580 DISTRIBUTION STORAGE-DISCUSSION
enhancing their beauty, particularly in
municipal installations. The factory
look is being taken away from elevated
tanks and standpipes. The use of
welding as a means of construction has
allowed the creation of designs that
were not considered possible in the
days of the riveted tank. Structures
are now of all-welded construction
columns are streamlined and tubular'
standpipes have ornamental dome 0;
ellipsoidal roofs, and some are orna-
mented with architectural pilasters,
cupolas, and so forth. This does not
add greatly to the cost, but goes a long
way in establishing better public rela-
tions. Since A WW A has issued
standard specifications (1) for elevated
steel tanks and reservoirs, their design
has become stabilized.
Some rule of thumb figures for the
cost of large steel reservoirs and ele-
vated steel water tanks may be of in-
terest here. On the basis of flat-bottom
all-welded reservoirs, about 40-ft high:
completely erected and painted, but
not including foundations, piping, or
valves, the cost of a 500,OOO-gal storage
vessel would be in the neighborhood of
4-5 cents per gallon; and of a 5-mil gal
ground reservoir, about 2i-3 cents per
gallon. Elevated steel water storage
tanks, 100-ft high to the bottom, cost
about as outlined by Aldrich: a 100,000-
gal elevated tank would run from 25
to 28 cents per gallon; a Lrnil gal ele-
vated tank, from 15 to 18 cents; and a
2,}-mil gal elevated tank, from 10 to 11
cents. On such tanks the entire capac-
ity is usable, of course, with a mini-
mum head of 100 ft.
Reference
1. Standard Specifications for Steel Tanks,
Standpipes, Reservoirs, and Elevated
Tanks, for Water Storage-AWWA
DIOO. Am. Wtr. Wks. Assn., New
York (1952).
George M. Haskew·--- _ Design of Wye Branches for Steel Pipe
Engr. & Sicpt., Plainfield-Union Water
Co., Westfield, N.J.
~t is quite a task to discuss a paper
which reflects engineering ability and
experience. Distribution storage is of
great importance in the operation of all
water companies. The writer thor-
oughly agrees with Aldrich about the
kind and proper location of containers.
It is certainly becoming very difficult
to secure sites, because of the rapid
growth and development of the nation
especially metropolitan areas, alon~
WIth the public's resistance to havine"
construction near its homes. b
The writer believes that steel con-
tainers are the best to use for either
underground or overhead storage, tak-
Ing mto account original cost and main-
tenance. The writer's firm has an
underground storage reservoir of re-
inforced concrete, with roof of the
same material, but all later additions-
standpipes, overhead tanks, and other
containers-have been constructed of
steel, following the trend for better • of this work were published (2) in
service and economic operation.
. A very interesting part of the paper
IS Table 1 on the ratio of period rates
to the average annual rates of produc-
tion (see p. 574). The use of water
is constantly increasing due to lawn
watering, air conditioning, and other
water-using devices. The figures in I they established a system of basic de-
Table 1 will not apply exactly to any
system, but if they are compared with
th.e data .of a particular utility, they
will fit fairly well and can be a guide
for future operations and additions.
H. s. Swanson, H. 1. Chapton, W. 1. Wilkinson,
--------C. L. King, and
E. D. Nelson--------
A contribution to the Journal by H. S. Swanson, Structural Eng.
Assoc.; H. J. Chapton, Project Design Engr.; W. J. Wilkinson; C. L.
King; and E. D. Nelson; all from Los Angeles Dept. oj Water &
Power, Los Angeles, Calif.
broader program was then instituted,
D URING the past 30 years a num-
ber of countries have been doing and large-diameter pipes were brought
independent work on wye branch de- into the investigation. The results of
sign. Work in Europe has been car- this program were published (3, 4) in
194(r47.
ried on primarily by the Swiss and
the English. In 1928 the Swiss firm :prior to World War I the Southern
of Sulzer Brothers began using a California Edison Co. had a large-
circular-collar stiffener on wye branches diameter wye branch fabricated at the
to be installed in the Puente Nuevo Krupp Works in Germany. A forge-
power station in South America (.1). welded penstock fitting with a cast-
A patent for this type of reinforcement steel crotch section, it has been in serv-
was granted Sulzer Brothers in 1931 ice in Big Creek Plant No. 1 since
by the Swiss government. Several 1913. In the following 10 years, sev-
years later the countries of central Eu- eral other wye branches were built for
rope 'began using two- and three-plate this company by M. W. Kellogg Co.
, reinforcing designs for wye branches, The static pressure required of these
with the Swiss taking the lead in ex- sections varied from 1,700 to 2,400 ft
perimental work. The results of some of head.
Early in 1936 the Metropolitan
1941. Water Dist. of Southern California
The English began experimenting began work on a series of large-
with wye branch reinforcement in 1934; diameter wye branches. The first was
without the knowledge that Sulzer a 10-ft diameter pipe branching into
Brothers were already acquainted with three 6-ft diameter pipes. The re-
the problem. This early work of the inforcement was unusual in that it was
British was done on small-size pipe; mostly inside the section and so placed
that it did not encroach on the area
sign for one- and two-plate reinforce- of flow to any great degree. It was
ment for wye branches. J. S. Blair installed in the Hayfield Plant in 1936,
was mainly responsible for carrying on and since that time a number of similar
this work. In 1938 the investigators. sections have been fabricated and put
were granted a British patent. Dur-·· into service.
ing World War II, work on the ex- In 1930 the Los Angeles Dept. of
periments halted, and it was not until Water and Power began investigation
1946 that it began. again. A much and construction of stiffener-reinforced
581
582 H. S. SWANSON ET AL. lour.AWWA hme 1955 WYE BRANCH DESIGN 583

wye branches. In the following year a
system of basic design and stress analy-
sis was devised, and in May 1935 the
first major hydraulic test was made.
The wye branch, a 36 X 36-in., 45-deg
section, was checked for stress distri-
bution by strain gages placed on the
reinforcing plates, and the results were
analyzed in a report. This unit was
later installed in the city's distribution
system. Since 1935 this method of
using plate stiffeners to reinforce the
as a consequence, there was no oppor-
tunity to correlate the data. When
the need for building a number of wye
branches at the same time .arose, it
was decided to consolidate the design
work and testing analysis into a report
which covered a range of sizes and
pressures.
General Purpose
The general purpose of this paper is
to present a practical and economical
The general assumptions in this re-
port are based on the use of a tested
model that is similar to the actual wye
branch to be designed (5). The actual
pipe and the model must be fabricated
from the same material so that the mod-
ulus of f:lasticity, modulus of rigidity,
and Poisson's ratio may be given con-
static load into the stiffener beams.
Temperature is ignored in the design
because the effect is negligible when
the section is underground. If the see-
tion is to be laid aboveground, how-
ever, temperature must be included in
the design or allowed for through the
use of expansion joints.

This juncture of three 6-ft diameter pipes

entering a 10~ft diameter line was fabri-
cated in 1936 for the Hayfield Pumping
This band-reinforced wye was also fabri- Plant on an aqueduct of the M etropolitom
1913 Installation cated for the Southern California Edison Water Dist. of Southern California. The
Co. by Krupp. It was placed in opera- vertical wye stiffener plates continue
The above sections were fabricated for the Southern California Edison Co. by the tion in Big Creek Plant No. 2A in 1928. . throwqk the pipe as tension diaphragms
Krupp W orks of Germany prior to World War I. They were installed in Big Creek The large penstock is 3 in. thick and the and are tied together on the uphill ewd by
Plant No.1 in 1913 and have been in continuous service ever since. The breech joint two 48-in. branches are 21-in. thick. the circumferential stiffener ring .."
is cast steel, and the penstock pipe is forge welded.
sideration in the prediction equations. Design and Test Methods
crotch seams has been used in the dis- method to be used in designing large-'
The free ends of the unequal-length The design method developed for
tribution system whenever possible. A diameter wye branches. In attempting
I stiffeners are assumed to be affected by the two-plate wye section involves esti-
postwar construction program has re- this, the authors have sought to estab-
a balancing force which causes equal mating the amount of plate needed for
sulted in the installation of more than lish standards and good practice in de-
100 stiffener-reinforced wye branches termining the proper type of reinforce- deflections. This assumption is made the reinforcement and applying the re-
since 1947. ment and other basic facts connected so that the principles of statics can be sultant circumferential stress from the
In the past each wye branch at Los with the design. Tables, charts, and applied to the problem. The main con- pipe shells on it. In a tee section, the
Angeles was designed individually with nomographs have been included to cern is finding a means of transmitting curved-beam equations could then be
respect to sizes and conditions, and, simplify the work as much as possible. circumferential stresses due to hydro- used to calculate the critical stresses.
584 H. S. SWANSON ET AL. Jour.AWWA WYE BRANCH DESIGN 585

The 24 X 24-in., 45-deg wye
left shows the two stiffener
place. The inside curve of
match "the inner surface of the
When the section is other than 90 deg,
however, one plate will have a larger
load than the other, so that the calcula-
tions must include balancing the deflec-
tions of both plates at the junction in
order to arrive at the amount of load
which the Ipnger plate will transfer to
branch at
plates in the shorter plate. In dealing with the
the plates larger-diameter pipe, many designs
pipe shell. which would hold the deflections down
(6). Upon examination of the prob-
lem, it was concluded that the plate
could vary anywhere from a ring
girder to a two-plate design, depend-
ing upon the ratio of pipe diameters.
Equations were set up under this as-
sumption and some designs tested. As
a result of this work, a graph was made
which is used in conjunction with the
nomograph.

At right, the branch pipe is being lowered

into position.

1935 Installation

At left, the assembled unit is being fitted
for welding. If a third plate were to be
added, it could be welded on when this
portion is completed.
Fabrication Methods
In 1935 this section was tested hydrostaticallyfor stress distribution by the Los An-
geles'Dept. of Water and Power. Following the test, the stresses were analyzed in a
report and the section was put into service in the distribution system.
were tried. The most economical of As an aid in devising a practical
these proved to be the three-plate de- method of design, several examples are
sign in which a half-ring plate was given, with accompanying tables and
added to a two-plate design to hold sketches. The nomograph section of
down the free ends. the report contains simplifications
The design of the single-plate wye which give results within the range of
began with the supposition that the accuracy necessary, but, to use the
plate could be treated as a ring girder nomograph (Fig. 15) properly, a few
 'C7,,-
586 H. S. SWANSON ET AL. Jour. A W W A rJ lune 1955

TABLE 1

T
Explanation of Symools

Symbol Explanation Units

.:
- ."
t;. deflection angle of pipeline deg

T
\ D diameter of pipe in.
\ A in.
\A" RB radius of large pipe
R. radius of small pipe in.
RJ centroidal radius of reinforc- in. Q,;
ing plate
a inside radius of reinforcing in.
plate ~
d depth of reinforcing plate in.
d, d at 1-in. thick plate (as in in.

