ARTICLE IN PRESS

International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404

Contents lists available at ScienceDirect

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

A new energy-absorbing bolt for rock support in high stress rock masses
Charlie Chunlin Li
The Norwegian University of Science and Technology (NTNU), Norway

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

Article history: An energy-absorbing rock support device, called a D bolt, has been recently developed to counteract
Received 17 May 2009 both burst-prone and squeezing rock conditions that occur during underground excavation. The bolt is
Received in revised form a smooth steel bar with a number of anchors along its length. The anchors are firmly fixed within a
10 December 2009
borehole using either cement grout or resin, while the smooth sections of the bolt between the anchors
Accepted 7 January 2010
Available online 1 February 2010
may freely deform in response to rock dilation. Failure of one section does not affect the reinforcement
performance of the other sections. The bolt is designed to fully use both the strength and the
Keywords: deformation capacity of the bolt material along the entire length. The bolt has large load-bearing and
Rock support deformation capacities. Static pull tests and dynamic drop tests show that the bolt length elongates by
Rock bolt
14–20% at a load level equal to the strength of the bolt material, thereby absorbing a large amount of
Ductile rock bolt
energy. The impact average load of a 20 mm D bolt is 200–230 kN, with only a small portion of the load
Energy-absorbing rock bolt
Yield support device transferred to the bolt plate. The cumulative dynamic energy absorption of the bolt is measured to be
47 kJ/m. D bolts were tested in three deep mines. Filed measurements show that D bolts are loaded less
than rebar bolts. This paper presents the layout and principle of the D bolt, and corresponding results
from static, dynamic, and field tests.
& 2010 Elsevier Ltd. All rights reserved.

1. Introduction gravitational rock falls. The main task of the bolt in a shallow
tunnel is to equilibrate the deadweight of loosened blocks that are
A challenge to underground excavation is stabilising the reinforced by the bolt. Therefore, the strength of the bolt is more
country rock that surrounds openings at depth. The essential important than its deformation capacity in low in situ stress
difference between rock at depth and rock near the surface is an conditions. At depths where in situ rock stresses are high,
increase in the in situ rock stresses. As a consequence of this however, it is no longer appropriate to use rebar bolts as rock
increase in the rock stresses, rock burst may occur in hard rocks, reinforcement devices. When used with weak and soft rocks, it is
or large squeezing deformations may appear in soft and weak frequently observed that either the face plate of the rebar is
rocks. It has been observed in many mines that such phenomena heavily loaded, or the thread of the bolt is pulled to failure [2]. In
start at a depth of about 600–800 m below the surface and the case of hard rocks, the rock spalls and sheds pieces like
become more obvious below 1000 m. Many metal mines, for onionskin behind the bolt plate, leaving a short section of the bolt
instance those in Sweden, Canada, West Australia, and South extruded out of the rock surface. In a deep cut-and-fill metal mine
Africa, at present conduct mining operations at depths below in Sweden, old rebar bolts are sometimes exposed on the work
1000 m and even down to 3000 m. At these depths, conventional faces of subsequent cut slices. It is observed that some of those
support devices [1] do not adapt well to the severe rock bolts fail due to only small amounts of shear and opening
conditions. Nevertheless, fully encapsulated rebar bolts are still displacement at rock joints/fractures [3]. The premature failure of
used in some deep mines, for example in Canada and Sweden, to the rebar bolts implies that rebar is too stiff to sustain rock
deal with rock burst problems. Fully encapsulated rebar has a dilations in high stress rock masses. To accommodate large rock
high load-bearing capacity; however, rebar cannot accommodate dilations, Split Sets have been used as reinforcement devices in
large rock dilations. A fully encapsulated rebar bolt would only many deep mines in Australia. Split Sets are indeed capable of
withstand a deformation of 2–3 cm when subjected to a fracture tolerating large rock deformations, but their load-bearing capacity
opening. The rationale of using rebar bolts is that they can provide is very low. Both rebar and Split Sets are low energy-absorbing
sufficient support to unstable blocks to prevent rock falls. This devices. In the early 1990s Ortlepp pointed out that support
rationale is considered appropriate in shallow locations, where devices used in deep mines should be able to carry high loads and
in situ rock stresses are low and the main risk is from also accommodate large deformations; that is, they should be
capable of absorbing a large amount of energy prior to failure [4].
The so-called cone bolt may be the first energy-absorbing device
E-mail address: charlie.c.li@ntnu.no designed to counteract rock burst problems [5]. The cone bolt has

