Anda di halaman 1dari 10

Geotechnical Testing Journal, Vol. 32, No.

4
Paper ID GTJ102051
Available online at: www.astm.org

H. B. Nagaraj,1 M. Mohammed Munnas,2 and A. Sridharan3

Critical Evaluation of Determining Swelling


Pressure by Swell-Load Method and Constant
Volume Method

ABSTRACT: For any construction activity in expansive soils, determination of swelling pressure/heave is an essential step. Though many at-
tempts have been made to develop laboratory procedures by using the laboratory one-dimensional oedometer to determine swelling pressure of
expansive soils, they are reported to yield varying results. The main reason for these variations could be heterogeneous moisture distribution of the
sample over its thickness. To overcome this variation the experimental procedure should be such that the soil gets fully saturated. Attempts were made
to introduce vertical sand drains in addition to the top and bottom drains. In this study five and nine vertical sand drains were introduced to experi-
mentally find out the variations in the swell and swelling pressure. The variations in the moisture content at middle, top, and bottom of the sample in
the oedometer test are also reported. It is found that swell-load method is better as compared to zero-swell method. Further, five number of vertical
sand drains are found to be sufficient to obtain uniform moisture content distribution.
KEYWORDS: clays, expansive soils, heave, swelling pressure, vertical drains

Introduction model method (Smith 1973); Mississippi State Highway Dept.


method (Teng et al. 1972, 1973; Teng and Clisby 1975); Controlled
For all construction activities it is essential to rest the foundations strain test (Porter and Nelson 1980) University of Saskatchewan
of structures at a suitable depth and also on a volumetrically stable (Fredlund et al. 1980); Restrained swell test and Double oedometer
soil. Soils that exhibit volume change from variations in soil mois- swell test (Basma et al. 1995). All these methods use an oedometer
ture are referred to as expansive or swelling clayey soils. In case of with varying conditions of allowing the soil to undergo swelling.
a structure rested on expansive soils (which are present in many Recently a method to determine swelling pressure using computer
parts of the world) though are expected to have volumetric changes controlled equipment in the form of a Linear Displacement Trans-
in the presence of water, because of its presence, this volume in- ducer (LDT) has been developed by Thompson et al. (2006). The
crease (swell) is prevented and as a consequence, leads to swelling continued research interest in developing new and more refined
pressure. This swelling pressure has serious consequences in the methods of determining swelling pressure indicates the need for
form of cracks and distress on the structures founded on expansive evaluating the swelling pressure of expansive soils. Hence, the
soils. Lightweight structures are severely affected due to high study of determining swell and swelling pressure of expansive soil
swelling pressure exerted by these soils. assumes importance.
For an engineered construction activity on expansive soils, the It is evident that through a laboratory determination of swell and
first step is to assess the degree of expansiveness and the likely swelling pressure using the oedometer, one can determine the
swelling pressure on the structure. Swelling pressure is defined in amount or degree of swell and swelling pressure. Though many
many ways and is dependent on the testing procedure. To assess the methods of determining the swelling pressure have been developed
degree of swell, many procedures, both simple and elaborate in- by various researchers, only three methods have been standardized
cluding laboratory methods of determining swell pressure have and also popularly used as documented in the literature (Brackley
been developed by geotechnical researchers and engineers, e.g., 1973; Sridharan et al. 1986; ASTM 1995; Rao 2006) as listed
Double oedometer method (Jennings and Knight 1957); Volu- below:
menometer method (DeBruijin 1961); Sampson et al. method 1. The swell-consolidation/swell-load method (Method 1)
(Sampson et al. 1965); Noble method (Noble 1966); Sullivan and 2. The constant volume method/zero-swell method (Method
McClelland method (Sullivan and McClelland 1969); Komornik et 2), and
al. method (Komornik et al. 1969); Navy method (Navy 1971)— 3. The method of equilibrium void ratio at different consoli-
Wong and Yong method (Wong and Yong 1973); USBR method dation pressures (Method 3).
(Gibbs 1973); Simple oedometer (Jennings et al. 1973); Direct Though many attempts have been made to develop laboratory
procedures mainly using the laboratory one-dimensional oedom-
Manuscript received July 31, 2008; accepted for publication March 25, 2009; eter to determine swelling pressure of expansive soils, the same ob-
published online April 2009.
1 tained by different procedures vary to a great extent as reported in
Assistant Professor, Department of Civil Engineering, BMS College of En-
gineering, Bangalore 560 019, India, e-mail: hbnraj@gmail.com
the literature by Sridharan et al. (1986), Dhowian (1990), and Th-
2
Post Graduate Student, Department of Civil Engineering, BMS College of ompson et al. (2006). Dhowian (1990) reports that the volume
Engineering, Bangalore 560 019, India, e-mail: munnas85@yahoo.co.in change behavior and the swell parameters are dependent on the
3
Professor Emeritus, Department of Civil Engineering, Indian Institute of stress path and the wetting sequence in the oedometer test. Invari-
Science, Bangalore 560 012, India, e-mail: sridharanasuri@yahoo.com ably the ISO method (Method 1) reveals higher magnitude for the

Copyright © 2009 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 1
19428-2959. Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
2 GEOTECHNICAL TESTING JOURNAL

TABLE 1—Physical properties of soil used in the present study.

