Yonatan Sisay
Yonatan Sisay
Yonatan Sisay
Dam Breach Analysis & Inundation Map for Melka Wakena Dam
BY
October, 2016
Certification
This to Certify That This Thesis Entitled Dam Breach Analysis & Inundation Map for Melka
Wakena Dam, Done and Submitted
By
At
Addis Ababa University
_____________________________
October, 2016
Addis Ababa University
Dam Breach Analysis & Inundation Map for Melka Wakena Dam
A Thesis Submitted to the School of Graduate Studies of Addis Ababa University in Partial
Fulfillment of the Degree of Master of Science in Civil Engineering
By
_________________________
Advisor Signature
_________________________
Internal Examiner Signature
_________________________
External Examiner Signature
_________________________
Chairman (Department of Graduate Committee) Signature
ABSTRACT
This study presented dam breach analysis & inundation map for Melka Wakena Dam. Melka Wakena
Dam is an earth-rock fill dam with 42m height and 2000m long. The dam was designed to produce
Hydropower of 153 MW. In this study, the Melka Wakena Dam breaching outflow hydrographs and
downstream flood propagations were simulated by applying computer programs to recognize the possible
relationships among the peak flows, dam breach parameters and downstream river parameter.
Dam breach analysis & inundation map involved with reservoir routing and river routing techniques. The
key inputs required in the flows routing processes include time to dam failure in hours (TFH), side slope
of dam breach (SS), shape of opening, downstream channel geometries, Manning roughness coefficients
and inflows to reservoir. The reservoir components of routing were performed applying HEC-RAS
computer program. Similarly, the HEC-RAS computer program was applied to conduct unsteady flow
routing through the river components of the testing waterways and produced various peak flows for given
conditions at specified downstream reach stations.
The maximum breach discharge resulted from HEC-RAS model was 36,527.15m3/s which results in
overtopping the dam by 18cm and the maximum breach discharge results for piping 32,627.70m3/s. The
influences of dam breach and river parameters on maximum breaching outflow discharges were analyzed
at the dam site and downstream stations. Dam breach analysis & inundation map analyses results showed
that the maximum discharge and downstream routing. The breach discharge for overtopping maximum
than piping.
Key Words: - Dam Break, Inundation Map, Dam Breach Parameters, HEC-RAS.
i
ACKNOWEGEMENT
First of all I would like to give my deepest gratitude to my advisor Dr.Ing. Asie Kemal for his continuous
supports and advices; without him this paper would not be accomplished.
Secondly, I would like to thank Ministry of Water, Irrigation and Energy Employees starting from their
archive department up to the Engineers who did not hesitate to help me while I was in their company to
collect information for the paper. And I would also like to thank employees of the Ethiopian Electric
Power Corporation employees for providing me the geographical information data of the Wabi Shebelle
River without any delaying.
At last but not the least I would like to give my deepest gratitude to my family and friends for their endless
support and loving me unconditionally, without their support I would not stand where I am right now.
ii
TABLE OF CONTENTS
ABSTRACT……………………………………………………….…………………………………........i
ACKNOWLEDGMENT…………………………………………………………...................................ii
LIST OF TABLES....................................................................................................................................vi
LISTOFFIGURES...................................................................................................................................vii
1. INTRODUCTION………………………………………………………………...............................1
1.1. Background…………………………………………………………………………………….....1
1.2. Objective……………………………………………………………………………………….....3
1.2.1. General Objective………………………………………………………………………….3
1.2.2. Specific Objective…………………………………………………………………………3
2. LITERATURE REVIEW ON DAM BREACH ANALYSIS………………………………………4
2.1. Background…………………………………………………………………………………….....4
2.2. Types of Dames…………………………………………………………………………………...4
2.3. Dam Breach Analysis Purpose………………………………………………………....................7
2.4. Causes of Dam Failures…………………………………………………………………………...7
2.5. Dam Failure Examples…………………………………………………………………………....8
2.6. Dam Breach Analysis Study Approaches………………………………………………………..10
2.6.1. Event-Based Approach…………………………………………………………………...11
2.6.1.1. Fair Weather (Non-Hydrologic) Failure………………………………………….11
2.6.1.2. Hydrologic Failure……………………………………………………………….15
2.6.2. Risk-Based (Consequences-Based) Approach…………………………………………...15
2.6.2.1. Inflow Design Flood and the Incremental Hazard Evaluation…………………....16
2.6.2.2. Loss of Life / Population at Risk……………………………………………….....17
2.6.3. Tiered Dam Breach Analysis……………………………………………………………..19
2.7. Dam Breach Parameter………………………………………………………………………….20
2.7.1. Dam Breach Parameter Definitions……………………………………………………....21
2.8. Breach Mechanisms for Embankment Dams…………………………………………………....22
2.9. Available Approaches…………………………………………………………………………...25
2.9.1. Empirical Models for Predicting Breach Parameters…………………………………….27
2.9.2. Empirical Models for Predicting Breach Outflows…………………………………….....32
iii
2.10. Overview of Dam Breach Hydrograph Model…………………………………………………35
2.10.1. Dam breach hydrograph and peak outflow generation tools…………………………….35
2.10.2. Breach hydrograph generation and downstream hydraulic routing……………………...38
2.10.3. Recommendations for selecting modeling software……………………………………..41
3. GENERAL DESCRIPTION OF THE STUDY AREA…………………………………………..42
3.1. Location………………………………………………………………………………………....42
3.2. Climatic Characteristics…………………………………………………………………………44
3.3. Wabi Shebelle River…………………………………………………………………………….44
4. DAM BREACH ANALYSIS AND INUNDATION MAP METHODOLOGIES………………45
4.1. Data Gathering and Processing………………………………………………………………….45
4.2. Dam Breach Analysis Procedures………………………………………………………………45
4.2.1. Predicting the Outflow Hydrographs…………………………………………………….46
4.2.2. Defining the River Geometry…………………………………………………………….47
4.2.3. Describing Reservoir Characteristics…………………………………………………….47
4.2.4. Physical Descriptions of Dam……………………………………………………………48
4.2.5. Determining Inflow Hydrograph to the Reservoir……………………………………….48
4.2.6. Estimating Dam Breach Characteristics…………………………………………………49
4.2.7. Routing Breach Outflow Hydrographs through Downstream Reaches………………….51
4.2.8. Defining Channel Geometry and Boundary Conditions…………………………………51
4.2.9. Selecting Manning Coefficients, “n” values……………………………………………..52
5. RESULTS AND DISCUSSIONS…………………………………………………………………..53
5.1. Breach Parameter Estimates…………………………………………………………………….54
5.2. Outflow hydrograph due to dam breach………………………………………………………...57
5.3. Routing of the hydrograph through the downstream……………………………………………60
5.4. Inundation map for the downstream…………………………………………………………….65
5.5. Emergency Action Plan…………………………………………………………………………66
6. CONCLUSIONS & RECOMMENDATION……………………………………………………..69
6.1. Conclusions……………………………………………………………………………………...69
6.2. Recommendation………………………………………………………………………………..70
REFERENCES………………………………………………………………………………………….71
v
LIST OF TABLES
Table 2.2: Range of Initial Reservoir Pool Levels for a Fair Weather (Non-Hydrologic) Analysis
(FEMA, 2013)…………………………………………………………………………………………....13
Table 2.3: Recommended IDF Requirements for Dams Using Prescriptive App. (FEMA, 2012)……...17
Table 2.4: Tiered Approach Dam Breach Inundation Mapping for use in EAPs (FEMA, 2013)……….20
vi
LIST OF FIGURES
Figure 2.3: Erosion on the downstream face of a cohesive soil embankment dam……………………...23
Figure 5.3: the discharge flowing out of the dam during the dam break for overtopping……………….57
Figure 5.4: the discharge flowing out of the dam during the dam break for piping………………………57
Figure 5.5: the discharge flowing out of the dam during the dam break for overtopping……………….58
Figure 5.6: the discharge flowing out of the dam during the dam break for piping……………………..58
Figure 5.7: the discharge flowing out of the dam during the dam break for overtopping……………….59
vii
Figure 5.8: the discharge flowing out of the dam during the dam break for piping……………………..59
Figure 5.9: the discharge flowing out of the dam during the dam break for overtopping……………….61
Figure 5.13: the discharge flowing out of the dam during the dam break for piping……………………63
Figure 5.17: inundation map for overtopping failure and depth of water level in meter ………………..65
Figure 5.18: inundation map for piping failure and depth of water level in meter ……………………...66
viii
Dam Breach Analysis & Inundation Map for Melka Wakana Dam
CHAPTER ONE
INTRODUCTION
1.1. Background
Dams are an important part of this nation’s infrastructure, providing flood control, water supply,
irrigation, hydropower, navigation, and recreation benefits. Despite their many beneficial uses and
value, dams also present risks to property and life due to their potential to fail and cause catastrophic
flooding. Catastrophic flooding occurs when dam fails and the impounded water escapes through the
breach into the downstream valley. When dams fail, property damage is certain, but loss of life can vary
dramatically with the extent of the inundation area, the size of the population at risk, and the amount of
warning time available.
