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

net/publication/270426236

Densification of granular soil by dynamic compaction

Article  in  Proceedings of the Institution of Civil Engineers Ground Improvement · January 2009

DOI: 10.1680/grim.2009.162.3.121

CITATIONS READS

19 3,880

4 authors:

Myint Win Bo Yung Na

Bo & Associates Ltd Larsen and Toubro
122 PUBLICATIONS   2,621 CITATIONS    9 PUBLICATIONS   105 CITATIONS   

SEE PROFILE SEE PROFILE

Arul Arulrajah Ming-Fang Chang

Swinburne University of Technology 37 PUBLICATIONS   1,087 CITATIONS   
355 PUBLICATIONS   6,045 CITATIONS   
SEE PROFILE
SEE PROFILE

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

A review of the recent studies on innovative bricks View project

Evaluation of Fly Ash Treated Reclaimed Asphalt Pavement (RAP) for Base/Subbase Construction View project

All content following this page was uploaded by Arul Arulrajah on 23 March 2015.

The user has requested enhancement of the downloaded file.

Proceedings of the Institution of
Civil Engineers
Ground Improvement 162
August 2009 Issue GI3
Pages 121–132
doi: 10.1680/grim.2009.162.3.121

Paper 800004
Received 14/01/2008
Accepted 12/11/2008
M. W. Bo Y. M. Na A. Arulrajah M. F. Chang
Keywords: land reclamation/ Director (Geo-Services), DST Chief Engineer, Hyundai Senior Lecturer, Faculty of Associate, Shannon & Wilson, Inc.,
densification/dynamics/granular soil Consulting Engineers Inc., Engineering and Construction, Engineering and Industrial USA.
Thunder Bay, Ontario, Canada Seoul, South Korea Sciences, Swinburne University of
Technology, Melbourne, Australia

Densification of granular soil by dynamic compaction

M. W. Bo DUC, MSc, PhD, FGS, FICE, CEng, CGeol, CEnv, CSci, PEng PGeo, EurGeol, EurEng, Y. M. Na MSc, PhD, PEng,
A. Arulrajah MEng, PhD, CPEng, FIEAust and M. F. Chang MEng, PhD, PE

Land reclamation often involves the placement of loose 2. CHANGI EAST RECLAMATION PROJECT AND
granular soil by means of hydraulic filling. Sand fill SAND FILL DENSIFICATION
formed by hydraulic filling generally does not allow The Changi East reclamation projects include 2000 ha of land
densification by surface compaction methods because of reclamation, which was carried out under five phases because
their limited depth of influence. Loose granular soil is vast quantities of fill material and ground improvement were
susceptible to liquefaction upon the impact of dynamic required. The project site was located in the eastern part of
forces. Even under static conditions, loose granular soil Singapore, as shown in Figure 1. A major portion of the site
may be subjected to bearing capacity failure and large was reclaimed for future expansion of the Singapore Changi
settlements, because of its low shearing resistance and airport, and the remaining areas were for industrial and other
high compressibility. Various densification methods are usages.
used for improving such soils to increase the friction
angle and elastic modulus. Several methods of deep These projects included an extensive amount of soil
compaction are available for such applications; among improvement works for treating the underlying compressible
these, dynamic compaction is one of the most effective clay as well as the granular fill. The hydraulic fill at Changi
ways of densifying granular soils to a significant depth. consisted mainly of sand with less than 10% fines to a
However, the success of dynamic compaction is affected thickness of up to 20 m. The grain size distribution of the
by many factors, several of which are not yet fully granular fill is shown in Figure 2.
understood. This paper deals with the dynamic
compaction densification method utilised at the Changi As the granular fill was deposited by means of hydraulic filling
East reclamation site in Singapore for the improvement using sand dredged from a borrow source, it was in a loose
of reclaimed sand fill. Field data collected are used as a state, with the range of cone resistance falling between 5 and
basis to investigate the effectiveness of the densification 7 MPa. In order to avoid excessive settlement of the fill, and to
method and the effect of various influencing factors satisfy the requirements of runway pavement design, the cone
critical to the success of dynamic compaction treatment. resistance specified for the granular fill after densification was
15 MPa for the runway and 12 MPa for the taxiway areas:
1. INTRODUCTION these are approximately equivalent to relative densities of 75%
Reclamation by means of hydraulic filling generally results in and 70% respectively.
loose granular fill. In addition, the granular soil mass formed
by hydraulic filling cannot be densified using surface An area of about 114 ha was improved by deep compaction
compaction, which requires lift-by-lift application of methods covered granular fill 7–10 m thick. Three types of
compacting energy. Therefore deep compaction is often deep compaction method were deployed: dynamic compaction,
required. vibroflotation and Muller resonance compaction (MRC) . The
areas where the three different types of compaction method
Several methods of deep compaction are available for the were used are shown in Figure 3. The dynamic compaction
densification of granular soils; among these, dynamic method was deployed in the area where the required depth of
compaction is one of the effective methods for densifying compaction was 5–7 m. The vibroflotation and MRC methods
granular soil in situ to a great depth. However, the success of were adopted in the areas where the required thickness of
dynamic compaction is affected by many factors, several of compaction was 7–10 m.
which are not yet fully understood. This paper describes the
dynamic compaction densification method used at the Changi Each of the three compaction methods has its own advantages
East reclamation site in Singapore for the improvement of and disadvantages, depending on the site and soil conditions in
reclaimed sand fill. Field data collected were used to the various areas. Nevertheless, the specified degree of
investigate the effectiveness of the densification method and compaction was achieved in all areas. This paper emphasises
the effect of various influencing factors critical to the success the densification of granular fill by applying dynamic
of dynamic compaction treatment. compaction.

Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al. 121
 N

Runway 1 Republic of Singapore

Singapore Changi Airport Location

Changi Coastal Road Length of

Volume of vertical
Phase 1C
Project Area: ha drain: mm
sand: mm3

Phase 1C Phase 1B Phase 1A 501 65 –

Proposed runway 3 Phase 1B 520 75 28
Phase 1C 524 68 49
Phase 1A Area A – North 91·0 12 13·0
Area A Area A – South 450 52 50·0
Area A – South
North Total 2,086·0 272 140·0

Sea

0 500 1000 2000 3000 m

Figure 1. Location of Changi East reclamation project

however, it is not practical or economical to measure the in

100
situ ö9, and it is also difficult and time consuming to measure
Passing percentage: %

80 the value of E at various levels along the depth of a soil

profile. In some projects the increase in modulus measured by
60
pressuremeters is specified, but there are practical difficulties in
40 measuring the modulus with pressuremeters in the sand. The
required degree of densification can be specified in terms of
20
relative density, in view of the fact that there is a strong
0 correlation between the relative density of granular soil and its
0·01 0·1 1 10
peak angle of shear strength and modulus of elasticity of
Particle diameter: mm
granular soil. The required peak angle of shear strength and
Figure 2. Grain size distribution of granular fill used in Changi modulus of elasticity were determined based on acceptable
East reclamation project settlement and bearing capacity requirements for the future
loading expected on the improved land. The improved granular
soil parameters after densification are generally measured by
3. DETERMINATION OF DENSIFICATION means of the standard penetration test (SPT) or the cone
REQUIREMENTS penetration test (CPT). Dynamic probing tests (DPT) were also
frequently carried out for preliminary assessment of the
3.1. Degree of densification increase in resistance. Relative density can be correlated with
The densification requirements for a project are often field measurements, such as those from SPT and CPT. 1–6
ascertained based on the required bearing capacity, and/or the
allowable magnitude of tolerable settlement. The bearing 3.2. Depth of compaction
capacity of a shallow foundation is usually dependent upon the Determination of the required densification depth is also
geometry of the foundation and shear strength parameters of critical for a densification project. If there is no design
the soil, especially the peak angle of shear strength (ö9). The requirement for liquefaction (i.e. no anticipated seismic or
magnitude of settlement is inversely proportional to the dynamic forces), the densified granular soil mass is required
modulus of elasticity (E) of granular soil. It is desirable to only to meet bearing capacity and settlement criteria. In this
avoid bearing failure, which can be achieved by improving the case, the densification can be determined based on the
peak angle of shear strength of the granular fill, and to estimated pressure bulb for a particular type and geometry of
minimise settlement by enhancing the modulus of elasticity of foundation. However, if liquefaction is a consideration, the
the soil through deep densification treatment. In practice, depth of treatment required may increase significantly,

122 Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al.
 CT9 CT10CT11CT12 CT13CT14CT15 CT16 CT17CT18CT19 CT20 CT21 CT22 CT23CT24CT25 CT26 CT27CT28 CT29 CT45 CT46 CT47CT48 CT49CT50CT51 CT52 CT53 CT54CT55 CT56 CT57 CT58

CX56CLA10

CLA11
CLA8
CLA9
CLA2
CLA3

CLA4
CLA5
CLA12
CLA13

CX61
CX57
CX55

CX58
CX59
CX60
CX1 CX2 CX3 CX4 CX5 CX6 CX7 CX8 CX9 CX10CX11CX12CX13CX14CX15CX16CX17CX18CX19CX20 CX21CX22CX23CX24CX25CX26CX27CX28CX29CX30 CX46CX47CX48CX49CX50CX51CX52CX53CX54

CLB2
CLB3
CLB4
CLB5

CLB11
CLB12
CLB13
CLB14
CLB7
CLB8
CLB9
CLB10
CLB1 CLB6 CLB15
CLB26 CLB27
Pilot area

CLC10

CLC21
CLC22

CLC23
CLC24
CLC11

CLC2
CLC1 CLC5

CLC6

CLC9
CLC3

CLC7
CLC4 CLC8 CLC8
CR-2
CR-1
CR-0

Legend:
MRC ⫹ surface tamping (15 MPa)
MRC ⫹ surface rolling (12 MPa)
Vibroflotation (15 MPa)
Vibroflotation (12 MPa)
Dynamic compaction (12 MPa)

