Downloaded from http://egsp.lyellcollection.

org/ at Monash University on June 3, 2015

/ . R o c k fill

7.1. Introduction this chapter. In June 1990 a NATO Advanced Study

Institute was held at LNEC Lisbon on rock fill
Where natural ground is excavated and used as bulk fill structures. The proceedings of this two week meeting
in civil engineering works, it is categorized as either (Maranha das Neves 1991) provide a useful account of
earth fill or rock fill. The distinction is important from many aspects of the subject.
two points of view:

9 from a technical standpoint the difference in perfor-

mance of the two types of fill is significant, rock fill
7.2. Functions and required properties
generally exhibiting superior engineering properties to 7.2.1. Applications
earth fill
9 from a contractual standpoint the difference in the The properties required of rock fill are directly related to
effort required to excavate the materials is significant, its function in a particular application. Rock fill may be
rock being more difficult to excavate and generally used in a variety of engineering structures including the
requiring blasting or ripping following:
9 highway embankments
The concept of rock fill construction has undergone some
9 embankment dams
radical changes over the years. Before heavy compac-
9 foundations for buildings
tion machinery was available, loose dumped rock fill was
9 gabions
used to build dam and railway embankments. The 22 m
high South Fork dam built in 1852 in Pennsylvania had The most common use of rock fill is probably in highway
a downstream shoulder of rock fill; its failure due to embankments. These are often of modest height but may
overtopping in 1889 ruined a large part of the city of be of considerable length. The use of rock in highway
Johnstown and 2209 people died (Jansen 1980). In the embankments tends to be governed by locally available
middle of the 19th century dams were built by gold material.
miners in California entirely of dumped rock fill. Sub- Rock fill has also been used extensively in the con-
sequently, in the first half of the 20th century, dams struction of embankment dams and some of the highest
100 m high were built in California using dumped rock dams in the world are built of rock fill. Nurek dam in
fill. Large settlements of the rock fill frequently occurred Tajikistan is 300 m high with an embankment volume of
during construction and reservoir filling. Techniques 58 • 106 m 3. As rock fill has a relatively high perme-
adopted to reduce settlement involved the use of sound ability it is usually necessary to provide a watertight
rock with a minimum of fines and sluicing with large element such as a clay core within the embankment (see
volumes of water. Fig. 7.9), or a concrete or asphalt membrane on the
With the advent of modern compaction plant it became upstream slope, see Fig. 7.11.
feasible to place rock fill in layers of a thickness which Shortage of good building land increasingly leads to
could be adequately compacted. Using this approach a construction on filled ground and it is not uncommon
well graded rock fill with some fines was found to give to place engineered fill to an appropriate specification to
superior performance. Since large highway embankments form the foundation for buildings. Large volumes of
are a relatively recent development, nearly all highway opencast mining backfills, some of them rock fills, have
rock fills have been constructed in this way as have most been placed and compacted to engineering specifications
modern rock fill dams and railway embankments. so that buildings and roads can be built over them when
High-quality rock fill is usually quarried at depth the opencast operation has been completed.
and there may be considerable quantities of overlying Gabions are steel wire or plastic mesh baskets filled
weathered overburden. There are obvious economic and with rock and they are usually used in situations where
environmental advantages if such weathered material land is at a premium or soil protection is required. They
can be used in the fill operation. Low grade rock fill, or have many applications in civil engineering, particularly
soft rock fill as it is sometimes called, can be con- on highways where soil retention is necessary.
sidered a transition between earth fill and rock fill and The different applications of rock fill have signifi-
has found increasing use. In a soft rock fill the fines fill cance for the degree of selection and control which it is
the voids between the coarser particles and a maximum either necessary or practical to exercise during fill place-
density is obtained after compaction. ment. In most applications it will be i m p o r t a n t that
The use of rock fill in civil engineering structures is a the parent rock material is strong, durable and chem-
maior subject which can only be briefly summarized in ically inert.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

258 ROCK IqLL

7.2.2. Slope stability

Where rock fill is used to form an embankment above
ground level, it is important to ensure that the rock fill
has sufficient strength for the slope to remain stable. 1000
Slope failure involves shearing of the rock fill in which
rock fragments must move over or between each other.
This motion is opposed by friction and obstruction,
giving rise to 'shear strength'. The shear strength will be
influenced by the size, shape, gradation and strength of
the rock fragments that make up the fill. These factors ~m /I constant volume
g- ')
can be conveniently expressed in terms of the internal
angle of friction or shearing resistance, qS. The shear 500
strength of a particular rock fill is also a function of both
the packed density of the fill and the stress level. In
simple terms the shear strength (r) of the material is b
related to the normal effective stress (or') and the angle of .g
shearing resistance (47) by the expression: a

r = or' tan r

A full explanation of shear strength and its use in slope 0 5 10

stability is given by Bromhead (1986) and Hoek & Axial strain ~1 (%)
Bray (1981). Fig. 7.1. Drained triaxial compression test on heavily
Rock fills are generally characterized by high shear compacted basalt rock fill with confining pressure of 100kPa
strength, particularly when they have been heavily (after Charles & Watts 1980).
compacted and, therefore, they have the ability to be
constructed at steep slopes. For a heavily compacted
rock fill the angle of shearing resistance, 05', will be The ability of a slope to remain stable is assessed by
particularly large at low confining stresses where the the ratio of the restoring force to the disturbing force
maximum, or peak, strength is associated with strongly which is known as the factor of safety (F). It would seem
dilatant behaviour. After the peak strength has been that some modern rock fill structures may have slopes
reached, further displacent reduces the strength to the with high factors of safety that do not fully utilize the
constant volume, or critical state, angle of shearing shear strength properties of the fill. An acceptable value
resistance, VYcv.This is illustrated in Fig. 7.1. For a loose for the factor of safety obtained in a stability analysis,
rock fill the maximum shear strength will be close to the say, F = 1.5, may have been based on conservative
constant volume angle of shearing resistance. The angle assumptions about the shear strength of the compacted
of repose of the loose rock fill, which is the slope rock fill, say, ~,~'= 4 0 .
angle that loosely tipped rock fill naturally adopts, is The pronounced curvature, particularly at low stres-
also similar to the constant volume angle of shearing ses, exhibited by Mohr failure envelopes for compacted
resistance and is typically 35 ~ to 40 ~. rock fills, which is shown in Fig. 7.2, has formed an
The dams built of dumped rock fill in California in the obstacle to the use of realistic shear strength parameters
second half of the nineteenth century were up to 30 m in in stability calculations as analyses have usually been
height and had slopes as steep as 1 vertical in 0.5 based on linear failure envelopes. It is now feasible
horizontal (63 '') (Galloway 1939). The construction of both to measure the shear strength parameters of a
slopes steeper than the angle of repose of the loose rock compacted rock fill (Charles & Watts 1980; Perry 1994)
fill, was achieved by hand placement of stone to form a and to use these realistic parameters in slope stability
rubble retaining wall. In recent years different construc- analyses (Charles & Soares 1984; Perry 1994).
tion techniques have been adopted for the construction
of rock fill embankments. The use of modern heavy
7.2.3. D e f o r m a t i o n
earth-moving machinery has led to rock fill being placed
in layers typically 0.5m to 1.5m deep and compacted Some of the most complex deformation situations
with heavy vibrating rollers. This method results in very involving rock fill occur with embankment dams where
dense well graded rock fills with strength and deforma- there are several different types of fill materials in
tion properties greatly superior to those of the uncom- various zones within the embankment. However, the
pacted, uniform sized rock fills previously used. Yet, as most difficult situations occur in those applications
de Mello (1977) pointed out, slopes have tended to where even small movements may cause problems such
become flatter. as the following:
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK FILL 259

9 downward percolation of water escaping from drains

in the crest of an embankment
The effect of inundation on an unsaturated rock fill can
400 be examined in an oedometer test carried out on a
sample of the rock fill compacted to the field density at
Cq
E the field moisture content.
z The compressibility of rock fill can be measured in
v
I:--
large oedometer tests and Fig. 7.3 shows the compres-
sion of some heavily compacted rock fills measured in a
one metre diameter oedometer (Charles 1991; Charles
& Penman 1988). Large oedometer tests have led to
$ 200 the following conclusions about the compressibility of
.lE
CO rock fills (Kjaernsli & Sande 1963; Sowers et al. 1965):
9 compressibility decreases with increasing strength of
the rock particles
9 compressibility decreases with increasing density of
the fill
9 compressibility is smaller for smooth surfaced
materials
200 400 9 compressibility is smaller for a broad grading
Normal effective stress o-' (kN/m2) 9 the major mechanism is the crushing of highly stressed
points of contact between particles which in turn
Fig. 7.2. Curved Mohr failure envelope for heavily compacted results in some reorientation of particles
slate rock fill (after Charles & Watts 1980).
Oedometer tests and field monitoring have also shown
that the rate of continuing settlement of rock fill under
9 where settlement sensitive structures are built on the constant applied stress is similar to the secondary com-
rock fill pression behaviour of clays. This conclusion is impor-
9 where a road passes from a rock fill embankment onto
tant as, in many situations the long term compression of
a rigid structure such as a bridge the rock fill will be more important than the immediate
In these cases it is imperative to reduce post placement compression that occurs during rock fill construction.
fill settlement.
Applied stresses within the rock fill will increase during
fill placement due to the weight of overlying layers of Vertical stress % (kPa)
rock fill, but the compression of the rock fill that occurs 500 1000 150
at this stage will normally be of little practical con- 0
sequence. It is the movements that occur subsequent to ~---------'~i Sandygrave I '
fill placement which are normally of most significance.
Compression of rock fill is associated with crushing of
points of contact and rearrangement of rock fragments
and particles which may be caused by a change in 2
effective stress or changes in moisture content. Changes
in effective stress can be caused by, for example: ~>
C
9 the rise and fall of the groundwater level within the -~ "-,~andstone
rock fill "~
9 the weight of buildings constructed on the rock fill ~ 4
Poorly compacted, unsaturated rock fill is likely to be
susceptible to collapse compression when first inundated
gravel 21.5
~ - N ~ d 3 )
as an increase in moisture content may weaken the
parent rock material and lead to further crushing at 6 Slate rockfill 21.9
points of contact. An increase in moisture content could Sandstone rockfill 20.2
occur in the following situations: I I

9 submergence of rock fill by a rising groundwater level Fig. 7.3. Compressibility of heavily compacted rockfills
9 downward percolation of surface water through the measured in one metre diameter oedometer (after Charles &
r o & fill Pemnan 1988).
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

260 ROCK FILL

Heavy compaction of the rock fill during placement can will be to volume reduction when subjected to any
greatly reduce the creep rate of the rock fill at low subsequent dynamic loading which could be associated
stresses, with, for example, a seismic event.

7.2.4. Dynamic behaviour 7.2.5. Permeability

Vibration can cause compression of a rock fill by the There is usually a general requirement that permeability
rock particles being rotated and rearranged into a denser should be sufficiently large to allow water to drain away
packing. Therefore, compaction of rock fill during and prevent the creation of excessive water pressures
placement can be achieved by the use of a vibrating (pore pressure) during construction by the saturation
roller. Such compaction is beneficial and the more of voids associated with the finer material in the fill.
effective the compaction, the less vulnerable the rock fill If water cannot escape the pore pressure reduces the

Fi~. 7.4. Types of gabion cage (after BS 8002: BSI 1994).

 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK PILL 261

effective stress level (or') thereby reducing the shear 9 to provide protection against erosion protection for
strength (T). In some applications, such as embankment watercourses and to provide protection against ero-
dams, there may be more specific permeability require- sion by the sea or rainfall
ments. While heavy compaction should produce a fill of
Where gabions are used to construct retaining walls, for
relatively low compressibility it also creates more fines
example to retain soil for a widened highway or to
and it is important that the percentage of fines is not so
provide extra ground for a development, the gabions
large that the permeability (hydraulic conductivity) of
take the form of baskets and are placed one on top of
the fill is reduced below, say, 10 5 m/s. The permeability
the other. Some designs require the gabions to be tied
of a rock fill can reduce very rapidly as the percentage of
together while others rely on the friction between the
fines increases and it is usual to specify an upper limit of
gabions to prevent movement. The walls should be
the order of 10% for the percentage of particles finer
constructed to a batter to enhance their appearance and
than 0.06 mm.
to resist overturning. Gabions can be used to strengthen,
stabilize, repair and reinstate existing slopes as well as
7.2.6. Gabions being used as a method of new construction.
For revetments, the lining of rivers and canals, or
Gabions are large cuboid shaped baskets or mattresses
coastal protection, gabions in mattress form are used.
which are made of steel wire, square welded rod or
These gabions have a large areal extent compared to
plastic mesh as illustrated in Figs 7.4 and 7.5, and filled
their height and have a low cellular structure. Their
with rock. The size of gabions vary but they are
permeability allows water behind the structure to drain
commonly 2 m to 6m long, l m to 2 m wide and in
away when water levels fall. In some cases a filter
depths of 0.3 m, 0.5 m and 1 m.
material may be used between the soil and gabion to
Gabions are permeable and flexible. These two
prevent wash-out of the soil behind the gabions.
qualities make them ideally suited for retaining materials
The rock pieces used in the gabion have to be of a
and for construction on soft ground. They can be used
grading which prevents them passing through the mesh.
for the following purposes:
The size of the mesh opening is, therefore, taken as the
9 to build retaining walls minimum size of rock fill. The maximum size of fill is
9 to stabilize or repair earthwork or natural slopes usually two thirds the minimum dimension of the gabion
9 to construct revetments or 200 mm whichever is the smaller. This requirement is

Fi~. 7.5. Oabions in use for road construction (Courtesy P. G. Fookes).

