Stockpile Reclamation and Grade Blending for Processing Plant Feed: A Systematic Review of Methods, Models, and Research Gaps

Stockpiles and run-of-mine (ROM) pads are critical control points between mine production and processing plant feed. The grade, quality mix, and variability of material delivered to the crusher and mill are largely determined by how these structures are designed, built, and reclaimed. Despite the operational significance of stockpile management, the field remains fragmented across five distinct research streams—physical blending theory, stockpile state modelling, reclaim sequencing and equipment scheduling, plant-feed and stockpile blending optimization, and sensor-driven reconciliation and closed-loop control—with limited integration between them. This paper presents a systematic review of 27 sources published between 1976 and 2025, including peer-reviewed journal articles, conference papers, a preprint, a book, and one industry publication. The literature search was conducted in May 2025 using Scopus, Web of Science, and Google Scholar, with records screened by title/abstract and full text for direct relevance to stockpile reclamation or grade blending in mining operations. A structured coverage matrix identifies that studies combining high spatial fidelity with strong optimization rigor are consistently absent from the literature, and that uncertainty handling and sensor-driven or real-time capability remain substantially underdeveloped. Six research gaps are identified and prioritized by practical significance, implementation readiness, and literature maturity. The four most operationally critical gaps concern: spatially explicit reclaim scheduling under live ROM-pad constraints; tractable multi-attribute blending formulations for polymetallic operations; uncertainty propagation to plant-feed predictions; and field-scale closed-loop validation. The review provides a structured development roadmap for ROM-pad optimization frameworks and identifies the specific integration challenges that must be addressed to move the field from static stockpile monitoring toward adaptive, sensor-updated decision support.