Flume Deswelion !

Example flume for chapter 8 II Revision12 11

j/ Absolute Distance __ - -
li
I I.j Freeboard Requirementis Expreied as an&sohte Distance

Figure 8.13 Defining freeboard requirements for the upstream channel.

8.8 Flume Design

The user may approach the problem of flume design in two ways:

Automated analysis of alternative designs derived from the user's initial

structure, and potentially some refinement of a selected design with assistance
from the design review report and rating tables; or
Iterative use of the design review report to manually change the design until the
design criteria are satisfied.

Before discussing WinFlume's design tools, we will first review the design criteria
and other issues common to both approaches.

8.8.1 Design criteria

WinFlume uses six design criteria to evaluate the suitability of a given structure for
flow measurement. In the automated design module, four of these are considered
primary criteria that must be satisfied within the design module:

Froude number at maximum discharge 5 0.50,

Freeboard at maximum discharge 2 required freeboard,
Submergence at minimum flow (actual tailwater 1.allowable tailwater), and
Submergence at maximum flow (actual tailwater 5 allowable tailwater).

298 Using the WinFlume Software

In addition, there are two secondary design criteria related to the flow measurement
accuracy of the structure:

Flow measurement accuracy at minimum flow meets design requirement and

Flow measurement accuracy at maximum flow meets design requirement.

These are considered to be secondary criteria because a structure that does not meet
the accuracy requirement can be improved by choosing a better head-measurement
method, which does not affect the design of the structure itself. The design review
report evaluates all six design criteria, without making a distinction between primary
and secondary criteria.

Froude number at maximum discharge

To obtain a reasonably smooth water surface whose elevation can be measured

accurately, the Froude number in the approach channel, Fr,, should not exceed 0.5
over a distance of at least thirty times h , upstream from the structure. If feasible, the
Froude number should be reduced to 0.2 within the approach section to the flume.

Freeboard at maximum discharge

Unless a measuring structure can be located at an existing drop in the canal system,
the installation of a measuring structure in an existing canal will increase the water
level upstream from the structure due to the head loss required to maintain modular
flow conditions. This increased water level will reduce the available freeboard in the
upstream channel. For the reasons discussed in Section 2.3, the freeboard
requirement upstream from a flume or weir can usually be relaxed, but some
freeboard must still be maintained. The user specifies the freeboard requirement on
the Flume Properties, Canal Data & Design Requirements form, and WinFlume
evaluates the design to ensure that this freeboard is maintained at maximum flow.

Submergence (actual versus allowable tailwater)

To maintain modular flow through a measuring structure (i.e., critical depth at the
control section), there must be some head loss through the flume due to the energy
losses caused by friction and the expansion of the flow in the sudden or gradual
transition downstream from the throat. WinFlume computes these required head
losses and uses them to determine the maximum allowable water level in the
tailwater channel. WinFlume’s design review compares the actual tailwater levels
specified by the user (see Section 8.7.7) to the allowable tailwater levels at minimum
and maximum flow. The design review assumes that if the tailwater levels are
acceptable at Qminand Q, they will also be acceptable at intermediate flow rates.
This will be true in nearly all cases, and can be verified when a detailed rating table
is generated. The allowable tailwater level is computed for each flow rate in the
rating tables created by WinFlume.

Chapter 8 299
Required discharge-measurement accuracy

WinFlume evaluates flume and weir designs to ensure that they will meet the
required discharge-measurement accuracy specified by the user (see Section 8.7.8).
The accuracy of any one measurement of discharge is a function of the combined
uncertainties in the Q versus h , rating generated by WinFlume and the uncertainty in
the measurement of the upstream sill-referenced head, h,. The uncertainty in the
rating table is assumed to be *1.9% if the ratio of H,/L is in the range 0.07 to 0.7 and
increases to +4% at H,IL ratios of 0.05 or 1 .O. For H,/L ratios below 0.07 or above
0.7, the rating table uncertainty is (Equation 6.44)

X, +
( 1.9 742( 0.07 -
=i 2).5) H,/L < 0.07

X,=k
( (:
1.9+12.78 L - 0 . 7
,"i H , / L > 0.7

The rating table uncertainty increases rapidly for H , / L values outside of the range
0.05 to 1.O, thereby discouraging the design of flumes or weirs that routinely operate
outside of that range. This is consistent with the results from laboratory testing.
discussed in Section 6.4.1.

For a given flume and a particular upstream head, the difference between the actual
discharge and predicted discharge obtained from the Q versus h , rating table is a
systematic error. However, this systematic error is different at different heads, varies
from flume to flume, varies with construction anomalies, and is otherwise
unpredictable for an individual flow measurement. For this reason, the systematic
rating table error is considered to be a random error when evaluating the combined
uncertainty of the flow measurement.

