Construction and Building Materials 93 (2015) 775781

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Investigation on conditions of hydraulic fracturing for asphalt concrete

used as impervious core in dams
Yingbo Zhang, Weibiao Wang , Yue Zhu
Xian University of Technology, 5 Jinhua South Road, 710048 Xian, China

h i g h l i g h t s
 Model tests carried out to simulate the conditions for very high asphalt core dams.
 The fractures of asphalt in the model tests due to the induced large tension strains.
 Hydraulic fracturing can be excluded for asphalt core in dams in common conditions.
 Hydraulic fracturing is discussed for asphalt core in dams in very extreme conditions.

a r t i c l e

i n f o

Article history:
Received 28 October 2014
Received in revised form 13 April 2015
Accepted 2 May 2015
Available online 23 May 2015
Keywords:
Hydraulic fracturing
Asphalt core
Model test
Finite element analysis
Embankment dam

a b s t r a c t
Asphalt concrete is used as impervious core in embankment dams. This paper investigates the conditions
under which hydraulic fracturing could be possible to occur for an asphalt core in dams. Asphalt concrete
test model was developed and tests were carried out on the models under tension, restraint and compression conditions corresponding to nearly-impossibly extreme conditions of an asphalt core in dams and
nite element analysis was performed on the model test results. Test and analyzed results show that
asphalt concrete models at fracturing are caused by large tensile strains in asphalt concrete under the
imposed conditions rather than the so-call hydraulic fracturing. The imposed conditions are extremely
impossible for an asphalt core in dams in reality and therefore the so-call hydraulic fracturing may be
excluded.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
The problem of hydraulic fracturing is one of major concerns in
earth core embankment dam design. The void content of an earth
core is commonly in a range of 1020% and a continuous hydrostatic pressure develops in the earth core due to water owing
through the core when reservoir is impounded. When the hydrostatic pressure or pore water pressure is greater than the vertical
total stress in the core a phenomenon of hydraulic fracturing
may occur [1].
In contrast with the void content of an earth core in dams, the
air void content of an asphalt core is less than 3% and mineral
aggregates are bounded by bitumen with high viscosity. There is
no system of voids in the asphalt core that would allow the
entrance of water and thus the generation of a pore water pressure
in the core. The main prerequisites for hydraulic fracturing are
Corresponding author. Tel.: +86 29 82313133; fax: +86 29 83230217.
E-mail addresses: zhangyb68@163.com (Y. Zhang), wangweibiao59@hotmail.
com (W. Wang), ztt021210@163.com (Y. Zhu).
http://dx.doi.org/10.1016/j.conbuildmat.2015.05.097
0950-0618/ 2015 Elsevier Ltd. All rights reserved.

non-existent in the case of impervious asphalt concrete. The applications of more than 300 asphalt facings on banks and bottoms of
reservoirs and on upstream slopes of embankment dams since the
1920s have been no indication of a hydraulic fracturing yet,
although there are considerably low effective stresses (even some
tensile strains) in the impervious asphalt layer of 510 cm thickness in comparison with the external reservoir water pressure.
Therefore, as early as in 1992 the International Commission on
Large Dams (ICOLD) excluded the effect of hydraulic fracturing
for an asphalt core in dams [1]. Late in 1999 Schnian [2] also stated the problem of hydraulic fracturing does not exist for core
walls made of asphaltic concrete.
In last century the dam height of most embankment dams with
an asphalt core or asphalt facing was less than 100 m and most
dam embankments seated on rock foundations in valleys with gentle abutments. In this century with the experience obtained from
researches and applications on asphalt concrete [36] the dam
height of some asphalt core dams is more than 150 m and some
dam embankments seat on deep compressive overburdens in valleys with steep abutments. For instance, the maximum dam height

776

Y. Zhang et al. / Construction and Building Materials 93 (2015) 775781

for the Cetin Dam in Turkey is 165 and 171 m for the Quxue Dam
which seats in a narrow valley with a steep abutment of 1V:0.33H
(72) in China [7]. In the special cases asphalt core may be subjected to more embankment and foundation differential settlements which may result in suspension of the core on the more
rigid zone of embankment and cause more reduction of the vertical
stress in the core. Furthermore, if an asphalt core dam is located in
a narrow valley with steep abutments arching effect may exist on
the core in the longitudinal direction. Such complexity of geological conditions for an asphalt core dam has caused the professions
uncertainty and concern about that how the effect of hydraulic
fracturing for an asphalt core could be excluded in such special
cases. This was the main incentive for the authors to undertake
the following study: (1) determining the conditions of hydraulic
fracturing on asphalt concrete with model tests; (2) calculating
the tensile strains in the asphalt concrete for the model test results
with nite element method; and (3) discussing the possibility of
hydraulic fracturing in an asphalt core in embankment dams.

