caseweste:rn-reserveuniversitv · quite different from the 30 psi q07 kpa) applied in the cell....
TRANSCRIPT
(
NSF/CEE-82007
PB82-21SS 49
FINAL REPORT
AN EVALUATION OF LABORATORY TESTINGTECHNIQUES IN SOIL DYNAMICS
National Science FoundationGrant No. PFR79-026l2
by
Adel S. Saada*
:--~-----------------'.,
CASEWeSTE:RN-RESERVEUNIVERSITV
.CI..EVELANDJ OHIO~':-:-~-'-'---------'-~----~r
INFORMATION RESOURCES \NATIONAL-SCIENCE FOUNDATION.
50272-101
REPORT DOCUMENTATION 11. REPORT NO.
PAGE NSF/CEE-82007.t. Title and Subtitle
Evaluation of Laboratory Testing Techniquesin Soil Dynamics, 1982 Final Report
/--------------- ---------- ------------7. Author(s)
A.S. Saada, G. Fries, C.C. Ker9. Performing Organization Name and Address
Case Western Reserve UniversityDepartment of Civil EngineeringCleveland, OH 44106
3. Recipient's Accession No.
5. Report Date
March 19826.
8. Performing Organization Rept. No.
10. Project/Task/Work Unit No.
11. Contract(C) or Grant(G) No.
(C)
PFR7902612(G)
14.
13. Type of Report & Period Covered
Fi nalf----------------
12. Sponsoring Organization Name and Address
S.C. Liu, CEEDirectorate for Engineering (ENG)National Science FoundationWashington, DC 20550
I----'---=-----:...-=-.:..-=--~--.::.-------- -------------------------'- -----------------115. Supplementary Notes
Submitted by: Communications Program (OPRM)National Science FoundationWashington, DC 20550
---------- ----- ------ ---------16. Abstract (Limit: 200 words)
---------------
An extensive soil testing program was conducted on clay and sand materials usingthree different devices: the standard triaxial; the N.G.I. simple shear device;and the thin long hollow cylinder. Both monotonic and cycling loadings were usedand led to results emphasizing the influence of the boundary conditions present ineach device. The use of the simple shear test to simulate the conditions that prevail during landslides or earthquakes is found to lead to erroneous results whencompared to the thin long hollow cylinder. For both static and dynamic tests, thetriaxial test is definitely an improvement over the simple shear one. The thinlong hollow cylinder is shown to be the most desirable configuration to be used inthe soils laboratory for studies related to strength and stability under both staticand earthquake situations.
1--------------------------------------------------117. Document Analysis a. Descriptors
SoilsSoi 1 testsSandsCl aysb. Identifiers/Open·Ended Terms
Ground motionTri axi al testSimple shear test
c. COSATI Field/Group
Cl ay soil sLandslidesCyclic loadsEarthquakes
A.S. Saada, /PI
Bounda ri esTesting equipmentIns trumentsShear properties
18. Availability Statement 19. Security Class (This Report) 21. No. of Pages
20. Security Class (This Page)NTIS I----------------j---------
22. Price
(See ANSI-Z39.18) See InstructIons on Reverse OPTIONAL FORM 272 (4-77)(Formerly NTIS-3S)Department of Commerce
AN EVALUATION OF LABORATORY TESTING
TECHNIQUES IN SOIL DYNAMICS
by
Adel S. Saada
This report is the condensation of two Master's degree disser-
tat ions by Messrs. G. Fries and C. Ker working under the direction of
the principal investigator. The proposed investigation has been fully
completed and significant results obtained. The report, with some edi-
torial adjustments will soon be submitted for publication in an
international journal.
The following pages contain:
1 - An abstract of the report g~v~ng a very shortsummary of the research conducted.
2 - A list of personnel involved in this research.
3 - The actual report to be published soon.
One of the master's thesis by Mr. Fries is enclosed as well as
two papers already published which have resulted from this work. The
thesis by Mr. Ker is being written and when completed in a couple of
months will be sent to the National Science Foundation. The two thesis
contain all the details and test results.
Within the framework of this grant the Principal Investigator
attended:
1 - The Symposium on Laboratory Shear Strength of Soilheld in Chicago in June, 1980, where he presentedthe state of the art on soil testing (enclosed).
} I
2 - The International Conference on Recent Advances inGeotechnical Earthquake Engineering and Soil Dynamicsheld in St. Louis in April, 1981, where he presentedtwo papers, one supported by a previous NSF grant andone supported by this grant (enclosed).
3 - The Tenth International Conference on Soil Mechanicsand Foundation Engineering held in Stockholm in June,1981, where he presented a paper summarizing a studysupported by a previous NSF grant.
We trust the National Science Foundation finds this report
satisfactory.
Respectfully submitted,
/')' / / l' . ,/,/./7' -" ..' /," l'/.. "< /' (.. /... :. \/ ' .... -/.. \,;,:' (. ,-'\
Adel S. SaadaProfessor and ChairmanDept. of Civil Engineering
, . ,I I t
ABSTRACT
An extensive soil testing program was conducted on clay and sand
materials using three different devices, namely the standard triaxial,
the N.G.I. simple shear device and the thin long hollow cylinder.
Both monotonic and cycling loadings were used and led to results that
dramatize the influence of the boundary conditions present in each
device. The use of the simple shear test to simulate the conditions
that prevail during landslides or earthquakes is found to lead to
erroneous results when compared to the thin long hollow cylinder. For
both static and dynamic tests the triaxial test is definitely an im-
provement over the simple shear one. The thin hollow cylinder is shown
to be most desirable configuration to be used in the soils laboratory
for studies related to strength and stability under both static and
earthquake situations.
ii
<
I V
r
I /
/
PERSONNEL
First Year:
Principal Investigator: Adel S. Saada
1 Graduate Student: Gerardo Fries
Student help and shop technicians
Second Year:
Principal Investigator: Adel S. Saada
1 Graduate Student: Ching-Chang Ker
Student help and shop technicians
iii
v
AN EVALUATION OF LABORATORY TESTING TECHNIQUES
IN SOIL MECHANICS
by
Adel S. Saada*, Gerardo Fries** and Ching-Chang Ker ***
A Paper
To be submitted for evaluation and possible publication in
Soils and Foundations
*Professor and Chairman, Department of Civil Engneering, CaseWestern Reserve University, Cleveland, Ohio.
**Engineer, Woodward-Clyde, Consultants, Houston, Texas.
***Graduate Student, Case Western Reserve University, Cleveland, Ohio.
Vi
INTRODUCTION
The influence of the boundary conditions on the numerical values
of the results obtained in laboratory soil tests has been the subject
of many dissertations. The object of laboratory testing is to study
the behavior of a given soil under conditions similar to those en
countered in the field and to obtain those parameters which describe
this behavior in a set of constitutive equations. In a laboratory
test the specimen is intended and generally assumed to represent a
single point in a soil medium. The validity of this assumption de
pends on the uniformity of stress and strain distributions within the
soil samples. This uniformity will depend on the configuration of
the specimen and the control and measurement of stress and strain on
its surface.
In the present study, after a discussion of the stress and strain
distributions in the simple shear device, the standard triaxial and
the thin long hollow cylinder, the results of actual tests on clay and
sand materials are given. Both static and dynamic tests were con
ducted and attempts were made to obtain equivalent conditions in the
various apparatuses.
In their State of the Art on Laboratory Strength Testing of Soils
SaadaandTownsend (20) discussed in great detail a large number of
devices under static conditions; but they did not give or compare
1
2
results obtained on the same soils simply because such data was not
available. Here this comparison is made for normally consolidated
as well as overconsolidated materials and for stress controlled as
well as strain controlled tests.
STATE OF STRESS IN THE VARIOUS DEVICES
A - The Simple Shear Device
Two types are in use: The Roscoe type with a square cross section
and rigid boundaries and the N.G.I. type with a circular section and
a reinforced membrane. The N.G.I. type, minus the reinforcement has
recently been placed inside a triaxial cell and the sample pressuri-
zed hydrostatically. The stress distribution when the specimen is
subjected to shearing stresses has been studied at length theoreti-
cally and using finite elements (20). A closed form solution was
found for the Roscoe type (13,16) with and without slippage at the
boundaries. A photoelastic study was conducted for both types (25):
Fig. 1 from Ref. 25 shows the shearing stress distribution in the
Roscoe type and Fig. 2 that in the N.G.I. type.
