COMPARISON OF EVOLUTIONS OF INTERNAL EROSION IN HOMOGENEOUS EARTH-FILL DAMS BUILT WITH MEDIUM SAND AND CLAY MIXTURE WHEN THE SEEPAGE IS AT UPPER- MIDDLE OR AT UPPER-CORNER PART OF THE DAM

Abstract

One of the most significant reasons for earth-fill dam failures is internal erosion resulting from piping. This research was carried out as a part of a project supported financially by the Scientific and Technological Research Council of Turkey (TÜBİTAK). This paper involves the comparison of the experimental findings related to two different scenarios to look into the breach process and to provide the data allowing the realization of more realistic numerical analyses. A circular tunnel of 2 cm diameter located 6 cm below the dam crest was created to induce the seepage. The experiments were conducted at Hydraulics Laboratory of Civil Engineering Department within İzmir University of Economics. The homogeneous earth-fill dams having a height of 0.60 m and a bottom width of 2 m were built in a flume 1.00 m wide, 0.81 m high and 5.44 m long. Some common soil mechanics tests were carried out before the dam was built. The dam bodies were constructed by using a mixture of 15 % clay and 85 % medium sand. High-precision cameras were used to record the temporal development of the breach resulting from the piping. The pump flow rate was measured by a magnetic flowmeter and the flow rate values outgoing from the breach were determined from the continuity equation. Gauss area formula was used to obtain the time-varied values of the breach areas. The temporal changes of water depth in the channel were also recorded. The so obtained experimental findings are presented and commented.

Full text

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
357
COMPARISON OF EVOLUTIONS OF INTERNAL EROSION IN HOMOGENEOUS EARTH-FILL
DAMS BUILT WITH MEDIUM SAND AND CLAY MIXTURE WHEN THE SEEPAGE IS AT UPPER-
MIDDLE OR AT UPPER-CORNER PART OF THE DAM
Prof. Dr. Mehmet Şükrü Güney, Ar. Gör. Merve Okan, Ar. Gör. Emre Dumlu, Yiğit Kalyoncu
ORCID: 0000-0003-1441-4784, 0000-0001-6095-2992, 0000-0003-4311-3040, 0000-0002-5955-963X
1sukru.guney@ieu.edu.tr, 2 merve.okan@ieu.edu.tr, 3emre@ncche.olemiss.edu,
4yigit.k@std.izmirekonomi.edu.tr
1,2,4 İzmir University of Economics, Faculty of Engineering, Civil Engineering Department, İzmir, Turkey
3 The University of Mississippi, National Center for Computational
Hydroscience and Engineering, Oxford, USA
Abstract
One of the most significant reasons for earth-fill dam failures is internal erosion resulting from piping. This
research was carried out as a part of a project supported financially by the Scientific and Technological
Research Council of Turkey (TÜBİTAK). This paper involves the comparison of the experimental findings
related to two different scenarios to look into the breach process and to provide the data allowing the realization
of more realistic numerical analyses. A circular tunnel of 2 cm diameter located 6 cm below the dam crest was
created to induce the seepage. The experiments were conducted at Hydraulics Laboratory of Civil Engineering
Department within İzmir University of Economics. The homogeneous earth-fill dams having a height of 0.60 m
and a bottom width of 2 m were built in a flume 1.00 m wide, 0.81 m high and 5.44 m long. Some common soil
mechanics tests were carried out before the dam was built. The dam bodies were constructed by using a
mixture of 15 % clay and 85 % medium sand. High-precision cameras were used to record the temporal
development of the breach resulting from the piping. The pump flow rate was measured by a magnetic
flowmeter and the flow rate values outgoing from the breach were determined from the continuity equation.
Gauss area formula was used to obtain the time-varied values of the breach areas. The temporal changes of
water depth in the channel were also recorded. The so obtained experimental findings are presented and
commented.
Keywords: Earth-fill dam; Homogeneous dam; Piping; Breach development; Discharge from breach.

