Document
copy n S
53K05 A miiq m.
-—
RESEARCH MEMORANDUM
STARTING CHARACTERISTICS AND COMBUSTION PERFORMANCE 03? MAGNESIUM SLURRY m 6.5-INcH-D~TER RAM-JET ENGINE MOUNTED IN CONNECTED-PIPE FACILITY By James B. Gibbs Lewis Flight Propulsion Laboratory Cleveland, Ohio , BY .. .
G- =iJ)q53-
...............
............ . ~N . . . . . . . . . . . . . ... . . . ..”— GRADE W brt 1~.b MARII:G CHANG~?”...’ . . . . . ..lh..llk.k.h..md...
ml. .
NATIONAL ADVISORY COMMITTEE
AERONAUTICS
FOR
WASHINGTON January 28, 1954
I
. ..-. ..-— -. ....-. . ... . .. -e NACA M E53K?J5 NATIONAL AIYVISQRYCOWITTEE FOR AEEmNAumcs ST!ART~G CHARACTERISTICS AND COMBUSTION PERFORMANCE OF MAGNESIUM SLURRY cONNEcmD-Pm! FACILITY IN 6.5-nK2H-DIAMETER RAM-JET ENGINE MOUNTED J3 By James B . Gibbs The starting characteristics and cutibust ion performance of 50 per- cent _sium powder in a hydrocarbon carrier were investi~ted in a — flight-type, 6 .5-inch-diameterram-jet engfne mounted in a connected- pipe facility. Sttiim.g disks, metal plates mounted in the couibustion chembm and blocking part of the burner area, were developed that provide quick, &pendsble starting of the engine over the starting equivalence ratio range investigated, 0.36 to 0.69. After the engine was started, the disk was expelled to permit normal operation of the eng~. One of the disks ns expelled within 0.1 second after fuel-flow initiation. No -- explosive starts were ~erienced.
A fLame:holder protection plate, -ales i~ed to cover the cavity in the flame-holder mounting inibe,permitted operatIon without flame-holder damage for test durations of about 20 sec- mds. The ccxdmstion tests were made with a short, 8.5-inch fuel-air mix- ing length) because conibustioninstability was encountered in previous free-Jet tests wtth a longer mixing length.
The ccmibust ion efffciency was above 77 percent for the equivalence ratlo range fr~ 0.5 to 1.0 and reached a maximum of 81 percent at an equivalence ratio of 0.7. ~- creases In inlet-air temperature from about 60° to 370° F caused Increases in combustor efflclency of 12 to 22 percent over the equivalence ratio range investigated.
Performance data obtained with the slurry system having the short fuel- air mixing length were compared with performance data obtained in the free- .
~et and flight tests of similar ethylene-tiled rem-jet engines at the WA The slurry fuel provfded over twice the fuel volume LEmf@y laboratory.
speclflc @tisej however, the fuel weight s~cific Impulse was slightly higher for the ethylene fuel.
An alr specific impulse of 187 seconds was obtained with the slurry fuel, while the maximum obta~d with ethylene was 159 seconds.
Ih the Langley fllght-test vehicle, a greater fuel load and a greater thrust would be possible if the ethylene fuel were replaced by a magnesium slurry.
Therefore,.h@her flight speeds, higher altitudes, .
and longer flight durations should be attainable with the slurry fuel.
.
.
2 MICA RM E53K05 .@” ,:.
+.A .-. .- PH mOJXJCTIOli the IOICALewis laboratory, As part of a high-energy-fuels program’at the ccmibustion performance of a magnesium sIiiiry was investigated in a 6.5-tich-diameter ram-jet engine mounted h ~ a connected-pipe facility (ref. 1). The fuel consisted of 50 percent:atomized magnesisunpowder by weight suspended in a hydrocarbon carrier fuel. The slurry performance was compared with the performance of ethylene in free-jet (ref. 2) and flight (refs. 3 and 4) tests of simil= ram-~et engines. As a result, appreciable performance gains were predicted for the slurry system over the ethylene-fueled, eupersonlc-fli.ght vehicle de~cribed in refererice3.
—. .-. .
— The slurry-engine comblnation is being preflight tested h the free- Jet faclllty at the NACA Wallops Island, Vtiglnla station. In the initial tests, flame holder end conkustor failures were encountered because of Intermittent couibuet ion upstream of the flame holder. Since the two types of testIng d~f ered mainly h the method of air diffueion, the un- stable ccuibuetion In the free-Jet was believed to have been caused by a — more irregular velocity profile near the station of fuel injection than was present in the connected-pipe tests. ti subsequent free-set tests~ ““” the.fuel inJection station was moved downstream ~ Inches to a region .-
r—
with a more rex velocity profile. The gxtion upstream of the flame holder was eltiktedj however, the shortened fuel-air mtiimg lei@th_- resulted in perfommmce lower than that predicted from the connected- . :“ plpe teats for equivalence ratios greater ~n about 0,.6 ~’, ~ou@: — --- out these tests, the engine starting characteristics were erratic with .
the flare igniter used in the connected-pipe tests. In addition, flame- .- holder life with stable combustion was considered ~ginal.
The present Investigation was directed. toward improving the starting characterlst Ics of the engtie, tireas Ing flame-holder durability, and Increasing the couibustlon performance at him equivalence ratios while using a fuel-air mixing length ~ inches sheer than that reported In reference 1. A series of short couibustion tests was conducted with a 50 percent ~gnesiun slurry in a rem-Jet engine siml.ler to that used in ref- erence 1 in a connected-pipe facility.
