it CONSIDEPITIOHNS F-.SRD13iNG COtMUSTION OF HIG1I PRESSURE SPRAYS E' T, Vincent Professor of recha.ical Engineering University of Miehigan Prqpared for the Conference on Fuel Sprays Univercity of Yilchigai Ann Arbor, Michigan March 30-31, 1949 Sponsored by Thn i Air aterial Co.amnd, USAF and 7~i Engineering Research Institute of the lnivkersity of Michigan

2Of: CONSIDERATIONS RBEARDING TW COMBUTSTION OF HIGH PRESSURE SPRAYS It'must be admitted at once that the writer has little t;*: c:fger to this gathering in the way of new theories or exper....mental results directly connected with the art of spray foT.mation, However. due to some forty years contact with spray:uo'ir:ations and characteristics for Diesel Engines, it is believed thaia. a useful purpose would be served by presenting some of the omeperience gained in that field. It is true that the Diesel field is totally different frori that with which the majority of persons present are concensrd since that field demands an intermittant spray of high freequencyp accurately metered and controlled, particularly as g.-.~ads the beginning and end characteristics, Interest today tentwrs mainly on a continuous spray in which the starting and s-to -pping periods are of negligible magnitude and of no influence on the resulting combustion process, This difference permits impeltification of a number of design and construction problems Le.n: continuous injection is desired and is mainly associated i:.ith the effect of fuel compressibility which results in unmaitoLiS.ed fuel at the starting and stopping periods together.ith the production of secondary sprays so harmful to fuel Let us examine some of the major requirents for the ro~t'de:rn type of jet mechanism in order to approach the problem in a logical manner, It is believed that the following represent the more important requirements:

lo Tle spray should be finely and uniformly atomised at all rates of flowo 2, The atomised fuel should be uniformly distributed throughout the mass of air when operating at full loado 3, The products delivered by the combustion chamber should be at a uniforn temperature, without stratification, at any loado 4, The combustion process should be capable of being maintained in an efficient manner in air streams of high velocityo 5. It should be possible to initiate and continue combustion at high altitudes where pressures and temperatures are low, There are many other items which go to make up the oreiall performance characteristics of an injection and corn bueiion chamber system. It will be observed that the injection aystfem and chamber have been coupled together; this is believed nec.essary at the present time since the individual characteristi-s of each affect one another to an unknown extent. The sman is true of the Diesel Engine combustion process, where thsai? are advocates of high and low pressure injection systems, quirttoeent and turbulent chambers, etc. However, if the engine perfcrmance characteristics are studied, it will be found that despite the vast differences in the basic principles employed for the injection and combustion process, the output and specific fuel consumption remains comparatively unchanged, the

advocates of high or low pressure injection, etc, turning out one HP for roughly the same amount of fuel in each case, in comnarable engines. It is the combination of injection system and combustion chamber that matters, not one of the other despite the advertisements. It is believed the same will be true in the case of the jet engine. However, at the present time, combustion chambers for various jet applications bear a closer similarity to one another than is the case in the oil engine, In ihe early days of the oil engine all chambers were very similar to one another, thus the possibility of advance with con:siderable change in principle should not be ruled outo Consider some of the above itemso First, the question of'tomisation on which considerable work has been done on engL ne sprays. The following represents the general observations: It appears necessary to divide the discussion into two parts, depending upon the type of nozzle employed in producing the spray, the plain hole nozzle, and centrifugal or whirling spray nozzles The characteristics of these two seem totally different. Taking the plain orifice, the atomisation appears to be a function of pressure of injection which, in turn, could be considered one of relative velocity between the fuel and the air into which injection occurs According to tests at the NoACoA. (Refo Report #425), the main effect of pressure is to produce uniformity of atomisation; loi( pressure srrays have very small drops present, as small as high pressure sprays, but few in number. These are Secompanied by larger droplets. The size of the largest droplet increases as the

pressure is reduced. Increase of injection pressure has the effect of reducing the large drops and leaving the small ones almost unchanged. Thus a more uniform spray is produced, but not necessarily one which can be burned at a faster rate, since there is a relation between the drop size and the distance of travel from the nossle~ The gas density in the Diesel chamber is high and the effect of resistance to motion of the droplet results in an increasing deceleration of the droplet as the size reduces. It follows that the length of travel must also reduce with drop site in quiescent air; it follows that fine atomization is not necessarily the one and only solution to a spraying problem. A large bore Diesel engine will definitely require a coarse atomisation to enable the fuel to travel from the nozzle to the outer reaches of the combustion chamber, unless some form of air motion is relied upon to carry the fuel droplet to the oxygen of the charge after it has lost its initial velocity resulting from the nressure difference of injection. Such a mixing process requires time which is usually not available, at least in high speed engineso The spraying characteristics of an injection system are of secondary importance, combustion is the primary obJective. Here some discussion of the elementary factors controlling efficient combustion can be considered as a starting point. When considering the combustion of a sprayed liquid fuel with air it is generally believed that the phenomena is a surface characteristic, the surface liquid producing a vapor