The relative uncertainty, X h l , in the measurement of h , is the ratio of the expected

error (see Section 8.7.8) for the chosen head measurement method, and the predicted
value of h , for a given flow rate. WinFlume uses Equations 4.1 and 4.2 to calculate
an acceptable value ofX,, and h , required to meet the user's accuracy requirements
at Qminand Qma.r. If the accuracy requirement is not met, WinFlume will suggest
narrowing the width of the control section, which increases h , and reduces X h l ,
thereby improving the accuracy.

8.8.2 Head loss objectives a n d tradeoffs

When designing a new measurement structure, a range of possible designs may

satisfy the six design criteria discussed in Section 8.8.1. This range will be deter-
mined by two limiting design criteria out of the six. One of the design criteria will
establish a lower limit on the amount of contraction required in the control section of
the structure (i.e., the reduction in flow cross-sectional area from the approach

300 Using the WinFlume Software

channel to the control section), while the other will set the upper limit on the amount
of allowable contraction. The maximum allowable contraction will always be
established by the freeboard requirement, while the required minimum contraction
will usually be established by one of the three other primary design criteria:

The Froude number must be less than 0.5 at Q,,

The flume must not be submerged at Qmin,and
The flume must not be submerged at Q,,.

In some cases the measurement accuracy criteria may control the required
contraction in the design, but more often these criteria only determine the shape
andor minimum required width of the control section, not the overall contraction.

The most common case is for the range of possible flume designs to be controlled by
the required freeboard and the allowable tailwater at maximum flow. If we consider
the characteristics of the possible designs falling between these two extremes, we find
that the design having the minimum contraction will have the minimum possible head
loss. This may be desirable when designing a flume to be added to an existing canal
system with minimal available head. However, this design is also the most susceptible
to submergence if 3ctual tailwater levels at the site are higher than those assumed by
the designer. In fact, if the lower contraction limit was established by one of the
submergence criteria, then the actual tailwater and allowable tailwater are the same at
either (Inlinor Q, (whichever flow controlled the design) and any increase in
tailwater levels will cause the flume to be submerged. By contrast, the design having
the maximum amount of contraction will have the greatest head loss of all the designs
and will have no additional freeboard at Q, beyond the required freeboard (see
Section 8.7.9). This design will be the least susceptible to submergence in case of
higher than expected tailwater levels, because the additional contraction causes the
allowable tailwater to be much higher than the actual tailwater. One disadvantage of
the maximum head loss design is that it creates the deepest pool upstream from the
structure and the backwater influence from the structure extends for the largest
distance upstream. The difference between the allowable and actual tailwater levels is
listed in WinFlume’s design evaluation table (see Section 8.8.3) as the submergence
protection. It can be thought of as a safety buffer protecting the flume from
submergence caused by higher than expected tailwater levels.

In addition to the minimum and maximum head loss designs, there are two other
specific head loss objectives that the designer may wish to consider. The first is an
intermediate head loss design, which balances the submergence protection and the
additional freeboard at maximum flow so that they are equal. This design is probably
the preferred design in most cases because it allows for some uncertainty in tailwater
levels without creating unnecessary backwater upstream from the structure. The
other specific head loss objective at a site where there is an existing drop in the
channel bed is to match the head loss across the flume to the bed drop. Such a design
will cause no increase in upstream water level and is thus a good choice in a channel
conveying sediment. WinFlume will identify the design for which the head loss at
Q,, matches the bed drop, if such a design exists (the bed drop must be at least as

Chapter 8 30 1
Figure 8.14 The Design Alternatives form, showing input required on the first tab of the form

large as the minimum required head loss across the structure). If the designer wishes
to also have the head loss match the bed drop at Qmin,this must be accomplished by
trail-and-error modification of the control section shape.

8.8.3 Analyzing alternative designs

Once the user has entered initial dimensions, flume and canal properties, and design
requirements, the design evaluation module of WinFlume can be used to evaluate
alternative designs derived from the initial design. These alternative designs will have
control section shapes similar to the initial design, but with more or less contraction.
The product of the analysis will be a report listing the range of possible designs that
meet the design criteria. The design module only addresses the question of what
control section shape and dimensions are needed to produce a structure that meets the
design criteria. Other refinements of the design are made separately (e.g., adjustments
to the lengths of the approach channel, throat, and transition sections of the structure).

The Flume Design form (Figure 8.14) is opened bv selecting Evaluate Alternative
Designs from the Design menu, or by clicking the button on the toolbar. On the
first tab of the form, the user must specify a method of contraction change and the
length increment by which the contraction should be adjusted to create each
alternative design. The second tab on the form displays the results of the analysis. If
the entered value for the design evaluation increment is increased, fewer acceptable
results will be shown within the range of possible designs. It is recommended to use
an incremental value that corresponds with convenient construction dimensions for
the flume. WinFlume will not accept increments greater than O.O5y,.

