Properties of pumped tailings fill

by G. E. BLlGHT*, Pr. Eng., D.Se. (Eng.) (Land.), Ph.D. (Land.), MASCE, MAAPT, F(SA)ICE (Visitor)
SYNOPSIS
The mining industry is making increasing use of pumped tailings fills. The properties that affect the performance
of tailings fills are described by means of typical examples. They are the particle-size grading for maximum density,
the water content for pumpability, settlement after placing, compression and drainage under stress, shear strength,
and resistance to shock loading.
SAMEVATTING
Die gebruik van slyk vir vul doeleindes in die mynbedryf nee m toe in gewildheid. Die eienskappe wat die werk-
verrigting van slyk vu! belnvloed word beskryf aan die hand van tipiese voorbeelde. Die eienskappe is die partikel-
groote-gradering vir maksimum digtheid, die watergehalte vir pompbaarheid, versakking na plasing, verdigting en
.
dreinering onder spanning, skuifsterkte en weerstand teen skokbelasting.

Introduction When the relative merits of a number of possible fill

materials are considered, an attempt should be made to
There is a growing world-wide interest in the mining
industry in the use of underground fills of tailings maximize the above factors, all of which, as it happens,
material. These may be cemented1 or uncemented2; and are compatible with one another.
may be placed mechanically, pneumatically3, or hy- Fills under Stress
draulically1, 2. The purposes of these fills are numerous.
When fills are subjected to stress, their behaviour is
For example, fills are used to
govem.ed by the principle of effective stress, which
(i) improve the stability of pillars2,
applie3 to all particulate materials7t. This principle
(ii) provide access to assist in the mining of widc ore-
states that changes in volume or in strength of a particu-
bodies2, 4,
late material are governed by changes in effective stress,
(iii) reduce surface subsidence in the case of shallow
and that volume and strength can change only if the
mining, or stabilize the surface in the case of lon£-
effective stress in the material changes. The effective
abandoned shallow workings,
stress is defined by the equation
(iv) act as structural support to stopes in deep mines1, 2.
a'=a-u, (I)
In addition to these purely functional uses, the dis-
in which a is the total or applied stress in the material,
posal of tailings in the form of underground fill reduces
u is the pressure in the water contained by the material
(though only slightly) the pressures on the surface
6 by (the pore-water pressure), and a' is the effective stress.
environment5,
Changes in the volume of a fill are governed by the
(a) releasing area at the surface for more productive
equation
use than the disposal of tailings, and
(b) reducing problems of air, water, and visual pollution, 6Ev=-C6a', (2)
in which 6 Ev is the change of volumetric strain corres-
both in the short and the long term.
ponding to a change in effective stress of 6a', and C is
Properties of Pumped Tailings Fill the compressibility of the fill material. Changes of shear
It appears that the following are the properties re- strength are controlled by the equation
quired of a successful pumped tailings fill. 6T=tan<p'6a', (3)
(I) The tailings should become pumpable at the lowest where 6T is the change in shear strength corresponding
possible water content. In South Africa's semi-arid to a change in the effective stress of 6a', and tan <P'is
to arid climatic conditions, water supply is often a the angle of shearing resistance of the fill material.
major problem. In addition, much of the water Integrating equation (3) yields the shear-strength
used to place a fill eventually drains out of it and equation
must be handled within the mine. However, the T=c'+tan<p'a', (3a)
most important reason for seeking pumpability in which c' is the shear strength when a' =0.
at a low water content is that, the lower the water Changes of volume, shear strength, and water content
content at which the slurry is placed, the less are inextricably linked in that the shear strength cannot
compressible the fill subsequently. change without a change of volume, the volume cannot
(2) During placing, the fill should settle into as dense change unless water is expelled from the fill, and the
a state as possible and after that should be as compressibility depends on the shear strength (as com-
incompressible as possible under applied stress. pression usually involves internal shearing).
(3) The fill should be capable of resisting the effects of The compressibility C and the angle of shearing
shock loading produced by blasting or seismic resistance <p' depend on the density at which the fill is
events. placed and increase with increasing placed density. As
the dry density Yd of a saturated fill is related to the
*Department of Civil Engineering, University of the Witwaters- water content w by
rand, Johannesburg. Gsyw
tThe reader is referred to this work for further details of the well- Yd (4)
accepted concepts of soil meChaniQs discussed in this section. l+wGs

