SCADA, the historian and the ERP disagree on production tonnes because each was built to answer a different question: SCADA runs the plant this shift, the historian stores compressed time series, and the ERP infers production from despatch and stock movements. Production reconciliation fixes this with a documented mass balance that produces one figure, with its working shown, every period.
Key takeaways
- None of the three systems is wrong. None of them, on its own, is the reported answer.
- The AMIRA P754 Code of Practice for Metal Accounting already defines how to reconcile measured sources. Most sites never built the plumbing to apply it.
- Production figures now feed Safeguard Mechanism baselines and AASB S2 emissions intensity, so a spreadsheet rebuilt by one person each month is a compliance risk.
- The fix is a reconciliation layer between operational and reporting systems. It does not replace SCADA, the historian or the ERP.
Why do SCADA, the historian and the ERP report different tonnes?
At month end the metallurgist reads 48,200 tonnes off the historian. The SCADA HMI shows a running total near 49,000. Finance closes the period in the ERP at 47,600. (These figures are illustrative, but the pattern is common.) The gap is wide enough to change a royalty payment, an emissions intensity figure and the board's confidence in its own numbers.
The divergence is structural. Each system makes a defensible measurement choice for its own purpose, and each choice produces a different total.
SCADA totals are running counters on weightometers and flow meters, sampled at scan rate. A weightometer drifts between calibrations. A counter resets when a PLC is power cycled. The HMI figure is right for the operator and wrong for closing a month.
The historian, whether OSIsoft PI, AVEVA Historian, Citect Historian or an open platform, returns an integral of a flow tag over time. That integral depends on the compression deadband, the interpolation method and how data gaps were backfilled. Two engineers querying the same tag with different summary settings get different tonnages, and both are defensible.
The ERP knows what was despatched, invoiced and surveyed. It does not know what happened between the sensors and the stockpile, so it infers production from movements and adjustments, usually reconciled to the last stock survey.
| What it measures | Why it differs | Impact | |
|---|---|---|---|
| SCADA / HMI | Live running totals off weightometers and flow meters | Calibration drift, counter resets, no system of record | Real-time |
| Historian | Time-series integral of flow and weight tags | Compression deadbands and interpolation change the total | Configurable |
| ERP | Despatch, invoicing and surveyed stock movements | Infers production from inventory, reconciled to stock count | Financial |
| Metallurgical balance | Mass and metal balance across the flowsheet | Reconciles all sources to a single best estimate | Reconciled |
The problem is that the business keeps asking each system the question in isolation.
Who usually owns the reconciliation today?
On most Australian sites, one person. The reconciliation happens by hand in a spreadsheet a few days after month end, and the person who builds it is the only one who understands why the numbers move. That is key-person risk sitting on the most important operational figure the business produces.
Across our consultants' careers, including direct and indirect engagements at BHP, Rio Tinto and Senex Energy, the pattern repeats: the tonnes disagree because three systems were designed for three jobs, and nobody built the layer that reconciles them.
What is metallurgical accounting, and what does AMIRA P754 say?
Metallurgical accounting is the discipline of measuring, reconciling and reporting metal production to a defined standard. The industry reference is the AMIRA P754 Code of Practice for Metal Accounting, developed under AMIRA Project P754 in the 2000s. It is an industry code, not law.
It sets out principles covering mass measurement, sampling, analysis, data management, and the balancing and reconciliation methods that turn many measured sources into one reported figure.
The principle that matters most here: every reported production figure should carry a known accuracy, use the best measurement available at each point, and come from a documented mass balance rather than whichever system is closest to hand. A weightometer with a stated uncertainty is worth more than an uncalibrated running total.
On most sites that model lives in one metallurgist's spreadsheet, rebuilt monthly, with SCADA and historian data pasted in and the ERP figure typed at the bottom. The discipline is sound. The plumbing is missing.
What does a production reconciliation layer do?
A reconciliation layer sits between operational systems and reporting systems, runs the mass balance the same way every period, and shows its working.
- CollectPull tags from historian, totals from SCADA, movements from ERP
- CleanseFlag calibration gaps, resets and bad-quality data
- BalanceRun the mass balance, weight each source by uncertainty
- ReconcileProduce one figure with a stated accuracy and an audit trail
- ReportPublish to finance, compliance and the board from one source
Collect. A monthly tonnage from a historian means a time-weighted total over the exact reporting period, handling hours when the instrument was in calibration or data quality was bad, using the same summary method every month. Do that once, in code, and you remove the largest source of unexplained variance: two people querying the same historian two different ways.
