HYDRAULICS BRANCH OFFICIAL FILE COPY

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

UNITED STATES DEPARTMENT OF THE INTERIOR

BUREAU OF RECLAMATION

HYDRAULi C LABORATORY REPORT NO. i 50 I

CONT!HUATION Of STUDIES TO DETERMINE SUITABLE METHOD FOR STARTING AND STOPPING THE PUMPS

AT GRANBY DAM PUMPING PLANT

COLORADO-BIG THOMPSON PROJECT, COLORAuO

by

J. N, BRADLEY, ENGINEER

Denver, Colorado, Sept. 20, 1944

* * * * *

* * * * * * * * * * * * * * * * * * * *

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

.'A

•

•

UNITED STATES

DEPARTMENT OF THE INTERIOR

BUREAU OF RECLAMATION

HYDRAULIC LABORATORY REPORT NO. 150

SUBJECT: CONTINUATION Of STUDIES TO DETERMINE SUITi,BLE METHOD

FOR STARTING AND STOPPIHG THE PUMPS

AT GRANBY DAM PUMPII\G PU.1-'.'f

COLOR1�DO BIG THOMPSON PROJECT, COLOPJtDO

By J. N. BRADLEY, ENGINEER

Under Direction of

J. E. WARNOCK, SENIOR ENGINEER

and

R. F. BLANKS, SENIOR ENGINEER

Denver, Colorado, Sept. 20, 1944.

..

•

UNITED STATES DEPARTMENT OF THE INTERIOR

BUREAU OF RECLAMATION

Branch of Design and Construction Engineering and Geological Control and Research Division

Denver, Colorado September 20, 1944

Laboratory Report No. 150 (Supplements laboratory

Report No. 113) Hydraulic laboratory Compiled by: J. N. Bradley Reviewed by: J. E. Warnock

Subject: Continuation of studies to determine suitable method for starting and stopping the pumps at Granby Drun pumping plant - Colorado-Big Thompson project, Colorado.

1. Purpose of investigation. The tests described in this report

constitute a continuation of those previously described in hydraulic

laboratory report 113, 11 Studies to determine suitable methods for

starting and stopping the pumps at the Granby pumping plant," by J. N.

Bradley, Fred I.ocher, w. A. Morgan, and T. F. Hammett, June 15, 1942 •

As the pumping plant and the problems involved have been described in

the above report, these will be treated briefly here.

The plant ultimately will house four pumps, of approximately 6,000

horsepower each, which will be subject to discontinuous operation re­

quiring frequent starting and stopping. The starting or the stopping

of pumps of this size under full load would cause dimming of lights and

interference with the operation of automatic equipment throughout the

vicinity, as the rate of power input to the pump motors would, in some

cases, exceed the rate of response of the generators at Green Mountain

Dam. The tests described both here and in the previous report were per­

formed in an effort to alleviate this condition.

The purpose of the tests described herein was to study the feasi­

bility of starting the pumps while throttling the flow in the intake

lines. The main objective in this case was to eliminate the purchase

of valves for the discharge lines. The information desired is best

stated in a memorandum to the Chief Electrical and Mechanical Encineer

by Mr. I. A. Winter, 11 Proposed laboratory study of startine; the Gr9.nby

•

pumps while throttling flow in the suotion pipes - Granby Pumping

Plant - Colorado-Big Thompson Project," dated May 2, 1944. Investiga­

tion of the following points was requesteda

(a) Surg1ng,1n the discharge line when the inflow is

leu than the normal capaoi ty of the pump under heads rang­

ing from the minimum to the maximum.

(b) Observation of the tendency of the pump to oavi­

tate excessively during the starting oyole.

(o) Observation of the behavior of the water entering

the suction tube of the pumps at high velocities.

The initial pump house valve layout is shown on.figure 1, HYD-113.

For the valve and the pipe layout used in this second series of

studies, shown on figure l of this report, the discharge valve origi­

nal� used was eliminated and two intake valves were provided instead

of one. One valve was to be used for throttling and the other was to

provide for emergency closure. It was proposed to start a pump with

one intake valve closed and the discharge line full of water. With the

pump up to speed, the intake valve would be opened at a constant rate

until full load was imposed on the pump. The procedure was to be re­

versed when stopping the pump.

