Isc Vennes,Hni Mitri,Dmodr Reddy Chinnsne,Mike Yo

a Department of Mining and Materials Engineering,McGill University,Montreal,Quebec H3A 0E8,Canada

b Copper Cliff Mine,Vale Canada Ltd,Sudbury P0M 1N0,Canada

Keywords:Destress blasting Preconditioning Rockbursts Strainbursts Numerical modeling

ABSTRACT Destress blasting is a rockburst control technique where highly stressed rock is blasted to reduce the local stress and stiffness of the rock,thereby reducing its burst proneness.The technique is commonly practiced in deep hard rock mines in burst prone developments,as well as in sill or crown pillars which become burst-prone as the orebody is extracted.Large-scale destressing is a variant of destress blasting where panels are created parallel to the orebody strike with a longhole,fanning blast pattern from cross cut drifts situated in the host rock.The aim of panel destressing is to reduce the stress concentration in the ore blocks or pillars to be mined.This paper focuses on the large-scale destress blasting program conducted at Vale’s Copper Cliff Mine (CCM)in Ontario,Canada.The merits of panel destressing are examined through field measurements of mining induced stress changes in the pillar.The destressing mechanism is simulated with a rock fragmentation factor (α)and stress reduction/dissipation factor(β).A 3D model is built and validated with measured induced stress changes.It is shown that the best correlation between the numerical model and field measurements is obtained when the combination of α and β indicates that the blast causes high fragmentation (α=0.05)and high stress release(β=0.95)in the destress panel.It is demonstrated that the burst proneness of the ore blocks in the panel stress shadow is reduced in terms of the brittle shear ratio (BSR)and the burst potential index (BPI).

1.Introduction

In the past 20 years,large-scale destressing has been applied successfully in both coal and hard rock mines.While in principle all destress blasting programs aim to reduce the potential of rockburst,its application is primarily one of two types,namely rigid roof destressing for stiffness reduction of the overlying strata in longwall coal mining,and panel destressing in hard rock mines for stress reduction in the ore pillars to be mined.

A feature of a rockburst prone coal mine is the proximity of the working face to thick and rigid host rock stratum [1,2],where abrupt caving of stiff overlying strata could result in dynamic loading of the mine,posing a serious threat to the safety of the working face.Instances where this rockburst mechanism is found is the retreat longwall mining at the Lazy Colliery in the Czech Republic,and retreat room and pillar during pillar recovery where it is necessary to fracture the roof periodically as the coal mining face advances [3-5].Reduction of stress at the face itself is not explicitly sought.The depth of the Lazy Colliery is 700 m,with coal seams that are 3.1-5 m thick.The longwall face is prone to rockbursting as it tends to accumulate high strain energy due to a stiff massive sandstone roof.The goal of the destressing program is therefore to reduce the strength and massiveness of the roof strata by creating a network of fissures.The destress horizon is targeted with long boreholes that are drilled from the gate roads.The entire length of a panel is destressed in roughly 20 stages averaging 5 drill holes per stage.The seismic effect parameter developed by Konicek et al.is the primary means of evaluating the effectiveness of a destressing stage,with most blasts yielding at least a very good seismic effect [3].Based on stress change measurements from compact conical-ended borehole monitoring (CCBM)probes,the range of influence of the site is measured to have decreased,demonstrating that the competency of the overlying strata was reduced by blasting.The longwall panel was extracted without any further rockbursts following the destress blasting program.A variation of the Lazy Mine destressing technique was applied in the Polish part of the Upper Silesian Coal Basin,referred to as torpedo blasting[6].In contrast to the Lazy Mine,the charge per stage is much lower ranging from 288 to 576 kg due to the lower charge length.Based on the seismic effect,most stages provided at least a very good destressing effect.

