International Journal of Mining Science and Technology 28 (2018) 573–582

Contents lists available at ScienceDirect

International Journal of Mining Science and Technology

journal homepage: www.elsevier.com/locate/ijmst

Reinforcement selection for deep and high-stress tunnels at preliminary

design stages using ground demand and support capacity approach
Reza Masoudi ⇑, Mostafa Sharifzadeh
Department of Mining & Metallurgical Engineering, Curtin University, Western Australian School of Mines (WASM), Perth 6102, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Underground mining is going to be deeper gradually because near surface resources are going to be
Received 25 October 2017 depleted. Therefore, risk of seismic events in underground mines is escalating. Additionally, existence
Received in revised form 19 January 2018 of the large ratio of horizontal to vertical stress, could be a potential reason for high-stress condition
Accepted 29 January 2018
and occurrence of dynamic activities. Depending on various parameters such as the level of induced
Available online 13 February 2018
stress, rock properties, etc., ground demand changes and it is difficult to estimate. On the other hand,
under seismic condition, energy dissipation and deformation capacity of supports is the most important
Keywords:
factors, however, rock support performance factors in dynamic conditions are still under investigation.
High-stress tunnels
Support system
Expanding the knowledge of reinforcement behaviour and capacity, specifically that of the rockbolt as
Ground demand a primary element in seismic conditions, would help to develop a suitable, safe and economic support
Reinforcement capacity design. This paper contains various methods to estimate ground demand including the intact rock prop-
Rockbolt erties approach, failure thickness and ejection velocity estimation, and rockburst damage potential
method. It also covers measurement methods of rockbolts energy dissipation capacities such as drop test,
blasting simulating, back calculation and momentum transfer measurement methods. A large-scale
dynamic test rig is also explained. Based on the findings, a table and a graph to show the applicable range
of each type of rockbolts were presented. Suitable rockbolt types for various ground energy demand and
deformation capacity range were categorised in the table and the graph. The presented support selection
method facilitates the selection of a suitable reinforcement system at the preliminary stages of design
and guides the designer to adjust the support reinforcement system based on observed ground and sup-
port reaction.
Ó 2018 Published by Elsevier B.V. on behalf of China University of Mining & Technology. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction rockmass energy demand and rock support energy dissipation

capacity. Numerous unknowns, uncertainties in geomechanical
Deeper underground mining exploitation is increasing world- parameters and randomness occurrence of seismic events increase
wide because near surface mineral resources become gradually the complexity of the rock demand determination and conse-
depleted. In-situ stress increasing in rock is the main difference quently extend the complication of an effective support design.
between rock stresses at depth compared to the rock near the sur- Though a significant amount of work has been done to estimate
face, and dynamic activities are direct consequences of such a con- energy dissipation capacity of support elements, this subject is not
dition. Seismic events such as the rockburst might occur below much known. Additionally, the role played by other mechanisms of
600–800 m depth and more likely passing 1000 m depth. Such loading, like dynamic shear loading, in the support system is also
phenomena are not limited only to deep mines as many shallow not clearly understood.
mines in Australia experience such events due to the presence of To achieve stability and safety at deep and rockburst prone con-
high horizontal to vertical stress ratios. ditions, appropriate support and reinforcement design is neces-
Hard rock mining is experienced at a depth of about 2 km in sary. The support system should not only be able to tolerate the
Australia, more than 3 km in Canada, and a depth of about 4 km static rock load and potential dynamic load due to induced stress,
in South Africa highlight the importance of ground stability at such but it should also not lose strength over a wide range of deforma-
depths. Finding a practical support design requires determining the tion. It could be concluded that the energy dissipation capacity of
support elements individually, as well as the ground support as
⇑ Corresponding author. an integrated system, needs to be found. Ground energy demand
E-mail address: [email protected] (R. Masoudi). cannot accurately be determined or calculated, but some

https://doi.org/10.1016/j.ijmst.2018.01.004
2095-2686/Ó 2018 Published by Elsevier B.V. on behalf of China University of Mining & Technology.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
574 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582

