0% found this document useful (0 votes)
35 views47 pages

【Side Channel Spillways Hydraulic Theory, Economic Factors, And Experimental Determination of Losses(Hinds,1926)】

Uploaded by

305773253
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
35 views47 pages

【Side Channel Spillways Hydraulic Theory, Economic Factors, And Experimental Determination of Losses(Hinds,1926)】

Uploaded by

305773253
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 47

AMERICAN SOCIETY OF CIVIL ENGINEERS

INSTITUTED 1852

TRANSACTIONS
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

This Society Is not responsible for any statement made or opinion expressed
in its publications.

Paper No. 1587

SIDE CHANNEL SPILLWAYS:


HYDRAULIC THEORY, ECONOMIC FACTORS, AND
EXPERIMENTAL DETERMINATION OF LOSSES*
BY JuLIAN Hrnns,t M. AM. Soc. 0. E.

WITH DrscussrnN BY M ESSRS. 0. H . HowELL, H. B. M ucKLESTON, I. GUTMANN,


A . J. WILEY, IVAN E. Ho ur,, J. 0. STEVENS, AND J ULIAN Hrnns.

SYNOPSIS

The design of a side channel spillway _involves certain special hydraulic


ind economic factors which, it is thought, have not been previously discussed.
On account of the disturbed conditions of flow, there are large internal fric-
tional losses, the relative amounts of which are largely dependent on the
magnitude and form of the individual installations. Flow can be computed by
Bernoulli's theorem, using an experimentally determined coefficient of loss,
,imilar to that of friction in ordinary flow, but such a coefficient would not be
constant, and would need to be determined separately for many different sets
of conditions. This situation can be avoided by utilizing the law of conserva-
tion of linear momentum for determining the flow. Formulas required for
this purpose are developed.
The question of the proper location of the spillway on the ground is dis-
cussed, and a systematic method of determining the most economic design is
suggested. , . .
Although the fundamental principles utilized are well established i~
physics, the question arose as to the propriety of applying these principles to
NOTE.-Tbe subject of this paper is one of ten selected by the Special Committee· on.
Irrigation Hydraulics for study and resea rch . The Committee, a fter studying the paper, bas
recommended Its publication. (See Progress Report of the Committee, Procooilings, : Am . .Soi:;
J. E., March, 1925, Society Affairs, p. 137).
• Published in P1·oceedings, Am. Soc. C. E., September, 1925. :, , .. ,
t Engr., Bureau of Reclamation, Denver, Colo.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


882 SIDE CHANNEL SPILLWAYS

the conditions under discussion. A series of experiment s was planned and


carried out for the purpose of establishin g the correctness of the assumption
used. The results of experiment s at the Arrowrock Spillway, and of laboratory
experiment s at Bellvue, Colo., are given in the paper. .
The two most important conclusions reached are, as follows:
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

1.-Bernou lli's theorem is not convenient ly applicable, because of lack of


uniformity in the coefficient of loss.
2.-The law of the conservatio n of linear momentum is directly applicable
without an experiment al coefficient, subject on1y to a small correction for
swell in volume due to entrained air and unequal distributio n of velocities.

INTRODUCTION

The proper design of a spillway of the type shown in Fig. 1 involves a


number of more or less novel problems, which, although simple if taken sepa-
' ' '

==ao . .
::::==3000
::::3-0.59

GENERAL PLAN AND PROFILE


OF SIDE CHANNEL SPILLWAY
Scare of Fe et
100 100 200 300

Q= 30 000 c.f.s.
L= 420 It. PROFILE
'n=o.75. ·a= 0.2 5
+ I El. 2690.00
~I
.;
FIG. 1.

rately, form a complicated whole. The ultimate economy of constructio n


depends on the form and dimensions of the spillway channel, its location, both
in plan and elevation, its hydraulic properties, and other factors. · The
hydraulics presents a special problem, due to loss by shock and other irregular-
ities caused by the constant' addition to the flow in the channel of the water
coming in over the spillway crest. Although this type of structure is very
inefficient hydraulical ly, physical conditions sometimes make its use desirable.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE CHANNEL SPILLWAYS 883
It is best suited to spillway discharges of moderate quantities. The determina-
tion of the quantity of flood discharge to be provided for is beyond the scope
of this paper. It will be assumed that the required spillway capacity is known.
The depth of flow over the crest has an important influence on the cost of both
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

the spillway and the dam. Generally, and within certain limits, the cost of the
spillway channel is reduced by a short crest with a deep overflow, but if
the spillway is of the open or uncontrolled type the cost of the dam will be
increased, due to the excessive free-board required. If spillway gates are pro-
vided, their cost increases with their height. The factors involved in this part
of the problem are too many and too varied to permit of a general solution.
The most practical method of solving this problem is to assume a number of
depths of overflow and prepare preliminary designs and cost estimates for the
spillway, "dam, and other related features.
Accuracy of Oomputations.-The quantity of flow to be provided for in the
design of a spillway can be only roughly approximated at best, and extreme
precision in computation is not justified. As the preparation of this paper
progressed, arithmetical simplifications were developed, and the method of
procedure finally recommended differs from that used in some of the typical
designs and the reduction of experimental data. However, all computations
are based on the same fundamental theories, and sufficient checking has been
done to show that the various methods of procedure yield results in sufficiently
close agreement for all practical purposes.
Units of Measure.-Although the equations deduced are independent of any
particular system of measurements, units of 1 ft. and 1 sec. will be used to
avoid confusion of language. The weight of 1 cu. ft. of water will be taken as
a unit of force, to eliminate the necessity of multiplying all forces and
momenta by 62.5 to convert them into pounds.
Symbols.-The letters and characters used in the formulas are, as follows:
A = area of water prism.
a = arbitrary coefficient of x in velocity equations.
b = inflow p@r foot length of weir crest.
d = depth of water in channel.
d M, d x, d V, etc. = derivatives of M, x, V, etc.
g = acceleration due to gravity.
H = head on crest (in weir formulas).
+
H = d y (in channel-flow formulas).
h v = velocity head.
M = momentum.
M., = momentum at up stream of two adjacent sections.
Md= momentum at down stream of two adjacent sections.
L1 M = Md - M.,, or change in momentum over a short length, L1 x .
n = arbitrary exponent of x in velocity equations.
Q = discharge, in cubic feet per second.
Q 1 = discharge at upper of two adjacent sections.
Q2 = discha'rge at lower of two adjacent sections.
T = width of channel at water surface.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


884 SIDE CHANNE L SPILL W .A YS

t = time, in seconds.
V = velocity, in feet per second.
V1 = velocity at upper of two adjacent sections.
V2 = velocity at lower of two adjacent sections.
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

LIV= V 2 - VP or change in velocity in a short distance , LI x.


x = distance along axis of channel.
LI x = distance between consecutive cross-sections of the channel.
y = ordinate to the water surface curve iµ the channel.

THE HYDRAUL IC THEORY

Referrin g to Fig. 1 it will be noted that the flow moves over the crest
in a direction approxim ately at right angles to the axis of the spillway channel.
At the moment it comes in contact with the body of the water already
in the
channel, the incomin g water has an appreciable velocity in a vertical
plane
normal to the channel axis. This transverse velocity is of no assistanc
e in
moving the water along the spillway .channel.

f - · - - · - - · - O v.ed.l.o.w-Cre sl - - - - - - - - - - 1

Lake Level

-
...H_... _____ _
EFFECT OF FORM
OF SURFACE CURVE

Fm. 2.
J
- -
--
----- -
The force producin g axial motion in the channel is derived from the falling
of the particles of water along the water surface curve. The useful fall
of
any particle is measure d by the drop in the water surface below its elevation
at the point at which the particle entered the stream. Fig. 2 represen
ts
diagram maticall y a section along the center line of a side-channel spillway
. The
water surface is assumed to fall along some curve from the point, B,
at the
upper end of the channel to a point, D, opposite the down-stream end
of the
crest. Each particle coming into the channel contribu tes to the producti
on of
velocity and the overcoming of resistanc es between B and D an amount
of
energy represen ted by its effective fall, or the drop in water surface from
the
point at which it enters, to D. Thus, a pa:rticle coming in at B will have
an

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


S]DE CHANNEL SPILLWAYS 885
effective fall .equal to PD, whereas one coming in at D will have no effective
fall. The total applied energy down to the point, D, is equal to that produced
by the entire fl.ow falling through the average drop for all the particles. If the
surface curve from B to D is a straight line, and if the inflow per foot of crest
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

is uniform, the average fall will be one-half the total fall, PD. If the curve
is convexed, upward, as BOD, the average fall will be greater than ½ PD,
and if the curve is concaved, the proportion will be less than½ PD.
Only a part of the average fall is available for the production of velocity
head at D. The remainder is used to overcome frictional and impact resist-
ances. The frictional resistances are relatively small and may be neglected
or estimated by methods used for other cases of variable fl.ow. Impact is of
greater importance.
Application of the Laws of Motion.-Wa ter fl.owing in a channel must
conform to each of the two laws of motion-the law of the conservation of
energy and the law of the conservation of linear momentum. The first law is
commonly expressed in hydraulics in the form of Bernoulli's theorem, which
states that the velocity head plus the static head at any point is equal to the
velocity head plus the static head at any other point, plus or minus the inter-
vening "losses." The word "loss" in this sense refers to the transformati on of
kinetic or static energy into some less available form, such as heat. The
energy is not actually lost in the sense of being destroyed. The law of the
conservation of linear momentum requires that the momentum of any system
of particles can be changed only by the application of an external force, the
amount of change being proportionate to the magnitude and duration of the
force. This law is commonly expressed in the form: "Force is equal to mass
times acceleration. " In hydraulics, it is more convenient to make the equiva-
lent statement that "force is equal to rate of change of momentum with respect
to time."
Neither of these laws is inherently superior to the other, each having its
particular field of usefulness. The results obtained from the proper applica-
tion of the two laws always correspond. In special cases one or the other may
be more readily applicable.
A good example of the special usefulness of the momentum theory is
afforded in calculations for the hydraulic jump iii a straight uniform channel;
where violent surges cause heat losses which make the use of Bernoulli's
theorem impracticabl e. However, the external forces acting on the water, in
the direction of fl.ow, are readily isolated, and the amount of deceleration can
be computed with precision. The only force to be approximated is the friction
between the_water and the channel, and this is usually small. After fl.ow con-
ditions below the jump are computed from the momentum equation the loss
of head can be found from Bernoulli's theorem. A loss of head, in such a case,
is unavoidable, and is caused by the fact that after the fast moving particles
coming into the jump and the slow moving particles below the jump collide
they must eventually move away with equal velocities.
Necessity for Loss of Headl.-The fact that a loss of head is necessary in
such a case may be illustrated by the collision of two simple bodies, such as

