Department of Civil Engineering

Soil Mechanics Laboratory

Laboratory Exercise 9: Drained Triaxial Compression Test on

Sand
References

Holtz, R. D.

Introduction
The following discussion is taken directly from Holtz and Kovacs (198 1).
During the early history of soil mechanics, the direct shear test was the most popular shear test.
Then, about 1930, A. Casagrande while at M.I.T. began research on the development of
cylindrical compression tests in an attempt to overcome some of the serious disadvantages of the
direct shear test. Now this test, commonly called the triaxial test, is by far the more popular of
the two. The triaxial test is much more complicated than the direct shear but also much more
versatile. We can control the drainage quite well, and there is no rotation of the principal
stresses like the direct shear test. Stress concentrations still exist, but they are significantly less
than in the direct shear test. Also the failure plane can occur anywhere. An added advantage:
we can control the stress paths to failure reasonably well, which means that complex stress
paths in the field can more effectively be modeled in the laboratory with the triaxial test.

Drainage conditions or paths followed in the triaxial test are models of specific critical design
situations required for the analysis of stability in engineering practice. These are commonly
designated by a two-letter symbol. The first letter refers to what happens before shear--that is,
whether the specimen is consolidated. The second letter refers to the drainage conditions during
shear.

The test you will be performing in lab is a consolidated - drained (CD) triaxial test. This test is
typically used for stability analysis of granular soils under constant (or long term) loading
conditions, such as the stability of a retaining wall after construction.

University of Washington

Department of Civil Engineering

Soil Mechanics Laboratory

PRINCIPLES OF THE TRIAXIAL COMPRESSION TEST

The triaxial compression test is used to measure the shear strength of a soil under controlled
drainage conditions. In the conventional triaxial test, a cylindrical specimen of soil encased in a
rubber membrane is placed in a triaxial compression chamber, subjected to a confining fluid
pressure, and then loaded axially to failure. Connections at the ends of the specimen permit
controlled drainage of pore water from the specimen. The test is called "triaxial" because the
three principal stresses are assumed to be known and are controlled. Prior to shear, the three
principal stresses are equal to the chamber fluid pressure. During shear, the major principal
stress, 1 is equal to the applied axial stress (P/A) plus the chamber pressure, 3. The applied
axial stress, 1 - 3 is termed the "principal stress difference" or sometimes the "deviator stress".
The intermediate principal stress, 2 and the minor principal stress, 3 are identical in the test,
and are equal to the confining or chamber pressure hereafter referred to as 3.

TYPFS OF TESTS
There are 3 types of tests:

1. Unconsolidated--undrained test which is also called the quick test (abbreviations commonly
used are UU and Q test). This test is performed with the drain valve closed for all phases
of the test. Axial loading is commenced immediately after the chamber pressure 3 is
stabilized.
2. Consolidated-undrained test, also termed consolidated-quick test or R test (abbreviated
CU or R).

In this test, drainage or consolidation is allowed to take place during the

application of the confining pressure 3. Loading does not commence until the sample
ceases to drain (or consolidate). The axial load is then applied to the specimen, with no
attempt made to control the formation of excess pore pressure. For this test, the drain valve
is closed during axial loading, and excess pore pressures can be measured.
3. Consolidated-drained test, also called slow test (abbreviated CD or S). In this test, the
drain valve is opened and is left open for the duration of the test, with complete sample
drainage prior to application of the vertical load. The load is applied at such a slow strain
rate that particle readjustments in the specimen do not induce any excess pore pressure.
2

University of Washington

Department of Civil Engineering

Soil Mechanics Laboratory

Since there is no excess pore pressure total stresses will equal effective stresses. Also the
volume change of the sample during shear can be measured.

University of Washington

Department of Civil Engineering

Soil Mechanics Laboratory

COMPARISON OF THE TRIAXIAL AIND THE DIRECT SHEAR TEST

The advantages of the tria xial test over the direct shear test are:

Progressive effects are less in the triaxial.

The measurement of specimen volume changes are more accurate in the triaxial.

The complete state of stress is assumed to be known at all stages during the triaxial test,
whereas only the stresses at failure are known in the direct shear test.

The triaxial machine is more adaptable to special requirements.

The advantages of the direct shear test are:

Direct shear machine is simpler and faster to operate.