The interest of researchers has primarily been directed at the
nonuniform state of stress resulting from the application of shear-
ing forces to the specimen. However, it was found ( 7 ) that the
lack of uniformity in the axial (0 ) radial (0 ) and circumferentialz r
(Oe) normal stresses was just as, if not more, serious when the specimen
is subjected to axial loads alone in a wire reinforced membrane or
to axial loads in a pressurized cell. The top and bottom platens
of the simple shear device are rigid and often have ribs (or similar
arrangements) to avoid slippage. Under such conditions there is
3
4
often very little relation between the normal stresses acting on
the lateral surface of the sample in a pressurized cell and ° andr
08 within the specimen. The physical dimensions as well as the
boundary conditions prevent any sort of uniformity to prevail.
Fig. 3 shows the results of a linear elastic finite elements
analysis conducted on a specimen of standard size subjected to an
axial displacement of 0.0238 in (0.6045mm) and a lateral cell pres-
sure of 30 psi (207 Kpa) . Young's modulus is chosen to be 2844 psi
(19h09 Kpa). On this figure one notices that for an imposed vertical
displacement the distribution of the three normal stresses is a
function of Poisson's ratio V and varies from the edge to the cen-
ter. This is true on the upper and lower boundaries as well as
the central section. The average value of ° also varies with V.z
However, more important is the fact that or and 08 are indeed
quite different from the 30 psi Q07 Kpa) applied in the cell.
Actually, depending on v, those two components of the state of
stress can be in the larger part of the sample closer to ° thanz
they are to the laterally applied stress. As seen in Fig. 3 only
a very thin layer close to the lateral faces "sees" the applied
30 psi Q07 Kpa) .
Of equal interest are the radial shearing stresses l imposedrz
by the end platens to the specimen because slippage is not allowed
there. Fig. 3 shows the distribution for various values of
v. The magnitude of l due to axial normal force alone could berz
5
higher than the one due to the application of the shearing force
supposed to place the specimen in a state of simple shear. Notice
that depending on the value of V the effect of the lateral stresses
may overcome that of the vertical ones so that T changes in sign.rz
The results of a complete finite element analysis for various
height to diameter ratios and a wide range of Poisson's ratios are
given in the appendix. This analysis applies to both the simple
shear and the triaxial tests.
The discussion above refers to linear elastic materials and does
not introduce any interaction between the shearing and normal components
of the stress and strain tensors. As pointed out by Bishop et al.
(2), when the end platens are rigid a nonuniform shearing stress
distribution will result in nonuniform tendencies to expand or
contract; thus leading to nonuniform normal stresses. The end
result is a totally intractable stress distribution with extremely
high concentrations that lead to premature failure when data based
on average values are compared to that obtained in other devices.
Indeed Bjerrum and Landva (3) concluded that the undrained shear
strength of specimens consolidated in the N.G.I. apparatus was
approximately equal to 2/3 of the average value measured in corres-
ponding triaxial tests. Similar results were found by LaddandEdgers
(8).
6
Finally, Bjerrum and Landva (3) write that the constant volume
test is equivalent to an undrained test and the change in applied
vertical stress on the specimen is equivalent to the change in
pore pressure which would have occured in the specimen if the
specimen had been prevented from draining for a condition of con
stant applied vertical stress. They also write that the validity
of their assumption was demonstrated in one single truly undrained
test in which the pore pressures were measured at the base of the
specimen. The experiments conducted in this research in the N.G.I.
device show that the drop in the vertical stress in the constant
volume test was several times larger than the increase in pore
pressure in the constant vertical stress test for the same consoli
dation pressure and the same shearing strain. The results will be
shown in the coming sections.
B - The Triaxial Test
The nonuniformities of the stress distributions within the
specimen are primarily due to the effects of the friction on the
end plates. The problem of the influence of end restraint during
uniaxial or unconfined compression of cylinders has been under in
vestigation since the latter part of the nineteenth century. Ref.
20 reviews most of the studies conducted on this subject. Of parti
cular interest are the papers by Balla (1) who summarizes previous
work and brings new insight to the problem, Moore (11) who examined
7
the effects of six Poisson ratios, a height to diameter ratio of
one and two and lateral confinement, and Radhakrishnan (18) who
used finite elements to study linear and nonlinear problems for
height to diameter ratios of one and two.
A feature that is common to all the linear solutions mentioned
above is the drop in the normal contact stress as one moves towards
the center, and a high concentration onthe edges. However, as shown
in the appendix the situation becomes different as the height to
diameter ratio decreases. The concentrations at the ends remain,
followed by a drop, then a substantial increase in the stresses as
one moves to the center of the sample. This is brought about among
others by the zones of influence of the end platens interfering with
each others. More complete information can be found in the references
cited in the State of the Art of Laboratory Strength Testing of Soils
by Saada and Townsend (20).
End restraints can be minimized by using rubber sheeting and
silicone grease to obtain frictionless contacts (17) • However,
Norris (12) showed that one could obtain different stress-strain
curves for sands depending on the number of layers of greased rubber
sheeting and the size of the grains.
C - The Thin Long Hollow Cylinder
When the inner and outer pressures on a hollow cylinder are equal
the radial and circumferential normal stresses across the thickness
8
are equal. Under a superimposed axial stress the nonuniformities
in the distribution of the state of stress are due to the radial
frictional forces that the end platens generate when the specimen
is prevented from expanding or contracting. Those forces are self-
equilibrating and their influence vanishes as one moves away from
the ends. References 20 and 25 contain detailed analyses of the
conditions within the hollow cylinder. In a hollow cylinder the
thickness as well as the mean radius play an important part in the
pressure distribution. The previous references give proportions
for whichfue end effects are reduced to a minimum. If the length
is 2 and the inner and outer radius are r. and r it is recommended1 0
that
,,> 5.44 /r 2o
and nr.
1 ~ 0.65r
o
Such proportions were checked by photoelasticity and by finite element
analysis (4) as well as experimentally (9) and found quite sat is-
factory.
Ifhen the hollow cylinder is subjected to torsional stresses both
shearing stresses and strains vary across the thickness. The larger
the ratio n, the smaller the nonuniformity and the more justified
is the assumption of a uniform distribution equal to the average.
In addition, as the deformation increases one moves towards a uni-
formity of stress. A combination of axial and torsional stresses,
added to a change inthe hydrostatic stress (the same inside and
outside the cylinder ) allow one to rotate the principal stresses
9
at will as well as generate practically any stress path desired.
By keeping the ratios of axial to torsional stress constant any
inclination of principal stress can be maintained during a given
test; and changed from test to test. If this is coupled with a
synchronized change in the cell pressure the inclination may remain
constant and the mean stress varied at will. Saada and his coworkers
have used hollow cylinders to study the properties of anisotropic
clays and sands for years. Ref. 20 lists many of their studies
and describes various stress paths made possible by this particular
configuration.
Finally, it ought to be mentioned that when studying cross
anisotropic materials the hollow cylinder device is rather unique.
Because the system of stress used is axially symmetric cross aniso
tropy around the axis will be maintained even though its degree
might change due to induced strains. This is not the case for
prismatic samples, be they cubic or parallellepedic: Indeed
as soon as the planes of the prism parallel to the axis of symmetry
move a new kind of anisotropy is created, most probably of the
orthotropic type. States of pure shear stress and simple shear
strain can be studied using hollow cylinders without many of the
boundary effects that plague other devices.
D - Some Comparisons Among the Results Obtained in Various Devices
Until very recently, few systematic studies were made to compare
results obtained in the various devices. Investigators have tried
to match the stress conditions on a horizontal plane in a simple
10
shear apparatus to those on a 45 g plane in the triaxial test. It
is important to remember that all the components of the stress
tensor affect the behavior of the sample and not just what happens
on a particular plane. As mentioned previousl~ Bjerrum and Landva
(3) and Ladd and Edgers (8) found very substantial differences
between the shearing strengths measured in the simple shear device
and the equivalent triaxial tests. Seed and Peacock (23) found that
cyclic strength in simple shear were about 30 to 50 percent smaller
than comparable cyclic triaxial strengths. Park and Silver (14)
attempted to compare the shear moduli and damping ratios measured
in cyclic triaxial tests with results obtained by Thiers and Seed
(24) using cyclic simple shear. They found close agreement when
the comparison was made under conditions of equal mean effective
stress; however, it must be remembered that the mean effective
stress is not known in the simple shear device. This brings into
the picture the whole subject of normalization. While normalizing
has obvious advantages it seems to be used primarily as a curve
fitting technique. When normalizing with respect to some parameter
does not work another is used until the desired coincidence is
obtained _ In view of the results shown in Fig. 3 and in the
appendix, which value of the normal stress is one supposed to take?