302 Using the WinFlume Software

 Methods of control section adjustment
1
Contraction of the flow cross-section in the form of a reduced width or raised sill
height is needed to produce critical-depth flow in the throat section of a flume (see
Section 5.4.1). Starting from the initial design, WinFlume’s design module will further
increase and decrease the control-section contraction by varying the sill height or the
width of the control section. Three different methods for varying the sill height are
available, making a total of four different methods of contraction change for
stationary-crest flumes. For movable weirs the only contraction change option is to
vary the width of the control section (vary side contraction). None of the four
contraction change strategies will vary the side slopes of the control section. There
may be situations where a change in side slope will be needed to find a feasible design;
such a change must be made manually. The design module also will not add or modify
a diverging transition, so if you have difficulty obtaining a satisfactory design, you may
wish to manually add a diverging transition and repeat use of the design module.

The choice of the contraction change method will determine the configuration and
dimensions of the final structure. For most design situations there is a range of
dimensions and configurations that will produce acceptable designs, so the designer
should choose the contraction change method with some consideration for the type of
structure that is desired and that can be conveniently constructed at the site. In some
cases, the designer may have to try more than one contraction method or a different
initial structure design before arriving at an acceptable structure. For example, if you
start with no contraction, simply raising the sill or varying the side contraction from
that starting point may not provide an adequate design, but manually adding a low sill
and then varying the side contraction may give an acceptable solution.

The contraction change methods and some possible applications for each method are
described below, and illustrated in Figure 8.15.

Raise or lower height of sill

This option moves the bottom of the control section vertically while all the other
parts (e.g., side slopes) remain in the same position relative to the approach channel.
If the side slopes of the control section are not vertical, the base width of the section

Raise or Lower Sill Raise or Lower Entire Section

Raise or Lower inner Section Vary Side Contraction

Figure 8.15 Methods of contraction change supported by the design module.

Chapter 8 303
will vary as the sill height is changed. This is the option to use for designing a broad-
crested weir in a lined canal. In some cases, the shape actually changes with this
option. For example, a sill in a U-shape could become a rectangle (i.e., bottom
moved above circular part of the U) or a trapezoid in a circle or in a parabola would
become a sill in a circle or parabola, respectively.

Raise or lower entire section

This option moves the entire control section up or down relative to the approach
channel bottom. The shape and the dimensions of the control section will never
change under this option. Hence, the amount of contraction is changed only by
raising or lowering the sill height,p, @, will not be reduced below zero). This option
is useful if a standard design (or prefabricated) flume is to be placed in a channel.
Such flumes are common for flow-survey work. This option could be used to test the
feasibility of using standard-sized structures at different locations throughout an
irrigation system.

Raise or lower inner section

This option only applies to four shapes: a complex trapezoid, a trapezoid in a circle,
a trapezoid in a U-shape, or a trapezoid in a parabola. Under this contraction-change
strategy, the inner trapezoid (the part that moves up and down) remains intact, with
no changes in its dimensions. The outer shape remains fixed relative to the approach
channel as the inner trapezoid moves up and down. This option is useful for
designing a complex flume inside of an existing channel, where the existing channel
becomes the outer shape, or for designing a trapezoidal control within an existing
channel that is circular, U-shaped, or parabolic. This option allows an efficient
design without disturbing the existing channel section. WinFlume will not reduce the
inner sill height below zero.

Vary side contraction

This option holds the bottom of the control section in a constant position relative to
the approach channel, while the bottom width, b,, changes to provide the proper
amount of contraction. Some shapes do not have a defined bottom width. For those
shapes WinFlume varies the diameter (for the U-shape and the circle) or the focal
length (for the parabola). For the complex trapezoid and the trapezoids inside of
other shapes, the inner bottom width is varied. For a sill inside of a circle, U-shape,
or parabola, the diameter or focal length is varied. For the V-in-a-rectangle, the width
of the rectangle is changed. This option is not allowed for V-shaped control sections,
and this is the only contraction change method that can be used for movable weirs.
WinFlume will not design a structure if the control section is wider than the
approach section.

Choosing the increment of contraction change

WinFlume will evaluate altemative designs at a fixed increment of contraction

change, specified by the user. The contraction change can be specified using several
different length units, and these do not need to match the data input/output units
choices selected for the structure (see Section 5.4.1 and 8.7.3). For example, the

304 Using the WinFlume Software

I
structure might be set to use units of feet for length dimensions, but a contraction
change increment of 1 inch could be chosen. This allows the user to choose a
contraction change increment that will yield convenient construction dimensions.
The contraction change increment should be large enough that WinFlume is not
forced to evaluate too many designs; an increment that might lead to evaluation of 1 O
to 30 designs is probably appropriate. If a very large contraction change increment is
chosen, WinFlume may not find an acceptable design because the range of
acceptable designs might be relatively narrow and is thus missed with a large
contraction change increment. However, in this case, WinFlume will automatically
reduce the contraction change increment in an attempt to find a range of acceptable
designs (see Section 8.8.10).

