You are on page 1of 46

21.

HYDRAULIC DESIGN OF FLOW MEASURING


STRUCTURES
John A. Replogle and Albert J. Clemmens

USDA–ARS Water Conservation Laboratory

Phoenix, Arizona

Clifford A. Pugh

US Bureau of Reclamation,

Denver, Colorado

21.1. INTRODUCTION
Experienced water providers and users can use this chapter as a quick review of hydraulic principles related to water
measurement and its relation to hydraulic design for environmental considerations.

The hydraulic design of flow measuring structures usually confronts the engineer with two opportunities. One is the design of
measurement structures in a retrofit situation and the other is in original project design. The retrofit mode is usually difficult and
requires much innovation just to obtain passable function within the space and sizing limitations and other constraints usually
imposed. Because of the increasing emphasis on quantifying flow rates and volumes in most aspects of water resource
planning and management, the retrofit applications currently dominate the design problems.

Most textbooks deal with recommending ideal installation situations and retrofit projects appear to be unable to comply
without great economic impact. This too frequently can lead to arbitrary compromises that produce poor measurement
performance. Even new installations may be limited by space requirements. This may force design decisions into the final
construction that compromise accuracy. This chapter will strive to show the design concepts available, particularly those useful
for designing both new and retrofit installations, and will point out measurement behaviors to be expected from various
compromises. This chapter suggests those deviations that cause least impact and guides the designer to choices that may be
hydraulically acceptable and still meet structural goals.

Of the numerous flowmetering methods available to the hydraulic engineer, most are based on well-established hydraulic
principles and are amenable to design manipulations of size, shape, and response. While this aspect of flow measurement is
documented in several handbooks and texts, the design and retrofit of sites to accommodate and facilitate measurement is not
as well described or is described in a scattered assortment of books and articles.

Pipeline flows of water are usually less complicated to measure than open-channel flows, most obviously because the flow
area does not change significantly with flow rate. Consequently, many applications of pipeline flows are held to stricter
accuracy standards than channel flows can reasonably achieve. Thus, channel flows and their measurement are usually limited
to large delivery volumes and to accuracies acceptable to the related activities, such as sewer flows and irrigation deliveries.

The purpose of this chapter is to consolidate design information for evaluating a flow measurement site, selecting a flow
measuring system, and adapting the measuring site to optimize measuring and other functions that may be desired from the
site. Emphasis will be on open-channel flow measurements because that is a likely need of the hydraulic engineer. Pipe
flowmeters in water supply will also be discussed in lesser detail because the major application of the many types of pipe
flowmeters is well covered in the chemical and petroleum industry literature.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Experienced readers may wish to further investigate and seek more advanced references in hydraulics and fluid mechanics.
Extensive information on fluid meter theory and detailed material for determining coefficients for tube-type meters is given in
American Society, of Mechanical Engineers (ASME) (1959, 1971) and revisited with modern updates in books by Spitzer (1990)
and Miller (1996). Brater and King (1982) have a thorough discussion of general critical depth relations and detailed
relationships for most common hydraulic flow section shapes in open channels. Bos (1989) covers a broad segment of
open=channel water measurement devices.

21.2. HYDRAULIC CONCEPTS RELATED TO WATER


MEASUREMENT
21.2.1. Basic Concepts for Pipe and Channel Flows
Flow can be classified into closed conduit flow and open-channel flow.Open-channel flow conditions occur whenever the
flowing stream has a free or unconstrained surface that is open to the atmosphere. Flows in canals or in vented pipelines that
are not flowing full are typical examples.

In hydraulics, a pipe is any closed conduit that carries water under pressure. The filled conduit may be square, rectangular, or
any other shape, but is usually round. If flow is occurring in a conduit but does not completely fill it, the flow is not considered
pipe or closed conduit flow, but is classified as open-channel flow.

Flow rate in a pipeline responds mainly to the pressure gradient or head difference that exists between two points along the
pipeline, modified by the frictional resistance to flow caused by pipe length, pipe roughness, bends, restrictions, changes in
conduit shape and size, the nature of the fluid flowing, and the cross-sectional area of the pipe.

In open-channel flows, the pressure gradient, or energy grade line, is controlled mainly by the force due to gravity, which is
influenced by the channel slope, resistance from the channel wall roughness, the channel shape, and the flow area. The fluid is
usually water.

Basic flow metering in both pipe flow and open channels depends on determining an average flow velocity by some means and
combining it with the flow cross-sectional area. For open channels, a common means involves current meter measurements
where metered point velocities are applied to their applicable subareas and summed over a flow cross section. Exceptions
include tracer-dilution techniques that do not require flow area or velocity. The uses of tracer techniques are applicable to
special pump calibrations and some difficult channel flows (mountain streams). They are avoided for most city water
distribution systems, sewer flows and irrigation applications because of the general expense with handling the equipment and
doing the analysis. The most used techniques applicable to openchannel systems, including sewer flows and irrigation canal
flow measurements, depend on exploiting the special velocity properties of critical flow, as discussed in a section 21.2.3.

Continuity equation. The first basic equation for water flowing in either pipes or channels is the continuity equation, which
simply states that discharge rate (volumetric flow per unit time), Q, is equal to flow cross-sectional area, A, times flow mean
velocity, V, through the flow cross section, or

(21.1)

Bernoulli energy equation. Another basic equation involves energy relations and is also applicable to both pipe and channel
flows. The most familiar form is for closed pipe flow, wherein the basic energy principles are described by the Bernoulli energy
equation. For two locations along a pipe at stations 1 and 2,(Fig. 21.1), the Bernoulli equation can be expressed as

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
(21.2)

where the terms are expressed in length dimensions as z = the height from an arbitrary reference plane (datum) h = the pressure
head V = average velocity through the pipe cross-section at the designated locationV2 /2g = the velocity head g = the
gravitational constant 1 ,2 = subscripts denoting the respective locations along the pipeline.

This equation is based on uniform velocity across the conduit area and no energy losses. However, in real fluid flows,
nonuniform velocities exist and friction causes energy conversion to heat. Typically, these velocities are zero at the walls and
reach a maximum profile velocity near the center of the flow. If the flow is viscous flow in a round pipe, the flow profile is
parabolic, that is, “bullet-shaped.” If the velocity is fully turbulent, the bullet-shape is much flattened, with steep velocity
gradients near the wall and nearly uniform profile across the remainder of the pipe. These idealized profiles can be skewed
drastically by regulating valves, structures, conduit bends and other flow obstructions. Therefore, application of these equations
depends on knowing, or controlling, the velocity profile so that the average velocity in the conduit cross section can be inferred.

Figure 21.1 Energy balance in pipe flow.

Equation (2.2) requires some adjustments to convert it to the energy equation, which is useful in analyzing flows in pipes or
open channels with a small slope (Chow, 1959). First we introduce correction factors, α1 and α2 , called the velocity distribution
coefficients, to account for the computational expediency of using the average velocities, V1 and V2 , to compute the kinetic
energy term, V2 /2g, at the respective locations 1 and 2 along a channel. These values for the usual range of turbulent flows in
water usually range from about 1.01 to 1.05, although for thick petroleum products in pipe flows and low velocity flows, the
value can approach a value of 2. Second, a term, hf, for the loss of energy between the two points is included. The result is

(21.3)

21.2.2. Pipe Hydraulics


Reynolds number. The behavior of flow in pipes is governed primarily by the viscosity of the fluid. In pipeline flows, the ratio
between the dynamic forces and the viscus forces is important for defining the limits between laminar and turbulent flows and
other functions of pipe flow. This ratio is called the Reynolds number, Rn, and is defined as

(21.4)

where V = the velocity of the flow, Lc = characteristic length, typically the pipe diameter, D and v = the kinematic viscosity.

Headloss characteristics in pipes. The Reynolds number, Rn, defined above, represents the effect of viscosity relative to inertia
and is used to define appropriate flow ranges for headloss equations in pipe flow. For example, headloss is proportional to the

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
square of the velocity, when the velocities and pipe size combinations defined by a pipe-diameter-based Reynolds number, Rn,
greater than about 1000. Most of the flows of interest in general hydraulic engineering have Reynolds numbers greater than
1000. Some exceptions are found in drip or trickle irrigation systems common in agricultural and urban landscape settings.

The headloss, hf, for Rn greater than the minimum value of about 1000 is traditionally expressed in terms of a friction factor, f,
the pipe diameter, D, pipe length, L, and the velocity head, V2 /2g, where g is the gravitational constant, and V is the average
velocity, as

(21.5)

The value for f is usually obtained from a Moody diagram which is a graphical representation of the f value in terms of the
Reynolds number, the roughness height of the pipe wall material, , and the pipe diameter, D. The Moody diagram is a graphical
solution of the Colebrook function

(21.6)

The values range from 0.0000015 m for smooth plastic pipe to 0.00026 m for cast iron pipe. Concrete pipe ranges from
about 0.0003 m to 0.003 m (Daugherty and Ingersoll, 1954). The equation can be readily solved by iteration techniques using a
computer spreadsheet.

21.2.3. Channel Hydraulics


Hydraulic mean depth. The hydraulic mean depth, Dm [U.S. Bureau of Reclamation (USBR), 1997] is the flow cross-sectional area,
A, divided by the flow surface width, T,or

(21.7)

For conduits such as pipes flowing nearly full, the surface flow width may be narrow, and Dm may be a larger value than the
physical water depth. For the usual natural channels and most canals, Dm is interchangeable with average depth. Sometimes it
is simply called the hydraulic depth (Chow, 1959).

Froude number. Open-channel flow behavior is governed primarily by gravity forces. The ratio of the inertial forces to the gravity
forces is called the Froude number, Fr, and is defined by

(21.8)

where V = the velocity of the flow, g = the gravitational constant, Dm = the hydraulic mean depth.

The Froude number applies to most open channel flows and is used for defining model scale ratios and estimating stable flow
characteristics in open channels.

Specific energy. It is useful to define the energy equation in terms of the local channel bottom instead of an arbitrary datum.
This is called the specific energy, E, and is given by:

(21.9)

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
That is, the specific energy is equal to the sum of the depth of flowy and the velocity head (Fig. 21.2).

Critical flow and critical depth. In open channels a flow phenomenon occurs that does not happen in closed pipe flows. The
process is called critical flow. Critical flow is defined for open=channel flows as the maximum discharge for the minimum
specific energy, that is, critical flow represents the minimum combination of potential energy (depth of flow, y) and kinetic
energy (velocity head, V2 /2g) for the given discharge (Chow, 1959). The depth of flow then is the critical depth. By virtue of the
continuity equation, for a constant discharge at critical flow, an increase in depth must necessarily be accompanied by a
decrease in velocity, which is called subcritical velocity. Conversely, a decrease in depth for the same flow rate necessarily
requires an increase in velocity, which is called supercritical velocity.

When critical flow occurs in an open channel it can be shown (Chow, 1959) that

(21.10)

where Vc mean flow velocity, g = gravitational constant, Dm = hydraulic mean depth, and α = velocity distribution coefficient
(Chow, 1959).

This can further be combined with the continuity equation, Eq. (21.4), to express the critical flow discharge rate, Qc, as

(21.11)

Figure 21.2 Specific energy balance.

In practice, the water surface slope in a contraction is relatively steep and the precise plane of the critical flow section is not
easily or reliably located. Thus, the data for accurately evaluating the hydraulic depth, Dm , is not readily obtained. For critical
flow flumes, the flow depth is therefore not measured at this critical section, but instead a depth is measured in the upstream
channel, where the velocity head is computable or is minimal. The critical depth is then mathematically derived based on energy
principles described by Bos et al. (1991). These flumes, sometimes referred to as the computable flumes that rely on critical
flow theory, will be discussed in more detail in Section 21.7.

For maximum discharge for minimum energy, the condition described above for critical flow in open channels, it can be shown
that

(21.12)

where: A = the channel flow area, T = the top width of the channel flow, Dm = the hydraulic mean depth, and Vc = the critical
velocity.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Thus, the velocity head at critical flow is equal to half the hydraulic mean depth, sometimes calledhydraulic depth, Dm = A/T
(Chow, 1959).

From the above,

(21.13)

where Fr is the Froude number defined above. Thus, at critical flow the Froude number is unity. Also note that the Froude
number can be defined by Fr = V/Vc for velocities other than critical.

Normal depth. Yet another depth is associated with open channel flows, thenormal depth. When the flow in an open channel
does not change from station to station, the flow is said to be uniform and the bottom slope, the hydraulic grade line, and the
energy grade line are all parallel to each other. Figure 21.2 shows the condition when the flow is not uniform.

Modular limit. If the downstream depth in a channel is too deep, the backwater will prevent critical depth from occurring. The
flow is considered to be submerged whenever the downstream water surface exceeds the crest elevation of a channel control,
such as a weir or flume. For flumes, particularly, this submergence has little effect on critical depth, and free flow exists until a
certain limiting submergence for that particular flow module called the modular limit is reached. At some point of
submergence, the upstream flow depth is affected. and the modular limit is exceeded, and free flow does not occur. The
modular limit is defined as that limiting submergence ratio, and is based on the ratio of the downstream depth to upstream
depth. The modular limit occurs when the downstream backwater causes more than1 percent change in the calculated
discharge in a particular flow module, or device (Bos, 1989). When the modular limit is exceeded, the flow is called nonmodular.

21.2.4. Energy Balance Relationships in Channels


Hydraulic problems concerning fluid flow are commonly described in terms of conserving kinetic and potential energy, and are
conveniently expressed using the classical Bernoulli equation in combination with the Continuity equation. The applications of
these equations are generally well documented, particularly for pipe flows, in texts and handbooks and are not repeated here
(Brater and King, 1982; Miller, 1996). The case for open channels is less complete, but is given considerable treatment in Brater
and King (1982), Chow (1959), and Herschy (1985). The computational uncertainties evolve from the effects of friction and
viscosity that distort the classic assumptions of a uniform velocity profile across the fluid stream. When accountings for friction
and flow profile are successfully applied, the results for discharge computations are usually good to excellent for both pipes
and open channels (Bos et al., 1991).

Headloss characteristics in channels. In terms of frictional headlosses, the wetted perimeter, Pw, of the flow is important.
Hydraulic radius, Rh, is defined as the area of the flow section, A, divided by the wetted perimeter, Pw, or

(21.14)

Conversely, the wetted perimeter times the hydraulic radius is equal to the area of an irregular flow section. The hydraulic
radius of a channel can be compared to the radius of a pipe, r, with a cross-sectional area A = πr2 and a circumference or wetted
perimeter Pw = 2 πr. Under these conditions, the hydraulic radius compares to the pipe radius, and to the pipe diameter,D, as

(21.15)

The Manning’s formula. Canal and stream discharge rates are usually estimated with use of the Manning’s formula. Many open-
channel flow equations have been proposed, but the most used is the Manning’s formula. This expression is partly rational and

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
uses an empirical coefficient, n, that is used in both the SI and American unit systems. In general form it becomes

(21.16)

where V = average velocity, n = the Manning’s roughness coefficient, Rh = the hydraulic radius, S e = the energy-line slope, and Cm
= conversion of units: 1.0 for metric units and 1.486 for American units.