1365-1609/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijrmms.2010.01.005
 ARTICLE IN PRESS
C. Chunlin Li / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404 397

been accepted as a dynamic support device in some Canadian load capacity was sensitive to the contact stress at the bolt–rock
deep mines. The use of energy-absorbing elements for dynamic interface.
rock support has been gradually accepted by many ground control Rock bolts can be classified according to their performance as a
engineers, particularly in the metal mining industry. strength bolt, a ductile bolt, or an energy-absorbing bolt. Strength
The D bolt is a new type of energy-absorbing support device bolts are those that provide a support load equal or close to the
invented in 2006. This bolt differs from other energy-absorbing intrinsic strength of the bolt material. Fully encapsulated rebar
bolts in that it is multi-point anchored in a fully grouted borehole. belongs to this category. Ductile bolts are able to accommodate
All other energy-absorbing bolts are two-point anchored in large rock deformations. Split Sets are an example of a typical
boreholes. The multi-point anchoring mechanism is very reliable ductile bolt. Energy-absorbing bolts are characterized by their
in the D bolt. Failure of one section (or one anchor) will not high load capacity and also their large deformation capacity.
negatively affect the rock reinforcement of the other sections of These three categories of bolts are illustrated in Fig. 1.
the bolt. In this paper, a brief overview of the history of energy-
absorbing rock bolts is provided. The D bolt is then introduced in
detail, including its layout and principle, and the results from 3. A brief review of energy-absorbing rock bolts
static, dynamic, and field tests.
The concept of an energy-absorbing support device was first
proposed in South Africa in the early 1990s [4]. The first energy-
absorbing rock bolt, the so-called cone bolt, was invented in South
2. Concept of the ideal bolt Africa [5]. The cone bolt consists of a smooth steel bar with a
flattened conical flaring forged onto one end. The smooth bar is
coated with a thin layer of lubricating material such as wax, so
All conventional rock bolts can be classified into three types in
that it easily detaches from the grout under pull loading. The bolt
accordance with their anchoring mechanisms: two-point an-
is fully encapsulated with either cement grout or resin in a
chored mechanical bolts (e.g., expansion shell bolt), fully
borehole. The dilation of the rock between the cone and the bolt
encapsulated rebar bolts, and frictional bolts (e.g., Split Set). The
plate will induce a pull load in the bolt shank. The cone is
pull and shear performances of conventional bolts have been
designed so that the conical end ploughs through the grout when
thoroughly studied in the laboratory [6,7]. Fully encapsulated
the pull load exceeds a predefined value. Thus, the bolt does work
rebar bolts are bonded to the grout/rock along their entire length
and absorbs energy from the rock. The cone bolt was initially
with an interlock between the bolt ribs and the grout. Rebar has
designed for use with cement grout. It was later adapted to resin
the advantage of a high load-bearing capacity; however, it
grout [9]. The modified cone bolt (MCB) has a blade at the cone
tolerates small deformations prior to failure, making it a strong
end to serve as a resin mixer. MCBs are used in the dynamic rock
but stiff rock bolt (Fig. 1). Frictional rock bolts interact with the
support systems of many burst-prone deep mines in Canada. Field
rock via friction between the cylindrical surface of the bolt and
observations have shown vulnerabilities in the exterior anchoring
the borehole wall. They can accommodate large rock
point of the cone bolt at the bolt plate. The anchoring may be lost
deformations, but their load-bearing capacity is quite low. For
if the rock fractures behind the plate and falls down [10], causing
instance, a standard Split Set bolt may not be able to bear load
the bolt to completely lose its reinforcement function. It was also
higher than about 50 kN (Fig. 1).
observed that the resin sometimes fails to harden, implying that
An ideal reinforcement device should have the strength of
the mixing quality of resin cannot always be guaranteed for MCB
rebar and the deformation capacity of Split Set bolts (Fig. 1), with
[11].
the ability to be rapidly mobilised to a load level similar to the
Durabar is another energy-absorbing bolt that was invented in
strength of the material. It should be capable of deforming over a
South Africa. Like the cone bolt, it is a two-point anchored device.
long distance while the load remains high. This concept has been
The interior anchor of the bolt is a kinked section of the smooth
recognized previously [8], but it has been difficult to technically
bolt shank which slips at a predefined pull load. The exterior
achieve such devices. One of the trials performed in the early
anchor is the bolt plate.
1990s combined two friction devices, Split Set and Swellex, to
Additional energy-absorbing bolts have appeared in the
enhance their load-bearing capacity [8]. The load capacity of the
market in recent years, for example the Garford bolt, yielding
combined assembly was indeed higher than that of Split Set;
Secura bolt, and Roofex [12,13]. All existing energy-absorbing
however, its anchoring mechanism remained frictional and its
bolts use two-point anchoring in the boreholes.

