Transfer learning for efficient meta-modeling of process simulations

•An efficient meta-modeling method is developed based on transfer learning.•The transfer learning strategy is based on the Bayesian migration technique.•The effects of different base models are discussed and compared. In chemical engineering applications, computational efficient meta-models have bee...

Full description

Saved in:
Bibliographic Details
Published in:Chemical engineering research & design 2018-10, Vol.138, p.546-553
Main Authors: Chuang, Yao-Chen, Chen, Tao, Yao, Yuan, Wong, David Shan Hill
Format: Article
Language:English
Subjects:
Artificial intelligence
Bayesian analysis
Bayesian inference
Chemical engineering
Chemical processes
Computational fluid dynamics
Computational fluid dynamics (CFD)
Computer simulation
Data acquisition
Data transfer (computers)
Fluid dynamics
Machine learning
Mathematical models
Meta-model
Migration
Model migration
Organic chemistry
Simulators
Training
Transfer learning
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•An efficient meta-modeling method is developed based on transfer learning.•The transfer learning strategy is based on the Bayesian migration technique.•The effects of different base models are discussed and compared. In chemical engineering applications, computational efficient meta-models have been successfully implemented in many instants to surrogate the high-fidelity computational fluid dynamics (CFD) simulators. Nevertheless, substantial simulation efforts are still required to generate representative training data for building meta-models. To solve this problem, in this research work an efficient meta-modeling method is developed based on the concept of transfer learning. First, a base model is built which roughly mimics the CFD simulator. With the help of this model, the feasible operating region of the simulated process is estimated, within which computer experiments are designed. After that, CFD simulations are run at the designed points for data collection. A transfer learning step, which is based on the Bayesian migration technique, is then conducted to build the final meta-model by integrating the information of the base model with the simulation data. Because of the incorporation of the base model, only a small number of simulation points are needed in meta-model training.
ISSN:0263-8762
1744-3563
DOI:10.1016/j.cherd.2018.07.008