Nonlinear model predictive and flatness-based two-degree-of-freedom control design: a comparative evaluation in view of industrial application

Abstract

Two advanced nonlinear model-based controller design methods - nonlinear model predictive control (NMPC) and two-degree-of-freedom (2DOF) control with flatness-based feedforward control design and decentralized PI-controllers - are compared in view of industrial application. The comparison is carried out on a setpoint-transition of the Klatt-Engell benchmark reactor model. Based on an analysis of simulation scenarios, the controllers are compared with respect to controller performance and several robustness criteria. Furthermore, implementation issues of both controllers are discussed. The choice of the control task as well as the tuning of the controllers and the comparison methodology are strongly oriented on industrial practice.