Pyke (15) presented a comparison of shear moduli and damping
ratios of sand similar to that made by Park and Silver. He tried
to match the static and cyclic shear stresses on the horizontal
11
plane in the simple shear test and the 45 0 plane in the triaxial
test; realizing that there are differences in the overall states of
stress in the two tests. For example the normal stress in the simple
shear test is constant while its value on the 45 0 plane of the tri-
0daxial test varies by ± :f where 0 d is the difference in the principal
stresses. Pyke points out the difference in the loops obtained in
the two apparatuses: While that of the simple shear is sYmmetric,
the one obtained in the triaxial test is not.
Finally Lee (10) found that, while not perfect the cyclic tri-
axial test offered a reasonable compromise for obtaining dynamic
properties of saturated sand.
The only comparison involving the thin hollow cylinder subjected
to torsional stresses has been made by Saada (22); it is a part of
the study reported herein.
EQUIPMENT AND SPECIMEN PREPARATION
A - Testing Machines
The simple shear tests were conducted using the standard wire
reinforced norwegian membrane. The top and bottom caps were modified
to allow for saturation and measurement of pore water pressure. They
were fitted with thin very short blades to avoid any possible slippage.
The triaxial tests were conducted in norwegian cells with rotat-
ing bushings. Enlarged and greased caps were used for clay samples
and regular caps were used for sand samples. Positive attachment was
provided between the piston and the cap to allow for tension during
cyclic loading.
The hollow cylinder tests were conducted using a specially de-
signed cell allowing for axial and torsional loads to be applied
to the sample. A transducer can be placed both inside and outside
the cell for accurate measurements. A special bushing and piston
assembly allows for the proper balancing of the forces so that both
extension and compression can be applied to the specimen.
K consolidation in the triaxial cell and the hollow cylindero
cell was obtained by means of a servo-system (19) which applies to a
sample a vertical displacement such that it keeps the same cross sec-
tion during consolidation.
12
13
Both controlled stress static and dynamic tests on solid and
hollow cylinders were conducted using a stress controlled SPAC.
This pneumatic analog computer can be programmed to subject speci
mens to static and slow dynamic axial extension and compression
alone. and pure torsion alone)or to all combined in a synchronized
way (20). The cyclic nature of the loading is controlled by an
electro-pneumatic signal generator that can induce various wave
shapes. A sinusoisal shape was used in this investigation. The
computer and its loading frame are shown in Fig. 4.
Dense soils have a peak stress beyond which the stress strain
curve falls off to the so called residual value. To go beyond this
peak, strain controlled machines are needed. ~fhilea standard Whykam
Farrance loading machine was used for axial loading, two special
frames and a strain controlled SPAC were designed and built for
use with the hollow cylinder. In the first,a torsional displace
ment is applied to the specimen and the resulting torsional stress
measured with a transducer. The output of this transducer is
amplified or reduced, transformed to a pneumatic signal and fed to
an axial actuator to give a proportional axial stress. One can
thus twist a sample beyond the peak and have principal stresses
which maintain the desired constant inclination on the axis of
symmetry; the torsional mechanical resistance of the sample being
the controlling one. In the second,an axial deformation is applied
to the specimen and the resulting axial stress measured with a
14
transducer. The output of this transducer is amplified or reduced,
transformed to a pneumatic signal and fed to a torque actuator to
give a proportional shearing stress. Here one can go axially
beyond the peak and the axial mechanical resistance is the con-
trolling one. The two frames and their SPAC are shown in Fig. 5.
B - Materials and Specimens Preparation
Both clay and sand materials were used in this comparative study.
The clay was Edgar Plastic Kaolin which has been extensively used in
research on cohesive soil. Its liquid limit is 56.3 percent an plastic
limit is 37.5 percent. The specific gravity of the solid particles is
2.62. Slurries at 125 percent water content were one dimensionally con-
solidated in an 8 in. (203.2 mm) conso1idometer. From the resulting blocks
samples for the simple shear, the triaxial test and the hollow cylinder
were cut and stored. Before each test the specimens were trimmed to:
1.4 in. (35.6mm) diameter and 3.25in. (82.55mm) height for the triaxial
test; 3.16 in. (80. 26mm) diameter and 1 in. (25.4mm) height for the simple
shear; 2.8 in. (71. 12mm) outside diameter, 2 in. (50.8mm) inside diameter
and about 6 in. (152.4mm) height for the hollow cylinder. Those specimens were
then K consolidated in their respective devices. A back pressure of 18 psio
(124.1 Kpa) was used for triaxial and hollow cylinder specimens.
Two sands were tested. One is a well known sand called Reid
Bedford sand and the other a south american sand which will be called
Latin Sand. The Reid Bedford sand is a typically uniform fine sand
with a specific gravity of 2.65 a coefficient of unformity Cu ~ 1.52
15
and a DIO
= 0.125mm. Its minimum void ratio as given by Durham
and Townsend (5) is 0.529 and its maximum void ratio is 0.816.
Its granulometric distribution is shown in Fig. 6. The Latin
Sand contains some fine shell fragments and shows a larger than
usual volumetric compressibility under hydrostatic stresses. Its
granulometric curve is shown in Fig. 7. Samples were prepared to
a void ratio of 0.72 by vibration on a shaking table for the Reid
Bedford sand and to a void ratio of 0.68 by tamping in moist con-
dition for the Latin Sand. Saturation was obtained using vacuum
followed by percolation of CO2
and deaired water; and back pressure
in case of triaxial and hollow cylinder specimens.
c - Convent ions Used in J:-.~~_Ev_alu~~ion of the Experimer~tal Reslil t~
When comparing the tests results, the current conventions will
be followed. This means that the average shearing stress and strain
in the simple shear device will be compared to the maximum shearing
stress and strain in a triaxial compression test:
T45
= Tmax
~(O - 0 )1 31,(02 1 1,0
2 d
where 0d is the excess axial stress also called the deviator.
Y45 = Ymax
where E is the axial strain and no volume change takes place.z
The shear modulus G is the secant modulus equal to T/Y and also
Eequal to 3 where E is Young's modulus. In the hollow cylinder the
16
shearing stresses and strains due to torsion are the ones in questiC:T]
In dynamic testing, the specific damping ratios and the secant
moduli obtained from the various devices will be compared. The
specific damping ratio A is defined as the specific damping capacity
divided by 4n; the specific damping capacity being defined as the
energy loss per unit volume (represented by the area of the hysteresis
loop) divided by the maximum potential energy, ~TY or ~Gy2. The
shear modulus is given by the ratio of T from peak to peak and
y from peak to peak. This definition is necessary because in the
triaxial test the loop is not symmetric and centered at the origin.
Fig. 8 and its caption shows the various definition.
~orm21iza io~ which is the process used to bring down results
to the same common denominator will be discussed as the data is
presented.
STUDY OF CLAY SOILS IN THE VARIOUS DEVICES
A - Testing Program for Clay Soils
All the tests conducted on clay soils were stress controlled.
The following tests were conducted in the standard triaxial cell CST):
Static compression
Cyclic axial
The following tests were conducted in the hollow cylinder device (HC):
Static compression
Cyclic axial
Static torsion with constant length
Cyclic torsion with constant length
Static torsion with constant axial force
Cyclic torsion with constant axial force
The following tests were conducted in the NGI simple shear device (SS):
Static shear with constant length
Cyclic shear with constant length
Static shear with constant axial force
Cyclic shear with constant axial force
The constant length tests in the simple shear device are also
known as constant volume tests. The changes in the axial stress are
measured with a very stiff diaphragm type transducer whose displacement
is negligible.
A test in the standard triaxial or hollow cylinder has an
"equivalent" test in the simple shear device. The equivalence is
17
18
obtained by using the same effective consolidation stress, monitoring
the water content and following similar testing patterns. Once the
static strength is known, the cyclic loading is applied in percentages
of the failure load, 20, 30, 40 percent until failure occurs. 60
cycles per level of stress were used for nearly all of the tests at
about one cycle per minute. During all the tests constant recording
was made of stresses, strain and pore water pressures. Hand calcu-
lations were complemented by a computerized data acquisition system
which reduced immensely the processing time.
Three different consolidation pressures were used for the clay
soil. They are referred to as 58-18, 48-18 and 38-18. The first
number represents the cell pressure and the second the back pressure
during consolidation and just before the undrained shear process be-
gain; both in pounds per square incn Ko consolidation was obtained
through an additional axial load automatically generated by SPAC.
Whenever axial loads were involved in the static or cyclic loading
procedure the excess load producing K consolidation was released,o
equilibrium allowed then the sample tested undrained from a hydro-
static state of stress.
Pure torsion tests in the hollow cylinder be they static or
cyclic started with the axial stress which produced the Ko consoli-
dation. This stress was kept constant during the constant axial
force tests, while static or cyclic torsional stresses were applied
to the sample. On the other hand this stress constantly dropped
19
during the constant length tests, while static or cyclic torsional
stresses were applied to the sample.
