Model predictive control with exogenous auto-regressive model to improve performance in the CO2 removal

Model predictive control (MPC) is used in the CO2 removal process Subang field to improve its performance. MPC maintain concentration at sweet gas output by controlling feed pressure (PIC-1101), makeup water flow rate (FIC-1102), and amine (FIC-1103). The empirical model applied represent auto-regressive exogenous (ARX) model. ARX compared with first order plus dead time (FOPDT) based on root mean square error (RMSE) between actual process, then parameters are tuned which include sampling (T), prediction horizon (P) (M) for three variables. Improved performance measured integral (ISE). results show that best an RMSE value of 35%-91% smaller than FOPDT optimal Prediction Horizon (P), Control Sampling Time (T) 75, 25 1 PIC-1101, 25, 10 FIC-1102, 30, FIC-1103. MPC-ARX (MPC using model) can 33% servo 6-56% regulatory control. However, not all them showed increase improvement from previous studies even though they had (ARX). This due parameter setting yet appropriate, so it needs be retuning.