演講快訊

2024
03.29

Physics-Inspired State Space Model Applications of Chemical Engineering: Process Control

康嘉麟 教授

時間:2024年03月29日 (五) 15:10
主持人:許蘇文 教授
講者:康嘉麟 教授
服務單位:雲林科技大學 化學與材料工程學系
講題:Physics-Inspired State Space Model Applications of Chemical Engineering: Process Control
地點:93406

摘要 :

Traditionally, dynamic prediction in chemical processes has required the development of first-principle models followed by parameter adjustment to fit the data from actual operations as closely as possible. However, developing first-principle models is extremely cost-intensive because of the difficulty in gathering critical domain knowledge. The application of state-space models in the field of chemical engineering has sparked significant interest, especially in the area of chemical process control. Recent advancements in state-space models have been notable, propelled by integrating machine learning techniques, particularly Autoencoders, sequence-to-sequence neural networks, and Transformers.

State-space models have extensive applications in chemical engineering, especially in process control and monitoring. These models, capable of representing complex systems with inputs, outputs, and state variables, offer a robust framework for understanding the dynamic behavior of chemical processes. The fusion of physics-inspired and data-driven approaches within state-space models has further amplified their ability to portray and predict the operational states of chemical processes accurately. This study reviewed developments in applying state-space models focusing on chemical process control and highlighted our recent efforts that employ these models for improved process understanding and control.

The research discusses the methods for constructing soft instruments using state-space models at different sampling frequencies and how to integrate the gain directionality of the physical interpretability into the models. It also discusses developing data-driven digital twin models using state-space models, allowing them to replace traditional first-principle modeling and achieve interactive digital twin models.

學經歷 :

EDUCATION
NATIONAL TSING HUA UNIVERSITY                                        
Ph.D. in Chemical Engineering                                                                  09/2011-09/2016
Academic Advisor: Dr. Shi-Shang Jang
Dissertation: Modeling Rotating Packed Bed on Absorption of CO2 by Aqueous Amines and Ammonia solutions

TUNG HAI UNIVERSITY                                                                    
M.S. in Chemical and Materials Engineering                                            09/2008-06/2010
Academic Advisors: Dr. Hong-Wei Yen
Thesis: Lactic acid production directly from starch in a starch-controlled fed-batch operation using Lactobacillus amylophilus

TUNG HAI UNIVERSITY                                                                   
B.S. in Chemical Engineering                                                                    09/2004-06/2008

 

PROFESSIONAL EXPERIENCE      
NATIONAL YUNLIN UNIVERSITY OF SICENCE AND TECHNOLOGY 
Associate Professor                                                                                    08/2021-Present
Main research:
*    Application of Artificial Intelligence (AI) on Chemical process
*    CFD Simulation and Industrial Scale-Up Modeling
*    Process System Engineering

NATIONAL YUNLIN UNIVERSITY OF SICENCE AND TECHNOLOGY 
Assistant Professor                                                                                    08/2019-07/2021

TAMKANG UNIVERSITY                                                    
Assistant Professor                                                                                    08/2017-07/2019

NATIONAL TSING HUA UNIVERSITY                                          
Postdoctoral Research Fellow                                                                   09/2016-07/2017

THE UNIVERSITY OF TEXAS AT AUSTIN                                           
Visiting scholar(Graduate Student Study Abroad Program)                     07/2015-07/2016