Approximate d,.
nomograph)
d at acute crotch in.
,
e ••• 12R
h
2
[
1 + 4 ( h
15 2R
)2j r = ------~--------
d
A
c
Centroid db d at obtuse crotch in.
'I
b210ge -;; +b l loge d
e, d: at junction of reinforcing in. (a)
Reinforcing plates e=R-r
Plates L, length of axis of large rei n- 1Il.
forcing plate (b)
L2 length of axis of small rein- in. b, b,
forcing plate
Fig. 1. Intersection of W, total water load on large re- Ib
Diameter Plates inforcing plate
W2 total water load on small re- Ib
tJ.s is unit length along centroidal axis r inforcing plate

ill
x is horizontal distance to center of tJ.s. Y1 theoretical reaction on large Ib
reinforcing plate
facts must be noted. First, the depth Y2 theoretical reaction on small Ib
of plate read from the nomograph is
for 1-in. thick plate; when this is used
with pipe of a very large diameter, the
Y3

T=P
reinforcing plate
theoretical reaction on ring
plate
resultant load on horizontal
Ib

Ib
)I ' .~~ 1 lL
. ,,111 ~ , lJ~'
curved-beam J( factor (7) becomes axis of plates (equal to
W±y)
very great, increasing the critical
P internal pressure in pipe psi
stress. If the value from the nomo- t thickness of reinforcing plate in. A A
r = ------------------------ r = ------~------
graph is converted to a thicker plate A cross-sectional area sq in. d g C d
with a smaller depth, however, the K M moment in section due to in.-lb 01 log. -
a
+
b : log. -d b3 loge -
g
+ bl log, -
a
+ C
b2 log, -
g
factor is lessened, giving a better de- load
m. moment at any point due in.-lb e=R-r e=R-r
sign. It is recommended that the depth to load of unity where de- (e) (d)
of plate, dw, should not exceed 30t (t flection is to be found Fig. 2. Typical Cross Sections of Reinforcing Plates
equals thickness of crotch plate). E modulus of elasticity for steel psi
Second, under some very high head It Poisson's ratio m order to approximate an allowable the use of a factor based on the data.
conditions in smaller-diameter pipe, the stress of 20,000 psi. Another factor The various designs were tested by
depth of plate has a tendency to ap- usually include a safety factor which to be noted is that, although the de- placing the section under hydrostatic
proach the diameter of the pipe. This will keep stresses below the yield point signs used in the report are based on load and using electrical strain gages
also will increase the K factor. It is of structural steel. Because the maxi- a rectangular cross section for the to indicate the amount of stress to
recommended, in such a case, that the mum allowable stress of the steel used crotch plates, the tests showed that the which the section was subject (8).
depth of plate be made less than the in this report is 30,000 psi, however, welding in the crotch, plus 'a portion These data were compared with .the
diameter of the pipe. it should be understood that the design of the pipe shell, acted as a flange design stresses, and the method of
If these recommendations are fol- pressure used in the nomograph was which decreased the inside fiber stress. design was corrected to conform to the
lowed, the nomographic design will kept to 1.5 times the working pressure This was taken into account through results of the tests.
588 H. S. SWANSON ET AL. WYE BRANCH DESIGN 589
4
II
3
\

2
\
~

1.0
<;
r--
---- ~ Inside Fiber

O. 9

O.8

O. 7
,/
....1-- - outside Fiber

o. 6 /
/

o.5
2 3 4 6 . 7 8
s.
c
Fig. 3. K Factor for Curved Beams (Rectangular Cross Section)

R· is ra d'tUS of centroid;
.. c ts one-halj
l '.
depth of plate; f max = K Me
-r-'

The use of single-direction SR-4 SR-4 electrical strain gages to be at-

strain gages in the testing of the wye tached to the inside periphery of the
sections was desirable because this type crotch stiffeners without shorting out
of gage required less mounting room under the high hydrostatic pressures
and fewer connecting wires than the used. Single SR-4 gages applied in
rosette type and gave stress results this manner gave test results in sec-
while the section was still under test. tions that had not previously been
The. Baldwin-Lima-Hamilton Corp., available. (See photos on pages 590
Philadelphia, developed a method of and 594.)
using these gages on sections with bi-
On one large wye branch, the Mag-
directional stress, but the direction of
naflux Corp., Chicago, was called in
the principal stress must be known.
to use the "Stresscoat" method for de-
The rosette gages are essential for ob-
taining stresses in locations where the termining the stresses, in conjunction
direction of the principal stress is not with the electrical strain gages which
known. The equations used in com- were normally used. It was consid-
puting stresses from rosette strain ered advisable to try this method be-
readings are so involved that the calcu- cause of the complete coverage that it
lations had to be made in the office gives. The results of the two methods
after the tests had been run (8, 9). compared very favorably and gave a
During the tests a waterproof ce- good picture of the stresses in the sec-
ment was found which allowed the tion. The "stresscoat" test confirmed
590 H. S. SWANSON ET AL. JOllr.AWWA JUlIe1955 WYE BRANCH DESIGN 591

the locations of the critical stresses
that had been determined in the design
analysis.
Conclusions and Recommendations
Throughout these tests, it was found
that the design of the actual section
could be predicted from the model to
a degree within the range of accuracy
6
was apparent that there would be diffi-
culty in proportioning the plates so as
to have each one carry its share of the
load. By the use of the method of
balancing deflections, a balanced design
for a three-plate section was found;
because the reduction in the size of the
plate depths, compared with that of a
two-plate design, was very small, how-

I
5
\ cen~;Q
Neutral Axis
III
~~
____
!

4
\ ~~
~"I
'e ......:~

\ '\
~c--c-' I.
I

'" <,
r-,
~ r-.....

I
~ 1.0..

o
001 0.02 0.03 0.04 0050.06
008
/=-
~I[ /,2
1+-
J 0.2 0.3 0.4 0.5 0.6 0.8 1.0

12R 15R2

Fig. 4. Determination of Eccentricity, e

Based on Tirnoshenho (10) equation:

~: e = 1~~ [1 +1
4
5( 2~ Yl e = fh;

in which e is eccentricity (distance from centroidal axis to neutral axis in curved

beam); h is depth of curved beam,' and R is radius of controidal axis of curved beam.

required. Because of the difficulties ever, it was concluded that, if a third

encountered in analyzing the type of
curved beam used in these designs,
however, it became necessary to apply
stipulations to the design procedure in
order to conform to the test results.
It was found relatively easy to ob-
tain good balance in the two-plate de-
sign, but, in the three-plate design, it
plate was to be used to hold down de-
flections on larger-diameter pipe, it
should be added to the two-plate de-
sign and the dimensions should con-
form to the top depth. Also, it should
be welded to the other reinforcement
plates only at top and bottom, being
left free from the pipe shell, so that
592 H. S. SWANSON ET AL. J01!r.AWWA lune 1955 WYE BRANCH DESIGK 593

TABLE 2 none of the shell stress would be trans- actual pipe is maintained, the safety
General Formulas (See Fig. 1) ferred to the ring plate. It appears of the structure is assured.
practical to begin adding the ring plate The principal symbols used in the
R R W = p sin ilL22
L1=--; L2=~; 2 to a two-plate design on high-head, design equations and sketches are
. ~ 2
equal-diameter wye branches with pipe listed in Table 1; other symbols will
sm C05
2 2 diameters greater than 60 in, be explained when used. The general
W at any point x from origin (triangular load) : During early testing of the models, formulas employed are shown in Table
psin ~X2 it was found that, in the critical sec-; 2. Following the section on analysis
W" = --2- tions, the measured stresses were lower and design procedures is a description
than the design stresses without ex- of the nomograph method, including a
Moment at any point :t from origin (triangular load) : discussion of the development of the
ception, giving an additional safety fac-
p sin ilx3 .
tor. It was concluded that the flange curves (pages 617-623). Test results,
111 = 6 ;M = m ± xy (use minus for larger plate, L1; use plus for smaller plate L,)
effect of the weld metal and a portion in the form of drawings and tables, are
shown on pages 624-627.
of the pipe shell in the crotch had ac-
a = Rpipe sin % (for acute angle of two- and three-plate design)
cepted some stresses which it had been
. Appendix A (page 628) contains a
comparison between a Sulzer Brothers
/l . assumed would be in the reinforcement
a = Rpipc cas 2 (for obtuse angle of two- and three-plate design) design and one based on the authors'
plates. By correcting for this in fur-
nomograph method. Appendix B
Two-Plate Design ther testing, the design methods were (pages 628-629) illustrates the deflec-
improved to give more economical tions that occurred on two 4S~deg wye
P = total shear at any point x from origin
proportions. branch pipes. The deflections are
p sin 6,..,;2 P P sin /lx'
PI = 2 - y; ,= --2- +y A review of the test data evaluated plotted on a nomograph, to provide an
.for a practical and economical design indication of what might be expected
y = load necessary to cause equal deflection of plates shows that the dimensions for the re- ~n other 4S-deg sections. A discussion
. m,x~s yx2~s inforcement plates for a wye branch of the results of overstress in wye
Deflection at x = 0 for Plate 1 = 1:"""liI - 1: EI may be chosen from the nomograph. branch pipes is found in Appendix C
If similarity between the model and the (pages 628 and 630).
Deflection at x = 0 for Plate 2 = 1: m2x/ls + 1: yx'/ls
EI EI
--------Analysis and
Design
Procedures--------
As deflections must be equal:
.., m,xils _ l: yx ils 2
= l: tn,x/ls 1:yx t.,s
2 In the designs described in this por- used in Seely's method can be found
.<J EI EI EI + EI tion of the paper, a trial depth and from the curves in Fig. 3 and 4.
thickness of stiffener plate were selected A series of typical designs are pre-
Three-Plate Design and the stresses in the critical sections sented: single plate, D.. =90 deg; two
were computed. Then a new depth and plate, D.. = 90 deg; two plate, D.. = S3
PI 1 k 1: YIX2~S
ate ,Ii
= mlxM
EI
_
EI
.
' thickness were chosen, as indicated by deg; two plate, D.. = 45 deg; three
the resulting stresses. These stresses plate, D.. = 90 deg; three plate, D.. = 4S
were evaluated by using the loading deg; and three plate (alternate de-
assumptions shown in Fig. 1, along sign), D.. = 45 deg.
Maximum Stresses with a carryover load, y, found by a In an early effort to use the model
method of balanced deflections. The theory as an aid to future design, a
Mmax= W [ % + (0. 2) ~ ] ±y [L + (0.2) ~ ] solution equations were taken from number of tables were computed, such
Timoshenko (10) and Seely (7). as the. one in Fig. 6 (page 597). This
The various cross sections of curved table was based on the two-plate 24-in.
beams which co.uld apply to wye branch wye branch shown in Fig. 7 and was
design are shown in Fig. 2, although retained as an example only. The
f nux
= K Mm."
I
C + W A'± Y. f.
mID
= Mm."
I
C _ W ± Y
A
t
only the rectangular section will be nomograph (Fig. 15) has greater ver-
*Timoshenko (10); c equals one-half depth of plate; for hI and h" see Fig. 4 (page 591).
considered in this report. The eccen- satility and lends itself better to corre-
t Seely (7). tricity, e, and the K factor which is, lation with test data.
594 H. S. SWANSON ET AL. ]our.AWWA lune1955 WYE BRANCH DESIGN 595

One-Plate Design

For one-plate design:

0.3 s ~~~0.6
W, = st; = PRB

W, = WI =PRB
2 2
At crown:

Moment due to W, = +
W R'*

Moment due to WI = -4
- W,R
. 2 '
'*
Slll £!,.

M, WI )R2
= ( W,- -.--
Sl1l'£!,.
~
4
Two·Plate Wye Under Test
This photograph of a 3D X 3D-in., 90-deg wye under test shows the mercury switches
r3 = L,' +'!:!
R, 2
and deflection gages.
At sides:
W,R.'
Moment due to W, = 4-----:--2
ITIllIIDIDI
Sill £!,.

-W,R'
Moment due to W, = --4--'

~ - W2)R.'
M _ ( Sl1l26.
, - 4

r
R.'
=-+-
L,
d",
2 ITJJJIIIIIJJ
R,'
r =-
L, +-db2
Fig. 5. General Equations for One-Plate Wye Design

Example Maximum moment at side, M2= (WI- W,) ~'

Main pipe radius, RB = 30.5 in. = 114,400 in.-lb

Branch pipe radius, R, = 10 in. As a trial, use a 6 X I-in. collar to stiffen the
Deflection angle, £!,. = 90 deg
joint. Then:
Design pressure, p = 300 psi
Me T
Uniform load at top, W, = pRB = 300(30.5) fmax =KtT+:A
= 9,150 psi
6M WR,
Ioa d at SIid e, W 2 = -2-
pRB 300(30.5)
U norm
if = 2 = Kt bd2 + bd sin £!,.

= 4,575 psi At the side, on the inner curve:

Maximum moment at top, M, = (W2- W,) ~' ~ = 4.2

e
= -114,400 in.-Ib
----
• Paris (11).
t K is Seely's constant,
of the collar.
dependent on the curvature

Three-Plate Wye Branch

596 H. S. SWANSON ET AL. lillie 1955 WYE BRANCH DESIGN 59i
K = L.185 = (-0.87) -114,400(6) +4,575(10)
(f)
()
. = 1 185 114,400(6) + 9,150(tO) I max 1 (5.5)2 1(5.5)
f, max • 1 (6)2 1 (6)
= 19,700 + 8,300 = 28,000 psi
= 22,600 + 15,200 = 37,800 psi (tension)
(tension)
At the top, on the inner curve: In the first trial, the outer stress at the side
was ignored because it could not exceed the
-B.
= 4.3 inner stress. At the top, however, the direct
c stress is additive to the outer stress due to
K = 1.18 moment, increasing the total stress to the
point where it must be investigated along bd2 For top section, K equals:
(h ) mnx 118 -114,400(6) + 4,575(10) with the inner stress.
S = 6" (section modulus)
=. 1 (6)2 1 (6) 0.0521 at 6.25, 1 in.
The final design is a I-in. thick collar,
'S = N3Sm* d = KDJ
0.0425 at 6.25, 1tin.
-22,500 + 7,600 = -14,900 psi welded at the junction of the pipes, with a 0.0368 at 6.25, 2 in.
depth of 7 in. in the crotches and a depth of
(compression)
At the top, on the ou ter curve:
5t in. at the top and bottom. II' .t:
Dm
Data: 0.0301 at 6.25, 3 in.

