13 Compaction

Source: GEOTECHNICAL ENGINEERING
13 Compaction
13.1 HISTORICAL BACKGROUND
13.1.1 Paths of Least Resistance

Compaction is the oldest and most common method for soil stabilization, at first
being accomplished automatically as herd animals followed the same trails—some
of which are followed by the routes of modern highways.
Compaction has some obvious benefits. A compacted soil is harder, and can
support more weight and shed water better than the same soil that has
not been compacted. The extensive road system of the Roman Empire consisted
of stones laid flat on top of compacted soil. Roman road builders even
attempted to compact soil with elephants, which is not very efficient because
elephants prefer to step in the same tracks. The modern version of the elephant
walk is a ‘‘sheepsfoot roller’’ that emulates tracking by feet or hoofs, but does so
randomly.
13.1.2 The Twentieth Century

Compaction was largely intuitive and hit-or-miss until the 1930s when the
Los Angeles county engineer, R. R. Proctor, devised a laboratory test that
allowed systematizing and optimizing the compaction process. Proctor’s
procedure revolutionized the construction of roads and earth dams, respectively
the longest and the largest human-made structures in the world. Without these
control methods, costs would be higher and failures commonplace.
280 Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2007 The McGraw-Hill Companies. All rights reserved.
Any use is subject to the Terms of Use as given at the website.
Compaction
Compaction 281
13.2 PROCTOR’S DISCOVERY
13.2.1 A Direct Approach

Proctor recognized that moisture content is a major variable influencing
compaction, and therefore devised a test that isolates the role of moisture content
by holding other variables constant. Compaction was accomplished in the
laboratory by ramming a soil into a standardized steel mold using a standardized
amount of energy, and it was discovered that there is an optimum moisture
content that gives the highest density for a given compactive effort. If a soil is too
dry, more energy is required to attain a particular density, and it is too wet, no
amount of energy will compact it to the same density. The optimum moisture
content is designated by OMC.
Proctor’s test was designed to simulate the action of a sheepsfoot roller. An earlier
field standard was if a roller ‘‘walked out’’ of a soil layer as it was compacted from
the bottom up, and some contractors still use this as a criterion for adequate
compaction. However, ‘‘walking out’’ measures bearing capacity, not density, and
a roller will walk out of a dry soil much more readily than a wet one. In a properly
compacted soil the air content is so low that the soil has very little capacity for
additional water.
The hand tamper devised by Proctor is shown in Fig. 13.1. The original tamper,
shown on the left, involved dropping a 5.5 lb (24.5 N) weight 12 in. (0.305 m) to
compact soil in three layers in a steel mold, using 25 blows per layer. The volume
of the mold is 1/30 ft3 (944 cm3). The lifting height is controlled by housing the
rammer inside a pipe; the bottom of the pipe is placed in contact with the soil and
the rammer lifted by hand until it contacts the top, when it is dropped.
13.2.2 AASHTO T99 Procedure

Four or five tests are performed with soil at different moisture contents, and
if the soil is not re-used (ASTM 698) about 5 lb (2.25 kg) will be required.
Some laboratories re-use soil after it has been thoroughly pulverized between
tests, but this is not an ASTM-approved procedure and if done should be so stated
in the test report. Soils containing significant amounts of mica should not
be re-used as the mica breaks down. Soils containing halloysite clay mineral
are irreversibly changed by drying, and should not be air-dried prior to testing.
Air-dried soils are moistened and allowed to age overnight to allow clay
expansion. The soil is compacted in three approximately equal layers using a
collar temporarily attached to the top of the mold to hold the soil. The collar is
removed, the top of the soil struck off with a knife or straight edge, and the net
weight and moisture content of the soil determined. The test then is repeated with
Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Compaction
282 Geotechnical Engineering
Figure 13.1
Hand-operated
Proctor and
modified Proctor
density rammers.
The mold holding
the soil is
standardized, but
the modified
Proctor uses a
heavier rammer
that is dropped
farther on thinner
soil layers. The
modified test was
developed to meet
the need of air-
field construction.
more or less added water. The AASHTO T180 (ASTM D-1557) modified Proctor
procedure is similar except for a higher ramming energy and the use of five instead
of three layers (Fig. 13.1).
13.2.3 Results
Figure 13.2 shows some typical results from standard compaction of a silty sand
(SM) soil. A smooth curve is drawn through the data points, from which the
optimum moisture content (OMC) and maximum compacted density are read
from the graph.
An important feature of all such graphs is the theoretical ‘‘zero air void line,’’
which is a curve that is calculated from the soil unit weight and specific gravity of
the solid components. As can be seen, one leg of the density curve approximately
parallels the zero air void line, indicating a persistent air content. It is physically
impossible for the curve to cross the line, and the horizontal departure from the
line indicates the amount of air that still is remaining in the soil after compaction.
Measurements of air permeability indicate that as a soil is compacted and air
voids are constricted, permeability decreases so that the last bit of air is retained in
the fine pores.

Compaction
Compaction 283
Figure 13.2
Compacted
moisture-density
relationship for an
SM soil and a
standard Proctor
compactive effort.
Modified Proctor
compaction moves
the curve upward
and to the left to a
higher density and
lower OMC.
13.2.4 Not Lubrication

It once was assumed that the optimum moisture content was optimum because it
provides just enough water to lubricate the soil grains. However, the effect of
water on non-clay minerals is to increase instead of reduce sliding friction. As
previously discussed, water weakens ionic bonding in a mineral surface so that it
scratches more readily, increasing sliding friction.
The anti-lubricating effect of water also is shown by compaction curves such as

Fig. 13.3, which is the same as in the previous figure but with additional data on
the dry side of the OMC. Compacting the dry soil resulted in a higher density
when the soil was dry than was obtained at the OMC.
The explanation for this behavior can be seen in Figs. 11.6 and 11.7 and the
accompanying discussion in Chapter 11: annular rings of water linking soil grains
Figure 13.3
Additional test data
for the soil
in Fig. 13.2
demonstrates that
friction and
capillary tension
define creates
optimum moisture
contents for
compaction.

Compaction
have negative pressure that draws the grains together, increasing sliding friction.
Unless augmented by the presence of clay minerals, capillary tension and negative
pressure disappear at zero moisture content. Capillary tension also disappears as
the soil moisture content approaches saturation, giving the two highest compacted
densities in a granular soil.
13.3 ZERO AIR VOIDS LINE
The zero air voids line normally is included on moisture-density (or moisture-unit
weight) plots because it defines the edge of prohibited territory, which is the
saturation moisture content for a particular density. If voids are completely filled
with water there is no air left to be expelled. The saturated unit weight can be
calculated from a block diagram or from the following formula:
max ¼ Gs w =½1 þ Gs ðm=100Þ ð13:1Þ
where max is the zero air voids unit weight, w is the unit weight of water, Gs is
specific gravity of the soil grains, and m is the moisture content in percent.
Solutions for several values of Gs and m are shown in Table 13.1 for plotting on
compaction curves.
13.4 COMPACTING IN LAYERS
13.4.1 A Critical Requirement

Compaction requires time and energy, and it may seem a huge bother to have
to spread and compact a soil in a series of layers. Occasionally an inexperi-
enced operator will decide that it is not worth the bother, with grim and
predictable consequences. The reason for this requirement is because pressure
applied over a limited area at the ground surface spreads and dissipates with
depth. This is illustrated in Fig. 13.4, where stress contours are drawn on the
basis of elastic theory and decrease as the square of the distance from the
contact area.
13.4.2 A Virtual Compactor

As a soil is compacted layer on layer, an important function of a previously
compacted lower layer is to in effect cause stresses to ‘‘bounce up’’ and intensify.
This is illustrated in Fig. 13.5. The argument is as follows: if the previously
compacted soil layer is sufficiently rigid that no deflection occurs, its influence can
be exactly duplicated by replacing the lower layer with vertical uplift pressures
in the same pattern as the downward pressures caused by the weight of the
compactor. This ‘‘virtual compactor’’ therefore doubles vertical pressure in the
lower part of the soil layer that is being compacted.

