MODELING OF NITROBENZENE REDUCTION TO ANILINE IN AN UNDIVIDED ELECTROCHEMICAL REACTOR USING ARTIFICIAL NEURAL NETWORKS

In order to establish the correlation between the input and the output variables fbr the electrochemical process of nitrobenzene reduction t o aniline, a Backpropagation Artificial Neural Network was used. The N etwork was trained with all experimental data, the total error of the learning process occurring with a very low value in a small number of epochs. The trained Network was used to predict the output variables f or patterns specially chosen to study the influence of each input vari able on aniline current yields. Considering these predictions, the opt imum conditions for the studied process were chosen and argued.