2. The test equipment. Arrangement of the test equipment for

these studies wa.s essentially the same as for the initial investiga­

tion, as is evident from a comparison of figures 6 and 7, HYD-113,

and figure 2 of this report. The same 8-inch vertical pump was used

in both oases. In the former tests a tank reservoir was employed to

supply water to the pump, while in the latter tests a booster pump was

connected in series with the intake line, ma.king p�ototype pressures

possible. The model approximated a general scale of 117 (table 1, HYD-

113) except for the discharge line. The confining walls of the labora­

tory restricted the height of the discharge pipe to one-third of its

scaled value and the length to only a fraction of the scaled distance

2

--...

l <t. �_to. 2.1+ $Q.�

� --�--- __ -- - - - •

� 0 '";.c

1�·-0·· .;:>4 ·

I :

- -------·- -- • •

,

----•-·--- g', --- _ _ . . ---- - - �

,- El g.3 G &, r

· Ii ; __ ·-'<P -r· -- -----·-- '· Pumprno P! +- ___ ---.i:'. __ . ___ .'0 __ 2._ ·

- ' P.- ..., I · · · 1 . -- - , on � -- -- 1

� £ -"" I \ ,·r ·

---,-- ,� ·-- · · · ' 00 ::l . • . . • ' j - . ./ / �1· - -- - -- ---- . : "- -,, ' • • • • •

• ' ,. , ../ •• ,

• ' • I

c:. t o 1 • , . • --r�- . · . ; ·.. ,

-,, --· . - ' J_ _ . . I· .. > . . . . . ' '

a . 2' -1---· ---l II .

. _- ,

1' l ' q_Max [L_i2�QO(Ji • -· ' ' -·(,[\,3290.00 l1 1l1

I

I

I

Supp-ressor 2.45 f P.>F .Zll

g'

;;;wJtY' - E�---'----t---- _-,- . '-:f �--·-----,--. Line continue$! :l

a

I

/

I

, , I n+ak.e

c-•··:e'··-�\\

2. 3

Gppf"Ol'-. Z1oOO ft. � , r \_ ·ro" I. D. steel pipe

- - I 1.D.Tunrie! ;a !

·Q -= 2.00 to 3-50 c.f s. I

;;;;-�---� 2.5 H, 27 _____ __,_ _______ ��-7. i 29

Valve. to be throttled/,

GRANBY P-UA

M Pl NG PLANT

SECTION TH.RU _INTAKE. UNE. 1 PUMP HOUSE AND

DISCHARGE. . LINE.. "TJ Q C ;o r1

MODEL TEST LAYOUT FOR GRANBY DAM PUMPING PLANT

t\'.l

to be traversed by the prototype conduit.

Instrumentation consisted of a voltmeter. an ammeter, and a watt­

meter connected as shown in figure 13, HYD-113, between a single phase

and the ground. Three electric, strain-gage type, pressure cells

(P. C. 4, P. C. 5, and P. C. 6) were located as shown on figure 2 of this

report for recording pressure changes in the intake. the pump suction

tube, and the discharge line. In addition to the above instruments,

electric tachometers were provided to measure the speed of the motor

and the rate of opening of the intake valve. Instantaneous records of

the pressure cells, the motor speed, the valve position, and the elec­

tric current were recorded by the laboratory oscillograph. An attempt

was ma.de to record the voltage, but, due to the fact that the oscillo­

graph contained no shielded channels, the 60-cycle frequency was super­

imposed on the record of every channel. After discovering that _the

voltage fluctuation amounted to no more than 5 percent, this measure­

ment was not considered important. A record of the power required by

the motor was also desired, but a watt element was not available for

the oscillograph. The primary interest in this study concerned surges

in power rather than actual values, and thes e are clearly manifest in

the current trace on the oscillograms. It should be kept in mind, how­

ever, that the general shape of the power record would vary considerably

from the current trace •. Although the voltage remained relatively con­

stant, the power factor du.ring these tests varied from zero to 85 per­

cent, as can be observed from figure 3. This curve shows the power

factor plotted with respect to the rated capacity of the motor for the

actual range over which it operated during these tests. In no case

did the motor on the test pump operate at greater than half of its rated

capacity.