Compaction
Compaction 285
Specific gravity of solids Table 13.1
Moisture content, % 2.5 2.6 2.65 2.7 3 Densities and

moisture contents
0 156.0 162.2 165.4 168.5 187.2 that define zero air
5 138.7 143.6 146.0 148.4 162.8 voids lines from
10 124.8 128.8 130.7 132.7 144.0 eq. (13.1). The upper
15 113.5 116.7 118.3 119.9 129.1 part of the table is in
20 104.0 106.7 108.1 109.4 117.0 lb/ft3, the lower part
25 96.0 98.3 99.5 100.6 107.0 in kN/m3
30 89.1 91.1 92.1 93.1 98.5
35 83.2 84.9 85.8 86.6 91.3
40 78.0 79.5 80.3 81.0 85.1
45 73.4 74.8 75.4 76.1 79.7
50 69.3 70.5 71.1 71.7 74.9
60 62.4 63.4 63.8 64.3 66.9
70 56.7 57.5 57.9 58.3 60.4
80 52.0 52.7 53.0 53.3 55.1
90 48.0 48.6 48.9 49.1 50.6
100 44.6 45.1 45.3 45.5 46.8
Specific gravity of solids
Moisture content, % 2.5 2.6 2.65 2.7 3
0 24.51 25.49 25.98 26.47 29.41

5 21.78 22.56 22.94 23.32 25.57
10 19.61 20.23 20.54 20.84 22.62
15 17.82 18.34 18.59 18.84 20.28
20 16.34 16.77 16.98 17.19 18.38
25 15.08 15.45 15.63 15.80 16.81
30 14.00 14.32 14.47 14.62 15.48
35 13.07 13.34 13.48 13.61 14.35
40 12.25 12.49 12.61 12.73 13.37
45 11.53 11.75 11.85 11.95 12.51
50 10.89 11.08 11.17 11.26 11.76
60 9.80 9.96 10.03 10.10 10.50
70 8.91 9.04 9.10 9.16 9.49
80 8.17 8.28 8.33 8.38 8.65
90 7.54 7.63 7.67 7.72 7.95
100 7.00 7.08 7.12 7.15 7.35

Compaction
Figure 13.4
Compaction
pressure
dissipates as the
square of the
distance from the
contact area, so
soil must be
compacted in thin
layers or ‘‘lifts.’’
Figure 13.5
Concept of a
‘‘virtual compac-
tor’’ that doubles
the pressure in a
compacting layer
over a rigid
previously
compacted layer.
13.4.3 The Importance of Roughness

Vertical stress is not the only requirement for compaction, as soil also must be
confined laterally or it will simply squeeze out. This effect can be demonstrated by
pouring a small mound of dry sand on a glass plate and pressing it down with the
palm of a hand; the sand squeezes out laterally. The experiment then is repeated
with the sand supported on a sheet of sandpaper, and friction with the sandpaper
reduces the amount of spreading so that more force can be applied to the sand and
it is compacted.
Without a rough base, major principal stress vectors radiate outward from an
applied surface load and tend to push the soil laterally, as shown at the right in
Fig. 13.6. At the left in the figure, frictional restraint shown by the horizontal
arrow tends to confine the soil and prevent spreading.

Compaction
Compaction 287
Figure 13.6
Frictional contact
with a rough base
layer helps to
confine stress and
aids compaction.
Principal stress
vectors on the
right, without a firm
base, are from
Jürgenson (1934).
Frictional restraint also can be added on top of the compacting layer by

substituting a grid or mesh pattern for a smooth-wheel roller, but it is at the
bottom of the layer that restraint is most needed.
13.4.4 Tamping-Foot Rollers

Tamping-foot rollers have projections that intensify contact stress. The objective
is to punch through and prevent compaction of the upper soil in each layer so that
on repeated passes of the roller the soil compacts from the bottom up (Fig. 13.7).
The wedge shape of the feet on tamping-foot rollers now is preferred over the
knobby shape of the sheepsfoot as being less destructive to the compacted soil as
the projections or feet walk out.
Design can be based on the same theory that is used for design of shallow
foundations, presented later in this book, and in at least one case a roller weight
was designed based on this theory because it is necessary for the bearing capacity
to be exceeded by the individual projections for this type of roller to be effective.
Bearing capacity, or support for a surface load, depends on several factors
Figure 13.7
Foot pressure
exceeds the
bearing capacity of
the uncompacted
soil, penetrating
through to compact
the soil layer from
the bottom up.

Compaction
including the size of the contact area, the soil cohesion, and its angle of internal
friction.
Successive roller passes on clay compact the soil and increase its cohesion, which
increases the bearing capacity so that the roller walks out. Without the contri-
bution of cohesion, as in cohesionless sands, there is very little walk-out so the
projections only knead and stir the soil. Sands are compacted with smooth-wheel
rollers or plates, and can be aided by vibration.
13.4.5 Proof Rolling

After a soil is compacted it may be subjected to ‘‘proof rolling’’ with a heavy roller
or even a heavily loaded truck. Contrary to popular opinion, a track vehicle such
as a bulldozer is not appropriate for proof rolling because it is designed to spread
pressure and be supported on soft soils. The soil pressure during proof rolling
equals the tire pressure in the vehicle.
Proof rolling can reveal pockets of soil that are insufficiently compacted or are
compacted too wet, but the reliability of this procedure depends on the soil
moisture content. If the moisture content is on the dry side of the OMC the soil
will support a heavier load because of capillary tension or ‘‘apparent cohesion’’
that is destroyed when the soil becomes wet.
13.4.6 ‘‘Oreo Cookie Effect’’

Even with a firm base and compaction in layers, stress distribution is not uniform
through the depth of a single layer, which results in a nonuniformity that has been
called the ‘‘Oreo cookie effect’’ (H. Allender, personal communication). Density
measurements therefore should involve an entire layer or multiple layers in order
to obtain reliable averages.
13.5 OUT GOES THE AIR, NOT THE WATER
13.5.1 A Critical Distinction

Compaction means to squeeze air out of the soil, which is rapid and is in contrast
to consolidation, which means to squeeze out water. The viscosity of water is about
1 102 poise, or about 50 times that of air. Compaction therefore is rapid and
consolidation is at least 50 times slower. A saturated soil can be consolidated but
not compacted.
As compaction proceeds and air is squeezed out, the soil eventually will reach the
point where there is very little air left, and the soil is close to saturation as a result
of compaction. The remaining air is trapped in tiny pockets, and compaction stops

Compaction
Compaction 289
regardless of the number of additional passes of a compactor. If processing

continues past that point, the result is called overcompaction.
A Conflict over Nomenclature

‘‘Consolidation’’ was first described by a geologist, Sir Charles Lyell, in a
textbook published in 1851. Lyell noted that sediments on the sea bottom are
compressed and consolidate under their own weight and not from the weight of
hundreds or thousands of feet of sea water. Despite this prior definition by an
eminent geologist, geologists in the petroleum industry call the process
‘‘compaction’’ and refer to shale as being ‘‘highly compacted.’’ The Oxford
English Dictionary favors the engineering usage, defining consolidation as
bringing together and compaction as packing together. It is important that
geotechnical engineers not confuse the two terms and their distinctive meanings.
13.5.2 Overcompaction
A continued application of compactive effort after soil reaches near-saturation
point not only wastes energy, but the energy is redirected into shearing and
remolding the compacted and nearly saturated soil. Shearing smears clay particles
so that they are oriented parallel to the shear surfaces and permanently weaken
the soil. This must be carefully guarded against, as the soil then must be dried and
pulverized before it can be recompacted.
The effect of overcompaction on strength of a clay is shown in Fig. 13.8, which

shows stress-strain curves for two samples of the same soil at the same moisture
content and same density, the difference being that the weaker soil was compacted
at a higher moisture content than the stronger one.
Clues to overcompaction are vibrations that can be felt under foot, and an
occasional ‘‘thunky’’ sound when stomped on. A bulge of soil sometimes may be
Figure 13.8
Stress-strain
curves show the
reduction in
strength after a soil
becomes
dispersed through
overcompaction
and shearing.
(After Seed and
Chan 1959, with
permission of the
American Society
of Civil Engineers.)

