Grind Finer, Feed Less: A Concentrator Trade-Off
A 2.4% Cu porphyry concentrator treats 900 t/h at P80 = 150 micrometres with 88% rougher-scavenger copper recovery. Locked-cycle pilot work shows that grinding to P80 = 106 micrometres lifts recovery to 91.0%, but Bond work index testing and the installed power indicate throughput would fall to about 780 t/h at the same draw. Plant availability gives 7,600 operating hours per year. Concentrate assays 26% Cu and is sold under a treatment charge plus a 10% price participation. The mill is the bottleneck; the mine and haulage can supply 900 t/h. A single ball mill circuit with cyclones closes the grind; the scavenger tail carries visible coarse middlings.
1. Quantify the trade-off in payable copper per year and state whether the finer grind should be adopted plant-wide.
Read the worked answer
Coarse case (P80 150 micrometres)
- Throughput = 900 x 7,600 = 6.84 Mt/y
- Cu in feed = 6.84 Mt x 0.024 = 164.2 kt
- Cu recovered = 164.2 x 0.88 = 144.5 kt/y
Fine case (P80 106 micrometres)
- Throughput = 780 x 7,600 = 5.93 Mt/y
- Cu in feed = 5.93 Mt x 0.024 = 142.3 kt
- Cu recovered = 142.3 x 0.91 = 129.5 kt/y
Result A loss of about 15 kt of contained copper per year, roughly 10%. The recovery uplift is far too small to pay for the 13% throughput loss.
Decisive check What recovery would break even at 780 t/h? R = 144.5 / 142.3 = 101.6% - impossible. No recovery gain can compensate for the lost tonnes, so the answer is not 'it depends on the metal price'. Do not adopt P80 = 106 micrometres plant-wide.
2. How would you capture the liberation benefit without losing the tonnes? Give concrete options.
Read the worked answer
The loss at 150 micrometres is almost certainly in the coarse composite/middlings fraction, not spread evenly through the circuit - confirmed by the visible middlings in the scavenger tail. So treat the stream, not the whole plant.
Options, roughly in order of cost-effectiveness
- Regrind the rougher/scavenger concentrate only. A stirred mill (or a small ball mill) to P80 40-45 micrometres on a stream that is only 5-8% of feed mass costs a fraction of grinding 900 t/h. This is the standard answer and it directly targets the middlings.
- Fix the classification. Review cyclone geometry, apex/vortex sizing, pulp density and pump speed. Bypassing coarse material straight to overflow is a common, almost free cause of a coarse P80 and poor liberation.
- Flash flotation or a coarse-particle recovery unit on the cyclone underflow, recovering liberated values before they are overground or lost.
- Test an intermediate grind (for example P80 130 micrometres). The grade-recovery curve is usually steep at the coarse end and flat near liberation; do not assume linearity between the two pilot points.
- Reagent and pH optimisation, and check whether the locked-cycle recovery actually translates to plant - locked-cycle results are routinely optimistic.
Before committing capital, confirm the benefit is real with a plant trial on one circuit, and confirm the ore-domain variability in hardness and liberation size - the pilot is a single composite.
3. If the company can fund only one $40M project - a pre-crush/pebble circuit adding about 8% throughput, or a regrind circuit - how do you frame the comparison?
Read the worked answer
Compare payable metal and cash, not tonnes or recovery points
- Pre-crush: +8% throughput on 6.84 Mt/y = +547 kt/y x 0.024 x 0.88 = +11.6 kt Cu/y
- Regrind: +3 recovery points on 164.2 kt = +4.9 kt Cu/y
On that arithmetic pre-crush is worth roughly twice as much, and it is the simpler metallurgy. But that conclusion rests on assumptions that must be tested:
- Does the mine sustain the extra tonnes? The prompt says yes to 900 t/h, so confirm haulage and reserve life; if the mine is the real constraint, the pre-crush value evaporates.
- Does pre-crushing change the grind? A coarser mill feed can coarsen the product and lower recovery unless classification is managed - the two projects interact rather than compete.
- Concentrate quality. At 26% Cu, extra throughput may push arsenic, or penalty elements, or exceed smelter allocation, and the 10% price participation changes the marginal value per tonne of concentrate.
- Frame as NPV and payback, with a downside case on metal price and recovery, and check the incremental water, power and tailings capacity.
The judgement point: a 3-point recovery gain is seductive and easy to sell, but a 13% throughput reduction or a small throughput gain moves far more metal. Always convert metallurgical metrics into payable metal before ranking projects.