Balance. Measured inputs, outputs and stockpile changes rarely close to zero on the first pass. A proper reconciliation distributes the residual across streams in proportion to their uncertainty. A trusted despatch weighbridge moves very little. A drifting in-plant weightometer absorbs more of the correction. The residual stays visible.
Report. The reconciled figure flows to the monthly production report, to climate disclosure and emissions intensity calculations, and to the board pack, from one source. Finance and the plant stop arguing because they read the same number.
Why does the production figure now matter for compliance?
For years a rough reconciliation was tolerable because the only consumer was internal reporting. That has changed.
The reformed Safeguard Mechanism commenced on 1 July 2023. It covers facilities emitting more than 100,000 tonnes of carbon dioxide equivalent a year, with baselines declining by 4.9 per cent a year to 2030 (DCCEEW). Many baselines are production-adjusted, so the emissions a facility is allowed depend on how much it produced. Your production figure is now an input to an obligation reported to the Clean Energy Regulator.
The same logic applies to AASB S2 climate disclosure, where emissions intensity divides emissions by output, and to any royalty tied to tonnes. When the number feeds a regulator, it has to be repeatable, documented and auditable. Our guide to Safeguard Mechanism emissions data covers the reporting side.
- Month-end reconciliation effortDays to hours
- Unexplained production varianceMade visible, not buried
- Audit position for Safeguard and AASB S2Documented and repeatable
- Single source of production truthFinance, plant and board aligned
Do you have a production reconciliation problem?
Some plants already reconcile well. The test is how the monthly number gets made and how it holds up when questioned.
How is last month's production figure produced today?
- One person rebuilds a spreadsheet each monthHigh risk: single point of failure, define the layer
- Whichever system finance can accessNo reconciliation, numbers will diverge unpredictably
- A documented mass balance, run the same way each periodSound: focus on automating and auditing it
- Nobody is quite sureStart with a data source review before anything else
If you sit in the top two rows, the figure your board and regulator rely on is more fragile than it looks. The fix does not mean ripping out SCADA or replacing the historian.
How does a reconciliation programme run?
Treat it as incremental work. Each stage produces a usable output before the next begins.
- Weeks 1-2Source reviewMap every tag, total and movement that feeds the number, and its trustworthiness
- Weeks 3-5Historian queriesBuild repeatable, quality-aware period totals from the historian
- Weeks 6-8Mass balanceCodify the AMIRA-style reconciliation with uncertainty weighting
- Weeks 9-10Reporting and audit trailPublish one reconciled figure with its working to finance and compliance
The source review is the step sites are tempted to skip, and it decides whether the rest works. Until every meter, weightometer, assay and stock movement is written down with how far it can be trusted, you cannot build a defensible balance. People inside the plant often stop seeing the assumptions they make every month, which is where an outside view helps.
For the reporting side of the same stack, see getting historian data into Power BI without building a second historian.
Why does buying a product rarely settle it?
Some metallurgical accounting products are good. The trap is believing the product is the answer. A package fed by the wrong historian queries produces reconciled nonsense, confidently. The value sits in integration decisions: which tag, over which period, summarised how, weighted against which despatch record. Those are specific to your flowsheet and instrumentation.
That is system integration work, and it belongs to whoever understands both the plant floor and the finance ledger. It sits naturally beside asset health programmes that draw on the same historian data.
Production, grade and reserve data is commercially sensitive, and critical infrastructure operators also carry Security of Critical Infrastructure Act obligations. See our guides to keeping regulated industry data onshore and CIRMP obligations before deciding where the layer runs.
Where should you start?
Write down how the number is made today, source by source, and ask how it would hold up if an auditor pulled the thread. That review usually takes about a week. It tells you whether you have a reconciliation problem, a data quality problem or a governance problem, and each needs a different fix.
How Solve8 can help
Our consultants and engineers have worked on historians, SCADA and production reporting across major Australian resources sites, and we build reconciliation layers that run inside your own systems with a documented audit trail. See our OT, mining and manufacturing data work or talk to us about mapping your production data sources before anyone proposes a build.
