THE FAN

I N C LUD I N G TH E

TH EORY A N D P RA C TI C E OF C EN TRI F U GA L
AN D A XI A L F A N S .
EM

THE FA N
I N CL U D I N G

T H E T H EO RY A N D P R AC T I C E O F
C E N T R I FU G A L A N D A X I A L FA N S

BY

C H A RL E S H . I N N E S, M . A .

“
I GAL
C ENT R FU P U M PS ,
' ‘
A UT HO R OF

I U RB I N E S, A N D W AT E R MOTORS,
"AND
“
P R O B L E MS I N MAC H I N E D ESI G N ”

R EV SE I D BY

w M WA L L A C E WH
. .
, . SC .
,

AND

F . RA B Y JO L L E Y M ,
. Sc . T EC H .

SE C O ND E DITI O N

LON D O N , EN G L AN D
T H E T EC H N I C A L P U B L I SH I N G C O L I M I T E D .

1, GO UGH SQ U A RE , F L E ET ST R E ET , E C .

1 91 6
 A U T H OR S P REFA C E
’
.

IN the following pages I give a theory of th e fan which

di ff ers considerably from anything that I have seen i n
print and which may therefore meet with some criticism
, .

I n all wo ks in this or any other language that I have read

r

I have found the equation

where H is the head of air against which the fan works 02 ,

'

is the tip speed a is a constant for any particular fan and 0

, ,

is the equivalent orifice .

That this equation cannot be that of all fans is obvious

when we remember that in many fans the manom etric
effi ciency increases at first as the orifice increases from z ero ,

while thi equation states that it decre ases ; nor can it be

s

mainta ined that the equation applies to fans whose mano

metric efficiency is greatest at z ero orifice for if the curve
,

of manometric e ffi ciency be drawn with this as ordinate

and orifices as absciss ae it will be found that the tangent at
,

the point where the curve cuts the vertical axis is hori z on
tal and that this is not the case may be seen from fig 4 9 .
,

which gives curves for eleven fans diff ering widely in con
struction H aving already studied the centrifugal pump
.
,

it occurred to me about ten years ago that its theory m ight

be applied t o that of the fan and except that the f an does
not actually lift air as the pump lifts water but acts like
, ,

a centrifugal pump that pumps against the resi stance of

hori z ontal piping only I consider that the same theory may
,

be applied to both .

I n C hapter V III I have endeavoured to show that my

.

theory ag rees with the results of e x periment as far as the se

,
v
.
vi A U TH OR S
’
P R EF A C E

may be trusted Th e following pages commence with the

.

theory of the centrifugal fan following which are e xpe r i ,

ments with and descriptions of this type ; and in Chapter

X is given a description of P rof B ateau s high pressure
’
-
. .

fans in whose design it may be mentioned the variation of

,

the density of the air must be taken i nto account I n .

C hapter X I I will be found an imperfect theory of propeller

.

fans imperfect because I cannot find all the information I

,

require fr m published experiments Followin g this will

o .

be found descriptions of this type and of P rof B ateau s .

’

screw fans the theory of which closes the book

, .

I hope this book will be of service to those who have to

design or who wis h t o understand the working of fans
,
.

T here are many very inferior fans used whic h do their

work wastefully and it would be satisfactory to see these
,

replaced by others of scientific design .

C H AS H IN N ES . . .

RU T H E RF ORD C O L LE G E ,
E N WA
C S T L E ON T Y N E ,
- -

Oct ob er , 1 9 04 .

N OT E .

IN this edition the work has been thoroughly revised and ,

various alterations and ex pa nsions of the text have been

made Th e calculations have been checked and a revised
. ,

and uniform system of notation has been adopted .

A lmost all the figu res have been redrawn C hapter XII I .
,

dealing with recent practice mainly in the con st r uct l on o f

fans and centrifugal compressors is quite new ,
.

I t is hoped that in its altered form the work may be

increasingly useful .

W M WA LL ACE
. , ,

F RABY J OLL EY
. .

S ept em b er , 19 1 6 .
 C ON T E N T S .

C H A P TER I .

e x a ms
Th e C o n se r v a t ion of li c ti o n f t h L w f C
En e rgy — A pp a o e a o on
se r ti o n f En rg y t
va o I n c o m pr i b l F l u i d i n M oti o n
e o an e ss e

R easo n f t r ti n g A i
or if it h d
ea Co n t n t V o l u m
r as a a s a e

A ppli c t i o n f F o r m u l ( 1 ) Th Lo f H d b y Su r f c
a o a — e ss o ea a e
F ri c t i o n f F l u i d u nd r S t d y F l o w i n P i p o r P
o a e
g ea a e as s a e
f U n i f o rm S c t i o n Dr H A Gi b o n F o r m u l C o nd i t i o n '
o e — . . . s s a—

for M i n im u m Lo f H d d to Su dd n C h n g i n t h
ss o ea ue a e a e e
D i r c t i o n f M o ti o n C o nd i t i o n f
e o M xim u m G i n f
— or a a o
P r u r H d Lo
ess e f H d d
ea — to C h n g i n V l o c ity
ss o ea ue a e e
w i tho ut C h n g i n D i r c ti o n
a e e 1 —9

C H A PTER I I .

M an om e t e r an d P i to t Tu b e A n e m om e te r — D e t e r m i n ati o n of
—

H ead f r o m t h e Wat e r G au g e an d t h e Cal c u l ati o n o f Ai r

D e n s i ty f o r a g i v e n Hu m i d i ty — Sa tu ra t i o n P re ss u r e s an d
V o l u m e s o f Aq u e ou s V ap ou r 1 0— 1 6

C H A PT ER I I I .

Law C h a n g e o f M o m e n t o f M o m e n tu m —
of Wo r k d o n o n e t he
i
A i r n i t s P as s a g e th r ou g h a Ra d a F i l low Fan — Lo ss e s of
E n e r gy o r H e ad w h il e p ass i n g th r ou g h t h e Fan — Eq u a ti o n
f o r find i n g t h e M an o m e t r i c an d M e c h an i cal i
E ffi c e n c es i in
Te r m s of — C as e s A , B , an d C 1 6— 3 2

C H A PTER I V .

Th e o r e t ic l C h r c t ri tic
a a a e s s of F an s — Tab l es 2, 3, an d 4— V ol u
me t r i c Effi c i n c y e 33 3 9
-

C H A PT ER V .

D e si g n o f Rad i al F l o w F an s — M ixe d F l ow F an s— Sc rew o r

P ro pe ll e r Fan s— D e si g n o f M ixe d F l o w F an s — Tab l e 5 40 -
57

C H A P TER V I .

V ar i at i o n of P r e ss u re i n a i
C e n t r f u g a Fan — S m l i il a r F an s 5 7 64
-

vii
viii CO N TE N TS
C H A P TER V I I .

P AG ES
H e e n an an d xp r i m n tGil b C n tr i fug l F
e rt s ’
E e e s on e a an s
T bl 6 T t f
a e M i n V n t il t i n g F
—
es o Th o ry f t h
a e e a an — e o e
P i to t Tu b Exp r i m n t o n C n t r i f u g l F n b y B ry n
e— e e s e a a s a

Donkin D ki —Ex p r i m n t l A pp r tu
on n s
’
T bl 7 8 e e a a a s -
a es
d 9 Th El
, ,

an — n Ty p f Fe u d b y D ki
eve E p i es o an se on n s
’
x er

m nt e B l g i n C o m m i i o n ( 1 8 9 2 ) Ex p r i m n t
s— e a on Mi n ss e e s e
F n T b l 10 1 1 12 13 14 15 16
a s— a es d 17 E p i
, , , , , , ,
an — x er
m n t wi th R t u F n
e s T bl 1 8 1 9 2 0 2 1 22
a ea d 23 a s— a es , , , , , an
C n t r i fu g l P u m p E p r im n t
e a b y C h rl H I nn x e e s a es . es

T bl a 24 d 2 5 C n t r i f u g 1 P u m p Ex p r i m n t b y C R
es an —
e a e e s .

P r o n T b l 2 6 d 2 7 Op n r u nn i n g F Ex p r i m n t
.

—
a s s— a es an e an e e s
-

— T bl 2 8 M ri n
a e V n t il t i n g F
— Ex p r i m n t b y M
a e e a an e e s .

L lon g T b l
e 29 30
—
d 3 1 M o r tim r D i m t r l F
a es , ,
an — e a e a an

E p ri m n t
x e Kl y F E r ci
e s T bl
—
32 e d 33 P l r an xe se s — a es an e ze

D o r tm u nd F Bu m t d d C h nd l r F
an — Ex p r i m n t s ea an a e an e e s

T bl a 34 eS F Ex p r i m n t
— T bl
er 35 B ckan d e e s— a e —
e an

H k l F nen cAll n F n a Wo r k xp nd d i n o r c o m i n g
s— e a s— e e e ve

Di F ti sc G n t H r ch r F
nc
'

on — e e s e-65 e s e an —1 7 6

C H AP TER V III .

C o m p r i o b t w n Th o r y
a s n d Ex p r i m n t ~T b l
e ee 38 39 40e an e e — a es , , ,

4 1 , 4 2 , 4 3 , 4 4 , 4 5 , 4 6 , 4 7 , 4 8, an d 49 1 77 194 -

CH A P T ER IX .

H ig h P r e ss u re F a n s
- —
Tab l e 50 1 9 4 — 2 03

C H A P TER X .

Th e Th e o r y of P r o p e ll e r V e n til a t i n g F an s Wal k e r s ’
E x pe r i
me n ts — Tab l es 51 52 , . an d 53 2 03 2 1 8 -

CH A P TER XI .

H eli cal P r o p e ll e r s — Tab l es 5 4 , 55 , an d 56 2 1 9 24 3

-

C H A P T ER XI I .

O th r P r o p e ll e r V e n t il a t i n g F a n s an d Rat e a u Sc r e w F an s Tes t o f
e —

an Axi al o r Sc r e w Fan h o ry Of Rat e au Sc r e w F a n s 2 4 3

— Th e T e —
2 58

C H A P TER XII I .

F an s an d C o m pr e ss o r s : Re c e n t P r ac t i c — I m pr o v e m e n t s i n Fan e

D e s i g n C o n s t r u c ti o n o f C e n t r i f u g al F l u i d P u m p s — S t e pp e d
V an e s f or C e n t r i f u g a l Fan s— Arr an g e m e n t o f Fan s— C on
s t r u c t i o n o f C e n t r i f u g a l F a n s —
C e n t r i f u g al Gas P ur i fi e r — F an s

f o r D u st La d e n A i r — I nd u c e d Dr a u g h t f or F u rn ac es — Tu r b i n e
-

Dr i v e n V e n t il a t i n g Fan s C e n tr i f u g al C o m pre s s o r f or Gas —

H ig h P r e ss u r e Ce n t r i f ug al C om pr e s sox— C e n t r i f u g al A i r Co m
pr e ss o r s— C o m pr e sso rs wi th I n c li n e d St rai g ht B l ad es 258 -
291

N OTE S 292

I D
N EX 2 93 302 -
 C E N T R I F U G A L FA N S

C H AP TER I .

The Cons er vati on o f E ner gy — E nergy is

indestructible ,

although it may appear in a n u m b e r of forms,

some of which
are useful t o man whilst others are not Thus steam at
,
.

a high pressure contains energy i n the form of heat part of ,

which can b e convert ed into useful work whilst the r e ,

m ai n d e r is wasted in overcoming fri ction or is rej ected with ,

the conden sed steam at a low temperature and is of no ,

service to man B ut the law of conservation of energy

.

supplies us with equations which are of the utmost service

in correctly designing machines in which a flow Of fluids
takes place because w e know that changes of pressure
, ,

volume and velocity are accompanied by alterations of the

, ,

fo r ms i n which the initial energy e xi sted but that the ,

quantity of energy is unaltered Al though we are now .

dealing with air a compressible gas and should therefore

, , ,

i n strict accuracy take into account the alteration of

,

volume that accompanies change of pressure yet in all

, ,

except the high pressure typ e of fans since this change of

-

pressure is so small it need not be considered especially

, ,

if w e denote the volume of air pas sing through t h e fan

as the volume occupied at a pressure which is the arith
metic m ean between that at suc t i on and discharge L e t
'

P lb per square foo t be the diff erence between the suction

.

and discharge pressures and Q the mean volume i n cubic

,

feet per second of the air pas sing through the fan measured
as above ; then P Q foot lb is the useful work don e by the
.

fan per second .

2 CE N TR IF UG AL F AN S
A pp li cation of the L aw of Con ser vati on f
o E ner gy to an

I ncomp r essi ble Fluid i n Motion ~

— Suppose a
liquid of density
.

8 lb per cubic foot is contained in a vessel ( fig

. whose .

free hori ontal surface is maintained at a constant height

z

H above any assumed level while the liquid is being dis

,

charged from a pipe connected to the vessel Consider .

a point b i n the surface of the liquid and a point a in the ,

pipe h feet above the given level L e t 22 feet per s e e . .

h e the velocity of a particle at a where the pressure is ,

p lb per square foot and let A be the sectional area of

.
,

the pipe in square feet .

FIG . 1 .

T hen ,
assuming a piston to move with the water at a a ,

displacement of 1 foot would involve p A foo t pounds of work

being done upon the piston at the e x pense of A cubic feet
,

of liquid or p foot pounds at the ex pense of each cubic

,
-

foot ; or as 1 cubic foot contains 8 1b p + 8 foot pounds of

,
.
,
-

work are done at the expense of each pound Of liquid .

Taking atmospheric pressure as z ero the energy of the ,

water at b is complete ly potential i n nature due to a level ,

head H feet ; whilst it follows that the energy at a e x ists

in three diff erent forms v i a potential energy h kinetic
—
, ,

energy e 2 g and energy due to pressure or pressure

2 -

, ,
 A PP LI C ATI ON OF FORMU LA 3

energy p + 8 H ence w e say that a liquid under pressure

, .

p,
velocity v,
and at a height 11 h as a pressure head
p +8 , ,

v e loci t he ad 12 —2 2 g and actual or level he ad h while the

2

Zan t head d ue to all th ree is H or more simply we

-
, ,

e q u i v al , ,

may state that 0

2
H h+ + P
( )
1
8 2g
.

Reas on f or tr eati n g A ir as if it had a C on stan t Volum e .

— practice the highest water gauge against which a fan

In
works is about 1 2 i n ches As the water barometer is about
.

3 4 feet this amount corresponds to a compression of 3 7 t h

,
1

of the o r iginal volume of the air Thus if the discharge .

,

of a fan is measured by the mean volume the ac tual ,

volume at any instant can only be gg t h part of the mean

volume greater or less than this H ence for practical pur .

poses the ai r may be treated as e u incompressible fluid

obeying equation I n exact experiments on fans the ,

hygroscopic state Of the air m ust be considered as well as ,

its temperature and pressure .

App li cation of For m ula Suppose for example that —

, ,

air weighs 00 7 5 lb per cubic foot and moves with a

.
,

velocity of 40 feet per second in a strai ght pipe which

grad ually enlarges i n section so that the velocity is reduced
to 2 0 feet pe r second what will the change of pressure per
,

square foot be if the pipe is hori zonta l ? L e t p l 2 1 be the ,

2

pressure and velocity when the latter is 4 0 feet pe r second ,

and p g 222 similar quantities when it is 2 0 Then putting

, .
, 0
'

for the density of air we have ,

2 2

’
0
'
2g 0
'
29

and h l = h2 ;
" 2
P2 P1 v
the “ 0
'
2g
2

_
00 7 0
13 2 p,
64
( 1600 4 00)

sq ft lh pe r . . .

This correspo nds to a change of he ad : feet of air .

4 CE N TR I FUG AL FAN S
I t will be seen later by the application of the principle
,

here involved that the mechanical e fficiency of a fan can

,

be considerably increased by using various means for

gradually reducing t h velocity of the air and thus
e ,

increasing its pressure .

The L oss of Head by Sur face F r i cti on of a Fluid u n de r

S teady F lo w i n a P ip e or P assage of Un ifor m Secti on — The
following analysis , due to D r N icolson , is Off ered as a .

solution of the problem

Th e irregularities or obstacles on the surface of a pipe
create an eddying motion i n the fluid in the nature of
vortices These i ncrease in si e as they pass to the centre
. z

o f the stream where their velocity is made up out of that

,

of the remaining uneddied fluid C onsider the flow i n a .

FIG . 2 .

pipe of circular section assume the speed loss to be or

, ,

where v is the mean velocity across the whole section of

the pipe ; further let 0 be the mean thickness and b the
, ,

mean circumference of the vorte x eddied at each obstacle

,

on the surface of the pipe as represented in fig 2

, . .

T hen the volume of fluid eddied per second at each

obstacle b 672 and the energy lost by eddy formation
,

§6
per Obstacle per se eon d = E2 where I S the dens i ty 0
‘

g
of the fluid .

I f there are n eddies per foot length of pipe then the ,

b evn o
kinetic energy lost per foot length 2 (w ) 2

9
 LOSS OF H E AD 5

Th e work done per second on a block of fluid l feet

long (p p 2 )a where 191 and p, are the pressures at points
l
— v,

l feet ap rt and a is the sectional area of the pipe

a .

N o w it follows that the work done per f oo t length per

,

second is equal to the kinetic energy lost per foot length ,

s i nce the forward motion is constant and the lost energy

is m ade good out of the pressure head thus ,

—
71
1
p2
l
an
2g
( ) cv 2 ,

so that loss Of head

“
5 35
2
“
PI Pz —
‘

( )
e
h1 —
n a .
— o
4 2 . 2
0
'
a 2g

where m is the mean hydraulic depth of the pipe

section divided by circumference Thus it follows that ( .

is pro portional to the n umber of obstacles per foo t length ,

to the square of the fractional speed loss and to the thic k ,

ness of wat er eddied which depends on the height o f the

obstacles it is also evident that in no way depends on
,

the nature of th e fluid .

P rofessor U nwin in D evelopment and Transmission o f

P ower gives

00 02 7 +3

Examples from Weisbach show

1
1+ for 9 163 0 P ipes :

1 for encrusted pipes :

where is the diameter of the pipe i n feet
d .

Fo r a pipe of circular section 177 = d 4 so that equation , .

,

( ) becomes
2
4t 122 e2
hl =
f E — —

53 9
a o .

29

f is the coefficient of resistance of the pipe referred to

t he velocity An d as the losses of head due t o bends
6 CE N TR IF UG AL F AN S
and elbows and passages of any form are proportional to
, ,

t h e square Of the velocity in a given m achine w e can say ,

that the loss of he ad is v , z

i n addition to that part of the loss of he ad due to sudden

changes of direction and velocity where F1 is the coefficient ,

of resistance of the machine referred to the velocity 72 .

Dr A H
. Gi bson s For m ula
. . I n the cas e of air
’
.

D r Gibson has shown 1 that formula ( 2 ) only applies if

.

the coefficient is varied n ot only with the physical con ,

dition o f the interior surface of the pipe but also with i ts ,

diameter with the mean velocity of flow with the mean

, ,

pressure and with the temperature of the air For either

, .

cast iron or wrought iron pipes laid under normal con

- -

di t i on s as to jointing etc he gi ves the following relation

,
.
,

ship for the drop in pressure in lb per sq in ( 8p ) in a . . .

le n gth l ft of pipe of diameter d f t

. .

n l n
vm l
-

8p = 0 000001 2 5
' P m
’

6 6 7‘
d 3 n ‘ ‘

pm being the mean pressure in lb per sq i n and v " . . .

, ,

the mean velocity of the air in ft per sec Th e above . .

relationship holds for a mean temperature of 6 6 F Th e °

.

value of n depends upon the diameter of the pipe as

follows

Die .
( in ) : 3 . 5 . 7 . 9 . 12 .

n 17 75

For a 3 in pipe the formula becomes

.

3p 00 000002 1,

and t h e drop i n pressure in lb per sq i n on 1 00 ft length . . . .

of 3 in pipe when the mean pressure is atmospheric

.
, ,
 L OSS OF H EAD 7

say 1 4 7 lb per sq . . in .
,
is for th e following values of e
( in
ft per sec )
. .

v 10 20 30 4O 60
00 125 8 00 4 4 7 4 009 3 9 3 01 5 9 1 03 3 4 5

a m ea n pressure of 4 0 lb per sq in the drop i n

Fo r . . .

pres sure at these velocities becomes

22 10 2O 30 40 60
8p 00 2 8 8 9 01 02 7 02 1 5 6 03 65 1 07 6 7 8

I t is evident that the loss of head or pressure drop is

afi ee t e d very much by the mean pressure of the air .

Th e values of n i n the formula were d e duced from the

e x periments of Riedler and Gutermuth on the P aris air
B

FIG . 3 .

mains the length of pipes ranging from

, ft to .

f t the mean pressures from 9 2 to 1 18 lb per sq

.
, . .

in and mean velocity from 8 5 to 2 8 6 f t per sec

.
, . .

U nfortunately the mean t emperature which would aff ect

,

the drop somewhat w as unknown but assumed at 65 F ,

°
.

E xperimental results on smaller and shor t er pipes by

M St ock alpe r D r Bri x and the author were also con
.
, .
, ,

s u lt e d though i n some of these the mean temperature had

,

to be assumed .

Cond i ti on f or Mi ni m um Loss of Head due to a S udde n

C han ge i n the D ir ection of Moti on — VVh e n a sudden change
of d i rection of motion and of velocity takes place , s uch as
that represen t ed in fig 3 from A B to A C, where A B r e pr e
.
8 CE N TR IF UG AL F AN S
sents a velocity c l A C a velocity , v ,
2
and the angle B A C is
called 6 then the loss of head
,

2g

This expression obviously has its least value when B C coin

cides w ith B D the perpe ndicular on A C so that
, ,

22
2
cos 6 ( )
4

T hus the minimum loss of head

h
e 12 @21 cos 6 2 2

Con di ti on f or Maxi m um Gai n of Pr essur e H cad — L e t 191 be

the pressure before the change and 192 after the n , ,

2 2
p 2 + 91
f

101 v1
+ loss of h e ad = ~ + 3
0
'

29 0 2 57 '

J7 2 79 1
gain of pressure head
a

@21
2
i2
z
z
711
2
12
2
2
2 1) e cos 6
29 29

12 72 cos 6 o
z _
2
AC CD
( )
1 2
5
6 6
Th e maximum gain of pressure head occurs when
d
dyz
( vl r 2 cos 6 12
2
2
) 0

re , when 2
1
2
'
01 cos 6,

or when A C=C D .
10 CE N TR I FUG AL F AN S

C H AP TER II .

Manom eter and measure the work done

P i tot Tube .
— To
by a fan we require to obtain the pressure d ifi e r e n ce
,

between suction and discharge an d the q uantity of air

pas sing through the fan in unit time A n instrument .

te rmed a manometer is used to obta in diff erence of

pressure I f a bent tube (fig 4 ) contains water and the
. .
,

end 0 is exposed to greater pressure than the end b the ,

FIG . 4 .

liquid will rise on the latter side to a height propo rtional

to the diff erence of pressure At the average te m perature
.

a cu ft of water weighs 6 2 3 1h so that each inch of

. . .
,

gauge registers a pressure of 5 1 9 2 lb per sq ft . . .

A distinction must be made between the dynamic and

the static pressure of a gas When the mouth of the tube
.

is placed parallel to t h e direction of flow of the air th e ,

static pressure is measured at that point ; but if the tube

 MAN OM ETE R AN D P I TOT T U BE 11

is be nt so that its mouth faces upstream and is at right ,

angles to the lines of flow there is i n addition to th e sta t ic

, ,

pressure a dynamic pressure due to the bombardment of

,

the mouth by the gas molecules This dynamic pressure

1
.

is taken advantage of in order to measure the vel ocity of

air in ac cura t e fan work Th e amount by which the
.

press ure is raised is

2
v
0
'
lb per sq
. . ft .
,
29

so that if h is the di ff erence of level of the water mano

meter ( fig 4 ) when the end 0 is exposed to the dynam ic
.

and b t o th e static pressure ,

2
12 72 0
'

h o
1D .

2 g

I n order to measure the dynamic pressure or head

accurately i n practice P i t ot s tube is used one form of
,
’

FI G . 5 .

which is shown i n fig 5 I t consists essentially of two

. .

concent r ic tubes containing air Th e end a of the inner or .

dynamic tube faces upstream and this tube transmits the ,

dynamic pressure through c to one leg of the mano m eter .

Th e static pressure is transmitted from the oute r tube

through tube b to the other leg of the manometer small ,

holes 00 2 in dia in the ou t er tube allowing it t o com

. .

m u n i ca t e with the outside air .

 12 CE N TR I F UG AL F AN S
A fter much experimenting it has been established that ,

the exact si e or shape of the mouth a does not materially

z

aff ect the res ult but that the exact si e shape and position
, z , , ,

of the static opening are important considerations I f the .

3

static opening is large or is in the shape of a slot in the

side of the P itot tube the pressure registered is too large
, .

P laced in positions where the flowing air produces a

suction eff ect the pressure reg isters too low ; for instance
, ,

an ordinary manometer with an Open mouth parallel to the

flow might register too high or too low according to the ,

si e of the mouth and the velocity of flow For these

z .

reasons a P itot tube of the form illustrated in fig 5 is to .

be recommended .

F or small pressure di ff erences in order to magnify the ,

readings the manometer is inclined to the hori ontal or a

, z ,

light liquid such as gasoline used .

A n em om eter For rough wor k the quantity of air dis

.
—

charged b y a fan is measured by an instrument called

‘

an anemo m eter I t consists of a small wheel carrying

.

vanes set at an inclination to the plane of rotation so ,

that when a current of air passes through t h e wheel in

the direc t ion of its axis the wheel will rotate and com
m u n i cat e its motion to gearing which works a coun ter
showing the number Of revolutions I t was originally .

supposed that the number of revolutions w as propor

t i o n al to the velocity of the air but this is not the case ,

even when the current of air has a uniform velocity ,

because of the friction of the apparatus and is very far ,

from being the case when the velocity is variable I the . n

latter case the anemometer which is usually graduated by ,

rotating it i n still air at the end of an arm driven at

uniform velocity greatly exaggerates the quantity of ai r
,

passing through it That this is so will be readily under

.

stood i f we consider the case of an anemometer which is

alternately placed during equal S hort periods in a current
of air and in still air ; when in the former it attains a
,

speed very nearly proportional to that of the air and when ,

i n the latter it slows down very gradually so that the total ,

number of revolu t ions is considerably grea t e r than would

 AN EM OM ETER 13

have been obtained had the instrument been placed in the

current of air for hal f the time .

D eter m i na ti on of H ead fr om the Water Gaug e and the Cal ,

cu lat i on of A i r De ns i ty f or a Gi ven H um i di ty A water —

gauge of h in corresponds to a pressure P lb per sq f t

. . . .
,

Where 8h
P — lb per sq ft ( )
9
fi
. . .

and 8 weight o f 1 cu ft Of water at the te mperature

. .

at which the experiment was mad e I f is the weight of . (r

1 cu ft of air at that same temperature ( to fi nd which

. .

the height of t h e b ar om e t e r and the moisture contained

'

by the air must be known ) then the equivalent head of air

,

against which the fan is working is

h8
H — ft of air
.
( )
1 0
l2 0
‘

and as 8 : 62 3 and
H = 7Oh or
"
if?
" at 6 2 Fah

h a . of
.
,
we may t ake
air ( )
1 1

where rough calculations only are necessary Where .

greater accuracy is desired we require the barometer ,

thermometer and hygrometer to obtain the correct value

,

of From D alton s L aw of P artial P ressures it is known

’

that the tota l pressure exerted by a mixture of two gases

filling a space is equal to the sum of the pressures that they
would produce if they filled the spac e alone I n the case o f .

unsaturated moist air we can find the pressure of the vapour ,

because by the hygrometer we can find the dew point or

temperature at which the amount Of moisture in the air
would just saturate it .

Th e variation in the saturation pressure and volume of

aqueous vapour with temperature has been carefully i n
v e s t i at e d by R egnault but no simple physical law ap
g ,

par e n t ly e x i sts between these quantities so that they are ,

most conveniently presente d in tabular form .

K nowing th e saturation volume for a given temperature ,

w e can find the weight of moisture per cu f t . .

14 CE N TR I F UG AL F AN S
L e t P = P ressure due to the air and moisture combined ,

in i nches of mercury .

P, = P ressure of aqueous vapour at the dew point

,

in inches of mercury .

[N ote P is slightly greater than the pres

—
,

sure at the temperature considered but ,

this change i n pressure may be neglecte d ]

D ensity o f dry air at the standard pressure
2 9 9 2 i n and at any absolute te mperate 7
.
, .

o , D ensity of aqueous vapour .

cr D ensity Of air and vapour combined .

0 D ensity of dry air at pressure and temperature

considered .

T hen the weight o f dry air in 1 cu ft . . of the atmosphe r e at

the g iven temperature 7 F is °
.

further ,

: 53 2 F
19 0 ;

where p is in lb per sq ft and v is the volume of

. . . 1 lb i n
.

cu ft
. .

29 92 '
X 13 6‘
x 8 39 8 -

Th l1s Ud
12 x 53 2 -
x r 2

whence by substitution
39 s (P '
— P,)
2

T his must be added to the density of the water vapour ,

obtained from the saturation volume given in the table

below Thus the density of moist air at the given tem
.

p e r at ur e and pressure is
' ’

0 os .
 C AL C U LATI ON OF A I R D E N SITY 15

I f very g reat accuracy i s desired account must be tak en of

the slight increas e in length of th e b arometric column due ,

to the fact that the temperature f at which it is read is not

3 2 deg Fah
. at which temperature only the standa r d
.
,

ba romete r is 2 9 22 To allow for this w e must write

.

398 (P — P s)
ax ’

2 9 9 2 ( 1 + 0 0001 {F
- -
32 } )r
the constant being the specific heat o f mercury .

Th e table of saturation pressures and volumes of aqueous

a o ur
p given on p 1 6 has been taken from D r J ude s
. .
’

P hysics
”
4
.

Th e following example will make the method of obtain

in h e value of clear
gr i e temperatur e of the atmosphere is 7 7 deg Fah and
0

. .
,

the dew point as obta ined by the hy g rometer is 4 1 Fah .

,

what is the weight of 1 cu ft of air if the barometer is . .

2 9 i n ? Th e pressure of the moisture is 0 2 5 7 2 i n so that

. .
,

3 '
8 P — P 8)
P and

39 8 ‘
x 28 7 4 °

538 x
0 07 1 3 lb per cu
. . ft .
,

because 2: 7 7 4 6 1 5 3 8 deg absolute Fah . .

From the table w e find that at 4 1 Fah it requires .

cu ft of vapour t o form 1 lb
. . .

1
a, lb per cu
. . ft .
,

so that the to tal weight of 1 cu ft of air is

. .

-
0 07 l 7
°

lb .
pe r cu ft
. .
16 CE N T RI F UG AL F AN S
SAT U RATI ON P RESS U RES AN D V OL U M ES OF AQ U EO S U
V A P OU R .

Sa tu r a ti o n pr e ssu re . tu r tio n o l u m
Sa a v e .

T e m p e r a tu r e Fall .
Inch es of m e r c u ry . No
. of cu . f t p e r lb
. .

1
Pg .

32 0 18 1 1 3 3 90
41 02 5 7 2 2 4 06
50 0 3 6 08 17 32
59 05 000 1 264
68 06 8 4 6 935
77 0 9 27 9 69 9
86 1 2 4 20 5 29
95 1 6 47 0 4 05
1 04 2 1 62 0 3 13
1 13 2 8 110 244
122 3 6 2 10 19 2
13 1 4 6 2 60 1524
1 40 5 8580 1 22
14 9 7 3 5 80 98 45
158 9 17 70 8 0 02
1 67 1 1 3 600 65 4 7
176 1 3 9 600 53 92

C H AP TER III .

Law f
o Change f
o Mom ent o
f Momof the
en tu m — On e

most important mechanical laws that applies to the fan is

that the change of moment of the momentum of a mass
acted upon by forces is equal to the moment of the impulse
of the external forces or to their angular impulse I f the .

weight of a body is W and its velo city is 12 its momentum

,

, ,

is Y
V v
and if r is the perpendicular from any point A
—
, ,
9
fig 6 upon the line indicating the direction of motion
.
,
,
18 CE N TR IF UG AL F AN S
corresponding radii i n feet L e t c be the absolute velocity .
2
of discharge and c1 that of inflow j ust before the vanes act
,

upon each particle while c is the absolute velocity i m m e d i

, a

ately after further let u1 and n be the velocities of inflow

, ,
an d outflow relative to the wheel Th e parallelograms .

of velocity of outflow and i nflow are L M N B and A E F G .

I f a1 and a are the tangential components of c1 and c2 at

2
entry and exi t from t h e wheel and 8 w is the weigh t of ,

8 10
the element of ai r under con s1d e r at 10n I t follows a r
2 , z
9
is the angular momentum of all forces acting on the ele
ment since its moment of momentum is z ero before reach
,

8
ing the wheel and 1 9 a r 2 after leavi n g the wheel I f
, ,
.
,

9
then is the angular velocity of the wheel in radians and
(0

W is the total weight passing through the wheel per sec

, ,

,
.

Wa r w = T w
2 2 ,

g
Wa 2’ =Tw
?
2 2

work per sec transmitted to t h e wheel if T to tal

.

twisti n g moment in lb ft H ence . .

,

a v + g = w o r k done by wheel per lb of air ( )

1 2 .
2 2

and neglecting the friction of the bearings which is never ,

a very great quantity this is the work done per lb Of ,

.

air on the fan Shaft however the air may approach the
,

fan For if no force acts on the air before it reaches the

.

fan it can have n o moment of momentum and therefore ,

must approach the fan radially or if inflow is axial , , ,

axially H ence the work done by the wheel and there

.
,

fore that done on the shaft by the mot or must be a2 02 + 9 -

,

but if the friction of shaft or arms acts on the air and gives
it angular momentum before reaching the wheel then ,

a u + y is the work done by the wheel vanes arms and

2 z , ,

the friction of the shaft upon the air or is the useful work ,

done by the motor on the shaft T hus equation ( 1 2 ) not .

only applies to fans in which the flow through the wheel

L OS S ES OF H E AD WH I L E P A SS I N G T H ROUG H TH E F AN 19

is wholly in a plane perpendicular to the a x is of the shaft ,

or r ad ial flo w fans but also to those in which the flow is

,

changed from an axial direction to a radial direction or ,

m i r e d f low fan s.

Los ses o f En er gy or Head whi le p a ssi ng thr ough the Fan .

In pas sing through a fan there are several losses of head ,

FI G . 7 .

which by proper design may either be entirely avoided or

reduced to a minimum when the orifice or
E
L
Jg H
has the value for which the fan has been designed where ,

Q cu ft or metres per sec H head of air in ft or metres

. . .
, .

given by formula ( 10) or and g = 3 2 2 for B ritish °

units or 9 8 1 for metric units Fi g 8 is an outline draw

,
. .

ing of a fan showing two sectional elevations I n some .

fans there are two eyes A at which the air enters ; i n

, ,

others one eye only is employed Th e air passes from .

2O CE N TR I F UG AL FAN S
the eye through the wheel B which rotates clock ways , ,

into the diff user C This di ff user is sometimes made with

.

parallel sides but Often with a slight taper of about 7 deg

,
.

I ts inner surface is always cylindrical but its ou t er sur ,

face is sometimes Of a spiral form as in the Rateau venti ,

lator t o be described later I t is however usually cyli n

.
, ,

d r i cal. Most fans are constructed without any diff user at

all and the wheel B disc h arges directly into the volute D
, ,

w hich is usually of rectangular or circular section i n cr e as

ing from the beak E according to a formula to be deal t
,

with later and having its greatest section at the base of

,

the chimney F Th e latter gradually increases in section

.

as shown so as to reduce the velocity of discharge Re

,
.

ferring to fig 7 w e see that when th e air enters the wheel

.
,

at A its direction may h e suddenly changed from c1 to

c
a, so that the loss Of head is
2
_ ( K F)
LI
29

where b is the radial component of the absolute speed

l
immediately after entry to the wheel and 6 is the angle ,

made by the vane A B at A with a tangent to the circle ,

through A I n order that this loss of he ad may be avoided

.
,

w e must make 6 such that

which we shall presently Show can only be the case for one
, ,

value of
Q — '

JgH
A fter passing t hrough the wheel the air enters either the
atmosphere if the fan has no casing the diff user or in the , ,

case of a fan without any diff user it passes directly into the
volute ; in the first case the head lost is c + 2 g as the 2
2
,

kinetic energy at discharge is all lost in the second case

LOSS ES OF H EAD WH I L E PA SS I N G H R OUG H
T THE F AN 21

there is no loss when entering the diff user whilst in the ,

third the loss is ( fig 9 )

.

bi
< 13 )

where a is the velocity in the volute which has a direction

,

very nearly tangential to the wheel I f the fan has no .

FIG . 8 .

chimney or expa nding discharge pipe c should = § a2 so , ,

as t o obtain a maximum gai n of pressure head ; but if it

has then a should a giving a m inimum shock loss for
,

the reasons given i n Chap I equations ( 4 ) and

, ,

.
,
If
the fan has a diff user and B D is its inner and C E its
,

oute r circumference ( fig the latter having a radius r

.
3,
then the change of the angul r momentum of each part icle
a

of air therein is nil b ecause no force acts on it during its

,
22 CE N TR I F UG AL F AN S
passage through the diff u ser Further if as b3 are the .
, ,

tangential and radial components at discharge and 3 2 3 3 , ,

the breadths of di ff user at inflow and discharge then ,

“ a
2 3

a r
3 2
a r
( 14)
2 3

and 2 -
7z r s
2 2
b2 ==2 7r r
3
3
3
b3

be 32
“5 ’
as bi

and if the sides of the diff user are parallel or ,

b3 ’2
as

This ,however neglects the thickness of the vanes the

,

path B C is then an equi angular spiral But 3 is usually

-
.
3
made slightly greater than 32 so that the sides are inclined
,

to one another at an angle of about 7 deg .

Th e air next passes into the volute and the loss Of ,

head is
bf ( a3
L3
2 9
( 17 )

and if there is an expanding discharge pipe we should

have c = a3 and if n ot c = % a3 according to C hap
v , , v ,
.

equation s ( 4 ) and I n addition to the above there I s

the loss due t o surface friction and bends which may be
written
2 2
u c,
~
( 18 )
2
L 4 = F1 + F2
gg 29
2
where F1 accounts for friction and bends In the wheel ,

and E2
39 the loss by friction i n the di ff user and friction “
 E FF I C I E N CY EQ U A T I ON S 23

and bends in the volute There is also the loss of head

.

due to the kinetic energy con tained by the air at discharge .

2
Cd
’
2 0

where cd is the speed in the discharge pipe .

B c If P
FI G . 9 .

T he values o f c in terms of
, a
2
and a
3
can only be Oh

t ai n e d at one orifice and make

,
2

2
I) 2 4
L2
29
, or 2g
19)

2
b
and L3 = i
b ?
or
9
( 2 0)
g g
Equa ti on f or fi nd i n g the Man om etr i c and Mechani cal
Eflici en ci es i n Ter ms of ( 1) — Th e work done by
th e wheel
per lb of air delivered is equal to the head H given t o
.

the air together with the work absorbed by losses of head ;

,

hence
losses of head H .

Case A .
—I n the case of a f an without any cas m g these
losses include
( 1 ) Shock loss at entry to the wheel .

( 2) L osses by friction and bends i n pas sing through

the wheel .

( 3 ) T h e leavi ng loss depending on the absolute spe ed

from the wheel .

24 CE N TR I F UG AL F AN S
Thus we have
a n c
2
u
f ( v1 -
b1 cot
2
H ( n
’

z 2
F1 2
a 2g 29 2r

1being t h e angle between tangent to t h e curve A B at B

< ; ,

and the circle through B measured clockwise ( fig In ,

.

some fans l is greater than 90 d e g so that cot I is

c) ,
( )

negative .

e z
z
a
2
b 2
b2
2
b 2
2
cot p)
2
2
(

u
,
2
b2
2
cosec 2‘

I b 2 ( 1 cot 2
( )
2

.
7
and vl v
2

so that it is evident that the equation f or a given fan can

be thrown into the form ,

n
2
2
+P v
2 Q — R Q
2 —
Sg H= O ( 22 )
where P R and S are constants containing F I and 6 of
, , ,
( ) , ,

which R and S are positive and P may be positive or

negative .

C ase B I f the fan h as a di ff user and no volu t e then

.
-
,

the losses comprise

hock loss at entry to the wheel
( 1) S .

( )
2 osses by friction and bends in passing through
L
the wheel .

( )
3 F riction loss i n the diff user .

( 4 ) L eaving loss depending on the absol ute speed

from the di ff user .

T hus we have
a. n
, 2 ( v1 b1 cot
a 2 9 2 9
26 CE N TRI F UG AL F AN S
angles as the discharge becomes small Th e reader must .

remember that 6 and d are the angles made with the tan
)

gents to the inner and outer circumferences of the wheel ,

by the mean direction of flow to and from each passage

( fig . I t is such points as these which account for the
diff erence usually of small magnitude however which is
, , ,

found between a series of experimental results and a

theory which is not absolutely correct owing to the di ffi
, ,

culty encountered i n expressing mathematically the exact

variation of the quantities i nvolved
We shall now consider the eff ect of varying 1 in equations
.

< )

( 2 1 ) to ( 2 5) on the mechanical efficiency of the fan and on ,

the manometric eff iciency

N 9 H
ut 25a—

" 2
2 ( )

Th e mechanical efficiency does not diff er much from the

ratio of the useful work done per lb H to the work done .

by the wheel per lb a2 122 + g Th e diff erence i s due to

. .

the friction Of the wheel bearings and to the friction of the ,

outside casing of the wheel with the air which w e have not ,

included in the above equations and which is Only notice

able at very small orifices We shall therefore find the
,

ratio which we shall call the air efficiency

, ,

— 9 H m
n

and we shall throughout take

52 J gTI ( 27 )
when b1 cot 6
— that is when the air glides on to the wheel vanes without
,

S hock loss .

Equation ( 1)
2 thus becomes
2 2
'
0 b2 l
CO t ( ib b2 w t b2 e o se c i
( )
2
F1 H
94
3 9 29
9
 E FFI C I E N C Y E Q U AT I ON S 27

2 e
2
2 —
2 v2 b2 cot t
< i —b 2
2
— '
0
2
2
+ 2 b2 0 c ot < >
’
2 l — b2 2
c0 t
2
¢

from which
v
2
2
b2 2
(1 cot 2
( t) b2 F12
cosec 2 2 g H ,

and 0
2
2
b2 2
( 1 + F 1)
cosec 2
d ) : 2
g H;
whence by equation ( 2 7 )
@
2
2 —
i ( 1 + F) cosec 2
¢g H = 2
g H .

Eliminating v
2

Nm
1 gH 4
m 2
8 Fl ) cose c2 < f> ( 8)
2

also the air efficiency

7
g H
)
v
2 ( '
v
2
b2 cot qb) ’

w h ich from equation ( 2 7 ) becomes

(m cot
( 29 )
II

C onsidering next case B in which th ere i s a diff user but no

, ,

volute let r the external radius of the di ff user be h r 2 and

, 3 ,

suppose t h sides parallel to on e another ; then from equa

e

tion ( 2 3 )

since the shock loss at entry to the wheel

01 bl cot 6 ’

0
2 9

N ow , in the case of a parallel di ff user

a r
3
s
r and b3
r b
(1
2
r
2 2
.
3 2 ;

and b3 112
28 CE N TRI F UG AL F AN S
H ence we get

2 v
2
2
2 '
0
2
b2 cot l
c) Fl (22 2
cosec 2 ( )
I
2 b2 00 ‘ ¢ + bg
2
Co t
2
26

1 Fz
2
’

2
v
2
2 1 —
2

1 F?
cosec
2 H
2
b2 2
I
( ) F1 2
2g ,

and putting we
1 F2 1
;
'

2_ 9 H 1_
2
.
0
2 2
co t c
p

1 + F2
gH
z =0
— F1 cos e c l
< >
( )
3 0
k 2

Giving 2
as a function
'
0 of H whence substituting for ,

o
f i n w e obtain an expression for manometric
e ff iciency i n terms of the variable I ( ) .

I f the fan has a volute and chimney but no di ff user and , ,

we suppose that c = a then equation ( 24) may be put inv 2,

the form
2 e
2
2
2 v
2
b2 cot < )
1 F1 b2 2
cosec 2
l
c) b2 2
F2 ( n
2
b2 cot
g H
z 9
cd .

Le t cd =
§ J g H . Then , since b2 = § x/ g H ,

e
2
2
( 2 — F2 ) -
F2 )

F2
—
gH cosec 2
4) cot 0 ( )
3 1
4
 E F F1C I EN C Y EQ U A T I ON S 29

Thus w e find in terms of H and subsequently

"
ca n v
2 ,

N , in terms of I ( ) .

I f on the other hand cv =l a

2,
and there is no chimney ,

then equation ( 2 4) becomes

2 1
2
2 —
2 0
2
“ WT — Fl b2 2
COS BC
Z
‘ P 00 b ( N
Z

i 00 17 —
1( v
2
— b2 co t

since cd is now the velocity of discharge from the volute ,

and therefore
:
4( v
2
co t

— v
2
bz co t d) 1

— 62 2 =2
9 H,

and putting g
E
13 G ol i)
g
2
v2 v2 1

i gH + F1 ( 2)
" 3

A quadratic in v2 which ultimate ly gives in terms

of I by the application of equation ( 2 5 a)
( ) .

I f there is a diff user volute and chimney and if r 3 2 , , , ,

b
so t h at a = and
%
then from equation
a —z
3
b3 ( )
25 s
,
k

2 v2 2
2 v2 b2 cot d ) F1 b2 2
cosec f 2
( )

ur 9 H_ 2
2
b
_ . F2 emi a ?
gH ’
2 2
64

S ince and ed : JgH

12
1 b2
o
f 2 — — 2 v2 cot ¢ 1
k g

g
2
1 CO t E
-
b2 2 -

2
+ F 1 cos e c
2
¢ + FZ — 2g H = 0,
16 g
3O CE N TR I F UG AL F AN S
and putting b2 AJ n
we obtain 2 — v
2 Al y H cot qt 1
k 2

1 Fl cosec 2
c)
l F2 cot 2
4) 9

Which gives 12 in terms of l and H H ence we can find

2
( ) .

the manometric effi ciency i n terms of e .

