ENERGY DISSIPATION ON FLAT‐SLOPED STEPPED SPILLWAYS:

PART 2. DOWNSTREAM OF THE INCEPTION POINT

S. L. Hunt, K. C. Kadavy

ABSTRACT. Many aging watershed dams require hazard classifications changes. As a result, these dams may no longer meet
state and federal dam safety regulations because of inadequate spillway capacity and flood protection. Rehabilitation options
for these embankments are often limited due to encroaching urban development. Roller compacted concrete (RCC) stepped
spillways are a popular choice in these situations because the spillway capacity can be increased with little or no additional
changes to the embankment dimensions. RCC stepped spillways are also selected because of the cost and time savings in the
construction of these structures. Design engineers require more information about the inception point location and the
approach velocity and energy dissipation in the spillway chute. These elements are important for properly dimensioning the
spillway training walls and stilling basin. Research and more specifically design guidelines for RCC stepped spillways
applied to small earthen embankments have been limited. A two‐dimensional, physical model was constructed to evaluate
the inception point, velocities, air concentrations, and energy dissipation in a 4(H):1(V) slope spillway chute having 38 mm
(1.5 in.) high steps. Model unit discharges ranging from 0.11 m3 s-1 m-1 (1.2 cfs ft-1) to 0.82 m3 s-1 m-1 (8.8 cfs ft-1) were
tested. The findings from this research show that a relationship developed by H. Chanson can be used to determine the
inception point location on stepped chutes with Froude surface roughness (F* ) ranging from 10 to 100 for slopes as flat as
4(H):1(V). Additionally, air concentrations near the inception point are approximately 0% and rapidly increase to 10%
slightly downstream of the inception point. These air concentrations continue to increase gradually to a constant as the flow
descends the chute. The study results show that energy losses increase from 30% when a normalized length (L/Li and L/Li* )
equals 1 to 73% when L/Li and L/Li* equals 3.5. A first attempt at providing an energy loss relationship at any point
downstream of the inception point is provided. This research will assist engineers with the design of stepped spillways applied
on relatively flat embankment dams.
Keywords. Air entrainment, Dam rehabilitation, Energy dissipation, Flood control, Inception point, Physical modeling,
Roller compacted concrete (RCC), Stepped spillways, Stilling basin.

R
oller compacted concrete (RCC) stepped spill‐ way is the same as the downstream slope of the dam. Typical
ways are not a new technology. According to face slopes of earthen embankments are 2(H):1(V) or flatter.
Chanson (2002), stepped chutes have been around Stepped spillway research for these flatter applications is
for centuries. In recent years, their popularity has limited, and design guidelines are scarce. Without model
increased for the rehabilitation of aging flood control dams. studies, the safety and integrity of a flat stepped spillway may
This technology is expected to be applied to thousands of be questioned, or the design may be overly conservative,
small flood control dams. RCC stepped spillways provide a leading to a more expensive project. The USDA Agricultural
means for increasing the capacity of existing flood control Research Service (ARS) Hydraulic Engineering Research
structures. Modifications to the dimensions of the existing Unit (HERU) is conducting research on stepped spillways
dam are often not required in conjunction with stepped spill‐ constructed on existing embankment dams. Air entrainment
ways, thereby making them more desirable over other rehabi‐ and energy dissipation in flatter (2(H):1(V) or flatter) stepped
litation designs. Additionally, RCC has cost and construction spillways may affect the design of the spillway training walls
time advantages when compared to conventional concrete. In and the stilling basin differently than steeper (2(H):1(V) or
most dam rehabilitations, stepped spillways are placed over steeper) stepped spillways. Therefore, the USDA‐ARS
the existing earthen embankment; thus, the slope of the spill‐ HERU is currently conducting research to evaluate the ef‐
fects that a 4(H):1(V) stepped spillway chute has on the in‐
ception point location, air concentrations, velocities, and the
energy dissipation for a specified range of skimming flows.
Submitted for review in April 2009 as manuscript number SW 7985; This article reports on research results on spillway perfor‐
approved for publication by the Soil & Water Division of ASABE in mance downstream of the inception point. This research will
December 2009. assist engineers in the design of these structures.
The authors are Sherry L. Hunt, ASABE Member Engineer, Research
Hydraulic Engineer, and Kem C. Kadavy, ASABE Member Engineer,
Agricultural Engineer, USDA‐ARS Hydraulic Engineering Research
Stillwater, Oklahoma. Corresponding author: Sherry L. Hunt, USDA‐
ARS Hydraulic Engineering Research Unit, 1301 N. Western St., Stillwater,
BACKGROUND
OK 74075; phone: 405‐624‐4135, ext. 222; fax: 405‐624‐4136; e‐mail: Design engineers of stepped spillways have expressed in‐
Sherry.Hunt@ars.usda.gov. terest in the affects of air entrainment and energy dissipation

