Sampling 2018 Esbensen Optimal Grade Comparison Experiment 201018.pdf

Optimal grade control sampling practice in open pit mining – A full-scale Blast Hole vs Reverse Circulation comparison variographic experiment Karin Engström Kim H. Esbensen

PH.D. Thesis

–

Defence: December 7.th 2018

“Sampling in iron ore operations Evaluation and optimisation of measurement systems to reduce sampling variability” Karin Engström

Due to unfortunate absence ... Instead of Karin:

PH.D. Thesis

–

Defence: december 7.th 2018

Supervisors

Professor Kim H. Esbensen Associate professor Sergey Kucheryavskiy Assessment Committee

Professor Reinhard Wimmer (chairman) Dr. Ralph J. Holmes Professor Pentti Minkkinen Doctoral thesis

Doctoral School of Engineering and Science: Biotechnology, Chemistry and Environmental Engineering, Aalborg University, campus Esbjerg,Denmark

PH.D. Thesis

–

Defence: december 7.th 2018

Paper I (published) Engström, K. (2017) A comprehensive literature review reflecting fifteen years of debate regarding the representativity of reverse circulation vs blast hole drill sampling. TOS Forum, 7:3646. DOI: 10.1255/tosf.99 Paper II (published) Engström, K. and Esbensen, K.H. (2017) Blasthole sampling (replicate and variographic experiments) in LKAB open pit iron ore mines: fit-for-purpose representativity? In: Dominy SC, Esbensen KH, editors. WCSB 2017. In Proceedings of the 8th World Conference on Sampling and Blending; May 9-11; Perth; AusIMM. pp 85-96. Paper III (published) Engström, K. and Esbensen, K.H. (2017) Optimal grade control sampling practice in open pit mining – A full-scale Blast Hole vs Reverse Circulation variographic experiment. Applied Earth Science - Transactions of the Institutions of Mining and Metallurgy: Section B, 126(4):176-187. DOI: 10.1080/03717453.2017.1414104. Paper IV (revised and re-submitted) Engström, K., Olausson, L. and Esbensen, K.H. (2018) Experimental evaluation of surface water sampling variability for environmental monitoring in iron ore operations using concepts from the theory of sampling (TOS). Submitted to: Mine Water and the Environment. Paper V (published) Engström, K. and Esbensen, K.H. (2017) Evaluation of sampling systems in iron ore concentrating and pelletizing processes – Quantification of Total Sampling Error (TSE) vs. process variation. Minerals Engineering, 116:203-208. DOI: 10.1016/j.mineng.2017.07.008 Paper VI (published) Engström, K., Mickelsson K-O., Töyrä, S. and Esbensen, K.H. (2018) Improvement of sampling for moisture content analysis in iron ore pellet feed – variographic characterization and online IR-analysis. In Proceedings of the 29th International Mineral Processing Congress; September 15-21; Moscow. Paper VII (published) Engström, K. and Esbensen, K.H. (2018) Variographic Assessment of Total Process Measurement System Performance for a Complete Ore-to-Shipping Value Chain. Minerals 8(7):310. DOI: 10.3390/min8070310.

Optimal grade control sampling practice in open pit mining – A full-scale Blast Hole vs Reverse Circulation comparison variographic experiment

Karin Engström Kim H. Esbensen

Narvik

Harbour Harbour

Kiruna Kiruna Under Under ground ground mining mining and and processing processing

Malmbanan Malmbanan Railway Railway Narvik--Luleå Luleå Narvik

Svappavaara

Open Open pit pit mining mining and and processing processing

Malmberget Under ground mining and processing

Luleå Harbour and Head Office

IRON ORE OPERATIONS

LKAB Production sites Mines

Sorting

Concentrating

Pelletising

Products

NARVIK

Kiruna Kiruna Pellets

40 km Gruvberget Leveäniemi

Malmberget Malmberget

Harbours

Svappavaara

Svappavaara

Pellets

Fines

LULEÅ

12

Open pit mine grade control • Reliable grade control is important to increase productivity and optimise resource utilisation • No universally agreed best practise for open pit mine drill sampling • Common methods for short term planning are manual Blast Hole sampling and Reverse Circulation drill sampling • Both methods have been shown to be associated with several sampling problems • Both methods have also been shown to be fit-for-purpose representative if applied with appropriate care and regard for local ore characteristics and mining conditions 13

“A COMPREHENSIVE LITTERATURE REVIEW REFLECTING 15 YEARS OF DEBATE REGARDING THE REPRESENTATIVITY OF REVERSE CIRCULATION VS. BLAST HOLE DRILL SAMPLING” Karin Engström1 and Kim H. Esbensen2

1. Quality development engineer: sampling and test methods, Luossavaara-Kiirunavaara AB (LKAB), Kiruna, Sweden. Industrial PhD student, Aalborg University, Denmark. Email: [email protected] 2. Sampling consultant. Adjunct professor, Aalborg University, Denmark. Adjunct professor, Geological Survey of Denmark and Greenland (GEUS), Copenhagen. Assoc. Professor, Université du Québec à Chicoutimi, Quebec. [email protected]

Blast hole samples are often regarded as inferior with comparison to ‘proper drill sampling’ like Reverse Circulation and Diamond (core) drilling and is accused of lacking representativity by the sampling community. This literature review covered a total of 31 publications, from which the conclusion is: it is not possible to draw a general overall conclusion as to the superiority of one drill sampling method over another! Both RC and BH have advantages and disadvantages and the choice of system needs to be related to the ore type and to the mining conditions. Thus it is always necessary to evaluate the specific sampling system to be used in the light of TOS and with respect to the characteristics of the ore to be mined. It would appear that the mining industry is doomed to continue to follow local, often economy-driven objectives and sampling solutions, even if these can be documented as inferior when seen in the light of the representativity imperative. A call is made for universal adherence to the principles laid down by TOS for representativity in the primary sampling stage, before economic, logistical or other (local) factors are allowed to intervene. What is the objective to analyse and to make decisions in the mining industry, based on samples that can be documented not to be representative?