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


886 SIDE CHANNEL SPILLWAYS

two billiard balls. In order to eliminate consideration of external forpes the


balls may be assumed to roll or slide without friction over a horizontal surface.
It will also be advantageous first to assume the two balls to move in the same
straight line. At the instant of first contact the centers of the two balls are
moving toward each other. The materials of the balls at the point of contact
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

immediately begin to be compressed. This compression results in a reaction


which accelerates the forward ball and retards the rear ball. The compres-
sion increases until the velocities of the two balls are identical, at which time
considerable energy has been expended in the compression of the balls. I£ the
balls are perfectly elastic this energy is available, through the resilience of
' the balls, to force them apart again. I£ the balls are entirely inelastic no force
will be available to separate them, and they will move together wi._th identical
velocities. As a result, there will be a reduction in the combined kinetic
energy of the two balls equivalent to the elastic energy of the perfectly elastic
balls at the instant of greatest compression. The important conclusion results
that i£ all the energy of impact is absorbed the two balls will move away
together and, conversely, i£ the balls are to move away together all the energy
of impact must be absorbed.
Momentum and Energy Laws Both Required!.-!£ the elasticity is perfect,
the relative velocity of the two balls after impact will be exactly the same as
before impact, whereas i£ they are perfectly in-e lastic the resultant relative
velocity will be zero. With intermediate degrees of elasticity the resultant
relative velocity will be determined by the proportion of the energy of impact
absorbed.
The individual action of the balls after impact may be determined only by
the application of both the laws of conservation of energy and of momentum.
There is only one combination of individual velocities that will satisfy both
the energy and momentum requirements, and that combination is controlled
by the energy absorbed in impact. That part of the energy of impact not
absorbed is utilized in the production of relative velocity between the balls
after impact, that is, the velocity of one ball with respect to the other. The
average velocity can be determined from the equation of momentum, and if
the relative velocity is known, the motion is completely determined. I£ the
relative velocity is zero, the actual and average velocities are identical, and in
this special case only, the factor representing the energy equation may be dis-
pensed with, the law of the conservation of momentum affording a complete
solution.
Impinging Streams of Water.-When two streams of water flowing at
different velocities come together and flow away down a common conduit, the
two systems of particles eventually intermingle and the particles move with
approximately equal individual velocities. Consequently, the energy of impact
must be lost, or changed into some form not readily available for the produc-
tion of velocity. As in the simpler case of the impinging balls, flow for this
special condition iE' completely determined from the momentum equation.
This is because the momentum always- precisely determines the average
velocity, which, in this instance, is also the actual velocity.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


sr,DE CHANNEL SPILL w AYS 887
The term "average velocity" should not be confused with the "mean
velocity" as ordinarily used in hydraulic computations . Average velocity, as
here used, is the arithmetical or algebraic average of the velocities of all the
particles. The mean velocity as ordinarily used is the average velocity over
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

the cross-section of the channel taken by areas. Unless the velocity is uniform
over the entire cross-section of the channel at a given point, these two terms
are not identical. Where the velocities vary, more particles pass through an
area of high velocity than through an equal area of low velocity, and the
average velocity is greater than the mean. The assumption that the velocity
is uniform over a given cross-section introduces an error into all ordinary
hydraulic computations , as this assumption is never strictly true. This factor
is relatively of greater importance for side-channel spillways than for other
more usual types of structures, because velocity variations are more pro-
nounced. However, the relation of the actual and average velocities is too
complicated to be given practical consideration , and the usual assumption of
uniformly distributed velocities will be made.
The effect of this approximatio n on the hydraulics of impinging streams
is interesting. As previously stated, the average velocity is completely and
accurately · determined from the momentum equation. If the velocity is
uniform over the cross-section there will be no relative velocity between
particles and all the energy of impact will be lost. If the velocity is not
uniform over the section there will be some relative velocity, and a part of the
energy of impact will not be lost. However, as shown, the average velocity is
greater than the mean velocity, and water-prism areas and depths computed on
the basis of the known average velocity will be too small. Thus, although for
a given average velocity, the theoretical impact loss computed on the basis
of a uniform distribution of velocities is greater than the actual impact loss,
nevertheless, the theoretical depth computed on the same basis is smaller than
the actual depth.
Hydraulic Formulas for Side-Channe l Flow.-A condition of impact and
shock loss exists at every point along a side-channel spillway from the begin-
ning to the end of a spillway crest, as from B to D in Fig. 2. Flow in such a
case is completely determined by making the momentum after impact equal
to that before impact plus any acceleration due to external forces, subject only
to the approximatio ns mentioned previously. The same result may oe obtained
by placing the energy after impact equal to that before impact minus all the
energy lost in impact. However, equating the momenta will be found to be
more convenient.
Consider conditions at two consecutive sections, a small distance, .tJ x, apart,
somewhere between B and D, Fig. 2. If the weir discharge per unit length
is b, the inflow between the two sections will be b LI x. The velocity and the
discharge at the up-stream section may be designated by V.and Q, respectively,
and at the down-st].'.eam section by V +
.tJ V and Q +
b .tJ x. The momenta at
the two sections, therefore, will be :
Up stream,
QV
- ............ ............ ..... (1)
g

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


888 SIDE CHANNEL SPILLWAYS

Down stream,
Q+bLlx
Md= - - - (V+ LIV) ................. .... (2)
. g
Subtracting Equation (1) from Equation (2):
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

QLIV
,
LI M = --
g
+ -bLlx
g
(V + LI V) ...... ............ (3)

As .d x approaches the limit zero, LIV becomes infinitesimal as compared to V,


and disappears, giving the differential equation:
-dM
dx
= -QclV
- +
g dx
b
V - ........................ (4)
g
The rate of change of momentum with respect to time being V_times the
rate of change with respect to x, we may write:
d M Q d V b V2
d t = V g
dx + -g- . .. ........... ·· · · .... ( 5)
The rate of change of momentum with respect to tim!l is equal to the
accelerating force, which in this case is the component of the weight of the
water acting down the water-surface slope. If the discussion is confined to the
horizontal part of the velocity and if x is taken as the horizontal distance
along the channel, the accelerating force will be Q d y, in which, d y is the
d X d X ,

tangent of the water-surface slope. Substituting. this force for dd~ in Equa-

tion (5), and reducing, gives:


bV2
- gVdV
dy
dx
dx + Q g ........................ (6)
Integrating,

y = glf"'( V dx
dV + Q
b V 2) d x . .... . .. ... .. . .... (7)
O

in which, y is the ordinate to the water-surface curve.


Special Formula for Simple New Designs.-!£ the relations of Q and V
to x are known in a given case Equation (7) can be ·integrated and the form
of the surface curve determined. These relations depend on the form and
dimensions of the channel, and will be algebraically complicated unless the
channel is purposely designed to make them simple. In preparing a new
design there appears to be no objection to choosing a form of channel that
will simplify the computations, or, what is the same thing, choosing the
'hydraulics and computing the dimensions of the structure to correspond.
The inflow per foot of spillway crest for the purpose of design will gen-
erally be uniform, and 'the total · discharge at any section x distant from the
upper end of the crest will be given by an equation of the form:
Q =bx ... ;'......... , ................ . (8)
An equation of the exponential type will be found convenient for ~xpress-
ing the velocity relation, and by properly choosing constants such an equation

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE CHANNEL SPILLWAYS 889
can be made to conform to a wide range of physical requirements . The fol-
lowing form .is suggested: .
v = axn . ...... ,; ............ ........ (9)
In Equation (9), a and n are arbitrary constants and V and x denote,
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

respectively, velocity of flow and distance from the upper end of the crest.
Substituting these values for V and Q in Equation (7), integrating, and
reducing:
a2 (n+ 1)
y = X
2n
• •••••••••••• .• ••• •• •••• (10)
2gn
·Substituting a 2 x 2 n for V2 , Equation (10) is simplified to:
n+l
y = - - h v. . .... ............ . . ...... (11)
n
in which, h v is the theoretical velocity head.

THE PREPARATION OF AN UNRESTRICTED DESIGN

Economic Factors.-A spillway channel is completely determined by Equa-


tions (8) to (11), if its shape and the values of a and n are chosen, The
proper choice of these factors is controlled by economic considerations. The
discussion of this part of ~he problem will be confined to a trapezoidal channel

ECONOMIC
BOTTOM WIDTH ECO!jOMIC
SlOE SLOPE

FIG. 4. FIG. 5.
on a . comparativel y steep hillside, which is a usual case. The conclusions
reached can be revised readily to suit other conditions.
The effect of the shape of the channel on the excavation is illustrated in
Figs. 3 and 4. Safety usually demands that the channel be set well into the
original formation. It may be required that the waterway be entirely in rock.
If the water-surfac e elevation, channel side slopes, area of water prism, and
location of point of outcrop, A, are fixed, it is evident from Fig. 3 that the
excavation is reduced by a narrow bottom width of channel. It is similarly
evident from Fig. 4 that, other things being constant, the side slopes should
be made as steep as feasible. The minimum practical width of bottom will
depend on the equipment to be used for removing the material from the
trench. If the excavation is to be done by machinery, a width of 15 or 20 ft.
may be required. For team work a somewhat narrower base may be used.
The reduction in excavation for extremely narrow widths is not great. The
side slopes should be trimmed to the steepest angle at which the materials
will safely stand.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


890 SIDE CHANNEL SPILLW.AYS

In many cases it will be necessary to line the spillway channel with con-
crete. Other things being constant, the cost of lining, which is an important
item, is least when the bottom width is such that the wetted perimeter is a;
minimum. With steep side slopes this will require an average width of water _
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

prism equal approximately to twice the depth of the water. The bottom of
the channel may be made somewhat narrower without greatly increasing the
amount of lining, but the cost of lining should be considered in the final
selection of channel width.
Figs. 3 and 4 are drawn to represent that part of the channel down stream
from the crest structure, but the same principles apply to the part of the
channel opposite the crest.
With a movable crest of the drum type an additional factor is introduced,
due to the necessity for supporting the crest on a concrete base. It will be
found advantageous to set the channel back into the hill a certain distance
to reduce the quantity of concrete required for this purpose. The distance, z,
in Fig. 5, should be chosen to make the combined cost of the shaded part of the
concrete and the shaded part of the exc~vation a minimum·. After a tentative
bottom profile has been determined, the selection of which will be discussed
later, the correct value of z can be calculated for a number of cross-sections,
and a theoretical plan of the structure plotted on a contour map of the site.
Unless the topography · is unusually regular, it will not be possible to fit a
practicable structure to the computed points with accuracy, but a location
approximating the computed locations may be selected for trial.
After the cross-section of the channel has been selected, the profile is con-
trolled by the values of a and n. Assuming a specified drop in the water
surface from B to D, the effect of varying n is illustrated roughly in Fig. 2.
If n = 0.5 the surface curve will be straight and the channel bottom will fol-
low some such line as BI. If n i& greater than 0.5 the water-surface curve
will be convexed upward, as B OD. If n is exactly 1.0 the surface curve will
be a parabola and the bottom will be a parallel curve, as K H . If n lies
between 0.5 and 1.0 the bottom line will start at B, but if the value of n is
nearly unity, it will drop rapidly nearly to the line, K H, which line it will
cross up stream from H. If n exceeds unity, the bottom line theoretically
drops to an infinite depth at the upper end of the channel, then rises rapidly
approximately to the line, K H, which it crosses before reaching H. These
statements can be readily proved by making a few trial computations from
hypothetical data.
The effect on the profile, of varying both n and a, is shown in Fig. 6. If
the value of n is known or assumed, the greatest economy in excavation at
any point is obtained by making a such that the total dro:Q from tli.e sub-
mergence limit at B to the bottom of the channel at the point considered,
expressed as y +d, is a minimum. Choosing H to represent y + d, the mini-
mum value can be found by an equation, derived as follows:
H - d + y . ... . . . .. .. . - . . .. . .•. .. •. . •• . ... . . . ....... . . (12) _

= d + (n 11) h v (from Equation (11) ) . . .. . ........ (13)

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE CH;ANNEL SPILLWAYS 891
v2 Q2
As h v is equal to and V 2 is equal to .A 2 :
2 g

H = d + (n +n 1) v2
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

2 g

- d+ (n: 1) 2 :~ 2 ................... (14)

l<-- - - - ---Overflow-C rest-----


1
I

N
FIG, 6.

