You are on page 1of 18

Tunnelling and Underground Space Technology 93 (2019) 103098

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology


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

Controlled blasting in underground construction: A case study of a tunnel T


plug demolition in the Neelum Jhelum hydroelectric project
Abubakar Sharafata,c, Waqas Arshad Tanolia,d, Georgios Raptisb,c, Jong Won Seoa,

a
Dept. of Civil & Environmental Engineering, Hanyang University, Seoul 04763, Republic of Korea
b
Independent Geological Engineering Services, Akyaka Mah., Kadirler Sok. No: 7/2, 48640 Ula, Mugla, Turkey
c
Former Employee of Neelum Jhelum Consultants, Malik House, Lower Chatter Rd., Muzaffarabad, AJK, Pakistan
d
Communication &Works Department, Khyber Pakhtunkhwa, Pakistan

ARTICLE INFO ABSTRACT

Keywords: Blasting is a popular method of excavation in underground construction. Blast-induced vibrations, however, if
Controlled blasting not properly controlled, can negatively affect the rock mass surrounding the tunnel. During the drill-and-blast (D
Critical particle velocity &B) excavation of twin headrace tunnels through a major fault zone below the Jhelum River, a 180-m3/h
Vibration attenuation model groundwater inflow occurred from the tunnel invert. A 610-m3 concrete plug and high-pressure grouting around
Permissible instantaneous charge weight
the tunnel periphery were formed to stabilize the tunnel and control inflow. To complete the tunnel, the tem-
Blasting damage
Tunneling through major fault
porary tunnel plug was ultimately demolished using controlled blasting. Ground vibration monitoring data from
the regular D&B operations were used to develop a site-specific vibration attenuation model. Using geotechnical
data from the feasibility stage and tunnel mappings during construction, a critical particle velocity above which
tensile cracking could be initiated was determined for the rock mass surrounding the plug. The critical particle
velocity was incorporated into the vibration attenuation model to obtain permissible charge weights to limit
blast damage to within a short distance from the charge. Using this approach, an innovative blasting pattern was
adopted in which the tunnel plug was sequentially demolished by detonating a burn cut and separate contour
blast rows. During blasting, rigorous supervision and data collection helped to record, evaluate, and when
needed, adjust the blasting program. Overall, the tunnel plug demolition was completed in a safe and cost-
effective manner.

1. Introduction by the explosive detonation and propagate through the medium as


ground vibrations. When vibrations (and stresses) exceed a certain
Blasting is used in tunnel construction as a common and cost-ef- limit, they can damage the rock mass or structures (Dey and Murthy,
fective excavation method. Moreover, blasting can be used to demolish 2012). Therefore, predicting ground vibrations is important for pre-
temporary or out-of-service structures. However, when uncontrolled, venting excess damage.
blasting can lead to negative effects such as excessive flyrock or over- Contour blasting techniques have been widely used since the mid-
break, rock mass loosening, damaged rock support, and tunnel failures. twentieth century (Oriard, 1982a, Olofsson, 1990) to control overbreak.
The likelihood of such potential adverse effects can be reduced if con- These techniques involve small-diameter and decoupled charges in a
trolled blasting techniques are properly implemented. closely spaced and lightly weighted blasting pattern. Consequently,
Several predictive blast damage models rely on parameters such as contour blasting produces less intense ground vibrations and sub-
the borehole radius; relative weight strength, detonation velocity, and stantially less blast damage (Van Eeckhout, 1987). Among several
gas internal energy of explosives; loading density; coupling factor; and contour blasting techniques, smooth-wall blasting has shown un-
rock specific charge (Bastante et al., 2012). However, a single equation disputed success in producing very stable rock surfaces in tunnels and
that describes ground vibration attenuation as a predictive damage large caverns.
model has acquired significant popularity because of ease of application Line drilling is another contour blasting method in which holes are
in the field (Nateghi et al., 2009, Yilmaz and Unlu, 2014, Trigueros drilled and left empty in the perimeter of the excavation. Therefore, in
et al., 2017). Blast damage is induced by stress waves that are generated line drilling, “pre-treatment” of the excavation is attempted with empty


Corresponding author at: Jaesung Civil Eng. Bldg. 510, Hanyang University, Hangdang1-Dong, Seongdong-Gu, Seoul, Republic of Korea.
E-mail address: jseo@hanyang.ac.kr (J.W. Seo).

https://doi.org/10.1016/j.tust.2019.103098
Received 28 August 2018; Received in revised form 10 June 2019; Accepted 20 August 2019
Available online 06 September 2019
0886-7798/ © 2019 Elsevier Ltd. All rights reserved.
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Nomenclature Q rock mass quality rating value (Barton, 2002)


R distance of the blasting source to the point of particle
List of abbreviations and symbols velocity measurement (m)
R/(Wmax)n = SD scaled distance
B burden, distance from an explosive charge in a blast hole R/√Wmax square-root scaled distance (m/√kg)
to the nearest free or open face (m) RRS reinforced ribs of sprayed concrete (shotcrete)
BBF Balakot–Bagh fault S spacing, distance between blast holes in a row (m)
CP longitudinal wave (or P-wave) propagation velocity (km/ STA tunnel station, location
s) V particle velocity at any point in a transmitting medium
D&B drill-and-blast tunneling method during propagation of wave (mm/s)
E Young’s modulus of the transmitting medium (GPa) Vcr the critical value of particle velocity above which tensile
f frequency of ground vibration at PPV (Hz) failure is assumed to occur (mm/s)
FRS fiber-reinforced sprayed concrete (shotcrete) W instantaneous explosive charge weight (kg)
JRCF Jhelum River Crossing fault Wact actual instantaneous charge weight applied during the
K ground transmission factor, a site-specific coefficient in blast design implementation in the field (kg)
empirical ground attenuation laws. Wmax maximum instantaneous charge weight in a blast round
L length of blast hole (m) (kg)
LHTT left branch of headrace twin tunnel Wperm permissible instantaneous charge weight in blasting that
M longitudinal (or P-wave) modulus of the transmitting produces only minor damage in the form of fine (up to a
medium (MPa) couple of millimeters wide) tensile cracking and loosening
n a site-specific coefficient in empirical ground vibration (small inelastic displacement) of unhealed joints in the
attenuation laws. It is assumed to be associated with the rock mass (kg)
symmetry of the explosives (Oriard, 1982a) β a site-specific constant in empirical attenuation laws,
N number of blast holes in a blast which represents the effect of attenuation with distance
NJHEP Neelum Jhelum Hydroelectric Project εd dynamic strain
PF powder factor expressed in terms of charge weight per ν Poisson’s ratio of the transmitting medium
cubic meter of blasting material (kg/m3) ρ density of the transmitting medium (kg/m3)
PPV maximum peak particle velocity, the highest between the σd dynamic stress (MPa)
peak particle velocities of the three components of motion σc uniaxial compression strength of intact rock (MPa)
in a blasting event (mm/s) σtd dynamic tensile strength of the transmitting medium
PPVL, T, or V peak particle velocity of the longitudinal, transverse, (MPa)
or vertical component of motion, respectively (mm/s)

holes to dissipate the explosive gases and attenuate vibrations. Line between particle velocity and blasting damage, and the use of the
drilling is a costly method because it requires extensive and accurate powder factor (PF) in blasting designs are reviewed. Geological and
drilling along the excavation perimeter, and as such, is not used ex- geotechnical details of the NJHEP are subsequently discussed including
tensively. Nonetheless, it is an attractive solution to reduce the poten- details of the groundwater inflow remediation measures. Next, we de-
tial for even minor unwanted damage. scribe how permissible instantaneous charge weights based on a site-
In this study, we describe the demolition of an unreinforced mass specific ground vibration attenuation model and a critical particle ve-
concrete tunnel plug using controlled blasting. The temporary plug was locity value are developed to mitigate blast damage to the rock mass
erected to stabilize, and perform high pressure remedial grouting in, a during plug demolition. An innovative blasting pattern is described
section of the headrace tunnel of the Neelum Jhelum Hydroelectric based on contour blasting concepts. Finally, implementation, inspec-
Project (NJHEP) following a major groundwater inflow while crossing a tion, vibration data collection, and tunnel and rock mass evaluation to
fault zone below the Jhelum River. To develop site-specific metho- validate the approach are described.
dology to demolish the plug using controlled blasting, a theoretical
background on ground vibrations, attenuation models, the relationship

Fig. 1. Example of velocity time histories recorded during a blasting event. PPVL, PPVT, and PPVV denote the peak particle velocities of longitudinal, transverse, and
vertical component of motion, respectively. PPV is the maximum value between the PPVL, PPVT, and PPVV.