The pressure cells were a type used for the first time in the lab­

oratory, and they will therefore be described here as a matter of rec­

ord. A cell (figure 4) consisted of two steel arms, one stationary and

the other movable. The movable arm was rigidly connected to a diaphragm

5

. .

• ' .

I

')

I ·.· J-�···

. L.'. 'f,QJ--....,.'-'--!-'+-+---,:--r---¥--,-,-�+---,---+----t-'=-+---,-,.:--,....t---'--...,......,....,.....

0.:

t­z 501-+--i--�+-------,l---+---+----,-l--'---+---+--_c...;....:..,-i-:--:--t----'--+--......,-,;�� Lu u � t..J .. Cl,,

. I :A�

,. I

.... -.... �--{-,-.........c.,.--- --�---·-·

_____ i .... ____,_

·10 20 30

PE.RC£-NT OF

ON 15 l

..

3" Zs 0. D. Cap----

f Pipe tap--'-

I I

',_ /" 0. D. tube 3/g" long

1 '--Gasket '--2B!eeders(#4-40) ASSEMBLY - SIDE VIEW

STRAIN METER UNIT - TOP VIEW SHOWING WIRING OF EXPANSION COIL

STRAIN METER UNIT - BOTTOM VIEW SHOWING WIRING OF CONTRACT/ON COIL

NOTES

FIGURE 4

Sealinq chamber filled with GE. No.227 insulafinq compound Sf rain meter unif filled with casfor oil leuvinq Zc.c. air space. Pressure chamber filled w/fh wafer before embedmenf. Sfrain mefer unit same as in Carlson stress mefer.

LABORATORY ELECTRIC HYDROSTATIC

PRESSURE GAGE

which flexed as pressure changes were applied to its opposite side.

Two coils of wire, one consisting of four loops and the other of five

loops, were strung over porcelain insulators mounted between the steel

arms. The ooils were strung to oppose one another in such manner that

a movement of the diaphragm would increase the tension on one and re­

duce the tension on the other. Movement of the diaphragm in the oppo­

site direction would reverse the strain, making the instrument adapt­

able for the measurement of intermittent, positive, and negative pres­

sures. Piano wire, 0. 0025 inch in diameter, possessing a fairly high

strain sensitivity, gave most satisfactory results of a number of wires

investigated. The dual coil served to equalize the variation of resis­

tance produced in the two wires by temperature changes. Upon final

assembly, to limit temperature change the coils were encased in oil.

The output of these cells was sufficient to make possible the elimina­

tion of vacuum-tube amplifiers at the input to the oscillograph.

3. Test procedure. Referring to figure 2, the first test was

performed with the 2-inch bypass valve open, the 8-inch discharge valve

open, the 8-inch intake valve closed, and both lines filled with water.

The booster pump on the intake line was started, thus producing a defi··

nite prototype pressure on the upstream side of the intake valve. Then

the test pump was started, ai'ter which the 8-inch intake valve was

opened at a steady rate. With this valve �ully open, the 2-inch bypass

valve was closed, completing the starting cycle. The stopping cycle

constituted the same procedure in reverse.

It was desired to investigate this procedure of operation, both

with and without aeration. Air could be supplied to the suction tube

of the pump through the 2-inch vent, shown on figure 2. This would

correspond to approximately a 14-inch vent in the prototype. A pipe

of the same size (figure 2) served as a bypass between the suction and

the discharge lines of the pump. The purpose of the bypass in these

tests was to provide circulation for cooling while the intuke valve

was closed.