Th e values of F F that agree best with practice are

2
1 ,

both S ubstitut i ng this in we Obtain

222
2 — 02 8 1 c o se c
2 < =
1 2g
> H ( 4)
3

from which th e following table is calculated

=
a 15 v
2
= 2 49 °

JgH Nm = O16 -

30 0 32
45
90 04 4

so that the effi ciency of an Open running fan without dif -

fuser or volute is very low unless the vanes are curved

backwards and it must be remembered that t o get the
,

mechanical effi ciency of en g ine and fan w e m ust multi

ply 7 by 0 8 5 on the average so that the greatest m e ch n i
7 ,
a

cal e ffi ci e n cy of the combination is 0 5 6 per cent Taking .

next the case of the open running fan with a diff user whose
-

sides are parallel and whose e x ternal radius is 1 % that of

th e fan that is

~
,

=1
r
3 3 ,
r 2,

equation ( 3 0) becomes
v2
2
122 /g H cot ( I)
( 3 5)
which gives the following table
6 5 11 -
17
30 072
45
90
 EF FI C I EN C Y EQ U ATI ON S 31

so that the highest possible mechanical e fli ci e n cy of engine

an d fan would be 6 1 per cent I t must clearly be under .

stood that i is not the vane angle but the angle of flow
c> , ,

and this will probably be from 1 5 deg to 3 0 deg greater .

than the vane angle When the fan has no di ff user but a
.

volu t e and chimney and the for mer of such a section that
,

c =a
v 2
then ( )
3 1
,
becomes
1 87 5
-
0 87 5
-
cot a02 6 511
2

gH 2 =0 ( 3 6)
—

from which we obtain

(
b : 24 3 #9 11 N HL :
30
45
90
1 20 08 0
1 35 0 78

giving a max imum effi ci ency of engine and fan of 7 0 pe r

cent .

I f however there is no chimney but the air is dis

, , ,

charged direct from the volute and a

2
then ( ) be
3 2 , ,

comes
0
2
12 cot 1 ’ 2
< )

(37)
which gives us the following table
c = 2 45
,
-
/g H
.

30 0 36 0 74 -

45 047 -
0 71
90 0 65
-
0 65-

giving a maximum efficiency of engine and fan Of 63 p e r

cent .

Finally if there are diff user volute and chimney

, , , ,

122
2
cot av2 J g H
( )
3 8
32 C E TR I F N UG AL F AN S
and we get the following table
= 2 41 JgH Nm = 0 172 o q-

q
a : 15 v
2
° l

o
30 1 59-
9 o
45 1 35
-
¢ m
60 1 22
-
9 4 O
O

90 10 7 ¢ O4
C

1 20 094 -
¢ O4
C

135 0 87-
9
O
O#
:
giving a maximum efficiency of engine and fan of 7 4 per
cent .

A fan designed i n this manner would require a very lo n g

chimney as the taper of a chimney cannot be very great in
, ,

order to reduce the velocity of the air to so low a value as

We shall therefore consider the case in which the
'

external radius of the diff user is 1 % that of the wheel and ,

the velocity 0 in the volute is a3 Equation ( 2 5 ) then

1, .

bec omes when w e put

,

v 61 cot 6 a
I
( )

1 , s ,

and Mm ,

v
2 —
2 g oz Jc I
ot ( )

cot2 ( 3 8 a
)
from which we obtain the following table
"
— Jg H Nm = 0 17 :
J

15 N
'

F n

—
aO 0 85
I
s
3O
H
45 r wU I

90 0 86
1 20
—
135 0 82
34 CE N TR I F UG AL FAN S
because if A A etc are t h e sectio n s of passages in a
I , 2,
.
,

m ach n e th rough w h i ch a liquid flows and V V

i etc t h e I 2,
.
, , ,

correspond i n g mean velocities of flow then ,

A 1 V 1 = A 2 V 2 = A 3 V3 = ( 4 1)
—

A g ain , if when b 2 = 3z1~J g H,

c
d
1
I b2 ’

so t h at ( 2 4 ) becomes
2
2 12
2
9

33 f
1
b -

cosec 2
1
( )
(C 2 62 cot < )
1 b2 ) 2
]
0 28 9 62 2
-
0 06 25 by
-
2g H .

$02
2
+ ( 3 04 cot 6) 52
gH
2
= 7
( 4 2)
‘

So that
1 b2 1
N an
+2 8
3 04 7
3
2
CO t6 . _
H m
'

,
b 2
— _
g( cot 2
cot
P utti n g qt : 3 0, cot this becomes
1 b Z) 2
+ 4 23
-
1 — — 2 =0
Nm x /g H x /\ mi ,
4 ’

and th us

~
b2 52

/N m 2

—
b2
“ v2
9
2 t /fl
'

f
'
m (43 )

F rom we can find m and thence 7 so

v h i ch ,
7
,

Table 2 is readily obtained ; only it m ust be remembered

that for small orifices the fan efficiency is really l ess
 T H EO R ETI C AL C H ARA CT E RIST I C S 35

than 9 H + a v2 owi ng to external friction between the

2 ,

o ut e r surface of the wheel casing and the air or if the , ,

wheel i s o pen at the sides betwee n the fan casing and the ,

air which friction :C i n any fan and the work wasted

,
0
2
2
,

cc 12 I f the e fli ci e n cy were equal to g H + a 0 then at

2
9
.
2 2,
z ero ori fi ce i t wou ld be equal to N Whereas its real value m ,

is zero .

T A BL E 2 V ARI ATI ON
.
—

E FF I C I E N C Y WI T H
or
v
2 + ~
MANC METRI C
/
g II AN D (22 + v g
A N_D_ A I R
H .

ba
— Nm
y tr
\f
.

04 6 0 86
'

08 02 1
04 6
0 07 4
00 43

If t
c) then when there is no shock at i nflow
, ,

v
2 J g H and therefore
,

~
v1
—
C Ot
1

also at th i s orifice : / H;
(1
2
v2 g

c” : b2 : I) 2 at all orifices ,

52
and ed :
a ,
36 CE NTRIF U G AL F ANS
so that ( 24) becomes
2 v, ”
3
» - '
v2 b2 cot 0+ b 2 2 co t ? — b2 2 cose cz t
cr

[ 2 + (v2
b2 -
O 6 16 b2
- ” - 23 2
29 H
16
.

62 — 2 g H=o
§ v2
2
+ 4 93 °
7 2 8 62 2—

'
02 b2 b2 2 —
%9 H = 0

l 2
2
t h at I s, + 5 54 8 19 o
gH
I
m

1
4
( )4

from which we obtain Table 3 .

T A BL E 3 — V A R I AT I ON 0F A131
3 A I R
:9 3
.

E FF I C I EN C Y WI T H b2 “ AN D 02 + «’
g H .

d
i ff

0 86 0

0
 TH E O R ETI C AL C HARA CT E RISTI C S 37

If 1
< 2 then when there
= 1 35 , ,
is no shoc k at inflow ,

«TH and th erefore

,

co t 6 — 3 ’

b,
‘

1 5
-

Also at this ori fice

—
J é II T
,

so tha t cv = 2 °
8 2 l) 2 at all orifices and ,
ca = bz + 4, so that
( 2 4) beco mes
2 v2 2 — 2 v2 bz cot c
t —
§ v2 b2 c0 t 0+ b2 2
00 t
2

b2
b2 = 2g H
[ (
,
— 2 —
— 1
3
bz co s e cz < > I b2
2
+ v2 — l 82 b 2 ) J
' 2
8 l6
8 2
+ 6 043 bz 2 g H=o ;
-
122 a
,

v2
2
+ 6 80 vz b 2
'

2g
—— H = 0,

and thus

9 9
m: ( )
4 5
# 517;
fr om which Table 4 is calcula te d

~
.

T A BL E 4 — V ARI ATI ON
. 0 F MA N C METR I C AND AIR
E FF I C I E N C Y WI T H
’
V
62 + g H AN D 122 g H .

s/ H
g

1 5 00
1 0 45
O 9 15
0 9 11 0 78
1 0 15

02 8
38 CE N T R IF UG AL FAN S
H ence i n this type of fan there is a considerable advantage
in making the vanes curve forward at discharge that is —
,

i n making 1 equal to or greater than 9 0 deg I t f r e

( ) .

quently happens that the orifice at which a fan has to

work during the time it is in use is not constant ; for
, ,

e x ample owing to the enlargement of a mine more air

, ,

may be required at the same gauge N o w w ith vanes .

,

having 95 greater than 9 0 deg the manometric effi ciency .

,

between the or ifi ce at which .

9 H and that at
which 52 : 0 8 /g H is more nearly constant than when
'
x

deg Further this type of fan must have an

.
,

efficient chimney in which the gain of pressure head is

,

equal to the loss of velocity head I t will be seen later .

that this can only be obtained with a uniform flow and

a proper inclination of the sides of the C himney .

I f the fan has a di ff user its e fli ci e n cy at a given orifice

,

will be practically the same when deg as when .

f 3 0 d e g because the principal cause of di ff rence is the

< > = ,
e

loss of head in t he vol ute due to friction the eff ect of the
,

diff user being to reduce the loss of head owing to the r e

Where there is not space to
,

d uce d speed in the volute .

fit a good chimney on board ship ) t hen a fan must ,

be used i n which and l = 3 0 deg so that a high < > .

,

e fli ci e ncy may be obtai ned at a constant orifice the v ane

angl e at discharge might then be about 1 5 deg .

Th e reader may possibly imagi ne that Tables 3 and 4

might be altered by a change in the ratio b z bz This how l
—- .
,

ever is not t h e case because IJ ] cot 6 v, when 62

, , Mg H 2

hence the te rm t 1 cot 6 will be unchanged whatever

may be I n the above w e have assum ed that the external
.

radius of the wheel is t h r e e t i m e s the internal and a

.
.

change in this will slightly aff ec t the co e fli ci e n t s of v2 62 ‘ 2

,

and g H But what ver th e ratio r + r may be the real

. e 2 l ,

manometric efficiency at z ero orifice should not diff er much

from 1 because the air withi n the eye is really rotating

~
,

with the same angular velocity as that within the

fan and under these circumstances
, —
J2g H When .

b2 : 05 /g H the mechanical and manometric efficiencies

are entire ly independent of + 73 because vl b cot 6 = 0 , 1
.
 V O L U M ETR I C E FF I C I E N C Y 39

a fan were made very large i n

Volumetr ic E lli ci c ncy — . If
p p
r e o r tto its discharge it would give a hi g h mecha nical
i un
,

ef ciency owing to the fact t ha t the loss due to the friction

fi ,

of air in the passages would be extremely small ; but the

cost of suc h a fan would probably more tha n outweigh the
advantage thus obtained I t i s therefore necessary for
.
, ,

designers to have another ratio whic h is usually called the ,

volumetric e ffi ciency which compares the discharge with

,

the tip s peed and the di m ensions of t h e fan L e t Q be .

the discharge the t i p speed and the external radius

, ,

of the wheel ; then if the fan has one eye the v olum e t r i c
, ,

efficiency is
Q 4 6)
'
02 T2
2 (

and if it has two eyes one on ach side e

Q ( 46a)

A radial or mixed flow fan having a small diameter and

moderate breadth at o utflo w may ha v e a higher volumetric
e fficiency than one whic h has a large diameter and small
breadth although the circumferentia l area through which
,

the air flows may be t h e same i n both and they may be so ,

designed that Q z r z r 2 2 may be the sam e for both Fo r

——
.

this reason we consider t hat a better measure of the

v ol l m e t r l c e fli ci e n c
l
y is
N = Q v1
7 32
2

where 3 2 is t h e breadth of the wheel internally at the

e xternal radius .

I n the following p ges w e shall mean by volumetric

a

e fli i e n c
c
y t h e quantity N and shall denote the more
a,

correct value by the letter N 1 These e fficiencies can be v .

greater than u nity .

 40 C ENTRIF U G AL FANS

C HAP TER V .

D es i g n f Radi al F lo w Fans Fans

o may be divided into —

three classes , v i z Rad i al Flow Fans

those i n which the -

direction of flow through the wheel is perpendicular to the

ax i s .

M i xed F low Fan s those which have an axial inflow and

—

radial outflow .

Scr e w 07 P r ol eller Fans those i n which the flow is

'
i —

wholly axial
We shall now consider the desig n of the Ra dial Flow
.

Fan .

6
We shall first consider the value L
JgH
H eenan and Gilbert found that with a fan 1 7 in 5 .

diameter and 8 in broad the discharge w about .

'

as

cu ft of air per min at the most suitable orifice ;

. . .

the water gauge was i and consequently n .

,

b2 Q 3 4 00 x 1 4 4
’
2 -
72 7
2
9
2
7r x 17 x 8 x 60

x
and gH x 8 2 0,

being taken as th e mean relative density of the air at

an a v erage temperature and pressure
52
hence —
0 l34
'
.

I n Bryan paper a Rateau fan with an ex t e rnal

D o n ki n s
’ 6

radius of in has its maximum efficiency at an e q ui v a

lent orifice of sq ft T h e equivalen t orifice
. .

01 Q Q
1 088 ( )
4 8
/m
_

o ak N /g H
42 CE N TR I F U GAL FAN S
Le t be the radius of the eye then ,

em f x 3 3 2 = 1 000 = Q
. 2
5 00
7 0
33 2
'

1r

Th e internal radius of the vanes should be made slightly

larger than this say ft L e t the velocity of inflo w be
, .

Th e thickness of the vanes may be neglected

/
v ,

and w e have
2 7 1 s1 x J fi = 1 000 ;
-
72
'

Where s1 is the eff ective breadth of the wheel at radius r ,

3
1
ft .

lf t h e radial component at outflow b 2 = 0 5 J H and '

g ,

t h e v an e s are mad to touch the circumference at t h e outer

e

r ad l us then we ha e alr ad
o

, y shown
v that n 0 64 5 and
r 0 '
2

16 5 ft per s e e I f the revolutions N per

. .

m i n are fixed we can calculate

.
, from t h e formula

but if not let

,
2 r
,
5 ft . then
N :
6 0 x 1 65 :3 1 5 rev .
per min .
,
10 '
71

and the eff ective breadth at the outer radius

s2 ft .

I t must be remembered that if 52 does Jg H n ot = 0 5

'

another val ue of v2 must be deduced from equation

I n the case under consideration ,

f 1
vl =v
z
JgH;
 S CR E W 0 H P R o r E L LE R FANS 43

Th e brake horse power required to drive th e fan is

Q It 6 2 3 00 x 2
: 29 3,
131 1 '

l 2 x 5 50 X ?) .
12 x 5 50 x 0 645

so that the indicated horse power of the engine i s

B HP
34 5
. . .

N ext suppose our fan is provided with a d i fl us e r , the '

vane should be a tangent to the outer circumference so ,

that allowing for the divergence of the air

,
—
d e g v2 , ,

J g H N 3 1 2 and , ,

cot X 2 46 6 7 11 2 46 .

b1

may be made wi th its sides diverging from

Th e d i fluse r
one another about 7 d e g corresponding to an inclination ,

to the vertical of about 1 3 and its external radius is 1

,

Th e diff user may either rotate with the wheel or form part
of the casing the sides of the wheel should appear concave
on a rad ial section so that the air may have no motion
,

parallel to the axis when discharged from the wheel into

the d iflll s e r Th e bre adth of the diffuser at the external
'

radius is
ft .

Th e brake horse power is now

62 8 x 1 000 x 2 18 88
-

BHP ‘ ' ’ ‘
c)
‘ 7 6’
12 x 5 5 0 x 0 68 5 0 68 5

and I HP. . .

08 5

I f the fan has 11 6 diff user or chimney b ut a volute i n , ,

which J
2 a2 then as the highest e fficiency is obtained with
, ,
44 CE N TR I F UG AL FANS

"f
c)

2
: 30 de
g__
s/
g
t
,

H
h
,
e

N :
vane angle at discharge should b e 1 5 d eg
2 12 rev per min and cot 6 : . .
,
,

62 cot l = —
,
c>

/ H /
67 section o f the n Th e
'
x
x
g
volute is proportional t o the ang u ar distance from i ts beak '

E fig 8 because the discharge from the fan is uniform all

,
.
,

round the circumference and if it is this angular distance in ,

radians s the section and S is its section at discharge then

, , ,

3 8
2W

and S __Q 2 Q 2 000 _ o

6” a
2

corresponding to a diameter o f 6 9 ft Th e section h ow .

,

ever need not be circular 3 a rectangular form is frequently

,

used I n some cases the wheel discharges in t o the centre

.

of this section 3 in others the volute is wholly outside t h e ,

wheel I f the fan has a chimney then

. but i n this ,

case it is possible to obtain a high manometric efficiency

We shall consider the design in two cases :firstly when
.

deg 3 secondly when it is 1 2 0 deg I n the former

.
, .

case we believe that it is necessary to make the vane angle

,

about 1 05 d e g so that t h e average flow may be rad ial if

, ,

the number of vanes are few say six 3 but if twelve vanes ,

are u sed then the vanes m ay be radial at discharge I n

, .

either case w e may as sume v2 = 0 2 = 1 1 1 J 9 H N = I 4O

,
'

rev per min Th e section of the volute at discharge is

. .

S _ Q_ 1000
13 6 sq ft
. .
,
a
2

while the larger end of the chimney has a section

S1 — Q 1 000
1 24 Sq a .

Cd J9 H
BHP
. . . 23 3 -
 SC REW 0R P ROFE LLER FANS 45

BHP
and I HP . . .
. . .

27 4

co t 6=

If deg the vane angle should be 1 3 5 deg if

. .

there are six vanes and 1 20 deg if there are twelve ; ,

.

then v2 = 0 9 8 # g H N = 1 2 4 and cot

'

, ,

27 8, -
a, = 98 +
-

J g H 1 27
= 6 5H s
a233
9 sq ft S, = I 2 4
— —
11 9
- -
. . .

sq ft as be fore
. . .

We now come t o the case where there is a diff user ,

volute and chimney and shal l assume it : 1 2 0 d eg making

, , ( ) ,

the vane angle 1 3 5 deg if there are si x vanes and 1 2 0 deg .

, .

if t here are twelve The external radius r 3 of the d i fluse r

.

is t o be 1 } that of the wheel and therefore ,

r
2 9 } ft .

J 9 H N 11 9 cot 6 , BHP , . . .

I HP. . Th e breadth of the diff user

is a
s be fore ft .
, and si nce
—

a
2
0
2
b2 co t c) l 02 8 8 ) s/g H H .

=
g Mg H : 68 6
g =
3 1 228
-
a a x
3 ,
.

3
1 000
H ence S sq f t and the large end of the . .
,

chimney 1 24 sq ft as before . . .

I t is the practice of some fan mak ers to mak e the sides -

o f the wheel parallel because no doubt it is very much ,

ch e aper ; there must be however a considera b le loss of , ,

h ead at inflo w owi ng to eddy formation On the other .

hand there is less loss of head at outflow from the wheel

,

owing to 62 + 2 g being very m uch less than 6 —9 2 9

2
1
2
.

A nyone w h o has read the above designs will have no

 C E N T RI F UC A L FANS
difficulty i n calculating t h e dimensions of such a fan ; but

~
to be strictly accurate he m us t not assume any of the
’
27
2
tabulated values of ' but must fi nd them from equa
/g H
’
U
—2
tions ( 2 1 ) to substituti n g the correct value of ,

N /E fi
2
01
and i n these equations a term F 2 should b e added to the 3
6
left hand to represent the additional loss of head at inflow
and in the wheel .

D esig n of Mi xed flow Fans I n the previous designs we

-
.
—

h ave supposed that inflow takes place fro m both sides and ,

when a fan is forcing air i nto closed passages there is n o

reason w h y this shoul d not be the case but when the air .

is being drawn from passages it is more convenient to erect

the fan with in flow from one side when the area of the eye ,

must be made twice and consequently the radius ,

times as great as that of a fan taking air from both sides

, .

Th e ratio of the external to the internal radius of the

wheel cannot be rigidly fixed Th e vanes sei z e upon t h . e

FI G . 11 A SC W
.
— D E V E L OP M E N T OF RE .

air immediately it enters the eye of the fan altho ug h there ,

is a mouthpiece before the eye whose obj ec t is to mak e the

velocity of inflow as uniform as possible over the whole area .

Th e a n gle 6 must therefore vary with the radius because

.

f’
cot 6 _ _1
bl
a

where w angular velocity of w heel in radians per sec .

2 1r
x N, N being the number of rev per min . .

60
 D ES I G N OF M I X ED F LOW
-
FANS 47

But 7 varies from z ero at the centre to 7 1 at t h e outer

'

radius of the eye N o w if a spiral of pitch p fig 1 1 and

.
, ,
.
,

radius 2 be traced round a cylinder and t h e cylinder

7r

developed one turn of the spiral will give us th e line A B

, ,

whose inclination to a line perpendicular to the axis

a

gives us

N o w, if a number of these spirals are traced with diff erent

values of it is clear that the co t ange nt of the an gle
increases as and the surface generated by joining all
these spirals by lines perpendicular to the ax is is called a
helical or screw surface A straight line which passes .

through and is perpendicular to the axis and which rises a ,

distance p with uniform v elocity while it makes a complete

turn with uniform velocity will also trace a helical surface ,
.

A true helix is usually generated i ll this w ay but a surface ,

traced by a curve whose plane contain s the axis and

, ,

whose outer end is on a cylinder having this axis th is end ,

tracing a spiral on t he cylinder will also trace a surface , ,

any cylindrical section of which will give a spiral i n

whi ch

and t h e end of each vane of a mixed flo w fan must be -

a surface of this description I t should then guide the air .

round so that it may flow i n planes perpendicular to the

,

axis of the fan .

As the Ra teau fan is the best illustration of a mixed flow -

fan wi th which w e are acquainted w e shall now describe it ,

and give its proportion s ( figs 1 2 Th e fan centre .

,

consists of a cast iron wheel A upon which the vanes are

-

fixed A is formed by the revolution of the arc of a circle

.

about the ax is A t the eye it is conical and at the

.
,

periphery it is normal t o the axis I n small fans the vanes

’
.

are placed in the mould when A is cast and in larger si es , z

they are held to it by angle irons Th e number of vanes .

is 20 to 2 4 for small fans and 24 to 30 for large Th e ,

.
50 C E NTRIF U G AL FAN S
Th evanes are of wrought iron and ar e stamped in a ,

mould They are very rigid in con sequence of their curva

.

ture i n every direction which is such that the traj ectory of

,

every particle of fluid has an almost constant curvature .

T his is of great importance i n lessening eddies

Fi g 1 5 shows the geometrical construction of the vanes :
.

K O is the axis of the fan shaft and circles passing through ,

S P Q R have r ad ii equal to t h e external radius of the

, , ,

fan while P R is the bread t h of the vane at the tip Tak e

,
.

a point B on the circle S and through it draw arcs B e , ,

FIG . 14 .
-
PE R ~
FE C T I v E V I EW OF R A T EA U F A N .

BC ; the angle m ade by the arc B e with the tangent at

B is 4 5 deg for a fan of type A ( see Table
. which is
generally the angle made by the vane at the external
rad ius 7 2 with the tangent Th e rotation i n the figure is .

supposed to take place from B to the right so that if vanes ,

curved backwards are used the centre of B e would lie on

the opposi t e side Th e circular arc B C makes an angle
.

C B D with the line B D parallel to the axis such t hat if , ,

b , is the intended velocity of the air at normal speed and ,

v, is the peripheral speed at the e x treme radius then tan ,

 DE S I G N OF M I X ED F L OW
-
FANS 51

C BD and this will ensure that at any point on

the arc B e the air will meet the edge of the vane when the
fan is running at the orifice for which it is design ed I n .

t ra cing out the surface of the vane the ar e B e moves with ,

B on the arc B C and e o n the cylindrical surface with

, ,

FI O . 15 .
— G E OM E T R I C A L CO N S T R U C T I O N OF V A N Es .

:
e aas radius and K O as axis 3 the whole of this surface is ,

however not required for the vane I magine b d t o

, . a c

sweep round the a xis K O and to cut out a portion 3 this

forms the vane d being the part near the eye and b c that
, a
 52 C E NTRIF U G AL FANS
near the periphery while K ( 17 forms the cone at the eye
, .

This construct ion may therefore be used for any type O f

-

mixed flo w vane whatever the angles 6 and f ha v e to be

-

,
c) .

I n Table 5 the proportions of these fans are given and ,

i n fig 1 4 is show n a perspective view of a Rateau fan with

.
,

part of the casing removed to s how the vanes .

To design a fan O f this type for a water gauge of 8 in and .

cu ft per mm we shall assume that the veloci t ies

. . .
,

of flow 172 are 0 5 J g H thus '

—
,

SO that the area 0 O f the suction orifice (Table 1 )

0
Q 3 000 4 5 2 sq f t
— —
b
.
.
,
1

ft .
, and the radius of cone ft .
,

p so that the external diameter of the fan is ft .

Th e vanes are curved forwards so that d e g and , ,

1 2 vanes are used so that we may assume the angle of flow

,

practically coincides with the angle Of the vane .

What is the best number of vanes for a given fan is a

question that we are unable to answer as no experiments ,

have been published upon this subject and designers d i fle r ,

widely on t h e poi nt Up to a certain limit the greater the

.
,

number of van es the higher the water gauge but a point ,

must be reached at which additional surface friction not

only reduces the e fli ci e n cy but reduces the gauge , .

I t will be noticed that the width of the vane at outflow ,

which we denote by 3 is so that the cylindrical

2,
surface through which the air leaves the fan at radius is
which is very little greater than Th e
final depth of th e di ff user Spiral D is f t so that the .
,

radius at that point is ft : 7 3 and as the tip speed of .

,
-

the wheel from ,

N : 1 65 rev per mi n . .

a : H : 1 8 2 ft per sec . .
,
2

7
'
182
r1
3
= a
2
— 2
121 ft per sec
. .
,
 P ROP ORT I ON S OF RAT EA U FAN S 53

TA B E 5 L .
— P ROP ORTI ON S OF RATEA U FA N S .

( Th e r a d i us o f t h e f a n is de n o t ed by

T Y P ES .

B . C .

Fo r l rga e m l‘‘m e s
Fo r m od r te e a Fo r m ll s a
wd S m a ll
f volum d e s an v o l um d e s an
p
g
ggi gz
l u e ee
p r ur e ss es . h i gh p r ss u r e es .

S U CT I O N O R I F I C E
Ex te rn a l r ad i u s 7
Rad i us o f c o n e p 0
Ar ea Of s u c t i o n o r i fic e 0
VAN ES
[n i t i l i n c li n ti o n t t h
a a a e

c i cu m f r n c
r e e f th e o e
u c t i o n o ri fi c wi th t h
s e e
° °
39 45 45
Fi n l i n c li n t i o n
a at th a e
c i rc u m f r n c e f th
e e o e
f w i th t h r o d ° °
an e s a. 60 60
D E PT H OF V A N E
At i n flo w 3,
A t ou t flo w 32 0 05 4
'

D I FF U SE R
Wid th s
3
0 05
D E PT H OF S P I R A L
I n i t i al d
Fi n a l D 05 0 7 2
I n c li n a t i o n Of on e f ac e to
a n oth e r
° ° ° ° ° °
7 4 3 4 3 3 2

9 2 9 2
' '

VO L U T E
Fo rm u la i i
g v n g t h e se c
i i
t o n s as f u n c t o n s Of t h e
ar e o f t h e s p r a a r
’
i l : 3 a:
(1 0 +x ) 0 0 4 0x ( 1
'

12 x )0

:
C HI M N EY
i
He g h t L 5 to 7 5 to 7 5 to 7
An g l e ofc hi n ym e 6
°
5
°

p l
Th e e n d o f t h e s i r a i s a t a i s a e f r om t h e d t nc
i w i on r uis a ad i s po nt h ch
pr
.

n t h rou h
pas s i g g t he or g
i i n of t h e S i al
:
.

r to
f a is t he a i o f t h e e g l n th r l o
of t he ar e of t h e s i a o r g n to
fr m the i i t h e p om t
at w ih ch t on
t h e Se c i c l c u l t d to
is to b e a tot l
t he p
e ng t h o f t h e s i r al A
:
a e a
p n t h rou gh
.

2 m ay b e c l cu l t d
a a e as a f unc t on
i n gl
of t he a t n
e to b e w ee t wo r ad i i as s i g
t he or g n p l
i i o f t h e s i r a , an d t h e
b y t h f o ll o wi g f o rm u la : 5
e n
4 80
po nt
i prl h r
t h e a n gl
" to
o f t h e S i a w e e t h e se c i n i sto c l cu l t d
»b i n g e p re s d i d e gre e
be a a e

e I e x se n s.

1 Ca l c ula t d f ro m t h e d f t h e sp i l
e e n o ra .
54 CE N TR I F UG AL FANS
and according to our theory this is the velocity in the
, ,

volute for a f an with a very long chimney which will

reduce the vel ocity to a low value H ence following our .
,

theory the section of the volute at discharge

,

3 000
S
121
sq ft
. .
3

while according to Table

, 5,

s :
:
1
( 1 x
)
and 1 when 3 S x x sq ft Accord . .

I ng to our theory the section of the chimney would be

Q 3 000
S1 _ _ 18 0 sq f t ,
% f 1 6
. .

v fi 6

and assuming the section of volute and chimney square

, ,

the chimney would require to be long enough to alter the

side of the square from ft to f t which with a . .
, ,

total taper of 7 d e g would require a chimney whose height

,

w as

L
tan 69 f t ,
nearly
7
.
0

i nstead of 5 to 7 times ( Table I n this case the brake

horse power
x 8 x 1 80000
12 x 3 3 000 x
270
I HP
. . . 3 18 .

I f however we assume a diff user in which

, ,
and
that c = % a3 the chimney will be greatly reduced in
v ,

height Th e values O f are the same as before but

~
. , ,

from ( 3 8a) ,

N : 1 6 7 rev per min . .

—
a : / H= 18 3 ft per sec
2 g . .
,

0 4 6 V g H = 6 1 ft per sec '

. .
,
 P R O PO RTI O NS OF RAT EA U FANS 55

hence the section of the volute at discharge

3 000
T
= 49 2 sq . ft .
,

which would form a square whose side is ft .

,
so that
the length of the chimney would require to be
13 4 2 6 4
L ° {t
t an 7
'

1 8 0000 X 8 x
H ere B HP 2 7 7, I HP. . . 3 27 .

3 3 000 x 1 2 08 2
. .

I n the first case t h e angle 6 that the vanes at the outer

,

radius O f the eye make with a ta n gent to the direction of

motion wi t h a line pe rpendicular t o both radius and
ax is is given by
—

772 06 x
( 30 5 6= = — —

ZI
7
06 l
g II
‘

an d i n the second ,

06 x 08 8
cot 6: = 1 06,

so that 6 is as nearly as possible 45 deg .

The or y o f the Sp ir a be the

l D ifl user ar i d Volute .
— Le t r 3

radius of the diff user at an angle t d from its commencement l ,

and let
r
a
=r
a + W +a d

where p and q are constants .

a
Then s
7
0
2 +0 6 a

Since the moment of a particle of air i n

Of \ m o m e n t um
i t s passage through the di ff user is constant that is the —
,

veloc ity at any point varies i nversely as the radius so that ,

the curve O f flo w is an e qui an g u ar spiral ‘

.
56 C E NTRIF U G A L FANS
L et A, fig 16 be the point where the particle leaves the
.
,

wheel and let the radius

,
at this point make an angle
3b with the radius passing through the point of commence
,,

ment of the diff user Further let B be the point where

.
,

the particle enters the volute so that A B is an equiangular ,

spiral L e t A B subtend an angle at C 3 then if r is the

. a.

radius at any poin t X on A B and 8 r the radius afte r ,

the particle has moved a distance corresponding t o 8 s along

a circular arc subtending 8 at C it follows tha t
,
a ,

4
'
7
ar d 3 di '

and 80 . . T hus, da
r
'
7
2

“ r a
a
2
log a_ _e
log T
2

FI G . 16 .
— V E L OC I T Y IN TH E V O L U TE .

Let be the section of the volute m ad e by a plane

s

ing through the axis and C B Then the section 8 has a .

quantity O f air passing through it per sec represented by .

,

the quantity
58 CE N TR IF UGAL FAN S
If B is
the i nclination of th e relative direction of flow i n
t h e wheel with the direction of motion of the wheel at the
,

radius then the above head

2
o -
o b co t 8 a
2
+ l 2

29
H
l
- -

2é(
2
v — b
‘’
-
cosec2 ,8 — 2 9 H) ,

where b and c are the radial component and absolute

velocity of flo w at radius r corresponding to 62 and c2 ,

( fig 7 ) at radius r 2
. .

I n the above consideration friction is neglected and i n ,

w
flo is supposed to take place without shock
rad i us r 2 the pressure head becomes
A t the outer .

1
29
1722 cosec? 2g H) .

Th e actual pressure in lb per sq ft may be Obtained

. . .

from the head b y mul t iplying it by 00 7 5 Th e total pres .

sure Ou the wheel i n feet of air is therefore

B 9 H) d r
’ 2
-
b co s e 0 ,
— 2

in radial flow fans and since b and 8 all i nvolve r the

, , , ,

total pressure is best found by dividing the wheel into a

number of small rings so that r d r can be ob t ained for 77
'

each and ,

y b? cosec 8 2 g H )
a— 2
e ,

can be calculated at the middle of the ring whose mean

radius is r and thickness d r T his will give a result which
.

will be practically correct .

Th e above ass u mes that the wheel although there is ,

only one eye through which air enters has an Opening O f ,

equal diameter at the other side I t also assu m es that the .

disc of the wheel is on on e Side so that there is ax ia l

thrust 3 by making the wheel with a disc on each side end ,

thrust can be e n t i r e lv avoided and the end thrust due to,

 V AR I ATI ON OF P RESS U R E IN A CE N TR I FUG AL FA N 59

suction at the eye comes on th e fan casing and n o t on the

wheel I f however we are dealing with a m i x ed flo w fan
.
, ,
-

w e must proceed somewhat di ff erently Our explanation .

will b e clearly unders to od by reference to fig 1 3 \ve . .

want t o find the axial component of the pressures on the

cone a f a and the curved surfaces a b a b ( neglecting vane ,

thicknesses ) .

L e t r 1 = r 0 be the r adius of the eye then at the point f ,

the head is
-
H
2g

but as we get further to the left on the surface O f the cone

the velocity of the air increases and the section O f passage,

should be calculated perpendicular to the mean direction of

the flow shown by the arrows .

L e t A be this section then the head is,

2
1
H
29

F urther, we may suppose the cone divided up into a num

ber of rings,and a central circle and areas projected on to a
plane pe rpendicular to the ax is Th e total axial pressure .

on the wheel is then given by the summation O f these

areas and their corresponding pressures which is ,

6 2
cos e cz fi 2 9 H) d r ,

where 7 is the radius Of the cone and b cosec B is the

'

o
, ,

velocity of flow of the air relative to the wheel Th e latter .

is most easily estimated by drawing lines perpendicular to

the arrow representing the mean direction of flo w and
, ,

calcu lating the areas of the frustra O f cones swept out by

these Then if a is the area swept out by one of these lines
.
, ,

b= Q
9
a

and B must be measured from a drawing of the vanes if it

cannot be calculated .
60 C ENTRIF U G AL FANS
T his pressure can be balanced by putting straigh t radial
vanes at the back of the wheel 3 the eff ect of the vanes is to
set the air in motion and thereby reduce the pressure I f .

there are no vanes at t h e back a pressure corresponding to ,

that exerted at the circumference will act over the whole

of the back of the fan We know O f no case in which fans
.

have been thus balanced but Show in fig 1 7 a sectional

,

elevation of a centrifugal pump bala nced i n this manner 7 .

Th e holes shown at 0 are u nnecessary .

I m agine a space such as that at the back of the wheel

, ,

in which a number O f radial vanes cause every particle of

FIG . 17 .
— BA L AN C ED C E N T R I F U G A L P U M P .

air torotate at th e same angular velocity C onsider a ring .

o f air of m ean radius r and radial thickness d r and let the

, ,

p ressure outsid e be p + 8 p and that inside the ring

, 7) ,

and take the axial width as unity I f is the weight of . a

1 cu ft of air the weight of a s t rip of the ring of length l

. .
,

is l 8 and the centrifugal f or ce due to its rotation is

0
'

"
2
v

R e7
and th e component normal to the diameter X Y is
~ 2
8F = o l8r
g sin 6
— .

}
 V A RIAT I O N OF P R E SS U R E I N A C E NTRI FU G AL F AN 61

resultant force due to half the ring is perpendicular

Th e to
X Y and ,

1 80 2 e
w ,
2
=F — =2 a 8
gr 7
.

Th e resul tant air pressure i nwards is also normal to XY

1 80 1 80
8 p l s 1 n 6 = 8p .

FI G . 18 .
— P R ESS U R E AT B A CK OF WH E E L .

But P F for equilibrium so that ,

2 0’2
7
2 0 d r = 2 r dp
, 9
i “
(9
2 1 dp
0 d
g
'

2
I nt e
(0
p C,
g r at l n g
29 0
'
62 C ENTRIF U G AL FANS
where C is the constant of integration so that if r b is the , ,

radius O f the boss that is the internal radius at which

—
,

rotation ceases i t follows that —

(7 2
2 2
M 2

33
el p o= h _ hb,
l
o

where hbare the pressure per sq f t and head at . .

,

radius r b 3 and at an y radius 7 if p h are pressure and '

head ,

~7 2) « 8
8 = hl = h2 _ 9
q 2g

where 172 in the cas e

, Of the fan i s ,

h, 72
2
2
62 2
cosec 2 I
( ) 2g H .

calculate the whole pressure on the back O f the wheel

To
w e must proceed as follow s :Th e pressure head on a ring of
m ean radius r and radial thickness d
, is

2 7r 7
'

d7 '

h
( r
2
2
,
2
)
29

so that the total pressure head between the radii r b and

is

2
2 2 2 7r co 2 7r w
2 7r h2 r d r r
2
2
7
'

d + r r
3
dr
7b 29

To this we should have to add hb t h e pressure 7r

on the flat annular end of the boss 7 being the radius of ,

'

s
2

the shaft and being h2 ) I f the vanes

3 2
, ( r r 1 2 .

at the back do not extend to a radius r 2 but to a smaller ,

radius r then the total pressure head on the back of the

a,

disc is
 V A R I A T I ON OF P RE SS U RE IN A CE N TR IF UG AL F AN (i 3

2
[l b 77
( 777

where hb h .

( r a
2

S i m ilar F ans — Suppose two fans t o be made from the

same drawing but to a d iff erent scale, so that the d i m e n
sions of the one are 77 times that of the other and suppose
.
,

them t o be driven so that the velocity of flow through

corresponding parts of them is the same Then the quan ~ .

tity delivere d by the one will obvious ly be 71 times that by

2

the other and if the water gauge prod uced by each is the
,

same then the orifices will be as is to 1 and the losses Of

, ,

head due to surface friction are the same because the areas
of surface divided by the sections of the fan pass ages are
the same the roughness Of the surfaces of course being
,

supposed t o be the same I f the manometric e ffi ciencies

.

are the same the work per lb done by the wheels and
, .

the losses due to shock at inflow to th em and outflow from

these are the same and therefore the air efficiencies or
,

“ 02
2
L
9 H 9
,
7

where L is the losses of head are the same An d we may ,

.

reasonably suppose the manometri c eff iciency of the one to

equal that of t h e othe r because the equation connecting 7 2 7 ,

H and internal velocities depends on the p r op orti ons of a

fan and not on i ts absolute dimensions and the propo rtions ,

are the same i n both 3 hence we have manometric e fli ci e n cy

and mechanica l efficiency are the same for orifices i n the
ratio of to 1 N ow if w e call
.

Q = 03
x/
gH
z
r
z
64 C E NTRIF UG A L FANS
the reduced orifice then a diagram can be d rawn with these
,

as absciss ae and the m anometric an d mech anical efficiencies

,

as ord i nates and w e shal l get two curves which will apply to
,

f an s O f a given type but of di ff erent si z es Th e volumet r ic .

i is Obviously th e sa me at equal reduced

O

e fl C I e n ci e s
Q
2
+ oz r ,
or l fi c e s . I n fi g 1 9 are shown characteristic curves of this
.

nature for Rateau fans We need not enlarge on the .

'

1 -0 11 4

RE D UC E D O R I FI C ES .

FI G . 19 .
— C H A R A C T E R IS T I C C U R V E S F OR R A T E AU F A N S .

assistance such curves should g i ve to the intelligent de

sig n er but he must remember that engine friction must not
,

be included in the mechanical efficiency and that if there ,

are sharp angles before the eye of t h e fan which disturb

the inflow bot h mechanical and m anometric e fficiencies will
be much reduced .
66 C E NTRIFU G AL FAN S
being connected to a pipe soldered into the side of the
delivery tube while the other branch is attached to a pipe
,

passing to th e centre of the delivery tube the end being ,

suitably bent to face the stream Th e pressure in t he .

pipe C ( inductive action being prevented ) will be that due

to the compression of the air only but the pressure in D ,

will be t hat due to the pressure and velocity combined .

H ence the gauge A which i ndicates the diff erence between ,

the pressures i n the pipes C and D will record the pressure ,

due to velocity only Th e compression of the air can be .

FI G . 20 —
, 28
”
FA N T ES T ED BY H EE N A N G I LBE RT .

measured by the second gauge B First if the outlet be .

,

closed S O that no air is delivered t h e g aug e A w ill remain ,

at z ero since there is no flow of air through the tube but

,
,

the second gauge B will indicate a considerable compres

sion about 1 1 % in of water if the tip sp e e d o f the fan be
'

, .
,

ft per min N ext let the end Of the d elivery tube

. .

be Opened to give say half the area of the outlet K Th e, , .

fan now passes a considerable quantity of air about ,

cu ft per min and on account Of this flo w the velocity

. . .
,

gauge will indicate nearly i n Of water Th e com . .

pression as shown by the gauge B will have fallen t o

, ,
 H EE NAN AND G ILBE RT S ’
EX P E RI M ENTS 67

8 in . D uring
the time the outlet of the delivery tube
w as closed with no air being delivered the efficiency w as
, ,

of course z ero But when the fan is passing

. cu ft . .

of air under a compression Of 8 in of water the efficiency

,
.
,

reckoned on the compression alone will be about 66 per

cent 1 5 horse power being required to drive the fan .

When the outlet of the delivery tube is fully opened the

,
.

Out l e t t
p
ar

Out l t e Close d ,

Out l e t O
p en

FI G . 21 .
— F A N D E L I V E RY T U BE .

fan delivers freely t o the atmosphere and gauge B shows ,

that the air is under 116 compression whatever but the ,

amount of air has increased to about cu ft per . .

min and the passage of this air through the delivery tube
.
,

shows a velocity head on A Of nearly 5 in of wate r Since . .

the air is not compressed but merely expelled at atmo ,

spheric pressure the effi ciency reckoned on the compression

,
68 C E NTRIF U G AL FANS
is now z ero T here is n ecessarily a deli ery intermediate
. v

between no deli v ery and n o compression at this tip speed ,

of 1 2 000 ft per min for which the mechanical e ffi ciency

, . .
, ,

reckoned on the compression produced on th e air is a ,

m ax I m u m .

Heenan and Gilbert presen te d their results on diagrams

in which the absciss a:were cu ft of air per min for . . .

FI G . 22 .
— FO R M S OF B L A D E T ES T ED .

a fixed tip speed a d ordinates were effi ciencies and com

n ,

pressions as shown by the gauge at that speed ; other

curves were drawn showi g the brake horse power required n

by the fan and the total e ffi ciency the kinetic e n ergy i n

, ,

the air discharged being added to the work done in com .

 H EE NAN AND G ILB E RT S ’
EX P E R I M ENTS 69

pressing the air This is sometimes called dynamic e ffi

.
.

ci c ncy and is misleading except in cases i n which the fan

, ,

can be fitted with an expanding chimney and even then the ,

kinetic energy left in the air at its mouth must be deducted .

A fan that has a high dynamic e fficiency may be Of very

low real efficiency and if quantity of air is Of more co
,
n

sequence than t h e smallness of brake horse power then ,

the volumetric e fli ci e cy Should be given and not an

n ,

imaginary mechanical effi ciency which is not attainable .

T h e horse power of t h e fan considering compression alone

was calculated by the formula ,

8h Q, h Q1
Fan horse P°w °r _
3 3000 x 1 2 63 5 2
’

where h water gauge i n inches and Q = discharge in cu f t . .

per min and 8 weight i n lb of 1 cu ft of water T hree

.
, . . . .

forms of blade were tested ( shown in fig 2 2 numbered 2 .

, ,

3 and
,
mak ing a n gles of 3 5 d e g 60 d e g and 90 d e g , , ,

with tangents to the outer circumference and 2 5 deg ,

.

with tangents to the inner circumference the fan centre ,

being 1 7 in diameter and 8 in wide Figs 2 3 2 4 25

. . . .
, , ,

give curves of efficiency water gauge and brake horse , ,

power for each Of these blades for a constant tip speed of

,

ft per min Blade N O 4 gives the best results

. . .
,

although not much superior to those of blade N O 3 whilst .

,

blade N O 2 havi n g a tip angle of only 3 5 d e g gives a low

.
, ,

e fficiency owing to t h e rapid drop of the compression curve

,

as the discharge of air increases T h e arrangement of .

apparatus for measuring the brake horse power volume of ,

air compression and speed are shown i n fig 2 6 Th e fan

, , ,
. .

is driven from the counter shafting B which derives its -

motion from a spherical steam e gine at C Th e outlet of t he -

n .

fan is connected by a Short circular iro n delivery tube with

a boiler flue E E 30 in diameter and 1 8 ft long At
,
. . .

the centre Of the flue a pa t i t i on F is fitted to which can be

r ,

attached a series of plate s having circular orifices of various

si zes varying between 4 % in and 1 8 in diameter A well
, . . .

ou t outlet of known diameter is placed at th e end Of the

flue G where the velocity of the air was measured by an

,

anemometer This opening was made much larger in

.
70 C ENTRIF U G AL FANS
diameter than the outlet of the fan i n order to avoid ,

i j uring the anemometer by a violent current of air I n

n .

fig 2 6 it is shown i n end elevation together with the levers

.

whereby the anemometer w as moved over all portions of

the outlet the i nstrument being kept truly perpendicular
,

to the axis of the fi ne Th e pulley driving the fan was .

not keyed to the shaft but was driven by it t hrough t he ,

Emerson power scale a form of transmission dynamometer

,

i n which the moment o f the d riving eff ort is balanced ,

t hrough a system of levers by a pendulum moving over a ,

graduated scale A speed counter records the r e v o lu

.

tions of the shaft A tachometer coupled to the counter

.

shafting B by a spiral spring enabled the speed of that ,

shafting to be regulated and the proper tip speed to be ,

approximately maintained by the man i n charge of the

engine A small band counter carried in a sliding frame
.
, ,

and applied when necessary to the fan spindle enabled ,

the exact n umber of re v per min Of the fan to be . .

obtained Measurements Of the pressure and velocity

.

of the air stream were taken at t he section M of the

-

delivery tube Th e velocity varied greatly in diff erent

.

positions on the s ame cross section of t h e tube Readings -

.

were taken at several points i n a cross section by means of -

t wo gauges each of which could be traversed ove r a

,

diameter at right angles to the other I t was afterwards .

found unnecessary to measure the velocity since it can be ,

readily calculated when the discharge as measured by the ,

anemometer is know n 3 but at the same time the V elocity

, ,

as measured by the velocity gauge gives a check on the ,

ane m ometer readings I f the cross section of the delivery

.
-

t ube be divided into a number Of imaginary areas an d the

.

square root of the mean gauge reading for each area be

multiplied by that area and by a suitabl e constant then ,

these results added together give the total discharge of the

fan Th e annexed table illustrates the variation of velocity
.

referred to 3 it gives the gauge reading due to the velocity

i n four positions along a diameter of the cross section of -

the delivery tube where the gauge was fitted as taken

, ,

simultaneously .
 HE E NAN AND
’
G I L B E RT S E X P E R I ME N TS 71

TA BL E 6 .
— I
V AR AT ON OF I V E L OCI TY .