Simple shear tests were conducted in a manner similar to that
used for the hollow cylinder tests. The vertical consolidation
pressure was the same as the one indicated by SPAC during the Ko
consolidation ofthe equivalent hollow cylinder test. Then depend-
ing on whether the test was a constant force (CF) or a constant
length one (eL) this stress would be kept constant, or the length
fixed and the stress allowed to drop during shear. Pore water pres-
sures were also measured during simple shear tests. This is not
adviseab1e since the flexibility of the norwegian membranes is quite
large. It was decided, however, to make the measurements in the
most popular devices.
Only static tests were conducted on one type of clay with over-
consolidation ratios of 2, 4 and 8.
The reason for conducting constant length (also called con-
stant volume) tests and constant force (also called undrained) tests
in the simple shear device was to check the validity of Bjerrum's
assumption regarding the equality of the drop in the vertical stress
in one to the increase in the pore water pressure in the other. For
comparison purposes the same type of tests were conducted on hollow
cylinders subjected to static and cyclic torsional stresses.
B - Results Obtained from Static Tests
Fig. 9 shows the results of static tests obtained with the 48-18
clay and normalized with respect to an effective normal stress 0 .n
20
The shearing stress T in the ordinate is the shearing stress on a 45~
plane in triaxial compression test on solid or hollow cylinders; and
the normal stress 0 is the effective normal stress on that plane.n
For the simple shear test and hollow cylinder subjected to torsion,
T and 0 are the shearing and normal effective stresses acting on then
horizontal plane. The shear strain y corresponds to the stresses de-
fined above. In Fig. 9 as well as in all the ones to follow, (HC)
refers to hollow cylinder, (SS) to simple shear, (ST) to standard
triaxial, (C) to compression, (CL) to constant length and (CF) to
constant force. One notices that for strains as high as five per-
cent the triaxial compression tests give the steepest curves. The
lowest strength was obtained from the simple shear test at constant
force and the largest from the hollow cylinder at constant length.
Identical results were obtained for the 58-15 and the 33-18 clays.
Fig. 10 shows the secant shear modulus G(= !) normalized withy
-respect to 0. One notices that the results of the simple shearn
test, whether they are conducted with constant length (constant
volume) or constant force (undrained) are still the lowest. The
differences are very noticeable for small strains where the simple
shear test gives results two to three times smaller than corresponding
hollow cylinder and standard triaxial tests. Here too the three con-
solidation pressures yielded similar results.
The changes in the effective vertical stresses and in the pore
water pressure which take place during constant length (constant
21
volume) and constant force (undrained) tests were recorded during
simple shear and hollow cylinder tests. Fig. 11 shows that whether
it is the simple shear test or the hollow cylinder test the pore
pressure in the undrained test is very different from the drop in
the axial effective stress in the constant volume test. The measure-
ments of the pore water pressures in the simple shear device were
certainly not accurate because of the flexibility of the reinforced
membrane, but it was decided to use the most popular device in this
investigation.
Various normalization parameters were used on the K consoli-o
dation clays (6) but none of them brought any improvement to the
data shown in Figs. 9 to 11. A normalization of the shear modulus
with respect to the undrained shear strength which is sometimes
used in practice was also found to be totally unacceptable.
c - Results Obtained From Dynamic Tests
For each level of stress a minimum of 60 cycles was applied to
the samples. The results given herein refer to the lath cycle. The
shear moduli and damping ratios were calculated as shown in Fig. 8.
In the constant force cyclic hollow cylinder tests HC(CF) only the
lrysteresis loops connected to torsion were considered. Fig. 12
shows the shear moduli obtained for the 48-18 clay plotted against
the peak to peak value of the shearing strain 2y. Similar results
were obtained for the 55-18 and the 33-18 clays. For small values
of 2y « 2 percent) the differences in the magnitudes of the moduli
22
obtained in the various devices are quite noticeable. The hollow
cylinder seems to yield the highest values and the simple shear the
lowest ones.
It is well known that for small strains the damping ratio in-
creases with strain. However, as the strain becomes larger this
ratio remains more or less constant (21). This seems to be true in
this investigation too. Figure 13 shows the data obtained in this
investigation at the tenth cycle. Again, each device gives a different
answer for the same material.
Finally, it must be noticed that the hysteresis loops kept a
reasonable symmetry with respect to the origin for the simple shear
and torsional tests on hollow cylinder. On the other hand there
was a constant shifting of the loops during axial cyclic tests since
the behavior of K consolidated clays is different in tension ando
compression (21).
D - Analysis of Results for Overconsolidated Clay
Tests on overconsolidated clays were conducted only on the
48-18 clay and for 3 overconsolidation ratios, namely 8, 4 and 2.
Only static tests were conducted; static compression in the standard
triaxial cell, pure torsion at constant length in the hollow cylin-
der device and simple shear at constant length (constant volume).
The results are similar to those obtained for K consolidated clays.o
Various normalizing procedures did not bring the results to coincidence
(6). Fig. 14 shows the value of the shear moduli versus strain for
the tests conducted.
STUDY OF SAND SOILS IN THE VARIOUS DEVICES
A - Testing Program on Sand Soils
Stress controlled tests were conducted on the Reid Bedford sand
and strain controlled tests were conducted on the latin sand.
The simple shear tests and hollow cylinder tests involving torsion
conducted on the Reid Bedford sand were identical to the ones conducted
on K consolidated clay; with K assumed equal to 0.5. The specimenso a
tested in the standard triaxial cell were consolidated under a hydro-
static stress equal to the mean stress acting on the corresponding
simple shear or hollow cylinder test. If for example a specimen at a
void ratio of 0.72 is subjected to a vertical consolidation stress of
10 psi (68.9 Kpa) in the simple shear device, the corresponding conso-
1idation effective stress in the triaxial test would be 20/3 psi. One
void ratio, 0.72 and three consolidation pressures were used for the
Reid Bedford sand. All the tests will be identified in terms of Ko
and the vertical consolidation pressure. For example (0.5-10) means
that during consolidation the vertical effective stress was 10 psi
(68.9 Kpa) and the 1attera1 was 5 (34.47 Kpa~and (1-15) means that
the consolidation was hydrostatic at an effective normal pressure of
15 psi (103.4 Kpa). This latter case applies to the standard triaxial
tests. In all three vertical consolidation pressures of 10 (68.9), 20
(137.8) and 30 psi (206.8 Kpa) were used.
In the strain controlled tests conducted on the Latin sand one
void ratio 0.68 and one consolidation pressure 6.1 psi (42 Kpa) were
23
24
used. The hollow cylinders were all tested under a hydrostatic
consolidation pressure of 6.1 psi (42 Kpa). In the simple shear test
two vertical consolidation pressures were used namely 6.1 (42) and 9.1
psi (62.74 Kpa). In the standard triaxial test a hydrostatic pressure
of 6.1 psi (42 Kpa) was used. For dynamic tests the amplitudes of
the shearing strains were 0.15 percent and 0.34 percent.
B - Results Obtained From Static Tests on the Reid Bedford Sand
Fig. 15 shows the results of static tests obtained with the (0.5-
30) sand normalized with respect to the effective normal stress 0 asn
was done for the K consolidated clay. On the same figure the curveo
obtained with the (1-20) sand is also shown. From those curves one
can see that depending on the type of test one obtains different stress-
strain curves; and that normalization with respect to a mean or a
normal effective stress does not bring them into coincidence. The
differences are quite pronounced for values of strain less than 5 per-
cent. Similar results were obtained with the (0.5-10) and (0.5-20)
sands (27).
Fig. 16 shows the secant shear modulus G for the (0.5-20) sand.
The results vary from device to device with the simple shear test at
constant force showing the smallest G at strains below 5 percent.
Similar results were obtained for the (0.5-10) and (0.5-20) sands.
As in the case of clays the change in the effective vertical
stresses and in the pore water pressure which take place during con-
stant length (constant volume) and constant force (undrained) were
25
recorded during simple shear and hollow cylinder tests. While one
would expect that a drop in the vertical effective stress of a constant
length test would correspong to "at least some increase" of pore water
pressure in an undrained test, as happened for clays, the opposite
took place in the hollow cylinder. Fig. 17 shows the drop in vertical
effective stress of the constant length hollow cylinder test and the
"decrease" in pore water pressure in the undrained test. Fig. 20 shows
the results recorded for the simple shear tests: There the trends are
what they are expected to be but the magnitudes are totally different.
Part of the above is, of course, due to the free 1attera1 movement that
can take place in the hollow cylinder and the "relative" rigidity of
the N.G.I. reinforced membrane. Indeed Fig. 21 shows the measured
volumetric strains in the constant length hollow cylinder and the so
called constant volume simple shear test.
