Statistical modeling and optimization of itaconic acid reactive extraction using response surface methodology (RSM) and artificial neural network (ANN)

In this paper, regression models were proposed to predict the degrees of extraction (%Y) for the reactive extraction of itaconic acid using response surface methodology (RSM) and artificial neural network (ANN). The prominent design parameters like itaconic acid concentration, extractant (tri-n-octylamine), and modifier (dichloromethane, an active diluent) composition were considered, and their impact on the extraction efficiency was determined. RSM and ANN fitted the experimental data with a correlation coefficient of 0.970 and 0.993, respectively. The statistical significance of the models (RSM and ANN) was ascertained by ANOVA analysis. The optimal design factors were determined to be 0.072 mol center dot L-1 acid concentration, 16.075 %v/v extractant composition, and 62.15 %v/v modifier composition at which the values of experimental and predicted %Y of 98.86% and 100.69%, respectively, were obtained by RSM model.