Tiegel home Talk to a metallurgist

Ore blending for the mill: blend grade, the lever rule and throughput on mixed ores

Blending ore for the mill: tonnage-weighted grade, the two-source lever rule, the reason blend throughput differs from the average, and implementation of a blending plan.

· 4 minute read · By

Most mines feed the mill from several sources at once: different pits, stockpiles of different grades, or ore types that grind and leach differently. Deliberate blending evens out the feed, so that the plant operates steadily and gold production is predictable.

The blend grade calculation is straightforward, but blend throughput is less intuitive. This article covers both calculations and the operating practice required to achieve the planned blend in the mill. The ore blend grade calculator performs the grade calculation for up to four sources.

Blend grade and the lever rule

gblend=∑Migi∑MixA=gtarget−gBgA−gB

where

Mi, gi
tonnes and grade of each source
xA
fraction of source A in a two-source blend that hits the target grade
Blend grade against the share of the high-grade sourceBlending a 2.4 g/t source with a 0.9 g/t source: the blend grade rises in a straight line from 0.9 to 2.4 g/t. A 1.6 g/t target needs 46.7% of the high-grade source.1.01.52.02.5020406080100High-grade source in the blend (%)Blend grade (g/t)46.7%target 1.6 g/tBlend grade
Grade blends linearly by tonnes, so the mix that hits a target grade comes straight from the lever rule. Illustrative figures.

Worked example: 60,000 t at 2.4 g/t and 40,000 t at 0.9 g/t blend to (60,000 × 2.4 + 40,000 × 0.9) ÷ 100,000 = 1.80 g/t. To achieve 1.6 g/t from the same two sources, the high-grade proportion is (1.6 − 0.9) ÷ (2.4 − 0.9) = 46.7%.

Throughput of blended ore

Qblend=P∑xiEi

where

Qblend
mill throughput on the blend (t/h)
P
mill power available (kW)
xi, Ei
mass fraction of each ore type and its specific grinding energy at the target grind (kWh/t)
Mill throughput against the share of hard oreWith 3,000 kW of mill power, throughput falls from 300 t/h on all soft ore (10 kWh/t) to 231 t/h at a 50:50 blend and 188 t/h on all hard ore (16 kWh/t).150200250300020406080100Hard ore in the blend (%)Throughput (t/h)Throughput
Energy, not throughput, blends by weight, so throughput on a blend is not the average of the two throughputs: 231 t/h, not 244. Assumes the ores grind independently. Illustrative figures.

Worked example: a mill with 3,000 kW available treats a soft ore requiring 10 kWh/t at 300 t/h and a hard ore requiring 16 kWh/t at 188 t/h. A 50:50 blend requires 13 kWh/t, so the mill treats 3,000 ÷ 13 = 231 t/h, not the 244 t/h average of the two throughputs. This assumes that each ore type grinds as it would alone, which is a reasonable first estimate; in practice hard components can accumulate in the circulating load, so the estimate should be checked against plant data on the blend.

Recovery can also blend non-linearly. An ore containing preg-robbing carbon, or one with high cyanide or oxygen consumption, can lower the recovery of everything with which it is blended. Blends should be tested in the laboratory before a weighted average of single-ore recoveries is relied upon; see diagnostic leach and preg-robbing tests.

Blend implementation

Implementation requires knowledge of the grade, hardness and any problem minerals of each pit, stockpile and ore type, and stockpile records of tonnes and grade placed and reclaimed, confirmed by regular survey. The feed ratio is set for each shift, together with the feeder or loader rates that achieve it.

Plant feed grade and throughput are compared with the plan every shift, and any drift investigated promptly. Each month, the mine-reported grade is reconciled against the plant-measured grade; see gold reconciliation and F-factors.

Feed variability

Blending also reduces short-term variation. When the mill is fed from several sources at once, a poor zone in one source is diluted by the others, so grade and hardness vary less from hour to hour. This stability has value in its own right, because grinding, leaching and reagent dosing all perform better on a steady feed than on a variable one with the same average.

Glossary

Blend grade
The tonnage-weighted average grade of the sources in a blend.
Lever rule
The rule that gives the mix of two components needed to reach a target value between them.
Specific energy
Grinding energy needed per tonne to reach a target product size (kWh/t).
Ore type
A domain of ore with similar grade, hardness and metallurgical behaviour.
Stockpile
A pile of mined ore held for later feeding, often by grade or ore type.
Reconciliation
Comparing grades and tonnes predicted by the mine with those measured by the plant.
Preg-robbing
Adsorption of dissolved gold by carbonaceous matter in the ore.
Variability
How much grade or hardness changes from hour to hour or day to day.

Frequently asked questions

How is a blend grade calculated?

Multiply each source's tonnes by its grade, sum the products, and divide by the total tonnes. 60,000 t at 2.4 g/t and 40,000 t at 0.9 g/t give 1.80 g/t.

How is the mix for a target grade determined?

With the lever rule: proportion of source A = (target − grade B) ÷ (grade A − grade B). For a 1.6 g/t target from 2.4 and 0.9 g/t sources, the result is 46.7% of source A.

Why is throughput on a blend lower than the average?

Grinding energy is additive by mass, not by throughput. A blend requiring 13 kWh/t gives 3,000 ÷ 13 = 231 t/h, below the 244 t/h average of 300 and 188 t/h.

Does recovery blend linearly?

Not always. Ores with preg-robbing carbon, high cyanide consumption or high oxygen demand can lower the recovery of the whole blend, so blends should be tested in the laboratory.

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 metallurgist

References and further reading

  1. Wills, B.A. and Finch, J.A. (2016). Wills' Mineral Processing Technology, 8th edition. Butterworth-Heinemann. Chapter 2, Ore handling.
  2. Bond, F.C. (1961). Crushing and grinding calculations. British Chemical Engineering, 6.

Worked examples and charts marked illustrative use invented figures, not data from any operation.