Similar results were obtained with the (0.5-10) and (0.5-20)
sands.
c - Results Obtained From Dynamic Tests on Reid Bedord Sand
Fig. 20 shows the results obtained from dynamic tests on the (0.5
30) sand for the 10th cycle. Here too the results obtained with the
(1-20) sand are shown; and again depending on the type of the test the
moduli can vary very widely.
The damping ratios for the 10th cycle are shown in Fig. 21 and
they vary from test to test. Similar results were obtained with the
(0.5-10) and (0.5-20) sands.
26
D - Results Obtained From Tests on Latin Sand
Fig. 22 shows the shear moduli versus the number of cycles for
shearing strains y = 0.15% and 0.34% respectively. Here again, for
the smaller number of cycles the modu1ii vary very widely depending
on the cycle number and consequently on the degradation of the reacting
stress: The triaxial and hollow cylinder tests give the higher magni
tudes of shear moduli when compared to the simple shear tests.
The damping ratios as a function of the number of cycles are shown
in Fig. 23. The simple shear constant force test is once more at the
bottom of the set of curves.
CONCLUSIONS
The analysis of the states of stress in the most popular testing
devices in soil mechanics has shown that the boundary conditions affect
the stress distribution in a very pronounced way. The extensive testing
program conducted on both clay and sand soils supports this analysis.
The standard triaxial test is still the favorite when it comes to static
or dynamic testing in axial extension or compression; but extrapolations
made to simulate shearing conditions leave a lot to be desired.
The simple shear test originally conceived to simulate soil condi-
tions during landslides and earthquakes underestimates the soil proper-
ties by very substantial amounts. The strength, the dynamic and static
shear moduli, the damping ratios it yields are consistently a small
fraction of their true values as measured in the thin long hollow
cylinder. This test underpredicts every parameter it is supposed to
measure. The contention that the drop of the vertical stress is a so
called constant volume test is equal to the increase in porewater
pressure in an undrained test has absolutely nothing to support it either
theoretically or, as shown in this study, experimentally. The volume
in this constant volume test is not constant at all and the reinforced
membrane has substantial flexibility. Using a standard membrane and
placing the device in a cell does not result in any improvement. since,/
as shown by the finite element analysis the sample in general does not
see the applied latteral stresses beyond a thin skin. No calibration
constant could be found to justify the continued use of this device in
27
28
research or commercial laboratories to study either static or dynamic
problems. As stated by Saada and Townsend (20) its proper place is
in the hands of designers who have calibrated their thinking in terms
of the results of those tests and have successfully applied those re
sults in their practice.
In the thin long hollow cylinder the natural cross anisotropy is
maintained and shearing stresses can be applied to horizontal planes
without resulting in the kind of stress concentrations and noninformities
present in the simple shear device. It does require a special cell
which)although not as simple as that of the standard triaxial test,is
much simpler than the ones required for prismatic elements.
Making a hollow cylinder of sand is not more complicated than
making a solid cylinder, once the cell is made. Any clay that can be
cut with a piano wire can provide a hollow cylindrical specimen. It
is obvious that many natural clays cannot be easily trimmed; but the
same critisim holds when one is trying to prepare a standard triaxial
specimen. The thinner the specimen the closer to uniform are the
shearing stresses and strains.
In conclusion it is felt that the thin long hollow cylinder with
the same inner and outer pressues and subjected to axial and torsional
stresses is the most reliabile and most versatile tool available today
to test soil. Phenomena of cyclic mobility and liquefaction can be
studied with it far more accurately than any laboratory tool of its kind.
It is recommended that it be adopted for standard laboratory investi
gations of earthquake related problems.
REFERENCES
1. Balla, A., "Stress Conditions in Triaxial Compression", Journalof the Soil Mechanics and Foundation Division, ASCE, Vol. 86,No. SM6, Dec., 1960.
2. Bishop, A.W., Green, G.E., Garga, V.K. and Brown, J.D., "A NewRing Shear Apparatus and its Application To the Measurement ofResidual Strength", Geotechnique, Vol. 21, No.4, pp. 273-328, 1971.
3. Bjerrum, L. and Landva, A., "Direct Simple-Shear Tests on a NorwegianQuick-Clay", Geotechnique, Vol. XVI, No.1, 1966, pp. 1-20.
4. DeNatale, J.S., Shen, C.K., Herrmann, L.R. and Vrymoed, J.L.,"Analysis of Stress Distribution in Torsional Shear Testing", Proceedings of the International Conference on Recent Advances inGeotechnical Earthquake Engineering and Soil Dynamics Vol. 1, pp.145-148, St. Louis, Missouri, 1981.
5. Durham, G.N. and Townsend, F.C., "Effect of Relative Density on theLiquefaction Susceptibility of a Fine Sand Under Controlled-StressLoading", Evaluation of Relative Density and its Role in Geotechnical Projects Involving Cohesionless Soils, ASTM, STP523, pp. 319-331,1973.
6. Fries, G., "Strength Characteristics of Clays Tested in DifferentLaboratory Devices", Thesis submitted to the faculty of Case WesternReserve University in partial fulfillment of the degree of Master ofScience in Civil Engineering, 1981.
7. Ker, C.C., "Strength Characteristics of Sands Tested in DifferentLaboratory Devices", Thesis submitted to the faculty of Case WesternReserve University in partial fulfillment of the degree of Master ofScience in Civil Engineering, 1982.
8. Ladd, C.C. and Edgers, L., "Consolidated-Undrained Direct-SimpleShear Tests on Saturated Clays", Massachusetts Insitute of Technology,Research Report No. 72-82, Soils Publication 284, July, 1972.
9. Lade, P.V., "Torsion Shear Apparatus for Soil Testing", LaboratoryShear Strength of Soil, ASTM, STP 740, pp. 145-163, 1981.
10. Lee, K.L., "Fundamental Considerations for Cyclic Triaxial Tests onSaturated Sand", BOSS 76, Behavior of Offshore Structure, The NorwegianInstitute of Technology, pp. 1-19, 1976.
11. Moore, W.M., "Effect of Variations in Poisson's Ratio on Soil TriaxialTesting", Highway Research Record No. 108, pp. 19-25, 1965.
29
30
12. Norris, G.M., "Effect of End Membrane Thickness on the Strenghtof Frictionless Cap and Base Tests", Laboratory Shear Strengthof Soil, ASTM, STP 740, pp. 303-314, 1981.
13. Prevost, J.B. and Boeg, K., "Reanalysis of Simple Shear SoilTesting", Canadian Geotechnical Journal, Vol. 13, No.4, 1976,pp. 418-429.
14. Park, T.K. and Silver, M.L., Dynamic Triaxial and Simple ShearBehavior of Sand", Journal of the Soil Mechanics and FoundationsDivision, ASCE, Vol. 101, No. GT6, pp. 513-529, June, 1975.
15. Pyke, R.M., "Settlement and Liquefaction of Sands Under Multidirectional Loading, Thesis, University of California, 1973.
16. Roscoe, K.B., "An Apparatus for the Application of Simple Shear toSoil Samples", Proceedings, Third International Conference on SoilMechanics and Foundation Engineering, Vol. 1, pp. 186-191, Zurich,1953.
17. Rowe, P.W. and Barden, L., "Importance of Free Ends in TriaxialTestings", Journal of the Soil Mechanics and Foundations Division,ASCE, Vol. 90, No. SMl, pp. 1-27, 1964.
18. Radhakrishman, N., "Analysis of Stress and Strain Distributionsin Triaxial Tests Using the Method of Finite Elements", TechnicalReport S-73-4, U.S. Army Engineer Waterways Experiment Station,Vicksburg, Mississippi, 1973.
19. Saada, A.S., "One Dimensional Consolidation in Triaxial Cell",Journal of the Soil Mechanics and Foundations Division, ASCE,Vol. 96, No. SM3, May, 1970.
20. Saada, A.S. and Townsend, F.C., "State of the Art: LaboratoryStrength Testing of Soils", Laboratory Shear Strength of Soil,ASTM, STP740, pp. 7-77, 1981.
21. Saada, A.S. and Shook, L.P., "The Behavior of Clays Subjected toSlow Cyclic Loading", Proceedings of the International Conferenceon Recent Advances in Geotechnical Earthquake Engineering and SoilDynamics, Vol. 1, pp. 369-376, St.Louis, Missouri, 1981.
22. Saada, A.S., A Discussion on the "Strength of Soils", Proceedingsof the International Conference on Recent Advances in GeotechnicalEarthquake Engineering and Soil Dynamics, Vol. 3, pp. 911-913,St. Louis, Missouri, 1981
31
23. Seed, H.B. and Peacock, W.H., "Test Procedures for Measuring SoilLiquefaction Characteristics, "Journal of the Soil Mechanics andFoundation Division, ASCE, Vol. 97, No. SM8, pp. 1099-1119, August,1971.