8

•

Three variables were involved in this method of o�eration: (1)

the static pressure on the intake line, which varied from 10 to 120

feet in the prototype, (2) the rate of opening and closing of the in­

take valvea and (3) the area of the air vent. As stated previously,

the discharge line was not to soalea however, the results showed this

factor to be unimportant in this case. Testing was performed over

the complete range of the three variables. In compiling the results

it was found that two tests sufficed to show that throttling of the

flow in the intake line was not a desirable method of operation.

4. Starting and stopping oyoles with air vent open. The oscil­

logram on figure 5 shows the conditions in the starting and the stop­

ping cycle with the 2-inoh air vent (figure 2) open to the atmosphere.

With the test pump at rest, the pressure on the intake line (P. C. 4,

figure 2) was 83 feet of water, the suction-tube pressure (P.c. 5) was

10.5 feet, and the pressure on the discharge line (P.c. 6) was 9.5

feet. The starting of the test pump is indicated by the current trace

on figure 5 and also by the motor tachometer which registered each_

revolution of the pump. The frequency of the osoillograph galvanom­

eters was not sufficient, upon starting of the motor, to produce an

accurate traoe of the current from the standpoint of amplitude.

Approximately one seeond after the test pump was started (figure

5), a positive pressure surge amounting to 32 feet of water occurred

in the discharge line, which was likewise reflected in the current

trace, the latter registering a surge of 36 to 23 amperes. Although

negative in sign, a corresponding pressure surge of 12 feat of water

was recorded in the suction tube of the pump. The seoond surge, the

peak of which oocurred at a valve opening of about 14 percent, regis­

tered a positive pressure of 26 feet in both the suction tube and the

discharge line. As the intake valve was opened at a uniform rate,

repetition of the surging occurred in the discharge line and the suc­

tion tube of the pump until the intake valve reaohed approximately 40

percent open. For valve openings of 40 to 100 percent the pressures

9

FIGURE 5

� "' I.

i,t ,- Intake Line PC.4

•.• -Discharge Line P.C.6

C: <I)

_,__,,__ , � .-Motor Tachometer ·-�-12 .... , ..

73'

,-P. C.4 25'

-- P.C.6

ii I 1d11-111)IH• ""ii . .-J...-

rirt' ('l"l"/if\r 1 /"NV"f'f'll'V'\ty'(" .fil'ff'v'/i1v'l/ifV\l"f\lVY'V'.fifV\1'l/i/il"i'V'V"'/ifV\1"fV'\rY'"l'VVi/if'l/'·l"l"r f' � 11 l"f' /" /" /" l"r/" I" 1 rl/' r rf'l"f'f'1rf'rr/"/"f'rrr/"f'f'/"f'

I\ ln;okeVolve Position lndicctor··, ( . i \ f\.r V V I V \_ "-r- � ,,---. v-._f' �.

i5 l�

�---------5.2 --- • :,. 10 _ 15 20

INTAKE VALVE OPENING IN PERCENT STARTING CYCLE

,111\l1il i\111 l 1li j1 Ll 11 ld/nl 1 \ 111 1 \111 1li 11 ill 11 1 1ll11 .iLi11\111 l1\ n1d1 11 1\1,1 .\111 111111d 1 1 1\111.1Pd ,i 1 1 , ,i1 1 1 ,\1 1 1 I\ 11.,( 1 1 d 11l,L11L\ 1i11\l1ildl 1 i1\11d,1 1o1,L1 l,\,1d111111,lli 11\1u 1L1,l,l11, ,ll111iL.i1,L111 1ll11 1ni1, 1 111 1 , LI 1 11 1 1L.....__i.1 • 1 11, d. 1 d, ,,L

1e' ___ ?C.6

I IJ 1�1 i I Ill I .j 11

P.C.4 .!.,'

� l'l'l"f"I" r I l"rf\r r l"l'rrl"' r r 11 l"r rl" rv'l"l"l"·r 1r'r'firrr I" rf' f'v I" f"f"'l""l'l"r /" /"1-I" /"f"'r IV"' /irl" /"/"I"/" l"f' /" l"l"v l"f'l"r, /"l"rl"'r f' f' I"' r r v r f"'r r r r r r r,, r r r, r, �,�V--\.,_/"' I vr v--�,.rrv·· v v "'),,----Y------v----\r I \.,r r-"' r-"'/j "'-� � � r