 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

262 ROCK H L L

to allow the rock in the gabion to be placed to fill the particle size distribution for laboratory tests need to be
gabion well. As gabions should be filled by hand to examined. These matters are not easy to resolve as
prevent damage to the gabion mesh, handling of the particle fracturing can influence the behaviour of rock
material is also a consideration. In environmentally fill and the particle size distribution may change
sensitive areas, the rock fill in the outer zone of the significantly during the course of a test. The simplest
gabion can be placed in patterns or placed to give the approach is to remove all the oversize material and test
semblance of a dry stone wall. To allow for settlement of what is left although this procedure assumes that the
the gabions, especially in a wall, the gabions should be finer fraction is representative of the whole. It has been
slightly overfilled. suggested that it is preferable to test a material with a
For the long-term stability of an individual gabion, grading parallel to that of the field grading. However,
the rock fill should meet an appropriate strength this latter approach requires far more work in sample
requirement. This is specified in terms of the ten-per- preparation and has the disadvantage that the tested
cent fines test. The minimum value required in the UK sample may have an excessive quantity of fines.
for highway construction is 50 kN. This ensures that the
rock fill will not Pail due to point contact of the rock
pieces as the weight of gabion, and gabions above, is
7.3.2. Index and classification tests
increased. Durability requirements are assessed using a
crystallization test (Ross & Butlin 1989). With the difficulty and expense of carrying out strength
It is important that consideration is given to the and compressibility tests, there is much interest in
corrosion resistance of the gabion. For temporary works correlating index and classification test results with
the wire used to form the gabion cage may be uncoated. engineering behaviour.
However, for permanent structures the wire is usually
either galvanized or polyvinyl chloride (pvc) coated. Particle strength, har~hwss and durability
Galvanized wire gabion cages are subject to corrosion The properties of the individual fragments of rock will
attack if aggressive soils are in contact with the cage. have a significant influence on the behaviour of the rock
Tests should, therefore, be undertaken to determine the fill. Particle strength and hardness will control the
aggressivity of the soils in terms of chloride ion con- amount of particle breakdown and crushing of points
tent, sulphate content, pH, redox potential, soil resis- of contact which occur during compaction and subse-
tivity and soil moisture content. If the soil is going to quent loading.
cause corrosion of the gabion cage then pvc coated wire Long-term performance of the fill will be affected by
should be considered and the pvc coating should be at the durability of the rock. Rocks are affected by wetting
least 0.25 mm thick and bonded to galvanized wire. The and drying; the strength of saturated rock fragments may
bonding should be sufficient to prevent capillary water be much less than the strength of the fragments when dry.
entering between the coating and the wire. Both pvc and Weak rocks such as shales will swell or disintegrate when
galvanized wire can be damaged by abrasion and their exposed to atmospheric wetting and drying and are thus
use should be avoided where granular materials may fall permanently weakened.
or wash against them. Properties of strength and durability can be measured
in several ways:

i the uniaxial compressive strength (q,) of cylindrical

7.3. Laboratory testing specimens of rock can be measured but considerable
time and effort is required to prepare the cylindrical
7.3.1. General
sample of rock for this standard test
The large size of rock fragments used in typical rock fills 9 in the point load strength test (Broch & Franklin
causes particular difficulties in determining the engineer- 1972) cylindrical specimens or irregular lumps of rock
ing properties of rock fill. Rock fragments with a are loaded to failure between the conical platens of a
dimension as large as lm or more have sometimes been portable tester
used in rock fills. It is clearly impractical to design 9 in the slake durability test (Franklin & Chandra 1972)
equipment that could test such material. The largest the resistance of a rock to wetting and drying cycles is
triaxial testing equipment in the world can only assessed by immersing samples in water and noting
accommodate samples of up to about l m in diameter the rate of disintegration; this test is useful for
and is thus only suitable for testing rock fill with a relatively weak rocks which are sensitive to the test
maximum particle size of" 150 200 mm. Even this size of 9 the aggregate crushing test has been used to assess the
equipment is not widely available and a single test at this suitability of a material to form a rock fill (Perry &
scale is extremely expensive. Parsons 1986)
The unavailability of large-scale testing equipment 9 the saturation moisture content or water absorption
means that possible approaches to scaling down the field can give an indication of the weathering potential
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK I-ILL 263

Particle size distribution For a rock fill containing large rock fragments, there will
The behaviour of rock fills will be influenced by their be great difficulties in measuring the three dry densities
particle size distribution (PSD). They can be described needed to compute ID.
as well graded where there is a wide range of particle
sizes or uniformly graded where all the particles are
within a narrow range. The determination of the particle 7.3.3. Shear strength tests
size distribution of a representative sample by sieving Triaxial testing equipment is commonly used to measure
will be a major task where the fill contains large rock shear strength. Drained compression tests are often the
fragments. simplest tests to perform and most closely resemble field
The percentage of silt and clay size particles (i.e. finer conditions, but in some situations undrained testing may
than 0.06ram) is important as, when this percentage be required. Strain rate should not greatly affect the
is high, the rock fill will cease to perform as a granular measured strength. Figure 7.1 shows the deviator stress
soil. The critical magnitude of the fines content is likely plotted against axial strain for a drained triaxial
to be in the range of 10% to 25% by weight with regard compression test on a heavily compacted rock fill at
to shear strength and 10% to 15% with regard to low confining pressure.
permeability. The direct shear test is appropriate for testing free
draining granular soils. However, the failure plane is
Moisture content very closely defined in the test and an unrepresentatively
Water is invariably present in fill materials; the amount high value of shear strength may be measured for a rock
of water, expressed as percentage by mass of the dry fill due to the failure plane being constrained to pass
solid particles, is termed the moisture content (w). through an artificially large number of rock particles.
BS 1377:1990 Part 2 describes the definitive method of Most field conditions will approximate to plane strain
moisture content determination by oven drying at 105~ rather than the axisymmetric stress conditions in the
to 110~ As rock fills usually have a very wide range of triaxial test. Plane strain testing equipment also has been
particle sizes, the moisture may be very non-uniformly used to measure drained shear strength parameters of
distributed throughout the fill. rock fills but the equipment is not widely available.
As the shear strength of rock fill is a function of both
Compactness density and stress level, it is essential that test samples are
The compactness of a fill is a function of method of compacted to an appropriate density and sheared under
placement and the subsequent stress history. The follow- an appropriate confining pressure if strength parameters
ing parameters are related to the compactness of the fill: measured in the laboratory are to be applicable to the
9 voids ratio (e) field situation. When triaxial tests are carried out to
9 porosity (n) derive shear strength parameters for a limit equilibrium
9 dry density (Pd) slope stability analysis, values of cell pressure should
9 dry unit weight (Ta) be used such that the stresses in the triaxial test cor-
9 air voids (V~0 respond to the range of stress which will be encountered
9 density index (Io) on potential critical failure surfaces in the embankment
slopes. Figure 7.2 shows a curved failure envelope
These parameters are closely related as the following typical of a heavily compacted rock fill.
relationships demonstrate:
9 n = {e/(1 + e ) } 100(%)
9 pa = {ps(100 - n)}/100 7.3.4. Compressibility tests
9 ~/d = Pdg Rock fill compressibility is most commonly measured
9 Va = {| - (Pd/Pw)[(Pw/Ps) + (W/100)]} 100 (%) under conditions of one dimensional compression in a
where Pw is the density of water (1.0 Mg/m3), w is the large oedometer as the test has a number of advantages:
moisture content (%), ps is the particle density (for many 9 the deformation of a rock fill structure due to a change
rocks 2.6 to 2.7Mg/m3), g is the acceleration due to in stress or moisture content can often be related to the
gravity (9.81 m/s2).
behaviour of the rock fill measured in the test
The density index (ID), which is sometimes termed 9 the test is relatively simple to perform even at a large
relative density, is a useful way of relating the in situ dry scale
density of a granular fill (Pd) to the limiting conditions of 9 it is easy to carry out creep tests at constant applied
maximum dry density (Pd . . . . ) and minimum dry density load
(Pdmin).
A considerable number of large oedometers tests has
(Pd ~ Pdmin)Pdm ) been carried out on rock fill materials (Kjaernsli & Sande
ID = (Pdmax-- Pdmin)Pd X 100 (%) 1963; Sowers et al. 1965; Fumagalli 1969; Marsal 1973,
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

264 ROCK I~ILL

Penman & Charles 1976). The rock fill is compacted in show a dramatic increase in discontinuity dilation which
layers with an electric vibrating hammer to the field can be the principal cause of the resulting instability.
density. The one dimensional compression behaviour Standard seismic refraction techniques (Matheson
of most rock fills on first loading is non-linear and so 1985) have been used successfully to determine the
it is important to determine the compressibility over a extent of the blast disturbance. This application depends
range of stress that is applicable to the field situation. on a low velocity layer being formed in the disturbed
Figure 7.3 shows the compressibility of a number of rock with its dilated discontinuities in contrast to the
heavily compacted rock fills as measured in a one metre relatively high velocity layer in the natural, undisturbed
diameter oedometer. rock mass with tight discontinuities and acoustically
In some cases the unloading and reloading stress strain coupled rock blocks. The width of the zone of dis-
relationships can be of interest. This will be the case for turbance can vary from two to four metres with smooth
deformations in the upstream rock fill of a central core and trim blasting, and from six to eight metres with
dam due to reservoir drawdown and refilling. The stiff- bulk blasting. Increasing instability in the final face is
ness is much greater than on first loading. associated with increasing widths of disturbance. There
Poorly compacted unsaturated rock fill is likely to be is no disturbance in most natural slopes and those
susceptible to collapse compression when first inun- formed by presplit blasting.
dated. The effect of inundation on an unsaturated rock Closed circuit television (CCTV) techniques have been
fill can also be examined in an oedometer test by used to measure directly the nature and extent of the
flooding the sample at an appropriate stress level. disturbance and to verify the findings of the seismic
refraction profiling (Gunning 1992). A skid mounted
camera is inserted into holes drilled into the rock face
and the depth of insertion is measured using a tape
7.4. Excavation attached to the camera head. Observation and video
7.4.1. General recording are performed simultaneously through either a
forward or a side viewing head. The discontinuity posi-
At present excavation of strong rock is accomplished tions and dilations can be determined from a detailed
economically by blasting and weak rock is excavated by scan of the borehole wall. Results verify that dilation
machine. Modern explosives used in rock excavation of discontinuities is the main feature of the zone of
by blasting almost entirely fall into the high explosive disturbance measured by seismic refraction techniques.
category and can be divided into gelatins, emulsions, The choice of excavation method involves predicting
slurries and ANFOs (Ammonium Nitrate/Fuel Oil the damage that may be induced, then minimizing the
mixtures) in approximately decreasing velocity of deto- damage to the final face whilst ensuring that the rock is
nation. Black powder is the only surviving example of a fragmented sufficiently to allow mucking out. The two
low explosive which deflagrates rather than detonates methods commonly used are presplit blasting and frag-
and, although used extensively for rock before the early mentation blasting, but only presplit blasting appears
20th century, finds no application in modern highway to eliminate damage entirely.
rock excavation, see w
Ripping is used on weaker materials which can be
broken up using steel tynes or rippers mounted on a 7.4.3. Presplit blasting
crawler tractor. The material is then scraped or dug This form of controlled blasting is designed to minimize
ready for loading and transportation to the fill site. damage in the final face (Matheson 1983, 1986). Closely
As well as the need for an excavation to produce rock spaced holes drilled along the proposed design slope are
of a grading suitable for use as a fill material, there relatively lightly charged. The presplit charges are deto-
will also be a requirement in an engineering application nated simultaneously forming discontinuities along the
such as a highway cutting that the excavation itself design slope. Bulk blasting techniques are then used to
remains stable for a long period of time. The follow- fragment and loosen the rock in the excavation area.
ing sections on excavation by blasting are based on The presplit discontinuities limit disturbance arising
Matheson (1995). from the bulk blasting and leave the rock forming the
design slope in as near as possible its natural condition,
thus minimizing induced instability in the final face, see
7.4.2. Instability induced by excavation using explosives
w A presplit face is shown in Fig. 7.6.
~onsiderable damage can be caused to a rock slope by The type, charging and detonation of the explosives
the uncontrolled use of high explosives. In civil engineer- used in presplitting require careful control. Hole diam-
ing applications the major effect of blasting is one of eter, hole spacing and explosive charging have proved
increasing disturbance of pre-existing natural disconti- to be fundamental, dependent variables which should
nuities. Presplit faces show little change from those of not be varied unnecessarily. The explosive charges must
natural exposures unlike bulk blasted faces which tend to be of sufficient strength to ensure that presplit fractures
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK PILL 263

