Fuzzy logic based control for membrane oxygenator in blood purification process / Hany Hazfiza Manap

Hany Hazfiza , Manap (2020) Fuzzy logic based control for membrane oxygenator in blood purification process / Hany Hazfiza Manap. PhD thesis, Universiti Malaya.

	PDF (The Candidate's Agreement) Restricted to Repository staff only Download (241Kb)
Preview	PDF (Thesis PhD) Download (9Mb) | Preview

Abstract

Membrane oxygenator in Extracorporeal Life Support (ECLS) is commonly used as a main component in Cardiopulmonary Bypass (CPB), Extracorporeal Membrane Oxygenator (ECMO) and Extracorporeal Carbon Dioxide Removal (ECCO2R). In this study, the focus is on Carbon Dioxide (CO2) gas exchange process in membrane oxygenator, which the CO2 is automatically controlled by adjusting sweep gas flow rate that entering the membrane oxygenator in achieving the desired pCO2 required by the particular patients. This automatic control action is controlled by the implementation of three controllers, Proportional-IntegralDerivative (PID) that is tuned by Ziegler-Nichols continuous cycling method, Fuzzy Logic Controller (FLC) and Fuzzy-PID controllers. This study is conducted in two stages: simulation and experimental work for two tasks, namely setpoint tracking and disturbance rejection. For evaluation of setpoint tracking, desired setpoint of arterial partial carbon dioxide (pCO2) is set at 40 mmHg at the beginning of the process, and been increased by 5% to 42 mmHg at the middle of the process, and been adjusted back to 40 mmHg at the end of the process. For disturbance rejection, an external disturbance is applied to the process plant (in membrane oxygenator) by a sudden shutoff of mass flow controller (MFC) to stop the sweep gas flow rate until 0 L/min at the middle of process. All the simulation works are done using level-2 SFunction in MATLAB/SIMULINK environment. An application of NI-USB 6009 DAQ in corporation with MATLAB/SIMULINK environment is used for both pCO2 sensors and mass flow controller as an actuator to control sweep gas flow rate in the experimental work. FuzzyPID recorded the best performance in both setpoint tracking and disturbance rejection tasks during simulation. The control performance become worst with loss of control action in experimental setting using similar PID and FLC parameters/rules, due to some assumptions and defined parameters in mathematical modelling that are differ than actual parameters used in experimental work. Re-tune process is conducted to calculate newly parameter for PID and rule for FLC and Fuzzy-PID. The proposed controllers then able to automatically control CO2 gas transfer in membrane oxygenator according to the desired setpoint. As a conclusion, all the controllers produce a good ability to automatically control CO2 removal from membrane oxygenator for both tasks, with Fuzzy-PID outshines the other controllers, PID and FLC. This can be seen from the lowest performance indices and the time taken to achieve the setpoint for Fuzzy-PID is lower by 21.15% compared to PID and 65.25% lower compared to FLC during setpoint tracking task. For disturbance task, Fuzzy-PID also recorded the fastest time to adapt with the disturbance, which 89.33% and 51.52% faster than PID and FLC, respectively. Hence, this in-vitro bench set-up has great potential to be further extended study.

Actions (For repository staff only : Login required)