444753 Aspen Model for Propane/Propylene Separation By Pressure Swing Adsorption Using Carbon Molecular Sieves

Tuesday, April 12, 2016: 10:15 AM
343B (Hilton Americas - Houston)
Zubin Kuvadia1, Yujun Liu1, Junqiang Liu2, Ted Calverley2, Marcos Martinez3 and Mark Brayden4, (1)Engineering & Process Sciences, The Dow Chemical Company, Freeport, TX, (2)Inorganic Materials & Heterogeneous Catalysis, The Dow Chemical Company, Midland, MI, (3)The Dow Chemical Company, Plaquemine, LA, (4)Hydrocarbon Research, The Dow Chemical Company, Plaquemine, LA

Cryogenic distillation is the current method for separating propylene/propane mixtures. It is a difficult separation requiring high reflux ratios, high energy input, and tall columns, because of the small difference in the volatilities of propylene and propane. Adsorption based processes have been explored in recent years as potential replacement of or augment to the incumbent cryogenic distillation process to reduce the energy intensity or debottleneck existing assets.  

Dow has developed new carbon molecular sieves (CMS) that show a high separation factor for propane and propylene. This presentation shows the development of a PSA model using ASPEN ADSORPTIONTM. The adsorption isotherms of propylene and propane on the CMS were measured at different temperatures that cover the commercial operating range. A temperature dependent, four-parameter Langmuir model was selected to fit the single component isotherm data and the model parameters were obtained by using the Aspen ADSORPTIONTM equilibrium parameter estimation tool.  Different multicomponent isotherm methods including the Ideal Adsorbed Solution Theory were explored. The heat of adsorption of each component was determined from the isotherm models according to the Clausius – Clapeyron equation. The mass transfer coefficients of propylene and propane were estimated using Aspen ADSORPTIONTM simulations by fitting breakthrough curves at different flow rates. Models for a simple Skarsktrom cycle as well as alternate cycle configurations from literature were set up using ASPEN. Model predictions are compared to PSA experimental results. Propylene purity and recovery are the key output variables. The validated PSA model can then be used for future process optimization and scale-up.

Extended Abstract: File Uploaded