2. Traditions in soil mechanics - an asset
or a liability?
- Following traditions gives us a feeling of
security, which is an asset,
but:
- Some traditions may become established
without a secure base and become a means of
perpetuating errors and misunderstandings
This paper looks at a few of these
4. How stable is a vertical clay bank?
- especially the sides of a vertical trench?
- and what can soil mechanics tell us on this
question?
• Deaths from trench collapses are regularly
reported in newspapers worldwide.
• It appears that there are still one or two
deaths each year in New Zealand in what are
wholly preventable circumstances.
5. Can geotechnical engineers predict the
stability of vertical banks in clay?
• Soil mechanics text books give the following
expressions for the maximum (critical) height of
unsupported vertical banks in clay:
In terms of effective stress parameters:
a
H c
ʹ′
=
4
c γ
K
In terms of undrained shear strength
H 4S
γ
u
c
=
6. These expressions are of theoretical interest
only, and very rarely are of practical relevance.
• Unfortunately, the cold hard fact is that this
is one area where geotechnical engineers
cannot provide reliable answers.
• Neither theory or experience can ever enable
reliable predictions to be made about the
height or the length of time at which a
vertical clay bank will remain stable.
7. 24m
2.4m
8m 7.2m 1.5m
Total stress
analysis
Effective stress
analysis
Regulation
safe depth
Critical depth
“Safe” depth
(S.F. = 3)
70o
What does theory tell us?
A typical residual clay:
Su = 100kPa γ = 16kN/m3 cʹ′ = 15kPa φʹ′ = 35o
Theoretical heights are quite unrealistic.
Site safety authorities limit the height of banks or depths of
trenches where people work to 1.2m to 1.5m
24m
5m 10m
45o
60o
Field observation
8. BANK OF AN OPEN EXCAVATION
- pushes sideways, buries, crushes.
TRENCH
- traps, buries, crushes.
Vertical banks on their own are dangerous
But trenches are much more dangerous - there are no
escape routes, and they immediately trap anyone in
them; death is normally from suffocation, but may also be
from internal injuries.
9. Government regulators appear to know more
about the stability of vertical banks than
geotechnical engineers
Codes of Practice allow people to work in an
unsupported trench only to a depth of 1.5m.
Beyond this height, the trench must be battered
back, or shoring or a shield used.
For soft “normally consolidated” clays, even
1.5m is too deep.
10. The appearance of a
trench tells us very
little, if anything
about its stability,
except that it is
standing up at the
time of observation
13. Erosion by rainfall
and runoff
Residual soil
Rock
Sea or lake level
Sedimentary soil
Delta
deposits
Transport by stream
and river
Residual soils – mode of formation does not involve a sedimentation
and consolidation process
Why then does the profession still interpret their consolidation
behaviour as though they have undergone a consolidation process -
using e-logp graphs and calculating Cc and Cs parameters?
14. Professor Nilmar Janbu (Norway):
“--- it remains a mystery why the
international profession still uses the
awkward e-logp plots, and the incomplete
and useless coefficient Cc which is not
even determined from the measured data,
but from a constructed line outside the
measurements ---”
- an observation based in his experience with
sedimentary soils – it is even more valid with residual
soils.
15. 0.8
Stress range of interest to
geotechnical engineers
C ? c
Maximum stress level
in most oedometer tests
0.6 0.6
0.4
0.2
Cs
25 100 1000 10000
Pressure (kPa)
Void ratio
0.8
0.4
0 5000 10000 15000 20000
Pressure (kPa)
Void ratio
0.2
(a) logarithmic plot (after Lancellotta, 1995) (b) linear plot
Typical consolidation test result from an over-consolidated sedimentary clay
- illustrating the points Janbu makes and the uncertainty of the parameters
Cc and Cs
16. Pressure (kPa)
1 10 100 1000
Pressure (kPa)
0. 500 1000
Log and linear scales – the so called pre-consolidation
pressure is close to the strain hardening pressure
0
6
0
6
2
8
2
8
Compression units
Compression units
Pre-consolidation
pressure = 90kPa
Stiffening pressure
(start of strain hardening) = 200kPa
a
b
(a) log plot (b) linear plot
4
10
4
10
17. The e-log(p) graph:
There are severe defects
with this plot regardless
of whether the soil is
residual or sedimentary
- examples in text books
- log scale for pressure
provides a severely
distorted picture of
compression behaviour
