CBlowoff Velocities of Spherical Flameholders by Alexander Weir, Jr., Donald E. Rogers, and Robert E. Cullen Project MX- 772 USAF Contract W33-038-ac-21100 Willow Run Research Center Engineering Research Institute University of Michigan UMM- 74. September 1950

WILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN l~. —----------- UMM-74 TABLE OF CONTENTS Section Page Foreward ii Abstract iii List of Illustrations iv I Introduction 1 II Description of Apparatus, Materials, and Method 4 of Testing III Spherical Flameholder Experimental Data 16 IV Miscellaneous Experiments 55 V Correlation of Blowoff Data 75 A. Previous Attempts 75 B. Development of Empirical Correlation 83 Bibliography 99 Distribution 102

WILLOWX RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 FOREWARD The experimental work reported here is part of the experimental work performed by the University of Michigan Engineering Research Institute, Project M772, at the Willow Run Research Center, Willow Run Airport. This work was performed under Air Force Contract No. W33-038-ac-21100, which was initiated in July, 1948, by Mr. G. L. Wander, Fuels and Lubricants Branch, Power Plant Laboratory, Air Materiel Command. In January, 1950, this contract was transferred to the jurisdiction of Mr. D. G. Samaras, Office of Air Research, Wright-Patterson Air Force Base. The experimental work on Project M772 was supervised by Mr. R. B. Morrison, whose aid and encouragement is sincerely appreciated. The experimental work (and the calculations involved) reported here was performed by A. Weir, D. E. Rogers, and R. E. Cullen. This report was prepared by A. Weir. A. W. Stohrer aided materially with the computations and plots required for this final report. Faculty advisors to Project M772 were Professor E. T. Vincent, Chairman of the Mechanical Engineering Department, and Assistant Professor J. W. Luecht, Aeronautical Engineering Department, University of Michigan. ii

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 ABSTRACT The effect of different variables upon the mass velocity at blowoff of a spherical flameholder was determined. The experiments were performed in an unconfined jet under the following conditions: Variable Range Propane-Air Ratio 0.060 to 0.100 lb propane/lb air Spherical Flameholder Diameter 0.0625 to 0.189 inches Ambient Pressure 0.4 to 1.0 atmospheresNozzle Diameter 1/8 to 1 inch Height of the Flameholder above Nozzle 0 to 1-1/8 inches An empirical equation was developed which represented the experimental data obtained under the above conditions. In addition to the study of the effect of the above-mentioned variables on the mass velocity at blowoff, several miscellaneous experiments and observations are reported. Flames were extinguished by imposing external sounds over a wide range of frequencies and intensitites. The mass velocity at blowoff was considerably increased by inserting glass wool upstream of the nozzle, probably because of the reduction of resonance between the flame and the upstream vessel, since shadowgraphs indicated the flow conditions were essentially unchanged. In the small "pilot" flame obtained just before blowoff, it was observed that a luminous region in the pilot flame always moved towards the flameholder before blowout occurred under vacuum. Static and total pressure measurements were'made throughout the flame region, as well as electrical conductivity measurements. It was found that a cylinder inserted at a certain critical distance downstream of a spherical flameholder increased the flame stability markedly. It was established that removal of the flameholder boundary layer considerably decreased the flame stability limits, and under certain conditions extinguished the flame. iii

WILLOW RUN RESEARCH CENTER- UNIVERSITY OF MICHIGAN UMM-74 LIST OF ILLUSTRATIONS NO. PAGE 1 Variables Affecting Mass Velocity at Blowoff of Spherical Flame Holders 3 2 Schematic Diagram of System 1 5 3 Schematic Diagram of System 2 6 4 Typical Flame Photographs 7 5 Photographs of Burning from a Spherical Flame Holder under Vacuum 9 6 Photographs of System 1 10 7 Photograph of System 2 11 8 Photograph of Burner Assembly in System 7 12 9 Mass Velocity at Blowoff of Spherical Flame Holders in 1 in. Jet at Atmospheric Pressure 17 10 Mass Velocity at Blowoff versus Absolute Pressure for 1/16 in. Diameter Spherical Flame Holder 18 11 Mass Velocity at Blowoff versus Absolute Pressure for 3/32 in. Diameter Spherical Flame Holder 19 12 Mass Velocity at Blowoff versus Absolute Pressure for 1/8 in. Diameter Spherical Flame Holder 20 13 Mass Velocity at Blowoff versus Absolute Pressure for 0.189 in. Diameter Spherical Flame Holder 21 14 Mass Velocity at Blowoff for 1/16 in. Flame Holder under Vacuum 22 15 Mass Velocity at Blowoff for 3/32 in. Flame Holder under Vacuum 23 16 Mass Velocity at Blowoff for 1/8 in. Flame Holder under Vacuum 24 17 Mass Velocity at Blowoff for 0.189 in. Flame Holder under Vacuum 25 18 Mass Velocity at Blowoff versus Nozzle Diameter - No Glass Wool Upstream of Nozzle 27 19 Mass Velocity at Blowoff versus Nozzle Diameter with Glass Wool Upstream of Nozzle 28 20 Shadowgraphs Showing Variation of Height of Completely Laminar Zone with Different Mass Velocities - for 5/8 in. Diameter Nozzle 29 i iv

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 LIST OF ILLUSTRATIONS (CONT'D) NO. PAGE 21 Shadowgraphs Showing Variation of Height of Completely Laminar Zone with Different Mass Velocities - for 1/2 in. Diameter Nozzle 30 22 Shadowgraphs Showing Variation of Height of Completely Laminar Zone with Different Mass Velocities - for 3/8 in. Diameter Nozzle 31 23 Effect of Nozzle Diameter on Total Pressure at Blowoff 32 24 Effect of Insertion of Glass Wool on Mass Velocity at Blowoff 34 25 Mass Velocity versus Height of Completely Laminar Zone 35 26 Effect of Reynolds Number on Height of Completely Laminar Zone for Various Diameter Jets 36 27 Effect of Flame Holder Height on Mass Velocity at Blowoff 3/8 in. Nozzle, No Glass Wool, in Test Chamber 37 28 Effect of Flame Holder Height on Mass Velocity at Blowoff 3/8 in. Nozzle, with Glass Wool, in Test Chamber 39 29 Shadowgraphs Showing Effect of Height of Flameholder Above 3/8 in. Nozzle on Combustion for a Mass Velocity Range of 3.1 - 3.2 40 30 Shadowgraphs Showing Effect of Height of Flameholder Above 3/8 in. Nozzle on Combustion for a Mass Velocity Range of 5.7 - 6.0 41 31 Shadowgraphs Showing Effect of Height of Flameholder Above 3/8 in Nozzle on Combustion for a Mass Velocity Range of 9.4 - 9.7 42 32 Effect of Insertion of Glass Wool on Flameholder Height versus Mass Velocity Plot - 3/8 in. Nozzle 43 33 Effect of Jet Confinement on Mass Velocity at Blowoff 3/8 in. Nozzle - No Glass Wool 44 34 Effect of Net Confinement on Mass Velocity at Blowoff 3/8 in. Nozzle - with Glass Wool 45 35 Effect of Flameholder Height on Mass Velocity at Blowoff - 5/8 in. Nozzle, No Glass Wool, Unconfined Jet 46 36 Effect of Flameholder Height on Mass Velocity at Blowoff - 5/8 in. Nozzle, Glass Wool, Unconfined Jet 47 L V

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 LIST OF ILLUSTRATIONS (CONT'D) NO. PAGE 37 Shadowgraphs Showing Effect of Height of Flameholder Above 5/8 in. Diameter Nozzle on Combustion for a Mass Velocity Range of 3.83 - 3.86, with Glass Wool Upstream of Nozzle 48 38 Shadowgraphs Showing Effect of Height of Flamholder Above 5/8 in. Diameter Nozzle on Combustion for a Mass Velocity Range of 3.70 - 3.71, without Glass Wool Upstream of Nozzle 49 39 Shadowgraphs Showing Effect of Height of Flameholder Above 5/8 in. Diameter Nozzle on Combustion for a Mass Velocity Range of 8.75 - 9.17, with Glass Wool Upstream of Nozzle 50 40 Effect of Insertion of Glass Wool Upstream of 5/8 in. Nozzle on Mass Velocity at Blowoff 51 41 Effect of Jet Confinement on Mass Velocity at Blowoff of 5/8 in. Nozzle 53 42 Effect of Insertion of a Cylindrical Rod into a Pilot Flame 56 43 Surface Suction Flame Holder 59 44 Photographs Showing Effect of Boundary Layer Removal on Flame 60 45 Blowff Velocity of Flameholder Shown in Figure 43 61 46 Separation Experiment 62 47 Photograph of Flame at Conditions Used for Pressure Traverses 66 48 Vertical Tube Pressure Traverses of Flame 67 49 Horizontal Tube Pressure Traverses of Flame 68 50 Pressure Measurements of Flame along Center Line 69 51 Pressure Measurements of Flame 71 52 Electrical Conductivity of Flame 73 53 Fuel-air Ratio versus 1 at Maximum Blow1 Reynolds No. off Velocity 76 54 Reynolds Number at Maximum Blowoff Velocity versus Flameholder Diameter for Various Pressures 77 vi

WILLOWN RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 LIST OF ILLUSTRATIONS (CONT'D) NO. PAGE 55 Development 56 Development 57 Development versus B 58 Development 59 Development 60 Development Ratio of Correlation Log G versus Log P of Correlation Fuel-air Ratio versus A of Correlation Flameholder Diameter P 1 of Correlation Log G versus Log ( O) 100 a of Correlation Log D versus C of Correlation Log D Versus Log Fuel-air 84 85 86 88 89 90 92 94 96 98 61 Effect of Nozzle Diameter on Mass Velocity at Blowoff 62 Development of Correlation G versus G - 4 (hD 1-84)2 - 0.6 65 Development of Correlation Dj versus f 64 Correlation of Experimental Blowoff Data vii

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 I. INTRODUCTION This paper is a study of some of the factors which change the flame stability limits of a flameholder. This study is important because of the combustion problems experienced with the operation of jet-propulsion devices. A ramjet combustion chamber is simply a duct with a flameholder inserted. The fuel-air mixture enters at one end, is partially burned in the wake of the flameholder, and leaves the combustion chamber with increased momentum due to the addition of heat. However, experiments with these burners indicated that adequate flame stability was necessary to secure operation of the ramjet. In addition, operation at high altitude conditions indicated that pressure was an important variable to be considered in the design of a ramjet combustion chamber. Hence, in order to secure efficient operation of a ramjet, it is necessary to have some understanding of the variables affecting the flame stability limit or blowoff velocity limit of a flameholder. There are a great number of variables which affect the blowoff velocity of a flameholder. These may be divided into two groups, those variables which are purely physical in nature, and variables whose effect is due to some physico-chemical aspect. These variables are not all independent of each other and there is no clear cut distinction between the two groups. In the first group might be considered variables, such as the shape (sphere, cone, cylinder, grid, orifice, disc, or flat plate) and orientation of the flameholder, the diameter or some characteristic dimension of the flameholder, the distance of the flameholder from the nozzle or jet exit, the diameter of the nozzle, the degree of confinement of the jet, the velocity distribution of the jet, and the turbulence of the jet. The second group might include such items as the type (paraffin, olefin, etc.) and state (gaseous or liquid) of the fuel, the heating value of the fuel, the droplet size and distribution of a liquid fuel, the homogenuity of the fuel-air mixture, the temperature of the fuel-air mixture, the temperature of the flameholder, the surface material of the flameholder, the presence of combustion products and water of humidity in the fuel-air mixture, the intensity and frequency of external sounds as well as resonance effects due to the flame itself, and the ambient pressure of the atmosphere. While there are probably other factors which also affect the blowoff velocities of flameholders, all of the above experimental variables have been investigated to a greater or lesser extent, either in 1

WILLOW RUN RESEARCH CENTER- LNIVERSITY OF MICHIGAN UM-74 this laboratory or at laboratories of other institutions. In the experimental work reported here, many of the above mentioned. variables have been minimized or eliminated in order to reduce the problem of flame stabilization to a more basic level. To that end, a single paraffin fuel, gaseous propane, was used to eliminate the problems inherent with liquid fuels. This fuel was mixed thoroughly with air reasonably free from water vapor or other combustion products. Spherical flameholders were used to eliminate the shape and orientation effects. An unconfined jet with a rectangular velocity distribution was used. Those parameters which were intentionally varied were fuel-air ratio, sphere diameter, ambient pressures below one atmosphere, nozzle or jet diameter, and the height of the flameholder above the nozzle. Some idea of the range of mass velocities* used. in this investigation may be seen in figure 1, which shows the effect on the mass velocity at blowoff caused by changing each of the above variables in turn. It is the purpose of this paper to show quantitatively the effect on mass velocity at blowoff of the five variables mentioned above, i.e., flameholder diameter, fuel-air ratio, ambient pressure, nozzle diameter, and flameholder height. * Mass velocity is defined as the mass flow per unit area of the nozzle (lbs/sec/sq.ft. of nozzle area) and is numerically equal to the product of the average linear jet velocity (ft/sec) and the average density (lbs mass/cu.ft.) of the fuel-air mixture issuing from the jet. It can be regarded as proportional to the thrust of a ramjet. I 2

WILLOW RUN RESEARCH CENTER- UNIVERSITY OF MICHIGAN I —- -------- UMM-74 I I L I I I.1 I! 0:,oL J o u-. J1 I 0.1 R I I I I FUEL-AIR RATIO no i i i — -I — - - -- -.08 Ii D=.062' P=1 ATM. Dj.375' h.187' -t 0& iu |i i - j'O E o E.07.06 FLAMEHOLDER DIAMETER F-A. PRESSURE F A.- 0.080 _____ DJ-a.375' h hu.187' 400-C~'y - - - - - - - - - - - - - - - - - - - -.125.0I 62" 800 h.187' FLAMEHOLDER HEIGHT.375 --— I F/A 0.080.50 1^^._ P=1 ATM. 0 ~~~~~~~~-.375 12 5 — z ~Il ATM.0.250 - - - - - -- - DJ a.375' - ~55.1252 0 5 10 I 5 20 I u UJ O Z ~i ur O.'T' U MASS VELOCITY AT BLOWOFF (LBS./SEC./SQ. FT.) FIG. 1 VARIABLES AFFECTING MASS VELOCITY AT BLOWOFF OF SPHERICAL FLAME HOLDERS

WILLOW RUN RESEARCH CENTER UNIVERSITY OF MICHIGAN UMM-74 II. DESCRIPTION OF APPARATUS, MATERIALS, AND METHOD OF TESTING 6 The equipment used during this investigation was designed to measure the mass velocity at blowoff for various fuel-air ratios and ambient pressures, the physical variables (i.e., spherical flame holder diameter, nozzle diameter, and height of the flame holder above the nozzle) being held constant. These physical variables could then be changed and the experiments repeated. Two different systems were used during this investigation. In System 1, a schematic diagram of which may be seen in figure 2, the fuel mass rate and the air mass rate could be varied during operation, the ambient pressure (atmospheric) being held constant. Blowoff of the flame was obtained in this system by setting some air flow and then changing very slowly the fuel flow rate until blowoff occurred. Identical values were obtained if the fuel flow rate was set and the air flow rate changed slowly. A schematic diagram of the second system used, System 2, may be seen in figure 3. In the operation of this system under vacuum, it was found advantageous to let the fuel-air ratio be held constant during an experiment by using a premised fuel and air mixture of known composition. The total mass flow and the ambient pressure could be varied during the experiment. Blowoff would occur in this system if the pressure were maintained constant and the mass flow increased, but operation was simpler if a certain flow of fuel-air mixture was set on the rotameter and then the pressure decreased slowly until blowoff occurred. In both System 1 and System 2 a stable pilot flame was observed just before blowoff except when blowoff occurred at low jet velocities or when the flame was accoustically disturbed. In figure 4 may be seen photographs of a flame burning from a holder at jet velocities of 186 and 268 ft/sec. As evidenced by these photographs, it may be seen that at the jet velocity of 186 ft/sec, steady state combustion is maintained far downstream of the holder while at the higher jet velocity combustion is confined to a very small region immediately downstream of the spherical flame holder. The photograph on the right in figure 4 (i.e., at the higher jet velocity) is a typical photograph of the pilot flame seen before blowoff. An observation concerning this pilot flame was made during the experiments performed in System 2 under vacuum. This observation, which enabled 4

