(537h) Model Predictive Control of Integrated Energy and Chemical Manufacturing Systems

Integrating energy and chemical production is a promising synergetic approach to increase sustainability and resiliency in energy-intensive chemical manufacturing systems [1-3]. Such a combination gives the flexibility of using chemicals as an energy storage medium to suppress electricity supply and demand fluctuations [1]. Excess electricity can be converted into green electrofuels used as energy sources during high electricity demand periods. One of the most critical challenges in automation and control of this system of systems is the complexity associated with the integration [4, 5], where model predictive control (MPC) appears to be a promising approach to address this issue due to its relative conceptual simplicity, flexibility, robustness, and ability to efficiently handle complex multivariable systems with the hard path and terminal constraints [6, 7].

An integrated energy and chemical manufacturing system can be considered as a network of combined lumped parameter systems (LPSs) (e.g., well-mixed reactors, staged separators, compressors, turbines), which can be described by ordinary differential equations (ODEs) and distributed parameter systems (DPSs), which are described by PDEs (e.g., heat exchangers, plug-flow reactors, packed beds). For such a system of systems, the MPC’s solvability becomes more crucial because the underlying optimization problem must be solved in the presence of PDE constraints. To bypass this issue, an on-demand data-assisted model order reduction framework is developed to construct low-dimensional reduced order models (ROMs) in the form of ODEs, which can approximate the spatiotemporal dynamics of the governing PDEs. The resulting discrete ROMs are then combined with the governing ODEs of the LPSs and used as the basis for a combined MPC and moving horizon estimation (MHE) design, where MHE estimates the unmeasurable states and unknown parameters required by MPC using the feedback from the measurement sensors. Note that the closed-loop performance of the proposed control architecture hinges on the accuracy of the ROM. To preserve the ROM accuracy, a supervisory structure is employed to monitor the controller performance and revise the ROM as needed. The proposed aperiodic supervisory strategy employs an event-triggered method where sensing, revision command, and the associated computation are only performed on-demand to save measuring and computational costs. The event-triggered method is reactive and revises the ROM when, for instance, the plant state deviates more than a certain threshold from the desired value. An integrated gasification and solid oxide fuel cell combined cycle, which integrates a gasifier with an air separation unit, a gas turbine, and a solid oxide fuel cell cycle, is used as a case study to illustrate the application and computational advantages of the proposed estimation and control strategy.

[1] P. Daoutidis, A. Allman, and M.J. Palys. Smart manufacturing: A sustainable energy perspective. Smart Manufacturing, 2020.

[2] M. El-Halwagi. Sustainable Design through process integration. Elsevier, 2010.

[3] N. Pistikopoulos, M.C. Georgiadis and V. Dua. Energy systems engineering. WILEY-VCH, 2008.

[4] P. Daoutidis, A. Allman, S. Khatib, M. Moharir, M.J. Palys, D.B. Pourkargar, W. Tang. Distributed decision making for intensified process systems. Current Opinion in Chemical Engineering, 2019; 25:75-81.

[5] P. Daoutidis, M. Zachar and J.S. Jogwar. Sustainability and process control: A survey and perspective. Journal of Process Control, 2016; 44:184-206.

[6] Q. Mayne. Model predictive control: Recent developments and future promise. Automatica, 2014; 50(12):2967-2986.

[7] J.B. Rawlings, D.Q. Mayne, and M.M. Diehl. Model Predictive Control: Theory, Computation, and Design. 2nd Edition, 2019.