NOTE: Bier (6) states that the assumption S« = 36 in., 6.25 For wye section, K equals:
of a uniform load is on the side of safety, but
-B. = 4.3
it should be noted that, at smaller deflection Dm = 24 in. 0.125 at 36, I in.
c
angles, the ring becomes egg shaped and the 0.102 at 36, It in.
b = I, I!. 2, 3 in.
K = -0.86 ring girder equation for moment becomes less * Subscript II m" refers to model.
0.0884 at 36, 2 in.
-086 -114,400(6) +4,575(10) applicable. Also, when the branch pipe diam- p = 350 psi 0.0722 at 36, 3 in.
(Jl ) max =. 1(6)' 1(6) eter approaches the diameter of the main pipe,
the assumption that the collar lies in one plane
16,400 + 7,600 = 24,000 psi causes inaccurate results. The conclusion is Plate Depth, s-« ••.
(tension) that this method of one-plate design can apply Diam.

1-;-
For Trial 2, use 5t in. as the top depth and only to limited cases of wye branches where D Di Wye Side Top
7 in. as the crotch depth. By inspection, the deflection angle is close to 90 deg and the
tension will govern at both the top and diameter of the branch pipe is substantially b = 1 in. b = I! in. b = 2 in. b = 3 in. b = 1 in. b = It in. b = 2 in. b = 3 in.
the crotch. At the side, on the inner smaller than that of the main pipe.
curve: The ASME Boiler Code (12) covers a good 41.5 5.19 4.23 3.67 3.00 2.16 1.76 1.53 1.25
portion of the one-plate design in the lower 24 117.6 14.7 12.0 10.4 8.49 6.12 5.00 4.33 3.54
-B.
= 3.8 ratios because the reinforcement pad is close 30 164.4 20.5 16.8 14.5 11.9 8.56 6.99 6.05 4.94
c to the plane of stress. It appears that this 36 216.0 27.0 22.0 19.1 15.6 11.25 9.19 7.96 6.50
method is good up to a ratio of diameters of 42 272.4 34.1 27.8 24.1 19.7 14.2 11.6 10.0 8. I\.'
K = 1.21
pipe of 0.3: 1. By the same token, the ring 48 332.6 33.9 29.4 24.0 14.1 12.2 10.0
= 12 114,400(6) + 9,150(10) girder assumption appears to be satisfactory 54 396.9 28.7 16.9 14.6 11.9
(J)
2mox .1 1(7)2 1(7) 40.5 35.1
up to a ratio of 0.7:1, but above this ratio, 60 465.0 47.4 41.1 33.6 19.8 17.1 14.0
= 17,000 + 13,000 = 30,000 psi the curvature of the plate is such that it has 66 536.0 47.4 38.7 19.7 16.1
(tension) been found advisable to use an equal pipe 72 611.0 54.0 44.1 22.5 18.4
diameter design and a truncated cone to reduce 84 770.0 68.1 55.6 28.4 23.2
At the top, on the outer curve: the diameter of the outlet pipe to the desired 96 941.0 67.9 28.3
size. In the one-plate design curve in the 108 1,123.0 81.1 33.8
!!:. = 4.63 nomograph section, an attempt was made to 120 1,313.0 94.8 39.5
c
join the methods of design together so that
K = -0.87 they merge into the two-plate design.
Fig. 6. Two-Plate Design (Example I)-Plate Depth Calculations
Two-Plate Design

Example 1 Cmnputationfor Selection oj Plate Thickness

NOTE: As the two plates are identical, it is
RB = 12 in.
R.=12in.
not necessary to set up a table for the compu-
tation of force y, which automatically reduces
WI = W2 = (!!.)
2
(L)2 sin A
2
= 350 (1697)2
2'
= 50400 Ib
,
A = 90 deg to zero. The selected depths for crotch and
p = 350 psi crown were obtained from the prediction Maximummoment:
Pipe shell thickness, i in.
Depth at crotch, 12 in.
tables for pipe tees which ranged in diameter
from 12 in. to 120 in., at a pressure of 350 psi. M = WI
( '3L + (0.2) 2'd) = 50,400
[16.97
-3- + (0.2) '2
12]
= .
345,500 in.vlb
Depth at crown, 5 in.

j
598 H. S. SWANSON ET AL. Joltr.AWWA ilmel955 WYE BRANCH DESIGN 599

By Seely equation:
R 14.48
-;; = -6- = 2.413; K= 0.77

KMc W 1.40(345,500) 50,400 ..

lmax = -r- + A = 0.>167(1.5)(12)2 + 12(1.5) = 16,200 PSI (tension)
KMc W (0.77)( -345,500)
Imi. = -r- + A = 0.167(1.5)(12)2 + 2,800 = -4,600 psi (compression)

Example 2
R8 = 18 in.; R. = 18 in.; t:. = 45 deg; p = 385 psi
Pipe shell thickness, fin.; depth at crown, 12 in.
Table 3 shows the corrected values as constructed : it also provides a simple expedient for
#12 converting the design value of y to a new y for any other pressure desired.
0
I
5
I
Scale in Inches
10
I
16f%1I-------'l7-+i-..;----
11-01.;-.--- Calculated Stresses for Larger Plate

Depth at crotch, 30 in.; plate thickness, 2! in.

NOTE: Owing to the flange effect of the pipe shell, a factor of 0.2 is applied to ~ . The cen-
Stress"-lb Stress*--lb troidal axis is taken as being parallel to the curve where the stresses are being computed.
Item Item
Computed Measured Computed Measured Maximum moment:

if12 9,760 9,260 :if17 -8,180 -5,710 M = W, [L'3 + (0 2) ~]

. 2
- y [~
3·
+ (0 2) ~]
2
if13 7,980 9,250 :if18 -5,525 -3,090
lf14 3,760 6,490 :if19 -3,550
lfl5 1,760 B·2 -4,020 -5,650 = 277,000 [16.13 + (0.2) 320] - 45,800 [48.38 + (0.2) 320] = 2,946,000 in.-Ib
lf16 -4,880 -4,000
a = R sin ~ = 18 sin 22tO = 18(0.3827) = 6.9 in.
'" At 350·psi pressure.
R 21.9 ..
Fig. 7. Two-Plate Design (Example 1)-Plan of Stiffener Plates h = 30 = 0.73; e = 0.128(30) = 3.84 In.; hi = 11.15 m,

For any pressure other than the design oalace (350 psi in this instance), after the Maximum fiber stress (tension):
critical. dimensions have been selected from the nomograph (Fig. 15), the outside
curue of this plate should conform geometrically with the shape shown. Timoshenko :
Mltl T 2,946,000(11.15) 231,000 .
1max = Aea +A = 30(2.125)(3.84)(6.9) + 30(2.12S) = 19,9S0 + 3,600 = 23,600 pSI

a = R sin ~ = 12 sin ~O = 12(0.707) = 8.48 in. Seely:

R 21.9
- = ~5 = 1.46; K = 2.01
c 1
!!:. = 14.48 = 1.207; e = 0.071 (12) = 0.852 in.; 1£, = 5.148 in.; It2 = 6.852 in.
h 12 2.01 (2,946,000) (6) .
Im•.x = 2.12S(900) + 3,600 = 18,600 + 3,600 = 22,200 pSI
By Tirnoshenko equation:
Calculated Stresses for Smaller Plate
MIt, ion) f . I
!mox = --
Aea
+ A-W =
345,500(5.148)
( )(
12 0.852 8.48
)
50,400
+ -1-2- =
. (
24,700 pSI tension or 1-m. pate Depthat crotch, 20 in.; plate thickness, 2t in.
Maximum moment:
Use 1!-in. plate:

!ms.oc = }(24,700) = 16,500 psi M = W2 [

L2
"3 + d. ]
(0.2)"2 +y "3 + (0.2) "2d2 ]
[ L,

- Mlt2 W 345,500(6.852) 50,400

!mi •• = Aec +A = - 12(1.5)(0.852)(20.48) + (12)(1.5) - 4,750 psi (compression) = 52,500 [7.02 + (0.2) 220] + 45,800 [21.06 + (0.2) 2 0]
2
= 1,530,000 in.-Ib
600 H. S. SWANSON ET AL. WYE BRANCH DESIGN 60L

TABLE 3
#61-- #51+-13j{!'
Two-Plate Design Calculations (Example 2)

R 18 R 18
L, = ~ = 0.3827 = 47.04; L2 = --.1 = 0.9239 = 19.48 Plate Thickness -::::2~"
sl11
2
cos
Z ------- - .
LI (scaled from Plate 1)
d1 = 30 in.; a, = 20 in.
= 48.38; L, (scaled from Plate 2) = 21.06
/ ;:'ICO

!
C\I

p sin .1(L,)2 385(0.707)(48.38)2 8500 Ib

WI = 2 = 2 = 31 ,
2St.:l # 10
A-2

W2
_ p sin
-
.1(L2)2 _ 385 (0.707) (21.06)2 _
2 - 2 - 6 ,
360 Ib ° 1
i+-lll

-;..-----78~"--------
3D" _I

o 5 10
I I I
px3 sin 6. 385 (0.707)x' Scale in Inches
1n, = 6 = 6 = 45.36x3

x
in.
1.\4. = ml -
in.-lh
xy Z
in..4
x (As)
in.
x (Ils)
~I-
x(lls)
I
(ml - xy)
Depth of
Plate
in.
~
=.»
s = 9.5± in.; t = 2! in.

4.70 5,000- 4.7y 313 44.66 0.143 1,000- 0.67y 12.1 Stress*-lb Stress*-lb
14.1 127,000 -14.1y 371 134.0 0.361 46,000- 5.09y , 12.8 Item Item
23.5 588,000 -23.5y 486 223.0 0.460 270,000 -10.80y 14.0 Computed Measured Computed Measured
32.8 1,601,000 -32.8y 767 311.5 0.406 650,000 -13.32y 16.3
42.0 3,360,000 -42.0y 1,416 399.0 0.282 947,000-l1.85y 20.0 A-I 9,220 10,620 H-2 -8,320 -6,710
51.2 6,036,000 -51.2y 3,111 487.0 0.157 948,000 - 8.04y 26.0 A-2 -6,520 -5,420 J-1 13,300 11,070
51.38 6,093,000 -51.4y 4,875 488.0 0.100 609,000 - 5.14y 30.2 B-1 7,540 11,140 J-2 -7,470 -7,300
51.38 6,093,000 -51.4y 4,875 488.0 0.100 609,000- 5.14y 30.2 B-2 -5,270 -4,260 #1 9,560 12,250
C-1 -4,600 980 #2 2,460 3,930
4,080,000 -60.05y C-2 -7,890 -6,450 #3 -7,690 -270
D-1 -2,800 -930 j:f4 -6,020 780
D-2 7,560 5,900 #5 -6,120 -6,320
s = 7.5 in.; t = 2.5 in.
E-1 -2,780 -2,860 #6 -5,120 -3,760
E-2 3,210 820 #7 370 210
3.7 2,000+ 3.7y 378 26.25 0.0694 0+ 0.26y 12.2 F-1 2,010 2,850 #8 10,270 8,600
10.9 59,000+ 1O.9y 559 81.75 0.1465 9,000+ 1.60y 13.9 F-2 -2,290 -1,480 #9 -6,320 -5,750
18.0 264,000+ 18.0y 886 135.0 0.153 40,000+ 2.75y 16.2 G-1 6,980 10,970 #10 -8,290 -6,780
23.06 544,000+23.1y 1,235 173.0 0.1401 76,000+ 3.23y 18.1 G-2 -6,610 -5,690 #11 -9,140 -7,050
23.06 544,000+23.1y 1,519 173.0 0.1139 62,000+2.63y 19.4 H-1 8,680 11,920
23.06 544,000+23.1y 1,664 173.0 0.1040 57,000+ 2.40y 20.0
* At 33S-psi pressure.
244,000+ 12.87y
Fig. 8. Two-Plate Design (Example 2)-Plan of Stiffener Plates
4,080,000 - 244,000 = 60.05y + 12.87y; y = 52,600 lb'
See italic note below Fig. 7.
* To obtain new value for y for any other pressure, multiply y by new pressure and divide by 385.
602 H. S. SWANSON ET AL. JUlie 1955 WYE BRANCH DESIGN 603
I!. 45°
a = R CDS "2 = 18 cos 2= 18(0.9239) = 16.63 in. Table of Stress Values (see Fig. 9)

,
!!:.
II
= 26.6 = 1.33; e = 0.064(20)
20
= 1.28 in.; hi = 8.72 in. Stress-lb

Item Compo Meas. Camp.

