ce7 - l. wesley

GEOTECHNICAL PROPERTIES OF RESIDUAL SOILS PROPIEDADES GEOTECNICAS DE LOS SUELOS RESIDUALES Laurence D. Wesley University of Auckland (retired) New Zealand

Abstract An account is given of the geotechnical properties of residual soils. It is pointed out that their formation process is very different from that of residual soils, and some aspects of sedimentary soil behaviour have no relevance to residual soils. In particular, the behaviour of residual soils cannot be related to stress history, as is the case with sedimentary soils. It is shown that the use of the conventional e-log(p) graph for portraying compressibility can easily lead to misunderstandings of compression behaviour and that a linear graph gives a much more reliable picture. The value of Atterberg limits and the liquidity index for indicating the properties of residual soils is described. An account is given of shear strength of residual soils, both in term of undrained strength and effective strength parameters

Resumen Se analizan las propiedades geotécnicas de los suelos residuales. Se señala que su proceso de formación es muy diferente de la de suelos sedimentarios y que algunos aspectos del comportamiento de los suelos sedimentarios no tienen relevancia en los suelos residuales. En particular, el comportamiento de los suelos residuales no puede atribuirse a la historia de sobrecarga, como es el caso de los suelos sedimentarios. Se muestra que el uso de los gráficos e-log (p) convencionales para representar la compresibilidad pueden conducir fácilmente a interpretaciones incorrectas del comportamiento a compresión y que un gráfico lineal proporciona una descripción mucho más confiable. Se describe el valor de los límites de Atterberg y el índice de liquidez para indicar las propiedades de los suelos residuales. Se analiza la resistencia al corte de los suelos residuales, tanto en términos de resistencia no drenada, como de parámetros de esfuerzos efectivos.

1 WEATHERING PROCESSES AND SOIL FORMATION

The weathering processes that convert rock into soil are generally a combination of physical and chemical processes.

1.1 Physical weathering Physical weathering may be one of two types:

(a) Erosion – by the action of glaciers, water of wind.

(b) Disintegration – caused primarily by wetting and drying, or by freezing and thawing in cracks in the rock.

(c) These processes produce a range of particles of varying sizes, which are still

composed of the same material as the parent rock. The grinding action of glaciers tends to produce a very fine grained material, generally known as rock flour.

1.2 Chemical weathering Chemical weathering is much more complex,

and results in the formation of groups of particles of varying sizes and properties, known as clay minerals. These particles are generally crystalline and are of colloidal size, ie less than 0.002mm. The most common clay minerals are kaolinite, illite and montmorillonite, but less well known clay minerals of special importance in volcanic areas are halloysite, allophone, and immogolite.

XIII Congreso Colombiano de Geotecnia - VII Seminario Colombiano de Geotecnia - ISBN 978-958-98770-2-9

A.7-1

1.3 Influence of topography It is worth noting that the weathering process is

different in hilly, well drained areas, to flat poorly drained areas. Weathering in well drained areas tends to produce soils with good engineering properties while weathering in poorly drained areas tends to produce soils with undesirable engineering properties. This is illustrated in Figure 1.

Water table

Well drained hilly and mountainous areas: Downward seepage results in deep weathering and soils tend to havegood engineering properties

Downward seepage

Poorly drained, flat, low lying areas:Absence of vertical drainage results in shallow weathering andsoils of poor engineering properties

No regular seepage Figure 1 Influence of topography on weathering.

In flat poorly drained areas, there is no regular seepage pattern. During wet weather, water will enter the soil from the surface but little horizontal flow will occur. During dry weather, water will be lost through evaporation and there will be upward seepage flow

1.4 Weathering profiles Weathering profiles from residual soils are

shown in Figure 2. The most well known weathering profile is that of Little (1969). This is shown in Figure 2(a). However, this profile is for weathered igneous rocks, and different profiles are obtained from other types of rocks, as shown in Figure 2 (b), (c), and (d). Little stated that the classification system he proposed was intended for the weathering of igneous rocks in the humid tropics, and not for weathering profiles generally. As shown in Fig. 2.1(a), the profile consists of a series of thick zones of not greatly differing thicknesses. With different parent rock, however, this may not be the case at all.

With some soils, especially igneous rocks at the basic end of the range such as basalt or andesite, the boundary between rock and soil is likely to be quite abrupt, without the thick transition zone that is typical of acidic rocks such as granite. Volcanic ash layers produce a pattern that is different again. This is because the parent material is deposited in layers of varying age, so that some deep layers may be more highly weathered than those closer to the surface. Weathering of soft sedimentary rocks tends to produce a layered soil reflecting the characteristics of the parent rock.

Soil:clay or silt

Soil: clay or silt

Completelyweathered

Highly weathered

Moderatelyweathered

Slightly weathered

Fresh rock Fresh rock Fresh rock

Fresh rock: inter-bedded sandstone and clay-stone

Weathered rock

Soil: Inter-bedded clay, silty clay, silt,and dense silty sand

(a) Gradual weathering profile - typical of weathered granite

(b) Sharp transition from rock to soil - typical of weathered basalt

(d) Stratified nature of parent rock reflected in soil profile - typical of weathering of soft sedimentary rock, especially sandstone.

(c) Uniform layers, degree of weathering not necessarily related to depth - typical of volcanic ash

Silty clay layers,almost homogeneousbut distinguished byslightly differentcolouring

1

2

3

4

5

6

Figure 2 Variety of weathering profiles, depending on the nature of the parent rock.


A.7-2

RESIDUAL SOIL:-produced by physical and chemical weathering of underlying rock.

Erosion by rainfall and run-off.

Transport by stream and river.

Delta deposits

SEDIMENTARY SOIL- later tectonic movement may raise this above sea level

Re-depositoin in layers in lakes or the ocean.

Sea or lake levelRock

Soil

Figure 2 Formation of the two main soil groups: residual and sedimentary soils.

1.5 Formation of residual and sedimentary soils.

The action of rainfall is to erode the surface of the weathered rock and transport the weathered material by stream and river until it reaches “still” water, in the form of an ocean or a large lake. Here it will settle out and form a deposit on the sea bed or lake bed. The process is illustrated in Figure 3. The soil formed directly from the weathering of the underlying rock is called a residual soil, and the soil built up in layers by deposition in the sea or a lake is called a sedimentary or transported soil.

Sedimentary soils may undergo a great deal of compression or “consolidation” as additional layers are deposited above them; they may also experience uplift as a result of tectonic movement and end up again as dry land many metres above sea level. A second cycle of erosion may then occur and much of the upper layers may be removed. As we are well aware, soils which have not had any removal of overburden material are called “normally consolidated”, while those that have experienced unloading are known as “overconsolidated” soils.

Soil mechanics grew up in northern Europe and North America, and most of its concepts regarding soil behaviour were developed from the study of sedimentary soils, especially remoulded soils. The influence of stress history received a lot of attention. Most textbooks and university courses on soil mechanics place

considerable emphasis on stress history – and soils are divided into normally consolidated and over-consolidated on this basis, and behavioural frameworks are developed around this stress history concept.

This might be appropriate if all soils were sedimentary soils. This is clearly not the case. Large areas of the earth consist of residual soils, and the application of concepts coming from sedimentary soils may or may not be relevant to these soils. It is interesting to note that very few textbooks, and probably very few university courses on soil mechanics, even mention residual soils, let alone give an account of their properties.

The irrelevance of concepts developed from the behaviour of sedimentary soils for explaining the behaviour of residual soils are particularly those related to their mode of formation. The use of the log scale for portraying compressibility is a specific example; this will be examined in a later section on oedometer test results.

Figure 4 is a further attempt to illustrate the important differences in the formation factors influencing the behaviour of sedimentary and residual soils. The formation process with residual soils generally weakens the material, as we would expect in a change from rock to soil. The material becomes more porous as the process advances. With sedimentary soils, the reverse is true. After deposition, the soil is consolidated, and also “ages” or hardens from chemical processes.


A.7-3

Figure 4 Formation influences on sedimentary and residual soils.