Fig. 7.6. Presplit face.

are formed between boreholes and these fractures must most surface excavations where precision drilling is
be in existence before the disturbance caused by the required. Its use has considerably increased the accuracy
bulk blasting reaches the proposed design slope. There of drilling presplit holes and it is usually specified on
must, therefore, be a delay between the detonation of trunk road projects.
the presplit and the bulk charges and a lower limit Successful presplit faces have been formed in a wide
of 50 milliseconds has proved satisfactory. There is no variety of sedimentary, metamorphic and igneous rock
upper limit and time intervals of days and weeks are types, ranging from intensely fractured phyllitic schists
commonly used. to block fractured granitic intrusions. Individual presplit
Accurate drilling is vital and the setting up and close lifts of up to 27 m and presplit faces up to 62 m high are
monitoring of the initial stages of drilling have been being used with great success. Nevertheless, caution is
found to be extremely important. The orientation of necessary in the use of presplit blasting as the technique
drill rods and drill holes can be defined in terms of a dip alone is not the answer to all slope stability problems.
(the angle from the horizontal) and an azimuth
(direction of dip). Deviation of drill holes with depth
7.4.4. Fragmentation blasting
can prove to be a problem, even when the initial
alignment has been accurately set. Deviations tend to Fragmentation blasting is employed to break up the
relate to the type of drill and the drilling technique, blocks naturally present within a rock mass to a size
rather than the rock conditions. However there is a consistent with the intended use. In highway cutting
tendency for drills to deviate towards the normal of any excavations, after blasting the blocks must have a size
hard and soft layering that exists in the host rock. distribution that is able to be easily loaded and
A simple drill orientation device has been developed to transported by mechanical plant out of the excavation
help accurate alignment of drill rigs and measurements area and, if required as fill, must be within the specified
of drill hole orientation (Matheson 1984). The design of grading. The degree to which fragmentation occurs
the device makes it largely independent of rig type and when bulk blasting depends on the range of block sizes
drill rod diameters and the accuracy is adequate for originally present in the rock mass, the natural block
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

266 ROCK I*ILL

size distribution (NBSD) and the blasting technique. which will fragment the rock mass to a BSD consistent
This bulk use of explosives is in complete contrast to with efficient loading and transportation, and a size
presplit blasting which employs carefully controlled range suitable for use as rock fill if required. However,
blasting to form the design slope, see w these methods of prediction need to be fully tested in
the field.
The techniques of minimising natural and induced
7.4.5. Ripping
instability on a rock face have been incorporated into the
Ripping is an inexpensive method of providing rock fill Optimum Design Technique for Rock Slopes (Matheson
where the in situ material is discontinuous or relatively 1992). This dual approach is essential as instability is
soft. A crawler tractor is fitted with one or more steel cumulative and one method cannot compensate for
tines or rippers which breaks up the rock as the tractor deficiencies in the other. For example, presplit blasting
moves forward. The output of the ripper will depend not cannot compensate for a rock slope design orientation
only on the material it is working but also on the power out of harmony with the host geotechnical conditions.
of the machine and the number and dimensions of the The objective is to improve whole life costs by increas-
tines, see w The fragmented material is removed ing stability and reducing costly remedial work and
by a bulldozer and loaded by a backactor onto a dump maintenance.
truck for transport to the fill area. The output of a ripper Whether to rip or blast will depend on the intact rock
is, therefore, also influenced by the capacity of the strength, the spacing of fractures and the abrasiveness of
bulldozer. Inadequate capacity for the backactor and the rock. These different parameters have led to classi-
dump trucks can be overcome by stockpiling material. fications based on seismic velocity (Caterpillar Tractor
However, it is good practice in this type of construction Company 1982) or point load strength and fracture
to match the different types of plant for excavation, spacing (Pettifer & Fookes 1994), see also w
haulage, placement and compaction to ensure maxi- Seismic velocity is the velocity at which a ground-
mum efficiency. borne vibration, usually the longitudinal or compression
wave component, travels through a rock mass. The
velocity is dependent on the stiffness of the rock and in
7.4.6. Assessment of excavatability by blasting or ripping
particular the number, spacing and width of disconti-
If a rock mass could be excavated without introducing nuities. The higher the number of discontinuities the
new discontinuities then the distribution of sizes slower the wave will travel. The measurement of seismic
characteristic of the volume of the excavated rock
would identical with the Natural Block Size Distribution
(NBSD) within the rock mass. The NBSD is analogous
to the Particle Size Distribution (PSD) of a soil. The
actual Block Size Distribution (BSD) in the excavated
material depends on both the original NBSD and also
the method of excavation, see w In strong rock
having a coarse NBSD explosives are required to reduce
blocks sizes to a level at which they can be handled by
mechanized plant and, if required, used as rock fill.
If this is not accomplished then secondary breakage of
the larger rock block will be required which is time
consuming and expensive.
Techniques of estimating both the NBSD and the
BSD have been developed (Wang et al. 1992) which
rely on the use of field data on the spacing and orienta-
tion of discontinuities within the area to be excavated.
The data can be obtained by extrapolation from out-
crop mapping, from scanlines on surface exposures, or
from CCTV scanning of holes drilled into the rock
mass. BSD can be estimated using either the Bond-Ram
model (Bond 1959; Cunningham 1982) or the Kuz-
Ram model (Kuznetsov 1973). Both models have been
incorporated into computer programs that allow the
effects of varying the blast pattern and explosive
charging to be estimated.
In theory the highway engineer is able to calculate the Fig. 7.7. Estimated ripper production (after Caterpillar Tractor
NBSD of a rock mass and choose a technique of blasting Company 1982).
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK HLL 267

velocity is non-destructive, fast and relatively straight- The discontinuity spacing index is the mean number
forward. Consequently its use in excavation has been of fractures per metre of a line, known as a scan line,
promulgated by excavation equipment manufacturers, across the site. A three-dimensional discontinuity spa-
notably the Caterpillar Tractor Company, as a means cing index, calculated on a unit volume basis or from
of quickly assessing the rocks mass characteristics actual spacings, should be used wherever possible since
and hence the best use of their plant. Figure 7.7 is this gives a realistic impression of average block size.
an example for one particular type of ripping plant, Figure 5.21 does not consider other factors such as plant
a Caterpillar D10 crawler tractor, and shows how the selection and cost and is purely based on geotechnical
assessment of the productivity of ripping is made. property considerations. Also it is stressed that predic-
Figure 7.8 illustrates another method by which seis- tions made using the graph need to be compared with
mic velocity and/or rock type can be used to assess the performance actually achieved and revisions will
productivity for the same machine. be required as more powerful excavation equipment
Graphical methods are particularly useful for rapid become available.
assessment of excavatability during initial site surveys or A statistical approach based on productivity has been
where test data are not available. The graphical method used in Australia (MacGregor et al. 1994). A large
of Pettifer & Fookes (1994) as shown in Fig. 5.21 is database of information on geological, ripping and
particularly suited for assessing whether blasting, rip- laboratory data has been interrogated and regression
ping or hard digging is required. In particular it gives lines plotted through the data. Correlations have been
guidance on the type of excavation plant that is needed calculated which allow a prediction to be made of the
for ripping. The assessment is based on the point load likely productivity of ripping plant. The calculated
index and the discontinuity spacing index, which are the productivity is used for assessing the output of the
two geotechnical properties which most effect excavat- site and determining whether or not ripping is eco-
ability. The point load test is explained in Appendix B. nomically feasible.

Velocity in metres 0 1 2 3 4
per second x 1000 I i [ I , I , I
Velocity in feet
3er second x 1000 0 1 2 3 4 5 6 7 9 10 11 12 13 14 15
i
Glacial till
Igneous
Granite
Basalt
Trap rock
Sedimentary
Shale
I
Sandstone '
Siltstone
Claystone
Conglomerate
Breccia ! !
Caliche
Limestone
Metamorphic
Schist
I I
Slate
Mineral & ores
Coal ,
Iron ore
I I I I
Rippable Marginal I I Non-rippable F////////M
Fig. 7.8. Multi or single shank D10 ripper performance (after Caterpillar Tractor Company 1982).
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

268 ROCK FILL

7.4.7. Treatment of instability of cutting slopes Increasing use is being made of stainless steels for similar
Treatment of instability should only be required where it reasons. Although extended claims regarding stress
has not been possible to ensure stability by design, or retention and corrosion protection are made, there is
where there have been deficiencies in the design process, considerable doubt over the ability of some systems to
sustain high levels of stress over their design life and
or poor construction practice. General guidance has been
given by Fookes (1976). In highway rock engineering to provide the expected degree of corrosion protection.
four basic techniques are available; removal of unstable Often strengthening techniques appear to be used for
rock, strengthening of the rock forming the face, support peace of mind rather than as primary aids to stability.
of the rock face, and containment of falling rock. In
extreme cases reprofiling of the slope may be necessary.
7.4.7.3. Support of rock face
If key areas of instability are given support early enough
The use of any particular treatment technique depends
then face stability can be maintained, and upward
mainly on the nature and scale of the instability, the geo-
progressive failure of a rock face is halted, see w
technical conditions present, the geometry of the slope
The type and extent of the treatment depends on
area, and the risk to road users. Techniques of hazard
whether material is being lost through weathering of
and risk assessment for slopes have been described by
the minerals in the rock or through block fall, or both.
McMillan & Matheson (1997, 1998).
Weathering can be arrested by placing a protective layer
of sprayed concrete, mass concrete or masonry on the
7.4.7.1. Removal of unstable rock face; the rock is then protected from weathering and the
Where practical, the first approach to instability in a face supported against further failure. The extent of such
rock slope is to consider removal of the unstable rock treatment can vary according to the geotechnical
using either hand scaling, mechanical plant, or the local circumstances from small local areas of 'dentition' to
use of small explosive charges. Removal is low in cost, that of the entire face.
usually fast to perform, and can be totally effective in the Progressive deterioration through block fall can
short term. However, problems can arise if block sizes or sometimes be prevented by providing judicious support
volumes are large, access difficult, or the risk particularly before the volume involved in the failure becomes
high. All techniques require careful consideration of too large. The loss of key areas on a face can be
slope stability, geotechnical conditions and an evalua- compensated for by installing support buttresses, walls,
tion of the implications in the long term. or beams. These generally are of concrete with a local
masonry tinish and are usually keyed in or pinned to the
7.4.7.2. Strengthening of rock .face rock mass using dowels or bolts. The technique is
In general strengthening techniques are expensive and, particularly effective at treating small areas of failure,
in some situations, of questionable cost-effectiveness. where further loss of material would increase the risk of
Their use tends to be confined to circumstances where larger failure occurring. Where the techniques do not
other methods cannot be employed and where block involve bolting they tend to be inexpensive and effective,
sizes are relatively large. Where unstable rock blocks and can easily be made to blend into the rock face.
cannot be removed without adversely affecting the A comprehensive description of available rock support
stability of the rock slope, techniques of strengthen- systems is given by Hoek & Brown (1988).
ing the face can be employed. All situations require a
careful analysis of the geotechnical conditions, the 7.4.7.4. Containment of falling rock
possible failure mode, the most appropriate technique The principle of containment involves allowing a certain
of strengthening, and the most effective method of level of rock fall to occur but preventing such falls from
installation. being a danger to road users while in the highway
In highway projects, the common strengthening cutting. The extent of rock removal, strengthening, or
techniques are rock bolts, anchors and dowels. With support required for full stability is thus often con-
rock bolts, holes are drilled by hand or machine, bolts siderably reduced, and costs are minimized. Techniques
inserted, an anchorage established, the bolts stressed vary according to the block sizes involved, the profile of
to their working load, and the holes back grouted to the rock face, and the geometry of the side slope. They
prevent corrosion. Depending on the situation, bolts can include netting of the face to contain falls of small sized
be from 3 m to 10 m in length, and be either anchored at material near the face, rock traps, fences, barriers, and
a particular point or continuously bonded. Anchors are earth or rock bunds to prevent larger blocks or volumes
similar to rock bolts but generally longer, composed of of rock reaching the carriageway. In using such tech-
multiple steel strands, and are invariably anchored at a niques maintenance is not entirely avoided and regular
particular point only. Dowels are generally short, and checks on the integrity of netting and fences, and
are continuously bonded to the host rock, see w periodic cleaning of debris filling in the area between the
A long design life is required for highways, up to 120 face and containment barrier, are still required. The
years, and so double corrosion protection is specified. lower cost of such containment measures compared to
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK I~ILL 269

