You are on page 1of 23

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/43485860

Effect of blast vibration on slope stability


Conference Paper November 1999
Source: OAI

CITATIONS

READS

684

5 authors, including:
Nenad Djordjevic

Peter Cepuritis

RAND Australia

Amec Foster Wheeler

30 PUBLICATIONS 274 CITATIONS

13 PUBLICATIONS 17 CITATIONS

SEE PROFILE

SEE PROFILE

All content following this page was uploaded by Peter Cepuritis on 29 September 2014.
The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.

Effect of Blast Vibration on Slope Stability


by
Nenad Djordjevic, Ian Brunton, Peter Cepuritis, Gideon Chitombo & Glen Heslop

Introduction
There is a significant amount of past evidence that links blasting and occurrences of
slope failure. Particularly of great concern at Telfer Gold Mine is the effect of open pit
blasting on the stability of the footwall slope above the underground portal and the
stability of underground structures. For this reason, Newcrest Mining Ltd. invited
JKMRC to conduct necessary field measurements and numerical modeling with the
following objectives:
1. assess the dynamic stability of the footwall slope above the portal;
2. assess the characteristics of current and future blasts in the open pit;
3. prepare open pit blasting guidelines that will minimize the detrimental effect of
blasting on slope stability.
To achieve these tasks it was necessary to develop a scaled distance attenuation curve
for blast vibrations. In order to validate the developed vibration attenuation curves it
was necessary to monitor vibrations from a production blast. Using JKMRC blast
design software tools, the vibration of the production blast was also modeled from the
blast design and vibration results from single hole blasts. The intention was, assuming
that the observed and modeled vibrations were close in terms of frequency and
amplitude, to use the software tools to model the vibrations of future blasts and assess
their likely effect on slope stability.
Numerical modeling of the dynamic stability of the footwall slope was done using the
FLAC finite difference code and the UDEC distinct element code. The results of the
numerical modeling and of the field measurements of the blast induced ground
vibrations were combined into the blast design rules that relate distance between the
slope and the blast, charge weight per delay and number of blasts necessary to cause
failure of the slope.
Development of the vibration attenuation curves
A vibration attenuation curve was developed by monitoring vibrations induced by
single hole blasts. In total, ten single hole blasts were fired and observed with tri-axial
geophones at six sites. The first set of blastholes was located at the north part of the pit,
while the second set of blastholes and observations was located in the south part of the
pit above the portal. Each monitoring site consisted of three Geosource SM-6b uniaxial
geophones mounted in an orthogonal configuration and bonded into a plastic sonde with
resin. Each sonde was lowered and grouted into a 2m deep borehole. Blast vibration
was recorded using a nine-channel Blastronics digital blast monitor. Each blast hole

EXPLO 99
1

was charged with ORICA Energan 2640, an explosive used regularly in production
blasting in the Telfer Open pit.
These results of vibration measurements are used to develop the vibration velocity and
acceleration scaled distance attenuation curves. Due to different geo-morphological and
geo-structural conditions, the vibration attenuation curves from the north and south part
of the pit differs significantly.
For the north part of the pit, the vibration attenuation curve can be represented by the
formula:
PPV (mm/s) = 137.67 (D/Q0.5)-1.188

(1)

correlation coefficient 0.981


PPV
D
Q

Peak particle velocity (mm/s)


Shortest distance through the ground (m)
Charge weight of Energan 2640 (kg)

The vibration acceleration attenuation curve expressed in units of gravitational


acceleration (g) is given by the formula:
PPA (g)= 9.115 (D/Q0.5)-1.747

(2)

correlation coefficient R = 0.986


In the southern part of the pit, the vibration velocity attenuation curve is characterized
by the formula:
PPV = 92.52 (D/Q0.5)-1.713

(3)

correlation coefficient R= 0.83


while the peak acceleration attenuation curve can be described by the equation:
PPA(g) =71.62 (D/Q0.5)-1.756

(4)

correlation coefficient R= 0.84


where
PPV
PPA(g)
D
Q

Peak particle velocity (mm/s)


Peak particle acceleration (gravitational acceleration)
Distance (m)
Quantity of Energan 2640 (kg)

The established vibration attenuation curves are presented in Figures 1 and 2. The
attenuation curve developed for the north part of the pit was used to provide sitedependent attenuation constants as inputs for the modeling of production blast induced
vibrations. The purpose is to validate the blast vibration modeling technique used by the
JKMRC. If the comparison between modeled and observed production blast vibration

EXPLO 99
2

shows reasonable agreement, then this modeling approach and the vibration attenuation
constants could be used for back analysis of blast induced failure in this part of the
footwall slope and to establish the relationships of the extent and nature of the peak
particle velocity, acceleration and vibration frequency produced by a particular blast
and their influence on slope stability.