2
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

2. Theoretical background (Charlie, 1985). The two-thirds SD has been recommended by the
Bureau of Indian Standards (1973). According to Tripathy and Gupta
2.1. Ground vibration and attenuation models (2002), however, field data suggest that two-thirds scaling is not sui-
table for accurately predicting ground vibrations both at long and short
When an explosive charge is detonated in a blast hole, the released distances. In this case study, the square-root SD is used because the
energy generates a shock wave that radiates away from the blast hole in explosives were arranged in row and line charges.
all directions. The energy, in the form of body and surface waves,
causes ground particles to oscillate in three directions (Stiehr, 2011). 2.2. Blasting damage and PPV
Assuming blast holes are drilled in a horizontal plane perpendicular to
the tunnel face, particle motions can be resolved into three mutually Singh and Lamond (1993) refer to the blast damage in a rock mass
orthogonal components—a longitudinal component where movement is as “…the visible alteration to the appearance of the rock structure in
horizontal along the axis of the hole, a transverse component also in the the form of cracking, slabbing, and overbreak…” adding, however, that
horizontal plane, and a vertical component. “…significant reduction in the strength of the rock mass can take place
Vibrations are usually recorded and reported as particle velocities well before the appearance of these obvious signs of damage.”
over time for each component of motion in plots called velocity time Apart from visual inspection surveys and methods such as the half
histories (Fig. 1). The absolute maximum zero-to-peak value of each cast factor, various investigative methods exist to identify blast damage.
component (longitudinal, transverse, and vertical) is the peak particle Borehole core logging, borehole imaging, ground penetrating radar,
velocity for the particular component of motion, called PPVL, PPVT, and cross-hole seismic surveying, permeability testing, acoustic emission
PPVV, respectively. Each component has a peak particle velocity that and ultrasonic surveying, electric resistivity and seismic refraction
generally occurs at a different time, and the maximum value of all three surveying, as well as borehole jacking and plate load testing have been
components is called the maximum peak particle velocity (PPV). used to study blast damage (Van Eeckhout, 1987, Singh and Lamond,
The amplitude, frequency, and duration of ground vibrations at a 1993, Raina et al., 2000).
given location depend on the energy of the blast (expressed by the Over the years, many researchers have measured PPV values using
blasting design and the explosive parameters), the pathways of ground observations of damage in the field to estimate the type and onset of
motion through the rock mass, and local site characteristics (Stiehr, blast damage caused by vibrations. Table 1 summarizes the studies on a
2011). The two most important factors influencing ground vibration range of rock types and conditions. Despite the differences in testing
attenuation are explosive charge weight and distance from the charge. conditions, including different blasting techniques and charge con-
In general, particle velocity increases as the charge weight increases finements, and different approaches for classifying damage, some
and distance from charge decreases. general trends of damage and PPV limits can be deduced from Table 1.
A predictive model of the PPV can be developed with relative ease No fracturing of intact rock or visible damage is observed for PPV
in the field using the charge weight and distance from the charge by values less than approximately 250 mm/s. When PPV values are in the
statistically analyzing recorded blast data. Many empirical vibration range of 250–600 mm/s, minor damage occurs in the form of fine
attenuation laws have been developed, and they generally follow the tensile cracks (i.e. as much as a couple of millimeters in aperture),
form indicated in Eq. (1) (Tripathy and Gupta, 2002): loosening of unhealed joints, or falling of rock from unlined tunnels.
Moderate blast damage is manifested in the PPV range of
PPV = K·(R (Wmax ) n) (1)
600–1000 mm/s when wider tensile cracks, possibly along weakness
where R is the distance from the blast source to the point of measure- planes, are formed or existing unhealed joints are enlarged and further
ment, Wmax is the maximum instantaneous charge weight in the blast, weakened. We define major blast damage as damage occurring in the
and K, n, and β are site-specific constants. PPV range of 1000–2000 mm/s in which intense tensile and some radial
Eq. (1) indicates a negative exponential relation that represents a fracturing occur, affecting healed joints with hard infills as well.
linear relationship in a log–log coordinate system with R/(Wmax)n on Finally, PPV values greater than 2000 mm/s may fracture intact rock,
the x-axis and PPV on the y-axis. result in complete disintegration of rock masses, or cause overbreak in
massive rock masses.
log(PPV) = log(K) ·log(R (Wmax ) n) (2)
Based on rock properties, the onset of blasting damage can be ap-
The constant K is the ground transmission factor (Olofsson, 1990) proximated by defining a PPV value above which fracture is initiated.
and depends not only on the confinement of the charge but also on the This, however, requires the assumption that the shock wave is trans-
immediate geology and the presence of faults and fractures in the mitted as a longitudinal wave (P-wave) in a simplified one-dimensional
ground. Higher values of K indicate either well-coupled explosives or a homogeneous, isotropic, elastic medium similar to a thin elastic bar
relatively homogeneous and massive rock mass with few fractures. The (Brady and Brown, 1993).
constant β represents the effect of attenuation with distance and de- The P-wave generates a dynamic longitudinal stress in the rock mass
pends on rock mass quality. Higher values of β indicate poorer quality along the direction of propagation, which induces dynamic uniaxial
rock mass that attenuates vibration energy more quickly and within a strain at all points along that line of motion and about the equilibrium
shorter distance from the charge, whereas lower values of β indicate position of the point. The particle velocity, V, at any point relates to the
more competent rock mass with less fracturing that transmits the vi- dynamic uniaxial strain, εd, as follows:
bration energy for a longer time and over a greater distance from the
V= d·CP (3)
charge (Yilmaz and Unlu, 2014).
The ratio R/(Wmax)n is called the scaled distance (SD) and is nor- where CP is the P-wave propagation velocity and is defined by Eq. (4)
malized to compare blasts of different intensity at different distances.
CP = (M ) (4)
Experience has shown that the constant n is associated with the sym-
metry of the explosives (Oriard, 1982a). The most commonly used where ρ is the density of the elastic medium and M is the elastic
values are 0.5, 0.33, and 0.67 (Tripathy and Gupta, 2002), and R/ longitudinal (or P-wave) modulus, which defines the stiffness of the
(Wmax)n is referred to as square- root, cube-root, or two-thirds SD, de- medium in a uniaxial strain state (Mavko et al., 2003). The P-wave
pending on the respective selected exponent. modulus can be further related to Young’s modulus (E) and Poisson’s
The square-root SD typically matches results from row charges, line ratio (ν) as in Eq. (5).
charges, or near-surface charges that generate surface waves. The cube-
M= E·(1 ) ((1 + )·(1 2 )) (5)
root SD mostly matches results from deeply buried point charges

3
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Table 1 Table 2
Suggested PPV levels for blasting damage in hard rocks and rock masses. Typical powder factor values for bench blasting (Stiehr, 2011).
Reference, supplementary information, and PPV limit (mm/s) Breakage difficulty (Blastability) Powder factor (kg/m3)
blasting effect
Very high 0.57–1.14
Bauer and Calder (1970) High 0.34–0.57
No fracturing of intact rock < 254 Medium 0.18–0.34
Minor tensile slabbing 254–635 Low 0.10–0.18
Strong tensile and some radial cracking 635–2540
Complete breakup of rock mass > 2540

Langefors and Kihlström (1973) per blast hole, B is the distance from an explosive charge in a blast hole
Fall of rock in unlined tunnels > 305 to the nearest free or open face (burden), S is the distance between blast
Formation of new cracks > 610 holes in a row (spacing), and L is the length of blast holes.
Tunstall et al. (1977); observations from typical smooth blasting Because most commercially used explosives have somewhat similar
Minor damage: loosening of unhealed joints > 500 energy values and thus similar rock breaking capabilities, PF is mostly a
Significant damage: opening and extension of > 900
function of the mechanical properties of the blasted rock (e.g., density
calcite filled joints
Major damage: fracturing of intact rock > 2000 and compressive and tensile strength) and the geological structure (e.g.,
frequency, distribution, and condition of joints). Weak, low density
Holmberg (1982); observations in granite
Rock mass damage: cracks are induced or 700–1000 rock requires less explosive per cubic meter than strong, dense rock.
enlarged Similarly, massive rocks with few joints or weakness planes require
Oriard (1982b)
greater PF than a rock mass that has numerous closely spaced joints.
Rock mass damage > 635 Table 2 presents typical PF values for bench blasting. Shale is an
Rustan et al. (1985); observations from contour blasting using low-detonation-velocity
example of a typical intact rock with medium PF. Rock types such as
charges in hard intensely jointed rock masses sandstone and limestone generally require high to very high PF values
Rock mass damage 1000–3000 for efficient rock fragmentation. Massive granites require very high PF.
Yu and Croxal (1985); observations in meta-volcanic rocks at Kidd base metal mine in Plain concrete also requires high to very high PF values for suitable
Canada breakage. In this case study, mostly plain concrete was blasted.
No visible damage < 250 The more number of free faces a blast has, the lower is the PF value.
Minor slabbing failure 250–500 In tunneling, however, only the rock face is free, which means the rock
Possible formation of cracks along weakness 500–1000
planes
is more constricted. Additionally, greater PF values are caused by
Moderate slabbing failure 1000–1200 greater blast hole deviation, the need to overcharge floor holes to en-
Major slabbing failure 1200–1800 sure swelling (i.e. to compensate for less space available for fragmented
Meyer and Dunn (1996); observations in metamorphosed volcaniclastic rocks and lavas rock to move because of its increased blasted volume), and the lack of
at Perseverance nickel mine in Australia greater number of instantaneous detonations between adjacent holes.
Minor damage > 300 Thus in tunnel blasting, PF is 3–10 times greater than that used for
Rock mass damage > 600
bench blasting (Olofsson, 1990). Generally, explosive consumption is
Bogdanoff (1996); near-field (≤1 m) measurements using conventional smooth greater in smaller tunnel sections than that for larger tunnel sections.
blasting explosives in perimeter and perimeter helper holes in hard rock conditions Similarly, the explosive charge is greater in the blast cut area and de-
Rock damage 2000–3000
creases toward the tunnel periphery.
Murthy and Dey (2003); observations in compact basalt
Overbreak > 2050