6" INCLINED /MANOMETER H9 MANOMETER GLASS THERMOMETER HzO MANOMETER FINE NICHROME WIRE AIR SPHERICAL FLAMEHOLDER CONTROL 4000 cfm _ U I \~~~~~~~~~~ yl~/^ WT~ BirjiBir i FCENTRIFUGAL BLOWER 1 " NOZZLE.__......... I. JOOiii~[;~iii[ 100 MESH Hg... Cu SCREENS MANOMETER ~ Ji *9"iglJiii — aii ii,,.. AIR ORIFICE vi~iaffi~ii ii:ts *ii~w *AR ORFIC 4 ROTOMETER 25 HP ~/ rtttit~tirir''''''ii^^24i J~ g MOTOR P ii0 |~5#BiBBS~:i|i ~ GLASS 60 Cu. Ft. SURGE TANK 5 ~iiiiiii~L':ilpirr;liS l THERMOMETER PROPANE PROPANE CONTROL CYLINDER VALVE PLENUM 1 CHAMBER""" FIG. 2 SCHEMATIC DIAGRAM OF SYSTEM 1 "'..............."..............~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..~~~~~::::..........~~~~.. ~ FIG. 2 SCHEMATIC DIAGRAM OF SYSTEM 1 C Ir 0 to z C ti1 01 O. O.O 1

S= |3f~~~'Z W31SAS 40 WVtVaOV 31IVW3HDS c'O1 0 1-1: 3A1VA -- ---- ------- _1OSINO ) Czq J,[a,t^^ "anl a~N~ne c' S3AnlXIWM IV-3NVdOMd a3XIW3id JO S3QGNIIAD 6 2Q *~!C.%1 33ni 83Nan9 — ^ 1 ||||| || |. sls3l wnnDVA o J a3snA*!a%'-. 0(3 O I. 1 1 1%1 i —I. \t(~~ S31ZZON 31aV3ONYHD331NI' ~,, 1 x,Z L x ao 1W N W 1L \ dD r:..~:!:i i....i:::i:!::ii:9.{:VH I lS 1.........-.o.^ g~awvH~is~i inriS=J I n (1 U N^ ^^ ^s. S0 -.1 * l^fi vrtgI -U, 1 yh313WOW 3H1 SSV1 / O ( ^ " < @ 8313WZoNV' AIO1VO WOI'I1 zL >, o |WOV8IHdVIQ JoxlOlS 1 OVBO$1S 3QNIlAD W wnnsveA _\, fY HSVN 3OA1NOV O"~~~~ H3 IS31~~SW10af0 z ~ ~~~~~~~~~wnnvn O cl ol

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 CO I M I (4a) Typical intermediate velocity flame Jet velocity - 186 ft. /sec. (4b) Typical pilot flame Jet velocity - 268 ft./sec. FIG. 4 TYPICAL FLAME PHOTOGRAPHS Propane-air ratio - 0.05 Nozzle dia. ~ 1 in. Height of flameholder = 1/2 in. Flameholder dia. a 1/8 in. Pressure: 1 atm. TYPICAL FLAME PHOTOGRAPHS. FIG. 4 7

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 the experimenters to eliminate premature blowoff points, was that a zone of violently agitated and burning gas moved toward the flame holder before blowoff. A sequence of six pictures approaching the blowoff point is reproduced in figure 5. Photographs A, B, C, D, E, and F were taken at increasing velocity and decreasing pressure from left to right with photograph F taken just prior to blowoff. In these photographs of the pilot flame region, the agitated region appears luminous (or white). The agitated region is very close to the holder just before blowoff and moves about quite rapidly. At the instant this region touched the downstream side of the flame holder, blowoff occurred. This phenomena was noted during the atmospheric blowoff runs, however there was not enough contrast between the wake and the agitated zone to allow a photograph or a close observation of the occurrence. During early experiments, an electric spark was passed through the mixture, after blowoff had occurred, to determine whether or not the mixture was combustible. This "spark test" was not used a great deal during the later experiments, since the appearance of the pilot flame aided materially in the determination of the blowoff point. A schematic diagram of System 2 was previously shown in figure 2. Several photographs of this system are shown if figure 6. The upper photograph shows the 4000 cfm. blower, the 60 cu. ft. surge tank and the propane cylinder, while the lower photograph shows the metering instruments and the plenium chamber. The rotameter used to meter the propane, manufactured by Fischer and Porter, was calibrated with propane by a volumetric displacement method as well as by measuring the change in weight of a propane cylinder over five minute intervals. The flat plate, square edged, brass orifice used to meter the air was designed in accordance with the Report of the ASME Research Committee on Fluid Meters and was then calibrated, with the section of six inch pipe and the radius taps used, against a standard orifice in the Ford Research Laboratory, Dearborn, Michigan, through the courtesy of Mr. W. G. Was. The values obtained in the calibration agreed with the values predicted by the ASME Formulas. A schematic diagram of System 2 was previously shown in figure 3. A photograph of this system is shown in figure 7 while figure 8 shows I 8

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN X______________ UMM-74 A B C D E F FLAMEHOLDER HEIGHT - 3/16" FUEL AIR RATIO - 0.080" NOZZLE DIA. - 3/8' SPHERE DIA. - 3/16" PHOTOGRAPH No. ABS PRESSURE mm of Hg A 680 B 465 C 412 D 359 E 333 F 307 FIG. 5 PHOTOGRAPHS OF BURNING FROM A SPHERICAL FLAME HOLDER UNDER VACUUM. 9

WI VI.:...OW I U N RES;EARICI C()ENTVEl:.:15R: -. UNIVVE.1"R S I1TY OF MXCIOtG.AN EUMMA..7h FIG. 6 -CPHOTOGRAPI IS OF SYSTEM 1. 10

I NILLOW RUN RESEARCH CENTER UNIVERSITY OF MICHIGAN MM-74 U1! E-:::'":. -.... + CZ:.........,,, i~............... -: _. i E.....:.;.E E * h..... "|:::':?. -! _,-.-~;.:~ ":~ {...' I t l:: t ].l: 0 >::: r FL -.................: i i;i: l!ii?!!{?,'~ii',ii: 0.i~....U-. ccttJ,.,. I LU f: {i.;,. j E... jij j; jij.j j, nn~ t' v = i x iiE l iiiiip h:: i i | ci.i _. i -- --..........[-............i L. iH..'fo liji..;;... i.:. iijiiiiiii-iiiii::- ii ii 2 iiili I | r..........~s i,iiiiii,:,;.;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...;;i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i:i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...................... liiii'iiiiii'i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~w....... 1 1 1,. i::iiiiiiiiiii~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..................:.-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~................ it.......................................... ~ ~ ~ ~ ~ ~ ~ ~ ~ i.............................................ii i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.............. II Ilil~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iiiiiiiiii:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~............::;::.;;;___:i I@8~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.......................;ilBU3Y~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~; C;~~~~~~~~~~~~~~~~~~~~~..... Uj~~~~~~~~~~~~~~~~~~~~~~~~~~~...... l~~~~~~ii~~~~~~~i::ii liiiii~~~~~~~~~~~~~~~~~~~~~~~~~~~i~~~~~' i.!" ~ ~ ~ ~ ~ ~ ~ ~~~....... i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........... i e ii:;i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~liiliiiiiiiii~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...................................................................:""""""'':':' a ii::::::iill-i~~~~~~~~~~~~~~~~~~~~~iiii~~~~ii~~~iiii::il:......................... iii;; ti~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~......... iii;::I;~~~i~~ii~i,::::i:i~~~iiiiiiiiiiiiiiiiiiiiiiiii i I~~~~~iiii...........::ici::: *Cf Lar::iiii~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iii~~~~~~~~~~~~~iii~~~~~~~~~~~~~~ii~~~~~~~~~~iiiiii~~~~~~~~~~~~~...... iii i- -.:iil:::::iiiiiiiiii: ~ ~ ~ ~ ~~~~~~~~~~~~~~~!::;::_:iiiiii~~~~~~~~~i..................:illF:,.!.{ %,=~;~,..~.? \!;D~,~,..~.:i~i~i~iiiii::;iiiii.;',i iiiiii~ii:i:'~"~::.:;............_...; ^., i i,,^ i "!,,.........,,, I>7~f 1 03'~ E 1/')~s~ I d LL 11

'r 0 Q):: M I:l::; i if E;Ei: A; l i T:;i:iiE'i;........... i.....'.:: ii.::'....:,;:;.. ~j i.-...:........ E L - iiE ^iiiii^ ^ ^ W|r j jI'S^ I i'S! fi: j'Eir' _! E! " -'!:i"::i'",' " SiS^:'i,''";', Z:' J;'/'; i:i _!! iEjE0i, i t. 0 ---------. ~ ~:!- m:.; i.j w;i:: E f.~_:......................:..,!.i'!:;.E E.; E; E.".g.::;itS.E.:...... -: i::li:ii: i: F'4:; g I @: - O m ni L N | i j - 0, - hiij..iijiiiijji iL -:a m,iijji~ijijiij~iijjjl! i~jjijj~- i jg:E!giiiigg~iiiiiiiiii iiigs! i | Z s~ijiiiiiiij iiii~ii.ii F o iwjjjj'jjiji jjij i.i~jjj.-jjiii ulji >;iGiiiiii iiiiiiiiiiiiii~iii is qi:ii - j'qjIjljjiiiiiiiiiii!!!!r! i C~hN x j jiiiiiiiiiiii~ii;. @ t t_ A x jj iii i ibii:'' NVoIHOIwI 0 aI ISI aAINIn-l- HLNhI HOIV3SSSHI NnfI xO'ItIIA

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN,,______________ UMiM-74 an exploded view of the burner tube and the nozzles. The nichrome wire shown on the inner tube was not used during this investigation. The rotameter used, manufactured by Schute and Koettering, was calibrated with an 0.080 fuel-air mixture by a volumetric displacement method at atmospheric pressure. This calibration was checked against previously calibrated rotameters in this laboratory. There is some indication that at very low pressures, the square root of the density ratio method of computing the rotameter flow is slightly in error. Also, at very low pressures, the mass velocity at blowoff is also very low, giving a reading near the bottom of the rotameter scale where the accuracy of the rotameter is in doubt. However, at mass velocities over 2 lbs/sec/sq.ft. of nozzle area it is believed that the rotameter calibration is correct. The fuel used in this investigation was commercially pure propane (95 per cent propane and propylene) obtained from the Flamegas Corporation, Detroit, Michigan. The majority of the data was obtained in System 2, and the propane used in this system was obtained from a single cylinder. The cylinder contents were not analyzed. Several cylinders were used in the investigation in System 1. However, no difference in blowoff velocities was observed when different cylinders were used. The air, supplied in System 1 by means of the 4000 cfm.blower shown in figure 6, had a relatively low humidity when measured in the 60 cu.ft. surge tank shown in figure 6. After the blower had been in operation for several minutes, the air temperature was ca. 100~F which probably accounted for the low humidity (0.01 lb. water/lb. bone dry air) of the air. As was previously stated, in the operation of System 2 under vacuum, it was found advantageous to use a premixed fuel and air mixture of known fuel-air ratio. Accordingly, cylinders containing propane-air ratios (by weight) of 0.060, 0.0637 (stoichiometric ratio), 0.070, 0.075, 0.080, 0.085, 0.090, 0.095, and 0.100 were prepared. The cylinders were filled with gaseous propane, evacuated to 15 mm Hg abs., filled with propane and evacuated a second time, and then filled with propane to a predetermined pressure. The temperature of the cylinder wall was measured at this time with a Leeds and Northrup Galvanometer and a chromel-alumel thermocouple. Air, from a 600 psig source, was then charged into the cylinder in such a fashion, so that when the cylinder was at the same temperature as that recorded during the propane charging, the total pressure was 240 psig. Measurements of the humidity of the air to be charged (which was at a 15

WILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN UMM-74 600 psig and ca. 110~F) indicated that the humidity was less than 0.01 lb. water/lb. bone dry air. The compressibility factors for air and propane, used in calculating the amount of propane to be charged for some desired fuel-air ratio, were obtained from a plot of compressibility factors with parameters of reduced pressure and reduced temperature. The cylinders of propane-air mixtures were placed in a shed ca. 75 ft. from the laboratory building where they were connected to System 2 inside the laboratory. The flame holders used in this investigation were steel ball bearings obtained (before the hardening process) through the courtesy of the Hoover Ball Bearing Co., Ann Arbor, Michigan. Four different size spheres were used, 0.0625, 0.0937, 0.1250, and 0.1890 inches in diameter. No. 80 drill holes were made through an axis of the sphere and the spheres were mounted on 0.005 in. diameter nichrome wire. Various methods were used to mount the sphere on the wire, including knotting the wire in the center of the sphere as well as the use of adhesives, such as sodium silicate ("water glass") or "X-pando" pipe cement. The latter methods were preferable. The spherical flame holders were mounted in the center of the jet as is shown in figure 4. Various heights of the flame holder above the nozzle were used. At low jet velocities, with nichrome wires larger in diameter than 0.013, the flame tended to burn from the wire as well as from the flame holder, but this tendency was not observed with the 0.005 diameter wire used. The diameter of the nozzle used in System 1 was 1 inch. A 3/8 inch diameter nozzle was used in the vacuum experiments in System 2, but nozzles with diameters of 1/8, 1/4, 7/16, 1/2, and 5/8 inch were also used with this system. The radius of curvature of each of these nozzles was equal to the diameter of the nozzle in each case. It is believed that the usual criterion for testing nozzles, i.e., a rectilinear velocity distribution, was met with all of these nozzles. In Table 1 is shown nozzle coefficients which were determined by means of a total head tube. Table 1 follows on the next page. 14

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 TABLE 1 Nozzle Coefficients of 3/8-inch Diameter Nozzle at 1/2 Nozzle Diameter Above Jet L Mass Velocity lb/sec/sq.ft. of Nozzle Area Nozzle Coefficient Average Velocity Maximum Velocity 4 6 10 14 18 22 0.954 0.965 0.964 0.966 0.968 0.970 Comparable values were obtained with the 1/2 inch diameter nozzle in System 2 and the 1 inch diameter nozzle in System 1. 15

WILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN 1X1M-74 III. SPHERICAL FLAME HOLDER EXPERIMENTAL DATA The data obtained with System 1 at atmospheric pressure is shown in figure 9. The diameter of the nozzle used was 1 inch and the flame holders were tested at a height of 1/2 nozzle diameter (1/2 inch). The experimental data was smoothed on this mass velocity plot since it is essentially a plot of the two experimental variables measured during this set of experiments. The area inside of the curves shown in figure 9 represents a condition in which it is possible to maintain stable burning, while the area outside of these curves represents a condition in which it is not possible to secure combustion under the test conditions. As may be seen from figure 9, the fuel-rich blowoff limit is the same curve for the four spherical flame holders tested. Because of the experimental technique used, greater weight was given to points near the outside of the curve since a large number of variables (gusts of wind, vibrations or noises in the laboratory, etc.) may cause premature blowoff, but only errors in observation or measurement will cause higher values of the blowoff velocity. The experimental data which was obtained under vacuum in System 2 is presented in figures 10, 11, 12, and 13. The diameter of the nozzle used in these experiments was 3/8 inch and the height of the flame holder above the jet was 1/2 nozzle diameter (3/16 inch). The data obtained with the 1/16 inch diameter spherical flame holder is shown in figure 10, which contains plots of the absolute pressure versus the mass velocity at blowoff for the nine fuel-air ratios tested. Similar data for the 0.0937, 0.125, and 0.189 inch diameter flame holders are shown in figures 11, 12, and 13. Data obtained from the smoothed curves shown in figure 10 was used to obtain the mass velocity plot shown in figure 14. This plot is similar to figure 9 except that parameters of pressure rather than flame holder diameter are used. It is to be noted that, as in the case involving flame holder diameter, the curves are concurrent on the fuel-rich side. Curves similar to figure 14,except for other diameter flame holders, are shown in figures 15, 16, and 17. However, the data obtained in System 2, although showing the same trends, did not have the same numerical values as the data obtained in I

FUEL-AIR RATIO (BY WEIGHT),,) sO o9 0o 0' 0' 0. 0' 0- 0. 2.5 // // ooF SPHERICALFLAMEHOL0.07 0.0637 -:-.~ / | QV0 \ 0.06 2.0 0! 55:^ 1 I 20 2500.05 04,1.5 <AN - 0.04 0:0d 1.0, cn ] IFIGURE-9 MASS VELOCITY AT BLOWOFF w -OF SPHERICAL FLAMEHOLDERS ( 4) -0.0625 INCH DIA. SPHERE ^*~m @ -0.0937 INCH DIA. SPHERE w-. Q -0.1250 INCH DIA. SPHERE 0.5 0 -0.1890 INCH DIA. SPHERE NOZZLE DIAMETER * I INCH HEIGHT OF FLAME- I HOLDER ABOVE NOZZLE eINCH PRESSURE * I ATM. COMPLETELY UNCONFIN JET I I I I I I I I I I /1 I i I I 0 5 10 15 20 25 30 35 AIR MASS VELOCITY AT BLOWOFF (LB&/SEC./SQ. FT. OF NOZZLE AREA FIG. 9 MASS VELOCITY AT BLOWOFF OF SPHERICAL FLAMEHOLDERS IN 1" JET AT ATMOSPHERIC PRESSURE r r C 0 Z z I C) ^ o4 xr?0 I I)

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 E CLJ I it) U)J 0 0 F/A 0.100 J61k ^^""""0 400 0 0 4 8 12 16 800 F/A 0.095 400 0 0 4 8 12 16 800 F/A - 0.090 400 0 0 4 8 12 16 800 F/A. 0.085 400 0 4 8 12 16 800 - F/A -0.080 400 - 0 4 8 12 16 800 F/A - 0.075 400 1 0 0 4 8 12 16 800 F/A - 0.070 0 400 0 4 8 12 16 800 0 / |' F/A - 0.0637 400 0 0 4 8 12 16 800 F/A ~-0.060 400 0 0 4 8 12 16 FIGURE 10 MASS VELOCITY AT BLOWOFF vs. ABSOLUTE PRESSURE SPHERICAL FLAMEHOLDER DIAMETER * 1/16" NOZZLE DIAMETER * 3/8" HEIGHT OF FLAMEHOLDER * 3/16" MASS VELOCITY AT BLOWOFF, LBS./SEC./SQ.FT. OF NOZZLE AREA FIG. 10 MASS VELOCITY AT BLOWOFF VERSUS ABSOLUTE PRESSURE FOR.0625'IN. DIAMETER SPHERICAL FLAME HOLDERS.

