Key Takeaways
- Real-time geometallurgy is the umbrella term for fusing ore characterization proxies (crush and grind hardness proxies, mineral liberation, %solids) into live setpoint recommendations for the plant, rather than waiting for periodic lab testwork.
- It runs on the same closed-loop AI platforms used for crushing and grinding optimization. It is an application pattern, not a separate piece of software.
- The core idea: instead of running a static geomet model built from quarterly composite samples, you stream ore-property predictions into the control room as material actually arrives at the plant.
- It is available today through general-purpose industrial AI and APC vendors, not as a mining-specific off-the-shelf product. Expect a systems-integration project, not a shrink-wrapped purchase.
TL;DR
Real-time geometallurgy means piping your ore characterization proxies (hardness, liberation, clay content, %solids) into a closed-loop AI or soft-sensor platform like Imubit or ABB Ability for Mining. The plant then adjusts itself to the ore in front of it, instead of running on a fixed setpoint tuned for average ore.
What Is Real-Time Geometallurgy?
Geometallurgy links what the geology says about the ore to how that ore will behave in the plant. Traditionally it works on a slow cycle. Composite samples go to the lab, testwork measures hardness and recovery, and a geomet model is updated every few months. The plant then runs on setpoints tuned for the average ore that model describes.
Real-time geometallurgy shortens that cycle to minutes or hours. Ore properties are estimated continuously from sensors and upstream data, and the plant responds to the ore actually arriving. A soft band of ore coming through the crusher gets different settings from a hard band, instead of both being treated as average.
How Do I Apply AI to Real-Time Geometallurgy?
Real-time geometallurgy is less a single tool and more an integration pattern on top of your geomet and plant-control stack. It has three parts.
Gather the ore proxies
The inputs are proxy measurements you are probably already generating elsewhere:
- Blasthole grade data from hyperspectral or XRF sensing.
- Hardness and clay proxies from core scanning.
- Ore liberation estimates from automated mineralogy and ore texture analysis.
The output is a continuously updated prediction of how the ore currently entering the mill or flotation circuit will behave.
Connect proxies to plant control
The AI layer that makes this real time, rather than a quarterly report, is the same class of soft-sensor platform used for grinding and flotation optimisation. Imubit and ABB Ability for Mining both build models that ingest live plant and lab data continuously. They translate it into setpoint recommendations. Vendors increasingly frame this as “real-time geometallurgy” or “dynamic ore characterization” in their mining pitches.
The pattern runs in four steps:
- Characterise the ore upstream through drilling, sampling and sensors.
- Stream those proxies into a data historian.
- Feed a model trained to map ore properties to the best plant behaviour.
- Write setpoints back to the control system.
Roll out real-time geometallurgy like any closed-loop system
Use the same shadow-mode-then-closed-loop rollout as standalone grinding or flotation AI. The model recommends first, operators compare, and closed-loop control comes once the recommendations prove reliable.
What to Expect From Real-Time Geometallurgy Today
Be honest with stakeholders: this is still an emerging integration pattern, not a single named product you can buy off a price list. You assemble it from your existing geomet data pipeline plus a general industrial AI or APC platform. The project looks more like systems integration than a software purchase.
The payoff comes where ore variability hurts most. That is usually grinding circuit optimization, where hardness swings change throughput, and flotation, where mineralogy changes recovery. Upstream, dynamic block models keep the ore predictions current as new data arrives.
FAQs
What is the difference between geometallurgy and real-time geometallurgy?
Traditional geometallurgy builds a model of ore behaviour from lab testwork on composite samples and updates it periodically. Real-time geometallurgy uses sensor data and upstream ore information to estimate ore behaviour continuously. The plant can then respond to the ore actually arriving rather than to an average.
What data does real-time geometallurgy use?
It uses ore proxies such as hardness, clay content, mineralogy and liberation from blasthole sensing, core scanning and automated mineralogy. It combines these with live plant data from the historian, such as throughput, power draw and particle size. The model maps ore properties to the best plant settings.
Can you buy a real-time geometallurgy product?
Not as a single off-the-shelf product today. Industrial AI and APC vendors provide the modelling and control layer. The ore data pipeline comes from your own geology and geomet systems, so implementation is mostly integration work.
Where does real-time geometallurgy deliver the most value?
It delivers the most where ore variability is high and costly. Grinding circuits gain from adjusting to hardness changes, and flotation gains from adjusting to mineralogy changes. Operations with very consistent ore see less benefit.
Try It With Geocluster
Working out which of your ore-characterization data streams (hyperspectral, XRF, automated mineralogy) are good enough real-time proxies is a research problem before it is an engineering one. So is choosing which AI or APC vendor fits your plant’s control architecture. Geocluster is built to help you work through exactly that kind of multi-source technical evaluation.