Deep learning of interfacial curvature: a symmetry-preserving approach for the volume of fluid method
From MaRDI portal
Publication:6158093
DOI10.1016/J.JCP.2023.112110arXiv2206.06041OpenAlexW4362608082MaRDI QIDQ6158093FDOQ6158093
Authors: Asim Önder, Philip L.-F. Liu 
Publication date: 31 May 2023
Published in: Journal of Computational Physics (Search for Journal in Brave)
Abstract: Estimation of interface curvature in surface-tension dominated flows is a remaining challenge in Volume of Fluid (VOF) methods. Data-driven methods are recently emerging as a promising alternative in this domain. They outperform conventional methods on coarser grids but diverge with grid refinement. Furthermore, unlike conventional methods, data-driven methods are sensitive to coordinate system and sign conventions, thus often fail to capture basic symmetry patterns in interfaces. The present work proposes a new data-driven strategy which conserves the symmetries in a cost-effective way and delivers consistent results over a wide range of grids. The method is based on artificial neural networks with deep multilayer perceptron (MLP) architecture which read volume fraction fields on regular grids. The anti-symmetries are preserved with no additional cost by employing a neural network model with input normalization, odd-symmetric activation functions and bias-free neurons. The symmetries are further conserved by height-function inspired rotations and averaging over several different orientations. The new symmetry-preserving MLP model is implemented into a flow solver (OpenFOAM) and tested against conventional schemes in the literature. It shows superior performance compared to its standard counterpart and has similar accuracy and convergence properties with the state-of-the-art conventional method despite using smaller stencil.
Full work available at URL: https://arxiv.org/abs/2206.06041
Multiphase and multicomponent flows (76Txx) Incompressible viscous fluids (76Dxx) Basic methods in fluid mechanics (76Mxx)
Cites Work
- Validation of volume-of-fluid OpenFOAM{\circledR} isoAdvector solvers using single bubble benchmarks
- A continuum method for modeling surface tension
- An accurate adaptive solver for surface-tension-driven interfacial flows
- Reconstructing volume tracking
- A second order coupled level set and volume-of-fluid method for computing growth and collapse of vapor bubbles.
- Second-order accurate volume-of-fluid algorithms for tracking material interfaces
- Interface reconstruction with least-squares fit and split advection in three-dimensional Cartesian geometry
- Quantitative benchmark computations of two-dimensional bubble dynamics
- Interface reconstruction with least-square fit and split Eulerian-Lagrangian advection
- A wave generation toolbox for the open‐source CFD library: OpenFoam®
- Title not available (Why is that?)
- Motion of two superposed viscous fluids
- On the properties and limitations of the height function method in two-dimensional Cartesian geometry
- A mesh-decoupled height function method for computing interface curvature
- Numerical models of surface tension
- On the combined effects of surface tension force calculation and interface advection on spurious currents within volume of fluid and level set frameworks
- Neural networks and deep learning. A textbook
- Consistent and symmetry preserving data-driven interface reconstruction for the level-set method
- NPLIC: a machine learning approach to piecewise linear interface construction
- Computing interface curvature from volume fractions: a machine learning approach
- A hybrid inference system for improved curvature estimation in the level-set method using machine learning
- A Deep Learning Approach for the Computation of Curvature in the Level-Set Method
- Accurate and efficient surface reconstruction from volume fraction data on general meshes
Cited In (1)
This page was built for publication: Deep learning of interfacial curvature: a symmetry-preserving approach for the volume of fluid method
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6158093)