Transactions of the ASABE

Vol. 53(1): 111-118 2010 American Society of Agricultural and Biological Engineers ISSN 2151-0032 111
on spillway and stilling basin design. Design guidelines are Growing
limited for stepped spillways, especially on the flatter Boundary Non-aerated Aerated Flow
stepped spillways associated with small embankment dams. Layer Flow Region Region

ÓÓÓÓÓÓÓÓÓÓÓÓ
Peyras et al. (1992), Rice and Kadavy (1996), Yasuda and Inception

ÓÓÓÓÓÓÓÓÓÓÓÓ
Ohtsu (1999), Chanson (2002), Chanson and Toombes Point
(2002), Boes and Hager (2003a, 2003b), Takahasi et al.

ÓÓÓÓÓÓÓÓÓÓÓÓ
Flow
(2006), Hunt et al. (2006), Hunt and Kadavy (2007, 2008),

ÓÓÓÓÓÓÓÓÓÓÓÓ
and Felder and Chanson (2008), among others, have re‐
searched flat (2(H):1(V) or flatter) stepped spillways. Each Li
has contributed to the understanding of these structures in
some fashion. For instance, Peyras et al. (1992) discussed the
flood flows that gabion‐style stepped spillways can with‐ Figure 1. Schematic of the inception point in relation to the stepped spill‐
way.
stand, and the cost savings for using this material in spillway
design. Rice and Kadavy (1996) reported on a specific model
Another key component in the design of the stepped spill‐
study of a stepped spillway on a 2.5(H):1(V) slope and indi‐
way is the energy dissipation that occurs in the spillway
cated that energy dissipation was significantly more in a
chute. Extensive research has been conducted on energy dis‐
stepped spillway when compared to a smooth spillway with
sipation, including work by Christodoulou (1993), Boes
the same chute slope. Chanson (2002) compiled extensive lit‐
(1999), Chanson, (2002), Chanson and Toombes (2002),
erature regarding stepped spillways concerning scaling ef‐
Chatila and Jurdi (2004), and Barani et al. (2005). The major‐
fects that occur in stepped spillway modeling, differences in
ity of this research was conducted on steep (2(H):1(V) or
nappe and skimming flow regimes, and development of rela‐
steeper) stepped spillways. Christodoulou (1993) discovered
tionships for inception point location and energy dissipation.
that the most important parameters governing energy dissipa‐
Chanson and Toombes (2002) reported energy dissipation
tion are the ratio of the critical depth to the step height and
rates where the ratio of head loss to total head was approxi‐
the number of steps. Boes (1999) and Chatila and Jurdi (2004)
mately 0.8 for slopes of 11(H):1(V) or flatter, and reported
indicate that energy dissipation is significantly larger in
that the air concentration distributions were similarly shaped
stepped spillways than conventional smooth spillways and
for smooth and stepped spillways under skimming flow re‐
that energy dissipation increases with increasing step height
gimes.
for a given spillway. Chanson (2002) suggests that greater en‐
Modeling air‐entrained flows can be difficult due to scale
ergy dissipation occurs in stepped spillways under nappe
effects. Scale effect describes slight distortions that are
flow conditions (i.e., water flows down a stepped spillway in
introduced in modeling when viscous forces and surface ten‐
a succession of free fall jets), which most likely occurs in
sion in highly air‐entrained flows are ignored. Boes (2000),
stepped spillways with large step heights on relatively flat
Chanson (1994a), Boes and Hager (2003a, 2003b), and Taka‐
slopes. However, the maximum design flow in these particu‐
hashi et al. (2006) have provided guidance to other research‐
lar spillways is rarely nappe flow.
ers on modeling techniques for reducing scale effects
The objective of this article is to provide velocity as ob‐
associated with air‐entrained flows. Generally, a scale of 10:1
tained by multiple measurement techniques, air concentra‐
or larger is an accepted scale for modeling stepped spillways
tion, and energy loss data downstream of the inception point
(Chanson, 2002). Boes and Hager (2003a) recommend a Re‐
for a range of flows. This generalized model study was con‐
ynolds number of 105 and a minimum Weber number of 100,
ducted on a 4(H):1(V) slope stepped spillway. Specifically,
respectively, to minimize scale effects.
this article provides a generalized relationship for determin‐
The inception point (fig. 1) is valuable in the design of
ing energy losses at any point downstream of the inception
stepped spillways because it indicates the location where sig‐
point for a 4(H):1(V) stepped spillway. For small embank‐
nificant flow bulking first occurs. The inception point is the
ment dams, the design flow is expected to be large, and the
location where the turbulent boundary layer reaches the free
stepped spillway chute is expected to be relatively short
surface. Flow bulking is the increased flow depth created by
compared to large gravity dams. In these instances, fully de‐
the mixture of air and water. As shown in figure 1, this flow
veloped air‐entrained flow may not be achieved within the
bulking occurs downstream of the inception point. If the in‐
spillway chute. In limited cases, air entrainment may fully
ception point occurs upstream of the expected design tailwa‐
develop at the free surface; consequently, flow bulking
ter and stilling basin, then the engineer should consider
downstream of the inception point would be expected. The
increasing the training wall height to account for this increase
air‐entrained data are the primary focus of this article. Flow
flow depth. Stepped spillways applied to existing embank‐
bulking directly impacts the design of the training walls for
ments typically have a relatively short spillway chute and
the spillway chute and the stilling basin. Additionally, energy
high design flow. Consequently, the inception point often oc‐
dissipation and velocities are important to properly size the
curs downstream of the expected tailwater or in the stilling
stilling basin. Therefore, this research is useful in the design
basin. In these cases, flow bulking has little impact on train‐
of stepped spillways and their stilling basins applied to small
ing wall design; however, the inception point location should
embankment dams.
be examined for each situation. Chanson (1994a) developed
a relationship to determine the inception point location for
primarily steep (2(H):1(V) or steeper) stepped spillways.
Hunt and Kadavy (2007, 2008) determined that Chanson's EXPERIMENTAL SETUP
relationship can be applied to stepped spillways with chutes A two‐dimensional model of a stepped spillway was
as flat as 4(H):1(V) for a Froude surface roughness (F*) rang‐ constructed for a generalized study to evaluate the inception
ing from 10 to 100 (Hunt and Kadavy, 2009). point, air concentrations, velocities, and energy losses asso-

