ePrints

A Bayesian Data Modelling Framework for Chemical Processes using Adaptive Sequential Design with Gaussian Process Regression

Lookup NU author(s): Dr Liam Fleming, Dr Jie ZhangORCiD, Dr Shirley ColemanORCiD

Downloads

• Published version [.pdf]

Licence

This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).

Abstract

Accurate simulators are relied upon in the process industry for plant design andoperation. Typical simulators, based on mechanistic models, require considerableresources: skilled engineers, computational time and proprietary data. This paperexplores the complexities of developing a statistical modelling framework for chemicalprocesses, focusing on inherent non-linearity in phenomena and the difficulty ofobtaining data. A Bayesian approach to modelling is forwarded in this paper, utilisingBayesian sequential design to maximise information gain for each experiment.Gaussian Process (GP) regression is used to provide a highly flexible model class tocapture non-linearities in the process data. A non-linear process simulator, modelledin Aspen Plus is used as a surrogate for a real chemical process, to test the capabilitiesof the framework.

Publication metadata

Author(s): Fleming L, Emerson J, Stitt H, Zhang J, Coleman S

Publication type: Article

Publication status: Published

Journal: Applied Stochastic Models in Business and Industry

Year: 2022

Volume: 38

Issue: 5

Pages: 787-805

Online publication date: 10/08/2022

Acceptance date: 25/07/2022

Date deposited: 27/09/2022

ISSN (electronic): 1526-4025

Publisher: Wiley

URL: https://doi.org/10.1002/asmb.2709

DOI: 10.1002/asmb.2709

Altmetrics

Share