020
Turbulence: The Legacy of A. N. Kolmogorov journal November 1996
Smoothed particle hydrodynamics: theory and application to non-spherical stars journal December 1977
Sixth problem of the millennium: Navier-Stokes equations, existence and smoothness journal April 2003
Two-Dimensional Turbulence journal January 2012
Kernel Flows: From learning kernels from data into the abyss journal July 2019
Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations journal February 2019
Stability of two-dimensional Navier–Stokes motions in the periodic case journal March 2015
Solving and learning nonlinear PDEs with Gaussian processes journal December 2021
Kernels for Vector-Valued Functions: A Review journal January 2012
Compression, Inversion, and Approximate PCA of Dense Kernel Matrices at Near-Linear Computational Complexity journal January 2021

Similar Records

Physics-constrained Gaussian process model for prediction of hydrodynamic interactions between wave energy converters in an array
Journal Article · 2023 · Applied Mathematical Modelling · OSTI ID:1967499

Progress in smooth particle hydrodynamics
Conference · 1998 · OSTI ID:663210

Mapping closure for probability distribution function in low frequency magnetized plasma turbulence
Journal Article · 1995 · Physics of Plasmas · OSTI ID:165176

Related Subjects