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN _UM-74 I U) U) w en w U) 3 4 F/A 0.100 4 00 400 40C 0 4 8 12 16 800 F/A - 0.095 400 0 0 4 8 12 16 800 F/A 0.090 80C0 4 8 12 16 800 F/A- *0.085 400 — _ --- -- -- -- -- - 0 4 8 12 16 8O 0^ ^^ ^ F/A - 0.080 00 4 8 12 16 Soo 800 F/A -0.07 5 400 0 0 4 8 12 16 800 -- F/A -0.070 400 0 __ -_ 0 4 8 12 16 800 - -- - F/A ~ 0.0637 400 0 0 4 8 12 16 800 — F/A - 0.060 400 0 0 4 8 12 16 FIGURE II MASS VELOCITY AT BLOWOFF vs. ABSOLUTE PRESSURE SPHERICAL FLAMEHOLDER DIAMETER - 3/32' NOZZLE DIAMETER HEIGHT OF FLAMEHOLDER - 3/8" * 3/16" r- Arlr-A MASS VELOCITY AT BLOWOFF, LBS./SEC./SQ.FT. Of NOZZLE AREA FIG. 11 MASS VELOCITY AT BLOWOFF VERSUS ABSOLUTE PRESSURE FOR DIAMETER SPHERICAL FLAME HOLDERS. 19.094 IN.

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 O00 - 800J^ ~ SWF/A - 0.100 400 _______ 0 4 8 12 16 800 F/A - 0.095 0 4 8 12 16 400 -- LJ co I -____ _ ____ _ _ __ ___ ____ (0) 4 8 12 16 0/ U) C 800 0 4. 8 12 16 Soo,,, _ ____ ____ ____ ___ _ ^ - F~";~-~A - o0.080 400 ~0 4 8 12 16 800 —-- F/A - 0.075 400 - 0~__ _ _ _ - 0 4 8 12 16 o00 F/A - 0.070 400 0 t 1 — -1 11- 11 0 4 8 12 16 80 F/A 1-0.0637 400 o/ 0 4 8 12 16 800 F/A -0.060 400 0 4 8 12 16 FIGURE 12 MASS VELOCITY AT BLOWOFF vs. ABSOLUTE PRESSURE SPHERICAL FLAMEHOLDER DIAMETER * 1/8' NOZZLE DIAMETER 3/8" HEIGHT OF FLAMEHOLDER * 3/16 MASS VELOCITY AT BLOWOFF, LBS./SEC./SQFT. OF NOZZLE AREA 20

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 a. 09 I w U) w or mJ G4 800 F/A -0.000 F/A - 0.075 400 0 4 8 12 16 20 0 4 8 12 16 2C 800 1 I F/A 0.095 F/A 0.070 400 0 4 8 12 16 20 0 4 8 12 16 2( F/A * 0.090 F/A 0.0637 400 - 0o 0 4 8 12 16 20 0 4 8 12 16 2( 800 - I I — ^ F/A 0.085 F/A ~0.060 400 0 4 8 12 16 20 0 4 8 12 16 2N 80HEIGHT OF FLAMEHOLDER /A ~f.oso FIGURE 13 o^ "'~""'~ ~ MASS VELOCITY AT BLOWOFF vs. ABSOLUTE PRESSURE 400 SPHERICAL FLAMEHOLDER DIAMETER - 0.189" NOZZLE DIAMETER - 3/8" HEIGHT OF FLAMEHOLDER - 3/16" /I 3 v 0 4 8 12 16 20 MASS VELOCITY AT BLOWOFF, LBSJSECJSQ.FT. OF NOZZLE AREA 21

2.C0 ______________FUEL-AIR RATIO uwgf () 1.5 LL 0u 0 I0 -J w cn I^ — ---.I atm. c) 0.5 ^w y^ /^FIGURE 14 z',' y- -0.8 atm. a<. at. ^MASS VELOCITY AT BLOWOFF OF a. y^_^^ INCH DIAMETER SPHERICAL FLAME0.6 atm. HOLDER UNDER VACUUM IN A INCH DIAMETER NOZZLE 0 5 10 15 20 25 30 AIR MASS VELOCITY AT BLOWOFF (LBS./SEC./SQ. FT.) FIG. 14 MASS VELOCITY AT BLOWOFF FOR.0625 FLAME HOLDER UNDER VACUUM. r r 0 z 2 Xd H 0 I=1 r./] 0 P _2 I I

gQ_2. _______0 _ _____________ __ ___________ _____FUEL-AIR RATIO 2.0 I~~ci~ ~ ~ ~ ~ i 1.5 In n 0/.5 ------------- I atm. I/ -/i —-^ - ------- 0.8 atm. u,) <n = 0.5'w / J^X FIGURE 15 z 0.6 atm. a<. ^ry^ /MASS VELOCITY AT BLOWOFF OF 0 32 a: /y> / INCH DIAMETER SPHERICAL FLAMEHOLDER UNDER VACUUM IN A INCH DIAMETER NOZZLE 0 5 10 15 20 25 30 AIR MASS VELOCITY AT BLOWOFF (LBS./ SEC./ SQ. FT.) FIG. 15 MASS VELOCITY AT BLOWOFF FOR.094 FLAME HOLDER UNDER VACUUM. 0 C) Z I1 z tr 11 ~3 mI z3 Cf Og

FUEL-AIR RATIO 2.0 cj 8; o 2 U) 1.5 U. io 0.a 0 1.0 -JI cn ~ / /88 /6 --— mlotm. (L 0001v 0.4 otm. HOLDER UNDER VACUUM IN A.INCH DIAMETER NOZZLE 0 5 0 15 20 25 3016 AIR MASS VELOCITY AT BLOWOFF (LBS./ SEC./ SQ. FT.) FIG. 16 MASS VELOCITY AT BLOWOFF FOR.125 FLAME HOLDER UNDER VACUUM. r r o z z t O'11 04 mI 0 PIl cli

2.0 FUEL-AIR RATIO U.. o U. u) 4. I atm. --- 0.8 atm. w ~ 0.5 w ~5y^T J^"T"FIGURE 17 4 MASS VELOCITY AT BLOWOFF OF A 0 5 10 15 20 25 30 FIG. 17 MASS VELOCITY AT BLOWOFF FOR.189 FLAME HOLDER UNDER VACUUM. 1-4 r r 0 z z 0 a 0 I —q 0 50 (I c zz ff h^

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 System 1. The essential difference between the two systems was that the nozzle diameter in System 1 was 1-inch, while the nozzle diameter in System 2 was 3/8-inch. As was previously stated, the velocity distribution in both jets was essentially rectilinear. Accordingly, an investigation of the effects of nozzle diameter was performed. Tests were made in System 2 at atmospheric pressure and with a fuel-air ratio of 0.080 with nozzle diameters of 1/8, 1/4, 3/8, 7/16, 1/2, and 5/8 inch diameter nozzles. In each case the flame holder was placed 1/2 nozzle diameter above the nozzle. The data obtained with this series of experiments is shown in figure 18. The data obtained in System 1 with a 1-inch jet is also shown in figure 18. In figures 20, 21, and 22 are shown shadow photographs of the various jets at their indicated conditions. The height of the completely laminar zone may be seen to decrease as the mass velocity increases,and these shadowgraphs indicate qualitatively the degree of turbulence in the jet at the various conditions. It may be seen in figure 18, that for the four flame holders tested in System 2, the largest values of the mass velocity at blowoff were obtained with the 3/8 inch diameter nozzle. However, the values obtained in System 1 with a 1-inch diameter nozzle were greater than any obtained in System 2. In System 2 the increase in blowoff velocity with increasing nozzle diameter up to the 3/8 inch diameter nozzle can be explained logically by blocking of the jet by the flame holder. However, the decrease in blowoff velocity in System 2 with nozzle diameters greater than 3/8 inch diameter cannot be accounted for by this explanation. As was previously shown in Table 1, the usual criterion for testing nozzles, i.e. a rectilinear velocity distribution, was met. This would seem to indicate than an unsteady state, rather than a steady state, phenomenon was causing lower blowoff velocities in the 7/16, 1/2, and 5/8 inch diameter nozzles. To confirm this supposition, total head measurements were made upstream of the nozzle at the moment blowoff occurred. This data is presented in figure 23. The similarity of the curve in figure 23 to the curves in figure 18 is evident. Since the phenomenon appeared to be of an unsteady-state nature, an external sound was imposed on the flame with an audio oscillator, amplifier, and a speaker. It appeared that at low jet velocities, ca. 100 ft/sec, the 26

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN | —--------- U 7UMM-74 - 1 1/8 1 7/8 3/4 u a uJ L12 LU N N 0 Z 5/8 1/2 3/8 I 1/4 1/8 FLAMEHOLDER HEIGHT 50Dj - PROPANE - AIR RATIO =.080' PRESSURE = 1 ATM. - COMPLETELY UNCONFINED JET S WOOL UPSTREAM OF NOZZLE 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) MASS VELOCITY AT BLOW OFF VERSUS NOZZLE DIAMETERNO GLASS WOOL UPSTREAM OF NOZZLE. FIG. 18 27

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 11/8 1 7/8 3/4 5/ 0) U CN N UJ 0 z 3/8 1/8 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) MASS VELOCITY AT BLOW OFF VERSUS NOZZLE DIAMETERWITH GLASS WOOL UPSTREAM OF NOZZLE. 32 FIG. 19 28

j i;;;;,; r r 0 c 2 tt n 3t Without glass wool and 100 mesh screens upstream of nozzle Mass velocity (Ibs./sec./sq. ft. of nozzle area) Reynolds' Number 16.1 6.7 71,400 29,700 3.8 16,600 1.8 7,800 ro I"-0 With glass wool and 100 mesh v screens upstream of nozzle il Mass velocity (Ibs./sec./sq. ft. of nozzle area) Reynolds Number - 13.7 6.9 60,800 30,400 3.9 2.8 17,200 8,061 0 g1 1z IO H 0 t 4 0 2 O r) M~ Propane-air mixture 0.080 (by weight) Pressure 1 atmosphere FIG. 20 SHADOWGRAPHS SHOWING VARIATION OF HEIGHT OF COMPLETELY LAMINAR ZONE WITH DIFFERENT MASS VELOCITIES - FOR.625 IN. DIA. NOZZLE.

WILLOO-W RUN RESEARCH CENTER — UNIVERSITY OF MICHIGAN UM-74 oo Zo <N _J o~~~~~~~~~~~~~~~~~- -=u -~~~~~~~ ~~~,J, o*~~~~ 0 o LU;04 CQ4 oN~~~~~ >. -o..:.:ii;:isl~~i-i~~iii~iir~i~;i:::i l ~ ~........................................... Z~~~: ~;;:::~i::. iji~i.':iii::;::: ~li~;l~j..'-..: o' ~ 2 ux o~~~~~~~~~~~~~~ ~~~~~~~~~~~~=o (/~.._..... o. - d~. d ~ ~~~~~~~~~~~~~~~~~~~~~ —............. iiiiii iiii:;:::il im:iiw..........................................iiil:::iiii;;-iiii ii iiisii: i;;.............................................iiv E............ >. O:ii~ii:i;;::iiiiiis~:~:I 0u.........I.........~...... 0 4 d..........;~~ ~ w S'''i:.................1: i~ii.,.................................................................... I...............=m = i......................................................................... ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Q L r.......... lilllillli"""""~uU Ln W~~~~~~~....................;.:.;. ii-,o o............. i~;;: -.;.. ~t:il~~i~~iil~~;:;-i::;j;~~~i~~iii:~ ~lri::~ ~iiiii; i ii~~~~ii;~~i;; ll::iiiiii::;EN -;:;;::::-:-::::.;:;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...............................i li:i.:i~.~~iii jli? 1 /................ ~ ii: Q i__jr~~~~~~~~~~~~~~~~~~~~~~~...............iiiii~ iiiiiiiii iiiiiiiiiiiii::' i::;ii;:':;:::;iii~~~~~~~~~~~~~~~~iiiii........ I:~~~~~~~~~~..... o;;liiiii:I O~~...... LL. i'iiiiii~~~~iii Q,~ cl) ih" L W Lu CY C3 cu ~o~~~~~~~~~ oo ~ ~?,ii: iiiiiii~~i: ii;j~il'lii~~::~i ill~::;'ll.;'~il~li:i';Cii Oi~ i-;i~ii~iii:;;~ iii<i E 1140~~~~~~~~~~~~~~~~~~~~~~~:-"~iii:;-:iaiiii'q~acii~w: ~i:i;i~li~ii:..l:i ~.:i"iii: 0 0 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i E cr - O d- d 0 Z~ ~ ~ E N r~~~~~~~~~~~~~~~~~~~~~:Ed: L L 30

z l^i 0o Z: Without glass wool and 100 mesh screens upstream of nozzle Mass velocity (Ibs./sec./sq. ft. of nozzle area) Reynolds' Number 24.7 15.8 65,700 42,000 7.5 3.2 19,900 8,600 kH With glass wool and 100 mesh screens upstream of nozzle Mass velocity (Ibs./sec./sq. ft. of nozzle area) Reynolds' Number z C/) cor p t C) 2: H 2 O I.-.O 25.4 67,600 16.4 7.8 43,500 20,600 3.3 8,700 Propane-air mixture 0.080 (by weight) Pressure 1 atmosphere FIG. 22 SHADOWGRAPHS SHOWING VARIATION OF HEIGHT OF COMPLETELY LAMINAR ZONE WITH DIFFERENT MASS VELOCITIES - FOR.375 IN. DIA. NOZZLE.

---- -- - -------—::: ---—:-_,^ -1^;^^^^^^^^^^^ ^ ON I -F-K-:J IPB _~~~~~~~~~~~~~~~~~~OA _ 44 _I _i- _ _~ FFEC —- -OF — 3Z1E-I ---—; —— 1 -ONT — 1-~-.r... i- f~OT~t PRE US~R AT 0 F / W ASS 001..I~-;OIR 1-00'1 A AS A IL e~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~!U 1 ----—:- ---- -- - -` —- -- - ----'t —- -:- - —' —--— "-t- -— i~tt -- ---- -t — 1 —-~ -~ —-t -~-~I — - i —- ---—: —— ~ —--—; ——;-;+-: —-;- -~ ——;ORO By W fi - 1 2 I!L{W If -I —sur )re ssure,F" -ibe 1.: ^~ LAMrUt W L -Ha n- w ~./ jw-r ------ -w i I w ~4 R -- -VI I I flTr+ -.-tai..r A:-1 at. LT -— cs;I ~i rlr ~r~t- - - fsirt~rrrrur- ~r — -- f-i-: - - - ---—: - -~ i A --- #t-AINW —-#4UtflE hi ttt.. Lcn~::jlm"" tbt::_-t tt I tP2t FIW - - ------ - r ------—;-' —-~ —'i — —: —-— t — ~ —: —-rit — —--— ttCT-t —-—. —-L —.. ~ —— ii~._i-~~-:~~_C~_~ ~-_ ~~.~~::_~_-I~ ~~~ ~C ~ ~___~~~_?~_'4-~ iK~ 4't I 4-I - 1 —-I- ~- -- -— 4 -- 1 -- -: — -~ ---- — 1 - -I r I I lI -i I i 4- - 4..... 1i - 4e 4 -I z r o M C) z z 0 n3 cI, H C! O C3 -- — 44-~ 7 7 i —-- Tt |t t -1I —. - - -4 I 7- I I 1 r i di i I - I~~~. -_ _ - _ _J _, A.- - - - -..I t rOTAL PR4ESURE UPS}REAM OF -f - -- — 7_77'47 - 7 7 oz"1 4?( - FIG. 23 EFFECT OF NOZZLE DIAMETER ON TOTAL PRESSURE AT BLOWOFF.