Figure 3. Layout showing areas where different methods of deep compaction were applied

depending on the liquefaction potential under the predicted The basic mechanism underlying dynamic compaction in
dynamic and seismic forces. granular soils is relatively well understood. When the pounder
impacts on the ground surface, the impact energy is
4. DYNAMIC COMPACTION: CURRENT PRACTICE transformed into seismic radiation, which subsequently
AND COMMON UNDERSTANDING transmits into the underlying soil mass. At the moment of
Dynamic compaction (DC) is a technique for improving the impact with the pounder, the impact energy is transmitted
mechanical properties of granular soil at depths by repeatedly mainly in body waves that consist of compression and shear
lifting and dropping a heavy weight (pounder) onto the ground waves, although surface waves are also generated in the soil.
surface. Repeated impacts over split-second durations are Whereas the body waves propagate radially outwards from the
imparted to the granular soil when the heavy weight hits the source along a hemispherical wavefront, as shown in Figure 4,
ground surface. The impact energy causes the soil particles to the surface waves propagate horizontally along the surface.
be rearranged into a denser state. The selection of impact The influence of these shock waves on the soil is dependent on
spacing and the number of drops per impact point is essential the soil types and the degree of saturation. For dry deposits the
if the specified density is to be achieved. compressive and shear waves induced by the impact overcome
the interlocking stresses within the loose strata, resulting in a
The compaction process is usually repeated in several passes reduction of voids. For a saturated granular deposit the
until the required post-treatment relative density has been mechanism of densification is quite different. The compressive
achieved. The spacing for the first pass of impact points is stresses induced by the DC impact result in a sudden increase
usually equal to the thickness of the densifiable layer, in order in pore water pressure, thereby forcing the soil into a state of
to allow the impact energy to reach the lower part of the layer. temporary liquefaction. The shear waves and Rayleigh waves,
The second pass is generally
made at the centroid prints of
the first pass. During each
pass, several drops are made Circular footing
Geometrical
at the same point. There is a damping law ⫺2 ⫺0·5
r⫺2 r r
maximum number of impacts ⫹
that leads to the closure of Rayleigh wave
voids in order to achieve the ⫹
minimum void ratio; after Horizontal comp.

this there is usually no Vertical comp.

further closure of voids in the Shear wave Relative amplitude

treated soil mass. After each ⫹ ⫹
pass, the craters created by
the dropping pounder are r⫺1
Geometrical
usually backfilled with damping law
surrounding materials prior Sh ⫹ r
ea ⫹
to the next pass. Finally, an rw
ind ⫺1 Compression wave
‘ironing’ pass with a low- ow r
energy impact and reduced
drop height is performed to
compact the shallow surface Figure 4. Wave propagation due to dynamic compaction7
layer.

Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al. 123
 which are slower, travel through the soil skeleton. The carried out at the Changi East reclamation site. The site was
combination of a temporary loss of contact stresses and underlain by a recently reclaimed sand fill with an average
dynamic oscillation forces the soil particles to rearrange into a thickness of 10–12 m. The granular soil has a D50 of 0.4 to
dense state. 1 mm (Figure 2), and a narrow range of qc , between 5 and
7 MPa (Figures 14 and 15). The soil was loose, but fairly
4.1. Influence depth homogeneous.
The common understanding of dynamic compaction of
granular soil is that the degree of improvement increases with The results of numerous tests obtained from Changi during the
the applied energy, and the influence depth increases with the early phases of reclamation suggest that the n factors for
pounder weight and drop height. The pounders are usually various energy weights and drop heights 16 vary from 0.33 to
square, circular, hexagonal or octagonal in shape, and made of 0.44, similar to the value proposed by Leonards et al. 10 In this
steel or concrete. Their weights normally range from 5 to 40 t, study, influence depth was determined from the CPT tests
and drop heights could be up to 25 m. Menard and Broisc 8 carried out after compaction. Table 1 shows that different n
proposed a formula that allowed an estimation of the influence values are obtained using different pounders in Changi. It is
depth D, in metres, as worth noting that the depth of influence is also dependent
upon the size and shape of the pounder. Data collected from
1 D ¼ ð w 3 hÞ
1=2 Changi suggest that the same weight of pounder with the same
energy may result in different influence depths if the geometry
of the pounder is different.
where w is the weight of the pounder in tonnes, and h is the
drop height in metres. A more appropriate and accepted form From Equation 2, the required pounder weight and height of
of equation is given by Lukas9 as the drop can be selected to achieve the required depth of
compaction. The effectiveness of dynamic compaction is
2 D ¼ nð w 3 hÞ
1=2 dependent on the combination of weight and geometry of the
pounder, the height of drop, the spacing, the number of drops,
and the total compactive energy applied. Details of the
where n is an empirical coefficient factor that varies between equipment and the energy applied, together with the achieved
0.3 and 1.0. An n value of 0.5 was proposed by Leonards et densification in the dynamic compaction work carried out at
al. 10 on the basis of compilation of field data for a number of Changi East, are summarised in Table 2.
conditions. The effectiveness of dynamic compaction is
strongly affected both by the soil condition and by the energy 5.1. Shape of pounder
configuration. Pounders of several different shapes, including square,
hexagonal and circular, have been used in dynamic
Van Impe 11 pointed out that the depth of influence depends compaction. The thickness of pounders may also vary. Some
upon the surface area and the shape of the pounder. Lukas 12 pounders have foot studs, or nuts and bolts used to hold the
stated that multi-tamping improved only the zone of influence, steel plates together. Pounders are usually made up of steel
and not the depth of influence. The degree of granular soil plates, although a few consist of concrete block. Figure 5
improvement by dynamic compaction peaks at a critical depth, shows two types of pounder used in the dynamic compaction
which is roughly one half of the maximum depth of influence. works at Changi. Generally, a pounder with a smaller base area
Mayne et al. 13 proposed a useful correlation between the will penetrate deeper than a pounder with a larger base area.
normalised crater depth, Dc =(wh)1=2 , and the number of drops. This creates additional depth of influence and vertical
displacement in the soil, which will be discussed in the next
Poran and Rodriguez 14 reported that there is a consistent section.
relationship between the trial specific energy and the
dimensions of the densifiable soil mass. 5.2. Lifting and dropping mechanism
In dynamic compaction work, lifting of the pounder is usually
Based on numerical analyses, Chow et al. 15 proposed a method achieved by using a crane with a winch system. High-capacity
for predicting the crater depth by applying a wave equation cranes with various boom lengths are used in dynamic
model. They validated their
predictions with two case
studies, and found that their Pounder mass: t
proposed method worked
15 14 23 23
well.

