estimacion de finos generados por voladura

This paper introduces an engineering approach toestimate the proportion of fines generated during theblasting process. The proposed framework is based onthe combination of two Rosin-Rammler baseddistribution functions to model the full range offragments expected to be produced during this process.This particular approach, which has been successfullyapplied for a number of years by the Julius KruttschnittMineral Research Centre (JKMRC), has beenimproved with the introduction of a new model to predictthe potential volume of crushed material resulting fromthe crushing and shearing stages of blasting. Othersources of fines including liberation of infilling fromdiscontinuities, particle collisions and post-blastprocesses have been excluded to simplify the modellingprocess. Validation analysis of the proposed frameworkhas shown that there is good agreement between modelpredictions and the measured distribution of fines. Inthree distinct cases, results verified the hypothesis that asingle index of uniformity can be used to describe thedistribution of fragments in the range of 1 mm throughto the expected post-blast mean fragment size (x50).Although some limitations have been noted, theapproach appears to provide useful approximations for

continuous improvement analysis and applications. Thepractical application of the proposed modelling frameworkis demonstrated with an engineering study aimed atassessing the impact of blast fragmentation on the overallproduction of fines in a hard rock quarry. Results fromsimulations showed that less crushing requirements due toan overall increase in fragmentation contribute to adecrease in the specific crushing energy and hence areduction in power consumption requirements. Thisanalysis helped demonstrate the importance of addressingthe impact of blast fragmentation distribution on overallquarry productivity requirements; and highlights theimportance of adopting a holistic approach whenaddressing the blast optimisation problem.I. Onederra (for correspondence I.Onederra@ uq.edu.au), S.Esen and A. Jankovic are at the Julius Kruttschnitt MineralResearch Centre, The University of Queensland, St Lucia,Brisbane, Queensland, Australia.

2004 Institute of Materials, Minerals and Mining andAustralasian Institute of Mining and Metallurgy. Published byManey on behalf of the Institutes. Manuscript received 18August 2004; accepted in final form 20 October 2004.

Keywords: Fines, blasting, volume, particle size, Rosin-Rammler model

Mining Technology (Trans. Inst. Min. Metall. A) December 2004 Vol. 113 A237DOI 10.1179/037178404225006191

Estimation of fines generated by blasting applications for themining and quarrying industries

I. Onederra, S. Esen and A. Jankovic

INTRODUCTIONBlasting activities in mines and quarries have beenplacing significant emphasis on the ability to tailorfragmentation to improve downstream processes. Inmany of these operations, the impact of fines has beenclearly identified. For example, the generation ofexcessive fines in operations adopting in situ leaching asthe main ore processing method, may hinder recovery ascertain fines tend to affect the permeability of leachingpads. Leaching performance may be affected if theproportion of material that is less than 150 m exceeds12% in the feed to the agglomerators.24 Similarly, theefficiency of coal processing is strongly related to thegeneration of fines of less than 05 mm. Increased finescontent in run-of-mine feed leads to higher handling andprocessing costs, low yields, increased product moisturecontent, and in many cases a reduced product value.10

The same can be said of quarry operations wherematerial of less than 2415 mm in size may beconsidered to be of no value and hence wasted.

Whilst fines may be detrimental to some operations,in large-scale metalliferous mining there is evidence to

suggest that by providing an appropriate sizedistribution to crushing and grinding circuits, ameasurable increased throughput and/or reduced powerdraw can be obtained.12 This may entail a requirement toincrease the proportion of finer material in productionblasting activities.

The need to be able to predict the amount of fines fromblasting has driven the development of an improvedengineering model. The model being proposed is based onfindings from both model-scale and full-scale data.Model-scale data were obtained from a comprehensiveexperimental programme,3 whilst full-scale studies havebeen compiled from surveys conducted over a number ofyears by researchers of the Julius Kruttschnitt MineralResearch Centre (JKMRC).

FRAGMENTATION MODELLINGFRAMEWORKFor a number of years, the JKMRC has been applyingtwo empirical models to estimate muckpile frag-mentation distributions in surface blasting operations.

These empirical models are the two-component model(TCM)8 and the crushed zone model (CZM).15 Thesemodels are a hybrid between the well-known Kuz-Ramapproach6 and two specific fines predictive models. Acomparison of these two models conducted by Hall andBrunton14 has highlighted some deficiencies in theirpredictive capabilities, due principally to some limitationsin their estimations of fines and intermediate sizefractions. Their conclusions indicated that there was aneed to review and further improve ways of predicting thedistribution of fines and, in particular, estimations ofcrushing in the vicinity of detonated blastholes. This studyhas been part of that process.

