Mass transfer correlations for air stripping towers

The overall system-specific liquid phase mass transfer coefficient, K(L)a, is the most important parameter in the design of air strippers. Designers frequently estimate K(L)a using a mass transfer correlations to avoid the cost of pilot tests. This paper evaluates the well-accepted Onda correlations using a newly developed data base. Predictions based on the Onda correlations systematically overestimated K(L)a for packings typically used in full-scale applications. They had an average error of +/-40% and an average standard deviation of +/-47%. The overestimation was a function of the packing diameter. A simple equation was developed to correct the Onda overestimation for packings larger than 1 inch (2.54 cm) in diameter, and it yielded predictions within +/-20% range. The experimental K(L)a values obtained in this study varied with packing depth contradicting a basic assumption of the mass transfer theory, i.e., that K(L)a spatially is constant. Also, the effect of measurement errors of the variables involved in the calculation of both the experimental and the predicted K(L)a values were investigated. A large part of the observed difference between experimental and Onda predictions was attributed to a lack of fit of the Onda model. Thus, the quality of the predictive correlations can be improved.