Experimental Study of the Evolution of the Breach and the Discharge Through the Breach Resulting from Piping due to Seepage at the Earth-Fill Dam Top

Abstract

Internal erosion, also known as piping, is one of the most important causes of earth-fill dam breaks. Many researchers dealing with numerical analyses in this area make some simplified assumptions about the shape of the breach and the discharge of water flowing through the breach. This study was conducted in the scope of the project supported financially by the Scientific and Technological Research Council of Turkey and it consists of experimental study which aims to provide data needed to perform numerical analyses with more realistic approaches. A dam with a height of 0.6 m, a bottom width of 2 m and a crest width of 0.20 m was built in a flume 1 m wide, 0.81 m high and 6.14 m long. Before the construction of the dam, some common soil mechanics tests were carried out. The dam was constructed by using a mixture consisting of 85 % sand and 15 % clay. A circular tunnel with a diameter of 2 cm was created along the centerline at 6 cm below the dam crest. In the closed system, water was pumped from the lower reservoir to the upper channel. Six cameras located at different locations recorded the evolution of the dam failure. Gauss Area formula was applied to determine the time-varied of the breach areas at upstream and downstream sides. The discharge of water through the breach and average flow velocity were determined by using the continuity equation. The changes in water depth in the channel were also recorded.

Full text

Proceedings of the 39th IAHR World Congress
19-24 June 2022
Granada, Spain
DOI number
©2022 IAHR. Used with permission / ISSN-L 2521-7119
Experimental Study of the Evolution of the Breach and the Discharge Through the
Breach Resulting from Piping due to Seepage at the Earth-Fill Dam Top
Mehmet Sukru Guney(1), Merve Okan(1), Emre Dumlu(2), Aslı Bor(1), Pelin Aklık(1), Gökmen Tayfur(2)
(1) İzmir University of Economics, İzmir, Turkey,
sukru.guney@ieu.edu.tr, merve.okan@ieu.edu.tr, asli.turkben@ieu.edu.tr, pelin.aklik@ieu.edu.tr
(2) Izmir Institute of Technology, İzmir, Turkey,
emredumlu@iyte.edu.tr, gokmentayfur@iyte.edu.tr
Abstract
Internal erosion, also known as piping, is one of the most important causes of earth-fill dam breaks. Many
researchers dealing with numerical analyses in this area make some simplified assumptions about the shape
of the breach and the discharge of water flowing through the breach. This study was conducted in the scope of
the project supported financially by the Scientific and Technological Research Council of Turkey and it consists
of experimental study which aims to provide data needed to perform numerical analyses with more realistic
approaches. A dam with a height of 0.6 m, a bottom width of 2 m and a crest width of 0.20 m was built in a flume
1 m wide, 0.81 m high and 6.14 m long. Before the construction of the dam, some common soil mechanics tests
were carried out. The dam was constructed by using a mixture consisting of 85 % sand and 15 % clay. A circular
tunnel with a diameter of 2 cm was created along the centerline at 6 cm below the dam crest. In the closed
system, water was pumped from the lower reservoir to the upper channel. Six cameras located at different
locations recorded the evolution of the dam failure. Gauss Area formula was applied to determine the time-
varied of the breach areas at upstream and downstream sides. The discharge of water through the breach and
average flow velocity were determined by using the continuity equation. The changes in water depth in the
channel were also recorded.
Keywords: Earth-fill dam; Piping; Breach geometry; Breach development; Discharge through breach
1. INTRODUCTION
Piping is one of the main problems which threatens the stability of earth-fill dams. Soil erosion can be
experienced in earth structures, especially in earth dams and levees, through embankment, foundation or from
embankment to foundation. This kind of erosion can occur in three stages: a) initiation and continuation of
erosion, b) progression to form a pipe, and c) formation of a breach (Fell et al., 2003). The FP5 IMPACT
(Investigation of Extreme Flood Process and Uncertainty) European project (2001-2004) revealed the
assessment and reduction of risks from extreme flooding caused by natural events or failure of dams and water
defense structures (Zech et. al.,2007). Chen et al. (2019) pointed out that between 1954 and 2018, 3541 dam
breach accidents had occurred and more than 30% of them were due to piping. Sparmos homogenous dam in
Greece is one of the recent examples (Dounias, 2019).
The ICOLD Bulletin B164 (2013) had analyzed the internal erosion of existing dams, levees and dikes, and
their foundations. Greco et al. (2008) used a two-dimensional depth-averaged (2DH) numerical model to
simulate the evolution of a breach in an earth-fill dam. Sharif et al. (2015) constructed a dam in a laboratory
flume by using a mixture of sand, silt, and clay with different compaction rates and examined the changes in the
depth, area, and volume of erosion during the piping evaluation by utilizing an image processing technique.
Most of the researchers realizing numerical analyses make some simplified assumptions concerning shape of
a breach and discharge of water flowing through the breach. Morris et al. (2008) revealed that instead of
simplified approaches, more realistic approaches are required about the breach mechanism as well as the
breach geometry and flow through the breach. Dhiman and Patra (2018) investigated the influence of the soil
properties on the breaching process by performing 13 experiments in the hydraulic engineering laboratory of
National Institute of Technology, Rourkela, India. Further, a multivariable regression was performed using the
test data of 25 embankments to obtain the relationship between the breach parameters and the soil properties.
New nondimensional control variables, such as embankment soil factors, relative compaction effort, relative
particle size, and erodibility, were proposed for developing the relationships. Damme (2020) presented a
process-based breach widening relation for levees constructed of dilatant soils. The process-based relation was

