C E NT R IF U GAL PU M PS

T H E IR

D ESIG N AN D C O N ST R U C T IO N

L O U IS C . L O E W E N STE IN , E E . PH D . .

AN D

C L AR E NC E P . C R IS SE Y
, M E .

320 IL L US TBA TI ON S
8 F OL DI N G P L A TE S

N E W Y OR K

D V AN N OSTRAND COMPANY
.

2 3 M UR R AY AND 2 7 W AR RE N STS .

1 91 1
 W WW!

P R E F AC E .

T HE ra pid subst itution of rota ting for re c ipro ca tin g m achinery h as

been one of the most striking engineering developments of recent
yea rs The m ar ked success of the ste a m turbine no doubt stimu
.

l a ted the development a d introduction of the high pressure a n d

n -

high s peed centrifuga l pump Although centrifuga l pump ha ve . s

been u sed for ma ny yea rs they were as a rule employed only for
very l a rge ca pa cities a n d low hea ds During the l as t ten yea rs .
,

however high pressure a n d high speed centrifuga l pumps h a ve co me

,
- -

into genera l use a n d centrifuga l pump design h a s passed through a

s urpri ing sta ge of development
s It i s a s igni fi c a t fa ct th a t lmost
.
,
n a

a ll l a rge m a nufa cturers of pumping m ac hinery a e buil ding c e t if r n r

uga l pumps a n d th a t to d a y centrifuga l pumps compete with recip

,
-

r o c a t in g pi ston pumps for a ll service s This is chiefly on a ccount of .

their lower initi a l cost sma ller spa ce requirements grea ter a da pta
, ,

b ility lower cost a n d e as e of opera tion a d their gre a t a dv a nt age of

,
n

being ca pa ble of direct conne c tion to high speed drivers -

.

An e a min a tion of centrifuga l pump litera ture publi shed in the

x ,

Engli h l a ngu a ge yield little inform tion of v a lue to those inte res ted
s , s a

in this subject The initi al development a n d introduction o f high

.

pressure centrifuga l pumps were due to the energy a n d enterpri se of

Europea n engineers as was the c as e in the development of the ste a m
,

turbine It is quite n a tura l therefore th a t the bes t litera ture upon

.
, ,

this subject is publi shed a bro a d a lthough even there we fin d few ,

complete a n d sa ti fa ctory works the inf orma tion as a rule being

s ,

distribute d in the technic a l m a ga ines z .

The succes of Europea n in ta lla tions of high pressure centrifuga l

s s -

pumps led quickly to the ma nufa cture of such pumps in this country .

Ou r m a nufa ct u re rs adopte d the E urope a n de ign a n d until a short s s

time ago relied wholly upon the experiments a n d da ta fur ished by n

the fore ign builders Experience d resea rch h a ve brought a bout

. an ,

during the l ast fe w ye rs rema rka ble progress in the desi gn of cen
a ,

t rifu g a l pump s both in thi

, s count y a d a bro a d A t the pre sernt n .

time the mos t progressive America n pump ma nufac turers a re pro

iv PREFAC E .

duc in g new designs base d upon their own experience a n d da ta an d ,

nea rly a ll the l a rger pump builders a e developing centrifuga l pumps r

of hi gh e fi c ie n cy a n d grea t relia bility This concentra ted study by .

a l a rge number of a ble enginee rs is producing an immedi a te an d ra pid

g ro wth of the centrifug a l pump in d us try .

The obj e ct of this book is to present to those interested in cen

t ri fug a l p ump m a nuf a cture a n d de s ign an d to te chnic a l student s a

clea r prese nta tion of the funda menta l principles involved a full ,

expla na tion of the c a lcul a tions necessa ry in securing the best e ffi

c ie n c ies an d perform a nce a pro per underst a nding of the co n st ru
, c

tive deta ils of the va rious types the best methods of ma nufa cture ,

a n d a correct k no wledge of the proper proportions of pump p a rt s in

order to secure sa fe a n d smoo th opera tion The a uthors h a ve m a de .

use of m a teri a l furni shed from ma ny so urces M uch of the theo .

re ti c a l prese nt a tion was obt a ined from Di e Zen t ifu l

g p u m pe w ith r a n ,

the kind permission of the a uthor M Frit Neuma n n Dr H ,

r . z . . .

Lorenz s recent public a tion Nw e The ori e un d Ber chn un g der Krei
’
e

selrd i er h as a l so been quite helpful The theory pre se nted by D

g . r
.

Lorenz is b as ed upon the ca lcul a tion of the flow of fluid by me a ns of s

three dimensiona l e qua tions as compa red to the preva lent pra ctice of
usin g two dimensiona l equations A e w a n d welcome point of view . n

is secured by thi s theo y but as yet it h as not been s uffi ciently

r ,

developed to suppl a nt the older a n d more genera lly a ccepted methods

of c a lcul a tion The con structive deta ils of the va rious types an d
.

m a kes of pumps were obta ined in most pa rt from m a nufa cture rs

a n d the technic a l m ag a ine s The methods of ca lcul a ting the
z .

strength of impellers c as in g a d di a phra gm s a n d the critic a l speed

,
s n ,
s

of high s peed pumps h a ve been u sed by the a uthors for a number of

-

yea rs an d a e p a rtly their own work an d pa rtly ba sed upon the

r

re ults obta ined a n d published by other investiga tors

s .

In conclu sion the a uthors ca n but expre ss a hope th a t the p ublic a

tion of this work written during the spa re time of a very a ctive
,

engineering life ma y te nd to ma ke cle a r some of the technic a l an d

,

pra ctica l cons idera tions necessa ry in the de sign an d con truction of s

centrifuga l pump a n d so help to promote in a slight degree the

s, , ,

a pplic a tion of s cienti fi c principles to some of the problem s pre sente d

to engineers of to day -
.

LYNN, M a s s , Ja n uary, 1 9 11 .
 C ON TE N TS .

PL O!

C HAPT E R I .

THE ORY OF CENTRIFUGAL PUM PS .

TH E CENTR I F UG AL PU MP
A IFICATION C ENT IF G AL P M PS
C L SS or R U U

DE I VATION
R THE F U N DAMENTAL EQ UA ION
or T .

F IC ION AN D SH O C
R T L S ES IN C E T IF G AL K O S N R U PUM PS
(a ) Lo sses i n E n tran c e or Suc t i o n Pi pe
(b) Lo sse s i n Impe lle r
(c ) Lo sses in Gui d e V a n es .

(d ) Lo sse s i n Guid e V a n e C asi n g

5 . PR E S E D IFFE ENCE IN
S UR C LEA ANCE S ACE AN D C LEARANCE
R TH E R P

SPA C E LO S E S S .

DIMENSIONS O F THE SU C ION PIPE T

DETERMINATIO N OF THE EX IT D I ENSIONS FOR 6 90 M .

°

I N FL E CE
U THE I MPELLE AN D GU I DE VANE AN G LES U ON THE
N or R P

PE IPHE AL R R

9 . I N FL ENCE
U THE IMPEL E AN D GU
or I D E VANE AN G LES U ON THE
L R P

P ESS U RE D IF FE
R NCE AN D SHORT C IRC IT LOSS THE C LEAR
RE - U or

ANCE SPACE
THE A N G LES 3 AN D 5 AN D THE PERI HERAL V LOCITY A C
0 , P E or EN

IF AL P M P WITHOUT PRESS RE D IFFE ENCE IN THE C L AR

TR OC U U R E

NCE SPA CE
A
DET RMINATIO
E E X IT V AL U ES BY M EANS O F
N O P THE A C U R V ES x AN D

G APHICAL D E TE MINA T ION O F

R E X I T V ELOCITY D IA G RAM
R TH E

M ATHEM ATICAL AN D GRA HICAL D ETE M INATION O F THE E NT NCE P R RA

V E LO ITY DIA G AM WH N 8
C R E ,

THE E N TRANCE AN D E X IT D IMENSIONS WHEN

THE E T ANCE AN D E X I T DIMENSIONS FOR EQ U AL H EA D WI T H V A Y
N R S R

IN G Q ANTIT IES U

C O NST U CT ION OF T HE E N ANCE AN D E X I V ELOCITY D IA G AMS

R TR T R AN D

D ETE MINATION OF SU C ION PIPE V ELOCI Y FOR

R 90 T T
°

THE I N VOLUTES IMPEL E AND GU I D E VANES

F OR L R

(a ) Ge n e ra l C o n s ide ra tio n of t h e Use of t h e In vo lute

( b) T h e In vo lute s at Impe lle r E n tra n c e
(c ) Th e In vo lute s at Im pe lle r E xit .

(d ) Th e In vo lute s at Guid e V an e En tra n c e

TH E G U I D E V ANE C ASI G N

AX IAL HRUST
T .

T HE MU LTI STAG E C NT IFUG AL P MP

-
E R U

THE C EN I F G AL PU M P WI THO U GU I D E V ANES

TR U T
vi C O NTENTS

CH APTE R II .

ONSUM PTIO N OF POWER AND EFFICIE NCY

C .

GENE R AL C O NSI D E R A TION O F C O N S UM PT ION O F PO WE R AN D E FFI C IE NCY

E F FI C IENCY WH E N THE N U MBER OF R E V OL UTIONS IS V ARIE D .

D ECR EASIN G F RICT ION H EA D IN IMPELLER BY D ECR EAS IN G TH E N UMBER

OF BLAD ES
I N C EA E OF H YD A U LIC E FFICI N CY WHEN a B AN D BY US N G
R S R E . , I

DI E TIN G V ANES AT E N ANCE TO IMPELLER

R C TR

C OM PA I O OF THE FRIC TIO H EAD S OF C ENT IF UG AL PUMPS WITH

R S N N R

IM PELL AN GLES B
ER B AN D flu a AN D THE a

I NFL ENCE O F C U V ATU RE OF B LAD ES U PON E FFIC IENCY

U R

27 . T HE E FFI C I N C DETE M INE D BY E X PE IME

E Y AS R R NT .

APTER III CH .

TH E RE GULATIO N AND CLA SSIFI CATIO N OF CENTRIFUGAL PUM PS .

28 R E GU LATIN G
.
Q U AN I TY OF WATER BY T
TH E INFLU EN C E
T H ROT FL IN G .

OF THE QU ANTITY OF WATE U PON H EA D AN D LOSSES D UE R To

T TLIN G H R OT

29 AD U ST ABLE G I D E V A E
. J U N S

30 C LA IFI CATION O F C E T IF UG AL PU M S
. SS N R P

CH APTE R IV .

A ION OF IM PELLER AND GUIDE

TH E C LCUL T A VANE S .

IMPELLE C HAN EL SECTION OF

R O DINARY T YPE
N S TH E R .

(a ) C a lc ula t io n o f T wo H igh press u re C e n t ri fu

-
ga l Pu m p s

( )
b C a lc ula ti o n a n d B la d e Fo rm o f a Lo w pre s s u re C e n t ri u
-
ga l Pum p f .

THE C HANNEL C O S SE IONS AN D DE VELO MENT OF THE BLAD E E N DS

R S CT P

U N THE SU RFACE OF THE C O E

PO N .

( a ) In Ge n e ra l
b
( ) D es c ri pt i o n of t h e Ch a n n e l C ro ss Se c tio n s , w i t h t h e C lcul
a a tio n

an d Des i gn of t h e Bla d es fo r T wo Lo w- press ure C e n t ri fuga l

Pum ps .

PR ESS URE D ISTRIB UTION IN A C ENTRIFUG L P M IN WHICH E NT ANC A U P R E

TO IMPELLE
TH E PA ALLEL THE SHA
R Is N OT R To FT

C H A PT ER V .

DES GN OF I IM PORTANT PUM P PARTS .

34 . BEA IN G R S

35 . PAC IN GSK

36 . IMPELL S ER .

37 . DIAPH A GMS C R ,
AS IN OS , AN D H E D A S

38 . C I CAL SPEE D
RIT S .
 CONTE NTS . vii

CHAPTE R VI .

TYPES OF CENTRIFUGAL PUM PS .

PAGI

H I O IC L R VI W
ST R A E E .

MOD N C ENT I F GAL PU MPS

ER R U

SULz E R P MP U S .

R A A P MP O F SA T
TE U UH ARLE C
S U T E R, o

JAE G E P M PS R U .

K GE G
U PU MPS O F E SCH
L- E LPK E WY C E R, SS O

WO HIN G O N PU MP
RT T S .

P M P O F THE A M T EN U N D M
U S K A G
R A UR AS C B IN E N F AB R I .

I P MO IS
. . RR

Sx O DA P MP U S

ALLI CH ALM P M PS
S- E RS U

WEISE AND M PU MP ON BKI S

B O N JAC K ON PU MPS
YR S

PU MP O F JOHN M C DO UG ALL C AL D ONIAN I ON W O

S TH E S C OMPANY E R RK ,

L MI D I TE

A PU MPS
LB E R GE R .

M ATHER & PLA I '

T

D LAV AL P MPS
E U .

R D WOO D P MP
. . U S .

B RO DN ITz SEY D EL PU MP S .

BE LINE M ASC NENBA U PU MP

R R HI S

W R OBI O N PU MPS
ILL AN S NS

DA ON PUMPYT S

B P M PS
ORS IG U

B FF ALO P MPS
U U

GEBA U E PU MPS R

A LL N PU MP
E S .

N OV E L M I CE ANEO U S D ESI G NS
AN D S LL

A E
C H PT R V II .

TESTING OF CENTR FU GAL PUM PS I .

66 . M ETHODS OF T ESTIN G PUMPS

( ) T h V
a t u i W t M te en r a er e er

(b) T h e We ir Ta n k
(c ) T h e N o z z le .

( d ) Th e Pi to t T ube .
 C E NTR IF UGAL PUM PS

I
. THEORY OF CENTRIFUGAL PUMPS .

ENTRIFUGAL PUMP
1 . TH E C .

A C E N T R I FUGAL pump is the re verse of a wa ter turbine In .

a turbine the kinetic energy of the w a te r flowing t hrough the whee l

is ch a nged into mech a nic a l work W hile in a centrifuga l pump

,

mechan i ca l work is ch a nged into kinetic energy of the wa te r A .

centrifuga l pump us ua lly consists of four ma in pa rts as shown in

Fig 1 .

1The pipe S c a lled the suction pipe when the pum p is a bove
.
,

the level of the supply wa ter a n d ca lled the entra nce pipe when the
,

pump is be low the level of the supply wa ter .

2 The impeller L
.

3 The guide v a nes L whi ch a re s ometim es o mitte d

.
., .

4 The c as ing G
. .

The wa ter flows through a centrifuga l pump from pa rt 1 to pa rt 4 ,

while in a wa te r turbine it flows in the oppo site direction .

As the n a me ce trifug l pump indic a te

n centrifuga l force pla y
a s, s

a n import a nt p a rt The wa te r pa ticles ente r the impeller (se e

. r

Fig 2 ) a t a di a meter D a n d le a ve a t a gre a te r di a meter

.
. The
impeller is rot a ted by a motor a t a n a ngul a r veloc ity ; th e wa te r «1

particles a re thrown outwa rd by centrifuga l force an d thus exert

a press ure or po ss e ss kinetic energy which c an be expressed by the

equa tion

in which repres ent the periphera l velocities a t the di amete rs

a n d 11 .

D a n d D respectively
. . .

The wa ter pa rticles flowi ng through the suction or entra nce pipe
with a velocity w enter the im peller a t the mea n di a mete r D with
‘

., .

a re l a tive velocity v a d lea ve the im peller a t the oute r di a meter D

. n ,
T HE ORY OF CENTR IFUGAL U M PS
P .
 T H E C E NTRIFUGAL PU M P .

w ith a rela tive velocity In order to Obta in di s ch a rge without

,

shock the resulta nt a bsolute exit ve lo c i tyQF of a an d 0 must form

‘
, c .

I a di a gon a l of a pa ra llel ogra m wh os e sides a e y an d This pa ra l

r a

le lo gra m is c a lled the exit vel ocity di a gra m In thi s exit velocity .

di a gra m a an d form an a ngle B which is c a lled the exit a ngle of

c a,

the impeller The a ngle 6 formed by u a d w mus t be tha t of the

. 0 , ,, n ,

guide van e The Objec t of the guide va nes is to reduce gra dua lly
s .
,

with minimum shock losse the a bsolute exit velocity w a d thus

s, , n

tra nsform a much as poss ible of the velocity hea d

s into pressure .

F ig 2
. .

The entra nce velocity di a gra m must be dra wn in a ma nner simil ar

to tha t of the exit velocity di a gra m in order to reduce shock losses
a t entra nce to the impeller The entra nce velocity w must form a
.
.

di a gon a l of a pa ra llelogr a m whose sid es:a e u a d The entra nce

r . n

a ngle of the impeller w ill be des i gna ted by an d the an gle between

R an d w by
. . The entra nce a d exit velocity di a gra ms a re dra wn
n

in Fig 2 Thes e a e simila r to the velocity dia grams drawn for

. . r

wa ter turbines but wi th directions reverse d

, .

Afte r the wa te r flows from the guide v a nes it is gen era lly led by
th e c as ing to the pressure or d is ch arge pipe The velocity a t which
.
4 H
T E OR Y OF C E NT RIFUGAL P M S U P .

the wa ter lea ves the cas ing will be design a ted by w In the follow .

ing dete rmina tion of the e fi c ie n c y Of centrifuga l pumps it must be

remembered th a t only the pa th of the wa ter from entra nce to t h e
s uction pipe to exit from the gui de v an es will be cons idered .

2 . C LA SSI F IC ATIO N OF C E NTR I FU GA L P UMPS .

Centrifug l a p umps
clas s i ed as high or low pressure pumps
a re fi -

a ccording to the m a gn itude Of the press ure or w a ter he a d delivered .

The di viding line be tween thes e two type of p umps i ta ken by mos t s s ,

builders to be a bout 30 to 50 feet hea d The sa me rul es a re em

, .

ployed in c a lc ul a ting both types .

C entrifuga l pumps a a l so cl ass i fied a hori onta l or vertica l

re s z

pump depending upon t h p ition of the h ft The choice Of

s, e os s a .

a rr ngement d p nd up n loc l conditions Fig 3 how a hori

a e e s o a . . s s

z o t a l Sh a ft pump built by Sul er B ros

n of Winterthur Swit er z .
, ,
z

l a nd a d Fig 4 show a v rtica l sh a ft Sul er pump Vertica l

,
n . s e z .

pump a e chiefly u ed in mine h a ft

s r s S s .

There a e a l so the enclo ed a d open types of pump depending

r s n s,

upon whether the guide va nes a re or a e not enclo ed in a c as ing r s .

Ope n type pumps a e only us ed for very sm a ll he a d a d in gener a l

-
r s n , ,

w here l a rge volumes of wa ter a re required They a e con tructed . r s

with vertica l sh a fts ( Fig Axi a l p umps in which the wa ter

.
,

enters an d lea ves axi a lly were formerly use d for thi s service but
, ,
 CLASSI FI CAT IO N OF C E NTRIFUGAL PU P
M S .
5

this method of construction h as bee n entirely a ban doned a n d there

fore will not be considered An enclosed type pump h as its guide
.
-

va nes surrounded by either a spira l or a round

c as ing a s shown in Fi gs 6 an d 7 . .

Di tinction is a ls o m a de be tween pumps with

s

an d w ithout guide v a ne Pumps with gu ide

s .

va nes ha ve a much higher effi ciency tha n those

without guide v a ne a n d the fo rmer should
s,

a l w a ys be u ed when the le as t expenditure of

s

power is a vita l fa ctor It is asto i shing th a t

. n

the importan ce of guide va nes was not recog

n i ed until high press ure centrifuga l pumps
z -

were built It was found th a t the use of

.

guide va ne w ith high press ure pumps pro

s -

du c e d a higher e ffi ciency in s pite of the f a ct ,

th a t they worked under considera bly less

fa vora ble conditions th a n when used in low
pressure pump ; a nd therefore guide va nes
s

ca me into u a lso with lo w pressure pumps

se -
.

The increas ed cost of a pump with guide

va ne is very quickly o flse t by its grea ter
s

effi ciency a d hence sma ller expenditure of

n

power F thi rea on pumps with guide

. or s s

va ne will be principa lly con idered in thi

s s s

book .

The centrifuga l pump h as either a n entran ce

or a suction pipe depending upon the m a nn er
,

in which the w ter is led to it Fig 9 shows

a . .

a n a rr a ngement with an entra nce pipe th a t is , ,

wa te r enteri ng under its own head This ar .

ra ngement is often found with low press ure -

pumps h a ving very sma ll head s The pump .

in thi s c as e is pl a ced below the level of the

Fig 4
supply w a ter d is therefore a l wa y re a dy
. .

an s

for service The ina ccess ibility of thi type of pump is a decided
. s

disa d va nta ge In clea ning a v a lve pl a ced in the entra nce pipe
.
,

must be closed a d the wa ter in an d a bout the pump removed

n .

I most c as es a s uction pipe i s u sed th a t is the pump is inst a lled

n , ,

a bove the supply level ( Fig Thi s h as the a dvan t age of a o

.
6 T HE O RY OF CE NT R IFU GAL U MPS
P .

beca us e the pump c an be quickly d ra ined without an y

c ess ibi lity ,

grea t trouble A d isa dva nt age of the suction pipe type is tha t in
.
-

Le ve l

most cas es the pump must be filled with wa te r before it c a be n

ope ra ted Sometimes it is po ible to lift the w ter to the impeller

. ss a

by speeding up the pump .

 DE RIVAT IO N OF T H E F N U DAM ENT AL EQ UATION . 7

ha ve one impeller while multi st age pumps ha ve two o r more im

,
-

pe lle rs a rr a ged s ucce

n ss ively upon the s a me Sh a ft Fo ins t a nce . r ,

Fig 8 shows a sixS ta ge high press u

. re Sulzer pump
-
.

Fig 8. .

Th e construction of centrifuga l pumps va ries cons idera bly There .

are m a ny di ff erent a rra ngements of both high a n d low press ure -

pumps which will be trea ted l a te r under the hea d of centrifug l

,
a

pumps of v a ri ous typ es .

3 . D E R I V ATI O N O F TH E FU N DAM
EN TAL EQ UATI ON .

Fig 9 s hows s chem a tic a lly a n a rra ngement of a low press ure cen
.
-

t rifuga l p M p h a ving a vertic a l s h a ft In every p art of the

. pump
proper there exists a press ure a bove the a tmosphere due to its ,

depth below the level of the supply wa te r Im a gine sma ll pipe s .

conn ecte d to the ins ide Of the pump a t points 3 e a l a n d d AS , , , ,

.

these points a ll lie in a hori zonta l pl an e there exi ts a t a ll points s

the sa me pressure hea d h when the pump is sta nding still

.l .

When the pump is run a t its norm a l speed the press ure a t ea ch
of these points will va ry a n d the ma gni tude of the press ure hea d
,

a t these points w ill now be inve stiga ted .

The following deriva tion of the so c a lled condition equa tion -

is an old method used for determining the fun d a men t a l equa tion of
wa ter turbines While this method is ra ther lengthy it gives a clea r
.
,

conception of the press ure a n d velocity distribution in the interior

of a centri fuga l pump The pump shown in Fi g 9 is pla ced below
. .

the level o f the supply wa te r in order th a t all th e pressure heads

 O Y
THE R I U A U P
OF C E NTR F G L P M S .
10 T HEORY OF CENT R IF UGA L P UM PS .

individual wa ter pa rticles a mongs t themselves there is lost a fric t i o n

2

hea d At point a a press ure hea d h a d a velocity hea d

2 a
a n
Pa

exist 0 being the rela tive exit velocity The following equa t i
, . . o n

ma y therefore be written

Ita ‘ ‘
l ha i ‘ ‘
l
‘ ‘
P H
55
a
11
2g

At point I which is directly a t entra nce to the guide va nes t h e

, , re
2

exist a pressure hea d h a d a lso a ve locity hea d 3 Due to sh o k

s
; , n c

losse a t entra nce to the guide va nes a friction hea d s u is los t i

s n

passing from a to 1 At point a the pre sure hea d h a d the ve locity

. s a n
"
w
_
hea d 2 exist w being the a bsolute exit velocity Hence
-
,

, . .

g
w.
- _
_
2

hl
w?
l n
2g 2g

order to convert as much as po ible of the velocity hea d

In
3; ss

into pre ure the velocity w Should be gra dua lly reduced in pas ing
ss , s

from point I through the guide v a ne L t w represent the velocity s . e d

a d
n
3
3 1 the
;
velocity he d a t exit from the c as ing a d let d
a be t h

friction hea d expended in c a rrying the wa ter through the guide

,
n n e

passages a d e as in g Thenn .

2 ”
10 1 we!
Pd H n
2g

The first form of the ca lled fund a menta l equa tion is obt ined
so —
a

by a dding equ tion 1 to 5 which giv s equ tion 6 Fo r conve i ence

a s , e a . n

equ a tions 1 to 5 a e here rest ted r a

 DE R IVATION OF T HE F UNDA M E NTAL EQUATION . 11

2a

hd

Fr om Fig . 9 it follows tha t

hd hw H 11

AS lrea dy expl a ined H i the desired hea d a t exit from the c as

a ,
s

ing The friction he a d necessa ry to overcome the los ses in the

.

delivery pipe h as n o t been considered beca use thes e losses m a y ,

vary betwe en wi de limit a n d are dependent upon the length a n d

s
”
w
cross section of the pipe The velocity hea d 2 a t the exi t of the
.
e
—

g
pump is a certa in fra ction of the des ired hea d H This frac tion
c an be repre sented by the coeffi cient so th a t a ,

The coe fficient gives the ma gnitude of the exit losse s an d va ries
a ,

w ith the des ired pre sure he a d w hen the velocity w is const an t
s ax .

Thi w ill be shown more clea rly l ter

s a .

I order to s implify equ tion 6 let

n a

Subs tituting e qua tions 7 8 an d 9 in equa tion 6 the genera l form of

, ,

the funda menta l equa tion is obtai ned ,

—
va
z
+ wa + z
u a
= 2 H 11 ( 1
9 + P+ a )

In order to obtai n the tot l hea d H which i used in the c a lcul a tion
a b, s

of centrifugal pumps a d repre ents the theoretic l hea d which the

n s a

pump c an deliver co s idering no losses w h a tever the de ired hea d

n , s

H c a n be multiplied by the fa ctor 7

. The hydra uli c effi ciency of 1

the pump c a be represented by

n 1 H ence e
77

UH » Hb
12 TH EORY OF CE NT RI FUGA L U MPS
P .

o
Fr m th e entra nce velocity di a gr a m ( g 0)
Fi 1 .

2 uw o . 00 8 5.

Fi g . 10 .

an d from the exit velocity di a gra m ( g 1 1)

Fi .

v} u “
2
w, 2
2 n o w“ c o s 8,

Fig . 11 .

Sub tituting in equa tion 10 the va lue of v3 v a d ( 1 [ + )

s s , 0
“
,
n .
1 a

from equa tions 1 1 12 a n d 13 the second form of the funda menta l

, , ,

equa tion is derived a n d is ,

n o w, c o s 8“ uw . . co s 6. n gH ,

B efore these equa tion s a re tra nsformed a n d other equa tions deter
mined for the ca lcul a tion of the dimens ions of pumps we h a ll in v s es

t ig te by wh a t me a ns high e ffi ciencies ma y be obta ined th a t is how

a , ,

to ma ke ) as sm a ll as possible With a View of obta ining the highest

I .

tota l effi ciency the va rious pa rts of centrifug l pumps will be studi ed a

in the followin g order : 1 the uction or entra nce pipe ; 2 the im

, s ,

peller ; 3 the guide va ne ; 4 the cas ing

,
s , .

4 . FR ICTI O N AN D S HO C K Lo ss Es IN C E NTR I FUGAL PUM PS .

( )
a . Lo sses i n E n tra n ce or Suc ti o n Pi pe .

In equation 1 the friction he d lost by the wa te r in p as s ing from a

the inlet of entra nce or uction pipe to the inlet of the impeller is
s

represe nted by If the pump h a a s uc t ion pipe thi pipe will s s

conta in a non return va lve w hich clo

-
utom a tic lly when the se s a a

pump i stopped Thi v lve is in t ll d to prevent the loss of

s . s a s a e

s uction he ad w hich would occur if the w ter flowed b a c k a d out

,
a n

o f the s uction pipe The resi tan ce to the flo w of wa ter hould be

. s s
 FRI CT ION AN D SH OC K LOSSES IN C E NTR I FUGAL PU M PS .
13

brought to a minimum a n d therefore it i best to m a ke the cross .

s

section of the pipe a mple The velocity of flow through the va lve ’

s ho u ld not be grea ter th a n the mea n suction pipe velocity This .

c an b a ccompli hed by en l a rging the s uction pipe a t the v al ve s e a t

e s .

The device for preventi n g return fl w ma y be of light cons truction o ,

beca use due to the continuity of flo w it rem a i n s ope n when the

, ,

pump is opera ting a n d only clos es every time the pump is brought to
,

rest It is therefore subject to ha rdly a y wea r

. n .

The friction hea d p H is a l so dependent upon the length an d

. ,,

di a meter of the uction pipe as well as upon the s uction pipe velocity
s

w .. With high velocities it is des ira ble to enl a rge the pipe conica lly
downwa rd in order to decreas e the friction lo es a n d to keep the
s ss

s hock losses which occur in the wa te r a t entra nce to the uction

, s

pipe as sm a ll as poss ible

,
.

Th e s uction pipe genera lly termina te s a t the c as ing a round the

center of the impeller through which the sh aft is ca rried A st ufl .

ing bo x mus t be used where the s ha ft pas ses through the cas in g .

Grea t ca re mu t be exerci sed in i ts des i gn in order to prevent on the

s

one ha nd too grea t a friction loa d being pl a ced upon the Sha ft when
the stuffi ng gl a nd is tightened a d on the other ha nd to gu a rd a ga inst n

a i r le a k ag e into the c asing .

( b) . Lo ss es i n Impe ller .

The friction hea d p H is lost by the w ter a t entra nce to the im

. ,. a

peller ch a nnel O a cco unt of the thickness of the impe ller bl de

. n a s

a di minution of cro ss section occurs c a u s ing the s uction pipe velocity ,

w to incre as e s u
. ddenly to a ve
loc ity w. .

Fig 1 2hows the entra nce to

. s

an im peller ch a nnel The bl a de .

thickness i the ch a nnel heigh t

s s .,

is an d the pitch of bl a d s a t
a ., e

entra nce i The diminution of

s
Fi g 1 2
cross ection d ue to the thickness
s
' '

of blades is proport iona l to the ra tio hence as the velocities ,

a re

inverse ly proportiona l to the cross section of flow the a bsolute . en

tran ce velocity w will equa l .

14 E
TH OR Y OF CE NT R IFU GAL PU M PS .

F rom Fig . 12 it is se en tha t

’
t.
a:

hence

At entra nce to the impeller the velocity is suddenly increas ed to

the velocity w which na tura lly is a ccompa nied by shock loss es In
.,
.

order to h a ve the ch a nge of velocity occur gra dua lly the bl a des a re
s h a rpened a n d the length of thi s s h arpening is gener a lly equ a l t o
,

+ a. s.
twice the bl a de thickness It is a dvisa ble in order to h a ve
.
,
as

a s s m al l as po ss ible to de i gn the bl a des th in a d to h a ve as fe w

,
s n

bla des as po s ible thereby increas ing

s ,

The friction losses occurring while the wa ter flows through the
impeller ch a n n els i exp re sed as p H To ma ke thi a minim um
s s a n . s

the impeller sho uld ha ve as fe w ch a nnels as poss ible ; a n d grea t c a re

sho u ld be ta ken a l o to h a ve very smooth cas tings on a cco unt of
s

the na rrow pas sa ges which m a ke m a chining a lmost imposs ible

, .

This is e peci a lly true of high pressure pumps When the impellers
s -
.

a re l a rge bl a des of s heet steel ma y be us ed which c an be ground

,

s mooth before be ing se t in the impeller The friction losses a e con . r

sid e a bly reduced by s uch a con struction

r .

The coeffi cient p is a lso dependent upon the design of the entire
a

bla de cha nn el C a re should be ta ken tha t the rela tive entra nce
.

velocity decreas es gra dua lly to tha t

of the rel tive exit velocity In a 00 .

Fi g 1 3 the pa th ea represent the . s

mea length of the ch a nnel At n . e

the velocity is v a d a t a it is . n

If pos ible the bl a des should be so s ,

constructed tha t the velocity of the

wa ter in p assing from e to a decreas es
linea rly th a t is equa l a mounts in , ,

Fig 13
equ a l dis t a nces a d not as s ho wn by n

the dotted line in the figure where

. .

po or bl a de form h as c a used the velocity to increas e a ga in a fter it

h as st a rte d to decrease .

A p a rt of the friction he a d s is lo st a t exit from th e impe ller by

n
 FR ICTION AN D SH OC K LOSSES IN CENTR I FUGAL U
P M PS .
15

the sudden i creas e of cross se ction If a represents the height of

n .
,,

the cha nnel a n d s the thickn ess of bl ad e a t exit from the impe ller
o ,

then the a bsolu te exit velocity 10 will be decreas e d a fter lea ving the
,

impeller by the va lue a

,
A s m a ller velocity w therefore exi sts a
’

in the clea ra nce spa ce between the impe ller an d guide va nes an d
equa ls

Th e bl a des should be sh arpened as a t entra nce in order th a t the

, ,

ch a nge of velocity from w to w will not occur to o suddenl y

e a
’
.

(c) Lo sses i n Guide Va ne s .

Another portion of the friction hea d n is lost during the pas n

sag e of the wa ter through the cle a ra nce spa ce an d a t entra nce to the

guide va nes As the impeller bla des a e consta ntly moving pas t
. r

the guide va nes there is a const a nt ch a nge of the clea r guide ch a nnel
cross sect i on At a certa in i nsta nt the rot ating impeller is in the
.

Fig 14
. . mg . 15 .

pos ition shown by Fig 14 with res pec t to the guide va nes an d in
.

w hich pos ition no decrease of t h e guide ch a nnel are a occu rs Wh en .

the impeller rea ches th e po s ition hown in Fig 15 the guide ch a nnel
s .

a re a is dec reas ed by the bl a des Of the impeller By sh a rpening the .

bla de ends a d h a ving sufficient clea ra nce spa ce ( Fig 16 ) the Shock
n .

loss es c a n be dimini shed The number of guide an d of i mpeller bl a des

.

should not be equa l in order to a void pe riodic impulses or vi bra tions

, .

These imp ulses are due to the fa ct tha t with a n equa l number of
guide a n d impe ller bl a des the clea r cross sections of a ll guide cha nnels
a re incre as ing or decre as ing in unis on .

In spite of a ll prec a utions it is a lmost im poss ible to prevent the

,

entra nce of foreign m a tte r into the pump It is therefore a dvi a ble . s

to a voi d sm a ll cle a r a n c e spa ces for there is d a nger th a t the foreign

,
16 T HEORY OF CE NT RI FU GAL P U MPS .

ma tter ma y come betwee n the guide a n d impeller bl a des Such an .

occurrence ma y brea k the bla de ends a d fi na lly destroy the pump n ,

whereas suffi cient clea ran ce spa ce might h a ve a ved it As h as s .

bee n sta ted shock losses occur a t entra nce to the guide va nes d ue
, ,

to the decre as e Of cro ss section jus t as a t the entra nce to the i m

,

peller Le t w be the velocity a t entra nce to the g uide cha nnel

.
; ,

a the height of ch an nel a t the beginning of guide v a nes a n d s t h e

; , ;

Fi g . 16 .

thickness of the va nes . No further proof is necessa ry to Show th a t

as before ,

This shock loss ma y be minimi ed by m a king the gui de va nes as

z

thin as po ss ible a d sh a rpe n ing them decidedly a t the entra nce as

n

previous ly des cribe d for the impeller bl a des .

A pa r t of the friction he a d d i lo t by the wa ter in flowing n s s

through the guide va nes The Object of the guide va nes is to cha nge
.

the a b olute e it velocity w into a m ller velocity which sho uld

s x a s a

equa l if possible the velocity of flo w i the c asing surrounding the

, ,
n

guide va nes In other words by the a id of guide va e as much

.
, n s
2
w
as po ss ible of the velocity he a d s hould be ch a nged into press ure
0
.

?2 9
The diminution of velocity must occur gra dua lly The guide .

ch a nnels should be constructed with a ea y divergence of cros n s s

section a n d be of s u ffi cient length a d wh a t was a id rega rding the

, n s

impeller cha nnel s is a lso true of the guide ch a nnels ( e Fig s e .

To decreas e the friction coeffi cient the guide v ne must be ma de a s

very smooth Their proper sh a pe will be taken up la ter

. .
18 TH EORY OF CE NTR I FU GAL PU MPS .

The error ma de in cons idering w w is however sma ll tha t . a

’
, , ,

i
s, the thickness of the bla des is neglected Th is as sumption is .

ma de i order to keep the eq ua tions in as simple a form as possible

n .

The press ure di fle re ce in the clea ra nce spa ce c an therefore be

'

expresse d as
ha kg

From Fig . 9 it c an be seen tha t

ha hd “H3 Pd Hn Pl Ha

F urther ,

ha kw l n

The va lues of h a an d h. s ubstituted in equ a tion 19 g ive

Pc ) +
-

Assume (P a p. ) H

w hich is a pproxima tely true . C on idering equa tion s 9 a d 1 1 an d

s n

s ub stituting the v a lue H n ha from equa tion 7 equa tion 22 ,

bec omes
H op " Ha

Th ewa ter will flow through the clea ra nce pa ce between impeller s

rim a d n i g b ck to the uction pipe

e as n a L t 2 be the velocity
s . e ,

coe ffi cient a n d the velocity Of the wa ter in thi clea ra nce spa ce s .

Then
5 V 2, i
g 7 H , we 2

the a mount of clea ra nce between c as ing a

If i is nd rim of impeller ,

the cleara nce a rea f is a pproxima tely

a

f a Da n i

The qua ntity of wa ter passing through this spa ce is

The substitution of the va lue for w from equa tion 24 in this equa tion “,

gives
 PRESSURE D I FFE RE NC E IN T HE C LEARANCE SPACE . 19

An l thrus t occurs beca use the pressure on the si de Of the

axi a

i mmlle r oppo s ite the uction pipe is grea ter th a n th a t in the pipe
s .

In order to neutra li e this thrust holes co n necting the t wo s pa ces a e

z ,
r

often pro vided in the impeller Such a a rra nge ment of course . n , ,

sa cri fi ces e ffi ciency for the qu a ntity of s hort circuite d wa ter flo wing
,
-

thro ugh the clea ra nce s pa ce to the s uction pipe is doubled or equa ls
2 Qw
By the a rra ngement of suita ble rims or rings the spa ce through
which lea ka ge c a n occur is m a de as s ma ll as po ss ible This de .

cre as es the coeffi cient £ The des ign of such rims wi ll now be
4 .

discusse d .

M a ny pumps are des i gned with a very close clea ra nce between t h e
ro ta ting impeller a n d the s uction pipe connection so th a t the wa ter ,

a fter pass ing through a cle a r a nce s pa ce i a t di a meter D F

( g i 1 7 ) .
, c
.

mus t a lso flow thro ugh a clea r

a nce spa c e i a t di a mete r D
. .

This small cle a ra nce is obta ined

by suita ble rims or rings cast on
the impeller a n d suction pipe
con nection In order to a void
.

a xi al thru s t it i s necessa ry to

ha ve a simil ar clea ra ce between n

impe ller a d casing on the o p

n

p os i te s ide of the impeller a t the

s a me di a mete r D ,

The qua ntity of wa ter short

circuited is determined as fol Fi g 1 7
lows :
. .

Le t f be the outer clea ra nce a re a f the inn er clea ra nce a re a

. , ,

the press ure existing in the pa ce between the rims a t di a meters D

s ,

an d D w , the velocity of fl w a t D w the velocity of flo w a t D

.” o c , .p . ,,

an d £ 5 t h e corre ponding velocity coe ffi cients

0 , , s The following con .

dition equa tion ma y o w be written n :

Further ,
20 TH EORY OF C ENT RIFUGAL P M PS U .

The va lues of w .” an d pl a ced in equa tion 27 give

Substitute for its va lue from equa tion 23 , an d as If, it

follows th t a

This va lue for s ubstituted in equ a t ion 2 9 gives the velocity of

flow exi ting in the inner cle ra nce spa ce d i

s a an s

an d the qua ntity of wa te r flowin g through thi clea ra nce spa ce is s

=
2 " 9H n _w a
2

Qp c f i £i

The Velocity of flo w exist ing in the outer clea ra nce spa ce i therefore s

only
5. 1 11
( 19 .
2

an d the qua ntity flowing through the outer cle ra nce spa ce a is

Qp
c
= fa £a ( 9 71
2 77 11 —
10 0
2
) 1

Hence by using two sma ll clea ra nce spa ce the qua ntity of short s

circuited wa ter c a be cons idera bly decreased

n .

The rota tion of the wat e r between the im peller a d i g a l so n e as n

a i ds in decreasing the sh ort circuit loss Ass uming frictionless

-
.
 PR ESSURE DIFFE RE NC E IN T H E C LE A RAN CE A
SP C E .
21

O pera tion centrifuga l force woul d crea te a rota tin g para boloid h a v
,

2
1
ing a hea d of 5 a t di a meter D On a ccoun t of the friction of the
3
—

wa ter a ga in s t the wa ll s a n d the individua l wa te r pa rt icles a mongst

themselves the wa ter will not rota te with the full periphera l velocity
,

of the im peller ; consequently the rota ting pa ra boloid c a nnot rea ch

its full h ea d A ume on a ccount of friction losses th a t the w a ter
. ss , ,

rotat es with a velocity of onl y qi u an d let 113 be the head of the rota t
a

ing pa ra boloid at di ameter D Then “.

This pressure k is oppo sed to the press ure

g If onl y one cle a r a nce
ring is used a d the rota t ing pa ra boloid i ta ken a ccount of there
n s ,

w ill be a diff erence of press ure in the cleara nce sp a ce of

H arm a

Hence the velocity of wa ter p assing through the cleara nce spa ce is
wt ? s
o
vz g ’l H n
“
wa z ' ”

d
’
z
u a
z

an d the qu a ntity of wa ter i s

Q =
fE 2 g ll Hn ¢ a
u
2
OD a a V ' —
wa ' " z 2

ing two clea ra nce rings a d con idering the rota ting pa ra boloid s
Us n s ,

the qua ntity of wa ter flowing through the clea ra nce spa ces ma y be
obta ine d as follows :
Aga in assume th a t the rota ting p a ra boloid rota tes with a periph
eral velocity of 4m Th en there will be a di fference of pressure bet ween
.

u u? z

a point a t di a meter a d a point a t di a meter D Of ¢

a
2
n ,
2a
The di fference of pressure Hm will o w ha ve only the va lue n

H apa " Ha

As before the condition equa tion ma y be written

, ,

Q 10 .
.
m wc p i f i
22 TH EORY OF CE NT RIFUGAL U
P M PS .

C onsidering equ a tion 39 the velocity w , .” is found to be

H op i

The qua ntity of wa ter expres e d in terms of di a meter s D is, foun d

from equa tions 2 7 a d 40 to be n

Q a E
a
[
2 ga —
w,, 2
1

an d expressed in terms of di a meter D , is found to be

Q '
=
fi £f

In determi ni ng the dime ions of the impeller the qua ntity of ns

Short circuited w a ter mus t be t ken into a ccount bec a u e thi

-
wa te r a s s

pas ses through the impeller It i therefore nece a ry to design the. s ss

impeller for a qua ntity of wa ter Q the va lue of which is ’

,

e = Q+Q

As be seen from these derived equ a tions the short circuited

can ,
-

wa ter depend upon the va lue of the cle a ra nce a re as f a d f

s the a n .

velocity coefficient 5 which decre as e with decreas ing clea ra nce ; an d

,
s ,

if the rot ting pa r boloid i l o cons idered upo the coeffi cient
a a s a s ,
n

The rot a tion of H a d the clea ra nce spa ce losse to the periph
" , n s

er l velocity u a d the a ngles 6 a d 6 wi ll be discussed in Article 9

a . n , n . .

6 . D I M E N SI O N S OF TH E S U CT I O N P I P E .

F rom the desired qua ntity of wa ter Q a n d a as umed s ucti o n n s

pipe velocity w the clea r a rea of cross section F of the uction pipe
. . s

can be determined as follows

 D I M E NSIONS OF T H E S C T U ION PIPE .
23

an d the suction pipe di ameter D . is foun d from

The va lue of the velocity w through the suction pipe depends

. .

upon the des ired qua ntity of wa ter Q a n d the desired head H It
2

is bes t to ta ke the velocity he a d &

1
as a cert a in fra ction 7 of the
2a
tota l head ; H so tha t
I

V2 7 11 9 11 .

The va lue of the coeffi cient 7 is dependent on the to ta l hea d H .

an d the des ired qu a ntity of w a ter With l w pres s ure pum ps h a ving
. o -

lo w hea ds a d l a rge qu a ntities of wa ter 7 is t a ken as

n a n d some

times grea ter w hile with high press ure pumps h a ving l a rge hea ds
,
-

a d s m a ll qu a ntities of wa ter 7 equ a l s

n a n d a t time still less s .

The coeffi cie n t 7 i of grea t im porta nce when it is de ired to com

s s

pa re a pump of e w des i gn with exi ting pumps

n s .

Substituting from equa tion 4 7 the v a lue of w in e qu a tion 44 . ,

Q
_
;_
x/n H 7
.

The va lue K will be c lled the ch a ra cteristic of centrifuga l pum ps

a .

A pump a lrea dy d esi gned c a be u sed for a ll desired he a ds a n d

n

quantities of wa ter for which the expression Q rem a ins the sa me

V aH
,
n

tha t is h as the sa me v a lue of K (see Article

,

Eq u a tion 46 c an a l so be written ,

we n
2 7 12 9 11 3 CH ”

This is the equa tion of a p a ra bol a The suction pipe velocity w will
.
.

therefore increas e for equa l va lues of 7 a d for different desired hea ds n

a ccording to a p ara bol a Fi g 18 s hows such a ve locity p a ra bo l a

. .

for 7
24 TH E O RY OF C E N T R F I UGAL PU M PS .

The a tmo pheric press ure A permits theoretic a lly the us e of a

s

suction hea d h of a bo ut feet so th a t considering the velocity

.. , ,

2
i

s; an d a friction hea d p .. H

?
w“ <
h p0 0 H 11 A
2g

H wo ever the wa ter conta i n s more or less

,

ai a d it i
r, therefore a dvis a ble to ta ke
n s

7 . DE TE R M I N ATI O N O F THE EX IT
D I M E N SI O N S F OR 6 .

The econd form of the funda ment a l

s

equa tion a s lrea dy derived is a

n a wa c o s 6. uw . . co s 6. 17 H
g . ( 15)
a
0 1 2 8 0 6 6 7 0 3 mfl n l l
the a bsolute entra nce velocity w be
Le t .

Fi g 18
t ken a t right a ngle to u with which
a s .,

umption the member uw c o s 6 0

. .

a ss . . . ,

bec a use 6 . so th a t equa tion 15 becomes

uw
. . co s 6. 11
gH .

I
UP

l
1 n
ta n fl,

Fi g . 19 .

From Fig . 19 , w hich repre ent the exit velocity di a gra m of

s s a low ~

s peed pump , 6 . an d
26 THEORY OF CENTRIFUGAL PU M PS .

” 911 3 +

therefore

Then for B .
°
90
“a 90 °

Atte ntion isca lled to the fa ct th a t i the future a ll simil a r va lue

n s

which refer to a velocity dia gra m ha ving 6 90 will h a ve the .

°

s ubs cript

With a given hea d an d a fter assuming the velocity m which c a n be ,

ca lled the ra dia l exit velocity the periphera l velocity c an be deter

,

mined from equa tion 58 for an y impeller a ngle

A function of the guide va ne a ngle 6 ca be s ubs tituted for v in . n ,

equa tion 57 From Fig 19 or 2 0 the following equa tion ma y be

. .

written

0, ta n 6.

Ifthe v a lue of u from equation 59 is substituted in equa tion 6 1

. an d

B.i s t ken
a equ a l to equ a tion 6 1 becomes

Ur t a D 5a 90 ° V 11a va so°

The va lue of v from equ a tion 6 1 pl a ced in equa tion 57 gives

,

rom these equa tion a fter as uming 6 a d 6 the periphera l

F s, s . n

velocity u c a n be ca lcul a ted The va lue v c a then be determined

. . . n

from equ a tion 6 1 .

If the periphera l velocity u is fixed by the outer impeller di a meter

.
 DETER M IN AT ION OF T H E E IT X DI M E NSIONS
. 27

D an d
. the the a ngle 6 can be ca lcul a ted by assumi ng v
. , an d

s olv in g equ a tion 57 for t an thus :

ta n 6 .

or by as s umin g 6 . an d s olving equa tion 63 for ta n 6 we fin d .

t a n 6.
u
.
2

71O H

the periphera l velocity a n d the outer impeller di a meter D

If . a re

gi ven the,revolution s pe minute n a e found

r from the equ a tion r

u 60 .
n
D . 1r

or b y ass uming n an d D. the periphera l velocity is

D. 1rn
60

or by ass uming n an d u
. the impeller di a meter is
u 60 .

M r

The v a lue of the rela tive exit velocity v . is

”r

si n B .

or s ubstituting th e a ngle function of 6 from equa tion 6 1 .

ta n 11
9 .
sin mu .

From Fi gs . 19 an d 20 it ca n be se en tha t the va lue of the a bso lute

e xi t velocity w. is

or it ma y be obta ined from equa tion 52 thus

a
71

14. c o s 6.
 28 T HE R O Y OF C ENTRI F UGAL P U M PS .

the periphera l velocity is not known

If ,
w. is ca lcul a te d from equa
1
tion 6 3 a d 7 2 by ass umi n g flu a d
s n A n s
2
se c
?
6. 1 ta n 2
co s 6.
equa tions 63 a n d 7 2 give
17 9 11 . (1 ta n 2
wa
t a n 6.
ta n 6 .

The impeller width b. is determined when the ra di a l exit velocity

vr is as umed for
s ,

D . 1rb. v, Q

Q
D . 1r v,

in tead of v the a ngle 6 is used then b is found by substituting

If s , . .

the va lue of v from equ tion 6 1 in equ a tion 7 5 with the re ult th t
, a , s a

Qu a

D . 1r ngH ,. t a n 6.

an d then from equa tion 7 5 v , can be found as

t ted before these equ a tions for the e it dimen ions a re va lid
As s a ,
x s

only when 6 th a t i w hen the a b olute entra nce velocity

. s, s

is perpendicula r to u I the follo wi g up to A ticle

.. 1 4 6 i
n t a ken n r . s

equa l to I Article 14 a d ub q u t ch a pter equ tion a

n n s se en s a s re

derived for th d t rmina tion of t h it di m n io for y d ir tio

e e e e ex e s ns an ec n

of the b olut e ntra n e velocity tha t i wh n 6

a s e c ,
s, e .

8 . I N FL U E N CE OF TH E I M P ELLE R AN D GU I DE VAN E AN G LE S U PO N
TH E PE R I P H E RAL V ELO CI TY .

the pr vious cha pter two equa tions were derived for the periph
In e

era l velocity

77 a
 I N FLUE NC E OF I M PELLER AN D G U IDE VAN E ANGLES . 29

T h e firs t equa tion can be ca lled the eq ua tion of periphera l veloci

ti e s fo r simil a r impe ller profiles for a fter as s uming the ra di a l exit ,

vel o c ity v the outer impe ller profile i determined for a y given
. s n

d i a m e ter D I this equa tion u i depe ndent upon 6 a n d v

.. n F . s . , . or

I) , t h e a ngle 6 from eq ua tion 6 2 ma y be substituted SO tha t u is

. 90
°
, .

a l o dependent upon 6 a d 6
s The a ngle 6 va ries with u
. n . . . .

E q u a tion II is the eq u tion of periphera l velocitie for impe lle rs

a s

o f d i ss i mil a r pro files Fo w ith co ta nt a ngle 6 a d the sa me

. r, ns . n

i m p eller di a met r D the impeller width b a d hence the impeller

e . ., n

p r o fi le ch a
, nge wi th v a ryi ng u I th i eq u a tion is dependent . . n s 11 .

up o n 6 a d . n

E q u a tion I is the more usef u l bec a use for a certa in profile the ,

p e ri pher a l velocity c an be determined by as uming while w ith s

e q u a tion II thi s c a not be done directly

n .

B ec a u e of its grea ter usefu

s lne equa tion I will be more cl os ely ss

e x a mined a n d the cours e of the curve for peripher a l velociti es as

d erived from thi equa tion will be inves tiga ted

s .

There a e three ch r cteristic points for the u curve as giv n by

r a a . e

e qu tion 58 a

1 When 6. 0 u will equ a l 0 th a t is the u curve passe through

. , . , , . s

the origin of the co ordin te y tem '

a s s .

2 When 6 . for which a gle u

. the curve become n .
so
, s

as y mptotic with a ngle of lope of 90 '

°
n a s

3 When 6 9 0 the c urve h a point of inflection an d t h

°
.
. as , e

a ngle of the t a ngent a t thi point i s s

or subs tituting the value Of v from equa tion 6 2

, , ,

tan 90
°
V ngH .

The subta ngent S a t the in flection point i s

or substituting for v the a ngle function of 6 r .

S 2 cota n 6 . 90
°
80 T HEORY OF CE NT RI FU GAL U P
P M S .

ua tion 59 shows tha t the peripheral velocity is indepe nden t of

Eq
v or 6 when 6
, . hence the ordi n a te a t the po int of in flection
.

h as the sa me val ue for every or 6 provided Of course e q u a l 0, . , ,

hea ds a e assumed Three u curves an d their ta nge nts a t the po in t

r .
.

of inflection a re dra wn in Fig 2 1 for the ass umpti on .

Fo r vario us ass ume d va lues of 6

lcula tion gives . ca

I 6 5 . 0 . 90 fo oto
4 23
°
, ;
°

II 6 15 . v . feet
90
° 53
°
,
°

III 6 25 . v . 90 feet °
66
°
,
°

The ta ngents a t the point of i flection clea ly Show the di rec t ion n r

of the u c uves a t tha t point ( Fig

. r .

It c a n be se e n from equ tions 78 a n d 7 9 th a t the t a nge nt of thea

a ngle 6 i directly propo rtio a l to v a n d ta n 6

s H ence the a ngl en , .

increas e or decreas es with v a n d ta n 6 respecti vely

s . . 90
° .

While the u curves a re conc a ve towa rds the base line from 6
. 0 .

9 0 they a re convex from 6 9 0 to 6 H ence wi th

° °
t 6
o . . .

incre as ing or 6 be tween 6 90 an d 6 0 the pe riphe ra l °

0, .
°
90 . . ,

velocity decreas es a n d from 6 90 to 6 180 it incre as e s

, .
°
.
°
.

I high press ure centrifuga l pumps we a re forced to m a ke t h e

n -

velocity v a n d therefore the a ngle sma ll in orde r to get the im

,,

peller width b as l a rge as poss ible With a s ma ll guide va ne a ngle

. .

5 to as is us ua l in high press ure pumps it c a be se e n

°
6.
-
, n

from the course of the u curves th a t a increa e or decreas e of th e . n s

impeller a ngle 6 h as very little influence upon the periphera l velocity

. ,

a t le as t within the u eful limit s s .

In lo w pre ure pumps the guide va ne a ngle 6 va ries wi dely up t o

-
ss .

If a high peed pu mp is to be designed the a ngle 6 should be

-
s , .

ta ken as l a rge as po s ible compared to the a ngle s

9 . I NFL U E N CE OF T H E I M PELLE R AN D GU I DE VAN E AN GLES UPO N

THE PRE SS URE DI FFE RE N CE AN D S H O RT CI R C U IT -

LO SS O F TH E C LE AR A N CE S PA CE .

rom equa tion 2 3 the press ure difference in the clea ra nce spa ce
F , ,

without con sidering the rota ting pa ra bo loid is ,

Hm " H “
 I NFLUENC E OF I M PE LLE R AN D G UIDE VAN E ANGLES
. 31

Subs tituting the va lue for 10 from e qua .

tion 73 , the a bove equa tion becomes

1 ta n 6.2

The va lue of the press ure d ifle re n ce is

therefore dependent upo n 6 a n d in
.

creas ing with grea ter va lues Of 6 a n d 6 . .

an d vi ce versa To show this rel a tion

.

clea rly th ree press ure difference curves are

,

dra wn in Fig 22 assuming as before

.
,

= 10 feet These curves a re for 6

.
.
o
90

a n d a re m a rked H

Fig 2 1
. .

It h as lrea dy been hown (equa tion 35) th a t the decrea e of

a s s

press ure difference due to the rota ting pa ra boloid is a pproxim a tely
32 T H E ORY OF CE NTR IFUGAL P U M PS.

Assume th a t the rota ting pa ra boloid revolves with a velocity

Fig 22
. .

ons idering these va lues a second series of pressure di fference

C ,

curves a re dra wn in Fig 2 2 by subtra cting the va lue of g from the

.
i
g
curves first foun d Therefore for
.
5 thes e curves a re in a ccord
a nce with the e qu a tion
34 THEOR Y OF C E NTRIFUGAL P U M PS .

With l a rger guide a ngle s uch as is used for low pressure c e n

a -

t ri fug a l pump s the qu a ntity of s,

hort circuited w a ter v a rie gre a tly -
s ,

depending upon whether the impeller is low or high peed F o s . r

in s ta nce if the qua ntity of short circuited wa te r is equa l to

,
fo
-
r

2 5 wi th 6 then it is only when 6

° °
6. . 45 an d f o . r

13 5 it is
°
6.

When investiga ting the influence of the short circuit loss up o n -

pump efficiency it is necessa ry to expre the qu ntity of Sh o t

, ss a r

circuited wa t r in per cent of the tota l qua ntity desired The l a g r

e . r e

the ra tio of outer di a meter D to qua ntity of wa ter desired t h e . ,

l a rger will be the qua ntity short circuite d when express ed in per c e n t-

of the qua ntity des ired This percent ge for the a me v lues o f . a ,
s a

” H .” a d ni inver s ely propor t ion l to the qu a ntity of w a t e r

s a

de ired If for insta nce a pump delivering 100 ga llons h as a sh o rt

s .
, ,

circuit loss of 3 per cent thereof th a t i a hort circuited qua nt i ty , s, s -

of 3 g llons then a pump with the sa me clea ra nce a d va lues o f

a ,
n

D . , an d a d delivering only 50 ga llons h as a l so a short cir c u

n it ,
-

loss of 3 ga llons ; still it is no longer 3 per cent but 3 per cent X ,

6 per cent of the qu a ntity des ired .

2
The ra tio is genera lly sma ller with low press ure th a n with hi gh
9
-

Q
press ure pump a d therefore the short ircuit loss referred to t h e
s, n -
c ,

qu a ntity delivered is grea ter with high press ure th a n with lo w

,
-

press ure pumps .

It is for thi re as on th a t the cle a a nce p a ce in high press u re

s r s -

pumps mu t be brought to a minimum Thi i be t ccomplished

s . s s s a ,

as a lre a dy des cribed by the s c a lled clea r nce rim or ring T h e

, o -
a .

rim is often grooved a n d tongue d (Fig This a rra ngemen t .

,

Fig 2 3 . . Fig 2 4 . .

however does not decrease the loss to a n y grea t extent for the
, ,

direction of the velocity of the short circuite d wa ter is not axi a l -

 TH E PE RI PHER AL VE LOC IT Y OF A C E N TR F I U GAL PU M P .
35

but combines with the periphera l velocity in producing a res ultan t

velocity the va lue of which is ( Fig 24 ) .

w “
W
The direction of this velocity forms the a ngle 6 with u ., so tha t

A veryclever device for decreas ing the clea ran ce pa ce is shown in s

the Germa n pa tent 142 2 14 Fig 25 shows this device which con . .
,

s ists of a loose pa cking ring D c a rried

by the impe ller Segments of a ring R
.

a e so pl a ced th a t their centrifuga l

r

force press es the ring D a ga inst the

c as ing a n d thus reduces the clea ra nce
s p a ce to a minimum I Germa n . n

pa tent 1 17 2 18 the clea ra nce rim is

,

s hown a djus t a ble which in it elf is , s

very good but thi construction grea tly

,
s
Fi g 2 5
increas es the cost of the pump .
. .

Unf ortun a tely there a e no publis hed repo rts of experiments

,
r

determining the a ctua l qua ntity of hort circuite d wa ter C a l s - .

cu la ti o n s give only very ppro i m a te v a lue b c a u se the coe ffi cient

a x s, e

of fl w is not e a ctly known a n d beca us e the clea ra n e i very difli

o x c s

cult to meas ure .

The hort circuited wa ter ma y be t a ken as from 4 to 6 per cent

s -

of the des ired qua ntity a n d the impeller ca lc ul a te d for a corre

spondingly l a rger qu a ntity .

10 . TH E AN G LES 6 AN D 6 AND THE PER I PH E RAL VELO CITY

. . OF A

C EN TR I F UGA L PU M P W I TH O U T PR E SS U R E D I FFE R E N CE
IN TH E C LE A R A N CE S PA CE .

In the following for simplicity the oppo ing e fi e t of the rota ting
, , s c

para bo loid on the press ure di ff erence in the cle a ra nce spa ce wi ll
not be cons idered Accordi n g to equ a tion 2 3 the pressure difleren ce
.

wi ll be

H0 9 ” Ha
36 TH E ORY O F C E NT R F I UGAL PU M PS .

If H .. is ta ken eq ua l to 0 thi equa tion becomes

,
s

2
19
The velocity hea d 2 represents the enti re theo retic l press ure
3 a
-

hea d The a ction of centrifuga l force h as been utili ed to a ccelera te

. z

the rel a tive entra nce velocity v to the rela tive e it velocity The
. x

a bso lute exit velocity w mus t be tr a ns formed entirely into press ure
. .

Fi g 2 6
. .

F rom Fig 26. it c an be seen tha t

Substituting the va lue of 0 from equa tion 6 1 in the a bove gives

,

Solving for u ., we h a ve

Thi equ a tion c a n a lso be eas ily derived from the fi st funda menta l
s r

equ tion When the b olute entra nce velocity w i perpendicul a r

a . a s . s

to u equa tion 10 as sumes the form

.,

2 11 911 .

Then substituting w .
x/2 n gH . from equa tion 90 we h a ve
“a = v¢
 DETERM INAT ION OF TH E E X IT VALUES . 37

H ence for the speci al cas e of H

,
0 the exit velocity di a gra m will ” , ,

be a rhombus (Fig From this fi gure the former equ a tion c a n

.

be directly obta ined ,

F rom the rhombus

v. 0.

Si n 6 . sin 2 6.

To derive the periphera l velocity for which H 0 c a lcul a te from ,

equa tion 9 1 the a ngle 6 a d then from equa tion 96 h the corresponding
. n

periphera l velocity u ..

11 . D E TE R M I N ATIO N O F TH E I T VA L U E S
EX BY M EA N S OF

x AN D A C UR V E S .

The periphera l velocity u. from equa tion 6 3 is

u x/” H”
ta n 6.

I
I
a

so th a t the a bove equa tion ta ke the simple fo s r m

u . x V n gH .

On Pl a te I a serie of curve a hown ; the va lue of a e c a l

s x s re s s x r

cula t d for v riou v lues of the a ngl 6 a d

e a s B y the a id of the e
a es . n s

curves v a riou problems c a be quickly olved

s n s .

F r inst a nce a fter determining t h impeller di meter a d p

o , ib e a n re s c r

ing a certa in r p m o th a t the pe riphera l velocity u i s 7 5112

. . . s . 1 ,
88 TH EORY OF CE NT R I FUGAL P U MPS .

the coefficient K ma y be determined by ca lcul a ting the necessa ry

pe ri phera l velocity u an d solvi ng the equa tion
.

rom the curves it is possible to rea d directly which combina tions

F x

of 6 a n d 6 a re permiss ible for the des ired periphera l v elocity or by

. . ,

ch a nging the a ngles 6 an d 6 whether it is poss ible to obt a in the

. .

desired r p m . . .

I genera l as h as a lre ad y been sta ted it i s a dv a nt ageous to assume

n , ,

the ra di a l exit velocity v instea d of or if 6 is as sumed the value

, . ,

of v should immedi a te ly be determined ; for only a fte r determining

.

v a n d ass uming 6 c a
. the impeller width b be determined from
. n .

equa tion 7 5 In order quickly to determine v for a y given guide

. , n

va ne a ngle or vice versa the A curves of Pla te I ha ve been drawn

,
.

These curves were derived in the following ma nner

Eq ua tion 6 2 is

v, t an 6. °
90 V n gH a

Le t the velo city head be A times the theoretica l hea d " H ," so

tha t
V A2 ngH .

S ubstituting this va lue of v in e qua tion 6 2 we h a ve ,

t a n 6. °
90

A ta n ?

The A curve which a ccording to equa tion 102 i a pa ra bol a h as

,
s ,

be en c a lcula ted for va riou va lues of a ngle F rom thi curve

s s

plotted for 6 a fter ass uming

. the corres ponding v lue o f A a

ca be rea d a n d then v determined by solving equ tion 100 The

n r a .

coe fficient A is independent of the desired hea d as hown by , s

equ a tion 102 ; hence the A curve may be u e d for all desired s

heads .
 DETERM I NAT IO N OF T HE E XIT VALUES 39

If v. is given , A is ca lc ula ted from the equ a tion

an d the a ngle 6 corres ponding to the coe fficient A c a n be determi ned

.

from the curve .

Fo r dete rmining the v a lue of v when 6 2 9 0 the curves must °

. . 1:

a l o be us ed
s Substituting in equ a tion 6 1 the v a lue of u from equ a
. .

tion 98 an d the value of ta 6 from 101 the va lue of v is n . , ,

or sol ving for A ,

Fo r 6. x equa l
equa tions 104 an d 105 ta ke the form
s 1, an d

of equa tion 100 a d 103 res pe ctively

s n .

The exit velocity v is a g in c a lcula ted by as suming

r or 6 by a .

ass uming before expl ined but o w the coeffi cient must be
as a , n K

ta ken into con ider tion s a .

A very imple equ tion c a n be derived for the a bsolute exi t ve lo c

s a

ity w by mea of the a d A curve as follow

. , ns x n s s

Sub titute the v lue of v from e q ua tio 1 04 a n d u from 98 in

s a , n .

equa tion 7 1 which bec o mas

,

A very simple cons truction of the velocity di a gra m a t exit is n o w

pos ible I equa tion 54 ubstitute the va lue u from equa tion 9 8
s . n s .

so th a t the hori ont a l com ponent of w is

z .

The t hree following va lue which a easy to c lcul a te s, re a , a re us ed

in con tructing the exit velocity di a gra m
s
40 T HEOR Y OF CE NT R F G L P I U A U MPS .

In Fig 2 7 . ab is dra wn equa l to u . x V an d a line para l lel

to u . at a dista nce v . La y Off on ab the dista n c e
V 7I9H
dra w 30 perpendic ul a r to b C o necting c with
K
a
an d a . n a

a d b a d dr a wi ng a d p a ra llel to be the p a r a llelogra m

n n b d gives t h e ,
a c

desired exit velocity di a gra m .

k a

Fig 2 7 . .

When the di a gra m derived in this ma nner it is well to check

is ,

the a ngles 6 a n d 6 in order to ma ke sure th a t no mista ke in c a l e u

. .

l tion h as occurred By a s u ming w perpendicul a r to u tha t is

a . s . ., ,

6. a n d by the id of the a d A curve

a every pump c an be x n s,

c a lcul a ted without the u of ta ble of trigonometric functions

se a .

While grea t exa ctness is not Obta ined by use of the curves it must ,

be considered tha t ; i more or less a rbitra ry The use of these

1 s .

curve will how th a t they

s s u fficiently exa ct a re s .

I order to determine the be t dimension of a pump it is impo r

n s s ,

ta nt to rea lize tha t a fter obt a ining t h e curves p rmit a choice of x e

combina tions of 6 a d 6 w hich is very va lu ble in obta ining the

. n . a

mo t a d va nt a geous dimen ions of a pump

s s .

12 . GRAPH IC AL DETE RM I N ATI O N IT VELO CITY D IA G RAM OF THE EX .

The following gra phica l method of de termining the exit velocity

di a gra m a n d thereby u c a be use d for hi gh pre sure pumps
. n -
s

beca u e the error inherent i this method is very sm all for s uch
s n

c as e s .
42 T HEORY OF C E NT R IFUGAL P U M PS .

Fo r the specia l c as es in which 6 . 45

°
an d 6 .

13 . M ATH EM ATIC AL AN D GR AP H IC A L D ETE RM I N ATI O N O F THE

EN TRAN CE VE LO CITY DIAG RAM W H E N 6.

After uming the impeller di a meter D which extends t o the
ass .,

center of gra vity of the a b (s e Fig the periphera l velocity

rc . e .

u c a n be dete rmined for the periphera l

. ,

velocities a e directly proportiona l to r

the cor es po nding dia met or r e rs ,

D .

D.

The l ength of the entran ce a c b r .

mu t be chose n tha t D b F s so .1r

Fig 29 _
.

.
w hich i s equ a l to the cle a r cro ss

ectiona l a rea of the uction pipe

s s

directly t the e ntra nce to the impeller The va lue of the a bso lute
a .

entra nce veloc ity h as a lre dy bee n given as a

in which denotes the height of ch a nn el a n d s the thickness of

a. .

bl ad e .

Ta king into a ccount the ho t circuit loss the impeller must be

s r -
,

c a lcul a ted for a qu a ntity Q Q The velocity w in th

’
. e

suction pipe is increas ed to a gre a ter velocity w a t entra nce to the .

’

impeller due to the a mount of w a ter short circuited ; hence

,
-

Q+Q. .

It istherefore nece
.
ss a ry to substitute for w. in equa tion 16 the
velocity so th a t
 DET ERM I N AT ION OF E NTRA NC E VELOC IT Y DI AGRAM . 43

Af ter ass um ing

the numbe r of impeller bl ades z., the pitc h t. at

the in n er dia me te r D of the impeller is .

The bl ades a t their in n er end mus t form an a ngle 6 . with u ., so

th a t (see Fi g 12 )
. .

the a bsolute entra nce velocity is to be ca lcul a ted from equa

If
'

a 8
tion 1 1 1 a n ass umption of the ra tio a must be ma de When
,
. .
.

90 the rel a tive entra nce velocity v is dete rmined from the
°
6. .

equa tion
0. V u ”
. w. ’

an d the impeller a ngle 6 from th e eq ua tion .

The va lue of can a. s. n ow be be t determined gra phica lly from a

s

c on s idera tion of equ a tion 1 13 .

a. + 8.
If incorrect as umpt ion of the ra tio
an s w as m a de the c a l
0 .

cul a tion must be repea ted This method of c a lcul a tion n e ce ssita t
.

3.
in g the ass umption of i s ve y cumbers ome By the a id of r .

equa tions 1 1 1 1 13 an d 1 15 a might be c alcul a te d but the equa tion

, , . ,

would be of a complica ted form .

A very s imple gra phic a l method for determining w a d . n a. s.

w ill n o w be shown This met hod is due to Ad m wh a pplied it in

. a , o

a s imil a r form for the determin a tion of the e xit velocity di a gra m of

a w a ter turbine ( D ingl P lyt h Jour a l . o ec n . n ,

Eq ua tion 1 1 1 an be written c
44 T HEORY OF CE NT R I FU GAL PU MPS .

The gra phical determi n a tion of w a n d a 3 c an be se en from . . .

Fi g 30 A right a ngled tri a ngle doe was firs t d r wn whose s ides

. .
-
a ,

d o u a n d cc w ; the di ta nce f was then ad ded equ a l to the

. . s c

bla de thicknes s A lin e p a ra llel to de was dra wn through f Th e

s. .
.

Fi g 30
. .

a rcha ving t as a ra dius an d struck from the point d as a center cuts

.

the p ra llel line a t point 9 From 9 a line is dra wn perpendicul a r to

a .

d cutting the l a t ter a t h a n d cutting dc a t i

c, The extens ion of do
. .

mee ts the ext en ion of bf a t Then i h

s 12
9 s an d ac (I . a ., .,

H ence equ a tio n 1 1 3 i s sa tis fied bec a use ,

!g
a. 8.

do

an d a s l o equ a tion 1 16 is sa ti sfied bec ause

a. + s.

0.

The p a ra llelogra m abcd is then the desired entra nce velocity

dia gra m .

14 . T H E EN TRAN CE AN D EX IT DI M EN SI O N S W H E N

the equa tions a lrea dy derived it h as been as sumed t ha t th e

In
ab olute entra nce velocity w i perpendic ul a r to u a d hence th a t
s . s . n

the ngle 6 a With this as s umption the member uw o s 6

. 0 . . c .

e
( q u a tion 1 5 ) a n d therefore it h a d no e fle c t upon the peripher a l
velocity It often occurs however for reas on to be di cu ed l a te r
.
, , s s ss ,

th a t the a ngle 6 is not for in t a nce it i genera lly sm ller th a n

. s ,
s a

90 for high pre ss ure centrifug l pumps The following discuss ion
°
-
a .

will expl a in the method of ca lcul ating exit an d entra nce dimens io n
When 6 2 .
 T HE ENTRAN C E AN D E X IT DI M ENSIONS WHE N 90 °
45

B egin ning aga in with the se c ond form of the funda menta l equa tion ,

we ha ve
uw
. . 00 8 6. uw c c 00 8 50 7] H
O ;

P g
: ” r

“a ma d
Fi g 3 1
. .

F rom the exit velocity di a gra m Fi

( g 3 1 ) . it c a n be seen tha t
as w. co s 6. u+ .

Fi g 32. .

Fro m the entra nce velocity di a gram Fi

( g 3 2 ).

f k w. co s 6.

The calc ula tion of the entra nce a ea of the immlle r is bas ed upon
r

the vertica l component of the a bsolute entra nce velocity w which .,

will be c a lled w ( ra di a l entra n ce velocity )

r If w a n d 6 a e given w .
. . r , r

is ca lc ula ted from the equa tion

w. Sin 6.

or if we ass ume w . an d eq ua tion 1 19 ma y be written

46 T HEORY OF CE NTR IFU GAL U P
P M S .

The dista nce fk ma y be expresse d as

so tha t combin ing equa tions 1 18 an d 12 1 gives

w. c os 6.

F rom equa tion 109

S ubstituting equa tions 1 17 122 an d 109 in equa tion , , 15 , we obta in ,

so lving for the periphera l velocity u .,

D w . v.
+ " 9H n 123
D 2 ta n 6.
a 2 t a n 6. tan 6. 2 ta n

If F. is the decreas ed entra nce a re a due to thickness of bl a des an d

F. the exit a rea of the impeller then ,

Substituting this va lue of v. in equ a tion 1 23 the periphera l velocity

w. D. F.
(
1
2 D t a n 0.
a Fa t a n fin

De Fe 2

/
”
1

(
w. 1
H
l T D.t an 6. F. t a n
g .

The determina tion of the peri phera l velocity by me s of this equa an

D F
tion is r ther cumbersome beca u e the ra tios D d 17 must firSt be
a ,
s
?
an —
6

as sumed If the an gle 6 devi a tes but slightly from

. . only a very
s m a ll e rror is introduced by u ing the s impler equ a tion 58 or 63 s .

Fo l arge devi a tion of the a ngle 6 from

r equa tion 12 5 a n d the .

following must be used for the determina ti n of u o . .

In pl a ce of v in equ a tion 1 2 3 a function of the ngle 6 c a n be

. , a .

s ubs tituted Sub stituting the v a lue for w co s 6 from equ tion 12 2
. . . a
 T H E E NTR N E AN D E A C XIT DI M E NS IO NS W H E N a? 90
°
47

an d the v a lue for u from equ a tion . 1 09 in equa tion 15 we obta in for
w cos 6
. .as the va lue

H
ng . D w . ,
w. cos 6“
u . D tan 6.
.

O
Fr m Fig 3 1 . it follows th a t

11 gH . D
_ w . .

u . D tan 6 . .

7] a D w c r
v, ta n
u . D t an 6.
.

F urther ,

fla D w c r ”r

u. D. ta n 6 . ta n 6 .

After ub tituting the va lue for v from equa tion

s s . 128 in equa tion 1 29
the periphera l velocity u will be .

ta n 6
-

ta n fi

Fo r the pecia l cas e of

s 6 . 90
°
equ tion a 12 3 or 1 30 ta kes the
form
+ V(D .

u tituting in thi equa tion 6 9 0 we a ga in obta in the simpli fied

S bs s .
°

equa tion u
’

v . gH
. . .

The curves (Pl a te 1) ma y be used in order to determine quickly

K

the express ion 1 i equ tion 130 With the a id of the

n a . x

curves it is Simpler to determine the periphera l velocity from equa

tion 130 th a n from equ tion 12 5 a .

Fink designed a impeller bl de h a ving the rel a tive entra nce

n a

velocity v ra di a l tha t i v perpendicul a r to u a d 6

. ,
s, . With . n .

this as s umption the periphera l velocity an d a ngle 6 a found as . re

follows :
48 T HEORY OF CENT R I FU GAL P U M PS .

°
Fo r 6 . 90
w, D.
ta n 0. Da
i
This va lue for “ f; a if subs tituted in equa tion 129 will give the pe riph
era l velocity for the ass umption 6 . 90
°
as

.
90

The a ngle 6 . is then found from equa tion 13 2 to be

w.D.
tan
“a b .

The ques tion n o w a rises Wha t is ga ined by m a king the a ngl , e

6 2
. Ass uming 6 the a bs olute entra nce velocity w i s
. .

perpendicul a r to u an d the wa ter will flo w through the suction pipe

.,

in a a xi a l direction a d therefore will tra ver e the shortest pa th

n n s .