24. Thiers, G.R. and Seed, H.B., "Strength and Stress-Strain Characteristics of Clays Subjected to Seismic Loading Conditions", ASTM,STP450, pp. 3-56, 1969.
25. Wright, D.K., Gilbert, P.A., Saada, A.S., "Shear Devices for Determining Dynamic Soil Properties", Proceedings of the GeotechnicalEngineering Division Specialty Conference on Earthquake Engineeringand Soil Dynamics, ASCE, Pasadena, pp. 1056-1075, June, 1978.
!.=oa
0.6
1.0
0.2
1.2
0.4
5
4-
,-~\'- 0.8'I
. ~
.
X
;z:
~-><--
--2a~/.36~
tJ.l
~
, .00.5o-0.5o ' ! ! ! ;
-, .0
xa
Figure 1. Shear Stress Distribution in the Roscoe Simple Shear Device.
1.00.'5a
~.--~' ............. z~ " ........... £R=O.90
\" 1:.. - """'-' •.\ '-LR-O.60
", ~ ~ .---
-0.'5
FINITEELEMENT
1.4
1.2
I .0
0.8
>- 0.6~l~
0.4
0.2
0-1.0
x
'(
..
a... 1/ '1>
'00')7
/'" I ....'" .9~ ~.6
L t= l·~
~
\~
xR
Figure 2. Shear Stress Distribution in the NGI Device.
11 q;v11 ~v H/D-1/3 H/D-1/32 2 0.499 184.3001 0.499 184.3176 0; . 30 psi 0.45 135.9994 C; • 30 psi2 0.45 135.9994 0.25 81. 9463 0.25 81.946 0.10 70.62034 0.10 70.6203
00 · 1 .2 .3 .4 .5 .6 .7 .8 .9 00 . 1 .2 .3 .4 .5 .6 .7 .8 .9I"/R I"/R
2 11 ~v H/D=1/3 2 IJ ~v H/D=1/30.499 184.3176 OJ = 30 psi
0.499 184.300 OJ • 30 psi
0.45 135.9994 0.45 135.99940.25 81.946 0.25 81. 9460.10 70.6203 0.10 70.6203
.. ..I 4 I
c1r-1c1av
• 1 .2 .3 .4 .5I"/R
.6 .7 .8 .9 • 1 .2 .3 .4 .5I"/R
.6 .7 .8 .9
2 lJ 6...... H/D=1/32 IJ C{.V H/D=1/3
0.499 184.3176 0; • 30 psi0.499 184.3176 0; . 30 psi
0.45 135.9994 0.45 135.99940.25 81. 946 0.25 Hl. 9460.10 70.6203 0.10 70.6203
::: :~ :~: 0;.~1 -1
dv ~v
3 • ....• 3
.00
· 1 .2 .3 .4 .5 .6 .7 .8 .9 00 . 1 .2 .3 .4 .5 .6 .7 .8 .9I"/R
I"/R
TOP .6 11 ~vH/D=1/3 CENTER
.5 0.499 184.3176 "3 • 30 psi0.45 135.99940.25 81.946
.4 0.10 70.6203
.3
.2Trz • 1~v
11'-
-. 1-.2-.3
-.40 • 1 .2 .3 .4 .5 .6 .7 .8 .9
r/R
Figure 3. Results of Finite Elements Analysis for Different \).
~4-<.-" !
•
•
oC I 19
I.. ~
Ij
•
\.l-. I
Figure 4. Controlled-Stress Pneumatic Analog Computer SPAC.
35
Figure 5. Controlled Deformations Pneumatic Analog Computer SPAC.
36
100
80
...Q) 60c:
c:Q)()...Q)
a.. 40
20
o
Gravel Sand Fines
Coarse to Fine Silt Claymedium
U.S. standard sieve sizesI I
0 f3 8 ~.!: '<t ~ Sf ~
'<t ci ci ci ci ci ci..... zC') z z z z ZI I I I
f1~
I1I I !I
"I
I II11\ I 'I
I II I I
I I I II
I! I II I I II 1 1 I 11I I ! I I!I I I I II I I I II II
II II
I Ii 1 I II I I I
III I I I IiI I I I II I I ! III [l 1
I III Ii I III I 1\1I I I
~I
I I! I !IiI I I I 1
0 <0 ~:;r~
O>~ '<t (; lI'l
8~ ": :!o r-- 8co '<t 0 c:i'<t c:i c:i c:i c:i c:i c:i
Grain diameter, mm
Figure 6. Granulometric Distribution of Reid Bedford Sand.
37
gci
..,8ci
oci
Gravel Sand Fines
Coarse to Fine Silt Claymedium
U.S. standard sieve sizesI I
= 0 ~ 0 ~ 008 &'Ot ..... ,.., ..,.,<.0 .....'Ot 0 0 0 .. o 0 <:> 0 0M z z z z z Z. 2.% Z
J. I t[T '"
-, III I r-.. I !I II II I III d I I!1 I I \
1"I,! II I I I
I I T III I I \ I I!I I I \ I II I I I II II
II I
I d I \t II I
~II
I I I I!I I I I II !I I I II 11 I1 I I
I 1\I II I I I I
~'"I ,! I I I! ,.......I
0 I T I
20
80
100
...Q) 60c--cQ)o...Q)
a. 40
Grain diameter, mm
Figure 7. Granulometric Distribution of Latin Sand
3d"
21
1 (or 0)
1max
YmaxI
1
II
1 . Im1.nI I
I
-'J'k"'------- 2y ----------'1.'\,..-
y (or E)
1 - 1max min
1 =2
1 - 1
Gmax min 1-
Y - Y Ymax - Y Ymax min minY 2
(4n) (~ 1Y)
Where lI'0
6'0(4n)(~ Gy2) = Specific Damping Ratio
Total Energy Lost per Unit Volume,and represented by the Area of theHysteresis Loop.
Figure 8. Hysteresis Loop and Definitions of Specific DampingRatio (A) and Secant Modulus (G).
~.~
L LQ rrJE' OJo .C
U (J)___. 06 1 I
OJC CC11 0 . g
23
------lI
JHC(CFl--
0: )
---1 _i.0
Sh e a l' S t r ad n
(48-18) STATIC .- HOLLOH CYLINfJLR
STRNDRR.D TRIRXIAL , 8- SIMPLE SHEAR.
• HC (C)~__ z-< . -~_- . ST(C)
~ ~",S(CL)ti
· 1
.2
.5
· 3
· 7
.8
.G
0-(!)
s.-o
rn rOJ ---"
"TI11l
(!)
-{.I' ~c: L
\.:)
o +JI tJl
(!)
(1ill +JL
L+J OJ(fj>
'I-\1-
'+- '+-W W'. "-(1 C1
(!)C1
QJ OJ
LL
+J +JUJ (Jj
LL
I" "QJ a)-t;..c(Jj) (fJ
Figure 9. Normalized Shear Stress vs. Shear StrainCurves from Ko-consolidated Clay, (48-18) series
------_.--,
2.00 rc-----
5
d
<f:3I,I
._----~
-~~-=::-=:-.-
z
Shea,- Strai.n c-n
(48-18) S"fRTIC- HOLLOW CYL..INDt.R
STANDARD TRIRXIAL . ~ SIMPLE SH~RR
II
Sv ,_. ; Iljl 1
(3
100
~L
L roQ.. vE ...c() lf)
U'-' 00
(lJ c::c 0(Ij .-- (JlQ r...
0-ן"
m .....QJ
-0VIIf)Q!
'" (lJ
LC
+'I a (J1-+-(I')
"l +'QJI LL.
<1J~ >(fl
" q-'+ 4-4- i..JLJ """-au (')0 'L:L
L'- irJttl
II)V'...c~U1(fI
Figure 10_ Normalized Shear Modulus vs. Shear Strainfrom Ko-consolidated Clay, (48-18) series
(48-18) STATIC SIMPLE SHEAR
2010
Vl....
VV
V
V/
VI
/ -pp
j -~ I--D
D
10
20
30
40
50
~
~.::
e ..o atL c:o •
io...J
• •a.. ••
• La. ..II)
c:- ...e ~• •II>eL •o ...c: ...
1"41&1
....• •c: c:88
Shear S~rafn (~)
(18-18) STATIC TORSION HOLLOW CYLINDER
2010
V J....-~
~~
~~--- I--'
l.-- t---I--
/,..."....
./
V".,.,
//
/
It ~~
V ~II>
oo
10
20
40
30
50
• •a.. ••• La.. ..(J)
c:- ...e L• •II>•L •0 ...c: ...
