The influence of pillars on the Merensky Reef horizon … · shallow UG2 Reef mining could encounter conditions similar to ... Kroondal West Rustenburg Impala BRPM Crocodile Pandora

Transaction

Paper

427The Journal of The South African Institute of Mining and Metallurgy VOLUME 105 REFEREED PAPER JULY 2005

Introduction

Extensive UG2 Reef reserves underlie thepartially mined Merensky Reef (MR) withinthe Bushveld Complex. The middling betweenthe two reef horizons varies from 4 m to 400 m. Unmined ground in the form of pillars,potholes and remnants has been left duringthe mining of the MR horizon. The adverseinfluence of the unmined portion of MRhorizon has become a major concern whenmining on the UG2 Reef horizon. This paperevaluates the effect of the overlying unminedground on the MR horizon, on mining of theUG2 Reef. General recommendations formining and support strategies to achieveefficient and safe extraction of the UG2 Reefhorizon are also provided. The extent ofmining in the Bushveld Complex is shown inFigure 1.

Approximately 1400 MR remnants(including potholes, regional pillars, etc.) wereanalysed from platinum mining operationsacross both the eastern and western limbs ofthe Bushveld Complex. During the analysis,regular in-stope pillar systems (i.e. crushpillars) were not taken into account, as theireffect on UG2 Reef mining would benegligible1.

The study on the various shapes and sizesof the unmined reef (commonly referred to asremnants) left on the MR horizon from thevarious platinum mines selected in thisexercise revealed that the remnant sizesranged from 16 m2 to 96 000 m2, with anaverage size of 6 400 m2. A cumulativefrequency analysis indicated that 90% of theremnants studied were less than 17 000 m2 inarea with 50% being less than 2 000 m2 inarea.

Evaluation of mining on the UG2 ReefTo clearly understand the issues affecting themining of the UG2 stopes under the MRremnants, both underground testing andconceptual studies were undertaken. Theformer involved underground inspectionstogether with a detailed instrumentation andmonitoring programme. The conceptual studiesrevolved around literature studies andnumerical modelling. The numerical modellingcodes used for this assessment includedMINSIM 2000, FLAC 2D and DIGS.

The size and shape of the MR remnantsvaried significantly and an analytical approachto determine a generic method of evaluatingthe effect of the overlying MR remnants on thestoping on the underlying UG2 Reef horizonwas investigated (Urcan et al, 2004)2. MINSIM2000, an elastic boundary element program3,was used to study the stress regimes on theUG2 Reef horizon under the MR pillars. Aseries of circular, square and rectangularpillars with different dimensions were createdon the MR horizon and modelled usingMINSIM. In addition, the influence of k-ratioand dip angle were also studied.

The results of these conceptual models,shown in Figure 2, should be taken as a guideto estimate field stresses on the UG2 Reef inthe vicinity of the MR remnants. Such a guidewill be useful at mine planning meetings as it

The influence of pillars on theMerensky Reef horizon on stopingoperations on the underlying UG2 Reefhorizonby N. Singh*, H. Urcan†, K. Naidoo*, J. Ryder*, B.P. Watson*, A.M. Milev*, and M.K.C. Roberts*

Synopsis

Merensky Reef is mined extensively within the Bushveld Complex.In previously mined-out areas on the Merensky Reef horizon thereare pillars, potholes and remnants that have not been mined. Theseunmined blocks of ground on the Merensky Reef horizon caninfluence the mining of the underlying UG2 Reef. The extent of theinfluence will depend on the size of unmined ground and themiddling between the two reef horizons. This paper quantifies theinfluence of partially extracted Merensky Reef on mining of the UG2Reef, and aims to provide recommendations for mining strategies,as well as regional and local support to achieve efficient and safeextraction of the UG2 Reef, for various middlings beneath theMerensky Reef horizon. In some cases it was found that relativelyshallow UG2 Reef mining could encounter conditions similar tothose experienced in deep and ultra-deep mining environments.

* CSIR Mining Technology.† Groundwork Consulting Pty (Ltd).© The South African Institute of Mining and

Metallurgy, 2005. SA ISSN 0038–223X/3.00 +0.00. Paper received Jan. 2005; revised paperreceived May 2005.

