STORMWATER CONVEYANCE

MODELING AND DESIGN

Authors
Haestad Methods
S. Rocky Durrans

Managing Editor
Kristen Dietrich

Contributing Authors
Muneef Ahmad, Thomas E. Barnard,
Peder Hjorth, and Robert Pitt

Peer Review Board
Roger T. Kilgore (Kilgore Consulting)
G. V. Loganathan (Virginia Tech)
Michael Meadows (University of South Carolina)
Shane Parson (Anderson & Associates)
David Wall (University of New Haven)

Editors
David Klotz, Adam Strafaci, and Colleen Totz
HAESTAD PRESS
Waterbury, CT USA





Click here to visit the Bentley Institute
Press Web page for more information

C H A P T E R
9
Culvert Design
A culvert is a hydraulically short underground water conveyance conduit that allows
water to flow through an obstruction such as a highway or railway embankment (see
the preceding page). The orientation of a culverts longitudinal axis is usually at a
right angle, or nearly so, to the centerline of the obstruction under which it passes. In
some instances, structures constructed as culverts may be considered bridges if their
spans exceed a certain dimension. Transportation departments in the United States
usually consider any culvert with a total span (counting all culvert barrels, if there are
more than one) greater than 20 ft (6 m) to be a bridge. It may sometimes be appropri-
ate to hydraulically model a structure classified as a bridge by using culvert analysis
methods; however, the techniques used to model bridge hydraulics generally differ
from those used with culverts and are beyond the scope of this book. [The reader is
referred to Floodplain Modeling Using HEC-RAS (Dyhouse et al., 2003) for informa-
tion on bridge modeling.]
Typically, a culvert is designed to convey the peak flow from a specified design storm,
and its performance is checked for a storm having a longer recurrence interval. The
purpose of this check is to assess the magnitude of roadway or embankment overtop-
ping expected to occur for the larger event. A local or regional review agency gener-
ally specifies the recurrence interval for a design. Chapters 4 and 5 present hydrologic
procedures for calculating discharges for both the design event and check event, and
Section 9.1 of this chapter presents a discussion of design flow estimation as it relates
specifically to culverts.
As stated in Section 2.2 (page 22), culverts are available in various shapes, sizes, and
materials. Materials and shapes are described in detail in Section 9.2. Section 9.3 pro-
vides a discussion of various types of culvert end treatments and their hydraulic per-
formance characteristics and describes special inlet configurations that can be used
when allowable headwater depths are limited.
The horizontal and vertical alignments of a culvert should closely follow those of the
channel or stream. This typical alignment is sometimes modified for reasons relating
to hydraulic performance, cost, or other constraints. Possible modifications include
the use of a doglegged alignment or the use of a drop at the culvert entrance to reduce
324 Culvert Design Chapter 9
the barrel slope. Issues associated with such modifications are presented in Section
9.4.
Section 9.5 presents the detailed hydraulic performance characteristics of a culvert.
The discussion includes the controls that affect hydraulic performance, the potential
effects of tailwater elevation, and detailed computational procedures for flow profile
estimation. Also discussed are methods for estimating outlet velocity, which is neces-
sary for design of energy dissipation measures.
During storm events having recurrence intervals longer than that of the culvert design
storm, or in cases where the culvert becomes clogged with debris, overtopping of the
roadway or embankment is expected to occur. Prediction of the weir flow characteris-
tics of roadway overtopping is presented in Section 9.6. Section 9.7 describes the
development of culvert performance curves, which provide a convenient way of
assessing the headwater elevation at a culvert for any discharge of interest.
Because the flow exiting the downstream end of a culvert often has a velocity higher
than what can be tolerated in the downstream channel, or because of the acceleration
of flow at the entrance, erosion protection measures are generally needed at the outlet
and/or inlet of a culvert. This protection may consist of a concrete apron, protective
riprap, or a geosynthetic material for reinforcement of soil and vegetation. Procedures
for riprap erosion protection design at storm sewer outlets are presented in Section
11.7, and these procedures can be applied to culverts as well. Section 9.8 discusses
outlet protection considerations specific to culvert design.
Because culverts often convey water in natural streams that may support fish and
other aquatic wildlife, minimizing the disruption of migration pathways may be a
design consideration. A discussion of designing for some of these environmental fac-
tors is provided in Section 9.9.
Section 9.1 Culvert Design Flows 325
9.1 CULVERT DESIGN FLOWS
In most cases, culverts are designed to convey a discharge corresponding to a chosen
recurrence interval. In such instances, one must estimate only the peak runoff rate.
However, for cases in which a culvert causes significant ponding upstream of its
entrance, the reduction of the peak discharge caused by detention effects may need to
be considered. In such cases, the engineer should develop and route a complete runoff
hydrograph. Chapter 5 outlines procedures for peak flow and hydrograph estimation.
Section 12.7 discusses flow routing procedures.
A government review agency usually specifies the recurrence interval for which a
design discharge or hydrograph should be determined. This criterion and others are
provided in the interest of promoting consistency from one design to another. In high-
way design applications, culvert design flows are often estimated for a 50-year event
based on FHWA design frequency recommendations (Brown, Stein, and Warner,
2001). Lower or higher recurrence intervals may be justified in other applications.
In addition to developing the design discharge, the engineer should estimate a check
flow. The recurrence interval associated with the check flow is longer (and thus the
discharge greater) than that of the design flow, because the check flow is used to
assess performance when embankment overtopping is expected to occur. Again, local
or state review agencies usually specify the check flow recurrence interval that should
be used; in highway design applications, it is often the 100-year event.
For cases in which a performance curve is needed for a culvert design, a range of
discharges corresponding to various recurrence intervals should be estimated. The
low end of the discharge range should be lower than the design discharge. The high
end of the range should be equal to or greater than the check flow discharge. In
performance curve development (Section 9.7), the headwater elevation is estimated
and graphed for each of these estimated discharges.
An alternative to basing design discharge on recurrence interval criteria is to base it
on an economic analysis. Figure 9.1 shows the costs associated with a circular culvert
as a function of its diameter. All costs are expressed as a present worth using appro-
priate discounting formulas. The capitalized cost curve represents the costs of design,
construction, and maintenance of the culvert over its lifetime. Culverts with large
diameters are clearly more costly than culverts with small diameters, and thus the cap-
italized cost increases with the diameter. The damage curve represents the present
value of flood damages anticipated to occur when the culverts capacity is exceeded.
Because the capacity of a smaller culvert will be exceeded more frequently, the result-
ing damages will be higher than those expected for a larger culvert. Thus, the damage
curve has a negative slope. The total cost associated with a culvert is the sum of the
capitalized cost and the damages. It is shown as a third curve in Figure 9.1. The diam-
eter corresponding to the minimum cost on the total cost curve represents the least-
cost culvert design.
326 Culvert Design Chapter 9
Figure 9.1
Graph of culvert costs
vs. diameter
An economic analysis such as that just described is usually not accomplished in
practice because of difficulties associated with estimating the damage curve. The
analysis for an area can be further complicated by the presence of multiple culverts
providing differing levels of protection. Nevertheless, performing an economic analy-
sis is an enlightening exercise, as the true total cost of a culvert becomes evident.
Such an analysis also shows that most culvert designs are likely less than optimal in
economic terms.
9.2 CULVERT TYPES AND MATERIALS
Culverts, like storm sewers, come in a variety of shapes, sizes, and materials. Cross-
sectional shapes associated with pipe culverts are circular, elliptical (horizontal or
vertical), and arched. Culverts may also be constructed from precast or cast-in-place
rectangular box sections, or from field-assembled structural-plate-arch materials.
These latter shapes are common for large-capacity culverts. Materials used for culvert
design are most commonly reinforced precast concrete, cast-in-place concrete, and
corrugated steel. Section 11.1 discusses available materials and sizes. Information
that is more detailed can be obtained from pipe manufacturer associations such as the
American Concrete Pipe Association (ACPA) and the American Iron and Steel Insti-
tute (AISI). Other materials used for culverts include corrugated aluminum, polyeth-
ylene, and polyvinyl chloride (PVC). Aluminum and plastic materials offer
advantages over steel in terms of corrosion, but are more limited in size availability.
They also do not offer the structural strength of concrete or steel.
In many cases, to address a situation where there is limited cover over the top of a cul-
vert, or to approximate the full width of a stream channel, multiple culvert barrels
may be placed side by side. Such a configuration can minimize flow contractions at
the culvert entrance and make the installation more hydraulically efficient. For
instances in which multiple culvert barrels are used, it is usually assumed during
Section 9.3 Culvert End Treatments 327
design that the total discharge is divided equally among the individual barrels. Excep-
tions to this practice are cases in which the individual barrels have different sizes, or
where they have different invert elevations or slopes. In these latter cases, a trial-and-
error procedure is required to determine the discharge through each barrel. Computer-
based modeling tools are particularly useful in such instances.
In the absence of review agency criteria, which would otherwise govern, the mini-
mum recommended diameter for a culvert is 12 in. (300 mm). A higher minimum
may be desirable for relatively long culverts, which are more difficult to maintain than
short ones. Culverts are susceptible to clogging, especially in areas where sediment
transport is high and where debris such as fallen leaves and branches is significant.
When non-circular materials are used, it is recommended that the shortest cross-
sectional dimension be at least 12 in. (300 mm).
9.3 CULVERT END TREATMENTS
The entrance and exit of a culvert are locations where flow is accelerated or deceler-
ated, often causing erosion of the channel bed and embankment side slopes. Erosion
protection measures are generally required in these locations and usually take the
form of a riprap blanket and/or special culvert end treatments. The section focuses on
the various types of end treatments that are used; Section 9.8 and Section 11.7 discuss
riprap protective measures.
Culvert end treatments do more than provide protection against erosion; they help to
stabilize embankment slopes around culvert entrances and exits and blend the culvert
ends into the embankment slope. The type of culvert end treatment used in a design
also has hydraulic ramifications, as some types may actually enhance the hydraulic
conveyance capacity of the structure. For instance, all other factors being equal (cul-
vert dimensions, material, discharge, tailwater elevation, etc.), some culvert entrance
types will result in lower upstream water surface elevations (that is, lower headwater
depths); these entrance types are considered to be more hydraulically efficient.
Figure 9.2 illustrates the most commonly used types of culvert end treatments. Each
of these types can be used at either end (entrance or exit) of a culvert. A plain or
projecting culvert [Figure 9.2(a)] has no special treatment. The culvert barrel simply
projects from the embankment slope. This type of end treatment offers no protection
against erosion of the embankment slope; its primary advantage is that it is
inexpensive. Hydraulically, a projecting end is neither efficient nor inefficient if the
wall thickness of the culvert pipe is significant (as in the case of concrete pipe); it may
be viewed as a middle-of-the-road treatment. If corrugated metal pipe (which has a
small wall thickness) is used, a projecting end is hydraulically inefficient.