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
358
ORTA DERECELİ KUM VE KİL KARIŞIMI İLE İNŞA EDİLEN HOMOJEN TOPRAK DOLGU
BARAJLARDA SIZMANIN ÜST-ORTA VEYA ÜST-KÖŞEDE OLMASI DURUMLARINDA OLUŞAN
İÇSEL EROZYONLARIN KARŞILAŞTIRILMASI
Özet
Toprak dolgu baraj yıkılmalarının en önemli nedenlerinden biri borulanmadan kaynaklanan içsel erozyondur.
Bu araştırma, Türkiye Bilimsel ve Teknolojik Araştırma Kurumu (TÜBİTAK) tarafından desteklenen 119M609
nolu projenin bir bölümü olarak gerçekleştirilmiştir. Bu çalışma, gediğin gelişme sürecini incelemek ve sayısal
analizler için daha gerçekçi veri sağlamak amacıyla oluşturulan farklı senaryolardan iki tanesi ile ilgili deneysel
bulguların karşılaştırmasını içermektedir. Sızma, baraj kretinin 6 cm aşağısında 2 cm çaplı tünel oluşturularak
başlatılmıştır. Deneyler İzmir Ekonomi Üniversitesi İnşaat Mühendisliği Bölümü Hidrolik Laboratuvarı’nda
gerçekleştirilmiştir. Taban genişliği 2 m ve yüksekliği 0,60 m olan homojen toprak dolgu barajlar 1.00 m
genişliğinde, 0.81 m yüksekliğinde ve 5.44 m uzunluğunda bir kanalda inşa edilmiştir. Baraj inşa edilmeden
evvel gerekli bazı zemin mekaniği deneyleri gerçekleştirilmiştir. Baraj gövdeleri %15 kil ve %85 orta dereceli
kumdan oluşan karışım kullanılarak inşa edilmiştir. Borulanmadan kaynaklanan gediğin zamana bağlı
gelişimini kaydetmek için yüksek hassasiyetli kameralar kullanılmıştır. Pompa debisi bir manyetik debimetre ile
ölçülmüş ve gedikten çıkan debi değerleri süreklilik denklemi kullanılarak belirlenmiştir. Gedik alanlarının
zamana bağlı değişimini elde etmek için Gauss alan formülü kullanılmıştır. Kanaldaki su derinliğinin zamansal
değişimleri de kaydedilmiştir. Elde edilen deneysel bulgular sunulmakta ve yorumlanmaktadır.
Anahtar Kelimeler: Toprak dolgulu baraj; Homojen baraj; Borulanma; Gedik gelişimi; Gedikten çıkan debi.
INTRODUCTION
Piping is a significant issue for earth-fill dam failures. In earthen constructions, especially in earth levees
and dams, soil erosion may originate through the foundation, the embankment, or from the embankment to the
foundation. Phases of this type of erosion include: a) initiation and continuation of erosion, b) progression to
form a pipe and c) formation of a breach (Fell et al., 2003). When investigating the failure of an embankment
dam experimentally and numerically, the evolution of the breach caused by the piping is a crucial factor. In the
literature, there have been many researches on dam failures, particularly those caused by overtopping, but there
have not been as many studies on dam failures caused by piping because it is difficult to investigate erosion and
conduct controlled experiments (Chen et al., 2019; Greco et al., 2008; Sharif et al., 2015). A two-dimensional
depth-averaged (2DH) numerical model was used by Greco et al. (2008) to simulate the development of a
breach in an earth-fill dam. According to Chen et al. (2019), 3541 dam breach accidents occurred between 1954

