Settlements and Damage Caused by Construction Induced Vibrat
Settlements and Damage Caused by Construction Induced Vibrat
Settlements and Damage Caused by Construction Induced Vibrat
K. Rainer Massarsch
Geo Engineering AB, Bromma, Sweden
ABSTRACT: Construction activities can affect structures in different ways. Damage can occur
as a result of distortion of the ground support and associated differential settlements. Ground
vibrations can also cause settlements in non-cohesive soils as a result of cyclic loading.
Dynamic and cyclic laboratory tests indicated that soil stiffness and strength decrease when a
critical strain level is exceeded. The magnitude of settlement depends primarily on shear strain
and number of significant vibration cycles. The relationship between strain (vertical and shear)
and ground vibration velocity has been established, based on extensive laboratory studies
performed for earthquake analysis. The results show that it is possible to estimate critical shear
strain levels at which loss of soil stiffness and strength can occur. A method for quantitative
determination of settlements due to surface vibrations is presented.
1 INTRODUCTION
Most types of construction activities, such as soil and rock excavation, driving of piles and sheet
piles or compaction work, generate vibrations in surrounding soil layers. Construction work is
often carried out in densely populated areas and close to sensitive structures and installations.
This aspect is enhanced by the increasing awareness of the public to environmental issues,
vibration-sensitive electronic equipment and machinery. Further, in many countries new
environmental regulations have been introduced and are stringently enforced. As a consequence,
authorities require frequently that a risk analysis be carried out prior to the start of construction
work. A risk analysis, when properly implemented, can be a complex task, where the impact of
many factors and the consequences on humans, buildings, and installations need to be
considered, Figure 1. Therefore, a risk analysis requires that the engineer has extensive practical
experience and a good theoretical knowledge in a number of technical disciplines, such as
vibration analysis, geotechnical and construction engineering etc. The risk analysis can have
important economic and technical consequences for a project. If unnecessarily conservative
assumptions are made, costs will increase. It may also, limit the choice of construction methods
and delay the project. On the other hand, if important factors are neglected in the risk analysis,
structures may be damaged or authorities stop or interrupt construction work.
Most risk analyses deal with the environmental effects of vibrations and noise on human
comfort, assessed primarily according to subjective criteria based on human perception. Codes
and regulations provide guidance for the measurement and interpretation of ground vibrations
with respect to environmental effects. Although such environmental consequences can be
important, they are small compared with damage to structures and installations in the ground. It
is therefore surprising that in most codes, limiting vibration values with respect to damage to
structures are empirical or expressed as “rules of thumb”, Studer and Süsstrunk (1981), Head
and Jardine (1992).
While such “rules of thumb” may be relevant for the geotechnical and structural conditions
in the particular area where they were developed, their general applicability is limited.
Therefore, such guidelines must be used with great caution. Clearly, representatives of
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 1
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
authorities as well as designer and constructors need a better understanding of the mechanism of
damage caused by ground vibrations.
Construction work can cause different types of damage to structures and installations in the
ground, some of which may not be related to, but often are attributed to ground vibrations.
Figure 2 identifies four different categories. Category I comprises soil movements, which are
due to “static” soil displacements, such as heave or lateral movements. In this case, structural
damage is primarily the result of soil displacement which result in differential settlement; the
mechanism is well-known and documented in the geotechnical literature. Soil heave occurs
usually in cohesive soils during installation of displacement piles (either by static or dynamic
installation methods), e. g. Massarsch and Broms (1989). Lateral soil movements are an
additional cause of structural damage, often due to excavations or slope instability (creep
movements). The extent of the problem may be aggravated by ground vibrations but damage is
primarily caused by static soil movements.
Problems belonging to Category II are less well known but have been discussed in the
geotechnical literature, for instance by Holmberg et al. (1984), Massarsch and Broms (1991),
and Massarsch (1993). Horizontally propagating vibrations cause a temporary distortion of the
surface layer to a depth corresponding to approximately one wave length. The magnitude of the
distortion depends on the wave length of the propagating wave (in most cases the surface wave)
and the vibration amplitude (displacement amplitude).
