U. S. Department of Commerce National Bureau of Standa rds

Research Paper RP2035 Volume 43, November 1949

Part of the Journal of Research of the National Bureau of Standards

Laboratory Flow Tests of Fixed Spray Nozzles with Hydrocarbons and with Air

By M. R. Shafer and H. 1. Bovey

The metering characteristics of fixed spray nozzles of the type used in some t urbo-jet

engines have been investigated. Some of the nozzles supplied by the Navy Depar t ment

co ntained burrs, metal particles, and improper machining, which ca used erratic fluid me ter­

ing. After be ing reco nditioned, a g roup of 26 nozzles was flow- tested wi t h fiv e differen t

fluid s to dete rmin e t he effects of fluid density, viscosity, and suppl y press ure upon the rate

of discharge of t he nozzles. The res ul ts indica te tha t it is impracticable to correct for

differe nces in t he physical pr operties of the tes t fluid . A co mparative method of nOII' ­

Lesting fi xed nozzles with a ir is described. AILhoug ll Lhis method leaves mu ch to be des ired,

it appears use ful for safe ancl rapid sizill g of nozzles to within ± 3 per ce nt of Lheir act ua l flow .

1. Introduction

For seve ral years the Bu reau of AeronauLics, D eparLment of Lhe Navy, has sponsored aL LllC National Bureau of Standards a program of [esLing and research on devices fo r handling and metering fuels for aircraft. A recent phase of Lh is p rogram is concerned wiLh the flow characLeris tics of fi xed spray nozr, les oJ the type thaL have been used in some turbo-jet engines. Depending upon ll le sup­plier, Lhese may be' des ignated as Mo nard l nozzles, Rago nozzles, eLc., and for the presenL pu rpose all arc essentially alike.

For the present work, three sets of s ixty nozzles each were procu red. Each seL bore a colo r desig­nation indicating that all of the nozzles of each set had been matched in flow to within ::1:: 2.5 per­cent a,t a pressure of 100 Ib jin.2

As a first step, all nozzles were flow-tested as received with Varsol at five different pressures. It was found that a large proportion exhibited significant changes in flow characteristics before and after being tested at a pressure of 250 lbjin.2

Subsequent disassembly showed that these changes were caused by bUlTs, metal parLicles, and poor machining.

From the lot of 180 nozzles, 26 were selected, cleaned, recond it ioned , and assembled for usc in the remainder of the LesLs. The How characteris­tics of Lhe 26 have remained eonsLant over several

Flow Tests of Fixed Spray Nozzles

months. In order to study Lite eHects of Lll e density and v iscosity of Lhe test fluid , they have beon tesLed repeatedly over Lbe press ure ra nge of 5 (,0 250 Ib / in. 2 with Lhe following liquid hydro­carbo ns : Varsol, pure n-!tepLane , Apco- 467 oil , a comme rcial mix ture of isooctanes, and Sol trol- l 00 .

In add iLion, a n aLLemp t has been mad e to develop a method using air, instead of a flammable hydrocarbon, as Lhe LesL flu id. This reporL pre­sents Lhe res ulLs Lhat have bee n obtained to elate with the s ix clifrerenL Les L media .

II . Description of Nozzles

As shown in figure 1, the nozzles co nsist essen­tially of a body, an inse rt, and a st rainer. They arc designated by Navy Pa r ts List No . 14G320- 4.

After passing through the s trainer, the fuel is directed by tangential slits in the insert into the swirl chamber formed between the end of the insert and the body. Th e kinet ic energy of the fuel in the swirl chamber is effective in atomizing it as it escapes Lh rough the orifice in the body. Th e dimensions and relative locations of the tan­ge nLial slits, Lhe s" 'i1'l chamber, and the orifice delerm in e the press ure-flow characteristics, the spray angle and distribution, and the drizzle point of the nozzle. T he latter may be defined as the lowest fuel pressure at which the nozzle produces

449

a spray, and below which it discharges large drops or a stream of fuel.

III. Flow Tests with Varsol of 180 Nozzles as Received

Figure 2 is a diagram of the apparatus used for flow-tes ting the three sets of 60 nozzles each, as they were received . The Varsol was circulated by a pump through a low-pressure circu it from a storage tank, through a heat exchanger, and back to the tank at a constant rate of about 200 gal/hr. Fuel to the nozzle was bled from this line, and passed through a Rotameter to a second pump having appropriate valves in a by-pass and in the discharge line. From this pump the Varsol passed through a 10-micron filt er to the fit ting bearing the nozzle. Fuel pressure was measured at this fit ting.

