Sunteți pe pagina 1din 10

Tunnelling and Underground Space Technology 43 (2014) 113–122

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology


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

An application of the modified Holmberg–Persson approach for tunnel


blasting design
Ozgur Yilmaz a,⇑, Tugrul Unlu b,1
a
Department of Mining & Mineral Extraction, Zonguldak Vocational School, Bülent Ecevit University, Zonguldak 67500, Turkey
b
Department of Mining Engineering, Engineering Faculty, Bülent Ecevit University, Zonguldak 67100, Turkey

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

Article history: In this study, the applicability of the modified Holmberg–Persson approach for a tunnel blast design was
Received 6 February 2014 investigated. For this purpose, firstly, a detailed description of the modified Holmberg–Persson approach
Received in revised form 7 April 2014 was made. After that, data collection program for the statistical analysis of the ground vibration param-
Accepted 28 April 2014
eters induced by the tunnel blasting was presented and then the determination of the critical peak par-
ticle velocity level associated with the rock damage was described. Finally, site-specific limiting peak
particle velocity values obtained by measuring the overbreak after blasting were given. The results of
Keywords:
the field investigations were discussed and a new equation to calculate the crushed zone radius was
Holmberg–Persson approach
Tunnel blasting
derived. The studies have shown that Holmberg–Persson approach which was originally suggested for
Blasting damage perimeter control design technique, can also be applied for a full blasting round to select the proper
Overbreak explosive-hole combinations.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction ology for blasting engineering applications which include the


development of site specific indices for damage prediction and
Drilling and blasting application is generally inevitable for rock damage control. However, an error in the mathematics behind
excavation activities in mining, quarrying and civil construction the H–P approach was discovered by Hustrulid and Lu (2002).
works. Therefore, the use of explosives is probably the most widely Iverson et al. (2008) later introduced a modification in the H–P cal-
used means of crushing rock, as well as the most cost-effective rock culation procedure so as to mitigating the problem, at least
excavation method in mining and construction operations. Predict- partially.
ing and limiting the blast induced damage zones are very impor- H–P approach is one of the most widely used engineering meth-
tant for the long term stability of underground openings which ods to model the site specific attenuation of the blast waves in the
are excavated by drilling and blasting method. Perimeter control rock mass. The main purpose of this study is to show the applica-
blasting techniques are commonly used in civil construction pro- bility of the modified H–P approach for a tunnel blast design. Even
jects to overcome difficulties arising from drilling and blasting though this approach is a perimeter control design technique, it
operations. In mining applications, however, these techniques are was used to design the whole blasting round in this work. In that
not so commonly used. Therefore, in underground mine galleries, regard, a limiting PPV value was obtained using the relationship
poorly designed blasting operations may result in overbreak and between peak particle velocity, linear charge concentration, dis-
unwanted structural damages. tance from the charge and the observed extent of damage, and then
In the late 1970s, Holmberg and Persson (H–P) introduced a it was applied to the design curves for different linear charge con-
perimeter control design technique based on the peak particle centrations to select explosive-hole combinations.
velocity (PPV) generated by the detonation of a charge
(Holmberg and Persson, 1978, 1979). H–P approach gained quite
wide acceptance due to the logical basis and the relative ease of 2. Description of the modified Holmberg–Persson approach
the application. This method provides a unique practical method-
Since seismic waves decay with distance in a fairly regular man-
⇑ Corresponding author. Tel.: +90 3722574010x1525; fax: +90 3722574023. ner, they are predictable with the acceptable accuracy (Nateghi,
E-mail addresses: ozgyilmaz9@hotmail.com (O. Yilmaz), unlutugrul@hotmail.
2011). The H–P approach is based on the fact that rock damage
com (T. Unlu). due to blasting being related to the peak particle velocity associ-
1
Tel.:+90 3722574010x1197; fax: +90 3722574023. ated with the blast produced seismic waves (Holmberg and

http://dx.doi.org/10.1016/j.tust.2014.04.009
0886-7798/Ó 2014 Elsevier Ltd. All rights reserved.
114 O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122

Persson, 1978). The seismic wave energy associated with the blast to the arrival of the seismic waves from the different elemental
induced ground vibration decreases with distances due to decay of charges is to be determined at an observation point (r0–z0) located
both the amplitude and the frequency of vibration. The ground along the r axis. Since the waves from the elemental charges travel
vibration recorded at a certain location depends principally on different distances to reach the observation point, their amplitudes
the amount of explosive charge per delay, the distance between will be distance dependent. Also, in general, the arrival times and
source and measurement points and the elastic properties of the wave orientation will vary depending on the velocity of detonation
transmitting medium (Tripathy and Gupta, 2002). Many authors of the explosive and the wave velocity through the rock mass.
used the scaled distance concept to estimate peak particle velocity Holmberg and Persson assumed that the entire charge detonates
of the ground vibration (Duvall and Petkof, 1959; Langefors and instantaneously, the amplitudes are simply summed without con-
Kihlstrom, 1963; Ambraseys and Hendron, 1968). The scaled dis- sidering arrival direction and PPV is proportional to the dynamic
tance is a normalized factor that combines the distance with the strain experienced by the rock mass. It was also assumed in this
explosives energy to give a single number which can be used for approach that the peak particle velocity due to each small element
further calculations. Holmberg and Persson (1978) does not take of charge within the blast hole is numerically additive. Besides, the
a particular charge symmetry into consideration to predict peak effect of free face boundaries and the velocity of detonation of the
particle velocity and uses the following general equation: explosive charge have been neglected for practical purposes
(Holmberg and Persson, 1978, 1979). This simplifies the situation
PPV ¼ K  W a =Rb ð1Þ
considerably and the resulting PPV is obtained by summing the
where PPV is peak particle velocity (mm/s), W is the maximum contributions from the different elemental charges:
charge per delay (kg), R is the distance between blast location and X
n X
n
vibration monitoring point (m), and K, a, b are site specific con- PPV ¼ DPPVi ¼ Kqa DLa Rbi ð4Þ
stants which can be determined by multiple regression analysis. 1 1