There may be many reasons for dam failures, among them the floods occurring in river basins near
existing dams, triggered by intensive rain is the most responsible one. When the flood hydrograph
entering a dam reservoir reaches a peak value of unusual magnitude, the amount of water exceeding the
capacity of the dam reservoir should be diverted downstream of the dam. If a spillway built for just that
function was not designed for that kind of magnitude, excess water may spill over the dam crest. In case
this happens, a breach may form in the dam body in minutes or hours depending on the type of material
used in the dam body. As this breach gets larger and larger in time, the enormous amount of water stored
in the reservoir upstream of the dam may start its motion as an uncontrolled flood wave downstream of
the dam. A flood caused by a dam failure may occur in a much bigger magnitude compared to those
floods generated by rain or snow melt. In the downstream river bed, the fast moving flood wave with its
great power having the potential to destroy whatever comes in its way, may provoke deadly
consequences should there be residential areas on its course.
Flood induced by dam breach happens occasionally throughout the world. Floods can induce serious
loss of life and significant economic losses. To recognize the possible effects of dam breaks, a detailed
knowledge of dam breakage processes and flood propagation is required. Warning time is the most
important parameter affecting potential loss of life due to dam failure. Numerical and hydraulic models
can be used to predict flood wave propagation and provide the information about the wave front arrival
time, area to be flooded and water depth. Therefore, models are useful tool for developing evacuation
plans and warning system for areas having potential flood risk. When population centers are located
close to dams, accurate prediction of breach parameters is crucial for development of effective
emergency action plans. The development of a dam breach is a complex process involving numerous
uncertainties.
The outflow flood hydrograph from a dam failure is dependent upon many factors. The primary factors
are the physical characteristics of the dam, the volume of the reservoir and the mode of failure. The
parameters which control the magnitude of the peak discharge and the shape of the outflow hydrograph
include: the breach dimensions; the manner and length of time for the breach to develop; the depth and
volume of water stored in the reservoir; and the inflow to the reservoir at the time of failure. The shape
and size of the breach and the elapsed time of development of the breach are in turn dependent upon the
geometry of the dam, construction materials, and the causal agent for failure.
Ethiopia can be considered as the water tower of Africa because of its high water resources availability.
The country has 12 river basins. The total mean annual flow from all the 12 river basins is estimated to
be 122 BMC (MoWR, 1999). Construction of dams has been commenced since the first dam was built
in 1939 and the dam was constructed on Akaki River to generate hydro-electric power. The dams built
so far are being used in order to alleviate the water related problems of the population. However, it
should be pointed out at the outset that the development of dams is being threatened by (1) sedimentation
problems arising from the degradation of catchment areas fueled by four pressure indicators namely
agricultural production, rapid population growth, poverty and wood energy demands, (2) in appropriate
runoff estimation methods resulting in over sizing or under-sizing of dams, (3) and unreliable spillway
flood estimation methods.
Large dam projects are prone to delays. The dams built in Ethiopia are no exception to the rule and all
have been delayed by at least one year. A complex geology has been one of the reasons for the delays,
leading to landslides and tunnel collapses. The Gibe II dam has been affected by such problems even
after its completion, when a tunnel collapsed and put the hydropower plant out of service for several
months. The construction of large dams entails many tangible and intangible costs. The financial cost
itself is already substantial. Resettlement adds to the social costs of the dams. Sedimentation from
unchecked erosion in the upper watershed of rivers reduces the lifespan of reservoirs. Environmental
costs are imposed on communities living downstream of the dams in Ethiopia (Wikipedia, 2015).
The purpose of the dam breach analyses has been to illustrate how the flood wave propagates and
attenuates along the river valley from Melka Wakena dam. In the present analyses the HEC-RAS model
is used for simulation of the flood wave caused by dam failure. This model is one of the most widely
accepted model of its kind.
1.2. Objective
1.2.1. General Objective
The term dam break analysis usually relates to the process of studying a dam failure phenomenon and
analyzing the resulting consequences at the downstream region. This generally deals with simulation of
assumed failure for existing dams and analyzing the resulting consequences. The prime objective is
prediction of the reservoir outflow hydrograph and the routing of that hydrograph through the
downstream valley to determine dam failure consequences and to facilitate effective emergency action
planning.
CHAPTER TWO
2.1.Background
Floods resulting from dam failures led to catastrophic and tragic consequences in the past. Researchers
have been working to develop computer programs that would help to design new dams or evaluate
existing dams.
The actual failure mechanics of dam failure have not been well understood for either earthen or concrete
dams. In earlier attempts to predict downstream flooding due to dam failures, it was usually assumed
that the dam failed completely and instantaneously (Abinet, 2010). Some investigators of dam-break
flood waves assumed the breach encompasses the entire dam and that it occurs instantaneously. Others,
such as Army Corps of Engineers (1960), have recognized the need to assume a partial failure rather
than complete breaches.
Several researchers have devolved regression equations to estimate breach size, shape, and time to dam
failure from historical dam breach information. A few researchers have tried to develop computer
models to simulate the physical breaching process.
2.2.Types of Dames
Dams may be classified by purpose, type, size, and hazard potential, the latter of which varies greatly
between States and Federal agencies. This section describes the most common types of dams.
There are numerous intended purposes for man-made dam structures, such as flood retarding,
diversionary, irrigation and water supply, hydroelectric power generation, and recreational. Recreation,
flood control, and fire protection are the three most common applications. Fire protection, as defined by
the NID data dictionary, includes stock ponds and small farm ponds.
The NID classifies dams by the type of construction material used with the majority listed as either a
concrete or embankment type dam. Concrete dams include arch, buttress, concrete, gravity, masonry,
multi-arch, and roller-compacted concrete (RCC) and are typically constructed of concrete or other
masonry components. Embankment dams are made of earthen materials and may be filled with rock,
earth, or other materials resistant to erosion.
Concrete Dams
There are several types of concrete dams ranging from conventional design styles such as gravity, arch,
multi-arch, and buttress dams to newer design approaches such as RCC dams.
Embankment Dams
Embankment dams are made from compacted earth. There are several types, as shown in Figure 2-1.
The two most common types of embankment dams are rock-fill and earth-fill dams.
Earth-fill dams are composed of suitable soils obtained from borrow areas or required excavation that
are spread and compacted in layers by mechanical means. Earth-fill dams may be constructed with
homogenous layers (homogeneous dam) or zones of different materials of varying characteristics
(zoned-earth dam). Earth-fill dams are typically trapezoidal in shape and rely on their weight to hold
back the force of water, similarly to concrete gravity dams. Typical zones include a clay core and filter
and drain zones.
A unique category of earth-fill embankment dams are tailings dams used by the mining industry.
Tailings dams are often constructed of coarse tailings produced by the mine but may also consist of
other soils obtained near the construction site. Tailings dams often rely on the stored tailings to control
seepage, but otherwise include many of the same design features as conventional water storage dams.
Rock-fill dams are constructed from compacted earth fill that contains a high percentage of rocks and
other larger particles. The fill typically drains easily and therefore no drainage layer is required. To
prevent seepage, rock-fill dams have an impervious zone on the upstream side of the dam or within the
embankment. The impervious zone can be made from a variety of materials including masonry, concrete,
plastic, steel pile sheets, timber, or clay. If clay is used, it is often separated from the fill by a filter to
prevent erosion of the clay into the fill material.
Earth-fill dams may include a water-tight core can also be made from asphalt concrete. Dams with this
type of core are called concrete-asphalt core embankment dams. Most concrete-asphalt dams use rock
and/or gravel as the main fill material. These types of dams are considered especially appropriate for
areas susceptible to earthquakes due to the flexible nature of the asphalt core.
Structural 12 1.8%
Spillway 11 1.7%
Erosion/Slide/Instability 13 2.0%
Unknown 32 4.9%
Other 18 2.7%
The southern embankment of the Lake Delhi Dam in Delhi, IA, failed on July 24, 2010, due to heavy
raining and flooding. The dam failed after receiving about 10 inches of rainfall in 12 hours. Before the
breach, river levels upstream of the dam had reached 24.22 feet, 10 feet above flood stage, breaking the
May 2004 record of 21.66 feet.
