Identifying first-principles models for bubble column aeration using machine learning

Peter Jul-Rasmussen; Arijit  Chakraborty; Venkat Venkatasubramanian; Xiaodong Liang; Jakob Kjøbsted Huusom

doi:10.1016/B978-0-443-15274-0.50174-8

Identifying first-principles models for bubble column aeration using machine learning

Peter Jul-Rasmussen, Arijit Chakraborty, Venkat Venkatasubramanian, Xiaodong Liang, Jakob Kjøbsted Huusom

Columbia University

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

Mass transfer of oxygen is investigated in this work using a pilot-scale bubble column unit with a two-fluid nozzle for aeration. First-principles models for the bubble column unit are identified by utilizing concepts in artificial intelligence (AI) and machine learning (ML), and applying the same to experimental data. By combining process knowledge with data-driven modeling, we discovered interpretable models for oxygen transport phenomena in bubble columns. By virtue of obtaining symbolic models, it is possible to perform post-hoc analyses on the same in order to gain physical insights into the mechanisms occurring in the system -- a convenience lost when using black-box models such as neural networks. This provides valuable understanding which can be applied when modeling more complex systems such as fermentation processes.

Original language	English
Title of host publication	Proceedings of the 33rd European Symposium on Computer Aided Process Engineering
Editors	Antonis Kokossis, Michael C. Georgiadis, Efstratios N. Pistikopoulos
Volume	52
Publisher	Elsevier
Publication date	2023
Pages	1089-1094
ISBN (Print)	978-0-443-23553-5, 978-0-443-15274-0
DOIs	https://doi.org/10.1016/B978-0-443-15274-0.50174-8
Publication status	Published - 2023
Event	33rd European Symposium on Computer Aided Process Engineering - Athens, Greece Duration: 18 Jun 2023 → 21 Jun 2023

Conference

Conference	33rd European Symposium on Computer Aided Process Engineering
Country/Territory	Greece
City	Athens
Period	18/06/2023 → 21/06/2023

Series	Computer Aided Chemical Engineering
Volume	52
ISSN	1570-7946

Keywords

Artificial intelligence
Machine Learning
Hybrid AI
Interpretable models
Mechanistic modeling

Access to Document

10.1016/B978-0-443-15274-0.50174-8

OpenUrl availability

Full text

Cite this

Jul-Rasmussen, P., Chakraborty, A., Venkatasubramanian, V., Liang, X., & Huusom, J. K. (2023). Identifying first-principles models for bubble column aeration using machine learning. In A. Kokossis, M. C. Georgiadis, & E. N. Pistikopoulos (Eds.), Proceedings of the 33rd European Symposium on Computer Aided Process Engineering (Vol. 52, pp. 1089-1094). Elsevier. https://doi.org/10.1016/B978-0-443-15274-0.50174-8

Jul-Rasmussen, Peter ; Chakraborty, Arijit ; Venkatasubramanian, Venkat et al. / Identifying first-principles models for bubble column aeration using machine learning. Proceedings of the 33rd European Symposium on Computer Aided Process Engineering. editor / Antonis Kokossis ; Michael C. Georgiadis ; Efstratios N. Pistikopoulos. Vol. 52 Elsevier, 2023. pp. 1089-1094 (Computer Aided Chemical Engineering, Vol. 52).

@inproceedings{44e39711db224107b8b61a3fe671ea08,

title = "Identifying first-principles models for bubble column aeration using machine learning",

abstract = "Mass transfer of oxygen is investigated in this work using a pilot-scale bubble column unit with a two-fluid nozzle for aeration. First-principles models for the bubble column unit are identified by utilizing concepts in artificial intelligence (AI) and machine learning (ML), and applying the same to experimental data. By combining process knowledge with data-driven modeling, we discovered interpretable models for oxygen transport phenomena in bubble columns. By virtue of obtaining symbolic models, it is possible to perform post-hoc analyses on the same in order to gain physical insights into the mechanisms occurring in the system -- a convenience lost when using black-box models such as neural networks. This provides valuable understanding which can be applied when modeling more complex systems such as fermentation processes.",

keywords = "Artificial intelligence, Machine Learning, Hybrid AI, Interpretable models, Mechanistic modeling",

author = "Peter Jul-Rasmussen and Arijit Chakraborty and Venkat Venkatasubramanian and Xiaodong Liang and Huusom, {Jakob Kj{\o}bsted}",

year = "2023",

doi = "10.1016/B978-0-443-15274-0.50174-8",

language = "English",

isbn = "978-0-443-23553-5",

volume = "52",

series = "Computer Aided Chemical Engineering",

publisher = "Elsevier",

pages = "1089--1094",

editor = "Kokossis, {Antonis } and {C. Georgiadis}, {Michael } and {N. Pistikopoulos}, Efstratios",

booktitle = "Proceedings of the 33rd European Symposium on Computer Aided Process Engineering",

address = "United Kingdom",

note = "33rd European Symposium on Computer Aided Process Engineering, ESCAPE33 ; Conference date: 18-06-2023 Through 21-06-2023",