A thinner soil sample is used in the direct shear test, thus facilitating drainage of the pore
water from a saturated specimen.

See Section 10.5 in Holtz and Kovacs (1981) for additional comments about these types of tests.

PROCEDURE
You will conduct a CD test on sand. Soil specimens will be loaded to failure under 3 different
confining pressures (one confining pressure and thus one specimen and test per laboratory
group). Failure will be defined as the peak or 3 maximum value of principal stress difference
reached. Steps numbers 1 through 19 will be done by the instructor prior to class.

1. Obtain the thickness of the membrane. This thickness is best obtained by measuring the
membrane doubled and then halving the measurement.
2. Saturate the sample drainage line to the lower platen then close the valve for this line.
3. Place a saturated porous stone on the bottom platen.
4. Attach a rubber membrane of the proper diameter to the bottom platen with rubber "O"
rings. The membrane should overlap the platen at least 1/2".
5. Weigh to 0.1 g a dish with the dry soil which is to be tested.
6. Place a specimen mold around the rubber membrane. Fold the top portion of the membrane
down over the mold, taking care that the membrane is not twisted or pinched.
4

University of Washington

Department of Civil Engineering

Soil Mechanics Laboratory

7. Apply a vacuum to pull the membrane against the side of the mold.
8. Place the sand in the membrane and mold by tamping each spoonful of soil, taking care not
to pinch the membrane with the tamper. Scarify the top of each layer before placing the
next one, to reduce stratification. The amount of tamping depends on the denseness of soil
desired.
9. Again weigh the dish of soil. The difference in masses is the mass of soil used.
10. Place a porous stone and then the top platen onto the sand. It may also be necessary to
coat the outer rim of the top pla ten with silicone grease to make a good, leakproof seal at the
top. Roll the membrane off the mold and onto the top platen and seal it to the platen with
rubber "O-ring". Take a small level and level the top platen.
11. Release the vacuum to the mold.
12. With the drainage line to the lower platen still closed, attach the line from the top platen to
the vacuum and apply a vacuum of about 30 kPa.
13. Now remove the specimen mold and observe the membrane for holes and obvious leaks. If
any are found, the sample must be rebuilt using a new membrane.
14. Obtain four height measurements approximately 90 apart and use the average value for the
initial specimen height L0. Take two diameter readings 90 apart at the top, at midheight,
and at the base using a pair of calipers. Take these measurements to the nearest 0.1 cm.
Compute the average diameter of the specimen at each height location and then compute a
final average specimen diameter as
davg = (dt + 2dm + db)/4

where dt is the average diameter based on the two top measurements and taking into
account membrane thickness, etc. Compute the corresponding value of initial sample area
A0 using the average diameter just computed.
15. Place the lucite cylinder on the cell base. Be sure the base is free of soil grains so that an
airtight seal can be obtained. Place the cell in the compression machine.
16. Fill the chamber with water through the cell chamber line with the bleed off valve at the top
of the chamber open.

University of Washington

Department of Civil Engineering

Soil Mechanics Laboratory

17. Attach pressure line to the cell chamber line and slowly apply a cell pressure of 20-30 kPa
while slowly decreasing the vacuum. Disconnect the top specimen drainage line from the
vacuum and connect to a burette.
18. To saturate the specimen, slowly flow water through the bottom specimen drainage line until
the sample and the top specimen drainage line are saturated.
19. The water level in the burette should be brought to a reading about the midpoint of its scale
in order that either expansions or contractions can be measured. (Water can be added to the
burette during the test if necessary.)
20. Raise the chamber pressure to 50 kPa.
21. When the chamber pressure is applied, it can be seen that the load dial deflects slightly due
to the difference between the weight of the load piston and the upward chamber pressure on
the piston base. Carefully lower the piston until the load dial just flickers as recontact of the
piston and soil specimen is made. Now set the load dial to zero to approximately tare the
system for the weight of the load piston and the pressure acting on its base so that the load
dial readings are the actual (stress difference) loads being applied to the soil specimen.
22. Attach a deformation dial (reading to 0.001 in. or 0.0025mm) to the piston so that the
deformation of the sample can be obtained. Set the dial gauge to zero; then very carefully
manually compress and release the dial plunger several times and observe the zero reading.
Readjust the dial gauge to zero if necessary. Recheck the load dial for zero reading.
Recheck the chamber-pressure gauge.
23. Set the compression machine to the desired strain rate (between 0.02 and 0.1 mm/min, as
specified by the instructor).
24. Turn on the compression machine. For the first 2% of axial strain, take a set of readings
about every 0.2% of strain. For the rest of the test, take readings every 0.5% to 1% strain.
Time observations need only be made every third or fourth set of regular readings. Load
dial, deformation dial, and burette readings must be taken at each regular reading.
25. Continue the test until an axial strain of about 20% is reached.
Note: It is possible to run a multi-stage triaxial test (a test at more than one confining
pressure). This saves time in specimen and apparatus setup, etc, and eliminates some of the
variability associated with obtaining a failure envelope with three different specimens. If a
multi-stage test is to be run, the steps are the same until this point (step 25). At this point,
6