Differentiati ng with respect to d:

~: = 1 - (n: 1) g~3 ~! ···· ............... (15)


dH
Making dd equal to zero, the minimum value of H is found to obtain when :

g A3 = (n +n 1) Q2 d A ............ ........ (16)


dd
d .A
The factor, d d, is equal to T, the top width of the water prism. Inserting

this value and recalling that


2
:~2 = h v, Equation (16) simplifies to:

A
hv = -n - - ........ . . . . . . . .· ....... . .. (17)
n + ·1 2 T
or, from Equation (11) :
A
y = 2 T' ... ............ ............ .. (18)

Having chosen a value for either a or n, the correspondin g value of the


other required to give the greatest economy at a specified point in a given
channel, can be found from these equations.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


892 SIDE CHANNEL SPILLWAYS

The correct combination of values for a and n ,to give the greatest economy
for the entire structure is not so :readily found. A general solution involves
the equation of the surface of the hillside which can not be practically ex-
pressed algebraically. The problem is best solved by trial.
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

Anticipating the results of the study, it may be stated that n is a much


more constant factor than a, and is more easily guessed. A start may be
made, therefore, with an assumed value of n. Having selected the trial value,
choose some station along the channel where the excavation is unavoidably
heavy and assume that this section will be a controlling factor in the cost
of the spillway. Usually, as a ·safety precaution, the channel must be set
well into the hill opposite the end of the dam, with the result that the excava-
tion is heayy near the point, D, Fig. 6. Below some point, N, the channel
need be set into the rock only as far as may be required to afford secure side-
walls. The section at D, therefore, influences the economy of construction
more than any other single section, and the maximum economy of excava-
tion for a given value of n, will be approximated if d +
y is made a mini-
mum at D.
Therefore, unless special conditions indicate that the controlling -section
is elsewhere, find the economic value of y for the point, D, from Equations
· (17) and (18) by trial* and the corresponding value of a from Equation (10).
The velocity and the water-surface curve may then be found from Equations
(9) and (10), or Equation (11). Knowing the velocity and the form of the
channel, the depth of flow and the bottom profile at points above D are
readily obtained. Below D it is economical to have the bottom of the channel
as high as possible without interfering with flow above D. Below N (Figs.
1 and Ci), the channel will usually be steep and flow will occur at a velocity
in excess of the critical velocity.
The velocity at D, determined as outlined, will be less than the critical.
A "control", or point of critical flow, therefore, may be provided at N or
between D and N. The control may be anywhere within this reach, but should
be a sufficient distance below D to permit the necessary change in velocity
to take place gradually. In the examples illustrated, the control is assumed
to be at N. Having located the control, the hydraulic computations may be
completed and the profile platted down stream from B to the point, N. The
grade from D to N may be level, or even reversed.
Comparative Estimates.-As soon as a bottom profile has been completed
the principles of Fig. 5 may be applied to several sections of the spillway
crest, and the resulting channel platted on a topographic map of the site. A
cost estimate for the spillway may then be prepared, which will locate one
point on one of the curves in the left:hand part, (a), of Fig. 7. The correct-
ness of the first trial value of a should be checked by assuming additional
trial values, with the same value of n, and making more cost estimates. The
result will give a number of points from which one of the curves in Fig. 7
• Assume a trial depth, d, and compute the area, A, and the top wi!lth, T , From Equa-
tion (17),findhvandthecorrespondingvelocity, V. Then, if AV is equal to the required dis-
charge, the assumed depth is correct. Otherwise, assume a new depth and repeat the process,
continuing until an agreement is secured. Having thus arrived at the economic depth, the
corresponding value of y is found from Equation· (18).

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE CHANNEL SPILLWAYS 893

(a) may be drawn. The lowest point on this curve will give the correct value
of a for use with the assumed value of n. A new value of n should then be
assumed, and the process repeated, platting other curves as shown in the
diagram, until it is clear that the most suitable combination of a and n has
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

been found. The curve in Fig. 7 (b) is obtained by platting the low points
of the curves in Fig. 7 (a) against their respective values of n.
In the example given it is seen that 0.75 is the best value for n, the cor-
responding value for a being approximately 0.25. This value of a is some-
what larger than the value corresponding to the value of y computed for least
depth at D. In all usual cases, with a lined channel, this will be true, and
trial values of a should never be less than are required to give the minimum
value of H at D unless the estimates show sma1ler values to be necessary.
220
(la) lb
~

r--. I-- n-½


" '-
r---." ""1.0
'
'\ \
rs...,_ 17
"
['... ~"-
-zr; -- _/
1__,..
[\
I\ )/
I/

200
n=½ 0.8 0.9 1.0 0,5 0.75 1.0
n-½ 0.3 0.4 0.5 Value of n
n= ¾ 0.2 0.25 0.3 Note:-These Data apply only
n=1.C\ 0.05 0,06 0.07 to Spillway shown in flg, ..1
Value of ·a
FIG. 7 .-RELATIVE COSTS.

Subm ergence of Crest.-An important matter not yet discussed is the


amount the weir can be submerged at B without seriously reducing the dis-
charge. It has been the gemlral practice heretofore to assm;ne that Herschel's
coefficients for ordinary submerged weirs are approximately applicable. The
accuracy of this assumption will be discussed subsequently in connection with
the experimental data. The economic importance of this point is apparent
from the fact that the bottom profile is obtained by measuring d y down- +
ward from the permissible submergence line. In Figs. 1 and 6 the submerg-
ence is assumed to be such that the inflow will not be reduced, that is,
according to Herschel, about 0.13 of the head on the crest. This small allow-
able submergence is chosen in these examples principally to simplify the
theoretical discussion by making the inflow per foot constant. According to
experiments, submergence reduces the discharge only slowly up to a fairly
considerable depth. Furthermore, in a spillway the back-water falls away
below the point; B. The crest may be submerged at B by one-half or two-
thirds of the head without seriously reducing the total discharge. Equation
(S) and Equations (10) to (18), inclusive, are not strictly applicable to such
a partly submerged condition, as _the inflow is variable. However, -they apply
with sufficient accuracy to any ordinary case, and the results may be checked
by methods to be given subsequently. It may be desirable sometimes to

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


894 SIDE CHANNEL SPILLWAYS

assume a heavy submergence, increasing the depth of spill to compensate for


the reduced discharge factor. The limit to which this process may be car-
ried can be determined by cost estimates.
Allowance for Swell.-In estimating the free-board required to prevent
slopping over the edges of the channel or the top of the lining, allowance
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

must be made for turbulence and "swell" due to entrained air and unequal
distribution of velocities. As previously pointed out the average velocity is
always greater than the mean velocity, which fact is ignored in the computa-
tion of depths and areas. Consequently, the actual depth may be expected
to exceed the computed depth by an amount depending on the extent of longi-
tudinal eddies. The volume will be further increased by air drawn · into the
stream by the infalling water. It is believed that the only information avail-
able on this subject is that given subsequently in connection with the experi-
mental data. The average bulking observed for a small model was about 4%
with a variation fr~m zero to about 10 per cent. Observations on full-sized
structures under normal flow conditions are needed.
COMPUTATIONS FOR A RESTRICTED DESIGN

Hydraulic Fortwulas.-The information already given is intended for use


in the preparation of new designs not hampered b? special limiting condi-
tions. The same fundamental hydraulic principles may be applied to the
analysis of an existing channel, or to the design of a channel which must
conform to certain prescribed dimensions and grades, or where inflow is non-
uniform due to excessive submergence or other cause. Equation (6) can be
rewritten to make it applicable, approximately, to finite values of L1 x. Equa-
tion (3) divided by L1 x gives:
LlM QLJV b
L1 x = g L1 x + g (V + L1 V) .............. ; ... (19)

Multiplying by the average velocity, V + 21 L1 V,

LlM Q .. 1 LlV b , l
~ = g (V + 2 L1 V) L1 x. + g (V + LlV) (V + 2 L1 V) ... . (20)

LlM Lly.
As~ = accelerating force = L1 x times average discharge, Equation (20)
becomes:
Lly 1 Q l LlV b i
LI x (Q + 2 L1 Q) = g (V ~
2 L1 V) L1 x + g ( V + L1 V) (V + 2 L1 V) . .. (21)
From which,

Q V+2LIV LIV b ·
Lly - l1 ) [- + -Q (V + LIV)] Llx . ........ (22)
g ( Q+ - LIQ Llx
2
or:

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE CHANNEL SPILLWAYS 895

or:

L1 y = gQ
2 (V1
(QI
+
+ VQ2)
2) [
L1 V + b VQ2L1 x]
1
.•........•.... (24)
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

in which, Q1 and V1 are the discharge and velocity at the up-stream end of
the reach, and Q2 and V 2 are the same functions at the down-stream end. It
is interesting to note that when b is zero, Q1 = Q2 and Equations (22), (23),
v.22 - v21
and (24) reduce to L1 y = 2
g , which is the energy equation for ordi-

nary variable flow, neglecting friction. Equations (22), (23), and (24) are
identical, except as to form, and may be used interchangeably, as found
convenient.
Location of Gontrol.-These equations are not as cumbersome as might
at first appear, but before they can be applied, a starting place, at which the
velocity is known, must be found. Such a starting place will be located at a
point of control, where the depth passes through the critical point, from
above to below.
If the channel slope between D and N, Fig. 1, is insufficient to support
flow at the critical depth, and below N, more than sufficient, the control will
be located at N, and the computation of the hydraulics may be started from
that point. If the slope between D and N is greater than is required to sup-
port flow at the critical depth, the control will come at D, or at some point
above. Up stream from D there is, in addition to the force of friction, a
resisting force due to impact. This force has the same effect on the forma-
tion of a control as a flattening out of the grade. If the slope of the channel
is insufficient to overcome losses and maintain flow at the critical depth
immediately above D, then th~ control will be at D. If the slope immediately
above D is more than sufficient for maintaining critical flow, then the control
will come at .some point farther up stream, where the slope becomes insufficient
for this purpose.
The actual location of a control above D is complicated by the fact that
the critical depth, the impact resistance slope for critical depth, and the dis-
charge are all variable. If the equation of the critical velocity can be written,
the slope can be found from Equation (6), but this can seldom be done. A
suggested method of attacking the problem is illustrated in Fig. 8 and Tables
1, 2, and 3. First, compute the critical velocities and discharges correspond-
ing to a number of depths, as shown in Table 1. The hydraulic .radii may
also be computed and recorded for use in estimating friction losses. From
Table 1 values of the critical depths arid velocities may be taken, by inter-
,polation, for use in Table 2, which is self-explanatory and gives in Column
(13) the drop in water surface necessary to maintain flow at the critical
depth throughout the full length of the channel. Starting from an arbitrary
water-surface elevation at some station, a profile of the channel, for critical
flow at all points, can. be platted, as shown by dashed lines in Fig. 8. A tan-
gent parallel to the bottom of the actual channel may then be drawn to the
resulting bottom line of the · critical depth channel, and the point of control

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


896 SIDE CHANNEL SPILLWAYS

will be located at the point of tangency. Referring to Fig. 8, it is evident


that the slope required to maintain critical flow to the left of Station 1 64 +
is greater than the actual slope, and to the right it is less, which is the con-
dition necessary to the formation of a control. If more than one point of
tangency is possible, that one giving the lowest position of the tangent will
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

be likely to control. It is possible to have two or more control points with


hydraulic jumps between.
Discharge in Cubic Feet per Second
4000 8000 12 000 16 000
', ...... I .
' ,, I'...: '- , For m oJ.w. ~ . a n Botto m_ o r c t ti ca JAIi
-- -
De i th i oint :s
', I'--.(

100
' ~,
',i :---
-- -- -- -
,~
B
1--::::: 2:::
............
-~ f¾ ._
D' I'¾ """'--
i rContro ---
·on P. T . Sta . H 64
__w~ s.
--
~ ~~~ ~ ~
1TlngeLP, rall e to Ohan e l Floor

ci--- --- -- ~dtto,,,


// '?;,;;,;,; .c::::-
~
-.._
~~
~ .,.,_ ~ ~ ""--
7777,
¾
r--:::::: ~
I»-... ~ ~
11,:
~ er s --- --- --
fa ce
c1;,,
' /7?-, K ~~~~
07,;~1 " 1oo,
40 , /i
~ r,,,~ -- ~~
~
LOCATION OF CONTROL ~
ABOVE D ~~
20
/ 77;
~
0 100 200 300 400
Distance al ong Crest in Feet
Fm. 8 .

