You are on page 1of 8

International Journal of Rock Mechanics & Mining Sciences 60 (2013) 381–388

Contents lists available at SciVerse ScienceDirect

International Journal of
Rock Mechanics & Mining Sciences
journal homepage: www.elsevier.com/locate/ijrmms

The effect of rock decompaction on the interaction of movement zones


in underground mining
F. Vivanco n, F. Melo
Laboratorio de Fı́sica no Lineal, Departamento de Fı́sica, Universidad de Santiago de Chile, USACH, Avenida Ecuador 3493, Casilla 307, Correo 2 Santiago, Chile

a r t i c l e i n f o abstract

Article history: An increased interest in scientific applications for underground mining, mainly to extend the produc-
Received 1 December 2011 tive life of open pit mines such as Chuquicamata in Chile, has motivated a growing effort to model
Received in revised form experimental and theoretically phenomena found in these mines as well as the processes involved in
29 December 2012
their operation. There is a general consensus that contamination as a result of dilution, a critical
Accepted 15 January 2013
Available online 1 March 2013
problem found in the operation of underground mines, might be reduced by an adequate design of
draw point grids and the appropriate handling of them. This requires understanding of the flow of rock
Keywords: fragments and the evolution of the movement zone created by the interaction of multiple draw points.
Sublevel caving In this paper, we present a theoretical study focused on determining the movement zone created by the
Kinematic model
interaction of two neighboring draw points operating in alternate mode that simulate those found in a
Isolated movement zone
sub-level caving mine. We employ a modified 2D kinematic model that includes a dilation front and
Draw points
assumes that rocks are restricted to move only along streamlines so that we may determine the
modification of an isolated movement zone that results from the extraction of material from a
neighboring draw point. The volume of extracted material required to initiate the interaction and the
location where it occurs are predicted in terms of the material’s previously extracted volume, diffusion
coefficient, density variations, and extraction rate. The results show that the top surface of the
previously isolated movement zone is modified in order to permit the surface to reach greater heights
and displace its maximum position closer to the operating draw point. We also find that the regions
outside of the operating draw point’s isolated movement zone are affected by the interaction and this is
confirmed by the deflection of tracer lines. This could have significant negative effects in underground
mining operations because dilution, initially located out of range of an operating draw point, might be
carried to either the neighboring draw points or the operating draw point’s opening, consequently
increasing pollution. The results presented can be extrapolated to 3D systems and generalized to other
type of flows described by more complex models than a kinematic model.
& 2013 Elsevier Ltd. All rights reserved.

1. Introduction main concerns in underground mining is dilution, the contamina-


tion of minerals by useless material. Draw points contaminated
Underground mining is under serious consideration to replace by dilution must be shutdown, incrementing the mining operat-
the longly, exploited open pit mine. Chilean mining company ing cost. In the mining community, it is thought that an adequate
Codelco plans to transform Chuquicamata, a milestone as one distribution of draw points and good handling of the material
of the biggest open pit copper mines, into an underground mine inside the mine might prevent or reduce the draw point pollution.
over the next few years. However, underground mines require a Therefore, a thorough understanding and characterization of the
considerable initial investment and have inherent risks. Within flow created by multiple draw points is required [1–6]. Despite
the techniques applied in underground mining is sub-level caving, the complexity found in underground mines, the main features of
a method that employs a lattice of draw points built below the flow of broken rocks during the extraction process have been
the ore. The extraction of material from these points induces a described by simple models based on empirical observations [3,5]
downwards motion and creates a flow of broken rocks. One of the and hypotheses that produce inaccurate predictions [2]. More
recently, kinematic and plasticity models have been introduced to
describe these flows in a more grounded way [7–9]. Although
n
Corresponding author. Tel.: þ56 996791090.
these models are different in nature, it has been shown that
E-mail addresses: francisco.vivanco@usach.cl, the main features of the flow can be reproduced regardless of
fvivanco@gmail.com (F. Vivanco). the model employed [7]. In a previous article, we studied the

1365-1609/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijrmms.2013.01.013
382 F. Vivanco, F. Melo / International Journal of Rock Mechanics & Mining Sciences 60 (2013) 381–388

