The use of Artificial Neural Network (ANN) for modeling of Cu (II) ion removal from aqueous solution by flotation and sorptive flotation process

Abstract

Abstract A three-layer artificial neural network model was developed to predict the removal of copper ions from simulated wastewater by flotation and sorptive flotation systems using Sodium Dodecyl Sulfate (SDS) as surface active agent and sunflower seed husk as low-cost locally available adsorbent based on 197 experimental data. The effect of initial copper ion concentration, flow rate, sunflower seed husk dosage and surfactant concentration at 40 min contact time, pH of 5 and at room temperature, were investigated. The best value of these parameters that achieved maximum removal efficiency of Cu(II) ion were 30 g/l initial Cu(II) ion concentration, 500 cm3/min flow rate for both systems and 7 g/l sunflower seed husk for sorptive flotation while the SDS concentration was 60 mg/L for flotation and 30 mg/L for sorptive flotation. The present model was capable to predict the removal efficiency with a tangent sigmoid transfer function (tansig) at hidden layer with 20 neurons and a linear transfer function (purelin) at output layer. The linear regression between the corresponding targets and network outputs was confirmed to be satisfactory with a correlation coefficient of 0.99955 and 0.99549 in flotation and sorptive flotation system, respectively. Sensitivity analysis shows that the time and pH was found to be most significant parameters of relative importance 24.22 and 23.02% in flotation system, and the sunflower seed husk dosage and time with relative importance of 26.86 and 24.23% were found to be the most affected one for sorptive flotation system.