Sl. No. Item Values/Description


Black Cotton Soil from
1 Soil
Belgaum
2 Specific Gravity, G 2.7
3 Liquid Limit (%) 87.9
4 Plastic Limit (%) 36.3
5 Plasticity Index (%) 51.6
6 Shrinkage Limit (%) 11.1
7 Grain Size Distribution Gravel (%) 1
Sand (%) 3
Silt (Size) (%) 27
Clay (Size) (%) 69
8 Unified Soil Classification CH
9 Activitya 0.75
10 Free Swell Indexb (%) 100
11 Free Swell Ratioc 2
12 Principal Clay Mineral Montmorillonite
FIG. 1—Schematic diagram showing the methods of tests to determine swelling
pressure (after Sridharan et al. 1986).
a
Activity, A = 共Plastcity Index兲 / 共% ⬍ 2 µm兲.
b
Free Swell Index, FSI= 共共Vd − Vk兲 / Vk兲 ⫻ 100 where Vd = sediment volume of
10 g of oven dried soil passing a 425 µm sieve placed in a 100 mL graduated
swell parameters, namely swell and swelling pressure than the CVS measuring jar containing distilled water. Vk = sediment volume of 10 g of oven
dried soil passing a 425 µm sieve placed in a 100 mL graduated measuring jar
method (Method 2). The discrepancies in the swell parameters are
containing kerosine.
due to differences in loading and wetting conditions in the follow- c
Free Swell Ratio, FSR is defined as the ratio of equilibrium sediment volume of
ing way: in ISO test soaking the sample under a low confining 10 g oven dried soil passing a 425 µm sieve in distilled water to that in carbon
stress promotes water penetration most efficiently. Therefore the tetrachloride.
swell parameters are relatively high and indicate higher magnitudes
of potential swell as compared to the CVS method (Method 2)
where water entry is restricted by a relatively high value of vertical zero. This method is quick, and again only one specimen is suffi-
stress. Earlier Sridharan et al. (1986) and more recently Thompson cient. However, the method is sensitive to load increment and rate
et al. (2006) have reported similar variations in the swell param- of loading (Fig. 1, curve 2).
eters by these two methods.
With this variation in the swelling pressure obtained by various The Method of Equilibrium Void Ratio at Different
methods mentioned above being present, one needs to explore the
reasons for the same. In this paper an attempt has been made to Consolidation Pressures—Method 3
reason out these variations in the swelling pressure obtained by In this method, three or more specimens are loaded to different
various laboratory methods using an oedometer. pressures around estimated swelling pressure, allowing the speci-
mens to imbibe water, and swell or compress to reach equilibrium
positions, which lie in a straight line on the swell versus log pres-
Different Laboratory Methods of Determining sure plot. This method requires at least three identical specimens
but is less time consuming (Fig. 1, curve 3).
the Swelling Pressure In the present study Method 1 and 2 have been adopted. Method
3 needs at least three or more identical specimens, and preparing
The Swell-Consolidation/Swell-Load Method— three identical specimens is quite difficult and time consuming. It is
also very difficult to ascertain whether the specimens prepared are
Method 1 identical or not without conducting the test. Hence, Method 3 was
This method permits complete swelling of the specimen upon satu- not used in this study.
ration at seating pressure of 6.25 kPa, and then, subsequent loading
it to bring it back to its original volume, which yields the maximum
value of swelling pressure. This method is time consuming, but one Material and Methods
specimen is sufficient (Fig. 1, curve 1).
This test procedure has been used by practicing geotechnical en- One natural expansive soil, namely Black Cotton soil which is ex-
gineers in many parts of the world for many years (Sridharan et al. tensively involved in construction with high value of liquid limit
1986; Reyad 1990; Alanazy and Edil 1992; Xin and Ling 1992). obtained from Belgaum, Karnataka State, India was selected and
used in this study. The soil was characterized for their physical
properties according to ASTM Standards and the results are sum-
The Constant Volume Method/Zero-Swell Method— marized in Table 1. It may be seen that it has a free swell ratio of two
Method 2 (Prakash and Sridharan 2004), and the primary clay mineral being
montmorillonite.
Continuous loading is done in this method, allowing water to be While other properties viz., liquid limit, plastic limit, shrinkage
imbibed by the specimen and keeping the volume change nearly limit, and activity are presented in Table 1; they no way reflect the

