TRANSPORTAT/OfV RESEARCH RECORD 1473

Effect of Vertical Confinement on Dynamic

Cone Penetrometer Strength Values in
Pavement and Subgrade Evaluations
MOSHE LIVNEH, ILAN ISHAI, AND NOAM A. LIVNEH

The Dynamic Cone Penetrometer (DCP) has become one of the most drilling of a small-diameter hole in the asphalt structure. These
useful testing devices in pavement evaluation in Israel and other parts issues are considered to be related to vertical confinement.
of .the world. Consequently, a reliable true meaning and in~erpretation
Since the DCP has become one of the most useful testing devices
of the results are needed. Research work dedicated to studying the ef-
fects of vertical confinement on the DCP strength values of the granu- in pavement evaluation in Israel, a reliable true meaning of the re-
lar pavement layers and subgrade is summarized. Specifically, four sults and an interpretation of the results were required. Conse-
major effects were studied: vertical confinement of granular layers, ver- quently, research work was dedicated to studying the effects of ver-
tical confinement of cohesive layers, vertical confinement of rigid struc- tical confinement on the DCP values. Specifically, the following
. tural layers, and the effect of vertical confinement of upper asphalt lay- effects were studied:
ers on the DCP strength values of the granular layers below them. Based
on engineering analysis and experimental testing in the laboratory and ·
in the field, the following conclusions were made. (a) No vertical con- I. Effect of vertical confinement of granular structural layers on
finement effect by rigid pavement structure, the upper granular layers, clay and silt subgrade DCP values;
or the upper cohesive layers on the DCP strength values of lower cohe- 2. Effect of vertical confinement of cohesive layers on clay sub-
sive subgrade layers was found. Any differences in the results between grade DCP values;
the confined and unconfined DCP values can be explained by the fric- 3. Effect of vertical confinement of rigid structural layers (an all-
tion that developed in the· rod during tilted penetration. (b) However, asphaltic structure) on clay subgrade DCP values; and
vertical confinement effects by upper asphalt layers on the DCP values
of the granular pavement layers exist. Since this is the true effect of the 4. Effect of the vertical co.nfinement of upper asphaltic layers on
pavement structure, any DCP measurement for pavement evaluation the DCP values of the granular layers below them.
purposes should be performed through a narrow boring in the asphalt
layers and not after removal of a wide strip of asphalt. Generally, these The study of these effects is important for use of the DCP in the
confinement effects decrease the DCP values, and. thus increase the reliable evaluation of existing pavements. This paper presents the
structural strength measured. These confinement and friction effects, theoretical background for vertical confinement, as well as an analy-
which can be evaluated quantitatively, should be taken into considera-
sis of the results obtained from both laboratory and field tests
tion when using the DCP method to evaluate existing pavements.
designed to study the effects mentioned earlier.

During the last decade, the Dynamic Cone Penetrometer (DCP) test
has been increasingly used in many parts of the world for pavement DCP TEST IN PAVEMENT EVALUATION
and subgrade evaluation through its relationship with the in situ Cal-
ifornia Bearing Ratio (CBR). This is because of its economy, sim- The DCP, as developed in South Africa (J 1), consists of a steel rod
plicity, and capability of providing rapid measurements of the in with a cone at one end. It is driven into the pavement or the sub-
situ strengths of subgrades and pavement layers without excavation grade with a sliding hammer, and the material resistance to pene-
of the existing pavement, as in the in situ CBR test (1-9). tration is measured in terms of millimeters per blow. The cone is an-
Extensive work with DCP testing and the experience gained in gled at 30 degrees, with the larger diameter of the cone being 20
Israel have shown that, in addition to the CBR-DCP correlation, mm. The hammer weighs 8 kg, an~ the dropping sliding height is
575 mm (Figure 1).
some other factors that have an influence, such as vertical confine-
ments, should be taken into consideration (J 0). For example, a ques- The DCP was originally designed and used to determine the
tion occasionally arises, that is, whether DCP results, obtained from strength profile of the flexible pavement structure and subgrade
the subgrade by means of the rod's penetration through the struc- (12-14). Usually, pavement testing at a given point involves the ex-
tural layers, are identical to the results obtained from the same sub- trusion of a 4-in. circular core from the top asphalt layers only and
grade after removal of the pavement structural layers. In other penetrating the DCP from the top of the base course layer down to
words, are the subgrade DCP results affected by the presence of the the required pavement or subgrade layer. The properties of the as-
flexible structure? Similarly, the question can be applied to an phalt layers are directly evaluated in the laboratory by a proper me-
all-asphalt pavement whose s~bgrade was examined following chanical test (resilient modulus, diametrical test, splitting test, Mar-
shall test, or others). The pavement parameters are continuously
measured and recorded with depth by the DCP. Immediately at the
test's conclusion, the shallow 4-in. hole is easily filled with either
portland cement concrete (regular or fast curing) or a proper cold
Transportation Research Institute, Technion, Haifa 32000, Israel. asphaltic mixture. In case of only subgrade evaluation for pavement
2 TRANSPORTATION RESEARCH RECORD 1473