A mitered end [Figure 9.2(b)] provides a visual improvement over the projecting end
and usually incorporates an erosion protection blanket or apron around the mitered
end. However, from a hydraulic capacity point of view, a mitered-end configuration is
inefficient. The culvert entrance loss coefficient associated with a mitered end is often
higher than that of other types of end treatments (see Table6.5 on page203).
328 Culvert Design Chapter 9
Figure 9.2
Typical culvert end
treatments
Concrete and corrugated metal pipe manufacturers can supply, in addition to pipe
materials, specially formed end sections such as the precast end section shown in Fig-
ure 9.2(c). These end sections are usually tapered to gradually accelerate or decelerate
the flow, and are shaped to facilitate the blending of the culvert end into the embank-
ment slope. Entrance loss coefficients shown in Table6.5 indicate that special end
sections have approximately the same efficiency as a square-edged headwall or
square-edged projecting end.
Headwalls and wingwalls [see Figure 9.2(d)] offer perhaps the best slope stability and
erosion protection, but are generally the most expensive of treatment types of those
shown in Figure 9.2. A headwall is a concrete wall at the inlet or outlet of a culvert
through which the pipe projects. A headwall located at the outlet end of a culvert is
sometimes referred to as an endwall. Wingwalls may be attached to either side of a
headwall to provide additional stability or improve hydraulic performance. Headwalls
and wingwalls, which may be precast or cast-in-place, are essentially retaining walls
and should be designed as such by considering factors such as earth pressure, loading,
and soil properties. The engineer should consider requiring a pipe joint near the head-
wall to allow for possible differential settlement or movement between the headwall
Section 9.3 Culvert End Treatments 329
and pipe. Headwalls and wingwalls should be provided with weep holes to allow
seepage water trapped behind them to escape.
Any type of end treatment illustrated in Figure 9.2 can be enhanced hydraulically by
rounding or mitering the edges of the entrance to the culvert barrel. Placing the socket
or groove end of a pipe joint at the entrance has essentially the same effect as round-
ing or mitering. The hydraulic enhancement derives from a reduction in flow contrac-
tion effects as the water enters the culvert. A square-edged entrance causes a large
contraction of the flow streamlines, resulting in an effective reduction of cross-
sectional flow area. Rounding or mitering the entrance edges as shown in Figure 9.3
reduces the contraction. Table6.5 shows that loss coefficients for rounded, mitered, or
socket-end edges are always less than those for square-edged entrances.
Figure 9.3
Square-edge and
curved-edge culvert
entrances (Norman,
Houghtalen, and
J ohnston, 2001)
At locations where available headwater depths are too limited to allow adequate cul-
vert performance, special culvert entrance configurations may be used to increase the
hydraulic efficiency of the culvert inlet. Side-tapered or slope-tapered inlets, as illus-
trated in Figure 9.4, may be employed to fulfill this need. (Note that the slope-tapered
inlet in this figure also has a side taper.) These special inlet configurations are more
costly to construct but can be advantageous in locations where other alternatives fail
to provide the required performance. Harrison et al. (1972) provide further informa-
tion on the hydraulic performance for these types of culverts.
Regardless of the type of end treatment used for a culvert, it is recommended that a
concretecradle or cutoff wall (see Figure 9.5) be installed around and under the ends
of the pipe. The cradle is especially important at the culvert outlet to prevent under-
cutting of the pipe and subsequent collapse. Of course, if a headwall end treatment is
used, the headwall itself performs this function. If a concrete apron is used at the
entrance or exit of a culvert, a cutoff wall should be integrated into it to prevent under-
mining.
330 Culvert Design Chapter 9
Figure 9.4
Side-tapered and
slope-tapered culvert
entrances
Credit: FHWA, 1985
Figure 9.5
Concrete cradle at a
pipe outlet
Credit: FHWA, 1985
It is frequently desirable to provide trash racks (sometimes called safety racks)
secured to the wingwalls at culvert entrances and/or exits to prevent unauthorized
entry and/or clogging by debris. Trash racks should have clear openings between bars
no greater than about 6 in. (150 mm) and should be constructed in modular sections to
facilitate removal for maintenance of the culvert. The rack itself should receive peri-
odic maintenance to clear trapped debris. Installing the rack so that the bars are
Section 9.4 Alignments of Culverts 331
inclined from the vertical can help keep the rack clear because the flowing water will
push floating debris up the slope of the rack. Inclining the rack also makes it easier for
a person trapped against the rack to climb out.
Although guidelines and recommendations vary, a trash rack should be structurally
designed to withstand the hydraulic loads that would occur under partial or full clog-
ging of the rack. Hydraulically, a trash rack causes a head loss to occur, but the loss
essentially can be neglected if the total clear opening area of the rack is much greater
(about four times) than that of the cross-sectional area of the culvert barrel. Racks
should be placed far enough upstream of the culvert entrance to ensure that the veloc-
ity of the accelerating flow is low enough for a person to escape.
9.4 ALIGNMENTS OF CULVERTS
Generally, the horizontal and vertical alignments of a culvert should conform as
closely as possible to those of the channel. However, it may be necessary or desirable
in some cases to depart from this ideal. Discussions of alignment problems and solu-
tions are presented in the subsections that follow.
Horizontal Alignment
If a stream channel is very sinuous, it can be difficult or impossible to develop a cul-
vert alignment that provides a smooth transition of the flow at both the upstream and
downstream ends of a culvert. In such instances, it may be desirable to realign the
channel to provide for smooth transitions. However, such efforts should be undertaken
with great care because the newly aligned channel may be unstable, resulting in sedi-
mentation and/or erosion problems upstream or downstream of the culvert. Brice
(1981) concluded that channel realignments are generally successful, except in
instances where the channel was already unstable.
An alternative to channel realignment is to use a curved or doglegged alignment for
the culvert barrel (see Figure 9.6). Of course, such a configuration will result in addi-
tional energy losses due to the curve or bend. Another disadvantage is that debris
tends to build up and cause culvert blockages at bend locations. If a culvert operates
under inlet control [see Section 2.2 (page 24) and Section 9.5 (page 335)], additional
head losses caused by a curve or bend will not affect the headwater elevation. The
headwater elevation will be affected if the culvert operates under outlet control.
Changes in the horizontal alignment of a culvert should ideally be accomplished
using a curved alignment. In cast-in-place concrete box culvert construction, curved
concrete forms may be used. In pipe culvert construction, small deflections at individ-
ual pipe joints according to the manufacturers recommendations can be accom-
plished to approximate a smooth curve. If a sudden bend is used instead of a curved
alignment, the bend deflection angle should be limited to no more than approximately
15 degrees, and such bends should be separated from one another by a distance of at
least 50 ft (15 m) (ASCE, 1993).
332 Culvert Design Chapter 9
Figure 9.6
Curved and doglegged
alignments
Vertical Alignment
In a sag culvert, one or more of the straight barrel segments has an adverse slope, as
illustrated in Figure 9.7(a), which usually results in part of the culvert flowing full and
under pressure. A sag culvert is especially prone to clogging by trapped sediments
and debris, but it can be the best option if its use prevents, for instance, a costly utility
relocation. Sag culverts are also used frequently at roadway crossings of irrigation
canals.
Broken-back culverts [Figure 9.7(b)] have bends in their vertical alignments. The
bend in a broken-back culvert may cause the upstream part of the culvert to have a
smaller slope than the downstream part, or vice versa. A broken-back culvert may be
used to minimize excavation, avoid a rock outcropping, or reduce the exit velocity
from the culvert.
Section 9.4 Alignments of Culverts 333
Figure 9.7
Sag and broken-back
culverts (Norman,
Houghtalen, and
J ohnston, 2001)
If a broken-back culvert has a steeper slope downstream of the bend, the possibility
exists that the hydraulic grade will drop below the crown of the culvert in the vicinity
of the bend although it is otherwise above the crown. This phenomenon causes sub-
atmospheric pressure to develop within the culvert at the bend location and can poten-
tially lead to structural and hydraulic problems. If this possibility exists, a vent should
be provided to allow atmospheric pressure to exist at the bend location.
Hydraulic analyses of broken-back and sag culverts are complicated by the fact that
they usually involve regions of both full (pipe) flow and open-channel flow. Further,
open-channel flow may be subcritical in one part of the culvert and supercritical in
another. Thus, the potential exists for many flow controls, and each situation must be
treated differently. Computer-based modeling tools can quickly evaluate flow regimes
for complicated culvert geometries, and hence can be used to optimize designs.
Skewed Culverts
Positioning a culvert alignment to approximate the flow path of the stream channel
often involves placing the culvert at an angle that is not perpendicular to the embank-
ment centerline, as shown in Figure 9.8, parts (b) and (c). This culvert alignment is
referred to as skewed in relation to the normal embankment alignment. In the case of
skewed culverts, headwalls may be skewed with respect to the barrel alignment, as
334 Culvert Design Chapter 9
illustrated in Figure 9.8(c), to better align with embankment slopes (that is, the culvert
is skewed with respect to the embankment, and the headwalls are skewed by an offset-
ting amount with respect to the culvert so that they will still align with the embank-
ment slopes). If the headwall on a culvert is skewed with respect to the barrel, an
additional entrance loss occurs; however, this loss is usually small and may be
neglected in many cases. This additional loss can be estimated by using the design
charts published in FHWAs Hydraulic Design Series Number 5, Hydraulic Design of
Highway Culverts (Normann, Houghtalen, and J ohnston, 2001), also known as HDS-
5.
Figure 9.8
Skewed culvert
alignments (Normann,
Houghtalen, and
J ohnston, 2001)
9.5 CULVERT HYDRAULICS
The hydraulics of a culvert are generally complex because of the short conduit length
and the several flow control types that may exist. A culvert may flow less than full
over part or all of its length, or it may flow full (closed-conduit flow) over its entire
length. Culverts are frequently not long enough for uniform flow to be achieved and
maintained; therefore, the entire flow profile often consists of gradually and/or rap-
idly varied flow. The flow control may also change with changes in discharge (that is,
the culvert may operate under one hydraulic control for some discharges, and under
one or more other controls for other discharges).
In design, the head loss (which approximately equals the difference between the water
surface elevations upstream and downstream of the culvert) that can be tolerated in a
culvert is often limited. The most widely used design approach is to evaluate the per-
formance of a trial culvert design for several different flow control types. The worst-
performing control condition (the condition resulting in the greatest head loss) is then
used to evaluate the acceptability of the proposed design. This method is a conserva-
tive, worst-case design approach that ignores conditions that may result in better per-
formance. The benefits of this approach are its ease of use and the assurance that the
culvert will perform adequately under the most adverse conditions.
Section 9.5 Culvert Hydraulics 335
Two general methods can be used to predict the hydraulic performance characteristics
of a culvert, and both are presented in this section. The graphical method involves the
use of nomographs that relate the headwater elevation at the culvert entrance to fac-
tors such as the length, size, and roughness of the barrel and the tailwater elevation
downstream of the outlet. This method is relatively easy to use and provides results
that are adequate for most purposes. It should not be used for culverts containing hor-
izontal curves or bends unless the additional losses are accounted for separately. It
should almost never be used for a broken-back or sag culvert.