Compaction
seen moving as a wave ahead of tires on heavy equipment, increasing energy

consumption because the equipment always is moving up out of a hole of its own
making. A relatively dry crust can develop that if broken through causes miring
of the equipment. Overcompaction commonly occurs on construction haul roads
on wet, fine-grained soils after repeated passes of heavy construction and earth-
moving equipment.
The remedy for overcompaction obviously is not more compaction. Instead

the soil must be ripped out and aerated or kiln-dried. If the soil contains active
clay minerals, in particular smectite, an alternative procedure that may be less
expensive and saves time is to add several percent hydrated lime or lime-rich fly
ash, which raise the plastic limit and the optimum moisture content.
13.6 COMPACTION OF GRANULAR SOILS
13.6.1 Free-Draining Soils

Water content is not critical for compaction of free-draining soils such as gravel
and crushed rock. Vibratory compaction is most effective with granular soils
including sands, as vibrations create oscillations of intergranular contact stresses
and friction.
13.6.2 Bulking and Flooding

Intergranular friction caused by capillary tension causes a granular soil to ‘‘bulk
up’’ or decrease in density after it is disturbed. Thus, a bucket filled with dry sand
will weigh more than a bucket filled with wet (not saturated) sand, even though
the latter includes some additional weight from water. This effect is called bulking,
and must be taken into account during batching operations for materials used in
concrete.
Bulking also occurs when sand or gravel is spread as a base for a floor or
foundation. Since the cause is capillary action, the density can be increased
by saturating the soil with water. This has led to a somewhat haphazard practice
of ‘‘compacting’’ granular soils by flooding with water, but it should be
emphasized that the result is equilibrium with the self-weight of the soil, and not
to any additional load. Flooding induces consolidation, not compaction. A soil
that is consolidated to equilibrium with its self-weight is still compressible under
an additional load. It therefore is said to be normally consolidated. The low
densities achieved by flooding may go unnoticed with light structures as
settlement is almost immediate, hopefully occurring before the concrete has set.
Although flooding is better than no treatment at all, it is not a substitute for
compaction.

Compaction
Compaction 291
13.6.3 Relative Density Test

The Proctor density test originally was intended to simulate the action of a
sheepsfoot roller, but as previously mentioned, sheepsfoot rolling is not effective
in sandy soils containing little clay. Furthermore, the coarser the soil, the less
effective are capillary forces at preventing densification, so moisture content
becomes less critical.
Because sands are most effectively compacted with a vibratory roller or plate
vibrator, a different reference density is used based on densities achieved with
a vibrating table (Fig. 13.9). Both maximum and minimum densities are
determined, and the position of the compacted density in relation to these two
limits is called the relative density, which is expressed as a percent.
The formula for relative density is

Dd ¼ 100ðemax eÞ=ðemax emin Þ ð13:2Þ
where emax and emin are the maximum and minimum void ratios determined from
the test, and e is the void ratio of the soil.
A more convenient form of this equation that uses unit weights is

max ð min Þ
Dd ¼ 100 ð13:3Þ
ðmax min Þ
where max and min are, respectively, the maximum and minimum dry unit
weights, and is the dry unit weight of the soil.
13.6.4 Defining Minimum and Maximum Density

A minimum density, min, is determined by oven-drying and pulverizing the soil to
a single-grain structure, then slowly pouring it through a funnel into a container
of known volume. The soil must be dry to prevent bulking.
Figure 13.9
Vibrating table for
obtaining a
standardized
maximum density
of granular soils.

Compaction
A maximum density, max, is more difficult to define because it depends on

the compaction mechanics and confinement. The determination has been
standardized and involves putting loose soil into a cylindrical mold that is set
on a vibrating table (Fig. 13.9). A surcharge weight is placed on top to confine the
soil, and the table is actuated with electromagnetic vibrations at 50 or 60 Hz
depending on availability. The vibration time is adjusted to give a uniform
number of cycles.
Tests have shown that max also depends on the geometry of the equipment, the
applied load, and the acceleration ratio, which is the ratio of acceleration from the
vibrations to that of gravity, so the conditions under which the test is performed
must be carefully controlled (ASTM Designation D-4253).
Minimum and maximum densities often are determined with the soil wet and after
oven-drying, and the higher sets of values used for determining relative density.
While capillary effects reduce the efficiency of compaction, they also tend to hold
grains together after densification instead of letting them fly apart from the
continuing vibrations.
13.6.5 Applications of Relative Density

The relative density determination is limited to soils with 15 percent
or less noncohesive fines passing the No. 200 sieve. Soils containing 5 to
15 percent noncohesive fines also may be impact-compacted for comparative
purposes.
Example 13.1
A field density determination gives ¼ 110 lb/ft3 (17.3 kN/m3). Laboratory tests give
max ¼ 127 lb/ft3 and min ¼ 102 lb/ft3 (19.95 and 16.0 kN/m3). What is the relative density
of the field soil?
Answer:
127ð110 102Þ
Dd ¼ 100 ¼ 37%
110ð127 102Þ
13.7 COMPACTION SPECIFICATIONS FOR FINE-GRAINED SOILS
13.7.1 The Proctor Relationship

Most compaction specifications are written around an optimum moisture content
and maximum density from a Proctor density test (Fig. 13.2). The moisture
content of a soil in the field is adjusted by adding water or aerating and drying
until it is within a specified range of the optimum moisture content.

Compaction
Compaction 293
A 100 percent density is difficult to achieve unless the moisture content is on the
dry side of the OMC, so the density requirement in the field normally is set at
95 percent of the maximum density obtained in the laboratory test. A target area
may be defined by a moisture content within 2 percent of the OMC and a
compacted density to equal or exceed 95 percent of the maximum is indicated by
the specification area in Fig. 13.10. This area can be modified to meet special
requirements.
Figure 13.10 also shows ‘‘compaction growth curves’’ for different percentages
of the standard compactive effort. As the compactive effort increases, the OMC
is reduced along a line that roughly parallels the zero air voids curve.
13.7.2 Too Dry for Comfort

Proctor discovered a problem that still sometimes is overlooked or ignored—that
compacting a soil on the dry side of the OMC leaves too much air in the soil. The
soil remains permeable to water, and when saturated may become so weakened
that it densifies or collapses under its own weight—even though the compacted
density can meet a 95 percent specification requirement. Some potential
consequences are illustrated in Fig. 13.11. The problem is most common in
trench backfill where the compaction moisture content often is not monitored or
controlled. As a result a strip of pavement that is over a sewer running down the
middle may be unsupported from one manhole to the next.
A particularly serious consequence of settlement of soil backfill in a trench is if the

soil weight is carried by a pipe crossing through the trench so that the pipe breaks.
Figure 13.10
Compaction
growth curves for a
silty clay. The
shaded area is for
a specification
requirement of
95 percent
minimum density
with the moisture
content within
2 percent of the
OMC. (Data
courtesy of Prof.
J. M. Hoover.)