A disk, blocklng part of the burner exea, was mounted in the couibusticm chanber as an engine-starting aid. The disk was desigued to provide hi@. combustor pressures and low velocities downstream of the flame holder prior to IgnltIon end then to fail from mechanical stresses at elevated temperatures shortly after ig- nition. Seven design variations of the starting disk were studied, and the effects of three fuel-distribution control-sleeve configurations on conibust ion performance are compared.
Also, the effects of Inlet-air tem- .
perature on performance and of a flame-holder protection plate on com- bustion performance and flame-holder durability are described.
l - , I IWICARM .E53K05 t FUEL AND APPARA!IUS - weight of ma~esium powder Fuel . - The fuel contained equal parts by and a~~ocarbon fuel. An analysts of the hydrocarbon carrier, ~-F- 5624A grade JP-4 fuel, is given h table I. Currently, only a limited emount of small-particle-size magnesium is available; this material is stored in drums and the purity of the -r varies from one drum to another. b order to conserve the purer powder for preflight and flight - - - testing, the powder used in this investigation was obtained from those drums contatifng the greatest amount of _ities. !l!hIs material was ..- made into two 200-pound hatches of slurry and representative samples of each batch were analyzed. In each case the purity of the powder was measured at 93 +1 percent. Thus, the ratio of actual uncouibinedmag- nesium to magnesium plus hydrocarbon was 0.48, the stoichiometric fuel- air ratio was 0.1109, and the lawer heat of couibustion was 14,900 Btu per pound of maguesium plus hydrocarbon. The slurry density including _ities ns measured at 1.05 grems per ctiic centimeter, the ssme value as computed for a 48-percent-pure magnesium slurry. The mean size of the nearly spherical powder particles was 1. S microns as determined with - a Fisher S* -Seive Sizer. Because of the small powder stie, the slurries were stable (apparently homogeneous) for over 24 hours, ad for this ti- vestigation no stabilizing additives were required.
.
Fuel system and ram-jet installation. - A diagram of the fuel system .
is shown in figure 1.
Fuel was supplled to the engine frcm a 2-cLibicfoot fuel tank pressurized with nitrogen. The rate of fuel flow was governed by the flow-restricting orifice located ~stream of the fuel injectors and by the controlled fuel-tank pressure. The maximum fuel-tank pressure was 260 pounds per square inch gage.
A diagram of the ram-set installation Is also shown in figure 1.
..
The combustion air, from the laboratory air s~ly, was passed thruugh a tube-type heat exchanger, metered, and then throttled by a remotely con- trolled butterfly valve. l%e ctiustor shell was cooled by diverting a fixed portion (approx. 36 percent) of the conitmstion air through a l/2- 3nch annulus between the shell and a cool- Jacket. The high external pressure of the cooling ah on the combustor necessitated the use of four -- .
longitudinal reinforcement bars and a cmibustor shell thickness of 0.033 inch.
T9E cooling air recouibined with the min portion of the combustion air in the inlet plenum and then entered the enghe. A 3-foot-long shroud was mounted on the ram-~et diffuser lip in sn effort to obtain a flat velocity profile at the entrance to the diffuser. The couibust ion products - were discharged into the atmosphere Just outside the test cell.
Rsm-~et ezwtie. - A detailed description of the basic engtie is pre- sented in reference 3. A diagram of the engine, as mod~led for the pre- sent investigation, is shown in figure 2. The diffuser-lip diameter was 4.42 hChSS. The inside diameter and the length of the Inconel couibustor NACA RME53K05 v. .: -.. .
%!zE-
was6 inchesin were 6.5 and 18 inches, respectively. The exhaust nozzle diameter.
.
!lheinner body of the basic en@ne of reference 3 was umdflied _- dlately downstream of the support struts to accommodate the slurry Tnjec- ..— tor and the flame holder. A spring-loaded variable-port-area injector, ..— having four longitudinal slots spaced 900 apart, is shown in fIgure 2.
.-— The piston area and the spring were designed to provide a pressure drop of about 50 pounds per sqgare inch across the fuel slots for the fuel- flow-range tivestigated. An O-ring seal and silicone grease were used to prevent seizing of the piston. Atomization of the fuel was achieved by inptnging the fuel jets on a cyl~ical fuel-distribution control sleeve mowted ti the air stream. The fuel slots were located ~ Inches from the face of the flame holder (fig. 2).
~ fh holder -S COUQOSed of V-gutters and funnels (surfaces of revolution of a cone), and it blocked 46 percent of the cmibustor cross-sectional area. me matertil used for the flame holder was 3/32-inch-thick Inconel. The aforementioned fuel I.nJector and flame holder gave the best performance of the several types tested during the develmnt progrsm described in reference 1; how- . ..- ever, the distance fran the fuel ti~ectlon slots to the flame-holder face .
(fuel-air mixing len@h) was 12 inches fn the investigation reported in k reference 1.
Ignition wae provided by a ma~sium f~e 4 inches long and 2 inches in diameter with a nomhal burrdng time of 20 seconds. The flare was cemented into a l/16-inch-wall Inconel ttie which fitted into the flame- holder cavity (flg. 2). !l%edownstream (i~it ion) face of the flare was coincident with the downstream edge of the flame holder.
.
wart ~ disks. - A total of seven design variations of starting disks of two general types were studied.
The two types dtff ered only in the locatfon of the disk in the cmibuetton chauiber and the u&hod of mountfng. Dfagreme of both types.and a t~le listing the dimensions are .—-.
shown in fQure .3.
The flame-holder-mounted disks, nsmhrs 1, 2, 3, and 4, were located near the center of the couibust ion chauiber and were held In place by means of straps hooked over the funnel part of the flame holder.
The nozzle-mounted disks were slipped edgewise in the nozzle at the disk flats, turned, and butted agabt the upstream edge of the nozzle.
.— A nozzle mounted disk is shown in position in figure 2.