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UM-74 flame was visibly affected most by certain frequencies, but at higher jet velocities, the flame could be extinguished well below its normal blowoff velocity by any frequency between 500 cycles/sec and 10,000 cycles/sec, provided the intensity was great enough. In an attempt to isolate this unsteady state phenomenon, a packing of glass wool retained by several layers of 100 mesh copper screen was placed in the tube immediately upstream of the convergent nozzle. This reduced the noise level of the flame (probably caused by resonance between the upstream tube and the flame itself) and probably tended to decrease the turbulence of the jet. The data obtained with this glass wool inserted upstream of the nozzle is presented in figure 19 which is a plot of nozzle diameter versus the mass velocity at blowoff for the four flame holders tested. The blowoff data shown in figure 23, together with the data presented in figure 18, is replotted in figure 24, which shows the increase in mass velocity caused by the insertion of the glass wool for the four flame holders tested. In figure 25 is shown the height of the completely laminar zone plotted against the mass velocity for the three jets tested. This height was obtained from the shadowgraphs shown in figures 18 and 19, while figure 26 shows the same data plotted as h/D versus Re. From figure 26 it may be seen that insertion of the glass wool does not have an appreciable effect on the height of the completely laminar zone, but it does have a great deal of effect on the mass velocity at blowoff. This would seem to indicate that the action of the glass wool in decreasing the noise level of the flame was more important than any effect the insertion of glass wool had on the turbulence of the jet. Since another investigator in this laboratory, Mr. R. A. Dunlap, was studying the effects of resonance on flames, no further investigation of noise level was performed by this group. Since the shaddwgraphs shown in figures 20, 21, and 22 indicated that the height of the completely laminar zone was a function of the nozzle diameter, as well as the mass velcoity, it became apparent that the height of the flame holder above the nozzle would be an important variable. All of the preceding experiments had been performed at a value of 1/2 of the ratio height of the flame holder above nozzle. Accordingly, nozzle diameter various experiments were performed using this flame holder height as one of the variables, while keeping the fuel-air ratio, ambient pressure, and flame holder diameter constant. In figure 27 may be seen a plot of mass velocity at blowoff versus flame holder height for the 3/8 inch diameter nozzle without the glass wool inserted upstream of the nozzle. A similar 33

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 I L 0 1 z I w N 4 N 0 Z ri -- -V i -- -- -- -- - /f__ _-_ 16 24 MASS VELOCITY AT 32 BLOWOFF 0 1/8" / 4 v, 4 0 4 8 16 32 LBS./ SEC./ SQ. FT.) D 0. 189" / O8~~~~~~~~/ 4 / / I (/) I 0 z 0r w IIJi N N 0 Z 0 8 (LBS./ SEC./ SQ. FT.) 16 24 32 MASS VELOCITY AT BLOWOFF --- NO GLASS WOOL UPSTREAM OF NOZZLE PROPANE AIR RATIO FLAMEHOLDER HEIGHT PRESSURE - 0.080 I1/2 Dj - I ATM. --- GLASS WOOL AND 100 MESH SCREENS INSERTED UPSTREAM OF NOZZLE COMPLETELY UNCONFINED JET FIG. 24 EFFECT OF INSERTION OF GLASS WOOL ON MASS VELOCITY AT BLOWOFF

WILLOW RUN RESEARCH CENTER —UNIVERSITY OF MICHIGAN UMM-74 -.5.4.3.2 L. I u z z LU z 0 N w z -J LL ILU 5: 0 U U0 IUJ I.1 0.5.4.3.2.1 0.5.4.375 in. jet..5 in. jet i iiii.625 in. jet 1,,.,.,. _ 111<~~~~~~im~~ mmmmm 1~mm.3.2.1 0 0 2 4 6 8 10 12 14 16 18 20 MASS VELOCITY (LBS. SECS. SQ. FT. OF NOZZLE AREA) 22 24 * With glass wool and 100 mesh screens upstream of nozzle. * Without glass wool and 100 mesh screens upstream of nozzle. FIG. 25 MASS VELOCITY VS. HEIGHT OF COMPLETELY LAMINAR ZONE. 355

1 —. -:?_1~:E r -::-.__t:^~-^:'~ ~HEIGHT OF CO T LA INAR _....^^^^ ^^ ^^^^^ ^^^^^ ^^ ^^^ FO'R VARI US DI METER JETS ^:...........:^:-.... / I~.^^^..... ^r^T^-'/:_.............:........'....' -. SSURE IATM.'~1: -::._;^::; —::' ^^^ .^;^1^^'*:'. —:.. pROPANE- ilR RAT-1O 0.080 (BY- EIGHT) 1/2 INCH D1AMETER NOZZI. 5/8 NCH DIAMETER NOZZLE _GLA WOOt a0 -MESHrgm 4-i^::;:^: ~^-:_:c.r ^ ^^^^^^ -7 ----— Y."' i SCREEN UPSTREAM OF NO ZLE ___^-:,/..,^^::.;:!^'';^^_ _____ ______ it_ — - _ __ _ —-.._ _. -......-. - - Ct4 ~ ~ - ON e r 0 c z t~1 m pi. -r)cn -ooo' —0 0 -. oo: — -: - -4090 L50 _:- ~ t-:-;;.U: ---. 1 ~ RE YWOLDSI NUMB 100 jj.'.':60 j0 1:810 1:1;R~~~~~ aoo -^:':.^ ^' \^ FIG. 26 EFFECT OF REYNOLDS NUMBER ON HEIGHT OF COMPLETELY LAMINAR ZONE FOR VARIOUS DIAMETER JETS.

NVr I UUONX R U RESEARCH CENTER-UNIVER S ITY OF MICHIGAN UMM-74 1/8 1 7/8 -c U._ -L u N N 0 z LU co O I LlU IZ 0 I U0 2: 5/8 1/2 NOZZLE DIAMETER -.375 FLAMEHOLDER DIA..062" PROPANE - AIR RATIO =.080 " PRESSURE = 1 ATM. NO GLASS WOOL UPSTREAM OF NOZZLE BURNING IN 10x12 x 18" TEST CHAMBER 8 -* I - - - - - - - a at - -- -------— rarSr —Izzz Izz)i zz 1/4'/8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) EFFECT OF FLAME HOLDER HEIGHT ON MASS VELOCITY AT BLOWOFF.375 IN.NOZZLE, NO GLASS WOOL, IN TEST CHAMBER. FIG. 27 37

L WILLOWR RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN....___________ UMM-74 picture is shown in figure 28, except that in these experiments the glass wool was inserted upstream of the nozzle. Shadowgraphs showing burning at indicated conditions with the 3/8 inch diameter nozzle are shown in figures 29, 30, and 31. The data shown in figures 27 and 28 may be compared in figure 32 which shows the effect of adding glass wool on the height of the flame holder versus mass velocity curve. The data shown in figures 27, 28, and 32 was obtained with the 3/8 inch jet inside of the 10 x 12 x 18 inch test chamber, as was the data obtained under vacuum. In order to determine the effect of this confinement of the jet when using the 3/8 inch nozzle, blowoff experiments were performed in and out of this test chamber, and, as may be seen in figure 27, there is no appreciable difference between the two cases when no glass wool is inserted -upstream of the nozzle, but there is some effect of confinement when the glass wool is inserted upstream of the nozzle, as may be seen in figure 34. The data in figure 33 seems to indicate that jet confinement effects were not present to any appreciable extent in the data obtained under vacuum. The vacuum data, as was previously stated, was obtained in the test chamber with a 3/8 inch diameter nozzle with no glass wool upstream of the nozzle. While the data presented in figure 27 was obtained at only one pressure, atmospheric, one fuel-air ratio, 0.080, and with one flame holder diameter, 1/16 inch, it might be assumed that the data obtained with other fuel-air ratios and pressures would not be altered appreciably if the 3/8 inch jet were completely unconfined. This statement, of course, applies only to the 3/8 inch diameter jet with no glass wool upstream of the nozzle. Experiments similar to those performed with the 3/8 inch nozzle were also performed with the 5/8 inch diameter nozzle. Figure 35 shows the mass velocity at blowoff as a function of flame holder height in an unconfined 5/8 inch diameter nozzle. Figure 36 is similar to figure 35, except that in these experiments, glass wool is inserted upstream of the nozzle. Figures 37, 38, and 39 are shadowgraphs of burning in the 5/8 inch diameter nozzle. Figure 40 allows the data presented in figures 35 and 36 to be compared. However, when the 5/8 inch diameter jet is confined in the 10 x 12 x 18 inch test chamber, a large effect of jet confinement is observed. When no glass wool is inserted upstream of the 5/8 inch nozzle, it is impossible to ignite the fire in the test chamber at any flame holder height. L 38

WILLOW RUN RESEARCH CENTER- UNIVERSITY OF MICHIGAN UMM-74 I 1 11/ 1 7/8 3/4 U) u u N ee I14 N0 LUJ 0 I -LU 0 I i. 0D wj 1/2 NOZZLE DIAMETER =.375 FLAMEHOLDER DIA. =.062 PROPANE - AIR RATIO=.080 PRESSURE = 1 ATM. GLASS WOOL UPSTREAM OF NOZZLE BURNING IN 10x 12 x 18 TEST CHAMBER. - - - - - - - - - - - 1 1-, X X X ++Sit~~~~~~. X X XL 4nAL~~~~~~~e 1 LL LL LL1 ALL~~~~~~~~~,LA~I -i 1/4'/8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) FIG. 28 EFFECT OF FLAME HOLDER HEIGHT ON MASS VELOCITY AT BLOWOFF 375 IN NOZZLE, WITH GLASS WOOL, IN TEST CHAMBER. 39

Height of flameholder above nozzle (inches) 0.062.125.187.250 0 E 0 oE a U C E,.. 0 Oc o 0.312 e r r 0 C) Z I z'm H 0 C) -i.c.. ri r: tj af cl CC O Cr r,- SHADOWGRAPHS SHOWING EFFECT OF HEIGHT OF FLAME HOLDER ABOVE NOZZLE ON COMBUSTION FOR A MASS VELOCITY RANGE OF 3.1 TO 3.2 LB./SEC./SQ. FT. OF NOZZLE AREA REYNOLDS NUMBER- 8,500 Spherical flameholder dia. 1/16 in. Nozzle diameter 3/8 in. Propane-air mixture 0.080 (by weight) Pressure 1 atmosphere FIG. 29 SHADOWGRAPHS SHOWING EFFECT OF HEIGHT OF FLAMEHOLDER ABOVE.3751 NOZZLE ON COMBUSTION FOR A MASS VELOCITY RANGE OF 3.1 - 3.2.

Height of flameholder above nozzle (inches) 0.062.125.187.250.312 E ~ C o O 0 J 0 CO,.0 4) E * H o O C o~ -e 8, Cs o. Flame will not burn under these conditions 0.062.125.187.250.312 0 r 0 So c z 0 m I?z 0 i-I t = H 0 zf h^4 SHADOWGRAPHS SHOWING EFFECT OF HEIGHT OF FLAME HOLDER ABOVE NOZZLE ON COMBUSTION FOR A MASS VELOCITY RANGE OF 5.7 TO 6.0 LB./SEC./SQ. FT. OF NOZZLE AREA REYNOLDS NUMBERS 15,000 Spherical flameholder dia. 1/16 in. Nozzle diameter 3/8 in. Propane-air mixture 0.080 (by weight) Pressure 1 atmosphere FIG. 30 SHADOWGRAPHS SHOWING EFFECT OF HEIGHT OF FLAMEHOLDER ABOVE.375" NOZZLE ON COMBUSTION FOR A MASS VELOCITY RANGE OF 5.7 - 6.0

NILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN UMM-74 (N C1 (, CII) L 0 tN E 0___.'' ~~~~~n 5 O 2 -Cf 0 t Q I, C C: 2 0 E cip O I._ c; * X~~L o %n ("......................................................... LU oz O - U O: O 2 LL O oO Vl o 0 OOLu ICO - - 0~ >~ *' 0 tn > u _ LL L L 0 Z c.. 0. o. U0 ~ ) N Z 0 > "o O 0 N3 coO to 0( C4 P7 (N 01 Co a C('o O.C u c.D N N c.0 -t 0 4I.6 X, o o im elzzou 0o woeUDsdn suaeejs 4seW OOL puo iOOM so16 4noll4!M alzzou jo woesDadn sueaJns lsauW 00 1 pUD lOOM SSolB6 q!M

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 1/ 8 1 7/8 U, 0 u rc LU I N 0 LU -J 0 so I I 0 LUJ 3/ NOZZLE DIAMETER =.375 FLAMEHOLDER DIA. =.062 PROPANE - AIR RATIO =.080 PRESSURE = 1 ATM. BURNING IN 10x 12x18" TEST CHAMBER. X GLASS WOOL AND 100 MESH SCREEN UPSTREAM OF NOZZLE A NO GLASS WOOL OR 100 MESH SCREEN -_ UPSTREAM OF NOZZLE - -- - - _- - -- -- -- -- -x — J ~ -— x\ 3/8 1/4 /8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) FIG. 32 EFFECT OF INSERTION OF GLASS WOOL ON FLAME HOLDER HEIGHT VS. MASS VELOCITY PLOT -.375 IN. NOZZLE.

WILLOW RUN RESEARCH CENTER- UNIVERSITY OF MICHIGAN 1UM-74 1/8 1 7/8 0 -c U r LU -J N 0 z LU 0 U..'r' I 0 -r (D ui 3/4 5/8 1/2 3/8 1/4 NOZZLE DIAMETER:.375 FLAMEHOLDER DIA. =.062 PROPANE - AIR RATIO =.080 PRESSURE = 1 ATM. A BURNING IN 10 x 12 x 18" TEST CHAMBER. I I I I I I i _ * BURNING IN COMPLETELY UNCONFINED JET NO GLASS WOOL UPSTREAM OF NOZZLE — I — I-I-I-I-I-I-I- >. 1 fc-I —I —I — 1/8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs/sec/sq. ft.) FIG. 33 EFFECT OF JET CONFINEMENT NOZZLE, - ON MASS VELOCITY AT BLOWOFF.375 IN. NO GLASS WOOL. 44

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN ______________ MM 7-74 1 1/8 1 7/8 U U__ -a ui N N O z IJ LJ O LU I O II LU ULA UJ 3 3/4 5/8'/2 3/8 1/4 NOZZLE DIAMETER =.375 FLAMEHOLDER DIA. =.062 PROPANE - AIR RATIO =.080 PRESSURE 1 ATM. GLASS WOOL AND 100 MESH SCREEN UPSTREAM OF NOZZLE I I I I O BURNING IN COMPLETELY UNCONFINED JET.. - - I I I l I I I - A BURNING IN 10x 12x 18" TEST CHAMBER. ----------------— A^ —-^^^ —------ 1/8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) EFFECT OF JET CONFINEMENT ON MASS VELOCITY AT BLOWOFF.375 IN. NOZZLE, WITH GLASS WOOL. FIG. 34

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 1/8 7/8 I I I I I I I I I I I I I' -c N N 0 z U 0 co Ul LU -i 0 m I UJ LL 0 ILUJ I i-I1 -i NOZZLE DIAMETER =.625 FLAMEHOLDER DIA. =.062 PROPANE-AIR RATIO=.080 PRESSURE = 1 ATM. NO GLASS WOOL UPSTREAM OF NOZZLE BURNING IN COMPLETELY UNCONFINED JET. I I I I I I I 3/4 1/2 3/8'/4 NOTE: IT WAS IMPOSSIBLE TO IGNITE THE.62" --- -- - NOZZLE (Without glass wool upstream) IN THETEST CHAMBER AT ANY FLAMEHOLDER HEIGHT. ^^^^~~~~~ — ^ —Tr cr rrr T- --— I-L 1/8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) EFFECT OF FLAME HOLDER HEIGHT ON MASS VELOCITY AT BLOWOFF.625 IN.NOZZLE, NO GLASS WOOL, UNCONFINED JET. FIG. 35

WILLOVW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 1/8 1 7/8 I -C v c LU N N O z O 0 etY UJ co 0 _J I Lu I U. I 3/4 5/8 V2 3/8 /4 NOZZLE DIAMETER =.625" FLAMEHOLDER DIA. =.062" PROPANE - AIR RATIO=.080" PRESSURE = 1 ATM. GLASS WOOL AND 100 MESH SCREEN UPSTREAM OF NOZZLE H I BURNING IN COMPLETELY UNCONFINED JET. J I........ T T T T~~~~~rl4CL -','/8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) FIG. 36 EFFECT OF FLAME HOLDER HEIGHT ON MASS VELOCITY.625 IN NOZZLE, GLASS WOOL, UNCONFINED AT BLOWOFF JET.

