Protection of aircraft fuel tanks against explosion hazards using inert combustion products

LC% *-m '

^ S DELFT

College Report No. 85

E HOGESCHOOL

GBOUWKUNDE O - DELFT

2 8 J'üll 1955

THE COLLEGE OF A E R O N A U T I C S

CRANFIELD

PROTECTION OF AIRCRAFT FUEL TANKS

AGAINST EXPLOSION HAZARDS USING

INERT COMBUSTION PRODUCTS

by

E. M. GOODGER, M.Sc.(Eng.), A.M.I.Mech.E., A.F.R.Ae.S.. A.F.Inst.Pet. F/Lt. P. CADMAN, D.C.Ae., and

Kanaalstraat 10 - DELFT REH3RT HO, 85

OCTOHER, 1 9 5 4 .

T H E C O L L E G E O F A E R O r T A U T I G S C R A N F I E L D

The Protection of Aircraft Fuel Tanks Against Ejqslosion Hazards using Inert Combustion Products

-by-E.M, Goodger, M.Sc,(Eng), A«M.I.MeGh,E,, A.F.R.Ae.S., A.F.Inst.Pet.

P / L t , P . Cadman, D . C . A e , ,

and S/Ldr„ IoT,A, i ï a r c h i e , D.C.Ae,, A . F . R . A e . S ,

SUIvaiARY

YiTithin certain ranges of temperature and pressure, the air-vapour mixture produced above the .liquid fuel in aircraft fuel tanks v.dll support combustion, axid vn.ll b u m explosively on the addition of the necessary energy for ignition, Althou^ the danger zone for ea.ch type of aviation fuel can be clearly defined under equilibrium conditions, the many transient factors involved in practice make it difficult to find completely safe operating conditions for any fuels

A T/ar-time approach to the solution ^Tas the continuous purging of the tank free space v/ith nitrogen to reduce the

oxygen content of the admixed air below the minimijmi permissible for combustion. In order to avoid the weight and bulk of s±r-borne nitrogen storage equipment, combustion products hc^ve been

suggested as an alternative and continuously-available supply of inert gases. The source of these gases may be cither the

primary zone of some combustion chambers of the main gas-tiorbine engines, or a separate inert gas generator located vd-tliin the aircraft,

This report describes experiments made on sampling from the primary zone of a Den/ent I chanber, and on the testing of various automatic control devices suitable for an inert gas

generator. The conclusions dravm from the results are that both systems of purge gas supply rre feasible, and that the critical condition vrill be the high altitude dive Tn.th engine idling, v,'hen the purge gas requirement is a maximum, and the supply a miniraLmi,

Tliis report is based on sections of a thesis submitted by F/Lt. Cadman in Jtine 1953.» and a thesis submitted by S/ldr. IvLirchie in June 1954, as part of the requirements for the av/ard of the Diploma of The College of Aeronautics. The investigations v.'-ere made under the supervision of I-ir, Goodger,

„2"

COIMTENTS Summary

''« The Explosion Problem, and Safety Requirements 1,1, Introduction

1 ,2o Explosive limits in Aircraft Fuel Tanks

1.3. Combustion Products as a Possible Inerting Fluid 1.4. Plow Requirements for Purged Tanks

2. Tapping; the Liain Engine

2.1, Exploratory Sainpling with ïfater-Cooled Probe 2,2c Sampling with Uncooled Probe

2,3o Sampling in Combustion Chamber Dane Head 2.4. Temperature Traverses in the Low Oxygen Zones 2.5. The Practical Fuel-Cooled Probe

2.6. Sampling with the Fuel-Cooled l^obe 3 • The Separate Inert Gas Producer

3.1. Thermal Conductivity 3.2, Combustion Temperatxare

:),'i. Exhaust Gas Density 3,4. Oxygen Paramagnetism 4» Conclusions

4,1 • Primary Zone Sampling

4,2, Separate Inert Gas Producer

Appendix A - Calculation of Purge Gas Plow Requirements Appendix B - Notes on a Proposed Chanber Probe Installation,

1 • The Explosion Problem, and Safety Requirements 1,1, Introduction

The fire hazard in aviation is a problem which has called for continued research, and the advances in aircraft design and fuel handling techniques have done much, in civil aviation, to reduce fire to a fairly remote possibility. In military aviation, however, fire initiated by enemy action is the

direct cause of most aircraft losses, and further active develop-ment is therefore vital.

The general problem can be subdivided into the various aspects of airframe and engine fires occ\irring in the air or on the groiind, and includes the initiation of fires v/ithin the air-craft fuel tanks. This latter aspect is particularly serious in that it almost invariably leads to explosion and the total loss of the aircraft. The conditions under which such disasters may occur, together i.dth a possible method of protection, foiTA the basis of this report.

As in all combustion processes, explosion is only poss-ible when the tvro requirements of ignition energy and inflammable niixtiore strength ore met. Sufficient energy for ignition may be provided if the tank is pierced by incendiary missiles, or by metallic particles Vi/hich may spark on striking intomeil baffles. The strength of the vapour-air mixture above the liquid fuel in an aircraft fuel tank is influenced by a number of factors and, under certain circumstances, lies v/ithin the inflammable range for the fuel,

As it is impraxïticable to counteract completely the ignition danger by means of extensive arrao-ur-plating, recent dev-elopment work on explosion safety has proceeded mainly upon the following alternative

lines.-(a) Using nitrogen for continuously purging the vapour-air mixttire in order to render it inert, V/here necessary, the nitrogen also acts as the pressurising fluid.

(b) Projecting extingiiishant into the vapour as soon as a small rise in tank pressure indicates an incipient explosion,

The explosion suppression method has been shown to be markedly effective but, being a single shot system, cannot be reset for repeated use. The nitrogen purging system has been developed to give explosion safety for the tanks of large air-craft, but is unsuitable for small aircraft due to the buUc of equipnent required. In order to overcome this dra\7back, com-bustion products have been suggested as an alternative and

continuously-available supply of inert gas. The source of these gases may be either the primary zone of some ccxibustion chanibers

-2^

of the main gas tiurbine engines, or a separate cccibustor, autonaatically-controlled to operate at or near stoichiometric conditions, located v.ithin the aircraft. The design of a sep-arate canbustor_ for inert gas purging is described by Hill

(ref, l ) , but little information has been found regarding primary zone sampling. It was decided at the College of Aeronautics, therefore, to investigate the possibilities of both methods of inert gas production,

1,2. Explosive Limits in Aircraft Fuel Tanks

Ifcst published results for the inflammable limit mix-tures of fuel vapour in air (or oxygen) relate to homogeneous conditions within spheres (ref. 2) or cylindrical tubes (ref. 3 ) . For the hj'-drocarbons contained in average commercial fuels, the inflammable range extends from approximately 1.2 to 7.0 per cent by volijme of fuel vapour in the mi:>:ture vn.th air at atmospheric pressure and temperature, and these limits gre.dua-lly approach each other as the temperatirre level is reduced. Reduction in pressure has little effect imtil a pressure level of about

200 m^m. merciory (= 32,600 ft, I,C,A,N,) is reached, after viiich the limits approach each other to meet at a minimtim pressure of about 50 m«m, mercury (= 6l ,500 ft,), depending upon the energy of the igniting agent,

Results have no\/ been presented (ref, 4) to effect the calculation of vreak and rich explosive limit mixtures for aviation gasoline in aircraft fuel tanlcs together with the temperatures at which these mxtures occur. Further results have been presented in the form of curves (fig, 1) shoi,-óng the variation of the limit

temperati-ires \rith. altitude for most aviation fuels (ref, 5)« ï^e

conditions leading to a tank explosion are similar to those obtaining during a laboratory deterLiination of flash point, when an igniting flame is introduced into the vapour space above the liquid surface, Ho"ivever, complete agreement between laboratory and tank tests cannot be expected, due to differences in dimensions and scale, and it is noticeable that the lov.'er ejrplosion limit temperatures at sea-level are some 7 C belo\7 the respective flash point values. The conventional flash point is thus no true criterion of safety from explosion in a tank containing tmdis-turbed fuel and air»