446 OCTOBER 197~ ,JQI,IRNAL, QF TH~ $QI,JTH AfRICAN INSTITUTE OF MINING AND METALLURGY
(in which Gs is the particle specific mass of the fill and of application of a stress increment, and z is a distance
y w the density of water), low water contents are associ- related to the drainage path or distance through which
ated with high dry ,densities and vice versa. water must percolate in order to escape from the fill;
An increment of total stress applied to a fill will c v, the coefficient of consolidation, depends on both the
result in the expulsion of water via a time-dependent permeability and the compressibility of the consolidating
process of consolidation, which is governed by the material.
differential equation The interchange between pore pressure, total stress,
8u 82U and effective stress according to equation (5) is shown
Tt=cv 8Z2 . . . . . . . . . . . . . . . . (5) schematically in Fig. 1: when the increment of total
stress is imposed, it is entirely transferred to the pore
In this equation, t is the time measured from the instant water, i.e. at time t=O, Lu=La and La'=O.
The pore water immediately starts to drain from the
v,.,-
'I fill, and, as it does so, stress is transferred from the pore
L,
'" water to the solids in the fill, i.e. at ~ime t>O, Lu+ La' =
~.'
Vi La.
Finally, after the elapse of a period of time that is
-"J
0
h..
L1(J' I related to the dimensions and drainage conditions of the
fill (z) and the coefficient of consolidation cv, the entire
~ b total stress increment is carried by the solids in the
-c:ClJ "J fill, i.e. at t~oo, La'=La and Lu=O.
E:
ClJ
Gradin~s for Maximum Density
.....
I.J The placement density, and hence the compressibility
~ and shear strength, of a fill greatly depend on the grading
or particle-size distribution of the material. In what
0 follows, guidance will be given as to what constitutes a
I. ./ Time ...
satisfactory grading and what factors are more important
Fjg. I-Interchange between total stress, pore-water in ensuring a high placement density.
pressure, and effective stress during the consolidation process Grading curves can be idealized to two main types:
(i) uniform or Fuller gradings, for which the grading
100 d, curve (the relationship between the logarithm of
the particle size and the percentage finer than that
size) is an inclined straight line, and
80 (ii) gap gradings, which consist of a series of inclined
straight lines joined by horizontal size 'gaps' in
which there are no particles present.
~ 60 Uniform Gradin(Js
c::
'Ll:
~ 40
0 Most tailings materials include a relatively narrow
range of particle sizes (see, for example, Fig. 8) and can
be approximated by a straight-line grading curve. The
two main variables in a straight-line grading curve are
20 the slope (or ratio of maximum to minimum particle
size) and position (size of the largest particle present).
Fig. 2a shows a series of straight-line or Fuller gra-
0 dings in which the ratio dI/do of the largest to the smallest
0,006 2
particle size present ranges from 8 to 128 (gradings FO
to F4). Fig. 2b shows the effect of these gradings on the
settled densities of aggregates made up from two differ-
Particle Size mm ent source materials. The same compactive effort was
used for each material. Fig. 3a shows a series of parallel
straight-line gradings FO, F5, F6, and F7, and Fig. 3b
0
~ .....
105 compares the settled densities of aggregates having
~.E these gradings.
\I)
'" 100 It is clear from Figs. 2 and 3 that the slope of a uni-
'I! -:::
b~ 0 form grading curve has a relatively minor effect on the
-~CCl> settled density but that the position of the curve has a
Cl> 95 greater effect. For the same slope of grading curve, the
c
1 4 16 64 256 density may decrease by up to 8 per cent when the
d,/da maximum particle size is decreased by a factor of 4.
Gap Gradings
Fig.2-(a) Series of continuous gradings FO, FI, F2, FJ, F4 Fig. 3a also shows a series of gap gradings GO to G3 in
(b) Settled densities of two materials made up in
gradings FOto F4 which the gap bridges between the uniform gradings