NUMBER OF BLOWS

/HAMMER
(NIL HAMMER MASS = 8 kg.)

TWO PARTS SCREW TOGETHER

/

MEASURING TAPE

jf l6mm

TEMPERED CONE
FIGURE 2 DCP test output.

JI 2omm

FIGURE 1 General details of South African DCP. basis of 135 tests the total equation, which was based on old and
new data, is formulated as follows (Figure 3):

design purposes, the DCP is penetrated dowri from the top of the log CBR = 2.14 - 0.69 (log DCP)1.5 (R2 = 0.876, N = 135) (2)
natural soil or compacted subgrade.
During testing the number of blows versus depth is recorded. The From a practical standpoint, both Equations 1 and 2 yield almost
DCP value is defined as the slope of the blows-versus-depth curve identical results (10).
(in millimeters per blow) at a given linear depth segment (Figure 2). Several other agencies and researchers around the world have
also tried to develop correlations between DCP and CBR values
(3,8,11,16). Webster et al. (8) compared some of these correlations
CORRELATION BETWEEN DCP AND CBR (Figure 4). It is evident that general agreement was reached between

To be able to relate DCP values to the structural parameter of the

pavement under the local pavement design and evaluation technol-
ogy, an extensive controlled laboratory and field test was carried out
to correlate DCP to CBR (10,15). In the laboratory conventional 100
and DCP tests were performed on a wide range of undisturbed and
compacted fine-grained soil samples with and without saturation:
Compacted granular soils were tested in flexible molds with vari-
able controlled lateral pressures. Field tests were made on natural
and compacted layers representing a wide range of potential pave-
ment and subgrade materials. Pavement evaluation tests were also
performed for pavement and material evaluation and for correlation
r:i
ci:::i
10
with pavement condition.
u
The correlative laboratory and field testing program resulted in a
quantitative relationship between the CBR of the material and its
DCP value as follows (Figure 3):

log CBR = 2.20 - 0.71 (log DCP)1.5 (R2 > 0.95) (1) 1
1 10 100
where the DCP is the penetration ratio in millimeters per blow. DCP, mm/blow
This relationship, which was initially formulated in 1985, was
based on 56 comparative test results. Later, this correlation was FIGURE 3 Relationship between CBR values and
checked as data accumulated over several years (10). Finally, on the DCP values (10,15).
 Livneh eta!. 3

100 c:::=====r:==~:::=~··:::c:::c=i=r:::r:::i:::===1===i====:::i==i:==i===i:=i=1=+========i
.. .. ··... ··.·.....

r:i 10 1--~~~-+~~-+-~~--+--1t--t-t-t--+-~~-t--~•~~~·~··~·+-----t
j::Q ..............
u

2 3 4 5 10 15 20 30 40 50 100 200
DCP, mm/blow

.
WES DATA LIVNEH (1987) HARRISON (1987)
---------- ····················
VAN VUUREN (1969)
-----
KLEYN (1975)
-··-··-··-··-··
FIGURE 4 Universal data for relationship between DCP and CBR (8).

the various sources of information. On the basis of these results the in which the ratio DIB is larger than 4 to 10. For the DCP device this
following equation was selected as the best correlation: ratio means depths of more than about 10 cm. ·
In this context it is important to show that the state of deep foun-
log CBR = 2.46 - 1.12 (log DCP) (3) dation is defined by Meyerhof (19) as follows:

It can be seen that a universal correlation exists between the DCP D = 4VNq,B (4)
· and CBR for a wide range of pavement and sub grade materials, test-
ing conditions, and technologies. where

