Design of Multi Model Fractional Controllers for Nonlinear Systems: An Experimental Investigation

Abstract

Abstract Control of nonlinear processes is challenging when compared to those of linear processes. Conventional PID controllers can be used to control nonlinear processes based on a condition that these controllers must be tuned in such a way that they should provide a very stable behaviour over the entire range of operating conditions. Inspite of tuning the PID controllers, it still results in a degradation of the control system performance. In order to stabilise nonlinear systems, multi model control approaches are found suitable. Multi-Model Approaches (MMA) rely on the problem decomposition strategy where a nonlinear system is segregated into set of many linear models based on their operating points. These reduced models are used to tune model based PI/PID controllers. In the literature, different methods are proposed to design these controllers in MMA framework and they are all integer order controllers. In this work, fractional controllers based MMA framework is developed for enhanced control of nonlinear systems. Gap metric based weighting methods are used with proper weighting functions to obtain the global controller. The weighting functions are formed by using gap metric to combine local controllers. Three different nonlinear processes (CSTR, conical tank and spherical tank processes) are considered and implemented the proposed methodology. For the purpose of comparison, MMA framework with integer order controllers is considered and it is observed that MMA framework with fractional controllers provide improved closed loop performances. Experimental investigation is also carried out to verify the applicability of the proposed method for level control on conical tank process and it is observed that the proposed method provide enhanced closed loop responses.