112 TRANSACTIONS OF THE ASABE

 crest 152 mm
sta sta
-2.4 m 0.0 m sta 38 mm
broad-crest
step 0 1.8 m
sta
stepped chute step 12 3.7 m
1 sta
1.5 m sta sta
4 5.5 m
6.1 m 8.4 m
step 24 basin floor
step 36

Figure 2. Schematic of stations on a stepped spillway model.

ciated with flood flows. The model was constructed across the
full width of a 1.8 m (6 ft) flume. The flume walls are 2.4m (8
ft) tall, and the spillway model has a vertical drop of 1.5 m (5
ft). The model was constructed with a broad‐crested weir, and
the chute slope is 4(H):1(V). Step heights of 38 mm (1.5 in.)
were placed in the model. The downstream edge of the weir cor‐
responds to station 0.0 m (0.0 ft) and step 0. The downstream
toe of the spillway is at station 6.1 m (20ft) at step 40. Figure
2 illustrates the two‐dimensional model.
To minimize scale effects associated with the collection of
data in air‐entrained flow environments, the criterion sug‐
gested by Boes and Hager (2003a) was followed, and a scale
of 10:1 or larger is recommended. Reynolds and Weber num‐ (a)
bers for this generalized model study were reported by Hunt
and Kadavy (2008, 2010). Water surface elevations and ve‐
locity measurements were collected. A moveable carriage set
atop rails on the flume walls allowed manual point gauge
readings of the centerline water and bed surface elevations.
The carriage was also used to collect centerline velocity mea‐
surements at the crest. A second set of rails was attached to
the flume walls and set parallel to the sloping chute. These
rails were used for velocity measurements along the center‐
line of the chute. Figure 3a illustrates the use of the carriage
for recording flow measurements, and figure 3b illustrates
velocity measurement collection normal to the chute slope.
Unit discharges of 0.11, 0.20, 0.28, 0.42, 0.62, 0.82 m3 s-1
m (1.2, 2.2 3.0, 4.5, 6.7, 8.8 cfs ft-1) were tested. Cross‐
-1
sectional velocity profiles were collected with an acoustic
Doppler velocimeter (ADV) on the broad‐crested weir to de‐ (b)
velop a calibration curve of the flow versus upstream head.
Velocity profiles and flow depths were collected along the
Figure 3. (a) Data collection using a mobile carriage along the top of the
centerline of the spillway using three separate measuring de‐ test flume walls, and (b) data collection using secondary rails set parallel
vices: (1) an ADV, (2) a pitot tube (PT) coupled with a differ‐ to the sloping chute.
ential pressure transducer, and (3) a two‐tip fiber optical (FO)
probe and data acquisition system. Each device has a unique Velocity profiles were taken normal to the spillway crest
range of velocities that could be captured during testing. The surface along the centerline at horizontal stations -2.4, -1.8,
ADV was limited by a maximum velocity of 4.6 m s-1 (15 ft -1.2, -0.61, 0.0 m (-8, -6, -4, -2, and 0 ft) (fig. 2) with the
s-1), and it could not be used in highly turbulent flow. The PT ADV. Velocity profiles normal to the spillway chute slope
coupled with a pressure transducer has the advantage of col‐ along the centerline were taken at horizontal stations 0.0,
lecting higher velocities than the ADV. The PT also provides 0.61, 1.2, 1.8, 2.4, 3.0, 3.7, 4.3, 4.9, and 5.5 m (0, 2, 4, 6, 8,
verification for some of the velocity measurements collected 10, 12, 14, 16, and 18 ft) (fig. 2) with the ADV when the ve‐
with the ADV. Matos et al. (2002) found that the PT can be locities were within its limits, with the PT for all stations on
used in air‐entrained environments when the air concentra‐ the chute, and with the FO probe when air entrainment was
tions in the flow are less than 70%, thereby giving it another present. These horizontal stations correspond to the down‐
advantage over the ADV. To achieve more accurate velocity stream edge of the step. Two velocity profiles were taken with
measurement in air‐entrained flows, the PT was back‐flushed the PT and the FO probe along the stilling basin floor. These
before each measurement to remove air bubbles in the tube measurements were taken normal to the floor at horizontal
and the lines to the pressure transducer. The two‐tip FO probe stations 6.6 and 8.1 m (22 and 27 ft), respectively.
was used to measure velocities and void fraction (C) in air‐ PT velocity measurements were collected starting near the
entrained flows. The void fraction data collected by the FO step edge surface and incrementally moved up the depth pro‐
probe was used to determine the equivalent clear water depth file. The ending measurement was the highest point in the
that would result without the presence of the entrained air. depth profile at which the flow still appeared as a coherent

