Review of algorithms for modeling metal distribution equilibria in liquid-liquid extraction processes

Abstract

This work focuses on general guidelines to be considered for application of least-squares routines and artificial neural networks (ANN) in the estimation of metal distribution equilibria in liquid-liquid extraction process. The goal of the procedure in the statistical method is to find the values of the equilibrium constants (K_j) for the reactions involved in the metal extraction which minimizes the differences between experimental distribution coefficient (D_exp) and theoretical distribution coefficients according to the mechanism proposed (D_theor)- In the first part of the article, results obtained with the most frequently routine reported in the bibliography are compared with those obtained using the algorithms previously discussed. In the second part, the main features of a single back-propagation neural network for the same purpose are discussed, and the results obtained are compared with those obtained with the classical methods.