The Back-Water Ourve.-Having located a control, the back-water 'curve


may be calculated both ways, as illustrated in Table 3, and as platted in the
lower part of Fig. 8. Equation (23) is used in the upper part of Table 3
and Equation (24) in the lower part, in order to take advantage of as many
constant terms as possible. The accuracy of the computations will depend
on the length and number of subdivisions ·assumed. The number of divisions
used in the examples shown could be increased to advantage.
EXPERIMENTS UPON A MonEL STRUCTURE AT BELLVUE, OoLO.
In the lat.t er part of December, 1923, the U. S. Bureau of Reclamation
constructed. a model spillway at the Bellvue Laborator y of the U. S. Depart-
ment of Agriculture, near Fort Collins, Colo., and conducted a series of
experiments for the purpose of establishing, if possible, the correctness of the
foregoing theory. The experiments were made by the writer under the direc-
tion of J. L. Savage, M. Am. Soc. 0. E., Designing Engineer of the Bureau
of Reclamation. A large part of the observational work was done by Ivan E.
Houk, M. Am. Soc. 0 . E., and Samuel Judd, Engineer, Bureau of Reclama-

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SJ1DE CHANNEL SPILLWA YB 897
tion. Valuable assistance in operating the laboratory was given by R. L.
Parshall, Assoc. M. Am. Soc. C. E., Irrigation Engineer, U. S. Department of
Agriculture.
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

TABLE 1.-COMPUTATIONS FOR CRITICAL DEPTHS.


FoR CHANNEL IN Fro. 8.
(Bottom width, 10 ft.; side slopes, ½ : 1)

DeEtb, in Area, in Topwidt.b, Velocity bead, Critical veloc- Hydraulic


eet, square feet, in feet, A ity, in feet Discharge, radius,
d. A. - T. 2T' per second. Q. in feet •.

2 22 12 0.92 7.68 169 1.52


4 48 14 1.71 10.49 504 2.53
6 78 16 2.44 12.52 978 3.33
8 112 18 3.11 14 .15 1 585 4.01
10 150 20 3.75 15.58 2 330 4.63
12 192 22 4.36 16 .75 3 216 5.22
14 288 24 I
4.96 17.86 4 252 5 .76
16 288 26 5 .54 18.88 5 440 6.29
18 342 28 6.11 19. 82 6 780 6.82
20 400 80 6.67 20 .71 8 284 7.31
22 462 8:1 7.22 21.55 9 960 7 .81
2, 528 34 7,76 22.34 11 800 8 .29
26 598 86 8.31 2-1. 12 18 820 8,77
28 672 38 8.84 28.84 16 020 9.26

TABLE 2.---'COMPUTATIONS FOR LOCATING CONTROL.


FoR· CHANNEL rn Fro. 8.
(Bottom . width, 10 ft.; side slopes, ½: 1.)

bV2 a x bV2+ii x
ax Q Q1+Q2 Ve V1+V2 aV Ql .. a y
X d~ - hf* ay+ht
Q1 - aV+
(1) (2) (3) (4) (5) (6) (7) (8) (9) (IQ) (u) (12) ( 13)
- - - - - - --- - - - - - -- -- --
10
0
10 .. "400· . . . .400° 8 .4 10.0 10.0 ·io:r ······· ·· · ··········· ···· ·· ..., ... 0.03
........
25 15 l 000 1 400 . 6.2 12.5 22 .5 2.5 .. ·iS:s · .. ··· ············ ·
21.3 '<i:2i; 0.04 .. .i: 29 ..
50 25 2 000 8 000 9 .2 14 .9 27.4 2.4 14 .9 17 .3 4.91 0.08 4 .99
100 50 4 000 6 000 13.5 17.6 32.5 2.7 17 .6 19.3 6 .77 0.15 6.92
150 50 6 000 to 000 16.9 19 .3 36.9 1. 7 9 .6 11.3 5 .21 0.15 5.86
200 50 8 000 14 000 19 .7 20.6 . , 89.. 9 13 6.8 8.1 4 .33 0 .15 4.48
250 50 10 000 18 000 22.l 21.6 :: 42 .2 1.0 5.4 . 6.4 3.74 0.15 3 .89
300 50 12 000 22 000 24.2 '22:4. · M.0 0.8 4.5 5.3 3.30 0.15 .3.45
350 50 14 000 26 000 26.2 23.2 45.6 0.8 3.8 4 .6 3.04 0.15 8.19
400 50 16 000 30 000 28.0 23.8 47 .0 0 .6 3.4 4.0 2.78 0.15 2.88
- - - - - - - --- - - - - - - - - -- ---
..... .... .. ..... ... . ..... ...... ........ ... ... .......... j .. .. .. .......... 38.28 1.17 39.45

*hf= f.riction loss, computation not shown .

. The general arrangement and principal dimensions of the Bellvue Labora-


tory and the model spillway are .shown in Fig: 9. The intake to the flume
is adjacent to the head-works of the Jackson Ditch, and is controlled by a
single 4 by 4-ft. sliding gate. Two 2 by 2-ft. waste-gates are available for
additional control, and an 8-in. pipe Jine, with a gate-valve near the weir
pool, may be utilized for the final adjustment of flow.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

·OC
<:.::
oc

TABLE 3.-COMPUT.ATIONS FOR B.ACK-W .ATER CURVE.


FOR CHANNEL IN FIG. 8.
(Bottom width, 10 ft.; side slopes, ½ 1.)
Ul


o ·~
...,~
;;,
<l
a)§
'"·~
"~
-ti
i
..,j
cS
...hD
'" .
s
·g~ t
6i\+
"
O> ~ :,.:
~1 .
::: O>
*
• ~
,,,
"'+ ....~ ,,;
~
~
X; I A.c. ...,o'"'" ol .....
...,,! ri
... .CO> + + l\Oi
.c ;:,
.£ ~ ~I>
.
<I)
~ 1;l a1 Qi
' ::: <l
+ i;:i
~ 6 <l
IX!~
" ""
t/.l,i A < Q i> ~ <l
z
<l ztrj
_<1> <2> ,_ _ _ ,__ , ___ ,___ ,___ ,___ ,___ ,___ ,___ ,___ ,_ _ ,___ ,_____
_ I,__
(3 ) (4) (5) (6) (7) (8) (9) (10) (11) (12 ) (13) (14) ( 15 ) ( 16) (17) ( 18) (19) (20 ) ~
i- - - 1 - - - 1 - - - - 1 - - - , - - - -
Ul
>ti
U4 G.ro 66 .80 17.70 33R.6 6 560 19. 70 .. .. .. .. ....
;;:si is::i2 ....~
~IN 58.70 7.90 74 .70 16 .00 288.0 4 000, 13,89 10 560 0.0118 .88.59 12.61 7.29 0.15 7.44 0.46
7.10 73.90 15 .20 267.5 .. ... 14.95 ...... .. .. .. 84.65 4,75 17.36 7.09 7.24 0.14 ~

~
7.27 74.07 15.37 271.8 .. .. . 14.72 .. .. .. .. ... . 84.42 4,98 17.5U 7.13 7 .28 0.01 0. K.
00 00 W .00 5.00 79.07- 12.87 211.5 · 2 000 9.46 6 000 0.0104 24.18 5.26 14.72 19.98 5.02 0:01 5.09 0.09
5.08 79.15 12.95· 213.5 .. .. . 9.37 .. .. .. .. .... '2 4.09 5.85 20.07 5.01 5.08 0.00 0. K .
•ro •u ~.95
12.~0
1,80
2.06
0.59
80.95
81.21
81.80
11.00
11.26
9.60
170.5
176.0
142,1
1 000
.. .. .
400
5.86
5.68
2.82
3 000
. .. . ..
1 400
0.0104
.. ....
0.0089
15.28
15 .05
8 . 00
3.51
8 .69
2.86
'id1
·s:1i2
12.88
13 .06
11.88
2.03
2 .04
0.86
u:02
0.01
2.05
2, 06
0 .87
0.25
0.01
0 .29
0 . K.
Cl.I

0.87 82,(18 9,88 147.6 .. . .. 2.71 .. .. .. .. .... 8.3U 2.97 11.49 0.86 ll.87 0.00 0 . K.
0 ~ 73.ffl 0.23 82 .31 8.61 A y assumed = 2 hv at x = 10

*
AY- g\Q,+Q•)
+
_ Q, ( v, V2) { A ir
.
+ b VQ,2 AX r = Column (11) X Column (12) X Column (15),, · · · · \24)

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

TABLE 3.-(Continued.)

x. 11.x. ...S§
o -~
""'
01>
~.$
~

".,'
;:;, _

<1
ea
...
·,:
't:~
..,.;-~
<.)

:s >
"'~.,
0
-ci
'5.,.
A
o;
-.,j
.,,i...
-<l
".,... . ·g:,.:
bD
~c,
<.)

i5"'
...,,.-;
.;
I>
~
+
0:
I~
of:C,
~
+
~
~
<1
~I ~I
<1
~O>
"
<1
~c,

+
:::..
"
...
"'
<1
t.
~
....
~

+
"'•
<1
..:
e...
r.1
.;
~
0
z 00
8
~
"' <1
a
( 19) (20)
~
(1) (2) (3) ( 4) (5) ( 6) (7) . (8) (9) ( 10) ( 11 ) (12) (13) (1 4 ) ( 15) ( 16) ( 17) ( 18)

-- -- --- -- --- --- - - - -- - - - - - - --- - - -- - - - - - - -- --- --- z


..... . z
164
200 36
49.10
43.70 -i:io 66.80
62.70
17.70
19 .00
333 .6
870.5
6 560
8 000
19.70
21.59 14 560 0.0171 41:29 1.89 3.55
.....
5.44 a:s2 0 .12
....
s:il-i o: iii ~
4.50 62 .30 18.60 359 .0
··· ··· 22.27 ... ... ...... 41.97
41.90
2.57
2.50
6 .12
6.05
4.40
4 .34
4.52
4.46
0.02
0.01 0 . K.
ts
62 .35 18.65 360.4 22.20
250 50 36 .20
4.45
6.15
6.22
56.20
56.18
20.00
19.98
400.0
397.9
10 000 25.00
25.18
1s·ooo 0.0173 47.20
47.33
2.80
2 .93
'.j:.j.j 7.24
7.37
5 .92
6 .03
0.18 6.10
6.21
0.05
0.01 0. K.
00
>ti
H
0:0110 ',i:iil 6.92 0.29 ts
800 50 28 .70 6.68 49 .50
49.73
20.80
21.08
424 .3
431.4
12 000
......
28.29
27 .81
22 000
...... ...... 53.42
52.94
3.16
2.68 .... . 7.35
6.87
6.68
6.18
0.24
.. ..
- 6.42 0.02
ts

~
6.40
21.04 481.7 27.79 52.92 2.66 6.85 6:16 6.40 0.ill O. K.
350 50 21.20
6.89
6 .51
6.48
49.74
43.23
48 .26
22.03
22 .06
463.0
463.9
i.i·ooo 30.24
80.18
26 000 o:oi3s 58.03
57.97
2.45
2.39
'S:ili 6.42
6.36
6.24
6.18
0.30 6.54
6.48
0.03
0.00 0. K.
400 50 13.50 6 .78
6.92
86 .53
86.84
23.08
22.81
495.5
489.2
iii'ooo 32.29
32.71
30 000
......
o:oitiil
......
62.47
62.89
2.11
2.53
°3:78
. ... .
5.89
6 .31
6.10
6 .58
. 0.35
....
6.45
6.98
0.28
0.01 0. K.
00

·· ····

*II. y= ~\~ 1
: d:~' { v + 11. b vQ: X} = Column (11) x Column (12) X Column (15) ... . .. (25i
thf = friction head, computation~ not shown.