The term K reflects the source energy and the coupling efficiency The above assumptions allowed the derivation of a simple non-
of the explosive to the blasthole wall. Higher values of K indicates linear relationship to describe the peak particle velocity attenua-
high energy and well coupled explosives. The term b represents tion in the near field. Under these conditions Holmberg and
the loss of vibrational energy with distance. Higher values of b rep- Persson replaced the summation by the following integral
resents less competent rock mass which attenuates vibrational expression:
energy more quickly while lower values represents a competent 2 3a
rock mass with little fracture which transmits the vibrational Z zf
energy with little attenuation (Scott, 2009). 6 dz 7
PPV ¼ K 4q h ib=2a 5 ð5Þ
The Eq. (1) above is based upon the assumption that the deto- zi 2 2
ðr 0 Þ þ ðz  z0 Þ
nation occurs at a single point and hence it is only valid when
the distance R is large compared to the length of the charge. When However Hustrulid and Lu (2002) pointed out that the step
the point under consideration is close to a long charge, like the case from PPV which was expressed as a summation in Eq. (4) to PPV
in the tunnel blasting, the peak particle velocity must be obtained which was expressed by the integral in Eq. (5) is not correct. The
by integration over the charge length (Hoek and Brown, 1980). exponent a was moved from inside the summation sign to outside
Holmberg and Persson (1978) divided the long charge into a num- of the integral sign. This error can be corrected by simply reverting
ber (n) of small elemental charges of equal length DL. Assuming back to Eq. (4). Since DL is the same for all of the elemental charges
that the charge concentration per unit length is q, the charge at the summation expression, this term can be removed from
weight for each element can be expressed as: under the summation sign to yield (Iverson et al., 2008; Tesarik
and Hustrulid, 2009):
DW ¼ qDL ð2Þ
X
n X
n
where DW is the charge weight of each elemental charge, DL is the PPV ¼ DPPVi ¼ Kqa DLa 1=Rbi ð6Þ
length of each elemental charge and q is the charge concentration 1 1
per unit length. Consequently, the peak particle velocity at a given
It is clearly seen in Eq. (6) that the PPV depends on the length of
observation point due to the arrival of a particular elemental charge
the elemental charge raised to the power a. Only for the very spe-
denoted by the subscript i may be expressed by the following
cial case, a = 1, the equation is stable. For the case a > 1, the PPV
equation:
decreases to zero as the elemental length decreases. For the case
DPPVi ¼ KðDWÞa =Rbi ¼ KðqDLÞa =Rbi ð3Þ of a < 1, the PPV increases to infinity as the elemental length
decreases. Thus, this procedure cannot be followed. NIOSH (The
Assumed problem geometry to simplify the discussion is shown National Institute for Occupational Safety and Health) has revisited
in Fig. 1. As can be seen, the charge lies along the z axis with the r the basic concepts that Holmberg–Persson approach involved and
axis passing through the mid-point of the charge. The total PPV due correct the mathematical problems. The new solution involves
determining the average travel distance to the observation position
for all of the elemental charges. In this approach, the PPV is given
by Iverson et al. (2008):
 b ¼ KðqLÞa =R
PPV ¼ KW a =R b ð7Þ
where R  is the average travel distance. According to Martin (2007)
the average travel distance can be defined by the following indefi-
nite integral expression (Iverson et al., 2008):
Z zf qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1

R ðz  z0 Þ2 þ ðr  r0 Þ2 dz ð8Þ
zf  zi zi

The value of the indefinite integral may be written as (Weast,


Fig. 1. Schematic representation of the simplified charge geometry. (1983), from Iverson et al. (2008):
O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122 115