DMtmmentat ion. - The conibustton air flow was measured by a square- edge ortilce conforming to A.S .M.E. standards. me orifice differential pressure and the couibustorstatic pressure at station 8 (flg. 2) were sensed by Statham pressure plclqm and recorded continuously on a four- .
channel oscill.ograph. me frequency response of the electrical portion of the system was flat up to 25 cycles per second.
However, for conven- ience, the ccmibustorstath-pressure tap and sensing element were Dhv- ___ sically separated so that an appreciab~ time lag -“s noted in &-” .
.. .. . m----
l L kF+?!3B . . . . . . . . . .
NACA RM E53K05 a measurement. The combustion-air-orificeupstream pressure was recorded at 2-second titervals by mans of a pressure pickup and a self-balancing .
potentiometer.
The fuel-flow rate was mtered by means of a rotating-vane flowmeter and an ~icatlng potentiometer. This measurement was checked with Bourdon gages which hdicated the pressure drop between the fuel tank and a point downstream of the restrtctlng orlflce (fig. 1) .
The orifice and ccmibuatw-inlet alr temperatures were measmd with a record~ potentimneter at 12-second intervals.
PROCEDURE The two fuel flowmeters and the pressure-recordd.ng eqyipment were sub~ected to a comprehensive calibration approx~tely every 6 runs. An additional check on the air-ortitce and conibustor pressures was made prior to each run by obtafn~ steady-state data at several air flows and com- paring the record data to ~~ter and gage readings. The absolute accuracy of the measured values 1A believed to be within +3 percent.
.
The slurry was transferred to the fuel tank innediately after mlxlmg by pressurizing the mixing barrel with nitrogen.
Tn order to ensure that . the fuel ti the tank was nearly homogeneous, the runs were usually mde .— .
within 2 hours after filling the tank. ~ a greater period of time elapsed after the tank was filled, the fuel tank was removed from the apparatus and agitated on a barrel roller prior to the run.
A snmll qmntlty of heated combustion ah was passed through the apparatus before each high-inlet-air-temperaturerun to permit the lnlet- ah ducting to approach equilibrium teqerature.
Pressure was applied to the fuel tank and the recorder-chart drives were turned on.
In the flight vehicle, the flare is ignited before the vehicle Is launched; therefore, the flare was ignited at low air-flow rates in this investigate Ion. The ah flow was then rapidly Increased to the starting condltian, whereupcm the fuel valve was opened.
From this point, two procedures were folloued: ~ the Initfal tests, emphasis was placed on the engine starting char- -..
acteristics, and the fuel valve was turned off wlthti 5 seconds after the start of the test.
The fuel and alr flows remdned nearly constant throughout these tests.
Ih the later tests, the ah flow was changed in steps of about 5-seconds~ duration in order to COV= a range of fuel-ah ratios. The fuel flow changed slightly during the later tests because of a changing codbustor pressure.
After the run, the fuel system downstream of the fuel tank was drained and flushed with gasollne. The flame holder and the fuel injector were - removed, cleaned, and Inspected for material fallure and oxide deposits.
,.
lWICARM E53K05 IIA!lA REDUCTION’?
Tn.order to allow as umch time as possible for the coqknzstor ti..,_tie ,* - irmtrumentat iom to reach equlllbrtum bef~ data were recorded, data points were chosen from the records at each air flow near the start of ~ the next step air-flow change.
: — computing the air specific im- The fuel performance was evaluated by pulse, fuel weight specific impulse, and *1 volume specifIc Impulse.
These parameters ewress the total stream momentum per unit weight of i!!
air, per unit weight of fuel, and per unit,volume of fuel, respectively, referenced at the exhaust-nozzle throat for a Wch nuuiberof 1.
They are defhed by the following expressions: —
- Total stream mcnnen
Air specific impulse Air flow
‘m w
Fuel weight specific impulse = ‘ir~p~~~cr~~~ (see) u .Fuel volti speclftc -se (-~ \uu ../ = Fuel weight spec&ic Impulse X Fuel density .
Figure 7 of reference 1 shows a straight-lige relation between the total stream momentum, determined by means of a thrust barrel, and the combustor ‘ -. .
static pressure at station 8j the data scatter was less than W percent of the mean when “theexit nozzle remained nearly free of oxide deposits.
Since In the present investigation, the range of operating conditions and the fuel were similar to, and the combustor and the cmbustor statfc- pressure-tap location were identical to thdse used in reference 1, the .. -_ stream momentum was ccmveniently determined by means of the measured ..- couibustor -exit static ~essure.
-.
The combustw efficiency for each datum point was determined by Fuel-air ratio, ideal ~ ~W Combustor efficiency (percent) .
Fuel-air ratio, actual — at constant air speciflc impulse. An effic.i.ency defined in this manner — includes exit-nozzle efficiencies and the heat losses to the burner walls as well as the cm.dmstion. efficiencies based on the enthalpy rise across the combustor. The ideal “=lues of equivalence ratio were obtained from l _ references 6 and 7 for the proper inlet temperature aud a couibustor- pres- sure of 2 atmospheres. Because of the Impurities in the magnesium powder . .
used to prepare the nominal 50 percent.S1-y, each pound of the fuel _ - .
, ., WA RM E53K05 7 actually consisted of 0.5 pound of ~-4, 0.465 pound of magnesium, end 0.035 pound of lnqnrrities. This mixture is equivalent to 0.965 pound of a 48 percent slurry and 0.035 pound Of impurities.
