Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

RESOURCES, RESERVES AND RECONCILIATION

By Harry Parker Kadri Dagdelen

BEST PRACTICE PRINCIPLES


Resource and Reserve Estimation is a serial
process: Data collection: geology, assays 3-dimensional geological solids model Exploratory data analysis Grade, density interpolation Resource classification/statement Mine design criteria Life-of-mine plans Reserve statement

DATA GATHERING DRILLING PROGRAMS


Wide-spaced drilling to identify size of the
prize Infill drilling to refine geometry, confirm ore controls, establish continuity, metallurgical ore types, contaminants Tighten spacing in risk areas: - Complex structure - Toe of ultimate pit - Apparent high-grade starter pits, stopes

GEOLOGICAL MODELING
3D solids modeling software continues to advance at rapid pace But basic problem is we continue to interpret what is on 2D plan or section without regard to adjacent sections A big problem where data are missing; under projection is common Painstaking manual adjustment still required.

GEOLOGICAL MODELING
When constructing other models for grade, alteration, consider local geological controls Ensure that modeler understands the controls to prevent time consuming editing

ZINC IS STRONGLY RELATED TO PROXIMITY TO SKARN/MARBLE CONTACT AND LESS TO SKARN TYPE --- 0.25 ---- 1.00 ---- 2.00 --- 3.00 --- 4.00 --- 5.00 ---

ZINC DOMAINS ----- 0.25 % Zn < 2.5 % Zn ----- 2.5% Zn

RESOURCE MODELING CHOICE OF INTERPOLATION METHOD


Simple methods preferred -Weighted averages more accurate than polygonal -Kriging marginally more accurate than inverse distance -Consider when CV < 1.5

CHOICE OF INTERPOLATION METHOD OTHER CRITERIA


To be useful, the resource model should reflect selectivity of mining Choose kriging neighborhood so distribution of kriged grades = distribution of selective mining units (SMUs) SMU is smallest practical volume that can be segregated to ore or waste Var SMUs = Var Comps Average( (h) inside SMU)

KRIGED 20 X 20 m Blocks % Cu
--- 0.25--- 0.50 --- 1.00 --- 2.00 --- 3.0 ---

KRIGED 20 X 20 m Blocks % Cu
--- 0.25--- 0.50 --- 1.00 --- 2.00 --- 3.0 ---

RECONCILIATION
Big Problem No One Wants to Talk About Often Done In Pieces May be Obscured/Invalidated by Stockpile
Accounting Hard to Determine Measurement Problems Accounting Problems People Problems

RECONCILIATION

Resource Model to Blastholes Reserve Model to Ore Control Ore Control to Mill Reserve Model to Mill

COMPARISON OF RESOURCE MODEL AND BLASTHOLES ----- 0.25 ----- 0.50 ----- 1.00 ---- 2.00 ---- 3.00 ----- % Cu

BLASTBLOCKS VERSUS RESOURCE MODEL BLOCKS FOR COPPER INDICATED BLOCKS (50 M NOMINAL SPACING)

BLASTBLOCKS VERSUS RESOURCE MODEL BLOCKS FOR COPPER MEASURED BLOCKS (25-30 M NOMINAL SPACING)

CHOICE OF INTERPOLATION METHOD CONDITIONAL BIAS


Kriged estimates that match SMU distribution may be locally conditionally biased Blocks estimated to be high grade will actually have lower grades; blocks estimated to be low grade will actually have higher grades We give up accuracy and accept this bias so that globally tonnage-grade curve matches SMUs. Internal dilution predicted correctly.

CONDITIONAL BIAS CHECK


Some conditional bias for Indicated No conditional bias for Measured Recognize that block model good for quarterly, semi-annual planning. Final ore/waste selection done later using blasthole samples. Alternatives: Use non-linear method OR Better: drill more close-spaced holes

RESOURCE TO RESERVE CONVERSION ULTIMATE PIT LIMITS AND MINE PLANS

RESERVE MODEL TO ORE CONTROL RECONCILIATION

10

FACTORS
F1 = Grade Control Depletions
Reserve Model Depletions

F2 = Received at Mill (heads)


Delivered to Mill (grade control)

F3 = Grade Control Depletions


Reserve Model Deplet.

x Received at Mill

Delivered to Mill

F1 = Grade Control Depletions Reserve Model Depletions

F1

F2
F2 = Received at Mill (heads) Delivered to Mill (grade control)

F3

F3 = Grade Control Depletions x Received at Mill Reserve Model Deplet. Delivered to Mill

11

F1

F1 = Grade Control Depletions Reserve Model Depletions Measures Local Accuracy of Reserve Model Checks to See if Right Support Implicit in Reserve Model Measures Bias of Grade Control Assays

F1

F2

F2 = Received at Mill (heads) Delivered to Mill (grade control) Measures Efficiency of Mining (unplanned ore loss, dilution) Measures Bias Between Grade Control and Mill

12

F1

F2

F3
F3 = Grade Control Depletions x Received at Mill Reserve Model Deplet. Delivered to Mill Measures Ability of Reserves to Predict Recoverable Tonnage, Grade, Metal to Mill

RECONCILIATION FOR COPPER ONLY ORES 2003-2004


Statistics Resource Model Depletions Ore Control Polygons (OCP) Received at Mil (RAM) Factors F1 = OCP/Res. Mod. Depl. F2 = RAM/OCP F3 = RAM/Res. Mod. Depl. 1.06 0.93 0.99 1.11 0.91 1.00 0.89 1.11 0.99 Mt 30.5 32.3 30.1 % Cu 1.51 1.67 1.52 % Zn 0.25 0.22 0.25

Possible high bias in blast hole grades Some shift of Cu-only to Cu-Zn ores at Mill

13

RECONCILIATION FOR COPPERZINC ORES 2003-2004


Statistics Resource Model Depletions Ore Control Polygons (OCP) Received at Mil (RAM) Factors F1 = OCP/Res. Mod. Depl. F2 = RAM/OCP F3 = RAM/Res. Mod. Depl. 1.13 1.02 1.15 1.11 0.94 1.05 1.05 0.81 0.86 Mt 23.9 27.1 27.6 % Cu 0.96 1.07 1.01 % Zn 3.04 3.21 2.61

Possible high bias in blast hole grades/OC Model Amount of Cu-Zn ores more than expected Selectivity of Cu-Zn ores worse than expected

RECONCILIATION
Review Past Production Versus Models Ideally Within 5% (Cu), 10% (Au) - Grade Control to Model Check Planned Dilution/Ore Loss (Aim for 0%) - Plant to Grade Control Check Unplanned Dilution/Ore Loss (Within 5-10%) If You Do Not Measure It, You Cannot Control It!!!!!!!!

14

You might also like