Robust predictive controller for nonlinear uncertain process based on S-PARAFAC Volterra models

Abstract

This paper proposes a new approach for synthesizing a robust predictive controller for nonlinear uncertain process by exploiting a reduced complexity discrete-time Volterra model known as S-PARAFAC-Volterra model applied to the quadratic case. This proposed robust control copes with physical constraints and geometrical constraints due to parameter uncertainties and leads to the min-max optimization problem. The major advantage is that the objective function is convex with respect to the parameter uncertainty set which simplifies the min-max optimization problem and minimizes the worst-case value of the objective function taken over the set of uncertain models. A quadratic criterion is optimized and a new optimization algorithm, formulated as a quadratic programming (QP) under linear and nonlinear constraints, is proposed. The developed nonlinear robust control algorithm for uncertain process is illustrated on a benchmark as a continuous stirred-tank reactor system (SCTR).