Therefore, to compute realistic conimetor efficiencies> it was necessary to multiply the fuel flow rate by a factor c& 0.965 to obtain the flow rate of the “pure fuel.” It was also necessary to use ideal equivalence ratios based upm alr- specific-impulse data for the actual unccnibined magnesium concentrate ion, 48 percent. These two correcticms affect the cmputed combustor effi- . .
ciencies in opposite directkms j and the net effe& Is that the values reported herein are about 2 percent higher than the values would have been .. —- & the fuel impurities had been tgnored.
RESUILl!S AND DISCUSS- Starting Characteristics The engine start= cliff iculties encountered in the free-Jet tests, which s-te flight-test starting condlttons, were probably caused by the use of higher air-flow rates and, therefore, cdbuetor velocitIes higher than those used in the connected-pipe tests reported h reference # 1. Ih order to achfeve the higher air fluws h the present investigation with the available air supply, low inlet-air temperatures were used for the Initial series of tests.
Although the low Inlet temperatures do not .
simulate the flight-test starting condition, the resulthg starting con- ditlcm created a combuetim environment which was 8s severe as could be achieved with the existing test apparatus and was adequate to determine the beneficial @Yect of a start- aid.
The slmlfkant results of all the starting tests m presented in table II.
The engtie started wtthout a s~tlng disk In the first test, where the madman air flow was 14.2 pounds per second end the hlet -air temperature, 300° F. In the second test, the temperature was reduced t~ 1350 F and the open-throttle air flow was 18.4 pounds per second. The engine did not start at this ccmdition without a starting aid. It was assumed that the engtie wuul.dnot start without ald at ccmibuetor-imlet conditions near, or me severe than, those of run 2.
disks. - Stitfng tisks 1, 2, 3, end l?lsme-holde~mmnted starthg 4 (fig. 3) were fastened to the flame holder by straps, and the disks were axially located near the central part of the ccmbustbn chdber. The -..
central locatiom was.chosen in order to reduce the quentity of fuel In the co?ibust ion chauiber at the time of ignitIon and thereby to reduce the possibility Of SX@OSiVe St8rtS. A typical start with a flame-holder- .
mounted disk Is shown in figure 4(a) which is a photograph of the oriflce- ati pressure dlfferentIal Ap and the ccuibuetor-exit statit-pressure traces. A possible criterion of engine starting ia the tIme required from .
fuel fluu initlathm (time zero) to the point where the engine Is operating 8 NACA RM E53M15 ( at full thrust. Full thrust was assumed to be the petit where the ccmibmtor-exit ~essure rose to the msximum steady state at the starting .
Condi.t ions. Since the exact point of full thrust or fuU conibustor-exit pressure was difficult to determine because of the slope of the pressure trace, a more reproducible engine starting criterion was the the requtred — to achieve 90 percent of full thrust. Both the t- to f~1 ~t ~d .
the time to 90 percent of full thrust are listed in table II for the tests without a starting disk and for those with flame-holder-mounted disks.
d As Indicated h table II starting disks 1 and 2 (fig. 3) failed either at the straps or at the fastening between the straps and disks prior to # Ignltim during 3 of the 12 start tests. -Thernmiberof fasteners was in- creased for starting disk 3 and the starting aid provided satisfactory ignition In four tests covering a range of equivalence ratios from 0.* .- to 0.69. The 69 percent blocked area of starting disk 3 was reduced to 57 percent for starting disk 4, and essentially the same time to 90 percent end 100 percent full thrust was obtained for starting disk 4 as for 3.
The starting results indicated that disks .3and 4 would be satlsfactOryj — however, the nature of failure of one of the flame-holder-mounted disks tested, disk 2, indicated a potential haz@rd in the use of this type of stsrting aid. ~ this test the straps on”only one side of disk 2 failed; this allowed the disk to flop over to me side of the combustor out of f.
the active burning zone. The resultant combustor performance was low_ — durtig the entire 4-second-duration run p~mmablybe cause of the skewed . .
flow resulttig from the lingering disk. Since there was no guarantee egalnst the recurrence of this type of failure, exit-nozzle-munted disks ., .
were next considered.
Exit-nozzle-mounted sttiing disks. - The exit-nozzle-mounted disks, mmibers 5, 6, and 7, were held in place at the qpstream edge of the exit nozzle by the alr forces.
In the runs with nozzle-mounted disks, the combustor-exit pressure tap was.upstream of the disk rather than down- ,..
streamj therefore> the couduetor pressme .~stream of the disk rose to a _ peak after ignition then dropped to the steady-state full-thrust value.
The time from fuel-fluw initiation to disk e-ion, which occurred at the peak conkm.stor pressure, was used as an engine starting criterion and is listed along with the peak pressure for all tests-in table II. A.photo- -- graph of the co-tor pressure and the orifice pressure-drop traces : during a typical start with a nozzle mounted disk is shown in figure 4(b).
Disk 5 (fig. 3) was not strong enou@. to resist the force of the air prior to i~iti.ono Disk 6 was made thicker and of harder almti~~ No . . ..
.- ~tart fail&es were encountered with disk 6 in SIX tests covering a r=ge of equivalence ratios from 0.36 to 0.57 S@ a range of air flows-from - m ~imum.time required.to e~l Msk ““.1 13.1 to 18.7 pounds per second.
..-.
6 in the six tests was 0.8 second. This time was not excessive but, in free-Jet or flight tests, the required peak ccmibustor pressure WOW not be reached because of d~user buzz) am the t- to ewel a 8iven ““ -
kiai-t%hiiim
NACA RM E53K05 disk might be increased. Therefore, disk 7 was’designed with a thickness of 3/32 Inch Instead of the 1/8 Inch thickness of disk 6. This disk was .
able to withstand air flows at least up to 18 pounds per second during a cold test and was expelled in O.1 second durbg a start test (run 31, table II). Ho eI@osive starts were experienced throughout the investiga- tion.
