REC-ERC-B5-7

December 1985
Engineering and Research Center

U. S. Department of the Interior

Bureau of Reclamatwn
7-2090 (4.61)
Bureau ot Reclamation TECHNICAL RE PORT STANDARD TITLE PAGI
3. RECIPIENTS CATALOG NO.

5. REPORT DATE

Hydraulic Model Studies December 1985

of Fuse Plug Embankments 6. PERFORMING ORGANIZATION CODE

D-l 532

7. AUTHOR(S) 6. PERFORMING ORGANIZATION

REPORT NO.

Clifford A. Pugh REC-ERC-85-7

9. PERFORMING ORGANIZATION NAME AND ADDRESS 0. WORK UNIT NO.

Bureau of Reclamation
1. CONTRACT OR GRANT NO.
Engineering and Research Center
Denver, CO 80225
3. TYPE OF REPORT AND PERIOD
COVERED
12. SPONSORING AGENCY NAME AND ADDRESS

Same
4. SPONSORING AGENCY CODE

DIBR
15. SUPPLEMENTARY NOTES

Microfiche and/or hard copy available at the E&R Center, Denver, CO

Editor: RNW(c)
16. ABSTRACT

Hydraulic model studies were conducted to help develop guidelines for designing fuse plugs where
these structures would be appropriate for controlling reservoir outflows from large floods with long
return periods. Model embankments at scales of 1: 10 and 1:25 simulated prototype fuse plugs from
10 to 30 ft (3 to 9 m) high. Eight tests were conducted for a variety of embankments and flow
conditions.
The erosion rates and discharge coefficients determined in this study can be used in computer flood-
routing programs to aid in the design of fuse plug embankments and to assess the effects of various
options.
The sand filter, embankment material, and material gradation were found to have significant effects
on the rate of erosion. It was also found that the configuration of the approach channel has a significant
effect on the hydraulics of the flow through the fuse plug and on the erosion rate.

17. KEY WORDS AND. DOCUMENT ANALYSIS

a. DESCRIPTORS-- fuse plugs/ erosion/ flood control/ modeling/ sediment/ auxiliary spillways/
embankments/ dam breaching/ discharge coefficients/ hydraulics

b. IDENTIFIERS-- Oxbow Project, Idaho

c. COSATI Field/Group 13M COWRR: 1313 SRIM:

16. DISTRIBUTION STATEMENT 19. SECURITY CLASS 21. NO. OF PAGE
(THIS REPORT)
Available from the Notional Technical Information Service. Operations 33
UNCLASSIFIED
Division, 5285 Port Royal Road, Springfield. Virginia 22161.
20. SECURITY CLASS 22. PRICE
(THIS PAGE)
(Microfiche and/or hard copy available from NTIS)
UNCLASSIFIED
REC-ERG85-7

HYDRAULIC MODEL STUDIES

OF FUSE PLUG EMBANKMENTS

by
Clifford A. Pugh

December 1985

Hydraulics Branch
Division of Research and Laboratory Services
Engineering and Research Center
Denver. Colorado

UNITED STATES DEPARTMENT OF THE INTERIOR BUREAU OF RECLAMATION

 ACKNOWLEDGMENTS
A team of engineers from the Engineering and Research Center guided the investigations, which were
conducted in the hydraulics laboratory at the Engineering and Research Center, Bureau of Reclamation,
Denver, Colorado. In addition to the author, the team comprised the following persons:
l Edward W. Gray, Jr., embankment dams specialist (principal investigator)
l Raymond G. Acciardi, geotechnical engineer
l Chih T. Yang, civil engineer
l Steven Higinbotham, supervisory civil engineer
l Greg G. Hammer, civil engineer

Cheryl States and B. A. Callow, civil engineering technicians, placed the model embankments. Cheryl States
did the computations for model placement and composed many of the figures in the report. John C. Craig
videotaped the tests and compiled and edited a video tape presentation on the research. Wayne Lambert
and John Norquist were the photographers. The technical advice given by Russell A. Dodge on sediment
transport was very valuable during the study.
The technical advice on modeling techniques given by Henry T. Falvey is greatly appreciated.
Edward Gray, Jr. provided valuable suggestions and review comments throughout the study
The technical review by T. J. Rhone is greatly appreciated.

As the Nations principal conservation agency, the Department of the

Interior has responsibility for most of our nationally owned public
lands and patural resources. This includes fostering the wisest use of
our land and water resources, protecting our fish and wildlife, preser-
ving the environmental and cultural values of our national parks and
historical places, and providing for the enjoyment of life through out-
door recreation. The Department assesses our energy and mineral
resources and works to assure that their development is in the best
interests of all olir people. The Department also has a major respon-
sibility for American Indian reservation communities and for people
who live in Island Territories under U.S. Administration.

I I
The research covered by this report was funded under the
Bureau of Reclamation PRESS (Program Related Engi-
neering and Scientific Studies) allocation No. SO-3, De-
velopment of Criteria for Use in Design of Fuse Plug
Embankments in Auxiliary Spillways.
L I

The lnformatil,n x:or~i.s:r)--d in thus report was developed for the Bureau
ot H?clamatici:. r,i: \r~ar:an:y as to the accuracy, usefulness, or complete-
ness is exoresse:i or ,!npi~ed

ii
 CONTENTS
Page
Glossary of symbols ..__. .._. ._. _. ._. .___. V

Purpose _. _. _. _. _. _. 1

Introduction.. _. _. _. _. 1

Conclusions _. _. . _. . _. _. . . . _. _. ._. ._. . . .._. ____. _. _. 1

Literature on fuse plug erosion .... .... ... ... .... ... .... ... ... .... .... ... ... ..... ... ..... ... ..... ... ..... ..... ... ..... ..... ... 2

The model ................................................................................................................................. 2

Description ............................................................................................................................. 2
Similitude ............................................................................................................................... 3
Hydraulic similitude .............................................................................................................. 3
Sediment transport ............................................................................................................. 3
Settling-velocity adjustment ................................................................................................. 4
Erosion-rate scale ratio ........................................................................................................ 5
Embankment design,. .......................................................................................................... 5
Structural similitude, ............................................................................................................ 6
Model measurements .............................................................................................................. 6
Model results ......................................................................................................................... 6

Analysis ..................................................................................................................................... 7
Effect of embankment design features ...................................................................................... 7
Core .................................................................................................................................. 7
Pilot channel ..................................................... ................................................................. 7
Sand filter .......................................................................................................................... 8
Size of embankment ........................................................................................................... 8
Hydrauiics of flow through the opening .................................................................................... 8
Broad-crested weir ............................................................................................................. 8
Discharge coefficients ........................................................................................................ 8
Projection to prototype .......................................................................................................... 9
Comparison with Oxbow field test ........................ ........................... ..................................... 9
Example calculation, ............................................................................................................... 9

Bibliography .___. ___. _.. ._. ___. __ _._. __. _. . . .__. 10

TABLES
Table

1 Fuse plug model test data, . . .._. __. __. __. __. ___. __. _... ..__. ..__. ._. 7
2 Recommended discharge coefficients __.._.,,..,,,...,_..._...._...__..._..............................__........... g
3 Oxbow field-test data ___. _._, ___.__. _._, __. __, __._. .__. _.... ______. ..__.