that which would be required for removal, strengthening the surface of which consists mainly of the finer
and support in a full treatment of face instability, make material (Williams & Stothard 1967). The spread-
them increasingly attractive. Although rigorous engi- ing of a separate layer of fine material on the
neering techniques are available to design containment surface of the rock fill should be avoided as it could
structures, considerable subjective judgement is needed substantially reduce the effective depth of compac-
to estimate both the size and timescale of the failures. tion by vibrating rollers (Forssblad 1981).
An evaluation by a qualified and experienced engineer- (b) The fines mixed with the coarser material must be
ing geologist is an essential requirement. non-cohesive so as to be capable of being vibrated
or washed through the layer. The presence of
7.4.8. Environmental impact cohesive fines makes the achievement of a satisfac-
tory state of compaction extremely difficult.
Increasing emphasis is being paid to the environmental (c) A heavy vibrating roller with a large amplitude of
impact of highway construction. Attention is being vibration is required to ensure that vibrational
focused on rock engineering works to reduce environ- stresses extend to the bottom of the thick layer.
mental impact and make them aesthetically acceptable. Towed vibrating rollers or self-propelled single roll
Pressure is also being exerted to design and to construct vibrating rollers are generally most appropriate for
remedial work such that it blends into the local this purpose. Forssblad (1981) recommended
countryside. Unsightly concrete support walls, conspic- amplitudes of 1.5 to 2 m m and frequencies of 25
uous bolts, or obvious netting in full view to road users to 30 Hz for rock fill compaction. The shell of the
are becoming unacceptable where alternative options vibrating roll must be of high grade steel to
exist. Road alignments, side slope profiles, the appear- accommodate the large stresses that arise during
ance of rock faces, and the conspicuousness of remedial the vibratory compaction of rock fill.
treatments can all be engineered with existing techni-
ques. Cosmetic issues can be considered at both the Prior to the development of the heavy vibrating roller, it
design and the construction stages, but are likely to add was common practice to sluice the fill with an excess of
considerably to the costs of highway construction. water from high pressure hoses so that the water passing
A similar situation arises in the case of blast vibration downwards through the layer conveys fines until the
which, although seldom producing measurable damage, voids were filled. This process could cause damage to
is often intrusive to those living nearby, see w the surrounds, however, and could be detrimental to the
long term performance of the rock fill embankment if
the foundation soil was cohesive and so weakened by the
7.5. Placement and compaction increase in moisture content.
When rock fill is relatively uniformly graded and
7.5.1. General completely devoid of fines, it can be assumed to be
already in a well compacted condition as placed, in that
Rock fill is a very coarse material containing particles inter-granular contact will prevent any further rearran-
with maximum dimensions typically of about 500ram gement of the particles. Material with these properties is
and, where a fill has both these large particle sizes and in often used for drainage layers which are placed at the
addition fine material to fill the interstices between the bottom of fills, particularly where large horizontal flows
coarser particles, only certain types of compaction plant of water are expected, e.g. on flood plains.
will be effective in attaining the highest possible density Although engineered rock fills are normally selected,
in the layer. The thickness of the layer placed will be placed and compacted to minilrfize settlement and
dictated by the maximum particle size; normally it is ensure adequate bearing capacity, there are certain
considered that the thickness should be such that the situations where this may not be feasible and occasion-
maximum particle size is no more than two thirds of that ally there may be applications where it is not necessary.
thickness (Forssblad 1981). For the interstices to be For example, end dumping has to be used to fill below
filled by migration of the fines downwards through the water level. The rock used in this situation should have
layer, a number of conditions need to be fulfilled: the following characteristics:
(a) The rock fill should be spread so that the coarser 9 it should be durable, especially resistant to degrada-
material does not accumulate on the surface of the tion by wetting and drying cycles (protection against
layer. Experience has shown that this is best wave action is provided by armourstone)
achieved by dumping the rock fill on top of and 9 it should have a coarse grading with very little fine
behind the edge of the advancing fill surface and material which would tend to be removed by the water
using a heavy bulldozer to push the material
beyond the edge. This causes the larger particles In some cases it may be necessary to densify the end
to roll to the bottom of the layer and fines to fill the dumped rock fill by some deep compaction technique
voids between them, thus creating a dense layer, such as dynamic compaction. This can be carried out as
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

2'J0 ROCK IqLL

soon as a platform of rock fill has been established deteriorate. A compaction-degradation test (Hale
above water level. Above this level the rock fill can be et al. 1981) is used to indicate the degree of difficulty
placed and compacted in layers as an engineered rock with which this would be accomplished. When the
fill. In remote areas of the world where a less demanding shale strongly resists efforts to break it down, special
standard of performance can be accepted, there may wetting and heavy rolling procedures may be required.
be circumstances in which it is considered that end 9 Hard, durable shales can be treated as rock and
dumping will provide a technically adequate solution. compacted accordingly.
In a review of laboratory test methods aimed at
7.5.2. Low grade rock fill classifying shales according to their potential breakdown
Embankments are often constructed of materials which during and subsequent to embankment construction,
are intermediate in character between rock fill and earth Bailey (1976) discussed problems of site control. It was
fill and which are variously described as low grade recommended that shales that are resistant to mechan-
rock fills or soft rock fills. Quarries may yield large ical breakdown but are prone to weathering, leading to
quantities of such materials before sound, strong rock collapse problems when contained within an embank-
fill is reached and there are clear economic advantages ment, should be degraded before reaching the fill area
if such materials can be used. and undergoing compaction. Thus, a more satisfactory
state of compaction can be achieved than would be
Shale the case with larger aggregations of clay particles.
Clay shale has characteristics between those of a clay On occasions water has been added to the shale to aid
soil and intact rock. Although initially intact and often breakdown and improve compaction.
difficult to excavate and handle, the dominant clay Bailey (1976) summarized experiences in the United
mineral makes shale sensitive to the effects of water States with construction trials involving shale. Such
and initially rock-like material can degrade on exposure trials have served primarily as compaction equipment
to wet weather to a very weak plastic clay. The extent to trials and often have provided a measure of the vari-
which clay shales weather varies considerably, however, ability in the behaviour of the shale. It was recommended
and the variations in their properties cause problems in that only one factor should be varied at one time, and
selecting compaction specifications which ensure that sufficient test samples taken to allow for the variability
adequate densities are attained. The two principal tests of the material and the imprecise control of the compac-
used in the United States for durability classification tion energy. No conclusion was possible regarding the
of shales, listed by Strom (1980), are the jar slake test relative merits of method or end product compaction
(based on observed behaviour of shale when soaked in specifications for shales, or what the end product should
water) and the slake durability test (measuring the be except in very general terms. Experience with com-
degradation when the shale is placed in a wire-screen paction trials on shales in Indiana has indicated that the
drum rotated under water). The latter test is specified following factors should be considered:
as ASTM Test D4644, American Society of Testing 9 shale type and initial condition
and Materials (Annual publication) and described by 9 moisture content and its variability
Brown (1981). 9 type and size of compactor
The classification tests can be used to divide the shales 9 number of passes
into three main categories, (Strom 1980): 9 layer thickness and the total number of layers
9 Soft non-durable shales are excavated and treated as 9 foundation material type and stiffness
cohesive soil. Sufficient breakdown of the aggrega- 9 weather conditions
tions usually occurs during the normal earthmoving Currently, in the United Kingdom, the use in highway
operations so that the material is compacted in thin construction of shales and other argillaceous rocks as
layers to typical requirements for soil compaction. rock fill for starter layers is prohibited ( M C H W 1 1994).
9 Hard, non-durable shales need to be compacted to the A starter layer forms the base of an ernbankment where
requirements for soil but, to carry out the compaction construction is over soft ground and creates a stable
processes effectively, additional operations are needed platform on which to construct and support overlying
involving the breakdown, removal or separate exca- weaker embankment layers. Coarse granular material
vation of the coarser, hard particles that would be for use as general bulk earthworks rock fill in the U K is
incompatible with spreading in thin layers. Materials required to have a 10% fines value in excess of 5 0 k N
in this category also include weak, non-durable shale and this value is likely to preclude the use of most, but
interbedded with rock layers such as limestone. Thus, not all, shales.
hard but nondurable shale has to be intensely
degraded during excavation, placement and compac- Rock fill with cohesive fines
tion and finally densified to a specification appro- If significant quantities of cohesive fine material are
priate to the soil type to which it would eventually associated with large rock particles it is impossible to
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK FILL 271

achieve a migration of the fines downwards through the placement trials prior to commencement of the main
layer by the use of vibration. The alternative of sluicing earthworks. As an example, rock excavation and com-
with water would not be successful because of the paction trials were carried out as part of the preliminary
cohesive nature of the material. Thus it can be expected investigations into building the Pennine section of the
that the cohesive element of the fill can only be com- M62 motorway which included the 73m high Scam-
pacted in relatively thin layers as per normal compaction monden dam (Williams & Stothard 1967). As a result of
practice. As the presence of the large particles precludes the trials which took place in 1964, a specification was
the use of such thin layers, a satisfactory compaction written which included detailed grading limits for the
could not normally be carried out. The only solution is rock fill, see w
to remove the coarse particles such that thin layer A trial to determine the procedures for the con-
compaction can be achieved. The type of compaction struction and control of compaction of rock fill for
plant in use will determine the thickness of layer that can embankments on the Rabat to Casablanca motorway is
be compacted, and the maximum particle size remaining described by Jakani (1987). Rock blasting trials were
in the fill should not exceed two-thirds of the layer included to establish a drill-hole pattern that yielded an
thickness. The maximum thickness of stony cohesive appropriate grading with minimal amounts of exces-
material that may be compacted by a vibrating roller is sively large particles. The material was spread in layers
275mm, provided that the vibrating roller has a mass 800 mm thick by a heavy bulldozer which was also used
per unit width of vibrating roll exceeding 5000kg/m. to tow the heavy vibrating roller for compaction.
With such a combination of machine and layer thickness Various methods of compaction control were tried:
the maximum particle size allowable in the fill material
would be (2) x 275 = 183 mm. 9 plate bearing tests with a 600 mm diameter plate were
Forssblad (1981) took a different view, however, in used but the results were too scattered for a reliable
discussing the compaction of morainic soils based on interpretation to be made
experience in Sweden. These materials often contain 9 in situ density measurements were also made but the
stones with a very large particle size and include a tests were time consuming and, because of the large
significant clay fraction; Forssblad contends that com- particle sizes, it was impossible to establish a
paction in thick layers is much more economical than laboratory maximum dry density
compaction in thinner layers involving the removal of 9 levels were taken of the surface layer to deduce
the large stones. Morainic soils have been compacted whether or not there was any significant settlement
successfully in Sweden in layer thicknesses of 0.7 to but none was registered
1.0m using heavy vibrating rollers with a mass on the
vibrating roll of 10 to 15 t. Such large vibrating rollers It was concluded that careful planning of the rock
are uncommon in the United Kingdom but where the blasting would be required to ensure the most effective
scale of the work is sufficiently large and the fill grading of the fill material and that a heavy vibrating
materials are similar in nature to morainic soils their roller with a roll loading of more than 3500 kg per metre
use could well be justified. width should be used. The compaction should be to the
In France, compaction of large rock particles with method specification developed in France for soils
cohesive fines is included in the specifications for (Ministere de l'Equipement et al. 1992).
earthworks construction, Ministere de l'Equipement
et al. (1992). The material is considered unsuitable for
use as fill if the fines are excessively wet (strength 7.5.4. British practice for specification and control
too low) or excessively dry (compaction too difficult). Because of the large particle sizes involved and the
At intermediate moisture conditions compaction is thicknesses of layer that these sizes necessitate, the com-
considered viable using either vibrating rollers, pneu- paction of rock fill is normally to the requirements of a
matic-tyred rollers or tamping rollers. Specified layer method specification, that is, the type of compaction
thicknesses vary from 200ram to 650mm depending plant, the maximum thickness of layer and the minimum
upon the material class and the type of compactor; number of passes are laid down. Where the in situ state
however, the maximum particle size is restricted to two- of compaction has to be determined by achieving a
thirds of the layer thickness and this will obviously often specific end product requirement, large scale replace-
dictate the use of a thicker layer with a commensu- ment tests may be performed. These involve the exca-
rate increase in the size of compactor. vation of large holes with depths equal to the thickness
of the compacted layer and of sufficient volume to
ensure that representative samples of the fill material
7.5.3. Trial fill placement are obtained. All material removed from each excavation
has to be weighed and its moisture content determined.
Where large quantities of rock fill are to be placed, it The volume of the excavation is determined either by
will usually be advisable to carry out excavation and lining the hole with a plastics-sheet and determining the
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

272 ROCK I~ILL

Table 7.1. Compaction of rock fill for highway works in the of layer and numbers of passes for the limited number of
United Kingdom using single roll vibrating roller categories of vibrating roller specified in the United
Kingdom for highway works.
Mass per metre width Maximum final thickness Minimum Information about rock fills used in some UK
of vibrating roll of compacted layer number of embankment dams is presented in Tables 7.2 and 7.3.
( kg ) (ram) passes
Table 7.2 illustrates embankment dam practice in rock-
up to 2300 unsuitable fill placement in terms of layer thickness, the weight of
over 2300 up to 2900 400 5 the vibrating roller (VR) and the number of passes.
over 2900 up to 3600 500 5 Table 7.3 gives an indication of average porosity (n) and
over 3600 up to 4300 600 5 air voids (Va). It was not possible to include all the
over 4300 up to 5000 700 5 relevant details in these two tables. The layer thickness
over 5000 800 5 and method of compaction may have been changed
during construction of the embankment, some water
may have been added to the rock fill and there may have
volume of water required to fill the cavity to surface level, been more than one type of rock fill used in the
the water-replacement method (British Standards Insti- embankment. Furthermore, some of the quoted density
tution 1990), or by determining the weight of a fairly results may be less reliable than others, being based on
single-sized granular material, pea gravel for example, far fewer test data or incomplete information.
required to fill the cavity and converting the weight to a In addition to information on achieved densities in
volume from a separate calibration of the material in heavily compacted rock fills in embankment dams, some
a container of known volume (Williams & Stothard information has been provided in Table 7.3 for two rock
1967). The mass of material involved can amount to fills used as opencast mining backfills. At Horsley the
several tonnes for the normal type of rock fill and backfill consisted mainly of mudstone and sandstone; at
the laborious nature of the operation discourages its Tranent of mudstone, siltstone and sandstone. Neither
application other than in trials or research situations. fill was systematically compacted. The densities were
The more usual way of controlling the compaction of determined from small samples and can only be
rock fill is by a method specification. The current a crude guide to the condition of these two very vari-
specification for highway works in the United Kingdom able fills.
( M C H W 1 1994), requires the use of single-roll vibrating The five dams had porosities in the range 15% to 27%
rollers for compaction of rock fill to ensure that a high and this could be regarded as typical for modern well
amplitude of vibration is achieved. Only the heavier graded heavily compacted rock fills. In contrast the two
machines available for civil engineering works are opencast backfills had much higher porosities, 33% and
approved for such work. Table 7.1 gives the thicknesses 37% respectively.

