Flotation on Eigenform AI Geo Tooltips

ABA Models

Mon, 21 Sep 2026 00:00:00 +0800

Key Takeaways

Acid-base accounting (ABA) itself is a standard, decades-old static/kinetic lab test — there’s no AI in the test protocol.
PHREEQC, the free USGS geochemical modeling engine, is the standard open-source tool for extending raw ABA results into predictive geochemical models.
The INAP GARD Guide is the industry-standard methodology reference for how to run and interpret ABA/ARD programs.
This is genuinely “Emerging” for AI specifically — the AI opportunity is in predicting long-term drainage chemistry from ABA + mineralogy data, not in the test itself.

TL;DR

Run standard static and kinetic (humidity cell) ABA testing per the INAP GARD Guide methodology, then use PHREEQC to model the resulting geochemistry — with genuine AI upside still emerging in using ML to predict long-term drainage behavior from that data rather than relying purely on kinetic cell extrapolation.

AI-Optimized Ore Blending, Stockpile & Plant Feed Strategy

Mon, 21 Sep 2026 00:00:00 +0800

Key Takeaways

MineMax by NTWIST is explicitly AI-driven — it sits as a supervisory layer over existing mine and plant systems, continuously learning from operational outcomes to make real-time blend and feed recommendations.
Deswik Blend, BOLT, and GO are the leading optimization-based planners for stockpile management, multi-commodity blending, and joint mine-to-market feed decisions.
Dynamic blend-consistency dispatching (routing trucks to maintain a target feed blend in real time) has been reported to raise truck cycle efficiency by roughly 11% in deployed systems.
MineMax’s four core models — OreMax, DynaMax, PlanMax, MillMax — cover ore tracking, stockpile intelligence, feed forecasting, and optimization as one connected decision layer.
All three items here (stockpile, plant feed, and blending strategy) are really one problem viewed from three points in the material flow, and increasingly solved by the same platforms.

TL;DR

Layer an AI optimization platform (MineMax/NTWIST, or Deswik’s Blend/BOLT/GO suite) on top of your existing mine and plant systems to continuously recommend stockpile allocation, plant feed mix, and dispatch blending targets from real-time ore-tracking data, rather than planning blends on a fixed weekly/monthly schedule.

Building AI Soft-Sensor Models for Flotation Performance

Mon, 21 Sep 2026 00:00:00 +0800

Key Takeaways

Throughput, recovery, flotation kinetics, and concentrate grade can each be modeled as a “soft sensor” — an ML model that continuously predicts plant performance from live sensor/historian data instead of a static simulation.
JKSimFloat and METSIM remain the standard purpose-built flotation simulators for offline circuit design and what-if analysis.
Imubit is the clearest AI-native product in this space — closed-loop reinforcement-learning optimization already deployed across process industries, including flotation-adjacent applications.
Published plant studies use NSGA-II multi-objective genetic optimization combined with ML feature selection to lift recovery beyond what static simulators achieve alone.
Concentrate-grade ML prediction is the least mature of the four (Emerging) — grade modeling still runs mostly inside the simulators rather than as a standalone real-time model.

TL;DR

Use JKSimFloat/METSIM for offline flotation circuit design, then layer a continuously-retrained ML model (or a closed-loop platform like Imubit) on top of live plant sensor/historian data to predict and optimize throughput, recovery, and grade in real time rather than relying on a static simulation.

Flotation

Mon, 21 Sep 2026 00:00:00 +0800

Key Takeaways

Flotation is one of the most mature applications of industrial AI in mineral processing — this isn’t speculative.
Imubit’s closed-loop AI uses reinforcement learning to write optimal setpoints directly to the plant’s existing control system, not just recommend them to an operator.
Reported gains: up to ~15% recovery improvement and ~20% reagent-use reduction — real deployments, not lab results.
Froth-camera vision models (including ConvLSTM-based research systems) are the sensing layer that feeds these controllers real-time froth texture and bubble-size data.

TL;DR

Feed a froth camera’s real-time image stream and plant sensor data (reagent dosage, pulp density, pH, airflow) into a closed-loop reinforcement-learning controller like Imubit, and let it write setpoints back to your DCS in real time instead of relying on operators to react to trends.

Mill Surveys

Mon, 21 Sep 2026 00:00:00 +0800

Key Takeaways

Traditional mill surveys are manual, periodic snapshots — a team samples the circuit for a shift and calculates a mass balance after the fact.
That’s being displaced by continuous capture: 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 are still layering 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

Replace periodic manual mill surveys with continuous historian-based data capture, then apply industrial-AI analytics (like Imubit) on top to get survey-grade mass-balance insight in real time instead of once a quarter.

Operational Variability & Plant Integration

Mon, 21 Sep 2026 00:00:00 +0800

Key Takeaways

The core problem: your design model (from testwork) and your actual plant behavior drift apart as ore feed varies — this is what “operational variability” means in practice.
Digital twins and industrial-AI platforms are purpose-built to reconcile that drift continuously, rather than catching it in a quarterly review.
Imubit builds its process model directly from plant historian data, so it naturally captures real operational variability rather than idealized design assumptions.
Rockwell’s PlantPAx and equivalent DCS platforms from Honeywell/Emerson are the control-layer backbone these AI/digital-twin tools plug into.

TL;DR

Reconciling plant variability against your design model in real time — instead of after the fact — is now a solved problem at the platform level: connect a digital-twin or industrial-AI layer (Imubit, or a Honeywell/Emerson digital twin) to your DCS historian and let it flag and adapt to drift continuously.