Today
One recipe across very different cells
- Control parameters are applied at line or section level.
- Cell differences are handled mainly by operator attention.
- Energy intensity is reviewed as a line average.
Every pot is different.They all run the same recipe.
A pot line runs hundreds of cells under largely common control parameters, even though cell age, lining condition and thermal history differ. INSUS learns how each cell actually behaves, so feeding and voltage recommendations follow the cell rather than the line average. Day one is advisory above the existing pot control system.
01
Per-cell models, not a line-level average.
02
Read-only above your existing pot control system.
03
Cell life stays part of how success is measured.
The problem
A process held stableby running it conservatively.
A Hall-Héroult reduction cell is a coupled thermal, electrical and chemical system. Run it too cold and it can sludge; run it too hot and it consumes more energy and can accelerate lining damage. Operating with margin is rational, but that margin also hides how much of it each individual cell is carrying.
01Energy intensity varies cell to cell
02Anode effects start with too little dissolved alumina
03Cell age and lining change behaviour
- 01Energy intensity varies cell to cell
- 02Anode effects start with too little dissolved alumina
- 03Cell age and lining change behaviour
The shift
From line-level parametersto per-cell behaviour.
The stability constraint does not change. What changes is knowing how much operating margin each cell is carrying, and where that margin can be reduced without touching thermal, chemical or cell-life limits.
With INSUS
Recommendations that follow the cell
- Recommendations reflect each cell's learned behaviour.
- Individual drift is visible before it needs an intervention.
- Energy intensity is attributed per cell to operating conditions.
How it works
From cell signalsto a safer operating point.
This fits OptiStudio as process and energy decision support for primary aluminium pot lines. Day one is advisory and read-only. It does not replace the incumbent pot control system.
Connect
Bring the cell history together
Connect per-cell pot control data, voltage and resistance, alumina feed events, anode-cathode distance adjustments, anode-effect history, bath chemistry and temperature, metal height, anode changes, tapping records, cell age and lining history.
Understand
Learn how each cell behaves
Build an operating model per cell, learning how alumina feeding, resistance behaviour, thermal state and measured bath chemistry relate on that cell, including how the relationship shifts as it ages.
Predict
See instability before it becomes an event
Estimate near-term alumina concentration, approach to instability and rising anode-effect risk for each cell, rather than relying only on line-level indicators or the event log after the fact.
Advise
Recommend the lowest-energy stable point
Generate feeding and voltage recommendations that target lower energy input while holding modelled thermal and chemical stability margins. Flag cells drifting toward instability or abnormal anode consumption early.
Verify
Measure against what actually happened
Compare recommendations with realised DC energy per tonne, anode-effect frequency, bath chemistry and cell-level outcomes. Longer-term cell and lining results stay part of validation, so efficiency is not bought with cell life.
Getting started
Start with one cell groupand the history you already keep.
For primary aluminium smelters that already instrument pot lines densely. Delivered through OptiStudio as per-cell stability and energy decision support. Expert Services can help with data assessment and validation. Read-only above existing pot control. No new sensors. Closed loop is only a later question after validation.
- What we need. Pot control and historian data covering cell voltage, resistance, noise indicators, current, feed events and anode-cathode adjustments, with bath and metal analysis as ground truth, cell and lining history, and energy and production records.
- What you see first. One defined group of cells spanning different ages, run in shadow mode. The first objective is an honest per-cell baseline and a clear answer on whether measurable headroom exists at all.
- Where the limits are. Existing pot control can already be highly optimised, so available headroom is plant-specific. Historical data retention can limit modelling depth. Any later supervised stage retains hard stability constraints and operator oversight.
Related for metals
Other metalsstarting points
Pot-line energy sits beside scrap charge advice and CBAM documentation on the same metals value chain.
- 01Scrap Charge Blend AdvisorRecommend scrap charge blends against chemistry, cost and process constraints using the material and melt records you already keep.
- 02CBAM Embedded Emissions AssemblerAssemble embedded-emissions evidence for CBAM declarations from production and energy records, with figures traceable back to source.
- 03Metals industryThe wider picture for nonferrous plants: smelting energy, scrap and charge chemistry, and carbon evidence around the tonne.
Questions potroom teams ask us
Cells on the same line differ in age, lining and thermal history, yet often run a common recipe. This use case builds per-cell models, flags anode-effect risk early and recommends feeding and voltage adjustments within stability limits.
Per-cell control history, voltage and resistance, alumina feed events, bath and metal analysis, anode-effect logs, cell age and lining records, production information and energy data. No new sensors.
No. During the pilot it is a read-only layer above the existing control system. The aim is to find where headroom exists, not to replace control that is already doing its job.
No. Recommendations remain subject to potroom engineering and operator review. It adds cell-level information and decision support, not authority.
A supervised or closed-loop stage can be considered for a defined cell group after validation. Any such step retains hard thermal and chemical stability constraints and operator oversight.
Through OptiStudio as process and energy decision support for the pot line. Expert Services can support data assessment, integration and validation with your potroom team.
Headroom is plant-specific and may be small where pot control is already sophisticated. Historical data retention can limit modelling depth. This is operational optimisation. It does not change anode technology or the materials system.
Primary smelting
Your best cell and your worst cell run the same recipe.
Start with the spread. Use your own per-cell data to see how energy intensity, stability and anode-effect behaviour actually vary across the potroom, before deciding whether there is headroom worth pursuing.