Design of Wye Branches for Steel Pipe

H. 5. Swanson, H. J. Chap ton,W. J. Wilkinson,

C. L. King, and E. D. Nelson
A contribution to the Journalby H. S. Swanson,StructuralEng.
Assoc; H. J. Chapton,ProjectDesignEngr.; W. J. Wilkinson;C. L.
King; and E. D. Nelson; all fromLos AngelesDept. of Water &
Power,Los Angeles,Calif.

thepast30 yearsa num- broaderprogramwas theninstituted,

DURING
berof countrieshavebeendoing and large-diameter pipeswerebrought
independent workon wye branchde- intothe investigation. The resultsof
sign. Workin Europehas beencar- thisprogram werepublished(3, 4) in
ried on primarily by the Swiss and 1946_47.
the English. In 1928 the Swiss firm Priorto WorldWar I theSouthern
of Sulzer Brothersbegan using a CaliforniaEdison Co. had a large-
circular-collar eneronwyebranches diameter
stiff wyebranchfabricated at the
to be installedin the PuenteNuevo KruppWorksin Germany.A forge-
powerstationin SouthAmerica(1). weldedpenstockfittingwitha cast-
A patentforthistypeofreinforcement steelcrotchsection,ithas beenin serv-
was grantedSulzer Brothersin 1931 ice in Big Creek Plant No. 1 since
by the Swiss government.Several 1913. In thefollowing 10 years,sev-
yearslaterthecountries ofcentralEu- eralotherwyebrancheswerebuiltfor
ropebeganusingtwo-and three-opiate thiscompanyby M. W. KelloggCo.
reinforcingdesignsforwye branches, The staticpressurerequiredof these
withthe Swiss takingthelead in ex- sectionsvariedfrom1,700to 2,400ft
perimentalwork. The resultsofsome ofhead.
of this workwere published(2) in Early in 1936 the Metropolitan
1941. Water Dist. of SouthernCalifornia
The English began experimentingbegan work on a series of large-
withwyebranchreinforcement in 1934, diameter wyebranches.The firstwas
withoutthe knowledgethat Sulzer a 10-ftdiameterpipe branching into
Brothers werealreadyacquainted with three6-ftdiameterpipes. The re-
theproblem.This earlyworkof the inforcement was unusualin thatit was
Britishwas done on small-sizepipe; mostly insidethesectionand so placed
theyestablished a systemof basic de- thatit did not encroachon the area
sign forone- and two-platereinforce-of flowto any greatdegree. It was
mentfor wye branches. J. S. Blair installed in theHayfieldPlantin 1936,
was mainlyresponsible forcarryingon andsincethattimea number ofsimilar
thiswork. In 1938 the investigatorssectionshave beenfabricated and put
weregranteda Britishpatent. Dur- intoservice.
ing World.War II, workon the ex- In 1930 the Los AngelesDept. of
periments halted,and it was notuntil Waterand Powerbeganinvestigation
1946 that it began again. A much and construction ofstifferier-reinforced
581
582 H. S. SWANSONET AL. Jour.
AW WA

wye branches. In the followingyear a as a consequence,therewas no oppor-

systemof basic designand stressanaly- tunityto correlate the data. When
sis was devised,and in May 1935 the the need forbuildinga numberof wye
firstmajor hydraulictest was made. branches at the same time arose, it
The wye branch,a 36 X 36-in.,45-deg was decided to consolidatethe design
section,was checked for stress distri- workand testinganalysisinto a report
bution by strain gages placed on the which covered a range of sizes and
reinforcing plates,and the resultswere pressures.
analyzed in a report. This unit was
later installedin the city'sdistribution GeneralPurpose
system. Since 1935 this method of The generalpurposeofthispaperis
using plate stiffenersto reinforcethe to presenta practicaland economical

1913Installation
The above sectionswere fabricated for the SouthernCaliforniaEdison Co. by the
KruppWorksof Germany priorto WorldWar I. Theywereinstalledin Big Creek
PlantNo. 1 in 1913and havebeenin continuous serviceeversince. The breechjoint
is caststeel,and thepenstockpipe is forgewelded.

crotchseams has been used in the dis- methodto be used in designinglarge-

tributionsystemwheneverpossible. A
postwarconstructionprogramhas re-
sulted in the installationof more than
100 stiffener-reinforced wye branches
since 1947.
In the past each wye branchat Los
Angeleswas designedindividuallywith
respect to sizes and conditions,and,
diameterwye branches. In attempting
this,the authorshave soughtto estab-
lish standardsand good practicein de-
terminingthe propertypedf reinforce-
ment and other basic facts connected
with the design. Tables, charts,and
nomographs have been included to
simplifythe workas muchas possible.
June1955 wye BRANOU DESIGN 583

The generalassumptions in thisre- staticload into"the stiffener beams.

portare based on the use of a tested Temperature is ignoredin tfcedesign
modelthatis similarto theactualwye becausethe effectis negligiblewhen
branchto be designed(5). The actual thesectionis underground, ffthesec-
pipeand the model must be fabricatedtion is to be laid aboveground,how-
from thesamematerial so thatthemod- ever,temperature mustbe includedin
ulus of elasticity,
modulusof rigidity,fhedesignor allowedforthrough the
and Poisson'sratiomaybe givencon- mseof expansionjoints.

1936 Installation
phis juncture of three6-ftdiameterpipes
1928 Installation Enteringa 10-ft diameterline was fabri-
cated in 1936 for the Hay field Pumping
This band-reinforced wye was also fabri- Plant on an aqueduct of the Metropolitan
cated for the Southern California Edison Water Dist. of Southern California. The
Co. by Krupp. It was placed in opera- vertical wye stiff ener plates continue
tion in Big Creek Plant No. 2A in 1928. through the pipe as tension diaphragms
The large penstockis 3 in. thickand the and Aretied togetheron the uphill end by
two 48-in. branches are 2'-in. thick. the circumferential ener ring.
stiff

siderationin the predictionequations. Design and Test Methods

The freeends of the unequal-length The
stiffeners
areassumedto be affected design methoddevelopedfor
by thetwo-plate wyesectioninvolvesesti-
a balancingforcewhichcauses equal matingtheamountof
plateneededfor
deflections.This assumption is made thereinforcement and applying there-
so thattheprinciplesof staticscan be sultantcircumferentialstressfromthe
appliedtotheproblem.The maincon- pipeshellson it. In a tee section,the
cernis findinga meansoftransmitting curved-beamequationscould thenbe
circumferentialstressesdue to hydro- used to calculatethe criticalstresses.
584 H. S. SWANSONET AL. Jour.AWWA*

The 24 X 24-in., 45-deg wye branch at

left shows the two stiffenerplates in
place. The inside curve of the plates
match the inner surface of the pipe shell.

At right,the branchpipe is being lowered

into position.

At left,the assembledunit is being fitted

for welding. If a thirdplate were to be
added, it could be welded on when this
portionis completed.

Fabrication Methods
June1955 WYE BRANCH DESIGN 585

When,the sectionis otherthan90 deg, (d). Upon examinationof the prob-

however,one plate will have a larger
load thantheother,so thatthecalcula-
tionsmustincludebalancingthe deflec-
tions of both plates at the junctionin
order to arrive at the amountof load
whichthe longerplate will transferto
the shorterplate. In dealingwiththe
larger-diameterpipe, many designs
whichwould hold the deflections down
lem, it was concluded that the plate
could vary anywhere from a ring
girderto a two-platedesign,depend-
ing upon the ratio of pipe diameters.
Equations were set up under this as-
sumptionand some designstested. As
a resultofthiswork,a graphwas made
whichis used in conjunctionwiththe
nomograph.

1935 Installation
In 1935 thissectionwas testedhydro forstressdistribution
statically by theLos An-
gelesDept. of Waterand Power. Followingthetest,thestresseswereanalyzedin a
reportand thesectionwas putintoservicein thedistributionsystem.

were tried. The most economicalof As an aid in devising a practical

these proved to be the three-platede-
sign in which a half-ringplate was
added to a two-platedesign to hold
down the freeends.
The design of the single-platewye
began with the suppositionthat the
plate could be treatedas a ringgirder
methodof design,severalexamplesare
given, with accompanyingtables and
sketches. The nomographsection of
the report contains simplifications
whichgive resultswithinthe range of
accuracy necessary,but, to use the
nomograph(Fig. 15) properly,a few
586
( 11
J/ro
ftsinaL,- »^f
^r^-^J^T
Load Due to ^-^
InternalPressure y* W**A
l__f__
Reinforcing '^ x^^
Plates
' *-^
_ . . .
Section A -A
Fig. 1. Intersectionof Equal-
Diameter Plates
As is unitlengthalong centroidalaxis;
x is horizontaldistanceto centerof As.
factsmust be noted. First, the depth
of plate read fromthe nomographis
for 1-in.thickplate; when this is used
withpipe of a verylarge diameter,the
curved-beamK factor (7) becomes
very great, increasing the critical
stress. If the value fromthe nomo-
graph is convertedto a thickerplate
with a smallerdepth,however,the K
factoris lessened,giving a betterde-
sign. It is recommended thatthedepth
of plate, dw,should not exceed 30t (t
equals thicknessof crotchplate).
Second, under some very high head
conditionsin smaller-diameter pipe,the
depth of plate has a tendencyto ap-
proachthe diameterof the pipe. This
also will increase the K factor. It is
recommended, in such a case, thatthe
depth of plate be made less than the
diameterof the pipe.
If these recommendationsare fol-
lowed, the nomographicdesign will
H. S. SWANSON ET AL. Jour.AWWA
TABLE 1
Symbol
Explanation of Symbols
Explanation Units
A deflectionangleof pipeline deg
D diameterof pipe in.
Rb radiusoflargepipe in.
Rs radiusofsmallpipe in.
R' centroidalradiusofreinforc-in.
ingplate
a inside radius of reinforcingin.
plate
d depthofreinforcing plate in.
«-/>sinAL2
d' d at 1-in.thickplate (as in in.
_^_M^ s^ nomograph)
dw d at acute crotch in.
• ] db d at obtusecrotch in.
_K__ '
dt d at junctionof reinforcingin.
ft; y I plates
/ L' lengthof axis of largerein- in.
forcing plate
k Plate Junction L2 lengthof axis of smallrein- in.
Line or Origin forcing plate
W' totalwaterload on largere- lb
inforcing plate
Wi totalwaterload on smallre- l|>
inforcing plate
y' theoretical reactionon large lb
reinforcing plate
yi theoretical reactionon small lb
reinforcing plate
3>3 theoreticalreactionon ring lb
plate
T=P resultantload on horizontal lb
axis of plates (equal to
W±y)
p internalpressurein pipe psi
/ thickness ofreinforcing plate in.
A cross-sectional area sq in.
M momentin section due to in.-lb
load
m momentat any point due in.-lb
to load of unitywherede-
flectionis to be found
E modulusofelasticity forsteel psi
u Poisson'sratio
usually include a safetyfactorwhich
will keep stressesbelow the yieldpoint
of structuralsteel. Because the maxi-
mumallowable stressof the steel used
in this reportis 30,000 psi, however,
it shouldbe understoodthatthe design
pressure used in the nomographwas
keptto 1.5 timesthe workingpressure
June1955 WYE BRANCH DESIGN 587

-■ , i [t
w ^
« -LLXI
.
T
* 1U k I

_ i 11 i mm

f = R- r
(b)

-- I <; 1 __. I

-J ,-Ll . .1 11
bi f1 •« 6i f t»
I I I

A A
r= r=
6l loge~ + b2loge-d + &3loge- b' loge~ + &2loge-
a
a g g
e = R- r e = R- r
(c) (d)
Fig. 2. Typical Cross Sections of ReinforcingPlates
in orderto approximate an allowable theuse of a factorbased on thedata.
stressof 20,000psi. Anotherfactor The variousdesignsweretestedby
to be notedis that,althoughthe de- placingthe sectionunderhydrostatic
signsused in thereportare basedon load and usingelectricalstraingages
a rectangular cross sectionfor the to indicatethe amountof stressto
crotchplates,thetestsshowedthatthe whichthe sectionwas subject (8).
weldingin the crotch,plus a portion These data were comparedwiththe
of the pipe shell,acted as a flange design stresses,and the methodof
whichdecreasedtheinsidefiberstress. designwas corrected to conformto the
This was taken into account through resultsof thetests.
588 H. S. SWANSON ET AL. Jour.AW WA