Piping and seepage failure: - In 1976, the failure of the Teton Dam in Idaho led to flooding in the cities
of Sugar City and Rexburg (Figure 2-2). The dam failure killed 14 people and caused over $1 billion in
property damage. Over 2,000,000 cubic feet per second of sediment-filled water emptied through the
breach into the remaining 6 miles of the Teton River canyon, after which the flood spread out and
swallowed on the Snake River Plain. Study of the dam’s environment and structure placed blame for the
collapse on cracks in the permeable soil (loess) used in the core and on cracks in the foundation bedrock
that allowed water to seep under the dam. The combination of these flaws is believed to have allowed
water to seep through the dam, which led to internal erosion, called piping, which eventually caused the
dam’s collapse.
Structural failure: - The Kingston Plant coal waste dam failed in Harriman, TN, on December 22,
2008. This was a 40-acre pond used by the Tennessee Valley Authority to hold slurry generated by the
coal-burning Kingston Steam Plant. The dam gave way just before 1 a.m., burying a road and railroad
tracks leading to the plant. Although no one was seriously injured or hospitalized, 5.4 million cubic
yards (> 1 billion gallons) of sludge damaged 12 homes and covered hundreds of acres.
Spillway gate failure: - A spillway gate of Folsom Dam in California failed in 1995, increasing flows
into the American River significantly. The spillway was repaired and the USBR carried out an
investigation of the water flow patterns around the spillway using numerical modeling,
Earthquake failure: - The Lower San Fernando Dam in California failed during an earthquake in 1971,
causing the fill in the dam wall to liquefy which resulted in the collapse of the upstream part of the dam.
A disastrous flood was only prevented because the reservoir level happened to be low at the time of the
earthquake and no water escaped downstream.
Poor design/construction failure:- In August 2008, the Redlands Ranch Dam located in Havasu, AZ,
failed due to neglect and poor design and construction. No loss of life was reported, but 426 people were
evacuated by helicopter and there was significant damage to the landscape.
Dam breach inundation studies are used for multiple purposes, including:
Evaluating and establishing the hazard potential classification for a dam
Evaluating dam safety risk and prioritizing dams within a dam safety portfolio
Selecting the appropriate SDF or IDF for dam and spillway design
Developing EAPs and exercise planning associated with dam safety permitting
Developing breach inundation zone mapping for flood warning systems and flood evacuation
planning
Developing breach inundation zone mapping for dam breach consequence studies and for flood
mitigation planning
Developing dam breach inundation zone mapping for risk communication to inform the public
of the risk living downstream of dams.
The greatest advantage to using an event-based approach is that it is a direct approach, less complicated
to perform and regulate, and produces more conservative breach inundation zone mapping when
compared to a risk-based approach. High-hazard potential dams are typically evaluated using a full PMF,
and significant- or low-hazard potential dams are evaluated on a percentage of a PMF or some more
frequent storm event.
from a storm event. A fair weather breach is typically used to model piping failures for hydrologic,
geologic, structural, seismic, and human-influenced failure modes.
Base flow conditions for a fair weather failure are typically ignored because of the small discharge and
volume compared to that of a dam breach. As a general guidance, base flow can be ignored if the dam
breach flow is two times greater than the base flow. Where base flow is considered, the discharge is
typically estimated based on reported base flows through the dam’s outlet works or from stream gage
records. The three most common initial water level elevations for fair weather breach analyses are as
follows:
Normal Pool Elevation (invert of the highest elevation of the primary outlet)
A breach at the normal pool elevation of the reservoir is used to estimate the volume and associated
breach discharge that would result from a failure event during fair weather conditions. For an
embankment dam, this type of event is modeled as piping/internal erosion failure, whereas for a concrete
dam, this event is modeled as a monolith collapse resulting from sliding, foundation instabilities, or a
seismic event.
Table 2-2. provides the recommended water surface elevation of a reservoir for used in dam breach
modeling based on published documents from Federal agencies and dam safety resource groups. The
normal pool elevation is recommended as the default volume for the fair weather failure. States should
consider a larger storage volume for dams where the primary and emergency spillway systems are
considered susceptible to blockage resulting in a higher water surface elevation and volume during a
non-hydrologic event.
Table 2.2: Range of Initial Reservoir Pool Levels for a Fair Weather (Non-Hydrologic) Analysis
(FEMA, 2013)
Initial
Referenced Initial Failure Supporting
Reservoir Supporting
Name in Inflow to Mode Federal
WSEL Documentation
Publication Reservoir Organization
Federal Guidelines
Federal for Dam Safety:
Emergency Selecting and
Normal full Normal None
Management Accommodating
reservoir(1) stream flow specified
Agency Inflow Design Floods
(FEMA) for Dams. pp. 17.
2004b.
Federal Engineering
Energy Guidelines for the
Normal full Normal None
Regulatory Evaluation of
reservoir stream flow specified
Commission Hydropower Plants.
(FERC) Ch 2. pp. 2-7. 1993.
Mine Safety Engineering and
Normal pool Normal and Health Design Manual: Coal
Piping
elevation stream flow Administratio Refuse Disposal
n (MSHA) Facilities. 2009.
National Dam Simplified
Normal pool Normal Safety Inundation Maps for
Piping
elevation(2) stream flow Review Board Emergency Action
(NDSRB) Plans. 2009.
Normal pool Reduce Dam Safety
Risk Modernization
Blueprint /
U.S.
Implementation
Top of active Normal Department of
Piping Phase 1: Launch Risk
conservation stream flow the Interior
Reduction /
(USDOI)
Inundation Mapping /
Modeling Subproject
Report. 2011.
Piping/inte Natural National Engineering
Normal
Seismic rnal Resources Manual: 210-V.
stream flow
erosion Conservation 1982.
Invert of Service
(NRCS)
auxiliary
Reduce Dam Safety
spillway Risk Modernization
Blueprint /
Top of joint Implementation
Normal
use(3) (auxiliary Piping USDOI Phase 1: Launch Risk
stream flow
spillway) Reduction /
Inundation Mapping /
Modeling Subproject
Report. 2011.
Reduce Dam Safety
Between Below top Risk Modernization
of dam Blueprint /
normal pool and Hydrologically (piping); Implementation
Hydrologic
top of dam induced static above top USDOI Phase 1: Launch Risk
event
failure of dam Reduction /
(overflow Inundation Mapping /
breach) Modeling Subproject
Report. 2011.
(1) Normal reservoir level: “For a reservoir with a fixed overflow sill the lowest crest level of that sill.
For a reservoir whose outflow is controlled wholly or partly by moveable gates, siphons or other means,
it is the maximum level to which water may rise under normal operating conditions, exclusive of any
provision for flood surcharge”.
(2) For small and intermediate-sized dams, it may be appropriate to use a single fair weather failure with
the initial elevation set to the top of the dam instead of the rainy and fair weather situations. This
“eliminates the need for expensive watershed and spillway studies and provides a reasonable upper limit
estimate for warning and evacuation”
(3) Joint use is a designation for dams with gated spillways. In these cases, the top of joint use is not the
invert of the spillway but rather some elevation that places water up on the gates.
A disadvantage of the risk-based approach is that by reducing the SDF or IDF to less than the full PMF
based on downstream consequences, new development in the downstream breach inundation zone could
alter the consequences, resulting in the need for future dam rehabilitation measures to increase spillway
capacity. Effective risk communication as a component of the local development approval process can
assist in reducing the occurrence of “hazard creep,” an occurrence where new downstream development
in a dam breach inundation zone increases the dam’s hazard potential classification or SDF/IDF design
requirement.
Once the appropriate IDF for the dam has been selected, the IDF is then routed through the dam to
determine whether the flood can be safely passed without failure. Should the IDF pass safely, then no
further evaluation or action is required; however, if the IDF cannot pass safely, then measures must be
taken to enable the project to safely accommodate all floods up to the IDF to alleviate the incremental
increase in unacceptable additional consequences a failure may have on areas downstream.
New guidance in Selecting and Accommodating Inflow Design Floods for Dams includes the
recommendations shown in Table 2-3.
Table 2.3: Recommended IDF Requirements for Dams Using Prescriptive App. (FEMA, 2012)
Hazard Potential
Classification Definition of Hazard Potential
Inflow Design Flood
Classification
described in (USBR 1986, 1999) as it is the most currently and widely used procedure for estimating
loss of life resulting from dam failure.
Probable loss of life is an important factor used in hazard potential classification systems and emergency
action planning. USBR (1999) presents a risk-based method to estimate the number of fatalities that
would result from dam failure. This method was developed using data from about 40 floods, many of
which were caused by dam failure. These publications outline the following seven steps to complete an
analysis for loss of life:
For each failure scenario and time category, the population at risk must be calculated. Population at risk
is defined as the number of people occupying the dam failure floodplain prior to the issuance of any
warning. The method developed for estimating loss of life provides recommended fatality rates based
on the flood severity, amount of warning time, and a measure of whether people understand the severity
of the flooding. Recommended fatality rates for estimating loss of life may be determined based on a set
of criteria that includes 15 different combinations of flood severity, warning times, and flood severity
understandings.