5. INVESTIGATION AND Drop height: m 20 20 12.5 25

Pounder surface area: m2 3.87 2.25 5.5 5.5
OBSERVATIONS AT Energy: t–m 300 280 287.5 575
CHANGI Influence depth: m 7.5 7 6 8
In order to investigate the n* 0.433 0.418 0.354 0.334
effectiveness of the dynamic
compaction and factors * Determined from CPT.
affecting influence depth and Table 1. Value of n for various pounders (after Choa et al. 16 )
crater depth, a study was

124 Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al.
 Scheme pounder used at Changi East.
Even in free-fall situations,
1 2 3 4 energy losses arising from
friction caused by rapid
Pounder weight: t 23 15 18 18 movement of the pounder in
Drop height: m 25 20 24 24
the air can still be expected.
No. of drops per pass 5 10 10 12
Energy per drop: t 575 300 432 432 Figure 9 shows how the
Spacing at each pass: m2 636 636 8.5 3 8.5 10 3 10 measured deceleration,
No. of passes 2 2 2 2 velocity, vertical
Effective area of improvement: m2 5.5 3.87 3.4 3.4 displacement and input stress
Energy per m2 : t-m 160 166 120 105
compared with the calculated
Compacted depth: m 7 7 7 7
Cone resistance achieved: MPa ,15 ,15 ,12 ,12 values.

Table 2. Details of dynamic compaction at Changi The theoretical impact

velocity of a falling pounder,
v1, can be calculated as

compaction works. However, if the pounder is too heavy, ð

tripods will need to be used instead of cranes. The drop point is 3 v1 ¼ a t dt þ C v
just in front of the crane, or at the centre of the tripod. Figure
6 shows a crane and a tripod used in the dynamic compaction
works at Changi East. In most cases, the cable follows with the
where a1 is the deceleration, t is the elapsed time and Cv is an
pounder when the pounder is released. Significant friction
integration constant.
between the pulley and cable can be expected, resulting in
some energy losses. This is taken into account in the empirical
The impact stress (ó t ) applied by the pounder over the duration
coefficient n in Equation 2.
of the impact can be calculated using the measured
deceleration record, as
A system that allows the pounder to drop freely without
significant energy losses from friction was also used in an early ma t
phase of the Changi East reclamation project. It comprises a 4 ót ¼
A0
clip holder, as shown in Figure 7. Figure 8 shows two types of

2500
350 1000 350 1850
260 1330 260
Eight 325 mm holes
320 mm bolt
350

170 350 170

100 mm Lifting eye

250 100 250

1850
1000
2500

350

Plan
Plan

Two 50 m thick steel plates

75 mm thick plate Eleven 50 m thick steel plates

1000 lifting head Four 250 mm dia. bolts
100 mm lifting eye
Eight 325 mm holes
320 mm bolt

500
75
100 275
100
200
Elevation

Elevation

Figure 5. Geometry of two pounders used at Changi (dimensions in mm)

Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al. 125
 (a) (b)
(a)
Figure 6. (a) Crane and (b) tripod used at Changi

(b)

Figure 8. Two types of pounder used at Changi

Figure 7. Lifting mechanism with clip holder

where st was determined based on the actual ‘at rest’
incremental pounder displacement for each drop. Chow et al. 5
also have demonstrated that the crater depth caused by
where m and A0 are the mass and contact area of the pounder pounder displacement can be estimated by applying wave
respectively. equation analyses.

Several different field measurements were made at the Changi Figure 11 shows a typical relationship between the peak impact
reclamation project to check the energy losses, surface vertical deceleration and the drop number. For a pounder of 18 t
displacement, and input stress and pore pressure due to dropping from a height of 10 m height for various numbers of
dynamic compaction in the sand. Figure 9 shows a comparison drops, it was found that the deceleration peaked at the third
of the theoretical and measured velocity of the pounder at the drop. After the sixth drop it was maintained at a constant
moment when it touches the surface. It can be seen that the calculated deceleration. This means that after the sixth drop the
measured velocity is typically about 80% of the theoretical effectiveness of subsequent compaction diminishes. The
velocity. pounder vertical displacement was also found to approach a
final stable value of 0.5 m shortly after the number of drops
5.3. Vertical displacement induced by pounding reached six.
The vertical displacement of the pounder, or the crater depth,
can be estimated. For example, the impact of an 18 t pounder 5.4. Pore pressure response in soil
dropped from 10 m height is about 1200 kPa. By integrating Pore water pressures in the soil mass at a depth of 5 m at two
the calculated velocity (v1 ), the displacement of the pounder locations, 2 and 3 m from the pounder drop point, were
(s t ) can be estimated from Equation 5. Figure 10 shows a measured during a trial test. It was found that a piezometer
comparison of the measured and estimated pounder vertical 2 m away recorded an excess pore pressure of about 140 kPa,
displacement, which is found to be comparable with that and that at 3 m away 60 kPa was recorded (Figure 12). These
calculated from the following equation pore pressures are almost ten times smaller than the impact
stresses imposed at the fill surface. The excess pore pressure
ð was found to increase in two cycles. First, the peak magnitude
5 s t ¼ v1 dt þ C d of excess pore pressure registered within less than 0.2 s, and
dissipated immediately. Then the excess pore pressure again

