Optimal Control of Carbon Dioxide Exchange Process in a Membrane Oxygenator Using Particle Swarm Optimization Approach

Abstract

The aim of this study is to evaluate the performance of Zeigler-Nichols continuous cycling and particle swarm optimization (PSO) method in tuning the optimal gains for Proportional-Integral-Derivative (PID). PID controller is implemented to control the rate of CO2 elimination from a membrane oxygenator during extracorporeal blood purification process. The sweep gas flow rate is chosen as the manipulated variable to control arterial CO2 partial pressure (pCO2) in blood. The Zeigler-Nichols continuous cycling tuning method is employed for tuning purpose and the performance of each controller (P-only, PI and PID) are evaluated based on three performance indices, namely integral absolute error (IAE), integral squared error (ISE) and integral time absolute error (ITAE). Next, the optimization algorithm known as PSO is used to calculate the gain parameter that can produce the best control action. The robustness of these tuning methods is assessed for set point tracking and load disturbance rejection tests. Results indicated that the PID is seen as the best controller compared to the classical controllers such as P and PI when Zeigler-Nichols continuous cycling as the tuning method is implemented. However, further tests highlighted the PSO-PID strategy (PID parameters that are optimized by PSO) showed even better control responses compared to PID alone. Thus, it can be concluded that optimization strategy by PSO method is the best tuning method to be used in determining the controller parameters for the automation of extracorporeal circulation control for both set point tracking and load disturbance rejection tests.