1. Introduction
The increasing demands for water security and the global transition toward renewable
energy underscore the critical importance of dam safety. Dams are indispensable infrastructure,
providing water for irrigation, drinking, and hydropower generation. However, as these
structures age and face extreme hydrological events, the risk of failure grows significantly.
Among various types, embankment dams are particularly susceptible to the safety issues
like overtopping and internal erosion (Kim et al., 2023; Lee and Choi, 2019).
Despite advancements in modern engineering and safety practices, dam failures continue
to occur in the 21st century. Notable examples include the Oroville Dam spillway incident,
the Taum Sauk Dam breach, and the Edenville Dam failure. The Oroville incident, while
not a complete dam failure, involved severe erosion of the main and emergency spillways
due to design flaws, construction defects, and geological weaknesses (France et al., 2018). The Taum Sauk Dam, a pumped storage hydroelectric facility, failed catastrophically
after its upper reservoir overtopped due to instrumentation malfunctions, leading
to extensive erosion and a complete breach (Rogers et al., 2010). Similarly, the Edenville Dam failure, caused by prolonged intense rainfall and historic
flood levels, demonstrated the risks associated with aging dams lacking adequate spillway
capacity, particularly in the context of climate change (France et al., 2022).
Learning from these incidents is imperative for improving dam safety management and
mitigating future risks.
Earth-cored embankment dams, characterized by a central clay core that serves as the
primary water barrier, are widely used due to their versatility and cost-effectiveness.
Five case studies of earth-cored embankment dams in Korea were performed in this study,
all of which experienced core material degradation leading to seepage-related safety
concerns. Through a detailed analysis of the dam safety incidents shown in this paper,
the study seeks to derive actionable lessons for enhancing dam safety protocols and
ensuring the long-term integrity of these vital structures.
This paper provides a concise summary of technical analyses of each dam incident including
geotechnical and geophysical investigations to infer potential causes and propose
remedial measures. It highlights construction material issues, inadequate compaction,
and arching effects during the initial filling process as key contributors to core
degradation. The study also evaluates low-pressure permeation grouting as an effective
remediation method for restoring the core's impermeability.
The findings of this study aim to serve as a resource for dam safety professionals,
offering insights into hazard mitigation technologies and informing strategies for
rehabilitating aging dams. The subsequent sections delve into the technical details
of the case studies, identify the causes of degradation, and discuss the implications
of the proposed solutions.
2. Case Studies of Dam Incidents
2.1 Dams in This Study
This study investigates the safety incidents at five earth-cored embankment dams in
Korea, denoted as W, A, D, Y, and H dam. A summary of their dimensions, construction
timelines, observed degradation, and remediation measures is provided in Table 1. Fig. 1 illustrates their typical cross- sections.
These dams, built with central clay cores serving as primary water barriers, exhibited
varying degrees of core material degradation. Observed issues included sinkholes,
wet zones, slope failures, and increased seepage rates. Initial remediation efforts,
such as compaction and grouting, were often insufficient, necessitating further intervention.
Ultimately, low-pressure permeation grouting was implemented in all cases as the definitive
remediation strategy to restore functionality and safety.
Fig. 1 Typical cross-section of dams in the study
Table 1 Comparative analysis of construction, safety incidents, and measures for five
dams
Dam
|
W
|
A
|
D
|
Y
|
H
|
Construction period
|
Dec. 1985 - Jun. 1993
|
Aug. 1968 -
Dec. 1971
|
Nov. 1986 -
May 1990
|
Dec. 1977 -
Dec. 1979
|
Built in Dec. 1984
Raised in Dec. 2004
|
H (m)
|
55
|
32.5
|
55
|
24.5
|
35
|
L (m)
|
407
|
223.5
|
326
|
120
|
315
|
Reservoir capacity (million m3)
|
135
|
18
|
8
|
5
|
0.9
|
Incidents
|
Apr.-Oct. 1998
Three sinkholes found near crest,
locally fluidized cores found,
excessive leakage
|
Jun. 1985
Wet zone found on the downstream slope,
Jul. 2003 Downstream slope sliding after intensive rainfall
|
Aug. 2013
Locally fluidized cores found during borehole investigation, Cores with coarse-grained
materials near crest
|
Aug. 2013 Locally fluidized or weakened cores found during borehole investigation
|
Oct. 2014
Leakage found on the downstream berm and extensive wet zones found on the surface
|
Remediation
|
May - Aug. 2000.