FIGURES
Figure Page

1 Model fuse plug embankment, test No. 7.. ....................................................................... II

2 Model operation, test No. 4 ............................................................................................
3 Flow-measurement weir calibration curve ......................................................................... ii
4 Initial breach viewed through end wall, test No. 8. ............................................................
5 Forces on a sediment particle .......................................................................................... 15
6 Dimensionless unit sediment discharge curves versus dimensionless shear
and grain Reynolds number ........................................................................................ 15

Iii
 CONTENTS - Continued

Figure Page

Settling velocity of sand and silt in water.. ...................................................................... 16

Fuse plug embankment and pilot channel cross sections .................................................. 17
9 Prototype gradation curves ............................................................................................ 18
10 Gradation curves, sand filter.. ......................................................................................... 18
11 Gradation curves, sand and gravel .................................................................................. 19
12 Gradation curves, slope protection ................................................................................. 19
13 Gradation curves, rockfill.. .............................................................................................. 20
14 Model fuse plug embankment placement.. ....................................................................... 20
15 Model fuse plug embankment.. ....................................................................................... 21
16 Flow through the pilot channel showing the failure mode of the impervious core.. .............. 21
17 Definition sketch of geometric features of model fuse plug embankment ........................... 22
18 Lateral erosion rate, test No. 1 ....................................................................................... 23
19 Lateral erosion rate, test No. 2.. ..................................................................................... 23
20 Lateral erosion rate, test No. 3.. ...................................................................................... 24
21 Lateral erosion rate, test No. 4 ....................................................................................... 24
22 Lateral erosion rate, test No. 5.. ..................................................................................... 25
23 Lateral erosion rate, test No. 6.. ..................................................................................... 25
24 Lateral erosion rate, test No. 7.. ..................................................................................... 26
25 Lateral erosion rate, test No. 8 ....................................................................................... 26
26 Views of the washout process (3 sheets) ....................................................................... 27
27 Schematic of the lateral erosion process.. ....................................................................... 30
28 Broad-crested weir flow profiles, tests No. 6, 7, and 8 .................................................... 30
29 Weir formula discharge coefficients, test No. 7.. .............................................................. 31
30 Relative erosion rates after breach.. ................................................................................. 31
31 Lateral erosion rates (after initial breach) for a fuse plug embankment with the
geometric features of tests 5 and 7.. ........................................................................... 32
32 Embankment design for Oxbow field test.. ....................................................................... 33

iv
 GLOSSARY OF SYMBOLS
A Cross-sectional area
B Width of base of fuse plug
b Width of base of fuse plug downstream from the core
Discharge coefficient
E Cauchy number
6 Correlation coefficient
D Reservoir water depth above fuse plug base
d Sand grain diameter
d 90 Grain diameter at which 90 percent of the grains have smaller diameters
E Modulus of elasticity
6 Euler number
ER Lateral erosion rate
f Function of
f Friction factor
6 Froude number
FC Critical tractive force
FkV Weight force
9 Gravitational acceleration
H Height of fuse plug
h Pilot channel depth
J Crest length
k Equivalent sand grain roughness
L Distance along fuse plug crest (breach length)
I Length
M Structural merit number
P Top length of pilot channel along crest
P Pressure
Water discharge
:, Sediment discharge
9 Unit water discharge
4. Unit sediment discharge
4.* Dimensionless unit sediment drscharge
R Hydraulic radius
R Reynolds number
R Boundary or grain Reynolds number
S Water surface slope
SS Sand specific gravity
T Impermeable core thickness
r Clay core thickness in the model
r Time
u Shear velocity
V Average water velocity
W Sediment settling velocity
W Width of fuse plug crest in direction of flow
We Weber number

Specific force of water

Specific force of sediment
Angle of core with horizontal
Kinematic viscosity of fluid
Density of fluid
Density of sediment grain
Surface tension
Standard deviation of particle sizes
Critical shear stress
Bed shear stress
Dimensionless shear stress
Friction angle

Subscript r refers to the ratio between prototype and model values

Subscript m refers to the model.
Subscript p refers to the prototype.

V
 PURPOSE can be regulated as the fuse plug is washing out by
controlling the flow through the gated structures.
The increasing size of design floods is causing dam
designers to investigate more economical methods The rate of lateral erosion as the fuse plug washes
of providing additional spillway capacity. In many out is of primary importance. The rate of increase in
cases, auxiliary spillways with fuse plug embank- downstream flow depends not only on the rate of
ments can provide an economical alternative to pass- lateral erosion, but also on the elevation of the res-
ing all the flow through concrete structures. The ervoir. The rate of lateral erosion depends on the
Bureau of Reclamation undertook the research de- gradations and the types of materials used to con-
scribed in this report to help develop design guide- struct the fuse plug, the depth of flow above the base
lines to be used where a fuse plug embankment of the fuse plug, and the geometry of the fuse plug
would be appropriate. section (crest width, angle of the upstream .and
downstream slopes, and configuration of the zoning).

INTRODUCTION The total discharge through an auxiliary spillway with

a fuse plug embankment is controlled by the eleva-
A fuse plug is a zoned earth and rockfill embankment tion of the grade sill or nonerodible foundation be-
designed to wash out in a predictable and controlled neath the fuse plug, by the width of the spillway
manner when the flow capacity needed exceeds the channel, by the depth of water above the grade sill
normal capacity of the service spillway and the outlet or foundation, and by the width of the grade sill in
works. A fuse plug embankment is designed to pre- the direction of flow.
clude use of the auxiliary spillway during minor
floods, much the same as spillway gates. In many The rates of lateral erosion and coefficients of dis-
cases, auxiliary spillways with fuse plug embank- charge obtained from the research described in this
ments can provide an economical alternative to pass- report can be used in a computer flood-routing pro-
ing all of the flow through concrete structures. gram to predict the downstream flows and reservoir
water surface elevations. The effects of varying the
A fuse plug is designed as a dam, stable for all con- rate of lateral erosion and the elevation of the pilot
ditions of reservoir operation except for a flood that channel on the reservoir elevation and on the auxiliary
will cause it to breach. The washout of a fuse plug spillway discharge can also be determined.
should begin at a preselected location; general ov-
ertopping of the entire fuse plug should not occur. The auxiliary spillway should be designed according
to standard practice. The flow velocities must be suf-
The preferred method is to initiate breaching of the ficient for the water to carry the eroded fuse plug
fuse plug with reservoir water. When the reservoir material downstream to avoid clogging of the return
level reaches a predetermined elevation, a low spot channel to the natural stream course. If excessive
in the embankment crest, called a pilot channel, will sediment deposition is anticipated, a site-specific
be overtopped. model study may be required to design the return
channel. If the approach channel to the fuse plug is
By placing highly erodible materials in the pilot chan- relatively shallow and long, losses may limit the dis-
nel section, breaching will occur rapidly, and the rest charge and further study may be required.
of the fuse plug embankment will wash out laterally
at a constant, predictable rate without overtopping.
The auxiliary spillway flow will increase at a constant CONCLUSIONS
rate. This automatic breach feature is desirable be-
cause it reduces the possibility of mechanical or hu- l A properly designed fuse plug embankment will
man error when operation of a flood-relief wash out in an orderly and predictable manner
mechanism is critical. When a wide auxiliary spillway when additional flow capacity is needed to pass a
is required, it may be desirable to sectionalize the large flood through a reservoir. The fuse plug will
fuse plug with concrete separation wails. By using preclude the use of the auxiliary spillway during
successively higher elevations for the fuse plug small floods.
crests and pilot channels in each section, the wash-
out process can be matched with successively less l The lateral erosion rate (after the initial breach) is
frequently occurring floods. The entire fuse plug primarily a function of the erosion rate of the em-
would not wash out unless the full capacity of the bankment material and not a function of the
auxiliary spillway was needed. strength of the impermeable core.

