Periodically time-varying economic model predictive control with applications to nonlinear continuous stirred tank reactors

Abstract

The economic cost functions can be periodically time-varying due to various economic factors such as the variations of product prices, energy consumptions, operating costs and customer demand changes. These time-varying parameters can lead to instability because the optimal steady state may not be asymptotically stable. In this paper, a periodically time-varying economic model predictive control algorithm is proposed. As the time-varying parameters in the economic cost functions vary, the average constraints are used to enforce the convergence without modifying the original economic cost functions so the transient economic performance can be improved. The time-varying terminal regions are included in the economic optimization problem in order to ensure recursive feasibility. The asymptotic convergence is proved by using the Lyapunov-like analysis. The developed predictive control algorithm is applied to the nonlinear continuous stirred tank reactors in order to gain the highest economic profits.