If the a ngle 6 2 the wa ter in the suction pipe c a n no longe r

.

flow in a n a xi a l direction but will be se t in rota tion by the co m,

ponent w c o 6 a d tra ve se a pira l pa th Fo thi reas on not onl y

. s . n r s s . r s

the suction pipe velocity but a lso the pa th described by the wa te r i s

increas e d B oth of these occurrences ca use grea ter friction losses In
. .

order t minimi e thes e lo e it i a dvisa ble to pl a ce directing va ne s

o z ss s s

a t entra nce to the impeller The Object of the e va nes is to ke p . s e

the flo w axia l in the suction pipe d further to give the wa ter a t an , ,

impe ller entra nce its correct b olute entr nce velocity a n d di e c a s a r

tion The e directing v ne ha ve therefore a n a ngle of 90 a t e n

. s a s
°

tra nce a d pproxim tely the a ngle 6 a t exit They decrease t h e

n a a . .

losse in the uction pipe con ider bly in spite f the fa ct th a t friction
s s s a O

a d hock lo ss es occur due to the w a ter p as s ing throu gh them

n s , Ex .

p e ri m t s h
ena ve proven th t the l tter lo ss e e m a l ler th a nathe a s ar s

losses in the suction tube when the directing va ne a o mitted s re .

While on the one h a nd the lo es up to the ntr nce of the impeller ss e a

a re l a rger w hen 6 9 0 as compa red with 6 on the other

°
. .

ha nd a ma ller loss occu i the fi rst pa rt of the impeller cha nnels

s rs n ,

beca use by decreas ing 6 the ch a nn el height is i creas ed there by

. a. n ,
50 TH EORY OF C ENT RI FUGAL P U M PS .

di rection of the bla de i rever ed a n d woul d a ga in dec re a se

s s a. 3. .

Such a bl a de would not of cour e be u ed for then the a bs olu t es s ,

entr nce velocity would be come to o grea t

a .

Fi g 33
.

The b t va l ue of the ngl

es w h n it is m a de less th a n 9 0 for
a e e
°

t h r as on j ut t t d wi ll o w be determin d
e e s s s a e ,
n e .

The friction losse s a re a p

pro im t ly propo tiona l t x a e r o

th qua re of the velocity a d e s n

w ill b come a mi nimum a t e

e i t from the di ecting va nes x r

l d in t h b ginning f the an e e O

1 imp ll r ch a nnel when t h e e e

F ig 3 4
e xpr ion w} v i a min e ss 2
. s

imum Thi occurs w hen

. .

. s

v. w When v. . w the a gle 6 . Fig 34 Sho ws a n

. n . .

entr nce velocity di a gra m in whi h 6

a Here c .

”
f
uD . .
w. c o s 5
2 2 D.
 E N T RA N C E AN D E X IT DI M E NSIONS .
51
Ur
Substituting in equ tion a 15 the va lue of w. co s 6. u .
ta n 6.
eq ua tion 1 17 a d for w c o s 6 its va lue from equa tion
n . . 1 35, we fin d
the periphera l velocity for the specia l ca se w v or . . 6. 6 to be
.

If 6 . the equa tion ta kes the form

no

When 6 hence w. 6 . v., the following va lues a re found for

the exa mple quoted a bove
Fo r 6 .
u . feet ,
w. v. 13 2
2 feet , 6 . 6. 48
°

Fo r 6 .
u . feet ,
w. v. feet , 6 . 6. 45
°

In Fig 33 the e va lues a d lso the ma gnitude of a

. s be
n a . s. c a n

rea d directly beca u e the ordina tes must pass through the point of
,
s

intersection of the w c urve with the corresponding v cur ve

. . .

15 . T H E ENTR AN CE AN D IT DIM EN SIO N S FO R

EX EQ UA L HE ADS WITH
VARY IN G QUANTITIE S .

In the foregoing a rt icle it was assumed th a t it was necessa ry

s

to cons truct a w impeller for ea ch hea d a d qua ntity of wa ter

ne n

d es ired The as s umption will o w be m de th a t an impeller with

. n a

known entra nce a d exit a ngles 6 an d 6 exists If this impeller

n . . .

is to deliver v a riou s qu a ntities of w ter a t the s ame hea d it is e c e a n s

sa ry to inves tiga te w h a t the entra nce a d exi t velocity di a gr m s n a

s ho u l d be for oper tion W ithout hocka It will be proved th t one s . a

a n d the sa me impeller m y be us ed for v a ry i ng qu a ntities a t cons ta nt

a
52 TH EO RY OF C E NT RI FU GAL PU MPS .

hea d by cha nging the guide va nes a t exit an d the directing va n es

a t entra nce provided the l tter a e us ed In pra ctice this is ofte n
a r .

of grea t va lue .

The second form of the funda menta l equa tion from which it is
a ga in be st to st a rt s t a tes th a t

uw. . c os 6. uw . . co s 6. n gH . ( )
1 5

It h as been found from the exit velocity di agra m ( see Fig 3 2 ) h a

. t t
v
w. eo e 6. u+
.
tan 6 .

The entra nce velocity di a gra m (see Fig 32 ). gives

therefore the equa tion for w . c os 6. ma y be writte n

w
w. co s 6. u
.
t a n 5.

ubstituting in equa tion 15 the va lues of

S w. co s 6. an d w. co s 6.
from equation 1 17 an d 1 39 we obta in
s

( u‘
tai ls . )
Fo r u an d
. ub titute their va lues as given in equa tions
v, s s 1 09 an d

124 . Then the periphera l velocity c n be expre ed as a ss

w.
F. t a n 6 . D tan 6 . .

The sam e impeller pro file was umed therefore the a ngl es 6
as s , . an d

6 . an d a l so the v a lue F F D s ., . , an d D a consta nt Hence

. re .
 E NTRAN CE AN D E X IT DI M ENSIONS .

F. tan B. Da t an fic

an d then equa tion 14 1 ta kes the very imple form of s

rom this equa tion a fte r determining C an d C the periphera l

F , . s,

vel o c ities c a be ca lcul a ted for va rious velocities w hence for va rious
n .,

qua ntities of wa ter .

The periphera l velocities for the two special cas es 6 90 an d .

°

6. 6 .a er de ived asr follows

It has been s hown th a t uw c os 6 when 6 90 an d tha t °
; gH . . . 1 . .

w. co s 6. u .

F rom eq ua tion 1 24

Fo r 6 .

“ 0 . u ta n 6
. . ra
gi ta n 6 .

This value of w substituted in equa tion

. 1 24 gives
D F
. .

D F.
.

l cing this va lue of v in equa tion

Pa , 1 17, the va lue of the periphera l
veloc ity for the ca se 6 90 is .
°

“a °
s. 90

To determine the qua ntity of wa ter for the speci a l c e 6 as .

D
so lve for u in the equ a tion u d from equ a tion 145
1
an fin d
.
D .
54 T HEORY OF CE NT R IFU GAL U
P M PS .

the velocity w The qua ntity of wa ter delivered is then determi ned
, .

from the equ ation Q F w . . .

Fo 6 r 6 or w . . .

w. co s 6.

rom Fig 34 which repre ents the entr nce velocity dia gra m for
F .
,
s a

the sp ci a l cas e 6
e w h ve . e a

Therefore
ta n

This va lue of v . is s ubstituted in equa tion 1 17 , g iving

u D F
. . . ta n 6 .

2 D F ta n 6

Afte rubstituting equa tions 135 a d

s n 1 49 in equa tion 15 the
periphera l velocity for the cas e 6 6 is . .

77 a
1 D. F. tan fl. 1 D 2

( )
.

2 D. F . t a n 6 . 2 D ..

The velocity w is d t rmined from equa tion 148 a d the qua n tity
. e e n

of wa ter delivered from the equ tion Q F w AS the entra nce a d a . .. n

exit cros section must rema in cons t nt the rel tive entra nce a n d
s s a , a

e it velociti mut v ry in dir ct proportio t t he qua ntity of

x es s a e n o

wa ter The rel a tiv ntr nce v locity i found from the equa tio n
. e e a e s

w.
Sin 6 .

and the rel tive e it velocity from the eq u tion

a x a

vr

si n 6 .

With ch a ge of w a d v the a ngle 6 a d 6 a l o ch a nge a d in

a n . n , s . n . s , n

the a me m a nner the guide v ne cro ection t entr ce a d t h e

s a ss s a an n

directing va ne cross ection a t exit (when directing v ne exi st ) w ill

s a s

a l o va y
s r .
 E NT RAN C E AN D E X IT DI M ENSIONS . 55

The a ngle 6 . c an be found a fter determini ng u (Fig . .

tan 6.
v

t an 6.

an d the a ngle 6 from the equation

. ( g
Fi 3 2 ) .

w.
t a n 6.
w
tan 6 .

Fo r the a bsolute exit velocity w ., Fi g 3 1 . gives

)
2

(
v.
3+ u+ .
ta n 6

an d from Fig . 32 the a bsolute entra nce velocity is

2

(
w,
m
tan

most cas es the determina tion of these v a lues ma thema tic a lly is
In
unnecess a ry a d it i a dvisa ble to dete rmine them gr phic lly
n s a a .

The equa tion just derived do not how the rel a tion b t wee n the
s S e

va riou va lue of the velocity di a gra m a d the qu a ntity of wa ter

s s s n .

I order to m a ke thi rel tio

n clea r evera l ex mples h ve been
s a n s a a

worked out a d the res ult are hown gra phic a lly on Pl a te II
n sThe s .

velocitie c a lcul a ted from the derived equa tion s a d velocity

s a re ,
n

hea d a plotted as ordina tes to the qua ntities of wa ter I thi

s re . n s

work we us ed the first form of the funda ment a l equ a tion ,

fi +
t
2
"
2 2
w. 2
v.
2
u 2

2a 29
i
29
u
H . +
29

To Show the influence of the impeller a ngle 6 on thes e va lue the . s

thre e following exa mpl h a ve be en c a lcul a ted es

I.
6 . II .
6 . III .
6 .

Th e sa me as s umptions w ere mad e as in the previous exa mple ,

na mely
x/ n . ft ; . hence ,
17 11 . ft ; D. . ft ; D. . ft .

Fo r v, fe et a nd w, feet let 6 .
56 T HEO RY OF C E NT R IFUGAL P M S U P .

With the va rious impeller a ngles the consta nt C will v a ry 1

a ccording to equ a tion 142 The a ngle 6 must be determi ned b y

. .

finding the periphera l velocity from equ a tion 58 for the s peci al c as e
w
an d then 6 is obt a i ned from the equa tion t a n 6
.
6
7
. . .
1 .

The consta nt C is independent of the a ngle 6 a n d therefore h as

. .

the sa me va lue in a ll three exa mpl es .

Fo r the v a rious qua ntiti es of wa ter the pe ripher a l velocities a re

determined a ccording to equa tion 144 .

Th e qua ntity of wa ter for 6 90 was t a ken equ a l to u n ity in

°
.

ea ch exa mple an d the va rious velocities dete rmined for the qua n
tities of wa ter between the limi ts a d These velocity head s n

as found from equa tion 1 0 for the three e x a mples 6 6 . .

13 5 a re dra wn as ordina tes to the corres pondi ng qua n

°
a d 6
n .

tities of wa te r in Figs 1 2 a n d 3 of Pl a te II an d the poin ts found a re

.
, ,

connecte d by curves In Figs 4 5 an d 6 the velocitie s h a ve be en

. .
, ,

dra wn as ordina tes to the qu a ntities of wa ter .

These curve show clea rly the rel a tion between the va rious val ues
s

a n d the qu a ntity of wa ter Fo r genera l des i gn the qua ntities of

.

wa te r between an d a e mo st u seful Fo r qu a ntities of wa te r

r .

grea ter tha n the rel a tive as well as the a bsolute velocities a n d
therefore the friction los e become too l a rge The intersection of
s s .

the w an d v curves of Figs 4 5 an d 6 gives the specia l c ase w v

. . .
, , . .

an d therefore 6 .

2
g
1
The course of the 2 curves in Figs 1 2 a n d 3 is of Speci a l inte r
? .
, ,

est for neglecting the rot a ting p a ra boloid the pressu

, ,
re d ifferen ce ,

in the clea ra nce spa ce a ccording to equa tion 2 3 is H ) H 1

The va lue of the existing pressure difle en c e in the clea ra nce spa ce r

c a n be seen directly from thes e fi gu res It is evident th a t for .

6. 1 35 the
°
pre ss ure d i fle re ce in the cle a ra nce sp a ce decre as e s
'

very ra pi dly wi th the qua ntity of wa te r while for 6 45 it rem ai ns , .

°

a pproxi m a te ly cons ta nt for v a rious qu a ntities of w a te r Fo r re a .

so s which will be more fully cons idered l a ter the an gle 6 is a lmost
n , .

a lwa ys l a rger th a n 9 0 in high pressure centrifuga l pumps When

°
-
.

6. a d if the a ngle 6
n the pressure di fference in t h e
.

clea ra nce spa ce an d therefore the short c ircuit lo sses will be c o n -

si de ra bly dec re as ed Thi a dva ntage which i n its elf is quite va l

. s ,

ua ble is in a ddition enh a n ced by th e fa ct th a t with 6

, , ,
90 t h e .
°
58 THEO R Y OF CE NT R I FU GAL PUM PS .

If a protra ctor is not at h nd the value

a
ta ? 6
is determined an d

tl’r
la id o ff a long the line b u ( Fig 35) so tha t ac t a 6 From a . .

ri

e erect a line w perpendicul a r to b co nect the poin ts

cc .
,a d a , n a n

c a d the line
n Thi l a tter method of con tructi g th
ac s s n e

entrance velocity d ia gr m i pref ra ble beca use of the exa ctne i a s e ss n

constructing the a ngle

Fig 3 5 . .

Whichevermethod i u ed it is be t to determine the entra nc s s s e

height gra phica lly After as uming the number of bl a de

a. s. . s s

a. 8.
2. the pitch t i determined d con idering th a t si 6
. s
t
,
an s n .
.

(equa tion the v lue of a be det rmined gra phica llya a. s. c n e

a s hown in Fig 35 The uction pipe velocity a be dete rmined

s . . s c n

a.
a na lyt c lly from the r l t on w
i a u r more imply by e a i . .
a.
o s

the gra phic a l method if (a ft r determi ing in Fig 35 where

,
e n a. s. .

a . hi d ant h li d i i dra wn a d prolonged until it cu ts

s. e ne s n

i
wr at 7
0 . Th en f k w.
’
b e c a use 8 “
6 n
hence it" a.

wr ac 8.

The uction pipe v loci ty w i then d termined a ccording to

s e . s e

eq u ti n 1 10 whi h t t th t w
a o
Q c s a es a
Q Q
, .

I t h det rmin tion of t h p riph r l v lo i t y a cording to eq u a

n e e a e e e a e c ,
c

tion 136 for the p i l c v w h nc 6 6 t h only as umps ec a a se . ., e e . e s

D:
tion necessa ry is th t Of t h ra tio D ther fore i thi as e the r d i al
a e e n s c a

e tra nce velocity w

n b t k i d p d tly of t h r di l exit
. can e a en n e en en e a a

velo ity c Th uc tio pi p v lo ity w w be a umed as

e s n e e c . can no ss

before d ft r d t rmini g the periphera l velocity u t h velocity

,
an a e e e n ., e
 A
E N T R NC E AN D EX T I V E LOC ITY DI AG RAM S .

w. an d the va lue a. 8 . ca n eas ily be determined gra phica lly as

follows
u
D ra w the right a ngled tri a ngle whi ch e

( Fi g . 36 ) -
dflc in df -

2
an d f k let kc s be the ass umed bl a de thicknes s With pitch
w.
’
. . a

t. as a rad ius de cribe n , a bout I whi h cut a line dr a wn throu gh

s a a re ( c s

" a

e pa ra llel to dk a t the point g . P rolo until it cut the prolon

ng dg s

ga t i o of
n f l a t point
c d a an la y off b u p ra llel to df A lin
a . a . e

“
dra wn perpendicul r to b a a at a di ta nce from equa l to 5 cuts
s a
°

the prolonga tion of If t c so th a t the p r ll logra m b d i the

( a , a a e a c s

de ir d ntr ce di a gra m The p rpendicul a r gh give the v lue

s e e an . e s a

a. d f the v lue of the r di l entr nce velocity

s. a n c a I the a a a n

di gra m thu found the condition ful filled a r

a s s e

Fig 37
. .

The determina tion of the e it velocity di a gr m ( Fig 37 ) i i mil r x a . s s a

to th t Of the ntr nce v locity di gr m Aft r l ul t i g t h

a e a e a a . e ca c a n e

p eripher a l velocity u l y ff the gle 6 w i

., t h u a b At o an . . a . a
 T HEORY OF CE NT R IFU GAL U
P M PS .

distan ce v perpendicul a r to u dra w a line p a rallel to u cutting t h e

. . .

l ine forming the a ngle 6 a t point d The p a ra llelogra m a bcd is . .

the desired exit velocity di a gra m .

Without a protector the di a gra m c an be cons tructed gra phica lly

”
by ma ki ng be tan a n d l a yi ng it o ff if positive on the prolonga
a

, ,
6 .

tion of u to the right of b or if nega tive to the left of point b

. ab , ,

( Fi g . The perpendicul a r to a b from c meets the line dra wn

pa ra llel to b from c a t di ta nce v so tha t the pa ra llelogra mabcd
a a s .,

i a ga in the des ired exit velocity di a gra m

s .

After ass uming the n u m ber of bl a des z t is dete rmined ; an d then ., .

8 a.
considering the equa tion i 6 t
we fi d gra phic a lly a s n .
.
n .
.

The e it velocity di a gra m was determined for a point j ust before

x

the exit from the impeller hence the veloc ity refe s to tha t velocity ,
r

exi ting in the contra cte d e it a rea F due to the thi c kness of t he
s x .

bl a des To be exa ct a velocity

.
,

Go

a. 8.

must be us ed for the ca lcul a tion of the a ct ua l exit a rea F D b .

’
. 1r . .

With bl a de of l a rge pitch it i po ible to ass ume for simplicity

s s ss

without introduci g grea t error th t v Wit h s ma ll pitc h a n d

n a .

correspo ndingly sma ller exit height it is bes t to determine t h e a ., .

velocity v an d then ca lcula te the impeller Width b from the equa

.
’
.

tion (see equa tion 7 5)

17 . TH E I NV OL UTE S FOR I MPELLE R AND GU ID E VA N E S .

( )
a . Ge n era l Co n si de ra ti o n o f the Us e f
o the Ia vo la te .

The fi t genera l funda me t l equa tion ( qua tion 10) shows tha t
rs n a e

the theory of centrifug l pump depend m inl y on the fa ct th a t a s s a

the w t r in flowing through t h impeller a n d guide va nes is forc ed

a e e

to assume definite velocity by me ns of the form given to th e

a . a

bl de ch a nne l The d irection of the wa ter is determined by t h e

a s .

entr n e d exit velocity di gr m The equ a t ions ubsequentl y

a c an a a s . s

deriv d determined t h ma gnitude of th e e velocities an d wit h

e e s
 T HE I NVO LUTES FOR I MPE LLE R AN D G UIDE VA NES .
61

w h a t a ngles the impeller an d guide va nes mus t be cons tructe d in

order to give the desired hea d an d quan tity of wa ter for cert a in a

number of revolutions I order to ful fill th e e equ a tion the first

. n s s,

condition is to give the impell er a d guide van e ch a nnel such a

'

n s

form tha t the prope r velocities an d their directions ma y be correctly

o bt a ined .

A gre a t mist a ke in the cons truction of centrifuga l pumps which

c a n be found to d a y in m a ny des igns i th a t the cross section s de ‘
-
s .

r ived from the c a lcul a tion of the e formul as is not a dhered to in the s

c ons tru ction of the bl ad e The re ult i th a t the supposed veloc ity
s . s s

di a gra ms do not e i st a n d the pump will not give wh a t is desired

x

a t the correct number of revolutions Des igners then determine .

expe rimenta lly the n umber of revolutions necess a ry to gi ve the

des ired he d d then cl a im th a t the equa tion dete rmined by theory
a an s

do not give correct resul ts, C hi fly by the id of experiments the e a

num ber of revolutions for a certa in qu a ntity a n d hea d was deter

)

mined an d one pump a ft r the other was cons tructe d until a seri es
, e

was obt a ined w hich in some m a n er ful filled the des ired conditions n .

It is to be regrette d th a t s uch incorrect des igns a e foun d to d a y r -

in pra ctice a d in text books Such designs will o w be discuss ed

n . n
‘
.

From equ a tion of form imil r to tho e lre a dy derived entr nce
s s a s a , a

an d e it velocity di a gra m
x determined The chief fa ct in s a re . o rs

dete mining the form of the bl a de a the impeller a ngles 6 a d 6

r s re . n .

an d the gu ide a ngle After determini ng the outer a d in er n n

di am eter D a d D the de ign of the impeller bl de was very

s . n . s a s

s imply a ccompli hed by dr a wing the a gles 6

s d 6 a t the di a meters n . an .

D a d D a d connecting the l i nes formi n g these a ngle by an arc

. n . n s

of as l a rge a ra dius as po ssible Recent text book even specify wh a t . s

the a ngle between point of entra nce n d exit should be in order to a

O bt in a fa vor ble connecting a c

a a r .

B l a de of thi s fa l e form a re Sho n dotted in Fig 38 com

s
w s B y . .

pa ri son with bl a de of the correct form hown by the hea vy full

s s

lines of the sa me fi gure it wi ll be Shown wh y the e fa lse forms are s

i ncorrect .

In the des ign of the bl a des a d bl de Ch a nnel the following con n a s

di tio n s mus t be strictly a dhered to in gener l a .

The bl a de ch a nnels must be form d a t beginning d end th t e an so a

the entra nce an d e it reas obt in d fter det rmining the rel a tive
x a , a e a e

velocities a ctually e i t in the de ign Further th a t th e direction

,
x s s .
,
62 THE O RY O F C E NT R FUG L PU I A M PS .

of the wa ter velociti as determined by the a gle 6 a n d ar es, n s . e

a t right a ngle to the entra nce an d exit areas The follo wing equ
s . a

tion mu t be sa tisfied
s

z a. b. v. z a. b. v.

where b a d b represent the entra nce a d e it a reas of ea c h

a. . n a. . n x

bl a de ch a nnel a d th ir numb r The cha n el heightn d

2 e e . n s a. an a.

mu t be t ri ght a ngl to bo th neighboring bl a de while t the m

s a es s a sa e

time a voiding contra ction or e l rgement of the tre a m N t

a n a s . o

Fig . 38 .

u til thi
n will the direction of t h w ter velocity be a t righ t
s is so e a

angl t v ry poi t of the

es o e ntra nce a d it a re
e n The rel tiv e n ex as . a e

v lo i t i
e c um d i t h
e s as s lc ul a t i will the a ct u lly be obt i ned
e n e ca ons n a a

a d th
n um d veloci ty di gr m will be correctly formed
e ass e a a s .

I or d r t f u
n l fill the dition th bl a de t entra nce a d e i t
o e se c o n s e s a n x

m u t f rm q ui d i t t curv
s o e t di t c d
s an from one an other es a a s an e a. an a.

an d form t h g l 6 a d 6 w ith
e an t h p riph ry f the
es circle
. wh n . e e e O s o se

di a m t D dD
e e rs a re A i volute will b t ful fill th se condition /
. an . . n n es e s.

In Fi g 3 8 t h bl d . r pr t d by he vy olid lines a e dra w

e a es e e se n e a s r n

as inv lu t t nt
o c a d i t Th cr
es a e tion of the bl d
ra n e n ex . e o ss se c a e

ch a nn l t th poi t
e s a p o d e t
e se ly t th giv n in t h e
n s c o rre s n s xa c o o se e

c lcul ti
a C m p ri g t h h vy a d do tt d bl d
a on s . o it an cle ly
a n e ea n e a es, c ar

be seen th t the l a tter h v i orr ct entr ce cro ections t h e

a a e nc e an ss s —
 TH E I N VO LUTES FO R I M PE LLE R AN D G UID E VA NES .
63

entrance cross ection to o la rge the exit cross section too sma ll
s , .

With uch cros se ctions rel a tive velocities other tha n those given in
s s

t he di a gr a m will occur bec a use it i impo ss ible for the c a lcul a ted
, s

di gra m to form A number of revolutions quite different from

a s .

the c lc ul a ted number is nece sa ry to give the de ired hea d a n d

a s s

qu a ntity of w at er .

Ex a ctly t h sa me error is m a de in the de ign of the guide v a nes

e s

as i m a de in the des ign of the impeller bl a de It i bes t to h ve

s s . s a

the beginning of the guide v nes form involutes a .

It will be ho wn in the follo wing h w to form the in volutes a t

s o

impeller exi t a n d entra nce a d in the guide va nes n .

( )
b . The In vo lutes at I mpe lle r En tra n ce .

entran ce wi thout shock the bl a des at entra nce must form an

Fo r
a ngle 6 with The circle who e di a meter is D an d to which the
. a .. s .,

entr nce dia gra m refers should pass through the middle of the
a ,

entra nce height It wi ll be sho wn l a ter th a t di a gra ms a t v a rious

a. .

point a cross the entr nce height va ry The mea n di a gra m upon
s a a. .

which the c a lcul a tion i b a ed will be ta ken a t the middle of s To s a. .

a void a contr a ctio or enl rgement of the trea m the bl ad e height

n a s ,

a mu t be t ke t right ngle to the neighbori ng bl ad e a d the

. s a n a a s s n

bl de m ust form a a gle 6 to the periphery a t the di a meter D

a n n . ..

The bl a de a t entra nce will w be con tructed a an involute a d no s s n

it pplica bility investiga ted

s a .

If i the number of impeller bl a des the di a mete r d of the b as e

z. s , .

circle of the involute i found from the equa tion s

implicity ass ume th a t the sides of the impeller so fa as the

Fo r s ,
r

bl des go rema in pa ra llel th a t i the entra nce width b is equa l to

a , ,
s, .

the exit width b After uming D a d the entran ce velocity

. . as s . n z.

di a gra m a d the va lue of n ca be found The method of a. s. n .

cons tructing the entra nce involute w ill now be shown .

The di meter of the bas e circle i dete mined a ccording to equ a tion
a s r

16 0 I order to con truct two ch a nnel s step Off the h a lf pitch

. n s ,

i
Six time on the circle who se di a meter is D ( Fig Through
s
s . .

2
64 TH EO RY OF C E NT RI FU GAL PU M PS .

the points thus found a d a lterna tely des igna ted by 1 an d 2 dra w
n

ta ngents to the bas e circle On both sides of poin ts I lay 0 6 a long th e

.

£ 3 9
ta ngent the dista nces 3an d 2 s ; an d on both Sides of 2 la y o ff t h e
3
.

8
dista nce 2
Fo r ea ch bl a de six points a, b, c and d, e, f are foun d .

The portion a bc of the involute ma y be repl a ced by a n arc of a circ l e

F ig . 39 .

w ho e ce ter lies a ppro im tely the i ters ection of the two t a n

s n x a on n

ge t dra wn through the t wo neighboring points 1 on the bas e

n s .

circle In order to be ure of guiding the wa ter correctly it is a d

. s

vi ble to prolong the involute beyond c a bo ut ha lf an inch The

sa .

pa rt d f is ma de pa ra llel to b when heet meta l bl a des a re use d b ut

e a c s -
,

wi th c as t bl a des it is give a form s uit a ble to the s h a pe of the bl a d e

n

cha nnel .

It c a n be se e n from Fig 39 a d prob a bly needs no fu

.
,
nrther proo f ,
66 TH E ORY OF C E NT R F G L PU I U A M PS .

w., an d w. . the middle di a gr m the rel a tive entr nce velocity

Fo r a a

at point 1 can be expressed a s

wr
Si n 6 .

and S imil a rly for the point c ,

w ith corre ponding indice s s ,

w. ,

s in 6 .

Divi ding the e t w equa tions

s o and t a ki g into a ccount equ tio
n a ns

16 5 an d 1 6 6 we h a ve
,

1 hence v.,

Simil rly for poi t d w h a ve v

a v n e ., ..

It i th refore proven tha t by us ing

s e involute for the beginn i n g an

of a bl de the rel tive entra nce velocity v rema ins con s ta nt a cro ss
a a .

the entire ch a nnel height a ..

( ) c . Th e In vo lute s at Impe lle r Exi t .

The h a pe of the impe ller bl a de a t exit wel l as a t entra nce is

s as

prefer bly mad e to conform in outline to the involute Le t a be

a .
.

the bl a d ch nn l h ight the bl a de thickne a d the numb

e a e e , 3. ss , n 2. er

of bl de a t impeller exit The di a meter of the bas e circle for t he

a s .

involute will then be

or from Fig 40, .

d. D. Si n 6 .

The v a lue a. s. is ca lcul t d from the equa tion

a e

m+ a

The co truction of the involute t the e it i exa ctly the sa me a

ns a x s s

a t the entr nce N O further proof i ne d e d th t the rel a tive exi t

a . s e a

velocity v rem a i the me cro the entire width if the involute

. ns sa a ss

is u ed or th a t th g ner l equ a tio

s ,
e e a n

D
’
si n 6 . .

s in D.
holds .
 TH E I NVO LUTES FO R I M PELLER AN D GUID E VA NES . 67

The velocity di a grams v ry a cross the entire exit height a n d the

a

va riou va lues of w
s an d v a r Obt a ined from equ a tion 1 7 0
., , e .

Only the e it velocity di a gra m c lcul a ted for the pump w ill a ti sfy
x a s

the equa tion d rived for the periphera l velocity All other di a gr m
s e . a s

a cro s t h
s bl de ch nnel hei ght give a incorr ct periph ra l v l
e a a a. n e e e oc

ity for the ctu lly exi ting ngle 6 d 6 b c a u e the re ulting
a a s a s . an e s s

ca lcul ted periphera l v locities increa e as the diam eter decreas es

a e s ,

which is of cou e impo ss ible rs .

F ro m eq u a tion 17 0 a d a l o Fig 40 it i een th a t the a ngle 6

n s . s s .

become m ll r as the di a met r decre

s s a e The a ngle 6 will however
e a se s .
. , ,

increas e towa d the center of the impeller in spite of the fa ct tha t the
r s

a ngle 6 decreas e to ward the ce ter a n d vice ver towa rds the
. s s n , sa

outside The ta ngent of ngle 6 i found from equa tion 153

. a . s ,

t a n 6.

Th is equa tion does not give a good ide of the decreas e d i a an n

crea e f 6 bec u e the right h d member cont i tw v i bl

s O a s -
an a ns o ar a es .

I order to obt in
n cle r ide of th r l tio e i ti g b t w 6
a a a a e e a n x s n e e en .

an d 6 with th corr ponding di m ete s we h ll olv the e it

. e es a r s a s e x
68 TH EORY OF CE NTRI FU GAL P U MPS .

velocity di a grams for va rious dia mete rs ta king a concrete exa mple ,
.

Fo re as on to be di cussed l a ter the periphera l velocity is c a lc u

r s s l a ted
for the di a meter D a t .

Le t V 7 H
)g .
= feet an d the di a mete r D a t 5 foo t .
9 5 equ a l .

Fo r this di meter let the impeller a ngle 6 135 a n d the guide a ngle
a .
°

°
6. 25
rom the curves ( Pl a te I) for 6 135 a d 6
F x so .
°
n . x

tha t for the assumption 6 9 0 u x fee t .

°
. .

From the A curves for 6 the ra d i a l exit velocity °

25 a n d . x

v. feet .

A s uming 12 bl a des a t impe ller exit equ a tion 1 6 9 gives a

s s , . .

inches Le t . inch ; then a inches Fro m

s. . .

equa tion 16 7 or 168 the di a meter of the base circle of the involute
i fo u
s nd to be d inche . s .

The involute h a ve been dra wn in Fi g 40 The outer dia mete r

s . . .

D .
“
fe t is found gra phic a lly as is a lso the di a meter D
e , .
’

foot a t the beginni ng of the involute .

The va lues found for the three points a m an d i are given in t h e , ,

following ta ble .

the l t column Of thi ta ble the peripher l velocities given a re

In as s a

derived from equ tion 58 a d r pr nt the v a lue nece a ry in ord

a ,
n e e se s ss er

th t no hock lo
a s h ll occur with the a ctu lly existing a ngles
ss e s s a a

6. a d n It c be e n th t th e c lcul a
anted p ripher a l velociti
s e es a es a e

v ry con ider bly from the a ctu a l e i ting velocitie which a re given
a s a x s s

in the fir t column s .

I order to neutr a li e thi error in the di a gr m as much as poss i

n z s a s

ble it i nece ry in c lcul ting the peripher l velocity for the d i a

s ssa a a a

gram not to t a ke the outer but to t ke the mea di a meter D If a n . .

a pump is d igned with a di a gra m b d on the outer di a meter the

es as e

p e ripher a l velocity of the impeller w i ll be too m a ll an d in ord e r t o s ,

 THE I NV O L U TE S FOR I M PE LLER AN D G U I D E V AN ES . 69

obta in the des ired hea d a n d qua ntity of w a ter it will be necessa ry .

to increase the number of revolutions beyond those for w hich it

was des igned .

Attention is c a lled to the fa ct th a t the a bove is equ a lly true i n

the des i gn of wa ter t urbines In the des ign of these wheel s for both .

lo w an d high s peed the error of b as ing the velocity di a gra m upon

s

the out r di ameter D i sometimes m de B ra ke experiments

e .
“
s a .

upo n such fa lse designs h ve hown th a t the turbine developed its a s

grea tes t power a t a number of revolutions quite di fferent from the

‘

ca lcul a ted number Pfa rr of Da rm ta dt Germa y in h is l ectures

. s ,
n ,

on hydr ulic ma chinery ca lled a tte ntion to this mista ke

a , .

It is u ua l to h a ve the outer di a meter D a whole number bec a use it

“
s .

i s a dimen ion used in the hops

s F this reas on the oute r dia meter
s . or

must be as umed an d the mea n di meter c a lcul a te d a fter as suming

s a

the outer impe ller a ngle The mea n di a meter is fou d ma th e ma t n

i a lly a fter as s u ming the outer di a meter D a d the oute r impeller “

c , . n

an gle from the equa tion (see Fig 40) .

D D D.
“
.
2
. a.
2
2 a. c os ( 80
1
4 4 4

D a V D a
d,
00
2
2 “up o
n
COS B a
a

The a ngle 6 i then determined from equ tion 1 7 0 If the mea n

. s a .

dia meter D i given the out r dia met r D i found from (se e Fig 40)
. s , e e .
a
s .

D D 2
s in
2 2

)
. . 6.
4 4

The oute r an d inner dia meters however c a n be found more , ,

quickly gra phica lly .

If the outer di a meter D a n d the pe riphera l velocity u a re given

“
“
. . ,

the l a tte r being derived a fter as uming the number of revolutions s ,

then the exit involutes c a be cons tructed a fte r ass uming n a d n

z. .C a l cul a te the di a mete r of the b as e circle from equ a tion 16 8 .

G ra phic a lly or from equ a tion 1 7 1 determine the me a n di meter a

D a d from equ a tion 1 7 0 fi d the a ngle

. n Fo r the c as e 6 n .

“
is dete rmined from the equa tion From the curve a
x K x s
V 1:a
of Pla te I read the angle 6. co re ponding
r s to the derived ngle a
70 T HEORY OF C E N T RIFUGAL PU M PS .

By of the A curve fi d the corresponding ra di a l xit velocity

a id s n e

The impeller width b i d termin d from equa tion 1 58 If the width

. s e e .

b i not uit ble for the impeller pro file it i necessa ry to m ke a

. s s a s a

more uit a ble a sumption of the a gle

s s n

If the me n di a meter is gi v n the out r di a meter i determined

a e , e s

in imila r m nner
a s a .

A di a dv a nt ge in t h u of t h i volut i t h t the v a lue of the

s a e se e n e s a s

a b olut e it velocity w v a ry cro

s e x the it h ight as b s ee . a ss ex e a. ,
can e n

from the ta ble on pa g 6 8 The qu t ion a ri wheth r it i o t e . es se s e s n

a dvi a bl to ma ke th a b olut e it v lo ity cro the e it heigh t

s e e s e x e c a ss x

a lw a y the a me th t i
s to llow it t f llow
s ,
quidi ta nt cu e
a s, a o o an e s r v
,

henc involut Should thi be done the r l a t iv xit velocity

e an e . s e e e

would cha nge a cro the exit height ss .

F ig . 41 .

Fo r th efo m r mpl t h b lute p th of th w te r as a n

r e e xa e, e a so a e a

invol ut h as b d r w i Fig 4 1 Aft

e een l ul ting the di a gra m
a n n . . e r ca c a

for t h indivi du l poi t th r l ti v p th f t h wa t r is determ ined

e a n s e e a e a o e e .

I th i c as the r l t i v
n s it v l i ty v i
e r the exit width
e a e ex e oc a r e s a c o ss .

T h w t r gu id
e a ei t good i th bl d h nn l s as wh n
a n ce s no as n e a e c a e e

t h r l t i v w t r ou
e e a i e i vol ut but it i
a e dva nt a ge to
c rs e s an n e, s an a

h a ve the a b l ute it v l i ty f ll w a i volut i p a i n g to t h e

so ex e oc o o n n e n ss
 V
THE IN OLUT ES FOR IM PE LLE R AN D G UIDE V ANES .
71

gui de va nes which as will be shown l ter hould lso be con tructed
, , a , s a s

of involute curve If the rel a tive velocity follow the p th of a

s . s a n

involute the a b olut wa ter cour e will not follow a involute but
, s e s n

will de cribe ome ort of a imil a r curve (

s s Fi g If in thi
s s se e . s

c e the wa ter in th guide ch nnel i m de t follo w the p th of a

as e a s s a o a n

involut then the b olut e it velocity i forc d t ch nge from it

e, a s e x s e o a s

fi r t as um d p th to th t of a n involute p th w hich i a ccomp a nied

s s e a a a ,
s

by ma ll losse I the following t ble the va lue of the di a gra m

s s . n a s s

a t point m ad i a given for the as s umption th a t the b olu te

s a, ,
n re a s

wa ter pa th is an involute As before the l as t column how the .

,
s s

p riphera l velocitie nece sa ry for opera tion without hock with the
e s s s

exi ting a ngles 6 a d 6 a ccording to equ a tion 58

s . n . .

A compa ri on of the e va lue with tho e giv n i the t ble on

s s s s e n a

page 68 sho w th t in the l as t c e th di ff erence i not gre t

s a as e s so a

b twe n the periph r l v lo iti c lcul a ted from the ngl 6 d

e e e a e c es a a es . an

6 a d tho e a ctu a lly

. n i ti g Hence the di gra m gre more
s ex s n . a s a e

clo ly w ith the a ctu a l condition if the b olute w ter p th follow

se s a s a a s

an in volute th a n if the rel a tive wa t r pa th i m a de to follow a e s n

involute .

B y very c refully conducted experiments it c a n be proved th t

a a

the p ump e ffi ciency i lightly increase d for the reas ons a lr a dy s s ,

e

g iven when the

,
a b olute w ter p a th i a involute A the
s l tter a s n . s a

method i somewh a t complic ted in the following it will be con idered

s a ,
s

for implicity th a t the rel a tive wa ter pa th i a involute lea ving the
s s n ,

more re fined cons truction to the con cie tiou de igner s n s s .

(d) . The In vo lute s at Gui de Va n e En tra nc e .

The bsolu te e it v locity w whi h le a v t h imp ller a t

a x e . c es e e an

a ngle 6 i decre as ed to a v locity w a ft r it from the imp ller Its

’
. s e . e ex e .

va lue i s
72 T H EORY OF C E NTR IFUGAL PUM PS .

It is umed tha t the middle wa ter threa d flowing from the point a
as s

( se e F i g 42
.
) ret a in the velocity
s w until it ente r
.s the
’
guide v an e
s

a d th a t the pa th of the s a me will be an involute which will h a ve a

n n

a ngle of slope 6 a t the point

. where t he di a mete r is D
a ..

Fig 42
. .

The di a mete r of the b as e circle of the involute is

d; D.
, 8 111 64

The height of the bl a de ch a nnel is a t, the bl a de thickness

tra nce a d the number of bl a de
s t, n s 21 . The value of a, is the n
s;

found from
al + 81

the impeller width a t exit is b the height of the ch a nnel c an

If ., be
determined from the equ tion a

az + 8z

an dthen inver ely the b as e circle of the involute

s can be determine d
from equa tion 175 .
74 T HEOR Y OF C E NT RIFUGAL PUM PS .

The an gle 6 a t whi ch the wa ter lea ves the periphery of the circle
« is
foun d from the e qua tion

Fi g 43 . .

To guide the wa te r correctly in direction an d velocity into the , ,

c as ing it is necessa ry to slope the ends of the guide va nes a t exi t

,

so th a t the wa ter s trea m in le a vi ng the di a meter d will ente r t h e

c as ing a t the a ngle d ) .

The mea n velocity of the wa ter in the casing is ta ken if possible , ,

s m a ller th a n one th ird of the periphera l velocity a

-
At exit fro m . .

the spira l cas ing the wa ter is led through a conica l pipe to the pres
s ure pipe in which as a r u
,
le the wa ter velocity is sm a ll Frequently
, , .

the exit from the c asing is in it elf conic a l s .

In a c as ing of circ u l a r cross section the a rea s of the cross sections

must decreas e linea rly in order to h a ve a consta nt mea n velocity .

The cons truction of the bounda ry curve of the cas ing is prefera bly
ma de by dra wing sma ll circles of dia meters d d 1 etc ta ngent t o l, z, ( 3 , .
,

the common circle of radius (see Fig With this assumption

r; .

a n equ a tion c a n be derived for s u ch a bounding curve .

The quantity of wa ter an d the a reas of the circles of dia mete s r

d d d etc are proportion a l to the a ngles h w

., 5, 4, .
,
res pectively , ., ,

bec a use the mea n wa ter velocity should rema in con tan t Le t s .

Q Q
1, 2,Q etc be
3,the qu a ntitie
. s,
of w a ter flowing through the cros s
 THE GUIDE V ANE C ASING .
75

e tions of the circles whose dia meters

s c a re d l , d z, d . res pe ctively .

Then

so th a t

an d genera lly

Fig 44 . .

The i dividua l circles are to be tan gent to a common circle of

n

ra diu r o tha t the pol a r o b din a te of the bounda ry curve a t a n y

s . s c r

point of the ca ing h as the va lue

s

fa “a? r,

This i the pol a r equa tion f a pa ra bo lic pira l sta rting a t ra diu
s o s s r.

Often for chea pness of cons truction the bound y curve of the ar

cas ing is m a de circul a r th t i imply pl a ced concentric to the

, a s, S

guide van es (se e Fig With thi co truction c re must be

. s ns a

t a ken to m a ke the cross section a l a re a of t h e c as i g quite liber a l

-
n .

There exi ts a pl a ce in thi cas ing t which the direction of the flo w

s s a

mu t be divided If the cross ection i too ma ll t thi point grea t

s . s s s a s

s hock lo sse occur It i a dva nta geou wi th uch a c as ing to let

s . s s, s ,

the wa te r flow from the guide ch a nnel a nea rly ra di a lly as possible s s .
76 TH E ORY OF C E NT RIF UGAL PUM PS .

l cing a circul a r cas ing eccentrica lly a bout the guide va nes sho uld
Pa
be trictly a voided Very l a rge shock an d eddy losses occur a t t h e
s .

n rrowe t pl a ce bec a u e the w ter is s uddenly reve ed in direction

a s s a rs .

The pump effi ci ncy i grea tly reduced by such a c as ing

e s .

19 . AX IAL TH R U ST .

been sta ted cert in pre ure will e ist in the spa ces betw ee
As h as , a ss s x n

the impeller a d its surrounding c ing a d if the e pre ure a e not

n as ,
n s ss s r

of the a me ma gnitude a d a e not exe t ed

s n r r

upon equ l re of the rota ting element a a as

an a i a l thru t w ill be produced

x The e s . s

pr ure a ca u ed fi r t by the pre ure

es s s re s s ss

di ff rence in the clea r nce p c e a d s c

e a s a n e

o n dly by t h ction of c ntrifuga l forc e a e es

formi g rot ting pa ra boloid n a a .

A inv tig tion w ill n w b m a de t o

es a no e

det rmine the pr ure in p a ce 1 a d 2

e e ss s s s n ,
Fi g 46
Sho w n in F ig 4 6 h w they are formed
. .

.
,
o

a d how l a rge the res ult a nt

n xi l thru t w ill bea a s .

Sp a ce 1 is not connected with the sp ce t entran ce to the im a a

peller so th a t neglecting the rota ting pa r boloid the pressure in

, ,
a ,

the clearan ce spa ce wil l rea ch the full v lue H 7 H

2
a 1
—

9
 AX IAL T H RUST . 77

leeting the pa ce occupied by the ha ft the Side of the impeller in

s s ,

sp a ce 1 w ill be lo a ded w ith a pre ure ss

D .
2
1r

in which W is the weight of a unit volume (cubic feet ) of wa ter .

Accordi g to the la w of communic a ti g tubes the pre ure di ffer

n n , ss

ence in the cle ra nce pa ce even when con idering the influence of
a s s

2
w
the rota ting pa ra boloid will still be 7 H 2g
If it i s g a in 7
.
a

ass umed th a t the rota ting pa ra boloid on a ccount of friction losse ,

s,

i rot a ting a t the outer di a meter with a velocity

s then the loa d
on the impeller will be decre ed by a a mount equ a l to the weight as n

of qua ntity of wa ter who e volume is th a t of the rota ting p a ra b

a s

o lo id The volume of the p a ra boloid is

.

1 D. 2
1r
2 y
¢ ,
2 4 29

so tha t the loa d P on the impeller in pa ce 1 i equa l to

. s s

D “ “

)
f
z
. 1r w. ¢ E
f
2a 2 g

Inorder t m ak e thi phenomenon cle r

o s a

imagine a hollow cylindric a l vessel entirely

fi lled with wa ter (see Fi g The height .

of the cylinder is h a d the di a meter of n

the bas e D The pivot bea ring will be

.

loa ded in a dd ition to the weight of the

,

cylinder with a weight of wa ter equ a l to

,

p h 2 W I f
1
the ve el i o w rot a ted ss s n
Fi 4 7
.

4 g . ,

wi th a pe riphera l velocity f u ce t ifug l force will throw out a qu a n

O . , n r a

tity of wa ter equ a l t the vol ume of the rota ting pa ra boloid thu formed
o s .

1 D u 2 2

The vol ume of th pa a bo l d 2 4 2 is r The p o t bea r ng W ll

oI 18
1r .
i l
g
1 D u 2 “
therefore be reli eved of the loa d 5 W The lo a d on the p
7r .

ot
a
.

be arin g due to the wa te r rema ini ng in the cylinder will then be

D 2

W
1r
.

4
78 T H EORY OF C E NT R IFUGAL PUM PS .

The condition s in spa ce 1 a re simil a r to those of the rota ti ng

cylinder .

I p a ce 2 a press ure occurs a t the i nner clea r a nce ri ng t a kin g into

n s ,

considera tion e qua tions 29 an d 42 the va lue of which is ,

in which f is the outer a d f the i ner clea ran ce spa ce a rea

. n . n .

A uming th a t the press ure in s pa ce 2 i everywhere of the

ss s sa me
ma gni tude p the impeller would be subjected to a loa d of
i,

Pi w
=
( a
B a —
D ) 5
2

the forma tion of a rota ting pa ra boloid the pressure p increase s

By .

from D to D a ccording to a pa ra bol a which a t di a meter D h as a

. . ,

9 z 2

hea d ¢ 2 a d a t di ameter D a hea d ¢ 3 The lo d on the im

1
;
2 2
n . a
g
peller will therefore be increas ed by the weight of the pa ra boli c
end whose volume J is represented by the equa tion
2
Dg z
w
2a 4

so tha t the loa d on the impeller in spa ce 2 n o w becomes

2
w.
2g “
P2 P. JW D? I
1

In the a rr ngement hown in Fig 46 the pressure P

a S .
1 P 2, so th a t
there is a re ulta nt axi a l thrus t A the va lue of which is
s

A = P1 —
P3
 T HE M ULTI ST AGE C E NT RIFUGAL PUM P
—
. 79

Substituting the va lues of P an d P from equa tions 1 . 1 84 an d 188 ,

the resulta nt axi a l thrus t is found to be

various wa ys a tte mpts ha ve bee n m ad e to do a wa y with this

In
troublesome a xia l thrust but it h as never been wholly ac complished
,
.

In the d es cri ption of the v a rious types of centrifuga l pumps this

question will receive further a ttention It is necessary especi a lly .
,

w ith high press ure centrifuga l pumps to s upply s trong thru

-
,
s t bea r

ings c a pa ble o f ta king thrust in either di rection The spheri c al .

bea ring h as recently been aga in a ba ndoned bec a us e a pparently the , ,

numbe r of revolutio is too high for this type of beari ng An

ns .

a ccura te ly fi tted thrust coll a r bea ri ng h as a ga in come into u s e be

c u e this construction is simpler an d therefore increa ses the reli a

a s

bility of the p u mp .

20 . TH E M ULTI—STAGE C E NTRIFU GAL PUM P .

onsiderin g only the question of highes t possible effi ciency the

C ,

design would be limite d to a single im pe ller th a t is a single sta ge , ,

-

pump Pumps w hich a re to deliver l a rge heads are divided into a

.

num be r o f sta ges ; tha t is multiple impellers a e mounte d success ively

, r

on a sh a ft ea ch one of these impellers or sta ges delivering like fra c

,

tions of the hea d The a rran ge ment of such multi stage pumps
.
-

va ries cons idera bly as will be see n l a te r Fo the pres ent a three stage
,
. r -

high press ure pump des igned by Jager of Le ip ig Germ a ny shown

-
z , ,

in Fig 48 will be cons idered

.
,
.

The hea d for ea c h sta ge i limited by effi ciency considera tions ; th a t

s

is the cle a ra nce an d friction losses mus t not be too gre a t

,
The .

a llowa ble periphera l vel o city of the impeller limits the he a d obt a in

a ble per s t a ge .
80 TH EORY OF C E NT R IFU GAL PUM PS .

With equa l clea ra nce spa ce areas a d the use of a clea ra nce ring
n ,

the short c ircuit loss will increas e with increas ing hea d a ccordi n g to
—

the ra tio

w here H.

But it i often necessa ry

s , w hen increas ing the he d to decreas e
a s,

the numbe r of revolutions an d to incre as the im pell r di a meter D

e e .

F ig 4 8
. .

an d therefore a l o incre e the cle a ra nce pa ce a rea The s hort

s as s .

circuit lo will th n be increas d co idering the grea te r clea ra nce

ss e e ,
ns

s p a ce a re a by the v a lue
,

If the hea d which is too grea t for a Single sta ge is divided a mongst
sever a l st a ges the press ure in the cle r nce
,
p a ce w ill decreas e on
a a s

a ccount of the decre as ed he d a d for the sa me number of revolutions

a ,
n
82 TH EOR Y OF C E NT RIFUGAL PUM PS .

may be cons idered as divided in , a multis t a ge pump , a mong a ll t h e

t ges
s a .

The tota l hea d of a stage c a n not be given off ha nd bec a us e it

depends upon the qua ntity of w ter a n d hence the size of the pump a .

The l a rger the qua ntity of wa te r the l a rger the hea d tha t ma y be
ta ken per sta ge In diff erent des igns the hea d varies between 30
.

a n d 2 00 feet per t a ge s .

The cal cul a tion of multis ta ge pumps is conducted in exa ctly th e

sa me m a n ner as h as been s ho wn for s ingle s t a ge pumps ex cept th a t -

the to ta l hea d delivered ) H is divided betwee n the number of ,

7

st a ges a d the c a lcul a tions for the guide v a ne

, n s a n d im peller are

made for this fra ctiona l hea d .

21 . TH E C E NTRIFU GAL PUM P WITH OU T G UIDE VAN E S .

Up to thi s point the as sumption h as a l wa ys been ma de th a t th e

wa ter flowi g from the impeller with an a bsolute velocity w w as
n .

c a rried into guide va nes the object of which was gra dua lly to d e

creas e this velocity a d ch a nge it into pressure No t until recently

n .
,

when high pre ure centrifuga l pumps were de igned h as the im

-
ss s ,

porta nce of guide va nes been recogni ed Formerly centrifuga l z .

pumps for sm a ll hea ds were des igned wi thout guide van es Even .

to d a y there a firms who do not use guide va nes in their des igns
-
re .

Their reas ons a e th a t a pump w ithout gui de va nes is simpler a n d

r

sa fer in oper a tion th a t the guide va nes a re subject to very grea t

,

deteriora tion on a ccount Of high wa ter velocities an d th a t they c a n

be easily clogged up on a ccount of the sm a ll cross section al a rea -
.

N doubt a pum p without guide v an es is simpler an d a l so chea pe r to

o

cons truct but the sa fety of opera tion h as little to do wi th the guid
,
e

v nes as no c as e are known in which a shutdown was ca used by

a ,
s

their use Re ferring to the wea r of the guide va nes it may be sa i d

.
,

th a t with ba ckwa rd turned impeller bl a de th a t is 6-

th e s, , .

rel a tive entra nce velocity v a d the a bsolute e it velocity ma y be . n x

ta ken a ppro im a tely equ a l so th a t with the me m a teri a l t h e

x , sa

we r due to friction of wa ter a t high velocities w ill be no l a rger i n

a

the guide v an e th a n in the impeller Cloggi g of the guide van e s

s . n

can be a voided by correct design for it i po s ible to m a ke the ,

s s

entra nce height of the impe ller ch a nnel m a ller tha n the entra nc e
a. s

heigh t a of the guide ch a nnel so th a t foreign substa nces whi ch ca n

, ,
 THE C ENTRIF UGAL PUM P WITHOUT GUIDE VANE S .
83

pass through th e im pe ller will n o t clog the guide ch a nn els The .

only a dva nta ge of a pump without guide va nes is its simplicity of

construction which pe rmi ts clea ning an d repa irin g to be easily a o
,

complished These a dva n ta ges will proba bly a lwa ys be in fa vor of

.

centrifug l pumps without guide va ne an d the ma jority of firms

a s,

un derta king the construction of centrifuga l pumps will fi d it n ec es n

sa ry to build chea p p umps without guide va nes as well as pump s

h aving guide va nes By exa ct an d very ca reful construction pumps

.
,

without guide van es c an be ma de to give fai rly sa tisfa ctory e fli

c ie n c ies .

The cons truction of centrifuga l pumps without guide van es i s

a lmos t entirely confined to lo w press ure pumps ; while th i

-
con s

struction c a n never fin d a pl a ce w ith multi s t a ge high press ure -

pumps .

The c lcul a tion of a pump without guide va nes is ca rried through

a

with the sa me equa tions as before except tha t in determining the

peripher l velocity a sma ller efficiency mus t be used The loss
a . es

while the w a ter is flowing through the impeller are equ a lly l rge a

whether the pump h as or h as not guide va nes C o n sidering Fi g 9 . .

,

the pre ure hea d h a t point a will be equ a l in both desi gn but the
ss . s,

loss e during the pas sa ge from e it of the impeller to entra nce of the
s x

c asing w ill be l a rger when the guide va nes a re omitt d ; in other e

word in a pump without guide va nes a sma ller pe cent of the

s, r

”
w.
velocity hea d 2 will be tra ns formed into press ure tha n i a pump n
7;
h a ving guide v a nes .

In spite of the fa ct th a t the friction loss es a long the guide v a nes

a re a b ent in a pump without guide v nes
s the shock a d eddy a , n

losses a e grea t th a t the total lo a t exit from the impe ller to

r so ss ,

exi t of the c asing is much l a rger as h as been conclus ively proved by

, ,

experiments The wa ter lea vin g w ith velocity w mu t fi rs t fi d

. a . s n

its p a th an d the decre as e of veloc ity an d hence the tra nsform a tion
,

in to press ure wi ll not be so irregul a r .

The fa ct th t the losses up to the impeller entra nce a e eq ua lly

a r

l a rge whether or not the pump possesse guide va nes lea d to ma ki g

,
s , s n

the press ure hea d h as l a rge as poss ible in order th a t the velocity
.

hea d g ma y be as sma ll as possible By thi s mea ns the los es fter

? s a
.

exit from the impeller are not of such gre a t import a nce On the .
 THEOR Y OF C E NTR IF UGAL PUM PS .

other ha nd with increas ing pressure hea d h the pressure in t h e

, .

cleara nce sp a ce a n d therefore the short circuit losses a e increas ed -

r .

I l w pres ure pumps delivering l a rge qu a ntities of w a ter not

n o -
s ,

o nly the a ngle 6 but a lso the a ngle 6 ma y be increased an d it i

. . , s

possible to secure such peripheral velocities th a t the opposing pres

s ure he a d formed by the rot a ting pa a boloid in the cle a ra ce s p c e r n a

w ill equa l the excess of press ure in the cle ran ce s pa ce an d th u s a

prevent hort circuit losses Attention is ca lled to Article 9 a n d

S -
.

e peci a lly to Fig 22 E perimenta l work a lo g the line of reduc ing

s . . x n

the hort circuit losses hould be very fruitful

s -
s .

The form of e as i g surrounding the impeller of pump without

n a

guide va ne i of such import a nce to the effi ciency th a t only t h e

s s

Spira l h a pe c a be cons idered The a bsolute velocity w will lea ve

s n .
.

the impeller a t a certa in a ngle a d the spira l cas ing must be con n

s truc te d o th a t the form a tion a ngle of the bound a ry curve of t h e

s

c ing (des igna ted by Fig 43 ) does not devi a te grea tly from th e
as .

a ngle o th a t the velocity w is m a de t

s ch a nge its direction as. o

little pos ible From the im p ller outward the width of t h e

as s . e ,

c as ing i increas ed conic a lly in order to decreas e the velocity a n d

s

obta in a tra nsform a tion into pressure in this pa rt of the cas ing .
 II . C ONSUM PTION O F POWER AN D EFFICIENCY .

22 . G E N E RAL C ON SIDE R ATION OF C ON SU M PTION OF POWE R

AN D E FF IC IE NCY .

To ma ke the centrifuga l pump economic a l it is of the fi rst impo r

ta nce to increas e the e fli c ie y as much as possible by proper des ign
nc .

The effi ciency of a pump de igna ted by the coeffi cient 5 is under
s

stood to repres ent t h e r a tio of the work obta ined from the pump

(output ) to the work delivered to the pump (input ) .

The work obta ined A a lso ca lled the wa ter work is expresse d
, w, ,

in horse power (h p ) a d i found from the qua ntity of w t r lifted

n s a e

per mi nute a d the hea d delivered by mea ns of the equa tion

n

‘
QH n Y
33 000

in which 7 is the weight of a cubic foot of the liquid Q the qu a ntity ,

of liquid in cubic feet per minute a d H the desired hea d in feet,

n . .

Le t A be the work deli vered to the pump sh a ft

, The effi ciency 5 .

ma y then be expre ssed by the equ a tion

A.

The work delivere d to the sh a ft is found from the equ a tion

QH . 7
33 000 E

The effi ciency is dependent upon the v a lue of the hydra ulic
a d mech a nic a l loss es
n In Article 4 it wa p a rti a lly e pl a ined how
. s x

the hydra ulic losses occur an d how they e minimi ed in order to ar z

m a ke the hydra ulic effi ciency as l a rge as po ible ss .

The hydra ulic losse a re composed of the friction shock a d eddy

s , ,
n

current losses The l a tter a reduced to minimum when the

. re a

velocity dia gra m upon whi h the ca lcul a tions a e bas ed a re rea li ed
s, c r ,
z

a t t h e given number of revolutio n s This condition is recogn i ed . z

by the quiet running of the pump .

If the pump rot a tes a t a n incorrect number of revolutions it is ,

85
86 CONSUM PTION OF POWE R AN D EFFIC IE NC Y .

at once known by the rumbling noi e ins ide the c as ing By s impl y s .

touching the pump cas ing these disturba nces ma y be felt If the .

velocity di a gra ms a e a ctua lly rea li ed only a quiet an d regul a

r z ,
r

sound of flo w ing w a ter is he a rd from the ins ide of the pump .

Fo a des ired qu a ntity of w a ter the correct velocity di a gra m is

r ,

forme d only a t a cert a in head an d it depends upon the skill of ,

the de igner whether or not the correct combina tion of qua tity a d
s n n

hea d is obta ined The c a reful construction Of the bl a des a d v a ne

. n s

in the hop is a l o of the utmo st importa nce as will be seen from

s s ,

the following The finis hed impeller pass es through so m a ny ta g

. s es

of construction th a t a e ror in the dra wi ng room a t once preclude

n r s

all po ss ibility of a correct impeller The entr nce a d exit cross . a n

se ction a l a re as a d a ngle m us t be e xa ctly a dhered to in the p a ttern

n s

s hop an d foundry Above a ll the dimens ions a t entra nce of the

.
,

guide ch a nnel mu t be exa ctly those found by c a lcul a tion In high

s s .

pres ure centrifuga l pump the entra nce heigh ts of the guide ch a nel
s s n s

va ry a pproxim a t ly between the limits 5 inch a n d 1} inches If for

e .
,

exa mple the c a lcul a ted va lue for thi height i

,
of a inch s s n ,

an d it i a ctu a lly m a de
s of a inch the qua ntity of wa ter delivere d n ,

for the ra ted revolution s will be 56 6 0 of the des ired qua n

tity Of wa te r ; hence this difference of only of a n inch intro
duce an error of 7 per cent in the qua ntity of wa ter to be deliver d
s e .

Thi exa mple show with wh t a ccur cy we mu t m a ke the entra nce

s s a a s

a d exit cross sections of the guide v a nes a d impeller in order to

n n

obta in the c a lcul ted re ult a s .

The correct qu a ntity of wa ter c a of cour e be obta ined by in c re as n s

ing the number of revolution o f the driver but such a procedu e s , r

u nece sa rily incre e the pre sure hea d thereby giving a ddition a l
n s as s s ,

work which m y not bene fi t the purch as er

a .

To decreas e the hort circuit loss e the clea ra nce betwee n the
s -
s,

clea ra nce ring must be reduced to a mini mum

s .

C on cientiou work in the dra fting room a n d c a reful cons truction

s s

in the hop a the chief requi it in order to obta in good hydra ulic
s re s es

e fli c ie n c ie s .

trictly spea king only the hydr ulic effi ciency is used in the c a l
S ,
a

cul a tion of a pump as c a be e n from equ tion 1 1 an d thos e

,
n s e a

following B ut to determine the power to drive the pump not only

.

the hydra ulic effi ciency but a l so the mech a nic a l a d short c ircuit n -

losses must be cons idered .

 C ONSUM PT ION OF POWER AN D E FFIC IE NC Y .
87

The work necessa ry to run the pump light is a cert a in fra ction of
‘
the tota l driving power H ere as in a ll other form of power ma .
,
s

chinery this fra ction becomes sm a ller as the si e of the ma chine

,
z

( tot a l power ) incre as es It is therefore e v ident th a

. t the over a ll -

e ffi ciency Of a centrifuga l pump increas es as its si e (driving power) z

increas es .

The mecha nica l losses an d therefore the no loa d work c an be ,

-
,

l a rgely incre as ed by certa in circumsta nce The friction work of s .

t h e be a rings is quite sm all w ith good lubric a tion for the weight of ,

the h a ft an d impeller i not grea t If the pump is not direct con

s s .

n e cte d but h as a belt bet w een it a n d the driver a n a ddition a l lo ss

, ,

i s c a us ed by the belt pull lo a ding the be a rings It is open to ques .

tion whether in determining the effi ciency the increas ed friction loss ,

due to the belt should be included a s work delivered to the sha ft .

The stufli g b o x through poor design unequa l tightening or too

n , ,

tight a a dju tment is c a pa ble of cons idera bly increas ing the mech a n
n s ,

ica l loss es Stuffi g bo x friction i a t tim a seri ous ma tter a n d m a ny

. n -
s es

ca ses might be quoted in which the stuffing bo x c a us ed complete

fa ilure of the pump It is sometimes necessa ry when us ing high .
,

pressures to relieve the loa d on the tuffi ng bo x by mea ns of some

,
s

s uita ble a rr a ngement which a llo ws only s m a ll p a rt of the tot l a a

pres sure to a c t .

Th thrust be rings provided to t a ke c a re of the a xi a l thrus t c a us e

e a

a further friction loss th a t v a ri es more or less This bea ring va ies . r

in des ign a n d c a be either a coll r bal l or step bea ring

,
n a , , .

H igh press ure centrifuga l pumps w hen fi rst des igned were not
-

constructed with uffi cient provision for t a king c re of axi l thrust

s a a .

On a ccount of the thrust be a ring the centrifuga l pump was looked

upon as unreli a ble in oper tion Sa fety of opera tion is of first a .

importa nce when a pump i used for lifting wa ter in mine work The s .

effort is now m a de in a ll types of high pressure centrifuga l pumps to -

b a l a nce the a xi a l thru t as nearly as pos ible by uit a ble design

s s s .

The pump in which the ba l a nce is most complete ma y be c a lled

su rest of opera tion O a ccount of the high number of revolutions
. n ,

the thrust bea ring cannot st nd exce sive press ure no m a tter how
s a s ,

well they a e constructed r .

Wh ile a complete a xi a l b a l a nce is never re ched by a y type a t the a n

present time va rious de ign s very nea rly a tt in a complete ba l a nce

, s a ,

a n d therefore the thrus t be a ri ng h as such sm a ll lo a ds th a t it can no

88 C ONSUM PT ION OF POWE R AN D EFFIC IE NC Y .

longer be co ns idered a limiting fea ture as far as sa fety of opera tion

is concerned .

The exit losses from the pump cas ing will be cons idera ble in the
c ase of low pressure pumps ha ving sma ll hea ds In order to ca rry the
-
.

w?
wa ter a wa y from the pump a cert a in velocity hea d is nece ss a ry
e
——

, ,
2g
which is expressed in equa tion 8 as a fra ction of the tota l hea d H (1

With high hea d h as a v a lue

s a to w ith lo w hea ds a d n

l a rge qua ntitie of wa ter equa ls

s ato In the l a tter c as es
s peci a l considera tion must be given before choosing the coeffi c ient n
a .

It is d ifli c ult to di ff erenti te the hydra ulic from the mech an ic a l

a

losses when given the tota l effi ciency of a cent ifuga l pump an d there r

fore the curves which follow a e only for tota l e fli c ie c y of pumps

r n .

23 . EFFIC IE N CY WH E N TH E N UM B E R OF REVOL UTION S IS VAR IE D .

Inthis a rticle it will be show n how the hydra ulic a n d mech a nic a l
loss e hence the tota l loss ch a nge when running a pump a t va rying
s, ,

speeds At ea ch number of revolutions the correct di a gra m (opera

.

tion without shock ) is formed only a t a certa in hea d an d qua ntity of

wa ter .

Fo r simplicity assu me th a t the impeller h parallel ide a d

, as s s n

th a t the bl a de from entr nce to exit describe an involu t w ith a

s a e

comm on bas e circle as shown in Fig 49 Le t the consta nt cha nel . . n

height be the ch a nnel width be b a d the mea n length of the

a, ,
n

ch a nnel be I .

Firs t it will be determined how the friction lo sse s ch a nge w hen

referred to the e i ting hea ds The hydra ulic ra dius R for the
x s .

impeller ch a nnel is

If v isthe velocity in the im peller ch nnel the friction hea d due to a ,

the flo w of wa ter in the ch a nnel a d c u i g a lo of pre sure he a d n a s n ss s

w ill be

The revolution of the pump a so ch o en th a t the rel a tive ve loc ity

s re s

v lies between the limit f 30 an d 6 5 feet

s o .
90 C ONSUM PTION OF POWE R AN D E FFIC IE N CY .

S ubstituting this va lue for In e qua tion 1 9 5 , we ha ve

H ence the friction hea ds va ry as the des ired hea ds ; th a t i i o s, n n e

an d the s a me impeller the friction lo es if expre ed in per ce t f ss ss n o

the exi ting hea d will be equa lly l a rge for he d within quite w i d
s a s e

limit s .

It is pl a in th a t thi is true for bl a de ch a nnel of other fo m s

s s r as

well as th a t shown in Fig 4 9 a d a l o tha t the sa me la w a ppli e s t .

,
n s o

the friction lo e in the guide ch nnel ss s a s .

Further lo es occur in the udden incre as e or decre as e of t h

ss s e

w a ter velocitie a t entra nce to a d exit from the impeller a d g ui d

s n n e

va ne due to the thickne of the bl d e w hich produce shoc k

s ss a s, s s .

According to e periment by Wei b ch x d Fli e g w ith p i p s s a an ner es

h a ving udden an d gr dua l enl a rgemen t the friction he d c o

s a s, a s n

sumed by hock a a pproxima tely proportiona l to the q u a re f

s re s o

the velocitie The sa me losse s occur in impe ller a d guide c h

s . n an

nel s when the wa ter ent r or lea ve the ch a nnel in the dire c t i e s s s on

of the bl a de or when the pump opera te With correct velocity d i

s s a

gra ms Therefore the shock losse t entra nce to or ex it fro m the

. s a

impeller if e pre ed in per cent of the de ired head will b e eq u a lly

,
x ss s ,

l rge for a ll hea d

a s .

;
The velocity he d } w hich i necess a ry to ca ry the w a te r a wa y
i t
a s r

from the pump is repre ented in equ a tion 8 as , s

The velocitie w a t exit from a pump c ing will va ry under a

s . as ,

ch nge of wa ter qu ntity as the qua r root of the hea d b c a us e the

a a ,
s e e

c ro ss ection rem in the me H nce the coeffi cient w ill h a v e

s a s sa . e a

the a me va lue for a ll hea d

s s .

C ollecting the re ult of ll t hi s inve ti g tion it i found th a t the

s s a s a ,
s

friction hock a d it lo
,
s if e pr d in p cent of the e x i t
,
n ex ss e s , x e ss e er s

ing hea d wi ll be independent of the l tter The hydra ulic effi ci e ncy
,
a .

of a centrifuga l pump will rem a in a ppro im a tely con ta nt o v e r a x s

Wi de r nge of he a d pro vided the p ump run a t the correct u

a s m ber s n
 E FFIC IE NC Y WITH VARY IN G SPE ED . 91

of re volutions for the existing hea d so th at the correct di a gra ms

wi ll be formed .

The vari a tion of mecha nical losses will be investiga te d for Simi
lar condition s .

The tota l work N necessa ry to drive the pump i composed of

. s

the w ork cons um ed in overcoming the wa ter Work N the hydra ulic w,

fri ction work N a d th e mech a nica l friction work N so th a t

n

Na = Nw + Nh + N a

N»
1 N
¢ w

Fr om the preceding inves tiga tion it follows th a t the exi ting frac s

tio of the wa te r work exp ressed as hydra ulic friction work is inde
n

pe d ent o f the hea d hence the coeffi cient rema ins cons ta nt
n , .

F o the o e fli c ie t wthe conditio ns a e di ff erent

r c n If we as s ume
r .

th a t the bea ring thrus t loa d etc rema in equa lly l a rge for va rious
,
.
,

hea ds then the power necessa ry to overcome the exi ting friction
,
s

wo rk va ries as the periphera l velocity or number of revolutions ;

th e refore
N» 1 ; wl g
” 1

Nat, s vg ” z

The wa ter work now varies as the squa re root of the third po wer
of the hea ds a n d the number of revolution s as the squa e root of
,
r

th e hea ds ; hence

71 : VH

F rom these equa tions an d equa tion 2 0 1 we ha ve

mecha nical friction va ries inversely with the co res ponding hea ds r .
92 CO NSUM PTION OF POWE R AN D EFFIC IE N C Y .

The la rger the hea d a n d therefore the driving pow e r the smalle r

the coeffi cient 4 7.

Wh ile the va lue of the friction an d Shock losses w hen refer ed to r

the hea d i independent of the a me the mecha n ic a l losse s will

s s ,

decreas e with increas ing hea d This wi ll now be s hown by means .

of a exa mple
n .

G iven a four t a ge high press u

-
re centrifuga l pump d eliveri g 100
s -
n

cubic feet per minute un der a hea d of H 400 fee t a d requiri g . n n

a driving power of 1 00 h p The wa ter work N i .

w, s

QH . 7 1 00 X 400 X
“9 °

33 000 33000

Le tit be umed tha t e pe riments h a ve hown the over a ll fli

as s x s -
e

N
c ie c y to be 5
n a d th a t the power req uired to ov n er
N a.

come mecha nica l friction i 5 hp Le t u w se e wh a t th

s . s no e

effi ciency of thi me pump w ill be if it delivers only a hea d H

s sa a,

2 00 feet .

With H 400 feet a d N 100 h p the po wer req uired t n ., . o

overcome th h ydra ulic lo e N

e 1 00 5) m m ss s 1.

N
Therefore t h e coeffi cient ¢ N F urther as th
,“
1 e
75 8 ,

d riving po wer to overcome mech nic a l friction N 5 hp w h v a ,

e a e

N m,5
sh
Nw .

The w a te r work N w, is fou d ccording to equ tion 2 o2 a

n a a as

Nw ,
VHIE h p . .

( b
i ( P2, l
l /z V /

h e ce
n

N., X h p . .

N 6. x h p . .

Therefo re th e tota l driving power is

94 CONSUM PT ION OF POWE R AN D EFFIC IE NC Y .

i mpeller width b is inches then h k ; th a t is the fr i c t i o n

, . g ,
2

hea d per ch a nn el is times as l a rge when 2 cha nn e ls are us e d a s

2
when 2 ch a nn el a re u ed
.

s s .

The friction hea d refers to one cha n el a n d as the numb e o f n ,

r

ch a nnels w s twice as ma ny in the fi t impeller as in the s e c o d

a rs n

the tot a l friction hea d will be times la rger in the former th a i n n

the l a tter .

By decreas ing the number of bl a des the friction he a d c a us e d b y

bl a de shock will lso be decreas ed The res ista nce hea d h c a us e d
a .
.

by sudden cha nge of cross sectio is n

Ew —
WP

or letting v
,
’
( 8 being the bl ade thickness ) ,

h”
a i y s
’

so tha t if the coefficient If is as s umed eq ua lly la rge in ea ch c as e the

following equa tion holds

R eferring a g in to the exa mple u ed a bove an d ta king the b l a d e

a s

thickne 3 for bo th impellers as

ss inch an d solving equa tion 2 05 ,

h = 3 2h .

This friction he a d r fer to one ch a nnel so th a t the to t al res i st a n c e

e s ,

he a d c au d by bl a de hock i
se times as grea t in a n impeller w ith
s s

z bl a d es as in one h a ving 3bl a des .

Thi exa mple Shows clea rly the ga in in effi ciency obta ina bl e b y
s

d ecre as ing the number of bl a des Only as m an y bl a de houl d b e

. s s

u ed in the impeller a d guide va nes as is a bso lutely necessa y fo

s n r r

good wa ter guid a nce .

 INCREASE OF H Y DRAULIC E FF IC IE N C Y .
95

25 . IN CR E ASE OFHYDR AULIC EFFICIE N CY WH E N 6. 6 . AN D BY

USIN G DIRE C TIN G VAN E S AT ENTRAN CE TO IMPE LL E R .

The friction loss c a n be decreas ed in a nother wa y which should be

employed when the highest effi ciency is sought .

I Article 14 it was Sho wn h o w the rel a tive entra nce velocity

n

could be decreas ed by decreas ing the a ngle Fig 33 Shows pl a inly .

.

the decre ase of v wi th the decrease of the a ngle 6 for every 6 the
. . .

corres ponding velocities h a ve been plotted a ordina te s In order to s .

s horten the p a th Of the w a ter in the uction tube when ass uming s

6 2 90 directing v a nes mus t be used before entr a ce to the im

°
. n

peller These directing v ne erve two purposes : fi s t they guide

. a s s r ,

the wa ter in the suction pipe in a axi a l direction ; second they give n ,

to the wa ter directly before entra nce to the impeller the proper
entra nce velocity an d di ection r .

When 6 90 thes e directi g v an e a re s uperfluous bec a use the

.
°
n s ,

wa ter in ente ring the impeller a t right a ngles wi ll of itse lf flow axia lly
through the suction tube .

The use of directing va nes will of course somewh a t increas e the

cost of the pump but it will be hown th a t for high press ure cen
, s
-

t i fu
r ga l pump in w hich the w s
a ter velocities are high there i s a

decided ga in in the decreas e of friction hea d .

A y ch a nge of the an gl e 6 ch an ge a l so the periphera l velocity

n . s ,

bu t for implicity the l a tte r will be cons idered const a nt in the follow
s

ing discussion

Fig 50 . .

Fig . ows in full lines th e entra nce velocity di agra m when

50 sh

9 0 a n d i n dotte d lines the entra nce velocity di a gr a m when

°
6.

90 let the m a gnitude of the

.

i which 0
°
6. nw Fo r 6 . .. .

velocities be u w v an d fo 6 6 let them be

., ., ., The r . .

friction he ad s for both c as es will o w be compa red n .

Le t th e entra n ce w idth b be the sa me in both c as es ; then the

.

entra nce heights de igna ted a a a d will be in the inverse ra tio

s s . n

to the rel a tive velocities ; therefore

c .l

E.
96 CONSUM PT ION OF POWE R AND E FF IC IE NC Y .

Ac cordi ng to e qua tion the friction hea d for 6 °

1 94 . 9 0 is

a. b. v.
’
h. A l
2 a .b. 2g

the case of 6 6 the wa ter mu t flo w through two ch a nels fir t

In . . s n ,
s

in the directing va nes a n d then through the im peller The velocities .

in bo th cha n nels a re e qua l beca use v L t h be the fricti o ., e n

hea d los t in eac h chann el ; this c an be expres sed as

= k
a. +b . v.,
z

h 1
2 a .b. 2g

Assum ing
tha t the ch an n el length l a n d the coe ffi cient A a re of
equa l va lue in equa tions 194 an d 2 07 the following equa tion is fou d ,
n

by di vidin g the former by the la tter

he as be ac: ”0
2

b. . a .. b. a.

