Experimentally validated predictive dual parametric fractional-order tuning approaches for time delayed chemical processes

Abstract

Designing effective controllers for integrating processes with significant time delays remains a challenging task, as conventional control strategies often struggle to maintain stability and performance. To address this, the present work proposes a fractional-order tilt/ tilt-integral–tilt-derivative (FO-T/TI–FO-TD) control framework, built upon the Smith-predictor structure, for both time-delayed integrating and stable chemical processes. The proposed architecture consists of a primary controller for set-point tracking and a secondary controller for load disturbance rejection. The controller gains are determined analytically using stability margins and maximum sensitivity ($M_s$) criteria. The effectiveness of the design is evaluated using several benchmark chemical process models, including robustness tests with a 20% perturbation in plant parameters. The results demonstrate improved servo performance and disturbance rejection, with the added advantage of requiring only two tuning parameters. Furthermore, the proposed scheme is validated in real time on a two-tank level control system, confirming its practical applicability.