Taguchi-based multi-response optimization of lead flotation from a complex Zn–Pb sulfide ore

This study presents a systematic statistical optimization framework for improving the selective flotation performance of a complex Zn–Pb sulfide ore by maximizing Pb recovery and concentrate grade while minimizing Zn misplacement. A Taguchi orthogonal array was first employed to efficiently design the experimental matrix, followed by Multiple Linear Regression (MLR) modeling combined with stepwise F-test analysis to quantify the effects and statistical significance of flotation variables. Subsequently, multi-response optimization using the Desirability Function was applied to identify the optimal operating conditions while resolving the inherent grade-recovery trade-off. The optimum conditions corresponding to the maximum global desirability (0.98) were determined as: pH = 11, [ZnSO 4 ] = 500 g/t, [NaCN] = 335 g/t, [PEX] = 30 g/t, and [2-MBT] = 10 g/t. Under these conditions, the developed models predicted a Pb recovery of 88.68%, a Pb concentrate grade of 72.06%, and a minimized Zn content of 2.91% in the concentrate. The statistical analysis further revealed that increasing collector (PEX) and collector-aid (2-MBT) dosages beyond critical levels did not provide significant improvements in flotation performance within the investigated range. This integrated statistical approach not only demonstrates significant improvements over baseline metallurgical performance but also provides a scalable, data-driven blueprint for optimizing complex mineral processing circuits.