ESTIMATION OF DISTILLATE COMPOSITION & ITS EXPERIMENTAL VERIFICATION

Abstract

Distillation is a method of separating the components of solutions which depends upon the unequal distributioni,i, of its components between a gas & a liquid phase, applied to cases where all components are present in both phases. The new phase is created from the original solution by vaporization or condensation. In the present work an ANN based estimator using backward error propagation algorithm, to estimate the compositions of the distillate by taking temperature profile of trays & reflux drum has been developed. A dynamic semi rigorous mathematical model of laboratory set-up of distillation column consisting of trays has been developed and simulated. The developed estimator is tested for a two component mixture of methanol and water. Estimated results are compared with the results obtained from simulation software developed. Experimentation is also performed on the laboratory set-up to validate the results of the developed estimator.