Panel destressing practice in hard rock mines is exemplified by 2 Canadian case studies.For these mines,the relative stiffness of the ore and host rock is not necessarily a major factor,and the purpose of destressing is primarily to create a stress shadow that shields to ore to be mined from high stress.At Fraser Mine,the objective of the destress blast was to directly reduce stress in the ore sill pillar being mined out[7].The extraction of the subsequent cuts was therefore facilitated,but with the expectation that global failure of the sill pillar would be accelerated.To create a destress panel,two rows of 14 blast holes were fanned from a cross cut drift,yielding a 1×107kg panel with an average explosive energy of 1617 J/kg of targeted mass.The blast took place in December 2001,and vibrating wire stress cells recorded a horizontal stress decrease perpendicular to the panel,and a vertical stress increase in the stress shadow.Similarly,a large-scale (2.7 × 107kg mass)choked panel destress blast was conducted at Brunswick Mine in October 1999[7,8].The blast pattern consisted of two parallel rows of 16 holes drilled pillar,with no free face.The blast pattern yielded an explosive energy of targeted mass of 840 J/kg.A stress drop was detected by the downhole gauge in the stress shadow,and a minor stress increase was detected in the uphole gauge.In both above-mentioned case studies,the measured stress changes indicated stress wrapping around the panel.

Overall,the Lazy Colliery case study demonstrates a successful destress blasting program where the goal is to fracture stiff overlying strata to facilitate its caving.A lower density of explosives is required compared to the Canadian case studies.The explicit goal of the blasts was not to reduce the major principal stress at face since the global failure of the roof is expected to occur as the mining face advances.The purpose is rather to reduce the dynamic effect of caving when it inevitably occurs.However,for the Canadian case studies,stress decrease was the explicit goal.Accelerated failure of the hanging wall was not an objective of the program but rather an accepted consequence of destress blasting.The blasts were choked as well,but with much higher explosive energy density.In the case of Fraser Mine,the numerical modelling back analysis where the panel material was simulated as a cavity (after destressing)conformed the most with in-situ measured stress cell changes,indicating that high displacement of panel material contributed to the high stress release.

The blast parameters of the case studies presented above are summarized in Table 1.The focus of this paper is Copper Cliff Mine(CCM),whose destress blasting program follows the philosophy of Fraser and Brunswick Mine destress blasts.The merits of destress blasting are examined through field measurements from uniaxial stress cells installed in the ore pillar,seismic data,and a minewide 3D numerical model of the destress panel and remnant pillars.The destressed zones in the model are simulated with the rock fragmentation factor (α)and the stress reduction factor (β).Themodel is validated based on the stress cell field measurements and the best conformity between the models and the stress cell measurement.

Table 1 Large-scale destress blast parameters.

2.Case study mine

The case study mine is Copper Cliff Mine(CCM),located in Copper Cliff near Sudbury,Ontario,Canada.The mining method employed is longhole open stoping with vertical retreat mining(VRM).The orebody of interest,100 orebody (OB),is pipe-shaped and nearly vertical,extending 1200 m vertically and 150 m horizontally.The ore consists of inclusions of massive to disseminated sulphide mineralization,while the massive quartz diorite host rock is composed of amphibole,biotite,and chlorite.Two geological structures are present near the orebody:a strongly sheared Olivine Dyke and a stiff Quartz Diabase Dyke.Ore below 3880L was mined out in a bottom-up,center out sequence,ending in March 2009.Ore above 3500L was mined in the same sequence,with the last stope mined in October 2009,leaving a remnant pillar between 3880L and 3500L under high mining induced stress (Fig.1).

Based on a separate numerical modelling study,it was determined that the proper sequence to reduce stress in the remnant ore is a North-South retreat from the Olivine Dyke,alternating between the sill and crown stopes.Nonetheless,the rockburst potential of the sill pillar was still known to be high based on computed mining induced stresses and increased seismic activity being experienced.A dynamic support system has been widely implemented in all active mining areas and rockburst control techniques were explored.

The ore pillar is composed of 18 stopes between 3880L and 3500L.There are 11 ‘‘sill stopes”between 3550L and 3880L,and 7 ‘‘crown stopes”between 3550L and 3500L.The sill stopes are mined with the VRM method,while the crown stopes are mined with upholes from 3550L.The drilling horizon for the sill stopes is 3550L and muck is drawn from 3710L.For the crown stopes,the stope is mucked and drilled from 3550L (Fig.2).