The factor S e is the slope of the energy line. Note that the bed slope of the channel,S o , and the slope of the water surface, S w,
are not to be used. These parameters are, however, equal to Se when uniform flow, with the resulting normal depth occurs. As
defined above, normal depth occurs when a channel flow approaches uniformity from station to station along the channel
(Chow, 1959).

For design purposes, the n value for concrete lined canals is usually about 0.014. A good finish can lower it to 0.012, while
concrete in poor condition and channels constructed with shot crete or gunite, usually have n values from 0.016 to 0.018. In
some instances, concrete lined canals, with significant algae growth, have experienced n values as high as 0.032. This latter
value approaches the values usually experienced with unlined channels, 0.03–0.04. Thus, for reliable application, the use of
Manning’s formula requires field experience and on-site inspection of the channel being computed.

21.2.5. Modeling Characteristics for Open Channels


For flowing water in open channels, fluid friction is a factor as well as gravity and inertia. This would seem to present a problem
for hydraulic scale modeling, because both dynamic and kinematic similarity are difficult to achieve simultaneously. Fortunately
for most open-channel flows, there is usually fully developed turbulence. Thus, the fluid friction losses are nearly proportional to
V2 , and are nearly independent of Reynolds number, Rn, with rare exceptions.

This means that in open–channel flows, inertia and gravity forces dominate over viscous forces (associated with pipe flows)
and are a function of the Froude number, Fn, alone. Geometric similarity between a model and a prototype then provides
kinematic similarity. For kinematic similarity the ratios of the respective velocities are everywhere the same. The velocity ratio,
Vr, is the velocity in the prototype, Vp, divided by the velocity in the model, Vm , or

(21.17)

For Froude modeling, and from the definition for Fn, we note that V is proportional to the square-root of a length,L (for open
channels we used the hydraulic depth, Dm ) with the gravitational constant, g, assumed to be constant. Thus, the above equation
can be written as

(21.18)

where Lr = the length ratio between prototype and model dimensions, Lp :Lm

Because the velocity varies as and the cross-sectional area as L2 r it follows that

(21.19)

or

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
(21.20)

This equation is valid when all the physical structure dimensions and the heads are of the same ratio. For example, it can be
used to convert a flume rating for one size to that of a similar flume of another size. Scale modeling works best for determining
calibrations in a range of Lp:Lm less than about 10:1, although ranges exceeding 50:1 have sometimes been used for studying
special situations.

21.3. BASIC PRINCIPLES OF WATER MEASUREMENT


Flow is usually measured by determining an average flow velocity and using the flow area to compute the volume discharge.
Flow meters then have the function of detecting this velocity and combining it with the physical information of the conduit to
produce a useable readout. This is easily demonstrated for closed conduits. Propeller meters, ultrasonic meters, laser-Doppler
velocimeters, electromagnetic meters, Venturi meters, and orifice meters all are based on inferring a basic velocity
measurement applied to a flow area for a discharge rate.

For open channels, many flumes depend on determining the velocity based on energy principles of critical flow. Weirs are
usually described in terms of orifice flow integrated over the weir width and the crest depth. Again these are basically velocity
expressions for flow through a defined area.

Dilution techniques applicable to both closed pipe and open channel flows depend on detecting the amount of fluid added to a
known starting amount of tracer material. The dilution ratio determines the discharge ratio, in the case of constant injection of a
tracer. The tracer may be a chemical or even injected heat or heated fluid.

Electromagnetic meters depend on generating voltages by flowing a conductive fluid, usually water, through a magnetic field to
produce a velocity indication.

21.3.1. Water Meter Classification


Flow measuring devices are commonly classified into those that arerate meters and measure discharge rate as the primary
reported indication and those that are quantity meters and measure volume as the primary indication. The latter include
weighing tanks and batch volume tanks and are used mostly in laboratory settings as flow rate standards.

Devices in either of these broad classes can again be divided according to the physical principle that is used to detect that
primary indication (ASME, 1959). The meter part that interacts with the flow to produce the primary indication is referred to as
the primary device. This interaction exploits one or more of a few physical principles, such as pressure force, energy conversion,
weight, electrical properties, mixing properties, sonic properties, and so on, to generate a signal. Primary devices are thus
limited in number and variety. Secondary devices convert the primary interaction into useable readout. These secondary devices
are numerous and relatively unlimited in configuration and variety. The function of one class can be converted into the
response of the other with suitable secondary devices.

Some water measuring devices particularly suitable to municipal water supply, waste-water treatment, agricultural irrigation,
and drainage applications are the historical rate meters that are treated in most hydraulic text. These include (1) weirs, (2)
flumes, (3) orifice meters, and (4) Venturi meters.

Head, h, or upstream depth, commonly is used for the open channel devices such as flumes and weirs. Either pressure,p, head,
h, or differential head, Δh, or differential pressure, Δp, is used with tube-type devices, such as Venturi meters and orifice meters.

Venturi meters in pipelines and long throated flumes in open-channel flows are examples where the energy principles and the

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
flow accountings mentioned above give good to excellent computational results with minor dependency on empirical
coefficients (Bos, 1989; Bos et al., 1991).

21.3.2. Installation Requirements


Special difficulties arise in applying velocity profile and friction accountings when insufficient pipe or channel exists upstream
from a flow measuring device. This is needed to ensure that predictable and acceptable velocity profiles are presented to the
meter. Frequently pipe or channel lengths can be significantly shortened by special structural flow conditioners. These
structural measures then become a design option. Some of these are discussed below and in section 21.3.3

Designs for pipe discharges are well described in textbooks and in standard handbooks. The design difficulties center around
selecting appropriate metering candidates for accomplishing the measuring function and in providing an appropriate
environment for economical, accurate, and serviceable operation.

In the case of pipe flows, recommended straight pipe lengths, in terms of pipe diameter, are to be provided upstream of the
meter to assure reasonable operating accuracy. These lengths depend on the flow pattern presented to the meter primarily
caused by valves and pipe elbows upstream from the meter. The number and orientation of elbows greatly influence the
circulation patterns and flow profile distortions presented to the meter.

Open-channel flow water measurement generally requires that the Froude number of the approach flow be less than 0.5 to
prevent wave action that would hinder or possibly prevent an accurate head determination.

Energy concepts are used to describe Venturi meters in pipe flows based on theBernoulli equation in which part of the pipe
forms a contracted throat that necessarily changes the flow velocity and hence converts some of the static pressure to velocity
head. The decrease in static pressure is the basis for flow detection. A similar concept can be applied to open channels. A
historical version is the so–called Venturi flume (Brater and King, 1982) that detects the change in water surface elevation
between an upstream station and in a contracted section. However, this small change is difficult to accurately detect, so the
direct concept is not used. Rather, contractions are designed to be severe enough to force critical flow velocities in the
contracted section. Thus, only an upstream head is needed to define the flow energy and flow area which can be converted to
discharge rate. These are generally called critical-flow flumes. The flow condition where only one head measurement is needed
is called free flow.

The critical-flow flumes themselves consist of those called long-throated flumes that force parallel flow in the contracted, or
control, section, called the throat, and those that have curvilinear flow in the throat and are calledshort-throated flumes. The
limiting throat control section is the sharp-crested weir consisting of a thin plate. Thus, for flumes and weirs one unique head
value exists for each discharge, simplifying the calibration procedure.

However, if the downstream flow level submerges critical depth enough to affect the upstream reading, the modular limit is
exceeded, and free flow does not occur. When exceeded, separate calibrations at many levels of submergence are then
required, and two head measurements are needed to measure flow. This condition generally is to be avoided in meter site
design because it reduces the accuracy of the measurement and increases the difficulty of flow determination. The modular
limit for sharp-crested weirs, in practice, is less than zero, requiring full clearance of the overfall nappe of at least 3 cm, while
short-throated flumes can usually tolerate 65 percent to 70 percent submergence. Longthroated flumes can tolerate from 70
percent to 90 percent depending on flow conditions and flume size.

Designing flumes for submerged flow beyond the modular limit decreases the accuracy of the flow measurement. Sometimes
flumes and weirs can be overly submerged unintentionally by poor design, construction errors, structural settling, attempts to
supply increased delivery needs with increasing downstream heads, accumulated sediment deposits, or weed growths.
Sometimes use of the submerged range beyond the modular limit is an economic compromise.

Approach flow conditions for pipes. Water measurement devices are generally calibrated with certain approach flow conditions.
The same approach conditions must be attained in field applications of measuring devices. Poor flow conditions in the area

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
just upstream of the measuring device can cause large discharge indication errors. For open channels, the approaching flow
should generally be subcritical. The flow should be fully developed, mild in slope, and free of curves, projections, and waves.

Pipeline meters commonly require 10 or more diameters of straight pipe approach. Fittings and combinations of fittings, such
as valves and bends, located upstream from a flow meter can increase the number of required approach diameters. Several
references (ASME, 1971; ISO, 1991) give requirements for many pipeline configurations and meters. These are discussed in
detail by Miller (1996).

Flow conditioning options. Many installations, especially in retrofit situations, do not provide for sufficient lengths of straight
pipe to remove velocity profile distortions and swirl to an acceptable level. Therefore, the designer may need to use flow
conditioners in combination with straight pipe lengths. Swirl sensitivity varies widely. Some meters are particularly sensitive to
swirl, such as the propeller and turbine meters. Magnetic flow meters are somewhat less sensitive to radial velocities than
single-path ultrasonic flow meters. Venturi meters are less sensitive than orifice meters. For a swirl angle of 20° , the discharge
coefficient changes by about –1 percent for a Venturi meter with β = 0.32 (β is the ratio of meter throat diameter to the pipe
diameter) and about +10 percent for a similar orifice. Thus a swirl can increase the discharge through an orifice for the same
differential head reading (Miller, 1996).

In pipeline flows, contractions can produce a central jet and also increase an incoming swirl, while expansions tend to slow
swirls and produce enough secondary flow to restore flow profiles to some semblance of acceptability. These characteristics
can modify the straight pipe lengths needed or the type of flow conditioner to recommend (Miller, 1996). Rough pipes also tend
to reduce a swirl.

For flows, such as that encountered in sewage discharges and irrigation pipeline deliveries that originate from open channels,
many of the tube-bundle types of flow conditioners can gather trash and cause maintenance problems. Many meter providers in
these situations use fins or vanes that protrude from the wall and have sloped upstream edges that shed trash. The vanes
protrude about one-fourth of the pipe diameter into the flow, leaving the center core of the flow open. While these vanes can
vary in number and length, the logic being that the fewer the vanes the longer they should be in the direction of flow, common
configurations are four vanes that are about two or three pipe diameters long. Vanes in themselves do not condition wall jets
well. Field experience, has shown that troublesome flow profiles can be conditioned significantly by inserting an orifice into the
pipe. The orifice diameter is about 90 percent of the pipe diameter and is used to control wall jets and force them to mix with
the general flow. The orifice in itself tends to cross-mix the jets and would appear to reduce spin. However, if the jets are
symmetrical and an initial swirl exists, orifices tend to increase the swirl. Inserting an orifice appears to be supported by recent
recommendations of Miller (1996) where it is stated: “To achieve a fully developed profile, it is important that the flow be
blocked or restricted close to the wall, with the central core having the larger flow area.”

The addition of vanes when space permits is recommended. Because orifices, in general, tend to force the flow to the pipe
center while increasing spin, it appears best to place the vanes upstream from the orifice. If they are placed downstream, the
spin not only may be increased, but the spinning central flow may not be touched by the vanes.

21.3.3. Examples of Flow Conditioning in Field Situations


Flow conditioning in an irrigation delivery pipeline. As mentioned previously, measuring devices frequently must be installed in
flow situations that are less than optimal. A field example occurred in Arizona where a large pipe was used as an outlet to a
secondary canal and a single-path ultrasonic meter placed in it was subjected to flow profile distortions. The pipe was about
0.75 m in diameter and delivered approximately 400 L/s. The flow rate readout was unstable, with fluctuations varying by about
15 percent. The problem appeared to be caused by slowly spiraling flow induced by the bottom jet from a partly open pipe inlet
gate and a 45° elbow. This is similar to two closely spaced pipe elbows that are not in the same plane, which can cause a spiral
flow pattern (ASME, 1971).

A successful attempt to modify the jet and cause it to cross mix so that the jet effects and the strength of the spiral flow were
reduced, was accomplished by inserting a large β-ratio orifice in the pipe (Fig. 21.3). This consisted of an annular metal ring

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
with the outside radius approximately that of the pipe and an inside diameter about 10 percent less, or an orifice with β = 90
percent. The orifice was installed about three diameters downstream from the elbow. The slight increase in headloss was
compensated by increasing the upstream gate opening. The orifice can be constructed by cutting notches from an
appropriately sized piece of angle iron or aluminum and bending it to a polygon that approximates the circle diameter of the
pipe interior. Some leakage around the ring is acceptable. For propeller meters, additional vanes projecting from the walls may
be needed to further reduce spiral flow. These vanes would be placed upstream from the orifice. In this installation, the
fluctuation was reduced to within about 3 percent.

Figure 21.3 An orifice plate with a large opening is used to condition a flow profile.

Flow conditioning in channels. By analogy and using a minimum of 10 pipe diameters of a straight approach channel, open
channel flow would require 40 hydraulic radii of straight, unobstructed, unaltered approach, based on the calculation of
hydraulic radius for circular pipes being equal to one-fourth the pipe diameter, (Eq. 21.15). This would translate for very wide
channels into approximately 40 times the flow depth. For narrow channels that are as deep as they are wide, this would
compute to be about 13 channel depths or top widths.

Other recommendations on approach channel criteria are presented by Bos (1989) and USBR (1997). Major features of that
criteria follow:

If the control width is greater than 50 percent of the approach channel width, 10 average approach flow widths of straight,
unobstructed approach are required.

If the control width is less than 50 percent of the approach width, 20 control widths of straight, unobstructed approach are
required.

If upstream flow is below critical depth, a jump should be forced to occur. In this case, 30 measuring heads of straight,
unobstructed approach after the jump should be provided.

If baffles are used to correct and smooth out approach flow, then 10 measuring heads (10h1 ) should be placed between the
baffles and the measuring station.

