Advanced Strategies for Optimal Design and Operation of Pressure Swing Adsorption

With expanding areas of applications, increasing needs for efficient cycles, and growing demands for efficient modeling, it has become essential to develop new systematic strategies for optimal design and operation of PSA systems. Although industrial usage of PSA is widespread, we observe a drought of any systematic methodology to design PSA cycles in PSA literature due to inherent complexity of cyclic PSA processes. We present a generic PSA superstructure to synthesize optimal PSA configurations. The superstructure is rich enough to predict a number of different PSA operating steps, and their optimal sequence by solving an optimal control problem. Because of low operating costs and high performance, PSA is considered as a promising option for both post-combustion and pre-combustion CO2 capture. Since commercial PSA cycles consider CO2 as a waste stream, cycle development specifically targeted towards high-purity CO2 separation is essential. We utilize superstructure approach for this purpose and succeed in synthesizing optimal cycles which can separate CO2 at a purity as high as 95%, or with a low power consumption of 465 kWh/tonne CO2 captured, for post-combustion capture. When applied for pre-combustion capture, superstructure approach yields cycles with an extremely low power number of 46.8 kWh/tonne CO2 captured. Large number of spatial nodes required to capture steep adsorption fronts lead to a large set of DAEs, and thus to a challenging PSA optimization problem. We generate reduced-order models (ROMs) which are not only orders of magnitude smaller, but also reasonably accurate. Consequently, replacing DAEs with these ROMs yields a cheap optimization problem. How-