If the project has a gated outlet works or a service l The erosion rates and discharge coefficients de-
spillway, the total outflow and the reservoir elevation termined in this study can be used in flood-routing

1
 computer programs to help design fuse plug em- to 0.38 m) high and 8.8 ft (2.7 m) long, at scales of
bankments. 1:lO and 1:25, (fig. 1). The model size was based
on the maximum flow available in the laboratory. A
Ratios of depth of water to embankment height and flow of 21.5 ft3/s (0.61 m3/s) was made possible by
depth of water to weir width have significant ef- using two pumps operated in parallel.
fects on erosion rates.
The overall model was 46 ft (14 m) long by 26 ft
The sand filter, embankment material, and grada- (8 m) wide. Flow entered the model through two l-
tion have significant effects on erosion rates. ft (0.3-m) pipes and passed through a rock baffle into
a 17-ft (7.6-m) by 5-ft (1.5-m) deep headbox (fig. 2).
A model design method is described that com- The headbox simulated a reservoir in a prototype
pensates for the fact that the Reynolds number is structure. The water surface level in the headbox was
normally too low to properly simulate sediment controlled by a 25-ft (7.6-m) long adjustable-height
transport in a Froude scale hydraulic model. This weir along one side. Water flowing over the adjust-
method uses settling velocity adjustments and di- able weir plunged into a side channel, then it passed
mensionless unit sediment discharges to adjust the through a flow-measurement weir. The flow-
model grain sizes and/or the model sediment den- measurement weir, a combination-type weir, was
sity. calibrated for flow versus water surface elevation in
the side channel. The lower 1 ft (0.31 m) of the weir
had a 90 V-notch. Above the V-notch were 2-ft
LITERATURE ON FUSE PLUG EROSION (0.61 -m) extensions on each side at a 15 angle with
horizontal. Above these 15 extensions were 1.5-ft
Fuse plug embankments have been designed and (0.46-m) long vertical sides. The calibration curve for
constructed for mine tailing dams, for levees, and for the measurement weir is shown on figure 3. The cal-
controlling the flow in auxiliary spillways. However, ibration was done in three parts corresponding to the
there has not been a documented case of a fuse plug three different sections of the weir. During each test,
controlled spillway actually operating. Most of the the calibration curve shown on figure 3 was used to
information in the literature is associated with studies compute the discharge through the measurement
conducted in 1959, to design a fuse plug-controlled weir. Water surface elevations were monitored with
spillway for the Oxbow Project, on the Snake River capacitance-type water surface probes. The data
between Idaho and Oregon [ 1, 2, 31. were recorded on disc with a microcomputer.

The Oxbow Project has two spillways, each with a The fuse plug embankment platform was located at
design flow capacity of 150,000 ft3/s (4250 m3/s), the end of the headbox, 2.5-ft (0.76 m) above the
which is on the order of the loo-year flood. The total headbox floor. One end wall of the platform was con-
discharge capacity of 300,000 ft3/s (8500 m3/s) cor- structed from transparent plastic to observe the in-
responds to the inflow design flood. The original de- itial breach and lateral erosion process (fig. 4). A
sign required three radial gates to control each sloping platform downstream from the horizontal
spillway. Later, the Idaho spillway, on the right abut- fuse plug platform led to a tailbox where the sediment
ment, was changed to fuse plug control [3]. The stud- was deposited before the water returned to the lab-
ies conducted to confirm the design assumptions oratory supply reservoir. A typical test used the fol-
included 1:20 and 1:40 scale model tests in the lab- lowing procedure:
oratory and a 1:2 scale field test at the damsite.
1. With the adjustable control weir at a low level,
Another study of erosion mechanics and washout the valves controlling the two inlet pipes were
time rates of erodible control embankments was opened; the entire flow of 21.5 ft3/s (0.61 m3/s)
made using hydraulic models at the University of entered the headbox, passed over the control weir
Windsor, in 1977 [4]. This study analyzed theoretical and through the measurement weir.
equations and compared calculations with model re-
sults.
2. The test was started by raising the reservoir
water surface with the control weir to a predeter-
THE MODEL mined level where water began flowing through the
pilot channel.
Description
3. As the fuse plug embankment washed away,
The model was designed to simulate typical proto- more water passed through the breach. The water
type fuse plugs from 10 to 30 ft (3 to 9 m) high. The surface was kept at a constant level by gradually
model embankments were from 0.5 to 1.25 ft (0.15 raising the control weir.

2
 4. The flow through the measurement weir, the in the model. Therefore, a model can usually simulate
level of the water surface in the reservoir, and the the critical prototype forces for a certain type of flow.
time were recorded continually. Each test was
videotaped and photographed. Hydraulic similitude. - The flow of water through
a fuse plug is primarily determined by gravity and
5. Flows through the breach were computed by inertia forces; the other forces may be neglected. The
subtracting the measurement weir readings from ratios between the model and prototype are deter-
the total (initial) flow. mined from the Froude law (equation 1). The scale
relations according to the Froude law are as follows:
Similitude
Ratio Scale relation (model/protototype)
Hydraulic model studies are used because of the Length = L, (geometric ratio)
large number of variables involved in hydraulics and
because of the differences in boundary configura- Area = L,
tions. The physical behavior of a model simulates, in
a known manner, the physical behavior of the pro- Volume = Lr7
totype.

There are several types of similarity. Geometric sim- Time = L, 1/2 (for gr = 1)
ilarity exists when the ratios of all homologous di-
mensions between the model and the prototype are Velocity = v, = L,/t, = L 112
the same. The geometric scale ratio, or length ratio,
is denoted by L, = L,/L,, where the subscripts m and Discharge II Q = Lr 512
p refer to the model and the prototype, respectively.
Sediment transport. - Vanoni [6] discussed the im-
Kinematic similarity, or similarity of motion, implies
portant variables involved in the present knowledge
that the ratios of velocities and accelerations be-
of sediment transport in a section on Fundamentals
tween the model and prototype are equal.
of Sediment Transport. He reduced the sediment
discharge rate, Q,, to the following relationship.
Dynamic similarity requires that the ratios of homol-
These symbols are defined in the glossary.
ogous forces between the model and prototype be
the same. In the study of hydraulic phenomena, the
0, = f (Q, R, d, u, p, ps, oLI, w, g) (6)
primary forces that influence the flow are the forces
due to gravity, viscosity, pressure, surface tension,
Models involving erosion of noncohesive bed ma-
and elasticity. The inertial force is the vector sum of
terial must simulate tractive stress (5,) because the
all the forces [5]. The following dimensionless num-
tractive stress causes the drag force required to
bers relate inertial force to each of the forces listed
overcome the forces holding a particle in place
above.
(fig. 5).

Froude number (inertia/gravity), F, = -$ (1) The tractive stress on a particle fluctuates because
of the turbulence. The drag force and turbulence are
a function of the Reynolds number (equation 2).
Reynolds number (inertia/viscosity), I?, = t (2) Therefore, a model operated according to Froude
scaling does not necessarily simulate the tractive
forces and sediment erosion accurately. In some
models the sediment sizes must be adjusted to com-
Euler number (inertia/pressure), f, = pv* (3)
AP pensate for a Reynolds number that is too low.