V Hi ^

^ 5 -^

I
03

Its
6
Ci ^ **^
^ Ck) tu

I III
?5 ^
^
« o ^

x
Si
bJO
111
i s r*
*-§ s

■Cx, ^

§: ^> e

si
is
 lllllllll
June1955
2 S^-
^ - -
1.0
0.9
^,
0.7 _^
0.6
/
0.5I I 1 1 1
12 3 4
Fig. 3. K Factor for Curved Beams (RectangularCross Section)
R is radius ofcentroid;c is one-halfdepthofplate; fmax= K - - .
The use of single-directionSR-4
straingages in the testingof the wye
sectionswas desirablebecause thistype
of gage required less mountingroom
and fewer connectingwires than the
rosette type and gave stress results
while the sectionwas still under test.
The Baldwin-Lima-HamiltonCorp.,
Philadelphia,developed a method of
using these gages on sectionswith bi-
directionalstress,but the directionof
the principal stress must be known.
The rosettegages are essentialfor ob-
tainingstressesin locationswhere the
directionof the principalstress is not
known. The equations used in com-
puting stresses from rosette strain
readingsare so involvedthatthecalcu-
lations had to be made in the office
afterthe testshad been run (8, 9).
During the tests a waterproofce-
ment was found which allowed the
WYE BRANCH DESIGN 589
._ Inside Fiber
~ " ====
|
"
outsideFiber
:
1 ' 1 ' 1 ' ' ' '
5 6 7 8
_/?
c
Me
SR-4 electricalstrain gages to be at-
tached to the inside peripheryof the
crotch stiffenerswithoutshortingout
under the high hydrostaticpressures
used. Single SR-4 gages applied in
this manner gave test results in sec-
tions that had not previously been
available. - (See photos on pages 590
and 594.)
On one large wye branch,the Mag-
naflux Corp., Chicago, was called in
to use the "Stresscoat"methodforde-
terminingthe stresses,in conjunction
with the electricalstrain gages which
were normallyused. It was consid-
ered advisable to try this methodbe-
cause of the completecoverage that it
gives. The resultsof the two methods
compared very favorablyand gave a
good pictureof the stressesin the sec-
tion. The "stresscoat"test confirmed
590 H. S. SWANSON ET AL. Jour.AW W A

I•
•is"
It
II
*i
ji
ii
a*
•
§

!!

|
^^
^u o
^>

I ^> e

! 11
I«

II

lg

"
S
5s
June1955 WYE BRANCH DESIGN 591

the locationsof the criticalstresses was apparentthattherewouldbe diffi-

thathadbeendetermined in thedesign
analysis.
Conclusionsand Recommendations
Throughout thesetests,it was found
thatthe designof the actual section
couldbe predictedfromthe modelto
a degreewithintherangeof accuracy
cultyin proportioningtheplatesso as
to haveeachone carryits shareofthe
load. By the use of the methodof
balancing a balanceddesign
deflections,
for a three-platesectionwas found;
becausethereduction in thesize ofthe
platedepths,comparedwiththatof a
two-platedesign,was verysmall,how-

r'i 11iii111^^
5
'
inun Axis
Neutral -/' '£ ^ ^- '

4 V' f%;
A^-H j

* ' ■*+-
I I

J
0 01
I 1 I I 1I111
0 02 0.03 0 04 0 05 0.06 0 08 0 1
II
0.2 0.3
rrrm-
0.4 0.5 0.6 0.8 1.0

of Eccentricity,
Fig. 4. Determination e
Based on Timoshenko
(10) equation:
h* [, 4 / h VI Mhi P Mh2 P

in whiche is eccentricity(distancefromcentroidalaxis to neutralaxis in curved

beam); h is depthof curved axis of curvedbeam.
beam;and R is radiusof controidal

required. Because of the difficultiesever,it was concludedthat,if a third

encountered in analyzingthe typeof platewas to be usedto holddownde-
curvedbeam used in these designs, flectionson larger-diameter pipe, it
however, it becamenecessary to apply shouldbe added to the two-platede-
stipulationsto thedesignprocedure in sign and the dimensionsshouldcon-
orderto conform to thetestresults. formto thetopdepth. Also,it should
It was foundrelatively easy to ob- be weldedto the otherreinforcement
taingoodbalancein thetwo-platede- platesonlyat top and bottom,being
sign,but,in the three-plate design,it leftfreefromthe pipe shell,so that
592 H. S. SWANSON ET AL. Jour.AW WA

TABLE 2
General Formulas (See Fig. 1)
R R ^sinALl2- P Sin ALg2
T -
Li-- •, L2=-
T -, ^=
W , W2 =
W
sin - cos -

W at any point x fromorigin (triangular load) :

sinAx2
Wx= P
Moment at any point x fromorigin (triangular load) :

m= ; M = m ± xy (use minus for larger plate, L' ; use plus for smaller plate L2)

A
a = Rpipesin - (for acute angle of two- and three-plate design)

A
a = ^pipe cos - (for obtuse angle of two- and three-platedesign)

Two-Plate Design
P = total shear at any point x fromorigin
_ p sin Ax2 _ p sin-Ax2 ,
^i=^-2 y' P2=^ +y

y = load necessary to cause equal deflectionof plates

Deflection at x = 0 for Plate 1 = 2 ^^ - 2 ^^

El El

Deflection at x = 0 for Plate 2 = 2 ^^ + 2 ^^

As deflectionsmust be equal:

v ntiXAs v yx2As ^ w2xA5 ^ yx2As

Three-PlateDesign

y,=y1+y>; Plate i, a = S -
^ S^;

Plate2,J = S ^ - S Plate3, a = Jf^ (halfring)

iii *g£;
iLi El

Maximum Stresses

=
Mm^ w[-3+ (0.2)^] ± y[l + (0.2)^]
, _ Afma,fti , W"d=y
~ , -_ M^x h<,- W±y*
/max - ~A I I /min ~.
Aea A Aec A~A

- ~ ~
/max - A -j ; /min -
-

* Timoshenko(10) ; c equals one-halfdepth of plate; forhi and hi, see Fig. 4 (page
591).
t Seely (7).
June1955 WYE BRANCH DESIGN 593

none of the shellstresswould be trans- actual pipe is maintained,the safety

ferredto the ring plate. It appears of the structureis assured.
practicalto beginaddingthe ringplate The principal symbols used in the
to a two-plate design on high-head, design equations and sketches are
equal-diameterwye brancheswithpipe listed in Table 1 ; other symbolswill
diametersgreaterthan 60 in. be explainedwhen used. The general
During early testingof the models, formulasemployedare shownin Table
it was foundthat,in the criticalsec- 2. Following the section on analysis
tions,themeasuredstresseswerelower and design proceduresis a description
than the design stresses withoutex- of the nomographmethod,includinga
ception,givingan additionalsafetyfac- discussion of the developmentof the
tor. It was concludedthat the flange curves (pages 617-623). Test results,
effectof the weld metal and a portion in the formof drawingsand tables,are
of the pipe shell in the crotchhad ac- shown on pages 624-627.
ceptedsome stresseswhichit had been Appendix A (page 628) containsa
comparison betweena Sulzer Brothers
assumedwould be in the reinforcement
design and one based on the authors'
plates. By correctingfor this in fur- method. Appendix B
thertesting,the design methodswere nomograph illustratesthe deflec-
(pages 628-629)
improved to give more economical tionsthatoccurredon two 45-deg wye
proportions. branch pipes. The deflectionsare
A review of the test data evaluated
plottedon a nomograph,to providean
for a practicaland economicaldesign indicationof what
mightbe expected
shows that the dimensionsfor the re- in other45-deg sections. A discussion
inforcement plates for a wye branch of the results of overstress in wye
may be chosen from the nomograph. branchpipes is found in Appendix C
If similaritybetweenthemodeland the (pages 628 and 630).

Analysis and Design Procedures

In the designsdescribedin this por- used in Seely's methodcan be found
tion of the paper, a trial depth and fromthe curves in Fig. 3 and 4.
thicknessofstiffenerplatewereselected A series of typicaldesigns are pre-
and the stressesin the criticalsections sented: single plate, A = 90 deg ; two
werecomputed. Then a new depthand plate, A = 90 deg ; two plate, A = 53
thicknesswere chosen,as indicatedby deg; two plate, A = 45 deg ; three
the resultingstresses. These stresses plate, A = 90 deg ; threeplate, A = 45
were evaluated by using the loading deg; and three plate (alternate de-
assumptionsshown in Fig. 1, along sign), A = 45 deg.
with a carryoverload, y, found by a In an early effortto use the model
methodof balanced deflections. The theoryas an aid to futuredesign, a
solution equations were taken from numberof tables were computed,such
Timoshenko(10) and Seely (7). as theone in Fig. 6 (page 597). This
The various cross sectionsof curved table was based on the two-plate24-in.
beamswhichcould applyto wyebranch wye branchshown in Fig. 7 and was
design are shown in Fig. 2, although retained as an example only. The
only the rectangularsection will be nomograph(Fig. 15) has greaterver-
consideredin this report. The eccen- satilityand lends itselfbetterto corre-
tricity,e, and the K factorwhich is lationwithtest data.
594 H. S. SWANSON ET AL. Jour.AW WA

Two-Plate Wye Under Test

This photographof a 30 X 30-in., 90-deg wye under test shows the mercuryswitches
and deflectiongages.

Three-PlateWye Branch
June1955 WYE BRANCH DESIGN 595

One-Plate Design
For one-platedesign:

0.3 = 0.6 rv
^ i
1 Flow 1I
Wi= stP= pRB ' **
y
-

' sin2A / 4
U2 , dt
rz = +
R: 2 _________
'TTTTTTTTTTIT
At sides: J[
Momentdue to W1 = s^ ^' H
j^^ -j- >'dt
. __ - /, /^^ *A ^X '' -*
-^2i?,2
Momentdue ^to Wi = ^_ f 4-^l 11 J^c**i ' -*

_ Vsin2A / Le-I ' / L-jJ

^2,^ II I I I I I I I I I
Lx I fYYYYYYYYtff

r2= L7+T
Fig. 5. General Equations for One-Plate Wye Design

Example Maximum momentat side, M2 = {W' - W2) ~r

4
Main piperadius,Rb = 30.5 in. = 114,400 in.-lb
Branchpiperadius,Rs = 10 in.
As a trial, use a 6 X 1-in. collar to stiffenthe
Deflectionangle,A = 90 deg
joint. Then:
Designpressure, p - 300 psi
Uniform load at top, Wx = pRB = 300(30.5) /max - -^*-T j 1 a
= 9,150psi
^5 300(30.5) 6M WRS
Uniform1 a at -j XT, ^
load side, W2 = £-r- = ^
' bd2 ~^ bd
sin A
= 4,575 psi
At theside,on the innercurve:
Maximum momentat top, Mi = (PF3- PFi) -j- ^ = 4.2
c
- -114,400 in.-lb
♦ Pans {11). t K is Seely'sconstant, on thecurvature
dependent
ofthecollar.
596 H. S. SWANSON ET AL. lour. AW WA

K = 1.185 no7x -H4,400(6) 4,575(10)