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY OCTOBER 1979 447

1
I'
FO and F5. The settled densities corresponding to the ~
various gap gradings are shown in Fig. 4 (for the same 'vi F5
two basic source materials). A comparison of Fig. 4 with ~
Figs. 2b and 3b shows that a far greater range of den- Cl
sities is obtainable via a gap grading than via a uniform ~
grading. For example, the addition of about 20 per cent ~0 ~ 100
fines to a material with grading FO can cause the settled V) 0"'-
tU
density to increase by 4 per cent, whereas the density of
the fines alone would be 5 per cent less than the density .~ 95
obtainable with grading FO. ~
The experimental results presented in Figs. 2 to 4 do C)
not provide a set of rules for producing the greatest 90
possible settled density, but do provide a framework for 0 20 40 60 80 100
experimenting with a combination of available materials
to produce a satisfactory high settled density, and hence % Fines
a fill of low compressibility and high strength. It can be Fig. 4-Settled densities of two materials made up in gradings
concluded that, in order of increasing importance, the FO. GO to G3, and FS
slope of a uniform grading curve, the size of the maxi-
mum particle present, and the percentage of fines in a
Once this has been established, all the laboratory speci-
gap grading all have an effect on the attainable settled
mens for further testing are prepared at that water
demity of a pumped fill.
content.
Water Content for Pumpability The pump ability of a slurry depends on its viscosity*,
which can be measured in a variety of ways. One suitable
One of the first requirements in any programme of method is by means of a rotating-cylinder apparatus in
testing to ascertain the properties of a potential pumped which the slurry is contained in the annular space be-
tailings fill is to estimate the minimum water content tween a pair of co-axial cylinders. The outer cylinder is
at which a slurry of the tailings will become pumpable. rotated at a fixed rate, and the torque transferred to the
inner cylinder is measured to establish the viscosity
of the slurry.
100
The pumpability characteristics of the slurry can then
be investigated via the Hagen-Poiseuille equation,
80 which for viscous flow in a circular pipe or a semi-
circular open channel flowing full has the form
yiD2
60 (6)
v=32/L
'-CIJ
c:
~ 40,
0 in which v=the velocity of flow
y=the density of the slurry (mass per unit
~ volume)
i=the flow gradient or the loss of head per
20 unit length of conduit
D=the diameter of the conduit (pipe or semi-
circular channel)
F5
/L=the viscosity of the slurry.
0,006 0,02 0,6 2 The rate of transport of dry solids can be found from
Q=vAPd' (7)
in which A = cross-sectional area of conduit
Particle Size mm Pd = dry specific mass of slurry (mass of dry
solids per unit volume of slurry).
It appears that a viscosity of 100 P (poise) or 10 Pa.s
100 represents the upper limit to pumpability, but lower
0~IJ..a
viscosities, and hence higher water contents, may be
0

'-
1110
<11"" 95
required, depending on the specific application.
b~ CB The water content at which a slurry has a viscosity
iii :g 2 of 100 P can be estimated by means of an adaptation
c: CIJ
of a simple soil mechanics test, using the liquid-limit
C C} 90
apparatus7. This consists of a hemispherical cup that can
0,1 1,0 10
be raised and dropped through a fixed height onto a
d, mm
Fig.3-(a) Series of continuous gradings FO, FS, F6, F7, and *Tailings slurries are here assumed to act as Newtonian fluids.
gap gradings GO, GI, G2, G3 This is probably not strictly true (see reference 8, for example),
(b) Settled densities of two materials made up in but is a sufficiently accurate first approximation to their true
gradings FO, FS, F6, F7 iJehaviour.

448 OCTOBER 1979 JOI,JRNAI., OF THE SqUTH AFRICAN IN~TlTVTe OF MINING AND METALLURGY
 hard-rubber base by means of a rotating cam. The slurry of the problem posed by the disposal of this bleeding
is placed in the cup, and its surface is levelled and a water will clearly depend on the volume of the fill being
groove cut in it with a special tool. The water content at placed, on its rate of placing, and on the configuration
which the groove just closes with one blow of the cup on of the fill relative to available drainage in the mine. The
the base corresponds to the water content for a viscosity more bleeding that occurs, the better will the fill ulti-
of 100 P. A typical relationship between viscosity mea- mately perform, as bleeding is an indication of settlement
sured by means of the rotating-cylinder test and the into a denser state.
number of blows in the liquid-limit test is illustrated in The amount of bleeding water is also very dependent
Fig. 5. on the grading of the fill. In Fig. 6, the increase in bleed-
The grading or particle-size analysis of the tailings has ing water with increasing cement content also arises be-
a considerable influence on the water content required cause the cement content acts as an addition of fines,
for the 100 P viscosity. For example, a typical gold-mine thus improving the grading and allowing the particles
slimes requires a water content of 55 per cent, whereas to settle into a closer packing.
a coarser gold-mine sand requires only 33 per cent. The
cement content of the slurry also influences the 100 P Compression of Fills under Stress
viscosity water content slightly, e.g. for the same gold- Most of the compression undergone by a pumped fill
mine sand: subjected to stress occurs at low stresses. This is illus-
lOOP trated by Fig. 7, which shows how a typical fill becomes
Cement content water content less compressible as the stress increases. Hence, it is the
0 ~% compressibility at low stresses that is all-important-
5 31,9% a fill that is incompressible at low stresses will remain
10 31,6%. incompressible at high stresses and vice versa.
This effect appears to result from an improvement in It will be noted that the compression curve has been
the grading of the fill from the addition of the cement presented in terms of decreasing water content. This
fines. The fines convert the approximate straight-line
grading of the sand (sand 2 in Fig. 8) to a gap grading
bI
that is similar to G3 in Fig. 3a, thus improving its flow c:: 6
properties at a given water content. '-"tJ
(l)
(l)
Settlement of Slurries after Placing --
Cl)
Sand 2
A pumpable slurry always contains more water than is 4
required to fill the voids in the settled material. If this .5'
were not so, interparticle friction and interlock (equation V)
V)
3a) would prevent the material from flowing. Once the 0 2
-..J
pumped fill is in place, this surplus water drains or
l...
bleeds off and must be disposed of. Fig. 6 shows the (l)