McKenzie and Holley (2004); observations from production blasting and (not more than 3. Case study project overview
4 rows at a time) trim blasting in large open-pit mine operations
Fine cracking > 300
Open cracking > 350
The NJHEP is a recently constructed hydroelectric project in
Significant damage > 400 northeastern Pakistan and the Muzaffarabad district of Kashmir. The
Intense damage > 700 project generates 969 MW of electric power by diverting water from the
Neelum River into the lower branch of the Jhelum River taking ad-
vantage of a gross hydraulic head of approximately 430 m.
Using Hooke’s law for the dynamic uniaxial stresses and strains The main features of the project (Fig. 2) include a diversion dam
(σd = M·εd) in combination with Eqs. (3) and (4) and assuming brittle and intake system; a 28.6-km-long headrace tunnel designed to carry
failure for the elastic medium, a critical particle velocity (Vcr), which if 283 m3/s of water; an underground powerhouse complex with a
exceeded will result in tensile failure, is defined by Eq. (6) 25 × 57 × 137 m3 (span × height × length) cavern, a surge chamber
Vcr = ( td·CP ) M = td ( ·CP) (6) and tunnel system, and four 100-m-long penstock tunnels; and a 3.6-
km-long tailrace tunnel. All these features required the setup of three
where σtd is the dynamic tensile strength of the medium. different construction lots (Fig. 3).
The headrace tunnel comprises single (31%) and twin (69%)
2.3. Powder factor horseshoe-shaped tunnel sections, whereas the tailrace tunnel is an
entire single horseshoe section. Tunnel spans range from 10.7 to 11.8 m
PF or specific charge provides a measure of the breakage difficulty for single tunnel sections and from 7.75 to 8.55 m for twin tunnel
and the anticipated fragmentation of the blasted material. Using PF as a sections. The tunnels were excavated using D&B methods, with the
basis for blasting calculations, we can calculate the necessary explosive exception of 10.5 km of twin tunnel sections mined using two main-
charge for a blasting operation (Olofsson, 1990). beam-gripper tunnel boring machines driving upstream.
PF is usually expressed as the weight of explosives needed to break Depending on rock mass quality, the typical tunnel support in D&B
1 m3 of rock (Olofsson, 1990) and can be calculated using Eq. (7): tunnels comprises an initial lining of 40- to 150-mm-thick fiber-re-
inforced shotcrete (FRS) with 4- to 5-m-long, radially installed, fully
PF = (N·W) (N·B·S·L) (7)
grouted rock bolts, followed by an additional 300- to 1350-mm-thick
where N is the number of detonated blast holes, W is the charge weight reinforced concrete lining. Below the Jhelum River, in an 800-m-long

4
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Fig. 2. 3D representation of NJHEP main features in the dam and powerhouse complex.

twin tunnel section called the Jhelum crossing, the final lining, which fault gouge of various proportions and predominantly dry. The Q-value
followed the initial support, consists of a 30-mm-thick, 6.6 m in nom- of the rock mass within the BBF zone ranged from 1.0 to 0.1, which
inal diameter, steel lining to provide hydraulic confinement due to in- describes very poor ground conditions.
sufficient overburden height (180 to 230 m) and insufficient confining The second fault zone, which was previously unknown, was en-
stress compared to the gross internal water pressure head of 430 m. countered when the tunnel headings reached approximately 150 m
Located at the foothills of the northwestern Himalayas, the geolo- from the northern bank of the Jhelum River and was therefore named
gical setting of the NJHEP is characterized by intense tectonic de- the Jhelum River Crossing Fault (JRCF) (Fig. 4a). The JRCF zone was
formation and the presence of extensive regional faults (Fig. 3). The not identified during site investigations because the surface trace is
entire tunnel system is located in the Murree formation, which is a concealed by fluvial deposits and young river terraces. In addition, site
sequence of continental clastic sedimentary rocks of the early Miocene investigations along the Jhelum River were limited because of access
age (Hussain et al., 2009). restrictions imposed by local landowners. Within the twin tunnels, the
The Murree formation appears mainly as alternating beds of sand- JRCF is a 50-m-wide zone, encompassing all three rock types of the
stone, siltstone, and mudstone. Table 3 presents intact rock properties Murree formation and oriented 071°/51° (dip direction/dip) with an
of the Murree formation. Numerical values were determined by la- apparent dip of 31° relative to the headrace tunnel alignment, which is
boratory tests undertaken during the feasibility and design stages of oriented N042°–222°.
NJHEP. The BBF and JRCF zones have a geotechnical nature of a block-in-
In the Lot C2 D&B headrace tunnels, the Murree formation exibits a matrix rock or bimrock, that is, a chaotic heterogeneous geological
high degree of compression in the form of tight folding often with mixture of blocks of different types and sizes surrounded by weaker,
steeply inclined to upright (80° to 90°) or even overturned rock strata often sheared, fine-grained matrix (Medley, 1994). Tunneling through
approximately normal to the tunnel alignment. Minor faulting, which these major fault zones resulted in several complications including
was easy to handle, was common, particularly along the bedding con- occurence of mixed-ground conditions with strong rock mass quality
tacts between competent sandstone and weaker siltstone or mudstone variation (even in a single excavation round); frequent stress redis-
layers. However, two major fault zones posed significant challenges tribution (particularly in sharp transitions from soft to stiff rock mass or
during the excavation of the tunnels near and at the Jhelum crossing. vice versa); several different failure mechanisms including excessive
The larger fault zone is located at the tunnel intersection with the overbreak that was frequently of a chimney type (Marinos, 2012),
Balakot–Bagh fault (BBF, see Fig. 3), also called the Muzaffarabad collapse of the face and crown, squeezing ground and long term creep;
Thrust or Tanda fault (Kaneda et al., 2008, Hussain et al., 2009, Turab, and heterogeneous distribution of groundwater.
2012). The BBF was responsible for the catastrophic 2005 Mw 7.6 Although the BBF zone is practically dry, the JRCF is located di-
Kashmir earthquake, which has a recurrence interval of approximately rectly below and hydraulically connected to the permeable fluvial de-
625 ± 125 years (Jadoon et al., 2015). posits. Hence, the JRCF directly linked the river with the tunnel ex-
During the design stage of the NJHEP, the surface trace of the BBF cavation resulting in an abundance of groundwater under a hydrostatic
was identified by pronounced fault scarps as well as tilted and dissected pressure of approximately 1.8 to 2.3 MPa (Kizilbash et al., 2017). Be-
old river terraces approximately 1.4 km away from the northern bank of cause of this unlimited supply of groundwater under pressure, the rock
the Jhelum River. Inside the headrace twin tunnels, BBF is a 190-m- mass Q-value of the JRCF zone ranged from 0.06 to 0.001, demon-
wide zone of moderately to intensely fractured rock, fault breccia, and strating extremely poor to exceptionally poor ground conditions.

5
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Fig. 3. NJHEP layout and regional (NW Himalayan) faults (modified after Turab, 2012).

Table 4 summarizes the typical rock mass characteristics en- of the excavation (Fig. 5). Waxed paper cartridges, 25 mm in diameter
countered in the Lot C2 D&B tunnel excavations and the JRCF zone. The and 200 mm in length, consisting of WABOX, a gelatinous 80% ni-
rock mass parameters were assessed from the geological tunnel map- troglycerin-based dynamite (Wah Noble, 2015), was the typical blasting
pings using the suggested methods for the quantitative description of agent used in 42-mm-diameter drill holes. Non-electric (shock tube)
discontinuities in rock masses (ISRM, 1978) and the Q-system (Barton, millisecond and half-second delay detonators, detonation cord, safety
2002). The P-wave velocity was measured during the feasibility stage of fuses, and plain detonators No. 8 were used to initiate the blast.
the project using seismic reflection and refraction methods near the During D&B tunneling, ground vibrations were monitored as part of
Jhelum river crossing. The seismic velocity for each rock type is at the quality control procedures. Vibration measurements were collected
depth of the headrace tunnel. The velocity applied for the JRCF zone is from a total of 37 blasting operations using a data logger with three
based on measurements of decomposed rock mass directly beneath the geophones measuring velocity, frequency, and acceleration in the
Jhelum river bed. longitudinal, transverse, and vertical direction, respectively. To avoid
Each blasting round in the Lot C2 tunnels was typically 3 to 4 m and decoupling, the geophones were fixed with cement on the tunnel invert
involved a wedge-cut pattern and smooth-wall blasting at the perimeter according to international specifications (ISEE, 2015). In addition,

Table 3
Intact rock parameters of the main rock types.
Parameter Sandstone Siltstone Mudstone

Color Gray Brown–Reddish brown Reddish brown


Grain size Fine–Medium Very fine–Medium Very fine
Bedding Thick–Massive Thin Very thin–Thin
Bulk density, kg/m3 2730 2771 2722
Uniaxial compressive strength, MPa (Mean and data range) 86.0 (20.0–114.6) 56.5 (20.0–86.5) 33.0 (12.0–54.1)
Brazilian tensile strength, MPa (Mean and data range) 8.1 (3.9–11.0) 6.8 (4.7–6.8) 5.3 (5.2–8.2)
Young’s modulus, GPa 32.0 21.0 12.6
Poisson’s ratio 0.27 0.22 0.26

6
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Fig. 4. (a) Cross section of the Jhelum River valley (looking upstream) at the location of the headrace tunnels. (b) Close up of LHTT section at the location of the
major groundwater inflow.

instruments were protected from flyrock using concrete blocks. The than 0.1 MPa, a pre-excavation grouting fan was constructed around
vibration data are tabulated in the Appendix Table A1. For every re- the tunnel perimeter and in some cases at the tunnel face. The grouting
corded blast, the data include the maximum instantaneous charge aimed to reduce inflows in post-grouting check holes to 2 to 3 L/min
weight, the distance from the blast to the geophone, the recorded peak (0.1–0.2 m3/h) per linear meter of hole and the hydrostatic pressure to
particle velocities of each ground motion component, and the PPV and less than 0.1 MPa (Kizilbash et al., 2017). The pre-excavation grouting
its respective frequency. fans typically extended to 15 m with a look-out angle of 5° to 10°. Type
II ordinary Portland cement was mixed in a water-cement ratio of 0.5
4. Tunneling in the Jhelum crossing (by weight) and injected at pressures up to 10 MPa. Successive fans
overlapped 3 to 5 m.
The headrace tunnel was excavated by advancing multiple headings The JRCF zone was encountered by the headings advancing up-
simultaneously in the upstream and downstream directions from mul- stream after the downstream tunnel headings had already been con-
tiple access adits. The excavations in the Jhelum River Crossing were nected to the powerhouse. Tunneling through the unanticipated JRCF
undertaken from opposite sides of the river in both tunnels and four meant the excavations proceeded through extremely to exceptionally
headings working toward each other were employed. poor ground conditions subjected to significant groundwater inflows
Probing and pre-excavation grouting was implemented under the and pressures and an unlimited source of water. The JRCF posed serious
Jhelum River as a general precaution against groundwater inflows. challenges and risks, ranging from potential tunnel instabilities to
Probing typically involved drilling three or more 15-m-long holes at the flooding the tunnel system downstream. One of the most significant
tunnel face. Probe holes were drilled ahead of the excavation to assess risks was that excessive seepage from JRCF could overwhelm the
the geology and degree of fracturing and to identify zones of potential pumping system and flood the powerhouse where electro-mechanical
high groundwater inflow. When the cumulative inflow from the probe work was being performed.
holes exceeded 50 L/min (3 m3/h) and a hydrostatic pressure greater To avoid flooding, large amounts of seepage water had to be