2. Test model and specimen preparation

Fig. 1 shows the schematic diagram of the developed equipment
for asphalt concrete model tests for hydraulic fracturing.
The asphalt concrete specimen with dimensions of 150 mm in
diameter and 250 mm in height was prepared in the cylindrical
steel mold. Standard asphalt concrete core mix design criteria have
been developed and, with relatively small variations, used for most

Table 1
Asphalt concrete mix used in this study (% of mineral weight).
Sieve size (mm)
13.2
19

9.5
13.2

4.75
9.5

2.36
4.75

0.075
2.36

00.075
(Filler)

13

10

18

14

31

14

Bitumen
content (B70)

7.0

of asphalt core dams for more than 50 years. The aggregate grain
sizes usually lie between 0 and 16 or 18 mm. The mix design of
the asphalt core for a high embankment dam in China was used
in this study and shown in Table 1. The aggregates were crushed
limestone. The added ller consisted of limestone powder and
the total ller content (<0.075 mm) was 14% of the mineral weight.
The bitumen was of grade B70 and the bitumen content was 7.0%
of the mineral weight (6.5% of the total weight).
The dry aggregates, added ller and bitumen were heated and
mixed and then the hot asphalt mix with a temperature of about
150 C was placed in three consecutive layers in the cylindrical
steel mold with a diameter of 150 mm. The top side of each layer
was compacted with the standard Marshall hammer with a at circular tamping surface of 98.4 mm diameter in a uniform distribution blows on the top area to simulate the eld roller compaction
of the core [8]. The compacted height of the asphalt base-layer
was 70 mm and a copper tube of 6 mm in diameter was pre-set
in the center of the base-layer. Before placing the middle-layer
asphalt mix a thin cylindrical steel tube of 40 mm in diameter
was temporarily xed in the center and lled with natural sand
in the tube and a lubricating paper was placed around the inner
perimeter of the steel mold wall. After placing asphalt mix in the
cavity between the tube and the steel mold wall the tube was
pulled out and the asphalt mix together with the natural sand
was compacted. The height of the asphalt middle-layer was
80 mm. The compacted height of the asphalt top-layer was
100 mm. The blows of Marshall hammer for the base-layer,
middle-layer and top-layer asphalt were 150, 170 and 210, respectively, after having been trying to compact a few specimens to
reach air void content of less than 2% and with air voids randomly
scattered in the specimens. The differences among the three layers
in blows were in correspondence with the different layer heights of
70, 80, and 100 mm from the base to top-layer with am approximately same compaction energy input. To measure the vertical displacement (vertical strain) of the asphalt specimen two
displacement base bars were installed in the base- and top-layer
asphalt, respectively. A steel cover was bound on the top of the
specimen.
3. Model test results under tension, restraint and compression
conditions
Asphalt concrete presents a viscoelastoplastic behavior and its
stressstrain
relation
is
time-dependent
and
temperature-dependent. Therefore, the model tests were carried
out in a stepped-loading creep manner at different temperatures
to reach relatively stable stressstrain states.
The selected test temperatures were 5, 10 and 20 C, respectively, considering the temperature of an asphalt core being 5 C
for embankment dams located in sub-arctic climate, being 20 C
in sub-tropical climate and being 10 C between the two extreme
climates. Prepared test model was kept at specied temperature
for at least one day before a designed test was carried out.

Fig. 1. Schematic diagram of the equipment developed to carry out asphalt

concrete model tests for hydraulic fracturing. 1. Asphalt concrete; 2. natural sand
(<2.36 mm); 3. pressure water; 4. lubricating paper; 5. displacement base bar of
8 mm in diameter; 6. displacement meter; 7. water outlet; 8. displacement
adjustable screw.

