Enhanced Design of a PI-PD Based Smith Predictor for Industrial Plants
G. Lloyds Raja
Abstract
A novel design of a Smith predictor with an interior-loop PD controller and outer-loop PI controller is presented. A set of challenging unstable/integrating plants with dead-time and positive zero is considered. For this class of plant models, parameters of PD controller are derived using Routh stability criteria and maximum sensitivity considerations. For the stabilized plant model, a PI controller is designed by matching derivatives of actual and expected closed-loop transfer functions. Guidelines for adjusting the speed and robustness of the control action are provided. Computer simulations are conducted using some standard plant transfer functions. Quantitative performance evaluation is carried out to illustrate the advantages of the suggested work over some contemporary strategies. The robustness of the design is studied with the help of some perturbed plant models. Though the proposed method is simple, it yields improved closed-loop control action than some reported works.