126 Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al.
 60 250
D⫽3m
Deceleration, a: g

Deceleration amax: g
40 200 D⫽7m
D ⫽ 10 m
20 150 D ⫽ 15 m
D ⫽ 20 m
0 100
0.2 0.4 0.6 0.8 1.0
⫺20
50
(a)

16 0
0 5 10 15 20
Drop number
Velocity: m/s

12
Calculated velocity
8 Theoretical maximum velocity Figure 11. Deceleration measurement after pounder drops
4 Measured maximum velocity

0
0.2 0.4 0.6 0.8 1.0
(b)
Section A–A
2·50 m
0·4

A
Displacement: m

7·0 m
3·65 m 1·0 m 2·35 m

m
35
4th 1st

2·
Calculated displacement

m
2·50 m
0·2

0
1·

P1
2nd drop 1st drop

m
65
Measured displacement

P2
3·

5·0 m
0 2nd 3rd
0 0·2 0·4 0·6 0·8 1·0
(c)

A
P2 P1

m
0
7·
1200 Plan
(a)
Stress: kPa

800 150
Excess pore pressure: kPa

P1
400 100

0 50
0 0·05 0·10 0·15 0·20 0·25 0·30 0·35 0·40
Vertical displacement: m
(d) 0
0 0·2 0·4 0·6 0·8 1·0
Time: s
⫺50
Figure 9. (a) Measured deceleration; (b) velocity
measurement; (c) displacement; (d) stress. Drop height ¼
10 m; drop mass ¼ 18 t; drop number ¼ 1 150
Excess pore pressure: kPa

P2
100

0·4
50
Pounder displacement: m

0·3
0
0 0·2 0·4 0·6 0·8 1·0
0·2 Measured
⫺50 Time: s
Calculated (b)
0·1

Figure 12. (a) Layout of measurement points; (b) measured

0 excess pore pressure against time
1 3 5 7 9 11 13 15 17 19 21
Drop number

Figure 10. Comparison of measured and calculated pounder

displacement other words, reduced void ratio of the granular soil after each
drop.

increased, to a smaller magnitude that generally took 2–3 min 5.5. Degree of improvement
to dissipate fully. In general it takes 3–4 min to complete the In the design of densification work, selection of the required
lifting and dropping process of dynamic compaction. Therefore, spacing and number of drops per point is essential in order to
for dynamic compaction in granular soil excess pore pressure achieve the specified density requirement, as both factors affect
may not be a significant issue. Excess pore pressures were the total compaction energy per unit surface area. Leonards et
measured after varied rounds of pounding: it was found that al. 10 reported that the degree of compaction correlated well
the excess pore pressure increases with increasing number of with the energy product, which equated to the total energy
drops. This can be attributed to increasing densification or, in applied per unit surface area times the energy per drop. Test

Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al. 127
 results at Changi East, as shown in Figure 13, support Several trial compaction tests were carried out with spacing
Leonard’s findings. The upper bound of maximum attainable between 6 and 10 m with various combinations of energy
cone resistance (qc ) is about 180 kg/cm2 (18 MPa) for the soil (Table 2). Based on these results, a spacing of 6 m by 6 m
in Changi, as illustrated in Figure 13. square was eventually used in the Changi East reclamation
project. In both methods two passes of 6 m by 6 m spacing
In the Changi East reclamation project, energy products of were applied, with drop point in the second pass at the centre
92 000 and 48 900 m2 t2 were applied in order to attain of the first pass. Therefore the net effective area of treatment
maximum cone resistances of 18 MPa (180 kg/cm2 ) and (i.e. the compacted area covered by one drop point after two
12 MPa (120 kg/cm2 ) respectively. The effective spacing of the passes) becomes 17.97 m2 , as shown in Table 4. Two methods,
pounding points and the numbers of drops per point calculated termed method A and method B, were applied using intensities
for particular pounders and cranes are shown in Tables 3 and of energy per drop of 575 and 300 t m respectively. These
4. To achieve the selected effective spacing, the sequence of energy levels were achieved using 23 t and 15 t pounders
pounding can be arranged into two or more passes to allow for dropped from 25 and 20 m and repeated for 5 and 10 drops
significant pore water pressure dissipation between passes. respectively. Methods A and B were applied respectively for
areas where the specified qc was 15 and 12 MPa. The resulting
qc profiles after compaction are shown together with those
prior to compaction in Figures 14 and 15 for methods A and B
Average qc in zone of max densification: MPa
0 5 10 15 20 respectively.
0 ⫻ 104
In the study, several combinations of types of pounder and
drop height under various soil conditions were used. For the
Energy ⫻ Energy/drop: m2 t2

4 ⫻ 104 same pounder and the same initial soil condition, the trend of
the relationship between the normalised crater depth and the
number of drops is similar, regardless of the drop height or the
8 ⫻ 104 location of drops. This can be seen in Figures 16 and 17.