250 4. D bolt
Ideal bolt – Energy-absorbing bolt
200 4.1. Layout
Pull load (kN)

150 Rebar – Strength bolt

The D bolt is made of a smooth steel bar that has a number of
integrated anchors spaced evenly or unevenly along its length
100
(Fig. 2). In the current design, the anchors are stronger than the
shank of the bolt. This design uses both the strength and the
50 elongation capacity of the shank when the bolt is subjected to
Split set – Ductile bolt
rock dilation. The shank, rather than the anchors, will yield and
0 plastically elongate until failure under extreme loading
0 20 40 60 80 100 120 140 160
conditions. The anchors can have different shapes; two
Displacement (mm) currently adopted shapes are shown in Fig. 2. The anchors are
Fig. 1. Concept of the ideal bolt and definitions of strength, ductile, and energy-
wider than the bar shank so that the bar is automatically
absorbing rock bolts. The curves for the rebar and Split Set are redrawn from centralised in the borehole after installation. This guarantees
Stjern [7]. that the bolt shank is fully capsulated in the grout. The size of the
 ARTICLE IN PRESS
398 C. Chunlin Li / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404

Deformable sections
Short section
Anchor

Paddle anchor

Wiggle anchor

Fig. 2. Layout of the D bolt.

paddle anchor is only slightly larger than the bolt shank, allowing
it to be inserted into small boreholes. The paddle anchors are
suitable for both resin and cement grouts. Resin grouts are mixed
using the D bolt anchor paddles. Fig. 3 shows two types of D bolts
that have undergone field testing, one with paddle anchors and
the other with wiggle anchors.

4.2. Principle

The D bolt is fully encapsulated in the borehole using either

cement grout or resin. The anchors are firmly fixed in the grout;
however, the smooth bar sections between anchors have no or
very weak bonding to the grout. When rock dilates between two
adjacent anchors, the anchors will restrain the dilation so that a
tensile load is induced in the smooth bar between the anchors.
The section elongates elastically at first, but quickly yields after
only a small amount of rock dilation. After that, the bar section
elongates plastically until the ultimate strain limit is reached.
Both the strength and the deformation capacity of the bolt
material come into play in this process.
Two tensile tests were performed in the laboratory to examine
the elongation capacity of the rebar steel material. Two samples
were prepared from a 20 mm rebar made from steel B500C. They
had a dog-bone shape, with a 12 mm diameter and 90 mm length
for the middle stretch section. The results of the two samples
were similar; Fig. 4 shows the results from one of them. Under
tensile loading, the sample elongated elastically until the strain
reached about 0.25%, and then yielded at 450 MPa. It became
hardened afterward, reaching the ultimate tensile strength of the
material (610 MPa) at a strain of about 10%. The sample continued
to elongate at the level of the tensile strength until the initiation
of necking at a strain of about 18%. The sample finally failed at an
ultimate strain of 24%. The area under the stress–strain curve is a
measure of the energy absorbed by the sample. The rebar material
is able to absorb a significant amount of energy if both the
strength and the deformation capacity of the steel material are
fully utilised.
The D bolt is based on the principle of full use of the strength
and deformation capacity of the bolt material along its entire
length. As an example, assume a bolt made of steel B500C is 2.7 m
long, 20 mm in diameter, and has four anchors along its length.
The effective deformable length of the bolt would be approxi-
mately 2.3 m, excluding the total length of the anchors. The bolt
would be able to elongate for 2.3  18% = 0.41 m. Assuming that
the average ultimate load is 180 kN, the static energy absorption
capacity of the bolt would be approximately 74 kJ.
For D bolts, it is vital that the smooth bar be able to detach Fig. 3. D bolts: (a) 2.2 m long paddle-anchor bolts and (b) 2.7 m long wiggle-
from the grout, permitting full use of the bar’s elongation anchor bolts.
 ARTICLE IN PRESS
C. Chunlin Li / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404 399

700 290 – 355 mm

Tensile strength
600
500 Yield strength
Stress (MPa)

Paddle-anchor bar, 20 – 22 mm in diameter

400
∅12 90 150 – 210 mm
300
200 Necking

100 Wiggle-anchor bar, 17.5 – 20 mm in diameter

0 Fig. 5. Bar samples for pull tests.
0 5 10 15 20 25 30
Strain (%)

Fig. 4. Stress–strain behaviour of the rebar steel material B500C under pull load. Table 1
Round bar samples for static pull tests.

Sample Steel Anchor Diameter Length of stretch Coating Grout

capacity. Due to Poisson’s effect, as the axial tensile load
type (mm) section (mm)
increases, the diameter of the smooth bar decreases. This cross
section contraction allows the bar to detach from the grout. To PS1 SS Paddle 22 290 No Resin
further guarantee satisfactory separation, the surface of the PS2 SS Paddle 22 290 Yes Cement
smooth bar can be coated with a lubricant media, such as plastic PS3 SS Paddle 20 355 No Cement
WS1 SS Wiggle 22 150 No Cement
pipes or wax. However, the results presented below show that WS2 B500C Wiggle 17.5 210 No Cement
coatings have a limited effect on the detachment at the bolt–grout
interface.
Every smooth section of the D bolt between two adjacent
anchors acts independently. The failure of one section (or loss of
one anchor) has only a local effect on the bolt’s reinforcement
capability. The other sections (or anchors) still provide reinforce-
ment to the rock. This independency is a significant improvement
compared to the failure characteristics of two-point anchored
energy bolts. For instance, if the bolt plate is lost because of rock
fall behind the plate, the D bolt only loses reinforcement in its
short thread section. The other sections are unaffected by the loss
of the bolt plate. Steel tube
The bolt plate provides the necessary surface retaining
support, and its loss would have a direct negative impact. Field
observations show that bolt plates are quite often heavily loaded
and bolt threads sometimes fail prematurely [2]. To avoid Bar sample
overstressing the plate and thread, the first anchor of the D bolt
is located very close to the bolt thread. This anchor can then carry Split
a large portion of the ground pressure that otherwise would be
transferred to the plate/thread, which significantly improves the Grout
plate/thread’s loading condition.