Building on the hybrid approach incorporated in theCZM model, the expected distribution of fragments inthe fines and coarse regions is modelled by two separatefunctions. These two functions are based on the well-established Rosin-Rammler distribution23 and given by:

R x e1 . xx

0 693

fn

50= - -_ ai k Eq. (1)

for values of x less than or equal to x50

R x e1 . xx

0 693

cn

50= - -_ ai k Eq. (2)

for values of x greater than x50

where R(x) is the proportion of the material passing ascreen of size x, x50 is the post-blast mean fragmentsize, nc is the uniformity index for the coarse end of thedistribution and nf is the fines uniformity index whichis given by:

.n

LN x

LN x

1

0 693f

n

50

f

=-

cc

mm

Eq. (3)

where fc is the proportion of the material passing ascreen of size 1 mm or the fines inflection point.

The modelling framework is graphically illustrated inFigure 1 by comparing the standard Rosin-Rammler or

Kuz-Ram based distribution with the proposed combineddistribution functions. This figure also highlights the mainmodelling parameters, namely the fines inflection point(fc), the expected mean fragment size (x50) and the coarseuniformity index (nc), also referred plainly in literature asthe uniformity index (n). In this paper, a thoroughdescription of a revised approach to determine the finesinflection point fc is given. The determination of the othertwo key modelling parameters (i.e. the mean fragmentsize, x50 and the coarse uniformity index, nc) follows thewell-documented Kuz-Ram approach2,5,6,8,15 and is,therefore, not covered here.

The fines inflection point (fc)Literature indicates that fines present in a muckpiletend to originate from the near field crushing zone,fracturing (shearing) zones as well as possibleliberation from rock mass discontinuities.9,11,25 Thefines inflection point is introduced to consider thesesources and is given by:

%f FinesV

V V F100( )c mm c b r11

#= =+

+- c m 7 A Eq. (4)where Vc is the volume contribution of the crushedzone, Vb is the volume contribution from breakage(major radial cracks), Vt is the total volume beingblasted and Fr is a rock mass fines correction factor.

The fines inflection point is based on thehypothesis that, for most conditions, the coarsestparticle size expected to be generated during thecrushing and shearing stages of blasting would be 1mm, and that the percentage passing fraction wouldbe directly proportional to the volume of crushedand/or sheared rock material surrounding a detonatedblasthole.

As depicted in Figure 2, the estimation of thevolume of crushed and/or sheared rock materialfollows simple geometric calculations given by (i) theradius of crushing and thus the volume of a cylinderof crushed rock; and (ii) the distribution of major

A238 Mining Technology (Trans. Inst. Min. Metall. A) December 2004 Vol. 113

Onederra et al. Estimation of fines generated by blasting

1 Key parameters of the proposed fragmentation modelling framework

radial cracks, which are assumed to be evenlydistributed around a borehole, planar and alsocontinuous along the length of the explosive charge.These two components define the total volume of astar-shaped crushed region (i.e. Vc + Vb).

In the proposed modelling framework, a rock masscorrection factor (Fr) has been introduced to addressthe hypothesis that fines may also be liberated fromrock mass discontinuities.9 However, an approach todetermine this parameter has not been developed, asthere is insufficient quantifiable evidence to supportthis. To simplify the modelling structure, the Frparameter is, therefore, currently disregarded and themodelling process involves only the determination ofVc and Vb.

The crushed zone model to determine VcThe determination of Vc is based on an improvedmodel to predict the radius of crushing generated by adetonated blasthole reported by Esen et al.11 Thismodel is given by the empirical relationship:

.r r CZI0 812.

c o

0 219

= _ i Eq. (5)where rc is the crushing zone radius (mm), ro is theborehole radius (mm) and CZI is defined as thecrushing zone index. This is a dimensionless indexthat identifies the crushing potential of a chargedblasthole and is calculated from:

CZIK

P

c

b

2

3

#=

v__ii

Eq. (6)

where Pb is the borehole pressure (Pa), computed fromnon-ideal detonation theory, K is the rock stiffness(Pa) and c is the uni-axial compressive strength (Pa).Rock stiffness K is defined assuming that the materialwithin the crushing zone is homogeneous andisotropic and is given by:

K E1 d

d=+ y Eq. (7)

where Ed is the dynamic Youngs modulus (Pa) and dis the dynamic Poissons ratio.

As discussed by Esen et al.,11 this model has beenshown to have better predictive capabilities than otherdocumented approaches. Its validity has also beenconfirmed with data obtained from full scale blastingconditions.

Given this improved modelling approach, the relativerole of the crushed zone in the overall contribution of fineswas investigated with the back analysis of several casestudies. Table 1 gives a summary of the physical andmechanical properties of the rock types at each site, andTable 2 presents a summary of the explosive and blastdesign parameters adopted in each case.

