DESIGN AND TEST OP A STEAM JET AIR EJECTOR

A Thesis

Presented to

the Faculty of the College of Engineering

The University of Southern California

In Partial Fulfillment

of the Requirements for the Degree

Master of Science (Mechanical. Engineering)

by

Philip James Charley

June 1947
 UMI Number: EP60470

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

UMI
Dissertation PyblisNng

UMI EP60470
Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code

ProOuesf
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106- 1346
 T h is thesis, w ritte n by

. . . . . . Pfelllp...^ames. ^I
under the guidance of h.Àz^.. F a c u lt y C o m m ittee,
a n d a p p r o v e d by a l l its m e m b e r s , has been
presented to a n d accepted by the C o u n c il on
G ra d u a te S tudy and Research in p a r t ia l f u l f i l T
m e n t o f the r e q u ir e m e n t s f o r th e d e g re e o f

Master of Solenoe (Mechanical Engineering)

D ean

Secretary

Faculty Committee

Chairman
 TABLÎî’ OF CONTENTS

CHAPTER PAGE

I G E N E R A L ...............................................1

Ejector defined ................................. 1

History of the ejector • • • « • • ........... 1

Uses of ejectors • • • • • • « ............... 2

Theory of a jet ejector .................. 5

Reason for a thesis on this subject ........... 7

II DESIGN OF THE E J E C T O R ............................... 9

Analysis of flow in an e j e c t o r ..................9

Design of the mixing tube .............. 11

Design of the nozzle • « • • • • • • • • • • • 1 4

Design of the secondary c h a m b e r .................17

III CONSTRUCTION OF THE E J E C T O R ........... 19

IV DATA DESIRED FROM THE S T U D Y ........................22

V RESULTS OF THE E X P E R I M E N T ..........................23

VI C O N C L U S I O N S ......................................... 25

VII BIBLIOGRAPHY .....................................26

VIII A P P E N D I X ........................................... 27
 LIST OP CURVES AND DRAWINGS

TITLE PAGE

Comparison of constant area-constant pressure mixing. • 28

Entrainment ratios for various positions of mixing tube 28

Vacua produced at various positions of mixing tube . • .29

Nozzle for steam jet air ejector • • • • • • • • • • • .30

Mixing tube holder for steam jet air ejector # .31

Nozzle holder for steam jet air e j e c t o r . ................32

Steam jet air ejector. ............. 33

Calibration curve for pressure gage. . . . . . . . . . . 3 4

Pictures of the ejector. . . . . . . . . . . . . . . . . 3 5

 EJECTOR DEFINED

An ejector is an apparatus which removes air, liquids,

or loose materials (such as sand, gravel, liquid clay and

chemicals) by means of the educive force of steam, compressed

air, or water. An injector is similar to an ejector but the

former is an induced current apparatus which forces feed

water into a boiler by means of a steam jet.

HISTORY OF THE EJECTOR

The ejector-injector has been used for many years

both in industry and in the laboratory. Probably the first

instance of their use was in 1859, when the invention of the

injector by Henri Giffard caused great interest. It seemed

impossible at that time that a jet of relatively low pressure

steam could force both itself and feedwater into the high

pressure region of the steam boiler. Even exhaust steam

could be used for this purpose. Though exhaust steam at

atmospheric pressure emitted to the atmosphere through a jet

would have no velocity, when emitted to a vacuum it acquired

a high velocity. The vacuum in the above case is obtained

by the steam condensing on contact with the feed v/ater.

The high velocity steam mixes with the feedwater, imparting

sufficient energy to it to carry it into the boiler. At

present it is possible to pump against a boiler pressure

of 150 p.s.i. with steam at atmospheric pressure.

 In 1884 Jacob Johann presented a paper to the Amer­

ican Railway Master Mechanics* Association^ on the results

of a test of a steam-jet air blower used in a locomotive

to increase the efficiency of fuel consumption and for the

abatement of objectionable smoke. He concluded that Its

use Was uneconomical as it did not noticably improve the

fuel consumption although it decreased the amount of cin­

ders and smoke. The necessity of constant supervision and

the loud noise attending its operation were so objection­

able that its use was not justified.

Since those times there have been great improve­

ments made and this type of equipment is now in wide use.

