See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/268589293

The Deformation Behavior of Muratli Asphalt Faced Rockfill Dam

Conference Paper in Geotechnical Special Publication · March 2012

DOI: 10.1061/9780784412121.055

CITATIONS READS
3 1,855

3 authors, including:

Y. S. Ünsever Mehmet Yener Özkan

Uludag University Middle East Technical University
24 PUBLICATIONS 122 CITATIONS 37 PUBLICATIONS 452 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Y. S. Ünsever on 05 May 2015.

The user has requested enhancement of the downloaded file.

GeoCongress 2012 © ASCE 2012 526

The Deformation Behavior of Muratli Asphalt Faced Rockfill Dam

Y. S. Ünsever1, M. Y. Özkan2 and S. G. Yıldız3

1
Department of Civil Engineering, Middle East Technical University, P.O. Box
06800, Ankara, TURKEY; PH (90312) 2105486; FAX (90312) 2105401; email:
[email protected]
2
Department of Civil Engineering, Middle East Technical University, P.O. Box
06800, Ankara, TURKEY; PH (90312) 2102414; FAX (90312) 2105401; email:
[email protected]
3
Department of Civil Engineering, Middle East Technical University, P.O. Box
06800, Ankara, TURKEY; PH (90312) 2102414; FAX (90312) 2105401; email:
[email protected]

ABSTRACT

The deformation behavior of Muratli Dam, the first asphalt faced rockfill dam
in Turkey, is investigated for the “end of construction” and “reservoir impoundment”
loading conditions. Two dimensional plane strain finite element analyses are carried
out in order to assess the total stresses, displacements and pore water pressures.
Hardening soil model is used to represent the non-linear, inelastic and stress
dependent behavior of rockfill material. Material model parameters are selected
mainly referring to the previous studies on the dams consisting of similar materials
and then back analyses are performed to determine the parameters which represent
the behavior of the dam. Calculated stresses and displacements by utilizing these
parameters are compared with the observed values for both end of construction and
reservoir filling conditions. It is observed that the overall agreement is quite
encouraging. It can be stated that two dimensional finite element analyses may be
used to predict displacements and stresses with a reasonable agreement.

INTRODUCTION

Rockfill dams with impervious face have been preferred since late 1960’s due
to significant reduction in leakage rate and post-construction deformation (Hunter and
Fell, 2003). Although impervious faced rockfill dams have become popular in recent
years, the design of these structures is largely based on past experience rather than
theory (Cooke, 1984). The study on such type of dams has a significant importance to
investigate the dam safety and define new design criteria for dams which will be
constructed in the future. Finite element method is used to assess the stresses and
deformations in earth and rockfill dams, where the soil behavior is modeled by
nonlinear, inelastic models.
In this study, stresses and deformations both for end of construction and for
full reservoir loading conditions obtained by the finite element (FE) analyses of
Muratli Dam (first asphalt faced rockfill dam in Turkey) are compared with those of
actual observations (Ünsever, 2007). Due to the fact that the dam has a large
length/height ratio, 2D FE analyses are carried out by using Plaxis v7.2 program and
GeoCongress 2012 © ASCE 2012 527

utilizing Hardening Soil Model to capture the nonlinear, inelastic and stress
dependent behavior of rockfill material (Schanz et al., 1999).

GENERAL PROPERTIES OF MURATLI DAM

Muratli Dam is the first asphalt faced rockfill dam in Turkey and it is located
in East Black Sea Region on Çoruh River. Height of the dam is 42 m from the river
bed and the slope of the dam is 2:1 (H:V) for both upstream and downstream sides.
Volume of the dam body is 1,981 million m3 and crest length of the dam is 280 m.
Muratli Dam has 36 cm thick asphalt impervious face. This technique, applied
mainly on medium height dams, decreases the cost of the project by considerable
amount. Asphalt face provides excellent water-tightness, durability and flexibility
(Singh and Varshney, 1995).
Since the basic geological formation under the dam is approximately 60 m
thick alluvium lying on rock formation, a cutoff wall is constructed down to the rock
surface. Cutoff wall is constructed of impermeable plastic concrete and is designed
with compatible deformability with the subsoil (alluvium). Main rock beneath the
subsoil is agglomerate and lithic-andesite tuff which is hard and sound. The cross
section of the dam is given in Figure 1 with material zoning (all the dimensions are in
meters). The legend for the materials and detailed information about the main zones
of the dam are presented in Table 1.