Z qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
equipment at such short and narrow galleries, the typical v-cut
ðz  z0 Þ2 þ ðr  r 0 Þ2 dz ¼ 1=2 ðz  z0 Þ ðz  z0 Þ2 þ ðr  r 0 Þ2
blast pattern is used at the mine as a cut design.
 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Since the nominal diameter and length of the used drill at the
þðr  r0 Þ2 loge ðz  z0 Þ þ ðz  z0 Þ2 þ ðr  r0 Þ2 ð9Þ mine is 0.38 m and 1.6 m, respectively, all blastholes are drilled
to a depth of 1.6 m. Besides, explosives must not exceed half-
length of the blasthole according to the safety regulations. There-
fore, blastholes were charged to a length of 0.8 m. The uncharged
3. Test site description and data collection
part of the blastholes were stemmed by a plastic water cartridge
and mud. At the each double v-cut pattern, 20–24 blastholes in
Site investigations were conducted at an underground develop-
4 rows were drilled depending on the site conditions.
ment roadway in a hard coal mine located in Üzülmez district
Methane safe permissible dynamite produced by a state owned
which belongs to Zonguldak province in Turkey. Turkey’s main
company (i.e. Mechanical and Chemical Industries Corporation) is
hard coal deposits are located in the Zonguldak basin, between Ere-
used in the TTK. The use of this explosive is primarily suitable for
ğli and Amasra on the Black Sea coast in north-western Turkey.
underground coal blasting operations and for the blasting of soft
Turkish Hardcoal Enterprise (TTK) is the biggest hard coal producer
rocks. Since the weight strength of this explosive is low, it is essen-
in Turkey and it is a state owned enterprise being responsible for
tially not suitable for hard rock blasting operations. However, it is
the operation and administration of all hard coal activities in Tur-
the only reliable product that is tested for mine gases, methane in
key. Because of the widely dispersed nature of the Zonguldak hard
particular, and coal dust in underground coal mines in Turkey
coal basins, TTK has established five production areas from west to
(MKEK, 2013). The technical properties of this methane safe per-
east; Kozlu, Üzülmez, Karadon, Armutçuk and Amasra. The first
missible type dynamite are given Table 2. Electric-delay blasting
three districts are located within a 10 km radius of Zonguldak, on
caps (copper) are also used at this mine according to safety
the northeast range of the basin. Mining operations are concen-
regulations.
trated in six collieries of the five districts. Extraction is carried
out simultaneously at several horizons within the collieries and
all of the production comes using underground mining methods. 3.2. Vibration Measurements
In the coalfield the average thickness of coal seams is 2 m and
the inclination of these seams vary between 0° and 90°. Ground vibration components induced by the cut blasting
The Zonguldak coalfield comprises a belt of carboniferous coal rounds during the face advancing were monitored by using a triax-
measures which contain more than 20 mineable coal seams. The ial ISEE (International Society of Explosives Engineers) geophone
geology is characterized by steeply dipping seams which in certain (Instantel™). For each blast round, geophone locations were cho-
areas lie beneath the Black Sea. The strata is highly disturbed by sen as close as possible to the drift face. This distance is practically
major and minor faults and folds and contain large volumes of determined as 10–60 m.
methane. The coal also has a high tendency to spontaneous com- It may not be important to anchor the transducer (sometimes it
bustion and is susceptible to outbursts. These ever present prob- is called as geophone because it holds three geophones inside) to
lems has meant that mining has developed in a highly labour the measurement surface when particle acceleration is less than
intensive manner, for which the opportunities for mechanization 0.3 g. However, when the particle acceleration is greater than
are limited. In the basin, stratigraphy includes palaeozoic, meso- 1.0 g, bolting or cementing of the transducer is required. The stan-
zoic, tertiary, and crystalline units. Rock units are found as volca- dard transducer can be used not only horizontally on the floor but
nic, volcano–sedimentary, and sedimentary of origin. also as wall-mounted in a vertical position. The geophone was
The application roadway has been planned as 250 m in length mounted to the sidewall surface of the roadway by drilling a short
and 10 m2 cross-sectional area with the nominal dimensions of hole into the rock mass, installing a dowel and an anchor bolt, and
4.065 m wide and 3.360 m high (crown). The overburden above then attaching the geophone onto the anchor bolt using a washer
the roadway is about 200 m. The rock mass in this area are mostly and nut with the arrow pointed in the direction of the event.
medium to coarse-grained sandstone. Table 1 gives the summary Fig. 2 shows an installed dowel and an anchor bolt onto the side
of the average static intact rock properties of this rock formation wall of the development roadway. Precautions were taken to pro-
obtained from the laboratory tests. tect the geophone cables from flying rock from the blast rounds
and impacts from mobile equipment.
3.1. Drilling and blasting round design A standard transducer may be bolted to a surface directly by
itself or with the aid of an optional levelling plate. However, before
The cut holes were only blasted during the measurement stud- installing the standard transducer, the position of the vertical and
ies in each blasting round and resulting PPV values of the ground transverse geophones must be switched. By this means, the geo-
vibration were recorded to associate the rock damage with the phones are repositioned to their natural axes compensating for
resultant PPV. Remaining part of the roadway face were drilled the vertical wall installation (Instantel, 2005).
and blasted afterward by the miners without measurement.
Because of the inefficient use of the electro-hydraulic drilling
Table 2
Technical properties of methane safe permissible type dynamite (MKEK, 2013).
Table 1
Intact rock properties of formation. Parameters Values
Density 1.10 gr/cm3
Parameter Value ± stdev
Cartridge length 200 mm
Unconfined compressive strength, rc (MPa) 65.2 ± 17.2 Cartridge weight 125 gr
Brazilian tensile strength, rtB (MPa) 8.46 ± 0.28 Bullet Block Test 280 cm3/10 g
Deformation modulus, E (GPa) 34.9 ± 3.12 Detonation velocity (Ø 125 mm PVC Tube) 5000 m/s
Poisson’s ratio, t 0.17 ± 0.06 Weight strength (S) 0.57
Density, qr (kg/cm3) 2630 ± 20 Relative weight strength compared to ANFO (SANFO) 0.655
P-wave velocity, Cp (m/s)a 3345 Linear charge concentration, q 0.625 kg/m
a
Equivalent linear charge concentration compared to ANFO, qA 0.409 kg/m
This property has been taken from Karakul and Ulusay (2012).
116 O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122

maximum charge per delay, and distance between blasting loca-


tion to vibration monitoring point, give an equation on site specific
velocity attenuation. In order to establish a relationship among
that parameters, multiple regression analysis was performed using
the data given in Table 3.
If the dependent variable is a linear combination of the several
independent variables, it is called as multiple linear regression. The
basic multiple regression model for two independent variables is
given in Eq. (11) below:

Y ¼ a þ bX 1 þ cX 2 ð11Þ

where Y is the dependent variable, X1 and X2 are the independent


variables and a, b, and c are the constants.
Fig. 2. Installation of a dowel and anchor bolt to the sidewall. To apply the H–P approach for roadway blasting design at a par-
ticular site, values for K, a, and b are required. In the basic H–P
Instantel Minimate Plus™ data logger was used to collect parti- equation which was modified by Iverson et al. (2008) (see Eq.
(7)), PPV is the dependent variable, W and R  are the independent
cle velocity data and Blastware™ (ver. 10) software was used to cal-
culate vector sums of the transverse, vertical, and longitudinal wave variables and K, a, and b are the constants and therefore, this equa-
forms. Measured magnitudes are tabulated for six different times tion is suitable for multiple regression analysis. Basic equation is
and they represent velocities in each of the three axes. Since the linearized by taking the natural logarithm of both sides in Eq. (12):
peak vector sum (PVS) reflects the effect of the other two compo- 
ln PPV ¼ ln K þ a ln W  b ln R ð12Þ
nents it is often preferred over the PPV. In most blasting operations,
the PVS occurs at about the same time as the PPV of one of the com- In linear regression analyses, the SPSS Statistics (ver. 19) is used
ponents, but the addition of the other two components increases its for statistical analysis (IBM, 2010). Summary of the multiple
magnitude (Alcudia et al., 2007). The maximum of the sums is the regression output is given in Table 4. It presents the standard sum-
peak vector sum and does not necessarily occur at the PPV of an mary output of the SPSS multiple regression analysis performed
individual waveform. The peak vector sum represents the resultant within the data analysis tool. The R square (coefficient of determi-
particle velocity magnitude it is computed as follows: nation) quantity is the proportion of the variation in the dependent
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi variable explained by the regression model, and is a basic measure
PVS ¼ T 2p þ V 2p þ L2p ð10Þ of the quality of fit of the model. In this case, a value of 0.889 indi-
cates that 88.9% of the PPV variability is explained by the linear
where Tp is the particle velocity along the transverse plane, Vp is the regression. The analysis of variance (anova) table divides the total
particle velocity along the vertical plane and Lp is the particle veloc- variation in the dependent variable into two components, of which
ity along the longitudinal plane. The MiniMate Plus™ calculates the one can be attributed to the regression model (regression) whereas
peak vector sum for each point of the sampled waveforms and dis- the other one cannot (residual). If the significance level for the f-
plays the largest value. Moreover, the magnitude of the resultant test is less than 0.05, then the hypothesis that there is no (linear)
vector is always greater than the magnitude of its individual com- relationship can be rejected. Since the significance level of vari-
ponents. The results of the ground vibration measurements and ables was computed as 60.006, multiple correlation coefficients
date are given in Table 3. The statistical evolution of these measure- can be called statistically significant. The intercept coefficient is
ments will later be discussed in the following section. obtained from the linear regression in the log–log transformed
space. It should be noted that e5.254 equals to 191.3. Finally, the site
4. The statistical analysis of the measurement results specific coefficients values of a = 1.016 and b = 1.445 are easily
extracted from the summary output.
When statistical analysis techniques were applied to blast As can be seen in Table 4, the coefficients of the modified H–P
vibration data, relationship among the peak particle velocity, the model K, a and b were determined as 191.3, 1.016 and 1.445

Table 3
Results of ground vibration measurements.

Round No. Round date PVS (mm/s) Frequency (Hz) Maximum charge per delay, W (kg) Distance, R (m)
1 06.02.2013 60.6 228 13.125 14.5
2 08.02.2013 25.8 205 7.250 16.0
3 12.02.2013 10.1 – 5.000 25.3
4 26.02.2013 12.2 256 3.000 11.7
5 04.03.2013 8.72 – 1.800 11.8
6 13.03.2013 4.69 >400 6.000 36.4
7 15.03.2013 8.53 >400 7.125 22.0
8 21.03.2013 42.0 76 3.750 10.2
9 28.03.2013 3.65 – 6.25 60.5
10 05.04.2013 8.65 – 3.750 26.0
11 11.04.2013 17.1 186 4.500 16.0
12 25.04.2013 4.64 – 5.250 43.2
13 30.04.2013 16.7 193 5.000 18.5
14 06.05.2013 19.7 102 5.500 12.0
15 20.05.2013 14.4 – 6.250 14.5
16 22.05.2013 23.8 341 6.125 16.7
17 28.05.2013 27.4 – 7.500 16.5
18 11.06.2013 56.9 44 6.65 10.0
O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122 117

Table 4
Summary of multiple regression output from v. 19 SPSS statistical software (IBM, 2010).

Regression statistics
Multiple R R square Adjusted R square Standard error Observation
0.943 0.889 0.874 0.2979 18
Analysis of variance (Anova)
DF Sum of squares Mean square F Significance F
Regression 2 10.655 5.328 60.006 0.000
Residual 15 1.332 0.089
Total 17 11.987
Variables in the equation
Parameter Coefficients Standard error Beta t-test Significance
K (constant) 5.254 0.481 10.914 0.000
R variable 1.445 0.153 0.821 9.464 0.000
W variable 1.016 0.152 0.578 6.666 0.000

respectively. If there is no a vibration monitoring device, the for-


mula can also be used to estimate particle velocity for similar oper-
ations as a preliminary approach. Table 5 gives a comparison for
the values of these coefficients obtained by different researchers
for various type rocks and the coefficients obtained in this study.
The relative weight strength of the permissible explosive com-
pared to ANFO is very low (see Table 2). These may be the reason of
the fact that the site specific K value obtained in this study is rela-
tively low. Furthermore, a slightly low value obtained for the coef-
ficient b may be considered as an indication of the strength of the
sandstone of the Zonguldak Basin. However, it should be kept in
mind for this type of comparison that the explosives used vary in
different cases. Therefore, it would be more accurate that results
from the same rock type and the same amount of explosives have
to be used.
As a result of the simplification process in the coefficients, the
final design equation of the site specific H–P model for the field
application is as follows:

PPV ¼ 191ðqLÞ=R1:4 ðR2 ¼ 0:889Þ ð13Þ

where qL is equal to W that refers to the maximum explosive per


delay. In this design technique, the problem which immediately
Fig. 3. Schematic representation of one set of possible paths for the seismic waves
arises is the selection of the distance to be chosen. As can be seen traveling from the cut holes to the side wall-mounted geophone.
from the schematic plan view of the v-cut holes shown in Fig. 3,
the waves from the different holes must travel different distances
to arrive at the monitoring point. Here, it can be bear in mind that
the pulse from the nearest hole to the side wall would be the one Since half of the length of the blasthole depth was loaded with
arriving first, because it has the shortest distance to travel. Accord- explosives during the applications, charge length L was chosen as
ing to Iverson et al. (2008), the solution is that, for a given charge, 0.8 m. The average travel distance R  was obtained by the analysis
the highest amplitude should be associated with the closest holes. of the indefinite integral given by Eq. (9) according to charge length
This is considered to be the most important reason for the apparent and distance from the source point. After that, the site specific
‘‘spread’’ in amplitudes. It is considered that taking an ‘‘average’’ modified H–P design curves can be obtained. The modified H–P
value to represent the holes is not correct. Rather, the most regular design curves obtained using the site constants determined for
curve of highest amplitude was chosen to represent each blasthole Üzülmez sandstones are shown in Fig. 4 with respect to different
sequence. In the analyses, the travel distance was assumed equal to linear charge concentration. The explosives differ in energy content
the distance from the geophone to the face of the roadway and it per unit weight and thus normalization with respect to ANFO is
was also assumed that R  R  for the charge lengths and distances. performed. Traditionally, if a limiting PPV value is known, based

Table 5
Comparison of the H–P coefficients obtained from different studies.

Researchers K a b Rock formation


Lundborg et al. (1978) 700 0.70 1.50 Hard igneous swedish rocks
Scott (1997) 456 1.12 2.24a Urquhart shales, Queensland
Bilgin and Esen (1998) 108 1.05 0.52 Concrete block (rc = 8 MPa)
McKenzie et al. (1995) 400 0.78 1.56 Strong sedimentary strata
Onederra (2004) 400 0.94 1.88a Igneous rocks, Salt Lake City
Iverson et al. (2008) 171 1.40 1.00 Magmatic rock (Norite), montana
Presented in this study 191 1.00 1.40 Medium-grained sandstone, Zonguldak
a
This equations have been simplified by assuming b = 2a for the case of square root scaling.
118 O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122

wave or compressional wave velocity of the rock (m/s) and qr is


the density of the rock (kg/m3).

6. Calculation of the PPV at the end of the crushing zone

Kutter and Fairhurst (1971) indicated that there are different


zones of varying destruction and deformation around the borehole.
These zones are; the hydrodynamic zone, the non-linear zone, and
the elastic zone. In the first zone, the radial compressive stresses
generated from the shockwave exceeds the dynamic compressive
strength of the surrounding rock, and develops complete crushing
as rock fail in compression. For this reason, this zone is also called
as crushed zone. To calculate the PPV at the end of crushing zone,
the simple equation for the stress in a plane sinusoidal stress wave
Fig. 4. The modified H–P design curves for the site constants determined for the is adopted (Persson et al., 2001):
Üzülmez mine.
Peq C p
PPVc ¼ ð18Þ
upon this design curves given in Fig. 4, linear charge concentration Ed
for the expectable overbreak can be determined.
where PPVc is the PPV at the end of crushing zone, Peq is the equi-
5. Determination of the PPV associated with the rock damage librium pressure at the end of the crushing zone and Ed is the
dynamic deformation modulus of the rock. The PPVc value at this
Blasting damage in a rock mass is a result of the borehole strain point is independent of charge length and it is assumed to occur
developed by an explosive in a blasthole. According to Hook’s law, at the midpoint of the charge (Onederra and Esen, 2004). Because
the axial stress is related to the axial strain. This relationship can of the extremely high pressure and temperature, the measurement
be expressed as: of the dynamic pressure around a blasthole is usually costly and dif-
ficult to achieve. Therefore simple empirical equations can be used
r to serve as a useful tool to estimate the magnitude of such dynamic
e¼ ð14Þ
E pressures. This equilibrium pressure at the end of crushing zone can
where E is the elastic modulus of the material and r and e are the be computed from an empirical equation given by Hino (1956) as
axial stress and strain respectively. If the compression or extension follows:
of the rock mass is by an elastic axial wave motion, the same equa-  h
rc
tion can be applied, relating the dynamic stress and strain to each Peq ¼ Pb ð19Þ
other. The peak strain in the stress wave in the rock mass is related r0
to the PPV and compressional elastic wave velocity Cp. In this case, where Peq is the equilibrium pressure at a distance rc from the cen-
the relationship between PPV and peak strain in an idealized case, tre of the blasthole (Pa), Pb is the borehole wall pressure (Pa), r0 is
when a plane shock wave passes through an infinite elastic med- the blasthole radius (m) and h is the pressure decay factor. The pres-
ium, can be expressed as follow (Persson, 1996; Persson et al., sure decay factor is a function of rock and explosive properties. It is
2001): a negative number that has been found to be in the range of 1.24
PPV to 1.65 for a wide range of explosive and rock combinations and it
e¼ ð15Þ is generally assumed as 1.5 for an average value (Liu and Tidman,
Cp
1995). However, a first approximation can be obtained with the fol-
According to Eq. (15), measurement of the higher levels of PPV lowing empirical relationship given by Liu and Katsabanis (1993):
obtained at a fixed distance behind blasthole charged with equal  1=3
amounts of different explosives, in the same rock type with the Cp
h ¼ ð0:0085Ed þ 0:9955Þ ð20Þ
same burden and spacing, can be used to identify explosives show- Vd
ing a higher generation of the borehole strain. The higher levels of
where Vd is the detonation velocity of explosive (m/s) and Ed is the
PPV can therefore be used as an indicator of greater fragmentation
dynamic deformation modulus (GPa). Eissa and Kazi (1988) pro-
potential, thus allowing a more appropriate choice of explosive for
posed the following equation for the relationship between static
application in rock types requiring a high degree of fracturing
and dynamic deformation modulus of the rock material:
(McKenzie, 1993). Eq. (15) presents the relationship between peak
1=0:77
vibration levels PPV and induced peak strain e for a rock mass of Ed ¼ ðEs =100:02 Þ =qr ð21Þ
compressional wave velocity. Assuming a brittle failure mode of
rock, a threshold of critical peak particle velocity PPVcrit which where Es is static deformation modulus (GPa), Ed is the dynamic
can be withstood by the rock before tensile failure occurs can be deformation modulus (GPa) and qr is density of the rock material
computed from Eq. (16) given by Persson et al. (2001): (g/cm3).
The detonation pressure needs to be calculated in order to esti-
rt C p mate the pressure acting on the borehole wall since detonation is
PPVcrit ¼ ð16Þ
E the beginning phase of the explosive process. For the determina-
where rt is the tensile strength. In addition, the vibration level tion of the detonation pressure in an explosive wave propagating
above which some damage may be expected can be estimated from from a cylindrical charge, following simplified equation given by
following equation given by McKenzie et al. (1995): Fickett and Davis (1979) and Henrych (1979), which was derived
  from the equation of state of ideal gases, can be written as:
rt
PPVmax ¼ 1:2 ð17Þ
C p qr Pd ¼ qe V 2d =ð1 þ cÞ ð22Þ
where PPVmax is the vibration level above which minor damage may where Pd is the detonation pressure (Pa), Vd is the detonation veloc-
be observed, rt is the ensile strength of the rock (Pa), Cp is the p- ity (m/s), qe is the density of the c explosive (kg/m3) and c is the
O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122 119