Vol. 53(1): 111-118 113

stream of water. This ending point normally corresponded to h = step height
a point void fraction of approximately 0.5. For the velocity q = unit discharge
profile analysis, velocities above this point were assumed to g = gravitational constant.
be equal to the velocity at the last measured depth. Hunt and Kadavy (2010) identify the downstream edge of
FO velocity and void fraction measurements were col‐ the broad‐crested weir as the origin point for length to the
lected starting at the upper surface of the depth profile where inception point (fig. 1). Table 1 shows F* for all tested flows
the void fraction was approximately 0.95 and incrementally and how closely the inception point (Li*) as calculated using
moved down the depth profile to the step edge surface or to equation 1 compares to the inception point (Li ) observed by
the depth at which the void fraction was so small that mea‐ Hunt and Kadavy (2008). These results are valid when F*
surements could not be collected. The FO void fraction data ranges from 10 to approximately 100 (Hunt and Kadavy,
was used for the FO and PT velocity profile analysis. 2009). Some of the differences between the observed and
predicted inception points are likely due to the subjectivity
of the visual observations. In particular, the observed and
RESULTS AND DISCUSSION predicted inception points for the unit discharge of 0.62 m3
s-1 m-1 (6.7 cfs ft-1) showed a 14% difference. The inception
An important element in stepped spillway chute training wall
point for this discharge was near the break in slope from the
design is whether to account for flow bulking. As the turbulent
spillway chute to the stilling basin, so this slope change may
boundary layer nears the free surface, the flow appearance
have influenced the observed inception point location. The
changes. The inception point is detected when the boundary lay‐
predicted inception point location for unit discharge of
er reaches the free surface. Figure 1 illustrates the inception
0.11m 3 s-1 m-1 (1.2 cfs ft-1) is 21% larger than the observed
point development. The growing boundary layer is generated by
inception point location, and this difference may be partially
the friction of the stepped surface. In the case of the stepped
attributed to F* equaling 10, the recommended minimum for
spillway, the growing boundary layer typically begins at the
the use of equation 1.
spillway crest. For a broad‐crested weir spillway, the growing
The results summarized in table 1 are valuable to design
boundary layer, as Hunt and Kadavy (2010) define it, begins at
engineers. The inception point location in relation to the
the downstream edge of the weir. Although a small amount of
stilling basin and the design tailwater provides the necessary
boundary layer development may occur upstream of this point
information for the design engineer to make an informed
on the smooth crest surface, it is considered insignificant
decision on whether flow bulking should be accounted for in
compared to the development that occurs downstream of this
the design of the stepped spillway training walls.
point on the stepped chute surface. Chanson (1994a, 2002) char‐
Additional information valuable to the design engineer is
acterizes the inception point as the location where the turbulent
the energy dissipation that occurs in the spillway chute.
boundary layer reaches the free surface. Upstream of the incep‐
Stepped spillways dissipate a significant amount of energy
tion point, the flow is described as non‐aerated, as shown in fig‐
when compared to traditional smooth chute spillways.
ure 1. Downstream of the inception point, an aerated flow
Consequently, the stilling basin size can be reduced for a
region is established, and this region becomes more fully devel‐
stepped spillway as compared to a stilling basin for a
oped as the flow moves further downstream.
traditional smooth chute spillway of similar size. To
As previously reported by Hunt and Kadavy (2008), flow
determine the energy loss in the spillway chute, velocity must
was observed to change from a glassy, smooth appearance to
be determined. Figures 4a through 4e present typical depth‐
a slight undulating appearance to a white water or frothy ap‐
velocity profiles measured with the PT and FO probe and the
pearance. The slight undulating appearance was noted during
depth‐void fraction (C) measured with the FO probe at
testing as the turbulent boundary layer approaching the free
stations 3.0 (10 ft), 3.7 (12 ft), 4.3 (14 ft), 4.9 (16 ft), and
surface. When the flow at the surface appeared frothy across
5.5m (18 ft) for a unit discharge of 0.28 m3 s-1 m-1 (3.0 cfs
the full width of the flume, the turbulent boundary layer was
ft-1). Figure 4a represents a velocity and void fraction (C)
thought to have reached the free surface, and that location
profile just upstream of the inception point at L/Li = 0.96,
was recorded as the inception point. Hunt and Kadavy (2007,
while figures 4b through 4e represent profiles downstream of
2008) found that an inception point relationship developed
the inception point at L/Li = 1.2, 1.4, 1.6, and 1.7,
by Chanson (1994b, 2002) can be used to determined the in‐
respectively. Figures 4a through 4e show similarities in the
ception point location in stepped spillway slopes as flat as
velocity data from the PT and the FO probe. In figures 4a and
4(H):1(V) even though the original relationship was devel‐
4b, the velocities near the spillway chute bed surface were
oped from steep (2(H):1(V) or steeper) stepped spillways.
undetectable by the FO probe because the air concentrations
This relationship was verified for Froude surface roughness
were significantly low or nonexistent. The PT provides a
(F*) ranging from 10 to 100 (Hunt and Kadavy, 2009). Chan‐
son's relationship, as indicated by equation 1, is based on the Table 1. Calculated and observed inception
spillway chute slope, unit discharge, gravitation accelera‐ point data for each of the flows.
tion, and step height (0.038 m in this study): Chanson (1994b) Observed