Proceedings of the 39th IAHR World Congress
19-24 June 2022, Granada, Spain
©2022 IAHR. Used with permission / ISSN-L 2521-7119
derived from the weir flow equation and a process-based erosion equation. The breach widening relation can
account for the effects of variations in soil parameters.
The aim of this study is to realize experiments to study the evolution of dam failure resulting from piping
along the centerline at 6 cm below the dam crest of the earth-fill dam.
2. EXPERIMENTAL PROCEDURE
The dam was constructed in a rectangular flume (Figure 1) with a width of 1 m, height of 0.81 m and length
of 6.1 m. At the upper channel of the flume, a homogeneous dam possessing 0.6 m height, 2 m bottom width
and 0.20 m crest width was constructed with the slope at the upstream and downstream sides of 1:1.5. Water
was circulated between the lower reservoir and the upper channel by means of a pump.
Figure 1. Experimental flume.
Some soil mechanics experiments were carried out before building the dam. The soil mixture utilized at the
construction of the dam was prepared by using 85 % sand and 15 % clay. The grain- size distribution of the
mixture obtained from the wet sieve and hydrometer analyses is shown in Figure 2.
Figure 2. Grain size distribution of the dam material.
From Figure 2, some characteristic diameters were obtained as D10= 0.006 mm, D30= 0.057 mm, D50=
0.099 mm, and D60= 0.3 mm. The uniformity coefficient Cu equals 54.5 and the curvature coefficient Cc is
equal to 1.969.
The specific weight of the mixture was found to be as Gs = 2.63, from the test ASTM D854 – 14.
The permeability of the mixture was found as k= 4.66x10-4 cm/s from the falling head permeability
test.
From the direct shear test, it was found that the soil has a cohesion value of 15.33 kPa and an internal
friction angle of 33.93°.
According to the consolidation test results, the compression index (Cc), recompression index (Cr) and
swelling index (Cs) were found to be as 0.100, 0.009 and 0.007, respectively. The oedometric modulus of
deformation (Eoed) was obtained as 35714 kN/m².
0.00
10.00
20.00
30.00
40.00
50.00
60.00
70.00
80.00
90.00
100.00
0.00050.0050.050.5
Percent Passing, (%)
Diameter (mm)

Proceedings of the 39th IAHR World Congress
19-24 June 2022, Granada, Spain
In order to determine the water content, the standard proctor test (ASTM-698) was executed by applying
13 drops and the so obtained curve is plotted in Figure 3. The reason of applying reduced energy (13 drops
instead of 25) was to increase the probability of the piping occurrence. The maximum dry density and optimum
water content were obtained as ϒdrymax= 1.794 g/cm3 and wopt = 12.5 %. The void ratio (e) was calculated as
0.469.
Figure 3. Dry density - water content relationship.
In Figure 4, some construction stages and completed shape of the dam body are shown. The dam was
constructed at 6 layers, each layer being 10 cm high. After homogeneous placement of the soil material, the
mixture was compacted using a flat plate and a proctor hammer (Figure 4a). When the compaction was
completed, the molds were extracted, and then the sides of the dam were trimmed by using a trowel.
(a) (b)
(c) (d)
Figure 4. Some construction stages: (a) Compacting by Proctor Hammer, (b) After compaction of the first
layer, (c) After compaction of the last layer, (d) final shape.
1.74
1.76
1.78
1.8
10.00 11.25 12.50 13.75 15.00
Dry Density (gr/cm³)
Water Content (%)

