Oconnell Type Correlation for Prediction of Overall Efficiency of Valve Tray Columns

Abstract

The prediction of overall column efficiency directly affects the design and cost of a distillation column. Therefore, there is an economic incentive to predict column efficiency as accurately as possible. The O'Connell correlation is widely recommended for efficiency predictions and was developed using bubble-cap and sieve tray data. In this work, a new O'Connell type correlation for prediction of overall efficiency of valve tray columns is developed using valve tray data collected at Fractionation Research, Inc. (FRI) that have been made public. The correlation was developed using data for ortho xylene / para xylene, isobutene / normal-butane, cyclohexane / n heptane, and n octanol / n decanol systems at pressures ranging from 0.2 to 164.7 psia. The data were collected on the 4 feet diameter FRI column between 1956 and1968. All of the FRI data corresponds to total reflux operation. An O'Connell type correlation (The OSU-FRI valve tray correlation) for prediction of overall efficiency of valve tray columns has been developed. The form of the correlation is similar to that of the O'Connell correlation. The mean absolute relative error obtained for the new valve tray correlation is 5.6 %. This model is expected to predict the efficiency of valve tray column in reasonable agreement, thereby help design engineers economically design valve tray distillation columns. The new OSU-FRI valve tray correlation is expected to predict efficiency over a wide range of pressures, as it was developed using data that were available at a wide range of pressures from vacuum to super atmospheric. In addition, the correlation is directly applicable to commercial column designs without any scale-up factors, as data used for the development of the correlation were taken from a commercial column.