(126e) A Novel Mixed Integer Dynamic Optimization Model for a PSA Process

Pressure swing adsorption (PSA) is an industrial process used in the separation of gas mixture under high pressure (Delgado, Rodrigues, 2008). There have been several studies that examined the most dominant phenomena taking place in the absorption columns. (Khajuria and Pistikopoulos, 2011) Despite the added value of key contribution in the field the development of advanced optimization strategies still remains a great challenge. Due to the periodic nature, the performance of PSA systems highly depends on the switching time. That, together with the existence of both discrete and distributed variables increase the system complexity.

In this work we study the development of dynamic optimization strategies using binary variables to declare the phase transitions (Ko, Siriwardane et al, 2003). The algorithm is based on if-else logical conditions that determine the current process phase. The use of binary variables compensates for the discrete phenomena taking place and enables the development of optimization strategies.

The majority of PSA models are characterized by rapid changes in the output and process disturbances (Dantas, Luna et al, 2011). Consequently this profile is translated to model instability. Many of these works do not consider the valve equation as part of the mathematical formulation. In order to increase the model reliability, the valve equation has to be included in the equation set (Choi and Wen-Chung, 1994). In this work we consider the valve equation and following the strategy described above we transform it in a continuous form in order to enable the optimization solution. The newly introduced equations are determined using the last square method in MATLAB and the dynamic optimization studies are carried out in gPROMS.

The objective of the optimization problem is the maximization of both recovery and productivity under the constraint of purity.

The optimization model described in this research can be applied to most PSA process. However, the focus of this study is on PSA process for bio-gas upgrading. The results of the optimized procedure from this model show significant improvements in productivity (13.85%) and recovery (3.57%) compared to current operating procedure.

Delgado J. A. and Rodrigues A.E. (2008). “Analysis of the boundary conditions for the simulation of the pressure equalization step in PSA cycles”, Chemical Engineering Science 63(18): 4452-4463.

Khajuria H. and Pistikopoulos E. N. (2011). “Optimization and Control of Pressure Swing Adsorption Processes Under Uncertainty”, AIChE Journal 59(1): 120-131.

Ko D., Siriwardane R. and Biegler L. T. (2003). Optimization of a pressure-swing adsorption process using zeolite 13X for CO2 sequestration. Industrial & engineering chemistry research, 42(2), 339-348.

Dantas T. L., Luna F. M. T., Silva Jr I. J., Torres A. E. B., De Azevedo D., Rodrigues A. E. and Moreira R. F. (2011). “Carbon dioxide–nitrogen separation through pressure swing adsorption” Chemical Engineering Journal 172(2): 698-704.

Choi C. T. and Wen-Chung H. (1994). “Incorporation of a valve equation into the simulation of a pressure swing adsorption process” Chemical engineering science 49(1): 75-84.

Ko D. and Moon I. (2000). “Optimization of start-up operating condition in RPSA”.Separation and purification technology 21(1): 17-26.