Also, for the same drop weight with the same pounder and the
12 ⫻ 104
same drop height, the trend of normalised crater depth against
number of drops may vary if the initial soil condition is
different. Figure 18 shows the different trends observed after
16 ⫻ 104
the first pass and the second pass using the same energy per
drop. The crater depths are smaller with the same number of
Figure 13. Correlation between average cone resistance and
energy/m2 times energy/drop blows in the second pass, as the soil has been densified to a
certain degree after the first pass of pounding.

For pounders of different

Step Method geometries, the trend of
normalised crater depth may
A B
vary, even if pounders of
similar weight, dropping from
1 Required CPT qc : MPa 18 15
2 Pounder mass: t 23 15 the same height, are applied
3 Drop height: m 25 20 in the same type of soil. This
4: 233 Available energy per drop: t-m 575 300 can be seen in Figure 19. The
5: Figure 12 Energy/m2 3 Energy/Drop: t2 92 000 48 900 information gathered at
6: 5/4 Required energy/m2 : t-m/m2 160 163 Changi is useful for guiding
7: 4/6 Effective area/Drop: m2 3.59 1.84
site supervision, since the
trend of the normalised crater
Table 3. Calculation of effective area
depth against the number of
drops is independent of the
drop height for the same
Step Method pounder, as long as the base
A B area is the same. So direct
supervision can be kept to a
minimum, done solely by the
1 First pass effective: m2 36 36
2 Second pass effective: m2 36 36 measurement of crater depth
3 Net effective area:* m2 17.97 17.97 after pounding. Physical
4: Table 2 Effective area/Drop: m2 3.59 1.84 counting of the number of
5: 3/4 Required no. of drops/Pass 5 10 drops and observation of
drop heights may not be
* Net effective area after both passes (i.e. 6 m 3 6 m square spacing, two passes).
required during the field
Table 4. Calculation of required number of drops observation, once the trial
tests have been carried out on

128 Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al.
 Cone resistance: MPa Number of drops
0 5 10 15 20 25 0 2 4 6 8 10 12 14
0 0

1
0·05

Dc
2
Öwh 0·10
Location 2
Location 3
3
Location 4
0·15
Depth: m

Figure 16. Normalised crater depths at different locations for

5
a given energy per drop: w ¼ 15 t; h ¼ 15 m; Dc is in metres

6
Number of drops
0 2 4 6 8 10 12 14
7 0

8
0·05
Figure 14. Pre- and post-compaction qc profiles from method A
Dc
Öwh 0·10 w ⫽ 23 t, h ⫽ 25 m

w ⫽ 23 t, h ⫽ 12·5 m
Cone resistance: MPa
0 5 10 15 20 25 0·15
0

1 Figure 17. Normalised crater depths for different drop

heights: Dc is in metres

Number of drops
3
0 2 4 6 8 10 12 14
0
Depth: m

4
0·05
Dc
5
Öwh 0·10
1st pass
2nd pass
6
0·15

7
Figure 18. Normalised crater depths measured after first and
second passes for the same energy per drop: w ¼ 15 t;
8 h ¼ 20 m; Dc is in metres

Figure 15. Pre- and post-compaction qc profiles from method B

point often turns out to be the most compacted point, and the
location exactly under the pounder is actually the least
the particular granular soil type to establish the relationship compacted point.
between the normalised crater depth and the number of drops.
Chow et al. 17 proposed the use of an improvement ratio, which
5.6. Most compacted point is defined resistance expressed as function of the ratio of the
Most specifications in a dynamic compaction contract require peak angles of shear at any point in the soil and that
that verification for acceptance of the compaction works be underneath the pounder (˜
/˜
b ) ratio X/D. In their proposal
based on an in situ test performed at the centroid point, on the ˜
represent, the extent of improvement at a distance X from
assumption that soil at the centroid location is the least the centre of impact, ˜
b is that underneath the pounder and
compacted. However, if a correct spacing is used, the centroid D is the diameter of the pounder. According to Chow et al., 17

Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al. 129
 Number of drops measurements at the Changi project are shown in Figure 22.
0 2 4 6 8 10 12 14 The relative densities shown in the figures were calculated
0 based on the in situ measured densities and the maximum and
minimum dry densities measured in the laboratory for
representative samples. Note that, for the single pounding, the
0·05 degree of densification reduces as the distance from the
pounding point increases. However, for the multiple pounding,
the soil mass located at the centroid of the compaction grid is
Dc
generally well compacted owing to multiple pounding effects
Öwh
0·10 from all four adjacent pounding points if the correct grid
w ⫽ 15 t, h ⫽ 20 m, base area ⫽ 3·87 m2 pattern is applied. From Chow et al.’s predicted contours for
X/ of 2.5 and 4, 17 a greater extent of improvement was
w ⫽ 14 t, h ⫽ 20 m, base area ⫽ 2·25 m2
achievable at the centroid point, where X/D becomes smaller.
0·15
Therefore the centroid point may not be necessarily the ideal
point for verifying the extent of improvement.