CAW = ((-0.87)
m
(*'
(/2)max ,0,114,400(6) + 9,150(10)
= i1.185 1(5.5)2 +-[(5^-
1(6)2 1(6) = 19,700+ 8,300 = 28,000psi
= 22,600+ 15,200= 37,800psi (tension)
(tension)
At thetop,on theinnercurve: In thefirsttrial,theouterstressat theside
was ignoredbecause it could not exceed the
* = innerstress. At the top,however,thedirect
4.3
c stressis additiveto the outerstressdue to
K = 1.18 moment,increasingthe total stressto the
point whereit must be investigatedalong
(*' < 1O -114,400(6) + 4,575(10)
(/i)max- 1.18 withtheinnerstress.
^ 1(6) The final design is a 1-in. thick collar,
= -22,500 + 7,600 = -14,900 psi weldedat the junctionof the pipes,with a
(compression) depthof 7 in. in the crotchesand a depthof
5Jin. at the top and bottom.
At thetop,on theoutercurve: Note : Bier (6) statesthat the assumption
ofa uniform load is on theside ofsafety,but
*=4.3 it shouldbe notedthat,at smallerdeflection
c
angles,the ringbecomesegg shapedand the
K = -0.86 ringgirderequationformomentbecomesless
«' - n o.-n4,400(6) + 4,575(10) applicable.Also,whenthe branchpipe diam-
-0.86
(/i)max eterapproachesthediameterofthemainpipe,
^ 1(6)
theassumption thatthecollarliesinone plane
= 16,400+ 7,600 = 24,000psi causes inaccurateresults.The conclusionis
(tension) thatthismethodofone-platedesigncan apply
For Trial 2, use 5J in. as the top depthand onlyto limitedcases of wye brancheswhere
7 in.as thecrotchdepth. By inspection, thedeflection angleis closeto 90 deg and the
tensionwill governat both the top and diameter of the branchpipe is substantially
the crotch. At the side, on the inner smallerthanthatof themainpipe.
curve: The ASME BoilerCode (12) coversa good
portionof the one-platedesignin the lower
* =
3.8
ratiosbecause the reinforcement pad is close
c to the plane of stress. It appears that this
methodis good up to a ratioof diametersof
K = 1.21
pipe of 0.3:1. By the same token,the ring
tf' 114>400(6) 9,150(10)
(Mnax-l^li ?i + girderassumptionappears to be satisfactory
1(7)2 1(?) up to a ratioof 0.7:1, but above this ratio,
= 17,000+ 13,000= 30,000psi the curvatureof the plate is such that it has
(tension) been foundadvisable to use an equal pipe
diameter designand a truncated coneto reduce
At thetop,on theoutercurve: thediameterof the outletpipe to the desired
size. In the one-platedesign curve in the
^ = 4.63 nomographsection,an attemptwas made to
c
join the methodsof designtogetherso that
K = -0.87 theymergeintothe two-platedesign.

Two-PlateDesign
Example1 Note : As thetwo platesare identical,it is
Rb = 12 in. notnecessaryto set up a tableforthecompu-
R8 = 12 in. tationof forcey,whichautomatically reduces
A = 90 deg to zero. The selecteddepthsforcrotchand
p = 350 psi crown were obtained from the prediction
Pipe shellthickness,f in. tablesforpipe tees whichrangedin diameter
Depth at crotch,12 in. from12 in. to 120in.,at a pressureof350 psi.
Depthat crown,5 in.
June1955 WYE BRANCH DESIGN 595

yS ^^rT°P Section ( J

ly WyeSection
vj ^f ir^ ^-^T - ^

bd2, . , , x For top section,K equals:

S„ = - (sectionmodulus) d= '
/6S
(D*)
6 ^A»8 0.0521at 6.25, 1 in.
S = N*Sm* d = #£* 0'0425at 6'25' !* in'
0.0368at 6.25,2 in.
Data: 0.0301at 6.25,3 in.
^ = D_
m
5m= 36 in.,6.25 For wyesection,K equals:
</2= - = = Dm = 24 in 0.125at 36, 1 in.
b b hD™ 0.102at 36, li in.
a_i
" ii ''o ?•^
<V refers
♦Subscript tomodel. at 36, 2 in.
' = 350
p ^ ^ psi ^.0884
0.0722at 36, 3 in.

PlateDepth,d- in.
Diam.
D D* WyeSide Top
in.
b = 1 in. 6 = 1Jin. 6 = 2 in. b = 3 in. b = 1 in. 6 = 1Jin. 6 = 2 in. 6 = 3 in.

12 41.5 5.19 4.23 3.67 3.00 2.16 1.76 1.53 1.25

24 117.6 14.7 12.0 10.4 8.49 6.12 5.00 4.33 3.54
30 164.4 20.5 16.8 14.5 11.9 8.56 6.99 6.05 4.94
36 216.0 27.0 22.0 19.1 15.6 11.25 9.19 7.96 6.50
42 272.4 34.1 27.8 24.1 19.7 14.2 11.6 10.0 8.19
48 332.6 33.9 29.4 24.0 14.1 12.2 10.0
54 396.9 40.5 35.1 28.7 16.9 14.6 11.9
60 465.0 47.4 41.1 33.6 19.8 17.1 14.0
66 536.0 47.4 38.7 19.7 16.1
72 611.0 54.0 44.1 22.5 18.4
84 770.0 68.1 55.6 28.4 23.2
96 941.0 67.9 28.3
108 1,123.0 81.1 33.8
120 1,313.0 94.8 39.5

Fig. 6. Two-Plate Design (Example 1) - Plate Depth Calculations

forSelectionofPlate Thickness
Computation

Wi-W- (f )W*L* = f (16.97)*

= 50,400
1b
Maximummoment
:

M"Wx i) =50'400
(§+(0-2) t ] ^345?500
[~F+(0*2) in-"lb
598 H. S. SWANSON ET AL. Jour.AWWA

^V = lj"'
*«" PlateThickness
^| j ;$O^
^- to*' ^s. i65"_4"#i3 '

#12 £-2
) |
o s ip |^ ioi-'-^rUh1"
'*
(( 4* 12" A
Scale inInches 16if"

Stress*- Ib Stress*- Ib
Item Item .
Computed Measured Computed Measured

#12 9,760 9,260 #17 -8,180 -5,710

#13 7,980 9,250 #18 -5,525 -3,090
#14 3,760 6,490 #19 -3,550
#15 1,760 E-2 -4,020 -5,650
#16 -4,880 -4,000
* At 350-psipressure.

Fig. 7. Two-Plate Design (Example 1) - Plan of Stiff

ener Plates

For any pressureotherthanthe designvalue (350 psi in thisinstance),afterthe

criticaldimensionshave been selectedfromthe nomograph(Fig. 15), the outside
curveof thisplateshouldconform withtheshape shown.
geometrically

a = R sin^ = 12 sin^ = 12(0.707) = 8.48 in.

R = 1A48 =
1207. e = OO71(12) = 0.852 in.; hi = 5.148 in.; h2= 6.852 in.
h 12

By Timoshenko
equation:
Mhi W 345,500(5.148) , 50,400= . . _n . . . w -.
f- = ^ +, A - + ,
fOrUn' pkte
24'7°°PS1(t6nSl°n)
12(0.852)(8.48) "IT
Use lj-in. plate:

/m« - 1(24,700) = 16,500psi

. -Uh* , W 345,500(6.852) 50,400 * " .
/-to = ~AiT + A = " 12 + 4>7S0psl (comPression)
(1.5)(0.852)(20.48) (12) (1.5)
June1955 wye branch DESIGN 599

By Seelyequation:

*=Mf8 = 2.413;
K- 0.77
c 6
KMc W
/„,„ = - + = 1.40(345,500) +, 50,400 = «*„„„.,
16-200 Psl (te««on)
j 0.167(L5)(i2), i2(T5)

/-. = *f + J = ^-j^y = -4,600psi Oppression)

+ 2,800
Example2
i?B = 18 in.; Ra = 18 in.; A = 45 deg; p = 385 psi
Pipe shellthickness,f in.; depthat crown,12 in.
Table 3 showsthe correctedvaluesas constructed ; it also providesa simpleexpedientfor
converting thedesignvalue ofy to a newy forany otherpressuredesired.
CalculatedStressesfor Larger Plate
2' in.
Depthat crotch,30 in.; platethickness,
Note : Owingto theflangeeffectofthepipe shell,a factorof 0.2 is appliedto - . The cen-
troidalaxis is taken as beingparallelto the curvewherethe stressesare beingcomputed.

Maximummoment:

^=^[f+(0.2)f]-,[f + (O.2)f]
= 277,000 - 45,800 = in.-lb
[l6.13+ (0.2)y] [48.38+ (0.2)y ] 2,946,000
a = R sin = 18 sin22J° = 18(0.3827) = 6.9 in.
|

- =- = 0.73;e = 0.128(30) = 3.84 in.; hi = 11.15 in.

h 30
Maximumfiberstress(tension):
:
Timoshenko

,
/™' = Mh + T = 2,946,000(11.15) + 231,000 = 19>95°+ 3'6°° = 23'6°° PS1
.
-JTa A 30(2.125)(3.84)(6.9) 30(2ji25)
Seely:
= 1.46^ = 2.01
^=^
c 15

CalculatedStressesfor Smaller Plate

2j in.
Depth at crotch,20 in.; platethickness,
Maximummoment:

^ =w,2[| + (o.2)f]
+ (o.2)f]+,[f
= 52,500 + (0.2) ] + 45,800 = in.-lb
[7.02 y [2I.O6+ (0.2)y ] 1,530,000
600 H. S. SWANSON ET AL. Jour. AW W A

TABLE 3
Two-Plate Design Calculations (Example 2)

" " 47'04' U ~ ~ " 19*48

Ll = . A 0.3827 A 0^239
sin - cos -

Li (scaled fromPlate 1) = 48.38; L2 (scaled fromPlate 2) = 21.06

d' - 30 in. ; &i - 20 in.

^ = ^sinAO^ _ 385(0.707)(48.38)* _ ^^ ft

^ _ ,«,A(LQ* _ 385(0.707)(21.06P = ^^ Jb

/>x3sinA = 385(0.707)x3
mi = -- -- - = 45.36jc3
o 6

x M = mi - xy I x(As) x(As) x(As) ~ B^tl} of

in. in.-lb in* in. "T~ ~~T~ (wi xy) P-ate

5 = 9.5 ± in.; / = 2| in.

4.70 5,000- 4.7y 313 44.66 0.143 1,000- 0.67? 12.1

14.1 127,000-14.1? 371 134.0 0.361 46,000- 5.09y 12.8
23.5 588,000 -23.5;y 486 223.0 0.460 270,000 -10.80? 14.0
32.8 1,601,000-32.83; 767 311.5 0.406 650,000 -13.32? 16.3
42.0 3,360,000-42.0? 1,416 399.0 0.282 947,000-11.85? 20.0
51.2 6,036,000-51.2? 3,111 487.0 0.157 948,000- 8.04? 26.0
51.38 6,093,000-51.4? 4,875 488.0 0.100 609,000- 5.14? 30.2
51.38 6,093,000-51.4? 4,875 488.0 0.100 609,000- 5.14? 30.2

4,080,000-60.05?

5 = 7.5 in.; t = 2.5 in.

3.7 2,000+ 3.7? 378 26.25 0.0694 0+ 0.26? 12.2

10.9 59,000+ 10.9? 559 81.75 0.1465 9,000+ 1.60? 13.9
18.0 264,000+ 18.0? 886 135.0 0.153 40,000+ 2.75? 16.2
23.06 544,000+23.1? 1,235 173.0 0.1401 76,000+ 3.23? 18.1
23.06 544,000+23.1? 1,519 173.0 0.1139 62,000+ 2.63? 19.4
23.06 544,000+23.1? 1,664 173.0 0.1040 57,000+ 2.40? 20.0

244,000 + 12.87?

