Grants and Contributions:

Grant or Award spanning more than one fiscal year. (2017-2018 to 2022-2023)

Model predictive control (MPC) rapidly became the advanced control method of choice in the chemical process industry after its introduction about three decades ago, and has since been widely adopted across a wide range of industrial sectors. The use of an economic performance objective has given rise to so-called economic MPC (EMPC) algorithms that may be applied either as a single-level control algorithm, or in a cascade manner with EMPC providing set-points to an inner regulatory MPC system. MPC relies on a dynamic model for prediction of future process behavior in order to calculate the control input. No model is perfect, and depending on the level of uncertainty, poor control performance can result if the uncertainty is not accounted for in the control calculation, particularly with regard to process constraint violations. This has led to the development of several robust MPC strategies over the past several years. However, these methods for the most part do not rigorously capture the closed-loop evolution of the process under the assumed uncertainty.

The primary objective of the proposed research program is to develop robust regulatory and economic MPC strategies based on uncertainty propagation through use of rigorous closed-loop prediction under the assumed uncertainty. The research will utilize a recent formulation for the closed-loop response of a plant under constrained MPC developed within our group in the context of dynamic real-time optimization (DRTO). By more accurately representing the uncertainty propagation, this approach would potentially provide superior performance over current robust MPC methods. The research program will involve the investigation of a number of areas consistent with the overall program objective. This will include performance evaluation over selected alternative robust MPC paradigms, assessment of the computational effort required and methodologies for reducing the computational load, its use in single-layer and cascade EMPC structures, performance analysis under various uncertainty types and characterizations, application to nonlinear MPC, and stability analysis. The efficacy of developed algorithms will be evaluated through application to several case studies, ranging from relatively simple processes to isolate specific features of the algorithm, to more complex, industrially-oriented case studies. Due to the ubiquitous application of MPC technology and its role in the economics of process operation, the potential impact of the proposed research is significant.

The research involves elements of control, optimization, numerical computation, dynamic modeling, and simulation. HQP trained during the course of the research will consequently be exposed to a range of process systems technologies that will prepare them to make high-level technical contributions to the sectors in which they will be employed.