Figure 1: Blast Vibration Footwall Slope North

Figure 2: Blast Vibration Footwall Slope North

EXPLO 99
3

Modeling of the production blast induced ground vibration


Modeling of the ground vibration induced by the production blast was done using the
3x3o-PRO blast design editor and VibPac, a software tool for analysis and modeling of
blast induced vibrations. The first production blast design was entered into 3x3o-PRO,
which then created an initiation file used for input into VibPac. The model uses
algebraic addition of vibration waveforms, for each blast hole in the production blast,
shifted by appropriate delay and travel times, to build up the total blast waveform. The
validity of this approach relies on the principle of superposition, where the effect of a
large multi-hole blast may be represented as the sum of simpler independent events.
Modeling of the production blast induced ground vibration was done using single hole
blast vibrations and the vibration scaled distance attenuation curve, developed along the
north profile. The vibration attenuation curve for the north profile is given by the
equation-1. The results of blast vibration modeling are presented in Figure 3. Observed
blast vibrations are presented in Figure 4. In Table 1, are presented parameters of the
observed and modeled production blast vibration.

Table 1
Vibration Component

Observed (mm/s)

Modeled (mm/s)

Radial

46.6

37.4

Transversal

69.8

76.5

Vertical

39.1

56.4

Vector Sum

86.1

92.6

The results of production blast vibration modeling are in good agreement with the
observed blast induced vibrations. In particular, the horizontal components of modeled
ground vibrations closely mimic the observed vibrations in terms of amplitude,
dominant frequency, form and duration.

EXPLO 99
4

Figure 3: Results of Blast Vibration Modelling

EXPLO 99
5

Figure 4: Observed Blast Induced Vibration

EXPLO 99
6

Development of criteria for dynamic stability of footwall slope


The stability of the slope under dynamic load can be evaluated in several ways. The
simplest approach is the pseudo-static approach. In this approach the dynamic
acceleration (earthquake or blast generated) acting on the mass of a potential landslide
is treated as a permanent static body force in a classic limit-equilibrium analysis.
Different dynamic load accelerations are applied until the factor of safety is reduced to
1.0.
This procedure is relatively simple and requires no more information than is needed for
a static factor-of-safety analysis. This approach is useful for identifying the yield
acceleration and hence peak ground acceleration, below which a slopes permanent
displacement will not occur. In the case where ground acceleration exceeds the yield
acceleration, a pseudo-static analysis tends to be very conservative, because slopes can
be exposed to a transient ground acceleration significantly above their yield
acceleration, with little or no permanent displacement.
It is important to highlight that pseudo-static methods consider seismic force acts in one
direction for an infinite time, when in reality inertia forces due to seismic loading are
reversible after a period of 0.025-0.1sec, and so the factor of safety can drop below
unity a number of times, but only for a short period of time. During such periods
permanent displacement will occur. The magnitude of permanent displacement depends
on the duration, intensity and frequency of the seismic motion.
Critical or yield acceleration (pseudo-static) can be calculated using the following
formula (Newmark, 1965):
(5)

Ac = (FS - 1)g sin


where:
Ac
FS
g

critical yield acceleration (in units of gravitational acceleration)


static factor of safety
acceleration of gravity
thrust angle

The thrust angle is defined as the angle from the horizontal in the direction that the
center of mass of the potential landslide block first moves. For a planar slip surface
parallel to the slope (which is the case similar to the Telfer slope above the portal), this
angle is the slope angle. For simple planar block sliding , the thrust angle is the
inclination of the shear surface. Applying the above equation to the slope above the
portal where the static factor of safety is FOS = 1.2 (Dames & Moore, 1990), with the
average depth of the shear surface of = 32, the critical or yield acceleration is:
Ac= (1.2-1) g sin(32)
= 0.105g