▲

The influence of pillars on the Merensky Reef horizon on stoping operations

would provide the rock engineering practioner with a tool todetermine the possible stresses that would impact on the UG2mining without undertaking protracted numerical modellingexercises. This should be used only as a first pass methodand once there is agreement on the possibility of mining, amore detailed assessment must be undertaken prior to anymining.

The MR remnants were modelled as isolated pillars in a980 m x 980 m area. Hence, these model results may beconsidered as the worst-case scenario. In actual mininglayouts, there will be other remnants in the vicinity, whichwill share the overburden load, resulting in reduced stresseson the remnants. Average Pillar Stress (APS) for themodelled MR remnants for various sizes and depths is givenin Table I.

To show the method of using Table I, an example ofdetermining the first pass APS on an MR remnant is providedas follows: for a pillar of 20 m radius, at a depth of 600 mbelow surface, the APS is estimated to be 321 MPa.

The findings of the MINSIM 2000 modelling are asfollows:

➤ The magnitudes of peak vertical field stress on the UG2reef were not significantly influenced by shapes of theremnants, for remnants with similar areas

➤ k-ratios of 0.5 and 2.0 were used in the modelling andit was found that the influences of the different k-ratios are minimal on the vertical stress field acting onthe UG2 Reef mining

➤ The effect of reef dip angles (10° and 20°) is minimalon the vertical field stress for UG2 Rreef mining

➤ Depending on the size of the remnant and the middlingbetween the Merensky and UG2 reefs, relativelyshallow depth UG2 Reef mining under the MRremnants will experience conditions that are similar todeep and ultra-deep mining conditions.

Potential failure mechanisms when mining UG2stopes under MR pillars

The ability to mine the UG2 Reef horizon safely andefficiently is determined by:

➤ Depth of mining

▲

428 JULY 2005 VOLUME 105 REFEREED PAPER The Journal of The South African Institute of Mining and Metallurgy

Figure 1—Mining within the Bushveld Complex (from Mintek)

Figure 2—MINSIM 2000 conceptual model design results

Polokwane

Mokopane

Pretoria

Lebowa

Messina

PPRust

Schematic Locality Planof the main

South AfricanPlatinum Mines and Projects

EastPlats

WestPlats

KroondalRustenburg

ImpalaBRPM

CrocodileRiverPandora

Northam

0 kilometres 125

Approximate Scale

Thabazimbi

Armandelbult

Union

Rustenburg

Marikana

Styldrift

Marula

DerBrocken

Bushveld ComplexMain mines & projectOther projectsMain Towns

Ga-PashaTwickenham

Modilowa

TwoRivers

Kennedy’sVale

ShebaRidge

BlueRidge

EverestSouth

Booysendal

DEPTH (m) 200 400 600 800 1000 1200 1400 1600 1800 2000

Middling (m) 10 20 30 40 60 100 150 200 300

Remnant Area (m2) 314 1257 2827 5027 7854 15394 31416 96211

Circular Remnant 10 20 30 40 50 70 100 175Radius 9m)

Equivalent Square 20 35 55 70 90 125 180 310Remnant dimension (m)

Equivalent Rectangular 10 20 30 40 50 70 100 180Remnant dimension 30 65 95 125 160 220 315 535(m). Approximately 3:1

Young’s Modulus (GPa) 50

k—Ratio 0.5 &2.0

Poisson’s Ratio 0.2

Grid Size (m) 5

Stoping Width (m) 1

Reed Dip (°) 20 &10 Off-R

eef Sheet

➤ The size of the overlying MR pillar➤ The distance (middling) between the Merensky and

UG2 reefs horizons.

Ground conditions for a stope mining on the UG2 Reefhorizon at relatively shallow depth changes drastically, dueto the increasing stress, as the stope face approaches the firstedge of the overlying MR remnant. This remnant acts as aconduit for the principal stress and there is a concentration ofstress ahead of the UG2 stope face. As the mining progressesbeneath the pillar, this concentration of stress increasessignificantly. The increase in the stress concentrationscontinues until the UG2 stoping has progressed beyond thefar edge of the MR pillar. As a result of this increase in stresson the underlying UG2 stope, conditions change from a low-stress, shallow mining environment to a high-stress miningenvironment with conditions similar to those experienced atdeep-level mining operations. Refer to Figure 3.

