Experimental coupled predictive modelling based recycling of waste printed circuit boards for maximum extraction of copper

The recycling process of materials from used and wasted printed circuit boards plays an important role in electronic waste management. These waste printed circuit boards (PCBs) hold metals such as copper, aluminium, nickel, and magnesium. The efficient recovery process of such metals from waste PCBs is needed for recycle and possible reuse for manufacturing of products. The metal recovery process is complex and, multidimensional and costly to perform. In addition, the efficient (maximum) recovery of metals exhibit higher dependence on determination of optimum combination of inputs in the recovery process from waste PCBs. Therefore, this work illustrated the ability of four predictive modelling methods (Analysis of Variance, Genetic Programming, Artificial Neural Network and Generalized Neural Network) to model complex suspension electrolysis process (recovery process) and their comparative analysis on recovery of copper metal from waste PCBs. Experiments were designed based on variations of three design/input parameters such as concentration of sulfuric acid, concentration of copper sulphate and current density. The comparative analysis of the four methods mentioned above reveals that Generalized Neural Network performed the best with coefficient of determination value at 0.92. 3-D surface contour analysis showed that a lower concentration of sulfuric acid (0.3–0.6) coupled with higher concentrations of copper sulfate (∼0.8–1) yielded a higher percentage of copper. However, current density brought an increase in copper recovery only when there was a corresponding increase in the concentration of sulfuric acid or copper sulphate. Thus, the optimal conditions to gain the maximum yield of copper obtained through this method were concluded as follows: concentration of copper sulphate value at 31.4 g/L, concentration of sulfuric acid at 112 g/L and current density at 3 A/dm2.