Surrogate submodel generation for an MEA-based post-combustion CO2 capture absorber

The absorber is the core operating unit of the MEA-based post-combustion CO 2 capture process. Research on absorber modeling has been ongoing over the past few decades. However, the mechanistic model based on complex thermodynamics and reaction kinetics is computationally expensive and time-consumin...

Full description

Saved in:
Bibliographic Details
Published in2023 IEEE 18th Conference on Industrial Electronics and Applications (ICIEA) pp. 1603 - 1608
Main Authors Wang, Yurun, Huang, Yi, Ye, Lingjian
Format Conference Proceeding
LanguageEnglish
Published IEEE 18.08.2023
Subjects
CO 2 absorber
Computational modeling
Data models
Electric potential
Hammersley sampling
Industrial electronics
Kinetic theory
neural network
Neural networks
simulation
surrogate model
Thermodynamics
Online AccessGet full text

Cover

More Information
Summary:The absorber is the core operating unit of the MEA-based post-combustion CO 2 capture process. Research on absorber modeling has been ongoing over the past few decades. However, the mechanistic model based on complex thermodynamics and reaction kinetics is computationally expensive and time-consuming. For example, the reaction balance typically needs iterative computations within the rigorous absorber simulator, for the computation of the true component concentrations distributed along the absorber. The objective of this paper is to develop a surrogate submodel for the reaction equilibrium, by establishing an explicit function to replace the equality relationships, the intensive iterations are avoided. The surrogate submodel is trained via the neural networks with data generated from the simulator using the Hammersley sequence. The results show that the surrogate submodel can substantially reduce the computation loads, while maintaining satisfactory accuracies compared with the rigorous simulator.
ISSN:2158-2297
DOI:10.1109/ICIEA58696.2023.10241779