Modeling on Transition of Heavy Metals from Ni–Cd Zinc Plant Residue Using Artificial Neural Network

Abstract

The assessment of heavy metal transition from Ni–Cd zinc plant residue (filtercake), as concentrations of zinc, nickel, cadmium and lead, requires quantification and a mathematical model that predicts their relative concentrations. Variability of filtercake characteristics may change the availability of heavy metals to the leachate or environment. In this study, a novel artificial neural network (ANN) model was constructed to predict Zn, Ni, Cd and Pb concentration leached from Ni–Cd filtercake in the leaching column. A three-layer backpropagation neural network was optimized and developed based on the Bayesian training algorithm. The inputs of this network are pH, flow rate of acidic influent, particle size and time. The geometry of the network giving the minimized mean square error (MSE) and sum of squared error (SSE) was a three-layer network having 18 neurons in the hidden layer (4:18:4) with a tangent sigmoid transfer function (tansig) at the hidden layer and linear transfer function (purelin) at the output layer. The fitting, regression, error and histogram plots for each response illustrate that there is a good agreement between the experimental data and the predicted values. Finally, a generalization of the developed model was carried out as 3D plots to evaluate the interactions of the input parameters on the transition of heavy metals to the leachate. With respect to these results, the effect of particle size on concentrations of zinc, nickel and lead are less (<3 mg/L) than that of cadmium (<3 mg/L). Furthermore, it was found that, at low flow rate, the concentrations of extracted metals are high due to enhancement of exposure residence time (between particles and leach solution).