7
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

efficiently pumped from the excavation out to the surface, which in-
volved a vertical lift of up to 230 m. The tunnel dewatering system had

Zones of disintegrated or crushed rock with various proportions of angular rock

decay under ~2.0 MPa hydrostatic pressure with considerable washout of joint
Large (180 m3/h) to exceptionally high (720 m3/h) inflows without noticeable
to be upgraded with multiple pumping stations over a distance of 5 km.

fragments, gravel, sand, silty–sandy clay (non-softening), or hard–soft clay


One prolonged breakdown of one pumping station or a power outage
could overwhelm the system and begin flooding the powerhouse.

fillings or collapse of rock mass structure leading to tunnel failures


To secure tunnel stability in the fault zone and reduce water inflow,
a number of additional pre- and post-excavation measures were taken.
Jhelum River Crossing Fault (comprising all main rock units)

Pre-excavation grouting increased in significance and occurrence with


the overlap of successive fans reaching up to 8 m in some cases. The
Irregular, Tabular due to fracture cleavage, Crushed

tunnel profile was enlarged to accommodate heavier support, and its


roof was pre-supported using spiling or forepole umbrella arches.
Finally, excavation round lengths were reduced to less than 2 m.
Smooth–Slickensided, Undulating–Planar

Post-excavation measures in the JRCF zone included installation of


4+ to Earthlike (fault breccia/gouge)

stiffer initial support. Single or double layers of reinforced ribs of


shotcrete (RRS), typically anchored with 5-m-long fully grouted rock
bolts, or 3-rod lattice girder sets embedded in 250-mm-thick FRS were
installed while taking care to complete the rock support one round
behind the tunnel face. To mitigate floor heave and improve overall
tunnel stability by forming a ring closure, the tunnel invert was also
0.03 [0.001–0.06]
Slightly–Highly

Extremely poor

reinforced with RRS or lattice girder sets that tied in with the structural
elements of the tunnel walls.
1.8–2.5
30–60

5–150

4.1. Major groundwater inflow in the JRCF


Dry or minor inflow, often medium

The measures for support and watertightness were not employed all
at once during the excavations in the JRCF; some were implemented
inflow near bedding contacts

after an incident prompted further action. For example, invert closures


Tabular, Blocky, Irregular

were introduced in the construction after a major groundwater inflow


Silty or clayey coatings,
occasionally soft clay

from the tunnel floor occurred in the JRCF at station (STA) D4L + 418
of the left branch of the headrace twin tunnel (LHTT).
3 to 3+ random
Smooth, Planar
Fresh–Slightly

2.2 [0.7–3.7]

The inflow occurred approximately 4 m behind the face in a tunnel


Mudstone

section that had a pre-excavation grouting fan, was under a 12-m-long


0.25–5

forepole umbrella arch, and its initial tunnel support was completed.
3–25

Poor

The tunnel face had been blasted and sealed with a layer of 60-mm-
thick FRS 24 h prior to the sudden water ingress event during in-
stallation of RRS and rock bolts. Within the first twelve hours of the
Dry or minor inflow, often medium
Blocky, Tabular, Locally Irregular

event, the observed inflow measured was 100 m3/h, and the water was
Clean, sandy or silty coatings

inflow near bedding contacts

contained, channeled away, and pumped out.


The ground conditions at the location of the inflow comprised an
Rough–Smooth, Planar

intensely fractured and water-bearing siltstone overlying a highly


3.0–3.6

sheared 5-m-thick siltstone/mudstone layer (Fig. 4b). The sheared


3 to 3+ random
Fresh–Slightly

layer, which marked the downstream end of the JRCF zone, was not
8.9 [0.7–17]
Lot C2 D&B tunnel rock mass characterization of Murree formation and JRCF.

exposed in the particular tunnel location. Instead, the layer was located
Siltstone

0.1–10

approximately 2 to 3 m beneath the invert dipping with a high angle


3–25

Fair

toward the upstream and forming a low permeability water barrier for
groundwater. Although this area had been pre-grouted, the subsequent
blasting led to rock mass relaxation around the tunnel. Since the invert
Clean, sandy particles, or hard calcite

Dry or minor inflow, often medium

was unsupported, the joints in the siltstone dilated, creating ground-


Massive, Blocky Locally Irregular

water flow paths. Additionally, because of high substantial ground-


inflow near bedding contacts

water pressures of 1.8 MPa, grout was likely eroded from some joints by
2+ random to 3+ random
Rough, Planar–Undulating

the flows.
Early remediation measures involved construction of an approxi-
mately 8-m-long, particularly thick (1.5 m), heavily reinforced, con-
Fresh–Slightly

crete invert tied in with the tunnel walls and with embedded flow
17 [1.1–33]
Sandstone

< 0.1–10

through drainage pipes equipped with valves. Vertical and sub-vertical


grouting holes, 3 to 8 m long, were drilled through the invert and grout
Good
1–22

was injected under gradually increasing pressure to reach 10 MPa re-


fusal pressure to grout the rock mass to a depth of 8 m below the invert.
Average Q-value [min–max

Results at pressures less than 5 MPa were initially promising; however,


Joint aperture or thickness
Joint roughness, waviness

P-wave velocity (km/s)

the concrete invert cracked when 5 MPa was reached, allowing the
Volumetric joint count

Number of joints sets

Average rock quality

grout to escape to the invert. Furthermore, the water discharge rate


classification

increased to 180 m3/h.


(joints/m3)

The deteriorating conditions and greater seepage necessitated a


Joint filling
Weathering

range]
Parameter

(mm)
Structure

more robust solution. Thus, a tunnel plug was constructed above the
Seepage
Table 4

invert, as shown in Fig. 6, to provide invert confinement pressure up to


10 MPa to complete remedial grouting. The plug was placed in six 1.5-

8
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Fig. 5. (a) Typical wedge-cut blasting pattern used in the C2 D&B tunnels. (b) Plan view of the wedge-cut showing geometry of blast holes with #1 to #4 delay
detonators. (c) Charging of tunnel face on site. (d) Installation of geophones of vibration data logger UBOX-501(TDEC, 2008) shown in window.

m-high lifts of mass concrete totaling 610 m3. At every lift, 0.5-m-wide Following the construction of the tunnel plug, grouting with pres-
benches were made at the downstream side of the plug to facilitate sures up to 10 MPa to depths of 8 to 10 m below the invert successfully
work for the successive lifts and for remedial grouting. The plug was reduced groundwater inflow from the JRCF to approximately less than
unreinforced, with the exception of the wire mesh between the concrete 0.3 m3/h per linear meter of grout hole. Despite this success of the plug
lifts to improve bonding and some shear reinforcement bars placed and grouting, the tunnel plug (Fig. 7) needed to be demolished for the
along the tunnel perimeter to provide better resistance against uplift at heading to continue.
the interface between the plug and the tunnel walls.

Fig. 6. Longitudinal and perpendicular cross sections of concrete tunnel plug from STA D4L + 423 to STA D4L + 414.

9
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

and delivery on site of expansive agents or a diamond saw.


Furthermore, diamond sawing would have required specialized per-
sonnel to safely operate the equipment near the tunnel support.
Consequently, to minimize delays and use the most cost-effective
and well-known local construction practices, drilling and blasting was
preferred to demolish the tunnel plug. A blasting design that would lay
out drilling, charging, and firing patterns to break the mass concrete in
an efficient manner without causing a new tunnel breach and water
inflow became immediately crucial.
Poorly designed and executed tunnel plug demolition involved the
risk of reactivating the groundwater inflow. This had occurred in the
opposite tunnel branch within the JRCF where another plug had been
installed, and during its removal, a poorly executed blast had damaged
part of the initial support. After the face had advanced approximately
20 m, groundwater inflow was reactivated through a breach in the re-
stored tunnel support.
According to Hoek and Brown (2019), very poor blasting, in terms
of control of drilling, charging, and firing sequence design, can lead to
Fig. 7. View of the downstream end of tunnel plug near STA D4L + 423. severe damage 2 or 3 m into the surrounding mass in hard rock tunnels.
However, this did not apply to the JRCF because the rock mass in the
5. Controlled blasting design fault zone was not hard rock but a heterogeneous ground as mentioned
in Section 3. Furthermore, although pre-excavation grouting was gen-
To avoid damaging the tunnel periphery and the grouted rock mass, erally efficient in sealing the rock mass in the JRCF, locations existed,
four methods were considered to demolish the plug, including drilling particularly in zones of clayey fault gouge, in which groundwater was
and blasting, expansive chemical agents, machine-mounted diamond lurking as close as 5 m behind the tunnel periphery. The drilling of rock
sawing, and mechanically using hydraulic breakers, hand-held rock bolts revealed this during typical support installations in the twin
splitters, and jackhammers. Drilling and blasting would be the most tunnels.
efficient method, but it posed the greatest risk of damage to the tunnel Considering the adverse ground conditions, extensive blasting da-
support and rock mass, while the other methods were less invasive. mage in the tunnel periphery around the concrete plug, such as major
However, chemicals, sawing, and breakers were considered to be time cracking of the rock mass and loosening or widening of grout-healed
consuming. Extra delays would have incurred also from procurement joints, could have developed favorable conditions for the groundwater

Fig. 8. Controlled blasting design approach.