Li * = 9.719(sin θ)0.0796 (F* )0.713 h (cosθ) (1) q Li* Inception Li Inception

(m3 s‐1 m‐1) F* (m) Point (step) (m) Point (step)
where 0.82 75 7.0 Basin 7.1 Basin
Li  * = distance from the start of growth of boundary layer 0.62 57 5.7 36 6.6 Basin
to the inception point of air entrainment 0.42 38 4.3 28 4.6 29
 = channel slope 0.28 26 3.2 21 3.5 22
F* = Froude number defined in terms of the roughness 0.20 18 2.6 17 2.7 17
height (F* = q/[g(sin){h(cos)}3]0.5) 0.11 10 1.7 10 1.4 9

114 TRANSACTIONS OF THE ASABE

Figure 4. Depth‐velocity and depth‐void fraction (C) for a unit discharge of 0.28 m3 s-1 m-1 (3.0 cfs ft-1): (a) station 3.0 m (10 ft), (b) station 3.7 m (12ft),
(c) station 4.3 m (14 ft), (d) station 4.9 m (16 ft), and (e) station 5.5 m (18 ft).

record of the velocities in these lower portions of the profiles, s-1), respectively. Based on this observation, the air
giving it an advantage over the FO probe. In figures 4a concentration in the flow appears small enough that it does
through 4c, the velocity profiles measured with the PT and not affect the average velocity when velocities are measured
the FO probe yield similar results, indicating that the air with the PT. Additional observations were made regarding
concentration likely does not influence the PT results. the velocity profiles. For instance, in previously reported
Separation between the two profiles is noticeable in research, Hunt and Kadavy (2008) showed that the velocity
figures4d and 4e near the flow surface. The difference in the profiles transition from uniform at the crest to approaching
profile is likely attributed to the more fully developed air in a one‐sixth power law distribution at the inception point for
the profile, yet upon further examination, the average all tested flows. Boes and Hager (1998) and Chanson (2000)
velocities for the PT and FO probe data for figure 4d are also determined that the velocity profiles tended to follow a
3.70m s-1 (12.2 ft s-1) and 3.69 m s-1 (12.1 ft s-1), one‐sixth power law distribution. Figures 4a through 4e show
respectively. For figure 4e, the average velocities for the PT that the velocity profiles follow a one‐sixth power law
and FO data are 3.71 m s-1 (12.2 ft s-1) and 3.73 m s-1 (12.3ft distribution at and downstream of the inception point.