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 2
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
I. Differential settlements or heave II. Damage in structure
due to static soil movements due to ground distortion
The propagating waves expose buildings or installations in the ground to repeated distortion
cycles (“sagging” as well as “hogging”). This effect is fundamentally a cyclic loading problem
and not a “dynamic effect”. Distortion problems can also occur at very slow distortion rates, for
example in connection with tunnelling work or as a result of seasonal ground water variations
(swelling and shrinking of foundation soil). However, in connection with construction activities
(for instance soil compaction work or pile driving), the number of distortion cycles can be high.
The distortion problem can be analysed using a “static approach”.
Burland and Wroth ( 1974) have shown that “static damage” can occur in load-bearing walls
as a result of hogging at a relative deflection d/B > 1.5 10-4, where d is the vertical deflection
(displacement amplitude) and B is the building length. In the case of ground vibration
propagation, distortion is critical when the wave length L becomes shorter than the building
length B. This is the case when the surface wave propagation velocity is low, which is typical
for soft clays and silts below the ground water level. Based on an extensive literature survey,
Massarsch (1993) proposed a critical relative deflection d/B > 1.5 10-5. Assuming sinusoidal
wave motion, a simple relationship can be used to estimate a critical vibration velocity vcr.
where C is the wave propagation velocity (e. g. the surface wave velocity). This simple
relationship is valid if the wave length is smaller than twice the building length B. Massarsch
and Broms (1991) conclude that a large number of cases of vibration damage to structures can
be explained by ground distortion rather than “dynamic effects”.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 3
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
Category III covers permanent settlement (total and differential) and strength loss due to
cyclic loading. It resembles at first glance Category II but the problem is fundamentally
different. Permanent ground distortion due to settlement occurs mainly in granular soils (cf.
Category II, ground distortion creates temporary differential settlements) and is the consequence
of strain-softening soil behaviour. Again, settlement problems are primarily related to cyclic
loading effects (degradation of soil strength and stiffness).
Category IV considers the case, where damage is generally caused by “dynamic effects” due
to ground vibrations. This aspect is discussed extensively in the literature and will not be
addressed in this paper.
During the past three decades, as a result of extensive research in the areas of earthquake and
off-shore engineering, major progress has been made in understanding the static, cyclic and
dynamic stress-strain behaviour of soils. While these concepts are now generally accepted in
geotechnical earthquake engineering, they have not yet been applied to other types of vibration
problems. Therefore, as a first step in this paper, an introduction to the behaviour of soils
subjected to static, cyclic loading and dynamic loading will be presented. This concept can be
applied to the behaviour of soils subjected to ground vibrations caused by construction
activities. Solutions will be presented for assessing vibration problems quantitatively.
C = (G / ρ)0,5 (2)
where ρ is the total soil density. Using equation (2), the variation of shear modulus as a function
of strain level can be shown in Figure 3. It is apparent that also the shear wave velocity
decreases, in the present case from 205 m/s at 10-4 % to 120 m/s at 10-1 %., corresponding to a
decrease by 42 %. It is not generally recognised that also the shear wave velocity is affected by
strain level. The shear wave velocity at a given strain level can be normalised by its maximum
value and expressed as a shear wave velocity reduction factor, RC. Massarsch (1985) reported
results of resonant column tests on soil samples with different values of plasticity index, IP,
where the shear modulus is shown as a function of shear strain.
Figure 4 presents the results of these tests in a linear diagram, but shown as normalised shear
wave velocities versus shear strain. In the opinion of the author, a linear diagram reflects better
the engineering properties of soils, compared to the semi-logarithmic presentation, which is
normally used in earthquake engineering.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 4
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
80 220
Shear Modulus
70 200
Shear Wave Velocity
50 160
40 140
30 120
20 100
0,0001 0,001 0,01 0,1 1
Shear Strain ( %)
Figure 3. Result of resonant column test on medium dense sand, showing the variation of
shear modulus and shear wave velocity with shear strain (note shear wave velocity
right scale).
1,0
0,9
0,8
C
Velocity Reduction Factor, R
0,7
0,6
0,5
0,4
0,3
Gulf (IP: 74 %) Välen (IP: 67 %)
0,2 Skå-Edeby (IP: 45 %) Norrköping (IP:36 %)
Bäckebol (IP: 28 %) Drammen (IP: 18 %)
0,1
Kentucky (IP:12 %)
0,0
0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35 0,40 0,45 0,50
Shear Strain, %
Figure 4. Variation of normalised shear wave velocity with shear strain, determined from
resonant column tests, data from Massarsch (1985).
