(144d) Packing Characterization for Post Combustion Capture: Model Development

Packing is widely used in post-combustion CO₂ capture because of its low pressure drop, good mass transfer efficiency, and ease of installation. In the CO₂ capture process, absorber and stripper performance are highly dependent on the effective mass transfer area of the packing (a_e). The stripper performance also depends on the liquid film mass transfer coefficient (k_L). Gas cooler and water wash performance depends on the gas film mass transfer coefficient (k_G).  In this work, three mass transfer models predicting a_e, k_L, and k_Ghave been developed based on measurements in the 0.428 m diameter PVC column, including the data for eleven structured packings and three random packings.

The effective mass transfer area model is based on previous researcher’s model (Tsai, 2010). The effective area is assumed to be only a function of liquid flow rate, packing surface area, surface tension, and is independent of gas flow rate and liquid phase viscosity. In this work, the improvement from Tsai’s model is to use liquid superficial velocity over packing total area (u_L/a_P) as the liquid flow rate per perimeter. Therefore, the area model can be applied in situations where channel dimensions are not known or hardly defined. The effective area model is:

a_e/a_P = 1.41[(ρ_L/σ)g^1/3(u_L/a_P)^4/3]^0.116

The k_L and k_G models are based on three factors that influence packing mass transfer: liquid or gas superficial velocity (u_L/G), the packing surface area (a_P), and the mixing point density (M). Mixing point density is a new concept proposed in this work, which is the number of contacting points between packing corrugated metal sheets. It represents the packing geometry effect on mass transfer coefficients. Mixing point density can be calculated by packing channel base B, crimp height h or packing surface area a_Pand corrugation angle in a more general case:

M = 6/(BhB*tanθ) = (3a_P³sinθcosθ)/[16(sin²θ+1)^3/2]

The k_L and k_G models devloped in this work are:

k_L=3.08E-3u_L^0.72M^0.42a_P^-1.15  

k_G=9.6E-3u_G^0.54M^0.29a_P^-0.5

The mass transfer models developed in this work can predict experimental data well. The average deviation is 10% for effective area model, 22% for k_L model, and 13% for k_G model. The models are also compared with literature models and have good comparison.