The friction hea d h occurs tw ice first in the directing v a nes t

., , a

exi t second in the impeller a t entra nce Beside thi a ma l l fricti

,
. s s s on

head is lost while the wa ter is flowing through the fi st pa rt of th r e

directing va nes As the cross section s in the first pa rt of the d ire t

. c

ing va nes a re l a rger th a n in the l as t pa rt the velocity of the w a t r is e

s m a ller in the fo rmer th a n in the l a tte r Therefore we h a ll ass um . s e

tha t this sma ll friction hea d is equa l to The total of a ll th s e e

friction heads is h Ta king thi into a ccount a d substitutin g s n

5 of equ a tion 2 06 in equ a tion 2 08 the l a tter t a kes the form

3
a
‘
.I
t

example I Article 3 1 the following impeller d imens ions at

In , ,

entra nce an d va lues a re given for the entran ce velocity di a gra m when
6 0 60

v. ft .
,
a. in .
, s. in .
, b. in .

Wh en 6. 90
°
the following v a lues were found
w. ft .
, 0. ft .
,
a. in .
, s. in .
, b. in .
98 C O NSUM PT ION OF POWER AN D EFF IC IE NC Y .

When directing
va n es a re used the sa me res ista nce hea d occurs at
exit from the va nes as a t entra nce to the impeller Further t .
, a

entran ce to the directing va nes a nother but sma ller resista nce head , ,

occurs The l a tter is sma ller beca u e the cross sections of th

. s e

ch a nnel s a e l arger a d therefore the velocitie corre po ndingly

r n s s

s m a ller As s uming a g a in th a t thi s re ista nce hea d a t entra nce i

. s s

h
of th a t a t exit from the directing va nes then h
,
.

Therefore when directing v an e a re us ed a n d w hen 6 6 the

, s . .

resista nce hea d c a u ed by a ll the hock losses is s time as small s s

as for the c as e w hen 6 in which the res ist a nce hea d i ca used
. s

by only one Shock loss a d th a t a t entra nce to the impeller n .

Without doubt this is suffi cient proof th a t m a king 6 .

e peci a lly for the cas e 6

s c a uses the friction a n d hock l o sses at
. s

entra nce to the impeller to be considera bly decreas ed n o tw ithsta d n

ing the fac t tha t the lo es in the directing va nes h a ve been add dss e .

E peci a lly in the des ign of high pre ure centrifuga l pu mp s dir t
s -
ss ec

ing va nes a t entra nce to the impeller hould be used even though s ,

they a d d to the cos t in order th a t the highes t hydra ulic effi cie cy
,
n

m a y be Obt a ined .

The condition a t entra nce to the directing van es a re s i mil ar t

s o

those a t impeller entra nce a d w sho uld be ma de as small a ,

n .
2 s

pos ible s .

26 . C OM PAR ISON O F TH E FR ICT ION H E ADS OF C E NT R IFU GAL PUMPS

WIT H IM PE LLE R AN GL E S 6 . B 6
“

a AN D . AND

TH E IN F LU E N C E OF C U R VATU RE OF B LADE S U PO N E FF IC IE NCY .

The friction lo e occurring during the flo w of the wa ter through

ss s

imp ller a d guid v a ne as me tioned in Article 4 re depend t

e n e s, n ,
a en

l o upo n the ha pe of the bl a de a d the form of the ch a nnel It was

a s S n .

found e perimenta lly th a t the b t effi ciencie were obta ined when
x es s

the impeller bl des were turned b ckwa rd th a t i when 6 90

a a s, s, .
°

The poorest effi ciencie were obt a ined with impeller bl a des decidedly
s

turned forw rd or wh n 6 a Impeller bl a des w ith 6

,
e90 . .
°

were much poorer i ffi i cy th a n impeller bl a d h a ving 6 90 °

n e c en es .

Therefore a t pr ent in t h b t desig s of centrifuga l pumps im

es e es n

peller bl de bent ba ckwa rd with 6

a s 90 e a lmo s t excl usivel y .
°
ar

used .
COM PAR ISON OF FRICTION H EADS OF C E NTRIFUGAL PUM PS . 99

An ttempt was made to investiga te somewha t more close ly the

a

rel a tion of the va rious types of impeller bl a des to the effi ciency ,

but of course it was not po ss ible to dete rmine for a y one impeller n

the m a gnitude of the friction hea ds c a used by the wa ter flo wi g n

through the ch a nnels bec a us e the coeffi cient of friction a e unknown

,
s r

a d c a n only be fou d experiment a lly

n Still the rel a tion of the
n .

vario us friction hea ds c a be a pproxima tely determined by mea n

n s

of a equ a tion imila r to equa tion 2 08 by as s uming th a t the c o e ffi

n S

c ie ts of friction an d the lengt h s of the ch a nnel s a re a l wa ys equa l

n .

The velocities of the wa ter th a t a e dea lt with in this cas e va ry r

between such small limits tha t this as sumption c a n ma ke no a ppre

c ia ble error .

Fi gs 5 1 52 a d 53 Sho w three i mpellers with guide va n es

.
, ,
n Th e .

impellers ha ve the follo wing a ngl es :

1 .
a .

II .
a .

III .
a .

Ass ume as before

D . fee t ,
D . foot ,

0. feet ,
w. fee t,

v. gH . feet ,

l o assume the cas e 6

an d a s .

Fo r the c as e 6 6 the periphera l velocities a re found from

. .

equa tion 136 The other velocities a d ch a nnel heigh ts a found

. n re

pa rtly a na lyt ica lly an d pa rtly gra phic a lly The bes t wa y to do .

this wi ll be expl a ined in a l a ter a rticle In Ta bles I a d II are . n

given the va lues thus derived for these three impe llers an d their
guide va nes .

Impeller I a d II th a t is 3 ha ve te n °
s n 135 a n d 6
, , a .

bl a des ea ch Impeller III with 6 45 h ad to h a ve fourteen

.
.
°

bla des bec a use for a fewer number of blad e the involute a t entr nce s a

could not be ma de to join the involut a t exit Thi impeller could e . s

not be des igned with bl a des of cons ta nt thickne bec a use the bl a de ss

curva tures would h a ve been too grea t n d would h a ve resulted in a

a completely useless ch a nnel All three guide va nes h a ve nine

.

bla des eac h .

1 02 C ON SUM PTION OF POWER AN D E FF IC IE NC Y .

If the coeffi cient k be ta ken equa l throughout the rel a tion of the ,

res ista nce hea ds ca used by shock los es a t the impeller entra nce c a n s

be expressed by the equa tion (see equa tion 2 10)

I
)
:
II

Fo r impeller III the righ t ha nd side of thi equa tion must a ga in

-
s

be multiplied by H ? the r a tio of the number of bl a de in e ch c ase s a .

The rela tions of resistan ce hea ds for im peller exit a d guide van e n

entra nce c be determined exa ctly like the a bove equa tions by us i n g
an

the correct ub cripts s s .

The short circuit losse will not be the sa me for the three im
-
s

p e lle rs bec
, a u s e by ch a nging the a b olute e x it velocity the e x c ss s e
2
w
pressure exi ting in the clea ra nce pa ce ; H 2 va ries
s s ,
1 .
.
.

a
In thi c as e a l o the hort circuit lo
s s s i for impeller I
- ss s ,

a n d the hort circuit loss es of impellers II a n d III w ill be comp a re d

s -

to thi v lue
s a .

As u ming equ a l va lues of the coeffi cients of friction the follo wing
s

equa tion ma y be written :

Q‘
an d s imil a rly for the ra tio pm

q‘
pl

In Ta ble III the va lues of the e va rious rel a tions a

s re given .

T ABLE III .

Im pe lle r E n t ra n c e Impe lle r E xi t

Gui d e Va ne Eu
tra nce
Sh o rt

F ri c t io n Sh o c k l F ri c t io n S ho c k F ric t io n Sh oc k Lo w
Lo s se s Lo sse s Lo ss e s Lo sse s Lo se s
COM PAR ISON OF FRICT ION H E AD S OF C E NT R IF UGAL PUM PS . 1 03

According to this ta ble the resista nce hea d ca us ed by friction

lo sses a t entra nce to the guide va nes is ti mes grea ter for guide
va ne III th n for guide v ne I The resista nce hea d for impeller
a a . s

III flu ,
ar a t every point con idera bly gre a t er th a n t h o se
e s

for impeller I only the short circuit lo se a e sma ller in c as e III

,
-
s s r ,

a n d these do not compensa te for the other lo ss e which e con s, ar

s id a bly gre a ter

er The tota l efficiency of the forw rd turned bl a de
. a -
s

will therefore be less th a n the effi ciency o f_ the ra di a l a n d the ba ck

wa rd turned bl a des
-
.

The re i ta nce he d a t impeller entra nce a d exit a e sma llest for

s s a s n r

impeller II 6 but a t guide va ne entra nce they a e la rger

, . r

th a n those of impe ller I while the short circuit lo e a e sma ller ,

-
ss s r .

It must be borne in mind ho wever th a t the res i st nce he a d a t , , a

entr nce to the guide va ne forms the l rger per ce t of t h tot a l

a s a n e

resi ta nce he a d beca use the l a rgest velocities occur t thi point ;
s , a s

the resista nce head a t impeller entra nce is much ma l le a d is s r n

s m a ller a t impeller ex it It ma y be correctly as umed t h t the ga in . s a

of frictio hea d a t impeller e t r nce a n d exit will nea rly Off et the
n n a s

l a rg r loss of friction he d a t guide va ne e tr nce s tha t con

e a n a , o ,

s ider n i g only tho e fa ctor ju t di cu sed it would ppe a r th a t the

s s s s s ,
a

s mal ler hort ircuit lo e of c ase II c u e the ra di l bl a de to give

s -
c ss s a s a s

a g in i effi ciency ov r tho e b t b ckw rd

a n e s en a a .

B e ide the f ctor

s s bove m ntioned uch as frictio
a s a d hock e ,
s n an s

losse the form of the bl de ch nnel must be considered in deter

s, a a s

mining the hydra ulic effi ciency o f the e impe lle s rs .

I gi v i g the bl a de ch a nnel its proper form c a re must be t ken in

n n a

order th a t the ch an ge from rel a tive entra nce velocity to rel tive e it a x

velocity sha ll occur gra du a lly a d if pos ible linea rly The bl a de n ,
s ,
.

c urva ture Should a lso h a ve the l a rge t poss ible ra dii These con
s s .

d it i o a e be t ful filled by
ns r bl a de ch a nn el formed w ith b a ckwa rd
s a

turned bl a des th t i impeller I I thi c as e it i easy to construct

,
a s, . n s s

the bl a des wi th l a ge ra dii of curv ture a d a lso to construct the

r a n

impeller ch a nnel o t h a t the ch a nge from v to v hence the ch a nge of

s . .,

cross ection sh ll occur as nea rly linea rly as poss ible The de igner
s ,
a . s

ca n convi ce himself th t thi is not eas ily po ible with the ra dia l
n a s ss

bl a des of cas e II I c e II the fi r t pa rt of the impeller ch a nnel

. n as s

cross ection diverge more ra pidly th the l a tt r portion nea r the

s s an e

impeller e it a d for thi re o the d cre of velocity w hen ra di l

x n ,
s as n, e a se a

bl a des a e used occur in ma er imil a r to the curve how dotte d

r s a nn s s n
1 04 CO SU
N M PT IO N OF POWE R AN D E F F IC IE N C Y .

in Fig 54 The full line shows the ide l de re of velocity hich

. .
a c ase w

c an be obt a ined w hen b a ckw a rd bent bl a de s a r e u s ed I thi c ase - . n s

the tra sforma tion of velocity into press re occu

n u rs w ith le hoc
ss s k

th a n when ra di a l bl a des a e used r .

Due to a phenomenon w hich w ill be briefly a lluded to a h a r p s

curva ture c a a l so grea tly influence the fli ie c y of flo w

n e c n .

It i genera lly kno wn th a t the re i t nc to

s a fl o w thro gh b n
u a e d s s e

increas e with decreas ing ra diu of cur ture

s v a B u t the incre s e d s . a

friction loss is o t a lone present bec u e it n s h a been found th a t a s

a nother phe omenon t a ke s pl a

n ce w hen there i s fl o w through h a r p s

curva ture It is ca lled the phenomenon of

s .
w h i li n g w t e ) r a r
“

Fig 54 . . Fig 55 . .

publis hed the re ul ts of ome very intere ting inve tiga

Isaa c h e n s s s s

tion t hrough curve d cha nel in Zi vi l I g i u 189 4

s n s -
n en e r,

I the e rticle I aa h
n s proves th a t in a curved pipe of rectilinea r
a s s c en

cro ection the velocitie in r di l pl ne a e inve ely propo rtion l

ss s s a a a s r rs a

to the corre ponding ra dii of curva ture o tha t if repres ents the
s ,
s ,
r.

out r d the inner r diu of curv ture a d w an d w the corr

e an r. a s a n . . e

s po d i g velocitie a t these r dii (

n n Fig the follo wing equ a tion
s a se e .

ma y be w ri t t n : e

w. r . C

The velocities in ra di a l ection of a curv d pipe va ry a ccording s s e

to a equil a tera l hyperbol a

n .

To det mi e the co n sta nt C a ra diu i found where a velocity

er n s s

w occurs w hich h as the a me v a lue

.
t h me a n velocity in the s as e

P f Pf f D m t dt G m y t
s so r
ro e t fu
ll y f t h i ph
arr o m ar s a , er an ,
re a s o s e no eno n

of wh i li g w t ri h i l tu
n w t t u bi
a er
”
p i ll y w
nh ids ec re
s on a er r n es , es ec a en co n s

e ri n g t h e F ra n c is t urbin e .
1 06 C ON SUM P TION OF POWE R AN D E FFIC IE NC Y .

Incons idering the frictio n s hock a n d short circuit losse s th e , ,

-

conclusion h as been rea ched th a t the friction d shock losses a r an e

equ a lly l a rge for the ba ckw rd bent an d ra di a l bla des While the short
a -
,

circuit losses a e l a rger with the ba ckwa rd bent bla des Thi grea te r
r -
. s

s hort circuit loss of the b a ck wa rd bent bl a de

-
i fu lly compensa te d -
s s

for by the more a dva nta geou form of the bl de ch a nnel s a .

Experiment h a ve proved th a t the be st effi ciency i rea li ed w ith

s s z

the ba ckw a rd bent bl a des a d for this rea on impeller bla de e o w

-
,
n s s ar n

a lmos t ex clu ively des igned a ccording to thi form

s s .

The losse c a used by the h a rp bl a de curva ture of im peller III

s s

become very noticea ble i the tota l effi ciency a d for thi reas on
n ,
n s

s uch bl a des a entirely out of pl a ce in centrifuga l pump design

re .

27 . TH E E FFIC IE N CY As DET E R M IN E D BY EXPE R IM E NT .

In the foregoing it h as been shown ju t wh a t points must be espe s

c i a lly co s idered in the de ign of centrifuga l pump in order to ob t a i n

n s s

the highes t e ffi ciency The following experimenta l d ta w ill be

. a co n

80 90 [ 00 [ IO 72 0 13 0 I4 O 150

Fig 56
. .

their influe ce upon the tota l e ffi ciency will be a na ly ed

si de re d a n d n z .

The excellent results obta ined by Sul r B ro of Winterthur a d ze s . n

C H Ja eger
. . C o of Leip ig h a ve been obt ined only by c a reful
. z a
 TH E E FFIC IE N C Y AS DETE R M INE D BY E XPE R IM ENT .
1 07

des ign of the pump in the drafting room an d by fa ultless construe

ti on in the shop .

Speci a l cons ider a tion is due to the e xpe riments m a de on Sul z er

pumps a t the H c o j a Hydra ulic Pl a nt in Spa in w hich will la t r
or ,
e

be more fully de cribed The res ult of the experiments are shown
s . s

in Figs 56 a d 57
. n .

I both tes ts the qu a ntity of wa ter w as the sa m e but the ma n o

n ,

metric hea d a ft r one yea r s ervice w as 12 76 feet while a fter five

e
’
s ,

80 00 too 110
c a 49 7: a n ? M /M

Fig 57
. .

y e a rs se
’
rvi ce it was 1 5 75 feet W ith 1 27 6 feet he a d the hydr a ulic .

pl a nt opera ted with three pumps in series th a t is ea ch pump dc liv , ,

ered a bout 42 5 feet head while with 1 57 5 feet hea d a a ddition l

,
n a

pump h ving 357 feet hea d wa a dd d Therefore the experimenta l

a s e .

re ult refer to th ree a d four pumps r pectively ea ch of which

s s n es ,

con isted of four t ge The mea n numb r of revolutions in bo th

s s a s . e

c as e was 87 0 per minute A the curve Show the effi ciency of the
s . s s ,

pumps under a mea n loa d of 4 14 h p was 7 6 per cent a n d a ft r . .

,
e

fiv ye a s of service deliverin g a l rg r he a d corr pondin g to a lo a d

e r a e es

o f 4 9 2 h p it w as a g a in 76 per c nt
. . Accordin g to the report of e .

T a k en fro m H e rzo g, Ele c tri c Ra ilro a d s an d Th e ir Ope ra ti o n , 1905 .

1 08 C ONSUM P TION OF POWE R AN D EFFIC IE NC Y .

the s upe rintendent the pumps were only out of service an a verage
of sixteen hours per month so th a t during five yea rs of a lmos t ,

uninterrupted service no decreas e of e fli c ie n c y h a d ta ken pl a ce .

This is a striking proof of the f ls ity of the frequently ma de sta te a

ment tha t ra pid deteriora tion of impeller a d guide va nes on a ccount n

of high wa ter velocitie a l wa y s occur in these pump Such goo d s s s .

a d n usta ined e ffi ciencie w re obt ined only by very c a reful con

s s e a

struction in w hic h the correct rel a tions of a ngles cro ss s ection ,

s,

n umber of revolutions to qu a ntity of w a ter a d he a d des ired were n

rea li ed produc ing a pump which worked completely without shock

z ,
.

‘
CU fi f .
’
” if M IN

Fi g 58 . .

The experiment recorded in Fig 58 are for a twos tag e mi ne

s .

S h a ft pump h ving a v rtic l h ft The e e periment how th t

a e a s a . s x s S a

in ge er l the l rger the c pa city f the pu mp t he higher will b

n a a a O e

t h e ffi cie cy
e bt ined Thi t w t ge pump deliver 56 5 ubi
n O a . s o-s a s c c

fe t f w ter p minute g in t m nometric h d of 148 feet t

e o a er a a s a a ea a

1 02 5 r p m F the gu r t d lo d the fli i y was 83 pe cent

. . . or a an ee a e c enc r ,

but und r other working condition the c urv e how th t the e fli

e s s s a

i
c en c y incre d to 8 4 per cent as e .

Equ a lly good re ults h a ve been obt ined by C H Ja eger C

s a . . o .

wi th their high pre sure centrifuga l pum p Fig 59 give xpe i

-
s s . . s e r

menta l r ult obt ined from a ix t ge pump which d livered 7 1 cubi

es s a s -
s a e c

feet Of wa ter per minute g i t 36 1 f t m a nometric he d a t 14 00 a a ns ee a

r p m The a b ci ae r pr e t t h qu n tity of w a ter per minute

. . . s ss e es n e a

in 100 pa rt of th orm a l q ua ntity d ired ; the upper curve hows

s e n es s
1 10 C O NSUM PT ION OF POWE R AN D EF FIC IE NC Y .

greater hea d Obta ined by throttling does not fully bene fit the c us
tomer beca use in mo t c e only one certa in hea d is of use The
, s as s .

cus tomer will be in terested in effi ciency curve wh ich refers an

to the one hea d he de ires If this i considered in the a bove s . s

experiments the effi ciency curve will decre as e ra pidly with sma ller
qu a ntities of wa te r but for the norma l loa d the effi ciency given
,

is correct .

The experimenta l results given bove for high pressure centrifuga l a -

pumps Sho w a tota l e fficiency be tween the limi ts of 70 a n d 80 per cent ,

v rying a ccording to the si e a d po wer of the pump Such high

a z n .

/2 IO If I.

va lue of e ffi ci ncy however c a only be rea ched by proper con

s e , , n

struction .

A lre dy me tioned t h tot a l los con i sts of the hydra ulic a n d

s a a n , e s s

mech a nic l lo es I the hydra ulic lo e the friction hock a n d

a ss . n ss s , s ,

exit losse e s epa r ted from the short circuit lo

s ar a The former h a ve -
ss .

a i fluence on the hea d de ired bec a u e c rta in pr sure he a ds a e

n n s s e es r

nece a ry to overcome them wh ile the short circuit losses require

ss
,
-

more work bec u e the pump must a l so deliver the qu a ntity of wa ter
a s

Short circuited —
.

Fo the c a lcul a tion of peripher l velocity number of revolution

r a , s,

etc only tho e lo ses a e t ken which influen e the he a d desired ;

.
, s s r a c

hence the tota l workin g hea d mu t be m ultiplied by the coeffi cient s

 T HE E FFIC IE NC Y AS DETER M IN ED B Y E X PE R IM E NT . 111

17 1 p a equa tion
(se e
’

I determining the dimens ions

n

of the impeller cha nnels the qua ntity of short c ircuited wa ter m ust -

be co n sidered .

The um of the short c ircuit a d mech a nic a l loss s is cons idered

s -
n e

to be a bout 6 pe cent of the to t l lo

r a d then the coe ffi cient ; i
a ss n 1 s

determined I order to h a ve a m ll excess of hea d t norma l loa d

. n s a a

it is well not to ta ke thi coeffi cient to o sma ll

s .
 III . T H E RE GULATION AN D C LASSIFICAT IO N OF
C ENTRIF UGAL PUM PS .

28 . RE G U LAT ING T HE Q U A NT IT Y OF WAT E R B Y TH ROT TL ING . IN

FL U E N C E O F TH E Q UA NT IT Y OF WAT E R U PO N H EA D AN D LO SSE S
DU E TO TH ROT TLING .

It h as bee n proven th a t a centrifug l pump will work witho ut a

s hock loss es a t a certa in number ,

f revolution only for one c m o s, o

bina tion of qua ntity a d hea d a d a t t h i point will rea ch it highes t

n n s s

e fli c i n c y
e If the qua ntity of wa ter delivered by a pump running a t
.

a co ns t a nt number of revolutions i regul t d which in mo t case s s a e ,

s

is brought a bo u t by throt tling then o pe ra tion W ithout shock ceas ,

es .

This throttling i a ccompli hed by me ns of a va lve in the disc ha rge

s s a

pipe Therefore every regul a tion by mea ns of a throttle va lve occ urs
.

a t the e xpens e of e fficiency .

Shock a d eddy curr nt lo e en ue a t entra nce to the impe ller

n -
e ss s s ,

beca use the qua ntity of w ter decre e c a using the a bsolute n a as s, e

tra nce velocity t th e i mpeller to as s ume a direction other th a n th a t

a

ca lled for by the slope of the involute a ngle a t the e it fro m the x

directing v ne At impeller exit a correct di a gra m might occur i f

a s .

it were not for the fa ct tha t with decre ing qu a ntiti of wa ter t he as es

a b solu te exit velocity tha t i the entr nce velocity to the guide
,
s, a

va es form a ma ller a ngle 6 with u tha n i dem a nded by th e

n ,
s s . . s

entra nce involute of the guide va nes .

The ma gnit ude of the lo which occur when the pump is throttled sse s

a d therefore opera ting w ith hock

n h a rdly be determ ined m a the s can

ma t i a lly c It i the duty of the xperiment l l a bora tory to conduct

. s e a

e periments long thi line W h ll only ttempt to ho w h o w t he

x a s . e s a a S

lo es ma i ly occur when throttling a d h o w they ma y be decreas ed

ss n n

by suita ble de i gn s .

Evi dently the loss es a t entra nce to a d in p ing through th e n as s

g uide v ne w ill
a be of thes gre a t r m a gnitude A ss ume pump of e . a

infinite di ameter o th a t the impeller d guide va nes an be rep e

s an c t

sented in Fig 6 1 The p rt of the bl a de which were involutes are

as . . a s

n w tr a ight lines bec u e the di met r of their b as e circle is infi nite

o s a s a e .

1 12
114 R E GULAT ION AN D C LASSIFIC AT ION .

of the entra nce velocity di a gra m ; the rel tive velocity 0 wi ll enter the a .

impeller a t a sma ller ngle a d c u e eddy current losse Sim il a r a n a s -

s

to those a t entra nce to the guide va ne If on the other h a nd the s .

, ,

clea r nce pa ce i m de ufli ie tly la rge the a b olu t velocity in

a s s a s c n s e

the cle a r nce p ce c ch a nge its direction Thi is due to the

a s a an . s

influence of the periphera l velocity a d the rel a tive veloc ity v n .

w h ich strives to enter the impeller t a a ngle so th a t the a ngle a n

6. at impel ler entr a nce c a till h a ve the proper v a lue B y h a vi g n s . n

a s u ffi cient cle a r nce it might be poss ible to obta in a veloc ity d i a

a

gra m a t impeller entra nce which i pr ctica lly free from shock for s a

qu a ntities of w ter withi wide limi ts a n .

When a pump is throttled not o ly is the qu a ntity of w a ter n

ch nged but the hea d a l o is more or less influenced I designing

a s . n

the bl des of pump thi mu t be es peci a lly c n idered for i mo t

a a s s o s ,
n s

c a e it i till ne
s s s s ry to obta in the norm l he a d when the pump
c e s sa a

i throttled
s To wh a t d gree the va riou velocities a d a ngles i
. e s n n

flu e e the he a d by decre s ing the qu a ntity of wa ter w i ll be i

nc a n

ves tiga te d in the follo w ing I order to obvi a te the cons idera tio n . n

of shock losse etc occurri g through throttling as ume pum p

s, .
,
n s a

w ith guide d directing v ne which ma y be ch nged a d which

an a s a n

a re a lw y correct for the e i ting qua ntit ie of w ter s th a t the

a s x s s a o

coeffi ci nt 7 m y be t ken of ppro i ma tely equa l v lue for va riou

e 7 a a a x a s

qua tities of wa te r
n .

I the follo wing di scuss io

n i t i be t to ta rt w ith the firs t n s s s

funda ment l equ tion equa tio 10 wh ich m a y be w ritten if

a a ,
n , ,

(1 + p ) 7 ( equ tion
a as
7 a

The indi vidu l member u pon the left h a nd side of the equa tion
a s -

w ill h a ve di ff re t influenc upon the v lue of the he a d as th e

e n es a

qu ntity of w t r c ha nges
a a e .

2
“0
The fir t member s ,
2a
wi ll rema in cons ta nt a n d independen t
of the qua ntity of wa ter beca use the number of revolutions remain s
the sa me .

The second member 2 c a n influence the hea d va riously ,

9
wi th ch an ge of qua tity Accordingly as this member wi l l
n .
 R E G ULAT ING Q UAN T IT Y OF WAT ER BY T H ROTT LING .
1 15

be po itive or nega tive ; for v

s it is zero a d therefore h as n o . 0. n

influence on the head .

The case v v in which the member 2 . i po s itive w ill fi rs t

. s
g
be cons idered Referring to Fig 9 the wa ter when flo wing from
. .
,

point t impeller entra nce to point t i mpeller e it will ha ve i ts

e a a a x
’
v
pressure incre ed due to the influence of the va lue 2
as ,
.

9
the velocity 0 decre e to a velocity v If the qua ntity of wa ter
. as s . .

now decrease t h v locitie v d will a lso decreas a n d the pre

s, e e s . an 0. e ,
s

sure hea d g ined by the cha nge of velocity from p o int to will
a e a

decreas e corre pondingly s .

Therefore if v v the memb r 2 .become ma ller with . e s s

9
dimini hed qua ntitie of w ter a d hence ca uses a decrease of hea d
s s a n .

If v v a ce rt in pre ure he d i necessa ry i

. . order to a c c e le r
a ss a s n

a te the velocity v to the l a rger velocity v The di ff erence of these

. . .

pre ure he d must to pe k b ta ke from t h pre ure e n s uing

ss a s , so s a , e n e ss

through the a ction of the centrifuga l forces The m ller the qua n . s a

tity of w t r the ma ller will be the va l ue of t h pre ure hea d e

a e s e ss x

tr cted to overcome the velocity increas e from v to v Hence if

a . ..

2 2
v
the me mber 2 0.
i neg tive it hel p to incre e th e he d when
.
s a s as a
g
the qua ntity of wa ter i decreas ed s .

Fo the specia l c as e
r v the v lue 0 will h a ve no 0. . a
29
influence upon the hea d provided the friction hea d neces a ry to over s

come th lo e of the wa ter a long the pa th from c to i neglected

e ss s a s .

2 2
v
The va lue a
0.
be tra n formed by ubstituting v
.
c
if
i n s s .
r

S 6

an d 0. so th a t

2
The third member of equa tion 22 1 is g the velocity he a d of the i
g
-
i

a bsolute exit velocity The influence of thi member upon the he d . s a

when the qua ntity of wa ter is cha ged c a best be seen from Figs 6 2 n n .
,

6 3 an d 6 4 which represe nt exit veloci ty di a gra m for 6

, ,
s .

6 . d 6 Ifanthe an gle formed .between w a n d v . .

1 16 REGULAT ION AN D C LASSIFIC AT ION .

isobtuse for the norm al dia gra m (thi s being the di a gra m for norma l
loa d s e Fig ,
the velocity w wil l increas e with decreas ing
e . .

qu a ntities of wa ter O the other ha nd if the an gle is a cute (see

. n ,

Fi g 6 2
. .

Fig . will decrease with decreas ing qu ntity of wa ter If

w. a .

the an gle 6 is not much grea t r tha n

. the c as e ma y a rise in which e

w fi r t decre ases to a minim u

. s m va lue when w i perpendicul a r to . s

v a d then a g a in incre
. n e I ge era l it ma y be sa id th t wi th as s . n n a

6. 90 the velocity
“
w incre as es wi th decre as ing qu a ntitie
s of .

wa ter while with B 90 it dec reas es

°
, . .

Fig 6 3 . . Fi g 6 4. .

Fo r 6. 90
°
( se e Fig 64 ) w.
. will decreas e very li ttle as the qua n
tity of wa ter is decreas ed .

The eff ect of the membe r 3 i uch therefore th a t for decreas ing
3 9
s s
5
, ,

qu a titie of wa ter the hea d will increas e if 6 90 an d decreas e

n s .
°

if 6 .

The following equ a tion m y be written from the exit velocity a

di a gr m a

2
Thet member on the left h a nd ide of equa tion 22 1 is
l as -
s

If th a ngle 6 90 for the norm a l di a gr a m then the a ngl e

°
e . ,

formed by w a n d v become an cute a ngle ; o th t w ith decreas ing

. . s a s a

qu a ntities of wa ter w firs t decreas e to a minimum va lue when w . s .

is pe rpendicul a r to v a d then increas es Fo reason a lrea d y

. n . r s
118 R E GULATION AN D C LAS SIF IC ATION .

wa ter otherwise the hea d can be increas ed only by increas ing the
,

number of revol ution the la tter being an i mpos ibility in mos t ins ta l
s, s

la t io ns If it i d ired to Obt a in the norm a l he a d over a wide r g e

. s es an

by mea of throttling the cons ta nt C mus t be ta ken nega tive

ns , . .

This i poss ible o l y when the a ngle 6

s 9 0 an d fu ther w hen the
n .
°
r

u u F
member t F
Hence the a ngle 6 a d 6 mus t b
. . “
s . n . e
t an Ba an fi e e

ta ken as la rge as po ible a d the ra t o as much ma ller th a n uni ty

ss n i s

as po ibless .

I order till to obt in the n orma l des ired he a d with de re ing

n s a c as

qu a ntitie of w ter the ba ckwa rd bent bl a de with 6

s a 9 0 mu t b -
.
°
s e

u ed I the previou rti le this bl a de form h as b en proved to

s . n s a c e

be the mo t effi ci nt B l de bent forw a rd

s pplic ble only to
e . a s a re a a

the ra re ca e in which it i de ired to obta in decreas ing he d w ith

s s s a

decreasing qu a ntity of wa ter .

I the bove di cu ion of the perform a nce of centrifug l p umps

n a s ss a

while throttling the umption was m de th a t to secure op ra tio as s a e n

w ithout hock di ff erent guid v a nes w ere to be u d co r ponding

s e se r es

to the diff erent qu ntiti of wa ter for one a n d the a m pump a n d

a es s e ,

therefore the coe ffi cient 7 w umed a pproxim tely con t nt 7 as a s s a s a .

I Fig 7 2 th eo retic l pre s ure curv

n . a hown they w ould a s es re s as

be derived for the ex m ple l ul ted in Article 3 1 The e c urve a ca c a . s s

a e of cour e
r tra ight li in ccord ce with qu tion 22 8 Th
s s n es a an e a . e

idea l curves s deriv d c nno t b obt ined in pra cti e b c u e the

o e a e a c e a s

pump h a s guid v nes th dim ion of which e o dete mined

a e e ns s a re s r

th a t the pum p oper te t norm l lo d without hock It i of a s a a a s . s

cour e impo ible to ch g the guide v

s ss to uit the dimini h ing an e an es s s

qu ntiti of w t r Wi th incre e or decreas e of the qu ntity of

a es a e . as a

w t r ho k
a e d d d y curr t lo
S c occur a t entr nce to d e it
an e -
en sse s a an x

from the impell r T h va lue of the e lo c not be de termi ed

e . e s ss e s an n

by c lc ul tion b ut only by experiment Only from eri of uch

a a s . a s es s

e perim nt i it po ible to bta in a cle r conc pt ion to the v a lu e

x e s s ss O a e as

of th e lo e es ss s .

It i not po ible for th e a uthor to give the r ult of uch e xpe ri

s ss s es s s

m e nt The experim t l r ul t
s . pl tt d in num rou curv es uch en a es s a re o e e s s

as a hown for i t nce in Fi g 56 to 6 0 B ut in orde to di cuss

re S ns a s . . r s

these c urve the exa ct d ta as to th d ign d dim n ions of t h e

s a e es an e s

pumps a nec a ry The r as o w h y thi i form tion is withhel d

re e ss . e n s n a

by m nufa ctur
a b r dily u d r tood e rs c a n e ea n e s .
 AD JUST ABLE GUIDE V AN ES . 11 9

It is only po ss ible to sa y th a t the pum p from whi ch the e xpe ri s

menta l da ta here given were deriv d h a ve nega tive va lues for the e

consta nt C Further the e fli i y curve how tha t the tota l

g. ,
c enc s s

e fli c i c y becomes less w ith decre as ing qu a ntities of w a ter

en O the . n

one ha nd thi diminution i due to the pump opera ting with shock
s s s ,

but on the other h a nd it i due to the fa ct tha t the hort ircuit a d s s -

c n

the mech a nica l lo e a e a pproxima tely equa l for a ll qua ntities of

ss s r

wa t r so th a t for m a ller qu a ntities of w ter the pe rcentages of loss

e ,
s a

increas e when referred to the tota l po wer cons um ption Fo r in .

s t a nce if the co n s umption of po wer necessa ry to overcome these

,

losses a t norm al loa d is 7 per cent of the to ta l power then it will

a , ,

t k en a b olutely rema in a pproxim a tely the sa me for h a lf the qua ntity

a s ,

of wa ter but refe red to the consumption of po wer n o w occurring it

,
r

w ould a mount to a bout 14 pe cent The decided fa lli g ff of the r . n o

effi ciency curve with decreas ing qua ntities of wa te r is due chiefly to
thi fa t
s c .

It should be noted th t the equ a tion derived for the rel a tion ofa s

press ure he d when throttling a gree with pra ctice only within m a ll
a s

limi ts of ch a nge of wa ter qua ntity Fo very sma ll qu a ntitie of . r s

w a ter a d by complete thro t tling the he ad ris es to a gre a ter v a lue

n ,

th a n the c lcul a tion giv a Fo compl te throttling thi c a be

es . r e ,
s n

u u 2 2
expl ined by the fa ct th t in det rmining the v lue
a the a e a
9
velocity u no lo ger refer t the di am eter D but to the out r
. n s o . e

impeller di ameter

G U IDE VAN E S 29 . ADJU STAB L E .

When the qua nti ty of wa t r p ass ing through an impeller is ch a nged e ,

the guide va ne a g le or a gle of lope of the involute will no lo ger

n ,
n s ,
n

correspo nd to tha t a ngle w ith w hich the wa ter lea ves the impeller ;
th a t is the direction of the a bsolute velocity of the wa ter lea vi g the
,
n

impe ller ch a nges with the qua ntity of w a ter flowing Fig 6 1 illus . .

tra tes h o w lo ses occur due t this ch a nge of qua ntity of w a ter A
s o . n

idea l condition would occur if the guide va ne could be a djusted o s s

tha t the ch a nnel cro ctions d va ne a ngle would corre pond to

ss s e an s s

the qua ntity of wa ter flowing This would give opera tion without .

s hock w hich,
imply me ns th a t for a y cert in qu ntity of w ter
s a n a a a

the correct velocity di a gra ms will correspo d to a cert in etting of n a s

the guide va ne Thi however i h a rdly pr ctic a l

s . s, , s a .
1 20 RE GULAT ION AN D C LASSIFIC ATION .

There is a form of a dju ta ble guide va ne which is genera lly use d s

for regul a ting wa te r turbine a d w hich could be used in centrifuga l s n

pumps The guide va ne an be so a djusted th a t the a ngle 6 n d

. s c . a

the cros s ection of the guide ch a nnels ma y be increase d or decre ed

s as .

The guide va nes a r o a rra nged tha t ea ch va ne c an be imulta neou ly

e s s s

tu ned by a suit a ble mecha n ism controlled outside of the pump as

r c

ing By the ue of uch a djusta ble guide va nes it i po ible to

. s s s ss

ch a nge the directing a ngles an d entran ce heights so th a t they cor e r

spo d to a cert ai n qu a ntity of w a ter

n .

Adju t ble gu ide va nes c a nnot be used for high press ure centrifuga l
s a -

pumps bec u e the guide cha nnel height a d cro ss sections are qui te
a s s n

s ma ll a d l o beca u e a dj us ta ble guide va nes a e d iffi cult to in s ta ll

n a s s r

in a multi ta ge d ign I l w pre ure pumps where l a ge cross

s es . n o -
ss r

se ctions occur there is no reas on wh y a d ju t a ble guide v a ne s c a nnot s

be used The only reas on wh y such regul a tion h as not been a dopted
.

i on ccount of the gre a ter cos t of con truction

s a C omm erci a lly it i s . s

diffi cult to h a ve a costly pump competing with che per types even a ,

if the former posses e some a dva nta ges a n d especi a lly whe the s s ,
n

gua ra nteed effi ciencies a t norma l loa d are the sa me B ut with the .

con ta ntly increas ing ue of centrifuga l pum ps the e w ill be man y

s s r

dema nds for a high effi ciency Should such high effi ciency be de .

m an d d with v a rying qu a ntitie of wa te r ad just a ble guide v a n e s

e s ,

mu t come into use in pite of the higher cost This a pplies more
s s .

e peci a lly t lo w press ure pumps

s o -
.

Fi g 6 6 sho ws a genera l rra ngement of a s cheme for a djus t a ble

. a

guide va ne I th is c ase the guide va nes c a be turned a bo ut the

s . n n

so c a lled gu
-
ide va ne bolt w hich a l o incident a lly sti ff en the ide s a, s s s

of the pump c as ing A m ll lever b ri gidly connects ea ch va ne to a

. s a

ring which i guid d by a pa rt of the pump cas ing a d which c an

0, s e n

be turned by ome uita ble mech a ni m Every position of the ring

s s s .

corr ponds to a cert in po ition of ll the guide v ne a d as suming

es a s a a s, n

c r ful as mbly y position of the ring will lwa ys give uni form
a e se , an a

e ntr nce h ight a d entra nce a ngle throughout

a e s n s .

I nother a rra ngement the a dju ta ble guide va nes are conn ecte d
n a s

ri gidly to the bolt The l a tter a e c rried through stuffi g boxes

s . r a n

in the cas ing The bolt conn ected to ma ll leve which ar

. s a re s rs e

a ctu a ted by a ri g upport d on the out ide of the cas ing This
n s e s .

ri g i l o tur ed by a uit ble mech ni m Although thi a

n s a s n s a a s . s r

ran gement is more costly th a n the former it is to be preferre d

1 22 RE GULAT ION AND C LASSIFIC ATION .

an d negligible error of a ngle be ca use with sma ller entra nce heights ,

the involute should ha ve a different rad ius of curvature .

When the entra nce cross section of the guide va nes is decreased
the head necessa y to overcome the friction a d Shock l osses i
r n n

creases With b a ckwa rd turned bl ad es o w genera lly used the

.
-
,
n ,

a bsolute velocity incre as e w ith decre as ing qu a ntities of wa ter ; a l o

s s

b at
with decreas ing height a the va lue 2 l) increases so tha t a ccord.
al
,

ing to equa tion 194 the fri ction los in the guide va nes incre as es s

with decreas ing qua ntities of wa ter With decreas ing heights of .

guide ch a el the ra tio of contra ction becomes l a rger on a ccount of

nn

the thick ess of the va ne According to equa tion 2 12 the hea d

n s .

lost through bl a de shock will increas e with decreas ing qua ntiti es of
w a ter bec a u e the w a ter velociti e incre as e
s s .

Fo r con tructive reas on the directing v a ne a t im pe ller entra nce

s s s

ca n not well be ma de a djusta ble A was mentioned in Article 28 . s ,

the losses due to cha nge of qua ntity of wa ter will not be very grea t
if the clea ra nce be tween va nes a d impeller i ma de suffi cient n s .

30 . C L ASSIFIC AT IO N OF C E NTR IFU GAL PUM PS .

When m a nufa cturin g centrifugal pum ps it becomes nec ess a ry to

decide quickly whether a y e i ting des ig s of impeller a n d gui de n x s n

v a ne c an be us e d for certa in s t of condition s In thi a rti c le

s a e . s

it will be shown h o w centrifug l pump c a be cl ass ifi ed by certa i n a s n

cons ta nts .

Le t u s ass ume th a t serie of pump h as been desig ed for va ri

a s s n

ou s he a ds d qu a ntiti s f w a ter a d th a t the pump h a ve be e n

an e O n s

found sa ti f tory by e p rim nt It i w necessa ry to sel ec t one

s ac x e e . s no

of th e e pump or to d ig
s w one to mee t a cert a in se t of c o n
s es n a ne

d it io s
n B efore d i g in g a
. w pump w e mu t decide w h e ther w e
es n ne s

can utili e old p a t t rn e peci lly tho e of the impeller a d the

z e s, s a s n

guide va nes without cri ficing t h effi i ncy of the pump

,
sa e c e .

Of course a t fi rst we inve ti g te wh e ther the entire model of s a

some xisting pump c a be u d Thi i b t ccomplished by

e n se . s s es a

determining the consta nt de ig a ted in Article 6 as the ch a ra cter s n

isti c K Accordi g to equ tion 49 K n 9 Fo multi s ta ge a ,

r
V iH
.

i n

pumps H is of course the hea d of one t ge a d in c a lcula ting such

.
s a ,
n
 C LASSIFIC ATION OF C ENT R IF UGAL PU M PS . 1 23

pumps it is nece a ry to determine immedi a tely the number of stages

ss .

The a me pump c a be used for all combina tions of hea ds an d

s n

qua ntities of wa ter for which Q h the sa me v lue hence the as a

Vn
,

sa me ch a ra cteri tic K This va lu K i a common consta nt for

s . e s

impeller n d guide va nes In order t obta in the v lue 7 H the

a . o a 1

coe ffi cient ; mu t first be as sumed Fo Q the qua ntity of wa ter in

1 s . r

cubic feet pe minut mu t be bt ined r e s O a .

The number of revolutions per minute of a pump va rie as the s

square roo t of the tot a l he d delivered by th a t pump a d therefore a ,

n

the sa me impeller a be u ed for diff erent hea d by imply ch a nging

c n s s s

its spee d I order to de t rmine quic kly the number of revolutions

. n e

per minute nece a ry for a y tot l hea d it i be t to de igna te ea ch

ss n a s s s

de ign of pum p by the number of revolutions pe minute correspond

s r

ing to some fixed tota l hea d The mo t conve ient v lue of thi . s n a s

fixed tot a l he a d 7 H i ten fe t a d t hi v a lue will be us ed in the

1 . s e n s

following problems Th t i y de s ign of pump . w ill be cl ass

a i fi ed s, a n

under the numb r of revolution per min ute wh ic h is necessa ry

e s .

to give a to ta l he a d a H equa l to 10 feet . .

The pump mus t a l o be cl as si fied a to it importa nt dimens ions

s s s ,

a n d it i be t to cl ass ify it in te rms of its outer impeller di meter

s s a

D expre d i n feet
“
. ss e .

A h been expla ined in Article 15 the sa me impeller n b

'

s as , ca e

u ed for di fferent qua ntities of w ter provid d the h d rema in

s a , e ea s

con ta nt by cha nging the guide va nes to uit a y given qu a ntity

s , s n .

H ence by ch a nging the guide v a nes v a rious co t a nt K a oh ns s re

t a in e d for the sa me impeller When the qu a ntity of w ter i ch a nged . a s

the periph ra l velocity a l o cha nges on a ccount of the influence f

e s o

the ngle 6 a d 6 hence the number of revolution per minute

a s . n s

ch a nge It i therefore poss ible to ca lcula te for ea ch w con ta nt

s . s ne s

the corresponding number of revolutions per minute referred of

course to a tota l hea d of 7 H 10 feet 7 . .

I l ay ing out a new de i gn of pump the con t a nt K a d the n u

n s m s n

be r of revolution determined for 7 H s 10n B e ide thi it

a re 7 . . s s s

i well to determine immedi a tely b t ween w h t limit the qu a ntitie s

s e a s

of wa ter may va ry for thi im pe ller by i mply ch nging the guide s s a

va nes Thi lea d to the determina tion of

. s s an d a d fro m n

them the va lues of a d be determined n ca n .

The grea t t d l e ast qu ntitie of w ter th a t c a be used with

es an a s a n
1 24 RE GULAT ION AND C LASSIFIC ATION .

good effi ciency for an y one impeller a re most s imply dete rmined by
dra wing the entra nce velocity dia gra m In this dia gra m we ass ume ,

wh a t ch an ge of a ngle 6 i permiss ible an d fin d the velocity w

”

. s .

corre ponding to the limiting va lue of

s Fro m thi s veloc ity th e s

qua ntity of wa ter ma y be read ily found The li miting va lues o f .

the a ngle 6 a e dependent upon the impeller a ngle

. r In the c a l
cu la t i o n s of a centrifuga l pump given l a ter this will be more fu lly
described .

Fo a n exi sting impeller pro file the periphera l velocity u is c a l

r .

cu la te d a ccording to equ a tion 14 1 a n d when th is val ue of u is s u

, b .

60 u
stituted in the equ a tion n the nec ess a ry revolutio ns per
.

D . ir

minute are found from the equa tion

F. D .

60 Fa t an Ba D t an Boo

D. 1r D
'

2 1 I D
.

then equ a tion 22 9 ma y be wri tten in a simplified form as

n —
w, A V w. 2
A +q 2
H.

To determine the exi ting ra di a l entra nce velocity w the entra nc e

s .

a re a F mu t be known an d then the velocity c an be fo u

. s ,
nd fro m t h e
Q
1

rel a tion w F .
 IV . T H E CALC ULATIO N OF IMP ELLER AND
GUIDE VAN ES .

31 . IMPE L LE R C H AN N EL SE CT ION S O F T HE ORDINAR Y TYPE .

(a ) Ca lcula ti on of
. Two H i gh press ure Ce n tri fuga l Pump
-
s .

In the preceding a rticl it was shown h o w to determine the

es .

entrance an d exit velocity di a gra ms for centrifuga l pump so t ha t a

it would work without hock an d a ls o h w the beginning a d end

s ,
o n

of the bl a des should be formed from i volute curves In the fol low n .

ing it will be hown h o w the impeller a d guide v ne cha nnel sho ul d

s n a s

be de igned The best wa y to do this is a ctu a lly to c a lcul a te sev eral

s .

p ump s a d fin d the cronss ection of the ch a

s nnel s .

First t a ke the Simples t c as e in which the entra nce width b is p a ra llel

, .

to the a xis of the sh a ft Show in Fig 6 7 If cross sections a re

as n . .

taken in pl a ne a a an d b b t right a ngles to the sh a ft axis the

s — —
a ,

se ctions will ll be the sa me a d therefore the bl a des a e a t ri g ht

a ,
n r

a ngles to every hori ont l cro z ection The cha nnel in thi as e
a ss s . s c

is very s imply determined bec a u e the end s of the bl a de a t xit

,
s s e

a n d entra nce a re i nvolutes a d they a re connected by a s uit a ble n

curve so th a t the ch a nnel h as a ,

good form I doing thi it m us t . n s

be borne in mind th a t the num be r

of bla des Should be as ma ll as s

i possible “
f .

Severa l exa mples will n o w b e

c alcul a ted a d the construction o f n

the bl a de cons idered The fi rs t s .

Fi g 6 7
e xa mple will be for a n impe ller
profile as Shown in Fig 6 7
. .

. .

EXAM PL E I See Pl a te III

. An impeller a d guide vane are t o
. n s

be designed for a multis ta ge high pressure centrifuga l p ump to me e t -

the following conditions :

H. 65 feet .
Q 70 cubic feet per minute .

12 6
 IM PELLER CHANNEL SECT IONS OF OR DIN AR Y TYPE .
1 27

1
The hydra ulic effi ciency w ill be as sumed as so tha t 71 H.
;
7

The quan tity of wa ter deli vered per second is Q £8

cubic feet .

The qua ntity of hort circ uited wa ter will be a umed as 5 pe r

s -
ss

cent of the qua ntity delivered o th a t the qua tity of wa ter passing ,
s n

through the impeller an d u ed in the c lc ula tion mus t be s a

Q
’
g cubic feet .

The Impe lle r . The di a meter of the suction pipe i ta ke as s n

6 inche (D s . foot) The hub of the impeller is 3 % inches in

.

di a meter (d . foo t) The clea r cro s ectiona l a rea F a t

. ss .
’

impeller entra nce is

F.
’
( D .
2
s qu re
a fe e t ,

so tha t the velocity w through this a rea .

’
is

Q
F
feet per econds .

the out r impeller di met r be 12 inches ( D

Le t e 1 foot ) a n d
a e .
“

t h outer im p ller
e glee an

L t the number of bl a de z
e 10 s .

According t equ a tion 1 6 8 the di meter of the bas e c ircle for the
o a

involute a t impe ller exit is

d D si foot
. .
“
n ,

an d a ccording to e qua tion 16 9 the dimen sion a . s. is

a. s. foot ,

The bl a de thickness is inch ( foot) ; therefore the height 3.

of the i mpeller ch el fo t
an n a. o .

The impeller dia mete r D upo wh ich the c a lcul a tio o f t he periph
., n n

era l velocity is based c a o w be determi ed ma thema tica lly by

,
n n n

equation 17 2 as

This sa me va lue c a n a s l o be fo und gra phica lly by dra wing a ta ngent

1 28 C ALCULAT ION OF IM PE LLER AN D GUIDE VANES .

to the base circle an d la ying o ff upon thi s ta n ge nt from its point o f

i nte rse ction with the circle of di a mete r D the dis t a nce 2 as can b e .
“ 9

2
eas ily understood by referring to Pla te III .

The impe ller an gle a t D which is the a ngle used in subs eq uent
.,

ca lcul a tions is found from equa tion 1 7 0

, ,

D
“
. si n

D .

“
6 . 151

By uming a n impeller width of .g inch (b

ass fo ot) . , th e
crosss ection a l area a t exit from the impeller a t di a meter D is .

1r Db . . s qua re foot .

This exit a rea i not a clea r a rea beca use the bl a de thickness
s ,
h as
not been ta ken into a cco unt The a ctua l clear a rea wo uld be
.

an

a. 8.
s qua re foot .

Fur ther the rel a tive exit veloc ity v through this a rea is
, .

feet per second .

The i nner impeller di a meter was dete rmined as inches (D .

foot ) Thi dia meter is fixed pa rtly by the dia meter of sh a ft

. s

necessa ry a d the thicknes of impeller hub a n d pa rtly by the proper

n s

opening nece a ry to a dmit wa ter to the impeller By as sum ing

ss .

th a t 6 6 th t i w v the periphera l velocity u is found fro m

. a s, . ., .

equa tion 136 which is as follows

,

2 t a n 6.

from which
i t. feet per second
1 30 C ALC ULATION OF IM PE LLER AN D G UIDE V AN ES .

expla ined on page 59 The a ctua l cons truction of this gra phic a l
.

method i shown in the entra nce velocity dia gra m Fig 69 a d

s , .
,
n

ha rdly needs furt her expl a n a tion It was found tha t .

w . feet per second w v feet per sec ond

, . . ,

a . foot or inch 8 foot or inch , . ,

°
30 60 22

Fig 6 9 . .

To check the g a phica l determina tion the va lue of w c a n be found

r .

2
a.. 8“
from the equa tion w . fee t per second .

The di a meter of the b as e circle of the entra nce involute c a n be

determined from equa tion 16 2 or perha ps more simply from eq ua tion
1 60 as

foot or inches .

The entra nce width b . at the inner im peller dia me te r D . is

foot or inch .

Afterdetermining a ll the impeller dimens ions given abo ve the ,

involutes a t entra nce d exit of the impeller a e dra wn as des cribed

an r

in Article 1 7 The two involutes a e then conn ected by a suita ble

. r

curve This i b st a ccomplished by dra wing the entra nce involute

. s e

on tra cing pa per a d moving it in rel a tion to the exi t involute until
n

both involu te a be connecte d by a curve which a ppears s ui ta ble

s c n

for the form of the bl a de ch a nn el .

I determining the impe ller profile c a re mus t be t a ken to ad he re

n

exa ctly to the width b fe e t a t di a meter D an d the wi dth

. .

b. fe e t a t di am eter D ..

Th e elbow fo ming the entra nce to the impeller must be so formed

r

tha t every cross section of this elbow will ha ve the area so th a t

 IM PELLER C HAN NE L SEC TIONS OF ORD INAR Y T YPE . 1 31

th ere will be no a ccelera tion or reta rd a tion of the wa ter flowi ng

through the elbow .

The Guide Va n es Th e di a meter d of the b as e circle of the i

.
. n

volut for the guide va nes is determined from equa tion 174
e

d; D4 8 111 04 .

It was foun d tha t

D. foot a nd 6.

so tha t
d. foot or inches .

After uming the number of va nes

ass z. 8, the va lue of a. s . is

dete rmined from equa tion 1 7 5 ,

a. 3. foot or inch .

The bl a de thickness 3 is as sumed as foot inch ) ; so

.

tha t the height of the cha nnel foot i ch ) a. n .

The va lue of an a l o be found from equ a tion 1 7 6

a. 3. c s ,

l
I

wil $
a. 8.
)
in w hich w
z.
(equ a tion The .
’

a. 8.
.

qu ntity of w ter Q flowing through the impeller does not flo w

a a
’

through the guide v ne bec u e the va lue Q includes the qu ntity

a s, a s
’
a

of short circuit d w te r If it i ass umed th a t the dire tion of the

-
e a . s c

w ter velocity in the clea ra nce p a ce de cribe

a involute then the s s s an ,

velocity w decreas es proportionally w ith the qua ntity of w te r o

.
’
a ,
s

velocity
necessa y in the a bove ca lcul a tion to ta ke into a ccount the decre ed
r as

qua ntity of wa te r due to the a bsence of the hort circuited wa ter S -

an d i ts co re pondingly decre as ed velocity a d w

r s ma y us e quation , n e e

1 7 7 which includes the qua ntity of wa ter Q a d its corre ponding

’
,
n s

velocity w .
’

The a ngle of the guide va ne a t exi t depend ! upon the h pe of s s s a

the cas ing urrounding the va ne If the c as ing i spira l in form the
s s . s

guide va ne a re give a n exit a ngle t ng nt to the involute of the

s n a e

cas ing If the guide va ne

. pl c d concentric to a circul r s a re a e a

c as ing as shown in Fig 7 it i dvi a ble to ma ke the e it of th

,
.
,
s a s x e

van es ra di a l so th a t the w ter will enter the c ing ra di a lly If the

a as .
1 32 C ALC ULATION OF IM PE LLE R AN D GUIDE VANIB '
.

w a ter a fter lea ving the guide chan n els is led through a connecting or
reversing ch an nel to a nother impeller as is the case wi th multi ,

s t a ge high press ure centrifuga l pumps

-
the exi t a ngle of the guide
,

va ne i m a de such th a t the wa ter lea ving the guide cha nn el flows

”

s s

in the direction of the conn ecting ch an n el The l a tte r is so formed .

th a t the a bsolute entran c e velocity to the next impell er re ceives its

proper di rection .

It would be a d vis a ble to check the c a lc u l a tions of the vel o c ity

di a gra ms by inves tiga ting whether the va lues of the velo cities j us t
found ful fill equa tion 22 which sta tes tha t ,

u.
”
u
.
’
v.
2
v.
”
w. ’
w. ’

+
20 2g 29 2a 2a 29

As it was ass ume d tha t 6 . it follows tha t v . w. therefore

Substituting the ca lcul a ted velocities in equa tion 22 we fin d ,

1) H. fe et ,

w hile the total hea d upon which a ll the c alcul a tion s were bas ed w a s
a l so fee t .

A further check upon th ese cal cul a tions an d a lso a n inves tiga ti o n ,

of the ch a nge of hea d when the pump is throttled c a be obta i ne d , n

from equa tion 22 5 which ta tes th a t

, s

u I

u 3
S

The va lues of the consta nt s a re

C.

Fo rthe normal qua ntity of wa ter delivered v .. fe et pe sec ond ;

r

hence
7 H feet 1 . ,

w hich checks a pproxima tel y with the initi a l va lue .

1 34 C ALC ULAT ION OF IM PE LLER AN D GUIDE V AN ES

The impeller an d guide va nes of the a bove pump c a n therefore be

used wi thout cha nge for all cas es in which the ch a rac teristic K h as
the va lue
By dra wing velocity di a gra ms it is found th a t the s a me impe ller
ca be use d for a ma ximum ca pa city of 8 5 cubic f et p minute d
n e er an

a mi n imum c a pac ity of 55 cu bic fe t p minu te wi thout grea tly d e er e

c e as ing the hydra ulic effi ciency ; the only cha nge neces a ry be i n g
r s

the redesign o f the guide va ne which m ust b m de uita ble for s, e a s

ea ch quan tity If the coeffi c ie t ? rem i n s cons ta nt we ha ve

. n 7 a

To dete rmine the n umber of revolution for a n y va lue of K the s

con st a nts A an d B a e found from equ a tions 230 an d 23 1 I th i

r . n s

exa mple they h a ve the following va lue :

B 850 .

The number of revolution for a y qua ntity a s n nd hea d of w t r a e can

n o w be fo und from equa tion 2 32 which i , s

To determine the velocity w the entr c . an e a rea must be give

F. n ,
!

a d then w i found from the equ tion w

n s a
Q To clas s ify the c l
. . a
F
cula t d pump exa ctly we should w ite
e r

H igh pre ure pump

-
ss ,
D
K n 53 0 .

K
F. A B 8 50 .

The con t a t A h as quite a sm ll va lue so th a t the number of

s n a ,

r volutio ch ng v ry littl with v rying velocitie w I the

e ns a es e e a s .. n

next a mple thi

ex co t t h s l rger v lue a n d the qu a ntity of
ns an as a a a ,

w t r w ill then h v e
a e import t influe ce upon the number of
a an an n

revolutions .

It w as found in the bl a de de si gn for e a mple I th a t the he a d would x

decrease by throttli g th t i by d creas ing the qu ntity f w a ter

n ,
a s, e a O .

The que tion o w a ri e If a oth r bl de form i t ken how c a n w

s n s s, n e a s a , e
 IM PE LLE R C HANNEL SECTIONS OF OR DINARY T PE Y .
1 35

be sure tha t the pump will deliver a t leas t the norma l pressure when
it is throttled in order to va ry the qua ntity of wa ter delivered
In Art icle 2 8 this question was b riefly dis cussed
,
Attention is .

a ga in c a lled to a point which concerns the choice of the outer a n d

inner di a meters o f a i mpeller I order th a t n H ma y increasen . n .

with decreasing qua ntity of w a ter the right h a nd p a rt of the se ond -

c

member of equa tio 22 5 mus t be nega tive The l rger thi negan . a s

u .
2
u 2
.

neces sa ry to mak e the member as l a rge as po ss ible w hich ,

9
is ac complished o n the one h a nd by a s uming l a rge value for the s s

a ngles 6 a n d on the other h a nd by a suita ble as sumption of the

.

inner an d outer impeller di a meters Pla cing u g in equ a tion

:
.
.

22 6 we h a ve
,

D
H ence the sm aller the ra tio 7.
1
.
the l a rger the constan t C1 an d the
higher will be the hea d w hich the pump c a deliver under t h rottling n .

If the inner di a meter f the impel ler of the a bove exa mple i O s

de re d the ntr nce line b

c as e e no lo ger be kept p r llel to th
a . can n a a e

a xis, but mu t be t ken s curve omew here in the entr nce elbow
a as a s a

of the impeller The determin a tion of the proper bl a de surfa ce or

.

s h a pe w ill be di s c ussed i n the followi ng article .

D
Another wa y of decre as ing the ra tio is by increas ing the outer
.

75 .
di a meter I order to Show wh a t influence an increas e of the
n

outer di a meter of an impeller wi ll h a ve on the va rious ca lcula ted

va lues a second bl a de form will be c a lcul a ted for the a me exa mple
, s

but by ass umi g the outer impeller di a meter D

n feet .
“

inches ) .

E XAM PLE II See Pl a te IV The outer impeller di a meter D “

. .
.

feet The impeller profile a t entra nce h a ll be the sa me as

. s

in exa mple I th a t ,
so

D . ft F sq ft
.
,w ft
.
’
D ft . .
, ,
’
.
, . .

The number of impeller bl a de will o w be z 9 I the fi r t s n . n s

exam ple the number of bl a des could not be ta ken le th a n 1 0 b c u se ss ,

e a
1 36 C ALC ULATION OF IM PE LLE R AND GUIDE VANES .

otherwise the entra nce a n d exit involutes coul d not be c onn e c te d .

With a l a rger outer di a mete r the cha nnel heights c an be ta ken gre a t er
on a ccount of the correspondingly grea ter di fi e re n ce of D D a n d . .,

therefore the number of bla des c an be dec reas ed .

We sha ll a ga in c a ll a ttention to the fa ct th a t the min imum numbe r

of impeller bl a des is determined by the poss ibil ity of con ne cting the
entra nce a n d exit involutes In Article 24 the a dva ntage of obta in
.

in g a mini mum num ber of impeller bl a des was di s c u ss ed .

With the bl a de form h a ving the l a rger impeller dia mete r the
cha nnels w ill of course become somewh a t longer but the height of ,

ch annel becomes l a rger on a ccount of the sma ller number of bl a des .

The hydra ulic losses for both bla de forms will be a pproxi m a te ly
equal Beca use of this the hydra ul ic effi ciency
.

as be fore s o th a t ag a in " H
is ass u me d
5
, .

The oute r impeller a ngle a t di a meter D is Th e n .

“

d. foo t an d a. s. fo ot .

Therefore the heigh t of ch a nnel a foot an d the bl a de th ick

.

ness s . foot .

The oute r impeller di a mete r D upon which most of the impe ll er .,

calcul a tions are bas ed c an be obta ined from equa tion 1 7 1 o r c a n

,

be determined gra phic a lly It was found tha t D . feet a n d .

tha t the bla de a ngle a t this di a meter is 6 15 1 42 .

“ ’

The impeller width b is now ass umed Le t b

. foot ; fro m . .

this we fin d
F .squ a re foot an d v fee t . .

Assuming 6 . 6 the periphera l velocity is found from eq ua

tion 1 36 as

u
. feet per second an d u . feet per sec on d .

The number of revolutions per minute now becomes n 1 3 3 0 .

At the inn er impeller di a meter D the pitch t foot . . .

According to the cons truction now well unders to od we dete rmi n e ,

gra phica lly

a. s. ft .
,
a. ft .
, 3. ft .

6 . 6. w. ft per s
. ec .
, w. v. ft per s e c
.
1 38 C ALCULAT ION OF IM PELLER AN D GUIDE VAN ES .

Fo r the norma l quan tity of wa te r v fee t per se cond fro m . ,

which ” H . as a ga i n s t which was ass umed in the ca lc u

la tion .

The theo retica l pressure curves for the bl ad e forms I an d II a re

sho wn in Fig 7 2 . The a bsciss a are quan tities of wa ter a n d th e
.

ordin a tes the tota l pressure hea ds n H It c a n be see n tha t t h e

s ligh tly l a rger i mpeller used with bla de form II c ons ide ra bly a i ds

in inc reas in g the press ure h ea d .

JO I0 0 I”
a . fl ” ! m in .

Attention migh t a ga in be dra wn to the fa ct tha t the theo retic a l

pressure curves shown in Fig 7 2 would a ctua lly show a diff erent
.

course bec a use the pressures would become l a rger with decreas in g
qu ntities of wa ter But in spite of this the representa tion of both
a .

pre ure curve i of intere t in so far as it gives a clea r conception

ss s s s

D
of the influence of the va rious assumptions of the ratio D upon t h e .

head .

The ch a ra cte ristic of the pump with bl ade fo rm II is found a s

be fore K ,
but the number of revolutions referred to q H .

10 is in thi s c ase n 467 so th a t the pump ma y be classi fied as

,

H igh -
pre ure pump
ss ,
D
K n 46 7 .
 IM PELLE R CH AN NE L SEC T IONS OF OR DINAR Y TYPE .
1 39

As h as been previo usly found ,

K
T he consta n ts A a n d B were found to be
A B 54 1 .

T he necessa ry entra nce a rea for c a lcula ting w i F . s . qua re s

f oot so th a t the e xa c t cl ass i fi c a tion of thi s pump ma y

,
n o w be w ritten

H igh press ure pump

-
,
D .
“

K n 46 7 .

Km a x .
K
F. A B 54 1 .

It is dvisa ble to collect these va lue derived for v a rious pumps

a s

in the form of ta ble a .

Fo r exa mple II the cons t a nt A h as a l a rger v a lue th a n in exa m

ple I The number of revolution corre ponding to the ma ximum
. s s

a d minimum qu a ntitie
n of wa ter (85 cubic fe et an d 55 cubic feet )
s

ca n be found as follows :

The a ngle of the imp ller bl de a t entra nce e must of coursea ,

rema in co ta nt bec use the me impeller i used This m ke

ns a sa s . a s

the v a l ue of v directly proportion l to the qu ntity of w ter p as ing

. a a a s

through the impe ller or di ch rged by the pump F the norma l s a . or

qua ntity (70 cubic fee t per minute ) v feet per econd ; hence . s

for the maximum qua ntity (85 cubic feet ) v feet per second .

a d for the minimum qu a ntity ( 55 cubic f et ) v

n feet p e . er

second As w
. v i the va lue of w a be found for e ch
, . s n , c n a

v a lue of v Fo the ma ximum qua ntity of w a te r w

.. r si 25 . n
“

feet per second ; for the minimum q ua ntity w i 25 . s n

°

feet per second The number of revolutions for ea ch qua ntity

.

c a n now be found from e qu a tion 2 3 2 so th a t the following ma y ,

be written :
Fo r corresponding to w feet 1 3 58 . ,
n ,

Fo r co res ponding to w rfeet 1 3 10 . ,

n ,

w hile the number of revolution for K ( for w hich the guide s

v a nes a d impeller of the pump were d igned ) was

n 1 330 I es it . n

Fig 7 1 a re sho wn the entr nce velocity di gr m for K

. a d K
a a a s n

a d their corre spondi ng velocitie s a e dimens ioned in the fi gure

n r .
140 C ALC ULAT ION OF IM PELLE R AN D G UIDE V ANES .

(b) . Ca lcula ti o n a nd Bla de F o rm f

o a Lo w pressure Cen trifu
-
ga l
Pump .

The bl a de form of a impeller a n d guide va es for a low press ure

n n -

centrifug l pump will n w be des igned The pump is to comply

a o .

with the following speci fic tion a

Q 6 36 cubic feet per minute ,

H. feet .

The hyd ra ulic effi cie cy i umed to be o th t n H

n
5 s as s s a

The qua ntity of wa ter delivered per second is Q cubic fe et .

The a mount of short circuited wa ter is as u med to be 6 pe cent -

s r

of the qua tity delivered o th a t the qua ntity pass ing through the
n ,
s

impeller is
Q
’
cubic feet pe r s econd .

The Impe lle r The uction pipe dia meter D will be as s um d as

. s . e

fe t e inches ) With a hub dia meter of d . fo ot .

( h
3 ?
i chen ) the c l
s ea r cro s sectiona l a rea of the uction pipe a t s -
s

entra nc to the impeller will be F

e qu re feet a n d there .
’
s a ,

fore the suction pipe v locity a t this pl a ce will be w e fee t .

’

p e r econd
s .

Le t the outer impeller di a meter D f et inches) e

a d the impeller a ngle a t thi dia meter be 155 a n d let t h

°
n s e

numb r of impe ller bl a de be z 12 The di a meter of the ba se

e s .

circle for the outer involute is found from equa tion 16 8 a n d is d .

f et By assu ming thickness of bl a de

e . foo t a 3.

inch ) the height of the ch nnel is found from equ a tion 16 9 t a o

be a . foot inche ) s .

The out r impeller dia m ter upon which the c lcul a tion s a r
e e a e

ba d i found from equ tion 1 7 2 to be D

se s feet The pitch
a . .

a t thi di a m t r is t
s e e foot . .

The a gle of the impell r bl de a t thi di a meter i found fro m

n e a s s

equa tion 1 7 0 It i 6 .152 s .

°

Th e friction los e a e much m a ll e r in lo w pre ure centrifuga l

s s r s -
ss

pump th a i high pr ure pu mps bec a u th former h a ve lo w

s n n -
e ss se e

velocities a d l a rger cro s ctio of ch a nnel I order to a void

n ss e ns . n

directing v a nes a t impe ller entr n e 6 i t ken equ a l to s o th a t a c . s a

th e a bsolu te entr a nce v e locity w i s perpendicul r to u . a ..

142 C ALC ULAT ION OF IM PE LLE R AN D GUIDE VANES .

ra phic a lly or a na lytica lly the rel a tive

G an d a bsolute exi t veloci
ties a r found as
e

v feet per econd w

. s , . feet per econd s .

The inner imp ller di a meter was ass umed s D e feet a . .

With z 12 the pitch t foot . .

The periphera l velocity a t this d ia mete r is u feet pe . r

second .

The velocity in the uction pipe a t impeller entra nce w found s as

to be w f et per cond
.
’
e se .

H a ving the v a lues u d t given the entra nce velocity d i., an . , a

gra m an be dra wn as h a lrea dy been expl ined ( e e Pla te V ) an d

c as a s

the va lue a be gra phica lly determined It h as been foun d tha t

a. 3. c n .

a. foot ,
8. foot ,

0. feet per second ,

w. fee t per econd s .

The di meter of the bas e circle of the entra nce involute is d

a .

foot The width b a t impeller entra nce is

. .

b foot ; .

further the clea r entra nce re a t king into ccount bl a de thickness

,
a , a a , is

F qua re foot . s .

Af ter the ca lcul a ted v lues for entra nce to a n d exit from the
a ll a

impeller a e determi ed the i vol ut a dra wn a d the bl a de

r n n es re n

pro file is formed (Se Pl te V ) I dra wing the impeller or bl a des

. e a . n

c re must be ta ken th a t a t the outer dia meter D the width of the

a .

impeller b foot be a dhered to a d a l o t the inner di a me

. n s a

te D the impe ller w idth b

r . foot mu t be kept unch a nged . s .

Further the elbo w t entra nce to the impeller bl a des mus t be so

,
a

formed th t every cro ectio t right a ngles to flo w mus t h a ve

a ss s n a

the va lue F s qu a re feet

.
’
.

The Gu id V T h number of v a ne w ill be t a ken as z

e a ne s . 14 e s . .

The di meter of the base circle of the i volute is fou d from equa
a n n

tion 1 7 4 as
d foot z ,

a n d from equ a tion 1 7 5 the height of ch a nnel a was found to be ,

foot as uming a thickne of va ne

,
s foot ss s. .

Pl a te V Shows the entire form of these v a nes .

 IM PE LLER C H ANNE L SEC TIONS OF OR DIN ARY T YPE .
1 43

The tota l ‘
Co n su
mpti o n o f Power .
y of the p ump will be e fli c ie n c

t a k e n as 7 5 per cent so th a t the cons umption of power as dete rmined

,

f r o m equa tion 19 2 is N h p . .

A a check upon th es e ca lc u
s la tions it would be well to se e if the
c a lc u l a te d velocities ful fill e qua tion 22 1 a n d further in order to , ,

h a v e a check as to wh a t the pre ure hea d will be when the pump ss s

i s t h rottl e d it i s a d vi s a ble to dete rmine the const a nts 0 an d C

, 1 .

fro m equa tion 228 .

T h ese cons ta nts a re found to be

01

so tha t equa tion 228 rea d s

uH . v. .

Fo r the norma l qua ntity of wa te r v which when s u b .

s tituted in the a bove formul gives ; H w hich cor res ponds

a 1 .

l sely with the va lue initia lly ass umed From the va lue of the
c o .

co n s t a nt C it c a n be seen tha t the hea d do e not fa ll below the

1 s

v a lue of the norm a l he d when the pump is throttled

a .

Atte ntion i c lled to the fa ct th t the bove pump c a lcul a tion

s a a a

w as m a de by id of the a a n d A cu rve w ithout us ing

x ta ble of s a

t ri gonometric functions .

I order to cl ass ify the pump the ch a r cteri stic K

n
Q
a
V II H
found an d a s l o the numbe r of revolution per minute referred to s

11 H 10 .

Fi g 7 4
. .

The grea test qua ntity of wa ter w hich c a be delivered by the n

pump a d still reta in good effi ciency was found by dra wing the
n

velocity d ia g a m Kse Fig 7 4 ) to be 777 cubic feet pe minute a d

r e . r n

the sma llest qu a ntity to be 459 cubic feet per minute This gives .
1 44 C ALC ULAT ION OF IM PELLER AN D G UIDE V AN ES .

The velocity dia gra ms for thes e

an d

ma ximum a n d minimum va lue a hown in Fig 7 4 From eq ua s re s . .

tions 230 a n d 23 1 the cons ta nts A a d B a found to be A n re

B
The exa ct cl assi fic a tion of the pump ma y now be written as

follows
Lo w -
pre ure pump
ss ,
D .
“

K
Kn g.

F. A

32 . T H E C H AN N E L C RO SS SE CT IO N S AN D D E V E LO PM E NT OF THE

B LADE E N DS U PO N TH E SU R FAC E OF A C ON E .

( ) a . In Ge n e ra l .

In genera l the weight of a centrifug l pump depend upon the a s

impeller dia meter I order to ve w igh t d a lso to i cre as e

. n sa e an n

the number Of revolution it i dvi ble to ke p t h imp ll r di a meter s s a sa e e e e

a s s m a ll a s po ible Thi i e peci lly tr ue of l w pre ure t ifu

ss s s s a o ss ce n r
-
.

g a l pump s b c a u e
,
l a rge
e impell r di
s m t me n con ider a ble weight e a e e rs a s .

In the wa ter turbine indu try the d ire to get high number of
s es a

revolutions a d minimu m weight of impell r l d to the decre e of

n e e as

the impe ller di meter The imp ller of a Fr ci turbine w a t

a . e an s as

fi rst des igned wit h a pro file imil r to the lo w pre ure centri fuga l s a -
ss

pump imp ll r Show i Fig 7 5 The outer di a meter was

e e n n . .

inche but tod y it would b b uilt w ith a outer di meter equa l to

s, a e n a

the di m t r of th ucti n pip or 1 4 3 i ch which na tura lly

a e e e s o e, . n es,

incre the nu mb r f r vol utio

a se s d consider bly decre as es t h
e o e n s an a e

weight of the i mp ll r e e .

E ven if t h co ditio e m w h t di ff re t i w t r turbine

n ns a re so e a e n n a e s

from tho e in centrifug l pump ther i no doubt th t the l w

s a s e s a o

pr ure centrifug l pump will pas through a i mil r dev lopment

e ss a s s a e .

I order to ful fill the condition p rh p mo t d ir d in

n c ntrifuga l e a s s es e a e

pump tha t of ma inta i ing t le t it norm a l hea d when the pump i

,
n a as s s

23
throttled the ra tio 1 mu t a l wa ys b t ke n gr t r th n unity
,
s e a ea e a ,

w hich i s not a n a ry condition w

e c e ss ith w a ter turbin es .
1 46 C ALC ULATION OF IM PELLER AND GUIDE VANES .

( )
b . Des cri pti o n of the Cha n n el Cro ss Secti o n s , wi th the Ca lc ula ti on a n d
De si gn of the B la de s for Two Lo w pressure Ce n trifu ga l P u m p s - .

obvious tha t an y one cross section through the impeller show n

It is
in Fig 7 6 c a n no longer determine the sh a pe of the bl a de To
. .

represent correc tly the surfa ce of

the blade it mu t be shown as a s

development of the ends of the

blade upon the surfa ce of a cone .

This will be more clea rly expl a ined

in the follow ing ‘ .

Ass ume th a t the impeller dime n

s ions a t entra nce an d exit were

found an d an impeller profile

dra wn .

