Moisture Content and the CBR Method of Design

Download Moisture Content and the CBR Method of Design

Preview text

Moisture Content and the CBR Method of Design
F. L. D. WOOLTORTON, Planning Engineer, Roads Branch, Ministry of Works, Nairobi, Kenya Colony ·
Introductory Remarks by the Chairman
The essence of science is generalization or, as Poincare says, to give the same name to different things. The essence of engineering is to use science, experience and constructive genius toward an economical solution of a specific problem. While there will never be a substitute for constructive genius, considerations of economy move us to expand science in such a way that it will cover more and more of the area for which experience is required at the present time. In road construction, experience still holds a large sway in the adequate understanding of soil and other construction materials and of the effect of climate on the performance and service life of material structures. One of the outstanding pioneers in expanding science into the area that is still essentially dominated by empiricism is Colonel Wooltorton, who has a large and intensive experience as a road builder in various climates. In his long and distinguished career, Colonel Wooltorton has always endeavored to understand scientifically the complex phenomena encountered by him and others and to coordinate them into a logical system as a first and necessary step toward a true science of rational road construction applicable to all climatic regions. In his present contribution, Colonel Wooltorton makes available to us his conclusions on the relationship between the moisture economy in soils and the proper use of the CBR method of design based on long experience and much good thinking.
e THE CBR test involves the compaction of a sample, in a manner specified by one of
several standardized techniques, at a predetermined or ascertained moisture content and to a predetermined or ascertained density; and the penetration of the sample, in its existing state or after soaking, by a plunger at a given rate to give a specified deformation under a measured load. The value measured, for use in design, is the load or the resistance of the plwiger to penetration. It may be considered as an indication of the strength of the material in the form of its resistance, under the conditions of test, to shear deformation or plastic flow. The measured load is expressed as a percentage of the resistance obtained from a like penetration of a standard material and is referred to as the California Bearing Ratio.
The required pavement thickness is determined from this CBR percentage by a method which is essentially empirical. The bearing ratio value is entered into an appropriate loading curve to give the required thickness.
Originally there was only one curve which became known as the curve for "light traffic" or the 7, 000 lb wheel load curve. With accumulated experience a second curve, for "medium heavy traffic," or for a 12,000-lb wheel load, was added (22, 23).
These took the form of the more modern variations shown in Figure 1. - -
The test was devised by the California Division of Highways in or about 1929 for the original purpose of comparing base and subbase course·materials and only later for the evaluation of subgrades, but it was not until 1940, when the U.S. Corps of Engineers adopted the test as the most convenient one for use in connection with its wartime airfield construction program, that the method was intensively developed.


An extensive program of investigations was put in hand with a view to extrapolating the design curves from the 12,000-lb

100 ----------~----~

wheel load to very much higher limits and to check the validity of the curves so ob-

so ~: ---1-4\.\-----------'------1
so ~; --'.__,___ __ _ _ _- - - ' - - - - - - - "

tained by experiment. As a result of the embracing laboratory
and field work undertaken, more is known

; 40 ~ , -+-+\.\-----'-, ---------~

I 4,000-lb
30 1 - - -+4-1.\---_.__

wheel load _ __

_,__ _ _ _-I

about this system of testing and method of design than of any other approach. From this research sprung much of the present-day knowledge of the design prob-

20~' ~---1t-\.\-\--'-~~~~~~~~----"
1 9,000-1 b wheel load
a: 15 i-----1.---4.......,.__.__ _ _ _- - + - - - - - - - - - !

lem and of the many and involved properties of soil under strain.
As far back as 1946, there appeared, in very different parts of the world (South

u 12,000-lb wheel load
10 ~------\--\-!.-->~-----+---------!

9 e

~ .~ • - _ - _- _- ___\ ,,..- ......~ .,...~ ....,.- ....,._- _- _- -----11 -- - -- - -- - -~ 1

Africa and Texas), two most remarkable

7 f------++--'t-+-'<-----+------~

research papers on the various relation-

& f------\t---T-"c-'~--+-------1

ships encountered in the bearing" value

method of design.