4,080,000 - 244,000 = 60.05? + 12.87?; ? = 52,600 lb *

* To obtainnewvaluefory foranyotherpressure, y bynewpressure

multiply anddivideby385.
June1955 WYE BRANCH DESIGN 601

^/d o ... rentroidalAxis ... z n|2 %'«H S

= 2i"
PlateThickness .^ r S,« - 2 2 I
M^2

A -2 i1^! T552 2 • ^|*~ PlateThickness

= 2£M 28i"-?S. '

~l£±I 30" >' 1 J-i-i ^ ^^^ ^'->l

U 78|" -^^r- ^ J+c-i - ' /^-^
0 5 10 IM^^-mJCJ X/34l"4'
I 1 1 N' 7« w 1 y's o '
Scale inInches '
' -*-F~/np '% '
( n I 'y# ' A#11

' /^ I |X ' ^j1 I /-2

4
H 21^» 4» 20" >

Stress* - Ib Stress* - Ib
Item Item
Computed Measured Computed Measured

A-l 9,220 10,620 H-2 -8,320 -6,710

A-2 -6,520 -5,420 /-I 13,300 11,070
B-' 7,540 11,140 J-2 -7,470 -7,300
5-2 -5,270 -4,260 #1 9,560 12,250
C-l -4,600 980 #2 2,460 3,930
C-2 -7,890 -6,450 #3 -7,690 -270
DA -2,800 -930 #4 -6,020 780
D-2 7,560 5,900 #5 -6,120 -6,320
E-l -2,780 -2,860 #6 -5,120 -3,760
E-2 3,210 820 #7 370 210
.F-l 2,010 2,850 #8 10,270 8,600
F-2 -2,290 -1,480 #9 -6,320 -5,750
G-' 6,980 10,970 #10 -8,290 -6,780
G-2 -6,610 -5,690 #11 -9,140 -7,050
H-' 8,680 11,920
* At335-psipressure.

Fig. 8. Two-Plate Design (Example 2) - Plan of StiffenerPlates

See italicnotebelowFig. 7.
602 H. S. SWANSON ET AL. Jour.AW W A

a = = 18 cos =-=18(0.9239) = 16.63 in.

#cos^ ^-

= = 1.33; e = 0.064(20) = 1.28 in.;fci = 8.72 in.

^ ^

Maximum fiberstress (tension) :

Timoshenko:
.
t
;max " Mhl ^i T- ~ ^530,000(8.72)
^
98,300 ,,-nn
' ^
Aea A 20(2.50) (1.28) (16.63) 20(2.50)

Seely :
* = 1.35
= = 2.66;*
c ^ 10

_ KMc +, T _ 1.35(1,530,000) + 98,300 ~ 14>400 .

/max- j A- 0-167(2.s)(20)*
PS1
20(2.50)

72" *
^ :

- plate
1 66|^Kf
I §£8~K-r? y-2^**
ti "S >'^eThickness,
if '
22°"^d^37^+#6

*-27rl" .. §.
=,."i7-i'5r ' ' 1

^
PlateThickness= / /v-

bk2 i 7 /
if

^y
^K^

1"~il^I - 13^" >| I 0 5 10 15

< -19H H I L- 1 1
^ 43-" ► Scale in Inches

Fig. 9. Two-Plate Design- Plan of StiffenerPlates

Data: RB= 21 in.; Rs = 21 in.; A = 53 deg; p = 192 psi. See italicnotebelowFig. 7.
June1955 WYE BRANCH DESIGN 603

Table of Stress Values (see Fig. 9)

Stress- Ib

Item Comp. Meas. Comp. Meas. Comp. Meas. Comp. Meas.

192 psi 285 psi 385 psi 510 psi

Plate 1

A-' 11,600 12,000 17,240 16,896 23,350 21,030 29,300 27,500

A-2 -5,600 -7,400 -8,310 -12,288 -11,150 -16,450 -13,900 -21,000
B-' 8,100 8,052 12,025 17,480 16,240 24,010 22,700 26,900
B-2 -5,000 -7,800 -7,470 -9,248 -10,090 -11,590 -14,000 -13,900
C-l -300 -4,200 -429 6,816 -485 9,840 -140 13,000
C-2 -5,500 -5,350 -8,100 -10,016 -11,000 -9,290 -15,880 -15,140
D-l -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-l -2,740 5,652 -4,060 7,184 -5,430 10,430 -7,270 6,800
£-2 4,150 2,110 6,190 -1,864 8,330 -1,616 10,160 -1,240
#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
#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-' 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-' 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
JA 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 30,720
#14 5,600 9,000 8,053 12,150 11,000 17,100 -14,500 22,050
604 H. S. SWANSON ET AL. Jour.AW WA

Three-PlateDesign

^Plate3 ^l(=^2)^ir<Tl

, V n *> ,«-rtT''''

<'
PlateH
Fv PlateV
"■/f/' i
I ' A^ Ve^
^' '
' '

L ^
/
III /
/ '| J" II 'I , 4.
_ , _ '1
Li(=L2)__^|
^- ^r*--^ 7/77777 'x//////////
I |*
Half-RingPlate Plates 1 & 2
(Plate 3)

Fig. 10. Three-PlateDesign (Example 1)

Data: RB = 12 in.; Rs = 12 in.; A = 90 deg; p = 350 psi; pipe shell thickness,■§tw.;
depthat crown,5 in.

DesignAssumptions =0
1. Rotationoftangentto centroidalaxis of Jf,(^)-2yr(l)
plate at crownequals zero; that is,
half-ring 4vr
thisplateis freebut guidedat thecrown. Jlfa= -
x (Eq 1)
2. Momentsat crownof Plates 1 and 2
V2timesthemomentat crownof half- , _ r Mmds
at crown= JI - - -
8S = deflection
equal - El
ringplate,fromvectorconsiderations.
~ r (- Mz + 2yx)xds
_
MethodofSolution J El
The three unknownquantities,Mi, Ms, (- M3 + 2yrsin0) (r sin0)rd0
and y, requirethreeequationsfor solution.
El
Two oftheseequationsare obtainedfromthe
two design assumptions,and the third is - 'yr fa r sin0d0
derivedfromthecondition thatthedeflections =
7T Jo El
ofall threeplatesat thecrownare equal.
Solution
For half-ring
plate:
- 4:yr3 iryr3 yrz/ w 4'
=
ttEI +2£/ £7'2~x/
El
r(- Ms + 2yx)(-l)ds = (0.2976)
J El ^

pi (- M3 + 2;yrsin0)(-l)rd0 ComputationofMaximumStresses
Jo £7 plate:
Half-ring

' =lW-3-28
-
~_ rl {Mzr 2yr2sin0)d0
~ ~~ _°
Jo El h 4.5
(M3rS- 2yr2
cos 9)r = 0 e = 0.1151 in.
June1955 WYE BRANCH DESIGN 605

TABLE 4
Three-PlateDesign Calculations (Example 1)

Ll=-^ u =u
=ra =16-97in-:
sin -

Wi = Li2 sin A =
^ (16.97) (0.7071) = 50,400 lb
|
A 350x3
m = -px*s'm
- = - - - = 58.33x3
6 6

_ V2j^ _ Vjgg) = 2V2(14.75)y_

2 2?T 7T

Plates 1 and 2 (/ = 1.5 in.; As = 4.17 in.; dw = 9 in.):

2.10 540- 2.10^ - 13.283/ 5.0 15.62 8.76 0.561 300- 8.63y
6.22 14,037- 6.22;y- 13.28^ 5.15 17.08 25.94 1.519 21,300- 29.62?
10.27 63,175-10.27y-13.28y 5.5 20.80 42.83 2.059 130,100- 48.49?
14.09 163.157 -14.09?- 13.28? 6.1 28.38 58.76 2.070 337,700- 56.66?
17.50 311,976-17.50?-13.28? 7.0 42.88 72.97 1.702 531,000- 52.39?
20.32 366,912 -18.59?- 13.28? 8.0 64.00 77.52 1.211 444,300- 38.59?
22.06 366,912-18.59?-13.28? 8.9 88.12 77.52 0.880 322,900- 28.04?

l,787,600-262.42? = £5i

For half ring:

0.2976?r3 e e ,_ _,
53 =■ =ry^-; «i = 53 (Eq 3)

= °-^
1,787,600-262.42, = «-2^ft209) _^
i li .»5y

4r(5>162) =
? = 5,162 lb; Mi = 96,940 in.-lb; Mi = 68,550 in.-lb

Maximum moment:

2?r - ikT3= 2 (5,162) (14.75) - 96,940 = 55,340 in.-lb

Maximum tensile stress:

Mln T 2(5,162)
_ 55,340(2.25-0.115) + t^.1'" I" ^ _ i" • V
4ea .4 1.5(4.5)(0.1151)(12.5) 1.5(4.5)

Maximum compressivestress:
606 H. S. SWANSON ET AL. Jour.AWWA

= 11" '
PlateThickness
/ I*"14!" / V on JL A 012345 '
/ / (jA 0 ii iiii '' 1 '' '
/ *M J^" Scale in inches

-3H'H ^^ M U- 6|"- iJ
^-4I"4^ 121" A U 16}|M *U 9" A

Stress*- Ib Stress*- Ib
Item Item .
Computed Measured Computed Measured

AA 7,360 6,840 #4 8,380 6,150

A-2 -3,160 -3,260 I #5 -8,860 -8,850
B-2 -2,080 -730 [ #6 -20 -1,650
C-2 7,780 5,090 #7 4,000 6,510
D-' 8,610 11,780 #8 12,360 11,350
#1 6,950 7,350 #9 2,750 3,600
#2 -2,030 3,380 #10 -6,000 -3,450
#3 -4,400 -3,450 #11 -8,380 -6,900
* At 350-psipressure.

Fig. 11. Three-PlateDesign (Example 1) - Plan of StiffenerPlates

See italicnotebelowFig. 7.

Horseshoeplate (Plates 1 and 2) :

a = = 12.0(0.7071)= 8.49 in.

Rsm^

-r = 8.485+4.5
- .... .
= 1.443; e = 0.536 in.

Maximummoment:

= 50,400 + 0.72 + ~ - 5,162 + 0.72 + - ~ (96,940)

(j^jt- ) (l6.97 ^)
= 366,760- 95,920- 68,550= 202,290in.-lb
* This additionaldistancewas made necessaryin the fabricationof the wye by the requirementsforfittingthe
plates togetherproperly.
June1955 WYE BRANCH DESIGN 607

Maximum tensile stress:

Mhi ,T_ " 202,290(4.5 - 0.536) 50,400 - 5,160

+ +
~Ae~a A 1.5 (9.0) (0.536) (8.49) 9.0(1.5)
= +13,050 + 3,350 = +16,400 psi

Maximum compressive stress:

s +1-mmmm+*»--«»+*»--«•>-
f Q A i M] I
I ^^

Half-RingPlate (Plate 3) Plate 1 Plate 2

Fig. 12. Three-PlateDesign (Example 2)

Data: RB=J0.5; Rs = 30.5; A = 45 deg; ^ - 300 psi; pipe shell thickness,
-^ in.;
depthat croivn,12 in.

DesignAssumptions Solution
1. Momentat crownofPlate 1 (Mi) equals For half-ringplate:
zero.
2. Rotationof tangentsto centroidalaxes -
Fr~~
ofPlates2 and 3 at crownequals zero; thatis,
theseplatesare freebut guidedat thecrown.
r(- Mz+yzx)(-l)ds
Methodof Solution J El
- yi, yi, yi,
, The five unknownquantities IT

M2,and Mz- requirefiveconditionsforsolu- f2 - Md + y3rsin0) ( - l)rd0

tion. Theyare: J0 (

1. Rotationoftangentto centroidalaxis of X IT

Plate 2 equals zeroat crown. = 2 - f2 yzr2sin 0d0 = 0

f
2. Rotationoftangentto centroidalaxis of Jo M3rd6 Jo
Plate 3 equals zeroat crown.
3. Deflectionsof Plates 1 and 2 at crown
IT V

are equal. (MirO)* + (y*r2cos $)* = 0

4. Deflectionsof Plates 2 and 3 at crown
are equal. Mzr - - yzr2= 0
5. Shear at end of Plate 3 equals sum of
shearsat endsof Plates 1 and 2.
2yzr = 2 (y3)(36.875)
M3=
The methodof virtualworkis used to ob- 7T 7T
tain equationsbased on the firstfourcon-
ditions. = 23.475^3 (Eq 2)
608 H. S. SWANSON ET AL. Jour.AW WA

TABLE 5
Three-PlateDesign Calculations {Example 2)

sin - cos -

Li (scaled from Plate 1) = 82.72 in.; L2 (scaled fromPlate 2) = 33.53 in.