and leads to routine
misinterpretation of true
behaviour
σʹ′c
σʹ′c
σʹ′c
σʹ′c
A
σʹ′c
σʹ′c
σʹ′c
Pressure (linear scale)
Void ratio
Void ratio
Void ratio
A
B
C
D
Horizontal line
Tangent
D F
C B
Cr
Cc
Void ratio
b
d
c
h
a
g
f
Void ratio
Void ratio
after Craig (1992)
after Budhu (2000)
after Das (1997)
18. Oedometer tests on
volcanic ash clays
Log plots create the
impression that all soils
have similar compression
behaviour
The linear scale shows
three distinctly different
types of stress vs
deformation behaviour
B9-3M OCR = 4.0
B9-4M OCR = 3.6
B7-5M OCR = 3.4
B8-7M OCR = 3.4
B8-8M OCR = 3.3
B7-9M OCR = 1.1
1.6
1.2
0.8
0.4
(a) log scale
10 100 1000 10000
Void ratio
0 500 1000 1500 2000
10
20
Compression (%)
30
Pressure (kPa)
(b) linear scale
19. Oedometer tests on
Auckland residual clay
- from the weathering of
sandstone
Apparent pre-consolidation
pressures
from the log plot
disappear when the
curves are re-plotted to a
linear scale
Pressure (kPa)
(a) log plot
0 500 1000 1500
(b) linear plot
Vertical strain (%) Vertical strain (%)
5
10
14
20. Oedometer tests on
volcanic ash clays
- the log plot suggests
that the compressibility of
each sample is the same
- the linear plot shows
very clearly that this is
not the case
- Sample A shows a clear
“yield” pressure, Sample
B shows linear behaviour,
and Sample C shows
strain hardening.
Pressure (kPa)
10 50 100 500 1000 5000
0 100 200 300 400 500
2
Void ratio
Pressure (kPa)
Compression (%)
A
A
B
B
C
C
(a) log scale
(b) linear scale
4
6
8
10
12
14
21. A more realistic
portrayal of soil
compressibility
- valid for all soil
types
Note also that the
linear parameter mv
is normally a much
more appropriate
parameter to use
for settlement
estimates than the
log parameters Cc
and Cs
Pressure (linear scale)
Strain
Strain hardening
Yielding (strain softening)
Linear
Vertical yield
pressure
Yield from
structural breakdown
Strain hardening is typical
of dense soils with low
liquidity index
Strain softening is typical
of non-dense soils with high
liquidity index
23. Square root of time (min0.5)
0 1 2 3 4 5
Average degree of consolidation (%)
C v
= 0.001 cm /2
sec C v
= 0.01 cm 2
/sec C v
= 0.1 cm 2
/sec √ time √ min.
0 2 4 6
Volcanic ash soil
Waitemata clay
(w
Tropical red clay
eathered sandstone)
Oedometer sample thickness = 20mm
Some consolidation graphs – actual and theoretical
- the absence of a linear section of the graph indicates that compression is
not governed by the rate of pore pressure dissipation – the pore pressure
has dissipated within a few seconds of the load being applied
2
4
6
8
Compression (%)
10
20
40
60
80
100
24. Valid and invalid
constructions
The highest value of
the coefficient of
consolidation cv
that can be
measured in a
conventional
oedometer test is
about 0.01 cm2/ sec.
Many residual soils
have higher cv
values than this
0 2 4 6 √ time √ min.
2
4
6
8
Compression (%)
10
√t √t (??) 90 90
Sample A
Sample B
Valid
Invalid
25. Estimation of the rate of consolidation of
surface foundations.
Text books and soil mechanics courses
normally only cover one dimensional
consolidation – which cannot be applied to a
surface foundation.
So what do geotechnical engineers do in this
situation??
26. Uniform load of infinite width Load of limited width
Known boundary
One-dimensional conditions assumed
by the Terzaghi consolidation theory
Conditions applicable to most foundations,
especially on deep residual soils
Terzaghi one dimensional consolidation and
the situation at most foundations
seepage paths
Remote, and possibly unknown boundary
1D vertical consolidation & drainage only
3D consolidation
& drainage
27. Solution for an
impermeable
foundation on a layer
overlying an
impermeable
boundary – adapted
from Davis and
Poulos, 1972
0
0.2
0.4
0.6
0.8
1.0
Degree of consolidation U
Time factor T = c t / b
s v
2
0.1 0.3 0.5 0.7 1 3 5 7 10 30 50 70 100
10
5
2
0.5 Values of h 1
Strip footing
b
h
b
0
0.2
0.4
0.6
0.8
1.0
Degree of consolidation U
Time factor T = c t / a
c v
2
0.04 0.1 0.3 0.5 0.7 1 3 5 7 10 30 40
5
2
0.5
h 1
Values of a
Circular footing
a
h
20 (=50)
10
(a) Strip footing
(b) Circular footing
Impermeable base
impermeable layer
Impermeable base
impermeable layer
28. SOME OBSERVATIONS ON THE WATER TABLE AND
SEEPAGE STATE IN UNCONFINED FLOW IN COARSE
GRAINED MATERIALS AND CLAYS
the subject is taught as though there is no difference, but
this cannot be true in the case of unconfined flow, where
there is a profound difference.