Ima gine n o w the impeller sub

divided into a number of pa rti a l
impellers (se e Fig 77 ) by di viding .

the exit width b equa lly into a .

number of pa rt The entra nce s .

a c i s di vided into the sa me number r

of pa rts in such a ma n ner th a t the

rota ting body forme d by e h of the ubdivided a rcs will represent ac s

equa l pa t of the tota l entra nce a rea Design ting these sma ll a rcs
r s . a

cons ecutively as etc a d letting 2 be the number of sma ll .

,
n

a rcs then for e a ch s ubdivi ion the following condition mus t be ful fi lled
,
s

d., n b., ir b.2

We wi sh to c a ll a tte n tio n to t h e fa c t th a t th e i o llo wi n g me th o d o f de te rm i n

i n g th e s urfa c e of th e b la de e n ds by me a n s o f c ross se c ti o n s t h ro ugh t h e chan n e l

an d bla d e s and by d e ve lo pin g th e blad e e n ds upo n th e s urfac e of a co n e , w as

o ri gi n a te d i n th e des ign ofFra n c is t urb i n es by A P . f a rr o f Da rms tad t , Ge rma n y ,

while c h ie f e n gi n e e r wi th Vo it h o f He id e n h e i m, i n co n ju n c ti o n with C . H utz e l

sied e r, n e w c h ie f e n gi n ee r wi t h Vo i t h T h is me t h o d was fi rs t ma d e pu
. bli c b y
E Spe id e l
. and W Wa ge n ba c h
. in Ze i ts c hrift des Verei n s de u ts c her In gen i e ure ,

No 20, Vo l
. . 1899 . These two me n h a ve onl y th e h o n o r of fi rs t b rin gi n g t h e
ma tte r to pub lic it y , wh ile a muc h high e r h o n o r is d ue t h e o rigi n a to r, A P . f a rr .

It is to be re gre t te d th a t t h e li te ra t ure upo n t h is s ub j e c t a l wa y f

s re e rs to t h e

c h an n e l c ro ss sec tio n s o f E Spe id e l

. and W Wa ge n bac h
.
, be c a use th e pri o ri ty of

t h is me th o d un do ub te dl ybl e o n gs to A P . farr .
 C H ANNE L C ROSS SECTION S .

The impeller shown in Fig 7 7 is divided into four pa rtia l impellers . .

The point b c d e on the entra nce a

s a, , d a t impeller exit were
, , re a n

determi ned a ccording to the method jus t described The points so .

found were co nected by se ca lled sectio lines which a e supposed

n -
n r

to be as ne a rly a t right a ngles to the e tra nce a as po ss ible so th a t n re ,

the direction of these lines ma y be co sidered to represent the di e c n r

tion of the w ter a t entra nce to the impe ller a d a l o th direction

a n s e

of the w a te r through the impeller Furt her the section lines a e .

, r

so co tructed tha t a y curve dra wn a t ri ght ngle s to them as

ns n a ,

th a t shown dotted in Fig 7 7 will be so divided th a t the rot a tion .

,

a re as of a ll p a rt of the curve between a y two section line will

s n s

be equ al The ction line will then divide the impeller into a
. se s

n umbe r of pa rtia l impellers through which equa l qua ntities of wa ter

flow .

The periphera l velocities an d pitch will va ry directly with the

di ameter a t the points a b c d so tha t for ll point a long the , , , , e, a s

entra nce a the velocity di a gr m a d a lso the heigh ts of ch a nnel

re a s n

wi ll be d iff erent To det rmine the heights of ch a nnel the velocity

. e ,

di a gra ms must first be determined a t the points a to e It is best to .

ss ume th a t the velocity w exi t a t entr n e to the impeller a d is ’

a . s s a c n

a t ri ght a ngle to the rot tion a re a formed by the entra nce a c b

s a r ..

With the following de cription of the ch nnel cro s sections a s a s n

ex mple w ill be c a lcul a ted Fo this exa mple the sa me conditions

a ,
r

will be assumed as for e a mple 111, th a t is the bl a de form for a x ,

de w pressure centrifuga l pump i a ga in to be de igned for

-
s s

Q 6 36 cubic feet per minute an d H. 26 feet .

As before the hydra ulic e fli i cy i s s umed as

,
so th a t '

c en s a

n H . The qu ntity of w a ter per econd p a ss ingathrough s

t h e impeller which includes the short circuited w a ter s h a ll a ga in

,
-
,

be Q cubic feet
’
.

The suction pipe di a meter sh a ll a l so be the sa me as before hence ,

D . feet ; th a t the clea r cro s ectiona l a rea of the suction

so ss

pipe is F squ a re feet a n d the velocity in the pipe w ill be

.
’

w .
’
feet per econd In order a t fi r t o t to complic te the
s . s n a

ch a nnel sections to o grea tly the out r impeller a ngle w ill be ssumed e a

as 6 .

The outer impe ller di a meter sha ll now be D .

“
fe et an d the
im peller width b foot . .
1 48 C ALC ULATION OF IM PE LLER AND GUID E VAN ES .

The periphera l velocity is eas ily dete rmi ed beca use 6 n . an d

therefore equa tion 59 c a be used n ,

1) 9 11 feet .

The number of impeller bl de i 10 hence the pitch a t the oute r a s s 2 ,

di am eter is t foot . .

By uming the bl de thick e as s

as s foot the cha nne l
a n ss

height will be foot a. .

The thickne of the bla de dec reas e the entra nce a re a to

ss s s

s quare fe t e ,

an d the ra dia l exit velocity through thi re s a a is

feet per second .

The exit velocity dia g am a o w be dra wn ( Pl a te VI) r c n n se e an d th e

following v lue dete mined gra phic lly or a n lytic lly
a s r a a a

w. feet pe econd r s an d 6. 17
°

The i n ner bound ry of the impeller pro file i o w dra wn tenta

a s n

t ive ly a n d a entr nce dra wn which will be as nea rly a t right

n a a re

a ngle to the ection line a po ible This entra nce a c is ta ken

s s s s ss . r ,

if a t a ll po ible as a of a circle b ca u e it i eas i r to determi ne

ss , an rc ,
e s s e

i t cent r of gr vity
s e d therefore it me a inner di m eter D
a an Often s n a ..

a p a r a bolic cu rve i u ed The mea n di a meter Of the ss umed

s s . a

entr nc a i d termined by y of the v a rious method u d i

a e rc s e an s se n

finding t h c nter of gr vity of a a

e e a d t a king t wice the pe pe n
a n rc n r

di uc l di t c betw n thi point a d the pump a xi s Then th e

ar s an e ee s n .

a re w hich w ould b
a we pt throu gh by rot ting t h i a ss um d entra nce
e s a s e

a rc i fou d by
s l vi g th t rm D b If the a re a thus found
n so n e e . 1r ..

i gre t r or l
s a e th n t h d ir d re a e ss the entra nce
a i eithe
e es e a a rc s r

short n d or l
e gethe d until the rot a tion
en a re equ a l
ne F In a s

s hortening or lengthe ing b the inner p a rt Of the impeller pro fil e

n .

i ch a ng d
s e .

I Pl a te VI the entra nce a

n found i hown by a he vy d hed re so s s a as

a d dott d line a d b w
n e fo und to be f t The po ition of t he
n . as oo . s

ce ter of gr vity of th entr ce a re f d t the is of the pump

n a e an re e rre o ax

was found gra phic lly a d the impell r di me ter D determined ; it

a n e a .

was found to be
D0 fOOt .
1 50 C ALC ULAT ION OF IM PELLER AND GUID E VANES .

a re engthened a bout of a i ch These point a lso a ma rked

l 1 .
7
n n . s re

upon the section lines A curve t h rough t he point of the bl ades

. s

then gives the lower limit of the bla des in the pl a n vie w .

The entra nce a re b which represents the line connecting the.,

middle points of the ch a nnel heights is now as s ume d t lie in a plan e , o

p ra llel to the impeller axis o tha t its hori zonta l projection is re p e

a ,
s r

sented as a s tra ight line w hich if poss ible is ta ken ad i a l These , , r .

ra di a l lines which represent the entra nce a e in the pla n view are
,
r ,

s how n in Pl a te V I by he a v y d as hed line s .

The involute c an n w be dra wn in the pl a n view by projecting

s o

s uffi cient points from the elev a tion The poi t ta ke on the sect ion . n s n

lines ha ve been chose n for thi purpose I Fig 78 the co truc tion s . n . ns

of the involute for the section line bb is shown in deta il .

Fi g 7 8 . .

the present e xamfle the di a meters of the b a e circles for the

Fo r s

involute for every po int upon th entra nce a

s e found from e re ar

e qua tio 160 an d their va l u

n pl a ced in the ta ble on Pl a te VI e s a re .

T h di a m te f th e b as circl for the involute to b dr wn upo n

e e rs O e es s e a

t h surfa ce of a cone a re de t rmi

e d gr a phic a lly The involutes a re e ne .

dr wn c ording to the me thod j ut given a d

a a c then tra nsfe rred s n a re

to t h pl a n view The curv which connect the point of the i

e . es s n

volute for eac h section line a e dr wn das hed in the eleva tion on r a

Pl te V I
a .

The furt her course of the ctio line in the pl a n view is n o w se n s

as umed a ccording to good jud gme t a d i t i

s be t to begi wit h n n s s n

s ctio
e line n This a w ll as the ot her s ctio n l ines m us t turn
n o .
,
s e e ,

r di l towa rds i t end b c a u e 6 9 0 a d t h co ec ti n g cur ve

a a s ,
e s .
°
n e nn

b twee the en tran ce involute d the ra di l exit li e i bes t dra wn

e n an a n s
 C H ANN EL CROSS SEC T IONS .
1 51

by an of a circle of as l a rge ra diu as poss ible I order to

a rc a s . n

se e m o e cle a rly the fo rm of the ch a nnel three

r uch section lines a e s r

d ra w in the pla n vi e w
n .

T o c o nnect the beginning of the bl a de with the ra di a l end of the

b l a d e for point a loping po ition of the bl a de exit height b i
e, s s . s

us ua ll y chos en w hich was done in the pre ent bl a de form Thi s

,
s .

s l o p i g po s ition of b
n n o w co mmon w ith wa ter turbine
., h as a s, n

a d v a t a ge over a
n tra ight position beca u e the bl a de hock a t exit
s s s

o f t h e w a ter from the impeller doe s not occur a t the a me moment s

a c ro s the entire ch a nnel width

s .

T h e other section lines were dra wn in the pl a n view in the s a me

ma r as the ection line ao a d ca re mu t be ta ken tha t they show
n n e s ,
n s

a s m oth cour e from one to the other

o s In Pl te VI t h e section line . a s

a re d r a wn in ful l line Viewing tw ucce ive bl a des in t he pl a n

s . o s ss

vi w the form of the ch a nnel c a b recogni ed to ome degree

e n e z s .

F or the further determina tion of the developed surf ce of the a

b l a d e so c a lled ax ia l a d hori onta l ections

-
n t a ken The l tter z s a re . a

a e a l o c a lled p ttern m a ker s ections bec a u t h e p a ttern m a ker

’
r s a s ,
se

us e the ho i onta l ctio

s rz hown in the pl a n view for ma king the
se ns s

b l o k upo n which the bl de i

c h ped a s S a .

I order to
n if the ection lin which were p rtly a rbitra rily
see s es a

t k en in the pl n view h ve a corr ct cou e a d l o to ob ta in

a a a e rs ,
n a s

acc ura te point of inter ction of the hori ont l cro ection w ith
s se z a ss s s

cu r ve upon the urfa ce of the bl de

s s i a l ection a e used The e a , ax s s r . s

s e c tion str ight lin

s a re in the pl a n view It i be t in the present
a es . s s

e x a mple to t a ke the e a xi a l ectio a r a di a l line in the pl a n vie w in

s s ns s s

ord r tha t the a xia l sectio ma y how in the eleva tion a cour e simil a r
e ns S s

to the entra nce a c r .

Thes e a xi a l section ca n be loca te d in the eleva tion by bringing

s

them to inte r ect the section lines in the pl a n view a n d by tra ns ferring
s

the points of inter ection to the elev tion T he points so found in

s a .

the elev tion a e connected by smooth lines

a r .

It ma y h p pe n th a t the a xi a l sections do not Show a regul a r con

a

to ur in the eleva tion which is ccou t d for by the fa ct th a t the

, a n e

s e ction lines were o t correct ly um d in the pl a n vie w T h

n a ss e . e

cou rse of the section l ine would the n h ve to b ch a nged s th a t the

s a e o

a xi a l sections i the eleva tion ma y h a ve a regul a r contour but

n ,

a l wa ys keeping in vie w th t the section lines in the pl a n vie w mus t

a

a l o h a ve a regul r contour
s a .
1 52 C ALC ULAT IO N OF IM PELLE R AN D GUIDE V AN ES .

Afte r dr wing the xi l cro s ection the templ at e for the bl a des
a a a s s s

can be ma de from the va riou ectio To produce thi bl a de ur s s ns . s s

f ce the hori onta l ction

a pa ttern ma ker ections a re used
z se s or
’
s s .

The e ections re pl a ced i the elev tion a t certa in di t nces through

s s a n a s a

the impeller a t righ t a ngles to the axis a d then their intersections n

with the bl de surf a ces a s ho wn as curve

a in the pla n view It is re s .

a dvi a ble to t a ke the e

s ctio through the point b c d e at
s se ns s a, , , ,

exit bec a use one point of the hori ont l ction in the pl a n view i z a se s

immedi a tely determined The other ction e spa ced a bout i to . se s ar

g inch p a rt depending
a upon the
,
i e of the impeller The hori S z .

zo n at l ections in the pl a n vie w a re determined by tra ns ferri g t h

s n e

inter ection of the ection lines d the r di l lines from the eleva
s s s an a a

tion into the pl a n view If the r di a l line a d the ection li es . a s n s n

h ve been fa vora bly cho en the hori ont l ection s i the pl an

a s ,
z a s n

view will how mooth n d regul r c ontour

S a s a a .

O Pl a te V I the hori o t a l ection a e shown in the pl a n an d

n z n s s r

elev tion number d con ec utively 1 2 3 etc The curves in the

a ,
e s , , ,
.

pl a n view h a ve a ve y r gul a r contour It will be noted th a t th e

r e .

points of inter ection of the hori onta l ectio s with the ection line
s z s n s

cc c a nnot be determined from the elev a tion To determine the . se

point of intersection hori onta l ections a dra wn through the

s z s re

bl a de ends upon the development of the cylinder with the sa me

s pa cing s in the elev tion whereupon the interse cting points on th e
a a ,

end of the bl a de c a be tra ferred to the pl n view n ns a .

The developed u fa ce of the bl a de i determined from th s r s s e

hori onta l ections hown in the pl a view a d by representi g thi

z s S n n n s

bl de urfa ce upon a SO c lled bl a de form To m a ke thi bl a d

a s -
a . s e

form boa rd a e t k n of a thick ss exa ctly equa l to the dista nce

,
s r a e ne

b twe the hori o ta l ction Upon these boa rd the curve of

e en z n se s . s s

the hori ont l tio z dr wn d the boa rd a e cut to the e

a se c ns a re a an s r s

curve Th e bo rds . th e n ile d together in th eir prope r

es a s a re n a

order a d the proj cting edg removed by a dr w h a ve thus givin g

n e es a s ,

upon thi wood n block a urf ce equiva lent to the developed sur
s e s a

f a ce of the bl de I order to d t rmine the

a . ct po ition of the
n e e e xa s s

boa rd o to the other it i be t to give th m the form shown on

s, ne ,
s s e

Pl t V I
a e They will a ll th h a ve a common right a ngle If then
. en .

the bo rd a pl a c d co cutiv ly with the ide of this right

a s re e n se e s s

a ngl xa t ly in li e th n t h
e e c h ori t l t io n ,a c tly in t h
e e zo n a se c n s a re e x e

s a me po ition as given on th e dr a wing

s I as sembling the s e bo a rd s . n
1 54 C ALC ULAT ION OF IM PELLE R AN D GUIDE VANES .

the exa mple given there still rema ins the dete rmina tion o f
Fo r
the dimens ions of the guide va nes .

The dia meter of the base circle for the involute i d fo o t s . .

Ass uming the number of bl a des z 9 a d a thickness of bl a de . n

3 . foo t the ch a n nel height was found to be a

,
foot . .

As lo w pre u re centrifuga l pump would h a rdly be con tructed

-
ss s s

with a impeller an gle

n the dete rmina tion of the cons ta nts
a n d the fu rther inve tiga tion of the pump would be of no i nteres t
s

an d w ill therefore be omitted .

EX AM PLE V Se e Pl a te VII . Le t us now cons ider the c a l e u .

la tion a d bla de des ign of a high spe ed pump h a ving 6 90 fo r

n -
.
“

whose re gul a tion a djusta ble guide va nes a e to be provided The r .

pump sha ll be de ig d for the sa me conditions as in exa mples III

s ne

a d IV
n tha t is Q 6 36 cubic fe et pe minute H
, ,
26 feet r , . ,

11 H . fe t The dimen ion a t entra nce sh l l be the sa me as

e . s s a

in exam ple IV D feet F , squa re fo ot w

. , .
’
, ,

feet per second The entra nce a c sh a ll be the sa me as before an d

. r ,

the dia meter a t its center of gra vity is

D. foot .

The impeller a ngle a t the outer dia meter sh a ll now be 6

The dia meter of the ba se circle for the exit involutes i d s .

foot .

B y ass uming the number of bl a de z 10 a d the thickne s f s n s o

the bla de foot the exit height

s s foot ,
a. .

Gra phic a lly n d a na lytic a lly the outer impeller di a mete r upo n
a

which the c a lcul a tion a e b as ed wa fou d to be s r s n

D . fee t ,

an d the impeller a ngle a t this di a meter

°
6 . 1 51

The impeller width sh a ll a ga in be ta ken as in exa mple IV a s

b . foot a d let 6 ,
n .

The r di l exit vel o city a t di meter D

a a feet c an be found a .

from equ tions 157 a d 158 This ta kes into a ccount the decreas ed
a n .

exit area due to the thickness of th e bl a des This ra dia l exit velocity is .

v. fe et per second .
 C H AN NE L CROSS SE CTIONS .
1 55

The clea r exi t area a t di a meter D is F s qu a re foo t . . .

The peripheral vel o city c an o w b foun d from equa tion 58 n e ,

be
ca use v an d 6 a e known
, . r .

_
t my
2
"

1
”

V 7) 11
9 . fee t per second .

B
a ta n

The number of revolutio per minute is ns n 590 .

Fro m equa tion u V ; gH we fin d . x 1 .

F rom the curves 4: (

se e Pa l te 1) the a ngle 6 is found corresponding .

to x an d 6 . 15 1
°
The va lue found is
6.

In order to dra w the exit velocity di a gra m '

we determine the
V 1; gH
K
fe et . Gra phica lly or a na l yt ica lly
w. feet per econd s an d v. fee t per econd s .

The va lues foun d for impeller entra ce for point a b c d e a e n s , , , ,

r

a g in a rr nged in a t a ble on Pl a te V II
a a d a fte r dra w ing the en t ra ce , an n

velocity di a gr ms for ea ch of these points the cha nnel height are

a s

determined g a phic a lly r .

Af ter dr w ing the section line the involute a e dra wn upon the
a s s r

s urfa ce of the cone a d then the v a riou , points a tra nsferred

n s re

upon the as sumed ection lines From thi the involutes ca n be

s . s

dra wn in the pl a n view The entra nce a c could not b ta ken as . r e

r di l on the pl a n view b ut h a d to be ta ken ta ngent to a circle wh os

a a , e

dia meter was d foot If the entra nce a rc h a d been ta ken .

as a ra di a l line in the pl a n view it would not h a ve bee n poss ible to

connect the bl a de end a t point c with the exit involute It mu t . s

be remembered tha t in the eleva tion the perpe ndicul r dis t nce of a a

the po int a t entra nce to the hori onta l ection p ing through the
e z s as s

po int t e it de t rmine w hether

e a x proper connection of the e
e s a s

points is poss ible B oth point a be only a cert a in di t nce from

. s c n s a

one a nother i the pl a n view in order to obt in a pos ible o

n a s c n

n e t io
c be tween the beginning d end of the bl a de There
n an . a re

va rious wa y in w hich to fi d ju t w hen uch a connection i or i

s n s s s s

not po ss ible Perha ps the be t wa y would be to t a ke

. ection s a s
1 56 C ALCULAT ION OF IM PELLER AN D GUIDE V ANES .

through the pl a n view as nea rly pa ra llel as possible to the section

line a d to project into a spec i a l elev tion the a c tua l vi ew of
63 n a

the begi n nings a d ends of the bl ade as they would a ppear upon
n s a

rot ting body formed by the section line a in the eleva tion
a e .

After the involu tes for impeller entra nce a d e xit h a ve bee n n

d ra w n in the pl a n view the section lines a e as s umed in thi s view r .

I order th a t the ra di a l s ctions in the elev a tion m a y show a

n e

course simil a r to the entra nce a c they a r dra wn in th e pl a view r e n

not ra di al but ta ngent to the circle to which the b line was ass ume d .

to be ta ngent The axia l sections a then a ga in tra n ferred to the

. re s

eleva tion a n d if they do n t h a ve a regul a r contour the section lines

,
o

in the pla n view must be ch a nged a ccordingly C on idering the . s

previous di cu ion the hori ont l section a re pl c d in the eleva

s ss ,
z a s a e

tio a d the curves of in tersection with the s urfa ce of the bla de a r

n n e

dete rmined in the pl a n view Fo r pro duci n g the bla de form the .

co c ve bl a de surfa ce is a g in u ed C re must be ta ke n th a t the

n a a s . a

sectio n s a re cho en through the correct bl de s urfa ce In orde not a s . r

to complica te the pl a n view on Pla te VII on ly the sha pe of o of

'

ne

the boa rd of the wooden bl de form is there given

s a .

Gui de Va The guide va nes a to be m a de a dj usta ble A

n es . re . s

before the dia meter of the b e circle of the involute is found for
, as

the no ma l qua nti ty of wa ter Fo r thi qu ntity d

r foot . s a , .

Ass uming 12 guide va ne a d the thickness of va ne 3 s n 1

foot the entra nce height i determined

, foot s as a , .

The entra nce involutes dra wn as before Pla te VII show the a re . s

further con struction of the guide v ne Speci a l c a re should be a s .

ta ken tha t the ch nnel height a increas es very gra du a lly w hich
a , ,

s hould a l so be the c ase for a y other po ition of the v a nes I order n s . n

to det rmine the limiting position of the va nes three va nes a r cut
e e

out of pa per which a the be turned until a ll three va n es become

,
c n n

ta ng nt to a common circle
e .

I d termining the di a meter of the bolt or pin a bout w hich the

n e

guide v e ro t t th li mi ti g po ition of the v a ne must be ta ken

an a es e n s

i t ccount When the guid va ne

n o a . clos d the ent ire pre ure e s a re e ss

he d plu suction hea d is e erted upon t h urfa ce of the va nes If

a s x e s .

t i the pitch in feet of the va ne in their limiting or end po ition

s d s s , an

b the width of the v a ne in f t th e the lo a d in pounds on e a ch

. s ee ,
n

va ne due to the pres ure h d of th w t r will be s ea e a e

P b i H pou ds . . n .
158 C ALC ULAT IO N OF IM PELLER AND GUIDE ‘
VANIB .

consta nt C h as even a l a rger va lue tha n in exa mple III a d therefo re

1 ,
n

the outer di a meter in the last exa mple could b somewha t reduce d e .

It a therefore be a id th a t the pumps c a lcul a ted in e xa mples III

c n s

a d V a
n a pproxi ma tely Of equ a l perform a nce
re Bu t the fa ct m u st .

be emphas i ed tha t the l ast pump is much lighter a d chea per to

z n

produce a ccount of its

on ma ller outer dia meter The weights of s .

the completed pump of exa mples III a d V a e to ea ch other as s n r

a bout 5 to 3 Another point wo rthy of cons idera tion is th t the l a s t

. a

pump opera te a t 59 0 revolutions per minute while the other pum p

s ,

ope ra tes a t o ly 380 revolutions per minute

n .

The a dju ta ble guide va nes of exa mple V permit a decrea se of the
s

quantity of wa te r over very wide limit a n d still gi ve a h igh effi cien c y s .

33 . PR E SS U R E D IST R IB UT IO N IN A C E NTR IF U GAL PUM P IN W H IC H

ENT RAN CE To TH E IM PE LL E R IS N OT PAR ALLE L To TH E SH A F T .

Inorder to decreas e the mea n i n ner impeller dia meter to a minimum

it is necess a ry to pl a ce the line of entra nce within the elbow of the
impe ller By doing this there will be different periphera l velociti e s
.

a t e ch point of the entra nce line a d therefore di fferent entra nce

a n

velocity dia gra ms The consta nt C a d C of equa tion 225 were

. s 1 n a

determined in exa mple V for so c lled mea n di a meter which wa s a -

a

a ss umed to be a t the cente r of gr vity of the entra nce a c These a r .

consta nts a e correct o ly for a threa d of wa ter pas sing through th is

r n

cente r of gra vity while for a y other point of the entra nce a e there
,
n r

would be di ff erent v a lue for the const a nt C a d C Ass uming th a t

s s 1 n g .

the impeller a d guide va ne a e divided into a numbe r of equa l pa

n s r r

tia l impellers d guide v nes there would result a num ber of pa rti a l
an a ,

pump d e ch of th e would influence the he ad di ff erently if the i r

s, an a es

disch rge w ere throttled as will be expl a ined in the following

a , .

Ass ume pump divided into a number of equa l pa rti a l pumps

a .

Equ a tion 10 mu t i g era l be true for e a ch of these p ti a l pumps

s n en ar ,

only the ep a ra te velocitie would ha ve different va lues The equa

s s .

tion rea ds
u +
.
2
v.
2
v.
2
+w .
2
w} 2 n gH .

cro ctions of the pump be determined L t f f f be th e

Le t a ll ss s e . e ., ., g

cro s tio f pa rt i l pump in which the velocities v v w occur

ss ec ns O a a ., . , .

( the bl de thick e i
a s as umed in finitely ma ll
n s o thss
a t w s s .
 PRESSURE DISTRIBUTION IN A C ENTR IFUGAL PUM P . 1 59

The condition eq ua tion c a now be writte n if q re prese nts the quan

n ,

tity of water passing through a pa rti l pump a ,

hence
I
.

Further a ccording to equa tion 13

,

w. 2
u 2
. v. 2 c os B. ( 13)
Su
bstitutingthese va lues of v v an d w in equa tion 10 ., ., . , an d lett ing
u.
2 the a bsolute exi t velocity is found to be
3

D. f . co s B .

In this equa tion 1 H is still unknown Its va lue depends ac cording

1 . .
,

to equa tion 22 5 upon the ra tio g which h as a diff erent va lue for
:

_
,

ea ch pa rti a l pump .

Le t F be the exit are a of a pa rti a l pump so tha t

.

Substituting the a bove in equa tio n 225 we ge t

“a Do F a.

ta n fl. D F.
.

In this equa tion is the entra nce a rea of a pa rti a l pump . Substi
tutin g this va lue of q aH in eq ua tion 2 35 we get
.
16 0 C ALC ULAT IO N O F IM PELLE R AN D G UIDE V ANES .

The v lue upon the right ha nd ide of this equa tion a equa ll y
a s -
s re

l a rge for ea ch pa rtia l pump therefore the veloc ity w an d the qua ntity
, .

of wa ter delivered a r the sa me for ea c h pa rti a l pump O the o th r

e . n e

ha nd the pre s ure hea d will va ry a ccording to equa tio 2 3 7

,
s s n .

A the velocity for e a ch p a rti a l pump i the sa me th a t is it i equ a lly

S s , ,
s

l a rge over the entire exit width there wi ll be unequa l pressure hea d
, s

only while the wa ter i pas s ing through the impeller B ut a ccord
s .

ing to the funda menta l l a w of hydra ulics it is h a rd ly possible for

s

va rying pressures to occur a t exit from the impeller when the v locity e

v is const a nt
. .Therefore a pres ure equ li tion mu t t ke pla ce
s a za s a

in the impeller which will not permit of a furt her investiga tion .

It might be well to t a te th a t the cond itions existing in wa t er

s

turbines a re jus t the reve se of those e isting in centrifuga l pum p

r x s .

With t h e former a certa in press ure he a d is given Fo determin i ng . r

the a bsolute entra nce velocity (corresponding to the a bsolute e xit

velocity of a pump ) equa tion 2 35 c be used from which it c a n be an ,

d irectly seen th a t the velocity w a n d therefore the qua ntity of wa ter

.

va rie for ea ch ingle pa rti a l pump Hence the ingle pa rtia l

s s . s

i mpellers of wa ter turbine under con t a nt pre s ure he a d work w ith

s s s

v ryi ng qua ntities f wa ter while centrifuga l pumps deliver eq ua l

a O ,

qua ntities Of wa ter under va rying pre sure hea ds s .

Fo the nor ma l qu a ntity of wa ter a d he a d for w hich 5 was as u m ed

r n . s

to be 90 this phenomenon does not occur bec a u e for this s pe ci l

°
,
s a

case the va lues of the velocities a t entra nce d o not influence the
pre ssure hea d as c an be s e fro m equ tion 2 25 when 5
, e n It a .

then ta kes the form

a
2
77 “0 +
16 2 DESIGN OF IM POR T ANT PUM P PARTS .

in which R i a coefficient of friction The v lue of this coeffi cient

s . a

was found by Las che a d St i b c k to depend U po the press ue p n r e n r

a d the temper ture t of the be a ring provided p rem a i ns w ithin the

n a ,

limit of 15 a d 2 00 pound per qua re inch a d the tempera ture t be

s n s s n

tween the limit f 80 F d 2 12 F If such is the cas e the fol

s o
°
. an
°
.

lowing imple formul a hold good

s s

y p ( t 32 ) cons ta nt
The hea t ra di a ted pe r hour from a be a ring c an be expresse d in
B t
. . u as.

k l‘
dl (t —
to )

in which t is the temper ture of the su roundi ng ai r an d k is a

. a r

coeffi cient depending omewh a t upon the design of the bea ring s .

If the be ri g c as ting i thin a d prese nts a l a rge ra di a ting su

a n s rfa ce n ,

k c a be t a ken s high as 150 but if the ra di a ting s urfa ce is c on fin e d

n a ,

a n d the a i circul a tion poor k mus t be t a ken as lo w as 50

r .

When no a rti fici a l cooling of the be a ring i provi ded Q mus t s ,

equa l Q d the fina l t mpera ture of the be a ring t ca n be c a l e u

, , an e

la ted from equ tions 240 a d 242 If the fi a l tempera ture is

a n . n

excessive eith r the Oil or the be a ri n g mu t be wa te r cooled L t

,
e s -
. e

Q be the hea t in B t u per hour impa rted to the cooling wa ter

. . . .

either by the Oil or by the be a ring Then .

Q Q!
“
Q 0

The wei ght of oil necessa ry per hour is

Q .

S peci fi c he a t Of the Oil (,

t t, )

in which
t; the tempera ture of the Oil entering the bea ring in degrees F ,

t. the tempe ra ture of the Oil le a ving the bea ring in degree s F .

Inc a lc ul a ting the weight Of wa ter nece a ry for cooling the ss

bea ring or the oil a very liber l a llowa nce mu t be m a de on a ccount

, a s

of the i effi ciency of thi method In mo t c lcul a tions the tem

n s . s a

pe a t
r ure a fi r s t
s as umed
rea d the te m pe r a ture
s t Of the Oil le a ving
,
n g

th e bea ring m u s t of course be t a ken much lo we r th a n the t em pe ra

 BE AR IN GS .
163

ture t of the bea ring SO as to be sure tha t the oil will a t least rea ch
,

the tempera ture t before it lea ves the be a ring

z .

When the be a ring is to be cooled by wa t r circul a tion the wa ter e ,

can either be introduced i to a cor d p ce in the lini g d pede ta l

n e S a n an s

or a n be forced through a coil of pipe embedded in the lining In

c s .

most cases ba bbitted li n ings a u ed a d the cooling wa ter pipe re s n s

ca n be a rra nged w ithin the b a bbitt Fo r high s peed work it is .

-

f e m a re d Wl f fi b e 55 M " r e m a 7 3 4

Fig 7 9 . .

best to use ba ll s ea ted linings which a utoma tica lly adjus t them
-

selves wi th the d e fl c ti of the ha ft or as s ist the sha ft in securing

e on S

a co rrect line u p -It is very

. import a nt e peci a lly in high peed ,
s -
s

work to des ign the be a ring SO th a t the ba bbitt will not become
,

lo ose after be ing in servi ce for some time The lining Shell i usua lly . s

provi ded with doveta iled recesse for holding the ba bbitt as hown s , s

in Fig 7 9 These recesse s hould be ma d d p enough a g od

. . s e ee , o

rule b ing to h ve th depth Of th r c

e a e o m wh a t gr a t r th n e e e ss a s e e e a

the mi imum thick e of the b bbit t b Fig 7 9 ho w d v lop

n n ss a . . s s a e e
1 64 DESIGN OF IM PORT ANT PUM P PARTS .

ment of the uppe r a n d lower ha lves of the linin g without the ba b

bitt a d gives the correct wa y of arr anging the doveta iled recesse s
,
n ,

the funda menta l idea being th a t a ll the rece sses of the lini n g a re
intima tely con nec ted by a deep wa y co nnection .
”

35 . PACKIN GS .

The effi cie cy of centrifuga l pumps depends pa rt ly upon the prope r

n

desi gn of the stuffing boxes a d pa ckings The pa cking on the s uction

n .

end of the pump must usu lly prevent a ir from lea king into t h e
a

F ig 8 0. .

pump a n d the pa cking on the high press ure end mu t prevent wa te r

,
-
s

from le king out Of th pump The pa cking in both c as e must be

a e . s

t ight enou gh to prev nt uch le ka ges a n d s till not cre a te e xcess i ve

e s a

friction lo e by binding the h a ft

ss s s .

To prevent a ir from ent ring the pump on the uction s ide a wa ter
e s

sea l is general ly u d of d ig somewh at i mil r to th a t Shown

se a es n s a

in Fig 80 The p a c king c i t of two di tinct l ngth w ith a spa cer

. . ons s s s e s

or l ntern betwee The bj t of t h l tt r i to kee p the p a cki ng

a n . O ec e a e s

a wa y from the Opening through whi h w a ter und e r pre ure i intro c ss s

due d
c A y le a k a ge of the i n r lengt h of p cking pe rmi t thi s
. n n e a s

wa ter to flow into the pump The upply Of w ter being a mple n o
. s a ,
166 DESIGN OF IM PORT ANT PUM P PAR T S .

An y wa ter which esca pes through thes e close clea r nce s is t a ken by a

the dra in into the waste pipe .

I multi st a ge centrifuga l pumps a p a cking must be used betwe en

n -

st a ges to prevent wa t er le a ka ge I m os t c ases a long close clea ran ce

. n

Fig 82 . .

fit between the Sha ft or slee ve a d the cas ing is employed Th isn .

des ign c a be seen in the cro ectiona l view Shown in Fig 82 So m e

n ss - s . .

designer use a l a byrinth pa cking imi la r to Fig 83 which h as given

s s .
,

Fig 83 . .

excellent results wi th stea m turbines but is useless in centrifuga l

pump The effec tiven of thi l byrinth p a cking is bas e d on the
s . e ss s a

exp a n ion of stea m in the enl a rged spa ces Incompressible fluids
s .
,

on the contra ry give only de d wa ter a t these points a d the fluid

,
a ,
n
 IM PE LLERS .
16 7

passes through the clo se cleara nce wi thout a n y disturba nce a t the
e la rged pl ce
n a s

Fig 84 illustra tes a nother method Of preventing le a k a ge between

.

st a ges It con i ts of a close fitt i g ring C a b ut the h ft or impeller

. s s - n o s a

hub a d a tta ched loo ely to the di phra gm B or c as ing between

n s a

st ge by me a ns of bolt s w hich permit of ome ra di a l pl a y

a s The ring s .

i c a pa ble of a djus ting itse lf to

s y po s ition of the h a ft a n d i s held an s ,

Fig 84 . .

aga in t the di a phra gm or cas ing by the difference of w ter press ure
s a

in the two sta ge Sometime provi ion i ma de for oil or grea se

s . s s s

lubrica tion Of this ri g as hown in the figuren ,

s .

36 . IM PE LLE RS .

Almost impellers for centrifuga l pumps whether Of the single

a ll ,

or double inlet type a re cas t in one piece ; tha t is the hub a n d we b

, ,

ar cas t integra l with the bl a des a d outer Side or shroud In the

e n .

ca e of very hi gh speed cas tings c a nnot be used on a ccount of the

s s,

high stresse produced in the i mpeller The designer mu t then ue

s . s s

a forged s te el hub a d w b with the bl a de either m a chined out of the

n e s

solid forging or the bl de m a de of sh t teel pl te

a d riveted to
s ee -
s a s an

the w b Of the imp ll r So metime

e e hroud ring i u d d ome
e . s a S s se an s

times it i omitted ; if ued the bl d

s rivet d or c t to thi ri g
s , a e s a re e as s n .

In nea rly a ll des ign of impellers it i import nt to inve tig te

s s a s a
168 DESIG N OF IM PORT ANT PUM P PART S .

the stresse s produce d by centrifug l force At a y poi t wit h i a . n n n an

impeller w b both r d i a l a d ta nge t ia l stress e exi t By di l

e a n n s s . ra a

s tre e a mea nt tresse imil a r to those existing in a tr igh t th in

ss s re s s s s a ,

rod under t ns ion Ta ngenti a l stresse s ar stresse s sim il r t t h o

e . e a o se

occurring a round the circumferen c e of a thin tube ubjec t d t o i s e n

te rn a l or bur ting pre ure s ss .

The stre produced by centrifuga l force in a di k the thi kn e ss

s se s s c

of which doe not va ry to o ra pidly a e determine d w hen a y t w

s ,
r n o

s tresse both a t the a me ra di u

s s or a t di ff erent r dii a re kno wn
s The a .

t w stresse us u a lly kno wn

o s the r di l tre a t the bore o f the a re a a s ss

hu b a d the ra di a l tre
n a t the periphery of the i m pe ller
s ss The .

r dia l stre a t the bore Of the hub i either nothing if the im p e ller
a ss s

ha s no forced fit on the h ft or is equ l to the tre produced by the S a a s ss

forced fit If the impeller i loose upon the h a ft the ra di a l s tress

. s s

a t the bore v a rie wi th the peed Of the impeller bec us e as t h

s S p ee d
s , a e

increas the tretch a t t he bore a l o incre

es s The fit i u u l ly s ase s . s s a

m de hea vy enough o th a t t h im pe ller will not be loo e upon t h

a s e s e

s ha ft wh running t it ma i mum speed I ca lcul a ting the

en a s x . n

stre in a impeller the ra di l tr a t the bore when the im pe ller

ss e s n a s e ss

i running a t it m a ximum peed hould be used

s s The ra dia l s tre s s s . s

a t the p riph ry Of e im peller is u ua lly nothing bec a u e no bl d

e an s , s a es

or ma teri l e tend beyond the periphery of the di sk

a x .

It i di ffi cult to t a ke i to a ccou t the s tre

s produced in a n n ss e s n

impeller w b by the bl a d a tta ch d th reto beca u e the ce n trifug l

e es e e s a

force of the bl a de i distributed a long the ide f the impe ller w e b

s s s s O

a n d c nnot pr perly be ch a rged to a y p r ticul a r r a dius

a o It i n a . s

po ible to fi d th e stre s exa ctly but for pr ctica l work it i

ss n es s se , a s

s uffi cient t pro e d o foll w The ma ximum tre in the w b

c e as o s . s ss e

or di k Of th impe ll r i fou d as if no bl a de e isted ; then the a ver

s e e s n s x

a g t
e f t h
s re ss e s di k w itho d w ithoute bl a de a rs found T h e an s e .

m ximum tr previou ly found for the di k w ithout bl a des i

a s e ss s s s

then corre t d o t h c e umptio th t the m aximum tr

n e a ss in n a s e sse s a

di k wi th a d w ithout bl a de v ry d ire tly with the vera ge stress

s n s a c a es

O f a di k w ith a d without bl a d
s Thi a umption a lthough n o t
n es . s ss ,

correct errs in t h d ir tion of a f ty b c a u t h a ctu a l m xi m um

,
e ec s e , e se e a

stre ss i s a lw y lightly low r A imil r proc i follow d with

a s s e . s a e ss s e

rega rd to the tr e in t h hroud ri g Whe n the bl a de re

s e ss s e s n s . s a

riv t d to the imp ller di k the c l ul ti n f tr

e e e rema ins the
s a c a o O s e ss e s

a m a d the stre ss in the rivets a re eas ily compute d

e, n es .
170 DESIGN OF IM PORTANT PUM P PART S .

This method i very l a boriou e specia lly for di ks Of l a rge ra di a l

s s, s

dimens ions a n d va rying thicknesses a n d Should be resor ted to o l y , n

in extreme cases Eq ua tions 2 45 a d 246 a re more e as ily ha n dl d

. n e

if the profile Of the di k se ction c a be represented by one or mo e

s n r

hype rbol as The hyperbol a c a n be expressed by the equa tion

.

t ox
“

in which
t isthe thickness of the disk a t ra dius 33 ,

c is a co ns ta nt ,

a is con t nt either pos itive or nega tive ; nega tive when the dis k
a s a

decreases in thickness outwa rd .

The two con sta nts c a n d a a re dependent upon the form Of the
hyperbola If this form i given the va lue of t a t a y two ra d ii x
. s ,
s n

a d x a re kn own from which c an d a a be olved by tri a l a d error

’
n ,
c n s n .

When the disk i of un iform thickness a 0 s .

Equ a tions 24 5 an d 2 46 be come integra ble when t ox A the “

. s

di sk decreas es in thickness outw rd a is nega tive or t car I a s ,

“
. n

te g a t i g equ a tions 24 5 a d 2 46 w ith t expressed as a bove the follo w

r n n

ing va lues a e Obta ined r

) ”
) bl l
(l )
v bfi h
fl
]
‘
24 8
'
l
v ax v l /2

0 2
(1 ¢ 1 V) b1xf l
(1 MP) bzxf
]
—
3 1 0 35 “ z"
249

2
N 3 (1 w)
E [8 ( 3 ) ]
v a

a v + l

a v + 1

b; an dequal consta nts

bz .

The consta nts b a d b a be found when two gi ven stresse s i n

l n 2 c n

the section Of the di k considered a e kn own These given stresse s

s r .

F th d i
or tie f th f mul
e r va uth ton l ti of St d l es e or as se e a or s
’
ran s a on o o o a s
’

Stea m T u
”
rbi n e s .
 IM PE LLE RS .
171

may be both ra di a l or both ta ngenti a l or one of ea ch I disks with . n

p a ra llel side a d with na rrow hubs the entire profile of the di k

s n s

m a y be represented by s ingle hyperbol a a I thi s c ase as h as been . n ,

expl a ined the two given tre es a the r di a l tresses a t the bo re

,
s ss re a s

a d a t the periphery
n In mo t c a e however it will require from
. s s s, ,

t wo to four hyperbol to represent the impeller pro file a ccura tely

as .

A formul bas ed on equ tion 2 48

a d 24 9 h
a been derived for the
s an as ,

c as e of four hyperbolas ,to olve for the ta ngenti l tr a t the bore

s a s e ss

( this being u u lly stheam a ximum tre in the di k ) w hen the r a di l s ss s a

st ress e a t the bore

s d periphery a
an given Thi formul a h o w re . s ,

ever is very cumber ome d requires much gre ter a ccura cy f

,
.
s ,
an a O

Fig 85
. .

computa tion th n th a t aff orded by a slide rule The process gen

a .

e ally follo wed i s therefore one

r imil a r to th t previously de cribed S a s

e cept th t no ma ll di t nce of nee d be ta ken but o ly v lues

x a s s a s 1: n a

a t the be ginning d d Of e a ch hyperbolic curve

an en I order to . n

i llu tra te thi method of lcul a ting tr

s s i centrifuga l pM p
ca s e ss e s n

impellers a typic l exa mple wil l be t k

a a en .

Fig 85 repre ents

. pump imp ll r d i gn ed for very high pe d
s a e e es s e s .

The ma ximum number Of revolution per minute i to be 42 00 s s .

We sh ll fi r t determine the di tribution Of ra di l

a s d t n ge t i ls a an a n a

stresses throughout th e d i s k I order to do thi s the profile of

. n
172 DESIGN OF IM POR T ANT PUM P PARTS .

the section of the impeller d i k is divided into a number f hyper

s O

boli e cu ves By tria l a d error we fi d tha t fo ur hype rbol as w ill

r . n n

very close ly a pproxim a te the impeller d is k section Fig 86 Show s . .

how closely the assume d a n d true ections a gree The l a t ter a re i n

s .

full line a d the former in dotted line The firs t hyperbolic c ur ve

s n s .

i from the bore (a:

s 3 inche ) to x
s 4 inches This is a stra igh t
.

line d repre ent a di k Of c ta t thickn

an s s s ons n The second hy per
e ss .

bo li c curve exten ds from x 4 inches to a: inches the third ,

1 74 DESIGN OF IM PORT ANT PUM P PARTS .

the stre
As a t the e d Of the fi
sse s t hyperbolic section (a: 4 in ) n rs .

must be the a me as the tresse s a t the beginning f the seco d

s s O n

hyperbolic ection (a: 4 in ) the e two equa tions c a be equa t d

s . s n e

equ l t ea ch other a d the va lues of the cons ta nts b a d b Of the

a o n l n g

second hyperbolic ection be determine d s .

U ing the va lues Of equa tion 24 7 for the second hype rb lic se ction
s o

i equ a tion 24 8 the l a tter a n be w ritten

n , c

—
52 9 4 86 8 0 X 1 0 bl 5
106 5 X 10 bz 3

53 15 X 10 b; 5
2 7 82 X 1 0 b2 3

B oth of these va lue of r di a l a d ta ngentia l stresses must equ a l

s a n

corresponding v lue expr ed in equa tions 248b a d 249 b beca u e

a s e ss n , s

they refer to tre se a t the me point Equa ting equa tion 2 48 b

s s s sa .

eq ua l to equ tion 248 a d q ua tion 249 b equa l to equa tion 249

a c n e c

determine the v lue of the w cons ta nt b a n d b for the second

s a ne s l 2

hype rbolic tion H ence equa ting the e equa tions a n d olving for
se c .
,
s s

b, an d bz we ge t
— 9
bl 4 82 9 X 0 3, 5540 X 10

b2 1 885 x 10
—
10
a . 2 98 7 x 1 0
— 8

Fo r a: inches , o r” an d a be found by substituting the e

,” can s

n ew v lu of b a d
a es l n b2 in equ tions 24 8 a d 249 M a ki ng these
a n .

s ub titution we g t
s s e

8 194

a c, 4 13 0

Proceeding with the third hyperbolic se ction as w ith the second ,

-
12 8 9 0 1 2 32 X 10 bl °
3 155 X 10 02 2

—
7 42 1 8086 X 1 0 bl 5
7 1 18 X 10 bz 2

As th v lue of the r di l d t ng nti l stre ses a t the beginning

e se a s a a an a e a s

of the third hyp rbolic tion mu t be qu a l to the va lue a t the

e se c s e s

end of the econd h y p rbolic ectio (for x

s inches) equa tione s n ,

24 8d qu a l equ a t ion 2 48 a d equ tion 2 4 9 d equ a l equ a tion 249 e

e s c n a s .

Equ a ting th e equ tion the new v a lu

es of the consta nt b a d b
a s es s l n ,

for the third hyperbolic ection be found Solving for these s can .

va lues we get
bl 1 1 78 X 1 0 4 2
38 7 0 X 1 0 9
-

0 .

b. 1342 10
— wa
. 2 2 55 x 10
—7
 IM PELLERS . 1 75

Fo r a: 13 inches be found by substituting these

, o r“ a n d a ,“ ca n

new va lues Of b a n d b in equa tions 24 8 an d 249 Solvin g after

l 2 .

ma king these substitutions we h a ve

0 1,

a ,“ 0 1
,

Proceeding as before the va lue of s a n d a ,“ for the fourth hyper

boli e section ma y be expressed a s
7 7 13 6 42 6 X 1 0 1°
b1 1 13 9 X b2

4440 2 865 X 1 0 1°
b1 6 77 4 X 10 be 2

As before equa tion 248f equa l equa tion 248g a n d equa tion 24 9 f
,
s

equa ls equ tion 249g Solving th ese equa tions for the e w va lues
a . n

of the consta nts b a d b for the fourth hyperbolic section we h a ve

; n z

—
bl 3848 X 17 ’

Og
,

52 7800 10 w —
a .

Fo r a: inche s, a , “ a n d on ,” a re found by substituting these

new va lues Of b l an d bz in equa tio ns 2 48 a d 2 49 Solving we get
n .

Si n ce the ra di a l tre s t the periphery of the impeller a n d

a ,“ is s s a

this stre i equ l to nothing the va lue for

ss s a ca be obta ined by
, a t, n

eq ua ting equa tion 2 48 b equa l to nothing Hence .

0 a t, 8 1 87 8 ,

10 2 6 2 pound per qu a re inch

s s

Ha ving Obta ined the va l ue of the consta nts b a d b for a ll l n ,

hyperbolic ectio s c a n be found from equa tions 2 53a b c an d d

s n , , , ,

a n d 2 54 a b c , d d, ,
an .

Fo r the fi rs t hyperbolic section

bz 2 02 5 X
176 DESIGN OF IM PORTANT PUM P PARTS .

Fo r the secon d hype rbolic section

bl 1050 x 10
—3

— 6
b: 19 04 X

Fo r the third hyperbolic section

bz 1602 X 10

Fo r the fourth hype rbolic s ction e

In 57 98 X

bz 392 7 X 10

Ha ving the va lues of the e const nt the tresses a t the d iff ere n t
s a s, s

rad ii of the impeller a n w be Obta ined by ubstituting in equ ti o n s

c n o s a

2 48 a n d 24 9 the v a lue of b
s d b ju t found
l an Ta ble I was c o m
2 s .

p ute d in thi s m a nner a dngive s the r a di a l a d t a ngenti a l s tre ss e s a t n

va rious dista nces from the center of the impeller These va lues a re .

plotted in the form Of a curve in Fig 87 . .

T ABLE I .

Ra d ia l St re ss T a n ge n t i a l
St re ss 0 3.

lbs pe r
. sq . in . lbs pe r s q
. . in .

10 262

The ra di a l stretch or elonga tion of the impeller at a n y dista n c e a:

from the center c an be expre sed by the formul as

°
bl ba “
E
"
0 23 it
178 DES IGN OF IM PORTANT PUM P PARTS .

in which
S is the spe ci fic den sity of the ma te ri a l in pounds per cubic in c h , ,

I is the moment of i ne rti a of the tot a l cross section of the i m

peller a bout the a xis Of the sha ft ,

A is the a rea of the tota l cross section of the impe ller ,

N is the number of revolutions per minute of the impeller .

hyperbolic cross se ctions of the form

Fo r t “
6x the moment of
inertia of cross section is
d fa
)
' '

xl

an d the a re of cross section is

a

+1
)
0
x1

where x a n d z a e the termina l ra d ii of the sectio cons idered

. ; r n .

U ing e qu a tions 2 58 a n d 2 59 for the four hyperbolic se ctions

s

the problem a n d substituting the umm a tion Of their v lues s a

equa tion 2 57 the a vera ge stress c a n be found The summa tion

,
.

the four va lues of I in thi s problem is

74 1
I
l— fi
2 1 (6 4 27 ) 4)
1 4
°
X 10 1 1
1 106 6 09

z I 22 7 1 in .
“
pounds .

The summa tion of the four va lue s Of A in this problem is

)
17 6 1 1 1
(4 3) T 4
x 10 6

2 A s qua re inches .

The a vera ge ta ngenti a l stress at 42 00 rpm

. . . is

22
Aver . a, N S ’
44 3; 7 2 10 lbs pe r sq i n . . . 26 0
 IM PE LLERS .

This a vera ge ta ngenti a l tress a gree well with the va lues founds s

previous ly an d plotted in Fig 8 7 I the l a tt r the ta genti a l . . n e n

stress es va ry from 3 6 2 6 pound per qu a re inch a t the periphery of s s

t h di k to 10 2 6 2 pound per qu a r inch a t the bore

e s If a me a n
s s e .

line of stress were dra wn it would very nea rly equa l 7 2 10 pounds
per squa re inch .

If the tot a l weight of the bl a des i W pounds a n d the ra d i us of s

their center of gr vity is R inches the ver ge ta ngenti a l tress in

a ,
a a s

the impeller di k n d hub whose tota l cross sectiona l rea is A due

s a -
a ,

to the bl a des is ,

Avera ge or , N f
W

The weight of the sixteen bl a des (eight on ea ch side of the disk ) is

45 po u nds The ra dius Of their cen ter Of gra vity is
. i nches .

Their centrifuga l force a t 42 00 r p m i 28 6 500 pound H ence . . . s s .

the a dditiona l a vera ge tress in the impeller di k due to the weight

s s

of the bla des is a ccording to equa tion 26 1

, ,

Avera ge 6 , 1 030 pounds per squa re inch . 262

The tot a l a vera ge stre in the disk is 7 2 10 1030 8240 pounds

ss

per qua re inch As uming previou ly e p l a ined th a t the ma xi

s . s as s x

mum t a ngenti a l tress i the impeller is incre ed in the sa me ra tio

s n as

as the a vera ge s tress es with a d without bl d es the ma ximum ta n

,
n a ,

ge n t i a l s tre ss w ill be
M a ximum 6 , 10 2 6 2 1 1 7 30 lbs . per sq in. .

The stresses in the Shroud ring a e ca lcul a ted simil a rly to those
,
s r

in the impeller disk The profile of the shroud ring ca be very

. n

closely defi ned by three hyperbol as The first hyperbol a ext ends .

from a: inches to inches the se cond h yperbol a from

a: ,

a: inches to x 15 inche a n d the third h yperbol a from s,

a: 15 inche to s inche Fo the firs t curve the va lues of

a: s . r

a a d c of equa tion 24 7 a a
n an d for the second
re c

curve a d c for the third curve

an a

a d c
n 1 24 7 X 10 These va lues were a ga in found by tri l a d
5
. a n

error Proceeding as be fore by the id of equa tions 248 to 252

.
, a
1 80 DES IGN OF IM PO RT ANT PUM P PARTS .

inclusive the ra di a l an d ta ngenti a l stresses a t a y point of the Shroud

,
n

ring ca n be found Keeping the speed a t 42 00 r p m a n d be gin

. . . .

ning with the ra di al stress equa l to nothing a t the inner dia meter of
the shroud ri ng the following va lues c an be obta ined :
,

Fo r the fi rst hyperboli c section ,

—
12 368 86 7 4 X 10 515
2 76 0 X
—
7 122 59 1 5 X 1 0 01 5
57 4 7 X 10 03 2

b; 4 17 0 X X
13 7 1

02 13 1 1 X 17 20 X

Fo r the second hyperbolic section ,

7
6 7 33 0 38 53 x 10 51 159 5 x
a
.“ 38 760 2 1 30 x 10 b 52 6 9 x
51 1 1 18 x 1 88 5 10

52 1079 x 10
-
9
12 3 8 x 10

a .. 1 3 45 1
57 7 0

Fo r the third h y perbolic section ,

10 2 6 7 13 05 X 1 0 11
b1 9 2 58 0 02

59 1 0 57 9 5 X 10 10
61 5509 X 10 2

46 9 8 X a ,” 183 1 X
9 4 16 X 10
—
10
a t
“
19 4 1 X
mm

a t”

Since i the ra di l tre

o r,” s a t the periphery of the Shrou d
a s ss

ring the stress is equa l to nothing ; hence

,

0 a .. 2 5 7 1 1,

2 8 36 8 pounds per squa re inch

1 82 DESIGN OF IM PORT ANT PUM P PARTS .

T ABLE II .

The a verage tre in the hroud ring due to its o w weight

s ss S n ca n

be found by mea ns of equa tions 2 57 2 58 a d 2 59 The s umm tio , ,

n . a n

of the three v lues of I of equa tion 2 58 for the three hype rbo lic se
a , c

tions of the hroud ring is

S

21 9
12 ( L 9 6 7 L 88 0)

x 10 5

2 I in .
‘
3
pounds .

The summa tion of the three v lue of A from equ tion 2 59 i a s a s

1 1 1 1

x 10 5
1 1
x 10 5
x 10
.

2 A s qua re inches .

S ub tituting the e va lue in equ a tion

s s s 2 57 , the a vera ge ta ngenti l a

s tre in the Shroud rings i

ss s

I
Avera ge 6 , N S z

A
2 1 730 pound per squ a re inch
s

This a vera ge t a ngenti a l stre s checks well with those plotted in

s

Fig 88 an d ma y therefore be considered a check on the previous

.
,

c a lcul a tion .
 IM PE LLERS .

The stresses in the Shroud rings a re somewha t high when com

pared to those in the impeller a n d therefore the thickness of the
,

sh oud rings Shoul d be ch a nged so as to decre ase the m a xim um

r

tan genti a l stress This ca n be a ccompli shed by increas ing the

.

thickn ess of the ring a t the inner di a meter a d proba bly decreas ing n

i ts thick ess a t the outer di a meter

n The ra di a l elonga tion of the .

Shroud ring Should be a pproxim a te ly the sa me as the ra di a l elonga

tion of the impeller a t equa l ra dii If this is not the cas e the pa rt
.
,

ha ving the sma ller elonga tion will ca rry the grea ter sh are Of the
bla de loa d This fa ct is sometime ta ken a dva ntage of when it is
. s

desired to ha ve the impeller disk ca rry the bl a de loa d an d relieve

the sh roud ri ngs of this dea d loa d In the present problem if the .
,

di mens ions of the shroud rings a e not ch a nged the impeller disk will
r

a ls o be stre ssed some wh a t by h a ving to s upport p a rt of the Shroud

rings themselves the loa d being tra nsmitted through the bl de

, a s

bec a use one end of the bla de is held by the Impeller disk while the ,

other end wa nts to move with the Shroud rings ; the l a tter ha ving the
grea ter ra dia l elonga tion .

If a impeller i des i gned without Shroud rings it is import a nt to

n s ,

investiga te the bending stresses or more correctly t h e crumpling

, ,

s tress es in the bl a des

,
In Fig 89 the bl a de is Sho w n as str a ight Up
. .

to a point C The weight of the tri a ngul a r piece AB C is at in

.
,

which S i the w eight of the m a teri a l per cubic inch

s The ra dius Of .

003
the center of gra vity of this tri a ngle is 7 c os The centri fuga l ; '

¢z
force of this piece at N rpm . . . is

F N 93‘S 2
r.
f ‘
2 c os ¢2

The force tending to bend the piece AB C a bout the line AC is F c o s 65 2,

ab
a n d it a cts a t a di st a nce of from th a t line giving a moment .

g
a
M 0 00002 842 N 2
Sr l co s <
9;

The se ction modulus a t AC i s

1 84 DES IGN OF IM PORT ANT PUM P PAR TS .

an d hence the bend in g or crumpling stress produced is

M ’
a
z z
b
N Sn co s ¢1
7 t ( a
2
b)
?

in which
r . is the ra dius at impeller inlet in inches , ,

( 01 i s the a ngle between the bl a de an d a t a gent to the circle wh o se n

ra dius is r; ,

a i s the width of the impeller bl a de in inches , ,

b is the lengt h of bl a de from impeller tip to point C in inches , ,

t is the thickness of bl a de in inches ,

.

Fi g] 89
.
.

37 . DIAPH RAGM S , C ASIN GS, AN D H E ADS .

i a phr gms c as ing

The d a d he a ds of centrifuga l pump s m u
,
s, an st

Often with sta nd high w ter pre ure a n d their trength c lcul a ti o n s
a ss s, s a

a e im port
r t Th e c as ting
an . u u lly of a complic a ted s h a pe
es s a re s a
,

a d it i r ther di ffi cult to d termine the m

n s a imum stresses to w h ich
e ax

these pa rts a e subjected w hen under hydra ulic press ure T h e

r .
1 86 DESIGN OF IM PORT ANT PUM P PAR TS .

if the cross section is well chosen this method Of as cert a ining the
stre sses w ill be sa ti s fa ctory bec a us e the m a x imum s tresse s w ill not

grea tly exceed the mea n tre se W Sh ll di cu l ter the proper s s s . e a s ss a

choice of cross se ction but will fi rst give a gra phica l method of finding
,

the modulus of section of a y cro ection n ss s .

Le t the he vy line in Fig 9 1 repre ent one h lf of the cro s section

a s . s a s

of a di shed di a phra gm of which AB i the vertic a l center line Firs t ,

s .

the origina l cross section is mas sed a bout the c nter line AB ; tha t is e ,

a ll hori ont a l w idth s Of the origin a l cross section a e moved over

z r

hori onta lly to the center line as for in t nce di t nce CD equ ls
z , ,
s a ,
s a a

dista nce C 6 etc This w cro ection a bout the center line
’
,
. ne ss s

h as the sa me modulu f ection as the origina l cross ection s O N ext s s .

ass ume a rbitr arily the center line f gr vity of the cro se ction O a ss .

O both s ide Of thi center line Of gr vity la y ff a uffi cient number

n s s a o s

of meridi a n lines p a ra llel ther to t king c re th a t there i meridi a n e ,

a a s a

line a t all point where there i a sudden ch a nge of cro ection

s s ss s .

Assume t wo a bitra ry line p a r llel to a n d equidi t a nt from the

r s a s

center line Of gra vity out ide f the cro section but prefera bly s O ss

close thereto La y o ff point 1 to 10 so th a t di st a nce 0 1

. s
’ ’ '— ’

equa ls dista nce 0 1 a l o di t a nce 0 2 equ l to di ta nce 0 2 etc

—
,
s s
’— ’
a s —
, .

From point 5 dra w a line to point which i a t the inte s ection

'
a, s r

Of line 5 w ith the outline of the m a sed cro ection Line 5 ’

s ss s —
. a

intersects the a rbitra ry line t b C onnect b to 0 by a tr ight line a . s a .

This line b O intersects line 5 a t —

Simil rly con ect points 6 . a n

1 to 10 to corresponding point s of intersectio of the meridi a n lines

’ '
n

1 to 10 wi th the bound a ry of the m as ed cro ection Where s ss s .

the e line inter ect the a rbitra ry line a re points to be connected to

s s s

the origin 0 The inter ection of e ch of these lines to the origin 0

. s a

with the origin l meridi a n line w ill give inters ection point on ea ch
a s s

meridi n line imil r to th a t of 6 By connecting a ll these points the

a S a .

cro s h tched ection a hown in the figure i found If the center

s a s s S s .

line of gra vity was correctly chosen the a rea of the cross h a tched
portion a bove the ter line of gr vity i equa l to th a t below If
cen a s .

thi i not the c ase a w cent r li mu t be t ken a d the work

s s ne e ne s a n

repea ted The modulus of ection c a n then be Obtained from the

. s

formula
AK 2

in w hich A is the a rea of the cross h a tched portion on both sides of

1 88 DESIGN OF IM POR TANT PUM P PAR TS .

the center line Of gra vity only one h a lf of which a rea is Shown i ,
-
n

the figure bec a use only one h a lf of the ection of the dia phrag m -
s

is Shown F is the di t a nce of the extreme fi ber from the cent r

. s e

line of gra vity a d K the dista nce of the arbitra ry lines from the
n

Ofl g n O
.

Ha ving found the mod ulus of section the me a n stress a t the e x ,

treme fi ber a be fou d If the dia phra gm i supported a ga ins t a

c n n . s

ra bbet a t it outer circumference a n d i uniformly loa ded the

s s ,

equiv lent concentra ted loa ds on ea ch Side of the center line w ould
a

be a t the center of gra vity of the a rea of e a ch se micircle or a t

times the ra dius of the e micircle The support a round ea ch emis . s

circumference an be repl a ced by a equiv lent rea ction a t the center

c n a

of gra vity of the emicircumference or a t s times the ra dius Of

the se micircle The tre a t the extreme fiber is equa l to the
. s ss

bending moment divided by the modulus of section If W is the .

tota l press u e on the di a phr gm in pound then one h a lf W c an b e

r a s, -

co idered concentra t d a t the center of gr vity of the a rea Of ea c h

ns e a

se micircumference The di a phra gm c a n n o w be repla c d by a beam

. e

whose cro section issequ a l to the cross section of the di a phra gm

s

supported a t t wo point time the ra dius of the di a phragm

s s

di tan t from the cent r of the bea m a d loa ded wi th two cone n
s e n e

t a te d lo a d
r times the ra dius of the di a phra gm from the center
s

Of the be m a .

The mea n stress a t the extreme fiber of a be a m loa ded as sta t d e

W
a bove w ith t wo concentra ted lo a ds
2
equ a lly di t a nt from a n d o n —
s

both sides of the center of the be a m a d supported a t two points n

equa lly di ta t f om a n d on both sides of the center line of the beam is

s n r

in which L is the dista nce from the point of upport to the nea rest c o n s

c n
e t a t
r d lo a
e d or in thi s c ,
e L R as R ,

R being the r a diu of the di a phr gm s a .

I f thi s me n tress ias kept belo

s w 3 000 pound s per s qu a re inch for
cast iron or 9 000 pound for c t steel the a bo ve method wi ll give
,
s as ,

sa fe de ign s B ut thi
.
method mu t be u
ss ed with ome judgment s s ,

for it is easily see n th a t the cross section a be chosen so Slight a t c n

 DESIG N OF IM PORT ANT PU M P PARTS .

line the a rea

,
is s qu a re i ches
n . H ence the modulu of s se c

tion is
AK 2
X 6 2

The tes t loa d to which this c a i g is to be ubjected i 240 pounds

s n s s

per squ re inch ; hence the tot l lo a d is 3 10 000 pounds The lo a d

a a .

is ta ken on the entire a re a of t h e c as ing up to the bolt ci rcle which ,

is 3 inches di a m e ter mea n stress . ex treme

Fi g 9 2. .

fi be rs of the c as ni g is f The dista nce L equa l s 0 2 12 2 R

.

inch es . He n c e

3 10 000 X
8 2 10 pounds per squ a re nch I .

2 X 91
 C R IT IC AL SPE E DS .
1 91

38 . C RITICAL SPE E DS .

The phenomenon of critic a l spe eds is Of grea t importa nce in the

de ign of high peed centrifuga l pumps The sh ft a n d the impellers
s -
s . a

form a rot ting element which in pra ctice is not in a bsolute mass
a

ba l an ce When such sha ft is rota ted it experiences mo e or less

. a r

violent vibra tions The intensity of these vibra tions depends

.
,

a mong other things on the peed Of rota tion the vibra tions rea ch
, s ,

ing a very s udden ma xi mum a t certa i n pa rticul a r speeds known ,

a scritica l spe ed It is therefore importa nt to h a ve the revolving

s .

element in as good ba l a nce as possible a n d to a void run n ing the ,

Sh a ft in the vicinity of i ts critic l s peed s a .

A revolving element ( ha ft an d impellers ) ma y be out of ba l a nce S

either dy a mica lly or both dyna mica lly a d sta t ica lly Assume a
n n .

rota ting element mounted on a hori ont a l Sha ft the l a t t r resting z , e

upo n pa ra llel a d level w y If the center of gra vity of the enti re

n a s .

rota ting element lie xa ctly in the cent r line of the Sh a ft the mas s
s e e

will then be in indi fferent equilibrium ; th a t i the sha ft will h a ve s,

no inherent tende cy to rota te to some pa rticul a r position In a ll

n .

pra ctic a l c ase s the cen ter of gra vity does o t lie in the center line n

of rota tion a d w h a ve a c as e imil a r to th t hown in Fig 9 3

n e s a s .
,

where m repr ent a n unb a l a nced mas s a t a di st a nce

; es s from the r,

a xi s The sh a ft will o w be in
. n

equilibrium in only two positions .

I n ll other po ition it will be

a s s

s ubject to a turn ing moment Of

m m s i t where 65 is the a ngle m a de
n c ,

by a norma l from m to the axis ;

w ith a vertic l pl a ne through the a

a xi s If a nother m a s m is a tta ched to the revolvi ng element a t

. s .

a di sta nce from the a xi s so th a t mm m r m a n d m bei n g

rz , g z, l g

in the sa me pl a ne pa ing through the a xi but upon opposite ides

ss s s ,

the indi ff ere t equilibrium of the sh a ft will be re tored or to use

n s , ,

the common phrase the h a ft will be in ta tic ba l a nce Evidently

,
s s .

a vertic a l sh a ft is a l wa y in st tic ba l a nce s a .

A h a ft while t a tic lly in b l a nce ma y be ubject to evere

s s a a s s

deflecting forces when se t in rota tion a in the cas e ju t illu tra ted ,
s s s .

The center of gra vity of the t wo unb l d ma lie in the axi s a an ce s se s s

Of rot a tion ; but their centrifug a l force s a e not b a l nced but c o n r a

1 92 DESIGN OF IM PO RT ANT PUM P PARTS .

s ti t moment ca us ing oppo ite force in the bea rings a d sub

ute a s s n

j e t
ci g then h a ft to s evere deflecting
S force s which m a y be high

enough to rupture it The correct wa y to neutra li e the eff ect of

. z

s uch unba la nced m s ses i to b l a nce the rota ting element dyn a mi
a s a

ca lly by a dding ba l a nce weight in an y two pl a nes perpe ndic ul r t s a o

the axis of the element prefera bly the t wo end pl nes Referrin g ,
a .

to Fig 94 the e ff ect of the unba l a nced mas s m i neutra li ed by

.
, ; s z

a dding two m a sses m an d m a t ra dii in the two end

’ " ’ ”
an d r l . r, .

Fi g 9 4. .

pl anes E
'
an d pl ced norm lly to the axi provided m m a
E a a s, l, l
'
nd

m. a ll in the s a me pl a ne through the i an d provided th a t

”
a re ax s

‘
m;
H II
7 " 3 17 1

m1 ml
, ’ ' Il II ll
rl a l rl al

When the h a ft i rota ting th e three mas es crea te centrifuga l

S s es s

moment which a in quilibrium ; th t is they con titute a sy tem

s re e a ,
s s

of mom n t who e a rithm tic l um a t y point i equa l to nothing

e s s e a s an s .

Simi l rly t h m as be b l nced by tw m e m an d ' "

a e s 771 2 c a n a a o a ss s 2 171 2

a t r a dii a d like w i a y other unb l nc d m e in the

'
an d rz n se n a a e as s s

rota ting element c a n be b l a nced If de ired a ll the ba l a ncing SO a . s ,

m as se in the two end pl a ne ca be combined into re ulta nt mass

s s n a s

in ea ch pl a ne Ther fore proper ma ba l a nce necessit t not only

. e ss a es

h a ving the center Of gra vity of the revol ving m as coincide with it s s

center of rot tion ( t tic ba l a nc ) but l o elimin ting the O c lled

a s a e a s a S -
a

centrifuga l mome t Of the dyna mic a lly unb a l a nced m e (dyn a mic
n s as s s
1 94 DES IGN OF IM PORT ANT PUM P PARTS .

e itivene
s ns d the s lightes t di sturba nce ma y se t it in ve ry
ss an

violent vibra tion These critic a l peed a n be determined with

. s s c

con idera ble a ccura cy from the di mension s of the sha ft an d th

s e

ma n er in which it is lo a ded a n d s upported The lowe t critica l

n . s

speed i usua lly c a lled the firs t critica l spe e d a n d the rest a re n um
s

bered uccess ively I the present desi gns of highs peed centrifuga l
s . n

p umps the Sh a ft seldom rea ches the third critic l s pe ed It runs a .

either between the second n d third critic a l Speeds or between t h a e

fir t an d second an d frequently below the fi rs t ; s o th a t in ne a rly a ll

s

c as es the first c i tica l peed sho uld be determined to ma ke sure th a t

r s

the no m l speed is either well below or well a bove this critica l spee d
r a ,

a d only i
n some des i gns i it necessa ry to determi ne the se con d
n s

critica l speed .

Fig 95 . .

ma gine a di k mounted on a h ft (Fig 95) so th a t its center

I “
s S a .

of rot tion is a t C d let the di k be Slightly out Of ba l a nce so th a t

a ,
an s

i t center Of gra vity i

s t S Le t the dotted a d d as hed line fro m
s a . n

s upport A to upport B repr ent the center line of the sh a ft when

s es

s ta tiona ry a d bent by the weight of the di k (weight of h a ft being

n s s

neglected ) When the h ft i t into compa ra tively slow rota tion

. s a s se

the centrifug l force f the unb l nc d mass will Spring the sha ft
a O a a e

8 di t ce 1] a d c a u e it to whip If m i the ma s Of the di s k

s an n s . s s ,

on it a gul a r v locity
s n d the di t a ce of the center of gra vity
e , an e s n

from the center of rot tion then m (y ) < i equa l to the c e n t ifu
a ,
c i1
z
s r

g a l force e ert d on the sh ft a d is b l a nc d by the el as tic oppo s ing

x e a n a e

S St d lee St m T ub i
o f full d i u i f th th y f iti l
o a s
’
ea r n es or er sc ss o n o e eor o cr ca

s pe e d s .
 C R IT IC AL SPEEDS . 1 95

force of the Sh a ft The el as tic opposing force Of a sha ft is propor

.

t io a l to the deflection ; hence if a i a const a nt depending upon

n s

length of Sh a ft tyle of support etc the el as tic resi ting (opposing)

, s s, .
, s

force f the Sha ft i P a y C onditions of equilibrium dema nd th a t

O s .

2 74

H ence the deflection is

If the a ngul ar velocity is increas ed y also i ncrease s an d become s

infinitely l a rge when a m 2

e:0 or when ,

This va lue of w is lled the critica l a ngul a r velocity From this it

ca .

ca be seen tha t the Sh a ft ca nnot rea ch t h e critica l spe ed except

n

Fig 96 . .

when it is con stra ined by guides or by bea rings which a c t as gui des ,

or w hen the velocity increases ra pidly th t the Sha ft while pas s ing
so a

through it critic l peed h as no time to deflect excess ively

s a s .

Theory a d pra ctice gree th t fter the critica l peed h as b en

n a a a s e

pass ed a e w condition Of st ble equilibrium e i t a d th a t the

n a x s s n

center of the Sh a ft C w e ch a ge it p sition w ith t he center of

no x n s s o

gra vity S a s how in Fig 9 6 The deflection y is o w found from

s n . . n

the formula
m (y e ) to
2
a y
1 96 DESIGN OF IM PORTANT PUM P PAR TS .

The more is increase d the sma ller y becomes an d with infin itely
00 ,

high rota tion y tha t is the dis k will rota te a bout its cent r Of
e, ,
e

g vity In pra ctice ca reful b lan cing will ma ke 3 very s ma ll so

ra .
, a

th a t the ra nge of the critic a l spe ed will not be very gre a t ; th t is a

noticea ble vibra tions us ua lly do not begin until the critic a l peed s

i s nea rly rea ched an d die wa y oon a fter the critic a l pe ed i s p as ed

, a s s s .

Good run ni n g c a therefore be Obt a ined when the opera ting s peeds
n

a r removed from the proximity Of the critic a l s pe ed s

e .

A method of c a lcul a ting the critic l peed of a rot a ting s h a ft with a s

a y nu
n mber of di fferent di a meters an d wi th y di stribution of an

masses a d for a y number of suppo t h as been devi sed by one of

n n r s

the a uthors a n d us ed successfully for a number Of yea rs The c a l .

cu la t io n s for critic a l peeds h a ve checked very well with the observed

s

c itica l p ed The meth d will be e pl a ined by both gra phica l a n d

r s e s . o x

ma thema tic l solution The gra phic l method a i ts the c alcul tor
a s . a ss s a

in quickly detecting errors in the computa tion a d helps gre a tly in n

unde ta nding the method It i therefore trongly recommended

rs . s s .

The genera l theo y of thi method i as follows A Sh a ft with an y

r s s .

a rra nge ment of support s ( be a rin gs ) is lo a ded w ith forces W W 1, 2,

W a t right a ngle to it a xi w hich c a u e de fle c ti my

-

3, s s s, s ons , ,,

y a, t the point
a w h e re the e force s a re pplieds I f a ll the e s a . s

forces a increas ed time their ma gnitude the deflec tions will

re 2: s ,

a sl o be increa d tim seThi gener l t tement i pe rfectly o b

at es . s a s a s

vi u o s Suppo
. the h a ft i lo ded by centrifug l force in tea d of
se s s a a s s

st a tic force as i the c e of rot t ing h ft with impeller not in

s, s as a a S a s

a b olute ba l n
s A long a the p ed f rot tion i s l w the d e fle
a ce . s s s e O a o , c

tion due to the centrifuga l force a ma ll but as these forces

s s re s ,

increase the defl ction l o increas e in like a mount It is not until

e s a s .

the critic a l umb r of r volution i r a ched th t the centrifuga l

n e e s s e a

forces become qu l to the l astic r i ting force of the sh a ft a n d

e a e es s s

a st a te of un t ble equilibri u m e i t for hould the deflections

s a x s s, s

from a y c a u b li ghtly i r d th e n t h e c trifuga l forces

n se e s n c e ase en

would incre as e in the a me propo tion I other words a t the s r . n ,

critic a l number of revolutio s the sh ft is in neutr l equilibri um for n a a

198 DESIGN OF IM POR TANT PUM P PAR TS .

a pproxima te methods for ca lc ul a ting the cri tica l speed use the sta ti c
deflections A good physica l conception of w ha t a ctua lly oc curs when
.

a Sh a ft i s rot a ting a ids m a teri a lly in unders t a ndi n g the phenomenon

of critica l speed A loa ded sha ft held between support deflect

. s s

an d i ts a xi s forms a curve of t a tic deflec tion When the sha ft i s . s

s lowly rot a ted it revolves a bout thi s be t a xi s simil a rly a would a n s

rota ting rope held between tw up port with con sidera ble sa g If o s s .

the rope i turned equ a lly a t both end it will rota te a bout thi cu rve d
s s s

axi s .This represents exa ctly the rot a tion of a loa ded haft If s .

the rota ting sh a ft is in perfect ba la nce it will a l wa y continue t o s

rota te in this curved position Ea ch unba l a nced mass on the rota t .

ing sh a ft will on the other h a nd exert a certa in ce trifuga l for

, , n ce

upon the Sha ft tending to deflect it from it ta tic position Whe

, s s . n

the e centrifuga l forces become la rge enough they will c a use the
s

sh a ft to whip a bout i ts curved a xi s of rot a tion We an im a gine th e . c

sh a ft re a ching s uch a s peed th a t the centrifuga l forces become l a rge

enough to ba l a n ce the el astic res isting force of the sha ft exa ctly a d s ,
n

this will crea te a sta te of neutra l equilibrium for a n y deflection If .

the rot ting element cons ist of a ha ft Of uniform di a meter a d

a s s n

c a rrying a single concentra ted lo a d the st a tic deflection curve a d , n

the el as tic curve of the sh a ft a t the critic a l spe ed must be simil a r

beca u e the sta tic weight a d the centrifug l force of the unba l anced
s n a

m as a exa ctly proportion l to one a nother In this cas e t h e

s re a .

st a tic deflection curve c a n be u d for det rm ining the cri tic al s peed se e .