movement zone that resulted from the interaction of multiple diffusion equation including their solutions can be found in
draw points operating simultaneously [10]. We found that the Ref. [18]. For simplicity we considered the solution of diffusion
interaction of draw points breaks the symmetry of the movement equation in the case of a narrow aperture, see Ref. [11]. Then, the
zone, increasing its height within the interaction region. However, vertical velocity is given by
for this study the draw points operate in alternate mode, meaning !
dz Q0 ½xx0 2
that the material is extracted by a single draw point at a time. uz ¼ ¼  pffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp  , ð1Þ
Here we present a theoretical study based on a modified 2D dt 4pDP z 4DP z
kinematic model that includes a dilation front and considers two and the horizontal velocity is obtained from ux ¼ DP @uz =@x, it
draw points operating in alternate mode in order to simulate reads
those found in a sub-level caving mine. It is important to mention !
that the results shown below can be generalized to three dimen- dx Q0 ½xx0 2 xx0 
ux ¼ ¼  pffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp  , ð2Þ
sions and that they can be derived for flows arising in more dt 4pDP z 4DP z 2z
complex geometries with the help of numerical methods. Here we
limit the study to present the simplest case which illustrates the where DP is the diffusion coefficient, measured in meters, which
general rules. In Section 2 we show the basics of 2D kinematic magnitude is of the order of the size of a typical rock fragment, x0
model and dilation front. Section 3 describes the interaction of is the location of the center of the draw point and Q0 is the
two draw points operating in alternate mode. In Section 4 we sectional flow rate, i.e. the section of removed material per unit of
briefly discuss the effects introduced by models that reproduce time, measured in m2/s. In three dimensional systems it corre-
the isolated movement zone more accurately and in Section 5, we sponds to the flow rate or volume of removed material per unit of
summarize the results. time, measured in m3/s. The system of coordinates is shown in
Fig. 1. Eqs. (1) and (2) represent the velocity field of steady flow of
grains where dilation or local volume changes are negligible.
2. Kinematic model, dilation front and isolated movement The position of grains or rock fragments in time is determined
zone by the pathlines or trajectories. These pathlines are obtained by
solving the differential Eqs. (1) and (2) simultaneously. In general,
In recent articles we have shown that simple granular models find the analytical solution to these equations is not possible
such as 2D kinematic model could be considered as a starting when velocity components depend on time. However, as in our
point to study more complex situations like that observed in case, in a steady state the velocity components are independent of
underground mining when sublevel caving technique is applied, time and therefore the pathlines and streamlines coincide, then
see Refs. [7–10]. Although, these models do not consider effects the trajectories are determined by introducing the streamlines
like humidity or grains shape and roughness the resulting predic- equation into Eqs. (1) and (2) and integrating in time. The stream-
tions are consistent with experimental measurements on scale lines are calculated from the tangent condition dz=dx ¼ uz =ux , it
models, see Ref. [10]. It is important to note that kinematic model reads
is successful in describing the velocity distribution, represented dz z
¼ : ð3Þ
by the streamlines, in a rectangular hopper under stationary dx xx0
conditions, when the material is in a loose packing state [11].
The solution to this equation is given by
However, the agreement becomes poor if material is in a nearly
compact state [12,13], because dilation takes place when densely z ¼ cðxx0 Þ2 : ð4Þ
packed granulate starts to flow. These finding have been recently
This equation which represents the trajectory of rock fragments in a
confirmed by experimental results [14,15] showing that stream-
steady flow created by the extraction of material from a narrow draw
lines are correctly predicted by kinematic model in loose packing
regime. Different approaches have been developed to modeling
this type of granular flows, Mullins [16] modeled the granular
flow as the upwards diffusion of voids and Litwiniszyn [17]
considered the probability of motion of grains as a random
process. Following similar ideas Nedderman and Tuzun [11]
developed a model in which particles located immediately above
the extraction orifice fall down, letting the particles in the upper
layers slide into the vacant space. This model is based on the
study of the trajectories of rock fragments, that is the geometrical
study of the motion of rock fragments, and therefore does not
require to make reference to the forces involved in the process
in particular the gravity, i.e. is a pure kinematic model and is
currently referred to as the kinematic model. If the sliding of
particles in the upper layers is viewed as a quasi-stationary
process, then it is expected that the horizontal velocity depends
on the gradient of the vertical velocity. According to Nedderman
and Tuzun [11] the simplest relationship that can be considered
between vertical and horizontal velocities in a 2D system is
ux ¼ DP @uz =@x, where ux and uz are the horizontal and vertical
velocities and DP is a parameter representing the lateral mobility
of grains called diffusion coefficient which has units of distance.
Assuming constant density throughout the system and replacing
Fig. 1. A schematic illustration that shows the dilation front moving upwards with
the last expression into the stationary mass conservation equa- velocity vF and the IMZ at two different times, t1 and t2. Behind the front, the rocks
tion @ux =@x þ @uz =@z ¼ 0 it is found a diffusion-like equation for move downwards with velocity vP while the rocks located ahead of the front
the vertical velocity, @uz =@z ¼ DP @2 uz =@x2 . A complete study of the remain static. At rightmost is shown the coordinate system.
F. Vivanco, F. Melo / International Journal of Rock Mechanics & Mining Sciences 60 (2013) 381–388 383