Compaction
Figure 13.11
Serious problems caused by collapse of soil compacted on the dry side of the OMC: (a) earth dam
piping; (b) pavement collapse; (c) fracture and leak of a gas or water pipe: gas flowing through loose
trench backfill soil and into a building can cause fire or explosion.
Broken water mains can cause locally severe underground erosion, so the
dramatic collapse of the street is sure to make the 6 o’clock news.
Gas line breaks can be more tragic, as escaping gas follows the path of least
resistance through loose trench backfill into a building or basement, where it can
cause a fire or explosion. Pipes crossing through soil in a trench should be
supported by some type of beam.
Collapse of soil within an earth dam can lead to serious consequences by allowing
‘‘piping,’’ or flow of water through the dam, leading to erosion and failure.
13.7.3 Preventing Collapse of Compacted Soil

Figure. 13.12 shows volume changes after compacting and wetting. The area of
zero volume change is shown shaded, and for the most part falls within
a box drawn for the specification limits of OMC 2 percent, and a compacted
unit weight between 95 and 100 percent of the standard Proctor density. Thus,
compaction within these specification limits normally will result in a soil that is
relatively stable against future volume changes. An exception is expansive clays,
which are discussed later in this chapter.
Where even a small amount of collapse cannot be tolerated, as in the case of

the core material in an earth dam, the specification limits for moisture content
often are increased to between the OMC and the OMC plus 2 percent to as high as
4 percent. However, compaction on the wet side weakens the soil and makes it
more susceptible to overcompaction and development of excess pore water
pressure.

Compaction
Compaction 295
Figure 13.12
Contours show volume changes from wetting of a compacted silty clay. Compaction within usual
specification limits (shown by the heavy arrows) results in negligible volume change if the soil is
not highly expansive. (After Lawton et al., 1992.)
Figure 13.13
Relation of
compaction energy
to moisture
content,determined
from energy data
along the
95 percent density
line in Fig. 13.12.
13.7.4 Compacting for Low Permeability

In applications such as in the confined core of an earth dam where low
permeability is more important than strength, the moisture content requirement
usually is 0 to þ2 percent or þ4 percent above the OMC.
Compacting on the wet side of the OMC also can save time and fuel by reducing
the required compactive effort. In Fig. 13.13 about 65 percent of the standard

Compaction
compactive effort is required for soil to reach the specification density at the
optimum moisture content. This is reduced to about 50 percent at 2 percent above
optimum. However, compacting at above the OMC narrows the target range and
increases the possibility for overcompaction.
13.7.5 Relaxing Density Requirements

For applications where the compacted soil is not heavily relied on for strength and
where some settlement is permissible, such as for landscape grading that does not
involve steep slopes or high retaining walls, the density requirement may be
reduced to 90 percent of the maximum. In the example in Fig. 13.13, 90 percent
density could be achieved with about 40 percent effort at the OMC, or compared
with 65 percent for 95 percent density. The same criterion applied to the soil in
Fig. 13.12 would result in about a 3 percent reduction in volume upon saturation,
so saturating the lower 1 m (3 ft) of a fill placed by this criterion would generate a
settlement of 30 mm or a little over an inch.
13.7.6 Modified Proctor Density

The trend toward a higher density and lower optimum moisture content conti-
nues with modified Proctor compaction, which moves the density curve higher
and farther to the left (Fig. 13.14). The energy for laboratory compaction is
56,00 ft-lb/ft3 (2700 kN-m/m3) for the modified test compared with 12,400 ft-lb/ft3
(600 kN-m/m3) for the standard test, an increase of over 450 percent. The
modified density requirement therefore is energy-intensive and is used only in
special circumstances.
Figure 13.14
Modified Proctor
density is
achieved with a
higher compactive
effort and a lower
moisture content,
and results in a
denser, stronger,
and less
compressible soil.
It is used in
heavy, duty
applications such
as airport landing
strips and
runways.

Compaction
Compaction 297
Question: Can a modified Proctor density be specified with a standard Proctor

moisture content? What about the reverse situation?
Answer: No and no. Why?
13.8 SPECIAL PROBLEMS WITH COMPACTION OF EXPANSIVE CLAYS
13.8.1 So There’s a Problem; Can We Cut a Deal?

Expansive clays do not give up easily. As discussed in the preceding chapter,
Section 12.8.9, expansion may be reduced or prevented by burying under 1 m (3 ft)
of compacted nonexpansive clay (n.e.c.) that can be manufactured by adding
and mixing several percent hydrated lime. What other methods, if any, are
available?
13.8.2 Adjusting the Compaction Moisture Content

Logically, if expansive clay is compacted wet enough it should be fully expanded,
but unfortunately it then will shrink if it is dried out. On the other hand,
compaction on the dry side of optimum will leave it permeable and subject to
expansion on wetting. The latter is illustrated in Fig. 13.12, so the moisture
content for compaction of an expansive clay often is specified to be between the
OMC and OMC þ 2 percent.
The two options, compacting on the dry side or on the wet side of the optimum
moisture content, are further illustrated by data in Fig. 13.15 for an expansive
sandy clay. From the upper graph it may be seen that expansion is over 2 percent
if the soil is compacted to maximum density at the optimum moisture content, is
reduced about one-half if the soil is compacted at 2 percent above optimum, and
becomes zero if compacted at 4 percent above optimum. However, such a soil will
be greatly weakened by dispersion and possible overconsolidation, and if it dries
out it will shrink excessively. On the other hand, if the soil is compacted at
2 percent below the OMC, shrinkage is reduced one-half compared to that which
will occur if the soil is compacted at the OMC, but shrinkage is not completely
eliminated.
Most problems with expansive clays occur when they are under low pressure
and subject to either an increase or a decrease in the moisture content, so
the compaction moisture content can be selected in anticipation of future
changes. For example, moisture tends to accumulate in soil that is sealed off from
the atmosphere, such as under pavements or floor slabs, so to reduce the
likelihood of volume change, expansive clay may be compacted on the wet side of
the OMC.

Compaction
Figure 13.15
Controlling the
volume change of
expansive clay by
adjusting the
compaction
moisture content.
(After Seed and
Chan, 1959, with
permission of the
American Society
of Civil
Engineers.)
Seasonal moisture changes in expansive clay can be reduced by (a) periodically

watering around the outer perimeter of a house or other structure, (b) extending a
concrete slab or membrane horizontally outside of the structurally loaded areas,
or (c) installing a vertical moisture barrier around the perimeter. Complete
isolation is difficult because of the high suction forces exerted by dry expansive
clay and the disruption caused by shrinkage cracking. In cooler climates such as in
Canada, interior heating may drive moisture out of expansive clay during the
winter, causing shrinkage that is very difficult to control.
Because some variations in moisture content are unavoidable, another option

that may be economical for small structures is to support them on a structurally
designed slab-on-grade that can bridge between support areas. Larger structures
can be supported on deep foundations that extend below the active zone
of the clay, in which case floor joists are supported on bails of straw or cardboard
‘‘crush boxes’’ so that the beams will not be lifted off the foundations, and
floors are structurally supported high enough off the ground to allow for clay
expansion.
A procedure that should not be overlooked is to replace the upper layer

of expansive clay with a soil that is nonexpansive and treat it with hydrated lime.
Treatment of the upper 1 ft (0.3 m) can reduce expansion, of the upper 2 ft (0.6 m)
can reduce it even further, and of the upper 3 ft (1 m) may bring it under control.