Ga u g e
Al r v e lo
D is t
n ce a
ti gég
P os i o n r e ad in g d ue y ’
fr om c e n t r e Re m ar k s
f t pe r
. .

to vl
e oc i ty . o f t ub e .

In h s c e .

fi f
am ’t e ’ ‘’
2 se
-
1 12 l f t l 0:
33
e ‘
1 1 -

In
1 70
.
.

T his we think is a proof o f imperfect design of the f

, ,
an ,

because if well designed there is n o reason why the velocity

should not be almost uniform over t h e whole section of the
tube if anything it should be g reater at the centre than ,

the sides I t will be noticed i n fi g 2 0 that some of the

. .

discharge fro m the fan is thrown against t h e upper surface ”

o f the discharge pipe instead o f t h e whole stream flowing ,

out parallel to the a x is of the tube Th e measurement o f .

the compression presented some diffi culty owing t o the ,

fact that the air flowing across the end of the side g uge a

caused a large amoun t o f induction a vacuum being often ,

recorded where a pressure was known to e x ist P rofessor

W C U nwin found that a plate placed across the end o f
.

. .

the tube prevents this inductive action Th e form o f side .

gauge used for measuring the static pressure is shown i n

fi g 4 2 I t was tested by H eenan and G ilbert and found to
. .
,

give very good results To draw the characteristic curves .

from the e perimental results seven resistance plates

x , ,

A B C D E F with graded circular ori fi ces were arranged

, , , , , , ,

t o fasten on to the centre F of th e boiler flue T wo o b se r .

v at i o n s were taken with each plate o f the discharge the ,

compressi on and the horse power supplied to the fan at

or near tip speeds of ft f t and ft . .

per min T h e authors veri fi ed the following laws :

, , .

1 Th e air discharge varies as

o r a constant
.

2 Th e
, g varie resist a nce
3 Th e
.
 72 C EN T RI F U G A L FA N S

r
e
t

W
a

f
o

s
e
h B la d No 3
e . .

V l u m dis h rg d i th us n ds
o e c a e n o a of c u. ft .
p er m i n .

F I GS . 23 A N D 2 4 .
— EFF I C I E N CY GA U GE AND P OWE R D I A GR A MS .

C u r v :III h ow ffi i n cy ;
es s s e c e P, b r ak e h o rse p o we r H, co m pr e s s i o n ;
0, t t lg
o a au
ge A to t al e ffi ci e n cy
’
, .
74 CE N TR I F U GA L FA N S

Q
m
m
m
m

5
3 c
. z
5
9 E
.
p
o m
o
o
m
.
é
S
€
0 m
o
o
m
m
a
.

G
w
r .
fl
G
3
w
a
4
n
o M
o
m a
n
a
m
a

7
3
2

5
2
.

o
m

d
a

m
 HE E NAN AN D G I L B E RT S ’
E X P E R I M E N TS 75

tapering side plates Th e diameter of t h e inlet was 5 ; i n .

.
,

the widths of the fan centre at inlet and outlet being 7 % in .

and 1 % in respectively This fan ran i a concentric ease

. . n ,

the clearance b ei ng l g in (which in o ur opinion made the .

comparison of tapering ve r us parallel sides worthless as

’

s ,

the preceding theory very clearly shows that a well des igned -

volu t e is nec essary t o obtai n high effi ciency and a su i table

water gauge ) Th e compression obtained with closed outlet
.

was 9 4 i n against 1 1 2 in with blade N O 4 fig 2 5

.
,
. .
,
. .

Further t h e compression fell o ff very rapidly with increase

,

s
g
" 0

3
y.
n

V lum o e d i h rg e d i th us n ds
sc a n o a of c u . ft .
per m i n .

FI G 2 7. .
—
C H A RA C T EI I s r I C : '

C UR VE F o r. FA N w r r H
'
T A P ER I N G
SI D E S .

FI G 2 8 . .
—
C H A1 :
A C T E1:
I T I c C U RV Es F OR FA N WI TH I 2 -
I NC H
C EN T R E D I A ME TE R .

Cu r v es :Ill s h o ws e ffi ci e n cy ; b rak e h o rs e p owe r ; co m pr e ss i o n .

of output so that the working compression would not be

,

more than 8 in Th e effi ciency was also low ( as might .

reasonably be e pected ) F i g 28 sho w s the characteristic

x . .

o b tained at ft p er min from a fan centre 1 2 in . . .

diameter the ma x imum width being 2 % in T h e centre

, .

rotated in a whirlpool chamber o f 2 3 in diameter Th e . .

centre was o f cas t iron and the tip angle of the blades was
.

3 0 deg O f co urse the effi ciency measured which was less

.
,

than 3 0 per cent would not be representative for so small .

,

a f n but the ma x imum compression did not e x ceed 4 4 in

a , .
76 C E N TRI F UGA L FANS
Th e e fl e ct of a whirlpool chamber ( usually called a diff user )
'

is to produce a very quiet running fan .

Test of a Mi n c Ven t i lati n g Fc m Th e fan selected for illus

.
-

t r a t i o n ( fi g 2 9 ) w as supplied to the P arkend C olliery C o

. .
,

S outh Wales by H eenan an d G ilbert and works in con nec

, ,

tion with the approach tunnel and ventil ting shaft of th e a

mine Th e wheel is 7 ft diameter and 2 ft wide Th e upper

. . . .

portion of the case and evas e chimney is built up o f wrought

iron p lates the lower portion being formed i n brickwork
,

in thick Th e fan was driven by a hori z ontal n on

. .
,

condensing engine the cylinder being 1 2% in diameter and

, .

1 7 % in stroke . To provide a variable resistance for the

.

FI G 2 9. . MI N E V E N T I L A T I N G
— FA N .

fan three 9 in by 3 in planks were placed across the

,
. .

mouth o f th e air drift and boards nailed to these planks

,

restricted the flo w o f air to the f more o less as required an r .

T h e folding doors at the top o f the ventilating shaft were

open during the whole o f the test A tachometer driven by .
,

a belt from the fan shaft enabled the appro imate speed o f ,
x

the fan to be j udged and regulated by the man in charge o f

the engi ne ; the number o f revolutions in a t w o minute -

reading being obtained by a hand counter held to the f an -

shaft Th e e gine speed was obtained by a coun ter applied

. n

to the shaft i n the same manner Th e air discharge was .

measured by an anemomete r at the top o f the fan chimney .

Th e area of the top o f the chimney was divided into eigh t

 T E ST OF M I N E V E NTILAT I N G FAN 77

equal rectangles by tightly stretched wires and the ane ,

m o m e t e r att ached t o a small i ron tube was held f or a

, ,

quarter of a min i n each division I n this fan the flo w o f

. .

air was fairly uniform over the whole of the outlet area o f
the chimney but i n some cases where t h e fan was run
, ,

slowly f or e perimental purposes guide vanes had to be

x ,

fi tted in the base o f the chimney to secure the resul t men

t i on e d F our degrees of O pe n ing were arranged at the
.

adj us tab lc ori fi ce and f or each of these readings were taken

, , ,

wit h tip speeds of ft ft ft ft . . .

,
.
, ,

an d ft per min T h e vacuum produced by the fan

. .

was measured by a side gauge placed in the air drift close ,

to the fan inlet and a pipe led from this tip to a water
,

gauge placed on a table outside ( I n our O pinion this is o n e .

o f the reasons f o r the app ren tly high eff iciency obtained a

by the fan which at a tip speed o f

,
ft per min . .

re ached 7 0 per cent ) We have already stated that if the

.

manometer is placed i a strong cu ent o f air the water n rr

gauge will be increased and that at inflo w t o the fan this ,

will o t be the correct gauge When the fan was running

n .

at a tip speed of ft per min t h e discharge fro m th e . .

fan at the effi ciency of 7 0 per cent was 1 4 cu ft per sq in . . . .

o f diametral wheel section per min so that the flow into .

,

the eye assuming i t to be uniform and neglecting the

,

obstruction of the h e ar i ng was at a velocity o f ,

14 24 x s4
x —
4i 6
60 1r

because eye was 3 % ft diamete r and the centre section

the .

24 x 84 in This velocity would increase the water gauge

.

by
12
=057
‘
-
x 1 11
o

8 20
.

It was inand should therefore be reduced to at least

.
,

ih .
,
so that the e fli c i e ncy could n o t b e more than
x =
7o 64 4
-
per cent .
78 CE N TR I F U GA L FAN S

( ass umi ng that the anemometer did not e x aggerate the

discharge which it invaria b ly does ) Three side gauges

,
.

C C fi g 2 9 were also placed at the root o f the chimney so

,
.
, ,

that the vacuum produced and consequently the effi ciency ,

V l u m dis ha rg d i n
o e c e cu . ft .
p er m i n p er
. sq , in . o f ce n t r s ti n
e ec o

FI G S . 30 AND 31 ,
— C HA n A 0T E R I STI c CU R V E S ; M I N E FA N S .

30 — 7 f t He e n a n Fan
.
-
. . Ti p sp d ee ,
ft.
pe r m i n .

31 f t H e e n a n Fa n
.
-
. . Ti p s pe e d , f t pe r m i n
. .

C u r v :M h
es s o ws e ffi ci e n cy ; 0 wat e r g aug e , P, h orse p o we r
( i nd i cat e d ) .

of the chi m ney could be determined I t was found that the

, .

vacuum was practic lly the same at all three so that only a ,

o n e was read Figs 3 0 to 3 3 give the characteristic curves

. .

obtained from this test Th e dotted lines 3 6 correspond . (

with the resistance o fle r e d when the choking boards were '

 TE S T OF MI NE V E N TI L AT I N G F AN

removed and t h e fan took air from the mine only .

ma imum effi ciency increases with the speed and is 6 7 2

x
,

6 8 10 12 14 6 16 18 20 22 2
V l u m di s h rg e d i n c u f t
o e c a . .
pe r m i n p e r
. sq . in .

FI G S . 32 AN D 33 .
— C H A P A CT E I
. :
I I G C U R V ES :
ST M I N E FANS .

32 .
— 7 f t Fan
-
. . Ti p s pe e d ft .
pe r m i n
ft
, .

33 . .

pe r m i n .

Cur ve s : e ffi c i e n cy ; 0 , wat rg e au
ge ; P , h orse p o we r
( in d i t d )
ca e .

cent at a tip speed of

.
ft per min and 7 0 3 per cc . .
,

at ft per min Th e fan was d esigned to pass

. .

cu ft pe r min with a water gauge of 3 % i n at a sp

. . .
.
,
80 CE N TR I F U GAL FANS

h
c
n
I

n
o
i
s
s
e
r
h
m
o
C

V l u m dis h rg d i n
o e c a e cu . ft .
p er m i n p er . sq . i n . of c e n t r s ti n
e ec o .

FI G S . 34 AN D 35 .
— C H A R A GT E R I ST I G C U R V ES M I N E FA N S .

H e e n an Fan He e n an Fan
”
34 . . 35 28 ,
— .

Ti p s pe ed , ft .
pe r m i n . in e ac h case .

Curv :P h es s o ws b rak e h or s e p o we r ; 0 co m pres s i o n , T, tt l

o a

g aug e ; M e ffi c i e n c y ; N to t al e ffi c i e n cy
, , .

of revolutions A t
3 00 . ft per min and 3 1 8 rev
. .
,

the water gauge was and the disch rge a

cu f t so that the fan is

. .
,
a mply large enough for the
work .
82 CE N TR I F U G A L FAN S
and if A is the section of the j e t
2
Av P ’
0
2
p
0
'
— P, or _ H
o A
,
g g a

where p is the pressure per sq f t is the density . .

,
0 ,

and H is t h e head equivalent to the pressure 17 He n ce we .

might at fi rst imagine that the reading o f the P itot tube

would be 22 9 9 instead of v + 2 g in feet of air ; but the
2 —
,
2

air deviated by the shock upon the e n d o f the P itot tube

does ot lose the whole of its m omentum parallel to the
n

a x is o f the tube so that in reality

,

H=m g
9

where m is a coeffi cient smaller than unity depending ,

probably u pon the section of the ori fi ce and of that of the

conduit i n which it is situated When the ori fi ce is small .

compared with the conduit e x periment s h ows that m = § ,

very n early I f h is t h e water g uge in inches

. a ,

2 2
1 2 12 0 12 x 39 8 3 °

29 8

where 8= weight of 1 cu ft . . of water ,

2
13 11
25 1 T
f or dry air ( 49)
where B is the height of the barometer in inches of mercury ,

and 7 is the absolute temperature i n Fahrenheit degrees or ,

7 = F + 4 61 ;

so that if we put B : 29 9 and F : 62 deg

2
v
72 —
4 390
1n . of water .

test a tip as described in H eenan and G ilbert s

To
’

paper the water gauge con nected with a revolving tube is

,
 TH EO RY O F TH E P I TO T T U BE 83

O bserved and the reading compared with that calculated

,

from the equation

2
12 0 0
h
29 8

the co rect values O f and 8 being used ; a correction

r 0
'

must however be introduced f or the vacuum produced by

, ,

the centrifugal f o ce of the air i n the revol v ing hori z ontal

r

tube carrying the tip This vacuum we have already .

shown to be v + 2 g at the centre ; for in a mass O f air

o
2
,

rotating we proved that

,

2
(0
1
3
— +d
-

29 0

SO that if be the e x ternal radius and ,

2
P1
’
0
=P 2
‘

2g 0
’

where 17 2 101 is the diff erence of pressure between the

—

centre and radius H ence if the tip measures th e

pressure due to the air velocity correctly there should be ,

no re ad ing on the water gauge f or any speed as the ,

vacuum due to the centrifugal force j ust balances the

pressure due to the velocity of the movi ng tip against t h e
air As a matter of fact small readings were Observed o n
.
,

the manometer but these were accounted for by supposing

,

that the air in the tank w as drawn round by the rota tion O f
the tip and that the ma x imum velocities o f the air were
,

f o r gauge fi g 3 7 2 7 ft per sec

, . f or gauge fi g 3 8 5 9 ft
,
. .
,
.
, .

per sec ; and f or gauge fi g 39

. ft per sec so that the
, .
,
. .

velocity O f the air may be calculated by formul a

E xp er i m e n ts on Cen tr if uga l Fans by B r yan l
D on/oi n .
”

A lthough these e periments were made o n small fans con

x ,

s i d e r ab le value is attached t o them as it is e x tremely ,

robable that the quantities o f air recorded are correct

pn accordance with H eenan and G ilbert e x periments
.

D onk i n found that when the passages remained the same

84 GE N TR lF UGA L FANS
the discharge varied as the speed I n each set of e x pe r i .

ments upon a given type O f fan t h e quantity O f air pas sing

was varied and a change in the pressure was produced by
,

throttling the flow in the delivery pipe at some distance

FIG . 37 . FI G . 38 . FIG . 39 .

L M,

FI G 4 0. . FI G . 41 . FI G . 42 .

T Y P ES OF G A UG E TI PS .

”
D E; K. 1 i ;
”
A. B,
L , M ,

from the fan and allowing th e air to pass successively i n

,

each e x periment through wove wire of 3 8 3 0 and , , ,

5 0 mesh es to an inch equivalent ori fi ce was also ”

Th e .

varied by the insertion o f one to four pieces O f perforated

z i nc superposed Th e wove wire of 3 and 8 meshes to an
.
 B RYAN P E RI M E NTS
’
D ON K I N S E X 85

inch gave respectively by calculation a n eff ective area of

8 0 pe r cent and 5 6 per cent O f the area O f the pipe
.
, O ne . .

piece O f perforated z inc gave an area O f 40 per cent Ex .

e i m e n t s under these several conditions as well as with

p r ,

no ba ffle were m ade upon each fan the end o f the delivery
, ,

pipe be ing O pen to the atmosphere i n all cases This end .

was i n a fi nal e x periment co m pletely stopped and the air

, , , ,

i nstead Of passing through the fan was churned u p i nside ,

it Th e I H P was t aken in each case as well as the

. . . .
,

pressure of the air and the speed o f the fan A bout ten .

e periments were made o n each fan each occupying about

x ,

1 5 min after all conditions had become consta t and as a

. n ,

simil r series of e x periments were made upon each with

a ,

the same pipe and apparatus the results are comparable , .

Eleven diff erent types o f f an illustrated in Fi g 4 5 ,

.
,

were tested Of diameters varying between 1 6 in and

, .

2 5 % in Th e number and shape Of the vanes diff ered

.

considerably Each f was driven by a strap from

. an

the same steam engine which was indicated to give the —

,

power absorbed Th e I H P required to drive th e engine

. . . .

at diff erent speeds was accurately known and was i n each ,

case deducted from the total I H P A large quantity O f . . .

air is required at low pressure in some cases and in others ,

a small quantity at high pressure Th e volume o f air pass .

ing at the maxi mum pressure with a given speed of the ,

fan and equivalen t ori fi ce was determined in the e x pe r i ,

ments Th e terms volum e t r i c and manometric (or pres

.

sure ) effi ciencies already e x plained are , ,

N Q 46 . ( )
,
2

Nm ( 2 5 a
)
and the mechanical effi ciency of the fan itself is
62 3 Qh
'

N:
12 x 550 x B H P . . .

where B H P is the horse power supplied to the fan shaft

. . .
-
.

T his is appro x imately the same as the e x pression we

86 C E NTRIF U G AL F A N S

FIG . 45 .
— FA N S T ES T ED BY B RY A N DO N K I N .
 B RYA N D ON K I N S E X
’
P E RI M ENTS 87

BH
FI G S 4 3 A N D 4 4 — E L E V A T I O N S
. . F A N B LA DE S.

B lad es o f f a n s sh own i n fig 4 5 . . A c e n t r e l in e o f f an s pin dl e ;

,

a rr o w I ! s h o ws d ir e c t i o n o f o u tl e t ; a rr o w 0 s h o ws d ir e c t i on Of
’

ro tati on .
88 C E NTRIF U G AL FANS

termed the air effi ciency gH

and which we also denot ed
by N .

T h e number O f vanes o f each fan together with th eir ,

shape and direction of curvature are given in Tab le 8 ,

.

Th e casing in which the vanes revol ved di ff ered con sider

ably in shape but was always of cast iron Th e vanes
,
.

revolved i n some cases with the concave in some with the ,

conve x an d i n others with the flat side to the outlet

, .

T hey were set both radially and inclined to the centre of

the Shaft Tests were m ade to determine the eff ect o f
.

driving the fan with the blades revolvi ng i n the O pposite

direction to that indicated by the makers and an increase ,

in delivery sometimes resulted although other conditions ,

remained the same A drawing of each fan is given i n

.

fi g 45
. Th e ma x imum effi ciency o r the best e x periment
.

o n each is given in T able 8 and the volumetric mano , ,

metric ( or pressure ) and mechanical effi ciencies are r e pr e

,

sented graphically in fi gs 4 8 4 9 and 5 0 A S regards .

, ,
.

possible errors the speeds pressures and quantities O f air

, , ,

are probably correct to within 3 and 4 per cen t Table 9 .

gives the results Of the e x periments on the diff erent fans

with ma x imum and minimum equivalent ori fi ces together ,

with particulars as to speeds etc O bserved i n the tests ,

.
,
.

D onki n s E xp e r i m en tal A pp ar atus A conical piece O f pipe

’
—
.

X Y (fig fi tting the outlet O f each fan was bol ted as

.
, ,

shown to the fan under test on on e S ide and to a wrought

, ,

iron pipe 1 4 3 ? in i n diameter at the other This latter was

,
1
.
,
.

used f or all e x periments Each fan was driven from a .

small engine by a strap from a rigger fi x ed o n the crank

Shaft . Tw o assistants started the counters by signal at the
same instant and at the end O f the e x periment they were
,

S imilarly thrown out o f gear B y this means the speeds.

and sli p O f t h e strap were ascertained A t B and B two .

pipes o f 4 in diameter were fi x ed to the air pipe Y Z and

.
-

to t hese were attached two U water gauges by which the ,

static pressures o f air were O bserved Th e dynamic pres .

sures were O btained by means O f a dial gauge o f special

design Th e circular pieces o f wove wire or perforated
.
90 CE N TR I F U GAL FA N S

w
.
.

8
. .

8 8 L6
u
8 8 a
9 9 9 9 u 8 8 8
8

2 3 § 8 8 0
8 9
.
c 0
2 8 2 6 9
H
e

8 a 2 8 8 8 8
o
w m m 2 3
9 8
0
.
0 9
9 M
M H m A
5
H G o
$ 8 «
a 9
H 9
3o
8 $ 8
o S o 8 o 8 0 8 .
0 0 0
8
S 8
8 88 83 o 9
2 2 9
m 2 8 8
8 0 8 5
2 E 2 3 2 3 3 8 8 8
9 : : £
9 a 9 9 a 8 8 6 H

:0 :0 :5 5
.

m 5 »
0 0 8 8
2 2 2 9 0
E 3
:
o 8 8
2 n G 2 m 3 s 5 8 0
% e o 9
. 9
. 9

:
I b8 £
2

m o
3
n S N V H o

m
.

aa e o u
H
. .

90
T
a
m
l
n
.
Ao u 9I0
. ms ou 0
1.
9 m mA
.
m H o
H
a 3 8
a
n m
v
H
H H

« ne9 ma
m w W 9 a 9 N « m 9
7
b m 3
.

u
a

Ao
.

m m m m H
8 t %
m
m .o 3 ? sA a
e

m m v m n “ .

u9 e J
o H m 9 F m T a
3
l
.

u R H
.

.
9 wm m m m d
Ao . m s mO
. qw N H H m m 3 m
o
m v a

m
m m m

1
mM g 3
H
1 19 8 1 1 m e; e m e m m H
S
I e AAOd GS AO
m
.

H
.

p aq aos q e
.
.
m ” m N H

H H H H H H o
o
m
'
Am e ae m s aq o ur
w W m N m H m m H
10 .

é
. . . .
.

m o o o o o o
In
'
s se I d 49 111 04 123 m w m m m m
o
m m
0
m a
s
a
2
4
C

a
m

'
u r ur
I s d ar e
.
IO 3 } m o

at e qe m
'
go s aq o u u r ; m o

{ asI '
s se l d o t uue u lfi q m Q

E
.
H
. . .

50 s aq o u r 2 0
9 3 a 3 «
m E
u; re
v a
'

s s e ad 3 » 3 ;
2 a
g M
m m m 9

w H
°

qu e o I e d de n s go d ug
.
u
v
9

w
r
m
o
h
AA m m u
3
g g
a
H
S K
'
u t uz e o m m m
J ed u s ; w m m H m n m m
JO A GH 3
‘
s

H H H H H H
 B RY AN D ON K I N S E
’
X P E RI M E N T S 91

n d
o
o
z
o 5
o 3
fi w
g e

oo o o o

m o m o o v
a; b 9 I n
“ '

m m m
9 :)n A gn b g
’
m
é b
v
.

o
m
.

o b b
" w m

:
w m m m
c
$ S
v
n
m oo xod . Kauat o m o u
z a m 9 v m
b
O

b 2
.

h d
.

qam o w m
g; a
2 2
-
op n lo A
o m m m m

:
m
n o c m a
3
v
m
d H w m
o S o w u w fi

—
.

m A % ; n
n
é b é o
w m
w g S
m
a
6
H «
r
fi
r e
a

r m
mw
:
u
.

u A é ;
.

m ”
” m m m m

: :3
w m m m
w fiw
c o
.
o
o 3 3 $ w m a m w
m w u m w m m 3
3 3 2
“

g w a fi

:
« o
up u 1 9 d
$ 3 3 $ 3
a
$ 3 3
'

m o n
m a} ; 0 m ag m m 0
9 0
2 : H fi 3 2 ? 2 m

"gg
‘
saq ou ; ( 1 ; “ # m m
99 m m J OA O
2
3 a o m
e
m 3 m 2 m a m 3
. o

a
. . . o . . . o

5 5 5 5 n
5 n
38 n
;
“ “

3 m u 8
: :
m 0 a
r
:
. .

u .

3
.

:
n

: :
.

: :
a
H
3 7
o

5 2
c

g
o

a a a a a H
.
H a h a H a a a
s 5 ?
5s 3 s o
.

s o s
o

5
c
.

> B g
.

w
.
0
2 3
“
3 ? 3 32 S
.

3 3 2
.
. . . .
3
8 o 3 3 9
a
8 8 0 £ 0

53 3
5u 52
0 0 0 0 55 0 0

wo
0
32 52 2 2 3 2 2
t» p
3
o
3
3 3
3 3
w 8 3
3 3 3
w
o 3 3
n S c w
«
m
.

0 v “

S :
H
>
H
H

>
Z
. :
a
N
u
C
m
94 C ENTRIF U G AL F A N S

mean dynamic pressure at Z having been thus obtained ,

the mean velocity of the air was deduced from the formula
( 2 ) given i n the list of formul ae used in the e periments x .

Th e mean velocity having been determined the quantity ,

o f air i n cu ft per sec was obtained by multiplying the

. . .

velocity by the area of t h e pipe Th e latter was gauged .

inside the end of the pipe Z allowance being made f or the ,

W ire template Th e temperature o f t he air at Z was noted

.

in each e periment with a wet and d r y bulb thermometer

x ,

and the barometric pressure was also observed .

’
The ele ven typ es of f an us ed i n Donkin s exp er i m en /s wer e
No .fi gs 4 5 , 4 4 , was made with twenty specially shaped
.
-

wrought iron vanes curved i n two directions and revolving

—
,

with the concave side to the outlet Th e casing of volute .

,

form gradually increased in cross sectional area towards the

,
-

outlet and there was only o n e inlet for air o f special bell
,

mouthed form Th e driving shaft was fi x ed in the centre

.

of t h e inlet with a cone on its end to guide the air to the

,

revolving blades ; there was no bearing or obstruction to

preven t the air from entering freely Th e two brass bear .

ings for the shaft on e with three thrust collars and one
,

plain measured 1 % in diameter by 5 % in long This fan

, . . .

worked very quietly D etails of the ten e x periments per .

formed o n this fan are given in Table 7 L ittle attention .

was paid to ensuri n g any particular speed in each e x pe r i

ment as it was found in the fans tested that all other
, ,

conditions bei n g the same the quantity of air delivered ,

was proportional to th e speed Th e three effi ciencies o f .

this fan are all high Two special e periments were made . x

with a sheet o f perforated z inc in the pipe one with the —

vanes varnished and covered with coal dust to represent

dirty vanes i n a coal mine and the other with vanes clean ,

and bright Corrected f or speed the results of the com

.
,

parison showed that 1 0% per cent more air was delivered .

by t h e clean vanes but t he mechanic al efficiency was about

,

the same Th e pressure e ffi ciency was 1 1 per cen t and

.
,

t h e volumetric effi ciency 6 per cent higher with clean than .

with dirty vanes Tw o e x periments were made o n e with

.
,

the large bell mouthed inlet fi x ed i n place as designed

-

and o n e with i t removed Corrected for speed the result .

,
 B RY A N D ON K I N S E
’
X P E R I ME N TS 95

showed that 3 1 per cent more air was p assed when the .

inlet was used and t h e mechanical e ffi ciency was 9 per

,

cent higher Th e pressure e ffi ciency was 3 3 per cent

. . .
,

and the volumetric 1 5 per cen t higher This proves the . .

ad vantage of admitting the air without shock n d i n the a

right direction I n fi g 4 7 is given the variation of the

. .

pressures at t h e end o f the pipe taken with a P itot tube ,

connected with the dial water gauge in an e x periment ,

o n this fan Th e pressure was taken in e ach o f the eight

.

equal areas I n this e x periment with t wo sheets o f per

.
,

f or at e d z inc in the pipe the mean o f the eigh t readi n gs ,

FI G 4 7 . .
—
SuBD I V I s 1 0 N OF P I PE F R EE -
EN D .

En d v i e w o f pipe a t 3 , s h o w Cr o ss - se c tin o f pip e s h ow

o ,

i n g p o in t s a t wh i c h pr e s i n g pr e ss u r e s a t t h e e l e v e n
s u re s we re t a ke n . p os i t i o n s i n in ch es o f
wa te r .

gave a velocity o f ft per min ( Table Th e . .

velocity obtained from t h e pressure at the centre of the

pipe taken t the same time was
a ft per m i n and . .
,

that taken from th e mean o f the pressures at points two

thirds of the rad ius from the centre w s ft per mi n a . .

No . fi gs 4 5 4 4 had twelve short curved vanes east

.
, ,

i n one piece revolving concave to the ou tlet S ome e x pe r i

,
.

ments however were made with th e vanes running i n the

, ,

O pposite direction to ascertain the eff ect on the pressures

and quantities o f air T he outer cas t iron casing was o f
.

volute form gradually increasing i n ar ea towards the outlet .

96 CE N TR I F UGA L F AN S

It was provided with two central air i nlets T wo e xpe r i .

ments were made with the vanes reversed and revol ving
conve x t o the outlet Th e increase in the quantity o f air .

delivered due to the vanes revolving concave to the outlet

was found to be 1 1 per cent Th e mechanical effi ciency .

was 5 per cent less i n the latter case but in the pressure
.
,

e ffi ciency a gai n of 5 % pe r cent and i n volumetric effi ciency .

a gain o f 2% per cent were eff ected . .

No . fi gs 4 5 4 3 was the simplest tested and had only

.
, , ,

s i x short straight radial wrought iron vanes with two inlets

,
-

for air ; the cast iron casing was eccentric to the shaft so
-

that the cr oss se ct i on al area gradually increased towards

o

the outlet Th e fan worked quietly and the three e ffi ci e n

.

cies occupied a relatively good position .

N o I V fi gs 4 5 4 4 was of a special duple x type the air

. .
, .
, , ,

passing from an outer to an i nner casing and from thence ,

to the outlet Th e vanes were mounted on a central plate

.

o f wrought iron so that o n e half o f them were in the outer

,
-

and the other half in the inne r easing There were e i ght .

large and eight small vanes on each side of the centre plate
revolving concave to the outlet Th e bearings of white .

metal were three in number 1 % in diameter and 5 in long ,

. . .

N o V fi gs 4 5 4 4 was made with twenty four short

. .
, .
, ,
-

curved wrought iron vanes intended to revolve with the con

-

cave side to the outlet I t was of the double flow type being .
,

provided with two cones mounted o n the shaft and bell ,

mouthed i nlets to direct the entering air Th e outer casing .

was of cast iron of volute form the area increasing gradu

—
,

ally to the outlet Th e two bearings arranged to allow f or

.
,

any de flection of the shaft were in in diameter by ,

.

4 % i n long and were placed well away from the i nlets

.
,
It .

will be seen that the three effi ciencies in the best e x per i
ment are all relatively high O ne e x periment was made .

with the two e x ternal bell mouthed inlets removed the -

,

directions o f the vanes remaining the same Th e gain .

due to these inlets was 3 % per cent in the quantity of air

— .

delivered 4 % per cent in mechanical effi ciency and 5 % per

,
.
,

cent in pressure effi ciency

. .

N 0 V I fi gs 4 5 4 3 had s i x vanes but their c urvature was

. .
, .
, , ,

not continuous 3 at about the middle they were set b ack and
98 C E N TR I F U GA L FANS
N o I X fi gs
. represents a type used i n consider
.
, . 45, 4 4,
able numbers o n the C ontinent Four vanes an unusually .
,

small number of th in cast iron are provided and are

, ,

intended to revolve concave to the outlet There are t wo .

inlets f or the air placed centrally to the shaft and also to

,

th e outer casi n g which was of cast iron Th e two bear , .

ings with white metal linings were 1 in i n diameter by

, ,
.

5 in long
. These were situated close to the inlets
. Th e .

fan was always noisy especially at small ori fi ces , .

N o X fi gs 4 5 4 3 was designed especi lly to deliver a

. .
, .
, ,
a

small quantity of air at a high pressure There were in all .

eighteen vanes s i x of whic h were whole vanes with twelve

, ,

half vanes i nterposed They were all cast i n on e piece .

and straight but not set radially to the centre T here

, .

were two inlets centrally with the shaft coned to guide ,

the entering air Th e boss of the revolvi g part was also

. n

curved to assist the entrance of the air Th e outer casing .

was o f cast iron set concentric with the shaft and of ,

rectangular section gradually increasing i n area to the ,

outlet Th e bearings were 133% in diameter and 6 in long

. . .
,

placed well away from th e air i nlets Thirteen e p e r i . x

ments were made with a diff erent number of baffles I t .

will be seen that this fan gave fairly good results com
pared with others but it must be remembered that it was ,

designed to produce pressure O ne e x periment was made .

with the vanes running in a contrary direction to that

shown in t h e fi gure Th e results corrected for speed .

showed that the quantity o f air deli v ered was 3 1 per cent .

m ore the mechanical e ffi ciency 4 per cent less the

,
.
,

pressure effi ciency 1 % per cent less and the volumetric .

,

effi ciency 3 per cent higher ( This last is diffi cult to . .

u nderstand for if the discharge was 3 1 per cent greater

,
.
,

the volumetric effi ciency which varies as th e discharge ,

and inversely as the speed for a given fan at a given

ori fi ce should also be 3 1 per cent more ) Th e fan worked
, . .

quietly A n attempt was m ade t o obtain the pressure and

.

velocity o f the air between the casing and the r e v olv

ing blades when running at rev per min and . .
,

with three sheets o f perforated z inc superposed in the

pipe A P itot tube was held against the current close t o
.
 B RYAN D ON K I N S
’
EX P E R I ME NTS 99

t he edge of the revolvi ng blades and also as near th e ,

inside of the casing as possible Th e pressure was found .

t o be higher when close to the vanes At D No X .

,
. .
,

fi g 4 5 the dynamic pressure when t h e tube was held

.
,

close up to the vanes was 1 1 % in of wate r corresponding .

,

with a velocity o f ft per min and with the tube . .

,

as near the outer casing as possible t h e pressure was

in o f water or
. ft per min A t E the opposite
,
. .
,

point on the circumference t h e dynamic pressure when ,

the tube was held close t o the vanes was 1 1 % in head .

o f water or ft per min and with the tube as

, . .
,

near the oute r casing as possible 1 1 ;3 i n or

1
ft per .
, .

min Thus it will be seen that the velocity of the air

.

increased round the i nside of the casing with the direction

of rotation for the same speed o f fan ( I f we can j udge .

from the fi gure the increase o f sectio n round the casing

,

was not proportional to the angle so that the ratio of the ,

section t o the quantity of air passing through it gradually

decreased towards the outlet and this would account for ,

the increase o f velocity mentioned Th e less velocity .

near the outside of the casing was probably due to ski n

friction ) Th e static pressure at A in the e x perimental
.

apparatus fi g 4 6 at the same time was 9 % in o f water

, .
,
.
,

and the quantity of air delivered was cu ft . .

per m i n .

N o X I fi gs 4 5 4 4 was made with blades fi x ed on one

.
,
.
, ,

side only of a disc having ten cas t iron slightly curved ,

-

thick vanes revolving concave t o the outlet I t had only .

o n e inlet which w as so constructed that the air entered

, ,

n o t parallel with the fan shaft as i n all the other fans ,

tested but at right angles to it I n this way a rot ry

, . a

motion was given to th e air before i t came in contact with

the revolving vanes Th e obj ect of this arrange m ent was .

t o m inimi se friction Th e e x periments however show .

, ,

that it was of little u se Th e outer cas ing was o f cast .

iron of volute form There were two cast iron bearings .

-
,

in diameter about 8 in long N ine e x periments were

.
, . .

made with the usual baffles I t wil l be seen that the .

three effi ciencies occupy a low position in the plotted

results shown i n fi gs 4 8 4 9 5 0 .
, , .
 1 00 C E N TR I F U GA L FA N S

H
.

6
8
.
.

9 p
4
.

G
I

"
3 a
i
.

I s m
T
l s U

I
l E S
l T $
/
S
5
/
3
'

N
5
I

K
N d
O e
n
D f
: o
on 31 3 1
“g S
N
A
F
e
r
o h
t

m
fi 8
.

5
o
z
fi 2
o
£
z 3
.
a
H
: m
0
l
1 O :
8

8
4
.

G
I

F
1 02 CE N TR I F U G AL FA NS
space between them and the outer casing e x ercise a con
s i d e r ab le influence o n the various effi cie cies Th e fi nal n .

inclin tion or angle of the vanes at their circumference has

a

more eff ect on the pressure of the air and less on the ,

mechanical and volumetric e ffi ciencies Th e revolving .

portion o f the fan should always be accur tely balanced a .

D onkin recom m ends continuous lubrication at high speeds

to reduce j ournal friction and allowance should be mad e ,

for the deflec t ion of the spindle Th e pulley should not .

be t o o small or narrow so as t o reduce the slip of the ,

strap Th e friction o f the air inside the casing is often

.

e x cessive and care should be taken to allow its entrance

,

and passage through the vanes and out of the fan with a
minimum o f skin friction C hanges o f direction and .

shocks should be avoided as much as possible .

For m ulr r used i n D on ki n s E xp e i m en ts These formulae

'
’
r —

are e pressed in metric measures which were used by

x ,

D onkin Th e results are tabulated i n English measures to

.

accord with the te x t h

.

H:
w
( l a)
where H = m e t r e s o f air h = p r e ssur e in millimetres of,

water w = w e i g h t o f 1 cu b ic m etre o f air in kilograms at

,

the atmospheric pressure and temperature .

V = 4 x/IT ( )
2 a

where V = v e loci t y of air in metres per sec at the end o f .

the pipe Z and h = pr e ss ur e in millimetres of water

,
.

T heoretical H P Q X wx H . .
*

in chevau x vapeur 75 -

Th w ri t r d
e t m k i t l r w h th r
e oe s n o t h H P r q ir d
a e c ea e e or no e . e u e

t d ri v t h f
o e h ft i l in h v u x v p r O
e an s a s a so h v l v p ur=
c e a -
a eu . ne c e a -
a e

09 8 6 En g l i s h H P . .

where Q : quantity of air delivered in cubic metres per sec .

Mechanical theoretical H P
c a)
. .

e ffi ci ency
P ressure effi ciency 2
v
2
 PR U S SI A N MI N I N G C O MM ISSI O N T ESTS 1 03

where 9 : 98 1 metres per sec per sec and

,
.
,
.
,
v
2
: pe r ipheral
speed .

V olumetric e ffi ci e n cy
Q ( )
6 a
2
v 7
s 2

where e x ternal radius i n metres .

O ri fi ce i n square metres
(
L ( 7 a)
x /
g H
Equivalent ori fi ce ( in square metres )
32 g
O rifi ce x 1 0 8 8 .

A ppended to this was a short account of the e x

e r i m e n t s made by the P russian Mining C ommission i n
p
1 8 8 4 o n the measurement o f air in a pipe by diff erent
methods from a large air holder at the B reslau gasworks
,
-
.

P r u s i n Mi ni ng Com m iss ion Tesi s

s a A spare gas —

holder was used f or measuring the volume of the air I t .

was 8 5 3 ft in dia contained

. .
, cu f t and served to . .
,

ch e ck the other methods adopte d by the C ommission Th e .

tests were probably the best that have been p ublished o n

the mea surement of air by the following methods ( 1 ) b y
anemometers ( 2) by P itot t ubes ( 3 ) through circular and
square ori fi ces Th e practical questions the C ommission
.

e n deavoured t o solve by using this holder and causing the

air t o pas s through a pipe were the following
1 D o the formul ae generally used for standardising
.

anemomete rs in a circular path in still air give correct

results 1 '

2 C an the instrument known as the P itot tube be

.

applied practically f or measuring the speed of ai r ; and if ,

so what formula should be used for calculating the speed

,

and quantity o f air ?

3 May the fall i n pressure between o n e side and the
.

other of a thi n ori fi ce interposed in a pipe be used for cal

c ula t i n g the quantity of air and if so what formu la should , ,

be applied ?
4 What is the loss of head due to friction i n regard to
.

length and diamete r of pipe used ?

 1 04 C E N TR I F U G AL FA N S

A bout eighty careful e periments were made and the x ,

res ults and calc ulations appear t o have been well checked .

Th e cast iron pipe was 1 4 3 in dia and 3 3 ft long

-
Fo r . . . .

st e
pp g and starting the anemometers quickly and accu
i n

r at e l
y an electrical arrangement was adopted Th e vertical .

fall of the air holder i n several places was carefully deter

-

mined electrically Th e fi rst series of e periments i n 1 8 8 4

. x

were made with the air holder at a water pressure o f 2 % i n

-
.
,

but in the 1 8 85 tests the holder was loaded and the pres
sure was increased to 4 % in o f water Th e density and . .

temperature of the air were noted and the e x periments ,

were made during the autumn to avoid the heating eff ect
of the sun upon the wrought iron holder U gauges were -
.
-

fi x ed at di ff erent parts o f the pipe A P itot tube was .

used for measuring the dynamic pressures of air n ot only ,

at the centre and at two thirds of the radius distant from

-

it but also round the inner circumference o f the pipe

, .

Th e circular ori fi ces used in these e x periments measured

7 0 3 in and 9 9 6 in dia
. T h e square ori fi ce measured
. .

6 2 6 in along the side

. Th e rec t angular ori fi ce was
.

i n by 4 4 5 in Th e e x perimental coeffi cient deter

. .

mined f or the circular ori fi ce was 06 4 and for the square ,

and rectangular ori fi ces 0 6 1 Four Casella anemometers .

and one Robinson anemometer were tested Th e C asella .

anemometers previously tested in the usual way at the

,

end of a radius bar and compared with direct measurement

o f air from the holder showed variable errors the e cess
, ,
x

ranging between 7 and 1 3 per cent A nemometer readings .

should t herefore be accepted with caution I n the 1 4 3 in . .

cast iron pipe a considerable diff erence i n speed was found

-

at diff erent parts and in the same vertical plane Th e .

centre gave the ma x imum speed and pressure and the inner
circumference the minimum ; the mean speed of the air
was found to be at two thirds of the radius from the centre
-

o f the pipe With regard to the resistance to the move

.

ment of the air in the pipe used the f ollowing are the con ,

elusions deduced from these e x periments and given in the

report :( 1 ) that the resistance of the air increases as the
square of its speed in the pipe ; the resistance to t h e
air in the pipe decreases as the d iameter of the pipe to the
1 06 C EN TRI F U G A L FANS
although the earliest used may be found ventilati n g many ,

mines at the present day both in this country and abroad,

.

I ts wheel is generally of considera b le diameter carrying a ,

number of vanes and enveloped over most o f the circum

,

ference allowing the air to escape by a single O pen i ng

, ,

regulated by a shutter to suit the ori fi ce of the mine Th e .

air enters the eye and by its centrifugal action it reaches

,

the circumference and passes out at the chimney Th e .

vanes o f Gui b als are sometimes plane and inclined i n the ,

O pposite direction to that of rotation but are usu lly ,

a

FI G . 52 .

GU I RA L FA N .

curved near the outer e x tremity until they become radial .

Th e number of vanes lies between s i x and ten for si z es

varying between 1 9 ft and 4 0 ft Th e breadth of wheel
. .

f o these diameters lies between 4 % ft and 1 0 ft

r Th e . .

chimney e pands from the wheel to the mouth i n order to

x

reduce the velocity and increase the pressure of the air .

Figs 5 1 5 2 5 3 5 4 show two e x amples of Gui b als the

.
, , , ,

fi rst o which is
f 3 9 1
3; ft in diameter and the.l tter 1 9 ft ,
a .

Th e breadth of wh eel o f the fi rst is 8 4 f t and o f th e latter .

,

6 4 ft Th e S er fan fi gs 5 5 5 6 was designed in 1 8 7 8 by

.
, .
, ,
 L IAN C O MMISSI O N
BE G 1 07

P rofessor S er
of the Ecole C entral e o f P aris Th e theory
,
.

of this fan is published in the Mémoires de la S oci ét é des

I ngenieurs C ivils for I t consists o f a circul r plate a

fi ed to the shaft and c rrying o n each side 3 2 curved

x a

vanes each o f which forms part of cylindrical surface

,
a

whose generatrices are parallel to the shaft and whose ,

transverse section is circular Their wi d th is constant and

.
,

it is arranged that inflow shall take place without shock .

FI G 5 3
. . Pm . 54 .

GU I B AL FAN .

The relative direction of outflo w is at an angle o f 4 5 with °

the radius I nflow takes place at both sides and the f n

.
,
a

is pro vided with a vol ute and e x panding chimney Th e .

volute is so designed that the loss of energy at entrance

from t h e c ircumference of the fan is a minimum and the ,

sides of the chimney are inclined at not more than 1 i n 8

to avoid the loss due t o sudden enlargemen t of passage .

Th i s type o f fan is made in si z es varying from 4 6 ft t o .

8 2 ft .Th e C apell fan ( fi gs 5 7 is formed of two

.
,
 1 08 C E N TRI F U GA L FANS
fans one outside the other
, . Th e inner fan consists o f a
a
drum of steel plate closed o n
, on e side if there is a single

FI GS . 56 .
—
SE R FA N .

FI GS . 57 AND 58 .
—
CAP E I . L FA N .

eye I ts diameter is that o f the eye Th e cylindrical

. .

surface contains si x openings usually rectangular spaced

, ,

equally round the circumference having an area n o t less

,
1 10 C E N TR I F U G A L FAN S

'
qu e o
oy ae m nro
A

q eo
'
u
1 9 d Kau ar o gg e

{B q q em n e on

'
qu e o
I ed
. Ao uo r my e
sp as m o u c m u e e m

'

q ue o
. [ B or u e q oe w
S
L
A 4; b
' '
9
B II I s e g u e
'
Ar nb g

G
I n e d Bo
'
n B A oq o
H u; '
uo
T
I
.
Ie
'

W
e uop d H I n gos n

S
T
N
E
M
R oe s I ad
'
.

P
X
°
oe s r e d
' 6 C 2 8 2 6 2 0 5 9 4
.
4} °

0
.
HI 4 18 10 XI I OO IGA 0 3 0 3 7 4 6 7 9 0 2
4 5 2 2 2 1 1 1 1 2 2

E
L
B

n o;
uI m I e d H
‘ ’
AO
go .

p e ac e
) 0
‘
u I m I. ed m ay

m e ; ; 0 ON '
 L
BE G I A N C OMM I S S I O N 11]

o m e w n IOA
o

3 0 co
v
H
s
A r u ba po on pe u Q 6 7
a
0 1 H

"
T a c o I. e d m us i ng s

I c a ru e q aa m uee m
.

S '
qu e o J ed Ke me r o g e
N o II iem o u B uI um ”
A
F
1 us e 1 9 d
'
1
R
E
S

H .

3}
.
b9
T u; 9 0n
'
Ai nb g
W
J n ed B A
'

S
T
N
E
u o e uo p
'

d H
'

w ee n H 01 m —4 V! H O1

M
R
E
P
X
E
was I e d .