These temperature curves shov: that the vulnerability of aircraft tanlcs to explosion depends upon the initieJ. fuel

tempera-ture, together vrlth the rate of clijnb and subsequent behaviour in

the air. For example, kerosine would be the safer fuel to use when operating wdth a low ground temperature and rate of climb

(e.g, heavy boiriber in the arctic), and gasoline preferable for a high-speed aircraft in the tropics. This is in direct opposition to the high specific gravity requirement for the modern high

Nominally, these temperature limit boundaries enclose the equilibrium danger areas, but explosion is also possible beyond both limits under the following circui-.istances, ViTien

liquid fuels are sprayed into the air, the greatly increased surface area results in a correspondingly increased rate of vaporisation, and a greater effective volatility. Explosion thus £xlso becomes possible beyond the low temperature side of the normal boundary curve, and may occur when a missile enters a tank and causes the liquid fuel to splash,

Explosion on the high temperatiire side of these bound-aries is possible when the oxygen content of the air in the mixture is increased, the effect being to displace the boundary curves bodily in the high temperature direction. As hydrocarbon fuels absorb atmospheric gases, ^nd the solubility of oxygen is greater than that of nitrogen, the release of o:^gen-rich 'air' at e.ltitude displaces the explosion boundaries in fig. 1 through about 10°C, The ingress of atmospheric air through the tank vents when diving into zones of higher pressure also provides explosive mixtures at temperatures higher than the noxmal boundary values. This follows as the fuel vapour, normally too rich to b u m , diffuses relatively slowly into the incoming air, end

creates near the vent a region of mixture whose strength gradates into the weak explosive range. The above three effects are illustrated for aviation kerosine in fig. 2. Y/hen atmospheric air is used as the pressurising fluid to prevent fuel boiling at altitude, the lovrered effective altitude vd.thin the tank brings the operating conditions dovm towards the explosive range during operation on the high temperature side of the boundary, (ref, 6 ) ,

It is evident, tlxerefore, that the many factors involved make it difficult to find completely safe operating conditions for any aviation fuel, and effective and continuous protection of some kind is essential when flying under conditions likely to lead to tank explosion.

1.3. CCT.ibustion Products as a Possible Inerting Fluid

It is not necessrry to remove completely the oxygen from air in order to inhibit combustion, and rosultc from ref. 7 show that the oxygen content of air need be reduced only to 12,1 per cent by vol-umc to give inertness (fig, 3 ) . This figure also

shows that inert exhaust gases (85 per cent Np and 15 per cent COp) from an automobile engine are more effective than pure nitrogen in suppressing combustion, the m.inimum oxygen content being

13*4 per cent. However, as free nitrogen is likely to be present in addition to that contained in the combustion products, the 12,1 per cent limit should be used as the safety threshold, The first requirement for any inerting gas, therefore, is that its oxygen content should be apjpreciably less than 12,1 per cent so that the oxygen content of the 'air' portion of the jxirged

-6-mixtiire does not exceed this figure, \rhen allcvring for air release,

tanlc breathing, and so on. An arbitrary maximum of 6 per cent has been assir.-ied in the tests a^t Cranfield,

The theoretical concentration curves for dry exhaust gases shov/ the oxygen content to be v;ell below this limit even vd-th mixtures considerably wealcer then stoichiometric (fig, 4)» It appears practicable, therefore, to produce engine exhaust gases of a nature suitable for tank purging. As the aim in the design of gas turbine combustors is to binm tlie fuel almost con-pletely \?ithin the prinaary zone, a continuous source of these gases immediately becomes apparent. Accordingly, tests v/ere carried out with a single gas turbine combustor in order to determine the nature and rate of flow of prixiary zone samples, Secondly, checks were made on the effectiveness of various auto-matic mixture control devices, in order to detenaine the most promising system for application to a combustor vrorking independ-ently from the m.ain engine caibustion system,

1.4. Flow Requirements for Purged Tanks

In a climbing aircraft, the purge gas flow requirements of a fuel tank are dictated by the release of oxygen-rich air from the fuel. Requirements for equilibrium conditions can be cal-culated from a knowledge of the rate of air release and the maximura permissible oxygen content, althougli a higher purge rate would be desirable in practice in order to cover the sudden releases of air due to supersaturation,

Of the data available on the saturation content of air in hydrocarbon fuels, the expressions from ref, 8 are convenient, giving rates of air release for a given volume of fuel at a given rate of climb. These expressions have been used in Appendix A to determine the flov/ rates of ca.ibustion products required to render inert the free space above 100 gallons of aviation kerosine, climbing at 5,000 ft ./min. The cor±)Ustion products are assumed to contain 6 per cent by volur.ae of oxygen, and the reqiiiremcnts are shovm in fig. 5A for diluting the released air to 12,1 per cent (the safety threshold), and to 10 per cent. The nitrogen

threshold curve is included for caiparison, Tliese gas flo-./s are likely to be more than adequate to occupy the space left by the liquid fuel consumed during the climb,

In the descent case, a tanlc free-spa.ce of 100 gallons is taken as a basis and, as kerosine exerts a very low vapour pressure, particiilerly at the lev/ temperatures at altitude, the

space can be assumed to contain air only, vrith its oxygen content

lii:iited to 12,1 per cent due to previous purging. .7ith increasing air density on descent, this air is ccmpressod into a sr.ialler

volume, the volume change being occupied by both purge gas and inhaled air. For the threshold condition, the proportion of

these trro incoming fluids must be such that their total oxygen content also equals 12.1 per cent. The purge gas and nitrogen requirement curves for descent are shown in fig. 5B.

2, Tapping the Main Engine

The ti70 major problems presenting themselves vrhen con-sidering sampling from the primary zone of combustion chambers are the high temperature (l600°C,) involved, end the possible changes in ccmbustion pattern occasioned by variations in the many controlling factors. Even -Vihen a suitable cooled probe is developed and located to sample gases of acceptable oxygen con-tent at a given condition, the need is recognised for exhaustive tests at other conditions to ensure that the maximum permissible oxygen content is still not exceeded. In addition, the presence of the probe must not in itself seriously eiffect the combustion process or the main gas flo^.v. Further, the sampled gases must be sufficiently cooled, and the condensed v/ater disposed of.

The equipment available for these tests consisted of a Dervvent I combustion chai-iber, which was considered to be reason-ably similar in design to current chexibers, together vri.th

associated air and fuel supply systems, igniter plug and booster coil, and an Orsat apparatus for the absorption of oxygen from the samples, (fig.6)(ref,9). Air was delivered to the test rig fran a Bristol Hercules type supercharger, the air temperature at the chamber inlet being about 40°C, The chamber discharged directly to atmosphere, so that no control of chamber pressvire was availeble. Aviation kerosine to specification D.Eng.R.D.

2482 v/as used throughout the tests.

In order to reproduce an actual inrning condition, it was necessary to select one at which the combustion chamber pressure equals the sea-level atmospheric value. The maximum altitude at v/hicn this condition holds was estimated to be 38,000 ft, with an aircraft speed of 500 m.p.h,, a chamber mass flov/ of about 1,5 lb/sec, and an air/fuel ratio slightly richer than 6o/l , In order to study the effects of changing both air/ fuel ratio e^t constant mass flo\/, and mass flow at constent air/ fuel ratio, the follov/ing test conditions were

selected.-(A) M r mass f lev = 1,5 lb/sec, air/fuel ratio = 6o/l

(B) ' ' ' = 1 . 5 lb/sec, t i l ^ 50/1

(C) ' ' • = 1 . 0 lb/sec, t < ' = 60/1 (D) ' ' ' = 1 . 0 lb/sec, » • t = 90/1

The weakest mixture used in this type of chamber Tender steady running conditions in practice is 90/I, and the same

-8-burner pressure v/as used for tests B end C. Test D was considered to be severe, as the low fuel pressure involved could lead to poor atcoisation, unbumt fuel, and a high oxygen content of the primary gases. As the oxygen absorbing agent (potassiur.i hydroxide) in the Orsat apparatus also absorbs carbon dioxide, it was necessary to analyse the samples first for carbon dioxide and then for oxygen. ''iflien likely sources of inert gas were discovered, the gas samples were also analysed for carbon monoxide. In order to save tine, it was decided to record results to the nearest 0.5 per cent of the total sample, and to the nearest 0.1 per cent when the result T/as less than 1 per cent.

