THE CALCULATIONOFLABORATORYANDIN-SITUVALUESOF
CALIFORNIABEARINGRATIOFROMBEARINGCAPACITYDATA*
bY
W. P. M. BLACK, BSc.
SYNOPSIS
The Paper briefly reviews the factors which affect Cet article esquisse les facteurs qui influent su*
the results of in-situ and laboratory California les rCsultats des essais CBR in situ et au laboratoire.
bearing ratio tests. A relation between CBR and On developpe un rapport entre le CBR et la capacit6
bearing capacity is developed which enables the portante qui permet de calculer la valeur CBR in
in-situ CBR value to be calculated from a knowledge situ si on connait la coh6sion, l’angle de frottement
of the cohesion, true angle of internal friction, and
suction of the soil. A method of correcting the interne veritable et la suction du sol. On propose
in-situ value to take into account the confining une methode pour corriger la valeur in situ afin de
action of the mould used in laboratory tests is tenir compte de la contrainte exercee par le moule
proposed. employ6 dans les essais au laboratoire.
Laboratory investigations made on a single-size On decrit des recherches faites au laboratoire avec
sand and on a heavy clay are reported in which close un sable Blementaire et avec une argile lourde et on
agreement was found between computed and meas- constate un bon accord entre les valeurs calculees et
wed CBR values. les valeurs mesurees.
IFTRODLCTION
The original procedure for determining the California bearing ratio of soil was concisely
defined and involved no ambiguity (Porter, 1938). 1 With the development of the California
bearing ratio method of flexible-pavement design and its wider adoption, alternative test
procedures have been introduced. The equivalence of these procedures and the conditions
under which they should be used have not been sufficiently closely investigated and this has
led to increasing confusion.
CBR tests can be conducted:
(a) on recompacted samples of soil in the standard laboratory mould;
(4 on “undisturbed” samples cut from the ground and trimmed to fit as closely as
possible into the standard mould; and
(c) in situ on the surface of a soil formation.
Even where identical average moisture contents and dry densities are achieved, tests (a)
and (b) do not necessarily give the same CBR value. This is in part because soil samples
compacted to nominally the same conditions may have very different pore-water pressures and
this is one important factor affecting the test rest&. Some methods of compaction lead to
positive pore pressures in the remoulded soil and this can result in abnormally low values.
The same considerations apply in comparing tests (a) and (c) but here another important
factor is the confining effect of the rigid mould in which the laboratory tests is carried out, an
effect which is absent in the in-situ tests. This may result, particularly in granular soils, in
laboratory results many times greater than the in-situ values, even where the moisture
content, dry density, and initial pore-pressure conditions are as closely identical as can be
achieved in practice.
Table 1 gives a comparison between laboratory tests carried out on remoulded samples and
corresponding in-situ values determined on a range of soils. The tests were in each case
carried out at the same nominal moisture contents and dry densities. A close examination of
these data suggests that all the factors reviewed above affecting the relation between laboratory
* Crown copyright reserved.
r The references are given on p. 2 I.
14
Downloaded by [ Syracuse University] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
BEARING CAPACITY AND CBR 15
and in-situ tests have been operative, their relative importance depending not only on the soil
type, but also on its moisture condition and dry density.
It is not proposed in the Paper to review in detail all the factors affecting the CBR test.
The penetration test itself is, however, considered from the point of view of bearing capacity
I
Table 1. Comparison of laboratory (remoulded) and in-situ CBR values
Soil type I Source of data I Dry Moisture CBR: per cent
density: content: _
lb/cm ft per cent remoulded in situ
____
Heavy clay Road Research Laboratory 95 243 8.9 7.9
(LL 69 PL 25)
Clay do. 96 25.1 3.9 3.0
(LL 59 PL 22)
Silty clay U.S. \Vater\l-ays Exp. Stn 109 19.3 2.0 12
(LL 37 PL 23) 107 19.0 , 5.0
104 j 16.1 22 :;
Sandy clay Road Research Laboratory 95 19.2 2.2 3.1
(LL 30 PL 18) I
Clayey sand U.S. JVaterways Exp. Stn 116 12.2 14 7.0
114 , 12.6 10 9.0
109 10.0 12 18
Single-size sand Road Research Laboratory 98 7.5
Crushed slag do. 140 4:; 44
- - - -
theory, and it is shown that this approach leads to a closer understanding of some of the differ-
ences that occur between the values obtained from the different types of test.