point. In this case the trajectories are parabolas centered at the defined by the dilation front, we integrate any of the two equa-
middle point of the draw point, x0. The arbitrary constant c deter- tions (7) or (8). For simplicity we choose Eq. (8), then replacing
mines a specific trajectory or similarly identifies a particular rock the streamlines equation, Eq. (4), into Eq. (8) and integrating
fragment during the flow. we obtained
In a recent publications [8,9], we have shown that neglecting !
4 pffiffiffiffiffiffiffiffiffi 3=2 ½xx0 2 r
the local volume changes in kinematic model results in a pDP z exp ¼ Q t, ð9Þ
misprediction of motion of tracer lines, measured experimentally
3 4DP z Dr 0
in scale models. This is a consequence of the packing decrease or
where t is the operating time of the draw point and x and z define
loosening that is not included in the original formulation of the
a point on the boundary of the IMZ. The set of all points (x,z)
kinematic model. We showed that at first approximation it can be
that are solution of Eq. (9) defines the boundary of the IMZ or
corrected by introducing a dilation front propagating upwards
dilation front. We observe that increasing Q 0 t and r=Dr produce
from the draw point. This dilation front is referred to a moving
more elongated IMZ boundary while increasing DP creates more
surface that propagates the increase in local volume or equiva-
rounded shapes. It is important to mention that the boundary of
lently packing decrease or loosening. It is identified with the
the IMZ obtained with modified kinematic model differs from
boundary of the Isolated Movement Zone (IMZ) mentioned in the
the one observed in gravel scale models. However, as mentioned
contemporary literature, see for instance Refs. [19–21]. According
by Nedderman and Tuzun, the kinematic model is suitable to
to the literature the IMZ represents the volume of mobilized and
reproduce the experimental results obtained in the flow of
dilated material. The dilation front separates the IMZ from the
granular material in a flat-bottomed hopper, see Ref. [11]. Despite
surrounding stationary material called stagnant zone. This dila-
the differences between the modified kinematics model and the
tion front is initially located in the vicinity of the draw point
experimental observations in gravel scale models, we consider
and starts to move upwards when material is removed from this
the modified kinematic model to as a first approximation in
point. As shown in previous work [8], the IMZ is significantly
determining the movement zone created by two draw points
affected by the packing changes because the sliding capacity of
operating alternately.
rock fragments is strongly dependent of the available empty
In highly compacted systems, the rock fragments slide down
space. Assuming a steady state flow, the mass balance in the
after available space is created underneath them due to the
volume element defined by two consecutive positions of the
propagation of the dilation front. Therefore, part of the removed
dilation front separated a time interval dt, as shown in Fig. 1, at
material from the draw point contributes to dilation front reaches
the point (x,z) located on the front can be written as
the rock fragment’s position and the rest of the extracted material
J  dA dt ¼ Dr dsF  dA, ð5Þ to the downwards motion of the rock fragment. According to
2
Appendix A, we have the following expressions for the horizontal
where J ¼ rvP is the flow of rocks in kg/m s, vP is velocity of rocks
and vertical displacement of a rock fragment, initially located at
in m/s, dsF is the displacement vector of the dilation front in
ðxS ,zS Þ, when Q 0 T material is removed from the draw point
meters which is parallel to the rock fragments velocity, Dr ¼
2 31=3
r0 r is the density change introduced by the rock motion in   !
Kg/m3, r0 is the density outside the IMZ, r is the density inside 6 Dr 3 1 ½xS x0 2 7
x ¼ x0 þ 4 1 þ  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp  Q 0 T 5 ðxS x0 Þ,
the IMZ and dA is the area element whose normal is parallel to the r 4 4pD z3 4DP zS
P S
front propagation vector dsF . In deriving Eq. (5) we assumed that
ð10Þ
the mass, or number of rock fragments, inside IMZ remains
constant for an infinitesimal displacement of the dilation front, and
such that the local volume changes can be identified with the "  ! #2=3
local density changes. Then, the expression for the velocity of the Dr
3=2 3 1 ðxS x0 Þ2
z¼ 1þ z  pffiffiffiffiffiffiffiffiffi exp  Q 0T : ð11Þ
dilation front reads r S 4 pDP 4DP zS
dsF r In the particular case where the rock fragment is initially located
¼ vF ¼  v : ð6Þ
dt Dr P in the axis of symmetry of the draw point, these expressions
According to this equation, the dilation front moves upwards simplify to x ¼ x0 and
faster than the rock fragments downwards. Its velocity depends   2=3
Dr 3=2 3 1
on the local density change such that for a very small change in z¼ 1þ zS  pffiffiffiffiffiffiffiffiffi Q 0 T : ð12Þ
r 4 pDP
density the front extends along the entire system almost instan-
taneously, recovering the behavior described by the original Taking into account that the rock fragments move towards the
kinematic model. However, the dilation front has a finite exten- draw point, we observe from Eqs. (10) and (11) that large local
sion for sufficiently large density contrast and it is initially located density changes increase the amount of material removed in
in the vicinity of the draw point. By introducing the components order to rock fragments slide down and create a steady flow. Thus,
of rock fragments velocity, Eqs. (1) and (2), into Eq. (6) we the final position of the rock fragment depends non-linearly
obtained the components of the dilation front velocity on the total removed volume, the diffusion coefficient, the local
! density change and the initial position of the rock fragments, even
r Q0 ½xx0 2 xx0  in the simpler case where the fragment is located along the
vx ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp  , ð7Þ
Dr 4pDP z 4DP z 2z symmetry axis of the draw point. It is worthy to mention that all
the results obtained with the modified kinematic model can be
and
generalized to more complex models like the plasticity model [7].
!
r Q0 ½xx0 2 As mentioned earlier, one of the main concerns of under-
vz ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp  : ð8Þ ground mining is controlling of dilution. This involves the thor-
Dr 4pDP z 4DP z
ough understanding of the rock flow in order to predict dilution
Note that the streamlines equation corresponding to the velocity relocation. Hence the importance of highlighting the result
field of the dilation front is the same as for the rock fragments, see expressed by Eqs. (10) and (11) which allows, in first approxima-
Eq. (4). On the other hand, to determine the boundary of the IMZ, tion, estimate the final position of rock fragments, either ore or
384 F. Vivanco, F. Melo / International Journal of Rock Mechanics & Mining Sciences 60 (2013) 381–388