To drive stress away from active stopes in the pillar,a panel destressing program is implemented in 4 phases (Fig.3).The panels are drilled and blasted from extensions of the stope cross-cuts that drive across the orebody into the hanging wall.The overall extraction sequence of the pillar,including the destressing phases,is provided in Table 2.The stope nomenclature is explained in Fig.4.The scope of this paper is the Phase 1 panels that were blasted on September 21,2015,and the extraction of stope 9631 that took place from October 2016 to January 2017.A plan view of the Phase 1 blast with respect to stope 9631 is shown in Fig.5.

Fig.1.Mining induced stress in the 100 OB pillar (outlined in white)in November 2014,before destressing is implemented,calculated by volumetric averaging.In red are zones with a computed major principal stress ＞90 MPa.The bulk of the pillar major principal stress is 55-60 MPa.

Fig.2.Section view of the remnant pillar before destressing.

Fig.3.Planned destress panels at CCM,looking East.Mined and backfilled stopes are shown in blue,mine developments are shown in green.Planned stopes in red are completely shielded by the destress panels.

The destress blast panels consist of both up and down holes with 2 rings of 114 mm diameter holes with a spacing of 1.8 m between the rings.An isometric view of typical drill pattern of up and down holes is shown in Fig.6.The up and down holes are collared such that a toe spacing of 2.6 m is maintained between the holes within the ring.Furthermore,the up and down hole rings are staggered such that a 2.4 m spacing is maintained between nearest rings.The Phase 1 destress blast required approximately 3000 m of drilling for both up and down holes.The holes were loaded with 23484 kg of emulsion using 370 detonators with a maximum charge of 223 kg per delay.However,the collars of both up and down holes were not loaded up to a length of 2-3 m.The holes were fired center out with 18 ms delay between the holes.Based on the design,the amount of emulsion yields an average energy of 2070 and 2154 J/kg for up and down holes,respectively.

Table 2 Planned mining sequence.

Fig.4.Plan views of sill and crown stopes in the 100-900 OB remnant pillar.Sill and crown stopes are located between 3550L and 3710L,and 3500L and 3550L,respectively.

Fig.5.Plan view of Phase 1 blast on 3550L.

3.Methodology

The Phase 1 destress blast is simulated holistically with the rock fragmentation factor (α)and stress reduction factor (β).The destress blast simulations are then validated based on uniaxial stress cell data,where the most conforming combination of α and β is retained.The first stope in the sequence (stope 9631)is then mined.The burst proneness of the ore in the stress shadow is quantified with the brittle shear ratio (BSR)and burst potential index(BPI).The incremental BSR and BPI are also evaluated.

Fig.6.Elevation view of 3550L panel blast pattern,with different coloring for each blasthole ring (marked in white and green).

3.1.Pillar instrumentation

Eight uniaxial stress cells were installed by CCM in the ore pillar before the Phase 1 destress blast.The stress cells were preloaded with a wedge and platen assembly to approximately 7-8 MPa during the installation.The stress cell measurements was set as initial as soon any stress change was measured.Subsequent readings were then analyzed with reference to the initial reading to determine the changes in stress from that point onwards.The stress cells were installed in vertical holes in the roof and floor of the sill drifts,in the North-South or East-West direction.The orientation of the stress cells and the measured stress change for Phase 1 destress blasts and the extraction of stope 9631 are shown in Table 3.The position of the downhole and uphole stress cells installed from 3550L are shown in Figs.7 and 8,respectively.Uphole stress cells(SC1,SC2,SC3,SC4,and SC7)were installed 7.5 m above the roof of the 3550L and downhole stress cells (SC5 and SC8)were installed 6 m below the floor of 3550L.It is notable that SC6 is located in a horizontal hole in 9550 sill,on 3550L.

Stress measurements were logged more frequently a few hours prior and after major production and destress blasts.Otherwise,the readings were logged at least once a day to once a week depending on the rate of change and mining activity in the area.The reported ‘‘immediate”stress changes in Table 3 is the difference between the last stress measurement before the blast and the first stress measurement after the blast.This immediate stress change is assumed to represent only the elastic stress redistribution caused by the stress release and stiffness reduction of the panel.Therefore,the computed stress changes in the numericalmodel after the destressing step can be directly compared to the measured immediate stress change.Subsequent stress changes that are measured in the hours to days following a blast are understood to be caused by progressive yielding of the rock mass,which eventually stabilizes as seismic activity subsides.Stabilized stress changes for Phase 1 are reported 29 days after the blast.Given that the model is linear elastic,the stabilized stress changes are not used for model validation.