Compaction
Compaction 299
13.9 STRENGTH AND MODULUS OF COMPACTED SOIL SPECIMENS
13.9.1 Another Emphasis

Compaction specifications not only are intended to control future volume changes
of a soil, they also may be intended to increase the soil strength. It sometimes
is assumed that higher density means higher strength, but this trend is trumped by
the moisture content: compacting on the dry side of the OMC may leave the soil
such that it can collapse under its own weight when wet with water, or too
vigorous compaction on the wet side can shear and remold the soil.
Most compacted soil is used in embankments for roads, highways, or earth dams.
In these applications, soil in the embankments not only should resist volume
changes, it must have sufficient shearing strength that side slopes are stable and do
not develop landslides. Railroad embankments, many constructed a century ago,
usually were not compacted and, significantly, still require continual maintenance
and repair. It therefore is important that the shear strength characteristics of the
compacted soil should be determined.
Compacted soils must be strong enough to support structures and minimize

settlement that not only can affect the integrity of the structure but also that
of connecting utility lines; a broken gas or water line or sewer that slopes the
wrong way is more than just an inconvenience.
13.9.2 Strength of Compacted Soils

Since compacted soil is a manufactured, quality-controlled product, its suitability
for particular applications often is based on experience with the same or similar
soils under similar applications. However, if a compacted soil is to be used for
founding heavy structures that are outside of the range of experience, the soil
strength and compressibility should be measured at the worst-case densities and
moisture contents that are within specification limits, and those values used in
design. After a soil is compacted it can be drilled, tested, and analyzed using
methods presented in later chapters.
The simplest option is to subject compacted density specimens to an unconfined

compression test, where ends are loaded until the specimen fails. This test
is discussed in Chapter 18.
13.9.3 Laboratory ‘‘K Test’’

Specimens compacted for laboratory density determinations usually are
discarded, but they can easily and quickly be tested for strength and

Compaction
Figure 13.16
The expanding
steel mold of the
‘‘K test’’ enables
rapid estimates of
soil strength and
compressibility
parameters from
Proctor density
specimens, but
drainage may be
incomplete.
compressibility in an expandable steel mold (Fig. 13.16). A specimen from a

Proctor density test is inserted in the mold and compressed at the ends, which
causes the mold to expand and automatically exert a gradually increasing lateral
confining stress on the sample.
In geotechnical engineering, K denotes the ratio between horizontal and vertical

stress. In a ‘‘K test,’’ vertical stress is measured and the lateral stress is obtained
from a suitable calibration from expansion of the steel mold. As both stresses
simultaneously increase it is possible to obtain running estimates of both the soil
cohesion and the angle of internal friction, as well as a compression modulus and
soil-to-steel sliding friction. Although the soil is confined between porous stones
to allow drainage, compacted specimens contain sufficient air that little drainage
occurs. K-test results have been used with bearing capacity theory for design of
tamping-foot rollers.
13.10 GRADATION AND COMPACTED DENSITY
Density after compaction is closely related to the range of particle sizes, as

progressively smaller particles fill in the voids, whereas density also is influenced
by the largest particles present, as a solid particle is more dense than a cluster of
many smaller particles.
A variety of moisture-density curves representing different gradations of soil is

shown in Fig. 13.17, where it will be seen that, in general, the coarser the soil, the
higher the compacted density and lower the OMC. However, it also may be noted
Compaction
Compaction 301
Figure 13.17
Some
representative
tests showing the
influence of
gradation on
compacted density
(Johnson and
Sahlberg, 1962).
that the lowest density is obtained with a sand that is nearly all one size, which is
the reason such a soil is said to be poorly graded and classified as SP or A-3.
13.11 FIELD DENSITY TESTS
13.11.1 Density as a Means for Compaction Control

Measurements of soil density and moisture content are the primary means for
quality control of compaction. In-place density measurements also are used for
determining cut-and-fill quantities. For example, if the density of soil in a borrow
pit will be increased 15 percent as a result of compaction operations, this
15 percent reduction in volume (which is loosely termed ‘‘shrinkage’’) must be
taken into account when calculating borrow quantities.
13.11.2 Direct Measurement Methods

The classic methods for measuring compacted density involve hand-excavating a
hole and measuring the volume of the hole, and the weight and moisture content
of the soil taken from the hole. The volume of the hole is determined by expanding
a rubber balloon into the hole with water and measuring the volume of water
required (a volumeter test), or by filling the hole with a standard sand and
measuring the amount of sand required (a sand-cone test). Construction traffic
must be halted during sand-cone tests because vibrations affect packing of the
Compaction
sand. Details of these tests will be found in laboratory testing manuals or standard
methods of the ASTM.
Another method involves driving a thin-walled steel tube into the soil and
measuring the weight and volume of soil extracted. This is considered less accurate
than the methods described above because of sample disturbance and compaction
during sampling.
A thin-walled steel tube called a Shelby tube can be pushed instead of driven, by
use of a drilling machine. The sample then is extruded in the laboratory, and the
weight, volume, and moisture content of representative samples measured.
Strengths also can be determined. This is the same procedure that is used in
exploration drilling for foundation soils, and has the advantage that it also is used
for sampling and testing soil underneath a compacted layer. Routine samples for
purposes of soil identification are obtained by driving a thick-walled Standard
Penetration Test (SPT) device, but the soil is so disturbed and compacted that it
is not acceptable for density determinations. Details of these tests are found in
Chapter 26.
13.11.3 Nuclear Density Measurements

The most common method for control of field compaction operations involves
measuring both the soil density and moisture using a nuclear density-moisture
meter. Two modes of operation are available, backscatter, which is as the name
implies, and transmission. For transmission measurements a probe containing a
radioactive source is pushed down into a hole in the soil and radiation
is transmitted upward at an angle to detectors in the base of the device. In the
backscatter mode, both the source and counter are at the ground surface, which
gives lower penetration and is less accurate. The backscatter mode is particularly
useful for measuring densities of pavements.
For the transmission mode a probe is driven into the soil to the required depth
and withdrawn, or a hole is made with an auger. The nuclear unit is placed over
the hole and a rod that is slightly smaller than the probe hole is lowered from
within its shield down into the hole, as shown in Fig. 13.18. The rod contains a
radioactive source that emits gamma rays. Best accuracy is obtained if the source
is against the side of the hole closest to the detectors. Gamma radiation
transmitted through the soil up to the bottom of the device is counted for
15 seconds to 4 minutes, depending on the accuracy desired and the transmission
characteristics of the soil.
Although nuclear devices measure density in kg/m3, they also are calibrated to
read unit weights in lb/ft3 or kN/m3. Test units require daily calibration, which
can be done by testing in concrete blocks of known density or other materials such
as limestone, granite, or aluminum (ASTM Designation D-2922). Operation
requires trained and certified personnel.
Compaction
Compaction 303
Figure 13.18
Nuclear density
gauge. For
transmission
measurements the
probe is lowered
the desired
distance into a
hole previously
prepared in the
soil. (Photo
courtesy Dr. David
White.)
Figure 13.19
Principle of nuclear
density-moisture
content testing
device for
monitoring field
compaction
operations.
13.11.4 Nuclear Determinations of Moisture Content

Moisture contents are determined using a fast neutron source that is incorporated
in the same instrument, shown diagrammatically in Fig. 13.19. As neutrons
Compaction
encounter hydrogen ions in the soil they are slowed, and those returning to the
unit are counted. The count includes all hydrogen ions including those that are
structurally held as OH ions in mica and the clay minerals, plus hygroscopic
water that is not driven off by normal oven-drying, plus hydrocarbons in organic
matter. Calibration therefore is required by counting similar soils with moisture
contents that are determined by oven-drying (ASTM Designation D-3017).
13.11.5 Accuracy of Nuclear Determinations

With modern instruments and accurate calibrations, density determinations are
accurate to within about 1 percent, and moisture contents to within about 2.5
to 10 kg/m3 (0.2 to 0.6 lb/ft3), depending on the counting time. The upper part of
the tested layer is emphasized, as about 50 percent of the counts come from the
upper 25 percent of the soil layer tested by the transmission method.
13.11.6 Operating Nuclear Devices

The U.S. Nuclear Regulatory Commission requires that nuclear density testers
have radiation warnings and be kept under lock and key when not in use.
Operators must be specially trained and certified, and wear film badges to ensure
that dosage does not exceed prescribed levels. Training normally is available from
instrument manufacturers. Units are periodically checked for radioactivity
leakage, and require periodic replacement of nuclear source materials. With
proper safety precautions the radiation danger is negligible, particularly when
compared to hazards from construction traffic.
13.12 FIELD COMPACTION EQUIPMENT
A broad catalog of compaction equipment is available and is continually being

improved as manufacturers seek a competitive edge. Most common are heavy
rollers, vibratory rollers, and tamping rollers. In close quarters such as trench
backfills, hydraulic hammers can be mounted on backhoes. Hand-operated
vibratory or ‘‘jumping jack’’ equipment may be used for moderate needs close to
buildings. Applicability of different equipment for different soil types is indicated
in Table 13.2.
13.13 CHEMICAL COMPACTION AIDS
Many different chemical compounds have been marketed as ‘‘compaction aids.’’