dilution, through simple expressions. It is important to mention those created by a single isolated draw point, see left panel of
that in sublevel caving mines the material is removed discretely, Fig. 2. If the removed material from draw point b is such that
therefore the continuous time may be replaced by a number of Mb ¼ M nb , the boundaries of IMZb and IMZa are tangent at the point
extractions, as will be shown later. In the next section, it will be A, defined by ðxn ,zn Þ. Then, there exists a streamline defined by the
determined the movement zone created by the extraction of points xb and A that intersects the top of IMZa boundary at the
material from two neighboring draw points, operating alternately, point B, see center of Fig. 2. The region affected by the interaction
and it will be shown how this interaction modifies the tracer lines is extended by additional extractions of material from draw point
displacement. b, as shown in the right panel of Fig. 2. We expected that the
boundary of the IMZa located between limit points B and B þ
be distorted due to the removal of material. To calculate this
3. The movement zone of two neighboring draw points distortion it is required to determine the new positions of the
operating in alternate extraction points on the IMZa boundary following the streamlines defined by
their initial positions. If ðxS ,zS Þ are the generic coordinates of the
In the next paragraphs, we focus our attention in determining
the boundary of the movement zone created by two draw points
operating alternately, where the same amount of material is
extracted from each draw point operating one at a time. We
considered a two dimensional system in the plane (x,z) consisting
of two neighboring draw points with infinitely small apertures,
located at ða,0Þ and ðb,0Þ and to hereafter referred to as draw
points a and b. Let Q a Ma and Q b M b be the amount of material
extracted from draw points a and b after Ma and Mb extractions.
Since these draw points are operating in alternate mode, the
material is removed first from draw point a while b remains
closed creating an IMZa boundary centered at x ¼ xa . Then, draw
point a is closed and material begins to remove from draw point b
creating an IMZb boundary centered at x ¼ xb . If draw points are
far enough apart we obtain two IMZ such as those created by a
single isolated draw point. However, if draw points are close to
each other so that individual IMZ overlap, the two IMZ interact
changing the shape of IMZa. This interaction begins after a certain
number of extractions Mnb from draw point b, and at point ðxn ,zn Þ
both IMZ are tangent to each other. According to Appendix B, this
point is given by
x þ x  x x qffiffiffiffiffiffiffiffiffiffiffiffiffi  Fig. 2. (Left) Two IMZ boundaries, IMZa and IMZb, created by Ma extraction from
a b a
xn ¼ þ b 9þ C 2 3 , ð13Þ draw point a and M b o M nb from draw point b. (Center) When M b ¼ Mnb , both
2 2 IMZ meet at point A of coordinates ðxn ,yn Þ. Subsequent extractions create a
perturbation on top of IMZa located in the neighborhood of point B. (Right) When
and M b 4 M nb , the IMZ boundaries intersect at two points, A and A þ , the region on
qffiffiffiffiffiffiffiffiffiffiffiffiffi 2
1 xb xa 2 1 top of IMZa boundary, delimited by points B and Bþ , is affected by the removal of
zn ¼ 9 þ C 2 3 : ð14Þ material from draw point b.
6DP 2 C2
And the number of extractions from draw point b required to
initiate the interaction is given by
 
Q 6un
M nb ¼ a M a exp  , ð15Þ
Qb 1un
n
wherepuffiffiffiffiffiffiffiffiffiffiffiffiffi
is the solution of Eq. (B.8) and C is determined from
u  ð 9 þ C 2 3Þ=C. As shown in Appendix B, the number of
n

