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

net/publication/353494408

Ground Vibrations Caused by Soil Compaction

Chapter · July 2021

DOI: 10.1201/9781003211372-4

CITATIONS READS

0 73

1 author:

K. Rainer Massarsch
Geo Risk & Vibration Scandinavia AB
67 PUBLICATIONS   403 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Ground treatment View project

Vibratory compaction of granular soils View project

All content following this page was uploaded by K. Rainer Massarsch on 05 August 2021.

The user has requested enhancement of the downloaded file.

Ground Vibrations Caused by Soil Compaction
K. Rainer Massarsch
Geo Engineering AB, Stockholm, Sweden

Massarsch, K. R., 2002. ”Ground Vibrations Caused by Soil Compaction”. Wave 2002. Proceedings, International Workshop,
Okayama, Japan, pp 25 - 37.

ABSTRACT: Presently used empirical methods do not account for many important aspects of ground vibra-
tions, for, instance the dynamic interaction between the energy source and the soil. Also the influence of dy-
namic soil properties on wave propagation is generally neglected. A concept is presented, which describes the
dynamic resistance of the compaction source. Two cases are discussed, the impact of a rigid plate on an elas-
tic half-space and the energy transfer to the soil from the shaft of a cylindrical compaction probe. Soil imped-
ance and rate of loading determine the dynamic soil response and thus ground vibrations. There is an upper
limit to the vibrations, which can be transmitted to the soil. Based on a closed-form solution it is possible to
assess the relationship between static and the dynamic soil resistance. In the vicinity of the compaction
source, shear wave velocity and thus soil impedance are strongly affected by strain level. A relationship is
proposed for estimating the reduction of shear wave velocity with shear strain. It is possible to estimate the
vibration intensity at the compaction source and to predict vibration propagation from the base and the shaft
of a compaction probe. The variation of force with time (loading rate) of an impacting plate, and thus the pre-
dominant vibration frequency can be estimated from time factors. Resonance effects can occur during com-
paction, which amplify ground vibrations. A case history is presented, which illustrates the compaction of a
deposit of sand using a falling weight. The results of extensive vibration measurements on the impacting plate
and in the soil are reported and compared with the proposed analytical solutions.

1 INTRODUCTION formation has helped to better understand the com-

paction process, as well as vibration propagation in
Compaction methods are used increasingly to im- the ground, (Massarsch, 1993 and Massarsch &
prove loose, granular soils to great depth more effi- Broms, 2001).
ciently and at lower cost. In the recent past, new, The main objectives of soil compaction are to
more powerful compaction machines and construc- eliminate total and differential settlements during
tion equipment have been developed, which make static loading, and to reduce the susceptibility of wa-
the densification process more efficient. However, ter-saturated sands to cyclic loading (liquefaction),
also the risk of excessive ground vibrations in- for instance during earthquakes.
creases. Electronic monitoring systems have been Soil compaction systems can be classified accord-
developed, which can control the compaction proc- ing to two aspects: the location of the point of en-
ess and monitor the performance of the equipment, ergy application (at ground surface or at depth) and
thereby aiding the machine operator during the dif- the type of applied energy (impact or vibration).
ferent phases of project execution. Sensors on the These considerations are important for ground vibra-
compaction machine an in the ground can be used to tions caused by soil compaction.
measure different parameters, such as compaction
frequency, power consumption and ground vibration
velocity, which are important for the compaction 2 EMPIRICAL RELATIONSHIPS
process, Massarsch (2002). This information can be
used to document and verify that the anticipated
compaction effect has been actually achieved. Moni-
2.1 Energy-based vibration prediction
toring provides also valuable information about the
performance of the equipment, and the effect of soil A commonly used concept to assess ground vibra-
compaction on the surrounding soil, Massarsch and tions from pile driving or soil compaction is based
Fellenius (2002). This increasing database of site in- on the energy of the compaction device. The follow-
ing empirical relationship has been proposed to pre- Perhaps the most important limitation is that
dict ground vibrations from pile driving, (Attwell Equation 1 does not consider the effect of ground
and Farmer, 1973) conditions on vibration propagation. It is not surpris-
ing that a large scatter is obtained, even when the
W measurement results are plotted in a double-
v=K (1)
D logarithmic diagram. Therefore, empirical relation-
ships as proposed in Figure 1 or Equation 1 should
where K is an empirical coefficient, W is the energy
be treated with caution and used only for site-
applied to the pile head (or compaction probe) and D
specific calibration of vibration measurements. Ex-
is the lateral distance from the pile measured on the trapolation and application of measurements to other
ground surface. A coefficient K = 0.5 is often as-
site conditions can be misleading.
sumed but its actual value depends on many factors
and can vary within wide limits (Massarsch, 1992).
Figure 1 illustrates the problems associated with 2.2 Effect of Compaction Depth and Distance
this simplified empirical approach. The compaction
Another aspect, which is important when assessing
energy actually applied to the soil is usually lower ground vibrations caused by soil compaction, is the
than the nominal energy of the compaction device.
depth of the vibration source, in relation to the dis-
However, in practice it is difficult to account for
tance to the point of vibration observations or meas-
these energy losses. Also, as will be shown below, urement, Figure 2.
the maximum energy, which can be transferred to
the soil and generates ground vibrations, is limited
by the dynamic response (impedance) of the soil.
It is also common practice to use in Equation 1
the lateral distance, measured at the ground surface.
This causes an error, which can be significant when
ground vibrations are measured close to the compac-
tion source, as the depth increases while the lateral
distance remains almost unchanged.