Approach flow conditions should be continually checked for deviation from these conditions as described in Bos (1989) and
USBR (1997).

The baffles described above can become unacceptable maintenance problems in open channels. Some field expediencies are
therefore described that have been found to work in specific instances, but have not been studied for assured design
generalizations. Nevertheless, these constructions are but small extensions to currently accepted practices in pipe flows.

Applications for open=channel flow conditioners include abrupt channel turns, sluice gate outflows, and channels downstream
from a hydraulic jump. The abrupt turns may benefit from floor and wall mounted vanes or fins. Based on pipe flow experience,
and assuming the channel is half of a closed conduit, these fins or vanes would probably be about 10 percent to 15 percent of
the channel depth.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
As in pipe flow, wall jets that can develop downstream from sluice gates appear to need treatment. This can be in the form of a
structural angle bolted on the channel floor and up the walls. Suggested size, based on the pipe flow analogy, is for the angle to
be about 5 percent into the channel flow depth. Whether the sidewalls need larger angles when the channels are wide has not
been tested.

21.3.4. Wave Suppression


Of special concern in open channels is wave suppression downstream from a sluice gate, hydraulic jump, or an abrupt turn.
Thus, the flow conditioners in channels have the additional task not present in pipe flows of surface wave suppression.
Excessive waves in irrigation canals make reading sidewall gages difficult. These waves are usually caused by a jet entry from
a sluice gate or by a waterfall situation. The unstable surface can be 10–20 cm high and extend for tens of meters
downstream.

Wave suppression in canals. A surface wave suppressor was tested by Schuster as reported in USBR (1997). It basically was a
constructed roof over the canal for a distance equal to about four times the flow depth. The roof structure is inserted into the
flow about one-third the flow depth. All flow is forced to pass under the structure. Wave suppression is between 60 percent and
93 percent (Fig. 21.4). For canals that usually flow at one level, this wave-suppression method is appropriate. The wave
suppressor shown in Figure 21.4 has been successfully used in both large and small channels (USBR, 1997). An important
aspect is that the structure is fixed and not allowed to float. Floating suppressors are not effective.

Figure 21.4 Wave suppresor design (From USBR, 1997).

Successful field applications of wave suppressors include some installations in trapezoidal irrigation channels, with 1:1 side
slopes and 60-cm bottom width. They were flowing about 400 L/S at about 45 cm deep. While the velocity was not high, about
0.8 m/s, the agitation from a flow entry gate was producing waves about 15 cm high. The suppressor “roof” was only about 60
cm in the direction of flow, and penetrated the flow by about 15 percent.

Another version that has worked in small channels is illustrated inFigure 21.5. This can work with a single cross–member if
the flow is usually at a fixed discharge rate and becomes similar to the suppressor described above. In severe jet cases an
additional floor sill, about 10 percent of the flow depth in height, has been used successfully.

The length of the roof in the flow direction has not been well studied, but field observations seem to support a length greater
than two lengths of the surface wave, if that can be estimated, otherwise, use two to four times the maximum flow depth as
described above.

To suppress waves in canals that do not always flow at the same depth, a staggered set of baffles may help (Replogle, 1997).
Because these will be submerged part of the time, they must have a thickness that overlaps slightly to accommodate the
vertical depth of interest. To avoid obstructing the channel severely, these baffles probably should not obstruct more than about
20 percent of the channel at any particular location. Staggering them as shown in Fig. 21.5 would accomplish this without
excessive obstruction. Rounding the upstream edges will help shed trash, but may be less effective in suppressing waves.
Observe in the sequence of drawings in Fig. 21.5 that the staggering is upward in the downstream direction. Note that the next
baffle slightly overlaps the horizontal flow lines so that flow passing over the top of one baffle is not allowed to free-fall and
start another wave. Fig. 21.5 a–c illustrate, the general behavior as the flow becomes less deep.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Figure 21.5 Wave suppressor for variable-depth flows in a canal. (From Replogle, 1997)

21.4. MEASUREMENT ACCURACY


Accurate application of water measuring devices generally depends upon standard designs or careful selection of devices,
careful fabrication and installation, good calibration data and adequate analysis. Also needed is proper user operation with
appropriate inspection and maintenance procedures. During operation, accuracy requires continual verification that all
measuring systems, including the operators, are functioning properly. Thus, good training and supervision are required to attain
measurements within prescribed accuracy bounds. Accuracy is the degree of conformance of a measurement to a standard or
true value. The standards are selected by users, providers, governments, or compacts between these entities. All parts of a
measuring system, including the user, need to be considered in accessing the system's total accuracy.

As mentioned above, a measurement system usually consists of a primary element, which is that part of the system that
creates what is sensed, and is measured by a secondary element. For example, weirs and flumes are primary elements. A staff
gage is a secondary element.

Designers, purchasers, and users of water measurement devices generally rely on standard designs and manufacturers to
provide calibrations and assurances of accuracy. A few water users and providers have the facilities to check the condition and
accuracy of flow measuring devices. These facilities have comparison flow meters and/or volumetric tanks for checking their
flow meters. These test systems are used to check devices for compliance with specification and to determine maintenance
needs. However, maintaining facilities such as these is not generally practical.

Various disciplines and organizations do not fully agree on some of the definitions related to measuring device specifications,
calibration, and error analysis. Therefore, it is important to verify that a clear and mutual understanding of the specifications,
calibration terminology, and the error analysis processes is established when discussing these topics with others.

21.4.1. Definitions of Terms Related to Accuracy


Error. Error is the deviation of a measurement, observation, or calculation from the truth. The deviation can be small and
inherent in the structure and functioning of the system and be within the bounds or limits specified. Lack of care and mistakes
during fabrication, installation, and use can often cause large errors well outside expected performance bounds. Because the
true value is seldom known, some investigators prefer to use the term uncertainty. Uncertainty describes the possible error or
range of error which may exist. Investigators often classify errors and uncertainties into spurious, systematic, and random
types.

Precision. Precision is the ability to produce the same measurement value within given accuracy bounds when successive
readings of a specific quantity are measured. Precision represents the maximum departure of all readings from the mean value
of the readings. Thus, a single observation of a measurement cannot be more accurate than the inherent precision of the

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
combined primary and secondary precision. It is possible to have good precision of an inaccurate reading. Thus, precision and
accuracy differ.

Spurious errors. Spurious errors are commonly caused by accident, resulting in false data. Misreading and intermittent
mechanical malfunctions can cause discharge readings well outside of expected random statistical distribution about the
mean. Spurious errors can be minimized by good supervision, maintenance, inspection, and training. Experienced, well-trained
operators are more likely to recognize readings that are significantly out of the expected range of deviation. Unexpected spiral
flow and blockages of flow in the approach or in the device itself can cause spurious errors. Repeating measurements does not
provide information on spurious error unless repetitions occur before and after the introduction of the error. On a statistical
basis, spurious errors confound evaluation of accuracy performance.

Systematic errors. Systematic errors are errors that persist and cannot be considered random. Systematic errors are caused by
deviations from standard device dimensions, anomalies to the particular installation, and possible bias in the calibration.
Systematic errors cannot be removed or detected by repeated measurements. They usually cause persistent error on one side
of the true value. The value of a particular systematic error for a particular device may sometimes be considered as a random
error. For example, an installation error in the zero setting for a flume might be + 1 mm for one flume and −2 mm for another.
For each flume the error is systematic, but for a number of flumes it would be a random error.

Random errors. Random errors are caused by such things as the estimating required between the smallest division on a head
measurement device and water surface waves at a head measuring device. Loose linkages between parts of flowmeters
provide room for random movement of parts relative to each other, causing subsequent random output errors. Repeated
readings decrease the average expected error resulting from random errors by a factor of the square root of the number of
readings.

Total error. Total error of a measurement is the result of systematic and random errors caused by component parts and factors
related to the entire system. Sometimes, error limits of all component factors are well known. In this case, total limits of simpler
systems can be determined by computation (Bos et al., 1991). In more complicated cases, it may be difficult to confidently
combine the limits. In this case, a thorough calibration of the entire system as a unit can resolve the difference. In any case, it is
better to do error analysis with data where entire system parts are operating simultaneously and compare discharge
measurement against an adequate discharge comparison standard.

Expression of errors. Instrument errors are usually expressed by manufacturers as either a percent of reading or a percent of full
scale. The secondary devices based on electronic outputs are more frequently expressed in terms of percent full scale. The
designer must be aware that a probable error value of say ±1 percent full-scale can exceed ±10 percent for small value readings
on the output device. When used with weirs, for example, the head reading of h1.5 in the weir equation can increase this 10
percent head measurement error to a 15 percent flow measurement error.

Terms Related to Measurement Capability

Linearity. Linearity usually means the maximum deviation in tracking a linearly varying quantity, such as measuring head, and is
generally expressed as percent of full scale.

Discrimination. Discrimination is the number of decimals to which the measuring system can be read. Precision is no better than
the discrimination.

Repeatability. Repeatability is the ability to reproduce the same reading for the same quantities. Thus, it is related to precision.

Sensitivity. Sensitivity is the ratio of the change of a secondary measurement, such as head, to the corresponding change of
discharge.

Range and Rangeability. Range is fully defined by the lowest and highest value that the device can measure without damage and
comply within a specified accuracy. The upper and lower range bounds may be the result of mechanical limitations, such as
friction at the lower end of the range and possible overdriving damage at the higher end of the range. Range can be designated

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
in other ways: (1) as a simple difference between maximum discharge (Qmax) and minimum discharge (Qmin), (2) as the ratio
(Qmax/Qmin), called rangeability, and (3) as a ratio expressed as 1:(Qmin/Qmax). Neither the difference nor the ratios fully define
range without knowledge of either the minimum or maximum discharge.

Additional terms (hysteresis, response, lag, rise time). Additional terms related more to dynamic variability might be important
when continuous records are needed or if the measurements are being sensed for automatic control of canals and irrigation.
Hysteresis is the maximum difference between measurement readings of a quantity established by the same mechanical set
point when set from a value above and reset from a value below. Hysteresis can continually get worse as wear of parts
increases friction or as linkage freedom increases. Response has several definitions in the instrumentation and measurement
field. For water measurement, one definition for response is the smallest change that can be sensed and displayed as a
significant measurement. Lag is the time difference of an output reading when tracking a continuously changing quantity.Rise
time is often expressed in the form of the time constant, defined as the time for an output of the secondary element to achieve
63 percent of a step change of the input quantity from the primary element.

21.4.3. Comparison Standards


Water providers may want, or may be required, to have well-developed measurement programs that are highly managed and
standardized. If so, water delivery managers may wish to consult American Society for Testing Materials Standards (ASTM,
1988), Bos (1989), International Organization for Standardization (ISO, 1983: ISO, 1991), and the National Handbook of
Recommended Methods for Water Data Acquisition (USGS, 1980).

Research laboratories, organizations, and manufacturers that certify measurement devices may need to trace accuracy of
measurement through a hierarchy of increasingly rigid standards.

The lowest standards in the entire hierarchy of physical comparison standards are called working standards, which are shop or
field standards used to control quality of production and measurement. These standards might be gage blocks or rules used to
ensure proper dimensions of flumes during manufacturing or devices carried by water providers and users to check the
condition of water measurement devices and the quality of their output. Other possible working standards are weights, volume
containers, and stopwatches. More complicated devices are used, such as surveyors’ levels, to check weir staff gage zeros.
Dead weight testers and electronic standards are needed to check and maintain more sophisticated and complicated
measuring devices, such as acoustic flow meters and devices that use pressure cells to measure head.

For further measurement assurance and periodic checking, water users and organizations may keep secondary standards.
Secondary standards are used to maintain integrity and performance of working standards. These secondary standards can be
sent to government laboratories, one of which is the National Bureau of Standards in Washington, D.C., to be periodically
certified after calibration or comparison with accurate replicas of primary standards. Primary standards are defined by
international agreement and maintained at the International Bureau of Weights and Measurements in Paris, France.

Depending on accuracy needs, each organization should trace their measurement performance up to and through the
appropriate level of standards. For example, turbine acceptance testing, such as in the petroleum industry, might justify tracing
to the primary standards level.

21.5. SELECTION OF PRIMARY ELEMENTS OF WATER


MEASURING DEVICES
21.5.1. General Requirements
Design considerations involve the selection of the proper water measurement device for a particular site or situation. Site-

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
specific factors and variables must be considered in extended detail. Each system has unique operational requirements and
installation concerns. Knowledge of the immediate measurement needs and reliable estimates on future demands of the
proposed system is advantageous. Possible selection constraints may be imposed by laws and compact agreements and
should be consulted before selecting a measurement device. Contractual agreements for the purchase of pumps, turbines, and
water measuring devices for water supply, sewage and drainage districts often dictate the measurement system required for
compliance prior to payment. These constraints may be in terms of accuracy, specific comparison devices, and procedures.
Bos (1989) provides an extensive and practical discussion on the selection of open channel water measurement devices. Miller
(1996) provides a recent compilation of selection criteria for pipe flow-meters suited to liquids and steam and other gas flows.
Bos (1989) provides a selection flow chart and a table of water measurement device properties to guide the selection process
for the open channel devices. Miller (1996) describes each meter in detail for the pipe systems, but is more general in leaving
the selection to the designer.

Because the design engineers for civil engineering projects are most likely to be dealing with irrigation water supply, waste
water, or drainage and flood flows, the emphasis is placed on the measuring systems deemed most appropriate to these
processes. Large closed-pipe systems for water supplies are frequently encountered, so installation situations appropriate to
these will also be included. Gas flows, including steam, are more likely to be encountered by mechanical and chemical
engineers and those readers are referred to Miller (1996) and ASME (1959, 1971).

21.5.2. Types of Measuring Devices


System operators for water supply, drainage, and waste water commonly use many types of standard water measuring devices,
usually in open channels with limited applications in closed conduits. Particularly prominent uses of open channel devices are
found in irrigation delivery systems and farm distribution systems, although these measuring devices are frequently used for
sewer flows and even flood flows. However, the latter two areas of application are frequently more difficult because of the
likelihood of heavy bed loads and floating debris.

In pipe flowmeters, the most commonly installed devices in industry are the orifice meters, accounting for up to 80 percent of
all industrial meters (Miller, 1996). Venturi meters and flow tubes provide much of the remainder. In absolute numbers, the
household meters, based on various technologies from nutating disks to paddle-wheel turbines, dominate.

