Professional Documents
Culture Documents
A case study on the cavity effect of a water tunnel on the ground vibrations MARK
induced by excavating blasts
⁎
Xiang Xia , Haibo Li, Yaqun Liu, Chong Yu
State Key Laboratory of Geomechanics and Geotechical Engineering, Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan 430071, China
A R T I C L E I N F O A B S T R A C T
Keywords: The cavity effect means the amplifications of blast vibrations when propagating through a cavity inside the rock
Cavity effect mass. It has been investigated by a series of blast tests in an underground water tunnel. Monitoring points were
Tunnel set up in pairs on the ground surface and symmetrically about the tunnel face. The amplifying coefficient of blast
Blast vibration vibrations was defined as the PPV (peak particle velocity) ratio of the two symmetry points. It is found that the
PPV
amplifying coefficient first rises to a maximum of 3.5, and then falls down with the distance in the tunnel axial
direction. Meanwhile, tunnel radius and its buried depth are two other influencing factors. The coefficient
linearly increases with tunnel radius, while decreases with the depth in a power function. These results suggest
that the cavity effect only works within a limited scope on the ground, and mainly occurs in shallow tunnel
excavations. By introducing a ratio of the tunnel depth to its radius, an empirical formula was proposed to
calculate the amplifying coefficient. It can be used to predict the PPV amplifications and therefore define a
vibration control area for the safety of ground structures and facilities. It could also be applied to more site
conditions by choosing appropriate parameter values in the formula.
1. Introduction 2014; Zhang et al., 2000), respectively. However, few of them can deal
with the problems of underground tunnel blasts. In this case the ex-
Until now, the blast technique has still been the most common cavated section of the tunnel turns into a void space, which will change
method in tunnel excavations due to its low cost, high efficiency and the structural integrity of rock mass and thereby leads to more com-
easy operation. Equally remarkable are its side effects, such as the blast- plicated transmissions and reflections of blast waves (Zhang et al.,
induced damage and vibrations. Generally, damage only occurs in a 2005). Singh et al. (2015) monitored the blast vibrations at an open-pit
limited range around the explosive source, while vibrations have a mining and found that the PPVs recorded on the ground surface are
much wider range of influence. They will propagate not only along the much higher than those inside the underground opening at the same
tunnel axis, around its periphery, but also upward far to the ground. scaled distance. Same conclusions have been drawn by Blair (2014) and
The ground vibrations may have an adverse impact on the property Feldgun et al. (2014). Such a phenomenon may be attributed to the
safety and people’s health in the neighborhood, arising civilian com- following reasons. First, in comparison with external free-air explo-
plaints (Lee et al., 2016; Ozer, 2008). Their influences on the ground sions, tunnel blasts are performed in interior and confined conditions,
structures and sensitive devices are also a significant consideration in which will generate higher intensity of shock waves and in turn cause
the excavation site (Nateghi, 2011). Therefore, severe restrictions on more damage to nearby residential and industrial structures (Wu et al.,
blast vibrations have been implemented and the limits are becoming 2004b). Second, as stated above, the underground void space created
increasingly strict in recent years (Arora and Dey, 2010; Kim and Song, by the tunnel excavation plays an important role in the propagation of
2015). blast waves.
To exercise control over the blast vibrations, the prerequisite is to Moreover, there also exist interesting and important differences
investigate their propagations in rock mass and thereby make accurate among rock behaviors on the ground (Basarir et al., 2010; Tao et al.,
predictions of PPV (peak particle velocity) at varied locations. Many 2013). For example, Zhang et al. (2005) have conducted blast tests
PPV predictors have been proposed for level ground and undulate ter- inside a railway tunnel and measured the ground vibrations at several
rain conditions (Alvarez-Vigil et al., 2012; Khandelwal and Singh, points. These points were arranged symmetrically on both sides of the
2007; Kumar et al., 2016; Lu et al., 2010; Nateghi, 2011; Xia et al., tunnel face. It is found that there is a discrepancy between the PPVs of
⁎
Corresponding author.