2 OBSERVATIONS ON CLASSIFICATION OR INDEX TESTS

2.1 Appropriate Pre-Treatment before Laboratory Testing.

It has long been known that some residual soils undergo irreversible chemical changes when they are air or oven dried. This is particularly true of soils containing either allophane or halloysite. Allophane clays may be quite plastic at their natural water content, but after oven drying may become completely non-plastic. Halloysite clays will undergo less drastic change on air or oven drying. It is most important therefore that these soils be tested without pre-drying them more than is necessary to carry out the test. It is good practice with residual soils to always avoid drying them any more than is necessary to carry out each particular test.

2.2 Particle size Particle size measurements can be made on

most residual soils in the usual manner. Two exceptions should be mentioned: 1. Soils which contain substantial amounts of

highly weathered coarse particles. These particles may break down easily during preparation for testing, and the particle sizes measured may thus reflect the treatment of the soil prior to testing as much as the actual properties of the soil. With such materials, particle measurements are less important than with hard grained material, as the properties

will be governed more by the strength of the particles than their size distribution.

2. Clays containing a high concentration of allophone area likely to flocculate if sedimentation methods are used for particle size measurements. Normal dispersing agents are often not effective with such material. This is probably not a great disadvantage as knowledge of their particle size is not of much significance.

2.3 Atterberg limits. These can be carried out in the normal manner,

but it is useful to realise that some residual soils are not very plastic in their natural state. In appearance they may be like sandstones. However, manipulation and re-moulding may destroy their structure and break up particles so that the material becomes moderate to highly plastic.

The usefulness of Atterberg limits as a guide to engineering behaviour lies in the position they occupy on the Plasticity Chart. There is no direct correlation between behaviour and either the Liquid Limit or the Plasticity Index – it is the position on the Plasticity Chart that is important.

Figure 5 The plasticity chart and residual soils. Those plotting below the A-line generally have

good engineering properties and those that plot above the A-line are likely to have poor engineering properties. Fig. 5 shows the position on the Plasticity Chart of three distinctive residual soils - the "Black Cotton" soils, the tropical red (halloysite) clays, and the volcanic ash (allophone) clays. The black cotton soils plot well above the A-line and have poor engineering properties, while the red clays and volcanic ash soils plot below the A-line and have good engineering properties.

Some organisations create a number of sub groups on the Plasticity Chart by drawing vertical lines on the basis of Liquid Limit values. This

Pressure

Void

ratio

(=

) A

BC

At deposition, the soil is very soft with a high voids content

Continuing deposition compresses the soil

Uplift and erosion may reduce pressureon the soil, allowing it to swell

Voids

Voids

Voids

Voids

Condition at deposition (Point A)

Eventualcondition(Point B & C)

Point B Point C

(a) Residual soil

Solids

SolidsSolids

Solids

Parent rock (little or no voids)

Weathering changes the solid matter and increases the void space. Residual soil

(substantial void space)

Vol.

void

sVo

l. so

lids

Volcanic ash soils

(allophane)

A-Line

0 50 100 150 200 250

150

100

50

Liquid Limit

Pla

stic

ity In

dex

“Black cotton” soils (montmorillonite)

Tropical red clays(halloysite)

Poor engineering properties (clay)

Silty clay

Good properties (silt)


A.7-4

does not seem to be a sensible approach for residual soils; rather than a subdivision based on the liquid limit, a subdivision along the lines shown in Fig. 5 would be most relevant to residual soils. The lines drawn parallel to the A-line divide soils into three types labelled clay, silty clay, and silt. However, the Plasticity Chart should be used as a means for evaluating residual soils rather than a means of classifying them. The classification of allophone clays as silts is very questionable, as many of these clays show only very limited evidence (or none at all) of the :quick” or dilatant behaviour normally associated with silts. However, their position below the A-line clearly is a good indicator of their generally good geotechnical properties.

It follows from what has been said above that correlations between Atterberg Limits and other soil properties on the basis of either Liquid Limit or Plastic Limit alone may well not be valid for residual soils, and should be treated with caution. Other uses for the Atterberg limits are that together with the natural water content of the soil, they tell us something important about the “density state” or porosity of the soil, as discussed in the next section.

2.4 Liquidity index, “density state” and porosity.

There are two methods used to express the “denseness” of soils, both of which make use of reference density states. They are each a measure of the position the soil occupies in relation to these reference “density states”, namely Atterberg Limits in the case of clays, and maximum and minimum densities in the case of sands.

These two concepts are illustrated in Figure 6. The manner in which the “denseness” is defined is different between the two methods. A Liquidity Index of 0 or 1 indicates a dense or “non-dense”

clay respectively, while a relative density of 0 or 1 indicates a loose or dense sand respectively. These indices have proved very valuable as indicators of likely behaviour of clay and sand respectively, and the Liquidity Index is a very useful parameter for evaluating residual soils.

Figure 6 Reference “compactness” states for clay and sand.

It is probably better to think of the Liquidity

Index as a “compactness” index rather than a liquidity index, because it is only an indicator lf liquidity for remoulded soils.

It should be noted in passing that the liquidity index is particularly useful if earth works are contemplated, as it is an indicator of likely handling difficulties, and the amount of drying that will be necessary. The Plastic Limit is normally reasonably close to the optimum water content from a standard compaction test.

Table 1 Coefficient of permeability values from several residual soils.

Soil Type Coefficient of Permeability (m/sec) Parent rock

Description Young (saprolitic) Mature (true soil) Remoulded

Granite 4x10-3 to 5x10 4x10-9 -6 to 5x10 - -9 Gneiss 5x10-6 to 1x10 5x10-7 -6 to 1x10 - -6 Basalt 3x10-6 to 1x10 - -9 -

Sandstone Grey clay Andesitic lahar/

volcanic ash Tropical red clay

(halloysitic) 1 x 10 0.3 –3 x 10-9

Volcanic ash

-10

Volcanic ash clay (allophane)

10-6 to 10 5 x 10-7 -7 to 10 to 5x10-8

-10

LL

PL

wnen

emax

emin

LL -

PL

LL - PL

e -

em

axm

ine - emax min

e -

em

axn

e - emax n

w -

PLn

w - PLn

0

1

1

0

= Liquidity Index =

Density Index

= Relative Density =

Density Index

CLAY SAND

Non-compact (loose)state

Compact (dense) state Vo

id ra

tio o

r wat

er c

onte

nt a

s

mea

sure

s of

com

pact

ness

Wat

er c

onte

nt

Void

ratio


A.7-5

3 PERMEABILITY

Generalisations are always risky in soil mechanics, but it is certainly true that residual soils tend to have substantially higher permeability than sedimentary soils. This is due to micro-structural features, such as the aggregation of clay particles into clusters, and the ability of bonds between particles to create a very open structure.

Remoulding and compacting residual soils tends to destroy this structure and generally results in a significant decrease in permeability. It should be noted also that correlations between permeability and particle size developed from the study of sedimentary soils often do not apply to residual soils.

4 COMPRESSIBILITY AND CONSOLIDATION BEHAVIOUR

4.1 Magnitude 4.1.1 Consolidation (oedometer) tests

The results from a number of conventional one dimensional oedometer tests are presented in this section to illustrate compressibility behaviour of residual soils. Figure 7 shows results from three samples of tropical red clay. They are plotted using both log and linear scales for pressure. The reason for doing this is that the use of the log scale easily leads to false interpretation of what exactly the curves reveal about the soil behaviour. The curves on the log graph appear to have the typical shape of an over-consolidated soil, and it would be possible to apply the normal procedure to determine pre-consolidation pressures.

However, when the same data is plotted using a linear scale for pressure, they present quite a different picture. It is seen that there is very little evidence of “over-consolidation”. The graphs are close to linear, although one sample shows a faint suggestion of a pre-consolidation pressure at a stress of about 170 kPa.

The other two curves are slightly concave upwards. In Figure 8 the graphs are re-plotted between zero and 350 kPa, the stress range likely to be of most interest to foundation designers. This emphasises the behaviour is indeed close to linear over this stress range.

Figure 7 Oedometer tests on three tropical red clay samples

Figure 8 Oedeometer tests on tropical red clay from 0 to 350kPa.