In order to study the UG2 Reef mining conditions underthe MR remnants, numerical modelling of a series ofscenarios was undertaken. Numerical modelling usingMINSIM was undertaken on seven existing Merensky andUG2 reefs mining layouts. Under selected MR remnants,extraction on the UG2 Reef horizon was modelled. No in-panel pillars (generally crush pillars) were modelled forconceptual UG2 Reef mining. The middling between the MRremnants and UG2 Reef mining varied between 20 m and 38 m. Stoping widths of 1.5 m were assumed for the models.Elastic constants, i.e. Young’s modulus of 50 GPa andPoisson’s ratio of 0.2, were assumed.

Figure 4a represents the model results for three differentremnants at a mining depth of approximately 1 450 m. Themiddling between UG2 and MR horizons was set at 20 m.The approximate size and APS of three MR remnants prior toany of the simulated mining occurring on UG2 Reef horizonare also shown.

To verify the change in the stresses acting ahead of theUG2 stoping, a section line was drawn through all three areasof concern. For discussion purposes only, the face stressesfor each of the conceptual UG2 mining steps, along sectionline A-A’, is shown in Figure 4b. For comparison, a facestress profile that is equivalent to single reef mining in solidground at a 1 550 m depth is also shown.

In all cases, the UG2 Reef face stresses under the MRremnants are in the region of 300 MPa or higher. These

values are almost double those that would be experienced bya stope on a single reef horizon mining in solid ground at a 1 550 m depth. When UG2 Reef mining is outside the area ofinfluence of the MR remnant, the stoping occurs in adestressed environment and face stresses are equivalent tothose experienced at very shallow depths. Similar findingswere achieved from other models that were analysed. One ofthe most significant findings was that the middling betweenthe two reef horizons is the most influential in determiningthe magnitude of the stress acting ahead of the underlyingUG2 stoping horizon.

In the course of MINSIM modelling of multi-reef miningscenarios as mentioned above, it was noted that the modelledvertical stresses were high, as expected. The horizontal stressresults within the immediate hangingwall of the UG2 Reefmining, for the 20 m middling, indicated that the once theUG2 Reef mining is within the abutment zone of the MR

The influence of pillars on the Merensky Reef horizon on stoping operationsTransaction

Paper


Table I

Average pillar stress (APS) for circular Merensky Reef remnants

Merensky Reef—depth below surface (m)

Merensky remnant 200 400 600 800 1000 1200 1400 1600 1800 2000radius (m)

175 15 30 44 59 73 89 104 118 133 148100 25 50 74 98 123 147 171 196 220 244 Merensky remnant70 35 69 103 137 171 205 239 273 307 333 APS (MPa)50 48 94 140 186 232 278 324 370 414 43440 60 116 172 229 285 342 400 455 502 52530 78 149 222 294 367 439 518 584 632 66620 112 215 321 425 530 624 746 842 887 93910 214 410 610 809 1007 1205 1404 1568 1611 1641

Figure 3—Stress changes as UG2 stope as the face approaches anoverlying MR pillar

Merensky Pillar

Stress Trajectories

Potential High Shear Stress

Stress Trajectories

High Stress Concentration

Mined out

De-stressed by Merensky Mining - Conditions for UG2 mining will be similar to those at relatively shallowdepths (i.e. low face stress, low closure rate, lessfracturing, low ERR levels and low seismic risk)

High stress due to Merensky remnant - Conditions for UG2 mining will be similar to those at greater depths (i.e.Increades face stress, high closure rate, increasedfracturing, high ERR levels and increased seismic risk)

UG 2 Stope Undermining the Merensky Reef (early stages)

UG 2 Stope Undermining the Merensky Reef (later stages)

▲


remnant, the horizontal stresses were highly compressive onand ahead of the face. These stresses become tensile withmagnitudes of up to 200 MPa (tensile) in the mined-outareas as shown in Figure 5.

