D3 · Publication Volume 19

Mine Geology, Grade Control and Reconciliation

Connects daily mine geology and grade control to material movement, reconciliation and model learning.

Purpose and boundary of this book

Mine geology connects an evolving geological interpretation to short-horizon material decisions. It asks what material is present, where a boundary is likely to be, which observation supports that interpretation, what can be selected by the mining method, where broken material moved, and why later measurements differ from earlier predictions. Grade control is therefore not a final assay attached to a truck. It is a chain of sampling, interpretation, classification, communication, movement and verification decisions.

This book develops that chain from long-term and short-term models through grade-control sampling, rapid model updates, ore–waste boundaries, dilution, ore loss, stockpiles, material tracking, dig lines, reconciliation, root-cause analysis and feedback to the resource model. The emphasis is on comparable quantities and traceable states. A model block, a blast parcel, a surveyed excavation, a stockpile inventory and plant feed are not automatically the same population, support, location, time or material condition.

The book does not approve a mining plan, certify a resource or reserve, prescribe a blasting or sampling procedure for a real site, determine a cut-off policy, perform metallurgical accounting, or replace the judgement of appropriately qualified geology, mining, survey, processing, laboratory, safety and environmental practitioners. Site controls, legal duties, hazards, equipment limits and reporting requirements must be established for the actual setting.

General and institution-neutral scope

This is a general, institution-neutral tutorial. It has no relationship to, affiliation with, sponsorship by, endorsement from or curriculum dependency on any company or individual. It is not written for a named owner, operator, consultancy, university, regulator, software product, property, deposit, mine or private database. Every unnamed bench, heading, stope, face, blast, sample, block, dig line, truck movement, stockpile, plant batch and grade value is synthetic teaching material.

Names of standards bodies, public agencies, researchers and technical publications occur only in source notes when needed to identify evidence. A citation does not make a named person or organisation the author, publisher, sponsor, provider, partner, endorser, scientific authority or subject of this tutorial. The website carrying these pages is only a host and delivery surface. It is not the tutorial's author, publisher, sponsor, provider, owner, scientific authority or curriculum subject, and it asserts no institutional ownership of the curriculum.

Institutional neutrality is also a technical control. A recognised name cannot make an unrepresentative sample representative, align two different time windows, recover an unrecorded stockpile movement, or explain a residual by assertion. Credibility comes from observable boundaries, declared measurement bases, versioned interpretations, explicit transfer events, reproducible balances and conclusions that remain proportional to the evidence.

The material-state contract

Before comparing any two numbers, define a material-state contract:

  1. Spatial boundary: the exact polygon, solid, excavation, stockpile or flow boundary included.
  2. Time boundary: event timestamps, reporting cut-off, opening inventory and closing inventory.
  3. Material state: in situ, designed, broken, loaded, hauled, stockpiled, reclaimed, fed or processed.
  4. Mass basis: volume-derived or weighed mass, density method, moisture basis and included contaminants.
  5. Quality basis: analyte or property, unit, sample support, analytical status and compositing rule.
  6. Classification rule: ore, waste, marginal, blend class or other decision category and its effective version.
  7. Measurement uncertainty: precision, bias evidence, missing coverage and known correlations.
  8. Lineage: source records, transformations, approvals, revisions and superseded outputs.

Two quantities become reconcilable only after these dimensions are compatible or an explicit conversion bridges them. The contract prevents a monthly plant feed total from being compared directly with the same month's blasted model without accounting for stockpile residence, unhauled broken material and the plant's measurement basis.

Learning outcomes

After completing the book, the learner should be able to:

  • distinguish strategic, tactical, short-term and shift-scale geological decisions;
  • design a grade-control sampling plan around support, representativity, contamination and turnaround constraints;
  • update a short-term geological model without erasing interpretation history or uncertainty;
  • convert a geological and grade interpretation into selective mining units and auditable ore–waste boundaries;
  • separate planned dilution, unplanned dilution, ore loss and material misclassification;
  • maintain a mass-and-quality ledger for stockpiles and material movements;
  • issue coordinate-safe, time-valid and version-controlled dig-line instructions;
  • reconcile long-term model, short-term model, mined geometry, movement records, inventories and plant measurements on comparable bases;
  • diagnose variance through a causal evidence tree instead of forcing a balancing factor; and
  • return operational evidence to the resource model through controlled learning loops.

Prerequisites, quantities and notation

The book assumes familiarity with geological domains, three-dimensional models, sampling, descriptive statistics and resource estimation. Mass is written M, grade or quality as g, and contained quantity as Q=Mg when g is a compatible mass fraction. For a parcel assembled from components i, the mass-weighted grade is


g_{\mathrm{blend}}=\frac{\sum_i M_i g_i}{\sum_i M_i}.

For a comparison in which A is the reference and B is the later comparable quantity, a factor may be written F_{B/A}=B/A. A percentage variance is 100(B-A)/A when A\ne0. Neither number is meaningful until numerator and denominator share boundary, time, state and basis. A residual balance for a bounded period is


R=M_{\mathrm{open}}+M_{\mathrm{in}}-M_{\mathrm{out}}-M_{\mathrm{close}}-M_{\mathrm{known\ loss}}.

A non-zero residual is a diagnostic signal. It is not automatically physical loss, measurement error or an adjustment to be hidden.

Synthetic teaching system

Worked examples use a fictional polymetallic lens cut by a synthetic fault and mined through two unrelated teaching scenarios: a bench-scale surface parcel and a small underground extraction panel. The local coordinate system, geology, grades and operating records were invented. No coordinates locate a real place. The example register contains a long-term block model, close-spaced control observations, surveyed shapes, shift movement events, three stockpile classes and daily feed composites.

The examples deliberately contain contradictions. One blast-hole interval has wet contamination, a channel sample crosses a geological contact, a revised boundary arrives after loading begins, two truck events share an identifier, a stockpile survey uses a different moisture assumption, and plant feed includes material reclaimed from an earlier period. These defects create investigation tasks; they do not describe any organisation or imply typical performance.

Workflow and minimum evidence artefacts

The operational chain is represented by linked artefacts: a sampling protocol, sample and assay status register, interpretation change log, short-term model manifest, selectivity definition, boundary uncertainty layer, dig-line release, excavation survey, movement-event ledger, stockpile balance, feed measurement record, reconciliation workbook and corrective-action record. Each artefact needs a stable identifier, effective time, spatial reference, units, material state, input versions, transformation method and limitation note.

Preserve observations separately from interpretations and decisions. A mapped contact is an observation; a three-dimensional surface joining contacts is an interpretation; an ore polygon derived from that surface and a decision rule is an operational decision. A later model may supersede a surface without deleting the older version that governed an earlier movement. Reconciliation depends on reconstructing what was known and instructed at the time, not applying today's interpretation retrospectively.

Assessment and completion standard

Completion requires a synthetic model-to-mine-to-feed review package. The learner must define comparison contracts, test sample representativity, issue a versioned short-term interpretation, generate and communicate selective boundaries, maintain material lineage, close stockpile balances, calculate comparable reconciliation factors, quantify important uncertainty and diagnose one multi-cause variance.

The final review must distinguish an observation, estimate, allocation, conversion, correction and unexplained residual. It must also show at least one alternative boundary, one timing sensitivity, one moisture or density sensitivity, and one case in which apparent variance disappears after the comparison bases are aligned. A satisfactory conclusion states what is known, what is inferred, what remains unresolved and which new observation would discriminate between competing causes.

Core sources