Design of robust MPC with integral action for a laboratory continuous stirred-tank reactor

Abstract

The paper presents a design of the robust model predictive control (RMPC) for a laboratory continuous stirred-tank reactor (CSTR). A neutralization reaction ran in the CSTR, and the reactants were acetic acid and sodium hydroxide. The controlled variable was pH of the reaction mixture. The control input was the volumetric flow-rate of the base. The system was modeled using experimental data of several step-responses. Measurement noise was reduced using the Hebky filter. The robust model-based control strategy was implemented to assure good control performance. The offset-free reference tracking was ensured by the implementation of RMPC with integral action.