Table 7.2. Field placement of rock .fill (Charles 1991)

Dam Date Rock.fill Laver (m) Pass no. VR Wt tonne

Balderhead 1965 shale 0.8 2 8.5

Scammondon 1969 sandstone/mudstone 0.9 5 11.5
Llyn Brianne 1971 mudstone 0.9 4 13.5
Winscar 1974 sandstone 1.7 4 13.5
Marchlyn 1979 slate 1.0 4 13.5

Table 7.3. Field density of rock.fill (Charles 1991)

Location Structure Pa (Mg/m3) w (%) #s (Mg/m~) n (%) V, (%)

Balderhead embankment dam 1.94 9 2.66 27 10

Scammonden embankment dam 2.02 7 2.69 25 I1
Llyn Brianne embankment dam 2.35 3 2.75 15 7
Winscar embankment dam 2.03 6 2.60 22 10
Marchlyn embankment dam 2.25 4 2.81 20 11
Horsley opencast backfill 1.70 7 2.54 33 21
Tranent opencast backfill 1.54 9 2.45 37 23

[On= dry density of fill, Ps = particle density, w = moisture content].

 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK FILL 2'/3

7.5.5. Other national practice for specification thoroughly wetted, and then compacted with heavy
and control equipment. Finer material must be applied to the top of
the layer being compacted to fill any voids.
The previous sub-section has dealt with UK practice and
the situation in some other countries is now briefly
South Africa
reviewed.
Specific mention of rock fill procedures is made in
National Institute for Transport and Road Research
Germany (1982) where it is recommended that the rock is spread
A useful description of the requirements for rock fill is by bulldozing to its correct position to avoid arching
given by Rosenheinrich (1978). Rock fill has to be placed within the dumped material. Rock should not exceed a
evenly over the whole area to be filled and only well maximum dimension of two-thirds of the layer thick-
graded material should be placed in the upper 1 m of ness. Materials which fracture easily or weather readily
the fill. The largest particle size used should not exceed should be grid-rolled in order to break them down to
two-thirds of the allowable layer thickness. Minimum satisfactory dimensions for the layer, since this may
requirements for compaction of rock fill is in the form of otherwise result in loose zones within the layer.
an end-result specification and is expressed in terms of
relative compaction with the normal Proctor test (2.5 kg Belgium and France
rammer method) as the laboratory standard. Minimum The compaction of embankments in rock fill poses
relative compaction values vary from 100 to 103% for particular problems, according to Centre de Recherches
material placed within 0.5m of the formation level, Routieres (1978), because of the difficulties of determin-
and 95 to 97% for material placed below that level. ing whether the required density has been achieved.
The actual values within these ranges depend on the The design engineer in Belgium is advised to specify the
classification of the rock fill. thickness of layer to be placed, the type of compactor
The difficulties of in situ measurements of density and and the number of passes. This form of specification is
laboratory compaction tests with coarse granular similar to that used in France (Ministere de l'Equipe-
material has led to an additional specification in terms ment et al. 1992).
of the results of platebearing tests. The determination is
made of E2, the modulus of deformation on the second
loading cycle of the plate. Escario et al. (1980) give the
equation for the determination of the modulus of 7.6. Case studies
deformation as follows:
E2 = 0.75d(Acr/As) Six case histories are presented to illustrate the
performance of rock fill.
where d is the plate diameter, normally 300 or 600 mm,
A a is 0.7Crma x --0.30"max, O'max is the maximum load
applied and As is the increment of deformation between 7.6.1. Scammonden dam
0.3Crmax and 0.7O-max. The specified maximum thickness Construction of the 73m high rock fill embank-
of layer depends on the type of material to be compacted ment commenced in 1967 and was completed in 1969.
and on the required level of compaction. Dynamic types The embankment, which is located some 10kin west
of compaction are regarded as the most effective, of Huddersfield, is a dual purpose structure, forming
particularly vibrating rollers and vibrating-plate com- both a motorway embankment and an embankment
pactors, with dynamic forces sufficiently large to transfer dam. A cross section of the embankment is shown in
the smaller particles such that they fill up the interstices Fig. 7.9. The wide crest carries the M62 Pennine motor-
between the larger particles. It is also considered most way and an upstream sloping soft clay core forms
effective to compact fills from the outside towards the the watertight element of the dam. Reservoir impound-
middle, with the compactor moving at a steady, uniform ing commenced in 1969 and the reservoir was full
speed, avoiding excessive speeds of travel which may in 1972.
result in inadequate densities in the layer. Overlapping In 1964, prior to embankment construction, extensive
of the adjacent tracks traversed by the compactor should trials were carried out to study methods of quarrying,
be maintained at about 100 to 200mm placement and compaction (Williams and Stothard
1967). It was found that multi-row blasting produced a
USA well graded fill that could be placed and compacted to
In the construction of rock fill, it is recommended a high density. Ripping and single row blasting were
(Transportation Research Board 1990), that the correct unsatisfactory.
sequence of operations is to dump the rock on to the As a result of the trials a specification was written
layer under construction. The material is then pushed for the rock fill and grading limits were specified for
by a bulldozer over the leading edge of the layer, is construction. No single piece of rock was to have a
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

2"/4 ROCK IqLL

M62 Motorway
/ 0 50 lOOm
Rip-rap and filter 1 in 2.7 / I I I
l in 3.1 I ~X ~ ~ ~ 1

in 1.8
Compacted
rock fill

Compacted I Filters ~ ~ , ~

Relief
C l a y core Illl ill well
Grout curtain

Fig. 7.9. Cross section of Scammonden dam (after BNCOLD 1983).

volume greater than 0.4 m 3 or a dimension greater than mudstone zones in the middle of the embankment.
0.gm. The percentage passing the 76#m sieve was Grading tests were regularly carried out on 25 tonne
specified as 0-5%. Rock fill was to be placed by tipping samples from the blast face. The rock fill was placed on
on the surface of the layer at least 4 m back from the the embankment in 0.9m thick layers and compacted
advancing face and then pushed forward over the face. with 5 passes of an 11.5 tonne vibrating roller. Oversize
No rock fill was to be placed when the moisture content pieces were fragmented by a drop ball on the embank-
of the material finer than 19mm exceeded 12%. ment. Density tests were carried out using the pea gravel
Compaction of each layer was specified as either 8 replacement method in an excavated hole having a
passes of an 8 tonne vibrating roller or 12 passes of a volume of approximately 1 cubic metre.
5 tonne vibrating roller. At times of heavy rain the dump truck tyres churned
The 3 x 106 m 3 of rock fill for the dam were excavated up the fines to form a slurry and it was feared that this
from the adjacent motorway cuttings. The valley is in might lead to weaker layers in the fill. However when
the Millstone Grit series of the Carboniferous system trial pits were taken down through two layers of fill, no
and the strata consist of alternating horizontal layers of evidence of these relatively smooth surfaces was found.
sandstone, mudstone and shale. The embankment was There was an absence of voids with densely packed fines
principally constructed of sandstone, but there are three between the larger rock fragments.
BRE measurements of the deformations of the
embankment during construction have been described
O -,,~ , , ' ' '''I ' ' ' ' ' ' ''I by Penman et al. (1971). Monitoring of the move-
ments has continued over the last 20 years and the
settlement of the crest of the high embankment is
shown in Fig. 7.10 plotted against logarithm of time
(Charles 199 i ).
~" 100
E OD

E
%~ 7.6.2. Roadforddam
E O Roadford dam forms a 37 • 106 m 3 reservoir which sup-
200 plies southwest and north Devon (Wilson & Evans
O ~ o ~0\( ~ 1990). The embankment was built in 1988 and reservoir
impounding commenced in October 1989. The reser-
voir was full early in 1991.
The 41 m high dam is constructed of low grade rock
300 i I I I IIIII I I I i iliil fill. An asphaltic concrete membrane on the upstream
0.1 1 10 face of the embankment forms the watertight element.
A cross section of the embankment is shown in Fig. 7.11.
T i m e since end of construction (years)
The Upper Carboniferous strata at the dam site and
qg. 7.10. Settlement of crest of Scammonden dam (after quarry consist of rapidly alternating mudstones, silt-
Charles 1991). stones and sandstones. Near the surface the weathering
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK I~ILL 2/5

~7 2.25 ~ 2.25
Asphalticconcretemembrane 17
& drainagelayers - - ~ - ~ . ~ ~ "
Cut-offstructure
with inspecti

7~ a ~^a
l =t~7nn~ T o e ~
" I

,-- Groutcurtain

Fig. 7.11. Cross section of Roadford dam (after Evans & Wilson 1992).

of the mudstones is extensive. When an exploratory than 0.5 m. From the gradings of material below 106 mm
quarry was opened the material was classified into three it was concluded that the fill had a slight excess of fines.
types related to the depth below ground level and It was expected that the material would be weather
weathering characteristics. It was considered that an susceptible but there were only a few occasions when it
embankment could be constructed of the less weathered was necessary to stop fill placing after heavy rain. Com-
materials obtained from depths greater than 4 m. paction was to a method specification which required
Samples of the rock showed that there was an average a 450mm finished layer thickness. Some 149 density
of one per cent by weight of iron pyrites (Millmore & cubic test holes with a dimension of 800 mm were exca-
Bell 1994). The mineral was considered to be potentially vated during embankment construction and the gravel
degradable and leaching tests were carried out to replacement method was used to measure the volume.
simulate the effects of weathering. Both laboratory On completion of the density measurement, the test hole
tests and trial embankments confirmed that the chemical was filled with water and a falling head permeability
constituents of the formation could lead to the oxidation test was carried out. The average permeability (hydrau-
of iron pyrites to produce dissolved metals and acidic lic conductivity) was 1.5 x 10 -4m/s. By July 1989 the
drainage water. Measures were taken to minimize the dam was ready for impounding and the reservoir was
oxidation process and mitigate polluting discharges. full early in 1991.
The dam construction contract was let early in 1987 Embankment deformations have been monitored
and further compaction trials were undertaken which during construction, reservoir filling and subsequently.
showed that the addition of water caused the compacted Since reservoir impounding, settlements have continued
surface to become slurried. Embankment construction at a reasonably constant rate apparently unrelated to
commenced in June 1988 and the 1 x 106 m 3 embank- water load. There has not been the steady reduction in
ment was completed by October. The embankment was settlement rate observed, for example, at Scammonden.
formed of equal quantities of sandstone and mudstone This has been attributed to additional settlement in the
quarried on site (Evans & Wilson 1992) and the speci- embankment associated with an increase in moisture
fication required that the embankment fill should be built content (Evans & Wilson 1992).
of rock defined in Table 10 of BS5930:1981 as fresh,
slightly weathered or moderately weathered. Excavation
7.6.3. Queen's Valley dam
was initially carried out by face shovel loaders assisted by
a ripper but to maintain output more powerful rippers The Queen's Valley dam was planned to form a 1.2 x
were introduced and the material was pushed towards 106 m 3 reservoir on the island of Jersey (Bridle 1988).
the loaders. This method of working produced a well The embankment was completed in September 1991.
blended and consistent material (Wilson & Evans 1990). Impounding commenced in December 1991 and the
Samples from the quarry were tested for particle reservoir was full by June 1992. The 34 m high dam was
density (specific gravity), moisture content and grading. built from ignimbrite rock fill with a watertight element
The fill was generally drier than expected with a mean provided by a 600 mm wide central dense bituminous
moisture content of only 4.4%. After ripping and concrete core. Figure 7.12 shows a cross section of
loading the maximum rock size was usually not greater the embankment.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

2)6 ROCK H L L

= Rip-rap
2 " ~,""f~Bit uminous =
1~ , ~ ~ concrete
j core ,ll~li Transition ~ %
C~ e~ I .........omoac,eO

Groutcurtain i=1 I
I I I
I 0 10 20 30m
9 I I I
Fig. 7.12. Cross section of Queen's Valley dam (after Bridle 1988).