00


Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

~
0
0

u,

, • . I . \ "'1<-nuuuer Hose; ~
~ o· 'u "'I' , r1.-,
,, '• I ' ..u- i • . , ti JI --9-6-~ ',;· 9 G;;~,~i"nize:~rt
- ->j
I ·n
p:j
f<-" · 9•1•-·- - i -·-l-2-JO½-·---r---1·2'-I0½--+--..,..,..-1-2'-10½- -4-·-l2-IO½- -->jf-<-,-·· - -- - - -23-9-------
I>
PLAN z
z
t,J
t<
u,
>,;
H
t<
t<
:;,~
::el
I>
><l
u,
s" ,• s' SECTIONC·C
I 11
r-·-·-t4-o-----1r
G11lvani __ .. •-- -
-Stillln
w
Amii u!:f .'i,,l• ff1i--oi.,
..:g
!_! ! BELLVUE LABORATORY

... ..
r3-0"-j+---·-14·0-•-·-•-+3-0::..I
..rb
f SECTIONB:B
EXPERIMENTAL SPILLWAY
DETAILS AND DIMENSIONS
SECTION A-A
}•'10 . 9

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE - CHANNEL SPILLWAYS 901
A 10-ft. Francis weir for measuring the discharge and baffles for stilling
the flow are located as shown. Fig. 10 is a general view of the laboratory and
Fig. 11 shows the spillway structure.
The spillway crest was of sheet iron rounded to a 9-in. radius, and was
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

approximately 16 ft. long. The trough was of 1-in. planed boards carefully
joined on the edges and made water-tight with fiber cement. The leakage
was negligible.
That part of the structure carrying -the rounded crest, correspondin g to
the part, B D, of Fig. 1, was located above a tight bulkhead across . the test
flume. In order _to simulate conditions from D to N, Fig. 1, a short length
of channel was provided below the end of -the spillway crest. This part of
the channel was located below the tight bulkhead. A second stop-plank bulk-
'head below the end of the channel was available for controlling the submerg-
ence at the outlet. A 2½ by 2½-ft. slide-gate· in the center of this bulkhead
made it possible to control the back-water accurately.
The quantity of water which could be passed over the spillway was limited
by the flow in the river and by the height at which the flow could be delivered
into the flume. Flash-boards were used on top of the diversion dam to increase
the head. On account of the lateness of the season, the river discharge was
affected by ice, and a steady flow was difficult to obtain.
Hook-gauges for reading depths in the spillway channel were placed in
metal stilling-wells anchored to the concrete walls of the test flume. Water-
tight connections from the spillway trough to the gauge-wells were made by
short lengths of garden hose. Nipples for the hose connections were screwed
into holes bored partly through the wooden sides of the trough, reinforcing
cleats being used where required. Smaller holes were bored through to the
inside surface to permit the passage of water. Five trough-gauge s were used,
one at each end of the crest and three at equally spaced intermediate points.
The up-stream gauge was connected to the end of the channel, on the center
line, 8 in. above the floor. The other gauges were connected through the
sides, 8 in. above the floor, as shown in Fig. 9. Two additional hook-gauges
set in wooden stilling-boxes on the wall of the test flume were used for reading
the head on the spillway crest.
Heads on the 10-ft. sharp-crested, rectangular weir were taken with two
hook-gauges in wells outside the weir box, as shown on Fig. 9. Discharges
were read from a large scale curve platted by Mr. Parshall, and are undoubt-
edly more exact than would be required for the present purpose.
Observation_s .-When the structure was put into operation it appeared that
a sufficient amount of air was being entrained to reduce materially the weight
of the flowing water. It was thought, therefore, that the water deptl_is meas-
ured in the stilling-wells might not furnish accurate information as to the
actual depth of flow in the channel, and it was decided to take direct level-
rod measuremen ts of the surface profile. On account of the disturbed condi-
tion of flow, it was not expected that these readings would be accurate, but
later they were found to be consistent and reliable.
The conditions shown in Figs. 12 and 13 and Figs. 14 and 15 are typical,
respectively, for non-submerged and submerged flows. The level-rod was held

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


902 SIDE CHANNEL SPILLWAYS .

opposite each gauge connection and at such elevation that, as far as could be
judged, it was in and out of the .water equal intervals of time. Generally, two
readings were taken for each station, one on top of the boil shown in the
photographs, and one in the depression on the crest side of the channel. In
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

some cases, where the flow was less regular, extra readings were taken near
the middle of each section. Level-rod readings were also depended on for
determining the back-water level in the pool into which the channel discharged.
The program for making the observations was arranged to eliminate, as
far as practicable, personal equation and chance. Tests with varying dis-
charges were run on three settings of the trough, two with a flat bottom slope
and one with a considerable grade, as shown on the profiles, Figs. 16, 17, and
18. Fifteen separate runs were made. Preparation for each run was made
by setting the head-gates and by-passes to approximate the flow required.
As soon as the depth in the gauge-boxes at the Francis weir became con-
stant, it was assumed that steady flow had been established. The level-ro·d
profile of the water surface was then taken as rapidly as possible, after which
the rodman beginning at the up-stream end read all the channel gauges, the
Francis weir gauges, and the down-stream and up-stream spillway crest
gauges in turn, making three complete circuits'. The levelman recorded all
readings. The observers then changed duties and repeated the series, reading
the gauges first and ending with the level-rod readings. The level of the
water in the submergence pool was read only once, after all other readings
were completed. Frost in the connections to the stilling-wells for the spill-
way crest gave considerable trouble, and these gauges could not always be
read. The gauge-wells for the measuring weir were outside the structure and
were more accessible for clearing out ice. It was necessary each morning to
drill the ice out of the connections and to thaw out the wells with hot water
or by b~ilding fires around them.
Comparative Wat er-Surface Profiles .-The test was planned primarily as a
check on a previously devised theory, and the most advantageous method o~
presenting the results seems to be in a series of profiles showing measured
and computed water surfaces. (See Figs. 16, 17, and 18.) The high, low,
and average water surfaces, determined from level-rod · readings, are shown.
Where available, the results of the trough-gauge readings are also platted.
The do~n-stream part of the channel being on a level grade, the location
of the control was not difficult. Where a control existed, it was rourid
theoretically and actually to be close to the outfall end of the channel. In
Runs N os. 3, 9, 10, 12, 14, and 15, the submergence at the outfall was such
that the critical velocity was not attained at any point and no control existed.
In these cases, the measured water surface in the lower end of the channel
was taken as a starting point. ·
Apparent Swell.-The computed profile generally is of the same form as the
average observed profile, lying slightly below it. As noted in the theoretical
discussion, it was expected that the measured depths ~ould exceed the com-
puted depths on account of entrained' air and swell due to the discrepancy
between mean and average vel?cities. Many factors may influence the amount

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

OFFIG . 10.- VIEW


B ELLYUE, COLO., LABORATORY.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

FIG. 1 2.-,-RUN No . 1 . . D ISCHA.RGE . 8.35 S ECOND-FEET. T YPICA.L FOR Fro. 1 3 .-RUN No. 6. DIS CHARGE, 22 .85 SECOND-FEET..'
NON - S UBMERGED FLOW. NON-SUBMERGED F LOW, .

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

FIG. 15.-RUN NO. 7. DISCHARGE, 27.19 SECOND-FEET,


FIG. 14.-RUN No. 3. DISCHARGE, 13.89 SECOND-FEET. TYPICAL FOR SUBMERGED . FLOW.
:,unMERGED FLOW.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

'
E
I I LL UN.,l-,t.J,,,1. y
~ W,, S.a bove1Spillway-t..... y-fW, ~- 1bof.'e SP,lllway
~ 99.5
0 Cres fof Spillwar-.... Ba f) k ROd-...,..1 lcreJ1 of lpmw!av Ga~ge caJj;,e, r ~ t~ '-<'. Ban R0<
'D
.,-
~
;J~JSY ~ '><
Bank Rod Average od W, S. below
"E99.0 Cref l ol Spill hanlle l
-r---= ~ :i::---. ver Rod- A' _l
<
~
~
~
Rivt r Rok j
L--<
JAvera9eRod
..::--
I _
W. iS, beieiw
Ct,a ,,nel\ R ✓

I C lcul,ted ,~~ W. S. below


-~
Calcul'ated~.,,,;, ,.._
I t
t98.5
I -- Channel _l !
1-Tg
~
..." I - I~ ~
.,,!. ~
C
·;98.o t---
C

·~ -
SETTING NO. 1 1
RUN NO . 1
Discharge 8.35 Cu. ft, per Sec.
Crest Not Submerged
'T-
1
~

~
"
SETTING NO . 1,
RUN NO. 2
Discharge 10.3~ Cu. ft. per Sec .
Crest Not Submerged
~ i
"-
'?~
'.3
-
<-
SETTING NO. 1 ,
RUN NO. 3
Discharge 13.83 Cu , Ft, per Sec.
Crest Submerged
I
~-
·c
~

Free Outfall Free Outfa ll ·.;:; Submercml Outfall u

"
~97,5
iii I
I I Floor of Chan~e l.. ._
:'E ~
'I
+--
F!o~rof i~
u 1-1, f l or of Channel i
I I
97.0
0 4 8 12 16 20 24 -. --
28
Distance along Channel in Feet
W.S, above Spillway W. S. \abov~ Spl lway D' '

--
ankRod
100.5
,Gauges

-~--~, "
Cr.e s 0L+~1~f, ....::-r--f ;:- ~- . "'es·- I'....
I I
Crest of S!}i llway
I

Etoo.o CalculatedJ Ba nk Roi:I Calc ulat ed I>,, ~"' '


~
0
.
.,, 99,5
River Rod k~ i:::-<
'V
/ Ave Jage.Rod
1" I W• .S. be) ow
Channel
"'
River Rod
~
~ .... ~
" I Ave'rage Rod
- --
W, S. below
Chann1el

"E
\ ~ ~~ er Rod.....,
~ \ \ / ..;::..,
": \
~

:l 99.0 1 I'("' u ...l.