USES OP EJECTORS

Although this study was to be made on an air

ejector, it is well to consider both the injector and

blower since they are very closely allied. Injectors

are widely used in boiler installations and are manu­

factured by several companies. They are very inefficient

as pumps but all the excess heat in the steam used to ac­

tuate them is recovered in the feedwater and therefore

the inefficient pumping action can be overlooked. Their

1 E. D. Benton and R. B. Engdahl, "Steam-Air

Jets for Abatement of Locomotive Smoke," Transactions
of the A.S.M.E., Vol. 69, January 1947 p. 35
 3

use is especially indicated for small low pressure boilers.

The principle of the jet blower can be exemplified

in the operation of a gas burner on the kitchen stove,

or the chuffing of a steam locomotive. Nevertheless

the use of these blowers as such is relatively rare.

2
They may be used with compressed air as the actuating

medium to force air into a room for air conditioning

purposes or to supply air for fuel consumption purposes

(steam being the working fluid) as in the overfire air

jut. In the latter case they are particularly adapted

to the locomotive as the installation is small and com­

pact and relatively little duct-work is necessary. A

very desirable point about its operation for supplying

overfire air is that the steam has the effect of an air

preheatcr. The principal objections to their use as air

blowers is as follows:

(1) An excessive amount of noise is produced

resulting in discomfort to the operator.

(2) No automatic control exists so they ar not

used when most needed and it is difficult for the operator

to regulate them in accordance with burning rates or steam

demand at various boiler loads.

^ G. L. Tuve, G.b. Priester, and D. K. Wright, Jr.,

"Entrainment and J et-Pump Action of Air Streams," Journal
the A.S.H. V.E.. Vol. 13, November 1941 p. 708
 4

(3) Design data on their construction and data

on their use has been lacking.

g
Benton and Engdahl have attacked these problems

with promising results. By using varied insulating m a ­

terials and control systems and conducting numerous tests

a reasonably high state of efficiency has been reached.

Many railroads are now installing or considering the

installation of these blowers and it is thought that

eventually they may become standard equipment especially

on engines which otherwise would contribute to the smoke

nuisance of a city.

Applications of ejectors can be seon in most large

industrial plants where they are used on vacuum stills

and evaporators, condensers, de-aerators, filters and

dryers, to mention but a few. The ejector which is also

knoYvn as a vacuum pump, air pump, evacuator, exhauster,

etc: has many advantage over reciprocating, centrifugal

or rotary pumps. Its outstanding features are its sim­

plicity, compactness, ability to handle large volumes

at high vacua and above all, the lack of moving parts.

Purthemnore the ejector is low in initial cost, light in

weight, and requires no heavy foundations. Hov/ever it is

still very inefficient when compared to other types of

3 Benton and Engdahl, 0 £. cit., p. 35

 5

pumps and may be attended by the usual objectionable noise

unless steps are taken to prevent it. Even with its low

efficiency, because of its other advantages, jet type

pumps are rapidly replacing practically all other types

of vacuum pumps throughout industry, for both high and

low vacuum service.

THEORY OP A JET EJECTOR

The principle of a jet pump is very easily illus­

trated by fig. 1.

The steam, air or motivating medium is discharged

from the nozzle. A, at a high velocity

r
fig. 1.

Between the end of nozzle A and the beginning of the

mixing tube B, the motivating medium picks up or "en­

trains " the medium being pumped. Once the jet stream

impinges on the mixing tube wall enhainment ceases and

in a straight mixing tube, the mixture continues on at

constant velocity. For purposes of illustration, suppose

 ô

the motivating force to be steam at some pressure greater

than atmospheric and the medium pumped air* It should

be noted at the outset that, so long as the pressure in

the steam main exceeds that outside the mixing tube,

some air will be driven from the chamber C. The action

of the pump depends on the air being diffused into the

stream of steam and thus the apparatus could be called

a diffusion puijip (indeed in the case of a vacuum produced

by mercury vapor this term is often used). However, a

better name might be a steam-stream or vapor-stream pump

as this more accurately describes the action. As the

steam leaves the nozzle at A it forms a "wall" at an angle

to the centerline equal to the nozzle angle. This "wall"