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
NAGARAJ ET AL. ON SWELL-LOAD METHOD AND CONSTANT VOLUME METHOD 3

swelling nature of the soil (Sridharan 2005). The reason being that pattern of holes to be drilled for making vertical drains was placed
these properties are determined on remolded soil samples, and on top of the compacted soil specimen. Using a manually operated
hence, do not include in situ soil conditions such as natural dry drill fitted with a drill bit of diameter 2.8 mm, vertical holes were
density, soil fabric, and stress conditions. Thus a meaningful reflec- made over the thickness of the statically compacted soil specimen
tion of the swelling behavior is not possible other than possible ex- in the oedometer rings. Care was taken to drill holes for vertical
istence of problematic swelling soil conditions. drains without disturbing the sides as typically shown in Fig. 2(a)
and 2(b) for nine vertical drains in a compacted soil specimen. The
required number of vertical holes were drilled in the similar man-
Experimental Program ner. At this stage, an air-dried, smoothly ground porous stone was
positioned into the base of the oedometer in a dry condition. A filter
In order to experimentally reason out the cause for the variations of
paper was placed on top of the porous stone. The ring with the
the swelling pressure by two of the three popular methods (Method
1 and Method 2), an experimental program was devised taking into specimen having a requisite number of vertical holes was placed on
consideration the important factors influencing swelling of soils. top of the bottom porous stone and filter paper. The vertical holes
The swell and swelling pressure of an expansive soil is primarily were carefully filled with fine sand of particle size less than 425 µm
dependent on the initial dry unit weight or void ratio and also on the by using a paper cone with a small opening at the bottom (Fig. 2(c)
initial moisture content (El-Sohby and El-Sayed 1981; Sridharan et and 2(d). The sand density was found to vary between
al. 2006, Yevnin and Zaslavsky 1970). Further, the experimental 1.15 to 1.2 Mg/ m3. Sufficient care was taken to fill the sand into
system of conducting swell test should be such that the soil has the vertical drains without spilling around the top of the specimen.
access to imbibe water, so as to get fully saturated. In this study, an An air-dried filter paper was placed on top of the specimen with
attempt has been made to introduce varying numbers of vertical vertical sand drains, above which a porous stone in dry condition
sand drains for providing better drainage to ensure complete satu- and the loading pad were placed. The oedometer was secured to the
ration across the thickness of the specimen in the swell test. This base by means of screws. Thus the test specimen in the oedometer
would overcome the deficiency of the wetting sequence which is of was mounted and positioned on the loading frame with a vertical
major concern as reported in the literature (Dhowian 1990). Also, deflection dial gage with sensitivity of 0.01 mm, properly adjusted
an attempt has been made to study the effect of initial dry density and fixed in position to give the dial gage reading. A seating load of
on the swelling behavior and its associated parameters like swell- 6.25 kPa was applied on the hanger. The initial dial gage reading
time behavior and swelling pressure without and with five and nine was adjusted and recorded.
vertical drains, with two varying densities. Results obtained from
Method 1 and Method 2 are compared and discussed.

Swell and Swelling Pressure Test


Experimental Setup The prepared specimen was inundated with water. The time-swell
The apparatus used in this study is essentially the same as that of observations were recorded for Method 1, for various test condi-
the conventional oedometer with fixed type ring. The rings were of tions, i.e., varying the density and number of drains, (without any
60 mm in diameter and 20 mm in height. Smoothly ground porous drains, five and nine number of drains). For Method 2, continuous
stones have been used in the oedometer to minimize seating dis- loading was carried out such that neither swell nor compression
placements. Porous stones used before placing inside the oedom- was noticed. Care was taken to see that variations from the defor-
eter are air dried and it fits close to the oedometer ring to avoid mation reading at the time of the specimen is inundated at stress ␴1
extrusion or punching at high vertical pressures. Standard filter pa- was kept preferably within 0.005 mm and not more than 0.001 mm
pers were used along with the porous stones to avoid soil penetrat- as outlined in the ASTM Standard D4546-90 (1995). The swelling
ing into the pores of the porous stones. pressures were determined at various densities and with varying
drains by Method 1 and Method 2. After the completion of deter-
mination of swelling pressure, unloading was done in stages until
Preparation of Specimens the seating load was reached. Under the seating load the specimen
was left to reach equilibrium for sufficient time. Then the specimen
It was decided to maintain the height of the specimen approxi- was completely unloaded and removed for determining the water
mately two-thirds the height of the ring to facilitate swelling of the content over the depth of the specimen.
soil on imbibing water. Hence, an initial height of 14 mm was used
for all the tests in this study. Oven dried soil ⬍425 µm was used to
prepare the soil specimens for the swell test at three desired densi-
ties using static compaction. All swell tests in this study were con- Water Content Determination
ducted at three selected dry densities of 1.20 (loose), 1.35 (me-
dium), and 1.50 (high) Mg/ m3. After the completion of the swelling pressure test, the sample was
Before placing the dry soil into the ring, the inside surface was removed from the oedometer ring. The soil sample was sliced into
lubricated with high vacuum silicone grease to reduce side wall three portions, namely, top, middle, and bottom and put into sepa-
friction. At each of the selected dry densities, the requisite amount rate containers, taking care to place the sliced sample from each
of dry soil ⬍425 µm was placed in the oedometer rings and stati- portion into two containers each to obtain the average water content
cally compacted by placing a metal spacer specially fabricated for of each portion. Water content was determined by the oven drying
the purpose to achieve a height of 14 mm on compaction. A plain method. Care was taken to remove the sand being used for vertical
paper (template) having a diameter of 60 mm with the number and drains before slicing into three portions.