Results from this preliminary back analysis aregiven in Table 3. The analysis shows that following theISRM rock classification system, categorised byYoungs modulus (), in the soft(low strength) rock types, the proportion of fines dueto the crushed zone relative to the total amount offines, are in the range of 91196% (Cases 1, 3 and 5).In the medium-to-hard (high strength) rocks, therange is 8592% (Cases 2, 4 and 6). This analysis

Mining Technology (Trans. Inst. Min. Metall. A) December 2004 Vol. 113 A239


Table 1 Physical and mechanical properties of rock types

Case study Rock type sc (MPa) T (MPa) (kg m3) Ed (GPa)

1. Coal blasting26 Coal 20 20 1440 962. Mount Cootha18 Hornfeld 200 16 2730 8303. Ok Tedi Mine1 Monzodiorite 55 78 2600 3404. Cadia Hill Gold15 Monzonite 127 90 2600 7705. Escondida Mine7 Porphyry ore 22 30 2616 1526. Porgera Gold Mine4 Hornblende Diorite 138 130 2725 700

T, tensile strength; , density.

2 Volume of crushed material around a blasthole

clearly supports the hypothesis that in full scaleblasting operations, the crushed zone around ablasthole is not the only significant source of fines,and that in most cases, the contribution of breakagecan be expected to be significant.

As the proposed approach seeks to estimate theproportion of fines only present in the muckpile, post-blast sources such as excavation, handling, mechanicalsieving and crushing are not relevant to the modellingframework being proposed. However, fines generatedby the breakage process itself must be considered. Inorder to address the issue of incorporating thecontribution of overall breakage on the proportion offines (i.e. Vb in the proposed framework), a simplisticcrack model has been adopted. This model isproposed as a preliminary engineering tool to estimatethe parameter Vb as discussed below.

The crack model to determine VbThe following approach assumes that the source offines from overall breakage is directly proportional toa volume of crushed material bounded by major blastinduced fractures. The number of near field radialcracks (C) around the blasthole is estimated followingthe approach proposed by Katsabanis:16

CTP

sd

b= f c m Eq. (8)where s is the strain at the blasthole and Td is thedynamic tensile strength of the rock (Pa), which isassumed to be in the range of 48 times the static value.

The strain at the blasthole wall, s, can beapproximated by Katsabanis:16

P

P

2 1 2 3 1

1s

p b

b

2= - + -

-f

y to y c

y

_ __i i

iEq. (9)

where Pb is the explosion or borehole pressure (Pa); is the rock density (kg m3); vp is the P-wave velocity(m s1); is the adiabatic exponent of the detonationproducts; and is the Poissons ratio of the rock.

The length or radial extension of cracks isdetermined empirically with the stress attenuationfunction proposed by Liu and Katsabanis,19 assumingthat the crack will arrest when the induced stress isequal to the static tensile strength or the rock material.In this case, the following relationship is proposed:

C rPT ro

eq

sc1

1

= -zc m Eq. (10)

where Ts is the static tensile strength of the rock (Pa),ro is the blasthole radius (m), rc is the radius ofcrushing (m), is the pressure decay factor and Peq isthe equilibrium pressure (Pa), or the pressureexperienced at the end of the crushing zone which isgiven by:

P P rr

eq bo

c=

zc m Eq. (11)The pressure decay factor is a function of rock

and explosive properties. It is a negative number thathas been found to be in the range of 124 to 165 fora wide range of explosive and rock combinations.19 A



Table 3 Role of the crushed zone in the generation of fines

Case study rc Vc Blast % 1 mm % 1 mm Relative proportion(mm) (m3) volume (crushed (measured Measurement of fines from the

(m3) zone) total fines) technique crushed zone (%)

1. Coal blasting26 746 934 4655 200 170 Mobile screening after excavation 118 (Vb, Fr, E&H, S)and handling

2. Mount Cootha18 122 045 19723 021 23 Primary crusher product 91 (Vb, Fr, E&H, C)3. Ok Tedi Mine1 573 983 9446 100 110 Split* with fines correction from 91 (Vb, Fr, E&H, C)

crusher product4. Cadia Hill Gold15 301 292 6318 046 50 Split with fines correction from 92 (Vb, Fr, E&H, C)

crusher product5. Escondida Mine7 1085 273 9446 290 180 Split with fines correction from 161 (Vb, Fr, E&H, C)

crusher product6. Porgera Gold Mine4 229 071 4066 017 20 Split with fines correction from 85 (Vb, Fr, E&H, C)

crusher product

* Split refers to the Split Engineering image analysis technique.17

Vb, contribution from the fracturing or breakage process; Fr, contribution of fines liberated from rock mass discontinuities; E&H, excavation and handling; S, mechanical sieving; C, crushing.