' '

JO 3} “ O

E was xe d
L
1 1 19 10 Kai oor e A
'
21 “
B
A
T

oe s 1 9 d
'

1 1 1 1 H
' '
ue } 6 1 5 2 5 7 2 5 N (N d ‘ co
'

Io
'
III III 1 9d
'
A og
3
3
9
3
3
4
0
3
4
3
7
3
0
3
6
3
CD
CO HHH
o u n ce

a ui fi u a
'

5 5 5 3 3 0 7 7 «p ea ce
)0
'
u I Iu l e d me} ; 6 7 8 6 7 8 5 6 b co w
°

’0 0N
11 C E N TR I F U G A L FA N S

'
qu e o
Op r qe m n I OA

'
q ue o
I ed
. Aou e r o gg e
[ B oi u e q oe ui ue e m
'
qu e o I ed .
9
Ko u e r o gge or ai e m 0
-
o u e u1 I m a m 5

'
qu ao
.

S
[ e oi u e q oe N
N
A
F
4; b e II I
‘ “

L oog gxo A gn b g
‘

L
E
P
A I ne de a
'

C -
x nm e qo
II I 1 1 1 ! n o s n o p
H
.

' '

T I H
c

W '
s e q ou I u ;
S
T
N
E 8 5 6 9 2 6 6 7 0
M M
'
03 9 T 9 8 5 6 1 7 7 6 0
4 6 4 6 6 6 8
' '
1 19 “ O 1 7
10 11
R 1 1 1
E
P
X
°
oe s 1 9 d
'

1; u;
E I
.
pe I o i oo le A

u ned e a

L au r fi u a
B
A
T
oe s I od
‘
.

um } 10
ed Ae a
'
Tu rn I .

SII e I o
'
cu I
u i ux 1 9 d A e g
' '
1 14 CE N TR I F U GAL FANS

TAB L E 1 4 .
— I
D ME N S ON S, I C OST , A N D S ERV IC E OF
GU I BAL FAN S .

NO . OF FA N .

D at e in s t all e d J an .
’
77 Au g .
’
82 Se p .
’
86 Ap l .
’
91
P e r o d o f s erv i c e t o
i 15 y rs . 9yrs . 5yrs . 9 mth s .

J anu ary 1 1 89 2 ,
4 mt h s . 4 mth s .

Total c ost of p l ant i n

C ost o f f an al one i n

D i a of
. w h e l i n ft e . 39 3 29 5
B r e ad h t of w h ee l i n f t . 6 89
D i a of
. ey e 13 1 13 1
2 e y es
H e igh t f c h i mne y b ove
o a 35 7 22 9
ce ntr f s h ft e o a
L e n g t h o f m ou t h o f 11 2 1 0 65
c h i mn y e

B re d t h o f mout h o f
a

c h i m ne y
Di o f s h f t i n i nc h es 12 5 1 02 11 C 114 an d
°
a . a

L en g th s h ft b etween
of a 14 5 12 5 12 3 15 5
b i g s i n ft
e ar n .

P u ll y d i
e i n ft a . . No ne No n e

W i gh t
e of m ov i ng p ar t s
i n lb .

N umb er of st m cylea

in der s
Dia . pi s t o n in i n c h e s
of 26 7 24 4 24 4 165
Strok e i n i n ch e s 33 4 39 4
Dia of . d r i v i n g p u ll e y N one No ne 13 1 N one
i n ft .

D i st ance b e t ween sh aft N one No n e 26 2 N on e

c e nt re s i n ft .

Si z e o f e n g i n e h ous e i n 42 6 x 55 7 x 10 6 x
°

13 1 16 4 32 8
 B E LG IAN C O MMISSI O N 1 15

TA B L E 1 5 .
— D I ME N SI ON S C OST , , AN D SER V I CE OF
SE R F AN S .

N0 . or FA N .

Da t i st ll d
e n a e J un e
’
84 J ly 8 7
u
’
J un e
’
8 8 J an 9 1 .
’

Pe riod f s rvi ce t o e o 7 yrs . 4 y rs . 3yrs . 1 yr .

J nu ry 1 1 89 2
a a ,
6 mth s . 5 m th s . 6 m th s .

T t l
o a st f p l t i n
co o an

C st f f
o l n in
o an a o e

D ia of . he e l i ft w n .

B r e a d h of th el in w e ft . 09 1 7 0 7 86
D i a o f e ye
.

t
H e ig h of ch i mn ov
e y ab e 29 5 19 6 5 226 369
ce ntr s t
e of h af
en t
L g h of mout of h 5 15
ch i mn y e
B r d th
ea f mo u th of o
c h im y ne
D i o f s h f t i n i n h es
a . a c
L en g t h o f s h f t b t w n a e ee

b ri n g s i n f t
ea .

P u ll y d i e i ft a . n .

We ig h t of movi n g p a r ts
in lb .

N umb r e of ste m a c yl
in d rs e ta nd me

Dia . of pi ston i n i n h es
c 19 6 13 8 118 1 5 1 an d
19 7
St r o k e i n in ch e s 27 5 29 5 25 6 81 5
D i a of . d ri v i n g p ull e y 17 3 1 05 13 1
i n ft .

D i s ta n c e b e tw e en s h af t 19 7
c e nt r e s i n f t .

Si z e o f e n g in e h ous e i n 26 2 x x x x
19 6 19 6
116 C E N TR I F UGA L F A N S

TA BL E 1 6 .
— D I ME N SI ON S C OST , , A N D SE R V IC E OF
C A P E LL F AN S .

No . or FA N .

e nst ll e d
D at i a M ar .
’
89 J uly ’
89 J u ly ’
91
er
P i od f s vi o er ce to J an u 2 y rs . 2 yrs . 6 mt h s .

ar y 1 , 1 89 2 10 m th s . 6 m th s .

To t al c o st
o f p l an t in
Co st o f f an al o i ne n
D ia f . O wh l i n ft
ee . 12 3 118
r
B e ad t h of w h e e l i n ft . 52 5
D i a of . ey e
2 e ye s 2 ey es
H ei g h t f h i mn y b ove
o c e a 12 45 72 14 2 5
c tre f s h ft
en o a

L n gth of m ut h
e f o o 1 05 119
h i mn y
c e
Bre dth f mouth of
a o

h im n y
c e
Di a f s h ft i n i n h s
. o a c e 1 0 4 an d an d
78 6
L e ng t h s h ft b twee
of a e n

b ri n g s i n ft
ea .

P ull y d i e i ft a . n . 2 95

We i g h t of m ving o p ar ts
in lb .

Nu mb er of ste am c yli n 2 ( c r an k s
d rse at r ig h t
a n gl s) e

.s n n n es
D i a o f pi t o i i c h 2 05 15 7
t n es
S r ok e i in c h 31 5 23 6
. vn n t
D i a o f d ri i g p u ll e y i f . 14 4 13 1
s n w e s t
D i t a ce b e t e n h af 29 5 24 6
n s t
ce t r e in f .

n e us n t
Si z e o f e g in ho e i f . 23 x 29 5 36 1 x 12 1 26 2 x 22 9
1 18 C E NT R IF U GAL FANS
only e x periments
E xp er i me nts wit h Rateau Fans — Th e
that we know of made with a variable number o f vanes
, ,

are those by M Rateau described in his work C on

.
,
“

siderations sur les Turbo Machines in which a small fan -

,
”

08 2 ft in diameter was tested for water gauge with 18 2 4

.
, ,

and 3 0 vanes Th e number of vanes that gave the best

.

result was 1 8 but the di ff erence w as trifling Th e highest

, .

water gauges were 5 3 4 5 3 and 5 08 millimetres and the, , ,

TAB L E 1 8 — R ATE AU FA N TEST WI T H D I FF US ER ON L Y .

D I A ME T E R , 4 59 FT '
.

a
45

s E
d

s
.
5
g ag;
a
;
Be lg .

'
:
5: 3
o

3
H
as

3
2
°

0 “
t
o
# u 5 3
3 é
‘ “ M

‘ 5 2
5 es s
Q . In .

1 4 05 3 17 19
2 4 05 4 54 076
3 4 05 5 49
4 510 690 45
5 411 64 1 18 2 37
6 5 18 802 35 0 55 8
7 4 14 68 2 26
8 4 81 7 91
9 4 05 7 52 15 6 2 16
10 458 850 23 1 65
11 4 02 827 141
12 4 25 872 168 2 47 2 01

lowest 4 6 5 4 8 9 and 4 4 with 1 8 2 4 and 30 vanes

, , , , ,

respectively U nfortunately no tests were made of

.
,

mechanical e ffi ciency I t is e x tremely probable that a

.

fewer number of vanes would give a better mechanical

effi ciency although the water gauge would probably fall
,
.

Th e e x periments i n Table 1 8 were made with a Rateau fan

having a diff user but no volute I nflow took place from
,
.
 EX P E R I ME NTS WI TI I R AT E AU FANS 1 19

one side and the wheel diameter was 4 5 9 ft ( fi g

,
. .

Th e trials were very caref ully m ad e as we see that , ,

although e x periments 3 and 4 5 and 6 7 and 8 9 and 1 0

, , , ,

1 1 and 1 2 were made at diff erent revol utions but each pair ,

with the same baffle the reduced ori fi ces for 3 and 4 are
, , ,

very nearly the same as also their manometric effi ciencies

,

FI G . 59 .
—RA TEA U FA N .

and the same may be said o f each o f the other pairs Th e .

mechanical which is the indicated effi ciency i ncreases

, ,

with the revolutions and power i n the three fi rst pairs and ,

is apparently the same in the last two Th e increas e is .

j ust what we might e x pect as the effi ciency of the fan

,

alone is constant and that of th e s w i m increases with the

,

power so that at a constant ori fi ce the effi ciency o f the two

,

should i ncrease with the power .

1 20 C ENT R IF U GAL FANS
Table 1 9 shows th e results of a series o f e x periments
m ade with a Ra teau f n of 4 5 9 ft diameter o f type A
a .
, ,

T able 5 tested O ctober 1 8 t h 1 8 9 1

,
Each e x periment , .

lasted three minutes Th e water gauge was taken i n an

.

ou t let from the fan drif t sheltered from the curren t Th e .

discharge was measured at t h e top of t h e chim ney w hich ,

was divided i nto 3 6 equal areas Tests were also made to .

TAB L E 1 9 .
— RATEAU FAN T EST :TY P E A , T AB L E 1 .

FT . D I A ME TE R .

: 5 8
:
"
g
'
a.
Q a g
a 76 8 3 d d 3
g Re
ri fi
u ce
E
55
Eg w:
‘
g
53g g3 2
O
5 o ce

:
, I. .

£ 5 a 0
O
was Cr o
35 5 %
s m a 2

— .

7 22
V II fi

1 26 9

17 0 5

3 72 2 65 5

0 28

fi nd the density o f the air so th t the correct value H ,

a

might be found from the water gauge I t will b e seen i n .

t h e table that the water gauge h s been corrected “ ”

a .

T his means that the gauge due to the velocity o f discharge

has been added and that due to the velocity o f inflow
,

subtracted ; while the latter was correct the former was ,

wrong because the v elocity of discharge being of no

, ,

value ought not to be reckoned to the credit o f the

,
 1 22 CE N TR I F U GA L FAN S
'

q u e o r ed .

AO u a e mei
op i e m o ue m

q eo
'
u
I
. ed Ao u e r org e

19 0 111 9 119 9 K

1
31 b 9
' '

E “I 99
9 11 0
L zu e j uA I n b g
B
A
T
,

A
E a r e n o e uo
p
P 4 5 9 4 9 0 8 0 0 1
) j xo m I n ge s n 3 3 3 4 3 4 2 1
Y
T .

J n ed P A
'

I C
R
E a a;
A9 qo
-
S T
T
S
E '
00s s ad
'
In
T 4}
[ Ga m ma at“
N ape 10
A [ 1 300 19 11 {r oo m
F
43 b a 6 vs
' '

U II I Ao u m r q o

T '
no
A n;
R
.

0 Pozoau oo
2

E
'
se q ou r II I
L e fi n e fi J eanna
B s sq o
p e AI
A
T
m e ; To
u r m 1 9 d A og
‘ '

I Bu e go
'
e II
u i ux J e d A e g
' '

0a m ‘1 ! 9 10 11

3 93 3 JO ON
'
H O‘ W V‘ IQ CO P Q OD O
 E X P E RI ME N TTS WI TH RAT EA U FA N S 12 3

m
'
aog I-I o H m m b fi
w m co Q
- CD O
O O O O O H

O L
‘ CD N N CN
—I
J ed Kau ai o tg a
—
OD OD O v- I v r -I

or n o q
H H H I l
uB N

'
q n oo
I
. ad Ao n e r o gi e

[ Boi ue q oe j q

J OM OCI s ar o q
'

P ‘WMP U I

' '

d H
.
‘
J I E u o on o p
S 111
. 0 am [ up s n
T
S
E '
eq e j d o g r e ] III
T (O O H IO N H CO O
OI O q oi oo m xo
m m m m fi m cfi
g
q
N
A
R
.

:
qu s j c ai n b a
0 o ne
s
H H H H H H H H
E
F T
E
U
A
M
A at
'
1} b9
'
a
m
m
m
m
m
o n ia
O
“)
N — i
I IO CO CD N LQ OD v- I lo
E qu e IBA r n b fl H v- I v-I C
N N
T D
A .

T H r -I v- I O O
R F co m m en ce
w
1
.
7
1
H
v- f
H
fl ‘
H H
lo co
H
2 9
Ke u m r q o
'

L eqi
B 9 31 9
2 09 1? ) 0
A KQI OO I GA U B GN
T

( p e i oe u )
'
oo

e u gfi ue
'

JO
u r m I. ad A e g
' '

H N CO V‘ IO CD L
‘
QD
 1 24 C E N TR I F U G A L FANS

3 11 9 0 I od
'
.

Ko u s r o gj e
r -I v- i v- ‘l v- l

quoo I e d
'
.

Kau aj o ry e
[ e o r u e q oe pq

J OAAOd
'

ae xo q

. P e ze oi p ui
S
T
S J e m od
'

E
e sao q u;
T
1 19
. n o e rr op
.

R
A E
F T
E
b8 O
‘
.

3;
.
l Ol O 0 0 0 0 0
U
’
05 9 ] v 0 4 4 1 2
II I s o g n o
co m m e r 4 0 0 1 3
qu e I B A I n b g H H CN CN 3 2 2 2 2
T
A
R '
Oe s 1 e d 10 1 63 a 9 0 5 5 8
.
co oo v- I co oo 9 8 3 0 2
© w
’

31 “ I
G O ¢ Vfl 8 6 7 3 9
2
2
.

238 1 q q H H H H 1 1 1 1

U
L
B 9 11 1 1 13 1 11
3
A go Ki r oo le A
T

( p e qoe u )
»
'
oo CO P- H OO G3 0
2 00
9 . 9 c 6
m co m m m 1

o m b oo s fi 8 0 2 6 6
w
H H H H H
OZ ‘ CD ‘
t f‘ lQ 4
1
8
1
9
1
3
1
8

2 991 :H 0
°
O
126 C E NT R I F U GAL FANS
plate I t will be noticed that th e equivalent ori fi ce is
.

always greater than the O pening i n the baffle plate i n the

ratio of about 1 2 t o 1 1 his is partly due to the passages .
'

behind the ori fi ce but also to the fact that the coe ffi cient
,

of contraction may be greater that I t probably

increases with the si z e of the ori fi ce .

Table 2 2 gives e x periments on a fan o f the sam e si z e

at Mon t r am b e r t made on D ecember 6t h 1 8 9 1 Th e
, , .

engine had a cylinder 1 4 % in diameter an d a stroke of .

,

23 6 A ll the horse powers given with fans of this type

.

are chevau x vapeur and consequently 09 8 5 of 1 Eng lish

-
,

horse power .

Table 2 3 contains e x pe r iments o n a Rateau fan 1 3 ft , .

1 % in diameter at the C o n solidation Mines Westphalia

.
, ,
.

T h e fan was guaranteed to give cu ft of air per . .

min with a water gauge of 6 in and an equivale n t ori fi ce

.
, .

o f 2 8 sq ft w h ich corresponds to a reduced orifi ce of

. .
,

0 6 very nearly te —
,

Q 06 — — -

r 2 «(I H
.
,

In making this calculation we must remember that the

equivalent ori fi ce is 1 0 88 the ori fi ce Q z g H — .

Centr if ugal P ump Exp er i m en ts by Char les H . L i nes .

”

These e x periments were made by the present writer on

March 1 3 th 1 8 9 7 and J anuary 1 5 th 1 8 9 8 at Wallsend
, , , ,

S lipway Tw o pumps are provided to empty the com

.

s dry dock but o n e only was used i n th e e x p e r i

’
an
p y ,

ments Th e diameters of suction and discharge are 3 6 i n

. .
,

that Of th e wheel 66 i n while its internal breadth at out .

,

flow is 5 % i n ; the internal diameter is 3 9 i n and th e

. .
,

vanes are radial at the inner circumference and curve back ,

in the arc o f a circle until they become t ange n ts to the

ou t er circumference I n both ex periments special care .

was taken in closing the gates to m inimise leakage which ,

was found to be about 3 0 cu ft per mi n ; this quantity . . .

i s small in comparison with the total quantity delivered ,

hence it i s neglected in the calculations Th e suction .

pipe is 1 5 ft 6 in long and 3 6 i n diameter and the dis

. . .
,
 C E NT R I FU GAL P U MP EX P E R I M E NTS 1 27

charge pipe enlarges with a be nd t o 5 4 in diameter at the .

j unction with the discharge pipe o f the second pump .

Th e remainder o f the discharge pipe is 9 5 ft i n all 5 4 in .

,
.

diameter with a right angle bend We do n ot thi nk that

,
-
.

the introduction of these e x periments needs any apology ,

as e x cept for the fact that i n a centrifugal pump I should

,
( )

not b e greater than 9 0 deg the rules for designing pumps .

,

and fans are precisely the same and deductions fro m ,

e x periments with the former apply to the latter Th e .

weight o f water was assumed as 6 2 % lb per cu ft in the . . .

"
fi rst e x periment and 6 4 i n the second I n the former the
eff ect o f t e tide on the density of the wate r w as forgott en
.

and i n the latter a sample o f the wate r in the dock whose

density it was intended t o test did not reach the Ruther
ford C ollege owing to an accident and the author desired ,

t o m ke every allowance for the pump whose e fii ci e ncy

a

was lo w ; but it must be remembered that the effi ciency

takes into account the friction of the pipes so that that of ,

the pump alone is greate r Th e losses of head in pas sing

.

through the pump are

( )
1 A t inflow ,

( v1 b1 cot
2
0
1
’
2g 29
since 6 : 90 deg .

( )
2 At entrance in to the volute ,

62 2 + ( a2
2 g

( 3) L osshead other than shock loss and losses due

of ,

t o bends and pipe f r iction .

Th e coeffi cient of resistance of the pump alone is here

t ken as F = 3 referred to the veloc ity of discharge so
ti at
,

’
c,
2

Th e losses due to bends and pi pe friction together with ,

t he loss due t o the radial speed namely b, —2 g ,

”

2
0 02 3 3 00 .
 1 28 CE N TRI F U GA L FA NS

TA BL E 2 4 . C EN TR I F UGA L P UMP TR A I

1 F all of w ter i s i d e d
a n oc k 0 1 2 3 4 5
in ft .

H e a d outsi d e d o ck in ft . 19 8 19 8 20 O 20 0 20 C
an d in .

H e ad i ns i d e d o c k at en d 18 O 17 0 15 0 14 C
of i nte r v al i n f t . an d

in .

A I B o i l e r p r e ssu r e 1 20 1 13 1 05 1 10
I
O I n t e r v al i n m i n . 73 73 7%
O
O Me an h e a d 2g , 4,
Q I H P 2 20 7 23 9 0 2 3 3 2 5 2 4 0 3
- -
. . .

m 12 12
-
22 1-
34 1 46 1
-

— F H P 9 14 10 1 10 15
-
D
C . . .
- -

O S H P 211 5 3 2 2 9 5 2 2 3 -1 23 0 2
l
-
f . . .

-H H yd r au l i c e ff i c i e n cy , 5 74 9 3 2 1 5 2 7 20 4
‘ - -
-
1
]
WE
l

P
S H P
en t
.
. .

. .
.

p er c

H0 [ C l u l t d e ffi i n y
a c a e c e c , no 14 15 17 9 211

13 7
H L
— 7
p
- c n t

0
E er e .

7 7
14 R e v pe r m in . . 1 25 5 13 65 1 4 03 144
15 V e l o ci t y of d i s ch ar g e 12 7 2 1 2 7 5
fr o m p um p
16 r
O i fi ce sq . ft . 16 25 10 8 86 6
9
Ti m e 9 45 9 52% 1 0 0 1 0 7 % 10 1 5 10 2 2
18 Q u anti ty d i s c h a r g e d dur
ing i n te rv al i n cu . ft .

g H
19
*
x 1 00
@ 20
2

Th e an g e o f l r e l ati v e d is ch a rg e fr om t h e f an i s as s u m ed to be at
°
26 2
v ari e s b e t we e
 130 C EN TR I F U GAL FA NS

TA B L E 2 5 .
— C EN TRI F UGAL PU MP T

F all w t i ns i d of a er e 2 3 4
d oc k i n ft .

H a d ou t s i d e d ock
e 21 0s 2 0 9 20 2 19 9 l9 6
in f t d in . an .

H e d i ns i d e d o k t
a c a 19 0 18 0 17 0 1 6 10 15 0
d f i t erv l i n
en o n a
ft d in . an .

h B o i l r p essure e r 1 10 1 10 1 10 1 10 1 10
U I nt rv l i n m i n
e d a . an 114 5 ll O '

se c .

>
O Me a n h e a d i n f t .

Q I H P
. . . 1 23 6
m 19 7

— F H P
O
C . . . 7 08
O S H P
I

t . . . 7 6 69
HH H ydr au l i c e ffi c i en cy 21 3
WH P
,

per
. . .

’
7
S H P . . .

c ent .

12 Ca l ul te d effi
c a c e i nc y ,
22 5
H '
r
; c p a1
H L
c ent .

n
-
nc +l '
l
14 Re v per mm
. .

15 Ve l ci y o t o f di s ch ar ge 86 5
f rom p um p
Q
16 Or i fi c e i n sq ft
/g H
. .

N
17 Tim e i h ou an d n r s 9 41; 9 5 3; 10 5 10 1 6
m i ut s n e
18 Q u nt it y d i s h g d
a c ar e 38
d u i g i nt rv l i n r n e a

en . it i .
° '

v t
I n t h e abo e t h e w e ig h of wa er per cu t .

1
‘
Thi s d oe s n ot i n cl ud e l eak ag e , wh ich at t h
 C EN TRI F U GAL P U MP E X P E RI M E NTS 13 1

BY C HARLES H . I N N ES

2 18 10 15 1 1

4
5

6 6 7 07 7 37 5 8 04 1 9 3 7 5 1 05 8 2
7 1505 17 6 0 2 62 6 27 6 0
49 9 59 6
10 4 13 1 3
1 7 8 2 5 2 06 6 24 9 4 7 2 59 0
3 26 37 9

29 5 5 32 0 33 2 336 3 56

128 0 132 5 1 4 01 1 5 02
103 5

4 99 4 72 4 50 4 32 4 21
- -
3 77
- - - -

1 0 4 5; 1 0 5 4 ; 1 1 3 ; 1 1 1 24 1 1 2 1; 1 1 3 8 ; 1 1 5 5
5 12 12;

t
ak e n at 64 lb .

of the ti lw
r a as f oun d t o b e 3 06 cu . ft .
pe r m i n .
132 C EN TR I F U G A L FANS
C ombining these , w e have —
loss ( 3 ) together with the
loss d ue to radial speed = 0 07 I n both sets of trials '

only o n e pump and engine were u sed Th e friction of the .

engi ne was taken by run n ing the engi ne unloaded at

speeds varying between 1 2 2 and 1 8 5 rev Th e H P . .

absorbed by the engine under these conditions is termed

the F riction H P and the horse power transferred t o the
. .
,

pump shaft is
SHP . . . I HP . . .

T hisis not e x actly correct because the friction of a loaded ,

engine is a little more than that of on e running light but ,

the error is n ot very great S ee Tables 2 4 2 5 pp 1 2 8 1 3 1 .

, ,
.
-
.

Th e following is the method o f comparing the effi ciencies

obtained from e x periment and calculation I n the fi rst tri l a

at the 1 4 th foot we fi nd the useful work ‘

done by the pump i n raising Q cu ft of water i n the given . .

62 5 x 3 6, 3 4 1 x 1 3 87 5 '

x
rev per min
. .
( these were taken b y a counter
read every minute ) mean head this was measured ,

by two posts outside and inside giving heights i n feet ,

above the sill of the dock and each i nterval commenced ,

at a foot on the inner scale O n the outer scale inches .

were m easured by a pole which h ad alternate inches at its

end painted in black and white f o r a foot length and this ,

held against the post enabled very accurate readings t o be

taken as the motion of the water was not more than 4 in
,
.

o n March 1 3 th and n o t more than 2 in o n J anuary 1 5 th . .

We mention these details as we believe there have been no

,

e xperiments on pumps made w ith greater care than these .

Th e quantity o f wa t er discharged was cu ft during . .

an interval o f min This gives a velocity of discharge

.

o f 8 1 7 ft per s ee from the pump and as F H P

. . at , . . .

rev was 10 4 and at 1 60 rev was 1 6 by interpolation at

. .
,

it was
H ence SHP
. .I HP F HP. 24 1 5 . . . . . .
'
;

P ump effi ciency 0 37 7 5 .

13 4 C EN TR I F U G AL FA N S

N N b
a a
c m a
s
o e e

m w
.

m O 9

Z d é m m
o o o m
O
m
m
a
fi
H
.

O
.

m
W
m
w
B
Z
Q
S
m
m
m
x
m c
.

5 a m o
A 5
€
.

S £
3
.

8 0
3
N
D
.
9 6. 3 5

Q
Q
d C
c
"
o
m
b m
a
5
.
8
£
8
2
0

6
N

E
d
wi
n

a
.
 MR . P A RS ON S
’
E X PER I M EN TS 135

at 4 10 rev ( the mean of the number duri ng the e x

.

p e r i m e n t s )
,
while the pump was not disc harging I t w as .

found that f t lb per min were required and this

.
-
. .
,

amount was deducted from the work done by the engine ,

and the work done b y the pump being divid e d by this was
called the corrected e fficiency .

Th e corrected efficiency 73 is more than the hydraulic 7

efficiency because the work required to drive a pump when

,

n ot discharging is very much g reater than the work

expended in overc oming the friction of the fan shaft and
the s urface friction of the disc Th e real hydraulic .

efficiency obvi ously is between this corrected efficiency “

and the ratio of the work done by the pump t o that done
by the engine I t will also be seen that the calculated
.

efficiencies do lie between the latter and P arsons corrected ’

efficiencies and are therefore close to the real e ffici e n ce s of

,

the pump T h es e e x perimen t s give us reasonable confidence

.

i n asserting that the efficiency of a pump is

_ 9 H

"
7 _
71 (1 v2
2
H
or
P IT I E

where L are the loss es of head in ft

Th e method of calculating n, and n2 is as follows :
.

Th e
dimens ions of the pump are g iven in a pape r on C entrif ugal
P umps by P rof U nwin “ Th e e x te rnal radius o f the fan
. .

is r : the i nternal r adius T h e breadths s2 .

,

an d 3
1
at the exte rnal and i nt e r n al radii are both 5 7 5 in
, , .

T here were eigh t vanes and as their thickness is not ven

'

, ,

they are assum e d to be f i n at their ends Th e v e oci t y . .

c i n the volute is given b y P rof U nwin as 3 b 2 but as he

u .
,

evidently neglects the vanes this must be modified , .

As suming a coefficient o f contractio n of at discharge

from the f an which is also the custom in rad ial flo w tur
,
-

bines 15 we get
,

cosec
0 = K = x 3 b2 2 3 5 b2 ;
°
v
136 CE N TRI F UG AL F AN S
and since
( 2 '

71 3
2
n 3
2
12 cosec K : 61 ( 2 s
l
a 3
1
t1 cosec 6) K,
72 t cosec 1)
6
(2 7r r
2 2
< 1

where W ,

n : 8, the number of vanes ,

6
°
15

L e t G : gallons per min .

,

then bl =
60 x 6 2 5 ( g m . -
3 -
71 3 t1 cosec 0) K
1 1 1

and
Gr .

Taking the first experiment as an examp le G = 1 01 2 , ,

62 1 4 9 C , v0
2
02 0
2
6 2
cot
’
I and assuming ’
( ),

the relative angle of flow coincides with the angle of vane ,

g H
100 nl =
a
x 1 00 = 57 5 per cent .

2 v2 X

Th e losses of head L in passing through th epump are

Th e loss at entry to the wheel
( v 6 cot Of
‘

1 1
2 37

the loss at ent ry to the volute

)
2
b2 +
( 0
2
00

29

and the surface f i ct i on ==F r Th e most suitable value

of F is 2 5 so that,
138 CE N TR I F UG AL F AN S
having an expanding rim ( fig I n other respec t s it is .

a s i mple open running fan with the vanes curved so as t o

-

be convex to the outlet Th e following are the principal .

d i mensions of the fan

D ameter to periphery of divergent outlet 3 6 ft 4 i n
ii . .

D ameter to t h e extremities o f the blades 3 5 ft .

D iameterof i nlet ring 1 3 ft 6 in

Width at outlet
. .

1 f t 1 in
Width at periphery of fan
. .

2 ft 2 in . .

Capacity of fan cu ft . .

Th ee x periments in Table 28 have been selected fro m

those given .

T AB L E 2 8 .
— TESTS RESU L TS :WA DD L E FAN .

No . of e x p ri m n t
e e .

Re v p e r m i n
. f an . of 50 57 2 59 7 3 372 67 4 7
C u f t of ai r p e r m i n
. . .

Wate r g au g e 26 7 1 1 0 82 3 27 4
U sef ul wo k r d one by 45 6 8 64 5 973 3 18 51
f an , H P . .

I H P
. . . 27 7 2 139 94
E ffi c i n
e c y 06 5 3 0 64 1 0 67 3 0 668 0 64 1

In t h i ri m e n t the se p ar ati o n d oo r s we r e o pe n ; t he o t h e rs
tpe m i n e
s ex e

"
we r e m ad e on .

Mamine Ven ti lati ng Fan E xp er i me n ts by M . L elong .

M L. elong made a number of very interesting experiments

in order to obtain data for the calculation of the dimensions
o f ventilators for warships Th e first quantity requiring .

formul ae w as the resistance o f t h e circuit through which

the fan discharged its air This is made up of ( 1 ) surface .

f r i ction changes of section ( 3 ) changes of direction

, ,
.
 M . L ELON G S E
’
X PER I M EN TS 13 9

( 1 ) Surface F riction loss of head due to friction

.
— Th e
may be ex pressed by the formula (put i n our own notation ) ,

l 02
'

Hf = f
77i 2 ¢

Fr o 61 —
. . Wanna :FAN .

where l is the length in ft m the mean hydraulic depth

.

equal to the area of section a divided by the perimeter s

, , , ,

and v the velocity in ft per sec is a coe ffi cient whose

. .

mean value is about 00 06 C alling .

2 9 H!
Q 2
1 40 CE N TR IF UG AL F AN S
the r esistance due to this ,
we get
29
w l8

( ) C
2 hanges of Section When a passage ends in a very .
—

large space the kinetic energy of the current is completely

,

lost and the corresponding loss of head is

,

’
29
while the resistance due to this
2 g H} ; 1
Q 2 ?
a

I nversely if the current of air flows from a large space into

,

a cylindrical pipe we usually al low a coe fficient of con

,

tractio n c 0 8 3 for the vein entering the pipe so that the

c , ,

loss of head here becomes .

2
v 1
H6 : " 1 ’
29 (0 29
and the corresponding resista nce
2 9 H6 0 45
Q
2 z
a

M L elong
. made several experiments i n order to find the
value of the coefficient applicable to rectangular passages of
large dimensions such as one finds on board ship Th e
, .

e x periments Were made with passages of two si z es Th e .

smaller w as made by dividing on e f t by 2 6 2 ft . .

section into two parts by a longitudinal partition They .

were 9 8 4 ft in length and were made of carefully planed

.
,

wood one end being connected with the atmosphere whilst

, ,

the other was enclosed in a chamber receiving the air

delivered by the fan Th e static pressure i n this chamber
.

was given by a manometer ; the discharge was given by an

anemometer Th e total resistance includes not only that at
.

inflow but also that due to friction and the loss of the
,

kinetic energy at discharge ; we have therefore deducted

these two last using 0 004 for the coefficient of friction
, .

Th e results obtained are given in Tables 2 9 3 0

.

.
,
142 CE N TR I F UG AL F AN S
Fo r t hesecond passage th e results given are the means
between those obtained for each of the two passages .

By connecting the e x it or entry of a pipe to the larger

space by means of a cone the l oss of head is much reduced , ,

and can become z ero I t is negligible if the angle at the

.

verte x of the cone does not e x ceed 3 0 deg when the passage .

is to be reduced and 7 deg when t h e passage must i ncreas e

,
. .

( )
3 R esistances due to C hange of D irection — Th e loss of .

head at elbows is probably due to a contraction formed by

the stream Taking the loss as that given by P eclet
. ,

‘ 2
U
H e
r- 1
92
5

where I being the angle made by one pipe with

( )

the prolongation o f the other

Weisbach gives the
.

Fo r bends the loss is much less .

following formula
2
v
Hb = ( b
where
{b 1 8 47

for pipes of circular section diameter d and mean radius of ,

curvature p and ,

:
(b 0 12 4
-

for rectangular sections where s is the length of the side of ,

the section parallel to the radius of curvature p According .

to these formulae the loss of head does not depend on the

,

total angle of the bend but on the ratio of s to p , .

Several e x peri men t s were made by M L elong t o see if

Weisbach s formula coul d be applied to large rectangular
.

sectioned ventilation passages such as are found in warships ,

.

W h v u u ll y
e a e t h f o r m u l m u lti pl i d b y t h f r c ti o n
s a se en e se ae e a, e a
o f t w r ig h t n g l
o f t h b nd b t v
a es o t h i m o d i fi c ti o n d
e e t u e en s a oe s n o

g r m n t b tw n W i b c h f o r m u l
,

b r i n g b ou t d t h r ult
’
a an a ee e e ee e s a s a an e es s
i th t bl
n e a e .
 M . LE LON G S E
’
X P ER I M EN T S 14 3

Th e passages used were s i milar to those describe d above .

A t the ends of the bends it w as necessary t o add a length

o f pas sage o f 3 me t res , so as to
obtain a uniform outflo w
of air which appeared t o be e x tremely i r regular after
,

leavi ng the bends Table 3 1 gives the results of e x pe r i

.

m ents and compares the coefficients ( b thus obtained

with those deduced fro m Weisbach s formula Before the ’
.

air entered the bend it had to pass through 3 metres

of passa ge and after l eaving the bend through the same
,

dista nce .

T h e e xperimental values of ( b are generally less than

those g iven by Weisbach Th e greatest discrepa ncies are .

to be found i n e xpe riments ( 4 ) and T hese may be

e x plained by the s m all angle of the bend of which Weis ,

bach s formula as given by M L elong does not tak e account

'
. .

Th e figures S how however that the total resistance is very

, ,

nearly indepe ndent o f the discharge and that the loss due ,

to bends is less than that due t o sharp corners if we accept ,

the formula 0
2
H = sin ¢ ,

I n experiments ( 5 ) to ( 8 ) the passages were divid e d by

vertical longitudinal partitions and expe riments ( 5 A ) and ,

( 7 A) refer t o those having the greater radii of curvature p .

Th e principa l d i fli cult y in making such experiments as

these se ems to b e that the resistance of th e bend i s only a
small pa rt of the whole and therefore as it is obta ined by
, ,

diff erence a large percentage error is possible

, .

Fan Exp er im e nts by M Lelong —Th e first fan tested was

. .

one designed for the D a Cha yla I ts dimensions were as

follows :Ex ternal diameter of wheel 5 2 5 ft diameter of
.

,
.

eye 3 2 8 ft number of vanes 24 ; width of vanes at the

, .
,

oute r circumference 0 49 2 ft Th e casing was a volute

,
.

whose sections were calculate d by the formula

8: a s ,
2 11

where Q d i s ch ar e in cu ft per sec ¢ = angle measured

I n radi ans f r om tge commencement of t h e volute t o the
a
. . .

.
,

l
144 CE N TR I F UG AL FAN S

00 00 00 0)

H
m m
H w w
e e
a m
m m . ::
a
a
;
a

—I H
lQ L’ D l Q lO

—
9
L
o
h
O
m m
o o
w
o
n
o
n
o
n

H H H r H H v I

H v- I
‘ v- { v —l r-f r -t r- i r -I ‘

H H V —I H H H H

.
a
w
. . .

3 m 2 m m me b
2 o 2
8 m8 o
w
3 a s ®
8w 3
0 w m o
5 ;
o
3 s n
o
m 5
w
.

8 a
3
0 o 0 0
3 2
o o
m
“
8 w n n
3 fi 2 3
o
w e 3 3

~
m% m % mc w% % 3
3 3
a a
3 H
d a d s
a
x 3 a a a o a
- 3 3 3
3 fi -
fi m
6 £ s s 6 6 $ 6
M
u i m
o £ m
o § fl
0 0
g m 9 5
m m A m m m fl 5

o 0
0
5
3
o
s ? 3 3 m m m E m€ fi 5
3 o i i
0 0 0 0 0 a
9 S 3
E E 8 S 3 fi 3
a o
m
m 5 g 3 g q B
.

3
m H
. o 8 o o R
. 8 o o 9 8
1 46 N
C E TR IF UG AL F AN S
fig 63 represe n ts the characteristic of this fan Curve ( 4 )
.
,
.

shows the e ff ect of doing away with the diffuser I n both .

these fans M L elong considers that the manometric e ffi

.

ci s ney is the same as the mechanical efficiency on account

of the o utlet angles of their vanes being 9 0 deg This . ,

0 -
1 -2 ~
23 4
' '
5 -6 «
8 1
-
3
-

Re d uc d Or i fi ce s :
e O+ r 3\/2g H
FIG . 63 .
— CH A ACTR I
E R S T C OF I F AN S T E STE D BY M L E L ON G
. .

FI G . 64 — SE CON D FA N T E S TE D BY L E L ON G .

however is doubtful Th e latter is probably more because

,
.
,

I t I s qu i te poss i ble that az < v2 as the angle of flow is n o t

always the same as the angle of vane .

Th e third fan tes t ed shown in fig 65 had two eyes ,

.
,
.

I ts vanes were inclined forward at 4 5 deg to the radius . .

 M . LE LON G S E X P E
’
RI M EN T S 14 7

Th e sections of the volute were the same as in the last case ,

but there was no d ifl use r and this we believe is the reason '

that the m ax imum manometric effic i ency was only about

65 per cent Th e external diameter of the vanes was
.

4 5 9 f t the diameter of each eye 02 9 5 f t

.
, the n umber of .

vanes was 1 6 and their width at the outer diameter of the

,

Wheel 04 5 9 ft Th e characte r i stic curve of this fan

.

fig 6 6 is much higher than the last but its mechanical

.
, ,

efficiency is not any greater than that of the first or second .

I f we assume that the angle of relative outflow from the

~~
wheel is the angle of the vane the e fficiency of the fan ,

alone i n this case should be very nearly

7
9H J 72; 5 1? 6 7 .
’ a '
b2 1
2 v2

7 m 7 m
7 7

Q V 77 m OR 7
‘
V 7 m
7
1 1
2

2 -
11 7
3
3
2n
w /2 se

7 m
7
1 1 1 26 OR

where 03 —
Q and 82 0272
J2gH
2
7
2

T his gives us from cur ve ( 6) the following table

R educed orifice O 02 R

7
)

which shows very clearly the mistake of not having a

p r Op e r diff user to receive the air discharged from the
wheel .

C urve 7 shows the dynamic manometric e fficiency of the

fan Fi g 6 7 represents an open running fan wit h radial
. .

vanes tested with a casing I ts dimen ions were :D iameter

, . s

of wheel f t ; diameter of eye 2 6 2 f t ; number of

, .
, .

vanes 3 4 ; breadth of wheel at discharge 0 3 5 4 ft I t

, , .

gave the characteristic curve 8 C urves 9 and 10 (fig 68) . .

represent the characteris t ics with and without an inflow ,

148 CE N TR I F UG AL F AN S
mouthpiece of a fan constru cted for experimental purposes
, ,

and only di ff ering from that shown in fig 6 7 in the .

F IG . 65 .
— TWO EY E D FAN
-
T EST E D BY M L E L ON G
. .

g
1
1

hc 3
a
e
B 2

6
- 0

Or i fi ce s .

FI G . 66 .
— CH A E AC TE R I ST I C OF T WO EY E D FA N
- .

its vanes at the outer radius ( fig

cli n at i on of I ts .

manometric e fli ci e n cy was less than that of the fan pre «

150 CE N TR I F UG AL F AN S

FIG . 7 0 — M0RTI E R D I A M E T R A
. L FAN .

Eq ui valent Or i fi ce s i n Sq ua r e Me t r e s .

FIG . 71 .
— RE SU L TS OF EX P E R I M E N TS W I T H M ORT I E R FA N .
 M OR T I E R D IA M ET R A L FA N E XP E R I M EN T S 151

v i o u sl discussed , as i t had bent back vanes I f we increase

y
-
.

the height of curve 8 in the same proportion as 9 is

above 1 0 we get curve 1 1 which M L elo g co snle r s
.

. n n
, ,

would have been the characteristic of t h e fan ( fig 6 7 ) if an .

inflo w mouthpiece h ad been ad ded .

Mor ti e r D i a m ctm l Fan Exp er i m en ts I fi g 7 0 i s sho w n — n .

a Mortier diametral fan Th e direct i on of rotat i on i s.

counter clockwise and the air enters and leaves as shown

,

F IG . 72 .
— I LL U ST R A I G M OV I N G 0 F CAS I N G
T N 0F Mo a ER FA N
T o S U I T T II E M I N E .

i
Or fi c e R Ra c k
. an d pi nio n f or m o v in g c ing as .

by th e arrows Th e vanes are bent forward so that inflow

.

takes place without shock and the air at outflow is thrown ,

forwards as well as outwards I ts velocity head at dis .

charge must be considerable this however is partly con , ,

verted into pressure head by the chimney These fans are .

m ade by L ouis Galland of C halon sur Sa6n e Th e results

,
-
7 .
15 2 CE N TR I F UG AL FA N S
of ex periments with a fan of this type 6 5 6 ft diameter by
,
.

ft broad are shown in fig 7 1 i n which the m e ch an i

.
,
.
,

cal e ffici e n cy of fan and engi ne and the manometri c

E N
I
T
L M
T I

N U
S

i
s
o
p
l
N E a
I I n
T T i

A R i
g
R O r
T
M

efficiency are given Th e quantity of air that would be

.

discharged per sec in cu metres at 22 5 rev per min or a

. . . .
,

peripheral velocity of 7 ft per sec is also shown

. .
, .
1 54 CE N TR IF UG AL F AN S

9
2
3
2

R
E
M

G
I

F
 K EY L F AN E X E R C IS E S 1 55

TABL E 3 2 .
-
I I
E FF C EN C Y OF A K L EY FAN .

Ap pr ox m ape
Re v of
Wate r g au o c
C f f
m a n om e t r i c M cc h an i c a1
géf fi c i e n cy per
D
‘

in t h e
gir
ef e ffi c e n c y o f
ii
uc t o n p p c e n t a ss u m i n g e n g n e an d
.

s i i e .
,
p er m i n .

i n inch se c
f an p e r c e n t
es .
l oooo .

TAB L E 3 3 .
— TESTS OF K L EY F A N S ( S M A LL ER TY P E ) .

D i s c h a rg e
in cu ft , ,
Wate r g au g e s i n i n c h e s .

3 ,H
'

5,
min 3
pe r .
'
3
‘ "
“
Co re
s Po n n In
5 . 5
c :aE .
5
92 o
7 89 o
11 s 1 5 75 23 60 . m
m “ .

D ischarge 7 05 880 29 6 8 87 -

Revs .

D ischarge 1 18
Revs . 9 60 1, 3 5 8 1, 5 2O
Discharge 49 2 5 -

Revs . 7 68 94 2
61 5 18 4
Revs . 615 755 87 0 974
D ischarge 87 5 -
26 2
Revs .
156 CE N TR I F UG AL FA N S
These fans are constructed with inflow at one or both sides 3
for forges and foundries from 1 1 8 in to 7 2 ft d i ameter .
'
.
,

and for mine ventilation from 1 6 4 to 3 9 3 f t diameter . .

Table 3 2 gives a series of e x periments upon a venti

lator of this type of the following dimensions :Ex te rnal
diameter 2 9 % f t internal diameter 1 9 6 5 f t ; e xt ernal
,
.
, .

breadth 2 6 2 ft internal breadth 3 9 4 ft

, .
, .

Table 3 3 with figs 7 4 and 7 5 ( for which w e have to

, .

than k C Mehler of Aachen ) gives particulars of the

.
, ,

smaller fans of this type .

P elzer D or tm und F an T his is shown in figs 7 6 t o 7 9

— . .

I t is largely used on the C ontinent and has in late years ,

been greatly improved Fi g 7 9 shows the wheel fitted . .

with twelve curved van es which receiv e the air without ,

shock A fter ente ring the wheel axially the air is received
.

by vanes whic h are plane and radial and with this arrange ,

ment the manometri c efficiency is usually about 5 0 pe r cent

but by alterations i n the con struction a considerably higher
value can be obtained .

F igs 7 6 7 7 and 7 8 S how the diff user vol ute and

.
, , , ,

chimney together with the plant for driving the fan Th e

,
.

aerial pressure on the wheel is balanc e d by allowi n g the air

from the diff user to flow into the conical spaces surrounding
the circ umference of the wheel on the suction side A .

thrust t o the right is thus obtained to balance the thrust

that naturally acts towards the left or suction side .

T hese fans are made with diamete rs between 1 1 8 in .

and 1 9 7 i n While the breadths of the wheel are

.
,

of the dia for a water gauge of 04 inch

. .

0 .

3’ 1, 3, 3’

0 75 2 i nches .

07 3
4
6
8
10
12
14
1 58 CE N T RIF U GA L FA N S

B um stea d an d example of
Chan dler Fan E xp er i m en ts — An
this type of fan is shown in sectional plan and elevation ,

figs 8 0 8 1 I t is fitted with a pair of tandem engines on e

.
,
.
,

on each side of the fan which can be run separately if ,

desired they have high pressure cylinders 1 6 in diameter —

, .
,

low pressure 2 4 i n and a 1 6 i n stroke These engines

-

, .
,
-
. .

when running at 2 20 rev indicate 3 2 0 horse power Th e . .

fan is 1 5 ft in diameter and 6 ft 6 in Wide and is capable

. . .
,

of discharging cu ft o f air per min but as it was . . .