Mass
100
Drop height
Particle Velocity

Distance
10

Figure 2. Influence of depth of energy source on vibrations on

1
ground surface.
1 10 100
To the right, surface compaction, or the initial phase
(Energy / Distance) α of deep compaction is shown, where the energy
source is near the ground surface, compared to the
distance to the observer. Vibrations can be assumed
Figure 1. Estimation of ground vibrations caused by soil com-
to emanate from a point source and that waves
paction or pile driving based on applied energy and scaled dis- propagate mainly in the surface layer. In the near-
tance. field, vibrations will be caused predominantly by
body waves (compression and shear waves, propa-
Another important aspect is vibration amplification gating along the ground surface). At larger distances
due to the dynamic response of the compaction (corresponding to at least two wave lengths), surface
probe or pile. If the impedance of the compaction waves will start to dominate.
probe is low, ground vibrations can be amplified To the left in Figure 2, the case of deep compac-
significantly. This effect has been measured by tion is shown. The energy source has now penetrated
Heckman & Hagerty (1978). The practical implica- into the soil deposit. It is often assumed that vibra-
tions of the impedance effect on pile driving were tions emanate mainly from the base (lower end) of
discussed by Massarsch (1992, 1993). It is prefer- the compaction probe. However, in the case of deep
able to use a compaction probe with a low imped- compaction using impact or vibratory probes, a sig-
ance, as this will increase vibration transmission to nificant amount of compaction energy can also be
the soil and thus the compaction effect. generated along the shaft, Massarsch (2002). Thus,
vibrations can be generated by several energy 3.1 Dynamic Soil Resistance of Plate in Elastic
sources (base and/or shaft) simultaneously. Half-space
Even in the case of a deep vibration source, body
The action of a compaction probe can be analysed
waves will dominate in the near-field, i.e. at a dis-
assuming a plate impacting on the ground surface.
tance less than the depth of the energy source. Sur- Figure 3 illustrates the difference between the static
face waves usually begin to manifest themselves at a
and the dynamic soil resistance. When a rigid plate
critical distance, which can be seen as the location is loaded statically, it can be assumed that the soil
on the ground surface, where body waves emerging resistance increases gradually with deformation. Of
at depth are refracted as surface waves. The critical
course, ground vibrations will not be generated due
distance depends on several factors, such as the to the slow loading rate. However, if the plate is
wave length of vibrations (i.e. the vibration fre-
dropped on the ground surface, or loaded dynami-
quency and wave velocity in the soil layers), the
cally, an additional dynamic resistance will occur
variation of wave velocity with depth (soil layering), and the total soil resistance will be higher than in the
but also on the depth of the energy source. As can be
case of static loading. Further, an observer will no-
seen from Figure 2, the general assumption that
tice ground vibration. Thus, the problem of ground
ground vibrations are mainly due to propagation of
vibrations is linked to the dynamic soil resistance. It
surface waves, is not generally applicable in the case
should be noted that Figure 3 gives a simplified pic-
of deep soil compaction. Only at a distance, which is
ture of a complex problem, especially, when the ap-
several times large than the depth of the energy
plied stress reaches the strength of the soil (non-
source, surface waves dominate. At large distances,
elastic range).
it can also be assumed that vibrations are generated
by a point source, even in the case of vibrations SOIL RESISTANCE
emanating from the base and shaft.