extractions from draw point b needed to begin the interaction,


and the point ðxn ,zn Þ depend on the diffusion coefficient DP, the
density variations Dr=r, the volume extraction rates Qa and Qb,
the initial number of extractions Ma from draw point a and the
separation distance between the center of draw points xb xa .
Note that, the interaction is primarily affected by the total volume
of removed material and the separation distance between draw
points. Eq. (15) allows to predict the time lapse that an active
draw point can operate without altering the surrounding areas,
previously dilated by operating neighboring draw points, pre-
venting dilution contaminates the draw point. From the point of
view of mining operation, this prediction might be useful in
programming the extraction of material from neighboring draw
points in sublevel caving underground mines, increasing the ore Fig. 3. Two interacting IMZ boundaries centered at xa and xb, the gray region R is
recovery and reducing dilution. the dilated zone of IMZa because of extraction of material from draw point b. Effect
of the amount of material extracted or number of extractions, (left) slightly above
As mentioned above, when the amount of material removed the minimum required to create an interaction, (center) about half of the total
from draw point b is smaller than the minimum required to initiate initially extracted material from draw point a, and (right) about two-thirds of the
the interaction, M b oM nb , it is observed two IMZ boundaries such as total initially extracted material from draw point a.
F. Vivanco, F. Melo / International Journal of Rock Mechanics & Mining Sciences 60 (2013) 381–388 385

points located between B and B þ and Q b Mb is the total volume and


of removed material from draw point b, then the new positions on " ! #1=3
the distorted boundary are given by 3 r 1 ðxb xS Þ1
z¼ z3S þ pffiffiffiffiffiffiffiffiffi exp  n
ðQ b M b Q b M b Þ : ð17Þ
4 Dr pDP 4DP zS
2 31=3
!
6 3 r 1 ðxb xS Þ2 The region on the boundary of IMZa distorted by the extraction
n 7
x ¼ xb ðxb xS Þ41 þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp  ðQ b M b Q b M b Þ5 , of material from draw point b is shown in Fig. 3. When the
4 Dr pD z3 4DP zS
P S
intersection between the IMZa and IMZb boundaries is small the
ð16Þ top of IMZa is slightly perturbed, as shown in left panel of Fig. 3.

Fig. 4. Effect on tracer lines due to the alternate extraction. (Top) Non-interacting IMZ boundaries, the tracer lines are deflected inside the zone delimited by each IMZ.
(Bottom) The interaction affects zones located outside the IMZ created by draw point b increasing deflection of tracer lines previously deviated by extractions from draw
point a. Amount of material extracted from draw point b; (left panels) 20% and (right panels) 80% of the total amount extracted from draw point a.
386 F. Vivanco, F. Melo / International Journal of Rock Mechanics & Mining Sciences 60 (2013) 381–388

By increasing the volume of removed material from draw point b, conditioned sub-level caving mines with small fragments of rocks
the top of the IMZa boundary is modified moving upwards and to and relatively low dispersion in fragment size. Thus, the results
the right of its original position, as can be observed from center obtained here can be considered as a first approximation to the
and right panels of Fig. 3. As a consequence of the interaction, the optimization of ore recovery in sub-level caving mines. The
IMZa boundary is modified moving from its original position even results presented can be extrapolated to 3D systems and general-
though no new material is extracted from draw point a. The ized to other type of flows described by more complex models
disturbance observed on the IMZa may contaminate the neighbor- than a kinematic model. To our knowledge, there are no previous
ing draw points affecting the planning of the future extractions. theoretical attempts to account for interacting movement zones.
From the point of view of underground mining, the interaction Lastly, experiments documented in existing literature have
between neighboring draw points might contribute negatively by chosen to focus on the tracer lines deflection created by the
redistributing unexpected dilution nearby the operating draw simultaneous extraction instead of alternate extraction of mate-
points. Conversely, this might be beneficial because the interac- rial from neighboring draw points.
tion could relocate ore closer to the active draw points and
facilitate its extraction. In the experiments the effects of the
interaction of draw points are visualized through deflection of 5. Conclusions
tracer lines. As mentioned in Section 2, the rock fragments,
forming part of the tracer lines, will start to slide down after The effects on a IMZ boundary due to the extraction of
the dilation front passes by their position. Therefore, the tracer material from a neighboring draw point have been determined
line positions are given by Eqs. (10) and (11). The resulting through the application of modified 2D kinematic model by
displacement of tracer lines along with the IMZ boundaries including a dilation front. We found that strength of the interac-
created by a couple of neighboring draw points operating in tion depends on the volume of removed material, the lateral
alternate mode are shown in Fig. 4. An initial amount of material, mobility of the rock fragments represented by the diffusion
Q a M a , is extracted from left draw point creating an isolated IMZa coefficient and mainly on the separation distance between draw
boundary and deflecting the zone of tracer lines located inside it. points. If the draw points are located far enough to avoid the
After material is removed from right draw point at Qb ratio, the overlap of the individual IMZ, we obtained two IMZ as the one
IMZb boundary and deflection zone of tracer lines are created, created by a single isolated draw point. On the other hand, if these
as shown in left panels of Fig. 4. Two different situations are draw points are close enough such that the individual IMZ over-
observed depending on the separation distance between draw lap, then the interaction mainly depends on the volume of
points. First, the separation distance is large enough such that the removed material, or equivalently on the number of extractions
IMZ boundaries do not overlap, it is observed two IMZ boundaries at constant mass rate extraction. There is a critical value for the
similar to the one created by an isolated draw point. Conse- number of extractions, from the second operating draw point,
quently, the tracer lines deflection is only localized inside of each where both IMZ share a common point. At this point pass a
IMZ boundary, see top panels of Fig. 4. Second, if draw points are streamline connecting the center of the operating draw point
close enough, the IMZ boundaries overlap. In such case, the model with a point on top of the previously created IMZ boundary. Any
predicts that shape and position of limiting surface located above increment in the number of extractions beyond its critical value
the left draw point change, showing a deflection of tracer lines creates perturbations on this top surface, modifying its shape and
outside of the IMZb created by the right draw point, compare right position. The initial shared point is transformed into two points
panels of Fig. 4. As a consequence, material located out of range of delimiting the region affected by removing of material from the
right draw point might be extracted because of the interaction. operating draw point. This region is located out of the range
The effects on the shape of the IMZa modifies the initially dilated covered by the movement zone created by a single isolated draw
region increasing the number of rock fragments allowed to move. point. This effect is observed in the deflection of tracer lines
This might have positive or negative consequences depending on located inside the previously created IMZ, as shown in Fig. 4. From
the sort of material affected by the interaction; ore or dilution. the point of view of the underground mining based on sublevel
Therefore, the interaction of neighboring draw points operating in caving, the modification of the boundary of an IMZ due to the
alternate mode should be taken into account to improve predic- extractions of material from a neighbor draw point could have
tions on relocation of ore and dilution. significant negative effects, because dilution initially located far
from an operating draw point might be carried to the vicinity of
its opening and may eventually contaminate it. Conversely, the
4. Discussion interaction might be beneficial by relocating ore closer to the
region of influence of a draw point which eventually might be
In previous sections, it has been determined the modifications recovered. Finally, neglecting the effects created by the interac-
of the IMZ that results from extraction of material from a tion of neighboring draw points might lead to wrong predictions
neighboring draw point and the effect on deflection of tracers of rock fragments motion, with negative impact in preventing
lines. The application of these results are limited to a stationary dilution.
systems where the granular material is in a loose packing state.
The calculations are based on a kinematic model modified with a
Acknowledgments
dilation front because it has been proved that streamlines found
in stationary loose packing granular systems are correctly repro-
duced by the kinematic model [11]. However, the shape of the This work was supported by program ANR-Conicyt ANR-011.
IMZ obtained with the kinematic model with a dilation front F. Vivanco thanks the support from Conicyt through Proyecto
might deviate from the observed in experiments in scale models. PBCT PSD-54.
More complex models that more accurately reproduce the shape
of the IMZ can be introduced in the calculations and it is expected Appendix A. Displacement of rock fragments
that these models will modify the results quantitatively but not
qualitatively. Based on these requirements for the kinematic Let ðxS ,zS Þ and (x,z) be the initial and final positions of a rock
model it is expected that the results might be applied to pre- fragment after a time T. If T S o T is the time required by the
F. Vivanco, F. Melo / International Journal of Rock Mechanics & Mining Sciences 60 (2013) 381–388 387