The second method for predicting culvert performance is to compute the flow profile
using the gradually varied flow techniques presented in Section 7.8. This method is
much more laborious than the first but can yield much more accurate predictions of
culvert performance in many cases. This method is usually required for broken-back
and sag culverts. Because of the number of calculations involved, computer programs
such as CulvertMaster are especially beneficial when using this method.
Flow Controls
Research by the National Bureau of Standards (NBS) and the FHWA in the United
States divides culvert flow controls for a straight, uniformly shaped culvert into two
basic classes depending on the location of the control section: inlet control and outlet
control (Normann, Houghtalen, and J ohnston, 2001). With inlet control, the culverts
entrance characterisics determine its capacity. Conversely, with outlet control, the
inlet can accept more flow than the culvert can carry, because of either the friction
losses in the barrel or the tailwater elevation. Descriptions of the various flow regimes
encountered under both control types follow. Additional flow controls may exist in
broken-back or sag culverts, and in those with improved inlet configurations.
Inlet Control. For a culvert operating under inlet control, the culvert barrel is
capable of conveying a greater discharge than the inlet will accept. The flow control
section is just inside the culvert barrel at its entrance, and the flow profile passes
through critical depth at this location. The culvert is hydraulically steep, and the flow
regime downstream of the control section is supercritical with an S2 profile [see Sec-
tion 7.8 (page 274) and Figure 7.24]. Conditions downstream of the entrance have no
effect on the culvert capacity.
Figure 9.9 illustrates four cases in which a culvert operates under inlet control. Note
the drawdown of the S2 water surface profile near the upstream end of the barrel in
each case.
Case A in Figure 9.9 represents a situation in which neither the inlet nor the outlet of
the culvert are submerged by the headwater or tailwater. The flow approaches normal
depth as it moves along the culvert, and, depending on the flow depth in the down-
stream channel, a hydraulic jump may form downstream of the culvert outlet.
Case B illustrates a situation in which the outlet is submerged but the inlet is not sub-
merged. The culvert is hydraulically steep. An S2 water surface profile forms inside
the culvert entrance, and flow approaches normal depth. Submergence of the down-
stream end of the culvert results in pressure flow near the outlet end of the barrel. A
hydraulic jump forms inside the culvert barrel and provides a transition between the
supercritical and pressure flow regimes.
336 Culvert Design Chapter 9
Figure 9.9
Inlet control flow
conditions (Norman,
Houghtalen, and
J ohnston, 2001)
In Case C, the inlet is submerged by the headwater, but the outlet is not submerged.
As in Case A, the flow approaches normal depth as it moves along the culvert. A
hydraulic jump may form in the downstream channel.
Case D shows that a culvert may flow partly full even for cases in which both the inlet
and the outlet are submerged. The hydraulically steep nature of the culvert promotes
the S2 profile near the entrance end of the barrel. A hydraulic jump forms a transition
between the upstream supercritical flow and the downstream pressure flow. A vent,
Section 9.5 Culvert Hydraulics 337
such as a median drain, should be provided in a culvert anticipated to operate in this
mode to permit atmospheric pressure to exist within the culvert.
Other than the discharge, factors affecting headwater depth for a culvert operating
under inlet control include the cross-sectional area and shape of the barrel and the
inlets edge configuration. The barrel slope also influences the headwater depth, but
its effect is small. A sharp inlet edge, such as that occurring on a thin-walled metal
pipe with a projecting end, causes a large flow contraction at the culvert entrance. A
square-edged inlet, or better yet, an inlet with a rounded or pipe socket edge, would
reduce the headwater elevation in this case.
For a given allowable headwater elevation, the capacity of a culvert can be enhanced
by depressing the invert of the culvert barrel below the channel bed (with the fill
retained by wingwalls), thereby increasing the headwater depth. The height of the
depression T, as shown in Figure 9.10, is called the fall.
Figure 9.10 shows that most of the fall T occurs over the front portion of the culvert
approach apron, and this portion of the apron should have a minimum horizontal
length of 2T. The remainder of the fall occurs over the second portion of the approach
apron, which consists of a projection of the culvert slope for a distance equal to half
the height D of the culvert.
Figure 9.10
Fall at culvert
entrance (Norman,
Houghtalen, and
J ohnston, 2001)
If not done properly, provision of a depressed inlet can lead to headcutting and degra-
dation of the upstream channel. Grade control structures in the form of buried riprap
or concrete walls placed at various intervals along the channel bottom may become
necessary to provide a measure of protection against upstream headcutting and degra-
338 Culvert Design Chapter 9
dation. Also, the approach apron should be paved. When a depressed inlet is used,
sufficient elevation change must exist to ensure a positive slope between the upstream
and downstream ends of the culvert.
Outlet Control. Outlet control occurs when the slope of a culvert is hydraulically
mild or when the tailwater elevation is high enough to affect the headwater elevation.
In such cases, the inlet is capable of accepting more flow than the barrel and/or down-
stream flow conditions will permit. The control section is at or near the outlet point,
and the flow condition in the barrel near the outlet is critical, subcritical, or full (pres-
sure) flow.
For a culvert operating under outlet control, many factors affect the headwater depth.
These factors include discharge; barrel cross-sectional area, shape, length, roughness,
and slope; inlet edge configuration; and tailwater depth. Figure 9.11 illustrates seven
cases in which outlet control may occur.
Case A, shown in Figure 9.11, represents a situation in which the culvert barrel flows
full and under pressure over its entire length. Both the inlet and the outlet are sub-
merged. This condition is often assumed to exist in outlet control calculations, but sel-
dom occurs in actual culvert performance.
Case B is a situation in which the outlet is submerged but the headwater elevation is
only slightly higher than the crown of the inlet. In this instance, because of the low
submergence of the inlet and the flow contraction that occurs at that location, the
crown of the inlet is exposed where the water enters the barrel.
Case C represents a rare situation in which the inlet is submerged, the outlet is not
submerged, and the barrel flows full over its entire length. A large headwater depth is
required to cause a culvert to flow full over its entire length when the outlet is not sub-
merged. The exit velocity in this case is quite high and can cause severe erosion in the
downstream channel.
Case D occurs more commonly than does Case C. Like Case C, it represents a situa-
tion in which the inlet is submerged but the outlet is not. If the tailwater elevation is
below critical depth, the flow passes through critical depth at the outlet end of the cul-
vert barrel. If the tailwater is between critical depth and normal depth for the barrel,
profile calculations begin at the tailwater depth. Because of the hydraulically mild
slope of the barrel, an M2 water surface profile forms upstream of the critical depth
(control) section. The M2 profile intersects the crown at some distance from the out-
let, and upstream of that point the culvert flows full and under pressure.
Case E is similar to Case D. If the tailwater is below critical depth, flow passes
through critical depth at the outlet and profile calculations begin at critical depth. If
the tailwater depth is between normal depth and critical depth, profile calculating
begin at the tailwater elevation. An M2 water surface profile occurs upstream of that
point. The normal depth is smaller than the height of the barrel, and the flow profile
approaches it asymptotically. Flow in the culvert remains less than full throughout its
length.
Section 9.5 Culvert Hydraulics 339
Figure 9.11
Outlet control flow
conditions (adapted
from Norman,
Houghtalen, and
J ohnston, 2001)
340 Culvert Design Chapter 9
Cases F and G can occur when the tailwater at the outlet is above normal depth for a
culvert operating with subcritical flow conditions. Depending on the culvert entrance
characteristics, the resulting backwater profile may take precedence over inlet control
conditions.
Graphical Analysis of Culvert Performance
A traditional approach to culvert performance analysis is graphical in nature and
makes use of nomographs published in HDS-5 (Normann, Houghtalen, and J ohnston,
2001). The nomographs are published for both inlet and outlet control conditions, for
different barrel shapes and materials, and for different entrance edge configurations.
Because of the large number of possible combinations of these influential factors,
many nomographs are needed.
The following subsections describe and illustrate the use of the charts for both inlet
and outlet control conditions. When designing a culvert, the headwater depths com-
puted using inlet and outlet control assumptions are compared, and the larger value is
used to determine whether the design is acceptable.
Inlet Control Charts. Figure 9.12 is an inlet control nomograph for concrete
pipe culverts with circular cross sections. It is only one of the many inlet control
charts available for different combinations of pipe material, shape, and entrance edge
configuration. This particular chart illustrates the relationship that exists between the
culvert diameter, discharge, and headwater depth for a circular concrete culvert with
three possible entrance types: (1) square edge of pipe with headwall; (2) groove end
of pipe with headwall; and (3) groove end of pipe projecting from fill.
For a specified discharge, culvert diameter, and entrance edge configuration, the
nomograph in Figure 9.12 can be used to estimate HW/D, which is a dimensionless
ratio of the required headwater depth to the diameter. As illustrated with the example
given in the figure, a straight line can be drawn from a point on the pipe diameter
scale, through the appropriate point on the discharge scale, to a point on HW/D scale
(1). The value on the HW/D scale defines the required depth ratio if the culvert inlet
configuration consists of a headwall with a square-edged barrel entrance. If the inlet
configuration is instead a groove (socket) pipe end with a headwall or a projecting
groove pipe end, a horizontal line should be projected from the HW/D scale (1) inter-
section point to scale (2) or scale (3), respectively.
Example 9.1 Inlet Control for a Circular Culvert. A circular concrete pipe culvert has
a diameter of 36 in., a length of 100 ft, and a slope of 0.50 percent, as illustrated in Figure E9.1.1. The
invert of the culvert will match the channel bottomelevation, which is 125.30 ft at the upstreamend.
Assuming that the culvert operates under inlet control, use Figure 9.12 to estimate the headwater ele-
vations corresponding to discharges of 30 and 60 cfs. The culvert inlet consists of a headwall with a
square edge at the barrel entrance.
Solution: A line projected to scale (1) fromthe points representing 36 in. and 30 cfs on the diameter
and discharge scales yields HW/D =0.90. The headwater depth is therefore
HW = 0.90D =0.90(36/12) =2.70 ft
The headwater elevation is the sumof the channel bed elevation and the headwater depth, or
125.30 +2.70 =128.00 ft
Section 9.5 Culvert Hydraulics 341
For the discharge of 60 cfs, HW/D =1.65, and the headwater depth is therefore
HW = 1.65(3.00) =4.95 ft
The headwater elevation is
125.30 +4.95 =130.25 ft
Figure E9.1.1 Culvert for Example 9.1
The next example shows that simply changing the entrance edge configuration can
turn an otherwise unacceptable design into an acceptable one, thus illustrating the sig-
nificant influence that the entrance edge has on the performance of a culvert operating
under inlet control.
Example 9.2 Inlet Control for a Circular Culvert with Different Entrances.