10 100 1000

1.6

1.4

1.2

Pressure (kPa)

Pressure (kPa)

(a) log scale

(b) linear scale

0 250 500 750 1000 1250

2

4

6

8

10

Void

rati o

Co m

pres

sion

(%

)

100 200 3000

2

4

5

Com

pres

sion

%

Pressure (kPa)

Linear scale up to 350kPa


A.7-6

Figure 9 Oedometer tests on Piedmont clay, USA.

Figure 10 Oedometer tests on clay derived from soft sandstone & mudstone (Auckland).

Figures 9 to 12 show the results of oedometer tests on a number of residual soils, plotted respectively using log and linear scales for pressure. Figure 9 shows tests on a clay derived from a weathered sandstone known as the Piedmont formation. The curves in Figure 9 (a) have been used to determine pre-consolidation pressures and over-consolidation ratios, but when re-plotted on a linear scale there is clearly no evidence of a pre-consolidation pressure.

Figure 10 shows results from a clay also derived from a weathered sandstone and mudstone formation known locally in Auckland, New Zealand as Waitemata sandstone. In this case the curves have the traditional form on the log plot, but when re-plotted using an arithmetic scale they are close to linear.

Figure 11 Oedometer tests on sensitive silt derived from sandstone and mudstone (Auckland).

The curves in Figure 10 are typical of

oedometer tests on this clay. The compression behaviour on the arithmetic graph is close to linear over a wide stress range. However, there are more silty layers within the clay that are of

B9-3M OCR = 4.0B9-4M OCR = 3.6B7-5M OCR = 3.4B8-7M OCR = 3.4B8-8M OCR = 3.3B7-9M OCR = 1.1

2.0

1.6

1.2

0.8

0.410 100 1000 10000

Void

ratio

0 500 1000 1500 2000

10

20

30

Com

pres

sion

(%)

Pressure (kPa)

(a) log scale

(b) linear scale

0 500 1000 1500

Pressure (kPa)

Pressure (kPa)

(b) linear plot

(a) log plot

Verti

cal s

train

(%)

Verti

cal s

train

(%)

10 100 1000

0 250 500 750 1000

Pressure (kPa)

Com

pres

sion

(%)

Void

ratio

(a) log plot

(b) linear plot


A.7-7

moderate sensitivity and show different compression behaviour. Figure 11 is an example of this. The samples appear to indicate “pre-consolidation” pressures in the log plot, and the linear plot confirms this. The curves show increased compressibility at about 200kPa. It should be clearly understood that this so called “pre-consolidation” pressure is not really a pre-consolidation pressure at all, as it has no connection with the stresses the soil has previously been subject to, ie it is not the result of stress history. It is the result of the structure created in the soil by the weathering process that formed it from its parent rock.

Figure 12 Oedometer tests on volcanic ash (allophone) clay.

The results given in Figures 7 to 12 demonstrate very clearly that the use of the traditional e-log p graph can easily present a misleading picture of soil compressibility, and should be treated with caution. In fact it should simply be rejected for residual soils. The only way to get a true picture of their compression characteristics is to plot the data on both a log and a linear graph. This is not surprising as there is no reason at all to assume that the log graph is appropriate for residual soils. The use of the log graph for presenting oedometer test results came about as a result of the study of sedimentary soils, especially artificial “slurry” samples prepared in the laboratory. When such soils are consolidated from a very soft or “slurry” state, the graph of compression (or void ratio) versus pressure is approximately linear when pressure is plotted using a log scale.

A number of authors have pointed out over the years the need to plot compressibility data using linear pressure scales, especially Professor Janbu of Norway (Janbu & Senneset, 1979). However, the profession as a whole is still bound by tradition and continues to use the e-log p graph almost exclusively (and blindly).

The three stress deformation curves in Figure 12 represent the three main types of compression behaviour found in residual soils (and in fact also in sedimentary soils). Figure 13 summarises this behaviour, and is a much better representation of the compression behaviour of all soils than the traditional e-log(p) graphs normally found in text books and courses on soil mechanics.

Figure 13 A general representation of the compression behaviour of soil.

3

2

1

10 100 1000 5000

0 100 200 300 400 500

2

6

10

Void

ratio

Pressure (kPa)

Com

pres

sion

(%)

A

A

B

B

C

C

(a) log scale

(b) linear scale

4

8

12

Pressure

Stra

in Linear

Strain softening

Strain hardening

Vertical yield pressure

Yield fromstructural breakdown

Strain hardening is typical of dense soils with low liquidity index

Strain softening is typical of non-dense soils with high liquidity index


A.7-8

4.1.2 Field behaviour The existence of a yield stress in some residual

soils has been confirmed by field records. An example is given in Figure 14, which shows settlement records of the foundation material beneath the Guri Dam (Prusza et al, 1983). The graphs all show a yield stress at about 20m of fill, which would be a vertical stress of about 350 kPa. The soil involved was weathered gneiss.

Figure 14. Plate loading tests showing evidence of a yield stress (after Prusza et al, 1983)

4.1.3 Values of compressibility parameters

Figure 15 Coefficient of compressibility (mv) for several residual soils.

To conclude this section, Table 1 shows some

representative values of Young’s Modulus (E) and coefficient of compressibility (mv

) for a range of residual soils. These values are also represented graphically in Figure 15

Table 1 Compressibility parameters for a range of residual soils.

Soil Type Young’s Modulus E (MPa)

from plate loading tests

Coefficient of Compressibility mv (10-4 kPa-1

Reference source )

Parent Rock Soil Description Oedometer test

0-300 kPa Estimated from plate

loading test

Andesitic lahar/ash

Tropical red clay 1 to 2 Author (Wesley)

Volcanic ash

Volcanic ash clay 0.7 to 2 Author (Wesley)

Sandstone Grey clay 1 to 2 Author (Wesley) Basalt Basalt saprolite 0.7 to 3.3 De Mello (1972)

Granite Granite saprolite 0.5 to 5 Lumb (1962) Granite Granite saprolite 0.4 to 1.0 Hui (1972) Gneiss Gneiss saprolite 47 to 84 0.1 to 0.17 Sandroni (1981)

Gneiss Gneiss saprolite 10 to 13 0.7 to 0.9 Werneck et al (1979).

Gneiss Gneiss saprolite 20 to60 0.13 to 0.42 Garga & Costa (1977)

Gneiss Gneiss saprolite 9 to 10 0.8 to 0.9 Napoles Neto (1954)

Gneiss Gneiss saprolite 28 0.30 Vargas (1979)

Gneiss Gneiss saprolite 0.30 to 0.45 Azevedo (1972)

Gneiss Gneiss saprolite 0.9 Campos (1980)

0 10 20 30 40Height of fill placed (m)

1

2

3

4

Verti

cal s

train

(%

)

KEYU Upper layerM Middle layerL Lower layer

2U4U

3U

2L3M

3L

0 1 2 3 4 Coefficient of compressibility, m (10 kPa ) v

-4 -1

Red (halloysite) clay

Volcanic ash (allophane) clay

Weathered Sandstone

Basalt saprolite

Granite saprolite

Gneiss saprolite

Piedmont soil


A.7-9

Soil Type Young’s Modulus E (MPa)

from plate loading tests

Coefficient of Compressibility mv (10-4 kPa-1

Reference source )

Parent Rock Soil Description Oedometer test

0-300 kPa Estimated from plate

loading test

Gneiss Gneiss saprolite 0.6 to 2.5 Werneck et al (1979)

4.2 Time rate and estimation of the coefficient of consolidation

Residual soils are frequently of significantly higher permeability than sedimentary soils, which means that their coefficient of consolidation (cv) values are also higher. This influences their behaviour both in the field and especially in standard oedometer tests. In particular, it makes the determination of cv values of residual soils from standard odometer tests somewhat problematical. The most common method for determining cv

This figure illustrates that when the coefficient of consolidation is below about 0.01cm

is probably that of Taylor, which uses a square root of time plot. The theoretical shape of root time plots for samples 20mm thick, based on Terzaghi’s consolidation theory, is shown in Figure 16. The points on the graphs indicate the time intervals at which readings are normally taken, at least when the recording is done manually.