is maintained until the stream of steam impinges on the

wall of the mixing tube B after which it follows the

contours of the tube. The mixing tube may be of cylin­

drical shape or may be in the form of a venturi or nozzle

itself. During the passage of steam from A to B, a certain

amount of air will diffuse into the steam and a certain

amount of steam will diffuse into the air. If the pres­

sure of the air is high (and consequently the mean free

path of the molecules of air small) collisions between

molecules v/ill be relatively great in number and little

diffusion will take place. This of course can be con­

trolled principally by the distance A-B and the

 7

temperatures involved, and also by the type of molecules

involved — in our case air. If however, the mean free

path of the molecules be large the air wèll diffuse read­

ily into the stream and vice-versa, only the latter will

be controlled somewhat by the great velocity of the steam

molecules away from the mouth of the nozzle. In any case

the diffusion of air molecules into the nozzle stream will

depend on the partial pressure of the air in relation to

the steam. Besides the above there is: (1) An acceler­

ation of the air particles by primary impact with the

steam particles, (2) Entrainment of secondary fluid by

viscous friction at the periphery of the primary jet and,

only in the case of a steam jet, (3) Change of state dur­

ing which a reduction of volume is made by condensation

and the energy of latent heat is made available.

REASON FOR A THESIS ON THIS SUBJECT

Although, as previously stated, ejectors have been

in wide use for over half a century, their design is still

largely empirical, and generally not clearly understood by

the engineering profession. Many firms manufacture ejec­

tors but their design data, because of the competition

is a closely guarded secret. Thus by constructing an

apparatus of this type, valuable data could be acquired.

After the data was obtained, the ejector could be used by

 8

undergraduate engineering students in order to become ac­

quainted with a type of apparatus which is not covered in

any other course.

Several papers have recently been vrritten on the

theory and design of the ejector, principally by Keenan

and Neumann^; Tuve, Priester, and Wright^; Gtoff and Googan^;

and Elrod . In most cases the approach has been theoret­

ical and in all cases where practical checks w^ re made,

the apparatus was of a small size.

For this thesis it was decided to construct an

apparatus of such a size to obtain large quantities of

air ejected at low vacua. Thus it would be possible to

check previous work while obtaining new data for a large

ejector.

4 J. H. Keenan and E. P. Neumann, "A Simple Air

Ejector," Transactions of the A.S.M.E. (J. of Applied
Mechanics ) Vol. 9, June"%9%5"P. A-75

5 Tuve, Priester, and Wright, 0 £. cit., p. 708

6 J. A. Goff and C. H. Googan, "Some Two-Dimen­

sional Aspects of the Ejector Problem, "Transactions of
the A.S.M.E. (J. of Applied Mechanics) Vol. 9, December
I9l2 P. Â-Ï51

7 H. G. Elrod, Jr., "The Theory of Ejectors,"

Transactions of the A.S.M.E. (J. of Applied Mechanics).
Vol. 12, Septeniber, 1945, F. A-170
 DESIGN OP THE EJECTOR

ANALYSIS OP PLOW IN AN EJECTOR

To obtain the flow through the nozzle the author

8
used Grashof *s empirical formula:

0.97
: —S?' ---- / A \I 1
S - .000651g) ------
1 . 5.6606ëT.

where Wg = weight of steam flowing, lb. / sec.

a r area of nozzle throat, sq. in.

p = steam pressure, p.s.i.a.

X s quality of steam if v/et, as decimal.

Tg = amount of superheat, deg. P.

This formula gives results which are correct to within

1.5 percent.^

Through the mixing tube, the principles of con­

servation of mass, energy, and momentum apply:

8 Marks '* Mechanical Engineers * Handbook 4th

Edition, McGraw Hill, 1941 p. 356

9 R. B. Engdahl and W. C. Holton, "Overfire Air

Jets," Transactions of the A.S.M.E. Vol. 65, October
1943, p7 741
 10

fig. 2.

Conservation of mass:

(2 )

Conservation of energy:

•» ’ ' 3^3 = «2 (^2 •» 2 ) (3)

2g
Conservation of momentum:

W-.V
1 1 + P3&2 = (^1 4. ^3 )Vg * p^ag - - - - -(4)
g S
where w = weight of flow, lb ./see*

a = area, sq* ft*

V = velocity, ft./sec*

V = specific volume, cu. ft./lb.

h B enthalpy, ft. lb#/lb.

g = acceleration of gravity ft./sec./sec.

p m pressure, Ib./sq. ft. abs.

velocity at exit of nozzle, ft./sec.
 11

These equations assume conditions which though not

strictly true, are fairly close. The steam is assumed

to be at zero velocity at the entrance to the nozzle,

and is assumed to expand reversibly and adiabatically to

the nozzle exit. The flow is very nearly adiabatic since

it occurs at a great speed. However the effect of wall

friction and supersaturation prevent the assumption from

being perfect.