The two authors, Kleyn (South Africa)

and McDowell (Texas), had adopted the same line of thought and had come to substantially the same conclusions. Their works were, in a most remarkable way, complementary for jointly they covered a very wide range of soil materials thus enabling in conjunction with similar data

3 f - - - - - - + - - - - ' . . , __ _,,__-'4------""---- --I
' ' '' ' ' '' ' ' ' 2 4 & e 10 12 14 1& 1e 20 22 24 2s
Pavement Thickness, in.
Figure l. CBR-pavement thickness curves for various wheel loads.

from elsewhere, for example, Kenya, the

deduction of general rules of the behavior of compacted soil under strain.

Their data were presented by·plotting on a family of density-moisture content curves,

for various compaction efforts, the iso-lines for such properties as bearing value

(after soaking), expansion, air-voids content. The techniques were based on compac-

tion by the impact method and an absorption procedure involving capillary rise. Kleyn

(10) used the CBR value as an indication of stability whereas McDowell (14) used the

results of the Texas Modified Bearing Value Punch Test.


The combined results of their work with deductions therefrom supported by modern

research, carried out for the most part after 1950, are referred to later in this paper.

The design curves are observed correlationships between CBR values and pavement thicknesses.
Their implication is that if a field or an "in place" CBR value is determined for a base or a not too highly swelling (27) subgrade soil in or under an established, aged and provenly durable pavement (high quality base and wearing course) so designed as to be free from frost action (27), then, for average conditions, the thickness of the pavement above, necessary for achieving such permanency, will generally be indicated by the appropriate curve. And that is all.
The method assumes that a waterproofed, durable and permanent type of pavement is required and that the design and construction of each course will be planned accordingly; but note that the CBR concerned is the ultimate in-place CBR for the undisturbed lJase or subgrade soil, at some moisture content generally unknown at the time of design and existing in a structural state controlled not only by the way in which the layer was compacted but by the chemical and physical structure developed with time. If climatic conditions are such that the moisture content always fluctuates then the appropriate moisture content will be the maximum occurring under the ultimate range of fluctuation.


The success achieved in using this method of design depends entirely upon adapting a successful correlation between testing requirements and the ultimate conditions of the prototype.
Testing procedure must be standardized and strictly controlled and the data required and obtained must be planned and interpreted with discretion and considerable acumen.
It is evident that ii the conditions governing the penetration test are in any way varied; then the value of the CBR will change and a different thickness will be obtained from the design curves. It is therefore essential that the accepted test specifications should be rigidly adhered to until such time as locally gained knowledge can be collected to justify the validity of any variation contemplated.
Nevertheless the aim must always be to develop a technique which will more and more nearly simulate the structural and controlling moisture content conditions of the subgrade to be eventually reached in the field. When this is achieved it will be possible to introduce a constant factor of safety into the design calculations.

The design correlationship, as given by the CBR curves, is thickness versus inplace CBR after a number of years.
The problem is how to arrive at this ultimate CBR value keeping in mind that the W.uot ~l.:;v t~ .;;tii,bl~ wh.::n .l:i.n:1i: openeci i.o iraiiic.
The ultimate in-place CBR, or stability, depends on (a) the method of compaction, (b) the ultimate density, (c) the ultimate maximum or equilibrium moisture, and (d) any structural effects developed in the soil with time.


Under the original California test method of obtaining design data, compaction was achieved under a static load of 2,000 psi 1 but subgrade experience eventually indicated

that the stability of the material compacted in this manner sometimes differed consid-

erably from the stability of the soil after compaction to the same density in the field.

In the meantime and as a result of this procedure of compacting by the static method,

the belief grew that stability always increased with increase in density where this is

now known to have limitations as can be deduced directly or indirectly from reference


-In an endeavor to rectify this discrepancy, between the stabilities attained in the

laboratory and in the field, other forms of compacting in the laboratory were investi-

gated by the U.S. Corps of Engineers. This resulted in a modified form of the AASHO

compaction test or compaction by impact (1942); more recently this has been replaced,

in some laboratories, by compaction by kneading action. There is not a great deal of

difference in the stability or moduli of deformation of samples compacted by these two

latter methods to the same density at the same moisture content though the results of

kneading action compaction have been found to be closer to those attained in the field

(26). Both compaction by impact and by kneading action give far more realistic prop-

erties, as evidenced by triaxial and shear test data, than static compaction, when com-

pared with the stabilities obtained after compaction by sheepsfoot rollers and rubber-

tired rollers. Compaction involving kneading action is believed to produce a soil struc-

ture or particle shape and arrangement most nearly simulating those achieved during

construction and thus most likely to produce comparable shear strength.