In a ll other c as e the sta tic deflection curve will not give a true indi
s

ca tion o f the critica l s pee d a lthough in m an y cas es it gives a very

,

close a pproxima tion The more nea rly the sta tic deflection cur e
.
v

is imil a r to the el tic curve Of the Sh a ft a t the critic a l s peed the

S as

more nea rly correct will be the critica l pe d c a lcul a ted from th e s e

st a tic deflection s .

Another formul a for critic a l peed of a y sh a ft bas ed on the s n

equ a lity Of the potenti a l a d ki etic energies Of the h a ft c an b n n s e

derived The pot nti a l ene gy of a h a ft rryi g loa d W W W

. e r S ca n s 2, 3,

etc (centrifuga l forc s be i g con idered a lo ds) a n d h ving de fle c

. e n s s a , a

tions y y y etc a t th se l d is
l, g, a, .
,
e ea s

4 (l l l
‘
'
W yz z Ways

If the e s le a ds while m a inta ining their deflections y ] , y a, etc .

,
 C RITICAL SPEE DS . 199

are rota ting wi th an a ngular velocity of w ra di ans pe r s c e ond their ,

kinetic e ne rgy is

(W i ll/1
2
Way z
z W ay s
a

At an y s peed the ha ft will fina lly ass ume s uch a Sha pe tha t
s

28 1

so th a t when W W W etc an d y y y etc a e known the

1, 2, 3, .
, l , z, a, .
, r

corre ponding c an be computed But a t the critica l pee d the

s to . s

Sh a ft i in indi ff erent equilibri um ; th a t i a t a n y particula r point

s s,

on the sh ft the deflection may be a s umed a t ra ndom The de fle

a s . c

tions a t a ll points will Of cour e h a ve to a da pt themselve to the s s

as s umed deflection in a ccord a nce with the di me ns ions of the sh a ft

, ,

distribution of loa d ki nd of support etc If the sta tic deflection

,
s, . s

of the Sh a ft a e computed the l a rge t deflection or the l a rge t Wy

r , s , s ,

ma y be as umed to be exa ctly co rect a t the critic a l speed The

s r .

sm a ller t a tic d flection s will be more or le s in error but thi s error

s e s ,

need not b very l a rge beca use the express ion for w i fa irly i n sen itive
e s s

to modera te errors in the sma ller deflections Therefore the critic l . a

spe ed ma y be express ed a pproxima te ly as

N
7 l
2 ‘ ‘
1 f Wy sz z
2
Wei/3 2

W1y 1 + W2y2 + W3y3

187
” 1 171
7 2
Wzysz
” 3 113
7 2

This formul a ometimes gi ves clo e res ults but should not be used
s s ,

except as a tri l va lue The more nea rly the sta tic deflection curve
a .

i simila r to the el as tic curve Of the h a ft when a ppro a ching the

s s

critic l peed the more nea rly correct will thi s formula be
a s .

From the foregoing it is clea r th a t before the critic a l peed of a s

sh a ft i computed it i d e ir a ble to kno w h o w to c a lcul a te the t tic

s s s s a

deflections of a sh ft a t va riou point for a n y given di t ribution f

a s s s O

loa ds a d suppo rts The underlying principles for obta ining st a tic
n .
2 00 DESIGN OF IM POR TANT PUM P PAR TS .

deflecti ons are expla ine d in most te xtbooks on mec ha ni cs an d will ,

be reviewed here very concisely If 3 denote s the curva ture or .

,

cha nge of ta ngent a t a n y point on a sha ft M the be nding moment

, ,

a t the point I the moment of inerti a of the cro ss section of the s h a ft

, ,

a n d E its coe ffi cient of el as ticity then it c a n be prove d th a t

,

1
E
53 ?
Integra ting this we get
dz 01

an d inte gra ting a ga in we get

dz C a: C:

the a bove C is the ta ngent a n d C the defle ction a t the point

In 1 :

where 0 Since a t th a t point a ll integra l mus t necess a ri ly be

a: ,
s

equ l to nothing I computing C a n d 0 let the deflection s for

a . n 1 2

a: l a dx , l be 11
n 0 ; 1 a d I being the loc a tions Of t he be a r
g, 1 n :

ings which a re l 1 l inches a pa rt Then f om equa tion 2 8 6

, a 1 . r

dx + Cxla + Cz

dx + C l l l + CQ

S ubtra cting equ ation 288 from equa tion 287 ,

12

4 C l (12
“
ll )

w hence

an d 02 be solve d from either 287 or 288

c an .

Genera lly the origin i so chosen th a t y s 0 for a: 0 . Then

%f f
1
flI
Cz = 0 an d 01 —
dx
T
.
2 02 DESIG N OF IM PORTANT PUM P PAR TS .

l
M 1271 ]
dic
i l
-

[
Ma r l s
'

z M 3171

fi fl M I + M r) $ z

MI + M2 $ 2

2 I

B efore the second integra l of equa tion 29 3 is ta ken it is a dvisa ble

to solve for the v lue Of the cons ta nt C In order to do this let
a .

a: in equa tion 2 9 3 Then .

2 61 13 1232 2 96

hence
I
2 6 .
[2 M me (M . M .)

S ubstituting this va lue of C in equa tion 29 3 we get

1

The integra tion of equa tion 2 98 between $ 2 an d x1 is

M
( IQ32 M 2131 )

2 M a w: (3 2 x1

—
xl ) + ( z
x ( ii 2 99

1
67 [
3 (31 1132 M 255 1 ) (332 271 )

l (M 2 M I) 6
]
‘ ‘ 2
“
171

“ M
)
0
 CRITICAL SPEEDS .
2 03

2 M 1 + M2
+ ( x2 _

The pra ctica l use of the a bove formula will be eas ily understo o d
in the following illus tra tion of the method of Obta i n ing the sta tic
defle ct ion curve .

To illu tra te the method of ca lcul a ting critica l s peeds it would be

s

well to calcula te the critica l speed of a pM p Sh a ft Fig 9 7 pre . . re

s ents a pump s h a ft c a rrying four impelle rs W W W a d W an d 1, -

2, 3, n 4

s upported a t e a ch end in a be a ring Th di st a nce be tween cent r . e s e

lines Of be a ring a n d impellers a e a l so Shown Th e di a meters of the

s r .

Sh a ft a as follo ws :
re

Di a meter of Sh a ft between R ; an d S1 is inches .

Di a meter of Sh a ft between S ; an d 8 2 is inches .

Di a meter Of Sh a ft between S 2 an d 8 3 is inches .

Di a meter Of h a ft between S
s . an d S . is inches .

Di a meter of Sh a ft between S . an d S. i s inches .

Di a meter of sh a ft between S . an d S. is inches .

Di meter Of h ft between S
a s a . an d S7 is inche s .

Di a meter of s h ft b tween S a e , an d R2 is inch es .

The ta tic deflection of the h a ft i first determined by trea ting the

s s s

s ha ft as a be a m supported a t the cen ter lines Of the bea rings a d n

loa ded with concentra ted loa d a t the center line of ea ch impeller s .

The weight of the ha ft wi ll be considered uniformly di tributed in

s s

thi cas e bec a use the sha ft is nea rly of unifo m di a met r If the
s r e .

di a meters of the h a ft va ry grea tly the weight c a be cons idered

S n

as a serie of concentra ted lo a d a pplied a t the center of e a ch length

s s

of uniform di meter ea ch loa d being the weight of uch length

a ,
s a .

The c a lcul a tion s for t tic deflection c an be conveniently ta bu

s a

l a ted a d T ble I gives the res ult of the ca lcul a tions for the pump
, n a s

S h a ft illustra ted in Fig 9 7 . .

C olumn 1 gives the v a rious po int s a long the h a ft a t w hich there s

is either a ch a nge of cro s ction or a point of a pplic tion of loa d

s se a .

C olumn 2 gives the di t a nce (in inche ) between succe ive point
s s ss s,

a s,for exa mple the dista nce from R to S i 7 inche th a t from S

, ; 1 s s, ;

to S is : i che etc n s, '

DES IGN OF I M PORTANT PUM P PARTS
.
 CRIT IC AL SPEE DS .
2 07

C o lumn 3 gi ves the (

dis t a n c ein inches ) from the o ri gi n R to the 1

point in ques tion as , ,

for ins ta nce the dista nce from R to S is
, ; 3

inches .

C olumn 4 give the bending moment ( in inch pounds ) due to

s s

the weight of the sha ft The weight of the h a ft is 2 14 pounds a n d

. s ,

con idering this weight uniformly di tributed ea ch rea ction a t the

s s ,

bea rings is 107 pounds The bending moments under the loa d .

point a given in Ta ble II

s re .

T ABLE II .

B di
en ng Mo me n ts i n Inc h Po ud n s .

‘
—
1o7 x 36 + 2 11
96
- -
x ac ! <
‘
1
- -
X 22 X —
18 10
6

The curve of bending moment due to the weight of h a ft is shown s s

a
s curve M in Fig 9 7 d i a pa ra bol a pas ing through the ordi
’
. an s s

n t a given i T ble II
a es s n a .

C olumn 5 of T ble I give the bending moment (in inch pounds)

a s s

due to the co centr t d lo d The rea ction a t the uppo rts due
n a e a s . s s

to the concentra ted lo a d determined in Ta ble III s a re .

T ABLE III .

Lo a d s.

Wl 2 17

W; 2 12

W; 200

W. 18 6 186 X

T o ta l re a c t i o n s 37 2 7

The bending moments due to concentra ted loa ds under the loa d
points a re given in Ta ble IV .
2 08 DE SIG N OF IM PORT ANT PUM P PAR TS .

T ABLE IV .

Un d er
B di en ng Mo men t in Inc h Po ud
n s .

The curve of bending moment due to concentra te d loa d i hown s s s s

as curve M in Fig 9 7 . .

C olumn 6 of T a ble I gives th combin d or tot a l bending moment e e s

( in inch po u nd s ) under the v riou point on the h a ft C urve Ma s s s .

in Fig 9 7 gives the curve of combined bending moment The

. s .

a ddition of the ordina te of the p ra bolic curve M s d the ordi a an

na te of curve M will give very li ght curva ture to the line of

s a s s

curve M but the e lin h a ve b en t ken a tra ight line for im

,
s es e a s s s s

p lic i t y t h,
e er or beingrentirely n gli g i ble e .

‘
C olumn 7 give s the momen t of ine ti ( in inche ) of the v rious s r a s a

cro ection of the h a ft The moment of in rti a of the cro

ss s s s . e ss

se ction of the h a ft from R to S is a d from S to S it i

I . s n I 2 s

etc .

M
C olumn 8 give s the v a lues of (in inch pound s ) a t the v a ri
3
—
—

I
ou s point a long the
s h ft M b ing the tota l b nding mom nt t
s a , e e e a

the point a n d I it moment of i erti a of cro ection C urve I in

s M n ss s .

?
7
F ig 9 7 . gives the 3 cu ve r . At point s S I, 3 2 , S 3 , S7 the cro ss

s ection ch nge a s s udde ly n an d ther fore the

e
gcurve a l s o ch a ng es

Al
sudd nly th a t t e ch of the point th re a t wo va lue for
e ,
so a a se s e re s
l
C olum 9 d 1 0 give t h r u lt of t h int rme di t
ns an tp n e es s e e a e s e s ecos

sa ry to c a lcul a te the v a lue of the fi r t int gr l of gdx From s s e a .

equa tion 2 95 we fi d th a t n

AI 51 2
)
272

I 2 I

$ 2
C olumn 9 gives the v a lue of s
I
an d column 10 the va lues of
M1 + Mz
2
210 DESIGN OF IM POR T ANT PUM P PARTS .

81 an d 82
4 tha t
or this va l ue
is , 56 60 X —
1 1 6 000 ;

betwee n S a d W i found by multiplyin g the value of ! dx

at S
2 n

6 1 2 8 0) by the di t a nce from S to W

—
1 s

th a t is s
f
z 2 ; ,

6 1 2 80 X 2 7 5 7 00 C ol umn 14 gives the va lues of the

—
.

he rof the equa tion from R to S is found by squa r ng the dista nce , 1 i

RS l a d dividing it by the moment of inerti a of the cr o

n section ss

19
2

of the h a ft from R to S
s so th a t
18
C ol u
, mn 15 1

va lue of thi member b tween R a d S is found by multi plying the

s e ; n 1

moment a t R (0 ) by 2 a dding it to the moment a t S

; 3 1 00) a d , 1
—
n

0 X2 3 100
6
from S to S i fou d by multiplying the moment t S
I 2 s 3 100 )
n a I
—

by 2 a dding it to the moment a t S

,
12 190 ) a d divi d ing by 6 or 2 n ,

C olumn 1 7 is derived from column s 13 , 14 , 15 , an d 16 , an d rep

f
Jll
resents (in inch pound ) th I f dx so lved as shown in equa
2
"
l -
s e ,

tio n 30 1 .

A
ff
l
C olumn 18 repre ent the dx from the origin (R ) to the
s s
—
1
l 0

point in qu tion Thi i determined by a ddi n g ll the va lues given

es . s s a

in column 17 b tw t h origin R a d the point in question C urve

e een e I n .

M
dx i Fig 9 7 re p r ent th e va lues
n B y dra wing a t r i ght
. es s es . s a
I
l in fro m R to th other d of this curve w c a dete mine the
e ; e en e n r

va lue E Y a t a y point by me uring the v rtic a l di ta nce between

n as e s

s uch a point on the curve a d thi tra ight line Thes e vertica l d i n s s . s

t a nces be plot t d from t h hori onta l b e line Rl a d the

can e e z as ,
n

re ulting curv B Y through t h ordi te would give a n E Y curve

s e e se na s
’

which would r p nt the d fl c tion ( Y) of t h h a ft multiplied by

e re s e e e e s

the co ta nt E ( ffi i t of l ti i ty for te l ta ken as 2 9 000

ns coe c en e as c , s e

Another w a y of fi ding thi d fl ction cu rve by purely gra phic l

n s e e a a

M
method would be to t a ke the fi r t i tegra l curve s n — —

I
d:c an d dete r
 CR IT IC AL SPEE DS 21 1

mine i ts consta nt of integra tion Thi i found by a tri a l a n d . s s

error method by c h osing a horizont l line like C C so th a t th e a rea

o a
’ ”

M
be tween the fi rs t integra l curve 7 dx a n d the consta nt line C C ' ”

is equa l on either side of thi s co t nt line th a t i the a re a C R C

’ ’”
ns a ,
s, l

is equa l to Thi rea b tween the li ne C C a n d curve ’ "

s a e

IlI
1l be quickly
da: determined by a pl a nimeter a d line C C
’ "
can n
—

shift d until the re as on either side of the consta nt line a re equa l

e a .

The econd integra l curve E Y a then be found by a llowi ng thi

s c n s

f
I
}
curve to repre ent the umma tion of the a reas betwe n curve ;
s s e

a n d line C C up t o a y point in que tion Fo in ta nce the ordin te

’ ”
n s . r s ,
a

on the E Y curve a t S i found by ma king it equa l to the a rea I s

between l ine C C the g curve from R to S the ordi

' "
and dx 1 I, a n d

na te for the EY curve under S 2 is found by ma king it equa l to the

‘
f;
4
a rea between line CC
’ ”
an d the dz curve fro m R l to Sz , an d

so on .

When no gr phic a l method i u d the ordin t of the E Y curve

a s se a es

ca be det rmi d m them a ti a lly from the v lu in column 18

n e ne a c a es

by ta king t h l as t v lue in th t column for poi t R d dividing

e a a n 2 an

1 1 1 51 480

multiplied by the va lue of x a t a y point will giv the va lues of CX

.
"

n e

in column 19 The v lue E Y or the o di a t s o f t h 9 cu ve a

. a ,
r n e e r ,
c n

then be found by a dding the va lue of the econd integr l in column 18 s a

to t he corre ponding va lue of CX in column 19 Thi i doing

s . s s

m thema tica lly exa ctly wh a t was done gra phica lly by d ra wing a
a

stra ight line from R of curve ? dz in Fig 9 7 to the other end

;
If .

of thi curve a d obt ining th e va lue of E Y by measuring the verti

s n a

c a l di t a nce froms y point of thi curve to t h e tr

an ight line B y s s a .

letting E 29 000 000 (for teel ) Y or the a ctua l deflection in inch s , , es,

can be determined for y point F u problem it is only e an . or o r n ec s

s a y to d t ermi e the d fl
r e t ion und r t h lo d
n e ec s e e a s .

Aft r d t rm ining the t tic defle tio

e e e ritic l pe ed c a be s a c ns, a c a s n

t ent a ti vely assumed but as sta ted before it is b tt e r to u

, se in mo st , e
21 2 DESIGN OF IM POR T ANT PUM P PAR TS .

cas es the a pproxima te estima te of critica l peed represented s by

equa tion 283 which rea ds ,

N 18 7 -7
W2y2 + 77 3y3 +
7
2 2

In the a bove equa tion W W W etc represent concentra ted 2 , 3, .

,

loa d a d therefore the weight of the sh ft which is a di stributed

s, n a ,

loa d c a nnot be t a ken di rectly into a ccount Some a llowa nce c a

,
. n ,

however be m a de by ub tituting for the ctu a l lo a d W W W

,
s s a s 2 , 3 ,

etc the equiva le t loa ds which would produce the tota l or combined
.
,
n

bending moment in column 6 of Ta ble I The equiva lent c o n ce n

s .

t a te d lo a d s a d re a ction c a be found as follows

r n s n

14 1 9 0

16 6 6 0 44 3 5 X

15 7 2 0 X X

16 6 60 X X

15 7 2 0 X

Th equiva le t conce tra t d lo a d an d their corres pond ing

e se n n e s d e fle c
tion ub titut d in q u t i 2 83 give
s s s e e a on

X 0 10 4 3 + 2 3 1 5 X 0 09 19 + 22 6 5 X0 06 4 9
. . . . .

N = 187 7

H ence the a pproxim a te or tri al critic l p ed i 6 18 revolutions pe r

, ,
a s e s

minute .

The centrifugal forces c ued by the equiva le t concentra te d a s n

lo a ds W W W etc revol ving a t 6 18 revolutions per minut

1 , 2, 3, .
, e
2 18 DESIG N OF IM PORT AN T PUM P PART S .

TABLE V III.

“ —

T u;
T 5
i n lb s
. .

1 018 3 540 4 500 -

5 800

—
3 180

7 285 —
7 840 4 7 100 !

4 2 500

1 132 4 3 530 4 5 300

4 4 550 —
78 200 '

1 887 4 5 205 —
28 700

4 5 850 4 00 000

1 187 4 5 405 4 0 500

4 5 050 4 2 5 400

1 187 4 5 520 4 0 700

4 5 280 4 45 100

1 108 4 5 800

_ 15 500 4 52 300

4 4 580 4 0 000

_3 500
1 4 75 200

4 2 235 4 1 100

4 0 8 10 4 84 300

0 401 — 0 705 — 4 700

_
8 500 4 80 000

2 8 14 —
5 775 1— 23 500

— 2 050 4 12 000

4 475 -
4 500

— 2 17 200
 C R ITIC AL SPE E DS .
21 9

T ABLE VIII .

lbs . in lbs
. . in ." l bs . in .
“
l l bs . in .
“ l bs . in .

4 3 500 4 3 500

—
13 500

— 1 18 900 -
3 143 4 50 300
—
601 700

-
6 593 —
33 500 —
3 15 500

— 918 200 3 773 800

—
586 500 — 7 495 4 05 100

— 1 6 10 800

—
8 1 10 4 2 500 — 757 500

— 2 368 4w 5 424 800 3 056 400

— 8 366 -
54 500 —
885 500
—
3 255 000

— 1 096 (l l ) 1 155 400

-
8 010 —

—
4 7 075 800 2 66 4 400 0 0919

— 893 (I! ) 7 443

- — 30 500 4 23 500

—
5 335 000

—
1 472 (KI) — 48 800 — 1 520 800

— 6 855 800 8 726 800 1 87 1 000 0 0646

4 20 000 —
5 037 4 0 000

—
7 696 3ll )
—
2 173 000 — 3 365 4 58 300 — 2 33 1 300

—
10 02 7 500

— 1 275 000 4 8 500 -

1 203 500

—
1 1 32 1 100 11 32 1 100
22 2 DESIG N OF IM PORT ANT PUM P PART S .

sti ff hubs or h a ve less forced fit on the sha ft It is necess a ry for t h e .

designer to ca lcul a te the di ffe ence between the moment of inert i a of

r

the Sha ft a lone a n d the equiva lent moment of inerti a of the sha ft
wi th forced n impelle s This is ra ther di ffi cult be ca us e the origi
-o r .

n l forced fi t of the impe lle rs no longer h o ld s w hen run ing nea r the
a n

critica l spe ed ; the impeller bore of course stretches as the spee d of

the Sha ft increas es The obliquity of the impellers when running
.
,

due to the de flection of the sha ft a lso influences the critica l s peed
‘
,

to some extent .

mo re co m ple te disc u uth o r s t ra n sla t io n

’
Fo r ssi o n of Sto d o ln
’
a see a a

Ste a m Turbi n es , or t he o ri gi n a l bo o k , f
o urth ed i tio n .
 VI . TYPES OF CENTRIFUGAL PUM PS .

39 . HISTO RICAL RE V IE W .

In giving the his to ry of the centrifuga l pump it is necessa ry to

in clude the hi to ry of the wa ter turbine be ca use the former is a
s ,

n a tura l a n d logi ca l outcome of the l a tter Wa ter wheel s were in .

use in very early times but the first wa ter turbi e which a ttra cted
,
n

the a ttention of scientist was a rea ction wheel invented an d built

s

by Professor Segn er in 1750 This turbine wheel a llowed the wa ter

,
*
.

to esc a pe from its peri phery in a ta ngenti a l direction an d thus rotate

the wheel by the ea ction of the di sch a rged wa ter Although Pro
r .

fesse r Segner a ttempted to esta blish a theory for the flo w of wate r

,

th rough thi s turbine it re mai ned for the grea t ma thema tici a n
, ,

‘
Le o nh ard Eu le r to expl a in the fund a ment a l l a ws of flow
r The .

la tter a lso suggested a s t of guide van es for Segner s wa ter turbine

e
’
,

which increas ed its effi ciency .

Fo nea rly ni nety ye a s a fter Euler s work m a them a tici a n s a n d

’
r r ,

physici t p id no a ttention to his theories nor m de y improve

s s a a an

ments upon Segn er s turbine I 182 7 Fou neyron de igned severa l

’
. n r s

ra di a l flo w turbines w hich ga ve such decidedly good resul ts t ha t

the interest of scientists wa a wa kened in the theory of these wa ter s

me te rs Poncelet i
. 1838 presented a pa p r before the French
n e

Ac a demy }? in which he di cuss e d the shock lo e a t the ends of the

s ss s

impe ller bl a des H e however limited h is discussion by cons idering

.

the str am of flowing wa t r divided into a number of pa ra llel

e e

strea m a d di cu
s n ed only the occurrences in one of these streams
s ss .

H e s ugges ted cert a in co fli c ie ts of wa ter contra ction to a ccount for

e n

the diff erence betwee n h i theory a n d the res ults obta ined in pra ctice
s .

Se gn e r : Th eo ri a m a ch i n a e c u ju sd a m h d ra u y
li c s s e t c o mpu ta tio o rmae f
a tq ue vi ru m ma c h i n ae h y dra u li c a e n u pe r d esc ri ptae
”
G6 t tin gae , 1 7 50
. .

1 Eu
'
le r : Re c h e rc h es su r l e ffe t d u
’ ’
n e ma c hi n e h dra u li q u y
e pro po s e par é
M Se gn er a G6 ttin gue
.
”
His to ire d e l Ac a dé mi e Ro a le , Be rli n , 1 7 50 ; a lso
.
’
y
Prin ci pes g n ra u éé x d umo u ve men t d es flu
”
ides , 1 755 .

t Po n ce l e t : S ur l a t h o ri e d es e flée t s m c a n i q ue
'

d é
e l a t urb i n e F o urn e y ro n .

Co mptes re n d u s , to m e III, 1 838 .

224 T YPES OF C E NT R IFUGAL PU MPS .

In 1843 C ombes completed this theory by e ta bli hing coefficients o f s s

resistan ce imil a r to those e ta bli shed by Wei ba ch C omb w

S s s . es as

a l s o the fi rs t to refer to E uler work a ga in The theory of wa ter ’

s .

turbines was developed e ctly as was the e c a lled pra ctica l theory xa s -

of hydra ulics by limiting the inve tiga tion of wa ter flo w to t w

,
s o

di mens ions in a pl a ne whereas a ctua lly t he l a w for the flo w of ,

s

w a ter s hould h a ve been e xp ress ed in equ a tions cont a ining thre e

dimen io s The o e fli ie ts found were unre l but they corre ted

ns . c c n a ,
c

the error of neglecting the use of three dimens ion It i u ely s . s s r

worth noting th a t thi Poncelet C ombes theory which was dev loped s -
, e

into a pra ctica l method by Re d t ba h fi contributed very little en c er

towa rds t h e dva ncement of turbine de ign a s .

N w types of turbines were e w introduced by enginee rs wh o

e n

designed them to co form to their pr ctica l ide a d from their n a as n

experi menta l re ults without a y of them pa ying much a tt ntion to

s n e

theoretic a l consider tion or e ta bli hed l a w 1 W sh ll merely a s s s s . e a

mention the a i a l flow turbine of Hen chel ( 183 7 ) d Jonva l the

x s s an ,

a i a l ra di l turbines of Fra nci

x -
a the impul se turbines of s

Zu pp gi n Sc h
er wa mm k ug a d G ir a rd as well r n

as the centrifuga l pump of App ld ( 1848) a d Gwyne s o n

Theorist tried to e ta bli h the l w of fl w for these e w m a chin s

s s s a s o n e

a s fa as the Poncelet method pe rmitted a n d ucceeded in es t a bli s h

r , s

ing some gene a l rule of design r s .

Expe rimen t on the Sh a pe of the bl a de were m a de in 187 6 b y

s s

R C P rsons a d publi s hed i the Proceedings of the I s titute o f

. . a n n n

C ivil En ginee s Vol 4 7 Unwin in 18 77 derived formul ae from

r ,
. .
, ,

experiment l re ults of Pa r on s I 1885 P ill fl published a the o

a s s . n , o on

re ti c a l di s cus s ion of ne a rly ll the types of pumps kno wn a t th a t a

time .

In the me a ntime the st a m turbine bega n i t conques t (a bo ut e s

Zeuner wh h d begun i 1 856 to inv tiga te the Fra nci

,
o a n es s

turbine sta rted to collect a ll a va il a ble re ults of experiments an d

,
s ,

C mb o R h h th é i q u
es : t pé i m t l ec u l
e rc u a é ti
es or es e ex r e n a es s r e s ro es r ac on

oua t uy u P i 1 84 3
a x .
”
ar s, .

T R d t b
e h T h
en i u d B ud
ac T u b i
e r: u d V t il t M
eo r e n a er r n en n en a o re n . an

h e im, 1844 .

I Fo r d esc ri
a ptfull
i o n o f th
er e o ri gi n o f w a te r t urb i n e s an d pumps , se e R u
m
ma n n , Allge me in e M as c hi n e n l e h re , Vo ls I a n d IV, 1 8 7 5
”
. .

1] Po i llo n : Tr a i t t h o ri q ue e t pra c t i q u é é
e d es po mpes e t ma c hi n es a é le ve r

les ea ux .
"
2 26 TY PES OF CE N TR IEUGAL PUM PS .

s urpris ing t ge of development It rema ined for Sulz er B ros to

s a . .

demonstra te by mea ns f their high press ure pumps ins ta lled at

O -

the silver m ine in H orca jo Spa in a d a t the Geneva wa terworks

s , ,
n

in Swit erl a nd th a t it was poss ible to obta in goo d effi ciency with
z ,

centrifuga l pumps delivering la rge qua tities Of wa ter under high n

hea ds The excellent re s ults obta ined a t these installa tions led
.

to the genera l introduction Of centrifuga l pumps in mine s an d

l a rge wa terworks To da y centrifuga l pumps s uccess ful ly o m
.
-
, c

pete with piston pumps a d the centrifuga l pump ca n not be dis n

rega rded in co s idering new insta ll a tions of an y wa te r ca pa city

n

wh a tever .

It is interesting to note th a t a ll the l a rger pump ma ufa tu r n c

e rs are either building or contempl a te building centri fuga l p umps ,

chiefly b ca use they rea li ze tha t centrifuga l pumps a re s upe rsedin g

e

the pis ton type of pump in all servic The reciproca ting pump es .

s itua tion tod a y resemble th a t of stea m engines some yea rs ago

s

when stea m turbines beca me a ctive competito rs of the s team

engine .

The chief a dva nta ges of centrifugal over rec iproc a ting pum ps are
lower ini tia l co t less sp ce requi rement grea ter a da pta bility to
s ,
a ,

a ll poss ible conditions less a ttention while opera ting an d fi na lly

, , ,

the grea t a dva nt ge Of being ca pa ble Of direct connectio to a

a n

prime mover .

Ce ntrifuga l pumps a re n o w found in every servi ce im a gin ble a .

Recently they h a ve bee n m a ki ng gre a t trides in mine work o wing s ,

to the sma ll spa ce required their gre t reli a bility an d the light ,
a ,

founda tion permi sible The centrifuga l pump opera tes without
s .

a n y vibra tion wh a tever w hen direct co n nected to a motor a n d ,

therefore hea vy founda tion a not necessa ry a is the c ase wi th s re ,

s

reciproca ting pump This is no me a n a dva nta ge beca use hea vy

s .
,

founda tions a re very diffi c ult to cons truct in mines .

As the centrifuga l pump delivers wa ter continuo u s ly a n d w ithout

s hock the ch a nce of bre a ki ng pipes i

, reduced to a minimum By s .

c a reful design of the impeller bl a des it is po ss ible to close a va lve at

the end of the di sch a rge line without h a ving the pre ure rise thus ss ,

obvi a ting the da nger of brea king the pipe or of Shutting down the
pump .

Th a t all these a dva nta ges h a ve been fully recogni ed in pra ctice z

is Shown by the peedy introduction Of centri fuga l pu

s mps .
 M ODERN C E NTR IF UGAL PUM PS
. 227

The reciproca ting pump when direct connected to a motor still

reta ins the grea t a dva nta ge Of con uming from 5 to 10 per cent l ess
s

Fi g 100 . .

Fig . 1 02 . Fi g 1 03
. .

power B ut if t h reciproc ating pump is driven by a be lt or by gea r

. e

ing it tot l e ffi ciency will not be grea ter tha n tha t Of a direct con
, s a -

ne c te d ce ntrifuga l pump .
2 28 T YPES OF C ENTRIFUGAL PUM PS .

No t only is the ini tia l cos t of a centrifuga l pump less th a n th a t O f

a recip ro c a ting pum p but the motor for the former will a ls o be co n
,

si de ra bly che a per th a n for t h e l a tter Di rect connected cent rifugal .

-

pumps requi re h igh s pee d moto rs where as rec iproc a ting pum ps
-
,

req u i re s lo w s peed moto

- rs .

In order th a t centrifuga l pum ps m y fin d a still l a rger field it i a ,

s

of first impo rt a nce to Obta in the highest effi ciency a n d gre a tes t
reli a bil ity poss ible TOO much s tress c a nnot be l a id on t h e f ct
. a

tha t centrifuga l pumps re quire very c a re ful design in the dra fting
room a n d the best workm a ns hip in the shop a nd onl y then is it ,

Fig . 1 04 .

possible to Obt in a ti sfa ctory effi ciencies This is es peci a lly t ru

a s . e

of high press ure centrifuga l pumps Above a ll we Should a void

- .

using old pa tterns a t the a cri fi ce of e ffi ciency ; e a ch c as e s hould be

s

ca refully tudied a d ev ry nd a vor ma de to Obta in t h highe t

s n e e e e s

e ffi i ncy It m y truthfully b a id th a t the striking success d uring

c e . a e s

recent y rs of the high pr ure c ntrifuga l pump is chiefly d ue to

ea -
e ss e

the fa ct th a t pump ma nuf cture rs h ve i n i ted upon de igns o f the

a a s s s

highe t e ffi ciency a d the ve ry b st workma nship irrespective Of the

s n e ,

increas ed cost of s uch design s .

There a re va ri ou type s of high pre ss ure ce nt ifuga l pumps F ive

s -
r .

of the b st known types a show s h e m tic lly in Figs 1 00 to 104

e - re n c a a . .

Fig 100 repre s nt s a hi gh pre ur c ntrifuga l pump fi r t des i gned

.
e -
ss e e s

by Sul er a d known as the Sul er ty pe Fig 101 repre e nts the

z n z . . s
23 0 TY PES OF C E NTRIF UGAL P U MPS .

c anbe ma de exa ctly equa l to a reas F an d F the a xi a l thrus t c a n 4 3,

be elimin a ted when (p p ) (p p ) The di ff erence of thes e , l ; a .

press ure is nothing more th n the exi ting clea ra nce spa ce pre ure
s a s ss s

expressed by equa tion 23 provided tha t the rot ting pa r boloid i ,

a a s

not considered This equa tion i . s

H OP " Ha

By dime ioning ea ch se t of guide va ne so tha t they a e a ll imila r

ns s r s ,

2
w
it i po ible to keep the velo i ty he d 2 equa l fo e very t of guide
s ss c a
.
,
r se
0
va nes If t his is the c as e the e ce pres ure in the cle ra nce pa c e
.
,
x ss s a s

i the d pendent only upon the va lue f the e xi sting pr

s n e ure he d O e ss a

I de ig i g the pu mp the press u

n s n n he d of ll impellers re a s a

are ta ke n t h sa me but through light error in con truction the

e , S s s

ca lc ul ted v lue of t h pre ure he d will not be exa ctly rea li ed

a a s e ss a s z

nor will they be e ctly equa l The re ulti g xi l t h ru t c used by

xa . s n a a s a

thi inequa lity Of pre ure hea d i t ken c re of by a rigid colla r

s ss s s a a

thr ut be a ring
s .

The a rra ngem nt of the i mpellers in pa ir nece ita te c rryi g th

e s ss s a n e

wa ter from the exit of one impell r to t h t r n e of the nex t e e en a c

impeller i a novel m a nner Thi rran g m t i h wn in Fig 105

n . s a e en s S o .
,

which repre ent a i ta ge Sul r pum p in r

s s s x- s t io I t hi ze c o ss s e c n . n s

pu mp the wa ter e ter through t h uction pipe into the fi r t im

n s e s s

pell r d then passes through it into the guide ch a n ne l Fro m the

e an .

l a tter the wa ter i led through rever ing ch nel wh ich ma y b

s a s an , e

co id red continu tion f the guide ch a nnel to the entra nce of

ns e a a O ,

t h ne t i mpeller
e x Thi rever ing ch a nnel p . betwe n the ucs s as s e s e s

cee d i g guide ch nel

n Thi i ccompli h d by cas ting th guide
an s . s s a s e e

a pp tu
a ra th t a lt rn te p as ge a provided for the guide d
s SO a e a sa s re an

the r ver ing ch nnel Fr m th

e s o d imp ller t h w ter p
a s . o e se c n e e a as se s

thro ugh t h s ond t of guide v e ec into a S sh a ped ch a n el

se an es n -
n ,

w hich l a d it to thee tr nce f th e th ird impeller

s en a O .

I n p i t
s e f t h complic O t d w t r p e g the w a ter guid an ce is
a e a e a ss a es

very ffi cient as there

e no very ha rp turn u ed
,
a re s s s .

Anoth r a dva nt g e f rra ngi g the i mpell r in pa i rs is th a t

a e o a n e s

clea ra nce pa ce lo occur only upon one ide of the impeller i m

s sse s s r .

This type h as a di sa dva nta ge in not being a ble to use a Odd um n n

be Of impelle rs bec a use the impelle rs m u

r ,
st a l wa ys be a rra nged i n
 Z
SUL E R PUM PS
. 2 31
23 2 Y
T PES OF C E NT R IFUGAL PU M PS .

p i in order to ba l a n e a xi a l thru t d therefore t hi ty pe loses

a rs c s ,
an s

in a d pta bility for d iff e t pre ue hea d Another di dva n tage

a er n ss r s . sa

i the need
s f m a king t wo pa tte
O for the impe lle one ight rn s rs , r

ha nded a n d the other left h a nd d -

e .

This a l so dema nds c rrying dou ble a

the numb r Of i mpelle rs in s tock f e or

p re or repa ir pa rts s a .

The p um p ho wn in Fig 105 h s . as

three pa ir f imp lle The wa ter s O e rs .

a f ter p as ing throu gh the e s ix im s s

p ll a d guid v e i led to cir e e rs n e an s s a

ula ing to wh ich i co nec t d c r cas ,

s n e a

conic lly divergi g deli ve y pi pe T h a n r . e

i i m pe lle wi t h guide ppa r tu as x rs a a s re

a embl d i c as ing t in one piec ss e n a eas e .

The pu mp ha ft h t w b ing s as o ear s .

O one end the h ft i c ried i a b ri g mou t d u po n the el bo w

n s a s ar n ea n n e

O f the uction pipe a d pro vi d ed wit h ri g lub ic a tio

s n O th n r n . n e

other end t h be ri g i f t d to the hea d of the c ing a d t h i

e a n s as e n e as n s

h i g l o
ear n ri t h ll t h ut h i g T h t uffi g b
a s car es e co ar t r s ear n . e s n ox a

t h hi gh pr
e ure end i
-
e ss rr g d ( Fig 106 ) th t the w t r
s SO a an e se e . a a e

F ig . 107 .

l ki
ea ng thro ugh it at a en t e rs a h o llo w s pa c e a rou nd the h a ft li i S n ng

and flo w s o ut O f t h e d ra i n pi pe b
Th i g m e t h s t w d v . s a rra n e n a o a an

t a ge s : fi rs t , t h e st u ffi g b i gr tly u l d d b
n ox u it n d onl y
s ea n oa e eca se ee

h e ld a gi t
a ns a v ry m ll u b l
e s d pr ur whi h is the di ff er
a n a an ce e ss e, c
2 34 T YPES OF C E NT R IF UGAL PUM PS .

The pump cas ing is us ua lly ma de of c ast iron but for high pre s ,

sure s it is m a de of c as t s tee l The impe ller a n d guide va nes a

. re

ma de of speci a l co mpo ition bron e a d the Sh a ft is nickel s t ee l

s z n .

Fo r speci a l c as es cert a in p a rt h a ve a cid proof linings The c as i g

s -
. n

Fig . 1 10 .

is ma de in one piece a n d i p rma nently connected to the founds

s e

tion In di as mbling thi pump the high pre ure hea d i fi r t r

. s se s -
ss s s e

moved Thi hea d conta i n s the bea ring a d i on the free end Of
. s n s

Fi g . 111 .

the pump Th impelle a n d di a p hragm conta i n ing the reve rsing

. e rs s

ch a nn el c a n then be di as mbled suc e ively from the sh a ft The

s s se c ss .

a dv a nt a ge of thi s co truction is th a t the suction a n d disch a rge pi pe

ns s

need not be di conne t d when the pump is ope ned for inspection
s c e .

Fi gs 109 1 1 2 Sho w other de ign of Sul er pumps

.
—
The p um p s s z .

shown in Fig 109 h a . u tio pipe elbow the suc tion s pa ce

s no s c n ,

b i g a concentric ring such as is Often used when the suctio n c o n

e n ,
 Z
SUL E R PUM PS .

n ec tio n mus t be ma de a t o Side Fig 1 10 shows a norma l type of ne . .

Sul zer pump with electric drive a n d Fig 1 1 1 one wit h belt drive .
.

The Sul er high pressure centrifuga l pump

z -

for a mine sh a ft is moun ted vertic lly a n d a

directly con nected to a moto r The unit .

is assembled upon a uit a ble fra mework s .

The genera l a rra ngement of such a pum p is

s hown in Fig 1 12 The fra mework i ma de
. . s

of structura l ste el an d wooden shields a e ,

r

pla ced below a d a bove the unit to protect then

pump a n d motor from fa lling bo dies or from

bodies thrown up by explosions The shea ve .

for the h a ft rope used in ra ising or lowering

s

the pump i fas tened to the upper crosspiece

s

of the fra mework The Sides of the fra me .

work are provided with slippers for proper

guida nce The motor is genera lly of the e n
.

clo ed type a d is conn ected to the pump by a

s n

flexible coupling A l a dder pla ced a long the .

s ide m a ke the pump a n d the motor more

s

a cce s ible
s On Of the firs t l a rge ins t ll a tio n
. e a s

O f centrifuga l pum ps was for the s ilver mines

of the C mpa ma M inera y M et lurgi a del

o a

H orc a jo Spa in ,
At fi rs t only three pump .
*
s

were insta lled with equa l hea ds of a bout 430

feet a n d ope ra ted one b ve the other in dif a o

fe re n t ga lleries of the mine La ter a fourth .

pump Of equa l hea d was i n sta lled in a still

lower ga llery Fig 1 13 shows the a rra ngement
. .

of the pumps in the mine The lowe t pump . s

dra ws the wa ter from a pit a n d ra ises it to

the next pump The other pumps receive the .

wa ter under pressure a n d therefore work with

out suction he a d SO tha t no a ir c an fin d a way ,

into the pump AS a ll the pumps of the

. a re

s me Si e th y t a ke c re
a z f qu a l qu a ntities of
e a O e
F 8 1 12
w ter w h n running a t t h a me s pe d
1 .

a e e s e .

Th i i t ll t i
s h b
ns a d u d i f ull by D F H w g i t h
a on as ee n lsc s se n r
. . ee r a en n e

Z i t h i ft d V i d ut h I g i u 190 1
e sc r es e re n es e sc er n en e ne , .
236 T YPES OF C E NTR IF UGAL PUM PS .

While in this case the pumps a re pl a ced a t va rious heights it is ,

bes t in mos t cas es to ins ta ll a ll the pumps in the lo wes t ga llery .

This a rrange ment is Shown in Fig 1 14 .

,

which represent an ins ta ll a tion by s

Su l er B ros for the Vikt r M ine in

z . o

Ra u xel Westph a li a Germa ny Tw

, ,
. o

four sta ge centrifuga l pumps we re con

-

n ct d in series At 1040 r p m
e e . . . .

the e pumps deliver 2 50 cu ft of

s . .

wa ter per min a ga ins t a tot a l he d . a

of 1 720 feet a d require 56 2 h p to n . .

drive ea ch pump La ter two Six sta g .

* -
e

pump were a dd d of more mod r

s e e n

cons truction e ch c a pa ble of delive ,

a r

ing 230 cu ft of wa ter per min a d . . . n

ea ch connected to a 1 100 h p mot . . or

running t a bout 1500 r p m Th a . . . e se

tw unit pumping 460 cu ft f

o s, . . o

wa te r per min occupy pra ctic a lly .

,
no

more p ce th a n the older uni ts s up

s a

plying only 2 50 cu ft of w a te r p . . er

min The four ta ge pumps h a ve im

.
-
s

p ll rs 2 4
e inche e in di a met e r wh il s ,
e

the Six st ge pumps h a ve im pelle

-
a rs

2 0 inche in di a meter O e S ix t ag
s . n -
s e

pump d liver the me hea d as th e s sa e

t w four s ta ge pump in serie s

o -
s .

The Boiler Insp ction Society of E e s

n Germ a ny in 1 9 03 m a de a very c r
se , , a e

ful t t of the t w four s ta ge pum p r

es o -
s .
’

Ifi g 1 1 3
. .
T h follo w ing r ults were bt a ined
e es O

0m m
C
d U d Un erTes t
o n d I t io ns . ing
n er
.

Pe r ce n t . Pe r c e n t .

E ffi c ie n c y o f t h e po we r pl a n t , i n c l ud in g t h e l o s s of

t ra n s m i s s io n t o m o t o rs
y
E ffi c ie n c o f pump in s t a l l a t io n m o t o r a n d pump
y
E ffi c ie n c o f t h e pump
E fli c ie n c y o f t o t a l in s t a l la t 1o n

H e rzo g, E le k t risc h e Ba h n e n un d Be t rie be , 1 90 5 .

f Z ei ts c h ri f
t d es Ve re i n es d e uts c h e r In ge n i e ure ,
1 904 .
238 T YPES OF C E NT RIFUGAL PUM PS .

inch in the disch a rge pipe The suction line is provided with a .

s a fety a n d a foot va lve a n d the di s ch a rge piping cont a ins a throttl e

and a n return va lve Thi prevents the entire wa ter col umn
on -
. s

fro m exerti g a press ure on the pumps Should they be stopped

n .

Sho u ld o non retur va lve be in ta lled then both pump cas ing mus t
n -
n s s

be des igned to ta nd press ure f 7 6 5 pounds pe r squa re inch

s a O .

Fig . 1 15 .

The pu m p dire t conn t d to the moto rs by me a ns of a

s a re c ec e

l ther b d co upling The t rting a d topping f th e pumps

ea an . s a n s O es

a re t ll d from th pow r t tion

co n ro e th a t only a emergency
e e s a ,
so n

swit h d t wo
c mm t r n d be in t lled a t the pumps Th e
an a e e s ee s a .

sta ting of th pump i as follo w

r e se s s s:

The m a in thr ttl v lve in the di h a rge piping rem a ins closed
o e a sc .

The pum p a fill d w i th w te r by m

s re of a u ili a ry connection
e a eans an x

to the di ch rg pipe d a ll a i in the pump c ing i a llowed to

s a e an r as s s

e pe T h g r tor a t th po wer hou e i then ta rted a d its

sc a . e ene a e s s s n

e i t tio lowly incre d A oon the pumps Show a pressure

xc a n s as e . s s as
 Z
SUL E R PUM PS .
2 39

of a bout 407 feet 17 6 pounds per squ a re inch ) grea te r th a n the wa te r

hea d i the d is ch a rge pi pe the throttle va lve i Slo wly opened Then
n , s .

thi excess of pre ure Slo wly decreas es as the qua ntity of discharge
s ss

increase When the throttle va lve i closed the loa d on the genera
s . s

tor i n the po w er t a tio i a bout 30 to 3 5 per cent of the no m a l lo a d

s n s r

under opera ting conditions As the va lve slowly opens the lo a d on

.

the genera tor incre as es The entire time necessa ry to t a rt this

. s

pl a nt from the ti me current is turned on until wa ter is di ch a rged s

from the press ure or delivery pipe i a id to be only one minute If s s .

the time interva l between sta rt ing the pumps a n d opening the
throttle va lve is to o grea t the wa ter in the pump c asings becomes

Fi g . 1 16 .

hea ted a d it is necessa ry to shut down the pumps for so me minutes

n

a n d re fi ll them w ith wa te r The hutting down of the pump

. s s

ta kes pl a ce in reverse order The throttle va lve is fi rst slowly .

closed a n d then the genera tor i unload ed by cutting do wn the s

excita tion The a ttenda nts control these ope ra tions by mea ns of
.

a telephone .

Another very inte resting ins t a ll a tion of Sul er B ros is the pump z .

ing sta tion of the Kas imir M ine in N iemce owned by the Wa rs ch a uers ,

Gesell ch a ft fii Kohlenbergba u und H ii tte be trie b Fig 1 16 Shows

s r n . .

this ins ta ll a tion Thre e centrifuga l pumps a e used Ea ch pum p

. r .

delive rs cubic feet of wa ter per minute a g in t 538 feet hea d a s

at 9 7 5 r p m
. .The pumps a s a rra nged tha t they c an ope ra te a t
. re o
2 40 T Y PES OF C ENT RIFUGAL PUM PS .

times a ga ins t high pressure a n d normal qua ntity or a t times aga ins t ,

lo w press ure a n d a l a rge qu a ntity of disch a rge When opera ti ng .

a ga ins t high pre s ure the p u mps a re connec te d in seri e so tha t the
s s

firs t pump dis ch a rges into the second a n d the second i nto the third .

B y this a rra ngement the three pumps deliver cubic fe et Of

wa te r per minute a ga ins t a hea d of 3 times 538 feet or a ga ins t a tota l ,

hea d of 16 14 feet which lifts the wa ter out Of the mine If the
,
.

pumps a e to deliver a l a rge qua ntity the va lves are so e t tha t th e

r ,
s

pumps disch a rge into a common discha rge line a d lift the wa ter n

out of a common well The three pumps then deliver 3 times .

cubic feet of wa te r per minute or a tota l qua ntity of ,

cubic feet of wa ter per minute a ga inst a hea d of 538 feet which lifts ,

the wa ter to the next higher ga llery of the mine Two di cha rge . s

pipes a re u ed one lea ding to the top of the mine an d the other to th e
s ,

next higher ga llery .

C ntrifuga l pumps are of es peci a l v a lue as h a ft pum ps in mine s

e S .

The limited spa ce given to mine h a fts necess ita tes a pump occupying s

minimum sp a ce Electric a l power is a lso a grea t a dva ntage in this

.

cl as s of work .

Sul er B ro s inst a lled the fi rst l a rge Sh a ft pump ins t a ll a tion in a

z .

mine h a ft in Sile i a Germa ny This ins ta ll a tion is Of grea te r

S s ,
.

intere t th a n ordina ry bec a use reciproc a ting pum ps were not ca pa ble
s

of ta king the w ter out of the submerged Sha ft The se rvice which
a .

the pumps h a d to ta ke c a re of was briefly as follows .

A sh a ft 1 3 12 feet deep a t the Do e sma rc k M ine was to b e nn r

dra ined At a depth Of 328 feet prings were encountered which

. s

floo ded the entire Sh a ft After the wa ter in t h h a ft o ve rflo we d it . e s

was found th a t there was a cons t a nt flow of a bout 530 cubic feet per
minute A a ttempt was ma de to dra in this h a ft with reciproca ting
. n S

pumps but only three pump could be insta lled in the sma ll cross
,
s

section of the Sh a ft a d the e could only deliver 3 50 cubic feet O fn s

wa ter per minute not con idering the time lo t on a ccount of th e

,
s s

g re a t number of rep a irs needed The pump s could not dr a in the .

s h a ft a n d fi n a lly the mine w as a b a ndoned

,
In order to sa ve thi . s

va lua ble property Sul er pumps were ordered as a e periment

,
z n x .

Three sh a ft pumps were built e a ch c a p a ble of deliv ri g 283 cubic , e n

feet Of wa ter per minute a ga inst 52 5 f e t hea d wh n ru n ning a t e e

970 r p m . The three pumps in p a ra llel could therefore deliver

. .

84 9 cubic feet per minute The spa ce occupied by the e pumps in . s

24 2 T Y PES OF C E NT R IF UGAL PUM PS .

Of wa te r per minute a ga ins t a hea d of 46 0 feet cons umin g a bout

,

1000 h p . .

ul er ros h a ve inst a lled a very l a rge pumping sta tion for th e

S z B .

city of M il n Ita ly At first three pumping sta tions were insta ll e d

a ,
.

in which direct connected induction motors furni s he d the po wer for

-

the pumps All three sta tions h ad power supplied from a common
.

power pl a nt La ter a fourth pumping sta tion was in sta ll ed in whic h

.

the pumps were belte d to gas engines .

Fig . 1 18 .

Th e ource Of wa ter supply in M il a n consisted Of wells sunk a t

s

va rious pl a ces a bout forty five in a ll From these wells the wa te r

,
-
.

is piped into l a rge receive rs in which v a cuum is held by a i pum ps

,
a r .

From these v a cuum receivers the pumps dra w the w a ter The .

l a rgest pumping ta tion is in Loreto d is h ow in Fig 1 19 The

s an S n . .

pumps dra w the wa ter from a commo s uction pipe out of the va cuum
n
 Z
SUL ER PUM PS .
243

re ceiver SW The pipes S bring th e wa te r to the receiver All

. .

four pumps deliver into a common dis ch a rge pipe D The four .

pumps deliver 509 cubic feet Of wa ter aga ins t 187 feet head wh ile
ru n ning at 82 0 r p m .Fig 1 2 0 give a photo gra phic view of the
. . . s

interior of the pumping sta tion a t Loreto .

Although Sul er B ros built a l a rge umber of pumps Of the type

z . n

s ho wn in Fi g 100 they h a ve n o w pr a ctic a lly a b a ndoned the b a ck to

.
,
-

ba ck cons truction a n d h a ve arran ged the impe llers as in the Ra tea u

Fig . 1 19 .

an d J a eger types A pump of their n e w type was exhibite d a t the

.

Ba va ri a n Ind us tri a l a n d Tr a de Expo s ition in N uremberg duri ng

19 06 . The reas on for this ch a nge evidently lies in the disadva ntages
mentioned previously The disa dva n ta ges of h a ving to a rra n ge the
.

impellers in pa irs which m a de such a n a rra ngement less a d a pta ble

,

for diff erent hea ds a d a lso necessita ted the m a king of righ t a d left
n n

h a nde d impellers was a grea t h a ndic a p The complic a te d ti n g

,
. c as

necessa ry for the wa ter guid a nce was a nother grea t di sa dva ntage .

Besides a ll thi s experience proved th a t a fter a cert a in ti me of oper

,

a tion axi a l thrus t developed which in high spee d pum p s cou ld not
,
-
 244 T Y PES OF C E NT RIFUGAL PUM PS .

be prope rly c ared for The reas o n for thi increas ed a xi a l thrus t
. s

was the we a r of the p a cking be t ween the two impellers Of a group

s .

All thes e di a dva nt a ge were s uffi cient to c a us e Sul er B ros to

s s z .

a b a ndon their old type in pite of the excellent results they h ad

s

obta ined with it .

In Sul s l a tes t con truction an a xi a l thru t mut occur a d thi

’
ze r s s s ,
n s

thru t is ta ken ca re f by a ba l n cing pi ton Thi pi ton Fig 12 1

s O a s . s s ,
.
,

i pl a ced b hind t h l as t imp ll r on t h high pr ss ure

s e e d of t h e
e e e -
e en

p ump The di a m t r of t h pi t n i o ch n t h t it c ba l an e
. e e e s o s s o se a an c

F ig . 12 0 .

the i l thrust developed by the pump The pi ton i s ho wn in

ax a . s s

Fig 12 1 as p a rt P a d i k e y d to the s h a ft
. n s It i held fi rmly on
e . s

one ide by the impeller L a d on the other ide by a nut N Th e

S n s .

oper tion of thi ba l a cing pi ton is as follows

a s n s .

In the cle a ran ce spa ce a high pre s ure e xi s ts which v a ries with
a s ,

the fina l pressure of the pump The wa ter in spa ce c a pass . a n

through the a n nul a r clea ra nce d betwee n the piston a d the lining an d n

flo w to sp a ce b where it i gre a tly reduced in pr e ss ure bec a us e of i t

,
s s

being throttled by the clea ran ce d From b the wa ter flows through .

th e cle a ra nce s p a ce 0 into a p a ce connected wi th a dra ins Should .

the pump pressure incre as e for a y reas on the press ure in spa ce n , a
2 46 TYPES OF C E NT R IFUGAL PUM PS .

to releas e some Of the pres sure in spac e b thus a ga in ba l an cing axi a l ,

thrust .

This ba l a ncing a rra ngement is upposed to work a utoma tica lly s

an d doe s not eces ita te the man ipula tion Of a y va lves Th e

n s n .

movement of the pi to n a d h ft need be but very light Pl a cing

s n s a s .

the b l a ncing pi ton a t the high pre ure end of the pump rende rs
a s -
ss

unnece a ry a tuffi ng bo x pa cking a ga ins t high press ure

ss s .

42 .
‘
RATE AU PU M PS OF SAU I I E R ''
HAR LE Co .

.ho ws chema tic lly the a rra ngement of a Ra te a u type

F ig 1 0 1 S s a

high pressure centrifuga l pump This typ of pump w as fi r t buil t

-
. e s

by Sa utter Ha rle C O of Pa ris The wa ter guid a nce in thi type

. . s

Fig . 122 .

is in a wa ve like dire tion Aft r l a ving th impe ller the wa ter is

-
c . e e e

ca rried in a ra di a l dir tion t h rough t h guid v

ec a d th e n t urn d e e an es n e

in the rever i g h nn l throu gh ppro im t ly

s n c a e s The exit a x a e

e n d of the rever i g h a nn l cont a in dire c ti g v a ne s

s n c e ( S e Fi g s n . e .
 RAT E AU PUM PS .
24 7

Th ese directing va nes a re so formed th a t one end Of the va ne lies

pa ra llel to the direction Of the wa ter in the revers ing cha nnel pe ,
r

mitting the wa ter to enter without shock a n d the other end of the ,

va ne a t entra nce to the impeller is given the proper direction desire d

for the a bsolute entra nce velocity To neutra li e the end thrust . z

the t wo ide f the impeller a hrouded to different di a meters

S s O re S .

Thi i cle rly shown in Fig 1 0 1 The impe ller bl a des extend to
s s a . .

dia meter D the left ha nd ide f the impeller i hrouded to t hi

-
S O s s s

dia mete r while the right h a nd ide i hrouded only to dia meter D
,
-
s s s ..

A lo wer pres ure e is ts a t di a meter D th a n a t di a meter D bec a us e

s x , ,

Fig . 12 3 .

2 _ 2

the va lue 2 is s ma ller a t the former th a n a t the l a tte r di a me

g
te r
. This a rra ngement therefore gives a higher press ure in the
clea ra nce spa ce a th a n in the clea ran ce pa ce b s .

It is quite di fficult to m a ke a correct choice in the dimens ions of

the impeller dia meters D a d D O t ha t the tota l press ures on both
o n , S

s ides Of the impeller w ill be b l a nced This a rra ngement permits

a .

of only a p a rt i l b l a nce C omplete ba l a ce i fina lly obta ined by

a a . n s

me ns of a ba l a ncing pi ton keyed to the h a ft beyond the l as t

a s S

i mpeller on the high press ure Side -

Thi pi ton rota te in y li . s s s a c n

dri c a l c as in g with very sm a ll cle a r c e s ( S e F i g O one

an . e . n

Side Of the pi ston full wa ter press ure e xis ts while the other Side c n ,
a
248 T Y PES OF C E NT RIF UGAL PUM PS .

be connec ted to a n y press ure stage or to the suction end of the pump .

The axi al thrus t is neutra li ed by the d iff erence of the press ures o n
z

the oppos ite sides of the pis to n Sa utter Harle Co found tha t
. .

the surfa ces of the piston were worn by the grit ca rried through with
the w a ter They therefore provided a specia l lubrica tin g a rran ge
.

ment to prevent the ra pid wea r of the piston .

By us ing a p a rt of the impe ller bl a de unshrouded on one Side i t

s ,

can be seen th a t the flo w of wa ter through the impeller is in terrupte d

a n d th a t eddy a d imp a ct lo ss es mus t t a ke pl a ce
n .
2 50 T Y PES OF C E NT RIFUG AL PUM PS .

There is a lso insta lled t the Hua nch a c a M ines a pump ca pa bl e

a

of delivering 150 cubic feet Of wa ter per minute under hea ds of 100 ,

2 00 3 00 a n d 400 feet
, , This pump is illus tra ted in Fig 12 5 The
. . .

division of this pump into four pre ure hea ds is very usefu l in dra i ss n

ing the mine a n d where it is necessa ry to lift the wa ter from the
lo wer ga llery ei t her entirely out of the mine or to an y one of t h e
inte rmedia te ga lleries .

Fig 1 2 6 illus tra tes a turbine d riven pump ins t a lled in a mine i n
.
-

Euro pe a n T u rkey The turbine i a lo w pre ure unit receiving

. s -
ss

s te a m a t a press ure O f a bout 10 pounds per qua re in ch The se t s .

Fi g . 12 6 .

runs a t 4500 r p m a d deliv r 6 5 cubic feet of wa ter pe minute

. . . n e s r

a ga i t a h d of 2 2 3 fe t A t t of thi t wa m a de with
ns ea e . es s se s

p und per qu r inch t m pr ure t t h e turbine throttle an d a

o s s a e s ea e ss a

va cuum of i h of merc ury ; the number of revolutions was

nc es

4460 per minut The tot l h d d livered by the pump was 22 8

e . a ea e

feet a d the qu ntity of di h rge

n a cubic feet of wa ter pe minute
sc a r .

The to ta l effi ciency wa a bout 3 1 p t orresponding to a con

s e r cen ,
c

sumption of 36 pound of tea m p hor e po wer hour Of wa te r s s er s -

pumped .

Fi g 12 7 ho w
. turbin drive pu mp for the Le ns M ines The
S s a e -
n .

unit c a n deli ver cubic fee t of wa ter per minute a ga i s t a hea d of n

 RAT EAU PUM PS . 2 51

39 4 feet the st e a m press ure a t the turbine being 1 70 pounds per

,

squa re inch ga uge Sa utter Ha rle C h a ve been quite successful

. o .

in getting very good effi ciencie even with very sm a ll qua ntities of s

di scha rge They built a pump delivering o ly

. ubic feet of n c

wa ter per minu te a ga in t a hea d of 13 12 feet The pump ran a t

s .

2 9 00 revolutions per minute .

Fi g . 12 7 .

repr ent
F ig 12 8
. in t lla tion of both hi gh d l w pr ue
es s an s a a an o -
e ss r

turbine direct connected to a Ra tea u centrifug l pump There i a a . s

sm a ll pump u sed for the ejector conden er dir ct co ect d t thi s e nn e o s

unit Fig 129 gives a pl a n a d elev a tion view of thi i t all a tio
. . n s ns n ,

in which A is the multi st ge pump B the high pr s ure t a m t ur

-
a ,
-
e s s e

bi n e C the low pressure tea m turbine D the pump for the cond er
,
-
s , en s ,

a d E the ejec tor conden er

n Fig 130 give a pl a n a d elev a tion
s . . s n

V iew f the reciproc a ting pumping unit wh ich wa r pl ced by the

O s e a

turbo pump just described The l tter occupie j ut one tenth th . a s s -

e

spa ce Of the former There i s l so a decided di ffere c e in the fo u d

. a n n a

tion requirements i fa vor of the turbo unit When the expen

n . se

of opera tion d m int na nce of the two unit a co id er d th

an a e s re ns e e

c ontra s t is ve y m a rked r Tests of the turbo unit give the followi g

. n

re ults
s
25 2 T YPES OF C E NT RIFUGAL PUM PS .

Qu ti tyan of wa t er pumpe d pe r min ue t . c u ft. .

1 194 ft .

2 200
10 1 1bs per sq . . in .

Exha u st pressu re (a bso ue ) lt . lbs pe r sq in

. . .

Co n su mpt io n o f ste a m pe r h o rse po we d e i ve red r l

( n o t i n c lud i n g c o n d e nsa t io n )
To ta l e ffi c ie n cy Of th e un i t . .

This fi m a lso builds vertical pump for mine Sh a fts Fig 1 3 1

r s . .

illustra t a n electric a lly driven vertic l h a ft mine pump for t h e

es a s

H u a nch a c a C omp a ny f B olivi a It deliv e rs 1 5 1 cubic feet Of w a te r

O .

per mi nute a ga inst a hea d Of 12 1 feet The revolutions per m inute .

F ig 1 2 8
. .

a re 7 37 the e ffi cie cy of the c mpl te unit is

and n pe r cent Th e o e .

motor is three ph e 500 volt a ltern a tor a d it co sumed

a -
as kilo
- n n

w tts when delivering the qu a ntity of wa ter a d he a d bove t a te d

a n a s .

Ra te a u in conjunction with Picc a rd l so de igned a multi sta ge a s -

ce trifuga l pu mp drive by a hydra ulic turbine Picca rd i know

n n . s n

through his engineering work in conn ction with the hydra ulic e

turbine a t Ni a ga ra F lls The hydr ulic turbine is mounte d on a

s a . a

hori ont l sh a ft c rrying the impellers of the multi st a ge centrifuga l

z a a -

pump The complete unit i illustr t d in Fig 132 The uni t

. s a e . .

entirely di assembled i shown in Fig 133 The di a meter of the

s s . .

wheel of the wa ter turbine is 12 inches a d the turbine develops n

more th a n 120 h p The pu mp was built to work a ga ins t a hea d o f

. .

5 18 feet a t 2 300 r p m The tests ma de in Geneva by Ra te a u an d

. . .

Picc a rd ga ve ext remely fa vora ble res ult At the e tes ts the highes t s . s
254 Y PES
T OF C E NTRIFUGAL PUM PS
.
 R AT E AU PUM PS .

hea d obta ina ble was 42 7 feet which was the height of the highes t ,

res ervoir in the city The res ults of the tests were as follows :
.

Q ua nt i ty f w t r d l i
o d t th e h y d uli tu bi
a e e ve re o ra c r ne

p
er mi u t n e

H d f w t r f h yd u
ea o a e li tu bi
or ra c r ne

Qu tity f w t d liv r d by pump pe mi ute

an o a er e e e r n .

H d d liv
ea e d by pump
e re .

To ta l effic ie cy n

Fi g . 13 1 .

The e fi c ie n cy of the centrifugal pump was 6 8 per cent an d th a t

of the turbine 7 1 per cent which is very sa tisfa ctory for a ma chi ne
,

of s uch s ma ll dime ion Thi s pump was con tructed for th e

ns s . s

D a rien Gold M ining C O a d pl a ce d a t the bottom of their mine

.
,
n .

The hydra ulic turbine wa driven by w a te r ca rried to it under a

s

very high hea d which furnished enough po wer to lift thi s sa me wa ter
,

to gether with wa ter in the min e up to the surfa ce of the mine .

25 6 T Y PES OF C ENT RIFUGAL PUM PS .

This kin d of hy dra ulic turbo pump is very use ful for lifting wa te r
from a sma ll wa terfa ll to hotels or residences sit ua ted a t h igher levels,

Fig 132
. .

or even for supplying w ter to cities in mounta inous regions In

a .

thes e c as es the he a d a va ila ble for the hydra ulic turbine is less th a n

Fig 133
. .

the he a d delivered by the pump but the qu an tity delive red to the
,

turbine is much gre a ter th a n th a t delivered by the pump .

25 8 T Y PES OF C E NT RIFUGAL PUM PS .

s pa ces 1 a n d 2 This would crea te an a xi a l thrust Ja ege r h as

. .

found this to occur a d h as pa tented an a rran gement which will n

a utoma tic a lly prevent a n a xi a l thru t when there i s unequ a l w ea r s

Of the p a cking rings Th is arran gemen t .

is Shown in Fig 1 34 . .

The sh a ft w ith the impellers h as a ‘

s m a ll a xi a l pl a y the periphery Of th e ,

im peller is cylindrica l while the c as in g ,

s urrounding the pe riphery of the impell e r

is m a de conic a l as c an be seen in t h e ,

figure If the impellers or Sh a ft receiv e .

for a y reas on a x i a l thru st s a y fro m n an ,

right to left the impeller will move to ,

the left The clea ra nce spa ce on th e .

left will be enl a rged a d its press u e in n r

creas e d a bove th a t exi ting in the right s

Fig 13 4
h a nd cle ra nce sp a ce a d the impeller i
.
a n s

forced ba ck tow rds the right J eg r

.

a . a e

a l so us es a th ru t be a ring on the su ction side of t h e pump to t ak e

s

up a n y a xi a l thrust which ma y e i t Thi s th rut bea ring must b x s . s e

of uch de ign a to t ke up a xi a l thru t but till permit Of ax i a l

s s s a s s

end pl a y The gener l fe tures of thi h e a i g a hown chem a ti

. a a s r n re S s

c a lly in Fig 135 The h a ft a t A h as a num ber Of coll ars s imil a r

. . s

to a m a rine type thrust bea r

ing Thes e coll a rot te within
. rs a

a Split cylindric l bu hing B a s

m a de Of b bbitt a d e ured a n s c

a ga in t rot a tion by the f ther

s ea

F . Thi s bu hing is fre to s e

a dju t it lf a xi a lly within

s se
Fig 135
c rt in limit between the
. .

e a s

c as ing C a d the pump c as in g Ordin a rily if n o e n d th rust exi ts

n . s

there i a clea ra nce a t ea ch e d of the bushing an d thrus t

s 2 n ,

c an be t a ken up only when the bu hi g re ch e the Shoulder D s n a s .

In t h i wa y the thru t b a ring doe

s t i terfere with the a u
s t oma tic
e s no n

a rra ngement Sho wn in Fi g 1 3 4 unle s such a l a rge u n b a l an ced thru s t. s

occur as to m a ke it neces a ry to h a ve the a id Of the th rus t

s s

bea ri gs n .

It i quite a a dva nt a ge to b l a nce e a ch i mpeller sep a ra te ly in a

s n a
 JAE GE R PUM PS .
25 9

high press ure centrifuga l pump for uch a n a rran gement permits
-
, s

the us e of impellers h a ving Slightly different press ure hea ds an d

,

Fig 136
. .

Fi g 13 7
. .

still allows the a xi a l thrust to be ba l a nced Fo r furt her deta ils

. Of

the cons truction of this type of pump se e Figs 136 an d 137

.
26 0 TY PES OF C E NT R IFUGAL PUM PS .

The wa ter ente rs through the suction pipe into the fi st im pel l er r ,

then pas s es through the guide va nes lyin g in the sa me pl a ne as t h e

impeller Both the impeller an d guide van es a e ma de Of bron z e
. r .

The guide v a ne a e upported by the c as i g an d a Open on t h e

s r s n re

s uction Side w hich permi ts m a chining the guide v a nes an d th u

, s

procuring exa ct dimensions of the ch an nel The guide va n es l e a d .

the wa te r ra dia lly a t exit The wa ter then passes through t h

. e

Fig . 1 38 .

S-
h ped ch nel which c a rries it to the next impeller To gui d e
S a an , .

the wa ter correctly thi rever ing ch a nnel h as directing van es or

,
s s ,

rib In thi s m a nner the wa ter go s from sta ge to ta ge until it

s . e s

fi na lly enter round c asing h a ving a conic l con ection to the

s a a n

press ure pipe .

The c asing surrounding the guide va nes consists of a number o f

like p a rts corresponding to the number of sta ges The sep ara te .

pa rts or sections a re combined into a complete unit by mean s o f

, ,

through bo l ts This subdivision of the c as ing Offers severa l impo r

.
26 2 TYPES OF C E N TR IFUGAL PU M PS .

The h aft is supporte d on the suction a d pressure ends in bea r

s n

in gs wi th rin g lubrica tion The bea ring a t the s uction end us ually
.

conta ins the thrus t coll ars .

C H Jae ger
. .Co b u ild three distinct types of centrifuga l
.

pumps the low pre ssure pump the modera te press ure pump an d
— -
,
-
,

the high press ure or turbine pump The lo w pressure cons tructi on
-
.
-

Fig 140
. .

h as bee n a dopte d recently a n d ta kes c a re of head s up to 80 feet

,
.

The pump cas ing is cylindric a l a n d the pump h as no disch a rge va nes .

Fig 1 38 ho w the g nera l a rr a ngem e nt of

. S s e l w press ure pump a o -
.

The e tra nce to the impeller is on both s ide w hich gives a per
n s,

fe c tly ba l a nced im peller provided of course th a t the outer impe ller

, , ,

clea ra nce is m de su fficiently l a rge Fig 139 illustr tes a lo w

a . . a

pressure centrifuga l pump of thi s type direct co nected to a motor n .

 JAEGE R PUM PS .
26 3

Th ese pumps h a ve a so mewh a t un fa vora ble wa te r guid nce a t a

entra nce to the impeller hence they a e built only in sma ll Si zes
,
r .

Fo r l a rger low press ure pumps the a rra ngement s ho wn in Fi g 140

-
.

Fig 141
. .

is u ed W te r a l o enters on both ide of the impeller but it is

s . a s S s ,

led through more gra dua l bend Thi rra ngement nece sit a tes
s . s a s ,

F i g 142
. .

of cou rse two suction pipe The pump shown in

,
s . 140 d elive rs

1 58 5 to 1 7 6 5 cubic feet of w ter per minute a ga i n s t

a he d of 43 fe t a a e

when ru ning t 36 5 r p m
n a . . .
26 4 TYPES OF C ENTRIF UGAL PUMPS
.
26 6 TYPES OF CEN TRIF UGAL PUMPS .

146 .

Fi g 142 . hows a model of a th ree

S

s t ag e pump in w hich the p a ckin g

rings a n be clea rly see n

c d , an

Fig 143 gives a view of t h

. e

sa me pump with the ction se s

s ep a r a ted I this ty pe Of pu m p
. n

it is impo rta nt th a t the pump

s hould never be run dry be c a u se ,

the pa cking would burn out At .

the highes t point of ea ch pump

s ection a sm a ll v a lve i nece ry s ss a

to permit a i to es c a pe when the r

pump i firs t filled with w ter

s a ,

an d dra in va lve hould b

a s e

provid d th a t the pump

e so can

be dr ined for di as sembly or to

a s

prevent free ing of the wa ter z .

Fig 144 ho ws a three—st a g e

. S

pump d l ivering 283 cubic f te ee

of wa ter pe r minute aga inst 4 1 0

fe t hea d The pump i con
e . s

n e te d to a
c induction motor by n

me an s f a le ther b a nd coupli g
O a n .

In thet welve st a gepump Sho w n -

in Fig 145 the drivi g motor i

. n s

pl a d in the middle with s ix

ce ,

mg
. 1 47 .
st a ges on e a ch s ide Thes e pum ps .
 JAEGER PUM PS .
26 7

a renot a rra nged oppositely to ba l n ce axi a l thrus t bec a use there

a ,

are flexible couplings between the pu mps a d the motor A pump n .

with twelve sta ges wo uld require a Sha ft of cons idera ble length a n d ,

if the impellers a e not c refully b a l a ced the high numbe r of e vo lu

r a n r

tions will ca u vibr tion of th h ft C h i fly on ccount f t h

se a e S a . e a o e

l a tter reas on the pump ho uld be divi d d

s e .

Fig 146 ho w a intere ting a rr g ment of

. s s n s i
an t g d
e a s x- s a e an a

two stage pump which c an be co nected by mea ns Of a clutch to

-
,
n
26 8 TYPES OF C E NT RIFUGAL PUM PS .

a common sh a ft which revolves a t 1500 r p m Fig 147 s hows a . . . .

J eger pump for Sh a f t or mine work

a .

On e of the l a rges t ins ta ll a tio n s of C H Ja eger CO i s a t t h e . . .

mine KO ig Lud wig in Recklingha us n Germa ny The pum p

n
”
e , .

cons i ts of eight sta ges four sta ge being pla ced on ea ch side Of the
s , s

electric mo tor Fig 148 giv s a cross s ction through one of t he

. . e e

four stage pumps of this insta ll a tion The pump delive rs 24 7

-
.

cu b ic feet of wa ter per minute a g ins t a he a d of 1903 fee t whe a n

in c a f it
/ o r mi n .

Fig 1 49 .

ing 268 cubic feet Of wa ter per minute a ga in t a hea d of 1843 s

fee t the pump run ing a t 14 75 r p m The power cons um ption

,
n . . .

a t thi l a d was 1 2 3 0 h p
s e With the throttle on the di sch a rge pipe
. .

entirely clo ed hence no outpu t the pr ure ro e to 2 028 fee t

s , , e ss s
,

t king a bout 42 5 h p to drive the pump

a . . .

An intere ti n g serie of curves ta ken fro m te t re ults on a n

s s s s

eleven sta ge pump for the W th a u en M ine in Germa ny is

-
“
es s

Shown in Fig 1 50 The norm a l r ti g of thi pump i 106 cubic

. . a n s s

feet of wa ter per minute when worki n g a g in t a he a d of 1375 feet a s .

The motor receives its current from a line supplying l a rge n um a

ber Of unit a d therefore its speed ca nnot be kept cons ta nt T h e

s, n .

highest e fficiency is 75 pe r cent an d was obta ined when the pump ,

 KUGE L —
G ELPKE PUM PS .
27 1

44 . K U G L GELPKE
E -
PUM PS OF E SC H E R , WYSS CO .

The Kugel Ge lpke high pressure centrifuga l pum p first built by

- -
,

Es cher Wy , C O of Zurich Swit erl a nd di ff ers cons idera bly from

ss .
,
z ,

the types lrea dy des cribed for in this type bo th the impe ller an d
a ,

guide va nes a e S Sha ped (se e Fig

r The wa ter enters the .

impeller ch nnels in aa xi a l direction a n d a fter p ass ing through a

an ,

double bend of 9 0 it lea ve the impeller ch a nnel in an axi a l di e c

°
s r

tion The w t er flo w in a simil a r ma nner through the guide va nes

. a s .

