Stiction combating intelligent controller tuning: A comparative study

Abstract

This paper investigates the effects of different controller tuning approaches on an intelligent controller, namely Stiction Combating Intelligent Controller (SCIC) earlier developed by the authors. The SCIC controller is inherently a variable gain fuzzy Proportional-Integral (PI) controller based on Takagi-Sugeno model and was specifically designed to handle the stiction nonlinearity in a control loop in presence of a sticky pneumatic control valve. Three different tuning methods, viz. Ziegler-Nichols, Tyreus-Luyben and Direct synthesis (DS) tuning approach, which are extensively employed in process industries to tune Proportional-Integral-Derivative controllers, are tested in this work to find the gains of SCIC and PI controller. The performance of both, SCIC and PI controllers, are rigorously evaluated experimentally on a laboratory scale nonlinear flow process for setpoint tracking, disturbance rejection, and robustness testing. Based on extensive experimental analysis it can be concluded that the SCIC controller tuned using DS approach performed best for almost all cases.