• Corpus ID: 166228192

Physics-informed Autoencoders for Lyapunov-stable Fluid Flow Prediction

@article{Erichson2019PhysicsinformedAF,
  title={Physics-informed Autoencoders for Lyapunov-stable Fluid Flow Prediction},
  author={N. Benjamin Erichson and Michael Muehlebach and Michael W. Mahoney},
  journal={ArXiv},
  year={2019},
  volume={abs/1905.10866}
}
In addition to providing high-profile successes in computer vision and natural language processing, neural networks also provide an emerging set of techniques for scientific problems. Such data-driven models, however, typically ignore physical insights from the scientific system under consideration. Among other things, a physics-informed model formulation should encode some degree of stability or robustness or well-conditioning (in that a small change of the input will not lead to drastic… 

Figures and Tables from this paper

Physics-Informed Probabilistic Learning of Linear Embeddings of Nonlinear Dynamics with Guaranteed Stability

TLDR
This work formalizes the problem of learning the continuous-time Koopman operator with deep neural networks in a measure-theoretic framework and enforces a structural parameterization that renders the realization of the Koop man operator provably stable.

Promoting global stability in data-driven models of quadratic nonlinear dynamics

TLDR
This work proposes that this theorem should be a standard diagnostic for the stability of projection-based and data-driven models, and illustrates how to modify the objective function in machine learning algorithms to promote globally stable models, with implications for the modeling of fluid and plasma flows.

Physics-informed machine learning

TLDR
Some of the prevailing trends in embedding physics into machine learning are reviewed, some of the current capabilities and limitations are presented and diverse applications of physics-informed learning both for forward and inverse problems, including discovering hidden physics and tackling high-dimensional problems are discussed.

Investigation of Nonlinear Model Order Reduction of the Quasigeostrophic Equations through a Physics-Informed Convolutional Autoencoder

TLDR
The investigation using the quasi-geostrophic equations reveals that while the PI cost function helps with spatial reconstruction, spatial features are less powerful than spectral features, and that construction of ROMs through machine learning-based methods requires significant investigation into novel non-standard methodologies.

Enhancing Computational Fluid Dynamics with Machine Learning

TLDR
This Perspective highlights some of the areas of highest potential impact of machine learning, including to accelerate direct numerical simulations, to improve turbulence closure modeling, and to develop enhanced reduced-order models.

Nonlinear proper orthogonal decomposition for convection-dominated flows

TLDR
A nonlinear proper orthogonal decomposition (POD) framework, which is an end-to-end Galerkin-free model combining autoencoders with long short-term memory networks for dynamics, which not only improves the accuracy, but also significantly reduces the computational cost of training and testing.

Physics-Guided Deep Learning for Dynamical Systems: A survey

TLDR
A structured overview of existing methodologies of integrating prior physical knowledge or physics-based modeling into deep learning and discuss the emerging opportunities is provided.

Physics-informed regularization and structure preservation for learning stable reduced models from data with operator inference

TLDR
Numerical results demonstrate that models learned with operator inference and the proposed regularizer and structure preservation are accurate and stable even in cases where using no regularization or Tikhonov regularization leads to models that are unstable.

Learning continuous models for continuous physics

TLDR
A convergence test based on numerical analysis theory is developed that verifies whether a model has learned a function that accurately approximates a system’s underlying continuous dynamics.

Deep Learning for Stability Analysis of a Freely Vibrating Sphere at Moderate Reynolds Number

TLDR
A deep learning-based reduced-order model (DLROM) for the stability prediction of unsteady 3D fluid-structure interaction systems and investigates the underlying mechanism and stability characteristics of VIV of a freely vibrating sphere via an eigenvalue selection process.
...

References

SHOWING 1-10 OF 41 REFERENCES

Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders

Deep Hidden Physics Models: Deep Learning of Nonlinear Partial Differential Equations

  • M. Raissi
  • Computer Science
    J. Mach. Learn. Res.
  • 2018
TLDR
This work puts forth a deep learning approach for discovering nonlinear partial differential equations from scattered and potentially noisy observations in space and time by approximate the unknown solution as well as the nonlinear dynamics by two deep neural networks.

Long-time predictive modeling of nonlinear dynamical systems using neural networks

TLDR
This work introduces Jacobian regularization in the loss function of the ANN, which suppresses the sensitivity of the prediction to the local error and is shown to improve accuracy and robustness.

Construction of reduced-order models for fluid flows using deep feedforward neural networks

TLDR
The DNN approach is able to learn transient features of the flow and presents more accurate and stable long-term predictions compared to sparse regression and the robustness of the current ROMs constructed via DNNs is demonstrated through a comparison with sparse regression.

Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations

TLDR
This two part treatise introduces physics informed neural networks – neural networks that are trained to solve supervised learning tasks while respecting any given law of physics described by general nonlinear partial differential equations and demonstrates how these networks can be used to infer solutions topartial differential equations, and obtain physics-informed surrogate models that are fully differentiable with respect to all input coordinates and free parameters.

Physics Informed Deep Learning (Part II): Data-driven Discovery of Nonlinear Partial Differential Equations

We introduce physics informed neural networks -- neural networks that are trained to solve supervised learning tasks while respecting any given law of physics described by general nonlinear partial

Deep learning algorithm for data-driven simulation of noisy dynamical system

Learning Deep Neural Network Representations for Koopman Operators of Nonlinear Dynamical Systems

TLDR
This paper introduces a computational framework for learning Koopman operators of nonlinear dynamical systems using deep learning and shows that this novel method automatically selects efficient deep dictionaries, requiring much lower dimensional dictionaries while outperforming state-of-the-art methods.