University of Washington

Department of Civil Engineering

Soil Mechanics Laboratory

the test is carried out not until 20% strain, but until the compressive force remains constant
for two readings. At this point, strain is stopped, and the cell pressure is raised to a higher
level (the second desired cell pressure). Step 24 is then repeated until, once again, two
consecutive load readings are the same. At this point, the cell pressure is raised once again,
and the specimen sheared. For the final cell pressure applied, the test should be carried out
to approximately 20% strain.
26. Apply a vacuum of 50 + kPa through the top specimen drainage line, at the same time you
lower the cell pressure.
27. Drain the chamber water.
28. Disassemble the apparatus (removing the chamber nuts one-half to one revolution at a time).
29. Sketch the failed specimen.

On the sketch, dimension the maximum and minimum

diameters, the length of the specimen, and the angle of inclination of the failure plane, if
there is one.
30. Release the vacuum and remove the specimen.

Calculations
1. Compute the unit strain from the deformation-dial readings as
ax = L/L0
and fill in the appropriate column of the data sheet.

Also compute the adjusted

instantaneous area
A = Ac[(1 - V/Vc) / (1- ax)]
and place this in the appropriate column of the data sheet. The c subscript above refers to
the fact that these should be the dimensions not at the start of the test, but after
consolidation of the specimen.
2. Compute the axial load using the load-dial readings. If a load (proving) ring is used, the load
P is
P = DR x load-ring constant
where DR is the load-dial reading in units of deflection.
column of the data sheet.
3. Compute the principal stress difference
(1 - 3) = P/A
7

Put these data in the appropriate

University of Washington

Department of Civil Engineering

Soil Mechanics Laboratory

and fill in the appropriate column of the data sheet.

4. Knowing that
1 = 3 + (1 - 3)
compute the principal stress ratio: 1/3
5. Compute V and fill in remaining information on data sheet.
6. Compute the volumetric strain as
vol = V/Vc
7. Plot principal stress difference vs. axial strain, principal stress ratio vs. axial strain and
volumetric strain vs. axial strain on the same piece of paper, to the same horizontal (axial
strain) scale.
8. Plot Mohrs circles for the three tests from your section on the same set of axes and fit a
tangent to the circles. Measure the slope of the tangent as the angle of internal friction of
the soil and show on the plot. Determine the cohesion, c. For cohesionless soils, if you
observe a large value of the intercept, something is probably wrong and should be discussed
in the report.
9. Plot the stress paths for all three tests and show the Kf line. Measure the slope of the Kf
line, , and check that sin = tan.
10. Compare the results of as determined by the direct shear test and the triaxial test.
Comment on the validity of these results.

Report
1. Memo format report.
2. Visual classification.
3. Five plots:

Principal stress difference vs. axial strain

Principal stress ratio vs. axial strain

Volumetric strain vs. axial strain

Mohrs circles for all three tests on the same plot with c and .

Stress paths for all three tests on the same plot with Kf and .

4. Idealized soil elements showing the states of stress on the specimen during a triaxial test.
8

University of Washington

Department of Civil Engineering

Soil Mechanics Laboratory

5. A sketch of the specimen at failure, showing the failure plane and its angle with the
horizontal, and approximate specimen dimensions.
6. As usual, a discussion of the results, including, but not limited to, specific errors and
inaccuracies witnessed and how these would affect the results, the plotsare they what
youd expect? , a comparison of the internal angles of friction obtained from this test and
the direct shear test.