Table 3 Measured stress changes for Phase 1 blast.

Fig.7.Position of downhole stress cells installed from 3550L with respect to the Phase 1 panels.

Fig.8.Position of uphole stress cells installed from 3550L with respect to the Phase 1 panels.

Fig.9.3D numerical model of the remnant ore pillar in 100 OB.

3.2.Numerical model

In the present study,a 3D model of the remnant pillar in the 100 OB is constructed in Rhino and discretized in Kubrix.The destress blasting program is simulated with FLAC3D.Fig.9 shows the final model geometry excluding the Trap Dyke and Olivine Dyke.

The model surfaces were constructed in Rhino and exported to Kubrix to generate a 3D grid for FLAC3D.For unmined stopes,the stope boundaries built in Rhino were based on the planned geometries.On the other hand,for mined stopes the surfaces obtained with a laser cavity monitoring system (CMS)were used and simplified.Similarly,drift as-builts were used to build the mine developments such as the ramp,3550L,3710L,and 3880L.The panels were also constructed in Rhino based on provided drill hole patterns.To find the panel geometry,it was assumed that a destress blast hole damage zone is 16 times the hole diameter[7].Therefore,with 2 rings of blast holes,the total panel targeted mass (Me)can be estimated as:

where d is the blasthole diameter;h the height of the panel;l the strike length of the panel;and ρrthe density of the rock.The resulting panel thickness is 3 m.

A mesh sensitivity analysis was conducted in a simplified pillar model.Different grid point spacings were applied to each domain surface depending on the desired accuracy.The final grid spacing for the panels and the stopes was set at 0.75 and 3 m,respectively,where higher accuracy was needed.Model boundary grid spacing was set at 25 m,and geological domain boundary grid spacing was set at 10 m.

The elastic material properties and rock mass characterization parameters obtained from previous laboratory results are shown in Table 4.The rock mass properties for each geological unit are calculated from the intact properties with the Hoek-Brown failure criterion parameters.For the ore and host rock,the rock mass stiffness applied to the model is calculated based on a geological strength index (GSI)of 60.For the Trap Dyke,a GSI of 100 is assumed.No intact data is available for the Olivine Dyke.However,since the Olivine Dyke is strongly sheared,it is unlikely to undergo brittle failure,and the failure envelope parameters are therefore not needed herein.

The bottom boundary is constrained in the z-direction,while all other faces are free to displace in any direction.The in-situ stress varies linearly with depth,and is given by the following empirical Sudbury regional stress equations:

where D is the depth below ground surface;σ1the major principal stress in the N12°W direction;σ2the intermediate stress in the N78°E direction;and σ3the vertical minor principal stress.The model is loaded with external boundary tractions,yielding different in pre-mining stresses in each geological domain in relation to the rock mass stiffness [9].The model x-faces and y-faces are loaded in the x-direction and y-direction,respectively,with shear stress boundary tractions in the x-y direction.Overburden pressure is applied to the top boundary.

3.3.Numerical simulation of destress blasting

Destress blasting is understood to have three effects on the targeted mass,which are the reduction of targeted mass stiffness due to fragmentation,the dissipation of stress due to fracture propagation,and the modification of the failure mechanism from brittle elastic to plastic yielding.In this paper,the first two effects are modelled holistically with the rock fragmentation factor (α)and the stress reduction factor(β).This methodology has been applied to single blasthole to assess the effectiveness of drifts destressing[10].The size of the damage zone was determined with a blast damage model.It was shown that for drift destressing pattern,the assumption that face is uniformly destressed is too optimistic and overestimates the volume of blast zones.However,drift destressing typically uses small diameter holes,with a low hole density in the blast pattern [11].In the case of panel destressing,a much higher explosive energy per targeted mass is applied,with a pattern hole density and hole diameter of a production blast.The assumption of uniform fragmentation and stress release in targeted mass is therefore valid.A previous parametric study of panel destressing strategy in a simple pillar showed that stress change magnitudes obtained with uniformly destressed panels can replicate the measured stress changes at Brunswick and Fraser Mine[12].