estimation might be achieved to help engineering judgment. Some Fig. 1b shows the ejection of a volume of rock by the mecha-
of the methods, based on intact rock properties, have attempted to nism of sudden buckling or spalling of rock in the wall or even in
find a relationship between rockmass properties and their poten- the face due to induced or concentrated stress on the boundaries
tial to burst, and the real condition of rockmass under stress of the opening where foliation of the rockmass is nearby vertical.
[1,2]. Some other methods are based on the estimation of probable This mechanism applies to strong to extremely strong rocks.
failure volume, ejection velocity and the travelling distance of Fig. 1c shows the ejection of a volume of rock in the wall due to
ejected materials [3]. Another recent method relies on the defini- a seismic event near the boundaries of a stope or a tunnel which is
tion of the effective parameters on the potential of rockburst and due to slip or energy transfer on an adjacent discontinuity. Initial
its likely damage [4]. On the other hand, some researchers believe or secondary discontinuities can bound the volume of ejection so
that there is not a precise method to determine rockmass demand it can be estimated if the location of such an event is known.
with any degree of confidence [5]. Fig. 1d depicts the mechanism of instability in the back due to a
Along with ground demand during dynamic events, much effort combination of the effect of loosening of discontinuous blocks,
has been expended in determining the rock support energy dissi- gravity, and/or a seismic event. Loosening of the blocks in the back
pation capacity. Rockbolt as the primary element to transfer the could be a result of the lack of enough confining stress or previous
energy of the displaced volume of surface rock to the ground in blasting. The seismic event can accelerate the phenomenon under
depth has been the focus. Several approaches including the drop the effect of available gravity.
test, blast simulation, back calculation and momentum transfer Therefore, considering the wide range of rockmass and dynamic
method have been developed in order to examine rockbolt perfor- load conditions, various types of failures such as spalling, rock ejec-
mance [4,6–10]. Another so-called large-scale dynamic test rig has tion and block fall can be expected.
been constructed in 2012 by Geobrugg in Switzerland in order to
investigate the whole support system as an integrated system 2.2. Ground seismic energy demand
[11,12]. Despite several research studies on different ground types,
support systems in a wide range of loading, rockbolt types, etc., When a dynamic load propagates in the excavation, rock defor-
there are limited comprehensive studies on this subject. mation occurs and cause an energy release. Estimating the magni-
In this research, at first, a short explanation of different mecha- tude of released energy is important to design a suitable
nisms of rockburst and rock ejection and various methods of reinforcement system. Although several methods have been devel-
ground demand estimation and rockbolt energy dissipation capac- oped to estimate the ground energy demand, they can be cate-
ity, are illustrated. Then, suitable rockbolt type selection is recom- gorised into three groups namely, Intact rock property approach
mended for different ground demand levels. The method is simply (IRPA), Estimation of failure volume and ejection velocity, Rock-
presented by table and graph which is easy to use in practice. The burst damage potential. A brief illustration of each method is given
presented methods can assist the selection of appropriate rockbolt in the following subsections.
type at the preliminary stages of mine design. Additional to the
rockbolt selection, some further considerations for the selection 2.2.1. Intact rock property approach (IRPA)
of other support elements is given as well. When a volume of energy that should be tolerated within the
rockmass exceeds its capacity (Strength), sudden failure happens,
and energy is quickly released. Although all factors such as discon-
2. Deep underground and high-stress mining tinuities and their infilling material properties, and the presence of
underground water and its effects are important, intact rock prop-
Seismically active underground mines are those which are erties have significant roles in this phenomenon. As a matter of
prone to dynamic rockmass failure. As mining progresses, the nat- fact, the intact rock energy absorption capacity could determine
ural stress equilibrium of the rockmass is disturbed. Stresses con- the upper limit of energy absorption capacity or in other words,
centrate around the edges of an excavation or in pillars of rock the potential releasable energy of the rockmass. Some criteria have
between excavations left unmined for support, due to low grade been defined to estimate the potential of rockburst based on intact
or other reasons. Stress may also be increased or relaxed on pre- rock properties including Index of strain energy, Potential energy
existing planes of weakness such as faults, shears or lithological of elastic strain [1,13], rock brittleness [14], and ratio of tangential
contacts. These stress changes cause the accumulation of potential stress to compressive strength [15].
energy in the unmined rock. This energy may be gradually dissi- An excess of energy during the post-peak deformation stage
pated, or it may be released suddenly during the process of inelas- conclude in violent rock fracturing [16]. Energy release rate (ERR)
tic deformation and radiates detectable seismic waves. has been developed as a basis for mining exploitation pattern
design. Rock subjected to the compression process experiences
elastic and plastic deformation. Elastic deformation (strain) of the
2.1. Ground behaviour in seismic conditions rock can be recovered if unloading occurs before peak strength.
At brittle failure, the elastic strain releases suddenly and causes a
Rockmass varies from massive, layered and jointed to heavily rockburst. Therefore, by applying a cyclic compressive strength
crushed conditions. In addition, dynamic loading has a broad range test, the energy storage capacity of rock can be estimated. As it is
of frequency, amplitude, and wavelength. Therefore, ground beha- shown in Fig. 2a, Uds is the portion of energy which is dissipated
viour varies widely considering the rockmass and dynamic loading due to initiation and propagation of micro-cracks in the rock sam-
conditions. The most common types of strain burst and seismic ple (plastic deformation). Uel is the portion of energy which is con-
failure mechanisms in different ground types are categorised into sumed for elastic deformation and stored in the rock. This portion
four primary ejection types based on various factors as shown in of energy stored during the loading process up to point A could be
Fig. 1. released gradually by unloading or suddenly by failure. The ratio
Fig. 1a shows the mechanism of strain burst during ejection of a between elastic strain energy and dissipated energy (index of
volume of rock due to stress concentrations or induced stresses. In strain energy) could be used as a criterion or an indicator of rock-
this condition, discontinuities have a minor effect on ejection, so it burst potential.
is difficult to predict the volume of rock to be ejected and even
sometimes the likelihood of an ejection. F ¼ Uel=Uds ð1Þ
 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582 575

Discontinuities Discontinuities

Seismic event

Tunnel
Tunnel

(a) Induced stress rock burst (b) Buckling and spalling caused by induced stress

Discontinuities
Discontinuities

Seismic event Seismic event

Tunnel Tunnel

(c) Rock ejection caused by a seismic event (d) Instability in back due to loosening
and/or a seismic event

Fig. 1. Failure mechanisms for underground deep and high-stress tunnels due to induced stress and seismic events.

σ σ The fourth criterion considers both the state of in-situ stress in

σ max σ max the rockmass and the mechanical property of rock is expressed by:

σA A σA A T s ¼ rh =rc ð4Þ
In Eq. (4), rh is the tangential stress in the rockmass surround-
ing the openings or stopes (MPa), and rc is the uniaxial compres-
sive strength of rock (MPa). A larger Ts indicates a more violent
Es probable rockburst [14].
φel
φds A summary of these criteria is shown in Table 1.
φelm Four indexes are available in this table indicating whether a
O εp εel ε O ε cr ε
rockburst event will be low, strong or violent based on estimated
εt or calculated amount of each index. The indexes on the left side
(a) (b) of the range indicate low potential, and on the right side of the
range indicate strong or violent potential of rockburst.
Fig. 2. Analytic calculation of energy in the rock sample cyclic loading of after
Kwasniewski, Szutkowski) (a) [13] and calculation of potential elastic strain energy
(b) [2]. 2.2.2. Estimation of failure volume and ejection velocity
Estimation of failure thickness and ejection velocity will allow
Investigations demonstrate that the potential energy of elastic the estimation of ground demand by calculating the potential
strain (PES), in other words, the elastic strain energy which is energy release (stored energy in flying rock) by the prospective
stored in a unit volume of rockmass, is another criterion that could volume of ejected rock and the estimated velocity of ejection.
scale the shock and rockburst occurrence [13]. As it is depicted in The energy demand on ground support due to a block ejected
Fig. 2b, the maximum elastic strain energy which could be stored from the backs, wall or floor could be calculated by the following
in a sample of rock before the peak strength is given by: Eq. [3]:

PES ¼ Uelm ¼ r2c =2Es ð2Þ Energy Demand ¼ 1=2mv 2e þ qmgd ð5Þ
where rc is the uniaxial compressive strength (MPa), and Es is the In this equation: m = the mass of the ejected block (kg); ve = the
unloading tangential modulus (MPa). ejection velocity of the block (m/s); g = acceleration due to gravity
The third criterion is the index of rock brittleness which is (m/s2); d = distance the ejected block has travelled (m); and q = 1, 0
defined as following: or 1 for ejection from the backs, wall or floor respectively.
The second term in Eq. (5) contributing to the energy demand
B ¼ rc = rT ð3Þ
(qmgd) represents the influence of gravity. Gravity adds potential
In which rc is the uniaxial compressive strength (MPa), and rT energy to rocks ejected from the backs and reduces the energy of
is the tensile strength of the rock (MPa). Based on this criterion, the a block ejected from the floor, while not contributing to ejection
lesser index indicates the probability of the more violent rockburst. from the wall [3].
576 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582

Table 1
Rockburst potential based on intact rock property.