<
~-
" 98. ;-
~
SETTING NO. 2 1
RUN NO. 4
.j

t-
-
:a:;-
~
SETTING NO.
RUN NO . 5
2,
1
-.).
~ r-- ~
BELLVUE LABORATORY
EXPERIMENTAL SPILLWAY
PROHLES-RUNS 1 TO 5
. -~ I--
Oi.scharge 23.02 Cu . Ft. per Sec.
Crest Sligh tly Submerged
~ ~ t--
Di sc harge 27.01 Cu. Ft, per Sec .
!- t--

!-
~
Crest Submerged
:§ 98. )
Free Outfal l Free Outlall
~-
.
>
I
u-
"
iii 97.
'
!.
~

floo r of Channel, .I flo or of Chan~ek I


97.0
0 4 8 12 16 20 24 28 0 4 8 12 16 20 24 28
Dis tance along Channel in Feet Oistanc.e along Channel in Feet
FIG. 16 .-EXPJ:RIJdENT.lL SPILLWAY, BELLVUJ: LABORATORY, PROFILE S OF RUNS Nos. l TO 5.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

W. $. abfo1e ~pillV,:ay
W• . above Spillway? Iw l ,_abp,.e..$oiJbiay. ~
100, ;
Ga ge Crest 1of sJmwJy CalGuJat~d~~Av~ragJRod;
.....
0
-~-_--., ,,_ i...- Bank Rod Clest di Sp(ll~aJJ 1"'-T"~' Cr~st o1 Spil)way}'I

0
~100, )

"
-,, 99, ;
~ I',,.. K
Calcu la ted /
Avefag e Rod

~ I'---,.___~
' ~ V-
IIIIII ~$1}f!t•~k Rpd SETTING NO. 3 1
RUN NO. 8
Discharge 8,55 Cu , Ft, per Sec.
t
r-::,,; ~lm~od~ H---1...JIAver!ige Rod I I Crest Not Submerged
"
E
~
River Rod
- ~

ff$ ~ '1or1
' 1 free Outfall

gi 99. )
Rl~rRbd..Jf~
.,
<{

Q) SETTING NO, 2, ~·ri •


S. below SETTING NO. 2, Calc'.u lat~~./ l6aube.)'f~.. ,._
~ 98. ; _ - RUN NO, 6 ; hill11ilL RUN NO. 7
Discharge 22.85 Cu. F1. per Sec. 0ischarge27.19 Cu, ft, per Sec 1--. -;:~
.!: - -
I 11tJ I I I I I I I 1n~;·r]
Crest Slightly Submerged ~ Crest Submeq:1ed f--1--l--f-'>'-+-+~
~
'f;::
00
Free Outfall i:r-- n .... _.,
g 98. )
~ r- -~-+--+- C/1
~
~
:,:,
~~
"> 1' .. 1
<,;>: !
~ 97. ;.._ 1-- ,;; f - f -
l-~-4---1- C
+--~ floor of Channel J Floor Pt Chann ' I
, ' ~ \. ~ ~flooqoftjhan1ei,,~l
~
:,..
97. )
0 4 8 12 16 20 24 28 z
Distance along Channel in Feet zt;j
V. s. abo·.le Spillwa~ W. S. above Spillway t"
100. ; 1

- -,...._, Bank Rod -Cre t ot Spill~ay C/1


>d
- ---C rest f-S1illwa~
W. S. below ,._ - verate RJd
H
t"
Rod Ctiannet -i-
.,,._ ~ v-
- tt ,__

---
~100. ) W.,S. below t"

--
....- Bank I Iver Rod
1Average Rod 1 Chan ·et
:oil
Iver Rod
"
0 · -- "TlC .:-:. -- Cale latei-' --~- l.J :>
~ · 99. ; ><1
Giugej Calculated \R.fVer rod Avera!}e anc
I C/1
E
~
!lank Rod SETTING NO, 3 1 t BELLVUE LABORATORY
! 99. )
SETTING NO, 3,
f- RUN NO. 10
Discharge 23.80 Cu. ft, per Sec ' EXPERIMENTAL SPILLWAY
..,- RUN NO. 9 f-
Crest Submerged 0

if" 98. ;
Discharge 10.07 Cu. Ft. per. Sec. Submeraed Outfall
"' PROFILES-RUNS 6 TO 10

.
Crest Not Submerged ,!
s --... Submerged Outfall
"' ~~ ~

g 98. ) '--
00
j~- I'---.. .!.

:,:,
>"
Q)
'--
~ '--
r----..
...,_I'--. r~ !--

iii 97. ; r---..,


'--... Floor of Channel 't Flo r of Channel l
97.0
-- _L _L -28-
0 4 8 12 ; 15 20 24 0 4 8 12 16 20 24 28
Distance along Channel in Feet Distance along Channel in Feet
FIG.17.-EXPERIMENTAL SPILLWAY, BELLVUE LABORATORY, PROFILES OF RUNS Nos. 6 TO 10.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

100.
.r---F=~
Riv r: Rod-:-
1:::--r--,...
~ r:: .,:::_ ~~
'- W. S, "'°ve Split ay
Ave age ft od
-i-.,.~ :---
-r -
w S, apove!Splf w11y
A eraJe-Rdd .I ~~ " ' ... ~pi/lw~y

"
> Calculated¥ Balk Rod ~ui Ro~ · Cres of :;~lllv ay
Crest o Spil way .'\. 8anJc Rod ___!,_ (;;,... ---- ~.J-:..:;,
j100 .o
I'(" C e,t f SP!,tlwa)'/ \ '

"'
Calcu lated--'I ~ \ W.c~i ~~~w R ver •d --,../
w.J!~:A:~ - Rlv1 r Roi:I_J ~ ~ !:'-,
.Cl
/\>-='-' ~ :o: ~ ~ II k.;auges- I - ~ f'- ~ ""'- w.l;nbl~r
,, 99.5
"
E
:,
I/ ' I T I
! "E' SETTING NO. 81
t Ca cula ted
1 -·I/ ~
t l.:\j
-
] 99.0 >-- >-- RUN NO. 12
$ETTI.NG NO. 3, N O\scb.a:rge 29.77 C• . f t. per Sec ·J. SETTI NG NO. 3 1 .i
~- +--~+-- RUN N0. 11 ,. >-- >-- Crest Submerged ~~
,-. ~ +-- RUN NO. 13
Discharge 23,58 Cu , Ft. per Sfc
if" o!!!)!_u :~
98 .5,1--__,_-1-_,Dlschaf1e31.10
Ca , fl, per See~ ..,_ 1,....'g-1--1-- S&Jbr,i H
!f---- ,-.
-~
+-- ~
~ Crei:'i Su~."!~!~ed 9 .._ Crest SMghMv Submerged
Free OutfaJt
~

.: " ~,.... F::::


t-- -
f'-...,.__ :E -'-
0
!~ <-'- u,
~ 98.0
.
:;:; f;;: i I'----- ~>--
>--
I'-----
r+- +--
~
> ·' 1'-- I ' --.._ '7'..
~ 97.5
I--.. ,-__ Flo, of hanj,el , i ~, 'Floo of han ,e1\ l 1, ..._ floo of ha1 nel ~ 1 C
~
I>
97.0
0 4 8 12 16 20 24
-'---"::::-
2J! 21
Oi~tance along Channel in Feet
21
t,j

--- 5l- ~~od !•A~t~W


"'1. W, $. above $p1 Rw&y - n 0o S, above WC t"
100. ;
~ ~c,1,~ = ::- ~ Ba~ Ro, w"i.• low
cfiaan I
611 ges'/ ~tvb.- Rod
•-• ,-- I]
8


u,
>,:I

E
.;l100. )
C st ol Sp ·uwaY_,., I....._ ·,, ~ res.I.of pllt ey..1
A 8laqe Re
~
. Ive, odl-" Average I De
:i:l
~
,0
,, 99.
' u,
"E:,
: 99.0I
SET TlN G NO. 3,
RUN NO. 14
1-:r ~ I--I--
SETTrNG NO. ll,
RUN NO. 16
1-r ' - BELLVUE LABORATORY
<(
~
Discharge 22.42 _Cu. ft, per Sec
. Crest Submerged >-- +-- -.:'iL- ,-. I - - >--
Discharge 21.1 1 Cu.. ft. per Sec
: Crest Submerged
~
EXPERIMENTAL SPILLWAY
~" 98. '
-= "
· S bme ed Outfall
.
~

~- >--
.._
I'-----
Submerged Outfa ll

~I-- I--
. .
PROFILES-RUNS 11 TO 15

m I"- •;'.t-- >-- :i:


~ 98.0
.
>
I"-
'-
;;;_
I
I'-----
--- 1'..
;zl- L -
I
2 97.5
"' -.. I" r-...F oor I Channel'\__ J. '-- ..__ Flolxot Cha "'' I
97.0 - ··"'--- (0
20 24 28 0 ·4 8. - ,2 f6 -20 24 28
0 4 8 )2 16
[;)lsta ne:8 a long Channel in Feet Di st an ce along Cha nnel i n Feet ""'
""'
FIG. 1 8 .-lllXPBlll><BNTAL S PILLWAY , B ELLVUE L ABORATORY. PROFILE S OF RUNB No s . 1 1 TO 15.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


912 SIDE CHANNEL SPILLWAYS

of swell. The volume of air taken into the stream probably increases with the
quantity and £all of the inflow. The proportion of this air held in suspension
probably depends on the volume and velocity of the channel stream. The
velocity swell is probably influenced by the same conditions. The data avail-
able are neither sufficiently accurate nor sufficiently extensive for developing a
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

law for the determin.ation of the increase in volume due to these causes. The
apparent swell for the :fifteen runs made at Bellvue is shown ·graphically in
Fig. 19. The points on the graphs were determined by comparing the channel
area for the computed depth with that for the average depth from rod read-
ings. Small uncertainties in the rod readings render the results erratic, but
there is a general tendency to substantiate the expected results. For example,
the weir crest is unsubmerged in Runs Nos. 1 and 8, and the channel outfall is
free, allowing high velocities and small storage volume. As might be expected,
20

-- -crLt n1 subit>rgj I
18
APPARENT SWELL IN VOLUME
Attributed to Entrained Air

--- - itly ] "
-- ;J bmerged
slig
SC
/
/" /
--
16
and Non-uniform Flow V
14
· Runs 1 to 15 inclusive --/- -..t.
, -
l' I /' /

12 ' '
oi
~
I/
,,, I
// ~I
v- ---
,'3

---- --~--
"'10 __ ...
""' '"6' -- 7 -
::t:
~ ' --
"i ::~--...: :~ --7· -----
C 8
- - ---- --12
v--- ~ ---;;
F-"':,.c-: ·l-,-- 7-
_1. ---~,
,, !\q.,''
0 1--;:: :,~
'\ /~. .,<:i""
1~,.......
a. 6

4
~ ..-:··- ' ~~
i-4- -~ ..,
.. ____
~
.. 2 -
-a
1

-- ,...2- :.-
- - / -a-
/ -
.n:::.::
s_~

'"""--- ---- -------~o
~----
,...,, v
,, c.- ;.:-;;'. ~--9.
10·· - -
',, ,7.,,,,.. .. ·
- _,,
__,..~
~i~

-
'<-5. ~>,<;,\
·u-- -....,..,. 7.;,~ ....- ;..:-"rs-- --- :,\···
;,.,t'- \'3' '
---- ;-- -" >-
.. ? i--• • • -/--~
2
'l-;;,
--- -.z-:;
14
--- ~
~9.=
\, ..,,,
--- -\..-; -c."
·- ~:3
;,,-- -9
---- ·-- i -. ---,.,
~k-14-
'
-:,,.-:..- -- ~-
0

i:::.- ..-10·· ----


i--- '.>. ...,
';
.fa ..............
2
:::,/ ----,''
4
0 4 8 12 16
Distance ~long Crest in Feet
Fro. 19. '

the swell is comparatively high. In Runs Nos. 2, 4, 6, and 13, the crest is
slightly submerged and the outfall is free. At Station 16, Rans Nos. 2 and 4
show fairly high air contents. Otherwise, these four runs show only a moderate
swell. Runs Nos. 5, 7, and 11 show a heavily submerged crest with free outfall.
Runs Nos. 5 and 11 show moderate and consistent swells. Run No. 7 is con-
sistent, but higher than might be expected. The crest in Run No. 9 is unsub-
merged, but the outfall is obstructed so that channel velocities are low. The
swell is small. In Run No. 10, both the crest and the outfall are slightly sub-
merged, and the swell is small, Runs Nos. 3, 12, 14, and 15 are fairly heavily
submerged, .both at the crest and the outfall, and, in general, show small or

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE CHANNEL SPILLWAYS 913

moderate swells. Run No. 13, which probably· approximates usual design
conditions as closely ,as any of the runs, shows an average sweH of about 5 per
cent. The effect of the scale of the structure on the swell is undetermined,
and the test results should be applied to actual design with caution.
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