Instrumentation:
Muratli Dam is heavily instrumented to observe the behavior of the dam. The
instruments used in the dam body are:
• Inclinometers (IC): Measure horizontal and vertical deformations in dam body and
subsoil (alluvium).
• Fill-extensometers (EW): Measure longitudinal displacement between two points in
fill such as horizontal and vertical deformations behind the asphalt face, vertical
deformations at upstream and downstream of cutoff wall and horizontal deformations
along the dam body.
• Earthcells (EC): Measure earth pressures (total normal stresses) in dam body.
• Piezometers (PW): Measure excess pore water pressures to check the cutoff wall’s
impermeability.
In the dam body and cofferdam, 48 fill-extensometers, 4 inclinometers, 51
piezometers and 34 earthcells are placed at two different cross sections (Section A1
and A2). Figure 2 shows the placement of fill-extensometers and Figure 3 illustrates
the piezometers, inclinometers and earthcells locations at cross section A1 (all the
dimensions are in meters).

Table 1. Materials Used in the Construction of Muratli Dam.

Zone Type
1 River Alluvium
2A River Alluvium or Rockfill (fine material)
2B River Alluvium or Rockfill (coarse material)
3A Rock Material (Type 1)
GeoCongress 2012 © ASCE 2012 528

Table 1. (Continued)
Zone Type
3B Rock Material (Type 2)
3C Random Rockfill
4 Impervious Material
5 Filter Material
6 Riprap
7 Surface Protection Layer

Figure 1. Typical Cross Section and Zoning of Muratli Dam.

Figure 2. Location of Fill-extensometers at Section A1.

GeoCongress 2012 © ASCE 2012 529

Figure 3. Locations of Piezometers, Inclinometers and Earthcells at Section A1.

FINITE ELEMENT ANALYSIS OF MURATLI DAM

As mentioned above, Muratli Dam is constructed as a compacted rockfill dam

having nonlinear, stress dependent and inelastic behavior. This behavior of rockfill
material can be idealized by hardening model which is based on plastic theory rather
than elastic theory (Schanz et al., 1999).
In this study, Plaxis v7.2 software is used with hardening model to reflect the
behavior of Muratli Dam. Parameters for hardening model such as Failure Ratio, Rf,
Stress Dependency Parameter, m, and Reference Stiffness Modulus, E50ref, are
estimated on the basis of previous studies made for Cethana Dam (Khalid et al., 1990)
and for Kürtün Dam (Özkuzukıran et al., 2006) at the start of the study and back
analyses are performed to assess the best fit with the measurements since no
laboratory test results on dam’s fill materials are available. Based on laboratory
experiments, the subsoil is alluvium mainly consisting of boulder sized particles.
Parameters of subsoil are determined (utilized) from test results: Unit weight, γ = 22
kN/m3, Cohesion, c = 0 kN/m2 and Friction angle, φ = 350, whereas hardening
parameters for subsoil are assessed from the back analyses. Rock beneath the subsoil
is assumed as rigid and parameters for rock are selected according to this assumption.
After the preliminary analyses, the range for the stress strain parameters for
each material is selected (Table 2). Cohesion, c, for both rockfill and alluvium is
taken as “1.0 kPa” in order to provide stability for the numerical calculations.
Dilatancy angle , Ψ, can be estimated by the formula Ψ ≈ φ - 300 and Reference
Oedometer Modulus, Eoedref, is taken equal to the E50ref (Plaxis Material Manual,
2004).
Elastic beam elements are used in the model for both the cutoff wall and the
asphalt face. Cutoff wall is designed to deform compatibly with the subsoil.
Generally, cutoff wall has a modulus of elasticity of 4-5 times the subsoil modulus
GeoCongress 2012 © ASCE 2012 530

(ICOLD, 1985). Asphalt face material properties are taken as Young’s modulus,
E = 1.0x107 kPa and Poisons ratio, υ = 0.15 based on the studies of Lollino et al.,
2005.

Table 2. Ranges of Stress Strain Parameters for the Model.

Material c (kPa) φ (0) E50ref (kPa) m Rf γ (kN/m3)
Rockfill 1.0 35 - 40 60000 - 80000 0 - 0.3 0.75 17 - 20
Alluvium 1.0 32 - 38 30000 - 50000 0 - 0.3 0.75 17 - 22
Subsoil 1.0 35 200000 - 400000 0 - 0.3 0.75 22

Loading conditions at Muratli Dam can be handled in two parts: the end of
construction (EOC) condition and the reservoir impounding (RI) condition. In the
first case, the dam deformed under its own weight and in the second case, the
deformation is due to the application of water load on the impervious face. In finite
element analyses, stage of construction has a significant effect on stress distributions,
horizontal deformations and especially on vertical deformations (Clough et al., 1967).
Therefore, analyses are carried out in stages to reflect the construction condition. The
finite element mesh with 15-node elements used in the analyses for cross section A1
is given in Figure 4.