Maximum fiber stress (tension): I I Meas. Camp.
I Meas. Camp.
I Meas.

Timoshenko: 192 psi 285 psi 385 psi 510 psi

Mill T 1,530,000(8.72) 98,300 . Plate 1

fmnx = Aea +:Ii = 20 (2.50)(1.28)(16.63) + 20(2.50) = 14,500 pSI
A-I 11,600 12,000 17,240 16,896 23,350 21,030 29,300 27,500
Seely: A-2 -5,600 -7,400 -8,310 -12,288 -11,150 -16,450 -13,900 -21,000

Z = 2~~6 = 2.66; K = 1.35

B-1
B-2
C-l
8,100
-5,000
8,052
-7,800
12,025
-7,470
17,480
-9,248
16,240
-10,090
24,010
-11,590
22,700
-14,000
26,900
-13,900
-300 -4,200 -429 6,816 -485 9,840 -140 13,000
KMc T 1.35(1,530,000) 98,300 . C-2 -5,500 -5,350 -8,100 -10,016 -11,000 -9,290 -15,880 -15,140
fm= = -1- +:Ii = 0.167(2.5)(20)2 + 20(2.50) = 14,400 pSI D-1 -8,130 1,779 -12,100 1,768 -15,700 3,880 -21,780 3,080
D-2 8,070 8,150 11,980 11,984 16,600 16,810 19,900 22,240
E-1 -2,740 5,652 -4,060 7,184 -5,430 10,430 -7,270 6,800
E-2 4,150 2,110 6,190 -1,864 8,330 -1,616 10,160 -1,240
23it4 ;.. #1 -5,700 -7,050 -8,490 -10,650 -11,020 -14,400 -15,900 -18,000
#2 -9,000 -14,130 -19,100 -25,000
D.2+T -
#3 -2,100 -1,200 -3,157 -2,850 - 4,125 -3,600 -7,210 -2,100
#4 13,010 9,150 19,300 14,070 26,100 17,520 34,800 23,400
#5 6,100 6,150 9,020 10,500 12,210 16,500 16,830 22,200
#6 5,400 8,250 7,930 10,500 10,700 15,150 15,250 20,100
#7 -5,930 2,850 -8,779 1,950 -12,104 3,150 -15,600 5,400
#8 -7,280 -10,800 3,150 -14,600 4,950 -19,300 6,450

Plate 2

F-1 7,940 16,260 11,520~ 20,968 15,910 31,590 +18,730 41,210

F-2 -8,800 -9,940 -12,480 -15,728 -17,640 -20,080 -18,570 -26,900
G-1 5,280 15,500 7,810 20,232 10,540 27,260 +12,740 35,980
G-2 -8,940 -8,970 -12,690 -13,888 -17,940 -18,710 -20,400 -24,000
H-l 7,140 10,740 10,560 14,032 14,300 19,900 +19,300 25,000
H-2 -9,840 -6,520 -14,578 -9,832 -19,690 -26,500
J-l 390 1,200 977 2,472 1,310 -2,850 440 2,010
J-2 -530 -1,315 -1,765 -1,580
#9 -9,250 -9,600 -13,130 -13,800 -18,480 -19,200 - 21,430 -21,800
#10 -8,050 -8,700 -10,270 -13,200 -16,070 -17,250 -19,130 -21,900
#11 -9,600 -6,450 -14,300 -9,600 -19,300 -13,050 -25,600 -17,250
~ #12 9,540 13,200 13,700 20,550 19,060 28,350 23,150 37,650
#13 8,600 12,450 12,700 16,500 17,220 23,820 11,550 30,120
o 5 10 15
5,600
I I I I #14 9,000 8,053 12,150 11,000 17,100 -14,500 22,050
Sca Ie in Inches

Fig. 9. Two-PlateDesign-Plan of Stiffener Plates

Data: Rn = 21 in.; R, = 21 in.; t:.. = 53 deg; p = 192 psi. See italic note below Fig. 7.
604 H. S. SWANSON ET AL. WYE· ·BRANCH DESIGN 605

Three-Plate Design TABLE 4

Three-Plate Design Calculations (Example 1)

R 12.0 7' L
L, =~ = 0.7071 = 16.9 10.; L, = 2
sm
2

WI = ~Lhin {I, = 3~0 (16.97)(0.7071) = 50,400 lb

3
m = px ~n {I, = 35~X3 = 58.33x3

Half-Ring Plate M, = ."fiM, = ."fi(4yr) = 2{i(14.75)y = 13.28y

(Eq 2)
(Plate 3)
2 2... .•.
Fig. 10. Three-Plate Design .(Example 1)
Plates 1 and 2 (t = 1.5 in.; {l,s = 4.17 in.; d., = 9 in.):
Data: Rn = 12 in.; R, = 12 = 90 deg;
in.; D.. p = 350 psi; pipe shell thickness, i} iw.;
depth at crown, 5 in. Plate
x M = m - xy - Ml 1 x (.as) Mx(.as)
Depth in.J,. x (as) -1- -1-
in. ;n.-lb
in.
-- --- ---
Design Assumptions
1. Rotation of tangent to centroidal axis of
M. (n - 2yr(1) = 0
2.10
6.22
540- 2.10y-13.28y
14,037 - 6.22y-13.28y
5.0
5.15
15.62
17.08
8.76
25.94
0.561
1.519
300-
21,300-
8.63y
29.62y
10.27 63,175 -10.27y-13.28y 5.5 20.80 42.83 2.059 130,100- 48.49y
half-ring plate at crown equals zero; that is, 14.09 163.157 -14.09y-13.28y 6.1 28.38 58.76 2.070 337,700- 56.66y
this plate is free but guided at the crown. M. = 4yr (Eq 1)
n: 17.50 311,976 -17.50y-13.28y 7.0 42.88 72.97 1.702 531,000- 52.39y
2. Moments at crown of Plates 1 and 2 20.32 366,912 -18.59y-13.28y 8.0 64.00 77.52 1.211 444,300- 38.59y
equal ~ times the moment at crown of half- .
.h = deflection at crown = fMmds
~
22.06 366,912 -18.59y-13.28y 8.9 88.12 77.52 0.880 322,900- 28.04y

=Eo1
ring plate, from vector considerations.
Method oj Solution.
= J (- Ma + 2yx)xds 1,787,600 -262.42y

EI For half ring:

The three unknown quantrties, MI, M3,
and y, require three equations for solution.
= r~ (- M3 + 2yr sin 0) (r sin lJ)rdO
03 =
0.2976yr3
EI ; 01 = 03 (Eq 3)
JD EI
Two of these equations are obtained from the
two design assumptions, and the third is -4yr rf r sin OdO 8760 -262 = 0.2976yr3 = 0.2976y(3,209) = 8385
derived from the condition that the deflections =r;:», =sr: 1, 7 ,0 .42y I 11.39 . Y
of all three plates at the crown are equal.
Solution
rf ra sin' OdO Y = 5,162 lb; M3 = 4r(5,162) = 96,940 in.-Ib;
... MI = 68,550 in.-lb
+2y.Jo EI
For half-ring plate:
_ -4y,3 .•.y,3 _ yT3(:,,: _~)
a = .
rotation at crown JMmds
-m = - .•.EI + 2EI - EI 2 "
Maximum moment:

= J (- 114"3 + 2yx) (-1)ds = 0 = ;; (0.2976) 2yr - M, = 2(5,162)(14.75) - 96,940 = 55,340 in.-Ib

EI
Maximum tensile stress:
r~(- Ma + 2yr sin 0) (-I)rdO Computation of Maximum Stresses
JD EI Half-ring plate: Mhl + '£ = 55,340(2.25 - 0.115) + 2(5,162) = +12170 + 1 530 = +13700 psi
Aea A 1.5(4.5)(0.1151)(12.5) 1.5(4.5) , , ,
= r~ (M3r - 2yr sin O)dO = 0 ! = 14.75 = 3.28
Jo EI It 4.5 Maximum compressive stress:
7C

(MarO - 2yr2 cos 0)2 = 0 e = 0.1151 in. Mhz + '£ = 55,340(2.25 + 0.115) + 2(5,162) = -9910 + 1 530 = -8380 psi
Aec A 1.5(4.5)(0.1151)(17.0) 1.5(4.5) , , ,
606 H. S. SWANSON ET AL. Jour.AWWA WYE BRANCH DESIGN 607

Maximum tensile stress:

2" + 1"
C-2
Mil, + I = 202,290(4.5 - 0.536) + SO,400 - 5,160
Aea A 1.5 (9.0)(0.S36)(8.49) 9.0(1.5)
17&,~ +13,050 + 3,350, = +16,400 psi
S-2 1"
Maximum compressive stress:

MII2 + I = 202,290(4.5 + 0.536) 3350 = -8050 3350 -4,700 psi

Aee A 1.5(9,0)(0,536)(17.49) + , , +,
o 1 234 5
I I II II
Scale in inches
A-I A-2
' . Plate 3
I ~6~"~
1+------]6~"-------.l.~9'1~
Plate 2

Plate 1
Stress*-ll> Stress*-lb ~
ltem ltem
Computed Measured Computed Measured
Half-Ring Plate (Plate 3)
A-1 7,360 6,840 #4 8,380 6,150
A-2 -3,160 -3,260 j #5 -8,860 -8,850 Fig. 12. Three,Plate Design (Example 2)
B-2 -2,080 -730 1 #6 -20 -1,650
C-2 7,780 5,090 #7 4,000 6,510 Deta: Ru = 30.5; Rs=30.5; 6.=45 deg; p=300 psi; pipe shell thickness, 7
1 6 in.;
D-1 8,610 11,780 #8 12,360 11,350 depth at crown, 12 in.
:if] 6,950 7,350 #9 2,750 3,600
:if2 -2,030 3,380 ~10 -6,000 -3,450
*3 -4,400 -3,4S0 #11 -8,380 -6,900
Design Assumptions Solution
• At J50-ps; pressure.
1. Moment at crown of Plate 1 (M,) equals For half-ring plate:
Fig. 11. Three·Plate Design (Example I)-Plan of Stiffener Plates zero.
See italic note belotn Fig. 7.
2. Rotation of tangents to centroidal axes
of Plates 2 and 3 at crown equals zero; that is,
'" = .
rotation at crown = f -m
Mmds

theseplates are free but guided at the crown.

