Design Stiffness of Subgrades Using The CBR Test
Design Stiffness of Subgrades Using The CBR Test
Design Stiffness of Subgrades Using The CBR Test
https://www.issmge.org/publications/online-library
Burt Look
Connell Wagner, QLD, Australia
Keywords: CBR, Elastic Modulus, Resilient Modulus, Repeat Load Test, Subgrade Strength
ABSTRACT
The design of pavements relies on the subgrade modulus, which is principally derived from
the California Bearing Ratio (CBR) test. This very common and simple test has many pitfalls
in its application. This is an index test and not a fundamental strength parameter. A
modulus conversion is then used and can be highly variable depending on the material type,
strain level, and its relation with surrounding material.
This paper illustrates the above issue by characterisation of a site using the CBR. This
involved zonation of areas and obtaining the relevant subgrade stiffness values. The
resilient modulus obtained from repeat load triaxial testing is compared with the simplified
modulus values derived by other traditional means.
1. INTRODUCTION
correlates to correlates to
[Resilient Modulus + Critical Strain + Pavement Life ] = Pavement Design Actual
Process
Modulus conversions are used to convert the CBR (a strength index) to the stiffness. These
conversions can be highly variable depending on the material type, loading conditions and
its relation with the surrounding material. The resilient modulus accounts for repeated
cycles of loading and unloading, while other modulus derivations are generally based on
one cycle of loading. The use of moduli derived from various tests and relationships are
discussed to show its variability, application, and limitations in characterisation of this site.
2. SITE ASSESSMENT
The Caboolture to Beerburrum rail site upgrade is approximately 13km in length, and is
located north of Brisbane, in southeast Queensland. The results of the initial 4-day
soakad CBR tests are tabulated in Figure 2, which also shows the moving average. This
moving average assessment provides a visual assessment of the changes occurring across
the site for zonation purposes. This provided the subgrade assessment for preliminary
design of the required capping and ballast layers. Design zones are compared in Table 1.
Site variability does not lend to the practical differentiating between a CBR 4% and 5%.
Therefore the refinements of Tables 1 were replaced with the simplified Table 2 zones.
55
cOmmOn grOund 07
RCR Cut North of
50 Blackburn Road
45
Mansfield RCR Cut South of
Road Cut Blackburn Road
40
35
30
CBR (%)
25
20
15
10
0
51440
51940
52440
52940
53440
53940
54440
54940
55440
55940
56440
56940
57440
57940
58440
58940
59440
59940
60440
60940
61440
61940
62440
62940
63440
63940
64440
Chainage (m)
The CBR test is a relative strength index that compares the penetration value with that of a
well-graded crushed stone of CBR 100%. It is a non-fundamental strength parameter that
provides a comparison of strength between layers of materials. The laboratory CBR value
may not be achieved in the field even with the same compaction, moisture and material
depending on the base / subgrade strength ratios (Hammitt,1970). This is an empirical test
carried out at a high strain level and low strain rates, while subgrades below pavements
normally experience relatively low strain levels and higher strain rates.
Pressures on pavement surfacing varies from 250 kPa to 550kPa for heavily loaded cars to
truck tyres, respectively, but in the past decade truck tyre pressures can be of the order of
750 kPa (comparable to train loadings). Aircrafts can be over 1500kPa. The subgrade
experiences a lower load as the stress is distributed through the overlying pavement.
During construction the subgrade experiences its most critical loading condition. In all
cases of pressures for pavement or subgrade materials, the same CBR static pressure of 6.9
MPa to 10.3 MPa is used for the 2.5mm to 5.0mm penetration of the plunger, respectively.
The pavement design requires modulus as an input, with the assumption that by having
suitable soil support strength values then the pavement’s required stiffness can be derived.
However strength and stiffness are not the same. The permanent shear deformation
cOmmOn grOund 07
induced by the plunger suggests the test relates more directly to strength than stiffness.
Thus there are significant sources of errors introduced into the design even when the
samples are correctly tested
1. The CBR load does not represent the current applied loads
2. Stiffness is not the same as strength
3. The CBR - modulus relationship has a significant variability depending on the
material, stress and strength range
4. Static and dynamic modulus values are different
5. The relationship between the soaked CBR value and the equilibrium CBR value
Nevertheless the CBR is one of the most common tests available with many supervisors and
engineers with a vast experience on its application, although the above illustrates that it
persists beyond the sophisticated analysis required in elastic layered pavement designs.
The CBR value only in a test certificate discards much useful data that can be used by the
designer. Both strength and movement must be used to characterise a site as failure from
movement can occur despite an adequate CBR strength. At CBR values less than 10%, the
swell characteristics needs to be considered. Look (2005) discusses these issues and the
relevance of equilibrium moisture conditions. The issue of designing for swell is not
discussed further in this paper.
The following methods of analysis were used to compare various modulus conditions
derived from the CBR test.
1. Using the CBR test results for a number of tests, the typical E = 10 CBR value was
adopted (Heukolm and Klomp, 1962). This is the standard approach for CBR values
< 10%. Other relationships were also used as shown in Figure 5 (AASHTO, 1993;
Angell, 1988; NAASRA, 1980; Powell et al., 1984).
2. Using data from CBR test results across south east Queensland, the stress strain
curve was derived, and the modulus values were derived from these plots (Figure 3)
3. For 8 specific samples for this site (representing 4 terrain conditions) and with 2
different levels of compaction the Resilient Modulus Tests were carried out. These
8 test results were compared with the above modulus conditions.
4. Additionally, tubes were pushed in these same 8 CBR samples and tested under the
similar confining pressure to derive the unconsolidated undrained test strength.