Description Index Potential of rockburst

Low Strong Violent
1 Index of strain energya [13] F = Uel/Uds 2 5
2 Potential energy of elastic strain (kJ/m3) [13] PES = rC 2/2Es 50.0 100 150.0 200 250.0
3 Rock brittleness [14] B = rC/rT 40.0 26.7 14.5
4 Ratio of tangential to compressive strength [15] Ts = rh/rC 0.3 0.5 0.7
a
Based on tests on coal specimens to provide the intensity of shocks or coal bombs.

If we consider the energy demand per square meter of excava- The empirical EVP index proposed makes use of these two
tion surface and substituting tq for m, the equation becomes [3]: components:
1 EVP ¼ ðdamage initiation factorÞ
Energy Demand per m2 ¼ t qV 2e þ qtqgd ð6Þ
2
 ðdepth of failure factorÞ ¼ ðE1=E2Þ  ðE3=E4Þ ð10Þ
In which: t = thickness of failed rock at the excavation surface (m);
and q = rock density (kg/m3). In order to consider the distance and magnitude of the seismic
Therefore, the critical factors required for energy demand are: event involved in each case history, the EVP data was compared to
peak particle velocity, which is assumed to equal the velocity of the fifth factor, peak particle velocity (PPV) to create a single index
ejection (Ve); excavation closure or ejection distance (d), and the called rockburst damage potential (RDP), as shown here:
mass of ejected material, which is a function of the failure thick- Rockburst damage potentialðRDPÞ ¼ EVP  PPV ð11Þ
ness (t) and the rock density (q).
The excavation closure (or ejection distance) ‘‘d” is used in the The respective distributions of these factors show that, in gen-
gravity component of the energy demand equation and is only eral, an increasing level of rockburst damage is associated with:
applicable when the design is being undertaken for the backs. It
represents the work done by the support system to halt the down-  Increasing stress conditions (E1);
ward movement of the rockmass. An approach is to use the dis-  Decreasing ground support system capacity (E2);
placement capacity of the ground support elements in the backs  Increasing excavation span (E3);
as a guide. In practice, the displacement capacities of the support  Decreasing geology factor (E4); and
element that fails first in a rockburst can be used for ‘‘d”. The  Increasing peak particle velocity (PPV).
results of drop weight dynamic testing of support elements can
be used to assist in determining appropriate ‘‘d” values. The above-explained procedure can be used to predict the level
The fracturing due to induced stress, blast damage, geological of rockburst potential. This method needs more experiments and
structure or a combination of all these three factors can form the practical feedback to prove or modify.
failure volume or mass of ejected rock which loads the support sys- In most cases, it is difficult to carry out a specific design because
tem. Failure volume can be estimated by various methods in an the rockmass factors that define demand cannot be dependably
excavation. A borehole camera survey can help to find the potential evaluated. Therefore, the rockmass demand can be described qual-
discontinuities for ejection and hence the probable volume of rock. itatively. As explained in Table 3, qualitative demand categories of
Numerical modelling also can be used for the estimation of the fail- rockmass could be defined in terms of low, medium, high, very
ure mass by measuring the overstressed zone surrounding an exca- high, and extremely high energy demand per square meter as well
vation, in other words, the zone around an excavation in which the as surface displacement and reaction pressure. Similarly, such a
stress exceeds the rock strength. Empirical estimation methods are rating can classify the reinforcement system in order to satisfy
also available. the rock demand [19].
Table 2 summarises the methods of estimating the failure vol-
ume for use in design calculations. The thickness should be calcu- 3. Dynamic rock support and reinforcement classes and tests
lated via as many possible as the mentioned methods in the table,
and the maximum thickness should be used in the calculation. The reinforcement and support system is a critical measure to
prepare a safe workplace as well as increase the longevity of a
2.2.3. Rockburst damage potential stable opening. An effective support system influences the safety
Heal [4] has established a method for assessing the likelihood of of workforces and equipment along with the economical mine
rockburst damage occurring at particular excavations in seismi- extraction. Different sorts of reinforcement and support systems
cally active underground mines. In this approach, five factors are are required for a particular application rely on a few elements
combined into a single index for determining the potential for including: the geometry of the excavation, the strength of the rock-
rockburst damage at a given location in an underground mine. mass, stresses present in the rock, corrosion and weathering pro-
Excavation vulnerability potential (EVP) is proposed as an index cesses, and blasting practices.
to empirically quantify the effect of local site conditions on rock- The primary method to lighten the impacts of mine seismicity is
burst damage. It makes use of four of the five mentioned factors, the design of a practical geometry and appropriate mining
those not related to the source of the seismic event: sequence. A rock support plan would be a complementary step
intending to mitigate the rockburst impact. A ground demand-
 E1: The stress conditions (r1T/UCS); energy dissipation capacity approach is a vital step in such circum-
 E2: The energy capacity of the installed ground support system stances. Therefore, acquiring the knowledge of energy dissipation
(in kJ/m2); capabilities of elements of a support system including the rein-
 E3: The excavation span (in m); and forcement, surface support, connecting elements and faceplate &
 E4: The presence of seismically active major geological nut is necessary as well as a whole support system as an integrated
structure. system.
 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582 577

Table 2
Failure thickness estimation [17,18] (after Heal [4]).