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 5
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
The most important conclusion from Figure 4 is that the plasticity index, IP has important
influence on the reduction of shear wave velocity. The shear wave velocity decreases more
rapidly in soils with low plasticity index. Therefore, sands experience a more pronounced
decrease of shear wave velocity even at small strain level than clays with high plasticity index.
The author has reviewed results from resonant column tests in different soil types and at varying
strain levels, as reported by Dörringer (1998). Based on a regression analysis of the data, the
modulus reduction factor is shown in Figure 5 for different shear strain values. The reduction of
shear wave velocity is most pronounced in soils with low plasticity index and increases with
shear strain level.
0,9
0,8
0,7
Reduction Factor, R C
0,6
0,5
0,4
0,10%
0,2
0,25%
0,1
0,50%
0
0 5 10 15 20 25 30 35 40 45 50
Plasticity Index IP, %
Figure 5. Shear wave velocity reduction factor as function of plasticity index and shear strain
level.
The shear wave velocity can be determined by field and laboratory methods. Hardin (1978)
proposed the following, widely used semi-empirical relationship for estimating the shear
modulus at small strains, Gmax for normally consolidated soils
where σ0´ is the effective confining pressure (σ1+σ2+σ3)/3, pa is a reference stress level (100
kPa) and e is the void ratio. Combining equation (3) and (2), the shear wave velocity at small
strain level, Cmax (elastic wave velocity) can be derived
Figure 6 shows the variation of the shear wave velocity as a function of vertical effective stress
for normally consolidated, dry non-cohesive soils with different void ratio. The calculation
assumes a K0-value of 0.57, which corresponds to a sand with a friction angle of 35 degrees.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 6
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
300
275
250
Shear Wave Velocity, m/s
225
200
175
150
Void Ratio, e
0,5 0,6
100
0,7 0,8
75
0,9 1
50
0 25 50 75 100 125 150 175 200 225 250
2
Vertical Effective Stress, kN/m
Figure 6. Variation of shear wave velocity with vertical effective stress in dry sand, cf.
equation (4).
It is thus possible to estimate, for most practical problems with sufficient accuracy, the
maximum shear wave velocity and the effect of wave velocity reduction as a function of shear
strain level. The above data suggest that the shear strain level γ is an important parameter, which
can not be neglected once it exceeds about 0.001 %. The shear strain level can be determined
from the following relationship if the vibration amplitude (particle velocity) v and the shear
wave velocity CS are known
γ = v / CS (5)
For example, if a shear wave velocity (medium dense sand) of 210 m/s, and a particle velocity
of 20 mm/s, are assumed, the shear strain level is about 0.01 % , cf. Figure 3.
One of the most important soil properties governing the behaviour of sand is density. When
subjected to cyclic loading, loose sands tend to densify until they have reached a “critical
density” or critical void ratio. On the other hand, dense sands dilate when subjected to repeated
cycles of loading until they approach approximately the same critical void ratio. The effect of
cyclic loading has been investigated extensively in the area of earthquake engineering, and
especially with respect to the pore water pressure build-up in loose saturated sands (liquefaction
potential). Design methods for assessing liquefaction by laboratory and field methods have been
published and are now state-of-practice. However, during the past three decades, only few
investigations have studied volume change in dry soils during cyclic loading. About thirty years
ago, during the same year three studies were published by eminent researchers, Brumund and
Leonards (1972), Seed and Silver (1972) and Youd (1972). These papers present important
information regarding the compaction effect during cyclic loading and also describe the
fundamental mechanism which causes soil densification. It is also important to point out that the
findings were obtained independently of each other but are in excellent agreement.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 7
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
Many investigators and practitioners have in the past attempted - and still attempt - to
correlate compaction behaviour of sands with stress fluctuations and with the values of
acceleration and frequency of vibration, associated with the compaction process. This approach
has not and cannot be substantiated by carefully performed laboratory tests but is still widely
used and referred to. The most important conclusions in the above three papers are summarised
below.
1. Fundamental concepts and published data show that shear strain is the primary factor
causing compaction of granular material.
2. Compaction (volume change) increases with shear strain amplitude.
3. The parameter that governs the ultimate residual compression is the steady-state
transmitted energy. This is valid for a wide range of frequencies. The residual
settlement cannot be correlated to acceleration.
4. Compaction is not significantly affected by vertical stress (for strain levels exceeding
0.05 %).
