Assessment of Input-Output Dead Times Impact on Tuning Procedures and Grouping Policies for Output Soft Constraints in MPC

Model Predictive Control (MPC) is an Advanced Process Control technique able to handle constrained multivariable processes characterized by dead times. In these processes, dead times information covers a key role for control systems design and for the associated tuning procedures. This paper aims to assess the impact of input-output dead times on tuning procedures and grouping policies associated to output soft constraints in MPC based on linear models. An MPC problem characterized by hard constraints on the inputs magnitude and on the inputs slew rate and by soft constraints on the outputs magnitude is considered for the description of the proposed approach. Window parameters, tuning parameters associated to slack variables and grouping policies for constraints relaxation represent the key points investigated in the work. In addition, a characterization of the proposed methods based on different input-output configurations and on the control matrix to be used at each control instant is presented. Simulation results based on first-order plus deadtime models show how the proposed methods can improve the effectiveness and the efficiency of MPC systems design.