Vol. 53(1): 111-118 115

 Figures 4a through 4e also illustrate depth‐void fraction
profiles. At each station, the void fraction increases
dramatically from 0% at the bed surface to near 95% at the
water surface (figs. 4a through 4e). The void fraction data
were used to determine the equivalent clear water depth (ycw )
for each station. For air‐water flows, the ycw as defined by
Chanson et al. (2002) is:
y 90

ycw = ∫ (1 − C) * dy (2)
0

where y90 is characteristic depth (m) when the air

concentration is 90%, and C is air concentration, which is
also known as void fraction. Upstream of the inception point,
the measured depth is equal to the clear water depth because
there is very little if any air present. Downstream of the
inception point, air is visibly present. Therefore, the ycw
downstream of the inception point is the equivalent depth of
flow with no air present. Figure 5 illustrates the ycw ‐velocity
profile for station 5.5 m (18 ft). When figure 5 is compared
to figure 4e, the ycw is clearly less in figure 5 as compared to
the flow depth with air in figure 4e. The clear water depth
information is needed to determine the average air
concentration in the flow as well as the energy dissipation. As
defined by Boes and Hager (2003a), the average air
concentration, Cavg , is:
ycw
Cavg = 1 − (3)
y90 Figure 6. (a) Average air concentration (Cavg ) versus the normalized
length down the spillway chute (L/Li ) based on observed Li , and
The average air concentrations for each flow rate and (b)average air concentration (Cavg ) versus the normalized length down
station were plotted against the log of the normalized length the spillway chute (L/Li *) based on calculated Li *.
down the spillway chute (i.e., length from the downstream
crest edge to the location of interest parallel to the spillway The average velocity obtained from the velocity profiles
chute, L, normalized by the inception point location, Li ). was used to determine the energy loss on the spillway chute
Upstream of the inception point, the air concentration is 0%. downstream of the inception point. The average velocity was
By definition, the average air concentration is 0% at the based on ycw, as illustrated in a single example in figure 5.
inception point, L/Li and predicted L/Li *, equal to 1. Slightly With the average velocity at each station known, the total
downstream of the inception point, the air concentration energy loss to a given step relative to the step of interest can
increases to 10% (fig. 6). The air concentration continues to be determined:
increase gradually until it reaches a constant value as the flow ΔH = H o − H (4)
descends farther downstream of the inception point. where
Figures6a and 6b also demonstrate the similarity between
V2
the average air concentrations versus the normalized length H o = yo + o
down the spillway chute using both the observed and 2g
predicted inception point locations. V2
H = ycw cos θ + α cw
2g
Vcw = clear water mean velocity
Vo = approach velocity
yo = approach depth above datum
ycw = clear water flow depth
 = chute slope
g = gravitational acceleration
 = energy coefficient.
Figure 7 illustrates the energy loss parameters as they
relate to the stepped spillway. The datum line is at the
elevation of the step of interest. In many open‐channel
applications where the channel is of regular cross‐section
with fairly straight alignment, the energy coefficient () is
assumed to equal unity because the effect of non‐uniform
velocity distribution on the computed velocity head and
Figure 5. Clear water depth (ycw ) velocity profiles at station 5.5 m (18 ft)
for a unit discharge of 0.28 m3 s-1 m-1 (3.0 cfs ft-1). momentum is small (Chow, 1959). To determine the effect
that the non‐uniform velocity distribution has on the compu-

116 TRANSACTIONS OF THE ASABE

 Vcw
2 ycw
V /2g
o
h
H 
Ho yo
Vo l
L H
Datum

Figure 7. Energy loss parameters as they relate to the stepped spillway.