E-mail address: xxia@whrsm.ac.cn (X. Xia).
http://dx.doi.org/10.1016/j.tust.2017.08.026
Received 14 October 2016; Received in revised form 8 August 2017; Accepted 16 August 2017
Available online 19 September 2017
0886-7798/ © 2017 Elsevier Ltd. All rights reserved.
X. Xia et al. Tunnelling and Underground Space Technology 71 (2018) 292–297
each pair of symmetry points. The PPV of the point behind the tunnel the working face is close to the point right beneath several ground
face (see Fig. 1) is generally greater than that of the symmetry point structures and facilities (Fig. 2). Their safety and normal operation are
ahead of the face. The ratio between the two PPV values can reach as the primary concern and the adverse effect of tunnel blast vibrations
high as 3.5 (Cai et al., 2015; Yang, 2012). It is the so-called cavity effect deserves great attention.
or hollow effect and this effect will act up to 40 m on the ground (Yang, Blast tests and vibration measurements have been conducted in
2012). Due to its action, the general attenuation law of ground vibra- order to study the propagations and amplifications of vibration waves
tions does not apply to the cases containing underground blasts. Little after they pass through a void space. 3–5 pairs of monitoring points
reports have been found on this topic, though it is of great interest to (MP) were placed on the ground surface along the longitudinal axis of
experts who are supposed to carry out safety and stability analysis and the water tunnel, each installed with a set of vibration sensor and data-
design of surface structures (Feldgun et al., 2014). logging instrument. Here a pair of MP refer to the two symmetry points
In the present work, the cavity effect of a water tunnel and its in- about the tunnel face, like the ones of B1 and A1, B2 and A2, and so on
fluencing factors have been investigated at the construction site of a in Fig. 1. It is obvious that one of the pair (B1, B2, B3) is located behind
nuclear power plant in China. The blast vibrations are magnified by this the tunnel face and the other one (A1, A2, A3), ahead of it. The velocity
effect when they reach the ground. The problem is that, several time histories of all MP were recorded in each blast test by the vibration
building structures are being and have been built on the ground surface, monitoring system. The amplifications of blast vibrations can thus be
inside which vibration sensitive instruments have been installed. analyzed by comparing the monitoring results of each pair of points.
Besides, there are still several tower cranes in operation just above the According to the research findings of Tao et al. (2013) and Zhang
tunnel (see Fig. 2). The blast vibrations may have adverse effects on et al. (2005), an amplifying coefficient of ground vibrations has been
these facilities. Their safety and performance are the primary concern. introduced here to describe the cavity effects of the water tunnel, which
Therefore, the ground vibrations must be controlled under a certain is given by
limit in the whole process of tunnel driving. For this purpose, it is es-
f = PPVb/PPVa (1)
sential to investigate the amplifying effect of the tunnel, based on which
an accurate prediction of blast vibrations can be made and a vibration where PPVb and PPVa are the PPV values of each pair of symmetry
control area is thus determined for the sake of ground structures. points behind and ahead of the tunnel face, respectively.
293
X. Xia et al. Tunnelling and Underground Space Technology 71 (2018) 292–297
10 Table 1
A1 Tunnel blast test and vibration monitoring results.
5 B1
Velocity (cm/s)
0 1 23.2 B3 17 3.36
0 0.5 1 1.5 2 B2 10 4.91
B1 5 6.02
-5
O 0 4.56
Time (s) A1 5 3.41
-10 A2 10 2.91
A3 17 2.29
(a) Measurement results
2 24.4 B3 15 4.15
B2 10 5.75
10 B1 7 7.32
A1 O 0 5.13
5 B1 A1 7 3.75
Velocity (cm/s)
A2 10 3.22
A3 15 2.29
0
3 23.2 B3 24 1.28
0 0.5 1 1.5 2
B2 18 3.43
-5 B1 7.5 5.96
Time (s) O 0 4.39
-10 A1 7.5 3.14
A2 18 2.7
(b) Numerical results A3 24 1.06
Fig. 3. Measurement and numerical results of the velocity time histories for a pair of 4 14.4 B3 21.5 0.86
monitoring points A1 and B1. B2 10 3.43
B1 3 3.98
O 0 2.67
It is further inferred therefrom that no amplifications will occur A1 3 2.8
(i.e., f decreases to near 1.0) when the MP is far enough from the tunnel A2 10 1.89
A3 21.5 0.76
face, say, D > 30 m, for the current case. This suggests that the cavity
effect only appears within a certain range on the ground and it is 5 20.6 B3 15 3.27
negligible beyond this range. B2 8 5.47
B1 4 6.16
O 0 4.66
4. Further study on the influencing factors of cavity effect A1 4 3.84
A2 8 2.83
A3 15 2.22
In underground tunnelling, the cross-section size and buried depth
6 14.4 B3 15 2.34
of the tunnel are the most important factors that will influence on the
B2 9 3.56
deformations, vibrations and stress distributions in the surrounding B1 5 4.12
rock mass (Qian et al., 2009; Singh, 2002; Singh et al., 2015; Tao et al., O 0 3.52
2013). Similarly, they will also have impact on the cavity effect, which A1 5 2.46
A2 9 1.84
has been investigated by the combination of field measurements and
A3 15 1.66
numerical simulations.