35 40 4 5 50 55 60 65 .. ,

- -P.C.6 ·9'

I • I 11· • j ' I� I I I I J I f I I I I 11 HIii io11 1,1 1 11Jldl1'11dld.1 1l

INTAKE VALVE OPENING IN PERCENT STARTING CYCLE

I .. I I I . I I I I . I 1 dl 1 1i l 11 ,.�,L .. ,1 il,1,••1,11\ 1 l 1 !1 t •• .. \ 1i ,' 1i �11 11• I l11111lill1 1 I II 1 ii1h,.\l1 ,11 l 1 1ii 1l111l11111i11d11,l111. i 1_ili 11,1 1 l 1i 1i 1 11 i .i\l 1 , 111 1 id1d11ll�l1ddjj,IIJlll1i\l11111 1!11! 111 IIIIIIHl 1 i1 ldl ,,tl,11 1111

-------------------->

_Q_4 -

P.C.4-

Ii. li11 Ill 11_11; i1 ,:.� I ,,11, 1 11 •

,, I I Ir' r" r" /" r' I I I Iv' r r'r r' I Irr r' l""r\f'\r I IV"' r /"' /""/1f'\/"r1f"' r' If"' I f'f'rV .r' f"' I"" f'\r /"'"'r I' I I' f"f"'r I r'v' r·f'\/"'•r r' f'I r' f"/""' I I r'r'fit I'( I I f'V"' f"' f'\r Ir ff I Ir I Ir

"'-1 -- --- ,,... I\" v--v--v--1 \,�'�'v\.ir V V I V \"-r--'v-----y---,.r \"-,,-- ·I

45 40 3 5 3 0 2 5 2.0

\ � v \·

50

�---------- INTAKE VALVE OPENING IN PERCENT STOPPING CYCLE

82' .-P.C.4 :nroke Line

... 0 ... 0

.-?.C.6 o·schorgeLine

1��������/""�r���������l"'��N�-�C�r.rnn, , ,, �V--·v 1,,r' l \,,-V \I"" 0 \. r I \ . r! I \ r ' . \ '0�:"lr_.-.ii;.f�I 1111111•�-i--1 �r-..... ,-. _...,.... _ _._ ___ � '\.· · I ' > I\..., '- '--� 'Int-eke \/olvelFt>sition Ind1ca1or

15 10 5 O

( Ir-' r-: -

INTAKE VALVE OPENING IN PERCENT STOPPING CYCLE

GRANBY PUMP TESTS

OSC ILLOGRAMS OF STARTING AND STOP PING C YCLES WITH SUCTION TUBE VENTED AND BY-PASS OPEN

T H RO TTL ING ACCO MP LI S HE D WI TH INT A KE VALVE

at the three cells became quite steady with little change in magnitude.

For this reason, the oscillographic records have not been reproduced

for the complete starting and stopping cycles. The lower portion of

figure 5 shows the record of the pressures and the current for the

stopping cycle. In this case the operating procedure was the opposite

of the starting cycle. The pressure surges for this case were not as

severe as for the starting cycleJ however, distinct surges in the cur­

rent occurred throughout both cycles. For the valve partially open,

surging was audible in the model as air was intennittently pulled

through the air vent pipe. In no case was the flow through the vent

steady.

As a means of demonstrating that the rate of opening or closing

of the intake valve had little bearing on the pressure fluctuations,

a short oscillogre..m was made with the intake valve held at a critical

operating point. This is shown as the lower record on figure 6, which

indicates pressure surges in the discharge line of 25 to 5.5 feet of

water, with corresponding current surges of 40 to 25 amperes. Each

surge was accompanied by an audible intake of air through the vent pipe.

The condition shown on figure 6 persisted, with little variation for

valve openings of 5 to 30 percent.