EXPLO 99
7

or

Ac = 1.04 m/s/s

This is a relatively low level of critical acceleration, which indicates that the slope is
close to being unstable. According to the classification proposed by Legg et al. (1982),
a slope which is characterized by a critical acceleration Ac = 0.105g is on the border
line between moderate and low level of stability. The previously used blast design
criterion for limiting blast induced wall damage was an acceleration of 2g (19.62 m/s/s).
This criterion is based on pseudo-static analysis of footwall stability and does not
consider the transient nature of the blast vibrations and their frequency. In order to
more completely address this problem, we analyzed the stability of the footwall slope
above the portal (at the catch fence), under dynamic load.
The load was prescribed at the base of slope in the form of a sinusoidal vibration
velocity signal whose amplitude and frequencies were chosen to mimic the typical
observed characteristics of blast induced vibrations. We varied the vibration frequency
and amplitude and the observed modeled deformation of the section of the footwall
slope above the portal. This modeling was done using the dynamic version of the
FLAC finite difference code and the UDEC distinct element code. A cross section of
the investigated part of the slope is given in Figure 5.

Figure 5: Slope above South Portal catch fence


Results of numerical modeling of slope stability using FLAC and UDEC
Based on the previous investigation by Dames & Moore, the critical parameter in the
footwall slope stability is the existence and properties of the soft siltstone layers. The
properties of sandstone are much higher than siltstone, so any instability will be
initiated by failure of the siltstone. We modeled a simplified slope, represented by two

EXPLO 99
8

layers of massive sandstone with one layer of siltstone between. Sandstone was
modeled as the elastic material, while the mechanical behavior of siltstone was modeled
using an elasto-plastic Mohr-Coulomb constitutive model. Mechanical properties of the
sandstone and siltstone used in modeling are given in the Table 2.
Table 2
Mechanical property

Sandstone

Siltstone/Silstone Interface

Density (kg/m^3)

2170

1960

Bulk modulus (GPa)

17

5.5

Shear modulus (GPa)

12

4.1

Cohesion (Kpa)

98

Tensile strength (Kpa)

390/300

Normal Stiffness (GPa/m)

Shear Stiffness (GPa/m)

Friction angle

14

Modeling was done in two phases, static and dynamic. First, we modeled the static
behavior of the slope. The static phase of modeling resulted in a stable slope. The
second phase consisted of the application of the horizontal vibration velocity function at
the base of slope and observation of the deformation and accumulated displacement of
the siltstone layer. The duration of the sinusoidal loading was initially chosen to be
0.1 sec.
The significance of the accumulated displacement caused by yielding of the siltstone
must be judged by its probable effect on the slope. The mere occurrence of yielding and
displacement, without regard to its magnitude, does not imply failure. Wierczorek et al.
(1985) used 5 cm as the critical displacement leading to macroscopic failure of
landslides in San Mateo County, California. Jibson and Keefer (1993) used 5 to 10 cm
displacement as the criterion for failure of landslides in the Mississippi Valley. It is a
question of engineering judgment to decide what kind of blast induced accumulated
displacement will cause macro failure of the slope above the portal.
Legg at al. (1982) proposed the classification of damage induced by dynamically
induced slope movement as a function of permanent displacement. Their classification
of slope failure state stipulates that when permanent displacement of the slope exceeds
5 cm, heavy failure of the slope tends to occur with landslide or slope displacement.
Their slope failure state scale is reproduced in Table 3.

EXPLO 99
9

Table 3: Slope failure state scale (After Legg et al.)