= J (- JIt[, + YaX)(-l)ds =0
Horseshoe plate (Plates 1 and 2) : Afethod of Solution EI
The five unknown quantrties+-y., Y2, Y3, 1r

a = R sin ~ = 12.0(0.7071) = 8.49 in. M2, and M3-require five conditions for solu-
tion, They are:
102 (- M. + s« sin 0) (-l)rdO
!:h = 8,485 + 4.5 = 1443' = 0536 . 1. Rotation of tangent to centroidal axis of .. ..
9.0 " e . In.
Plate 2 equals zero at crown.
2. Rotation of tangent to centroidal axis of
= J,' M,rdO - J,' Y3r2 sin OdO = 0

Maximum moment: Plate 3 equals zero at crown. ")f" ••

3. Deflections of Plates 1 and 2 at crown

+ (Y3y2 cos 0)' = 0
w l
(Ll3,. + 0 . 72*' + d10
w
) _
Y
(L
1
+ 0 72* + d10
•
w
) _ "/2
2
M
3
areequal.
4. Deflections of Plates 2 and 3 at crown
(M3rO)'

areequal. M3r?: - Y3y2 = 0

S. Shear at end of Plate 3 equals sum of 2
= S0,400C6~97 +0,72 + :0) - 5,162(16.97 + 0.72 + :0) - ~ (96,940) shearsat ends of Plates 1 and 2.
M« = 2Y3Y = 2 (Y3) (36.875)
= 366,760 - 95,920 - 68,550 = 202,290 in.-lb The method of virtual work is used to ob- 7r 1f'
tain equations based on the first four con-
:tThis additional distance was made necessary in the fabrication of the wye by the requirements for fitting the (Eq 2)
plates together properly. ditions. = 23.475Y3
 1-
,
608 H. S. SWANSON ET AL. J01~r.A WWA lime 1955 WYE BRANClI DESIGN 609
TABLE 5 TABLE 5 (contd.)
Three-Plate Design Calculations (Example 2) Plate 2-Equation for Zero Ratation at Crown

R 30.5 7 7 . R 30.5 0 . x Plate

L, =~ = 0.3827 = 9.0 In.; L2 = ~ = 0.9239 = 33.1 In. in.
M
·in.-lb Depth
[ (-1)
--1-
(Il.s) M(-l) (6.s)
in.i
Sln cos in. [
2 2 -- --_·_-------1--- ---1- 1
5.4 5,600-- 5.4y-M2 12.5 163 -0.06564
L, (scaled from Plate 1) = 82.72 in.; L2 (scaled from Plate 2) = 33.53 in. 15.8 - 368 +0.354y. +0.0656M2
139,500-15.8y-M, 15.0 281 -0.03808
25.8 - 5,312 +0.602Y2+0.0381 M.
607,300-25.8y-M. 17.7 462 -0.02316
34.1 -14,065 +0.S98Y2+0.0232M.
1,398,800-34.1y-M. 19.9 657 -0.01629
35.2 -22,786 +0.S55y,+0.0163M2
1,525,900 -35.2y- M. 21.2 794 -0.01348
35.2 -20,569+0.474Y2+0.0135M2
1,525,900 -35.2y- M. 22.0 887 -0.01206 -18,402+0.425y,+0.0121M.

-81,502+3.008Y2+0.1687M2
m = px3 sin 1'1 = 300x3(0.707) = 35 36x'
66' _ M( -1)(l'1s) - 81,502 + 3.008y, + 0.1687 M2
ex - EI = E = 0

M. = 81,502 - 3.008Y2 _
x
in,
).,f = 1n

in.-lb
- XYI
I Depth
Plate I [
in.4
x(6.s) I x(~s) =r=
Mx(ll.s)
• 0.1687 - 483,100 - 17.83y, (Eq 1)
I 'In.
I I
Plate 1 (t = 1 in. (assumed); I'1s = 15.1 in.; dw = 54 in.)

7.6 15,500 - 7.6Yl 12.0 144 114.8 0.7972 12,400- 6.06y,

0, = deflection of crown = f Mmds
EI
= f (-M, + y"x)xds
EI
I
22.7 413,600 -22.7Yl 13.2 192 342.8 1.7854 738,400- 40.53Yl
37.8 1,909,800 -37.8Yl 15.5 310 570.8 1.8413 3,516,500- 69.60Yl 1 r~ . 1:":
52.9
67.7
5,234,600
10,971,800
-52.9Yl
-67 .7Yl
19.4
25.4
608
1,366
798.8
1,022.3
1.3138
0.7484
6,877,200-
8,211,300-
69.50Yl
50.67Yl
= EI J 0 (- M,T Sill O)rdO + EI f: (y,r' sin' 8)rd8
82.2 19,584,500 -82.2Yl 35.5 3,728 1,241.2 0.3329 6,519,700- 27.36Yl
85.1
85.1
21,538,200
21,538,200
-85.1Yl
-85.1Yl
48.0
53.3
9,216
12,618
1,285.0
1,285.0
0.1394
0.1018
3,002,400-
2,192,600-
11.86Yl
8.66YI
= ~
E1
(M r'
3 cos
8)~ + E11 [
y,r
3 (
:28 - sin ()cos ())
2
]f
---------
31,070,500 - 284.24YI = ~ (_ M,r' + 1rr3Y3) = ~ ( -2y,r' + 7T'r3y,) = r3Y3(~ _~)
E1 4 EI 1r 4 E1 4 n:

Plate 2 (t = 1 in. (assumed); I'1s = 10.7 in.; db = 22 in.) _ 0.1488r3Y3 _ 0.1488(36.875)3y,(12) = 62.88y,
- EI - E(1.0) (11.25)' E
5.4 5,600- 5.4Y2-M2 12.5 163 57.8 0.3546 2,000- 1.91Y2-0.3546M,
15.8 139,500-15.8Y2-M, 15.0 281 169.1 0.6018 84,000- 9.51y,-0.6018M, NOTE: In solving simultaneous equations for shears and moments at the ends of plates
25.8 607,300-25.8Y2-M, 17.7 462 276.1 0.5976 362,900 -15,42y,-0.5976M, (crow~), all plates ar~ taken to be 1 in. thick. As I occurs in the denominator of all three
34.1 1,398,800 -34.1Y2-M, 19.9 657 364.9 0.5554 776,900-18.94y,-0.5554M, deflection equ~tlOns, It.does not matte; what thickness is used, so long as all three thicknesses
35.2 1,525,900 -35.2y. - M,. 21.2 794 376.6 0.4743 723,700 -16. 70y.-0.4743M, areequal. VI.hen solving for stresses III the plates, the actual thicknesses will be used.
35.2 l,525,900-35.2y.-M, 22.0 887 376.6 0.4246 647,900-14.95y.-0.4246M,
Plate 1 :
2,597,400-77.43Y2-3.008M,
Etol = 31,070,500 - 284.24YI
Plate 1: tEal = 31,070,500 - 284.24Yl Plate 2 :
Plate 2: tEa, = 2,597,400 - 77.43Y2 - 3.008M. Eto. = 2,597,400 - 77.43y. - 3.008M.
Plate 3 : tEo, = 62.88y, = 2,597,400 - 77.43Y2 - 1,453,200 + 53.63Y2 = 1,144,200 - 23.80y,
Plate3:
su, = 62.88Y3
31,070,500 - 284.24Yl = 1,144,200 - 23.80y, (Eq 3)

,J,._
610 H. S. SWANSON ET AL. Jour. A W WA 1 J~ne19S5 WYE BRANCH DESIGN 611

284.24YI - 23.80Y2 - 29,926,300 = 0 Maximum moment:

1,144,200 - 23.80Y2 = 62.88y, (Eq 4)
W2( ~2 + 0.52* + ~~) - Y2(L2 + 0.52 + ~~) - M2
Y3 = Yl + Y2 (Eq 5)

G2.88YI+ 86.68y, - 1,144,200 = 0 = 115,600 (33~01 + 0.52 + ~~) + 59,560 (33.01 + 0.52 + 2~~O) - 1,545,000
{ - (52.88YI - 5.26y, - 6,620,300 = 0)
91.94y, + 5,476,100 = 0 = 1,586,000 + 2,127,500 - 1,545,000 = 2,168,500 in.-Ib
Maximum tensile stress (Seely):
y, = - 59,560 lb : Y3 = + 40,740 lb ; YI = 100,300 Ib
KMc !. = 1.28(2,168,500)(6) 115,600 + 59,560 = 3770 + 3 180 = '16950 si
M, = 23.47Sy, = 956,400 in.-lb J + A 2.5(22.0)2 + 22.0(2.5) 1 , , -r, p
M: = 483,100 - 17.83( -59,560) = 1,545,000 in-Ib Maximum compressive stress:

Computation of Moximum Stresses , -0.799(2,168,500) (6) 3 180 = _ 8 600 + 3 180 = 5420 .

2.5(22.0)2 +, , , -, pSI

Plate 1:
Piate 3 (half-ring):
a = R sin ~ = 30.5 (0.3827) = 11.68 in. r 36.875" T

~= 5.625 = 6.J6; K; = 1.109; Ko = 0.910

r 11.68 +
27.0 7 . Maximum moment:
k= 54.00 = O. 16; e = 7.098 In.
Y3(r) - M3 = 40,740(36.875) - 956,400 = 545,900 in.-lb
Maximum moment: Maximum tensile stress:

Tf:
>r 1
(L''3 + 3 .02* + 10 dw) - y, ( L, + 3.02 + s;
10 )
KiMc
[
+!.
A
= 1,109(545,900)(6) +
2.5(11.25)
40,740
2.5(11.25)
= 11,480 + 1,450 = +12,930 psi

Maximum compressive stress:

= 673,700 C9~70 + 3.02 + 5:~0) - 100,300 (79.70 + 3.02 + 5:~0)
KoMc !. = -0.910(545,900)(6) 1450 = -9420 1450 = -7,970 psi
[ + A 2.5(11.25) +, ,+ ,
= 23,573,000 - 8,838,000 = 14,735,000 in.-Ib
TABLE 6
Maximum tensile stress:
Three-Plate Design (Example 2)-Comparison of Results Obtained Using
Mh, T 14,735,000(27.0 - 7.1) 673,700 - 100,300 . Three Different Assumptions
Aea + A = 54.0(2.5)(7.098)(11.68) + 54.0(2.5) = 26,200 + 4,250 = 30,450 PSI (a) All plates free but guided at crown (Sl = S2 = S3 = 0).
(b) Plate 3 free but guided at crown, others simply supported.
Maximum compressive stress: (c) Plates 2 and 3 free but guided at crown, Plate 1 simply supported.

Mh, T - 14,735,000(27.0 + 7.1) ; Stress-psi

Aecf + A = 54.0(2.5)(7.098)(42.5) +4,250 = -12,340 + 4,250 = -8,090 psi
Plate Location Computed
Plate 2: Measured
(a) (b) (c)
r 35.0t
~ = 11.0 = 3.18;K, = 1.28;Ko = 0.799
1 ou tside of crotch - 5,500 - 7,050 - 8,090 - 5,850
1 inside of crotch +26,900 +28,400 +30,450 +34,300
:f: This additional distance was made necessary in the fabrication of the wye hy the requirements for fitting the 2 ou tside of crotch - 6,600 - 6,780 - 5,610 - 6,150
plates together properly.
t c is the radius of curvature to the outside fibers (Timoshenkc). scaled in this case (see Fig. 22, page 626) owing 2 inside of crotch +19,500 +16,950 +16,950 +16,700
to ex-tremely nonconcentrrc curves. 3 outside of crotch -10,100 -13,680 - 7,930 -10,350
t Scaled.
3 inside of crotch +16,490 +22,330 +12,970 +15,750
3 outside of crown +18,960 +25,700 +15,800 +16,200
3 inside of crown -23,300 -31,500 -19,300 -29,100
612 H. S. SWANSON ET AL. JoItY.AWWA JUlie 1955 WYE BRANCH DESIGN 613

From these and other comparisons it may one another and provide the fixity needed by TABLE 7
be seen that the best resul ts are obtained each plate for zero rotation at the crown. Three-Plate Design Calculations (Example 3)
when using the design assumptions in which Plate 1, however, is much longer than either
Plates 2 and 3 are considered free but guided of the others, has less rigidity, and has no R 18.0 7· R 18.0 1948 .
(that is, slopes of tangents are zero at crown) opposing plate to restrain its rotation at the L, = --. II = -827
0.3 = 4 .04 111.; L2 = ~ = 0.9239 = . 111.
and Plate 1 is considered simply supported. crown. Therefore, even though it is not corn- S1l12' cas 2'
The justification for these assumptions, aside pletely simply supported, its degree of fixity
from better agreement with test results, is is relatively low, and for practical purposes is L, (scaled from Plate 1) = 49.5 in.; L, (scaled from Plate 2) = 19.75 in.
that Plates 2 and 3 are almost in line with considered zero.
W, = pLhin II = 335(47.04)2(0.7071) = 262000 Ib
2 - 2 '
W2 = pLhin II = 335(19.48)(0.7071) = 44960 lb
22'

