Data-Driven Analysis of Dilution Factors in Dredge Mining : Addressing Fast Percolation Challenges from an Ore Body Knowledge Perspective at Grande Côte Operations (GCO), Senegal (ULiège)

Abstract

This study investigates the factors influencing water behaviour in the dredge pond at Grande Côte Operations (GCO), Senegal, with a focus on addressing dilution challenges. Maintaining the correct water level in the dredge pond is crucial for optimizing mineral extraction efficiency and ensuring the economic viability of mining operations. Variations in the water level can force the dredge to mine lower-grade sands, leading to increased dilution of valuable heavy minerals. In certain areas of the deposit, rapid percolation, associated with higher hydraulic conductivity, causes swift changes in the pond water level, complicating the maintenance of optimal dredging conditions. Economic losses related to the control of the dredge pond water level are estimated @25 kt of HMC per year for a half a meter water level variation, which translates to ~7.5M euros/year. Using various datasets encompassing geological, hydrogeological, and production data, this research employs a multilayered methodology to address the rapid percolation issue. Time series and different custom-made tools were combined to conduct exploratory data analyses aiming to identify periods of fast percolation and associate them spatially with geological variables. Once identified, the temporal and spatial trends within the information lead to the identification of significant lithological and hydrological features linked to hydraulic conductivity. By understanding these variables, the study aimed to develop predictive models that can forecast areas of rapid percolation, allowing for proactive water management strategies. Despite the predictive algorithms not achieving the desired accuracy, the study successfully identified key features linked to the issue of rapid percolation. The primary challenge in developing a robust model was the lack of systematic measurements of hydraulic conductivity and particle size analyses. To address this, two strategic methodologies were proposed, emphasizing the need for detailed particle size distribution analysis and, if feasible, the incorporation of additional intrinsic lithological properties. These enhancements are crucial for developing a more accurate and effective model to anticipate dilution caused by rapid percolation. This study addresses dilution challenges in the dredge pond at GCO by investigating lithological factors influencing water behaviour. The findings provide actionable insights for developing methodologies to anticipate pond water level variations, enhancing mining efficiency, water management, and profitability at GCO.