Experimental study on mass transfer and prediction using artificial neural network for CO2 absorption into aqueous DETA

Abstract

The volumetric overall mass transfer coefficient (KGav) for carbon dioxide (CO2) absorption into aqueous diethylenetriamine (DETA) was experimentally determined in an absorption column packed with Sulzer DX-type structured packing over a temperature range of 30–50°C and at atmosphere pressure. The effects of the main operating parameters (i.e., inlet CO2 loading, solvent concentration, liquid flow rate, CO2 partial pressure, inert gas flow rate, and liquid feed temperature) on KGav were investigated. The experimental results showed that KGav was influenced by inlet CO2 loading, solvent concentration, liquid flow rate, CO2 partial pressure, and liquid feed temperature, but the effects of inert gas flow rate were insignificant. In addition, an artificial neural network (ANN) model was designed to predict the mass transfer performance. The main operational and physical parameters were selected as input parameters, while KGav was chosen as an output variable. Comparison between the predicted values from the ANN model and experimental data demonstrated that the ANN model is suitable for predicting the absorption performance of packed columns.