dilation front to reach the initial position of the rock fragment, with
ðxS ,zS Þ, then TT S is the time that the rock fragments is moving 3 Q ðÞ M ðÞ r
due to the extraction of material from the draw point. This time is AðÞ  pffiffiffiffiffiffiffiffiffi , ðB:2Þ
4 pDP Dr
determined by integrating Eqs. (7) or (8). Replacing the stream-
lines, Eq. (4) into Eq. (7), and integrating between 0 and TS, where ðÞ ¼ a,b labels the draw points and AðÞ is proportional
Z TS   Z xS to the volume of the extracted material from each draw point.
Dr pffiffiffiffiffiffiffiffiffiffiffiffiffi3ffi 1
dt ¼ 4 pDP c exp ðxx0 Þ2 dx, ðA:1Þ Differentiating Eq. (B.1) with respect to x and assuming that DP
0 r 4DP c 0
and r=Dr are constant during the extraction process, we obtain
we obtain for the time TS, the tangent as
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi  
4 Dr pDP c3 1 !1
TS ¼ exp ðxS x0 Þ3 : ðA:2Þ dz xxðÞ 3 ½xxðÞ 2
3 r Q0 4DP c ¼  : ðB:3Þ
dx 2DP 2 4DP z
The value of c is determined by the initial position of the rock
fragment, xS, and the center of the draw point, x0, through Now, to determine M nb and ðxn ,zn Þ we use the fact that both IMZ are
c ¼ zS =ðxS x0 Þ2 . Then replacing into Eq. (A.2) we obtain tangent at this point. Then, we have the following conditions:
pffiffiffiffiffiffiffiffiffi !
4 Dr pDP ½xS x0 2 3=2 z ¼ 4DP C½ðxxa Þ2 ðxxb Þ2 , ðB:4Þ
TS ¼ exp zS : ðA:3Þ
3 r Q0 4DP zS
and
Similarly, replacing the streamlines equation into Eq. (2) and 1
integrating between TS and T z¼ ðxxa Þðxxb Þ, ðB:5Þ
6DP
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi  Z x Z T
1
4 pDP c3 exp ðxx0 Þ2 dx ¼  dt, ðA:4Þ where C  lnðAa =Ab Þ. Replacing Eq. (B.5) into Eq. (B.4) we obtained
4DP c xS TS
the value of xn as follows:
we obtain the final horizontal position, x, of the rock fragment, x þ x  x x qffiffiffiffiffiffiffiffiffiffiffiffiffi 
a b a
2 31=3 xn ¼ þ b 9 þ C 2 3 : ðB:6Þ
  ! 2 2
2
6 D r 3 1 ½xS x 0  7
x ¼ x0 þ 4 1 þ  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp  Q 0 T 5 ðxS x0 Þ,
r 4 4pD z3 4DP zS Then, introducing this value into Eq. (B.5) we calculated the height
P S
at the interaction point which reads
ðA:5Þ qffiffiffiffiffiffiffiffiffiffiffiffiffi 2
1 xb xa 2 1
where TS and c were replaced by the right hand side of expression zn ¼ 2
9 þC 2 3 : ðB:7Þ
6DP 2 C
(A.3) and zS =ðxS x0 Þ2 . We determined the vertical displacement of
the rock fragment using a similar method. This displacement is These two last expressions define the intersection point of both
given by IMZ in terms of C which in turns is a function of the unknown
"  ! #2=3 number of extractions from draw point b. To determine this
Dr 3=2 3 1 ðxS x0 Þ2
z¼ 1þ zS  pffiffiffiffiffiffiffiffiffi exp  Q 0T : ðA:6Þ number of extractions we rewrite Eqs. (B.6) and (B.7) introducing
pffiffiffiffiffiffiffiffiffiffiffiffiffi
r 4 pDP 4DP zS
the variable u  ð 9 þC 2 3Þ=C and replace the values of xn and zn
Eqs. (A.5) and (A.6) determine the total displacement of a rock into Eq. (B.1) with ðÞ ¼ a, we obtain an equation to determine the
fragment after Q 0 T volume of material is removed from the draw value of u, it reads
point. In the particular case where the initial position of the rock    2
1 þu 2
fragment is located on the axis of symmetry, xS ¼ x0 , the total ð1u2 Þ exp ¼ 6DP A2=3 a : ðB:8Þ
1u xb xa
displacement of the fragment is determined by x ¼ x0 and
  2=3 Let un be the solution of Eq. (B.8), then writing C in terms of u and
Dr 3=2 3 1
z ¼ 1þ zS  pffiffiffiffiffiffiffiffiffi Q 0 T , ðA:7Þ solving C ¼ lnðAa =Ab Þ for the number of extractions from draw
r 4 pDP
point b, we obtain an expression for the minimum number of
and the time required by dilation front to reach the initial extractions from draw point b in order to both IMZ intersect at one
position of the fragment simplifies to point, it is given by
pffiffiffiffiffiffiffiffiffi
 