A correct des i gn of va nes in the impeller an d guide a ppa ra tus i s

s uppos ed to give the wa ter proper gu id a nce .

Fi g . 1 52 .

This form Of wa ter guid a nce is perh a ps very efficient when the
pump is working under ra ted condi tions beca use the entra nce a d , n

exit velocity di a gra ms a e correct for these conditions But if the

r .

dis ch a rge i throttled thus decreas ing the qu ntity of wa ter delivered
s , a ,

the wa ter will no longer h a ve the proper velocity to form the correct
velocity dia gra ms a n d therefore the effi ciency will be grea tly
,

dimini shed by hock losses In Article 28 a ttention was ca lled

S .

to the necess ity of h a ving l a rge cle a ra ces bet ween the impe ller n

a d di s ch rge v a nes
n a The wa ter b ing forced through a double
. e

bend in the impeller a d se t in rota tion ma y become quite trouble

n

some for it is turned through 90 jus t when the grea test velociti es
°
,

occur .

This type h as an adva nta ge over others in its simplicity of con

s truct ion The guide va nes an d reversing bl a des re ma de in one
. a

piece The casing c an be ma de quite sm a ll in di a mete r

. .
27 2 TYPES OF C E NT R IFUGAL PUM PS .

Fig 1 52 . hows a longitudinal c ross section Of a se ven sta ge hi gh

s -

pressure Kugel Ge lpke pump The guide va nes h a ving the S Sha ped
-
.
-

cha nn els a r as sembled in units SO tha t they a but a ga ins t one a nother
e

in a cylindrica l cas ing These guide va nes can be disasse mbled by

.

s lipping them Off the sh a ft a t i ts free end I the l as t s ta ge t h e . n

wa ter lea ves the guide va nes In a ra di a l direction a fter which it ,

passes ra di a lly into the Spira l discha rge cas ing .

nI the des ign of Es cher Wy ss C O wa ter under press ure i

, . s

a dmitted into the s p a ce between the guide v nes a d impelle in

s a n rs

order to ba la nce the axi a l thrust The wa ter pres s ure is regula te d .

by mea ns of a va lve .

The axi a l thrust c a n be ta ken ca re of only by some pecia l ba l a nci g s n

a rran gement just as in the Ra te a u pump

, I the be ring there is . n a

a l so a e rie
s of thrust coll a r to t ke ca re of a y a xi a l thrust not
s s a n

provided for in the ba l a ncing rr ngement Thi thrust bea ring i a a . s s

a dju t a ble an d permits a

s a djus tment of the pos ition of the i m
n

p e lle rs in a n a xi a l direction .

45 . WO RTH IN GTON PU M PS .

Henry R . Worthington of
ork n w p a rt of the Inter New Y , o a

n a tion l Ste m Pump C o mpa ny h as been building high pres s ur

a a ,
-
e

centrifug l pump since 190 1 a d h a sold more centrifuga l pumps

a s ,
n s

Fi g 153. .

th a n other ma nufa cturer in the United Sta tes In 1 904 this

an y .

firm secured the rights to build the J a eger type Of pump in a ll

countries except Germa ny a n d Austro Hun gary This type of pump -
.
27 4 Y
T PES OF C E NTRIFUGAL PUM PS .

h asbeen us ed extens ively a n d gives good results but quite rece n t l y ,

Wo thington developed a n e w type Simil a r in principle to the dou b l e

r

suction type Shown in Fi g 1 04. This new type is the na tura l o ut

.

Fig . 1 55 .

come of the de ire to build high press ure pumps Of grea ter spee d
s -
s,

thus increas ing the hea d pe r stage decreas ing the di a meter of t h e
,

impellers an d reducing the number of sta ges The effi ciency of this
, .

m
fa i ly th i c k /" ra m
.

Fig 156
. .

n ew type is grea ter th a n th a t Of the Old for reas ons eas ily unders tood
from the theoretic a l di cuss ion given in thi s book The axi al ba l a n ce
s .

of this type is superior to th a t of the Jaeger .

2 76 T YPES OF C E NT R IF UGAL PUM PS .
27 8 TY PES OF C E NT R IFUGAL PUM PS .
 WORTHINGTON PUM PS
.
27 9

force . oth bea rings a re lubric a ted by Oil rings The impellers a re
B .

c onst ructed of bronz e an d a re of the enclos ed or Shrouded type ,

Fig 1 6 1
. .

h ing ba ckwa rd turned bl a des They a re c as t in one piece

av -
. an d

ma chined d polis hed on the outside The wa ter ch a nnel

an . s a re

F ig 1 6 2.
.

chipped an d fil ed to a smooth fin i sh As in a ll Jaeger design s th e

.
,

fla nges on both s ides Of the impe ller form a close run ning fit with the
-
2 80 T YPES OF C E NT RIFUGAL PUM PS .

clea ra nce rings ecured to the pump c as ing The disch a rge va n e s
s .

a e m a de O f bron
r with one s ide remov ble ( Fig 1 55 ) so th a t the
ze , a .

wa ter pas a ge s be c a refully m chined a d poli hed in order to

s can a n s

reduce the w ter friction to a mi imum The pump h ft i ma de

a n . s a s

Fi g . 163 .

Of s t ee l,
w i t h br o n e prot cti g
z leev The
e s h
n a ft iss a clo s eesSlidin
.
g
fit in the i m p ller ehub d i provid d w ith
s, a n a houlder
s a g a ines t s ,

w hic h the impelle a d l vrs a n ured by

s ee m ee n s of a
re nut
se c o n a

the uction d Th t uffi g bo

s en . e s a p a rt of the be a in g
n xe s a re r

bra cket a d a b tw
s, n re t h e c as ing
e d h e a ri gs to prevent sa n d
ee n an n ,
28 2 TY PES OF C E NT RIF UGAL PUM PS .

an d g rit from entering the be a rings W a ter s e a l s are provided .

for the stuffing boxes a n d they a re connected by mea ns of suit

,

a ble piping to the di s ch a rge s ide O f the pump Fig 156 gi ves t he . .

cha ra cteristic curves Of thi pump as dete rmined by tes t Whe

s . n

delivering 150 cubic feet ( 1 125 ga llon ) f wa ter pe minute a ga i nst s O r

a he a d of 6 00 fee t the pump ho wed a n e ffi ciency of 7 5 pe r cent

s .

F ig . 165 .

Fig .giv a photogra phic vi

157 es a fi ve s t a ge Wo thi n gto n e w Of -
r

pump of the Ja ger type u ed for boiler feed It delive 6 7 cubic

e s . rs

feet f wa ter (300 g llo ) per mi ute t a pres u Of 2 50 pound

O a ns n a s re s

per qu re i ch The di me t r of t h di h a rg pi pe i 4 inches

s a n . a e e sc e s .

The pump here Sh own i dir t o t d to 100 h p T rry ste a m

s ec c n n ec e a -
. . e

turbine running a t 2 100 revolutio per minu te Thi type of ns . s

pump for feeding b iler give o u interrupted a d tea dy d i

s s an n n s s

cha rge th us elimina ting a ll shock on the pipe system N0 a ir

,
.
2 84 TYPES OF C E NT RIFUGAL PUM PS .

ch a mbers or relief va lves a re necessa ry as the pressure will neve r ,

build up to a d a ngerous point even though the disch a rge va lve b

, e

entirely cl osed .

Fi g . 16 7 .

When l a rge number of t a ge i required Worthington divides

a s s s

them into t w pump a rr nged on both ide of the motor so th a t

o s a s s ,

the unba l a nced axi a l thrust of one pump is Opposed to tha t Of t h e

Fig 168
. .

other Fig 158 how uch a a rra ng ment The unit consi ts o f
. . s s s n e . s

ten ta ge c a p ble of d liveri g 2 7 9 c ubi feet of w a t er pe minute

s s a e n c r

( 3 000 000 g a llon p 2 4 hou ) gs i ter he d of 7 5 0 fe

rs e t The a a ns a a .

driver i a 6 00 h p W tinghou inducti motor Four of thes e

s -
. . es se on .
286 T Y PE OF C E NT RIFUGAL PUM PS .

uni ts were supplied to the contrac to rs for the Jac kso n Stre et R e
gra de Sea ttle Was hingto n where nearly two
, , , million cubic y a rds

Fi g 1 70
. .

Ofea rth were removed with these pumps by a method kno wn as

hydra ulic mini ng Th is method is less expens ive an d quicker th n

. a

Fig 1 7 1
. .

the stea m Shovel whenever it a n be employed Th e pumps opera te

c .

in series when a h a rd str tum is encountered a n d in pa ra llel for

a

removing the softer m a teri a l The sa me method is employe d

.
 WORTHINGTON PUM PS . 2 91

Fig 1 6 2 . ow two st a ge pump for lifting wa ter out of a hot

sh s a -

well The di a meter of the di sch a rge pipe is 5 in ches The pump is
. .

ca pa ble of delivering 7 0 cubic feet (52 0 ga llons) Of wa ter per mi nute

aga inst a hea d Of 80 fee t This pump is direct connected to a 25 h p
. - . .

Kerr te a m turbine run ing a t 12 00 r p m

s n . . .

This type Of pump is a lso built in vert ica l un its Fi g 163 is a . .

photogra phic vi ew of a four sta ge pump h a ving vertica l sh a ft

-
a .

The pump is insta lled t the wa terworks of the city of Athens Ga

a ,
.

Thi s unit delivers 24 0 cubic fee t of wa te r per minute ( 1800 gallons or ,

2 000 000 ga llons in 24 hours) a ga i nst a pressure of 1 50 po u

n ds Th . e

di a meter of the di scha rge pipe is 8 inches .

Fi g 176
. .

The Worthington pumps sodes cribed are high press ure pum p
far -
s

of the J a eger type This firm h as a lso built very good designs
.

of low pressure pumps of both the hori zonta l an d vertica l types

-
.

Fig 164 gives a ph otog a phic vi ew of a two un it low pressure se t

. r - -
.

Ea ch u n it h as a n a vera ge c a p a city of 66 85 cubic feet (50 000 gal

lons ) of wa ter pe r mi nute an d is used for emptying d y docks The r .

pumps work a ga inst a hea d varying from 0 to 35 feet The di ameter .

of the disch a rge pipe of ea ch unit is 45 inches an d the suction pipe ,

is c a rri e d on both ides of the impeller fo rming w h a t i s c a lled the

S ,

double s uction type These pumps are connec te d to Genera l Electri c

- .

motors Fig 16 5 is a ph o to gra ph ic view of a vertic l low press ure

. . a -

pump delivering 2 925 cubic feet (2 1 8 7 5 ga llons) Of wa ter per minute

29 2 T YPES OF CE NT RIFUGAL PUM PS
.
2 94 T YPE OF C E NTRIFUGAL PUM PS .
 WORTHIN GTON PUM PS . 29 5

precedi ng pump in not h a ving a y dis ch ge va nes a lthough the hea d n ar ,

is qui te hi gh Th e c as ing is so cons tructed as to form a s pec i a l

.

involute dis ch arge to a id in decreas ing the Shoc k losses a t exi t from
the impeller On a ccount of the dirty a n d gritty wa ter to be
.

pumped di ch a rge va nes were not u ed in this c as e Fig 17 0 pre

,
s s . .

sents the ch a ra cteri tic curves of thi pump At the normal loa d s s .

the effi ciency is 7 5 per cent which is excellent con ideri g th a t the ,
s n

pump h as no dis ch a rge va nes .

F ig . 180 .

Fig . illu tra te the co truction Of v rtic l bi roto r pump

17 2 s s ns a e a -

h a ving no di ch rge v nes The weight of the rot ting element is

s a a . a

ca r i d on a wa t r te p Thi pump c a n deliver 936 cubic fe et

r e e s . s

( 7 000 g llo ) of
a w a t r per
ns minu t e a g a in t a he a d of 1 80
e feet a t a s

s peed f 1500 r p m
O . . .

It m y b f i tere t to Sho w the rel a tion between he a d c a p a city

a e O n s , ,

po wer a d effi ci cy for va riou p ed of a l w press ure pump with

,
n en s s e s o -

o ut di ch rge v e
s a Fig 1 7 1 give a serie
an f curves ho wing thi
s . . s s O S s

r l tio hip t a ke from a ctu l t t on Wo thi gt on pump h ving

e a ns , n a es s a r n a

a 5 i ch di ch rge pip
-
n s The i ffi i y curv
a of pe i a l
e .
—
so e c enc es a re s c
2 96 T Y PES OF C E NT RIFUGAL PUM PS .

interest for they Show a t a gl a nce the region of an y pe r ce nt o f

,

effi ciency .

As has been sta ted Worthington h as recently a dopted the doubl e

,

s uction impeller for multi t a ge high press ure high s pee d a n d high
-
s - -
, ,

e fi c ie n c y de ign A recent ins t a ll a tion of this type was f r th e

s .
o

C ommonwe a lth Edi son C omp a ny of C hic a go Severa l units h a ve .

been insta lled for boiler f d d others a e being built A photo ee ,

an r .

gra phic view o f one of the e uni ts is given in Fig 1 7 3 The pump s . .

h as three st a ges a d is direct connected to a 150 h p C u rtis stea m

n -
. .

turbine built by the Gen r l Electric C ompa ny The pump c an e a .

deliver 6 0 cubi feet (450 ga llons ) Of wa ter per minute a ga ins t a

c

Fi g 18 1 . .

pressure of 240 pounds per qu a re inch when opera ting a t a spee d s

Of 2 400 r p m Fig 1 7 4 gi v
. . outline dra wing showing the genera l
. . es a n

dimen ions f the unit a d Fig 17 5 give the cross s ectiona l views
s O ,
n . s -

of the pump The wa ter is ad mit ted through the fiv i c h suction

.
~
e n

O p ning to b th s ide
e of the fi r t impe ller The discha rge va nes a re
o s s .

SO co tructed th a t the w t r c pas s through cored holes in the se

ns a e an

v nes a n d SO g in entra nc to the opposite uction side The impe l

a a e s .

le are of the hrouded typ e turned smooth a n d polished on the

rs S ,

out ide a d c a refully chipped d file fi i h d in the bl a de cha nn els

s n an -
n s e .

The di ch a rge v nes a re c t ope n on one side to permit Of a ccura te

s a as

ma chining of the guide ch n els A pl te is fas tened to this open a n . a

s ide as s ho wn in the s ection a l view The wa ter passes from th e .

29 8 T YPES OF CE NT RIFUGAL PUM PS .

a ga ins t hea d of 500 feet The stea m turbine ru s norma lly a t

a . n

1 600 but the governor c an be a djus ted over quite a l a rge

ra nge of speeds The turbine h as two bea ri gs a n d is connec te d
. n ,

to the pump Sh a ft by means of a flexible coupling des igned by th e

Genera l Electric C ompa ny a n d shown in Fig 17 8 This coupling . .

cons ists of two disks A a d B m ad e Of a number of thin heet teel

n s -
s

p u n c h i g
ns the outer
, e n d s o f which are secured to the ring C The
s
.

inn er end Of dis k A is riveted to the turbine h a ft coupling D s ,

while th a t of B is riveted to the pump sh a ft coupling E Th i . s

a rr a ngement permits of l a ck of a lign ment betwee n t h e t wo Sh a fts ,

but is perfectly rigid in tra n mitting the torque The pumps re s . a

s imil a r a d h a ve a be a ring fas tened to the s uction end of the c s ing

n a .

A pe des t a l bea ring is provided between the flexible coupling an d

the pump The coupling between the two pumps is of rigid con
.

struction a d is not fle ible n The thr ust bea rin g i on the o ut

x . s

bo rd end o f the pumps Fig 1 7 9 gives an outline view of this un it

a . .

a n d Fig 180 a cross section through the outbo a rd pump

. Th . e

cons truction Of the impelle a d discha rge va nes i Similar t o th t

rs n s a

Sho wn i Fig 1 75
n A thes e pump h a ve not yet been inst a ll d
. . S s e ,

complete te ts a e not a v il a ble but a shop t t h as been ma de on

s r a ,
es

one pump with one impeller only This test will not show as goo d .

a res ult as is expected from the complete unit Fig 1 8 1 gives in . .

full lines the cha ra cteristic cu ve Obta ined from the Shop test an d r s ,

the dotted lines represent the expected ch a r cteristic curves of a a

s ingle pump with t wo imp ller On the whole thes e curves a r

e s .
, e

s a ti sfa ctory for they represent e ffi cienci es Obt ined on commerci a l

,
a

m a chines a d not on s peci a l pumps built for tes t a n d then ra ted

n

a t the point Of their best e ffi ciency .

It m a y be of interes t fin lly to a d d a photogra phic view ( Fig 18 2 )

,
a , .

of the Worthington pump in ta lled on the N w York fire boa ts s s e .

Thes e units were the fi rst multi ta ge centrifuga l pumps driven by —

s

s te a m turbines ins t a lled on fi bo a ts There a e two units on boa rd

re . r

ea ch bo a t a n d ea ch unit is c a p a ble Of delivering 7 35 cubic feet (5500

,

ga llons ) Of wa ter per minute a ga inst 12 5 pounds pe squa re inc h r

pressure or 6 00 cubic feet (4500 g llons ) of wa ter pe minute aga i t

,
a r ns

150 po unds pe r squ a re inch pres s ure The norm l opera ting spe ed . a

is 1800 r p m .In c as e of emergency the wa ter pre ssure c a n b e

. . s

incre s ed by increas ing the p ed Of t h e turbine The two units o n

a s e .
TY PE OF CENTRIFUGAL PUM PS .
3 02 T YPE OF C E N T RIFUGAL PU M PS .

The high press ure centrifuga l pumps h a ve a wa te r guid a n c e

-

Sim il a r to the Ra te a u a d Ja eger typ es but a ls o h a ve several nov e l

n ,

Fig . 185 .

fe a tures their o wn Fig 187 hows a

Of . . S fi ve - t ge high pr ss ure
s a -
e

pump direct connected to a mo tor a d Fig , n . 188 gives a longitudin a l

Fi g . 18 6 .

cross section of a five ta ge pump The pump is built up in c

-
s . se

tion of one sta ge e ch in ord r th t it m y be e ily di as sembl d

s a e a a as s e .

Another a dv a nt a ge of this method of construction is th a t it is poss i

 PUM PS OF T H E AR M ATU RE N U M ASC H IN E N FAB RI . K A G . .
3 03

ble to a dd a num ber of sta ges by imply s ubstituting a longer Sha ft s

an d longer through bolt on the c as ing s .

The impellers a re of ra ther novel des ign Fig 189 Shows a cross . .

s ection a l view throu gh the bl a des Of an impeller The ntra nce a d . e n

Fi g 1 8 7
. .

e it end f e ch bl d ar involute in form a d are conn ected by a

x s O a a e e n

c uve of m imum r diu I t h u u l cons truction by other pump

r ax a s . n e s a

builder t h e id of the imp ll

s S es pl an e urfa ces of revolution ;
e e rs a re s

th a t i if a cross section i t ak en through the line X Y of the im

s, s

Fig . 1 88 .

pe ller shown in Fig 189 it will ppear as i Fig 190 the impeller
'

.
,
a n .
,

Sides b ing pl ne urf c of revolution th a t a t y cert in r diu

e a s a es , so an a a s

the bl de h i ght would l w ys be con t nt The impe ll r ch nnel

a e a a s a . e a

ca be repres ented s in Fig 19 2 which is developme t Of the

n a .
,
a n
3 04 TYPES OF C E NT RIF UGAL PUM PS .

ch a n el along the mea n pa th of the

n wa ter d Shows va rious
an

s ections t a ken a t right a ngl es to flo w . All thes e sections tak en a t

a, b, c, d, an d e , 189 an d 19 2 , at right a ngles to flow a re tra pe

z o i ds .

Fi g 1 90
. .

Fig . 10 1 .

The impeller des igned by Neum a n chief engineer of the Arm a n ,

turen und M as chinenfa brik A G is so cons tructed th a t ll cross

. .
,
a

sections of the impeller ch a nn el t a ken a t right a ngles to flow are

3 06 T Y PE OF C E NT RIFUGAL PUM PS .

cha nn el from a to e for both des ign s of impeller Th e dotte d lin e .

represents the decrease of velocity for the impeller shown in Fig 19 0 ,

.
,

a n d the full line the decre as e of velocity for the impeller Shown i ,
n

fi g 19 1
. In the l a tter the decre as e of velocity i line ar a n d cor
. s

res ponds to the theory pres ented in Article 4 The dotte d curve .

Shows a very ra pid decre as e of velocity from point a to c From 0 .

to d the decre as e of velocity is less At d the lo west velocity is .

rea ched tha t of

,
feet per second Fro m d to e the velocity .

a ga in increas es to feet per second Such a decrease an d in .

creas e of velocity na tura lly disturbs the prope r tra nsforma tion Of
velocity hea d into pres s ure he d which influences the e fli c ie n cy o f
a ,

the pum p Shock losses a lso occur to some exte nt in

. im pe ll er ,

as shown in Fi g 1 9 0 . .

Fi g . 19 4 .

Another fea ture Of thi type of pump is the method Of

s

the short c ircuited wa ter into the uction pipe The us ua l c o n

-
s .

struction is shown in Fig 19 5 Fig 1 96 shows the des ign of clea r

. . .

a nce rin g for this type f pump In Fig 195 the short circuited
O . .
-

w a ter lea v es the clea ran ce s pa ce a t a high velocity a d meets t h e n

w a ter of lower velocity entering the impeller perpendicul a rly This .

c a u es strong eddy currents which continue up to the entra nce o f

s ,

the impe ller a n d hinder a proper entra nce Of the wa te r to the im

peller Th e short circuited wa te r flows from the clea ra nce rings a t
.
-

velocities from 60 to 100 feet per second a d over To obvi a te this n .

fa ult the clea ra nce rings are ma de as shown in Fig 196 Th e

,
. .

wate r lea vi ng the cle ara nce rings a t high velocity meets th e m a in
 PUM PS OF THE ARM ATURE N U M ASCHINE NF ABRI . K A G . . 3 07

body Of wa te r entering the impeller a lmost parallel to the direction

of flow th us producin g a sort of ejector ac tion This pe rmi ts a
,
.

pa rt Of the energy Of the short circuited wa ter to be rega ined

- .

Fig . 105 . Fi g 100

. .

In this type of pum p a xi a l thrust is ta ken care of by b al a ncin g the

press ures upo n both sides of the impeller as in the J a eger pump
,
.

Fig 1 9 7 gives a cross sectio n a l view of the impe ller a n d c asin g of a

.
-

Fig . 197 .

multis tage pump Cle ara nce spa ces 1 a d 2

. ma de of the sa me
n

i e by us ing cleara nce rings of the sa me di a mete r a n d cha mbe r 3 is

s z ,

c onnec ted with ch a mber 4 by hole s in the Side of the impeller This .
308 TY PES OF C E NT R IFUGAL PUM PS
.
.

will produce equ a l pressure in ch a mbe rs 3 a n d 4 pe rfectly e limin a t

,

ing the sta tic pres sures A s uita ble form a tion of the clea ra n c e
.

Fig . 198 .

at the rings w ill a l o produce equa l pre sures in pa ces 1 a n d 2 ;

s s s

Since the a reas are equa l the tota l press ures will ba l a ce ea c h n

other .

The outbo a rd bea ring of the pump conta ins a thrust he a rin g
to ta ke up a y sma ll thrust ca u ed by unequ a l we r of the
n s a

clea ra nce rings The thrust bea ring is provided with forc e d
.

lubric a tion
.
31 0 TY PES OF C E NT RIF UGAL PUM PS .

Fig . 198 s hows

l a rge s in gle s ta ge high pressure pump di re c t
a - -

connected to a 2 00 h p Siemens Sc h a c ke rt motor It delivers 4 60

-
. .
-
.

cubic feet of wa ter per minute a ga ins t a hea d of 165 feet a t 12 00 r p m . . .

Fig 1 9 9 is a two s t a ge high press ure pump delivering 9 0 cubi c fee t

.
- -

of wa te r per minute a ga ins t a hea d of 150 fee t ; the pump r uns a t

1 100 r p m . . .

Fig 2 00 represents a seven st a ge pump of this des ign ins ta ll e d

.
-

as a fire pump for the Underground Electric Ra ilwa ys C ompa ny of

Ga llo n s pe r Min u
te

Fig 20 1 . .

Lo ndon It is de igned to deliver 325 ga llons (44 cubic feet ) pe r

. s

minute a ga ins t a hea d of 140 pounds per squa re inch (322 feet ) a n d ,

ru n s a t 9 6 0 r p m The pump is direct connec ted to 50 h p in

. . . a — . .

duction motor of 33 % cycles which expl a ins its low speed Fig 20 1
, . .

Shows the curve heet for thi pump which h as a n e fli c ie n cy Of

S s ,

per cent a t the peci fied hea d The e fli c ie n cy rise s to 7 4 pe r

s .

cent when the he a d is increased .

 I. P. M ORRIS PUM PS . 311

47 . I P M ORRIS PU MP S
. . .

The I P M orris Compa ny of Phil delphi a Pa do not m a nu

. . a , .
,

fa cture an y ta nda rd line of centrifug l pumps but build both high

s a ,

a d lo w pre
n -
ure pump s peci a lly des i gned to mee t the need of a y
ss s s n

custo mer Their peci a lty is lo w pres ure pumps of l a rge si e On e

. s -
s z .

of the l a rge t unit in t lled by thi firm is a lo w press ure pump

s s s a s -

for the pumping ta tion of the N w Orlea ns Dra ina ge Commiss ion
s e .

Fig 202 . .

The impe ller of this pump i 9 feet 4 inches in dia meter a n d the diss

ch a rge pipe is 6 feet i dia meter It i driven by 500 h p Genera l

n . s a - . .

El ec tric moto r The sha ft of the unit is vert ica l The pump is
. .

pl a ced in a concrete cha mber below the motor the l a tter being ,

ins ta lled upon the fl o o r of the sta tion The pump delivers 19 2 00 .

cubic feet f wa t r per minute a ga in t a hea d of 10 feet with a

O e s n

effi ciency of 7 4 per cent The maximum effi ciency is . per cent
w hen the pump i delivering 1 7 6 4 0 cubic feet of wa ter per minute
s

a ga ins t a he a d f 1 1 feet O .

Fig 2 02 gives a photo gra phic vi ew Of a recent d i gn of high

. es

pre ssure centrifuga l pump built for the wa terworks of the city of
Du luth M inn It is a two stage hori onta l sh ft pump a rra nged for
,
.
-
z a
31 2 TYPES OF C E NTRI F UGAL PUM PS .

direct connection to a Genera l Electric induction motor The tw o .

impellers are pl a ced ba ck to ba ck so as to reduce the a xi a l thrus t .

The unba l nced a xi a l thrust is ta ken c a re of by a b a ll he ari n g a t

a

t a c h ed to the t a il be a ring The bl a de surfa ces of the impe ller a d

. n

guide va es were m a de mooth in order to minimi e the frictio

n s z n

losses The pump i equipped with two ring Oiled ba bbitte d be a r

. s -

ings secured to the suctio a d disch a rge elbows The tuffi g bo x

n n . s n

on the suction si de Of the p ump is suppli ed with a bronz e spa cer ri n g ,

Fig 2 03. .

into which wa ter i fed This wa t r sea l prevents the a i from bein g
s . e r

dr wn through the tuffing bo a d into t h pump through the

a s x n e

s uction elbo w The perform a nce of thi pump is hown in Fig 2 03

. s S . .

It will be seen th a t when the pump deliv r wa ter a t the ra te of e s

1 2 53 c ub ic feet per minute ( 13 500 000 g llon per 24 hou rs ) a ga in s t a s

a tot a l he a d of 32 7 feet a d when running

n t a s peed of 7 55 r p m a . . .

the pump effi ciency is 82 per cent The combined effi ciency of the
.

motor tra nsformers a d pump when the l tter is Opera ting under
,
n , a

the conditions a bove sta ted was found to be 7 2 per cent Th e entire
,
.
314 TYPE OF C E NT RIFU GAL PUM PS .

A photo gra phic view Of one of these ets is shown in Fig 2 04

s . .

The se t h as three bea rings one between turbine a d generator

, n ,

a nother between genera tor a d pump a d a third on the out bo ard

n ,
n

en d o f th e pu mp Th e turbine is overhung on the sh a ft On e

. .

s h a ft exte nds through turbine d genera tor an d is direct con nected

an

to the pump sha ft The entire unit is mounted on a common bas e

. .

The bea rings a e upplied with oil under press ure Ea ch pump h as
r s .

t wo impellers pl a c d b a ck to b ck as in the origina l Sul er type Of

e a z

pump The di charge va nes a d revers ing p a ges a e c ast into

. s n as s r

the c a ing The di a meter of ea ch impeller i 18 inche Between

s . s s .
 I
. P
. M ORRIS PUM PS .

the impe llers is a di a phra gm holdi ng a pa cking ring which c an ,

slightly a djus t i tself centra lly as the p ump comes up to speed an d ,

which i s held in po s ition by wa ter pressure a fter the pump rea ches
i ts no rm a l s pe ed The two centrifuga l pumps on ea ch he a t c an be
.

opera ted in eries or in pa ra llel When runn ing in pa ra llel they

s .

c a n deliver 12 03 cubic fee t of wa te r pe minute (9000 ga llons perr

minute ) a t the no z les a t a press ure of 150 pounds pe r squa re inch

z .

When run ning in series whi ch is equiva lent to a cting as a single

,

Fi g 2 06
. .

four sta ge pump they c a n disch a rge

-
, cubic feet of w a te r per
minute (4500 gallons per mi nute ) a t 300 pounds per squa re inch
press ure The wa ter for the pumps is dra wn from a 14 in h hea der
.
-
c

whi ch extend cross the engine room a d i delivered to bul b h a ped

s a ,
n s -
s

turrets loc a ted on the deck a mid hips Ab ve the deck turret is a
s . o s

bridge on which are mounted two 4 inch monitor noz le with me -

z s

c h a n ica l control The bo a ts a re moored with three mooring pud

. s s,

which con i t Of he a vy s teel tubing 1 8 inches in di a meter a d h e d

s s n s

with hea vy c as t steel points to ta ke hold of the river bottom By

-
.
 TYPE OF C ENTRIF UGAL PUM PS .

means of stea m o pe ra ted ra c ks an d pistons they are raised a n d

-

lowered th rough wells in the botto m of the boa t .

After the boa ts were pl a ced in commiss ion they were gi ven c a p a city
te sts coveri n g a period Of eight hours d u ring which time a t frequent ,

an d regul a r inte rva ls re a din gs were ta ken The res ults Obta in ed .

are indi c a ted in the followi ng ta bul a tion :

Fig 207
. .

Tu rbin e spe e d (sta rbo a rd ) . 1660 prm

. . .

Tu rbin e spee d 1634 r p m

. . .

Bo ile r pressu re (s ta rbo a rd ) . lbs pe r sq in

. . .

Bo ile r pressu re lbs f pe r s q in . .

C o n de n se r va c uum . in .

l
C e n trifuga pu mp press ue (mea n r ) . lbs pe r sq in
. . .

Nu l
mbe r o f n o zz es ” 2
Dia me te r o f n o z z esl .

i
D s c ha rge pe r pump .

i
D sc h a rge , two pumps
To ta l disc ha rge pe r h o ur .

l r
Co a bun e d , o attl .

r
Co al burn e d pe r h o u

r
C o al bun e d pe r sq ua re e
.

f ra te s rf
.

f etOg uace ”

Stea m co ns umptio n pe r h o rse po we h o ur -

r .
320 T YPES OF C E NTRIFUGAL PUM PS .

an d as the lining of the spherica l sea te d self a ligni ng type the

is -
,

entire thrus t be aring a djus ts i tself to the position Of the sha ft The .

t hrus t bea ring co ist of a seri es of thrus t coll a rs keyed a ltern a tely
ns s

t o the s h a ft a d b a ring c as ing with roller be rings bet wee n them

n e , a .

The thru t coll a rs on both the h a ft a d e asi g a pa ced a ccura t ly

s S n n re s e

a d held in pl ce by pa cing leeves c a refully ground to equ al len gth

n a s s s .

Fi g 2 1 1
. .

This b ring w ill ta ke thru t in ither direction The roller be rings

ea s e . a

a e lubric a ted by
r O il under pr ure upplied through r di l holes
e ss s a a

lea ding to a hole in the center of the h a ft into which the oil i forced
s ,
s

by the oil supply syste m Thi thrust bea ring i highly sa ti f ctory
. s s s a

an d h as not given a y trouble in service

n The turbines a e upplied. r s

with stea m at 180 pounds per qua re inch pre ure a d exh ust
s ss ,
n a

into a va cuum Of 28 inches The pumps a e two t ge w ith impeller

. r -
s a ,
s

a ra nged b a ck to b ac k a n d with the di s ch rge v a e

r a d revers ing a n s n
32 2 TY PE OF C E NTR IF UGAL PUM PS .

thrust either between the pump a n d motor in a se pa ra te pedestal ,

or one of the motor bea rings is constructed to conta in the thrust

Fig 2 13
. .

The press ure obta ined from e h sta ge of a Ra tea u pump is sma ll
ac .

Ordi na rily one sta ge gives a press ure of 65 feet of wa ter

. Na tura lly

the impellers a re quite sma ll in di a mete r On a cco un t Of using

.

such sm a ll hea ds per st a ge a high pre ss ure pump requires m a ny

-
,

i mpellers an d therefore a long sh a ft which Often necess it a tes mount

,
 K
S ODA PUM PS .
323

ing the impellers in groups ea ch group in a sepa ra te casing The ,

.

driving mo tor in such cas es is pl a ced between the two pump c as ings ,

an d the a xi a l thrus t is a lmos t wholly b a l a nced by a rra ngin g the two

groups o f impellers so tha t the thrus ts are in opposite directions .

Such an a rr a ngement is Sho wn in

Fig 2 13
. .

Fig . hows a Ra tea u pump

2 14 s

driven by a stea m turbine The .

unit was built by the Skoda Works .

The pump h as four impellers runs ,

at 32 50 revolutions per minute ,

an d delivers 106 cubic fee t of wa ter

per minute a ga ins t a hea d of

fee t On a ccount of the high speed
.

of the pump ea ch sta ge ga ve a hea d

of a bout 17 0 feet It i intere ting . s s

t note the sm a ll spa ce required for

o

this unit ; it is feet long

'

feet wide a d feet high The

, n .

conica lly diverging dis ch a rge pipe -

permits the us Of a sm a ller c a ing e s

from w hi ch the wa ter lea ve t a s a

ra ther high velocity Thi high ve . s

lo c ity is gr a du a lly reduced in the

conica l disc h a rge pipe Fig 2 15 . .

S hows a mine ha ft pump Of the s

Skod a Works des ign c a pa ble of de

livering 14 cubic feet Of wa ter per
Fi g 2 15
minute a ga inst feet hea d .
. '

Fi ve impellers ru n ning a t 2 900 r p m a re us ed The pump c as ing . . . .

is so de signed th a t a h a lf section c a n be removed wit hout d isas sembly

of a n y other part Thi s permits of a n eas y i ns pection
. .

The Skoda Works built a pump for experimenta l purposes a ccord ,

ing to the directions Of Ra tea u with a impeller of only 3} in ches , n

dia meter which was direct connecte d to a ste am turbine runn in g

,

a t 1 8 000 r p m The pump delivered

. . . cubi c feet of wa ter per
minute a ga ins t 7 7 5 feet head The sh a ft effi ciency Of the pump .

a lone w as reported to be 6 0 per cent With a somewha t reduced .

qua ntity Of wa te r this sma ll pump delivered hea ds up to 985 feet .

3 26 TYPE OF C ENTRIFUGAL PUM PS
.
 ALLIS C HALM E RS PUM PS
— .
3 27

Allis C h a lmers
-
ty pe These motors a re of 800 h p woun d for
. . .
,

three phase 2 5 c ycle 6300 6 6 00 volt ope ra tion an d des igned to

-
,
-
,
—
,

run a t a bout 7 4 0 r p m
. Fig 2 1 7 s hows a photogra phic view Of
. . .

one of these pumps a n d Fig 2 18 gives a longitudina l cross section

. .

Ea ch pump is des i gned to deli ver 400 cubic feet of wa ter pe r mi nute

Fig 2 10
. .

a ga i ns t a press ure of 300 po unds per squa re inch AS can be see n .

from the figure axi a l thrus t is pa rti a lly ba l a nced by means of a

,

ba l a ncing piston a t the high press ure end Of the pump which oper
-
,

a tes in a m a nner s imil a r to tho se a lre a dy des cribed An y dditiona l . a

end thrust is ta ken up by ba ll h e ari gs in the outboa rd bea ring

n .

This ba ll bea ring consi sts of two rings of 1% inch di a meter ste el
balls wa ter cooled an d Oil lubric a ted
, The b a l a ncing pisto n is
.
32 8 Y
T PES OF CENTRIFUGAL PUM PS .

fi tte d very loosely in order to keep the friction losses sma ll an d as ,

a res ult a cons idera ble a mount of wa ter le a ks p as t it into a ch a mber

a t the end of the pump which is provided with a di s ch a rge v a lve ,

a d pipe lea ding into the s uction

n Adjusting the va lve in this pipe .

regul a tes the press ure on the piston The bea ring a re of the ring . s

O iled type Lubric a tion is furnis hed by Oil cups to a ll pa rts of

.

the impellers a d dis ch a rge va nes where there a re close run ning
n -

clea ra nces .

Prof R C C a rpenter m a de exh a ustive tests of these pumps an d

. . .
“
,

Fig 2 19 gives the ch a r a cteri stic curves of the pump as dete rmined
.

by h i tests The ma ximum effi ciency was 80 per cent which is

s .
,

very good .

50 . WE ISE AN D M O N SK I PU M PS .

Wei se of H a lle Germ a ny a fa vora bly known fo

and M o n ski , ,
re r

their reciproca ting pumps a d more recently they h ve ta ken up ,

n a

the m nufa cture of centrifuga l pumps for both hi gh a d lo w pre

a n -
s

s ure ervice Their high pressure pumps a intere ting from a

s .
-
re s

con tructive ta ndpoint

s s .

Fig 2 2 0
. .

Fig . ives schem a tic ally cro s ection through a multi

2 20 g a s s

s t a ge centrifuga l pump built by thi fi m The w t r ter throug h s r . a e en s

S f
eeull p t yb R C C p
re t i P
or d i g f t h Am.
i S i t y
. ar e n e r n ro c e e n s o e e r c an oc e

of M ec h a n ic a l E n gin ee rs, 1909 .

330 Y
T PE OF C E NTRIFUGAL PUM PS .

the origin a l Sulzer design it diff ers funda menta lly in th a t the Weise
,

an d M o ns ki des ign necess ita te

s only one ch a nge of di rection of the

wa ter in the pum p while the Sul er des i gn h as a pa th for the wa ter
, z

which reverses severa l times The l a tter necess ita tes somewha t
.

complica ted c as tin gs especi a lly for the dis ch arge va nes This firm
.

, .

st a tes th a t i ts medium s i ed pumps obt a in a

-
z efficiency Of 70 to 7 5 n

per cent a n d l a rge units 80 per cent an d a bove .

Fig 2 2 1 gives a photo gra phic view of a multi s t a ge pump c a pa ble

.

of delivering 2 10 cubic feet of wa ter per mi nute a ga in st a head Of

Fi g 2 2 1
. .

1 57 5 feet or 280 c ubic feet of wa te r per minute a ga inst a hea d Of

,

1 14 8 feet .

In 1907 thi s firm pl a ced in opera tion a ten st a ge double c as ing - -

pump c a pa ble of delivering 1 06 cubic fee t Of wa ter per minute

a ga i n s t a he a d of 2 13 0 feet a n d occ as ion a lly pumping up to 2 3 00

feet This is rea ching a condition Of high stres ses An other pump
.
.

w as built c a p a ble of delivering 2 12 cubic feet Of wa ter per min ute

a ga inst a he a d Of 1 7 00 feet The pump requi red a 1000 h p driver -
. . . .

51 . BYR ON JAC KSON PU M PS .

The Byron Ja ckson Iron Works of Sa n Fra ncisco Cal m a nu ,

.
,

fa cture high press ure centrifuga l pumps Of the Ja eger type They
-
.

a l so build l w pre s
os ure pump
-
s for a l a rge v a riety O f w ork Fi g 22 2 . .

gives a cross sectiona l view of a t wo st ge high press ure pump

- - a -
,

w hich i typic a l of a ll their multi st a ge design s

s AS c a n be see n
-
.
 Y
B R ON JAC KSON PUM PS . 331

from the figure the impeller bl a des are ba ckwa rd turned a n d the
, ,

discha rge va nes a d reversing ch a nnels are simil a r in sh a pe to those

n

of the Ja eger pumps al ready des cribed At ta ched to the l as t im .

peller is a revolvi g piston wh ich serves as a ba l a ncing piston for

n

b la ncing the e nd th ust This piston h as a Slight clea ra nce be

a r .

tween it a d the c as ing ; this permi ts pa rt of the wa ter flowing

n

from the pe iphe y of the impeller to lea k to the ch a mber on the

r r

other side of the piston The pressure in this ch a mber wi ll then

.

be slightly less th an the d is ch a rge press ure This pressure exerte d .

over the a rea Of the pi ton prod uce a pressure in the direction oppo
s s

site to th a t produc d by the wa ter in the rest of the pump The

e .

b e ca u s e .
a t ws u m s

Sun s -
o

. U Q NC N Em " n o“ St l f '” 0” AA

Fig 22 2. .

a rea Of the piston is so de ign ed th a t the w a ter pressure upon its

ca n a l w a ys be gre a ter th an the a xi a l thrust This exc of pressure . e ss

moves the impellers a d h a ft slightly to the left which pe mits

n s , r

the high press ure wa ter to esca pe through the di ch a rge holes into
-
s

the suction of the l ast im peller I this ma nner a xi a l thrust is . n

b a l a nced automa tica lly An outside view Of this two stage pump
.
-

direct connected by means of a flexible coupling to a n induction

motor is shown in Fig 2 23 M ore sta ges c a n be a dd ed by simply
. .

introducing a nother section of e as i g with impeller a n d using a n

longer Sha ft bas e a n d tie bolts The bea rings

, , ring lubric ted . a re a

a d the p a cking on the s uction end h as a w a ter e a l

n The impellers s .

a d di ch a r ge v a nes a e of bron e a d the cle a ra nce ring

n s r on the z n s

impellers are renewa ble .

33 2 T YPE OF C ENTRIFUGAL PUM PS
.

Inpump design s with sh aft vertic a l the ax i a l b al a nce is sec ured

by means Of a ba la ncing pisto n simil a r tha t Shown in Fig 22 2 . .

illustra tes a n a rra ngement for ba l a ncing the a xi a l thrust of

Fi g 2 2 4
.

vertic l pump The wa ter i dra wn into the impeller A th ough

a s . s r

the suction pipe B a d is delivered into the di s cha rge cas ing a t a
n

62 m

Fi g 22 4
. .

pressure equa l to the di sch a rge hea d a t which the pump is opera ted .

Thi di cha rge press ure will extend to the ba l a ncing ring on eith r
s s e

s ide of the impeller minus a decreas e of press ure c a used by the

,
3 34 T YPE OF CE NT RIFUGAL PUM PS .

so n ron Works to the city of Sa n Fra ncisco for i ts fire bo a ts As

I .

s ho wn in the illus tra tion e a ch unit cons is ts of t wo two s t a ge pum ps

,
-

direct connected to a 6 00 h p 1800 Cu rtis stea m turbine buil t

-
. .

by the Genera l Electric C ompa ny The pumps a SO a rra nged th a t . re

they ma y be run either in pa ra llel or in s eries ; in the former c as e ea c h

unit delivers 668 cubic feet (5000 ga llons ) of wa ter per minute under
a di ch a rge press ure of 1 50 pound
s per squ a re inch a n d running in s ,

s eries it delive rs 334 cubic feet ( 2 500 gallons ) of wa te r per minute

under 300 pounds per squa re inch discha rge These two

Fig 22 6. .

fire bo a ts a sta tioned in Sa Fra ncisco Ba y an d co n s equently

re n ,

are required to pump lt w ter necess ita ting a ll pa rts coming

sa a ,

in conta ct with wa ter except the c as ings being cons tructed o f

, ,

bron e ; the nickel steel h a ft

z bron e a rmored The turbine
s s a re z .
,

opera ting condens ing under ste m a t 150 pound per squa re inch a s

a d no uperhe a t h as a wa ter r te O f
n s , pound per hors e po we r a s -

hour delivered a t the pumps when opera ting under 2 6 inche s

va cuum .

Fig 2 2 6 Shows a four st ge Byron Ja ck son pu mp direct connected

.
-
a

to a 750 h p C urti s ste a m turbine built by the Gener a l Elec t ri c

-
. .

C omp an y Eight of these u its were ins t a lled for the city Of San
. n
 JOHN M C DOUGALL C OM PAN Y PUM PS . 3 35

r
F a nci s co for fi re prot e ction E a ch unit is req uired to deliver 400 .

cubic feet (3000 ga llons ) of wa ter per minute under a disch a rge
pres ure of 300 pounds per squa re inch a n d a sucti on hea d f 15 feet
s O .

The unit run a t 1800 r p m The illu tra tion shows the top h a lf of
s . . . s

the p ump c as ing removed expo sing the impeller a n d discha rge ,
s

van es which together with the pa ck ing boxes a d followers a re

, ,
n ,

of bron e Axi a l ba la nce is sec ured as a lrea dy described The

z . .

unit is a four bea ring se t the turbine being connecte d to the

-
,

pump by mea ns of a flexible coupling The turbine opera tes non .

condensing .

52 . PUM PS OF TH E JO HN M C DO UGALL C ALE DO N IAN IR O N WOR KS

C O M PAN Y , LIM IT E D .

The John M c Do uga ll C a ledonia n Iron Works Compa ny Limit ed , ,

Of M ontrea l C a na d a a re licens ed to m an ufa cture Worthington

, ,

pump in C a na da It is therefore not necessa ry to give a n y deta iled

s .

des cription of their centrifuga l pumps but it is of interest to note ,

s ome of t he novel pump ins ta ll a tions ma de by this fi m r .

At the m unicip a l w a terwork Port Arthur Ont a rio w a ter s, , ,

turbine driven centrifuga l pump a e ins ta lled Fig 227 give a s r . . s n

illustra tion of thi in ta ll a tion Two wa ter turbines a e insta lled

s s . r

in common c as ing a d connect d by a coupling At ea ch fre end

a n e . e

of the turbine three t a ge centrifuga l pump is pl ced There i

s a -
s a . s

a coupling bet w een e a ch pump an d the turbine It is po ss ible to .

d ive a ll as one unit or one pump with both wa ter turbines or one
r , ,

pump w ith one turbine The l a tter m kes pra ctica lly duplica te . a

units Ea ch pump is c a p ble of delivering 1 1 1 cubic feet of w ter

. a a

p e r minute ( 1 000 000 B riti s h g a llons per 2 4 hours ) a g a ins t 4 00 feet

head Ordina rily this se t opera tes with one pump driven by both
.

w a ter w heel s in order to get rid of the end thrus t of the wa ter
w heel the loss due to unning under lo w ga te not be ing Of a n y
,
r

con equence to the city The pumps opera te a t 1 100 r p m

s . . . .

when pum ping a ga ins t 400 feet he a d for fi e se rvice or to fill the r

high level sta ndpipe a n d the s peed is decreas ed when less pressure
-
,

i s required .

In the city Of Winnipe g M a nitob a there i an unusua l ins t all a tion , , s

of three vertica l three s ta ge pumps for the wa terwork system -

s .

Fi g 2 28 illustra te s one Of these pumps which h as a c a p a city Of 2 7 8

.
,
33 8 Y
T PE OF CE NTRIFUGAL PUM PS
.
 JOH N M CDOUGALL C OM PAN Y PUM PS .
3 39

the g round level The h a fts pas s through a i locks a d in order

. S r ,
n

to get a t the pumps it is nece a ry to go into a c a isson under a

ss

press ure of 35 po unds pe r square inch .

E a ch of these vertic a l pump s h a fts h as three se ts of thrus t be a rings ,

one for the motor one for the sha ft an d one for the pump with slip
, , ,

couplings be tween them The pumps run a t 68 0 r p m when deliver

. . . .

i n g full pres s u e a n d the s peed i red uced by res ist a nce in the motor
r ,
s

c ircuit for lo wer press ures The effi ciency Of the whole unit is lo w
.
,

o wing to o pera ting a t slow s peeds the us e Of res is ta nce an d the fa ct

, ,

th a t the couplings a d pa rt of the Sha fting a re opera ting in the wa ter

n .

The city own s its o wn hydroe lectric p la nt a d mecha nica l efficiency ,

n

was not SO much a n object as the cost of a ttend a nce a n d m a inte

na nce .

Fig 2 2 9 illus tra tes a in t a ll a tion for the high pres ure fire servi ce
. n s -
s

of the city of Toronto Ca na da This ta tion conta ins two pumps

,
. s

Fig 230
. .

of two sta ges ea c driven by W tinghouse Pa rsons tea m turbines

h, es -
s .

E a ch pump h as a c a p a city of 556 cubic feet of w a ter per minute

( 5 000 000 B riti hsg a llon s per 24 hour ) ga in t 7 00 feet he a d w hen s a s

run nin g a t 1700 r p m The pump a n d t urbine ea ch h a ve their

. . .
34 0 TYPES OF CE NT R IFUGAL PUM PS .

own thrust bea rings the coupling being rigid for a lignment but
,

flexible for end motion As these pumps Opera te only a fe w .

hours ea ch year effi ciency is not importa nt but relia bility a n d

, ,

quick sta rting are the principa l objec ts A press ure of 300 .

pounds c a n be obta ined in less than a minute a fter the a larm is

turned in .

Fig 2 30 shows a ra ther novel i ns t a ll a tion Of a three s ta ge centrifuga l

.
-

pump direct connected to a Belliss triple expa ns ion engine The -

.

ca pa city of thi un it is 1336 cubic fee t of wa te r per minu te ( 12 000 000

s

B riti sh ga llons per 24 hours ) a ga ins t 2 07 fee t he a d when ope ra ting

a t 340 to 35 0 r p m This unit is ins ta lle d in the lo w level sta tion
. . .
-

Of the M ontrea l w a te rworks The unit was offi cia lly te te d by P o f . s r .

R J Durley Of M cGil l Univers ity in N ovem ber 1 909 an d nette d

. .
, ,

the builders a bo nus of $ 10 000 No indica to r c a rds were Offi cia lly .

ta ken but a repre enta tive of the engine builder took c ards an d
, s ,

from the indica ted horse po wer ho wn by the e c a rds the to ta l over S s

a ll e ffi c ie c y was 7 0 per cent

n Thi included not only pump a n d . s

engine lo e but l o the power required to drive the a ir an d fee d

ss s a s

pumps which were a tta ched to the m a in engine a d for which

, ,
n

no a llowa nce was ma de The following figures give the results of .

the te st :

RE SU LTS OF T E ST M ADE ON B E L L ISS T RIP LE -

E X PAN SIO N EN GIN E
AN D W O R T H IN GTO N T H R E E - STAG E T U R B IN E PU M P , M ON TR E AL
W ATE R WOR KS LO W -
LE V E L PU M P IN G STAT IO N , 1 9 TH AN D 20T H
N O V E M B E R , 1 9 09 .

Pump Pe rfo rma n ce .

Dura t io n of te st . . . ho urs 13
To ta l wa te r pumpe d d urin g tes t Imp ga s . l . 6 5 15 000
r
Ave a ge wa te r d e live re d pe r h o ur . Imp gals . . 50 1 155
Ra te p um i
of
p gpn e r 2 4 h o ur s . Imp ga ls . . 12 02 7 7 0 0
Ave ra ge press u re ga u ge rea d i n g a t p ump . . fe e t
t
fee
Ave r
a ge s uct io n ga uge re a d in g at pump . f t
ee

r
Diff e i l l
e n ce n e ve of ga uges . f t
ee

T t l v r g
o a rr
a e a e co ec te d h ea d f t
ee

Average w rk d o on e on wa ter pe r ho ur . ft lbs

. . 1 02 7 2 6 7 52 0
342 TY PES OF C E NTR IF UGAL PUM PS .

Fig 23 1 g
.
ives the stea m consumption per pump horse power p e r
hour as worked up from these te sts The dots represent ho uly . r

read ings a t va rious superhea t C urve A represe nts the gua ra nte e d
s .

s te a m cons um ption The increas e of stea m consumption due to a

.
.

ch a nge Of uperhea t from 120 F to 1 00 F is from

s
°
pounds t o
.
°
.

pounds a n increase of
, per cent Ther fore 3 per c t w a . e en s

lb s S f t
. o m
/ vr M

Fig 23 1 .

used as a correction in the foregoing ta bl The duty of the pump es .

a d engine a llo w ing for a ll he a t return e d to the boile s

n , n d a pplying r a
,

the neces a ry correction for exce uperhea t was 1 10 15 1 000 foo t

s s s s ,

pounds per thou a nd pound of stea m upe rhe a ted one hun d re d
s s ,
s

degrees .

This pl nt is novel in its design on a ccount Of the fa ct th a t it

a
,

h as a lwa ys been suppo ed th a t it w as imposs ible to m a ke a commer

s
 ALBE RGER PUM PS .
34 3

c ia l success with an engine driven centrifuga l pump delivering wa ter

-

a ga i ns t high he a ds While the duty obta ined does not compa re

.

fa vora bly with the best pra ctice in high duty pumping engines yet
-
,

when we cons ider the fa ct th a t the contra ct price was only one
third of the price as ked for vertica l triple expa n ion pumping engines
-
s ,

the inte res t on this difference a dded to the di ff erence in the cost of
engine house a d founda tions would more tha n compensa te for the
n

extra fuel cons umption .

53 . ALB E R GE R PU M PS .

The Albe rge r Pump C ompa n y

of N w York build high press ure
e -

centrifuga l pump of the Ja eger type They a lso build a line of

s .

l w press ure pump for motor engine or te a m turbine drive

o -
s , ,
s .

Fi g . gives a cross sectiona l vie w through a single sta ge low

2 32 - -

press ure pump for motor d ive The sa me des ign but with l a rger
r .
,

impeller is used for di ect conn ected tea m engine drive a n d where
,
r -
s ,

th e spe eds are very lo w a sec ond s ta ge is a dded Fo r high s peed .

-
344 Y
T PES OF C E NTRIFUGAL PUM PS .

work an d low heads a design simil ar to th a t shown in Fig 233 i s

,
.

u s ed. The dividin g line betwee n low an d high pressure work i s -

a bout 7 0 feet he ad .

Fo r higher hea ds th a n 7 0 fee t guide va nes are employed an d t h e

design is then simil a r to tha t of J aeger Fig 234 gives a cross . .

section through a four st a ge high p ress ure Albe rger pump

- - Th e .

impellers a e m a de of bronz e a n d ha ve b a ckward turned bl a des

r .

The disc h a rge va nes a re a ls o of bronz e an d the sh a ft is protec ted

,

Fi g 2 3 3
. .

by bronz e sleeves The be a rin g supports a e c as t integra l with the

. r

pump c a sing a d the journ a l a e ring lubric a ted C lose clea ra nce
n s r .

p ckings a re provided betwee n the di a phr gms an d impellers a n d

a a

betw en the di phra gms a n d sh a ft le ves w hich a re rene wa ble The

e a s e , .

unba l a nced a xi a l thrust is t a ken ca re of by a wa ter cooled m a rine -

type thrus t bea ring The Alb ge r pump is built up in section s

. er

exa ctly as des cribed in Article 38 A photogra phic view of this

.

pump is given in Fig 2 35 The pump i hown direct connected to

. . s s

a s te a m turbine c a p a ble of developing 7 0 h p The pump delivers . .

34 6 T YPES OF C E NTR IFUGAL PU M PS .

47 cubic feet ( 3 50 g a llo n s ) of w a ter per minute a g a ins t a pre

s s ure of

2 25 pound s per s qu ar e inch when running a t 2 4 00 revolution s p e r

Fi g 2 35
. .

minute Thi pump is used as a boiler feed pump This firm h a

. s . s

built pumps from 2 inches di a meter dis ch a rge to 2 4 inches dia met er

Fig 2 36
. .

discha rge h a ving a c a pa city ra nge from

,
to 2 4 00 cubic feet pe r
minute (50 to 18 000 gallons per minute ) .
 ALBE RGE R PU M PS .

Fig 236 g
. ive photogr phic view of a th ree s t age pump Th e
s a a -
.

ompa ny furni hed a pump of this typ e for one of the

Al be rger C s

large power houses in C hica go for fire protection The pump is .

G a l /m s p er Min ut e .

Fig 237
. .

direct conn ecte d to a 6 0 h p induction motor opera ting a t a syn

-
. .

c hro n o us speed of 1800 revolution per minute The pump was

s .

designed to deliver cubic feet (400 ga llons ) of wa ter per minute

a ga in st a tota l hea d of 300 fee t The dia meter of the di sc ha rge
.

p ipe is 5 inch es. Fi g 2 3.7 gives in full li n es t h e c h ara c te ris ti c curv es

 3 48 TY PES OF C E NT R IF UGAL PUM PS .

of this pump from test da ta the s peed of the in duction motor be in g

,

1 7 88 e volutions per minute

r It is interes ting to note th a t th i
. s

pump when runn in g a t 1 700 revolutions per min ute will give a .

much hi gher efficiency ; the tota l head of course is cons idera bly de
creas ed The dotte d lines Show the cha ra cteristic curves for t h is
.

l a tter condition .

Fig 23 8 repres ents a t wo st a ge pump d es i gned to comply w it h

.
—

the underwrite r conditions for fi protection se rvice The pu m p

s
’
re .

delivers 134 cubic feet ( 1000 ga llons ) of wa ter per minute a g in t a a s

Fi g 238
. .

pressure of 100 pound p qua re i ch This pump is direct c o n

s er s n .

nec t e d to a 90 h p direct current motor a d the fi gure s ho w t h e

— . .
-
,
n s

fi ttings furni shed with the pump as c a lled for by the under wri t e r
,

s peci fi c a tions Fig 2 39 gives the ch a ra cteri tic curves of thi s pu m p

. . s

ta ken from a ctua l tests a d shows a effi ciency of over 69 per cen t
n n .

The norm a l speed is 1400 revolutions per minute .

Another se t of ch a ra cteri tic curv for t wo st a ge pump is give n

s es a -

in Fig 240 The pump when deliv ring 40 cubic feet (300 ga llon s )
. .
e

of wa ter per mi nute a ga inst a hea d of 12 0 fe e t a d running a t 1 7 2 0 ,

n

r evolution s pe minute rg a ve a e ffi ciency

,
of 7 0 per
n cent .
 ALBERGER PUM PS . 3 51

A ra th er novel des ign of a l a rge pump is shown in the cross

section a l view Fig 24 1 ,
The three impellers work in para llel all
. .
,

ta king wa ter from a suction pipe of 28 inches di a meter an d disch arg

ing into a pipe of 24 inches di a meter The novel fea tures of this .

design e th a t the pump is built up of section a l un its a n d can be

ar

built in either one t wo three or four se ctions dependi ng upon the

, , , ,

qu ntity of wa ter to be h a ndled an d the speed a va il a ble an d a l so

a

upo n the design of the c as ing which i not only in sections but split
,
s

hori on ta lly a n d wi th remov a ble suction an d di sch a rge pipe fittings

z .

Fig 2 42
. .

Fig 243. gives a outline view of thi pump direct con necte d to a
n s

400 h p C urti s s te a m turbine m a de by the Genera l E lectric C o m

-
. .

pa ny a n d Fig 242 shows a photogra phic view of this unit Two of

,
. .

these units h a ve recently been in ta lled for the Sta nda rd Oil Co m
s

pa ny for genera l w ter supply a t their refinery Point Rich mond

a , ,

C al E a ch pump i s c a pa ble of delivering 186 0 cubic fee t of wa te r

.

per minute (2 0 000 000 ga llons per 2 4 hours ) a ga i ns t a hea d of

80 feet The s peed of the unit is 12 50 revolutions per minute
.
.

The Albe rger C ompa ny h a ve a pplied for p a te nts on thi s general

cons truction .
3 54 TYPE OF C E NTR IFUGAL PUMPS
.
 M ATH ER AN D PLATT PUM PS .
3 55

pan y e pl an t The genera l design of the pump is shown clea rly in

’
.

the cross sectiona l view Fig 246 The bea rings a re ring lubrica ted
-
,
. . .

The unbal a nced a xia l t h rus t is a lmost wholly ta ken care of by the
ba l a ncing pisto n P on the dis ch a rge end of the pump C h a mbe r A .

is fil led wi th wa ter from the high press ure di sch a rge side of the -

pump a n d ch a mber 0 is connected th rough a bala ncing va lve a d

,
n

pipe to the suction end of the pump If the unb a l a nced axi a l thrus t .

is towards the l e ft the cle a a nce bet ween ch a mbe r B a n d cha mber
,
r

A is incre as ed a n d th a t between B a d C decre as ed be c a u

,
s e the n

piston P moves to the left This ra i ses the press ure in B until .

ba l a nce is secured Should the a xi a l thrust be in the opposite

.

direction the press ure in B decreas es This mea ns of ba l an cing is

, .

nea rly a utoma tic a d depend omewha t upon the setting of the
,
n s s

ba l a nce va lve D Should there be a y unb l a nced axi a l th rus t not

. n a

ta ken care of by this ba l a ncing a rra gement it will load the ba ll n ,

thrust bea ring provided for tha t purpose upon the extreme dis
ch rge end o f the Sha ft The rest of the detai l cons truction of this
a .

p u p i s cle a rly ho w n in the

s d r a w ing .

A ra ther inte re ting high press ure pump of speci a l design for
s -

c olliery si king or dr a ining purpo ses i s

n hown in Fig 247 This s . .

pump is electric a lly driven by a a lterna ting current motor of 2000 n -

volt a d 350 bra ke h p The pump delivers 96 ubic feet (6 00 Bri t

s n . . c

is h ga llons ) of w a ter per minute a ga ins t a he a d of 1 2 30 feet The .

Speed of the unit i 1 4 50 r p m Even for the hi gh duty for whi ch

s . . .

this pump is ma de the over ll dim ens ions are sm a ll due to the hi gh
-
a ,

s peed of rot a tion The pump is a ra nged for suspending vertically

. r

in the shaft during sinking opera tions a n d when the ultima te depth ,

of the sha ft i rea ched the pump ca n be l a id hori onta lly a d used as
s z n

a n ordin a ry st a tion a ry mine pump

,
Fig 248 i s a photogra phic vie w . .

of thi pump uspe nded vertica lly an d the feet for hori onta l in
s s ,
z

st a lla t i con be pla inly seen

an .

Referring to the pump section in Fig 2 47 it will be noticed th a t .

,

the impell r a a rra nged in two groups ba ck to ba ck th a t is with

e s re , ,

the suction ide of four impellers fa cing the opposite direction from
s

the s uction end of the other four This ins ures a a lmost perfect . n

a xi a l b a l a nce but in order to de a l with a n y s light unb a l a ncing there

,

ma y be due to we a r or the weight of the impellers when the pump

i s suspended vertic a lly the Speci a l b a l a ncing device ho w n on the
, s

su c tion end of the pump is provi ded The two spe ci a l views of this .
3 56 T YPES OF CENTRIFUGAL PUM PS
.
 M ATHER AN D PLAIT
'
PUM PS
. 3 57

bal a ncing device how the con S

n e c ti o n to the u pper an d lower

sides of the ba l a ncing va lve The .

principle of opera tion of this va lve

is simil a r to th t of the b a l a ncing a

pi ton de cribed before The up

s s .

per side o f the v lve is connected a

by mea ns of a g inch pipe to the -

sp ce C between the econd a d

a s n

third impellers when the pum p i s

running hori zonta lly a n d to the

sp ce B between the first a d
a n

second impelle when the pump rs

i run ing vert ic lly

s n The lo wer a .

side of the bala ncing va lve i s

co nected to p ce B betwe e
n s a n

impeller 1 a d 2 in both c ass n es .

When the pump i opera ting s

ve rtic lly the higher pre sure in

a s

spa ce C helps to upport the s

w ight of t h e rot a ting p a rt

e s .

The ma in portions of the casi g n

of the pump conta ining th dis ,

e

ch a rge v n a e co tructed of
a es, r ns

clo gra ined c t iron The d i

se -
as . s

ch a rge va e a e provided with n s r

renewa ble h a rd bron e tips so z

cons tructed th a t the whole of the

throa t of the di ch a rge passa ge s

i of h a rd met a l re i ting wear

s ,
s s

a d e
n ily rene wa bl ; a further
as e

a dv nta ge of the bron e tip lies

a z

in it a cid resi ti g propertie

s -
s n s .

The impellers a e a ll constructed r

of bron e a d a e a ccura te ly ma
z n r

chin d a ll over c a refully bala nced

e ,

an d ecured to the mild stee l

s -

sh a ft by mea ns of a n a ccura tely

fit ted fe a ther key It wi ll be .
3 58 T YPES OF C E NTRIFUGAL PU M PS .

noticed th a t there is a centra l bron e sleeve upon the sha ft rota ting z

in the centra l interna l bea ring of the pum p ; a t the outer ends of
the sha ft other bron e sleeves a d nut hold a ll of the impe lle z n s rs

firmly in position The centra l be a ring is wa ter lubrica te d but i

.
,
s

provided with a sm ll a mount of grease lubric a tion a t sta rting

a .

The cas ings of the pump th ough of cas t iron a e provided with , ,
r

wrought steel hoop shru n k on to ena ble the pump to withsta nd the
-
s

grea t hea d Ea ch section of c as ing i ra bb ted into the a dj a ce t

. s e n

one a d the whole is m de to form one complete unit by long ub

,
n a s

s t a n t ia l steel bolt passing from the delivery to the suction end

s

c asings The pump sh a ft is connected to the driving ha ft of the

. s

motor by mea s of a fle xible coupling of speci a l design which a llows

n ,

for a y possible inequ litie in lining up a d for a y d a mage the

n a s n n

pump sh a ft might suff er in the h ft of the mine s a .

The motor is of a entirely enclosed type with short circui ted n -

rotor thus giving a very impl e con truction The st a tor i w ter
,
s s . s a

cooled by t a pping the d i ch a rge pipe w ith a 3 inch copper connection s -

a n d v a lve which a llo ws w a ter to enter under pre ss ure into the wa ter
,

j a cket of the sta tor The l a tter i provided with a inch outlet To
. s —
.

effect sta rting a speci a l multi point a utom a tic tra n former sta te
,
-
s r r

is provided a d the unit is st a rted up with the delivery v a lve on the

,
n

pump closed so th a t the po wer t a ken under st a rting conditions i

, s

not more th a n a bout 30 per cent of the power t ken a t full lo a d a .

M a ther Pl a tt Lt d l o build high pre sure centrifuga l

,
.
,
a s -
s

pumps of th dou bl inlet type A ro e ction a l view of such a

e -
. c ss se-

pump drive by Z e lly st a m turbin i shown in Fig 249 A

,
n a o e e, s . .

number of th u it i t ll d t the M ontrea l Wa ter a d

e se n s a re ns a e a n

Power C omp a ny s pl t i C d a Ea ch pump d live s 16 80 cubi c

’
an n an a . e r

feet ( 10 500 B riti h ga llo ) of w t r per minute a ga in t a hea d of

s ns a e s

4 1 0 f t when running a t 1 500 r p m

ee The norm a l lo a d on t he . . .

turbi e is 1800 h p The pump h as two impeller working in p ra llel

n . . s a ,

th a t is both imp llers rec ive wa t r from the uction pipe an d

,
e e e s

deliver it to a common disch a rge pipe The wa ter enters the suc .

tion pipe under a h e d of 10 feet a d flows into the middle of t h e a n

pump c a si g From h re it divid into four tre ms ntering o n

n . e es s a e

both ides of e ch impeller In thi s a rra n gement n o a xi a l thrus t

s a .

ca n be produced Fig 2 50 gives a cross secti n th rough the pum p

. . o ,

i which the cored ope ings throu gh t h di s ch a rge v a nes a e sho w n

n n e r .

These Ope nings permit the pas sa ge of the water from the mid dle o f
36 0 TYPE OF CENTRIFUGAL PUM PS .

the pump c as in g to the opposite suction side of ea ch impeller The .

wa ter on lea ving the guide va nes ente rs the two circul a r cha mbe rs C ,

the l a tter being united a t the dis ch a rge pipe The end of the pump .

s h a ft is s uppo rted in a long be a ri g an d both the turbine be a ring

n ,

an d pump be a ring are provided with forced lub ric a tion Th e s te a m .

turbine h as sixteen sta ges which are divide d into two groups of
,

eight wheels ea ch The sec ond group is of l a rger di a meter th a n the

.

firs t . All the ne wer t urbines of this type a re n o w built in this m an ner .

The turbine is supplied with superhea te d tea m a t 165 po un ds per s

Fi g 2 50
. .

s qu a re inch pre ss ure. The len g th of the entire un i t is a bout 2 1 feet .

The governor of this unit is of a somewh a t unique cons truction .

It is kno wn as the C horlto n Whitehe a d governor

-
Fig 2 5 1 gives. .

sectiona l views through this governor The entire governor c as i g

. n

is fi lled with oil

. The governor s pindle p ass e s through a long s leeve
an d oper a te s the throttle v a lve directly Pin s
. A are f as tened to the
c as in g a d form centers for the double bell cra nk DC The vertical
n .

r
a ms 0 C a re a
,
’
tt a ched to the weight s E E an d, the ’
hori
,
z ont a l
m
ar s D D ,
’
tr n
a s mit motion to the governor s pindle The weight . s

a re held in po s ition by the s pri n gs F F , the l tter

’
, tied together
a by
me ans of t h rough bolt s,
preventing the s pring s from mov i ng out w a rd .
 M ATHER AND PLATT PUM PS .
36 1

Th e compression in the springs ba l an ces the centrifuga l forces of

t h e weight in a ll positions so th a t the governor is is ochronous
,
. The

Fi g 25 1
. .

t bili ty for regul a tion is obta i ned by mea ns of a pisto n G which

s a

produces a das hpot effect either by forcing the oil out through t h e

sm a ll openin g H or by dr a wing it in t h ro ugh thi s open ing

. The oil
36 2 TYPE OF C E NT R IF UGAL PUM PS .

flow t hrough H is controlled by mea ns of a needle va lve J The .

pisto ns s imply destroy more or less a pa rt of the forces produce d

, ,

by the springs F .

Fig 2 52 is a photogra phic view of a n electric a lly dri v

. en p ump
s peci a lly des igned for boiler feed purpo ses The t wo u ni ts of the .

pump a re a rra nged to work in series The ultim te pre sure require d . a s

is 2 00 pounds per squ a re inch a t the boilers plus friction in piping

a n d economi e rs The pump delivers
z . cubic feet of wa ter per
minute when running a t 1 100 r p m Fig 2 53 gives a cross sectiona l . . . .
-

view of one section of this pump It consists of si sta ges with . x

s ingle inlet impeller

-
It will be noticed th a t th e re is no extern a l
s .

h e a rin g a t the di s ch a rge end ; thi s m a n th a t troublesome gl a nd e s a

under high press ure is dispens ed with a n d furt her th a t the sh a ft , ,

Fig 2 52 . .

is much sti ff er due to its short r le gth The a rra ngement for
,
e n .

se curing a xi l b a l a nce i
a e xtremely s imple a d is of the l a tes t
s n

type develo ped by this firm It i pl a ced on the extreme high . s

pressure end of the pump The va lv V i cre we d on the end of the

. e s s

sh a ft a d forms a close cle a ra n c e p c b twe en ch a mbers A a n d B

n s a e e

a d be t wee
n ch a mbers A a d C t h tota l clea ra nce between the
n n ,
e

va lve a d f b ing
n a ce s inch C h mber A i upplied wi th wa te r
e . a s s

a t the pre ssure of ome intermedi t ta g a n d ch a mber C is con

s a e s e,

n e c te d to the s uc tio d of t h pu mp nThe v a rying pressure in

en e .

ch amber B a u d by a light mov me nt of the va lve V with the

,
c se s e

sh a ft crea t the nece a ry ba l a nci g pres ure At the suction

,
es ss n s .

end of the pump there is a ext rna l be ring of the self oili g n e a -
n

type At the point where the h ft pass es through the uctio n

. s a s

ch a mber a stuffi ng bo x a d gl a nd a provided d a e fitted n re an r

with a water supply to prevent a ir lea ka ge into the pump Fig 254 . .
36 4 TYPE OF CE NT RIFUGAL PUM PS .

gives a vi ew of the pump Opened up an d shows the eas e of dis

as sembly .

Fig 2 55 gives a cross sec tion a l view of an electric a lly driven pum
-
p
.

especially designed for working in a dee p well a n d ad a pted to be ing

entirely submerged in wa ter During no rma l working hours when
.
,

the pump is in Opera tion the wa te r i lower a n d well below the leve l
, s

of the pump ; it i not necessa ry therefore for the uni t to be a t work

s , ,

duri ng the week e nds an d during the stoppage the wa te r rise s a bove
-
,

fi g 2 54. .

the pump a n d motor to a height of 100 feet The motor is pro .

te c te d from d a m a ge by the a i tight dome or he ll which s urrounds it

r -
,

but to m a ke quite cert a in th a t no h a rm c a come to it w hile under n

wa ter compressed a i r is s upplied to the dome from the s urfa ce

, ,

a press ure of a i r being m a int a ined in the dome gre a ter th a n t h e

press ure due to the level of the wa ter a bove the unit The unba la nced .

a xi a l thrus t i ta ken ca re of by a b ll thrust h e a ri g The pump

s a n .

delivers 16 0 cubic feet ( 1000 Briti sh ga llon s ) of wa ter per minute

aga inst a he ad of 2 6 5 feet a t 9 7 0 r p m . . .
36 6 TYPES OF C ENTR IFUGAL PUM PS .

A very novel u it i shown in Fig 2 56 a d in section in Fi g 25 7

n s . n . .

It is a combin a tion of a high pr ur centrifuga l pump a n d a s ingle

-
e ss e

s t a ge st a m turbine
e pa rticul rly de igned for fi e purposes T he
,
a s r .

impeller of the pump a n d the whee l of the turbine are mounted o n

F ig 256
. .

the me h ft the l t ter being supported in subst a nti a l self oili g

sa s a , a -
n

be a rings thus i u i g t h utmost rigidity The pump c an deliver

,
ns r n e .

104 cubic fee t (6 50 B i ti h g a llons ) of wa ter per minute a ga inst 2 50

r s

feet hea d The unit ru a t 3000 p m A th ere a e o ly two

. ns r . . . s r n

rot a ting wh eel th e gre a test pos ible simplicity i obta ined a d th e
s, s s n

unit c a n run continuously without fe r of bre kdow Owing to th e a a n .

36 8 TY PES OF C ENT RIFUGAL PUM PS .

hi gh speed of rota tion the quantity an d hea d are ve ry grea t in p ro

portion to the si ze of the pump . The ste a m cons umption wi th o nl y

one row of buckets in the turbine is na t ura lly somewha t high , b ut

 M ATHE R AN D PLATT PUM PS .
36 9

as the pump is only requi red for fi e purposes the economy is not
r ,

a vit a l co ns ider a tion .

Sever a l ch a ra cteristic curv es of M a ther Pl a tt pum ps are of

interes t Fig 2 58 gives the res ults of tes ts on
. . sm ll high press ure
a a -

single st a ge pump des igned to deliver 7 2 cubic feet (450 B riti s h

-

ga llons ) of wa ter per minute a ga ins t a hea d of 7 0 fe et The norma l .

revolutions per minute a re 1 12 0 The ma xim um pump e mc ie n c y

.

of 80 per cent was obtai ned a t 390 ga llons per mi nute with a hea d

of 73fee t when running a t norm a l speed When delivering 450 gal

.

lons per minute the hea d obta ined a t 1 12 0 r p m was 7 1 feet a n d the
. . .

efficiency of the pump was 7 7 5 per cent . .

Fig 2 59 give the ch a r a cte i stic curves of a four s t a ge pump d riven

. s r -

by a Belli M o rc o m engine run n ing a t 400 r p m

ss The norm l . . . a

lo a d of the pump is 1400 g llon per minute ga in t a he d of 150

a s a s a

feet Speci a l notice should be t a ken of the cons ta nt hea d obta ined
.

over the entire ra nge of output Although the maximum pump

.

effi ciency does not occur a t the ra ted loa d the combin e d effi ci e ncy
,
37 0 Y
T PES OF C EN TR IFUGAL PUM PS

of pump an d engine h as its ma xi mum of 6 4 per cent a t the ra t e d

load Th e pump effi cienc y at this loa d is 78 pe r c ent a lthough i t
.
,

/2 0 0 5 00 2 000

Fi g 2 6 1
. .

rea ches a maxi mum of 80 per cent when di sch a rgi ng 1 100 gallons p e r
minute .

Fig 2 6 0 give s the ch a ra cte ris tic curves obt ai ned from a tes t o f a
.

four sta ge high pressure pump designed to deliver 1000 gallon s p e r

- -
37 2 T Y PE OF CEN TR IF UGAL PUM PS .

be interes ting to kn ow th a t the nece ity of discha rge va nes be co m es ss

less a ppa rent the higher the peed of the pump w hich corres po d s s , n

to a impeller of sm a ll di a meter In order to obta in the excelle t

n . n

effi ciencies se cured with De La va l pum ps the conto ur a d sh a p e n

of the whirlpool cha mber a n d volute must be ma de quite s mo o t h

a d a ccur a te ; a d to a ccomplis h thi s a ll p u
n n mp cas ings bo t h f ,
o

the singl e sta ge a d multis ta ge types a e c ast in dry sa nd mold

— n ,
r s ,

which ins ure s uch a ccura te contou s a n d s moo th su rfa ces an d a l s o r ,

m a ke poss ible the cas ting of the e pa rts in one piec e obvi a ti g s , n

recourse to remova ble or sectiona l di a phra gms ins erted in th e

cas ing .