specific heat ratio. Defourneaux (1973) gave an expression for the and photograph examinations. It was also observed during the
specific heat ratio with the density of CHNO (carbon, hydrogen, drifting that the space which occurred after the blasting was
nitrogen and oxygen) type of explosives as follows: mostly lined off with the last outer alignment of the cut holes.
However, it was encountered in some cases that the large crushed
c ¼ 1:9 þ 0:6qe ð23Þ
zones occurred over the last blasthole line.
In this way, the detonation pressure for many explosives can be Since the equivalent linear charge concentration compared to
obtained by using Eqs. (22) and (23). The detonation pressure gen- ANFO of the permissible explosive used in the TTK is 409 kg/m,
erates a gas pressure in the reacted portion of the explosive col- the site specific modified H–P design cure for the TTK mine would
umn. The gas pressure, often called explosion pressure, is the be as shown in Fig. 9. The final step in the process is the determi-
pressure that is exerted on the borehole walls by the expanding nation of the limiting PPV values. For the dynamite loaded cut
gases after the chemical reaction has been completed. Explosion holes, the size of crushed zone was on the order of 0.15 m. By
pressure is approximately one-half of the detonation pressure applying this information to the modified H–P design curve shown
(Konya and Walter, 1991). Besides, if the explosive charge was in in Fig. 9, limiting PPV value for the crushed zone is obtained as
intimate contact with the borehole wall, the explosion pressure follows:
would be the borehole wall pressure. The borehole wall pressure,
PPVcðcrusedzoneÞ ¼ 427 mm=s ð25Þ
thereby, can be estimated through following equation:
Pb ¼ Pd =2 ð24Þ PPV value at the borehole wall (R = 0) can also be calculated as
629 mm/s (Fig. 10). Applying Zonguldak rock mass properties
given in Table 1 for Eq. (16) results in a critical peak particle veloc-
7. Overbreak measurements and determination of the site- ity upper and lower levels, of which rock breakage or crushing
specific limiting PPV Values began, as follows:
PPVcrit ¼ 550—900 mm=s ð26Þ
In this study, the amount of overbreak by means of excavation
outside of the last line of cut holes was described as the measure of Holmberg and Persson (1979) used Eq. (16), to designate the limit-
crushed zone. Overbreak or crushed zone size measurements were ing PPV based on new crack formation in the surrounding rock mass
based on a series of photograph of the face that were taken before and they found that the critical peak particle velocity for new crack
and after the blasting operations (Figs. 5 and 6). At the each blast formation is between 700 and 1000 mm/s for Swedish hard igneous
round, outer boundary of the cut holes were lined with a luminous rocks. In this regard, it was observed that the site-specific PPVcrit
paint and then a photograph of the roadway face was taken (A). value of 550–900 mm/s obtained in this study is consistent with
Another photograph of the face was taken after blasting at the the previous study. However, Iverson et al. (2008) have reported
same frame (B). After that, two of them representing before and that the limiting PPV for new crack formation is 1850 mm/s for
after the blast were superposed with 50 per cent visibility and then the Montana magmatic rocks (Norite).
blasthole positions were marked (C). Finally, blasthole positions Scott (2009) obtained maximum PPV value of 465 mm/s that
before the blast and the formed crushed zone sizes were acquired the vibration level above which minor damage may be observed,
in the same photograph (D). applying the Bickham rock mass properties for Eq. (17) which were
The detailed photographs after the blast round including the given by McKenzie et al. (1995). In the same manner, applying Zon-
marked blastholes and hole alignments are given in Figs. 7 and 8. guldak rock mass properties for Eq. (17), the maximum PPV level at
In these photographs, the crushed zone limits were also signed which rock damage may began is obtained as follows:
as well. As a result, overbreak due to the dynamite-charged cut PPVmax ¼ 1154 mm=s ð27Þ
holes were determined as about 0.15 m by in situ observations

Fig. 5. Example of the overbreak determination by superposing the photographs (case 1).
120 O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122

Fig. 6. Example of the overbreak determination by superposing the photographs (case 2).

Fig. 7. Roadway face after blasting with marked blasthole positions and aliments. Fig. 9. Site-specific modified H–P design curve for permissible type dynamite.

Fig. 8. Roadway face after blasting with marked blasthole positions and aliments. Fig. 10. Site-specific modified H–P design curve for permissible type dynamite with
the observed damage.

