Dynamic adsorption Modelling of P-nitrophenol in Aquous solution using artificial neural network

This work aims at developing an artificial neural network to describe the dynamic adsorption of p-nitrophenol from an aqueous solution using an nda-100 resin as adsorbent under different conditions. nine neurons were used in the input layer corresponding to the nine input parameters (time,area surface,particle diameter,mass of adsorbent,bed high,apparent density,molar mass,initial concentration and flow rate) and one neuron in the output layer corresponds to the reduced concentration. after training/test and validation,the best ann was found with four neurons in the hidden layer with the transfer function (tangent sigmoid),levenberg marquardt retro-propagation algorithm and linear transfer function for the output layer. the optimized ann was obtained with a high correlation coefficient of r2 = 0.9992 and mean squared error of mse = 0.00028 during the generalization phase. an interpolation and a prediction stage have been done to test the reliability of the best ann. results obtained from the interpolation gave a correlation coefficient of r2 = 0.999 and mean square error mse = 0.00036,so for the prediction are r2 = 0.9187 and mse = 0.0478. these results show that the proposed ann model is able to accurately predict the reduced concentration of the dynamic adsorption phenomenon of p-nitrophenol using an nda-100 resin. © university of mohammed 1er oujda morocco.