ted velocity head and momentum, equation 5 was used to predicted inception points. When the observed ratio L/Li and
determine the energy coefficient: computed L/Li  * is equal to 1, then the energy loss is
approximately 0.30 for all tested flows. Downstream of the
α=
∫ v dA ≈ ∑ v ΔA
3 3
(5) inception point, the energy loss increases to approximately
3 3 0.73 for values of L/Li and L/Li * of 3.5. The data in figure 9a
Vcw A Vcw A
were fit with a power relationship using two constraints:
where v is the point velocity for an incremental area ( A) in (1)at L/Li = 1.0,  H/Ho = 0.30 and (2) at L/Li = ∞,  H/Ho =
the velocity profile, Vcw is the mean velocity of the 1.0. The resulting equation (eq. 7) represents the data well
equivalent clear water depth‐velocity profile, and A is the and is illustrated in figure 9a. Based on these findings, the
total area of the profile. Assuming two‐dimensional flow that relative energy loss at any point downstream of the inception
is uniform across the width, the area (A) could be replaced by point may be approximated by the following relationship:
ycw. As shown in figure 8, the energy coefficient ranges from −0.87
1.01 to 1.07. This energy coefficient range is reasonable ΔH ⎛ L ⎞
= 1 − ⎢⎢ + 0.51⎟⎟ (7)
when compared to the values of 1.05 to 1.1 reported by Boes Ho ⎝ Li * ⎠
(1999) for fully air‐entrained flows.
With the energy coefficient data determined, equation 4 Differences between the fit of equation 7 and the data
was used to determine the energy loss in the stepped spillway. illustrated in figure 9b are attributed to a breakdown in the
Hunt and Kadavy (2008, 2010) reported that the energy loss predicted inception point relationship provided in equation 1.
follows a linear trend upstream of the inception point. The
energy loss relationship upstream of the inception point
provided by Hunt and Kadavy (2008, 2010) is:
ΔH L
= 0.3 * (6)
Ho Li*
Figure 9 presents the energy loss data downstream of the
inception point. Figure 9a illustrates the energy loss verses
the normalized length down the spillway chute observed
during the test (L/Li ). Figure 9b shows the relationship of the
relative energy loss versus the normalized length down the
spillway chute using the inception point as calculated by
equation 1 (L/Li *). These figures demonstrate the similarity
between the relative energy loss versus the normalized length
down the spillway chute using both the observed and

Figure 9. (a). Relative energy loss versus the normalized length down the
spillway chute based on observations, and (b) relative energy loss versus
Figure 8. Energy coefficient (a) versus the normalized length (L/Li ) down the normalized length down the spillway chute based on the predicted
the spillway chute. inception point location.