Consider the same culvert described in the previous example, but assume that the maximumallowable
headwater elevation is 129.65 ft when the discharge is 60 cfs. Clearly, the square edge with headwall
design is not acceptable on this basis. Without changing the culvert diameter, determine whether an
alternative type of culvert entrance configuration can be used to lower the headwater elevation to
below 129.65 ft.
Solution: As described in Example 9.1, the value of HW/D found on scale (1) for the discharge of 60
cfs is 1.65. Extending a horizontal line on the chart fromthat point, one determines values of HW/D
on scales (2) and (3) to be 1.44 and 1.47, respectively. The respective headwater depths are 4.32 and
4.41 ft, and the respective headwater elevations are 129.62 and 129.71 ft. Comparing these values to
the allowable headwater depth of 129.65, it is concluded that use of a groove end with a headwall will
achieve the desired performance, whereas a projecting groove end (marginally) will not.
Outlet Control Charts. Anoutlet control nomograph such as the one shown in
Figure 9.13 is used to predict the culvert head, H, for a culvert operating under outlet
control. The head is equal to the total head loss through the culvert, but, in practice, it
can usually be taken as the difference between the headwater and tailwater elevations.
The variables that impact the outlet control head are culvert discharge, cross-section
geometry, length, entrance loss coefficient, and roughness. The nomograph presented
here is for circular concrete culverts with an n value of 0.012. As with the inlet control
nomograph from Figure 9.12, it is only one of many such charts published for differ-
ent combinations of culvert materials and shapes.
To use the nomograph shown in Figure 9.13, the engineer must first determine which
of the two length scales shown should be used based on the entrance loss coefficient
k
e
. The descriptions and entrance loss coefficient values provided in Table9.1 (which
is a repeat of Table6.5) can be used to determine which length scale (in this example,
k
e
=0.2 or k
e
=0.5) should be used.
342 Culvert Design Chapter 9
Figure 9.12
Nomograph to
compute headwater
depth for circular
concrete culverts with
inlet control
(Normann,
Houghtalen, and
J ohnston, 2001)
After the appropriate length scale has been determined, a straight line connecting the
applicable points on the diameter and length scales is drawn. Then, the head value is
determined by drawing a second straight line from the appropriate value on the dis-
charge scale, through the intersection point of the first line and the turning line, to
the head scale. An example of how these lines are constructed is shown on the chart.
Finally, an energy balance for full flow through a culvert with both the inlet and outlet
submerged can be written to express the headwater depth as
Section 9.5 Culvert Hydraulics 343
Figure 9.13
Nomograph for
computing head on
circular concrete pipes
with outlet control
(Normann,
Houghtalen, and
J ohnston, 2001)
HW= h
o
+ H S
o
L (9.1)
whereHW=headwater depth (ft, m)
h
o
=tailwater depth (ft, m)
H =total head loss through the culvert (ft, m)
S
o
=culvert barrel slope (ft/ft, m/m)
L =culvert length (ft, m)
This expression assumes that the velocity heads in the upstream and downstream
channels are equal, which is a reasonable approximation in most instances.
344 Culvert Design Chapter 9
Table 9.1 Entrance loss coefficients for pipes and culverts operating under outlet control
(Normann, Houghtalen, and J ohnston, 2001)
One method for determining the tailwater depth in Equation 9.1 is to approximate it
using a uniform flow formula such as the Manning equation. This approximation
requires information on the downstream channel cross-sectional dimensions, slope,
and roughness. Also, this approach assumes that the flow in the downstream channel
is approximately uniform and that flow controls downstream of the culvert do not
cause backwater or similar effects. An alternative (and more accurate) method for
determining the tailwater depth is to perform a flow profile analysis for the down-
stream channel. This method is clearly more laborious than the uniform flow method
just described, but is required when backwater or other effects cause the tailwater
depth to differ appreciably from that of uniform flow.
Example 9.3 Outlet Control for a Circular Culvert with Tailwater Effects.
Repeat Example 9.1 assuming that the culvert operates under outlet control. The invert elevation at
the downstreamend of the culvert is 124.80 ft, as illustrated in Figure E9.3.1. Backwater effects due
to an obstruction in the downstreamchannel cause the tailwater depth to be h
o
= 4.20 ft when Q = 30
Structure Type and Entrance Condition k
e
Concrete pipe
Projecting fromfill, socket or groove end 0.2
Projecting fromfill, square edge 0.5
Headwall or headwall and wingwalls
Socket or groove end
Square edge
Rounded (radius =D/12)
0.2
0.5
0.2
Mitered to conformto fill slope 0.7
End section conforming to fill slope 0.5
Beveled edges (33.7 or 45 bevels) 0.2
Side- or slope-tapered inlet 0.2
Corrugated metal pipe or pipe-arch
Projecting fromfill (no headwall) 0.9
Headwall or headwall and wingwalls (square edge) 0.5
Mitered to conformto fill slope 0.7
End section conforming to fill slope 0.5
Beveled edges (33.7 or 45 bevels) 0.2
Side- or slope-tapered inlet 0.2
Reinforced concrete box
Headwall parallel to embankment (no wingwalls)
Square-edged on 3 sides
Rounded or beveled on 3 sides
0.5
0.2
Wingwalls at 30 to 75 frombarrel
Square-edged at crown
Crown edge rounded or beveled
0.4
0.2
Wingwalls at 10 to 25 frombarrel
Square-edged at crown 0.5
Wingwalls parallel (extensions of box sides)
Square-edged at crown
Side- or slope-tapered inlet
0.7
0.2
Section 9.5 Culvert Hydraulics 345
cfs, and h
o
= 4.70 ft when Q = 60 cfs. By comparing the inlet and outlet control solutions, determine
whether the culvert operates under inlet or outlet control for discharges of 30 and 60 cfs.
Figure E9.3.1 Culvert for Example 9.3
Solution: FromExample 9.1, the circular concrete culvert has a diameter of 36 in., a length of 100 ft,
and a slope of 0.50 percent. The channel bed elevation at the upstreamend of the culvert is 125.30,
and the inlet configuration is a headwall with a square-edged barrel entrance.
FromTable9.1, the entrance loss coefficient for the this type of inlet configuration is 0.5. Using a
diameter of 36 in. and a length of 100 ft (on the scale representing k
e
=0.5), a straight-edge can be
used on Figure 9.13 to position a point on the turning line. Using that point and the discharge of 30
cfs, a line can then be projected to the head scale to determine H =0.60 ft. When the discharge is 60
cfs, the head loss is similarly found to be H = 2.45 ft.
Application of Equation 9.1 for the case when Q = 30 cfs yields
HW=4.20 +0.60 0.005(100) =4.30 ft
The headwater elevation is therefore
125.30 +4.30 =129.60 ft
Because this elevation is higher than the headwater elevation of 128.00 ft computed for inlet control,
conservative practice dictates that the outlet control headwater elevation of 129.60 ft governs for a
discharge of 30 cfs.
For the case of Q =60 cfs, Equation 9.1 yields
HW=4.70 +2.45 0.005(100) =6.65 ft
The headwater elevation is therefore
125.30 +6.65 =131.95 ft
Because this value is again higher than the headwater elevation determined for the assumption of inlet
control (130.25 ft), outlet control again governs and the headwater elevation is 131.95 ft for a dis-
charge of 60 cfs.
Figure 9.13 and other similar charts are based on the assumption that the culvert flows
full over its entire length and is submerged at both its upstream and downstream ends.
Nevertheless, such charts can be used to approximate outlet control behavior when the
tailwater elevation drops to or below the crown of the culvert outlet (Cases C through
G in Figure 9.11). The procedure yields the best approximation, however, when the
culvert flows full over at least a portion of its length. When the culvert flows only
partly full over its entire length, as in Case E, the method deteriorates as the headwa-
ter elevation drops further below the top of the culvert pipe. Adequate results may be
obtained for headwater depths as small as 0.75D, where D is the height of the culvert
barrel. Accurate calculations require flow profile analysis, as described in the next
section.
346 Culvert Design Chapter 9
If one chooses to use the approximate nomograph analysis of outlet control in Case D
or E, the tailwater depth used in Equation 9.1 should be the larger of (1) the actual
tailwater depth or (2) the arithmetic average of the critical depth in the barrel and the
height of the barrel [(y
c
+D)/2]. The critical depth in the barrel can be determined by
using the methods described in Section 7.4, page 246, or it can be estimated for circu-
lar pipes using Figure 9.14. As noted in this figure, in no case can the critical depth be
greater than the height of the barrel.
Example 9.4 Outlet Control for a Circular Culvert with Smaller Tailwater
Effects. Repeat Example 9.3 (page 344) assuming that the obstruction in the downstreamchannel
has been removed and lowering the tailwater depths to h
o
= 1.80 ft when Q = 30 cfs and h
o
= 2.70 ft
whenQ = 60 cfs, as illustrated in Figure E9.4.1. By comparing the inlet and outlet control solutions,
determine whether the culvert operates under inlet or outlet control for each of the discharges.
Figure E9.4.1 Culvert for Example 9.4
Solution: FromExample 9.3, H =0.60 ft when Q =30 cfs. Using Figure 9.14, the critical depth in the
culvert barrel for this discharge is about 1.75 ft. When using the outlet control nomographs with Case
D or E flow, the tailwater depth should be taken as the larger of the actual tailwater depth or the aver-
age of tailwater depth and barrel height. Averaging this depth with the culvert diameter yields 2.38 ft.
Comparing this value with the actual tailwater depth of 1.80 ft and taking the larger of the two leads to
h
o
=2.38 ft. Substitution into Equation 9.1 yields
HW=2.38 +0.60 0.005(100) =2.48 ft
This value is approximately 83 percent of the height of the culvert barrel, and hence the result, while
only an approximation, may be adequate. The headwater elevation is
125.30 +2.48 =127.78 ft
Comparing this elevation to the inlet control headwater elevation of 128.00 ft found in Example 9.1
(page 340) shows that inlet control governs the performance. The headwater elevation for this dis-
charge would therefore be 128.00 ft.
When Q =60 cfs, H =2.45 ft (fromExample 9.3). The critical depth for this discharge is about 2.50 ft
(fromFigure 9.14). Averaging the critical depth with the culvert barrel height yields
(3.00 +2.50)/2 =2.75 ft
Because 2.75 ft is larger than the actual tailwater depth, h
o
=2.75 ft. Substitution into Equation 9.1
yields
HW=2.75 +2.45 0.005(100) =4.70 ft
Because this depth is larger than the culvert diameter, the culvert flows full over part of its length;
therefore, the quality of the approximation is better than it is for the smaller discharge of 30 cfs. The
headwater elevation is
125.30 +4.70 =130.00 ft
Because this elevation is again lower than the headwater elevation determined for the assumption of
inlet control (130.25 ft), inlet control governs and the headwater elevation is 130.25 ft for a discharge
of 60 cfs.