With pressure-surge indications such as these manifest in a model

havinE a discharge line much undersealed, it is certain that pressure

changes in the prototype would be more severe than those indicated by

the model. As the pressure changes are accompanied by corresponding

fluctuations in the power required by the motor, the above plan for

starting and stopping the Granby pumps is not considered feasible.

5. Starting and stopping cycles without aeration. The oscillo­

gram on figure 7 is a record of the instantaneous pressures and the

current obtained for operation similar to that above-described except

that the valve in the air-supply vent was closed. With the pump at

rest, the pressure in the intake line (P. C. 4) was 82 feet of water,

the pressure in the discharge line (P. C. 6) was 9.5 feet, and the

11

Intp k e ]L i ne ] P.c,-. - ,

O S C I LL O G R A M O B TA I N E D AT C R I T I C A L O P E R A T I N G P O I N T - 1 4 P E R C E N T V A L V E O P E N I N G

S U C T I O N T U B E N OT V E N T E D - B Y - PA S S O P E N

O S C I L L O G R A M O B T A I N E D FO R C R I T I C A L O P E R A T I N G Z O N E - 5 TO 3 0 P E R C E N T V A L V E O P E N I N G

S U C T I O N T U B E V E N T E D - B Y- P A S S O P E N

G R A N B Y P U M P T E S T S

T H R O T T L I N G A C C O M P L I S H E D W I T H I N T A K E V A L V E

"Tl

G)

C

::0

ITI

en

1 1 I l r ' l l 1

1,

�-------

3 5

- tr.tok e L i n e P.C. 4

I N T A K E V A L V E S T A R T I N G

I I I

,. Di scharge Line F'.C. 6

F I G U R E 7

_#· - ;;.c.6 17

, - -P,C .4

' -?. C.5

1/YV"'�����l\l'VVVVVVVYVVV''�IV''IVVV'\IV\

�r----·,.r-rv�-�1',\f/4 I

1 \r I Y�'rt/�r-r�r 20 , _ 25 • 30

16'

I N P E R C E N T C Y C L E

;!l I N T A K E V A LV E O P E N I N G I N P E R C E N T

S T O P P I N G C Y C L E

G R A N B Y P U M P T E S T S

-------'1

S T A R T I N G A N D S T O P P I N G C Y C L E S W I T H O U T A E R A T I O N - B Y - P A S S O P E N

T H R O TT L I N G A C C O M P L I S H E D W I T H I N T A K E V A L V E

pressure in the suction tube of the pump (P. C . 5) was 10 feet. Upon

starting of the motor, the pressure in the sucti on tube dropped to a

negative 22 feet of water, while the pressure in the discharge line

fluctuated from 3 to 28 feet. A corresponding positive surge was re­

flected in the suction tube of the pump. As the intake valve was un­

seated; a ioud crackling noise occurred which persisted, although de­

creasing in intensity, unti l the valve was about 20 percent open. The

crackling noise was cavitation produced by the high-veloci ty jet shoot­

ing along the invert of the suction tube. For the smaller valve open­

ings, water was supplied to the test pump at a deficient rate, thus

produci ng an initial negative pressure in the sucti on tube . In addi­

tion, the high-velocity jet issuing from under the gate leaf tended to

skip over the invert of the suction tube because of gate-slot interfer ­

ence and imperfecti ons in the surface downstream, producing localized

negative pressures. The addi tive effect of these two actions caused

the pressure along the suction- tube invert to be reduced to the vapor

pressure of water, thus developing oondi-tions conducive to cavitation

although P. C. 5, located on the side of the suction tub e, did not indi­

cate pressures this low. As the intake valve continue d to open, the

static i ntake head dropped to 15 feet of positive pressure at 50 percent

valve opening, the pressure in the pump suction tube reversed to

slightly positive, and the pressure in the discharge li ne increased

approxi mately 7 feet over the original pressure in overcoming fricti on

caused by flow in the line. The oscillogram reproduced on figure 7 was

out at a valve opening of 50 percent , as pressure fluctuations beyond

this point were insignificant.