Results of design analysis - the design evaluation report

When the Review Possible Designs tab on the Flume Design form is selected,
WinFlume will execute an algorithm that generates and evaluates a family of design
alternatives derived from the starting design, using the user’s chosen method and
increment of control section contraction change (see Section 8.8.10 for details of the
algorithm). Figure 8.16 shows the results of the design evaluation as displayed on the
second tab of the design evaluation form. The upper part of the form contains a
spreadsheet that lists all of the evaluated designs that meet the four primary design
criteria (see Section 8.8.l), or in the case where none of the designs meet the criteria,
the spreadsheet lists all designs that were evaluated. The lower part of the form

EVALUATION OF DESIGN C R I T E R I A
Froude at Çmax = 0.151 Maximum a l l o w e d = 0 . 5 0 0
Freeboard at Çmax = 0.436 m Minimum a l l o w e d = 0 . 1 0 3 m
Tailwater at Çmax = 1.500 m Maximum a l l o w e d = 1 . 5 0 2 m
Tailwater at Çmin = 0.800 m MaXmUm a l l o w e d = 1 . 1 9 9 m
Head at Wax = 0.511 m Min for a c c u r a c y = 0.027 m
Head at Çpin = 0.188 m M i n f o r a c c u r a c y = 0.012 m I-

Figure 8.16 The Review Possible Designs form, showing the results of evaluating designs with varying
sill heights. The bottom portion of the form shows an abbreviated design review report for
the design altemative highlighted in the spreadsheet at the top.

Chapter 8 305
 shows an abbreviated design review report for the design alternative that the user is
highlighting in the spreadsheet. Buttons on the form allow selecting one of the
alternative designs as the new current design, or closing the design alternatives form
and returning to the initial design. The leftmost columns of the spreadsheet show the
control section shape and the key length dimension defining the amount of
contraction in each design. Columns to the right indicate whether the design meets
(OK) or does not meet (Not OK) each design criterion. The head loss comment
column will highlight the minimum, maximum, and intermediate head loss designs,
as well as the design whose head loss at Q,, matches the bed drop, if such a design
exists. These comments will not be provided if the design does not also satisfy the
two secondary design criteria for flow measurement accuracy.

By scrolling to the right side of the spreadsheet, additional columns can be seen that
provide more detailed information about each design. Table 8.3 shows a printout of
the complete design evaluation report. Key columns to note in the right half of the
table are the extra freeboard at Q, and the submergence protection. The minimum
head loss design will have zero submergence protection (see Section 8.8.2), while
the maximum head loss design has no extra freeboard beyond the required amount
entered by the user (see Section 8.7.9).

8.8.4 Refining the design after using the design module

The evaluation of design alternatives only varies the contraction in the control
section shape (i.e., the sill height, width, diameter, or focal length of the shape).
After a new design has been selected from the list of alternatives, it may be necessary
to adjust the lengths of the approach channel, converging transition, and control
section. The rating table report or the abbreviated design review shown on the third
tab at the bottom right corner of the main WinFlume screen may be helpful in
determining acceptable lengths for these flume components.

Table 8.3 Design evaluation report printed by WinFlume.

(~~~......DESIGN
.. ... CRITERIA
... ~~~......
>........ Actual Actual Extra
Sill Throat <------- Primary Criteria ---------) <- Secondary -> Head Head Broude Freeboard Submergence Estimated
Control-Se~fron Height Wldfh Froude Freeboard T a l l w a t e r Tallwafer Error Error Loss Loss N u d e r ct hnax Protection Random Error
Shape it ff Number a t Qmax et hnax af Qmin P Qmax P amin Comment ft @ Omax ft fi 1 Qmax/Qmln

Srmpie Trapezoid 4.017 26.111 Oir Ok Ok Ok Ok OX Minimum 0.22 0.13 1 03 0.00 *l.
Simple Tiaperord 4.1 26.3 OL OL Ok OL ox ox --- 0.27 0.13 o 98 0.05 *l.
Slmpie Iraperold 4.2 26.6 OX Ok Ok Ok OX
ox
ok ~~~

---
0.36 0.12 0.89 0.14 *,.
Simple Trapezoid 1.3 26.9 ox Di o* OL Ok 0.45 0.12 0.80 0.23 fl.
Simple Trapezoid 4.4 21.2 Ok OL oi Ok Ot oi --- O 54 0.12 0.72 0.32 fl.
Simple Trapezoid 4.5 27.5 OL OL Ok D i Ok Ok ~~~ 0.63 0.11 0.63 0.40 il.
S"le Trapeiold 1.6 2, 8 Ok ox OL Ok oir OL --- 0.72 0.11 0.54 0.49 *l.
Simple Trapezoid 4 628 27.884 OX OL OL Ok OX OX Intermediate 0.74 O.li O 52 0.52 *I.
Simple Traperoid 4.7 28.1 OL Ok Ok Ok OL Ok 0.81 0.11 o 45 0.18 fl
Simple Trapelold 4.8 28.4 OL Ot ox Ok Ok Ok 0.90 0.11 0.36 0.67 11.91-7.02%
Slmple rraperold 4.9 28.7 OL OL OX OX OL Ok 0.99 o 10 0.27 0.76 *i.
Simple Trapezoid 5. 29. Ok Ok OX Ok Ob ok 1 O8 0.10 0.19 0.85 *l.
s1mpple Trapelold 5.1 29.3 Oh OL oir o* oir OL --- 1.17 0.10 0.10 0.34 fl
S",e Trapelold 5.2 29.6 oir OL Ok o* Ok ox --- i 26 0.10 o o, 1.03 tl.
Simple Trapezoid 5.206 29.619 OL ok OX OL Ok OX M a x i m ~ 1.27 0.09 0.00 1.04 *i.