Furthermore the secondary fluid, (3) fig. 2., is

assumed to expand reversibly and adiabatically to the

entrance of the mixing tube after which it mixes and pass­

es out the end of the tube at a uniform velocity, the ef­

fect of wall friction being neglected.

DESIGN OP THE MIXING TUBE

There are two means by which the steam, being dis­

charged from the nozzle mixes with and entrains the air.

The first case, fig. 3a., almost universally used in

commercial practice is mixing at constant pressure.

Pig. 3b. shows a tube in which mixing occurs at constant

cross-section. Keenan and Neumann^^ have shown (Graph 1,

Appendix) that mixing at constant area will give better

results than at constant pressure.

TÜ Keenan and Neumann, o£. cit. p. A-75

 12

a.

b.

fig. 3.

The reasons for this are as follows:

(a) If the mixing is done at constant area (fig.

3b) the pump can deliver air to a pressure (2) vdiich is

higher than the pressure in the chamber (5). In the case

of a constant pressure mixing tube (fig. 3a) the pressure

at (2) cannot exceed the pressure at (3) unless a diffuser

is attached to the end of the mixing tube.

(b) Keenan and Neumann have also shown analytic­

ally that when serving as a blower (pressure at (2)

equals pressure at (3) ) the ejector using constant area

mixing delivers a higher secondary fluid flow for a given

exit area.

An ejector could be made combining constant pressure

and constant area mixing but Keenan and Neumann state that
 13

constant area mixing alone will still give better results

than any such combination. The only exception to the

statements made above is when the ratio of the area of

the mixing tube to the primary nozzle is very small.

The effect of the venturi diffuser placed on the

end of the mixing tube is to gently decelerate the fluid

so that a maximum of kinetic energy at the "throat" is

converted to potential energy in the form of static head

at the exit of the diffuser section. Thus the ejector

can operate at an even lower pressure in the secondary

chamber (3). A diffuser was not used in this study however.

From the above discussion, it seems evident that the

tapering portion of the air tube shown in fig. 3a is un­

desirable, even though this part is present in all commer­

cial ejectors. It was therefore eliminated from the de­

sign of the ejector for this study.

The length of the mixing tube is of some importance

ll 12
and has been shown * to produce the best operating

characteristics when equal to seven times the diameter.

A marked drop in efficiency is obtained if the length is

less than this, but an increase in length has little

effect. It was decided to make the mixing tube ten times

TI Keenan and Neumann, o£. cit. p. A-75

12 Engdahl and Holton, o£. cit. p. 741

 14

the diameter as thus portions could be removed to check

the above statement*

The shape of the mixing tube entrance doubtlessIj

effects the flow, as it serves to guide the secondary

fluid which should flow with as little turbulence as

T "I
possible* kngdahl and Holton obtained a relative per­

formance of 100 for a bellmouth entrance and only 89 for

a sharp entry# With a conical entry the performance was

98* It is interesting to note that these performance

characteristics are in the same ratio as théir discharge

coefficients when serving as simple orifices* Due to the

ease of fabrication and high (98) performance character­

istic it was decided to use a conical shaped entry, to be

held in place on the mixing tube by means of steel clips#

DESIGN OP THE NOZZLE

The nozzle was designed with a throat diameter of

0.5 inches, as in none of the previous studies had such

a large nozzle been used. It was made converging-diverg-

14
ing as better results are to be expected with this type

as compared to a straight converging. As the erosive

action of the (wet) steam was not important, since the

13 Engdahl and Holton, 0 £. cit# p. 741

14 Keenan and Neumann, 0 £. cit. P. A-75

 15

operation of the ejector was to be limited, the body of

the nozzle was made from low carbon steel.