5. In the 10 cycles/min to 115 cycles/min (0.17 – 1.9 Hz) range, frequency of straining has
no significant effect on compaction behaviour.
6. No significant behavioural differences were detected between samples tested dry and
similar samples tested in a saturated, but completely drained, conditions.
7. Even at static loading conditions, evaluations of settlement is subject to considerable
error (+/- 25 – 50 %). For complex conditions associated with cyclic loading, it is
unrealistic to expect that evaluations could be made with even this degree of accuracy.
However, an approximate evaluation of possible settlement is adequate for many
purposes.
In order to draw general conclusions of the above test results, the published data have been
redrawn and are below presented in a uniform manner. Youd (1972) performed cyclic shear
tests (NGI-type apparatus) on samples of dry and water-saturated Ottawa sand. Tests were
performed at different effective stress levels and shear strain was measured at increasing
number of load cycles, Figure 7. The initial void ratio was approximately eo = 0.545 and the
relative density was Dr = 75 – 79%.
Normalised vertical strain ε is expressed as vertical strain normalised by strain at the first
loading cycle ε1 to demonstrate that the relative settlement (compared to compression after the
first loading cycle) is largest at low strain level. Settlement does increase almost linearly with
the number of vibration cycles (in a semi-logarithmic diagram). Youd also showed that
settlement (and thus relative compression) is independent of vertical effective stress and of
frequency (strain rate) at the same strain level. The above findings are important for practical
applications. In order to estimate vertical compression, it is necessary to determine three
parameters: strain level, number of loading cycles, and initial density.
Seed and Silver (1972) presented results from important tests, where they correlated shear
strain level to vertical compression. Cyclic load tests were performed on dry silica sand at two
different densities, using a shaking table. The vibration frequency was 4 Hz, the maximum
number of load cycles was 300 and the maximum acceleration was 0.3 g (corresponding to a
maximum vibration velocity of 0.7 m/s). The test data for loose sand are shown in Figure 8.
The ratio between vertical strain ε and shear strain γ can be calculated from elastic theory for
small strain levels (< 10-4 %) and depends on Poisson’s ratio, ν, according to the following
equation
ε = γ (1 + ν) (6)
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 8
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
Number of Cycles, N
1 10 100 1 000
0
4
ε / ε1
6
Normalised Vertical Strain,
10
12
20
Figure 8. Vertical settlement – shear strain relationship for silica sand, Seed and Silver (1972).
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 9
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
The data published by Youd (1972) and Seed and Silver (1972) have been re-analysed in order
to study the relationship between vertical strain and shear strain ε/γ as a function of shear strain
amplitude and for different numbers of load cycles, Figure 9. In the case of sand with a
Poisson’s ratio of 0.33 vertical strain ε should be about 1.33 γ, which is in good agreement with
the data in Figure 9 at one load cycle and small shear strain level. A shear strain factor f1 = ε/γ
is defined, which has been plotted against shear strain γ and as a function of number of loading
cycles. According to Figure 9, vertical compression increases with the number of load cycles
but decreases with shear strain level. Vertical compression is higher in the case of loose sand,
by a factor of about 50% when comparing 45 and 60% relative density.
100
N =300
Dr: 45%
N =10
Dr: 45% Youd (1972)
N =300
10 Dr: 60%
f1 = ε / γ
Load
Seed & Silver N =10
cycles
Dr: 60%
(1972)
N =2
Dr: 45%
100
N =2 0
1
Dr: 60%
30
Elastic
1
0,1
0,01 0,1 1 10
SHEAR STRAIN, %
Figure 9. Shear strain factor f1 as function of shear strain for different values of load cycles
and relative density, data Seed and Silver (1972) and Youd (1972).
The agreement between the two independently performed test series is good. Figure 9 can be
used to estimate vertical settlement in sand if the shear strain amplitude and the number of load
cycles are known. One practical problem, however, is to translate field vibration measurements
with irregular cycles into equivalent uniform vibration cycles. A commonly used method in
earthquake engineering is to estimate the number of equivalent vibration cycles by visual
inspection of the irregular time history.