306 Using the WinFlume Software

 8.8.5 Using the design review reports

Although the design module makes it convenient to quickly identify a workable

flume design, there may be situations in which the designer needs to further refine
the design or wishes to develop a design on their own. The general procedure is to
modify the design or the design requirements and then check the design using the
I main design review report (Review Current Design command on the Design menu)
or the abbreviated design review shown on the third tab in the bottom right comer of
the main WinFlume screen. Errors or warning messages and design suggestions can
be considered and the design revised until the design review report indicates a
satisfactory design. Table 8.4 shows a printed design review report. The structure is
evaluated against each of the six design criteria, and advice, warnings, and error
messages pertaining to Qminand Q, are provided. The design requirements for the
structure are also summarized at the bottom of the report. If fatal errors occur at
minimum or maximum discharge so that hydraulic calculations cannot be completed
(e.g., flow in the approach channel overtops the canal banks or tailwater submerges
the control section so that a valid h , versus Q rating cannot be determined), the
warning and error messages will help the user determine appropriate modifications
that need to be made to the design.

8.8.6 Modifying the flume design t o satisfy the design criteria

When any of the six design criteria are not satisfied, there may be several different
design modifications that can be made in an attempt to satisfy the criteria. To assist
the designer, each of the criteria are discussed below and the modifications to
consider in each case are presented.

Froude number at Q,,

To avoid unpredictable flow conditions in the approach channel (standing waves, for
example), the Froude number in the approach channel must be less than 0.5. For most
channels and flumes, the Froude number increases as discharge increases. Thus, it is
only necessary to check the Froude number at maximum flow. If the approach section
Froude number is too high at the highest possible upstream water depth (zero
freeboard), the approach channel is too small to allow accurate flow measurement
with a flume or weir, and WinFlume will never find an acceptable design. The only
possible way to arrive at an acceptable design in this case is to increase the cross-
sectional area of the approach channel. If you have specified a non-zero required
freeboard, you might find an acceptable design by decreasing the specified freeboard.

If the Froude number in the approach channel is too high, but there is additional
freeboard available in the approach channel, you can reduce the Froude number by
increasing the amount of control section contraction, either by raising the sill or
reducing the width. If the design objective is to minimize the total head loss through
the structure, it should be noted that designs that are limited by the upstream Froude
number cannot achieve the minimum head loss condition (i.e., zero submergence

Chapter 8 307
Table 8.4 Flume design review report.

User: Clemmens/Wahl/Bos/Replogle WinFlume32 - Version 1.05

D:\WinFlume Example.Flm - Revision 3
Example flume for Chapter 8
Printed: 2/12/2001 12:22:54 PM

SUMMARY EVALUATION OF FLUME DESIGN

__.._____.___.___.__..--.------------.---

Design is acceptable, but improvements are also possible.

Three errors or warnings.

EVALUATION OF FLUME DESIGN FOR EACH DESIGN REQUIREMENT

.___..____..__..___________________

Ok. Froude number at Qmax = 0.124 Maximum allowed = 0.500

Ok. Freeboard at Qmax = 1.231 ft Minimum allowed = 0.334 ft

Ok. Tailwater at Qmax = 5.515 ft Maximum allowed = 5.652 ft

Submergence Protection at Qmax = 0.137 ft

Ok . Tailwater at Qmin = 2.724 ft Maximum allowed = 4.554 ft

Submergence Protection at Qmin = 1.830 ft

Ok . Head at Qmax = 1.669 ft Minimum for accuracy = 0.104 ft

Expected discharge measurement error at Qmax = k2.65 %

Ok Head at Qmin = 0.453 ft Minimum for accuracy = 0.039 ft

Expected discharge measurement error at Qmin = f2.01 %

ADVICE, WARNINGS, AND ERROR MESSAGES

__..____.___..___._________________

ERRORS AND WARNINGS AT MINIMUM DISCHARGE:

- Converging section length is too short (ramp is too steep).
Suggest converging ramp length >= 10.50 ft

ERRORS AND WARNINGS AT MAXIMUM DISCHARGE:

- Upstream energy head / control section length exceeds 0.7.
Suggest control length >= 2.43 ft
- Converging section length is too short (ramp is too steep).
Suggest converging ramp length >= 10.50 ft

CONTROL SECTION DATA

______..____.___..__.--.------------..----
Section shape = SIMPLE TRAPEZOID
Bottom width = 26.600 ft
Side slopes = 1.5O:l
Sill Height = 4.200 ft