Table 2 Explosive and design parameters

Hole Detonation Explosive Charge Burden diameter velocity density length spacing

Case study Explosive (mm) (m s1) (kg m3) (m) bench height (m)

1. Coal blasting26 Emulsion 150 5340 1180 5.4 7 7 952. Mount Cootha18 Emulsion 102 5447 1200 115 35 4 143. Ok Tedi Mine1 Emulsion 251 5000 1150 100 74 85 154. Cadia Hill Gold15 Emulsion 229 4950 1200 120 6 7 155. Escondida Mine7 Emulsion 270 5020 1050 75 74 85 156. Porgera Gold Mine4 Emulsion 200 4500 1250 53 55 62 12

first approximation can be obtained with thefollowing empirical relationship:

. .EVOD

0 0085 0 9955

.

dp

0 33

=- +zo

-

_ di n Eq. (12)where Ed is the dynamic Youngs modulus (GPa), vp isthe p-wave velocity (m s1) and VOD is the confinedvelocity of detonation of the explosive charge.

In order to make preliminary verifications of theproposed crack model, all of the case studiesdescribed in Table 3 have been re-analysed and newcrushed volume predictions have been made andsummarised in Table 4.

As shown in Table 4, there is now better agreementbetween the measured and the modelled proportion offines at the assumed cut-off point of 1 mm. Discrepanciesare within the expected errors associated with samplingand modelling assumptions. In general, in all rock types,the predicted fines inflection point is lower when comparedto the measured total fines. This is more pronounced in theweaker rock types where further degradation can beexpected from handling, transportation and crushing, ashas been previously demonstrated by Djordjevic et al.10

VALIDATION OF THE PROPOSED FINESMODELLING FRAMEWORKThe approach to predict the fines inflection point (fc)and the derivation of a single index of uniformity (nf)to describe the distribution of fines between this pointand the mean fragment size is validated in this section.A comparison between model predictions andmeasurements conducted in three full-scale blastsoutside the original database has been carried out.

The adopted criteria for validation involved a directcomparison between the measured and predicted 10%and 20% passing fractions (i.e. P10 and P20). Theadoption of P10 and P20 values as a comparisonbenchmark is justified by their common use as inputparameters in comminution simulation tools for thedesign and optimisation of crushing and grindingcircuits. Table 5 gives a summary of the inputparameters available and used in the analysis.

In case study 1, fragmentation assessment and blastmonitoring work have been reported by Hall13 andinvolved the application of image analysis techniques formaterial greater than 258 mm. Below this size, thematerial was initially screened in the field, followed by sub-sampling and laboratory screening down to 036 mm.

In case studies 2 and 3, reported by Onederra andCorder,22 two important blasting/ore domains wereidentified and surveyed in benches 2755S1 and2875N11 of the pit. In both of these domains, theproportion of fragments less than 1 mm in size (i.e.fines inflection point) was measured from laboratorysieving of the primary crusher product or what isreferred to as belt cut sampling.

For the conditions described above and followingthe procedures outlined earlier, the zones contributingto crushing given by the Vc and Vb parameters werecalculated and the fines inflection point (fc)determined in each case. Table 6 summarises themodel results together with the measured values.

As shown in Table 6, in all three cases the predictedfines inflection point values are reported below themeasured values. This is considered to be a logicaloutcome, as the proposed fines modelling frameworkdoes not consider the contribution of fines given by



Table 4 Comparison between new modelled results and measured values

Case study & rock type rc Number Volume crushed Blast % 1 mm % 1 mm (mm) of cracks (C) Vc + Vb (m

3) volume (m3) (model) (measured)

1. Coal 746 43 381 4655 82 172. Hornfeld 122 2 281 19723 143 233. Monzodiorite 573 6 605 9446 64 114. Monzonite 301 3 293 6318 46 55. Porphyry ore 1085 23 107 9446 113 186. Hornblende Diorite 229 2 485 4066 12 2

Table 5 Descriptive parameters of case studies

Case study

1 2 3

Quarry blast13 Open pit Blast 2755S122 Open pit Blast 2875N1122

Explosive Emulsion Heavy ANFO Heavy ANFOHole diameter (mm) 109 270 270VOD (m s1) 5345 5100 5100Explosive density (kg m3) 1150 1250 1250Charge length (m) 113 72 72Burden spacing bench height (m) 36 42 135 65 65 16 70 70 16Rock type Foliated phyllite Porphyry ore Porphyry orec (MPa) 71 (parallel to foliations) 58 61T (MPa) 110 60 60 (kg m3) 2700 2500 2500Ed (GPa) 300 370 370

particle collisions and degradation of the rock materialfrom loading, handling, transportation, and dumping.

