Comparative Study of Genetic Algorithm Optimized FO-PID and LQR Control Strategies Applied to a Piston Pump in a Volume Calibration System

Abstract: One of the key responsibilities of Getinge's ventilators is to deliver accurate gas volumes to patients. To ensure this precision, specially designed rigid steel tanks are utilized to evaluate the performance and precision of the ventilators in providing exact air volume. The intention of this study is to design and implement a suitable controller for actuating a servo piston pump in order to be used for the tank volume measuring and calibration process. Two controlling strategies were chosen for this purpose: Linear Quadratic Regulator (LQR) and Fractional Order Proportional Integral Derivative (FOPID). This work also aimed to establish a narrative of the two controlling strategies after optimizing them using genetic algorithm optimization (GA) and evaluating their effectiveness in controlling a brushless DC motor (BLDC) actuating a servo piston pump. This involved modeling the system in Matlab and Simulink based on the mathematical representations of the system's dynamics, specifically focusing on its pneumatic behavior. The nonlinear model was linearized and served as a basis for the controllers' optimization through the genetic algorithm. Both controller designs were then compared in both the Simulink environment and the actual physical system. The results show that the FOPID exhibits superior performance in the Simulink environment. Contrariwise, the LQR displays a far greater level of superiority in the physical system, whereas the FOPID performance significantly deteriorated upon implementation in the physical system. Furthermore, the study suggests implementing anti-windup techniques and ensuring the accurate digitization of fractional calculus for further research to enhance the performance of the FOPID controller on the physical system.