Section 9.5 Culvert Hydraulics 347
Figure 9.14
Graphs for estimating
critical depths for
circular pipes
Credit: FHWA, 1985
348 Culvert Design Chapter 9
Flow Profile Analysis of Culvert Performance
As shown in the previous sections and examples, culvert performance analysis using
the nomographs published in HDS-5 (Normann, Houghtalen, and J ohnston, 2001) is
convenient, but doing so yields only approximate solutions in many practical cases.
More accurate analysis requires computation of flow profiles and makes use of the
closed-conduit and open-channel flow relationships presented in Chapter 6 and Chap-
ter 7. The subsections that follow describe flow profile analysis methods for both inlet
and outlet control. The tedious nature of these methods makes the use of computer-
based modeling tools attractive. The methods described herein are the basis for the
algorithms used in many modern computer tools.
Inlet Control. The entrance to a culvert operating under inlet control behaves as a
weir when the headwater depth is smaller than the culvert entrance height. It operates
as an orifice if the headwater submerges the entrance. This behavior is shown by the
two curves in Figure 9.15. Between the two curves representing weir flow and orifice
flow is a transition zone representing an intermediate type of flow. A graph such as
Figure 9.15 is a key to inlet control performance evaluation using flow profile analy-
sis.
When the culvert entrance operates as a weir, the HW/D ratio can be expressed as
(Normann, Houghtalen, and J ohnston, 2001)
(9.2)
An alternative expression is
(9.3)
where HW =headwater depth (ft, m)
D =culvert barrel height (ft, m)
E
c
=specific energy (ft) corresponding to the barrel critical depth
(E
c
=y
c
+V
c
2
/2g), (ft, m)
Q =discharge (cfs, m
3
/s)
A =full cross-sectional area of the barrel (ft
2
, m
2
)
S
o
=barrel slope (ft/ft, m/m)
C
f
=conversion factor (1.0 U.S. Customary, 1.811 SI)
K and M are constants determined from hydraulic model tests
Values of K and M are provided in Table9.2, along with an indication of whether
Equation 9.2 or 9.3 should be used in the calculation. Equation 9.2 is theoretically
better than Equation 9.3, but Equation 9.3 is easier to apply in practice. The slope
term expressed as (0.5S
o
) in Equation 9.2 should be changed to (+0.7S
o
) if the barrel
entrance is mitered. Equations 9.2 and 9.3 are applicable for values of Q/AD
1/2
less
than or equal to about 3.5 for U.S. Customary units, or 1.93 for SI units.
1/ 2
0.5
M
f
c
o
C Q
E HW
K S
D D
AD
| |
= +
|
|
\ .
1/ 2
M
f
C Q
HW
K
D
AD
| |
=
|
|
\ .
Section 9.5 Culvert Hydraulics 349
Figure 9.15
Transition fromweir
to orifice control in a
culvert (Norman,
Houghtalen, and
J ohnston, 2001)
If Q/AD
1/2
is between about 3.5 (1.93 SI) and 4.0 (2.21 SI), transitional flow occurs,
and the headwater value must be interpolated as shown in the transition zone in Figure
9.15. Above this range, the culvert entrance acts as an orifice, and the HW/D ratio can
be expressed as (Normann, Houghtalen, and J ohnston, 2001)
(9.4)
wherec and Y are constants determined from hydraulic model tests
Values of the constants c and Y are provided in Table9.2. Equation 9.4 is valid for val-
ues of Q/AD
1/2
greater than or equal to about 4.0. Below this range, transitional flow
occurs. Again, the slope term (0.5S
o
) should be replaced with (+0.7S
o
) if the barrel
entrance is mitered.
2
1/ 2
0.5
f
o
C Q
HW
c Y S
D
AD
| |
= +
|
|
\ .
350 Culvert Design Chapter 9
Table 9.2 Constants for inlet control equations (Normann, Houghtalen, and J ohnston, 2001)
Culvert Shape
and/or Material
Inlet Edge
Description
Unsubmerged (Weir Flow) Submerged (Orifice Flow)
Equation
in text K M c Y
Circular, concrete
Square edge with headwall
(9.2)
.0098 2.0 .0398 0.67
Groove end with headwall .0018 2.0 .0292 .74
Groove end projecting .0045 2.0 .0317 .69
Circular, CMP
Headwall
(9.2)
.0078 2.0 .0379 .69
Mitered to slope .0210 1.33 .0463 .75
Projecting .0340 1.50 .0553 .54
Circular
Beveled ring, 45 bevels
(9.2)
.0018 2.50 .0300 .74
Beveled ring, 33.7 bevels .0018 2.50 .0243 .83
Rectangular box
30 to 75 wingwall flares
(9.2)
.026 1.0 .0347 .81
90 and 15 wingwall flares .061 0.75 .0400 .80
0 wingwall flares .061 0.75 .0423 .82
Rectangular box
45 wingwall flares, d = 0.043D
(9.3)
.510 .667 .0309 .80
18 to 33.7 wingwall flares, d = 0.083D .486 .667 .0249 .83
Rectangular box
90 headwall with -in. chamfers
(9.3)
.515 .667 .0375 .79
90 headwall with 45 bevels .495 .667 .0314 .82
90 headwall with 33.7 bevels .486 .667 .0252 .865
Rectangular box
45 skewed headwall; -in. chamfers
(9.3)
.545 .667 .0505 .73
30 skewed headwall; -in. chamfers .533 .667 .0425 .705
15 skewed headwall; -in. chamfers .522 .667 .0402 .68
10-45 skewed headwall; 45 bevels .498 .667 .0327 .75
Rectangular box
with -in. chamfers
45 nonoffset wingwall flares
(9.3)
.497 .667 .0339 .803
18.4 nonoffset wingwall flares .493 .667 .0361 .806
18.4 nonoffset wingwall flares; 30 .495 .667 .0386 .71
Rectangular box
w/ top bevels
45 wingwall flares, offset
(9.3)
.497 .667 .0302 .835
33.7 wingwall flares, offset .495 .667 .0252 .881
18.4 wingwall flares, offset .493 .667 .0227 .887
Corrugated metal
boxes
90 headwall
(9.2)
.0083 2.0 .0379 .69
Thick wall projecting .0145 1.75 .0419 .64
Thin wall projecting .0340 1.5 .0496 .57
Horizontal ellipse,
concrete
Square edge w/ headwall
(9.2)
.0100 2.0 .0398 .67
Groove end w/ headwall .0018 2.5 .0292 .74
Groove end projecting .0045 2.0 .0317 .69
Vertical ellipse,
concrete
Square edge w/ headwall
(9.2)
.0100 2.0 .0398 .67
Groove end w/ headwall .0018 2.5 .0292 .74
Groove end projecting .0095 2.0 .0317 .69
Pipe arch, CM,
18-in. corner radius
90 headwall
(9.2)
.0083 2.0 .0379 .69
Mitered to slope .0300 1.0 .0463 .75
Projecting .0340 1.5 .0496 .57
Pipe arch, CM,
18-in. corner radius
Projecting
(9.2)
.0300 1.5 .0496 .57
No bevels .0088 2.0 .0368 .68
33.7 bevels .0030 2.0 .0269 .77
Pipe arch, CM,
31-in. corner radius
Projecting
(9.2)
.0300 1.5 .0496 .57
No bevels .0088 2.0 .0368 .68
33.7 bevels .0030 2.0 .0269 .77
Arch, CM
90 headwall
(9.2)
.0083 2.0 .0379 .69
Mitered to slope .0300 1.0 .0463 .75
Thin wall projecting .0340 1.5 .0496 .57
Section 9.5 Culvert Hydraulics 351
Under inlet control conditions, flow in the barrel just downstream of the culvert
entrance is supercritical and forms an S2 water surface profile that asymptotically
approaches normal depth. The direct-step method, described in Section 7.8 on page
275, can be used to compute the profile.
If the tailwater level is high and submerges the culvert outlet as shown in Cases B and
D of Figure 9.9, one of two possible situations occurs. In the first situation (which is
illustrated for Cases B and D in Figure 9.9), pressure flow occurs in the most down-
stream part of the culvert and a hydraulic jump forms inside the barrel. In the second
situation (not illustrated), the S2 water surface profile extends all the way through the
culvert to its outlet and a hydraulic jump may form downstream of the outlet. In either
case, formation of a hydraulic jump is contingent on a momentum balance being
achieved (see Section 7.5, page 260).
Example 9.5 Inlet Control Equations. Assume that the culvert described in Example 9.1,
illustrated in Figure E9.5.1, operates under inlet control. Use Equations 9.2, 9.3, and/or 9.4 to deter-
mine the headwater depths for discharges of 30 cfs, 45 cfs, and 60 cfs. Compare the headwater depths
for the discharges of 30 cfs and 60 cfs to those determined in Example 9.1.
Figure E9.5.1 Culvert for Example 9.5
Solution: FromExample 9.1, a circular concrete culvert with a square-edge headwall entrance has
L =100 ft, D =36 in., S
o
=0.5%, and an upstreaminvert elevation of 125.30 ft. For each of the three
discharges, the values of Q/AD
1/2
and the corresponding flow types (based on criteria given in the
preceding sections) are as follows:
Consider first the discharge of 30 cfs, for which weir flow occurs. For a circular culvert with head-
walls and a square-edged barrel entrance, Table9.2 indicates that the constants K
1
and M
1
are equal to
0.0098 and 2.0, respectively, and that these values should be used with Equation 9.2. By trial and
error, the critical depth in the barrel corresponding to this discharge is y
c
=1.77 ft, and the corre-
sponding specific energy is E
c
=2.51 ft. Substitution of these values into Equation 9.2 yields
HW=2.51 +3.00(0.0098)(2.45)
2
0.5(0.005)(3.00) =2.68 ft
This value compares quite favorably with the depth of 2.70 ft determined in Example 9.1.
Orifice flow occurs for the discharge of 60 cfs; thus, Equation 9.4 applies. FromTable9.2, the values
of K
2
and M
2
are 0.0398 and 0.67, respectively. The computed headwater depth is
HW=3.00(0.0398)(4.90)
2
+3.00(0.67) 0.5(0.005)(3.00) =4.87 ft
Q (cfs) Q/AD
1/2
Flow Type
30 2.45 Weir flow
45 3.68 Transitional flow
60 4.90 Orifice flow
352 Culvert Design Chapter 9
Again, this compares favorably with the depth of 4.95 ft determined in Example 9.1.
To determine the headwater depth for a discharge of 45 cfs (for which transitional flow occurs) a lin-
ear interpolation procedure will be used. Equation 9.2 applies for Q/AD
1/2
values up to 3.5, whereas
Equation 9.4 applies for Q/AD
1/2
values greater than 4.0. Thus, the largest discharge that Equation 9.2
applies to for this culvert is 43 cfs, and the smallest discharge that Equation 9.4 applies to is 49 cfs.
UsingEquation 9.2, the headwater depth corresponding to a discharge of 43 cfs is found to be 3.48 ft.