Seed (1976) proposed a “liquefaction curve” from which the equivalent number of vibration
cycles can be determined for any given irregular sequence of vibration cycles. Also other more
sophisticated methods, such as the “cumulative damage approach” have been proposed but with
similar results. Since the settlement mechanism is closely related to the liquefaction problem, it
is proposed to use a similar approach when determining equivalent vibration cycles from an
irregular record of vibration cycles. Figure 10 shows a relationship between normalised
vibration cycles v/vmax and the normalised number of equivalent vibration cycles, N /Neq.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 10
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
1
0,9
0,8
0,7
0,6
V / Vmax
0,5
0,4
0,3
0,2
0,1
0
1 10 100
N / Neq
Figure 10. Determination of equivalent vibration cycles from vibration record with different
numbers and amplitudes of vibration cycles
From Figure 10 the equivalent number of vibration cycles of an irregular vibration record can be
determined. For example, one cycle of maximum vibration amplitude, vmax is then equivalent to
4,3 vibration cycles at amplitude 0.5 vmax or 10 cycles at 0.4 vmax. Experience from liquefaction
studies has shown that the shape of the correlation curve is not crucial for the outcome of an
analysis, provided that the approach is used consistently.
veq = vz mz = vx mx (7)
Figure 11 shows the variation of the shear strain factor mz with dimensionless depth z/L for use
with vertical peak particle velocity, where L is the wave length and ν is Poisson’s ratio. A
similar diagram has been developed for the case of horizontal peak particle velocity. The
influence of Poisson’s ratio is relatively small for non-cohesive soils (0.25 < ν < 0.33). It should
be noted that below about 1.5 wave lengths from the surface, the Rayleigh wave particle motion
becomes insignificant.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 11
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
mz = 0.9 - 0.6 z/L
Figure 11. Shear strain factor mz for use with vertical peak particle velocity, after Mohamed and
Dobry (1987) with indication of simplified relationship.
For most practical purposes it will be sufficient to use in non-cohesive soils a simplified, linear
relationship according to equation (8) when estimating the value of mz, cf. Figure 11
The wave length L can be calculated if the wave velocity and the vibration frequency f (or
period of vibration T = 1/f) are known
L=C/f (9)
The relationship given in equation (8) makes it possible to estimate the variation of shear strain
amplitude based on measurement of the vertical vibration amplitude at the ground surface. As a
first approximation, the shear strain factor for a homogeneous soil layer (vertical vibration
measurements) can be taken as mz = 0.5. This indicates that the equivalent vibration amplitude
in a soil layer corresponds to about half the vertical vibration amplitude measured on the ground
surface.
Clough and Chameau (1980) suggested that the effect of ground vibrations on settlement during
vibratory sheet pile driving in sands can be assessed qualitatively by the shear strain which is
generated due to wave propagation. Dobry et al. (1982) proposed that the susceptibility of
sands to liquefaction (liquefaction potential) due to earthquakes can be predicted by the “strain”
or “ stiffness” approach. It makes use of the shear wave velocity of the soil deposit to estimate
the maximum cyclic shear strains induced by seismic shaking and then compares these
maximum cyclic shear strains with the threshold strain of the soil, γt. The threshold strain is
defined as the value of cyclic shear strain such that the cyclic shear strains less than γt will not
cause any densification of dry granular soils, or any pore pressure build-up in water-saturated
granular soil. Mohamed and Dobry (1987) suggest that for most sands, the threshold strain is γt
∼ 10-2 %.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 12
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
Equation (5) makes it possible to estimate the shear strain level if the vibration velocity and
the shear wave velocity at a given strain level are known. The effect of shear strain level on
shear wave velocity can be taken into account, cf. Figure 5. The shear wave velocity in normally
consolidated soils can either be measured in the field or estimated using equation (4) or from
Figure 6. Based on the shear strain factor shown in Figure 11 it is possible to estimate shear
strain variation with depth from vibration measurements on the ground surface. Figure 12 shows
the relationship between vibration velocity (particle velocity) and shear wave velocity for two
different levels of shear strain. Based on extensive laboratory tests for different soils it can be
assumed that if the shear strain level of 0.001 is not exceeded, the risk of ground settlement or
strength loss is very low. However, if the shear strain level caused by ground vibrations exceeds
0.1 % there is significant risk of settlements or loss of shear strength in cohesive soils. Figure 12
does not include the effect of number of load cycles.
100,0
Shear Strain: 0,1 %
CLAY Risk for Settlements
IP = 50 or Shear Strength Reduction
SAND
IP = 0
Vibration Velocity, mm/s
1,0
CLAY Shear Strain: 0,001 %
IP = 50 Litttle Risk for Settlements
or Shear Strength Reduction
SAND
IP = 0
0,1
10 100 1000
Shear Wave Velocity, m/s
Figure 12. Estimation of risk for settlements or strength reduction from vibration velocity as
function of shear wave velocity for different levels of shear strain, cf. equation (7).