Anchor
5. Static pull tests

The design calls for the anchors to be stronger than the bolt
shank. Static pull tests were conducted to examine the strength of Adaptor
the anchors of the D bolt. Fig. 5 shows two bar samples used in the
pull tests, one with paddle anchors and the other with wiggle
anchors. Short samples had to be used due to the space limitation
of the test machines. The stretch section length is in the range of
290–355 mm for the paddle-anchor samples, and 150–210 mm
for the wiggle-anchor samples. The diameter of the bar is either
20 or 22 mm for the paddle-anchor samples, and 17.5 or 20 mm
Fig. 6. Arrangement of the pull test.
for the wiggle samples (Table 1).
Each sample is fully encapsulated in a split steel tube, with the
split located in the middle of the stretch section of the sample
(Fig. 6). The steel tube for the paddle-anchor samples is 63/ samples, the steel tube is 80/40 mm in outer/inner diameter and
31.5 mm in outer/inner diameter and 600 mm in length. The resin 680 mm in length. Ordinary Portland cement is used to grout the
cartridges used to grout the paddle anchor samples are 27 mm in wiggle-anchor samples in the tubes. This cement grout has a
diameter and 400 mm in length. Two pieces of resin cartridges are water/cement ratio of 0.33 and a 7-day uniaxial compressive
inserted into the tube hole, then the bar sample is spun into the strength of 68 MPa. The cement mortar is pumped into the tube
hole and the anchor paddles mix the resin. For the wiggle-anchor hole and then the bar sample is simply pushed into the hole. The
 ARTICLE IN PRESS
400 C. Chunlin Li / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404

Fig. 7. Sample PS1: (a) during testing, (b) and (c) after testing.

300
Pull tests of D bolts with paddle anchors
samples (PS2 and PS3) are presented in Fig. 8. These latter two
250
samples are cement grouted. The diameters of PS2 and PS3 are 22
and 20 mm, respectively.
200
Pull load (kN)

The test results of the two wiggle-anchor samples, which are

150
cement encapsulated, are presented in Fig. 9. Sample WS1 and
WS2 have diameters of 20 and 17.5 mm, respectively.
100
Sample PS1, 22 / 290 mm, resin, not coated
Sample PS2, 22 / 290 mm, cement, coated
50
Sample PS3, 20 / 355 mm, cement, not coated 6. Dynamic drop tests

0
0 10 20 30 40 50 60 70 80 Drop tests were performed to evaluate the dynamic energy
Displacement (mm) absorption capacity of the D bolt [14]. For these dynamic tests,
boreholes are simulated by a split steel tube with an inner
Fig. 8. Pull test results of three paddle-anchor bar samples. diameter of 32 mm and a wall thickness of 12 mm. The split tube
is placed in a jig to align it with a drill. Resin cartridges are then
300 slid to the end of the tube. The bolt is inserted in the chuck of the
Pull tests of D bolts with wiggle anchors drill, which is mounted on a sliding rail with an independent
250 advance drive system. The bolt is spun into the tube with a steady
advancement and constant rotation speed.
Pull load (kN)

200 The test arrangement is shown in Fig. 10. For every test, a mass
of 893 kg is dropped from a height of 1.5 m onto a plate connected
150 to the D bolt. To perform the drop, the mass is lifted with an
electromagnet, which in turn is lifted by a pair of cranes mounted
100 in parallel on the top of the machine. The mass freefalls onto the
sample when power is cut to the magnet. The nominal input
Sample WS1, 20 / 150 mm
50 energy of the first drop of every sample is equal to 13.14 kJ.
Sample WS2, 17.5 / 210 mm
The plate and end displacements of the bolt are measured by
0 line scan cameras. The plate displacement is measured at the bolt
0 10 20 30 40 50 60 70 80 plate, while the end displacement is the movement of the bolt
Displacement (mm) end. The impact load and the bolt plate load are measured by load
cells located below the impact plate and the bolt plate,
Fig. 9. Pull test results of two wiggle-anchor bar samples. Both of the samples are
cement encapsulated. respectively.
Dynamic drop tests were conducted on four D bolt samples
each having a diameter of 20 mm. The samples were installed in
32 mm holes with 28 mm resin cartridges. They were spun into
cement-grouted samples are cured for at least 3 days before the the hole at a rotation speed of 300–350 rpm and at an
pull tests are conducted. advancement speed of about 0.1 m/s.
The arrangement of the pull tests is shown in Fig. 6. The The drop test results for the four samples are summarised in
samples were loaded at a speed of 3 mm per minute. The opening Table 2. Fig. 11 shows the typical curves of the impact load and
displacement at the split of the steel tubes was recorded relative the plate load measured during a drop. It is seen that the impact
to the applied load during testing. peak load (in the range of 200–250 kN) is not much different from
Fig. 7 shows the resin-grouted paddle-anchor sample PS1 the impact average load (200–230 kN). Only a small portion of the
during and after testing. The sample is 22 mm in diameter. The impact load is transferred to the plate, with plate loads measured
load–displacement curves of this sample and of other two in the range of 10–110 kN with an average of 54 kN.
 ARTICLE IN PRESS
C. Chunlin Li / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404 401