Similarly, using Equation 9.4, the headwater depth corresponding to a discharge of 49 cfs is found to
be 3.91 ft. Linear interpolation between the coordinate pairs (Q =43 cfs, HW =3.48 ft) and (Q =49
cfs, HW =3.91 ft) produces a headwater depth of HW = 3.62 ft for a discharge of 45 cfs.
Outlet Control. The type of flow profile analysis performed for a culvert operat-
ing under outlet control depends on the profile condition from Figure 9.11 that is
being analyzed. Equation 9.1 can be applied directly if the culvert outlet is submerged
and the flow corresponds to either Case A (a full-flow condition) or Case B (full flow
except for the contraction at the culvert entrance) from Figure 9.11. Note, however,
that this equation assumes that the velocity heads in the upstream and downstream
channels are approximately equal and thus cancel one another out in the energy equa-
tion.
Case C is a case of full flow that occurs when the critical depth in the culvert barrel is
equal to the height of the barrel (it can be no larger) and the tailwater depth is smaller
than the height of the barrel. In this case, the hydraulic grade line elevation at the out-
let is equal to the elevation of the crown of the barrel. Assuming that the velocity head
in the upstream channel is negligible, the headwater elevation is greater than the outlet
crown elevation by an amount consisting of the sum of the velocity head in the barrel,
the frictional loss along the length of the barrel, and the entrance loss. The entrance
loss can be expressed as the product of the velocity head in the barrel and a minor loss
coefficient k
e
(see Table9.1). If the Manning equation is applied for evaluation of
frictional losses, the friction slope S
f
can be estimated using Equation 6.19. The prod-
uct of that slope and the barrel length is the total head loss due to friction. Thus, the
headwater elevation can be expressed as
(9.5)
where EL
hw
=headwater elevation (ft, m)
EL
o
=invert elevation at the outlet end of the culvert (ft, m)
D =culvert barrel height (ft, m)
S
f
= slope of the energy grade line (ft/ft, m/m)
L = culvert barrel length (ft, m)
k
e
= culvert entrance loss coefficient
V =full-flow velocity inside the culvert barrel (ft/s, m/s)
Cases D and E in Figure 9.11 occur when both the critical depth and the tailwater
depth are smaller than the height of the barrel. A flow profile analysis for either of
these cases begins with the determination of whether the tailwater depth is greater
( )
2
1
2
hw o f e
V
EL EL D S L k
g
= + + + +
Section 9.5 Culvert Hydraulics 353
than or equal to critical depth or less than critical depth. If the tailwater depth is
greater than or equal to critical depth, the depth at the control section at the down-
stream end of the barrel is equal to the tailwater depth, and computation of the M2
flow profile begins at this depth and proceeds in the upstream direction. If the tailwa-
ter depth is smaller than the critical depth, the M2 flow profile calculation starts at
critical depth and proceeds upstream.
A Case D condition occurs if the M2 water surface profile reaches the crown of the
culvert at some distance upstream of the outlet. Upstream of the point where the water
surface intersects the crown, the culvert flows full and the friction slope can evaluated
using Equation 6.19. An extension of the friction slope to the inlet of the culvert
yields the hydraulic grade line elevation just inside the culvert entrance. The entrance
loss and the full-flow velocity head can be added to this value to obtain the headwater
elevation.
A Case E condition occurs if the M2 flow profile extends all the way to the culvert
inlet without intersecting the crown. In this case, the hydraulic grade line elevation
just inside the barrel entrance is equal to the computed water surface elevation at that
point. The headwater elevation is determined by adding the velocity head just inside
the barrel entrance and the entrance loss to this computed water surface elevation.
Example 9.6 Determination of Outlet Control with Gradually Varied Flow. A
rectangular concrete box culvert has a span of 1.2 mand a rise (height) of 1.0 m. It is 60.0 min length
and has a slope of 0.0015, as illustrated in Figure E9.6.1. Assuming that the culvert operates under
outlet control and the tailwater depth is 0.7 m, determine the headwater elevation for a discharge of 2
m
3
/s. Use Mannings n =0.013. The invert elevation at the culvert outlet is 30.5 m, and the culvert
entrance has a loss coefficient of 0.2.
Figure E9.6.1 Culvert for Example 9.6
Solution: Because the tailwater depth is smaller than the barrel height, the outlet is not submerged.
The critical depth inside the rectangular barrel is found to be
y
c
={2
2
/[9.81(1.2)
2
]}
1/3
=0.66 m
Because this value is smaller than the barrel height, the flow corresponds to either Case D or Case E
inFigure 9.7. The normal depth in a rectangular channel with the width, slope, and roughness of the
culvert barrel is y
n
=1.06 m. Because this depth is larger than the culvert height, the flow might corre-
spond to Case D. However, profile calculations are necessary to confirmthis. Had the normal depth
been less than the barrel height, one could say without qualification that the flow would correspond to
Case E.
354 Culvert Design Chapter 9
Computation of the M2 flow profile using the direct-step method (see Section 7.8 on page 275) is
summarized in the following table. Because the tailwater depth of 0.7 mexceeds critical depth, calcu-
lations begin at this elevation.
It can be seen fromthis summary that the water surface profile does not intersect the crown of the cul-
vert; thus, the flow profile corresponds to Case E.
More detailed M2 water surface profile calculations indicate that the depth of flow at the upstream
end of the culvert just inside its entrance is about 0.9 m. The corresponding velocity is
V =2/[0.89(1.20)] =1.85 m/s
and the velocity head is 0.18 m. The headwater depth is the sumof the depth inside the entrance, the
velocity head, and the entrance loss:
HW=0.9 +(1 +0.2)(0.18) =1.116 m
The culvert invert elevation at its entrance is
30.5 +60(0.0015) =30.59 m
and thus the headwater elevation is
30.59 +1.116 =31.706 m
Outlet Velocity
It is necessary to evaluate the velocity of flow emerging from a culvert in order to
design measures to protect against erosion at the outlet (see Section 9.8). If the outlet
is submerged, or if the barrel flows full over its entire length, then the outlet velocity
is simply the ratio of the discharge to the cross-sectional area of the barrel. As shown
in Figures 9.9 and 9.11, this situation can occur for culverts operating under either
inlet or outlet control. When the outlet is not submerged, the procedure for determin-
ing the outlet velocity is more involved.
For a culvert operating under inlet control, an accurate estimate of the outlet velocity
requires a flow profile analysis. However, the outlet velocity can be approximated
using a simpler procedure that is conservative in that it always yields a value greater
than or equal to the actual velocity. This procedure assumes that normal depth exists
in the culvert at the outlet. Because normal depth may be smaller than the actual
depth, the cross-sectional area of the flow is assumed to be smaller than it may be in
reality, which in turn results in a higher calculated velocity. Determination of the nor-
mal depth and the associated cross-sectional area is a trial-and-error process that may
be accomplished using the procedures described in Section 7.3.
When a culvert operates under outlet control and the outlet is not submerged, the
depth of flow at the outlet should be taken as the larger of the tailwater depth and the
critical depth in the culvert barrel. The cross-sectional area of the flow in the barrel
Depth (m) Distance from Outlet (m)
0.70 0
0.76 12.18
0.82 28.70
0.88 51.07
0.94 81.71
Section 9.6 Roadway Overtopping 355
may be computed based on the chosen depth, and the outlet velocity may be deter-
mined.
Example 9.7 Outlet Velocity Determination. Estimate the outlet velocity for the 4-ft by
3-ft culvert fromExample 9.6 (page 353).
Solution: As described in Example 9.6, the flow is 2 m
3
/s, the outlet is not submerged, and the tailwa-
ter depth is 0.70 m. This depth is greater than the critical depth in the barrel, and thus is the depth at
which flow exits the barrel. The cross-sectional area of the flow at the barrel outlet is
A =0.70(1.2) =0.84 m
2
The outlet velocity is therefore
V =Q/A =2/0.84 =2.38 m/s
9.6 ROADWAY OVERTOPPING
If the actual discharge approaching a culvert exceeds the design discharge for the
structure, or if the culvert has become clogged by accumulated sediment or debris, the
headwater elevation at the upstream end of the culvert will rise and may lead to over-
topping of the embankment, or flow over the embankment. Overtopping can cause a
hazard to the traveling public and, in severe cases, can cause the entire embankment
to fail and be washed out.
Flow over a roadway embankment usually occurs at the low point of a sag vertical
curve and is similar to the flow over a broad-crested weir. Equation 9.6 relates dis-
charge to the head on the weir:
Q = C
d
L(HW
r
)
3/2
(9.6)
where Q =discharge (cfs, m
3
/s)
C
d
= discharge coefficient (ft
1/2
/s, m
1/2
/s)
L =crest length normal to the direction of flow (ft, m)
HW
r
= head on the weir (ft, m)
HW
r
is computed as the difference between the headwater elevation and the roadway
embankment elevation. The discharge coefficient is determined as
C
d
= k
t
C
r
(9.7)
where k
t
=correction factor for submergence
C
r
= base coefficient value (ft
1/2
/s, m
1/2
/s)
The value of C
r
depends on HW
r
, the breadth of the embankment in the direction of
flow (L
r
), and whether the embankment has a paved or gravel surface. The term k
t
is a
correction factor to account for the effect of potential submergence of the weir and is
a function of the ratio of the tailwater height h
t
to HW
r
. Numerical values of C
r
andk
t
may be obtained by using Figure 9.16. It is important to note that the value of C
r
obtained from Figure 9.16 will be in ft
1/2
/s. This value must be converted to m
1/2
/s
before solving Equation 9.6 using SI units.
356 Culvert Design Chapter 9
Figure 9.16
Graphs for
determining the
broad-crested weir
discharge coefficient
used in computing
flow over a roadway
(Normann,
Houghtalen, and
J ohnston, 2001)
The weirs crest length and elevation can be difficult to estimate when the crest is
defined by a sag vertical curve in a roadway profile. An approximate solution in this
situation is to subdivide the total crest length into a series of horizontal segments. For
the example shown in Figure 9.17(a), the roadway crest is approximated as three weir
segments of lengths X
1
, X
2
, and X
3
. The discharge for each segment can be computed
separately, and then the total discharge can be computed as the sum of the individual
segment discharges.
Normann, Houghtalen, and J ohnston (2001) reported that it is often adequate in cul-
vert design to represent a sag vertical curve by a single horizontal weir segment [see
Figure 9.17(b)]. The length of the weir X in this case can be taken as the distance
between the points on the sag vertical curve having the same elevation as the headwa-
ter. The head on the weir can be approximated as half of the difference between the
headwater elevation and the elevation of the lowest point on the vertical curve.
Section 9.6 Roadway Overtopping 357
Figure 9.17
Methods of
computing
overtopping
discharges at a sag
vertical curve
(Normann,
Houghtalen, and
J ohnston, 2001)
Credit: FHWA, 1985
If a total discharge exceeding the design capacity of a culvert is specified, a trial-and-
error approach is required to determine how much of the flow passes through the cul-
vert barrel and how much overtops the roadway embankment. Clearly, the sum of the
two individual discharges must be equal to the specified total. If a performance curve
has been developed and graphed for the culvert installation, it can be used to obtain a
direct, graphical solution to a problem of this type.
