KAUTZ

An Investigation of a il

Turbine Centrifugal Pump I

Mechanical Engineering

9 08

UNJVKRSITY

JT,L LIHRARY

UNIVERSITY OF ILLINOIS

LIBRARY I

t

Class Book Volume i

My 08-15M

Digitized by the Internet Archive

in 2013

http://archive.org/details/investigationoftOOkaut

AN INVESTIGATION OF A TURBINECENTRIFUGAL PUMP

BY

William Waddell Kautz

THESIS FOR THE DEGREE OF BACHELOR OF SCIENCE

IN MECHANICAL ENGINEERING

IN THE

COLLEGE OF ENGINEERING

OF THE

UNIVERSITY OF ILLINOIS

PRESENTED JUNE, 1908 ^^-

W\6

M

UNIVERSITY OF ILLINOIS

Jme 1> 190G

THIS IS TO CERTIFY THAT THE THESIS PREPARED UNDER MY SUPERVISION BY

WILLIMi WADDELL KA^TZ

ENTITLED -M INVEST KJAT ION OF A TURBINE CMTRIFUGAL PUMP

IS APPROVED BY ME AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE

DEGREE OF Bachelor of Soience in Mechanical Engineering

Approved:

HEAD OF DEPARTMENT OF Mechanical Engineering.

114588

' »

»

O

1.

INVESTiaATION 03? A TURBINE CE]\[TRII?UGAL PIMP.

INTRODUCTION.

The actual results olDtainable from centrifugal pumps

differ consideraul.7 from the results which should ohtain accord-

ins to theory. This is due to incorrect assumptions as to flow

thru the machine, and to indeterminate losses in the machine.

Desisners who have overcome the difficulties have apparently

tried to keep their information to themselves for their own

pecuniary advantage. Up to the i^resent time, few of the practif

cal methods of increasing centrif:ngal pump efficiency, or "better'

ing their operation, have "become very widely known, and a number

of tests must he run on any design of pump in order to deter-

mine its actual performance.

Within the last few years, centrifugal puraps'have cone

into use for fire service, city pumping, "boiler feed, etc. The

small floor space they occupy is the greatest point in their

favor, although their adaptability to pumping large quantities

of water against a low head is no slight advantage.

The object of this thesis is to investigate the opera-

tion of a turbine type centrifugal pump. The pump under test

in this investigation is a two stage, turbine type centrifugal

prnnpvmanufactured by Henry R. Worthington Co., which is owned

hy the department of Tlieoretical and Applied Mechanics of the

University of Illinois.

jiirETHOD OP TESTING:

The test of the pump investigated was made to deter-

2,

mine the range of operation and the corresponding efficiencies

for the pump. For each run the speed was kept constant and the

discharge and efficiencj/' determined for varioiis ?:eads. The

pump was placed over the sump in the Hj^'draiilic Laboratory of

the University of Illinois. For most of the runs the v/ater was

pumped through 80 feet of 12 inch pipe arid discharged into a

weir channel. In a few runs, liowever, the v/ater was pumped

through 110 feet of 8 inch pipe and discharged into a weir "box.

In all runs the head was varied "by throttling the discharge valve,

and the v/ater was measured over a weir, then allowed to return

to the SLimp to be used over again.

APPARATUS: The pump was belt-connected to a 100

horse power Ideal Engine. As stated before, the pump is a two

stage centrifugal pump of the turbine type. The runners are

15 inches in diameter and are made up of two circular discs

with the runner blades between them, all cast in. one piece. The

guide vanes surroimd the runner and their outside diameter is

28 inches. Figure 1 shows a section perpendicular to the shaft

through the runner and guide vanen. Figure 2 is a vert ical . sec-

tion througri the pump showing the path of the water from suction

to discharge pipe. Water enters the first runner at the center,

leaves at the periphery, entering and passing thru the guide

passages; it flows do^Nn from the first guide passages to the

inside of the second runner, and discharges from them, through

the second guide passages into the spiral trumpet shaped casing

and out at the discharge pipe. There are eigl-it vanes in each

runner, four of them extending from the outside to the center

3.

•

4.

and the other four extending only half way. l^e direction at

entrance is practically radial, and at exit the angle is twelve

degrees off the tangent.