WILLOW RUN RESEARCH CENTER - UNIVERSITY OF - MICHIGAN UMM-74 CN t) I U LU N C4 0 z 0 ID 0 LLJ LL I 0 I CO) 00 C-i Ce) N C i) N 0 Z 0 LL 0 t/) ILU C ) z aoou, O O Cl) C') 00 co C)z < 00 O IU LU o I0 < L c nZ o O Ua I o V) Z Ito C4'O'O W0 < >, 00 o 0<c6 Ci 0 _JO o LU LU 0 w 0 I__ U- Z UV U. 0 0 I z CE:::E o 0 in) C') oo C) IN dV d 48

IW0ILLOWV RUN RESEARCH CENTER,-UNIVERSITY OF MICHIGAN UM-74............iiii;;l'''',,- - t\ ~~~~~~~~~~~~~~~~~..........'-',:..,....~'4' ~ ~ ~..... ~~~~~~~~~~~~~~~~.....~, ".........'W ~ a..........~~~~~~~~~~~,,. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.':a ii::..i:i iiifiiiiiiiiai'~ifiiiififiiiifiriliiiiii iiii,~C'i'~.:~2~':,: ~i iiiiiiiiii i~ia ~!~!i'~;:~a:!ii:';il~~~~~~iiiiil~~~~~~~~~:ii;'i~~~~~~~~~i~~~~ii~~~~ii!;.':::!.:i~ ~ ~ ~ ~ ~ ~ ~ ~~ ~~~iiiiiaiiiiiiiiiiifi iii[giii;i'~;iia':i!i!iiiiliiiiiiiiiiiill,~iii'iiiii ~ ii'/~i~i' aii:: jiiiiiiiji[~ i~ii'~i;................... ~ ~ ~ ~ ~ ~........ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.......;i;'""iiiiiil~~~~~~~~~~~~~~~~~~~~~~~~~iiiiiiii~~~~~~~~~~~~~~~iii~~~~~~~iiiiii'iiii11~:!:iiiii,:: "'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... ~~~~:~::~::::;::;iiiiiiiiii;;;:j~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:i:!iiiii:i~:,:~:ii~ii O~~ 0 Niiii;-i:;~i iiil:.;''ii'ii:i'i.i::ii~:: C* --:iiiii;::iiiiiiii 00 O~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i'ii!;i~':iiiii[:~iii:!f: 0:iiiiiii~ii;...................;:; O q.. ~~~L I ~ LL,..-:. 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........ -t - ~ t O ~~~ ~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........ - -........................l~i~'~'i~::';s l Ij ZLU LU -, =:'::::: ~..........................................;;i I —-..............~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~................... M ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.1........ W o......... L 13 Z~~~~~~~~V liliirl~i~il~il'ill~lii~iiiiiiiii':l i....'i;;:i;:~:;ii~iiiiiiil............ U co......................................................... i::ii~~~iiiii-iiiili~~~i::iiiiliiiii~~~ii l~~ii:::dlL-:i"~~~~!"!"?!l'~ ~Liiiii;;l. ii:~~:;:fi~~~~~~i~:~~ii~~;j/ji:1;~ ~~l~ii;:.:':"":':':::::::':::":::'~.::::'::'::::::.:::'::' u,~~~~~~~o u ii;:::iiiiiiiii~:'":;;i~~~~~~~~~~~~~i:;::;i~~~~~~~iii"";:i:.!il':'li:~~~~~~~~~~~"':jciiii;;l::iii!:i~:~~~~~~~~~~~~"] I V O~0....................................LU LU..........................iiii jiii~i; i~~jj:i): ~ij'i ~~!'~:'-"i..................................................... """""""""""""""""";'" i-liij::~:ij!-':l'''iij;:li-...................................................... -- -.:.::::!:::!!::::::::;:::;;;:;;:::::i:::::: 1 11. poin t................. E x 0" ~ iiiii m O fC3 O~ h:;iiii;~~~~~~~~~~~~~~~~~~~~~~~................... i.........."C LO...........~i:i~iii;iiiiiiiiiii.iii~ i iiiiiiiiiiii:ii i iiiiiiii V'i:':ijililliii:~~~~iiiiiiiiliijji~~liiii; liiji:-;i~~~ii~ililiiiii'iii~~~ij.:...... I.,.............I U J W....................................... i i i i: V ):::iiiiiii;:i::.;i~i;:.......... iLi:; ii; Z Wi LU:::- ii~i~iiiiiiisiiiiiiiiiiiii D............................. W',liii~iiiiisiiliii I -~ilGi!; iiiTiliiiii l < Diii i~iiiii;'';l~~~~~~~~~~i~~~iL ii::iiijji~~~~~~~~~~~~~~~~~~~~~l'l:11'.:i:.i~~~~~~~~~~~~~~~~~~~~l:;;iiiii~~~~~~~~~~~~~~~iiii LLI~~~~~~~L V i;:;~:~;:iiiiii~i:~.~:::ii::i:::;i;:; ii ii —iii~ ~ i~i;::-:::::~:~:-:i: ~::i;:::; i-i;:i;: iii:; i~i iiiiiiiiii iiiiiiii iii~i;;:;i-::::-:........... i-f~~~i- -i-:i......................... i::-'~~i'i~;l.'li'''~~'ii'':""'iiiiii~i i':':::;';i;::'::::::'i;:;.................................. Z~; 4ai~ ~ R' UI! iiiiiij~iiliis -11..............................ii~j...... L sel~~~t~~i' iiiiilis i~~~~~ii;::iil~~~~iii:":""""i':""ii ~.........!::~ ~ -,::::!,;~~~;::.:.::p:;:::::;:::::;:7:::;:;L;:;: LLI:::::i;.iiii:;i~~iii:;i................. Lu...............................:'i-::;- i l:;:: il:1: i iiiiiiiiiii;:''-i''iiiiiil uJ:::~~~~~~~~~~~~~~............................... i'iiiiiii:; ii~~~~~~~~.......................... Abet!! I op P~:;;,,; i iii..::;:,;ii;ii: ~~~-~-~;;;: ~ ~~;i;;~;::: i;;ii ~ ~ ~ ~ ~:::L;!:-;;::;::::: 1::;:.;; 7::;:.................................. "~~~~::i'iii""""""'""""""i;:':.:."":;. i";':"...................................';'''';l-;;:lli:!;!'i~~~~~~~~~ii'........................... L ~ i~iii""""ii":'"""'i""""i'~::::i~i~lii/jjjj~liilii~i!:iililii, i;.ili~l;:iii:'':,e0 i P iiiii;;:i;;;;;;i;;i;;;;;;;;;;;;;;;;;;;;i:: e~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... oL h Y~~~~~~........ V! c3~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... ~~~~~L cn~~~~~~~~~~~~~~~................................ o;iiii~i:'''': ii~ir! iiii iiri~ii;;i':""~"i'iiiiiii..................................: iiLui~i~i 0 iii~ iii;:ii: i~~lii: ~~ilijililiiiil~iiiii:i:lliiiliiiliii-i.......... c oii:' i"""': w:ii,, iii;.i~~~~~~~liiiiiiiiiii.:i:::::::i:::::;.:::::::::::;:.....:.:::~~~~~~~~................................~ ~ ~ ~ ~~~~~~~i.:::.-~iiii~~iii l~i i~iii~i'iii;;iiiil~ijl~c"~i~: ~::;...........................iiiiiiiiiiiiiiie iiiii co < ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~;1:~;t:"' -~~:iiii:ii i:~'ii............................: i i~i~j:ii:::~jiii:; i~-l~i;:l; i........... <Yiiii'';:ii;:i:i~~:ji l;::i.........................................................:.:::...............;i!'~''':'." l~' i' """'"'';""';''""";"';';'""':'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1.......... iiiiiii~i iii~~...........: i:;:;:;:;: ii............. ii i i~ ~ii i';;ilii;:~;.~iii:;ii~i0 W............................................................... riii0.ili iiil;iiii:'iiiiiii::i::iiliiiiiii ~~i;'ii;:i:: iiiii:i::;; -~::~-:: ~ ~ ~ ~ ~ ~ ~ ~ ~ iiiiij iiiiiiiiiiiiiiiiiiliiii;iiiiiiiiiiii~iiiiiiiiiliiiiECA.........................::;:;:::;;:;:::::::!!::: I:........... iii i; i ii i ii i;:i1i iei: iiii. ii:;:ii:;::::::ii1iiii......................,iiii' iii ji~i::~iii;''iiii~iiiii:::i-i~i;iiiiiiiiiiii~iiiii~iii:;:iiii::iii ~iii~................ e;ji::.:..............................:li''::':'ii~:'~iii iiili I;':;n'"":.................................................................................................................;:;;;:;;;;; 49

WILLO UUN 1 F: S..AR': CI -I -(3VNT ER! UNLItVEIlZ; S ITY U~H —:'lh-.t OF MI X ICIIGAN N fi, 0,, 0 9 t 1%1 N 7.; I.:C+: 6t o O tt, Lt. 0... 0 N Q) V) N tt! N 0,.4 "t. LV).. M.* LX; t) Lii 0 o tlr O 0 i ) to. tre 1. O Wl L.... 0..A.7t,",: s N 9 ql 0 to FIG. 39 SHADOWGRAPHS SHOWING EFFECT OF HEIGHT OF FLAMEHOLDER ABOVE.625" NOZZLE ON COMBUSTION FOR A MASS VELOCITY RANGE OF 8.75 - 9.17 WITH GLASS WOOL UPSTREAM OF NOZZLE. Qo

WILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN I —--------- UMM-714 1 — l I 0 u _i LU -I N N 0 z LU 0 *u co al LU -J 0 I U-I 0 ILU I 1/ 1 7/8 3/4 5/8'/2 3/8 /4 i 1/8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs./sec/q. ft.) FIG. 40 EFFECT OF INSERTION OF GLASS WOOL UPSTREAM OF.675 IN. NOZZLE ON MASS VELOCITY AT BLOWOFF. 51

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 Blowoff velocities for these same conditions, except that experiments were conducted outside of the test chamber, were shown in figure 35. When glass wool is inserted upstream of the nozzle, the effect of jet confinement may easily be seen from figure 41. The effects of jet confinement may be summarized as follows: 1. When a ratio of test chamber area to nozzle area of 1100 to 1 was used, there was no appreciable difference between that ratio and an infinite ratio (i.e. an unconfined jet) when no glass wool was inserted upstream of the nozzle, and very little difference when glass wool was inserted upstream of the nozzle. 2. When a ratio of test chamber area to nozzle area of 400 to 1 was used, it was impossible to ignite the flame when no glass wool was inserted upstream of the nozzle, and when glass wool was inserted upstream of the nozzle, the mass velocity at blowoff was ca. 1/2 the value obtained in an unconfined jet. 3. When a ratio of test chamber area to nozzle area of 36 to 1 was used (in System 1) severe explosions (at a flame holder height of 1/2 inch) resulted as the flame flashed back through the nozzle. Five such explosions occurred in System 1 before System 2 was used. The effect of nozzle diameter may be seen in figures 32, 34, 40 and 41 which are plots of the mass velocity at blowoff versus height of the flame holder for the 3/8 inch and the 5/8 inch diameter nozzles under various conditions. The experimental data presented in this section has indicated that the following variables have an appreciable effect on the mass velocity at blowoff: fuel-air ratio, sphere diameter, pressure, nozzle diameter, height of the flame holder above the nozzle, the degree of jet confinement, and the noise level of the flame as affected by the insertion of the glass wool upstream of the nozzle. The last four variables are all interrelated and are not easily separable. For example, the nozzle diameter and the height of the flame holder are both important in determining how much effect confinement of the jet will have on the mass velocity at blowoff. The insertion of glass wool was believed to decrease the resonance between the flame itself and the upstream tube, but no attempt was made 52

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN I —-1 —------ UMM-74 - I 11/a 1.. UJ uJ - N N 0 z > 0 UJ co O -ri Uur I ui 3/4 NOZZLE DIAMETER =.625 _ _ FLAMEHOLDER DIA. =.062" BURNING IN COMPLETELY PROPANE - AIR RATIO=.080" UNCONFINED JET. PRESSURE 1 ATM. GLASS WOOL UPSTREAM OF NOZZLE BURNING IN 10 x12 x18"TEST CHAMBER.__ -.A4 _,, - - - 1/2 3/8 /4 /8 0 0 2 4 6 8 10 12 14 MASS VELOCITY AT BLOWOFF (Ibs./sec/sq. ft.) EFFECT OF JET CONFINEMENT ON MASS VELOCITY AT BLOWOFF.625 IN. NOZZLE, FIG. 41 553

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN.............. UMM-74 _ to determine quantitatively this decrease in resonance. in the mass velocity at blowoff was measured. Only the change It is not possible to eliminate these system effects in ramjet combustion chambers, since there must be some degree of confinement of the jet. It is believed, therefore, that these data indicate that the system variables have an effect on the mass velocity at blowoff comparable to the effect of fuel-air ratio, pressure, or flame holder diameter.

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 IV. MISCELLANIEOUS EXPERIMENTS Several experiments were performed during this investigation which were not connected directly with blowoff velocities of spherical flame holders, but were connected with the more general problem of flame stabilization. They are presented here with the hope of offering other investigators new methods of attack on the general problem of flame stabilization. A. Immersed Flame Holders This work was performed in November, 1948, and previously reported in a progress report.3 The appearance of a pilot flame at high jet velocities has previously been stated. From aerodynamic considerations, it seemed that a bluff body located downstream of the spherical flame holder would provide a low velocity region and might again stabilize the flame and allow it to propagate. In figure 42 are several photographs showing the effect of insertion of a cylindrical rod in the wake. All of the photographs in figure 42 are at exactly the same conditions; namely, atmospheric pressure, 0.125 inch diameter spherical flame holder, 1 inch diameter jet, flame holder height of 1/2 inch, fuel-air ratio of 0.05 and a jet velocity of 268 ft/sec. The effect of insertion of a cylindrical rod at the critical position into the pilot flame shown in Photograph 8 may be seen in Photograph 9. The appearance of the flame when the cylinder is below the critical position is shown in Photograph 10. Photographs 11 and 12 show the appearance of the flame as the cylinder is progressively moved downstream from the spherical flame holder. Photograph 13 is a photograph of a smaller diameter cylinder inserted at the critical position in the pilot flame shown in Photograph 8, while Photograph 14 shows the appearance of the flame with the same cylinder inserted further downstream. It should be emphasized that there is a critical position for the height of the rod above the flame holder, for example the 3/8 inch diameter rod in Photograph 10 is only 3/8 inch below the optimum position shown in Photograph 9, while in Photograph 11, the 3/8 inch diameter rod is only 5/16 inch above the optimum position. It also appears that there is a variation in the critical distance above the flame holder for a variation in cylinder size, as is shown in Photographs 10 and 13, the critical distance being 1-1/2 + 1/16 inch for the 3/8 inch diameter rod and 1-1/4 + 1/16 inch for the 1/8 inch diameter rod. 55

VWILI,,W R U N RE, S EEARCH C ENTER - UNIVEI S., ITY OF MIC HIG AN UMM-t 3/8" R OD - CRITICAL DISTANCE i 1/2"il/6"w ABOVE FLAMEHOLDER ROD 1/4 BELOW CRITICAL DISTANCE ROD 1/4 ABOVE CRITICAL DISTANCE FIG. 42 A EFFECT OF INSERTION OF A CYLINDRICAL ROD INTO A PILOT FLAME. 56