4 Dr pDP 3=2
Q a 6un

TS ¼ z : ðA:8Þ Mnb ¼


Q M exp 
, ðB:9Þ
1un

a
3 r Q0 S b

where 9  9 represents the integer part of the argument. If the


Appendix B. Intersection of two IMZ number of extractions exceeds the critical value, i.e. M b 4M nb , then
both IMZ intersect at two points. To determine these points we
Let a and b be two neighboring draw points with infinitely intersect the curves that define each IMZ. The curve for draw point
small apertures, located at ða,0Þ and ðb,0Þ in the plane (x,z), where a is given by
!
b 4a. Then, let assume that Q a M a and Q b Mb are the amount of ½xxa 2
material extracted from draw points a and b after Ma and Mb z3=2 exp ¼ Aa , ðB:10Þ
4DP z
extractions. If the draw points are close enough so that both
isolated IMZ overlap, then there is a point, ðxn ,zn Þ, corresponding
and for draw point b it reads
to a certain number of extractions from the draw point b, M nb , !
where these IMZ are tangent to each other. The IMZ created by ½xxb 2
z3=2 exp ¼ Ab : ðB:11Þ
each draw point are given by Eq. (9), rewriting we have 4DP z
!
½xxðÞ 2
z3=2 exp ¼ AðÞ , ðB:1Þ The values of z at the intersection points are determined by solving
4DP z
Eqs. (B.10) and (B.11) for x and combining the resulting equations,
388 F. Vivanco, F. Melo / International Journal of Rock Mechanics & Mining Sciences 60 (2013) 381–388

