Derivative Based Kalman Filter and its Implementation on Tuning PI Controller for the Van de Vusse Reactor

This work focusses on predictive PI (PPI) control law employing on a class of stable, nonlinear benchmark process. To facilitate controller parameter(s) updation, two different types of derivative based Kalman filter (KF) strategies like Extended Kalman filter (EKF) and Ensemble Kalman filter (EnKF) techniques were taken into consideration. The servo-regulatory performance of the PPI controller was found satisfactory even in presence of white Gaussian noise. From the extended simulation studies, it can be inferred that EnKF-PPI control logic implemented on the nonlinear dynamical systems are having slightly better performance over EKF-PPI control law. Demonstration and practical utility of the PPI control method in the presence of process uncertainty or process-model mismatch scenario have also been investigated in this work.