= px3 sin II = 335x3 (0. 7071) = 39 48x3

m
66·

x
in.
M
in.-lb
I Plate
Depth 1
in.4 x(~s)
x(~s)
-1-
Mx(~s)
=r:
in.
Fig. 13. Three-Plate Design (Example 3)
I
Data: RB = 18 iw.; R, = 18 in.; .6. = 45 deg; p = 335 psi; pipe shell thickness, ~ in.; Plate 1 (t, = 2 in. ; M = 9.48 in. ; d", = 30 in.)
depth at crown, 12 in.
4.7 4,100- 4.7Yl 12.4 318 44.6 0.1403 600- 0.66Y1
14.2 113,000 -14.2Yl 13.0 366 134.6 0.3678 41,600- 5.22Yl
Design Assumptions 5. Deflection at crown of Plate 2 equals 23.7 525,600 - 23. 7Yl 14.3 487 224.7 0.4614 242,500 -1O.94Yl
1. Moment at crown of Plate 1 (M1) equals that of Plate 3. 33.1 1,431,700-33.1Yl 16.5 749 313.8 0.4190 599,900 -13.87Yl
zero. The equations for virtual work are applied 42.5 3,030,700 -42.5Yl 20.2 1,374 402.9 0.2932 888,600 -12.46Y1
2. Rotation of tangents to centroidal axes S1.Q 5,302,900 - 51.6y, 25.2 2,667 489.2 0.1834 972,600 - 9.46y,
to conditions 1, 2, 4, and 5:
of Plates 2 and 3 at crown equals zero; that is, 52.5* 5,538,700 - 52.5Yl 29.2 4,150 497.7 0.1199 664,100- 6.29Yl
these plates are free but guided at the crown. Eq 1: 52.5 5,538,700-52.5y, 30.0 4,500 497.7 0.1106 612,600 - 5.81Yl
M2 = 206,000 - 15.17Y2
Method of Solution 4,022,500-64.71y,
The method of solution is similar to that Eq 2:
used in Example 2, the main difference being M3 = 16.47Y3 Plate 2 (t2 = 1 in.; lls = 6.52 in.; db = 12 in.)
that Plate 3 is elliptical, not circular. In the
following analysis, five unknowns will be con- Eq 4: 3.2 1,300- 3.2Y2-M2 12.3 155.1 20.9 0.1348 200 - 0.43Y2 -O.135M2
sidered: y], Y2, and Y3 (the shears at the ends 4,022,500 - 64.71Yl 9.5 33,800- 9.5Y2-M2 12.2 151.3 61.9 0.4091 13,800 - 3.89Y2 -0.409M2
of Plates 1, 2, and 3, respectively); and M2 15.3 141,400-15.3Y2-M2 12.1 147.6 99.8 0.6762 95,600-10.35y. -0.676M2
= 1,079,400 - 72.61y, - 4.028M.
and M3 (the moments at the crown ends of 20.1 319,700-20.1Y2-M2 12.0 144.0 131.0 0.9097 290,800 -18.28Y2 - 0.9lOM 2
Plates 2 and 3). M, is assumed to be zero, 2,943,100 - 64.71y, 20.95 357,900-20.9Y2-M2 12.0 144.0 136.6 0.9486 339,500-19.83Y2-0.949M2
while M, and M3 are such that the tangents = - 72.61y, - 4.028(206,000 - 15.17y.) 20.95 357,900-20.9Y2-M2 12.0 144.0 136.6 0.9486 339,500-19.83Y2-0.949M2
to the plates at the crown remain at zero slope
under load. =- 72.61Y1 - 829,800+ 6l.lOY2 1,079,400 -72.61Y2 -4.028M2
These unknowns are solved under the fol- 3,772,900 - 64.71Yl + 11.51Y2 = 0
lowing five conditions:
Eq 5: Plate 3 (t3 = 1 in.; lls = 6.56 in.; db = 12 in.)
1. Slope of tangent to Plate 2 at crown
equals zero. 1,079,400 - 72.61y, - 4.028M, 21.6
3.3 0+ 3.3Y3-M3 12.0 144.0 0.1500 + 0.50Y3-0.150M3
2. Slope of tangent to Plate 3 at crown = 91.44Y3 - 4.500M3 9.7 0+ 9.7Y3-M3 12.0 144.0 63.6 0.4417 + 4.28Y3-0.442M3
eq uals zero. 1,079,400 - 72.61y, 15.5 0+15.5y,-M3 12.0 144.0 101.1 0.7062 + 10.95y, -0.706M,
3. Y3 = Yl + Y2 20.5 0+20.5y,-M, 12.0 144.0 134.5 0.9340 +19.15y,-0.934M,
4. Deflection at crown of Plate 1 equals - 4.028(206,000 - 15.17Y2)
24.0 0+24.0y,-M, 12.0 144.0 157.4 1.0933 + 26.24y, -1.093 M,
that of Plate 2. = 91.44Y3 - 4.500(16.47y,) 25.8 0+25.8y,-M, 12.0 144.0 169.2 1.1753 +30.32y, -1.175M,

+91.44y, -4.S00M,

* Maximum % distance is to point one-tenth depth of plate from inside edge.

614 E:. S. SWANSON 1<:'1' At.. iour.AWWA lune 1955 WYE BRANCH DESIGN 615

TABLE 7 (contd.) 2111

4 14~~#8 25~1I 37~~#7 49~"
lll
I
".:i- ..••••••
Equation for Zero Rotation at Crown +~l" 1-'1"
E-2 D-2 C-2

x
in.
.Nt = 1n - XYt
iIZ.-lb
- M 2" Plate
Depth I I
zn.~
(-1) (.:\.s)
I
I M(-l)
I
(Ll.s)
"<-
N
o
-

";;00
;;;Iq-

~ ~ :;;'N ;:1:': Plate Thickness = 2" B-2~.

T 1
t
in.
I IE_I #4 l+D'1 1 #3 C-1 ~ _ 7ok"-;oj~1I
1+ g!"..;i 14,%'~ 22!".l 13"" }+#2 ~ =I~
Plate 2
~l - 4--30i6.:::t.....372'!.i I B-1 ~N-

3" 16"",-- :1:... N

3.2 1,300- 3.2Y2-M, 12.3 155.1 -0.04204 -50+0.135y,+0.0420M2 44.!."->{ !#1
9.5 33,800- 9.5Y2-M, 12.2 151.3 -0.04309 1,460+0.409y,+0.0431M2
2"'51~'~
15.3 141,400-15.3y, - M, 12.1 147.6 -0.04417 - 6,250+0.676y,+0.0442M,
20.1 319,700-20.1Y2-M, 12.0 144.0 -0.04528 -14,480+0.910Y2+0.0453M,
20.95 357,900- 20.9y,-M, 12.0 144.0 -0.04528 -16,200+0.946y,+0.0453M, A-I
1"
A·2
20.95 357,900- 20.9Y2-M, 12.0 144.0 -0.04528 -16,200+0.946y,+0.0453M, f4 22"
'-.\. 30" ,]

o 5 10
] ] \
Scale in inches

_ '" Mml!.s _ ~ M( -l)~s _ - 54,640 + 4.022y, + 0.2652M, _ 0

Item
Stress*-lb Stress*-lb
"" - .t.. ~ - EI - E - Item
Computed Measured Computed Measured

M2 =
54,640 - 4.022y,
0.2652 = 206,000 - 15.17Y2 (Eq 1) , A-I 6,950 9,310 L-l 4,710 4,100
A-2 -2,340 -3,730 L-2 -630 -800
"'" = ~ Mml!.s = M(-I)(c.s) = - 4.500Y3 + 0.2733M 3 = 0 B-1 8,000 10,130 M-l 13,760 13,550
c EI - B-2 -3,440 -4,060 M-2 -6,400 -6,220
C-1 -3,890 -5,180 #1 7,760 5,460
4.500Y3 C-2 -5,570 -3,190 #2 1,160 3,750
Ma = 0.2733 = 16.47Y3 (Eq 2) D-l -7,870 -5,580 #3 -8,450 -1,290
D-2 9,250 8,380 #4 -5,980 300
E-l -1,630 -4,390 #5 -2,740 -4,800
E-2 1,990 2,440 #6 -3,920 -4,350
1,079,400 - 72.61y, - 829,900 + 61.10y, Y1 = +54,060 lb F-l 14,500 12,500 1,520
#7 2,400
= 91.44y, - 74.12Y3 Y3 = +30,240 lb F-2 -3,000 -4,420 #8 11,360 8,550
G-l 14,070 10,900 #9 -4,360 -3,810
249,600 - 11.51Y2 = 17.32Y3 M, = 567,400 in.-Ib G-2 -4,000 -420 #10 -800 2,460
Eq 3: M3 = 498,000 in.-Ib H-l -4,520 4,800' #11 13,950 9,210
H-2 6,300 7,010 #12 7,220 6,450
Y3 = Y1 + y, Computation of Maximum Stresses J-1 -14,310 -5,150 #13 -5,840 -4,590
249,600 - 11.51Y2 = 17.32(Y1 + y,) Plate 1: J-2 17,700 11,800 #14 -5,530 810
K-l -12,180 -12870 ~ #15 10,380 12,150
249,600 - 28.83Y2 - 17.32Y1 = 0
{ -1,009,800 - 3.08y, + 17.32Y1 = 0 a = R sin% = 18.0(0.3827) = 6.89 in. K-2 13,110 9:850 I
-760,200 - 31.91y, = 0
~ = 6.89 + 15.0 = 21.89 = 1.459 , * At 335·psi nressure.

y, = -23,820 lb c 15.0 15.0 Fig. 14. Three-Plate Design (Example 3)-Plan of Stiffener Plates
See italic note below Fig. 7.
616 H. S. SWANSON ET AL. rune 1955 WYE BRANCH DESIGN
JOf£r.AWWA