Weber number Shields developed a diagram relating dimensionless

pLv2 shear stress (r*) to a boundary or grain Reynolds
(inertia/surface tension), W, = ~ (4)
number (R). Shields used this diagram to define crit-
ical shear stress (rJ (the stress required for incipient
motion of sediment). This concept has been ex-
Cauchy number (inertia/elasticity), C, = L!!L2
E (5) panded by others to include dimensionless unit sed-
iment discharge.
It is apparent that a model fluid cannot simulate all
of these properties concurrently. However, in most
cases several of the forces will be absent or negligible
Vanoni [6] used Taylors data to show that dimen- Substituting equation (15) into equation (14),
sionless unit sediment discharge at low transport lev-
els falls very close to the Shields
motion
transport,
(figure 6). To properly
the dimensionless
curve for incipient
simulate
unit sediment
sediment
discharge
?=($)(y,r.y)
rate (cJ~) must be the same in the model and the To determine f in an open channel, the Reynolds
prototype. number (I?,) is computed according to the following
equation.
For a model with a grain Reynolds number (R) greater
than 5 and less than 100, the unit sediment discharge R,, = 4RV/v (17)
rate for the model would be higher than that for the
prototype (if the model sand grains are sized ac- where R = the hydraulic radius. The relative rough-
cording to geometric scaling). Because the dimen- ness (needed to determine f) is defined as:
sionless shear stress (r*) is about the same in the
model and the prototype where: relative roughness = K,/4R (18)

ud where KS = rugosity.
Grain Reynolds number, R = - 03)
u
Kamphius [8] found that

and
KS = 2 d,, (19)

Dimensionless shear stress, r* = where d,, = the particle diameter at which 90 percent
(ys ?y)d () of the grains are smaller in diameter.

it can be shown that dimensionless shear stress is The hydraulic radius can be taken as the flow depth,
a form of the Froude number and the density ratio if the channel is relatively wide.
of the sediment and of the water.
If a model is scaled geometrically according to Froude
scaling (r,, = r,*), the model unit sediment discharge
The shear velocity, u* = (10)
rate (q,*) will be too great in the range 5 L R <. 100.
Therefore, the model should be adjusted to properly
simulate sediment transport in this range. A diagram
therefore, r,= pu** (11)
of settling velocity (w) of sand and silt particles in
water (fig. 7) illustrates that small particles (< 1 mm
and the unit force, y = pg (12)
in diameter) settle at slower velocities as the particles
become smaller. For particle diameters larger than
Substituting equation (12) into equation (11):
(1 mm), the settling velocity is a function of the par-
ticle diameter (d) to the l/2 power. This is consistent
r, = yu/g (13) with Froude scaling for velocity, V, = L, (see the
previous section on hydraulic similitude).
Substituting equation (13) into equation (9):
Settling-velocity adjustment. - By increasing the
size of a model sediment grain, the settling velocity
r*= ($) (&) (14) can be corrected to the proper value for Froude scal-
ing. According to geometric scaling, a 1: 10 scale
model of prototype sand 2.0 mm in diameter would
The first term in equation (14) is in the form of a use sand 0.2 mm in diameter. However, the settling
Froude number, the second term is the ratio of the velocity would then be about 0.066 ft/s (0.02 m/s)
densities of the water and the sediment, when grav- (see fig. 7). when it should be 0.161 ft/s (0.049
ity is factored out. m/s), according to Froude scaling. If the model par-
ticle size is increased from 0.2 to 0.4 mm, the set-
It is sometimes more convenient to compute r* by tling velocity is corrected to 0.161 ft/s (0.049 m/s),
using a form of equation (9) relating r++ to the Darcy- the proper value for Froude scaling.
Weisbach friction factor ( f). Rouse [7] has shown that
the shear velocity, The effect of settling velocity adjustment on the di-
mensionless sediment discharge rate (q,*) is shown
(15) on figure 6. Note that the model values of rc before
the settling velocity adjustment are about the same

4
 as prototype values they simulate. Tests No. l-5 sim- ence between a fuse plug embankment and a typical
ulate the 25-ft (7.6 m) high prototype embankment, rockfill or earthfill dam is the arrangement of the im-
and tests No. 6, 7, and 8 simulate the 12.5-ft (3.8 m) pervious core. The core of a fuse plug embankment
high prototype embankment. However, the value of is inclined so that when the downstream material is
CL* must be the same in the model and prototype to washed away, pieces of the core break off from
properly scale the time rate of sediment transport. bending under its own weight and under the water
When the model grain sizes are adjusted for settling load. The core material is normally silt or clay. The
velocity (as described above) the value of t* de- sand filter prevents piping through cracks that de-
creases, while the value of R increases. This brings velop in the core and keeps windblown silt and clay
the model value of 9, much closer to the projected from infiltrating the downstream embankment ma-
prototype value of qsc (upper pair of curved lines on terial. The compacted sand and gravel and com-
figure 6). In this study, the model grain sizes were pacted rockfill are designed to be noncohesive and
computed using this method of settling-velocity ad- easily erodible once the washout process begins.
justment to account for the low grain Reynolds num- The prototype gradation curves for each zone are
ber. This method applies to noncohesive materials in shown on figure 9. A range of acceptable sizes are
the model and in the prototype, and must be checked shown with the gradation simulated in the model
for each grain size and each model flow condition. If study indicated by a dashed line.
model Reynolds number (R) is less than 5, a lighter
sediment specific force (rs) could be substituted to The model and prototype gradation curves for each
match 9s*. If R is greater than 100, no adjustment zone are shown on figures 10 through 13. These
is necessary. model gradation curves were determined by making
settling-velocity adjustments to the grain sizes de-
Erosion-rate scale ratio. - After the settling- termined by geometric scaling. The adjustments in
velocity is adjusted as described above, the erosion grain sizes do not significantly affect the shapes of
rate scales according to the Froude law. Velocities the model gradation curves.
scale according to L, lj2; therefore, lateral erosion rate
is: The pilot channel section was designed to wash out
quickly when the water flowed through the pilot
(ER), = Lr* (20) channel. A slightly larger rockfill material with fewer
sand sizes was used in this section to ensure a rapid
The small-scale model tests conducted for the Ox- break. The prototype pilot channel was designed to
bow Project used prototype diameter uniform-sized be 3 ft (0.9 m) deep: 1 ft (0.3 m) of water depth and
materials in the model. For these materials, Tinney 2 ft (0.6 m) of freeboard. This depth of water was
and Hsu [I] concluded that the erosion rate ratio determined to be adequate to initiate a breach during
would be, (ER), = I!-,~. Chee [4] derived the following the Oxbow study. The width of the pilot channel was
equation for the erosion-rate scale ratio, investigated in this model study. The side slopes of
the pilot channel were set at 1: 1; however, this value
could be varied in the prototype. The gradation and
(ER), = Lro375 (S, - 1),2 d,Ol3 (21)
compaction of the noncohesive materials are impor-
tant factors in determining the erosion rate. As the
where S, is the sand specific gravity. If (S, - I), = 1 materials are compacted, more tractive force is re-
(the same density sand in the model and the proto- quired to remove the grains because there are more
type) and d, = L, (geometric scaling), then (ER), = grains per unit volume and more contact and inter-
Lp505. This is very close to the Froude scaling ratio locking between them. A well-graded mixture of
obtained in this analysis, (ER), = lrl/*. However, grain sizes requires more tractive force to erode than
Chees method does not account for a low grain Rey- uniform-sized material requires. Smaller particles fill
nolds numbers in the model. the voids between larger particles, making the mix-
ture more dense and creating more contact between
For (S, - I), = 1 and d, = 1 (prototype-sized grains the particles.
in the model), the erosion-rate ratio derived by Chee
is approximately that derived by Tinney and Hsu, For the reasons cited above, great care was taken
Lr0375 versus L,j3. in placing the model fuse plug embankments. Rela-
tive density tests were conducted on the model ma-
Embankment design. - The fuse plug embankment terials before they were placed. (See gradation
was designed with the same zones found in most curves on figs. 10, 11, and 13.)
zoned-material or rockfill dams. The arrangement of
these zones is shown on figure 8. The main differ- The method consisted of placing a predetermined
weight of material in a given volume to obtain 70
Numbers in brackets refers to bibliography. percent relative density. The proper proportion of

