Stokesian processes : inferring Stokes flows using physics-informed Gaussian processes
Open Access
- 20 October 2023
- journal article
- research article
- Published by IOP Publishing in Machine Learning: Science and Technology
- Vol. 4 (4) , 045013
Abstract
We develop a probabilistic Stokes flow framework, using physics informed Gaussian processes, which can be used to solve both forward/inverse flow problems with missing and/or noisy data. The physics of the problem, specified by the Stokes and continuity equations, is exactly encoded into the inference framework. Crucially, this means that we do not need to explicitly solve the Poisson equation for the pressure field, as a physically meaningful (divergence-free) velocity field will automatically be selected. We test our method on a simple pressure driven flow problem, i.e. flow through a sinusoidal channel, and compare against standard numerical methods (Finite Element and Direct Numerical Simulations). We obtain excellent agreement, even when solving inverse problems given only sub-sampled velocity data on low dimensional sub-spaces (i.e. 1 component of the velocity on 1D domains to reconstruct 2D flows). The proposed method will be a valuable tool for analyzing experimental data, where noisy/missing data is the norm.Keywords
Funding Information
- Kyoto University (SPIRITS 2020)
- Japan Society for the Promotion of Science (Kakenhi 20K03786)
This publication has 29 references indexed in Scilit:
- Active learning of constitutive relation from mesoscopic dynamics for macroscopic modeling of non-Newtonian flowsJournal of Computational Physics, 2018
- Hidden physics models: Machine learning of nonlinear partial differential equationsJournal of Computational Physics, 2018
- Numerical Gaussian Processes for Time-Dependent and Nonlinear Partial Differential EquationsSIAM Journal on Scientific Computing, 2018
- Machine learning of linear differential equations using Gaussian processesJournal of Computational Physics, 2017
- Inferring solutions of differential equations using noisy multi-fidelity dataJournal of Computational Physics, 2017
- Nonlinear information fusion algorithms for data-efficient multi-fidelity modellingProceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2017
- Particle Image Velocimetry for Complex and Turbulent FlowsAnnual Review of Fluid Mechanics, 2013
- Smoothed profile method for particulate flows: Error analysis and simulationsJournal of Computational Physics, 2009
- Particle-Imaging Techniques for Experimental Fluid MechanicsAnnual Review of Fluid Mechanics, 1991
- Stokesian DynamicsAnnual Review of Fluid Mechanics, 1988