Real Time Implementation Of Nonlinear Autoregressive With Exogenous Input Model Predictive Control For Batch Enzymatic Esterification Process

Zulkeflee, Siti Asyura (2017) Real Time Implementation Of Nonlinear Autoregressive With Exogenous Input Model Predictive Control For Batch Enzymatic Esterification Process. PhD thesis, Universiti Sains Malaysia.

Download (392kB) | Preview


The lipase catalysed esterification process is an important process in the food and pharmaceutical industry. Obtaining the optimum production for the esterification process is a big challenge due to numerous factors that affect the kinetics of the process. In this work, the MPC was designed and implemented to control the temperature and water activity of the lipase-catalysed esterification process. Prior to that, a kinetic model that followed an ordered Bi-Bi mechanism was developed to study the function of water activity and temperature. The kinetic parameters were estimated using the interp function in MATLAB® software. Then, the first principle model was developed and validated with the experimental data. The first principle model was solved using the 4th order Runge-Kutta method (ode45) by means of a Differential Equation Editor (DEE) block diagram developed using the MATLAB® software. The developed model showed a strong predictive capability to represent the real process. The validated first principle model was then used to study sensitivity and nonlinearity as well as to generate the input/output data for an empirical model. Based on the sensitivity study, it was found that the input variables, i.e. jacket flowrate, jacket temperature, and air flowrate, have significant effects on the output variables, i.e. reactor temperature and water activity. The nonlinearity study showed that the lipase-catalysed esterification process can be classified as a nonlinear process. The objective of the MPC control strategy was to control the reactor temperature and water activity of a batch esterification reactor. The empirical model, which was embedded in the MPC was developed using the Autoregressive with Exogenous input (ARX) and Nonlinear Autoregressive with Exogenous input (NARX) models and were known as the ARX-MPC and NARX-MPC, respectively. The parameter estimation and model validation for the empirical model were carried out using the recursive least squares estimation (RLSE) system identification toolbox in MATLAB®. The results showed that the NARX models fit the real data very well when compared to the ARX models. The MPC parameters were tuned to determine the best controller performance. The best-tuned ARX-MPC and NARX-MPC controllers were compared and evaluated in terms of set point tracking and disturbance rejection. The ISE results achieved in this study showed that the developed NARX-MPC fitted satisfactorily with the control system and it had outperformed the ARX-MPC controller. Additionally, the NARX-MPC was found to be more robust than the ARX-MPC in a robustness study. Finally, the NARX-MPC controllers were chosen and tested in real-time implementation. The results showed that the NARX-MPC was effective in controlling the temperature and water activity of the process in a real-time environment.

Item Type: Thesis (PhD)
Subjects: T Technology
T Technology > TP Chemical Technology > TP155-156 Chemical engineering
Divisions: Kampus Kejuruteraan (Engineering Campus) > Pusat Pengajian Kejuruteraan Kimia (School of Chemical Engineering) > Thesis
Depositing User: Mr Mohamed Yunus Mat Yusof
Date Deposited: 11 Feb 2020 07:45
Last Modified: 17 Nov 2021 03:42

Actions (login required)

View Item View Item