A processing plant separates ore containing a small proportion of a valuable metal, such as copper or gold, into a valuable product, the concentrate, and a waste stream, the tailings. Recovery is the proportion of the metal in the feed that reports to the concentrate: a recovery of 90% means that 90 of every 100 tonnes of copper fed to the plant report to the product and 10 are lost.
Recovery could be calculated by weighing every stream: the tonnes and grade of the feed and of the concentrate. In practice, plants weigh the feed accurately on a belt scale but often cannot weigh the concentrate or tailings every shift. All three streams are, however, sampled and assayed routinely. The formulas in this article calculate recovery from the assays alone, on the principle that every tonne of solids and every tonne of metal in the feed must leave in either the concentrate or the tailings.
This article sets out the two formulas most used by metallurgists and the limitations of each.
The two-product formula
For a single separation, the feed splits into a concentrate and a tailing. With F, C and T denoting the solids mass flows and f, c and t their assays, two balances must hold:
where
- , ,
- solids mass flow of feed, concentrate and tailings (t or t/h)
- , ,
- assay of feed, concentrate and tailings (% or g/t)
Eliminating T gives the mass yield to concentrate and, from it, the recovery:
where
- mass yield to concentrate (fraction of feed mass)
- recovery of the metal to concentrate (fraction; multiply by 100 for %)
Only the three assays are required. Tonnage does not enter the equation, which is why the formula is used every shift in flotation and gravity circuits.
Worked example
A copper rougher circuit has a feed of 1.20% Cu, a concentrate of 25.0% Cu and a tailing of 0.15% Cu. These are illustrative figures.
| Quantity | Calculation | Result |
|---|---|---|
| Mass yield to concentrate | (1.20 − 0.15) / (25.0 − 0.15) | 4.23% |
| Copper recovery | 25.0 × (1.20 − 0.15) / [1.20 × (25.0 − 0.15)] | 88.0% |
Sensitivity to assay error
A change of ±0.01% Cu in the tailings assay alone moves recovery from 88.0% to 88.8% (tailing 0.14%) or 87.2% (tailing 0.16%). A difference of one hundredth of a percent in a low-grade stream shifts the reported recovery by most of a percentage point. The formula is most sensitive when the feed grade is close to the tailings grade and when the concentrate is only slightly richer than the feed. For this reason, tailings sampling and assay precision deserve as much attention as the concentrate.
Recovery in a leach plant
In a CIL or CIP plant the solids pass through the leach with almost no change in mass, so the dissolution recovery on a solids basis reduces to:
where
- solids (dissolution) recovery, before soluble losses (fraction)
- head grade of the leach feed (g/t)
- grade of the leached solid residue (g/t)
With a head grade of 2.00 g/t and a solid residue of 0.15 g/t, the solids recovery is 92.5%. Two qualifications apply. Gold that dissolved but left the plant in tailings solution or on fine carbon has not been recovered, and must be subtracted to obtain plant recovery. The head grade may be the assayed head from the feed sampler or a calculated head built up from the residue and the recovered gold. Both should be tracked, because a persistent difference between them is often the first sign of a sampling bias.
The three-product formula
When the feed splits into two concentrates and a tailing, for example lead and zinc concentrates, a single element is no longer sufficient. With two assayed elements, a and b, the mass yields x = C1/F and y = C2/F are obtained by solving two balance equations simultaneously:
where
- ,
- mass yields to concentrate 1 and concentrate 2 (fractions of feed mass)
- ,
- feed assays of elements a and b
- ,
- concentrate 1 assays of elements a and b
- ,
- concentrate 2 assays of elements a and b
- ,
- tailings assays of elements a and b
where
- determinant of the two balance equations; a value near zero means the result is unstable
- recovery of element a to concentrate 1 (fraction); other recoveries follow the same pattern
The recovery of element a to concentrate 1 is then x·c1a / fa, and similarly for each element and product.
