Calculator D2

Ore Variability Characterization Framework

A system that tracks how ore quality changes across a mine and uses that information to adjust processing and blasting so nothing gets wasted and everything runs smoothly.

Typical Scale
Applied at block model resolution: 5–20 m × 5–20 m × 2–10 m
Industry Standards
Aligned with JORC Code Appendix 1, SME Best Practices for Grade Control
Data Latency Target
≤ 4 hours from assay to circuit adjustment (real-time loop)
Key Sensor Types
LIBS, XRF, NIR, gamma-ray density, on-belt elemental analyzers

⚠️ Why It Matters

1
Inconsistent ore feed composition
2
Variable liberation size distribution
3
Fluctuating flotation recovery
4
Increased reagent consumption
5
Reduced plant throughput
6
Lower overall metal recovery

📘 Definition

The Ore Variability Characterization Framework is an integrated engineering methodology that quantifies spatial and temporal heterogeneity in ore grade, mineralogy, texture, and metallurgical response, enabling closed-loop coordination between geological modeling, mine planning, blast design, and comminution–separation circuit control. It operationalizes grade uncertainty through statistical rock characterization, real-time assay integration, and dynamic process model updating.

🎨 Concept Diagram

GeologyMiningProcessing

AI-generated illustration for visual understanding

💡 Engineering Insight

Ore variability isn’t noise—it’s signal. The most expensive 'grade loss' isn’t in the tailings pond; it’s in the 5–10% of material that sits just below economic cutoff but could be recovered with minor circuit tuning—if you know *which* 5% and *when*. That requires linking geostatistical confidence intervals directly to process controller setpoints—not just dashboards.

📖 Detailed Explanation

At its core, ore variability characterization begins with recognizing that orebodies are not uniform blocks but stochastic assemblies of lithologies, alteration zones, and structural controls—each imparting distinct physical and chemical behavior. Traditional resource estimation treats grade as a scalar field; this framework treats it as a multivariate random function with correlated attributes (e.g., Au grade, pyrite abundance, clay content).

Deeper integration emerges when grade variance is coupled with metallurgical response surfaces: a 2 ppm Au zone may yield 92% recovery if d₅₀_lib = 60 µm, but only 68% if d₅₀_lib = 140 µm—even at identical grind size. This necessitates joint simulation of grade and liberation—not sequential estimation.

At the advanced level, the framework incorporates time-varying parameters: oxidation state evolution in near-surface ore, seasonal moisture effects on conveyor belt sampling bias, and sensor drift correction using reference standards traceable to NIST SRM 2582. True real-time application requires embedded Bayesian updating—where each new assay updates posterior distributions for all linked properties, feeding predictive digital twins that simulate circuit response under alternative blending scenarios.

🔄 Engineering Workflow

Step 1
Step 1: Domain delineation (geological domains → litho-structural units)
Step 2
Step 2: High-density geochemical sampling (composite drill core, 1–2 m intervals)
Step 3
Step 3: Mineralogical characterization (QEMSCAN/MLA on representative composites)
Step 4
Step 4: Metallurgical testwork (grindability, leach kinetics, flotation response per domain)
Step 5
Step 5: Statistical domain modeling (variogram analysis, conditional simulation, SMU definition)
Step 6
Step 6: Real-time reconciliation (blending model + online analyzers + DCS feedback)
Step 7
Step 7: Dynamic circuit adjustment (mill speed, reagent dosing, cyclone pressure)

📋 Decision Guide

Rock/Field Condition Recommended Design Action
High grade variance (σ²_au > 6 ppm²) + low liberation size (d₅₀_lib < 75 µm) Implement ultra-fine ROM sorting (XRT or LIBS) and reduce crusher P80 to 12 mm
AC > 15 kg/t + HGI < 45 Install dedicated acid pre-neutralization stage and increase primary crusher reduction ratio by 20%
d₅₀_lib > 180 µm + σ²_au < 1.5 ppm² Optimize blast fragmentation to achieve P80 ≤ 150 mm; defer fine grinding until secondary crushing

📊 Key Properties & Parameters

Grade Variance (σ²_au)

0.2–12.0 ppm²

Variance of gold assay values (ppm) within a defined domain (e.g., 10 m × 10 m × 5 m block), quantifying spatial grade heterogeneity.

⚡ Engineering Impact:

Drives sampling density, selective mining unit (SMU) size, and cut-off grade optimization.

Liberation Size (d₅₀_lib)

45–250 µm

Particle size at which 50% of target mineral grains are fully liberated from gangue, determined by QEMSCAN or MLA analysis.

⚡ Engineering Impact:

Directly determines optimal SAG mill discharge grind size and dictates downstream flotation circuit configuration.

Hardgrove Grindability Index (HGI)

35–85 (unitless)

Empirical measure of relative ease of grinding coal or carbonaceous shale; adapted for carbonate-rich ores via calibrated proxy assays.

⚡ Engineering Impact:

Predicts specific energy consumption in crushing and grinding circuits—low HGI correlates with high SAG mill power draw.

Acid Consumption (AC)

2–25 kg H₂SO₄/t

Mass of sulfuric acid (kg/t) consumed during leach test to reach pH 1.5, indicating reactive carbonate content.

⚡ Engineering Impact:

Triggers pre-leach neutralization dosing and dictates cyanide stability in CIL/CIP circuits.

📐 Key Formulas

Selective Mining Unit (SMU) Size

SMU = k × √(σ²_au / d₅₀_lib)

Empirical estimate of minimum economically minable block size balancing grade risk and dilution.

Typical Ranges:
High-grade narrow vein
2.5–5.0 m
Bulk oxide deposit
10–25 m
⚠️ SMU ≥ 2× blast hole spacing to ensure controllable fragmentation

Acid Demand Proxy

AC ≈ 0.32 × CaCO₃_% + 0.47 × MgCO₃_%

Estimates acid consumption based on carbonate mineralogy from QEMSCAN.

Typical Ranges:
Dolomitic skarn
12–22 kg/t
Siliceous quartz vein
2–5 kg/t
⚠️ Error < ±15% validated against bottle roll tests

🏭 Engineering Example

Telfer Mine, Western Australia

Carbonate-altered felsic volcanics with disseminated Au-Cu sulfides
Acid Consumption (AC)
11.3 kg H₂SO₄/t
Hardgrove Index (HGI)
52
Grade Variance (σ²_au)
4.8 ppm²
SAG Mill Specific Energy
12.4 kWh/t
Liberation Size (d₅₀_lib)
82 µm

🏗️ Applications

  • ROM ore sorting calibration
  • Dynamic SAG mill liner wear prediction
  • Cyanide dosing optimization in CIL
  • Pre-concentration circuit selection

📋 Real Project Case

Open Pit Gold Mine Blast Optimization

Large copper mine expansion in Chile

Challenge: High vibration levels affecting nearby structures
Read full case study →

🎨 Technical Diagrams

Geological Domain MappingVolcanicCarbonateAlteration
AssayQEMSCANLeach TestDCS Setpoint

📚 References

[1]
Guidelines for the Characterization of Ore Variability and Its Impact on Processing — International Council on Mining and Metals (ICMM)
[3]
SME Mining Engineering Handbook, 3rd Edition — Society for Mining, Metallurgy & Exploration (SME)