Wi = A = ~ (79.7O)2(O.7O71) = 673,700 lb
|z,i2sin

W2 = A = (33.01)2(0.707l) = 115,600 lb
|z,22sin ^

£x3sinA
m= £- = 300^(0.707) = 35.36x3
6 6

x M=m-xyx P^te / ^(A5) Mx(As)

t«. tn.-/6 in* x{As) ~T~ ~~i
DJPth

Plate 1 (/ = 1 in. (assumed); As = 15.1 in.; dw= 54 in.)

7.6 15,500- 7.6yi 12.0 144 114.8 0.7972 12,400- 6.06^i

22.7 413,600-22.7^1 13.2 192 342.8 1.7854 738,400- 40.53^i
37.8 1,909,800-37.83/1 15.5 310 570.8 1.8413 3,516,500- 69.60^i
52.9 5,234,600-52.9^1 19.4 608 798.8 1.3138 6,877,200- 69.50;yi
67.7 10,971,800-67.7yi 25.4 1,366 1,022.3 0.7484 8,211,300- 50.67yi
82.2 19,584,500-82.2yx 35.5 3,728 1,241.2 0.3329 6,519,700- 27.36;yi
85.1 21,538,200-85.1^! 48.0 9,216 1,285.0 0.1394 3,002,400- 11.86^i
85.1 21,538,200-85.1^1 53.3 12,618 1,285.0 0.1018 2,192,600- 8.66^i

31,070,500-284.24^!

Plate 2 (t = 1 in. (assumed); As = 10.7 in.; db = 22 in.)

5.4 5,600- 5Ay2-M2 12.5 163 57.8 0.3546 2,000- 1.91^2-0.3546M2

15.8 139,500-15.83;2-Ar2 15.0 281 169.1 0.6018 84,000- 9.5l3>2-0.6018Af2
25.8 607,300- 25.83/2-AT2 17.7 462 276.1 0.5976 362,900- 15.42^ -0.5976ikf2
34.1 l,398,800-34.l3/2-lf2 19.9 657 364.9 0.5554 776,900-18.94^-0.5554^2
35.2 1,525,900-35.2y2-M2 21.2 794 376.6 0.4743 723,700-16.703>2-0.4743 AT2
35.2 lf525,900-35.2y2-Jlf222.0 887 376.6 0.4246 647,900-14.95;y2-0.4246Ar2

2,597,400- 77.433/2
- 3.008M2

Plate 1: tEdi = 31,070,500- 284.243/1

Plate 2: tE82= 2,597,400- 77.43^2- 3.008Af2
Plate 3: /£53= 62.883/3
June1955 WYE BRANCH DESIGN 609

TABLE 5 (contd.)
-
Plate 2 EquationforZero Ratationat Crown

x M Platft / (-D(Af) M(-l)(As)

in. in.-lb D^>th in* J 7

5.4 5,600- 5Ay-M2 12.5 163 -0.06564 -368 +0.354,2 +0.0656ikf2

15.8 139,500- 15.8, -Af2 15.0 281 -0.03808 - 5,312 +0.602,2 +0.0381 M2
25.8 607,300- 25.8, -M2 17.7 462 -0.02316 - 14,065+0.598,2+0.0232M2
34.1 1,398,800- 34. 'y-M2 19.9 657 -0.01629 -22,786+0.555,2+0.0163ilf2
35.2 l,525,900-35.2,-Af2 21.2 794 -0.01348 -20,569+0.474,2+0.0135M2
35.2 1,525,900-35.2,- AT2 22.0 887 -0.01206 - 18,402+0.425,2+0.0121Af2

-81,502+3.008,2+0.1687Af2

M(-l)(As) - 81,502 + 3.008,2 + 0.1687ilf2

a= _ = _ =0 n

M,-***^^- 483,100-17.83,2 (Eq 1)

(" M> +
53 = deflection
ofcrown= f^ =
f ^)xAs
ir ir

= 4? P i-M^r sin6)rd0
+ 4? &**sin2
d>dd
111 **o jli uPo

= 1 /^ • cos6)>
J , if Je sinecose'V
Tl(M>r> +_[^(---^-)J
1 /-2y3r2 7T^3' rWTT 2'
X( ur2,^'

_ 0.1488r3y3_ 0.1488(36.875)3y3(12) ~_ 62.88y,

El £(1.0)(11.25)3 E
Note: In solving simultaneous equations for shears and moments at the ends of plates
(crown), all plates are taken to be 1 in. thick. As / occurs in the denominator of all three
deflectionequations, it does not matter what thickness is used, so long as all three thicknesses
are equal. When solving for stresses in the plates, the actual thicknesseswill be used.

Plate 1 :
EtSi = 31,070,500 - 284.24yi
Plate 2 :
E», = 2,597,400 - 77.43,2 - 3.008ikf2
= 2,597,400 - 77.43,2 - 1,453,200 + 53.63,2 = 1,144,200 - 23.80,2

Plate 3:
Etfa = 62.88,3
31,070,500 - 284.24,i = 1,144,200 - 23.80,2 (Eq 3)
610 H. S. SWANSON ET AL. Jour. AW W A

284.24?! - 23.80^2 - 29,926,300 = 0

1,144,200 - 23.80^2 = 62.88?3 (Eq 4)
y* = yi + y2 (Eq 5)
f 62.88?! + 86.68?2 - 1,144,200 = 0
i - (62.88?! - 5.26?2 - 6,620,300 = 0)
91.94?2 + 5,476,100 = 0

y2 = _ 59,560 lb; y3 = + 40,740 lb; yx = 100,300 lb

M3 = 23.475?3 = 956,400 in.-lb
M2 = 483,100 - 17.83 (-59,560) = 1,545,000 in.-lb

Computationof Maximum Stresses

Plate 1:

a = 12sin = 30.5(0.3827) = 11.68in.

|

Maximummoment:

= 673,700 + 3.02 + - 100,300 3.02 +

( 7-^ ^ ) (79.70 + ^ )
= 23,573,000- 8,838,000= 14,735,000in.-lb

Maximumtensilestress:
Mhi , T - 14,735,000(27.0-7.1) , 673,700-100,300 = .
1^ + I + 26'2°°+ 4'25° ' 3°'45°P§1
54.0(2.5)(7.098)(11.68) 54.0(2.5)
Maximum compressive stress:

Plate 2:

- = 3.18; UTi= 1.28; UTo= 0.799

=^^

* Thisadditional wasmadenecessary
distance inthefabrication
of thewyeby therequirements forfitting
the
platestogetherproperly.
t c istheradiusofcurvature scaledinthiscase (seeFig.22,page626)owing
to theoutsidefibers(Timoshenko),
to extremely curves.
nonconcentric
| Scaled.
June
1955 wye branch design 611

Maximum
moment
:

0.52* - yt + 0.52+ - M2
JP.(y + +^b) (l, ^)
= 115,600 + 0.52+^) + 59,560(33.01+0.52 +^tt) - 1,545,000
(^f^
= 1,586,000 + 2,127,500 - 1,545,000 = 2,168,500 in.-lb
Maximum tensile stress (Seely) :
KMc
- T 115,600 + 59,560 = 13'77° - 15in= ,lftOrn ■
+I = 1.28(2,168,500) (6)
+ 1,r7n, + 3'18° +16'95°PS1
2.5(22.0)* 22.0(2.5)
stress:
Maximumcompressive

- -8,600+ 3,180
+ 3,180 - -5,420psi
-aWgy(6)
Plate 3 (half-ring)
:
"= = X, = 1.109;Ko = 0.910
5.625 6.56;
c TT7T
Maximummoment :
y»(r) - Mz = 40,740(36.875)- 956,400= 545,900in.-lb
Maximumtensilestress:

- »■«+'■«-+».• »*
^^'■^S9^
Maximumcompressive
stress:

TABLE 6
Three-PlateDesign (Example2)- Comparison ofResultsObtainedUsing
ThreeDifferentAssumptions
(a) All platesfreebut guidedat crown(s' = s2 = 53 =0).
(b) Plate 3 freebut guidedat crown,otherssimplysupported.
(c) Plates 2 and 3 freebut guidedat crown,Plate 1 simplysupported.

Stress- psi

Plate Location Computed

■ • Measured
(a) (b) (c)

1 outsideofcrotch - 5,500 - 7,050 - 8,090 - 5,850

1 insideofcrotch +26,900 +28,400 +30,450 +34,300
2 outsideofcrotch - 6,600 - 6,780 - 5,610 - 6,150
2 insideofcrotch +19,500 +16,950 +16,950 +16,700
3 outsideofcrotch -10,100 -13,680 - 7,930 -10,350
3 insideofcrotch +16,490 +22,330 +12,970 +15,750
3 outsideofcrown +18,960 +25,700 +15,800 +16,200
3 insideofcrown -23,300 -31,500 -19,300 -29,100
612 H. S. SWANSON ET AL. Jour.AW W A

Fromtheseand othercomparisons it may one anotherand providethe fixityneededby

be seen that the best resultsare obtained
whenusingthe designassumptionsin which
Plates 2 and 3 are consideredfreebut guided
(thatis, slopesoftangentsare zeroat crown)
and Plate 1 is consideredsimplysupported.
The justificationfortheseassumptions, aside
frombetteragreementwith test results,is
that Plates 2 and 3 are almostin line with
each plate for zero rotationat the crown.
Plate 1, however,is muchlongerthaneither
of the others,has less rigidity,and has no
opposingplate to restrainits rotationat the
crown.Therefore, even thoughit is not com-
pletelysimplysupported,its degreeof fixity
is relativelylow,and forpracticalpurposesis
consideredzero.

'
Platel3/ / Cp|ate 2 Plate1 ' ' Plate2 I I Plate3 ' '
y/ | If

Fig. 13. Three-Plate Design (Example 3)

Data: RB = 18 in.; Rs = 18 in.; A = 45 deg; p = 335 psi; pipe shell thickness,f in.;
depthat crown,12 in.

Design Assumptions 5. Deflection at crown of Plate 2 equals

1. Moment at crown of Plate 1 (Mi) equals that of Plate 3.
zero. The equations for virtual work are applied
2. Rotation of tangents to centroidal axes to conditions 1, 2, 4, and 5:
of Plates 2 and 3 at crownequals zero ; that is,
these plates are freebut guided at the crown. Eql:
Methodof Solution M2 = 206,000 - 15.17^2

The method of solution is similar to that Eq 2:

used in Example 2, the main differencebeing Mi = 16.47^3
that Plate 3 is elliptical, not circular. In the
followinganalysis, fiveunknownswill be con- Eq4:
sidered: 3>i,3>2,and y$ (the shears at the ends
4,022,500 - 64.713/1
of Plates 1, 2, and 3, respectively); and Mi = 1,079,400 - 72.61^2 - 4.028M2
and Mz (the moments at the crown ends of
Plates 2 and 3). M' is assumed to be zero, 2,943,100 - 64.7l^i
while Mi and ilf3are such that the tangents = - 72.61y2 - 4.028(206,000 - 15.17y2)
to the plates at the crownremainat zero slope
= - 72.61^2 - 829,800 + 61.10;y2
under load.
These unknowns are solved under the fol- 3,772,900 - 64.71yi + 11.51y2 = 0
lowing fiveconditions:
Eq 5:
1. Slope of tangent to Plate 2 at crown
equals zero. 1,079,400 - 72.61^2 - 4.028M2
2. Slope of tangent to Plate 3 at crown = 91.44^3 - 4.500 AT3
equals zero. 1,079,400 - 72.61^2
3. yz = yi -'- yi - 4.028(206,000 - 15.17^2)
4. Deflection at crown of Plate 1 equals
that of Plate 2. = 91.44y3 - 4.500(16.47^)
June1955 WYE BRANCH DESIGN 613

TABLE 7
Three-Plate
DesignCalculations(Example3)

sin- cos -

L' (scaledfromPlate 1) = 49.5 in.; L2 (scaledfromPlate 2) = 19.75in.