The Rotameter was calibrated with the fluid used for the tests and at the temperature of the tests . Observed values of flow are believed ac­curate to within at least ± 0.25 percent. The pressure gages were also calibrated at intervals and are believed to be accurate to within ± 0.5 lb/in. 2 over the range from 9S to 250 Ib/in2 •

FIGURE 1. Spray nozzle, parts list No . 14G320-4 .

450

The ISO nozzles were tested as received with Varsol having a kinematic viscosity of 1.102 centi­stokes and a specific gravity of 0.779 at the test temperature of SO o F. The flow of each was measured at the following pressures and in the order stated: 9S , 150 , 250 , 50 , 9S, and 5 lb/in.2

gage. The results, except for the initial valu es at 9S lb/in. 2, are shown in figures 3, 4, and 5 for the sets of 60 nozzles color-coded red, green, and purple, r espectively. The flows shown at 9S lb/in. 2 were observed after the nozzle had been subjected once to the higher press ures.

PRESSURE

THERMOCOUPLE _ GAG7 ,

/cow PRESSURE PUMP

NOZZL~9

RETURN TO TANK

FIGURE 2. Schematic dia(fmm of nozzle fu el test.

72r-~-r~--r-~-r------'--'-'--'-'--'-'1 I I 250 psiq 0 6

70~+-~----~+--r-'~'-+--r0-J~_-A~~r~-,~~~ ~~3~~~~~~

68 o_0.,g.. 0 __ j-ot>~ rr~ 0-0"", o"TJ: - -: 0 0 _/'" ,/

~ -641-i'r."'-,~ I.c....o _'*_ ~=F,,-+--j--+-+-+--·--:LT I-f---

°r1 ....... 62W-~-L~ __ ~~-L~ __ L-~~~ __ L-~~~

o 0

5

5

50

4

1 150 psiq

2 - D 0 00 1"01 00 0 00 00 0 0 ° 0 0 0 0

00 'I I I 8

0 0

I 3

30

29

28

10.0

9.

9.0

j-----0

Oo~ 0 0 0 o 0

I

00

*' , ~ , 50 psiq , I

0 0 000 )00010::1 00 1 0 0 0

I I I I

I I -.L 5 psiq

9 0 0 00 0 + oJ oo .to 0 ooJ 0

oj ooLobooO I I

I I

I 0 0 40

0

0 0 0 o I

0

I I

I 0 0

ooo~ ooo 0 0

I 5~f~

000 i 0 I I I I 0

8 .50 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 NOZZLE NO.

FIG U RE 3. Flow test of 60 red nozzles wilh V'arsol.

Journal of Research

~H ""}f!"""I"" "H~TIt"I"" "j"""j"" "1"" "I

3 3 I I 3 2

11-3

30 p ooo

2 9

10.0

9.5

9.0

B.5

0 00

00 j

I 0 0 f."-o ~o . .. 0

o I

0 0 00 0

0 0 ° 0° .

50 psiq 00

0 ~

0 0 ° 0 0 00 0

j-o -o~ o 0 -

00 o 0

0

5 ps i~ 0 0 000 0

0 0 0 0

0·' 1 oO t o 0 t 00 0 0 o. 0 .

I I I 60 . 68 76 84 92 100 108 116

78

76

74

72

70

-

~ kO<> 0

NOZZLE NUMBER

Flow test oj 60 green nozzles with Y 01'8oi.

I 250 pSlq J I I · /

;M:j ' f I "I" rr --r--- 00 0 0 I

1-~ --- -.- -" '-oF o~ (0] 0 --f--. j

00 0 0 ___ _ __ -'

00 0 o _ __ ~ __ _ - - - -- ----t--... -- t I

34,-,-,--,-,-,-------------,--,-----.

33 -

32

30L-__ ~_L~~~L_L_i__L~~ __ L_ ____ ~

10.5,,.--,------,----,-----,----,--,---------,-----;--,.----,-"------,

9 .5 f<,--L,,'--+,,--,,*

9'?2!::0-"-'----;;:12"'"8 -'-~;------,-=-=--'-----~-'------:::-:::----'-----;:::;---'------;:;176;:----'

FIG URE 5. F low test of 60 purple noz zles with Varsol.