3 DYNAMIC SOIL RESISTANCE

”Dynamic
”Dynamic Effect”
Effect”
Causes Vibrations
In spite of the practical importance of ground vibra- Static loading
tions associated with soil compaction, little informa-
tion is available in the geotechnical literature, de-
scribing the fundamental aspects of this problem.
There is a need to better understand the factors,
which influence the generation of ground vibrations.
The problem is complex, compared to conventional
vibration problems, such as ground vibrations from
DEFORMATION
industrial activities (machine vibrations) or blasting.
During soil compaction, the location and the dy- Figure 3. Variation of the total soil resistance with deforma-
namic characteristics of the energy source change. In tion, consisting of a static and a dynamic component. The dy-
addition, as one of the objectives is soil densifica- namic soil resistance is responsible for ground vibrations.
tion, the dynamic and static soil properties will
change gradually. Thus, the soil conditions at the A simplified analytical model will be used to assess
end of compaction will not be comparable to those, the parameters, which influence the emission of vi-
determined by soil investigations prior to compac- brations from the source. This approach has the ad-
tion. For instance, the shear wave velocity of loose, vantage over more sophisticated models, that the
uncompacted sand ranges typically between 100 and relative importance of different parameters becomes
150 m/s. After compaction, the shear wave velocity apparent.
can have increased to 200 to 300 m/s. Also the soil It can be assumed, that a rigid, circular plate im-
volume decreases and thus the density increase. It is pacts on the surface of an elastic half-space. Bodare
thus necessary to assess how compaction affects the and Orrje (1988) have developed a closed-form solu-
dynamic and static soil properties. tion for the dynamic resistance at the interface be-
Vibrations can originate either at the base, and/or tween a rigid plate and the soil, which is based on
along the shaft of the compaction probe. While the the “Herlitz equation”. This solution is attractive, as
dynamic response of piles subjected to impact load- the results of a complex problem can be presented in
ing is relatively well understood, the mechanism a relatively simple form. Orrje (1996) has shown by
governing pile-soil interaction during vibratory driv- comparison with model tests, that this theoretical
ing or vibratory compaction is still not fully under- model, which is strictly applicable only in elastic
stood. In the following sections, the interaction of a material, describes many of the significant aspects of
plate and of a shaft with the surrounding soil will be dynamic plate-soil interaction.
discussed.
In the case of an impacting circular plate, the dy- mainly influenced by the loading rate (loading ve-
namic stress σdyn acting on the surface can be esti- locity) and the strain-dependent soil impedance
mated from the following relationship (strain-dependent shear wave velocity). Bodare and
Orrje (1988) have presented solutions of the Herlitz
σ dyn = z 0 v (2) (29for the elastic case.
equation
where the z0 is the specific impedance and v is the Figure 4 defines three parameters, which influ-
particle velocity. The specific soil impedance zs can ence the force-time relationship. The two time fac-
be determined from tors tp and ts express the relationship between the
plate radius, r and the compression wave velocity, cp
cs ρ z s and the shear wave velocity, cs, respectively. The
z0 = = (3)
s s time factor tp is the time required for the compres-
sion wave cp to travel from the centre of the plate to
where ρ is the bulk density, cS is the shear wave ve- the perimeter and tS is the time for the shear wave to
locity. As will be shown below, the shear wave ve- travel from the centre to the perimeter. The time fac-
locity cs is not a soil constant but is affected by tor T0 defines the time at which the dynamic force
shear strain. Consequently, also the soil impedance reaches its maximum.
is strain-dependent. The parameter s is the ratio be- Force, kN
tween the shear wave velocity, cs and the compres-
sion wave velocity, cp and can also be expressed as a
function of Poisson’s ration, ν
1 − 2ν c
s= = s (4)
2(1 − ν ) c p
The dynamic stress σdyn acting on the interface of a π ts ) / qs
T0 = (π
circular plate and the soil can thus be estimated from
the following relationship
2(1 − ν )
σ dyn = v cs ρ (5)
1 − 2ν tp = r / cp ts = r / cs T0 Time, s

The dynamic stress is depends mainly on the shear

wave velocity below the plate, the soil density and Figure 4. Determination of dynamic soil resistance of a circular
Poisson’s ratio, cf. Figure 3. From Equation 5, the plate on in an elastic half-space. The shape is defined by three
maximum vibration vmax during loading can be de- time factors.
termined
The time factor T0 can be determined according to
σ
vmax = mob* s (6) Figure 4 if the time factor ts and the parameter qs are
ρ cs known. The force-time relationship is thus defined
by three “characteristic time factors”, tp, ts and TO. It
The soil density ρ and Poisson’s ratio ν do not vary is apparent that the plate radius and the velocity of
significantly. However, the shear wave velocity c*s the compression wave and of the shear wave below
decreases with increasing shear strain. Thus, there is the plate influence the duration of loading (rise
an upper limit to the vibrations, which can be trans- time). If the plate radius is increased, the predomi-
ferred to the soil due to dynamic plate loading. nant frequency of ground vibrations will decrease.
Equations 2 and 5 confirm that ground vibrations Bodare and Orrje (1988) have published solu-
will be higher when the rate of loading (impact ve- tions, from which the relationship between the static
locity) increases. The shear strain level γ d can be and the dynamic loading resistance can be calcu-
estimated if the particle velocity v and the shear lated. Figure 5 has been developed based on this re-
wave velocity, cs are known lationship. The dynamic soil resistance is responsi-
v ble for ground vibrations to occur. The force ratio
γ = (7) Pdyn/Pstat depends on the factor qs, which in turn is
cs affected by the two time constants ts and To. The dy-
namic force increases with increasing qs In dry sand,
it can be as high as 6 times the static soil resistance
3.2 Evaluation of Dynamic Base Resistance and many times higher in sand below the ground wa-
As shown in Figure 3, the total soil resistance during ter. This is due to the high compression wave veloc-
dynamic loading consist of a static and a dynamic ity of water compared to dry soil.
resistance component. The dynamic resistance is
10
Groundwater
9

6
Pdyn / Pstat

3
SAND
2
ν: 0,3
1

0
0 1 2 3 4 5 6 7 8 9 10

qs = (ts π) / T0

Figure 5. Ratio between the dynamic and the static force in dry sand and in sand below the ground water, cf. Figure 3.