Table 3
capacity or energy dissipation capacity of the rock support is under
Typical rockmass demand for ground support design [19].
investigation by researchers [20]. The primary challenge in mea-
Demand category Reaction pressure Surface displacement Energy suring the dynamic capacity of the ground support including the
(kPa) (mm) (kJ/m2) rockbolt is to prepare repeatable loading conditions similar to
Low <100 <50 <5 what is experienced at a supported face during a seismic event.
Medium 100–150 50–100 5–15 Providing a good monitoring system and qualified data acquisition
High 150–200 100–200 15–25
Very high 200–400 200–300 25–35
apparatus along with well-controlled equipment are requirements
Extremely high >400 >300 >35 of a dynamic testing facility in order to acquire reliable data and
meaningful analysis.
‘‘Drop testing” has been under the attention of researchers to
convey kinematic energy to ground support elements in order to
3.1. Dynamic capacity of rockbolts measure energy dissipation capacity [3,9,21–27]. The momentum
transfer concept has been utilised by some other researchers [6].
Implementation of a dynamic resistance support system is the In this method, deceleration of a dropped reinforcement sample
most common method of stabilising an underground opening in attached to a mass is measured and the amount of energy con-
mines. Rockbolts along with surface support comprised of mesh sumed particularly for deformation and failure of sample is calcu-
and shotcrete, play a crucial role as one of the main elements of lated. Employing a simulated controlled blasting process as the
a support system. A tunnel that experiences seismic activities like dynamic load applied on a completely supported area along with
a rockburst needs to be supported by appropriate elements, cap- a well-instrumented system is another category of measurement
able of tolerating dynamic loading. This area in geotechnical engi- of dynamic performance of a support system as an integrated sys-
neering is still under development. In other words, the dynamic tem [7,28]. In addition, back calculation of support capacity has
578 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582

also been performed by Heal [4] which can be assumed as another characteristics, including wavelength, amplitude and frequency
method to estimate the dynamic support capacity. created by blasting are different to those produced by large seismic
events. Normally, the wavelengths in the seismic events are longer
3.1.1. The ‘drop test’ and frequencies are lower in comparison to those in blasting.
The drop test rig is a controlled laboratory facility to investigate Obviously, to investigate and understand the behaviour of rock-
the dynamic behaviour of ground support elements submitted to a mass and ground support elements, a reproducible or repeatable
seismic event simulated by sudden loading of a dropping mass simulated dynamic event would be a great success. Many research-
from a predetermined height [3,9,21–27,29,30]. This test has expe- ers have tried to employ the blasting method for simulation of a
rienced numerous amendments and has turned into a standard rockburst, but there are few or small number of successful exper-
testing technique for laboratory assurance of rockbolt energy iments. Distortion by gases and not enough generated energy to
absorption or dissipation capacity. There are also various difficul- produce premeditated destruction have been the main reasons of
ties required with this test including slow instrumentation reac- ineffective experiments. Nevertheless, the behaviour of the whole
tion, uncontrolled vibrations in the loading system, and other support system as an integrated system can be investigated with
sources of unmeasured energy losses [31]. The advantage of this this method.
test facility is its repeatability and cost effectiveness as soon as it
is assembled. A number of drop testing equipment has been con- 3.1.3. Momentum transfer method
structed during the last twenty years in Canada, South Africa and The momentum transfer concept has been employed by the
recently in Australia to be able to perform dynamic performance Western Australian School of Mines (WASM) via a dynamic loading
assessment of ground support elements. Although a standard system in order to find out the energy dissipation capability of the
method of testing has been available, these rigs have been con- ground support elements or system. This equipment utilises a sam-
structed with considerable dissimilarities which make the exami- ple of reinforcement attached to a mass to apply a dynamic impact
nation of their outcomes to some degree complicated or not to the sample by dropping them from a certain height and mea-
comparable [20]. surement of deceleration after impact. The testing facility is cap-
A rockbolt or cable bolt, cement or resin encapsulated in thick- able of testing different types of rockbolts, cable bolts, or
wall steel pipe to replicate the rockmass, is frequently used in the reinforcement systems, prepared sample of surface support or a
drop testing experiments. Despite the fact that a specific thickness mixture of both, to be able to assess the mechanism of dissipation
and measurement of steel tube were given to provide similar con- of the energy by a ground support system and interaction between
finement of the in-situ rockmass with the same magnitude, the the surface support and reinforcement and the mechanism of the
steel pipe cannot completely replicate the rockmass which may transfer of the dynamic load [6].
introduce an error of some degree into the estimation [8]. The concept of this facility is illustrated in Fig. 3. Using a
In spite of the fact that there have been critical enhancements dropped mass of 2000 kg as the simulated ejected rock with an
made to the drop testing mechanical assembly, it is still not illus- impact velocity of 6 m/s is a standard arrangement for testing of
trative of in-situ conditions. The drop test technique has numerous rock reinforcement. This arrangement provides a kinetic energy
presumptions that would influence the performance of the support of 36 kJ applied to the test sample and must be dissipated by the
elements contrasted with their genuine performance in the field. support element. The buffers have to absorb the energy of the
Moreover, the drop tests deliver results of individual support ele- beam as well as a portion of the energy of the dropping mass.
ments that need to be compiled and consolidated to design the The excess energy is applied to the test sample following the
support system. It is helpful to take the outcomes from the differ- impact because of the change in potential energy of the dropping
ent reinforcement elements and the surface support and assemble mass. Making a radial cut artificially in steel pipe simulates the dis-
them together. However, providing a cost effective, controlled and continuity in the rockmass typically situated 1.0 m beneath the
repeatable procedure for estimating the support elements’ proper- bearing plate [39].
ties in a laboratory is its outstanding advantage. Characteristically, the investigation of a sample of reinforce-
ment or support system has to be based on first impact loading
3.1.2. Blast simulation that can be a single large dynamic impact. Therefore, the testing
Blast simulation experiments have been performed in-situ try- equipment has to have sufficient energy or enough capacity to be
ing to recreate the seismic event via the blasting to measure the able to exceed the strength of the sample with a single impact.
consequences on most common ground support systems [7,32–
37]. In-situ simulated blasting testing to investigate the rock sup-
port behaviour and performance was innovated by Ortlepp [38].
Drop beam
In comparison to drop testing, the simulation of rockbursts by (simulated as rock which is
blasting has a large level of difficulty. Performing such a destruc- not ejected)