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
4 GEOTECHNICAL TESTING JOURNAL

Results and Discussion

Results of Swell Tests


It has been earlier reported by various researchers that initial place-
ment condition, namely, the dry density of the soil is one of the
major factors influencing swelling behavior of soil (e.g., El-Sohby
and El-Sayed 1981; Al-Mhaidib 1999; Azam and Abduljauwad
2000). In this study, three dry densities, i.e., 1.20 Mg/ m3,
1.35 Mg/ m3, and 1.50 Mg/ m3 were selected to study the influence
of initial dry density on swell behavior of expansive soils with a
specific interest in studying the influence of introducing vertical
drains in combination with varied initial dry densities.
Figure 3(a) shows the swell-time behavior of the soil selected at
three dry densities without any vertical drains. It can be seen from
the figure that the amount of swell has increased with an increase in
the dry density, this is as expected. In order to have a check on the
reproducibility of the test results, certain swell tests were repeated
in the entire experimental program. One such reproducibility test is
shown in the figure at a dry density of 1.20 Mg/ m3. It can be ob-
served that there is good reproducibility on the swell-time behavior
as well on the amount of swell.
Since efforts to drill holes for making vertical drains at the lower
density of 1.20 Mg/ m3 was found not successful. Further swell
tests at that density were not considered. Figures 3(b) and 3(c) show
the swell-time behavior having five and nine vertical drains, respec-
tively, at two dry densities, namely, 1.35 Mg/ m3 and 1.50 Mg/ m3.
The amount of swell has shown to increase with the increase in
density for the same number of vertical drains. Also the tests have
been shown to be reproducible.
Figures 4(a) and 4(b) show the influence of introducing vertical
sand drains on the swell-time behavior at particular density of
1.35 Mg/ m3 and 1.50 Mg/ m3, respectively. It can be observed
from Fig. 4(a) that the amount of swell has increased with the in-
troduction of vertical drains as compared to those without vertical
drains, showing very marginal variation in the swell behavior be-
tween five and nine vertical drains. Similar increased swell can be
observed at higher density of 1.50 Mg/ m3 with the introduction of
vertical drains as shown in Fig. 4(b). However, there is marked in-
fluence both on the swell and swell-time behavior between five and
nine vertical drains. Table 2 summarizes the amount of swell in di-
visions and in percentage without and with vertical drains at varied
dry densities. It is seen from Table 2 that as for the case of no ver-
tical drains being present, for an increase in the initial dry density
from 1.2 Mg/ m3 to 1.5 Mg/ m3, the percent swell increases from
14.4 % (Trial 1)/13.4 % (Trial 2) to 31.1 %, i.e., 2.22 times the
increase in swell. Similarly, for a density variation from
1.35 to 1.5 Mg/ m3, the swell has increased from 18.9 % to 31.1 %,
i.e., 1.65 times the increase in swell. When the vertical drains are
five in number, the increase in percent swell is from 22.1 % (Trial
1)/22.4 % (Trial 2) to 34.4 % for an increase in dry density of
1.35 to 1.5 Mg/ m3, i.e., it has increased by 1.56 times. When the
number of drains were increased from five to nine, the increase in
percent swell is from 22.6 % to 36.3 %, for an increase in dry den-
sity from 1.35 to 1.5 Mg/ m3, i.e., it has increased by 1.61 times.
FIG. 2—(a) Final stage of drilling nine vertical holes. (b) See through view of
nine vertical holes for vertical drains. (c) Filling of sand using paper cone to
Thus it is seen that irrespective of the presence or absence of drains,
make nine vertical drains. (d) Final view of specimen filled up with sand to make the density effect on swell is around a 60 % increase, for an increase
nine vertical drains. of density from 1.35 to 1.5 Mg/ m3, which is quite significant. It is
also clear that even in the presence of vertical drains, increase in
swell values with an increase in density has almost remained the
same, i.e., ratio of percent swell for ␳d = 1.50 to 1.35 Mg/ m3 is the

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
NAGARAJ ET AL. ON SWELL-LOAD METHOD AND CONSTANT VOLUME METHOD 5

FIG. 3—(a) Swell versus log time relationship for varying density with zero vertical drains. (b) Swell versus log time relationship for varying density with five vertical
drains. (c) Swell versus log time relationship for varying density with nine vertical drains.

same, as given in Table 2. At a particular density, introduction of 1.35 Mg/ m3, with and without vertical drains (Method 1). Figures
vertical drains has increased the swell values, being more signifi- 6(a) and 6(b) show typical plots of void ratio versus effective ver-
cant when the density is higher. When the density is less tical consolidation pressure (logarithmic scale) for the dry density
共1.35 Mg/ m3兲 the difference between five and nine drains is mar- 1.35 Mg/ m3 with five and nine vertical drains, respectively
ginal. (Method 2). Similar results have been obtained for the other density
of 1.5 Mg/ m3.
Table 3 summarizes the values of swelling pressure by Method 1
Results of Swell Pressure Test and Method 2 for all the placement conditions. It can be seen that
Figure 5 presents a typical plot of void ratio versus effective verti- by Method 1, and at a dry density of 1.35 Mg/ m3, with the intro-
cal consolidation pressure (logarithmic scale) for the dry density duction of vertical drains there is marked influence on the swelling