Partial compaction grouting (L80m)
Apr. - Sept. 2003 Permeation grouting
|
Jul. 1986 - May 1989
General grouting performed four times
Aug. - Oct. 2003 Downstream slope sliding and emergency slope protection
May - Oct. 2004 Permeation grouting
|
Apr. - Oct. 2014 Permeation grouting and upper crest bentonite injection
|
Sept. - Dec. 2014
Permeation grouting
|
May - Oct. 2015
Permeation grouting and upper crest bentonite injection
|
2.2 W Dam
W dam, a 55 m tall earth-cored rockfill structure, experienced significant core degradation,
marked by the formation of three sinkholes near the crest during normal operation
in 1998 (Fig. 2). According to Park and Oh (2018), excessive leakage at the downstream toe, reaching 2,340 m³/day at Normal High Water
Level (NHWL), was initially reduced to 860 m³/day following partial compaction grouting.
However, seepage levels rose to 1,007 m³/day within 2.5 years due to diversified flow
paths, prompting comprehensive remediation through low-pressure permeation grouting
in 2000.
Subsequent investigations in 2013 revealed localized cracks acting as seepage paths,
emphasizing the need for continuous monitoring despite the remediation efforts maintaining
core impermeability.
Fig. 2 W dam sinkholes incidents
2.3 A dam
A dam, constructed with a clay core containing significant sandy soil, exhibited wet
downstream zones shortly after impounding. Despite four rounds of grouting, heavy
rainfall in 2003 triggered a major downstream slope failure (Fig. 3). Emergency stabilization was followed by full-length permeation grouting, addressing
fluidized layers detected in geotechnical investigations.
Investigations in 2013, a decade after remediation, revealed new fluidized zones and
leaks, suggesting the ongoing development of flow paths. Continuous monitoring was
recommended to ensure long-term stability, and the downstream slope was reinforced
by with riprap to enhance slope stability.
Fig. 3 A dam safety incident
2.4 D and Y dam
Unlike W and A dams, D and Y dams displayed no visible signs of degradation. However,
geotechnical investigations revealed localized core degradation, including fluidized
zones, coarse granular particles, and high water content areas (Figs. 4-5). Electrical resistivity surveys identified low-resistivity zones indicative of internal
erosion (Fig. 6).
D Dam, completed in 1990, exhibited two fluidized layers at 50% and 75% of the dam
height from the crest, corresponding to construction phases with material changes.
Similarly, Y dam, completed in 1979, showed loose upper layers and fluidized zones
at lower sections, posing long-term risks of internal erosion. Remedial measures included
low-pressure permeation grouting for both dams.
Fig. 6 Electrical resistivity survey result of D dam
2.5 H Dam
H Dam exhibited extensive wet zones on its downstream surface, with leakage observed
near the berm at one-third of its height from the crest (Fig. 7). Investigations revealed high permeability and loose core material, with the permeability
coefficient exceeding design values. Construction deficiencies, including inadequate
core height and steep slopes, contributed to its inherent vulnerabilities.
Remediation involved permeation grouting and bentonite injection in the crest area,
addressing both structural and material deficiencies. These efforts effectively restored
the dam's impermeability, though continued monitoring remains crucial.