WVILLLOVW QRIUN RESEARZCHI C ENTER - UNIVEI4$:RSITY OF MICHIIGAN tum4- 7t4 1/8" ROD --- CRITICAL DISTANCE - 114" 1 tl/l' ABOVE FLAMEHOLDER - ALL PHOTOGRAPHS ARE AT SAME CONDITIONS JET VELOCITY 268 fr./sec. ON MASS VELOCITY 20 Ib./sec./sq. ft. of Nozzle Area FLAME)HOLDER IDIA 1/8" NOZZLE 01IA 1 lHEIGHT OF FLAMEHOLDER 1 /2' PRESSURE I aim. |COMPLETELY UNCONFINED JET FIG 42 B EFFECT OFINSERTIONOFA CYLINDRICALRODINTOAPILOTFL FIG. 42 B E;FFECT OF INSERTION: OF A CYLINDRItCAL ROD INTO A PILOT FLAME. 57

WILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN UMM-74 A few experiments were performed using a sphere instead of a cylindrical rod. The second sphere seemed to act just as an additional flame holder, that is burning occurred directly from the surface of the sphere, and its position was not critical, while the action of the rod. seemed to be a more involved aerodynamic process. Fastex shadowgraphs and titanium tetrachloride smoke tracers did not shed appreciable light on the phenomenon. B. Boundary Layer Removal Experiments A flame holder involving surface section was constructed as is shown in figure 43. The vacuum was applied with a Nash Hytor MP 673 vacuum pump. A flow tube in the vacuum line was calibrated so that the amount of mixture removed could be measured. At full vacuum the flow through the bleed holes was ca. 2.5 cu.ft. per mm of 0.073 lbs per cu.ft. air. This value was substantiated by both the calibrated flow tube and the pump specification curve. In figure 44 in the right photograph, a flame is shown burning from the cylindrical holder with no vacuum applied at a fuel-air ratio of 0.0865 and a jet velocity of 137 ft/sec. In the left photograph, the flame at the same conditions is seen after full vacuum was applied. The flow through the ten No. 60 (0.020) dull holes is ca. 2.5 cfm. The total flow through the jet is 42 cfm. Calculations based on the Blasius solution for the order of magnitude of the thickness of the boundary layer indicates a boundary layer flow between 0.5 and 2.5 cfm. The use of titanium tetrachloride smoke tracers showed little or no disturbance of the flow pattern with vacuum on. Therefore, it is believed that most of the boundary layer is being removed with no appreciable alteration of the flow pattern about the holder. Figure 45 is a comparative plot of the stability limits on the rich side for this type of flame holder with and without boundary layer removal. The curve for the flame holder shown in figure 45 without vacuum concurs exactly with the curve obtained for the four spheres previously tested in System 1. As may be seen in figure 45, with vacuum applied there is an appreciable decrease in the stability limits. A further extension of this experiment is shown schematically in figure 46. A ring of 0.026 inch diameter wire was used to artifically create separation at various points along the flame holder. Three distinct regions were found, as shown in figure 46. With the ring ~I located at any point in region 1, ignition was secured easily and the I - 58

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 NO. 60 (.040) DRILl HOLES 5 HOLES PER ROW, EQUALLY SPACED, \_I I A 1" J 1/4* DIA. COPPER TU&ING. / NOZZLE & DRUM COVER ASSEMBLY.'41M I 1' DIA. I I... TO VACUUM PUMP.................................... -.......... I DRUM WALL / DETAIL OF BLEED HOLES - TWICE SIZE FIG. 43 SURFACE SUCTION FLAME HOLDER. 59

WILLOW RUN RESEARCH CENTER- UNIVERSITY OF MICHIGAN UMM-74 With boundary layer removal Without boundary layer removal PHOTOGRAPHS SHOWING EFFECTS OF BOUNDARY LAYER REMOVAL ON FLAME Jet velocity 137 ft. /sec. Mass velocity =10.4 Ibs./secs. /sq. ft. of nozzle area. Nozzle dia. - 1 in. Cylindrical flameholder dia.-.25 in. Propane - air ratio = 0.0865 Pressure 1 atmosphere. FIG. 44 PHOTOGRAPHS SHOWING EFFECT OF BOUNDARY LAYER REMOVAL ON FLAME. 60

I I curve for all spheres.~/ 2.0 1.8 OC H a Z 1.6. *i 1.4 z 1.0 a. ~' 0.8 0 U 0.6 LU > uI ui < 0.4 el-\ I yr2y r0, I T10 0 lFlameholder with no vacuum. A Flameholder with full vacuum.,//ff r 0 z Z t H 0 C) =,4 I 0 t ^P 0.2 0 0 2 4 6 8 10 12 14 16 18 20 22 MASS VELOCITY OF AIR (Ibs. sec. sq. ft. of jet area.) FIG. 45 BLOWOFF VELOCITY OF FLAMEHOLDER SHOWN IN FIG. 43 24

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 1/8' f1/16 I - REGION 1 A J REGION 2 REGION 3 1/4" DIA. I! J OF SEPARATION RING REGION 1 IGNITION & CONTINUOUS BURNING POSSIBLE, FLAME BURNS FROM WIRE. REGION 2 IGNITION POSSIBLE, BUT STEADY BURNING IMPOSSIBLE WITHOUT USE OF SURFACE SUCTION. REGION 3 IGNITION & STEADY BURNING POSSIBLE WITHOUT SURFACE SUCTION, FLAME BURNS FROM HOLDER. SEPARATION EXPERIMENT FIG. 46 62

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 flame burned stably. The zone was somewhat larger due to the fact that the flow separated at the ring. Region 1 extended a distance of 0.125 inches measured from the end of the flame holder. At the extreme point of region 1 it became impossible to ignite the mixture using the initial combination of jet velocity equal to 137 ft/sec and fuel-air ratio of 0.0826, or any other combination of fuel-air ratio and jet velocity. Region 2 represents the length in which it was impossible to obtain burning off the flame holder. At the end of region 2, a distance of 11/16 inch from the downstream end of the holder, it again became possible to obtain ignition and steady burning. The combustion, however, appeared to be unsteady and very incomplete. When surface suction is applied with the ring located anywhere in region 1, the flame is altered in much the same manner as is shown in the right hand photograph in figure 44. The downstream combustion ceases and only a pilot flame remains. With the ring located anywhere in region 2, ignition and steady burning is impossible without the application of vacuum. In other words, surface suction allows the stabilization of a flame behind the holder where it was not previously possible. When the ring is located anywhere in region 3, the effect of vacuum is again the same as shown in figure 35. It is to be noted that the action of surface suction becomes more effective in restabilizing the flame as the location of the ring approaches the point dividing regions 1 and 2. The above mentioned experimental work was previously reported4 in a progress report in July, 1949. Investigators at Battelle Memorial Institute attempted to duplicate the results shown in figure 44 but did not succeed, and their conclusion was "that the boundary layer has little or no effect on the blowoff limits and consequently, profile drag can be considered to have little or no effect on flame holding."5 However, their apparatus6 differed in several respects from the apparatus shown in figure 43. Although the nozzle diameter in both systems was 1 inch, the distance of the straight section following the convergent section of the nozzle was 3/4 inch in the apparatus shown in figure 43 and 2 inches in the equipment used at Battelle. Also, the latter system contained four relatively large supports at the inlet of the nozzle. These differences might tend to make the jet in the latter system more turbulent than the jet of the apparatus shown in figure 45. I 63

WILLOWR RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 The notch in the cylindrical flame holder used by Battelle might have had a separation effect similar to the effect of the wire ring shown in figure 46. The experiments performed at Battelle Memorial Institute were designed to change the length of boundary layer build-up and not to completely remove the boundary layer immediately upstream of the flame, as in our experiments. Hence, their "full vacuum" and "no vacuum" designations did not have the same meaning as in the experiments previously described, i.e., in our experiments, full vacuum meant complete removal of the boundary layer immediately upstream of the flame, while in their experiments, full vacuum meant that the distance for boundary layer build-up was decreased by a factor of 3 or of 8, but not removed immediately before the flame. It was found necessary in our experiments to remove ca. 2 to 5/ of the total flow through the nozzle in order to obtain the results pictured in figure 44. In our 1 inch jet, this required almost the full capacity of a reasonably large vacuum pump (Nash Hytor MD 673), the use of a standard water aspirator not being sufficient. The amount of material removed in Battelle's experiments was not stated explicitly. The data presented in section 3 indicates how sensitive flame holding is to system variables, such as the height of the flame holder, etc. The reason for the disagreement of the two experiments is probably due to some combination of the above mentioned possibilities. Since the system and the technique employed at Battelle was not identical to that used here, it is believed that their failure to reproduce the experiment,shown in figure 44, is meaningless. Since the disagreement was reported by Battelle, the experiments have been repeated in this laboratory many times, not only with the cylindrical flame holder shown, but with conical and disc flame holders as well. A rotameter installed in the vacuum line indicated that reproducible results could be obtained over a wide range of fuel-air ratios and jet velocities, and that decidedly lower blowoff velocities were obtained when the boundary layer was removed.

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 C. Flame Traverses An attempt was made to obtain quantitative measurements in the region downstream of the flame holder during burning as well as in cold flow. It was believed that analytical work would be meaningless because of the length of time required to perform analysis by conventional methods, as well as the time consumed by sampling. It was also felt that a thermocouple traverse would not be very accurate because of the lack of knowledge concerning emissivity coefficients, etc., at the high temperatures involved. However, it was believed that physical measurements of pressure and electrical conductivity could be made in the flame region at various locations. For this purpose, an apparatus was constructed which permitted vertical and horizontal traverses to be made accurately. The flame holder used in the experiments was constructed from a 1/4 inch copper tube, plugged at one end and connected to a water manometer at the other. A hole, 0.040 inches in diameter, was drilled through the brass plug. The flame holder was installed axially through the nozzle (in a manner similar to the flame holder shown in figure 43), and thus a vertical cylindrical flame holder was obtained which permitted measurement of the pressure at the base of the flame. For the traverse measurements of pressure a piece of steel hypodermic tubing 0.042 inches in diameter, connected to a water manometer, was used. Measurements were made with burning and without burning using a fuel-air ratio of 0.0942 and a jet velocity of 109 ft/sec, the one inch diameter nozzle in System 1, and a completely unconfined jet at atmospheric pressure. The end of the 1/4 inch cylindrical flame holder was ca. 1-3/8 inches above the nozzle. A photograph of the flame burning under these conditions may be seen in figure 47. The pressure traverse tube orientated vertically may be seen in the right of the photograph. Horizontal traverses were made at different heights above the flame holder to obtain pressure measurements for each of the following cases: vertical tubing with burning; vertical tubing without burning; horizontal tubing with burning; and horizontal tubing without burning. The experimental data obtained with the vertical tube traverses is shown in figure 48, while the data obtained with the horizontal tube traverses are shown in figure 49; a cross plot of this data showing the pressure measurements versus height along the flame center line is shown in figure 50. Data obtained from figures 48 and 49 was used to obtain pressure contours, and these plots, superimposed on a photograph I 65

VWILLOW(X RIUN RESE-ARCH CJENT:ER' - UNIVEI.I'VSITY OF MICHIGAN U1M41-74(`~~c~~$Gp%[E~~~,$~~B~8*ti; ~~~tst> 7).. t~~~l~X~~Cr`I~~- lE~'7~`3~`t $T1s ~F~6C~~~f~rl ~~~~~~~~~~~~~~~~~~~UIv'i i v FiG. 47 PHOTOGRAPH OF FLAME AT CONDITIONS USED FOR PRESSURE TRAVERSES. 66

WILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN TUM-74 L H ~ 1.75 5 4 3 2 1 0 CM. FROM FLAME CENTER H - 3.025 6 4 —- 2- 1 - I4 3 2 1 CM. FROM FLAME CENTER H * 6.85 I 4 2 X X 5 4 3 2 5 4 3 2 1 1 CM. FROM FLAME CENTER CM. FROM FLAME CENTER H - 9.40 H - 10.675 4 to 2 2 *^' 0 -I I _A ~6~/ 4 2 0 o' VA-1 r ji L__ I___I _I ____ III ____ 4 3 2 CM. FROM FLAME CENTER 4 3 2 1 CM. FROM FLAME CENTER 4 3 2 1 CM. FROM FLAME CENTER x ~ BURNING 0. NO BURNING H - HEIGHT ABOVE FLAMEHOLDER BASE, CM. TRAVERSE TAKEN AT EVERY 1/8' (.32 CM.) OF HEIGHT:SELECTED POINTS SHOWN HERE. MASS VELOCITY - 8.3 LBS./SEC./SQ.FT. PROPANE AIR RATIO - 0.0942 PRESSURE - I ATM. NOZZLE DIAMETER I INCH CYLINDRICAL FLAME HOLDER DIA. - 1/4' FIG. 48 VERTICAL TUBE PRESSURE TRAVERSES OF FLAME. 67

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 H -.475 4 x f 11 A ---, 7 n.0~~ ==S / 2 0 H - 1.75 o 5 4 3 2 I CM. FROM FLAME CENTER I H * 3.025 4 0 2 0 5 A4 3 2 1 0 CM. FROM FLAME CENTER I 4 3 2 0 2 0 1 I::) 0 CM. FROM FLAME CENTER 6 H - 4.3 4 2 02... 5 4 3 2 I CM. FROM FLAME CENTER H * 5.575 4 I 6 H * 8.125 5 4 3 2 1 0 CM. FROM FLAME CENTER H 9.40 1* 1 1 1 1 H * 6.85 4 - 2 5 4 3 2 I C CM. FROM FLAME CENTER H * 10.675 2,0Oj 0 -------------------- 4 2 0 0 4 3 2 I 0 5 4 3 2 1 0 5 4 3 2 CM. FROM FLAME CENTER CM. FROM FLAME CENTER CM. FROM FLAME C x - BURNING MASS VELOCITY -8.3 LBS./SEC./SQ.FT. o - NO BURNING PROPANE AIR RATIO -0.0942 H * HEIGHT ABOVE. FLAMEHOLDER BASE, CM PRESSURE ~ I ATM. TRAVERSE TAKEN AT EVERY 1/8'(32CM.) NOZZLE DIAMETER ~ I INCH OF HEIGHT SELECTED POINTS SHOWN HERE. CYLINDRICAL FLAMEHOLDER DIA. 1/4" HORIZONTAL TUBE PRESSURE TRAVERSES OF FLAME.;ENTER 0 FIG. 49 68

TOTAL PRESSURE & NO BURNING 4,. __ ". _ __ __ __ __ _ ^ ^. _!__ __ _ __^" — ar -m7. TOTAL PRESSURE & BURNING 0 0 LU G I PR E, DISTANCE DOWNSTREAM FROM FLAMEHOLDER -(INCHES) Lu _ __ —.. FIG. 50 PRESSURE MEASURE S OF F E A STATIC PRESSURE & BURNING STATIC PRESSURE & NO BURNING5 - - V2 1 l z1 2 2!Y 3 3z 4 FIG. 50 PRESSURE MEASUREMENTS OF FLAME ALONG CENTER LINE. r r 0 c) Z CI) 4 1 z 0 m.11 C) t= c Ig H Il 0I I^

I WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 of the flame zone, are shown in figure 51. The reading of the manometer connected to the flame holder was -0.2 cm. of water gage in the case of burning and -0.9 cm. of water gage in the case of no burning. These readings corresponded exactly to those obtained when the hypodermic tubing, orientated both vertically and horizontally, was placed directly over the hole in the flame holder. A few experiments were performed with the tube placed vertically, but pointed in the downstream direction. Although vacuums were generally obtained, a positive pressure was observed in the center of the flame at specific heights with a few fuel-air ratios possibly indicating a flow reversal under these conditions. Velocity calculations were not made from these measurements for several reasons. The stagnation pressure, Pst, has been defined7 as the sum of the static pressure, P, and the pressure rise resulting on isentropic deceleration to zero velocity. The equation Pst- P. and the velocity, u, has been derived from a force balance for steady functionless flow as Ps= P + pudu st J go 0 The vertical tube measurements might be assumed to correspond to the stagnation pressure with the assumptions stated above. However, it is believed that the data obtained with the horizontal tube traverses does not correspond so closely to the static pressure because of the aspirating effect of the gas moving past the end of the horizontal tubing as well as the lack of a surface upstream in which to obtain boundary layer buildup. Had a conventional pitot tube (i.e. double walled) been used in these measurements, it is believed that the necessarily larger diameter would have seriously disturbed the flame and flow process. As seen from the above equation, it is necessary to know the gas density as well as the static and stagnation pressures in order to compute the velocity at any point. To compute the gas density requires a knowledge of the temperature and composition of the gas as well as its static pressure. The experimental work of Morrison and Dunlap8 L 70

