Imperial College London
Portrait Picture

Luca Magri

Faculty of Engineering, Department of Aeronautics

Professor of Scientific Machine Learning
 
 
 
//

Contact

 

l.magri Website

 
 
//

Location

 

CAGB324City and Guilds BuildingSouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@inproceedings{Ozan:2022:10.3397/IN_2022_0163,
author = {Ozan, DE and Magri, L},
doi = {10.3397/IN_2022_0163},
pages = {1191--1199},
publisher = {Institute of Noise Control Engineering},
title = {Physics-aware learning of nonlinear limit cycles and adjoint limit cycles},
url = {http://dx.doi.org/10.3397/IN_2022_0163},
year = {2022}
}
Download

RIS format (EndNote, RefMan)

TY  - CPAPER
AB  - Thermoacoustic oscillations occur when the heat released by a flame is sufficiently in phase with the acoustic pressure. Under this condition, the linear instability can saturate to a nonlinear self-excited oscillation with a large amplitude. A typical nonlinear regime is a limit cycle, which is characterised by a periodic orbit in the thermoacoustic phase space. In this paper, we develop a physics-aware data-driven method to predict periodic solutions using forward neural networks. The physics is constrained in two ways. First, the training is informed by a physical residual, which penalises solutions that violate the conservation of mass, momentum, and energy. Second, periodicity is imposed by introducing periodic activation functions in the neural network. We test the algorithm on synthetic data generated from a nonlinear time-delayed model of a Rijke tube. We extend our study to learning the adjoint variables of the Rijke system. Adjoint methods offer a cheap and easy way to calculate the gradients with respect to design parameters, We find that (i) periodic solutions of thermoacoustic systems can be accurately learned with this method, (ii) for periodic data, periodic activations outperform conventional activations in terms of extrapolation capability beyond the training range, and (iii) exploiting the physical constraints, fewer data is sufficient to achieve a good performance. This work opens up possibilities for the prediction of nonlinear thermoacoustics by combining physical knowledge and data.
AU  - Ozan,DE
AU  - Magri,L
DO  - 10.3397/IN_2022_0163
EP  - 1199
PB  - Institute of Noise Control Engineering
PY  - 2022///
SN  - 0736-2935
SP  - 1191
TI  - Physics-aware learning of nonlinear limit cycles and adjoint limit cycles
UR  - http://dx.doi.org/10.3397/IN_2022_0163
UR  - https://www.ingentaconnect.com/content/ince/incecp/2023/00000265/00000006/art00021
ER  -
Download