3.1. Model test results under tension

The test was designed with an attempt to investigate how large
tensile strain the asphalt specimen could undertake with a

777

Y. Zhang et al. / Construction and Building Materials 93 (2015) 775781

constant water pressure in the inner chamber in the specimen. The

test was carried out in the following procedures (see Fig. 1). The
vertical displacement of the top of the asphalt specimen was rst
restrained by xing the adjustable screws and colored water was
lled in the chamber and the water pressure was increased up to
0.5 MPa and kept for one day before testing. During the testing
the water pressure was kept at a constant value of 0.5 MPa. The
top of the asphalt specimen was pulled by rotating the adjustable
screws and the vertical displacements (tensile strains) between the
base bars on opposite side of the specimen were measured to calculate the average tensile strains. The tension was performed in a
step creep manner and carried out in two ways. The rst way was
to rotate the screws per half of an hour to result in an average tensile strain rate of 0.1%/h between the base bars while the second
way was to rotate the screws per day to result an average tensile
strain rate of 0.2%/day. The tests were stopped when a water leakage was observed which implied fracturing of the asphalt specimen. Figs. 2 and 3 show the test results in the rst way and in
the second way, respectively.
Table 2 shows a comparison of the tensile strains between the
base bars at fracturing of the asphalt specimen with two different
strain rates under a condition of a water pressure of 0.5 MPa in the
chamber in the specimen.
Table 2 shows that the signicant difference between the two
tensile strain rates in the step tension creep had an insignicant

Table 2
Tensile strains (%) between the base bars at fracturing of the asphalt specimen under
a condition of a water pressure of 0.5 MPa in the chamber in the specimen and under
different strain rates and temperatures.
Strain Rate

Temperature (C)

0.1%/h
0.2%/day

10

20

1.1
1.0

1.2
1.3

1.4
1.4

effect on the tensile strains between the base bars at fracturing

of the asphalt specimen, which is in agreement with the ndings
by Zhang et al. [6] that the bending strains at cracking of asphalt
concrete was not affected by the creep rate. Increasing test temperature resulted in a little tensile strain increase at fracturing of the
asphalt specimen in the tests.
3.2. Model test results under restraint
The test was designed with an attempt to investigate how high
water pressure in chamber the asphalt specimen could undertake
under restraint conditions. The vertical displacement of the top
of the asphalt specimen was restrained by xing the adjustable
screws during the testing (see Fig. 1). Water pressure in the chamber was step by step increased at a rate of 0.30.6 MPa per day in
the preliminary stage and 0.1 MPa per day when approaching possible fracturing of the asphalt specimen. Fig. 4 shows the test
results.
The test results indicate that the water pressure was 3.1 MPa at
5 C, 1.2 MPa at 10 C and 0.7 MPa at 20 C, respectively, under
restraint conditions for fracturing of the asphalt specimen and that
low temperature conditions respond high water pressure
conditions.
3.3. Model test results under compression

Fig. 2. Tensile strain between the base bars versus time under a water pressure of
0.5 MPa in the chamber in the asphalt specimen and with a strain rate of 0.1%/h
before fracturing of the asphalt specimen.

The test was designed with an attempt to investigate how low

compressive stress the asphalt specimen could undertake under
high water pressure in the chamber. The test model was put in a
compression testing machine and tested at 10 C. Two test models
were performed.
For the rst test model, the vertical force P was gradually
increased up to an average vertical compressive stress of
1.55 MPa on the asphalt specimen in 2 h which resulted in a 0.8%
of average compressive strain between the base bars (see Fig. 1).
It should be noted that the displacement adjustable screws were

3.5

Water Pressure (MPa)

3.0
2.5

5C
10C

2.0

20C

1.5
1.0
0.5
0.0
0

Time (Day)
Fig. 3. Tensile strain between the base bars versus time under a water pressure of
0.5 MPa in the chamber in the asphalt specimen and with a strain rate of 0.2%/day
before fracturing of the asphalt specimen.

Fig. 4. Water pressure in the chamber in the asphalt specimen versus time under
restraint conditions before fracturing of the asphalt specimen.

Y. Zhang et al. / Construction and Building Materials 93 (2015) 775781

1.2

4.0

2.5

1.0

3.5

0.8

1.5

0.6

1.0

Water Pressure
Specimen Stress

0.4

0.5

Strain

0.2
0.0

0.0
0

1.4
Water Pressure
Specimen Stress
Strain

3.0
Stress (MPa)

2.0

1.6

2.5

1.2
1.0

2.0

0.8

1.5

0.6

1.0

0.4

0.5

0.2

0.0

Time (Hour)

0.0
0

Fig. 5. Compressive stress on the hollow asphalt concrete and corresponding

compressive strain between the base bars as well as water pressure (2.5 MPa) in the
chamber versus time before fracturing of the specimen.