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 TRAVERSE TUBE ORIENTED VERTICALLY FUEL-AIR RATIO = 0.0942 JET VELOCITY = 109 FT/SEC READING OF MANMETER CONNECTED TO FLAMUOLDER = -0.9 CM LINES OF CONSTANT PRESSURE ARE SHOWN IN UNITS OF CM OF WATER GAGE TRAVERSE TUBE ORIENTED HORIZONTALLY FUEL-AIR RATIO - 0.0942 JET VELOCITY = 109 FT/SEC READING OF MANCOMTER CONNECTE TO FLAIEROIDER = -0.9 CM LINES OF CONSTANT PRESSURE ARE SHOW IN UNITS OF CM OF WATER GAGE I I I I FIG. 51 PRESSURE MEASUREMENTS OF FLAME. 71

I WILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN UMM-74 indicated that equilibrium conditions were not attained in propane-air flames at low jet velocities ca. 10 ft/sec and it is believed doubtful that equilibrium would be attained at all points in a flame burning in a jet with a velocity of 109 ft/sec. Thus the lack of knowledge of the gas temperature and composition as well as the uncertainty of the static pressure measurements makes a density, and,hence, a velocity, computation from these data inadvisable in the case of burning. It has been known for more than a century that gases from flames are conductors of electricity. Volta caused a discharge of electricity from the surface of a non-conductor by passing a flame over it.9 An attempt was made with the present equipment to measure the conductivity of the flame at different locations. Two wires, 3/16 inch apart, were traversed through the flame area and the voltage drop between them, compared to a 10 meg. ohm resistance, was measured. A reading of 0 volts was obtained when no flame existed. The data obtained (using the 1/4 inch cylindrical flame holder in the 0.1 inch unconfined jet of System 1 at atmospheric pressure) for a fuel-air ratio of 0.0942 and a jet velocity of 109 ft/sec may be seen in figure 52. Fuel-air ratio seemingly caused greater variations in the flame conductivity than did jet velocity. The values shown in figure 52, however, are integrated values and not point values, since the conductivity of a larger volume of gas than the volume included between the wire tips is measured. In an attempt to decrease the magnitude of this effect, it was decided to measure the resistance between two parallel flat plates with an area large in comparison with the wire area. Accordingly, two plates, ca. 2 mm by 2 mm were affixed to the wires and traversed through the flame. However, the plates were so large that they acted as secondary flame holders and no data was obtained. It should be noted that at locations above the pilot zone, the wires became incadescent, and, therefore, contribute an additional ionizing effect to the gases surrounding them. The combination of these two ionizing sources makes it virtually impossible to differentiate between them. It would be possible to obtain an empirical correlation between conductivity and fuel-air ratio or jet velocity for a given system which might have practical application, but at the present time it is believed that this would not add materially to a better understanding of the flame holding mechanism. i 72

"XVIIJLOWVV RItN I'4RE.:S)EARC:; A CE:NTIQ ER-UNIVERS ITY OF MICHIGAN U131'" I S,- V.. I I L FIG. 52 ELECTRICAL CONDUCTIVITY OF FLAME. 75

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 The only conclusion obtained from these and other experiments is that it is very difficult to determine experimentally absolute values of any physical property at different points in the flame region. The static pressure may be measured satisfactorily at one point, at the base of the flame, but the other measurements only indicate that the static pressure varies a great deal throughout the flame region, but no reliance is placed on the absolute values obtained. The lack of temperature and composition data as well as the uncertainty of the static pressure measurements prevents computation of the velocity from the total pressure measurements. A large number of shadow photographs were taken of flames, including Fastex shadowgraphs at 5000 pictures per second. Some of these shadowgraphs have been presented in previous progress reports10;112, as well as in figures 20, 21, 22, 29, 30, 31, 37, 38, and 39 of this report. However, these shadowgraphs do not permit any absolute determination of the flow pattern in the flame area except to indicate that the flow and flame process is a function of the many variables previously mentioned. Some work was done with smoke tracers of titanium tetrachloride, but no results were expected or obtained with their use, since the work of Langmuir and Blodgettl3 indicates that solid particles do not follow the gas streamlines in the range of particle sizes and Reynolds number with which we are concerned. An attempt to determine a composition distribution by the measurement of a physical property, electrical conductivity, was not successful because of the difficulty in differentiating between the ionizing effect of the flame and the ionization caused by the emission of electrons from the hot probes. The conclusion is that it is very difficult to obtain absolute values of the flow or composition distribution in the flame region, at least with simple apparatus. It is possible that some combination of optical methods, such as the use of an interferometer and an emission spectrometer might be of value in determining absolute values of the flow and composition distribution in a flame region. 74

WILLOWV RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UM-74 V. CORRELATION OF BLOWOFF DATA A. Previous Attempts It was previously reported14 that a straight line was obtained when the fuel-air ratio at the maximum blowoff velocity (for each spherical flame holder) was plotted against 1 Re DG (R = Reynolds number = - e kL D = spherical flame holder diameter G = maximum mass velocity at blowoff L = viscosity of fuel-air mixture issuing from nozzle) This straight line was obtained with the use of data from System 1 (1 in. nozzle) at atmospheric pressure. Because of the relationship between 1 4 — and the thickness of a boundary layer, this type of plot led to e the boundary layer experiments reported in Section IV. However, when experimental data was obtained under vacuum with System 2 using a 3/8 in. diameter nozzle, a series of similar curves (for different pressures) was obtained rather than one straight line. These may be seen in figure 53. As may be seen from this plot, the 1/16 inch diameter spherical flame holder data does not fall on the straight line obtained with the other three spherical flame holders. Straight lines were obtained, however, when the Reynolds number was plotted against sphere diameter, for maximum blowoff velocity. These lines all intersect at R = 0, D = 0.04 inches as may be seen in figure 54, which might indicate that sphere diameters smaller than 0.04 inches would not hold a flame. It was noted under vacuum with the 0.0625 inch diameter sphere, that burning occurred sometimes as a Bunsen flame, rather than from the flame holder. Experiments were not performed with smaller flame holders, however. Calculations were made to determine the amount of chemical energy of the fuel contained in the boundary layer flow. This "heat release I 75

WILLOW RUN RESEARCH CENTER- UNIVERSITY OF MICHIGAN UMM-74 0.05 0.045 0.04 0.035 U ILU > 0.03 u. u.3 0 o 0.025 < 0.02 I-' 0.015 0.t0 0.005 0.07 0.075 0.08 0.085 0.09 0.095 FUEL-AIR RATIO- LB. PROPANE/LB. DRY AIR 1 FIG. 53 FUEL-AIR RATIO VERSUS AT MAXIMUM BLOWOFF VELOCITY. /REYNOLDS NO. 76

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN I —-------- UTMM-74 - U LL. I0 9 I LU z M) 0 LU 0.05 0.1 0.15 0.2 0.25 SPHERICAL FLAMEHOLDER DIA. - (INCHES) FIG. 54 REYNOLDS NUMBER AT MAXIMUM BLOWOFF VELOCITY VERSUS FLAME HOLDER DIAMETER FOR VARIOUS PRESSURES. 77

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM- 74 of the boundary layer", (BTU/sec) was plotted against the mass velocity at blowoff to obtain a straight line from the origin for all fuel-air ratios. However, different lines were obtained for each sphere tested. The slope of these lines was extremely sensitive to the value of AH (enthalpy change) used to determine the "heat release of the boundary layer." Equilibrium values as well as values calculated from the experimental Pb/ P data of Morrison and Dunlapl5 were used. However, when the data was plotted as indicated by dimensional analysis Vb p (i H vs DP AH on log-log paper) four lines, one for each of the spheres tested were obtained rather than a single line for all spheres. Other investigators have obtained limited correlation of blowoff data. Scurlockle attempted to correlate his data as well as the blowoff data of Longwell, and the data of Lewis and Von Elbe by plotting the log of Vbo against a generalized oxidant fraction (i.e. a function Do.4 of fuel-air ratio so computed that a stoichimetric mixture of all fuels is 0.5). Different curves were obtained for the different fuels as well as the data obtained in different systems. Scurlock did not obtain correlation with different fuels in the same system by plotting V bo.o45. V (T - T ) vs. generalized oxidant fraction 0o.45 f b 0 V = blowoff velocity, ft/sec bo D = flame holder characteristic dimension V = laminar flame speed f Tb - T = temperature rise due to burning. b o Scurlock did, however, succeed in correlating the data obtained in his system at atmospheric pressure by plotting V versus fuel-air ratio b 0.45 to obtain a single curve for different flame holder sizes and spheres. This data was obtained in a confined test chamber, 3 inches by 3 inches by 17 inches long, the flame holder being nine inches above the 3 inch square nozzle. De Zubayl7 obtained a similar type of correlation, I 78

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 Vb but including pressure, by plotting fuel-air ratio vs. log 08 0.9 T090o p50.9 D P V = blowoff velocity, ft/sec D = disc diameter, inches P = pressure, psia to obtain a single curve for three different diameter discs (1/4 inch, 1/2 inch, and 1 inch diameter). This data was obtained in a 2-3/4 inch diameter cylindrical test chamber with the disc being held 2 inches from the inlet of the 4 inch long test chamber. A water spray nozzle was inserted in the flame ca. 2 inches downstream of the holder. Jensen and Miessol8 found that separate equations were necessary to characterize fuel rich and lean blowoff. For 1/4 and 1/8 inch diameter spheres and discs, for propane,isobutane, and ethane, the equation for the lean blowoff was found to be (Pef)2 1 Pe ( )0 CD while the equation for the fuel rich side was found to be 0.6 <q_ Pe = f 10,300 - in which Pe = Peclet number based on flame speed, Pe = Peclet number based on blowoff velocity, v CD = drag coefficient, CkL -= Prandlt number of fuel-air mixture, k D. = nozzle diameter = 1 inch, D = flame holder diameter. 79

WILLOW RUN RESEARCH CENTER - UNIVERSITY OF MICHIGAN UMM-74 However, the Peclet number is merely the product of the Reynolds number and the Prandlt number, and the Prandlt number is essentially a constant for the fuel-air mixtures used. The density and viscosity vary only slightly (since the data is obtained at atmospheric pressure); D is a J constant, since only one nozzle was used in the experiments, so that the lean equation may be reduced to D V D (D V )2 K -- K Vb or (V)o f 1.7 0.5 D CD while the rich blowoff equation may similarly be reduced, in terms of actual experimental variables to V $- V =K b f D0.1 0.25 D CD This data was obtained in a one inch diameter nozzle in an unconfined jet with a 3 inch diameter plenium chamber upstream of the nozzle. The height of the flame holder above the nozzle was not stated. As may now be seen, three different investigators obtained similar type correlations of blowoff data, i.e. Scurlock - F/A a(b / at 1 atm in partially a confined jet a = 0.45 De Zubay F/A aD ~ under vacuum in D P partially confined jet a = 0.85 b = 0.95 i 80

WILLOW' RUN RESEARCH CENTER- UNIVER SITY OF MICHIGAN UMM-74 Jensen and Miesso - at 1 atm in unconfined jet for lean blowoff (V)2 a (F/A)2 a. Vb f Da Cc D CD a = 1.3 c = 0.5 for rich blowoff V Xv a F/A - bI f a c D C D a = 0.1 c = 0.25 However, none of these investigators succeeded in obtaining any straight line correlation involving fuel-air ratio, nor did any of them include such variables as nozzle diameter or height of the flame holder above the nozzle. In addition, Scurlock and De Zubay obtained their data in a partially confined jet. The wide variation in the exponent of the flame holder diameter (0.1; 0.45; 0.85; 1.3) could probably be attributed in part to the differences in the testing systems, but is also a function of fuel-air ratio, as may later be seen. Some of the data obtained during this investigation could be represented by the correlation obtained by the 5 previous investigators ( i.e. F/A a Vb ) but a b c D P CD no one equation would represent the data when the system variables were imposed. A great many unsuccessful attempts were made to apply dimensional analysis to the experimental data obtained at this laboratory, and no one theoretical equation could be found which represented all of the data. It is difficult to evaluate such terms, for example as viscosity, which affects the flow pattern around the holder, but which varies from point to point due to temperature and composition changes. 81

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 Since previous correlations proved inadequate,and because no theoretical equation was found which represented all of the experimental data obtained, an empirical equation was developed which represented the experimental data obtained. This equation, for a completely unconfined jet, spherical flame holders, and propane-air mixtures, includes the following experimental variables G, the mass velocity at blowoff D, the spherical flame holder diameter F/A, the fuel-air ratio P, the absolute pressure D., the nozzle diameter and h., the height of the flame holder above the nozzle. This equation is P a 0.29 B -7.2 + 17.32 Dj - 4r] G = ( ) ( 8.42 (1- r) in which 1 1.15 D + 6.6 F/A - 0.255 (0.100/F/A)3 11 B = -0.91 1.42 D. - 1.54 r r = - 1.8 ( - DJ1 4)2 -0.60 D The development of this equation and its limitations are presented on the following pages. 82

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 B. Development of Empirical Correlation The mass velocity at blowoff (G, (lbs/sec/sq.ft. of nozzle area)) was plotted against the absolute pressure P, in mm Hg. on log-log paper to obtain a straight line of slope a. An example of this plot may be seen in figure 55 which is for a 1/16 inch diameter spherical flame holder, the 3/8 inch diameter nozzle in System 2, at a flame holder height of 3/16 inch. Plots similar to figure 55 were constructed for all of the vacuum data obtained in System 2, i.e. for flame holder diameters of 1/16, 3/32, 1/8, and 0.189 inches, each flame holder being tested at nine fuel-air ratios from 0.06 to 0.100. The values of a, the slope of the line on figure 55 and similar figures, were then plotted against fuel-air ratio as is shown in figure 56. As may be seen in figure 56, the slopes of the lines for the four spheres are the same, 6.6, so that the equation of lines on figure B is (1) a= 6.6 (F/A) + b in which b is the intercept at F/A = 0 and is a function of sphere diameter. The values of b are plotted against sphere diameter in figure 57 to obtain a straight line whose slope is 1.15 and whose intercept at D = 0 is -0.255. Hence, the equation of the line in figure 57 is (2) b= 1.15 D - 0.255 substituting this value of b in equation (1) we obtain (3) a= 6.6 (F/A) + 1.15 D - 0.255 Hence, referring to figure 55, we have (4) log (1) = a log G + c 100 p 1/a (5) or log G = log (1-) - c p 1/a Values of (-) were then calculated for all of the vacuum points 100 1/a obtained in System 2. Log-log plots of ( P ) vs. G were then drawn 100 83

7.60 I d) E E au W. w au: Of a 5.00 2.00 2.00 r 0 z c Z n o (fi tr 2 m; 0 ~ —, 0 F= a~ 1.00 1 2 5 10 20 D=1/16 F/A = 0.080 G - LBS./SEC./SQ. FT. OF NOZZLE AREA AT BLOWOFF FIG. 55 DEVELOPMENT OF CORRELATION LOG G VERSUS LOG P

0.6 0.5 0.4 0.3 D- D0.189,- D: 0.125 -1- D:0.09375 -. D: 0.0625 0O \JI 0.2 0 r r 0 o c z PI C() z m r4 0 t cl g H n 0 J.,. 0.1 0 < INTERCEPT-b -0.1 -0.2 F/A:FUEL-AIR RATIO - LB. C3H6/LB. AIR DEVELOPMENT OF CORRELATION FUEL AIR RATIO VERSUS a FIG. 56

-0.3 -' -0.2 -0.1 0 INTERCEPT -0.255 -SLOPE 1.15 - - - - -I - -- -- - - - ^ r r o 0 C z z C n 1 m 1,4 o tv g 14 t3 cn H O! 0 it t, 0.1 D= SPHERICAL FLAMEHOLDER DIAMETER - INCHES DEVELOPMENT OF CORRELATION FLAME HOLDER DIAMETER VERSUS b 0.2 FIG. 57

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN _MM-74 with a slope of 1 for each of the four spheres and their nine fuel-air ratios. An example of one of these plots may be seen in figure 58 which is for a fuel-air ratio of 0.080 and a sphere diameter of 1/16 inch. The intercepts at G = 1 (.. log G = 0) were obtained from figure 58 and similar plots. Equation (5) may be written as p 1/a (6) log G = log (-) - log c P l/a or (7) G= (10) c in which c is the intercept obtained at G = 1 from figure 58 and similar figures. The values of c obtained were plotted vs. D (sphere diameter) on log-log paper as is shown in figure 59. As may be seen from figure 59, the lines of constant fuel-air ratio intersect at a common point, D = 0.29 and c = 1.4. The equations for these lines may then be written (8) log (0.29) - log D = d (log 1.4 - log c) 0o29 14 or (9) log ( D) = d (log - ID c Values of the slope d were then plotted against F/A to give a straight line on log-log paper as may be seen in figure 60. The slope of this line is 35.11 and oAe point on this line is F/A = 0.100 at d = -0.91. Hence, the equation for the line shown in figure 60 may be written as (10) s = 3 log - 0.91 - log d (10) slope 5.11 (10) s lope = = log 0.1 - log F/^ 0.1 - 0.9 or (11) 3.11 log F/A = log F/A g d -0.91 or (12) d 0. 11 o.1 3.11 F/A Substituting this value of d in equation (9) and solving for c, we obtain 1.4 (13) 0.1/FA3.11,__.29 (-.8 _______ \ D/ 87