For open–channel flows, weirs, flumes, submerged and free orifices, and current meters dominate the flow measuring
methods. Pipe flow meters, propeller and turbine, acoustic, magnetic, and vortex-shedding meters are used on large water
supply wells such as those used in irrigation and municipal water supply. Differential head meters, such as orifice meters,
Venturi meters, and flow tubes, are also used in these applications. The meters considered herein are

1. Open-channel flow devices

a. Current metering (cup, propeller, and electromagnetic probes)

b. Weirs

c. Flumes

d. Acoustic (transonic and Doppler)

e. Tracers

f. Miscellaneous

2. Pipe flow devices

a. Differential head meters

b. Acoustic (transonic and Doppler)

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
c. Tracers

d. Turbine/propeller/other insert mechanical

e. Vortex-shedding

f. Miscellaneous

The main factors which influence the selection of a measuring device include (USBR, 1997):

a. Accuracy requirements

b. Cost

c. Legal constraints

d. Range of flow rates

e. Head loss

f. Adaptability to site conditions

g. Adaptability to variable operating conditions

h. Type of measurements and records needed

I. Operating requirements

j. Ability to pass sediment and debris

k. Longevity of device for given environment

l. Maintenance requirements

m. Construction and installation requirements

n. Device standardization and calibration

o. Field verification, troubleshooting, and repair

p. User acceptance of new methods

q. Vandalism potential

r. Impact on environment

Accuracy requirements. The desired accuracy of the measurement system is an important consideration in the selection of a
measurement method. Most water measurement installations, including the primary and secondary devices, can produce
accuracies of ±5 percent. Some systems are capable of ±1 percent under laboratory settings. However, in the field, maintaining
such accuracies usually requires considerable expense or special effort in terms of construction, secondary equipment,
calibration in-place, and stringent maintenance. Selecting a device that is not appropriate for the site conditions can result in a
nonstandard installation of reduced accuracy, sometimes exceeding ±10 percent.

Accuracies are frequently reported that relate only to the primary measurement method or device. However, many methods
require secondary measurement equipment that produces the actual readout. This readout equipment typically increases the
overall error of the measurement.

Cost. The cost of the measurement method includes the cost of the device itself, the installation, secondary devices, operation,
and maintenance. Measurement methods vary widely in their cost and in their serviceable life span. Measurement methods are
often selected based on the initial cost of the primary device with insufficient regard for the additional costs associated with
providing the desired records of flows over an extended period of time.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Legal constraints. Governmental or administrative water board requirements may dictate the water measurement devices or
methods. Water measurement devices that become a standard in one geographic area may not necessarily be accepted as a
standard elsewhere. In this sense, the term “standard” does not necessarily signify accuracy or broad legal acceptance. Many
water agencies require certain water measurement devices used within their jurisdiction to conform to their standard for the
purpose of simplifying operation, employee training, and maintenance.

Flow range. Many measurement methods have a limited range of flow conditions for which they are applicable. This range is
usually related to the need for certain prescribed flow conditions which are assumed in the development of calibrations. Large
errors in measurement can occur when the flow is not within this range. For example, using a bucket and stopwatch for large
flows that engulf the bucket is not very accurate. Similarly, sharp-edged devices, such as sharp=crested weirs, typically do not
yield good results with large channel flows. These are measured better with large flumes or broad-crested weirs, which in turn
are not appropriate for trickle flows.

Certain applications have typical flow ranges. Irrigation supply monitoring seldom demands a low-flow-to-high-flow range above
about 30, while this range on natural stream flows may exceed 1000.

In some cases, secondary devices can limit the practical range of flow rates. For example, with devices requiring a head
measurement, the accuracy of the head measurement from a visually read wall gage may limit the measurement of low flow
rates. For some devices, accuracy is based on percent of the full-scale value. While the resulting error may be well within
acceptable limits for full flow, at low flows, the resulting error may become excessive, limiting the usefulness of such
measurements. Generally, the device should be selected to cover the desired range. Choosing a device that can handle an
unnecessary large flow rate may result in compromising measurement capability at low flow rates, and vice versa. This choice
depends on the objective of the measurement. For example, in irrigation practice, usually choose a device that can measure the
most common flow range at the expense of poorly measuring extremes, such as flood flows. For urban drainage, the flood peak
may be important.

For practical reasons, different accuracy requirements for high and low flows may be chosen. This is reasonable when an
annual total is the primary goal and the low flows contribute a small percentage to that total. Also, if the inaccurate low flow
readings are truly a random error then this error approaches zero with large accumulations of readings. Thus the designer
needs to know if management decisions are made from individual readings or from long term averages.

Headloss. Most water measurement devices require a drop in head. On retrofit installations, for example, to an existing
irrigation project, such additional head may not be available, especially in areas that have relatively flat topography. On new
projects, incorporating additional headloss into the design can usually be accomplished at reasonable cost. However, a tradeoff
usually exists between the cost of the device and the amount of headloss. For example, acoustic flow meters are expensive but
require little headloss. Sharp-crested weirs are inexpensive but require a relatively large headloss. The head loss required for a
particular measuring device usually varies over the range of discharges. In some cases, head needed by a flow measuring
device can reduce the capacity of the channel at that point.

Adaptability to site conditions. The selection of a flow measuring device must address the site of the proposed measurement.
Several potential sites may be available for obtaining a flow measurement. The particular site chosen may influence the
selection of a measuring device. For example, discharge in a canal system can be measured within a reach of the channel or at
a structure such as a culvert or check structure. A different device would typically be selected for each site. The device selected
ideally should not alter site hydraulics so as to interfere with normal operation and maintenance. Also, the shape of the cross-
sectional flow area may favor particular devices.

Adaptability to variable operating conditions. Flow demands for most water delivery systems usually vary over a range of flows
and flow conditions. The selected device must accommodate the flow range and changes in operating conditions, such as
variations in upstream and downstream head. Weirs or flumes should be avoided if downstream water levels can, under some
conditions, cause excessive submergence. Also, the information provided by the measuring device should be conveniently
useful for the operators performing their duties. Devices that are difficult and time consuming to operate are less likely to be
used and are more likely to be used incorrectly.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
In some cases, water measurement and water level or flow control are desired at the same site. A few devices are available for
accomplishing both (e.g., constant-head orifice, vertically movable weirs, and Neyrpic flow module; Bos, 1989). However,
separate measurement and control devices are typically linked for this purpose and usually can exceed the performance of
combined devices in terms of accuracy and level control, if care is exercised to assure that the separate devices are compatible
and achieve both functions when used as a system.

Type of measurements and records needed. An accurate measure of instantaneous flow rate is useful for system operators in
setting and verifying flow rate. However, because flow rates change over time, a single (instantaneous) reading may not
accurately reflect the total volume of water delivered. Where accounting for water volume is desired, a method of accumulated
individual flow measurements is needed. Where flows are steady, daily measurements may be sufficient to infer total volume.
Most deliveries, however, require more frequent measurements. Meters that accumulate total delivered volume are desirable
where water users take water on demand. Totalizing and automatic recording devices are available for many measuring
devices. For large structures, the cost for water-level sensing and recording hardware is small relative to the structure cost. For
small structures, these hardware costs remain about the same and thus become a major part of the measurement cost, and
may often exceed the cost of the primary structure itself.

Many water measuring methods are suitable for making temporary measurements (flow surveys) or performing occasional
verification checks of other devices. The method chosen for such a measurement might be quite different from that chosen for
continuous monitoring. Although many of these flow survey methods are suited for temporary operation, the focus here is on
methods for permanent installations.

Operating Requirements. Some measurement methods require manual labor to obtain a measurement. Current metering
requires a trained staff with specialized equipment. Pen-and-ink style water-stage recorders need operators to change paper,
add ink, and verify proper functioning. Manual recording of flows may require printed forms to be manually completed and data
to be accumulated for accounting purposes. Devices with manometers require special care and attention to assure correct
differential-head readings. Automated devices, such as ultrasonic flowmeters and other systems that use transducers and
electronics, require operator training to set up, adjust, and troubleshoot. Setting gate-controlled flow rates by simple canal level
references or by current metering commonly requires several hours of waiting between gate changes for the downstream canal
to fill and stabilize. However, if a flume or weir is installed near the control gate, that portion of the canal can be brought to the
stable, desired flow level and measured flow rate in a few minutes, and the canal downstream of the flume or weir can then fill
to the correct level over a longer time without further gate adjustments. Thus, the requirements of the operating personnel in
using the devices and techniques for their desired purposes must be considered in meter selection.

Some measuring devices may inherently serve an additional function applicable to the operation of a water supply system. For
example, weirs and flumes serve to hydraulically isolate upstream parts of a canal system from the influence of downstream
parts. This occurs for free overfall weirs and flumes flowing below their modular limits. Acoustic, propeller, magnetic, and
vortex-shedding flowmeters do not provide this function without additional structural measures such as a downstream overfall.
If these meters are used, and the isolation function is desired, then the designer should be made aware of the requirement and
provide a free overfall. Isolating the influence of upstream changes from affecting downstream channels, is less easily
accomplished. However, it can be partly implemented with orifices that have a differential head that is large compared to the
upstream fluctuations.

The designer should be aware that a sharp-crested weir overfall requires a relatively high head drop and may need to be
excessively wide to provide the isolation function with low absolute head drop. While a long board can be used downstream
from a propeller meter to provide the necessary width of flow that will pass a required quantity of water at small head, that
small head, and the crude board would not be well suited for measuring flow rate.

The designer may wish to take advantage of broad-crested weir behavior and provide a thick crest that can withstand in excess
of about 80 percent submergence, which usually translates into low absolute head loss. When used with a propeller meter, for
example, the broad-crested weir need not be well defined and can be economically installed (Replogle, 1997).

Ability to pass sediment and debris. Canal systems often carry a significant amount of sediment in the water. Removal of all

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
suspended solids from the water is usually prohibitively expensive. Thus, some sediment will likely be deposited anywhere the
velocities are reduced, which typically occurs near flow measuring structures. Whether this sediment causes a problem
depends on the specific structure and the volume of sediment in the water. In some cases, this problem simply requires routine
maintenance to remove accumulated sediment; in others, the accumulation can make the flow measurement inaccurate or the
device inoperative. Sediment deposits can affect approach conditions and increase approach velocity in front of weirs, flumes,
and orifices. Floating and suspended debris such as aquatic plants, washed=out bank plants, and fallen tree leaves and twigs
can plug some flow measurement devices and cause significant flow measurement problems. Many of the measurement
devices which are successfully used in closed conduits (e.g., orifices, propeller meters, and so on) are not usable in culverts or
inverted siphons because of debris in the water. Attempting to remove this debris at the entrance to culverts is an additional
maintenance problem.

Flumes, especially long-throated flumes, can be designed to resist sedimentation. The design options available are to select a
structure shape that will maintain velocities that assure erosion of sediments, or at least continued movement of incoming
sediments through the flume, at important flow rates. In large broad-crested weirs (a class of longthroated flumes) for
capacities greater than 1 m3 /s per m of flume width, velocities greater than 1 m/s can be achieved for the upper 75 percent of
the flow range, and is usually erosive enough to maintain flume function even for high-sediment bed loads. At the lower flow
ranges and for heavy sediment bed loads, deposition is likely and frequent maintenance may be required.

Trapezoidal sections tend to retain low velocities into the upper ranges of flow and are less sediment worthy. Long-throated
flumes with flat bottoms throughout and side contractions maintain a high velocity for 0.5 m3 /s per m width, and higher, but
must have throat lengths that are 2 to 3 times the throat width in order to be accurately computable. The sediment worthiness
of a flume design depends more on these absolute velocities than on whether the flume floor is flat throughout or raised as in a
broad-crested weir. This prompts the designer to select shapes that can provide these velocities. One suggestion for broad=
crested weirs in a fixed sized channel is to construct a false floor in the head gage area to increase the velocity there and
prevent changes in area of flow there. Also, sediments can accumulate in the upstream channel to a depth of the false floor
without affecting the function of the flume. This can extend the time between mandatory channel cleaning.

Device environment. Any measurement device with moving parts or sensors is subject to failure if it is not compatible with the
site environment. Achieving proper operation and longevity of devices is an important selection factor. Very cold weather can
shrink moving and fixed parts differentially and solidify oil and grease in bearings. Water can freeze around parts and plug
pressure ports and passageways. Acidity and alkalinity in water can corrode metal parts. Water contaminants such as waste
solvents can damage lubricants, protective coatings, and plastic parts. Mineral encrustation and biological growths can impair
moving parts and plug pressure transmitting ports. Sediment can abrade parts or consolidate tightly in bearing and runner
spaces in devices such as propeller meters.

Measurement of wastewater and high sediment transport flow may preclude the use of devices that require pressure taps,
intrusive sensors, or depend upon clear transmission of sound through the flow. Water measurement devices that depend on
electronic devices and transducers must have appropriate protective housings for harsh environments. Improper protection
against the site environment can cause equipment failure or loss of accuracy.

Maintenance requirements. The type and amount of maintenance varies widely with different measurement methods. For
example, current metering requires periodic maintenance of the current meter itself and maintenance of the meter site to
assure that is has a known cross section and velocity distribution. When the flow carries sediment or debris, most weirs,
flumes, and orifices require periodic cleaning of the approach channel. As mentioned above, design and meter selection can
mitigate the maintenance problems with sediments, but are not likely to eliminate them. Electronic sensors need occasional
maintenance to ensure that they are performing properly. Regular maintenance programs are recommended to ensure
prolonged measurement quality for all types of devices.

Construction and installation requirements. In addition to installation costs, the difficulty of installation and the need to retrofit
parts of the existing conveyance system can complicate the selection of water measurement devices. Clearly, devices that can
be easily retrofitted into the existing canal system are much preferred because they generally require less down time, and
usually present fewer unforeseen problems.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Device standardization and calibration. A standard water measurement device infers a documented history of performance
based on theory, controlled calibration, and use. A truly standard device has been fully described, accurately calibrated, correctly
constructed, properly installed, and sufficiently maintained to fulfill the original installation requirements and flow condition
limitations. Discharge equations and tables for standard devices should provide accurate calibration. Maintaining a standard
device usually only involves a visual check and measurement of a few specified items or dimensions to ensure that the
measuring device has not departed from the standard. Many standard devices have a long history of use and calibration, and
thus are potentially more reliable. Commercial availability of a device does not necessarily guarantee that it satisfies the
requirements of a standard device.

When measuring devices are fabricated onsite or are poorly installed, small deviations from the specified dimensions can
occur. These deviations may or may not affect the calibration. The difficulty is that unless an as-built calibration is performed,
the degree to which these errors affect the accuracy of the measurements is largely unknown. All too frequently, design
deviations are made under the misconception that current metering can be used to provide an accurate field calibration. In
practice, calibration by current metering to within ±2 percent is difficult to attain. An adequate calibration for free-flow
conditions requires many current meter measurements at several discharges. Changing and maintaining a constant discharge
for calibration purposes is often difficult under field conditions.

