Economic analysis of model predictive control on dimethyl ether purification process

The dimethyl ether product (DME) can be used as an alternative energy that is more environmentally friendly and sustainable. At the DME purification plant, the mixture of DME, methanol and air in the feed will be separated to obtain a pure DME of 99% concentration. Model predictive control is used to control the processes in the DME purification plant using first order plus dead-time models (FOPDT). This study examines the use of MPC technically by comparing it with the control performance of the proportional-integral (PI) controller and analyzed economically. The result, the eight single MPCs (SMPCs) performances are better than the PI controllers in overcoming the disturbance with the decrease of the integral of absolute error (IAE) by 40 % to 96%. Thus, there is a saving of utility (water) used is 0,015%. Since the utility is water, the economic value is low. This led to a considerable payback period (PBP) of 14.5 years, and net present value (NPV) of USD -846 in the feed flowrate disturbance scenario of 5%, so that the use of SMPC in the DME purification plant is economically unfeasible. However, the use of two multivariable MPCs (MMPCs) instead of the eight SMPCs will decrease the PBP and will increase the NPV, therefore, the use of MMPC can be considered for further research.