To begin,the rock fragmentation factor introduced by Tang and Mitri considers the reduction of stiffness of the targeted mass due to blasting [13].The modulus of elasticity is reduced in the panel by the factor α which ranges from 0 to 1:where Edestressis the destressed modulus of elasticity.

In addition,Tang and Mitri proposed Eq.(6)to estimate the Poisson’s ratio (v)due to softening of the rock mass [13]:

If the solution to Eq.(6)is greater or equal to 0.5,the Poisson’s ratio is set to 0.49.Tang and Mitri also introduced the stress reduction/dissipation factor (β),which represents an instantaneous release of stresses due to the blast [13].The residual stress tensor(σr)in the targeted zones is given by the following equation:

where β ranges from 0 to 1,and

The balanced stress state in the panel prior to destressing is replaced with the residual stress state (σr)defined by Eq.(7).This removes a portion of the strain energy in the panel proportional to the factor β,causing an imbalance between the model boundary work and the total strain energy in model.A new equilibrium is reached after solving the model where the final stress tensor inthe panel lies between the initial stress tensor and the residual stress tensor.Table 5 shows the tested combinations of α and β for Phase 1 destress blast.It is assumed that α+β=1.

Table 4 Elastic material properties and Hoek-Brown failure envelope parameters.

Table 5 Tested combinations of α and β.

According to Liu et al.,choke destress blasting should aim for a swell factor of 5%-15%,which will release ground stresses but keep fragmented ‘‘frozen”rock in-situ [14].However,the computed panel volume will be reduced for all possible combinations of α and β.Given that panel volumetric reduction is desirable and that the targeted material will swell,rock must be ejected from the panel to obtain the destressed panel stress state.At CCM,the stope crosscuts from which the panels are drilled acted as a void,and ejected material in the panel access drifts was observed.However,the holistic simulation approach used for destressing does not directly consider rock displacement,but replicate the effect of rock displacement by further reducing the panel stiffness.Therefore,the destressed modulus of elasticity is not the actual field modulus of the panel,but instead a reflection of the effect of both fragmentation and ejection of material from the panel.

3.4.Evaluation of pillarburst potential

The pillarburst potential of the ore pillar needs to be assessed before and after destressing to evaluate if the panel destressing strategy sufficiently reduces the stress in the pillar to ensure safe mining.The brittle shear ratio (BSR)and burst potential index(BPI)are both suitable parameters for this study.To begin,the brittle shear ratio (BSR)is expressed as follows [15]:

where σ1is the mining induced major principal stress;σ3the mining induced minor principal stresses;and UCSintactthe uniaxial compressive strength of intact rock.In a study by Martin and Kaiser,rock was found to undergo brittle shear as the ratio of the deviatoric stress to the uniaxial compressive strength exceeded 0.4,and the risk of strainbursts was deemed significant when the ratio exceeded 0.7[16].Ore zones in the model with a BSR exceeding 0.7 are therefore termed ‘‘at risk”.

On the other hand,the burst potential index (BPI)relates the energy storage rate of the rock to its’critical strain energy,whereas a BPI of 100%is initially assumed to be the threshold for high risk of rockburst.It is expressed as follows [17]:

where ESR is the total strain energy in the rock;and ecthe elastic strain energy capacity.Since the model is loaded with the boundary traction method,all zones in the model have zero stress under zero strain.Therefore,the ESR of a zone ‘‘i”is calculated based on the total stress and total strain in a zone:

To calculate the critical strain energy of the ore,host rock,and Trap Dyke,the energy at failure of a triaxial loaded specimen was calculated analytically.The axial load at failure was calculated with

the generalized Hoek-Brown failure criterion for an intact specimen [18]:

Hooke’s law was used to calculate the axial and radial strain at failure under triaxial loading,where σ11=σ1,σ12=σ23=σ13=0,and σ22=σ33=σr.The ESR of the specimen at failure can then be calculated analytically,and treated as the critical strain energy of the rock under triaxial conditions:

The critical strain energy for the Olivine Dyke is not calculated as the material is too soft to fail in a brittle manner.Also,the BPI is only valid in compressive zones at risk of brittle shear failure.Therefore,for zones in a low stress regime where σ3＜0,the BPI is set to 0.