Field verification, troubleshooting, and repair. After construction or installation of a device, some verification of the calibration is
generally recommended. Usually, the methods used to verify a permanent device (e.g., current metering) are less accurate than
the device itself. However, this verification simply serves as a check against gross errors in construction or calibration. For
some devices, errors occur as components wear and the calibration slowly drifts away from the original. Other devices have
components that simply fail, that is, you get the correct reading or no reading at all. The latter is clearly preferred. However, for
many devices, occasional checking is required to ensure that they are still performing as intended. Selection of devices may
depend on how they fail and how easy it is to verify that they are performing properly.

User acceptance of new methods. Selection of a water measurement method must also consider the past history of the practice
at the site. When improved water measurement methods are needed, proposing changes that build on established practice are
generally easier to institute than radical changes. It can be beneficial to select a new method that allows conversion to take
place in stages to provide educational examples and demonstrations of the new devices and procedures.

Vandalism potential. Instrumentation located near public access is a prime target for vandalism. Where vandalism is a problem,
measurement devices with less instrumentation, or instrumentation that can be easily protected, are preferred. When needed,
instrumentation can be placed in a buried vault to minimize visibility.

Impact on environment. During the selection of a water measurement device, consideration must be given to potential
environmental impacts. Water measurement devices vary greatly in the amount of disruption to existing conditions that is
needed for installation, operation, and maintenance. For example, installing a weir or flume constricts the channel, slows
upstream flow, and accelerates flow within the structure. These changes in the flow conditions can alter local channel erosion,
local flooding, public safety, local aquatic habitat, and movement of fish up and down the channel. These factors may alter the
cost and selection of a measurement device.

21.5.3. Selection Guidelines


Selection of a water measurement method can be a difficult, time-consuming design process if one were to formally evaluate
all the factors discussed above for each measuring device. This difficulty is one reason that standardization of measurement
devices with-in water agency jurisdictions is often encouraged by internal administrators. However, useful devices are
sometimes overlooked when devices similar to previous purchases are automatically selected. Therefore, some preliminary
guidance on selection is offered to the designer so that the number of choices can be narrowed down before a more thorough
design analysis of the tradeoffs between alternatives is performed.

Short list of devices based on application. The list of practical choices for a water measuring device is quickly narrowed by site

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
conditions because most devices are applicable to a limited range of channel or conduit conditions. Economics also limits
applicable devices. For example, few irrigation deliveries to farms can justify expensive acoustic meters. Likewise, using
current meters for manual flow measurement in a channel is appropriate for intermittent information but is usually too labor
intensive for use on a continuous basis. Table 21.1 provides a list of commonly used measurement methods that are
considered appropriate for each of several applications. Table 21.2 provides an abbreviated table of selection criteria and
general compliance for categories of water measurement devices. The symbols (+), (0), and (–) are used to indicate relative
compliance for each selection criterion. The (+) symbol indicates positive features that might make the device attractive from
the standpoint of the associated selection criteria. A (–) symbol indicates negative aspects that might limit the usefulness of
this method based on that criterion. A (0) indicates no strong positive or negative aspects in general. A (V) means that the
suitability varies widely for this class of devices. The letters NA mean that the device is not applicable for the stated conditions.
A single negative value for a device does not mean that the device is not useful or appropriate, but other devices would be
preferred for those selection criteria.

Vortex-shedding flow meters are not specifically rated in this grouping. They are expected to compete with orifice meter
applications. They generally offer less head loss that orifice meters and can cover a wider discharge range for a particular
installation. Although they have been around for many years, they have only recently been offered in a configuration that makes
them competitive with orifice meters, which they are generally expected to replace because they can produce less pipe head
loss. Open-channel applications for vortex-shedding meters are not considered practical.

Table 21.1 Application–Based Selection of Water Measurement Devices

1. Openchannel conveyance system

 a. Natural channels (see Herschy, 1985)

  (1) Rivers

    Periodic current metering of a control section to establish stage=discharge relation

    Broad–crested weirs

    Long–throated flumes

    Short–crested weirs

    Acoustic velocity maters (AVM—transient time)

    Acoustic Doppler velocity profiles

    Float–velocity/area method

    Slope–area methods

  (2) Intermediate–sized and small streams

    Current metering/control section

    Broad–crested weirs

    Long–throated flumes

    Short–crested weirs

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
    Short–throated flumes

    Acoustic velocity meters (AVM—transient time)

    Float–velocity/area method

 b. Regulated channels (see USBR, 1997)

  (1.) Spil ways

   (a) Gated

     Sluice gates

     Radial gates

   (b) Ungated

     Broad–crested weirs (including special crest shapes, Ogee crest, etc.)

     Short–crested weirs

  (2) Large canals

   (a) Control structures

     Check gates

     Sluice gates

     Radial gates

     Overshot gates

   (b) Other

     Long–throated flumes

     Broad–crested weirs

     Short–throated flumes

     Acoustic velocity meters

  (3.) Small canals (including openchannel fluid conduit flow)

    Long–throated flumes

    Broad–crested weirs

    Short–throated flumes

    Sharp–crested weirs

    Rated flow control structures (check gates, radial gates, sluice gates, overshot gates)

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
    Acoustic velocity meters

  (c) Other

    Float–velocity/area methods

  (4) Irrigation delivery to farm turnout

   (a) Pipe turnouts (short inverted siphons, submerged culverts, etc.)

     Metergates

     Current meter

     Weirs

     Short–throated flumes

     Long–throated flumes

   (b) Other

     Constant head orifice

     Rated sluice gates

     Movable weirs

2. Closed conduit conveyance systems (see Brater and King, 1982; Miller, 1996)

 (a.) Large pipes

   Venturi meters, venturi tubes, nozzles

   Rated control gates (orifice)

   Acoustic velocity meters (transit time)

 (b.) Small and intermediate–sized pipelines

   Venturi meters, Venturi tubes, nozzles

   Orifices (in–line, end–cap, shunt meters, etc.)

   Propeller and turbine meters

   Magnetic meters

   Acoustic meters (transit–time and Doppler)

   Pitot tubes

   Elbow meters

   Vortex – shedding

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
   Trajectory methods (e. g., full-pipe trajectory; California pipe method for part full pipe)

   Other commercially available meters (household types)

Table 21.2 Selection Guide for Water Measuring Devices.

Source: Adapted from USBR (1997).

Symbols 1, 0, 2 are used as relative indicators comparing application of the listed water measuring device to the listed criteria

Symbol V denotes that situability varies widely.

Symbol NA denotes “not applicable” to criteria


*Venturi, orifice, pilot tube, etc.
†Propeller meters, turbine meters, paddle wheel meters, etc.

21.6. SELECTION OF SECONDARY DEVICES FOR DISCHARGE


READOUT AND CONTROL
While the emphasis of this chapter is on hydraulic design, it is important that the output of the design be translated into useful
information for the user.

21.6.1. Intended Uses


The secondary device that is used with a flowmeter depends strongly on the use of the information. Immediate management
decisions, such as adjusting a valve or canal gate, require nearly instant feedback to the operator at an accuracy and precision
that fully utilizes the available accuracy of the primary device. As discussed above, random errors over many measurements
tend to cancel, and long-term totals can often absorb large random detection errors that would not be tolerable for decisions

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
depending on a single reading. The designer should be cognizant of the user needs and be prepared to provide an appropriate
output.

21.6.2. Quality Assurance


Secondary devices provide a variety of functions, primary of which are data recording and data quality assurance. The
secondary devices necessary for these may not be the same. It is good practice to provide manual, instantaneous flow rate
output at the meter site so that the servicing personnel can quickly know that the main secondary instrument is functioning. For
weirs and flumes wall gages that show flow rates directly are recommended as a quick visual check of secondary
instrumentation. However, a wall gage that shows a head reading that is converted to a discharge rate by the technician using a
table or equation will usually suffice.

Sometimes it is practical to provide field check capability to a secondary device by special treatment of the installation. For
example, if a pressure transducer is used to detect head on a flume, the transducer can be mounted in a stilling well attached to
the flume. If it is further mounted on a movable rack, it can allow servicing personnel to raise and lower the transducer a
prescribed amount to verify that the output signal correctly reports the change (Replogle, 1997).

In pipe flows, differential head meters, such as Venturi and orifice meters should be fitted with manometer ports that can be
easily accessed by servicing personnel to verify the detection and transmission of data by electronic devices. Propeller, turbine,
acoustic, and magnetic meters are among meters that are usually closely integrated with their secondary devices so that
separate verification of the primary device function is more challenging.

These checks of the secondary devices do not, however, necessarily detect a malfunction of the primary device, such as scale
growth on Venturi surfaces, worn orifice plates, or sediment filled weirs and flumes. It is sometimes advisable to provide
measurement in a main line and in both lines of a bifurcation, thus providing redundancy that can help detect meter
malfunctions. Regular inspection is a necessary operational requirement and an important design and selection feature is often
the ease of inspection.

21.7. APPLICATIONS OF LONG-THROATED FLUMES


Long-throated flumes are replacing the older open-channel devices, such as Parshall flumes, because of their design flexibility
in terms of size, shape, accuracy and economics. Versions for large canals and small irrigation furrows are in use. Shapes vary
from rectangular to trapezoidal, with circular versions useful in partly full pipes and complex versions for exceptionally wide
flow ranges (Bos et al., 1991; Clemmens et al., 1993; USBR, 1997; Wahl and Clemmens, 1998).

Long-throated flumes and broad-crested weirs operate by using a channel contraction to cause critical flow. If there is not
enough contraction, critical flow does not occur. Flow is then nonmodular and gauge readings become meaningless. If there is
too much contraction, the water surface upstream may be raised excessively and cause canal overtopping or other problems.
The problem facing the designer is to select the shape of the control section, or throat, so that critical flow occurs throughout
the full range of discharges to be measured. Also, the designer must provide acceptable sensitivity and accuracy while not
causing too much disruption in upstream flow conditions (e.g., sediment deposition, canal overtopping). This appears to be a
difficult task, but existing design aids and rating tables make this task more manageable. A selection of these aids and tables
are presented herein. A multitude of possible designs have been sorted, based on practical experience and theory, to a relatively
few selected structures from which the designer may choose.

Table 21.3 Choices of Broad – Crested Weir Sizes and Rating Tables for Lined Canals, Metric Units. *,†

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Canal Max Canal [Range of Canal [Range of Canal Selection Weir Crest Weir Shape Min Head
Shape Side Bottom Depth† Capacities] Lower‡ Capacities] Upper Table Width b c Sill Ht. p1 Loss ΔH
Slope z1 Width b 1 d (m) (m3 /s) (m3/s) 21.4 (m) (m) (m)
(m)

(1) (2) (4) (4) (5) (6) (7) (8) (9)

Source: Adapted From Replogte et al., (1990).


aLa = Hmax:Lb = 2 to 3p 1:x = La + Lb > 2 to 3Hmax

L > 1.5H1max, but within given in Table 25.4

d > 1.2 h1max + p 1; ΔH < .0.1H1


†Limited by sensitivity
‡Maximum recomended canal depth
||Limited by Froude number, otherwise limited by canal depth

1.0 0.25 0.70 0.08 0.14|| Am 0.5 0.125 0.015

0.09 0.24|| Bm 0.6 0.175 0.018

0.10 0.38 Cm 0.7 0.225 0.022

0.11 0.43 Dm1 0.8 0.275 0.026

0.12 0.37 Em1 0.9 0.325 0.030

0.13 0.32 Fm1 0.1 0.375 0.033

1.0 0.30 0.75 0.09 0.21 Bm 0.6 0.15 0.017

0.10 0.34 Cm 0.7 0.20 0.021

0.11 0.52 Dm1 0.8 0.25 0.025

0.12 0.52 Em1 0.9 0.30 0.029

0.13 0.44 Fm1 1.0 0.35 0.033

0.16 0.31 Gm1 1.2 0.45 0.039

1.0 0.50 0.80 0.11 0.33|| Dm2 0.8 0.15 0.019

0.12 0.52|| Em1Em2 0.9 0.20 0.024

0.12 0.68|| Fm1Fm2 1.0 0.25 0.029

0.16 0.64 Gm1 1.2 0.35 0.037

0.18 0.46 Hm 1.4 0.45 0.043

0.20 0.29 Im 1.6 0.55 0.048

1.0 0.60 0.90 0.12 0.39|| Em2 0.9 0.15 0.021

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Canal Max Canal [Range of Canal [Range of Canal Selection Weir Crest Weir Shape Min Head
Shape Side Bottom Depth Capacities] Lower Capacities] Upper Table Width b Sill Ht. p Loss ΔH
Slope z Width b d (m) (m /s) (m s) 21.4 (m) (m) (m)
(m)

(1) (2) (4) (4) (5) (6) (7) (8) (9)

0.13 0.62 Fm2 1.0 0.20 0.025

0.16 1.09 Gm1 1.2 0.30 0.035

0.18 0.86 Hm 1.4 0.40 0.043

0.20 0.64 Im 1.6 0.50 0.050

0.22 0.43 Jm 1.8 0.60 0.049

1.0 0.75 1.0 0.16 0.91 Gm2 1.2 0.225 0.030

0.18 1.51 Hm 1.4 0.325 0.038

0.20 1.22 Im 1.6 0.425 0.047

0.22 0.94 Jm 1.8 0.525 0.053

1.5 0.60 1.2 0.20 1.3|| Km 1.50 0.300 0.031

0.24 2.1|| Lm 1.75 0.383 0.38

0.27 2.5 Mm 2.00 0.467 0.044

0.29 2.2 Nm 2.25 0.550 0.050

0.32 1.8 Pm 2.50 0.633 0.056

0.35 1.4 Qm 2.75 0.717 0.059

1.5 0.75 1.4 0.24 1.8 Lm 1.75 0.333 0.036

0.27 2.8 Mm 2.00 0.417 0.042

0.29 3.9|| Nm 2.25 0.500 0.049

0.32 3.5 Pm 2.50 0.583 0.055

0.35 3.1 Qm 2.75 0.667 0.062

0.38 2.6 Rm 3.00 0.750 0.066

1.5 1.00 1.6 0.29 3.4|| Nm 2.25 0.417 0.046

0.32 4.7 Pm 2.50 0.500 0.052

0.35 5.7 Qm 2.75 0.583 0.059

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Canal Max Canal [Range of Canal [Range of Canal Selection Weir Crest Weir Shape Min Head
Shape Side Bottom Depth Capacities] Lower Capacities] Upper Table Width b Sill Ht. p Loss ΔH
Slope z Width b d (m) (m /s) (m s) 21.4 (m) (m) (m)
(m)

(1) (2) (4) (4) (5) (6) (7) (8) (9)

0.38 5.1 Rm 3.00 0.667 0.065

0.43 3.9 Sm 3.50 0.833 0.081

1.5 1.25 1.7 0.32 4.1|| Pm 2.50 0.417 0.048

0.35 5.6|| Qm 2.75 0.500 0.055

0.38 7.2 Rm 3.00 0.583 0.061

0.43 5.9 Sm 3.50 0.750 0.074

0.49 4.5 Tm 4.00 0.917 0.084

0.55 3.3 Um 4.50 1.083 0.089

1.5 1.50 1.8 0.35 4.8|| Qm 2.75 0.417 0.051

0.38 6.5|| Rm 3.00 0.500 0.058

0.43 8.1|| Sm 3.50 0.667 0.071

0.49 6.6 Tm 4.00 0.833 0.083

0.55 5.1 Um 4.50 1.000 0.092

Table 21.4 Rating Equations and Ranges of Application for Trapezoidal Broad–crested Weirs * (Adapted from Replogle,
Clemmens and Bos, 1990) Discharge equation: Q = C1(h1 + C2) U Units: (h1, m; Q,m3/s)

Coef for Weir A m Weir B m Weir C m Weir D m1 Weir D m2 Weir E m1 Weir E m2 Weir F m1
Eq.