Fig. 7 Anomalies of H dam
3. Lessons Learned
Analyzing safety incidents from aging embankment dams provides critical insights for
improving dam safety management practices. This section outlines the key lessons learned
from the five case studies, highlighting strategic investigation approaches, proactive
diagnostics, mitigation of inherent vulnerabilities, and the importance of a lifecycle
perspective in dam safety management.
3.1 Lesson 1: Need for Strategic Dam Assessment to Address Varied Vulnerabilities
Tailored Investigation Approaches: Each dam's safety concerns were addressed through
comprehensive geotechnical investigations, including borehole logging, soil and field
testing, and geophysical surveys. Matching investigation strategies to specific objectives—such
as seepage path detection or material property assessment—is crucial. Techniques like
Standard Penetration Test (SPT), soil mechanics laboratory tests, geotechnical site
characterization, and electrical resistivity surveys proved instrumental as shown
in Park and Oh (2016b). Drone photogrammetry is another useful tool for deformation monitoring (Kim et al., 2021; Park et al., 2023).
Swift Responses to Indicators: Signs of safety issues, such as excessive leakage,
sinkholes, or slope instability, demand immediate, in-depth examination, even during
operation as indicated in Park and Lim (2016). Methods such as dry borehole drilling and no-rotary wash techniques are effective
in these conditions (Park and Oh, 2016a).
Multi-Tiered Assessments: A tiered approach, starting with basic surveys and advancing
to detailed investigations, enhances efficiency. For W, A, Y, D dams in this study,
initial electrical resistivity surveys provide a broad overview, while targeted drilling
and 3D tomography offer in-depth insights into dam integrity.
3.2 Lesson 2: Necessity of Proactive Diagnostics Beyond Visual Inspection for Aging
Dams
Proactive Diagnostics Essential: Aging dams require active investigation methods to
uncover hidden risks. While W, A, and H dams exhibited visible safety concerns, issues
in D and Y dams remained undetected for decades until geotechnical investigations
revealed core anomalies (Park and Oh, 2016b).
Bridging Data Gaps: Many older dams lack comprehensive construction or design records
as shown in A and W dams. Proactive diagnostics—combining direct investigations and
geophysical techniques—are critical for evaluating their safety and extending their
operational life as shown in Table 1.
3.3 Lesson 3: Importance of Minimizing Inherent Vulnerabilities to Prevent Dam Incidents
Addressing Core Deficiencies: The studied dams exhibited vulnerabilities such as fluidized
cores, saturated slopes, and sinkholes, often resulting from hydraulic fracturing.
Alternative theories suggest that internal erosion and material deficiencies, compounded
by inadequate compaction and saturation, amplify these risks.
Construction Practices Impact Longevity: Cases such as W Dam revealed that changes
in material sourcing during construction directly influenced long-term safety. H dam's
excessive leakage stemmed from insufficient core height, emphasizing the need for
rigorous quality control during design and construction phases.
Minimizing Inherent Vulnerabilities: By identifying and addressing design and construction
errors early, the risk of long-term failures can be mitigated. A lifecycle approach
ensures that inherent vulnerabilities are minimized at every stage.
Noteworthy incidents in the dams, including unevenly distributed fluidized cores with
elevated moisture content, saturated downstream slopes, sinkholes at the dam crest,
and downstream slope sliding, are often attributed to hydraulic fracturing within
the core zone. This phenomenon is generally observed in the core zones of embankment
dams, aligning with Sherard's theory (Sherard, 1986), which posits that dam damage or failure often results from hydraulic fracturing
or dispersive clays. Hydraulic fracturing is characterized by the initiation and expansion
of cracks from the upstream side of the core, driven by water pressure surpassing
the total stress on the crack surface and the subsequent widening of these cracks.
Despite the prevalence of Sherard's explanation, the pattern of incidents in the examined
dams raises questions regarding the actual mechanics of hydraulic fracturing. The
depth, thickness, and distribution of fluidized zones, alongside the occurrence and
dimensions of cracks relative to dam height, suggest a deviation from the classic
hydraulic fracturing model proposed by Sherard.