The stress strain response was used to obtain the static modulus for these samples.
5. RESULTS OF ANALYSIS
cOmmOn grOund 07
5.1 Modulus derived directly from CBR results
The CBR test involves plotting the force required for various penetration values. This curve
may be corrected depending on the shape of the curve to determine the force at 2.5mm
and 5.0mm penetration into the sample. Given that the test size is 178mm, then indicative
modulus values were obtained by determining the stress at the deflection into the sample
divided by the sample height. Strictly speaking this is not the stress strain response as the
CBR plunger is pushed into the sample rather than the sample itself deforming.
Figure 3 plots CBR data from the site and shows that at low strains (0.8%), the Modulus
derived in this manner is ~ 5 to 4 u CBR for 1.4% to 2.8% strains, respectively. This is
significantly lower than the to 10 u CBR relationship (Heukolm and Klomp,1962). A similar
ratio was obtained for the unsoaked Modulus / CBR for the samples tested for this project.
275
250 y = 4.49x
225 2
R = 0.99
y = 4.83x 2.5
200
Modulus (MPa)
2
175 R = 0.95 5
150 1.5
125 Linear (2.5)
100 y = 3.70x Linear (5)
75 2
R = 0.99 Linear (1.5)
50
25
0
0 10 20 30 40 50 60
CBR (%)
For this site, samples were referenced tested using the maximum dry density (MDD) and the
optimum moisture content (OMC) for standard compaction. Tests targeted the OMC and
95% MDD and 98% MDD. At the latter compaction level the strength was between 10% to 20%
higher strength than those compacted to 95% MDD. However soaked values tended to be
less sensitive to strength changes.
Following the CBR and repeat load tests, tube samples were obtained and tested in an
undrained triaxial condition using the same confining pressures as the repeat load test.
The variation with strain for samples at various peak stress values is summarised in Table 4.
These relatively low E / Cu ratio as compared to typical values given in Jamiolkowski et al.
(1979) are indicative of the high “overconsolidation ratio” for compacted samples.
The derived Elastic Modulus / CBR relationship of ~ 5 u CBR is consistent with the previous
cOmmOn grOund 07
approach, i.e. using the CBR raw data only, and accounting for plunger penetration. This is
derived from the stress strain curves, and is not representative of the resilient modulus
required for pavement designs. This modulus is discussed in the next section.
Australian Standard (1995) provides the method of testing for determination of the resilient
modulus (Mr) and permanent deformation of unbound materials. This testing was carried
out to characterise the subgrade material and for rail rather than pavement loading,
therefore some variation to the applied stress levels needed to be considered. The results
can be sensitive to the stress and stress ratios applied. This testing was budget limited in
not being able to apply various levels of deviator stresses to examine this effect. Alderson
and Robinson (2006) show the Mr test had a repeatability of 19% and reproducibility of 43%,
and suggest that due to such high variability that the test is not suitable for inclusion in
specifications in its current form. However CBR testing (from which correlations are
derived) has a coefficient of variation of 17% to 58% (Lee et al., 1983)
Initial samples exhibited early failure at the above loads, and a preconditioning using 250
kPa vertical stress for the first 50 cycles of loading was carried out for the subsequent
samples. The resilient strain varied from 0.2% to 0.8% for the samples tested, thus
adopting the low strain elastic modulus strength values in Table 4 may be justified. Table 5
provides a summary of the results for the samples tested.
These resilient modulus values are significantly higher than those obtained from the
previous methods, which applies for static elastic modulus to strength ratios. For this test
data using a resilient / elastic ratio of 2, was considered a conservative lower value. Figure
4 compares these modulus values with established relationships. This provided some
confidence in design for this site for not adopting the lower correlations shown.
6. Conclusion
The paper outlines the procedure adopted for a rail line upgrade, from zonation of areas in
terms of the CBR value to establishing the relationship between the CBR and modulus
values. Proper characterisation of the site subgrade properties involves also considering
the CBR test, with the appropriate level of compaction, the type of test and the
appropriate conversion factor to determine the design resilient modulus, which is different
from the static elastic modulus.
For this site the resilient modulus is approximately twice the elastic modulus. The soaked
strength results were 10% to 50% of the unsoaked result. The soaked to the unsoaked
condition E/Cu ratios were ~ 50% from various methods. The strength of the samples
compacted to 98% MDD is 10% to 20 % higher than the value derived from compaction at 95%
MDD (Standard).
cOmmOn grOund 07
450
Heukelom and Klomp (1962)
400
Powell et al. (1984)
Resilient Modulus (MPa)
350
AASHTO (1993) - Base
300 Course
Queensland Main Roads
250 (1988)
NAASRA (1980)
200
Test Data
150
7. REFERENCES
AASHTO (1993). Guide for design of pavement Structures. American Association of State
Highway and Transportation Officials, Washington, DC
Angell, D.J. (1988). Technical Basis for the Pavement Design Manual. Queensland Main
Roads
Australian Standard (1995). “Soil Strength and consolidation tests – Determination of the
resilient modulus and permanent deformation of granular unbound materials”.
AS1289.6.8.1 – 1995
Jaimolkowski, M., Lancellotta, R., Pasqualini E., Marchetti, S. and Nova, R. (1979). Design
Parameters for soft clays. General Report, Proceedings 7th European Conference on Sol
Mechanics and Foundation Engineering, No, 5, pp 27 – 57
Hammitt, G.M. (1970). Thickness requirement for unsurfaced roads and airfields, bare base
support. Report S – 705, U.S. Army Engineering Waterways, Experiment Station, Vicksburg
Lee, I.K., White W. and Ingles O.G. (1983). Geotechnical Engineering. Pitman Publishers