ie7b05024_si_001.pdf (91.59 kB)

Development of a Pressure Swing Adsorption Cycle for Producing High Purity CO2 from Dilute Feed Streams. Part I: Feasibility Study

Download (91.59 kB)
journal contribution
posted on 01.06.2018, 11:23 by Hanife Erden, Armin D. Ebner, James A. Ritter
Simulations of a 3-bed 9-step pressure swing adsorption (PSA) cycle were carried out to study the enrichment and recovery of 0.4 vol % CO2 from dry air using 13X zeolite with the feed pressure and flow rate fixed at 1 atm and 570 SLPM. The goal was to produce 97 vol % CO2 at greater than 62% CO2 recovery, less than 1 kW vacuum pump power, and small volume. The PSA cycle step sequence consisted of feed (F), idle (I), heavy reflux (HR), cocurrent equalization down (EqD), forced cocurrent depressurization (CoD), countercurrent depressurization (CnD), light reflux (LR), countercurrent equalization up (EqU), and light product pressurization (LPP). A parametric study revealed the following effects on the PSA process performance. Increases in the HR = LR = F step time (thus cycle time) caused significant increases in the CO2 purity and vacuum pump power, while the CO2 recovery exhibited a modest maximum. Increases in the LR ratio caused significant increases in both the CO2 recovery and power but only a modest increase in the CO2 purity. Increases in the CnD end pressure caused only slight changes in all three parameters, with the CO2 purity increasing and CO2 recovery and power decreasing. Increases in the LR pressure caused significant decreases in both the CO2 recovery and power, with only a slight decrease in CO2 purity. Increases in the CoD end pressure caused a significant decrease in the CO2 purity, a modest increase in the CO2 recovery, and only a slight decrease in power. The best PSA process performance produced a CO2 purity of 96.3 vol % at a CO2 recovery of 87.8% and a feed throughput of 1264 L­(STP)/h/kg, while consuming 572 W. This new PSA cycle was very effective at concentrating CO2 (over 242 times) because of a forced CoD step and a very long HR = LR step time relative to the other cycle steps, where the source of the HR was exclusively from the LR step. The LR step dominated the power requirement, so much that the power required by the CnD step and especially the forced CoD step were both insignificant in comparison.