2,1, Exploratory Sampling with Water-Cooled Probe

In order to derive an overall picture of the gas concen-tration vidthin the chamber, a v/ater-cooled probe was designed for

insertion from the chaLiber cjutlet. The probe consisted of tv-io

concentric stainless-steel tubes, the inner sanipling tube of l/8in, I/D, and the outer jacketing tub* of -g-in, O / D , ilains water was fed into the rear end of the jacketing tube and i/as allo\red to spray out througla two 2m.m. I/D pipes situated near the sampling end, the pipes being carried rearwards to discharge about 6in. dovmstream of the sampling orifice. This water, v/hich served to cool the probe mounting bracket, was therefore unlikely to affect the combustion process of the sampled gases. The mounting

bracket was located 5in, to the rear of the chcmber outlet, and adjustment was possible for sampling from any point \.dthin the central horizontal plane of the chamber, subject to the Id-iits provided by the flame tube v/alls,

A typical result from this test series is shcn in fig,7, and the most promising region for further investigation is seen to lie on the chamber axis just upstream of the secondary air dilution holes,

2,2, Sampling vdth Uncooled Probe

Having located the regions of low oxygen content, it \Tas

decided to make more thorough investigations using a probe more nearly resembling the type visualised for a permanent installation, A probe \7as designed for insertion through the side of the chamber, in a radial direction norraal to the chai.iber axis in order to

minimise the length exposed to the hot gases. As violent cooling of the sample might resvilt in chilling the caubustion reactions, and a higher oxygen content, the -g-in, O/D stainless steel probe Tra.3 lincooled, the gases being passed through a cooling coil before reaching the Orsat, The probe was marked vdth a scale at -g-in, intervals to shov/ the penetration into the chamber, and three sampling holes were drilled into the chamber at axial internrals of 2,l/4in, The sampling holes in both the air casing and flame

tube were blanked off -v/hen not in use,

The results obtained are shov.-n in fig. 8 and the

con-clusions, which confirm those drawn fra.i the previous

water-cooled probe tests, are as

follows.-(a) At constant air mass flo\7, the law-oxygen zone decreases

in size vd-th increase in air/fuel ratio, (figs, 8A and B ) ,

(b) At constant air/fuel ratio, the lov/-oxygen zone decreases

vd.th decrease in air mass flow. This follows as the reduced

tem-perattire level lowers the combustion efficiency, (figs, 8A and C ) ,

(c) The fuel burner pressiore directly controls the vddth and

length of the low oxygen zone, and appears to be the main governing

factor, (figs, 8A end B ) ,

(d) The carbon monoxide concentration increases vdth increasing

fuel pressure end decreasing air/fuel ratio, the maximur.i

concentra-tions being located near the burner. As a check on the combustion

danger of this carbon monoxide v/hen in the fuel tanlc, fig, 9 v/as

studied and replotted as fig, SA» This indicates that the

in-flammable region requires gas concentrations v^hich are unlikely to

be encountered in samples teken from the test chexiber,

(e) The beneficial reduction in oxygen content obtained by

not cooling the probe is negligible. Although the life of the

uncooled probe v/as sufficient for the tests described, the

metal-lurgical problems would be severe, and probe cooling is recanmended

in a practical installation.

2.3. Sampling in Combustion Chaj-jber Bone Head

In viev/- of the temperature difficulties anticipated in

connection vdth a permenently installed gas pick-up in the chamber

iixiedieLtely forward of the secondary air holes, it was decided to

investigate the possibility of obtaining suiteble gas from the dome

head of the chaLiber adje-cent to the sprey nozzle. It v/as realised,

hov/ever, the.t considerable quantities of carbon monoxide and u n b u m t

fuel might be picked up, and that the presence of the probe might

adversely ccffect the fuel spray pattern. Pig. 10 shovra a typical

result T/nich indicates that gases of 6 per cent maxiinum oxygen

concentration could be obtained vd.th a probe penetration of lin, as

against about 2-|-in. in the previous side-probe tests, at a

tempera-ture not exceeding 520°C, (see next section). However, the carbon

monoxide concentration \7as in the region of 9 per cent and

apprec-iable quantities of u n b u m t fuel vrere found to eiccumulate in the

liquid trap included in the gas sampling system. This method of

approach was therefore abandoned,

-10-2,4. Temperature traverses in the Low Oxygen Zones

In order to evaluate the magnitude of the probe cooling problem, temperature traverses were made in the more praiiising zones of low oxygen content. As temperatures up to 1700°C v/ere to be expected, a thermocouple of platinum/platintim 13 per cent rhodium v/as used, capable of satisfactory operation up to 1770°C, The thermocouple was mounted in a stainless steel tube of 5/32in, I/D and l/4in, O / D , and v;as shielded from radiation effects, About 95 per cent of the dynamic temperature rise v/as expected to be indicated, but this effect v/as found to be very small in the present tests. Compensating leads were used to connect to a

cold junction and a Negretti and Zambra Null Reading Potentiometer, The operating conditions used were similar to those in the sampling tests. Any errors in reading caused by evaporative cooling or catalysis vrore no doubt swamped by the difference bet\7een the test air inlet temperature of 40'^C and the calculated engine tempera.ture, at 38,000 ft,, of 118°C, As an approximation, this difference could be e.dded to the measured temperatures, so that the tempera-ture peak, measiored as 1475°C, would probably be 1553°C in practice,

The typical result in fig, 11 shov/s the temperature tribution patterns to bear a close resemblance to the oxygen dis-tribution patterns, the maximum temperatures occurring in the low oxygen zones. It is also to be noted that the probe should be located as far downstream (i,e, as near the secondary holes) as possible. As the v/idth of the high temperature zone increased v/ith fuel pressure, and as the probe v/ould have to penetrate to the chamber axis for sampling at low fuel pressure, the greater part of the probe length would be subjected to high temperature at high fuel pressures,

2,5. The Practical Fuel-Cooled Probe

As many heat-resisting materials suffer frc3m thermal shock, it v/as decided to revert to liquid cooling for the prac-tical probe, and to use as coolant the fuel destined for the

chamber, rather than call for special V7ater or air cooling systems, The design of the stainless steel probe is shovTn in fig, 12, and provision v/as made to meas\ire the flov; of the fiuel and its tem-peratiire on entering and leaving the probe. The probe was located in the rear (dovnstreara) of the three sampling positions provided, vvith the sampling orifice at the chamber axis. The temperature rise of the cooling fuel is as shov/n in fig, 13, and the flov/ v/as found experimentally to be suitably turbulent above a rate of about 1 ft,^/hour. This was considered to be prefer-able to laminar flov/, vdiere the formation of 'dead' zones might lead to hot spots» A light straw discolouration on the upstream outer surface of the probe shov/ed that the skin teniperature had not exceeded about 300°C,

No suction pump v/as available to simulate tank conditions at altitude, and the sampled gases discharged directly to atmosphere, By using the expressions for fluid flo\7 given in ref, 10, a spot calculation v/as me.de for the fla;/ of inert gases through the 5/32in, bore probe and gas delivery system serving a freely-vented tank in

an aircraft operating at the standard 38,000 ft, full pov/er condition, A pressure drop of 2 p,s,i, V7as assumed in the system, and the

successive approxixiations to the gas velocity gave a gas delivery

rate of about 7 ft^-'/Trdn., In a given system, the gas deliveiy rate

vn.ll vary vdth the cross-sectional area of the probe and the pi'essure difference across the system. As the pressure in the combustion chesLiber falls at a greater rate than the.t in the fuel tank, the effect of increasing altitude is a reduction in the gas delivery rate. As the maxixTum purge requirement occurs at the mex:ii;ium operating altitude, this becaaes the most critical condition to be considered in the design,