DESIGN CRITERIA
_______________________________

Structure Type: Stationary Crest

Freeboard design criterion: Freeboard >= 20% of upstream sill-referenced head
Allowable discharge measurement errors for a single measurement:
At minimum discharge: t8 %
At maximum discharge: t4 %
Head detection method: Transducer
Expected measurement error = f0.002000 ft
Design discharges and associated tailwater levels:
Minimum discharge = 25.000 CU. ft/s Minimum tailwater depth = 2.724 ft
Maximum discharge = 193.000 CU. ft/s Maximum tailwater depth = 5.515 ft
Tailwater calculation method: Power curve using 2 Q-H measurements
Q = 25.000 CU. ft/s - - - > y2 = 2.724 ft
Q = 193.000 CU. ft/s - - - > y2 = 5.515 ft

308 Using the WinFlume Software

 protection or actual tailwater equal to allowable tailwater) because meeting the
Froude number criteria requires some additional contraction beyond that needed to
satisfy the submergence criteria. If the objective is minimum head loss and the
designer finds that the design is limited by the approach channel Froude number,
then the approach channel should be enlarged (c.g. widened or the invert lowered).

Freeboard at Q,,

If this criterion is not satisfied, the upstream water level is too high. The cross-
sectional area of the control section is too small, and is thus restricting the flow at
Q,,. The flow area at the control section can be enlarged by reducing the sill height
I
or widening the control section. Since this criterion is only evaluated at Q, it may
I
I be possible to only widen the control section at the top by flattening the side slopes.
In this manner, the flow measurement accuracy at Qmincan be preserved. If the
control section area cannot be enlarged enough to meet the freeboard requirement,
then it may be necessary to raise the maximum allowable water level by decreasing
the required freeboard or increasing the canal depth.

Tailwater at Q,, (flume submergence)

For the flume to function properly as a flow measurement structure, critical depth
must occur at the control section (i.e., modular flow). WinFlume computes the head
loss through the structure at this condition and the corresponding maximum
allowable tailwater depth. If the actual tailwater depth is greater than the allowable
depth at e,,, then there is not enough head available for the flume to operate at
modular flow conditions. The flume will be submerged. Thus, the upstream sill-
referenced head will be influenced by the tailwater level, and the structure will not
measure the flow. To resolve this problem the designer must either increase the
available head at the site by increasing the control section contraction and thereby
increasing the water level upstream of the structure, or the designer must reduce the
head loss through the structure. Thus, three possible solutions are: increase the sill
height, narrow the control section at e,,, or add a downstream ramp to reduce the
total head loss.

Tailwater at Qmin(flume submergence)

This design criterion is similar to that described in the paragraph above, except that
submergence of the flume is occurring at the minimum flow condition. Possible
remedies are: increase the sill height, narrow the control section at Qmin(i.e., reduce
the base width of the control section), or add a downstream ramp.

Flow measurement accuracy at Q,, (required head)

The accuracy of the flow-measuring structure is mainly influenced by the

uncertainty associated with the device used to measure h , , the upstream sill-
referenced head in the approach channel. The user specifies the head-detection
method and its uncertainty and a desired overall accuracy for the flow measurement

Chapter 8 309
structure. The overall accuracy of the structure is a combined function of the rating
table uncertainty (approximately *2%) and the relative uncertainty of head detection
at any given flow rate. WinFlume calculates the minimum upstream head, h,,
required to achieve the desired accuracy for one single measurement at maximum
discharge. If the control section does not constrict the flow enough, this minimum
head will not be obtained and the measurement accuracy will be less than that
desired. To achieve the desired flow measurement accuracy at Q, the possible
remedies are narrowing the control section at e,, (i.e., reducing the top width),
using a more accurate head-detection method, or increasing the allowable
measurement error at Q,u~x. If the H,/L ratio for the structure is greater than 0.7,
increasing the control section length may also increase the accuracy of the structure.

Flow measurement accuracy at Qmin (required head)

This criterion is similar to that described in the previous paragraph, except that the
flow measurement accuracy is evaluated at Qmin.If the criterion is not satisfied,
possible remedies are narrowing the control section at Qmin(i.e., reducing the base
width, which also may require lowering the crest.), using a more accurate head-
detection method, or increasing the allowable measurement error at Qmin.If the H J L
ratio for the structure is less than 0.07, decreasing the control section length may also
increase the accuracy of the structure.

8.8.7 Using the rating table report to refine the design

In the design module and in the design review reports, WinFlume evaluates the six
design criteria at only Qminand Q,. In most cases this is sufficient, since the design
criteria will usually be satisfied at all intermediate flow rates as well. To ensure that
this is the case, a rating table can be generated, and the rightmost column of the
rating table checked to ensure that there are no errors or warnings at intermediate
flows. This may also reveal refinements to the design that need to be made, although
they are not related to the six design criteria, such as recommended lengths of the
approach channel, converging and diverging transitions, and the control section. The
lengths of these flume elements need to be within the recommended ranges to ensure
that the assumption of one-dimensional flow with essentially parallel streamlines is
satisfied at the control section. This requirement must be met to ensure that the rating
tables generated by WinFlume will have the stated accuracy of +2 percent.

