Integrating Dynamic Economic Optimization and Model Predictive Control for Optimal Operation of Nonlinear Process Systems

Economic optimization of chemical processes has traditionally been addressed through real-time optimization (RTO). In general, an RTO system carries out economic optimization and computes optimal process operation steady-states using steady-state process models. These steady-states are used by the feedback control systems, typically designed via model predictive control methods, that force the process to operate on these steady-states. In recent years, numerous calls for the development of the so-called “smart manufacturing paradigm” have led to several attempts to integrate MPC and economic optimization of chemical processes to deal with variable demand, energy prices, variable feedstock and product transitions. In previous work [1], we developed a Lyapunov-based economic MPC (LEMPC) method capable of optimizing closed-loop performance with respect to general economic consideration for nonlinear systems. In the present work, we propose a conceptual framework for integrating dynamic economic optimization and MPC for optimal operation of nonlinear process systems. First, we introduce the proposed integrated framework. This optimal operation strategy consists of an LEMPC that preforms upper layer dynamic economic optimization to compute optimal process operating time-varying trajectories which are used by the lower-level control layer that may use conventional MPC schemes or even classical control. Such a framework takes advantage of the LEMPC ability to compute optimal process operation time-varying policies using dynamic process model instead of traditional steady-state models. Second, we prove practical stability of the proposed integrated dynamic economic optimization and control framework including an explicit characterization of the stability region of the closed-loop system. Third, we demonstrate through extensive simulations using realistic chemical process models that such an integrated control paradigm can both (1) achieve stability and (2) perform economically better than traditionally RTO systems using steady-state models. 1. Heidarinejad, M., J. Liu and P. D. Christofides, "Economic Model Predictive Control of Nonlinear Process Systems Using Lyapunov Techniques,'' AIChE J., 58, 855-870, 2012.