2,6, Sar.Tpling \7ith the Fuel-Cooled Probe

As a final practical checi: on the suitability of the design and location of the fuel-cooled probe, gas samples v/ere taken and analysed using various probe penetrations. The results confirmed those obtained from previous sarapling tests in the same zones, and shovred meocijiuH:! concentrations of 6 per cent oxj^gen and 5 per cent carbon monoxide to be easily obtainable. The optimuti location of the sanipling orifice v/as found to be vdthin l/4in, of the chamber axis. By violently cooling the sampled gases vdth liquid oxygen, the v/ater and imburnt fuel were condensed and measured as 0,062 and 0,0057 lb, respectively per lb, of sampled gas. Removal of these fluids from the purge ge^ v/ould be essential in practice. General conclusions regarding the method of prir.iary zone sampling for the supply of inert purge gases are given at the end of the report, and notes on a proposed installation are included in Appendix B,

3, The Separate Inert Gas Producer

For large aircraft vdth large tanlc capacity, a separate combustor may be more suitable, and could be accoanodated more easily than in a smaller aircraft. The advantages of such a

scheme are the greater flexibility of operation, the closer control over the oxygen content, and the utilisation of the heat extracted from the inert gases for general heating purposes. In this

latter connection, it is interesting to note that a current type of space heater consuming 4 gallons of fuel per hour would produce stoichiauetric inert gases at the rate of about 100 ft,Vniin,

If the main engine compressors are used for the air supply, the separate coi-ibustor will suffer from the same

disad-TECHNISCHE HOGESCHOOL

VLIEGTUiGEOUWKUNDE Kanaalstraat 10 - DELFT

-12-vantage as the probe system, in that the inert gas delivery •;d.ll depend upon the engine operating conditions, and insufficient gas may be available in the dive vdth engine idling. The problem

could be eased, at the expense of added cOT.iplicatlon, by fitting

the coribustor with its ovm compressor, vdiich may be driven from

the main engine through a variable-speed gearbox, or from the ccci-bustor itself,

The autoiiatic control of air/fuel ratio vdthin the com-bustor is best effected by sensing the output, rather than the input, conditions in order to eliminate any effects of increased oxygen content due to combustion inefficiency. Pig« 4 shov/s that an operating air/fuel ratio range from about I3/I and 21/I is permissible, although closer control may be desirable in practice, Various properties of the combustion products vrere examined in order to determine the most convenient property for the indication of air/fuel ratio and, through a suitable relay device, for the necessary control of fuel and/or air supplies (ref, 11),

3,1, Thermal Conductivity

The thermal conductivity values of the individual caa-bustion products are of the same order, except that of carbon dioxide, which is lov/er, and of hydrogen, vvhich is very much higher. By means of Salamon's formula, involving the partial volumes of the individual products, the mean thermal conductivity values can be found for the varying concentrations shov/n in fig,4, and are plotted in fig, 14. This curve is for the exhaust products from a gasoline, but curves for other hydrocarbon fuels do not vary appreciably,

The form of the curve in fig. 14 shc(\7s that an air/fuel ratio indicator v/orlcing on the thermal conductivity principle v/ould reverse its reading on passing tlirough the stoichiometric mixture, and that its sensitivity v/ould be highest on the rich side. The Cambridge Mixture Indicator, as used in certain piston-engined air-craft, eiiTploys this principle, and a continuous measui-ement of thermal conductivity is made fra.! the cooling effect on a heated wire contained in a i.Tieatstone bridge circuit. As indicated above,the use of the instmament is restricted to mixtures richer than stoichiometric, but a rich working range is penuissible from the inert gas point of viev/. If the output signal of such an instrument can be i:iade to operate e. relay control device, it is clear that a method of automatically controlling the air/fuel ratio values of rich mixtures will be available,

In order to test the effectiveness of a thermal conduct-ivity control system, a combustor of some type is necessary, arranged so that the exhaust gases can be continuously sampled, and the fuel input to a given air flow mechanically controlled, The most convenient choice of combustor v/as the Fedden single-cylinder spark-ignition sleeve-valve engine which, although

buming gasoline type fuels oiiLy, V7as already equipped with an exhaust sampling system and a fuel-injection pump controllable by a rack. The existing instrumente.tion included, in addition, means of accurately measiiring air and fuel mass flows, and an Orsat exhaust gas analyser,

The electrical output of the Cambridge thermal conduct-ivity instrument v/as first deteri:iined over the rich mixtinre range (fig, 15), the air/fuel ratio values being obtained from indiv-idual readings of air and fuel mass flew. Gas analysis results were found to bear out quite well the theoretical values shov/n in fig, 4» The presence of oxygen on the rich side of stoich-iometric is usual in practice, due to dissociation and incomplete mixing, and is of importance in this connection,

The controlling system designed to operate on the out-put signal from the Cambridge instrument consisted of an adapted German electric-furnace temperature controller feeding to a 24 volt D,C, compo\ind-v/ound miniature motor (fig. I6), The furnace controller had been designed to receive the output from a thermo-couple, in C3rü.er to switch on or off the furnace input pov/er, and was unsuitable for aeronauticel use, but it v/as readily available

and lent itself to experimental purposes. The miniature motor v/as geared dov/n in a I5OO/I ratio to give a linear speed range

for the pump rack from 0.013 to 0,050 inches per minute. In general, the minimum speed i/a.s desirable in order to minimise hunting. The direction of rotation of the motor was determined by a tvdn gang relay, controlled by the output from the Cambridge

instrument,

Results fran a typical test run are shown in fig, 17 v/here it is seen that control has been effective to \7ithin + 0,15 units of air/fuel ratio, and that equilibriiJiTi was again reached 3«5 minutes after making a manual adjustment of the con-troller from 11.23/1 to 10.63/1.

3.2. Combustion Temperature

T/hen a fuel-air mixture is burnt in a chaniber, the temperature resulting from the canbustion is a function mainly of the heat released by the burning mixture, and the heat ebsorbed by the combustion products. Ydth no losses through the chcmber v/alls, these t\.7o quantities of heat are equal, and the temperature may be determined from the quantities of products produced and their individual values of specific heat. If canbustion is con-sidered as an irreversible process, the maximir.i teriperature vdll occur at the stoichiometric mixture, where the absence of excess air or fuel gives a maximum heat release, end enables the heat to be absorbed by the minixium number of products to the maximur.i extent,

-14-Hov/ever, above a temperature level of about 2,000 C, the ccmbustion process begins to reverse, and the products dissociate into cerbon monoxide and hydrogen. Less released heat is thus available for heating the products, and the resulting combustion temperature is lov/er. Once the temperature drops belov/ the 2,000°C level, recombination of the dissociated products occurs, and the combustion may be subsequently caipleted. The maximum temperature attained during the complete process, however, v/ill be less than that attainable had dissociation not occurred, The dissociation curve of combustion temperature versus air/fuel ratio thus lies belov/ that obtained v/hen ignoring dissociation and, because of differences in gradient of the latter, the pealc of the dissociation curve is displaced towards the rich side of stoichiaietric. Fig, 18 has been obtained fravi calculations for

an air-benzene mixtiire burnt in the cylinder of a spark-ignition

engine, and shov/s the effects of dissocie.tion on the temperature-mixture curve,

It is evident from this figure that an appreciable tem-perature gradient exists on both sides of the peek, so that a

sensitive method of mixture control appears possible. Use he.s been made of this temperature variation in piston engine v/ork for the development of sparking plugs and for investigations into mixture distribution. In the Rabezzana and Kalmar method of test, one sparking plug in each cylinder has a then.iocouple fitted into a drilling in the central electrode, in oirder to measure the

electrode temperature at a point l/8in, from the tip. On running the engine, the thermocouples indicate the iiean temperature level in each cylinder, and these temperatixres are sufficiently sensitive to the overall engine mixture setting for the determination of individual cylinder mixture strength, (ref. 12),

A sparking plug fitted iTith a chromel-alumel thermo-couple v/as used in the Fedden single-cylinder engine, and the resulting temperature-mixture curve in sho-/m in fig, 19. The temperature range of nearly 300°C infers a variation in thermo-couple output of some 12 mV, which is sufficient to operate a controlling mechanism. This tjrpe of instrument v/ould be more

satisfactory for controlling on the v/eak side of stoichiometric where the temperature gradient is steeper, and where the conditions

are sufficiently removed from the temperatiire peak to avoid reversed action,

The Car-ibridge instrument used in the control circuit previously described V7as replaced by the thermocouple sparking plug, and similar test runs v/ere carried out. The test results

given in fig. 20 sha./ that control v/as maintained to vdthin +_ 0,3 units of air/fuel ratio during the test, the lerge number of motor reversals indicating the high degree of sensitivity,

3.3. Exhaust Gc.s Density

The curve of exhaust gas density, as in fig, 21 for a 100/130 grade aviation gasoline, shows a marked variation over the rich side of stoichiometric and very little change on the v/eak. Theoreticdly, therefore, exliaust density offers a .means of air/ fuel ratio measiare.ment and control, in a similar menner to thermal conductivity. Hov/ever, measurement of gas density is at present only possible by means of laboratory type instruments, insuff-iciently robust for airborne installation, and no further experi-mental v/ork in this direction v/as undertaken.

