Performance evaluation for linear based model predictive control in controlling production rate and reactor temperature of ethylene glycol reactor

Abstract

Process Automation Systems are widely known to be a crucial element in processing plants worldwide. In an environment that is ever-changing, process automation is applied in industries when possible to mitigate complexities with advance process control (APC) systems specifically leading to higher efficiency, less operator interaction and increased profits. APC is a proven control and optimization technology delivering measurable and sustainable improvements in production yield, coupled with the added value of energy savings. Many researches focused on modelling the Ethylene glycol (EG) production process each with distinct reactors and reactions. On the other hand, the advances on its control system were poor. Thus, this research fills the gap on the implementation of APC for the EG reactor. In this work, a linear based model predictive control (LMPC) is developed and implemented in EG reactor. The aim of LMPC is to control the production rate and reactor temperature for an optimized hydrogenation reactor. The optimal set point used is based on multi-objective optimization approach using £-constraint method. A state-space model is used in the prediction control is constructed with the best fit of 91.51% and 83.34% for EG production rate and reactor temperature, respectively. The LMPC is developed in Simulink and tuned via the MPC design toolbox in Matlab. The LMPC performance was evaluated based on the analyses results obtained from set point tracking, disturbance rejection and robustness test. The results showed that the LMPC can operate efficiently under tight boundary constraint. This sums up that the LMPC can control both control variables under strict boundary conditions.