PLS sin A = 335(47.04)>(0.7071)_ ^^
m = ^

^=^sinA=335(19.48)(0.7071)=449601b

sin A _ 335^(0.7071) =
m=Px> ^^
6 6

L inM,b -? ••- •<"»^ ^F

Plate 1 (h = 2 in.; As = 9.48 in.; dw= 30 in.)

4.7 4,100- 4.7yi 12.4 318 44.6 0.1403 600- 0.66^i

14.2 113,000-14.2yi 13.0 366 134.6 0.3678 41,600- 5.22^i
23.7 525,600-23.7yi 14.3 487 224.7 0.4614 242,500-10.94^1
33.1 l,431,700-33.1yi 16.5 749 313.8 0.4190 599,900-13.87^1
42.5 3,030,700-42.5;yi 20.2 1,374 402.9 0.2932 888,600-12.46^1
51.6 5,302,900-51.6yi 25.2 2,667 489.2 0.1834 972,600- 9.46yi
52.5* 5,538,700-52.5yi 29.2 4,150 497.7 0.1199 664,100- 6.29^i
52.5 5,538,700-52.53fi 30.0 4,500 497.7 0.1106 612,600- 5.81yi

4,022,500-64.71^1

Plate 2 (t2= 1 in.; As = 6.52 in.; db = 12 in.)

3.2 1,300- S.2y2-M2 12.3 155.1 20.9 0.1348 200- 0.43y2-0.135ilf2

9.5 33,800- 9.5y2-^2 12.2 151.3 61.9 0.4091 13,800- 3.89;y2-0.409ilf2
15.3 141,400-15.3^-^2 12.1 147.6 99.8 0.6762 95,600- 10.35y2-0.676M2
20.1 319,700-20.1y2-ilf2 12.0 144.0 131.0 0.9097 290,800-18.28y2-0.910ilf2
20.95 357,900-20.9y2-M , 12.0 144.0 136.6 0.9486 339,500-19.83;y2-0.949ilf2
20.95 357,900-20.9y2-ilf2 12.0 144.0 136.6 0.9486 339,500- 19.83y2-0.949M2
- 72.61y2
1,079,400 -4.028M 2

Plate 3 (t3= 1 in.; As = 6.56 in.; db = 12 in.)

3.3 0+ 3.3y3-Mz 12.0 144.0 21.6 0.1500 + 0.50y3-0.150ilf3

9.7 0+ 9.7^3-^3 12.0 144.0 63.6 0.4417 + 4.28y3-0.442M3
15.5 0 + 15.5^3-^3 12.0 144.0 101.7 0.7062 + 10.95?,-0. 706Af3
20.5 0+20.5^3-^3 12.0 144.0 134.5 0.9340 +19.15y,-0.934M3
24.0 0+24.0y,-M3 12.0 144.0 157.4 1.0933 +26.24y,-1.093Jlf3
25.8 O+25.8y,-Jkf, 12.0 144.0 169.2 1.1753 +30.32y,-1.175ikf3

+91.44y3-4.500ilf3
* Maximum depthofplatefrominsideedge.
x distanceis to pointone-tenth
614 H. S. SWANSON ET AL. Jour.AW W A

TABLE 7 (contd.)
Equation for Zero Rotation at Crown

x M = m - xyi - M« Xlatfe. / (-U(As) M(-l)(As)

in. in.-lb DJnpth in* 7 7

Plate 2

3.2 1,300- 3.2y2-M2 12.3 155.1 -0.04204 -50+0.135:y2+0.0420ilf2

9.5 33,800- 9.5y2-M2 12.2 151.3 -0.04309 - 1,460+0.409^2 +0.043 'M2
15.3 141,400-15.3^2-iW2 12.1 147.6 -0.04417 - 6,250 +0.676;y2+0.0442 Af2
20.1 319,700-20.1^-^2 12.0 144.0 -0.04528 - 14,480+0.910y2+0.0453ilf2
20.95 357,900-20.9^2-M2 12.0 144.0 -0.04528 -16,200 +0.946^ +0.0453 M2
20.95 357,900-20.9^-^2 12.0 144.0 -0.04528 -16,200+0.946^+0.0453^2

-54,640 +4.022^2 +0.2652il/2

Plate 3

3.3 + 3.3ys-M3 12.0 144.0 -0.04555 -0.150^+0.0455^3

9.7 + 9.7^3 -Mz 12.0 144.0 -0.04556 -0.442y3+0.0456M3
15.5 +15.5y3-M3 12.0 144.0 -0.04555 -0.706y3+0.0455ikT3
20.5 +20.5^3-^3 12.0 144.0 -0.04556 -0.934y3+0.0456ikf3
24.0 +24.0^-^ 12.0 144.0 -0.04555 -1.093y3 +0.0455 Af3
25.8 +25.8^-^3 12.0 144.0 -0.04556 -1.175y3+0.0456M3

-4.500y3+0.2733ilf3

^MtnAs - 54,640 + 4.022y, + 0.26527lf2

a2 = s-Er = s^M(-l)As = =0n
ei e
- 15'17^ CEq1)
" 206'000
M2=54Ml;6s2022y2
_,MmAs = -
a3= 2^r -
M(-l)(As) =- 4.500y3 + 0.2733 M3 = n0
ei e
(Eq2)
^-s-1"7^
1,079,400 - 72.61y2 - 829,900 + 61.10y2 yx = +54,060 lb
= 91 My t - 74.12^3
y3 = +30,240 lb
249,600 - 11.51^2 = 17.32;y3 M% = 567,400 in.-lb
Eq3: Ms = 498,000 in.-lb
ys = yi + y* Computationof Maximum Stresses
249,600 - 11.51y2 = 17.32(yi + y2) Plate 1 :
r 249,600 - 28.83y2 - 17.32yi = 0
a = ^ = 18.0(0.3827) = 6.89 in.
1-1,009,800 - 3.08^2 + 17.32yi = 0 sin|
-760,200 - 31.91yj ^0 r _ 6.89 + 15.0 _ 2L89 _
~ " ~ *
3/2= -23,820 lb c 15.0 15.0
June1955 WYE BRANCH DESIGN 615

2fi' 14l%j#8 25^ 37JfV 49§'^ 6lI"-H

KH^i" ' ' -
f *M* ?qii -^l#6+vl
I E-2 D-2 C-2
™ ° ° =
2 = 2"
PlateThickness
nX
T p i 7 T 2 ^.ic* ,?^K

411 '"V-Z J f1" ' '

*ii,- ^nTTF*?*'^--^ #12 ' i» I
-
I yUr-^f xr ^^' !«
' 22" >il«
//-2 ^ 11 = 1" X!*
PlateThickness 30" ^
h-49i"^l.;

/"164 '
#9-/_3o() I 18nt^#15 <^^ y^^
/F-2 F-l/ 1
±1 lAf-1 A/-2| 0 5 10
Lj_f f_jL/ L| p
l'ij ^ 10" >1k-1" 1"*| h« 10" H kl" I I I
k 12" 4* 18}" A U 19" A* 12" »l scale ininches

- Ib
Stress* - Ib
Stress*
Item Item
Computed Measured Computed Measured

A-l 6,950 9,310 L-l 4,710 4,100

A-2 -2,340 -3,730 L-2 -630 -800
5-1 8,000 10,130 M-l 13,760 13,550
B-2 -3,440 -4,060 M-2 -6,400 -6,220
C-l -3,890 -5,180 #1 7,760 5,460
C-2 -5,570 -3,190 #2 1,160 3,750
ZM -7,870 -5,580 #3 -8,450 -1,290
D-2 9,250 8,380 #4 -5,980 300
£-1 -1,630 -4,390 #5 -2,740 -4,800
E-2 1,990 2,440 #6 -3,920 -4,350
FA 14,500 12,500 #7 1,520 2,400
7^-2 -3,000 -4,420 #8 11,360 8,550
G-l 14,070 10,900 #9 -4,360 -3,810
6-2 -4,000 -420 #10 -800 2,460
HA -4,520 4,800 #11 13,950 9,210
H-2 6,300 7,010 #12 7,220 6,450
JA -14,310 -5,150 #13 -5,840 -4,590
J-2 17,700 11,800 #14 -5,530 810
KA -12,180 -12,870 #15 10,380 12,150
K-2 13,110 9,850
* At335-psipressure.
Fig. 14. Three-PlateDesign (Example 3) - Plan of Stiff
ener Plates
See italicnotebelowFig. 7.
616 H. S. SWANSON ET AL. Jour.AWWA

Ki = 2.00; Ko = 0.641
Maximummoment
:

= 262,000 + 2.46*+ - 54,060 + 2.46+ ~) = 2,700,000

in.-lb
(^f^ yjj) ^47.04
Maximumtensilestress:
KiMc , T = 2.00(2,700,000)
(6) , 262,000-54,000 = «o/w, „ „ = „
-T + A 2.0(30)*
+ 18'°0°+ 3'47° + 21'470
2.0(30)
psl
Maximumcompressive
stress:

^p +f=~°ToxTo)(6) - -5-770
+3-470 --2-300
+3-470 psi
Plate 2:

a = = 18.0(0.9239)= 16.63in.
i?cos|

£ = !*"±«4 . = 3.77; * = 1.220;*. = 0.840

c 6.0 ^ 6.0
Maximummoment:

= -567,400 + 44,960 + 0.27* + - 23,820 + = 281,400in.-lb

(^j^ j^) (l9.75 j^)
Maximumtensilestress:

-KiMc +, T = 1.220(281,400)(6) +, 44.960 + 23,820= 14'3°°

4AOM r^
+ 5'73° = 20'03°PS1
A 1.0(12)* 1.0(12.0)
Maximumcompressive
stress:
-0.840(28M00)(6)
=E^+T_ + 5730 _ _9>850+ S73O _ _4>i2o ^

Plate 3
Ki = 1.220; Ko = 0.840
Maximummoment:
- Mi + yz = -498,000 + 30,240 0.52* + ^~' = 288,200in.-lb
(l« + y ^ (l9.48 +
Maximumtensilestress:
KiMc T -1.220(288,200)(6) +, 30,240 = 14'650
4Adegn = +17'170
+ 2'520 PS1
-t+a= i.o(i2)2 rofe
Maximumcompressive
stress:
~-K0Mc , T -0.840(288,200)(6)
+ 2'52° = -10'090 + 2>520= -7'570 Psi
*A= 1.0(12)'
* This added distancewas made necessary fabrication
during ofthewyeinordertofittheplatestogether.
June1955 WYE BRANCH DESIGN 617

NomographMethod
Duringthe periodof this studya them. Thereremainedone otherim-
meanswas soughtby whichthe data portantdimensionof the reinforcing
on similitude of designcould be pre- plate to consider - thatat the top of
sented. Curveshad beenmadein the the sections. When this was investi-
past, but these provedcumbersome.gated,it was foundto varyquiteclosely
Finallyit was decidedto presentthe ina gooddesignwiththesmaller crotch
materialin the formof a nomograph depth(db). By checking through all
(Fig. 15). After the authors had in- thepastdesigns,it was possibleto plot
vestigated thepossibility and developed anothersetofcurvesof topdepthver-
theform,a seriesof two-plate 90-deg sus base depth. One of thesecurves
wye branches were designed pres- was based on the one-platedesigns,
for
suresfrom100to 600 psi. The result- theotherson thetwo-plate designsat
ing crotch depths were found to follow variousdeflection
angles (Fig. 18).
a curvedpattern, and thiswas usedto In examiningtheequationsofsimili-
plottheinterpolated pointson theplate tudeand theirrelationto thiswork,a
depth scale. Later the pressurescale methodof converting the depthand
was extended to 1,200psi bythesame thicknessof plate for curvedbeams
procedure. When the 90-deg wye was sought. Usingthesectionmodu-
branchdesign appearedcomplete,a lus equationdid not give accuratere-
seriesof designswas madeat various sultsin thiscase, becauseit did not
angles of deflection.These results holdthestressesconstantas theplate
were comparedwiththe 90-deg de- was variedin cross section. By em-
signs,and thereappearedto be a con- piricalmethodsa conversion equation
stantproportion betweenthe two at was derivedby whichthe depthand
any particular deflection angle. This thickness of plate could be variedso
proportion, called the N factor, was thatthe depthwas nevermorethan30
plottedagainst the deflectionangle of times the thickness. This equation
thewyebranch(Fig. 16). was based on test resultsand was
Up tothispointtheauthorshadcon- checkedby being used on sections
sideredonlysections withequaldiame- whichwerelatertested. Fromthisit
ters and two-platedesigns. Next a was assumedto be reasonably accurate.
seriesof one-platedesignswas calcu- In orderto checkthe nomograph,
latedforvariousratiosof pipe diame- completethe curves,and derivethe
ters. When the resulting dimensionsequationsused,a greatnumberof ex-
of thesedesignswere comparedwith amples was required. These were
thoseof thetwo-plate designs,a rela- mostlybasedon theoretical knowledge.
tionship was found thatdependedon Test resultshad beenused throughout
thedeflection angle. In orderto graph as a basis forthe designdata, but it
the comparisons, it was necessaryto was stillnecessary to makeothertests
plota familyof curvesat variousde- to check the curves as they were
flectionangleson a graphof pipe di- formed. In this series of complete
ameterratioversusa proportion called testson sectionswhichwerebasedon
theQ factor(Fig. 17). Thesecurves the nomograph methodof design,the
fellso closetoeachotherthatit seemed stressesin the criticalsectionswere
simplerto plotonlythosefor90, 60, analyzedto ascertaintheireffecton
and 30 deg and interpolate between anycurvesused in thedesign. From
 °-qy
- ' |-2io
1-200
- ' | - 190
i2- ' b180
- |- 170
'
- ' | - 160
- ' |-150
24- ' [L 140
- 1-130
'
30 ' i-
- |-120
'
-
36 ' [ HO
- ^-100
' z-
42 '