A tiered study approach was developed by the USDOI and is presented in their report titled Reduce
Dam Safety Risk Modernization Blueprint / Implementation Phase 1: Launch Risk Reduction /
Inundation Mapping / Modeling Subproject Report (USDOI, 2011). The tiered dam breach analysis
approach presented in this document adapts the USDOI approach and provides additional detail.
The NDRSB EAP Workgroup (2009) noted that the cost of detailed dam breach studies is consistently
cited as the primary impediment to EAP development and, therefore, many States have adopted a form
of simplified and conservative inundation maps for use in EAPs. The NDRSB EAP Workgroup also
stated that although detailed studies often provide a more precise representation of potential flooding
for a given set of assumptions, a more accurate representation of dam failure flooding is not necessarily
provided (FEMA, 2013).
In their effort to increase the number of EAPs for dams, a tiered approach in dam inundation modeling
has gained popularity with many State and Federal dam safety programs. Instead, the tiered approach is
used to determine the appropriate level of complexity in the assessment, modeling, and mapping of a
dam failure based on a dam’s hazard potential, size, and the complexity of the downstream area under
investigation.
The level of analysis for the tiered approach should correlate the sophistication and accuracy of the
analyses with the scale and complexity of the dam and downstream area under investigation. Therefore,
analysis of high-hazard potential dams located upstream of populated areas or complex floodplains
should use more sophisticated modeling and additional sensitivity studies to properly assess the
consequences of a dam failure; whereas, analysis of low-hazard potential dams situated upstream of
sparsely populated areas may rely on more approximate methods of analyses.
In general, as the sophistication of the modeling increases, so does the level of effort, time, and cost
necessary to conduct the analysis. Table 2-2. Provides guidance to determine the tier level for analysis
for dam failure inundation modeling and mapping. The dam failure analysis should be continued
downstream to a point where the breach flood no longer poses a risk to life and property damage, such
as the confluence with a large river or reservoir with the capacity to store the flood waters.
Table 2.4: Tiered Approach Dam Breach Inundation Mapping for use in EAPs (FEMA, 2013)
Breach Peak Breach Downstream
Tier Level Applicable to Parameter Discharge Routing of Breach
Prediction Prediction Hydrograph
Simplified
• Low-hazard Models
GeoDam-BREACH,
potential / small (SMPDBK,
SMPDBK,
Tier 1 – Basic level size GeoDam-
Empirical DSAT,1D HEC-
Screening and • First level BREACH, or
Equations RAS Steady State,
Simple Analysis screening for Technical
or HEC-HMS
significant- or Release [TR]-
Hydrologic Routing
high-hazard dams 66) or HEC-
HMS
• Significant-
hazard potential / HEC-RAS (Steady
HEC-HMS or
Tier 2 – intermediate size Empirical or Unsteady
HEC-RAS
Intermediate • High-hazard Equations Modeling) 1-D or 2-
Unsteady Model
dams with limited D models
population at risk
• High-hazard
potential / large Empirical
size dams with Equations, HEC-RAS Unsteady
HEC-RAS
Tier 3 – Advanced sufficient NWS Model or 2-D
Unsteady Model
population at risk BREACH, or models
to justify WinDAM
advanced analyses
Tier 1 and 2 analyses are most appropriate for low-hazard potential / small sized and significant-hazard
potential / intermediate-sized dams with a limited number of structures. More detailed surveying or
modeling may be warranted for Tier 3 analyses for high-hazard potential / large-sized dams, those with
a large population in the evacuation area, or those with significant downstream hydraulic complexities
It has been noted by several sources that the selection of breach parameters for modeling dam breaches
contain the greatest uncertainty of all aspects of dam failure analysis and therefore a careful evaluation
and understanding of the associated breach parameters is necessary (Wurbs, 1987; USBR, 1998; Wahl,
2004; Gee, 2008, etc.).
A number of methods are available for estimating breach parameters for use in dam breach studies.
Since the selection of the breach parameters is specific to each dam, guidance is provided describing
methods currently applied by dam safety professionals without recommending a standardized method.
Breach depth - Also referred to as breach height in many publications. This is the vertical extent
of the breach, measured from the dam crest down to the invert of the breach. Some publications
cite the reservoir head on the breach, measured from the reservoir water surface to the breach
invert.
Breach width - The ultimate breach width and the rate of breach width expansion can
dramatically affect the peak flow rate and resulting inundation levels downstream from the dam.
Case studies typically report either the average breach width or the breach width at the top and
bottom of the breach opening.
Breach side slope factor - The breach side slope factor along with the breach width and depth
fully specifies the shape of the breach opening. Accurately predicting the breach side slope
angles is generally of secondary importance to predicting the breach width and depth.
The breach width is described as the average breach width (Bave) in several of the empirical equations.
The breach height (hb) is the vertical extent from the top of the dam to the invert elevation of the breach.
Many publications and equations also use the height of the water (hw), which is the vertical extent from
the maximum water surface to the invert elevation of the breach
When breach formation times are reported in case studies, there is often some question as to whether
the reported times are only for the breach formation phase, or if they might also include some portion
of the breach initiation phase. Distinguishing between the two during (or after) a failure is a difficult
task, even for a trained observer. In the interest of promoting more accurate reporting of breach initiation
and breach formation times, the following definitions are offered:
Breach initiation time - The breach initiation time begins with the first flow over or through a dam that
will initiate warning, evacuation, or heightened awareness of the potential for dam failure. The breach
initiation time ends at the start of the breach formation phase (see next item).
Breach formation time - The duration of time between the first breaching of the upstream face of the
dam until the breach is fully formed. For overtopping failures the beginning of breach formation is after
the downstream face of the dam has eroded away and the resulting crevasse has progressed back across
the width of the dam crest to reach the upstream face.
A dam breach usually occurs in two distinct phases starting with the breach initiation followed by the
breach formation.
Overtopping Failures
Overtopping failures can occur very differently depending on the composition of the dam. Perhaps the
simplest overtopping failure to discuss is failure of a cohesive soil embankment. According to a study
by Ralston (1987), a small head cut typically forms on the downstream face of a cohesive soil
embankment and progresses upstream as shown in Figure 2.3.
Figure 2.3: Erosion on the downstream face of a cohesive soil embankment dam
The breach is considered to begin when erosion occurs across the width of the dam crest. After the
breach initiates at the top of the dam crest, it enlarges to its ultimate extent. If there is no physical reason
to believe the embankment would fail at a certain location, the breach should be modeled as initiating
at the maximum section typically located at the centerline of the downstream main channel. A
generalized trapezoidal breach progression is illustrated in Figure 2.4.
The breach may stop growing when the reservoir has emptied and there is no more water to erode the
dam or the dam has completely eroded to the bottom of the reservoir or has reached bedrock (Gee, 2009).
The breach progression may be modeled as either a linear progression or a sine wave progression:
Linear progression: rate of erosion remains the same for the duration of erosion development)
Sine wave progression: breach grows very slowly at the beginning and end of development and
rapidly in between
In a study by the State of Colorado Department of Natural Resources, no significant difference were
found between linear and sine wave progression models when comparing one overtopping case study in
HEC-Hydrologic Modeling System (HMS) and HEC-RAS (2010). Both progressions should be
evaluated and the progression with the more conservative results should be utilized.
There are several possible options to identify the breach initiation time. For breaches associated with a
hydrologic event, the initiation can be considered to begin when the reservoir water level reaches a
certain elevation or after the water level has exceeded a certain elevation for a specified Duration. For
fair weather breach analysis, an initiation time should be specified regardless of pool elevation (Gee,
2010).
Other physical models like DAMBRK simulate the breach of the dam and the resulting reservoir
outflow. The geometry and time of formation of the breach should be given to this program as an input,
and the output will give the breach enlargement as function of time (e.g., linear increase of breach
dimensions). The required input parameters should be found from either comparative methods or from
prediction equations or other physical models.
Breach characteristics
When a small variation in one of the breach parameters (width, depth, failure time and overtopping head)
occurs, large changes in peak flows will take place especially for reservoirs with relatively small storage.
In 1984, Singh and Snorrason used some models such as DAMBRK and HEC-1 on 8 hypothetical
breached dams to assess which breach parameter affects mostly the peak outflow.
Failure Time
They found that if failure time were reduced by half its initial value, the peak outflow for a PMF
hydrograph would increase by 13 to 83 %. But for large reservoirs, the change in peak outflow was
much smaller showing a variation of only 1 to 5 %.
Breach Width
It seems that the changes in breach width is more effective for large dams because it produced larger
changes (35-87%) in peak outflow and smaller changes (6-50%) for small reservoirs.