The CBR' s obtained for samples prepared by static compaction may be very differ-

ent from those obtained by the other methods of compaction and they may be very mis-

leading if used as a basis for design.

There is some evidence that the unit pressure exerted during compaction in the

field also has some effect on stability.

1 This pressure was selected as that then believed to give densities most comparable with those found under old pavements.

Soaked Samples
An analysis of published stability data, for acceptable strains, bas shown:
1. For constant molding moisture content, an increase in the density to which a low to medium plastic soil (P.l. from 0 to about 12-15) is compacted is accompanied by an increase in bearing value for densities up to a value corresponding to an air void content which decreases with decrease in the moisture content of compaction after which the bearing value ceases to increase so rapidly and may decrease appreciably. This is illustrated by Figures 2 and 3.

5% Air Voids\

P.I.= 8 (Loam)
Passing No. 40 = 94% Passing No. 200 = 58%
- - - lso- CBR Lines - - - - lso- Expansion Lines - - - lso-Air Content Lines

10% Air Voids--i

-u ~
.c.. 120

Modified Compaction Curve

Zero Air Voids ans ion

' ',l ~0.1., Expansion
10., Air Voids

Figure 2.




Molding Moisture ContentL ,., dry weight

Iso-CBR lines for various conditions of compaction after four ~ soald.llg; a.f'ter (!2_) •


100 90 80 1-----+---- -- - --1----70
60 f - - -- - f - - -- - - - - - - - ;- -- -

CBR for c ompacl ion - - -" 1
al 1% dry of op!imum -----.--i......-----t---"---1 (modified l


a: 40

con I en!


20 15

Figure 3.

Proc!ar ap!imum mo i s!ure con!en!






Dry Oensi!y, pct

Variation in CBR with moisture content of compaction (data from Fig. 2).

For materials possessing a plastic index in excess of about 12-15, the lines of isobearing value change their pattern (see No, 2 below) a..11d the tendency cf the bearing
value to decr ease, for densities in excess of a critical value, seems to fade out as the P. I. inc reases (see Fig. 4).
2. The maximum bearing value for low to medium plastic soils 2 occurs, for any given compactive effort, at a moisture content of about 3 to 1 percent dry of optimum; while for medium to highly plastic soils the maximum bearing value moves to the wet side of optimum, at a P. I. of 12-15, where the deviation from optimum increases with increasing P.1:, and the effect is more pronounced for modified than for standard Proctor compaction (see Figs. 2 and 4).
3. For equal percentage saturation, bearing values increase with density for materials whose P. I. varies from zero to about 15.

Unsoaked Samples of Silty Clay and Sandy Clay Compacted by Kneading Action
1. For constant moisture content of molding, an increase in density caused an increase in stability depending on the moisture content and the range of densities involved (25) (see Fig. 6). -2. For a constant degree of saturation an increase in the density was accompanied by an increase in stability (25) (see Fig. 6).

Density may thus be a critical factor in design. A reference to Figure 3 shows, that for the particular medium-plastic soil concerned, an increase in density from 126 to 128 pcf for a sample compacted at modified optimum moisture content the CBR drops from 65 to 50.
When designing for a density, the value required must be sufficient to eliminate the possibility of appreciable consolidation, must be sufficiently high to achieve an economic value of the stability, should in conjunction with its moisture content give a condition of near minimum volume change, should be economically attainable.with the equipment available and must not be so high as to lead to the possibility of a decrease in stability
3 Some soils containing halloysite or fine material of volcanic origin appear to be exceptions in that while the iso-CBR lines are of the form associated with soils having a P. I. of less than 15, their P. I.' s are of the order of 30-35. The explanation may be in the moisture fraction representing the inter-planar water of halloysite or the vesicular moisture of porous volcanic material which are not indicative of plasticity.