3.5.Validation of destress blasting model

The stress change results from the stress cells and the model are not compared directly since the magnitude of the stress change error between the model and the stress cells does not adequately reflect model conformity.Model conformity is instead evaluated by measuring the distance error (ΔD)between the stress cell and the computed stress change contour with the same stress change.The distance error is calculated in the model by finding the closest zone in the stress cell where the computed stress change is within±0.1 MPa of the measured stress change.This process is illustrated in Fig.10.In this case,the effect of high contour density at the stress cell location is mitigated.

4.Results

4.1.Evaluation of destress blast input parameters

To begin,the holistic modelling approach described in Section 3.3 must be validated with the measured stress changes.Conformity between the stress measurements and the model is based on ΔD,defined as the distance between the stress cell and the closest zone in the model with a computed stress change equal to the stress cell measurement.Fig.11 shows the stress change shells superposed with their respective stress cells.For SC2,SC5,and SC7,the best match is obtained with model 1.For SC8,which reads a stress change of -1.5 MPa in the N-S direction,no clear trend is observed because the shell of a 1.5 MPa stress decrease intersects to position of the stress cell in all 4 models.The stress cell is therefore not well situated to measure the destressing effect of the Phase 1 blast,and by itself would have yielded inconclusive results.

Fig.10.Comparison between model stress changes and measured stress changes after destressing the Phase 1 panels(in red).SS7 (in yellow)reads a 6.2 MPa stress decrease in the N-S direction.The shell(in blue)shows where the computed stress changes in N-S equal -6.2 MPa.A small distance between the cell and the contour indicates good conformity between the numerical model and the stress change data.

Fig.12 shows the distance error (ΔD)results for all stress cells for models 1 to 4.SC2,SC5,SC6,and SC7 show a trend towards a high destressing effect,while SC1 and SC4 show a trend towards a lower destressing effect.Stress changes for SC3 and SC8 are not sensitive to destress blast input parameters,due to the positioning effect demonstrated in Fig.11d for SC8.

Finally,Fig.13 presents the spread of the ΔD values for the 4 destressing models.Model 1 has the lowest median ΔD value and the lowest maximum ΔD value.Considering the measured data from all other stress cells,model 1 is the most valid.This indicates that the panel is highly fractured,with the rock stiffness reduced from 25 to 1.25 GPa and a stress reduction of 95%.Therefore,the merits of the destress blasting program conducted at CCM will be evaluated with α=0.05 and β=0.95 as destress blast input parameters.

4.2.Evaluation of the merits of destress blasting

The abundance of stress change data obtained with the stress cells in the pillar provided a unique opportunity to validate the rock fragmentation factor and stress reduction factor constitutive models for a large-scale panel destress blast.In Section 4.1,the most suitable combination of destress blasting parameters was determined to be α=0.05 and β=0.95.

Fig.11.Comparison of Phase 1 blast stress change shells with the positions of 4 stress cells.

Fig.12.Distance error (ΔD)obtained with all 8 stress cells.

Fig.13.Distance error (ΔD)obtained for all 4 models and the quartile spread obtained with SC2,SC5,SC7,and SC8 for each model.

In this section,the first stope in the extraction sequence(stope 9631)is mined after destressing Phase 1 with these parameters.An overview of the modelled scenarios is given in Table 6.Stope 9631 is mined in 6 lifts bottom up.The final lift is named as ‘‘crown blast”,where the top half of the stope in terms of height is blasted.The stope is backfilled after the ore from the crown blast is mucked.The rockburst potential of the stope,and the stress shadow stopes are then evaluated with the BSR and BPI for each stope lift.The BSR and BPI are observed along two planes intersecting the stope (Fig.14).

