Data-Driven Gas Holdup Correlation in Bubble Column Reactors Considering Alcohol Concentration and Carbon Number

Abstract

Due to the complex relationship between various parameters affecting gas holdup in bubble column reactors, a unique correlation for gas holdup does not exist. The available correlations proposed in the literature for gas holdup prediction in aqueous alcohol solutions in bubble columns fail to predict gas holdup over a wide range of conditions. Therefore, based on around 1000 data points from the previous studies, an empirical correlation and a trained model were derived using the dimensionless numbers Reynolds, Froude, Eötvös to Morton ratio, and alcohol carbon number. The predictions were compared to experiments with different water–alcohol mediums at various concentrations to validate the correlation and trained model, and a good agreement was observed. However, the ML model was predicting more accurately, and it was indicated that the Reynolds number had the most significant impact on gas holdup, followed by the Eötvös to Morton ratio.