Failure State

Damage

Permanent
Displacement (cm)

light

ground shaking only effect

D < 0.5

moderate

small cracks likely to form

0.5 < D < 5

heavy

major ground failure,


slope displacement likely

5 < D < 50

severe

extreme ground failure,


large slope displacements

50 < D < 500

catastrophic

total failure,
slopes moves large distance

D > 500

We assumed, in the first approximation, a value of 5 cm as an indication of macro-scale


failure of the slope. Using this value of accumulated blast induced displacement it is
possible then to deduce the number of blast-induced ground vibration velocity
oscillations, of a given frequency and peak amplitude, that will cause macroscopic
failure of the portal slope.
This result indicates that the threshold of macroscopic failure is above the acceleration
level of 0.64 g, which is in apparent contradiction with the result of the calculation
using the Newmark equation (0.1 g). The Newmark equation predicts the acceleration
level that will cause the factor of safety to drop to 1, that is, the onset of failure. This
equation does not predict the extent of slope failure. Our results predict a level of
ground acceleration that will cause continuous yielding along the failure surface,
resulting in some permanent deformation of the slope.
For ground acceleration levels with an intensity of 0.64 g or less, there is localized
yielding at the base of slope, which does not propagate further and does not cause
permanent deformation of the slope. From the results of in situ measurements, under
most conservative estimates, horizontal ground acceleration will reach a level of 0.6 g at
scaled distances less than 5. Results of numerical modeling show that initiation of the
localized yielding of the slope occurs at accelerations close to 0.1 g, as predicted with
the Newmark equation.
In order to properly assess the influence of blasting on the slope stability it is necessary
to take into account the frequency of the ground vibration. This frequency is affected
by distance, in the sense that the principal frequency of the vibration pulse decreases
with distance. Another factor that influences frequency of the blast vibration is the
initiation sequence. This gives us the opportunity to alter the frequency of the
production blast induced vibration by changing the initiation sequence in the blast. The
frequency spectrum of the production blast induced vibration can be represented as the

EXPLO 99
10

product of the spectrum of the single hole blast induced vibration and the spectrum of
the blast initiation sequence.
We also noticed an apparently abnormal behavior of the slope when exposed to
horizontal vibration of constant amplitude (100 mm/s) but with different frequencies:
25, 35, 45, 50, 65 and 100 Hz. When we plotted the accumulated permanent
displacement of the slope per cycle of external vibration as a function of frequency of
external load, there is a strong peak at 45 Hz (1.85 mm/cycle), Figure 6. This can be
explained by the resonant frequency of the section of slope. Part of the slope between
the siltstone layer and the free surface is characterized by strong acoustic boundaries
that inhibit loss of seismic energy, leading to resonance. Resonant frequency is
determined by linear dimension and propagation velocity of the seismic wave. Very
significantly, the amplitude of the accumulated displacement decreases with a decrease
of frequency, and also, above 65 Hz, decreases with an increase of frequency of the
external vibration velocity function. This finding has significant implications for blast
design.

Figure 6: Induced Displacement of Slope Per Cycle of Loading (vel=0.1m/s)


UDEC is a distinct element code ideally suited for modeling deformation of
discontinuous media such as the rock mass in the footwall slope of the Telfer Open Pit.
Deformation of the siltstone interface (bedding plane), was modeled using standard
Coulomb linear elastic-perfectly plastic constitutive model.
Results of the UDEC modeling are presented in the form of horizontal vibration
velocity in the middle of the slope just below the Siltstone interface, displacement time
history at the base of the slope, and cumulative vectors of slope displacement. The
displacement value of 5cm was confirmed by the UDEC dynamic analysis of slope
stability. UDEC results show that when accumulated displacement reaches a value of
about 6.5cm there is sudden jump in the displacement of the slope and failure, Figure 7.

EXPLO 99
11

Figure 7: Displacement Vectors of the Portal Slop

Engineering Guidelines Regarding the Effect of Blasting on Slope Stability

EXPLO 99
12

Results of the in situ measurements of blast induced ground acceleration and results of
the numerical modeling were combined in the form of engineering guidelines. These
connect blasting parameters such as charge weight per delay and distance from the
footwall slope with the number of typical production blasts, at each particular distance,
that will create an accumulated displacement of 50 mm. First, we established a
relationship between the intensity of ground vibration acceleration (velocity) and
duration of the dynamic loading, required to generate an accumulated displacement of
50 mm. This relationship was established using the results of numerical modeling.
Second, the established relationship was then used with the observed attenuation curve
of ground vibration velocity to develop another relationship between distance, charge
weight per delay and number of blasts that will cause slope failure. The relationship
between duration of dynamic loading required to create an accumulated displacement of
50 mm and the intensity of horizontal ground acceleration (calculated from vibration
velocity and frequency) is presented in Figure 8.