Breach Depth
If breach depth is changed, little change in peak outflow has been identified, leading to the conclusion
that the change in peak flow is not really dependent on the reservoir size.
Other studies conducted by Petrascheck and Sydler (1984) also proved that change in the breach width
and breach formation time would significantly affect the outflow peak discharge, inundation levels, and
flood arrival time. For locations not far from the dam, both breach width and breach formation time will
have a great influence.
Some critical results have been found by Wurbs(1987). In large reservoirs, the peak outflow takes place
at the moment when the maximum depth and width of the breach are attained. Changes in reservoir head
are relatively slight during the breach formation period. In small reservoirs, a huge change in the level
of the reservoir takes place during the formation of the breach; consequently the peak outflow occurs
sometime before reaching the final breach. Here, the formation rate of breach is crucial.
Where:
tf =failure time (hr)
FERC (1987)
FERC proposed usually
2hd < B < 4hd
But B can range
hd < B < 5hd
Where:
B=is the breach width (m)
hd=dam height (m)
tf = 0.79(S)0.47
Where:
tf =dimensionless breach formation time=tf /(ghb)0.5
These equations were based on very specific dam characteristics like the presence of core, height of
water above breach bottom, the extent of overtopping and so on. He also realized that overtopping causes
the most breach extension and erode at a higher rate than any other failure mode.
In 1995, 8 years after his first study, he published new and revised equations based now on 63 case
studies. This time, the new equations are not non dimensional. These equations have better estimated
coefficients. These new equations are:
Bavg (m) =0.1803K0Vw0.32 hb0.19
Where:
K0=constant=1.4 if there is overtopping and 1 if else.
tf =0.000254Vw0.53 hb(-0.9)
Where:
tf =failure time (hr)
Z=1.4 if there is overtopping, if not Z=0.9
Reclamation (1988)
They develop these equations for earthen dams where:
B = 3hw
tf (hours) = 0.011B and B is in meters
Where:
hw =height measured from the initial reservoir water level to the breach bottom elevation
which is assumed to be the streambed elevation at the toe of the dam.
tf =failure time (hr)
B=is the breach width (m)
Reclamation uses these formulas in the SMPDBK model. The suggested formulas are conservative, and
thus they represent a factor of safety for the hazard classification procedure.
6.17*106-1.23*107 42.7
>1.23*107 54.9
They plotted the volume of the eroded embankment versus water outflow volume and water depth above
the breach invert, with upper bounds of reasonable breach geometry estimates. These methods are
dependent on the amount of erosion that occurs:
tf (hr) = 0.020hw + 0.25 (erosion resistant)
tf (hr) = 0.015hw (easily erodible)
Where:
tf should be in hours
hw= the depth of water at the dam at the time of failure (m)
AAIT, School of Civil and Environmental Engineering Page 31
Dam Breach Analysis & Inundation Map for Melka Wakana Dam
Moreover, they have suggested other equations that estimate the time of failure using the average lateral
erosion rate (the ratio of the final breach width to breach formation time) and depth of water above the
breach invert. They conclude that there is abettor estimation using these equations than the first ones
that they developed. These new equations are:
Kirkpatrick (1977)
Using data from 13 failed embankment dams and 6 other hypothetical failures; he related the peak flow
versus the depth of water behind the dam at failure. This equation is written as:
QP=f (hw)
Where:
QP = peak flow (m3/s)
hw= the depth of water at the dam at the time of failure (m)
But the flaw of this method is that among the case study failures he used is the St.Francis Dam in
California, which was a concrete gravity dam.
SCS (1981)
The Soil Conservation Service used the 13 cases studied by Kirkpatrick in order to develop another
method, for earth dam, that relates the peak dam failure outflow to the depth of water at the dam at the
time of failure. The equation is given by:
When Hw >31.4 m
Where:
Hw = the height of water directly at the reservoir before breach measured from the bottom
of the final breach.
Qp = peak outflow through the breach (m3/s).
When Hw <31.4m
Qp = 0.000421 (VwHw/WH) 1.35…..………………………… (2)
Where:
Vw=reservoir water volume at the time of failure (m3)
W=average width from the bottom of the final breach to the top of the embankment (m)
H=distance from the bottom of the final breach to the top of the embankment (m)
But the flow calculated in (2) should not exceed the value given by (1) and not less than
Qp =1.77Hw2.5
Where:
Hw = the height of water directly at the reservoir before breach measured from the bottom
of the final breach (m)
Qp = peak outflow through the breach (m3/s).
From the plot of the results of this method with that of the observed flows, it appears that there is a good
matching between calculated and measured peak flows except at the low peak flows.
The problem of this method is that it does not provide a way for determining a peak outflow that provides
a factor of safety when evaluating downstream flooding.
Reclamation (1982)
Used the work done by SCS and proposed a similar envelope equation for peak breach outflow using
case study data from 21 failed dams.
Where:
Qp = peak outflow through the breach (m3/s).
Vw= the total quantity of stored water at failure (m3)
Hw = the hydraulic height of water directly at the reservoir before breach. This formula
will exaggerate the peak flow for embankment dams (m)
They have also tried to establish similar relations on non-earthen dams, but this attempt did not succeed
because the standard deviation of the data was large.
Costa (1985)
This method is mainly based on regression analysis. It applies for both embankment and concrete dams,
because the 31 cases studied to develop this method were a mix of both embankment and concrete dams.
The peak outflow is given by:
Qp =0.763(VwHw) 0.42
Where:
Qp = peak outflow through the breach (m3/s).
Vw= the total quantity of stored water at failure (m3)
Hw = the hydraulic height of water directly at the reservoir before breach. This formula
will exaggerate the peak flow for embankment dams (m)
But this formula overestimates the peak outflow for the embankment dams because a concrete dam will
have bigger breach than a similar embankment dam having the same volume.
Froehlich (1995)
The equation is found by running a multiple linear regression on 22 dams where discharge data were
available. This equation is given by:
Qp =0.607Vw0.295Hw1.24
Where:
Qp = peak outflow through the breach (m3/s).
Vw= the total quantity of stored water at failure (m3)
Hw= the hydraulic height of water directly at the reservoir before breach. This formula will
exaggerate the peak flow for embankment dams (m)
This equation gives a good agreement with the measured computed peak flows over the entire range.
Tools that generate the dam breach peak discharge and/or hydrograph only; and
Tools that develop a breach hydrograph and perform downstream flood routing
Simplified numerical models typically relate the breach hydrograph (or breach peak flow) to simple
reservoir characteristics such as reservoir volume and dam height. These models may or may not include
hydrologic modeling to determine the envelope maximum water depths to calculate the breach flow.
Most simplified models do not consider complicated downstream conditions such as backwater effects.
Additionally, reservoir routing (if present) uses level pool routing methods; in other words, the reservoir
water surface is considered level during drawdown. This simplification is not applicable to all situations.
The main benefit of simplified numerical models is that substantially less time is required to set up and
execute these models.
WinDAM B
The ARS recently developed WinDam B in cooperation with the NRCS and Kansas State University,
which expands on the capabilities of WinDam A.
significance in dam breach studies and its ongoing use for some dam breach studies, a description of the
model is included in this document. The model was initially developed in 1987 with updates in 1988,
1991, and 2005. The BREACH program is no longer supported by the NWS and is not available for
download on the NWS Web site. It is still used because it is known to more accurately predict breach
progression than other available methods and perhaps because it has not yet been replaced by another
freely available, non-proprietary program that performs the same function.
BREACH couples the conservation of mass of the reservoir inflow, spillway outflow, and breach
outflow with the sediment transport capacity of the unsteady uniform flow along an erosion-formed
breach. The growth of the breach, as shown in Figure 4-6, is dependent on the dam’s material properties
and the assumed location of the downstream face of the dam. Sediment transport equations are used in
the model to compute the rate of erosion and size of a breach based on supplied soil characteristics of
the dam material and the inflow hydrograph. Enlargement of the breach is further evaluated by a sudden
collapse due to excess hydrostatic pressure and breach width expansion by slope stability (Gee, 2010).
The outflow hydrograph is obtained through a time-stepping solution.