In cases 2 and 3, differences between predicted andmeasured fc values of 34% and 51% may be perceived asgross underestimations of the proposed model. However,the friable nature of the complex porphyry ore wasexpected to be subject to further degradation fromloading, handling and dumping prior to crushing. Thissusceptibility to degradation was confirmed by themeasured differences between the primary crusherproduct and SAG mill feed fragmentation. This differencewas of the order of 58% for the 1 mm size fraction in theN11 and S1 domains, respectively.22

Figure 3 shows a comparison between the predictedand measured fines, highlighting the 10% and 20% passingfractions. A parity chart of measured versus modelledcumulative percentage passing for fragments in the rangeof 163 mm is also included. It should be noted that, in allcases, the fines uniformity index (nf) and the subsequentmodelled distribution of fines was determined with themeasured post-blast mean fragment size (x50). This wasconsidered adequate in order to assess the validity of thehypothesis proposing that a single index of uniformity canbe used to describe adequately the distribution offragments from 1 mm through to the expected post-blastmean fragment size (x50).



Table 6 Comparison between modelled and measured fines inflection point values for three full scale blasts

Volume Fines inflection pointNumber of crushed Blast % 1 mm % 1 mm

Case study rc (mm) cracks (C) Vc + Vb (m3) volume (m3) (model) (measured) Difference

1. Quarry blast19 240 5 836 20295 41 49 082. Blast 2755S113 606 8 542 676 80 114* 343. Blast 2875N1113 592 8 602 833 72 123* 51

*Primary crusher product (belt cut sampling).

3 Comparison between modelled and measured fines in three full scale blasts

Case study P10 Difference % P20 Difference % Model Model(mm) (mm) Error (mm) (mm) Error Measured Measured

1. Quarry blast19 45 4 05 125 155 13* 25 1922. Blast 2755S113 17 N/A N/A N/A 85 98+ 13 1333. Blast 2875N1113 20 N/A N/A N/A 95 11+ 15 136

*Laboratory sieving; +calibrated SPLIT image analysis system.

As shown in Figure 3, the proposed fragment sizedistribution function appears to be an adequatedescriptor of the expected distribution of fines generatedin full-scale blasting conditions. The parity chart showsthat although the model may in some casesunderestimate and overestimate the proportion of finesgenerated by blasting, overall trends are being capturedby this single uniformity index. In terms of the predictiveerrors associated with specific size fractions such as P10and P20, for case 1, the error is of the order of 125%and 192%, respectively, whilst for cases 2 and 3, theaverage predictive error for P20 is about 135%. This isconsidered adequate for engineering design purposesgiven the expected variability of rock material andexplosive performance within the blasted volume.

Whilst the above cases support the hypothesis that infull-scale blasting conditions the use of a single index ofuniformity is appropriate to describe the distribution offragments from the fines inflection point (fc) to the meanfragment size (x50). Recent large-scale field trialsconducted in more massive and competent rock masses20

have shown that an approach based on a singleuniformity value may produce an overestimation of theproportion of fines, particularly in the range of 10100mm. Results from sieved muckpiles under theseenvironments have shown that in the fines region therecould be at least two marked changes in uniformity whichdescribe the natural breakage characteristics of the rockmaterial. A preliminary comparison of these data withthe proposed fines distribution function has confirmedthis. However, it is important to highlight that for therepeatable documented cases, the proposed star-shapedcrushed model was able to estimate the fines inflection

point (fc) adequately. Further work is being conducted toaddress the above limitation and refine the overallpredictive capabilities of the model.

APPLICATION OF THE PROPOSED FRAG-MENTATION MODELLING FRAMEWORKTo demonstrate the practical application of the proposedfragmentation modelling framework, fragmentationdistributions for two blasts have been modelled and usedas input to processing simulations. This cross-disciplinarymodelling work was aimed at assessing the impact of blastfragmentation on the overall production of fines in a hardrock quarry. The analysis focused on quantifying therelative contribution of blasting to the total production offines or waste material and it is based on conditions foundat the Mount Cootha Quarry summarised in Table 1 anddocumented by Kojovic et al.18

Quarry processing simulations were conductedfollowing two specific requirements. The first related tothe production of fine aggregates (product range, 18 mmto 24 mm), with waste considered to be any material lessthan 24 mm. This required the implementation of threestages of crushing similar to that illustrated in Figure 4.The second simulation considered the production ofcoarser aggregates (product range 32 mm to 10 mm),with waste considered to be any material less than 10mm. This required the implementation of only twostages of crushing.

Table 7 describes the blast design parametersadopted in this modelling exercise. Geotechnicalproperties measured at the Mount Cootha Quarry aredescribed in Table 1. Results of the expected muckpile



Table 7 Blast design parameters for simulated conditions

Blast 1 Blast 2

Explosive Emulsion EmulsionHole diameter (mm) 109 109Detonation velocity (m s1) 4968 4968Explosive density (kg m3) 1180 1180Charge length (m) 113 113Burden spacing bench height (m) 36 42 135 30 35 135Powder factor (kg m3) 061 088

4 Example of a three stage crushing circuit at the Mount Cootha quarry (after Kojovic et al.18)

fragmentation distributions given by the proposedfragmentation modelling framework are shown inFigure 5.