B) Impact plate set-up

Impact Load cells

Anchor Impact plate

Plate Load cells

Grout Displacement target

Bolt shank

0.8 m
Steel tubes

Anchor

Drop load

Impact
load cell
A) General set-up
Plate load

Fig. 10. Drop test arrangement: (a) bolt, (b) alignment of the test and (c) test rig.

7. Field tests month are presented in Figs. 12 and 13 for the D bolt and rebar,
respectively.
Field tests were performed to examine the in situ performance In pillar J4, only one instrumented D bolt was installed in the
of the bolts and the manoeuvrability of the bolts on existing mine wall at a height of about 1.5 m above the floor. The strain gauges
bolting equipment. on the middle section of the bolt failed 40 days after installation.
Paddle-anchor bolts were installation-tested in two Canadian Fig. 14 presents the measurement results over a 4 months period.
deep and burst-prone mines. The bolts were 22 mm in diameter Among the three deformable sections, the outermost and
and 2.2 or 2.4 m long (Fig. 3a), and were installed on MacLean innermost sections of the D bolt installed in pillar J4 were the
bolters. Two types of resin were used in every bolt hole. A piece of most and least loaded, respectively. After 4 months, the axial load
fast resin cartridge with a setting time of 1 min was inserted in of the bolt was about 110 kN in the outermost section and about
the bottom of the hole, and slow resin cartridges with a setting 75 kN in the innermost section.
time of 2–4 min were used to fill in the rest of the hole. The bolt
was spun into the hole at a rotation speed of 350–500 rpm. The
insertion time was about 10 s for the 2.2 m D bolts. When the bolt 8. Discussion
reached the bottom of the hole, spinning continued for 5 s to mix
the resin at the bolt end. Some of the installed bolts were pulled to 8.1. Static pull tests
examine their load-bearing capacity. In one of the mines, five D
bolts were pulled to 190–200 kN without slippage [15], which Two 22 mm paddle-anchor samples were tested: PS1 was resin
indicated that the resin was well mixed and the paddle anchors encapsulated and PS2 was cement encapsulated. The two samples
were firmly fixed in the resin. were similar in both yield strength (195 kN) and ultimate strength
Fifty samples of wiggle-anchor bolts were field tested in a (250 kN), even though they were installed with different grouts
Swedish deep mine that has serious squeezing rock conditions at (Fig. 8). The stretch sections of these two samples were
depth. The bolts were installed with cement grout in mine stopes approximately 290 mm long. The ultimate elongation was about
and pillars at a depth of about 1100 m below the surface. The bolts 52 mm for PS1 and 58 mm for PS2, corresponding to a ultimate
are 20 mm in diameter, 2.7 m long, and have four wiggle anchors strain of 18% and 20%, respectively. The elongation prior to
(Fig. 3b). Strain-gauge instrumented D bolts and rebar bolts were necking was 42 mm (14%) for PS1 and 48 mm (17%) for PS2, with a
installed to measure the bolt load. The strain gauges were necking elongation of approximately 10 mm for both samples.
attached to the cylindrical surface of the D bolts in the middle Sample SP3 was cement encapsulated, and had a diameter of
of the deformable sections. The instrumented bolts were installed 20 mm and a stretch length of 355 mm. It yielded at 153 kN and
in two mine stopes and one mine pillar. The bolts in one of the failed at 209 kN. Its ultimate elongation was 63 mm, correspond-
stopes were damaged by blasting immediately after installation, ing to an ultimate strain of 18%. Its elongation prior to necking
so only measurements in one stope (J6) and the pillar (J4) were was 53 mm (strain 15%), and its necking elongation was also
available for analysis. approximately 10 mm.
One instrumented D bolt and one instrumented rebar Sample PS2 was coated with a thin shrink pipe, and it had an
were installed 1.5 m apart in the wall of the stope J6. The strain ultimate strain of 20%, slightly larger than the other two samples
gauges on the two innermost sections of the D bolt failed 17 days (18%). Although the increase in the ultimate strain may be due to
after installation. After 1 month, the stope was backfilled and no the coating, the increment of 2% is marginally small so the effect
longer accessible. The measurement results over a period of 1 of the coating may be minimal. The ultimate strains of the
 ARTICLE IN PRESS
402 C. Chunlin Li / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404

Table 2
Drop testing results of D bolts. Drop weight= 893 kg, initial drop height = 1.5 m, and initial input energy = 13.14 kJ [14].