As documented in the BREACH Manual developed by Fread in 1991, the BREACH model considers
the possible existence of the following complexities:
Core material having properties that differ from those of the outer portions of the dam
The necessity of forming an eroded ditch along the downstream face of the dam prior to the
actual breach formation by the overtopping water
The downstream face of the dam having a grass cover or being composed of a material of larger
grain size than the outer portion of the dam
Enlargement of the breach through the mechanism of one or more sudden structural collapses
due to the hydrostatic pressure force exceeding the resisting shear and cohesive forces
Enlargement of the breach width by slope stability theory
Initiation of the breach via piping with subsequent progression to a free surface breach flow
Erosion transport for either non-cohesive (granular) materials or cohesive (clay) materials
Wahl (2004) suggests that the BREACH model is constrained, as other similar models, in that it does
not adequately model head cutting erosion processes that dominate the breaching of cohesive soil
embankments. Another limitation of the BREACH model is that the breach hydrograph prediction is
simulated without incorporating downstream effects, such as tail water and dynamic effects on the flow
within the upstream reservoir, because it uses level pool reservoir routing. This program may be used in
conjunction with other programs to simulate downstream dynamic effects using the breach parameter
results (i.e., breach width and development time) as input into a separate flood routing model that can
determine the breach hydrograph itself, while accounting for dynamic water-level effects of the reservoir
and downstream tail water effects (Fread, 1988; Wahl, 2010).
The NWS DAMBRK and FLDWAV software contain a BREACH subprogram that simulates piping
and overtopping failures in earthen dams when users provide the typical dam and reservoir
characteristics, thus generating breach parameters.
A dam breach simulation in HEC-HMS may be computed through two breach methods: overtopping or
piping. For overtopping, the failure is simulated at a point on the top of the dam and expands in a
trapezoidal shape until it reaches the maximum size input into the program. The piping dam breach
function of HEC-HMS is used to simulate failures caused by piping inside an earthen dam. The failure
begins with the water naturally seeping through the dam core until it increases in velocity and quantity
enough to begin eroding fine sediments out of the soil matrix. The piping failure uses many of the same
user-input parameters as the dam overtopping breach; however, it also requires the initial piping
elevation and piping coefficient. The time growth curve may be specified in HEC-HMS as either linear,
non-linear (sine wave), or user specified.
Similar to the precursor program HEC-1, HEC-HMS uses a level pool routing procedure for the
upstream reservoir to estimate the breach hydrograph. The reservoir is represented as either a controlled
or uncontrolled water body with the assumption of level pool and a monotonically increasing storage-
outflow function. Hydrologic routing employs the continuity equation and an analytical or empirical
relationship between reservoir/reach storage and the discharge. Output results from HEC-HMS include
a resulting breach hydrograph that must be used in conjunction with other software, such as HEC-RAS,
for downstream routing of the generated flood wave.
The main advantage of using HEC-HMS to simulate a dam failure is the ease of program use. The
program does not suffer from the instability issues of its counterpart HEC-RAS. A major difference
between HEC-HMS and HEC-RAS, , is that HEC-HMS uses level pool routing whereas HEC-RAS uses
dynamic pool routing (full St. Venant equations of conservation of mass and conservation of
momentum) for reservoir drawdown. However, dynamic routing requires detailed bathymetric data for
the reservoir, which are frequently difficult and expensive to obtain. Level pool routing, on the other
hand, only requires a simple stage-storage curve for estimating reservoir drawdown. Goodell et al.
(2009) argued that dynamic routing is generally a more accurate method for estimating reservoir
drawdown. However, level-pool routing is often an adequate method for drawdown computation. This
is especially true for small reservoirs that are roughly equal in length and width and do not have a
considerably long fetch length.
Floodplain. The dam safety simulation differs from reservoir routing in that the elevation-outflow
relation is computed by determining the flow over the top of the dam (dam overtopping) and/or through
the dam breach (piping/internal erosion), as well as through other reservoir outlet works. The elevation-
outflow characteristics are then combined with the level pool storage routing to simulate a dam failure.
A dam breach is simulated in the HEC-1 program using the methodology incorporated by Fread in the
NWS DAMBRK program (Fread, 1979). Structural failures are modeled by assuming certain
geometrical shapes for the dam breach. The outflow from a dam breach may be reduced by backwater
from downstream constrictions or other flow resistances. HEC-1 allows a tail water rating curve or a
single cross-section (and a calculated normal-depth rating curve) to be used to reflect such flow
resistance. Submergence effects are calculated in the same manner as in DAMBRK. The dam-break
simulation assumes that the reservoir pool remains level and routes the flood wave downstream using
steady-state theory (USACE, 1998).
The following discussion on HEC-RAS is adapted from user support documents developed by the
USACE.
The steady-flow component of the modeling system uses a standard step method intended for the
solution of water surface profiles for steady, gradually varied flow. The basic computations are based
on the one-dimensional energy equation in which energy losses are evaluated by friction and
contraction/expansion of the channel. The momentum equation may be used when the water surface
profile is rapidly varied in conditions such as a mixed flow regime. The system can handle a full network
of channels, a dendritic system, or a single river reach. The steady-flow component is capable of
modeling subcritical, supercritical, and mixed flow regime water surface profiles. To perform a steady-
state analysis for routing a resulting breach flow downstream in HEC-RAS, an upstream boundary
condition must be provided in the model. This boundary condition is the peak outflow generated from
the breach hydrograph that has been determined externally, in such forms as HEC-HMS, NWS
BREACH, or an empirical equation.
The unsteady component of the HEC-RAS modeling simulates one-dimensional unsteady flow and can
perform subcritical, supercritical or mixed flow regime computations. The governing equations for
unsteady flow are the conservation of mass (continuity) and momentum equations derived from the full
equations of motion (St. Venant equations). Upstream boundary conditions typically consist of an inflow
hydrograph from the upstream watershed into a defined reservoir. For a dam breach analysis, the
reservoir outflow is dynamically routed downstream.
Failure modes integrated into the HEC-RAS model include overtopping and piping. Additional failure
modes may be approximated with variations to one of those two methods. Overtopping failures start at
the top of the dam while a piping failure can start at a specified elevation/location and grow to the
maximum specified extents. Breach parameters, such as breach width, depth, side slopes, and
development time are estimated external to the model. Values for the breach size and development time
are needed to produce a reliable estimate of the outflow hydrographs and resulting downstream
inundation areas.
In HEC-RAS, both steady-state and unsteady-flow analysis use the same set of geometric data. This
geometric data includes the reservoir storage volume, dam and downstream channel characteristics,
cross-sectional data, etc. Differences in results between these two routing methods are a result of the
computation procedures and inclusion of flow attenuation in unsteady-flow routing. The ASPFM has
noted a generally small computational difference of 0.1 to 1 foot between steady and unsteady-flow
analysis based on hypothetical event analysis (Altinakar, 2008). Further suggesting that while the
difference between the two methods can be outside of this specified range, these differences do not
necessarily mean that unsteady flow is more accurate than steady flow. The ASPFM has identified three
key features between the steady-state and unsteady flow that provide computation differences:
1. Losses: Steady-flow losses computations use absolute differences in velocity head at adjacent
cross-sections multiplied by an expansion or contraction coefficient, whereas unsteady-flow loss
computations are computed by the momentum equation.
2. Friction Slope: Average friction slope between cross-sections is determined by averaging the
conveyance method for steady flow. For unsteady flow, the average friction slope between cross-
sections is computed directly from a simple average of the computed friction slopes.
HEC-RAS can perform inundation mapping of water surface profile results directly using the RAS
Mapper or the external HEC-GeoRAS tool. Using the HEC-RAS geometry and computed water surface
profiles, RAS Mapper creates an inundation depth and floodplain boundary dataset. Additional
geospatial data can be generated for analysis of velocity, shear stress, stream power, ice thickness, and
floodway encroachment data. HEC-GeoRAS is a set of GIS tools that prepare the geometric date for
import into HEC-RAS and generate the flood inundation data from the HEC-RAS output.
CHAPTER THREE
3.1. Location
Melka Wakana Dam is located in the highlands of Bale Zonal Administration about some 280Km.south-
east of Addis Ababa. It lies 70N of the equator. The Dam is at the Wabe Shebelle River a large water
course, flowing along the south-east cost of the country towards the territory of Somalia.
The 2300m-2400m elevation of altitude has highly influenced the climatic situation of the vicinity. The
mean annual temperature is not greater than 13-14C (Max 28C).the dam is in the upper course of the
river where the mean annual flow is 827x106 m3 and the maximum high-water flow is 530 m3/s. The
terrain conditions of the region are favorable for creating a reservoir for the over-year regulation,
capacity 763 m3, and for installing a derivation hydro power plant.
The rock-fill dam with a length of 2,000 m and maximum height of 42 m is filled with local materials
with the central loamy core and the rock apron slopes. The areal cement grouting and the cement-grout
curtain to a depth of 25 to 30 m are provided in the dam foundation. The automatic flood gate without
shutters on the top edge of the discharge structure is designed for flood discharge with flows of up to
640 cubic m/s. The water diversion chute (horizontally curved) has a variable grade over the length and
ends in a ski jump spillway, which dumps water into the river channel. Construction of dam started on
1983 and commissioned on 1988.
CHAPTER FOURE
Reservoir characteristics: The reservoir characteristics consist of reservoir storage elevation curve and
reservoir surface area elevation curve.