0.23 < L < 0.34 0.30 < L < 0.42 0.35 < L < 0.58 040 < L < 0.58 0.30 < L < 0.45 0.38 < L < 0.56 0.38 < L <0.56 0.42 < L < 0.61

Source: Adapted from Replogle et al. (1993).

Calibrations developed with computer model (Clemmens et al., 1993).

C1 2.145 2.365 2.276 2.837 2.913 2.921 3.037 3.122

C2 0.0067 0.0079 0.0045 0.0125 0.0101 0.0090 0.0089 0.010

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Coef for Weir A Weir B Weir C Weir D Weir D Weir E Weir E Weir F 1
Eq.

0.23 < L < 0.34 0.30 < L < 0.42 0.35 < L < 0.58 040 < L < 0.58 0.30 < L < 0.45 0.38 < L < 0.56 0.38 < L <0.56 0.42 < L < 0.61

U 1.8667 1.8599 1.7597 1.8765 1.8555 1.8152 1.8252 1.813

h1min 0.030 0.030 0.040 0.040 0.040 0.030 0.030 0.030

h1max 00.220 0.280 0.340 0.390 0.300 0.370 0.370 0.370

Q1min 0.005 0.005 0.010 0.010 0.010 0.010 0.010 0.010

Q1max 0.140 0.240 0.380 0.520 0.250 0.520 0.520 0.680

Coef. for Weir F m2 Weir G m1 Weir G m2 Weir H m Weir I m Weir J m Weir K m Weir L m
Eq.

0.42 < L < 0.61 0.50 < L < 0.75 0.45 < L < 0.68 056 < L < 0.84 0.48 < L < 0.71 0.40 < L < 60 0.48 < L <0.72 0.58 < L < 0.87

C1 3.319 3.601 3.695 3.854 4.135 4.281 4.369 5.389

C2 0.0107 0.0130 0.0108 0.0067 0.0096 0.0048 0.0170 0.0151

U 1.8376 1.8271 1.8141 1.7522 1.7438 1.6708 1.8324 1.8581

h1min 0.030 0.040 0.040 0.040 0.040 0.030 0.060 0.060

h1max 0.410 0.500 0.450 0.560 0.480 0.400 0.470 0.580

Q1min 0.010 0.020 0.020 0.020 0.020 0.020 0.040 0.050

Q1max 0.680 1.100 0.920 1.500 1.200 0.940 1.300 2.100

Coef. for Weir M m Weir N m Weir P m Weir Q m Weir R m Weir S m Weir T m Weir U m
Eq.

0.65 < L < 0.97 0.75 < L < 1.10 0.80 < L < 1.20 0.85 < L < 1.20 0.85 < L < 1.28 0.95 < L < 1.40 0.95 < L < 1.40 0.85 < L < 1.20 0.68 < L
<1.00

C1 5.831 6.284 6.713 7.115 7.563 8.397 9.144 9.789

C2 0.0169 0.0268 0.0190 0.023 0.0243 0.0209 0.0129 0.0080

U 1.8528 1.8986 1.8397 1.8331 1.8477 1.8043 1.7301 1.6711

h1min 0.060 0.090 0.080 0.070 0.070 0.060 0.060 0.060

h1max 0.650 0.750 0.790 0.850 0.930 0.950 0.810 0.660

Q1min 0.050 0.100 0.100 0.100 0.100 0.100 0.100 0.100

Q1max 2.800 3.900 4.700 5.700 7.200 8.200 6.600 5.100

Stage-discharge equations or direct rating tables for three types of channel conditions are presented in this chapter. These are:
(l) lined trapezoidal channels, (2) earthen canals or lined rectangular channels, and (3) circular pipes and conduits not flowing

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
full. Some offerings for natural streams are presented in Bos et al., (1991) and Replogle et al., (1990).

Trapezoidal broad-crested weirs (i.e., having only a bottom contraction) were used for the development of selected standard
sizes for lined trapezoidal channels. Calibration equations developed from tables computed with the computer model
(Clemmens et al., 1993) are presented in the following series of tables (Tables 21.3 and 21.4). Trapezoidal flumes with side
contractions were not selected for general use in standard irrigation situations because they are usually more difficult to
construct and require more head loss. A design aid was also developed to help ensure sufficient sensitivity, limit the Froude
number, and otherwise assist in selection. The designer needs to only select a weir width with its corresponding sill height.

Rectangular broad-crested weirs were again chosen for unlined canals. However, for these weirs the designer must select a
channel (and throat) width as well as a sill height, and must be more aware of the other design considerations. For lined
rectangular channels, only the sill height must be selected. These tables and equations can also be used to determine the rating
for side-contracted, rectangular flumes by appropriate adjustments to handle changing velocity-of-approach problems (Bos et
al., 1991).

Circular pipes can also be accommodated with broad-crested weirs extending across the pipe. Convenient sizes have been built
using sill heights ranging from 0.2 to 0.5 of the pipe diameter. The designer selects a sill height and pipe diameter to handle the
desired flow rate and backwater conditions.

For natural streams, V – shaped flumes were chosen because of the wide range of discharges that must be measured. The only
design choice here is the throat side slope. The designer has the option of designing a flume shape or size, not presentad here,
by using the theoretically based computer program (Clemmens et al., 1993). A new version of this program for the Windows
environment is currently in the beta-test phase (Wahl and Clemmens, 1998).

Figure 21.6 Broad-crested weir in trapezoidal, concrete-lined canal.

Sediment carrying capabilities are sometimes important design characteristic. As mentioned in a previous section, flumes that
are V= shaped do not pass bedoad sediment well.

The rating tables are developed with the assumption of a particular known approach channel cross section (or flow area).
However, any particular control section size and shape can be used with any approach section size and shape. The discharge
can be adjusted with an approach velocity coefficient, as discussed in detail in Bos et al., (1991), or custom calibrated with the
flume program (Clemmens et al., 1993; Wahl and Clemmens, 1998). The rating tables given here automatically limit the Froude
number. If smaller approach areas are used, the designer must determine that the Froude number remains less than about
0.45.

Frequently, the site conditions may call for flumes that would have dimensions beyond the ranges provided by the ratings in this
chapter. To extend beyond these limits and for further information refer to Ackers et al., (1978), Bos (1989); Bos et al., (1991);

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Ciemmens et al., (1984, 1993); and Wahl and Clemmens, (1998).

21.7.1. Structures for Lined Trapezoidal Canals


Standard broad-crested weir sizes are recommended for use in slip-formed canals of convenient metric dimensions. In lined
canals, the canal itself furnishes the entrance channel section La, and the tailwater section (Fig. 21.6). Basic requirements
become a converging transition to the throat section and the throat section itself. Calibrations for these standard sizes are
included herein so that the subsequent designer may select one of these weirs to be built into an existing lined channel as
shown in Fig. 21.6. In selecting these standardsized canals and the related flow rates, consideration was given to proposals by
the International Commission on Irrigation and Drainage, to the construction practices of the U.S. Bureau of Reclamation, U.S.
Department of the Interior (USDI), and to design criteria for small canals used by the Natural Resources Conservation Service,
U.S. Department of Agriculture (USDA) (Bos et al., 1991).

Present practice leans or tends toward side slopes of 1:1 horizontal to vertical for small, monolithic, concrete-lined canals with
bottom widths less than about 0.8 m and depths less than about 1 m. Deeper and wider canals tend toward side slopes of
1.5:1. When the widths and depths are greater than about 3 m, the trend is more toward 2:1 side slopes. This is particularly
observed if canal operating procedures may allow rapid dewatering of the canal. In some soil conditions this can cause
hydrostatic pressures on the underside of the canal walls that lead to wall failure. Most of the lined canals used in a tertiary
irrigation unit or on large farms are of the smaller size. They have 0.3- to 0.6-m bottom widths, 1:1 side slopes, and capacities
below 1 m 3 /s (35 ft3 /s).

In Table 21.3 precomputed broad-crested weir selections are given for canals with bottom widths at quarter-meter increments,
with special insertions for 0.3 m (approximately 1 ft) and 0.6 m (approximately 2 ft). The offering of many precomputed sizes
will aid in retrofitting older canal systems and yet not prevent the adoption of standard sized canals as proposed by the
International Commission on Irrigation and Drainage (ICID). Canal sizes with bottom widths in excess of 1.5 m or 5 ft,
respectively, are avoided in the precomputed tables on the assumption that these sizes deserve special design consideration.

Table 21.3 shows a number of precomputed weirs in SI units that may be used for the various combinations of bottom widths
and sidewall slopes as given in the first two columns. American units are available in USBR (1997). The third column lists
recommended values of maximum canal depth, d, for each side-slope, bottom-width combination.

For each canal size, several standard weirs can be used (Column 6). Columns 4 and 5 give the limits on canal capacity for each
canal-weir combination. These limits on canal capacity originate from three sources:

1. The Froude number in the approach channel is limited to a maximum of 0.45 to ensure water surface stability.

2. The canal freeboard Fb upstream from the weir should be greater than 20 percent of the upstream sill-referenced head, h1 . In
terms of canal depth this limit becomes d = 1.2 h1 + P1 .

3. The sensitivity of the weir at maximum flow should be such that a 0.01 m change in the value of the sill-referenced head h1
causes less than 10 percent change in discharge.

Also indicated in the last column of Table 21.3 is a minimum headloss ΔH that the weir must provide. Excessive downstream
water levels may prevent this minimum headloss, which means that the weir exceeds its modular limit and no longer functions
as an accurate measuring device.

The required head losses for the various broad-crested weirs were evaluated by the method described earlier and, for design
purposes, listed for each weir size with the restriction that the computed modular limit will not exceed 0.90.

Thus, the design headloss is either 0.1 H1 or the listed value for ΔH, whichever is greater. For these calculations, it was
assumed that the weir was placed in a continuous channel with a constant cross section (e.g., p1 = p2 , b1 = b2 , and z1 = z2 ) and
that the diverging transition was omitted (abrupt expansion). Technically, the modular limit is based on the drop in total energy
head through the weir (i.e., including velocity head). In the above continuous channels, the velocity head component is usually

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
similar in magnitude upstream and downstream from the structure when p1 is approximately equal to p2 . This means that h
may be substituted for H in most cases.

Table 21.3 is primarily intended for the selection of these standard weirs. It is also useful for the selection of canal sizes. The
Froude number in the canal is automatically limited to 0.45. Selecting the smallest canal for a given capacity will give a
reasonably efficient section. For instance, if the design capacity of the canal is to be 1.0 m3 /s, the smallest canal that can be
incorporated with a measuring structure has b1 = 0.60 m, z1 = 1.0, and d = 0.90 m. Larger canals can also be used. The hydraulic
grade line of the channel should also be checked to ensure an adequate design.

Each standard weir can be used for a range of bottom widths. This is possible because the change in flow area upstream from
the weir causes only a small change in velocity of approach and a corresponding small change in energy head. The width
ranges have been selected so that the error in discharge caused by the change in flow area is less than 1 percent. This is a
systematic error for any particular approach channel size, and the extent of this error varies with discharge.

A weir suitable for several of the listed canal bottom widths is also suitable for any intermediate width. For example, inTable
21.3, Weir G m1 can be used in canals with bottom widths of 0.30, 0.50, and 0.60 m, or any intermediate width, for example,b1 =
0.45 m. The user will need to calculate the sill height, headloss, and maximum design discharge for these intermediate sizes.

The rating equations for the weirs are given in Table 21.4 and will reproduce the values presented in the original tables
produced by computer modeling (Bos et al., 1991) to within about 1 percent. The original tables were computed using the
following criteria (Fig. 21.6):

1. Each weir has a constant bottom width bc and a sill height p1 that varies with the canal dimensions.

2. The ramp length can be chosen such that it is between 2 and 3 times the sill height. The 3:1 ramp slope is preferable.

3. The gage is located a distance at least H1max upstream from the start of the ramp. In addition, it should be located a distance
of roughly two to three times H1max from the entrance to the throat.

4. The throat length should be greater than 1.5 times the maximum expected sill-referenced depth h1max, but should be within
the limits indicated in Table 21.4.

5. The canal depth must be greater than the sum of ( p1 + h1max + Fb), where Fb is the freeboard requirement, or roughly 0.2
times h1max.

Occasionally, a weir cannot be found from these tables that will work satisfactorily. The user must then judge between several
options, for example:

1. Find a new site for the flume with more vertical head available.

2. Add to the canal wall height upstream from the site so that more backwater effect can be created.

3. Try one of the other weir shapes.

4. Use the tables to interpolate and get a rating for an intermediate width, probably with some sacrifice in accuracy.

5. Produce a special design using the computer model.

Example. Given: An existing canal has a bottom width of b1 = 0.30 m, side slopes z1 = 1, and a total depth d = 0.55 m. The
discharge depth relationship was estimated for one flow rate by using surface flows, and using Manning’s formula (Eq. 16) to
estimate the depth at the maximum expected flow. For this example the maximum flow was estimated to be Qmax = 0.15 m3 /s
at a canal flow depth y2 = 0.43 m prior to installation of the measuring device.

Required: Select a broad-crested weir structure from Table 21.4.

Procedure: The basic strategy for selecting a flume for an existing canal is to cause enough rise in the existing flowing water
surface to create the necessary water surface drop needed to allow the flume to function properly. Because the rise in water

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
surface is nearly equal to the required energy loss, ΔH, this value is used to approximate the water surface drop, or, Δh ≈ ΔH.
This minimum water surface rise usually must be provided for the maximum expected flow rate, and is then safe for lesser flow
rates for this type of installation in a trapezoidal channel. (Note: Placing a rectangular structure in a trapezoidal channel may
require checking the low flow also, because as flow rate decreases, the depth drops faster in a rectangular channel than in a
trapezoidal channel, and discharge may be nonmodular at low flow rates.)

