Multi-objective optimization of graphite flotation via grinding–classification–reagent interactions: A modeling-based approach

Abstract

Graphite flotation is strongly governed by complex interactions between grinding conditions, particle size classification, and reagent chemistry, particularly in mica-rich ores where selectivity remains a major challenge. In this study, a multi-objective optimization framework is proposed to improve both fixed carbon grade and recovery through an integrated modeling-based approach. A series of progressive flotation experiments were conducted, incorporating variations in grinding time, classification strategy, and reagent regimes. Performance indicators, including fixed carbon content, recovery, and a composite optimization score, were systematically analyzed to capture trade-offs between product quality and yield. Statistical correlations and response surface models were developed to quantify the individual and interactive effects of operating variables. The results reveal the existence of an optimal operational domain characterized by a balanced compromise between grade and recovery. Classification was identified as a key lever for enhancing selectivity, while excessive grinding led to diminishing returns due to fine gangue entrainment. The proposed models demonstrate good predictive capability and robustness around the optimum conditions, providing a reliable decision-support tool for process optimization.