Since the quality of the steam was not known before

operation it was assumed, for design purposes, to be satur­

ated, as little error would be encountered if there was

a slight amount of moisture present. The actuating steam

pressure was taken as 100 p.s.i.g., an unfortunate assump­

tion, as it turned out, since a maximum of 92 p.s.i.g. was

obtained at the nozzle. This assumption was based on a

boiler header pressure of 105-110 p.s.i.g., which due to

a faulty gage, later proved to be Incorrect. It is doubt­

ful if there would be more than five p.s.i. pressure drop

in twenty-five feet of the two inch line vdiich was to

deliver the steam to the nozzle. Another unfortunate

assumption was that a maximum of twenty-five inches of

mercury vacuum would be obtained in the secondary chamber.

Actually the maximum obtained was less than four inches.

The flow #as calculated (using these conditions)

from Grashof*s formula to be 1170 lbs. of steam per hour.

Since the nozzle was to have a rounded entrance the coef­

ficient of discharge v/as taken as 1.0.

 16

I 2

fig. 4.

For saturated s t e a m , P ^ / P ^ = 0.575, so P^ = 66.0

p.s.i.a.

The enthalpy^^ of the steam at the various sections

15
corrected for friction losses v/as found to be as

follows:

= 1X89.6 Btu/lb.

hg : 1146.0 Btu/lb.

hg = 935.0 Btu/lb.

The critical velocity at the throat v/as calculated

from

V = 223.7 y - h.

and equaled 1480 ft./sec.

15 Kents* Mechanical Engineers* Handbook, op.

cit. p. 356

16 J. H. Keenan and F. G. Keyes, Thermodynamic

Properties of Steam, 1st Edition, John Wiley and Sons, 1936
 17

The velocity at section (3) v/as calculated from

Vg . 2 2 3 . 7 / (l-y)(hi-h 3 )
t
where y = ^3 - ^5
Kl - hg

Vg equaled 3420 feet/sec.

From the equations of continuity of flov/ the areas

of the various sections v/ere determined to be: Ag = 0.2

sq. in., Ag = 1.61 sq. in. These corresponded with

diameters of 0.5 in. and 1.433 in. Since it was decided

to use a nozzle included angle of 10^, this made the

length of the nozzle from throat to exit 5.35 in. The

upstream diameter was determined from the size of the

standard 1^ in. steam pipe used.

The nozzle was made with a tapering end section

so as to disturb the flow of the entrained fluid as little

as possible.

DESIGN OP THE SECONDARY CHAMBER

This chamber was rather unimportant in the design

of the ejector since it only had to contain the secon­

dary fluid (air) and provide a support for the nozzle and

mixing tube. It was built using standard eight inch pipe

as this was the smallest that could be used and provide

the necessary clearance for the mixing tube mouth. It

was built in three sections which could be designated;

 18

(1) Mixing tube holder (fourteen inches long), (2) Noz­

zle holder (seven inches long), (5) Main secondary cham­

ber (ninety-eight inches long), where the amount of air

flov/ing was to be measured. At the far end of the pipe

a flat steel plate was bolteb and this plate was drilled with

a series of different diameter holes so spaced that when

either fully or partially covered by a similar flat plate,

uniformly distributed air could flow into the chamber.

The amount of air flowing was controlled by this latter

plate which was fastened to the former at the center and

was free to rotate on this axis. Thus, by varying the

closure of the holes with this plate, any amount of air

could be allowed to flow. By using this system, flow

straightening vanes became unnecessary. At a distance

of ten pipe diameters from this plate a hole was cut in

the side of the pipe for the insertion of a pitot tube

rake. Two other holes were drilled and tapped to provide

for, (a) pitot static tube, and (b) mercury manometer.

 19
CONSTRUCTION OP THE EJECTOR

One-half inch steel flanges were welded to the three

sections of the secondary chamber. Care was necessary

that these w re welded normal to the axis of the eight

17
inch pipe as Engdalil and Holton proved that it was

important to have the nozzle centered with respect to the

mixing tube and these latter were fastened in different

sections of the secondary chamber. The flanges were held

together by six three-eights inch bolts. This small size

was used because the interior of the chamber was expected

to contain a vacuum and thus small stress would be placed

on the bolts. Asbestos gaskets were inserted between the

flanges to provide an air-tight joint.