Use Table 21.3 and find the rows for z1 = 1 and b1 = 0.30 m. For a flow rate of 0.15 m3 /s, weirs Bm through Fm1 can be tried
because Qmax falls within the discharge ranges for each of these weirs. Weir Bm is tried first because it has the lowest sill
height, p1 = 0.15 m, and should be the most economical to build. The minimum required head loss listed in column 9, ΔH, for
this weir is 0.017 m. The actual headloss through the weir should be the greater of the listed ΔH and 0.1h1 . To find the head on
the flume, h1 , use the discharge equation for long-throated flumes given at the top of Table 21.4:

From Table 21.4 the appropriate coefficients for the discharge equation for flumeBm are

Solving for h1 and with Q = 0.150 m3 /s:

In this case, 0.1 h1 = 0.0219 m exceeds the listed value of 0.017 m, and the larger value of 0.0219 m is used as the needed drop
in water surface.

For the flume to function properly, the difference in the upstream water depth,h1 + p1 , and the downstream channel depth, y2 ,
should be greater than the required water surface drop, or

The computed value is actually negative, which means this flume would be completely submerged

Now examine the next=sized weir. However, because the next size can be estimated to raise the water surface by something
slightly less than the sill increase of 0.05 m, it is unlikely that raising the sill by only 0.05 m will work because the previous sill
produced no drop, by virtue of the negative value calculated. Thus, move on to weir Dm1 with a p1 = 0.21 m.

In Table 21.4, we find ,with the corresponding values from Table 21.4 for C1 ,C2 and U, that h1 = 0.196 and ΔH = 0.025 m in
Table 21.3. However, 0.1 h1 = 0.020 m, so use the listed value for ΔH = 0.025 m as the needed drop.

Now again check the difference in the upstream depth,h1 + p1, and downstream channel depth, y2 to see if (h1 + p1 ) – y2 > ΔH
= 0.025 m (the required drop in water surface).

This time, (h1 + p1 ) – y2 = (0.196 + 0.250) – 0.430 = 0.016 m, which still does not provide the needed drop of 0.025 m.

Next try weir Em1 with p1 = 0.30 m. This time, h1 is calculated to be 0.186 m and, from Table 21.3, ΔH = 0.029 m.

Then (h1 + p1 ) – y2 = (0.186 + 0.30) – 0.430 = 0.056 m, which exceeds the needed 0.029 m value. Thus, this weir meets the
primary criterion and can be checked further.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Weir C1 C2 U h1 0.1 h 1 Table ΔH Req. Drop h1 +p1 –y2 Adequate

Bm 2.365 0.0079 1.8599 0.219 0.0219 0.217 0.0219 20.061 No

Dm1 2.913 0.0101 1.8555 0.196 0.0219 0.025 0.025 0.016 No

Em1 2.921 0.0090 1.8152 0.186 0.0186 0.029 0.029 0.056 Yes

The minimum required canal depth, d, including freeboard, can be calculated from dmin = 1.2 h1 + P1 = 1.2 (0.186) + 0.30 = 0.523
m, which is less than the 0.550 m canal depth given in this example. This weir is acceptable, and the remainder of the hydraulic
dimensions can be obtained from Tables 21.3 and 21.4. They are

The length from the gauge to the ramp La = H1max = 0.37 m

The ramp length Lb = 3 p1 = 3 (0.3 m) = 0.9 m

The throat length L = 3 p1 = 1.5 (0.186) = 0.28 m

However, from Table 21.4, L > 0.38 m, so use L = 0.38 m, to use the given calibration values.

A downstream expansion is not necessary.

The flume width bc = 0.90 m.

This example is modified from Bos et al. (1991). Other examples and detailed discussion of this example are presented therein.

Figure 21.7 Flow measuring structure for earthen channel with rectangular control section. (From
Bos et al., 1991)

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Table 21.5 Rating equations and Ranges of Application for Rectangular Weirs *

Source: Adapted from Replogle et al. (1990).


*Calibrations developed with computer model (Clemmens et al., 1993).

21.7.2. Rectangular Structures for Unlined Canals


Weirs and flumes for earthen (unlined) channels require a structure that contains the following basic parts: entrance to
approach channel, approach channel, converging transition, throat, diverging transition, stilling basin, and riprap protection.

As illustrated in Fig. 21.7, the discharge measurement structure for an earthen channel is longer, and thus more expensive, than
a structure in a concrete-lined canal (Fig. 21.6). In the latter, the approach channel and sides of the control section already exist
and the riprap is not needed.

The approach canal of Fig. 21.7 provides a known flow area and velocity of approach. The rating equations for the rectangular
weirs given in Table 21.5 assume that the approach section is rectangular and has the same width as the throat. If the

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
upstream sillreferenced head is not measured in a rectangular approach canal of this same width, but instead is measured in
the upstream earthen section, then these tables require correction to the discharge, Q, for the change in the approach velocity.
Procedures for this are given by Bos (1989).

The full-length structure of Fig. 21.7 can be further shortened by deleting the diverging transition or the rectangular tail water
channel. The diverging transition may be deleted if the available head loss over the structure exceeds 0.4 ΔH1 , so that no
velocity head needs to be recovered. The rectangular tail water channel may be deleted if, at maximum flow, the Froude number
at the beginning of the exit channel is less than 1.7 (Bos et al.,1991).

A rectangular broad-crested weir discharges nearly equal quantities of water over equal widths. The major differences are
associated with the friction along the walls. Thus the flow is nearly two-dimensional over the weir, so that rating tables can
provide the flow rate, q, in m3 /s per meter width of sill for each value of h1 . This allows a wide variety of sizes for rectangular
broad-crested weirs. For each width, bc, of the weir, an accurate rating table for the total discharge,Q, can be developed by
multiplying the table discharges by bc. Thus,

(21.21)

Table 21.5 presents rating equations for a series of rectangular broad-crested weirs that were developed from the computer-
modeled tables given in Bos et al. (1991). These equations will reconstruct those computer-derived tables to within about 1.5
percent. This adds an additional uncertainty to the ±3 percent error claimed for the original tables. That 3 percent includes the
added error caused by averaging small groupings of weir widths. The weirs were selected to keep the error of side wall effects
to within 1 percent. Ratings are given for several sill heights, p1 , to aid in design. Interpolation between sill heights gives
reasonable results.

If the approach area, A1 , is larger than that used to develop these rating tables, either because of a higher sill or a wider
approach channel, the ratings must be adjusted for Cv. To simplify this process, the discharge over the weir for aCv value of 1.0
is given in the far right column of each grouping. This column is labeled P1 = ∞, because that would cause the velocity of
approach to be zero and Cv = 1.0. This approximates a weir at the outlet of a reservoir or lake. The complete correction
procedure is given in Bos et al., (1991) and Bos (1989).

The design procedure for lined rectangular canals, is relatively straightforward. It consists of selecting a sill height,P1 , that
causes modular flow throughout the discharge range, and provides sufficient freeboard at the maximum discharge. For unlined
canals, an appropriate width must be chosen. There are usually several widths that will work. Extremely wide, shallow flows are
subject to measurement errors due to low head detection sensitivity. Extremely narrow, deep flows require long structures and
large head losses.

Because of the wide variety of shapes that can be encountered in earthen channels and in the range of discharges to be
measured, it is rather complicated to determine the interrelated values of h1max, p1 , and bc of the structure. While this makes the
design process somewhat more complicated, it allows the designer greater flexibility and expands the applicability of the weirs.

1. The discharges to be measured (per meter width) must be within the range of discharges shown in the rating table for the
selected weir if these flume dimensions and tables are to be used.

2. The allowable measurement error should not be exceeded. This allowable error may be different at different flow rates (see
Bos et al., 1991).

3. Flow should be modular at all flow rates to be measured.

4. Placing a weir in the canal should not cause overtopping upstream.

5. The measuring structure should be placed in a straight section with a relatively uniform cross section for a distance of about
10 times the width of the channel.

6. The Froude number should not exceed 0.45 for a distance of at least 30 times h1 upstream from the weir.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
The following criteria should be considered by the designer.

If these criteria are followed, the designer should obtain a satisfactory structure that will operate as intended.

For a rectangular structure in an earthen canal, the rectangular section need not extend 10 times its width upstream from the
flume if a gradual taper is used to guide the flow into the rectangular section. For the flumes given here, it is recommended that
the rectangular section extend upstream from the head measurement location (gauging station), as shown in Fig. 21.7. It is
also recommended that riprap be placed downstream from the structure for a distance of 4 times the maximum downstream
water depth. A step should be provided at the transition between the rectangular section and the riprap section to avoid local
erosion from floor jets. Sizing of riprap and filters is discussed by Bos (1989) and Bos et al. (1991).

An analysis of head measurement errors is presented by Bos (1989) and Bos et al., (1991) and will not be repeated here. For
lined channels, a freeboard criterion of 0.2 h1max has been used satisfactorily. For unlined channels it may be more appropriate
to specify a maximum water depth, y1max. Submergence or modular flow should be checked at both minimum and maximum
expected discharges.

Figure 21.8 Long-throated flume in a partly filled pipe.

If the channel is rectangular, or the length of the rectangular throated structure downstream from the weir sill is as inFig. 21.7,
then we can use the lower value of ΔH = 0.1 H1 or the ΔH value given at the bottom of Table 21.5. If a shorter length in an
earthen channel is used and the tailwater channel is significantly larger than the stilling basin would be, then considerably more
headloss will probably be required. The designer could use the headloss value for the discharge into a lake or pool, ΔH = 0.4 H1 .
This may represent a drastic difference in the value of headloss. The designer may decide to use the shortened structure and
calculate the actual modular limit by use of the computer model (Bos et al., 1993; Wahl and Clemmens, 1998). Another
alternative is to build a prototype in the field and set the crest to the appropriate level by trial and error.

For a more comprehensive background and design discussion for rectangular flumes, refer to Bos et al. (1991).

21.7.3. Structures for Circular Channels


The broad-crested weir has been adapted for flow measurements in pipes with a free water surface because of its relatively
simple construction, Fig. 21.8. Other shapes, such as side contractions and center line piers can be and have been used. The
reader is referred to the computer program and design procedure for most calibrations (Clemmens et al., 1993; Wahl and
Clemmens, 1998). There are two situations of particular interest for designing a weir in a pipe. The weir can be placed
somewhere in the middle of a straight pipe, or at the end of a pipe, such as the entrance to a deep manhole, which is probably
more common because the site is more accessible.

When the weir is placed in the interior of a straight pipe, the presence of the weir must cause a rise in the upstream water
surface to develop the required head loss. This increase in flow area upstream causes a proportional decrease in velocity and

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
may cause subsequent sediment deposition. Where this is a problem, a downstream ramp should be considered to reduce the
required head loss and thus accommodate a smaller sill height. Sediment problems are further aggravated at low flows
because the free-flowing (normal-depth) water surface drops proportionally faster than the water surface upstream from the
weir. Thus, the velocity difference becomes greater and greater with decreasing discharge. For situations where wide
fluctuations in flow rates exist and sedimentation is a problem, an alternative weir shape or location should be considered.

An alternative for the measuring location is the end of a pipe, particularly where there is a large drop in water surface. A weir
can often be designed so that the water level upstream from the weir either matches, or is below, the normal depth of water in
the pipe. This may avoid aggravating sedimentation problems.

Portable flumes using plastic pipe. Portable flumes for flow rates up to about 50 L/S with trapezoidal or rectangular shape are
described in Bos et al. (1991). More recently, calibrations have been presented for sills placed in circular pipe sections. Small
versions are convenient to construct and use as portable flumes. Dimensionless ratings for average roughnesses and profile
lengths were previously presented for partly full circular culverts and pipes fitted with similar sills (Clemmens et al., 1984).
However, direct computation by the computer model (Clemmens et al., 1993) for the specific roughness of the construction
materials provides a slightly more accurate rating. These portable versions can be used for measuring drain tile outlet flows or
to measure irrigation furrow flows.

An equation for these circular flumes was developed for a group with the following relative proportions:

where D = the diameter of the pipe.

The equation is limited to the following ranges of application:

The developed expression is

(21.22)

where

The forgoing expressions were derived by computing a series of flume sills and ramps (broad-crested weirs) in a pipe 1 m in
diameter with sill heights between 0.2 D and 0.5 D. The resulting direct computed tables for each sill height were curve-fitted to
generate values for C1 , C2 , and U for each sill height in the 1-m-diameter pipe. These several values forC1 , C2 , and u as
functions of p1 were further curve-fitted to obtain the above expressions. The D2.5 factor represents the Froude-modeling length
ratio applied to generalize the equation to a range that we recommend should not exceed about 10:1 (0.1–10 times the direct-
computed size of D = 1 m).

For the smallest size, deviations at the low flow ranges can be as high as 3 percent compared to direct computer-modeled
tables. For the larger sizes this deviation is less than 1 percent from the direct computed tables.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Example. Because of the relative complexity of applying the above equation to various sizes, an example for a common size in
Enghish units is presented. The diameter is 1 ft and the sill height is D/3:

Convert American measurements to meters:

Then

and

Thus, Eq. (21.22), for this pipe size and sill height becomes

or

where h1 is expressed in meters and Q is in m3 /s. ( Multiply Q by 35.31 to convert to ft3 /s.)

As with the previously discussed broad-crested weirs, the width, bc, is one of the two most important dimensions in the flume.
The other is the head measurement (Bos et al., 1991). For the head measurement, a translocated stilling well is recommended,
Fig. 21.8. This method transfers the upstream depth reading to an easily measured location above the sill at the flume outlet.
The translocated stilling well conveniently references the upstream head to the sill floor without the necessity of accurately
leveling the flume. This makes it truly portable (Bos et al., 1991). The upstream gage reading location should be used only if the
flume can be conveniently, or permanently, leveled.

If the weir is located in a section of pipe, the normal depth,yn, equals the downstream depth y2 . The maximum flow depth for
stable, nonpriming flow is about 0.9 of the pipe diameter. Thus, y2 would need to be less than this by an amount equal to the
needed head drop in the flume. Thus,

(21.23)

Equation 21.23 gives the limits on design to provide for modular flow y1 = y2 + h and to keep the pipe from flowing full y1 = 0.9D
at maximum flow. Here again we use Δh ≈ ΔH. If flow in the pipe is caused by downstream backwater effects (in excess of
normal depth), these criteria should be checked at low flows.

