(334v) Automated Optimization and Control of Modular Chemical Systems

Modular chemical systems for advanced manufacturing have attracted increased interest in recent years due to their ability to employ process intensification, as well as their potential for increased flexibility, and productivity (e.g., see [1,2] and citations therein). This interest is particularly prominent in the chemicals, pharmaceuticals, biotechnology, and energy industries (e.g., [3-5]). This Ph.D. thesis addresses modular chemical systems from a process optimization and control perspective that enables automation. Modular chemical systems are advanced manufacturing systems with characteristics such as (1) dynamics described by combinations of algebraic, ordinary differential, partial differential, and integral equations, (2) high to infinite state dimension, (3) time delays, (4) actuator, state, and output constraints, (5) stochastic noise and disturbances, (6) parametric uncertainty. Most methodologies are developed to handle dynamical systems with a subset of these characteristics and a focus on the specifics of the process.

However, to promote efficiency in industrial practice, autonomous design of optimization and control for modular chemical systems, connected in a plug-and-play manner, should be pursued. Such control and optimization systems are mainly conducted through a computer with little to no supervision. In the work conducted in the context of this Ph.D. thesis, we describe steps taken towards such automation, through modeling, control algorithms and optimization approaches tailored to address the characteristics and challenges of such systems. Such steps include (1) automating the construction of dynamic first-principles plantwide models obtained by plug-and-play interconnection of modules that are modeled in advance; (2) automating the hierarchical plant-wide control; (3) developing methodologies to expand the range of applicability of linear input-output models in linear model predictive control, to address the high state dimension of modular systems with low on-line computational cost [6]; (4) addressing the optimization of transient periods (i.e. startup) to minimize waste and increase efficient operation times, through hybrid dynamic optimization and linear model predictive control schemes [7]; (5) creating feedback controllers with guaranteed properties for dynamic neural network models, as an alternative to first-principles model control [8]. This Ph.D. work builds upon recent publications that propose ways to modify linear model predictive control formulations to handle significant process nonlinearities [6], with applications that extend to process startup [7]. It also demonstrates progress in designing controllers with guaranteed properties for dynamic artificial neural networks [8], which can be used in the place of first-principle models, providing an alternative to nonlinear model predictive control, which usually is a nonconvex problem that requires the solution of an NLP online, with unknown feasibility and performance properties.

Interests I developed during my thesis work are related to systems modeling, simulation, and control. My current research interests include, but are not limited to, first principles modeling, and simulation, empirical modeling of dynamical systems using data, surrogate modeling, dynamic optimization, model predictive control of high dimensional systems, and autonomous design of advanced control systems. I am also interested in expanding the above methodologies to address uncertainties in systems, such as parametric uncertainties and model-plant mismatch. My research interests find applications in numerous fields, such as in the (bio)pharmaceutical industry which is rapidly moving towards continuous processing, in the chemical industry where modeling and control are always relevant, as well as in the energy industry, with the evolving subfield of battery modeling and control.

References

1. -C. Bédard, A. Adamo, K. C. Aroh, M. G. Russell, A. A. Bedermann, J. Torosian, B. Yue, K. F. Jensen, and T. F. Jamison. Reconfigurable system for automated optimization of diverse chemical reactions. Science 361(6408):1220-1225, 2018.
2. A. Paulson, E. Harinath, L. C. Foguth, and R. D. Braatz. Control and systems theory for advanced manufacturing. In Emerging Applications of Control and System Theory, edited by Roberto Tempo, Stephen Yurkovich, and Pradeep Misra, Lecture Notes in Control and Information Sciences, Springer Verlag, Chapter 5, 63-80, 2018.
3. Lakerveld, P. L. Heider, K. D. Jensen, R. D. Braatz, K. F. Jensen, A. S. Myerson, and B. L. Trout. End-to-end continuous manufacturing: Integration of unit operations. In Continuous Manufacturing of Pharmaceuticals, edited by P. Kleinebudde, J. Khinnast, and J. Rantanen, Wiley, New York, Chapter 13, pages 447-483, 2017.
4. T. Myerson, M. Krumme, M. Nasr, H. Thomas, and R. D. Braatz. Control systems engineering in continuous pharmaceutical processing. Journal of Pharmaceutical Sciences, 104(3):832-839, 2015.
5. S. Hong, K. A. Severson, M. Jiang, A. E. Lu, J. C. Love, and R. D. Braatz. Challenges and opportunities in biopharmaceutical manufacturing control. Computers & Chemical Engineering, 110:106-114, 2018.
6. Nikolakopoulou, M. von Andrian, and R. D. Braatz. Plantwide control of a compact modular reconfigurable system for continuous-flow pharmaceutical manufacturing. 2019 American Control Conference (ACC), Philadelphia, PA, USA, pp. 2158-2163.
7. Nikolakopoulou, M. von Andrian, and R. D. Braatz. Fast model predictive control of startup of a compact modular reconfigurable system for continuous-flow pharmaceutical manufacturing. 2020 American Control Conference (ACC), Denver, CO, USA, in press.
8. Nikolakopoulou, M. S. Hong, and R. D. Braatz. Output feedback control and dynamic artificial neural networks using linear matrix inequalities. Submitted.