Dam characteristics: This category includes data about name of dam, dam type, dam size, location of
the dam, elevation of downstream toe of dam, design water storage pool elevation, maximum flood
surcharge elevation, spillway crest elevation, crest of dam elevation, and height of the dam measured
from downstream toe to the crest, and category of the dam.
General Information: This category of data is for general information purposes. It includes
jurisdictions of the dam owner (city, town, and country area), geographic information, watershed
boundary, and others.
Downstream Information: Data gathered under this category includes bank stations, reach stations,
downstream developments, cross section plots, Manning roughness coefficients, and other pertinent
hydraulic structures.
Inflow Hydrograph: The inflow hydrograph data category includes the flood events hydrograph
provided by the dam owner.
Several researchers have developed peak flow regression equation form historic dam failure data. The
peak flow equation were derived from data for earthen, zoned earthen, earthen with impervious core. In
general, the peak flow equation should be used for comparison purposes.
Once a breach hydrograph is computed in HEC-RAS, the computed peak flow from the model can be
compared to these regression equations as a test for reasonableness.
The maximum breach outflow that will be obtained from the analysis should be checked for its
reasonableness. Literatures recommended that one can check the reasonableness of the maximum breach
outflow obtained by one method with other methods.
For verification of the reasonableness of the value of breach out flow obtained by the analysis we have
to compare it with the value obtained by empirical formula as shown above or with the envelope. But
the envelope as discussed in chapter three will not be the true upper bound because it only taken in to
account fourteen historical dam failure incidents. Therefore, the value obtained using the empirical
relationship suggested by MacDonald-Langridge-Monopolis as show above will be used as an upper
peak breach outflow.
3579.313
6000
8000
10000
11067
12028.59
14000
16293.9
19000
wa
21010.96
e b
23259.54 s
ha
24878.96
b e
26669.94
l
34340.81
44281.12
47449.56
56280.88
62161.39
65000
72901.56
wabe shabel
MELK WAK
2,505.00
2,500.00
2,495.00
2,490.00
2,485.00
2,480.00
0.00 100.00 200.00 300.00 400.00 500.00 600.00 700.00 800.00 900.00
Volume (MCM)
Froehlich (1995a): Froehlich utilized 63 earthen, zoned earthen, earthen with a core wall (i.e., clay),
and rock fill data sets to develop as set of equations to predict average breach width, side slopes, and
failure time the data that Froehlich used for his regression analysis had the following ranges
Froehlich (2008): In 2008, Dr. Froehlich updated his breach equations based on the addition of new
data Dr. Froehlich utilized 74 earthen, zoned earthen, earthen with a core wall (i.e., clay), and rock fill
date sets to develop as set of equations to predict average breach width, side slopes, and failure time.
The data that Froehlich used for his regression analysis had the following ranges:
Height of the dams: 3.05-92.96 meters
Volume of water at breach time: 0.0139-660.0m3x 106
AAIT, School of Civil and Environmental Engineering Page 49
Dam Breach Analysis & Inundation Map for Melka Wakana Dam
Froehlich’s regression equations for average breach width and failure time are:
tf = 63.2(Vw/(ghb2))0.5
Where:
Bave = average breach width (meters)
Ko = constant (1.3 for overtopping failures, 1.0 for piping)
Vw = reservoir volume at time of failure (cubic meters)
hb =height of the final breach (meters)
g = gravitational acceleration (9.80665 meters per second squared)
tf = breach formation time (seconds)
froehlich’s 2008 paper states that the average side slopes should be:
1.0H:1V overtopping failures
0.7H:1V otherwise ( i.e piping/seepage)
hb (C+hbZ3/2)
Where:
Wb = bottom width of the breach (meters)
hb = height from the top of the dam to bottom of breach (meters)
Z3 = Z1+Z2
Z1 = average slope (Z1:1) of the upstream face of dam
Z2 = average slope (Z2:1) of the downstream face of dam
Zb = side slopes of the breach (Zb:1),0.5 for the Macdonald method
The implicit formulation of the St. Venant equation is well-suited from the standpoint of accuracy for
formulating unsteady flows in a natural channel. Therefore, HEC-RAS is chosen for unsteady state flood
routing, and this technique simultaneously computes the discharge, water surface elevation, and velocity
throughout the river reach. The following parameters are crucial in running HEC-RAS to perform
unsteady flow routing:
After the routing reach is established by the boundary locations, cross sections are obtained to represent
the reaches. Cross section locations are measured from downstream to upstream.
For the purposes of this study, default values of expansion and contraction coefficients are used
throughout the unsteady state analysis. The program by default assigns a value of 0.3 and 0.1 for
expansion and contraction coefficient, respectively.
By referring the Wabe Shebelle river bed and bank materials and Chow, Manning n values 0.048 and
0.042 are taken for the banks and flow channel respectively.
CHAPTER FIVE
The dam breach analysis and inundation map for Melka Wakana Dam as a testing basis involved testing
a number of dam breach parameters. The parameters defined for the reservoir and river component of
the analysis were prepared based on existing data and some empirical formulas. With the aid of
hydrologic and hydraulic modeling software, reservoir and river flow routings were carried out to
establish relationships among the characteristics influencing a peak flow at the dam and specified
location in the downstream. The findings were discussed in the following sub-sections.
Froehlich (1995a)
The Froehlich (1995a) method assumes a side slope of 1.4H:1V for an overtopping breach and 0.9H:1V
for a piping breach. Given the breach height of 42 meters, this yields a bottom width for the breach of
Wb = 307.76 meters for overtopping and Wb = 224.03 meters
Froehlich (2008)
The Froehlich (1995a) method assumes a side slope of 1.0H:1V for an overtopping breach and 0.7H:1V
for a piping breach. Given the breach height of 42 meters, this yields a bottom width for the breach of
Wb = 248.96 meters for overtopping and Wb = 194.42 meters
To compute the bottom width of the breach, the method says to use side slopes of 0.5H: 1V.The user
must also estimate an average side slope for both the upstream and downstream embankment of the dam.
In our case average side slope of 2.5H:1V were used for both upstream and downstream. The bottom
width equation is
Wb = Veroded-hb 2(CZb+hbZbZ3/3)
hb (C+hbZ3/2)
Where:
Wb = (2.45x106-422(10*0.5+42*0.5*5/3))/(42(10+42*5/2))
Wb = 493.76 meters
tf = 0.0179 (Veroded) 0.364
tf = 0.0179 (2.45x106 )0.364
tf = 3.79 hours
From here, all three set of parameters should be entered into HEC-RAS software and run as separate
breach plans. This will result in three different breach outflow hydrographs. However, once the
hydrographs are routed downstream, they will begin to converge towards each other.
For our case using Probable Maximum Flood (PMF) conditions analyzed overtopping and piping
Failures mode.
In this Dam Break Analysis, using mixed flow regime simulation, both upstream and downstream
boundary conditions (inflow hydrograph and rating curve, respectively) and dam dimensions were
identified.
EG Max WS
WS Max WS
2500
Crit Max WS
Ground
2400
Elevat io n (m)
2300
2200
2100
10000 15000 20000 25000
Main Channel Distance (m)
Plan: Overtopping 1995 River: wabe shabel Reach: wabe shabel RS: 23259.53
40000 Le gend
35000
Flow
30000
Flow (m3/ s)
25000
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Figure 5.3: the discharge flowing out of the dam during the dam break for overtopping
Plan: Froehlich 1995a River: wabe shabel Reach: wabe shabel RS: 23259.53
30000 Le gend
25000 Flow
20000
Flow (m3/ s)
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Figure 5.4: the discharge flowing out of the dam during the dam break for piping
Plan: Froehlic 2008 River: wabe shabel Reach: wabe shabel RS: 23259.53
30000 Le gend
25000 Flow
20000
Flow (m3/ s)
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Figure 5.5: the discharge flowing out of the dam during the dam break for overtopping
Plan: Froehlich 2008 River: wabe shabel Reach: wabe shabel RS: 23259.53
30000 Le gend
25000 Flow
20000
Flow (m3/ s)
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Figure 5.6: the discharge flowing out of the dam during the dam break for piping
25000
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Figure 5.7: the discharge flowing out of the dam during the dam break for overtopping
Plan: MacDonald piping River: wabe shabel Reach: wabe shabel RS: 23259.53
35000 Le gend
30000 Flow
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Figure 5.8: the discharge flowing out of the dam during the dam break for piping
From the above HEC-RAS model results we get three different peak flow, but selected the method which
have larger value than the other. Ones we select used selected method to make downstream routing,
inundation map and emergence action plane.
MacDonald and Langridge-Monopolis gives the largest peak flow, therefor use the result of MacDonald
and Langridge-Monopolis for downstream routing, inundation map and emergence action plane.