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
.
6 GEOTECHNICAL TESTING JOURNAL

FIG. 4—(a) Swell versus log time relationship for same density 共=1.35 Mg/ m3兲 with varying vertical drains. (b) Swell versus log time relationship for same density
共=1.50 Mg/ m3兲 with varying vertical drain.

pressure. The swelling pressure has increased from 315.5 kPa for sure than Method 1 for all the cases and it varies from 15.5 % to
zero Vertical Drain (VD) to 380.7 kPa for five VDs (20.7 % higher) 32.3 %. Thus, it can be observed that there is only a marginal varia-
and to 391.1 kPa (24 % higher) for nine VDs. However, for Method tion in swell and swelling pressure between five and nine vertical
1, between five and nine VDs, the increase in swelling pressure is drains by Method 1 as compared to Method 2.
only 3 %. The reason for this observation may be that the increase The effect of density on swelling pressure by Method 1 and
in VDs from five to nine had a marginal influence on the water im- Method 2 can be more clearly seen from Table 3. For Method 1, the
bibed over the thickness of the compacted dry specimens, and swelling pressure has increased from 315.5 kPa to 939.3 kPa for
hence the resulting swell and swelling pressure. At the same den- zero VD and for an increase of density from
sity, a similar increase in swelling pressure by Method 2 is 24 % and 1.35 Mg/ m3 to 1.5 Mg/ m3, the increase being 2.98 times. For five
43 %. However, for Method 2, between five and nine VDs, the in- and nine VDs, respective values are 2.92 times and 2.97 times, re-
crease in swelling pressure is 15.6 %. This may be because of the spectively. For all three cases the variation in swelling pressure is on
increased water imbibed with the increase in VDs from five to nine. the order of around 2.95 times for an increase in density of
Similarly, at a dry density of 1.5 Mg/ m3, swell tests by Method 1.35 to1.5 Mg/ m3. Similar values for Method 2 are 2.75 times
1, the swelling pressure has found to increase significantly with the (zero VD), 2.89 times (five VDs), and 2.93 times (nine VDs). Sum-
introduction of vertical drains. The swelling pressure has increased marily, for Method 2 the increase in swelling pressure is on the
from 939.3 kPa for zero VD to 1113.4 kPa for five VDs (18.5 % order of 2.86 times, being not very much different as compared to
higher) and to 1162.9 kPa (24 % higher) for nine VDs. However, Method 1. Thus it can be seen that, irrespective of the presence or
between five and nine VDs, the increase in swelling pressure is only absence of drains, the density effect on swelling pressure is around
4.4 %. By Method 2, the increase in swelling pressure for an in- 2.9 times by both the methods for a density increase from
crease in VDs to five and nine are 30 % and 53 %, respectively. 1.35 to 1.5 Mg/ m3. Thus, the density effect is quite significant and
However, for Method 2, between five and nine VDs, the increase in uniform.
swelling pressure is 17.4 %. Method 2 gives a lesser swelling pres- Further, the authors feel it is better to discuss the role of sand

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
NAGARAJ ET AL. ON SWELL-LOAD METHOD AND CONSTANT VOLUME METHOD 7

TABLE 2—Experimental values of swell.

Swell Values—Method 1
Ratio of
Dry Density Percent
1.20 Mg/ m3 1.35 Mg/ m3 1.50 Mg/ m3 Swell for ␳d
No. of =1.50 to
Vertical Swell Swell Swell Swell Swell Swell 1.35
Drains (divsa) (%) (divs) (%) (divs) (%) Mg/ m3
Zero 202 14.4 % 265 18.9 % 436 31.1 % 1.65
(Trial—1) (Trial—1)
188 13.4 %
(Trial—2) (Trial—2)
Five … … 310 22.1 % 482 34.4 % 1.54
(Trial—1) (Trial—1)
314 22.4 %
(Trial—2) (Trial—2)
Nine … … 316 22.6 % 508 36.3 % 1.61
(Trial—1) (Trial—1)
316 22.6 %
(Trial—2) (Trial—2)
a
1 division= 0.01 mm.
Note: All tests also include top and bottom drains.

