Optimal Predictive Control for a Pressure Oscillation Adsorption Process to Produce Bioethanol

Abstract

The PSA (Pressure Swing Adsorption) is considered ahighly non-linear process due to its cyclical nature and the interaction between variables. The control of this process is a challenging feature. This work focus on the design of an Optimal Predictive Control for the PSA process, since there are disturbances that alter the purity of the product (specified at 99% ethanol), The proposed control rejects certain (combined) disturbances in the input variables. The performance and stability of the controller is such as in simulations for temperature changes (disturbances), pressure drops and increases of valve opening time Asaresult, it was deonstrated that the designed control aids to maintain the desired purity (99% ethanol) in a smaller number of cycles, on the other hand, when applying the disturbances in open loop purity falls so far that it does not meet international purity standards.