As discussed earlier, the results of the blastfragmentation model shown in Figure 5 have beenused as input into two quarry processing simulationoptions using the JKSimMet program. This programhas been developed from extensive JKMRCexperience in the field of comminution and has nowbecome an industry standard tool for design andoptimisation of mineral processing activities.21 Asummary of the simulation results for both blastingscenarios and the two processing options are shown inTable 8 and Figures 6 and 7.

The application of the proposed fines modellingframework in conjunction with quarry processingsimulation models have shown that any increase in theproportion of fines generated during the blastingprocess does not translate directly into an equivalentincrease in the amount of fines or waste productdownstream (i.e. after crushing). This is because asignificant proportion of fines may be generatedduring the crushing stages of the production of therequired products.

As shown in Figure 6, the impact of crushing onfines (waste) generation is more pronounced in thefiner quarry processing circuit, where three stages ofcrushing are required. The amount of fines generatedin crushing is about 1626 times of that produced byblasting for blasts 2 and 1, respectively. In the coarserprocessing circuit (Fig. 7), fines generated fromblasting have a more significant contribution to thetotal production of waste material. As shown, theamount of fines produced in crushing are about0813 times of that produced by blasting for blasts 2and 1, respectively.

The analysis shows that in both cases, less crushingrequirements due to an overall increase infragmentation contribute to a decrease in the specific

crushing energy and hence a reduction in powerconsumption requirements. This suggests that, in thecase where finer aggregate products are required, theenvironmental and cost benefits of decreased powerconsumption must be weighed against the penalty ofincreasing the amount of fines generated afterprocessing. As shown in Figure 6, a 27% increase inthe amount of fines due to the higher intensity blast,translates only to a 1% increase in the total amount ofwaste, but at the same time a 4% reduction in thepower consumption requirements. For the coarse



Table 8 Processing simulation results

Processing method Fine processing Coarse processingWaste: less Waste: less

than 24 mm than 10 mm

Blasting scenario Blast 1 Blast 2 Blast 1 Blast 2ROM P80 (mm) 1925 1456 1925 1456ROM 24 mm (%) 66 93ROM 10 mm (%) 137 181 137 181PCP P80 (mm) 1355 1231 1076 1024PCP 10 mm (%) 149 186 163 197PCP 24 mm (%) 71 95Primary crusher (kW) 172 15 214 18SCP P80 (mm) 338 332 352 334SCP 10 mm (%) 307 332 159 158SCP 24 mm (%) 133 151Secondary crusher (kW) 538 497 864 834Tertiary crushers (kW) 1923 1887

Final products18+10 mm (%) 331 324112+68 mm (%) 165 16268+24 mm (%) 266 26624 mm (%) 238 24832+18 mm (%) 448 43218+10 mm (%) 242 23710 mm (%) 31 331

Total (kWh t1) 263 253 108 101

ROM, run-of-mine; PCP, primary crusher productSCP, secondary crusher product.

5 Modelled fragmentation distributions for blast 1 and blast 2

aggregate production option (Fig. 7), higher intensityblasting results in a 21% increase in waste product buta 6% reduction in power consumption. Because of thefiner fragmentation in blast 2, there would be anincrease in loading and hauling productivity whichhas not been quantified in this paper. It is importantto add that this will have an impact on the total cost ofproduction as documented by Kojovic et al.18

CONCLUSIONSAn engineering approach to predict the proportion offines generated during blasting has been presented.The improved hybrid approach introduces a newmodel to predict the potential volume of crushedmaterial resulting from the crushing and shearing

stages of blasting. Other sources of fines includingliberation of infilling from discontinuities, particlecollisions and post-blast processes have been excludedto simplify the modelling process.

Based on the back analysis of a number of full-scale blasting surveys, this study has confirmed thatupon detonation of an explosive, the region ofcrushing around a blasthole is not the only source offines. This has justified the inclusion of a factor thatconsiders the contribution to fines by the overallfracturing process.