Overtopping Piping
For overtopping
Plan: MacDonald ovtop River: wabe shabel Reach: wabe shabel RS: 23259.53
40000 Le gend
35000 Flow
30000
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Figure 5.9: the discharge flowing out of the dam during the dam break for overtopping.
Plan: MacDonald ovtop River: wabe shabel Reach: wabe shabel RS: 19000
40000 Le gend
35000 Flow
30000
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Plan: MacDonald ovtop River: wabe shabel Reach: wabe shabel RS: 12028.59
40000 Le gend
35000 Flow
30000
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Plan: MacDonald ovtop River: wabe shabel Reach: wabe shabel RS: 1000
40000 Le gend
35000 Flow
30000
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
For piping
Plan: MacDonald piping River: wabe shabel Reach: wabe shabel RS: 23259.53
35000 Le gend
Flow
30000
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Figure 5.13: the discharge flowing out of the dam during the dam break for piping
Plan: MacDonald piping River: wabe shabel Reach: wabe shabel RS: 19000
35000 Le gend
Flow
30000
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Plan: MacDonald piping River: wabe shabel Reach: wabe shabel RS: 12028.59
35000 Le gend
Flow
30000
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
Pla n: MacDonald piping River: wabe shabel Reach: wabe shabel RS: 1000
35000 Le gend
Flow
30000
25000
Flow (m3/ s)
20000
15000
10000
5000
0
1800 2400 0600 1200
12Aug2015 13Aug2015
Time
For overtopping
Figure 5.17: inundation map for overtopping failure and depth of water level in meter
For piping
Figure 5.18: inundation map for piping failure and depth of water level in meter
Warn and evacuate the isolated residences at risk. These procedures are to supplement and be
used in conjunction with County’s Emergency Operation Plan.
Failure of the Melka Wakana dam could cause significant damage to (all roads and isolated residences
downstream of the dam within the danger reach) located downstream of the dam.
AAIT, School of Civil and Environmental Engineering Page 66
Dam Breach Analysis & Inundation Map for Melka Wakana Dam
Operating procedure
The dam will be inspected periodically each year for maintenance and distress signals.
The dam observer will inspect the dam and Flood Warning for the area and complete the following tasks.
The dam observer will note & record water levels in reservoir and the rate at which the pool is
rising.
If the dam shows signs of internal piping (muddy seepage exiting the downstream embankment),
erosion, slope failures, blocked spillways, or other ominous distress signs, the dam observer send
massage to police to roadblock downstream roads and warn any isolated residences in the danger
reach. The dam observer may contact to designated engineer to provide assistance.
If the pool level rises too within one meter of the dam crest, the dam observer will contact the
County Emergency Operations Center to dispatch police to roadblock downstream roads and
warn any isolated residences in the danger reach.
Preventative actions
Listed below are potential emergency actions which may prevent or delay the failure of the dam. They
should be considered based on site-specific conditions, as well as the risk of failure and risk to
employees.
Possible Actions to be taken in the event of:
Provide erosion protection on downstream slope by placing riprap or other appropriate materials.
Plug the seepage with appropriate material such as (riprap, hay bales, bentonite, sandbags, soil,
or plastic sheeting if the leak is on upstream face of dam).
Lower the reservoir level until the flow decreases to a non-erosive velocity or stops leaking.
Place a sand and gravel filter over the seepage exit area to minimize loss of embankment soils.
Continue lowering the reservoir level until the seepage stops or is controlled. Refill reservoir to
normal levels only after seepage is repaired.
sandbags
riprap
pumps
pipe
Laborers.
CHAPTER SIX
6.1. Conclusions
Dam failure places populations at risk; however, tools exist to evaluate the contingencies. HEC-RAS
used in concert with HEC-GeoRAS provide the capabilities to create a river hydraulics model, simulate
a dam failure, and map the resulting flood wave. The main goal of this study was to create a flood hazard
map for Melka Wakana Dam along with a flood protection measure framework. Based on these flood
hydrographs, unsteady-flow simulations were performed in order to define areas where overtopping and
piping will occur during large flood events. The hydraulic modeling results were incorporated into a
representative flood hazard map. Based on the encountered hazard situation, a flood protection measure
framework was developed as Emergency Action Plan (EAP).
The Melka Wakana Dam breach has been simulated for overtopping and piping breach. With a breach
Bottom Width 493.76m and time of 3.79h.The simulated results reached peak discharges of
36,527.15m3/s and 32,627.70m3/s, for overtopping and piping breach respectively. The maximum
discharge at the lower end, 22 km below the dam, was reduced to 35,179.22m3/s and 30,159.74m3/s,
for overtopping and piping breach respectively and Dam overtopped by 18cm.
6.2. Recommendation
Dam breach analysis and inundation map for Melka Wakana Dam as a testing basis involved making a
number of assumptions based on literature reviews and historic data. In the real world, there is a large
degree of uncertainty associated with the breach parameters and breaching outflow estimation. It would
be helpful to minimize the ambiguities associated with breach parameters estimation using different
modeling software and analysis techniques for obtaining a wider range of dam and reservoir
characteristics and downstream river characteristics data.
In this study, only one dimensional unsteady flow routing technique was used to carry out dam breach
analysis. There are a number of assumptions in the modeling software. It would be helpful to utilize a
different version of the software and enhance the findings of this study.
Finally, I would like to recommend the dam owner to give special attention to the Dam breach analysis
and make a detail investigation by using the latest dam breach software’s.
REFERENCES
Predicting Embankment Dam Breach Parameters, A Needs Assessment, Tony L. Wahl, USBR,
Denver, Colorado, USA, August 1997.
Ram S. Gupta, 2nd edition (January 10, 2001). Hydrology and Hydraulic System, Waveland
Press.
Singh, K. P., and A. Snorrason, 1984, “Sensitivity of Outflow Peaks and Flood Stages to the
Selection of Dam Breach Parameters and Simulation Models,”
Smith, N. (1971). “A history of dams.” Peter Davies, LondonStedinger, J., D. Heath and K.
Thompson, Risk Analysis for Dam Safety Evaluation; Hydrologic Risk U.S. Army Corps of
Engineers, 1996.
U.S. Army Engineers (1960 and 1961) Floods Resulting from Suddenly Breached Dams,
Hydraulic Model Investigation. Corps of Engineers, Waterways Experiment Station. Vicksburg,
Mississippi.
U.S. Army Engineers (31 August 1994) Hydrologic Engineering Requirements for Reservoirs,
Engineering Manual (EM) 1110-2-1417, Department of the Army, U.S. Army Crops of
Engineers, Washington, DC.
USBR (1986), Guidelines to Decision Analysis.
USBR (1999) A Procedure for Estimating Loss of Life Caused by Dam Failure, DSO-99-06,
Sedimentation and River Hydraulics Dam Safety Office.
USDOI. (2011) Reduce Dam Safety Risk Modernization Blueprint / Implementation Phase 1:
Launch Risk Reduction / Inundation Mapping/Modeling Subproject Report. U.S. Department of
the Interior Enterprise Architecture. March, 2011.
Using HEC-RAS for Dam Break Studies, US Army Corps of Engineers, Hydrologic Engineering
Centers, August 2014.
Velleux, M., Takamatsu, M., and Yang, L. (2010). “Literature Review: Comparison of Common
Watershed Models for Rainfall, Runoff Sediment Transport, and Contaminant Transport and
Fate.” Mahwah, NJ.
Wahl, T.L (1998), “Dam Safety Research Report: Prediction of Embankment Dam Breach
Parameters, A Literature Review and Needs Assessment” U.S. Department of the Interior Bureau
of Reclamation Dam Safety Office, Denver.
Andrew Steininger, “Dam Overtopping and Flood Routing with the TREX Watershed Model”,
M.sc Thesis, Colorado University, 2014.
Bicknell, B.R., Imhoff, J.C., Kittle, J.L. Jr., Donigian, A.S. Jr. and Johanson, R.C. (1997).
“Hydrological Simulation Program—FORTRAN”, User’s manual for version 11. U.S.
Environmental Protection Agency (EPA), National Exposure Research Laboratory, Athens, GA.
Crawford, N.H. and Linsley, R. K. (1966). “Digital simulation in hydrology; Stanford Watershed
Model IV.” Technical Report No. 39, Department of Civil Engineering, Stanford University,
Stanford, CA.
DHI. (2005). “MIKE SHE Technical Reference, Version 2005.” DHI Water and Environment,
Danish Hydraulic Institute, Denmark.
Dams and hydropower in Ethiopia, October 2015, Wikipedia, the free encyclopedia.
Ewen, J., Parkin, G., and O’Connell, P.E. (2000). “SHETRAN: distributed river basin flow and
transport modeling system.”