Flow Tests of Fixed Spray Nozzles

--~- - ------

Specificat ions for th ese nozzles state that the flow at 250 lb jin. 2 shall be in the range from 154 to 160 percent of the flow at 100 lb/in2 . Correct­ing these limits to the pressure of 98 Ib/in .2 used in the present tests, the limi ts become 156 and 162 percent of the flow at the tesL pre sur e. Spcc i­fied flow limi ts for these nozzles at 98 a nel 250 Ib/in.2 are indicated by dashecllines.

It will be noted that some of Lhe nozzlcs failed to flow within the specified limits, partic:ulilr ly tiw red group (fig. 3) at 250 lb/ in.2 It is also np­parent that nozzles that are matched in How al one pressure frequently are unmatched at other pressures.

ylany of the nozzles flowed differently at 98 Ib/in.2 before and after being S ll bj eded to high et' press ures. As will be see n in fi g ure 6, thi s c' hange continued for many press ure cycles wi til cefta i n ]Jozzles. The zero Ii nes represe nt the Hows aL 98 lb/ in.2 shown in Lhe previous three figu res. F lows at this pl'C'ss ure cha nged by more Limn 1 percent for 46 of th e nozzles as a res ul t of subj ec t­ing th em to pressure cycl ing, Th i indicated t ll at something must Ju\,Ve bc'en moved flhout witl li ll Lhe nozzles durillg the tes ts.

0 3

2

+ 1 0 - I-t-o 0 .

o

1 - 0

0

-2

30 f-Z

~ 3 a:: '" a. 2 z 0+ ~

I

'> 0

10

I-0

0

I" '" C

~ 11--1-

~ ~O

2 • I

0

0

. 0

+1

o go

1,,0 0

- I .0~ f H

2 120

I

0

8&9.

8

0

68

0

0 0

128

I

~ I I --l- _ 0 '1- o . 0 :1 0 0

0 '0--0-..-8-1 0 o ·

8 o 0 0 0

..i o 0 o 0 8 a 8 §

o 0 0 o . I 0 ootoroo ~ 0 0 0 0 0 - r .

I

0

o 8 I- ~

I I I 0

n

16 24 32 40 48 56

I o 1 ! I I !

0 ~ - . . 8 0 _0 _ t-g 10 - : . • b 0 0 :0 0 00 10 o oj 88~ ~

I '1 t·. 00 i 0 t o ."

1 0"'0° lIP 0 .

o~Tl o . ; 0.0

I 0 . 1 I 0

76 84 92 100 108 116

I 0

0 0 -.-r--g-

tls 0 . o. o SO 0 ~o .0 .JI..

to ·0 I 0 . 0 0 0

. . 0

I I 0 -.-

136 144 152 160 168 176 NOZZLE NUMBER

FIG U RIC 6. Nozzle flow reproducibilit y with Varsol at 98 psi g a s received .

451

IV. Reconditioning of 26 Nozzles

Following the above tests 26 of the nozzles were taken apart, and the following defects were found in all to varying extents :

1. There were metal chips resembling filings in the tangential slits and swirl chambers; 2. Many of the seals between inserts and bodies seemed imperfect; 3. Most of the inserts were no t properly finished on the sealing end, and the edges of the slits were ragged ; 4. Large burrs left in cutting the sli ts remained attached and caused par tial blocking of the swirl chambers.

In the enlarged photograph shown in figure 7 such burrs are visible, as are the rough surface of the insert and the ragged edges of the slits.

The burrs and chips were removed, and each nozzle was r eassembled with its original insert and strainer. No significant change in flow has occurred subsequently , so that it seems safe to state that initial changes were due to movements of burrs and chips by the test fluid . It is obvious that the presence of such foreign particles cannot be tolerated in a metering device.