(5)

ENGINEERING BACKGROUND OF VERTICAL and

CONFINEMENT EFFECT
D = the foundation depth,
In the context of the basic bearing capacity approach, a plastic fail- B = the width or diameter of the foundation, and
ure mechanism can be adopted to describe soil behavior under cone- <!> = the material's internal angle of friction.
shaped penetrometers. This approach was adopted by Livneh (17) Here, for cohesive soils (where<!> is equal to 0), Equations 4 and
and Livneh and Greenstein (J 8) to find the effects of lateral pres- 5 lead to the ratio of DIB equal to 4, whereas for cohesionless soils
sure on CBR values in granular materials. Similarly, few analytical (where<!> is> 0) a deep foundation is defined at a depth of D greater
or numerical solutions for wedge and cone-shaped penetrometers, than 4B.
which account for both cone apex-angle and roughness, are given
in the literature (19-21).
Durgunoglu and Mitchell (20) proposed three types of failure
mechanisms relevant to deep foundations that can also be adopted
for penetration tests (Figure 5). On the basis of actual test results
with variable apex-angles and cone roughness, they concluded that
the failure mechanisms described in Figure 5(b) and 5(c) are inap- I I
propriate, whereas the failure mechanism represented in Figure 5(a)
closely represents the actual failure surface asso~iated with wedge
penetration.
f
I
I
I
I
I
I
I
T
D
On the other hand, Meyerhof (19) provided solutions for both co-
hesive and cohesionless sqils for certain conditions assuming the
failure mechanism shown in Figure 5(b), whereas Nowatzki and
lqf
f
1
Karafiath (21) used a finite-difference technique for limited condi-
(a) (b) (c)
tions and obtained some penetration resistance values for cohesion-
less soils assuming the failure mechanism shown in Figure 5(c). It FIGURE 5 Different failure mechanisms proposed for
should be noted that this mechanism is only possible in the situation penetration test (20). -
4 TRANSPORTATION RESEARCH RECORD 1473