358 Culvert Design Chapter 9
Example 9.8 Roadway Overtopping. For the culvert described in Example 9.1, assume
that the roadway embankment is paved and has a breadth of 70 ft, as illustrated in Figure E9.8.1. The
lowest roadway elevation at the sag point of a vertical curve is 130.70 ft. The sag vertical curve crest
of the roadway embankment is described by the parabolic equation
y =130.70 +0.000075x
2
wherey = crest elevation (ft)
x =distance fromthe lowest point in the sag (ft)
Estimate the discharge over the roadway embankment when the headwater elevation is 131.20 ft.
Assume that a single horizontal segment can be used to approximate the weir and the tailwater eleva-
tion downstreamof the culvert does not submerge the weir.
Figure E9.8.1 Culvert and roadway profile for Figure 9.8
Solution: The maximumdepth of submergence of the roadway embankment (weir) is the difference
between the headwater elevation and the minimumcrest elevation, or
131.20 130.70 =0.50 ft
The average head on the weir is HW
r
=0.25 ft (one-half the maximum depth of submergence). The
length of the weir crest across which flow occurs is found by computing the values of x in the para-
bolic equation that satisfy y =131.20. These values are 81.65 ft and +81.65 ft; thus, the total crest
length is L =163.30 ft.
The ratio of HW
r
toL
r
is
HW
r
/L
r
=0.25/70.0 =0.00357
Section 9.7 Performance Curves 359
FromFigure 9.16, k
t
= 1.00 and C
r
= 2.97 ft
1/2
/s, so
C
d
= k
t
C
r
=2.97 ft
1/2
/s
Substituting this value into Equation 9.6, along with the crest length and average head, yields the dis-
charge over the embankment as
Q =2.97(163.30)(0.25)
3/2
=61 cfs
9.7 PERFORMANCE CURVES
A performance curve for a culvert is a graphical relationship between the discharge
and the headwater elevation (or depth). As shown in Figure 9.18, the overall perfor-
mance curve for a culvert is constructed from individual performance curves repre-
senting inlet control, outlet control, and roadway overtopping. It therefore illustrates,
at a glance, the type of flow control that exists for any discharge of interest, the corre-
sponding headwater elevation, and the depth of flow over the roadway (if any). If
detention effects upstream of the culvert are significant and are thus taken into consid-
eration in optimizing the culvert design, then the performance curve is the same as the
stage-discharge curve required for the detention basin routing calculations (see Sec-
tion 12.6, page 505).
Figure 9.18
Construction of the
overall performance
curve for a culvert
(Normann,
Houghtalen, and
J ohnston, 2001)
360 Culvert Design Chapter 9
An inlet control performance curve such as the one illustrated in Figure 9.18 is con-
structed by assuming various discharges, determining the corresponding headwater
elevations, and plotting the headwater versus discharge data on a graph. Either the
nomographs or the inlet control equations presented in Section 9.5 (page 348) can be
used for this purpose.
Assumed discharges used in developing the curve should range from zero to a value
that is greater than the design discharge. If the inlet control equations are used instead
of nomographs, care should be taken to represent the transitional range of flows as
well as the weir and orifice flow ranges.
The outlet control performance curve is also graphed by assuming various discharges
and determining the corresponding headwater elevation for each. The chosen range of
discharges should span the design discharge. The curve can be developed by using
nomographs, but it is more accurate to use flow profile analysis as described in Sec-
tion 9.5 (page 352).
After the inlet and outlet control performance curves have been plotted on the same
graph, the overall performance curve is formed from those parts of the individual
curves that have the highest headwater for a given discharge. For example, at loca-
tions on the graph where the inlet control curve is higher than the outlet control curve,
the overall performance curve is the same as the inlet control curve. In a similar vein,
the overall performance curve is the same as the outlet control curve where it is higher
than the inlet control curve (see Figure 9.18). This procedure for delineation of the
overall performance curve applies only to the portion for which the headwater eleva-
tion is at or below the roadway crest.
For headwater elevations above the crest, roadway overtopping must also be consid-
ered. The head on the weir represented by the sag vertical curve is equal to the differ-
ence between the headwater elevation and the roadway crest elevation. Knowing this
head, the discharge over the weir can be determined by using Equation 9.6. This dis-
charge, added to the discharge that passes through the culvert for that same headwater
elevation, defines the overall performance curve for the region where the headwater
elevation is above the roadway crest. As shown in Figure 9.18, this portion of the
overall performance curve is generally quite flat compared to the rest of the curve,
indicating that the water surface elevation rises much more slowly with increasing
discharge once roadway overtopping has begun.
9.8 EROSION PROTECTION
The velocity at which water exits a culvert is usually much higher than the velocity in
the channel and can cause objectionable scour and erosion if the excess kinetic energy
associated with the flow is not dissipated. Section 9.5 (page 354) describes how to
estimate the velocity at which a flow exits a barrel. That exit velocity, along with geo-
metric information related to the culvert design, can be used to predict the size and
extent of required riprap erosion protection at the outlet (see Figure 9.19). Section
11.7 describes this process.
Section 9.8 Erosion Protection 361
Figure 9.19
Riprap outlet
protection
For instances in which the kinetic energy caused by the exit velocity of the flow is too
great to be safely dissipated using a riprap blanket, several options are available. The
exit velocity may be reduced by modifying the culvert barrel near the outlet. For
example, one might design a broken-back culvert with a smaller slope at the outlet
than at the inlet, thereby reducing the outlet velocity. Another approach is to increase
the barrel roughness near the outlet by using a material such as corrugated metal
instead of concrete in this area. The rougher corrugated metal increases the depth and
reduces the velocity near the barrel outlet.
If barrel modifications such as those just described cannot reduce the outlet velocity
to an acceptable limit, one must resort to use of a type of stilling basin structure at the
outlet, such as an impact-type energy dissipator (which uses a baffle in front of the
water jet to dissipate energy). Corry et al. (1983) provide further direction on this sub-
ject.
In culvert design, it is often desirable to provide erosion protection measures at the
entrance as well as the outlet. As flow enters a culvert, it must usually contract into a
smaller area than that of the upstream channel. The result is increased velocity, which
in turn leads to localized erosion of the channel bed and banks. The erosion protection
necessary at an entrance is typically not as extensive as that required at the outleta
riprap blanket, concrete apron, or slope paving is usually sufficient.
362 Culvert Design Chapter 9
9.9 ENVIRONMENTAL CONSIDERATIONS
Culverts can often be used in lieu of bridges for small streams; however, culverts can
pose obstacles to migration pathways if the stream supports fish and similar aquatic
organisms. For instances in which fish passage is a concern, it is essential that input
and requirements of state and/or federal fish and wildlife protection officials be
obtained early in the design process. In some instances, a bridge may be required even
though a less expensive culvert could provide the required hydraulic performance.
To ensure that culverts are suitable for fish passage, additional design criteria may be
introduced to govern minimum flow depths within the barrel and minimum and/or
maximum flow velocities. Seasonalities of fish migration and streamflow are impor-
tant considerations here. The species of fish present in the stream are also important,
as some species are less tolerant of certain flow depths and velocities than are other
species. Another consideration is culvert lighting. Fish may not enter dark culverts
even when conditions are otherwise favorable. Long culverts may be constructed in
sections to allow natural light to enter. Artificial lighting may also be an option in
some circumstances.
Perhaps the most significant requirements for fish passage relate to the need to simu-
late the natural streambed environment within the culvert structure. If an open-bottom
culvert is used, such as a field-assembled corrugated arch structure, natural channel
materials can be left in place along the culvert bottom. This type of culvert requires
adequate foundation support for the culvert arch bearing ends. Localized scour or
general channel bed degradation that could undermine the foundations is another con-
cern. If a culvert pipe is used instead of an open-bottom structure, the barrel may be
oversized as shown in Figure 9.20. The barrel can then be partially filled with natural
streambed materials.
Figure 9.20
Partial filling of a
culvert pipe with
natural streambed
materials
In some cases, special baffles and/or weirs are installed in culvert barrels to simulate
natural riffles and pools that may exist in undisturbed areas of the stream. Design pro-
cedures for these types of installations, and for fish passage facilities in general, can
be found in the Model Drainage Manual (AASHTO, 1991) and in Chang and Nor-
mann (1976). Fish and wildlife agencies also may provide specific design guidance.
Section 9.9 Environmental Considerations 363
Modeling Focus: Supplemental Culvert Design
An existing roadway streamcrossing consists of a 6-ft (span)
by 4-ft (high) concrete box culvert with the following charac-
teristics:
Upstreaminvert elevation =124.75 ft
Downstream invert elevation =124.55 ft
Length =60.0 ft
Headwall with 30 to 75 wingwall flares, K
e
=0.5
The minimumelevation of the roadway sag vertical curve is
132.55 ft. The culvert outfalls to a downstreamtrapezoidal
channel with the following characteristics:
Channel invert elevation matches culvert outlet invert
Bottom width =6 ft
Side slopes are =1.5 H to 1.0 V
Mannings n =0.045
Channel slope =0.0075 ft/ft
A computer programsuch as CulvertMaster can be used to
analyze the capacity of the culvert and generate a rating curve.
For the culvert and channel characteristics described above
and a maximumheadwater elevation of 132.55 ft (the road-
way overtopping elevation), the culvert capacity is found to be
269 cfs. However, allowing for 1 ft of freeboard between the
maximumheadwater elevation and the minimumroadway ele-
vation (i.e., the maximumheadwater elevation is 131.55 ft),
the culvert capacity is 236 cfs. The rating curve for the culvert
is shown below.
Improvements are being made to the roadway, and the ability
of the existing culvert to meet current needs must be evalu-
ated. The culvert was originally designed to convey the 50-
year peak flow fromthe NRCS (SCS) Type II stormwith 1 ft
of freeboard; however, substantial development has occurred
in the area since the original structure was installed.
Assuming that the watershed will be fully developed (subject
to zoning regulations) within 5 years, the 50-year design flow
is found to be 295 cfs. It is apparent that the capacity of the
existing culvert is insufficient to meet the design require-
ments. Because the existing culvert is in good condition, it
will remain in place but will be supplemented by a new cul-
vert in parallel.
The supplemental culvert must be capable of conveying
295 cfs 236 cfs =59 cfs at the allowable headwater of
131.55 ft. The engineer decides to evaluate the feasibility of
supplementing with a circular, concrete culvert section. The
existing headwalls will be extended to accommodate the new
culvert.
Assuming that the inverts and length for the new culvert will
match the existing culvert and using a square edge with head-
wall entrance type, the engineer enters the needed capacity of
59 cfs and allowable headwater elevation of 131.55 ft and
allows the programto solve for the minimumcommercially
available size that meets the criteria-in this case, a 36-in. pipe.