V. Flow Tests of 26 Reconditioned Nozzles with Hydrocarbons

The 26 reconditioned nozzles, renumbered in the order of increasing flow capacity with Varsol at 98 Ib jin.2, were next flow-tes ted at pressures of 98 and 250 Ib jin.2 wi th five different hydrocarbons having the following properties:

Propert ies at tes t

T est temperature Eqll i\--T est fluid Name tern· aleu t N o. per·

I Kine· sp. gr. at

ature s p . g. ma tic 60°/60° F ' viscosity

-- -----Centi·

OF stokes 1 YarsoL __________________ 80 0. 779 1.102 0. 787 2 ASTM n ·Hep tane. -.-- 80 . 679 .566 .6S9 3 Apc0467 ____ ___ _ ... _ ... _ SO . S02 2. 2S2 . SlO 4 Commercial isooctanes ___ 70 . 743 . 747 .747 5 So!trol-l00 .. .. ........... 75 . 743 1. 435

I . 750

• Values in this col umn obtained b y the a iel of table 3 in NTI S Oircula r 0410.

T est fluid No. 1, designated as Varsol, is a gen­eral utility solven t, stocked as a storeroom item at this Bureau . It is similar to cleaner's naphtha, m ade by many compani es.

452

FIGURE 7. Def ective nozzle in sert.

Test fluids No.2, 4, and 5 were supplied through the co urtesy of the Phillips P etroleum Co ., and test fluid No. 3 through the courtesy of th e Anderson-Prichard Oil Corp .

Fluids No.1 , 2, and 3 were selected because each is thought to be under consideration as a standard fluid for testing jet engine auxiliaries. Fluids No . 4 and 5 were selected because they have the same density but widely different viscosities . Of the five, n-heptane and the mixture of isooctanes have properties approximating those of aviation gaso­line, whereas Varsol and Apco resemble kerosene more elosely.

The results obtained with 26 nozzles at pressures of 98 and 250 Ib jin. 2 for each of the five liquid hydrocarbons are presented in table l.

Figure 8 is presented to show the reproducibility of the flow measurements and of the nozzles su b­sequent to r econditioning. E ach point represen ts th e deviation of one observation made wi th Varsol from the average of all observations made with the same nozzle, fluid , and tes t pressure. The maximum deviation does no t exceed ± 0.5 percent , and only 5 of 100 observations deviate from the m ean by more than ± 0.25 percent. Thus the nozzles themselves, as well as the measurements , appear satisfactory for presen t purposes .

The changes effected by reconditioning the nozzles are shown in fi gure 9 . As will be seen,

Journal of Research

1.'A l3T~E 1. Flow capacity (pounds per hOllr) oj 26 no zzles ,for 5 liquid hydrocarbons

Test flu id number Sozzle n umber ______ _

___________________ __ ' _1 ___ -=-___ ~ ___ ~_J _ _=_ ___ b _1 ___ 2 __ 1 __ 3 ____ 4 _1 __ "_-__ I 1

3 ________ ___________ __ _____________________ _ 4 _______________ __ __ ________ ________ ______ _ _ 5. _________ . ___ __________________________ __ ._

6 __________ _________________________________ _ 7 _____________ • ____________________________ _ 8 ______ _____ __ ____________________________ _ _ 9 - ___ ___________________________ ________ ___ _ _ 10 ____ _____________________________________ _

] 1 ______________________________ ____________ _ 12 _______________________________________ _ _ 13 _________________________________________ _ 1'L __________ __ _____________________________ _ 15 ________________________ _

16 _______________________ _ Ii __________________________ . ___ . __________ _ 18 _________________________________________ _ 19 _________________________________________ _ 20 _________________________________________ _

2L ________________________________________ _ 22 _________________________________________ _ 23 __________________________________________ _ 2·L ________________________________________ _ 25 _______________________________________ _ 26 _________________________________________ _

38.97 40.00 40. J4 40.26 41.30

41. 52 4L57

4 t. 57 4 L 63 42.07

42.34 42.40 42.44 42.60 42.64

42.67 43.10 43.67 44.02 44.73

45.07

45.20 45.53

45. G6 46. 10

47. GO

II. 'r est pressure in first f,,"c columns is 98 Ib/in2,

b 'rest pressure in last 0\'0 col u mns is 2.10 Ib/in:l.

