Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

The strength of well graded cohesive fills


The engineering behaviour of fill material derived from well graded Boulder Clay has posed problems for civil engineers concerned with road design and construction. Small increases in the moisture content of the fill material can result in significant loss of shear strength. This factor can have serious implications for pavement performance when upper layers of the subgrade material are affected. A series of compaction, undrained shear strength and CBR tests were performed on samples selected to simulate well graded Boulder Clay fill material. The reduction in strength of the fill with increasing moisture content was investigated; the concept of matrix moisture content was examined. A correlation between undrained shear strength and CBR is proposed.


The deleterious effects of water on clay fills are well documented. The combination of a well graded cohesive fill such as Boulder Clay and heavy intermittent rainfall during the earthworks season has a great influence on road design and construction activities in Scotland. The rate of deterioration of the strength of Boulder Clay during wet weather can be dramatic; the actual CBR value of the material at formation level at the time of construction can be significantly lower than the CBR values measured on samples obtained during the ground investigation.

This paper presents the results of laboratory tests which have been performed on simulated Boulder Clay fill material to quantify and explain the loss in strength due to increasing moisture content. A relationship between undrained shear strength (cu) and California Bearing Ratio (CBR) is suggested.

The CBR test

The essential features of the test apparatus are shown in Figure 1. The test, which is fully described in BS1377(1990) consists of loading a sample of soil at a rate of 1mm/minute with a circular plunger of cross sectional area equal to 1,935mm2. Load versus penetration curves are plotted and the actual load values at 2.5mm and 5mm penetration are noted. These load values are compared with the standard loads of 13.2kN and 20kN. The CBR of the soil is expressed as a percentage of the standard load, the larger value being used for design purposes. For a particular soil the CBR value generally increases as dry density increases and as moisture content decreases. The test is empirical and was adopted in the UK during the 1950s.

Unconsolidated undrained shear strength test

The essential features of the test apparatus are shown in Figure 2. Samples of soil 100mm in diameter by 200mm high were sheared to failure by applying an axial load at a rate of strain of 2% per minute, as described in BS1377(1990). Since no drainage was permitted dur- ing shearing, total stress

conditions applied. Pore water pres- sures were not recorded.

Relationship between CBR and other soil parameters

Considerable research has been carried out in an attempt to relate the CBR value of a soil to parameters such as moisture content, plasticity, suction and shear strength. One approach has considered the CBR test to simulate bearing capacity failure under a circular footing; since the load is applied rapidly, undrained shear failure is implied for cohesive soils and hence it may be assumed that u = 0.

Black (1962) investigated the relationship between CBR and plasticity data for cohesive soils and stated that experience had shown that at a penetration of 2.5mm bearing capacity failure of the test soil had generally occurred. Black demonstrated the relationship between CBR and moisture content for six remoulded inorganic clay soils and proposed a method which allowed engineers to estimate the CBR of a soil from its liquid limit, plastic limit and natural moisture content. Other relationships concerned suction and degree of saturation.

Black and Lister(1979) published results of an investigation into the strength of clay fill subgrades and discussed relationships between CBR, moisture content, plasticity, suction and undrained shear strength of cohesive soils.

CBR = (stress on plunger)69 {1}

at 2.5mm penetration where stress is in kPa. This is deduced from the area of the plunger = 1,935mm2 and the standard force applied at 2.5mm penetration = 13.2kN. Black and Lister referred to model footing studies by Skempton (1951) which had indicated that at 2.5mm penetration the stress on the plunger is equal to about half the ultimate bearing capacity [Qult] of the soil.

Hence CBR = (0.5*Qult)69 = Qult138{2}

The ultimate bearing capacity of a circular footing placed on the surface of a soil is Qult 6.0*cu

Hence CBR = cu23 {3}

or cu = 23*CBR {4}

This implies a linear relationship between undrained shear strength and CBR value. Black and Lister suggested a minor modification to the formula, based on the suction method :-

cu = 23*(CBR + 1) {5}

The soils considered in their study were silty or sandy clays such as London clay, brickearths and red coffee soils.

Black (1979) published results of an investigation into the strength of clay subgrades using a penetrometer. The relationship between CBR and undrained shear strength (cu) was considered with respect to theoretical ultimate bearing capacity of a circular footing and the relationship CBR = cu23 was restated.