7 14.4 B3 25 1.2
B2 16 2.7
4.1. Numerical model setup and validation B1 8 4.24
O 0 2.48
A1 8 2.35
First, a three-dimensional model of the water tunnel was built using
A2 16 2.15
the FLAC3D program. It is an explicit finite-difference code and was A3 25 1.03
employed here to analyze the dynamic behaviors of rock mass induced
8 13.8 B3 28 0.85
by underground blasts (ITASCA Consulting Group, 2003). Since the B2 13 4.15
model is symmetric about the vertical plane across the tunnel axis, only B1 1 3.97
one half is considered in the calculation. As presented in Fig. 5, the O 0 3.26
dimensions are 180 m × 40 m × 60 m (width × length × height), re- A1 1 3.15
A2 13 2.18
presenting the sizes in the axial, radial and vertical directions, respec- A3 28 0.75
tively. It consists of approximately 351 000 elements and 372 600 nodes
9 14.4 B3 18 2.27
(gridpoints), with quiet (viscous) boundaries on lateral sides and free
B2 12 2.95
boundaries on the ground. The vibration MPs were set up according to B1 6 5.21
the arrangements of the field tests. O 0 3.51
For simplicity, the explosive load took the form of a stress wave A1 6 2.92
A2 12 1.69
function proposed by Starfield and Pugliese (1968) and Jong et al.
A3 18 1.47
(2005), given by
294
X. Xia et al. Tunnelling and Underground Space Technology 71 (2018) 292–297
depth.
Test results
Numerical results
Bigaussian Įƫng
1.5
Dcri
0.5
0 5 10 15 20 25 30
Distance, D (m)
Table 3
Physical and mechanical properties of the target granite.
Table 3. In Fig. 3(b) are shown the calculated results of velocity time
Fig. 5. Numerical model of the water tunnel with buried depth B = 15 m and radius histories of the MP A1 and B1. The PPVs of these two points are 4.12
R = 3.5 m.
and 6.58 cm/s, respectively, and therefore the amplifying coefficient is
worked out as 1.60. A comparison between the test and numerical re-
tunnel face as well as on the wall within the section of advance per sults is shown in Fig. 4. It validates the preciseness of the present model
round. by the rather good agreement and provides a basis for the further study
The peak value of the bore-hole pressure Pb can be estimated by on the influencing factors of cavity effect.
Henrych (1979) and Persson et al. (1993)
3 4.2. Influencing factors: the tunnel radius and buried depth
d
Pb = ρe De2 ⎛ e ⎞ /8 ⎜ ⎟
295
X. Xia et al. Tunnelling and Underground Space Technology 71 (2018) 292–297
Tunnel radius
a 1.5–3.0
α 0.8–1.1
Dcri (m) 7.0–10.0
2 w1 (1/3–1/2) Dcri
w2 (1.0–1.2) Dcri
3
than that of its symmetry point ahead of the face. An amplifying
coefficient thus has been introduced to quantify such an effect, which is
defined as the PPV ratio of the two symmetry points. Numerical models
have been built accordingly and they take account of varied tunnel
geometries. It’s concluded that the amplifying coefficient first rises to a
peak point of approximately 3.5, then falls down with the increasing
2 distance from tunnel face, indicating the cavity effect only works within
a limited scope on the ground, i.e., 30 m in this study. Yang (2012)’s
research reveals a greater action range of the cavity effect, i.e., 40 m,
due to a larger tunnel radius. A few other observations (Cai et al., 2015;
Zhang et al., 2005) show that this range is generally less than 30 m. The
discrepancy is caused by the neglect of tunnel size, depth, rock prop-
1 erties and so on. The safety of ground structures and facilities requires
0 5 10 15 20 special attention when underground blasts are performed within this
B/R range.