F o r e c o n o m i c reasons it was essential to use all the the lower quality rock fill containing more fines being
suitable material that was f o u n d in the q u a r r y and the used d o w n s t r e a m of the core. In the d o w n s t r e a m rock
general excavation as fill in the e m b a n k m e n t . T h e design fill the material finer than 0.063 m m was restricted to a
of the d a m was based on two categories of rock fill, with m a x i m u m of 12%.

Fi~. 7.13. Rock fill placement at Queen's Valley dam; dumping and watering the fill.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK I~ILL 2//

Trials were carried out on both categories of rock fill to 7.6.4. Loch Lomondside A82 highway
establish satisfactory methods of placing and compacting
the fill and to measure mechanical properties. Density, The Camus nan Clais to Hollybank section of the A82
moisture content, grading, permeability and shear wave highway runs along the western shore of Loch Lomond
velocity were measured in field trials and the results were in Scotland. The improvement of this section of the A82
used to design appropriate laboratory tests to assess the has been completed and the road was opened in 1987.
strength and deformation properties of the fill. A 2.1 m At two locations there was not sufficient shore between
wide towed vibrating roller with a static weight of 11.7 the steep hillsides and the loch itself to construct the
tonnes was used in the trials. In general the trials indi- highway along a new alignment. To cut into the hill-
cated very satisfactory results in terms of the densities side at these two locations was not possible due to the
and gradings which were achieved. Figure 7.13 shows the presence of a landslip at one site and the need for an
rock fill being dumped several metres behind the edge excessively high cutting at the other. It was, therefore,
of the advancing fill surface during dam construc- decided to construct the new road along the shoreline in
tion. The fill is being pushed forward by a bulldozer in the Loch and on rock fill. Howison & MacDonald (1988)
Fig. 7.14. The compacted surface of the rock fill is describe the design and construction of this scheme.
shown in Fig. 7.15. Densities of both categories of fill in At the South Camus Bank location, see Fig. 7.16, the
the dam embankment were about 2.1 Mg/m 3. slope of the loch consisted of organic silt and peats
Deformations during construction are largely a overlying sands and gravels at a depth of about 4m.
function of the compressibility of the rock fill and the In order to displace the mantle of soft material below the
height of the embankment. The low compressibility of water level, rock fill was placed by bottom opening
the compacted rock fill ensured that there were no large dump barges in 8 m wide bands working from the shore
movements during construction or reservoir filling; to the embankment toe (Stage 1). Each band was com-
observations indicated a maximum movement during pleted before the next was placed. In Stage 2 rock fill was
construction of only a few centimetres. placed in 1 m layers at the toe of the embankment to

Fi~. 7.14. Rock fill placement at Queen's Valley dam; pushing fill over advancing face.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

2/8 ROCK I~ILL

Fig. 7.15. Rock fill placement at Queen's Valley dam; compacted surface of fill.

Proposed
Existing A82 Stage 1 fill in band widths
A82 CE

x h _ 1 in 4 0 11.0mOD
i ' ~o "7 Rip-rap _ Average water level (WL) 8.0m OD
. . . . . ~ ~i

4.ore o o
/ - - . ~ -,.-~--_%__ _ _-2,,4~~7 35m maximum
Stage 4 fill in~" U ~ ~ " " " " % , ~ - I2 1 Stage 1 f,9II below WL
0.45m layers J '~ ~ '~ /
9 10.0m Natural angle
Stage 3 ,il~ ""~/.~ ~ ~ [ .
end 9
tipped . 11 5m OD i of repose o
nd-tipped O.rgani.csandy " " ~ . " ~ '~i~" Approx. 38
silt and peat ~ .~~,~ ~_~_~.~

Ded
geS?e/dDense
sand " ~ - ~ ~~ ~ ~ ' - ~- ~- - .~
- ~ - - _ _

Stage 2 fill in 1m layers " ----- - ---__. _~.__..

Fi~. 7.16. Sequence of fill placement at A82 South Camus Bank (after Howison & Macdonald 1988).
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK FILL 279

improve the stability. Stage 3 required the end tipping of 7.6.5. Llanddulas to Glan Conwy section of the
the fill at around the average water level for the loch and A55 highway
finally Stage 4 was fill placed in 450 mm layers above the
The A55 highway runs along the north coast of Wales
water level. Rip-rap was placed at the mean water level
providing an important strategic link in the area. This
to protect the embankment slope. section runs between cliffs of boulder clay and the sea. Sea
At the location at Wades Bridge Bank the geology of
defences were incorporated within the highway earth-
the site was slightly different in that the organic silts and
works and both the sea defences and highway earthworks
peat were sandwiched between two granular layers.
contained a high proportion of rock fill. The design of the
Therefore, rather than displace the soft material the rock fill structures is described by Springett & Stevenson
design philosophy was based on containment and
(1984) and the construction by Rowland et al. (1988).
improvement of shear strength with time. The rock fill Figure 7.17 shows a typical cross section through
embankment profile was similar to that at the South
the highway works. The whole embankment structure
Camus Bank location but the rock fill was placed in
is formed on a 600ram bedding layer of crushed rock
layers, each 1 m thick from the base of the embankment (100 mm to 150 mm size) placed on a plastic filter mem-
to the top. The toe bund (Stage 2 of the South Camus
brane whose purpose is to prevent finer material con-
Bank) was constructed first. Fill was placed slowly to taminating the rock fill and weakening the foundation,
allow an improvement in the shear strength of the soft
and to provide a stable working platform. The sea
material to be achieved.
defences consist of a rock fill core of 250 mm to 500 mm
At a third location, Caravan Site Bank, rock fill was
size material. The primary armour on the slope of the
used to infill a small bay between two headlands. The
embankment is formed by two layers of 5 tonne Dolos
ground conditions were good, consisting of loose sands
units laid to a density of 45 per 100 square metres. Just
and silts overlying dense granular material. Conven-
over half were placed in the base layer and the remainder
tional rock fill embankment construction was under-
in the top layer. The 22 000 Dolos units were cast on site
taken in 1 m thick layers. in the open to allow easier access. Production was not
The construction below water level of the 189 •
delayed by one day due to bad weather during the whole
103m 3 of fill was undertaken with two barges of
of the 15 months casting period and frost was not a
120 tonne and 80 tonne capacity. The maximum rock
problem. When the gantry crane for lifting the units
size permitted on the barges was 450 mm and the was thought to be vulnerable to high winds a crawler
rock was placed at locations established using a grid of
mounted crane from the adjacent storage area was used.
surface marker buoys. Monitoring of the loch bed Primary armour at the embankment toe consisted
and embankment slope was carried out using a narrow
of 3 tonne rock boulders. Secondary armour consists
beam echo sounder with confirmation of these surveys
of 1 tonne or 2 tonne rock boulders. The rock armour is
by divers. keyed into the beach at the outer face of the defences to
At South Camus Bank the displacement technique
a depth of about 2.5m. Landward of the sea defences,
worked well with monitoring showing that 4 m of soft
the road embankment was constructed of earth fill
material was moved down slope and the final fill
overlying a base rock fill layer of 100mm to 550mm
quantity was less than that assessed at the tender
size material: Wave screen armour, with a nominal
stage. At Caravan Site Bank more settlement occurred
weight of 300 kg, was placed against the road embank-
than was expected and the quantity of fill was increased
ment. Carboniferous limestone was used throughout
accordingly. However, at Wades Bridge Bank problems
for the rock fill and was obtained from local quarries.
occurred both in the northern and southern sections.
Pulverized fuel ash (PFA) was used as a lightweight fill
In the northern section penetration of the loch slope
on the embankment, and road pavement foundation and
and base was more than expected as the toe bund was
layers constructed on the PFA. The road was opened
placed. Once stability of the toe bund had been achieved
ahead of time in two stages in 1984 and 1985.
subsequent placement of the embankment went much as
expected. In the southern section the toe bund was
placed with little difficulty. However, as the rock fill
7.6.6. New Hong Kong airport
behind the bund was placed a major displacement
occurred and about 15 000 cubic metres of material 7.6.6.1. Introduction
moved down and beyond the embankment toe. Filling The construction of the 1248ha platform for Hong
resumed at the toe and proceeded steadily to the sur- Kong's new airport at Chek Lap K o k was one of the
face, the displacement having removed much of the largest civil engineering projects ever undertaken in
underlying soft material. Settlement pins and inclin- South East Asia, see Fig. 7.18.
ometers have been installed in the embankments and The work consisted of dredging approximately
are showing movement similar to that that predicted in 67 million m 3 of soft marine clay, excavating and placing
the design. 118 million in 3 of soil and rock won from Chek Lap
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

Fig. 7.17. Typical cross section through A55 road embankment and sea defence works (after Springett &
Stevenson 1988).

Fig. 7.18. Chek Lap Kok, August 1996, showin~ completed platform.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK FILL 281

K o k , L a m C h a u , West and East Brothers Islands and 7.6.6.2. Geology

the dredging and placement of 76 million m 3 of landfill Hard Rock. C h e k Lap K o k and the n e i g h b o u r i n g
from m a r i n e b o r r o w areas. S u r r o u n d i n g the platform small island of L a m C h a u are f o r m e d p r e d o m i n a n t l y of
is 1 3 . 2 k m of sea wall, both sloping (formed using rock Mesozoic fine to m e d i u m grained granite with associated
fill) and vertical (formed using concrete blocks). rhyolite, intruded by Tertiary basalt and l a m p r o p h y r e
The levelled islands of C h e k Lap K o k and L a m C h a u dykes. The granite is generally closely to m o d e r a t e l y
form approximately one quarter of the airport platform; j o i n t e d (typical of H o n g K o n g ) , while the rhyolites tend
the r e m a i n d e r is reclamation formed from various to be very closely to closely jointed. A m a j o r n o r t h -
fill materials. south striking fault f o r m e d a p r o m i n e n t valley that
On 30 N o v e m b e r 1992 the Airport P l a t f o r m Con- bisected the f o r m e r island; m i n o r local faults, shear
tractors Joint Venture (APCJV) were a w a r d e d the zones and veining were c o m m o n . A variable thickness
contract on a fixed lump sum basis of approximately of completely d e c o m p o s e d rock f o r m e d a veneer over
H K $ 9 billion. m u c h of the island, generally 0.5 to 3 m thick but

Table 7.4. Fill types used at Chek Lap Kok (ref. to Fig. 7.19)

Fill type Description Source Use Areas

A Hard and durable rock, Slightly to moderately In areas requiring the Southern runway and
formed from bulk weathered rock, generally characteristics of durable taxiway.
blasting. The allowable grade II or III. rock fill i.e. low sensitivity
grading ranged from 2 m Obtained from bulk blasting to earthquake induced
boulders to silt. In practice of the former islands of settlement and
the as-excavated material Chek Lap Kok and Lam liquefaction, low creep
was generally a well Chau. and free draining.
graded rock fill ranging Seawalls and Southern
from 600 mm boulders to Runway.
a medium gravel.

The allowable grading Highly to completely In areas of piling, tunnelling East and west vehicle
ranged from 300mm decomposed rock and and deep excavations with tunnels and APM
boulders to clay with up to residual soils. associated dewatering. tunnel extension.
50% fines. In practice the The metasedimentary Capping layer for utilities General capping
as-excavated material was material obtained from the and pavements. layer.
generally a silty clayey, Brothers Islands was not
gravely sand with boulders considered hard and
and cobbles (20 to durable and was therefore
100 mm). Occasionally classified as Type B.
large boulders were Generally machine
present. excavated.

C Marine sand, with a small Marine borrow areas within In areas of general fill, piling, General areas.
gravel content. The Hong Kong Territorial tunnelling and deep Midfield.
mineralogy is waters. excavations with
predominantly quartzitic, associated dewatering.
but includes some Capping layer for utilities
calcareous content. and pavements.
A/B Mixture of Type A and B. Completely to moderately General fill areas. Stabilizing General areas.
decomposed rock. Both blanket of material Northern runway
blasted and machine bottom dumped on the landside.
excavated. dredged surface prior to
endtipping.