A table of results was included which summarised the work of various authors and this showed that


Undisturbed samples : cu/CBR = 8.6 to 11.0

Remoulded samples : cu/CBR = 24.5 to 27.6

The equation cu = 23.0 *CBR {6}

was proposed for remoulded soils and the equation

cu = 11.5*CBR {7}

was suggested for undisturbed overconsolidated soils.

Brown, Loach and O'Reilly (1987) considered the relationships between CBR and undrained shear strength although the main effort was directed towards investigating repeated loading of fine grained soils.

The relationship : cu (kPa) = 7.8*CBR {8}

was obtained using shear vane and pocket penetrometer equipment on samples which had been saturated anisotropically and overconsolidated.

In addition, samples were prepared by mixing dry soil and water and compacting them into a CBR mould. The undrained shear strength was measured using the pocket penetrometer and the relationship was found to be

cu (kPa) = 34*CBR {9}

A less satisfactory relationship was obtained using the shear vane, the correlation being of the form :

cu = k*CBR {10}

The value of 'k' is a function of soil type and varies between 11.5 for Gault Clay to 20 for Keuper Marl. Previously Black (1979) had suggested a value of 23 for remoulded soil.

Ervin (1993) presented results demonstrating the influence of moulding water content on the undrained shear strength of three compacted clay soils being placed in a dam. The moisture contents were varied from 2% below to 4% above standard optimum moisture content. The relationship between undrained shear strength and moisture variation from optimum demonstrated a steady near linear reduction in strength with increasing moisture.

Jones and Greenwood (1993) reported on relationship testing for acceptability of cohesive fill from three sites in eastern England. The soils were clays varying from low to high plasticity. Linear relationships between moisture content and the logarithm of undrained shear strength and CBR were obtained. A linear relationship of the form cu = 23*CBR was obtained for the clay of intermediate to high plasticity.

Laboratory testing of simulated Boulder Clays

Material preparation

Three grading curves were chosen to represent generally well graded boulder clay. Test samples were prepared by combining sub-rounded sand and gravel particles with a silty clay obtained from the brick pit at Errol, Tayside. The particle size distribution of the samples was chosen such that the materials were well graded and contained 25%, 35% and 45% silty clay respectively; the remainder of each sample consisted of gravel finer than 20mm and sand. The grading curves of the silty clay and the three test materials are shown in Figure 3. The silty clay content of the test samples henceforth is referred to as '% fines'.

Results of Laboratory Tests

The results of soil classification tests are summarised in Table 1 and plotted graphically in Figure 4. Since liquid and plastic limits are determined on soil particles passing the 0.425mm size sieve, the reduction in plasticity as the fines content reduces is directly related to the increasing amount of fine and medium sand in the test samples. The rate of reduction in the value of the liquid limit with the addition of fine and medium sand is much greater than the effect of sand on the plastic limit. All four test results plot above the 'A' line and are classified as clays of low to intermediate plasticity.

The relationship between dry density and moisture content (compaction curve) was established for each of the sample types and these are shown in Figure 5.

The undrained shear strengths and CBR values of each sample type were measured at combinations of dry density and moisture content representing points on the compaction curves. The samples were prepared at the required moisture content and to the required density by static compression in steel moulds, taking care to minimise the presence of density variations within the samples. It was inevitable that the density and moisture content of samples did not exactly match the target values from the compaction curves but close agreement was achieved. The undrained shear strength was determined from a single sample at a cell pressure of 50kPa. The value of cu has been taken to be half the deviator stress at failure. This is truly valid only when u is zero which theoretically occurs when the degree of saturation of the sample approaches 100%.

The relationships between CBR test results, density and increasing moisture content are shown in Table 2.

The relationships between undrained shear strength, density and moisture content are shown in Table 3.

Analysis of Laboratory Test Results

It can be seen from the data in Table 4 that the optimum moisture contents are 6% to 7% below the plastic limits of each sample. The air content of the samples at maximum dry density is low at around 3% and would be acceptable for fill material in embankment construction. The degree of saturation is more variable, ranging from 84% to 95%.

The relationship between undrained shear strength and CBR with increasing moisture content is shown in Figures 6 and 7; the reduction in both cu and CBR with increasing moisture content is marked for all three sample types. Typically a 2% increase in moisture content above optimum moisture content results in a 40% to 46% reduction in shear strength and a 50% to 60% reduction in CBR value.