With a weir at the end of a pipe and sufficient overfall, the weir can be lower to keep approach velocities high for better
sediment transport. Preferably, the normal depth, yn, should be larger than or equal to y1 . Hence at maximum flow:

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
(21.24)

All weir and pipe combinations can be considered to be Froude models of each other. This means that rating tables for sizes
other than those shown in the table can be readily developed without the computer model, as long as all dimensions remain
proportional.

21.8. FIELD SIMPLIFIED, EXPEDIENT MEASUREMENT


TECHNIQUES
Obtaining useful design information on an existing flow system in a retrofit situation is frequently a challenge. In open channel
flows, reliance on historical flow levels and the uniform flow equations can sometimes provide the needed information (Chow,
1959). Too frequently, true uniform flow does not occur even in canals that appear to have long straight reaches. Careful field
surveys of the canal slope and the water surface slope over an extended distance may be required in order to calculate the
needed energy slope for use in the required equations. Even then the friction value is usually nebulous. Good current metering
records are valuable for both rivers and large canals. Accurate current metering in small canals is often not suitable. However,
small canals can frequently be temporarily measured with portable flumes and weirs and the information used for permanent
designs.

Figure 21.9 Suggested method to obtain accurate flow profile information for computing the energy
slope of a flowing canal for accurately applying Manning’s formula.

21.8.1. Channel Roughness Measurement and Water Surface Profile


Measurements.
An accurate and quick field technique to obtain a channel flow rate using Manning’s formula, or determine a channel flow area,
or obtain data to calculate energy-line slope for computing the roughness value for Manning’s n is illustrated in Fig. 21.9. The
technique uses a standard surveying level and two static pressure tubes inserted into the canal water at approximately equal
distances upstream and downstream from the surveying instrument site. The static pressure tubes are described in Replogle
(1997), and Bos et al. (1991). The distances are typically 50–100 m. Using equal distances in both directions is good surveying
technique in case the surveying instrument is not in good repair. Stilling wells in the form of small cups placed slightly into the
water surface provide a damped water surface for the survey data. The cups need to be tall enough to survive splash from
velocity impact and surface waves. Manning’s formula (Eq. 21.16) can be used with an estimate of Manning’s n value and the
channel energy line slope, S e (Chow, 1959). As mentioned earlier, Manning’s n value typically is 0.012–0.014 for new concrete
canals in excellent condition, increasing to 0.018 for damaged concrete surfaces, and in excess of 0.030 when heavy grass-like
algae growths are present.

The method has been successfully applied to concrete-lined canals to evaluate channel roughness. For this process, an
independent measurement of the flow rate was obtained with portable broad-crested weirs as described in another section, and
Manning’s equation was solved for n. The channel flow areas and the flow velocities at the two points illustrated inFig. 21.9,

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
could be accurately calculated using the survey data for the water surface. These velocities then provided the velocity head that
was added to the respective flow depths to establish the slope of the energy grade line, S e.

21.8.2. Portable Flow Measuring Flumes


Portable measuring flumes can provide reliable site-specific data that can usually be translated into precise design information
for permanent structures. Portable flow measuring flumes are described in Bos et al., (1991) and in Replogle and Wahlin (1998).
The latter reference describes flumes with adjustable throats that are especially convenient to install in small unlined canals.
They allow accurate measurements with canal ponding of less than 2–3 cm.

The permanent structure design will often be one of the long-throated flumes described in earlier sections. This design process
has been expedited by several design-oriented computer programs, which aid in the selection and installation of these devices.
These have also been described in earlier sections.

21.8.3. Surface Floats


The velocity of surface floats is sometimes used to estimate flows in lined channels. Unless survey data exist to define a
natural channel, the surface velocity times an unknown area of flow produces an unreliable discharge rate. Lined prismatic
channels, after probing to determine sedimentation conditions, are a fair candidate for this method of flow estimation.

Rather than using a single float, multiple floats consisting usually of buoyant ditchbank trash, popcorn, or even numerous ice
cubes, provide a surface float method that is reliable to ±10 percent. Ice cubes are particularly useful in windy conditions. The
recommended procedure is to apply the floating material across the channel about two or more top widths upstream from a
designated point A. Measure the time for the initial front of the floating material to pass from pointA to a point B established
30–50 m downstream. Because of the many particles, the time for the front edge of the floating material to reach point B
should provide the velocity of the fastest surface flow element. This flow element is a function of the channel shape and the
channel roughness. Also, the roughness and the length between stations A and B contribute to the longitudinal dispersion of the
floating material. While this length dispersion pattern has not been established well enough to firmly define Manning’s n, it does
appear to be related.

For field flow estimates, when the channel is rough or large, the average velocity is about 90 percent of the maximum surface
velocity. When the channel is small or smooth, the value approaches 70 percent. While this trend is evident, precise values have
not been established. The current information does support an average value of 80 percent. These values differ from the rod
floats as used and reported in USBR (1997) where the average is represented by range from 66 percent to 80 percent with an
average of 73 percent. However, those floats sampled at least the top quarter of the flow depth as opposed to only the surface
velocity suggested here.

21.8.4. Checking a Flow Profile


Sometimes there is a need to inexpensively check how the velocity profile is behaving near a measuring device, and to check if
measures taken to condition it have been effective. One way to obtain quick and easy results is the rising bubble method.
Trickle irrigation tubing is weighted so that it will stay in a straight line across the bottom of the channel of interest. Pressurized
air or other gas is released at a rather fast rate from the many small holes, Fig. 21.10. The predominant larger bubbles rise
uniformly enough to define an undisturbed water surface area between the line of injection and the predominant emergence.
The smaller bubbles rise more slowly and emerge in the downstream bubble trail. One immediate observation is the symmetry
of the emergence line. A ragged, nonsymmetrical line in a prismatic channel indicates velocity profile distortions.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Figure 21.10 Air bubbles used to check velocity patterns.

Using rising bubbles as a flow measurement method. This same system can be used to measure discharge rate. The discharge
is calculated quite simply by the product of the surface area defined by the emerging bubbles and the release line, multiplied by
the average bubble rise velocity of 0.218 m/s (Herschy, 1985). A limitation of this rising-bubble method is the difficulty of
measuring the surface area accurately, but this method will give good discharge estimates in poorly defined earth channels and
automatically adjusts for both velocity profile and channel shape.

21.8.5. Low-Pressure Pipe Venturi


The Venturi tube, although old compared to the modern sonic meters, is still a viable method of measuring fluid flows. It has
good anticlogging characteristics and is sometimes used on pressure flows of sewage. In low-pressure pipelines associated
with agricultural irrigation (less than 30 m.of pressure head) Venturi tubes were fashioned from standard Polyvinyl chloride
(PVC) pipe fittings for 30.5 cm (12-in) main lines and a throat diameter of 20.3 cm (8 in) (Replogle and Wahlin, 1994). Figure
21.11 shows the basic construction. Standard pipe fittings were used to reduce to 10-in pipe then to the 8-in pipe.Fig. 21.12
shows a method of making suitable manometer taps.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
Figure 21.11 Configuration installed in the field.

Figure 21.12 Pressure tap showing glued-on boss and pipe fitting used in field units.

The basic expression for discharge, Q, is derived from the classical Bernoulli Equation and can be written in a form that is
applicable to round pipes or other conduit shapes as

where Cd discharge coefficient, usually 0.90 ≤ Cd ≤ 0.99, Ap = area of approach piping,,At = area of contracted throat section, G =
gravitational constant (9.81 m/s2 = 32.16 ft/s2 ), α = velocity distribution coefficient (estimated as 1.02), andhp – ht =
Differential head between upstream pipe tapping and throat tapping.

These Venturi meters fashioned from plastic pipe fittings of the kind usually used by the irrigation industry have an expected
accuracy of ±2 percent, not including the errors of the readout method. It is recommended that a throat length of three times the
throat diameter be used for these plastic Venturi meter constructions. Shorter throat lengths appear to cause difficulties in
pressure detection due to flow separation. Longer throat lengths produce excessive headloss. The rate of contraction of the
reducer fittings caused no significant change in Cd , and thus the meter calibration. However, commercial fittings vary in their
contraction angle and the fittings with a 25 contraction angle (measured from pipe centerline) exhibited a greater total head
loss through the meter than those with a less severe 15 contraction rate (18 percent and 25 percent of head reading
respectively). These constructions conform to expected Venturi meter behavior with a discharge coefficient about 2 percent to

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
5 percent less than “standard” Herschel–type Venturi tubes. The discharge coefficient, Cd , can be estimated by the following
empirical equation:

where m1 = 0.964; m2 = 0.0466; m3 = 254,000; and Rn = Reynolds number (based on pipe diameter).

The most important construction factor is the fabrication of the pressure taps and the immediate connections. They should be
drilled with appropriate backing blocks to reduce burrs and with a guide to assure that they are constructed perpendicular to
the pipe wall. A slight rounding of the piezometer tap holes will help reduce burrs. It is recommended that the pressure taps be
installed on the sides of the meter to prevent air bubbles from entering the pressure lines. It is not necessary that the pressure
taps be on the same horizontal line and the meter can be mounted at any angle. Slow-setting PVC cement is recommended to
allow workers sufficient time to uniformly assemble and adjust large pipe parts. For an accepted error of ±2 percent, one
properly constructed pressure tap at each of the two required piping locations is adequate.

21.9. ACKNOWLEDGMENTS
We hereby acknowledge M. G. Bos who encouraged and contributed to a significant cooperative effort with Replogle and
Clemmens to produce the first comprehensive treatment of long-throated flumes. That publications by Bos et al., (1991) formed
the basis for much of the discussion presented on the newer “computable” flumes, including the computer program developed
to compute them (Clemmens et al., 1993).

Much of the remainder of the text for this chapter was condensed from the recent U.S. Bureau of Reclamation publication
(USBR, 1997) to which the current authors contributed significantly, but to which we acknowledge the editorial efforts of Mr.
Russ Dodge, and other Reclamation employees, who undertook much responsibility to bring that publication to fruition.

21.10. REFERENCES
1. Ackers, P., W. R. White, J. A. Perkins, and A. J. M. Harrison, Weirs and Flumes for Flow Measurement, John Wiley & Sons,
New York, 1978.
2. ASME, Fluid Meters, Their Theory and Application, 5th ed., Report of ASME Research Committee on Fluid Meters,
American Society of Mechanical Engineers, New York. 1959.
3. ASME., Fluid Meters, Their Theory and Application, 6th ed., H. S. Bean, ed., Report of ASME Research Committee on Fluid
Meters, American Society of Mechanical Engineers, New York. 1971.
4. ASTM, Annual Book of ASTM Standards, Water Environment Technology, Section 11, Vol. 11.01 Water (1), American
Society for Testing and Materials, 1988.
5. Bos, M. G. ed.,. Discharge Measurement Structures. 3rd rev. ed., Publication 20,. International Institute for Land
Reclamation and Improvement/ILRI, Wageningen, The Netherlands. 1989.
6. Bos, M. G., J. A. Replogle, and A. J. Clemmens, Flow Measuring Flumes for Open Channel Systems, American Society of
Agricultural Engineers, St. Joseph, MI, (republication of 1984 edition by John Wiley & Sons), 1991.
7. Brater, E. F., and H. W. King, Handbook of Hydraulics, 6th ed., McGraw-Hill, New York, 1982.
8. Chow, V. Te, Open Channel Hydraulics, McGraw-Hill, New York, 1959.
9. Clemmens, A. J., M. G. Bos, and Replogle, J. A. “RBC Broad-Crested Weirs for Circular Sewers and Pipes,” Journal of
Hydrology (Netherlands), 68:349-368, 1984; Stout, G. E., and G. H. Davis, eds., The Ven Te Chow Memorial Volume. 1984
10. Clemmens, A. J., M. G., Bos, and Replogle, J. A., FLUME: Design and Calibration of Long-Throated Measuring Flumes (with
computer disk, Version 3.0), Publication No. 54, International Institute for Land Reclamation and Improvement/ILRI,
Wageningen, The Netherlands, 1993.
11. Daugherty, R. L., and A. C. Ingersoll, Fluid Mechanics, 5th ed., McGraw-Hill, New York, 1954.
12. Herschy, R. W., Streamflow Measurement, Elsevier Applied Science Publishers, London and New York, 1985.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.
13. ISO Standard 5167, Measurement of Fluid Flow by Means of Plates, Nozzles and Venturi Tubes Inserted in Circular Cross-
Section Conduits Running Full, ISO 5167-1980(E), Geneva, Switzerland, 1991.
14. ISO, Measurement of Liquid Flows in Open Channels, Handbook No. 15, International Organization for Standardization.
Geneva, Switzerland, 1983.
15. Miller, R. W., Flow Measurement Engineering Handbook, 3rd ed., McGraw-Hill, New York, 1996.
16. Reginald, W., Streamflow Measurement, Elsevier Applied Science Publishers, London and New York, 1985.
17. Replogle, J. A., A. J., Clemmens, and Bos, M. G. “Measuring Irrigation Water,.” in T. A. Howell, K. H. Solomon, and G.
Hoffman eds. Management of Farm Irrigation Systems, American Society of Agricultural Engineers, St. Joseph, MI, 1990.
18. Replogle, J. A., and B. T. Wahlin, “Venturi Meter Construction for Plastic Irrigation Pipelines,” Applied Engineering in
Agriculture,10(1):21–26, 1994.
19. Replogle, J. and B. Wahlin, “Portable and Permanent Flumes with Adjustable Throats,” Irrigation and Drainage Systems,
Vol. 12, No. 1, Kluwer Academic Publishers, Netherlands, págs. 23-34., 1998
20. Replogle, J. A., “Practical Technologies for Irrigation Flow Control and Measurement,” Journal of Irrigation and Drainage
Systems, 11(3):241–259, 1997.
21. Spitzer, D. W., Industrial Flow Measurement. Resources for Measurement and Control Series, Instrument Society of
America, 1990.
22. USBR, Water Measurement Manual, U.S. Department of Interior, Bureau of Reclamation, 3rd ed., Superintendent of
Documents, U.S. Government Printing Office, Washington, DC, 1997.
23. USGS, National Handbook of Recommended Methods of Water-Data Acquisition, U.S. Geological Survey, Office of Water
Data Coordination, Government Printing Office, Washington, DC, 1980.
24. Wahl, T. L., and A. J. Clemmens, Improved Software for Design of Long-Throated Flumes. in Proceedings of the 14th
Technical Conference on Irrigation, Drainage and Flood Control, U.S. Committee on Irrigation and Drainage. Phoenix, AZ,
1998.

© McGraw-Hill Education. All rights reserved. Any use is subject to the Terms of Use, Privacy Notice and copyright information.

You might also like