drains not reinforcing the swollen soil and hence have no influence Moisture Content Variations over the Depth of
on the swelling pressure. The volume of sand used in VDs was
the Specimen
about 1.1 % (five VDs) and 1.98 % (nine VDs) of the total volume
of the compacted sample used. This is a very small fraction of the In order to verify the accessibility of water over the depth of the
total volume of compacted soil used in the tests. Also, the density of specimen for various testing conditions, the sample was removed
sand in VDs was in a loose state varying between from the oedometer ring after the completion of the swell pressure
1.15 to 1.2 Mg/ m3. Hence the presence of sand drains would not test. The soil sample was sliced into three portions, namely, top,
reinforce the soil, and hence have influence on the compressibility middle, and bottom and the water content of each portion was de-
and swelling pressure. This aspect is clear from Fig. 5. From the termined by the oven drying method.
figure it can be observed that the presence of sand drains does not From Table 4(a) which summarizes the water content over the
show any significant change in compressibility properties. Further, depth of the specimen for various placement conditions, it can be
the differences in swell magnitude observed between soil with and observed that for tests conducted without vertical drains, the mois-
without drains is due to the better drainage through the VDs as
ture content varied over the depth by both the methods of testing.
clearly brought out in the experimental work and not due to addi-
This variation is observed to be less in Method 1 as compared to
tional confinement/reinforcement. Hence, the observed increase in
Method 2. The variation of moisture content between the center of
the swell and swelling pressure is because of better saturation of the
the sample to the average of the top and bottom of the sample is
soil specimen with improved drainage and is further explained
summarized in Table 4(b). In Method 1, the variation of water con-
below.
tent is observed to be 2 % and 2.85 % at 1.35 Mg/ m3 and
1.50 Mg/ m3, respectively, when there is no drains. Similarly, for
Method 2, for zero VD, the variation of water content is observed to
be 4.1 % and 10.9 % at 1.35 Mg/ m3 and 1.50 Mg/ m3, respectively.
The variations are observed to be reduced significantly with the in-
troduction of the vertical drains. In Method 1, the variations in
water content over the depth was observed to be almost the same
with five and nine vertical drains. This may be the reason for both
swell and swelling pressure being nearly the same for tests con-
ducted by Method 1 with five and nine vertical drains. Whereas, in
Method 2, the variations in water content with the introduction of
vertical drains were found to be quite less as compared to that with-
out vertical drains, but the difference being still present, the varia-
tion being about 2 % and 5.1 % at 1.35 Mg/ m3 and 1.50 Mg/ m3,
respectively, for five vertical drains and about 1 % and 2 % at
1.35 Mg/ m3 and 1.50 Mg/ m3, respectively, for nine vertical
FIG. 5—Log pressure versus void ratio (Method 1) relationship for same den- drains. This difference in the water content explains the differences
sity 共=1.35 Mg/ m3兲 with varying vertical drains. in swelling pressure found in Method 2.

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
8 GEOTECHNICAL TESTING JOURNAL

FIG. 6—(a) Log pressure versus void ratio (both Method 1 and Method 2) relationship for density 共=1.35 Mg/ m3兲 with five vertical drains. (b) Log pressure versus
void ratio (both Method 1 and Method 2) relationship for density 共=1.35 Mg/ m3兲 with nine vertical drains.

Concluding Remarks of swelling pressure than Method 2, and the difference in swelling
pressure determined using five or nine vertical drains is only mar-
From this experimental investigation done to determine swell and ginal, the reason being that in Method 1 the sample is allowed to
swelling pressure, it is evident that Method 1 (Swell-load method)
swell by imbibing water under seating load of 6.25 kPa, and later
and Method 2 (Constant volume method) give varied results of
loading the swollen sample to determine the swelling pressure. In
swelling pressures without vertical drains. Introduction of vertical
drains has not only improved the swell but also swelling pressure in this process, the number of five vertical drains was sufficient
Method 1 and swelling pressure in Method 2. Further, the differ- enough for the sample to imbibe water over the complete thickness
ences in the swelling pressure by the two methods have narrowed of the sample and has been reflected in the uniformity of the water
down. It can be further observed that Method 1 gives higher values content over the depth. In Method 2, introducing vertical drains has

TABLE 3—Values of swelling pressure by Method 1 and Method 2 at various dry densities.

Values of Swelling Pressure (kPa)


Ratio of
Swelling Pressure by
Method 1 Method 2 Method 1 to Method 2
Ratio of Ratio of
Swelling Swelling
Dry Density Pressure Dry Density Pressure Dry Density
Number for ␳d= for ␳d=
of 1.50 to 1.50 to
Vertical 1.20 1.35 1.50 1.35 1.20 1.35 1.50 1.35 1.20 1.35 1.50
Drains Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3
Zero 171.7 315.5 939.3 2.98 160.4 230.9 635.5 2.75 1.07 1.37 1.48
Five … 380.7 1113.4 2.92 … 286 827 2.89 … 1.33 1.35
Nine … 391.1 1162.9 2.97 … 330.5 970.9 2.94 … 1.18 1.19
Note: All tests also include top and bottom drains.

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
NAGARAJ ET AL. ON SWELL-LOAD METHOD AND CONSTANT VOLUME METHOD 9

TABLE 4(a)—Values of moisture content (%) determined over the depth of the specimen.