2/sec, the number of readings obtained will scarcely be sufficient to define the straight section of the graph which is fundamental for determining the value of t90 and thus the value of the coefficient of consolidation, cv

.

Figure 16 Theoretical square root of time plots for 20mm thick oedometer samples.

Figure 17 shows root time graphs from three different residual soils. None of these graphs shows the straight-line section predicted by consolidation theory illustrated in Figure 16.

Figure 17 Root time graphs from oedometer tests on three residual soil types.

The reason for this is simply that the pore

pressures dissipate so rapidly that the shape of the curves is no longer governed by the mechanics of pore pressure dissipation. It is probable that pore pressures are fully dissipated even before the first reading is made. With manual recording, it is not possible to take readings at closer intervals than those in Figure 17, although it is apparent from the shape of the curves that even continuous recording of deflection would not significantly alter their appearance.

There is thus an upper limit to the value of the coefficient of consolidation that can be measured in conventional consolidation tests, this limit being approximately 0.1m2/day (= 36.5m2/year = 0.012cm2/sec.). Readings taken in the first minute will only lie on a straight line if the cv

Volcanic ash soil Waitemata clay (wTropical red clay

eathered sandstone)

0 2 4 6√ √ . time min

2

4

6

8

10

Com

pres

sion

(%)

value is

0 1 2 3 4 5

20

40

60

80

100

Aver

age

degr

ee o

f con

solid

atio

n (%

)

Square root of time (min ) 0.5

C = 0.001 cm

/sec

v

2

C = 0.01 cm

/sec

v

2

C = 0.1 cm

/secv

2


A.7-10

less than this. If reliable values of cv are required for soils with higher values, it is necessary to use a different method of measurement, such as a pore pressure dissipation test in a triaxial cell. Values of cv

for the soil types in Figure 17 are shown in Table 2

Table 2 Values of coefficient of consolidation (cv) for the soil types in Figure 4.15

Soil type Coefficient of consolidation (m2

Waitemata silts /day)

and clays 0.01 to 10

Indonesian red (halloysite) clays: 0.07 to 0.7

Volcanic ash (allophone) soils 0.01 to 200

Because graphs of the shape shown in Figure 13

are common with residual soils, and not uncommon with sedimentary soils, their physical significance should be clearly understood. The important points are the following:

(a) If the oedometer test does not produce a linear deformation versus root time plot, it means that the deformation rate is not governed by pore pressure dissipation, and the coefficient of consolidation is greater than the limiting value mentioned above.

(b) In terms of conventional concepts of primary and secondary consolidation, nearly all the compression is secondary compression, as it is occurring under constant effective stress. The curves therefore reflect the creep behaviour of the clay when tested in this particular way.

(c) The shape of the curves is not primarily a property of the soil. It is a function of the drainage path length in the test. If the soils in Figure 13 were tested in an oedometer with a much greater sample thickness (say 1m instead of 20mm), then the root time plots would almost certainly show clear linear sections in accordance with consolidation theory, followed by a section representing secondary consolidation. The proportions of settlement made up of primary and secondary settlement would then be very different, and governed by the thickness of the sample tested.

(d) In practice, soil layers are generally thick, and the drainage path length may be several orders of magnitudes greater than in the oedometer. The absence of primary consolidation in the laboratory test is

therefore not an indication that there will not be primary consolidation in the field. Similarly, the presence of a large proportion of secondary consolidation in the oedometer test does not necessarily indicate that secondary consolidation will be important in the field.

Mention should be made at this point of what appears (in the author’s experience) to be a common mistaken interpretation of the root time plot. This is illustrated in Figure 18, which shows the results of tests on two samples. Sample A, with a low value of the coefficient or consolidation, conforms to expected behavour, and the construction for determining √t90

Sample B, on the other hand, has a very much higher coefficient of consolidation, and does not conform to expected behaviour; there is no linear section of the graph. Despite this, an arbitrary straight line has been drawn to “best fit” some of the points, and the standard construction applied to determine √t

has been correctly applied.

90. The value so obtained is about 1.5 min0.5

, giving a value of t90 of 2.25min. The true value of √t90 is probably no greater than 0.25min0.25 corresponding to a t90

of only 0.06min. The coefficient of consolidation calculated on this basis would be in error by a factor of nearly 40.

Figure 18 Correct and incorrect interpretations of root time graphs from oedometer tests.

0 2 4 6√ √ . time min

2

4

6

8

10

Com

pres

sion

√t 90√t ( ??)90

Sample A

Sample B

Correct

Incorrect


A.7-11

This practice occurs because of an expectation

that all soils should conform to “normal” behaviour. The proper interpretation in such cases should simply be that no straight line exists, and the test indicates that the value of cv is greater than the limiting value discussed above (0.1m2

4.3 Practical implications from consolidation behaviour

/day).

The following are the most significant factors that emerge from the behaviour shown in the figures above:

(a) Results of oedometer tests should be plotted using a linear scale and the linear parameter obtained from the graph is the most appropriate parameter to use in settlement estimates

(b) Some residual soils show a yield stress, often in the range or 200kPa to 400kPa. This is higher than normal foundation pressures, so does not present a problem.

(c) Determination of a reliable value of the coefficient of consolidation from standard oedometer tests is often not possible, because of the high cv values in many residual soils.

(d) The rate of primary consolidation during construction of buildings is likely to be high, so that most settlement will be complete at the time construction is finished.

5 SHEAR STRENGTH: UNDRAINED AND EFFECTIVE STRESS PARAMETERS.

5.1 Undrained shear strength For fine grained sedimentary soils there are

some empirical correlations relating undrained shear strength to other soil parameters. For example, Skempton (1957), has related the undrained shear strength of normally consolidated soils to the effective consolidation pressure and the Atterberg Limits.

His relationship is shown in Figure 17, and is

summarised in the expression:

( )PISu 0037.011.0 +=′σ

where Su σ′ is the effective overburden pressure

is the undrained shear strength

PI = Plasticity Index.

Figure 19 Undrained shear strength related to confining stress and PI (Skempton, 1957).

For fully remoulded soils, it is also possible to

relate undrained strength to the Atterberg Limits, and to the Liquidity Index of the soil. The undrained shear strength is considered to be about 170 kPa and 1.7 kPa at the Plastic and Liquid Limits respectively (Sharma & Padma, 2003). Slightly different values have been proposed by other authors. These values, and the curve relating undrained shear strength with Liquidity Index, are shown in Figure 20. This curve represents the lower limit of shear strength at which a soil can exist. Most soils will exist in nature with a higher undrained shear strength than that given by this curve. Only undisturbed soils that show no loss of strength on remoulding, (ie non-sensitive soils), will lie on this line.

Figure 20 Undrained shear strength versus liquidity index for several undisturbed residual soils, and fully remoulded soils.

Also shown in Figure 18 are approximate limits

for undrained shear strength and liquidity index for three residual soils. As expected there is no relationship between these values and those for remoulded soils. The undrained shear strength of undisturbed residual soils generally lies well

-0.2 0 0.5 1.0 1.4 (PL) Liquidity Index (LL)

250

200

100

0

Und

rain

ed s

hear

stre

ngth

(kP

a)

Fully remoulded soils

Volcanic ash clays

Tropical red clays

Weathered sandstone (Auckland, NZ)

0 20 40 60 80 100 120

0.6

0.4

0.2

Plasticity Index (PI)

Su

Su

S = undrained shear strength where effective vertical stress =

u

σ′

σ′

σ′ = 0.11 + 0.0037 PI


A.7-12

above the strength of the fully remoulded soil – a fact that arises because of the contribution which the structure of the soil makes to its shear strength.

It appears that the undrained shear strength of fine grained residual soils is seldom less than about 75 kPa, with the possible exception of black cotton clays, and is normally above 100 kPa. Even black clays are not particularly soft, though probably softer than most residual soils. The problem with black cotton soils is that they are prone to very large volume changes with water content changes, and also their effective strength parameters are generally very low.