To investigate the vertical confinement effect one should look at where de is the increase factor of Ne depending on the foundation
the Terzaghi's basic bearing capacity formula. Under this concept depth.
the penetration resistance qu can be defined as follows: Theoretically, vertical confinement is mainly possible because of
the bearing capadty .component -yDNq (Equation 6). In cohesive
(6) soils, when <I> is equal to 0, Nq is equal to 1.0, and thus, this compo-
nent is quite negligible. Hence, it can be seen that for clay subgrades
where the values of both Ne and Nq in Equation 6 eliminate the effect of
vertical confinement in deep penetration measurements of the DCP.
qu
the unit tip resistance,
=
On the other hand, for granular materials (where <I> is> 0), both
the unit cohesion,
C =
Nq and Ne possess significant values that increase with depth. There-
-y = the material density,
fore, the vertical confinement of the upper layers will be significant
D ·= the depth of the penetrometer tip,
and will increase with the depth of the DCP tip.
Nn N,., Nq = Terzaghi's bearing capacity coefficients (Figure 6),
This situation of a very deep foundation is similar to a position in
and
which the foundation is not as deep but the soil surface is bound by
Kc, K,. = the shape factors.
a rigid structure (as, for example, in an all-asphalt structure). On the
The question then asked is whether DCP test results obtained at basis of the preceding discussion, it can be concluded that the DCP
a depth of 10 cm and more below the clay subgrade, when the test results obtained from a penetration depth of 10 cm or more from the
is conducted in the clay subgrade after drilling into the pavement, surface of the clay subgrade will be only negligibly affected by the
are different from the results obtained from a similar subgrade at a presence of a rigid structure on the subgrade's surface or will be
depth of 10 cm and more below the surface, when the test is con- totally unaffected.
ducted following the removal of a wide strip of pavement. In other These conclusions are derived from the engineering analysis, and
words are DCP test results obtained at a depth of 10 cm and more they are compared with the empirical testing described later in this
still subject to the possible effects of vertical confinement? paper.
The answer according to Meyerhof (19) is negative when cohe-
sive subgrade is concerned. If one looks at Figure 6 (which presents
the bearing capacity coefficients as a function of the angle of inter-
EFFECT OF CONFINEMENT BY.
nal friction), it can be seen that for <I> equal to 0, the value of Ne re-
GRANULAR STRUCTURE
mains constant, commencing from a depth equal or greater than that
expressed in Equation 4. Brinch-Hansen (22) also provide a nega-
The effect of confinement by the granular structure on the sub-
tive answer since, commencing from a certain value of D, the
grade' s DCP.values was studied on the basis of an analysis of ac-
increase in Ne as a consequence of the continued increase in D is
tual DCP test data (10). The analysis presented in this section is
negligible. This is reflected in the following equation:
based on 11 DCP test locations. The DCP test was conducted fol-
0.35 lowing the drilling of asphalt cores at these points. The subgrade
de = 1.0 + B 0.6 (7)
- + - - - -4 - DCP values obtained by this method are referred to as the confined
D (1 + 7tg <!>) values, that is, DCPcon· These values are compared with the DCP
values derived from the subgrade values described earlier follow-
ing removal of all structural layers (i.e., after the digging of test
1000 I pits). The DCP values obtained from the exposed subgrade are
,
>O - - Strip ( D<.B) J ,,
z referred to as the unconfined values, that is, DCPunc·
--Square (D<B) ,I I
z
CT . The test results obtained are presented in Figure 7. It can imme-
u - - - Pile CQ/8>4-10) I
I ~~ diately be seen from Figure 7 that the two .test populations are not
z / ,, identical, as also verified statistically. Thus, for example, a statisti-
/A ~
I /
(/)
/ cal t-test of the results showed that the mean unconfined DCP (x) is
~
..._ 100 /
/
/ , J. ff larger than the mean confined DCP ()i) at a confidence level of 95
u / /,/ //:-7 percent.
Lt: Ne v / ~,,. IW' Moreover, a linear y = a + bx type of regression analysis leads
> v ... /
~/ to the following results:
!:::
u
~ ,,,' ""
/
~~ v
.. ~ v, ~
<
Ne /
~v a= 21.03 (Sa= 6.42), b = 0.103 (Sb= 0.158), R2 = 0.45,
u ,,/ and CT2 = 46.21
<!)
10..... ~ .........-·
././ / '/
z ~ ~ N' ~
. . ' v /,
a:
<
V'"
~ ~ 1/ where
IJJ
CD
~~
v N-g
VJ
I~ Sa and Sb
CT
2
= standard deviations for a and b, respectively;
= mean square error (MSE); and
~

~'
R = correlation coefficient.
I
/ A statistical F-test of this regression leads to the conclusion that
0° 10° 20° 30° 40°
at a confidence level of 95 percent, the regression is not significant
ANGLE OF INTERNAL FRICTION, JO
or, in other words, that there is no correlation between x values and
FIGURE 6 Bearing capacity factors for spread and pile y values. The analysis according to Grubbs (23) leads to the con-
foundations (16). clusion that the systematic error related to measuring x values is sig-
Livneh eta!. 5