A Crosby pressure gage was placed above t>ie piirap in

the discharge pipe at 45° to the axis of the pump, to indicate

the head at that point. A Crosby vacuurn gage placed in the

suction pipe showed the suction head. Calibration curves of

these gages are mot shown as they were calibrated by the depart-

ment and found to be correct. Two Crosby, inside spring, steam

indicators were used and cards were taken simultaneously from

both ends of the engine cylinder. The springs were calibrated

and found to be sufficiently correct.

The v;ater was discharged into a weir box and was

measured over a contracted weir. Por some of the runs an 18

inch weir was used and for the remainder a 3 foot weir was used.

The head on the weir was measured with a hook gage placed at the

side of the box and about four feet back from the weir crest.

Calibration curves for the two weirs are shown on curve sheets

I and II. By noting the head on the crest of the weir, the dis-

charge ma3'- be read directly from the curves. The curves were

plotted between head and discharge, using values calculated by

Smith's formula, using Smith's coefficients.

OBSERVATIONS; The engine v/as started and run at slow

speed until the pump was primed, and then the speed increased un-

til the desired R.P.M. was attained. The pump was primed by

filling from a standpipe in the Hydraulic Laboratory. In each

run the speed was kept constant and the head was varied; and the

5.

discharge measured for each change of head. Readings of the

head on the pump, the suction lift, the speed of engine and pump,

and the head on the weir, also indicator cards, were taken every

five (5) minutes. Tl-ie head on the pvsmx) at starting, was made

low enough that the pmnp operated at less than maximum economy;

four or five readings were taken with the head constant, and '

then it was increased by (10) or (20) feet, and four or five

readings taken at that head and this repeated until the maxiraurm

head was reached.

Water was pumped from the sicnp, through the twelve

inch main, to the weir box and flowed over the ?/eir back into

the sump and the water level in the sump was noted as soon as

the pipes and the weir box were filled, after which it remained

constant. Water was admitted from t?ie standpipe so that the

distance from the gage on the discharge pipe to the water ms

constant at ten (10) feet. The speed of the engine and pump,

were taken v/ith a Starrett speed counter and a stop watch. Speed

readings were taken on both engine and pump to determine the

amount of slip in the belt. The head on the weir crest was read

by means of a hook gage, tfiking a zero reading when the water

level was Just up to the crest, and subtracting that from the

readings when the water was flowing ^to get the actual head on

the crest.

Indicator cards were talcen from both ends of the engine

cylinder, by means of the two Crosby indicators, using (40) pound

springs. All readings were taken simultaneotisl^r and the head

was increased in each test until the point of maximum pump econ-

omy was well passed.

6.

The belt was thrown off and the engine run light to

determine the friction of the engine parts at various speeds.

Indicator cards were taken during this run with 20 pound springs

in the indicators. A curve of friction horsepower is shown

on curve sheet III.

TABLES

:

The tables on the following pages contain the cal-

culated data. On the test data sheets, column 1 is the number

of the readings col^jmn 2, the discharge gage reading; colTimn 3,

the distance from the discharge gage to the sump water level;

column 4, the total lift, being the sum of columns 2 and 3;

colurrm 5 is the calculated discharge, in cubic feet per second;

and colTimn 6, the calculated discharge in gallons per minute;

column 7 is the R.P.M. of the pump; coliiran 8 is the net horsepow-

er delivered to the pump, and column 9,^ the average net porsepower

delivered to the pump at each head; colTiimn 10 is the hydrau-

lic horsepower of the pump; col:imn 11, the calculated efficien-

cies of the pump, and column 12, the average efficiency of the

pump for each head.

Table No. 8 , p. 23 , shows the revolutions of the

ptunp required to give certain discharges against different heads.

It was calculated from the formula n ^ K V h -f- a which is ex-

plained later.

RESULTS Aim CONCLUSIONS:

The discharge from the pumiJ, column 5, was obtained

from the calibration curve for the weir used, curve sheet No.H

the discharge in gallons per minute, column 6, being calculated

"by the formula;

Gallons per minute =s w x Q x G x 60 (1)

w, "being the pounds of water per cubic foot (.-562.5), Q, the

discharge, cubic feet per second, and G the gallons per cubic

foot (-7.48).