1"41&1
· .....• •c: c:88.., ..,
....~.::·..o at~ c:o •io...J
c:- ~...Q.(J)o a..~ ..,
Q..
Shea,. S~rafn (~)
Figure 11. Comparison of the Increase in Pore Pressure to the Drop inEffective Vertical Stress in Simple Shear and Hollow Cylinder Tests. (1 psi = 6.9 Kpa)
(48-18) DYN8MICSTANDARD TRIRXIRL
,.
12000
11000
10000
;::; 9000(J)n. 8000....,
~7000
~ I:; 6000~
"00 50130L:
L 40001lJ0
.c. 3000Ul
2000
1000
00 1 2
HOLLOW CYL!NDER~ SIMPLE SHEAR
SS(CL)
3 i:l 5
Shear Stra1n (~) PEAK-TO-PEAK
Figure 12. Shear Modulus Corresponding to 10th Cycle vs. Shear StrainKo-consolidated Clay, (48-18) series. (1 psi = 6.9 Kpa)
~
(48-19) DYNAMICSTANDARD TRIRXIAL ,
HOLLOW CYLINDER •&. SIMPLE SHE~R
30. •
"I !P>'
'......., 20
ST(AX),
0 I.... / I.LV,v.L}.. ~
III
•
...J:. I':..-4:: c-Q..E 10III
Q
54321oI. I
o
She ar St r a' n 00 PEAK-TO-PEAK
Figure 13. Specific Damping Ratio (10th Cycle) vs. Shear StrainKo-conso1idated Clay, (48-18) series.
(48'-18) STATIC .- O\/[R-CONSOLIDATED CLAYHOLLOW CYLINDER, STANDARD TRIAXIAL. & SIMPLE SHEAR
5000 r- I
5421
~~~~_'w, __n ~
HC(CL).OCR-ti ST(C) ~SS(~ ~ '" "" .~o I v,",,&.'\.-v V\JI.'-U Uy \ y ..... I 'VY1.),-" ,ll\l \ \/J.,f I • vy,n y ,
o
She at" St r<l i '1 on
Figure 14. Shear Moduli vs. Shear Strain Overconsolidated Clay, OCR = 2.4 and 8. (1 psi = 6.9 Kpa)
,.. ,..L. L.Q."E aoocUUl\J
aalSc., c 1-0Q.-
" .9• L.moa .... • 8"t7\J
In " .7V'"a.Sl 1/c L. 7
O~
Ul
~ " .5" .",-."
.,~
L. L.+' a .4Ul>
.3ct-ct-ct-ct-a a .2
"" "" " • 1a aL. L.+'~ 0UlUl 0L. L.., .,a I).s:oC(n (n
(0.5-30) STATIC - HOLLOW CYLINDERSTANDARD TRIAXIAL & SIMPLE SHEAR
LegendNO. TYPE
1- HC(CF)2. HC(CL)3. ST(HY)
5 l~.• ST(Ko)
4- 5. SS(CF)
~
46. SS(CL)
7 7. . HC (C)2
10
SHEAR STRAIN (Y.)
Figure 15. Normalized Stress-Strain Curves for Various Devices.:
20
(0.5-30) STATIC - HOLLOW CYLINDERSTANDARD TRIAXIAL & SIMPLE SHEAR
20
6
10
8000
,....... 7000~Q...
'-' 6000tn:3 5000:::J~
~ ~ 4000
-....J~ 3000a:
~2000~~ ---- 31000 .5- ~ 7
1
SHEAR STRAIN (~)
Figure 16. Shear Moud1i for Various Devices. (lpsi = 6.9 Kpa)
(0.5-30) STATIC - HOLLOW CYLINDER
1098765
PP HC(CF)
4
HC(CL)
321
13 ....<? I I I I tit I I I
113
-213
-113
-313 ..o
c-,...-0. ~l
o Q.~ .....
(:I
olS ,..."'.cGl+OII en"- t:o III
Lt..-J
· ·........Co' Ii)C Co 0
r-.. UU""'t- ..... 'V
~ CIl
• (IIa. QJ· "-0.+0If)
c:- ·+'G) ...III G)Ill>III... ·O~
Ct+-..... W
SHEAR STRAIN (%)
Figure 17. Comparison of the Increase in Pore Pressure to the Drop ofEffective Vertical Stress in the Hollow Cylinder Tests.
(0.5-30) STATIC - SIMPLE SHEAR
e~""-o.~
00. 10L. ......
Cl SIoc'S 8""",I;;:.I) .... 7omL. e 6a mu...-l
5· ·........ 4~ ~
e e 3a a-\:. uu
'-J '-J 2~
~· ~ 1Q..m· "-n. .... 0(/)
c: -1·.... -2I) L.
" m«1> -3I)L. · -40 ....e ....HW -5
0 1 2 3 4 5
SS(CL)
6 7 8 9 10 11 12 13 14 15
SHEAR STRAIN (Yo)
Figure 18. Comparison of the Increase in Pore Pressure to the Drop ofEffective Vertical Stress in the Simple Shear Device.
(0.5-30) STATIC - HOLLOW CYLINDER & SIMPLE SHEAR
\r,~
.12112
.12111
.01
.I2II21S
.12108
.01217lSI'oJ
> .1211216"-,..H .12105'oJ
"'0.1211214>
.1211213
.1211212
.01211
121
l12I 1 2 3 4 5 s 7 8 9 1121
SHEAR STRAIN (%)
Figure 19. Volume Changes in the So Called "Constant Volume" Tests.
(0.5-30) DYNAMIC - HOLLOW CYLINDERSTRNDARD TRIRXIAL & SIMPLE SHEAR
5
5
4
LegendNO. TYPE
l. HC (cF)2. HC(CL)3. ST(HY)4. ST(Ko)5. SS(CF)6. SS(CL)
632
~~_ 9.1
12000 r73
\11000 ' ,
,..., 10000-II) 90000..-.J
(f) 8000:J...J 7000:Jr::l
~0 6000E-- 0:: 5000a:w 4000:r:(J1
3000
2000
1000
00
SHERR STRAIN (Yo) peak to peak
Figure 20. Dynamic Shear Moduli at the 10th Cycle forVarious Devices. (1 psi = 6.9 Kpa)
(0.5-30) DYNAMIC - HOLLOW CYLINDERSTANDARD TRIAXIAL & SIMPLE SHEAR
Legend
v,~
"~...,o....ta:0:
CJZ....~a:~
40
30
20
10
1 2 3
1
NO. TYPE1. HC(CF)2. HC(CL)3. ST(HY)4. ST(Ko)5. SS(CF)6. SS(CL)
5
4 5
SHEAR STRAIN (%) peak ~o peak
Figure 21. Damping Ratio for Various Devices.
LATIN SAND (1-6.1) ~ (0.5-9.1) DYNAMIC - HOLLOW CYLINDERLegendSTANDARD TRIAXIAL ~ SIMPLE SHEAR
NO. TYPE
l. HC (cF) 1).1
Y 0.15% 2. HC (eL) 9.15
20012l 3. ST(HY)6.14. ST(Ko)9.155. SS(CF)9.15
~6. SS(CL)9.157. SS(cF)6.1-
~-:SS(CL)6.161 8.
0.. :~
2til
3 :~-l 4~00 10001:
Q!: 6et
8~til
1
01 I I I I ! I I I I I
0 10 2~ 30 413 50 60 70 80 90 100
NO. OF CYCU:S
LATIN SAND (1-5.1) ~ (8.5-S.1) DYNAMIC - HOLLOW CYLINDERSTANDARD TRIAXIRL ~ SIMPLE SHEAR
2
1200
1100
11300-~ 90130......,
Ul 800:J-l 7130::J00 600L
Q: 500a:w 400J:Ul
300
2013
100
00
s
NO. OF CYCLES
y 0.34%
• 8 •
Figure 22. Variation of the Shear Moduli With the Number of Cyclesfor Controlled Strain Tests. (1 psi ~ 6.9 Kpa)
53
LATIN SAND (1-6.1) & (13.5-9.1) DYNAMIC - HOLLOH CYLINDERSTANDARD TRIAXIAL & SIMPLE SHEAR
•
y = 0.15%313 \
2.~_:=__-=._~_~.---tR~-4 .~ • •
oHto:IX
~HIl.:f:0:Q
._------------
l-__-L--~~---.ll---..ll---l::!__~!~--'!'=_-__:_:..I::_-_::::.':---::,1Be 113 213 30 0+13 sa 60 713 80 913 11313
NO. OF CYCLES
LATIN SAND (1-6.1) & (13.5-9.1) DYNAMIC - HOLLOW CYLINDERSTANDARD TRIAXIAL & SIMPLE SHEAR
y = 0.34%
o....to:It:
ClZ....Il.:f:0:Q 10
7~
• •5
a I I I Ia 40 513 613 79 80 S0 109
NO. Of CYCLES
Figure 23. Variation of the Damping Ratios With the Number ofCycles for Controlled Strain Tests.