Bolt no. Drop # Plate displ. (m) End displ. (m) Plastic elongat. (m) Impact load (kN) Plate load (kN)

Peak Avg. Peak Avg.

D1 1 0.056 0.003 0.053 214 204 54 40

Coating 2 0.047 0.000 0.047 240 204 75 55
3 Not available 0.000 0.041b 229 215 31 31
Total 0.003 0.141a

D2 1 0.056 0.006 0.050 225 217 60 70

Coating 2 0.049 0.001 0.048 220 221 123 111
3 Not available 0.000 0.045b 205 228 121 108
Total 0.007 0.143a

D3 1 0.096 0.094 0.002 230 179 193 197

No coating 2 0.057 0.000 0.057 252 218 223 213
3 0.048 0.002 0.046 255 221 234 217
4 Not available 0.000 0.030b 233 222 229 220
Total 0.096 0.133a

D5 1 0.056 0.003 0.053 266 217 101 57

No coating 2 0.051 0.002 0.049 246 221 22 10
3 Not available 0.000 0.011b 180 230 5 7
Total 0.005 0.113a

Notes:
a
Measured after testing.
b
Back-calculated by subtracting the stretch increments of the previous drops from the measured total stretch.

250 Peak load Average load 100

Impact load Rebar in Stope J6
200
80
31 days
Axial load (kN)
Load (kN)

150
60 26 days
100
40 17 days
50 Plate load
10 days
20
0
0 10 20 30 40 50
0
Time (msec)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Fig. 11. The impact and plate loads registered during the first drop of sample D1 Distance from wall (m)
[14].
Fig. 13. Loading process of the instrumented rebar bolt in stope J6.

100 D bolt in Stope J6 140 D bolt in pillar J4

80 120 111 days

Axial load (kN)

100
Axial load (kN)

60
31 days 80 63 days
26 days 39 days
40 60
17 days
40 10 days
20 10 days
20
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Distance from wall (m) Distance from wall (m)
Fig. 12. Loading process of the instrumented D bolt in stope J6. Fig. 14. Loading process of the instrumented D bolt installed in pillar J4.

samples were very close to the elongation capacity of the steel Each smooth section of the D bolt is designed to be 0.8–1 m long.
material, which implies that the smooth stretch sections of the Fig. 15 presents the load–displacement curve for a 1 m long smooth
samples were well detached from the grout and able to almost section of the bolt, calculated on the basis of the test result of
freely deform under pull. sample PS3. The curves of the other types of bolts are also included
 ARTICLE IN PRESS
C. Chunlin Li / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404 403

D bolt: Table 3
1 m long deformable section Elongations and strains of the dynamic test samples.

200 Sample D1 D2 D3 D5 Mean

Cement rebar Ultimate elongation (m) 0.141 0.143 0.133 0.113 0.133
Load (kN)

150 Ultimate strain (%) 17.7 18.0 16.7 14.2 16.8

Resin rebar Cumulative energy (kJ) 39.7 40.9 39.18 30.9 37.7
Super Swellex
100
Note: The length of the stretch section is 0.795 m.

