A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression

Aurélie Bellemans; Mohammad Rafi Malik; Fabrizio Bisetti; Alessandro Parente

A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression

Aurélie Bellemans, Mohammad Rafi Malik, Fabrizio Bisetti, Alessandro Parente

Research output: Chapter in Book/Report/Conference proceeding › Conference paper

160 Downloads (Pure)

Abstract

The intricate nature of detailed kinetic mechanisms for plasma-assisted combustion motivates the development of high-fidelity surrogates to simplify their use in extensive numerical simulations. This chapter presents a machine-learning approach based on the coupling of Principal Component Analysis (PCA) with Gaussian Process Regression (GPR) to produce a reliable reduced-order representation of the detailed plasma-combustion physics. The entire state-space is expressed in function of a selected number of principal components using a non-linear Gaussian regression model. This machine-learning framework allows for a superior dimensionality compression compared to conventional data-driven reduction strategies based solely on principal component analysis. The performance of the present technique is assessed for the simulation of ethylene-air ignition by nanosecond repetitive pulsed discharges.

Original language	English
Title of host publication	Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference
Subtitle of host publication	Space Technology Proceedings
Publisher	Springer, Cham
Pages	379-392
Number of pages	13
Volume	8
ISBN (Electronic)	978-3-030-80542-5
ISBN (Print)	978-3-030-80541-8
Publication status	Published - 16 Jul 2021

Access to Document

Bellemans_UQOPAccepted author manuscript, 921 KBLicence: Unspecified

Handle.net

Persistent link

SRP60: SRP-Groeifinanciering: A system identification framework for multi-fidelity modelling
De Troyer, T. (Administrative Promotor), Runacres, M. (Co-Promotor), Blondeau, J. (Co-Promotor), Bram, S. (Co-Promotor), Bellemans, A. (Co-Promotor) & Contino, F. (Co-Promotor)
1/03/19 → 29/02/24
Project: Fundamental

Cite this

Bellemans, A., Malik, M. R., Bisetti, F., & Parente, A. (2021). A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression. In Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference: Space Technology Proceedings (Vol. 8, pp. 379-392). Springer, Cham.

Bellemans, Aurélie ; Malik, Mohammad Rafi ; Bisetti, Fabrizio et al. / A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression. Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference: Space Technology Proceedings. Vol. 8 Springer, Cham, 2021. pp. 379-392

@inproceedings{007599a05c254ca4bc32db0489230c61,

title = "A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression",

abstract = "The intricate nature of detailed kinetic mechanisms for plasma-assisted combustion motivates the development of high-fidelity surrogates to simplify their use in extensive numerical simulations. This chapter presents a machine-learning approach based on the coupling of Principal Component Analysis (PCA) with Gaussian Process Regression (GPR) to produce a reliable reduced-order representation of the detailed plasma-combustion physics. The entire state-space is expressed in function of a selected number of principal components using a non-linear Gaussian regression model. This machine-learning framework allows for a superior dimensionality compression compared to conventional data-driven reduction strategies based solely on principal component analysis. The performance of the present technique is assessed for the simulation of ethylene-air ignition by nanosecond repetitive pulsed discharges.",

author = "Aur{\'e}lie Bellemans and Malik, {Mohammad Rafi} and Fabrizio Bisetti and Alessandro Parente",

year = "2021",

month = jul,

day = "16",

language = "English",

isbn = "978-3-030-80541-8",

volume = "8",

pages = "379--392",

booktitle = "Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference",

publisher = "Springer, Cham",

}

Bellemans, A, Malik, MR, Bisetti, F & Parente, A 2021, A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression. in Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference: Space Technology Proceedings. vol. 8, Springer, Cham, pp. 379-392.

A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression. / Bellemans, Aurélie; Malik, Mohammad Rafi; Bisetti, Fabrizio et al.
Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference: Space Technology Proceedings. Vol. 8 Springer, Cham, 2021. p. 379-392.

Research output: Chapter in Book/Report/Conference proceeding › Conference paper

TY - GEN

T1 - A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression

AU - Bellemans, Aurélie

AU - Malik, Mohammad Rafi

AU - Bisetti, Fabrizio

AU - Parente, Alessandro

PY - 2021/7/16

Y1 - 2021/7/16

N2 - The intricate nature of detailed kinetic mechanisms for plasma-assisted combustion motivates the development of high-fidelity surrogates to simplify their use in extensive numerical simulations. This chapter presents a machine-learning approach based on the coupling of Principal Component Analysis (PCA) with Gaussian Process Regression (GPR) to produce a reliable reduced-order representation of the detailed plasma-combustion physics. The entire state-space is expressed in function of a selected number of principal components using a non-linear Gaussian regression model. This machine-learning framework allows for a superior dimensionality compression compared to conventional data-driven reduction strategies based solely on principal component analysis. The performance of the present technique is assessed for the simulation of ethylene-air ignition by nanosecond repetitive pulsed discharges.

AB - The intricate nature of detailed kinetic mechanisms for plasma-assisted combustion motivates the development of high-fidelity surrogates to simplify their use in extensive numerical simulations. This chapter presents a machine-learning approach based on the coupling of Principal Component Analysis (PCA) with Gaussian Process Regression (GPR) to produce a reliable reduced-order representation of the detailed plasma-combustion physics. The entire state-space is expressed in function of a selected number of principal components using a non-linear Gaussian regression model. This machine-learning framework allows for a superior dimensionality compression compared to conventional data-driven reduction strategies based solely on principal component analysis. The performance of the present technique is assessed for the simulation of ethylene-air ignition by nanosecond repetitive pulsed discharges.

M3 - Conference paper

SN - 978-3-030-80541-8

VL - 8

SP - 379

EP - 392

BT - Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference

PB - Springer, Cham

ER -

Bellemans A, Malik MR, Bisetti F, Parente A. A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression. In Advances in Uncertainty Quantification and Optimization Under Uncertainty with Aerospace Applications: Proceedings of the 2020 UQOP International Conference: Space Technology Proceedings. Vol. 8. Springer, Cham. 2021. p. 379-392

A Machine-Learning Framework for Plasma-Assisted Combustion Using Principal Component Analysis and Gaussian Process Regression

Abstract

Access to Document

Handle.net

Fingerprint

Projects

SRP60: SRP-Groeifinanciering: A system identification framework for multi-fidelity modelling

Cite this