Figure 8: Duration of the slope loading that will induce failure


This diagram, which describes the relationship between the average acceleration at the
base of siltstone layer and the duration of shaking that will cause failure, can be
described by the equation:
Duration = 14.597 (PPA)-1.3497

(6)

Coefficient of correlation R=0.98.


where

EXPLO 99
13

Duration
PPA

duration of dynamic loading (sec)


peak amplitude of horizontal ground acceleration (g)

This equation is applicable when the horizontal ground acceleration exceeds a value of
0.64g. For accelerations of less than 0.6 g there will be no permanent displacement
along the entire slope, although some localized yielding in the base of slope is possible.
According to the results of measurements, an acceleration of 0.6 g will not occur at
scaled distances greater than 5. The previous equation can be combined with horizontal
acceleration scaled distance curves, established by in situ measurements along the
southern profile, and which can be described by the equation:
PPAg= 1.18 (Dist/Q0.5)-1.1316

(7)

or by the upper limit line which is parallel with the best-fit curve but passes just above
the point of the largest measured acceleration. This most conservative line, in terms of
slope stability, can be described by the equation:
PPAg = 2.81 (dist/Q0.5)-1.1316
PPA
(g)
Dist
Q

(8)

mean value of horizontal (radial and transversal) ground acceleration


distance (m)
charge weight of Energan 2640per delay (kg)

For scaled distances larger than 3, the upper limit attenuation curve from the southern
profile was used. For small scaled distances, the acceleration attenuation curves can be
described by the equation:
PPA = 7.845(dist/Q0.5)-2.056

(9)

coefficient of correlation R=0.99


These three equations (6, 8 and 9), can be combined to produce a relationship between
duration of the dynamic loading that will cause macro failure, distance from the blast
and charge weight per delay.
For scaled distances: 5 > (dist/Q0.5) > 3
Duration = 3.62 (dist/Q0.5)1.527

(10)

For scaled distances (dist/Q0.5) < 3


Duration = 0.905 (dist./Q0.5)2.775

(11)

Considering that a typical production blast is characterized by a relatively strong


vibration amplitude that lasts about 2 sec, we can convert the above equation into an
equation that relates the number of typical production blasts that will cause failure,
distance and charge weight per delay:
N = 1.81 (dist/Q0.5)1.527

(12)

EXPLO 99
14

for scaled distances greater than 3 and less than 5, Figure 9,


and
N = 0.452 (dist./Q0.5)2.775

(13)

for scaled distances less than 3, Figure 10.

Figure 9: Number of blasts required to induce slope failure

EXPLO 99
15

Figure 10: Number of blasts required to induce slope failure


These relationships were developed using single hole blast vibrations that were not
affected by the existence of the broken ground, that could be created by the blast holes
fired before in the case of large multi-hole blast. Hence, in their application it is
important to utilize distance between blast holes and slope that goes through the
unbroken ground.
As a criterion for initiation of yielding within the slope we can use a value of horizontal
ground acceleration of 0.1 g, derived from the Newmark equation, and confirmed by the
numerical modeling. Using the upper limit line from the acceleration scaled distance
diagram (Equation 7), it is possible to calculate the scaled distance at which localized
yielding of the slope may occur.
0.1 = 2.81 (dist/Q0.5)-1.1316
or

dist/Q0.5 = 19

This means that at a scaled distance of 19 or less localized yielding of the slope may
occur. As discussed previously this is not an indication of slope failure or permanent
displacement. At best we can interpret this as a threshold of damage.
Blast design modifications
The current blast design used in the open pit blasting is a symmetrical V2 pattern with
42 ms inter-row delay and effectively 25 ms inter-hole delay achieved by increase of
down-hole delays, in 25 ms intervals, in each hole within a row. All blast holes are
charged essentially in the same way and burden and spacing are essentially constant
within each blast.