8.8.8 Issues when designing structures with compound control sections

The design of structures having compound control sections sometimes presents

unique difficulties. These structures have a control section that is a complex trapezoid
or any of the trapezoidal shapes inside of another section shape. These structures are
intended to operate with flow only in the inner trapezoid shape at Qminso that small
flows can be measured accurately. At Q, the flow fills the outer part of the control
section shape, which is made very wide to minimize the upstream depth increase

3 10 Using the WinFlume Software

caused by the flume at maximum discharge. In the transition zone where flow just
begins to enter the outer portion of the throat shape, the accuracy of these structures is
poor. They should be carehlly designed so that this transition zone corresponds to a
flow rate that does not occur frequently and is not of special significance.

The construction of the converging and diverging transitions of flumes having

compound cross sections also presents unique problems. From the standpoint of the
hydraulic analysis needed to develop the h , versus Q rating, the only requirement for
the transitions is that they be smooth and gradual, without offsets at the upstream and
downstream ends of the throat section. These requirements can be met by a
multitude of different transition designs. If the flume is being constructed from
concrete, the transition surfaces can often be free-formed during construction.

8.8.9 Difficulties finding an acceptable converging transition length

We have stated previously that a converging transition slope of 2.5:l to 4.5:l is

desirable for both the floor and sidewalls of the transition. The converging transition
slope is implied by the transition length entered into the flume bottom-profile screen.
WinFlume evaluates the length of the converging transition when generating rating
tables and design review reports. The program evaluates the length of the converging
transition at both minimum and maximum flow, and for each flow rate bases the
recommended transition length on the maximum of three different measures of the
contraction:

The vertical contraction from the invert of the approach channel to the invert of
the control section,
The horizontal contraction from the side wall of the approach channel to the side
wall of the control section at the elevation of the control section invert, and
The horizontal contraction from the side wall of the approach channel to the side
wall of the control section, at the elevation of the water level in the approach
channel ('y,).

There are some special issues to consider when one or two of these measures of
contraction are dramatically different from the others or when the amount of
contraction changes dramatically between minimum and maximum flow.

To measure large ranges of flow (e.g., QmaxjQmin > 200), V-notched control sections
or other shapes with a very narrow base width are often used (see Section 2.4). A
common problem when designing such flumes is the presence of error message 22
(see Section 8.12), which states that the converging section length is too short (side
contraction is too abrupt). The recommended transition length that eliminates this
error message may seem unreasonably long to the user, since it will often produce a
relatively flat floor ramp in the converging transition. Figure 8.17 illustrates the
problem graphically. The governing contraction amount is the horizontal contraction
at the invert of the control section, which is much larger than the other contractions.
The converging transition length must be 2.5 to 4.5 times this length to eliminate

Chapter 8 31 1
Figure 8.17 Converging transition length is based upon the largest of the contractions shown (heavy
dashed lines) at minimum and maximum flow. When the control section is narrow at the
base, the converging transition length is often governed by the side contraction at the invert
of the control section.

error messages 22 and 23 (see Section 8.12). In this situation, we recommend

making the transition length equal to the recommended amount given in connection
with error 22 (2.5 times the maximum horizontal contraction). This may yield a
relatively flat floor ramp, but it will ensure that there is no flow separation at the
entrance to the throat section.

A related problem can occur when attempting to use an approach channel and
control section of dramatically different shape, such as a rectangular control section
and a trapezoidal approach channel having relatively flat side slopes, as illustrated in
Figure 8.18. At maximum flow the governing contraction length is the horizontal
contraction at the ylmax elevation, but at minimum flow the governing contraction is
either the vertical contraction, or the horizontal contraction at ylmin,which may be
much less than the value at maximum flow. In such a situation it is possible to
receive error message 22 (side contraction too abrupt) at maximum flow. Then, when
the converging transition length is increased, message 11 or 23 (ramp or side
contraction too flat) is displayed, but the error now occurs at minimum flow. It is
even possible to obtain both error messages simultaneously, with a single value of
the transition length.

To resolve this problem, one should recall the reasons for which a 2.5:1 to 4.5:l
transition slope was recommended. Slopes more abrupt than 2.5:l cause flow
separation at the entrance to the throat section, which can significantly affect the

Figure 8.18 When the control section shape is significantly different from the approach channel shape,
the goveming (largest) contraction may be significantly different at minimum and
maximum flow. This can cause conflicting warning messages that the converging transition
is both too short and too long.

3 12 Using the WinFlume Software

 accuracy of the head-discharge rating of the structure. On the other hand, slopes
flatter than 4.5: 1 yield a longer, more expensive structure, but have little effect on the
accuracy of the head-discharge rating. Thus, a transition that is too long is preferred
over one that is too short. As a result, one should lengthen the transition to eliminate
error 22 at maximum flow, and tolerate error 11 or 23 at minimum flow. A case
where this might not be recommended is the situation in which flow measurement
accuracy at maximum flow is relatively unimportant; in such a case one could
tolerate a transition that is too abrupt at maximum flow.

