Hard-constrained neural networks for modeling nonlinear acoustics

@article{Ozan2023HardconstrainedNN,
  title={Hard-constrained neural networks for modeling nonlinear acoustics},
  author={Defne Ege Ozan and Luca Magri},
  journal={Physical Review Fluids},
  year={2023},
  url={https://api.semanticscholar.org/CorpusID:258887417}
}
An acoustic neural networks are developed that learn from sensor data, whilst being constrained by prior knowledge on acoustic and wave physics by both informing the training and constraining parts of the network's architecture as an inductive bias.
View PDF on arXiv
5 Citations
Background Citations
3
Methods Citations
1

Figures and Tables from this paper

5 Citations

Current and emerging deep-learning methods for the simulation of fluid dynamics

This work describes a series of novel DL methods for replacing or augmenting conventional numerical solvers for simulating fluid dynamics, and broadly classify these methods into physics- and data-driven methods.

Data-driven computation of adjoint sensitivities without adjoint solvers: An application to thermoacoustics

Adjoint methods have been the pillar of gradient-based optimization for decades. They enable the accurate computation of a gradient (sensitivity) of a quantity of interest with respect to all system…

A Dual-Path neural network model to construct the flame nonlinear thermoacoustic response in the time domain

Deep learning algorithms are developed that can construct a comprehensive flame nonlinear response from limited numerical simulation data using a frequency-sweeping data type as the training dataset, and a Dual-Path neural network is introduced to enhance the precision in learning flame nonlinear response patterns from the training data.

A PINN approach for identifying governing parameters of noisy thermoacoustic systems

Abstract Identifying the governing parameters of self-sustained oscillation is crucial for the diagnosis, prediction and control of thermoacoustic instabilities. In this paper, we propose and…

Inverse flow prediction using ensemble PINNs and uncertainty quantification

68 References

Towards reconstruction of acoustic fields via physics-informed neural networks

A method based on physics informed neural networks (PINN) is proposed in order to reconstruct the entire acoustic field of a given acoustic element, provided some acoustic measurements at some few locations, which satisfy both measurements and the acoustic wave equation.

Physics Constrained Learning for Data-driven Inverse Modeling from Sparse Observations

Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations

Learning in Modal Space: Solving Time-Dependent Stochastic PDEs Using Physics-Informed Neural Networks

Two new Physics-Informed Neural Networks (PINNs) are proposed for solving time-dependent SPDEs, namely the NN-DO/BO methods, which incorporate the DO/BO constraints into the loss function with an implicit form instead of generating explicit expressions for the temporal derivatives of the Do/BO modes.

Physics-aware learning of nonlinear limit cycles and adjoint limit cycles

It is found that periodic solutions of thermoacoustic systems can be accurately learned with this method, and for periodic data, periodic activations outperform conventional activations in terms of prediction capability beyond the training range, and under the physical constraints, fewer data is sufficient to achieve a good performance.

Understanding the difficulty of training deep feedforward neural networks

The objective here is to understand better why standard gradient descent from random initialization is doing so poorly with deep neural networks, to better understand these recent relative successes and help design better algorithms in the future.

Physics-informed machine learning

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.

Short- and long-term predictions of chaotic flows and extreme events: a physics-constrained reservoir computing approach

It is shown that the combination of the two approaches is able to accurately reproduce the velocity statistics, and to predict the occurrence and amplitude of extreme events in a model of self-sustaining process in turbulence, and is shown to be robust with respect to noise.

Real-time thermoacoustic data assimilation

A Bayesian data assimilation method for a low–order model to self–adapt and self–correct any time that reference data, for example from experiments, becomes available to infer the thermoacoustic states, heat release parameters, and model errors on the fly without storing data (real–time).

Deep learning of vortex-induced vibrations

A new paradigm of inference in fluid mechanics for coupled multi-physics problems enables velocity and pressure quantification from flow snapshots in small subdomains and can be exploited for flow control applications and also for system identification.
...