High accuracy prediction of Cu(II) cations removal from synthetic aqueous solutions with ANNs using normal and modified biosorbent

Abstract

ABSTRACT This study aims to accurately predict Cu(II) removal efficiencies from synthetic aqueous solutions with Artificial Neural Networks (ANNs) using normal and modified Abies bornmulleriana cone. The predicted values of Cu(II) removal efficiencies for normal and modified Abies bornmulleriana cone were determined using a single ANNs model structure. In the present investigation, the neural network’s input parameters are biosorption time, biosorbate concentration, biosorbent amount, pH, particle size, agitation rate, and temperature. The removal efficiencies of the Cu (II) ions were selected as the experimental responses. The optimal network design is (7:7-9-2:2). Sensitivity analysis was employed to estimate the relative contribution of the input parameters to the ANN’s performance. The most critical parameter affecting the removal efficiencies of the Cu(II) was defined as biosorption time. The ANNs model for the normal and the modified Abies bornmulleriana cone had determination coefficients of (0.95, 0.98), standard deviation ratios of (0.22, 0.14), root mean square errors of (27.89, 26.02), and mean absolute errors of (2.21, 2.06), respectively. The outcomes from the network signify that the ANNs model could predict high accuracy Cu (II) removal efficiencies for both biosorbents.