APPENDIX I
Elastic finite elements analysis of
the stresses induced in a cylinder
subjected to axial deformation and
lateral pressures.
55
A linear elastic finite element analysis was made to investigate
themagnitudeof the stresses generated in short circular cylinders
subjected to axial displacement with rigid end platens and lateral
stress; no slippage being allowed between the specimens and the platens.
The use of a fine grid helped bring out the stress concentrations at
the edges (Fig. Al). Four Poisson ratios namely 0.499, 0.45, 0.25 and
0.1 were used, as well as lateral pressures of 0, 20, and 30 psi
(10, 137.9, 206.8 Kpa). For a Poisson ratio of 0.45 and no lateral
stress, the problem was solved for height to diameter ratios HID of
.167, 1/3, 1/2, 1 and 2. This last step was taken to compare the re
sults of this study with published results (1, 11, 18) as well as to
extend such results to smaller height to diameter ratios. The results
have been normalized with respect to the average vertical stress 0av
which is the total axial force divided by the area. Young's modulus
was chosen to be 2844 psi (19609 Kpa) and the average axial strain 2.38
percent. Under such conditions "if the ends are perfectly smooth and
do not have any effect on the sample" the average stress should be
67.55 psi (465.73 Kpa). This indeed is not the case and this average
stress varies with both HID and Poisson's ratio, it is written on each
of the figures so that the actual magnitudes of the elastic stresses
can be computed.
Figure A2 shows the distribution of Oz at the top and the middle
section of an unconfined specimen for a Poisson ratio of 0.45 and five
different HID ratios. The curves related to HID equal to 2 and 1 at
SG
azthe top of the sample are characterized by a ratio aav
along 70 percent of the radius followed by high stress
close to unity
concentration
along the edges. This is in line with previously published results.
However, as HID decreases the stresses increase substantially in the
central part then vary as shown in the figure. Notice that the proximty
of the curves may be misleading since each ordinate has to be multi-
plied by a different a to obtain a .av z
Figures A3 and A4 show the distributions of a , a , 0 8 and Lz r rz
for the same HID = ~ which relates to the NGI size of specimen, and
values of 0 3 equal to 0 and 20 psi (137.9 Kpa). Each figure shows four
different Poisson ratios. The common features of all these figures are
1that at a ratio HID = 3' the pressure distribution as well as the average
vertical stress are highly dependent on Poisson's ratio and vary very
substantially as one moves along the radius as well as across the thick-
ness of the sample. The values of or and 0 8 , for the higher values of
Poisson's ratio, have very little resemblance to the laterally applied
stress; in other words, except for a thin outer shell, the sample is
very highly affected by the friction of the platens. The value of Trz
is, of course, equal to zero on the middle section, but against the
platens it reaches values that can be higher than the applied shearing
stresses during a shear test; in addition, depending on the lateral
stresses and Poisson's ratio it can take positive or negative values.
The analysis above shows that it is practically impossible to
talk about some average representative value of the state of stress
57
within the simple shear device. The combination of a shearing force
and the normal displacement studied above results in a very strongly
nonuniform distribution of stress and strain in the sample. Normali
zation of test results appears to be an exercise in forced curve fitting.
\.l1(',- .......}
,----(------; - I
~ :: ~0: L._______ cf. psi
3 777777 7777 77 7/777777777/ 3
Diameter= D = 2R = 3.16 inInitial Height= H 2
~ E= 2844 psi (200 kg/em) »
I E= 2.4%
I
_,-1 211f- 1 r,nUTOP SECTION
2 LV
4221 40
3t---+_-+_-+--+--+--+--+--4--4-~-+-+-+--l--+--+--I--I--+---J '-;3'J
B/2 4141---j----t--t---t---t---t--+--1--+--l--l-+--+-HH-+-+---W
68I 34
5 61 80
t105
~ 81 gP r~IDDLE SECTION
1('-'i 1.26 <t.
,~ R .,
~
Figure AI. Finite Element Grid for Simple Shear Device
TOP
2
OJ = 0 psi
HID1 .1672 1/33 1/24 15 2
~v161. 7775115.9147
93.520878.218672.9414
324
~I
Figure A2. Distribution of Vertical Stresses In An Unconfined Condition.
60
C1<iV'154.3539115.914776.123868.3262
1/1 0.4992 0.453 0.254 0.10
4
2H/D-l/3C'J • 0 psi
C>UV154.3646115.9147
76.123868.3262
V0.4990.450.250.10
2
_3
_4
00 • 1 .2 .3 .4 .5 .6 .7 .B .913
0 • 1 .2 .5 .6 .7 .8 .9r/R r/R
2 1/ ~H/D=1/3 2 H/D"l/31/ ~
0.499 154.3646OJ • 0 psi
0.499 154.3646 "3 . 0 psi
0.45 115.9147 0.45 115.91470.25 76.1238 0.25 76.12380.10 68.3262 0.10 68.3262
=:: :~
c
~1 C>e 1dv C>dV
-••• --13
0 . 1 .2 .3 .4 .5 .6 .7 .B .9 °13 . 1 .2 .3 .4 .5 .6 .7 .B .<J -1r/R r/R
TOP CENTER
.6 1/ C>eW H/D-l/3
.5 1 0.499 154.3646 ~ • 0 psi
2 0.45 115.91473 0.25 76.1238
.4 4 0.10 68.3262
.3
.21rzC1' .1dv
13
-. 1
-.2
-.3
-.413 . 1 .2 .3 .4 .5 .6 .7 '.8 • S
r/R
00 • 1 .2 .3 .4 .5 .6 .7 .B .913
13 . 1 .4 .5 .6 .7 .8 .9
r/R r/R
V C>d\( H/D-l/3 2 V 0,;10' H/D-l/32
0.499 154.3646 OJ • 0 psi 0.499 154.3646 ~ • 0 psi
0.45 115.91470.25 76.1238
0.45 115.9147
0.10 68.32620.25 76.12380.10 68.3262
C>r C>r
d 1 -1
dv~v
dV
Figure A3. Results of Finite Element Analysis for Different V.
H/D=1/3C'j ° 20 psi
v 0CIV2 0.499 174.3359
Ir
0.45 12Q.2673 0.25 79.5674 0.10 69.9361
~, .........--...-::::'-----:~~ j, 4-dV
2V
1 0.4992 0.453 0.254 0.10
~v174.3256129.267
79.56769.9361
H/D=I/3OJ • 20 psi
""0 .2 .3 .4 .5 .6 .7 .8 .900 .5 .6 .7 .8 .9 . 1. 1 .2 .3 .4
r/Rr/R
H/D=1/3 2 II ~vH/D=1/3
2 J) q;v c.3 = 20 psi 1 0.499 174.3359OJ =20 psi
1 0.499 174.3359 2 0.45 129.2672 0.45 129.267 3 0.25 79.5673 0.25 79.567 4 0.10 69.93614 0.10 69.9361
~ ~ ~ :c5
~:6,-1
~ ~ ~ ~;1 ~v
;~~!<IV
•• •• • 3. 3 •.~4, ,
1 °0 .3 .4 .5 .6 .7 .8 .9°0 .5 .6 .7 .8 .9 . 1 .2• 1 .2 .3 .4r/Rr/R
2 J) cr.v H/D= 1/3 Eo 2844 psi
2 J)H/D= 1/3 Eo 2844 psi
1 0.499 174.3359 0; = 20 psi cr.v OJ = 20 psi2 0.45 129.267 0.499 174.32563 0.25 79.567 0.45 129.2674 0.10 69.9361 0.25 79.567
0.10 69.9361
~ :: ::5" :~0;.
Ci 1 --0:- 1,>I,~
• 3 • • • •• ....3..40
0 00
1 2• 1 .2 .3 .4 .5 .6 .7 .8 .9 . 1 .2 .3 .4 .5 .6 .7 .8 .9 'Ir/R
r/R
CENTER
.9.8.7.6.5r/R
.4.3
v ~v H/D=1/30.499 174.3359 OJ = 20 psi0.45 129.2670.25 79.5670.10 69.9361
.2• 1
.6
.5
.4
.3
.2r,.z 1
d • t~~;.l;;;~~=:'l'=::=:E=!:~=~~~~~~.av 0
-. 1
-.2
-.3
-.40
TOP
Figure A4. Results of Finite Element Analysis for Different v.