Exit-nozzle-mounted disks are recommended for the flight vehicle since thetr complete expulslon is more certain than is that of the flame-
E
holder-mounted dtsks. Disk 7 was expe~d very rapidly after Ignltion P and yet had adequate strength to withstand the ah flow forces antici- pated at the proposed flight startIng conditionsj this disk is therefore recommended for the current flight vehicle.
Couibust ion Performmce This part of the tivestlgatlon was directed toward improving the coudustIon performance of the slurry-fueled cunbustor, developed in the investigation reported in reference 1, by using a modified fuel-alr mti- ing length of 8.5 tithes instead of the l%inch length used In reference .
1. Because of the performance charactertstlcs of the slurry fuel and the particular flight application proposed, emphasis was placed on improving the rich performance, for example, performance at equivalence ratios above .
0.5. The requirements for the flight application w~e a conbustor that would exhibit hi@er thrusts (air specific Impulse) and more desirable fuel consuurpt ion than the ethylene couibustor at the foll.x conditions: combustor-imlet pressure, 30 to 60 pounds per square Inch absolutej inlet- air temperature, approximately 350° F j exit-nozzle throat dtameter, 6.0 inches.
!T!be results of all tests In which conibmtion was achieved axe presented In table III, and the significant results are discussed in the following paragraphs.
oxide deposits l - The oxide deposits In these tests are signlftcant with respect to both pro~ected flight performance e and titerpretat ion of the performmce data reported herein. Because the performance data are expressed In terms of ati speclftc impulse and are dependent upon the exit- nozzle area and nozzle pressure data, the combustor was carefully checked after each run. No deposits were found In the nozzle throat or blocklng the conibustor-exit static pressure tap and only thin scaly oxide deposits were found on the conibustor walJs.
Effect of Met-air temperature. - The effect of inlet-air temperature .
on the couibust ion effIclency of the couibustorconfigurate ion with a fuel- .
distrtbutIon control sleeve 2 inches long and 4 Inches in diameter and a fleme-holder protection plate is shown in figure 5. Approximately M to 20 percentage points in couibustirm efficiency were gained by ticreasing to 367° F over the range of the cmdxzstor inlet-air temperature from 63°- equivalence ratios investigated.
10 NACA RM E53K05 Effect of flmhe-holder protection plate on flame-holder durability performance. - The fl--holder design incorporates a flare mounted .- .
In the rear of the centerbody. Observations made in the investigation reported M reference 1 emd in * present @vesti@tion ~~cated t~t.
the failures which originated In the rear ~@ of the fwe-holder case .
and the flame-holder mounting tubing progressed forwexd with a resultant failure to the flsme-holder section. It was considered advisable to re- strict the open area at the end of the flar&holder ttie in order to mln- iud.ze the reclrculation of cmibustion products ti the crltIcal area.
This was attempted by the use of a flame-holder protection plate which conststed of an Inconel cap containing a 3/4-Inch-diameter hole for the discharge of the flare flame. When the flame-holder protection plate was used In the short tests, 20 seconds or less~ damage to the flame- — holder section was negligible although the p~tectlon pl+te was burned away h all tests that exceeded 10 seconds.
.- ~ the tests of about 25 seconds, damage to the flame holder was be- ghwl.ng as evidenced by declining perfo-qg level ~ the ~st fe!!. see-.
ends of one of the tests. Examination of the cmibustor after the test also Indicated the stsz% of flame-holder damage. me use of a flame- -— holder protection plate permitted “failure ~ee” ol=ation ~ the short.
_— ... - tests ~ to 20 seconds in duratlcn.
In the four tests made without a flame-holder protection plate at the IOW met temperatures, the performance was lower than in the tests Z@S _ However, In the two tests with a pro- wtth a protection plate (fig. 5) .
tectIon plate at the higher islet temperature, the protection plate was .- burned away dur~ the tests without noticeably affecting the perfomemce.
.
This tidicates that performance is unaffected by the use of the prutectlon plate at the high inlet temperatures which simulate fltght conditime. _ - Effect of fuel-distributim control-sleeve length and diameter. - A loss in performiince at high equivalence ratios was experienced in the free- Jet tests when, in order to @prove couibustion stability, the fuel-air mtxing length was reduced from 12.0 to 8.5 Inches. Part of the loss was probably caused by the decreased length available for fuel spreading which in turn caused a fuel rich region near the flame-holder center. LChere- fore, the fuel-distribution control-sleeve dimeneicms were varied h an effort to regain the perfomnance lost. CcabuetIon efficIency Is plotted .
as a function of equivalence ratio h figure’6 for three fuel-distribution For the 8.5-inch fuel-alr mixing length, control-sleeve configurate ions.
considerable Increases h performance were achieved by reducing the control-sleeve length from ~ to 2 fiches~ t@e largest gains occurred in — .
the low-equivalence-ratio region. Further i&ov&ents at the higher equivalence ratios resulted when the fuel-distribut icm control-sleeve diameter was increased from 4 to 4$ inches. .
Both of these changes cause ‘1 - NACA RM E531K)5 .
the fuel to be distributed further from the flame-holder center. The com- bustor efficlency was above 77 percent for the equivalence ratio range from O.5 to 1.0 and reached a maximum of 81 percent at an equivalence ratio of 0.7. Thfs performance Is considered acceptable for hit ial -.
fllght tests.