60 In other words when the DCP test is conducted through granular

layers, an error of 34 percent magnitude may occur. In this specific
a Clay
case the DCP values of the clay subgrade should be increased by
50 li!I Silt
about 34 percent to obtain the true reduced strength of the material
i:l';
0 beneath the granular structure.
::0 Finally, it should be mentioned that the effect of friction created
40
~ ca
in the rod (either by tilted penetration or by material collapse) can
be quantitatively evaluated by torque measurement in the DCP de-
Cl 30 vice during penetration intervals. This work is being done and_ will
~ be reported in the near future.
~
0
u 20
e; EFFECT OF CONFINEMENT BY CLAY LAYERS
Cl Constrained regression
10
through graph origin
It is also important to examine whether conducting the DCP test on
the surface of the clay subgrade and conducting the DCP test
0 through this sub grade (commencing from a certain depth) lead to
0 10 20 30 40 50 60
identical results. The results of the 27 tests are presented in Figure
DCP UNCONFINED , mm I blow 8. The method was to conduct a DCP test on the surface and then
dig a 0.5-m-deep test pit and. conduct the DCP test again, com-
FIGURE 7 DCP test results for subgrade beneath
granular structure. mencing from the pit bottom downward.
The DCP value in the first test, which corresponds to a depth of
0.5 to 0. 7 from the soil surface, is compared with a DCP value in
nificantly different from the systematic error related to measuring y the second test, which corresponds to a depth of 0.0 to 0.2 from the
values. At the same time, the precisions of the readings in both tests pit bottom. The first test provides the value of the confined DCP (the
(as expressed by the standard deviation of the random error) can be x values), and the sec;ond test provides the value of the unconfined
considered identical. This analysis proves once again that the two DCP (they value).
test populations are not identical, and thus indicates the possible To test whether the two test populations are identical, a statisti-
existence of vertical confinement. cal analysis similar to that described earlier was performed. Ac-
Now, attention should be given to the two silt points in Figure 7. cording to the statistical t-test results, both populations are identi-
Since in the silt 4> is > 0 on the basis of the discussion in the engi- cal. However, this result should be accepted with caution because
neering background, it is conceivable that the two test points in the there is no assurance that in the test population (derived from vari-
silt are indeed related to the phenomenon of vertical confinement, ous sites with a wide range of strengths) the condition requiring that
whereas there is no reason for the confinement effect to take place the standard deviation from the mean does not change with the
in the rest of the test points, which are in clay. For this reason the change in the number of sites is indeed obtained.
character of the two populations should be reexamined following
exclusion of the two silt test points (marked separately in Figure 7).
The hypothesis that the two silt points do not belong to the
population was examined by similar linear regression without the
two points described earlier. The regression led to the following 60
parameters: a

a= 12.10 (Sa= 3.199), b = 0.418 (Sb= 0.086), R 2 = 0.771, i:l'; 50

0
and 0'2 = 9.005 ::0
a
An F-test of this regression leads to the conclusion that at a ~ 40
a
confidence level of 95 percent the model is significant and a linear Cl
a
correlation between x and y does exist. ~ 30
The conclusion is that the two points do not belong to the same ~
0 a Line of equality
population, because without them there is a significant linear corre-
u
lation. Finally, it can be assumed that the difference between the z 20
two populations (x,y) is a result of the friction created in the rod dur-
:::>
ing tilted penetration of the DCP rod through the granular material eJ
Cl 10
(as will be discussed later or of the friction created in the rod dur-
ing a collapse of the granular material on the rod surface during pen-
etration. A constrained regression through the graph origin was con- 0
ducted to obtain the extent of the friction's effect on the results, and 0 10 20 30 40 50 60
the following correlation was obtained: y = 0. 73x. This leads to the
DCP CONFINED , mm I blow
following relationship:
FIGURES Correlation between confined and unconfined
DCPunc = 1.34 DCPcon (8) DCP within clay subgrades.
6 TRANSPORTATION RESEARCH RECORD 1473

A regression analysis of the linear y = ax + b type leads to the DCPs = 1.55 DCPv (10)
following results:
where CBRv is the CBR value calculated from the DCP test con-
a= 2.86 (Sa= 3.75), b = 0.98 (Sb= 0.16), R2 = 0.59, ducted through the all-asphalt structure, and CBRs is the direct CBR
and cr2 = 48.23 value.
To te~t whether the increase in strength pre~ented in Equations 9
Examination of the confidence interval for the parameters a and and 10 is the result of vertical confinement caused by the rigid all-
b at a confidence level of 95 percent leads to the acceptance of the asphalt structure, a special series of tests was conducted in the lab-
hypothesis that a equals zero and b equals 1, that is, that there is no oratory as described in the next paragraph.
additive or multiplicative bias. Hence, the calculated values for a In the test series clay samples (liquid liniit, 66 percent; plasticity
and b, 2.86 and 0.98, respectively, are not significant. index, 43 percent} were compacted by 12 blows (4 samples) and 26
An analysis by the method of Grubbs (23) also indicates that both blows (18 samples) in the standard CBR mold. The moisture con-
the systematic ~rror and the standard deviation of the one popula- tents were higher than the plasticity limit. These clay samples were
tion's random error are identical to those of the other population. not saturated with water. Each pair of samples was tested with the
The final conclusion is that despite the variations shown in Fig- DCP instrument both without vertical pressure (free top surface)
ure 8, it seems that the two populations are identical. In other words and with the surface restrained by means of a thick metal plate with
the effects of the penetrating rod's surface friction are not felt in this a 3-cm-diameter hole. At the end of the test it was clearly observed
case. This is logical, since in clay material (in contrast to granular that when the sample was unrestrained there were vertical move-
material) the chances of preserving the gap between the sides of the ments and cracking around the area of penetration, whereas in the
hole created with the cone's penetration and the penetrating rod's case of the restrained samples, the upper soil surface remained
surface, whose radius is smaller, are better. · smooth and uncracked and without any vertical movement.
All of the various statistical tests indicate that restraining the sam-
ples does not affect the DCP results. This can also be seen from the
EFFECT OF CONFINEMENT plotted results presented in Figure 10. The regression model leads
BY RIGID STRUCTURE to the following parameters:

An examination of the DCP values of a clay subgrade confined be- a= 1.26 (Sa= 3.51), b = 1.01 (Sb= 0.114), R 2 = 0.90,
neath a 50-cm-thick all-asphaltic pavement has yielded deviant and cr2 = 8.73
strength results compared with the CBR values obtained by direct
testing. The DCP test was conducted in the subgrade following For a confidence level of 95 percent, a equals zero and b equals
drilling of the core for the entire depth of the all-asphalt structure. 1. This analysis does not follow the trend of results presented in Fig-
The res.ults are presented in Figure 9. ure 9. The explanation for this can be given by assuming that fric-
When the regression is restrained through the graph zero of tion forces do develop along the penetrating rod as a result of non-
Figure 9, the equation is: vertical penetration rather than because of the artificial restraining
of the clay in the laboratory or its restraint by the all-asphalt pave-
CBRv = 1.75 CBRs (9) ment at the site. To test this assumption a series of tests was con-
ducted in which the rod was inserted both vertically and at an angle
or of 15 degrees.

20

~~
Constrained regression
through graph origin
U::r;:
0.
Q~ DD
::s0 0IXl ::::>1¥
e: or.I)
::r:: E-
15

~ o-
::::>
1¥ E-
u
< ::r:: ~
E- E- ::r:: 10
b4=
IXl Q ll..
0fl) fl)

~~~
>o~
5
uz
1¥ E-

uE_
IXl E-

0
0 5 10 15 20
DIRECT CBR VALUE IN % DCP RESTRAINED SURFACE , mm I blow

FIGURE9 CBR results from clay subgrade beneath 50-cm- FIGURE 10 Effects of restrained sample surface on
thick all-asphalt structure. DCPvalues.
Livneh et al.
7

The statistical analysis indicates that there is a difference between 16

the two tests. The mean CBRv value for the vertical penetration is
6.6 percent, whereas the CBR, value for penetration at a 15-degree
14
angle is 8.9 percent. A constrained regression through the origin ~
0
produces the following correlation (Figure 11): ::0
12
CBRv = 0.78 CBRr (11) ~
~ 10
0
The meaning of this above regression is that the DCP value de-
creases (the strength increases) when the penetrating rod is inserted
at an angle. This fact can be used to explain the results presented in
I
u
0
6
8

Figure 9, since in many cases under actual deep DCP penetration, z

the rod tends to tilt at an angle ofup to 15 degrees. It should be noted :::>
~ 4
that the tilted penetration can be avoided by penetrating the DCP u
0
rod through a vertical supporting frame. This is accomplished by the
2
regular DCP device or by an automated DCP device (24).

2 4 6 8 10 12 14 16
EFFECT OF CONFINEMENT
BY ASPHALT LAYERS DCP CONFINED , mm I blow
FIGURE 12 Effects of asphalt confinement on DCP
The effect of confinement by the asphalt layers on the granular
values in granular material.
pavement layers was examined by conducting DCP tests on the base
course materials after drilling an asphalt core and comparing the re-
sults with those obtained from performance of the DCP test on these
materials following the removal of a wide asphalt strip from the of the granular material. Since this effect has been shown to exist,
same spot tested previously. In contrast to the case described in the it is now appropriate that the strength values of the granular mater-
preceding section, one can expect here an increase in the normal ial in the CBR or DCP field tests should be determined only after
DCP value compared with the confined DCP values in granular ma- the drilling of asphalt cores and not after the removal of a wide strip
terials, since the bearing capacity factor, Nvq• is far greater than 1. of asphalt.
Indeed, the results obtained indicate the following correlation, as
can also be seen in Figure 12:
SUMMARY
DCPunc = 1.84 DCPcon (12)
The DCP has become one of the most useful testing devices in pave-
The difference between the DCP values is statistically significant, ment evaluation in Israel and other parts of the world. Conse-