Worked example
Illustrative lead-zinc assays:
| Stream | Pb % | Zn % |
|---|---|---|
| Feed | 3.0 | 5.0 |
| Lead concentrate | 65.0 | 6.0 |
| Zinc concentrate | 1.5 | 52.0 |
| Tailings | 0.3 | 0.4 |
The solution gives a mass yield of 4.02% to the lead concentrate and 8.48% to the zinc concentrate, with 87.5% of the feed mass reporting to tailings. Lead recovery to the lead concentrate is 87.0% and zinc recovery to the zinc concentrate is 88.2%. The same figures show the misplaced metal: 4.8% of the zinc reports to the lead concentrate and 4.2% of the lead to the zinc concentrate. Substituting the yields back into both balances closes them exactly, which is a useful check on any spreadsheet.
Limitations of the three-product formula
The determinant D measures how different the two concentrates are. If they have similar assays for both elements, D becomes small, and small assay errors produce large, even negative, mass yields. If calculated yields vary from shift to shift while plant conditions are steady, D should be checked before the circuit is blamed.
Formula recovery and metal accounting
Formula recoveries are well suited to shift control and trend monitoring, but they are not a substitute for month-end metal accounting. A defensible monthly figure rests on measured tonnes, regular stocktakes and a balance in which every input and output is accounted for. This check-in, check-out principle is set out in the AMIRA P754 Code of Practice for Metal Accounting. When formula recovery and accounted recovery disagree month after month, the difference is informative, and usually points to a weightometer, a sampler or a stock estimate that requires attention. Reconciliation of the data is covered in mass balance data reconciliation.
In Saldora, a recovery formula can be configured as a calculation on the daily measurement set, so that the shift figure and its inputs sit alongside the month-end reconciliation.
Glossary
- Feed
- The ore entering the plant or a part of it.
- Concentrate
- The valuable product stream, rich in the metal being recovered.
- Tailings
- The waste stream left after the valuable mineral has been removed.
- Assay
- The measured grade of a sample, such as % copper or grams of gold per tonne.
- Grade
- How much metal is in the material, for example 1.2% Cu or 2.0 g/t Au.
- Mass yield
- The share of the feed's mass that reports to the concentrate.
- Recovery
- The share of the metal in the feed that reports to the product.
- Residue
- The leached solids leaving a leach plant.
- Head grade
- The grade of the plant feed.
Frequently asked questions
Why calculate recovery from assays instead of weighing the concentrate?
Concentrate and tailings are often not weighed every shift, whereas samples of all three streams are taken routinely. The formula converts those assays into a recovery without the tonnages.
Is the two-product formula accurate?
It is exact if the assays are exact. In practice it is very sensitive to the tailings assay, so a small sampling or assay error can move the result by about a percentage point.
What is the difference between formula recovery and accounted recovery?
Formula recovery is calculated from shift assays and is used to operate the plant. Accounted recovery is based on measured tonnes, stocktakes and a reconciled month-end balance, and is the figure reported to the business.
When is the three-product formula used?
When the feed splits into two concentrates and a tailing, such as lead and zinc concentrates, and assays for two elements are available for every stream.
Discuss your plant with a metallurgist
Tiegel is a Perth metallurgy company. We help with testwork, plant optimisation, metal accounting and investigations, and we build the Tiegel Suite software.
Talk to a metallurgistReferences and further reading
- Wills, B.A. and Finch, J.A. (2016). Wills' Mineral Processing Technology, 8th edition. Butterworth-Heinemann. Chapter 3, Metallurgical accounting, control and simulation.
- Gaylard, P.G., Randolph, N., Wortley, C.M.G. and Ralston, I. (2006). Designing for metal accounting. Journal of the Southern African Institute of Mining and Metallurgy, 106. Link
- Applying the AMIRA P754 Code of Practice for Metal Accounting. AusIMM MetPlant 2008. Link
Worked examples and charts marked illustrative use invented figures, not data from any operation.