MODE OF FAILURE IN THE CBR TEST
Fig. l(a) shows the CBR plunger in relation to the soil sample contained in the standard
6-in.-dia. mould. Bearing-capacity failure occurs according to the plasticity theory when the
Rankine passive pressure in the soil wedges around the plunger (ABCD and EDFG) is reached,
allowing the soil to heave upwards and the plunger to move down at constant pressure. A
finite lateral strain must occur to bring the soil to its passive pressure state, the maximum
strain occurring when the initial lateral pressure is equal to the Rankine active pressure.
Therefore the deformation to failure will vary depending on the method of compaction which
largely controls the lateral pressure in the CBR mould. The possible extent of this variation
is illustrated in Fig. 1 (b), which shows the load/deformation curves for laboratory CBR tests
carried out on three specimens of an air-dry sand all compacted to the same nominal dry
density, but using three different methods of compaction. Compacted by the Proctor
rammer, the soil was particularly unstable and failed to produce any appreciable lateral
pressure in the mould, hence the large strain to failure. By compacting under a large static
load the penetration at failure was reduced to one-sixth of the value for the soil compacted by
the rammer. The third method of compaction used produced an intermediate result. In
the three tests the bearing capacity did not vary by more than 16%, and this variation was
probably due mainly to the frictional resistance associated with the depth of penetration to
achieve failure, when an unsuitable method of compaction (for this type of soil) was used.
Since the CBR test is a measure of the deformation modulus in a particular condition and
not of the ultimate bearing capacity it is subject to variations as indicated in the extreme case
illustrated by Fig. l(b). The relation between CBR value and ultimate bearing capacity
also depends on the type of soil as well as on the method of compaction used. In practice,
however, the stress applied to the CBR plunger to give the standard penetration of O-1 in. is
Downloaded by [ Syracuse University] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
16 W. P. M. BLACK
C.B.R. plunger
6in. dia.
PENETRATION - inches
Fig. 1. (a) Assumed failure surface beneath CBR plunger. (b)? CBR
load/penetration curves for air dry single-size sand at 100 lb@. ft
dry density compacted by three methods
often close to the ultimate bearing capacity (as in curve 2 of Fig. 1 (b)). When this is the case,
CBR=G
Hence,
CBR = 4ut
where qu is the ultimate bearing capacity.
CORRELATION OF BEARING CAPACITY AND CBR
It has been shown by Meyerhof (1951) and others that on a model scale the ultimate
bearing capacity for a circular footing can be estimated from a formula of the type given by
Terzaghi (1943) :
qu = 1~2cN, fyDN, +O-6yRN, . . . . . . (2)
* The 1,000 arises because the stress necessary to cause 0.1-k. penetration in the “standard” crushed
stone reference material used as the criterion for evaluating CBR values was 1,000 lb/sq. in.
t As previously used by Wilson, 1950.
Downloaded by [ Syracuse University] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
BEARING CAPACITY AND CBR 17
105 lb/cult/
. Mtasured C.0.R:S
- Estimated C.B.R,‘S
--- - Estimated in-situ C.aFf:S
0.1-
0 16 18 20 22 24
MOISTURE CONTENT - PEP cent
Fig. 2. Comparison of the estimated relations between CBR, moisture
content, and dry density of Sulehay sand with measured CBR’s in a
CBR mould
where c = cohesion of the soil
y = bulk density of the soil
D = depth of the footing
R = radius of footing
and N,, N,, and NY are bearing-capacity factors depending only on the friction angle of the
soil +.
The last term can be neglected for a footing of diameter as small as 2 in. Where the suc-
tion and effective stress are identical, the cohesion c in the first term can be replaced by s tan 4
where s is the suction of the soil at failure and 4 is the true angle of internal friction. In the
second term yD is equal to the pressure exerted by the surcharge weights when the expression
is applied to the CBR test. NC, N,, and N, are factors depending only on + and have values
given by Meyerhof (1955).
In applying equation (2) to the standard laboratory CBR test, the confining effect of the
mould must be considered. Table 1 shows that in the case of clays any such confining effect if
present is insufficient to mask the opposing influence of factors arising from disturbance or the
development of excessive pore pressures. For granular materials the apparent increase in
Downloaded by [ Syracuse University] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
18 W. P.M.BLACK
CBR value may be large, as is shown from the measurements made on the single-size sand and
on the slag.
