Key Takeaways
- A traditional mill survey is a manual, periodic snapshot: a team samples the circuit for a shift and calculates a mass balance after the fact.
- Continuous capture is displacing it. Plant historians like AVEVA PI System log the same circuit variables 24/7, at far higher resolution.
- This is genuinely an “Emerging” category, not a mature off-the-shelf product. Most operations layer analytics on top of historian data rather than buying a dedicated “AI mill survey” tool.
- Imubit and similar industrial-AI platforms are the closest thing to a purpose-built analytics layer for this data.
TL;DR
Keep the manual mill survey for calibration, but stop relying on it alone. Log circuit data continuously in a plant historian, then apply industrial-AI analytics such as Imubit on top. The result is survey-grade mass-balance insight in real time instead of once a quarter.
What Is a Mill Survey?
A mill survey is a planned sampling campaign around a grinding circuit. A metallurgical team samples the key streams over a shift: mill feed, mill discharge, and cyclone overflow and underflow. The lab measures solids content and particle size distribution for each sample. The team then reconstructs a mass balance to check whether the circuit performs to design.
Surveys need the circuit to run as close to steady state as possible. Otherwise the samples describe a moving target. Mass balance software then adjusts the measured data statistically to find the most consistent balance. If the adjusted numbers differ too much from the measurements, the survey has to be repeated. The Global Mining Guidelines Group publishes a guideline on how to survey and sample grinding circuits for this purpose.
This approach is accurate, but it is expensive to run often. As a result, most sites survey quarterly, or when something looks wrong.
How Do I Automate Mill Surveys With AI?
The shift underway is to treat the plant historian as a continuous survey. That takes three steps.
Instrument the circuit
Continuous surveys need online measurements at the same points a manual survey would sample. That usually means particle-size analyzers, density gauges and flow meters on the key streams. Most modern grinding circuits already have some of these. The gaps show you where to invest first.
Log everything in the historian
Next, make sure every measurement logs to AVEVA PI System or an equivalent historian at a useful resolution. At that point you already have the raw data for a mass balance every minute rather than every quarter. What is still missing is the analytics layer that turns that stream into an actionable circuit-performance signal.
Add the analytics layer for continuous mill survey data
This is where industrial-AI platforms like Imubit fit. Rather than offering a standalone product, they treat this data as one input stream among many. Those streams feed a broader optimisation model of the comminution circuit. The model can flag when classification efficiency drops, when the mill is overloaded, or when throughput and grind size drift apart.
Where Continuous Mill Survey Tools Stand Today
Be realistic: there is no single dominant “AI mill survey” product on the market, the way there is for automated mineralogy. Most sites doing this well built it as a custom analytics layer on their historian data. They did not buy it off a shelf. That means the real project is more data engineering than AI shopping.
Manual surveys do not disappear either. Online analyzers drift, and a periodic manual survey is still the best way to check them. In practice, the manual survey becomes the calibration point for the continuous one. The grinding circuit’s output also sets up everything downstream, so better survey data pays off in flotation too.
FAQs
How often should a mill survey be done?
Most sites run a full manual survey quarterly, or whenever circuit performance looks off. With continuous historian data and analytics in place, you can monitor the circuit every day and keep manual surveys for calibration and major changes.
What is measured in a grinding circuit survey?
A grinding circuit survey measures mass flow, solids content and particle size distribution at key streams. These typically include mill feed, mill discharge and cyclone overflow and underflow. Those values feed a mass balance that shows how well the mill and classifiers are performing.
Can AI replace a manual mill survey?
Not completely. AI analytics can turn continuous historian data into near-real-time circuit insight. However, online instruments drift and need checking against physical samples. The manual survey remains the reference that keeps the continuous data honest.
What is a mass balance in mineral processing?
A mass balance accounts for all material entering and leaving each part of a circuit, so flows and grades add up consistently. Measured data never balances perfectly, so software adjusts the values statistically. The balanced figures then show recovery, circulating load and classification performance.
Try It With Geocluster
If you are weighing whether to build a continuous-survey analytics layer in-house or buy into an industrial-AI platform, that build-vs-buy question is what Geocluster can help you work through, with real vendor and case-study grounding.