A Peculiar Air Oondition.-An interesting phenomenon was discovered in


studying air conditions. A ''rope" of air varying in size, but about· 2 in. in
diameter, was formed approximately at the gravity axis of the water prism,
and remained there throughout · all the tests, except perhaps in cases of deep
submergence. This "rope'' ended squarely against the up-stream end · of the
channel. It increased slightly in diameter and became less distinct toward
the down-stream end. If broken up with a hoe, it immediately reformed.
Discharge and Submergence Coefficients for Orest.-It was not anticipated
in planning the tests that any valuable data would be obtained on the
discharge coefficient of the rounded metal crest, and no .special · attempt
was made to obtain favorable. approach conditions. Nevertheless, , the results
roughly check, in an interesting way, a theory of compound weir discharge
previously proposed. According to Bazin's· measurements* the lower nappe
of the discharge over a sharp-crested weir rises a'bove the crest elevation by
an amount equal to 0.112 times the head on the crest. If a crest having the
exact shape of the nappe is put under the falling sheet, it 'may be assumed
that the discharge will n:ot .be altered. However, the summit of the new
crest will be above the original crest level by the previously mentioned per-
centage of the head on the crest. H for convenience it is desired to express
the discha~ge in terms of the depth on the new crest, the discharge coefficient

must be multiplied by ( -~ _
1 11 2
/=
1.195. This factor should apply to ;he
coefficient of any discharge formula in which the head appears as the three-
halves power, and is theoretically strictly applicable only to a perfectly shaped
crest. The crest can be shaped to :fit only one flow. For smaller flows, the
conditions of a flat-topped weir are approached, and the discharge coefficient
decreases. As the depth approaches zero, the crest becomes relatively very
flat and the weir tends to act as a control section. The discharge over a
control is:
(J = : H' ✓: g = 3.09 HH .........-.. . .... . . (25)

The coefficient in Equation (25) is 0.93 that of the Francis coefficient, 3.33,
for suppressed sharp-crested weirs.
In estimating the flow over compound weirs, the use of multipliers to be
applied to results from tables for sharp-crested weirs is very convenient. The
highest multiplier, if the flow does not clear the crest, is 1.195, and the lowest
for a very flat crest, with a rounded approach, is 0.93. Usually, the flat type
weir is subject to a transition loss and the multiplier can be less than 0.93 if
the approach is not properly rounded. Multipliers for a number of crest
forms· have been published by the U; S. Bureau of Reclamation.+ The highest
• See Parker"s "Control of Water", p. 400,
t "Hydraulic and Excavation Tables'"., p. 96.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


914 SIDE CHANNEL SPILLWAYS

value shown is 1.169. Crest multipliers for the test structure are platted in
Fig. 20 (a) over depths as abscissa, no allowance being made for- submergence,
The figures·: to the left of the hyphen at ea:ch point designate the number
of the run·· and those to the right show the percentage of submergence at
Station 0 ~n the basis of the computed depths. ·It will be noted that the
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

points for· zero submergence fall . quite regularly · along a curve joining the
theoretical points at the two extremes. The :submerged points are distrib-
uted in a fairly regular way, but are insufficient in number to permit the
derivation of a .law governing the effect of submergence.
It was eX.pected that the law of submerged discharge would difier materially
from that fo.r ai··Sharpccrested weir. In ·order ·to 1 test this assumption the
theoretical discharges were recomputed, making.allowance for the submergence
due to the computed water-surface profile in the channel, in accordance with
Herschel's coefficients.*· The results, are shown in Fig. 20 (b)~ The figures
designate the nmnber of the runs; The poi:ri.t s follow a definite line, indicating
that the; multipliers drop away ·from the theoretical . maximu:rn toward the
control section value as the head on the crest decreases. There is no char-
acteristic difference between submerged and .unsubmerged runs. Run No. 15
is the only one that is seriously erratic. • All multipliers are based on Bazin's
formula for suppressed weirs, with allowance for velocity of approach, except
for the lower. theoretical point. ·Bazin's fo:rmula ..does not e:xtendi to a zero
d,epth. Un#l more Qomplete data are a;vailable, it appears that a multiplier
of 1,15 may, be f'jafely ,used ,fQr a correctly shaped .crest, and that Herschel's
coefficients fo~ subme,~gence a:pp~y with .su:(licient .accuracy tq spillway con-
ditions.
1.25 r-~r-~--~--~--~---r~~~--~-.--~--,---,-~T-~~--,-~

0' 85o;;---L....C.-o:::-.';:-2--...L--o::-l.-:-4__.J,_
· -·...,
oo::l.G::---L---,oJ..s=--.....l.---=o.L,
.2,..-_L__-=-
o.L4__·-L
' __.J
o.-6--..L__Jo.s
.. Av~rage H ead on Qres:t in Feet,/rom Compyted Profil e _
' FIG. ,20;.,-c~·EST M!!L';IPLIEU AND EFFECT .OF .SUBMERGENCE: .

Qharacteristic ,t uns are :shown on Figs. 21; 22, 23, and 24. The effect of
submergence on the turbulence •of flow is illus.t rated in Figs. 21 and 23, Runs
13 and 15;: respectively. ·Conditions in these t wo views are approximately the
same, except as to submergen ce at the outfall.
• Se~ Merrim~n·s "'l're!ltiBe on Hyd·r~uJics.,"

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

FIG. 21.-RuN No. 13. DISCHARGE, 23.58 Si;coND-FEET, PRACTICALLY No SUBMERGENCE.

FIG. 22.-RuN No. 14. DrscHARGE, 22,42 · sEcoNn - FEE'l'. MoDERA'mLY sunMEHGEu FLow.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

FIG. 23 .-RUN No. 15. DISCHARGE, · 21.11 SECOND-FEE~'. HEAVILY SUBMRRGED.

FIG. 24.-RUN No. 12. DISCHARGE, 29.77 SECOND-FEET. SHOWING CROSS-FL(?W BE~9W
CREST FQI\ (:E;RTAp,r S,AGJ;;~, .

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Stl)E CitA:NNEt SPILt WA YB

EXPERIMENTS AT ARROWROCK DAM

The 1923 Overfiow.-A spillway structure of adequate capacity is rarely


required to carry the full discharge for 'which it is designed. For that reason
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

tests on full-sized installations are difficult to obtain. In the spring of 1923,


the Arrowrock Reservoir, on the Boise Irrigation Project, in Idaho, over-
topped the spillway by about 10 000 sec-ft. The crest of this spillway is
a movable type and might have been lowered to obtain full spillway dis-
charge. However, such a discharge is in excess of any flood ever recorded in
the river, even before the construction of the reservoir, and would be considered
more or less of a calamity. The designed capacity is 40 000 sec-ft., or about
four times the 1923 overflow. The crest gates were not fully lowered, but the
water .spilled over them in a partly raised position; The resulting conditions
were very different from those contemplated in the foregoing discussion,
especially in regard to height of fall, turbulence, and air content. However,
it was .considered desirable t~ make such rough observations as were possible
with simple equipment.
Preparations and Equipment. -The work was in charge of W. G. Steward,
Assistant Engineer, U. S. Bureau of R eclamation. Measurement s on two
different dates were made of the water surface profile in the channel, of the
draw-down to the weir, of the upper napp_e of the overfalling sheet, and. of the
s~rface velocities in the spillway channel.
The general arrangement and dimensions . of the spillway are shown in
}tig. 25. The crest is divided by piers into six se_ctions, each controlled by
a 62, by 6-ft. drum-gate. As shown · on the diagram the gate crests were
approximatel y 1 ft. below their maximum height at the time the tests were
made.
Before flood stage was reached the gauges were painted on the . side of the
channel opposite the crest to measure any flow to a depth of 12 ft. .The gauges .
were· locat ed half-way between Piers 1 and 2, 2 and 3, opposite Piers 3, 4, 5,
6, and 7, and at 100-ft. intervals below Pier 7. Opposite Piers 1 and 2 there
were heayy deposits of sand and gravel, washed in from the hill above, which
prevented the painting of the gaug~s at these points; Wires were stretched
across the ohannel from Piers 1, 2, 3, 4, 13, and 7 for .the purpose of running
out a trolley to which was attached a we1ght that could be lowered by a cord
to determine the distance to the water surface. A transit was used with the
weight and the results were satisfactory, but the work was slow. The wave
action on th~ land bank of the cha~nel was measured partly with a transit and
partly by gauge readings. One of the gauges was completely submerged and
the others were covered part of the time. The elevation of the water surface
on the reservoir side of the channel was determined by a weight suspended
over the parapet wall, making allowance for the inclination of suspension.
The curvatur.e of the .water surface above the weir crest was measured down
from the tops of the piers, and the curvature of the nappe of the falling water
was measured by an extension lever with a movable weight suspended near the
end.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

tO
~
0

· 3220 -- '-- -· ... ,.. .,.,

32101 1-- - 11 ·=r-1 1- --111 ----r--11:::::q--11 - - :::- ,

_;32001 11::Y~. I I -····-····--r····, -, -- ····1 :::-, SECTION A· A


Cl) u:,

4, 31901-_ _j~:31~~~~ <;~~~~~..., ....:..-f:...__ ---+--L-,--- :----~7--'--- -T~---l---:--


---7-7 ~
1
.=
c
0 Ci
~ 3180 Hydraulic Jump ~Approximate Localion ) iil
~
~ Center Line and Far Side of Channel Average
z
..jjj 3170 -- =·=-=·----=~=-- -=="'.!."" --···-"':':":':'":. .. ..·:.-=-·---:. - · Bank Side of C annel z
t,J
. {. Calculated Water Surf:ce ----- t'
3160 ···-, -, ·-·--
:;-_.:;::·;-;-:~,~--~.:::~'.:'\ :'.·: :.:-~:~ . . . ~.;. ·. ~ . ~:,. . ;·::-"-, • "" ----~;
':. "'"0....
3220 A ....... ,,,. (',.,,.,.1 c1 .... r- &♦ .... t ... i:- 'l?1n "·••• c: ... 'l'lnn n,1 t'
t'
1- -- 11-=r - 11-- - 111·-=r-- 11-rr-11::v- - 11 :il
32101
-~
.; 3200, - ,~~B•~Chan nel I -· I . .,., ..,. ., .. I · I 1- - · I I 1
~ 3190 . ,.,.,.. , .: ~ ·:~·-_. -~""'' '" . ,,,, - . .
I •.••• , .........., I ·I I I I I

r- "l . l
- /""•----··
t::--+,,..,,____

~~-:=
!''"'&,,.,,-,;_
. ·-t
--~., ,31sol
- •
I

" I:
Appro,umate Localio:n)

t
>
i J 31701 I ____ __ _ ~- - _ _ ___ "'.'~s:::: -:>:jrs:_
--------~~-r~--------~:,• - ·:__;~::~~~~~N:
i~
a . 2uc.;;~ ~~~~:~AV

i
:.:-~~..<---
3160
i
Sta.o+oo
J"une

1+00
12, 1923

2+00 3+00 4+oo


FIG. 211.
-~~·:.:.~~~i- ".-":J1;;:;~,E~~!\!,i;i\:.:
5+oo fi+OO 1+00

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE CHANNEL SPILLWAYS 921
The elevation of the movable crests was found by measuring down from the
tops of the piers. The velocity of the water in the channel was determined by
surface floats consisting of pieces of driftwood or bundles of kindling. The
velocity of approach to the weir crest was measured with a current meter.
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