}

Jul-Rasmussen, P, Chakraborty, A, Venkatasubramanian, V, Liang, X & Huusom, JK 2023, Identifying first-principles models for bubble column aeration using machine learning. in A Kokossis, M C. Georgiadis & E N. Pistikopoulos (eds), Proceedings of the 33rd European Symposium on Computer Aided Process Engineering. vol. 52, Elsevier, Computer Aided Chemical Engineering, vol. 52, pp. 1089-1094, 33rd European Symposium on Computer Aided Process Engineering, Athens, Greece, 18/06/2023. https://doi.org/10.1016/B978-0-443-15274-0.50174-8

Identifying first-principles models for bubble column aeration using machine learning. / Jul-Rasmussen, Peter; Chakraborty, Arijit ; Venkatasubramanian, Venkat et al.
Proceedings of the 33rd European Symposium on Computer Aided Process Engineering. ed. / Antonis Kokossis; Michael C. Georgiadis; Efstratios N. Pistikopoulos. Vol. 52 Elsevier, 2023. p. 1089-1094 (Computer Aided Chemical Engineering, Vol. 52).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Identifying first-principles models for bubble column aeration using machine learning

AU - Jul-Rasmussen, Peter

AU - Chakraborty, Arijit

AU - Venkatasubramanian, Venkat

AU - Liang, Xiaodong

AU - Huusom, Jakob Kjøbsted

PY - 2023

Y1 - 2023

N2 - Mass transfer of oxygen is investigated in this work using a pilot-scale bubble column unit with a two-fluid nozzle for aeration. First-principles models for the bubble column unit are identified by utilizing concepts in artificial intelligence (AI) and machine learning (ML), and applying the same to experimental data. By combining process knowledge with data-driven modeling, we discovered interpretable models for oxygen transport phenomena in bubble columns. By virtue of obtaining symbolic models, it is possible to perform post-hoc analyses on the same in order to gain physical insights into the mechanisms occurring in the system -- a convenience lost when using black-box models such as neural networks. This provides valuable understanding which can be applied when modeling more complex systems such as fermentation processes.

AB - Mass transfer of oxygen is investigated in this work using a pilot-scale bubble column unit with a two-fluid nozzle for aeration. First-principles models for the bubble column unit are identified by utilizing concepts in artificial intelligence (AI) and machine learning (ML), and applying the same to experimental data. By combining process knowledge with data-driven modeling, we discovered interpretable models for oxygen transport phenomena in bubble columns. By virtue of obtaining symbolic models, it is possible to perform post-hoc analyses on the same in order to gain physical insights into the mechanisms occurring in the system -- a convenience lost when using black-box models such as neural networks. This provides valuable understanding which can be applied when modeling more complex systems such as fermentation processes.

KW - Artificial intelligence

KW - Machine Learning

KW - Hybrid AI

KW - Interpretable models

KW - Mechanistic modeling

U2 - 10.1016/B978-0-443-15274-0.50174-8

DO - 10.1016/B978-0-443-15274-0.50174-8

M3 - Article in proceedings

SN - 978-0-443-23553-5

SN - 978-0-443-15274-0

VL - 52

T3 - Computer Aided Chemical Engineering

SP - 1089

EP - 1094

BT - Proceedings of the 33rd European Symposium on Computer Aided Process Engineering

A2 - Kokossis, Antonis

A2 - C. Georgiadis, Michael

A2 - N. Pistikopoulos, Efstratios

PB - Elsevier

T2 - 33rd European Symposium on Computer Aided Process Engineering

Y2 - 18 June 2023 through 21 June 2023

ER -

Jul-Rasmussen P, Chakraborty A, Venkatasubramanian V, Liang X , Huusom JK. Identifying first-principles models for bubble column aeration using machine learning. In Kokossis A, C. Georgiadis M, N. Pistikopoulos E, editors, Proceedings of the 33rd European Symposium on Computer Aided Process Engineering. Vol. 52. Elsevier. 2023. p. 1089-1094. (Computer Aided Chemical Engineering, Vol. 52). doi: 10.1016/B978-0-443-15274-0.50174-8

Identifying first-principles models for bubble column aeration using machine learning

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this