Artificial Neural Network Modeling for Biosorption of Pb (II) Ions on Nanocellulose Fibers

Abstract

This study investigates the efficiency of nanocellulose fibers (NCFs) derived from rice straw for the abatement of lead ions from aqueous solution. The maximum biosorption of lead (II) is found to be 94.21 % at pH 6.5 and equilibrium time of 40 min by using NCFs of 0.5 g/200 ml (test volume) and initial concentration of 25 mg/L of lead concentration. The data collected from laboratory-scale experimental setup are used to train a multilayer perceptron network combined with backpropagation and Levenberg–Marquadt algorithm. Different artificial neural network (ANN) architectures were tested by varying network topology as a function of number of neurons in the hidden layer of ANN model. A single-layer ANN model with ten hidden neurons at 1,000 epochs using sigmoidal transfer function has been found to have the best predictive performance. The network is found to be working satisfactorily with minimum mean squared error (5.8472E-05) and mean absolute error (1.38845310688678) during training phase. Comparison between the network results and experimental data gives a high degree of correlation coefficient (R 2 = 0.995) indicating that the model is able to predict the sorption efficiency with reasonable accuracy. The influence of each parameter on the variable studied was assessed, and pH, volume of the test solution, biomass dose, metal concentration, and contact time were found to be the most significant factors. Simulations based on the developed ANN model can estimate the behavior of the biosorption phenomenon process under different conditions.