5.2 Effective stress strength properties It is evident from observation of natural slopes

in residual soils that the shear strength of these materials is generally significantly higher than that of sedimentary soils. Their high effective shear strength arises from several factors, including the following: 1. Most residual soils contain clay minerals that

tend to have good frictional properties. The notable exception is black cotton soils, which contain montmorillonite.

2. Most residual soils have quite significant micro-structural effects, which contribute very positively to the shear strength of the material.

3. The micro-structure generally contributes a significant cohesive component to the shear strength of the material, ie a significant c′ value.

Figure 21 Influence of discontinuities on results of triaxial tests on residual (sandstone) soil.

In contrast to these positive factors, there is the

negative influence of discontinuities within the soil arising from its parent rock. These constitute planes of weakness, and make the determination of representative values of c′ and φ′ very difficult. Figure 21 shows results of triaxial tests on a clay derived from weathered sandstone containing

discontinuities. It is seen that there is a wide variation in strength reflecting the influence of discontinuities in the soil. Such behaviour makes the application of analytical methods extremely difficult.

Some residual soils rarely contain discontinuities. Volcanic ash clays are an example; Figure 22 shows results of triaxial tests on a very large number of volcanic ash samples. It is seen in this casse that there is only a fairly narrow scatter in the results

For intact materials, not influenced by the presence of discontinuities, the position of the soil on the Plasticity Chart provides a good guide to the likely φ′ value of the material, at least for predominantly fine grained soils. As indicated earlier, it is the position of the point in relation to the A-line that provides the best indication of likely engineering properties.

Figure 22. Triaxial tests on volcanic ash (allophone) clays (Indonesia and New Zealand).

Figure 23 The friction angle (φ′) related to position of the soil on the Plasticity Chart.

0 200 400 600 800 1000 Normal stress (kPa)

600

400

200She

ar s

treng

th (k

Pa)

c = 54kPa = 34

/

o

φ′

c = 5kPa = 25

/

o

φ′

“MIDDLE CLAY” from weathered sandstone

50

40

30

20

10

0-30 -20 -10 0 10 20Distance below or above the A-line = PI - 0.73(LL - 20)

Fric

tion

angl

e (d

egre

es)

φ′

Above A-line Below A-line

Sedimentary soilsResidual soils Volcanic soils

0 100 200 300 400 500 Normal stress (kPa)

300

200

100

She

ar s

treng

th (k

Pa)

c = 34kPa

= 35

/

/

o

φ

c = 14kPa = 34

/

/

o

φVOLCANIC ASH CLAYS(from Indonesia & New Zealand)


A.7-13

The φ′ value tends to be low for soils lying above the A-line and high for those lying below the A-line. Some limited data for residual soils is shown graphically in Figure 23. This shows φ′ values plotted against distance of the soil below or above the A-line. It is clear that there is a marked decline in the φ′ value as the soil approaches and rises above the A-line.

The soils labelled sedimentary in this figure are probably not strictly sedimentary – their parent material is believed to be sedimentary, but the weathering process they have been subject to has produced soils with the properties of black cotton soils, which are generally considered to be residual soils. The actual type of soil does not appear to be very important – it is likely that the trend shown in Figure 23 applies to all soils, regardless of whether they are sedimentary or residual. However, much more data is needed to confirm this statement.

5.3 Stress versus strain behaviour The following figures illustrate the behaviour of

two particular residual soils in triaxial tests. Figures 24 and 25 show results of tests on both undisturbed and remoulded samples of red clay, and the following figures 26 and 27 show similar results for a silt derived from the weathering of sandstone. These are the same soils for which oedometer tests were presented earlier.

Figure 24 Consolidated undrained triaxial tests on undisturbed and remoulded samples of tropical red clay.

Figure 25 Stress paths in the above triaxial tests on tropical red clay.

Figure 26 Consolidateed undrained triaxial tests on undisturbed and remoulded samples of sensitive residual silt.

Figure 27 Stress paths in the above triaxial tests on sensitive residual silt.

These tests show the following characteristics: 1. The red clay behaviour is similar in both its

undisturbed and remoulded state. When it approached failure, the deviator stress continues to increase – only slightly but

0 100 200 300 400 500 600

300

200

100

Tropical red clay Undisturbed Compacted

σ ′ σ ′1 3 +

σ′σ′

1

3-

2

2

Peak c = 14 kPa = 37

′

φ′

o

Residual

(kPa)

(kP

a)

In situ vertical effective stress

0 2 4 6 8 10 12 14

0 2 4 6 8 10 12 14

100

200

300

300

200

100

Strain (%)

Por

e pr

essu

re (k

Pa)

Dev

iato

r stre

ss (k

Pa)

Undisturbed Remoulded

50

200

400

400

200

50

Effective consolidation pressure (kPa)

Clayey silt: Liquidity Index = 0.5 Sensitivity = 10

0 100 200 300 400

200

100

σ ′ σ ′1 3 + 2

σ′σ′

1

3- 2

Clayey silt Undisturbed Remoulded

Peak c = 5kPa = 33

′

φ′

o

Residual

(kPa)

(kP

a)

In situ verticaleffective stress

0 2 4 6 8 10 12 14

0 2 4 6 8 10 12 14

500

100

200

300

400

300

200

100

Strain (%)

Pore

pre

ssur

e ( k

Pa)

Dev

iato

r stre

ss (k

Pa)

Undisturbed Compacted

500

400250

300

100

50

50100

250

300

400

500

Effective consolidation pressure (kPa)

Tropical red clay Liquidity Index = -0.17Sensitivity = 1


A.7-14

steadily. The pore pressure curves show a steady decrease indicating that the soil is behaving in a dilatant manner, ie it is tending to increase in volume. This is perhaps not surprising as it is a dense material in both its natural and remoulded state. It is still a little surprising as it is a very fine grained moderately plastic clay – such clays do not normally show dilatant behaviour.

2. The behaviour of the silt appears similar at first site, but it is significantly different. In its undisturbed state, its strength reaches a peak, and then declines at a slow but steady rate. The pore pressure shows a slight increase or remains steady. When remoulded, the behaviour is different – it now behaves in a typical silt manner, showing dilatant behaviour at all stress levels.

3. The stress paths indicate that both clays behave rather like moderately over-consolidated soils. At low stress levels they behave as overconsolidated materials, but their behaviour changes gradually as the stress level is raised and tends toward normally consolidated behaviour.

5.4 The cohesion intercept The value of the cohesion intercept, c′, usually

plays a significant role in maintaining the stability of slopes in residual soils, and this point is emphasised here. The true source of the c′ value is uncertain; the most common explanation is that it is due to some form of weak bonds between particles, and in most soils this is probably true.

Figure 28 illustrates the real nature or the c′ intercept in residual clays. This shows triaxial tests on the residual clays of Auckland that are formed from the weathering of sandstone and mudstone formations. A laboratory study of the soil involved drained triaxial compression tests and extension tests, and also a number of tension tests to investigate the behaviour of the clay at the origin of the Mohr-Coulomb failure envelope. Tension tests are a very special type of test that should not be confused with extension tests.

The latter are simple tests in which the lateral stress exceeds the vertical stress, so that the soil fails in extension, but all stresses are compressive. Both vertical and horizontal stresses are compressive and no part of the soil experiences tensile stress. Tensile tests are designed specifically to create a vertical tensile stress so that the Mohr-Coulomb envelope straddles the origin and failure takes place on planes of zero effective stress. The tension tests were drained and followed the procedure developed by Bishop

and Garga (1969). This involved the use of “dumb-bell” shaped samples as indicated in Figure 28. A full account of the study is to be found in Meyer et al (1999).

Figure 28 Tension tests to measure the cohesion intercept at zero normal stress (after Bishop and Garga, 1979).

The tests demonstrate clearly that the soil has significant tensile strength, the actual values ranging from 7.7 kPa to 12.0 kPa. They show also that the soil still has considerable strength on failure planes on which the effective normal stress is zero or less than zero. The average c′ value from these tension tests is in very good agreement with the intercept value from the Mohr-Coulomb envelope established by conventional methods. The Mohr-Coulomb parameters, c′ and φ′, given in Figure 28 are in good agreement with those obtained by back-analysis of actual failures in intact areas of the Auckland clay.