tive test in active mines during operation of other activities needs

sophisticated coordination with operative units while the logistic Simulated
discontinuity Loading mass as
of setting up and carrying out the tests is not straightforward, simulated ejected rock
and the cost is also high. The positive points of the method is the
testing of the support system as an integrated system which is
completely installed in place as opposed to individual support ele- Ultrasonic sensor to
record beam
ments. Issues, for example, installation procedures and the interac- Computer and movement during
data compression and
tion with the rockmass were also investigated, and shortcomings acquisition recovering the buffer
system (22
of the whole system were underscored [20]. channel 25
kHz sampling
It is worth mentioning that the movement of ground in blasting rate of each) Buffers
is not similar to that of a rockburst because of seismic events. The Camera 250 fps
gas pressure is not available in the rockburst condition while in
blasting it is accompanied by the shock wave, as sometimes the
generated gases quickly expand and may conclude to unpre- Fig. 3. Dynamic testing facility with momentum transfer concept (after player,
dictable results at the test location. On the other hand, the wave Thompson [39]).
 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582 579

Based on previous experiments, it has been proved that the multi- 3.1.5. Large-scale dynamic test rig for ground support
ple loading cause the measured result to overestimate the capacity In order to examine the ground support as an integrated sys-
in comparison to the results of a single large impact. The WASM tem, Geobrugg Company has constructed a dynamic test rig. Using
testing equipment is capable of applying 120 kJ of kinetic energy this rig, it is possible to apply a dynamic load to a sample of a com-
to the sample [39] which is more than the capacity of most com- plete support system containing a 3.6 m  3.6 m sample of surface
mon rock bolts. support combined with four dynamic rockbolts. Because the large
It is practical to calculate the dynamic force-displacement dia- sample includes all support elements, it is capable of demonstrat-
gram of the support element via a well instrumented and moni- ing the performance of the surface support and the reinforcement
tored system. The portion of the applied kinetic energy, which is in combination together as well as the connecting and terminating
dissipated by the prepared sample of support, would be deter- elements [11,12]. Fig. 4 shows the test setup.
mined by calculation of the area under the force-displacement As it can be seen in the figure, a horizontal chain link mesh is
graph. Another portion of energy that is absorbed by buffers can connected to the main steel frame using lacing wire ropes, while
be calculated separately for every test. The accelerometers assist the mesh is held by four dynamic rockbolts. Surface support simu-
in evaluating and computing the deceleration response of the sys- lated by shotcrete or concrete slab could be poured over the wire
tem. Alternatively, it can be calculated by a fast computerised mesh engaged with the four rockbolts via terminating and con-
video camera, measuring the relative displacement of a target by necting elements. Some natural rock boulders and gravel are
object tracking software. placed on top of the slab sample to simulate broken rocks during
Finally, the underground ejection velocity is considered as the a rockburst event contained by surface support. An impact plat-
relative velocity between the loading mass and the dropping beam. form made of steel is placed over the gravel to distribute and trans-
The ejection phenomenon happens and a block of rock, which was fer the impact of a dropped block to the gravel layer, natural rock
at rest or stationary under the stress at the wall or vault of the tun- boulders, and simulated surface support. The mass of the dropped
nel, quickly accelerates and reaches a peak velocity. The velocity block is 6280 kg and it can be lifted and dropped from a maximum
returns to zero if the ground support system tolerates the dynamic height of 3.25 m limited by a guiding rail. One of the four bolts is
impact. Compared to a strong ground support system, a weak or instrumented by two load cells at both ends. Two high-speed cam-
soft support system would be a reason for larger displacement eras are installed in front of the main frame, the upper one for the
and greater ejection velocity. The most important aspects of the filming of the test block movement and impact and the lower one
ground support design that has to be considered in a mining oper- to monitor the support with several measuring targets on the mesh
ation is the maximum permissible deformation of the reinforce- and bolts with a computer tracking program to evaluate the dis-
ment system and ensuring that the surface support has enough placement, velocity and acceleration of the targets. Dissipated
toughness to tolerate the displacement [39]. energy can be calculated by the difference in potential energy of
the test block before and after the impact [11,12].
3.1.4. Back-calculation Testing a large scale of the support sample as an integrated
Back analyses of the actual rock ejection and the associated sup- system submitted to a dynamic impact is the strong point of this
port system is potentially a way of estimation of the dynamic testing rig. Engineers, to some extent, can evaluate energy dissi-
capacity of the ground support. The problem is predicting the loca- pation capability of a ground support system exposed to dynamic
tion of an ejection due to its randomness and other uncertainties, impact and compare the compatibility of the elements in the pre-
and consequently lack of sufficient monitoring to collect enough pared sample. The result would help the designer to avoid leaving
data regarding the event, for example, velocity of the ejected mass. a weak link in a support system because the weakest link in a
Therefore, back analyses of driven events like blasting would be an support system affects and limits the maximum capacity of the
appropriate method to address this issue. whole system.
A comparison between the test results of simulated rockbursts One weak point of the system is that a single drop would not
with back analyses of absorbed energy in some case studies has cause the support system to fail under test and multiple drops
been performed by Heal [4]. A correlation has been found between can conclude in an overestimation of the energy dissipation capac-
the back calculation of case studies and the simulated rockburst ity of the rockbolts or even the whole support system.
results, but the method has not been proved yet nor used by other There are not many published results of this testing facility
researchers. It seems that this method with some modification can and perhaps this is due to a limited number of support systems
be an approach to calculating rock support dynamic dissipation tested. Therefore, the performance of it can only be evaluated
capacity at the real scale. after publication of more test results and comparison to real case

Crane with hook

Concrete block
6280 kg

Impact platform
Load cell

Rock bolts

Rock blocks

Concrete layer
Mesh Load
cell

Fig. 4. The large-scale dynamic testing rig of geobrugg [11].