* 0t-4 10 ( j O Su s U o t w < U U) 0 -A F / A:0.080 bD= i /16u ~~~~~~~INTERCEPT@G:I IS C~~~~~~~~~~~~~~~~~~~ /l I~ IISLOPE =,, Il INTERCEPTr G=I IS C _______ i I/\>_________________ __ e04 r r C z Z CI, C1= H z z 0 o~ gpi im z^ wp H3 P( 2 1 50 100 200 (P) I /a 100 P FIG. 58 DEVELOPMENT OF CORRELATION LOG G VERSUS LOG (-) 100 500 1 a

0.4 u 0.2 1 LL I I SL-PE z o 0.1 LU LU * n 9 - - _.04.060 I F o.080 _ _.075 F / - 0.1001.095.090 |.85.04 _____ 095.090 1.0 I I_ I_____ 1__ 1.0 10 100 FIG. 59 DEVELOPMENT OF CORRELATION LOG D VERSUS c FIG. 59 DEVELOPMENT OF CORRELATION LOG D VERSUS c r 2A r o z n L'q C3:0 2: I o g 0-4 0O;

-1.0 -0.5 -0 0 r r 0 0 z z rX C Z I) (/] ~ 124,) 2 d -0.2 0.06 F/A LB. PROPANE/ LB. AIR 0.100 FIG. 60 DEVELOPMENT OF CORRELATION LOG D VERSUS LOG FUEL AIR RATIO

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 and substituting this value of c in equation (7), we find P C1 0 B29 B (14) G 100 D 1.4 in which 1 1 a 1.15 D + 6.6 (F/A) -0.255 and 1 (0.1/F/A)311 B d -0.91 Equation (14) should be valid for spherical flame holder diameters (D) from 0.0625 to 0.189 inches in diameter, for propane-air (F/A) ratios from 0.06 to 0.100, and for pressures (P) from 760 to 200 mm Hg abs. It is, however, valid only for the above conditions when used in a 3/8 inch diameter nozzle at a flame holder diameter of 3/16 inch, when tested in a system in which the ratio of test chamber area to nozzle area is at least 1100 to 1. Thus, in order to determine the mass velocity at blowoff (G) which would be obtained if the system geometry was varied, additional terms must be added to equation (14). The effect of nozzle diameter and height of the flame holder upon the mass velocity at blowoff was presented in Section III. In this correlation, no attempt was made to incorporate the data obtained with the use of glass wool upstream of the nozzle or to consider the effects of jet confinement. The problems of jet confinement and the resonance change induced by glass wool are intimately connected with flame stabilization,but resonance effects in confined jets have been considered in a previous report2 from these laboratories, and not enough quantitative data was obtained in this investigation to include these effects in a correlation. Hence, the nozzle diameter and height of the flame holder are the only system variables which are to be incorporated into equation (14). Figure 61 shows the effects of the height of the flame holder upon the mass velocity at blowoff for a 3/8 inch and a 5/8 inch diameter nozzle in an unconfined jet. (The data used in equation (14) was obtained with a 3/8 inch nozzle in a 10 x 12 x 18 inch test chamber. However, as was previously shown, and,as may be seen in figure 61, there is essentially no difference between this degree of confinement, test chamber area = 1100/1, and a completely unconfined jet.) nozzle area 91 j

*- v E 1 1 1. X r C I NOZZLE DIAMETER:3/18 WITHOUT GLASS WOOL & 100 MESH SCREENS UPSTREAM OF NOZZLE. BURNING IN UNCONFINEDJET. NOZZLE DIAMETER=3/8' WITHOUT GLASS WOOL & 100 MESH SCREENS UPSTREAM OF NOZZLE. BURNING IN TEST CHAMBER 10" X 12 X 18" (ONE CURVE DRAWN FOR THESE POINTS, I.E., NO DIFFERENCE BETWEEN CONFINED AND UNCONFINED BURNING FOR 3/8' JET WITHOUT GLASS WOOL UPSTREAM OF THE NOZZLE) NO77ZLE DIAMETER:SI/'.......... -___ v r\) u) uj I U z -a " I I ILU I 0 o Vx s4< WITHOUT GLASS WOOL & 100 MESH SCREENS UPSTREAM OF NOZZLE. BURNING IN UNCONFINED JET. I I NOTE: IT WAS IMPOSSIBLE TO IGNITE THE MIXTURE UNDER THE FOLLOWING CONDITIONS: NOZZLE DIAMETER:5 /8 | WITHOUT GLASS WOOL & 100 MESH SCREENS UPSTREAM OF NOZZLE. BURNING IN TEST CHAMBER 10" X 12" X 18' ____ PRESSURE 1 ATMOSPHERE PROPANE-AIR RATIO 0.080 I^N SPHERICAL FLAME HOLDER DIAMETER 1/16' r r 0 Z z o c 20 m) 0 13 t3 ) O M) vT" *4* "Johmozo4ft, vv q*ftftft v -doI I I I I I I -- 142" - xA - I I I ,%^.If v v V 0 0 2 4 6 8 MASS VELOCITY AT BLOWOFF (LB./SEC./SQ. FT. OF NOZZLE AREA) 10 12 14 FIG. 61 EFFECT OF NOZZLE DIAMETER ON MASS VELOCITY AT BLOWOFF

WILLOWV RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 Figure 61 indicates that there is a maximum height for the flame holder, and that this maximum height is some function of the nozzle diameter. For the 3/8 inch and the 5/8 inch diameter nozzles, the optimum h was D. ^r J l.found to equal numerically found to equal numerically D i in which h is the height of the flame holder, and Dj is the nozzle diameter, both expressed in inches. It was found that the experimental data shown in figure 61 a parabolic type of curve (i.e. (g-k)2 = 4p (x-j)). obtained for both nozzles when G, the mass velocity plotted against I could be fitted Straight lines at blowoff, was to were G- 4 ih -(h) 2 _ 0.6 \ Di Djmax \ i fj G - 4 h [1;84 2 \ J Dj - i 0.6 G - HI Straight lines so obtained are shown in figure 62. As may be seen, the two lines intersect at G = 11.6, G - 4 = -12.2. If the slope of these H lines (f) (which is some function of nozzle diameter) were known, the equation of these lines would be represented as (15) f = 6 -12.2 - G -4 H I or rearranging (16) G = f 12.2f - 4H + 11.6 H 1 - f/H Although complete height vs. G data is available for only two nozzle diameters, 3/8 inch and 5/8 inch, data was obtained at h = 0.5 for DJ 7/16, 1/2, and 1 inch diameter nozzles. The use of this data would give one point for each of these three nozzles on figure 62. If the assumption were made that lines representing other h for these nozzles D IJ 93

10 s o LU m o z I- U. U 0 S > ) ) oU I i 8 6 0t r r c z z o Z C/' H C C) e M-] - HP7 z! 4 2 h h. -j - (.6 -06 i,-0. 0.6 k= Flameholder Dj= Nol33e Dia.(lnehes) -2 4 -6 -8 -10 -12 G - 4 FIG. 62 DEVELOPMENT OF CORRELATION G VERSUS (Dh 1.83 02 (6 — D. )- 0.6.1 i

WILLOW RUN RESEARCH CENTER UNIVERSITY OF MICHIGAN UMM-74 also passed through the intersection point of the 3/8 and 5/8 inch nozzle, then the slope (f) could be determined for these nozzles as well as for the 3/8 and 5/8 nozzles for which more complete data is available. This was done and the values of f were plotted against Dj in figure 63 to obtain a straight line whose equation is (17) f = 1.42 Dj - 1.54 This value of f was substituted in equation (16) to obtain the expression for G shown below -7.2 + 17.32 D - 4 (18) G = 1 - f/H in which 1.42 D - 1.54 f j__ _ _ H (D i ) - 0.6 The values of G. the mass velocity at blowoff, predicted by equation (18), are compared with the experimental values obtained, in Table 2. In general the highest experimental value obtained was used to draw the curves shown in figures 61 and other blowoff curves. As may be seen from Table 2, equation (18) represents the experimental data reasonably well, except in the 1/8 inch and possibly in the 1/4 inch nozzle. This deviation of the smaller nozzles could perhaps be explained by blocking of the nozzle by the flame holder. Thus, a correction for nozzle diameter and flame holder height may be applied to equation (14) to obtain the following expression P a 0.29 B (19) 1D -7.2 + 17.52 Dj - 4 F 1.4 8.42 (1 -r) I 95

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 TABLE 2 COMPARISON OF EXPERIMENTAL VALUES OF THE MASS VELOCITY AT BLOWOFF WITH THE VALUES OBTAINED FROM EQUATION 18 Sphere Diameter = 0.0625 inches F/A = 0.080 lb/C3H8/lb air Completely Unconfined Jet No Glass Wool Upstream of Nozzle Nozzle Diameter Inches Dj Height of Flame Holder Nozzle Diameter h/Dj Mass Velocity at Blowoff lbs/sec/sq. ft. G Experimental G from Eq. 18 1 0.5 1.5 135.5 5/8 0.5 6.4 - 7.0 7.8 1/2 0.5 7.1 - 8.9 9.o 7/16 0.5 7.8 - 8.2 8.7 3/8 o.5 7.8 - 8.9 8.4 1/4 0.5 6.5 - 7.1 8.2 1/8 0.5 4.2 - 4.6 7.1,,,, I 3/8 3/8 3/8 3/8 3/8 3/8 5/8 5/8 5/8 5/8 5/8 5/8 5/8 5/8 5/8 5/8 0.1667 0.333 0.5 0.667 0.833 1 0.1 0.2 0.3 0.4 0.5 0.6 0.8 1.0 1.2 1.6 10.5 9.6 7.8 6.0 4.3 3.5 4.0 4.3 6.0 6.0 6.5 5.8 4.9 3.9 3.8 3.5 - 10.9 - 10.2 - 8.9 - 6.5 - 4.9 - 3.7 - 4.4 - 4.8 - 7.3 - 7.3 - 7.0 - 6.4 - 5.3 - 4.7 - 4.0 - 3.8 10.9 10.2 8.4 6.5 4.9 4.5 5.0 6.0 7.3 8 7.8 6.5 4.9 4.3 4.0 3.8 96

WILLOWX RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 in which 1 1 a 1.15 D + 6.6 (F/A) - 0.255 B 1 [0.100/(F/AI 3.11 d -0.91 f 1.42 D. - 1.54 H j ( D "- 0.60 OJ Dj ) 6 when the variables are expressed in the following units G = mass velocity at blowoff lbs/sec/sq.ft. of nozzle area P = absolute pressure, mm of Hg D = spherical flame holder diameter, inches F/A = lb propane/lb air D = nozzle diameter, inches j h = height of flame holder above nozzle, inches This equation, then, is valid for fuel-air ratios from 0.06 to 0.100, sphere diameters from 0.0625 to 0.189 inches, and pressures from 200 mm to 760 mm Hg abs, in a 3/8 inch diameter nozzle with h/D = 0.5. It is also valid for a F/A of 0.080, D = 0.0625, for nozzle diameters from 1/8 to 1 inch, at h/D = 0.5, and for 3/8 and 5/8 inch nozzles from h/D = 0 to h/D = 1.6. It is conceivable that the equation would predict the mass velocity at blowoff for other conditions, but no experimental data is available to test the equation. The equation is valid for completely unconfined jets, or at least jets in which the area of the test section to the nozzle area is at least 1100 to 1. It is known that a greater degree of jet confinement reduces the mass velocity at blowoff considerably due to resonance and other effects. The size of the plenium chamber upstream of the nozzle, as well as insertion of glass wool immediately upstream of the noz-zle,also changes the mass velocity at blowoff, probably because of resonance between the flame and the upstream chamber. The representation of all the experimental data obtained by equation (19) may be seen in figure 64. 97

WILLOW RUN RESEARCH CENTER — UNIVERSITY OF MICHIGAN UMM-74 f 0 1/4 1/2 3/4 1 DJ FIG. 63 DEVELOPMENT OF CORRELATION D VERSUS f J 98

WILLOW RUN RESEARCH CENTER -UNIVERSITY OF MICHIGAN UMM-74 CORRELATION OF EXPERIMENTAL BLOWOFF DATA FOR IN. PROPANE-AIR RATIOS (F/A) FROM 0.06 TO 0.100 D = 3/8" D 3/8 SPHERICAL FLAMEHOLDER DIAMETER (D) FROM 0.0625 TO 0.189INCHES OD 0.(5 ABSOLUTE PRESSURES (P) FROM 200 TO 760 MM. HG. AND D 0.0625 F/A 0.08 P ~ 760 MM. NOZZLE DIAMETERS (q) FROM 1/8" TO I" FLAMEHOLDER HEIGHTS (h) FROM 0" TO 1-1/8" w 4 w z LJ 0 Ci N N 0 -J ILL o C> cn m LIJ (/) U) <r I PN ato.?IO (o\ D [- 7.2 + 17.32Dj -4r 1.4 8.42(1 - r) IN WHICH: I Cr = 1.15D + 6.6( F/A -0.255) 0. 1OO/F/A] 3.11. = -.091 1.42 Dj - 1.54 r = I -D y -~ s rj _ D jE4 0.60 CORRELATION OF EXPERIMENTAL BLOWOFF DATA FIG. 64 99

L WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 BIBLIOGRAPHY No. 1 Hougen and Watson, "Chemical Process Principles, Part 2, Thermodynamics," John Wiley and Sons, New York, New York, 1947, p. 489. 2 University of Michigan External Memorandum No. 43, "Resonance of a Flame in a Parallel Walled Combustion Chamber," R. A. Dunlap. 3 University of Michigan UMR-27, Progress Report No. 3, AAF Contract W33-038-ac-2100, pp. 3 - 12. 4 University of Michigan UMR-33, Progress Report No. 6, AAF Contract W33-038-ac-2100, pp. 3 - 10. 5 Battelle Memorial Institute, Progress Report No. 30, "Aerodynamic Conditions Affecting Blowoff of Burners," AMC Contract AF-33-038-2038 6 Ibid, figure 161. 7 Hottel, Williams, and Satterfield, "Thermodynamic Charts for Combustion Processes, Part One, Text," John Wiley and Sons, New York, New York, 1949. 8 University of Michigan External Memorandum No. 21, "Measurement of Flame Speeds with the V-Flame," R. B. Morrison and R. A. Dunlap, 1947. 9 Thomsen and Thomsen, "Conduction of Electricity Through Gases, Vol. I," Cambridge, 1928. 10 University of Michigan UMR-33, Progress Report No. 6, AAF Contract W33-038-ac-2100, figure 1, p. 5. 11 University of Michigan UMR-31, Progress Report No. 5, AAF Contract W33-038-ac-2100, figure 1. 12 Ibid, figure 2. 13 Langmuir and Blodgett, USAF Technical Report No. 5418, February 19, 1946, Graph Reported in Perry, "Chemical Engineer's Handbook," Third Edition, McGraw Hill Book Co., New York, 1950, p. 1022. 14 University of Michigan UMR-29, Progress Report No. 4, AAF Contract, W33-038-ac-2100. 15 University of Michigan External Memorandum No. 21, "Measurement of Flame Speeds with the V-Flame," R. B. Morrison and R. A. Dunlap, 1948, figure 17, p. 27. L 100

WILLOW/ RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UMM-74 BIBLIOGRAPHY (CONTINUED) No. 16 A. C. Scurlock, "Flame Stabilization and Propagation in High Velocity Gas Streams," Meteor Report No. 19, MIT Guided Missiles Program, May 1948. 17 E. A. De Zubay, "Flame Stability Characteristics of Circular Discs," Scientific Paper No. 1506, Westinghouse Research Laboratories, Westinghouse Electric Corp., March 9, 1950. 18 Summary Report to AMC, Contract AF-33-038-2038, Battelle Memorial Institute, Section on "Aerodynamic and Other Conditions Affecting Blowoff of Burners," R. A. Jensen and C. A. Miesso, May 15, 1950. 101

WILLOW RUN RESEARCH CENTER-UNIVERSITY OF MICHIGAN UrTI —74 DISTRIBUTION Distribution of this report is made in accordance with AMC letter dated 8 April 1949. 102