Plant Optimization Software for Mineral Processing Circuits
MillMax turns live plant data into operating targets your control room can act on
so every shift runs the circuit the way your best operators would, while your existing control system holds stability.
WHAT'S BROKEN TODAY?
Results depend on who is on shift.
Your most experienced operators consistently outperform their peers. That judgement lives in their heads, is never written down, and leaves the plant when they do.
The variables that matter most are not measured.
True grind state, effective hardness, mineral response, recovery — these either cannot be instrumented or come back from the lab hours after the decision needed to be made.
Problems surface after they have already cost you.
Alarms fire once the process has drifted. By then the off-spec material is downstream, and the shift is spent recovering rather than optimising.
WHAT MAKES MILLMAX DIFFERENT
It understands the condition, not just the numbers.
The same sensor reading means different things depending on feed, equipment state and what happened upstream an hour ago. MillMax builds an understanding of the operating condition first, so every recommendation is made in context rather than pattern-matched against a tag.
It learns from your operators.
From years of decisions in the historian, MillMax infers the operating priorities your best people were working to — so that experience can be applied consistently on every shift.
It is built to improve, not decay.
Conventional models are tuned once and drift as conditions change. Model performance is monitored continuously, with controlled retraining as conditions change, and recommendations bounded by approved operating limits.
It works on partially instrumented circuits.
Soft sensors carry a confidence range with every estimate, so the optimiser knows when to trust them — covering instruments that are offline, measurements delayed in a lab, or variables that simply cannot be taken.
It sits above your control system, not in place of it.
Advisory and read-only from day one. The loop closes only when your operations team decides it is ready.
HOW MILLMAX WORKS
1. INPUTS
- Plant historian data (throughput, power, recovery, reagent usage, key control tags)
- Circuit configuration and operating limits (constraints, setpoint ranges)
- Process flow diagrams and P&IDs
- Laboratory results (assays, grind size, mineralogy, recovery tests)
- Standard operating procedures and operator logs
- Equipment availability and maintenance signals
- Cost sensitivities (energy, reagents, penalties)
2. OUTPUTS
- Operating targets — throughput, grind, reagent strategy and the recovery trade-offs between them
- Soft sensor estimates for variables that cannot be measured directly, each with a confidence range
- Early warning of drift and developing upsets, before alarm thresholds are crossed
- Clear recommendations explaining what to change, why, and what to expect
THE DEPLOYMENT WORKFLOW
Ingest live plant data
— from your historian, control system and plant documentation.
Establish operating context
— what condition the circuit is actually in, and what is driving the behaviour you are seeing.
Estimate unmeasured process states
— using soft sensors and validated models.
Predict plant response
— to changing conditions, with uncertainty attached.
Evaluate economic trade-offs
— throughput against recovery against reagent and energy cost.
Recommend operating targets
— within your safe and defined constraints.
Present explainable guidance
— to operators, or send setpoints to the DCS/SCADA once your team is ready.
Learn from outcomes
— as the plant responds — including every operator override — refining recommendations over time.
TOP KPIs IMPACTED
Recovery improvement
— typically 0.5–5% depending on circuit and variability
Throughput stability
— fewer oscillations and fewer corrective interventions per shift
Reagent efficiency
— up to 30% savings demonstrated in pilots
Energy efficiency
— improved kWh/t through better operating-point selection
Consistency across shifts
— the same standard of decision-making at 3am as at midday
CASE STORY
Reaching the highest extraction on record
A nickel refinery ran a multi-stage kiln circuit in which several of the variables governing extraction could not be measured in real time. Operators were making dosing and temperature calls on partial information, and results varied between crews.
MillMax deployed soft sensors to estimate the missing states, then moved to closed-loop optimisation above the existing APC.
Correcting off-target conditions in a leaching circuit
A base-metals refinery operating a high-pressure acid leach circuit needed tighter control of acid and steam dosing under shifting feed conditions. MillMax ran recommendations in parallel with the DCS, identifying optimal dosing as conditions changed.
PRIMARY USERS
Metallurgists
Control Room Operators
Processing Managers
Plant Managers
Technical Services Managers
Ready to turn plant variability into predictable performance?
Book a walkthrough to see how MillMax converts live plant signals into stable circuit performance and better recovery.
45-minute discovery session
— understand circuit objectives, variability drivers, and current control strategy.
Opportunity assessment
— review available plant data to size the upside, using historical data only.
Tailored walkthrough
— see MillMax recommendations in the context of your circuit and constraints.