Optimization of PID controller for flexible link system using a pareto-based multi-objective differential (PMODE) evolution

Abstract

The conflict between the transient performance of the link position and tip vibration in a flexible link system has made the control of such system a challenging task. The system is required to obtain a fast transient position response together with minimal tip vibration. This can be viewed like many other real-life control problems as a multi-objectives optimization problem in which an optimal compromise between the design objectives is required. PID controller is noted with historical simplicity in terms of design and implementation when compares to other linear time invariant (LTI) control techniques. However, the shortcoming of PID lies in the tuning of the controller gains for a given problem. To overcome this, a Multiobjective Differential Evolution (MODE)-based PID controller is reported in this study for controlling a flexible link system. The gains of the PID controller are tuned using a developed MATLAB-based MODE to obtain pareto-solutions for both link position and tip vibration. The performance of the selected best PID controller from MODE-based design is benchmarked with the LQR controller provided by the manufacturer (QUANSER) of the laboratory scale flexible link plant. Though, the LQR shows better transient performance in the position responses, the developed MODE-PID gave better tip response performances as indicated in both the simulation and experimental responses obtained.