Split c as ings a e used throughout so a rr a nged th a t a ll movi g

r , n ,

wea ring a n d interna l pa rts a be removed after lifting the ca s i n g

, c n

cover a d removing the bea ring c a ps a d without disturbing t h e

n ,
n

s uctio or di ch rge connections the pumps thus diff eri g from t h e

n s a ,
n

s ide pl a t ingle t ge pump a d from multista ge pum p in w hi c h

-
e s -
s a s, n s,

t h interna l di a phr gms or p rtitio

e a olid disks threa ded on t h e
a a ns re s

s h a ft between the impeller s .

All the m a in members of the pump s ubject to we a r a re pro te c t e d

by ma ll repl a cea bl p rt uch as h a ft sleeve t he l a byrinth pa c k
s e a s, s s s,

ing ri g tt ched to t h e c as ing a n d to the impeller the bus hi gs

n s a a , n

protecting the di phr gm of multi ta ge pumps a n d the bearin gs

a a s s , .

The e p a rt s m nufa ctured to limit ga uges on a n interch a ngea b l e

s a re a

s ystem to fa cilit a te ra pid a d a ccur a te repl ac ement n .

B ron e is used for th e impell r

z d for ll w orking p a rts comi n g e s an a

in conta ct with the liquid except where other metal ar requi e d ,

s e r

by reas on of th cha ra ct r of the liquid h a ndl d Bronz e is us e d

e e e .

on a ccount of i t a bility to re i t corro ive a ction of wa ter an d to

s s s s

reta in it h a pe d poli h for indefinite period

s s an s an .

The speed of D La v l pump s much higher th a n th o s e

e a s a re

commo ly ued ; for in t n pump a re r gul a rly built to op e r

n s s a ce , s e

at edirect connected to D L v l i gle ta ge impulse turbin e s e a a s n -

s ,

a t s peed of 4000 r p m d a bove a n d a a extreme c as e i m

s . . . an ,
s n ,

p e lle h a ve been
rs cons tructed to gener t 7 00 fe t he a d a t 2 0 000 a e e

r p m By dint of c reful de igning a d sup rior finish chec k ed

. . . a s n e ,

by repea ted test these results h a ve been a ccompa nied by h igh

s,

efficiencies .

In ordina ry service re l a tively high he a ds a e genera ted per s t a ge r ,

single sta ge pu -
mm for ins ta nce being built for feeding bo i le rs
, ,
 DE L AVAL PUM PS .

aga ins t press ures of 1 7 5 pounds per squa re inch The a tta inment
.

of s uch h gh he a ds pe r sta ge with goo d effi ciency is m a de possible

i

Fi g 2 6 2
. .

by the use of l a byrinth pa cking ri g a tt ched to the cas ing a

n s a nd to
the impeller which prevent the l kage of w t r b ck from t h
,
ea a e a e d is
374 T YPE OF C E NTRIFUGAL PUM PS .

cha rge ch a mber into the s uction cha mber The la byrin th rin g .
s

give a longer lea ka ge pa th within limited spa ce a d introduce se v n

era l brupt turns in the course of the wa ter thus inc reas i g the
a , n

res is ta nce to lea ka ge As a cons equence the volume of lea kage a n d

.

the loss of po wer therefrom a re reduced Fig 2 6 2 s hows cle a rly the . .

des ign of these l a byrinth rings The rota ting pa rts a e s crewed
. r

upo both sides of the impeller a n d the ta tiona ry pa rts are fas te ned
n , s

to the pump c as ing both being renewa ble

, .

A c a reful exa mina tion of the ch arac te ris tic curves of De La val
pumps shows th a t the pumps a re des igned to be self regul a tin g -
,

Fi g 263 a
. .

a fea ture th a t h as been de cribed a n d a ttention c a lled to it in th

s e

theoretic a l di cussion in thi book The pump impeller is so de i gn d

s s . s e

th a t if the output is increased beyond the ra ted qu a ntity or po i t n

of m a ximum effi ciency a t a given spe ed the power t a ken by th e p M p ,

will drop ff ra ther th a n incre as e Thi s is import a nt s ince an y

o
’
.
,

pump may be overlo a ded as for i sta nce by fa ilure of the dis ch a rge
,
n

piping but in s uch c as e the power limiting ch a ra cteri tic pre vents
,
-
s

overlo a ding of the motor a d m a kes it feas ible to ins ta ll a m oto r

,
n

of ju t sufficient si ze to c are fo r the power required under n orma l

s

opera ting conditions Were a l a rger motor ins ta lled in order to t ake
.

c a re of a rising power ch a ra cteristic it would under ra ted cond ition

, , s,

no rma lly run underlo a ded an d hence a t lower effi iency c .

37 6 TYPES OF C E NT R IFUGAL PUM PS .

Fig 2 6 3 b
. how the ch a ra cteri tic of a ingle st a ge pump d e s i gn e d
s s s s s -

to deliver 5500 gallons per minute a ga ins t 45 fe et hea d or ,

ga llon per minute a ga in t 6 0 f t he d w h n driv n a t

s s ee a , e e r .
p . m .

by a 7 5 hor e po wer induction motor Thi pump was to

-
s . s se rv e as

a circul a tin g pump for a b a rometri c conden er ; the hi gher h

s ead be in g
 DE LAVA L PUM PS .

re quired to sta rt the ba ro metric condenser without assista nce from

the a i pump
r .

Atte ntion is c a lled to a nother peculi a rity of the he a d c a p a city

cu rves of these pum ps It will be noticed th a t they slope continu
.

ou sly from the shu t o ff point to the m aximum delivery th a t is

-
, ,

there is no rise of pres sure with initia l incre as e in delivery It is .

Fi g 2 65a
. .

cl a imed th a t such initi a l rise of press ure re nders a pump un suitable

under certa in conditions Fo r ins ta nce if the hea d genera ted a t
.
,

no delivery is 2 15 feet an d the hea d genera te d a t the delivery

,

corres ponding to the highest efficiency is 235 fe et the pump ,

could not sta rt delivering aga ins t a st a tic pres sure of 235 feet ,

such as th a t of a sta ndpipe without increas ing the speed o r

,

opening a by pass After bringing the delivery up to the des ired

-
.

a mou t the spee d co u

n , ld then be reduced the flow throttled or
, ,
37 8 T YPES OF C ENT R IFUGAL PUM PS
.
88 0 TY PES OF C ENTR IFUGAL PUM PS .

p ro tecting bus hing on the dia phra gms an d c as ing a n d the marine
,

type of thrust bea ring c a n be clea rly seen

,
.

Fig 2 66 s hows a photogra phic view of a single s t a ge mo tor

.
-

driven pump with the pump c as e cover lifted a n d bea ring ca ps

removed The sh a ft impeller a n d l a byrinth pa cking rings c a n
.
, ,

be removed bod ily without d i turbing the suction or disch a rge

s

piping .

Fig . gives a photo gra phic view of a direct connected five

2 67 -

s t a ge turbine driven pump - c a pa ble of deliveri n g

,
cubic feet
of wa ter per minute (500 g llo per minute) a ga in t a pressure
a ns s

of 600 pounds per squ a re inch when runni g a t 2 900 r p m n . . .

This se t i used for furnishing hydra ulic pressure ervice in steel

s s

mills .
 R . D . WOOD PUM PS . 38 1

56 . R . D . WOOD PU M PS .

R D
. . Wood
of Phil a delphi a build chiefly l a ge sin gle stage
Co . r

p umps for compa r tively lo w pre ures but h a ve a ls o built some

a ss ,

high pressure multis t a ge pumps The low pressure pumps are

-
.
-

usu a lly of the double i let type w ith a volute c as i g a lthough some
-
n ,
n ,

of the sma ller pumps h ve a suction pip a t one side of the cas ing
a e ,

with a single inlet impeller Fig 2 6 8 ho w a l rge single sta ge

-
. . s s a -

Fi g 2 68 . .

pump h a ving a c a pa city of 10 000 cubic feet (7 5 000 ga llons ) of

w ter per minute The hea d i 80 f et a d the di a meter of the dis
a . s e n

ch a rge pip i 54 inche This pump is for direct connection to a

e s s .

s te a m engine Fig 2 6 9 ho ws
. . s ingle t a ge pum p for h a ndling
s a -
s

se wa ge a t Sa nto s B r a il,
The pump di sch a rges a ga inst a hea d of
z .

6 0 feet . It is direct connected to a motor by mea n s of a flexible

coupling The pump be a rings a re supported upon the c a ing a d
. s n

a e ring lubric a ted

r Axi a l thru t is t a ken up by b a ll be a rings t the
. s a

extreme suction end of the sh a ft Fig 2 70 gives a cross se ctional . .

-
382 T YPES OF C E NT R IFUGAL PU M PS .

view of a simil a r pump but of nea rly double the c a pa city ; in th is

view the pump cons truction a be more clea rly see n The im c n .

peller in thi c as e is of the Ja eger des ign The clea ra nce betwe e
s . n

imp ller d cas ing is ma de cy lindrica l a d of the sa me di a mete r

e an n

on both ide of the impeller in order to as sist in equa li ing axi l

s s ,
z a

thrut Pa ckings a e provided t bo th ends of the h a ft a d are

s . r a s n

supplied with wa ter ea l to prevent a i lea ka ge into the s uction

a s r

pipe No dis ch a rge va nes a e employ d in these types of pumps

. r e .

Fi g 2 6 9 . .

Fig 2 7 1 gives a cro ss section a l view through a two st a ge pum p

.
- -
.

The Sul er type is here employ d The t w impellers a re pl a ced upon

z e . o

the sh ft ba ck to ba ck a n d a re definitely loc a ted by an a djusting

a , ,

nut a t the entra ce end of ea ch impeller The two impellers a e

n . r

sep a ra ted by a di a phr gm s upport e d in the c as ing by a r a bbet an d

a

held securely by studs The di a phra gm forms a close clea ra nce

.

with the hub of the impellers C lose clea ra nces are a lso formed .

between the impellers an d the c as ing upon cylindric al surfa ces The .

c as ing h as the usu a l cored pas sages of Sul er construction Th e z .

bea rin gs a re supported upon the pump casing an d re ring lubri a

3 84 T PE Y OF C E NT RIFUGAL PUM PS .

c at e d . Axi a l
t h rust is ta ken care of by ba ll thrust
on the sucti on end of the sh aft .

Fi g 2 7 2
. .

Fig . illustra tes a two s ta ge pump insta lled for town wa t r

272 -
e

s upply in Ju a re M exico Thi pump i motor driv n a d works

z, . s s e n

a ga ins t a hea d of 2 00 feet Its co n s truction i exa ctly as shown in

. s

the cross se ction of Fig 27 1

-
. .

57 . B R ODN IT Z SE YDE L PUM PS .

The centrifuga l pump built by B rod it s Seydel of Berlin n z ,

Germ a ny ,
noted for their ra nge of d pta bility a n d design
a re a a .

Thi firm builds di ff erent types for clea n a ndy a d limy wa ter
s ,
s ,
n s ,

a n d h as been bui lding centrifuga l pumps as a peci a lty since 186 0 s .

C uriously this fi rm h as not m a de u

,
se of guide v a nes even in thei r

high pressure centrifuga l pump desi gn

- .

Fi g 2 7 3 shows a type of high pr

. ure centrifuga l pump simil ar in
-
e ss

prin ciple to the Sul er pump The high pre sure centrifugal pum ps
z .
-
s

of Brodn it Seydel a e n a tura lly a much older type th a n the

z r

Su l er design
z .
 BRODNIT Z AN D SEYDE L PUM PS .

In th e pump illustra ted in Fig 27 3 the wa te r enters a t impeller B

.

an d flows th rough it di rectly into a spira l cas ing C It is then .

Fi g 2 7 3
. .

ca rri ed from C to D by me n of a connecting pipe loc ted out ide

a s a s

of the cas ing From D the w a ter ente t h ne t impeller E I

. rs e x ,
. n

the out ide view of thi pump i h own The co necting

27 4 s s s s . n

pipe c an be clea rly see n The wa ter a ft r pass ing through the
.
, e
386 TYPE OF C E NT R IFUGAL PUM PS .

e ond impeller a ga in enters a Spira l c asing The disch a rge end o f

s c , .

this pump i in the form of a conic a lly diverging pipe

s .

On the press ure ide of thi s pump the s h a ft is held in an enclose d

s

bea ring lubric a ted by a suita ble greas e while on the suction side the ,

s h a ft p as se through a ring oiled bea ring which i mounted in a

s -
s

pedesta l upported upon the bedpl a te

s .

As in the Sul er pump complete ba l a nce of the a xi al thrust by

z ,
a ,

arra nging the impeller ba ck to b a ck c be ecured o l y whe

s ,
an s n n

both sta ge deliv r the a me hea d a d when the outer rim cle r
s e s s n a

a nce is uffi ci tly l a rge compa red to the inner clea r nce To ta ke
s en a .

c re of the xi l thru t completely a thru t bea ring i neces a ry

a a a s s s s .

Fo l a rger h a d t w or three pu m p pl a c d on both ide of the

r e s o s e s s

driver a e u d I thi a rra nge ment only four st a ge c a be used

r se . n s s n ,
so

Fig 2 7 5
. .

tha t the e pump s t uit ble for high he d

s a re We h a ll not
no s a a s . s

discuss the qu ti wh th r t hi pump without guide v a nes a

es on e e s c n

compete comm r i lly wit h t h r pump h ving well designed guid

e c a o e s a -
e

va ne d ther f r hi gh r effi cien i

s an e o e e c es .

Fig 2 7 5 s how
. t rifug l pump d igned for w ter cont a ini g
s a ce n a es a n

a l a rge moun t of d or li my ma s Th e pump is quite simil a

a sa n s s se . r

to t h pump ju t d ib d e ce pt th t there i one ta ge on ea ch

e s e sc r e ,
x a s s

sid of the driving pull y The impellers a d le r nce rings a

e e . n c a a re

ma de of c t st l in ord r to wi th t d
as ee ive wea r Fo r clea n
e s an e x c e ss .

ing the mud drum d the id h a mb r of the impellers a upply

an s e c e s s

of clea n wa ter i provided Thi pump was m a de for the c a nto n

s .
“

of St G allen in Swit e rl a nd a d is u d for removing the dredged

. z n se

gr a vel a d cl a ynIt w ork s very s a ti sfa ctorily in spite of th es e

.

adverse conditio ns .
3 88 T YPE S OF C E NT RIFUGAL PUM PS .

as a speci alty the so c alled Liliputian centri fuga l pumps

-
. Fi g 2 7 8
.

s hows a two st a ge p u
-
mp of this type direct connected to a mo to r .

Fig 2 7 8. .

The illustra tion give a ide of h o w sm a ll this se t i Its use s

s n a s .

a re not very v a ried the pump se rving chiefly for opera ting s m a ll
,

wa te r founta ins .

58 . B E RL IN E R M ASC H IN E N B AU PU M PS .

The Berliner M chinenba u A G of Berlin Germa ny uccee de d

as . .
, ,
s

the firm of L S h w t ko pfl wh o built centrifuga l pumps simil a r

. c ar z
'

Kugel Ge lpk type

-
Recently the B erliner M as chin enb a u
e .

h a ve ch a nged the type to th a t hown in Fig 2 7 9 The unbal a nced

s . .

a xi a l thrus t of a ll the impeller i t a ken c a re of by a n a utom a ti c

s s

ba lan cin g piston K The a rra ngement of th is bal a ncing pisto n is

.
 BE R LINE R M ASC H INE N BAU PUM PS . 389

sh own more in deta il in Fig 28 0 The press ure preva iling in the . .

s p a ce b is s m a ller th a n the tot a l press ure developed by the p ump as a ,

pressure throttling ta kes pl a ce through the clea ra nce spa ce 0 If .

the axi a l thrust developed in

the impellers increases the pis ,

ton d move towa rds the left a n d

s

increas e the clea ra nce a t This

s c .

decreas es the throttling eff ect a t

c an d permit s the press ure in b

to increas e Consequently the

.

pressure on the piston increa e s s

a utom a tic lly corre ponding to

a ,
s

the increas e of the axia l thru t s

devel o ped i n t he i m pel le r s .

When this a xi a l thrus t de

creas e the pi tonmoves to wa rds
s, s

the right an d the regul a ting u s r

fa ces a t a pproa ch a ch other

c e ,

t h e reby increas ing throttling

a n d decre as ing the pre s ure in b s .

Thus the pre ure on t h ba l ss e

a c i g pi s ton a uto m a tic a lly i

n n n

cre e a d decreases with a

as s n

ch a nge of a xi l thru t Besides

a s .

this ba l ncing a r a ngement the

a r

pump i provid d wit h a coll a r thrust bea ring mounted upon the
s e ,

ext reme high press ure end of the ha ft d ta kes ca re of a n y u

-
s , an n

b a l a nced thrust not ta ken ca re of by the ba la nc ing piston Some .

a x i a l pl y is nece sa ry in thi s thru t h e a i g or el se it w ill not per

a s s rn

mit the slight xi l movement necess a ry when the ba l a ncin g piston

a a

opera tes The thru t be a ring is oil lubrica ted from the sa me reser
. s

voir as the a dj a cent bea ri g n .

The ca sing is ma de in one piece the hea ds a d di a phra gm s ,

n

being as embled in ide of this ca ing It i worth noti g tha t the

s s s . s n

guide v a ne a r ma de very short a d th t the revers ing ch nels

s e , n a an

conta in longitudina l ribs throughout their entire length preventing ,

the wa ter from whirling a d giving it proper guid nce a t entra nce n a

to the next impeller Fig 28 1 gives a photo gra phic view of a high
. .

pressure p ump ins talled in the Ro l an d M in e German y Th e pump , .

39 0 TY PE OF C E NT RIF UGAL PUM PS
.
39 2 T PEY OF C E NTR IFUGAL PUM PS .

conn ected to motors of their own man uf a cture Fi g 28 2 . .

a fiv t ge high pre ure pump Th e rigid cons truction o f

e s a
- -
ss .

the o uter be a ring is worth noticing T h e .

pump shown in Fig 283 h as the suctio n .

entra nce on both Sides which pa rtia lly ba l ,

a c es the axi a l thrus t

n Fi g 284 s ho ws a . .

mall high press ure pump with vertica l sh a ft

s -
.

The motor is mounted a t quite a di ta n s ce

from the pump beca use the la t te r often o p

c ra te s s ubmerged .

59 . WILLAN S R O B IN SO N PU M PS .

ob ins on of Rugby Engl a d

Willa n s R ,
n ,

build centrifuga l pumps to the des ign o f s

M Jens Orten B ovingr . The genera l design — .

of their multi s tage pump i shown in Fig -

s s .

285 Wa ter enters through the s uction pipe

.

the left a d passes through the impe llers

on n

u ively From the exit of an y impeller

s c c e ss .

the w ter ente rs guide va ne an d is then

a s

g u id d by rever s ing chea nnel to the entr a nce s

f the ne t impe ller

o Axi a l b l a nce is a uto
x . a

mat i lly ecure d by the b a l a ncin g pis ton n

ca s o

th high pre ure end of the pump The

e -
ss .

p r t io of
o thi epi to n
a c a n be e a s ily under
n s s

t d from the fi gure

s oo A proof th a t thi . s

b l ncing a rra ngement i e ffi cient lies in t h e

a a s

fa ct th t no thrust be a rings a e furni shed wit h

a r

F ig 2 8 4
. . th e pumps The impe lle rs a d guide va n es
es . n

a re m a de of bron e d a ll the wa ter p assa ges z , an

i
a re e t her m chin ead or fi i h e d by h a nd The
n s s h a ft i m a de of . s

nickel t el a d wh re it p as e throu gh the stu ffi ng boxes it is

s e ,
n e s s

cover d by r wa bl e bron e bu hi g The be a rings a e cas t in

e ene z s n s . r

one pi with t h e end cover a d a of the ring lubrica ted type

ece s n re -
.

The m i c i g i a i gl pi c lin d with bron e The guid

a n as n s s n e e e e z . e

v a ne di a phra gm fit ylind i lly into the c ing The stu ffi ng

s c r ca as .

boxes o n both the high an d low pressur e nds pa ck aga ins t lo w -

e
 DAYTON PUM PS .

pres sure o nly , an d th e on e on th e uction

s en d carries a wa ter

Fig 28 5
. .

This firm builds a sta nd a rd line of bo th high an d lo w press ure -

pum ps Fig 286 illu tra te s a high pressure pump direct con nec ted
. . s -

to a direct c urrent moto r

-
.

60 . DAYTO N PUMPS .

The Da yton Hydra ulic M ac hinery Co mpa ny of Da yton Ohio , ,

build high press ure centrifuga l pump with a suction pipe a t ea ch

-
s

end of the ma chine a d the disch a rge pipe loca ted in the middle of
n

the pump casing Fig 28 7 give a longitudina l cross section of a

. . s

two st a ge high press ure p u

-
mp Wa ter is a dmitted a t both ends a n d
-
.

pas ses t hrough the two end impellers a n d di sch a rge va ne into s re

versing ch a nnel s which lea d the wa te r to both ides of the middle

,
s

impeller The wa ter on le a vi ng this impeller pas se through a se t

. s

of disch a rge va nes to the disch arge pipe In this a rra ngement axi a l .
39 4 T YPES OF C E NT R IFUGAL PUM PS .

thrus t elimina ted ; but a two s t age m a chine virtua lly cons is ts f
is - o

four impellers bec a use the wa ter passe from either end of the p uffi p
'

,
s

to wards the middle Such a n arra ngement requires a some w h a t

.

longer pump The o ly a xi a l thrust possible in this a rra ngem e t

. n n

would occur if a y impeller should deliver a slightly h igher hea d th a n

n

Fi g 2 8 6
. .

th ecorre po di g i mp ller i t h th r h lf of the pump This some

s n n e n e o e a .

wh a t l rg r h e a d ma y b pro d u e d by m a ll error of con s tru

a e e ction
c s s

an d c a not be in a y w a y ri u
n A ot h r d cid e d adv nt a ge of
n se o s . n e e a

thi a rra g me nt of i mp ll i th a t no hi gh pres s ure tuffi ng b x

s n e e e rs s -
s o

i h e sa y
s ec s Th e only tu ffi ng b
r . ne sa ry a on the two su
s o xe s ces re e

tion ends an d these a re ma de air tight by introdu c ing a wa te r se a l

,
-
.
39 6 TYPE OF CE NTRIFUGAL PUM PS .

Fig 2 89
. represe nts a fire un derwri ters pump Thi s pump ca n’
.

deliver cubic feet ( 1000 ga llons ) of wa te r per minute aga in st

1 10 pounds per s qua re inch press ure The pump is direct connected
.

Fi g 2 88
. .

to a 1 00- h p . . Westinghouse typ e

motor run ni g a t 1400 revo
S n

lution s per minute The motor is entirely enclosed a d is venti

. n

l a te d by mean s of a blower driven from the pump sha ft .

Fi g 28 9
. .

It will be noted th at in these illustra tions the pump cas ing is divided
into two pa rt a long a hori ont a l pl a ne the t wo p a rts being doweled
s z ,

together a d th a t a ll the pipe connections both for the suction a d

,
n , n

disch a rge are in th e lowe r h a lf ca in g Thi s permi ts of quick a ccessi

,
s .
39 8 TY PES OF C E NTRIFUGAL PUM PS .

bility beca use the upper h a lf cas ing c an be eas ily an d quickly
removed The pa ckings be tween sta ges a re kept in pl a ce an d rigidly
.

held by long screws pas sing through the cas ing as c an be seen in Figs ,
.

2 88 a n d 289 The impellers guide va nes revers ing ch a n nels a n d

.
, , ,

clea ran ce rings a e not mad e in h a lves a n d a e a ll l ifted out of t h e

r ,
r

lower h a lf c as ing together with the sh a ft The sh a ft is supported .

either in ba ll bea rings or in ba bbitt lined bea rings The l a tte r are
-
.

Fi g 2 9 1
. .

ring lubrica ted Th e only pa rts of the pump subject to a n y a ppre

.

c i a ble we a r a re the cle a ra nce ri ngs between impeller a n d c as ing an d ,

these c an be very eas ily renewed .

Fig 290 ho ws to good a dva nt a ge the a ccess ibility of th es e pumps

. s .

The illustra tion hows a biple x volute pump This type cons ists of
s .

a s ingle sta ge pump h a ving t wo impellers working in pa ra llel The

-
.

pump c an deliver 802 cubic feet (6 000 ga llons ) of wa ter per min ute
a ga ins t a he a d of 3 5 feet when running a t 1 800 revolutions per

minute The dia meter of the di ch a rge pipe is 15 inches

. s .

This firm a lso builds highs peed pumps without di scha rge va nes ,

known as volute pumps Fig 2 9 1 Shows a high s pee d volute pump

. .
-

h a ving a disch a rge pipe 8 inches in di a meter The pump c a n de .

liver 2 00 cubic feet ( 1500 ga llons ) of wa te r per minute aga in st a

400 TYPE OF CE NTRIFUGAL PUM PS .

c an be remove d singly or by removin g the couplin g the entire in

,

te rio r of th e pump c an be removed asse mble d The impellers are

.

fas tened to the sha ft by mea ns of keys an d lock n uts which permi t
,

a ls o of eas y d isas sembly Axi a l thrust is ta ken care of by exte ndin g

.

the entran ce side of se vera l impellers so a s to form a rota ting dis k

or pla te beyond the impeller bl a des This exte nded Side of th e
.

impeller serves as a side pl a te for the dis ch a rge va nes The ends
.

of the di scha rge va nes form a close cleara nce with this rotatin g
di s k . As the other s ide of the impeller ext e nds only to the begin
n i ng of the dis ch a rge v an es there wi ll be a di fle re n c e of press ure of
,

th e wa ter es c a ping into the cle a ra nce s pa ce on either s ide of the

i mpe ller the exte nde d side being in a regi on of hi gh er pressure
, It.

is not po ss ible entirely to ta ke ca re of a xi a l thrus t by th is arra nge

ment beca use the press ure in the cleara nce s pa ce c a nnot be definitely
dete rmined Bo rsig therefore uses a th rust be a ring to prevent axial
.

movement of the s h a ft .

With very l a rge hea ds the a xi a l thrust is ba l a nced by arra nging

th e impe llers in oppos ing groups which of course necess ita te s th e
,

use of ri gid couplings Fi g 29 4 shows such a a rra ngement To

. . n .

ai d ass embly th es e groups of impellers are pl a ced on a common bed

pl a te with the driving motor This bedpl a te is extended in both

.

directions beyond the pump c as ings a n d is provided with s lides so

,

th a t the casi ng c a n be dra wn a wa y from the inn er pa rts of the pump

by means of j a ckscrews In Fig 294 the cas ing at one end of th e
. .
 BORSIG PUM PS .

pump ha s been drawn a wa y The pressure pipe conta ins a non .

return va lve an d the suction elbow is provi ded wi th a safety va lve

, ,

8 0 th a t even if the non return v a lve -

bec om e s i no p er a t i ve th e suction
piping is protected a gai nst high
pre ure
ss .

A Bo sig builds low pressu

. r re ce n -

t ifu
r g a l p ump s for qu an titi es of
wa te r from 7 0 to 1000 cubic feet pe r
minute a d for hea ds up to 100 feet
n .

Fig 29 5 shows the s t a nda rd fo rm of

.

their low pressure pumps which con

-
,

s i sts of a s ingle i mpe ller with wa ter

entra nce on both Sides of the impell er .

The middle piece of the spira l ca s ing

h as ve y broa d feet resting upon the
r

bedpl a te Att a ched on either side

.

of this cas ing are the suction hea ds

in the form of elbows poin t ing down
w a rd The e elbows h a ve c a st to
. s

them bra ckets supporting the bear

i ngs . The two suction elbows are
bolted to a common suction pipe .

The di ch a rge pipe i a pa rt of th e

s s

s pir a l ca s ing a n d is conic a l in form .

The pir a l form of the ca ing with

s s

i ts div e rgi g di c h rge pipe redu c es

n s a

gr adu a lly the velocity of the wa te r

a fter it le ves the im peller a n d no
a

dis ch a rge va nes are used Th e pa ck .

ing on both ends h as a wa ter sea l as ,

the e pa cking must prevent a i le k

s s r a

a ge into the pump c a s ing The be a r .

i ngs a e rin g lubric a ted a n d lined with

r

ba bbitt Wh enever these pump s a e

. r

driven by mea n s of belts a d pulle ys n

a third bea r ing i s provided w hich is

mounted upon the bedpl a te as s hown in th e illustra tion The entire .

design is ext remely rugged .

40 2 T YPES OF C E NT RIFUGAL PUM PS .

Fig 296
. s hows a sma ller low pr essure pump di rect con ne cted to
-

an ele c tric motor

. In the sm a ller low pressure designs th e ca sin g is
-

Fig 29 5
. .

m a de cylindric a l instea d of spira l an d no conic a l disch a rge pipe is

used bu, t in th is d e ign as well a in the previou s s peci a l a ttention
s ,
s ,

pa id to proper wa ter guid a nce The cross sections at entra nce

-

was .

to a d exit from the impellers a e of a mple dimen ions an d every

n r s ,

s h a rp turn is a voided as far as po ssi ble .

404 TYPE OF C E NTRIF UGAL PUM PS .

the thrust bea ring The oil fin a lly overflo ws through

. an o il drain
ba ck into the o il cha mber o f the bea ring .

Fig 29 8
. .

. gives the cha ra cte istic c urves of a t wo stage pump ca pa ble

Fig 299 r -

of delivering cubic fe et ( 1000 ga llons ) of wa te r per minute

Ger/l oo s p e r ” wri/f a .

fi g 299
. .

a ga inst a hea d of 230 feet The pump runs a t 1 140 r p m

. . . . Th e
e fli c ie n c y a t this point is the maximum a bout 7 2 per cent
,
.
 GEBAUEB PUM PS . 405

63 . G E BAUE R PU MPS .

Fr Geba uer of Berlin Germa ny builds both high pressure an d

.
, ,
-

lo w pre ss ure centrifuga l pump s The low press u

-
. re centrifug a l pump s
-

a re of ra ther novel d es i gn in th a t the b as e pl a te for both the pump

Fi g 300 . .

an driver is used as a p a rt of the pump s suct ion pipe Thi s a r

d its
’
.

ra ngement ma ke s these pump r ther che p a n d imple The w a te r

s a a s .

c a n be a d mitted a t either end or a t both end s of the b as e Fig 300 . .

Fig 30 1. .

Shows a pump of this de ign c pa ble of delivering 353 cubic feet

s a

of wa ter per minute aga i t he d of 33 feet Thi s pump is one of

ns a a .

the sm a ller unit a d imil r d esigns a e built up to ca p a city of

s, n s a r a

5300 cubic fee t per minute .

4 06 TYPES OF CE NTRIFUGAL PUM PS .

The p ump isof the d o uble s uc tion typ e h a ving one impeller .

Wa ter is admit te d through the end elbow s which a e bolte d directly

, r

to the b ase The cas ing is of spira l form

.

a n d cont a ins no dis ch a rge v a nes .

Fig 30 1 represe nts the norm a l type of a

.

high pressure ce n t iffi gal pump m de by thi s

-
r a

firm The c asing is in cylindrica l sections

.
,

o n e section for e a ch s ta ge of t h e pum p The s e .

s ec tions a ra bbete d one into the other

re

a n d a ll held together by me a ns of through

bolts If for a n y reas on a nother sta ge is to

.

be added a ll the old pump pa rt c an be s

u sed except the sh a ft a n d the through bolts ,

which would ha ve to be ma de longer This .

co ns truction of c as ing is s imil a r to the Jae ge r

des ign This pump h as of course disch a rge

.

va nes of the u sua l form The suction pipe .

extends downwa rd through the bedpl a te a n d

h as the outbo a rd be a ring a d t hrus t a tta ched n

to i t Th e motor is d ir c t connected to the

'
’

. e

p M b
p y me n s of a fle x ible coupling
a .

Fig 302 hows a mine s ha ft pump of

. s

Geb a uer design The suction pipe join s th e .

pump c s ing at two point 180 a pa rt an d

a s
°

the dis cha rge pipe is a l o tta ched to the s a

cas ing a t two points 180 a pa rt The moto r a d s tep bea ring are
°
. n

mounted directly a bove the pump a n d the entire structure is ca rried

,

in fra me m a de of ch a nnel irons

a .

64 . ALLE N PU MPS .

W . .llen So n C ompa ny of Bedford Engl a nd ma nufa cture

H A , , ,

centrifuga l pumps for a l a rge ra nge of service Their multis ta ge .

pumps h a ve a wa ter guida nce imil a r to the Ja eger type a n d the

s

pump is built up in section imil a r to most Ja eger types There is

s s .
,

however a somewh a t novel di ff erence in th a t no holes a e provided

,
r

through the impeller sides for securing axi a l ba l ance Fig 303 gives . .

a longi tud in a l cross section through a four s t a ge p ump It will be -

.

noted th a t the impeller hub is so a rra ged th a t it forms a pa rt of th e

n

wa ter passage a t entra nce to the succeedin g impe ller A possible .

408 TYPES OF C ENTRIFUGAL PUM PS .

objection to this design might be tha t the wate r on lea vin g the re
versing cha nn el enters a rota ting elbow before it rea ches the entra nce
of the next impeller This rota ting elbo w mus t produce some rota tion
.

of the wa ter depending pa rtly upon the smoothness of the elbow an d

,

pa rtly upon the velocity of the wa ter Even if the entra nce a ngle of
.

the impeller is des i gned to a llow for this rota tion there must be ,

considera ble dis turba nce a t the end of the revers ing ch a n nel whic h ,

cont a ins ribs to prevent the wa ter from rota ting It is extremel y
.

difficult correctly to design the entra nce edge of an impeller bl a de

when the wa ter is rota ting beca use the velocity of rota tion is n o t
,

un iform a cross the entra nce w idth .

Fig 304 . .

Axi a lbal a nce is secured by b l ncing piston on the hi gh pressure

a a a -

end of the pump The piston ma rked ( l ) in the figure rota tes wi th
.

a running cle a r a nce bet w e en it an d the bu s hing (2 ) fas tened to t h e

pump cas ing The cylinder ma rked (3) is a lso a tta ched to the pump
.

cas ing an d is provided with a fa cing a gains t which the ra di a l fa cing

on the bal a ncing pis ton run s The coll a r thrust bea ring on the
.

s uction end of the pump permits of a very slight a xi a l movement ,

which also give sZa sma ll varying clea ra nce between the ra di al fa cings
of the piston ( 1) an d cylinder Th e short circuited wa ter from
-

the periphery of the l as t impeller flows from the an n ular cha mber
 ALLEN PUM PS . 409

A to cha mbe r B The cleara nce between the pisto n ( 1 ) an d the

.

cylinder (3) regula tes the press ure of the wa ter in cha mber B When
.

this pressure is la rge enough to overba la nce the axi a l thrust the ,

Fi g 305
. .

clea ra nce be twee n pisto n a n d cylinder increases bec a use the sha ft
h asa Slight a xi a l pl a y which a ga i n dimini s he the pre
,
s ss ure in ch a m
be r B . When the axi al thrus t is gre a ter th a n the press ure exerte d on
41 2 T YPES OF C E NT R IF UGAL PUM PS .

impe llers an d guide va nes are ma de of bronze ma chine d an d polish ed ,

a ll over The guide va nes a re bolted to the sides of the revers ing
.

ch ann els a d a ll a re held in pl ce by a common c as ing The portio

n a . n

of the i mpeller which runs in the revers ing ch a mber is encircled by

a gun met a l liner or cle a r a nce ring
-
The wa ll s of the rever ing . s

ch a nnel a e fi tted wi th gun meta l bus hings where they fit into

s r -

the c as ing in order to prevent corrosion ; The sh a ft runs i

, n

ring lubrica ted bea rings a n d i kept in its correct pos ition by a
-
s

colla r thrust bea ring The uction gl an d is wa ter se a led an d the

. s ,

28 4 0 0

20 4 0

-
ao 4o

s t o

zoo o 30 0 0 40 00 500 0 00 0 0

Fi g 308.
.

gl a nd on the press ure side h as to wi thsta nd a tmospheric press u re

o ly
n .

W H Allen So n s an d Co build centrifuga l pumps to suit both

. .
,
.

the type of driver n d the required working conditions Th is i

a . s

best illustra ted by the two ch a ra cteristic curves shown in Figs 307 .

and 308 .

Fi g represents the perform a nce of a pum p for which the

. 307
workin g conditions require a sma ll va ri a tion in power with grea t
varia tions of hea d a n d quan tity the driver in thi s cas e being a three ,

phas e motor Fig 308 gives the ch a ra cteristic curves of a pump

. .

for which the working conditions a e consta nt hea d with grea t r

v ri a tion of qu ntity the driver being a t a m ngine

a s a ,
s e e .
 V
NO E L AN D M ISC E LLANEOUS DESIG NS .

65 . N OV E L AN D M ISC E LLAN EO US DESIG NS .

A novel type of centrifuga l pump h as been des ign ed by Profess or

N ova k for which he cl a im quite high effi ciencie
,
s Fig 3 09 give s . . s

t wo cross ectio n s of a ingle t a ge high pre ure N o va k pump built

s s -
s -
ss

for R M eyer M ii lh im Germ ny The pump cons ists of a single

.
,
e ,
a .

impeller 13% inches in dia meter surrounded by a se t of floa ting ,

Fig 309 . .

guide va n es which a re free to rota te a bout the i mpeller a n d which

,

a re centered upon it hub Surrounding this floa ting se t of guide

s .

va n es is a cas ing conta ining ta tiona ry guide va nes The w ters . a

enters the impeller as in the usua l designs but upon lea ving it strike , s

the floa ting guide va nes As the entra nce a ngle of these va nes i
. s

not co rrect when they a re ta tiona ry the wa ter impinges upon

s ,

them a n d sets them rota ting until the entra nce a ngle is nearly
correc t for proper entra nce of the wa ter The wa ter then lea ves .
41 4 T YPES OF C E NT RIFUGAL PUM PS .

the rota ting guide va nes a t a lo wer spee d a d higher pressure Th e n .

guide va nes in the cas ing a re intended to tran s form the re m a ining
velocity into press ure The genera l idea is excellent bec a use the
.

ma in object is to obvi a te shock a n d entran ce lo sses to the guide

va nes while the wa ter velocity is high The flo a ting guide va nes .

can be sa id to a dj ut their entra nce an gle a utoma tica lly by incre as in g

s

or decreas i n g the pee d of rota tion to s uit an y certa in qua ntity of

s

di ch a rge Te t h a ve been ma de with the floa ting guide va nes

s . s s

held ta tiona ry a d left free a d the t wo re ults compa red to pro ve

s n n s

the superiority of the l a tter a rr ngement Such compa rison is o f a .

course not convincing bec a u e the floa ting guide va ne held t a

,
s s s

t io a ry a e not of the prope r de ign for s t a tiona ry u

n r se The pump s .

Fi g 3 10 . .

should be compa red to one of t nd rd des ign h a ving the sa me s a a

d i a m eter of impeller d de igned for the a me qu a ntity a d

an s s n

s p ed
e .

It is quite doubtful if thi s de ign become commercial A s ca n .

certa in power is expended in keeping the va ne rota ting a n d the s ,

tra n form tion from velocity into pr sure in these va nes is not very
s a es

e fli c i e n t .

M ess rs Thomas Pa rker Limited of Wolverh a mpto n Engla nd

.
, , , ,

build the Ree Ro tubo pump the impeller of which is illustra ted
s r ,

in Fig 3 10 The six inner v nes a e m rely rib to prevent the

. . a r e s

whirling of the w a ter The centra l re ervoir of the impeller is full

. s

of wa ter a n d is kept under centrifug l pre sure when the pump o a s r

ta te The ma in idea of the inventor is the sm a ll di ch a rge no le

s . s zz s

at t h e periphery of the i mpe ller which produce a no zle e ff ect an d , z

 NOV EL AN D M ISCE LLANEOUS DESIGNS . 41 7

ness of form a d perfect smooth ness of e as in g very good effi ciency h as

n

bee n obta ined with pumps opera ting at high speeds This firm h as .

pa tented a design o n this principle for multis ta ge work .

Q uiri C i e of Schiltigheim G.erm a ny build a cre w pump which , ,

s

is a l o a centrifuga l pump
s The impeller A ( Fig 3 11 ) h as a coni c a l
. .

a ddition B w hich c a rries bl a des of a s cre w form the object of which ,

is to lea d the wa te r gra du a lly from a a xi a l to a ra di a l flo w The n .

wa te r flowi ng r dia lly strikes the a pex of the cone a n d is guided to

a

the screw w here it is se t in rot tion a n d lifted by centrifuga l force

,
a .

The a xi a l velocity gra dua lly decrease as the ra di a l velocity increas es s ,

un til fi na lly the wa ter ent r the a ctua l impeller A These pump e s . s

are u ua lly built in pa irs with the motor between them Axi l
s . a

thrust is then elimina ted .

Klein Sc h a li
, B ecker of Fr a nkenth a l Germ a ny build mul t i
nz n , ,

s t a ge centrifug a l pumps h a ving a wa ter guid a nce Simil a r to the

Ja eger type Fig 3 12 give a ectiona l view of a s ix t a ge pump

. . s s s

an d hows the novel fea ture of a cylindrica l c as ing m a de in ha lves

s ,

conta ining the guide va nes d rever ing ch a nnel but o con an s s, s

s tructed th a t the wa ter pre ure i not e erted upon it The cas ingss s x .

merely se rve to center the va rious unit of guide va ne n d reversing

s s s a

cha n nel a d a n be m a de very light The a xia l thru t is ta ken up

s, n c . s

in a thrust he a i g which erve both as a journ l d thru t be a r

r n ,
s s a an s

ing This be a ring is a utoma tic a lly oiled by the sma ll oil lifting
.
-

device shown on the extreme end of the bea ring The bea ring .

itself is wa ter cooled .

 VII . T ESTIN G OF C ENTRIFUGAL PUM PS .

66 TE STIN G PU M PS
. M E T H ODS OF .

The te sting of centrifuga l pumps is quite Simple a n d the metho d ,

us ua lly employed depends pa rt i a lly upon the qua ntity of wa ter di s

ch arged an d the hea d a ga ins t which the pump opera tes Severa l .

importan t points however must be t a ken into a ccou t when testing

, ,
n

a pump a n d determining i ts a v a il a bility to the s ervice in question .

If the pump is to be inst a ll ed wi th a long s uction line the test mu st ,

be conducted under s imil a r conditions or proper a llowa nce s must be

ma de The sa me thing holds true of the dis cha rge end of the pump
. .

If the f iction he a d i l rge the pum p w ill opera te quite di ff erentl y

'

r s a ,

from a pump without friction hea d All meas urements of pressur . e

s hould be m a de in a tr ight portion of the wa ter piping suffi ciently

s a

fa r removed from elbows turns or v a lves Q uite,erroneou s re

s ult
, s .

ma y be obt a ined by t a king read ings a t poin ts where the wa ter i s

dist urbed .

I me as uring the hea d or s uction pres s ure a mercury column o r

n ,

a c a refully c a libr a ted pre s ure g a uge is us ed A Pitot tube pl a ced i n

s .

the center of the pipe an d pointing aga in t the strea m is conn ected s

to the mercury column or press ure ga uge The rea dings will then
, .

give the pressure hea d plus the velocity hea d a t tha t point .

In determini ng the tota l he a d a ga inst which the pump opera tes t h e

da tum line hould be ta ken a t the center line of the pump If t h e
s .

disch a rge pre sures a e t ken a bove the center line the dista n c e
s r a ,

from the center line to the measuring point should be a dded to t h e

observed he d a d if the uction hea d i mea sured below the cen te r
a ,
n s s

line its dista nce from the center line should be a dded to the observ d
, e

suction hea d .

There a e four method genera lly used for obt a ining the qua ntity
r s

of wa ter pas sing through a pump

1 Th e Venturi wa ter meter .

2 The Weir t a n k
. .

3 The no z le
. z .

4 The Pitot tube

. .
4 20 TEST IN G OF C E NTR IFUGAL PUM PS .

l . Ve n turi Wate r M ete r

The Venturi wa ter meter when ca re
.
,

fully han dled gives very a ccura te results Fig 3 13 shows an e x

, . .

c e lle n t a rra ngement for te sting a centrifuga l pump by me a ns of a

Venturi meter Va lve D is used for regul a ting the req uired hea d
.

a ga ins t which the pum p is to work The Venturi meter indic a t . es

t h e qua ntity of wa ter pas s ing through it on a mercury U ,

ga uge the two limbs of which a connected to the Venturi met r

,
re e

at E a d F n The difle ce of press ure recorded in the two limbs

.
'

re n

Fi g 3 14
. .

of the ga uge i in proportion to the squa re of the di cha rge from the
s s

pump If the sc a le i gr du ted i proportion to the squa re root

. s a a n s

of the pressures so as to rea d in ga llons or cubic feet per minute the

qua ntity of wa te r di sch a rged a be rea d without the necessity of
c n

any ca lcul a tion wh a tever The figure a lso show . recordin g s a ,

registering an d indic a ting device con nected in pa ra llel with the

,

mercury column .

I m king the co n nection to the Venturi met er

n a s d di ch a rge an s

pipe it is a dvisa ble to con ect to four point a round the pipe 90
n s ,
°
 M ET HODS OF TEST ING PUM PS . 4 21

a part so as to get a vera ge rea dings The a ccura cy of the Venturi

, .

meters is gua ran tee d by the ma kers of the ins truments to be within
two per cent .

2 W i Ta n k
. e r Fo r me as uring qua ntity of di s ch a rge e peci a lly
.
,
s

for very l a rge uni ts the Weir ta nk c a be used Fig 3 14 illu tr tes
,
n . . s a

a Weir t a nk with the wa ter dis ch a rging through the c a libra ted

orifice C are mus t be ta ken th a t the ta nk is divided into two or

.

more ch a mbers so tha t the di turba nces of inflow do not aff e t the s c

wa ter a t di cha rge The Weir ta nk is ofte n used as a chec k upon

s .

the rea dings ta ken by one of the other methods especi a lly th a t of ,

the Pitot tube or no le beca use when using the Weir ta nk it doeszz ,

not require much grea ter l a bor to ta ke the noz le or Pitot tube z

meas urements .

To meas ure the hea d on the crest of a Weir a hook ga uge i gener s

a lly u ed
s This instrume n t is quite common a n d need no e xpla
. s

na tion The ca le o n the ga uge a n d it vernier c a n be ma de to

. s s

rea d directly in cubic feet or ga llons .

3 Th N
. le e Fig 3 15 shows a
ozz . .

form of no le which c an be em zz

ployed for meas uring qua ntity of

di sch a rge A y no le used should
. n zz

be c a refully c a libra ted a d a curve n

of disch a rge dra w giving the qua n n ,

Fi g 3 15
tity of di ch rge from the no l e for
s a zz
. .

va rious pressue a t the point where the pres ure ga uge i a tta ched
r s s s .

The no le shown in Fig 3 15 h as the following dimens ion

zz . s

D inches di a meter .

d inche di a meter
s

The di ta nce from the center of the no le to the press ure ga uge
s zz

is 18 inches Fig 3 16 gives the di c h rge curve of thi s no le

. . s a zz .

The qua ntity of w ter disch arge d in g llon per minute c an be e x

a a s

press ed by the formul a

Q 1 963 d z
c

in which 0 i a coeffi cient of di ch rge equa l to

s for this n o le s a zz ,

h the pre ure he d in the n o le in feet d the di a meter of the

ss a zz ,
42 2 TEST ING OF C E NT R IFUGAL PUM PS .

throa t of the no le in inches an d D the dia mete r of the nozzle a t

zz ,

which the press ure is me as ured .

If the pressure is meas ured in po unds pe r squa re inch by a pres

s ure ga uge then
,

h = 2 3 04 P + m
.
30 6

in which P i the pre sure in po und pe squ a re inch a n d m t h e

s s s r

distan ce from the center of the no le to the pressure ga uge in fe t

zz e .

t o 4 0 6 0 w M ” ” a
Pr e s s ur e 0”

Fig 3 16. .

Th is formul a ta kes into con idera tion the velocity of a pproa ch an d

s

the wa ter column betw een the no le n d ga uge zz a .

4 The Pi t t Tu
. o b Fig 3 1 7 illustra te
e . . an a rr a ngement fo r s

meas uring the qua ntity of disch a rge by me a ns of a Pitot tube .

The tube me s ures the velocity hea d a t the ori fice of the dis cha rge
a

nozzle a n d the qua ntity of di ch a rge

,
be found from the formul a
s can

307
4 24 TESTING OF C E NT RIFUGAL PUM PS .

the l a tter s effi ciency The pump e fli c ien cy c an th e n be eas ily

’
.

obta ined.

When us ing a stea m turbine or engine it is often diffi cult a n d

expens ive to determine the bra ke horse power input to the pump
-
.

Fo r this reas on the gu a ra ntee s are frequently m a de in te rms o f

work delivered to the wa ter an d a moun t of ste a m used by the driver .

It is us ua lly express ed in the number of million fo ot pounds delive re d

to the wa ter pe 1000 pounds of stea m at some specifie d se t o f

r

conditions a n d this is ca lled the duty of the uni t A be tte r

, .

wa y of express ing the duty of a pum ping unit is in the n umbe r

of foo t po unds delivered per of energy in the te a m use d

s .
 IN D E X .

Abso l u te pa th o f wa te r t h ro u gh i m p le r
e l .

Ad a m
Ad uj s ta b le gu ide va n es
Albergar pu mps .

Alle n , W H , Pu. mps

. .

Alli s C h a lme rs pumps

-

An gle s Ba a n d de te rm i n a t io n o f
Appo ld
Arma t ure n a n d M as c h in e n a b ri k A G pumps f . . .

f
Axi a l c e n t ri uga l pumps .

Axi a l t h ru st

see a lso Ba la n c in g pis to n .

i n d ic a to r .

Ba bbit te d be a rin gs .

Ba c k wa rd t urn e d i mpe lle r bla d es , e ffi c ie n c y of

Ba la n c i n g p is to n , Alle n , W H . .

Allis C h a lm e rs -
.

Be rli n e r M as c hin e n ba u
B y
ro n Ja c kso n

Ku ge l-
G e l p ke

M a the r an d Pla tt
Sulze r .

Willa n s and Ro bin so n

B a l a n c i n g, s ta t i c a n d d y na mi c
Ba lls e a te d li n in gs
Ba ll t h rus t be a ri n g .

Ba n k i .

Ban n i n ge r

Be a ri n gs , c o o le d by ra d ia t io n .

Be a rin gs , d es ign of

lub ric a ti o n of

me t h o d of ba bb it ti n g
r ubb in g s pe e d o f .

s ph e ri c a l se a te d .

te mpe ra t ures of .

t h rust , see T h rus t b e a ri n g .

wa te r c oo li n g o f .
4 26 INDE X .

Bla d e d es ign , i mpe lle r .

Bla des , umbe r o f i mpe lle r

n .

Bo rsi g pu mps .

Bro d n it z a n d Se y de l p umps
Bu ff a lo pumps
Bu ilt u
-
p im pe lle rs d es i gn o f
,

B y ro n J ac kso n pu mps

C .

C a lc ula t io n of a lo w press u
-
re ce n tri uga l pump f .

i mpe lle r a n d gu i de va n e s
t wo h igh pre ss u re -
ce n t ri fug al p umps
C as i n gs a n d h e a ds , s tre n gt h o f
c i rc ula r

t h e gu ide va n e
Ce n te r y
o f g ra vi t , d e te rmi n a t io n of

Ce n tri u f
ga l a c t io n in c e n t ri fug al pum ps
f o rc e o f w a te r

pumps , a xi a l

c a lc ula t io n of lo w pre ss u
-
re

c h a ra c te ri s t ic o f

c lass ifi c a t io n o f

d o uble in le t t pe y
su c t io n t y pe 22 9
dut y of 42 4
e ffi c ie n c y of 85
e n c lo s e d . 4
e xpe ri me n ta l e ffi c ie n c y of 106
fi r c ti o n lo sses i n 12
fo r fi re b o a t se rv i c e 30 1 3 13 3 17 , 3 34
e n gi n e se rvi c e 2 7 0 , 36 7
s t a t io n s . 3 10 325 334 339
y
h d ra uli c mi n i n g
fi r c t io n he a d of

h i gh p ress ure
-

h isto ri c a l re vie w of

h o rizo n ta l
y
h d ra uli c e fli c i e n e y o f

h yd ra ulic lo sses in .

lo w press ure
-

me c h a n ic a l lo ss e s in
4 28 IN DE X .

Ch o rlto n -
Wh ite hea d go ve rn o r .

Ci rc u la r pump c as i n gs .

C las s ifi c a t io n o f ce n t ri fug al pumps .

C lea ra n ce rin gs , i mpe lle r

s pa c e losse s .

press ure d iff e re n c e i n

f
C o e fi c ie n ts o f n o z z les .

Co mbes
Co n d it io n eq u a t io n

Co n s umptio n o f po we r a n d e fi c ie n c f y
C ri ti ca l s pe e ds
a p pro xim a te o rm u la fo r f
g pra h i ca l d e te rmi n a t i o n o f
ma t h e ma ti c a l d e te rmi n a ti o n of

Cu
rves x an d A

y
Da to n pu mps
fl
De e c tio n o f pu mp sh a ts , d e te rmin a tio n o f f
De La va l pumps .

Des ign o f i mpo rta n t pump pa rts

De te rmin a t io n o f e xi t v a lue s by mea n s o f K an d A c urv e s

Di a ph ra gm pa c kin gs , d es ign o f
pa c kin g rin gs
Di a ph ragms , s tre n gt h o f
Di me ns io ns o f s u c t io n pipe .

Di re c tin g va n es , use o f

Do uble s u c t io n pu mp
Du lut h fi re -bo a t pu mps
y f
Dut o f c e n t ri uga l pumps
y
D n a mi c ba la n c i n g

Effi c ie n c y a n d co n s umptio n of po we r
y
h d ra uli c .

w h e n 5.
B.
o f b a c k wa rd t u rn ed i mpe lle r bla d e s

c e n t rifugal p u m ps , e xpe ri me n t a l
wi t h o ut gu i de va n es

f
o rwa rd t u rn e d i mpe lle r bla d e s

mu lt is ta ge c e n t ri ug a l p um ps f
ra d i a l i m pe lle r bla d es
wi t h va r yi ng n umbe r o f re vo l ut io n s
Elbo ws , flo w of wa t e r t h ro ugh
E n c lo se d c e n t ri fug al pumps
E n t ra n ce an d e xi t d i me n s io n s fo r e q ua l h ea d s w i t h va ryi n g q ua n t i t i es .

v e lo c it y
d ia g ra ms w h e n a. 2 90
°

En tran ce pipe
velo c it y di agra ms
 IN DE X . 4 9

En t ra n ce v e loc it y di a gra ms fo r 6, 2 90
°

Esc h e r, Wyss a n d Co pu mps .

Eule r, Le o n h a rd .

Evo lven te n pu m ps .

Exi t an d e n tran ce d ime n sio n s fo r e q ua l h e a ds wi th va ry i n g quan ti ties

ve lo c it y
di agrams

°
8g 2
90
g p
ra hi ca l de te rmma tio n of .

E xpe ri me n ts upo n th e e ffi ci e n c y o f ce n t ri fug l pump

a s

F
Fire bo a t , ce n t ri fug al pumps fo r, B y ro n J ac kso n
M o rris , I P . . 3 13 3 1 7
Wo rt hin gto n . 301
en gi n e , ce n tri fug l pump f
a s or . 27 0, 33 7
sta t io n s , c e n r t i fug l pump
a f s or 289 , 3 10, 325 , 334, 339
so
Flo w o f wa te r t h ro ugh a c e n t ri fug l pump
a 1
pi pe s a n d e lbo ws . 104
Fo rc e d f
ee d lubric a ti o n of bea ri n gs . 16 1
Fo rward turn e d i mpe lle r b la d e s , e fli c i e n c y o f 98
FOt t in ge r

Fra n c is

lo sse s
i n i mpe lle r
suc tio n pi pe
Fun da me n ta l e q ua ti o n .

Ge n e ra l Ele c t ri c Co mpa n y , t h rus t be arin g

Gira rd
Go ve rn o r, C h o rlto n Wh i te h ea d -

Gra phi c a l de te rmi n a tio n o f c e n te r o f gra vi t y

d e fle c t io n o f sh a ts f
e xi t ve lo c i t y
di a gra m
mo d ulu s o f se c ti o n

Gui de va n es

ad j ut s a b le

an gles , in flu e n ce upo n p p
e ri h e ra l ve lo c i t y
an gles , in flu e n ce upo n sh o rt- c i rc ui t lo sses
c a lc ula tio n o f
430 IN DE X .

a t e n t ran c e

lo sse s i n .

wi t h o ut

H igh -p ress ure c e n t ri fug al pumps

c a lc ula tio n o f

H isto ri ca l re vie w o f c e n t ri fug l pum p

a s

H o t wa te r pac ki n g .

Hu t ze lsied e r, C .

y
H d ra u li c e ffi c i e n c y
lo sses i n ce n t ri fug
al pum ps
mi n i n g, c e n t ri fug l pump
a s fo r

a n gles , in flu en ce upo n pe ri ph e ra l spe e d

s h o rt- c i rc uit lo sses
bla d e d es ign
d e ve lo pme n t
e ffi c i e n c y
o f bac k wa rd t urn ed

bui lt up .

c h a n n e l s e c tio n

c l e a ra n c e ri n gs

d es i gn o f
i n vo l ute s
lo sses .

n umbe r o f bla d es of

s t re n gth o f

st ress es in h igh spee d -

s t re t c h o f, d u e t o c e n t ri fug l f
a o rc e

t a n ge n t i a l s tresse s in
wi th ri ve te d bl a d es
43 2 IN DE X .

M o d ulus o f sec tio n , de te rmin a tio n of

M o rris I P , pu
, m ps
. . .

M ulti ro to r c e n t ri uga l pumps

-
f
e fi c ien c y o f

Ne uman n , F rit z
Ne w Y o rk fi re -bo a t pu m ps .

N o va k pumps .

No ve l a n d misc e lla n e o u s d es ign s

f
No z z le co e fi c ie n ts o f flo w
meas ure me n ts by
Numbe r, c o rre c t , o f i mpe lle r bla d es
o f re vo l utio n s , e fi c ie n c y wi t h va r yi ng

O il fo r bea ri n gs .

Ope n ce n t ri fug al pumps

Orte n Bo vi n g,
-
Je ns

Pac ki n gs , d es ign of

d ia ph ra gm

h o t w a te r

lea t h e r c uppe d
w a te r se a l o f
Pa c ki n g ri n gs fo r d i a p h rag ms
Pa ra bo lo id , ro t a ti n g
Pa rs o n s , R C .

Pe rip h e ra l ve lo c it , d eyt e rmi n a t i o n o f

i n flue n c e o f i mpe lle r and guid e va n e an gles upo n
Pi to t t ube , me as ure me n ts by
f
P a rr, Pro f . A

Pi pes , flo w of w a te r t h ro ugh .

Po illo n .

Po n c e le t
Po rt a ble c e n tri fug al pum ps 2 7 0, 36 7
Pra s il .

Pres ure
s d ifi e re n c e i n c lea ra n c e s pa c e .

Press ure d is tri but io n i n c e n t ri uga l pu mps f

he ad o f c e n t ri fug al pumps
on bea rin gs .

Pro s ll
Pump pa rts , d es ign o f i mpo rt a n t
y
t pe s , see C e n tri u g p
a l u m p t pe s f y .
 IN DE X . 43 3
PAGE

Ra di al impe ll e r bla d es , e fi cie n c y of

st resses i n i m pe lle rs .

pumps
Re dte n b a c h e r .

Ro ta t in g para bo lo id .

Ro un d pu mp cas in g .

Sa n Fra n c isco fire -bo a t pumps

Sc h wa mmkrug
Sc re w pumps , Qui ri
Se gn e r

f
Sh a t pumps
Be rlin e r M as c hin en b a u
Ge b a uer .

Jae g er

M a t h e r a n d Pla t t 3 57

f
Sh a ts , c a lc ula tio n o f c ri ti ca l spe e ds o f 19 1
d e fle c t io n o f 1 99
f
Sh oc k a n d ri c t io n lo ss es . . 9 , 10 , 1 2
Sh o rt c irc u
- i t lo ss es . 33 , 86
in flu e n ce of gui de upo n
va n es 30
i mpe lle r a n gles upo n 30
Sh o rt- c irc uit wa te r 19
Sh ro u d rin gs , st re tc h o f, d ue to ce n t ri fug l f
a o rce 18 1
Simps o n pu mps 4 15
Sin gles ta ge pu m ps o
Sk o d a pu mps 32 1
Spe e ds , cri t ic a l . 19 1
Sph e ric a l se a ted be a rin gs 32 0
Spe ide l, E . 1 46
Spi ra l pump c as in g 5, 73
St a ti c b ala n c in g 19 1
1 7 0, 1 93,
4 34 IN DE X .

St re n gth of dia phragms .

i m pe lle rs

p u m p c as in gs an d h e a ds .

Stress e s , c a lc ula t i o n o f i n p ump d ia ph ra gms , c as i n gs a n d h e a ds

i n h igh -s pe ed impe lle rs .

pro d u c e d by c e n tri ugal f f o rce

St re t c h i n im pe lle rs d u e to c e n r a t i fu g l f o rce

s h ro ud rin gs d ue to c e n r t i fug l f a o rc e

St ri e bec k , e xpe rime n ts o n lu bri c a tio n of be a ri n gs by .

f
St ufi n g bo xes , d esign o f
Su bme rged pu mps
Su c ti o n h e a d , ma ximu m pra c ti c a l

d i me nsmns of

fi r c t io n in
v e lo c i t yf or 6. 33
Sulze r pumps 4 , 7 , 106 ,

T e mpe ra t ure of be a ri n gs
T es ti n g o f c e n tri fug al pumps
T h eo re tic a l h ea d of pump .

T h ro t tlin g, re g ula tio n by

T h rust , a xia l .

see Ba la n c i n g Pis to n .

bea rin gs
Allis C h a l me rs
-

Arma t ure n a n d M asc h i n en f

a b ri k A G.

Ge n e ra l E le c t ri c C o
ma ri n e t pe y .

ro lle r

se l f
-
a lig n i n g .

Wo rt hin gt o n
T ri -
ro t o r ce n t ri fug al pum ps .

T urbi n es ,
y
h d ra uli c .

y
T pes o f ce n tri fug al pumps
see C e n t ri fug al p um p t y pe s .
 F e b rua ry , 1 9 1 1

SH O RT T IT L E C A T A L O G -

O F TH E

Pub l i c a t i o n s an d Im p o r t a t i o n s
OF

D V A N N O ST R A N D C O M PA N Y
.

23 M U R R AY AN D 27 WA R R EN ST R EE T S, N . Y .

P ri c es m a rk e d w i th a n as te r i s k a re NE T .

A l l bi n d i n gs a re in c l o th un l e s s o th e rw ise n o te d .

A B C Co de .
( Se e Cla use n Th ue ) -
.

Abbo tt , A V Th e
. . El e c trical Tra n smi ss i o n
o f En e rg y .

A Tre a tise o n Fu el ( Sc ie n c e Se rie s No . .

Te sti n g Ma c h i n e s ( Sc ie n c e Se rie s No . .

Ada m , P Pra c ti c a l Bo o b i n d i n g
. Tra n s by T E Ma w k . . . .

Ada m s, 1 W .Se wers a n d Dra in s fo r Po pu

. lou s Distri c ts

Addyma n , F T Pra c ti ca l X R a y Wo r
. .
-
k
AI Co de . ( See Cla use n Th ue ) -
.

Aikma n , C M Ma n u . re s a n d th e Pri n c i ple s o f Ma n u

. ri n g 8 170 ,
Ale xa n de r, J H Ele me n ta r Ele c tri ca l E n g i n e e ri n g
. . y r z mo , .

Un i ve rsa l Di c ti o na r Of W e igh ts a n d Me a su re s y 8 vo , .

Alla n , W Stre n gth o f Be a ms Un de r Tra n sve rse Lo a ds

.
( Sc ie n c e Se ri e s .

No . 1 6 mo ,

Th e o r o f Arc h e s y ( Sc ie n c e Se ries No . I 6 mo , .

Alle n , H . M o d e rn Po we r Ga s Pro duc e r Pra c ti c e a n d Appli c a ti o n s I z mo ,

Ga s and Oi l E n gin e s . 8 vo ,
An de rso n , F A . Wa te r
Bo iler Fee d. . 8 vo ,
An de rs o n , Ca pt G L Ha n dbo o fo r th e Use
. . . k of Elec tri c i a n s 8 vc ,
An de rso n , J W
'
Pro spe c to r s Ha n dbo o
.
’
. k . Iz mo ,
An des, L Ve ge ta ble Fa ts an d Oi ls
. . . 8 vo ,
An ima l Fa ts a n d Oi ls Tra n s by C Sa lte r . . . . 8 vo ,
y
Dr in g Oils, B o il e d Oi l , a n d So li d a n d Li q ui d Drie rs . 8 vo ,
-
f
Iro n Co rro s i o n , An t i o uli n g a n d An ti c o rro s i ve Pa in ts -
. Tra n s . by
C Sa lte r
. . . 8 vo ,
Oil C o l o rs , an d Pri n te rs In k ’
. Tra n s by A M o rri s
. . and H R o b so n
.

8yo,
Tre a tme n t of Pa pe r fo r Spe c ia l Purpo se s Tra n s by C Sa l te r . . . ra mo ,
y
. .

Ann ua l R e po rts on th e Pro gre ss o f Ch e m i s tr .

Vo l . I
( 1 90 4)
.

Vo l . II ( 1 00 5 )
.
 D . V AN NOSTR AND COM PA NY S SH ORT TITLE CATALOG ’
3

An n ual R e ports o n th e Pro gre ss of Ch e mi str y .

Vo l III ( 1 9 0 6 ) . . 8 vo ,
( 1 9 0 7)
V o l IV . . 8 170 ,
Vo l V ( 1 9 0 8 ) . . 8 vo ,
Vo l VI ( 1 9 0 9 ) . . 8 vo ,
Argan d, M Ima gin ar . y Qu a n ti ti e s. Tra n sla te d f
ro m th e Fre n c h by
A S Ha rd . . y .
( Sc i e n c e Se ri e s No . 1 6 mo ,

Armstro n g, R .
, and Ide ll , F E . . Ch im n e y s fo r Furn a c e s an d Stea m Bo ile rs .

( Sc ie n c e Se rie s No . . r6 mo ,
Arn o l d, E Arma tu re W i n di n gs o f Di re c t Cu
. rre n t D - y n a m o s. Tra n s by .

F B DeGre ss
. . 8 vo ,
As h e , S . W .
, an d K il y J
e e , . D . Elec tric Ra i l wa y s. Th e o re t i c a ll y and

Pra c ti ca ll y T t d rea e . Vo l I R o lli n g St o c

. . k 12 mo ,
Ash e , S . W . E le c tric Ra ilwa y s. V o l II . . E n gin e e ri ng Pre li m i n a ri e s and

Dire c t Cu rre n t Sub — Sta ti o ns ra m o ,

y
Ele c tri c it : E xpe ri me n ta ll a n d Pra c t i ca ll y y I z mo ,

k
At i n s o n , A A Ele c tri c a l a n d Ma gn e ti c Ca lc ula ti o n s
. . 8 vo , .

k
At i n so n , J J Fri c ti o n o f Ai r in Mi n e s
. ( Sc i e n ce Se ri es No
. 1 6 mo , . .

k
At in so n , J J , a n d W illia m s , Jr , E H . Ga se s M e t wi th in Co a l Min e s
. . . . .

( Sc i e n c e Se ri e s No 1 6 mo , .

Atk inso n , P Th e Ele me n ts o f Ele c tric Ligh ti n g

. rz mo ,

Th e E le m e n ts of Dyn a m i c E le c tri c it y an d Ma gn e ti sm . ra mo ,
Po wer Tra n s mi tte d by Ele c tric ity . I e mo ,

Auc h i n c lo ss , W . S . L in k and Va l ve M o ti o n s Si mpli fi e d 8 vo ,

Ayrt o n , H . Th e El e c tri c Arc . . 8 vo ,

Ba c o n , F . W . Trea ti se on th e R ic h ards Stea m- E n gi n e In di ca t o r . 12 mo ,

Ba i le s, G . M . Mo de rn M i n in g Pra c t ic e . Fi ve V o lume s 8 vo , ea ch,

Ba ile y , R D . . Th e Bre we rs An a l st
’
y . . 8 vo ,
Ba k e r, A L . . Qu t a ern i o n s 8 vo ( In P re pa ra ti o n .
)
Ba k e r, Be n j . Pre ss ure of Ea rth wo r k .
(Sc ie n c e Se rie s NO . 1 6 mo ,

Ba k e r, 1 O .
( Sc ie n ce Se rie s No
. Le ve ll i n g . . mo ,
16

Ba k e r, M N Po ta ble Wa te r
.
( S.c i e n c e Se ri e s No . . 1 6 mo ,

Se we ra ge a n d Se wa ge Pu r i fi c a ti o n ( S c i e n c e Se rie s No . . 1 6 mo ,

Ba k e r, T T . . Te le gra phi c Tra n sm iss i o n of Ph o to gr a ph s I a mo ,

Bal e , G R M o dern Iro n Fo un dr Pra c t ic e

. . Two V o l u me s y . . I z mo .

Vo l I Fo u n dr. E q ui pm e n t , Ma te ria ls Use d

. y
Vo l 11 Ma c h i n e Mo uld i n g a n d M o u
. . ld in g Ma c h i n e s
Ba le , M P Pu m ps a n d Pumpin g
. . ra m o ,

Ba ll , R S . . Po pula r Gui de to th e H ea ve n s . 8 vo ,
Na tura l So urc e s of Po wer .
(We s tm i n ste r Se r i e s ) . 8 vo ,
Ba ll , W . V . La w Afi e c ti n g E n gi n e e rs . 8 vo ,
Ba n k so n , Ll o y d Sl ide Va l ve Dia gra ms
. .
( e n c e Se ri e s No
Sc i . . 16 mo ,
Barba , J . Use of Ste e l fo r Co n struc ti ve Purpo se s . I amo,

Ba r k A H
e r, Gra ph i c Me th o ds o f E n gin e De s ign
. . . . rz mo ,

k
Ba r e r, A Te xt ile s a n d Th e ir Ma n ua c tu
. re W
( e stm in ste r Se ri e s ) 8 vo , f . .

y
.

Barn ard , F A P Re po rt o n Ma c h in e r a n d Pro c e sse s Of t h e In dustria l

. . .

Arts and Appara tus of th e E xa c t Sc i e n c e s at th e Pa r i s Un i ve rsa l

E xpo s i ti o n , 1 86 7 . 8 vo ,
Ba rn a rd , J . H . Th e Na va l M il iti a m a n
’
s Guide 16 mo le a th er
Ba rna rd , Ma j or J
. G . R o ta r M o ti o n y .
( e n c e Se ri e s No
Sc i . 1 6 mo ,
4 D . V AN NOSTR AND COMPANY S SIIORT TITLE CATALOG ’

Barrus, G H . . B o ile r Te sts . 8 vo , *

3
E n gi n e Te sts 8 vo , *
4
Th e a bo ve tw o purc h a se d t o ge th er *
6
Ba rwi se , S . Th e Puri fi c a t i o n of Se wa ge 12 mo , 3 5°
B a te rde n , J . R . Ti mber .
( We stm i n ste r Se rie s ) . 8 vo , *2

k
. .

Be a dle , C . Ch a pte rs on Pa pe rma i n g. Fi ve V o lume s rz mo, ea ch ,

Bea um o n t, R . C o lo r in W o ve n De sign 8 vo , 7 50
Fi n ish i n g of Te xtil e Fa bric s 8 v0 ,
B e a umo n t, W W Th e Stea m E n gin e In d i c a t o r . .
-
8 vo , 2 50
Be de ll , F , a n d Pie rc e , C A Dire c t a n d Al te rn a ti n g Cu
. rre n t T . . eti s ng . . 8 vo ,
Be e c h , F . D y e ing o f Co tto n Fa bric s . 8 vo ,
D y e ing o f W o o le n Fa bri c s 8 vo ,
k
Be c w i th , A . Po tt e r y 8 110 , pa pe r,
B e gtrup, J . Th e Sl ide Va l ve 8 vo , .

Be n de r, C E . . Co n ti n uo us B ridge s .
( Sc ie n c e Se rie s No . 16 mo ,
Pro po rti o n s of Pi ers . use d in Bri dge s ( Sc ie n c e Se ri e s No
. .

16 mo ,
Be n n e tt , H G . . Th e Ma n ua c ture f of Le a th e r 8 vo ,
Bern th se n , A . A Te xt bo o k of Orga n i c Ch e mi stry . Tra n s . by G .

u
*2
50
Be rr y , W .
J . Di fi ere n tia Eq ti l
s o f t h e Firs t S e c i e s
p a on “ m o ( In Pre p
12 ara t io n .
)
Bersc h , J . f
Ma n ua c ture o f M i n eral a n d La e Pigme n ts k . Tra n s by A C . . .

. 8 vo , *
5 0 0

Bertin , L E Ma rin e B o ile rs

. Tra n s by L S R o be rts o n
. . . . . . 8 vo , 5
Be veri dge , J Pa pe rma e r s Po c e t B o o
’
k k k rz mo , *
4
‘
.

B i n n s, C F Ce ra m i c Te c h n o l o g
. . y . . 8 vo , 5 0 0

Ma n u
a l o f Pra c t i c a l Po tti n g . . 8 vo , *
7 50
Th e Po tte r ’
s Cra t f . . rz mo , *
2

Birc h m o re , W . H . H o w to Use a Ga s Ana l yi s s . . 1 z mo , *

1 2 5
B la i n e , R G . . Th e Ca lc ulus and Its Appl ica ti o n s . . 1 2 mo , *1
50
Bla k e, W . H . Bre wers Va de ’
Me c um . . 8 vo , *
4
Bla k e, W . P . R e po rt upo n th e Pre c i o us Me ta l s 8 vo , z

Bli gh , W . G . Th e Pra c ti c a l De s ign of Irriga ti o n Wo rks . 8 v0 , *

6 0 0

y
Bl th , A . W . Fo o ds : Th e ir C o m po si t i o n a n d Ana l yi s s . . 8 vo , 7 50
Po iso n s : Th e ir Eff e c ts a n d De te c ti o n . 8 vo , 7 50
'

k
Boc m a n n , F . 12 mo , *2
50
B o dm e r, G R . . H dra u y
lic M o t o rs and Turb i n e s 1 2 mo , 5
B o ilea u, J . T . Tra ve rse Ta bl e s 8 vo , 5
Bo n n e y , G E . . Th e E le c t ro — p la t e rs a o o
’
Hndb k 1 2 mo , 1 2 0

B o o th , W . H . Wa ter So f te n i n g a n d Tre a tme n t ” 8 vo , *2

50
Supe rh e a te rs and Supe rh e a t in g an d Th e ir C o tro l n . . 8 vo , *
1 so
B o ttc h e r, A . Cra n e s : Th e ir C o n struc t i o n , M e c h a n i ca l E q ui pme n t a n d
Wo rk i n g . Tra n s by A To lh a use n
. .
4 to ,
*
10

B o ttle r, M M o de rn . B le a c h i n g Age n ts Tra n s by C Sa l te r . 1 2 mo , . .

*2
50
B o tt o n e , S R Ma gn e to s fo r Auto m o b ili sts
. . Iz m o, *
1

Bo u l to n , S B Pre se rva t i o n o f Ti mbe r

. ( Sc i e n c e Se ri e s No 8 2
. . . . 1 6 mo , 0 50
Bo urgo ugn o n , A . Ph yi
s ca l Pr o ble m s .
( Sc i e n c e Se r i e s No . . 16 mo , o 50
Bo urr y, E . Tre a ti se on Ce ra m i c In d us tri e s . Tra n s by . W . P R ix
. . . 8 vo ,

Bo w , R H . A Tre a ti se o n ra c i n g
. B
Bo wi e , A J Jr A Pra c tic a l Tre a t i se
. . . on H dra u y
l ic M i n i n g
6 D V AN NOSTR AND C OM PA NY S SH O RT T IT LE C AT ALO G
’
.

Ca rte r, E T . . M o ti ve Po we r and Ge ari n g f o r E le c tr i c a l Ma c h i n e r y . 8 vo , *

5 00

Car te r, H A . . Ra m ie ( R h e a ) , C h i na Gra ss . 12 mo , *
2 00

Carte r, H R M o de rn Fla x, H e m p, a n d Ju
. . te Spi n n i n g 8 vo , .
*
3 00

Ca th ca rt , W L Ma c h i n e De s ign Pa rt I Fa ste n i n gs
. . 8 vo , . . . .
*
3 00

Ca th c art, W L , a n d Ch a fi e e , J I Ele me n ts o f Gra ph i c Sta tic s

. . 8 vo , . . . -
*
3 00

Ca ve n , R M , a n d La n de r, G D
. S ste ma ti c In o rga n i c C h e m i str 1 2 m o ,
. . . y y .
* 2
0 0

Ch a mbe rs Ma th e ma t ic a l Ta ble s
’
8 vo , . 1 75
Ch a rn o c , G F kW or sh o p Pra c ti c e ( We st mi n ste r
. . 8 vo ( In k . .

C h a rpe n ti er, P Ti m ber . 8 vo , *

6 0 0

C h a tle , H y Pri n c i ple s a n d De s ign s o f Ae ro pla n e s ( Sc i e n c e Se ri e s )

. . .