Moisture Content (%)


Number Method 1 Method 2
of Dry Density Dry Density
Vertical Specimen 1.2 1.35 1.5 1.2 1.35 1.5
Drains Portion Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3
Zero Top 47.1 45.4 36.4 46.3 41.5 31.4
Middle 46.9 43.5 33.5 45.4 37.5 20.6
Bottom 47.3 45.7 36.2 46.7 41.7 31.6
Five Top … 43.8 34.7 … 40.2 30.4
Middle … 43.2 33.3 … 38.2 25.4
Bottom … 43.4 34.6 … 40.2 30.7
Nine Top … 43.4 31.2 … 39.7 29.3
Middle … 43.2 31.1 … 38.3 27.3
Bottom … 43.4 31.3 … 39.2 29.4
Note: All tests also include top and bottom drains.

TABLE 4(b)—Variation of moisture content between center of the sample to average of top and bottom of the
sample.

Variation of Moisture Contenta


Method 1 Method 2
Number of Dry Density Dry Density
Vertical 1.2 1.35 1.5 1.2 1.35 1.5
Drains Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3 Mg/ m3
Zero 0.25 2.01 2.85 1.07 4.06 10.88
Five … 0.42 1.30 … 2.06 5.07
Nine … 0.20 0.16 … 1.19 2.08
a
Variation of moisture content = (Water content at top or bottom—Water content at middle).

helped to improve the swelling pressure, but has remained lower Azam, S. and Abduljauwad, S. N., 2000, “Influence of Gypsifica-
than Method 1 and also differs with the number of vertical drains. tion on Engineering Behavior of Expansive Clays,” J. Geotech.
This is evident by the variation in the water content over the thick- Geoenviron. Eng., 126, No. 6, pp. 538–542.
ness of the specimens. In Method 2, since the sample is not allowed Basma, A. A., Al-Homoud, A. S., and Husein, A., 1995, “Labora-
to swell by continuous loading, the presence of nine vertical drains tory Assessment of Swelling Pressure of Expansive Soils,” Appl.
has allowed to imbibe more water than five vertical drains and Clay Sci., Vol. 9, pp. 355–368.
hence increased swelling pressure and also less variation in the Brackley, J. J. A., 1973, “Swell Pressure and Free Swell in a Com-
water content over the thickness of the specimen. From the above pact Clay,” Proceedings of the 3rd International Conference on
discussion it can be inferred that introducing vertical drains cer- Expansive Clays, Vol. 1, Israel Institute of Technology, Haifa,
tainly increases the swelling pressure, and that the soil sample in pp. 169–176.
the swell test by Method 1 reflects a completely swollen condition. De Bruijn, C. M. A., 1961, “Swelling Characteristics of a Trans-
Hence, the swelling pressure determined by Method 1 can be con- ported Soil Profile at Leeuhof Vereeniging (Transvaal),” Pro-
sidered better when compared to Method 2. The results bring out ceedings, 5th International Conference on Soil Mechanics and
the fact that the existing procedure predicts lesser value of swell Foundation Engineering, Vol. 1, pp. 43–49.
and swelling pressure by both the methods and is not conservative. Dhowian, A. W., 1990, “Heave Prediction Techniques and Design
Considerations on Expansive Soils,” Journal of King Saudi Uni-
versity, Vol. 2, No. 2, pp. 355–377.
References El-Sohby, M. A. and El-Sayed, A. R., 1981, “Some Factors Affect-
ing Swelling of Clayey Soils,” J. Geotech. Engrg., Vol. 12, pp.
Alanazy, A. S. and Edil, T. B., 1992, “Lateral swelling pressure,” 19–39.
Proceedings of the 7th International Conference on Expansive Fredlund, D. G., Hasan, J. U., and Filson, H. L., 1980, “The Predic-
Soils, 3–5 Aug., Dallas, TX. tion of Total Heave,” Proceedings of 4th International Confer-
Al-Mhaidib, A., 1999, “Swelling Behavior of Expansive Shales ence on Expansive Soils,” ASCE and International Society for
from the Middle Region of Saudi Arabia,” Geotech. Geologic. Soil Mechanics and Foundation Engineering, Denver, pp. 1–17.
Eng., Vol. 16, pp. 291–307. Gibbs, H. J., 1973, “Use of a Consolidometer for Measuring Ex-
ASTM Standard D4546-90, 1995, “Standard Test Methods for pansion Potential of Soils,” Proceedings of Workshop on Expan-
One-Dimensional Swell or Settlement Potential of Cohesive sive Clays and Shales in Highway Design and Construction,
Soils,” Annual Book of ASTM Standards, Vol. 4, No. 8, ASTM University of Wyoming, Laramie, pp. 206–213.
International, West Conshohocken, PA, pp. 696–702. Jennings, J. E. B. and Knight, K., 1957, “The Prediction of Total

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009
10 GEOTECHNICAL TESTING JOURNAL