A variable adjustment of the mixing tube with

respect to the nozzle was made possible by means of the

design shovn in Drawing No. 2. By loosening the four

nuts holding the outer plate, the tube could be moved

to any desired position after which the outer plate was

pulled up close to the inner plate. A seal was maintained

by means of the packing placed around the tube as shown

in the drawing.

The nozzle was held in place by a streamline fix­

ture welded into the eight inch pipe. This was made

Tt Engdahl and Holton, o£. cit. p. 741

 20

smooth to disturb the air flow as little as possible, and

was brazed to the If- inch steam supply pipe.

At the upper end of the 1-^ inch pipe (which was

sealed by welding it into the 8 inch p ipe) a hole was

drilled and tapped to take a bourdon-type steam gage.

A"pigtail" was placed between the gage and the steam so

the steam temperature would not affect the gage readings.

The If inch pipe was welded to the 2 inch pipe

which was the source of steam from the boiler. A stand-

18
ard A.S.M.E. sampling nozzle was v/elded into this 2

inch pipe and was attached to a throttling calorimeter.

It was believed that in the range of qualities likely to

be encountered the throttling calorimeter would give the

best results obtainable.

The pressure in the secondary chamber was obtained

by means of a mercury manometer - - by far the simplest

and most accurate device available. Atmospheric pressure

was obtained from an accurate aneroid barometer.

The measurement of air flow in the secondary cham­

ber was of considerable importance and it was decided to

18 Gebhardt, Steam Power Plant Enp;ineering, 6th

Edition, John Wiley and Èons
 21

utllize^^ a pltot tube rake* This type of measurement

should be used where the velocity of the air is fairly

high and also should be calibrated* However the velocity

coefficient is usually close to unity and with the eleven

different readings obtained by the rake It was felt

calibration could be omitted* Actually it made no dif­

ference as the results of the test were such that very

little air flow was obtained - - so little in fact that

no readings could be made on the inclined pitot tube

boal5d* It is thought that much better results would have

been obtained by the use of a sharp edge orifice at the

low flow rates encountered*

It would also be necessary, in a complete run, to

obtain the relative humidity of the air entering the secon­

dary chamber as this has some (though slight) effect on

the volume of air entrained by the jet# A sling pschyro-

meter was believed best suited for this determination#

19 Moyer, Power Plant Testing, 4th Edition,

McGraw Hill, 1934
 22
DATA DESIRED FROM THE STUDY

As has been previously mentioned, much of this work

has already been reported upon, though for small size

ejectors only. It would therefore seem desirable to

repeat the other experiments to find out if the conclusions

obtained then would still hold for large apparatus. It

was also desired to test the theory of Keenan and Neumann

to see if good vacua could be obtained v/ith a constant

area mixing tube#

The questions to be answered were as follows:

(1) What vacua are obtainable with various steam

pressures?

(2) What vacua are obtainable with various posi­

tions of the mixing tube relative to the nozzle?

(3) What is the optimum position of the mixing

tube relative to the nozzle?

(4) What entrainment (air-steam) ratios are obtain­

able for various vacua?

(5) What is the optimum length of mixing tube?

(6) Is the use of a diffuser section necessary?

 23
RESULTS OP THE EXPERIMENT

The results of this experiment proved very disap­

pointing. Only two of the previously mentioned questions

were answered, the data being so poor that it was not

possible to attempt fux'ther experimentation with the exist*

ing equipment. However some data v/as obtained which

checks previous studies.

One series of three runs w a s made to determine

the vacua obtainable with various steam pressures and

positions of the mixing tube. In each of these three

runs the results were identical and are shovm in graph­

ical form at the end of this report. It was found that

the maximum vacuum obtained with this design was 3.4

inches of mercury whereas 25 inches had been expected.

In the author’s opinion this is due principally to the

lack of a diffuser section on the mixing tube. Then too

the fact that the nozzle was designed for conditions other

than those existing would help cause this poor result.

It will be noted that for lov; steam pressures, for

which the nozzle was not correctly designed, the shape

of the curves obtained is convex upwards at the most ef­

ficient point. This is probably due to the fact that at

these low pressures (less than the design critical flow

pressure) overexpansion of the steam jet takes place.

As the nozzle exit is placed further within the mixing

 24

tube, this overexpansion has less effect on the entrain­

ment, indicated by the "flattening" of the curve. If

this "hump" in the curve is to be eliminated, the nozzle

must be designed for and operated at one specific steam

pressure.