For example, if the shear wave velocity in a very loose sand close to the ground surface is about
100 m/s, it can be concluded from Figure 12 that settlements are not likely to occur if the
associated, equivalent vibration velocity veq is lower than 0.5 mm/s. However, the threshold
value, where there is a risk of settlement in loose sand is 6 mm/s. When the shear strain level
exceeds 0.1 % (at a vibratin velocity of approximately 50 mm/s) a significant risk of ground
settlement and loss of shear strength exists. Settlement increases with increasing number of
vibration cycles. However, the risk of settlement decreases with increasing shear wave velocity
and increasing plasticity index IP.
The information provided in the previous sections makes it possible to assess quantitatively the
magnitude of settlement in free-draining (non-cohesive) soil due to propagation of surface
waves if the thickness of the compressible layer and the vertical strain are known
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 13
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
∆s = ε ∆ H (10)
where ∆s is the settlement of a layer with thickness ∆H and ε is vertical stain. The shear strain
level γ at different depths can be estimated from ground vibration measurements at the ground
surface and taking into account the depth effect using the strain factor mz, cf. Figure 11. Ground
vibrations in the far-field from the vibration source (at distance of least 1.5 times the wave
length) are caused primarily by surface waves. The surface wave velocity in sand is only
slightly (approximately 7 %) lower than that of the shear wave. It can be determined from shear
wave velocity either measured or estimated from equation (3), and then be corrected with
respect to strain level. The shear wave velocity reduction for sand and silty sand is shown in
Figure 13. Indicated in the figure are the three shear strain levels suggested in Figure 12.
The effect of strain level on shear wave velocity must be determined by iteration, starting
with the small-strain shear wave velocity, then determining the wave reduction factor, Rc from
the calculated strain level. Already within two or three iteration cycles, a stable value of the
velocity reduction factor is obtained. The irregular number of vibration cycles can be translated
into an equivalent number of vibration cycles Neq from Figure 10. Finally, the effect of cyclic
loading (number of equivalent shear strain cycles) Neq on vertical strain can be assessed
quantitatively by introducing a shear strain factor, f1, cf. Figure 9. The process is expressed in
the following simple relationship
The calculation of settlements in sand is illustrated by the following example, involving a deep
layer of loose sand subjected to ground vibrations. It is assumed that the shear wave velocity in
the sand is 150 m/s, the vertical peak vibration amplitude at the ground surface (determined by
field measurements) is 15 mm/s and the dominant frequency of vibrations is 20 Hz. The
following vibration cycles are anticipated: two cycles at 10, four cycles at 7, four cycles at 6, six
cycles at 5 mm/s, respectively. The method of calculating the number of equivalent vibration
cycles, based on Figure 10 is shown in Table 1. The number of equivalent vibration cycles,
which corresponds to the peak vibration velocity is thus approximately 3,5.
The shear wave velocity reduction factor at an initial strain level of (0.015 %) can be estimated
from Figure 13, Rc ≈ 0.9. Assuming that the surface velocity is 7 % slower than the shear wave
velocity, a surface wave velocity is 130 m/s is obtained which results in a shear strain level γ ≈
1.2 10-2. The wave length L = 6.5 m and therefore, settlements are anticipated to occur down to
a depth corresponding to 9.5 m (1.5 L).
The soil layer can now be divided into, say, three layers (2.0 m + 3.0 m + 4.5 m) and
settlement can be estimated for each layer according to equation (11). The calculations are
shown in Table 2.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 14
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
1
0,9
0,8
C
0,7
Velocity Reduction Factor, R
0,6
0,5
0,4
Sand
0,3
Sitly Sand
0,2
Test data Medium
Dense Sand
0,1
0
0,0001 0,001 0,01 0,1 1
Shear Strain, %
Figure 13. Variation of shear wave velocity reduction factor for sands and silts, after Vucetic
and Dobry (1990).
The total settlement in the loose sand deposit caused by short-duration surface wave vibrations
with a vibration amplitude at the ground surface of 15 mm/s will be about 0.7 mm. It should be
noted however, that if the number of vibration cycles increases to, for example 300, then, the
settlement will increase by a factor of 20, resulting in settlements on the order of 6.5 mm. Also,
close to the vibration source, other wave types, such as shear and compression waves may
dominate, which can cause larger settlements.