and thus, the asphalt does indeed have a confining effect on the state quently, a reliable true meaning of the results and an interpretation
of the results were needed. This paper summarized research work
that was dedicated to studying the effects of vertical confinement on
~

- ~
~
50 the DCP values of the subgrade and the granular pavement layers.
Specifically, four major effects were studied: vertical confinement
of granular layers, vertical confinement of cohesive layers, vertical
confinement of rigid structural layers, and the effect of vertical con-
u:i 40
..J finement of upper asphalt layers on the DCP values of the granular
~
layers beneath them.
~ On the basis of the theoretical analysis, experimental testing in
z 30 the laboratory and in the field, and statistical analysis, the following
0
~ conclusions were made.

~ 1. No vertical confinement effect by rigid pavement structure or

20
~ by upper cohesive layers on the DCP values (or strength) of lower
~
cohesive subgrade layers exists. Also, no vertical confinement ef-
0
:!:! 10 fect by the upper granular layers on the DCP values of the cohesive
< subgrade beneath them exists ..Any difference in the results between
~ the confined and unconfined DCP values in the rigid structure case
~
or in the case of the granular layers can be explained by the friction
u that developed in the DCP rod by tilted penetration or by a collapse
0 10 20 30 40 50 of the granular material on the rod surface during regular penetra-
DCP AT A PERPENDICULAR PENETRATION, mm/blow tion. This friction effect, which is also a function of the ratio be-
tween the cone tip and the rod diameters, can be quantitatively eval-
FIGURE 11 DCP values from vertical penetration and uated by torque measurements in the DCP device during penetration
penetration at a 15-degree angle. intervals.
8 TRANSPORTATION RESEARCH RECORD 1473

2. A vertical confinement effect by the upper asphaltic layers on 7. Buncher, M. S., and D. J. Christiansen. USAF New Contingency
the DCP values of the granular pavement layers does exist. Gener- Soils/Pavement Testing Van. ASCE Special Publication on Road and
Airport Pavement Response Monitoring Systems. ASCE, 1991, pp.
ally, these confinement effects cause a decrease in the DCP values, 27-40. .
thus increasing the structural strength measured. Since this is the 8. Webster, S. L., R.H. Grau, and T. P. Williams. Description and Appli-
true effect of the pavement structure, any DCP measurements for cation of Dual Mass Dynamic Cone Penetrometer. Instruction Report
pavement evaluation purposes should be performed through a nar- GL-92-3. U.S. Army Engineers Waterways Experimental Station, May
1992.
row boring in the asphalt layers and not after removal of a wide strip
9. de Beer, M. Developments in the Failure Criteria of South African
of asphalt. Mechanistic Design Procedure for Asphalt Pavements. Proc., 7th In-
3. The vertical confinement effects and the friction effects de- ternational Conference on Asphalt Pavements, Vol. 3, 1992, pp. 54-75.
scribed earlier, which can be quantitatively evaluated, should be 10. Livneh, M., I. Ishai, and N. A. Livneh. Carrying Capacity of Unsurfaced
taken into consideration when using the DCP method in pavement Runways for Low Volume Aircraft Traffic. Phase III. Application of the
Dynamic Cone Penetrometer. Preliminary Report. Transportation Re-
and subgrade evaluation. search Institute, Technion, Haifa, Israel, April 1990.
11. Kleyn, E.G. The Use of the Dynamic Cone Penetrometer (DCP). Re-
port 2/74. Transvaal Roads Department, 1975.
12. Kleyn, E.G., and M. J. J. Van Neerden. Using DCP Soundings to Op-
ACKNOWLEDGMENTS timize Pavement Rehabilitation. Presented at Annual Transportation
Convention, South Africa, 1983.
This work was conducted as part of a comprehensive project on the 13. Kleyn, E. G., and P. F. Savage. The Application of the Pavement DCP
carrying capacity of unsurfaced runways financed by the U.S. Air to Determine the Bearing Properties and Performance of Road Pave-