Compressive Strain (%)

3.0

Compressive Strain (%)

Stress (MPa)

778

Time (Hour)
Fig. 6. Compressive stress on the hollow asphalt concrete and corresponding
compressive strain between the base bars as well as water pressure (3.5 MPa) in the
chamber versus time before fracturing of the specimen.

loosened during the testing to allow the vertical force P act on the
top of the specimen. Then the water pressure in the chamber was
gradually increased up to 2.5 MPa in 2.0 h and simultaneously the
force P was increased to compensate the water uplift force in the
chamber to keep a constant average vertical compressive stress
of 1.55 MPa on the hollow asphalt concrete in the middle part of
the specimen. The force P was calculated by the following
equation:

P 10001:55S  Sc rw Sc 

where: P = the compressive force on the hollow asphalt concrete in

the middle part of the specimen, kN, S = area of the asphalt specimen, pD2/4, D = 150 mm, Sc = area of the chamber in the asphalt
specimen, pD2c =4, D = 40 mm, rw = water pressure in the chamber,
MPa.
After a period of 2 h the compressive strain between the base
bars was increased to 0.9% from 0.8% due to the creep effect of
asphalt concrete. Fig. 5 shows the loading process. Then the compressive stress of 1.55 MPa was reduced keeping a constant water
pressure of 2.5 MPa in the chamber until leakage was observed.
Test result shows that the conditions at fracturing of the specimen
was when the vertical compressive stress was reduced as low as
0.2 MPa and the resulted vertical compressive strain between the
base bars was 0.6% returned from 0.9%.
Similarly for the second test model, the compressive stress of
1.55 MPa was reduced keeping a constant water pressure of
3.5 MPa in the chamber until leakage was observed. Test result
shows that the conditions at fracturing of the specimen was when
the vertical compressive stress was reduced as low as 0.6 MPa and
the resulted vertical compressive strain between the base bars was
0.8% returned from 1.0% (see Fig. 6).
4. Finite element analysis of model test results under tension,
restraint and compression conditions
The nite element analysis was performed to investigate the
strain conditions for the model test at fracturing under each condition. According the structural characteristics of the model shown in
Fig. 1, the asphalt concrete model is discretized into an axisymmetric nite element mesh (see Fig. 7). For the middle part of the
model smaller 8-noded elements with a size of 5 mm (width) by
5 mm (height) were used and for the top and base parts elements
with 5 mm by 10 mm were used. The boundary condition on the
center of the axisymmetric model indicates zero x-displacement
which will allow vertical deformation only. The boundary conditions on the bottom side and right side covering a height of
70 mm (base-layer) from the bottom are zero x- and

Fig. 7. Finite element mesh for the axisymmetric asphalt concrete test model.

y-displacement which indicate a binding condition between the

asphalt specimen and the steel mold. The boundary condition on
the right side covering a height of 180 mm up to the top is zero
x-displacement which will allow the asphalt specimen sliding vertically only along the steel mold. The top side can be applied for
tension, restraint and compression conditions. The chamber in
the axisymmetric asphalt model can be applied for a specied
water pressure.
The analysis was simplied with a linear elastic model using the
software SIGMA/W in GeoStudio 2004. The equivalent moduli were
used in the nite element analyses that accounted for the effects of
loading rate and temperature on the stressstrain behavior, without having to model the whole creep process itself. The Poissons
ratio for the asphalt concrete was assumed equal to 0.45 based
on the extensive laboratory long-term creep test results presented
by Wang [9]. The average equivalent Youngs modulus at 10 C was
taken equal to 34 MPa based on the back-analyses of the measured
tensile strain of 0.9% between the base bars for the model test
under compression condition with water pressure of 2.5 MPa in
the chamber (see Fig. 5) and the strain of 1.0% with water pressure
of 3.5 MPa (see Fig. 6). The Youngs modulus at 5 C and at 20 C
were estimated equal to 50 MPa and 17 MPa, respectively, based
on the test results of temperature effects on the modulus [9].

779

Y. Zhang et al. / Construction and Building Materials 93 (2015) 775781

Fig. 8. Cracks were observed on the middle part of the asphalt specimen.

0
0

01
-0 .

