Anglo American on real time ore selection with Active Hyperspectral Sensing
The geology of orebodies is inherently variable, and the ability to characterise that variability accurately and in time to inform operational decisions is critical to effective mine planning and processing performance. Conventional ore characterisation methods (including drilling, geological modelling and laboratory sampling) provide essential information, but can often involve delays between data collection, analysis and operational response. This may limit the precision of ore selection in the pit and impact the quality of material sent to the plant for processing.
In response to this challenge, Anglo American says it is developing sensing technologies that are bringing real-time ore intelligence into operations, enabling teams to make faster, better-informed decisions as material moves through the process.
The goal is to shift ore control from reactive to predictive. Instead of relying only on delayed laboratory analysis, sensing technologies can continuously assess ore characteristics directly in-pit, on conveyors, or within processing streams. To explore this opportunity, it has been studying systems to improve ore routing, reduce dilution, stabilise plant feed and ultimately, recover more value from the orebody.
At the Minas-Rio iron ore operation in Brazil, the presence of alumina-rich clay in the orebody was identified as a source of variability that could be more effectively and proactively managed to optimise operational performance, with a technoeconomic study indicating a significant value opportunity for sensing and selective mining. Initial test work was carried out at Anglo American’s Sensor Fusion Loop in South Africa – a full-scale testing facility designed specifically to evaluate sensing technologies in real-world operating conditions.
Ore from Minas-Rio was tested using multiple sensing approaches under both wet and dry conditions to determine whether problematic high-alumina material could be accurately identified.
Initial results demonstrated the effectiveness of Active Hyperspectral Sensing (AHS) – a new technology developed in partnership with Finnish researchers at VTT. Unlike conventional hyperspectral systems, which often underperform in mining environments, due to changing light conditions and operational complexity, AHS uses an actively controlled light source that enables more reliable sensing directly in the field.
During testing, results indicated that the new technology achieved a high degree of accuracy in differentiating key ore types from Minas-Rio, representing an important milestone in demonstrating the feasibility of real-time ore classification in operational mining environments.
The next challenge was determining whether the technology could work in-pit, under typical operating conditions.
Following successful testing at the Sensor Fusion Loop, Anglo American undertook an initial field test with the sensor technology in the mining pit. Field demonstrations at Minas-Rio successfully differentiated high-alumina material under both bright sunlight and wet, overcast conditions, demonstrating the technology’s ability to operate in the varying conditions of a working mine.
On the basis of this successful demonstration, Anglo set out to pioneer a breakthrough innovation: a mobile “Sensing Pod” capable of scanning ore directly in the mining pit, processing results in near real time and integrating with the mining fleet management system.
A proof-of-concept programme is now under way at Minas-Rio to evaluate how sensing can support selective mining, improve ore control and contribute to more stable downstream processing.
At the same time, Anglo American says it continues to explore broader applications for sensing technologies across multiple commodities and value drivers. Sensing it says represents a new way of understanding orebodies, which has the potential to move mining toward faster, more connected and more intelligent operational decision-making.
The post Anglo American on real time ore selection with Active Hyperspectral Sensing appeared first on International Mining.