Implementation of first order Model Reference Adaptive Control (MRAC) on regulating temperature of essential oil extraction process

This paper presents the performance of proportional-integral-derivative (PID) controller and Model Reference Adaptive Controller (MRAC) on regulating essential oil extraction process. Model of Auto-Regressive with Exogenous Input (ARX) from previous research on real time experiment data was used to represents the system dynamic of heating process. Heating process is important thing in the essential oil extraction and became challenging for industrial to control the temperature at the desired temperature. Besides, quality of essential oil will reduce if the extraction process expose to overheating of temperature. Unsuitable temperature regulation on extraction oil such as overshoot will give a large settling time and affected steady state response. Generally, conventional PID controller that commonly used in control system is not provided a desired output performance. Implementation of MRAC by using Lyapunov approach in the system gives better response than conventional PID controller. Model reference of the system was developed based on first order plus dead time (FOPDT). Parameter gains of PID were tuned automatically by PID tuner based on system performance with varied time response while adaptation gain for MRAC system chosen based on the best system performance by manually inserted into adaptation mechanism in simulation. In addition, adaptation gain ±0.1 is the best selection in adaptation mechanism for MRAC system. Result shows that MRAC controller was able to minimize its overshoot and robust in performance better than conventional PID controller in step test and set point tracking test. In addition, conventional PID controller provided faster time taken in rise time, 1760 second to initially reach the set point compared to MRAC, 2500 second but MRAC takes in short time to settle the response, 4460 second compared to PID, 7220 second. The comparison of output response between both controllers was obtained by using transient analysis and performance indices in MATLAB/Simulink.