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
359
and 2018, and more than 30% of them resulted from piping. Sharif et al. (2015) built dams with varied
compaction rates in a laboratory flume, and then used image processing technique to assess depth, area, and
volume change due to erosion during the piping progress. Numerous researchers who do numerical studies
make simple assumptions about properties of breach and water flow through it. According to Morris et al.
(2008) more realistic techniques are needed regarding the breach mechanism, as well as the breach shape and
flow through the breach.
In the scope of the project 119M609, supported financially by the Scientific and Technological Research
Council of Turkey (TÜBİTAK), it is aimed to carry out the piping experiments on homogeneous earth fill dams
and also earthen dams with clay core. These experiments were performed with different scenarios according to
the dam type and the location of the weak zone. Some of the experimental findings were presented in different
scientific meetings (M. Sukru Guney et al., 2023; Mehmet Sukru Guney, Dumlu, Okan, & Tayfur, 2022;
Mehmet Sukru Guney, Dumlu, Okan, Bor, et al., 2022; Mehmet Sukru Guney, Okan, et al., 2022). In addition
to these submitted papers, two master theses were also prepared and submitted (Dumlu, 2022; Okan, 2022).
This paper involves the experimental results concerning the evolution of dam failure due to the seepage
in the case of two scenarios corresponding to the dam bodies constructed with medium sand-clay mixture
having D50 = 0.30 mm. The objective of this research is to conduct experiments to investigate the development
of dam failure resulted from the seepage in homogenous earth-fill dams to provide insights into the breach
mechanism and data to the relevant researchers who deal with numerical analyses and emergency action plans.
EXPERIMENTAL PROCEDURE
The dam was constructed in a rectangular flume 1.00 m wide, 0.81 m high, and 5.44 m long (Figure 1).
The bottom of the flume consists of sheet metal whereas its sides are made of tempered glass for the purpose of
obtaining good records from the cameras located at different locations. The lower channel serves as water
supplying container. The upper channel which is at the upstream part of the dam corresponds to the dam
reservoir. A pump equipped with a check-valve and regulating valve was used to provide water circulation in
the closed system.
The details of the characteristics of dams are given in Table 1. The dams were designed and constructed
with side slopes of 1 vertical to 1.5 horizontal. In both of the scenarios, a circular groove of 2 cm diameter,
aligned from upstream to downstream was created at 54 cm from the bottom of the homogeneous dam bodies,
in order to initiate the formation of the breach. While initial groove was located at the middle of dam body in
the upper-middle case, it was located at the corner of the dam body in the upper-corner case.

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
360
Figure 1: Experimental set-up.
Table 1: Details of the characteristics of dams.
Height (cm)
60
Bottom width (cm)
200
Crest width (cm)
20
Length (cm)
96
Side Slopes (Vertical:Horizontal)
1:1.5
Initial groove shape
Circular d= 2 cm
Dam type
Homogeneous
Before building the dam, some common soil mechanics experiments were conducted. 85 % medium
sand and 15 % clay were used to prepare the soil mixture of the dam body. Wet sieve and hydrometer analyses
were conducted to obtain the grain size distribution of the soil mixture.
The particle size distribution of the mixture is given in Figure 2a. According to this figure; D10=0.006
mm, D30= 0.075 mm, D50= 0.3 mm, and D60= 0.4 mm. The coefficients of uniformity and curvature were found
as Cu =66.7 and Cc =2.34, respectively. The soil was classified as Clayey Sand with a corresponding symbol

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
361
SC according to Unified Soil Classification System. The specific gravity for the mixture was 2.63. Specific
gravity for the soil passing No.200 sieve was obtained as 2.72. 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 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 mixture has a cohesion value of 15.33
kPa and an internal friction of 33.93⁰.
Proctor test was performed to obtain the maximum dry unit weight and the optimum water content
(Figure 2b). From Figure 2b, ϒdry,max= 1.8 g/cm3 and wopt = 12.5 %. In the experiments, the energy was reduced
by 50 % in order to facilitate the occurrence of piping. Hence, the number of blows applied for each layer for
proctor test was 13 instead of 25.
(a) (b)
Figure 2: Grain size distribution and dry density-water content relationship.
In the experiments, the bulk density of 2 g/cm3 was satisfied for each layer. Before the compaction, each
layer was 14 cm thick and after the compaction it was reduced to 10 cm.
Figure 3 shows some construction stages of the dam bodies.
(a) (b) (c)
Figure 3: Some construction stages: a) Spreading out of the mixture, b) Compacting, c) After compaction of the
second layer