10
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

to force its way closer, and combined with rock mass relaxation, breach The P-wave velocity was further corrected, as suggested by Barton
the tunnel support. Given that hardened reinforced shotcrete requires (2002), for the effect of in situ stress or tunnel depth and the rock mass
vibration levels as high as 1500–2000 mm/s from nearby blasts (Ahmed porosity by considering a depth of 200 m and a porosity of 10%. Al-
and Ansell, 2014), the primary goal of controlled blasting was to pre- though the major water inflow was at a depth of 180 m, a 200-m depth
vent either new large cracks to develop or existing healed joints to was conservatively selected as more representative of the in situ stress,
reopen within 3 m of the tunnel periphery; in other words, induce only particularly of the major horizontal stress induced by the compressional
minor damage. tectonic environment (Ma et al., 2018).
Fig. 8 demonstrates the adopted control blasting design approach. The recorded Q-values from the geological tunnel mappings of the
To determine the maximum explosive charge that would cause only 20-m-long tunnel section represent the ground conditions after ex-
minor damage to the rock mass, the critical particle velocity described cavation but prior to the major water inflow. However, to estimate the
by Eq. (6) was used. Additionally, a predictive site-specific vibration actual rock mass quality of the section, particularly in the invert, after
attenuation model, similar to Eq. (1), was developed using the vibration the erection of the tunnel plug and the remedial grouting, Barton’s
monitoring program data from elsewhere in the D&B tunnels. By in- (2002) recommendations regarding the effect of high-pressure grouting
corporating the critical particle velocity into the vibration attenuation on the Q-parameters were considered.
model, a permissible instantaneous charge weight (Wperm) was defined Barton (2002) noted that a successful high-pressure grouting pro-
in relation to the distance from the charge to the tunnel periphery. In gram would result, beyond direct permeability reduction, in the con-
this way, firing blast holes having a charge weight of Wperm would re- solidation or stiffening of the rock mass which could be reflected po-
sult in only minor damage to the critical rock mass zone. sitively in the individual Q-parameters. Barton suggested an assessment
Having defined Wperm, the geometric characteristics and the firing of this effect on each of the Q-parameters. His recommendations were
sequence of blasts were deduced by selecting a PF for concrete and used to extrapolate rock mass quality to the pre-tunneling as well as the
applying Eq. (7). The selection of an appropriate PF for concrete was post-remediation conditions from the Q-values derived during tun-
based on technical literature and experience from previous small con- neling.
crete demolitions elsewhere in the project. By defining a relative small Thus, three cases were defined for the studied tunnel section. Case 0
shot length, Eq. (7) provided an estimate of the burden, the spacing, represents the original ground conditions prior to any tunneling. Case 1
and the number of allowable instantaneous detonations. represents the ground conditions following tunnel excavation, as re-
Successful controlled blasting also requires effective site im- corded in the actual tunnel mappings, and Case 2 represents the ground
plementation via robust quality control procedures. This was achieved conditions following tunnel plug construction and remedial grouting.
by strict oversight of the drilling, charging, and firing operations and For Case 2, the estimated Q-values refer to the 8-m-long tunnel section
application of an improved vibration monitoring program for plug de- where the tunnel plug was erected.
molition. Vibration data collected during the plug demolition blasts was To capture the effect of variations in compressive and tensile
fed back into the design, enabling blasting design refinements. strength, sensitivity analyses were performed for the three cases. The
compressive strength influences the calculated particle velocity by af-
5.1. Critical particle velocity fecting the P-wave velocity of rock mass in Eq. (8), while the tensile
strength has a direct effect on the velocity through Eq. (6). The
Laboratory data for the intact rock parameters of the main rock minimum, maximum, and mean values of compressive and tensile
types (Table 3) were used to derive a critical particle velocity value. strength (Table 3) of the three rock types were used to develop
Furthermore, geological tunnel mappings and the Q-values of an ap- weighted averages for each recorded mapping for input into Eq. (6) and
proximately 20-m-long tunnel section in the LHTT (from STA Eq. (8).
D4L + 433.9 to STA D4L + 414.7), including the location of the major Table 5 indicates that the average Q-value increased more than an
water inflow, were also considered. order of magnitude from Case 0 through Case 2. The P-wave velocity of
Inherent in Eq. (6) is that the rock mass is homogeneous, isotropic, the in situ rock mass (Case 0) is somewhat less than what the seismic
and elastic. Thus, the values for density, compressive strength, and surveying estimated during the feasibility stage (Table 4). The differ-
tensile strength of the intact rock were selected as weighted averages ence might be attributed to limitations with seismic testing, such as
based on the percentage of each rock type encountered within the 20-m interference of seismic waves along steeply dipping strata and velocity
tunnel section. Furthermore, instead of an intact rock dynamic tensile inversion cases whereby higher velocity layers overlie lower velocity
strength, the laboratory values of static tensile strength were used as a layers. These limitations and the lack of calibration boreholes in the
more conservative approach. Similarly, the density of intact rock was feasibility stage caused the authors to believe that the estimation of the
assumed to be the density of the rock mass. P-wave velocity of the JRCF zone based on the extrapolation of the
Although the P-wave velocity for the in situ condition of the rock recorded Q-values during tunneling represents better the original
mass had been estimated during the feasibility stage (Table 4), a more ground conditions.
representative value for the condition of the rock mass following the The critical particle velocity appeared to decrease from Case 0 to
remedial measures was calculated according to Barton (2002), who Case 2, which is expected since remedial grouting consolidated and
correlated rock mass P-wave velocity (in km/s) to the Q-value. strengthened the rock mass. In Case 2, which reflects the ground con-
ditions following the restoration of watertightness in the invert, the
CP 3.5 + log10 (Q·( 100)) (8)
mean, minimum, and maximum velocity values (516, 415, and
c

where σc is the uniaxial compressive strength of the intact rock in MPa. 584 mm/s, respectively) were in general accordance with the PPV-

Table 5
Results of critical particle velocity (Vcr) based on sensitivity analysis. SST, SLT, and MST represent sandstone, siltstone, and mudstone, respectively.
Case Participation of rock unit in the study section Q [min–max] Q·(σc/100) [min–max] CP (km/s) [min–max] Vcr (mm/s) [min–max]

SST (%) SLT (%) MST (%)

0 13 54 33 0.01 [0.003–0.02] 0.005 [0.0005–0.01] 1.5 [0.9–2.0] 1567 [1191–2572]


1 0.1 [0.04–0.4] 0.06 [0.008–0.3] 2.4 [1.8–3.2] 949 [628–1313]
2 – 94 6 8.3 [6.3–9.8] 4.4 [1.3–8.4] 4.3 [3.8–4.6] 516 [415–584]

11
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

based damage criteria of Table 1 for developing minor blast damage. remove the first four benches (that is, the first 2 m) of the plug. In each
The minimum and maximum values of critical particle velocity for Case bench blasting, eleven 42-mm-diameter blast holes were to be spaced
2 were used to develop the controlled blast design to demolish the plug. evenly along a single row with their ends at least 1 m away from the
tunnel walls. Blast holes ranged in depth between 1.9 and 2.4 m ac-
5.2. Vibration attenuation model cording to the height of the bench. Line drilling was also performed in
all vertical blasts by drilling two empty relief holes at each side, posi-
Vibration monitoring data collected from the typical rock tunnel tioned 0.3 m from the tunnel walls and drilled to the same depth as the
blasting (see Appendix Table A1) were employed to develop the ground blast holes.
vibration attenuation model for the rock mass conditions in Lot C2. The All bench blast holes were charged to the minimum Wperm con-
recorded data were plotted in log-log space, similar to Eq. (2), and a sidering their distance from the tunnel walls. The design required in-
linear regression analysis was performed using a square-root SD. The dependent detonation of the blast holes via a different half-second delay
constants, K and β, were determined with a coefficient of determination detonator per hole. Based on the blast hole patterns, PF values for the
R2 = 0.743. Reverting to the exponential form, as in Eq. (1), the vi- bench blasts were in the range of 0.61–0.73 kg/m3 for the lower (wide)
bration attenuation law was found to be one-third of the blast holes, which was deemed to be satisfactory ac-
cording to the typical breakage criteria for plain concrete.
PPV = 647·(R Wmax) 1.90
(9) Following vertical bench blasting, horizontal line drilling was per-
formed along the perimeter of the remaining 6-m-long plug for addi-
where R/√Wmax is the square-root SD in m/√kg, and PPV is the max-
tional damage control. The line drilling comprised 42-mm-diameter
imum peak particle velocity in mm/s.
relief holes, 200 mm off the tunnel walls and at 300 to 600 mm center-
The regression statistics are summarized in a tabulated format in the
to-center spacing.
Appendix Table A2. Fig. 9 presents the PPV data versus the square-root
An innovative solution was devised to remove the remaining plug
SD and the vibration attenuation law with a 95% confidence band. The
by horizontal contour blasting in three 2-m-long rounds. This design
confidence band was constructed by considering the combined dis-
included a 1 × 1 × 2 m3 (width × height × length) burn cut to form an
tribution of K and β of individual predicted values for 95% confidence
opening at the center and near the lower one-fourth of the plug. The
(see also the Appendix).
burn cut incorporated three uncharged drilled holes 127 mm in dia-
meter surrounded by eleven blast holes (42 mm in diameter) arranged
5.3. Permissible instantaneous charge weight
in two squares and detonated by millisecond delays. The concrete
around the burn cut was demolished in layers by firing of 42-mm-dia-
Rearranging the terms of the vibration attenuation law and solving
meter horizontal contour blast holes with hole-by-hole detonations
for the maximum instantaneous charge weight, Eq. (9) becomes Eq.
(Fig. 12).
(10)
Five blasting contours were developed around the cut with a burden
Wmax = (647 PPV) 1.05 ·R2.00
(10) of 1 m each. Based on the hole diameter and length, the 1-m burden was
considered adequate to achieve good fragmentation. Contour blast hole
Substituting PPV with the maximum and minimum values of the spacing was designed at 1.0 to 1.25 m with the exception of the lateral
critical particle velocity from Case 2, (Vcr = 584 mm/s and 415 mm/s, holes in contours 3, 4, and 5. Due to their closer proximity to the tunnel
respectively), a range of permissible instantaneous charge weights periphery, the hole spacing for contour 3, 4, and 5 was 0.6 to 0.8 m to
(Wperm) that would limit damage to “minor” is defined by Eq. (11). reduce possible overbreak.
Wperm = 0.63·R2to 0.90·R2 (11) Eq. (11) was used to evaluate permissible charge weights for the
burn cut and the blasting contour holes with respect to their distance
According to Eq. (11), a permissible charge weight, detonated at the from the nearest rock mass behind the tunnel walls and roof. Particular
center of the nearly 9-m diameter tunnel plug, of up to 12.6 to 18.0 kg care was taken for the contour holes nearest to the tunnel walls. Using
would produce only minor blast damage to the rock mass. A blast hole ten different numbers of half-second delay detonators plus an in-
detonated just 1 m from the rock mass would be limited to 630 to 900 g stantaneous detonator, up to eleven holes were allowed to be charged
of charge weight to produce only minor damage. Fig. 10 represents Eq. and were detonated at a time in each blast.
(11) and was developed for use in the field to quickly evaluate Wperm It should be noted that the burden and spacing of the contour holes
limits with respect to the distance from the blast source to the rock could have been determined by Eq. (7) using a single PF suitable for
mass. concrete. This, however, would have complicated the drilling of con-
tour holes because it would require different burden and spacing di-
5.4. Drilling charging and firing patterns mensions according to the location of the blasting contours and the

