【Side Channel Spillways Hydraulic Theory, Economic Factors, And Experimental Determination of Losses(Hinds,1926)】
【Side Channel Spillways Hydraulic Theory, Economic Factors, And Experimental Determination of Losses(Hinds,1926)】
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.
SYNOPSIS
INTRODUCTION
==ao . .
::::==3000
::::3-0.59
Q= 30 000 c.f.s.
L= 420 It. PROFILE
'n=o.75. ·a= 0.2 5
+ I El. 2690.00
~I
.;
FIG. 1.
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.
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.
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
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
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
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)
tangent of the water-surface slope. Substituting. this force for dd~ in Equa-
y = glf"'( V dx
dV + Q
b V 2) d x . .... . .. ... .. . .... (7)
O
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.
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.
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) _
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
N
FIG, 6.
A
hv = -n - - ........ . . . . . . . .· ....... . .. (17)
n + ·1 2 T
or, from Equation (11) :
A
y = 2 T' ... ............ ............ .. (18)
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.
(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
~
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.
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
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:
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
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
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
·OC
<:.::
oc
s§
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)
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)
~
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
'°
'°
~
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
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
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
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, .
'
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 ✓
·~ -
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
~
"
~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
"E
\ ~ ~~ er Rod.....,
~ \ \ / ..;::..,
": \
~
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.
'
!.
~
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~
.,
<{
---
~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~ !--
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
~
□
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
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_~
v·
'"""--- ---- -------~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
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
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.
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.
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.,"
FIG. 22.-RuN No. 14. DrscHARGE, 22,42 · sEcoNn - FEE'l'. MoDERA'mLY sunMEHGEu FLow.
FIG. 24.-RUN No. 12. DISCHARGE, 29.77 SECOND-FEET. SHOWING CROSS-FL(?W BE~9W
CREST FQI\ (:E;RTAp,r S,AGJ;;~, .
tO
~
0
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
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
FIG. 28.-ARROWROCK SPILLWAY ," JUNE 12, 1923, LOOKING U P STREAM I•'ROM PiER 4 .
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
/
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' ~
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.
<: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.
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.