These generally are detergents that are added in small mounts to water to decrease
capillary tension, which reduces intergranular friction so that a specification
density can be achieved with less effort. However, reducing capillary attraction

Compaction
Compaction 305
Table 13.2
Applications of
compaction
equipment.
Adapted from Caterpillar Performance Handbook, ß Caterpillar Inc
also reduces the strength of the compacted soil and can increase dispersion of
clays. In other words, a small amount of chemical can make a soil easier to
compact by making it weaker. In recognition of this, chemical ‘‘compaction aids’’
are not permitted in most compaction specifications.
Whereas hydrated lime flocculates an active clay soil, increasing the OMC and
lowering the compacted density, chemical compaction aids do just the opposite.
Field inspectors should be wary of water additions that have a soapy feel or tend
to foam. Many aids are rendered ineffective by expansive clay minerals that
remove organic chemicals by adsorption between clay layers.
CaCl2, calcium chloride, also is the reverse of a compaction aid as it increases the
surface tension of water, and densification occurs as a result of higher suction
forces when the soil dries out. Calcium chloride occasionally is used to increase
the surface density of granular unpaved roads and as a dust palliative, but has a
low pH that accelerates rusting.
Another chemical additive, lignin, which is a byproduct from paper manufactur-

ing, is used as a dust palliative and behaves as a glue until rain leaches it out.
13.14 STATISTICS OF COMPACTION
13.14.1 Variability and Accuracy of Testing

The more variable a material, the more tests are required to characterize the final
product, and soil can be quite variable. The most common statistical parameter is
Compaction
the arithmetic mean, or average, but this says nothing about variability. If the
average density meets a design specification, it means that one-half of the tests do
not meet the specification, so the average must be higher than the specification
requirement.
13.14.2 Scraper Mixing

Soil variability can be reduced by a simple procedure illustrated in Fig. 13.20,
where scrapers load while moving downslope across soil layers instead of picking
them up one by one. Additional mixing is obtained by disking and grading, or a
specialized soil mixer can be used.
13.14.3 Maxima and Minima

Maximum and minimum values indicate the range or spread of the data. If
minimum values fall below the design requirement, the engineer must make a
choice: should the area where the deficiency occurred be reworked and recom-
pacted, or can the deficiency be ignored? Usually the first option is selected because
no construction person cares to approve a test that does not meet specifications.
The engineer also has other choices: if one determination does not meet speci-
fications it may only reflect a natural variability of data, but if half of the tests do
not pass, the entire project is suspect and should be tested and, if necessary, torn
out and recompacted under more carefully controlled conditions.
If moisture contents are too high, as sometimes happens when compaction is

attempted during rain, the soil must be scarified and aerated before being
Figure 13.20
Mixing reduces
soil variability and
energy is saved
by cutting across
soil layers.
(Duwayne
McAninch,
personal
communication.)

Compaction
Compaction 307
recompacted, which can introduce construction delays and cost overruns.

However, the integrity and reputation of the engineer must remain above question.
13.14.4 The Control Chart

A useful tool used in manufacturing is a control chart (Fig. 13.21). The control
chart is simply an as-measured day-to-day plot of critical data such as soil
compacted density and moisture content. Adverse weather or other conditions
should be noted on the chart and can help to explain anomalies and decide on
corrective measures.
For example, the control chart may show a group of tests that gave marginal or
unacceptable density data. Is the reason bad data, or is it perhaps coincident with
an early-morning rain, a change in the soil, a change in operator or operating
procedure, a change in equipment, or is it none of those causes and attributable to
carelessness or inattention to details?
One of the first jobs for an engineer-in-training often is that of inspector. It is a

very important job, one that requires and develops self-confidence. The inspector
must not be intimidated or fooled into looking the other way, but should be
discreet about offering suggestions for improvement because of liability issues.
The duty of the inspector is to determine whether specifications are satisfied and
to report and if necessary close down a job if they are not. The inspector should be
wary of the contractor who knows in advance when tests will be conducted and
then suggests to ‘‘test over here.’’
13.14.5 Measures of Variability

If all tests are required to meet specifications, from the point of view of the
contractor one test that passes is better than 10 tests if one does not pass, and
the more tests there are the more likely it is that one or more will not pass. The
current practice of requiring additional compaction in areas that do not meet
Figure 13.21
Control chart for
compacted
density. A similar
chart can be
prepared for
moisture content
measurements and
can help pinpoint
problem areas.

Compaction
specifications gives no assurance that other similar or worse areas may not have
been missed by the tests. Ideally, a large number of tests will be randomly
performed and they all will meet specifications, but it is not a perfect world.
The most common statistical measure for variability is the standard deviation,
which unfortunately is given the same symbol, , that engineers use for stress. The
standard deviation is routinely calculated by computer programs that calculate
averages.
An average usually is designated by x, and the standard deviation by a number

following a symbol; therefore 5.5 1.2 means an average value of 5.5 and a
standard deviation of 1.2, unless a different definition is stated.
The formula for the square of the standard deviation, called the variance, is
X
2 ¼ ðxi x Þ2 =n ð13:4Þ
where xi represents individual data readings, x is the average, and n is the number
of data points. It will be noted that ‘‘variance’’ has a specific meaning and should
not be confused with ‘‘variability.’’
13.14.6 The Meaning of a Standard Deviation

Mathematically a standard deviation is the second moment of the data about the
mean. The mean is first calculated, and the departure of each data point from the
mean is determined. Each departure then is squared so that negative departures
become positive and additive. The sum of the squared departures is divided by the
number of data points to obtain an average called a variance. In order to return to
a linear scale, the square root of the variance is the standard deviation.
‘‘Standard error’’ is the standard deviation divided by the square root of the
number of observations. The term carries an unfortunate connotation and is not
popular with engineers.
Manufacturing tolerances may have limits of acceptability as x 3, where

is the standard deviation. This is satisfactory when the standard deviation is small,
but in soils it typically is much too large for this criterion to be valid or acceptable.
In the case of compacted soil, which is a manufactured product, the limits are set
by density and moisture content criteria obtained from laboratory density tests.
13.14.7 Coefficient of Variation

An important measure in engineering is the coefficient of variation, Cv, which is
defined as the standard deviation divided by the mean and expressed as a percent:
Cv ¼ 100=x ð13:5Þ

Compaction
Compaction 309
The coefficient of variation allows comparing the variability of apples and that of
oranges. Cv usually is in the range 10 to 25 percent for most soil properties. The
lower the coefficient, the more confidence one may have that a soil is uniform or a
job is uniformly constructed
13.14.8 The Normal Distribution Curve

Most statistical theory is based on the concept of random error, which
theoretically results in a ‘‘normal’’ or ‘‘Gaussian’’ distribution of data. This is
shown by the symmetrical, ‘‘normal’’ curve in Fig. 13.22. However, this
distribution is not strictly applicable to most engineering problems because the
ends trail off to infinity in both directions, which may be physically impossible.
For example, soil moisture contents cannot exceed limits imposed by saturation,
densities cannot exceed those dictated by a zero air void condition, and neither
can be negative. Furthermore it is the lower ‘‘tail’’ of a distribution, which is least
well defined, that defines a probability of failure.
13.14.9 Is There a Better Theoretical Distribution?

The normal distribution is based on probability theory that says that if the
number of measurements is very large, than the distribution is symmetrical
around an average value and follows a bell-shaped or Gaussian normal
distribution curve. This led to a concept of ‘‘probable error’’ that is based on
the area under a normal curve. The term has fallen into disuse with limited
measurements because they probably do not properly define the distribution. The
‘‘probable error’’ equals 0.6745, and means that there is a 50 percent probability
that a measurement will be outside of bounds determined by the mean and the
standard deviation. For example, if a mean is 100 and the standard deviation is
10, 50 percent of the measurements should fall in a range 100 6.745, or between
93 and 107.
The concept of probable error is used when there is a large number of tests that do
follow a normal distribution, and is a basis for control chart limits of 3 that
Figure 13.22
A beta distribution
may be preferred
over a normal
distribution
because it can be
adjusted to stay
within physical
limits that are
absolute.