It can be concluded that ground vibrations will be to the dynamic energy, which can be transferred
significantly higher in water-saturated soils. Also, from the compaction probe to the surrounding soil,
with increasing soil compaction, the shear wave ve- cf. Equation 8. This depends mainly on the shear
locity and thus ts increases and result in higher wave velocity at large strain (in the transition from
ground vibrations. the plastic to the elasto-plastic zone). However, with
increasing soil densification, the dynamic soil resis-
tance increases, and thus also the level of ground vi-
3.3 Evaluation of the Dynamic Shaft Resistance brations.
From equation 7 and 8 can be concluded that
3.3.1 Vertical Ground Vibrations ground vibrations generated along the shaft of the
The dynamic resistance acting along the shaft of a compaction probe will increase with increasing
vertically vibrating compaction probe can be esti- shear wave velocity. Thus, increasing soil densifica-
mated similar to the concept presented in the previ- tion will results in higher ground vibration ampli-
ous section. However, dynamic shaft resistance can tudes.
be estimate more easily if it is assumed that the Vibrations will be transmitted from the shaft of
probe moves as a rigid cylinder. Along the perime- the compaction probe to the soil as vertically polar-
ter, the dynamic shaft resistance can be estimated ized shear waves. This cylindrical wave front is
from similar to that of a Rayleigh wave and has the same
geometric damping.
τ dyn = v z s (7) It can be difficult to predict along which section
where v is the loading velocity and zs is the soil im- of the compaction probe (or pile) the maximum
pedance amount of vibration energy will be disseminated.
This depends on several factors, the most important
zs = c s ρ (8) of which are the rigidity of the compaction probe
(probe impedance) and the dynamic probe-soil inter-
where cs is the strain-dependent shear wave velocity action (impedance ratio between soil and probe).
and ρ is the soil density. The total shaft resistance
results from the combined effect of the static and the
dynamic component. Again, there is an upper limit
3.3.2 Horizontal Ground Vibrations erwise, too high impedance values will be obtained,
In the geotechnical literature is often assumed that in which over-predict the dynamic soil resistance and
the case of a vertically oscillating probe or pile, only thus vibration levels in the vicinity of the energy
vertical ground vibrations are generated from the source.
shaft, cf. previous section. However, in addition to The shear wave velocity can be measured in the
vertically polarised shear waves, also horizontal vi- field or in the laboratory, or estimated from empiri-
brations can exist in granular soils, Massarsch cal relationships. However, it is not generally recog-
(2002). This is due to the friction between the shaft nized that the shear wave velocity decreases even at
of the compaction probe and the soil. The horizontal relatively low strain levels. Figure 7 shows the result
stress changes result in a compression wave, which of a resonant column test on a reconstituted sample
contributes beneficially to the compaction of the of medium dense sand. The shear modulus and the
soil, as the lateral earth pressure increases. Figure 6 shear wave velocity decrease when a critical shear
shows the results of field measurements during vi- strain level (approximately 10-3 %) is exceeded.
bratory compaction, Krogh & Lindgren (1997). At a strain level of 0,1 %, the shear modulus and
the shear wave velocity have decreased from 76
50 MPa to 25 MPa, and from 205 m/s to 120 m/s, re-
spectively. This effect is significant, especially in
0
Velocity, mm/s

sandy soils and is so large that it can not be ne-

50 glected. The shear wave velocity at 1 % shear strain
(which corresponds to strain levels in the vicinity of
100
the compaction source) is about 50 % for plastic
150 soils (clays) and as low as 20 % for sands, when
compared to the shear wave velocity at small strain
200
0 0.1 0.2 0.3 0.4 0.5
(elastic wave). The shear wave velocity can be read-
Time, ms ily determined from the shear modulus if the soil
Figure 6. Horizontal vibration amplitude measured at different density ρ is known
depth levels during vibratory compaction, from Krogh &
G
Lindgren (1997). cs = (9)
ρ
Horizontally oriented vibration sensors (geophones) The decrease of the shear modulus and of the shear
were installed on, and at different levels below the wave velocity can be described by reduction factors
ground surface, at a distance of 2.9 m from the cen- RG and Rc, respectively. The shear modulus, G and
tre of the compaction probe. At the time of the vibra- the shear wave velocity, cs at a given shear strain
tion measurements, the tip of the compaction probe level can be determined if the respective maximum
had passed the lowest measuring point. It can thus values and the reduction factors RG and Rc are
be confirmed that strong horizontal vibrations are known
generated in spite of the fact that the compaction
probe is oscillating only in the vertical direction, G = RG Gmax (10a)
Massarsch (2002). The horizontal ground vibrations
amplitudes were of the same magnitude as the verti- c = Rc cmax (10b)
cal vibration amplitudes. It has also been shown that
as a result of vibratory compaction, the horizontal According to Equation 9, the following relationship
stresses increase in the soil. This compaction effect exists between the reduction factors of the shear
is of practical importance as it changes permanently modulus and the shear wave, respectively
the stress conditions after compaction. Rc = RG (11)
Extensive investigations have been carried out to
4 EFFECT OF STRAIN LEVEL ON SHEAR study the effect of shear strain on the shear modulus,
WAVE VELOCITY AND IMPEDANCE for instance Massarsch (1985), Vucetic & Dobry
(1985), Rollins et al. (1998). Döring (1997) has per-
The soil impedance (and thus the shear wave veloc- formed an extensive literature survey of resonant
ity) and Poisson’s ratio are of importance when as- column tests. Based on this compilation of numerous
sessing the dynamic soil resistance and ground vi- tests in cohesive and low-plastic soils (clays, silts
brations during soil compaction. As has been and silty sands), a relationship was developed be-
emphasised, the soil impedance must be determined tween the shear wave reduction factor and shear,
using the strain-adjusted shear wave velocity. Oth- which can be expressed in the following form
80 220