,

FIG
. 80 — BU
.
M STEAD A N D C H AN D FIG 81 — B U M ST E D A N D C H N D A A
: TOP SE CT I ON A L :
. .

LER FA N L ER FA N EN D SE CT I ON A I
VIEW
.

.
V I EW .

AA , C ou p li n g s . li
A A , C ou p n g s ; B , adj u s ab e t l
i
b e ar n g ; 0, e n g n e b ase ; D D , i
it
d r f ; E , ar c h c o nn e c n g dr f ts -
ti i .

found impossible to get this quantity through the mine at

t h e stipulated water gauge in the fan drift i t was decided
,

to permit some air to enter the top of the upcast shaft and ,

then to measure the total volume i n the fan drift Which was ,

of ample area and length to obtain accurate measurements .

Th e air from the mine enters the fan at both sides and i s ,

delivered upwards to the atmosphere Th e chimney is .

very large in order that the air may be discharged at low

,

velocity Th e casing is a volute Th e blades of the fan

. .

are mounted on a steel disc 10% ft diameter , . .

Ex periments to find the best form of blade proved to

 B U M STEAD AN D C H ANDL E R FAN F X P ERI ME N TS 159

a
fl m Q W w «
4
.
n
d a b w o e 5
0
w v v w n m
8 o
n

v
w W ©
.
0 . 0 9 9 «
r
2 8 8 &
.

o w a
? 2 3
.

m w m n
H
a 3 o
3
Z
s
» n

m
z
m E 3
m m: N m
m
a
. w A 9
x m a
s
m o

2
4 M s d m N « 5 w O

:
o
h
E a
b R b h é
.

g
5
g
m a m o o o m
a S s a w m m m
g

m n o n «
7
fl

h e b e
u w ». w
m
m
a
w

.
o
.

8 £
o
n 8
o
w
s
a 8
w
u
o

3
n
? 3
e
» 3
u m .

s
.

“ . ” ” .

M
S E
a

m S
m
1 60 N
C E TR IF UG AL FA N S
the satisfaction of the makers that a modified 8 form is
the best With the i nner end of the blade curved forward i n
,

the direction of rotation so as t o cut into the air and , ,

gradually raise its velocity as it passes outwards .

Th e fan shaft bearings are provided with a vertical

adjustment ; they are not placed in the air drift i n the
usual way but are isolate d therefrom by a sheet steel
,
-

cover They are t hus free from all dirt and dust that
.

passes through the fan Th e results of a series of tests.

made by M Strick manager of the C ossall C olliery are

.
, ,

given in Table 3 4 .

Th e manometri c e fli ci e n cy has been ad ded the values ,

o f H being calculated by the formula

1 0000
H= ' — _ ’

h
T4 4
o

S er Fan Exp er i m ents — T his fan has been describe d ( see

p 1 07 ) and illustrated ( see p 1 08 and figs 5 5 and . .

We add some e x periments w ith a small blowing fan

.

13 —

E xternal diameter 1 9 7 in .

I nternal diameter 1 18 in .

L ength of vanes radially 3 94 I n .

Widt h of vanes parallel to the ax 1 s 3 5 5 in .

C ross section of the discharge pipe

- x 9 8 4 in .

07 sq ft . .

TA BL E 3 5 .
— S M A L L SER FAN E XP ER I M EN TS .

r
e
gn u; P5 , ;
8 g
a2 35 % g “
5 5
E8
s 3 3 E2 E 8 52
'
'
u

:
5 a 53% g g
9
.

9 2 . a
; 2
In

935 99 5
9 10 62 6 83 2 09 7 5
930 67 0 770 09 9 6
950 57 0 64 5 0 9 95
16
'

CE N TR IF UG AL F A N S

2
3
m
q
m
z
z
m
m
3
2

5
m
m
BEC K A N D H E N K E L F AN S 163

z
<
m
.

3
3
2
3
*

K
C
E
B
164 CE N TR I FUG AL F AN S

7
8

G
F
.

z
fi
m
a
m
o
p
fi
a

m
a
m
o
u
c
h
.

S
S
z

E
L
B
U
O
D

6
1 66 CE NT RIF UGA L FA N S

WW
E u
O
P f
o

5
e 5
s 2
1 1000 2 000 3000
r 1

o 6 Cq J wr mm .

3 1
H 1
1

y
c
n
i n
c a

k 2 n
e
t

B s

e
r
P

0 0
0 I I 2 000 I
I I I I
0 64 5 0-5 4 1 1 15

F IGS 90 AND 91 .
—TE S Ts OF 4
’
6
”
LE
SI N G IN LE T FAN .

M OTOR D R I V E N
.

9 0 — 1 00 r
. ev .
p er m in . 91 .
— 200 r ev .
p er m i n .
 A LL E N FA N S 167

2
£
a"
0 0
'

‘
3
3
:
:
F 0
9
1 8
0
v Q:
'
6
p.
0

In

3 7

I.
0

~
r
U

8 5
5
Vi 0 3

: 4

. S
0
a.
3
0
fl

3
n

FI G S . 92 AN D 9 3. — TESTs OF 4
' "
6 SI N G LE IN LET FA N .

M OTOR D R I V E N .

pe r m i n
92 — 3 00 r m in 93 — 4 00 r
. ev .
per . . ev . .
1 68 CE N TRI FU GAL F AN S

C
L
C
L

T
0

L
Y
D

e
r
u
s
s
e
r

FI GS TE S T S S I N GL E I N L E T FA N
’
95
”
. 94 AND .
— 0F 4 6 .

MOT OR D R I V E N .

-
5 00 r ev. pe r m in . 95 .
— 600 r ev .
pe r . m in .
170 CE N TR I F UG AL F AN S
pages I n the first place it will be noticed that th e dis
.
,

charge when the water gauge is ero is a little more than z

thirty ti m es the number of rev per min and if the equiv . .

,

ale n t orifices are calculated when the discharge in cu ft is . .

fi fteen times the rev per min the following table results
. .
,

Re v per min
. . 2 00 3 00 400 6 00 7 00 800
r min 1 2, 000
Water gauge
.

in

inches 05
Eq uivalent

cu ft. .

per sec and . H is head of air in ft . Taking

1 0000
H ==h x
144

which gives a correct value for the average densities of air

and water we obtain the equivalent orifice

~
,

_
Q
0 92 x 60 x 1 2
/ h
g 2600 J ]?

~
Q being here taken i n cu ft per min from which the . . .
,

values i n the above table are calculated This shows that .

when Q 2 2 tip speed Qas 7

/h thus it follows that the
,
00

manometric efficiency at this equivalent orifice is constant .

I t can be calculated by the formula

g H = 9 3 2 7; 4 03 00 12
41 — 2
.

z
( y N 2
‘

g 1
e
z
2 w 2 6 6

where N number of rev per min He n c e th e mano . .

metric e ffi ci e n cy at this orifice is approximately 5 0 per

cent Th e m echanical efficiency decreases with t h e speed
.
,

as the brake horse po wer of the motor that drives the fan
is partly absorbed by the work done on the air by the
wheel and by the surface friction of t h e outside of the
,
 A LL E N FAN S 17 1

wheel two quantities of work which vary as the cube of

,

the revolutions but in addition to this there is the bearing

friction which although not strictly accurate we will take
, , ,

proportional to the revolutions .

I f we assume
BHP . . . A (W W N
N
c N
’
1 06 6

B HP .

=A
( i y C’
.

v —-
0r
N
l
T Z UU
Te e n

then for 800 and 4 00 r ev . we get

20
= 0 64 A + C
-

6
and
so that A : 36 5 and
This gives the following table :
B H P actual
. . . 13 7 8 50
B H P calculated
. . . 137 0 62 5
Re v per min
. . 7 00 600 300 2 00

~
which shows a very close agreement an d justifies our ,

assumptions Th e useful work varies as the cube of the

.

revolutions for Q /h : N at a given orifice and since

oc a ,

th e useful work Q h i t also N Th e air e ffi ciency

cc ,
cc
3
.

useful work done

1
7 work transmitted to wheel

is therefore probably a constant quantity .

Wor k expen ded i n Over c om i n g D i s c Fr iction I n the fore —

going analysis we asserted that that portion of th e brake

horse power required to overcom e the surface friction
between the air an d the outside of the wheel varied as the
cube of the revolutions A proof of th is statement may be .

conveniently introduced here We will assume that the .

surface of the disc is of such a nature that a film of air

172 CE N TR I F UG AL FANS
adheres to the surface so that if a certain ar ea 3 sq ft is
,
. .

m o v m g W i th a linear speed 0 ft per sec the retarding force .

F=co sv
2
1b .
,

where c is a constant and is the density of the air i n 0

'

cu ft per lb Applying this to an elementary annulus a

. . .
,

r ad l us 7 ft and thick ness dr fig 9 8 we have

‘
. .
, ,

s = 2 7r r dr
and u = w r ,

SO that dF = 6
2
2 1M d1 w r 2
' ' ’
’
0

lb .

dF is the resistance opposing the motion of the annulus .

FI G . 98 .
— Dr sc F R I C T I ON .

Th e resisting couple 7 dF c 2 ‘
7r 0
' 2
(0 7
“
dr lh f t .
-
.
,
so that
the total couple due to the whole disc
r
2
r dF

2
02
5

Th e work done per se e . by t h e i m peller in overcoming this

resistance 7
. 5

c (o c 2 7r 0 ( o
3' 3 *

D H P 3 5
— — 00 w
'

5 50
. .

00
'
174 CE N TR I F UG AL F AN S

z
s
m
m

m
m
a
m
é
a
m

z
m
w
17 6 CE N TR I F UG AL FANS
which gives the same reduced orifice as that of the
fan multiplied by
, to allow for e x aggeration by e
anemometer Then.

9 H

Q 0 x oe 2 n
B ut 1; 2 _
2 71 '
3
2
7. x x 05 8 4

0 55 -

08 5
0 647
-

'
— 22
the angle of outflow .

Th e velocity of whirl 52 cot qb

53 7 311 : 1 6 n H .

T hat i n the volute

144 Q
0v
17 6 x

144 x 33 8 x
JQH ‘
1 49 ‘
9H '

17 6 17 l
' °

I t is clear that a chimney is necessary to reduce

which would cause a loss of head if n ot reduced of , ,

H,
2 g
so that the e fficiency would be below 30 per cent since
"
.
,

2
c 1 11
2 a2 0
2
'
x 1 50°
178 C ENTRIF U G AL FANS
Tabl e 3 6 gives the calculated and experimental values
for H for the Rateau fan of Table 2 0 Th e quantities are .

in this cas e i n the m etric system Th e last two col umns .

show h o w very closely the fan obeys the law .

Table 3 7 gives a comparison of the manometric

efficiency obtained by experiment and that from
1 O
0 136 0 69 0,
f
fm N/Um
which is equation ( 3 9 ) for fan N o V III . .

TA BL E 3 7 .
— F AN N O . V III . B RY AN D ON K I N S
'

EX P E R I M EN TS .

Eq u i l n t o ri fic
va e e A c tu a m an ol
in q ft
s . . i
m e t r c e ffi c e n c yi .

P er c en t . Per c en t .

590 590
5 45
43 0 435
33 1
24 0
190

2 87

F ans V I .
, X .
,
an i XI . have been treated i n a similar
manner in Tables 3 8, 3 9 , and 4 0 respecti vely .
C OM P ARIS O N B ET WE E N TH EO RY AN D EX P E RI M ENT 179

T A BL E 38 .
— FAN N O V I
. . :D ON KI N S ’
E XP E R I ME N TS .

Eq u i l n t o r i fic
va e e Cal c ul t d m n o
a e a Ac tu l m
a an om e tr ic
in q f t
s . . m t r i c ffi c i n c y
e e e . ffi c i
e ency .

Pe r cen t . Per cent .

600 600
57 5 58 0
’

5 40
5 00 5 00
444 430

15 0

1
0 2 v
1 666 = 0
‘
.

7 m
7

T A BL E 39 .
— FAN N O . X . :D ON KI N ’

S E X P ER I ME N TS .

Eq u i l n t o r i fic
va e e Cal cul t d m n o
a e a Ac tu l m
a a n om et ric
in q ft
s . . m t r i c ffi c i n c y
e e e . ffi c i
e e n cy .

Per ce n t . P er ce n t .

57 0 570
63 0 60 ‘

59 8 59 8
5 10 52 C °

416 4 16
33 0 33 C °

17 7 17 8
12 1 13 5

1 o
O 0 .

7
1m
18 0 C E NTRIF U G AL FANS

T A BL E 40 .
— FAN N O . XI . D ON KI N S EX P E R
’
I M EN TS .

Eq u i l n t o r i fic
va e e Ca l cul t d m n o
a e a ‘
Ac tu l m
a an o m et ri c
m t r i c ffi c i n c y
e e e . effi c i en cy .

Per ce n t . Per ce n t .

28 5
32 8 26 0
0 24 19 0 19 0
155

2 78

Th e s m al l es t eq u i val e n t o r i fi ce at wh ich a t es t was m ad e , e xc p t e

ze r o o r i fice ,
was sq . ft .

Th e equation to the above fan is

1 2 64 0 2 =0 .

Fan N o 1 , . in the same paper has an equation ,

2
e
z
00 6 4 9 12 Q 4 4 40 h Q 0,
z 2

where 02 is the peripheral velocity in ft per sec and Q is

' . .
,

the number of cu ft of air per min It being the static

. . .
,

water gauge Th e observed head cannot be e x pected to

.

agree closely with that calculated from the above equation ,

as the static water gauges were often lower than they

should have been owing to i nduction in certain cases a ,

vacuum being shown where there was undoubtedly pressure ,

for how could the air flowing in a pipe of uniform section

pass through b afli e plates and yet with unchanged velocity ,

reach a space of greater pressure a man ifest contradiction ,

of the law of the conservation of energy 7 Th e reason for

the apparent vacuum was that induction took place as ,
1 82 C E NTRIF U G AL FAN S
T A BL E 42 .
— M A N OM ETRI C E FF I C I E N C I ES OF TH E
P A R KE N D M IN E FAN .

V ol u m i n c u
e . f t pe r
.

m i n pe r sq
. . in . M an om e t r i c
of di
am e t ri
c al s e c tion , e fli c i e n cy
.

Zero 06 04
Zero 06 02
06 00
Zero 06 03
06 nearly
06 03
05 8 7
05 9 4
05 6 3
0 559
05 6 2
05 6 5
5 91 05 5 4
0 483
04 9 5
04 4 7
1 0 39 03 7 5

0 3 27
155 0 23 7
160 O 2 24
23 5 0 1 25
0 100
272 00 9 9

We now come to a formul a much harder to prove ,

viz .
,

that the air efficiency 9H

7
7 z

a2 1) 2

Th edifficulty is due to the exag gerati on of the disch arge

and possibly also of the water gauge I t is al so i n many .
 C OM P AR I S O N B ET WEE N T HEO RY AN D E X E P RI M E NT 1 83

cases difficult to determine and therefore 0 further I is 2 , ( )

often a variable quantity I n Do n k i n s paper however .

’

, ,

where the records of the discharge can be taken as being

accurate experiments are submitted i n which the agree
,

ment between the air efficiency 7 i s almost perfect except 7

at very small orifices where the formula does not hold ,

good owing to the fact that losses of energy which we

,

neglect in the above formula become of consequence At .

these small orifices the friction between the wheel disc and
the air and that of t h e air against the fan casing during its
,

passage thro ugh the wheel become of importance Th e , .

former varies as of whilst the corresponding work w asted

,

varies as 12 T h e latter is a quantity of t h e second degree

2
3
.

i n 02 and Q i e it may be expressed by the formula

—
,

Th e importance of these terms becomes manifest when we

poi nt out that when Q is z ero
9 H L H
d 0
’
0
,
—
nm on the average ,
a 2 2

whilst at z ero orifice the actual air e ffi ciency is z ero .

Fan N o 1 i n D o n k i n s experiments is a Rateau fan and

’
.
,

ten e x periments with this are given i n Table 7 There .

are twe nty wrought iron vanes which are inclined forwards
-

at 4 5 deg to the tangent to the outer circumference A lso

. .

from Table 5 for fans discharging against a small head

, ,

I) 2
Q a o x 1m at e l
2W 3
ppr y .

cot 0 Also
2
b2
a
2
v2
'
,

as I : 1 3 5 d eg assuming that the angl e of discharge coin

() .

cides with th e vane angle and that there is no coe fli ci e n t ,

of contraction at discharge from the wheel T hese .

assumptions are j ustifiable as there are twenty vanes ( a ,

large number for a diameter of 1 9 6 and the vanes are

so desig n ed that uniform outflow probably tak es place I t .
 1 84 C E NTRIF U G AL FANS
has been proved by e x periment that the inflow to the fan
is uniform Then .

9 H 3 2 2 x 1 0000 It
_
a2 v
2
14 4 0
2

where h is the dynamic wa t er ga uge . As an example in

e x periment 6 Q : 2 7 00 ,
,

62 : 6 7 2, 0 = 1 7 8,
2

gH
'
0 5 4,
d 02
a

as compared with 05 9 4 i n the experiment .

Table 4 3 gives the experimental e fficiencies and th e

corresponding values of 7 : 7

T A BL E 4 3. — FAN N O . DON KI N S E XP ER
’
I M EN TS .

xp r i m n t l
E e e a Ca l cul t da e Ai r e ffi ci e n cy
ffi c i
e cy en . ai r e ffi c i n c y
e . d i ff e r e n c e
.

1
15 3
21 3'

339 3 09
470 438
53 8
599 593
6 06
49 9 I2 7
5 00

and shows that there is a very close agree m ent between

the two except i n the last two cases this diff erence how
, ,

ever was predicted

,
.

Fan N o 3 h ad plane radial vanes so that the outflow

.
,

must have been radial and whether uniform or no is of no

,

consequence because a2 = r 2 I t will be clear to the reader

,
.

that t he out flo w must have been radial because we are ,

j ustified i n supposing the wheel fi xed and a ce n t r i f ug al

accelerative force acting on each particle I t must not be .
186 C E NTRIF U G AL FANS
F rom Table 1 3 , taking the third e x periment with each
fan , we get Table 4 7 .

I n all the above except 7 8 1 0 of Table 2 0 1 an d 9 of , , ,

T able 22 and 3 of Table 1 3 the values of n are less than

, ,

the mechanical e fficiencies although these include en g ine ,

friction I f w e multiply n by 1 5 which is the highest

.
9
,

engi ne mechanical efficiency we are j ustified in assuming ,

all experiments except 3 of 1 3 8 of 2 0 and 1 of Table 2 2 , , ,

w hich are at very small reduced orifices give values ,

of ; below the corresponding mechanical efficiencies

7 .

Th e greatest discrepancies are those in Table 1 3 where the ,

experimenters acknowledge that their discharges are too

high I t is reasonable to suppose that the discrepancies
.

are therefore mainly due to exaggerated discharge because ,

i n this case it would reduce 7 for by an increase Of 62 w e 7 ,

increase a I n Table 4 8 we have supposed the actual

2
.

discharge i n Table 20 to be T v t h s of that given by the g

anemometer so that n is increased an d the mechanical

e fficiency reduced We get the following results
,

N o w except at small orifices such as 8 and 9 which is

, , ,

Omitted as it is practically z ero

, efficiency of fan = n very ,

nearly
efficiency Of engine and f an
efficiency Of engine
T A BL E 4 5 RATEA U FAN :T Y P E A T A B LE 1
.
—
, .

No of e xp r i m nt nt M e c h an i c al e ffi c i e n cy
. e e . 77
P er Ce
pe r cen t .

32 3
338
37 1

485
5 64

607
612
 C O MP A R IS O N B E T WEE N THEO RY AND EX P E R I ME NT 187

T A BL E 46 — . R A T EA U FA N : D I A M ET ER .

No of e x p ri m n t ce n t
M e c h an i cal e ffi c i e n c y
pe r c e n t
Pe r
'
. e e .
0
,

T A BL E 47 .
— R A T EA U F A N S :B ELG I A N C O M M I SS I O N .

NO of f an ce n t
M e c h an i ca l e fli c r e n cy
. . 7,
p er .

pe r cen t .

29 8 48 3
444
545 47 0
50 5 775

Hence efficiency of en g ine

e fli ci e n cy of engine and fan
7
)

and should be about 08 5 to 09 0 We infer from the .

above that the exaggeration of the discharge is greater

than what we have assumed v i z the ratio but not ,
.

very much and this agrees fairly well with the results Of
,

the P russian C ommission if we allow for the additional

exaggeration due to the variable velocity Of the air which
is always found in m i e s n .
 18 8 C E NTRIF U G AL FAN S

T A BL E 48 .
— R ATE A U F AN :
TY PE A, T A BL E 1 .

M e c h an i c al
igi fi
r
per ce n t e fli ci e r i
q
e xp n nt .
1; .
y pe r
c er l

1 29 7 08 7
2 35 8 33 3
3 392 38 7
4 4 14 09 4
5 509 476 09 4
6 58 9 540 09 2
7 649 549
8 62 0 38 7 06 2
10 63 5 53 1 08 4

Le t us ne x t consider H eenan and Gilbert s experiment ’

with a fan 1 7 in diameter and 8 i n wide . Th e e fli ci e ncy . .

h ere given is the ratio of the useful work done by the fan
to the work done on a shaft which drives the fan by a belt ,

so that it should not be m uch less than the efficiency of the

fan alone or n at orifices of moderate si e if we can assume
,
z ,

the trut h of the statement that the work done per 1b by .

the wheel is a, v2 g I n fig 2 3 we have the results Of . .

e x periments with a blade terminating at 3 5 deg to the cir .

c um f e r e n ce so that if the angle of relative out flow were

,

the same as the vane an gle we should have deg .

C onsider the e x periment i n which the discharge was

cu ft per min and the tip speed
. . ft per min
.
,
. .

Th e total water gauge is 9 6 i n so that we have .

,

:
1 2 000
=
2 000 Q
v
2
—
200’ Q ’
an d bz =
60 60 2 1r r
2
3
2

2 000 x 1 4 4
— 112 2 ft per sec
17
. .

60 X 7, x X 8
190 C E NTRIF U G AL FANS
I f w e co n sider the d ischarge of cu ft per min
. . .

3 5 00 3 5 00 x 1 4 4
_
Q 50
°
b2 “

50 x 71

as before and the dynamic , 01

°
total e ffi ciency ,

from fig 7 also
.

1 0000 32 2 x 7 6 000
"
f“ 80 that “ : 1 2 1,
14 4 2
2 00 x 1 4 4 x 07 -

2 00 — 1 2 1 79
C0 t
_
19 7 12 7
14 deg .

an d even if we suppose
coe fficient Of contraction of 0 8 a ,

9b 1 7 deg 1 8 min . .

Fi g 2 5 shows similar curves for blade N o 4 fig 2 2

. .
, .
,

which has double curvature and terminat e s radially at t h e ,

outer circumferen ce so that if the a n gle of flow was t h e

,

same as the angle Of vanes I would be 9 0 d e g Th e ,

( ) .

dy amic efficiency when the flow is

n , cu ft per min . . .
,

is 8 3 per cent and the dynamic gauge is 1 2 8 in

.
,
.

2 5 00 x 1 4 4
= 1 4 05 °
17 x 8

1 28 000 32 2 x 89 0
H ‘ = 890° “
2
:
144 2 00 x 0 83

2 00 — 172 7 -
27 3 -

7 1
C
‘’
fl
)
1 1 05 -
14 0 5

deg 1 4 min 2 7 . .
,

which w e consider extremely improbable as there are six ,

van es I n our Opinion the discharge is very much e agg er

. x

ated Th e manometric efficiency i n this case is

.

g H 32 2
89 0
°

x
nm
_

T
2
2 00 x 2 00
_ P
71 6 per cent .
,

wh ich is probably very much nearer the true value Of t he

 C OM PARIS O N B ETW E EN TH EO RY AN D E X E P RI M E NT 191

mechanical effi ciency . At the same discharge the compres

sion or sta tic gauge is 11 i n and this gives
.
,

322 x 1 1 0000
0 —
2 2 00 x 1 4 4 x

since the mechanical efficiency is 70 per cent .

f"
2 00 242
“0 ‘ 1
I
7
.

14 0 5 1 4 05
-

f
( ) 30 deg . 8 min .
,

while the manometric efficiency is

32 2 x 1 1 0000
7 111 :
7 200 2
x 14 4

Th ed scharge Of
i
cu ft per mi n gives the follow . . .

ing results : a = 152 7

2
cot qS= 25 °

,
°

the dynamic e fficiency being 08 5 and the dynamic water

g uge 1 1 6 i n while the manometric effi ciency is 6 4 7 per
a

cent and the mechanical efficiency 5 0 per cent We d o

.
,

.
,

not think that the mechanical efficiency is as low as the

manometric efficiency because it is quite possible that 1
,
< )

may be less than 90 d e g and w e believe that it is I n ,

.

support of this statement we mention some e x periments

with a Far co t centrifugal pump at Kh at at b e h Egypt 9
, ,
1

which w ere made with very great care and are pro b ably ,

accurate I n these t h e mechanical e fli ci e cy of engine and

. n

pumps w as 6 5 per cent corresponding to a probable .

,

efficiency of pump alone of between 7 2 2 and 7 6 5 per cent .

,

while the manometric efficiency w as 6 5 9 per cent

We shall n o w discuss s me experiments 20 w ith an open
o

running fan at t h e Seghill C olliery Before doing so w e .

may state that w e believe the e fficiency is exaggerated .

N o t knowing the section of the fan drift w e cannot find ,

the co r rect reduction of the water gauge due to the velocity

of the air therein ; but neg lecting this let us consider the
, ,

to tal energy of air rejected from t he outer circumference of

the fan Se e Table 4 9
. .
192 C E NTRI F U G AL FANS

T A BL E 49 .
— OP EN R UN N I N G FAN :S EG H I LL C OL L I ER Y .

A v r
e ag e o f e xp r i m n t
e e s . 2 to 9 . 10 to 1 7 .

Re v s of fan per
min . .

D ischarge i n cu
ft per mm 1 35 7 00
Water gauge in drift
. . .

2 46 °

H P i n the air
. .

I HP
. . . 96 1 153 7
Mechanical effi ciency per cent . 54 32
'

Se p ar at i o d oo r o pn s e n .

Taking the average Of experiments 10 to 1 7,

2 r : ft 3 :
2 , 2
.

N: rev per min so that 02 : 1 1 1 ft per . .

, .

the water gauge in the drift

;
1 3 5 00
the d i scharge Q '

6
2 2 62 cu ft per sec
. . .
,

22 6 5
neg1cct mg con t rac t I on
, ,

62 .

n X
,
x 1 2 92
,

g H
az =
'
,
where 7
7
:
0 7
7
2

and a will be least i f w e give ) its largest possible value

2
7 .

A ssume that the mechanical efficiency of the engine was

only 8 5 per cent and we get .
,

32 2 x 24 600 x 8 5
“2 ‘
7 7,
14 4 x 111 x

and assu m ing the density of the air as 00 7 6 1 , the energy

rej ected at outflow in H P is . .

b2 + a2 2 2
7 7 + 15 9 2 - 2
2 2 65 x 0 07 6 1
-

Q X x X 30
5 50
.

2g 550 64 4

Adding to this the impeller H P 5 2 6 we account for . .

,

82 6 H P Of the 9 6 1 . giving a. total dynam ic

194 C ENTRIF U G AL FAN S
Adding to this useful HP
. . we find the dynamic
efficiency is
14 45
9 4 1 pe r cent .
,

which is of course impossible I f we assume that the

, ,
.

mechanical efficiency of the engine alone is 90 per cent .

,

the dyna m ic efficiency becomes

14 4 5
97 per cent .

Th eonly conclusion is that the anemometer very greatly

exaggerates the discharge Of air .

C HAP TER I X .

H I G H P R E SS U R E
-
F A 1\ S .

TH IS chapter is a summary Of a paper by P rof A Rateau 1 . . .

2

H itherto fans have not been required to give a water

gauge Of more than 2 4 i n while it is only lately that
.
,

centrif ugal pumps have been used for heads over 5 0 ft By .

means Of steam turbines and a single pump a head of nearly

ft has been Ob t ained These turbines pumps and
. .
, ,

fans are the design of P rof R ateau and those mentioned

.
,

herei n were constructed by Sautter Har lé of P aris -

, .

I n the theory of centrifugal pumps and fans there are

four quantities Of importance :
Th e mechanical effi ciency
 H I G H P R ESS U R E FAN S
- 195

Th e coefficient of po wer transmitted to the shaft of the

pump
Wg 7 m
7
"
'
a
7
3
0
'

2
0
n

Th e above quantities are numbers independent of all

u nits I n the formul ae Q is the volume i n cu ft or cu
.
. . .

metres per sec H is the head in metres or f t i s the weight

.
,
.
,
o

i n lb or kilogrammes Of 1 cu ft or metre of the fluid .

pumped T2 is the external radius Of the wheel W is the

. .

, , 3

work done per sec i n f t lb or k i log r am m e t r e s a d 2 is

. .
-
.
,
n 12

the velocity Of the wheel at the outer radius P rof Rateau . .

FI G . 1 02 .
— C UR V ES .

used volumetric efficiencies as absciss ae instead of orifices ,

and drew three curves of ym 7 and as i n figs 102 and a

,
7
,
r , .

1 03 . Fi g 1 02 are those of a centrifugal pump and 103

.
,

of a Rateau fan For a given pump at a fi x ed number of

rev per min the curves and nm are those Of W and H to
.

. .
,
7 ,

u i t b le scales
a .

Figs 1 04 and 105 show a centrifugal fan drive n by a

.

ste am turbine and i n tended to produce a considerable

pressure Th e fan is m ade of steel of very good quality
.
,

capable of running at a peripheral velocity of over 800 ft .

per sec it turns in a cast iron casi n g h avi ng two openings

.
-

for suction and forming a di ff user and volute Th e turbine .

,

which is a steam P elton wheel is 1 18 in in dia while the fan , . .

,
19 6 C E NT R IF U GAL FANS
is 1 0 in the method of raising the Oi l from a lower to a
.

higher reservoir is Sho w n i n fig 1 06 in which the pipe A is .

,

connected to the lower reservoir and B to the higher Th e .

small tube M brings a small amount of air from the fan ,

which mixing in small bubbles with the column of Oi l B

, ,

lowers its specific gravity to an extent sufficient to enable

the column A to raise it to the higher reservoir Th e dis .

charge or pressure of the fan can be controlled by a

pneumatic governor I t consists of a cylinder D fig 1 04
.
,
.
,

containing a piston P the details of which are shown in

,

. l il°

F IG . 1 03 .
—
C UR V ES .

a ; 0 = nm .

fig 1 06 Th e rod Of this piston is connected by a short r od

. .

to the point C of the lever A B C which oscillates about the ,

point B and i s connected to the rod of the steam throttle

,

valve at A Th e cylinder D has its t wo ends connected by

.

the pipes 1 2 figs 104 and 1 05 with a straight tube ( 1)

, ,
.
,

and P itot tube both placed in the discharge pipe Of the

f an see also 3 and 4 fig 107 ,Thus the diff erence of the
. .

pressures on the two sides of the piston P is proportional to

the square Of the velocity Of discharge and exerts an upward ,

pressure which is balanced by the weight of the piston ,

assisted by an additional weight if necessary and a s p I ng S r ,

198 C E NTRIF U G AL FANS
friction a rubber diaphragm is sometimes used Th e ex .

1 m e t s were made at S autter H arl e s P aris works and

’
e
p r n -
,

the measurements by M R ateau and M C hatelain T h e

. . .

discharge w as measured by a convergent discharge pipe two ,

of whose faces were fixed and paral lel while the other two
,

were movable A t its larger end this was fi xed to the dis
.
 H I G H P R E SS U R E FANS
-
199

charge pipe of the fan 3 a mercury manometer was used to

measure the pressure in this convergent mouthpiece and the ,

discharge could be calculated from this T h e Speed w as .

changed from to rev per min Th e peripheral . .

speed Of the fan reached nearly 8 7 0 ft per sec while the . .

,

discharge pressure amounted to 1 6 7 5 in of mercury or 2 2 8 .

,

of water more than half an atmosphere Th e revolutions

—
.

were measured by a counter driven by worm gear which ,

reduced the speed one thirtieth On e two or three of the

-
.
, ,

steam turbine no zles were opened to give the power r e

z

quired by the fan and the steam pressure in the steam

,

chest was noted as by this means t he power Of t h e t urbine

,

FIG . 1 07 .

could be deduced fro m previous experiments so that the ,

total efficiency of engine and fan w as thus obtained .

By formul ae obtained by M Rateau the consumption Of

22
.

steam and consequently the power theoretically available

, ,

could be calculated Th e governor was not used during

.

these experimen t s so that the pressure i n the steam ches t

,

w as not a fle ct e d by it Th e discharge Q th e useful power

.
,

i n the air the theoretical po wer of the steam Tt and the ,

coefficients n and were calculated i n the following

man ner :the discharge Q i n cu ft per sec estimated at
,

. .
, .
,

atmospheric pressure is equal to the product of S the section

, ,

of the convergent discharge pipe i n sq f t and the velocity of . .

,

flow i n ft per sec Hirn s experiments prove that the

. .
’
2 00 C E NT R IF U GAL FANS
of discharge of a convergent pipe does not diff er
co e fli ci e n t
more than 1 or 2 per cent from unity M Rateau proves
. . .

in his paper that without serious error

2
cd

2 g 0
°

where A p is the diff erence of pressure per sq ft between . .

suction an d discharge and is the mean density of the

,
a

air Th e figures in the seventh column of Table 5 0 are

.

Obtained in this manner Th e si x th column gives H in ft

. .

of water and in the calculation of the temperature must

,
0
'

be assumed to be 3 8 deg C e n Th e reduction of volume

. .

during compression must be taken i nto account in calcu

lating the useful power T To Obtain the figures in the
u.

ninth column the variation Of pressure per sq ft must be . .

multiplied by the volume in cu ft at the mean pressure ,

"
. .

so that
32 8
Ai Q 3 2 8 +O5 H ° '

_
T
550

Th e work theoretically obtainable fro m the steam T, is cal

culat e d asfollows Th e discharge Of steam is Obtained from
the formula
I =s P log P ) ,

where s is the total section of the n o z les i n sq centi z .

metres and P is the pressure i n kilogrammes per sq centi

, .

m etre while I is the quantity of steam discharged per sec

,
.

in grammes if this is multiplied by it gives kilo

grammes per hour I f K i s the number of kilogrammes per
.

horse power hour when the steam chest pressure is P and

,

the e x haust p i n this case the atmospheric pressure then

, ,

log P
K = O85 + .

log P log p —

T hen the theoretical horse power obtainable from the

steam is
I
Tt
K
2 02 C E N TR IF UG AL F ANS
T he mechanical efficiency
—
n
Tt a

and is given i n the last column of T ab l e 5 0 on p 201 A s . .

the density of the air varies at diff erent speeds owing to ,

th e great compression the manometric efficiency is not

,

constant for a given opening of the convergent discharge

pipe A lthough the fan is only 1 0 in dia it develop
. . .

0t
'

005
-
015 -

FI G . 1 08 .

a maximum of 4 5 6 5 H P and a maximum pressure of

. .
,

1 9 ft of water
. F ig 1 08 gives characteristic curves a
. .
,

separate characteristic being gi v en for each speed as the

efficiency of the turbine increases with the speed T he
while the abscissa:are values
.

ordinates are values of

of I t m us t be r e m em ber ed that the efli ei en cy i s the r ati o of
the use ul f wor k done b y the f an to the id eal a m oun t ob tai n a ble
f r om the normal
s team i n th e R anki n e or Clausi ri s c cle
y At -
.

speed the efficiency passes 2 8 per cent showing t h t the .

, a

fan alone has an efficiency of 5 6 per cent as that of the .

,
 HI G H P R E SS U R E F AN S
- 2 03

turb ine is ab out 5 0 per cent according to exp eriments .

,

previously made .

H igh pressure fans driven b y steam turb ines can b e used

-

for cupolas blast furnaces and Bessemer converters and

, , ,

wherever a water gauge of more than 3 6 in is required-

. .

T hey can even be employ ed to compress air to 7 0 lb p e r .

sq in or over A sin g le wheel can increase the pressure

. . .

in the ratio of to 1 so that two wheels working in ,

series would give a pressure of a third 3 4 and a ,

fourth 5 atmospheres T heir mechanical efficiency i s .

slightly inferior to that of ordinary piston compressors ,

b ut superior to R oots blowin g machines whose efficiencies

’

are not more than 3 5 to 4 0 per cent F or the supply of .

air to a b last furnace whose capacity is 1 60 tons of cast

iron per day and req uiring ,
cu ft of air per min . . .

at atmospheric pressure and compressed to half an atmo ,

sphere the fan would be 2 f t 7 ; i n dia would run at

,
. 7 . .
,

rev per min and the steam t urb ine would b e ab out
. .
,

the same si z e T he effi ciency of the turbine fan for this

.

hig h power 5 00 useful H P would reach 1 0 per cent

— . .
— .
,

corresponding to a steam consumption of 4 9 8 lb of steam .

per useful H P hour if the turbi ne worked with con

. .

densation .

C H A P TE R X .

The The or y f
o T h e propeller
P r op eller Ven ti lat ing Fan s .
-

is the simplest form of fan ; it requires n either diff user nor

volute al t hough frequently provided with a chimney
, .

P ropeller fans are used whe n a large volume of air is

required at a very low pressure as for example in the , , ,

ventilation of buildings i n which case the volumetri c ,

efficiency becomes far more impor t ant than the mano

metric o r mechanica l efficiency I ts com plete theory .
,

however is ext remely complicated m ainly b ecause each

, ,

particle of air does not keep t o a cylinder concentric wi t h

the axis of the fan I n the following approximate theory .
2 04 C E N T R IF UG AL F ANS
we shall suppose that each particle of air moves on a
cylindrical surface and that the axial component of inflow
,

is the same as tha t of outflow I n fig 1 09 is shown the . .

section of a b lade A B 0 is t he relative angle of inflow

, , ,

and d that of outflow if we assume that the angle of flow

) ,

coincides with the a n gl e of vane L et b be the axial com .

ponent of the velocities of i nflow and outflow the peri ,

v

h e r al velocity at a radius r 12 that at th e extreme radius

p 2 ,

r a the component of the absolute velocity of the air at

2
,

outflow at radius r perpendicular to b oth radius and axis

, ,

or i n other words the velocity of the whirl T he motion

, , .

FI G
. 1 09 .
-
V E L OC I T Y D I A GR A M S F OR P R OP E L L E R FA N .

of the b lade is to the left . T hen from the triangle of

velocities at outflo w
a :v b cot I
( )

d v
also — work done by the b lade at that
J(

radius per pound of air and if the air enters the wheel ,

without sudden change of direction it follows from the , ,

triangle of velocities at inflow that ,

7)
( 30 b
2 06 C E N T R I F U GAL FAN S
is proportional to a and as t h e motion is a the total work
2
, ,

wasted is proportional to a and may be represented b y s

,

4 ’
29

which is i ndependent of the quanti ty passing through the

fan ; hence the loss of energy per pound or the loss of ,

head is represented b y
,

F3 ?
a
2g b

b ecause b is proportional to the w eight passing through

the fan .

L et us now s u ppose that the axial velocity is the same

at any radius ; then
2
Q b 122 # ,

and the total loss of head at inflow is

2
b 6>
.

2
( v -
COt
2 1r ¢ b on d f
LI
w oe —
m o

where r 1 is the internal radius I t must b e remem b ered .

that 9 is a function of r and not necessari ly a constant ,

.

T he loss of head at outflo w i s

(v M OW) 2
m
-

2m “ 2 7r r bo dr
b
.
2
29
’
’
l
L 2“ "
b r! 2 g b rr

where qt is a function of r , and v = r where w is the to

angular velocity in radians . I f there is a chimney the loss ,

of head at outflow becomes

W 2
7

fi
. 2
1 0w
2
2
'

bo dr
b
2
r
4
2 7r To bd7 ‘

b 71 '
0
 TH E T HE O R Y O F P R OP E LL E R V E N T I LA T I N G F ANS 2 07

T he total loss of head friction the fan b ecomes

'

by n

L3
5 11 '

and the total loss of head i n the chimney is

(v b cot i ) c>
3
2 1r b d7
b a dr
‘

1r r
'
r o
b2
L, = F2 + F3 2
2g b rr ( r
2
2
b 71

L et us first suppose that t he b lade of th e fan is a plane

surface and that
,
0: cl) constant . T hen L I b ecomes

1
LI +b
2
cot 2
4
)
3
4
b w co t qb
?
2 M
29 2 .

72 ]

T o simplify this put r 1 O , .

Ll ~ 2
w
+b
2
co t
2
i
< >
2g

I
§ vz b c o t
2
+b
z
cot <
b —
29

T he loss at outflow if no chimney is used is , ,

2
1 2 b
L ,a +b
2
e ot < >l -
v2 b co t +—
29 2g

b ut if a chimney is used it becomes ,

L 21,
To

gg R
3

Z
+b z
co t
2
¢
b2
29 R "
and the loss of head by friction i n the fan b ecomes
2 08 C E N T R I F UG AL FANS
T he work done by the fan per lb is .

( r
3 w2 — br 2 w cot qb ) dr
2 2
b 71
'

( r
2
r
1 ) 0
'

I ntegrating and putting r 1 : 0 the expression b ecomes ,

b wr
z _
l v 2 —
2 § v2 b cot qb
co
3 9
H e n ce since t h e work done b y the fan per lb of air
, .

delivered is equal to t h e head together with the work ,

a b sorb ed b y the losses of head it follows in the case of a , ,

fan with no chimney that ,

2 9 H v2= vz b 2
cot q (v 2 +
S 2
2 b 2
cot 1 g v b cot qt ) — 2
( ) —
.
2
.

bZ F1 b cosec 1 F b -
2 2
( ) —
2
2
,

neglecting the loss b y friction on the casi n g due to the ,

whirl speed C ollecting and rewriting the e q uation

.
,

b ecomes

2 g H § oz b cot 1
= b 2
cot q + cosec
l F 2 ( ) —
2
>
l

Walk r s Exp er i m en ta A s this type of fan usually dis

e
’
—

charges a g ainst very little pressure let us first suppose ,

H : 0 and find suitable values of P I F2 from experiments , ,

made b y Walker with a propeller fan whose b lades

,
23

were plane an d set at various angles When the blades

,
.

were set at 40 deg to a plane perpendicular to the axis .

,

the volumetric efficiency was 6 9 7 and when at 2 5 deg it , .

was 5 3 8 per cent S ubstituting these values in the above

.

equation putting
Q b “
2
b 7’

0 27 2
,
12 7
g ’
0
0 69 7 ,
2 2 2
1r
and —- — we g et
2

for
and for
so that F1 = 1 00 and F '

2
= O 9 0, '
2 10 C E N T R IF UG AL FANS
friction caused b y the whirling motion of the air b ecomes
o f importance and as w e have neglected it theory cannot
, , ,

b e expected to agree with practice .

T he only efficiency that we can consider i n this

i s the dynamic e fli ci e ncy so that ,

i-
o
Z -

g oz b co t gt

4
g
0
cot qt
2

which can be transformed to

2
m
w
y
s f ot o 40

w here is the volumetric effi ciency »

T he results ob tained from this formula diff er entirely
from those given b y experiment because the angle of flow ,

does not coincide with the vane an gle T he ab ove equation .

gives us
2
m
1 2
7 ”
7d
cot I
( ) :
7 ”
4 7
3 "
71

T he following gives the calculated values of I and () ,

shows that the direction of flow does not follo w the curved
b ack of the vanes si nce the an g le of flow relative to the
,

wheel di ff ers considera b ly from the vane angle

V ane angle 15
°
30
°
45
°
50
°

° ’ ° ’ °
I
°
1 0 30
’ ’
( ) 18 18 22 49 23 45
E xperimenta1 e fli ci
33 41 6 26 4 20 9
ency per ce n t .

I n the experiments with fan N o 1 6 with plane blades .

, ,

the mechanical effi ciency of the fan alone is n ot given b ut ,

 WAL K E R S EX P E R I M E N TS 211
’

it can b e calculated b y deducting horse power from

that of the motor which gives the S haft or b rake horse ,

power and the horse power i n the air divided b y this

,

gives us the e fficiency of the fan alone A pplying the .

above formula for the mean relative angle of outflow

we ob tain
A n gl e of va n e 1 50 20° 25 ° 27
°
3 00 3 5° 40°

E xp e r imn t l ffi
e a e
l n y p nt 3 08 461 46 0
'

c a c er ce .

o f f l n
an a o e

Vo l u m
e n yp
c
e tr
er
ic
ce
e
nt
ffi c i
.
2s 6 43 o 53 s 58 0
0

e4 7 es 4
o

09 7

C o t 4) 8 01

1 9 — 35
' '
7 ° -
7
'
10 ° -
46
'
13
° —
5 CV 1 5
°—
3 17
°-
3
’
18 — 44
° ° ’

I t will b e seen from the ab ove that 4 is approximately )

half the vane angle and also that the volumetric efficiency ,

of fan N o 1 7 with rounded back vanes i s li t t le better

.
, , ,

than that of fan N o 1 5 with plan e b lades for angles of .

, ,

30 deg 3 5 d e g and 4 0 deg

.
, , .

Considering next fans with curved vanes i n which 0 and

1 are constant b u t unequal we fi n d that in the equation
< )
,

for volumetric efficiency it is best to take 6 as the an g le of

the face and 4 as the mean b etween the angles of the face
,
)

and b ack T his is probably due to the fact that the dirce
.
;

tion of outflow depends on the change i direction which n

takes place b etween inflow and outflow the latter is deter

mined by both the back and front angles of the vanes ;
fu r ther at inflow a sudden change takes place from an
,

axial to a forward motion and this is produced by the

front of the blade T here are three fans in Walker s
,
’

paper wit h curved b lades N os 9 1 4 and 15 T he first ,

.
, , .

and third have curved faces and b acks while the second ,

h as a plane face b ut a curved b ack I f is the mean ,

. a .

angl e of i nclination of a b l ade ,

i n the 0= a — 20
;

in the second d > = a + 2 0, 6 = u

;

i n the third 6 =a — 20 .
2 12 C E N T R I FUG AL F ANS
N o w, in this type of fan we have the equation
2 g H= v
2
2 -
o
z
b co t qb + b z
co t 2 6— g
-
v b co t
2

:
2
v
—
s b 2 cos e c2 ¢

2gH= cot 6 — b2 ( cot2 6 + cot2

e
z
b 1
< ) 1 F2 F1 cosec 2

When H 0 this b ecomes ,

$2 cot0
( cot 6 cot 2 6 =b 2 c osec2 b) ( ,

and the volumetric efficiency

4 7T 6
( cot 6 cor.2 ( ,b F, F1 008 6 02
2
3
’
v
2
1 40
b ut the above w ill not give results agreeing with practice
for a constant value of F1 b ecause the vane b eing curved , ,

t h e relative velocity of the air over its surface is variab le .

H ence for the last term in the denominator we su b stitute

,

F1 ( cose c 1 cosec 2
and obtain very g ood results
<) 2
.