580 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582

studies. On the other hand, it seems that the monitoring data is and the quality of installation is another important factor. As it can
not enough to calculate the portion of energy dissipated by a sup- be seen in the graph, one of the Conebolts tolerates more than 300
port sample because the steel frame absorbs a part of the poten- mm deformation and absorbs or dissipates 60 kJ of the ground
tial energy of the testing block by deflection and vibration that released energy. In comparison, two other Conebolts tolerate less
cannot be measured or calculated by the predicted monitoring than 150 mm and less than 300 mm and can dissipate 20 kJ and
system. It is also worth mentioning that the testing facility does 35 kJ, respectively. Grout quality is a major factor for Conebolts
not completely replicate the seismic phenomenon that happens because a strong cement grout could lead to higher initial loading
in the ground. and early rupture while soft cement grout leads the rockbolt to
early sliding and not reaching its maximum load capacity. In both
3.2. Rockbolts energy dissipation capacity cases, energy absorption capacity of a rockbolt dramatically drops.
So before starting to implement a ground support scheme, it would
In this part, the most common types of rockbolts are discussed be necessary to plan a test program to determine the conditions for
and divided into different capacity categories. It is assumed that optimum performance of the rockbolts. Examples of influencing
the surface support system (including shotcrete, mesh and nut) parameters include grout mix design, curing time and preloading.
are acting appropriately and transfer the load to the rockbolt. Then The result of the test program should be used to develop a quality
the rockbolt would be the central element absorbing and dissipat- control plan.
ing energy. Considering rockbolts’ energy absorption as shown in Fig. 5 and
Typical load–deformation behaviours of different rockbolts discussed above, suitable rockbolt type selection for various
under the loading test are collected and illustrated in Fig. 5. ground demand categories are proposed in Table 4. This table could
According to the load-deformation capacity, the rockbolts are be an initial guideline to narrow the choices, and it is evident that
classified into five groups namely, stiff, medium yielding, high complementary studies such as dynamic tests are required for
yielding, very high yielding, and extremely high yielding rock- detail design. Although there are some newer types of rockbolt like
bolts. As shown in the figure, a category of rockbolts, such as Dynamic Omega-Bolt which can absorb 22–35 kJ in static and
expansion shell and resin/grout encapsulated rebars, are concen-
trated on the left side of the plot and represent stiff rockbolts
with less than 50 mm deformation capacity and less than 5 kJ Table 4
energy absorption capacity. The second category such as Split Demand–capacity based support selection.
set, Swellex, Roofex and Yield-Lok are the rockbolts which can
Ground demand Reinforcement selection
tolerate deformations between 50 mm and 100 mm with an
Surface Energy Recommended reinforcement Capacity
energy absorption capacity between 5 kJ and 15 kJ. The D-Bolt,
displacement (kJ/m2) category
Conebolt, Swellex, Roofex and Yield-Lok which are high yielding (mm)
rockbolts could lie in the next category. For deformation capacity
<50 <5 Expansion shell rockbolt, Resin/ Low/stiff
greater than 200 mm, Conebolt, Garford and Roofex (possibly cement steel rebar,
with small spacing) fall into the very high yielding category, 50–100 5–15 Split set, Swellex, Roofex, Yield- Medium
and just Conebolt and Garford are suitable for the extremely high Lok
yielding category. 100–200 15–25 Swellex, D-Bolt, Conebolt, Roofex, High
Yield-Lok
An important fact related to high yielding rockbolts is that they
200–300 25–35 Roofex, Conebolt, Garford Very high
show different behaviour depending on loading conditions and >300 >35 Conebolt, Garford Extremely
other environmental circumstances. Loading velocity is one factor high
that can change the load and deformation capacity of yielding bolts

Resin grouted 22 mm (5.7 kJ) Epansion shell 17.3 mm Swellex dowel (15 kJ) Split set stabiliser (7 kJ)
rockbolt (3 kJ)
Resin grouted 22 mm steel Cement grouted 20 mm steel Conebolt (35 kJ) Conebolt (20 kJ)
Rebar (4.5 kJ) Rebar (6.5 kJ)
Static performance for D-bolt Dynamic performance for Cement encapsulated
Conebolt (60 kJ) D-bolt (40 kJ) threaded bar (5 kJ)
(20 kJ)
Plain steel bar 20 mm B500C
(40 kJ) Garford bolt (50 kJ) Roofex (20 kJ) Yield-Lok (17 kJ)
300

Dynamic performance for D-bolt (40 kJ)

250
Resin grouted 22 mm (5.7 kJ)
Cement encapsulated threaded bar (5 kJ)
Conebolt (20 kJ) Conebolt (60 kJ)
200
Load (kN)

Cement grouted 20 mm steel Rebar (6.5 kJ) Dynamic performance for D-bolt (40 kJ)
Resin grouted 20 mm steel Rebar (4.5 kJ)
150 Conebolt (35 kJ)
Static performance for D-bolt (20 kJ)

Yield-Lok (17 kJ) Plain steel bar 20 mm B500C (40 kJ)

Garford bolt (50 kJ)
100

Epansion shell 17.3 mm rockbolt (3 kJ) Swellex dowel (15 kJ) Roofex (20 kJ)
50 Split set stabiliser (7 kJ)

Stiff Medium yielding High yielding Very high yielding Extremely high yielding

0 50 100 150 200 250 300 350 400

Displacement (mm)

Fig. 5. Load-deformation behaviour of different rockbolts (modified after [29,30,39–41]).

 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582 581

Capacity category

Capacity Low/Stiff Medium High Very high Extremely high

energy ˘5 kJ 5-15 kJ 15-25 kJ 25-35 kJ ˚35 kJ
displacement 0 mm 50 mm 100 mm 150 mm 200 mm 250 mm 300 mm 350 mm
Cement Rebar
Resin Rebar
Expansion shell

Rock bolt types

Split set
Swellex
Roofex
Yield-Lok
D-bolt
Cone bolt
Garford

Fig. 6. Energy dissipation capacity category of different types of reinforcement.

