Production scheduling in dynamic real-time optimization with closed-loop prediction

Abstract

Process plants are operating in an increasingly dynamic environment, fueled largely by globalization and deregulation of energy markets, resulting in fluctuating market conditions and large variations in electricity prices. Such conditions pose challenges for traditional hierarchical plant decision-making systems, leading to efforts toward integration across the decision-making layers. This paper proposes a formulation for integration of production scheduling decisions within a dynamic real-time optimization (DRTO) framework. The DRTO formulation utilizes a closed-loop prediction of the plant response under the action of constrained model predictive control (MPC). The integrated scheduling and DRTO system communicates decisions to the underlying MPC system through time-varying set-point trajectories, thereby permitting the standard MPC implementation to be retained. The efficacy of the prosed system is illustrated through application to both single-input single-output (SISO) and multi-input multi-output (MIMO) case studies.