A semi-empirical correction of the bearing-capacity approach, to take into account the
presence of the mould, can be made as follows. In Fig. 1 (a), the section of the failure surface
is assumed to be circular with centres at A and E. With this assumption the shear surface
does not intersect the mould for angles of 4 less than 30”, and no correction for mould restraint
is necessary. For greater values of 4 the shear surface intersects the pot on a circumference
through C. If it is assumed that the presence of the pot in this case does not affect the shape
of the shear surface in the soil (and this is unlikely to be other than only approximately true),
the “mould effect ” would be proportional to (a) the area 2 of the mould between the circum-
ference through C and the surface of the soil, (b) the pressure exerted by the soil on the inner
surface of the mould during the test, and (c) the shear strength of the soil, under the conditions
prevailing in the test.
MOISTURE CONTENT - per cent
Fig. 3. Comparison of estimated relations between in-situ CBR,
moisture content, and dry density with measured values-single-
sized sand
The pressure exerted by the soil at the rigid boundary of the pot would be of the order of
twice the passive pressure P (Weiskopf, 1945) defined by:
P = yD tans (45 + $/2) + 2c tan (45 + +/2) . . . . . (3)
The extra load (arising from the confinement of the mould) to be applied to the plunger to
cause bearing-capacity failure can be estimated by taking moments about the edge of the
plunger. If the additional load on the plunger is Lez then:
L,,xl = 2Zk(C + 2P tan 4)
and the additional item on the 3-sq.-in. plunger is given by:
S,, = 2/3Zk(c + 2P tan 4) . . . . . . . . (4)
where Z is the internal area of the mould within the shear zone and P is defined as in equation
(3), and k is constant.
Downloaded by [ Syracuse University] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
BEARING CAPACITY AND CBR 19
EXPERIMENTAL RESULTS
To establish the degree of accuracy with which laboratory and in-situ CBR values could be
estimated using the bearing-capacity approach, tests were carried out on a single-size sand
(of particle-size distribution between the No. 72 and No. 200 B.S. sieves). The tests included
shear box determinations of c and 4 and measurements of the suction, s.
The shear box measurements carried out at dry densities of 90,95, 100, and 105 lb/cu. ft,
on (a) air-dry soil and (b) samples prepared to a moisture content of lo%, showed that the
relation between 4 and dry density was not materially affected by the change in moisture
content. This relation is shown inset on Fig. 2. The cohesion was of course influenced by
30
f ’ E
510
z’ ::
I I ‘9
. 0
D
2, 30 35 40 45 50 55
MOISTVRE CONTENT-per cent PLASTlClTI INDEX- want
Relatm between 6, momture content and PlaSticitY index
(0 050”)
VT
v-
I
lD 12 14 m 18 2D 22 24
rl 26 28
I I 30 32
MOISTURE CONTENT - percent‘
Fig. 4. Comparison of estimated CBR’s of saturated London Clay with
the measured relation as given by Davis (1949)
moisture content and measurements of cohesion were made at various dry densities and
moisture contents. However, to reduce the amount of shear box testing the relation
c = s tan + was used where applicable to evaluate the cohesion, the suction, s, being measured
by the rapid method (Croney, Coleman, and Bridge, 1952). At low moisture contents where
the soil was grossly desaturated shear box determinations of c were made. These values of
c and 4 were used in conjunction with Meyerhof’s relationships between NC, N,, and + (Meyer-
hof, 1955), to deduce the bearing capacity, and the equivalent in-situ CBR was calculated as
O-1 qu. The equivalent plunger stress was then corrected in accordance with equation (4),
and the corresponding laboratory CBR values deduced.
A series of laboratory CBR tests was conducted on the same sand using dry densities of
90, 95, 100, and 105 Ib/cu. ft, a method of dynamic compaction being devised which gave a
****
Downloaded by [ Syracuse University] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
20 W. F. M. BLACK
load/penetration curve similar in shape to CBR curves measured in situ and to curve 2,
Fig. l(b). The experimentally determined CBR values are shown as points on Fig. 2. In
general, three separate determinations were made at each dry density and moisture content,
and the results show how closely reproducible the CBR test can be when adequate care is
taken with the sample preparation. The dotted curves show the computed in-situ CBR/
moisture content/dry density relations deduced from equation (1). In the full curves these
relations have been corrected for mould restraint, using equation (4).
To obtain the close agreement between the measured and estimated laboratory CBR
values, constant K = 1.4 was used in the right-hand side of equation (4). It is interesting to
note that when the Terzaghi equation and values of NC, N,, and 4 for a circular footing were
used, this constant approximated to unity.