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
362
During the compaction, the mixture was spread out evenly and compacted by using a plate and proctor
hammer. At the end of the construction, the L-shaped molds were removed and the excessive soil materials
were carefully trimmed by using a trowel.
A magnetic flowmeter was used to measure the flow rate. Six cameras were placed at different locations
to monitor the experiment and record the evolution of the dam failure. An electromagnetic sensor was utilized
to adjust the water level so that the pump starts and stops at pre-determined water depths in the channel.
EXPERIMENTAL FINDINGS
Figures 4, and 5 show the time-varied breach shapes as obtained by the cameras placed at downstream of
the dam, in the case of upper-middle and upper-corner, respectively. The time t=0 is the time at which the
initiation of the seepage occurs at downstream face.
(a) (b) (c) (d)
Figure 4: The temporal development of the breach at downstream in the case of the seepage at upper-middle a)
t=0, b) t= 127 s, c) t= 254 s, d) t= 380 s.
(a) (b) (c) (d)
Figure 5: The temporal development of the breach at downstream in the case of the seepage at upper-corner a)
t=0, b) t= 167 s, c) t=334 s, d) t=500 s.

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
363
Figures 6 and 7 show the time-varied breach shapes as obtained by the cameras placed at upstream of the
dam, in the case of upper-middle and upper-corner, respectively. The time t=0 is the time at which the initiation
of the seepage occurs at downstream face
(a) (b) (c) (d)
Figure 6: The temporal development of the breach at upstream in the case of the seepage at upper-middle a) t=0,
b) t=127 s c) t=254 s, d) t=380 s.
(a) (b) (c) (d)
Figure 7: The temporal development of the breach at upstream in the case of the seepage at upper-corner a) t=0,
b) t=400s, c) t= 800 s, d) t=1200 s.
Camera recordings were used to obtain the water depth values in the channel. The images recorded by
the upstream and downstream cameras were examined in order to assess the breach's geometry and determine
the change in its shape. The boundary coordinates of the breaches at the downstream and upstream sides were
obtained at Get-data Graph Digitizer 2.26 software along with the scaling of the images derived from the
records corresponding to a certain period. The temporal total breach areas were calculated by Gauss Area
formula. The discharge of water outgoing from the breach was determined by using the continuity equation Eq.
(1):

(1)
where  is the flow rate delivered by the pump,  is the discharge from the breach,  is the
storage in the dam reservoir during the time interval . The temporal water depths in the channel for different

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
364
scenarios are given in Figure 8a. The temporal discharges outgoing from the breach calculated by using Eq. (1)
for different scenarios are given in Figure 8b.
(a) (b)
Figure 8: a) The temporal water depths in the reservoir and b) time-varied discharge outgoing from the breach
for different scenarios.
The temporal variations of the total breach area at downstream and upstream are shown in Figure 9a and
9b, respectively. The comparative experimental findings are given in Table 2.
(a) (b)
Figure 9: Temporal variations of the total breach area a) at downstream and b) at upstream for different
scenarios.

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
365
Table 2: Comparison of the experimental findings for different scenarios.
Scenario number
Upper-middle
Upper-corner
Max. flow rate Qmax (L/s)
8.3
3.6
tQ= time to reach Qmax (s)
245
325
Max. breach area at downstream Bdmax (cm2)
2379
733
td=time to reach Bdmax (s)
370
1080
Max. breach area at upstream Bumax (cm2)
3129
1792
tu=time to reach Bumax (s)
520
1950
The graphs of the curves Q/Qmax = f(t/tQ), Bd/Bdmax = f(t/td) and Bu/Bumax = f(t/tu) are given in Figs 10a,
10b and 10c, respectively. Q, Bd and Bu denote time-dependent values of the discharge, the breach area at
downstream and breach area at upstream, respectively.
(a) (b) (c)
Figure 10: Graphs of the curves a) Q/Qmax = f(t/tQ), b) Bd/Bdmax = f(t/td) and c) Bu/Bdmax = f(t/tu).
RESULTS AND CONCLUSIONS
In all scenarios, the erosion started at downstream, developed and continued inward toward upstream.
As expected, the peak flow rates were found to be very small in the case of seepage located at the upper
parts of the dam. This finding was due to the small value of the initial water head over the breach causing small
flow velocities and a delay in the time of occurrence of the peak flow rates.