5
each of the material sizes was mixed to obtain the The ratio of the moduli of elasticity must equal the
desired gradation. A known weight of the mixture model scale ratio. However, it is difficult to find a
was then placed in a known volume in the model (fig. model core material that has a modulus of elasticity
14). The volume in the model was controlled by using low enough to satisfy this ratio and maintain a seal.
wooden forms 1% to 2 in (38 to 51 mm) thick. The Therefore, a mixture of 10 percent clay and 90 per-
main embankment downstream from the core (zones cent sand with a modulus of elasticity approximately
3 and 6 on fig. 8) was placed in layers using this equal to that of the prototype was used.
method. The compaction was obtained by using a
compressed-air powered vibrator and tamping tools. Because the model modulus of elasticity (E,,) was too
When the placement of this zone was complete, the large, the core thickness had to be reduced to com-
forms were removed and the corners were trimmed pensate. A clay core reduced in thickness (1/3of geo-
to the correct slope. The sand filter, core, embank- metric scaling) was used in most of the tests. This
ment upstream from the core, slope protection, and thickness resulted from computing the correct mo-
gravel surfacing were then installed using templates ment of inertia in the model to compensate for the
and tamping tools. A completed model fuse plug em- modulus of elasticity being too large in the model.
bankment is shown on figure 15. The remainder of the core thickness was built with
sand sprayed with a stabilizing agent (which added
Structural similitude. - The impervious core was no strength).
not simulated as part of the hydraulic modeling be-
cause the cohesive clay portion does not fail as a Model Measurements
result of sediment erosion. The core is designed to
break off in pieces from the weight of the water and During each test several measurements were made
embankment material above it, as the non-cohesive to document the washout.
material downstream washes away. Figure 16 is a
schematic diagram illustrating the failure mode of the 1. Bypass flow was recorded at 3-second inter-
materials in the pilot channel. The core fails in a sim- vals. The measured weir flow was subtracted from
ilar manner during the lateral erosion process, as the the total inflow to obtain the flow through the
material on the face of the embankment downstream breach.
from the core is washed away.
2. The reservoir level was recorded continuously.
The structural behavior of the core material was sim-
ulated qualitatively, because the prototype core ma-
terial strength will vary a great deal. A structural 3. Flags were placed on the top of the embank-
analysis of the prototype core as a cantilevered slab ment at 1-ft intervals, and a grid pattern was
indicates that only about 2.9 ft (0.88 m) of core painted on the downstream face of the embank-
would overhang horizontally before it would break ment. The lateral erosion rates (ER) were recorded
[assuming a high tensile strength in the core of 1000 by noting the time that the erosion reached each
lb/in2 (6895 KPa)]. flag. These rates were checked by viewing the
video tape.
The structural behavior of the core material is gov-
erned by gravity and elasticity forces. The structural 4. Each test was filmed using video tape cameras,
merit number(M) is the dimensionless ratio of gravity still photographs, and slides. One video camera
forces to elasticity forces was located downstream from the embankment,
and the other filmed the washout through the
acrylic plastic end wall.
(22)

5. The discharge, reservoir level, and the time

where E = modulus of elasticity of the core. were recorded on a floppy disk with the aid of a
microcomputer.
For structural similitude:
Model Results
(23) The erosion process and lateral erosion rate are par-
tially dependent on the geometric configuration of
If Y,~ = y,, , then the embankment. Figure 17 shows the configuration
of the model embankments tested and defines the
symbols. Table 1 lists the values of each of the per-
(24) tinent features for each test. These values are listed
as dimensionless ratios of the fuse plug height (If).

6
The rate of erosion was consistent throughout the To assess the effect of the core thickness on the
test for any one model configuration and flow con- lateral erosion process, one test was conducted with
dition (see figs. 18 through 25). The erosion rate a clay core thickness indicated by geometric simili-
varied as the model configuration or flow condition tude (test No. 2). This thickness was about three
was changed. The first constant in the regression times greater than that required for structural simil-
equations, shown on each graph, is the lateral ero- itude. The strength was about 25 times greater than
sion rate. required. During this test the initial breach proceeded
about the same as in the other tests up to the point
Before each test the reservoir level was held constant of the first core break, which did not occur naturally.
at a level below the pilot channel invert, with a long The first break was assisted by manually breaking
adjustable weir. The test was begun by raising the the clay, after which the washout process proceeded
water surface to a level equivalent to a 1 ft (0.3 m) much the same as in the other tests. The lateral ero-
water depth in the prototype pilot channel. The ma- sion rate was 1.52 ft/min (0.463 m/min), which was
terial downstream from the core eroded down to the only 2 percent less than that in an essentially identical
base of the fuse plug. When the support was re- test with a thinner clay core [test No. 4, ER - 1.55
moved from beneath, a piece of the core broke off. ft/min (0.472 m/min)]. During the lateral erosion
This process recurred until the material in the pilot process in test No. 2, the noncohesive material
channel was completely washed away. Figure 26 is downstream from the core eroded at a constant rate
a series of photographs illustrating the erosion proc- and the core broke off in bigger pieces than it did in
ess. the other tests.

After the initial breach, the embankment eroded lat-

erally. The flow eroded the face of the embankment Test No. 2 indicates that the lateral erosion rate (after
downstream from the core in a steady, continuous the initial breach) is primarily a function of the erosion
manner. As the noncohesive material washed away rate of the noncohesive material downstream from
and the support was removed, pieces of the core the core and not a function of the core strength.
broke off. Figure 27 is a schematic diagram illus-
trating how water flowed around the core and eroded During test No. 3 the core was installed at an angle
the downstream embankment. 30 above horizontal (table 1). The material down-
stream from the core was shielded more by the core,
and the initial breach took longer. The lateral erosion
ANALYSIS rate was about the same. It was decided to use a
core angle of 45 for the remaining tests to prevent
An analysis was made of the model results to de-
excessive shielding of the downstream materials.
termine the effect of the pertinent geometric and flow
parameters.
Pilot channel. - Various widths and positions of the
Effect of Embankment Design Features pilot channel were tested. The location of the pilot
channel did not have a noticeable effect on the lateral
Core. - The role and the effect of the impervious erosion rate. The erosion rate for test No. 4, with
core were analyzed in the model study. A previous the pilot channel located near the center of the em-
section on Structural Similitude discusses the bankment, was the same in both directions, and
structural modeling of the core. A clay core one-third about the same as the erosion rate of a similar test
the thickness indicated by geometric scaling was with the pilot channel close to one end of the em-
used during most tests. bankment.

Table 1. - Fuse plu,g model test data.

Test H, 8 Sand ER,
No. ft Scale WH B/H b/H deg T/H W L/H P/H h/H filter D/J D/H tt/mln

1 1.0 1:25 0.4 4.4 3.1 45 0.12 0.04 0.0 0.24 0 12 Yes 0 21 0 92 1 74
2 1.0 1:25 .4 4.4 3.1 45 .12 .12 .12 .36 .12 No 21 .92 1 52
3 10 1:25 .4 4.4 4.0 30 .12 .04 .32 .48 .12 No .21 92 1.53
4 1.0 1:25 .4 4.4 3.1 45 .12 .04 3.24 .48 .12 No .21 .92 1 55
5 1.0 1:25 .8 4.8 3.4 45 .12 .04 0.45 .74 .12 Yes 15 .92 1.60
6 0.5 1:25 .8 4.8 3.4 45 .12 .04 .91 1.48 24 Yes .07 .84 0.68
7 1.25 I:10 .8 4.8 3.4 45 .12 .04 .51 0.88 .24 Yes .I7 .84 1 66
.15 .73 1 43
8 1.25 I:10 .8 4.8 3.4 45 .12 .04 1.60 3.20 24 Yes * 12 .60 0 63

Refer to figure 17 for definitions of the symbols.

1 ft = 0.3048 m
* The upstream water level (0) was lowered.