This situation appeared to be implausible and thepossibility of some form of modelling or program error wasmooted. Confirmatory modelling exercise using the 2D DIGS4

program was undertaken, using the geometry shown inFigure 6. Stages of UG2 Reef mining were then modelled,starting from the 200 m position shown in Figure 6 andprogressing to a final face position of 260 m (ultimate spanof 60 m).

In the above model, the following inputs were used:

➤ middling of 20 m➤ E = 80 GPa➤ ν = 0.2➤ σv = 45 MPa (depth = 1500m), σh = 45 MPa (k = 1)➤ grid sizes: MR = 2 m, UG2 = 1 m.

The analytic APS on the 20 m MR remnant, prior to UG2reef stoping, was calculated to be 396 MPa.

Figure 7a shows that only a relatively small amount ofdestressing occurred on the MR remnant as understopingprogressed towards the centre of the overlying remnant. The

▲


Figure 4a and 4b—Conceptual modelling and associated results

Figure 5—Horizontal stress results for 20 m middling; mining depth = 1 450 m

Figure 6—Geometry modelled using DIGS

AREA = 300 m2

APS = 446 MPa

AREA = 500 m2

APS = 388 MPa

AREA = 12112 m2

APS = 292 MPa

-350

-300

-250

-200

-150

-100

-50

00 2 4 6 8 10 12 14 16 18 20

Mining Step - 7Inside the Remnant

Mining Step - 8Limit of Remnant

Initial toMining step 4

Mining Step - 6Edge of Remnant

Face stress profile forminng at 1550 m withoutany remnant above

Initial Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 Step 7 Step 8Step 9 Step 10 Step 11 1550m

VE

RT

ICA

L S

TR

ES

S (

MP

a)

B: Face stresses along A-AA: Mine layout used for modelling

FACE STRESSES ALONG SECTION A-A (20m MIDDLING) - MINING DEPTH = 1450 mAVERAGE PILLAR STRESS (20 m MIDDLING)

HORIZONTAL STRESS FOR UG2 MINING

HO

RIZ

ON

TA

L S

TR

ES

S (

MP

A)

0 2 4 6 8 10 12 14 16 18 20

-300

-200

-100

0

100

200

300

0 250 270 520

200 260

UG2 mining direction

modelling results, however, showed that the elasticconvergence in the UG2 Reef stope accelerated sharply.Shown in Figure 7b are the plots of the variation ofhorizontal (skin) stresses in the hangingwall of the UG2 Reefmodel as stoping progresses. The same high tensile stressesare present in this model as in the MINSIM models discussedearlier. This modelling served to demonstrate the very likelypresence of tensile horizontal skin stresses. Presented byRyder (2005)5

The following technical argument can be used to describethe mechanism of the origin of the high horizontal tensilestresses. The induced horizontal stress on the reef horizon(due to stope convergence in elastic ground) is the same asthe induced vertical stress. For example, when stoping on asingle horizon, the induced vertical stress must equal minusthe field stress of σv and so the induced horizontal (skin)stress is also –σv, which means the absolute horizontal stressis small and compressive once the field horizontal stress σh isadded back (provided k>1).

When stoping a second reef in fully destressed ground,the field stresses are low and compressive, and so theinduced horizontal stress is low and tensile, giving a nettabsolute horizontal stress likely to be low and compressive.Once the second reef transgresses into the rockmass wherethe field vertical stress is high, however, large tensilehorizontal stresses will be induced and the likely fieldhorizontal compressive stress is unable to offset this largelevel of tensile stress.

Apart from complications due to interaction between thetwo reefs (possibly fairly small, judging from the APS curveof Figure 7a), it is proposed that the further increase ofhorizontal skin tension is ascribable to shear-induced ride onthe UG2 Reef mined in the vicinity of the remnant. Thestresses illustrated in Figure 7b are in the hangingwall skinof the UG2 Reef, but further examination of the DIGS resultsshowed that the horizontal tension persisted typically 3 m to

8 m up into the hangingwall. In an even larger zone, morethan 10 m into the hangingwall, substantial tensile verticalstresses greater than 5 MPa were also present.