An absolute comparison of the effect of fuel-alr mixing length Is impossible because the conibustorewluated In reference 1, although sim- ilar to the conibustor tested in this investigation, had a different fuel- air mixing length, flame-holder protection plate, and fuel. As previous- ly discussed, the effect of a flame-holder protection plate on cozibustion performance at an inlet-air temperature of approximately 340° F was neg- ligible. !l!he fuel used in this investigation, although of lower purity, was of smaller particle size than the fuel used in reference 1 and} be- cause of the effect of pmticle size on ths couibust ian performance re- ported in reference 8, would prob@.y be more reactive than the fuel used Comparison of the cmibustor efficiency curves presented In reference 1.
inches long and 4 In figure 6 for fuel-distribution control sleeves 4+ Inches in diameter shows that a c~neiderdble loss in perfmmance results when the fuel-air mixing length is reduced from 12 to 8.5 inches. The maJor part of this performance loss was recovered by reducing the control- .
sleeve length to 2 inches and increasing the diameter to ~ inches.
C_ison of Slurry and Ethylene Performance The performance of slurry fuel in the best configuration with the 8.5-inch fuel-ati mixing length was compared with the performance of ethy- lene fuel k free-jet tests aF a slml~ rem-~et (ref. 2) . The slurry performance was obtained from figure 6 for the 2-inch-long, 4#+nch- diameter, fuel-distribution control sleeve. The dtier was =rsed in a supersonic alr stream h the free-jet tests, and this may have resulted in conitxmtor- inlet veloctty profiles different from those encountered in the present connected-pipe tests.
V&riations in velocity profile, by affecti.ngthe mixing of the fuel and air, could influence the combustion performance. The diffuser-entrance and nozzle-throat diameters for the ethylene tests were 3.96 and 5.75 inches, respectively, as compared to 4.42 and 6.0 for the slurry tests. Ccneequently, the codbuet= velocities associated with the ethylene data were lower than those reported for the slurry tests.
ison of fuel weight and air specific impulse data. - Fuel Coulpar weight specific impulse is plotted against air specific Impulse for slurry The Me- performance data for the and for ethylene fuels (fig. 7(a)).
slurry and octene-1 fuels were obtained frcm reference 6 and for ethylene frcm reference 9. Over the range of air specific Impulse obtained with NACA RM E53K?35 .
ethylene, the fuel.weight specific impulse of ethylene was greater than .. -.
that of the slurry. However, the experimental curves”@’n”d& converge‘“ .— ._ as the fuel-air rat10 was increased. At an ~air specific --se of”159 seconds, which was the ~imum obtdned with ethylene, the fuel weight .- .. . ...
specific impulse of the ethylene and the slurry were 2350 and 1.950sec- — onds, respectively.
In short-range-flight a~lications, the fuel wetght represents,a very small fraction of the total vehicle we I@ and the lower .
fuel wefght specific impulse values obtatied at low thrust levels is therefore of secondary importance. The thmist determines the attainable .. ~ .
flight speed and altitudej the higher thrust ,levelsobtainable with slurry
R
fuel is therefore @ primary imp~ce. The slurry fuel permitted op&r-
--
ation up to an ah specific impulse of 187 seconds.
.— . — Ccmpr icon of fuel voltme and alr spec~ic iqpulse data. - Fuel vol- .
ume spectiic impulse, which is a measure of the volume of fuel consumed, Is presented on figure 7(b) as a function of ‘the alr specific inpulse. A deneity of 18.3 pounds per cubic foot was used for the ethylene vol~ impulse computation. ~iS &sity WaS obtained when the flight fuel tank was pressurized to 1200 pounds per sq,e @h.
The fuel vohnue specific Impulse of the slurry was 2.6 and 2.9 times that of ethylene at air spe- cific Impubes of 140 and 159 seconds, respectively. At air specl.flcim-
.:-
pulse values of 1.!50 to 170, the ~erhen~l fuel volume @pulse of the d.
slurry was about 10 percent lower than the ~deal volume @@se for — octene-1.
--- .
.
The fuel volum specific impulse is a significant ~smeter when applied to vehicles that have a small rat10 of fuel weight to gross vehi- cle weight.
For exm@e, the Langley flight vehicle had an ethylene fuel — to gross weight ratio of 0.1.
Hence, a Mge increase b volume specific fuel conmnuption can make possible a correspond- large increase in fuel load with only a small increase in vehicle gross wei@t.
---- -.
SUMMARY OF REEmJJ!s 1. The results obta-d with a 50 percent magnesium slurry fuel in a flight type, 6.5-tich-diameter reM-jet engine in a connected-pipe .. ..-.
facility are as follows: a. Starting disks, mounted h ef~r &e c~tra~ or exit sta- ___ tions In the combustor, provided quick, dependable starting of the - engine over the equivalence ratio range imestlgated, 0.36 to 0.63. “- A disk mounted in front of the exit nozzles was expelled within “- - .
0.1 second eSter fml flow initiation. No ex@oslve starts were experienced. .
b. A flame-holder protection plate which covered most of the .
flare-holder.. lmibecavity prevented ~~ to the flame holder during - —.— __ .
NACA RM E53KD5 Slight damage tests of 20 seconds’ duration or less. occurred In plate, appre- two tests of longer duration. Without the protectlm ciable flame-holder demage occurred dur~ a test duration of 3 seconds.
c. A short, 8.5-inch fuel-air mixing length was used in order to eliminate the intermittent conknzeticm upstream of the flame holder which had been e~rlenced in prevloue free-Jet tests of the engtie. The short mixing length resulted in an appreciable decrease in combustion efflci-y frcm values previously achieved with a longer mixing length. The greater part of this performance loss was recovered, huwever, by redesigning the fuel-distribution control sleeve. The conbustor efficlency for the best configuration with the.8.5-inch ~ng length was shove 77 percent for the equivalence .
ratio range from 0.5 to 1.0 and reached a maximum of 81 percent at an equivalence rat10 of 0.7.
d. An increase in average Met-air temperature from 63° to 367° F ticreased the conibustor efficiency from 56 to 78 percent at en equivalence ratio of O.5, and the increase was from 58 to 70 per- cent at an equivalence ratio af 1.0.