influence of cement content on two typical sand -cement I'(J

slurries placed at the 100 P water content. The extent ~ 0

~ 0 2 4 6 B 10
Viscosity of Slurry Pas Cement Content %
60 1 10 100 Fig. 6-Relationship between cement content and water lost
by bleeding for two cemented sand slurries

55 60

- 0
~ 50 50
~
t:
CLo
1:
. 40
~ ~- 20 ~
c:: .S
<345 ~
~
'-
c::
<330 1 -
I1J
....
40 \ij

~ 40
....
..
~
'" I1J
lJ
~ 20 ~-60 'i:
..
;:,.
35
1 10 80
10 100
0 10 20 30 1,0 50 60 70
No of Blows in Liquid Limit Test Vertical EffectIVe Stress MPa
Fig. 5-Relationship between viscosity, blow count in the Fig. 7-Compression curve for slimes containing 10 per cent
liquid-limit apparatus, and water content for a typical cement and showing that most of the compression occurs
tailings slurry at Iow stresses

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY OCTOBER 1979 449
 tOD emphasizes the fact that a fill of saturated tailings can
compress only as water is expelled from it under in-
creasing stress.
80 As stated earlier, the compressibility of a fill is highly
dependent on its grading, both at small and large stresses.
For small stresses, this is illustrated by Figs. 8 and 9:
60 Fig. 8 shows the grading curves for three sands (sands
I 1 and 2 have been referred to in Fig. 6), while Fig. 9
'-~c: I
I .
shows the compression curves, at low stresses, for sands
i:l: 40 1 and 3 and the effect on compressibility of added
~ -,'-I-Sand 1 cement fines. (All the specimens were placed at a water
;' content corresponding to 100 P viscosity.)
" The water contents at zero effective stress indicate
20 4-/~i the effect of cement content on bleeding, since all the
1// specimens were placed at the same initial water content.
It will be noted that the materials that bled most (lower
0 ...,,"1'./
.- initial water contents) are also less compressible under
0,02 0,06 0,2 0,6 2 applied stress. This conclusion also applies to a com-
parison of sands 1 and 3: sand 1 had a lower initial water
Coarse Fine
I
Medium I~~~ content than sand 3 and proved less compressible under
Silt Sand
I
==:J stress.
Apart from altering the grading of the fill, the cement
Particle Size mm
content also decreases the compressibility by developing
Fig. ~Grading curves for three typical tailings sands
cohesive bonds between the solid particles of the fill.
25
-I

~""'13
E
20 c:
.~0
12
{J
:.:::
15
0
!I)

~ 11
0
l.>
Sand Cement: 2% 0 .....
" 0

~10 5% . c: 10
.~
10% e u
~.....
c:
C1J
.....
5 l.>
~ 9
c:
0
l.) 0 200 400 6 oo 8oo TO00
(a) 8
'- ~40
Effective Stress kPa
0 200 iOO 600 800 TOOO
Effective Stress kPa
40
~35
Sand 3, Cement: 2% 0 .. 10% Cement
5% . b
"'" 30
E