Two other solutions are also available. One is to change the shapes of the control or
approach channel sections to make them more similar, but this is not always practical.
The other solution is to recall that the exact geometry of the converging transition is
not critical to flume performance, as long as the transition is gradual without offsets
or large breaks in slope approaching the throat section (see Section 3.1). Thus, even
though WinFlume only allows the user to specify a single converging transition
length, which defines the angles for both the horizontal transition and the floor ramp
(vertical transition), one can use different transition lengths for the side walls and
floor when the flume is constructed. Examples of this type of construction were
shown in Figure 3.25 and 3.58. When using WinFlume to model a flume constructed
in this manner, the most important consideration is correctly locating the gaging
station. The distance from the gage to the upstream end of the throat (i.e., the sum of
the approach length and the converging transition length shown on the flume bottom
profile) should be the same in WinFlume and in the constructed flume.

8.8.10 Algorithm for finding and evaluating alternative designs

The algorithm used by WinFlume to find and evaluate alternative designs is

essentially a trial-and-error procedure. This is a departure from the FLUME 3.0
program, which attempted to optimize the flume design to meet all six design criteria
while also satisfying a single head loss objective. WinFlume simply presents the
range of designs that satisfy the design criteria, then allows the user to choose a
design that meets a head loss objective or other requirement that the designer may
have for the structure. This design algorithm has the advantage that it is more robust
when working with complex and circular control section shapes; the algorithm used
by FLUME 3.0 at times could not determine which direction to change the control
section contraction to move toward a satisfactory design. WinFlume simply analyzes
all of the possibilities. Although this method is more computationally intensive, this
is not a problem with the computational power of today’s personal computers.

Once the user has chosen a method and an increment of contraction change, the
program begins the design process by reducing the contraction in the initial control
section shape as much as possible (e.g., reducing the sill height to zero if the method
of contraction change is to raise or lower the sill). This becomes the base design for
the structure, and all design alternatives to be evaluated are derived from this base
design. Next, the program determines the maximum and minimum upstream water
levels that an acceptable design can produce at the maximum design flow rate, e,,,.

I Chapter 8 313
The maximum allowable upstream water level, before considering freeboard
requirements, is the top of the upstream channel. The minimum required upstream
water level at Q,, is either that corresponding to an approach channel Froude
number of Fr, = 0.5, or the tailwater level at the site, whichever is higher. Water
levels lower than these would either violate the maximum Froude number criterion
or require an energy gain through the structure, which is of course not possible. For
these maximum and minimum approach channel water levels, the program can
determine the corresponding contraction in the throat section shape required to
produce critical flow in the control section at Q,. A subroutine, using the procedure
described in Section 6.3.3, is used to determine the required contraction for each
case. The flume designs having these maximum and minimum amounts of
contraction become the upper and lower limits for the range of flume designs that
must be analyzed.

Once the range of designs has been bracketed, the program builds and evaluates
designs of “virtual” flumes between the upper and lower contraction limits. The
primary control section dimension for each design will be an even increment of the
contraction-change interval chosen by the user. Flume designs are evaluated against
the four primary design criteria (see Section 8.8.1) to determine the range of
acceptable designs. A bisection search (halving the contraction change increment
until the limiting case is reached) is used to fürther refine the minimum and maximum
amounts of contraction that will yield acceptable designs. The design with the
maximum contraction is the maximum head loss design (zero additional freeboard),
and the design with the minimum contraction is the minimum head loss design (zero
submergence protection), unless the Froude number is the limiting criterion for the
minimum-contraction design. Designs having intermediate head loss (extra freeboard
equal to submergence protection) and head loss equal to the bed drop at the site (if
such a design is possible) are also located using a bisection search technique.

The results of the analysis are presented to the user, as described in Section 8.8.3.
The user may choose to accept one of the designs as the new current design or
discard the results of the analysis. Only designs meeting the four primary design
criteria (freeboard, Froude number, no submergence at Qmin and Q,,) are shown to
the user, unless there are no acceptable designs, in which case all of the evaluated
designs are listed. Designs that meet the four primary criteria, but do not meet one of
the measurement accuracy requirements, may be improved by selecting a better
water level measurement method.

If the contraction-change increment specified by the user is too large, or if the design
criteria are too limiting, the design algorithm may not find any acceptable designs on
the first attempt. In this case, the program searches for two adjacent designs for
which the unsatisfied criteria in each design are satisfied in the adjacent design. This
indicates that with a smaller increment of contraction change, it may be possible to
find an acceptable design between these two designs. If this is the case, the program
will reduce the contraction-change increment by a factor of 10 and repeat the entire
process within this reduced range. This continues until acceptable designs are found
or until the program determines that no acceptable design is possible.

3 14 Using the WinFlume Software