K, = 2.00; Ko = 0.641
----------NomographMethod----------
Maximum moment:
During the period of this study a them. There remained one other im-
LL
WL ( 3' + dw)
10 - Yl
( t., + d10
w) means was sought by which the data portant dimension of the reinforcing
on similitude of design could be pre- plate to consider-that at the top of
47.04 + 2.46* + 10 30) - 54,060 ( 47.04 + 2.46 + 30 10) sented. Curves had been made in the the sections. When this was investi-
= 262,000 ( -3- = 2,700,000 in.-lb past, but these proved cumbersome. gated, it was found to vary quite closely
Maximum tensile stress: Finally it was decided to present the in a good design with the smaller crotch
material in the form of a nomograph depth (db). By checking through all
K,Me T 2.00(2,700,000)(6) 262,000 - 54000 (Fig. 15). After the authors had in- the past designs, it was possible to plot
-1- + A = 2.0(30)2 + 2.0(30) , = 18,000 + 3,470 = + 21,470 psi
vestigated the possibility and developed another set of curves of top depth ver-
Maximum compressive stress: the form, a series of two-plate 90-deg sus base depth. One of these curves
wye branches were designed for pres- was based on the one-plate designs,
-K.Me T -0.641(2,700,000)(6)
--[- + A = (2.0)(30)2 + 3,470 = --:5,770 + 3,470 = -2,300 psi sures from 100 to 600 psi. The result- the others on the two-plate designs at
. ing crotch depths were found to follow various deflection angles (Fig. 18).
Plate 2: a curved pattern, and this was used to In examining the equations of simili-
/';, plot the interpolated points on the plate tude and their relation to this work, a
a = R cos 2' = 18.0(0.9239) = 16.63 in.
depth scale. Later the pressure scale method of converting the depth and
was extended to 1,200 psi by the same thickness of plate for curved beams
r 16.63 + 6.0 22.63
~= 6.0 = (5.0 = 3.77;K, = 1.220;Ko = 0.840 procedure. When the 90-deg wye was sought. Using the section modu-
branch design appeared complete, a lus equation did not give accurate re-
Maximum moment: series of designs was made at various sults in this case, because it did not
angles of deflection. These results hold the stresses constant as the plate
- M. W + 2 ( ~2 + :~) .- Y2 (L. + :~) were compared with the 90-deg de- was varied in cross section. By em-
signs, and there appeared to be a con- pirical methods a conversion equation
= -567,400 + 44,960 C 9 48
3 + 0.27* + ~~) - 23,820 (19.75 + ~) = 281,400 in.-lb stant proportion between the two at
any particular deflection angle. This
was derived by which the depth and
thickness of plate could be varied so
Maximum tensile stress: proportion, called the N factor, was that the depth was never more than 30
KiMe T 1.220(281,400) (6) 44.960 + 23 820 plotted against the deflection angle of times the thickness. This equation
-1- + A = 1.0(12)2 + 1.0(12.0) = 14,300 + 5,730 = 20,030 psi the wye branch (Fig. 16). was based on test results and was
Up to this point the authors had con- checked by being used on sections
Maximum compressive stress:
sidered only sections with equal diame- which were later tested. From this it
-KoMe T -0.840(281,400) (6) ters and two-plate designs. Next a was assumed to be reasonably accurate.
--[- +A = 1.0(12)2 + 5,730 = -9,850 + 5,730 = -4,120 psi series of one-plate designs was calcu- In order to check the nomograph,
Plate 3 lated for various ratios of pipe diame- complete the curves, and derive the
ters. When the resulting dimensions equations used, a great number of ex-
K; = 1.220; Ko = 0.840
of these designs were compared with amples was required. These were
Maximum moment: those of the two-plate designs, a rela- mostly based on theoretical knowledge.
- M. + Y. (r.. + ~) = -498,000 + 30,240 (19.48 + 0.52* + 1~.O) = 288,200 in.-lb . tionship was found that depended on
the deflection angle. In order to graph
Test results had been used throughout
as a basis for the design data, but it
Maximum tensile stress: the comparisons, it was necessary to was still necessary to make other tests
plot a family of curves at various de- to check the curves as they were
K.Me +!. = 1.220(288,200)(6) 30,240 _ _ .
[A 1.0(12)2 + 1.0(12.0) - 14,6.)0 + 2,520 = +17,170 PSI , flection angles on a graph of pipe di- .formed. In this series of complete
ameter ratio versus a proportion called tests on sections which were based on
Maximum compressive stress: the Q factor (Fig. 17). These curves the nomograph method of design, the
-KoMe T -0.840(288,200)(6) fellso close to each other that it seemed stresses in the critical sections were
--[-.- + A = 1.0(12)2 + 2,520 = -10,090 + 2,520 = -7,570 psi simpler to plot only those for 90, 60, analyzed to ascertain their effect on
and 30 deg and interpolate between any curves used in the design. From
* This added distance was made necessary during fabrication of the wye in order to lit the plates together.
 o 1( June 1955 WYE BRANCH DESIGN 619
210
200
this information the curves were ad- and db should be converted to conform
190 justed by making the design pressure to a greater thickness, t, by use of the
12 180 that pressure which gave the section a general equation:
stress of 30,000 psi at its most critical
170
point. An effort was made to balance d = d1...1: t )0,9[7 - ~ 360
160
the stresses in the various parts so ( t
150 that each had a stress close to the criti-
24 cal stress in magnitude. in which d, is the existing depth of
140
On the following pages the nomo- plate; t, is the existing thickness ()1
130 plate; d is the new depth of plate;
30 graphic method of design is illustrated,
120 along with examples and test results. t is the new thickness of plate selected;
36
Superimposed on the drawings of the
110
tested sections are the stresses which
100 occurred on the inner and outer fibers
42
of the reinforcement plates. The ten-
90 sion stresses were plotted outward and
48 the compression stresses inward.
80

Nomograph Design
~ 54 70 .~
I Step 1. Lay a straightedge across
1i ~I

0:: ~ the nomograph (Fig. 15) through the

'060 60 a:
~ '0 appropriate points on the pipe diame-
s:
E ~ ter (see Step 2b) and internal-pressure
g 66 50 0
scales; read off the depth of plate from
its scale. This reading is the crotch
72 40 depth for l-in. thick plate for a two-
plate 90-deg wye branch pipe.
Step 2a. If the wye branch deflec-
30 tion angle is other than 90 deg, use the 1.0 2.0 3.0 4.0 5.0
N Factor
N factor curve (Fig. 16) to get the
84
factors which, when multiplied by the Fig. 16. N Factor Curves

20 depth of plate found in Step 1, will

give the wye depth, dw, and the base
depth, db, for the new wye branch.
96 Step 2b. If the wye branch has
unequal-diameter pipe, the larger-
10
diameter pipe will have been used in
Steps 1 and 2a, and these results
should be multiplied by the Q factors
108 found on the single-plate stiffener
curves (Fig. 17) to give d'w and d'b.
These factors vary with the ratio of
the radius of the small pipe to the
radius of the large pipe.
120 o
Cl Step 3. If the wye depth, dw, found
Fig. 15. Nomograph for Selecting Reinforcement Plate Depths
f so far is greater than 30 times the
For wyes with deflection angles [rom 30
to 90 deg, the N factors obtained from
the above curves are applied to the plate
depth, d, found from the nomograph (Fig.
15), in accordance with the equations:
dw = Nwd; db = Nbd.
and 6. is the deflection angle of the wye
branch.
Step 4. To find the top depth, dt
or d't, use Fig. 18, in which dt or d't
is plotted against db or d'b. This di-
mension gives the top and bottom
depths of plate at 90 deg from the

I
of Equal-Diameter Pipes thickness of the plate (1 in.), thend; crotch depths.

Plate thickness, 1 in.; deflection angle, 90 deg.

620 H. S. SWANSON ET AL. JOllr.AWWA June 1955 WYE BRANCH DESIGN 621

l-
r- V V V
I-
/ / /
40

-
-
Vv V/ V /

35
-
- ijl / IV/ V
~
V
l-
I-
I-
I~/WfV e>/
;;?' Q..,"b roC) oflJ~

7
-, Q:' / ~'
~'b q
30 'I. !'--- ,,'"
l- V 'I.'~

IIVvl/V
I-
~
l-
25

,,~
I-
I--

l-
II II V
£
~ 20
r-
I// ~

V
o
a.
~
r-
l- /
5
r-
r-
I V ~

.l-
~
r-
'-
I10 ~

0
r- 'j
Fig. 17. Q Factor Curves I-

For pipes of unequal diameter, find dw and db for the larger-diameter pipe (from Fig.
=
15 and 16); then: QW'dw d'w, crotch depth of single-plate stiffener; and Qbdb = d'b' ~hW
base depth of single-plate stiffener.
'f
Step S. The interior curves follow The important depths of the re-
the cut of the pipe, but the outside inforcement plates, dw, db, and dt (Fig. 0 I I I I I I I I I I I I I I I I
crotch radius in both crotches should 19) can be found from the nomograph. . 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Base Depth, d. or d '. - in.
equal d, plus the radius of the pipe, or, If a curved exterior is desired, a radius
in the single-plate design, d't plus the equal to the inside pipe radius plus dt Fig. 18. Selection of Top Depth
radius of the smaller pipe. Tangents can be used, both for the outside curve
connected between these curves com- of the wye section and for the outside d't and d' b are one-plate design dimensions; d, and db are two-plate design dimensions.
plete the outer shape. curve of the base section.
622 H. S. SWANSON ET AL. J01lr.A WWA WYE BRANCH DESIGN 623
Iune 1955

Step J. As dw is greater 'than 30 times the If a third plate is desired as an addi-

thickness of the plate, try t = 2 in. in the tion to the two-plate design, its size
conversion equation: should be dictated by the top depth
_
d - d,
( t,
-
)0.917 - 360 _-
LI

dt
(!)0.770 (de). Because the other two plates are
t 2 flush with the inside surface of the
Wyc Section Base Section
= d,(0.586)
pipe, however, the shell plate thickness,
plus clearance, should be subtracted
dw = 96.5(0.586) = 57 in.
from the top depth. This dimension
db = 53.4(0.586) = 31 in.
should be constant throughout, and the
Fig. 19. Plan and Layout
Step 4. Read the top depth, dt, from the plate should be placed at right angles
two-plate design curve in Fig. 18:
to the axis of the pipe, giving it a half-
Ex:ample 1-0ne-Plate Design Try a thickness of 1! in.: dt = 15
ring shape. Its thickness should equal
Final results:
RB = 30 in. _ (1- )0.917-3':0
d - d, = d)
(2- )0.792 Thickness of reinforcing plate, t = 2 in.
the smaller of the main plates.
R. = 21 in. 1.5 3
Depth of plate at acute crotch, dw = 57 in.
~ = 45 deg Depth of plate at obtuse crotch, db = 31 in. References
d = d,(0.725)
Depth of plate at top and bottom, d, = 15 1. Force Mains. Sulzer Brothers Catalog
Working pressure, 230 psi d'", = 63.4 (0.725) = 46 in. in. (1931).
Design pressure, 230 X 1! = 350 psi Outside radius of plate at both crotches,
d'b = 40.5(0.725) = 29 in. 2. Sulzer Patented Stiffening Collars on
51 in. the Branches of High-Pressure Pipe-
Step 1. With the larger pipe diameter 60 Step S. 'Find the top depth, d't, from the
in. and the design pressure 350 psi, read the lines for Hydroelectric Power Works.
curve for one-plate design in Fig. 18: Three-Plate Design Sulzer Tech. Reu., No.2, p, 10 (1941).
critical plate depth, d, from the nomograph
(t = 1 in., /1 = 90 deg): for d'» = 29 in., d', = 18 in. The preceding nomograph section 3. BLAIR, J. S. Reinforcement of Branch
Pieces. Engineering (Br.), 162:1
d. = 50 in. Final results: has covered the design of one- and
(Jul. 5, 1946).
two-plate wye branches without touch- 4. BLAIR, J. S. Reinforcement of Branch
Thickness of reinforcing plate, t = 1 tin.
Step 2. Using the deflection angle 45 deg, I ing on a three-plate design because of Pieces. Engineering (Br.), 163 :63
Depth of plate at acute crotch, d' w = 46 in.
find the factors on the N factor curve which (Jan. 17, 1947).
Depth of plate at obtuse crotch, d'b = 29 in. its similarity to the two-plate. The
will convert the depth found in Step 1 to apply 5. MURPHY, GLENN. Similitude in Engi-
Depth of plate at top and bottom, d', = 18 function of the third plate is to act like
to a 4.5-deg wye branch (t = 1 in.): neering. Ronald Press Co., New
in.
a clamp in holding down the deflection York (1932).
dw = K"d = 2.45 (50) = 122 in. Outside radius of plate at both crotches
equals the top depth plus the inside of the two main plates. In doing so, it 6. BIER, P. J. Welded Steel Penstocks De-
db = Kbd = 1.23(50) = 61.5 in. radius of the smaller pipe, d','+ R, = 18 accepts part of the stresses of the other sign and Construction. US Bur.
+ 21 = 39 in. . plates and permits a smaller design. Reclamation Eng. Monographs, No.3
Step 3. With the ratio of the smaller pipe (Jul. 1949).
radius divided by the larger pipe radius This decrease in the depths of the two
Example 2-Two-Plate Design 7. SEELY, F. B. Advanced Mechanics of
R' 21) . main plates is small enough to make Materials. John Wiley & Sons, New
( R8 = 30 = 0.70 and the deflection angle it practical simply to add a third plate
RB = R. = 36 in. York (1932).
(/1 = '105 deg ), use Fig. 17 to find the Q factors D. = 53 deg to a two-plate design. The additional 8. HETENYI, M. Handbook of Experi-
which give the crotch depths for a single- plate should be considered a means of mental Stress Analysis. John Wiley &
plate pipe wye stiffener (I = 1 in.): Working pressure, 150 psi Sons, New York (1950).
holding down the deflection at the 9. Graphical Method for the Determina-
Design pressure, 150 x i; = 225 psi
junction of the plates. The two fac-
Q" = 0.52 tion of Stresses From Strains on
Step 1. With the pipe diameter 72 in. and tors which dictate the use of a third Three Intersecting Gage Lines, and
Qb =

=
0.66
= 63.4 in.
a pressure of 225
psi, read the critical depth of plate are diameter of pipe and internal Its Application to Actual Tests. 1.
d'" 0.52(122) plate from the nomograph (t = 1 in., ~ = 90 Research N atl. Bur. Stds., 10: (May
pressure. When both of these are
d'b = 0.66(61.5) = 40.S in. deg) : 1953).
above a certain limit, a ring plate can 10. TIMOSHENKO, S. Strength of Materials.
d = 49 in. be used advantageously. During the D. Van Nostrand Co., New York
Step 4. As the depth, d'"" is greater than
30 times the thickness, t, the conversion equa- work on this report, it was concluded (1930).
Step 2. From the LV factor curve, find the
tion should be used: that these limits could be 60-in. ID 11. PARIS, J. M. Stress Coefficients for
two factors at D. = 53 deg; then, at t = 1 in.:
Large Horizontal Pipes. Eng. N etus-
and 300 psi. If either of these factors
d = d)..!.t )0.917-~ a. dw = 1.97 (49) = 96.5 in.
is below the limit, the designer should
Rec. (Nov. 10, 1921).
12. Unfired Pressure Vessel Code. Am.
(
t d, = 1.09(49) = 53.'10 in, be allowed to choose a third plate. Soc. Mech. Engrs., New York.
 U"'J
ti. S. SWANSON ET AL. WYE BRANCH DESIGN
!our.AWWA JUlie 1955

------------Test Results------------

One-Plate Wye Two-Plate Wye

A single curved plate serves as reinforce- This 30 X 30-in., 90-deg wye section has
ment for this 30 X 61-in., 45-deg w)'e two crotch plates (see Fig. 21).
section (see Fig. 20).