: i90
48- ^-'
%pOoA j-80
c 54 a00 ^K% l~70 .E

3 60- VA' ^6°l

72- a^°^' 5- 40

: %o° ' ^~30

~
84 '

A
<i>0 1- 20

96 ^^ A -
' - 10

108 '

120 * J} 0
0

forSelectingReinforcement
Fig. 15. Nomograph Plate Depths
of Equal-Diameter
Pipes
1 in.; deflection
Plate thickness, angle,90 deg.
June1955 WYE BRANCH DESIGN 619

thisinformation the curveswere ad- and dhshouldbe converted to conform

justedby making the designpressure to a greater thickness,
t, by use of the
thatpressurewhichgave thesectiona generalequation:
stressof 30,000psi at its mostcritical
point. An effortwas madeto balance
the stressesin the variousparts so
thateachhada stresscloseto thecriti-
cal stressin magnitude. in whichd± is the existingdepthof
On the following pages the nomo- plate; t± is the existingthicknessof
graphicmethodofdesignis illustrated,plate; d is the new depthof plate;
alongwithexamplesand testresults. t is thenewthickness ofplateselected ;
Superimposed on the drawings of the
testedsectionsare the stresseswhich 90c f 1 1 1
occurredon theinnerand outerfibers
of thereinforcement plates. The ten- 80 = +
sionstresseswereplottedoutwardand
thecompression stressesinward.
- -V
i7oi
NomographDesign
Step 1. Lay a straightedge across - ~~ '
thenomograph (Fig. 15) through the < 60= 1 V
appropriate pointson the pipe diame-
ter(see Step2b) andinternal-pressure q 50 = 1 i '-'-I
V^
scales; readoffthedepthofplatefrom
its scale. This readingis the crotch
depthfor 1-in.thickplatefora two- 40
= V -V

plate90-degwye branch pipe.

Step 2a. If thewyebranchdeflec- ^i nil in In ii |' 11 n nl n ii 11ii 11in*mi Iiiii
tionangleis otherthan90 deg,use the 0 1.0 2.0 3.0 4.0 5.0
N factorcurve (Fig. 16) to get the N Factor

factorswhich,whenmultiplied by the Fig. 16. N Factor Curves

of
depth plate found in Step 1, will For wyeswithdeflection anglesfrom30
givethe wye depth,dw,and the base to 90 deg, the N factorsobtainedfrom
depth,rf&, forthenew wyebranch. theabove curvesare appliedto theplate
Step 2b. If the wye branchhas depth,d,foundfromthenomograph (Fig.
unequal-diameter pipe, the larger- 15), in accordancewith the equations:
diameterpipe will have been used in dw= Nwd;db= Nbd.
Steps 1 and 2a, and these results
shouldbe multiplied by the Q factors
found on the single-platestiffenerand is
A thedeflection angleofthewye
curves(Fig. 17) to give d'wand d'b. branch.
These factorsvary withthe ratio of Step 4. To findthe top depth,dt
the radius of the small pipe to the or d'u use Fig. 18, in whichdt or d't
radiusof the largepipe. is plottedagainstdhor d'. This di-
Step3. If thewyedepth,dw, found mension gives the top and bottom
so far is greater than 30 times the depths plate at 90 deg fromthe
of
thicknessof theplate (1 in.), thendw crotchdepths.
620 H. S. SWANSON ET AL. ] our.AW WA

1.0p 1 1 1 1
T|

"I ~M~
"I
l2 =~ >A, / / / / /

~j6v7T
02 - / f *r r f

0 ^mTTmii i 11111111111111111111ii 11111111111111111111111111111111111111111111n 11111111111111111

0 0.2 0.4 0.6 0.8 1.0

Fig. 17. Q Factor Curves

For pipes of unequal diameter,finddw and db for the larger-diameterpipe {from Fig.
15 and 16) ; then: Qwdw= d'w, crotchdepth of single-platestiffener;and Qbdb= d'b,
base depth of single-platestiffener.

Step 5. The interior curvesfollow The importantdepthsof the re-

the cut of the pipe, but the outside
crotchradiusin bothcrotchesshould
equaldtplustheradiusofthepipe,or,
design,d' plus the
in thesingle-plate
radiusof the smallerpipe. Tangents
connectedbetweenthesecurvescom-
pletetheoutershape.
inforcement plates,dw,db,and dt (Fig.
19) can be foundfromthenomograph.
If a curvedexterior is desired,a radius
equal to theinside pipe radiusplus dt
can be used,bothfortheoutsidecurve
of thewyesectionand fortheoutside
curveof thebase section.
June1955 WYE BRANCH DESIGN 621

: I I I I I I I I I /I V' l/f

: / / / /
¥ / Z Z

30 / / -■vVriy
/ / / A

■fir-Am--
1 20 1-7-/7

10--H-H-

or 1 1 1 1 1I 11 1I 111I 111 11I 111I I 1 l

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Base Depth,db ord'b- in.

Fig. 18. Selection of Top Depth

d'tand d' are one-plate
designdimensions; designdimensions.
dtand dbare two-plate
622 H. S. SWANSON ET AL.
X ' '.«<s$*"/>
WyeSection ^O^ BaSeSection
' ~&j/
-* -:S=* - WyeSection
jr
Fig. 19.
Example 1- One-PlateDesign
Rb = 30 in.
Rs = 21 in.
A = 45 deg
Working pressure,230 psi
Design pressure,230 X 1§ = 350 psi
Step 1. With the larger pipe diameter 60
in. and the design pressure 350 psi, read the
critical plate depth, d, from the nomograph
(t = 1 in., A = 90 deg) :
d = 50 in.
Step 2. Using the deflectionangle 45 deg,
find the factors on the N factor curve which
will convertthe depth foundin Step 1 to apply
to a 45-deg wye branch (t = 1 in.) :
dw = KJ, = 2.45(50) = 122 in.
db = Kid = 1.23(50) = 61.5 in.
Step 3. With the ratio of the smaller pipe
radius divided by the larger pipe radius
(?±- = ?A = and the deflection angle
o.7o)
(A = 45 deg), use Fig. 17 to findthe Q factors
which give the crotch depths for a single-
plate pipe wye stiffener(t = 1 in.):
Qw = 0.52
Qb = 0.66
d'v, = 0.52(122) = 63.4 in.
d'h = 0.66(61.5) = 40.5 in.
Step 4. As the depth, d'w, is greater than
30 times the thickness,t, the conversionequa-
tion should be used :
Jour.AW W
TopSection
' *v
y*
U^dw-U Base Section
1 H'^'H
^ *•$ '
Plan and Layout
Try a thicknessof 1J in. :
d=di'u) -di'i)
d = <Ji(0.725)
d'w = 63.4(0.725) = 46 in.
d'h = 40.5(0.725) = 29 in.
Step 5. Find the top depth, d'tj from the
curve for one-plate design in Fig. 18:
ford'h = 29 in., d't = 18 in.
Final results:
Thickness of reinforcingplate, / = li in.
Depth of plate at acute crotch,d'w = 46 i
Depth of plate at obtuse crotch,d'h = 29 ir
Depth of plate at top and bottom, d't =
in.
Outside radius of plate at both crotch;
equals the top depth plus the inside
radius of the smaller pipe, d'*-'- R8 = 18
+ 21 = 39 in.
Example 2- Two-PlateDesign
RB = R8 = 36 in.
A = 53 deg
Working pressure, 150 psi
Design pressure, 150 X 1} = 225 psi
Step 1. With the pipe diameter 72 in. and
a pressureof 225 psi, read the critical depth of
plate fromthe nomograph (/ = 1 in., A = 90
deg):
d = 49 in.
Step 2. From the N factor curve, find the
two factorsat A = 53 deg; then, at t = 1 in,:
dw = 1.97(49) = 96.5 in.
db = 1.09(49) = 53.4 in.
 June1955 WYE BRANCH DESIGN 623

Step3. As dwis greaterthan30 timesthe If a thirdplateis desiredas an addi-

thicknessof the plate, try t = 2 in. in the
tion to the two-platedesign,its size
conversionequation: shouldbe dictatedby the top depth

'-*(?r"*-*ar
= dx(0.586)
(dt). Becausetheothertwoplatesare
flushwiththe inside surfaceof the
pipe,however, theshellplatethickness,
= =
dw 96.5(0.586) 57 in. plus clearance,should be subtracted
= =
db 53.4(0.586) 31 in.
from the top depth. This dimension
shouldbe constant throughout, and the
Step4. Read the top depth,dtyfromthe plateshouldbe at
two-platedesigncurvein Fig. 18: placed rightangles
dt = 15 to theaxis ofthepipe,givingit a half-
Final results: ringshape. Its thickness shouldequal
thesmallerof themainplates.
Thicknessof reinforcing / = 2 in.
plate,
Depthofplateat acutecrotch,dw= 57 in.
Depthofplateat obtusecrotch,db = 31 in. References
Depthof plateat top and bottom,dt = 15 1. Force Mains. Sulser BrothersCatalog
in.
Outsideradius of plate at both crotches, (1931).
51 in. 2. Sulzer Patented StiffeningCollars on
the Branchesof High-PressurePipe-
Three-PlateDesign lines for HydroelectricPower Works.
Sulser Tech.Rev., No. 2, p. 10 (1941).
The precedingnomographsection 3. Blair, J. S. Reinforcement of Branch
has coveredthe designof one- and Pieces. Engineering (Br.), 162:1
two-plate wyebranches without touch- 4. (Jul. 5, 1946).
Blair, J. S. Reinforcement of Branch
on
ing a three-plate designbecauseof Pieces. Engineering (Br.), 163:63
.l;itssimilarity to the two-plate.The (Jan. 17, 1947).
ftfunction ofthethirdplateis to act like 5. Murphy, Glenn. Similitudein Engi-
a clampin holdingdownthedeflection neering. Ronald Press Co., New
York (1932).
ofthetwomainplates. In doingso,it 6. Bier, P. J. Welded Steel PenstocksDe-
acceptspartofthestressesoftheother sign and Construction. US Bur.
platesand permitsa smallerdesign. ReclamationEng. Monographs,No. 3
This decreasein thedepthsofthetwo (Jul. 1949).
mainplatesis smallenoughto make 7. Seely, F. B. Advanced Mechanics of
Materials. JohnWiley & Sons, New
it practicalsimplyto add a thirdplate York (1932).
to a two-plate design. The additional 8. Hetenyi, M. Handbook of Experi-
plateshouldbe considered a meansof mentalStressAnalysis. JohnWiley&
holdingdown the deflection at the 9. Sons, New York (1950).
Method for the Determina-
Graphical
junctionof the plates. The two fac- tion of Stresses From Strains on
torswhichdictatethe use of a third Three IntersectingGage Lines, and
plateare diameter ofpipeand internal Its Applicationto Actual Tests. /.
pressure. When both of these are ResearchNatl. Bur. Stds., 10: (May
abovea certainlimit,a ringplatecan 10. 1953).
Timoshenko,S. Strengthof Materials.
be used advantageously.Duringthe D. Van Nostrand Co., New York
workon thisreport,it was concluded (1930).
thatthese limitscould be 60-in. ID 11. Paris, J. M. Stress Coefficientsfor
and 300 psi. If eitherofthesefactors Large HorizontalPipes. Eng. News-
Rec. (Nov. 10, 1921).
is belowthelimit,thedesignershould 12. Unfired Pressure Vessel Code. Am.
be allowedto choosea thirdplate. Soc. Mech. Engrs.,New York.
624 H. S. SWANSONET AL. Jour.
AW WA

Test Results

One-Plate Wye
A singlecurvedplateservesas reinforce-
ment for this 30 X 61-in., 45-deg wye
section{see Fig. 20).