50 increased after every drop due to the plastic elongation of the

Split set stretch section. Therefore, the input energy for each subsequent
0 drop increased by the amount of the drop weight times the
0 40 80 120 160 elongation increment of the previous drop. The energy absorbed
Deformation (mm) by the sample was equal to the drop weight times the total drop
height after plastic elongation, for every drop except for the last
Fig. 15. Load–displacement curves of different rock bolts under pull loading. one where the samples failed. In this last drop, the input energy
Notice that the curve for the D bolt is calculated in accordance with the test result
of sample PS3. The curve for Swellex is after Dahle and Larsen [17], and the curves
was only partially used for plastically stretching the material. This
for rebar, expansion shell, and Split Set are after Stillborg [6]. proportion of energy can be estimated using the stretch incre-
ment in the final drop with reference to the increments in the
for comparison. The energy absorbed by a bolt is represented by the previous drops. The cumulative energy absorption of the stretch
area under the load–displacement curve. It is seen that the energy section is the sum of the energies absorbed in all the drops. The
absorption capacity of the 1 m long section of the D bolt is very cumulative energies of the four samples are listed in Table 3. The
large compared to the other types of bolts in the diagram. The total average cumulative energy of the 0.795 m long samples was 38 kJ,
energy absorption capacity of a D bolt is the sum of the energies corresponding to 47 kJ/m of bolt length.
absorbed by all the deformable sections of the bolt.
8.3. Field tests
8.2. Dynamic drop tests
The D bolt installed in stope J6 was mostly loaded in the
Samples D1, D2, and D5 had similar behaviour in the drop outermost deformable section (Fig. 12). The axial load reached
number, impact load, and plate load. Sample D3 behaved differently about 50 kN in the section after 1 month. For the rebar bolt in
in the first drop, in that a ploughing displacement of up to 94 mm stope J6, the axial load was largest at the bolt plate, and it
occurred, accompanying a very small plastic stretch (2 mm). For the attenuated with the distance from the wall surface (Fig. 13). Two
subsequent three drops, D3 behaved similarly to the other three load peaks occurred at depths of about 1.3 and 2.5 m for the rebar
samples. The tests showed more load being transferred to the plate bolt, possibly due to rock fracture openings. It can be inferred that
for sample D3 than for the other samples. After cutting open the the rebar bolt would rupture in the positions of the peak loads
steel tube for sample D3, it was seen that the bolt slid through a when the fractures opened widely enough.
patch of soft resin, settled, and started to stretch. The patch of soft
resin might have been caused by a too fast advance when the
8.4. On the test methods
anchor paddles passed that position. Thus, a steady advance speed
is crucial to ensure that the resin is well mixed.
Both the static and dynamic tests presented here used one-
The D bolt’s impact peak load is approximately the same as its
section samples, even though a D bolt is multi-sectioned.
impact average load, which is slightly larger than the strength of
Although the testing of multi-sectioned samples is desirable, it
the steel material. These two factors characterise the D bolt.
is technically difficult to apply loads to all of the sections of such a
Conversely, with other energy-absorbing bolts that absorb energy
sample, particularly in the case of dynamic testing. Rock bolts are
through a slippage mechanism, the impact load drops rapidly
loaded in situ either statically by time-dependent rock dilation, or
from the peak level to a residual (called average in this paper)
dynamically by rock bursts. In the case of static loading by rock
level as soon as slippage starts. The impact average load of such a
dilation, every section of the D bolt is loaded independently, based
bolt may be only about 50% of its impact peak load [16].
on how the rock dilation is restrained by the anchors of the
The distances between the anchors were measured by cutting
section. Therefore, it is reasonable to use one-section samples to
open the steel tubes after the testing was completed. The total
study the static performance of the D bolt.
elongation of the bolts was obtained by comparing these
In the case of fault-slip rock bursting, the individual sections of
measured distances before and after testing. Plastic elongation
a D bolt would be loaded in a more complicated manner than with
only occurred in the section overriding the split of the steel tube.
static loading. It is not easy to simulate the burst-induced loading
The ultimate plastic elongations and strains of the four samples
process to a rock bolt in laboratory tests. Numerical modelling
are summarised in Table 3. The dynamic ultimate strain of the
may be a more appropriate way to study the loading condition of
bolt samples was 15–20%. The average ultimate strain of the
a multi-sectioned bolt sample in the future. The dynamic tests
uncoated samples (D3 and D5) was 2.4% smaller than that of
presented in this paper are only to demonstrate the energy
the coated samples (D1 and D2). Similar to the static pull tests,
absorption capacity of one section of the D bolt, should that
the effect of a coating on elongation did not appear to be
section be subjected to a dynamic load.
significant under a dynamic loading condition.
The input energy refers to the potential energy of the drop
mass at the beginning of every drop, and is equal to the drop 8.5. On the effect of confinement
weight (893 kg) times the drop height. The height of the first drop
was fixed at 1.5 m for all the samples. Thus, the initial input Mine operations will change the confinement to the bolts that
energy for the first drop was equal to 13.14 kJ. The drop height are already installed in the underground openings. The change in
 ARTICLE IN PRESS
404 C. Chunlin Li / International Journal of Rock Mechanics & Mining Sciences 47 (2010) 396–404

the confinement stress mainly influences the deformation of the Norwegian Research Council, Næringslivets Internasjonaliserings-
smooth shank of the D bolt, but does not affect the strength of the stiftelse, and Boliden Mineral AB, Sweden. The dynamic drop tests
anchors which are firmly fixed in the grout via mechanical were supported by Xstrata Nickel, Canada. The author would like
interlocking. An increase in the confinement stress would increase to thank Per-Ivar Marklund, Graham Swan, Scott Carlisle, Brad
the friction at the bolt–grout interface, which would cause an Simser and Chantale Doucet for constructive discussions in the
uneven distribution of load along the smooth shank, thereby course of the development of the technology. The cooperation of
reducing the shank’s ultimate elongation. Conversely, a decrease David Counter, Xstrata Copper, Canada, and Mike Yao, Vale Inco,
in the confinement stress would be favourable to mobilise the in the course of the field tests is appreciated. Xiangchun Tan is
deformation of the bolt shank. The effect of the confinement thanked for her contribution in designing the tool for anchor
stress on the performance of the D bolt has not been taken into forming. Thanks are also offered to Diyuan Li and Mahdi
account in the laboratory tests presented in this paper. Shabanimaschool for their assistance in some of the pull tests.