Only in coarse materials is the water table a sharp boundary
between two zones – a lower one having pore pressures and
seepage, and an upper one with none
29. Ground surface
Saturation boundary for clay
Saturation boundary
for sand or gravel
Water table
Pore water pressure
The water table in a level static situation
Negative pore pressure
Positive pore
pressure
Negative Positive
a
b
u = - γwa
u = γwb
Hydrostatic
(equilibrium)
pore pressure
30. Two possible seepage states for
the same water table
Only in coarse materials is the
water table the boundary of the
seepage zone. In clays seepage
occurs above the water table
according to the same laws as
below it.
In clays water cannot drain out
under gravity and the soil above
the water table remains fully
saturated.
The term “unconfined flow” is
not really correct for clays – the
ground surface is really the
upper boundary of the seepage
zone.
Ground surface
Measured water table
(a) Normal assumption of flow net for the given water table
- implies a coarse material and an external re-charge source
Zone of limited
rainfall re-charge
Zone of negative pore pressure
Measured water table
Zone of positive pore pressure
(b) A valid flow net for limited re-charge from rainfall on the slope
a
b
a
b
d
c
d
c
31. Homogeneous
clay embankment
Phreatic surface
Phreatic surface
Drainage layer
Drainage layer
Flow net in homogeneous earth dams
Water level
Water level
Impermeable rock
(a) Flow net as normally depicted - only correct for sand or gravel
(b) Correct flow net for clays taking into account seepage above the phreatic surface
32. Influence of rainfall on coarse materials and clay
Swell
Depth limit
of swell
Ground surface (unchanged) Initial ground
surface
Final ground
surface
Final
water
level
Initial
water
level
Clean sand or gravel Clay
33. Influence of
dewatering in
coarse materials
and in clay
The mechanics are
quite different in
each case
In coarse materials
the governing
parameters are k,
the permeability and
n, the porosity
In clays they are the
compressibility mv
and the coefficient
of consolidation cv
Ground surface (little change by water table lowering)
Initial water table
Air replaces water
in this zone
Pumped
excavation
Unsaturated zone
Final phreatic surface
Fully saturated zone
(a) Rigid granular material (sand or gravel)
Initial water table
Pumped
excavation
Ground surface settles as
groundwater is lowered
Negative pore pressure zone
Final phreatic surfaceFull saturation above and
below the phreatic surface
Postive pore pressure zone
(b) Compressible fine grained material (clay or silt)
34. Estimation of the stability of steep slopes subject to
prolonged rainfall
A “worst case” assumption can be made that the water
table rises to the ground surface
The results of a conventional slip circle analysis can
give very misleading results if the water table is put in at
the ground surface using a computer programme
– because the computer programme calculates the pore
pressure from the vertical intercept between the water
table and the slip surface.
A realistic flow net will give a much higher safety factor
35. Water table input: SF = 1.09
Flow net input: SF = 1.36
Soil properties:
γ
= 16.5 kN/m
3
c ʹ′
= 50 kPa
φʹ′
= 40
o
40m
Steady rainfall on
ground surface
metres
0 10 20 30 40
Influence of assumed pore pressure state on safety factor
36. Compacton of residual soils
- There are three problems:
1. Variability
2. Sensitivity
3. Flat compaction
curves(volcanic ash clays)
37. Steel mill site:
Weathered basalt and ashes
40 50 60 70
Water content (%)
1.3
1.2
1.1
1.0
0.9
Dry density (gm/cm )3
Zero air voids line
Highly
variable
properties of
many residual
soils
especially
volcanic soils
38. Basis of an alternative compaction control method:
By using undrained shear strength and
air voids, a similar quality of fill can
be obtained using a uniform
specification regardless of the
variability or the soil
39. Conventional
Proctor tests
for compaction
monitoring
- note the air
voids lines
1.6
1.5
1.4
1.3
1.2
1.1
1.0
18 20 25 30 35 40 42
Water content (%)
Dry density (tonne/m = gm/cm ) 3 3
Zero air voids (a ) line (S = 100%) v r
Modified
Standard
γD (modified)
γD (standard)
Optimum w/c (standard)
Optimum w/c (modified)
a = 5% v
a = 10% v
40. Shear
strength
versus
compaction
water content
- note the
shear
strength at
optimum
water content
300
200
100
1.7
20 25 30 35 40
Water content (%)
1.5
0
1.6
1.4
1.3
Vane tests
Unconfined comp. Tests
Dry density gm/cm3
Undrained shear strength (kPa)
Optimum water content
41. Basis of the alternative compaction control method:
1. The undrained shear strength of a clay at standard
Proctor water content is normally in the range of
150kPa to about 200 kPa. Specifying a minimum
undrained shear strength prevents the soil being too
wet.