Shear Modulus
70 200
Shear Wave
V l it

Shea Wav Velocity, m/s

Shear Modulus, MPa

60 180

50 160

40 140

30 120

20 100
0,0001 0,001 0,01 0,1 1
Sheaer Strain ( %)

Figure 7. Resonant column test on a reconstituted sample of medium dense sand. Shear modulus (left axis) and shear wave velocity
(right axis) as a function of shear strain level (semi-logarithmic scale).

compaction source, the soil is usually in a plastic, or

G 1
=
[ ( )] (12a) elasto-plastic state.
Gmax 1 + α γ 1 + 10 − βγ Three compaction zones can be identified adja-
cent to the compaction source:
794 1. Plastic zone: where the soil is in a failure
α= 0.36505
(12b) condition and subjected to large strain levels >10-1
IP
%. The source-soil interaction in this zone is com-
plex but of less practice significance.
β = 0,046 + 0,5475 log( I P ) (12c) 2. Elasto-plastic zone: where the strain level
where α and β are empirically determined coeffi- ranges between approximately 10-3 and 10-1 %.
cients and IP is the plasticity index. The relationship 3. Elastic zone: where the shear strain level is
of Equation 12 is shown in Figure 8. All soils exhibit below 10-3 %, and no permanent deformations can
strain softening and this effect depends on soil plas- be expected.
ticity. The shear wave velocity is reduced when a In the plastic zone the vibration velocity is lim-
critical shear strain level is exceeded. The reduction ited by the shear wave velocity at large strain (and
of the shear wave velocity is more pronounced in thus the strength) of the soil. In the plastic, and the
soils with low plasticity, i.e. sands and silty sands. In elasto-plastic zone, the wave propagation velocity is
soils with higher plasticity, the shear wave velocity strain-dependent. The shear wave velocity increases
reduction is lower but still of importance. with increasing distance from the energy source. In
the elastic zone, the wave propagation velocity is
constant (elastic wave velocity).
5 VIBRATION PROPAGATION IN SOILS Wave propagation in the elastic and elasto-plastic
zone can be analyzed using simple or more sophisti-
The propagation of vibrations from a source into the cated models. However, it is necessary to account
surrounding soil is complex. In the vicinity of the for the influence of shear strain, cf. Figure 8.
1,00

0,90

0,80

0,70

0,60
Rc = c/cmax

0,50 Plasticity Index, IP

0,40 5 10 15 20

0,30
25 30 35 40

0,20
45 50 80 60

0,10

0,00
0,001 0,01 0,1 1
SHEAR STRAIN, γ %

Figure 8. Effect of shear strain on shear wave velocity as a function of soil plasticity.

where A1 and A2 are the vibration amplitudes at dis-

tance R1 and R2, respectively. The exponent n de-
5.1 Wave propagation pends on the wave type as shown in Table 1.
The propagation of waves in an elastic or elasto-
plastic medium can be analysed using sophisticated Table 1. Wave exponent n for different wave types
methods. However, in many cases, also simplified Wave type Exponent n
solutions will be satisfactory, provided that source- Body Wave 1,0
Body Wave at surface 2,0
soil interaction is accounted for and appropriate soil Surface wave (or cylindrical wave) 0,5
parameters have been selected. This aspect appears
to be more important than the degree of analytical
sophistication.
The absorption coefficient α takes into account the
The vibration velocity at the interface of the plas-
effect of soil damping
tic and elasto-plastic zone can be estimated using the
relationships provided in Section 3. The maximum 2π D f
vibration velocity can then be used to predict the vi- α= (14)
cs
bration propagation from the shaft and from the base
of the compaction probe into the surrounding soil. where D is material damping, f is the predominant
Also in the case of vibration propagation the ef- vibration frequency and cs is the shear wave veloc-
fect of strain level on vibration velocity in the near- ity. Material damping at small strain levels is almost
field (elasto-plastic zone), must be taken into ac- independent of soil type and ranges typically be-
count. For most soil compaction problems, an esti- tween 2 and 4 %. However, also material damping
mate of vibration velocity using the simple wave at- increases with increases with shear strain, Vucetic
tenuation from a point source will yield satisfactory and Dobri (1985), Rolling et al (1990). At a shear
results strain levels between 0.5 and 1 %, the damping ratio
can be as high as 20 %.
A2 R
=( 2) −n
e −α ( R 2 − R 1 ) (13)
A1 R1
5.2 Predominant Frequency resonance, ground vibrations are strongly amplified.
The predominant vibration frequency can be deter- The probe and the surrounding soil vibrate in phase,
mined by field measurements. In the case of vibra- resulting in a more efficient transfer of compaction
tory compaction, the predominant frequency at some energy.
distance from the compaction point corresponds
usually to the operating frequency of the vibrator. 18
Compaction
However, in the case of impact loading, it is more