.
2. The following results compare the performance of the slurry fuel In the short, 8.5-inch fuel-air mixing length with that of ethylene fuel, . both evaluated under slmiler conditions. T!heethylene data were obtained from free-jet tests of a ram-Jet engine designed for a flight-test vehi- cle at the NACA Langley laboratory.
a. Tbe fuel volume specific impulse of the slurry was 2.6 and 2.9 times that of ethylene at alr-epecific-impulse values of 140 and 159 seconti, respectlvel.y.
b. The mexlmum alr spectiic impulse obtained with the slurry and with ethylene were 187 and 159 seconds, respectively.
c. The fuel weight specific -ses of ethylene and slurry fuels were 2350 and 1950 seconds, respectively, f= th =imum ah specific Impulse obtained with ethylene, 159 seconds.
CONCLUSIONS A starting disk was developed which Is rec~ed for use in a pro- .
posed NACA Langley flight-test vehicle. The disk provided reliable .
starting of the engtie ati was expelled rapidly after iguition.
— ~ the Langley flight-test vehicle, a greater fuel load and a greater .
thrust would be possible if the ethylene fuel were replaced by a magnesium - – 14. . NACA RM E53K05 bb?ry.
Therefore, higher flight speeds, +1.titudes, and longer flight durations should be attainable with the slqrry fuel.
— —.- Lewis Flight Propulsion Laboratory National Advisory Committee fcm Aeronautics Cleveland, Ohio, Noveniber9} 1953 REFERENCES .-..— 1. Branstetter, J. Robert, Gibbs, Jemes B.; and KWfbmn, Warner B.: Ma~eslum-Slurry Conibustion Performance in 6.5-Ihch-Dl-ter Ram- Jet Engtie Mounted In a Connected-Pipe Facility. NACARME53E27, 1953.
2. Faget, Maxlme A., Watsonj Raymnd S., mld Ba@lett, Walter A., Jr.: Free-Jet Tests of a 6.5-~ch-Diamter,Rem-Jet En@e at Mach””Numbers of 1.81 and 2.00. NACA RML50L06, 1951.
3. Faget, Maxims A., and Dettwyler, H. Rudolph: Initial Flight Mvestl- gation of a Twin-Engine Supersonic Rem Jet. NACARM L50K10, 1950. ‘“ 4. Dettwyler, H. Rudolph, .andBond, Aleck C.: Flight Performance of a Twin-Enghe Supersonic Ram Jet from 2,300 to 67,200 Feet Altitude.
..
. —.
NACARML50L27, 1.951.
.
5. Rudnick, ~ilip: Momentum Relations lnIWpul.eive Ducts. Juur. Aero.
Scl., vol. 14, no. 9, Sept. 1947, pp. :540-544.
6. Breltwleser, Roland, Gordon, Sanford, and Gammon, Benson: Summary — Report on Analytical Evaluation of Air and Fuel Specifi.c-Iqulse Characteristics of Several Nonhy&ocerbon Jet-Engine Fuels. NACA ,.
RME52108, 1953.
.— 7. Tower, Leonard K., and Gemmm, Benson E.: Analytical Evaluation of Effect of Equivalence Ratio, ~et-Air Temperature, and Ccuibustion Pressure on Perfonmnce of Several Possible Rem-Jet Fuels. NACA —.. ..
.— RME53G14, I-953. -.
8. Lord, Albert M., end Eveus, Vernida E.: Effect of Particle Size and Stabilizing Additives on the Caibustton Properties of bhgneslum .-. .— slurry. NACA RM E52K12, 1953.
9. Simon, Dorothy M., and Wong, Edgar L.: “FlameVelocities over a Wide Cowositionllange for Pentane-Air, Ethylene-Air, and Propyne-Air- .
Flames . NACA KM E51H09, 1951.
.
1 , ,-.
M NACA RM E53K05 .
TABLE I. - ANALY818 OF HU)ROCAK80N CARRIER FUEL Fuel properties -F-562U wade JT-4 A.S.T.M. distillation D86-46,% IMtial boiling petit 140 Percent evaporated 10 222 30 268 3m Final botling point Residue, percent 1.0 8pectiic gravity 0.768 Reid vapor pressure, lb/sq in. 2.5 Hydrog=-carbon ratio 0.167 Net heat of ccnibustion, Btu/lb 18,675 .
.
16 NACA RM E53K05 .
mm Dick mlet- camus- EwIva- Timeto Peak Time mmarke tlcm lealob ~t ~ ~ ~*- ratio fulltbmlet, expel. tit t%- a~ erature, flow, Seo atal’ii etatio % lb/Ma lli8k, ~sme, 60 lm eea. rl+lu h.
&m 1 molw 3m U.z 0.66 0.9 1.0 w tithout aiak 2 mm 156 16.4 .55 Hoe- m start s 1 b mtart railm b ~ nt~ta . Diek- 5 Imlmlg strapsfailed ‘1 2 Tim start fnilureo in el@t8tWt~. m8t0n- 9 ~m box alsk b 10 Smpe railed u.
u u .
M 3 78 15.s 0.69 1.!5 2.s Is 76 17.3 .56 1.1 1.6 17 76 15.8 1.1 2.2 16 75 1.6.2 l * :Z ‘kc&ocmllwlti-~mlm trace ls 5 Dti fail.e~ lWChtUliCd4 20 ~icm to fuelflow 21 initintimin ench atteqpt 23 & 72 17,4 0.57 1.s 2.0 24 6 66 16.7 O.1 65 0.5s 25 66 16.2 .4 26 6s 1.6.1 :: 63 .Z 27 375 15.7 .8 62 .56 28 .42 70 w X5.1 26 250 LS.7 .54 :: 62 -* so 7 366 X.7 0.62 57 %Amnmnt-ekwt drive notcmatstar’td run 0.1 51 557 3s.2 .51 56 .
