Ore Blending and Grade Scheduling

Ore blending combines material from different sources to meet a specified feed or product objective. Grade scheduling decides when those sources are mined, delivered, stored and reclaimed. The two decisions are linked: a blend calculation describes a mixture, while its schedule must make the required material available in the correct period. Published stockpile scheduling formulations consequently couple quality requirements with material quantities and operational constraints.^[1]

The engineering objective is a feasible sequence of feed deliveries that supports production and product requirements. Maximizing head grade alone does not define that objective. The relevant boundary might be a crusher feed, a concentrator feed or a saleable bulk product; its location determines which measurements, inventories and quality specifications belong in the calculation.

Establish the Quality and Processing Basis

A blend specification may include valuable components, deleterious elements, mineralogical characteristics and physical properties. Geometallurgical work connects geological variability with processing behaviour through representative testing and spatial models. Bulled and McInnes show why flotation planning must consider relationships among mineralogy, grind, throughput and metallurgical response. A grade model by itself cannot establish the recovery or concentrate quality of every proposed mixture.^[2]

For each requirement, state the measurement basis, receiving location and averaging period. A monthly average can satisfy a target while individual deliveries violate an operating limit. Conversely, imposing the same narrow target on every truckload may unnecessarily restrict a blend that is deliberately assembled over a larger batch. The appropriate interval follows the actual storage, mixing and processing arrangement.

Separate three kinds of requirement:

• Mandatory bounds: limits that the selected route, process or product specification must satisfy.
• Operating targets: preferred conditions within those bounds, such as a stable feed composition.
• Economic preferences: choices about value, handling cost or inventory use among feasible alternatives.

A numerical penalty for missing a target does not make an otherwise unacceptable blend permissible. Record which requirements allow negotiated deviation and which make a proposed schedule infeasible.

Use Consistent Mass and Grade Balances

Grades and tonnages must refer to compatible material bases. Where assay grades are reported on dry solids, wet truck or belt tonnages require a corresponding moisture determination before calculating contained metal. Metal-accounting practice also requires traceable mass measurements and representative samples; an assumed or poorly matched moisture content can bias the dry-tonnage estimate.^[3]

For a batch formed without material loss or reaction, with non-negative source masses and positive total mass, conservation gives:

M = sum(m_i)
C = sum(m_i * g_i)
G = C / M

Here m_i is the dry mass from source i, in tonnes; g_i is its component grade as a mass fraction, in tonnes of component per tonne of dry material. M is total dry mass, C is contained component mass in tonnes, and G is the resulting mass fraction. Percentage assays must first be divided by 100. Stockpile blending models use this weighted-composition principle.^[4]

Apply the balance separately to each conserved component. An unweighted arithmetic mean is justified by equal contributed dry masses; otherwise use the mass-weighted balance. Use dry bulk densities when converting volume proportions to mass proportions. The calculation predicts the batch average; it does not demonstrate uniform composition at every point in the pile or every instant of reclaim.

Contained metal is also distinct from recovered metal. Recovery is the fraction of a specified feed component reporting to a defined product. Even with fixed, independent recoveries, the component balance gives an overall recovery weighted by contained component mass, not simply by ore tonnes. Actual blending may change processing conditions, so recovery and concentrate grade require a response model supported by tests for the proposed mixtures and operating conditions.^[2]

Carry Inventories Through Time

A schedule needs a balance for every stockpile and period. In a simplified model without losses, closing dry inventory equals opening inventory plus receipts minus reclaim; the corresponding contained-component inventory must also balance. Stockpile blending models explicitly carry quantities and composition between successive decisions.^[4]

Timing within a period matters. Material delivered late in a shift cannot support an earlier reclaim instruction simply because both movements appear in the same daily total. Likewise, a positive closing inventory does not prove that inventory remained non-negative throughout the period. Where this matters, divide the period or introduce event timing rather than treating all receipts as available at its start.

The inventory record should distinguish what exists from what can be used. Material may be awaiting assay, segregated for a particular route, inaccessible beneath another parcel, or unavailable while a receiving area is being prepared. These conditions define availability restrictions in addition to the mass balance.

A stockpile average is an estimate with a physical basis. Surveyed volume, density, moisture, recorded deliveries and reclaim measurements can disagree. Gaylard and colleagues identify uncertainty in stockpile measurements and emphasize investigating measurement bias during reconciliation. Differences should be investigated and corrections retained in the data history, rather than concealed by altering grades to force a balance.^[3]

Represent How Material Is Stacked and Reclaimed

Pre-crusher stockpiles can provide storage, operational buffering, blending and separation of grades. Jupp and colleagues explain that these roles can conflict: retaining distinct sources helps blend selection, whereas mixing them can reduce that flexibility. Their analysis of iron ore operations also distinguishes short-term smoothing from changes in the underlying mean grade. A stockpile cannot indefinitely compensate for a sustained shortage of suitable feed.^[5]

Physical arrangement determines which mixtures can actually be delivered. A schedule should reflect the accessible reclaim face, the stacking and reclaim sequence, and the material layers or parcels intersected. Treating an entire pile as instantly and perfectly mixed can obscure the relationship between deposited material and crusher feed. The necessary model detail depends on the intended stockpile function and reclaim practice.^[5]

This creates a useful acceptance test: can an operating instruction identify where each contribution will come from, how it reaches the receiving stream and when it is available? If the answer is unclear, the calculated percentages are insufficient to release a blend. Rehandling might make the mixture possible, but it also introduces a separate movement, resource demand and timing decision that must be scheduled.

Select Blends as Part of the Production Schedule

Optimizing an isolated blend from a fixed list of materials can miss choices available when mining, routing and stockpile decisions are considered together. Souza and colleagues formulate stockpile scheduling around quantity and quality requirements while representing material availability and movement constraints. Such formulations are useful only when their mathematical options correspond to physical operating choices.^[1]

For a particular planning period, test the proposed blend against the released mining sequence, available loading and transport resources, stockpile access and receiving capacity. The grade schedule should identify both the source contributions and the inventories left for subsequent periods. A favourable blend today may consume the only material capable of correcting an expected quality imbalance tomorrow.

Consider a qualitative planning situation in which a lower-impurity source becomes temporarily unavailable. Keeping the previous percentages would create an impossible instruction. Substituting another source may preserve tonnage but breach a product limit; reducing throughput may preserve quality but miss a production commitment. The alternatives include rescheduling accessible sources, using suitable retained inventory, changing the delivery interval, or revising the production target. Each alternative requires its own material balance and a check of the remaining schedule. This is a decision comparison, not evidence that one response is universally preferable.

Reconcile Performance and Reassess Assumptions

Uncertainty concerns both the source estimates and their combination. Everett's iron ore simulation work emphasizes preserving relationships among component grades and the sequence of deliveries when preparing model inputs. Independent random grades can produce unrealistic mixtures. The same study links pre-crusher estimates to measured lump and fines compositions, illustrating why feed composition and saleable-product composition need separate, calibrated models in that operating context.^[6]

For planning, this implies testing plausible combinations of source qualities and arrival times, rather than applying unrelated variations to every input. A blend dependent on one uncertain corrective source deserves particular attention. Reporting only its expected average hides how strongly feasibility depends on that assumption.

Compare planned contributions with actual source movements, measured feed and product results, and reconciled inventories. Investigate whether a deviation arose from an incorrect source estimate, a different material movement, measurement bias or a processing response outside the assumed range. Updating only the next blend percentage leaves the underlying cause unresolved. A revised schedule should therefore carry forward both the corrected inventory and the reason the previous prediction changed.^[6]