VERTICAL PARTICLE VELOCITY, mm/s

Frequency
16
difficult to predict the predominant vibration fre-
14
quency, which depends on the dynamic interaction
12
between the source and the surrounding soil. The in-
formation provided in Section 3 can be used to esti- 10

mate the predominant frequency in the case of im- 8

pact loading. 6

The predominant period of an impulse, for in- 4

Penetration
Frequency
stance due to the impact of a plate on the ground sur- 2
face, can be estimated according to Figure 4. The
0
time factors tp and ts can be calculated from the plate 5 10 15 20 25 30 35
radius and the respective wave velocity. The pa- FREEQUENCY, Hz

rameter qs can be estimated, based on Figure 5, and

Figure 9. Vertical ground vibration velocity at a distance of 4
is for soil compaction typically in the range of 2 to m from the compaction probe during probe penetration and
4. With this information it is possible to estimate the resonance compaction
rise time to the maximum force. The predominant
vibration frequency can then be estimated from The dynamic response of the soil deposit during
1 compaction can also be used to monitor the compac-
f = (15) tion effect. With increasing densification of the soil
4 T0 layers, the resonance compaction frequency rises.
It is interesting to note that the predominant vibra- Also the ground vibration velocity increases and soil
tion frequency is mainly dependent on the plate ge- damping is reduced.
ometry and the wave velocity below the plate. The It can be concluded that ground vibrations are
frequency increases with decreasing time factor. significantly amplified if the compaction frequency
Thus a smaller plate gives a higher vibration fre- is close to the resonance frequency of the compac-
quency than a larger plate at the same ground condi- tion system. Ground vibration amplification can be
tions. as high as 10 to 20 times and should therefore be
taken into consideration.
5.3 Vibration Amplification
An important aspect of ground vibration prediction 6 CASE HISTORY – DYNAMIC COMPACTION
in connection with soil compaction is vibration am-
plification. Vibration amplification can be used in Changi airport was initially constructed during the
connection with vibratory compaction to enhance 1970’ s when 700 ha of land were reclaimed from
the compaction efficiency, Massarsch (2002). This the sea. This ambitious land reclamation project was
compaction method is known as “resonance com- continued to provide two additional landing strips
paction” (MRC). A compaction probe is inserted in and terminal buildings. The land reclamation com-
the ground, using a vibrator with variable frequency. prised 1 550 ha (210 million m3). After removal of a
During the probe penetration phase, the optimal soft marine clay layer from the sea bottom the exten-
penetration speed is achieved when the probe is vi- sion of the airport was constructed on a 5 to 15 m
brated at a high frequency (around 30 Hz), cf. Figure thick sand fill. The sand was loose below a dense
9. During the compaction phase, the vibrator fre- surface layer. The average cone penetration resis-
quency is lowered to achieve vibration amplification tance prior to compaction was 4 to 8 MPa. The fill
(resonance of the vibrator-probe-soil system). Figure consisted mainly of sand with an effective particle
9 shows how the vertical vibration velocity, meas- diameter d10 = 0.5 mm. The ground water level was
ured on the ground surface, varies as a function of located about 3 m below the formation level.
the vibration frequency. During probe penetration Extensive field investigations were carried out by
and extraction, a high vibration frequency (around the author in connection with the compaction pro-
30 Hz) is used, which does not cause significant ject. Two different soil compaction methods were
ground vibrations. During the compaction phase, the compared: dynamic compaction, using a heavy fal-
operating frequency of the vibrator is reduced to the ling weight, Figure 10, and a vibratory compaction
resonance frequency of the probe-soil system. At probe, Figure 11.
Dynamic compaction is probably the oldest MPa. The mass of the pounder was 21 ton and the
ground improvement method. In spite of its simplic- drop height was varied between 15 and 20 m. The
ity, the efficient execution requires careful planning degree of compaction increased from the ground sur-
and extensive field testing. Compaction is usually face and reached a maximum at about 4 m depth (15
performed in several passes. The compaction energy to 29 MPa). Below this level, the compaction effect
(mass of pounder times drop height) is usually cho- decreased again.
sen, based on past experience, taking into considera- During the initial compaction tests, comprehen-
tion the soil conditions and the required degree of sive dynamic measurements were performed, Krogh
densification. However, at large projects trial com- & Lindgren (1997). The pounder was instrumented
paction is usually carried out to determine the re- with an accelerometer and ground vibrations were
quired number of compaction passes, the optimal measured on and below the ground surface with ac-
pounder mass and drop height. celerometers (in the near-field) and with geophones.
Figure 12 shows vibration measurements during the
impact of the pounder.