, CJ-3 ?im, , .
TABL6 III. - S-Y OF COMBUSTION DATA F(lR 6.5-INCH-DIAMETKR RAM-JET ENGINE pllu%?y: 50 peroent magnesium powder In MIL-F-5624A grade JP-4 f’uelq s!
-d-u--- %mUnm-m.
“i i“ ..- . .
.
. . . .
I Ikhtnlg+alm (Imm9) KkmmtOr ay~ rmltQlkd9dJuul for yield Dtrem II Of momi ~.
o.%?4-In.-ti
n -“ /’
Rubimr di@na~ oil — —
&
mum E!
I .
l , ,, ;T60S ‘ ‘ ,:, ,.
:, , .
, (XT-3 baak 5001 , .
E
A14° 1“3.
Rmtmtim mite-l Mshnca h open area, Statim refemnm, aq ft.
in.
1 I 4.5 0.005 42.5 0 .nl 9 46.8 .IS6 I I 1 I mam-hcildm’ detail 2.
- Bs=jat mgina.
. .
-.
-.
NAch Ru E55K05 (b) Roazle-mnmted disk.
A“ Aluminum-rivet -.
J
fautxmam lhbar of Disk Bl@ked Msk A B c E fasteners material area, paroenta ~,, 2 6 33 ‘-” 35”- 69 +“ Al 52s 6s z 6 & ‘- 2 &.n m 52S 2 10 3 4 69 3 -- E $“ Al 24S-T 57 4 4 3 m & . -- -. 3“A152a 62 5 4 - = - .- -- 4 - i“ u ‘-’ “ + - &“ U 2&.T ~ . -- -.
(a) Flam-holdar-mounted disk. 7 -. ..— %ercant of combuetor =ea. .- - Mwama of sttil~ disks l .
Figure 3.
.
.
,. , 1 3091 ““ , , I .
. .
NACA RM E5?K05 .
, s , , , 1 Run: Inlet-air Flares-holder temperature, protection *150 F plate YeO o 5, 7, U, 13, 15, 63 16, 17, 19, 20, 23, 24, 26 q q 27, 28 367 Ye6 o 8, 12, 22, 25 63 No q L c L~ \ o % g 70 $ Q / ~ \ \ “ .
i = 60 .
E a B ti fl u o v v 30’ .2 .4 .6 .8 1.0 1.2 1.4 .
Equlvelence ratio Figlwa 5. - Effect of inlet-alr tqature and flame-holder protection plate . .
on combuetor performance. Fuel, 50 percent qagnesium IU MIL-F-562M grade JP-4; fuel-distribution controlsleeve,2 incheslang end 4 Inchesin diem- eter;fuel-alr tixhg length,8.5 Inches.
t14mRF-
,.
.
.. . .
.
r 1 1 1 1 t I I I I I Inlat-ais Pwl-dlatIabntilm nm.1.-ail’ Cmltrol deavu Kwng lawth, tq9r&ra# La&i% -&b ~“ . .
m A 29, SO 8.5 4; sl z a.5 237 v z —.— 527 87, 22 (cm’vu 2 4 8.5 frcm fig. 5) 34” --—— &f. 1 4 ‘ ‘“0 m - Tail.ad E@ol imlicates point Vhara gutial loon of flare holdar oc!~.
_– I I I I I I I I I J I /y
t-ttt’’’-x-’’’’’””
/
\
\,
0“-
u 1.6 1.s ?2 .4 .0 .6 1.0 1.2 1.4 Eqnivul.anca ratio --- .
.
W RM E55K05 .- —..- -..
.
.
.
Fuel-air : ratio ,0.03 \ \ \ / .04 ‘! ,\ \ / \ \ Experilmltal \, ‘, ———— Ideal \ \ \ ‘\\ \ { \ \ \ \ .
\ f~\ .06’ ~6 * \ \ \’ , .
\ \ \ \ ~ Msgnealua Blurry / \ / \ \ \ . my / “ \ / A / \ 160 160 m m la 140 Air epecffic iIW13,ee, oec .- (a) Fuel weight specific Iawlaa - fmLCWOII of afi ep-lfic *U-c . . - _ three fuele. Combuator-inlat Figure 7. - Comparison of impulse date for EItagnat Ion temperature, 330° 1?} aonlc dlachrge of exheuat product a.
.
~erlwntal elurry dateobtelnad - mm S1 of -e 6 end e~arimn- taiethylene date obtained frcm refarancd 2.
NACA RM E53K05 25 .
3’6)@c13 \ !
‘2 \ \
Erperilmlta
\
!
————Ideal ‘, \ ‘,\ t t < ‘, \ g s \’ m \’ \ \ ‘\ II \ —— ~m 1, Octene-1J ~ ‘r \ 1!
~ 16 ~ \ ~ \ ‘\ ~ \ \ \ ‘\ h 1’ \ Ethylene /A \ d \ \< Magneelum Bmrry \ \ \ \ 160 160 200 lm 140 A1.rapeciflc impulse, t3ec (b) Fuel volum? apeclfic lqulee as function of ah qecific impulse.
Ccmpariecm of lqmlse data for three fuele.
. -e 7. - concluded.
Combuetor-inlet stagnation ~erature, 3E410 F; ecmic diacherge of exhaust products. 3SxPerimd%al slurry data obtained from mu 31 of figure 6 and experimentalethylene data obtained fhcm .
reference 2.
.~ NAcA-LaI@9y - 1-M44 -an