No 1 z 6 ) . 1 6 m0 ,

H o w to Use Wa te r o we r r z mo , P
Ch i ld , C T Th e H o w a n d W h y o f E le c tri c i t
. . 1 2 mo , y
C h ri stie , W W . . B o i le r wa te rs , Sca le , C o rro s i o n , Fo a mi n g
-
8 vo ,
Ch i mn e yD e s ign a n d Th e o r y 8 vo ,
Furna c e Dra t f .
( Sc i e n ce Se ri e s No . 16 m0 ,
Ch urc h
’
s La bo ra to r y GuidRe wri tte n by E d wa rd
e . Ki n ch 8 v0 ,
Cla ppe rto n , G . Pra c tic a l Pa pe rm a i n g k . . 8 vo ,
Cla r k C H Ma rin e Ga s ( In P res s )
f r Mc h a ni
. . .
,

Cla r k , D . K . Rule s, Ta ble s and Da ta o e ca l En gi n ee rs 8 vo , 5

Fue l : Its Co m busti o n and Ec o n o m y 12 mo , 1 50
Th e M e c h a n i ca l E n gi n e er s Poc ’
k e tb o o k . 1 6 mo , 2

Tra mwa s : Th e ir C o n stru c ti o n y and W o rki n g . 8 vo , 7 50

C lar k , J M . . Ne w S ytm s e La i n g Ou oft R a i l wa Tu y
rn o uts y . . . 12 mo , 1

Cla use n Th ue, -

W . A B C Te le gra ph i c C o de Fo urth E d i t i o n . 1 2 m0 , *
5 0 0

f
Fi th E d i ti o n . 8 vo , *
7 0 0

Th e A 1 Te le gra ph i c Co de . 8 vo , *
7 50
Cle e ma n n , T M Th e Ra ilro a d E n gi n e e r s Pra c ti c e
. .
’
. Ia mo , *1
50
Cl e r k , D .
, a n d Ide ll , F E Th e o r o f th e Ga s E n gin e . . y .
( Sc i e n c e Se ri e s

16 mo , 0 50
Cle ve n ge r, S R . . Tre a tise on th e Me t h o d of Go ve rn me n t Surve i n g y .

16 m o , m o ro c c o .

Cl o uth , F . Rubbe r, Gutta -

Pe rc h a , and Ba la ta . 8 v0 ,
C o ffi n , J . H C . . Na viga t io n and Na uti ca l Astro n o m y . 12 mo ,
C o l burn , Z .
, an d Th urst o n R . H . Ste a m B o ile r E xpl o si o n s .
( Sc ie n c e
Se ri e s No . 16 mo ,
C o le , R S . . Trea tise Ph o to gra ph i c Opti c s
on 12 mo ,
Co le s - Fi n c h , W . Wa te r, Its Origi n a n d Use . 8 vo ,
C o ll i n s , J . E . Use ful All y o s an d Me m o ra n da fo r Go ldsm it h s , J e w e le rs .

1 6 mo 0 50
Mi Th e i Q u 8
. .

Co n sta n tin e , E . ar n e E n gi n e e rs , r a l i fi c a t io n s a n d Dut ie s . . vo , *

2

C o o m bs , H A . . Ge a r Te e th .
( Sc i e n c e Se ri es . . 1 6 mo , 0 50
C o o pe r, W . R . Pri ma r yB a tte ri e s . 8 vo , *
4
Th e E le c tri c ia n Pri m e rs . . 8 vo , *
5
C 0 ppe rth wa i te , W C Tun n e l Sh ie lds
. . . .
4 to ,
*
9
C o re y , H T . . Wa te r Supply E n gi n e e ri n g . 8 vo ( In P re s s .
)
Co rfi e l d , W . H . Dw e ll i n g H o use s .
( Sc i e n c e Se ri e s No . 16 mo , o 50
Wa te r a n d W a te r Suppl -
y .
( Sc i e n c e Se ri e s No . . I 6 mo , o 50
C o rn wa ll , H BMa n ua l o f B l o w pi pe An a lys i s
. .
-
. 8 vo , *
2 50
Co u rtn e , C F y
Ma s o n r Da ms . . y . 8 vo , 3 50
C o we ll , W B Pure Air, Oz o n e , a n d Wa te r 1 2 mo , *
. . . 2
 D VAN NOSTRAND COMPANY 8 SH ORT TITLE CATALOG
.
’
7

C ra i g, T Mo ti o n o f a So li d in a Fu
. el ( Sc ie n c e Se rie s No . . 16 mo ,
Wa ve a n d Vorte x M o ti o n ( Sc i e n c e Serie s No . . 1 6 mo ,

C ra mp, W Co n tin uo u s Cu rre n t Ma c hi n e De sign

. . . 8 vo ,
C ro c k er, F B Ele c tri c Li gh ti n g
. Two o lu . mes 8 vo . V . .

Vo l I . . Th e Ge n e ra tin g Pla n t
Vo l II . Di stri bu
. tin g S st e ms an d La mps y
C r o c k er, F B , an d Are n dt, M
. Elec tri c Mo to rs
. . 8 vo ,
C r oc k er, F B , a n d Wh ee l er, S S
. . Th e Ma na ge me n t . . of El e c trica l Ma
ra mo ,
C ross, C F . .
, Be va n , E .
J .
, and Sin da ll, R . W . Wo od Pulp a n d Its Appli ca
ti o n s .
( W e s tmi n ste r Se ri es ) . . 8 vo ( In Press ) .

C r o ssk ey, L R . . E le me n ta r yP ers pe c ti ve 8 vo ,

C ro ssk e y, L R . .
, an d Tha w , Adva n c e d Pe rspe c ti ve
J . 8 vo ,
Culle y J , . L . Th e o r y o f Arc h es ( Sci e n ce Se rie s No
. . . 16 mo ,

Da ve n port, C . Th e B o o k .
( W e stm i n st er Serie s .
) 8 vo ,
Da vi e s, E H . . yf
Ma c h in e r or Me ta l li e ro us Mi n e s f 8 vo ,
Da vies , D C . . f
M e ta lli ero us M in era ls a n d Mi n in g 8 vo ,
Earth y Min e ra ls a n d Mi n i ng 8 vo ,
Da vie s, F H Elec tri c Po we r a n d Tra c t i o n
. . 8 vo ,
Da wso n , P Ele c tric Tra c ti o n o n R a ilwa s
. y 8 vo ,
Da y, C Th e In dica tor an d Its Dia gra ms
. 1 z mo ,

De err, N Su ga r a n.d th e Sug ar C a n e 8 vo ,

De i te , C Ma n u a l o f So a pma k in g
. Tra n s by S T . . . . Ki ng 4 to ,
De la Co u x, H Th e In du s tria l Use s o f W a te r . Tra n s by A M orris . . . . . 8 vo ,
De l Mar , W A E le c tri c Po we r Co n du . c tors . 8 vo ,
De n n , Gy A De ep e- l ve l M i n
. e s o f th e Ra n d
.
4 to ,
Dia m o n d Dri llin g fo r Go ld .

De R o o s, J D C Li n a ge s ( Sc i e n c e Se ri e s No
. . . 1 6 mo ,k . .

De Va ro n a , A Se we r Ga se s ( Sc i e n c e Se.ri es No 1 6 mo , . .

De rr, W . L . Bl o c k Sig Opera ti o n

na l Obl o n g 1z mo ,
B e sa in t, A . Th re e Hun dre d Sh a de s an d H o w to M i x Th e m 8 vo ,
y
.

Di b di n , W J Publ ic Ligh tin g by Ga s a n d

. . E le c tr i c it . . 8 vo ,
Puri fi ca ti o n o f Se w a ge a n d Wa ter . . 8 vo ,
Di e teri c h , K . An a l yi
s s of R e si n s , Ba lsa ms, and Gum Re sin s 8 vo ,
Di n ger, Li e ut H C Care a n d Opera ti o n o f Na va l Ma c hi n e r
. . . ra mo , y
Di xo n , D B Ma c hi n ist s a n d Stea m En gi n e e r s Pra c ti ca l Ca lc u
.
’
.
’
la to r .

16 mo , mo ro c c o ,
Do ble , W A . . Po we r Pla n t C o n struc ti o n on th e Pa c i fi c Co a st ( In P re ss .
)
Do dd, G . Dic ti o n ar y of f
Ma n ua c ture s, M i n i n g, Ma c h in er y , an d th e
In dustria l Ar ts . rz mo ,
Do rr, B F . . Th e Surve y or s
’
Guide and Poc k et Ta ble- b o o k .

mo , mo ro cc o , 16

Do wn , P B Ha n d Co ppe r W ire Ta ble

. . y
1 6 mo ,

Dra pe r, C H E le me n ta r Te xt bo o k o f Ligh t, H e a t a n d So u
. . nd 1 2 mo , y -
.

H ea t a n d th e Pri n c i ple s o f Th ermo d na m i c s 1 2 mo , -

y
Duc k wall , E . W . Ca n n i n g and Pre se rvi n g of Fo o d Pro d uc ts . 8 vo ,
Dumesn y, P , . an d No y e r, J . Wo o d Pro duc ts , Di st illa te s, and E xtra c ts .

8 vo ,
Dun ca n , W . G .
, a n d Pe n ma n , D . Th e Ele c trica l Eq uipme n t of Co ll ierie s .

8 v0 ,
8 D . VAN NOSTR AND COM PANY S S H O RT TITLE C AT ALO G ’

Duth ie , A L De c o ra ti ve Gla ss Pro c esse s

.
( Westmi n ste r
. . 8 vo ,
y
.

D so n , S 8 Pra c ti c a l Te st in g o f Ra w Ma te ria l s
. . 8 vo ,

E c c le s, R G . .
, and k
Duc wa ll , E . W . Fo o d Pre se rva ti ve s
Pa pe r
Edd y , H T . . R e se arc h e s i n Gra ph i c a l Sta tics
Ma ximum Stre sse s un de r Co n c e n tra te d Lo a ds .

E dgc u m be , K
In dustria l Elec tri ca l M ea su r i n g In stru
. m e n ts
E i ssl e r, M Th e Me ta ll u rg o f Go ld
. y .

Th e H dro me ta llu rg o f Co ppe r y y .

Th e Me ta llurg o f Sil ve r y
Th e Me ta ll urg o f Arge n t i ero u s Le a d y f
C y a n ide Proce ss fo r th e E xt ra c ti o n of Go ld
A Ha n d k on Mo de rn E xpl o si ves
E ki n , T C . . Wa te r Pipe an d Se wa ge Di sc h a rge Dia gra m s f o li o ,

Eli o t , C . W .
, and Sto re r, F H . C o mpe n di o u
. s Ma n u al of Q u lit
a a ti ve
Ch e mica l An a l yi
s s . 12 mo ,
Elli o t, Ma j or G H . Eu. ro pe a n Li gh t h o use S ste ms -
y 8 vo ,
En n i s, Wm . D . Li n se e d Oil a n d Ot h e r Se e d Oi ls 8 vo ,
Appli e d Th e rmo d na mi c s y 8 vo
Er uf rt , J D e i n g o f Pa pe r Pul p
. Tra n s by J H u y bn e r . . . 8 vo ,
Ersk i n e- M urra , J A Ha n dbo o o f W ire le ss Te le gra phy . k y . . . 8 vo ,
E va n s, C A Ma ca da m i zed R o a ds
. . .
( In P re s s .
)
Ewin g, A .
J . Ma gn e ti c In d uc t i o n in Iro n 8 vc ,

Fe irie , J . No te s on Le a d Ore s 12 mo ,
No tes on Po tt e r y m 1 2 mo ,

y J Vntila ti on l Mi
.

Fa irl e , W .
, a nd An dre , Ge o . . e of Co a n e s. ( Sc ie n c e
Se ri es No . 1 6mo,

Fa irwea th e r, W . C . Fo re ign and C o lo n ia l Pa te n t La ws 8 vo ,

Fa nn i n g, J . T . H dra uli c y and Wa te r suppl y -
E n gin ee ri n g . 8 vo ,
Fa uth , P . Th e Mo o n i n M o d ern As tro n o m y . Tra n s . by J M c c a
. be .

*
8 v0 , 2 00

Fa y, I . W . Th e C o al - t a r Co lo rs . . 8 vo ( I n P re s s ) .

Fe rn ba c h , R L . . Glue a nd Gel a ti n e 8 vo ,
Fi sc h er, E . Th e Pre pa ra ti o n of Orga n i c Co mpo un ds . Tra n s . by R . V .

Sta n f o rd ” 12 m0 ,
Fish , J . C L . . Le tte ri n g W o rk in g Dra win gs
of Oblo n g 8 vo ,
Fis h e r, H . K . C .
, and Da rby , W C Subma ri n e Ca ble
. . Te s t i n g 8 vo ,
Fi s k e, Lie ut B A . . . E le c tri c it yi n Th e o r y and Pra c ti ce . 8 vo ,
Fle i sc h ma n n , W . Th e B o o k of th e Da i y
r . Tra n s by C M Ai ma n. . . k . 8 vo ,
Fle m in g, J . A . Th e Alte rna te - c urre n t Tra n s f o rme r . Two V o lumes . 8 vo .

Vo l . I . Th e In d uc ti o n of E lec tri c C urre n ts

V o l II . . Th e Util i za t i o n of In d uc e d C urre n ts .

Ce n te n ar y of th e Ele c tri ca l C urre n t . 8 vo ,

E le c tric La m ps E le c tr i c Li gh ti n g and 8 vo ,
Ele c tr ic a l La bo ra to r No te s a n d Fo rm s y 4 to ,
k
A Ha n d bo o fo r th e Ele c tri c a l La bo ra to r a n d Te sti n g R o o m Tw o y .

V o lume s 8 vo , e a c h ,
s
. .

Fl ue ry, H Th e Ca lc ulus W i th o ut Li m i ts o r In fi n i te s i ma l
. Tra n s by . .

C 0 Ma i ll o ux
. . .
( In P re s s ) .
10 D . V AN NOST R AND COMPANY S SH ORT T IT LE C AT ALO G ’

Go uld , E S . . Ca lc ulus .
( Sc i e n c e Se ri e s No . 1 6 m0 ,

H igh Ma so n r yD a m s. ( Sc ie n c e Se ri e s No . 16 mo,
Pra c t ica l H dro sta ti c s y an d y
H dro sta ti c Fo rmula s .
( Sc i e n c e Se ries
No . 16 mo , o 50
Gra n t, J . Bre wi n g an d Di st ill in g ( We stmin ste r Se ri e s ) 8 v0 ( In Pres s )
. . .

Gra y , J . E le c tri ca l In flue n ce Ma c hi n e s 1 z mo , .

Gre e n w o o d , E . Cla ss ifi e d Guide to Te c h n i c a l and Co m m e rc i a l B o o k s. 8 vo ,

Gre go ri us, R M in era l Wa xe s Tra n s by C Sa lte r
. . . . . 12 mo ,
Gri fi ths , A B A Trea ti se o n Ma n ure s
. . . r z mo ,

De n tal M e ta ll urgy . . 8 vo ,
Gro ss , E . H o ps . 8 vo ,
Gro ssma n , J . Ammo n ia and Its Co mpo un ds . 12 mo ,
Gro th , L A . . We ldi n g a n d C utti n g Me ta ls by Ga se s or E lec tr ic it y . . 8 vo ,
Gro ve r, F . Mod e rn Ga s and Oil En gi n e s 8 vc ,
Grun er , A . Po we r l o o m -
Wea vi n g . 8 vo ,
Gttldn e r, H ugo . In te rna l Co mbusti o n E n gi n e s . Tra n s by H Die d e ric h s
. . .

4 t0 ,
Gun th e r, C 0 . . In te gra ti o n . 12 mo ,
Gurde n , R L . . Tra ve rse Ta bl e s . f o li o h a l f m o roc c o ,
Guy , A E . . E xpe rime n ts o n th e Fl e xure of B e a ms. . 8 vc ,

Ha e der, H . H a n db o o k on th e Stea m -
e n gi n e . Tra n s . by H . H . P.

Po wle s rz mo ,
Ha i n ba c h, R Po tte r De c o ra ti o n Tra n s by C Sla ter
. y . . . ra mo ,

Hal e , W J Ca l c u la ti o n s o f Ge n e ra l Ch e mi stry
. . 1 2 mo ,

Ha ll , C H Ch e mi str o f Pa i n ts a n d Pa i n t Ve h i c le s
. . y . r z mo ,

Ha ll , R H . . Go ver n o rs and Go ve rn i n g Mec h a n i sm . 1 2 m0 ,

Ha ll , W . 8 . Ele m e n ts of th e Diff e re n t ia l an d In te gra l Cal c ulus 8 vo ,

De sc ripti ve Ge o me tr y 8 v0 vo l ume and a 4 to a tl a s,

Ha ll e r, G F . .
, an d Cun n i n gh a m, E T . . Th e Te sla C o il rz mo ,
H a lse y , F A . . Sl ide Va l ve Ge ars . 1 2 mo ,

Th e Use of th e Sl ide R ule .

( Sc i e n c e Se ri e s No . 1 6 m0 ,

W orm a n d Spira l Ge a rin g .

( Sc ie n c e Se rie s No . 1 6 mo ,

Ha m il to n W G Use fu l In fo rma ti o n fo r Ra il wa y M e n
,
. . 1 6 mo ,

Ha m me r, W J Ra diu m a n d Oth er Ra d i o a c ti ve Substa n c e s

. .
-
8 vo ,
Ha n c o c k , H . Te xtbo o k fM h o ec a n ic s a n d y
H dro sta ti c s . 8 vo ,
Ha rd y ,
E . E le m e n ta r y P i iplr nc es of Gra ph i c Sta tic s . . 12 mo ,
H a rper , W B Util i za t io n o f Wo o d Wa ste by
. . Dis tilla ti o n .
4 to ,
H a rri so n W B Th e Me c h a n ic s To o l b o o k
, . .
’ -
12 mo ,
H art J W
,
E xt e rn a l Plu
. m bin g W o rk
. . . 8 vo ,
H in ts to Plu m be rs o n Jo in t W ipin g . . 8 vo ,
Pri n c i ple s o f H o t W a te r Suppl y . . 8 vo ,
Sa n ita r y Plumb i ng an d Dra i n a ge . 8 vo ,
k
Ha s i n s , C H . . Th e Ga l va n o m e te r an d Its Use s 1 6 mo ,

H a tt , J . A H . . Th e Co lo ri s t sq uare 1 2 mo ,

Ha usbra n d , E . y
Dr in g by M e a n s of Air and Ste a m . Tra n s . by A C . .

Wrigh t rz m o ,

E va po ra tin g, C o n de n s i n g and C o o l i n g Appa ra tus . Tra n s . by A C . .

8 vo ,
Ha us n e r, A Ma n ua c ture o f Pre s e r ve d Fo o d s
. f and Swee t me a ts . Tra n s .

by A M o rri s a n d H R o b s o n
. . . 8 vo ,
 D . V AN NOSTR AND COMPANY S SH OR T TIT LE C ATALOG ’
ll

H a w k e, W
Pre m ie r Ci ph e r Te le gra ph i c Co d e
. H .
4 to , .
*
5 00

Wo rds Supple me n t to th e Pre mie r Co de 4 to ,

*
5 00

H a w k e s w o rth , J Gra phi c a l H a n dbo o fo r R e i n o rc e d Co n c re te De sign

. k f .

t *
4 ,o 2 50
R a y, A . Al te rn a tin g Curre n ts . 8 v0 , *
2 50
Pr in c i ple s of Alte rn a te- c urre n t W ork i n g 12 mo , 2 00

E l e c tri ca l Dis tr ibuti n g Ne two r k s an d Distributin g Li n e s . . 8 vo , *

3 50
C o n ti n uous Curre n t E n gi n e e ri n g . . 8 vo , so
H e a p , Ma or D P j
Ele c tri c a l Appl ia n c e s . . . . 8 vo , 2 00

H e a v i s id e , O . Ele c tro m a gn e t ic Th e o r y . Two V o l ume s . 8 va , ea c h , *

5 00

Heck, R C H . . Ste a m E n gi n e
.
-
an d Oth e r Ste a m M o to rs . Two V o lumes .

V ol . I . Th e rm o d n a m i c s y an d th e Me c h a n ic s . . 8 vo , *
3 50
V o l. II . Form , Co n struc tio n , and W o rki n g 8 v0 ,
5 00
*

f Abridge d e di ti o n o f a bo ve vo l ume s ( Ele me n ta r ) y 8 vo ( In P re pa ra tio n ) .

No te s on Ele me n ta r y Ki ne ma t ic s 8 vo , bo ards , *1 0 0

Gra phi c s of Ma c h in e Fo rc e s 8 v0 , bo a rds , *1

0 0

H e d ge s , K
M o dern Ligh tn in g Co n duc to rs
. 8 vo , 3 00

y
.

H e e rma n n , P D e rs Ma te ria ls
’
Tra n s by A C Wrigh t
. . . . . 12 m0 , *2
50
B e l l o t , Ma c q uer a n d D Apl ign y
’
Art o f D e in g W o o l , Sil . y k and Co tt o n .

*2
8 vo , 00

H e n ri c i , O . S k e le to n Struc ture s . 8 vo , 1 50
H e rma n n , F . Pa in ti n g on Gla ss and . 8 v0 , *
3 50
H e rrma n n , G . Th e Gra phi c a l Sta ti c s of Me c h a i s m n . Tra ns . by A P . .

Smi th . 12 m0 , 2 00

H e rz f e ld , J . Te s ti n g of Ya rn s an d Te xt ile Fa bri c s 8 v0 , *
3 50
H ilde bra n dt , A Airshi ps, Pa st a n d . 8 vo , *
3 50
Si sN
.

H il de n bra n d , B W Ca ble Ma i n g . .
-
k .
( Sc ie n c e er e o. 16 mo , 0 50
H il l , JW . . Th e Puri fi ca ti o n of Publi c Wa ter Suppl ie s . Ne w E di ti o n .
( In Pre s s )
Wa te r Ana lysis
.

In te rpre ta ti o n of ( In I
’
rc xx .

No
. .

H iro i, I . Pla te Girde r C o n struc ti o n .

( Sc ie n c e Se rie s .
95 . 16 mo ,
Sta ti ca ll yI -
n de term i n a te Stre sse s 1 2 mo ,

H irshfe ld, C F . . En gi n e e ri n g Th e rm o d y n a mi c s. ( Sc i e n c e Se rie s No .

16 mo , 0 50
Ho ba rt, H M . . y
H ea v E le c tri c a l En gi n e e ri n g . 8 vo , *
4 50
Ele c tric it y . 8 vo , *
2 00

Ele c tric Tra i n s . 8 vo , *2

50
Hobbs, W . R P . . Th e Ar i th me i c t of El e c tr i ca l M e a s ure me n ts 12 mo , o 50
Ho fi , J . N . Pa in t and Varn ish Fa c ts an d Fo rm ula s ” 1 2 111 0 ,
*
1 50
110 3 , Co m . W . B . Th e Avo ida n c e of C o lli s i o n s at Se a . 16 m0 , m o ro c c o , 0 75
Ho le , W . Th e Distri buti o n of Ga s 8 v0 , *
7 so
Holle y ,
A L . . R a ilwa yP ra c ti c e f o li o , 1 2 00

Ho lmes, A B . . Th e E le c tri c Ligh t Po pula rl E xpla i n e d y . . 12 m o , pa pe r, 0 50

Hopkins, N M . . E xpe ri me n ta l Ele c tr o c h e m istr y . 8 vo , *
3 00

Mo de l E n gi n e s and Sma ll B o a ts . 1 2 111 0 , 1 2 5

Ho pkin so n , J . Sh o o lbre d, J . N .
, and Da y, R E . . D y n a m ic E le c tri c i t y .

( Sc ie n c e Se rie s No . 1 6 mo , 0 50
Homer, J . E n gin e e rs ’
Turn in g 8 vo , *
3 50
Me ta l 12 m0 , 1 50
To o th e d Ge a ri n g 1 2 m0 , 2 2 5
Hou
gh to n , C E . . Th e E le me n ts of Me c h a n i c s of Ma te ria ls 1 2 mo , *2 00

Hou
lleviq ue, L . Th e E vo lutio n of th e Sc ie n c e s 8 vo , *2
00
12 D V AN NOST R AND C OM PA NY S SH ORT T ITLE C ATALOG
’
.

H o we , G . Ma th e ma tic s fo r th e Pra c tica l Ma n 12 mo ,

H o w o rth , J . R e pa iri n g and R i ve tin g Gla ss , Oh 1n a and Ea rth e n wa re .

8 vc , pa pe r,
H ubbard, E . Th e Uti li za t i o n of Wo o d -
wa ste . 8 vo , .

H umber, W . Ca lc ula ti o n of Stra i n s 1n Gird e rs . 12 m0 ,

Humph re y s, A C . . Th e Bus in e ss Fe a ture s of E n gi n e e ri n g P ra c t i c e . 8 vo ,
Hurs t, G H . . Ha n d b o o k o f th e Th e o r y of C o lo r . 8 vo ,
Di c ti o n a r y of Ch e m ic a ls and Ra w Pro duc ts 8 v0 ,
Lubri c a ti n g Oi l s , Fa ts and Gre a se s 8 vo ,
So a ps 8 v0 ,
Te xt i le So a ps and Oi l s 8 v0 ,
Hurs t, H E . .
, and La tte y , R T . . Te xt b o o -
k of Ph yi s cs . 8 vo ,
Hutc h in so n R W Jr
. . . Lo n g Dista n c e E le c tr i c Po w e r Tra n s m i ss i o n 12 mo ,
Hutc h in so n , R W Jr . .
, .
, and Ihl se n g, M C . . E l e c tr ic i t yi n Min i n g 12 mo ,
( In P re s s )
H utc h i n so n , W . B . Pa te n ts a nd H o w to M a k e Mo n e y Out o f Th e m .

1 2 mo ,
H utto n , W . S . Ste a m b o ile r C o n struc ti o n
-
. 8 vo ,
Pra c t i c a l E n gi n e e r s H a n d bo o
’
k 8 v0 ,
Th e W o r s Ma n a ge r s H a n d bo o
’ ’
k k . 8v
o,

y
H de , E . W . S k ew Arc h e s .
( Sc i e n c e Se ri e s No . . 16 m0 ,

In d uc t i o n C o i ls .
( Sc i e n c e Se r i e s N0 .

In gle , H . Ma n ua l of Agric ult ura l C h e m i s tr y

In n e s , C H . . Pro ble ms i n Ma c h in e De s ig n
Air C o m pre ss o rs and B l o w i n g E n gi n e s
Ce n tr i fug al Pumps
Th e Fa n
Is h e rwo o d , B F . . E n gi n e e ri n g Pre c e d e n ts f o r Ste a m M a c h i n e r y
Iva tts , E B . . R a ilwa yM a na ge me n t a t Sta ti o n s .

Ja c o b, A, . and Go uld , E . S . On t h e De sign i n g a nd C o n s tr uc t i o n of

St o ra ge R e se rvo irs .
( Sc i e n c e Se r i e s No . . 1 6 mo ,

Ja m ie so n ,
A . Te xt B o o k on Ste a m Ste a m E n gi n e s
and 8 v0 ,
E le m e n ta r yM an ua l on St e a m a n d th e Ste a m E n gi n e 1 2 mo ,
Ja n n e tta z , E . Guide t o th e De te rm i n a ti o n of Ro c k s . Tra n s . by G . W .

Pl m pto n y 12 mo ,
Jeh l, F . Ma n ua c ture o f Ca rb o n s f 8 v0

Je n n i n g ,
s A S Co m m e rc ia l Pa i n ts
. . and Pa i n t i n g . W
( e s t m i n s te r Se rie s ) .

8 v0 ( I n

Je n n iso n , F H . . Th e Ma n ua c ture f of La k e Pig m e n ts 8 vo ,

Ca m s and t h e Pr i n c i pl e s o f t h e i r C o n s truc ti o n 8 vo ,
Je ps o n , G . .

Me c h a n ic a l Dra w in g 8 vo ( In ) N IIYII IN I I .
)

Jo c k i n , W A r i th m e t i c
. o f th
'

e Go ld a nd Si l ve rs m i th rz mo ,

J o h n so n , G L . . Ph o to gra ph ic Opti c s a n d C o l o r Ph o to gra ph y 8 vo ,

Jo h n s o n W H , . . Th e C ul t i va ti o n a n d Pre pa ra t i o n o f Pa ra R ub b e r . 8 vo ,

Jo h n so n W M o
,
A . . Th e M e ta ll urgy of Ni c k e l p u
( In I
m l u
’
m )
ri .

Jo h n s t o n , J F W . . .
, an d Ca m e ro n , C . E le m e n t s o f Agri c ul tura l C h e m i s tr y
a nd Ge o l o gy 12 mo,

J ,J
ol y . Ra id o a c ti vi ty and Ge o l o g y 12 m0 ,
Jo n e s , H C . . E le c tri c a l Na t ure of M a t te r an d R a d i o a c t i vi t y 12 mo ,
14 D . VAN NOSTR AND COM PANY S SH ORT TITLE CATALOG ’

K o e st er, F . Stea m Ele c tr i c Po we r Pla n ts

-
4 to ,
$
5 00

y
H dro e le c tri c De ve l o pme n ts a n d E n gin e e rin g 4 to ,
*
s
K o ll e r, T Th e Utili za ti o n o f Wa ste Pro d uc ts
. . 8 vo ,
Co sm e ti cs . 8 v0 ,
K ra uc h , C . Te st in g of Ch e m i c a l R ea ge n ts . Tran s by .
J . A
. W ill i a mson
an d L . W . Dupre . 8 v0 ,

La mbert , T . Le a d an d i ts Co mpo un d s 8v0 ,

B o n e Pro duc ts an d Ma n ure s 8 v0 ,
La mbo rn , L L . . C o tto n se e d Pro d uc ts 8 vo ,
Mo dern So a ps , Ca n dle s , and Gl y c e ri n 8 vo ,
La mpre c h t, R . R e c o ve r y W k Aft or er Pit Fi re s Tra n s by C Sal te r
. . . . . 8 vo ,
La n c h es ter, F W Ae ria l Fli h t Two V o lumes 8 v0
g
. . . . .

V ol . I . Aero dyn a mic s ”

Vo l H . . Ae 1 0 d 0 n e ti c s
Larn er , E T . . Prin c iple s of Alte rn a ti n g Cu rre n ts 12 m0 ,
La rra bee , C S . . Ciph e r an d Se cre t Le t ter and Te le gra ph i c C o de . . . 1 6 mo ,

La R ue, B F . . Swi n g Bridge s .

( Sc i e n c e Se ri e s N0 . . 1 6 mo ,

La ssar - Co h n , Dr . Mo de rn Sc i e n tifi c Ch e m i str y . Tra n s by M M Pa tti. . .

so n Muir . . 12 m0 ,
La ti mer, L H . .
,
Fi e ld , C .
J . an d H o w e ll , J W . . I can n de s c e n t E le c tric
Ligh tin g ( Sc i
. e n ce Se r i e s N 0 57 . 16 mo ,
g
.

ngP
.

La tt a , M N . . Ha n d b o o o f Am e r i c a n kGas- En in eeri ra c t i c e 8 v0 ,
Ameri c a n Pro duc e r Ga s Pra c tic e .
4 to ,
Le a s k , A R . . Brea d o wn s k at Se a 1 2 mo ,

Tri ple an d Qu d upl a r e E xpa n s i on . 12 mo ,

Re f igr e ra ti n g Ma c h in er y . 1 2 m0 ,

k
Lec y , S T S Wrin l e s in Pra c ti c a l Na viga ti o n
. . . k 8 vo ,
Le Do ux, M Ic e- Ma i n g Ma c h i n es ( Sc
. i e n c e Se ri e s No 4 6 k . . . . 16 m0 ,
Le e d s , C C Me c h a n i cal Dra win g fo r Tra de Sc h o o ls
. . o bl o n g 4 to ,
H igh Sc h o o l Ed i ti o n
Ma c h in er Tra de s Edi ti o n y . .

f
Le evre , L . Arc hi te c tural Po tte r y . Tra n s . by H . K . B ird and W M . .

. .
4 to ,
Le h n e r, S . f a c tu
In k Ma n u re . Tra n s by A M o rris a n d H R o bs o n
. . . . . 8 vo ,
Le mstro m , S . Elec tric it yi n Agr ic ul ture a n d H o rti c ul ture 8 vo ,
Le Va n , W . B . Ste a m E n gi n e In di c a to r
-
.
( Sc i e n c e Se rie s No . 1 6 mo ,

Le wes , V B . . Li q uid an d Ga seo u s Fu e ls .

( We stmin ste r Se ri e s ) . . 8 vo ,
Li e be r, B F . . Li e be r ’
s Sta n dard Te l e gra ph ic C ode . 8 vo ,
C ode . Ge rma n E d i ti o n 8 vo ,
Spa n i sh Ed i ti o n 8 v0 ,
Fre n c h E di ti o n 8 v0 ,
Term i n a l In de x 8 v0 ,
Li e be r ’
e Appe n di x f o li o ,

Ha n d yT a bl e s t
4 ,o

Ba n k e rs a n d Sto c bro k k e rs
’
Co d e and Merc h a n ts and Sh ippe rs Bla n ’
k
Ta bl e s 8 v0 ,
C o m bi n a ti o n C o de 8 vo ,
E n gin e e rin g C o de 8 vo ,
Li ve rm ore , V P .
, and W ill i a ms J , . H o w to B e c o m e a Co m pe te n t M o to r
. rz mo ,
 D . VAN NOSTR AND COMPANY S SH ORT TITLE CATALOG ’
15

Li vin gsto n e , R . De s ign an d Co ns truc ti o n of Co mm uta tors 8 vo,

Lo bbe n , P . Ma c hin i sts ’
and Dra ftsme n ’
s Ha n db o o k 8 vo,
Lo c k e, A G . . an d C G
. . Ma n uf a c ture o f Sulph uri c Ac id 8 vo , .

Loc kwoo d, T . D . Elec tric it , Ma gn e tism, yan d Ele c tro—te le gra ph .

8 vo ,
Ele ctri c a l Me a sure me n t an d th e Gal va n o m eter 12 m0 ,
Lo dge , 0 .
J . Ele m e n tary Mec h a n i c s 1 2 m0 ,

Sign a ll in g Ac ro ss Spa c e w ith o ut W ire s 8 vo ,

Lo rd, R T De c o ra tive an d Fa n c Fa bric s
. . y 8 vo ,
Lo rin g, A E . . A Ha n dbo o k o f th e Ele c tro ma gn e ti c Te le gra ph . . 16 mo ,
Ha n dbo o k .
( Sc i e n c e Se ri e s No . . 1 6 mo ,

Lo w e n ste in , L C . .
, an d Cri sse y , C P . . Ce n tri fug l Pump a s . .
( In Press ) .

Luc k e, C E . . Gas E ngi n e De sign . 8 v0 ,

Power Pla n ts : th e ir De sign , Efi c i e n c y, an d Po wer C o sts . 2 vo ls .

( In P re pa ra ti o n .
)
Po wer Pla n t Pa pe rs Fo rm I Th e Stea m Po we r Pla n t . .
pa pe r,

Lun ge , G Coa tar a n d Amm o n i a

l -
. Two V o lu me s . 8 vo ,
Ma n ufa c ture o f Su lph ur ic Ac i d a n d Alk a l i . Th re e V o lumes . . 8 vo ,
Vo l I Sulph u
. ric Ac i d In two parts
. .
*1
5 00

Vo l . II . Sa lt Ca k e, y
H dro c hl o ric Ac id an d Le bla n c So
’

da In two

Vo l III .Ammo n ia .

Tec h n ica l Ch e m ists Ha n dbo o k

’
1 2 m 0 , lea th e r,

Te c h n ica l Me th o ds o f Ch e m ica l Ana l s is Tra n s by C A Kea n e y . . . . .

in c o lla bora ti o n wi th th e c o rps o f spe c ia lists .

Vo l I . In tw o . 8 vo ,
V o ls II . an d III ( In Pre pa ra ti o n ) .

Lupto n , A , Parr , G . . D A, . . an d Per in , H k . E lec tri c i t y as Appl ie d to

M in i n g . 8 v0 ,
Luq uet , L M . M in e ra ls in R o c. kS e c ti o n s . 8 vo ,

Ma ce we n , H A Fo o d In spe c ti o n . .

k
Ma c e n z ie , N F No te s o n Irriga ti o n . .

k
Ma c ie , J H o w to Ma e a Woo le n M ill Pa y
. k
Ma c kro w, C Na va l Arc hi tec t s a n d Sh i pbuilder s Poc k e t bo o k
.
’ ’ -
.

1 6 m 0 , lea th er,

Ma gu ire , Ca pt E Th e Atta c a n d De e n se o f Coa st Fo rti fi ca ti o n s

. . 8 vo , k f . .

Ma guire , Wm R Do m e st ic Sa n ita r Dra ina ge a n d Pl u

. . m bin g 8 vo , y . .

Ma lle t, A Co m po u . n d E n gin e s Tra n s by R R B u el ( S c i e n c e S e ri e s . . . . .

No . 1 6mo ,

Ma n sfi e ld, A N Elec tro-ma gn e ts .

( Sc i e n ce
. Ser i es No 1 6 m 0 , . .

k
Mar s , E C R . Co n stru c ti o n o f Cra n e s a n d Li fti n g Ma c hi n e ry
. . 1 2 m0 , .

C o n stru c ti o n a n d Workin g o f Pu mps 1 2 m0 ,

Ma n ua c tu f
re o f Iro n a n d Ste e l Tu be s 1 2 m0 ,

M e c h a n i c a l E n gi n e e ri n g Ma te ria ls 1 2 m0 ,

k
Ma r s , G C H dra uli c Po we r E n gin e eri n g
. . y 8 vo ,
In ve n ti o n s, Pa te n ts a n d De sign s 1 2 m0 ,

k
M ar h a m , E R Th e Am e ri ca n Ste e l W ork er
. . 1 2 m0 ,

M arl o w, T G Dr in g Ma c h in e r a n d Pra c ti ce
. . y y 8 vo ,
Ma rsh , C F C o n c i se Tre a ti se o n Re i n o rc e d C o n cre te
. . f . . 8 vo ,
Ma rsh , C F , a n d Dun n , W R e i n orc e d Co n cre te
. . . f 4 to ,
16 D VAN NOSTR AND COMPANY S SH ORT TITLE CATAL OG
.
’

Ma rsh , C F . .
, and Dunn, W . Ma n ual of R e in f o rc e d Co n cre te an d Co n
c re te Bl o c k Co n str uc ti o n . 1 6m0 , m o roc c o , *2
50
Ma s s i e , W . W .
, and Un derh ill , C R . . W ire le ss Te le gra ph y and Te l e ph o n y .

12 m0 ,
M a th e so n , D . Austra lia n Sa w M i ll e r -
’
s Lo g and Ti mber R ea d yR k ec o n e r.

12 m 0 , l e a t h e r,
Ma th o t , R E In tern al C o mbusti o n E n gi n e s
. . . 8 vo ,
Ma uric e , W Elec tri c B la stin g Appa ra tus a n d Expl o si ve s
. . 8 vo ,
Sh o t Firer ’
s Guide . 8 v0 ,
Ma xw e ll , J C Ma tt e r an d M o ti o n
.
( Sc i e n ce Se
. ri e s No . . 16 mo ,
Ma xwe ll , W H , a n d Bro wn , J T En c c l o pe d ia o f Mun i c i pa l
. . . . y an d Sa n i
ta r yE n gi n e eri n g 4 to ,
y
Ma er, A M Le c tu re No tes o n Ph si c s
. . y 8 vo ,
M cc ullo ugh , R S Me c ha n i ca l Th e o r o f H e a t . . y 8 vo ,
Mc In to sh ,J G Tec h n o l o g o f Sugar
. . y . 8 vo ,
I dut i l Al
n s r a c oh ol . 8 vo ,
M uf tu
an ac re o f Va rn i sh e s and Ki n dre d In d ustri e s . Th re e V o lume s .

8 vo .

Vo l . I . Oi l Crush i n g, Re fin i n g and B o il i n g
Vo l . II . Varn ish Ma terial s and Oil Va rn ish Ma i n g k
Vo l H I ( In Pre pa ra t io n )
D
. . .

M c Kn igh t , J .
, an d Bro wn , A . W . Ma rin e M ul ti tubula r B o ile rs
M c Ma s te r, B J . . Bri dge an d Tun n e l Ce n tre s .
( Sc i e n c e Se r i e s No . 20 .

16 m0 ,
Mc Mec h e n , F L Te sts fo r Ore s, Mi n e ra ls a n d Me tal s
. . 1 2 m0 ,

Mc Ne ill , B Mc Ne ill s Co d e
’
. 8 vo , .

Mc Ph e rso n , J A Wa te r wo r s Dis trib uti o n . . 8 vo , -

k
k
M e li c , C W Da ir La bo ra to r Guid e
. . y
1 2 m0 , y
Me rc , Ek Ch e m i ca l R e a ge n ts ; Th e ir Pu
. ri t a n d Te sts 8 v0 , y
Me rritt , Wm H Fi e ld Te sti n g f o r Go l d a n d Si l ver
. . 1 6 m 0 , l e a th e r, . .

y
Me e r, J G A , a n d Pe c e r, C G Me c h a n i c al Dra wi n g a n d Ma c h in e
. . . k . .

De sign t
4 ,o

M ic h e ll , S M i n e Dra i n a ge
. . 8 v0 ,
Mie rz in sk i , S Wa te rpro o fi n g . of Fabi r cs . Tra n s . by A M orris . an d H .

R ObSOll . 8 V O,
M ill er, E H ua n tita ti ve An a l si s fo r Mi n in g E gi n ee rs
. . Q y n 8 vo ,
Mill er, G A De te rm i n a n ts
.
( Sc ie n c e Se rie s
. . . . 16 mo ,
y
M il ro , M E W H o m e La c e ma k i n g
. . .
-
1 2 m0 ,

M in i fi e , W Me c h a n i ca l Dra win g
. 8 vo ,
Mi tc h e ll , C A , a n d Pri de a ux, R M
. Fibre s Use d in Te xt ile
. . . an d Allie d
In d us tr ie s . . 8 vo ,
M o de rn M e te o ro l o g y 1 2 m0 ,

k
Mo n c to n , C C F Ra d i o te l e gra ph
.
( W e stm in ster . . 8 vo , y .

M o n te ve rde , R D Ve st Po c e t Gl o ssa r o f E n gl i sh Spa n is h , Spa n ish

. . k y -

En gli sh Te c h n i ca l Terms 6 4 m o , lea th e r, *

1 00

M o ore , E C S Ne w Ta ble s fo r th e Co mple te So luti o n o f Ga n guille t a n d

. . .

’
K
utte r s Fo rmula 8 v0 ,
M o re in g, C A , a n d Ne a l , T Ne w Ge n e ra l a n d M i n i n g Te le gra ph Co de , 8 vo ,
. . .

Mo rga n , A P W ire le ss Te le gra ph Appa ra tus fo r Ama te urs

. . 1 2 m0 ,

M o se s , A J Th e Ch a ra c te rs o f Cr s ta ls
. . 8 v0 , y
Mo se s, A J , a n d Parso n s, C L Ele me n ts o f M in e ra l o g
. . 8 vo , . . y
 18 D . V AN NOSTR AND COM PANY S S H O R T T IT LE C AT ALO G ’

Pe rrigo , O E . . Ch a n ge Ge a r De vic es . 8y o , 1 00

Pe rri n e , F A C . . . C o n duc t o rs f o r Ele c tr i c a l Di s tr i buti o n . . 8 vo , *

3 50
Pe ti t, G . W h i te Lea d and Zi n c W h i te Pa i n ts . 8 vo , *
1 50
Pe ti t, R . H o w to B uild an Ae ro pl a n e . Tra n s by T O B H ubba rd ,
. .
’
. an d

J . H Le d e b o e r. 8 v0 , *1
50
Pe tti t , Li e ut J 8 Gra ph i c Pro c e ss e s
. . . .
( Sc i e n c e Se ri es N o .
76 . . 16 m0 , 0 50
Pe rr y , J . Appl i e d M e c h a n i c s .
’

8 vo , *2
50
k ( Sc i e c e Se rie s No
n
.

Ph ilbri c , P H . . Bea ms and Gi rd e rs . .

Ph i ll i ps , J En gi n e e ri n g Ch e mi s tr
. y . . 8 vo , *
4 50
Go ld Assa i n g y . 8 vo , *
2 50
Da n ge ro us Go o ds . . 8 vo , 3 50
Phi n , J
. Se ve n Fo ll i e s of Sc i e n c e . 1 2 mo , *
1 2 5
Ho use h o l d Pe st s , and H o w t o Ge t R i d of Th e m 8 v0 ( In P re pa ra ti o n ) .

k
Pi c w o rth , C N . . Th e In d ic a to r Ha n d bo o k . Tw o V o l ume s . 12 m0 , ea c h , 1 50
Lo gar ith m s for B e gi n n e rs 1 2 m 0 , b o a r ds , 0 50
Th e Slide R u le . 1 2 mo ,
. 1 00

Pla n e Ta ble , Th e . 8 v0 , 2 00

Pla ttn e r ’
s Ma n ua l o f Bl o w pi pe An a l -
yi s s . E igh th E di ti o n , re vi se d . Tra n s .

by H B C o rn wa ll . . . 8 vo , *
4 00

y
Pl mpto n , G W Th e An e ro i d Ba ro m e te r . . .
( Sc ie n c e Se rie s No .
35 . 1 6 mo, 0 50
H o w to b e c o m e an E n gi n e e r .
( Sc i e n c e Se ri es . 16 mo , o 50
Va n No stra n d ’
s Ta bl e Bo o k .
( Sc ie n c e Se ri e s No . . 1 6 m0 , 0 50
Poc h e t , M L . . Ste a m In j e c t o rs. Tra n sla te d f ro m th e Fre n c h .
( Sc ie n ce
Se rie s No . 1 6 mo , 0 50
Po c k et Lo gar i th ms to Fo u Pla c e s r .
( Sc i e n c e Se r i e s No 65 . . 0 50
l e a th e r, 1

P0 pe , F L M o de rn Pra c ti c e o f th e El e c tri c Te le gra ph

. . 8 vc , 1 50
Po pple we ll, W C Ele m e n ta r Tre a ti se o n H e a t a n d H e a t E n gin e s
. . y . 1 2 mo , *
3 00

Pre ve n ti o n o f Smo e k . 8 vo , *
3 50
Stre n gth o f Ma te ria l s . 8 vo , *1
75
*
Po tte r, T . C o n cre te 8 vc , 3 00

Pra c ti c a l C o m po un di n g of Oi ls , Ta ll o w a nd Gre a se . 8 vo , *
3 50
Pra c ti ca l Iro n Fo un di n g . 1 2 m0 , 1 50
Pra y ,
T .
, J r. Twe n t yY e a rs w i th th e In d i c a to r . 8 v0 , 2 50
Stea m Ta ble s and E n gin e C o n sta n t . . 8 v0 , 2 00

Ca l o ri m e te r Ta b le s 8 vo , 1 oo

(I
.

Pree c e , W . H . E le c tr ic La mps n P re s s .
)
Pre l i a i , C . Earth and Roc kE xc a va t i o n 8 vc , *
3 00

Gra ph ic al De te rm i n a ti o n of E a rt h Sl o pe s . 8 vc , *2 00

Tun n e li n g .
8 V0 , 3
Dre dgi n g . A Pra c tic a l Trea ti se .
( In P rrs s .

Pre sc o tt , A B Orga n ic An a l s is y 8 v0 , 5
Q u lit ti vCh m i l Analy i
. .

*
Pre sc o tt , A B . .
, and J o h n so n , 0 C . . a a e e ca s s . . 8 v0 , 3 50
Pre sc o tt , A B . .
, and Sull i va n , E C . . Fi t B k i Q u li t t i
rs Ch oo n a a ve e m i s tr y .

12 m0 , *1
50
Pritc h a rd , O G . Th e M a n ua c ture f E le c tri c li gh t Carb o n s 8 vo , pa pe r,
of - *
0 60
y
.

Pro st , E . C h e m i c a l An a l s is o f Fue l s , Ore s , M e ta l s Tra n s by J C . . . .

Sm i th 8 v0 ,
to
.

Pulle n , W W . . F . Appli c a t i o n of Gra ph ic M e th o ds th e De s ign of

Str uc ture s 1 2 m0 , *
2 50
In j e c t o rs : Th e o r y ,
C o n s truc ti o n and W o rk i n g . . 12 mo , *1
50
 D . VAN NOSTR AND COMPANY S SH ORT T ITLE C ATALOG ’
19

Pu W H
l s i fe r , . . No te s fo r a H i sto r y of Le a d
Purc h a se W R , . . Ma so nr y
Putsc h , A . Ga s and C o a l dus t Firi n g
-

P y n c h on , T R . . In tro duc ti o n t o Ch e m i c a l Ph yi s cs

f
Ra ter G . W . Mec ha n i c s of Ve n tila ti o n .
( Sc i e n c e Se ri es No .

Po ta ble W a te r .
( Sc i e n c e Se rie s No . 16 m0 ,
Tre a tmw t of Se pti c Se wa ge .
( Sc i e n c e Se r i e s No . 1 6 m0 ,

f
R a te r, G , . W Se wa ge D
.i spo sa l i
and
n th e Un i te Ba
d S ta te s 4 tko ,e r, M N . . .

ik
R a e s, H P Se wa ge Dispo sal Wor s
. . 8 vc , k
y
Ra il wa Sh o p Up- to- Da te 4 to ,
Ra mp, H M Fo un dr
. Pra c ti c e .
( In P res s ) y .

Ra n da ll , P M . . Qu t
Opera to r s Ha n dbo o k
ar z
’
. 12 m0 ,
R a n da u, P . En a me ls a n d E n a me ll i n g . 8 vc ,
k
Ra n i n e, W JM . . . Appl ied Me c ha n i c s 8 v0 ,
Ci vil En gin e e rin g 8 vc ,
Ma c h in e r a n d Mill wo rk y 8 v0 ,
Ra n k i n e , W JM . . . Th e Ste a m-e n gi n e and Ot h e r Pri me M o ve rs . . 8 vo ,
Use ful Rul es a n d Ta ble s 8 v0 ,
R a n k in e , W J M . . .
, an d Ba mber, E F . . A Me c h a n i ca l Te xt bo o -
k . . . 8 vo ,
Ra ph a e l , F C . . Loc a l i za ti o n of Fa ul ts i n Elec tri c Ligh t a n d Po we r Ma in s .

8 vo ,
Ra th bo n e , R L B . . .
Je we ll e ry
Si mple 8 vc ,
Ra te a u, A . Flo w o f Stea m th ro ug h No z z l es an d Orifi c e s . Tra n s by H . .

B Br do n . y . . 8 vc ,
Ra use n be rge r, F . Th e Theory o f th e Re c o il o f Gu s n . 8 vo ,
Ra ute n stra uch , W . No te s on th e E le m e n ts Ma c h i n e De s ign , 8 vo , bo a rds ,
of

Ra ute n s tra uch , W .

, an d Will ia ms J , . T . M a c hi n e Dra ft in g a n d E mpiric al
De sign .

Pa rt I Ma c h i n e Dra ti n g
. f . 8 v0 ,
Part II E mpirica l Des ign .
( In P repa ra t i o n .
)
y
.

Ra m o n d, E B Al te rn a ti ng Cu
rre n t E n gi n ee ri n g
. . . 12 m0 ,
Ra y n e r, H . Sil k Th ro wi n g a n d W a ste Sil k 8 vo ,
ysalt y T
.

R e c ipe s fo r th e C o l o r, Pa in t , Varn i sh Oi l , Soa p a n d Dr er ra d es 8 v0 ,

Re c i pe s fo r Fl in t Gla ss Ma i n g k 1 e mo ,
Re dwo o d , B Pe tro l e um ( Sc ie n c e Se ri e s N0 9 2
. . . 1 6 m0 ,

R e e d s E n gi n ee rs Ha n dbo o
’ ’
k 8 v0 ,
Ke y to th e Nin e tee n th E d iti o n of R ee d ’
s E n gin ee rs Ha n dbo o k
’
8 vo ,
ful Hi
. .

Use n ts to Se a - go i n g En gin e e rs . 12 m0 ,
Ma ri n e Bo ile rs 1 2 m0 ,

R e i n h a rdt , C .
. W . Le tte rin g fo r Dra f ts me n , E n gin ee rs , and Stude n ts .

o bl o n g 4 to , bo ards ,
Th e Te c h n i c M e c h a n i ca l Dra t in g
of f o blo n g 4 to , b o a rds ,
R e i se r, F . Ha rde n in g a n d Te m pe rin g o f Ste e l . Tra n s . by A M o rris a n d
.

H R o b so n m0 ,
W
. 12

R e i se r , N . Fa ul ts i n t h e Ma n ua c ture f of o o le n Go o ds . Tra n s by A .

M o rris an d H R o bs o n
. . . 8 v0 ,
Spi n n in g a n d W e a vi n g Ca lc ula ti o n s 8 vo ,
k k
. .

Re n wic , W G Ma rble a n d Ma rble W o r i n g

. . 8 v0 ,
y
R e n o l ds , O , a n d Ide ll , F E Tri ple E xpa n s i o n
. . . E n gi n e s .
( enc e
Sc i
Se rie s N0 . . 16 m0 ,
20 D V AN NOST RAND COMPANY S SH OR T TITLE C AT ALOG ’

R h ea d , G F . . Simple Struc tura l W o o dwo rk ” 12 m0 ,

hod o fObta i ni theDi fi
.

R ice , J . M .
, and J
o h n so n , W W . A Ne w Me t ng er

e n tial of Fun c ti o n s . 12 mo ,
Ri c ha rdso n , J Th e.M o de rn Ste a m E n gi n e 8 vo ,
R i c h a rds o n , 8 S Mag n e ti s m an d Elec tr i c it
. . y 12 m0 ,
R ide al , S . Glue and Glue Te sti n g 8 vo ,
R i n gs, F . Co n cre te i n Th e o r y and Pra c ti ce . 1 2 m0 ,

R i pper, W . Co urse of In struc ti o n in Ma c h i n e Dra wi n g . . f o li o ,

R o berts, F C . . Figure of th e Ea rth .

( Sc ie n c e Se rie s No . 1 6 mo ,

Ro berts, J , Jr . . La bora tor yW ki or n Ele c tric a l E ngin ee rin g 8 vc ,

R o berts o n , L S . . Wa ter tube Bo ilers
-

R o bi n s o n , J . B . Arc h i te c t u
ral Co mpo s iti o n . 8 vo ,
R o bi n so n , S . W . Pra c ti cal Trea ti se on th e Tee th o f Wh ee ls ( Sc ie n ce
Seri es No . 16 m0 ,
Ra ilro a d Ec o n o mi cs ( Sc ie n c e Se ries No . 16 mo ,
Wro ugh t Iro n Bri dge Me mbe rs ( Sc i e n ce Se rie s No . . 16 m0 ,
R oe bli n g, J A Lo n g
. a n d Sh o rt Spa n R.a il wa B ri dge s y f o li o ,

R o gers , A A La bo ra to r Guide o f In d ustria l C h e m istr

. y y . . 12 mo ,
R o ge rs, A , a n d Aubert , A B
. In du stria l Ch e mi s tr . . y .
( In P res s .
)
R o ge rs , F . Ma gn e ti sm of Ir o n V esse l s .
( Sc ie n ce Se ri es No . . 16 mo ,
R o ll in s , W . No te s on K—Ligh t . 8 vo ,
Rose , J . Th e Pa tte rn - ’
k
ma ers Ass ista n t . 8 v0 ,
Ke y to En gin es un n i n g 1 2 m0 ,
an d E n gi n e r

R ose , T . K
Th e Prec i o us M e tal s
.
( Wes tmin ste r Series ) 8 vo , . .

R o se n h a i n , W Gla ss Ma n ua c tu re ( Westm in s ter Series )

. 8 v0 , f . .

R o s s, W A Plo wpi pe in Ch e m str a d M ta ll u

. . i n e rg 1 2 m0 , * 2 y y .

R o ssiter, J T Stea m E n gi n es
. ( We s tm i n ster
. 8 v0 ( In P ress ) . . . .

Pu mps a n d Pum pin g Ma c h i n er ( Westm in ste r 8 vo ( In Pres s ) y . . .

*2
. 8 vo ,
R o uil li o n , L . Th e Ec o n o mi c s of Ma n u a l Tra i n i n g . . 8 vo , 2

R o wa n , F J. . Pra c ti ca l Ph yi s cs o f th e M o d e rn Stea m bo il e r -
8 v0 , 7 50
R o wa n , F J , a n d Ide ll ,
. . F . E . Bo iler In c ru
s ta ti o n an d Co rro si o n .

( Sc i e n ce Se r i e s N0 . 16 mo ,
R o xburgh , W . Ge n era l Fo un dr y P ti ra c ce . 8 vo ,
Ruh me r, E . Wirel ess Te l e ph o n y T . ra n s . by J Ers ki n e-Mu
rra . y . . 8 vo ,
Russe l l , A . Th e o r y of Ele c tric Ca bles a n d Ne twor s k . 8 vo ,

Sa bin e , R . H isto r y an d Pro gre ss of th e Ele c tric Tel egra ph .

Sa e l tze r, A . Trea tise on Ac o ust ic s .

Sa l o mo n s, D . E lec tri c Ligh t In sta ll a ti o n s 1 2 mo . .

Vol . I . Th e Ma n a ge men t o f Acc u mu la to rs

Vo l H . . Appa ra tus
Vo l III . . Appli ca ti o n s .

Sa n f ord, P G . . N1tro - e xpl o s1ves

Sa un de rs , C H . . Ha n dbo o k of Pra c tica l Mec h an ics

Sa un n i er , C . Wa tc h ma ker s ’
Ha n dbo o k
Sa y
e rs, H M . . Bra k es fo r Tra m Ca rs
Sc h ee l e , C . W . C h e m i ca l E ssa y s

Sc h e lle n , H . M a gn e to- e l e c tric a n d D y na mo -

e lec tri c Ma c hi n e s
Sc h erer, R . Case i n . Tra n s by C Sa lte r
. . .
22 D . V AN NOSTR AND COMPAN YS ’
SHORT TITLE CATALOG

Sin c la ir , A . De ve l o pme n t th e Loc o mo t i ve En gi n e

of 8 vo , ha l lea th er, . f
Sin da ll , R . W . f
Ma n ua c ture o f Pa pe r ( We stm in ster Se ri es ) 8 vc , .

Sl o a n e , T O C .
’
. E le me n tar y El e c tri ca l Cal c ula ti o n s 1 2 m0 ,

Sm ith , C F . . Pra c ti ca l Alte rn a tin g Curre n ts a n d Te stin g . . 8 vo ,

Pra c ti ca l Te stin g of D y na mo s and Mo to rs . 8 vo ,
Smi th, F E . . Ha n dbo o k of Ge n e ral In struc ti o n fo r Me c h a n i cs . 12 m0 ,
Smith , I . W . Th e Th e o r y of De fi e c ti o ns an d of La ti tude s and De pa rtures .

16 mo , mo ro c c o ,
Smith , J
Ma n ufa c ture o f Pa in t
. C . . s vc ,
Smi th , W Ch e mi str o f Ha t Man ua c tu
. rin g y f . 1 2 m0 ,

Sn e ll, A T Elec tric Mo ti ve Po wer

. . 8 v0 ,
Sn o w, W G Poc k e tbo o k o f Ste a m H e a ti n g a n d Ve n tila ti o n
. . .
( In Press ) .

Sn o w, W . G .
, an d No la n , T . Ve n tila ti o n of Buildi ngs .
( Sc ie n c e Se ri es
No 5 ) . 1 6 mo ,

So d d y , F Ra di o a c ti vit
. y . 8 vc ,
So lo m o n , M . Ele c tric La mps .
( We stmi n ster Se rie s ) 8 vo ,
So th e rn , JW . . Th e Ma rin e Stea m Tu rb i n e . 8 v0 ,
So xhle t, D H . . D y e in g a n d Sta i n i n g Marble . Tra n s by A Mo rris
. . an d

H R o bso n . 8 vo ,
Spa n g, H W . . A Pra c ti ca l Tre a ti se on Ligh tn in g Pro te c ti o n 12 m0 ,
Span ge n burg, L . Fa ti gue of M e ta ls . Tra n sla te d by S . H . Sh re ve .

( Sc i e n c e Seri e s No . 23 . n 16 m0 ,
Spe c h t, G J , H ard . . y , A S , M c M a s te r ,
. .
J B a . .
, nd Wa lli n g . To po gr a ph i ca l
Surve in g y ( Sc ie n c e Se rie s No 7 2
. . . 16 m0 ,
Sp y
e ers , C L . .
-
k
Te xt bo o o f Ph s ica l Ch e mistr y y 8 vo ,
Sta hl , A . W . Tra n s mi ss i o n of Po we r .
( Sc ie n c e Se rie s No . . 16 mo ,
Sta h l , A . W .
, an d Wo o ds , A T . . E le m e n ta r M ec h a n i sm y 1 2 m0
,
Sta le y , C .
, and Pie rso n , G S . . Th e Se pa ra te S ytm s e of Se we ra ge . . 8 vo ,
Stan da ge , H C . . Le a th e rwo r k e rs Ma n u
’
al . . 8 vo ,
Sea li n g Wa xe s Wa fe rs , , and Oth e r Adh e s i ves 8 v0 ,
Agglu ti n a n ts o f a ll in ds fo r a ll Pu rpo se s K . 12 m0 ,
Sta n sb ie , J H Iro n a n d Ste e l
.
( W e s tm i n
. ste r . 8 v0 ,
Ste ve n s , H P . . Pa pe r M ill Ch e mist 1 6 m0 ,

Ste ve n so n , J . L . Bla st- Furn a c e . 1 2 m o , le a th e r,

Ste wart , A . Mo d ern Po lyph a se 1 2 m0 ,

Ste wart , G . M ode rn Ste a m Tra ps . 1 2 m0 ,

Stil e s , A . Ta ble s fo r Fi e ld E n gi n ee rs 1 2 m0 ,

Still ma n , P Ste a m .
-
e n gi n e In dic a to r 1 2 m0 ,

Stod o la , A . Stea m Turbi n e s . Tra n s by L C Lo e we n ste i n

. . . 8 v0 ,
Sto n e , H . Th e Ti mbe rs of Co mm e rc e 8 v0 ,
Sto n e , Ge n R . . Ne w R o a d s and R o a d La ws . . 12 mo ,
Sto pe s , M . An c i e n t Pla n ts 8 vo ,
Th e Stud y of Pla n t Li f e 8 vo ( In P res s )
c
.

y
.

Sudbo ro u h
g , JJ . . an d J a me s , T C . . Pra ti c a l Orga n i c Ch e m i s tr . . 12 m0 ,
Sufl i n g, E R Tre a ti se o n th e . . Art o f Gl a ss Pa i n ti ng 8 vo ,
Swa n , K
Pa te n ts , De sign s a n d Tra d e
. Mar k s .
( We stm i n ster Si e r es. )
8 v0 ,
Swe e t, S H . . Spe c ia l Re po rt o n C o a l 8 vc ,
Swi n burn e , J, W o r
. d i n gh a m ,
C H , a n d M a rt i n ,
T C . . . . Ele tc ric C urre n ts .

( Sc i e n c e Se ri e s No . 16 mo ,
Swo o pe , C . W . Pra c t i c a l Le s s o n s i n El e c tr i c i t y 1 2 m0 ,
 D VAN NOSTR AND C OMPANY S S H ORT T IT LE C ATALOG 23
’
.

Ta ilf e r, L . B lea c hi n g Li n e n an d Co tt o n Yar n a nd Fa bri c s 8 vc ,

Ta te , J . S . Surc h a rge d and Difi e re n t Fo rm s of R e ta i n i n g wa ll s -
.
( Sc ie n c e
Se rie s No 7 16 m0 ,
lMh
. .

Te mple to n , W Pra c tic a . ec an i c’

5 W orksh o p C o m pa n i o n .

12 m0 , m o ro c c o ,
Terr y , H L . . In dia R ubbe r and i ts Ma n ua c ture f .
( We stmi n s te r Se rie s )
8 vc ,
Th o m, C .
, and J o n e s, W . H . Te le gra ph i c C o n n e c ti o n s o bl o n g 1 2 m0 ,
Th o ma s, C . W . Pa pe r ma - k e rs
’
H a n dbo o k .
( In P re s s .
)
Th o mpso n , A B . . Oil Fi e lds of R uss ia . t
4 ,o

Pe tr o le um M i n in g and Oil F ie l i De ve lo pm e n t 8 vo ,
Th o m pso n , E P . . H o w to Ma k e In vc n ti o n s 8 vo ,
Th o m ps o n , S P . . D y na m o Ele c tri c Ma c h i n e s .
( Sc i e n c e Se ri e s No .

16 mo ,
Th o mps o n , W . P . Ha n db o o k of Pa te n t La w of All C o un tr i e s 1 6 mo ,

Th orn l e y , T . C o tto n Co m bin g Ma c h i n e s 8 v0 ,

Co tt o n Spin n i n g . 8 vo .

F irst Ye a r .

Se c o n d Yea r .

u
Th ird Yea r .

Th urso , J W . . M o de rn T rbi ne Pra c ti c e svc ,

Tid y C Me ym o tt
, . Trea tme n t o f Se wa ge . .
( Sc i e n c e Se ri e s No 94 . . t6 mc ,
y
Ti n n e , W H G d m in in g Ma c h i n e r
o l -
. . y 8 v0 ,
Ti th e rle y, A . W . La bora to r yC o urse o f Orga n i c Ch e m i s try 8 vo ,
To c h , M . Ch e m is tr y and Te c h n o l o g y of Mi xe d Pa in ts ” 8 v0 ,
Pigme n ts an d Ar tisti c Pa i n ti n g ” . 12 mo ( In P res s ) .

To dd, J . an d Wh a ll W B Pra c ti c a l S e ama n sh ip , . . 8 vc ,

To n ge , J . Co a l W
( e stm in ste r Se ri e s
. . . 8 vo ,
To wn se n d , F . Alte rn a ti n g Curre n t E n gi n e e ri n g 8 vc , b o a rds

To wn se n d, J . Io n i za ti o n of Ga se s by C o l l i s i o n 8 vo ,
Tra n sa c ti o n s of t h e Am eri ca n In s titute o f C h e m i c a l E n gi n e e rs . 8 vc .

Vo l . I . 1908 .

Vo l II . . 190 9

Tra ve rse Ta ble s .

( Sc i e n c e Ser ie s No . 16 m0 ,
mo ro c c o ,
Trin ks, W .
, and H o us um , C . f
Sh a t Go vern o rs .
( Sc ie n c e Se rie s No .

16 m0 ,
Tro wbridge , W . P . Turbi n e Wh ee ls .
( Sc i e n ce Se ri e s No . 1 6 mo ,

Tuc k er , J . H . A Ma n ual of Suga r An a l yi s s . 8 vo ,

Tumlirz , 0 . Po te n tia l . Tra n s by D Ro be rts o n
. . 1 2 m0 ,

Tun n er , P A . . Trea ti se o n R o ll turn i n g Tra n s by J B Pe a rse

-
. . . . .

8 vo , te xt a n d o l i o a tla s , f
Turba yn e , A A . . Alph a be ts and Numera ls 4 to ,
Turn b ull , J r .
, J, . an d R o bi n so n , 8 . W . A Tre a tise on th e Co mpo u nd

Ste a m- e n gi n e , ( Sc i e n c e Se ri e s No . . 1 6 m0 ,
Turrill , S M . . Ele me n ta r yC o urse i n Pe rspe c t i ve . . 12 mo ,

Un de rhi ll , C R . . So le n o i ds, E le c tro ma gn e ts an d Ele c tro ma gn e ti c Wi n d

i n gs 12 m0 ,
Urq uh a rt, J W . . E l e c tric Ligh t . 1 2 mo ,

E lec tro - pla ti n g . 1 2 mo ,

 24 D V AN NOSTR AND C o mp a N s s S HO RT T IT LE ( tAT ALOt
.

Urq uh art , E lec tro t pi n g

J W . . y .
12 m0 , 2 oo
E le c tric Sh i p Ligh ti n g 1 2 m0 , 3 00

Un ive rsa l Te le gra ph Ci ph e r C o de .

12 m0 , 1 ao

V a c h er, F . Fo o d In s pec to r ’
s Ha n dbo o k . 1 2 mo , *
2 so
Va n No stra n d ’
s Ch e m ica l An n ual . Sec o n d i ss ue 1 90 9 12 m0 , *
2 so
Mec h a n ica l E n gi n ee rin g Da ta
Yea r B oo k First i ss u
of e 19 1 1 ( n Pre s s )
I .

y
. .

Va n Wa ge n e n , T F Ma n u a l o f H dra u lic M in in g . 1 6 mo ,
. 1 .

Ve ga , Baro n Vo n Lo gari th m ic Ta ble s 8 vc , h a l m oro c c o

. 2 so f ,

V illo n , A M . . Pra c ti c al Trea ti se on th e Le a th e r In d us tr y . Tra n s by F . .

T Addyma n
. ” 8 vo , *
10

V i n c e n t, C . Ammo n ia an dits C o m po un ds . Tra n s by M . .

J e
S
a l.t r 8 vc , *
z ao

Vol k , C . Ha u la ge a n d Wi n din g Appl ia n c e s 8 va , *

4
Vo n Ge o rgi e vi c s, G . Ch e mic a l Tec h n o lo g y of Te xt ile Fi bre s . Tra n s by .

C Sa lter
. 8 vo , *
4 so
y
.

Ch e m istr of Dyestufis Tra n s by C Sa lte r

. . . 8 vo , *
4 so
Qu
.

Vose , G L . . Gra ph i c Me th o d fo r So l vi n g Ce rta i n e s ti o n s i n Ari th me tic

and Alge bra .
( Sc ie n c e Se rie s No . . . 16 mo , 0 so

Wa bn er R , . Ve n til a ti o n i n Mi n e s . Tra n s by C Sa lte r

. . . . 8 vo ,
Wa de E J , . . Se c o n da r y B tt i a e r es . . 8 vo ,
Wa dsworth , C . P im y B tt y Ig
r ar a er n i ti o n 12 mo ( In P res s ) .

Wa gn er , E . Pre se rvin g Fru i ts , Vege ta bles , and Me a t . 1 z mo , .

Wa lk e r F , . Aeria l Na viga ti o n 8 vo ,
y
.

D na mo Buildi n g .
(Sc ie n ce Se rie s No . 1 6 mo ,

E le c tric Ligh ti ng for Ma rin e E n gi n ee rs . . 8 vo ,

Walke r, S F Ste a m Bo ile rs , E n gi n e s a n d Turbi n e s
. . 8 vc ,
Re frigera ti o n , Hea tin g an d Ve n tila ti o n on Sh ipbo a rd .

rz m o ( In Pre s s ) .

E lec tric i t yi n Min i n g . 8 vo ,

Stea m B o ile rs , En gi n e s a n d Turbi n e s . . 8 vo ,
Wa lk e r, W H Sc re w Pro pulsi o n . . 8 vo ,
y
Wai lis- Ta le r, A J Bea rin gs a n d Lubrica tio n . . . . 8 vo ,
Mo de rn C c le s y . 8 vo ,
M o tor Ca rs . 8 vo ,
Mo to r Ve h i c le s for Busi n e ss Purpo se s . 8 vo ,
Poc k et Bo o k of Re f igr era ti o n a n d Ice 12 m0 ,
Re f ig r era ti n g Ic e -m
a i n g Ma c hi n e r
an d k y . 8 vo ,
Re f ig r e ra ti o n a n d Co ld Sto ra ge . . 8 vo ,
Suga r 12 m0 ,
Wa n kl yn J A Trea ti seo n th e Exa m in a ti o n o f M il k
, . . . 1 2 m0 ,

Wa te r An a lysis
Wa n sbro ugh W D Th e A B C o f th e Di ffere n tia l Ca
,
l c u1
. . u s .
1 2 m0 ,

rz mo ,

mf th Ml
Sl ide Va l ve s .
1 2 m0 ,

Ward J H , . . Stea or e il i o n 8 vo ,
Wa ri n g, Jr G E . . y
Sa n i ta r C o n di ti o n s .
( Sc i e n c e Se ri e s No . . 1 6 mo ,

Se we ra ge La n d Dra i n a ge
and

Mo de rn Me th o d s o f Se wa ge Dis po sa l .

H o w to Dra i n a H o u se

Warre n , F D H a n dbo o k. o n R e i n . f o rc e d C o n c re te
26 D V AN NOSTRAND C OMPANY S SHORT T ITLE CATALOG
’
’

Zh
a n e r, R . Tra n smiss i o n of Po we r .
( Sc ie n c e Se ri es No . . 1 6 mo ,

Z idl
e er, J, . and Lusg arte n , J . El e ctri c Arc La mps 8 vo ,
Zu
e n er, A . Te c h n i c al Th ermo d y na m ic s. Tra n s by
.
J . F . Kl
e in . Two
8 vo ,
Z imm G Fer, . . Me c h a n i ca l Ha n dlin g o f Ma te ria l 4 to ,
Z ip J T
se r, . e xt ile Ra w Ma te ria l s Tran s by C Sa lte r
. . . . . 8 vo ,
Zu
r Ne dde n , F . k
Engi n eeri ng Wo r sh o p Ma c hi n e s a n d Proc e sses . Tra n s .

8vo
 Fo s t e r s Po c k e tbo o k i s
’
th e

b us y e n gi n ee r s
’
fi rst as s i s ta n t.

Wh e n h e wa n ts re a l l y re li a ble
ENGINEERS in f
or ma ti o n h e , turn s t o Fo e
t e r to g e t i t fo r h i m, an d he

POOKEI BOOK
'
n .
pl a c e s
- l ~n~ " 8 1 m Ti b hi m d a ta on th e l at e s t an d

be s t pra c ti c e in e l e c t ri c al

e n gi n ee ri n g w i th a mi n i mum e x pe n d i t reu of

ti me and e fi o rt.

Th e ma te ri al i n Fo s te r i s l o gically arra n ge d

an d i s i n d e xe d i n a full t ab l e o f c o n t e n t s an d

a vo l umi n o us i n d e x th a t a lo n e c o v e rs s i x t yp g a e s,

a n d to th i s is a d d e d a se t o f pa t e n t t h umb i n d e x
ta b s t h a t ma k e f
re e re n c e to any s e c t io n o f th e
b o o k pra c t i c a l l y i n s ta n ta n e o us . Th e i n d e x i s
mo s t th o ro ug h an d re l ia b l e . It po i n t s rig h t t o

th e s po t w h e re th e i n f o rm a t i o n s o ugh t i s .

If yo u e ve r n eed in f o rm a t i o n on e l e c t ri c i t y ,

yo u o ugh t to h a ve a co p y of th e n ew Sixt h
Ed i ti o n , c o m pl e te l y re vi s e d a n d e n l a rge d , w 1t n

fo ur fif t h
-
o f th e o ld ma t te r re pl a c e d by n e w,

n o - to -d a te ma te ri a l , and c o n t a i n i n g al l th e e x

c e ll e n t f eat ure s na me d a b o ve , wi th pa g e s ,
i l l us tra t i o n s a nd 7 1 8 ta b l e s . T h e pric e i s
 By ALBERT l . FRYE, M Am So c C E . . . . .

A C O M PREH ENSIV E tre a tme n t o f Civi En gin e e rin g in se ve n ty l ,

se c tio ns c o mprisin g a bo u
, t L400 pa ge s ; w ith 500 ta b e s an d l
l 000 il u
, l
stra tio n s l l
Also a c o mp e te g o ssa ry o f e n gin e e rin g te rms
. .

Ea c h man su bje c t re c e ive s e c o n o mic co n side ra tio n an d a n a sis an d ly ,

is re in fo rc e d w ith e xc e rpts fro m an d re fe re n c es to the mo st impo rta n t

,

c o st an d o the r da ta in o u l
r e a din g te c hn ica pu l
b ic a tio n s in c u din g l —
l
hu n dre ds o f il u l
stra tio ns o f u p - to -da te e n
g in e e rin g stru c tu re s an d de ta ils .

H o w to de sign e n gin e e rin g stru

“
c tu re s e co n o mica ha s re c e ive d lly "

c are fu stu l y
d in the pre pa ra tio n o f this w o rk The fu n da me n ta
.
prin l
l l
c iple s a id do w n ca n be fo lo w e d re a di y by e ve ry l ou n g e n gin e e r an d y ,

l
ca rrie d to prac tic al c o mp e tio n in the fin ishe d stru c ture .

l
Mo st o f the ta b e s a re n e w an d ha ve be e n pre pare d re ga rd e ss o f l
time an d e xpe n se Man y o f the m ha ve be e n arran ge d in a n o ve
. l
ll
ma n n e r w hic h w i a ppe al e spe c ia lly
to e n gin e e rs ; a n d a ll ha ve be e n
ll l
ca re fu y c he c ke d an d re c he c ke d to e imin a te the po ssibi ity o f e rro rs l .

Gra phic a l me tho ds ha ve be e n i u ll

s tra te d fre e thro u g h o ut the w ly
o rk ,

in c o nn e c tio n w ith an a tic a so uly l l

tio n s .

M e tho ds an d c o st o fw o rkha ve re c e ive d spe cia a tte n tio n ; a n d thus the re l , ,

is cro w de d in to the L400 pa ges o f te xt in sma t pe a n d in po c ke tbo o k

, ll y
fo rm a va st amo un t o f e n gin e e rin g da ta fo rmu
, las an d ta b e s tha t w i be
, l , ll
fo u n d a bso u te l ly l y
in dispe n sa b e to e ve r e n gin e e r c o n tra c to r an d stu , de n t .

D V AN NOSTRAND C O M PANY, Publi s h e r:

.

23 Murra y an d 27 Wa rre n Sts .

, NEW YORK