Tension
#1712" ffii2'Umr%jH
"HITS 12"4 21
forffi"
" ^19 j"j r-^^u^^.
P.ate 1" .
Thickness-
jS ^^T

#1 8 8 #12
^1 ^>'"" "c^ ^
#13^ A3 ' #25
111r"24" j 0 10 20
Scale in inches
Y^^^ J^^

Tfo_
Item Measured T-ttW1
Item Measured Tfot_,
Item Measured
Stress*- Ib Stress*- Ib Stress*- 16

#1 32,000 #11 22,000 #21 -8,000

#2 34,000 #12 23,000 #22 -9,000
#3 31,000 #13 -12,000 #23 -12,000
#4 12,000 #14 -12,000 #24 -12,000
#5 -8,000 #15 -11,000 #25 -12,000
#6 -20.000 #16 -10,000 #26 28,000
#7 7,000 #17 -11,000 #27 0
#8 21,000 #18 -4,000 #28 22,000
#9 25,000 #19 10,000 #29 3,000
#10 23,000 #20 1,000
* At 250-psipressure.

Fig. 20. One-Plate Design- Plan of Stiffener

Plate
Data: RB = 30.5 in.; Rs = 15 in.; A = 45 deg; p = 250 psi. See italic note below
Fig. 7.
June1955 WYE BRANCH DESIGN 625

Two-Plate Wye
This 30 X 30-in.,90-degwyesectionhas
twocrotchplates{see Fig, 21).

... cn #16 fill #17 #18 63ii #19

PlateThickness
=1" • ob

. im^ILL«,
'

,_'
y&>

mM^
W • #22
J^<v#^
If J# ^J
#104i-J #20 '•
i |-.Vj«l:fl #2
17" 2i£" »T
p| l-::-;j|«
Ȥ 1M^#1 0 5 10
i,c^ - - - ""■"""""""""" Scale in inches
#9JJi'^

"
Tt«m Measured T. _ Measured Tt.m
Item Measured
Item Stress*-^& Item Stvzss*-lb Stress*- Ib

#1 30,000 #9 -7,000 #16 -5,000

#2 30,000 #10 -8,000 #17 -4,000
#3 30,000 #11 -7,000 #18 -1,000
#4 24,000 #12 -7,000 #19 -1,000
#5 15,000 #13 -7,000 #20f 6,000
#6 11,000 #14 -6,000 #21f 15,000
#7 7,000 #15 -6,000 #22f 6,000
#8 14,000
* At 360-psipressure. t Principalstressesof rosettes.

Fig. 21. Two-PlateDesign-Plan of Stiffener Plates

Data: RB = 15 in.; RB= 15 in.; A - 90 deg; p = 360 psi. See italic note below
Fig 7. See also Table 8 (page 627).
626 H. S. SWANSON ET AL. Jour.AW WA

i
Tension *■
j-r^'T r^ ^r t>
n #2120" '££ 20" g| #27
2Q!!n^::tioM:!#i5^2bMr#26
Compress.cn =2},,
p,ateThickness ^M
. J^ ; J_

^'" 4.liptical *•
^ ,£M$* Curves-"; ♦laf&^SL

^fe ^- 54" p;>y£ 82.72" >+«- 33.53"fibgW::122"-^S

*31
#17^"""' Angle-mounted^

pension IIT/11^"

I () 0 10 20 30
^V>^/#36 (_ x^^^A
y'^-22" Scale ininches
#34j^>v ^^Compression ^^^
Compression*<3<3>^ vJa^u. ^' y^

- -3l^ Tj,
Tension
I
#32ip-Jt3t#35

If
ltem Measured Tt
ltem Measured T,
Item Measured
Stress*-^6 Stress*-^ Stress*- Ib

#1 25,000 #16 22,000 #30 -6,000

#2 34,000f #17 -8,000 #31 -7,000
#3 27,000 #18 -6,000 #32 16,000
#4 16,000 #19 -9,000 #33 3,000
#5 12,000 #20 -10,000 #34 -31,000
#6 4,000 #21 -10,000 #35 -10,000
#7 -4,000 #22 -15,000 #36 -2,000
#8 -8,000 #23 -13,000 #37 17,000
#9 -5,000 #24 23,000 #38 -3,000
#10 -1,000 #25 24,000 #39 -1,000
#11 5,000 #26 0 #40 4,000
#12 13,000 #27 -2,000 #41 28,000
#13 21,000 #28 -5,000 #42 1,000
#14 21,000 #29 -6,000 #43 11,000
#15 21,000
* At 335-psipressure. t 22,000 at 200 psi.

Tig. 22. Three-PlateDesign- Plan of Stiffener

Plates
Data: RB = 30.5 in.; Rs = 30.5 in.; A = 45 deg; p = 300 psi. See italic note below
Fig. 7.
June1955 WYE BRANCH DESIGN 627

Three-PlateWye
In this61 X 61-in.,45-degwye,theplate
on thebackis notweldedto thepipeshell
{see Fig. 22).

TABLE 8
MeasuredStresses
for30 X 30-in.,90-degTwo-PlateWye
{See Fig. 21)

Stress- Ib*
Gage -
100 psi 200 psi 300 psi 360 psi 400 psi

#1 9,400 18,500 26,500 30,400 33,000

#2 8,600 17,300 25,400 29,500 32,300
#3f
#4 5,600 13,100 20,000 24,200 27,000
#5 3,300 7,500 12,300 15,200 17,100
#6 2,500 5,600 9,400 " 11,300 12,500
#7 1,700 3,500 5,400 7,100 8,300-
#8 2,900 7,100 12,100 14,400 16,000
#9 -1,700 -3,800 -5,600 -7,100 -8,100
#10 -1,700 -4,100 -6,000 -7,600 -8,600
#11 -1,500 -3,600 -6,000 -7,400 -8,300
#12 -1,700 -3,600 -5,400 -7,000 -8,000
#13 -1,800 -3,800 -5,400 -6,800 -7,800
#14 -1,500 -3,600 -5,000 -6,300 -7,100
#15 -1,200 -3,300 -4,700 -5,800 -6,600
#16 -1,200 -3,300 -4,500 -5,400 -6,000
#17 - 800 -2,300 -2,900 -3,600 -4,100
#18 - 200 - 900 -1,100 -1,500 -1,700
#19 - 200 - 800 - 800 -1,200 -1,400
#20J 2,500 3,100 4,900 5,700 6,200
#21J 4,200 7,800 12,300 14,800 16,500
#22J 1,600 3,000 4,900 5,600 6,100
* Minus 6ignindicatescompression.
t Gage No. 3 failedbut was consideredto have the same stressesas No. 1 because the two weresymmetrically
placed.
XRosettes.
628 H. S. SWANSON ET AL. Jour.AW WA

APPENDIX A
ComparisonofSulzer and NomographDesign
TABLE 9
ofDesigns
Comparison
In 1941 Sulzer Brothersreported
(2 *) a wye branchwhichwithstood Item Sulzer Nomo-
an internal pressure of2,410psi before
it failed. For thesakeof comparison,Main pipe diameter - in. 23.6 23.6
Branch diameter - in. 9.8 9.8
theirdata were convertedfromthe pipe
Deflectionangle- deg 60 60
metricsystemintoinchesand pounds, Service pressure- psi 750
and the presentauthors'nomographTest pressure-psi 1,140 1,140
methodof designwas used to obtain Failurepressure - psi 2,410
valuesforoveralllengthand widthof Shellstressat failure-psi 64,000
thereinforcing plate. The resultsare Thicknessofreinforcing plate 0.79 0.79
compared in Table 9. - in.
Overalllengthofcollar- in. 31.5 35.2
Overallwidthofcollar- in. 26.3 27.4

APPENDIX B
DeflectionNomograph
The nomograph in Fig. 23, basedon amplerepresented bythebrokenline-
testdata,can be used forfinding de- a 54-in.ID welded-steelpipe at a de-
thatwilloccurat thejunction
flections signpressureof350 psi- thedeflection
ofthereinforcementplatesin two-and fora three-platedesignis foundto be
three-plate
wye branches. In the ex- 0.068in.; fora two-plate,0.098in.

APPENDIX C '
ResultsofOverstress
During the testingof the various at thecrotchappearedto act in an er-
wyebranchpipesin thisreport,there raticmanner.Whentheinternal pres-
was opportunity to observetheirac- surewas released,a numberof gages
tionswhenoverstress occurs. In the failedto returnto theirzero reading.
latertests,whenthedesignswere,cut Examinationof these gages showed
to a minimum and straingages were thatthesteelin theirvicinity was per-
placedinsidethe sections,the critical manentlydeformed. In still other
areas of stressreachedtheyieldpoint areas,thesteelwas undercompression.
verycloseto thetestpressure.When It appearedthat,underhighstresses,
additional pressurewas appliedto the the sectionacted in a plasticmanner
section,thestraingageson theinside and distributed stressesfromcritical
* See reference
list,page623. areas to less criticalones by yielding.
June1955 WYE BRANCH DESIGN 629

0 - * 0.40
I

™
' 0.55

' 0.35
' 0.50 _

24"" ' Z~
' 0.45 -
~
30 - ' 0.30

0.40 -~
'
"" -
42 - '

' I °25
48 - ' 0.35
"" "
' -
.5 M _ V
54 ' 5« - -
I '

I -
a: o.3O-
n
' ' |

'' I Z7"0
:-
° 66 -

x '
' '% &
V 0 25 - _
'

- j:
- °15
-
xA
'feCP'^v

'
">N
o.15- _01Q

' 0.10 ^S -
790
' :-
"' - 0.05
~
108- '°K

120J O-^~°
Fig. 23. Nomographfor Deflectionsin 45-degWyes
630 H. S. SWANSON ET AL. Jour.AW WA

Then,whenthepressurewas released, actuallybe stressedin tension.Thus,

the yieldedmetalwas put into com- the elasticlimitof the sectionas a
pression,along with nearby areas wholewillbe raised.
whichhad notbeenoverstressed.The Severalsectionsweretestedfarbe-
conditionthat remainedwas a pre- yondtheirtest pressurein orderto
stressedcriticalarea whichwas under watchfor visiblesigns of failureor
a permanent compression at zeropres- deformation in the steel. Lueder's
sure. When this sectionis nextput lines appearedin the reinforcement
underinternalpressure,the apparent plates at about twice the pressure
stresswillhaveto exceedtheresidual whichfirstbroughtthe criticalareas
compressive stressbeforethearea will to theyieldpoint.

Correction
forQuicklimeand HydratedLime-
The TentativeStandardSpecifications
AWWA B202-52T (January1953Journal) containan error. In Paragraph
4C.2.3,theamountof concentrated acid is givenas 37.5 ml. The
hydrochloric
correctfigureis 87.5 ml.