The pressure record for the stopping cycle is shown in the lower

portion of figure 7. As the valve opening reached 10 percent, the

pressure in the pump suction tube fell to a negative 28 feet, or the

vapor pressure of water at this altitude. The low pressure was accom ­

panied by a crackling noise which is indicative of the presence of

cavitati on. In addition to the crackling noise , the fluid passing

14

•

through the pump wae a mixture of air and water, the air being released

from the water by vaporiz ation. This air, as demonstrated in the pre­

ceding s ection, would be very undesirable in the prototype . Following

complete closure of the intake valve, a series of surges occurred in

both the dis eharge line and the auction tube of the pump, but little

fluctuation was evidenced in the current trace . Upon disengag�nent of

the pump motor fro m the power sour ce, water hammer developed in the

discharge line and the suction tube. This also would be undesirable in

the prototype when one considers that extremely long lines are involved.

Figure 7 i s interesting in that the current i s steady throughout both

oyoles in s pite of the pressure conditions .

A short osci llogram was made of the most critical operating point

for thi s condition of operation, which appeared to be at a valve open­

ing of approximately 14 percent. Although the suction-tube pressure

was negative 24 feet of water, the pressures in general were steady,

whi ch demonstrates that in this case the rate of opening and closing of

the intake valve would not material ly affect the pressures o-f figure 7.

Nevertheless, this type of operation i s not recommended for the Granby

pumps.

To remedy the above operating disadvantages, i t is suggested that

the throttling be accomplished by a valve placed in each dis charge line

as proposed originally . It is felt that these valves will more than

pay for themselves in reduced maintenance costs and fewer operational

difficulties, to say nothing of the advantages to be gained in the re­

duction of power and pressure surges during the starting and the stop­

ping cycles.

6. Throttling with discharge valve. Since the osoi llograms shown

in HYD- 113 were ma.de, the laboratory oscillographic equipment has been

greatly improved. For thi s reason the test shown on figure 18 A and B

of that report was repeated. In that test the pump was brought up to

speed with auxi liary water jets, whi le in the present tests the pump

was started electri cal ly, with the impeller rotating in air . Referring

15

to the photograph on figure 2 and the oscillogra.m on figure 8 of this

report - with the 8-inch discharge valve closed, the bypass valve open,

and the intake valve open, compressed air was fo rced into the suction

tube of the pump, thus unwatering the pump impeller . The test pump

was then started in air, electrically, at which point the oscillographic

record on figure 8 begins. The entrapped air then was expelled steadily

through a small valve on the discharge side of the pump, causing gradual

submergence of the impeller . The timing of this procedure is recorded

on figure 8. Upon completion of the air removal, the sudden rise in all

three pressure curves indicates the po int at which the test pump became

fully primed. The stopping cycle, shown in the lower part o f the figure,

was the reverse of the starting cycle, and, in this case, the pressures

for both cycles were similar. The stopping cycle consisted of closing

the di scharge valve at a constant rate. Upon complete closure, com­

pressed air was forced into the suction tube of the pump, unwatering the

rotating impeller at a steady rate. When fully unwatered,· the motor on

the test pump was deenergized. Pressure fluctuations can be controlled

to a large extent, in this case, by decreasing the rate of opening and

closing of the discharge valve and decreasing the rate of intake and ex­

pulsion of compressed air to the pump chamber. In the case of the model

pump, the operation was carried out at a moderate speed in order that

the reco rd wo uld not be to o long and cumbersome to reproduce. There is

no lower limit to the speed of this proces s in the prototype.

In addition to stabilization o f pressures, this metho d of operation

has the outstanding advantage o f requiring less power when starting and

stopping. Figure 8 shows that it required 20 amperes to turn the pump

impeller in air against 30 amperes to do the same thing in water ( fig­

ures 5 and 7) . Starting the impeller in air should therefore reduce

the initial power demand by perhaps one-third. However, no conclusions

can be drawn from the instantaneous current trace at the beginning of

each test, as the frequency o f the galvanometer elements was insufficient

to reco rd correctly the larger amplitudes.