This combination of high elastic convergences, very highhorizontal tensile stresses (able to rupture even the strongestintact rocks) and significant vertical tensions (able to rupturestrong parting planes) is highly unfavourable forhangingwall stability. Observations from undergroundconditions found increasingly blocky and unstablehangingwall conditions as stoping approaches andtransgresses a strong stress shadow in a low middlingsituation, which confirms these conclusions. Figure 8 showssome of the conditions observed when the UG2 stoping isclose to the abutment of the overlying MR faces.

Implications for support design are considerable,especially once the UG2 stoping face reaches the edge of thedestressing shadow, as shown in Figure 9. Substantialupgraded support systems become necessary, with both highareal coverage and the ability to accommodate high closure,when stoping operations are being carried out in the highabutment stress zone. In the modelled environment, therequired support resistance appears to be such as to supportat least 10 m of rock, that is, about 300 kPa. Suitable supporttypes would be, for example, backfill with supplementarystiff elongates, or appropriately designed grout packs at closespacing.

The poor UG2 hangingwall conditions encounteredunderground when mining under an MR remnant appear tobe worse than those that might be expected as a result of thehigh levels of vertical stress. In fact, strong induced tensilestresses (which are not normally encountered in ultra-deepsingle-reef mining) are likely to lead to poor hangingwallconditions, with upgraded support requirements. From thetwo scenarios investigated (20 m middling vs. 38 mmiddling), it was observed that the horizontal stresses havehigh magnitudes in both compression and tension. It is


Paper


Figure 7—Influence of APS, convergence and horizontal stresses on UG2 stoping

0

-50

-100

-150

-200

-250

220 230 240 250 260 270 280

Position on UG2 horizon (m)Fiel

d st

ress

es o

n U

G 2

hor

izon

(MP

a)

(a) Field stresses on the UG2 reef horizon, prior toUG2 Reef stoping (tension positive notation).

400350300250200

150100500

220 230 240 250 260Position on UG2 horizon (m)

Hor

.H/W

Ski

n S

tres

s (M

Pa)

(c) Horizontal hangingwall skin stresses in theUG2 (tension positive)

450400350300250200150100500

0 20 40 60 80

Field Sigz on UG2Field Sigx on UG2

UG2 span =40mUG2 span =50mUG2 span =60m

APS on remnant(MPA)Max. Cvgence onUG2 (mm)

Span on UG2 (m)

AP

S, M

ax. c

vgen

ce

(b) Variation of APS on temnant and max.convergence in UG2 Reef stope as UG2Rreefstoping progresses under the remnant

▲


suggested that the middling has a significant influence onthis aspect of the stress regime acting on the UG2 Reefhorizon. Further work is clearly required to confirm anddelimit these conclusions.

Underground monitoring programme

Continuous closure monitoring was conducted at two sites ata deep platinum mine in the Western Bushveld to obtain anunderstanding and a comparison between ground behaviourin destressed conditions and conditions when mining under aremnant. It was also the intention to compare measurementsof closure with convergence data obtained from numericalmodelling reported in the earlier section of this report. Twodifferent sites were monitored:

➤ Mining the UG2 reef under MR remnant conditions;and

➤ MR virgin ground conditions. This site was chosen tocompare the values obtained for mining operationsunder virgin conditions.

Telescopic closure meters, developed by the CSIR, wereinstalled in the panels to monitor the closure in the area.peak velocity detectors (PVDs) were installed adjacent to eachclosure meter in order to obtain information on peak particle

velocities (PPVs) that could result from seismic events undernormal and high stress conditions. Figures 10a and 10bshow the two types of instrumentation units that weredeveloped by the CSIR and installed in the stope.

A useful closure parameter is the closure ratio (CR). For asingle mining increment j, CR is defined as the ratio ofinstantaneous to total daily closure. Calculation of CR isdemonstrated in Figure 11 after Malan (2003)6.

From previous monitoring and analysis of closure rates, ithas been observed that there is a good correlation betweeninstantaneous closure and PPV data. In all the areasmonitored, PPV instruments were installed in the panel;however, not all instruments successfully recorded data dueproblems encountered underground. Malan and Napier(1999) found that closure ratios were a good indicator ofgeotechnical conditions7. From underground observations itwas noted that low closure ratio values were typicallyassociated with poor hangingwall conditions and a high riskof falls of ground, whereas high closure ratios wereassociated with areas prone to face bursting on the deep-levelgold mines. A plan of the area monitored is presented inFigure 12. The stope that was investigated was at a depth of1 556 m with a 20 m middling separating the UG2 andMerensky reefs.