Figure 29 Cohesion intercept from back analysis of irrigated terraced ricefields.

40

-20 0 20 40 Effective normal stress (kPa)

20

She

ar s

t ress

( kP

a)

Extension failure envelope from tria

xial

tests at higher confining stresses

c = 13.7kPa, = 29.7

′

φ′

o

T

T

Shape of triaxialsample to test soil in tension

H

A

B

C

D

Terraced slope for irrigated rice cultivation

For H = 3m = 35

= 16 kN/m = 35the required value of c = 8.4 kPa on plane A-B = 15.3 kPa on plane C-D

βγφ

o

3

/ o

/

.


A.7-15

An interesting example illustrating the role of the cohesion intercept is that of terraced rice fields that are permanently irrigated from a canal supply at the top of the slope. An idealised cross-section of such rice fields is shown in Figure 29. The flow net is for an infinite slope, and is a good approximation (at least theoretically) for the central terraces on such slopes. We can back analyse this slope in two ways. First, we can analyse individual terraces, and second, we can analyse the slope as a whole. The potential failure planes on which such analysis has been carried out are indicated in Figure 29.

To obtain the c′ value, it is necessary to make an assumption about the φ′ value. A value of 35o is adopted here, which is considered to be an average value for halloysite and allophone clays on which these rice-fields are normally built, at least in Indonesia and other parts of Southeast Asia. The terraces are formed on slopes as steep as about 40o

Plane A-B 8.4kPa

, and the maximum terrace height is about 3m, although most terraces are in the 1m to 2m range. This back analysis gives the following values of c′:

Plane C-D 15.3kPa Without c′ values in this range the individual

terraces, and the slope as a whole, would not be stable. It is clear also that because of the way the terraces are formed, at least half of each vertical face must consist of remoulded soil. In fact, a larger proportion probably consists of remoulded soil as the rice farmers clear the faces of weeds and other vegetation from time to time, and in doing so scrape a layer of soil from the faces and dump it on the surface of the terraces immediately below. In this way a slow remoulded of the whole slope appears to occur. The explanation for these significant c′ values may be that there are attractive forces between the very fine, and very unusual particles contained in these clays.

5.5 Residual strength The residual strength of residual soils varies

widely, as it does in sedimentary soils, but it does not follow some of the correlations with Atterberg Limits or clay fraction derived from sedimentary soils. Figure 30 shows a plot of φ′ against PI, which is often used to illustrate a correlation between the two, as indicated by the dotted line on the chart. However, the correlation is really only an upper limit, and it is clear that if volcanic ash clays are included in the data, the correlation no longer applies.

Figure 30 Residual friction angle (φr′) versus Plasticity Index

Figure 31 Residual friction angle (φr′) versus distance above or below the A-line.

6 VOLCANIC ASH (ALLOPHANE) CLAYS.

Allophane clays, perhaps more than any other soil, illustrate the misconceptions that can arise when conventional soil characterisation procedures are applied to residual soils. On the one hand, many engineering projects involving allophane clays, such as Cipanunjang dam in Indonesia have been successfully completed without difficulty, and on the other hand the clays have something of a reputation as “problem soils”. The designation “problem” soils has arisen primarily because their behaviour does not fit comfortably into what geotechnical engineers have come to regard as normal. Their behaviour may be “problematical” in the sense that it does not fit comfortably into normal patterns, but this does not necessarily qualify them as problem soils.

0 20 40 60 80 100

40

30

20

10

1

2

3

4

56

7

89

10

1113 12

14

30

31

32

3334

35

36 37

38

39

40 41

42

43

444546

4748

49

151619

18

1720

21

22

23

24

2526

27

2829

50

51

52

53

54

55

56

5758

59

60

61

6263

64

Plasticity Index

Res

idua

l fric

tion

angl

e (d

egre

es)

φ r/

Numbers beside points are sample identification(see Table 1, Wesley, 2002)

Clays in generalVolcanic ash clays

-100 -50 0 50 100 Distance above or below the A-line: PI = PI - 0.73(LL - 20)∆

40

30

20

10Res

idua

l fric

tion

angl

e

(deg

rees

)φ′ r

Above A-lineBelow A-lineVolcanic ash claysClays in general

Silty claySilt Clay

A-li

ne

0


A.7-16

The very fine grained nature of the soils, together with their extraordinarily high natural water contents, creates the expectation that they will be highly plastic, of high compressibility and of low strength, and that they may even be a source of shrinkage and swell problems. None of these is the case; in fact the opposite is generally true, so that conventional geotechnical characterisation “indicators” are shown to be faulty. Their unusual characteristics do, however, give rise to some problems, one of the most prominent of which has been in the area of compaction and compaction control.

6.1 Performance in natural hill slopes The stability of these soils on steep slopes has

already been commented on with respect to the steep, irrigated, terraced rice-fields of Southeast Asian countries. Rouse et al (1986) report that in Dominica slopes remain stable at inclinations of around 40o, even when used for crop cultivation. Belloni and Morris (1991) report on slopes in Ecuador with inclinations ranging from 35o to 55o

6.2 Formation of allophane clays.

. The wet climates in which these slopes exist, and their use as rice-fields mean that there must be significant seepage pressures within them for much of the year. Yet they remain stable despite the steep slopes and high seepage pressures.

The formation and composition of allophane clays is complex, and most of the research on the subject comes from the discipline of soil science rather than soil mechanics. This research shows that allophane seldom occurs by itself. Instead, it is almost invariably found with other clay minerals, especially a mineral called imogolite. Secondly, it shows that allophane is not strictly amorphous, as early literature asserted. Both allophane and imogolite have some crystalline structure, albeit of a very different nature to other well known clay minerals.

Allophane clays are formed by the in situ weathering of volcanic material, primarily ash, although they may be formed from other volcanic material. The essential condition for allophane formation is that the parent material consist of non-crystalline composition. Volcanic ash meets this criteria; the fine particles consist of volcanic “glass”, that is, they are non-crystalline (or amorphous), similar to the shiny black rock known as obsidian

In addition to the above requirement of non-crystalline parent material, it appears that the weathering environment must be well drained, with water seeping vertically downward through

the ash deposit. High temperatures also appear to favour or accelerate the formation of allophane clays. Allophane clays may be very deep; in Indonesia the writer has encountered cuts in these materials up to about 30m deep, while site investigation drilling has shown depths of up to almost 40 metres.

Allophane is believed to undergo further chemical weathering through the following sequence: Volcanic ash allophane halloysite kaolinite sesqui-oxides laterite

This weathering process is essentially one of chemical conversion and leaching out of silica by seeping pore water. As the silica content decreases, the concentration of iron and aluminium increases, in the form of sesqui-oxides, that is the hydrated forms of iron and aluminium oxide (geothite and gibbsite). These tend to act as cementing agents which bring about the formation of the hard concretions which make up laterite.

6.3 Structure of allophane

Figure 30 Structure of allophone, imogolite, and halloysite.

The precise structure of allophane clays is

somewhat problematic. Their extraordinarily high natural water contents and void ratios clearly indicate an unusual material, and call for an

Allophane spheres

Imogolite threads

(b) halloysite

(a) allophane and imogolite


A.7-17

explanation in terms of either structure or chemical composition (or both). Studies over the past 20 years (Wada, 1989) show that the material in its natural state does have an ordered structure consisting of aggregations of spherical allophane particles with imogolite threads "weaving" among them, or forming “bridges” between them, as illustrated conceptually in Figure 32.

This is based on an electron micrograph (Wada, 1989) of the material in its undisturbed state. Also shown in Figure 32 is the form of halloysite clay minerals, which frequently occur in association with allophone, and are a progression of the weathering sequence that forms allophane.

6.4 Natural water content and Atterberg limits. The natural water content of allophane clay

covers a very wide range, from about 50% to 300%. This corresponds to void ratios from about 1.5 to 8. It appears that water content is a reasonable indication of allophane content – the higher the water content the greater the allophane content. Atterberg Limits similarly cover a wide range, and when plotted on the conventional Plasticity Chart invariably lie well below the A-line. This means that according to the Unified Soil Classification System they are silts. However they do not display the characteristics normally associated with silt – the tendency to become “quick” when vibrated and to dilate when deformed. At the same time they are not highly plastic like true clays, so they do not fit comfortably into conventional classification systems. Figure 31 shows a plot of the Atterberg limits of allophone clays on the Plasticity Chart.