EXPLO 99
16

Optimization of the blasting practice can be achieved by adjusting initiation sequence to


exploit existence of the minimum in the frequency spectrum of the single hole blast. As
we discussed previously multi-hole blast induced ground vibration could be represented
as the result of interaction of two time-sequence; one is ground vibration induced by
individual blast hole, and second is time-sequence of blast holes initiation, where each
blast hole is represented with spike of constant amplitude separated in time by effective
delay times (sum of pyrotechnic delay time and travel time of the seismic wave from the
blast hole to the point of concern).
This relationship can be represented in the frequency domain, where the resultant
spectrum of the multi-hole blast induced ground vibration can be represented as the
result of multiplication of the spectrum of the single-hole blast induced vibration and
spectrum of the blast initiation sequence. Symmetrical operation in the time-domain is
given in the form of convolution of the two time-functions (Djordjevic 1995,1996).
An important feature of the spectrum of the single hole blast vibration is the existence
of the spectral holes, or regions in the spectrum below a cut-off frequency at which the
spectral amplitude locally decreases to a very small value. Their existence gives us
opportunity to adjust blast initiation sequence in such way that spectral amplitude of the
initiation sequence falls in the frequency region where single hole blast vibration has
very small value. It is possible to have a very large production blast, that will exploit
characteristics of the single hole blast vibration resulting in vibration with very low
amplitude.
Current blast design is close to the optimal in terms of initiation sequence. This is
confirmed by the vibration and spectrum of the production blast vibration, Figures 1112. The spectrum of production blast ground vibration is characterized with a peak at
about 16Hz. This spectral peak is very sharp, which is an indication that it comes from a
large number of vibration oscillations of moderate amplitude. In the case of one large
vibration pulse, the spectral peak would be relatively broad. The spectrum of the single
hole blast vibrations is characterised by a local minimum at about 16Hz. Proposed
changes in the initiation sequence will not alter this pattern, except that it will reduce
the number of holes initiated instantaneously.
Current blast design causes at least two blast holes, located at a relatively short distance
from each other, to be initiated simultaneously. This increases the charge weight per
delay causing an increase in the amplitude of the induced ground vibration. A better
result in terms of slope stability will be achieved by reducing the number of blast holes
initiated at each instant of time. This can be achieved by having a different inter-row
delay at one side of the blasts centerline. At the moment, both sides of the blast are
initiated after 42 ms from the center hole.
We propose that the side of the blast further from the slope be initiated 64 ms after the
center hole, while the first hole on the side of blast closer to the slope should remain
with 42 ms delay time. All other delays should remain as presently used. Such an
initiation sequence will reduce the effective charge weight per delay Figure 13, due to
increase in differential travel times of the seismic waves emanating from individual
blast holes, and consequently reduce the intensity of vibration hitting the slope.

EXPLO 99
17

From the results of production blast monitoring, Figure 11, we can notice that the
highest vibration amplitude originates from the blast hole closest to the slope. To
reduce the intensity of vibration even further we propose that the charge weight per
blast hole, located along two geometrical rows closest to the slope, be reduced by at
least 30%. These modifications should be conducted in conjunction with a similar
reduction of burden and spacing, in order to maintain satisfactory fragmentation.
These two modifications, initiation of the half of the blast further from the slope shifted
after the first half of the blast, and reduction of the charge weight per blast hole along
the two lines of holes closest to the slope, will significantly reduce the intensity of the
ground vibration coming into the slope and hence increase stability of the slope. These
measures are not likely to significantly change current fragmentation and muck pile
shape.

Figure 11: Radial vibration velocity and its spectrum


induced by the production blast

EXPLO 99
18

Figure 12: Single hole blast No. 5 vibration velocity and its spectrum

EXPLO 99
19

Figure 13: Charge weight detonated vs time distribution

Conclusions
Amplitudes of blast induced footwall vibration are determined by the charge weight of
explosive and distance and can be best described with scaled distance attenuation
curves. These curves were also influenced by the geological structure and differ from
site to site.
Dynamic numerical modeling of slope stability under the influence of blasting was
conducted with the FLAC finite difference code and the UDEC distinct element code.
Modeling was restricted to the portals slope above the catch fence. Results of
numerical modeling using both methods are in general agreement. Critical acceleration
above which slope deformation starts to occur is determined, by both numerical
methods, to be in the range 0.64-0.8g. This is much higher value than predicted utilizing
Newmarks equation (0.1g), that is based on the pseudo-static approach to modeling of
slope stability under dynamic load.