0
0.
01

0.02

-0.02

-0.0
2

(a)

-0 0
.0
1

1
.0
-0

(b)

When the cylindrical steel mold was opened after leakage was
observed, the horizontal cracks were observed on middle part of
the asphalt specimen (see Fig. 8).
Fig. 9 shows the typical calculated strain contours for the model
test under tension conditions with a tensile stain of 1.2% between
the base bars and with a water pressure of 0.5 MPa in the chamber
and at 10 C. From the asphalt concrete behavior characteristics
and the calculated results for the model tests with nite element
method it is judged that the water leakage was caused by the
cracking in the inner wall due to large tensile strains. A summary
of the test conditions and the calculated stains at the middle point
A of the inner wall of the asphalt specimen at fracturing are given
in Table 3.
Table 3 shows that the calculated loop tensile strains at the
middle point A of the inner wall of the asphalt specimen at fracturing under different conditions were equal to or larger than the
tested direct tension creep strain at cracking of 2.7%. It may be concluded that the fracturing of asphalt specimen in the model tests
was virtually caused by the cracking due to large tensile strains
under different conditions rather than the so-called hydraulic
fracturing. The tensile strains at fracturing of the asphalt specimen were affected by the imposed stress state. Table 3 shows that
the calculated loop tensile strains at fracturing of the specimen
under the compression conditions were around doubled that under
the tension and restraint conditions. Triaxial compression tests on
the asphalt concrete specimen also yielded large loop tensile
strains. During the triaxial testing and when cracks opened up in
the specimen face the pressure water (conning stress) could penetrate the cracks and cause asphalt concrete to fail [8]. Triaxial test
results of the asphalt specimen show that the axial strains at failure are in a range of 825% under a range of conning water pressure of 0.52.5 MPa. The calculated loop tensile strains in the outer
part wall of the triaxial specimen are in a range of 3.611.3%.

5. Discussions and practical implications

(c)

Fig. 9. Calculated (a) loop strain contours, (b) vertical strain contours and (c) radial
strain contours for the asphalt specimen at fracturing under tension conditions with
a tensile stain of 1.2% between the base bars and a water pressure of 0.5 MPa in the
chamber and at 10 C. Minus strain in the gure means tensile strain.

In order to evaluate the tensile strains at fracturing of the

asphalt specimen direction tension creep test of the asphalt concrete was carried out at 10 C. The specimen size was 40 by 40
by 220 mm3. The initial tensile stress on the specimen was
0.25 MPa and increased by 0.125 MPa per day until cracks
occurred. The average tensile strain of three specimens at cracking
was 2.7%.

Asphalt core embankment dams are constructed in layers and

the density of asphalt core is larger than that of adjacent transition
zones of gravel and rocklls. The core has a trend to settle more
than the transition zones due to its low equivalent modulus of
elasticity which may result in a suspension or so-called hanging
effect of the core on the more rigid transition zones. Field measurements have shown a maximum reduction of 36% for the vertical stress in the core of the 125 m high Yele Dam which rests on a
deep alluvial overburden [3]. Three-dimensional static analysis
showed a calculated maximum reduction of 37% for the vertical
stress in the core of the 171 m high Quxue Dam with a steep abutment of 1V:0.33H [10]. Even though the vertical stresses in the
core were signicantly reduced the vertical stresses were still at
least 50% bigger than the water pressure on the upstream face of
the core and vertical compressive strains were resulted for the
whole asphalt core wall.

Table 3
Calculated strains at the middle point A of the inner wall of the asphalt specimen at fracturing under different test conditions.
Tension
Temperature (C)
Water pressure (MPa)
Stress on specimen (MPa)
Strain between base bars (%)
Calculated strain at A point (%)

Note:

eh
ev
er

Restraint

Compression

5
0.5

10
0.5

20
0.5

5
3.1

10
1.2

20
0.7

1.0
3.0
1.1
3.4

1.2
3.8
1.4
4.4

1.4
6.3
1.7
6.3

0
4.7
0.3
5.5

0
2.7
0.2
3.1

0
3.1
0.2
3.7

eh , ev and er represent loop strain, vertical strain and radial strain, respectively. Positive strain in the table means tensile strain.