The blasting pattern adopted for the controlled demolition of the


tunnel plug drew advantages mainly from contour blasting techniques
in which a small number of independent charges were fired each time
using half-second delay detonators. The independent (hole-by-hole)
detonation of charges was dictated from the permissible instantaneous
charge weights calculated particularly for blast holes near the tunnel
perimeter. Additionally, the independent detonations were performed
to avoid constructive interference of seismic waves from the detonation
of more than one blast holes in the same delay interval (cooperating
charges).
The downstream stepped profile of the plug could negatively in-
terfere with horizontal blasting. Therefore, we decided to initially re-
move the benches up to the mid-height of the plug by vertical blasting
to develop a vertical face for working unobstructively with horizontal Fig. 9. Vibration attenuation model for Lot C2 D&B tunnels. PPV data versus
blasting. Fig. 11 shows the stepped profile of the tunnel plug and the square-root scaled distance, regression line, and 95% confidence band in a log-
three successive vertical blasting operations that were designed to log graph.

12
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Field supervision and data collection focused on implementing the


following:

• Drilling patterns: recording the number, diameter, burden, spacing,


and length of blast holes.
• Charging pattern: recording the explosive weight (number of car-
tridges) per hole, placement of cartridges within the hole, deto-
nating cord connections, and stemming.
• Delay sequencing; associating proper delay numbers with each hole;
controlling the instantaneous charge weight per delay interval.
• Ground vibration monitoring: installing a vibration data logger with
three geophones set 15 to 20 m behind the tunnel plug.

Wabox 80% cartridges, decoupled, evenly distributed, and con-


Fig. 10. Permissible instantaneous charge weight (Wperm) curves for minor nected with detonating cord, were used for charging the holes.
blasting damage with respect to distance (R) from blasting source. Stemming consisted of 1 m of drill cuttings. Although the blasting im-
plementation attempted to maintain the location of blasting contours
respective permissible charge weight relative to the tunnel perimeter. according to the design, small variations in the size and shape of the
To ease the field implementation of the blasting design, the simplified openings formed by the burn cut or the preceding contour blasts dic-
burden and spacing dimensions mentioned above were followed. Thus, tated subsequent alterations to the contour geometry. Nonetheless, the
average PF was allowed to vary between 0.42 kg/m3 for contour holes burden, spacing, and depth of contour holes were generally in ac-
near the tunnel perimeter and 1.60 kg/m3 for holes in blasting contours cordance with the design. After each blast, loose or hanging concrete
1 and 2. These values were deemed acceptable for the breakage of the was scaled manually or using a hydraulic breaker.
concrete, albeit the latter value caused high fragmentation and ex- Table 6 summarizes the plug demolition details. The nearest con-
cessive flyrock. tour holes to the rock mass had a distance of 0.7 m in the field and were
Since both burn cut and contour holes had a drilling length of 2 m, placed in contours 4 and 5. The actual instantaneous charge weights
the sequence of a burn cut and blasting contours was repeated in three applied in the field (Wact) were kept lower than the permissible ones
rounds for the demolition of the remaining 6-m-long tunnel plug. The calculated by Eq. (11) based on the early results of the vibration
bottom of the tunnel plug, approximately 1 m in height, was to be re- monitoring program and observations regarding the effectiveness of
moved after blasting demolition of the upper part by using a hydraulic blasts with lighter charges. Exception to this were one burn cut, where
breaker. the actual charge weight was near the lower limit of the permissible
range, and holes in contours 4 and 5, where the actual charge weights
were at the upper limit of the permissible range. The average PF for
6. Blasting design implementation on site contour blasting of plain concrete was kept around 0.4 kg/m3.
Ground vibrations were recorded for thirteen of the twenty blasts
Implementing the controlled blasting design on site required a during plug demolition (refer to the Appendix Table A3). The recorded
quality control program that included rigorous supervision and collec- vibrations are presented in Fig. 13 as PPV versus SD (green triangles)
tion of blasting data. These data served as a basis for a back analysis of plotted over the vibration attenuation model in a log-log graph.
the blasting procedures, and when needed allowed, operational changes As can be seen from Fig. 13, PPV values during the plug demolition
to be quickly applied to maintain the permanent structure. are located around the upper 95% confidence limit. One explanation for

Fig. 11. General layout of vertical blasting for the first four benches of tunnel plug. Plan view (i) shows the blasting pattern for the 1st and 2nd blasting, whereas plan
view (ii) details the 3rd blasting.

13
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Fig. 12. (a) General layout of burn cut and contour blasting. The delay numbers in contour blast holes are for half-second delays, whereas the numbers of the burn cut
holes represent millisecond delays. (b) Marking of the blasting layout on the tunnel plug.

Table 6
Tunnel Plug Blasting Details.
Blasting details 1st bench 2nd bench 3rd bench Burn cut Contour blasting
blasting blasting blasting
Contour 1 Contour 2 Contour 3 Contour 4 Contour 5

No. of blast holes1 11 11 11 11 14–8 11–10 11 11 6–5


No. of relief holes1 2 at each side of tunnel walls 72 along tunnel periphery
2
Burden, B (m) 0.25–0.75 0.25–0.752 0.70 N/A 3
1.0 1.0 0.8–1.1 0.6–0.9 0.8–0.9
Spacing, S (m) 0.62 0.58 0.56 N/A3 1.0 1.0 1.1–1.5 0.8–1.1 0.6–1.1
Height/Depth, L (m) 2.4 1.9 1.9 2.0 2.0 2.0 2.0 2.0 2.0
Distance, R4 (m) 1.2 0.9 1.3 3.6 2.6 1.6 1.0 0.7 0.7
Wperm5, (kg) 0.95–1.4 0.55–0.79 1.0–1.5 8.1–11.6 4.2–6.0 1.6–2.3 0.63–0.90 0.31–0.44 0.31–0.44
Wact6 (kg) 0.34 0.29 0.37 1.8/9.07 0.86 0.86 0.58 0.44 0.44 (0.30–0.44)
(0.65–2.6) (0.65–1.0) (0.30–0.72) (0.30–0.44)
Fired holes/delay8 1 1 1 1/57 1 1 1 1 1
Powder Factor9 (kg/ 0.91–0.302 1.05–0.352 0.50 N/A 3
0.43 0.43 0.18–0.33 0.22–0.46 0.22–0.46
m3)

Notes:
1
All holes were 42 mm in diameter.
2
In the upper two-thirds and lower one-third of the hole height, respectively.
3
Non-applicable.
4
Distance of blast holes closest to rock mass.
5
Permissible instantaneous charge weight, as calculated by Eq. (11), for blast holes closest to rock mass.
6
Mean values of actual instantaneous charge weight (Wact) recorded for blast holes closest to rock mass. Minimum and maximum values are in parentheses.
7
The typical maximum Wact for the burn cuts was 1.8 kg, and the typical firing was 1 hole per delay. Because of a mismanagement of delay sequencing in a burn
cut, however, 5 holes detonated instantaneously.
8
#0–#10 half-second delay detonators were used for bench and contour blasting. #1–#11 ms delay detonators were used for the burn cuts.
9
Mean values of powder factor for blast holes closest to rock mass.

this may be the greater transmission factor for blasted concrete than the Moreover, line drilling helped to reduce damage by scattering shock
fractured rock mass. Another explanation is that the 1.5-m-thick re- waves and dissipating the explosives gases. The authors believe that the
inforced invert and the extensive remedial grouting around the tunnel vibrations caused by blasting did not damage the surrounding rock
plug consolidated the rock mass, which increased the transmission mass, as groundwater inflow was not reactivated and no overbreak was
factor and the PPV values. Nonetheless, since the recorded vibrations observed behind the line drilling. A hydraulic breaker had to be em-
were near the 95% confidence limit, the vibration attenuation model ployed to remove the remaining strip of concrete behind the last
was considered representative. Therefore, the permissible charge blasting contour and re-profile the tunnel periphery. The lower 1-m
weights are also valid. part of the tunnel plug were additionally demolished using a hydraulic
As mentioned before, the overall blasting design was conservative to breaker. Judging by the final outcome, it was considered that line
minimize the risk of damage to the rock and reactivating the ground- drilling could have been avoided if the design was a bit less con-
water inflow. The extensive remedial grouting also improved the water servative.
tightness and further strengthened the rock mass in the invert. To expedite tunneling construction and to make up for the time lost