Published in TOS Forum, Issue 7 (2017)

Current open pit mine grade control at LKAB • Mine planning is based on a resource block model derived from Diamond Drilling • For more detailed short term mine planning, manual Blast Hole drill sampling assay data are added to the block model • The manual BH sampling method has previously been compared to segment sampling and full cone BH sampling

16

Current practice for manual BH sampling in which samples are extracted from the centre of blast hole cones. A correct sectorial cut is opened first, after which an equal amount of material from each cone level is extracted using a small shovel (this is incorrect acc. TOS) Note production drilling grid markers in the background. 17

19

20

Sample preparation: 60 Blast Hole samples (BH)

Primary Sampling

1. 1.st weighing 2. Splitting 3. Drying 4. 2.nd weighing

Manual BH sample (current protocol)

Manual BH

Segment BH sample

Segment BH

Complete coarse BH cone

Complete cone

Complete fines BH cone

Complete fines

Blast Hole #1-60

21

22

A Crushing D C B

Milling

A: Segment sample from radial bucket (only from coarse BH cone) B: Manual BH sample using a shovel (only from coarse BH cone) C: Complete coarse BH cone D: Complete fines cone 23

Study objectives • Result from a previous variographic experiment at LKAB indicated that the manual sampling method was concluded to be fit-for-purpose for the experimental transect studied. • The objective of the present study is to evaluate the alternative of using RC drill sampling for possible improvement of sampling representativity. • Comparative variographic analysis between RC and BH sampling results. • Experimental data will assessed with univariate statistics and with multivariate data analysis approaches (Principal Component Analysis, PCA), see Ph.D.. 24

25

Leveäniemi open pit mine • All experiments were performed in Leveäniemi open pit mine, located in Svappavaara in northern Sweden. • The Leveäniemi ore deposit of Kiruna type is hosted by Metamorphosed rocks of volcanic and sedimentary origin • The main ore type mined in Leveäniemi open pit is massive magnetite as well as magnetite veins/breccias in the trachy-andesites and biotite schist • The production planning is based on a 3D block model, in which the chemical assays from exploration drilling and manual BH sampling are combined.

Leveäniemi open pit mine

26

27

28

29

”Normal Operating Conditions (NOC)”

30

Open pit mine sampling at LKAB • Blast holes in Leveäniemi open pit are drilled using an Atlas Copco D65 BH drill rig • Dust collector separate coarse and fine material in two BH cones • Every other BH is sampled manually – NB sample extraction is applied to the coarse BH cone only • The manual sampling procedures cuts away a sectorial part of the BH cone. • The sample is collected as a vertical slice of uniform thickness in the exposed axial centre of the cone • Deliberate exclusion of the smaller fines BH cone is a clear violation of the Fundamental Sampling Principle (FSP) 31

BH vs RC drill sampling experiment • 53 Blast Holes were drilled and samples were extracted from the coarse BH cone • Samples were extracted using the current manual sampling method • Samples were also extracted with an alternative full segment approach • From 18 of the total 53 blast holes, the complete BH material was collected, weighed and split representatively in the field allowing an authoritative reference result

• RC drill holes were deployed along a parallel profile to the Blast Holes • Average distance of 1.5 m to the earlier drilled BH profile • Spacing in-between RC drill holes were 2.5 meters along the strike

• The set of parallel BH and RC sampling campaigns enable three variographic characterisations, one for each sampling method. 32

33

The complete BH cone material was processed ...

Complete BH cone material – unbiased reference 34

35

36

Chemical results from parallel drill experiments

37

Chemical results from parallel drill experiments

Correlations between parallel BH and RC samples for Fe and Si grade 38

Bias test results

39

Variographic characterisation

40

Decomposing the relative variogram

Observed total variability

Total Measurement system error

41

Variographic characterisation comparison

42

Variographic characterisation comparison

43

Variographic characterisation - comparison

44

Conclusions - for local mining conditions • RC sampling is associated with lower sampling variability compared to the current manual BH sampling method • The possibility of dividing the 15-20m benches into smaller intervals/ strata could be valuable in some areas of the open pit mine (very variable mineralisation/geology) • However the mining value gained from this does not appear to be able to offset the very significant increase in drilling costs and added resources for sample handling, preparation and analysis! (This is the considered LKAB conclusion) • The manual method can still be characterised, and used, as fit-for-purpose for short term mine planning at the Leveäniemi open pit mine • RC sampling is not appreciably superior to manual BH sampling, to motivate the sig. increased cost, in the Leveäniemi open pit mine! (The considered LKAB conclusion) 45

General conclusions (to the degree possible …) • There is no generally optimal drill sampling method that can be established as a universal best practice. • The drilling and sampling method needs to be carefully evaluated and validated with respect to each specific ore heterogeneity and mining situation. • A general recommendation would be to use pilot variographic sampling experiments, along a carefully selected profile with relevant ore heterogeneity • Allows mining geologists and production engineers to evaluate the specific total measurement system performance in direct relation to the full ore variability

46

Thank you for your attention

48

PCA results

49

PCA results

50

PCA results

51