7
The width of the pilot channel controls the amount Q = CLH 32 (25)
of water passing through to initiate the breach. Qual-
itative observations of the model tests indicate that For critical depth, d, = (2/3)H, and using the Froude
the pilot channel width (p) should be about /2 of the number at critical depth:
fuse plug height (p/H = 0.5) to ensure that adequate
breaching flow passes through the pilot channel.
(26)
Sand filter. - Tests were run with and without the
sand filter surrounding the main embankment down- Equation 26 can be expressed as:
stream from the core. (See fig. 8 for embankment
zoning.) It was found that the sand filter has a sig- q zr g/(2/3) 3,?/./12
nificant effect on both the initial breach and the lateral
erosion. Without the filter, the water flowing through or
the pilot channel infiltrates the downstream nonco- q -= 3.09 H312 (27)
hesive material and partially saturates this zone, thus
prolonging the breaching process. The lateral erosion where the unit discharge, q = Vd,
rate is also significantly slower when the down-
stream sand filter is removed. Test No. 1 and test The coefficient (C) in equation (25) is 3.09 ftljL/s
No. 4 have identical embankment designs, with the (1.70 m117/s)
exception of the sand filter. The erosion rate for test
No. 4, without the sand filter, is about 11 percent For a broad-crested weir, the critical depth point is
slower. The volume of the compacted sand and actually slightly upstream from the downstream end
gravel zone is greater without the sand filter because of the crest, and the losses reduce the theoretical
the sand and gravel also occupy the sand-filter zone. value of C slightly. Empirical data [9] indicate that in
The compacted sand and gravel particles require the broad-crested weir range 0.08 c D/J 5 0.5, the
more tractive force to move because these particles theoretical discharge coefficient is reduced by a fac-
are larger and their sizes vary more than those in the tor of 0.848. Therefore, equation 25 would be:
sand filter material. This causes more particle inter-
locking. Q = 0.848 (3.09) LH312

Size of embankment. - The relative size of the em- or

bankment was investigated in test No. 5. The width Q = 2.62 LH? (28)
of the fuse plug crest was doubled, increasing the
area of the eroding face of the embankment. The Discharge coefficients measured during test No. 7
erosion rate for test No. 5 was about 8 percent less are shown on figure 29. These coeffrcrents were
than that for test No. 1. This is about the same per- computed from the discharge and water depth meas-
centage as the increase in the area of the cross sec- urements made during the test. The discharge coef-
tion of the downstream compacted sand and gravel ficients increase to a maximum value of 2.78 ft112/s
zone. (1.53 m1/2/s) for a breach length (L) of 4 to 6 ft (1.2
to 1.8 m). The coefficient drops slightly, to C = 2.62
Hydraulics of Flow Through the Opening ft/*/s (1.44 m/*/s) when the embankment is com-
pletely washed away. This value matches the em-
Broad-crested weir. - Flow through the opening is pirical value for a broad-crested weir (equation 28).
similar to flow over a broad-crested weir. The broad- The higher discharge coefficient during the washout
crested weir flow depends on the depth of water can be attributed to a longer effective weir length
above the crest (D) and the length of the crest (J). caused by the flow coming around the face of the
In the range of 0.08 < D/J c 0.5 [9], flow over a embankment (fig. 27). During test No. 6 (D/J -. 0.08)
horizontal crest is in the broad-crested weir flow the flow on the crest was controlled by friction. In
range. Figure 28 shows broad-crested weir flow pro- test No. 4 with erosion in both directions the relative
files for tests No. 6, 7, and 8. For broad-crested weir weir length was increased during washout by the
flow, the top surface and flow streamlines become flow coming around the face of the embankment on
parallel with the horizontal crest. The flow discharge each side (fig. 27). The discharge coefficient was 3.1
is controlled by the critical flow depth near the end ft/2/s (1.71 ml//s) during washout. Table 2 sum-
of the crest [IO]. If D/J <. 0.08 the flow depth is marizes the recommended discharge coefficients for
controlled by friction on the crest. flood-routing studies.

Discharge coefficients. - The following equation These discharge coefficients can be used in a com-
expresses discharge over a weir as a function of puter flood-routing program after the initial breaching
water depth. has occurred. The lateral erosion rate is a function

8
 Table 2. - Recommended discharge coefficients (equation 25). tion 29 would be changed accordingly. For example,
if the cross-sectional area of the downstream com-
C pacted sand and gravel section is decreased by 10
mli2/s ft112/S percent, the erosion rate predicted by equation 29
In the broad-crested weir should be increased by IO percent. If the water depth
range (0.08 < D/J c 0.5) (D) or crest length (J) are different from those on
During washout in one direction 1.51 2.75 figure 3 1, the erosion rate would be adjusted using
During washout in both directions 1.71 3.10
After washout is complete 1.44 2.62
figure 30. This adjustment would also be used if the
reservoir level changes during the washout process.
of the flow depth to crest length ratio (D/J) and the
depth to fuse plug height ratio (O/H). Tests No. 6 and Comparison with Oxbow Field Test
7 have the same D/H ratio (0.84), but different D/J
ratios (fig. 28). The result is a different erosion proc- The only data available on a prototype-size fuse plug
in operation are for a field test performed as part of
ess. For (D/J) > 0.12, the flow surface was still draw-
the design of a fuse plug control for an auxiliary spill-
ing down as it reached the embankment. This caused
a longitudinal vortex along the face of the down- way on the Oxbow Project on the Snake River in
stream compacted sand and gravel that accelerated Idaho. The %-scale field test of a 27 ft (8.2 m) pro-
the erosion rate. For D/J < 0.12, the flow was parallel totype embankment showed the same pilot channel
to the crest as it reached the embankment. The ero- breach and lateral erosion process indicated by this
model study. The following are the geometric and
sion was similar to streambank erosion with no vor-
flow parameters for the Oxbow field test (as defined
tex caused by drawdown to aid the erosion process.
on fig. 17).
Relative erosion rates are shown as a function of
water depth (D) and crest length on figure 30. This The Oxbow embankment design [I] is shown on fig-
ure 32. The gradation curve for the Oxbow zone 4
figure illustrates that the relative erosion rate is
(concrete aggregate) was similar to that of the pro-
a power function of the water depth ratio for
totype zone 3 (compacted sand and gravel) simulated
D/J ,: 0.12. For D/J > 0.12, the erosion rate is much
in this study (fig. 11).
faster. The data on this graph are scaled to the same
size embankment by the erosion-rate scale relation- Although the embankment designed for the Oxbow
ship according to Froude scaling. field test was slightly different from the embank-
ments tested in this model study, the erosion rate
This analysis illustrates that the length of weir (J) has was close to the erosion rate predicted by equation
a major effect on the erosion process. 29 (fig. 31).

Projection to Prototype During the later stages of the Oxbow field test, the
entire embankment downstream from the sand filter
The results from this study can be used to predict was zone 3 (well-graded) material. The erosion rate
the behavior of a prototype fuse plug embankment reduced from 336 to 72 ft/h (102 to 22 m/hr), dem-
designed as described in this report. The discharge onstrating the importance of material gradation. This
coefficients discussed in the previous sections can drastic reduction in erosion rate was caused by the
be used to predict the flow through a given size open- cohesion of the well-graded zone 3 material.
ing.
Table 3. - Oxbow field test data (1 ft = 0.3048 m).
The lateral erosion rate (after the initial breach) for a
given embankment design and flow depth can be ER,
predicted from the model tests. Figure 31 shows the H. Sand w
erosion rates for an embankment with the geometric ft W/H S/H b/H 0 T/H L/H p/H h/H falter D/H mln)

configuration given in table 1 for tests No. 5 and 7. 13.5 0.37 4.12 2.59 45 0.11 0.37 0.81 0.15 Yes 093 5.6
The flow depth was determined by the pilot channel
design, with 2 ft (0.6 m) of freeboard and 1 ft (0.3 m)
of water depth, regardless of the embankment Example Calculation
height. Equation 29 is an empirical equation that can
be used to estimate erosion rates for embankments Find the erosion rate and discharge through a fuse
of this configuration that are from IO to 30 fi (3 to plug with the following dimensions:
9 m) high. H = 17.5 ft (5.33 m) 8 = 45
ER= 13.2 H + 150 (29)
l/V = 10.0 ft (3.05 m) h = 3.00 ft (0.91 m)
If the configuration of the embankment or the flow
condition is changed, the erosion rate given by equa- P = 13.0 ft (4.00 m)