30 10% e '-
.~ 20
.Q
11!
~
E
25 Cl..
~ 10

0
20 100 200 300 400 500 600 700
0 200 400 600 800 100n Effective Stress kPa
(b)
Effective Stress kPa Fig. IO-(a) Variation of coefficient of consolidation with
effective stress for a typical tailings fill
Fig.9-(a) Compression of sand I at three cement contents (b) Variation of coefficient of permeability with
(b) Compression of sand 3 at three cement contents effective stress for a typical tailings fill

450 OCTOBER 1979 JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY
 However, as will be seen later, at the high water contents decreases with increasing effective stress, typically as
and low cement contents used in pumped tailings fills, shown in Fig. lOb.
these bonds are relatively weak and have only a small Although the figures quoted above are typical, both
effect on the compressibility of the fill. c v and k can be drastically reduced by the presen-:e of
To sum up, Fig. 9 shows that materials that settle only a few per cent of clay-sized material in the fill.
into a dense packing on being placed will be less com-
pressible under subsequent loading. Shear Strength of Fills
Typical strength characteristics for cemented-tailings
Drainage of Fills fills are shown in Fig. 11. In these diagrams, the maxi-
As described earlier, the rate at which fills under mum shear stress at failure t(Ul-u3)r has been plotted
compression drain is influenced primarily by the co- against the average principal effective stress t(Ul' +u3')r.
efficient of consolidation, Cv. Fortunately, most tailings The diagrams clearly show the effect on shear strength
materials are relatively permeable, and values of Cv lie of an increasing cement content. Increasing cement
in the range of 10 to 50 m2 a year or more (in comparison contents affect both the slope 0/ of the strength lines
with clays, which may have Cv values as low as 0,1 m2 a (0/ is related to the angle of sh!Jaring resistance 4>' by
year). sin </>' = tan 0/) and the cohesion c' of the material. At
The value of Cv is influenced by the effective higher stresses, the major effect on shear strength
stress, and decreases with increasing effective stress as results from the increase in </>'. For example, for sand 3,
shown in Fig. lOa. However, the effect of stress on Cv, when t(Ul' +U3')r was equal to 3000 kPa, the strength
and hence on the rate of drainage, is usually relatively increase as a result of an increase in the cement content
minor. from 0 to 10 per cent was 550 kPa, of which 110 kPa
It may also be necessary to estimate the rate of per- resulted from the increase of cohesion and 440 kPa
colation flow through a fill if the fill is being used as a from the increase in </>'. This disparity increases pro-
water barrier. In this case, D'Arcy's law, gressively with increasing stress.
v=ki . . . . . . . . . . . . . . . . . . . (8) The effect of cement content on both c' and </>'for
(where v is the velocity of flow under a gradient of head sands I, 2, and 3 is summarized in Fig. 12. A comparison
i, and k is the coefficient of permeability) is applicable. of Fig. 12 with Figs. 8 and 9 shows that the least com-
The coefficient of permeability for tailings materials pressible sand (I) has considerably superior strength
typically falls in the range I to 50 m a year, and also characteristics to those of sand 3. Sand 2, which has a

200
2000

Sand 1
1500 150

1000 ~~100
c:
.52
11)
~..:c 500 QJ
.c: 50
0
-.;.'" l.>
tf>
,
b- 0 u
'-~ 0 500 1000 1500 2000 2500 3000 0
112(q'+ a;), kPa 0 2 4 6 8 10
!:! 2000 QJ Content
~ Cement %
u
li) c:
Sand 3 .~
:a 1500 11)
'-
" Cl>
..c: ,0 ID
lI) Q:: 3
tOOO CrI
.~
~It!
QJ
500 .c: san d 3
lr)
.....
0
QJ
0 10°
"5!
0 500 1000 1500 2000 2500 3000 c:
0 2 4 6 8 10
'12 (a;'+ 0;), kPa ~~Cement Content %
Fig. I I-Failure stress diagrams for triaxial shear tests on two Fig. 12-Relationship between cohesion c', angle of shearing
cemented sands resistance </>',and cement content for three cemented sands

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY OCTOBER 1979 451
compressibility intermediate between that of sands 1 100
and 3, displays intermediate strength characteristics. One
is therefore led to the conclusions that, in relation
80
to strength,
(i) the greater the initial density of the fill, the better
its performance; and 60
(ii) the cementitious bonds developed by any cement (1"
ei' 0
.,,.
that may be added to a fill play a relatively un- 40
important role in the strength of the material; the
effect of the cement on compressibility and frictional
20
strength via its effect on .p' is far more important.
This latter effect probably results from improved grading
(via higher settled density) rather than any chemical 0
0 50 100 150
bonding.
Tm~x % .
Role of Cement in Pumped Fills ao