The single-size sand used in these experiments was employed as a sub-base material in one
of the Road Research Laboratory’s full-scale pavement design experiments. The sand was
compacted using a smooth-wheel vibrating roller and a considerable number of in-situ CBR
measurements were made during the progress of the work. Fig. 3 shows the average moisture
content, dry density, and in-situ CBR values for each of thirty-three sections of sub-base.
The computed in-situ CBR relations given in Fig. 2 are replotted for reference: the computed
curves agree with the experimental data within the limits of accuracy of the in-situ tests.
Although no similar comparative tests have been carried out on clays, some information
on the accuracy with which laboratory CBR values can be estimated from shear strength and
suction data has been obtained using data published by Davis (1949) and Gibson (1943). Fig.
4 shows the relations between laboratory CBR values and moisture content for London Clay
at drv densities of 90. 95, 100, and 105 lb/cu. ft determined by Davis. With zero surcharge,
equaiion (2) becomes:
qu = 1.2 CAT,
0 4 6 12 lb 10 24
/ I I
0 IO 20 IO 40 50 60 70 00 90 100 I
MOISTURE CONTENT - ptrcrnt
Fig. 5. Suction/moisture content and shrinkage relation for London Clay
Downloaded by [ Syracuse University] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
BEARING CAPACITY AND CBR 21
or for a saturated clay,
qu = 1 a2 s tan #N,.
Suction tests carried out on the same clay are shown in Fig. 5. The suction corresponding to
remoulded samples compacted to zero air voids will be represented by points on the dotted
curve referring to the soil in a continuously disturbed condition. From this curve the suctions
corresponding to the moisture contents of the intersections of Davis’s curves with the satura-
tion line can be deduced. Further, Gibson has shown that for cohesive soils the true angle
of friction is approximately reIated to the plasticity index and that the value does not vary
greatly with moisture content, Fig. 4 (inset). Using Gibson’s value of 4 corresponding to the
plasticity index of the clay (13”) and the deduced values of suction, values of qu on the satura-
tion line were computed. CBR values, assuming CBR = 0.1 qu, are plotted on Fig. 4.
Because of the low vaIue of 4 no correction for mould restraint is necessary. Data are not
yet available to allow values off the saturation line to be calculated but the agreement is
sufficiently close to encourage further experimental work.
CONCLUSION
The principal conclusion from the work is that both for sands and clays laboratory and
in-situ CBR values can be calculated from measurements of cohesion, angle of shearing
resistance, and suction. This approach should assist in the interpretation of CBR measure-
ments and clarify the circumstances in which laboratory and in-situ tests can be used in the
design of flexible pavements.
ACKNOWLEDGEMENT
The work described in the Paper was carried out as part of the programme of the Road
Research Board of the Department of Scientific and Industrial Research. The Paper is
published by permission of the Director of Road Research.
REFERENCES
CRONEY, D., J. D. COLEMAN,and PAMELA A. BRIDGE, 1952. “ The suction of moisture held in soil and other
porous materials.” Dep. D.S.I.R. Rd Res. Tech. Pap. No. 24.
DAVIS, E. H., 1949. “The California bearing ratio method for the design of flexible roads and runways.”
Ghotechnique, Land., I : 4 : 249-263.
GIBSON, R. E., 1953. “ Experimental determination of the true cohesion and true angle of internal friction
in clays.” Proc. 3rd Int. Conf. Soil Mech. and Found. Engng, 1 : 126-131.
MEYERHOF. G. G., 1951. “The ultimate bearing capacity of foundations.” Ge’otechnique, Lond.. 2 : 4 : 301-
331.
MEYERHOF, G. G., 1955. “Influence of roughness of base and ground-water conditions on the ultimate
bearing capacity of foundations.” Gbtechnique, Lond., 5 : 3 : 227-242.
PORTER, 0. J., 1938. “ The preparation of subgrades.” Proc. Highm. Res. Bd., Wash., 18 : 2 : 323-331.
TERZAGHI, K., 1943. “Theoretical soil mechanics.” John Wiley, New York.
WEISKOPF, WALTER H., 1945. “Stresses in soil under a foundation.” J. Franklin Inst., 239 : 6 : 446-465.
WILSON, G., and G. M. J. WILLIAMS, 1950. “ Pavement bearing capacity computed by theory of layered
systems.” Proc. Amer. Sot. civ. Engrs, 76. Separate 16.
Downloaded by [ Syracuse University] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