it reads References
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 
3 3 [1] McNearny R, Abel J. Large-scale two-dimensional block-caving model tests.
4DP z ln Ab  ln z þ 4DP z ln Aa  ln z ðAb Aa Þ ¼ 0:
2 2 Int J Rock Mech Min Sci Geomech Abstr 1993;30:93–109.
[2] Bergmark JE. The calculation of drift spacing and ring burben for sublevel
ðB:12Þ
caving, LKAB Memo #RU 76-16; 1975.
This equation must be solved numerically. Assumed that zn [3] Janelid I, Kvapil R. Sublevel caving. Int J Rock Mech Min Sci 1966;3:129–53.
[4] Kuchta ME. A revised form of the Bergmark–Roos equation for describing the
represents the solutions of Eq. (B.12), then the values of xn at the gravity flow of broken rock. Miner Resour Eng 2002;11:349–60.
intersection points are given by [5] Kvapil R. Gravity flow of granular materials in hoppers and bins. Int J Rock
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ffi Mech Min Sci 1965;2:35–41.
n 3 [6] Rustan A. Gravity flow of broken rock—what is known and unknown. In:
x ¼ xa þ 4DP z ln Aa  ln z :
n n ðB:13Þ Conference proceeding of MassMin 2000: the Australasian Institute of Mining
2
and Metallurgy; 2000. p. 557–67.
Since the modified kinematic model assumes that the rock frag- [7] Melo F, Vivanco F, Fuentes C, Apablaza V. On drawbody shapes: from Bergmark–
Roos to kinematic models. Int J Rock Mech Min Sci 2007;44:77–86.
ments move along the streamlines, it is expected that the dilation [8] Melo F, Vivanco F, Fuentes C, Apablaza V. Kinematic model for quasi static
propagates, to the top of the IMZ of draw point a, following the granular displacements in block caving: dilatancy effect on drawbody shapes.
streamlines. The points ðxn ,zn Þ, solutions of Eqs. (B.12) and (B.13), Int J Rock Mech Min Sci 2008;45:248–59.
and the center of the draw point b, x ¼ xb , define the following [9] Melo F, Vivanco F, Fuentes C. Calculated isolated extracted and movement
zones compared to scaled models for block caving. Int J Rock Mech Min Sci
streamlines: 2009;46:731–7.
! [10] Vivanco F, Watt T, Melo F. The 3D shape of the loosening zone above multiple
zn
z¼ ðxxb Þ2 : ðB:14Þ draw points in block caving through plasticity model with a dilation front. Int
ðxn xb Þ2 J Rock Mech Min Sci 2011;48:406–11.
[11] Nedderman RM, Tuzun U. A kinematic model for the flow of granular
On the other hand, the streamlines defined by Eq. (B.14) intersect material. Powder Technol 1979;22:243–53.
the top of the IMZ created by the draw point a. The intersection [12] Bransby PL, Blair-Fish PM. Initial deformations during mass flow from a
bunker: observation and idealizations. Powder Technol 1975;11:273–88.
points are determined solving the equation [13] Levinson M, Shmutter BW, Resnick. Displacement and velocity fields in
!
hoppers. Powder Technol 1977;16:29–43.
ðxb xn Þ2 ðxxa Þ2 ðx xn Þ3
ðxb xÞ3 exp 2
¼ Aa b 3=2 , ðB:15Þ [14] Samadani A, Pradham A, Kudrolli A. Size segregation of granular matter in
4DP z ðxb xÞ
n
zn silo discharges. Phys Rev E 1999;60:7203–9.
[15] Caram H, Hong DC. Random-walk approach to granular flows. Phys Rev Lett
where xn and zn represent the solutions of Eqs. (B.12) and (B.13). 1991;67:828–31.
The IMZ independently created by each draw point are mod- [16] Mullins WW. Experimental evidence for the stochastic theory of particle flow
under gravity. Powder Technol 1976;9:29–37.
ified by the interaction between the draw points. Following
[17] Litwiniszyn J. The model of a random walk of particles adopted to researches
Appendix A, we obtain the new positions of the modified IMZ. on problems of mechanics of granular media. Bull Acad Pol Sci 1963;11:
Then, for each point ðxS ,zS Þ on the original IMZ, the resulting new 61–70.
positions are given by [18] Crank J. The mathematics of diffusion. 2nd ed. Bristol, England: J.W. Arrow-
smith LTD.; 1975.
2 31=3
! [19] Castro R, Trueman R, Halim A. A study of isolated draw zones in block caving
6 3 r 1 ðxb xS Þ2 n 7 mines by means of a large 3D physical model. Int J Rock Mech Min Sci
x ¼ xb ðxb xS Þ41 þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp  ðQ b M b Q b M b Þ5 ,
4 Dr pD z3 4DP zS 2007;44:860–70.
P S [20] Trueman R, Castro R, Halim A. Study of multiple draw-zone interaction in
block caving mines by means of a large 3D physical model. Int J Rock Mech
ðB:16Þ
Min Sci 2008;45:1044–51.
and [21] Pierce M. A model for gravity flow of fragmented rock in block caving mines.
" ! #1=3 PhD thesis. Sustainable Minerals Institute, The University of Queensland,
3 r 1 ðx x Þ1 Queensland; 2010.
z ¼ z3S þ pffiffiffiffiffiffiffiffiffi exp  b S ðQ b M b Q b M nb Þ , ðB:17Þ
4 Dr pDP 4DP zS

where Q b Mnb and Q b M b 4Q b M nb are the volume of material


removed from draw point b required to initiate the interaction
and the total volume of removed material from draw point b.