In view of earlier remarks, one is led to ask whether Fig. I4-Results of undrained repeated-load triaxial shear
cement is necessary in a pumped fill and whether, if the tests on a cemented slimes fill
major effect of the cement is to improve the settled
density and compressibility of the fill by impoving its of the applied shear stresses is relatively large, there is
grading, some other type of fine material (preferably a no time for drainage of the fill to allow the shear-
waste material obtainable at low cost) would not do induced pore pressure to dissipate. In terms of equation
almost as well. These are questions that should be (3),
explored in each individual case. There are certainly 6T=tan .p'6(a-u); ,. (3b)
indications, for example, that a mixture of pulverized a remains reasonably constant, but u increases with each
fuel ash (PFA) and lime added to a tailings fill can pro- application of shear stress, the net result being that 6T
duce almost the same effect as cement9. PFA is produced is negative, i.e. the shear strength of the fill decreases.
in vast quantities in South Africa2 as a power-station In the worst case, 6T becomes equal to -T, and the fill
waste product, and PF A from ash dams can be obtained loses all its strength and liquefies.
for the cost of loading and transportation. Fig. 13 com- The effect of seismic or blast loading can be simulated
pares the compression curves for three slurries of gold- in the laboratory by the application of alternating shear
mining slimes containing cement and lime- PF A fines. It stresses under undrained conditions to a specimen of
is evident that the use of lime- PF A mixes warrants tailings fill in a triaxial test 7.
further investigation, especially as the addition of this In tests of this type on specimens of cemented-slimes
type of fines to a coarser material than gold-mine slimes fill, it has been found that the pore pressure first of all
could be expected to have an even better effect. (The increases (indicating a tendency for the fill particles to
grading of the PF A was not much finer than that of the move into a denser packing) and then starts to decrease
slimes.) (indicating a tendency for the material to increase in
in volume or dilate). A set of results for tests of this typelO
Effects of Shock Loading is shown in Fig. 14. On the vertical axis, the ratio a'/ao'
Shock loading as a result of blasting or seismic events represents the effective stress after N cycles of repeated
has the effect of applying a rapidly alternating shear loading as a percentage of the initial effective stress
stress to a fill. Each application of shear stress causes an (ao') in the fill. On the horizontal axis, the magnitude of
increment of pore pressure and, because the frequency the applied repeated shear stress is also expressed as a
percentage of the initial effective stress. It will be seen "
0 that in this particular case, 100 cycles of a shear stress
equal to ao' caused the effective stress in the fill to fall
to 15 per cent of its initial value. The effective stress
20 (and hence the strength) did not fall to less than this,
~ however, and the fill was in no danger of liquefying.
.~ ~.
1,0 It is well established that dense particulate materials
~ do not liquefy when subjected to seismic-type loading,
If)
Iij
.~ 60
0 but that materials of low density are prone to lique-
faction. Tests such as those depicted in Fig. 14 should \
1:::
Cl>
([) certainly be undertaken if a fill is likely to be subjected
:::..
I to shock loading, but the best guarantee against lique-
80 faction is to ensure that the grading of the fill is such
0 10 20 30 1,0 50 60 70 that the fill will settle into as dense a state as possible
Vertical Effective Stress MPa
on being placed.
Fig. I3-Compression of gold-mine slimes containing Conclusion
(I) 10 per cent cement
(2) 5 per cent lime and 16 per cent PFA The density a fill attains on being placed is very
(3) 6 per cent lime and 12 per cent PFA important since it affects the subsequent compressibility,