EXPLO 99
20

Results of modeling demonstrated that there is resonant frequency of the investigated


slope. In the frequency range 40-45Hz, there is amplification of slope deformation for
constant input motion. From the point of view of blasting, this will mean that an
initiation sequence that will amplify these frequencies will be unsuitable from the point
of view of stability of the slope above the portals catch fence. Blast initiation design
that will result with effective delay times in the range 20-25ms will strongly amplify the
intensity of slope vibrations.
Using 5cm of permanent displacement as the criterion for slopes macro-failure,
confirmed with results of numerical modeling, we combined results of in situ
measurements of ground vibration and results of numerical modeling to produce blast
engineering guidelines that link blast design parameters, (distance from the slope and
charge weight per delay) with the number of blasts likely required to initiate slope
failure.
For scaled distances (distance/charge weight per delay0.5), greater than 5, the likelihood
of blast induced failure is negligible. For scaled distances less than 5 and greater than 3,
the number of blasts required to induce failure is given by the equation:
N = 1.81 (dist/Q0.5)1.527
For scaled distances less than 3, the number of blasts required to induce failure is
determined by the equation:
N = 0.452 (dist./Q0.5)2.775.
Localized damage to the slope starts for blasts at scaled distances 19 or less.
In terms of the initiation sequence, current production blast design is close to optimal.
However, in terms of slope stability, improvements can be achieved by having larger
inter-row delay time on the side of the blast, further from the slope. Currently
production blasts are initiated with a V2 pattern where both sides of the blast,
symmetrical to the center line, are initiated with the same inter-row delay (42ms).
Increasing inter-row delay to 64ms on the side further from the slope will not change
that two blast holes are initiated at the same pyrotechnic delay time, however due to
increasing distance between such two holes, effective delay time seen from the slope
face, will increase due to longer travel time of the seismic wave.
The largest amplitude of the seismic waves at the slope comes from the closest blast
holes. We recommend to decrease charge weight along two lines of blastholes closest to
the pitwall, by about 20-30%. This should be coordinated with similar reduction in
spacing, to maintain same fragmentation.
Further reduction in the intensity of the blast induced slope vibration can be achieved
by introducing a hole-by-hole firing sequence using surface Nonel with constant downhole delays. changing initiation system. Ideally initiation should start from the blast
hole closest to the slope, with effective delay time between initiation of successive blast
holes of about 60-70ms. Relatively short down-hole delays (say 300ms), will reduce the
severity of delay scatter, and hence minimize probability of out-of-sequence initiation
and generation of large amplitude ground vibration.

EXPLO 99
21

References
Djordjevic, N.,A. Kavetsky, A.Scott, (1990), Blast Design Optimization to Minimize
Induced Vibrations of Structures. Proc. of the Third International Symposium on Rock
Fragmentation by Blasting, Brisbane Australia.
Djordjevic, N. (1996). Good vibrations-control blastings impact-a new approach.
Quarry Australia. Vol.4 No.3, March 1996.
M.R. Legg, R.T. Eguchi, J.H. Wiggins and J.E. Slosson, Seismic hazard models for the
city of Oakland ,Technical Report, J. H. Wiggins Co., 1982.
Newmark, N.M. Effects of Earthquakes on Dams and Embankments. Geotechnique,
Vol. 15., No.2, 1965, pp. 139-159.
Wieczorek, G.F.,R.C. Wilson and E.L.Harp. Map Showing Slope Stability During
Earthquakes in San Mateo County, California. Miscellaneous Investigations Map I1257-E. U.S. Geological Survey, 1985.
Jibson, R.W. and D.K. Keefer. Analysis of the Seismic Origin of Landslides: Examples
from the New Madrid Seismic Zone. Geological Society of America Bulletin, Vol. 105,
No.4, 1993, pp. 521-536.
Dames and Moore, Report. Pit1A Footwall Slope, Geotechnical Investigations. Dames
and Moore, 1991.
Wilson R.C., D.K. Keefer. Dynamic Analysis of a Slope Failure from the 6 August
1979 Coyote Lake, California, Earthquake. Bulletin of the Seismological Society of
America, Vol. 73, No. 3, pp. 863-877.

EXPLO 99
22
View publication stats

You might also like