Proceedings of the 39th IAHR World Congress
19-24 June 2022, Granada, Spain
©2022 IAHR. Used with permission / ISSN-L 2521-7119
The flow rate was measured by a magnetic flowmeter. The evolution of the dam failure was recorded by
six cameras placed at different locations. In order to adjust the water level, an electromagnetic sensor was
attached which starts and stops the pump when water depths in the channel were 0.540 m and 0.555 m,
respectively.
In order to generate the formation of the breach, a circular hole of 2 cm diameter lying from upstream to
downstream was created at 54 cm from the bottom of the dam body. The experiment was started when the
water in the flume reached this level and passed through the hole.
3. EXPERIMENTAL FINDINGS
The temporal developments of the breach recorded by the cameras located at downstream and upstream
of the dam are given in Figure 5 and Figure 6, respectively. The time t=0 corresponds to the starting of the
seepage.
(a) (b)
(c) (d)

Proceedings of the 39th IAHR World Congress
19-24 June 2022, Granada, Spain
(e) (f)
(g) (h)
Figure 5. The temporal development of the breach at downstream a) t=0, b) t=180, c) t=230 s, d) t=280 s,
e) t=340 s, f) t= 380 s, g) t=410 s, h) t= 570 s
(a) (b)

Proceedings of the 39th IAHR World Congress
19-24 June 2022, Granada, Spain
©2022 IAHR. Used with permission / ISSN-L 2521-7119
(c) (d)
(e) (f)
(g) (h)
Figure 26. The temporal development of the breach at upstream a) t=0, b) t=180, c) t=230 s, d) t=280 s,
e) t=340 s, f) t= 380 s, g) t=410 s, h) t= 570 s
The water depths in the channel were attained from the camera recordings. In order to evaluate the
shape of the breach and survey the changes in its geometry, the upstream and downstream cameras
images were investigated. In order not to work with fisheye images, the videos taken from lateral camera
recording were edited and straightened with Hit-film-Express version 2021.1. Moreover, extra sensitive
solutions were implemented to ensure that the images are completely flat. The images taken from the
records corresponding to a certain time were scaled and the boundary coordinates of the breaches at
downstream and upstream sides were specified at Get-data Graph Digitizer 2.26 software. The surface
areas of the breach developed at different instants were computed by the Gauss Area functions.
The discharge of water through the breach was determined by using the continuity equation:
∆S=(Q
pump
-Q
breach
)∙∆t
[1]

Proceedings of the 39th IAHR World Congress
19-24 June 2022, Granada, Spain
where Qpump is the flow rate delivered by the pump, Qbreach is the discharge through the breach, ∆S is the storage
in the channel during the time interval ∆t.
The average velocity V of the flow through the breach was approximately calculated by using
V=
Q
breach
A
[2]
where A represents wetted area.
The temporal water depths in the channel and discharge through the breach calculated by Eq. [1] are given
in Figure 7 and Figure 8, respectively.
Figure 7. Time-varied water depths in channel.
Figure 8. Time-varied discharge through the breach.
The temporal variations of the breach area at downstream and upstream are shown in the Figure 9.
Figure 9. Temporal variations of the breach area at downstream and upstream.
0
5
10
15
20
25
30
35
40
45
50
55
60
0 100 200 300 400 500 600
Water Level (cm)
Time (s)
0
500
1000
1500
2000
2500
3000
3500
0 100 200 300 400 500 600
Surface Area (cm²)
Time (s)
Downstream Upstream

