🎓 Lesson 4
D3
Mapping the Mine-to-Mill Data Pipeline
The mine-to-mill data pipeline is the step-by-step flow of information—from drill-and-blast design through ore handling, crushing, and grinding—that connects mining operations to metallurgical processing so engineers can make better decisions.
🎯 Learning Objectives
- ✓ Analyze blast fragmentation data (e.g., Kuz-Ram output) to predict SAG mill throughput and specific energy consumption
- ✓ Design a traceable data handoff protocol between blast planning software (e.g., SHOTPlus) and plant DCS/PI System for real-time feed characterization
- ✓ Explain how variations in rock mass rating (RMR) and blast-induced damage propagate through the pipeline to impact grinding circuit availability
- ✓ Apply statistical process control (SPC) methods to detect data drift between upstream (drill logs) and downstream (screen analysis) measurements
📖 Why This Matters
In 68% of operating mines, unplanned mill upsets stem from unanticipated changes in ore hardness or fragmentation—originating not in the plant, but in the blast design or geotechnical model. Without a robust mine-to-mill data pipeline, engineers operate blind: blasting teams optimize for throw and cost; metallurgists react to erratic feed size distribution. This lesson bridges that gap—teaching you how to build, validate, and sustain data continuity that directly improves recovery, reduces energy use, and extends equipment life.
📘 Core Principles
The mine-to-mill pipeline rests on three interdependent layers: (1) Physical layer—the actual material movement (rock → muckpile → haul truck → crusher → mill); (2) Data layer—the digital representation (e.g., blast hole coordinates, RQD, P-wave velocity, fragment size distribution (FSD), Bond Work Index (BWI)); and (3) Semantic layer—the agreed-upon meaning and units (e.g., 'F80' must mean 80% passing size in µm, not mm, and be derived from laser scan + sieve analysis—not estimation). Critical enablers include spatial-temporal alignment (matching blast round ID to corresponding mill feed batch), metadata tagging (e.g., 'GeoDomain: Pit-Alpha, Bench-7, Blast-2024-087'), and error propagation modeling (how ±5% error in RMR propagates to ±12% error in predicted BWI).
📐 Fragmentation-to-Grindability Mapping
This formula links blast-derived fragmentation (via Kuz-Ram) to mill-specific energy demand using a calibrated empirical relationship. It enables predictive comminution modeling before ore enters the plant.
Predicted Specific Energy (kWh/t)
E_pred = α × BWI × (F80_cm / 10)^0.25Estimates SAG/ball mill specific energy consumption based on blast-generated fragment size distribution and site-calibrated factor.
Variables:
| Symbol | Name | Unit | Description |
|---|---|---|---|
| E_pred | Predicted specific energy | kWh/t | Estimated energy required to grind ore to target P80 |
| α | Site calibration factor | dimensionless | Empirically determined multiplier accounting for ore texture, moisture, and circuit configuration |
| BWI | Bond Work Index | kWh/t | Laboratory-derived measure of ore grindability |
| F80_cm | 80% passing size of blast muck | cm | Fragment size metric obtained from photogrammetry or sieving |
Typical Ranges:
Hard, low-structure porphyry copper: 12–18 kWh/t
Medium competency banded iron formation: 9–13 kWh/t
💡 Worked Example
Problem: Given: Kuz-Ram F80 = 125 mm (from image analysis of muckpile), measured BWI = 14.2 kWh/t, and site-specific calibration factor α = 0.93 (determined from 3 months of paired blast/mill data), calculate predicted specific energy for next blast round.
1.
Step 1: Convert F80 to cm: 125 mm = 12.5 cm
2.
Step 2: Apply calibrated mapping: E_pred = α × BWI × (F80_cm / 10)^0.25 (based on Morrell’s modified Bond equation)
3.
Step 3: Compute: E_pred = 0.93 × 14.2 × (12.5 / 10)^0.25 = 0.93 × 14.2 × (1.25)^0.25 ≈ 0.93 × 14.2 × 1.057 ≈ 13.97 kWh/t
Answer:
The result is 13.97 kWh/t, which falls within the safe range of 12.5–15.8 kWh/t for this ore type and circuit configuration.
🏗️ Real-World Application
At Newcrest’s Telfer Mine (Western Australia), implementation of a unified mine-to-mill data pipeline reduced SAG mill liner wear variance by 34% and improved copper recovery by 1.8% annually. Key actions included: (1) assigning unique blast round IDs linked to GPS-tracked haul trucks; (2) embedding F80 and BWI estimates into the PI System via OPC UA; (3) triggering automatic mill speed and ball charge adjustments when predicted E_pred deviated >5% from baseline. The pipeline cut manual reconciliation time from 16 hours/week to <2 hours/week.