Figure 4. Finite Element Mesh Used in the Analyses.

RESULTS OF THE ANALYSIS

End of Construction Case:

For the end of construction case, horizontal and vertical extensometer
measurements are compared with the calculated displacements. The comparison of
computed and observed horizontal deformations for cross section A1 and vertical
deformations for cross sections A1 and A2 are given in Figure 5. It is seen that, most
of the points are within the region bounded by the lines having slopes 2:1 and 1:2
(H:V).
Comparison of the observed and calculated settlements could not be done as
inclinometer readings were not available. Inclinometers were not working due
misplacement. In the dam body, maximum settlement is calculated at about mid-
height and midsection of the dam as 110 mm and 145 mm for section A1 and A2,
respectively.
Total stresses in Muratli Dam have been also measured with the help of
earthcells. All analyses are based on two dimensional assumption, therefore stress
GeoCongress 2012 © ASCE 2012 531

reduction factors should be used to account for the valley shape. However, for
Muratli Dam both stress reduction factors are “1.0” since river width to height ratio is
greater than 1.0 (Hunter and Fell, 2003). In Figure 6, calculated and observed total
stresses for cross section A1 are shown for EOC case. Almost all of the data fall in
the region defined by lines with slopes 2:1 and 1:2. But it is seen that some of the
earthcell readings are really smaller than the calculated ones which is assumed to be
arching effect, a common consideration in earthcell placement. Therefore, the contact
between the instrument and the earth fill during placement might have been poor.

Reservoir Impoundment Case:

Following the construction of an impervious faced rockfill dam, large
deformations occur during first reservoir impoundment period since water load acts
on the surface of the dam. Major part of the deformations due to reservoir filling
occurs at upstream part of the body. Therefore, deformations that occur during this
period affect the performance of the impervious face. Also, possible cracks which
lead to leakage in the dam body may occur as a result of these deformations if the
face cannot tolerate the excessive deformation (Khalid et al., 1990).
A comparison of calculated results and observed values from the vertical and
horizontal fill-extensometers for cross section A1 are summarized for RI condition in
Figure 7. The agreement between the observed and calculated values is generally
acceptable; however at some locations, there exists significant differences between
the measured and calculated values. It is assumed that this might be due to the
inhomogeneity of the alluvium at those locations or due to calibration errors of the
instruments. In RI analyses, maximum additional displacement due to reservoir
impoundment is calculated as 70 mm near the cutoff wall.
The earthcell measurements and calculation results are also in good
agreement. It can be concluded that total stresses increase with reservoir
impoundment as a result of water load applied on the impervious face. The effect of
the water load can be seen clearly at earthcells located at upstream region. However,
this impact disappears towards the downstream region.

Figure 5. Observed and Calculated Displacements for EOC Condition.

GeoCongress 2012 © ASCE 2012 532

Figure 6. Observed and Calculated Total Stresses for EOC Condition.

Figure 7. Observed and Calculated Displacements for RI Condition.

Calculated Stress and Displacement Contours:

Displacement and stress contours at section A1 are plotted for both end of
construction and reservoir impoundment loading conditions. Figures from 8 to 10 and
Figures from 11 to 13 depict end of construction and reservoir impoundment
contours, respectively. For vertical and horizontal displacements, negative values
indicate settlement and upstream movement, respectively, where negative values of
stresses indicate compression.
For the end of construction stage, the shear stress and vertical displacement
contours in the upstream part and in the downstream part are almost symmetrical,
whereas some discrepancies are observed for horizontal displacement contours.
Indeed, from Figure 8, it can be seen that the movement at upper part of the dam is
towards upstream direction whereas the lower part of the dam and subsoil move
towards downstream direction. From Figure 9, it can be seen that maximum
settlement occurs at about mid-height of the dam as expected. As it is mentioned
above, the settlement of rockfill dams range between 0.25% and 1.00% of the dam
height (Singh and Varshney, 1995). In Muratli Dam, maximum settlement is
GeoCongress 2012 © ASCE 2012 533

calculated to be 0.29% of dam height for section A1, which is in accordance with
Singh and Varshney’s suggestion. From Figure 10, it can be seen that the shear
stresses at dam axes are zero and increase significantly towards the faces. Maximum
shear stresses at upstream and downstream regions are 143 kPa and 93 kPa,
respectively.