Figure 19. Normalised crater depths measured with similar 5.7. Ageing effect
pounder weights and drop heights but different base areas
Dynamic compaction is often carried out in passes to allow for
pore pressure dissipation during the pause period. However,
because granular soil is highly permeable, dissipation of excess
the extent of improvement can be taken as zero at X/D . 3.5 pore pressure is generally quite rapid, particularly if fissures
and 1.0 at X/D < 0.5. develop. Therefore no significant increase in strength or
softening or ageing is expected after compaction. A minor
In the Changi East reclamation project, large numbers of CPTs increase in cone resistance may occur as a result of the slow
were carried out around and under the pounder locations for redeposition of soluble silica at grain contacts, which act as
6 m by 6 m and 7 m by 7 m square grid spacings, both with natural cementation. Increases in penetration resistance over
two passes of pounding. In these tests, the X/D ratios of the time after a densification treatment have been reported. 18–20
centroid locations are between approximately 1.5 and 1.75. Schlosser 21 has proposed a method for estimating increased qc
Selected comparative qc profiles for various distances from the after compaction. In the Changi East reclamation project CPTs
pounding points are shown in Figures 20 and 21. In general, at were carried out at 14 days and 3 months after compaction,
the location under the pounder point, the average qc value is but the change in the cone resistance is scarcely noticeable
the lowest and the soil is the least compacted, although a thin (Figure 23). Most ageing, if indeed such an effect exists, should
layer of highly compacted sand between 2 and 3 m deep is have taken place during the first 14 days.
present. The soils at the centroid point were found to be the
most homogeneously compacted with depth. Sample plots of 6. CONCLUSIONS
the contour of average values of relative density under the Field measurements and observations from dynamic
pounding point and the centroid point obtained from several compaction used for the densification of hydraulically placed

Cone resistance: MPa Cone resistance: MPa Cone resistance: MPa

0 5 10 15 20 0 5 10 15 0 5 10 15 20
0 0 0

⫺2 ⫺2 ⫺2

⫺4 ⫺4 ⫺4
Depth: m

⫺6
Depth: m
Depth: m

⫺6 ⫺6

⫺8 ⫺8 ⫺8

⫺10 ⫺10 ⫺10

⫺12 ⫺12 ⫺12

⫺14 ⫺14
⫺14 C1-2 (Blows: 15) C1-3 (Blows: 15)
C1-1 (Blows: 15)
B1-1 (Blows: 10) B1-2 (Blows: 10) B1-3 (Blows: 10)
A1-1 (Blows: 5) A1-2 (Blows: 5) A1-3 (Blows: 5)

(a) (b) (c)

CPT location
3
2
1 Weight ⫽ 23 t Height = 25 m

Figure 20. Cone resistance measured at various locations for 6 m 3 6 m grid spacing: (a) centroid point; (b) intermediate point;
(c) under the imprint

130 Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al.
 Cone resistance: MPa Cone resistance: MPa Cone resistance: MPa
0 5 10 15 20 0 5 10 15 20 0 5 10 15 20
0 0 0

⫺2 ⫺2 ⫺2

⫺4 ⫺4 ⫺4

Depth: m
Depth: m

Depth: m
⫺6 ⫺6 ⫺6

⫺8 ⫺8 ⫺8

⫺10 ⫺10 ⫺10

⫺12 ⫺12 ⫺12

⫺14 ⫺14 ⫺14

F1-1 (Blows: 15) F1-2 (Blows: 15) F1-3 (Blows: 15)
E1-1 (Blows: 10) E1-2 (Blows: 10) E1-3 (Blows: 10)
D1-1 (Blows: 5) D1-2 (Blows: 5) D1-3 (Blows: 5)
(a) (b) (c)

3
CPT Location 2 Weight ⫽ 23 t Height ⫽ 25 m
1

Figure 21. Cone resistance measured at various locations for 7 m 3 7 m grid spacing: (a) centroid point; (b) intermediate point;
(c) under the imprint

Cone resistance: MPa

35
9 35 35 0 5 10 15
35

75 75 0
55

35
El: mCD

55
8
55
35

35
7 35
2
6
⫺6 ⫺4 ⫺2 0 2 4 6
Distance from imprint point: m
4

9 35 35
Depth: m

70 70 50
El: mCD

6
50

8
50
50

50
35 35 25
7 25

6 8

⫺10 ⫺8 ⫺6 ⫺4 ⫺2 0 2 4 6 8 10
Distance from centre of grid: m
10

Figure 22. Contour of relative density under imprint and

centroid points from several tests at Changi 22 12

27-05-96 28-08-96

sandfill in the Changi East reclamation project have been Figure 23. No significant ageing after 14 days of compaction
described. Factors such as the degree of improvement, the most
compacted point and the ageing effect have been investigated,
based on field data collected in the project. mechanism. Suitable spacing, the required energy and the
number of passes can also be estimated from the same
The following conclusions can be drawn. empirical correlation.
(c) The geometry of the pounder and the initial soil conditions
(a) The centroid point within a compaction pattern is the most also affect the depth of the crater.
well-compacted point, and that directly under the pounder (d ) No significant ageing effect was found more than 14 days
is often the least compacted. Therefore the centroid point after compaction at the Changi East site.
should not be used as a quality control point. (e) Physical counting of the number of drops and observation
(b) The influence depth can be estimated by applying the well- of drop heights may not be required during the field
established generalised empirical correlation, although the supervision of dynamic compaction once trial tests have
components of the equation can be slightly different been carried out on a particular granular soil type to
depending upon the geometry of pounder and the dropping establish the relationship between the normalised crater

Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al. 131
 depth and the number of drops. The role of direct 11. VAN IMPE W. F. Soil Improvement Techniques and Their
supervision during actual work can in practice be replaced Evolution. A. A. Balkema, Rotterdam, 1994.
by the measurement of crater depth after pounding. 12. LUKAS R. G. Dynamic Compaction for Highway
Construction: Design and Construction Guidelines, US
Department of Transportation, 1986, Report No. FHWA/
REFERENCES RD-86/133.
1. JAMIOLKOWSKI M., LADD C. C., GERMAINE J. T. and 13. MAYNE P. W., JONES J. S. and DUMAS J. C. Ground response
LANCELLOTTA R. New development in field and laboratory to dynamic compaction. Journal of Geotechnical
testing of soils. Proceedings of the 11th International Engineering, ASCE, 1984, 110, No. 6, 757–774.
Conference on Soil Mechanics and Foundation Engineering, 14. PORAN C. J. and RODRIGUEZ J. A. Design of dynamic
San Francisco, 1985, 1, 57–153. compaction. Canadian Geotechnical Journal, 1992, 29, No.
2. SKEMPTON A. W. Standard penetration test procedures and 5, 796–802.
effects in sands of overburden pressure, relative density, 15. CHOW Y. K., YONG D. M., YONG K. Y. and LEE S. L. Dynamic
particle size and overconsolidation. Journal of Geotechnical compaction analysis. Journal of Geotechnical Engineering,
Engineering, 1986, 112, No. 3, 425–447 ASCE, 1992, 120, No. 8, 1141–1157.
3. SALGADO R., MITCHELL J. K. and JAMIOLKOWSKI M. Cavity 16. CHOA V., BO M. W., ARULRAJAH A. and NA Y. M. Overview
expansion and penetration resistance in sand. Journal of of densification of granular soil by deep compaction
Geotechnical and Geoenvironmental Engineering, ASCE, methods. Proceedings of the International Conference on
1997, 123, No. 4, 344–354. Ground Improvement Techniques, Macau, 1997, pp. 131–
4. CUBRINOVSKI M. and Ishihara K. Empirical correlation 140.
between SPT N-value and relative density of sandy soils. 17. CHOW Y. K., YONG D. M., YONG K. Y. and LEE S. L. Dynamic
Soils and Foundations, 1999, 39, No. 5, 61–71. compaction of loose granular soils: effect of print spacing.
5. CHOW Y. K., YONG D. M., YONG K. Y. and LEE S. L. Dynamic Improvement of granular soils by high energy impact.
compaction of loose sand deposits. Soils and Foundations, Journal of Geotechnical Engineering, ASCE, 1994, 120, No.
1992, 32, No. 4, 93–106. 7, 1115–1133.
6. BALDI G., BELLOTTI R., GHIONNA V., JAMIOLKOWSKI M. and 18. DEBATS J.-M. and SIMS M. Vibroflotation in reclamation in
PASQUALINI E. Interpretation of CPTs and CPTUs. Part 2: Hong Kong. Ground Improvement, 1997, 1, 127–145.
Drained penetration. Proceedings of the 4th International 19. MITCHELL J. K. Soil improvement: state of the art report.
Geotechnical Seminar, Singapore, 1986, 143–156. Proceedings of the 10th International Conference on Soil
7. WOODS R. D. Severing of surface wires in soils. Journal of Mechanics and Foundation Engineering, Stockholm, 1981,
Soil Mechanics and Foundation, ASCE, 1968, No. SM4, 4, 509–565.
DSI-098. 20. OSHIMA A. and TAKADA N. Relation between compacted
8. MENARD L. and BROISE Y. Theoretical and practical aspects area and ram momentum by heavy tamping. Proceedings of
of dynamic consolidation. Géotechnique, 1975, 25, No. 1, the 14th International Conference on Soil Mechanics and
3–18. Geotechnical Engineering, Hamburg, 1997, 3, 1641–1644.
9. LUKAS R. G. Dynamic compaction: engineering 21. SCHLOSSER F. Theme lecture: Soil improvement and
considerations. In Grouting, Soil Improvement and reinforcement. Proceedings of the 14th International
Geosynthetics, ASCE Geotechnical Special Publication No. Conference on Soil Mechanics and Geotechnical
30, Vol. 2, ASCE, New York, 1992, pp. 940–953. Engineering, Hamburg, 1997, 4, 2445–2466.
10. LEONARDS G. A., CUTTER W. A. and HOLTZ R. D. Dynamic 22. NYM. Institution characterization of reclaimed sandfill
compaction of granular soils. Journal of the Geotechnical with particular reference to dynamic compaction, Thesis
Engineering Division, ASCE, 1980, 106, No. 1, 35–44. Mangory Technology University, Singapore, 2002.

What do you think?

To comment on this paper, please email up to 500 words to the editor at journals@ice.org.uk
Proceedings journals rely entirely on contributions sent in by civil engineers and related professionals, academics and students. Papers
should be 2000–5000 words long, with adequate illustrations and references. Please visit www.thomastelford.com/journals for author
guidelines and further details.

132 Ground Improvement 162 Issue GI3 Densification of granular soil by dynamic compaction Bo et al.

View publication stats