9
 The nonerodible base for the fuse plug is cut into ER reduction ratio = 0.70
rock in the abutment of the dam. The pilot channel
is at one end of the fuse plug and the fuse plug em- Adjusted ER = 0.70 (400) = 280 ft/h (85.3 m/h)
bankment is 2000 ft (610 m) long. The fuse plug
base is an average of 200 ft (61 m) wide in the di- If the reservoir level remains about the same during
rection of the flow. The upstream and downstream the washout, the embankment would wash out at a
slopes of the embankment are 2: 1. constant rate in 7.1 h [2000 ft/(280 ft/h) or 610 m/
(85 m/h)]. If the reservoir routing of the design flood
The solution is: indicates a change in the water surface elevation dur-
ing the washout, the erosion rate would also change
W/I-/ = 0.57 and h/H = 0.14 according to figure 30. The erosion rate should be
updated during the routing to accurately predict the
If it is assumed that the downstream edge of the core inflow/outflow relationship.
intersects the top of the fuse plug at the center (fig.
The discharge coefficient during the washout would
8), then
be 2.75 ft1/2/~ (1.51 m1/2/s) (table 2). After the wash-
out is complete the flow would be:
b/H = 3.29 (from geometry)
Q = CLH32 = (2.62) 2000 (15.5)32
If the pilot channel breaches when the water is 1 ft
(0.3 m) deep, D = H - h + 1 ft = 15.5 ft (4.72 m), Q = 317,800 ft3/s (9000 m3/s)
and

D/H = 0.89, D/J = 0.08 BIBLIOGRAPHY

[I] Tinney, E. R., H.Y. Hsu, Mechanics of Wash-
The erosion rate (ER), is then computed for the ref- out of an Erodible Fuse Plug, Journal of the Hy-
erence case (fig. 31, tests 5 and 7) using equation draulic Division, Proceedings, American Society
29: of Civil Engineers, vol. 87, No. HY 3, May 1961.

ER = 13.2 H + 150 = 381 ft/h (116 m/h) [2] Oxbow Hydroelectric Development, Idaho
Spillway With Fuse Plug Control, Model Studies
Because the reference embankment is larger than of Fuse Plug Washout, R. L. Albrook Hydraulic
that in the example (W/H = 0.8 and b/H = 3.4), an Laboratory and International Engineering Com-
adjustment is needed. The area of the cross section pany, Inc., Washington State College, Pullman,
downstream from the core for the example is about WA, August 1959.
5 percent less than that for the reference embank-
ment (from the geometry of the cross section). [3] Hsu, H.Y., Design of Fuse Plug Spillway, In-
Therefore, the lateral erosion rate should be in- dian Journal, of Power and River Valley Devel-
creased from the computed value, because there is opment, Hydraulic Structures Special Number,
less material to erode. vol. 91-2, pp. 308-311 and 335, 1962.

Adjusted ER = 1.05 reference ER = 1.05

x X 381 [4] Chee, S. P., Mobile Controls for Water Re-
= 400 ft/h (121 m/h) sources Projects, American Water Resources,
Water Bulletin, paper No. 77072, February
Water depth to fuse plug height ratio (D/H) is the 1978.
same in the example and in the reference (D/H =
0.89). However, the water depth to crest length ratio [5] Morris, H. M., Applied Hydraulic in Engineer-
(D/J) is much less in the example (0.08 vs. 0.17). ing, the Ronald Press Co., New York, NY, 1963.

The relatively long approach channel has a significant [6] Vanoni, Vito A., Sedimentation Engineering,
effect on the erosion rate (fig. 30). To make the ad- American Society of Civil Engineers, ASCE Task
justment for a long approach channel, the relative Committee for Preparation of the Manual on
erosion rate for D/J < 0.12 is divided by the relative Sedimentation of the Sedimentation Committee
erosion rate for D/J > 0.12. This ratio is multiplied of the Hydraulics Division, No. 54, New York, NY,
by the calculated erosion rate for the example (fig. 1975.
30):
[7] Rouse, H., Elementary Mechanics of Fluids,
Relative eroslon rate (D/J < 0.12, D/H= 0.89) ~ 0.62 _ o,70 John Wiley & Sons, Inc., London, pp. 201-203,
Relative erosion rate (D/J > 0.12, D/H = 0.89) 0.89 October 1948.

10
[8] Kamphius, J. W., Determination of Sand [IO] King, H. W., Handbook of Hydraulics, Mc-
Roughness for Fixed Beds, Journal of Hydraulic Graw Hill Book Co., New York, NY, 1963.
Research 12 (1974) No. 2.
Pugh, C. A., and E. W. Gray, Jr., Fuse Plug Em-
bankment in Auxiliary Spillways -Developing De-
[9] Bos, M. G., Discharge Measurement Struc- sign Guidelines and Parameters, In: Darn Safety
tures, Publication No. 20, International Institute and Rehabilitation, Fourth Annual USCOLD Lec-
for Land Reclamation Improvement, Wagenin- ture, Bureau of Reclamation, Denver, CO, January
gen, The Netherlands. 1984.

1 I
Figure 1 - Model fuse plug embankment, test No. 7. P801-D80944.

Figure 2. - Model operation, test No. 4. P80VD80945

13
 Figure 3. -Flow measurement weir calibration curve (1 ft = 0.3048 m,
1 ft3/S = 0.02832 m3/s).

Figure 4. -Initial breach viewed through end wall, test No.8. P801-D80946

14
 FLOW
I

F, = weight force
Fc = critical tractive
force
0 = friction angle

Figure 5. - Forces on a sediment particle.

CONSTANT SEDIMENT DI:SCHhRCE RELATIONSHIPS

Prototype

T 12.5 f t. Prototype

qs :t = constant

(--Settling velocity adjustment

,Shields curve for incipient motion

qs t=,oo1
.qsk.oool

RI -GRAIN REYNOLDS NUMBER

Figure6. - Dimensionless unit sediment discharge curves versus dimensionless shear and grain Reynolds
number.

15
 I I

10000 .O

SETTLING VELOCITY OF
SAND AND SILT
= 2.65 / _1000.0
/
i9
o/
I/:;
3 E
/
/ 1 IOO.OE
I

--.I
W
/ I-
,
:
a
- 10.0ii
W
-I
0
c
ix
a
1.0 n

--I
----I
0.1

f
0.01
---I
i
I II I I I lllll I I I llllll I Illlfll
0.00 I 0.01 0.1 1.0 10.0 100.0 I000.0

w=SETTLING VELOCITY m/s (X 102)

Figure 7. - Settling velocity of sand and silt in water (1 ft = 0.3048 m).

16
0I CORE MATERIAL

0
2 SAND FILTER

03 COMPACTED SAND AND GRAVEL

04 SLOPE PROTECTION

05 GRAVEL SURFACING

EMBANKMENT

0
6 COMPACTED ROCKFILL

www --

PILOT CHANNEL
Figure 8. - Fuse plug embankment and pilot channel cross sections

17
 PE

.RMAETER
FINE
OF PARTICLE

Figure 11. - Gradation

I
IAND
MEDIUM I COARSE

curves, sand and gravel.

I76
I
I FINE

1
GRAVEL
1 COARSE COBBLES

Figure 12. - Gradation curves, slope protection.

19
 Figure 13. - Gradation curves, rockfill.

Figure 14. - Model fuse plug embankment placement. P801-D80947.