The hydraulic horsepower of the |)ump, column 10, is

calculated from the formula:

Hydraulic H.P. = Q x w x H x 60 (2)330dtr

Q and w being the same as in (1) etnd H being the total

head, column 4. The net horsepower delivered to the pump, col-

vnm 8, is the indicated horsepower of the engine, less the fric-

tion horsepower. The engine horsepowers are calculated from

the formula:

I. H. P. = P_J. AJT3;50bo"

P being the mean effective pressure on thepiston (the average

from the t wo cards) pounds per square inch, L is the length of

the stroke, in feet, A the area of the piston in square inches

and N, the number of strokes per minute (= R. P. M. x 2)

.

The pump efficiencies are calculated from the for-

mula.

' Hydraul ic H. P^H.P. deliTI Fo pump.

Prom the data of the tests curve sheet IV is plotted, expressing

the relation between R, P. M. and head on the pump, for various

discharges. Curve sheet V shows the relation between R.P.M.

and discharge; curve sheet VT, the relation between R.P.M. and

8

efficiency and curve sheet VII shows the relation hetweon dis-

charge and effioienc}^, for various heads. From a stud^r of the

results it is evident that there is a fairly wide range of dis-

charge for any head without much variation in efficienc3r, and

obviously the pump may operate with the sane econora^r under dif-

ferent heads. It is also evident that at any discharge the head

ma3'- be varied considerably'- without a great change in efficiency

but the best efficiencies are obtained at high heads. The rea-

sons for this appear when the variations in the term a_ are noted

in the formula n = K

The maximum speed for the engine was 300 R.P.H., it

being equipped with an inertia governor, which prevented ex-

ceeding that speed; this would run the pump at 1300 R.P.M. , but

a run could not be made with pump speed exceeding 1200 R.P.M.

,

on account of belt slippage.

Variation of the engine indicated horsepower is per-

haps mostly due to the poor quality of steam obtainable, but

no doubt the average of the several cards, for each reading at

constant speed and constant head, represent a true average. The

friction horsepower curve is subject to a small error, although

repeated trials were made to get it, and considerable care was

used in taking data for it.

The losses entering into the pump operation due to

various causes, are: (a) a loss due to Journal friction of ro-

tating parts, (b) a loss due to friction of the water in the

suction pipe, (c) internal friction of the pump, due to friction

of the water over the rotating parts of the pump; and (d), im-

9.

pact losses, due to impact on the runner vanes. The velocity*'

head [V-t^) in the suction piDe requires energy'' and that velocity2 g

head counts as a loss against the pump. The friction is perhaps

excessive through the small section in the runner, where the

water enters, since the velocity must "be necessarily high, when

large volumes of water are discharging. There will "be a loss,

due to eddien, if trie water has much of a velocity of whirl,

at entrance to the runner.

TT-ie Worthington Company rate this pump at 450 gallons

per minute against a 135 foot head, nmning 1500 R.P.M., this

requiring ahout 32 H.P. This given as pump efficiencj?- of ahout

(48) per cent. Much "better economy was obtained at lower speeds

and higher heads, as is seen from the data, and it is probable

that if a speed of 1500 R.P.M. could have been reached, the pump

would have operated at economy, better than fifty (50) per cent,

and against heads above two hundred (200) feet.

A close approximation to t}ie actual loss in the machine

can be arrived at by means of the following analysis:

(a) Velocity head in suction pipe

~ hs ^ Vi^ , where v^ is the velocity in the suc-2g

tion pipe, calcula. ted from the discharge and area of cross-

section: (_j2 ) .

(b) Loss at entrance to suction pipe =s .93 vf- and

this loss may be taken as v-j ^ , since there is a foot valveSg"

on the suction pipe.

(c) The friction loss may be calculated from the general

formula for friction in iron pipes, which is:

10.

dl 2 g

(d) The velocity head in the discharge pipe r Yl^, -v?.^ >

2 gwhere V2 is the velocity in the discharge pipe, calculated from

a2.