3.4. Oxygen Parairiagnetism

For purposes of tanlc purging, tlie most important con-stituent of the caibustion products is oxygen, and it happens that oxygen is the orly constituent exhibiting parar.iagnetism. The remaining constituents ere dLamagnetic, i.e. magnetic flux passes more easily through a vacuiu.i than through these gaseous constituents. Fig, 22 sho\/3 the overall magnetic properties of the exliaust gases fra.i IOO/13O grade aviation gasoline, indicating a large gradient over the \/ealc range. From the purging aspect, therefore, a control system operating on the paramagnetic principle VTould appear to be ideal, and experimental v/ork v/as initiated on

tv70 oxygen-sens ing devices, emplojdng this principle in different V7ays, Although neither instrument v/as fully successful, a brief

description is given of the lines of approach,

In the first, magnetic, method, a constant magnetic field was arranged so that the oxygen attracted to it would pass over a heeited spiral in a T.heatstone bridge circuit, the cooling effect giving a measure of the oxygen flov/. The instrument embodied a circulatory passage for the gases, and thus differed from tlie.t described in ref, 13, The instrument v/as calibrated using known mixtures of nitrogen and oxygen. Apart from the instability caused through heating effects of the electro-magnet used, the response v/as found to be linear only dov/n to 30 per cent oxygen, beyond v/hlch great inaccuracy arose. As oxygen contents no higlier than about 6 per cent are required, this type of

instruj-ient v/e.s abandoned,

In the second, magnetic-capacitive, method, a magnetic circuit was constructed incorporating tv/o gas gaps. The exhaust gases flov/ed through the first gap, the amount of magnetic flux passing across the gap depending upon the oxygen content, and

con-trolling the attraction exerted eucross the poles of the second, air, gap. This second gap was designed in the form of a vene-tjrpe capacitor fitted on the end of a cantilever vd.thin the magnetic circuit. The measurement of capacitance change in the air gap should tlius provide a measia^e of the oxygen content of the exliaust gases. Due mainly to constant 'drift' of the instrui-ient,

-16-and to the excessive ai:iplification necessary vdth an instrument of this type, this aspect of the investigations v/as also

abandoned,

4, Conclusions

The general conclusions that can be drav/n frcn the con-siderations of inert gas supply, and of tank reLjuirements, are as

follov/s,-4,1 • Primary Zone Sampling

(a) Inert gases containing a maximum of 6 per cent oxygen by volume are obtainable frai a fixed point vdthin the primary

zone of a Der\7ent I cccibustiün chaiiiber over a v.dde range of operating conditions,

(b) For the temperatures and mass flo\/s existing in the Der\TOnt I chamber, the probe could be suitably cooled by means of the fuel subsequently burnt in the chamber,

(c) The sampled gases need to be cooled before introduction into the tanks, and the condensed v/ater and unburnt fuel to be removed. The gas cooling could be effected by means of the fuel destined for one of the other combustion chambers,

(d) For the rates of climb and descent considered, the required purge gas flow rates are of the same order, and the curves of variation v/ith altitude are of similar form. The occurrence of the meocimum requirement at altitude, v/here the r-^inimut-l delivery is ei.Vu.ilable, confirms that the system should be designed for both climbing and descending at the maximur:

oper-ational altitude. In fact, the most critical condition vdll be the dive at liigh altitude vdth the engine idling, when the purge requirement is high and the gas deliveiy at its minimum,

4,2, Separate Inert Gas Producer

(a) Autcns.tic control to v/ithin + 0,3 units of air/fuel ratio is obtainable over the rich range by means of a therr.ial conduct-ivity type of control gear,

(b) ilutcEiatic control to sii.iilar accuracy is obtainable over the wealc range by means of a combustion tenperature type

of control gear. This method is vrdhably more promising,

although high temperature problems may be incurred in the design of a suitable thermocouple,

(o) Both the above methods are independent of exhaust gas pressure, temperature and mass flow,

(d) The use of exhaust gas density, or of oxygen volume content, is not recommended, for the reasons stated in the text.

-18-REPEREICES 1. Hill, J.P, 2 , Burgess and T.Tieeler 3 , T/hite

4.

5.

6 , Goodger, E,H, 7 , J o n e s , G,Y/, and G i l l i l a n d , ¥ , H . 8 . D e r r y , L,D, and o t h e r s 9 , Cadman, P , 1 0 , S p i e r s , H,M, ( e d . ) 1 1 , Ivfurchie, I . T . A , 1 2 , Rabezzana, H, and Kalmar, S , Tiny i n e r t gas g e n e r a t o r d e v e l o p e d f o r f u e l tanlc s a f e t y , J o u r n a l of t h e S o c i e t y of Automotive E n g i n e e r s , V o l , 6 0 , September 1952, p p , 4 2 - 5 1 . J o u r n a l of t h e Chemical S o c i e t y , V o l , 1 0 5 , 1914, p p , 2 5 9 1 - 6 , J o u r n a l of t h e Chemical S o c i e t y , V o l , 1 2 5 , 1924, p , 2 3 8 7 , C , F , R . Handbook, 1944 e d i t i o n . Coord-i n a t Coord-i n g F u e l R e s e a r c h Q^rxCoord-iCoord-ittee, Nev/ York, C,R,C, Handbook, 1946 e d i t i o n . Coord-i n a t Coord-i n g Research C o u n c Coord-i l , I n c , Nev/ York, E x p l o s i v e l i m i t s i n a i r c r a i ' t f u e l t a n k s , (To b e p u b l i s h e d I n I t e t r o l e u m , e a r l y 1955) U . S . Bureau of Ivïines R e p t s , I n v e s t ,

No. 3 8 7 1 , 1946.

Vapour and a i r r e l e a s e from a v i a t i o n f u e l s .

Journal of the Institute of Petroleuii, Vol. 38, 1952,

Combustion products as a possible source of inert gas for protecting the fuel tanlcs of military gas turhino propelled aircraft

aigainst fire risks arising fran eneijy

action,

Unpublished College of Aeronautics Thesis 1953.

Technical Data on Fuel,

British National Coniittee of Tforld Bav/er Conference, London, 1950,

An e::amine.tion of methods of automatic control of air/fuel ratio of combustors designed for fuel tank conditioning, Unpublished College of Aeronautics Thesis, 1954.

lïixture distribution studied by spark plug temperature,

References, contd.

13» Richardson, R.D, Continuous determination of oxygen concentration based on magnetic prop-erties of gases.

Trans. A.S,M,E., Vol, 70, 1948, pp,211-6.

APPENDIX A

Calculation of Purge Gas Flov/ Requirements 1. CLDIB

From reference 8,

X = 129 R k ^22*307111) > ^''^^^ ^ «Cl 1,000 metres, end

X = 129 R k ^^'^78) ^ ^^Yien H.>-11,000 metres, v/here

10^

X = rate of air release in ftvmin, from lOO/lmp, gallons at 60°F and at the pressure appropriate to the

altitude H in metres,

R = rate of climb (= 5OOO ft./min. = 1524 metres/min, in this case)

k = solubility ""j = vol. percentage gas dissolved in coefficient /' . li<TiiVgas partial pressure,

J = 0.018per cent/m,m. Hg. for kerosine.