The relationship between cu and CBR for the materials chosen to represent boulder clay is shown in Figure 8. The parabolic curve

cu = 27.15*CBR0.586 11}

is proposed; this curve has a correlation coefficient of 0.953.


The marked reduction in soil strength, measured either in terms of undrained shear strength or CBR value needs to be considered in more detail. Several researchers and practising engineers in the past have considered how the additional water is absorbed by well graded cohesive soils. Laboratory tests during this investigation indicated that the saturation moisture content of the sand and gravel fraction did not exceed 5%. Therefore at moisture contents in excess of 5%, water must be preferentially absorbed by the 'fine' particles - namely the silt and clay fraction. The matrix moisture content is defined as the moisture content of the particles finer than 0.06mm, assuming that the sand and gravel particles are saturated at 5% water by weight.

The matrix moisture content [mmc] can be calculated from :-

mmc = [100*(w - 0.05*(100 - %Fines))] %Fines {12}

where 'w' is the moisture content of the whole sample, expressed as a percentage.

Table 5 illustrates the relationship between moisture content and calculated matrix moisture contents for the three different test samples. The relationships between undrained shear strength and CBR with increasing matrix moisture content are shown in Figures 9 and 10. Small increases in the moisture content, as determined on the whole sample, result in large increases in the matrix moisture content. It appears that the optimum moisture content occurs when the matrix moisture content of the sample is approximately equal to its plastic limit. CBR values below 2.0 and undrained shear strengths below 45kPa are obtained when the matrix moisture content approaches the liquid limit of the sample.

The standard laboratory test to determine the CBR values of boulder clays are not wholly reliable for a number of reasons but the gravel content and the limited cross-sectional area of the plunger in relation to the particle size of the material can lead to significant errors. Erroneously high results may arise due to concentrations of gravel in the test zone beneath the plunger. Although the CBR test procedure has been standardised the CBR value of a soil is an empirical parameter. The test can be considered as a 50mm diameter circular model footing. In practice the CBR has proved to be an acceptable parameter on which to base road design.

The 100mm diameter unconsolidated undrained triaxial test is a routine laboratory test procedure in which the soil fails in shear. In theory the failure plane should be inclined at 45 to the horizontal, assuming that u = 0, which should be the case for full saturation of the sample. Non-uniform distribution of the granular particles should not be so critical due to the greater area of the failure plane [in theory an ellipse of 44,400mm2] compared to the area of the CBR plunger which is 1,935mm2. A set of three identical samples prepared to the same density and moisture content is required for testing at different cell pressures; multistage triaxial testing is not advised.


The rapid reduction in both shear strength and CBR values of well graded cohesive fills as the moisture content increases has been demonstrated.

It is tentatively proposed that the CBR value of well graded cohesive fills can be estimated from the undrained shear strength parameter using the equation :-

CBR = (cu27)1.7

A 100mm diameter unconsolidated undrained triaxial test on compacted material is likely to provide a more reliable measure of the strength of a well graded material; the design CBR may be estimated from undrained shear strength.

It is suggested that the optimum moisture content of a well graded cohesive fill material coincides with the matrix moisture content being close to the plastic limit of the fill material.


Black, WPM, 'A method of estimating California Bearing Ratios of cohesive soils from plasticity data, Gotechnique, December 1962.

Black, WPM, 'The strength of clay subgrades: its measurement by a penetrometer', TRRL Report, LR901, 1979.

Black, WPM and Lister NW, 'The strength of clay fill subgrades: its prediction in relation to road performance', TRRL Laboratory Report 889, 1979.

Brown SF, Loach SC and O'Reilly MP, 'Repeated loading of fine grained soils', TRRL Contractor Report 72, 1987.

BS1377:1990, 'Methods of test for soils for civil engineering purposes', BSI, 1990.

Skempton AW, 'The bearing capacity of clays', Building Research Congress, London, 1951, (Building Research Station).

Ervin MC, 'Specification and control of earthworks', Engineered Fills, Eds Clarke B.G., Jones CJFP and Moffat AIB, Thomas Telford, 1993.

Jones RH and Greenwood JR, 'Relationship testing for acceptability assessment of cohesive soils', Engineered Fills, Eds Clarke BG, Jones CJFP and Moffat AIB, Thomas Telford, 1993.

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Please note comments made online may also be published in the print edition of New Civil Engineer. Links may be included in your comments but HTML is not permitted.