8 SUMMARY
9 ACKNOWLEDGEMENT
Prof. Bengt H. Fellenius, Urkkada Technology Ltd. has reviewed the paper and made many
valuable suggestions for improvement. His support and encouragement is acknowledged with
gratitude.
REFERENCES
Brumund, W. F. and Leonards, G. A. 1972 Subsidence of sand due to surface vibration. Journal
of the Soil Mechanics and Foundation Division, Proceedings ASCE, Vol. 98, pp. 27 – 42.
Clough, G. W. and Chameau, J. 1980. Measured effects of vibratory sheet pile driving. Journal
of the Geotechnical Engineering Division, ASCE, Vol. 106, No. GT10, Oct., pp. 1080 - 1099.
Dobry, R. Ladd, R. S., Yokel, F. Y., Chung, R. M. and Powell, D. J. 1982. Prediction of pore
pressure buildup and liquefaction of sands during earthquakes by the cyclic strain method.
Building Science Series 138, U. S. Nat. Bureau of Standards, Washington, D. C, pp.
Döring, H. 1997. Verformungseigenschaften von bindigen Böden bei kleinen Deformationen
(Deformation properties of cohesive soils at small deformations). Royal Institute of Technology
(KTH), Stockholm, Examensarbete 97/8, 54 p.
Head, J. M. and Jardine, F. M. 1992. Ground-borne vibrations arising from piling. CIRA
Technical Note 142, 83 p.
Holmberg, R. et al. 1984. Vibrations generated by traffic and building construction activities.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 16
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.
Massarsch, K. R. and Broms, B. B. 1989. Soil Displacement Caused by Pile Driving in Clay.
International Conference on Piling and Deep Foundations, London, 15 - 18 May, 1989,
Proceedings, pp 275 - 282.
Massarsch, K. R., 1985. Stress-Strain Behaviour of Clays. 11th International Conference on Soil
Mechanics and Foundation Engineering, San Francisco, Proceedings, Volume 2, pp. 571 - 574.
Massarsch, K. R. and Broms, B. B., 1991. Damage Criteria for Small Amplitude Ground
Vibrations", Second International Conference on Recent Advances in Geotechnical Earthquake
Engineering and Soil Dynamics, St. Louis, Missouri, March 11 - 15, 1991, Vol. 2, pp 1451 -
1459.
Massarsch, K. R., 1993. Man-made Vibrations and Solutions, State-of-the-Art Lecture, Third
International Conference on Case Histories in Geotechnical Engineering, St. Louis, Missouri,
June 1 - 6, 1993, Vol. II, pp. 1393 - 1405.
Mohamed, R. and Dobry, R. 1987. Settlements of cohesionless soils due to pile driving.
Proceedings, 9th Southeast Asian Geotechnical Conference, Bangkok, Thailand, pp. 7-23 – 7-
30.
Seed, H. B. and Silver, M. L. 1972. Settlements of dry sand during earthquakes. Journal of the
Soil Mechanics and Foundation Division, Proceedings ASCE, Vol. 98, pp. 381 - 396.
Seed, H. B. 1976. Evaluation of soil liquefaction effects on level ground during earthquakes.
ASCE Annual Convention and Exposition, Liquefaction Problems in Geotechnical Engineering.
Philadelphia. pp. 1 – 104.
Studer, J. and Süsstrunk, A. 1981. Swiss Standard for Vibrational Damage to Buildings, X
International Conference on Soil Mechanics and Foundation Engineering, Stockholm, Vol. 3,
pp. 307 – 312.
Vucetic, M. and Dobry, R. 1990. Effect of soil plasticity on cyclic response. Journal of the
Geotechnical Engineering Division, ASCE Vol. 117, No. 1. Jan, pp. 89 – 107.
Youd, T. L. 1972. Compaction of sands by repeated shear straining. Journal of the Soil
Mechanics and Foundation Division, Proceedings ASCE, Vol. 98, pp. 709 – 725.
Massarsch, K. R., 2000. ”Settlements and damage caused by construction-induced vibrations”. Proceedings, Intern. Workshop 17
Wave 2000, Bochum, Germany 13 – 15 December 2000, pp. 299 – 315.