36.67 38.45 37.65 38.20

38.62

39. 00 39.80 38.37 38.82 39. 05

4L 39 ~9. 52 40. 03 40.33 40.42

39.92 41. 00 40.77 40.87 41. 37

41. 05 42.08 42.57 41. 8:5 42. '13

45. 07

~8. 47

39.50 39.97 39.92 40.93

41.80 40.27 41. 46

40.45 41. 10

41. 47 4 L 47

42. 17

41. 98 42.34

42.6 l 4:3. 82

42. 92 42.53 4:1.87

43.20 43.72 45. 1:3 4:3.97

H 13 45.83

~ 0.50 .----4--.-,-----,--,--,-----,,---,--,--,---,-,----, w

'" ~+ O.25 1-_4-+-9------j--o-+-+_-II-_+--t---'I~°c__f-.~--y

z l-~g~-~~_70~-x-+ ___ t--·~~O~-.4_~O~~·~-~i-~O~ ~o 8 0 "b 0 0

:; 0 0 0 -1° 0 i ll: t'J - 0.25 1--~'v;,..-.j-_+--"-+_o-<>--+-4_-t__-+-+-;;o--+-·- f-----

~ "- 0.50 L-----L_-L_l--.l_ -L_-L--1_-'-_ -'-----''----'--_--'----' o 4 8 10 12 14 16 18 20 22 24 26

NOZZLE NUMB ER

FIGURE 8. No zzle flow repTod1lcibili ty with VaI'sol at 98 psig after cleaning. Number of identi cal readings: 0 , one; x, two; . , three; .A. , jive.

12

.... ~ ~+ 4

Z 0 o ii :> -4 :'5

~ 8 "-

12

I I I I

I I I

o I I 0

0 0

0 0 0 0 -

1 0

0 0

I 1 o ~ ~ ~ ~ rn w " ~ ~

NOZZLE NUMBE R

FIG UHE 9. Comparison oj flows beJore and after cleaning nozzles, Vw·sol at 98 psig.

Flow Tests of Fixed Spray Nozzles

36.90 39. 40 38.57 38.80 40.02

40.4l 40.67 39.57 41. 13

40.55

41. 93 41.16 4 L 73 4 L 42 41. 92

41. 30 42. 12 42. 13

42.54 42.83

'136 1 43.59

'13.90 44. 45 44.48 46.42

48

46

44

~ 4 2

~ 40

38

36

34 o

. 0

. I

37.97 39.00 39.32 38.95 40.06

40.45 40.35 40.12

39.95 40. 70

41. 25 40.75 40.95 40.74 4 L 87

'11. 22 4 I. 7:l 4 I. 86 42.45

42.77

'13.87 4:3.03 4'1. 20 43.97 44.64

4,5.79

I

~

.

. 0

j

62.45 65.50 64. 12 65.45 66.85

66.85 67.45 66.05 66.85 67.30

69. 45 67.55 69.05 69.20 69.28

68.55 70. 10

69.65 70 . . \0 71.10

72.08 71. 85 72.7:3

72. 05 72. 60 7(1.88

e

. . . 2

58. ]0

61.20 60.05 61.50 61. 25

62.65 63. 75 6 1. 70 61.05 62.10

66.00 62.85 6:3.85 6,.30

63.70

63.60 65.80 65.25

6-1. 55 65. 70

66.30 (i6. 10

68.05 G(l. 8.5 65.80 71. :30

-

~ a

62. 10

63.80 (\5.30

64.25 06.60

67.2,5

66.50 67,25 66.50 67. 60

67.00 67. 95 68. 15 67.90 68.75

68.8" 70. j()

69.80

70. 30 71. 60

72. 00 70. (iO

73.65 72. 95 7:145 7(L 00

1 ~ I

'-rl . . . I

I I

0 .

.09.00 6.3.2,; 6 1. 50 G2.70 ()'3 . 50

61. 65 66.00 (,3.4 ,;

61. 9.\ 6·1. GO

67.3,)

65.15 66.15

65.85 66. 15

6,).90 6 7.5:')

(i7.35 G7.00 68.0.)

68. 90 68.5.5 70. 'I.; 70.50 li8.50 7:3. 4.5

. 0

• I •

I

I

. 0

.

I

I

60.65 G:3.35 62.00 H2.70 G·I. 20

61.lj,~

G·1. 7U

6:3.45 0·1. :l'; G5. 10

G6.70 G-1. 85

6G.20 65.85 66. G5

65. 85

67.30 67.05 G7.90 68.45

69. 80 69. 40 70.05 71. 20 70.90 7 I. ]0

i ;

.