20
 Figure 15. -Model fuse plug embankment. P801-D80945

Fw

?""7-
.7

Figure 16 Flow through the pilot channel showing the failure mode of the impervious core

21
Test H, e Sand ER,
No. ft Scale WH B/H b/H deg T/H VH L/H P/H h/H filter D/J D/H ft/min

1 1.0 1:25 0.4 4.4 3.1 45 0.12 0.04 0.0 0.24 0.12 Yes 0.21 0.92 1.74
: 1.0 1:25 .4 4.4 4.0
3.1 .12 .12 .12 .36 .12 No .21 .92 1.52
El .12 .04 .32 .48 .12 No .21 .92 1.53
4 1.0 1~25 .4 4.4 3.1 45 .12 .04 3.24 .48 .12 No .21 .92 1.55
5 1.0 1125 .8 4.8 3.4 45 .12 .04 0.45 .74 .12 Yes .15 .92 1.60
6 0.5 1:25 .8 4.8 3.4 45 .12 .04 .91 1.48 .24 Yes .07 .84 0.68
7 1.25 1:lO .8 4.8 3.4 45 .12 .04 .51 0.88 .24 Yes .17 .84 1.66
.15 .73 1.43
8 1.25 1:lO .8 4.8 3.4 45 .12 .04 1.60 3.20 .24 Yes .12 l .60 0.63

1 ft = 0.3048 m
l The upstream water level (0) was lowered.

t PILOT CHANNEL

Figure 17. - Definition sketch of geometric features of model fuse plug embankment.

22
 LATERAL EROSION RATE
6.0

7.0

6.0

a 58.0
;ii
W
&4-0..
l?i
= 3.0..

c
5 [
2.0 ..

+---+ TEST Cl
Y - 1.74 x- 3.98, co- la.997
1.0 ..

0.0 ::: :: :: :::::: : ::: :: :::: : ::: :: : : :: : ::: : :,

B m m 8 8
8 ,: ni : i ri f F: f

TIME Crnrn>

Figure 18. - Lateral erosion rate, test No. 1.

LATERAL EROSION RATE

a0

7.0 ..

6.0..

a
5.0..

?I

;;
E 4.0..
t
W

i$ 3.0..
i
s {
-I
2.0.. w TEST 12
Y - 1.92 x- 3848 I co- 0.998

1.0 ..

0.0 L
0.0/:::::::::::::::::::::::::::::::::::::::* 1
B
B
ss ii
i ii N
N ii ii i i

TIME Cmin>

Figure lg. - Lateral erosion rate, test No. 2.

23
 LATERAL EROSION RATE
a. 0

i
2.0 + -4 TEST Y3
Y - 1.53 x- 33.89 I cc- 0.995

Figure 20. - Lateral erosion rate, test No. 3.

LATERAL EROSION RATE

a0- TEST U4

7.0 __ et------f) EFJlSION TO THE LEFT

Y - 1.55 x- 28.58 , co- 1.090
+---+ EROSION TO THE RIGHT
Y - 1.55 x- 26.41 I co- 0.999

1.0..

a0 :::::::::::::::::::::::::::::::::::::::,
19 m e3 m m B
vi !? i;
TIME Cmfn>

Figure 2 1. - Lateral erosion rate, test No. 4.

24
 LATERAL EROSION RATE
a. 0

TEST 15
Y - 1.60 X- 11.24 , Co- 0.968

a0 ::::::::::::.:...*.,...,....,.,..,
m m ca B m m m m B
d < d 6 si , ni 6-i 4

TIME Cmin>

Figure 22. - Lateral erosion rate, test No. 5.

LATERAL EROSION RATE

U TEST #6
Y - a 88 x- a 77 , co- a Qsa

a0 c-H::::::::::::::::::::::::+::::::+
8 8 8 B B B 8 8 8 B B

Figure 23. - Lateral erosion rate, test No. 6.

25
 LATERAL EROSION RATE

TEST #7

+-mm+ Headcl.04 ft
Y - 1.66 x- 14.41 I co- 0.998
L---d H-ad-0.92 ft
Y - 1.43X- 11.60, Cc.- 1.000

Figure 24. - Lateral erosion rate, test No. 7

LATERAL EROSION RATE

8. 0

7.0 ._

6. 0 __ o-------o TEST 68
Y = 0.63 x- 2.55 I cc= 0.9!3s

Figure 25. - Lateral erosion rate, test No. 8.

26
(a) Water flowing through the pilot channel over the clay core. P801-D80949.

(b) After initial breach. PSO1-DSO950.

Figure 26. -Views of the washout process.

27
(c) The lateral erosion process is underway. PBO1-DBO951

Figure 26. -Views of the washout process. -Continued

28
(e) The reservoir elevation was held constant by the long adjustable weir in the background. PBO1-DBO953.

(f) The erosion process is almost complete. P801-D80954.

Figure 26. -Views of the washout process. -Continued

29
 Figure 27. - Schematic of the lateral erosion process. The water flows across the face of the embankment, around the
core, and erodes the noncohesive material downstream from the core.

Embankment for
Tests *7 and *8 oCritical depth
(D/J) Test #i (Scale I: lo)
0.20

0.15

0. IO

0 05

0.00 i

(S&Ii I :i15)

/
/ J c
/
/

,,i /

Figure 28. - Broad-crested weir flow profiles, tests No. 8, 7, and 8.

30
 Figure 29. - Weir formula discharge coefficients, test No. 7.

Em I.(

O.!

01 -oTEST#6,
I (D/J)=
I
q TEST#7; (D/J)=
A TEST#7; (D/J ) =
0.7 -eTEST#E;
-*TEST#E; (D/J)= .I2

0.6

E05
W
.
Wr
2
0.4

0.3

0.2

0.1

0.0 :
0.0 0.2 0.4 0.6 0.8 1.0
D/H

Figure 30. - Relative erosion rates after breach, used to adjust

computed erosion rates (see fig. 31) for varying water depths
and crest lengths (data values adjusted to the same scale).

31
 600

0 TEST NO. 5s (SCALE-la 25)

IJ PROTOTYPE , D/H-E. 92
500

400 :: ER = 13.2 H + 150

OXBOW FIELD TEST

A TEST NO. 7, <SCALE-l, 10)

0 PROTOTYPE , O/H-E. 84

100 &e----Model doto

f

In
.-I

l-i -HEIGHT OF FUSE PLUG <foot>

Figure 31. - Lateral erosion rates (after initial breach) for a fuse plug embankment with the geometric features of tests No. 5 and 7 (see table 1).
 0
I CORE

0
2 FILTER MATERIAL

0
3 WELL GRADED MATERIAL

0
4 CONCRETE AGGREGATE

Figure 32. - Embankment design for Oxbow field test

-IrS GOVERNMENT PRINTING OFFICE 1995--841432

33
 Mission of the Bureau of Reclamation

The Bureau of Reclamation of the U.S. Department of the interior is

responsible for the development and conservation of the NationS
water resources in the Western United States.

The Bureaus original purpose to provide for the reclamation of arid

and semiarid lands in the West today covers a wide range of interre-
lated functions. These include providing municipal and industrial water
supplies,. hydroelectric power generation; irrigation water for agri-
culture; water quality improvement; flood control; river navigation,.
river regulation and control; fish and wildlife enhancement,. outdoor
recreation; and research on water-related design, construction, mate-
rials, atmospheric management, and wind and solar power.

Bureau programs rnost frequently are the result of close cooperation

with the U.S. Congress, other Federal agencies, States, local govern
ments, academic institutions, water-user organizations, and other
concerned groups.

A free pamphlet is available from the Bureau entitled Publications

for Sale. It describes some of the technical publications currently
available, their cost, and how to order them. The pamphlet can be
obtained upon request from the Bureau of Reclamation, Attn D-822A,
P 0 Box 25007, Denver Federal Center, Denver CO 80225-0007.