Optimizing process economics and operational safety via economic MPC using barrier functions and recurrent neural network models

Abstract

In the present work, a control Lyapunov-barrier function (CLBF)-based economic model predictive control (EMPC) system is designed to optimize process economics, and ensure stability and operational safety simultaneously based on a prediction model using an ensemble of recurrent neural network (RNN) models. As accurate first-principles models are not available for many industrial processes, RNN models are utilized in this work to approximate the dynamics of a general class of nonlinear systems in an operating region. The ensemble of RNN models are incorporated in the design of CLBF-EMPC, under which guaranteed closed-loop stability and process operational safety are achieved for the nonlinear systems with two types of unsafe regions, i.e., bounded and unbounded sets. The application of the proposed RNN-based CLBF-EMPC method is demonstrated through a chemical process example with the case studies of a bounded and an of unbounded unsafe region, respectively.