10
2.5
1.550.2
0.90.6
5.8
0.5
6.7

3.5
1.550.6
1.00.8
7.4
0.3
8.6

780

Y. Zhang et al. / Construction and Building Materials 93 (2015) 775781

Total 15 model tests were carried out to investigate the different conditions of hydraulic fracturing for asphalt concrete used
as impervious core in dams. The test results at different conditions
are veried each other and general conclusions may be drawn out.
The model tests under compression indicated that the average
vertical compressive stresses on the hollow asphalt concrete in
the middle part of the specimen were 0.2 MPa for water pressure
of 2.5 MPa in the chamber and 0.6 MPa for water pressure of
3.5 MPa at fracturing of the specimen. The stress conditions at fracturing of the specimen are easily obtained as:

rw 2:5ra 2:0

where: rw = water pressure in the chamber in the asphalt concrete,

MPa, ra = average vertical compressive stress on the hollow asphalt
concrete in the middle part of the asphalt specimen, MPa.
It should be noted that the horizontal deection of asphalt core
along dam axial direction is very small and the so-call arching
effect caused by the steep abutments in a narrow valley may be
negligible on the asphalt core as asphalt concrete presents a significant stress relaxation behavior [3,6]. However, the calculated loop
tensile strains in the inner wall of the asphalt specimen are relatively large (67%, Table 3) for the model tests under the Eq. (2)
compression stress conditions. If the loop tensile strains in the
inner wall of the asphalt specimen were not so large the Eq. (2)
stress condition would not stand. If there were no tensile strains
in asphalt concrete face even the water pressure was so high up
to 7.5 MPa or 11.5 MPa on the face there would be no water leakage occurrence [1,2]. Test results show there was no water leakage
occurrence even for an only 2 cm thick asphalt concrete slice with
a water pressure of 1.0 MPa on the slice face [6]. Therefore, the
conditions of the asphalt specimen at fracturing under tension,
restraint and compression are extremely impossible for an asphalt
core in embankment dams in reality.
Asphalt concrete used as impervious core in embankment dams
consists of more ne aggregates, ller and bitumen. Compared
with a low viscosity of uid water in the earth core with void content of 1020%, the viscosity of the bitumen in the asphalt concrete
of the core is so high and there are almost no entrances for water to
penetrate in the impervious core with air void content of less than
3% and generate a pore pressure. The fracturing of the asphalt specimen in the model tests was virtually caused by the cracking due to
large tensile strains under different conditions rather than the
so-called hydraulic fracturing. The tensile strains at cracking of
asphalt concrete are large and are one or two orders of magnitude
larger than the corresponding cracking strains for compacted earth
materials used in the earth cores [11]. Asphalt core is designed to
be exible and can adjust to the deformations of the embankment
ll without cracking during construction, impounding, and operation. Therefore, the so-call hydraulic fracturing may be excluded
in the case of asphalt concrete used as impervious core in embankment dams in normal conditions.
The conclusions are valid strictly for the test results of the
asphalt concrete and model test conditions described above. The
prerequisites for excluding the so-called hydraulic fracturing
for the asphalt core in dams are that the air void content is kept
less than 3% and the induced tensile strains less than the tolerable
strains of the asphalt concrete materials. Most eld measurements
and laboratory long-term creep test results documented that the
initial air void content of asphalt concrete was double reduced in
application conditions [12]. However, the eld measurements for
the Finstertal Dam with an inclined asphalt core showed some
dilations or air void content increases for the asphalt core with a
bitumen content of around 6.3% [13]. The model test carried out
by Pircher and Schwab revealed that only the boundary zones of
the core wall will be affected by dilating, while the center will
maintain its initial low porosity and required imperviousness.

To avoid the dilation or the air void content increase of asphalt core
due to construction imperfections, reservoir impounding and dam
settlements, the bitumen content is more than 6.5% and the minimum thickness of the asphalt core is 60 cm in modern asphalt core
engineering.
For very high embankment dams in narrow valleys like the
171 m high Quxue Dam in China the connection of the core with
the plinth on the steep abutments is crucial. To prevent the connection from cracking and hydraulic fracturing due to large shear
deformations specic design measures are taken. In addition to
double widening the thickness of the core at the connection, two
copper water-stop strips are put at the connection and thick sandy
asphalt mastic layer (3 cm) with modied bitumen by adding 4%
SBS (Styrene-Butadiene-Styrene) is placed between the core and
the plinth. Furthermore, 0.2% more bitumen content is added in
the core for a 10 m horizontal length from the plinth at the steep
abutment. The joint model test results for the asphalt core of
171 m high Quxue Dam will be presented in a separate paper.