accompanying any future increment of densification under traffic or by over-compaction during construction.
Though many high-swelling soil subgrades may lose density with time, if over-compacted, the more general run of materials utilized in good construction would appear either to remain at a near constant (dry) density or evidence a slight increase in (dry) density.
They could remain at constant density because of the constancy of the subgrade moisture content or, if that moisture content increases, by virtue of the grading and plasticity properties of the material or possibly because the superimposed pavement loading was sufficient to suppress and prevent volume change. The slight increase in dry density would most probably be a result of additional compaction under traffic. It could, perhaps, be caused by "thixotropic set" or particle orientation under stress.
If, while using such a soil as illustrated in Figure 2, construction control is such that there is a danger of over-compaction or if there is a possibility of some increase in density under traffic, it would appear advisable to compact at a moisture content slightly dry of optimum.
Whether it will ever be feasible always to determine with close accuracy the equilibrium or maximum moisture content within the equilibrium range of moisture contents likely to occur in the subgrade, or in any other layer, under all conditions of internal and external climate cannot yet be foretold. It is, however, known that certain research organizations are actively engaged on this problem.

Zero Air Voids
2% Expansion

P. l. = 62 (Black Turf) Passing No. 40 • 90% Passing No. 200• 82%
- - Compaction Curves - - lso-CBR Lines ----Isa-Expansion Lines -·-lso-Air Voids Lines

....>80 c
0 >-

Proctor Compaction Curve

.__..._Zero Air Voids
'--1 % E~pan ion 1
2% Expansion

Figure 4.

20 22 24 26 28 30 32 34 36 38 40 42 44 Molding Moisture Content, % dry weight
Iso-CBR lines for various conditions or compaction a:rter rour after (10).


48 50


Ory Density, pcf .





--1 ~Maximum 5,.. Air Voids


/ 8 1--~~~~r-~~~--+~~+-~-+----:~·~~~+-~~~---'~~~~--1~~~~---I



1 . . - 7 1--~~~--jr-~~~--+~~-1--~~·~~~~~+-~~~--1~~~~--1~~~~-1 CSR-Density Curve for Compaction at

Optimum Moisture 6 ..... Content

v /

Optimum Moisture Content


~ '--.......~i.--.i- ~ 5 r--~~~~+--~~---:1~9--..---+~----=-.......,--~~~-+~~~~-+-~~~~-+-~~~-----

5% Air Voids / / / /

CSR-Moisture Content Curve

4 3

...._____r--._ r-----+-v -r-- /+-~-+-~~~~-+----I /

../ L Possible minimum value in the field


after development of structure


- - r - - . _----

Figure 5.





Moisture Content, "• dry weight

CBR-moisture content and density curves f'or Fig. 4).



Proctor compaction (data f'ran

For certain conditions, as of a high water table, such a forecast is possible, though perhaps laborious, by means of the pF or suction pressure approach (3a) but this meth-
od has not yet been proved applicable when the water table is below a depth of about 10 rt.
For other conditions, as for a low water table, there are a number of approximate methods, some of which, in view of the heterogeneous nature of soils, may be found to give a sufficiently accurate answer for predictive purposes no more deviating from the truth than that for any other soil test data. They may, in fact, give far more reliable predictive data than at present surmised if used in the right way and under the right conditions.
Consider the plastic limit guide evolved out of experience. Assume that the soil is not submitted to slaking forces or is able to resist such forces. It appears to be generally accepted that, no matter to how high a relative density a highly plastic soil subgrade is compacted, it will, in time, swell until it ultimately reaches a (dry) density corresponding very closely to Proctor maximum density, thit is, for the purpose of obtaining a correlationship between the maximum moisture content and the plastic limit, one should ensure that the material was compacted to Proctor maximum density or bas reached a value approaching that density.
For compaction under a given compactive effort, the greater the moisture content of compaction the less the moisture which can be absorbed after compaction and this value reaches a minimum when the moisture content of compaction slightly exceeds the optimum moisture content for that effort.
For more average subgrade soils, the amount of moisture which can be absorbed is, in the absence of any slaking, controlled by the air-voids content at the time of compaction. In more general terms, this air-voids content is, for Proctor compaction,


