ANN modeling of lead adsorption from water using biowaste material as adsorbent

Abstract

This research employed an Artificial Neural Network (ANN) approach to model the adsorption of lead from water using biowaste material as an adsorbent. To develop the ANN model, 39 experimental data point values for training and 15 experimental data point values for testing were used. The model involved the use of a tansigmoid transfer function for the input, purelin for the output, and a hidden layer consisting of 19 neurons to minimize the square error. The highest percentage of lead removal was achieved with optimal adsorption parameters (i.e., pH, concentration of lead, and biosorbent dosage). The experimental data closely matched the predicted values obtained from the ANN model with a R2 (regression coefficient) of 0.988. To determine the maximum percentage of lead removal and the optimal adsorption parameters, a pattern-search approach using a genetic algorithm was employed. In this study, adsorbent powder was prepared using biowaste materials, such as Borasus flabellifer coir.