(e) Another loss of head occurring is what might "be

termed the "machine loss", which is that due to friction, impact,

etc. inside the casing. The R.P.M. required to give a certain

discharge is given approximately/ hy the following equation:

n = K /h + a ,

where "n" is the R.P.M. of the pump, "h" is the total head,

represented by the sum of the gage reading and the distance

from the gage to the sump water level, "k" and "a" are constants

determined as follows: from the data in the tahles value's of

"a" are determined, for different discharges, using a constant

value of K for all discharges, which gives a constant value of

"a" for the different speeds and heads for any discharge. In

order to get at the "machine loss" in the pump, calculations

are made of the other losses, and their sura subtracted from "a"

in the above formula, this difference being the machine loss,

for the discharge considered. •

The machine loss iij then:

\ 2 E di 2g^2g/Taking account of all th^^e losses, that is, taking the head

pumped against in the formula for efficiency, as the sum of

and a , the efficiency is seen to be approximately^ 81*^

11.

, ..The 19ff not yet accounted for is

j owrnal and "belt friction, charged to the pimp..

BIBLIOGRAPHY.

Reference has "been made to the following:

Centrifugal pumps, TurlDines and Water Motors, "by

Charles H. Innes, M. A.

Proceedings of the Institiuition of Civil Engineers,

of London, 1878, col. 53, p. 249.

Treatise on Hydraulics. By Wm. C. Unwin.

EpFICIENCr

><

(0<S«0 t

«^ o

O^" •

CO t.(0

H

R5E

POWER I^, CD

MOb ^.

1

«5.

nCO

c

E

o

>

Q

V.'

><

V

r>

On

f

CO

(D

Ob t

—

-i

CD ^ \

cK0.00

Q- E

Q-(0

V

(Q

Q

(0 (0

r\\J

00

o(0

VJ

O(0

VJ V w VJ

o<Q

Discharge

c

j2 Eo

»^CO

t

K

t

CO toto

CO oCO CO

oCO

o

o ^ON

>^ k•

<r

•Si. <^

tu

lu

Ll.

1

h

"5

^2

o oCO

oo

oo

oo

oo

<u^ S^ o

o ^o o o

^5

5o

On ONo o

2 CDo

13.

CJ

2lU

IL

LU

i.

><

W t

t

^^w

<^

^'

CO Ci

^'

A^'

?

d:

o

I

~5

-a

X

V

\ CO (D «) <».

0)

\N

V

E

c2o

J-

Q

><

OS S o

Vrv

(0

\IN.

—

\

to

\\

\ \

R.RM£ Q o Q

OxQo\

Q Q oOS o

o ocv

Of\

y

<T

. c

6 1 \

Q\

v: •

OS•

o COto (*>^

Ui

LI

ll-

1

-J

o o OQQ

c o00

Q>

N,

^ w

o ^

o \ Q>

V.o\ N. N,

o 5Q) oo OS

o

z: o o

Ia;

Ooo

oz

PSO 0E)

1^

<

<

oCL

UCO

or

o

3

s-a

X5 On

o

Q

to

On

(0

to

o COo

2£

CL

o O N

<Xo

D o L

J2 i"a L

3 L

fry

v9

SI

oNO

os8>

Ov9

ID

JCD

<

I

-J

1^ 50^. o^

O oil C> On

O o o1-

o o\3 CO

yo

LJ

Ou.

Ul

>so'

^"S»

.

is- ^

or

GlUi

o

I

X

\r-

\ \ \\

n'r«i

E

c£

>^

D

ID

><

sD w

0^

V\̂'

CO

NO

CO

wrp>

h

«v

K CQ

ci

A.rk

Cl

-a

\S

Its

*fQ~\

(ft

N

w A

IIIvUv3

<\3 -NO v2

fs/

1^

i

DISCH

^ i

a.

"V\

"<"V -V.

<.

•v. 0^•

OS•

•

h-

lU

aj

On

w

0^

\->

wV.

M<

U.

/

li

« ft)

. « ft

Q

o\\, L V L V

s s V \

S s

->

n00 00

«'to

o2

liJ

a:(Li

oalu

<o

DC

o

X

(I)

><

3

XL'

>

C

J;

Q

J)

4 o

w

1-

o

CO

t

o

fo

"7

NO NO"

CO'

J:

00

O

CO

2q: E

a.

vo

o

<Io

GO CONO

OO

NO V9

IV-V. 1^

IV cv tQ

(

-J

o ^1

V VV V

i5 «o

n1

1-

CO QN On

<3N

oO

oo

o oo

MX

u

1

C><

N

K

1-

^<

t ^;

NO •4

<

IT

lU

Ou

X

X

V

CDCD' CO'

03

tS

/TV

K

E

>

LV>

CO* IN ^-

CO

P

R.PM,E3 \ \ •

—

woa:

<Xo

D

^ E S>s

CO CD (a T

. </

^» «> (A

•fY

•

CO

* ^. ^.

hlU

\

-J

«V*V So

. A•v.

