Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration

Digital twins are frequently discussed in a bio-manufacturing context, but actual realisations of digital twins are rare. To use digital twin instances, significant investments in digital infrastructure and high-fidelity mathematical models are required. This work presents a real-time implementation...

Full description

Saved in:
Bibliographic Details
Main Authors: Peter Jul-Rasmussen, Mads Stevnsborg, Xiaodong Liang, Jakob Kjøbsted Huusom
Format: Article
Language:English
Published: Elsevier 2025-03-01
Series:Digital Chemical Engineering
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2772508124000747
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850100953661833216
author Peter Jul-Rasmussen
Mads Stevnsborg
Xiaodong Liang
Jakob Kjøbsted Huusom
author_facet Peter Jul-Rasmussen
Mads Stevnsborg
Xiaodong Liang
Jakob Kjøbsted Huusom
author_sort Peter Jul-Rasmussen
collection DOAJ
description Digital twins are frequently discussed in a bio-manufacturing context, but actual realisations of digital twins are rare. To use digital twin instances, significant investments in digital infrastructure and high-fidelity mathematical models are required. This work presents a real-time implementation of an ensemble hybrid model with incremental learning for predicting dissolved oxygen concentration in a pilot-scale bubble column. A bootstrap-aggregated hybrid modelling framework is applied for constructing an ensemble of 1000 hybrid models using different partitions of the training/validation data, providing a measure of the parameter distributions and prediction uncertainty. Each model in the ensemble hybrid model has the same model structure relying on first-principles material balances and an Artificial Neural Network for prediction of the liquid phase volumetric mass transfer coefficient. Incremental learning is applied, efficiently enabling the model to adapt to new data acquired during runtime. The software implementation follows recent ISO issues using a modular structure allowing for flexible allocation of server resources and an intuitive User-Interface is developed for controlling the application. From a real-time prediction study, the models using incremental learning are found to have superior performance both at normal operating conditions, when interpolating, and when extrapolating compared to using only the pre-trained model.
format Article
id doaj-art-0a10a448323340e99c82144f0dbe82d8
institution DOAJ
issn 2772-5081
language English
publishDate 2025-03-01
publisher Elsevier
record_format Article
series Digital Chemical Engineering
spelling doaj-art-0a10a448323340e99c82144f0dbe82d82025-08-20T02:40:10ZengElsevierDigital Chemical Engineering2772-50812025-03-011410021210.1016/j.dche.2024.100212Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aerationPeter Jul-Rasmussen0Mads Stevnsborg1Xiaodong Liang2Jakob Kjøbsted Huusom3Process and Systems Engineering Center (PROSYS), Department of Chemical and Biochemical Engineering, Technical University of Denmark, Kongens Lyngby, 2800, DenmarkProcess and Systems Engineering Center (PROSYS), Department of Chemical and Biochemical Engineering, Technical University of Denmark, Kongens Lyngby, 2800, DenmarkCenter for Energy Resources Engineering (CERE), Department of Chemical and Biochemical Engineering, Technical University of Denmark, Kongens Lyngby, 2800, DenmarkProcess and Systems Engineering Center (PROSYS), Department of Chemical and Biochemical Engineering, Technical University of Denmark, Kongens Lyngby, 2800, Denmark; Corresponding author.Digital twins are frequently discussed in a bio-manufacturing context, but actual realisations of digital twins are rare. To use digital twin instances, significant investments in digital infrastructure and high-fidelity mathematical models are required. This work presents a real-time implementation of an ensemble hybrid model with incremental learning for predicting dissolved oxygen concentration in a pilot-scale bubble column. A bootstrap-aggregated hybrid modelling framework is applied for constructing an ensemble of 1000 hybrid models using different partitions of the training/validation data, providing a measure of the parameter distributions and prediction uncertainty. Each model in the ensemble hybrid model has the same model structure relying on first-principles material balances and an Artificial Neural Network for prediction of the liquid phase volumetric mass transfer coefficient. Incremental learning is applied, efficiently enabling the model to adapt to new data acquired during runtime. The software implementation follows recent ISO issues using a modular structure allowing for flexible allocation of server resources and an intuitive User-Interface is developed for controlling the application. From a real-time prediction study, the models using incremental learning are found to have superior performance both at normal operating conditions, when interpolating, and when extrapolating compared to using only the pre-trained model.http://www.sciencedirect.com/science/article/pii/S2772508124000747Digital twinHybrid semi-parametric modellingUncertainty quantificationIncremental learningReal-time predictions
spellingShingle Peter Jul-Rasmussen
Mads Stevnsborg
Xiaodong Liang
Jakob Kjøbsted Huusom
Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration
Digital Chemical Engineering
Digital twin
Hybrid semi-parametric modelling
Uncertainty quantification
Incremental learning
Real-time predictions
title Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration
title_full Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration
title_fullStr Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration
title_full_unstemmed Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration
title_short Implementation of real-time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration
title_sort implementation of real time incremental learning for ensemble hybrid model prediction in pilot scale bubble column aeration
topic Digital twin
Hybrid semi-parametric modelling
Uncertainty quantification
Incremental learning
Real-time predictions
url http://www.sciencedirect.com/science/article/pii/S2772508124000747
work_keys_str_mv AT peterjulrasmussen implementationofrealtimeincrementallearningforensemblehybridmodelpredictioninpilotscalebubblecolumnaeration
AT madsstevnsborg implementationofrealtimeincrementallearningforensemblehybridmodelpredictioninpilotscalebubblecolumnaeration
AT xiaodongliang implementationofrealtimeincrementallearningforensemblehybridmodelpredictioninpilotscalebubblecolumnaeration
AT jakobkjøbstedhuusom implementationofrealtimeincrementallearningforensemblehybridmodelpredictioninpilotscalebubblecolumnaeration