about 4 Ya percent as an over-all maximum. This 4 7'2 percent air-voids content
could possibly permit of an over-all maximum absorption of about 3% percent of moisture. On this basis, the maximum moistUre content possible would be of the order of (0. M. C. plus 4) percent. This is illustrated in a statistical way by the average curves which appeared in "Engineering and Construction Control of Embankments" published in the Proceedings of the American Road Builders Association in 1941.
In the same set of curves the Plastic Limit and Optimum Moisture Content are connected, in an average way, by a relationship which may be expressed for the
range P. I. = 28 to 6 as:
0. M.C. =(1.17 P. L. - 8)

so that the maximum moisture content might be expected to be of the order of:

(1. 17 P. L. - 4) percent

A similar expression is deducible from Marwick' s points (18a) for British soils.
Expressed as a percentage of the Plastic Limit:

Max. Possible Moisture Content P. L.


(1.17 - P. L_) 100

Sandy - Clay

120 t - - - + - - - + - - - t - - - t - - l - - + - - P ( ~ 13 oopro• . ,_____;

I !


100 ~--'--":'-~-~--'---'--~~--~







20 22

Moisture Cont1nt, % dry wtiQht







20 22

Moisture Content, '% dry wei9ht

a: ~ ~ t----+---t-- ~"'--t-+---:::1-"o:::;:,-..,.L+--+-----i
30 1-----1--Y:--!--;..c;.._+--_...,.__...,,'--+--l-- ---1 2o l----+--.,1..£:::f:::_..j.~'i:z:;t:z~--+--+-....:..:...___j

o ....__

104 I06 108 110 11 2 11 4 11 6
Dr, Donsity, pct

118 120


Figure 6. Variation in CBR with compaction effort and molding moisture content
for unsoaked samples; after (25).

for which when
P. L. = 16 : maximum moisture content= 92 percent P. L.
P. L. = 30 : maximum moisture content= 104 percent P. L.
and these are values of the order found by Kersten (9a). If the material were originally compacted at Proctor 0. M. C. to a densicy -of less
than Proctor maximum then the amount of moisture the material could absorb would be greater than considered above.
It would thus appear that in a statistical way the maximum moisture content of the subgrade might be expected to reach ultimately a moisture content which could be represented by the approximate range 92 to 104 percent of the plastic limit with values in excess of 104 percent for certain types of materials and when compaction was less than 100 percent Proctor maximum.
For P. L. of 16, the range would be 14 to 16 percent and for a P. L. of 30 the range would be 27 to 30 percent.
Though for constant density, the CBR, can vary very considerably over a range of 2 to 3 percent of moisture content one wonders whether in soil work a much greater accuracy in testing and during construction is indeed possible.
The real difficulty in interpreting such data lies in being able to estimate the time interval which will elapse before the maximum moisture content is reached should construction be carried out at a lower value. If it is four days or four years, the value of the maximum moisture content is of great importance but should it take forty years to accrue then it becomes of less importance in these returning days of lower standards of construction followed by some of the backward areas being developed in the face of inadequate funds.


Though any firm correlationship between the maximum moisture content and the plastic limit would be most informative, the relationship does not appear to be on such a rational basis as the field moisture equivalent or even the saturation percentage relationship both of which appear to mean something understandable.
An interesting sidelight on the F. M. E. correlationship arose during the study of data for a believed halloysite subgrade soil. It was noted that 0. 75 x F. M. E. represented a moisture content 9 percent lower than the modified optimum moisture content or one occurring well off the chart of equi-C BR lines for the moisture content range lying between the compaction curves for modified and Proctor compaction. Such moisture content did not appear to be likely equal to the maximum site moisture content under a pavement.
It was then noted that two of the data available applied to the subgrade under an old gravel road. In this area the rainfall was 80 in. and rain occurred throughout most of the year. The data were actually taken for a rainy period. The measured site moisture contents for those tests were 27 percent and 29 percent only.
The plastic limit was 35 and the Proctor optimum 45 percent. The field densities were 95 percent and 111 percent of Proctor maximum and the F. M. E. was 43.