The overall modelling framework has beenvalidated with three distinct case studies. Results fromthis analysis have shown that there is good agreementbetween model predictions and the measureddistribution of fines, verifying the hypothesis that a



6 Simulation results describing the proportion of fines (waste product) generated by blasting and processing in thethree stage crushing circuit

7 Simulation results describing the proportion of fines (waste product) generated by blasting and processing in thetwo stage crushing circuit

single index of uniformity can be used to describe thedistribution of fragments in the range of 1 mmthrough to the expected post-blast mean fragment size(x50). Some limitations have been noted, particularlywith regards to the possible overestimation of fines inmore massive and competent rock types. However, theapproach appears to provide useful approximationsfor continuous improvement analysis and engineeringapplications. Further work is continuing to identifyother possible limitations and thus improve theframeworks predictive capability.The practical application of the current frameworkhas been demonstrated with a cross-disciplinarymodelling study. The study was aimed at assessing theimpact of blast fragmentation on the overallproduction of fines in a hard rock quarry. Results haveindicated that any increase in the proportion of finesgenerated during the blasting process does nottranslate directly into an equivalent increase in theamount of fines or waste product downstream (i.e.after crushing). This is because a significantproportion of fines may be generated during thecrushing stages of the production of the requiredproducts. The impact of blasting will depend on thefinal product requirements. Further analysis alsoshowed that less crushing requirements due to anoverall increase in fragmentation contribute to adecrease in the specific crushing energy and hence areduction in power consumption requirements.

The application case study has helped demonstratethe importance of addressing the impact of blastfragmentation distribution on overall quarryproductivity requirements. This highlights theimportance of adopting a holistic approach whenaddressing the blast optimisation problem and the keyrole that engineering modelling tools, such as thoseproposed in this paper, can play in this process.

ACKNOWLEDGEMENTSThe authors would like to thank Prof. E. T. Brown,Mr Ian Brunton and Dr G. Chitombo for theircomments and suggestions. The authors also wish tothank Assoc. Prof. Dr H. Aydin Bilgin of the MiddleEast Technical University, Ankara, Turkey for hiscollaborations with the JKMRC.

REFERENCES1. C. BAILEY, D. THORNTON and M. DUNGLISON: Mine

to mill surveys of OK Tedi Mining Limited, InternalJKTech report, Job No 99244, 2000.

2. M. BAILING, Z. SHIQI, Z. DINGXIANG and G. CHUJI: Astudy of bench blasting in rhyoporphyry at a reduced scaleand the statistical analysis of the regularity forfragmentation distribution, Proceedings of the FirstInternational Symposium on Rock Fragmentation byBlasting, Lulea, Sweden, 1983, 857872.

3. H. A. BILGIN, S. ESEN and M. KILIC: Patarge Project,Internal Report (in Turkish), Barutsan A.S., Elmadag,Ankara, Turkey, 1999.

4. I. BRUNTON, A. JANKOVIC, S. KANCHIBOTLA, T.LICINA, D. THORNTON and W. VALERY: Mine to mill

optimisation at Porgera gold mine, JKMRC internalreport, 2001.

5. S. H. CHUNG and P. D. KATSABANIS: An integratedapproach for estimation of fragmentation, Proceedings ofthe 27th annual conference on explosives and blastingtechnique (ISEE), Vol. 1, 2001, 247256.

6. C. V. B. CUNNINGHAM: The Kuz-Ram model forprediction of fragmentation from blasting, Proceedings ofthe first international symposium on rock fragmentationby blasting, Lulea, Sweden, 1983, 439453.

7. D. DAVID, I. BRUNTON and D. THORNTON: Mine tomill surveys of Minera Escondida Ltd, Internal JKTechreport, Job No 99251, 2001.

8. N. DJORDJEVIC: Two-component of blastfragmentation. Fragblast 1999, South African Institute ofMining and Metallurgy, Johannesburg, South Africa,1999, 213219.

9. N. DJORDJEVIC: Origin of blast induced fines. Min.Technol. (Trans. Inst. Min. Metall. A), 2002, 112,A143A146.

10. N. DJORDJEVIC, J. ESTERLE, D. THORNTON and D. LAROSA: A new approach for prediction of blast inducedcoal fragmentation, Mine to Mill 1998 Conference, TheAustralasian Institute of Mining and Metallurgy,Brisbane, Australia, 1998, 175181.

11. S. ESEN, I. ONEDERRA and H. BILGIN: Modelling thesize of the crushing zone around a blasthole, Int. J. RockMech. Min. Sci., 2003, 40, 485495.

12. C. GRUNDSTROM, S. S. KANCHIBOTLA, A. JANKOVICHand D. THORNTON: Blast fragmentation for maximisingthe sag mill throughput at Porgera Gold Mine,Proceedings of the 27th Annual Conference on Explosivesand Blasting Technique, ISEE, Orlando, Florida, USA,2001, 383399.

13. J. HALL: A critical comparison of blast fragmentationmodels, unpublished Masters of Engineering Sciencethesis, University of Queensland, 2003.

14. J. HALL. and I. BRUNTON: Critical comparison ofJKMRC blast fragmentation models, Proceedings of theEXPLO 2001 Conference, Hunter Valley, NSW, 2001,207212.

