Robustness Analysis of A Class of MPC Tuning Strategy

The classical Model Predictive Control (MPC) is still considered in many industrial applications, although advanced control methods have seen significant development over the last few years. However, a lot of MPC strategies still suffer from robustness to cope with a variety of process dynamics. For the MPC controller to work effectively, it must be properly tuned. However, MPC is challenging and there is no ideal analytic method to obtain exact solutions that result in the best desired responses. Based on that, this paper investigate the robustness performance for a MPC tuning strategy. Three formulae were derived (Proposed
$\boldsymbol{\alpha}$
, Proposed
$\boldsymbol{\beta}$
and Proposed
$\boldsymbol{\gamma}$
). These formulae are used for calculating the suppression coefficient. The three formulae were derived by fitting the optimal empirical suppression coefficients for varies process dynamics commonly found in industry. Simulation results show that the use of the proposed strategy results in good performance compared to other strategies previously proposed. The proposed strategy also demonstrated a robust performance with respect to modelling errors of process parameters. The results demonstrate the effectiveness and validity of proposed strategy when compared with conventional strategies.