The .suction-pres:!iu.i·e metilou of uelel'lulning the equilibrium moisture content or the maximum moisture content within the equilibrium range of moisture contents likely to occu1~ in ariy layer ui a road structure under an impermeable pavement of infinite dimensions has been developed in England. The method has been described by Croney (3a) and elsewhere. The principles behind this method and the implications of the conclusions reached in its study are not, however, clear at all. The following approach is believed to present an understanding of the principles involved and the conclusions to be reached from such an approach.
Let the total free energy of moisture at a point within a soil profile at equilibrium moisture content and at constant temperature be represented by Afs


Afs = Af
Where and

Af Afos AfFs

= free energy due to surface tension and radius of curvature of the
air-water interface,
= free energy due to hydrostatic pressure of moisture adjacent to
soil particle surface caused by adsorption forces as well as by any pressure transmitted from external sources,
= free energy due to the osmotic pressure developed by any dis-
solved material,
= free energy due to water by virtue of its position in the particle
adsorptive force field as well as in the earth's gravitational field.

If "X" and "Y" are two clay particles each surrounded by an adsorbed moisture sheath, as shown dotted, and "a" and "b" are points in this system in which "a" is below the meniscus of the
capillary ring moisture but without the adsorptive force fields,
and "b" is within the adsorbed water surrounding particle "X",
then, assuming equilibrium conditions, the moisture is pure
water and there is no osmotic pressure, energy conditions give:

CJ"'•'\ ;.. - ..-"' b




..... ..::::·~

' _ _,~'




1. For point "a" lying outside the particle force field the energy due to the hydro-

static pressure in the moisture by virtue of its position in the particle adsorption field

and that due to the position of the particle in the adsorption field disappear.

Under such conditions:

Afs =Af

= -(gph + Pw)



Pw = effective superload pressure due to weight of the moist soil



h = height of the point above the free water surface

or resultant suction pressure =gravity potential - pressure potential


s =u - a.P


where and

s = suction pressure and is negative, u = gravity potential which may be negative or positive depend-
ing on position of the point relative to the ground water level, p = superimposed vertical loading,
a. = a fraction varying between 0 and 1 depending upon the effectiveness of intergranular contact.

2. For point "b" lying inside the particle force field the same general energy relationship holds, that is:


but ~fFs + ~fps now refers to the energies associated with both the gravity and particle force fields.
In British publications the gravity potential appears to be referred to as "the porewater pressure as determined by the position of the point relative to the water table."
The interpretation appears to imply that, if the water table is 100 ft below a point in the profile at moisture equilibrium then the pore-water pressure, or pressure deficiency, across the capillary meniscus (if there is one) would be as high as 43 psi and the radius of the meniscus would oe as low as O. 00002 in.
It seems doubtful, however, if capillary water can exist as water in such fine soil pores and it would appear that for such a low water table, frequently exceeded in the tropics, the equilibrium moisture content may well be controlled, as far as the lifetime of a road is concerned, more by adsorbed moisture than by capillary moisture as such. Further, it seems very likely that temperature gradients may, on occasion, influence such equilibrium values to a marked degree to give maximum and minimum equilibrium values of a perhaps clifferent order.
If the pavement is permeable and of limited width there will be further modifications for considerations.
The suction pressure approach to moisture distribution appears to lead to the possibility of confusion arising between the energy with which moisture may be held at a point with the energy available at the point to hold moisture. The difference is, of course, to be represented by the time interval necessary for the distribution of vapor flow to bring about that distribution suggested by energy conditions.
Though it may be convenient to consider moisture as attracted by suction forces, there is no obvious reason to believe that such forces are necessarily responsible for attracting moisture to a point or for controlling the amount of moisture at a point under a pavement when the water table is low.
The energy available to hold water at a point in a profile is made up of the following two components:

Energy associated with the particle force field or the adsorptive energy


The capillary energy available at the point to hold moisture when it arrives as may be modified by any swelling pressure or any super-load.

The first component has a maximum value as long as the relative humidity is not
appreciably less than 100 percent or for suction pressures numerically less than that
corresponding to a pF of about 4. O; while the second depends on the shape of the parti-
cles, that is, plate or rod-like or spherical, and the density to which the material is

Preparing to load PDF file. please wait...

0 of 0
Moisture Content and the CBR Method of Design