Purging vd-th Nitrogen

Total oxygen content of released air and purging gas

= 0,121

0.326 X

X + purge volume

Thus Hor^e volume _ 0.205 _ ^

m u s , ^ - 0.121 - - ^

i , e . 1 volui-ie of r e l e a s e d cdr i s required to be d i l u t e d vdth 1.7

volumes of n i t r o g e n .

-20-Purging with combustion products

_ . , J.i.r^.^r).1 0,326 X + 0.06(rUrge volume)

T o t e l oxygen c o n t e n t = 0.121 = —'-^ •::7 ?r—^ T ^'^ X + Rirge volume Thus R i r g e volume _ O.205 _ .

i J ^ s , X - 0.061 - ^ ^ ^

For improved s a f e t y , t h e oxygen c o n t e n t may be p u r g e d t o 10 p e r c e n t , t h e n

Purge volume _ 0.226 _ f- /-c • X - 0.040 ~ •2i£^

2 , DESCEM'

Purging with nitrogen

Toteil oxygen content of added gases _ n 121 - 0,21 (.inhaled air volume) ~ * ~ volume change of original air

But inhaled air volume = volume change of original air - purge gas volume

Thus, Purge gas volvime = 0,0679 ( V o l , change of o r i g i n a l a i r ) , f t ,

P u r g i n g v/ith ccmbustion p r o d u c t s

T o t a l oxygen ccjntent of added g a s e s

0,21 ( i n h a l e d a i r v o l , ) + 0 , 0 6 (piorge gas v o l , ) ~ Voliffiie change of o r i g i n a l a i r

•z

Thus, Purge gas volume = 0,095 (Vol, change of original air, ft , For 10 per cent oxygen,

APPENDIX B

Notes on a Proposed chember probe Installation A schematic layout of a complete purging system is

shown in fig, 23, and notes on the individuEil components are given below,

1. Fuel-cooled probe, and gas cooler

Results have shown sufficient heat capacity to be

available in the fuel destined for one chamber to cool the probe, and in a f\irther, equal, quantity of fuel to cool the sampled ga.s to 200°C, The probe and cooler should therefore form a single unit incorporating the two separate fuel flov/s. The temperature of the cooling fuel may be lev/ at altitude, and the gas cooler should thus be of the parallel, rather than the contraflow, type in order to avoid condensa.tion and freezing in the outlet gas, The high pressure difference existing between the fuel and gas passages suggests a tube design for adequate strength. Calcula-tions show that a 3/8in, l/D gas tube v/ould require a developed length of about 11 ft, to bring the gas temperatxore dov/n to 200°C, 2. On-off Cock

This has been introduced so that the purging system may be turned off if, at idling speed, the purge gas becomes

insuff-iciently inert, or the fuel flov/ unable to cool the probe. The cock could be controlled by either a direct interlinlc with the throttle, or a servo system operated by fuel pressure.

3. Condenser

In order to prevent the accimiulation of condensed water in the fuel tanks, the condenser temperature must be r^t least e.s low as that in the tanks. This suggests as cooling medium the fuel feeding from the tanks to the engine. The condenser must be able to dispose of the condensate, or to retain it vdthout

impeding the gas flow. The condenser should be suitably located in relation to the engine heat to prevent the formation of ice. Lagging and/or filtration may also be necessary.

4» Control VeJlve, and Fuel Tank Vents

Many alternative vcdve and vent arrangements are possible,

but a valve of seme type is essential to control the flov/ rate of the purge gas, to prevent return flov/, and to maintain the required degree of pressurisa^tion v/ithin the tanks. Any reduction in gas pressure effected by this valve will also assist in drying the flov/

of piirge gas. In e^ddition, the vs2.ve may be specially designed

to peraiit maximum purging of the oxygen-rich air during the climb, and, in a sustained flight at constant altitude, to maintain the

-22-tank pressure with purge gas at a slightly lov/er pressure level than the setting of the tank vent, T M s would obviate the

otherwise continuous purge flovr, v/hich is unnecessary at constant

altitude. The tank vents, v/hich would probably be of conventional design, should be loca^ted as far as possible from the pvurge gas inlet to obtain a maximum purging effect throughout the tank free space,

Kanaalstraat 10 - DELFT

7qooo

lOtOOO

- 6 0 - 4 0 - 2 0 o FUEL TEMPERATUBE t

ALTITUDE TEMPERATURE BOUNDARIES FIG. I

VOLUME 'o OXYGEN IN ATMOSPHERE

6 0 - 4 0 - 2 0 O 2 0 4 0 6 0 FUEL TEMPERATURE ° C

ADDITIONAL DANGER ZONES FOR KEROSINE FIG. 2

FUEL-AVIATION CASOUNE lOQ/IX) GRADE.C.H,. (NITROGEN NOT SHOWN)

67-4 7 9 0 6 0 8 0 VOLUME % NITROGEN IN " A T M O S P H E R E ' = J p ^ - ^

MINIMUM OXYGEN COMBUSTION LIMIT FOR GASOLINE

FIG. 3

CARBON DIOXDE

THEORETICAL EXHAUST GAS CURVES FOR 100/130 AVIATION GASOLINE

PURGE REQUIRMENTS FOR lOO I.G. FUEL DURING CLIMB AT 5 0 0 0 FT/MIN.

REQUIREMENT OF 6 % O , COMBUSTION PRODUCTS FOR D^oOt CONCENTRATK3N IN TANKS

REQUIREMENT OF 6%Oi COMBUSTION PRODUCTS FOR l3-l%Ot CONCENTRATION IN TANKS

(THRESHOLD SAFETY)

REQUIREMENT OF NITROGEN FOR 12 I ' A J O I CONCENTRATION IN TANKS ^THRESHOLD SAFETY) 30 4 0 SO ALTITUDE ipOO F T 6 0 FIG 5a.

PURGE REQUIREMENTTS FOR lOO IG FREE SPACE DURING DESCENT AT 1 0 , 0 0 0 FT/MIN

REQUIREMENT Of 6 % O j COMBUSTION PRODUCTS FOR I O % 0 t CONCENTRATKJN IN TANKS

REQUIREMENT OF 6%0» COMBUSTON PRODUCTS FOR 12 • I % O t CONCENTRATION IN TANKS ( T H R E S H O L D SAFETY) REQUIREMENT OF NITROGEN FOR ? l % 0 » CONCENTRATION IN TANKS JHRESHOLD SAFETY)

2 0 3 0 4 0 ALTITUDE I P O O F T ,

5 0 60

FLOW REQUIREMENTS OF PURGE GAS

MANOMETERS

BURNER PRESSURE CONTROL

<>

COMPRESSOR ELECTRIC MOTOR

URNER PRESSURE I 1 fIGNITER I 1 r PLUG BOOSTER COIL AIR TEMPERSrURE 0 0 P U M P PRESSURE < >PUMP PRESSURE ^ CONTROL • - ' BOOSTER COIL SWITCH

KR COCK

LAYOUT OF TEST RIG FUEL PUMP AND

ELECTRIC MOTOR MEASURlNd TANK Q FILTER ROTAMETER 25 >

ANNULUS AREA OF hSóD* 3 0 HOLES OF OIODIA. , 4 0 HOLES OF 0375DIA AREA-4-4 D' - 6 0 HOLES OF OIODIA AREA-0-47G' 8 HOLES OF O-50IA AREA-I-57D' 24 HOLES OF OSDIA A R E A - 1 2 - I G ' AR CASING COMPRESSOR CASING FLAME TUBE ENTRY ELBOW

HALF VEW ON FRONT OF FLAME TUBE WITH ENTRT ELBOW REMOVED.

INTERCONNECTOR SLEEVE PORT

OUTLET AREA OF I 2 - 6 G :

SWIRLER VANES

CENTRE UNE O f ^NG|NE_

HALF SECTION ON AA.