Vol. 53(1): 111-118 117

For a Froude surface roughness (F*) greater than 10, Boes, R. M., and W. H. Hager. 1998. Fiber‐optical experimentation
equation7 appears to fit the data well, but when F* is equal in two‐phase cascade flow. In Proc. Intl. RCC Dams Seminar. K.
to or less than 10, equation 7 begins to underestimate the Hanson, ed. Denver Colorado, Schabel Engineering.
energy loss. From this observation, it is concluded that Boes, R. M., and W. H. Hager. 2003a. Two‐phase flow characteristics
of stepped spillways. J. Hydraul. Eng. ASCE 129(9): 661‐670.
equation 7 should be limited for use with equation 1 for F*
Boes, R. M., and W. H. Hager. 2003b. Hydraulic design of stepped
greater than 10. This is a first attempt at providing an energy spillways. J. Hydraul. Eng. ASCE 129(9): 671‐679.
loss relationship downstream of the inception point. Chanson, H. 1994a. Hydraulics of skimming flows over stepped
Additional data from other tested step heights is expected to channels and spillways. IAHR J. Hydraul. Res. 32(3): 445‐460.
improve this relationship. Chanson, H. 1994b. Hydraulic Design of Stepped Cascades,
Channels, Weirs, and Spillways. Oxford, U.K.: Pergamon.
Chanson, H. 2000. Characteristics of skimming flow over stepped
spillways: Discussion. J. Hydraul. Eng. ASCE 125(4): 862‐865.
CONCLUSIONS Chanson, H. 2002. The Hydraulics of Stepped Chutes and
Many small earthen flood control dams are faced with Spillways. Steenwijk, The Netherlands: A. A. Balkema.
inadequate spillway capacities due to permanent pools filled Chanson, H., and L. Toombes. 2002. Energy dissipation and air
with sediment and sediment depositing in the flood pools. entrainment in a stepped storm waterway: An experimental
Additionally, many small flood control dams have study. J. Irrig. and Drainage Eng. ASCE 128(5): 305‐315.
experienced changes in hazard classification. Urbanization is Chanson, H., Y. Yasuda, and I. Ohtsu. 2002. Flow resistance in
the cause of most changes in the hazard classification of these skimming flows in stepped spillways and its modeling.
Canadian J. Civil Eng. 29(6): 809‐819.
structures. Many engineers are selecting RCC stepped
Chatila, J. G., and B. R. Jurdi. 2004. Stepped spillway as an energy
spillways as a rehabilitation design for these structures in dissipater. Canadian Water Resources J. 29(3): 147‐158.
order to comply with dam safety regulations. Chow, V. T. 1959. Open‐Channel Hydraulics. Boston, Mass.:
This model study provides details regarding the design of McGraw‐Hill.
a stepped spillway on a 4(H):1(V) slope. Based on Christodoulou, G. C. 1993. Energy dissipation on stepped
observations and data collected, a relationship developed by spillways. J. Hydraul. Eng. ASCE 119(5): 644‐655.
Chanson (1994a) may be used to determine the inception Felder, S., and H. Chanson. 2008. Turbulence and turbulent length
point on stepped spillway chutes as flat as 4(H):1(V) when F* and time scales in skimming flows on a stepped spillway:
ranges between 10 and 100. Velocity profiles recorded with Dynamic similarity, physical modeling, and scale effects.
a PT and FO probe downstream of the inception point Queensland, Australia: University of Queensland, Division of
Civil Engineering.
resulted in similar average velocities. Therefore, accurate
Hunt, S. L., and K. C. Kadavy. 2007. Renwick dam RCC stepped
velocities can be measured using a PT in air‐entrained flows. spillway research. In Proc. ASDSO Annual Meeting, CD‐ROM.
As reported by Hunt and Kadavy (2008), velocity profiles Lexington, Ky.: Association of State Dam Safety Officials.
transition from uniform at the crest to a one‐sixth power law Hunt, S. L., and K. C. Kadavy. 2008. Energy dissipation on a
distribution near the inception point. Velocity profiles 4(V):1(H) stepped spillway. In Proc. ASDSO Annual Meeting,
continue to follow a one‐sixth power law distribution CD‐ROM. Lexington, Ky.: Association of State Dam Safety
downstream of the inception point. Air concentrations near Officials.
the inception point are approximately 0% for all flows and Hunt, S. L., and K. C. Kadavy. 2009. Inception point relationship
rapidly increase to 10% slightly downstream of the inception for flat‐sloped stepped spillways. ASABE Paper No. 096571. St.
point. After this point, the air concentrations gradually Joseph, Mich.: ASABE.
Hunt, S. L., and K. C. Kadavy. 2010. Energy dissipation on
increase to a constant value as the flow descends the chute.
flat‐sloped stepped spillways: Part 1. Upstream of the inception
Additionally, a first‐generation relationship for determining point. Trans. ASABE 53(1): 103-109.
the energy loss at any point downstream of the inception point Hunt, S. L., K. C. Kadavy, S. R. Abt, and D. M. Temple. 2006.
was provided. Energy losses increased from 30% for L/Li and Converging RCC stepped spillways. In Proc. 2006 World
L/Li* equal to 1 to 73% for L/Li and L/Li * equal to 3.5. Environ. and Water Resources Congress, ASCE Conf.,
Additional data have been collected from this model to CD‐ROM. Reston, Va.: ASCE.
compare the effects of different step heights on the inception Matos, J., K. H. Frizell, S. André, and K. W. Frizell. 2002. On
point, velocities, air concentrations, and energy dissipation; performance of velocity measurement techniques in air‐water
these data are currently under analysis. This research will flows. In Proc. Hydraulic Measurements and Experimental
assist engineers with the design of stepped spillways and their Methods 2002, CD‐ROM. T. L. Wahl, C. A. Pugh, K. A. Oberg,
and T. B. Vermeyen, eds. Reston, Va.: ASCE.
associated stilling basins when applied on relatively flat
Peyras, L., P. Royet, and G. Degoutte. 1992. Flow and energy
embankment dams. dissipation over stepped gabion weirs. J. Hydraul. Eng. ASCE
118(5): 707‐717.
Rice, C. E., and K. C. Kadavy. 1996. Model study of a roller
REFERENCES compacted concrete stepped spillway. J. Hydraul. Eng. ASCE
Barani, G. A., M. B. Rahnama, and N. Sohrabipoor. 2005. 122(6): 92‐297.
Investigation of flow energy dissipation over different stepped Takahashi, M., C. A. Gonzalez, and H. Chanson. 2006.
spillways. American J. Applied Sci. 2(6): 1101‐1105. Self‐aeration and turbulence in a stepped channel: Influence of
Boes, R. M. 1999. Physical model study on two‐phase cascade cavity surface roughness. Intl. J. Multiphase Flow 32(12):
flow. In Proc. 28th IAHR Congress, Session S1. International 1370‐1385.
Association for Hydro‐Environment Engineering and Research. Yasuda, Y., and I. Ohtsu. 1999. Flow resistance of skimming flow in
Boes, R. M. 2000. Scale effects in modelling two‐phase stepped stepped channels. In Proc. 28th IAHR Congress, Session B14.
spillway flow. In Proc. Intl. Workshop on Hydraulics of Stepped International Association for Hydro‐Environment Engineering
Spillways, 53‐60. H. E. Minor and W. H. Hager, eds. Steenwijk, and Research.
The Netherlands: A. A. Balkema.

118 TRANSACTIONS OF THE ASABE