▲


Figure 8—Conditions observed in UG2 stopes close to MR abutment

Locality plan and site of photographs.

A view into panel 1N from position A and B.This position is close to the abutment of theoverlying Mrensky Reef pillar. At least 200mm of stope closure had accurred, thedeformation of the support units attest to this.

The view from C into the 1N strike gully. The blocky natureof the ground should be noted. The high stress environmenthas caused differential movement of blocks rather thangenerating stress fracture

A view from position D of the strike lines of pillars belowthe 1N gully. Such spalling would be normal below a strikeline of crash pillars cut at a width to height ratio of 2:1

Closure in excess of 300 mm at position F

The approximatesize of remnant is170 m by 170 m

1N

2N

4N

Merensky solid

3N

Figure 9—The 45° rule for destressing showing the high abutment stress zone

Near side of abutment Far side of abutment

UG2 stoping Future UG2 stoping under abutment

45°45°

Direction of mining

Merensky stoping Merensky pillar Merensky stoping

Destressed or stress shadow zoneHigh abutment

stress zoneDestressed or stress shadow zone

The closure meters were initially placed 1 m apart asmining advanced to obtain maximum data input. Aninterpretation of the data shows that the maximumdeformation occurs during blasting time. The closurebehaviour of the UG2 Reef is characterized by a small instan-taneous jump after blasting, followed by a low steady-stateclosure rate. The circles, in Figure 13, show the stopeproduction blasts. The average closure rate is 3 mm per day.The average closure ratio calculated for this stope was 0.36.The closure profiles from different positions within the stopeshowed similar patterns. The closure profile for the UG2 Reefpanel is shown in Figure 13.

Monitoring in the panel was stopped when the panel wassubjected to strong ground motion from a seismic event,resulting in a fall of ground. The average closure ratio of 0.35obtained from the UG2 Reef mining under a MR remnantpillar is similar to those obtained by Malan (2003) on thestudy undertaken on the Carbon Leader Reef, in seismicallyactive areas within deep-level gold mines. Further work isrequired to confirm this initial conclusion.

Underground seismic measurements were carried out at anumber of different positions within the stope. However, due

to the difficult underground conditions and high humidity,most of the PVDs were damaged or lost. Two data-sets wereavailable for analysis. The first one contained 484 seismicevents with a maximum PPV of 2.4 m/s, recorded over aperiod of 7 days. The second contained a limited number of 6events with a maximum PPV of 0.13 m/s, recorded over aperiod of 5 days.

Case studies

In order to evaluate the conditions of the UG2 Reef miningunder the MR remnants, various sites were visited. Thesesites were then modelled, using MINSIM 2000, to study thestress regimes and convergence levels on the UG2 Reefmining. This was an attempt to link observed conditions tothe numerical modelling. An example of one such case studyis discussed below.

Investigations were carried out in January 2004, in astope approximately 1 500 m below surface. The middlingbetween the UG2 and Merensky reefs was 20 m. The miningon Panel 4E extended halfway across the MR remnant afterPanel 4E had mined across the top segment of the remnant,as shown in Figure 14.


Paper


Figure 10—Instrumentation units developed for underground observations

Figure 11—Typical continuous stope closure after blasting (after Malan, 2003)

(a) Closure meter (b) Peak Velocity detector

CRn

S

Sij

Tj

j

n

==

∑1

1

∆∆

▲


Modelling results for Panel 4E are presented in Figure 15a and 15b. Prior to any UG2 reef mining in thisarea, the vertical field stress under the MR remnant isapproximately 120 MPa on UG2 Reef elevation. In June 2002the face position of the UG2 Reef was 12 m behind the MRremnant and lagging 26 m behind Panel 5E, the face stresswas 130 MPa. In September 2002 no mining took place atPanel 4E. Panel 5E was advanced by 54 m, resulting in Panel4E lagging 80 m behind Panel 5E. At this stage face stresses

in Panel 4E increased to 160 MPa. In December 2002, Panel4E was directly under the middle of the Merensky Reefremnant. At this position Panel 4E lagged behind the Panel5E by 64 m and the face stress increased to 370 MPa. Theresults shown in Figure 15b highlight the considerableincrease in convergence under and ahead of the MR remnant.The maximum convergence under the remnant is 95 mm.

