Iterative tuning of modified uncertainty and disturbance estimator for time-delay processes: A data-driven approach.

Uncertainty and disturbance widely exist in the process industry, which may deteriorate control performance if not well handled. The uncertainty and disturbance estimator (UDE) emerges as a promising solution by treating the external disturbances and internal uncertainties simultaneously as a lumped term. To overcome its limitation caused by time delay, a modified UDE (MUDE) has been proposed recently. However, its parameter tuning relies heavily on trial-and-error, thus being time-consuming in balancing the robustness and performance. To this end, this paper aims to develop an automatic tuning procedure for the MUDE-based control system. The quantitative relationship between system performance and the scaled parameters is empirically built. Iterative Feedback Tuning (IFT) is utilized to approximate the nominal model towards actual process. Through the empirical formula and optimized model, an automatic design procedure is proposed after taking into account the system robustness and output performance simultaneously. Simulation results show the superiority of the closed-loop performance over the original MUDE controllers. The experimental results validate the feasibility of the method proposed in this paper, depicting a promising prospect in the practical application.