14
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

principle aspect of the methodology was to determine the permissible


instantaneous charge weight that would maintain PPV below the pre-
scribed range near the tunnel periphery to limit any damage to rock
mass and tunnel support, from none to “minor” characterized as a
couple of millimeters of tensile fracturing. A site-specific vibration at-
tenuation model was developed to incorporate the critical particle ve-
locity above which tensile cracking initiates. The critical particle ve-
locity was based on the material properties of the main rock types
around the tunnel plug, as determined by laboratory tests and rock
mass classification methods during tunnel mappings. The vibration at-
tenuation model was obtained from statistical analysis of vibration data
collected during previous blasting operations in the headrace tunnel D&
Fig. 13. Plot of the vibration data from the blasting demolition of the tunnel B excavations.
plug (green triangles) on the Lot C2 D&B vibration attenuation model. (For Vibration monitoring data collected during the plug demolition in-
interpretation of the references to color in this figure legend, the reader is re- dicated that the predictive attenuation model was sufficiently accurate,
ferred to the web version of this article.) aligning along the upper 95 percent confidence limit, which might be
attributed to the higher seismic velocities of massive concrete.
Similarly, high-pressure rock grouting around the plug likely increased
during the mitigation of the water inflow, no testing was performed to the seismic velocities. Therefore, the authors believe that a correction of
directly assess or quantify blasting damage imparted to the rock mass. up to +2 standards deviations of the attenuation model should be
Following plug demolition, the tunnel surfaces were carefully in- considered in future similar works to better represent the ground
spected, and no notable surface cracks were observed. Furthermore, transmission factor due to the strengthening of the surrounding rock
delamination tests were performed on the tunnel shotcrete to assess for mass or the presence of massive concrete.
drummy, de-bonded areas; no problems were revealed. After resuming The controlled blast methods were successfully employed to de-
regular tunnel construction, no further instabilities or water inflows molish the plug in a safe, efficient, and economic manner, thus showing
were observed in this tunnel section. that blasting with selective use of mechanical means is an attractive
The plug was successfully and entirely demolished in eight days solution when compared to less invasive methods such as expansive
working two 10-hour shifts per day. Satisfactory results were achieved chemical agents or strictly mechanical methods. The methodology
using controlled blasting methods more economically and efficiently could have been improved by using a geophysical method such as the
than any of the other proposed methods considered. tunnel seismic prediction to estimate the P-wave velocities of the rock
mass that surrounded the tunnel plug more accurately.
7. Conclusions

The controlled demolition of a temporary tunnel plug in the NJHEP Acknowledgments


headrace tunnels by blasting was discussed in this study. The 610-m3
unreinforced concrete plug was constructed to restore watertightness The authors would like to thank the Water and Power Development
after a major groundwater inflow occurred while excavating through Authority (WAPDA) of Pakistan for allowing the publication of project
difficult fault zone ground conditions. data. The authors are also grateful to the two anonymous reviewers
To remove the plug efficiently while managing the risk of re- whose comments greatly improved the quality of the paper. This work
activating groundwater inflow, it was necessary to control blast vibra- was supported by the National Research Foundation of Korea (NRF)
tions. A preliminary review of published literature on blasting damage grant funded by Ministry of Science and ICT (MIST) of Republic of
and ground vibrations revealed general associations regarding degree of Korea (No. NRF-2018R1A5A1025137). The authors Abubakar Sharafat
damage to the rock mass and peak particle velocities. and Waqas Arshad Tanoli are extremely thankful to the Higher
A blasting methodology was developed to demolish the plug using Education Commission of Pakistan for Human Resource Development
contour blasting techniques, and a small number of light charges were Initiative-Universities of Engineering Science and Technology scholar-
fired independently each time to remove the concrete in layers. The ship.

Appendix

In Fig. 9, the 95% confidence band around the regression line was plotted using the following equation:
C.I. (individual predicted value) = Yi ± t (n 2)· (MSres ·(1 + (1 n) + ((Xi X)2 SS X )))

where

• C.I. is the confidence interval of an individual predicted value


• Y is the predicted value of log(K) + β·log(R/√W ) using the K and β coefficients predicted by the regression analysis
max i

• t is the (1 ± 0.025)-th quantile of the Student’s t-distribution with n − 2 degrees of freedom


i
(n−2)

• n is the number of observations in the sample


• MS is the mean square of the residuals (observed Y – predicted Y)
res

• XX == log(R/√W
log(R/√W ) is a given value of log(R/√W ) in the sample

i max i max
) is the mean value of log (R/√W ) in the sample
max mean max

• SS is the sum of squared deviations of log(R/√W ) in the sample


X max

15
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Table A1
NJHEP Lot C2 vibration monitoring data during standard blasting operations.
Blast No. Wmax (kg) R (m) PPVL PPVT PPVV PPV f (Hz)
(mm/s) (mm/s) (mm/s) (mm/s)

1 31.8 25.1 43.20 39.78 29.88 43.20 129.242


2 23.1 28.4 38.01 28.79 18.94 38.01 152.283
3 24.0 80.7 2.27 0.19 1.20 2.27 57.220
4 24.2 30.4 21.57 17.08 10.49 21.57 117.798
5 28.4 32.1 10.86 22.28 9.19 22.28 192.566
6 21.5 34.2 26.57 13.35 9.62 26.57 86.060
7 32.0 86.0 1.89 1.15 0.47 1.89 117.190
8 20.5 49.8 15.83 7.50 7.59 15.83 125.580
9 18.4 68.0 4.91 0.80 2.73 4.91 63.324
10 32.0 16.0 15.45 51.19 27.10 51.19 93.384
11 32.0 48.0 2.95 2.63 3.74 3.74 92.010
12 21.0 44.6 6.87 18.59 22.36 22.36 91.705
13 27.6 44.8 7.55 11.15 6.75 11.15 42.458
14 32.1 47.5 10.75 9.89 9.16 10.75 80.610
15 32.5 50.9 12.94 6.85 5.78 12.94 120.153
16 18.9 63.8 2.99 1.32 1.30 2.99 104.065
17 18.9 41.5 6.34 25.23 13.27 25.23 90.332
18 18.4 42.3 3.17 11.71 5.79 11.71 89.111
19 16.8 44.7 3.20 2.35 5.92 5.92 83.618
20 18.4 97.8 0.76 1.15 0.91 1.15 93.994
21 12.4 18.6 2.79 13.85 4.05 13.85 110.931
22 18.9 29.1 1.56 27.88 12.64 27.88 54.474
23 21.0 28.4 17.73 36.88 8.31 36.88 48.676
24 18.9 27.3 11.05 27.99 7.67 27.99 138.397
25 21.2 26.2 10.80 26.00 10.02 26.00 135.498
26 18.9 25.6 3.52 18.61 14.31 18.61 79.193
27 14.7 25.1 1.66 2.91 8.84 8.84 77.972
28 21.2 25.0 13.46 29.01 1.90 29.01 134.430
29 15.1 25.2 20.09 31.55 11.20 31.55 132.904
30 18.9 25.7 11.51 6.43 9.97 11.51 78.278
31 18.8 78.6 2.17 1.52 0.16 2.17 57.678
32 18.4 90.2 2.19 0.95 0.17 2.19 142.719
33 29.4 64.5 2.33 1.31 0.16 2.33 141.032
34 29.4 50.4 3.64 0.33 0.17 3.64 57.678
35 27.3 65.8 3.70 1.05 0.16 3.70 6.561
36 28.0 75.4 2.25 7.79 0.29 7.79 75.684
37 18.0 97.5 2.27 0.27 0.33 2.27 2.441

Table A2
Linear regression output of the standard vibration monitoring program.
Model summary (NJHEP vibration monitoring data in log-log space)

R R square Standard error of estimate Count of independent variables Observations Adjusted R square

0.862 0.743 0.557 1 37 0.735

Analysis of variance (ANOVA)

Sum of squares Degrees of freedom Mean square F Significance F

Regression 31.347 1 31.347 101.093 0.000


Residual 10.853 35 0.310
Total 42.199 36

Coefficients and 95% confidence intervals

Coefficient Standard error t-Statistic P-value Lower 95% Upper 95%

K (Intercept) 6.472 0.424 15.266 0.000 5.612 7.333


R/√Wmax (Slope) −1.897 0.189 −10.055 0.000 −2.280 −1.514

16
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Table A3
Vibration monitoring data during blasting demolition of tunnel plug in Lot C2 LHTT.
Blast No. Wmax (kg) R (m) PPVL PPVT PPVV PPV f (Hz)
(mm/s) (mm/s) (mm/s) (mm/s)

1 0.37 15 4.00 2.82 4.46 4.46 169.373


2 0.63 16 6.17 9.83 4.17 9.83 40.153
3 1.80 17 13.98 6.70 11.42 13.98 129.842
4 3.20 17 26.79 12.81 20.21 26.79 98.877
5 1.07 17 11.18 10.56 6.56 11.18 67.340
6 9.00 19 39.38 25.48 73.22 73.22 91.858
7 1.20 19 7.46 9.23 3.45 9.23 62.478
8 1.77 19 11.03 15.82 5.18 15.82 80.534
9 1.77 19 10.60 5.75 5.85 10.60 92.673
10 1.77 20 4.31 10.24 5.70 10.24 86.517
11 2.13 20 4.60 10.12 7.56 10.12 93.689
12 1.23 20 3.20 5.89 4.32 5.89 92.468
13 0.90 20 9.51 5.09 3.26 9.51 52.458