Sorted Large hard durable boulder Slightly to moderately Sloping seawalls. Sloping seawalls.
processed for armour rock or weathered rock, generally
rock durable graded rock fill for grade II or III. Obtained
filter layers and from bulk blasting. Small
underlayers, proportion imported.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

282 ROCK FILL

becoming over 20m deep in the vicinity of the major 7.6.6.4. Fill allocation plan
fault (Langford 1994). A simplified fill allocation plan is shown in Fig. 7.19.
The former Brothers Islands, a sedimentary inlier,- The rationale behind the fill allocation plan was influ-
were composed of metasediments, generally meta- enced primarily by the requirements of a fast track
sandstone and graphitic siltstone. High grade graphite project together with relating fill types to specific future
was mined from West Brother island until the late 1960s construction and fill performance requirements. This
for use in nuclear power stations. careful planning significantly reduced the time consum-
ing and costly requirement of extensive sorting and
Superficial Deposits. The offshore Quaternary geology processing of the blasted rock, the exception being for
comprises a complex sequence of interbedded clays and armour rock and underlayers for construction of the
sloping seawalls.
sands of an alluvial/marine origin, overlain by Holocene
marine clays (James et al. 1994). The onshore super-
ficials, usually found either at the base of valleys or in 7.6.6.5. Blasting, excavation and placement of fill
small bays, generally comprised either typical flood plain The bulk blasting and excavation to formation level of
sediments, sands and clays, or beach sands and gravels. Chek Lap Kok and Lam Chau, and the placement
of 118 million m 3 of rock fill, had more in common with
a quarrying or mining operation than normal earth-
7.6.6.3. Fill types works. At peak production, the works were reputedly
The reclamation was formed from three basic fill types, the worlds fourth largest operational open cast mine,
designated as Types A, B and C, the description and see Fig. 7.20.
use of which are listed in Table 7.4. A number of other Bulk blasting techniques were used to fragment the
fill types were also produced for specific uses, such as rock to an extent that would both allow excavation
in the construction of the seawalls that surround the and produce an as-blasted fill of the required grading.
reclamation. To cater for the large quantities of explosives required

Fig. 7.19. Fill allocation plan.

 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK FILL 283

Fig. 7.20. Excavation of rock fill from blasted benches, December 1994.

for blasting, a bulk A N F O , a mixture of fuel oil and Table 7.5. Typical blasting parameters
ammonium nitrate, and slurry manufacturing plant was
established on site. The explosive mix used comprised Blast hole diameter (ram) 165 to 216
30% A N F O and 70% emulsion for wet holes and 100% Burden (m) 5.0 to 6.5
Spacing (m) 6.0 to 7.5
A N F O for dry holes. Typical blasting parameters are set 1.5 to 2.0
Sub-drill depth (m)
out in Table 7.5 Total depth (m) 13.5 to 15
Blasting was generally carried out twice a day, using Weight explosive per hole (kg) 150 to 240
between 40 and 50 tonnes of explosives for each blast Bench height (m) 12 to 13
and producin~ about 80000m 3 of fragmented rock.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

284 ROCK FILL

In all about 40000 tonnes of explosives were used. effective in reducing creep within the fill. One area of the
Twelve percussive drilling rigs drilled over 2700km of site was subjected to dynamic compaction (DC) in order
blast hole, producing about 85 million m 3 of fragmented to improve the settlement characteristics. The D C was
rock. Sixteen excavators, with a bucket capacity of up to carried out on an area of rock fill directly adjacent to a
2 2 m 3 were used to load a fleet of 67 dump trucks, vibrocompacted sand fill in order to reduce the differ-
(Caterpillar 785s and 777s), with a capacities of up to ential settlement between these fill types.
136 tonnes. The saturated density of the Type A and A/B below
The majority of the fill obtained from Chek Lap Kok the water table has been estimated to be approximately
was directly end tipped. In areas where there was concern 2 2 k N / m 3 which equates to a bulking factor of 1.35.
over tip face stability, either a stabilizing blanket of rock Although no ground treatment of the upper layer of fill
fill or sandfill was first barge-dumped on the dredge (above the water table) has been carried out, the tracking
surface, or a policy of tipping short of the tip face and action of the considerable number and size of plant
using bulldozers to push the fill over the tip face was required to construct the platform has resulted in this
adopted. Adherence to this policy meant that only a zone being, in general, in a dense state. The equivalent
small number of minor failures occurred at the tip face. load per tyre contact area for the haul trucks greatly
exceeds conventional heavy compaction plant. Large-
7.6.6.6. Seawalls scale water replacement tests have been performed in
The construction of the seawalls (Neville-Jones & Type A/B fill with results typically in the range of 23
Leutchford 1994) required the production and selection to 24 kN/m 3.
of 450 kg to 5 tonnes armour rock, various underlayer The permeability of the Type A and A/B fills below
grades and seawall core material. The seawall core rock the water table is extremely high. In general it was
was selected at the blast face and consequently did not possible to measure the actual permeability using
not require additional processing. The finer underlayer pumping tests, due to the inability to create a significant
materials were produced by crusher and mechanical drawdown. The assumed value is thought to be in the
screening process. The remainder of the rock was pro- range of 10 -3 to 1 m/s. The permeability of the Type B
cessed using backhoes for armour rock smaller than (CDG) was typically in the range of 10 -7 to 10 -5 m/s.
900 mm, scalping buckets for those of intermediate size It was neither necessary nor feasible to carry out
and individual selection for 3 to 5 tonne armour rocks. strength tests on either the Type A or A/B fills. However
The scalping bucket consisted of a bucket frame with based on the side slope angle of fill stockpiles around the
spaced bars set to the required rock size. site an internal angle of friction of 45 ~ would be
In the early stages of the project, no particular effort applicable in the loose state, and considerably greater
was made to produce rock of the sizes required for use as strength could be mobilized when compacted or partially
armour rock. It later became obvious that a short-fall saturated.
was likely and the day-to-day blasting operation was For design of reclamations, the creep and stiffness
modified to maximize armour rock production. characteristics are extremely important. The instrumen-
tation for the airport platform has been described
7.6.6.7. Quality Control of rock fill by others (Newman et al. 1995; Ayson & Lang 1996).
The primary quality control of the rock fill was carried Approximately 40 extensometers have been installed
out by a visual assessment of the weathering state, through the fill and underlying alluvium and anchored
durability and grading of the rock. This was carried out into the underlying bedrock in order to monitor sub-
principally at the blast face, and dictated the final surface settlement of the platform. A number of these
destination of the fill. In addition, samples of the material were extended up through the surcharge fills constructed
were reviewed at the tip face prior to end tipping. at various locations on the platform. These have allowed
Bulk grading tests of the fill were carried out at a complete time settlement curves to be determined.
frequency of 1 test for every 10000m 3 of fill. The A typical result for compression of the fill due to sur-
grading tests were done using a 'bunk bed' of grading charging is shown in Fig. 7.21. It can be seen that
screens with apertures of 900, 300 and 75ram. Addi- settlement is a time dependent process. It is of note
tional testing was also carried out where the fill was that the initial compression of the fill is not immediate
required for use as an aggregate. but takes place over 2 to 3 weeks. Based on the known
high permeability of the fill the settlement is not con-
7.6.6.8. Properties of the rock fill sidered to be a consolidation process but more related
In general, no ground treatment was undertaken at to creep.
Chek Lap Kok during the placement stages and Creep is usually assumed to commence 0.1 years after
consequently the end tipped or bottom dumped rock completion of filling. This period is generally used as a
fill is in a relatively loose state. Limited areas of the site mathematical convenience. However, it is useful to
were surcharged, primarily to reduce settlement in the define the settlement which occurs in the first 0.1 years as
underlying alluvium. However, surcharging also proved bein~ the constrained modulus and settlement thereafter
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK PILL 285

OI

0.05

0.1

0.15

g
"~ 0.2

(~ 0,25 i

Surcharge removed

0,3

0.35

0.4

6 8 10 12 14

Weeks since Construction of S u r c h a r g e

Thickness of fill = 2 2 . 5 m Height of surcharge = 8m

Fig. 7.21. Compression of fill due to surcharge.

as creep settlement. It should be noted that neither 2 0 m thickness of rock fill below the water table was
of these values are elastic parameters, as there is little to typically in the range 200 to 500 m m for a surcharge load
no recovery on unloading. Compression parameters in the range 100 to 160kN/m 2.
derived in this way show that both Type A and A/B fill The Type A fill below the water table has a
have similar characteristics. The compression of the constrained modulus in the range 20 to 50 M N / m 2 and

0,6

0.5 -- 9 I~'E136!
I~'E137~
9 I~'E145|
I-x-sP12~
0.4 . . . . . . . . . . . . . . . . . 9 . . . . . . . . ~ . . . . . ~ . . . . . . . . . J. . . . . . . . . . . . . . . . . ~ . . . . . . . r . . . . . . . . . . . . .

I~ 0.3 . . . . . . . . . . . . . ! ....... ':i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

i i i
L
9
9 i
, i

i 9 i i
O
0.2

>O i i i 9 i ~ i i i

0.1 .... "- ~ ~- ~ . . . . . . . . 9 . . . . . . ~ . . . . ,4 . . . . . . . 9 . . . . . . . . ~ . . . . . . . . . . ~ . . . . . . . . ; ......... ,, . . . . .

, - . . ~ i = l9 = , p ,

, , ,- . . . . , , '+ : :
I I I " 9 , l l
0 . . . . . . . . . t- . . . . . . . . t . . . . . . . . ~ . . . . . . t . . . . . . . . . I. . . . . . . . . . . . . . . . ~ . . . . . . . f . . . .
: ', ', ~ = 0.2%

, ~ i , t , 9 i ~ "
-0.1

, J , , ' [1 = 1% 9 ' ,
-0.2

-0.6 -0.5 -0,4 -0.3 -0,2 -0.1 0 0.1 0.2 0.3 0.4

L o g (yr) after r e c l a m a t i o n

Fi~. 7.22. Lo~ time strain plot of fill type A.

 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

286 ROCK FILL

9 ! , ~ L
. . . . . . . ~;i 9. . . . . . . . ', . . . . . . . . . l'~. . . . . . . . . r, . . . . . . . . . . .~. . . . . . . . . . . -~. . . . . . . . . . .
~ 9 . . . . ~. . . . . . . . ~ . . . . . . . . --~-- IEl10

, 9 . . . . . [ ', --m--lE124
i 9 J i J i i i i
i 9 i i l i t i i i-*-J~12~ I
. . . . . . . . . ~ . . . . . . ~_~ . . . . . . . . ~ . . . . . . . ~. . . . . . . . . + .......... _, . . . . . . . + . . . . . . . , . . . . . . . . , . . . . . . . . . t + IE127

l l ~ ~ ll~ l l ~ i [ l
i 20 [
I ~ I I I d I i I --O---IE147
'L i 9 , l i i , , i
A
0.3
L 1 ~ , . . . . . -'m-- IE156

--x- IE 1 5 7
i , , 9 , , i i i dl
i , i 9 i , ~ i i i
---I.~ IE159

0.2
- - x. .-.-. . . . S
J~ . PIO6J

~ 0.1

', : a =0.2~

-0.1

-02

-0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4

Log (yr) after reclamation

Fig. 7.23. Log time-strain plot of fill.

the Type A/B fill in the range 10 to 40 M N / m 2. Care the peak of the works, 2 hectares of reclaimed land were
must be taken when considering these values, as fills had being formed each day.
been in place for variable periods prior to surcharge The rapid pace of construction was at least in part
loading, and were subject to different surcharge heights. achieved by the use of a carefully programmed fill
From the extensometers installed in the reclamation, allocation plan taking best advantage of the various
up to three years of monitoring data have been collated, fill types available. This minimized the requirements
and are available to enable creep parameters to be for fill processing and ensured that the fill placed in a
determined. The logarithmic creep parameter (c0 for particular area met the specific engineering requirements
Type A has been determined to be in the range 0.2 to of the future planned construction.
0.5% per log cycle of time and Type A/B in the range 0.3 This approach also minimized the requirement for
to 1.0% of log cycle of time. Typical results derived from ground improvement of the reclamation fill thereby
extensometers in the fill are shown in Figs 7.22 and 7.23. reducing both cost and time. To date the performance of
For comparison purposes the creep measured in each the reclamation has met with the design requirements
extensometer was determined on a common base of one for the follow-on facilities and is expected to do so
year after completion of the reclamation at the location throughout the operational life of the airport.
of the extensometer. Total creep was calculated from
0.1 years after reclamation. Acknowledgements. Section 7.6.3 on Queen's Valley dam is
included by kind permission of The Jersey New Waterworks
In addition to the above stiffness values, plate load
Company Limited.
tests have been carried out on the compacted Type A/B
fill above the water table. Subgrade reaction in the range
100 to 300kPa/mm has been determined for 300mm References
diameter plates.
AMERICAN SOCIETY OF TESTING AND MATERIALS. Annual Book
of A S T M Standards. ASTM, Philadelphia.
7.6.6.9. Conclusion ANON 1990. Tropical residual soils: working party report.
From start to the substantial completion of the reclama- Quarterly Journal o f Engineering Geology, 23, 1-101.
tion took approximately 31 months and involved the - - 1 9 9 4 . Manual of Contract Documents for Highway Works.
movement of about 350 million m 3 of material, equi- Volume 1: Specification for highway works (MCHW 1).
valent to two tonnes per second over the 31 months. At Her Majesty's Stationery Office, London.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