All measurements were made by Mr. Steward, and a report giving in detail
the methods used is on file in the office of the Chief Engineer of the U. S.
Bureau of Reclamation at Denver, Colo.
Comparative Profiles.-The observed water surface profiles, in comparison
with the computed depths, are shown on Fig. 25. The theoretical profiles shown
were not calculated from Equations (23) and (24), but from a previously
devised more complicated formula based on the same theory. On May 26,
1923, the water surfaces in the channel were measured along the land bank and
along the center line. The agreement between the observed center-line profile
and the computed profile is very close; closer, in fact, than would be expected
in view of the extreme turbulence. On June 12, ·1923, the agreement is not
so close, the observed profile coming well above the computed one at all points
above the down-stream end of the crest. Below the end of the crest, the com-
puted and observed depths agree quite closely for both sets of observations.
There is an apparent discrepancy in the observed depth for the two dates.
The discharge on May 26 was 10 380 sec-ft., and that on June 12, 9 500 sec-ft.,
yet the observed depths were greater on the latter date. It is probable that the
second set of measurements is more reliable. According to Mr. Stewa.rd, in
the .first set the water-surface elevations in the channel were obtained by
lowering the weight from the cross-wire to what appeared to be the average
surface elevation and then taking the angle to the weight with a transit. There
was, he states, a possibility of error in this method due to the continual splash-
ing and wave motion which caused the weight to swing if lowered too far. The
weight was more or less hidden by the spray. The second set of measurements
was made more cai·efully, advantage being taken of the experience gained in
the first trial. The distance and angle to the weight were measured while the
weight was at a sufficient height above the wate1· surface to be out of the
spray. The weight was then lowered to what appeared to be the mean eleva-
. tion of the water surface by paying out additional cord, the length of cord paid
out being measured. It will be realized that great accuracy was not possible
in either case. An idea of the turbulent conditions of fl.ow is ·afforded by Figs.
26, 27, and 28.
In computing the theoretical water surface profiles it was necessary to
proceed as in the case illustrated in Fig. 8. The control points were near the
upper end of the channel. An interesting theoretical situation appears to
exist near the down°stream end of the crest in both experiments. It is not
possible to follow the fl.ow through to the end of the inflow section without a
hydraulic jump. The theoretical jump apparently lies within the observed
water prism. There are many complicating influences at this part of the
channel, and any attempt to make accurate computations appears useless.
Among other things the horizontal curvature of the spillway channel gives
the inflow a component velocity in the direction of the channel, and, further-
more, a complete admixture of the last incoming water with channel flow

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


922 SIDE CHANNEL SPILLWAYS

probably is not accomplished for some distance. The approximate locations


<>fthe theoretical jumps are shown on Fig. 25.
Apparent Swell.-At the time the tests were planned it was thought that
by observing velocities of flow in the channel an estimate of the volume of
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

entrained air could be obtained. The only available method of measuring


these velocities was. by means of surface floats. The velocity being variable,
it was necessary to time the floats at numerous points along the channel. The
number of observers was not sufficient to permit the timing of a single float
at all required points. The time of passage from the upper end of the channel
to each of the various stations was determined separately by different-floats,
and the velocities for the various reaches were found by taking differences of
these time -intervals. No great' dependence can be placed on the results. Even
if the surface velocity along a given line were definitely known, it would give
very little idea of the actual average velocity. A comparison of the mean
velocity computed from observed depths with the surface velocity from floats
gave apparent air contents of from - 4% to +
130% of the volume of water.
The variations were erratic.
The results are shown in _Table 4, as is also the swell obtained from a
comparison of the observed and computed water prism areas. It is believed
that the percentages of swell shown are higher than need be provided for in
design.
'
TABLE 4.-AIR CONTENT OR SWELL, ARROWROCK SPILLWAY.
(Ratio of Air, or Swell, to Theoretical Water Volume.)

FROM OBSERVED FROM Co>1PU1'ED


SURFACE VELOCl1'1ES .. DEPTHS
Station*
numbers
lllay 26, 1923 June 12, 1928 May 26, 1923 June 12, 1923

1 .... . ... ....


2 -0.04 0.56 0.45".
3 0.21 0.46 0 .02 0.57
4 0.84 0.44 U.09 0.54
5 0.27 0.88 0.05 0,48
6 0.19 U.46 .... 0.38
7 0.08 0 .50 . .. ....
....
,.
.
.......
8 0.09 0.07
9 0.04 0.19 ....
10 0,04 0.32 .... ....
11 -0.01 0.29 .. .. ....
• Numbers of obse rvation stat10ns, not d1sta_n ces. (See Fig. 29.) Values show very rough approxi-
mations. ·

Approach Velocity to Crest.-The observed surface velocities and the


approach velocities_to the weir crest are shown on Fig. 29. The approach
velocity measurements were taken on June 12, 1923, from Pier 5. The velocity.
of approach is affected by the contour of the reservoir floor near the crest and
by the position of a log boom up stream from the crest. It is not the same
for all stations along the crest.
Shape of Upper Nappe.-An interesting result of these observations _ is
shown on Fig. 30. One of the simplest problems in experimental hydraulics
wollld i.ppear to be the determination of the path_ of a jet over a sharp-crested

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

Fra. 26.-Am:ownocK SPILLWAY, JUNE 12, FIG. 27.-ARROWROCK SPILLWAY , JUNE 1 2,


1923 , LOOI<ING DowN INTO CHANNEL 1 92 3 , LOOKING U l' CHANNEL ~' ROM
FROM. PIE R 1. DOWN-S T REAM END.

FIG. 28.-ARROWROCK SPILLWAY ," JUNE 12, 1923, LOOKING U P STREAM I•'ROM PiER 4 .

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Sil)E CHANNEL SPILLW.A.YS 925
weir, yet practically nothing toward this end has been done. Measurements
out to a limited distance were made by Bazin, and various writers have
theorized about the subject, generally using these measurements as a basis,
with the widely varying results shown in the diagram. The trajectories
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

shown are not all based on identical assumptions, and in order to plat some
of them it was necessary to make certain interpretations which may not
agree with the original intentions of the various authors. The majority of

Observation Station N~mber


2 3 4 5 6 7 8 g 10 11
' n ..,....-n ~ ;,. •
II

/
I
I " "-- 0

20 .I
I
'O
C
/v
0
0
Q)
o/ SURFAC E VELOCITI ES IN CH ANN EL
en
/ May 26, 1923
Q)
a. 10 I
-.;
if 40 .

-
0 :i..-
V
.---0- 'O
0/
/ SURFAC E VELOCI TIES IN CHANNEL
I J une 12, 1923
10
0 , 2 3 4 5 6 7 8
Distance in H u nd reds of Feet

0
1
1'µpstriam 1
lr~mt .e1t
12'
i" ~
4' ~

APPROACH V ELOC ITIES


J u ne 12, 1923
I I !I
4 3 2
Ve locities In Fee t per Seco nd
FIG. 29.-EXPERIMEN TAL DATA, A _R ROWROCK SPILLWAY. -SURFACE AND A .PPROACH VELOCITIBS.

the curves are based on an _assumed sharp crest with a vertical up-stream
face. These curves are presented, not as a criticism of any of the theor ies
proposed, but to show the lack of definite information on the subject. Mr.
Steward was fortunate enough to obtain measurements of the upper nappe for
both measured flows at Arrowrock. The observed nappe for June 12 is
shown superimposed upon the t4eoretical curves for comparison. I n order
to avoid too great confusion the observed nappe for May 26 is platted
separately.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

<:O
~ .
O'>

I 3216,f-- t - - t - - t - -t -- t - t - - t - + - : - - t - - t ---t
8216'
1 w, S. El, Reservoir 3213.61 3214LLW, $. El. R),ser,olr 3213;ao
8214 -.....!--
~
- - -- - - --- -r-,..._
8212!
",, __,, 3212

"-'•"~""'M'-
--+-.
Q
3210 !---."-
I W ,---1-- 't»J I I I
3210I ,_ .

/ 'ii' " Ul

3208
:~ \ ~ ='A
3208 I I 8
!;;I
, #
]3206 ,, ,.:I I \',\~
... ]3206
L l:?f I I \ I \\ I I ··-J\ I I I a
~
-.,, ~ ""'
/ I I \ LL

---
LL .•
zz>
Theo etlcal ertica Sharp Crest \ ,\ \ .!:3204
.!:3204
V \ \ I /
C
·: -~" 'le!
C
v ,,
. Observed U1 per Na pe
.
0

I
'33202 / '' :;:;3202 !;;I
/ , pproxlmate C. . Obse ved flow, Jur e 12, · 923 ObServed ~pper ~appe"IV t:,t
> >

-
Appro~lmatelC,L.Ob:served[Flow(l \
~ 3200 / pprox mate L wer Nappe
Creager
[5 3200 June 12, 192~ I ~
Ul
>,:j
Al)proxln:(ate LaWer NaP.pe / >-<
3198
Parke 1 "" 3198 t<
Bellass\s . \\
"" I t<
3196 JUN E 2 1923 Etche~erry " 31 96
::a
Curv s attributed to variot s Authorities Merriman ·. '.~ ~.
are for center lines as cO tnputeO,from Turneaure and Russell- ...l I \ ~
3194 Ul
3194 I
publ'shed Formul~s, and are in~ost f rlzze l
31.92

8190
...
easel for a theoretical sh rp Cre$t, with
~lcal f~nt wa~ as shown dotted
\

\1 I
\\'1 I
',ii
\ II
?!
l!.
'! 3192

3190
MAY 26, ) 923

\ I\\ I Ill m L ~
20 18 16 14 12 10 8 6 4 2 0 2 4 6 8 10 12 14 16 8 6 4 2 0 2 4 6 8 10 12 14
Distance s in Feet Distan ce s in Feet
FIG. 30 .- EXPERIMENTAL D ATA, ARROWROCK SPn.LWAY.. SHAPB OF NAPPE, MAY 26 A.ND
J UNE 12, 1 923.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927


SIDE CHANNEL SPILLWAYS 927
Although the discharge on May 26 was greater than that on June 12, the
trajectory for the latter date is the flatter of the two. The first observation
was made at Pier 3 and the second at Pier 5, where· the velocity of approach
appeared to be greater. Mr. Steward thinks this accounts for the discrepancy
Downloaded from ascelibrary.org by Texas A&M University on 09/29/24. Copyright ASCE. For personal use only; all rights reserved.

between the two nappes. The observed points are indicated on Fig. 30 by
circles. The lower nappe· curves shown are estimated.
The experimental nappe for June 12 conforms closely to that proposed by
Creager, the ni8st conservative of all the theoretical maps. When it is
remembered that the economy of an overflow dam is often largely influenced
by the thickness required to prevent the jet from leaving its surface, ·the
importance of pursuing this problem further will be appreciated.

EXPERIMENTS BY J. W. ELLMS
A situation similar to side-channel spillway flow is found in the design
of wash-water troughs for rapid sand filters. A discussion of this problem,
by 0. N. Miller, Assoc. M. Am. Soc. 0 . E., has been published.*
Mr. Miller develops a theoretical formula based on the energy equation,
and containing an undetermined factor for what he calls friction. This
factor, which is expressed as a percentage of the velocity head, includes the
friction of the channel and the internal friction, or impact loss.
The mathematica l discussion appears to be correct. In order to simplify
the formula deduced, the assumption is made that the water surface curve
is parabolic, making the average effective fall throughout the trough length
equal to · two-thirds of the total fall.
It is evident from the general discussion of the problem given in this
paper that in many cases this assumption may be incorrect. However, it is
probably applicable, within the necessary limits of accuracy, to the particular
problems discussed by Mr. Miller.
Applying the theoretically derived equation to some experimental data
obtained by Mr. Ellms at the Cincinnati, Ohio, Filtration Plant, a value of
0.75 of the velocity head was found for the friction br impact coefficient.
A coefficient found in this way is applicable only to the conditions under
which the experiments were conducted, the loss by shock being dependent on
the distribution of velocities throughout the length of the trough, rather than
on the velocity at the outlet end.
Mr. Miller's formulas are designed for rectangular channels and are not
directly applicable to the results obtained at Bellvue. However, the net loss
can be found by subtracting the velocity head from the average fall, which
latter is obtained from the water-surface curve. The average ratio of impact
and friction loss to velocity head at x = 16 ft., for the 15 runs at Bellvue, is
about 0.375, or about one-half the value found at Cincinnati.
• "Water Purification,"' by J. W. Ellms, M. Am. Soc. C. E., Appendix B, 1917 Edition.

Transactions of the American Society of Civil Engineers, 1926, 89(1): 881-927

You might also like