Keenan and Neumann obtained a curve (Appendix,

Graph 2) showing the effect of distance of mixing tube

from primary nozzle exit. Although, as has been prev­

iously explained, no runs were made to determine entrain­

ment ratios encountered vs. positions of mixing tube, it

can be argued that maximum secondary flow would be obtained

at the same position where maximum vacuum was noted.

This curve then may be compared with that obtained in the

present study.

As can be seen there is remarkable similarity in

the curves so it can be stated that for large equipment

the throat of the mixing tube should be placed between

0.5 and 1.0 mixing tube diameters from the exit of the

primary nozzle. This result was obtained for smaller

apparatus not only by Keenan and Neumann but by Engdahl

and Holton and others.

 26
CONCLUSIONS

Although the results of this study were not all

that was expected, sufficient evidence has been obtained

that, with several changes, the existing apparatus could

be made to operate to better satisfaction. It is sug­

gested that a nozzle be designed which will fit the con­

ditions obtaining better than the present installation.

The need of a diffuser section is clearly evident.

As all commercial types of ejectors have either

constant pressure mixing tubes (or combinations of constant

area-constant pressure) it would be desirable to run a

series of comparative tests on a mixing tube of this type

also.
20
For future work the application of Elrod’s

"Optimum Area Ratio" should be used in the design of the

mixing tube. Also an investigation should be made on

methods of reducing objectionable noise when an ejector

is operated with atmospheric exhaust.

It is to be hoped that a continuation of both

theoretical and practical design of ejectors can be made

in the near future.

20 Elrod, 0 £. cit. p. A-170

 26
BIBLIOGRAPHY

Benton, E. D. and R. B. Engdahl, "Steam-Air Jets for

Abatement of Locomotive Smoke," Transact ions of
the A.8.M.E., Vol. 69, January 1947

Tuve, G. L., G. B. Priester, and D. K. Wright, Jr.

"Entrainment and Jet-Pump Action of Air Streams,"
Journal of the A.S.H.V.E., Vol. 13, November 1941

Keenan, J. H. and E. P. Neumann, "A Simple Air Ejector,"

Trgnsactions of the A.S.M.E. (J. of Applied Mechanics)
Vol. 9, June 1942

Goff, J. A. and C. H. Coogan, "Some Two-Dimensional

Aspects of the Ejector Problem," Transactions of
the A.S.M.E. (J. of Applied Mechanics) Vol. 9,
December 1942

Elrod, H. G. Jr., "The Theory of Ejectors," Transactions

of the A.S.M.E. (J. of Applied Mechanics) Vol. 12,
September 1945

Marks* Mechanical Engineers * Handbook 4th Edition, McGrav/

Hill, 194Ï

Engdahl, R. B. and W. C. Holton, "Overfire Air Jets,"

Transactions of the A.S.M.E. Vol. 65, October 1943

Keenan, J. H. and P. G. Keyes, Thermodynamic Properties

of Steam, 1st Edition, Hohn Wiley and Sons, 1936

Gebhardt, Steam Power Plant Engineering, 6th Edition,

John Wiley and Sons

Moyer, Power Plant Testing, 4th Edition, McGraw Hill,

1934

Keenan, J. H., Thermo dynamic s, John Wiley and Sons, 1941

Kuethe, A. M., "investigations of the Turbulent Mixing

Regions Formed by Jets," Transactions of the A.S.M.E.
(J. of Applied Mechanics) Vol. 57, September 1935
 27

APPENDIX
j— ke

DJStJ^OF. -r JE. ToSë€. 7h.8£.

ZdF r

E
n

%
 29

<4

a//rrpk/f
 30
r

—. 99V i

1 1 i

9b
:
%
C kj
K ï
o
Q '^'l

ro ;
K
2! >r e

I
■i:
5
I J
N 3Q
lï O

T r I
 51
N

i
>- I

j !
r

I
K

.w
I
H

I
I
I

n i D

o/
fV
 i
LU
W

Im'i

LU
co
 34

■s

ttfcrt

7%
 55

'Nt i

The ejector (steam line disconnected)

Close-up of ejector showing air-tube in place

 36

Close-up of ejector showing nozzle (air tube removed)

Side view showing pitot tube board