This value was considered quite high to apply to the Holmberg ical equation to compute the equilibrium pressure at the end of
and Persson design curves. Therefore, it is clearly seen that this crushing zone in Eq. (18), had been given in Eq. (19) by Hino
cannot be a correct approach to design. (1956). If Eq. (19) is rewritten as it is:
Calculating the PPV at the end of crushing zone adopting the  h
simple equation for the stress in a plane sinusoidal stress wave rc
Peq ¼ Pb ð28Þ
had been given in Eq. (18) by Persson et al. (2001). Also, an empir- r0
O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122 121

If it is assumed in this equation that rock fails in compression in


the crushing zone due to the fact that the radial compressive stres-
ses generated from the shockwaves exceed the dynamic compres-
sive strength of the rock, and accepting equilibrium pressure at the
end of crushing zone (Peq) is equal to the dynamic compressive
strength of the rock (rcd), the crushed zone radius can be computed
from the flowing equation:
 1=h
rcd
rc ¼ r0 ð29Þ
Pb
where all terms are the same as described before. The dynamic
compressive strength of the rock material can be calculated from
the simple equation proposed by Yilmaz and Unlu (2013) as
follows:

rcd ¼ 0:4rc ðeÞ1=3 ð30Þ


where e is the strain rate (1/s). As the strain rates of rock material
near the charge hole may range from 103 to 105 1/s (Cai et al.
2007), the strain rate value of 104 1/s has been selected as an aver-
age value in Eq. (30). Applying the rock mass and explosive proper-
Fig. 12. Site-specific modified H–P design curve for different linear charge
ties given in Tables 1 and 2 for Eq. (29), the crushed zone radius is concentrations with various limiting PPV boundaries.
found as 12.4 cm. It can be seen that this value is close to the
obtained value from the site observations and measurements. 9. Conclusions

8. Discussion of results An alternative approach to design a tunnel blasting round for


proper explosive-hole combinations has been presented. The
The H–P design curve for permissible type dynamite used in approach uses the modified Holmberg–Persson approach which
TTK mine is given in Fig. 11 with the upper and lower levels of crit- was originally introduced as a perimeter control design technique.
ical peak particle velocity above which rock breakage may began. Eighteen blast rounds were used for the ground vibration and over-
According to the figure, upper breakage limit of Zonguldak sand- break measurements to apply this approach. It is proven in this
stone is found considerably above from the design curve which study that it is possible to design a tunnel blasting round with
was made for permissible type dynamite. However, approximately the modified Holmberg–Persson approach.
8 cm crushed zone radius value was obtained for the lower break- Holmberg–Persson site specific constants (K, a and b) used for
age limit. From the perspective of perimeter blasting techniques it the modelling near field peak particle velocity attenuation have
can be stated that overbreak was minimal and very smooth rock been obtained for the test site. Therefore, to predict peak particle
surface is obtained after blasting. However, it is also seen that velocity for controlled blast design, the obtained constants can
the permissible type dynamite is quite inadequate for blasting of conveniently be used for this site in the future.
the cut or stopping holes. It can also be concluded in this study that the roadway blasting
The modified H–P design curves for different linear charge con- applications performed with the permissible type explosive used in
centrations with the limiting PPV boundaries obtained by various TTK mine have a limited area of utilization, and may result in inef-
approaches are presented in Fig. 12. Based on these design curves, ficient blasting patterns. From this point of view, it is foreseen that
if GOM II type dynamite of which equivalent linear charge concen- in the blasting operations for cut or stopping, using a more power-
tration compared to ANFO is 1 kg/m, is used, the crushed zone (or ful explosive instead of permissible type dynamite with respect to
the overbreak) radius may be expected about 20–35 cm according linear charge concentration would yield more successful results in
to Eq. (16), 12 cm according to Eq. (17) and 44 cm according to TTK mine. In addition, it should be noted that careless use of these
site-observations. type powerful explosives for the perimeter blasting applications
can cause excessive overbreak easily.
A new equation to calculate the crushed zone radius has been
derived from an empirical equation that computes the equilibrium
pressure at the end of crushing zone. Site observations and mea-
surements were used to show the validity of this equation. It has
been found in the researches that there is a close agreement
between the results of suggested equation and site measurements.

Acknowledgements

The authors would like to thank the mining group staff at the
TTK Üzülmez specifically Birol KARAÇAM for providing an active
development drift to facilitate research in monitoring the blast
vibrations and measuring overbreak.

References

Fig. 11. Site-specific modified H–P design curve for permissible type dynamite with Alcudia, A.D., Stewart, R.R., Eliuk, N., Espersen, R., 2007. Vibration and air pressure
upper and lower levels of critical PPV. monitoring of seismic sources. CREWES Res. Rep. 19, 14.
122 O. Yilmaz, T. Unlu / Tunnelling and Underground Space Technology 43 (2014) 113–122