ments. Report L 7/82. Transvaal Roads _Department, South Africa, 1982.
Force Engineering & Services Center, Tyndall Air Force Base, 14. Kleyn, E.G., and P. F. Savage. The Application of the Pavement DCP
Florida. The authors thank that organization for the sponsorship and to Determine the Bearing Properties and Performance of Road Pave-
aid. Thanks are also due to the Israel Air Force, the Technion Soils ments. Proc., International Symposium on Bearing Capacity of Roads
& Roads Testing Laboratory, and also to A. Aines and F. Hirosh for & Airfields, Trondheim, Norway, 1982.
15. Livneh, M. Correlation Between the DCP and CBR Values. Publication
their technical and administrative assistance.
87-065. Transportation Research Institute, Technion, Haifa, Israel,
1987.
16. Van Vuuren, D. J. Rapid Determination of CBR with the Portable Dy-
namic Cone Penetrometer. The Rhodesian Engineer, Vol. 7, No. 5,
REFERENCES · 1969, pp. 852-854.
17. Livneh, M. The CBR Test with Lateral Pressures. Proc., 3rd Asian Re-
1. Kleyn, E.G., J. H. Maree, and .P. F. Savage. The Application of a gional Conference on Soil Mechanics and Foundation Engineering,
Portable Pavement Dynamic Cone Penetrometer to Determine In-Situ Vol. 1, 1967, pp. 293-297.
Bearing Properties of Road Pavement Layers & Subgrades in South 18. Livneh, M., and J. Greenstein. A Modified CBR Test for Granular
Africa. Proc., 2nd European Symposium on Penetration Testing, Am- Materials. Geotechnical Testing Journal, ASTM, Vol. 1, No. 3, 1978,
sterdam, May 1982. pp. 141-147.
2. Smith, R. B., and D. N. Pratt. Field Study of In-Situ California Bearing 19. Meyerhof, G. G. The Ultimate Bearing Capacity of Wedge-Shaped
Ratio and Dynamic Penetrometer Testing for Road Sub grade Investi- Foundations. Proc., 5th International Conference on Soils Mechanics
gation. Australian Road Subgrade Investigation. Australian Road Sub- & Foundation Engineering, Vol. 2, 1961.
grade Investigation, Australian Road Research ~oard, Vol. 13, No. 4, 20. Durnunoglu, H. T., and J. K. Mitchell. Influence of Penetrometer Char-
p. 198. acteristics on Static Penetration Resistance. Proc., European Sym-
3. Harrison, J. A. Correlation Between the California Bearing Ratio and posium on Penetration Testing, Stockholm, Sweden; 1974.
the Dynamic Cone Penetrometer Strength Measurements of Soils. 21. Nowatzki, E. A., and L. L. Karafiath. The Effect of Cone Angle on Pen-
.Proc., Institute of Civil Engineers, Part 2.83, Technical Note 463. etration Resistance. In Highway Research Record 405, HRB, National
Institute of Civil Engineers, 1987. Research Council, Washington, D.C., 1972.
4. Operating Instrunientationfor the TRRL Dynamic Cone Penetrometer. 22. Brinch-Hansen, J. A General Formula for Bearing Capacity. Bulletin
Information Note. Transport and Road Research Laboratory, 11. Danish Geotechnical Institute, Copenhagen, Denmark, 1961.
Crowthorne, Berkshire, United Kingdom, 1986. 23. Grubbs, F. E. Errors of Measurements, Precision, Accuracy and the Sta-
5. Livneh, M., and I. Ishai. Pavement and Material Evaluation by a Dy- tistical Comparison of Measurement Instruments. Technometrics, Vol.
namic Cone Penetrometer. Proc., 6th International Conference on the 15, No. 1, Feb. 1973, pp. 53-66.
Structural Design of Asphalt Pavements, Ann Arbor, Mich. July 1987, 24. Livneh, M., I. Ishai, and N. A. Livneh. Automated DCP vs. Manual
pp. 665-676. DCP in Carrying Capacity of Unsurfaced Runways. Project 115-101.
6. Chu, K. M. Determination of CBR and Elastic Modulus of Soils Using Transportation Research Institute, Technion, Haifa, Israel, 1992, 42 p.
a Portable Pavement Dynamic Cone Penetrometer. Proc., !st Interna-
tional Symposium on Penetration Testing-ISOPT-1, Orlando, Fla., Publication of this paper sponsored by Committee on Strength and Defor-
March 1988. mation Characteristics of Pavement Sections.