Figure 8. Horizontal Displacements for EOC (in mm).

Figure 9. Vertical Displacements for EOC (in mm).

Figure 10. Shear Stresses for EOC (in kPa).

When horizontal displacements at RI condition are considered, it is seen that,

the whole dam body and a part of the subsoil move towards downstream, as expected.
Horizontal displacements are concentrated at upstream face and they decrease
towards the downstream (Figure 11). Water load also causes vertical displacement of
GeoCongress 2012 © ASCE 2012 534

the upstream face, whereas downstream half of the dam body is not significantly
affected from impounding (Figure 12). Maximum horizontal and vertical
displacement is 50 mm and 70 mm, respectively.
When reservoir is filled with water, water load pushes the dam body towards
the downstream direction as shown in Figure 13. As a result, negative shear stresses
at upstream decrease considerably and positive shear stresses at downstream increase
significantly. Also, it is observed that shear stress increments are concentrated near
the dam axis.

Figure 11. Horizontal Displacements for RI (in mm).

Figure 12. Vertical Displacements for RI (in mm).

Figure 13. Shear Stresses for RI (in kPa).

CONCLUSION

In this study, deformation behavior of Muratli Dam, the first asphalt faced
rockfill dam in Turkey is investigated by two-dimensional finite element program
Plaxis v7.2. Nonlinear, stress dependent and inelastic behavior of the fill materials are
incorporated into the hardening soil model to assess the displacements and stresses
 GeoCongress 2012 © ASCE 2012 535

which are compared with the measured values for both end of construction and
reservoir impoundment conditions.
In general, measurements and calculation results are in acceptable range. It
can be said that two-dimensional finite element analyses may yield displacement and
stresses in a range of 50% to 200% of the observed ones. It may have been possible to
predict the stresses and displacements more realistically should there be sufficient
experimental data available.
The study indicates that the locations of the displacement and stress
measuring instruments are generally suitable for the purpose. Maximum displacement
values found at the dam center cannot be compared with the related instrument
readings since inclinometers at these locations have not operated properly. The
malfunctioning of inclinometer casings is attributed to the improper installation and
lack of experience of the maintenance team on these devices. Therefore, maintenance
of instruments during and after the construction is a vital issue.
For further studies, a 3-D analysis can be carried out to understand the effect
of the third dimension.

REFERENCES

Clough, R. W., and Woodward, R. J. (1967). “Analysis of Embankment Stresses and

Deformations.” Soil Mech. and Found. Div. J., ASCE, 93, SM4, 529-549.
Cooke, J. B., (1984). “Progress in Rockfill Dams.” (18th Terzaghi Lecture),
Geotech. Engrg. J., ASCE, 110, 10, 1381-1414.
Hunter, G., and Fell, R. (2003). “Rockfill Modulus and Settlement of Concrete Face
Rockfill Dams.” Geotech. Geoenv. Engrg. J., ASCE, 129, 10, 909-917.
ICOLD (1985). “Filling Materials for Watertight Cutoff Walls”, Bulletin 51,
International Committee on Large Dams.
Khalid, S., Singh, B., Nayak, G. C., and Jain, O. P. (1990). “Nonlinear Analysis of
Concrete Face Rockfill Dam.” Geotech. Engrg. J., ASCE, 116, 5, 822-837.
Lollino, P., Cotecchia, F., Zdravkovic, L., and Potts, D. M. (2005). “Numerical
Analysis and Monitoring of Pappadai Dam.” Can. Geotech. J., 42, 1631-1643.
Özkuzukıran, S., Özkan, M. Y., Özyazıcıoğlu, M., and Yıldız, G. S. (2006).
“Settlement Behavior of a Concrete Faced Rock-fill Dam.” Geotech. and
Geolog. Engrg. J., 24, 6, 1665-1678.
Plaxis v. 7.2 - Material Models Manual. (2004).
Schanz, T., Vermeer, P. A., and Bonnier, P. G. (1999). “The Hardening Soil Model:
Formulation and Verification.” Beyond 2000 in Computational Geotechnics-
10 Years of Plaxis, Balkema, Rotterdam.
Singh, B., and Varshney, R. S. (1995). “Engineering For Embankment Dams.” A.A.
Balkema Publishers, Brookfield.
Ünsever, Y. S. (2007). “An Analysis of Deformation Behavior of Muratli Asphalt
Faced Rockfill Dam.” MSc Thesis, Middle East Technical University, Ankara,
Turkey.

View publication stats