>we hi can then react with the oxygon of the air, releasing heat. It lfollows that the resulting envelope of combustion gases must b-e renmoved from the surface of the drop if combustion is to'ontimnue at a rapid rate, i.e., some relative velocity is necest. ry between the fuel particles and the air It is believed to be immaterial whether the fuel is moving faster than the air or i.c-versa, the object being to tear away the surrounding gaseous enverlope, bringing oxygen into intimate contact with the vaporizting fuelo The Diesel engine industry provides plenty of evidence to substantiate these observations, since the combustion has been e.ffected in so many different ways. l o The quiescent open combustion chamber In this, the common chamber for slow to moderate speed engines, the air during the combustion phase of the cycle has little motion but gives excellent outputo In general, good specific fuel consumptions are secured, accompanied with excellent smokeless exhaust. In this design the injection system and its spray characteristics become of dominant importance. The degree of atomisation, spray pattern, penetration, etc., must be matched accurately with the chamber shape, and thus the subject of this conference is seen in perhaps its most important phase. 2. The whirlol chamber of the single sleeve alve engine Due to Ricardo, this design is perhaps the exact

- C5 antithesis of the above open chamber engine In this case, the fuel is added in praotically a non-atomised state at one local point of the chamber, but the air, trough the action of the admission ports in the sleeve valve, can be given a tremendous whirl about the axis of the engine cylinder. Speed of whirl up to 30,000 RPM have been recorded in an engine cylinder of about 5" in diameter, which probably means a minimm tangential velocity of the air of some 450 to 500 f.p.s. It follows that, in this case, the air sweeps past a more or less unatomised column of fuel, combustion occurs on the surface of the fuel droplets, and the combustion products are oswpt away by the air stream, exposing a fresh surface of fuel to the continued action of a stream containing oxygen. The mistaken idea is often expressed that the air flow carries the fuel drops along with it, thus mixing the fuel and air together. This is very definitely not the case, since, with such a coarse saray, the large drops travel in a straight line, due to their high kinetic energy, and are deflected but little by the air motion. This is apparent by examining the piston top which shows the mark of the extreme outer fringe of the spray almost directly under the nozzle, which seems to oonfirm the above expressed scheme of combustion. The power output and specific fuel consumption achieved with this method of combustion are comparable with the

previous combustion system. Maximu efficiency is, in general, secured rwith an air velocity somewhat loss than the above 450 to 500 f.ps., i.e., about 350 f.pcs. In the oase under disoussion, the engine perormance characteristics are claimed to be praotically ifldependant of the spray haeracteristios, as regards atomisation, dispersion, etc. There are many other interomdiate types of oil engine combustion chambers, employing air velocities of the order of 300 to 350 f.p.9s or higher, in which good combustion characteristics can be secured; this Wans complete combustion in a time interval of 0.003 to 0.004 sooe Translating this to the flow combustion ohamber of the present Jet mechanism, if similar oonditions can be secured, the chamber need not be longer than about 1.5 to 2 feet, maeh shorter than present demands, which would enable considerable reduotion of space and meight to be effected. tTrue, the initial state of the air into which the fuel is injected in the oil engine is considerably different froi that of the present jet engine, being at a pressure of 0 ome 450 to 550 p.s9io and 1100 to 1200~ F, but the speeds of flight under discussion in some of the advanced thinking of today do not rule out the possibility of employing such values in the Jet engine of the future. Applying the above type of thinking to the problem at hand, i.oo, the jet engine fuel system and combustion chamber, it raises the question of whether the present day components of such injection systems and combustion

ambers are the best ombinaon it not possible c~ambers are tbest combination, la it not possible that the future will result in varying the principle in as many ways as was the case in the oil engine? What those principles may be the author cannot forecast; it remains for future developments to chart the course. There are indications that the frture may demand a considerable increase in combustion chamber velocity with reduction of external drag, etc., in the ease of the ram-Jet type of engine. Calculations made at this University, Ref UtJ 1 & 7 and E;fl 12, have shown that the overall performance of a ram jet ean be expected to improve with increase of entrance combustion chamber velocity to at least 600 f.p.s., approx., if efficient combustion can be secured simultaneously. Bringing the above observations down to the spray patterns and atoisation of the fuel, there is ample evidence that at the opposite end of the soale a well-dispersed spray from a low pressure injection system can be carried along by a high velocity air flow before combustion is complete, consequently requiring a long chamber for efficient combustion and good temperature distribution. Sprays such as those in common use at present, such as the whirling spray, using a few hundred pounds per square inch injection pressure, wuld fulfill the above conditions sinee such sprays lack penetrating power, are rather uniformly atomised but of large drop size with low