4/L7 12" tiS 12"

o 10 20
I I I
Scale in inches 21*"-------.-
o 5 10
1 I I
Scale in inches

Measured Measured Measured =

Item Stre.5s*-lb Item Stress"---lb Item
Stress*-lb Measured
Measured Item Stress*-lb
Measured Item Stress*--lb
;1 32,000 1*11 22,000 ;;21 -8,000 Item Stress*-1b
>::2 34,000 1*12 23,000 #22 -9,000 ;;16 5,000
jf9 -7,000
#3 31,000 ;;13 -12,000 ;:;23 -12,000 ~1 30,000 jf17 -4,000
#10 -8,000
;:;4 12,000 #14 -[2,000 #24 -12,000 jf2 30,000
-7,000 jf18 -1,000
-8,000 -12,000 30,000 #11 -1,000
~5 illS - 11,000 #25 jf3 -7,000 jf19
;::6 -20,000 :fi16 -10,000 jf26 28,000 j:f4 24,000 #12 6,000
-7,000 #20t
;7 :fi17 -11,000 15,000 #13 15,000
7,000 #27 0 li5 -6,000 jf21t
;8 21,000 -4,000 j;:28 22,000 11,000 #14 6,000
\:18 li6 -6,000 #22t
;9 25,000 10,000 3,000 7,000 #15
#19 il29 #7
:;110 23,000 #20 1,000 #8 14,000
I
* At t Principal stresses of rosettes.
* At 2S0-psi pressure. 360-psi pressure.

Fig. 20. One-Plate Design-Plan of Stiffener Plate Two-Plate Design-Plan of Stiffener Plates
Fig. 21.
See italic note below
Data: RB = 30.5 in.; R, = 15 ir~.; A = 45 deg; p = 250 psi. See italic note below Data: RB = 15 in.;
R =15 in.; 6,.=90 deg; p=360 psi.
SFig 7. See also Table 8 (page 627).
Fig. 7.
626 H. S. SWANSON ET AL. Jour.AWWA June 1955 WYE BRANCH DESIGN 621

#27

Three-Plate Wye

In this 61 X 61-in., 45-deg wye, the plate

on the back is not welded to the pipe shell
(see Fig. 22).

Tension
n:'i7:;'';/''3
""I,,'2ifj" >
""I I'
Plate Thickness =2t"
,~
o 10 20 30
I I I I
Scale in inches

TABLE 8
Measured Stresses for 30 X 30-in., 90-deg Two-Plate Wye
(See Fig. 21)

Stress-lb*
Gage
Measured Item Measured Measured
Item Stress*-lb Stress*-lb Item
Stress*-lb 100 psi 200 psi 300 psi 360 psi 400 psi

:jj:l 25,000 :jj:16 22,000 #30 -6,000 #1 9,400 18,500 26,500 30,400 33,000
:jj:2 34,000t #17 -8,000 #31 -7,000 #2 8,600 17,300 25,400 29,500 32,300
:jj:3 27,000 #18 -6,000 #32 16,000 #3t
:jj:4 16,000 #19 -9,000 #33 3,000
:jj:5 12,000 #20 -10,000 #34 -31,000 '1 #4
#5
5,600
3,300
13,100
7,500
20,000
12,300
24,200
15,200
27,000
17,100
:jj:6 4,000 #21 -10,000 #35 -10,000 #6 2,500 5,600 9,400 11,300 12,500
:jj:7 -4,000 #22 -15,000 #36 -2,000 #7 1,700 3,500 5,400 7,100 8,300
:jj:8 -8,000 #23 -13,000 #37 17,000 #8 2,900 7,100 12,100 14,400 16,000
:jj:9 -5,000 #24 23,000 #38 -3,000 #9 -1,700 -3,800 -5,600 -7,100 -8,100
~10 -1,000 #25 24,000 #39 -1,000 :jj:10 -1,700 -4,100 -6,000 -7,600 -8,600
~11 5,000 #26 0 :lf40 4,000 :jj:11 -1,500 -3,600 -6,000 -7,400 -8,300
~12 13,000 #27 -2,000 :lf41' 28,000 #12 -1,700 -3,600 -5,400 -7,000 -8,000
~13 21,000 #28 -5,000 :lf42 1,000 :jj:13 -1,800 -3,800 -5,400 -6,800 -7,800
~14 21,000 #29 -6,000 :lf43 11,000 :jj:14 -1,500 -3,600 _ -5,000 -6,300 -7,100
#15 21,000 :jj:15 -1,200 -3,300 -4,700 -5,800. -6,600
:jj:16 -1,200 -3,300 -4,500 -5,400 -6,000
* At 335-psi pressure. t 22,000 at 200 psi. :jj:17 -800 -2,300 -2,900 -3,600 -4,100
#18 - 200 ~ 900 -1,100 -1,500 -1,700
Fig. 22. Three-Plate Design-Plan of Stiffener Plates #19 - 200' - 800 - 800 -1,200 -1,400
#20t 2,500 3,100 4,900 5,700 6,200
Data: RB=30.5in.; R, = 30.5 in.; /:;,=45 deg; p=300 psi. See italic note below #2 It 4,200 7,800 12,300 14,800 16,500
Fig. 7, #22j 1,600 3,000 . 4,900 5,600 6,100

* Minus sign indicates compression.

t Gage No. 3 failed but was considered to have the same stresses as No.1 because the two were symmetrically
placed.
1: Rosettes.
 1-
628 H. S. SWANSON ET AL. Jour.AWWA I: Jfjne1955 WYE BRANCH DESIGN 629

o 0.40
-----------------AppENDIX A,-----------------
Comparison of Sulzer and Nomograph Design
TABLE 9
12
Comparison of Designs
0.35
In 1941 Sulzer Brothers reported
(2 *) a wye branch which withstood Item Sulzer
Nomo-
graph
an internal pressure of 2,410 psi before
it failed. For the sake of comparison, Main pipe diameter-in. 23.6 23.6 24

their data were converted from the Branch pipe diameter-in. 9.8 9.8
0.45
Deflection angle-deg 60 60
metric system into inches and pounds, Service pressure--psi 750 30 0.30
and the present authors' nomograph Test pressure-psi 1,140 1,140
method of design was used to obtain Failure pressure-psi 2,4,10
Shell stress at failure-psi 64,000 36
values for overall length and width of 0.40
the reinforcing plate. The results are Thickness of reinforcing plate 0.79 0.79
compared in Table 9. 42
-in.
Overall length of collar-in. 31.5 35.2 0.25
Overall width of collar-in. 26.3 27.4 48 0.35

.5
I
.S
I 54 ~
'"
----------------~APPENDIX B;------------------ c:'"
c.
<, il:
0.30

Deflection Nomograph
~. '0
60 <, ~ 0.20

I ~
E <, e
.2
~
The nomograph in Fig. 23, based on ample represented by the broken line- '"
0
66
<, w
c
test data, can be used for finding de- a 54-in. ID welded-steel pipe at a de- I <, 0.25
flections that will occur at the junction sign pressure of 350 psi-the deflection <,
of the reinforcement plates in two- and for a three-plate design is found to be 72
<, 00 0.15
three-plate wye branches. In the ex- 0.068 in.; for a two-plate, 0.098 in. ,0

<, <:l,
r::,O 0.20

-----------------APPENDIX C----------------- <,

84
Results of Overstress
0.15
0.10
During the testing of the various at the crotch appeared to act in an er-
wye branch pipes in this report, there ratic manner. When the internal pres-
was opportunity to observe their ac- sure was released, a number of gages 96
tions when overstress occurs. In the failed to return to their zero reading.
later tests, when the designs were cut Examination of these gages showed
to a minimum and strain gages were that the steel in their vicinity was per- 0.05
placed inside the sections, the critical manently deformed. In still other
108
areas of stress reached the yield point areas, the steel was under compression.
very close to the test pressure. When It appeared that, under high stresses,
additional pressure was applied to the .. the section acted in a plastic wanner
section, the strain gages on the inside and distributed stresses from critical
120
o
* See reference list, page 623. areas to less critical ones by yielding. o

Fig. 23. NomogJ'aph for Dellections in 45-deg Wyes

630 H. S. SWANSON ET AL.
Then, when the pressure was released,
the yielded metal was put into com-
pressiori., along with nearby areas
which had not been overstressed. The
condition that remained was a pre-
stressed critical area which was under
a permanent compression at zero pres-
sure. When this section is next put
under internal pressure, the apparent
stress will have to exceed the residual
compressive stress before the area .will
Correction
The Tentative Standard Specifications for Quicklime and Hydrated Lime-
AWWA B202-52T (January 19S3']oURNAL) contain an error.
4C.2.3, the amount of concentrated hydrochloric acid is given as 37.5 m!. The
correct figure is 87.5 ml.
Jou,,..AWWA
actually be stressed in tension. Thus,
the elastic limit of the section as a
whole will be raised.
Several sections were tested far be-
yond their test pressure in order to
watch for visible signs of failure or
deformation in the steel. Lueder's
lines appeared in the reinforcement
plates at about twice the pressure
which first brought the critical areas
to the yield point.
In Paragraph
New Method for Detection of Coliform
Organisms
--------A. A.HajnaandS. R.
Damon--------
A contribution to the Journal by A. A. Hajna, Bacteriologist, and
S. R. Damon, Director, both from the Bureau. of Labs., State Board
of Health, Indianapolis, Ind.
IN estimating the extent of coliform-
organism contamination
supplies, a presumptive
of water
test medium
fered broth containing glucose and in-
corporating sodium desoxycholate,
a presumptive medium, would not be
as
containing lactose is usually utilized in more suitable than that containing lac-
the examination, because the coliform tose for the examination of water.
group ferment lactose rapidly with the The present report is based on a com-
production of acid and gas, thus serv- parison of primary inoculation in lac-.
ing as a useful pollution index. tose broth; prepared according to
Hajna (1) proposed, in 1951, a buf- Standard Methods (3), followed by
fered desoxycholate lactose broth for confirmation in brilliant-green Iactose
use as either a presumptive or con- bile (2 per cent), with primary inocu-
firmatory medium in the examination lation in the new buffered desoxy-
of water and food. The inclusion of cholate glucose broth, followed by con-
sodium desoxycholate was to inhibit firmation in buffered desoxycholate
growth of sporeformers and other lactose broth.
Gram-positive bacteria without affect-
ing the growth of coliform organisms.
Media and Methods
Gas production in the broth was The formula of the glucose broth is
abundant in most cases, sometimes essentially the same as that of the me-
being even greater than in the standard dium previously reported by Hajna,
lactose broth. Occasionally, however, except that glucose is substituted for
only small amounts of gas were noted lactose. To 1,000 ml distilled water,
in both media when incubated for 48 add:
hr, owing to the presence of slow lac- Quantity
g
tose fermenters.
Bacto tryptose 20
: McCrady (2) has stated that the NaCl 5
majority of laboratory workers are op- K,HPO. 4
posed to excluding from the coliform KH,PO. 1.5
group the slow lactose fermenters, Glucose 5
Na desoxycholate 0.1
which usually ferment glucose with a
large amount of gas. In view of this, Adjustment of pH or filtration is not
the question arose as to whether a buf- necessary. The medium will be clear,
631