T hese three fans ( 9 1 4 1 6 ) were tested wi t h mean an g les

, ,

of 1 7 d e g 2 7 d e g and 4 0 d eg and we think that the

, , .
,

results obtained with the two first angles are of very little
use except to Show t hat t he last is better b o t h for ,

mechanical and volumetric e ff iciencies because the air at ,

i nflow was struck by the b ack of the blade ; in fact wi t h ,

a decrease in the angle of outflow the angl e of inflow is ,

decreased too rapidly b y turnin g round the blade Caleu .

lating the volumetric effi ciency from the equation

b '
71 4 cot 6 7r

1 ( cosec
9 co t 2 I F2 6 cosec qt )
9

3 l00 t 2 ‘ 1 1 1
2
( ) 4 2 2

and writing F1 = 1 OO, F2 = 0 9 0,

'
as i n the first case'

b 7r 4 cot 6 7r

3 {( cosec 2 6 + cosec t ) 0 1} 2 ( x

cot 6
2 14 C E N T R IF UG AL FANS
7
7 for 1
( ) per 7 6 12 0 14 8 14 0
ce n t .

f or < ) =
1 15
°
49 1 3 8 4 0 25 52 8
Mg H
r

/g H
2
or I
( ) : 30 0
5 36 45 4
N

which shows that without a chimney these fans cannot b e

used efficiently to produce pressure I t will also b e
, ,
.

noticed that the efficiency decreases as 1 increases < )

We may next consider the case of the ab ove types of

.

propeller with a chimney whose outlet has a diameter

three times that of the fan T he general equation of the .

fan is the n
2g H =r 2
§ b 6

:
— o co t
2 2

, 2
2 —
R2 2 R

b 2 F2 F 1 b2 ( cosec 2
I
( ) + cosec 2

which merely states that the head produced is equal to the

work per pou nd done by the wheel less the losses at ,

inflow to the fan outflow fro m the chimney and those , ,

due to friction T his b ecomes when Simplified

:
.
, ,

,
2 , 2
2 g H= —
1 — —
2 g co t 6 m

; co t c t 1
1

T
2 7
‘
F1
b 2
cot 2 2 2 —

2 ( co se c
z
1
< 2 + co s e c 2

an d when r - —
z R=§ this gives us
2 g [ cot 6 -
co t r >
l]

6 cot 2
Ef
l
F2 ~( cosec
F1
— 2
6 cosec2
2
 WAL K E R S EX P E R I M E N TS 2 15
’

F irst consider the cas e of plane b lades when H = 0 ; we

then get putting F1 : 0 90 and F2
,

e z+ 71 b cot 1
4
-
< )
z 7 2 2
b2 9 cot2 I 09 1 2 cosec2 35]
1 0
-
O ( ) <

:
3 z
e z
+ 0 1 4 8 1 2
2
b cot 1 b 2
[ 2 1 1 c 0 t 2 0
< ) —
;
'

~
so that the formula for the volumetric efficiency is
b rr 6 71 '

v
2
/ cot2 I
( ) cot 4)
which gives
f
c) : 30 deg . 45 deg . 60 deg . 90 deg .

pe r cent . 78 1 10 135 15 1
2
T he dynamic e fli ci e n c
y 7
7d 7 3
2 2 cot
#
"
71
p
(

and shall here suppose l to b e the vane angle

we c) .

: 3 0 deg 45 deg 60 deg 9 0 deg

I ( ) . . . .

per cent = 1 45 . 23 0 27 1 23 1

We should prob ably get in practice a very m uch h igher

efficiency than this as t would actually b e ab out half the
,
c

ab ove val ues ; in the third case for a vane angle of 60 deg ,
.
,

supposing t he angle of o utflo w is actually 4 5 deg we .

,

should have an e fli ci e ncy of 4 3 3 per cent

When H is not z ero w e get
.

12
2
b cot i b 2
C ot
z ( ) —

from which
b cot
/ H
5 4 5 '

= g
—
z
H
—
0

N 4 7
— 8 eot 2 0 C

g g ”
, Z

T his gi ves the peripheral speed for various values of

b + J g H and the corresponding values of air effi ciency
,

can be found from the equation

7
l
 2 16 C E N T R IF UG AL FANS
T he velocities and efficiencies are given in T ab le 51 .

TA BL E 5 1 .

f or
b

M7 11

01 2 13
2 15
03 22 0
04 22 9
05 28 8 2 40

77 f or
b
‘

J Zfi i

01 05 2 3 0 4 90
-
04 7 0 04 5 7
-

02 05 08 0 485 0 4 67
- -

03 03 9 8 04 8 8 04 7 6 04 62
03 08 04 5 1 04 5 9 04 4 1
05 0 239
-
0 4 03
-
0 429
-
04 1 5

which shows that the b est angle of inclination of the

vanes if w e consider volumetric efficiency as well as
,

mechanical is about 4 5 d e g but that this type of fan is

, ,

very unsuita b le as a means of produci g pressure because n ,

of its low manometric and mechanical efficiencies .

N ext co n sider the case of vanes i n which I is constant

,
( ) ,

b ut 6 varies so as to do away with the loss of shoc k at

,

inflow ; in this case the g eneral equation becomes

2
2 g H = o2 2 —
§ o
2
b co t

b
2
r
2
4

F. a r b 2
a c o s ec
z
R 4 .
2 18 C E N T R IF UG AL F ANS

TA B L E 5 3 .

n for
b
J ET? o o o o

¢ z 45 .
¢ 2 60 .

0 7 00 0 7 08 0 7 17 0 7 22
O GBO 06 60 0 67 5 0 7 01
'

03 06 5 7 06 1 2 06 3 9 0 6 69 0 67 8
04 15 0 5 00 0 544 05 85 06 89

N o w,the volumetric efficiency is b 7r ——

Z vz for these fans ,

and in a R ateau fan we have

volumetric efficiency T
2
2

since 2wr 3
z 2
2
7
2
2
, for fans of type A .

H ence even if we assume for the Rateau f an a mano

,

metric e ffi cw:_nc y of 9 0 per cent the volumetric efficiencies

L
.
,

for 6 : J g H of the two fans are 1 02 and per cent .

,

so that the propeller fan is the superior of the two in this

respect b ut of course much inferior i n mechanical and
, , ,

manometric efficiencies T he ab ove assumes f = 1 3 5 deg

. < > .

i n bot h cases and e ven if f = 9 0 for the propeller the

,
<> ,

volumetric efficiency is per cent N o propeller fan .

with d = 1 3 5 has yet bee n constructed as far as we know

> ,

and it is quite possible that such a fan might discharge the

air in t h e wrong direction .
 HE LI C AL P R OP E LL E RS 2 19

C H A P T ER XI .

H elica l P r og wller a — N ext

consider the case of helical
blades in which however the pitch at i nflow is not the
, , ,

same as at outflow b ut that 6 at every radius is so arranged

,

that inflow takes place without shock and therefore L I is ,

z ero L et P b e the pitch at outflow then

.
,

0 77 7 '

cot P ;

2 -
11
'
7 2 7r
’
i
’

r
so that wl) cot
z
and . . cot I
( L —
72
cot A r .

fig
,
co (

F or a fan with a chimney the loss at outflow is ,

2 — 2
( 7
2
7
1 ) 2 9 R4

( lr 4
7
2

9R
2 4
2 0R

2 2
A b) T
} b T
e
‘
2 20 C E N T R IF U GAL F ANS
T he loss b y friction b etween t he air and the surface o f
t he vanes is ,

F1 6 2
( cosec
2
g> l + c o se c 6) 2
2 « Mo d!

2
29 ( 7
2

dr
F1 b 2

T he loss of head by friction in the chimney is ,

M
r 3
Z
Z WT b U d T Z Wb T O d T
'

b2 T
I
L 4 = F2 2
2g b (7 2
‘

1r

f
4
dr
62
( w — A b) 3

F2 _
+ F3
§g gb

52
(w A b) 3
222 C E N T R IF UG AL FANS
.

0
P utting F1 = 1 00 °

and F2 = O 9O
' 2
: 3
$ the equatio n
,
R
b ecomes

) (0
: 2 2
2 gH 22
2 + 712
v
2
6 cot [ 08 (1 +
89

[ cot
2
( + )
2 2
b 1 m
and if m =§

2 g H = O7 7 2 '
0
2
2 —
0 9 88 cot 22
2
6
(O 34 cot 2
¢2 + 1 9 1 ) b 2

so that
0 ?
'
2
cot 12
2
b
cot ( 04 4
2
1
< 2 24 75) 6 2
9 H 0 .

:~
F urther the air efficiency
,

H g H
1
] w2 _ 1
( b w)
2
T his we can now determine with an assumed value of »
l,
c

for various values of 6 + /g H 3 since for $2 4 5 °

122 6 9 H

and for 5
92 : 90
°

b
2 4
J9 H Jg H
from which t he values of n are obtained as follows

— °

H v2

02 06 4 0

05 04 9 7
 H E L I CAL P R OP E LL E RS 2 23

T he ab ove shows that th e e fli e i e n cy decreases as th e dis

charge increases and increases with the angle
,
I t 18
prob a b ly even g reater for ¢ = 1 3 5 b ut as the air might b e 2 ,

discharged in the wron g direction with this arrangement ,

w e have not considered it .

T he volumetric e fli ci e n cy
Q 6 02 7
2
1 )
« -

2 2
T
2

S upposing H : 0 ,
then this b ecomes

which o b viously increases with pa and is 1 1 3 per cent for ( ,

.

I3
< >= 4 5 d e g and ,per cent for ¢2 9O deg T he
= . .

dynamic efficiency
52

7
1d :

2
7 12
2

( 1 +m
2
) (1 1r
2
l — é cot ( I) ,
v
2

$ cot
1 3
2
« 1
If

so that for l = 24 8 < >

2 ,

and for ia 364 < > .

H ence w e may conclude that if an effi cient chimney is ,

u sed 1 should b e 90 d e g
< ) and the inflow ed e of the b lade
2, g ,

should b e helical and of such form that if 02 is the value

,

of 9 when r = 7 2,
v
cot 02 2

sothat inflow may take place witho ut shock and a fair ,

mechanica l and a high volumetric e fli c i e n cy can b e ob tai ned .

 2 24 C E N T R IF UG AL F ANS
E xp er i m ents wi th P r op elle r Ventilati ng Fans —~
T h ese ex
p e r i m e n t s ( 2 4 ) were made b y Mr Walker during 1 8 9 5 6
.

at Westminster T he primary purpose was to ascertain

.

W
.

( )1 hether this kind of fan follows the ordinary laws

respecting the mutual relations of speed o f fan power ,

ab sorbed and amoun t of air discharged

, .

( )
2 T he general characteristics regardin g the speed of
fan power absorb ed and quantity of air discharged with
, , ,

diff erent angles of the b lades .

( 3 ) T he e ff ect of fans di ff ering from one another only in

the cross section of their blades
-
.

T he experiments showed that the ordinary laws hold

good and that the propeller fan is adapted to the discharge
,

of large volumes of air at small pressures and further that ,

volumetric e ffi ci e n cy i s more important than even mechanical

efficiency A ll the experiments were made with fan s
.

having a free discharge (except a few at the end ) the outlet ,

b eing t he same as the i nlet T he volumetric e flflci e n cy of

.

the propeller fan is greatest wi t h free discharge and falls ,

o ff rapidly if the discharge pipe is b aflle d S eventeen .

three bladed fans were tested all 2 3 % in dia T hey are

-
, . .

shown i n fig 1 10 T he fans were driven direct from the

. .

spindle of an electro motor fi xed centrally to a cas t i r on

-
r

frame in the rear of the fans ( fi g s 1 1 1 T he motor .

,

was a continuous current series wound machine of ab out

- -

one third of an electrical horse power T he air was

-
.

delivered through a tube 24 in b ore and 4 ft long ( figs . . .

1 14
.
,
made of stiff sheet iron and placed concentric -

with the fan axis and at t h e end of the frame Th e speed .

of the fan was indicated by a ta chometer attached by ,

a Hook e s j oint to the m otor spindle and was read to two

’
,

rev per min I n most of the experiments the fans were

. .

run at 600 rev per mi n 3 the speed b eing kept constant b y

. .

the adj ustment of a suitab le form of resistance T he .

velocity of the air was measured by an anemom eter of

i n dia placed a t the oute r end of the delivery tu b e ;
. .
,

the instrument was calibrated at K e w Observatory for

speeds from 5 00 to ft per min T he velocity varied . .

greatly in di ff erent positions of the same cross section of -

the tu b e A s m ooth b rass rod T ,; in dia was placed

.
5
. .
01
d
C E N T R IF UG AL FANS

n
m
v
m
dl
m

S
C
H
?

m
m
b
. e .

E2 8
m2 3
8
.

fi S
.

m
i a N

m
Q
i
q
o
m
h n
m
s
o
l
mb
m z n

5
s
o
3
. w
a
z E
fi
fi d
.

8
c
z
c
m
c o

fi
o
m
< fi
e

fi
:
. w
m
V.

w
u
b
m
EX P E R I M E N T S WI T H P R OP E LL E R V E N T I LAT I N G FANS 2 27

hori z ontally across the end of the tub e t o w h ich the

anemometer was attached so that the centre of the latter ,

moved in the hori z ontal diameter of the tu b e for all posi

tions on the rod and the instrument always moved in the
,

same plane across the current T he B H P of the motor . . . .

w as obtained b y a dynamometric brake fig 1 1 3 T he fan , . .

having been removed the b rake pulley was fixed in the ,

same posi t io n upon the spindle T he b rake was highly .

e fle c t i v e it was sensitive and ran without oscillation T he ,

.

pulley w s of cast iron

a in dia with smoo t h circum
,
. .
,

ference round the pulley was wound a fi ne silk cord t h e ,

upper end of which was attached to a S alter s balance ’

while the lower e d supporting a scale pan hung vertically

n , ,

benea t h I f \V lb = load i n the scale pan and w= r e ad i ng

. .

of the balance t hen the B H P ab sorbed b y the b rake

, . . .

( “I 2 RN
where R i s the rad i us of the pulley in feet ,

and N is the numb er of rev per mi n . .

F or each ex periment anemometer readings were taken

at each of the four following radii of the delivery tub e :
15 , 103 in T he cross section of the delivery tube
.
-

was divided i nto four imag inary concentric rings and each ,

of the above radii corresponded w ith the centre line of one

of these rin g s e ach of the three outer rings was equal i n
bread t h t o the diameter of the anemometer T he velocity .

of t h e ai r in ft per min as ascertained at each of the four

. .

radii was multiplied by t h e area of the correspondin g

ri n g s in sq f t and the prod ucts b eing added together gave
. .
,

the number of cu f t of ai r discharged per minute. . .

T he velocities g iven at each of t he four radii are g iven i n

T ables 5 4 and 5 5 for all the fans tried T he areas of the .

four imaginary rin g s were 1 2 7 5 09 4 5 0 6 1 4 03 07 sq , , , .

ft T he mea n velocity of t h e air was o b tained by dividing

.

the air discharg e i n cu ft by sq ft Re ad i g s . . . . n

of the anemome t er w ere t aken for t wo minutes at each of

the four radii for each experiment t ogether with volts , ,

amp eres hei g ht of barometer and temperature of air A

, , .

se ries of experiments were mad e with the motor runnin g

at 600 rev per min and the experimental readings ar e

. .
,
2 28 C E N T R IFUG AL F ANS

T A BL E 5 4 .

Ang l e V e l oc it y o f ai r pe r m i n at
l ts
.

of Re v s Vo .

N O of
.

"n
P9 1 03
f an
m in m o t or
’

. .

m o t 01
d eg s . 1g i . 5% in . 7; in . 10% in .

890 7 67 68 8

600
54 5
80 0
783 9 03
8 08

*
6 00
875
*
495 52 7 0
-

*
600

6 45
*
6 00

5 25

24 0 L 1 10
'

18 7

E xpe r i m e n ts on P r o pe ll er V e n ti l ati n g Fan s Fi gs 1 1 0 a n d 1 1 5

, .
,
of
th os e m ar k e d r e vol v in g o u t o f t h e t ub e B l ad e s se t at
. an
5% i n .
, 75 ih . , 1 03 i n fr om t h e ax is o f t h e tub e
. .
230 C E N T R IF U GAL F AN S
2
5
H w 0
.

m . a N
. He N
1 m N
. 3
.
H o m 0 H
e H H
u
o a o o
m H m w m m
V N
. N
. n n w f
o
m

3
o
fi
o

O o
m
.

E
w w m c w
5
H 0 N H, 3 a q
8
.
M a
m m m N H N m 0
H m
€ S
.

m m v H , 3 m m m m
8
. H
5

.
w
a
5
2
c
s
m .
o
8 8 p
o
m
3 m
8
0 m
9 e 8
2 8
o
m 8 %
m S a
m H a

m H s
x ?
m .

E
.

w 2
m
.
a .

o s
m z
v
H
3
:
.

H 8 3
<
3 a
m
.

S w
0 g
m E
.
5 >

3
H

5 o
D >
Em m S
m
v
.

d a
m
m b k fi
e
fl
.
a
m N
o
d
s 5
a .
c m
.

m h c
i fl m 3
a
r m m
v fl
H 2
.

:8
e
.
3 £
9
? 0 0
9
5

m m m O
o 3
6 8

3
a
n
P 2
N
mo w
.
a
m m
H
.
o
H H
n
H H
.

H H
H

m m m
,

N
n
o
5 H
Q
.
O
u o
O
H
’
H
o m m w 4 m
:5
3
a 2
3
N H
w 3 m . n m b N
. w o. m
5
.

5 a
s

m
5

.
N
3 8
6 8
o 8 S
5 H
A
H
 EX P E R I M E N TS WI I ' ' ‘

H P R OP E LL E R V E N T I L AT I N G FANS 23 1

0
0
6

8 6

L4
3
:
Q)
2. %
> 5
0
s: p
a
S
o
2 n
g
.

: :5
0 .

8
’
e

Ll i
4
. 2

H
(N

t
a f
o
e
2 s e
4 b

m:
s
In e u
t
d
-4

Z a
M h y
2 t
i
r
e
wv
O O O O 8
a.
1
a
mod
'
n un 3}
d

E s
d 7
a
1

g
*
3 a i
n

‘ 8 6
v

3 a
1
.
o
v
8 .
3
0
e
r
U
>3 Z
‘ 8 m
3 . 2
a
s
H
e

O
h .

«5
‘0 r —1 H m 9 1 H 1
um
c m
6
3

: 3
fi m
e

3 z
E
u: P 2

< 1 2 o
a
8
.

-1 fi
Z 3
O
r
3
:
4
h
a. 8
u o gqo gy e luo o
3
'

S p ai gub a t
oq 3 n
o
d H d H 8
. . . . .

8
.

= 1 010 m 30
l o o o o o o o N c
o c a
m
n
e
2
I 1 011 1} w
' ' .

q E
um 11 ;
p aq os i
p 1 1V S
'
1
r
0 e
0
e p
C
m J 0 a IBu V
'
se e
p n
a v
.

l
e
p r
 232 CE N TR I F UG AL F AN S
shown i n fig 1 16 I t appears from this di ag r a m that the
. .

current varies directly as the di ff erence of tensions in the

cord hence th e relation between the current and torque
,

can be expressed by a linear equatio n thus ,

T= a C —
h
where T : torque 0 : current and a an d b are constants
, , .

He n ce the B H P of the motor at 600 rev per min

. . . . .

Tm ( a C -
b) w

J ar E le ct r ic 2110 10 1 :

[ 00

50

5 lo :
5 20 25 30 35 so

Fr o . 1 16 .
— EX P E R I M EN T AL RE AD I N GS AT 6 00 RE V . P ER MI N .

( Se e Tab l e 5 4, p . 228 Se r i es E l e c t ri c M o t o r ) .

0, C u rr e n t in am p er e s E, e e l c t r o m ot i v e f o r c e T, t o r q u e ; r ad i u s o f
p u ll e y in i n c h e s x weig h t in o un ce s .

where T is in lb feet and w is i n radians per min With

.
, .

a given torque the amp eres were n o t quite constan t

for all speeds of the motor ; they increased slightly and
uniformly with the increase i n the number of rev per min . .

I t w as easy however to frame a formula which gave the

, ,

torque at any speed the motor might b e running at t aking ,

into account the small increase of the current due to speed .

A lthough not essential for the presen t experiments it w as ,

interesting to d ete rmine at what speeds the motor should

234 CE N TR I F UG AL F AN S
into account the moisture in the air t h e weight of , 1 cu ft ‘

of air
. .

1 3 3 04
( B g 6) approximately
F 461 ,

where b is the pressure due to the moisture in in of .

mercury see section 6

I f W weight of air discharged i n lb per sec and
.

,
. .

V = v e 10 01 t y of air in ft per sec then the kinetic energy of

. .
,

the air dis charged is and the horse power of air

g
discharged is
WW
2 X 5 50
V cons t ant 3
,
9

for the sa m e fan under same conditio n s Hence if Q .

cu ft of air per sec the horse power of the air discharged

.
.
,

V 2
Q 1 3 3 04 B
-
V2 Q B

h or se power I n an d i scharged
Th e mechanical e ffi ci e n cy = ,

volumetric effi ciency is as usual 1) 7

, 2,
and dynamic pressure
2 2
2

e ffi ciencies are evidently O the static pressure efficiency

T
2,
2
bei g ero
n z .

E xperiments were made with fan blades at diff erent

angles to the plane of rotation Th e results with fan 1 7 .
,

having plan e surfaces and rounded backs to the blades are ,

given in T able 5 5 and plotted in fi g 1 1 8 These may be

,
. .

termed the characteristic curves of the fan for varyi g n

a ngles .I t should be noted that maximum vol umetric

efficiency is not obtained with the same angle as maximum
mechanical e ffi ciency I n Table 5 6 fa s 1 6 and 1 7 have
. n

been compared and the latter certain ly has the better volu
,

metric e fficiency by a very small amount and the m e ch an i ,

cal efficiency of fan and motor which of course includes

‘

, , ,

motor friction is better for the latter 3 but whe n we deduct

,
E X P E RI M EN TS WI T H P R OP E LL ER V E N TI LA TI N G F AN S 23 5

01 0

0 09

1 7 50

0 05

0 03

15
‘
20 ‘
25 °
3 0‘ 35 ° 40‘ 45
°
5 0‘ 55° 6 0‘

F IG . 1 18 .
— P L OT 1 I N GS
’ ‘

or T A BL E 55 (p .
236 CE N TR I F UG AL F AN S
O O3SS
°
horse power for the bearing friction of the motor ,

the e ficiencies of fan alone are as below

f

A ngle of vane
° ° ° ° ° ° °
15 20 25 27 30 35 40
Mechanical e ffi ci
ency of fan 30 8 46 0 4 2 4 40 2 3 3 7 29 3
,

N o 16 .

Me ch an i cal e

ency of 3 3 C 38 7
°

426 42 8 416 37 O 3 14
°

N o 17.

So that on e is about as good as the other Seventeen th r ee .

bladed fans were tried in order to test the eff ect of the
cross section of the fan blades fig 1 10 They may b e
-
, . .

divided into four groups Th e first comprises fan s 1 to 4 .

,

the second 5 to 1 0 the third 1 1 to 1 5 and the fourth 1 6

, ,

and 1 7 Th e blades were of sheet i ron T 5 in thick and

.
1
.
,

excepting 1 0 their cross sections are either lines or arcs of -

circles Th e fans in each group di ff er e d from one another

.

only i n the cross section of their blades wh ich were linear

-
, ,

plano convex concavo convex of diff erent curvatures

-
,
-

,
.

Fan 1 h ad flat blades Fan 2 was formed by fixing a .

circular back to fan 1 Fan 3 was forme d by curving the .

blades of fan 1 Fan 4 was formed by fixing to the back

.

of fan 3 a still more convex surface Th e blades of the .

other groups were similar i n form but of diff erent area and ,

thickness T hese changes in shape produced considerable

.

e ff ect Fan 1 to fan 4 were all tried with their blades at

.

1 7 deg inclination . Th e superiority i n mechanical and

.

volumetric e f ficiencies of the last should be noticed .

I t appears that on exposing the perimeter of the fan air ,

is sucked into the outer circumference so that the volu ,

metric efficiency is largely increased see fan 8 at 40 deg — .

Some of the fans were tried with exposed perimeter by

moving the delivery tube 4 % in forward as sho wn i n .
,

dotted line in fig 1 1 4 Thus in the case of fan 9 the

. .

mechanical and volumetric effi ciencies were increased from

1 6 9 to 2 9 4 and 6 2 to 7 8 per cent respectively A much . .

wider form of bla de may be used i n fans arranged to feed

from the tips This type of fan should therefore where
.
,

possible be fi x ed with its circumference exposed

,
.
23 8 C E NTR I F U G AL FANS

l l9 .
l 20

A in

IQ ] .

FI G S 1 19 AND .A N D 4 8 I N F A N S FI G 1 2 1 — C U R V A
1 2 0 — 24 -
IN .
-
. . . .

B L A DES I N F A N FI G S 1 1 9 A N D 1 20 FI G S 1 22
.

T U B E OF ,
. . .

A N D 1 2 3 — F A N e I I S P E C I A LL Y D ESI G N ED E L E C T R I C Mor o n
. .

FI G 1 2 4 — FI E L D M A G N E T S ( I N S E C T I O N ) OF F A N FI G S 1 2 2
. .
, .

A N D 1 23 .

E FF E C T OF D I SC 0N F A N S I N FI G S . 1 22 AND 1 23 .
EX P E RI M ENTS WI T H P R O P E LLE R V E N T I LA T I N G FANS 239

I n figs 1 2 2 1 23 is shown a fan with a specially designed

.
,

electr i c motor Th e field magnets are cylindrical and are

.
,

shown in section in fig 1 2 4 ; they are made as small as .

possible in order to reduce the resistance off ered to t he

pas sage of the air Th e best position of the motor from this
.
,

point of view was generally found to be a little in front of

,

the fan the exact position depending on t h e si e of the

, z

ori fice There appears to be a central region immediately

.

in front of the fan where only a little stream of air is

delivered owing probably in fan s working with a free dis
, ,

charge to the centrifugal action on the front face of the

,

blades which is apparent near t h e centre

, .
23

Several experiments were also made with contract d out e

let and inlet with a fan 2 3 % i n dia the blades being set at . .
,

3 5 d e g to the plane of rotation

. Th e fan w as driven at .

8 00 rev per min from a shunt wound motor ; i t discharged

. .
-

i nto a 2 f t dia delivery tube 4 ft long with partially closed

. . .

outlet having central holes 6 i n 1 2 i n and 1 8 in dia

, .
,
.
,
. .

U nder these con d itions the fan was tried both for propelling
and exhausting air and its efficiency was i n both cases ,

much reduced Th e fact that this type of f a n is unable t o

.

maintain static pressure is probably d ue to the compara ,

t i v e ly slow speed of the blades near the centre in ,

consequence of which the air tends t o pass back agai n

through the centre of the fan Th e eff ec t of fixing a .

circular disc on the delivery sid so as to prevent the air e,

from returning through t he fan w as t o increase the ,

e fficiency to a great extent when working against resistance ,

whereby a static pressure w as obtained in the air delivered .

Experiments were made with discs of di f f erent d iameters ,

and it was found that in order to obtain a good efficiency ,

the si e of disc should increase with the contraction of

z

orifice Th e reason for this is evident because a and v are

.
,

g reater n earer the outer circumference and a 71 must always ,

exceed 9 H .

Figs 1 2 5 1 2 6 1 2 7 S how the eff ect of a disc with fan

.
, , ,

fig 1 2 2 1 2 3 on the circulation of the air ; in the case in

.
, ,

whi ch the disc is omitted the air to a reat extent returns

through the centre of the fan C o nsi gc r i g that this type . n

of fan is used for drawing air through a material to be

2 40 CE N TR I F UG AL FANS
dried like wool or through tortuous flue s as in refrigerating
, , ,

apparatus the adoption of the central disc becomes a

,

necessity Th e fan is more efficient when exhausting than

when producing pressure Without a circular disc 3 1 cu ft
.

. . .

FI G . 125 .
— WI T H D I S C ( DD ) , C ON T R ACTED O U T L E T P , RO EP LL I N G .

of air were dri ven through the 6 in orifice per m i n . .

,

when running at a speed of 8 00 rev per min Th e . .

delivery was increased t o 4 5 1 cu ft whe n e x hausting the . .

,

FI G . l 26 . WI TII D I SC
-
, C O N T R A C T ED I N LET ,
EX H A U S T I N G .

other conditions being identical ; the volumetric efficiency

w as thus increased nearly 15 times With the 1 2 in orifice . .

the d ischarges were 4 9 7 when blowing and when

FI G . 12 7 .
— WI T H OUT D I SC , C oN r '

R ACT ED O U T L ET , C I E CUL AT I N G .

exhausting With the 18 in orifice the volumetric efficiency

. .

was increased only from 5 8 % per cent when blowing to 6 7 .

when exhausting T he advantage of suction over blowing

.
»

is largely d ue t o the fact that the dynamic head is very

24 2 C E NTR I F U G AL FAN S
noticed i n the axial discharge of the air Th e angle at the .

tips was 2 7 deg with the plane of revolution Th e helical

. .

blades were made of in sheet bras s pressed on a wood .

mo uld Th e fan was run at 600 r e v and the anemometer

. .
,

w as placed 1 8 i n i n front of the fan N o delivery tube . .

w as employed and the axial velocities at radii of 2 i n

,
.
,

5 i n and 1 2 in were 7 7 0
.
, and 2 3 0 while at the .
, ,

centre the V locity was 6 6 5 Th e rotary velocities were

e .

also measured by placing t h e anemometer Wheel i n a plane

passing through the axis Th e velocities at radii of 2 i n . .
,

5 in 8 ih
.
,
and 1 2 in were 23 4 f t 5 6 2 f t 5 3 0 f t and
.
,
. .
,
.
,
.
,

1 85 ft per min . Th e e fficiency was increased by putti ng

.

rounded backs to the blades but the experiments showed ,

that helical blades did not possess any advantages over th e

ordinary non helical blades With fans of later design
-
,

having plano convex blades Walker obtained volumetric

.

efficienci es of 8 6 to 9 0 per cent but he does not give the .

,

angle of blade .

Th e horse power necessary for driving the fan to produce

a given discharge of air i s as follows Taking the barometer
at 3 0 i n the temperature of the air at 60 deg Fah and
.
,
.
,

the mechanical efficiency at 30 per cent let d be the .

,

diameter of the fan in feet a the ar a of the fan disc in ,

e

square feet V the veloci ty of the air in ft per sec and Q

, .

t h e quantity of air discharged i n cu ft per sec then H P . . .

,
. .

required to drive th e fan

:
H P in d s ch a g e d air V Q B 00 0003 7 5 6
. r
2

x
mechanica l efficiency I

and substituting 6 0 4 6 1 and V : 3 ,

3
Q
3

HP . . required to dri v e the fan = 1 '

17 x 10
-
5
83
.

Further Q : and taking the volumetric efficiency

at 9 0 per cent .

Q z

85
'

Where N rev per min Thus a fan 2 in dia running at

. . . .
,

6 00 rev per min discharges

. cu ft per min .
,
. . .
 S C R EW FANS 24 3

I t is evident from these two formul a that d should be as

large as possible Of course these formul a only apply t o
.

free discharge Th e eff ect of i ncrease of diameter is shown

.

by the fact that to discharge cu ft of air per min . . .

the 2 ft fan would require 0 7 3 horse po wer while a 4 ft

.
,
.

fan would need only 00 4 5 .

C H AP TER XI I .

OT H ER P R O P E LL E R V E NTILATIN G F ANS AN D R AT E A U
SC R E W F ANS .

TH E R E
are many types of ventilating fans i n use at the
present day Th e H attersley P ickard fan is designed not
.
-

only to deliver a large volume of air but to discharge ,

against a considerable static head I t is fitted with a boss .

o f large diameter ; this is because the central position of

the propeller is not only useless for movi n g air but is ,

absolutely harmful when working again st pressure Th e .

bl ades are inclined towards the intake so as to enable the

air to cross them at right angles and therefore with the ,

least possible friction They are of helical construction

.
,

wit h a longitudinally or a x ially increasing pi t ch f a I is —

,
( )

greater than 6 and at the outer circumference they app e ar

,

excellently adapt e d for drawing in the air radi ally as well ,

as a xi ally T his appears to be a most carefully designed

.

fan Th e following is a list of particulars of standard si z es

.

W Re v ol ut i n o s Cu f t
’

o f ai r
E_
Em s
5 33 2
s i saa:
e
g
gg
“

p r m in
.

g
. .

S
V
9 3 pe r m i n ut
-
e . e .

30 «H
a g
w
o

s
3 3 0

E
g s a g
0
0
B
“
O .a
. ,

Q ” 3 G O V

18 7 00 to to I to g 3} II 113
24 5 00 t o 900 to ,, to g
1
4 12 117
30 4 50 to 7 50 to I to I 5 2
36 4 00 t o 65 0 to I to 1} e 2;

48 3 00 to 350 to I to 2} 8 3; 1 18 94,
244 v
C E NTRIF U G AL FAN S
Th e Blac k man fan is very largely used i n this country .

Figs 1 2 8 and 1 2 9 Show a propeller fan made by B eck and

.

H enkel , of Cassel , designed to discharge large quantities of

air at low pressure Th e vanes are made of steel , and
.

revolve i n a conical casing These fans are made with

.

FI G . 128 .

diameters from 1 0 in to 1 1 8 i n they can be driven direct

. .

or by transmitted power .

Rateau Scr e w F ni a
a Th e propeller type o f fan is cer
—

t ai n ly that best suited for delivering a large amount of

air at an e x ceedingly small pressure ; but where a large
amount has to be discharged agai nst some considerable
‘

head a modification of the propeller fan in which only the

, ,

outer port ions of the blades the part Where 7) is great )

246 C E NTR I F U G AL FANS
cylindrical section through the guide and wheel blades of
a Rateau screw fan Th e direction of the air s m otion i s
.
’

S hown by th e arrows Th e guide vanes are m m and the

.
,

wheel vanes a b Th e motion of the wheel is upwards

, .
,

a n d the wheel vanes are so designed that inflow takes

place at th e normal orifice wi t hout shock Th e sectional .

elevation of fig 1 3 1 needs no co m ment except that the

.
,

chamber D has its inner side p artly conical in order to

FI G . 1 30
.

reduce the velocity of the air by increasing the section of

discharge Th e Wheel fig 1 3 0 has its vanes formed of
.
,
.
,

steel plate fi xed to the rim of a slightly conical wheel of

cast iron or bron e by means of angl e irons i n the case
z ,

of large fans or by embedding them into the rim in smaller

,

si es A general view is seen i n fig 1 3 2 A n alternative

z . . .

arrangement is Shown in figs 1 3 3 1 3 4 Th e spiral admis

.
, .

sion chamber gives the entering air velocity in the opposite

direction to that of rotation and after leaving the fan with
,
 S C R E W FAN S 24 7

a velocity Wholly axial the air is discharged through a

,

passage whose section i s increased b y making its inner

surface t h e f r ust r um of a con e A third m e thod is shown
.

in fig 1 3 5 ; here the air en t ers parallel to the axis and

.
,

FI G . 13 1 .

the vanes at in flow are so inclined as to receive it with

out shock Th e change of the moment of momentum is
.

eff ected by pr viding a volute to reduce the tangential

o

motion at discharge and so convert the kinetic energy of

, .
2 48 C E NTRIF U G AL FAN S

FI G . 13 2 .

FI G . 133 .
o
t
!

UQ C E N TR IF UG AL FAN S
where Q cubic feet of air per second ,

A = area of S ection in square feet ,

c = c oe ffi c i e t of contraction
n ,

H = pressure head i n feet of air ,

h water gauge in i n ches ,

3 = density of water ,

O
.
= density of air .

FIG 1 35
 S C R EW FANS 25 1

TE ST O F AN A X IAL 0 R SC R E W FAN O F A TY P E SI M I LAR

To T H AT S H O W N I N F I G 1 3 5 O F 4 6 9 FT E X TR E M E .
, .

D I A M E T E R 0 F V AN ES .

N o of
. r oluti
ev on s pe r
mi n ut e 5 26 5 32 552
.

W te r g g i
a au e n ins .

K i l o w tta f s o 9 85 101

of wo rk
se con d 23 90 3 030

Di sc h arge in c u. ft .

pe r se c nd
o 0 296 386

Me ch an i al e ffi ci e n cy
c
of f an an d m oto r O

Man o m e t r i c e ffi ci e n c y 0 21

Vo lu m e t r i c e ff i c ie ncy 0 0 42

I t s h ould be n t d th t
o e a m ax i m um v ol u m t ri c
e an d m e c h n i c al
a e ffi ci e nci e s
o c u r at
c th e sa m e ri fi
o ce .

The Theor y W shall first con

f
o Rateau Scr ew Fe m — e
sider the type shown i n figs 1 3 0, 1 3 1
and 1 3 2 Re . .

ferring to the sectional view of guide and wheel vanes ,

fig 1 3 1 let be the outlet angle made by the gu ide vanes

.
,
a

m at the mean radius of the wheel wi th a plane pe r pe ,

n

d i cular to the axis 0 qt the angles made with such a plane,

by the wheel vanes at inflow and outflow L e t e m be the .

speed of the wheel at the mean radius r and b 1 the a xi al m

componen t of the air at discharge from the guide wheel .

Then from the velocity triangl e ( fig if the air enters .

the moving wheel without shock ,

l
cot 6 — fi l + c ot
‘
a.
bi

Again , at outflow which is a xial i ,

— e ,
a
2
= is the
0— i f be
axial component of discharge which , in fig 1 3 6 is greater
.

than then
e 62 cot m
25 2 C E NTRIF U G AL FAN S
Th e work done per pound of air by the wheel is
a vm
l
cm 61 cot a

9 9

Th e loss by friction in the guides is

6“
F cosec 2 a
,

Gu i d e
FIG . 1 36 .

the frictional loss i n the wh eel

F2
_
cosec 2
9 52 2 2
00 8 00
49

and the leaving losses are

where ba = v e loci t y of discharge from the fan casing

left of D .
254 C E NTR I F U G AL FANS
we obtain an equation between H c m , , 62 , 3
2,
0
.
3 ,
and a

III the form of a quadratic i n c m and

F s

A c
1

F1 62 2
+ '2 9 H —
O
t 3 S
e
a 3 3

I n designing this type of fan ho wever it is best to com ,

mence by making certain assumptions We can safely

,

assume that t h e mechanical efficiency of t h e fan alone is

about 60 per cent and as this type of fan is intended to
.
,

give a h ig h volum tric e fficiency we should take this as

e ,

7 5 per cent ; then

.

volumetric efficiency 2

where v
,
peripheral speed ,

radius of the tips of the blades ,

z
2 71 '

8
1
01 2 7r Tm 8
1
01
8
2 ]

m
U
’
7
2
v m
'

l
‘

N ow , we shall suppose that 9 .

1

2 71
'
l) 1 b
so that 4 ,
x 13 6 am am ,

and putting nv
= 0 '

7 5 I) ,
,
v’ " vm

gH {1 H
a
1
vm
’
Um 01 cot a

{7 H
00 5
06 x vm
z

nm
3 b1 7
b2 1 I)
l
let us say ,
6
0 66 3 cm ,
 S GR EW FANS 25 5

so that cot ¢ > = 3fl = 1 -5 1

g2

I
( ) 33 deg . 40 min .
,

i ndependent of the manometric e fficiency and dependent ,

only on r m, and 82 A gain .

,

vm
cot 6 —
+ cot a
b1

Th e following is a list of values of and 6 for various a .

values of 7 calculated from the above equations

7 m,

per cent . 10 20 40
'
60
'
° ' ° ’ ° °
a. 7 3 40 59 3 8 40 29 29 38
° ’ ° ’ ° ' ° ’
0 25 5 8 23 5 1 8 50 15 52

I n the next type of fan with a spiral inflow passage but

no guide vanes th e manometric efficiency will i ncrease the
,

greater velocity of the i nflo wi n g air ; the speed of the air

i n this passage is the ta ngential component of the velocity
of inflow ,
a.L e t A be the sectional area of the passage ,

and Q the discharge i n cu ft per second then . .

a
Q c
1 l 3
2 o 7r o r 1n o s
l

62 Q and em b2 cot < 3

,
2 1r r m

v’” d
cot 6 l

b1

Then w e may proceed as follows

7
9 H a
9 H
} 1
“ 1 vi ii vm 7
)

J 7
7 "
:
k say,
1
?
256 C E NTRIF U G AL FANS
so that if 7 an d 1m are ass u med al and c m are k nown in
7 7 . ,

term s of H A ssuming. and n = 7 5 per cent and v .

,

v
3 : fl as before then ,
1
3
(l @
00 t 6
_ l 711,

05 68 cm

:
and is n ow k nown ,
and 62 0 6 63 cm as before .

Further, cot 1
and as before
° ’

qt 15 1 4) 33 40 .

We need not deal with th e type of fan in fig 1 3 5 as its .

,

theory is precisely the same as that of the radial flow fan -

with vol ute .

I n designing the type fig 1 3 7 we sh all suppose th e , .

,

volumetric e ffi ciency 7 5 per cent as before and that .

,

2 4
3 : r m
O
9 7m
E
' /
1 a 2 .

11

th
2 1r 7
‘

"6 1
2
8
2 1
7r 61

05 66 77, 0 4 2 5,
-

22 3
1
bl ~
o 77 s ,
vm 82 ’
Um

am being of course the velocity at the mean radius .

We can now find the vane angle at inflow Since

, ,

vm
cot 6 23 5
bl
and 6 23
°

1f is the angle made by the mean direction of motion of

4)
the air relative to the wheel at outflow then ,

cot <> i
" m

b2
a
2
12 86 1
§5 8 CE N TR I F UG AL FAN S
air leaving the wheel has an axial component b2 and
Th e
a tangen t i al component a and if the guide vanes are to 2,
receive the air withou t shock they must be i nclined at
angle such that
a

6 Um
tan a
‘

a a
2 2

7
]m

following
Th e is a list of values of i f) and a . for various
values of M
Mp ernt ce 10.
2 20 30 40 60 80
° ' ° ’ ° ' ° ' ° ° '
43 4 49 21 57 16 66 4 8 90 1 1 3 12
° ' ° ' ° ' ° ’ ° ' ° '
a
. 7 7 52 66 46 5 7 15 49 2 1 37 49 30 12

C HA P TER X III .

FANS AN D C OMP R ES S O RS :R E C ENT P R AC I I C E

'

I MP ROV EM ENTS I N FA N D ES I G N .

I N centrifugal fans especially of the drum type i n which

, ,

the blades are long i n an axial direction compared with the

radia l depth there is a tendency for the reduction in
,

pressure at the inlet t o cause the air on this side of the -

fan to be actually drawn from t h e discharge t o the suction

through the blades This was first noted by S D avidso n
. .

when e xperimenti n g with a fan wheel open to the air at

one end ( fig 1 3 8 I ) and rotating without an enclosing casing
.
, .

of any sort Th e outward discharge of air was strongest

.

at the back or closed end A but the strength of this ,

discharge gradually diminished and bec ame m l at a point

0 nearly mid way therefro m and from this point onwards ,

i t assumed a centripetal flo w Fo r this reason h e termed .

A the positive and B the negative end of t h e fan

'

,
When ‘

.
 IM P ROV E M E N TS IN FA N D ES I G N 25 9
'

FI G . 138 — I n r
. FA N S no v n n .

I .
, Ill u s t rat i n g t h e e uli l I cy c l o n i c m o v e m e n t o f ai r i n an o pe n d r u m
c
‘

ty pe f f a n I Pa r t i ti o n pl ac e d i n a n d ou t s id e a fa n to pre v e n t
. o . s

t h a i r m o v e m n t s s ho w n i n I
e e Des i g n o f c a i n g f r e n c ase d
. s o

f an to c o m p l t h e n e ga t i v e e n d o f t h e w h e e l to d ra w f r o m t h e
e

s u c ti o n l e g i o n. I V V V I Dru m ty pe o f c e n t r i f u g al f an i n
. , .
, .
,

w h i c h t h e a n g l e o f b l ad e a t t h e i n l e t e n d i s m ad e l ess th a n a t t h e
d i s c e n d to ca u e a m o r e e ve n l y d is t r i b ute d d i sc h arg e
s .
2 60 C E N TR I F UG AL F A I\ S
such inflow takes place it must be due to the reduction of
pressure at 0 being s o great that the pressure di ff erence
between the discharge and suction is large e nough to force
air against t he cen t rifugal action of the blades To .

counteract this he proposed to place a par t ition in the fan

wheel to separate the positive from t h e negative end and ,

also t o so place t h e fan with respect to the wall separating

the suction and discharge regions of the outside air that
the negative end would operate i n the suction region as
shown i n fig . H e even goes so far as to propose such a
design as shown in fig 1 3 8 for a fan of the encased type
.
, .

Th e curved part of the cas i n g approaches the fan wheel in

such a manner that it admi t s air from th e suction side to
the outside of the negative end of the wheel whilst ,

leaving only the positive end to t h e discharge These .

d ifficulties show the fruits of depar t ing for the sake of —

cheapness of construction from t h e more rational design —

of the Rateau fan in which the discharge area of the fan

,

wheel is limited to the positive end and is much educed r .

I t is evident that many designers have been well aware if ,

not of an actual i nw ard flow of a very unequal d i s t r i b u,

tion of discharge along the wheel Fo r examp le t he blades .

,

of the K eith fan are deeper i n a radial directio n at the

negative than at the positive end of the wheel and it is ,

claimed that the added centrifugal eff ect due to this

increases the discharge at t h e s o called n gative end -
e .

A gain Siemens and D A Hae k e t t have patented a form

,
. .

making the angle at which the fluid is discharged from the

vanes less at the intake t han at the disc or positive end of
the fan This angle of discharge is defined as t he angle
.

between the vane and the t angent to t h e circumference of

t h e wheel Thus the angle B( fig 1 3 8 I V ) at the intake end
. .
,
.

being less than at the positive end the air at the intake
a ,

end is given a greater velocity of discharge Th e blade .

consists of two parts namely A the blade proper and a lip B

, ,

( fig
. 1 3 8 ,
Th e lip B is attached to the disc C ( fig 1 38 V ) ,
.

by point welding Th e edge D of the part A ( fig 1 3 8 V I ) is

. .
,
.

straight the outer and i nner edges E and F are contained

,

in one plane ; the edge G appears concave when looking a t

the front side Th e direction of rotation is indicate d by
.
2 62 C E NTRIF U G AL FANS
is riveted to it When dealing with the rivets placed close
.

to the periphery of the wheel a suitable tool such as a ,

holdi n g up hammer E for holding the h eads of the rivets

-
,

is passed through t h e outer Opening Of the wheel When .

dealing with the rivets situated towards t h e inlet of the

w heel however the holding u p hammer cannot be con
, ,
-

v e ni e n t l
y held fast during the ri v eting Operation and it is ,

therefore necessary to pass t h e hammer thro ugh t h e i n let ,

as shown in the fi rst fig But in building up a small .

delivery whe el the two discs are so close to one another

as to render it impossible to pass the hold i ng up hammer -

through the inlet This difficulty is overcome i n a method

.

of construction recently patented by Rateau by forming ,

the rear disc of a flat or slightly conical member F having

a rearward extension 0 to form a h ub When the vanes
'
.

have been riveted to the disc F and t h e outer disc A has

been placed in position as before access to the front of the ,

d i s c F remains qui t e free Th e holding u p hammer can

.
-

then be introduced straight within t h e wheel for enabling

the rivet heads to be held For fi nish ing the wheel afte r
.

it has been m ou nted on the shaft a distance piece H is -

added i the form of a sleeve having the require d profile

, n ,

to constitute the bottom of the inlet This sleeve also .

serves to maintain adjacent discs of successive wheels at

the required distance apart i n cases where a nu m ber of
wheels are used on the same machine .