Figure 31 Allophane clays on the Plasticity Chart. For volcanic ash clays in Java, Indoneisia, there

is a reasonably clear correlation between the relative proportions of allophane and halloysite in the soil and the water content. Increases in natural water content and Atterberg limits accompany increases in allophane content and corresponding decreases in halloysite content. The relationship is illustrated in Figure 32.

Figure 32 Natural water content and Atterberg limits versus allophone content.

6.5 Degree of saturation, liquidity index, and sensitivity

Data from two profiles are shown in Figure 33. This illustrates two important points. First, the soils are essentially fully saturated except within the top one or two metres from the ground surface. Second, the natural water content fluctuates rather randomly with respect to the plastic limit and the liquid limit. This means that the liquidity index varies from about zero to greater than unity, with an accompanying variation in sensitivity

Measured sensitivity values in the upper profile in Figure 33 range from about 1 to 2.5. In the lower profile sensitivity was not measured directly, but the water content exceeds the liquid limit at a depth between 10m and 16m, indicating very high sensitivity soil. The author has encountered moderate to high sensitivity at other sites in Indonesia. The sensitivity of allophane clays in New Zealand appears to be generally higher than in Indonesia.

6.6 Compressibility and consolidation characteristics.

Typical results from oedometer tests on undisturbed samples from Indonesia and New Zealand are shown in Figure 34, plotted in the usual manner and also using a linear scale for pressure. As discussed earlier, the conventional e-log(p) plot presents a misleading picture of soil compressibility soil. The log curves in Figure 34 (a) suggest that all the samples have similar

0 40 80 120 160 200 240 Liquid Limit

80

40Pla

stic

ity In

dex

A-line 60

20

100Allophane clays from Indonesiaand New Zealand

250

150

50

0 20 40 60 80 100 Allophane (%)

Nat

ural

wat

er c

onte

nt a

nd A

tterb

erg

limits

(%)

Liquid limit Natural water content Plastic limit

Note:The total % of allophane plus halloysite is about 90%. The remaining 10% is made up of coarser particles.200

100


A.7-18

compression characteristics with yield pressures of varying magnitude.

Figure 33 Typical profiles of basic properties of volcanic ash (allophone) clays. However, when plotted using a linear pressure

scale this is clearly no longer the case. Only Samples 1 and 2 show a yield pressure. This yield pressure arises from the structure of the soil created by the weathering process, but why some samples show a yield pressure and others do not is unknown. It is likely to be related to the original denseness of the parent material, but this is uncertain. These graphs emphasise the points made earlier, namely that only the linear plot conveys a true picture of the compression behaviour, and also that the linear parameter mv is likely to be a more appropriate parameter for estimates of settlement.

Figure 34(b) suggests that the compressibility of these soils is not very different despite the large variation in void ratio shown in Figure 34(a). Figure 35 further illustrates this fact; it shows values of the constrained modulus D (the inverse of mv

Regarding the time rate of consolidation of these clays Figure 17 shows a root time graph from an oedometer test at a low pressure increment. It is not possible to determine reliable values of the coefficient of consolidation from such a graph because there is no initial straight section. To accurately measure c

) at both low and high pressure increments versus initial void ratio.

v

Water content PL and LL0 50 100 150

Water content, PL and LL0 50 100 150 200

Clay fraction (%)20 60 100

S (kPa)60 100 140

u

S (kPa)60 100 140

u

Sensitivity 0 1 2 3

Degree of saturation (%)90 100 110

2

4

6

8

10

Dep

th (m

)

Dep

th (m

)

10

20

30

LLwPL

LLwPL

5

15

25

it is necessary


A.7-19

to carry out pore pressure dissipation tests in a triaxial cell.

Figure 34 Oedometer tests plotted on both log and linear scales.

Figure 35 Constrained modulus (D) versus void ratio.

Figure 36 Measurements of k and cv in pore pressure dissipation tests in a triaxial cell.

A summary of the result such tests is presented in Figure 36. The tests were carried out on four

Pressure (kPa) 10 100 1000 5000

Pressure (kPa) 0 500 1000 1500 2000

5

4

3

2

1

10

30

20

40

Com

pre s

sion

(Ver

tical

stra

in %

)

Sample 123456

(a) log scale

(b) linear scale

20

15

10

5

0

1 2 3 4 5 6

15

10

5

(a) 0 - 200 kPa

(b) 1600 - 2000 kPa

Initial void ratio

Cons

train

ed m

odul

us D

(kPa

)

20 100 1000 2000 Pressure (kPa)

New Zealand samples - undisturbed

Indonesian samples - undisturbed

New Zealand samples - remoulded

20 100 1000 2000 0.0004

0.0004

50

10

1

0.1

0.01

0.001

50

10

1

0.1

0.01

0.001

Pressure (kPa)

Coe

ffici

ent o

f con

solid

atio

n (c

m/s

ec)

2C

oeffi

cien

t of c

onso

lidat

ion

(cm

/sec

)2

100

10

1

0.1

0.01

Coe

ffici

ent o

f per

mea

bilit

y (1

0 m

/sec

)-9

Coefficient ofpermeability

Coefficient ofconsolidation

Indonesian samples only - undisturbed

(a) Pore pressure dissipation tests

(b) Pore pressure dissipation and permeability tests


A.7-20

undisturbed samples, two each from Indonesia and New Zealand. On the New Zealand samples, the dissipation tests were repeated after thoroughly remoulding the samples. Figure 36(a) shows the coefficient of consolidation values obtained from all the dissipation tests. It is seen that the cv

The c

value decreases by approximately four orders of magnitude as the stress increases from 50 to 1000 kPa.

v value from the remoulded samples is consistently low and close to the end (high stress level) value from the undisturbed samples Remoulding the soil apparently destroys the open structure of the undisturbed soil which is believed to account for the high permeability With the Indonesian samples, permeability measurements were also made between each consolidation stage; the results are shown in Figure 36(b). The coefficient of permeability shows an identical trend to the cv values, as would be expected, confirming that it is the very high permeability of the soil which accounts for the high cv

6.7 Effective stress strength parameters

values.

Comment has been made earlier on the high values of friction angle found in these clays. A further feature is the small difference between peak strength and residual strength. This is illustrated in Figure 37.

Figure 37 Peak and residual strength of allophone clays.

This drop from peak to residual is very small

compared with that normally found with sedimentary soils. The exact explanation for this is not known but part of the explanation is that allophone and imogolite particles are not “plate-

like” and therefore particle orientation to form smooth, slickenside surfaces does not occur.

6.8 Compaction characteristics Allophane clays show unusual compaction

behaviour in two respects. Firstly, their compaction curves from Proctor compaction tests do not always show clear “peaks” of maximum dry density. There is thus no optimum water content at which to compact them. Secondly, their compaction curve is dependent on the extent to which they are dried prior to carrying out the test.

Tests were carried out by simply drying the soil from its original water content for each point on the compaction graph, and tests repeated after air and oven drying of the soil. For sample (b) a further test was done after drying only to a water content of 65%. Sample (a) hardly shows a peak dry density, but sample (b) shows a flat peak at a water content of about 135%.

Figure 38 Compaction tests on allophone clays.

7 COMPACTION CONTROL FOR RESIDUAL SOILS.

7.1 Difficulties in compaction of residual soils Control of compaction with residual soils by the

usual method of specifying water content and dry density limits can be difficult for several reasons. Firstly, many residual soils are extremely variable and there is no fixed value of optimum water content. Figure 39 shows an example of this situation. The tests were all done on samples from a relatively small industrial site where the soils were residual derived from volcanic materials.