6. Summary and conclusions

A cylindrical test model for asphalt concrete used as impervious
core in embankment dams was developed in an attempt to investigate the conditions of asphalt concrete at fracturing duo to water
pressure. The asphalt concrete models were tested under tension,
restraint and compression conditions and the model test results
were analyzed by applying nite element method. The conditions
for an asphalt core at fracturing due to water pressure in embankment dams were discussed and the following conclusions are
provided:
 Asphalt concrete used as impervious core in embankment dams
is designed with air void content of less than 3% and there are
almost no entrances for water to penetrate the core and generate a pore pressure. Therefore, the main prerequisites for
hydraulic fracturing are non-existent in the case of impervious
asphalt core.
 Analyses on the asphalt concrete model test results under tension, restraint and compression conditions corresponding to
extreme conditions of an asphalt core in dams showed that
the fracturing of the asphalt specimen in the model tests were
caused by large tensile strains rather than so-call hydraulic
fracturing.
 If an impervious asphalt core in embankment dams is designed
to be exible and the resulting tensile strains in the core under
application conditions are lower than the strains that the
asphalt concrete can undertake without cracking, the asphalt
core will be in safe conditions and the so-call hydraulic fracturing should be excluded.
For very high embankment dams in narrow valleys like the
171 m high Quxue Dam in China, the asphalt core and the connection of the core with the plinth on the steep abutments should be
carefully designed to avoid the dilation of the asphalt core and to
strengthen the cracking resistance of the connection.

Acknowledgments
The research work for the paper was supported by the National
Natural Science Foundation of China (No. 51179155) and by
Program 2013KCT-015 for Shaanxi Provincial Innovative Research
Team. The authors appreciate the assistances of Chao Lu, Liyan
Wu and Danni Wang of the Xian University of Technology,
China, during the experimental work.

Y. Zhang et al. / Construction and Building Materials 93 (2015) 775781

References
[1] ICOLD. Bituminous cores for ll dams. Bulletin 84. International Commission
on Large Dams, Paris, France, 1992.
[2] Schnian E. The shell bitumen hydraulic engineering handbook. Shell
International Petroleum Company Ltd.; 1999.
[3] Wang W, Heg K, Zhang Y. Design and performance of the Yele asphalt-core
rockll dam. Can Geotech J 2010;47(12):136581.
[4] Wang W, Heg K. Developments in the design and construction of asphalt core
dams. Int J Hydropower Dams 2010;17(3):8391.
[5] Wang W, Heg K. Cyclic behavior of asphalt concrete used as impervious core
in embankment dams. J Geotech Geoenvironmen Eng (ASCE)
2011;137(5):53644.
[6] Zhang Y, Heg K, Wang W, Zhu Y. Watertightness, cracking-resistance and selfhealing of asphalt concrete used as a water barrier in dams. Can Geotech J
2013;50(3):27587.
[7] W. Wang, K. Heg, Y. Zhang, Y. Zhu The asphalt core embankment dam: an
attractive alternative. In: International Symposium on Dams for a Changing

[8]
[9]

[10]

[11]
[12]
[13]

781

World-Need for Knowledge Transfer across the Generations & the World,
Kyoto, Japan, 2012, pp. 47/4-12.
Wang W, Heg K. Method of compaction has signicant effects on stress-strain
behavior of hydraulic asphalt concrete. J Test Eval 2009;37(3):26474.
W. Wang, K. Heg, A simplied performance model for asphalt concrete used
as an impervious core in dams under static loading. Canadian Geotechnical
Journal, (submitted).
Y. Chen, S. Li, W. Wang, Three-dimensional static analysis on the Quxue
asphalt-core rockll dam. Xian University of Technology, Report, 2008 (in
Chinese).
Leonards GA, Narain J. Flexibility of clay and cracking of earth dams. J Soil
Mech Found Div (ASCE) 1963;89(2):4798.
W. Wang, Research on the suitability of asphalt concrete as water barrier in
dams and dikes [Doctoral thesis]. University of Oslo, Norway, 2008.
W. Pircher, H. Schwab, Design, construction and behaviour of the asphaltic
concrete core wall of the Finstertal dam R.49-Q.61, Transactions, 16th
International Congress on Large Dams, ICOLD Press (1988), Paris, pp. 901924.