April 20, 2023 Thursday
Machine learning is creating new paradigms and opportunities in the design of advanced process control systems for chemical processes. Traditionally, model predictive control (MPC), a constrained optimization-based control problem formulation that is the gold standard employed in advanced control of chemical processes, is formulated with linear data-based empirical models and is used to compute control actions to maintain optimal process operation while accounting for process and control actuator constraints. However, chemical processes are inherently nonlinear and often require nonlinear models in order to be controlled efficiently. Nonlinear first-principles process modeling provides a direct way for accounting for nonlinear process behavior in the control system design but it is cumbersome and difficult to implement in complex industrial processes which are not well-understood. Machine learning (ML) tools like recurrent neural networks provide an efficient way to build nonlinear dynamic models from data that can be used in the model predictive control system, thereby improving control system performance. In this talk, we will present our research work on the use of ML tools in modeling nonlinear dynamics systems for MPC. Specifically, we will present: a) a general framework of using recurrent neural networks for modeling nonlinear systems within MPC, b) theoretical results on the generalization error of RNN models and closed-loop stability of RNN-MPC using statistical learning theory, and c) novel ML modeling methods including physics-informed ML, online ML, and reduced-order ML modeling to address practical challenges such as data scarcity, model uncertaincies, and curse of dimensionality, respectively. Throughout the talk, we will present applications of our methods to chemical and pharmaceutical processes of industrial interest to demonstrate their applicability, and provide new insights to the development of the next-generation manufacturing paradigm including digitalisation, automation, and digital twin in a more economical and efficient manner.
Prof. Zhe Wu
Assistant Professor, Department of Chemical and Biomolecular Engineering
National University of Singapore
Zhe Wu is currently an Assistant Professor in the Department of Chemical and Biomolecular Engineering, National University of Singapore (NUS). He received his B.S. in Control Science and Engineering from Zhejiang University, China, in 2016 and received his Ph.D. in Chemical Engineering at the University of California, Los Angeles, USA, in 2020. He worked as a postdoctoral researcher in the Department of Computer Science at the University of California, Los Angeles before he joined NUS. His research interests are in the general area of process systems engineering with a focus on process dynamics, optimization and control as well as on data science and machine learning and their application to chemical engineering. He has written more than 50 peer-reviewed journal articles, and received over $SGD 2 million in external research funding from the ASTAR, NRF, MOE and chemical and pharmaceutical companies. Dr. Wu is on the editorial board of Digital Chemical Engineering journal, and is handling two special issues as guest editors.