Indonesian samplesNew Zealand samples

Residual strength from ring shear tests:

Peak stre

ngth from tri

axial te

sts c

= 20kPa,

= 40

/

o

φ′

0 100 200 300 400 500 600 Normal effective stress (kPa)

500

400

300

200

100

She

ar s

tress

(kP

a)

Residual strength c = 0, = 36′ φ′ o

20 40 60 80 100 120 140 160 180 200

20 40 60 80 100 120 140 160 180 200

1.2

1.0

1.2

1.0

0.8

0.6

0.8

0.6

0.4

0.4

Water content (%)

Water content (%)

Dry

den

sity

g/c

m3

Dry

den

sity

g/c

m3

Zero air voids

Zero air voids

Natural Air driedOven dried Air dried to 65%

Natural Air driedOven dried

Sample (a)

Sample (b)


A.7-21

Figure 39 Compaction test curves from a small industrial site.

Secondly, there is the absence of clear

maximum dry density and optimum water content values, as indicated above for allophone clays.

Thirdly, there is the difficulty of some residual soils having water contents well above the optimum value in climates which make drying very difficult. It is important in these situations to properly evaluate the compaction characteristics of the soil, and if possible adopt methods for compacting the soil that minimise the need for drying.

Figure 40 shows an example of this. The tests here are on volcanic soils in Japan, and involve compaction using a range of compaction efforts corresponding to variation in the number of blows with the compaction rammer. The tests are done at the natural water content of the soils. The strength of the soil after compaction has been measured using a simple cone penetrometer. It is immediately clear that most of the soils become softer as the number of blows increases, although three samples show an initial increase in strength before the steady decline in strength.

These graphs suggest that there is an optimum compactive effort for compaction of these soils, and that it is a relatively light effort. It is important to recognise that compaction has two effects on soil; firstly to force the particles closer together and hopefully make the soil stronger, but secondly an unintended effect of destroying the natural structure of the soil and making it weaker

Whether compaction is beneficial depends on which of these two influences is stronger. If drying the soil is not practical, then tests of the kind illustrated in Figure 40 should be carried out to determine the feasibility of effectively compacting the soil without drying it. In situations

such as this a different method of compaction control is desirable, as described in the next section.

Figure 40 Influence of compactive effort on the strength of volcanic soils. (after Kuno et al,1978).

7.2 Compaction control using undrained shear strength and air voids.

An alternative to the conventional use of dry density and water content to control compaction is the use of undrained shear strength and air voids. This method was developed in New Zealand to cope with the rapid variations in properties that occur in many local residual soils, such as those illustrated in Figure 39. The method is described in detail by Pickens (1980), and only an outline of the method is given here. Figure 41 illustrates the basis for using undrained shear strength as one of the control parameters; it shows the results of a standard compaction test on clay, during which undrained strength has been measured in addition to density and water content. The measurements were made using both a hand shear vane and unconfined compressive tests. The two strength measurements give significantly different results.

It is seen that at the optimum water content the undrained shear strength is about 150 kPa from the unconfined tests and about 230 kPa from the vane tests. These values are to be expected; the optimum water content from a standard Proctor test on clay is normally close to the plastic limit, where the undrained shear sstrength is normally in the range of 170 kPa to 200 kPa. Conventional

Steel mill site: Weathered basalt and ashes

40 50 60 70 Water content (%)

1.3

1.2

1.1

1.0

0.9

Dry

den

sity

(gm

/cm

)3 Zero air voids line

16

12

8

4

0 20 40 60 80 100 120 Number of Rammer Blows

Con

e In

dex

qc Kanuma soil

w = 220%

Volcanicash soil w = 59%

Solid lines are various Kanto loams

w = 110%

w = 121%

w = 117%

w = 108%

w = 109%

A

B

C

DE

Arrows indicate “optimum compactive effort”


A.7-22

compaction specifications may allow water contents 2 or 3% greater than optimum, in which case the comparable shear strength values would be about 120 kPa and 180 kPa. Thus to obtain a fill with comparable properties to those obtained with normal control methods, specifying a minimum undrained shear strength in the range of 150 kPa to 200 kPa would be appropriate. This would put an upper limit on the water content at which the soil could be compacted.

Figure 41 Measurements of undrained shear strength during a standard compaction test.

However, this required strength could be achieved by compacting the soil in a very dry state, which would generally be undesirable, as dry fills may swell and soften during rainfall. To prevent the soil from being too dry a second parameter is specified, namely the air voids in the soil.

Figure 42 Compaction control using undrained shear strength and air voids.

At “standard” optimum water content the air voids in the soil is generally about 5%. If the soil is compacted 2 to 3 % drier than optimum, the air voids may be as much as 8 or 10%. Thus to prevent the soil from being compacted too dry an upper limit is placed on the air voids, normally in the range of 8% to 10%. Figure 42 illustrates how this method of controlling compaction relates to the traditional method.

The zero air voids line is the upper limit of dry density for all water contents, and thus applies to both methods. The traditional method involves an upper and lower limit on water content, and a lower limit on dry density, and thus encloses the area shown in the figure. The alternative method indirectly places an upper limit on water content, corresponding to the minimum shear strength, and a lower limit on dry density corresponding to the line parallel to the zero air voids line representing the upper limit of air voids. There is no specific lower limit of water content, but the air voids limit prevents the soil from being too dry, as normal compaction equipment will be unable to achieve the required air voids limit.

- Experience has shown that suitable limits for the two control parameters are as follows:

- Undrained shear strength (hand vane values): Not less than 150 kPa (average of 10 tests)Minimum single value: 120 kPa.

- Air voids (for “normal” soils): Not greater than 8%

These values have been found to be very satisfactory in producing firm, high quality fills. The undrained shear strength can be measured in situ by hand shear vane, or by taking samples for unconfined compression tests. The hand shear vane is the much simpler of the two methods. The air voids can only be determined by measuring the density, water content, and specific gravity, in the usual way.

REFERENCES Bishop, A.W. and Garga,V.K, (1969). “Drained tension

test on London cla”. Geotechnique, 19 (2): 309-313. Kuno, G. Shinoki, R. Kondo, T. & Tsuchiya, C. (1978).

On the construction methods of a motorway embankment by a sensitive volcanic clay. Proc. Conf. on Clay Fills, London. 149-156.

Little, A.L. (1969) “The engineering classification of residual tropical soils”. Proc. Specialty Session on Engineering Properties of Lateritic Soils. Seventh Int Conf on Soil Mechanics and Foundation Engineering. Mexico, 1969, (1): 1-10.

Meyer, V.M., Pender, M. J. and Wesley, L.D. (1999) “The very low effective stress behaviour of a residual soil”. Proceedings 8th

20 25 30 35 40 Water content (%)

1.5

0

100

200

300

1.6

1.4

1.3

1.7Vane testsUnconfined comp. Tests

Dry

den

sity

gm

/cm

3

Und

rain

ed s

hear

stre

ngth

(kP

a)

Opt

imum

wat

er c

onte

nt

Australia- New

Water content

Dry

den

sity

She

ar s

treng

th

Zero air voids

Air voids limit

Shear strength

Shear strength limit

Dry density limit

Water contentlimits from compaction test

Wat

er c

onte

nt li

mit

from

sh

ear s

treng

th c

riter

ia

Limits from water content and dry density criteria

Limits from shearstrength and airvoids criteria


A.7-23

Zealand Conference on Geomechanics, Hobart. (2): 877-883.

Pickens, G. A. (1980). “Alternative Compaction Specifications for Non-uniform Fill Material”. Proc. Third Australia-New Zealand Conf. on Geomechanics, Wellington (1): 231 –235.

Prusza, Z. Kleiner, De. & Sundaram, A.V. (1983). “Design, construction and performance of large dams on residual soils”. Proc. 7th

Skempton, A.W. (1957) “Discussion on the planning and design of the new Hong Kong airport”. Proceedings Institution of Civil Engineers, (7): 305-307

Pan Am. Conf. Soil Mechanics and Foundation Engineering. Vancouver, (1): 185-198

Wesley, L. D. (2002). “Geotechnical characterization and behaviour of allophane clays”. Proc. International Workshop on Characterisation and Engineering Properties of Natural Soils. Singapore, 2002, (2): 1379-1399. Balkema


A.7-24

ce7 - l. wesley

Documents