Both the tunnel radius and its buried depth have an influence on the
Fig. 7. The variations of maximum amplifying coefficient, fmax with the ratio of tunnel
buried depth to radius, B/R.
cavity effect. It linearly increases with tunnel radius, while decreases
with the buried depth in the form of a power function. When a tunnel is
located deeper than 60 m underground, nearly no PPV amplifications
It is the empirical formula used in the present construction site to will occur, suggesting the cavity effect can be safely ignored in this
predict the amplifying effect of ground vibrations. case. It is verified by Zhang et al. (2005)’s observations that the cavity
A general form of Eq. (8), which is given by effect only appears in shallow tunnels, generally referring to those less
2 than 30 m underground. The amplifying coefficient of the tunnel above
B −α −0.5 ⎛ D − Dcri ⎞ ⎧ w = w1 D ⩽ Dcri this level is probably greater than 1.5, according to the current study. In
f = 1 + a⎛ ⎞ e ⎝ w ⎠
R
⎝ ⎠ ⎩ w = w2 D > Dcri
⎨ (9) contrast, a maximum 50% of vibration amplification induced by deep
tunnel blasts is usually acceptable in engineering applications.
will probably apply to more site conditions, depending on the values of An empirical formula has been proposed to calculate the amplifying
parameters a, α, Dcri, w1 and w2. It reaches the peak point fmax at the coefficient of ground vibrations. It is used at the present site to predict
critical distance Dcri. More field tests and numerical researches in var- blast vibrations on the ground and thereby determine a control zone
ious site conditions are still needed to complete these parametric stu- where vibrations need special care. Moreover, the formula takes into
dies. However, a feasible range could be defined for some of them, since account the influences of the geometric features of the tunnel by in-
the amplifying effect will generally follow several rules. For example, it troducing a ratio of the buried depth to its radius B/R, therefore it is
only takes effect within a certain distance on the ground, say, 30 m at able to be applied to more site conditions by choosing appropriate
most from the tunnel face. Second, no cavity effect will appear if a constant values. These constants are determined by data-fitting method
tunnel is located below a certain level underground. It is 60 m, ac- based on blast test results, or by a rough estimate as recommended in
cording to the current study. Yang (2012)’s research shows that when the present work. The latter approach will bring about a substantial
the buried depth is greater than 50 m, fmax will drop down below 1.2. In saving in labor, time and cost, along with an acceptable accuracy.
addition, according to other similar observations (Yang, 2012; Zhang It should be stressed that only the amplifying effect in tunnel axial
et al., 2005), the maximum coefficient fmax will not be greater than 3.5 direction was discussed in the present study. The distribution of am-
in normal conditions. As to the critical distance Dcri, it will increase plifying coefficient on the entire ground surface still needs research. In
rapidly from approximate 5 m to 11 m with the buried depth B, then addition, on the basis of Feldgun et al. (2014), Liu (2009) and Wu et al.
remains unchanged afterwards. In fact, all these rules are the boundary (2004a)’s findings, there are still some other factors such as rock
conditions of Eq. (9). Possible ranges of the constant values can be properties, joint distribution, charge detonation manners and blasting
determined and are presented in Table 4, through which an accurate scheme, which will, to some extent, affect the amplifying effect of
prediction of blast vibrations is available for various tunnel sites. ground vibrations. Theoretically, the cavity effect will also have an
impact on the frequencies of the blast waves. For instance, Singh et al.
(2015) have concluded that the ground vibration frequencies are
296
X. Xia et al. Tunnelling and Underground Space Technology 71 (2018) 292–297
297