Figure 10. Deep soil compaction (“dynamic compaction”) at

change airport.

Figure 12. Vibration measurements during dynamic compac-

tion at Changi airport, Singapore.

Different types of seismic measurements were per-

formed in order to determine the shear wave velocity
prior to, during and after compaction. Geophones
were also installed below the ground surface, cf.
Figure 6.
The shear wave velocity varied within the strati-
fied soil deposit, also prior to compaction. The shear
wave velocity was relatively high (around 200 m/s)
in the upper surface layer. In the uncompacted sand
at the ground water level, the shear wave velocity
decreased to 160 m/s.
The results of the vibration measurements and the
interpretation of the measurements are presented in
Figure 11. Vibratory compaction (MRC), using vibrator with the following figures. Figure 13 shows the accelera-
variable frequency in order to achieve vibration amplification. tion (retardation) of the pounder during impact. The
Note the geophone for resonance measurements in the picture time factor T0 can be estimated from the acceleration
to the right. record. Due to practical reasons and because of the
air trapped below the pounder at impact, the rise
time is not as sharp. This aspect was taken into ac-
In the following, the results of field measurements count when selecting the time factor as 0.025 sec-
during dynamic compaction will be described. As a onds.
result of extensive field tests, the following dynamic
compaction procedure was chosen in order to
achieve a cone penetration resistance of at least 10
10

9
Sand
8 Groundwater

7
Dry
6

P dyn / P stat
5

4 qs = 2,09
Pdyn/Pstat = 0,52 ν: 0,3
3

0
0 1 2 3 4 5 6 7 8 9 10

qs = (ts π ) / T0
Figure 13. Acceleration measured on the pounder during im-
pact. Indicated is the estimate time factor, T0. Figure 15. Estimation of dynamic soil resistance and vibration
amplification, cf. Figure 5.
From the acceleration measurement the impact ve-
locity can be determined by integration. From the The penetration of the pounder can be calculated by
measured acceleration and the pounder mass the double integration of the acceleration record. Figure
force can be calculated. If the applied stress is plot- 16 shows the movement of the pounder during im-
ted against the velocity, the specificsoil impedance pact, and the calculated penetration depth was about
can be readily determined, Figure 14. 0.35 m. This value is in good agreement with field
observations and confirms the accuracy of the accel-
eration measurements.
6
2.5 10

z0 = σ/v = 1,47⋅106 / 4

2.0 10
6 = 375 000 N/m3⋅s.

6
1.5x10
Stress, N/m2

6
1.0x10

5
5x10

0
cs: 90 m/s

5
-5x 10
2 0 2 4 6 8 10 12 14 16 Figure 16. Pounder penetration into the ground as determined
from acceleration record, cf. Figure 13.
Velocity, m/s
Figure 14. Soil impedance determination from pounder impact
record. Figure 17 shows vibration measurements at increas-
ing distance from the vibration source on the ground
The specific soil impedance corresponds to 375 surface. The predominant frequency of 12 Hz can be
kN/m3s. From the measured impedance, the shear estimated from the vibration record and is in good
wave velocity in the zone of impact (elasto-plastic agreement with the frequency estimated from the
zone) below the pounder can now be determined. time factor, T0, cf. Equation 15. The wave propaga-
Assuming a soil density of 1,9 t/m3 and a wave ve- tion velocity determined from the arrival time inter-
locity ratio s = 0,53, a shear wave velocity of 90 to vals corresponds to approximately 400 m/s.
110 m/s is obtained. This shear wave velocity is ap-
proximately 50 % of the shear wave velocity meas-
ured at small strain. 7 SUMMARY AND CONCLUSION
From the acceleration record, T0 and the shear
wave velocity, the parameter qs can be determined. It Vibrations caused by soil compaction are important
is then possible to estimate the dynamic soil resis- when planning projects in urbanized areas or in the
tance and the vibration amplification, Figure 15. In vicinity of vibration-sensitive structures or installa-
the present case, qs = 0.52, which is a relatively low tions. In spite of its practical importance, little quan-
value, which is probably due to the large plate di- titative information has been reported in the litera-
ameter and the low shear wave velocity at impact. ture about this problem. Most correlations are of an
empirical nature and neglect important aspect, such
as interaction between the energy source and the
ground, and wave propagation in the ground.
larized shear waves. However, in granular soils, due
to soil friction, also horizontal vibrations are created.
Ground vibrations increase during soil compac-
50 tion and also the vibration frequency increases. An-
other important factor is vibration amplification,
10 m which occurs during vibratory compaction when the
100
operating frequency approaches the frequency of the
vibrator-probe-soil system.
150 Results of extensive field tests during compaction
trials at Changi airport are described. The energy
Particle Velocity, mm/s