452 OCTOBER 1979 JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY
 strength, and resistance to shock loading. The density 4. WOROTNtcKI,G., ALEXANDER,
L. G., WILLOUGHBY,
D. R.,
and ASHCROFT, J. F. Deformation and behaviour of high
attained by a fill material is very much a function of its rise filled stopes at C.S.A. Mine Cobar, N.S.W. Proceedings,
grading, and the inclusion of cement fines may improve 2nd Australian-New Zealand Conference on Geomechanics,
the grading of a .fill material and hence increase its Brisbane, 1975. pp. 48-55.
5. BLIGHT, G. E., MORE-O'FERRAL, R. C., and AVALLE, D. L.
density and strength and decrease its compressibility. Cemented tailings fill for mining excavations. Proceedings,
9th International Conference onSoil Mechanics and Foundation
Acknowledgements Engineering, Tokyo, 1977. vo!. 1. pp. 47-54.
6. CHAMBEROF MINES OF SOUTHAFRICA. Handbook of guide-
The unreferenced test results given in this paper are lines for environmental protection. vo!. 1/1979. The design,
published by kind permission of the City Engineer of management and closure of residue deposits. Johannesburg,
the Chamber, 1979.
Johannesburg, Messrs Cromore (pty) Ltd, and Messrs 7. LAMBE, T. W., and WHITMAN, R. V. Soil Mechanics. New
Samancor Management Services (Pty) Ltd. York, Wiley, 1969.
8. HANSFORD, G. S., LEVY, C. D., and DE KOCK, J. W. Rheo-
logical measurements on pulps from South African gold
References mines. J. S. Afr. Inst. Min. Metall., vo!. 76, no. 8. 1976.
1. MORE-O'FERRAL, R. C., and MACAULAY, R. J. Slime- pp. 363-369.
cement fill as a gold mine stope support. Proceedings, 9. AVALLE, D. L. Properties of weakly cemented slurries of
Symposium on Practical Support Methods. Johannesburg, gold mine sHInes. M.Sc. (Eng.) thesis, University of the
South African Institute of Mining and Metallurgy, 1978. Witwatersrand, Johannmburg, 1976.
2. WAGNER, H., and GALVIN, J. M. Use of hydraulically 10. CHAMBER OF MINES OF SOUTH AFRICA RESEARCH ORGANISA-
placed PF A to improve stability in bord and pillar workings TION. Instrumentation to measure various parameters of
in South African collieries. Proceedings, Symposium on the fill in deep-level excavations and a preliminary assessment
Utilization of Pulverized Fuel Ash. Pretoria, 1979. of the performance of a cemented slurry fill. Project
3. PATCHETT, S. J. Fill support systems for deep-level gold GMIUO3, Research Report no. 5/79. Johannesburg, the
mines. J. S. Afr. Inst. Min. Metall., vo!. 78. 1977. pp. 34-46. Chamber, 1979.

ProCessional development seminars

During the 1979-80 academic year McGill University Quality of working life in the mineral industry, 28th
will once again be offering professional development January to 1st February, 1980 (Seminar Leader:
seminars in various aspects of mineral engineering, T. E. Hawkins)
mineral management, and mineral economics. These Union-management-employee relations in the mineral
seminars have been designed to meet the practical needs industry, 4th to 7th February, 1980 (Seminar
of the Canadian mineral industry and of mining organiza- Leader: T. E. Hawkins)
tions in developing countries. Geostatistical ore reserve estimation, II th to 15th
Lectures, discussion groups, case studies, and work- February, 1980 (Seminar Leaders: M. David and
shops are conducted by the staff of McGill University M. L. Bilodeau
with the assistance and co-operation of recognized Mineral investment decision techniques, 18th to 29th
authorities from government agencies, mining companies, February, 1980 (Seminar Leader: B. W. Mackenzie)
independent consultants, and other universities.
Organizing for results in the mineral industry, 3rd to
The United Nations Centre for Natural Resources,
7th March, 1980 (Seminar Leader: T. E. Hawkins)
Energy and Transport and the Canadian international
Development Agency help in arranging the participation The financing and implementation of mineral projects,
of personnel from developing countries. 10th to 14th March, 1980 (Seminar Leader: P.
The following seminars are offered: Glenshaw)
Principles and practices of ground control, 5th to 17th The planning and execution of mineral development
/; November, 1979 (Seminar Leader: John E. Udd) strategies, 17th to 21st March, 1980 (Seminar
Extractive metallurgy of copper, 12th to 16th Novem- leader: P. M. T. White)
ber, 1979 (Seminar Leader: W. G. Davenport) Mineral processing systems, 24th March to 3rd April,
Blasting systems, 19th to 23rdNovember, 1979 (Seminar 1980 (Seminar Leader: J. A. Finch)
Leader: R. R. MacLachlan) For further information, write to Lorna McFadden,
Health and safety aspects of the mine environment, Department of Mining and Metallurgical Engineering,
21st to 25th January, 1980 (Seminar Leader: D. A. McGill University, 3480 University Street, Montreal,
Trotter) Quebec, Canada H3A 2A7.