LAYOUT OF TEST RIG

^ § 0 - 4 % OXYGEN ^ ^ 4 - r f i b OXYGEN ^ ^ 8 - 1 ^ OXYGEN

ALL FKÏURES SHOWN, SHOW % BY VOLUME

t ^ Ö 2

ïa-95 20-100 25-IK) ?iHK) JO 100 » M 0 «M05 «HOO 5<M00 75-90 , 5 ^ 80-80 ^o-eo l « ^ •O^IK) 80 "W 140-50 I&5-20 65-100 7-0-90 IVi-fr5 150*55 \e5-50 n-5-^0 90-75 105-70 14-0-50 \7-0-2-5 a-5-20 -12 IH)-« 130-4-5 (50-+5

AIR MASS FLOW-IfeLB'SEC. AIR FUEL RATIO - 6 0 = 1 AIR INLET T E M P = 4 0 ° C . FUEL PRESSURE = l4OPSl.feAUG0' FUEL DELIVERY - ll-2aP.H.

EXPLORATORY TEST RESULTS WITH WATER COOLED PROBE

FIG. 7

miïcüo

J0;li0' K)

9 0ilHVlf TOrtiW aOiiooo 70; 110 0

5-5ilJO0 50J5OIO JftlJ-OhO 100 4 0 liat5-0 I00)60 80 8 0 80- 8-5 « • 70 100 40140 45140 lJl85 2sli65 50I1&O •ll-O -l-S |-5il25_. IO,IOO--70 |-5,IOO*5 60-120 60-(llO Ï0-M-5 25 160 50 140 15175 l-0,IIO65 15, 6O|l00 55 25 120 150 25H55 iofoo CO CO. 0 . 20; lOO^'^^I^I^È^^'^f^' 20;l1O50 2-5,12O*0 20,l|-0t40 2 0 , 9 0 70 20,10060 20,11050 70 95 8 0 9 0 809^0 50125 20170 55125 55 145 20170 40140 20170 40140 101-185 20ti70 4Qti40

AIR MASS F L O W - IfcLB/SEC AIR FUEL RATIO - 6 0 : | AIR INLET TEMP = 40°C.

FUEL PRESSURE = I 4 0 PS.I. (GAUGÖ.

FUEL DELIVERY = 112 aPH.

a.

AIR MASS FLOW- IfeLB/SEC, AIR FUEL RATIO - 9 0 1 AR INLET T E M P - 4 0 ° C

FUEL PRESSURE - 60PSI (GAUGO.

FUEL DELIVERY = 7 5 G.PH.

AIR MASS FLOW - I LB/SEC. AIR FUEL RATIO - 60\ AIR NLET TEMP = 40°C.

FUEL PRESSURE -60PS.I.^VUJGO.

FUEL DELIVERY =7-5G.PH.

C.

TEST RESULTS WITH UNCOOLED PROBE

8 0 64 u _J bl O

^ 1

S

UI m O- < IA t UJ - I < < 48 32 16 IRICH LIMIT

1

/

LX*-_L K^^'

/

/

~~7

/ PEAK 1 1

RATIO INERT GAS

INFLAMMABLE GAS

1

_Ui U a UJ - l O "•

i

I-X 2 -15 -20 Z UJ

LIMITS OF INFLAMMABILITY OF CARBON MONOXIDE WITH AIR WHEN MIXED WITH VARIOUS PROPORTIONS

OF NITROGEN pic 9a.

O ; SAFETY THRESHOLD

VOLUME L O , IN MIXTURE

DERIVED LIMITS OF INFLAMMABILITY FOR MIXTURES OF CARBON MONOXIDE,

OXYGEN AND NITROGEN

ALL FIGURES SHOWN, SHOW % BY VOLUME.

AIR MASS FLOW - I LB/SEC.

AIR FUEL F^TIO - ÓO-I.

AIR INLET TEMP - 40°C.

DOME HEAD TEST RESULTS

WITH UNCOOLED PROBE

FIG. IG.

OVER I 4 0 0 C.

I200-I400°C.

I 0 0 0 - I 2 0 0 C .

800-IOOO°C.

6 0 0 - 8 0 0 C.

4 0 0 - 6 0 0 ° C .

2 0 0 - 4 0 0 C.

BELOW 200°C.

/'?>'>»?^-^^y» >yy/ y / J / / / / / / / / / / ^ / r j 7 / //// 7 J-7^-7- T 7 / / / / f ^ ^ }-3r7-7-7 f / / / ^^ I ii/nm<min/f r r / r-r- r r /•

AIR MASS FLOW

AIR FUEL RATIO

AIR INLET TEMP

I \ LB/SEC.

60-I.

40°C.

TEMPERATURE TRAVERSE

/"

SECTION ON AA, WITH FULL VIEW SECTION ON PLANE OF DIVIDING OF CENTBAL TUBE WALL

[ © J ^ B B A Z E SECTION O N B B . l|l2t»/KC KMSS FLOW FUEL INLET TEMPERATURE 90.I AFR 60t1 FUEL DELIVERY C. P H.

TEMPERATURE RISE OF COOLING FUEL FIG. 13

DESIGN OF FUEL-COOLED PROBE FIG. 12 t 0 082 2 o o t e "• i bj ID O § O ) O I C ( -THERMAL CONDUCTIVITY OF EXHAUST GASES FROM 100/130 AVIATION GASOLINE FIG. 14 e 13 l i 2 0 24 2 8 AIR FUEL RATIO ( W E I G H ^

A/F RATIO ^ E I G H T ) - 4 0 BRIDGE CHARACTERISTICS OF CAMBRIDGE INSTRUMENT FIG 15. MANUAL SELECTOR OF RWnCULAR AIR-FUEL RAHO

LAYOUT OF THERMAL CONDUCTIVITY

TYPE C O N T R O L L E : ^ FIG 16. POWER SUPPLY 24 V. DC CAMBRIDGE WHEATSTONE OUTPUT

FUEL PUMP MOUNTING FUEL IN VARIABLE DELIVERY PLUNGER BRYCE FUEL PUMP I I - 6 I I - 4 1 1 2 I 13 1 1 0 O I O - 8 a. A I O - 6

TEST RESULTS USING THERMAL CONDUCTIVITY

TYPE CONTROLLER FIG 17.

12 16 2 0 24 2 8 TIME OF TEST RUN ( M I N U T E S )

COMBUSTION TEMPERATURE

Cc)

3I00| 3 0 0 0 2 9 0 0 2 e o o 2700 2600 2 5 0 0 2 4 0 0 600r 5 50 SOO 3 4 5 0 Ü It bl a Z y 4O0

§

o 350 z 3 0 0 /

y\

K

X II ta 13 14 15 AIR-FUEL RATIO WEIGHT

THEORETICAL COMBUSTION TEMPERATURE CURVES FOR

AIR BENZENE MIXTURES

FIG 18.

o o 12 14 16 18 20

AIR FUEL RATIO (WEIGHT^

EXPERIMENTAL TEMPERATURE-MIXTURE CURVES FOR 100/130 GASOLINE IN FEDDEN ENGINE

FIG 19. 12-O i i a ' ^ M - 6

i

AIR-FUE L RAT » O M AVE AIR-FUE / ^ •

SINGLE CYUNDER FEDDEN ENGINE; FUEL, AVIATION

RAGE L RATIC

y

y

/ / — " ^

GASOUNE tOO/l30 GRAOE"

\

\J

^

6 8 lO 12 14 16 18 20

TIME OF TEST (MINUTES)

TEST RESULTS USING TEMPERATURE TYPE CONTROLLER

u. 7-0 z f 6 0

A

1 WEIGHT O F O O oiBic r r / OF AIR coNDrm UNDER 3NS / / SIMILAR _/ - —

B

c i t i i )

DENSITY OF EXHAUST GASES FROM 100/130 AVIATION GASOLINE

FIG.2I

13 16 2 0 34 AIR FUEL RATIO ( W E I G H T )

VOLUME SUSCEPTBt-ITY (X I 0 ' ' A T 2 0 ° C )

MAGNETIC PROPERTIES OF EXHAUST GASES FROM 100/130 AVIATION GASOLINE

FIG. 22 ISOO I6O0 MOO 12 OO 4 0 0 t ^ N A MRAMAGNETt DIAMAQNE 0 .TIC or^ c I 0

t

y 1: 8 s s p , InU 1 lO >l 1 / / / / / /

J

/ / / / 2 0 2S AIR-FUEL RATIO 1 1 / / 7 ^ 3 0 35 WEIGHT 1 1 PRESSURE REDUCTION VALVE CONDENSER INWARD CONTINUOUS VENTING,^^VE , BLEED VENT

SCHEMATIC LAYOUT OF CHAMBER PROBE PURGING SYSTEM FIG. 23

629.13.011.525-756.6

CoA REPOBT 85

The College of Aeranautics, Cranfield.