• ^

Q«^ CD c^ e

\o2: S t CO

fl

18.

z

o

\L

iU

t1

CO

t

a:

I

0.

Id

^CCrsU

I

Is

1.

\ no'\ \ON

V

KOn

I

£3

a

v.-

>

- r>

.0

CO

Q

SQ

—

,

o

Sft:

c:

E\

r\ wW\,

o o o<^

IS

<Xo<n

O

o 1(ft (Q

^orfi

ON ON'I-

ON

.0-

t-

O Q.«

u»CO

^.00>3 ^.

u.

1

V-

u

-J

£^o o

CO OD•-^

OCO

oOn On

«o-f^ E

a ^

o o o o o

o 5

d\o

o o o o o o o oCD

CO ON1-

19.

o

o

cc:

o

UJ

<Io«0

v9

>

5

£

o

Q

E

CL

E"i

O

o a.

CO

oo

v5i

On

oOn

A

o

>0

t

oON

o

•s9

OOn

t

t

-SI

QOS

so

to

1-

oOS

NO

ON

oo"

o

to*

oOn

o

en

NO

<N<

•Nf

I

f-

^2

oO

oo O

Oo

o1-

oso

oss St

E

• «J

oOn On On

1^ CD

o

On

o5

Xo.

>

z

lU

CC

Ui

^

D.

^CC/-V

I

V

£

a

. 4)

Q

>CO

0'

1-

«)' K<^

r

cc

E3

OSr\

o\r\

\V.

On On

\s>

cc

<:

Iu

Q

. f

'J hON

tO

si'

\ o

ta

Ui

1

-J

6-P CO CO

O o

QOS

CO ON r>

^1.

" >oZDoil

r«

Of"A ON*

rV

cf

oCL

Ui

(/>

o

I

.0

~3

X

r\VJ

«^

On^*

VJ n LCv.*\

>

aI-'

I

Si

R.P.M.

E «0

CO

03(ft

Q OS QsO N

\

o

\

Discharge

O ite;

Of) w

O Q_

«^

<O

V-

lU

Ll

1

Hli.

_J

"5

-H

\, \>

On

r 1

5 CO

ON On

1

ON

s s1

<^

^ 5

Q<S

o2 -1 1-

'

—

Table

Lost Head.

Values far Consjoinis.

iv 1 ^ V*n ~I \ ^^t- 1

k n - K Vhtoi

f

77 1 O >1 X 77 l^h+To

77 IS-

3oo 77 zr

Zfo 77 n - 77 I/Th^

^ oo 77 4-r n» 77 yfvM^

77

77 7-r

77 n * 77 ^hfsr

71 n = 77 ihHos'

"V.

vS

^,00v»\

r

EQ

s£)

f>f\

VU \ A\

Q)

CL

t ON —Cn

OOn N,

i-—

I,

1

IS

a:

'JCD

v)

D15CHA

oo(*)

O Os9

to ON vj

lo <v

CQ ON

o

NO

o \^

On

Head

Ll

O4*

\oO

o

i3.

UJ

2T ^

if^ V 03

<o Xa

1-V

k

2 i

£ ?

>s

On

LU Zw o

2 o

u

•0 o

< (jj

2

u

*CO

Discharge

'i

ua.

<^

to

i.

Vo.

3o

CO

*

U. OF 1. S. S. FORM 3

Curve No. Z.

UNIVERSITY or ILLINOIS

Computed from ^M/Thf^ FoHMULA

H^AO - FEE TJ. _Lj L

, Of I. S. S. FORM 3

Curve Ho.d

CUF^VE No 4.

Curve No. 5. Z9.

c

>

oo4

oo^1

4

'UtU4 ^9

I.

««I

c

, OF 1. 3. S. FORN

CU RVE No. 6 .

Curve No. 7-31.

ir

<Xo

Q

oo

OuRV£ No.S . U.