I

I

1 10 12 14 16 18 20 22 24 26

NOZZLE NUMBER

FIGURE 10. Flow oj 26 cleaned no zzles at 98 lJsig. 0 , Apeo; . , Varso); x, n-heptane.

-

th e changes in flow ranged from + 8 to - 8 percen t, and only three of the 26 nozzles changed less th an 1 percen t.

The daLa of table 1 for Varsol, Apco, and n-hep­tan e at a pressure of 98 lb/in .2 are compared in figure 10 . The n ozzles were numbered arbi trarily in the order of increasing flow with Varsol, so that a r elatively smooth curve is obtained for this fluid. However the corresponding curves for the other

453

two fluids are not smooth , indicating that nozzles matched for one fluid are not necessarily matched for fluids having difI:crent properties, and showing the need for caution in selecting a standard test fluid.

To further emphasize this point, consider the example furnished by nozzles 6,7, S, and 9. W'ith Varsol these nozzles show an average flow of 41.57 Ib /hr and a total spread of 0.25 percent. W'ith n-heptane the order is different, the average flow 6.2 percent lower, and the spread is 3.7 percent. l'Vith Apco the flows are in a still different order, the average flow is 1.4 percent lower, and the spread is 3.7 percent. There are many other similar examples in the data of figure 10.

From the data of table 1, the ratio of the flow at 250 Ib /in. 2 to that at 9Slb/in. 2 can be calculated for each nozzle and for each test fluid. For fluids No.1, 2, and 3 this observed ratio is as follows:

Varsol , ranging from 157 to 164 percent, aver­aging 160.2 percent; n -heptane, ranging from 156 to 161 percent, averaging 15S.7 per'cent; and Apco, ranging from 160 to 167 percent, averaging 163.2 percent.

As a further ill ustration, the ratios are plotted in figure 11 in the form of a frequency curve with

... o

10

1

~

o 15·1

I

1 1

I

II I -j 156

FIGUl~E 11.

I

I ~ 'I, I I

1 I , I ,

1 ! I \ I I I \ f-I I I \

I I ' / ,\ i- I \ I I I II

1 ~ / \ 1\ II 1\ \ r--v ~ II \ \ V II \ 1 \ / 1\ 1

/ \

"" \ i

158 160 162 164 166 168 (25Qpsig FLOWI 98 ps ig FLOW) 100, PERCENT

250 psig --.- flow ratlo wi th different hydrocarbons 98 ps~g

-0-, Apeo; - e--, Varsol ; - X- , ~ ·heptane .

each point showing the number of nozzles having a given rat io. It will be apparent from these results that the characteristics of the test fluid must be known before the ratio can be specified .

As stated previously, fluids No . 4 and 5 have the same density and were chosen to show the

454

effects of viscosity on nozzle performance. Figure 12 shows how the results obtained with Soltrol varied from those obtained with the mixture of isooctanes. I t is at once apparent that the eHect of viscosity varies in both magnitude and direction from nozzle to nozzle. Hence it does not seem possible to develop any means of correcting for the viscosity of the test flu id.

4

1/

I

I

3 o

u

• •

0 . I

.

• I • 0

0 0

0 • • •

0

~ ~ ~ ~ ~ ~ " M ~ NOZZLE NUMBER

FIGURE 12. V iscosity effect on spmy nozzles . O. 08 ps iR; e, 250 psig.

VI. Flow Tests With Air

Early in this nozzle test program it appeared desirable to attempt the development of equip­ment using a ir instead of a flammable hydrocarbon as a test fluid. Buch development seemed a logical extension of the previous successful evolu­Lion of the Navy Orifice Comparator, in which air is used to flow-test metering jets of aircraft car­buretors with greater speed, accuracy, and safety than could be attained by other methods .

In applying the method to spray nozzles, the apparatus shown diagrammatically in figure 13 was investigated. Briefly, air compressed to 50 Ib/in. 2 or more is passed through a pressure regulator and a filter to a test fixtUre consisting essentially of two chambers, each about 17~ in. in diameter by 4 in. in Jength, separated by a small orifice or bleed. The second chamber serves as a mounting for th e nozzle, through which all of the regulated air escapes to the atmosphere. Pressure taps and manometers provide for observing the pressure in the second chamber and the drop in pressure between the two chambers.