On these sections,3 observation points were identified.The BPI and BSR for points 1 and 2 are plotted in Figs.15 and 16.Figs.15 and 16 show that for all points,destressing reduces the BPI and BSR over all mining steps.At point 1,the BSR is lowered from 0.58 to 0.52 before mining,and from 0.7 to 0.63 after mining.For point 2,the BSR is also reduced,with a BSR of 0.93 before destressing,and a BSR of 0.87 after destressing.At the final mining step,destressing reduced the BSR from 0.98 to 0.93.Point 3,under low deviatoric stress,also measures a significant BSR reduction.The same pattern is observed with the BPI,where the greatest reduction in burst potential is measured at point 2,from 51% to 44% after mining stope 9631.

Ore at risk,defined as ore with a BSR ＞0.7,is measured at all mining steps for both scenarios.The proportion of ore at risk in stope 9631 before the crown blast is measured at 15.0% with no destressing.The proportion of ore at risk is reduced to 7.7% if the Phase 1 panels are destressed.For the stopes located in the stress shadow of the Phase 1 panels,destressing will decrease the ore at risk from 14.2% to 9.4%.

Table 6 Stope extraction mining steps.

Fig.14.Section views of stope 9631 after the crown blast.Contours show the BPI,where a BPI ＞0.25 is in red.Observation point 1 is shown in the short section and observation points 2 and 3 are shown in the long section.

Fig.16.BSR and BPI at point 2.

Fig.17.Mass of ore at risk in the pillar stopes,excluding stope 9631.

After extracting stope 9631,the destress blast reduces the proportion of ore at risk these to 8.2%,from 10.3% if no destressing occurs.Destressing also reduces the tonnage of ore at risk in the entire pillar over the extraction of stope 9631.Before mining,a BSR cut-off of 0.7 yields 8.3×106kg of ore at risk with no destressing.After destressing,the ore at risk volume in the pillar is lowered to 8.0×106kg.The effect of destressing is more pronounced after the extraction of stope 9631 however,where the ore at risk volume is reduced from 12.5×106kg with no destressing to 10.3×106kg with destressing.

Fig.18.Ore at risk in stope 9631 during its extraction.Step 2 is the destressing of the Phase 1 panels.

Fig.19.Ore at risk in the Phase 1 stress shadow stopes during the extraction of stope 9631.Stress shadow stopes are stopes 9631 sill,9632 sill,9631 crown,and 9671 crown.

Ore at risk can also be evaluated with the BPI.To determine a suitable cut-off,the total volume of ore at risk delineated with the BSR at mining step 8 with no destressing is set as a benchmark.It is found that the tonnage of ore at risk obtained with a BPI cut-off of 25% is equal to the tonnage of ore at risk obtained with a BSR cut-off of 0.7 (see Fig.17).This BPI cut-off is therefore used to assess ore at risk for all other mining steps and destressing scenarios.The proportion of ore at risk in the pillar over the mining sequence of stope 9631 is plot in Figs.18 and 19.With the BPI cut-off of 25%,the destressing step will reduce ore at risk from 8% to 4.4% in stope 9631,and from 8.6% to 6.4% in the combined stress shadow stopes.After extracting stope 9631,ore at risk in the remaining stress shadow stopes decreases from 7.4% to 6.2%.

5.Conclusions

In this paper,it is shown panel destressing is beneficial to stress dissipation in the ore pillar thus enabling more efficient and safe mining.The stress reduction is quantified with an orebody wide numerical model and validated based on measured uniaxial stress change data.Validating the model however requires multiple stress cells,as placement and measurement error can be significant depending on the position of the cell with respect to the panel.In addition,results from single cell can be inconclusive based on its position with respect to the panel,with the same stress decrease being read at the stress cell position in all models.Only by combining results from multiple cells is it possible to conclusively validate the numerical model.Four combinations of α and β were tested and it is shown that the most conforming values are α=0.05 and β=0.95.This indicates that the panel was highly damaged,releasing most of the stress.Given these input parameters,the immediate stress change caused by panel blasting provides a beneficial stress decrease in the stress shadow stopes for the Phase 1 blast.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work is financially supported by a joint grant from MITACS Canada and Vale Canada Ltd,and the MEDA fellowship program of the McGill faculty of Engineering.The authors are grateful for their support.Technical and operational input along with the collaboration in developing these de-stress programs at Vale’s Copper Cliff Mine are greatly appreciated.