2. The air voids in a clay compacted close to optimum
water content is normally about 5% to 8%. Specifying
an upper limit to air voids prevents the soil being too
dry.
3. Controlling the undrained shear strength and air voids
will produce fill of similar characteristics as the
conventional method of compaction control
42. Alternative
method and
the
conventional
method of
compaction
control – the
undrained
shear
strength/air
voids
method
produces a
fill of similar
properties
Water content limit from
shear strength criteria
Water content
Dry density
Shear strength
Zero air voids
Air voids limit
Shear strength
Shear
strength
limit
Dry density
limit
Water content
limits from compaction test
Limits from water
content and dry
density criteria
Limits from shear
strength and air
voids criteria
43. Softening during the compaction process
It is important to recognise that compaction of a soil
can have two effects:
(a) “Densifying” the soil, ie pressing the particles
closer together and squeezing out air.
(b) Remoulding the soil, causing it to soften.
Most natural soils lose some strength on remoulding
- compaction is a form of remoulding.
Compaction destroys bonds, crushes particles, and
releases water trapped in the structure of the soil.
44. 16
12
8
4
w = 110%
0 20 40 60 80 100 120
Number of Rammer Blows
Cone Index qc
Kanuma soil
w = 220%
Volcanic
ash soil
w = 59%
Solid lines are
various Kanto loams
w = 121%
w = 117%
w = 108%
w = 109%
A
B
C
D
E
Arrows indicate “optimum
compactive effort”
Many volcanic ash
soils are sensitive
and lose strength
as compactive
effort is increased
- there us thus an
“optimum
compactive effort”
at natural water
content
- and at other water
contents
45. With sensitive, highly structured soils, such as volcanic
ash clays, the traditional Proctor approach for
compaction which puts all the emphasis on density is
often inappropriate
Drying the soil may not be a feasible option, in which
case compaction at the natural water content is the only
possibility
To do this, as much of the soil structure should be
preserved – this means using relatively light
compaction equipment so that the soil is “pressed
together” rather than “rammed” in the normal way.
46. Typical
compaction
curves from
high
allophane
content soils
- in this case
the correct
procedure for
testing the
soil must be
followed,
and field
trials may be
desirable
1.2
1.0
0.8
0.6
0.4
Zero air voids
Natural
Air dried
Oven dried
20 40 60 80 100 120 140 160 180 200
1.2
1.0
0.8
0.6
0.4
Water content (%)
20 40 60 80 100 120 140 160 180 200
Water content (%)
Dry density g/cm3 Dry density g/cm3
Zero air voids
Natural
Air dried
Oven dried
Air dried to 65%
Sample (a)
Sample (b)
47. Control measurements
• Shear strength – various options, including in
situ vane tests and penetrometer tests, or
undisturbed sampling for laboratory tests.
Hand vane tests are the simplest.
• Air voids – in the usual way, by measuring
water content, density, and specific gravity
48. Falling weight
hammer
(a) DYNAMIC
PENETROMETER
(FALLING WEIGHT)
Handle to apply
manual push
(b) STATIC
Hand shear vane and hand penetrometers
PENETROMETER
(DIRECT PUSH)
Graduated scale
to measure penetration
Fixed fall
height
Proving ring to
measure force
Torque Guage
Vane pushed by
hand into soil.
49. Concluding remarks:
To become a good geotechnical engineer:
1. Be curious, even inquisitive
2. Take every opportunity to observe soil behaviour in the
field.
3. Don’t accept conventional wisdom before thinking it
through, and continue thinking about it each time you take
on a new project.
4. Don’t become hypnotised by theoretical knowledge –
recognise its limitations, and don’t seek to impose on a
soil preconceived ideas of how a it should behave.
5. Don’t be side-tracked by irrelevant theories such as
critical state soil mechanics. Stick with real soils, not
theoretical ones.
50. References from which the material in my
presentations is taken:
Fundamentals of Soil Mechanics for
Sedimentary and Residual Soils
Geotechnical Engineering in Residual
Soils
(both published by John Wiley and Sons)