16

I 1 1111 111i,1:i,1 1 I I !' I I I I I 1 · 1 I I I

I : :' :'· : ! I : : I : ! 01",j,,.:, , ,:. ,: , :, : : :, . . j::'f .'P :: -·: : : '': : ··� : : .. : : : : �= : : I I I I I :;

' I

F 1 G J F- � 8

=-= === -

Air E x liausl ing from Pum_e. \ia l u't e

1'0

S T A R T I N G C Y C L E

S T A R T I N G C Y C L E

D I S C H A R G E VA L V E O PE N I N G I N P E R C E N T S T O P P I N G C Y C L E

G R A N B Y P U M P T E S T S

S T A R T I N G A N D S T O P P I N G C Y C L E S W I T H P U M P I M P E L L E R R U N N I N G I N A I R - B Y- PA S S O P E N

T H R O T T L I N G A C C O M P L I S H E D W I T H D I S C H A R G E VA L V E

r r+-v,r-- ,r--,iv--

-,\ 15 2D

D I S C H A R G E V A L V E O P E N I N G I N PERC E N T

S T O P P I N G

----------;,.j

•

"

7. The bypas s valve. Although this subject was discus sed in HYD-

113, it will be mentioned again fo r emphasis. The bypas s is an es sen­

tial factor for smooth operation in the s tarting and the stopping cycles.

During the starting cycle the bypass makes pos sible circu lation for

cooling. During the stopping cycle it serves as a drain for water

trapped on the discharge side of the pump and as a regu lator for con­

tro l ling the rate of unwatering of the impel ler. Without the bypas s it

is not pos sible to relieve entirely the pressure developed on the dis­

charge side of the pump, which, if not relieved, wi ll reflect in surges

and an increase in the pow.er requirement to the motor previous to dis­

engagement.

Provision of an 8-inch pipe with gate valve should be co nserva­

tively adequate for the bypas s in the prototype. The valve can be set

to unwater the impel ler at the desir ed rate on the shut-down cycle.

Upon determination of this point, a stop or limit switch can be in­

stal led on the bypas s valve to limit its travel . It wi 11 be neces sary

to determine in the field the maximum valve setting, together with the

compr es sed-air pres sure and the rate at which air should be admitted to

the suction tube of the pump.

B . Summary and conclusion s . From an hydraulic standpoint , throt­

tling with the intake valve is fundamental ly unsound in that extremely

low pres sures are created unneces sarily . In high head design this is

to be avo ided wherever pos sible, for past exp erience has shown the re­

sults to be either unsatis factory or unpredictable. The application of

venting to relieve the condition, unneces sarily created by low pres­

sures, merely complicates the problem in this ease, as a mixtu re of air

and water is an unstable fluid and pres sures throughout such a fluid

wil l be found as unstable as the fluid itself. In the case of the Granby

pumps, whe re surges in pres sure are directly affected by the power re­

quirements of the motors, neither cavitation nor venting should be tol­

erated, as both produc e an unstable fluid.

In conclusion, the fo l lowing recommendations are suggested in

17

.,

..

•

in connection wi th the starting and the stopping of the Granby pump s =

( a) Throttling of dis charge through the Granby pumps

by use of valves on the intake lines is inadvisab le .

{b) The most satisfactory method , to date. constitutes

$tarting and stopping the pump impellers in air and thr�t­

tUng the flow with dis charge valves, as di1 eus sed in sec­

tion 6 .

( c ) Instal lation o.f the bypas s ShQUld not b e over­

looked when a valve is used in the diecharge line , as it pro­

vide s a means of coo ling the water by eiroulation duriIJg the

period that the discharge valve i s closed and the impeller is

turni� in wa't;•r . Also , it wi:11 se"e to 00ntro l the rate at

which the power deorease s d.uri� the stopping cycle .

( d) The compressed-air intake po rt functi ons best when

located midW$Y in the aucti on t.ube of the pump. The air­

exhaust port should be located on the dis charge side of the

pump, preferably in a small dome on top of the pipe where

water wi ll not interfere with the escaping air •

18