Regrind milling is the equipment-level answer to the liberation question raised on particle size, liberation and recovery, and it sits within the plant cluster described on gold processing plants for tailings as a targeted, second grinding stage rather than a scaled-up version of the primary mill. The two pages work together: the liberation page establishes why finer grinding sometimes matters, and this page covers the actual equipment and circuit design a project builds to deliver it.
Where does a regrind stage sit in the flowsheet, and what does it actually grind?
A regrind stage sits after an upgrading step, typically a gravity concentrate, a flotation concentrate, or a coarse middlings stream, rather than processing the plant's entire feed tonnage. This targeted placement matters because it lets a project apply the extra energy and cost of finer grinding only to the fraction of material that has already been shown to actually need it, rather than regrinding barren gangue that gains nothing from finer liberation. On a tailings retreatment flowsheet specifically, a regrind mill commonly appears between a rougher gravity or flotation stage and a cleaner stage, taking the intermediate concentrate up to the finer grind its subsequent processing actually requires.
Ball mills vs stirred media mills for regrind: how is the choice made?
Ball mills remain a common choice for moderate regrind targets, using steel or ceramic grinding media tumbling inside a rotating cylindrical shell, the same basic mechanism used in a primary grinding circuit but scaled down and configured for a smaller, already-upgraded stream. Stirred media mills, which use a rotating internal shaft to agitate fine grinding media within a stationary shell rather than tumbling the whole mill body, are generally favored for very fine regrind targets because they achieve fine grinding more energy-efficiently at that size range than a conventional ball mill operating outside its efficient zone.
The crossover point between the two technologies is a function of target grind size and total tonnage rather than a fixed rule, and vendors of both mill types can generally provide grinding-energy test data for a specific feed to support the comparison. For a tailings retreatment project, this choice is made from bench-scale grinding testwork on the actual concentrate stream in question, not assumed from what a fresh-ore operation elsewhere might use for a superficially similar target size. Media cost and consumption rate also factor into the comparison over a project's operating life, since stirred mills typically use much smaller grinding media than a ball mill and consume it at a different rate, a real ongoing operating cost that deserves the same scrutiny as the initial capital comparison between the two mill types.
How is a regrind target size actually set?
A regrind target size comes directly from liberation testwork, typically the same QEMSCAN or equivalent mineralogical analysis described on particle size, liberation and recovery, which shows the specific particle size at which the gold in question becomes meaningfully liberated for the recovery method waiting downstream. Grinding finer than that testwork-established target wastes energy without recovering additional gold; grinding coarser than it leaves gold locked that a small additional grinding investment would have freed. Neither error is acceptable engineering practice once the liberation data exists, which is why regrind circuits are properly specified from testwork rather than a generic finer-is-always-better assumption.
How is a regrind circuit typically configured?
A regrind circuit is usually built as a closed circuit, pairing the mill with a classifier, commonly a cyclone, that returns oversize material back to the mill for further grinding while allowing correctly sized material to pass forward. This closed-circuit arrangement prevents over-grinding of particles that have already reached the target size while ensuring under-ground material gets the additional grinding time it needs, and it is standard practice on essentially every modern regrind installation rather than an optional refinement. Open-circuit regrinding, where material passes through the mill only once regardless of the size it comes out at, is rare in practice because it either under-grinds a portion of the stream or over-grinds another portion trying to compensate, neither of which is acceptable once a specific liberation target has been established from testwork.
Power draw and mill sizing for a regrind circuit are calculated from standard comminution testwork, typically using the feed material's measured grindability alongside the target size reduction the liberation study calls for. This calculation determines the mill's required installed power for the tonnage a project intends to regrind, and getting it right at the design stage avoids the common and costly mistake of installing a mill that cannot actually hit its target grind at the intended throughput.
What does regrinding cost, and when does it pay for itself?
Grinding energy consumption rises sharply, not linearly, as target particle size decreases, since breaking particles into progressively smaller fragments requires disproportionately more energy input per unit of additional surface area created. This means the cost of pushing a regrind target even modestly finer can escalate quickly, and that cost has to be weighed directly against the value of the additional gold the finer grind actually liberates, established from the same testwork that set the target in the first place.
The economic case for regrinding is strongest where testwork shows a genuinely large liberation gain over a comparatively modest additional grind, and weakest where most of a stream's gold is already liberated at its current size, in which case further grinding mostly just consumes energy on already-accessible material. A feasibility study proposing a regrind stage should show this liberation-gain-versus-energy-cost comparison explicitly rather than including regrinding as an assumed, default step in every flowsheet.
Because regrinding only ever applies to a reduced-volume, already-upgraded stream rather than the full plant feed, its absolute energy cost in kilowatt-hours per tonne of original feed is generally modest even when the grinding energy per tonne of the concentrate itself is high, since the concentrate represents only a fraction of total plant tonnage. This is precisely why regrinding a concentrate is routinely economic even when regrinding the entire feed to the same fine target would not be, reinforcing the mass-reduction logic that runs through this recovery methods set generally: apply expensive treatment only to the smaller stream that actually needs it.
How does regrind interact with the rest of the plant's water and thickening circuit?
Reground material typically needs re-thickening or reclassification before it can return to a gravity or leach circuit, since the regrind process itself dilutes the stream with additional water and produces a finer particle size distribution than the downstream equipment was necessarily sized around. Planning this water and classification interaction at the flowsheet design stage, rather than treating regrinding as an isolated unit operation, avoids bottlenecks appearing later at the connection points between the regrind mill and the stages on either side of it.
Water added during regrinding also has to be accounted for in the plant's overall water balance, since a project managing water carefully for cost or environmental reasons needs to know exactly how much additional water each regrind stage introduces and where that water is recovered or discharged. This is a modest but real addition to the water management picture covered more broadly elsewhere in the plant cluster, worth including explicitly rather than treating regrind water demand as negligible.
Why regrinding tailings differs from regrinding fresh ore
Tailings arrive at whatever grind the historic plant left behind, often coarser than a modern liberation target by the standards of current grinding and recovery technology, which means a regrind decision on tailings is frequently a genuine retrofit question rather than a routine design parameter chosen freely from a blank sheet, echoing the same inherited-grind logic already established on the liberation testwork page.
Wear and maintenance considerations specific to regrind mills
Regrind mills, whether ball or stirred media, experience continuous media and liner wear from the grinding action itself, and because a regrind stage sits at a critical connection point between an upgrading stage and its downstream cleaning or leaching step, unplanned regrind mill downtime can stall the whole concentrate stream behind it. Scheduled liner and media replacement, tracked against measured wear rates rather than a fixed calendar interval alone, keeps a regrind circuit performing at its tested capability.
Why regrind decisions belong in the same testwork program as equipment selection generally
Because the regrind target size, mill type, and circuit configuration all flow directly from liberation testwork, the regrind decision cannot really be separated from the gravity, flotation, and leach equipment decisions made elsewhere in a project's flowsheet design. A single, coordinated testwork program covering the whole recovery sequence, rather than siloed studies for each equipment category individually, is what gives a project an internally consistent flowsheet rather than a set of individually optimized stages that do not actually fit together well.