Severely fractured face and evidence of fall of ground isshown in Figure 16a. This photograph was takenimmediately down-dip of Pillar 4 looking towards Panel 4Eface. High closure and hangingwall fracturing was presentunder the Merensky Reef remnant and is shown in Figure 16b. Pillar 2 showed minor fracturing, as shown inFigure 16c. The pillar was 4 m behind the Merensky Reefremnant and the size of the pillar was 4 m and 2 m wide onwest and east sides, respectively, with a length of 6.5 m. Thispillar was ledged (carried a siding) approximately 0.5 m to 1m from west to east. Figure 16d shows severe scaling thatwas present on the down-dip side of Pillar 4. This pillar wassited immediately under the MR remnant without a siding.

The numerical modelling for the eastern panels indicatedthat the MR abutment edge exerted a stress of 120 MPa onthe UG2 Reef elevation (prior to UG2 Reef mining in thevicinity of the remnant). These edge stresses indicate that theconditions may be similar to those at ultra-deep-level miningwith an added hazard of the high horizontal tensile stressesdescribed earlier adding to potential hangingwall instability.

Conclusions

The study on the MR remnant shapes and sizes (includingpotholes, regional pillars, etc.) from the various platinummines revealed that the remnant sizes ranged from 16 m2 to96 000 m2 with an average area of 6 400 m2. Cumulativefrequency analysis indicated that 90% of the remnantsstudied were less than 17 000 m2 in size with 50% being lessthan 2 000 m2 in size.

MINSIM 2000 studies indicated relatively shallow depthUG2 reef mining under the Merensky Reef remnants canexperience conditions that are similar to deep and ultra-deep

▲


Figure 12—Plan of the UG2 Reef stope panel that was instrumented (notto scale)

Figure 13—Continuous closure recorded for UG2 Reef stope

12m from face

11m from face

10m from face

9m from face

8m from face

7m from face

6m from face

03/0 04/07 05/07 06/07 07/07 08/07 09/07 10/07 11/07 12/07 13/07 14/07 15/07 16/07 17/07 18/07 19/07 20/07 21/07 22/07 23/-7 24/07 25/07 26/07 27.07 28/0700:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00 00:00

Time dependentbehaviour

Instrumentationset-up problems

24

22

20

18

16

14

12

10

8

6

4

2

0

Northam 6–24 4W closure 0361 20m middling 1556m depth installed 3/7/03 9:35am 6m from the face

Date and Time

Clo

sure

(m

m)


Paper


Figure 14—Locality plans of pillar on Merensky Reef horizon (in dashed outline)

Figure 15—MinSim model results for Panel 4E

Overlying Remnant

UG2 stress due toabove Merensky Remnant

MININGDIRECTION

RAISE LINEE

E

DISTANCE

VE

RTI

CA

L S

TRE

SS

(MP

a)

UG2 Pillar

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

2002/06 2002/09 2002/12 2003/03 2003/06 2003/09 2003/12 2004/01

FACE STRESSES FOR UG2 MINING - SECTION B-B’ - 5–30 PANEL 4E

OverlyingRemnantHigher convergence due to

overlying Merensky Remnant

MINING DIRECTION

RAISE LINE

DISTANCE (m)

CO

NV

ER

GE

NC

E (m

m)

UG2 Pillar

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100

120

100

80

60

40

20

0

2002/06 2002/09 2002/12 2003/03 2003/06 2003/09 2003/12 2004/01

Convergence FOR UG2 MINING - SECTION B-B’ - 5–30 PANEL 4E

▲


mining conditions. This finding is dependent on the size ofthe MR pillar as well as the middling between the two reefhorizons.

Conceptual MINSIM studies indicated the presence of highhorizontal stresses (compressive and tensile) in thehangingwall of the UG2 stope when mining within theabutment zone of the MR remnant. Numerical modelling ofthe hangingwall skin of the UG2 Reef using DIGS showedthat the horizontal tension persisted typically 3 m to 8 m upinto the hangingwall. In an even larger zone, more than 10 minto the hangingwall, substantial tensile vertical stresses inexcess of 5 MPa were noted.