-9absolute velocity and are thus easily capable of being oarried along by an air current. Complete combustion will then depend upon the relative velocity between the oil and atro If this is low, a long length of travel is necessary to bring sufficient oxygen in contact with the fiel for complete conbustion. Looking at the alternatives, there is the ease of high pressure l:ijrcotion with greatly decreased drop size, of increased velocity of t-i;noel, which is hardly affected by any reasonable air velocity up to abyot 500 f.p.s Su eh a spray could be formed in an air stream, the or-: an passing across the stream at any desired rate, exposing the fuel to o hmximum burning rate. The length of travel from the Jet to the ti. of the spray can be made alot anything desired, to suit the air veloity involvedt. Sah sprays have usually ee foled by the use of aa^.:ul holes and high injection pressures The resulting sray,,cxared with the whirl type, has very little dispersion. It follows.hat, for complete 0overage of any area, multiple orifices would be nocessary. Good filtration would be an essential, and a suitable high presAure pump would require development. Again, the present low d.:..sity medium into which injection occurs would perhaps produce some ooplications rgarding length of spray path due to lowered resistance to notion. Altogether, these are some rather diffeiult conditions to fulfill at the present time for existing types of chambers, but tire will undoubtedly contribute to easing the oondition which, together with the advancing sklls and knowledge available to handle tbh problem, will contribute to its solution.

-10Secondly, it is possible to produce high pressure sprays of fine atomisation, coupled vwith ride dispersion, by tho use of iring ing orifices* Such jets have proved to possess sufficient.;aiBration for high density air despite their wide dispersion a wivr ll thus withstand the action of high velocity air currents, g:i.ii high rates of relative velocity between fuel and air, t9is, p J;-sibly, quick and efficient combustion. The orifice si.ze in this c:ie is not so critical, and in addition the injection pressures a.eo 10t excessive A caamrehesive spray investigation should covear all phases ofL the problem, both high and low pressures, types of nozsles, etc,, so hlat the spray pattern, atcmisation, etc,, rtill be caoable of,rsadf.otion for any desired combination. This is a very lesirable state that still remains to be aohieved. So far as the writer is aTmro, it cannot be too strongly urged that the attaiment of such a oeal means absolutely nothing so far as the final end state is coacerned. The desired end is efficient combustion in the smallest coibustion chamber, under the widest variations nossible among ope.ating conditions, Ihether this is achieved by high or low pesasure, atomised or non-atomised fuels, cje, is purely incidental. As ec result the vriter believes that a spray investigation as such rrm trs little without its associated combustion study. This has been rnoved the case many times in the past, in the oil engine industry, w\hl: large sums of money have been expended on spray inestigations, Th^ final result is still a trial and error solution for each engine,

- 11 - so ftar as efficient combustion, which is the main object, is concerned, In conclusion, it can be stated that combustion is possible in high velocity air streams in very short time lntervalB without thwe aid of flame holders, if given suitable initial condittions, spray ffor:mations, etc. That spray pattern itself is no measuwe of the r-Xsulting combustion process; similarly air velocity is no criaerion for success or failure * The overall result actiseved is a ormbination of velocity, spray pattern, chamber shape, and state o~ the air. Regarding the actual application to particular problems, it vll be necessary to solve the rather complicated combination pvrsaanted by all the above variables, if it is desired to be able to d{.sign and put on paper a combination capable of operating over tEo Tange opresented by modern flight requirements from sa-level tc possible 80,000 ft. altitude. It can be reported that in addition to its spray investigation this University is conducting some fundamental and applied res);earoh in the field of combustion for the Ram-Jet and similar mjet mechanisms. It is hoped that it will be possible to coordinate t:ese two programs in such a manner that answers to some of the p',t)lems outlined above can be secured. Again it is emphasized that efficient omabustion is the g;al, Spray characteristics is one step along the path, but tIc.ar are many other steps to be taken before reaching the desired objective.