References Ma, C.S., Chen, W.Z., Tan, X.J., Tian, H.M., Yang, J.P., Yu, J.X., 2018. Novel rockburst
criterion based on the TBM tunnel construction of the Neelum-Jhelum (NJ) hydro-
electric project in Pakistan. Tunn. Undergr. Space Technol. 81, 391–402. https://doi.
Ahmed, L., Ansell, A., 2014. Vibration vulnerability of shotcrete on tunnel walls during org/10.1016/j.tust.2018.06.032.
construction blasting. Tunn. Undergr. Space Technol. 42, 105–111. https://doi.org/ Marinos, V., 2012. Assessing rock mass behaviour for tunnelling. Environ. Eng. Geosci. 18
10.1016/j.tust.2014.02.008. (4), 327–341. https://doi.org/10.2113/gseegeosci.18.4.327.
Barton, N., 2002. Some new Q-value correlations to assist in site characterization and Mavko, G., Mukerji, T., Dvorkin, J., 2003. The Rock Physics Handbook. Tools for Seismic
tunnel design. Int. J. Rock Mech. Min. Sci. 39, 185–216. https://doi.org/10.1016/ Analysis in Porous Media. Cambridge University Press, Cambridge, UK.
S1365-1609(02)00011-4. McKenzie, C., Holley, K., 2004. A study of damage profiles behind blasts. In: Proceedings
Bastante, F.G., Alejano, L., González-Cao, J., 2012. Predicting the extent of blast-induced of the Annual Conference on Explosives and Blasting Technique, vol. 2. International
damage in rock masses. Int. J. Rock Mech. Min. Sci. 56, 44–53. https://doi.org/10. Society of Explosives Engineers, pp. 203–214.
1016/j.ijrmms.2012.07.023. Medley, E.W., 1994. The Engineering Characterization of Melanges and Similar Block-in-
Bauer, A., Calder, P.N., 1970. Open Pit and Blasting, Seminar Mining Engineering. Matrix Rocks (Bimrocks). Department of Civil Engineering, University of California at
Department Publication, Queen’s University, Kingston, Ontario. Berkley (Doctoral dissertation) (accessed July 05, 2018).
Bogdanoff, I., 1996. Vibrations measurements in the damage zone in tunnel blasting. In: Meyer, T., Dunn, P.G., 1996. Fragmentation and rock mass damage assessment – Sunburst
Mohanty, B. (Ed.), Rock Fragmentation by Blasting. Fragblast-5. Balkema, Rotterdam, excavator and drill and blast. In: In: Aubertin, Hassani, Mitri (Eds.), Rock Mechanics
Netherlands, pp. 177–185. Tools and Techniques, vol. 1. Balkema, pp. 609–617.
Brady, B.H.G., Brown, E.T., 1993. Rock Mechanics For Underground Mining, second ed. Murthy, V.M.S.R., Dey, K., 2003. Predicting overbreak from blast vibration monitoring in
Chapman & Hall, London, UK. a Lake Tap Tunnel – a success story. Fragblast: Int. J. Blast. Fragment. 7 (3), 149–166.
Bureau of Indian Standards, 1973. IS 6922–1973 (Reaffirmed 2003): Criteria for Safety https://doi.org/10.1076/frag.7.3.149.16787.
and Design of Structures Subject to Underground Blasts. BIS Earthquake Engineering Nateghi, R., Kiany, M., Gholipouri, O., 2009. Control negative effects of blasting waves on
Sectional Committee, New Delhi, India. concrete of the structures by analyzing of parameters of ground vibration. Tunn.
Charlie, W.A., 1985. Review of Present Practices Used in Predicting the Effects of Blasting Undergr. Space Technol. 24 (6), 608–616. https://doi.org/10.1016/j.tust.2009.04.
on Pore Pressure. Geotechnical Branch, Division of Research and Laboratory Services, 004.
Engineering and Research Center, U.S. Department of the Interior, Bureau of Olofsson, S.O., 1990. Applied Explosives Technology for Construction and Mining, second
Reclamation, Denver, CO (accessed June 10, 2018). ed. APPLEX AB, Arla, Sweden.
Dey, K., Murthy, V.M.S.R., 2012. Prediction of blast-induced overbreak from uncontrolled Oriard, L.L., 1982a. Influence of blasting on slope stability: state-of-the-art. In: Brawner,
burn-cut blasting in tunnels driven through medium rock class. Tunn. Undergr. Space C.O. (Ed.), Third International Conference on Stability in Surface Mining, June 1, 2,
Technol. 28, 49–56. https://doi.org/10.1016/j.tust.2011.09.004. 3, 1981, Vancouver, British Columbia, Canada: [proceedings]. Society of Mining
Hoek, E., Brown, E.T., 2019. The Hoek-Brown failure criterion and GSI – 2018 edition. J. Engineers of AIME, New York, NY (accessed July 09, 2018).
Rock Mech. Geotech. Eng. https://doi.org/10.1016/j.jrmge.2018.08.001. Oriard, L.L., 1982b. Blasting effect and their control. In: Hustrulid, W.A. (Ed.),
Holmberg, R., 1982. Charge calculation for tunnelling. In: Hustrulid, W.A. (Ed.), Underground Mining Methods Handbook. Soc. of Mining Engineers of the American
Underground Mining Methods Handbook. Soc. of Mining Engineers of the American Inst. of Mining, Metallurgical, and Petroleum Engineers, New York, NY, pp.
Inst. of Mining, Metallurgical, and Petroleum Engineers, New York, NY, pp. 1590–1603.
1580–1589. Raina, A.K., Chakraborty, A.K., Ramulu, M., Jethwa, J.L., 2000. Rock mass damage from
Hussain, A., Yeats, R.S., MonaLisa, 2009. Geological setting of the 8 October 2005 underground blasting, a literature review, and lab- and full scale tests to estimate
Kashmir eartquake. J. Seismolog. 13 (3), 315–325. https://doi.org/10.1007/s10950- crack depth by ultrasonic method. Fragblast: Int. J. Blast. Fragment. 4 (2), 103–125.
008-9101-7. https://doi.org/10.1076/frag.4.2.103.7449.
International Society for Rock Mechanics Commission on Standardization of Laboratory Rustan, A., Naarttijarvi, T., Ludvig, B., 1985. Controlled blasting in hard intense jointed
and Field Tests, 1978. Suggested methods for the quantitative description of dis- rocks in tunnels. CIM Bull. 78 (884), 63–68.
continuities in rock masses. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 15 (6), Singh, S.P., Lamond, R.D., 1993. Investigation of blast damage and underground stability.
319–368. In: Peng, S.S. (Ed.), Proceedings of 12th International Conference on Ground Control
International Society of Explosives Engineers Standards Committee, 2015. ISEE Field in Mining, August 3-5, 1993. Department of Mining Engineering, College of Mineral
Practice Guidelines for Blasting Seismographs. International Society of Explosives and Energy Resources, West Virginia University, Morgantown, WV (accessed July 14,
Engineers, Cleveland, OH (accessed June 14, 2018). 2018).
Jadoon, I.A.K., Hinderer, M., Kausar, A.B., Qureshi, A.A., Baig, M.S., Basharat, M., Frisch, J.F. Stiehr (Ed.), 2011. ISEE Blasters’ Handbook. 18th ed. International Society of
W., 2015. Structural interpretation and geo-hazard assessment of a locking line: 2005 Explosives Engineers, Cleveland, OH.
Kashmir earthquake, western Himalayas. Environ. Earth Sci. 73 (11), 7587–7602. Trigueros, E., Cánovas, M., Muñoz, J.M., Cospedal, J., 2017. A methodology based on
https://doi.org/10.1007/s12665-014-3929-7. geomechanical and geophysical techniques to avoid ornamental stone damage caused
Kaneda, H., Nakata, T., Tsutsumi, H., Kondo, H., Sugito, N., Awata, Y., Akhtar, S.S., by blast-induced ground vibrations. Int. J. Rock Mech. Min. Sci. 93, 196–200. https://
Majid, A., Khattal, W., Awan, A.A., Yeats, R.S., Hussain, A., Ashraf, M., Wesnousky, doi.org/10.1016/j.ijrmms.2016.12.013.
S.G., Kausar, A.B., 2008. Surface Rupture of the 2005 Kashmir, Pakistan, earthquake Tripathy, G.R., Gupta, I.D., 2002. Prediction of ground vibrations due to construction
and its active tectonic implications. Bull. Seismol. Soc. Am. 98 (2), 521–577. https:// blasts in different types of rock. Rock Mech. Rock Eng. 35 (3), 195–204. https://doi.
doi.org/10.1785/0120070073. org/10.1007/s00603-001-0022-9.
Kizilbash, M.H., Dickson, P.A., Jaffery, N.A., 2017. Neelum Jhelum hydroelectric project: Tunstall, A.M., Djordjevic, N., Villalobos, H.A., 1977. Assessment of rock mass damage
pre-excavation grouting in twin headrace tunnels. Geotech. Special Publ. 287, from smooth wall blasting at El Soldado mine, Chile. Trans. Institut. Min. Metall. 106
238–248. https://doi.org/10.1061/9780784480786.024. (A1–50), A42–A46.
Langefors, U., Kihlström, B., 1973. The Modern Technique of Rock Blasting. John Wiley & Turab, S.A., 2012. Structural Geology of the Surroundings of Muzaffarabad with
Sons, New York, NY.

17
A. Sharafat, et al. Tunnelling and Underground Space Technology 93 (2019) 103098

Emphasis on Neotectonics (MPhil dissertation). University of Peshawar, Pakistan. 97–106.


Yilmaz, O., Unlu, T., 2014. An application of the modified Holmberg-Persson approach Van Eeckhout, E.M., 1987. Controlled Blasting and Its Implications for the NNWSI Project
for tunnel blasting design. Tunn. Undergr. Space Technol. 44, 113–122. https://doi. Exploratory Shaft. Los Alamos National Laboratory, Los Alamos, NM (accessed July
org/10.1016/j.tust.2014.04.009. 28, 2018).
Yu, T., Croxal, J., 1985. Kidd Creek blasthole stopping. In: Proceedings from the Wah Nobel (PVT) Ltd. (2015). Wabox – 80%. Nitroglycerin Based Dynamites. Product
International Symposium on large scale underground mining. Lulea, Sweden, pp. Specifications. Wah Nobel Group of Companies, Wah Cantt, Pakistan.

18

You might also like