15. S. S. KANCHIBOTLA, W. VALERY and S. MORRELL:Modelling fines in blast fragmentation and its impact oncrushing and grinding, Explo 99 A conference on rockbreaking, The Australasian Institute of Mining andMetallurgy, Kalgoorlie, Australia, 1999, 137144.

16. P. D. KATSABANIS: Explosives technology open pit andunderground blasting, Part D, Chapter 5, Canada,Queens University, 1996.

17. J. KEMENY, K. GIRDNER, T. BOBO and B. NORTON:Improvements for fragmentation measurement by digitalimaging: accurate estimation of fines, Sixth InternationalSymposium for Rock Fragmentation by Blasting,SAIMM, 1999, 103110.

18. T. KOJOVIC, S. MICHAUX and C. MCKENZIE: Impact ofblast fragmentation on crushing and screening operationsin quarrying, Proceedings of the EXPLO 1995Conference, Brisbane, QLD, 1995, 427435.

19. Q. LIU and P. D. KATSABANIS: A theoretical approach tothe stress waves around a borehole and their effect on rockcrushing, Proceedings of the Fourth InternationalSymposium on Rock Fragmentation by Blasting Fragblast-4, 1993, 916.

20. P. MOSER, M. OLSSON, F. OUCHTERLONY and A.GRASEDIECK: Comparison of the blast fragmentationfrom lab-scale and full-scale tests at Bararp, (ed. R.Holmberg), Proceedings of EFEE 2nd world conference



on explosives and blasting technique, Prague, CzechRepublic, 2003, 449458.

21. T. J. NAPIER-MUNN, S. MORRELL, R. D. MORRISONand T. KOJOVIC: Mineral communition circuits theiroperation and optimisation, JKMRC, The Universityof Queensland, Brisbane, Australia, 1996.

22. I. ONEDERRA and G. CORDER: Mine to mill atLaguna Seca, Minera Escondida Limitada, JKTechReport, Job No 03185, 2003.

23. R. ROSIN and E. RAMMLER: Laws governing finenessof powdered coal, J. Inst. Fuels, 1933, 7, 2936.

24. A. SCOTT, D. DAVID, O. ALVAREZ and L. VELOSO:Managing fines generation in the blasting and crushingoperations at Cerro Colorado Mine, Mine to Mill 1998Conference, The Australasian Institute of Mining andMetallurgy, Brisbane, Australia, 1998, 141148.

25. V. SVAHN: Generation of fines around a borehole: alaboratory study, (ed. Wang Xuguang), Proceedings ofthe 7th international symposium on rock fragmentationby blasting Fragblast 7, Beijing, China, MetallurgicalIndustry Press, 2002, 122127.

26. D. THORNTON, S. KANCHIBOTLA and J. ESTERLE: Afragmentation model to estimate run-of-minedistribution of soft rock types, Proceedings of the 27thAnnual Conference on Explosives and BlastingTechnique, ISEE, Orlando, Florida, USA, 2001, 4153.

Authors

Italo Onederra has over eight years' experience in the areasof rock breakage, excavation engineering and mininggeomechanics, with particular expertise in the continuousimprovement of both underground and open pit drilling and

blasting activities. He currently holds the position of SeniorProject Engineer at the Julius Kruttschnitt MineralResearch Centre (JKMRC), University of Queensland,Australia. He holds a BE (Hons) from the University ofMelbourne, a MEngSc from the University of Queenslandand has recently completed studies towards a PhD in miningengineering. Italo has worked in industry-funded R&D andconsulting projects involving more than 24 miningoperations in Australia and overseas including Chile,Argentina, Brazil, South Africa and Botswana.

Sedat Esen gained his BSc (1994) and MSc (1996) degrees inMining Engineering from the Middle East TechnicalUniversity, Ankara, Turkey and holds a PhD degree (2004)in Mining Engineering from the JKMRC, the University ofQueensland, Australia. He has applied research andconsulting experience in blasting engineering design,analysis and optimisation. He is currently a researcher at theSwebrec, the Swedish Blasting Research Centre at LuleUniversity of Technology (LUT). His research activitiesinclude detonation, fragmentation damage, blast designoptimisation, mine-to-mill and environmental problems dueto the blasting.

Aleksandar Jankovic is currently Metso Minerals ProcessTechnology Asia Pacific's Manager of Development andProcess Engineering. He gained his Master of TechnicalScience degree from the University of Belgrade in 1991 andhis PhD from the University of Queensland in 1997. He thenworked for Mount Isa Mines as a project metallurgist beforejoining the JKMRC in 1999 to work on Mine-to-Millrelated projects such as Century, Cadia, Alumbrera, MountIsa, Ernest Henry, Fimiston and the optimisation of SAG-Ball mill grinding circuits as well as tower mill modelling.



estimacion de finos generados por voladura

Documents