1. BİLSEL INTERNATIONAL GORDİON SCIENCE RESEARCHES CONGRESS, 29-30 SEPTEMBER 2023
https://bilselkongreleri.com/
366
The smallest breach areas were found in the case of the seepage at the upper-corner because of the
existence of the flume side preventing the development of the breach.
ACKNOWLEDGEMENTS
The authors thank the Scientific and Technological Research Council of Turkey (TUBITAK) for
supporting financially this study through the project 119M609.
REFERENCES
Chen, S. shui, Zhong, Q. ming, & Shen, G. ze. (2019). Numerical modeling of earthen dam breach due to piping failure.
Water Science and Engineering, 12(3), 169–178. https://doi.org/10.1016/j.wse.2019.08.001
Dumlu, E. (2022). Experimental and Numerical Investigation of the Evolution of Piping and Resulting Breach in Earth-
fill Dams. M.Sc. Thesis, İzmir Institute of Technology.
Fell, R., Wan, C. H., & Foster, M. (2003). Progress report on methods for estimating the probability of failure of
embankment dams by internal erosion and piping.
Greco, M., Pontillo, M., Iervolino, M., & Leopardi, A. (2008). 2DH numerical simulation of breach evolution in an earth
dam. River-Flow2008, 1, 661–667.
Guney, M. Sukru, Dumlu, E., & Okan, M. (2023). Experimental Study of the Evolution of the Breach and the Discharge
through the Breach resulting from Piping due to Seepage at the Mid-part of Earthfill Dam. International Journal of
Structural and Civil Engineering Research, 12(2), 43–51.
Guney, Mehmet Sukru, Dumlu, E., Okan, M., Bor, A., Aklik, P., & Tayfur, G. (2022). 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 Bottom. 7th International Conference on Civil Structural and Transportation Engineering (ICCSTE’22)
Niagara Falls, Canada – June 05-07, 2022, July, 115-1-115–118. https://doi.org/10.11159/iccste22.215
Guney, Mehmet Sukru, Dumlu, E., Okan, M., & Tayfur, G. (2022). Experimental Study of Breach Evolution and
Discharge Through Breach Resulting from Piping due to Seepage at the Upper Corner of in an Earth-Fill Dam. In S.
Elci & G. Bombar (Eds.), 14th International Conference on Hydroscience & Engineering (ICHE 2022) MAY 26-27,
2022 (Issue May, pp. 99–109). https://doi.org/10.11159/iccste22.215
Guney, Mehmet Sukru, Okan, M., Dumlu, E., Bor, A., Aklık, P., & Tayfur, G. (2022). 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. In M. Ortega-Sánchez (Ed.), 39th IAHR World Congress 19-24 June 2022 (Issue June, pp. 2735–
2743). https://doi.org/10.3850/IAHR-39WC2521711920221298
Morris, M., Hassan, M., Kortenhaus, A., Geisenhainer, P., Visser, P., & Zhu, Y. (2008). Modelling breach initiation and
growth. Flood Risk Management: Research and Practice, 581–591. https://doi.org/10.1201/9780203883020.ch67
Okan, M. (2022). Experimental and Numerical Investigation of Piping in Uniform Embankment Dam with Weak Layer at
the Upper Region. M.Sc. Thesis, İzmir Institute of Technology.
Sharif, Y. A., Elkholy, M., Hanif Chaudhry, M., & Imran, J. (2015). Experimental Study on the Piping Erosion Process in
Earthen Embankments. Journal of Hydraulic Engineering, 141(7), 04015012. https://doi.org/10.1061/(asce)hy.1943-
7900.0001019