THE HiOTBCTION OF AIECHAPT FUEL TANKS AGAINST EXPLOSION HAZAEDS Ü S D K INEHT COMBUSTION PRODUCTS, by Eric M, Goodger, Paul Cadman and Ian T.A. Murohie, Oct. 1954, 22pp. 23fig.

Within certain ranges of tanperature and pressure, the aiJ>-vapour mixture produced above the liquid fuel in aircraft fuel tanks trill sxqiport conibuBticn and will b u m explosively on the addition of the necessary energy for igjiition. Although the danger zone for each type of aviation fuel can be clearly defined under equilibrium conditions, the many transient factors involved in practice make it difficult to find completely safe operating conditions for any fuel. A war-time approach to the problem was the oontinuous jxirglng of the tank free space with nitrogen to redxice the oxygen content of the admired air below the minianm permissible for canbustion. Combustion products have been suggested as an alternative and continuously-available si^iply of inert ^ s e s . This report describes experiments made on sampling frcm the primary zone at

a Derwent I chamber, and «i the testing of various automatic control devices suitable for an Inert gas generator. The oonolusions drawn fron the results are that both systons of purge gas supply are feasible, and that the orlt-ioal condition will be the high altituie dive with engine idling, when the purge gas requironent is a maxinun, and the simply a minimum.

629.13.011.525-756.6

CoA EEPCET 85

The College of Aeronautics, Cranfield.

THE PROTECTION OF AIRCKAFT FUEL TANKS AGAINST EXPLOSION HAZAEDS USING INERT COMBUSTION PRODUCTS, by Eric M. Goodger, Paul Cadman and Ian T.A. Murohie. Oct. 1954. 22pp. 23fig.

Within certain ranges of taiperature and pressure, the air-vapour mixture produced above the liquid fuel in aircraft fuel tanks will support canbustion and will b u m explosively on the addition of the necessary energy for ignition. Although the danger zone for each tyi» of aviation fuel can be clearly defined under equilibrium conditiona, the many transient factors involved in practice make it difficult to find oanpletely safe operating conditions for any fuel. A war-time approach to the problem was the continuous purging of the tank free space with nitrogen to reduce the oxygen content of the admixed air below the minimisi permissible for canbustion. Conibustian products have been suggested as an alternative and continuously-available sx^rply of inert gases. Tliia report describes experiments made on sampling fron the primary zoie of a Derwent I chamber, and on the testing of various autonatic control devices suitable for an inert gas generator. The oonolusions drawn from the results are that both systems of purge gas supply are feasible, and that the crit-ical condition will be the high altitude dive with engine Idling, when the purge gas requirement is a maxiimin, and the sv^iply a mlnlmun.

629.13.011.525-756.6 CoA REPORT 85

The College of Aeronautics, Cranfield.

THE mOTBCTION OP AIHCKAFT FUEL TANKS AGAINST EXPLOSION HAZARDS USING nffiRT CCMBUSTION HiODUCTS, by Eric M. Goodger, Paul Cadman and Ian T.A. Murchie. Oct. 1954. 22pp. 23fi«.

Within certain ranges of temperature and pressure, the air-vapour mixture produced above the liquid fuel in aircraft fuel tanks will support combustion and will b u m explosively on the addition of the necessary energy for ignition. Although the danger zone for each type of aviation fuel can be clearly defined under equilibrium conditions, the many transient factors involved in practice make it difficult to find completely safe operating conditions for any fuel. A war-time approach to the problem was the continuous jxirglng of the tank free space with nitrogen to reduce the oxygen content of the admixed air below the minimum permissible for combustion. Combustion products have been suggested as an alternative and continuously-available supply of inert gases. This report describes experiments made on sampling fron the primary zcne of a Derwent I chamber, and on the testing of various autonatic control devices suitable for an inert gas generator. The oonolusions drawn fron the results are that both systems of pui^e gas svipply «"^ feasible, and that the crit-ical ccndition will be the high altitude dive with engine idling, when the purge gas requironent is a maximus, and the supply a mlniniun.

I CM g'

1^:

-^ -^ ^ -^ g 6 ^ A .u b •€ c o g , » , . . 5 s . ö c ï ^ p . c j a » 3 o B o s ! « H 4 - i e + > t 3

-°^h

Hulk

S^fe b d - S - P a r H O S 3 s e B « 5 w -• ^ 3-3 ' ^ ^ ? b g

' ï g ë f i j j

^ ë . ? g : 3 S o 8 > - i l ° S „ 3 8 S - Ö ' ' § o * * 2 g ë i .tl^' & 0 » 4 0 4 » J 3 . P ? 5 . 4 » 2 5 j « e H . ^ b . C . ^ ' H S H C ' d o X w b * t ' 3 £ * i . j . H u « ) _ , ö V " . ° a _ 2 ' B . S « > > i 3 v D t c q 8 4 ' H W

Bë

The College of Aeronautics, Cranfield.

THE APPLICATION OP MajO-MOTION TO INDUSTRIAL OPERATIONS, by Clifford J. Norbuiy. Doc. 1954. 49PP. 25fig.

An investigation of the field of application of Memo^Uotion study, with a view to the extension of its use into new fields by developing apparatus that could take photographs at Intervals varying fron two per second to one every four hours.

Experimental work has shown that Memo-Uotion heis the following applloatlois in which it has special advantages over other forms of study.

1. Single operator repetition work, for irtiich the term Maoro4otion has been introduced here.

2. Area studies, the study of a group of men or machines. 3. Team studies.

4. Utilisation studies. 5. Work measurement.

The main conclusion is that Memo^oticn could becane a valuable new tool of Work Study, and can be used where other forms would be impractical or uneocncmlo. It is a very versatile tool, and the addition of a Spaced Shot attachment to a cine-camera greatly extends its industrial applications. CoA REPORT 86

The College of Aeronautics, Cranfield.

THE APPLICATION OP MajO-ï,10TION TO INDUSTRIAL OPERATIONS, Norbury. Dec. 1954. 49pp. 25fig.

658.542

by Clifford J. An Investigation of the field of application of Memo^itotion study, with a view to the extensicn of its use into new fields by developing apparatus that could take photographs at intervals varying fron two per second to one every four hours.

E:q>erlmental work has shown that Memo-Motion has the following ajiplioatlans in which it has special advantages over other forms of study.

1. Single operator repetition work, for yrtiich the term Macromotion has been introduced here.

2. Area studies, the study of a group of men or machines. 3. Team studies.

4. Utilisation stvidles. 3. Work measurement.

The main conclusion is that Memo^otion could becane a valuable new tool of Work Study, and can be used where other forms would be impractical or uneccnanic. It is a very versatile tool, and the addition of a Spaced Shot attachment to a cine-camera greatly extends its industrial applicaticns.

658.542

CoA REPORT 86

The College of Aeronautics, Cranfield.

THE APPLICATION OP MIMO-I.!0TION TO INDUSTRIAL OPERATIONS, by Clifford J. Norbury. Deo. 1954. 49pp. 25fig.

An Investigation of the field of application of Memo-4Jotion study, with a view to the extensicn of its use into new fields by developing apparatus that could take photographs at intervals varying from two per second to one every four hours.

Experimental work has shown that Memo^otion has the following applicaticns in Trtiich it has special advantages over other forms of study.

1. Single operator repetition work, for which the term Maoro^otion has been Introduced here.

2. Area studies, the study of a groiq) of men or machines. 3. Team studies.

4. Utilisation studies. 5. T/ork measurement.

The main conclusion is that Memo-Motion could become a valuable new tool of Work Study, and can be used where other fonus would be impraotioal or unecononic. It is a very versatile tool, and the addition of a Spaced Shot attachment to a cine-camera greatly extends its industrial applications.