STE PP E D V AN E S F OR C EN TR I F UG AL F A NS .

In order to give greater rigidity to stepped vanes or

blades of centrifugal fans A lldays and Onions and G F . .

J enks Birmingham have formed on t h

, ,
back of the e

blades a rib or short longitudinal corrugation at the

part where one plain part passes by a step i to another n

plain part or rib A is a ring or disc to which the inner

.

ends of the blades B are fixed as is usual and C is the , ,

’

ring to which the outer e n ds of the blades are fixed .

Each of the blades is provided with two Shoulders or

steps D and rigidity is given to the blades at t h e stepped
,
 STE PP E D V A N FS F OR C EN T R I F UG AL FANS 2 03

parts by ribs or short longitudinal corrugations Th e .

vanes B may be arranged inclined t o the disc A in stead

of at right angles to them as shown .

FI G
. l 4 o — ST E PP E
. D BLAD ES F OR A C E N T R I F U G A L F A N WI T H
, R I BS
Fo r . s T i F FE N I N C T H EM A T T H E J OI N TS .

A RRAN G EM ENT OF F ANS .

It is extremely important in arranging fans that they

should be fixed i n such positions that the flo w at discharge
is interfered with as little as possible even at some distance
from the peri phery of t h e fan I n this work it is useful to
.

bear i n mind the molecular theory of gases and regard t h e

air molecules as solid particles travelling in straight paths ,

until they impinge on some surface when they rebound , .

following more or less the ordinary laws of impact li .

they meet a surface normally they w ill be dri ven directly

back agai n ; of cou r se no surface is perfectly regular o
, ,
s

that a certain amount of scattering is caused by this and

by the returning molecules colliding with each other A .

molecule returning to the fan at high velocity will require

2 64 CE NTRIF U G A L FANS
more work to be done upon it to induce it to discharge
again than was needed to discharge it i n the first instance .

Th e commotion caused by a multitude of such erratic mole

cul e s disturbs what may be called th e stream line flow t‘

Of the fluid T h e proof that this is not mere t h eori z ing

.

may be found in the recently developed molecular air pump

and i n recently protected methods Of arran ging fans For .

example in torpedo boat destroyers the fan for producing

,
-

force d draught is usually uncased and is placed in one of

the upper corners of the stok e hold close to the deck and
S ides Th e result is that the air which is forced out radi
.
,

ally all round the f an strikes against the deck and sides
,

and the beams supporting them Th e f an should be partly .

or wholly e n cased with two or more curved deli very outlets

,

so arra n ged as to guide the outgoing air directly into the

stokehold and away from any Obstructions Fi g 1 4 1 . .
,

shows a partly encased A llen fan with t wo delivery outlets

-

0 arranged so as to avoid the overhead girders shown in

,

section A gain in the case of fans for t h e stokeholds of

.
,

steamships where several fans have to be placed Side by

,

side i t is found that some O f the air delivered from one

,

fan obstructs the delivery of air from an adjacent fan To

overcome thi s W H and R
.

, A llen have patented

. . .

d eflectors to guide the air in t h e desired path Fi g 1 4 1 . .

,

shows the deflector 1) for two adjacent fans rotating i n

Opposite directions and the arrows indicate the paths taken
by the air at discharge the intake being as in the previous
case at the eye of the fan Without thi s deflector there
, ,

.
,

is a tendency for the delivery of one fan near the thinnest

part of the volute to enter the other fan Fi g . .

shows the deflectors D used where t w o adjacent fa s rotate n

i n the same direction the lef t hand deflector preventing

,
-

the molecules from bombarding the awkward faces of the

casing and the right hand d e fle et i g the stream of its fan
-
n

gradually into the delivery of the le f t h d fan o

an .
266 C E NTR I F U G A L F A NS

CON STR U C TI ON OF C ENTR I F U G A L F ANS .

Much attent i on I S apparently being paid by designers of

multi wheel centrifuga l fans to the principle of construe ~
-

t i on i n eff orts to Obtain good balance easy and rapid re pair , ,

efficiency and strength Fi g 14 2 l and I l how a t wo

. .
, s

stage fan designed b y P \V A llday and T H P lummer

. . . . .

FI G 1 4 2
.
— T wo STA G E C EN T RI F U G A L FA N
-
.

L o n g way s e c t i o n
. El e v a t i o n o f i nn e r h al f o f e n d c as i n g s ho w
.

i n g g u i d e v a n es .

The middle casing A h as axial inlets at opposite sides and

resembles generally the casing of an ordinary double inlet -

single fan the end casings B B are connected by set screws

or studs to the Opposite hal ves of the casing A Th e two .

end fans B B force the air into the axial inlets of A Th e .

casing of t h e middle fan and base of the compou nd fan

integral with it is divided into two symmetrical halves i n
a vertical plane per pendicular to the shaft and the two ,

halves flanged and bolted toget h er Th e runner of the .

middle fan casing receives axially the air compressed to a

 C ONSTR U C TI O N OF C EN T R I F UG AL FANS 267

certai n pressure by the e n d fans and d i scharges it i nto the

tangential discharge O pening A 1 opening preferably i n the ,

base Th e casings of the t wo outer fans are made i n two

.

sy m metrical halves meeting together i n a hori zontal plan

,
e

containing the axis of the shaft Th e i nner walls of the .

end casings have curved webs or vanes 0 ( fig 14 2 which .

,

guide the air f rced into t h e circum ferential s pa ce I n t o the

o

axia l eyes of the midd le fan T he bearings f o r the runner

.

shaft are carried by the lower halves of the outer casings ,

and the common h ub of t h e three runners is made i n two

h alves meeting together end to end at t h e middle of the
i nner fan By this arrangement a con venient means of
.

balancing the runner is obtained as the one half can be ,

partially rotated with respect to the other half when balance

of the runners is being tested When the best balance is .

ascertained the three runners are secured together by the

,

bolting of the t wo halves e nd to end To prevent the .

return of the compressed air to the inlet Openings of any

f n concentric flanges ( shown at CJ Of middle fan fig 1 4 2 I )
a ,
I , .
, ,

which w ork between annular concen t ric flanges O n the

casings an d constitute obstructions ar e fixed to the ru n ners
T hes e flanges have no doubt been suggested by the laby ~
, .

r i n t h packing method of steam turbine practice Th e fan .

is patented .

C ENTRIF U G A L GAS P U RIFI E R .

Gases containing tar or dust in suspension may be puri

fi e d by being passed through a centrifugal fan of special
construction By centrifugal action the solid particl es are
.

flung against the interior surface of t h e vol ute and collect

t h c e whilst the purified gas passes to th e discharge duct
r

When the fluid to be purified holds in suspe sion substances

, .

l ia b le to liquefy by the action of heat such as for example , , ,

when treating t ar lad e n coal gas a heating jacket for

.

de tarring is provided around the volute Since the fl uid

-
.

leaves the i m peller with a considerable velocity of whirl it ,

is evident that there is a loss of energy h e r e unless some

useful purpose can be served I n an ingenious arrangement .

by Rateau illu trated the gases after passing over t h e fan

, s , ,
2 68 A
C EN TR I F U G AL FA N S
blades B and through the v olute or whirli n g c h amber 0 ,

are made to pass over the blades E of an in ward radi al

flo w turbi e before being discharged the fixed guide blades
n
,

of this tu bine bei n g shown at D Th e turbine may be

r .

fixed o n t he same shaf t as the fan and its action red uces
‘
,

the power necessary to drive the fan or it may be fixed on ,

a separate shaft run n ing at a diff erent speed and used for
some external purpose Th e jacket 0 through which .
,

steam or suitable hot gas or liquid is passed is for de ,

tarri ng the water of condensation in the casing being with

,

FI G . 143 ,
— CE N T R I F U G A L G A S P U R I F IE R C O N S IS T I N G
,
OF A FAN
AND A T U R B I N E I N S E R I ES .

drawn through t h e outlets K an d t h e melted tar passing

out through an outlet L which should be immersed i n a ,

water seal in order that the extraction shall take place

automatically Whe n dust is to be extracted from the gas
.
,

water under pressure is inj ected into the fan inlet throug h
tuyeres M the particles of dust being thus caused to
,

agglomerate and then to be separated from the gas I n

chamber 0 by centri fugal action the dust laden water ,
-

passing out through the Opening N I n some case s instead .

,

of placing fan and recuperation turbine back to back as

shown i n fig 14 3 they are separated and the whirlin g
.
,
270 C ENTR I F U G A L FANS
B into the conduit C th e suction being produced by the
,

fan D the eye of which is connected through pipe E to 0

, .

A fter passing through fan 1) these gases are led by an

enveloping flange F and curved webs and vanes G to the
eye or inlet Opening of the pressure fan H Th e fumes or .

dust lad e n air u nd r pressure i n the casing of the fan H

-
e ,

passes into the main K from which t hey may be taken by

,

b ranch pipes L to the hearths of the black s miths forges A

’

or to furnaces or other appliances in which they may be

used Th e patentees ar e E A llday and T H P lummer
. . . . .

I N D UC i D D R A UG H T F OR F U RNA C ES .

an alternative to natural draught the use of a fan i n

AS ,

a duct forming a by p ass to the chim n ey is now common

FI G 1 4 5 — COM B I N E D F U R NA C E D R A U G H T
. . .

C o m b i n e d i nd u c e d an d n a tu r al c h i m n e y d r au g h t i n w h i c h t h e f u rn a c e
,

ou t l e t i s a t s o m e d i s t an c e ab o ve g r ou nd l e v e l D am pe rs all o w t h e
~
g a s e s to pas e i th e r d i r e c t l y u p t h e c h i m n e y o r b e d ra wn b y a f an .

enough but where the outlet fi n e is at some distance above

,

g ound level the a rangement is unsuitable b e cause the

r r ,

length of t h e requisite ducts becomes excessive and i t is ,

 T U RB I N E -
D R IVE N V E N T I LA T I N G FA N 27 1

undesirable to mount the fan at an elevation To meet .

these diff icul t ies Babcock and Wilcox and R Agar have .

patented the arrangement S hown i n fig 1 4 5 whic h are .

,

elevations at right angles to each other Th e outlet flue D .

from the furnace diverges as it meets the chimney and two ,

dampers E and F are placed here in branch ducts formed

by the outlet flue walls in conj unction with angularly dis
posed walls connected to partition A This partition A .

divides t he chimney into two longitudinal chambers B an d

C. Chamber B is connected by a short duc t G to t h e fan
H
( ) inlet the discharge from which passes t hrough duct K
to compartment C and thence up the chimney Wh n
,

. e

natural draught is desired damper F is closed and E opened ,

thus passing the gases directly up the chimney .

T U R B IN E
R I V E N V E NTILAT I N G F AN
-
D .

V arious combinations of fans and turbines have been

proposed in which a single vane wheel is used which draws -

air into a casing i n the direction of its axis and is itself

driven by an approximately t ngential stream of actuatinga

fluid Both flat vanes and curved vanes of the bucket type
.

have been used and it is usual to arrange the no zzle sup

,

plying the actuating fluid and the conduit through which

the fan delivers the stream of air in a s t r ai g h t li n e with
each other Th e association of the fan and tur b ine to form
.

a single unit makes the apparatu simple cheap and easily s , ,

porta b le Fig 1 4 6 illustrates a form with curved vanes

. . .

I t will be noticed that the buckets for the actuating fluid

are more numerous than t h e fan blades proper Th e .

buckets are bent at their tips li k e those of a l c lt o wheel ’

n .

Th e air sucked i nto the cas ing is mixed with the actuating
fluid and the two fluids enter the conduit A guide G is
, .

arranged in front of the no le to retain the jet i n its path

zz

toward the vanes Th e function of the jet is twofold ( 1 )

.

t o drive the wheel ; ( 2 ) to create suction in the condui t .

Th e P el t on bucke t s may extend across the entire width of

the wheel or across only part of it A n alternative pro
, .

posal which appears to be a good one fi xes them to a

, ,

separate ring member which is attach e d to t he wheel 3 the

,
272 C E NTRIF U G AL FA N S
latter may be a doubl e wheel consisti g of two parts side n

by side with the turbi ne ring between them Th e inventor .

,

however has missed the one th i ng which undoubtedly

would have made his i nvention an improvement on p r e ~
,

v i o us arrangements To have separated the turbine ring

.

of blades from the fan blades by a disc would have improved

the working considerably With P elton buckets arranged .

in the manner proposed the actuating fluid will to a e e r t ai

, ,
n

extent be d riven radially i nto the fan opposi g the out

, n

ward motion of the air I t will set up eddies i t h e fan

. n

wheel and generally dist urb the natural lines o f flow and
this will be very m rked with a guide vane such as G
a

directing the fluid partly in a radial direction Th e .

FI G . 1 46 . VE N T I LA T I N G
— FA N D RI V EN BY AS T R E A M OF F LU ID
I M PI N G I N G ON BU C K E I S
' ‘

AT THE P E R I P H E RY OF T H E FA N

efficiency would be much improved by attaching the

turbine blades to the outside Of the fan wh el disc so that e

the lines of flow from the fan would be undisturbed and

the suction eff ect of the actuating jet very materially
impro v ed Moreover it will be noticed that the inventor
.
, ,

w h o describes himself as a foreman writes of the vanes ,

“
being bent at their tips to form buckets like those of a
P elton wheel Th e buckets are n ot fixed o n the wheel i n
”
.

the manner usual i n P elton wheels h i t h e latter the .

strea m of actuating fl uid travels round the bucket i a n

plane more or less parallel to the axis of the wheel and ,

this arrangement would interfere less with the air streams

of the fan than the one proposed in which the stream is
directed i n a radial direction round the bucket and in w ards .
27 4 C E NTR I F U G AL FANS
owing to fail ure of some part of the system it becomes
necessary to pump gas of high tempera t ure through the
compressor This temperature may be sufficiently great to
.

m elt babbitt out of the S haft bearings thus n ecessitating a ,

shut down Often it is imperative to keep going at any

.

cos t so that t h e compressor requires t o be rugged in con

,

struction and capable of withstanding high temperatures .

l i repairs must be made the compressor should be of a ,

character to permit of them being easily and quickly made .

A gain the various condui t s carrying stea m or hot gases

,

expand unequally and sometimes impose great stresses on,

t h e casings of the machine To meet these di fficulties the .

General Electric C o of N e w Y ork has patented the form of

.
, ,

construction illustrated in fig 1 4 7 I t is a r ad i al vaned com . .

pressor dri ven by a steam turbine of the C urtis type all ,

completely enclosed in a casi ng Th e com pressor portion 1 .

O f the casing is divided i n a v e r t i cal plane in t o right and

left hand sections to facilitate casting and machining and
-
,

i n particular the surfacing of the annular walls of the gas

conveying passage immediately surroundi n g t h e impeller 2 .

Th e j oint between the casing parts is carefully machined ,

and after the machine i s completed will not or di naI i ly be

disturbed This portion Of the casing is also divided hori
.

z o t all
n
y in the plane of t h e shaft into upper and lower
s e ctions Th e various sections are fl anged and bolted
.

together Th e lower half Of th e casing ins t ead of being

.

supported on a foundation i s mounted d irec t ly on the gas

inlet conduit 3 and the gas discharge conduit 4 ( 11 L ) both ,

of which lie in a plane perpendicular to the shaft are i ,

n

cli ne d to each other and are capable of y ielding laterally ,

by a slight amoun t when for any reason the steam inlet

pipe o the steam exhaus t conduit 5 expands and contracts
r .

T h e conduits 3 4 5 should be long enough to permit

, , ,

these m ovements to take place without u nd uly straining

their supports or the j oints between them and the casing .

I mp ell r I some c ases the body of the impeller 6 is

e — n

m ade integral with the shaft in order to insure rigidity .

Th e vanes r e radial and the outer portions extend between

a

t h e side plates 7 which p roject i nto the a n nular discharge

,

ch mber 8 Gas is admitted to both sides of the impeller

a .
.
 C E NTRIF U G A L C OM P R E SS O R FO R GAS 27 5

I?

FI G . 147 .
— T ur . s o C O M P R ESS O R FOR GAS .
276 C E NTRIF U G A L FANS
B ear i n gs left hand side of the compressor portion
— Th e -

of the casing is provided with an openi n g of suitable Si e z

carefully finished and a shouldered and chambered head 9

secured to it Th e o ute member of the bearing is seated
. r

on a shoulder form ed i n the head P art 1 0 ( I V ) indicat e s . .

a Spacing d evice w hich engages a shoulder on the shaft

and is clamped by t h e nut 1 1 This serves to posi t ion th e .

rotating element Of the machine Access may be had to .

the nut by removing the cap 1 2 which is secured pressure ,

tight on the head Surrounding the spacer is an oil ring 1 3

.
,

which conveys Oi l from the chamber belo w to the spacer ,

from which it flows t hrough t h e downwardly inclined

passage 1 4 to the h ea r i n g T h e lubricant is main ta ined in
.

the h e ar i ng at a suitable level by the dam 15 the excess ,

being returned to th e Oi l chamber by the passage 1 6 .

Bet w een the chamber in the head and the impeller chamber
is a partition 1 7 and supported by this is a grooved m etal
,

packing l 7 of such character that it will not b e inj ured by

x

r latively high temperatures This packing does n ot

e .

norm ally make contact with the shaft so there is no wear ,

at this point I t acts to prevent hot gases from freely

.

entering the bearing ch amber w here it would soon destroy ,

th e lubricant I n a small chamber between t h e packing

.

and bearing is an Oi l thrower 1 8 of ordinary constructio n .

Th e right hand end Of the shaft is supported by a ball or

-

roller bearing 1 9 of the same character as the o n e on the

left I t is supported by a web 2 0 which forms an integral
.
,

part Of the turbine portion 2 1 of the casing This portion .

encloses the rotor 2 2 of the steam turbine and is seated

pressure tight on the compressor portion of the c asi n g
‘

-
.

T h e shaft does n o t extend through the outer walls of the

casing at either end so that there is no opportunity for gas
,

or steam to escape from t h e casing or for air to e ter I n ,

n .

order to prevent steam from the turbine portion entering

t h e compressor portion a packing 24 is provided composed
of a number of rings C arbo is satisfactory for this pur
. n

pose since i t will take a high polish as t h e shaft rotates

, ,

w ill not cut the shaf t and will stand relatively high tempera
tures This packing is car r ied by a housing 2 5 t hat has a
.

peripheral tongue 2 6 fitting pressure tight in a groove i n -

278 C E NTRIF U G AL FAN S
Th e blades are usu ally made radial i n order to avoid the
bending stresses due to curvature I n order to allo w the .

shaft to whip slightly without we akening the attach

”

ments special means have to be taken to insure a small

,

amount of free sliding action at certain points Th e various .

ro t ating parts mus t be accurately balanced .

A S t he air leaves the impellers at high velocity it is

necessary to gradually convert the kinetic energy to
pressure energy as first suggested by P rofessor Osborne
—

R eynolds b y passing the air through graduall y expanding

—

passages V anes that are radial at outlet are not quite as

mechanically e ffi cie n t as backward turned vanes and this -
,

supplies a f urt h er reason for the use of d i vergent passages

at outlet Radial vanes which are c ontinued straight t o
.

the inlet would be very inefficient as it is necessary to ,

c urve them at the inlet to suit the relative velocity of the

air to the vane at this point S ome makers divide the .

vane into t w o parts making the larger external part

,

straight and radial a for m that can be readily machined

—

—
and attaching to this a c ur v ed but radial por t io n at the
inlet which being small can also be readily machined
, , , .

A n example of recent construction i n h i g h ~ pr e ss ur c com

pressors is illustrated i n figs 1 48 I 4 SA 1 V l1 Referring .
, ,
.
-
.

to 1 and. the air suction is shown at 2 af t er passing over

the first two impellers 1 3 and 1 4 the air is discharged
through the divergent passages 1 1 to the space 1 7 ,

where it divides part passing along channel 1 8 to the

,

right eye of the next wheel and the remainder along ,

passage 1 9 to the left eye of the same wheel .

Constr uct i on of Casi n g Th e casing is divided into an

.
—

upper and lower part bolted toge t her i n an axial plane

,
.

Th e head 1 is integral with the i nlet c nduit 2 C losely o .

surrounding t h e shaft is an annular member 3 that forms a

continuation of the inlet co nduit T his member is divided .

i nto upper and lower parts and permits the removal of the,

head 1 without disturbing the shaf t coupling which is

e closed in the guard or housing 4 Th e dia m eter of the
n .

bore of the head is greater than the diameter of the coupling .

Th e g uard is made of S heet metal formed i nto a cyli n der

a n d fastened by strips 5 Wi t hi n the casing are dia
.
 H I G H P RE SS UR E
-
CE N TR I F UG AL C OM P RES SOR 27 9

h r a m s 6 seated i n suitable shoulders 7 formed on t h e

p g
inside O f the casing Th e diaph ragms are provided with .

shoulders 8 to rec ive the plates 9 and 1 0, between which e

are secured t h e discharge vanes 1 1 .

Cons tr uct ion of I mp elle r s — T h e impeller is provided wit h .

vanes 1 3 and 1 4 which a e strictly radial and i ntegral r

with the body 1 2 T he impeller is divided i nto t wo part . s

( see which are fi nished on their O pposi t e faces and

shouldered ( V I ) to insure proper fitting 1n assembling

. .

these parts the female member of the impeller is expan ded

,

by heating and then allowed t o shrink on t h e shoulder

proj e c tion of the other member T h e parts a e then united . r

by set screws I n order to direct t h e air into the channels

.

or passages between the impeller vanes at t h e proper angle ,

curved inle t o guide vanes 15 and 1 6 are provided which

r ,

rotate with t h e impeller These vanes ar e carried by a .

cast metal support 2 1 which is slipped over the hub

po rtion of the impeller body Each of these vanes is pro .

v i d e d with a pair of lugs 2 2 which engage oppo ite sides ,

s

of an impeller ( 11 and and prevent the outer ends from

.

moving out of their proper position These small i nlet .

vanes t h ough cu ved are strictly radial so th at there is

,
r , ,

no tendency to fold over under centrifugal action .

Mel/ d of F i xi g I mp ller t Swll Th e su ppo rt 2 1

ro n e o i —

carrying the inlet vanes is slipped over the hub 2 3 of t h e

im pell r and is secured in place by a locki n g ring 2 4
e ,

T his ring is p r ovided with an i nte rnal shoulder 2 5 that

is seated i n a groove formed i n the hub Th e ring is .

shrunk on t h e shoulder not being deeper than 7 19m; in or .

0
1
T6 7 ;
in F ormed in the bore of the impeller are grooves
.

26 each containing a S plit ring preferably Of hard ,

non ru t ing metal of high crushing strength T h e grooves

-
s .

are slightly dovetailed and the rings are rolled into the
grooves under high pressure w hich forces the metal under
the overhanging walls of the groove A fter t he rings are .

forced i n place they e machine d to a diameter about

, ar

00 1 2 i n smaller than the diameter of the shaft

. The .

impellers are then forced end wise o n the S haft und r e

pressure and prevent d fro m rotating b y the key 2 7 and e

rings i n 26 By t hese means t h e m t al of the im peller is

. e
CE N TR I F UG AL F AN S
o
$w
:
6 m

m
8 0
m 6
2
3 %
8
> h
fi

m
.
qm fi
o

O
m
E
w S
fi S
m A
m
z
.

m
o
o H o
a

Q
£
g
<

~
G 8 m o

D 8 u
M s
H
M 3g o
H 5 fi
Z
H
8c
O fi m
:
a
e
s
H c
m
b
e
w o 5
m s
?
m é
w fi s
fl m
fi b e
m c o
a
o
é n
o

:m3
e
fi o
e
g
m fi g
m w
a o
n
0

3
.

fi
»m
m ?
v s 9
fi
:
.

a
8 m
O a
H
? e

m
p
o
o

5 m
5 m
2 o >
H

:
%

5
3

$
s m
.

8
m
:
m
28 2 C ENTRIF U G AL FANS
put under radial compression which varies from t h e centre ,

outwards the idea being that when t h e centrifugal tensile

,

stresses are induced the resultant stress w ill be so smal l

that very little increase i n diameter ensues Mounted on .

the shaft are sleeves 2 8 made of some non rusting metal — -

—
which fit inside the ends of the hubs o f the impeller
Th e bore of the hub is 00 02 in i n diameter less than the .

external diameter of the sleeve and so forms a compres ,

sion fit Surr unding the end Of the hub is the shrunk

. o

ring 24 which in addition to preventing the inlet

,

vanes from moving axially serv s to hold down the , e

stretching eff ect of the end of the hub a 7 6 0 i n to — 2 —

0
.

in fit is found be t for this I n o r der t o permit a

. s .

litt le axial movement between the impeller and its S haft ,

and so avoid the impactive stresses which would otherwise

come on the impeller when the shaft w hipped it is n e c e s ,

sary to have approximately a sliding fit between the

impeller and shaft To this end the hub is slotted at
.

29 so that i n eff ect a ring 2 9 is formed whic h is

, ,
X
.

connected to the h ub proper By reason of the slot 2 9 .

( V 11 ) the ring 2 9 is relieved of ta ngential stresses s o

.
X
,

that its dia m eter will not be perceptibly increased by

centrifugal forces i n t he hub but will retain as nearly ,

as possi b le its original d am e t cr and thereby form a

'

support with a constant degree of tightness on the S haft .

l h e discharge vanes
'
Cons l ucli on of D i char g e Van e
‘
-
r s s .
—

1 1 ( 1 and. are made of heavy sheet stock and are

U shaped in cross section Th e down turned Sides of each
-
.
-

vane are united b y bolts screws or rivets to the ide , ,

s

plates 9 and 1 0 Th e side portions of one vane follow

.

at their edges the Shape of the ad jacent vane and ,

preferably rest on that vane so as to form good passages

and also gi ve support Th e vanes are stiff ened at their .

inlet ends by rivets 2 0 Th e spaces 8 between the walls .

of the impeller may b e used for circulating cooling w ater .

P ar ki ng I n order to space the impellers on the shaft

—
,

cylindrical members 30 are provided t hat have internal

‘

shoulders which e n gage the sleeves 2 8 ( V ) and so prevent .

axial movement Th e periphery of the cylindrical member

.

3 0 i provided wi t h proj ections 3 1 that extend between

s
 C ENTRIF U G AL A I R COM P R E SSORS 28 3

corresponding proj ections on cylinder 3 2 bolted to the —

diaphragm and thus form a labyrinth packing to prevent

—

the passage of air from one chamber t o another along the

shaft .

B la ci ng Although each of t h e impellers is carefully

a n —

balanced either statically 0 r dynamically it sometimes ,

happens that when they are all assembled on the shaf t

the machine is out of balance as a whole and in order to ,

pe r mit of balancing without taking o ff the top part of the

casing Openings filled by removabl e plugs as 3 4
, and
33 a e provided in the ends of the casing
r By taking .

out these plugs access can be had to the body of the,

impeller and metal added or taken away as desired T his .

me t hod of construction is patented by the General Electric

C o Of N e w Y ork
.
, .

CE NTRIF UG AL A I R C OM P R ES S O RS .

S ur g i ng Ej surging or pulsating of the air i n a

e cta — Th e
centrifugal air compressor may be due to two quite distinct
causes Th e cause which is more easy to comprehend and
.

to guard against is that of working agai nst a fl uctuating

head such as a blast furnace i n which the resistance
, .

occasioned by the charges of fuel ore and flux v aries con ,

.

s id e r a b ly from time to time I n such cases when the flo w .

,

is r tarded the air is liable to accumulate i n t h e delivery

e ,

conduit of the compressor until its pressure rises so high

a s t o cause a backward rush i n the conduit to be succeeded ,

by a forward rush to restore the lost pressure The surging .

sometimes becomes so violent as to endanger the apparatus .

I n order to reduce these su ging e fle c t s the N e w Y ork r

'

General Electric C o 7 place a sheet meta l no le A ( fig 1 4 9 . zz .

,

I ) i n the condui t preferably the intake pipe of the co m

.
,

pressor This permits the air to flow freely i nto t h e

.

machine but any reverse flow is throttled at the no le

,
zz .

T h e smalle end of t h e no le points toward the com

r zz

presso and its diameter i f r om one third to one quarter

r , s - -

that of the cond uit I the flange B are several ports . n

whi h are controlled by an annular damp r D having ports

c e

whic h are adap t ed to be brought i or out of regi stration n

28 4 C E NTRIF U G AL FANS

FI G 1 4 9 —S U R G I N G I N AI R C O M P R ESSO R S
. . .

Sh e e t m e t al n o zzl e i n i n t ak e t o pr e v e n t s u r g i n g D i a g r am o f
We ath e rc o c k g u id e v an e s
.

pr e ss u r e a n d v o l u m e r e la t i o n
s e lf ~
.

adj u s li n g to t h e d i r e c t i o n o f fl o w
.

with the ports in the flan ge These ports serve to regulate

.

the flo w somewhat .

Th e second cause is neither so easy to understand nor to

2 86 C E NTRIF U G AL F A NS
will cut the curve i two points 0 and D corresponding to
n

two quite diff erent flows L E and L H respectively Whe n .

the compressor is working therefore at a certain speed , ,

against this particular pressure the flow of fluid through

it may be either L E or L H Th e flow i s ambiguous and .
,

consequently unstable I t may change from one to the

.

other at its own pleasure and this fi c kle n e ss leads to

,

surgings or pulsations when working on this part of the

load I n seeking for a remedy it is best to study the
.

reason of this u p and down variation of the flow At a

- -
.

certain speed of running the blade of a compressor or

turbine pump gives a definite amount of energy to the
working fluid partly as pressure and partly as kinetic
,

energy N o w the ki netic e nergy to be of use mus t be

.
,

largely conv e rted i nto pressure e ergy by passing through n

the diverging discharge vanes Th e vanes are set at an .

a n gle to suit a given direction of discharge from the

impelle r that is to suit a given flow so that if the flow
—
, , ,

is not suitable a certai n amount of kinetic energy is lost in

,

ed d ies T he fl uid molecules are thrown hither and thi t her

.

and career around each other in little circular paths This .

frittering away of energy may be as serious at too low as

at too high a discharge hence the shape of the pressure
,

volume curve Obviously the correct remedy would be to

.

vary either t h e outlet angle of the impeller or the inlet

angle of the discharge vanes t o suit each flow Th e .

m aximum pressure would then occur at or near the z ero

discharge line ; it would probably be a little distance to
the left unless the inlet angle of the impeller w e re made
to suit a low discharge H C Hom e h am has pro
. . . r s

posed float i n g discharge vanes which automatically adjust

2 5

themselves to the direction of flow as a weathercock takes

the direction of the wind I n the illustration ( fig 1 4 9
. .
,

the straight guide bla des A are mounted by means of

a pi n B i n an annulus C which annulus is m ounted i n the
,

frame A bout this annulus a floating ring or annulus D is

.

mounted which bears by means of antifriction devices in

,

the framework T his floating annulus carries small

.

pivotal pins E having saw cuts i n their heads into which ,

the edges of the guide blades enter Th e floating annulus .

 C E N T RI FU G AL A I R C OMP R ES SORS 287

D is free t o move i n a circumferential direction A special .

device is a r ranged S O that the attendant may hold the

guide vanes i n any desired position or may assure himself
of t h e freedom of the floating ring .

A method proposed by the General Electric C o of

26
.
,

N e w Y ork to abolish this instability is to throttle the flo w

,

of air preferably at the intake to t h e compressor Th e .

instability i s due to the upward trend of the pressure curve

on ligh t loads and the idea is to i ntroduce such resistance
,

— and therefore le s s of pressure at the inlet that the

, ,
—

increase in pressure due t o the better conditions i n the

compressor on increased flow is slightly more tha n neutral
ised by this lo ss Th e curve then takes a down ward
.

direction as indicated by the dotted line A B ( fig 1 4 9

,
.
,

A t no pressure under these conditions are there two possible

flows and therefore there is no instability I t is plain
, .
,

however that the remedy is a somewhat wasteful one

, .

Th e method of throttling the intake is shown in fig 1 4 9 A 1V .

, .

Th e i ntake A of the compressor B is provided with a throttle

valve 0 mounted on a stem which has at its lower end a
, ,

dashpot Th e valve is a disc located i n an inverted frusto

.

conical section of the intake so that as it rises and falls it

,

enlarges and diminishes the annular opening between its

periphery and the wall of the section The stem passes up .

throug h the top of the elbow in the intake pipe and carries
a cam I ) whose back is parallel with the ste m and bears
,

against light guide rollers A lever having two arms at

.

right angles is pi voted to the upper end Of the standard .

T h e vertical arm carries a roller bearing on the cam A

weight Wis ad jus t able along the hori z ontal arm With
.

an i ncrease i n volume Of air passing throu g h the intake the

valve rises and the cam tilts the l ver throwing the arm e ,

wi t h the loa d upward and thereby shortening the eff ec t ive

,

distance of the weight IV from the pivot I t is hoped by .

some uch arrang ment to be able to produce j ust the

s e

amount of throttling necessary for any given flo w Th e .

loss of pressure necessary varies from ero at the outse t to z

a maximum an d then reduces again to z ero The vertical

. .

part of t h e acting face of the cam allows the valve to lift

against prac tically no load at the commencement and the ,
28 8 C E NTRIF UGA L F AN S
throttli n g action i s therefore small With increased flo w .

the curved part of the cam face comes into action gradually ,

increasing the resistance to the motion of the valve u nder

the bombardment of the inrushing ai r molecules Th e .

throttling action is thus increased for it must be remem ,

«

bered that the increased flow requires a larger val ve open

ing and t h e loss of pressure is proportional to the square
,

of the increased velocity due to any d eficiency i n valve

opening As the flow increases further the gradually
.

diminishing leverage of t h e weight 147 once again reduces

t he resistance to t h e upward motion of the valve .

T h e exact S hape of the cam face is however a matter , ,

for experimental determination I t cannot be obtained .

theoreti cally as suggested by t h e patentees i n their s pe ci fi

,

cation without muc h experimental data on t h e resistance

,

and throttling actio of such val ves I t is Obvious for

n .
,

instance tha t if the amou n t of throttling necessary for

,

point 0 (fig 1 4 9. has been determined and applied t h e

, ,

air then supplied to the comp essor i s in a more or less

r

disturbed state a d the charac t eristic pressure volume curve

n -

for air i n this condition

" wo u ld b e somewhat d i fle r e n t to
'

the curve A 0 I So that for any small increase i n the

'
.

flo w 8 o the i ncrease in press ure 6 p might be quite diff erent

to that ob t ained from a tangent to the curve A C F ’
.

C OM P R E SS O RS WI T H I N C L I N i D STRA I G H T B LAD E S .

Straight blades for compressors are cheaper and easier

to manufacture than cur ved ones but radial straight blades ,

cannot make the proper angle at inlet I t is necessary to .

incline th e blades in order to secure t h e proper inlet angle .

Straight blades have already been used in compressors but ,

i n this case the faces of the rotor destined to receive the

blades had to be straight C urved rotor faces however
.
, ,

have the advantage over straight faces of forming a better

guide for the air and of enabling a more favourable shape
to be given to the rotor to counteract strains and more ,

particularly to avoid the bending strains which are liable

to occur i n rotors with one sided ad m ission By t h e rota
-
.

tion of a straight line about an axis without changing its ,

29 0 CENT RI FU G AL F AN S
the wheel shown in plan by k a D raw A B fig 1 5 1
— .
— .
,

I V — equal i n length t o k a— fi g 1 5 1 111

. t o S how the .
,
.
-

thickness o f the roto r Make A L and B S perpendicular .

to A B fig 1 5 1 I V
— — equal respectively to k a and k b1
.
,
'
.

— fi
g 1 5 1. then L and
, S are t wo points o n the
rotor profile To obtain intermediate points m ark the
.
,

positions of any three points 1 2 3 along l a— fig 15 1 , , ,

c .
,

FI G . 151 . M E T H O D OF D R A WI N G T H E HY P ERROI D A L FA C E
— OF T H E
R OTO R OF A C O M P R ESS O R WI TH S T R A I G H T B LA DES .

III draw hori zontal lines 10 2d 3 6 to meet a b i n

— an d
, , ,

the points 0 d 6 respectively D raw vertical lines

, ,
dd .
‘
,

ee
l
t o meet a1 b1 i n the po ints d 6 ‘ respectively In 1
, .

fig 1 5 1 I V mak e C 0 D D E E B equal respectively

.
, .
, , ,

t o a l 1 2 2 3 3 k o f fig 1 5 1 III
, , D raw vertical lines
, ,
.
, .

through these points and make 0 M D N E R equal ,

'
, ,

respecti vely t o k cl lo d1 and k cl of fig 1 5 1 111 Then

, , .
,
.
 C OMP RESSORS WI TH I N C LI N ED I
STRA GHT B LA D ES 29 1

t he cur v e S R N M L through these points is the o ne

required N ote that the lengths k a1 k cl k
. k 6 1 I: b1 , , ,

a r e the true lengths o f lines drawn through the several

points of the generating line and perpendi cular t o the a x is ,

o f revolution ; consequently they represent the radii o f ,

t h e curve at these points .

Th e curve P Q for the rear face o f the rotor may be

o btained in an e x actly similar manner L e t the a x ial .

width o f bl ad e Q S— fi g 1 5 1 I V at outlet be less than

. , .
—

the W idth P L at inlet Make 113 o r b g— fi g 1 5 1 III

. .
, .

i n plan equal to Q S fi g 1 5 1 I V and p a fi g 1 5 1

— .
, .
— — .
,

III equal to V A — fig 1 5 1 I V
.
— A ssuming that the
.
, .

inlet and outlet circles of revolution are equal to those o f

t h e front face and that the planes o f the bl des are parallel
,
a

t o the a x is of revolution then d l b is the elevation of p q as

1
,

well as a b— fig 1 5 1 I II n d the shape is obtained by

.
, .
— a

proj ection from a1 b1 as in the last case Th e ver t ical lines .

f rom the intermediate points — fi g 1 5 1 I I l lie a little t o .

, .
-

the right o f the verticals cc dd l I t W ill be noticed ‘

, ,

t hat inclined straight blades cut the inner and outer circles
i n such a manner as to make the angle at inlet Y fig 1 50 — .
,

II between the blade and the r ad ius greater than the

.
—

sa me angle at outlet Fo r compressors running at high

.

speeds it is nece ss ary that this angle at inlet should be

large but it is not at all necessary that the outlet angle
,

should be large in fact in many cases the bl ade is m ade

,

r adial at outlet I t has already been demonstrated that

.

in fans and compressors the curving back o f the blade at

o utlet regarded as so essential t o effi ciency i n centrifugal
,

pumps secures only a trivial increase in mechanical effi ciency

,

a t the e x pense o f a real loss o f manometric effi ciency .

2 92 CE N TRI FU GAL FA N S

N OTE S

1 P h i l Ill ag . x v ii p 3 89 , v ol, ,

by W M Wall ace p 1 9 4
. . . .

Hyd r a u l i cs
”
2
.
, .
, . .

3
J ou r n A m er S oc E n g Se p t 1 9 1 2 p 1 3 5 4
.

P hys i cs :
. . .
,
. . .
, ,
4 E x p e rim e n t al an d Th e o r e t i cal by R H J ud e
“ ”

, . .

5 Se e p ap e r on C e n t rifu al F an s
.

g b y H e e n an an d Gi lb e r t i n t h e
P r o c I n s t i t of C i v i l E n g in e e r s v ol c x x ii fi g 8 p l at e 5
, ,

. .
, . . . .
, .

5
Se e fi g 4 1 fur th e r o n . .

7 B ri t i s h P a t e n t 1 9 1 3
,
3
P r oce e d in g s o f I n sti tu ti o n o f C i v i l En g in e e r s v ol c x x iii p art 1 ,
. .
,
.

9
Th e tabl e g i v in g th e se e x p e rim e n t s w i ll b e fou n d i n A pp e n d i x I I .

o f H e e n an an d Gil b e r t s p ap e r
’
.

10
P r oc e e d in g s o f t h e I n s t i t u t i on of C i v i l E n g in ee r s v ol c x x ii ,
. .

11
L e s V e n t i l at e ur s d e Mi n e s R e vue Un i ver selle d es M i ne s
“ ”

, ,
v ol x x
. 1 89 2
.
,
.

1 2 F r om
Th e Th e or y o f t h e C e n t rifu g al P um p an d Fan N EC
”
. .

I n st o f En g in e e r s an d Sh ipb u i ld e r s v ol x i v
, .

. .

P r o ce e d in g s o f t h e I n s t i t u t i o n of C i v i l E n g in e e r s v ol xl v ii
.
,
13
,

P r ocee d in g s o f t h e I n st i t u t i o n o f C i v i l E n g in e e r s v ol l iii
. .

14
. .

H y d r au l i c Mo to r s T u r b in e s e t c
,

15
B od m e r s
’ ”

P r oc I n s t Min in g E n g in e e r s v ol x 1
, .

13 . .
,

D u C al c u l d e s V e n t i l ate ur s by M L el on g I n g éni e ur d es
. .

”
17

C o n s t r u c t i o n s N a v al e s
.
, ,

1 8 F ro m D i e Ge bl ase b y v on I h e r i n g ”
.
,
1 9 A n n al e s
d e s P o n t e s e t C h au sée s 1 888 .

Mi n ute s o f t h e P r ocee d i n g s of t h e I n s t i t ut e o f
,
2 0 Se e v ol x l iii . .
,

Mini n g E n g i n e e r s .

Bull e tin d e la Soc i ét é d e l I n d ust r i e Mi n ér a l e

'
21 .
”

Se e Rapp or t au C on g r es I n te rn ati o n al d e Mécan i q ue A pp l i q ue

’
22
.

2‘
I t i s ob v i o u s t hat t h e ce n t r e d oes l e ss w o r k per p oun d t h an t h e
o u te r p art s i n ce a v + g i s t h e w or k p e r p o u n d d o n e b y t h e f an .

Hy d r au l i cs b y W M Wall ace p 1 3 4
,

"
24 ”
.
, . , . .

25
B ri ti s h P ate n t 1 9 1 3 ,
2 Br i ti sh P at e n t 1 9 1 3 .
2 94 I N DEX
Bl ow in g Mac hi n e s ts 203 R oo
’
Ch a g e n
i n V e l oc ity with out C h an g e i n
B D e str oy e rs Fan
, ,
D irec tion Lo s s f H ead d ue t o 9
.

oa , t pe d o
'‘
l or o
, , f or , , o ,

of M m e n t o f Mom e n tu m
.

( fi g 1 4 1 ) 265 . . o L aw of ,
B o d i n e r , 29 2 16
B r k Dyn m t ri ( fi g 1 1 3 ) 226
. .

a e, am o e c, C h arac te risti c C u rv e f or Fan w ith

B ri x D r 7
. .

ap e r i n g S id e s ( fi
g5
‘

g 27 )
B r n W l t n 1 37 ,
. . .
, .

ow a o
B y n D n kin S u n d r D nk in
.
, , .

i a ee e o wi th 1 2 l n C e n t r e D i a
B u k ts t h P ri ph ry f t h I
o . .
-
.

c e a t e e e o e “an m e t e r ( fi g 28 ) 7 5
C u rv e s :
,

V n ti l ti n g F d riv n b y a S t r m
, . .

e a an e ea Mi n e Fan s ( fi g 30 an d s.
f Fl uid i m p i n g i n g
,

o ( fi g 1 46) 2 7 2 on , 3 1 ) 7 8 , ( fi g s 3 2 an d 33 ) 7 9 , ( fig s
B u m pstead 1 58 159
. . .
.

, ,
. 3 4 an d 3 5 ) 80, ( fi g 36) 8 1 . .

ofF s t ste d by M L l n g ( fi g
an e e o
C l u l ti n of A i D e n sity f r G ive n
.
, .

a c a o r o a 63 ) 1 4 6
H u m idi ty D t rm i n tio n
.

f H d , e e a o ea f M to r S pe d ( fi g 1 1 7 ) 23 3
o o e
fr m t h W t r G u g a d t h e 1 3
, , .

o f Tw o E y d F
.

o e a e a e, n , .
(fig -
1 48 e an ,
C pe ll Fa
.

o f F n s Th r tical 33
.

a 1 03 1 07 ( fi g s 5 7 n d 58), , , . an ra t
a i ti
c er s cs a eo e
1 08 1 09 ( T b l 1 2) 1 1 2 ( T bl e 1 6) 1 1 6
,
t st d b y M L e l n g ( fi g 6 8)
.
,

, , a e ,
a . e e . o ,

t lli A n m m t r s 1 21
.

C asa r e e o e e , . 1 49 .

C s ll A n m m t r s 1 04
a e a e o e e Ch t l in M 1 98
a e a
u n d e r D u C h yl
.
, , .
, .

C si n g C n st r u ti n f 2 7 8
a o c o o Ch y la S a ee a a.
C i rc u l ti n g w it h ut D is
. .
,

f L arg M r ti r F suit
,

o t e o e an o a C n t r c te d o c, o a
Min O r i fi P rm n n t A l t r
e ce , e a e e a O utl t ( fi g 1 27 ) 240
e ,
,

C l usius C y l 2 02
.

ti n ( fi g 7 3 ) 1 5 2 o . . a c e,
C ll g
, .

f M r ti r F
o t suit t h e Min e
o e an o o e R uth rf rd 1 27
e, e o
C l l i r y C ss ll 1 60 ; P r k n d 7 6 1 81
, , .

M vin g f ( fi g 7 2) 15 1
o o , .
'
. o e ,
o a , a e
1 8 2 ; S g hi ll 1 9 1
, ,
C ss l G r m n y 2 4 4
,
a e e a e
,

D r ug ht ( fi g 1 45 )
.
, , .

F n s 16 1 d F u rn
2m
a o bi ne ac e a
C n t r D i m t r C h r t r isti c C u r v e
.
, , .

e e a e e a ac e 0
Pr ussi n
.
,

f F
or w ith 1 2 I an ( fi g 2 8) 7 5 -
n., . . C m m issi n T sts
o o e a
C e n tr i f u ga l A i C m p r ss or s 2 83 r o e Mi n i n g 103
C m p r ss r H ig h P res su r B elg i n
, .
, .

o e 277 o ,
-
e, ,
E xp ri m n ts
a e e on

(fig 1 4 8) 280, .
(fig . 1 4 8a ) 281 . Mi n F n s 1 05 e a , .

f or Gas , 27 3 C om pa ris n b tw n Th r y o d E e ee eo an x
t ppe d B l ad e s f or a w ith
.

