421054 Simultaneous Control of Safety Constraint Sets and Process Economics Using Distributed Economic Model Predictive Control

Tuesday, November 10, 2015: 9:10 AM
Salon G (Salt Lake Marriott Downtown at City Creek)
Anas Alanqar1, Fahad Albalawi2, Helen Durand1, Matthew Ellis1 and Panagiotis D. Christofides3, (1)Chemical and Biomolecular Engineering, University of California, Los Angeles, Los Angeles, CA, (2)Electrical Engineering, University of California, Los Angeles, Los Angeles, CA, (3)Department of Chemical and Biomolecular Engineering and Department of Electrical Engineering, University of California, Los Angeles, Los Angeles, CA

Maintaining safe operation of chemical processes while simultaneously maximizing economic profit and meeting environmental constraints constitute issues of paramount importance in the area of process systems and control engineering.  It has long been argued that process safety is fundamentally a process control problem, yet, few research efforts (e.g., [1]) have been carried out towards the direction of integrating the rather disparate domains of process safety and process control. Economic model predictive control (EMPC) [2] has attracted significant attention recently due to its ability to  optimize process operation accounting directly for process economics considerations. However, there is very limited work on the problem of integrating safety considerations in EMPC to ensure simultaneously safe operation and  maximization of process economic profit.

Motivated by the above considerations, this work will present a distributed economic model predictive control methodology for simultaneously coordinating  in real-time the size of the safety sets in which the process state should reside at all times in order to ensure safe process operation and feedback control of the process state to optimize economics via time-varying process operation. Initially, the safeness of a region of operation with respect to a potential failure (i.e., determine if stability of the process can be maintained for any point in the region of operation if a failure occurs in a sensor or an actuator or a software component) is characterized and a safety logic is created that determines what are the safe regions of process operation under both normal and faulty conditions. Using data on probability of the potential failures, measurement feedback of the process state and future process state trajectory information as determined by the EMPC over its prediction horizon, a control system is built to determine and modify in real-time the region of safe process operation. This information is in turn used as a state constraint in the EMPC that regulates the process state to optimize economics such that the two control systems coordinate their actions. Through a rigorous stability analysis, conditions for recursive feasibility and closed-loop stability are established. The proposed method, which effectively integrates feedback control, process economics and safety considerations, is demonstrated with a chemical process example.

[1] Leveson NG, Stephanopoulos G. A system-theoretic, control-inspired view and approach to process safety. AIChE Journal. 2013;60:2-14.
[2] Ellis, M, Durand H,  Christofides PD. A Tutorial Review of Economic Model Predictive Control Methods.  J. Proc. Contr. 2014;24:1156-1178.

Extended Abstract: File Not Uploaded
See more of this Session: Networked, Decentralized and Distributed Control
See more of this Group/Topical: Computing and Systems Technology Division