Improving the prediction of irrigated pressure drop in packed absorption towers

Various tools estimating irrigated pressure drop in gas-liquid counter-current randomly dumped packed beds are carefully examined through the perception of a comprehensive database. The reported measurements consisting of ca. 5000 experiments represent an important portion of the non-proprietary information released in the literature. Artificial neural network (ANN) modeling Is proposed to refine the accuracy and broadness in predicting the irrigated pressure drop across the bed. The ANN correlation [f(LGG) = f(Re-G,Ga-G,Re-L,Ga-L,St(L),S-B, chi)] yields an average absolute relative error (AARE) of 20.0% and a standard deviation on the AARE of 19.8% for the whole database and remains in accordance with the physical evidence reported in the literature.