F an , S e , p ri m n t 1 7 7
e e , .

R i b s f or sti f f e n i n g th e m at f V lu s f H f
o a Ra te u F
e o or a an ,
t h e J o i n t s ( fi g 1 4 0) 2 63 ( T bl a e
H i g h Pr ss r e C t i f u
.
, .

T w S tag e ( fi g 1 4 2) 2 66 o- mp r e ss o r - e u en r
V ari a tio n f P re ssu r i n a 5 7
o
,
. .
,

o e , . g al, 27 7 , ( fi g 1 48 ) 28 0, ( fi g 1 4 8a ) . .

F an s 1 28 1
b y B ryan D on k in E x p e r i
.

n t i fu l
.
,

,
. f or Gas , C e r g a , 273 .

m e n t s on 83 T u rb o , ( fi g 14 7 ) 275
ith I n l i n d t r i ht B l d s
. .
, .

C n st ru c ti on o f 2 66
o w c e S a g a e , 2 88
it h tr i ht B l d s
.
, .

S t p p e d V an e s f or 262
e w S a g a e , Me th d o of
Fl uid P u m p s C n st r u c tio n of 261
, .

o d r in
aw F
g t h e Hyp e r b o i d al ace of
P u ri fi r 2 6 7 , .
,

G as e t r
t h e R o o o f a , ( fi g 1 5 1 ) 290 . .

p r ss rs n tri fu l
.
,

o s i sti n g o f a Fan an d ac n Co m e o , C e g a A i r , 2 83
n t Pr ti
.

Tu rb i n e S e ri e s ( fi g 1 4 3 ) , . F ns
a an d , R e ce ac ce , 2 5 8 .

268 S g gu r in i n Ai r , ( fi g 1 4 9 ) 284 , ( fi g
P u m p B l n e d ( fi g 1 7 ) 60
. . .

a a c 1 4 9 a ) 23 5
xi u G in Pr s
. .
,

n diti n
.

E xpe ri m e n ts b y C h rl e s H
,

a Co o f o r Ma m m a of e
su r H d 8
.

I nn e s 1 26 e ea
b y R 0 P rs n s 1 3 3
.
, .
,

a o f orM i n i m u m L ss f H d d t o o ea ue o
,
Dir
.
,

S udd n Ch n g i n th
.

( Tabl e 26) 1 34 . a e a e e ce o
T r ial s b y C h arl e s H I nn e s .
, ti n f M ti n 7
o o o o , .

1 89 7 , ( Tabl e 24) 1 28 1 898, C on g é r I n t rn ti n l d Mé n iq u

s e a o a e ca e

( Tabl e 2 5) 1 3 0 . A p p li q u e 2 9 2 , .

C h nd le r
a 15 8 , 1 5 9 Co n s r v ti n f E n e rg y Th
e a o 1 o e,
Dir
.
. ,
I n m p r ssi bl F l uid i n
,

C h an g e i n tio n L ss o f H e ad ec o to an co e e
M ti n A ppl i ti n of t h e
,

d u e t o C han g e i n V e l oc ity w ith o o , ca o

ou t 9 L w of 2 a
in t h e D ir ec tio n o f Mo tion C o n di
, .

Mi n s W e st h l i 1 26
.
,

on s o li d t i a on e , a a, .

t V l um
,

ti ns f r Mi n i m u m L o ss o f H e ad
o o t
on s a n R eas n or t re ti ng e, o o a

d u e t o a Sudd e n 7 , . A i as i f it h d 3
r a .
 I ND EX 2 95

Co ns tr u c ti o n M e th ods of 277 D i fl user an d V ol ute Th eo of th e 5 6

Ra te au Fa n est wi th
.
, , . , ,

o‘Casin 2 7 8 o nl
of Ce n t r fiug al Fan s 266
,

abl e 1 8) 1 1 8
.

of Cen tr i fu gal m y! Pu m ps 261 Dim ensi ons C ost an d Ser i ce of Capell

. .

v
of Dischar ge V anes 282 F an s ( Table 1 6) 1 1 6
, .
, , ,

of I m pe llers 27 9 of Gui b al Fans ( Tabl e 1 4 )

, . , .

of V anes G eo m e tr i cal ( fi g 1 5 ) 51
, .
,
1 14
Co n tr ac t ed I nle t E xh austing wi th of Ra te a u Fans ( Tabl e 17 )
.
, , . .

, , ,
D isc ( fig 1 26) 240 1 17
Ou tl e t Cir cu la tin g Wi th o ut Di sc of Se Fans ( T abl e 1 5)
.
, . .

, ,
r ,
,

( fig 1 27 ) 240 1 15
Pr opell ing wi th Disc ( fig Di sc Co n t ra c ted I nle t E xhaus ti ng
. . .

wi th ( fi g 1 26) 240
, , .
, , ,

1 25 ) 2 40 , . .

Ou tle t Ci rc u la ti n g
.

C ossall C o ll i ery 1 60 , . , ,

Co st an d Se rvi ce of ( Dim e n sion s) ( fi g 1 27 ) 240

Pro pe llin g W th ( fig 1 25)
. .

Ca pell Fa ns ( T able 1 6) 1 1 6 , .
,
i , .

Gui b al F an s ( Ta ble l 4 ) 1 1
Ra tea u F ans ( T abl e l 7 ) 1 1 F ri c ti on Wo r k e x pen d e d in 1 7 1
,

Se r F an s ( T abl e 1 5 ) 1 1 5 Discharge V anes C on s t ru c ti o n of 282

, , , .

C r o s by I n di ca to rs 1 2 1
. , , .

Di rec tion L oss of H ead du e t o Chan ge

,

i n V lod t y w ith o u t C han g e i n 9

, . ,

C ur tis 27 7 e
Cu rva t ures of B la d e s in F an Figs 1 1 9 Do n kin B ryan 4 0 83 86 88 89 9 3
, . , .

, .
, , , , , , , ,

an d 1 20 ( fig 1 24) 238 1 83
Cu rv e fo r F an w i th Tape r ing S id es B est E x pe rim en t o n E ach of
.
, . .

’
D on ki ns
th e E le en T y pes of F an ( Tabl e 8)
,

( fig 27 ) 7 5 . . v ,

wi th l 2 I n Cen tr e Diam et e r -
91
B r y an E x peri m en ts ( Table 3 7 )
.
.
,

( fig 28) 7 5 . .
, ,
Cu r v e s ( fig 1 03 ) 1 9 6 1 7 8 ( T able s 38 an d 3 9 ) 1 7 9
fo r M ine Fan s Ch ar acte ris ti c ( fi gs
, .
, ,

( T able 4 0) 1 80 ( Ta bl e 4 1 ) 18 1
.

A ppara
.
, , ,

30 an d 3 1 ) 7 8 ( fi gs 3 2 an d 33 ) 7 9 . P r ess u m ur i n g r e o
cas

( fi gs 3 4 an d 35 ) 80 ( fig 36) 81
,

tus ( fi g 46 ) 89
fo r Ra teau F an s Ch ar acter isti c E x pe ri m en ts F an N o 1 ( Ta bl e 7 )
. . . .
, , .

.
, , , ,

( fig . 1 9) 64 . 90 .

Fo rm u l e u s e d 1 02 a
Th e E le ve n T y pe s of Fan use d
, .

Da l ton s L w 1 3
'

a , .

D e L l Tu b i n 27 3
a va r e, i 94 n,
E xpe ri m en tal Appa ra tus 88
. .

D eli ve ry Ai Tu be ( fi gs 1 1 4 an d 1 15)
r , .
, .

226 T es ts E ffi c i en ci e s o f Fan s ( fi gs 4 8
Tube Fan ( fig 2 1) 6 7
. .
, ,

an d 4 9 ) 1 00 ( fig 5 0) 101
D e si gn I m p rove m en ts in Fan 258 F an S ee u n d er Pelze r
, , . .
, . .

or tm u d n
o f M i x ed Fl ow F an s 4 6
, , . .

Do rt m u n d
of Rad ial F l o w F an s 40 D o u bl e I nl e t For c ed Dr au gh t Fan
, . .

D e si gned E le c tri c M o to r Fan wi th

, .
,

( figs 86 an d 87 ) 1 64
S pec i ally ( figs 1 22 an d 1 23 ) 238
.
, .

D ra ug ht Co m b i ned Fu rna ce ( fi g 1 45)

D es t roy er T o r p ed o B oa t Fan f or 265
, . .
, , .

-
27 0
D e te rm i na ti on of H e d fr om th e Wa t e r F an Do u ble I nle t Forced ( figs 86
.
, , , .

a
G a u ge an d th e C al c u la ti on f Ai
.
, ,
o r an d
D ensi ty fo r G i v en H u m idity 1 3 for F u rnac es I n du ced 27 0
,

9.
l m en t an d T an s m i si o n D ra w ing t h Hype b o i da l Face of th e
. .
,
of
, ,
D eve o r s e r
ow er 5 Ro to r of a Co m with S t aig ht r e s so r r
,

B lad es M e th o f ( fig 1 5 1 ) 290
.

of a S c re w ( fig 1 1 ) 46
Di agram s E ffi c i en cy G au ge an d Pow er
, . .
, , . .

, , ri v e A rra gem en t of El ec t ri c ( fig s
, n , .

(fi 2 3 an d 2 4 ) 7 2 ( fig 25 ) 7 3 , . . 1 11 an d 1 1 2) 226
{ or pe ller F an V el oc ity ( fig
.

, , .
“
D u C al c ul d e s V en t ila te u rs by M ,
.

L e l o ng 2 9 2
1 09 ) 204
Dia m e ter Ch arac teristi c Cu e { o r Fan
.

r v D u Cha y la 1 43 " , .

l 2 l n c h C e n t re ( fi g Fan f r th e ( fig 6 2) 1 45
.
,

28 )
-
, . o , . .

5 D us t lad Ai Fan s f 2 69
en r
E x t r m e Te st of an Ax ial
or
D yn am om e t ri c B rak e ( fig 1 13 ) 22 6
. o
, , .

of V an ce , e
or S cre w Fan of a Type Sim ilar
. .
,

to th at sh ow n i n Fig 13 5 f o E ff ec ts S (2 283 u
F t E x tre m e 25 1 E ffi ci en ci es T erm s ! E q u a ti on fo r
.
, , , .

0
fin d i n g the M an o m e tr i c an d
.
, .

D iam e tr al F an M or ti er ( fi g 7 0) 1 50
E x pe i m n ts M o rti e r M echani cal 28
, . . .

r e , , 1 51 .
,
.
2 96 I N D EX
ci e n c i es of F an s D nk i s T e sts o n
’
E xp r im n ts b y R C P rs n s C
e e a o en .

( figs 4 8 an d 4 9 ) 1 00 ( fig t i f g l P u m p 1 33 ( T bl
, , . .
.

r u a
of th e Par k en d M ne Fan M an o
a e
C n t i f ug al F n s b y B ry n
.
, .
, ,

i o n e r a a
m e t r i c ( T abl e 4 2) 1 82
E mc i e y G a u ge an d P o w e r Diagr m s
,
D n k i n 83 o
Mi n F n s B l g i n C m m issi on
, .
, .

nc , a on e a e a o
( fig s 2 3 an d 2 4 ) 7 2 ( fig 25 ) 7 3
, ,
.
, 1 05
o f a K le y F an ( Table 3 2) 1 5 5
. . .

, . W l k r s 208a e
’

V al u a t i o n o f Man o m e t ri c an d A i
, .

r Th E l v n Ty p s f F
e e use d i e e o an n
( a l 2 ) 3 5 ( Tabl e 3 ) 3 6 ( Tab le
,

Tb
'
e , , D ki 94 on n s, .

4 é w i th G b l ( T bl 1 0) 1 1 0 ui a s, a e
Mi n i m u m
.

Ve l u m e t r i e , 3 9 . w i th M x i m u m d a an

g yp t , K h at a t b e b , 19 1 . E q uiv l e n t O i fi E h F a r c e s on ac an ,
E le c tri cDr l ve , A rrange m en t o f ( fig s R su l s f ( T b l 9 ) 9 2
e o a e .

an d , .
,
111 1 1 2) 226 . w i h M r ti r F
t R su l ts f ( fi g
o e an , e o
M o to r F an w i th S p e ci all y D e
.
,

7 0) 1 5 0
w it h P r p ll r V n ti l ti n g F n s
, .

sig ne d ( figs 1 22 an d 123 ) 23 8 o e e e a a

E le v a ti o n o f F an B l d e s ( fi g s 4 3 an d
, . .
,
a , . 2 24 .

44 ) 8 7 . w ith R t u F n s 1 1 8 ( T bl 1 3) a ea a a e
E le v en T yp e s of F us e d i n D en k i n s
, ,
’
an 1 13
W i th S
.

E px cnm t 94 en s, F n s ( T bl 1 8) 1 1 8
er a a e
E x t r m D i m t r f V n s T st f
.
,

E nergy T h e C on s er v a ti o n of 1
.

e e a e e o a e e o an
A xi l
, , .

o r H ea d w hile passing th r ou gh th e
,

S r wF
a or f Ty p S i m i l r
c e an o a e a
F an L oss e s of 1 9 t t h t sh w i
o a F i g 1 35 f o Ft n n o
t o an I n com pre ssi ble F l uid i n
, , . .
,
.
,
25 1
M o t i on A p p h t i o f th e L w
.

, ca on a Ey d F e C h r t r isti
an , f Tw a ac e c o o
o f C o n s erv a tion o f 2 ( fi g 6 6) 1 4 8 .

E q u a ti o n s f o r fin di ng th e M an o m e t ri c
. .
,

t st d b y M L l n g T w
e e e o o
an d M e chan i ca l E ffi c i en c i e s i T erm s
o
.
, ,

n
of 45 23
H S o l i ngen G er m an y 27 3
, .

r b e, F ace o f t or of a C om p r sso r w ith

t h e Ro e
E s ch er Wyss an d C o f Z S t rai g h t B l ad e s M e th d o f d r w in g
.
, , , .

h 289 o . ur l c , ,
o a

K l e y F an 1 5 3
.
,
xer ci se s , t h e H yp b oi d al ( fi g 1 5 1 ) 29 0
er
E xhausti ng w ith D is c Co n t ra ct e d
. .
, .
,

, ,
F cfl i t i
a f o R e p ai r 27 7
es r , .

1 l t ( fig 1 26 ) 2 4 0
n e Far co t 1 9 1
E x peri m en t C om pa riso n be t w een
, . .

Fi l d Mag n e ts ( i n S c ti o n ) o f F n
.
,

, e e a ,
Th e o ry an d 1 7 7 . F ig s 1 22 a d 1 23 ( fi g 1 24) 238 n , . .

F i x m g I m p e ll e r t o S naf t Me t h od o f
.
,

o n E ach o f t h e E le v en T y p e s of
F an D en k n s B e s t ( T able 8 )
, ,
’
i 27 9
Fl w i n a P i pe P ssa g e o f U n i f orm
, ,
.

91 . o or a

p i m en tal App ara tus D n k i n s 88 S e c ti n Th e L o ss f H ad b y S u r f c e

’
x er e o o e a
F ri c ti n of a F l uid u n d r S t dy 4
.
,

F an ( fig 69 ) 1 4 9
, ,

o e ea
R eadings a t 6 00 R p er M i n B u k e ts t t h e
.
, ,

ev F l uid i m p i n g in g on c a
Pe ri ph ry of t h e F V n til at
. .
,

( fig 1 1 6) 2 3 2 e an e

E xperi m en t s B ry an D
o
.
,

d r iv e n b y a S tr m of
.

ki ( Table i g F
’
, on n s, n an ea ,

3 7 ) 1 7 8 ( Table s 38 an d 3 9 ) 1 7 9
i n Moti n A p pl i cati o n o f t h e L aw
,

( T able 4 0) 1 80 ( T able 4 1) 1 81 o
B u m pst a d an d Ch n d ler Fan f E n e rg y t o an
,
.
,

e a f C n co t o s a i on o
I n co m p r ssi b l 2
,
1 5 8 ( T able 3 4 ) 159 e e,
F an B l ad e s i W G Wal k er s P um ps C n st ru cti on f Ce n t i f u
.
, .

'
n . .
, o o r
,

( fig 1 1 0) 225 g al 261
P i pe o r
.
,

u n d e r S te ady F l w i
. .

F an N 1 D n k in s ( T abl e 7 ) 90
o
’
o n a
e
Pass ag e f U n i f rm Se c ti n Th
.
, , .

Fo rm ul e us ed i n D n k i n s 1 02 a e
’

,
o o o ,
e

L ss f H e d by S u rf ce F ri tion
.

H i rn s 1 99 ’

,
o o a a c

M or ti er s D i am e t ral Fan 1 5 1
.

of a 4
'

F rce d D r u ht F D o ubl e I nl e t
. .
, ,

Op en r u nn i ng F an 1 3 7 o a an ,

R atea u F an ( Table 23) 1 25

-
,
.
,

( fi g s 8 6 an 87 ) 1 64
. .

Fo rm u l us d i D o n k i
.
,

S er F an 1 6 0 . ae E xp e r ie n n s
’

m n ts 1 02
,

S m all S er Fan ( Table 3 5) 1 60 e

W G Wal k er s ( Table 5 4 ) 228 F r m s of B l d e Te ste d ( fi g 22) 68

,
.
, .

’
o a ,
. .

Fo rm ul A ppl i a ti f 3
.
.
, ,

( T abl e 5 5 ) 23 0 ( T abl e 5 6) 23 1 a, c on o

b y Ch arle s H I nne s C en t ri f u ga l F r e E n d S ubdivisio n of Pi pe ( fi g

.
,
.
,

e-

Pu m p 1 26
.
. , , ,

b y M L el ong M i n e V en tila tin g Fr i c tio n Wor k e xpe n d e d i n ove r

.
,

co m in g D i sc 1 7 1
.
, ,

Fan 1 3 8 , . , .
I
29 8 I N D EX

J e nk s, G F , B i rm i ngh am 262 M ehl er C 156

ed B l ad e for a C en t ri fu ga l M é moi re d e l Soc iété d e s I ngén i e u rs
. . .
, , .
, .

o i ts
J n , Ste p s a.
Fan wit R i bs for sti ffen ing t he m Ci il s for 1 87 8 107 v
a t t h e ( fi g 1 40) 263
,

M et hod of d raw ing t h e Hype r b i d l

.
,

o a
Fac e of t h e R o tor of a Com pre s
. .

J u d e Dr 1 5
, .
, .

R H 29 2
. . sor wi t h S t ra ig h t B lad es ( fi g , .

1 5 1 ) 290
K e i t h F an 260 of fi x i ng I m pell er to S h ft 27 9
.

a
M eth ods of Co n s t ruc ti on 27 7
.
, .
,

K w O bs er v a tory 22 4
e
M i ne M o i ng of C asi g of M orti er F
.
, , .

K h t t b h Eg yp t 1 9 1
a a e v n an
K l y F an 1 5 3
.
, , ,

e to s u it t h e ( fi g 7 2) 1 5 1
Fan s B e lg ian Commi ss i o ( 1892)
.
,
. .
,

E fii i e y of a ( T ab l e 3 2) 1 5 5
c nc n
F ans Tes t s of ( Small er Ty p ) , .
,
Exper imen ts on 1 05 e
Ch arac teri s t i c C u r ve s ( figs 30 , .
, ,

( T able 3 3 ) 1 55
V en tila tor ( fi gs 7 4 an d 7 5) 1 54 an d 3 1 ) 7 8 ( fi g s 3 2 an d 3 3 )
. .
,

7 9 ( fi gs 3 4 an d 3 5 ) 80 ( fi g
. .
, , .

. .
, ,

La v a l Tu rbi ne See u n d er De L av al 3 6) 8 1
Ori fi c e Permanen t A l tera tion
. .

T u rb ine
Law f C h ange of M omen t of M omen C as i g of Large M ortier Fan t ,
.

o n o
tum 16 s u i t ( fi g 7 3) 1 5 2
L aw of C on s erv ati on of Energy t o n V en tila ti g Fan ( fi g 29) 7 6
. . .
, ,
a n
I n c ompr ss ible F lui d i n M oti on Tes t ing of a 7 6
. .
,

e , , .

A ppl i ca ti on of t h e 2 B lan y 1 7 3 z
Con sol id ati on W e stph alia 1 26
. .
, ,

L elong M 138 1 4 0 1 4 2 1 43 1 46 1 48
M in i m u m Eq u i valen t O i fi on Eac h
.
, .
, , , , , , , , ,

r ce s
L o s s of Hea d b y S u rfac e F ri c ti on of a F an R es u l ts of E xperi men ts
Fl u i d u n d er S t ad y Flow i n w i t h M ax im u m an d ( Table 9 ) e
,

a Pi pe or P ass age of U n i form

,

92 .

Sec t ion Th e 4 L oss of Head d u e to a S u dd en

d u e to a S udd en Ch ange i n t h e C h ange i n t he D i rec tion of
, , .

D i rec ti on of M o ti on M o t i on C on d i t i on f 7 or
C on d i ti on for M in im u m M in in g C omm i s i on Tes ts P ru ss ian
.
, , ,

, s ,

7 1 03
Ch ange i n V elocity w ith o ut M xed F low Fan s 40
. .

i
C h a ge i n D irec t i on 9
.
,

n De s ig n of 4 6
t of M omen t u m L aw of Ch ange
.
, ,
.

L o ss e s of Energy or He d w h i le pa ss M
g 16
a o n en
ing t hro u g h t h e F 19
,
an 0
M o en tu m Law of Ch n g e o f M omen t
. .
,

M agne ts ( i n Sec tion ) of Fan F igs 1 22 m 16 0

,
a

an d 1 23 F i el d ( fi g 1 24 ) 238
. .
,

M t m b r t 1 26 on r a e
M anomet er an d Pi to t T u be 1 0 M ortier 1 5 0 1 5 1
. . .
, , ,

M anome t ri c an d Ai E ff i ci en c y Varia M o tion A ppl ic tion of t h e Law o f

.
, , , .

r a
t i on of ( Table 2) 35 ( Table 3 ) 3 6 Con s er vation of Energy to an , ,

I c ompre ssi bl e Fl u i d i 2
, , ,

( Table 4 ) 3 7 n n,
M e c h an i ca l E fii i i i n Con d i ti on for Min i m u m Lo ss f
. .

and c enc es o
Term s of E q u a ti on for fi n d i ng Head d u e t a S u dd en C ha ge i o n n
t h e 23 th e D i rec ti on of 7
Effi c i en c i e s of th e Park en d M ine M o tor F an w it h Spe c iall y Design d
.
,
.
,

e
Fan ( Table 4 8) 182 Elec t ri c ( fi gs 1 22 an d
,

23 8
Mar i ne V en tila ti ng Fan Experi men t s
. .
.
,

Spee d C ha rac t eri s ti cs of ( fig 1 1 7 )

,

b y M L elong 1 3 8
.
, ,
23 3
M axim u m d M i n i m u m E q u i v alen t d ri v en T e s t of a 4 F t 6 I n S i gl e
. . .
,

an - n
E ac h F an R e s u l ts o f I nl e t F an ( fi gs 90 an d 9 1) 1 66
. .
,
O ifi r ce s o n
( fi gs 92 an d 9 3 ) 1 6 7 ( fi gs 94
, , .
,

E xperi men t s wi t h ( Table 9 ) 9 2 d an

G a i n of P re ss ure Head C on d iti on 9 5 ) 1 68 ( fi gs 9 6 an d 9 7 ) 1 69
. .
, .
,

M o i ng of C as ing of M orti er F an t o , . .
,

for 80 v
M e s u ring A p para tu s B y an D k i n s s uit th e M i ne ( fi g 7 2) 1 5 1
.
,
'
a r on
Pre su re ( fi g 46) 89
. .
, ,
s
P ower V ol u me an d Speed B ee nan
. .
,

N i c h o l son Dr 4
N orth E a s t Coas t I n s titute of Engi~
, , .
,

an d G i lber t s A ppara tu s for ( fi g

, ,
'
.
-
,

26) 7 4 an d S hi pb ui l d ers 29 2 n ee r s
M e ch a ic al E ffi ci en ci e s i n Term s of f »
.
.
,

n i

E q u a ti on f o fi n d i g t h e M a ome t ri c O n ion s S ee un d e r Al ld ay s an d
,

r n n .

an d 23 ,
. O nio s n .
 I N DE X 299

O p en ~
r un n i n Ex pe ri men ts 137 Fan P re ss u re C en trif ugal C om pre ssor
Seg hgll Co lli e ry ( T abl e 4 9 ) 1 9 2
, .
, ,

H ig h 27 7 ( fi g 1 8) 2 0 ( fi g 4 8
w i th R a d ia l Vanes ( fi g 67 )
.
, .
, .
,

8 )14 1 a 28
Fan s H i gh 1 94
. .
,

149
Ori fi c e P ermanen t A lter t i on C a s in g T u rbine Fan Experi men ts R a tea u
. , .

a
of Larg e M ort i er F an t o s uit
, , ,

H ig h ( Table 50) 20 1
M in e ( fi g 7 3 ) 1 5 2 P ro fi le To d ra w R o tor 28
.

9
Pro pe ll er F an V e l oc i t y D i agram s for
.
, . . , ,

Re d u c ed ( fig 1 9 ) 6 4
on Ea ch F an
, ,

( fi g 109 ) 20
. .
,
R es ults of E x p e r i 4
men ts wi th M ax i m u m and M in i V en t i la t ing Fan s E xpe i men t
. .
,

, r s
m um E q u iv al en t ( Table 9 ) 9 2 wit h 2 24
Oth er Prope ll er V en t ila ting Fan s an d
.
, , .

Th e T h eory of 203
Ra tea u Sc rew Fan s 24 3 Pro pe ll er V en tila t in g an d R a tea u
.
,

Out l et Ci rc u la t i ng w i t h o u t Di sc C on Sc rew F an s Oth er 24 3

, .
, ,

, , , , , .

trac te d ( fi g 1 27 ) 240 Propeller s Heli c l 2 19 a

Propell i ng w i th D isc C on t ra c te d P i p lli g w i th D i sc C on t ra c t e d Out
.
, . .
, ,
c e n .

( fi g 1 25 ) 240
, , ,

le t ( fi g 1 25 ) 24 0
, ,

O v erc om i ng D i sc F ric ti on W or k ex Prus si an C omm i ss i on 1 87 1 89

. . , . .

,
M i n i g C omm ission 9 3 , , .

pen d e d i n 1 7 1 , . n , .

T es ts 1 03
P u mp Experi men t s by C R P ar s on s
.

P ac ki ng 282
P ari s Ex h i b i ti on 1 9 00 1 7 3
. .
, .
,

C t if
en r i 1 3 3 ( Table 2 6) 1 34
u a
T ri al s by C arl e s H I nnes C en t ri
. , , .

R at ea u an d Co 269
,

. .
,
.
,

Sa u tter Har lé 1 9 4 -
fu gal ( Table 24 ) 1 28
P ark en d Coll i ery C ompan y 7 6
.
,

( Table 25) 130

M in e Fan P u m p s C on s t r u c tion of C en t ri fu gal
. .
,

M n om e t r i c E ffi c i en c i es of th e
a
,
F l u i d 26 1 ,

P u ri fi er C en t ri fu ga l Gas 2 67
, , .

( T able 4 2 ) 1 82
Pars on s Hon C R C en trifu gal P u mp c on s isting of a Fan an d 9
.
.
, ,

. .
, ,
T u rbi n e i n Seri e s ( fi g
. .

Expe rim en ts 1 33 ( T abl e 26 ) 1 34 ,

.
, .

P assage Re s i s ta n c e of R e c tang ular

,

1 4 3 ) 268
Po w er V o lu me an d Speed Heenan
, , .

( Tables 29 an d 30) 1 4 1
t h ro u g h a R ad ia l Flow Fan W or k
.
, , ,

, an d G i lbert s A ppara tu s f '

or
d on e on th e Air i n its 1 7 meas ur i g ( fi g 26) 7 4 n
Pec l e t 1 42 , . .
, .

D i agram s Effi c i en c y G au ge an d
P el ton Bu ck e t 27 1 27 2 ( fi gs 23 an d 24 ) 7 2 ( fi g 25 ) 7 3
,
.
, ,

Pressu re a t Bac k of W h eel ( fi g 1 8) 61

. .
, , .
, , .

W h eel 1 05 27 1 27 2
P e l zer D ort m u n d F n 1 56 ( fi g s 7 6 Press u re Head C on d i tion for M
.
, , , .
, .

a , , .
,
an
an d 7 9 ) 1 5 7 m u m G ai n of 8
Per h ery of th e Fan V en til ating F an i n C en t i f ugal F n Va iati on of
,
. .

d ipv en by a S tr eam of Fl u i d i mping 5;

,
z r a , r ,

i n g on B u ck e t s a t t h e ( fi g 1 4 6 ) 27 2 -
mea s u ri ng A ppara t u s B ry an
P ermanen t Al te rati on C as i ng of L arge
.
, .
,

Den k i n s ( fi 4 6 ) 89 '

Mort i er F an t o s u it M i ne Or i fi c e
.
,

Pr e ssu re s an d V u me s of A q u eo u 0 s

( fi g 7 3 ) 1 52
,
V apo u r S t u t i ( T able 1 ) 16 a r a on ,
P ers pectiv e Vi ew o f Ra tea u Fan ( fi g P ropeller F an s S c rew or 40
. .
, .

P roportion s of R at eau F an s ( Tabl e 5 )

.
, , .

1 4 ) 50
Phy si c " Ex pe ri men ta l an d Th eo ,
.

s 53
P u m p B alan c ed C en t r i f u gal ( fi g 1 7)
.

r e t i cal by R H J u d e 29 2
Pi ck ard F an s See u n d er Hatte rs ley
, . . , , , .
.

60
Pi ck ard Experi men t s b y Ch arl es H I nn es
. .

Pi pe F r es En d S u bdi v is ion of ( fi g 4 7 ) Cen t ri fu gal 1 26 ,

. .

-
.
, , , .

96
Wo r k d o n e
.

or P assage of U n iform Sec ti on , R dial Fl o w F

a an , on
Th e Los s of Hea d b y S u rfa c e th Ai i its P sag e e r n as
Fri c ti on of a Flui d un d er S t ead y th ro u gh a 1 7 , .

Flow i n a 4 D esi gn o f 4 0
Pi to t Tub e Manome te r and 10 ,

R an kin e Cy cle 202

. .
,

Ra te u 4 0 4 1 4 7 105 1 09 ( T ble 13 )
, , .
, .

Th o
r y of t h e 81 e , . a , , , , , , u
Plot ti ngs of l able 55 p 230 ( fi g 1 1 8) 1 1 3 T b le 1 7 ) 1 1 7
( T abs 86 ) 1 7 7 1 7 8 1 9 5 1 9 8 1 99
, .
, . , a 145 , ,

0 , , , , ,
P lu m m et T H 266 27 0 200 218 2 60 26 2 2 6 7
Pr c t i ce Rec en t : Fan s an d Com pres
, . .
, .
, , , , .

a , F an ( fi g 1 2 an d 1 3) 4 8 ( fi g 5 9 )
, s .
, .

so rs 258 , . 1 19 1 2 1 , 1 88
. .
3 00 I N D EX

R a t ea u F an C on solid ation M ine s S c re w De v e l op men t o f a ( fi g 1 1 ) 4 6

F an Te s t of an A x ial of a Ty pe
.
, , , .

W e s tph li a 26 a 1
P er spec ti e V iew ( fi g 1 ) 50 S i m i lar t t h a t s h ow n i n F i g
.
, , ,
v 4 o
F t Ex treme Dia
. .
, .

Tes t T y pe A Table I ( T able 1 3 5 of

me t er o f V anes 25 1
.
, , .
, , .

1 9 ) 1 20 ( Table 2 0) 1 2 2
F an s Ot h er P ropeller V en til a ti ng
.
, , .

T e s t s ( Table 21 ) 1 23 ( Tabl e 22)

1 24 .
, ,
F an s a d Ra tea u 243
,
n .

Type A Tabl e I ( Table 4 5) R a tea u 24 3 ,

T h e T h eory of R a teau 25 1
. .
, , ,

1 8 6 ( Table 4 8 ) 1 88
or Propeller F a ns 40
.
, .
,
Q 1 7 F t D i ame t er ( T able 4 6)
’

Se c on d F an t e s t e d b y M L e l ong ( fi g
. .
, ,
1 87
F an s fi
. .
,
64 , (g 1 9) 6 4 64 ) 14 6
i an C omm issi on Segh ill C oll i er y 1 9 1
.
, . .

el g B ( Tab l e
O pen r u n ing F an ( Table 4 9)
, , .

4 7 ) 1 87 .
- n ,
Experimen t s w it h 1 18 ( Table 19 2
Ser F an 1 05 ( fi gs 5 5 an d 56) 108
.
, ,

P roport i on s of ( Table 5 ) 53 , .
, ,

( Table 1 1 ) 1 1 1 ( Table
Te s t W i t h D iff u s er onl y ( l ab l e
, .
,
‘
Experi men ts 1 60 .

Small ( Ta b le 3 5 ) 1 60
, ,

1 8) 11 8
S c rew Fan s 2 43 244 Ser Profe ss or 1 07
. .
,

V en tilator 1 09 2 46 Ser v i c e of D i men s i on s C o s t an d

.
, , , , .

Ra t ea u an d C P ar i s 269 C apell Fan s ( Table 1 6 )

.
, , , ,

o .
, , .
,

R ead i ng s at 600 R e v per Mi n E x p e r i 116

G u bal F an s ( Table 1 4 ) 1 1 4
. .
,
men tal ( fi g
.

23 2 l
R ea son for t rea t i ng Ai r as i f it h a d a R a t e u Fan s ( Table 1 7 )
.
, .
, .

a
C on stan t V ol u me 3 1 17
,

R e c en t Prac t i c e F an s an d C ompre s Ser F an s ( Tabl e 1 5 ) 1 15

.
, .

sors 2 5 8 S h aft M e th o d of fi x i ng I m p eller t o

.
, ,

R e ctang u lar P ass age R e si s tan c e of

.
, , ,

, , 27 9
( Table s 2 9 an d 30) 1 4 1 S i emen s 2 60
.

S ingle I nle t A llen F an ( fi gs 88 an d 89 )

. .
,

R e d u c e d O i fi ce ( fi g 1 9) 6 4 r s,
R egnau l t 1 3
. . .
,

1 65
R epa i rs Fa c il iti e s for 27 7 M o tor d ri v en Te st s of a
.
, .

Fa n,
4 Ft 6 I ( fi g s 90 an d 9 1 )
-
, , .
,
R es i s tan c e d u e to B en ds ( Table 3 1 ) n
1 66 ( fi gs
. . .
, ,
144 9 2 an d 93 ) 1 67
of R e c tang u lar Pa s s age ( Table s 29 ( fi g s 9 4 an d 95) 1 68 ( fi g s 9 6
.
, .
,

. .
,
an d 30) 1 4 1 an d 97 ) 1 69
,

R e s u l ts W a dd le F an T e s t s ( T able 28) Small Ser F an Exper i men ts ( Tab l e 3 5 )

. .

, , ,

138 1 60
of Experi men t s W it h M ax im u m
. .

é t é d e l I d u s t i e M i n érale 29 2
’
Soc 1 n r

an d M i n i m u m E q u i valen t O i So c i e t e d e s I n gé n i e u r s C i v i l s 1 8 7 8
.
,

r
fi s on Ea c h F an ( Table 9 ) 9 2
, ,
ce Mem oi d e la 1 07
r os
of E xper i men t s w i t h M or ti er Fan
, .
, .

Sol i ngen G e m an y 2 7 3 r

( fi g 7 1 ) 1 50 Spe ci all y D e si gned E le c t ri c M o tor F an

.
, , ,

R e v ol uti on s p M in u t e R ead i ngs a t w i t h ( fi g s 1 22 a d 1 23 ) 2 3 8

. .
,

er n
600 ( fi g 1 1 6 ) 2 3 2
.
, , .

Spee d Heenan an d G lb t A pparat u s 1 er

’
s
for mea s u ri ng P ower V ol u me
.
, .
,
R ue U w
ev nell d Jil i meas 29 2'
r es s,
R ey nol ds P rofe ss or Os borne 2 7 8
.
, ,

an d ( fi g 26) 7 4
R i b s for s ti ffen ing t h em a t th e J oi n t s Ch arac teri s ti cs of M o tor ( fi g 1 1 7 )
. . .
, , ,

S t eppe d B lad es for a C en tri f u gal Fan

, , .

233
w i t h ( fi g 1 4 0) 263 Spe c ial D i ff u s er a d V ol u t e T h eory of
.

n
th e 55
.

R i d ler 7
.
, ,

S t a ge ( Two C en t r i fu gal F an ( fi g 1 4 2)
. .
, ,

R obi n s on A nemome t er s 1 04 , .
,
.

R oo ts B low ing Mac hi nes 203

’
2 66
R uth erfor d C ollege 1 27 S t eam s h i p S tok e h ol ds F an s for ( fi g
. .
,

R o tor of a C ompre ssor w it h S trai gh t

.
, .
, ,
1 4 1 ) 265
M e th od of d raw i ng th e S t eppe d B lad e s for a C en t ri f u gal F an
.

B la d e s
H yp e b oi d al F a c e of t h e ( fi g 1 5 1) w i th R i b s for s tiff en i ng t h em a t
, ,

r .

29 0
,
th e J o i n t s ( fi g 1 4 0) 263
R o tor Pro fi le T o d raw 289 V ane s for C en tr i f ugal Fan s 262
. . .
,

St i fi e n i n g t h em a t t h e J o i n ts S t eppe d
.
, ,
.
,
'

Sat u rati on P re ss u re s an d V olu me s of B la d e s for a C en t ri f ugal F an w it h

,

A q u eo u s V apo u r ( T able 1 ) 1 6 R i bs for ( fi g 140) 263 ,

M
. . .
,

S a u tter H ar lé P aris 1 9 4 1 98
. -
, , ,
. S t ock alp e r , 7 .
 3 02 I NDEX

Ve l oc ity w ith ut C h n g i n D i r c ti n o a e e o W add l F n e a , 13 7 , ( Tabl e 28) 1 3 8 , ( fi g .

L ss f H ad d t C h ng e i n 9
o o e ue o a 6 1 ) 1 39
Pr p lle r F ( fi g
. .
,

Di g r m s f W lk r s W E G xp ri n ts
’
a a or o e an , a e e me , 208 . 2 2 4 ,
.

fig
.
, .
,
1 09 ) 204 . ( 1 1 0) 2 25 , ( Ta e 5 4 ) 228 , ( Ta e
. bl bl
i n t h e V l u t ( fi g 1 6) 56 o e, 5 5 ) 23 0, ( Ta e 5 6 ) 23 1 , 23 7 , 24 1 , 2 4 2, bl
V n ti la t u rs d Mi n s L 292
. .

e e e e ,
es , . 2 45 .

V n til ti n g F
e a Mi n ( fi g 29) 7 6 an , e, . . W l
a lace , W M 29 2 . .
, .

T st f Mi n 7 6
e o a e, . W lls n d Sl i p w y 1 26
a e a , .

T u rb d riv n 27 1 o- e W rshi ps 1 3 8
a
t r G u g D t rm i ati n f H ead
. .
, ,

E xp ri m n ts b y M L e l n g e e . o , W a e a e, e e n o o
1 38 fr m o d th C l u l ti n
an f Ai r e a c a o o
D n sity f G ive n H um idity 13
,
.

d riv n b y S tre m of F l uid e a a e or a

i m pi n g i ng B u ck ts t t h
.
,
on e a e W th rc oc k G uid e V n s ( fi g 1 49 )
ea e a e
P ri p h r y o f t h F ( fi g
.
,

e e e an , . 284 .

1 46 ) 2 7 2 W e stm in ste r 2 24
F n s E xp ri m e n ts w ith P rope ll r
.
, .

a , e e ,
W e stph l i C n s l id ti n Mi n s 1 26
a a, o o a o e , .

2 24 W is b h 5 1 4 2 1 43
e ac
Th Th eo ry f P r pe ll r 203 W h l P r ssu r t B k f ( fig 1 8) 61
. .
, , ,

e o o e ee e e a ac o
d R atea u Sc r w F n s O th e r W i l x S u n d e r Babc c k d Wi l
.
, , , . .

an e a co ee o an

P r p ll r 243 , .

o e e cox
V e n ti l to r K l y ( fi gs 7 4 W ith D is C n t rac t d I nl t E x h ust
, . .

a d 7 5) 1 5 4e an . c, o e e a
i g ( fi g 1 26) 2 40
, , .
,

R te a u 1 09
a n
V i w f Rat e u F Pe rs p ec tiv e ( fi g O ut l t P r pe ll ing ( fi g 125)
.
, .
, .

e o a an , , . e ,
o .

1 4 ) 50
W ith ut D is C n t racte d O utl t
.

V i l l s 1 21
a , . o c, o e ,
V l u c a d S p d H e n an a d Gi l
o a n ee e n C i r u l ti n g ( fi g 1 27 ) 2 4 0
c a
t h A i i n its P ss g
.
, , , . .

b r t s A pp tus f
e
’
m eas ur i n g
ar a or W rk d ne
o o on e r a a e
P w r ( fi g 26) 7 4
o e th r u g h R di l Fl w F n 1 7
o a a a o a
i n v r m i ng D is
. . .
, ,
V l u m s f A q u us V po u r S tu ra
o e o eo a a e xpe n d d e o e co c
ti n Pr ssu r
,

o d ( T bl 1 ) 16
e e an ,
a e . F ri ti n 1 7 1 c o , .

V l u m t ri E ffi i n y 3 9
o e c c e c Wyss d Co S u n d r Es h e r W yss
an ee e c
Spi ra l D i ff us e r
, . .

V l ute Th r y f t h
o , eo o e an d Co .

d 55
an
V l ity in t h e ( fi g 16) 5 6
.
,

e oc , . . Z u ri h c , 2 89 .

H Y D R A U L ICS 1N T H EORY A N D P RAC T I CE

By W M WA LL A C E W
.

. .
, h Sc . .
,

43 . ne t ; p o t fre e s , 43 4 d . . ne t ; ab ro ad 45 7 d n e t , . . .

EN G I N EERI N G P ROBLEMS Par t I By W M

WA LLA C E
, . . .

. Pri c e 35 . ne t ; p os t f r e e 33 4d n e t ; ab ro ad
, . .
,

3 5 6d . ne t .
[ 7 m l [ 1 i n p r ep a r a t i on
3
. .

WO RKS HO P C O S T S ( fo r En g in e e rs an d Manu fact ur e r s) .

By S I N C L A I R FEA RN an d FR A N K FEA R N Pri c e 12s 6d . . .

n e t ; p o st f re e 135 n e t , .
; ab ro a d , I 33 . 6d . ne t .

T H E T EC H NICA L P UBLI SH I N G C o .
, LI MIT ED ,

1 . C OUC H SQ UA R E. F EET L STR EET . LO N DON , EN G L A N D

B I L L I N G A N D SON S . LT D .
, P R I NT E RS , G UI L D FO R D , E N GL A N D.
 RAC [ CA
N(ll R
A N D E N G N EE RS ’ GA Z ETT E I .

An I ll ustr t d W kly N sp ape r

a e ee ew

gi v in g c o nc i se te ch nical info r m atio n

f o r all e n gage d in e ve ry b ran ch of
En g in e e r in g P r acti ce and A dm ini str a

SP EC I A L FE AT U R ES
W rksh o op N o te s, P ate n t N o te s an d
Spe cificatio n s , Boo k Re v ie w s , Eu
g in e e r in g P ro bl e m s an d a ll c u r re n t
topics .

O r de rfro m yo u r N e w sa ge n t Bo ok .

se ll e r o r i f any d iffic u l ty i s e it p e r i e n ce d
,

in o b tainin g t he p ape r pl e as e c o m
m uni ca t e w ith the P u b l i she r s .

P r e p aid nn u al su b scriptio n at ho m e
a ,

7 3 6d . . ne t C o lo nie s o r ab roa d .

lOs . ne t.

T HE P RA CT ICA L E N GIN EE R
'
A N D E N G I N E E RS G A Z ETT E .

l, G o u g h S q u ar e , F l e e t S tr e e t ,

LON DON E C EN GLA N D

"
, . .
, .

Te le p h o n e — 10 l rn
2 1 Ho b o . T c le . A dd — .
"T e p uc o l, F l ee t . L d on o n.
The Te chnic al P ublishing C o L f : .
,
‘i

l , C OUC H SQ U A R E , FL EET ST .
,

LON DO N , EC . .
, EN G LA N D
Te le g r ap h i c Ad d r es s

T EP U C O L . FL EET . L D ON ON .
’
102 1 H OLBORN .

N n u m e r ou s re q u e s ts f or lists of B OOK S on
c o n se q u e n c e of

vari ou s s u b j e c ts w e h av e d ec i d e d to i n c l u d e s u c h i n ou r tw o
,

P oc ke t B oo ks w h i c h ar e p u b lis h e d an n u ally ab out D e c e m b e r T h es e

, .

l is ts ar e b r ou g ht u p to d ate p e ri odic ally an d c o n tai n th e title s , ,

autho rs an d pr i c es of
, t h e l e ad i n g E n g li s h an d A m e r i c an te c h n i cal

p ub li catio n s I n add iti o n to th e val u ab le i n f o r m ati o n th e P o c ket B oo k s

.

c o n tai n th e se L I S T S f o rm a p e r m an e n t ready re f e re n c e catal o g u e of

,

C ON T E M P O R A R Y TE CH N I CAL L I TE R ATUR E .

P RI C E S of e i th e r “
TH E P R A CT I C A L E N GI N E E R " ,

ME C H A N I CA L o r E L E CT R I CA L P OC K ET B OOK S :
In C l oth Bi ndi n g l ,
s . n et ; p o st f r ee i s ,
. 4d . n et ab r o ad ,
l s 6d . . n et .

In P lt i eB i ndi n g l
ne , s . 6d . n et p o s t f re e ,
I s 1 0d
. . n et ; ab r o ad ,
25 n et .

O rd e r f ro m you r B oo ks el e r , l N e w sag e n t , o r d i re ct f rom th e

P u b li sh e r s , a ddres s as ab ove .

I f you ar e in w an t of an y b oo k on an y te ch n i cal s ub ec t j wr ite

f or a v d ic e to
Th e Li t e r a r y Edi t or ,

TH E TE C H N I C A L P U B L I SH I N G CO .
, L TD .

1 , G o u gh Sq u a r e, Fle e t St r e e t , L o n d o n , E C . .
, En g l an d .

Te le g r a ph i c Addr e ss :T “
E P UC O L , F L E ET , L ON D O N .
" T l ph o : H
e e ne 1 02 1 OL BORN
 FTO
IVERSlTY O RONTO
IJBRARY

c ar d

dfi s

P o c ke L

Acm e Lib r a ry Ca dr P oc k e t

Und e rP at . I d x Fi l
Re f . n e e .
”

Made b y L I BRARY BUREAU