12.5 m transfer from a plate impacting on the ground sur-

200
face was measured. From the acceleration measure-
ments on the pounder the specific soil impedance
250 and the predominant vibration frequency can be de-
termined. The field tests confirm that the soil imped-
15 m
ance is lower in the vicinity below the compaction
300
plate. The measured predominant vibration fre-
quency is in agreement with the predicted value,
350 based on the time factor T0.
17.5 m
400
8 ACKNOWLEDGEMENTS

450 The research on soil compaction was supported by a

0.2 0.4 0.6 0.8 1
grant from the Lished Foundation (Lished Stiftel-
Time, sec sen), which is gratefully acknowledged. Anders
Lindgren participated in the field tests at Changi and
Figure 17. Ground vibration measurements on the ground sur- performed part of the data evaluation. Dr. Anders
face caused by pounder impact. Bodare has developed the theory of dynamic plate
loading, based on the Herlitz equation. His com-
ments and suggestions have been most valuable and
An attempt has been made to describe quantitatively are acknowledged.
the interaction between the compaction source and
the surrounding soil. The interaction of a plate and
of a cylindrical shaft with the soil have been ana- REFERENCES
lysed and solutions are proposed for estimating the
dynamic soil resistance. The most important parame- Attwell, P, B. & Farmer, I. W., (1973), Attenuation of Ground
Vibrations from Pile Driving, Ground Engineering, Vol. 3,
ter is soil impedance, which determines the dynamic No. 7, pp. 26 - 29.
resistance of the soil during dynamic loading. How- Bodare, A., & Orrje., 1988. Impulse load in a circular surface
ever, the shear wave velocity and thus the soil im- in an infinite elastic medium. (Closed solution according o
pedance decrease with shear strain and this effect the theory of elasticity). Extended version. KTH jord och
must be taken into consideration. bergmekanik, JoB Report No. 23, Stockholm.
The soil impedance limits the dynamic energy Döringer, H., 1997. Verformungseigenschaften von bindigen
Böden bei kleinen Deformationen (Deformation properties
and thus the vibrations that can be transmitted from of cohesive soils at small strain). Royal Institute of Tech-
a compaction source. nology,Stockholm, Soil and Rock Mechanics.
A simple chart is presented, which makes it pos- Examensarbete 97/8, 54 p .
sible to estimate the ratio between the dynamic and Krogh, P., & Lindgren, A. 1997. Dynamic field measurements
the static soil resistance. The dynamic resistance, during deep compaction at Changi airport, Singapore.
Examensarbete, JoB, Institutionen för Anläggning och
and thus ground vibrations, increase with increasing Miljö, Royal Institute of Technology (KTH), JOB Report
shear wave velocity and increasing loading rate. 97/9, 88 s.
A semi-empirical relationship is proposed for the Massarsch, K. R., 1979. Closing remarks, Discussions, Session
assessment of the reduction of shear wave velocity 1, Seventh European Conference on Foundation Engineer-
as a result of shear strain. The strain-softening effect ing, Brighton, Proceedings , Vol. 4, p. 59.
is most pronounced in granular soils. Massarsch, K. R., 1985. Stress-Strain Behaviour of Clays. 11th
International Conf6erence on Soil Mechanics and Founda-
The propagation of ground vibrations from the tion Engineering, San Fransisco, Proceedings, Volume 2,
shaft and from the base of the compaction probe pp. 571 - 574.
have been described. A vertically oscillating com- Massarsch, K. R., 1992. Static and Dynamic Soil Displace-
paction probe generates along the shaft vertically po- ments Caused by Pile Driving, Keynote Lecture. Fourth In-
ternational Conference on the Application of Stress-Wave
Theory to Piles, the Hague, the Netherlands, September 21-
24, 1992, pp.15 - 24.
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.
Massarsch, K. R. & Broms, B. B., 2001. New Aspects of Deep
Vibratory Compaction. Proceedings, Material Science for
the 21st Century, JSMS 50th Anniversary Invited Papers,
Vol. A, The Society of Material Science, Japan, pp. 172 –
179.
Massarsch, K. R., 2002. Effects of Vibratory Compaction.
TransVib 2002 – International Conference on Vibratory
Pile Driving and Deep Soil Compaction. Louvain-la-Neuve.
Balkeama, Keynote Lecture, pp. 33 - 42.
Massarsch, K.M. & Fellenius, B.H., 2002. Dynamic compac-
tion of granular soils. Canadian Geotechnical Journal, Vol.
39, No. 3, pp. 695 - 709.
Orrje, O., 1996. The use of dynamic plate load tests in deter-
mining deformation properties of soil. Doctoral thesis, Di-
vision of Soil and Rock Mechanics, Royal Institute of Tech-
nology (KTH), Stockholm. 111 p.
Rollins, K. M., Evans, M. D., Diehl, N. B. and Daily, WD, III.,
1998. Shear modulus and damping relationships for gravels.
ASCE. Journal of Geotechnical and Geoenvironmental En-
gineering, Vol 124, No 5, pp. 396-405
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.

View publication stats