dynamic conditions [42], they need more laboratory and industrial tem are not entirely capable of simulating the real conditions
experimentation. occurring in the ground.
Fig. 6 shows the energy dissipation capacity of different types of Having an estimation of both factors, the ground demand and
reinforcement. Choosing a specific type of rock reinforcement, the the support capacity, is essential, therefore, even with a large
figure shows the range of energy dissipation and deformation amount of uncertainty, designers can compare these two factors
capacity under each named capacity category. to define a factor of safety. In addition, the methods could be mod-
Based on the expectation of the deformation and energy ified and calibrated in a certain area by the probable occurrence of
demand of a location, the ground demand relates to the relevant seismic activities similar to observational methods. Comparison of
categories in this table. The range of suitable reinforcement for the support systems tested by multiple facilities assists with pro-
the category is proposed in the ‘‘Rock Bolt Types” column. The moting the design for the next step.
expected deformation and ground demand are complicated though
and come from the methods explained in Section 2.2 as well as pre-
vious experiences and engineering judgments. 5. Conclusions

Under seismic conditions in mines, the idea of improving, con-

3.3. Considerations of linking and terminating arrangements of
serving, and mobilising the inherent strength of the ground to be
reinforcements
self-supported is not valid enough while energy dissipation capa-
bility and large deformation capacity of support system is the pri-
The reinforcement connects to the surface support by linking
mary objective. In this research, the ground demand and likelihood
and terminating arrangements like nuts and bearing plates, split
of a dynamic event have been estimated using different methods.
set rings, or the sealing weld and soft ferrule on Swellex. The
Despite the low accuracy of these estimation methods due to many
ejected mass applies the dynamic load to the surface support or
assumptions, they can assist in the selection of a relatively appro-
containment support. The load needs to be passed via the linking
priate support system at preliminary design stages. The design can
and terminating arrangements and transferred to the ground
be modified with observations during construction progress.
through the reinforcement. Everyone of these elements have to
Stiff behaviour at the beginning of the loading, along with high
be able to tolerate the applied dynamic load independently and if
strength and yielding capability by increasing deformation, are
any of them failed, the load would no longer transmit to the
essential qualities of the support components under dynamic load-
ground and ejection would occur from in-between the rockbolts
ing conditions in order to dissipate a sudden release of energy. To
[3,4].
estimate the capacity of rock support systems exposed to seismic
Some experiments show that the capacity of the bearing plate
events, a number of estimation methods including laboratory drop
under a dynamic loading condition is much less than their nominal
tests, simulated rockbursts, back calculation, momentum transfer
load capacity [43,44]. Therefore, in designating each ground sup-
concept and large-scale dynamic test were discussed. Although
port system, it is critical to be sure that the linking and terminating
various assumptions and interpretations are needed to employ
elements have adequate impact loading capacity to transfer the
the results of dynamic tests, more dynamic capacity measurement
load to the reinforcement and avoid of local failure of the surface
of support elements is required to cover the wide range of possible
support.
energy released and resulting deformation. On the other hand,
ground support reacts in different ways under different circum-
4. Discussion stances. The velocity of ejection (dynamic loading velocity), quality
of grouting of rockbolts and appropriate linking between all ele-
Ground support system design in a seismically active ground or ments are some of the known factors that affect the performance
rockburst prone area needs specific consideration regarding evalu- of the ground support system. The arrangement of a test program
ation or estimation of the released or transferred energy to the sur- before finalising the design is vital to ensure a successful design.
face of the opening on one hand, and knowing the energy Ground demand is estimated using the methods discussed
absorption or dissipation capacity of the support system on the along with an associated degree of uncertainty. However, to begin
other hand. Design of a support system at a certain location under- with, the potential for rockburst could be assessed through labora-
ground requires an evaluation of both ground demand and support tory tests on intact rocks. Estimation of failure thickness and veloc-
capacity, in order to design a reliable support system. The pre- ity of ejection could support the assumptions and results of the
sented methods in the evaluation of ground demand have a large laboratory tests. Using rockburst damage potential, the previous
degree of uncertainty while the testing methods of the support sys- result could be cross-checked, and this could also be summarised
582 R. Masoudi, M. Sharifzadeh / International Journal of Mining Science and Technology 28 (2018) 573–582