Compaction
theoretically include 99.73 percent of the data. Thus, in the manufacture of

1000 light bulbs the 3 control chart limits should pass 997 and reject 3.
In order to obtain the most efficient manufacture that is closest to the tolerance
limits, the product should be consistent so that the standard deviation is low and
3 is low.
13.14.10 Beta Distribution

The use of a beta distribution has been pioneered in geotechnical engineering by
Harr at Purdue University. A beta distribution has defined upper and lower limits
that customarily are adjusted to be 0 and 1. An example of a beta distribution is
shown in Fig. 13.22, and has been drawn to stop before it reaches the zero air voids
condition. The beta shape is variable, and in this case a lopsided shape is
reasonable as measurements tend to be concentrated close to the zero air voids line.
A beta distribution may be narrower or broader than that which is shown, may be
symmetrical or skewed one way or the other, may be higher at the ends than at
the middle, or may even be a horizontal line indicating an even distribution.
With that kind of versatility it may be difficult to assign the most appropriate
descriptive parameters. As a first estimate a beta distribution may be selected that
approximates a symmetrical normal distribution, but with discreet upper and
lower boundary values based on physical limits such as not allowing a negative
factor of safety.
13.14.11 Reliability and Factor of Safety

Engineers usually require that a design meet a certain minimum factor of safety,
which is a ratio between the design value and a failure value. For example, if the
calculated load on a beam to cause failure is 100, a factor of safety of 3 will mean
that the design value must be 300. This not only allows for variability of the beam
strength due to inconsistencies in manufacture, but also can to a limited extent
compensate for future overloading.
In the case of compaction, it is inevitable that some tests will not meet
specifications, and a beta distribution can help to translate a failure ratio into
reliability. A formula for reliability is (Harr, 1987):
aþ1
R¼ ð13:6Þ
aþbþ2
where a is the number of successes and b is the number of failures.
A beta distribution for eq. (13.6) is shown in Fig. 13.23, which in this case indicates
a reliability of 0.71—that is, based on this limited number of tests, 71 percent of
the product can be expected to meet specifications, instead of 80 percent indicated
by the number of tests that actually passed.

Compaction
Compaction 311
Figure 13.23
Beta distribution
calculated for 4
successes and 1
failure. (From Harr,
1987.)
Figure 13.24
Reliability
prediction based
on number of
successes with or
without one failure.
13.14.12 Number of Consecutive Tests with No Failures

Equation (13.6) can be written for no failures or b ¼ 0:
aþ1
R¼ ð13:7Þ
aþ2
where a is the number of tests. Some solutions for this and the preceding equation
are shown in Fig. 13.24. For example, if 20 tests are performed and one test does
not pass, R ¼ 0.91, but if all tests pass, R ¼ 0.95, which is more acceptable. This
does not necessarily mean that 5 percent of the job will fail because failure also
depends on other factors. According to this analysis, 10 tests without a failure

Compaction
indicate a reliability of 11/12 ¼ 92 percent. The higher the number of tests and
lower the number of failures, the higher the reliability.
A common procedure is to repeat a test that does not pass, in which case it will
have approximately the same probability of success as the first batch of tests.
The repeated test therefore might be regarded as somewhat superfluous—unless
it fails.
13.15 PARTICLE STACKING
13.15.1 Vertical Orientation of Voids

As discussed in Chapter 8, sedimentation leads to stacking of soil particles, as
the contact angle must be high enough to cause slipping. A similar observation
was made by Oda (1978) for sand held in a container that was tapped on one
side (Fig. 13.25). The phenomenon later was modeled by steel rods by Oda et al.
(1985). Such an arrangement is metastable and should be prevented by changing
directions of the major principal stress, as occurs under a roller or tamping foot.
13.15.2 Dilatancy or Compression

Vaid and Thomas (1995) found that when specimens of loose sand were
compressed vertically they tended to dilate, which would occur if there is bending
and opening of a columnar structure. Compressed horizontally, the sand
densified, which also is consistent with the concept of a columnar arrangement.
Pham (2005) vertically compressed a loessial silt, some specimens with and others
without lateral strain. Vertical compression alone resulted in a much lower degree
Figure 13.25
Preferred vertical
orientation of
sand grains after
self-weight
densification. The
dashed circles
show random
orientation.
(Modified from
Oda, 1978.)

Compaction
Compaction 313
Figure 13.26
Comparison of
void ratios
obtained by
vertical
compression and
vertical plus lateral
compression of a
silt initially
compacted to
95 percent
maximum density.
The change in void
ratio at 300 kPa
mean stress is
from 0.64 to 0.42,
a decrease of
34 percent. (From
Pham, 2005.)
of compression, shown by comparing the curves in Fig. 13.26. Specimens that

were compacted horizontally as well as vertically also had a much higher modulus
as a result of the additional densification.
While these observations may appear to be academic, they affect laboratory

compaction of soil specimens in preparation for testing. Laboratory compaction
should model the field process, which was the basis for the Proctor density test
and other methods such as a kneading compactor. Static loading of soil in a steel
mold does not duplicate these conditions.
13.16 SUMMARY
Compaction is the most frequently used method for improving soil properties
such as strength, low hydraulic conductivity, and stability from volume changes,
but requires careful tests and controls to achieve these purposes. Compaction of
non-expansive soil to 95 percent density at or near the optimum moisture content
should yield a product that is stable from future volume changes, whereas a soil
that is compacted on the dry side of optimum may collapse when wet even though
it meets the density requirement. Compaction on the wet side of optimum is useful
for some purposes but increases the danger of overcompaction that will damage
the soil.

Compaction
Expansive clays can be approached by replacing or treating the clay, or to a

limited extent by adjusting the compaction moisture content and preventing
future changes in moisture content.
Compaction normally improves strength and reduces compressibility, but these

properties must be measured or their adequacy is based on experience with
particular soils. Statistical variability indicates that occasional tests will not meet
specifications, in which case a reliability can be calculated based on the number of
tests that pass and that do not pass. Compaction is most effective if it produces
horizontal as well as vertical compressive stresses.
Problems
13.1. Suggest compaction specifications for the following: (a) base course for a
highway, (b) base course for an airport runway for large aircraft,
(c) residential landscaping, (d) trench backfill that will support pavement,
(e) clay core of an earth dam.
13.2. Calculate the compactive energy per unit volume of soil in the standard
and modified Proctor tests.
13.3. Explain the differences and applications between compaction and
consolidation.
13.4. (a) Why are soils compacted in layers? (b) Why is it important to have
a firm, rough base for compaction?
13.5. Soil in a haul road develops a soft, bouncy feel and a rolling swell ahead of
scraper wheels so that in effect the equipment is always going uphill. What
is the reason and what is your recommendation?
13.6. A bucket filled with dry sand weighs more than a bucket filled with wet
sand. Explain.
13.7. A nuclear gauge is used to test soil at the standard Proctor OMC and
maximum density in Fig. 13.2. What percent accuracy can be expected for
the moisture content?
13.8. Identify the soil in Fig. 13.2 by comparing the compacted density curve
with those of Fig. 13.17. Does this identification appear to be correct?
13.9. Repeat Problem 13.8 with the soil in Fig. 13.15.
13.10. Give two reasons why a field density might give data above the zero air
voids line. What should be done to mitigate this problem?
13.11. In Fig. 13.14, 90% modified Proctor density is higher than 95% standard
density, and there is a shortage of water at the construction site. Which
specification do you recommend? Is there any risk involved in specifying
90% maximum density?
13.12. (a) Relative density tests on a dry sandy gravel give a maximum and
minimum of 20.1 and 14.5 kN/m3 (128 and 92 lb/ft3), respectively. (b) If
G ¼ 2.65, find emax and emin. (c) Dry densities of this soil are found to be
Compaction
Compaction 315
19.2, 18.2, and 17.0 kN/m3 (122, 116, and 108 lb/ft3). Calculate the relative
densities.
13.13. What is the percent saturation at the OMC and maximum density in
Fig. 13.2? Assume G ¼ 2.70.
13.14. List some of the hypotheses that have been proposed to explain the failure
of Teton Dam in Idaho and comment on each. (Additional reading
required.)
13.15. A proposed earth dam will contain 4,098,000 m3 (5,360,000 yd3) of earth.
It will be compacted to a void ratio of 0.80. There are three available
borrow pits, which are designated as A, B, and C. The void ratio of the soil
in each pit and the estimated cost of moving the soil to the dam is as
shown in the tabulation. What will be the least earth-moving cost, and
which pit will it be most economical to use?
Cost of moving
Pit Void ratio per m3 per yd3
A 0.9 $1.67 $1.28