In operation, the pressure in the second chamber is held constant, which means that the drop in pressure across the nozzle is also constant, and the drop in pressure across the bleed between the two chambers is observed . The latter is determined by

Journal of Research

r AIR AT 50 ps i Q

FIGURE 13. Schematic diagram of nozzle air test.

Lhe volume rate of flow between chambers, which , in turn, is a function of the flow characteristics of the test nozzle.

Obviously the method is comparative rather than absolute, so t hat a set of nozzles calibrated by some other mcthod is required for thc calibration of the equipment using air . For nozzles having a flow capacity in Lhe range 40 to 50 Ib/hr at 100 Ib / in .2, it ha s bee n found by experiment LhaL the air-test method g ives best r esults when the bleed between the two chambers co nsists of a single hole made by a No. 76 dr ill (d iameter = 0.019 in. ), and when the pressure in Lhe second chamber is from 1.2 to 2 times the pressure drop across the bleed.

In developing the air- test method , t he 26 nozzles mentioned prev iously were tested wi Lh yarious constant press mes in t he seco nd chamber t lu'oughou t the range from lOin. of watrr to 50 in. of mercury. The brst correla tion betwrr n results with a il" a nd wiLit liquid hydrocarbolls is obtained in thc range 25 to 35 in. of mereury. An example of this correlation is givell in flgu1"e 14, in which the preSS Ul"e in thr srro nc/ chambrr (PI ) was 30 in. of mercury, and the observed drop in pressure across the bleed (P z- p] ) is plotted against the observed flow of Varsol at a press me of 98 Ib/in2• The best smooth cmve tlu'ough the observed points seems to be a straight line, from which the maximum devia t ion is less than 2 per­cent and the average deviation is less than 1 percent. The sensitivity of this air apparatus was about 1 in. of mercury per pound of fuel dis­charged through the nozzle per hour.

The dashed lin es in figw'e 14 are the limi ts of flow specified for nozzles coded r ed , green, and pmple. On the basis of the results shown in this flgmc it might be concluded that the ai.r-test method was satisfactory for flow-testing spray nozzles. If this were true, it is certainly to be preferred from the standpoint of speed and safety.

As already stated, figure 14 compares the results obtained with air and those obtained with

Flow Tests of Fixed Spray Nozzles

Varsol. Siruilarly figure 15 sho~vs a comparison of resul ts with a ir and Apco, and figUl"e 16 shows a comparison of those with air and n-heptane , all data for liquid fuels being [or a preSS Ul"e of 98 lb/in2 • D eviations from the straigh t lines a re all within ±3 percent except for one noz7.lc in figure 16. Co nsidering that Lbe nozzles clo no t perform consistently with the three liquid test fluid s, as was shown in figme 10, the corrcJatio n of the r es ul ts with air and those with li quids is bet ter than m ight have been expected.

48r---,----,---,----r---,----r---,----~--,

~46~--~----L--~----~--~---+~~~11--~

" ~ ~

t---+-----j----j -t-~ ------- ------ -- t-

~ ·30 In. Hq

16 17 IS 19 20 21 22 23 24 DROP ACROSS BLEED, In HQ

FIG U RE 14. Correlation of nozzle tests with ai,. (lnd Yarsol.

48

I I

....--V 6 -- - r-./

°V 4 -- - ....-- t--" 0

~ /0 o 0' 2

~~ J -- -

o 0

/" 0 "-

/ Vo

Pt , 30 In. Hg

38 IS 16 17 18 19 20 21 22 23 24

DROP ACROSS BLEED, In. HQ

F i eU R E 15. Correlation of nozzle tests with (ti,. and Apeo.

6

4 /

all 2

~ u

0

~ 0

0 0...9-

i5 0 4

~ 0

I ~o 0

3 6 15 16

FIGURE 16.

P," 30 In. Hg

17 18 19 20 21 22 23 24

DROP ACROSS BLEED, in. H~

Correlation of nozzle tests with air and normal heptane.

455

In its present state, the air-test method would seem to have considerable merit for preliminary tests of nozzles, particularly in production and after overhaul. It would certainly be valuable in rejecting nozzles that flow so far from the design value that tests with liquid would be a waste of time, and in grouping nozzles within reasonably narrow flow limits. It might also be useful in the matching of nozzle bodies and inserts to get desired performance.