Alternative perspectives offered by Mesri et al. (Mesri and Ali, 1988), Nonveiller (Nonveiller, 1988), and Lofquist (Lofquist, 1988) highlight the impact of water pressure and flow velocity on pre-existing weak areas,
leading to internal erosion, piping, water leakage, and slope destabilization. This
viewpoint, suggesting the amplification of inherent vulnerabilities due to factors
like saturation-induced settlement, variations in material sourcing, and inadequate
compaction, offers a more convincing explanation than the traditional hydraulic fracturing
theory like the Sherard's theory.
Case studies, such as the W dam, where fluidized zones and water inflow areas are
predominantly found at specific depths tied to changes in material sourcing during
construction, underscore the significance of construction phase decisions on long-term
dam integrity. Similarly, investigations into the H dam revealed discrepancies in
core height from design specifications, contributing to significant water leakage.
As demonstrated, design and construction phase errors or deficiencies serve as inherent
vulnerabilities that can precipitate failures or incidents throughout a dam's operational
lifespan. The identification of such inherent vulnerabilities and their impact on
the dam's aging process emphasizes the critical importance of minimizing these inherent
risks through meticulous quality control from the outset of dam design and construction.
This proactive approach is fundamental to ensuring the long-term safety and resilience
of dam infrastructure.
Fig. 8 shows the depth, thickness, and distribution of fluidized zones found in the study
dams, and Fig. 9 shows crack occurrence cases investigated by Foster and Fell (Foster and Fell, 1999) and the fluidized zones considered as cracks in the 5 study dams. This shows the
status of crack occurrence compared to the height of the dam. As shown in Figs. 8-9, considering the variation in location and size of the cracks, it is difficult to
believe that the pure hydraulic fracturing phenomenon claimed by Sherard occurred
as described in Table 2.
In fact, in the case of W dam, as shown in Fig. 10, the fluidized zone and the inflow water are concentrated at a depth of 20 to 45
m from the dam crest. According to construction records, the part where the material
source was changed during the construction process corresponds to this section. In
the case of the H dam, as shown in Fig. 11, as a result of the excavation investigation of the dam crest, it was discovered
that the core was not built to the designed height, which can be seen as contributing
to the excessive water leakage.
As confirmed in Figs. 8 to 11, incidents that occur during post-construction operation can be seen as being greatly
affected by various defects or errors in the design or construction phase. As such,
various defects or errors in the design or construction process exist as inherent
vulnerabilities throughout the entire life cycle of a dam and can become a major cause
of failures or incidents during the operation and management phases. Fig. 12 shows the general relationship between the aging process and inherent vulnerability
during the entire life cycle of a dam. In Fig. 12, path O-A-A’ has no inherent vulnerability, path O-B-B’ has some inherent vulnerability
and deteriorates faster and has a shorter lifespan than both paths O-A-A’ and O-C-C’.
It represents a life cycle pattern in which the vulnerability is so severe that an
accident such as collapse occurs during the initial freshening process.
Therefore, minimizing inherent vulnerabilities through thorough quality control from
the design and construction process is the best way to ensure the safety of the dam
throughout its entire lifespan.