Heave from the Double Oedometer Test,” Transact. S. African Laramie, pp. 189–205.
Inst. Civil Eng., Vol. 7, pp. 285–291. Sridharan, A., 2005, “On Swelling Behaviour of Clays,” Proceed-
Jennings, J. E. B., Firtu, R. A., Ralph, T. K., and Nagar, N., 1973, ings of International Conference on Problematic Soils, Eastern
“An Improved Method for Predicting Heave Using the Oedom- Mediterranean University North Cyprus, Cyprus, pp. 499–516.
eter Test,” Proceedings on Vertical Rise, 3rd International Con- Sridharan, A., Rao, A. S., and Sivapullaiah, P. V., 1986, “Swelling
ference on Expansive Soils, Haifa, Israel, Vol. 2, pp. 149–154. Pressure of Clays,” Geotech. Test. J., Vol. 9, No. 1, pp. 24–33.
Komornik, A., Wiseman, G., and Ben-Yaacob, Y., 1969, “Studies Sullivan, R. A. and McClelland, B., 1969, “Predicting Heave of
of In-Situ moisture and Swelling Potential Profiles,” Proceed- Buildings on Unsaturated Clay,” Proceedings, 2nd Interna-
ings, 2nd International Research and Engineering Conference tional Research and Engineering Conference on Expansive
on Expansive Soils, Texas A&M Univ. Press, College Station, Soils, Texas A&M Univ. Press, College Station, TX, pp. 404–
TX, pp. 348–361. 420.
Navy, Department of Naval Facilities Engineering Command, Teng, T. C. P. and Clisby, M. B., 1975, “Experimental Work for
1971, Design Manual—Soil Mechanics (DM 7.01), U.S. Depart- Active Clays in Mississippi,” Transp. Engrg. J., Vol. 101, pp.
ment of the Navy, Alexandria, VA, p. 256. 77–95.
Noble, C. A., 1966, “Swelling Measurements and Prediction of Teng, T. C. P., Mattox, R. M., and Clisby, M. B., 1972, “A Study of
Heave for a Lacustrine Clay,” J. Dev. Studies, Vol. 3, No. 1, pp. Active Clays as Related to Highway Design,” Research and De-
32–41.
velopment Divisions, Mississippi’s State Highway Department,
Porter, A. A. and Nelson, J. D., 1980, “Strain Controlled Testing of
Engineering and Industrial Research Station, Mississippi State
Soils,” Proceedings of 4th International Conference on Expan-
University, MSHDRD-72-045, p. 134.
sive Soils, ASCE and International Society for Soil Mechanics
Teng, T. C. P., Mattox, R. M., and Clisby, M. B., 1973, “Mississip-
and Foundation Engineering, Denver, pp. 34–44.
pi’s Experimental Work on Active Clays,” Proceedings of the
Prakash, K., and Sridharan, A., 2004, “Free Swell Ratio and Clay
Workshop on Expansive Clays and Shales in Highway Design
Mineralogy of Fine Grained Soils,” Geotech. Test. J., Vol. 27,
No. 2, pp. 220–225. and Construction, University of Wyoming, Laramie, pp. 1–17.
Rao, S. M., 2006, “Identification and Classification of Expansive Thompson, R. W., Perko, H. A., and Rethamel, W. D., 2006, “Com-
Soils,” Exapnsive Soils—Recent Advances in Characterization parison of Constant Volume Swell Pressure and Oedometer
and Treatment, Taylor & Francis, pp. 15–24. Load Back Pressure,” Proccedings of the 4th International Con-
Reyad, M. M., 1990, “Comparison between Two Methods for ference on Unsaturated Soils, Geotechnical Special Publication,
Swelling Pressure Determination,” Proceedings of the 1st Alex- ASCE, Reston, VA, No. 147, Vol. 2, pp. 1787–1798.
andria Conference on Structures and Geotechnical Engineer- Wong, H. Y. and Yong, R. M., 1973, “A Study of Swelling and
ing, Alexandria, Egypt, pp. 53–61. Swelling Force During Unsaturated Flow in Expansive Soils,”
Sampson, E., Schuster, R. L., and Budge, W. D., 1965, “A Method Proceedings, 3rd International Conference on Expansive Soils,
of Determining Swell Potential of an Expansive Clay,” Conclud- Haifa, Israel, Vol. 1, pp. 143–151.
ing Proceedings on Engineering Effects of Moisture Changes in Xin, J. Z. and Ling, Q. X., 1992, “A New Method for Calculating
Soils, International Research Engineering Conference on Ex- Lateral Swelling Pressure in Expansive Soil,” Proceedings of
pansive Clay Soils, Supplementing the Symposia in Print, Texas the 7th International Conference on Expansive Soils, 3–5 Aug.,
A&M Univ. Press, Collge Station, TX, pp. 255–275. Dallas, TX.
Smith, A. W., 1973, “Method for Determining the Potential Vertical Yevnin, A. and Zaslasvky, D., 1970, “Some Factors Affecting Com-
Rise,” Proceedings of Workshop on Expansive Clays and Shales pacted Clay Swelling,” Can. Geotech. J., Vol. 7, No. 1, pp. 79–
in Highway Design and Construction, University of Wyoming, 89.

Copyright by ASTM Int'l (all rights reserved); Fri Jul 17 04:00:46 EDT 2009

Anda mungkin juga menyukai