As a matter of fact, the uncertainties in the performance of nozzles determined with air are probably little, if any, greater than the present uncertainties in the actual performance of nozzles in engines. Consequently the further develop­ment of the air-test method will be carried along as rapidly as the resolution of the over-all problem of nozzle performance secms to warrant.

VII. Discussion and Conclusions

The nozzles used in these tests were received late in 1947, and are believed typical of nozzles of this type in production at that time. The presence of chips, burrs, and inadequate finish on certain parts were pointed out months ago, and it seems probable that such easily remedied faults may no longer be a matter of concern.

A Q\:ed nozzle performs two important func­tions, namely the production of a spray and the metering of fuel. This report deals only with fuel metering. In nozzles of the type tested, this is accomplished primarily at two locations within the nozzle. These are at the four tangential slits that operate in parallel, and at the discharge orifi"Ce in the nozzle body, which is in series with the four slits. It is not surprising, therefore, that the simple law of discharge through an orifice does not hold exactly for the nozzle as a whole. As examples, the relation between pressure and flow for one nozzle may differ somewhat from that of another, and the effect of a given change in viscosity of the test fluid may be in one direction for one nozzle and in the opposite direction for another.

Examined from this point of view, it is surprising that the results obtained with SL,( different test fluids including air agree as well as they do.

Perhaps the most obvious conclusion from these studies involves the choice of a fluid suitable for testing fuel nozzles and other engme auxiliaries.

456

It is believed that a fluid chosen as a standard for this purpose should have as many of the following characteristics as may be found attainable:

(a) It should be ch eap and free from highly strategic ingredients; (b) It should be available in quantity from many sources, and its physical properties should be readily reproducible from a production standpoint; (c) It should be safe, noncorrosive, and should have the same effects as engine fuels on packings, diaphragms, etc.; (d) Its physical properties should be the same as those of engine fuels and should not change with use by evaporation of light ends; (e) When exceptions are made to the above r equirements, it must be possible to interpret data obtained with the test fluid in terms of performance in engines.

If the test fluid differs in density and/or vis­cosity from the fuel used in the engine, the present resul ts show that:

(a) The actual rate of fuel delivery to the engine at any pressure may differ by several percent from the rate predicted from the test data; (b) Nozzles matched for the test fluid at a particular pressure may not be matched at this or any other pressure in the engine; and (c) it will no t be practicable to develop corrections which are generally applicable for differences between physical properties of the test fluid and the engine fuel.

Both density and viscosity of the fuel being metered by a nozzle are important in determining its rate of discharge at a given pressure. Both those properties change considerably with tem­perature. Thus it seems likely that turbo-jet operation might be improved by giving more attention to the temperature of the fuel entering the individual nozzles, and to possible methods for its control.

Much additional thought and experimentation are warranted in the development of nozzle test dat,a that will be tI'llly indicative of the perform­ance of nozzles in operating engines. Even though some fluid is selected as a standard for test purposes, there remains the development of satisfactory test method s and the evolution of significant test specifications. An obvious first step toward the latter is the determination of tolerances in flow that can be alIo,,·ed in engines, and more particularly of the unavoidable differ­ences encountered in operating engines .

Thus the over-all pl'oblem of determining, by means of bench tests, whether a given set of spray

Journal of Research

nozzles will gl\Te opti mum performance in an engine is highly complex. Equally difficult is the testing of a set of replacement nozzles so there win be no question that they will perform as well as the original set.

So long as spray nozzles arc used in engines, there is no doubt that solution of the afore­mentioned problems will pay dividends in im­proved engine performance, particularly at high altitudes. Employment of spray nozzles requires that the general level of fuel pressure be high, whieh in turn involves mechanical difficulties with

Flow Tests of Fixed Spray Nozzles

fuel pumps, lines, manifolds, and seals, and in­ercases the fire hazard in case a fuel line is broken. IL, therefore, seems legitimate to raise a question as to the relative meri ts of expending r csearch effort on the further development of tb e various components of h igh-pl'essUl'e fuel systems, or on the development of combustion chambers that will function with low-press ure fuel and without spray nozzles.

'VASHI1\GTON, January 10, 1949.

451