Table 2 Conventional hydraulic fracturing and alternative findings in this study
Aspect
|
Sherard's Hydraulic Fracturing Theory (1986)
|
Alternative Explanations found in this study
|
Failure Mechanism
|
Water pressure exceeds total stress, initiating and expanding cracks in the core
|
Water infiltrates weak zones, leading to internal erosion, piping, and seepage failure
|
Primary Cause
|
Hydraulic stress leading to cracking
|
Material deficiencies, heterogeneity, inadequate compaction, and saturation effects
|
Crack Location
|
Initiates from upstream side of the core
|
Found in fluidized zones in center & material transition layers
|
Crack Pattern
|
Line
|
Zone or Area
|
Main Contributing Factor
|
Water pressure-driven fracturing
|
Water flow exploiting pre-existing weak zones
|
Field Evidence
|
Cracks should be consistently found relative to dam height
|
Crack occurrence varies, correlating with construction defects and material changes
|
Case Study Support
|
Classic hydraulic fracturing models suggest cracks should be uniformly distributed
in core zones
|
W, D & Y Dams: Fluidized zones matched material transition points
H Dam: Core height discrepancy led to leakage
|
Recommended Mitigation
|
Control hydraulic stress in the core
|
Enhance material selection, compaction, and construction quality control
|
Fig. 8 Locations and thickness of fluidized zones of the study dams
Fig. 9 Depths of cracks in core and height of dams
Fig. 10 Embanking schedule and change of borrow materials for the core zone (W dam)
Fig. 11 Status of the top of core by excavation investigation (H dam)
Fig. 12 A generic aging model of dams taking into account inherent vulnerability
3.4 Lesson 4: Lifelong Commitment Required for Dam Safety – Addressing Root Causes
Lifecycle Perspective: Dam safety incidents during operation often originate from
earlier phases, such as design or construction. Following guidelines from ICOLD Bulletins
154 and 175, organizations should adopt comprehensive safety protocols that span a
dam's entire lifecycle, from planning to decommissioning.
Comprehensive Monitoring Systems: Developing and maintaining robust systems for issue
identification, monitoring, and response are essential. These systems must be adaptable
to address both internal and external risks effectively.
All incidents that occur during the operation phase of a dam originate from the previous
phase. Every incident encountered during the operational phase of a dam can be traced
back to earlier stages in its lifecycle, encompassing the plan, design, construction,
commissioning, operation, and eventual rehabilitation or decommissioning phases. As
outlined in the International Commission on Large Dams (ICOLD) Bulletin 154, the operational
phase represents the most extended period in a dam's lifecycle (ICOLD, 2017). Organizations overseeing dam operations are tasked with the development and maintenance
of comprehensive safety protocols. These protocols must be designed to effectively
identify, monitor, and rectify any potential or actual issues impacting dam safety
over the long haul, while also being robust against both external and internal disruptions.
ICOLD Bulletin 175 (ICOLD, 2021) further emphasizes that incidents during the operation phase are not spontaneous
but have roots in the dam's previous phases. This underscores the importance of a
holistic approach to dam safety, recognizing that effective incident prevention and
management stem from rigorous attention to detail throughout every phase of the dam's
lifecycle.
3.5 Lesson 5: Urgent Need for a Government-Level Incident Management System (IMS)
for Dams
Integrated Management: Effective dam safety requires government-level incident management
systems to support local authorities, especially in cases with limited expertise or
resources. South Korea's initiative to classify high-risk reservoirs and provide financial
support for reinforcement is an exemplary approach.
Enhanced Technical Oversight: Beyond financial assistance, governments must establish
expert-driven frameworks for incident prevention, investigation, and response. This
ensures comprehensive safety management across diverse dam infrastructures. The necessity
of a government-level incident management system becomes evident in cases like the
H dam, where local governmental management faces challenges due to limited dam safety
expertise and budget constraints. These limitations hinder effective safety management
and can lead to frequent catastrophic failures.
In South Korea, the government's initiative to support disaster-prone and high-risk
reservoirs through various projects showcases a proactive approach to enhancing dam
and reservoir safety under local government management. By 2023, out of 13,685 local
government-managed reservoirs, approximately 230 were identified as high-risk through
inspections and risk assessments (Fig. 13). These were subsequently classified as disaster risk reservoirs, qualifying them
for government-funded reinforcement measures.
Despite these advancements, the government's role is primarily focused on designating
disaster risk reservoirs and facilitating financial support. The technical aspects
of incident prevention, investigation, and response largely depend on a selected group
of experts. This reveals a significant area for improvement, indicating the need for
a more integrated and expert-driven approach within the government-led incident management
framework to ensure the comprehensive safety and integrity of dams and reservoirs.