9. Conclusion References

The D bolt is a multi-point anchor steel bar installed in a fully [1] Hoek E, Kaiser PK, Bawden WF. Support of underground excavation in hard
grouted borehole. The bolt absorbs energy through full use of the rock. Rotterdam: Balkema; 1995.
strength and the deformation capacity of the bolt material (steel) [2] Li CC. A practical problem with threaded rebar bolts in reinforcing largely
deformed rock masses. Rock Mech Rock Eng 2007;40(5):519–24.
along its entire length. The bolt is equally loaded in every [3] Li CC. Field observations of rock bolts in high stress rock masses. Rock Mech
deformable section. An opening of an individual rock fracture will Rock Eng 2009; online.
load the entire length of the bolt section that overrides the [4] Ortlepp WD. The design of support for the containment of rockburst damage
in tunnels—an engineering approach. In: Kaiser PK, McCreath DR, editors.
fracture regardless of where the fractures are located. Each Rock support in mining and underground construction. Rotterdam: Balkema;
deformable section works independently to restrain rock dilation. 1992. p. 593–609.
Loss of one section does not affect the functionality of the others. [5] Jager AJ. Two new support units for the control of rockburst damage
In: Kaiser PK, McCreath DR, editors. Rock support in mining and underground
Both static pull tests and dynamic drop tests have shown that construction. Rotterdam: Balkema; 1992. p. 621–31.
the maximum deformation of a smooth steel bar encapsulated in [6] Stillborg B. Professional users handbook for rock bolting, 2nd ed. Trans Tech
resin or cement grout is close to the ultimate elongation of the bolt Pub; 1994.
[7] Stjern G. Practical performance of rock bolts. PhD thesis, Universitetet i
material. This indicates that the smooth bar almost completely
Trondheim, Norway, 1995. p. 52.
detaches from the grout. Coating the cylindrical surface of the bar [8] Windsor CR, Thompson AG. A new friction stabilizer assembly for rock and
may slightly enhance the elongation of the bolt; the effect of the soil reinforcement applications. In: Kaiser PK, McCreath DR, editors. Rock
support in mining and underground construction. Rotterdam: Balkema;
coating needs to be verified by more tests in the future.
1992. p. 523–9.
For dynamic loading, the impact peak load of the bolt is very [9] Simser B. Modified conebolt static and dynamic tests. Noranta Technology
similar to the average load, with both slightly above the level of Centre Internal Report, Quebec, Canada, 2002.
the tensile strength of the bolt material. For the tested 20 mm [10] Li CC, Marklund P-I. Field tests of the cone bolt in the Boliden mines.
In: Nilsen B, Hamre L, Rohde JKG, Berg KR, editors. Fjellsprengninsteknikk/
bolts, the impact peak load is in the range of 200–250 N, and the Bergmekanikk/ Geoteknikk. Oslo: Norsk Jord og Fjellteknisk Forbund; 2004.
impact average load is in the range of 200–230 kN. The Article 35: 12 pp.
cumulative energy absorbed by the 20 mm bolt is 47 kJ/m of [11] Simser B, Andrieux P, Mercier-Langevin F, Parrott T, Turcotte P. Field
behaviour and failure modes of modified conebolts at the Craig, LaRonde
stretch length. Locating the anchor close to the thread provides and Brunswick mines in Canada. In: Potvin Y, Hadjigeorgiou J, Stacey D,
protection to the thread/plate, so that only about 25% of the editors. Challenges in deep and high stress mining. Perth: Australian Centre
impact load is transferred to the thread/plate. for Geomechanics; 2007. p. 347–54.
[12] Varden R, Lachenicht R, Player J, Thompson A, Villaescusa E. Development
Field trials of D bolts were conducted in three deep mines. The and implementation of the Garford dynamic bolt at the Kanowna Belle Mine.
in situ measurements show that the rebar bolt is unevenly loaded In: 10th underground operators’ conference, Launceston, Australia, 14–16
along its length, with the maximum load at the bolt plate/thread. April 2008. 19 pp.
[13] Charette F, Plouffe M. Roofex—results of laboratory testing of a new concept
Furthermore, peak loads occur at the opening fractures. The of yieldable tendon. In: Potvin Y, editor. Deep mining ’07. Perth: Australian
D bolts are equally loaded in each deformable section. D bolts Centre for Geomechanics. p. 395–404.
possibly can survive for a longer time than rebar bolts, since it is [14] Li CC, Doucet C, Carlisle S. Dynamic tests of a new type of energy absorbing
rock bolt—the D bolt. In: Diederichs M, Grasselli G, editors. 3rd Canada-US
possible to avoid premature failure due to peak loads in D bolts.
rock mechanics symposium. Toronto: Canadian Institute of Mining; 2009.
9 pp.
[15] Cai M, Champaigne D. D bolt pull test at Vale Inco’s North mine. New York:
Acknowledgements Mansour Mining Inc; 2009. 5 pp.
[16] St-Pierre L. Development and validation of a dynamic model for a cone bolt
anchoring system. MS thesis, McGill University, Montreal, 2007.
It is acknowledged that the static pull tests and the field tests [17] Dahle H, Larsen T. Full-scale pull and shear tests of 5 types of rock bolts.
were financially supported by Næringslivets idéfond for NTNU, Technical report, SINTEF, Trondheim, 2006. 27 pp.