The 36-in. pipe is inserted into the program's worksheet for
the existing system, and the programperforms iterative calcu-
lations to solve for a headwater elevation of 131.44 ft on both
culverts for design flow conditions. This design therefore
meets the minimum1-ft freeboard requirement. The program
may also report the exit velocities, which are 9.6 cfs for the
box culvert and 9.1 cfs for the 36-in. culvert. Outlet protection
such as a riprap apron (see Section 11.7) should be provided.
The engineer is also asked to check the design for the 100-
year stormevent, which corresponds to a peak discharge of
336 cfs. The headwater on the culverts is found to be 132.46 ft
for the 100-year storm, which is slightly below the minimum
roadway elevation (i.e., the road will not be overtopped for the
100-year check storm).
The individual and composite rating curves for the proposed
culvert systemare shown below.
124
126
128
130
132
134
0 50 100 150 200 250 300
Discharge, cfs
H
e
a
d
w
a
t
e
r

E
l
e
v
a
t
i
o
n
,

f
t
124
126
128
130
132
134
136
0 50 100 150 200 250 300 350
Discharge (cfs)
H
e
a
d
w
a
t
e
r
E
l
e
v
a
t
i
o
n
(
f
t
)
Supplemental
48-in. Culvert
Existing Culvert
Both culverts
364 Culvert Design Chapter 9
9.10 CHAPTER SUMMARY
A culvert is a relatively short conveyance conduit that passes through an obstruction
such as a roadway embankment. Culverts can have a variety of cross-sectional shapes,
including circular, elliptical, arched and rectangular geometries, and may consist of
multiple barrels. Available materials include concrete, corrugated steel, PVC, corru-
gated aluminum, and polyethylene.
Culverts are typically designed to convey the peak flow resulting from a design storm
of a particular recurrence interval (for instance, the peak flow resulting from a 24-
hour, 25-year storm event) with a headwater elevation that does not exceed an allow-
able level. If significant upstream ponding occurs, the detention effects will result in a
reduction of the design storm peak flow that should be considered in the evaluation.
As an alternative to the design-storm approach, culverts may be designed using eco-
nomic analysis. The goal of this approach is to minimize total cost if both flood dam-
ages and the capitalized cost of the culvert are considered.
Culvert performance may also be checked for a storm event of greater magnitude than
the design storm (often, the 100-year event). This evaluation typically involves an
analysis of roadway overtopping in addition to culvert hydrualics.
Various end treatments are used at culvert entrances and exits to provide erosion con-
trol and improve hydraulic efficiency. The culvert barrel may simply project from the
embankment, or it may be mitered to conform to the embankment slope. Precast end
sections are often available from pipe manufacturers. Alternatively, headwallswith
or without wingwallsmay be used at one or both ends. Finally, specialized inlet
types, such as side-tapered and slope-tapered inlets, may be necessary if certain geo-
metric constraints are present.
Section 9.10 Chapter Summary 365
Although culvert barrels are typically straight and roughly perpendicular to the road-
way or obstruction they cross, there are many exceptions. A skewed culvert is a cul-
vert that is not perpendicular to its obstruction, typically in order to better align with
the natural direction of the stream. A horizontal alignment may be curved or dog-
legged to avoid obstacles such as rock outcroppings. A broken-back culvert has bends
in its vertical alignment, and may be used to reduce exit velocities, for instance. With
a sag culvert, a portion of the culvert barrel is lower than the culvert entrance and exit.
Such culverts tend to clog, but they may be required in, for example, certain drainage
canal crossings.
Culvert hydraulics are complex, and several flow control types may exist for a single
culvert and discharge. The basic approach to hydraulic evaluation of a culvert is to
analyze its performance for each of the possible flow control types and allow the
assumption that results in the worst performance (that is, the highest headwater) to
govern the design. Two general methods of evaluation exist: (1) the graphical method,
which uses nomographs to find the headwater elevation for a particular set of condi-
tions; and (2) the method of calculating the complete flow profile, as presented in
Section 7.8. The second method is used by most culvert-analysis computer programs.
It is the more computationally intensive of the two approaches, but it is more accurate
in many situations.
Flow in a culvert can be classified as functioning under either inlet control or outlet
control. Under inlet control, the barrel of the culvert is capable of conveying a greater
discharge than the inlet will accept. For an outlet-control situation, the inlet is capable
of accepting more flow than the barrel characteristics or downstream tailwater condi-
tion will permit.
A performance curve graphically shows the relationship between discharge and head-
water elevation. For a culvert, the performance curve is constructed by plotting the
rating curves that would result for both inlet- and outlet-control assumptions. Which-
ever curve has the higher headwater elevation for a given discharge is assumed to gov-
ern the design and is used to define the performance curve. If the discharge is
increased to the point where the corresponding headwater is higher than the culvert
embankment, the resulting flow over the roadway (for example) should be reflected in
the performance curve. For hydraulic analysis purposes, the roadway is typically
assumed to be a broad-crested weir.
Culvert outlet velocities should be evaluated to determine what energy dissipation and
erosion protection measures are necessary, if any. Riprap can be used to both dissipate
energy and provide erosion protection. Another option is to modify the culvert barrel
to reduce the outlet velocity. Other protective measures include impact-type dissipa-
tors and stilling basins.
Other factors that should be considered in culvert design relate to possible environ-
mental impacts. For instance, a culvert may be an obstacle to fish passage. To mini-
mize these impacts, the culvert may be designed to maintain a certain minimum water
level and/or minimum or maximum velocity, and to simulate the natural streambed
environment.
366 Culvert Design Chapter 9
REFERENCES
American Association of State Highway and Transportation Officials (AASHTO). 1991. Model Drainage
Manual. Washington, D.C.: AASHTO.
American Society of Civil Engineers (ASCE). 1993. Design and Construction of Urban Storm Water Man-
agement Systems. ASCE Manuals and Reports on Engineering Practice, No. 77. New York: ASCE.
Brice, J . C. 1981. Stability of Relocated Stream Channels. Report No. FHWA/RD-80/158. U.S. Geological
Survey.
Brown, S. A., S. M. Stein, and J . C. Warner. 2001. Urban Drainage Design Manual. Hydraulic Engineering
Circular No. 22, 2d ed., FHWA-SA-96-078. Washington, D.C.: U.S. Department of Transportation.
Chang, F. M., and J . M. Normann. September, 1976. Design Considerations and Calculations for Fishways
Through Box Culverts, Unpublished Report, Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C.
Corry, M. L, P. L. Thompson, F. J . Watts, J . S. J ones, and D. L. Richards. 1983. Hydraulic Design of Energy
Dissipators for Culverts and Channels. Hydraulic Engineering Circular No. 14. Washington, D.C.: Fed-
eral Highway Administration, U.S. Department of Transportation.
Dyhouse, Gary, J . A. Benn, David Ford Consulting, J . Hatchett, and H. Rhee. 2003. Floodplain Modeling
Using HEC-RAS. Waterbury, Connecticut: Haestad Methods.
Harrison, L. J ., J . L. Morris, J . M. Normann, and F. L. J ohnson. 1972. Hydraulic Design of Improved Inlets
for Culverts. Hydraulic Engineering Circular No. 13. Washington, D.C.: Federal Highway Administra-
tion, U.S. Department of Transportation.
Normann, J . M., R. J . Houghtalen, and W. J . J ohnston. 2001. Hydraulic Design of Highway Culverts, 2d ed.
Washington, D.C.: Hydraulic Design Series No. 5, Federal Highway Administration (FHWA).
PROBLEMS
Note: Problems 1, 3, 4, 5, and 11 may be solved with the nomographs provided in this book. Problems 2, 8,
and 10 may be solved using the nomographs in HDS 5. Alternatively, the CulvertMaster software provided
with this book can be used.
9.1 Considering only inlet control, determine the flow in a 24-in. concrete pipe culvert with a square-
edge headwall. The headwater elevation is 6 ft. Assume Mannings n =0.013.
9.2 Considering only inlet control, what is the headwater elevation required to convey 5.7 m
3
/s in a
1200-mmcorrugated metal pipe with a mitered entrance? Assume Mannings n =0.02.
9.3 A 60-in. concrete pipe is to be used as a culvert to convey 400 cfs under a road. Calculate the reduc-
tion in headwater elevation resulting fromreplacing a square-edge inlet (K =0.5) at the headwall
with a groove-end inlet (K =0.02).
9.4 A culvert is to be designed to carry 250 cfs under a roadway. The culvert is 75 ft long, and the
upstreaminvert elevation is 100 ft. A 60-in., groove-end concrete pipe is to be installed at the natural
streambed slope of 0.01. Assume that there is a free fall at the exit (no tailwater). Determine if the
culvert is operating under inlet or outlet control and report the headwater elevation.
9.5 Repeat Problem9.4 for a culvert with a tailwater at an elevation of 4 ft above the crown of the down-
streamend.
Problems 367
9.6 A 1200-mmCMP culvert is operating under inlet control. The inlet type is a headwall, and the slope
is 0.008. Determine the flow type and the headwater depth for flows of 1.70 m
3
/s and 3.4 m
3
/s.
Assume Mannings n =0.02.
9.7 The 50-year flow at a design site is 11.3 m
3
/s. The maximumallowable headwater elevation for a
proposed culvert is 4.57 m. The culvert will have a 90 headwall with chamfers. Design the smallest
rectangular concrete barrel that will pass the peak flow without exceeding the allowable maximum
headwater. Use n =0.012. Because of structural considerations, the box culvert must be 1520 mm
high. Other pertinent design data are:
Streambed slope =0.01
Culvert length =61 m
Tailwater elevation =2.74 mabove the culvert outlet invert
9.8 Determine the outlet velocity of a 4-ftdiameter culvert conveying 150 cfs and operating under inlet
control. Use the conservative approximation method.
9.9 A performance chart for a roadway culvert is shown below. The roadway at the culvert location has a
low point of 37.2 m. The vertical alignment of the road is described by the parabolic equation y =
37.2 +1.52 10
-3
x
2
(x and y are in meters), where y is the crest elevation and x is the distance from
the location of the low point. Assume a 10-m-wide paved road surface. Calculate the total flow
through the culvert and over the roadway when the headwater elevation is 38.1 m.
9.10 Construct a performance curve for a CMP culvert with a diameter of 4 ft and a length of 150 ft. The
pipe entrance is projecting fromthe fill, and the pipe slope is 0.01. Assume that Mannings n =0.024
and that there is free flow away fromthe exit under all flow conditions. Use a flow range from0 to
200 cfs.
9.11 Construct a performance curve for a 60-in. circular concrete pipe culvert. The pipe is 100 ft long and
the slope is 0.008. The entrance is a groove end with a headwall, and Mannings n =0.013. The tail-
water is 8 ft above the invert of the exit end for all flow conditions. Use a flow range of 0 to 400 cfs.
This combination
inlet in a local
depression captures
runoff in a parking
lot.