In conclusion, for dams that have experienced incidents, as in the case studies examined,
the establishment of an Incident Management System (IMS) is paramount. This system
should be capable of methodically overseeing the entirety of the process—from the
initial detection of an incident through to its thorough investigation and subsequent
reinforcement measures.
Fig. 13 Number of designation of dams with hazards by year in South Korea
4. Conclusion
This study examines five central earth-cored embankment dam incidents in Korea, focusing
on core material degradation and seepage-related issues. Each dam exhibited specific
challenges, including sinkholes, fluidized cores, wet downstream surfaces, and slope
failures. Despite these variations, a common factor was the compromised impermeability
of the clay core, often attributed to material deficiencies, inadequate compaction,
and complex geotechnical interactions. The common feature among all dams was the degraded
impermeability of their core layers, and the degrees of degradation were largely inhomogeneous
and anisotropic.
Some key features and lessons to learn are described. According to the electrical
resistivity survey of all five dams, significant heterogeneity appears with very low
resistivity values. The common cause might be the interaction of a complex set of
reasons pertaining to permeable core materials, less compaction, and arching. However,
the consequences of these phenomena are exhibited differently depending on the degree
of heterogeneity of the causes and how much they are distributed locally or globally.
The case study draws valuable lessons from five dam safety incident case studies,
emphasizing the importance of strategic investigations, proactive diagnostics for
aging dams, minimizing inherent vulnerabilities, recognizing the lifelong nature of
dam safety, and the need for government-level incident management systems. Investigations
must match objectives with a tiered approach (basic to in-depth). Swift, detailed
investigations are crucial when critical indicators arise. Electrical resistivity
surveys provide a useful initial overview, with targeted drilling for further exploration.
Proactive diagnostics for aging dams are valuable. Older dams, often lacking design
records, require proactive direct investigations even more urgently. Inherent vulnerabilities
may lead to long-term incidents (e.g., changes in materials, deviations from design).
Operational incidents have roots in earlier phases of a dam's lifecycle. ICOLD Bulletins
emphasize this interconnectedness and the need for proactive safety protocols throughout
the dam's lifespan. A government- led system with integrated technical expertise is
essential.
For remediation of core layer as a water barrier, a low-pressure permeation grouting
method is adopted. The empirical case studies and lessons in this paper are expected
to offer an important reference for hazard mitigation technology in a wide range of
aging dam rehabilitation projects.
Regarding the specific implementation framework, a well-structured Incident Management
System (IMS) for dam safety must integrate a proactive monitoring framework with rapid
response capabilities to mitigate risks associated with seepage failures. The implementation
plan should be anchored in real-time data collection, predictive analytics, and a
structured response protocol. First, an advanced seepage monitoring system should
be deployed to detect early signs of core material deterioration. Additionally, remote
sensing techniques can enhance surface deformation monitoring, offering crucial insights
into potential slope instability.
The response framework should be structured around a tiered approach, categorizing
incidents based on severity levels and defining clear intervention protocols. For
minor seepage anomalies, preventive maintenance measures like controlled drainage
and localized grouting should be implemented. For moderate issues involving the development
of wet zones or sinkholes, targeted low-pressure permeation grouting can be deployed
to reinforce the core material without inducing additional stress. In severe cases,
such as large-scale slope failures or rapid water ingress, emergency measures—including
controlled reservoir drawdowns and rapid embankment reinforcement—must be enacted.
Furthermore, government- level support is critical to ensuring the success of this
IMS, requiring the establishment of regulatory guidelines, funding mechanisms, and
inter-agency coordination. Periodic dam safety training, scenario-based drills, and
public awareness campaigns should also be integrated to enhance preparedness and response
efficiency. By embedding these measures into a lifecycle-based dam management approach,
the IMS will significantly enhance resilience against seepage-related failures, safeguarding
water security and renewable energy infrastructure.