into a qualitative description. Using ground demand – support [21] Yi X, Kaiser P. Impact testing for rockbolt design in rockburst conditions. Int J
Rock Mech Mining Sci & Geomech Abstr 1994;11:671–85.
energy dissipation capacity (Table 4 and Fig. 6), the rockbolt type
[22] Ortlepp W, Stacey T. Testing of tunnel support: dynamic load testing of
selection was introduced. The selected rockbolt can be tested, ver- rockbolt elements to provide data for safer support design. Safety Mines Res
ified, and modified by proper dynamic testing or observation of Advis Committ 1998:1–49.
progress during construction. The reliability of the support ele- [23] Ortlepp WD. Rock fracture and rock bursts. In: Monograph Series M9.
Johannesburg; 1997.
ments would be monitored and back calculated after initial instal- [24] Ortlepp WD, Stacey TR, Kirsten HAD. Containment support for large static and
lation and following excavation progress. This will allow the dynamic deformations in mines. In: international symposium-4th rock
support selection and details to be modified based on monitoring support and reinforcement practice in mining. Kalgoorlie; 1999. p. 359–66.
[25] Stacey TR, Ortlepp WD. Retainment support for dynamic events in mines. In:
and back calculation, progressively and continuously. International symposium-4th Rock support and reinforcement practice in
mining. Kalgoorlie; 1999. p. 329–33.
References [26] Ortlepp W, Swart A. Extended use of the Savuka dynamic test facility to
improve material and analytical technology in deep-level stope support. Safety
Mines Res Advis Committ 2002:1–57.
[1] Kwasniewski M, Wang J. 3-D numerical modeling and study of mine tremors
[27] Gaudreau D, Aubertin M, Simon R. Performance assessment of tendon support
associated with coal mining in vicinity of major of faults. Pupls Inst Geophys
systems submitted to dynamic loading: École polytechnique; 2004.
1999; 22: 351–64.
[28] Tannant DD, Brummer RK, Yi X. Rockbolt behaviour under dynamic loading:
[2] Wang JA, Park HD. Comprehensive prediction of rockburst based on analysis of
Field tests and modelling. Int J Rock Mech Mining Sci Geomech Abstr
strain energy in rocks. Tunn Undergr Space Technol 2001;16:49–57.
1995;32:537–50.
[3] Kaiser PK, McCreath D, Tannant D. Canadian rockburst support
[29] Li CC. A new energy-absorbing bolt for rock support in high stress rock masses.
handbook. Geomechanics Research Centre: Laurentian University, Sudbury;
Int J Rock Mech Min Sci 2010;47:396–404.
1996. p. 314.
[30] Li CC, Stjern G, Myrvang A. A review on the performance of conventional and
[4] Heal DP. Observations and analysis of incidences of rockburst damage in
energy-absorbing rockbolts. J Rock Mech Geotech Eng 2014;6:315–27.
underground mines. Perth, Western Australia: University of Western
[31] Soleimani SM, Banthia N. A novel drop weight impact setup for testing
Australia; 2010.
reinforced concrete beams. Exp Tech 2014;38:72–9.
[5] Villaescusa E. Geotechnical design for sublevel open stoping. Boca Raton: CRC
[32] Archibald JF, Baidoe JP, Katsabanis PT. Rock burst damage mitigation benefits
Press; 2014.
deriving from use of spray-on rock linings. In: Proceedings the 3rd
[6] Player JR, Villaescusa E, Thompson AG. Dynamic testing of rock reinforcement
international seminar on surface support liners: thin spray-on liners,
using the momentum transfer concept. In: Fifth international symposium on
shotcrete and mesh. Quebec City; 2003.
ground support in mining and underground construction. Perth; 2004. p. 327–
[33] Espley SJ, Heilig J, Moreau LH. Assessment of the dynamic capacity of liners for
40.
application in highly-stressed mining environments at Inco Limited. In:
[7] Heal DP, Potvin Y. In-situ dynamic testing of ground support using simulated
Surface support in Mining. Perth; 2002. p. 187–92.
rockbursts. In: Deep mining. Perth; 2007. p. 373–94.
[34] Hagan TO, Milev KB, Spottiswoode DM, Hildyard AM, Grodner SM, Rorke MW,
[8] Li L, Hagan P, Saydam S. A review of ground support systems performance
et al. Simulated rockburst experiment - an overview. J S Afr Inst Min Metall
subjected to dynamic loading. Int Soc Rock Mech; 2014.
2001;101:217–22.
[9] Plouffe M, Anderson T, Judge K. Rock bolts testing under dynamic conditions at
[35] Hildyard M, Milev A. Simulated rockburst experiment:numerical back-analysis
CANMET-MMSL. In: Proc 6th int symp on ground support in mining and civil
of seismic wave interaction with the tunnel. J S Afr Inst Min Metall
engineering construction, Cape Town, S Afr Inst Min Metall Symposium Series
2001;101:223–34.
S; 2008. p. 581–95.
[36] Hildyard M, Milev A. Simulated rockburst experiment: development of a
[10] Player J, Villaescusa E, Thompson A. An examination of dynamic test facilities.
numerical model for seismic wave propagation from the blast, and forward
In: Australian mining technology conference; 2008. p. 349–79.
analysis. J S Afr Inst Min Metall 2001;101:235–46.
[11] Roth A, Cala M, Brändle R, Rorem E. Analysis and numerical modelling of
[37] Reddy N, Spottiswoode S. The influence of geology on a simulated rockburst. J
dynamic ground support based on instrumented full-scale tests. In:
South Afr Inst Min Metall 2001;101:267–74.
proceedings of the seventh international conference on deep and high stress
[38] Ortlepp WD. An empirical determination of the effectiveness of rockbolt
mining. Sudbury; 2014.
support under impluse loading. In: Int symp on large permanent underground
[12] Morissette PNR. A ground support design strategy for deep underground
openings. Oslo; 1969. p. 197–205.
mines subjected to dynamic-loading conditions. Canada: University of
[39] Player J, Thompson A, Villaescusa E. Dynamic testing of reinforcement system.
Toronto; 2015.
In: 6th international symposium on ground support in mining and civil and
[13] Kwasniewski M, Szutkowski I, Wang JA. Study of ability of coal from seam 510
engineering construction. Cape Town; 2008. p. 597–622.
for storing elastic energy in the aspect of assessment of hazard in Porabka-
[40] Galler R, Gschwandtner G, Doucet C. Roofex bolt and its application in
Klimontow Colliery. Sci Rept Silesian Technical University; 1994.
tunnelling by dealing with high stress ground conditions. In: ITA-AITES world
[14] Qiao CS, Tian ZY. Study of the possibility of rock burst in Dong-gua-shan
tunnel congress Helsinki, Finland; 2011.
Copper Mine. Chin J Rock Mech Eng 1998;17:917–21.
[41] Wu R, Oldsen J, Lamothe M. The Yield-Lok bolt for bursting and squeezing
[15] Wang YH, Li WD, Li QG. Fuzzy estimation method of rock burst prediction.
ground support. In: Proc 5th int seminar on deep and high stress mining,
Chin J Rock Mech Eng 1998;17:493–501.
Santiago, Australian; 2010. p. 301–8.
[16] Linkov AM. Rockbursts and the instability of rock masses. Int J Rock Mech
[42] Scolari F, Brandon M, Krekula H. Dynamic inflatable, friction rockbolt for deep
Mining Sci & Geomech Abstr 1996;33:727–32.
mining. In: Deep mining: eighth international conference on deep and high
[17] Wiles T. Reliability of numerical modelling predictions. Int J Rock Mech Min Sci
stress mining. Perth; 2017.
2006;43:454–72.
[43] Simser B, Potvin Y. The weakest link–ground support observations at some
[18] Haile A, Grave D, Sevume C, Le Bron K. Strata control in tunnels and an
Canadian Shield hard rock mines. In: Deep mining—proceedings of the 4th
evaluation of support units and systems currently used with a view to
international seminar on deep and high stress mining. Perth; 2007. p. 7–9.
improving the effectiveness of support, stability and safety of tunnels; 1998.
[44] Jan VS, Palape M. Behaviour of steel plates during rockbursts. In: Deep mining.
[19] Thompson A, Villaescusa E, Windsor C. Ground support terminology and
Perth, 2007.
classification: an update. Geotech Geol Eng 2012;30:553–80.
[20] Potvin Y, Wesseloo J, Heal D. An interpretation of ground support capacity
submitted to dynamic loading. Mining Technol 2010; 119: 233–45.