B 2.0 1.19 0.91
C 1.6 1.56 1.19
13.16. (a) Calculate the mean, standard deviation, and coefficient of variation of
the data shown in Fig. 13.21. (b) Two of the tests do not meet
specifications. According to the subcontractor all of the soil received the
same number of roller passes so he questions the validity of the tests. As
the inspector, what are your options?
13.17. The data points shown in Fig. 13.27 are obtained in a Proctor density test
of a silty sand. Which is more likely to be the correct interpretation of the
data, linear regression as in (a), conventional curve fitting (b), or curve
fitting as in (c)? Explain.
13.18. The following data were obtained during compaction of clay fill to
support a shopping mall. Use a computer spreadsheet to prepare a control
chart. The density is specified to at or above 95% of the maximum
Figure 13.27
Data
interpretations in
Problem 13.17.

Compaction
laboratory density, which is 98.6 lb/ft3 (15.5 kN/m3) and the moisture
content is specified to be between the optimum of 18.8% and 4% above
optimum. (a) With reference to Fig. 13.17, is this a reasonable maximum
density? (b) Prepare a control chart for density and for moisture content
showing the specification limits. Are any of the data outside of these
limits? (c) Calculate the reliability. If 500 tests were conducted, what
number might be expected to pass?
Test No. Dry density Moisture

lb/ft3 kN/m3 content,%
1 95.4 15.0 19.8

2 99.3 15.6 18.0
3 99.0 15.55 17.8
4 97.6 15.3 18.0
5 98.9 15.5 18.0
6 99.3 15.6 17.0
7 97.8 15.4 18.8
8 95.8 15.1 19.0
9 96.7 15.2 17.4
10 96.6 15.2 18.9
11 95.7 15.0 18.5
12 98.9 15.5 20.0
13.19. Calculate the mean, standard deviation, and coefficient of variation of the
density data in Problem 13.18. Does the uniformity of the data indicate
good, average, or poor quality control on compaction? Would a control
chart based on x 3 define acceptable limits? What are those limits?
Partial ans.: x ¼ 97.58x 1.497.
13.20. A statistical ‘‘t’’ test is based on a symmetrical distribution and is used to
estimate confidence limits on the mean. With 12 observations, one can be
95% confident that the actual mean is x 0.664. Calculate the limits on
the mean.
13.21. According to the area under a normal curve, 5% of the densities will
exceed x þ 1.645 and 1% will exceed x 2.875. Is this reasonable? Why
(not)?
References and Further Reading

Harr, M. R. (1987). Reliability-Based Design in Civil Engineering. McGraw-Hill,
New York.
Johnson, A. W., and Sahlberg, J. R. (1962). ‘‘Factors Influencing Compaction Test
Results.’’ Highway Research Board Bull. 319.

Compaction
Compaction 317
Lawton, E. C., Fragaszy, R. J., and Hetherington, M. D. (1992). ‘‘Review of Wetting-

Induced Collapse in Compacted Soils.’’ ASCE J. Geotech. Eng. 118(9), 1376–1394.
Leonards, G. A., ed. (1987). Dam Failures. Elsevier, Amsterdam.
Oda, M. (1978). ‘‘Significance of Fabric in Granular Mechanics.’’ Proc. U.S.-Japan
Seminar on Continuum Mechanics and Statistical Approaches in the Mechanics of
Granular Materials 7–26.
Oda, M., Nemat-Nasser, S., and Konishi, J. (1985). ‘‘Stress-Induced Anisotropy in
Granular Masses.’’ Soils and Foundations 25(3), 85–97.
Olson, R. E. (1963). ‘‘Effective Stress Theory of Soil Compaction.’’ ASCE J. Geotech. Eng.
89(SM2), 27–45.
Proctor, R. R. (1993). ‘‘Fundamental Principles of Soil Compaction.’’ Engineering News
Record (Aug. 31 and Sept. 7, 21, and 28).
Seed, H. B., and Chan, C. K. (1959). ‘‘Structure and Strength Characteristics of Compacted
Clays.’’ ASCE J. 8(SM5, Pt. 1), 87–128.
Vaid, Y. P., and Thomas, J. (1995). ‘‘Liquefaction and Post-liquefaction Behavior of
Sand.’’ ASCE J. Geotech. Geoenviron. Eng. 89(SM2), 27–45.
Jürgenson, L. (1934). ‘‘The shearing resistance of soils.’’ Jour. Boston Soc. Civil Engrs. July:
184–217. Reprinted in Contributions to Soil Mechanics 1925–1940. Boston Soc. of Civil
Engrs., Boston, Mass.
Pham, Ha (2005). ‘‘Support Mechanisms of rammed aggregate piers.’’ Unpublished Ph.D.
thesis, Iowa State University Library, Ames, Iowa.


13 Compaction

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

13 Compaction

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

13 Compaction

Uploaded by

Copyright:

Available Formats

Source: GEOTECHNICAL ENGINEERING

13.1 HISTORICAL BACKGROUND

13.1.1 Paths of Least Resistance

13.1.2 The Twentieth Century

280 Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

13.2 PROCTOR’S DISCOVERY

13.2.1 A Direct Approach

13.2.2 AASHTO T99 Procedure

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

282 Geotechnical Engineering

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

13.2.4 Not Lubrication

The anti-lubricating effect of water also is shown by compaction curves such as

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

284 Geotechnical Engineering

13.3 ZERO AIR VOIDS LINE

13.4 COMPACTING IN LAYERS

13.4.1 A Critical Requirement

13.4.2 A Virtual Compactor

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Specific gravity of solids Table 13.1

Moisture content, % 2.5 2.6 2.65 2.7 3 Densities and

Specific gravity of solids

Moisture content, % 2.5 2.6 2.65 2.7 3

0 24.51 25.49 25.98 26.47 29.41

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

286 Geotechnical Engineering

13.4.3 The Importance of Roughness

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Frictional restraint also can be added on top of the compacting layer by

13.4.4 Tamping-Foot Rollers

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

288 Geotechnical Engineering

13.4.5 Proof Rolling

13.4.6 ‘‘Oreo Cookie Effect’’

13.5 OUT GOES THE AIR, NOT THE WATER

13.5.1 A Critical Distinction

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

regardless of the number of additional passes of a compactor. If processing

A Conflict over Nomenclature

The effect of overcompaction on strength of a clay is shown in Fig. 13.8, which

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

290 Geotechnical Engineering

seen moving as a wave ahead of tires on heavy equipment, increasing energy

The remedy for overcompaction obviously is not more compaction. Instead

13.6 COMPACTION OF GRANULAR SOILS

13.6.1 Free-Draining Soils

13.6.2 Bulking and Flooding

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

13.6.3 Relative Density Test

The formula for relative density is

A more convenient form of this equation that uses unit weights is

13.6.4 Defining Minimum and Maximum Density

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

292 Geotechnical Engineering

A maximum density, max, is more difficult to define because it depends on

13.6.5 Applications of Relative Density