Proceedings of the 39th IAHR World Congress
19-24 June 2022, Granada, Spain
©2022 IAHR. Used with permission / ISSN-L 2521-7119
The time dependent wetted area and velocity values obtained by using Eq. [2] at upstream and
downstream are given in Figure 10 and Figure 11, respectively.
(a) (b)
Figure 10. (a) Wetted area and (b) velocity values at upstream
(a) (b)
Figure 11. (a) Wetted area and (b) velocity values at downstream
4. RESULTS AND CONCLUSIONS
In this study, time-varied evolution of the breach resulting from the piping at the top of earth-fill dam was
analyzed. The discharges through the breach corresponding to different instants were calculated using the
continuity equation. The boundary coordinates of the breach surface areas and wet areas of the breach were
obtained by using the Get-Data Graph Digitizer, and the areas at each time were calculated by applying the
Gauss-area function of these obtained coordinates. The time dependent velocity values through the breach
areas were also calculated. During the experiment, the breach initiated on the downstream side and then
evolved towards to upstream side. The maximum discharge through the breach was calculated as Qbreach=8.31
L/s at t=240 s. The maximum breach surface area at the upstream was found to be Aups= 3128.7 cm2 at t=520
s, while at downstream Adown= 2379.3 cm2 at t=370 s and remained unchanged afterwards. The maximum
wetted areas were Awetted-ups=274 cm2 at t=250 s and Awetted-downs=225 cm2 at t=270 s. The maximum velocity
values through the breach were calculated as Vups=51.1 cm/s at t=190 s and Vdown=55.8 cm/s at t=230 s for
upstream and downstream, respectively. The pump was turned off at t=380 s, and then experiment was
terminated.

Proceedings of the 39th IAHR World Congress
19-24 June 2022, Granada, Spain
In addition to the experimental studies, the numerical analysis also continues to be performed by using the
software PLAXIS-3D. It is also aimed to give comments of these experimental findings in the light of the
numerical analysis results during the oral presentation.
5. ACKNOWLEDGEMENTS
The authors thank the Scientific and Technological Research Council of Turkey (TUBITAK) for
supporting financially this study through the project 119M609.
6. REFERENCES
Chen S., Zhong Q., and Shen G. (2019). Numerical modeling of earthen dam breach due to piping failure.
Water Sci. Eng., 12 (3), 169–178.
Damme M. (2020). An analytical process-based approach to predicting breach width in levees constructed
from dilatant soils. Natural Hazards, 101, 59–85.
Dhiman S. and Patra K. C. (2018). Experimental study of embankment breach based on its soil properties. ISH
J. Hydraul. Eng., no. December, 1–11, doi: 10.1080/09715010.2018.1474500.
Dounias, G. and Bardanis, M. (2019). The failure of homogeneous dams by internal erosion -The case of
Sparmos Dam, Greece. Sustainable and Safe Dams Around the World – Tournier, Bennett & Bibeau (Eds)
© 2019 Canadian Dam Association, ISBN 978-0-367-33422-2
Fell,R., Wan, C. H. and Foster, M. (2003). Progress report on methods for estimating the probability of failure
of embankment dams by internal erosion and piping. University of New South Wales, Sydney, Australia.
Greco, M., Pontillo, M., Iervolino, M., and Leopardi, A. (2008). 2DH numerical simulation of breach evolution in
an earth dam. River-flow2008, Vol. 1, M. S. Altinakar, M. A. Kökpınar, I. Aydın, S. Çokgör, and S. Kırkgöz,
eds., Kubaba, Ankara, Turkey, 661–667.
ICOLD Bulletin No.164. (2013). Internal erosion of existing dams, levees and dikes, and their foundations, Vol
1 & 2
Morris M. W., Hassan M., Kortenhaus A., Geisenhainer P., Visser P., and Zhu Y. (2008). Modeling breach
initiation and growth. Floodrisk, Oxford, UK, 30 September-2 October,2008
Sharif, Y. A., Elkholy, M., Hanif Chaudhry, M. and Imran, J. (2015). Experimental Study on the Piping Erosion
Process in Earthen Embankments. J. Hydraul. Eng., 141 (7), 04015012.
Zech, Y. and Soares-Frazão, S. (2007). Dam-break flow experiments and real-case data. A database from the
European IMPACT research, Journal of Hydraulic Research. 45(1), 5-7, DOI:
10.1080/00221686.2007.9521827