ROCK H L L 287

AYSON, I. J. & LANG, C. 1996. Survey Monitoring and ESCARIO, V . , LOPEZ-CORRAL, A. & SERRANO, A. 1980.
Instrumentation Techniques Adopted During the Con- Correlation between plate bearing tests and deformations
struction of Hong Kong's New Airport at Chek Lap Kok. originated by a 10t axle load for compaction control of
Proceedings of the 8th FIG International Symposium on earthworks. International Conference on Compaction. Edi-
Deformation Measurements, 25-28 June 1996, Hong Kong. tions Anciens ENPC, Paris, Vol II.
Dept. Land Surveying and Geo-Informatics, Hong Kong EVANS, J. D. & WILSON, A. C. 1992. The instrumentation,
Polytechnic University. monitoring and performance of Roadford dam during
BAILEY, M. J. 1976. Degradation and other parameters related construction and first filling. Water Resource,; and Reser-
to the use of shale in compacted embankments. Joint voir Engineering. Proceedings of 7th Conference of British
Highway Research Project-76-23, Purdue University, Indi- Dam Society, Stirling. Thomas Telford, London, 157-165.
ana. FOOKES, P. G. 1976. Stabilization and control of local rock falls
BNCOLD 1983. Dams in the UK 1963-1983. British National and degrading rock slopes. Quarterly Journal of Engineer-
Committee on Large Dams, London. ing Geology, 9, 37-55.
BOND, F. C. 1959. Confirmation of the third theory. American FORSSBLAD, L. 1981. Vibratory Soil and Rock Fill Compaction.
Institute of Mining & Metallurgical Engineers, Annual Dynapac Maskin AB, Solna, Sweden.
Meeting in San Francisco. FRANKLIN, J. A. & CHANDRA, R. 1972. The slake-durability
BRIDLE, R. C. 1988. Selection and design of the waterproofing test. International Journal of Rock Mechanics and Mining
element for Queen's Valley dam, Jersey, Channel Islands. Science, 9, 325.
Transactions of 16th International Congress on Large FUMAGALLI, E. 1969. Tests on cohesionless materials for rock
Dams, Vienna, 2, 643-653. "fill dams. Journal of Soil Mechanics and Foundations
BRITISH STANDARDS INSTITUTION 1981. Code of Practice for Division. American Society of Civil Engineers, 95, January,
Site Investigations. BS 5930: 1981. BSI, London. SM1,313 330.
- - 1 9 9 0 . BS 1377: 1990. Methods of Test for Soils for Civil GALLOWAY,J. D. 1939. The design of rock fill dams. Trans-
Engineering Purposes (in 9 parts). BSI, London. actions of American Society of Civil Engineers, 104, 1-24.
- - 1 9 9 4 . BS 8002: 1994. Code of Practice for Earth Retaining GUNNING, A. 1992. CCTV borehole surveying and its applica-
Structures. BSI, London. tion to rock engineering. International Society of Rock
BROCH, E. & FRANKLIN, J. A. 1972. The point-load strength Mechanics: Eurock '92, Chester. British Geotechnical
test. International Journal of Rock Mechanics and Mining Society, Telford, London, 174-178.
Science, 9(6), 669. HALE, B. C., LOVELL, C. W. & WOOD, L. E. 1981. Development
BROMHEAD, E. N. 1986. The Stability of Slopes. Surrey Uni- of a laboratory compaction degradation test for shales.
versity Press. Transportation Research Record 790, Transportation
BROWN, E. T. 1981. Rock Characterization Testing and Research Board, Washington DC, 45-52.
Monitoring. ISRM Suggested Methods. Pergamon, Oxford. HOEK, E. & BRAY, J. W. 1981. Rock Slope Engineering. Institute
CATERPILLAR TRACTOR COMPANY 1982. Caterpillar Perfor- of Mining & Metallurgy, London.
mance Handbook. Caterpillar Tractor Company, Preoria, - - & BROWN, D. F. 1988. Rock support. Mining Magazine.
Illinois. October 1988.
CENTRE DE RECHERCHES ROUTIERERS 1978. Code of Practice HOWISON, J. A. & MACDONALD, A. 1988. Trunk Road A82: the
for the Geotechnical Investigations of Road Alignments. Loch Lomondside road: conception to implementation.
Volume 2 Survey of problems (in French). Recommanda- Proceedings o2["the Institution of Civil Engineers, 84(1),
tions CRR-R42/78, C. R. R Brussels. 497-517.
CHAIGNE, P., LEFLAIVE, E. & SCHAEFFNER, M. 1980. A new JAKANI, A. 1987. E a r t h w o r k s - d r a i n a g e - subgrade. Report
concept for compaction specifications for road embank- from Morocco. PIARC XVIII Worm Roads" Congress,
ments (in French). International Conference on Compac- Permanent International Association of Road Congresses,
tion. Editions Anciens ENPC, Paris, Vol II, 4 9 3 4 . Paris, 61-75.
CHARLES, J. A. 1991. Laboratory compression tests and the JAMES, J. W. C., EVANS, C. D. R., CROSBY,A. & HUMPHREYS, B.
deformation of rock fill structures. Chapter 5 of Advances 1994. An interpretation of the marine geology for the
in Rock fill Structures (ed. Maranha das Neves E), replacement airport at Chek Lap Kok, Hong Kong. British
pp. 73-95. Geological Survey Technical Report WB/93/34 R.
- - & WATTS, K. S. 1980. The influence of confining pressure JANSEN, R . B. 1980. Dams and Public Safety. US Department
on the shear strength of compacted rock fill. G~otechnique, of the Interior, Denver.
30(4), 353-367. KJAERNSLI, B. & SANDE, A. 1963. Compressibility of some
& SOARES, M. 1984. Stability of compacted rock fill coarse grained materials. Proceedings of European Con-
slopes. GOotechnique, 34(1), 61-70. ference on Soil Mechanics and Foundation Engineering,
- - & PENMAN, A. D. M. 1988. The behaviour of embank- Wiesbaden, 1, Deutsche Gesellshaft fur Erd und Grund-
ment dams with bituminous watertight elements. Transac- bau, 245-251.
tions of 16th International Congress on Large Dams, San KUZNETSOV, V. M. 1973. The mean diameter of the fragments
Francisco, 2, 693-705. formed by blasting rock. Soviet Mining Science, 9, 144-148.
CUNNINGHAM,C. V. B. 1982. Rock fragmentation related to LANGFORD, R. L. 1994. HKGS, Geology of Chek Lap Kok,
blast design. Quarry Management and Products, December, Sheet Report No. 2 Geotechnical Engineering Office,
820-824. Hong Kong.
DE MELLO, V. F. B. 1977. Reflections on design decisions of MACGREGOR, F., FELL, R., MOSTYN, G. R., HOCKING, G. &
practical significance to embankment dams. G~otechnique, MCNALLY, G. 1994. The estimation of rock rippability.
27, 281-354. Ouarterlv Journal of En~ineerin.~ Geology, 27, 123-144.
 Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

288 ROCK FILL

MARANHA DAS NEVES, E. (ed.) 1991. Advances in Rock fill Coastal Structures and Breakwaters. Institution of Civil
Structures. Kluwer, Dordrecht. Engineers, Thomas Telford, London.
MARSAL, R. J. 1973. Mechanical properties of rock fill. In: PENMAN, A. D. M. & CHARLES, J. A. 1976. The quality and
H~RSCHFELD, R. C. & POULOS, S. J. (eds) Embankment suitability of rock fill used in dam construction. Tran-
Dam Engineering-Casagrande Volume. Wiley, New sactions of 12th International Congress on Large Dams,
York, 109-200. Mexico, 1, 533-556.
MATHESON, G. D. 1983. Presplit Blasting in Highway Rock --, BURLAND, J. B. • CHARLES, J. A. 1971. Observed and
Excavation. Departments of Environment and Transport, predicted deformations in a large embankment dam during
TRL Report LR 1094. Transport Research Laboratory, construction. Proceedings oj" Institution of Civil Engineers,
Crowthorne. 49, May, 1-21.
- - 1 9 8 4 . A Device for Measuring Drill Rod and Drill Hole PERRY, J. 1994. A technique for defining non-linear shear
Orientations. Departments of Environment and Transport, strength envelopes, and their incorporation in a slope
TRL Report SR 817. Transport Research Laboratory, stability method of analysis. Quarterly Journal of Engineer-
Crowthorne. ing Geology, 27, 231 242.
1985. The stability of slopes exposing rock. Proceedings oj" & PARSONS, A. W. 1986. Assessing the Quality of
the International Symposium on Failures in Earthworks. Rock Fill." a Review of Current Practice jor Highways.
Institution of Civil Engineers, London, 295-304. Department of Transport, TRL Report RR60. Transport
- - 1 9 8 6 . The design and excavation of stable slopes in hard Research Laboratory, Crowthorne.
rock, with particular reference to presplit blasting. Inter- PETa'IEER, G. S. & FOOKES, P. G. 1994. A revision of the
national Symposium on Rock Engineering and Excavation in graphical method for assessing the excavatability of rock.
an Urban Environment. Institution of Mining and Metal- Quarterly Journal of Engineering Geology, 27, 145 164.
lurgy, Hong Kong, 271 283. ROSENHEINRICH, G. 1978. The placing and compaction of rock
- - 1 9 9 2 . Highway rock excavation. Explosives '92, Leeds. fill (in German). St u Tiefbau, 32, 4, 10-16.
Institute of Explosives Engineers, London, 139-154. Ross, K. D. & BUTLIN, R. N. 1989. Durability Testjbr Building
- - 1 9 9 5 . Aspects of highway rock engineering in the UK. In: Stone. BRE BRI41. Building Research Establishment,
EDDLESTON, M., WALTHALL, S., CRIPPS, J. C. & CULSHAW, Watford.
M. G. (eds) Engineering Geology of Construction. Geolo- ROWLAND, S. B., JOHNSON, K. A . L. & WENDES, J. 1988. A55
gical Society, London, Engineering Geology Special Llanddulas to Glan Conwy construction. Proceedings of
Publications, 10, 169 187. the Institution of Civil Engineers, 84(1), 965 988.
MCMILLAN, P. & MATHESON, G. D. 1998. Rock slope hazard SOWERS, G. F., WILLIAMS, R. C. & WALLACE, T. S. 1965. Com-
assessment: a new approach. In: MAUND, J. G. & pressibility of broken rock and the settlement of rock fills.
EDDLESTON, M. (eds) Geohazards in Engineering Geology. Proceedings of 6th International Conference on Soil Mech-
Geological Society, London, Engineering Geology Special anics and Foundation Engineering, Montreal, 2, 561-565.
Publications, 15, 177 184. SPmNGETT, M. & STEVENSON, T. D 1988. A55 Llanddulas to
- - 1 9 9 7 . A two-stage system for highway rock slope risk Glan C o n w y - design. Proceedings of the Institution of
assessment, International Journal of Rock Mechanics and Civil Engineers, 84(1), 939 964.
Mining Sciences, 34, 3 4. STROM, W. E. JR. 1980. Design and construction oJ" shale
MILLMORE, J. P. & BELL, J. E. 1994. Roadford reservoir- embankments: Summary. Report FHWA/TS-80/219. Army
environmental experiences. Transactions o[" 18th Interna- Engineer Waterways Experiment Station, Vicksburg,
tional Congress on Large Dams, Durban, 2, 659 672. Mississippi.
MINISTERE DE L'EQU1PEMENT DU LOGEMENT ET DES TRANS- TRANSPORTATION RESEARCH BOARD 1990. Guide to Earthwork
PORTS 1992. Construction of embankments and formation Construction. State of the Art Report 8 TRB, Washington.
layers (in French). Service d'Etudes Techniques des Routes WANG, H., LATHAM, J.-P. & MATHESON, G. D. 1992. The
et Autoroutes, Laboratoire Central des Ponts et Chaussees, design of fragmentation blasting in surface rock excava-
Paris. tion. International Society of Rock Mechanics: Eurock '92.
NATIONAL |NSTITUTE FOR TRANSPORT AND ROAD RESEARCH Rock Characterization, Chester. British Geotechnical
1982. Construction of Road Embankments. Technical Society, Thomas Telford, London, 233-238.
Recommendations for Highways TRH 9 NITRR, Pretoria. WILLIAMS, H. & STOTHARD, J. N. 1967. Rock excavation and
NEWMAN, R. L., PINCHES, G. M., LANG, C. & PEART, M. 1995. specification trials for the Lancashire-Yorkshire Motor-
Preliminary Aspects of the Instrumentation of the Recla- way, Yorkshire (West Riding) Section. Paper No 6972,
mation for the New Hong Kong Airport at Chek Lap Kok, Proceedings q[" Institution of Civil Engineers, 36, 607-631.
HKIE Annual Geotechnical Seminar 'Instrumentation '. WILSON, A . C. & EVANS, J. D. 1990. The use of low grade rock
NEWLLE-JoNEs, P. & LEUTCHFORD, O. 1994. The construction fill at Roadford dam. The embankment dam. Proceedings
of 13 km of seawall for Hong Kong's new airport at Chek of 6th Cof!['erence q/" the British Dam Society, Nottingham.
Lap Kok. Proceedings of Conference on Advances in Thomas Telford, London, 21-27.
Downloaded from http://egsp.lyellcollection.org/ at Monash University on June 3, 2015

Fi~. 8.2. Some examples of cross sections of rock armoured marine structures.