A closure ratio of 1 was obtained from the MR panels.This value of 1 implies that the rockmass is showing trueelastic behaviour. The average closure ratio is 0.35 in UG2reef stope panels that mined into the high abutment stresszone caused by the overlying MR remnants. A closure ratioof 0.36 is typical for the Carbon Leader Reef, which is foundin deep-level gold mines. It can be argued that, when miningunder the overlying pillars, the conditions experienced will besimilar to those of mining at depth. Further work will beneeded to validate this conclusion.

From underground observations, it is considered that the45° rule is applicable for estimating the zone of MR remnantstress concentration. In the proximity of these remnantsdifferential and time dependent movement occurs in the UG2Reef hangingwall between blocks constituting the jointed andat times fractured hangingwall. Stress-induced deformationof the rockmass by mobilizing joint-bounded blocks appearsto be the preferred method of stress relaxation in the stopeface rather than fracturing in the hangingwall.

When UG2 stoping operations are within the abutmentstress zone from the overlying MR remnant, the supportsystem will have to be changed so that it can cater for largehangingwall and footwall deformations while maintaining

the hangingwall integrity using support units that provideareal coverage. One such example would be backfill.

Within the zone of MR remnant stress concentration (45°rule) UG2 Reef in-stope pillar size is also critical. Oversizepillars result in high pillar stresses prior to failure, which canbe violent. The consequence of this can be pillar punchingand associated hangingwall damage.

Acknowledgements

The permission from the PlatMine Committee to publish thispaper is gratefully acknowledged.

References

1. OZBAY, M.U. and Roberts, M.K.C. Yield Pillars in Stope Support,Proceedings of Symposium Rock Mechanics in South Africa, 1988,SANGORM.

2. URCAN, H., NAIDOO, K., SINGH, N., ROBERTS, M.K.C., RYDER, J., WATSON, B.P.,SPOTTISWOODE, S., JAGER, T., GRAVE, M., and MILEV, A.M. Strategies for themining of the UG2 Reef below partially extracted Merensky Reef, PlatMine2004

3. NAPIER, J.A.L. and STEPHANSEN, S.J. 1987. Analysis of deep level Minedesign problems using the MINSIM-D boundary element program. APCOM’87. Proc. 20th Int Symp. SAIMM. pp 3–19

4. NAPIER, J.A.L. and HILDYARD, M.W. Simulation of fracture growth aroundopenings in highly stressed, brittle rock, J. S. Afr. Inst. Min. Metall. 921992. pp. 159–168.

5. RYDER, J. Internal memorandum on very high horizontal tensile stresses inmultireef environment, CSIR –Miningtek report. 2004.

6. MALAN, D.F and NAPIER, J.A.L. Guidelines for measuring and anlaysingcontinuous stope closure behaviour in deep tabular excavations, SIMRAC.Johannesburg, 2003. 67 pp.

7. MALAN, D.F. and NAPIER, J.A.L. The effect of geotechnical conditions on thetime-dependent behaviour of hard rock in deep mines. Amadei, B., Kranz,R.L., Scott, G.A. and Smeallie, P.H. (eds.) 37th U.S. Rock MechanicsSymposium, Vail Rocks ’99, Balkema, dated 1999 pp. 903–910. ◆

▲


Figure 16—Severely fractured face and evidence of fall of ground in panel

(a) View from B—down-dip side ofPillar 4, looking towards face

(d) View from E—down-dip side of Pillar 4

UG2 reef stope—references tolocation of photographs taken

(c) View from D—up-dip side of Pillar 2,4 m and 2 mwide on west and east respectively, length 6.5 m. Ledge0.5 m to 1 m from west to east. Minor fracturing

(b) View from C—abutment over panel 4E,immediately down-dip ofpillar, between pillar andbackfill

Merenskyremnant

Documents

The influence of pillars on the Merensky Reef horizon … · shallow UG2 Reef mining could encounter conditions similar to ... Kroondal West Rustenburg Impala BRPM Crocodile Pandora