Ambraseys, N.R., Hendron, A.J., 1968. Dynamic behaviour of rock masses. In: Stagg, Langefors, U., Kihlstrom, B., 1963. The Modern Technique of Rock Blasting. John
K.G., Zienkiewicz, O.C. (Eds.), Rock Mechanics in Eng. Practice. John Wiley & Wiley & Sons Inc., New York, p. 405.
Sons, London, pp. 203–227. Liu, Q., Katsabanis, P.D., 1993. A theoretical approach to the stress waves around a
Bilgin H.A. and Esen S., 1998. Sıkılama malzemesinin patlatma performansına borehole and their effect on rock crushin. In: Proceedings of the Fourth
etkisi,3. Delme ve Patlatma Sempozyumu, Sßubat 22–27, Ankara, pp. 59–70. International Symposium on Rock Fragmentation by Blasting, Vienna, Austria,
Cai, M., Kaiser, P.K., Suorineni, F., Su, K., 2007. A study on the dynamic behavior of pp. 9–16.
the Meuse/Haute-Marne argillite. Phys. Chem. Earth. 32, 907–916. Liu, Q., Tidman, P., 1995. Estimation of the dynamic pressure around a fully loaded
Defourneaux, M., 1973. Sciences of Techniques de L’armement, pp. 872 . blasthole, Technical Report MRL 95-014, CANMET/MRL, pp. 15.
Duvall, W.I., Petkof, B., 1959. Spherical Propagation of Explosion Generated Strain Lundborg, N., Holmberg, R., Persson, P.A., 1978. The dependence of ground
Pulses in Rock. US Bureau of Mines, RI 5483, pp. 21–22. vibrations on distance and charge size, Report R11:78, Byggforskning (in
Eissa, E.A., Kazi, A., 1988. Relation between static and dynamic Young’s moduli of Swedish).
rocks. Int. J. Rock Mech. Min. 25 (6), 479–482. Martin, P.A., 2007. Personal communication by Iverson, et al. 2008, Department of
Fickett, W., Davis, W.C., 1979. Detonation. University of California Press, Berkeley, Mathematical and Computer Sciences. Colorado School of Mines, August.
CA. McKenzie, C.K., 1993. Methods of improving blasting operation. In: Hudson, J.A.
Henrych, J., 1979. The Dynamics of Explosion and its Use. Elsevier Scientific Pub. Co., (Ed.), Comprehensive Rock Engineering, vol. 4. Excavation, Support and
Amsterdam and New York. Monitoring, Pergamon, London, pp. 71–94.
Hino, K., 1956. Velocity of rock fragmentation and shape of shock wave. J. Ind. McKenzie, C.K., Scherpenisse, C.R., Jones, J.P., 1995. Application of computer assisted
Explos. Soc. 17 (4), 236–241. modelling to final wall blast design. In: EXPLO 95-A Conference Exploring the
Hoek, E., Brown, E.T., 1980. Underground Excavations in Rock. IMM, London, p. 527. Role of Rock Breakage in Mining and Quarrying, September 4–7, AusIMM,
Holmberg, R. Persson, P.A., 1978. The Swedish approach to contour blasting. In: Brisbane, Australia.
Proceedings of the 4th Conference on Explosives and Blasting Technique, SEE, MKEK, 2013. <www.mkek.gov.tr/en/products.aspx> Product Catalog.
pp. 113–127. Nateghi, R., 2011. Prediction of ground vibration level induced by blasting at rock
Holmberg, R., Persson, P.A., 1979. Design of tunnel perimeter blasthole patterns to units. Int. J. Rock Mech. Min. 48, 899–908.
prevent rock damage. In: Jones, J.M. (Ed.), Proceedings of the Tunneling’79. Onederra, I., Esen, S., 2004. An alternative approach to determinethe Holmberg–
Institution of Mining and Metallurgy, London, UK. Persson constants for modelling near field peak particle velocity attenuation.
Hustrulid, W., Lu, W., 2002 Some general design concepts regarding the control of Fragblast 8 (2), 61–84.
blast-ınduced damage during rock slope excavation. In: Proceedings of the Persson, P.A., 1996. The relation between strain energy, rock damage, fragmentation
Seventh International Symposium on Rock Fragmentation by Blasting, Beijing, and throw in rock blasting. In: Mohanty, B. (Ed.) Proceedings of the Fifth
China, pp. 595–604. International Symposium on Rock Fragmentation by Blasting, August 25–29,
IBM, 2010. SPSS (Statistical Package for the Social Sciences) Statistics, Version 19, Montreal, Canada, pp. 113–120.
IBM Company. Persson, A., Holmberg, R., Lee, J., 2001. Rock Blasting and Explosives Engineering.
Instantel, 2005. Minimate PlusTM Operator Manual, 716U0101 Rev 14, Printed in CRC Press, p. 540.
Canada. Scott C., 1997. The Effect of Blasting on the Stability of Underground Excavations in
Iverson, S.R., Kerkering, J.C., Hustrulid, W.A., 2008. Application of the NIOSH- Bedded Strata. M.Sc. Thesis, University of Queensland, Australia.
modified Holmberg–Persson approach to perimeter blast design. In: Scott, A., 2009. Appendix 16 – Review of the Impact of Blasting on Rock Mass
Proceedings of the Thirty-Fourth Annual Conference on Explosives and Permeability, Bickham Coal WRA & Draft WMP, pp. 43.
Blasting Technique, January 27–30, New Orleans, LA, pp. 1–33 Tesarik, D.R., Hustrulid, W.A., 2009. A hydrodynamics-based approach for
Karakul, H., Ulusay, R., 2012. Prediction of strength properties of rocks at different predicting the blast damage zone in drifting as demonstrated using concrete
saturation conditions from p-wave velocity and sensitivity of p-wave velocity to block data. Blasting Fragm. 3 (2), 141–166.
physical properties. Yerbilimleri 33 (3), 239–268 (in Turkish). Tripathy, G., Gupta, I.D., 2002. Prediction of ground vibrations due to construction
Konya, C.J., Walter, E.J., 1991. Rock Blasting and Overbreak Control, National blasts in different types of rock. Rock Mec. Rock Eng. 35 (3), 195–204.
Highway Institute. U.S. Dept. of Trans. Federal Highway Administration Office of Weast, R.C., 1983. Handbook of Chemistry and Physics, Editor-in-chief. 64th ed..
Implementation. CRC Press Inc., 1983–1984.
Kutter, H.K., Fairhurst, C., 1971. On the fracture process in blasting. Int. J. Rock Mech. Yilmaz, O., Unlu, T., 2013. Three dimensional numerical rock damage analysis under
Min. 8, 181–202. blasting load. Tunn. Undergr. Sp. Technol. 38, 266–278.

S-ar putea să vă placă și