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY OCTOBER 1979 453
Rock-handling
.
techniques at Sishen iron-ore
mine by B. G. FORDYCE*. B.Sc. (Eng.) (Wits.). M.B.L. (S.A.) and
W. B. PARKER**. B.Sc. (Eng.) (Wits.) (Members)
SYNOPSIS
A description is given of the methods used in management planning, pit-equipment selection and maintenance-
operator training, and formal organization of mining personnel.

SAMEVATTING
Die metodes van bestuursbeplanning, die keuring en instandhouding van groeftoerusting, operate:.Jrsopleiding
en die formele organisasiestruktuur van die mynbouafdeling, word bespreek.

Introduction Outputs of the MPS

The production of iron ore at Sishen Mine began in Besidel.pit plans drawn to desired scales, tonnages of
June 1953, the ore being crushed and screened in a dry ore ana ~ste rock by level and by area are calculated.
state. In 1961 a crushing and wet-screening plant was One of the most useful outputs is the determination of
erected, and in 1963 the first heavy-medium beneficia- minimum and maximumt stripping curves3 (Fig. 2). If
tion plant was commissioned. the minimum curve is followed, the present value of the
The production of iron ore for the Sishen-Saldanha cost is minimized, but the pit equipment and service
Export Project began in 1976 after considerable expan- facilities must be continuously expanded. If the maxi-
sion of the pit equipment and facilities. An additional, mum curve is followed, the present value of the total
improved heavy-medium beneficiation plant was also production facilities is maximized, with a later redund-
built. The maximum total output of these two plants, ancy of these facilities. An average stripping policy
which supply lumpy and fine ore for domestic and ex- ensures at least stable production facilities over the
port markets, is rated as 27 million tons per year. life of the pit, which is advantageous to management
Run-of-mine ore and waste rock are at present being control.
mined at a rate of nearly 80 million tons per year.
Roads and Declines (Ramps)
Management Planning The main pit roads, which are 30 m wide, usually have
Long and Medium-term Planning by the MPSl a slope of 1 in 15, with a maximum slope of 1 in 12 in
Heavy reliance is now placed on a powerful com- special circumstances. Safety berms are built by the
puterized mine-planning system, MPS, which has been tipping of waste rock on the cliff side of these roads (see
partially installed at Sishen. The ore-body is modelled Fig. 7).
as follows: drawn geological sections are digitized and Access roads are built for the pit electrical-distribution
link-point sets between sections are indicated; inter- network and the pit dewatering pipelines.
mediate sections are interpolated linearly by the system; Declines are normally 50 m wide, with a minimum
and hence a three-dimensional model of the ore body is width of 30 m.
constructed. Any section that is perpendicular to the Waste Dumps
survey co-ordinate system can be viewed on the graphics
The first priority is that the waste dumps must be as
terminal or plotted on paper to the desired scale. Once
close as possible to the pit for minimum haulage dis-
these sections are checked by the geologists, pit design
tances, waste-rock dump and pit-wall stability and
can commence.
drainage having been taken into account. Dumps are
Long and Medium-term Pit Planning built at present in 20 m slices with a maximum height
The most economic pit design is determined by use of of 60 m.
the Lerchs-Grossman dynamic programming algorithm2 Medium-term Planning
constrained by the marginal volumetric instantaneous
Forecasts of ore sales are periodically reviewed, and
stripping ratio, the estimated maximum pit-slope angle,
this information is used in the scheduling of ore and
and the surface contours of the area.
waste-rock production according to the objective of
This layout is typically unpractical, but, by means of ..
minimizing costs, constrained by the present stripping-
the simplified interactiveness of the system, pit profiles
ratio policy and the ore specifications, which require
are smoothed and haul roads inserted. The tonnages of
blending of the ore from several areas. These decisions
ore produced from the final and optimum pit designs
are carried over to 1 in 5000 scale plans, where priority
are then retrieved from the model of the ore-body. Part
areas are indicated and ranked.
of the long-term pit plan is shown in Fig. 1.
Short-term Planning
*Iscor, P.O. Box 178, Durnacol, Natal. The object of the short-term planning is to translate
**Iscor Thabazimbi, Transvaal. the medium-term plan into weekly and daily plans so
tMaximum curve established by mining the pit level to final
pit limits. that the production objectives are realized efficiently

454 OCTOBER 1979 JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY