A Theoretical Investigation of the Possibilities of Internal Cooling of Aircraft Engines by Water Injection to the Cylinder Citation Losey, Robert M. (1937) A Theoretical Investigation of the Possibilities of Internal Cooling of Aircraft Engines by Water Injection to the Cylinder. Master's thesis, California Institute of Technology. doi:10.7907/3PXW-GF19. https://resolver.caltech.edu/CaltechETD:etd-04112008-132848 Abstract No abstract. Item Type: Thesis (Master's thesis) Subject Keywords: Aeronautical Engineering Degree Grantor: California Institute of Technology Division: Engineering and Applied Science Major Option: Aeronautics Thesis Availability: Public (worldwide access) Research Advisor(s): Millikan, Clark Blanchard Group: GALCIT Thesis Committee: Unknown, Unknown Defense Date: 1 January 1937 Record Number: CaltechETD:etd-04112008-132848 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-04112008-132848 DOI: 10.7907/3PXW-GF19 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 1346 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 21 Apr 2008 Last Modified: 03 Oct 2019 23:07 Thesis Files Preview PDF (Losey_rm_1937.pdf) - Final Version See Usage Policy. 3MB Repository Staff Only: item control page

GUULING Ur ATRCHAFT EAGINiS by WaT INJECTION TG Tak CLbinukk, By RM, GOSEY ist Lieut. Air Corrs, Acknowledgments, The writer wishes to express appreciation for the assistance rendered br the following rereons: Pr. C.B. Milii*en, Assistant Professor of Aeronautics, 2nd Mir. 7. Curtis Rockefeller, Instructor at the Guggenheim Aeronautics Laboratory, California Institute of Technoloey, for suggections and critical reading of this papers Major EAR. Page of U.S. Army Air Coros for certain facts and experimental data; Mr. Carlos Yood of Douglas Comnany FEneineering Steff for aid and advice, end to Mrs. Madeline “ood for t: ning of the a Introduction The purpose of this vaper is to investigate, theoretically, the several possibilities of internal cocling of engines by water injection to the cylinder, and the necessary condensor system for obtaining the required amount of wat er from the exhaust gases. The radial air cooled engine is inherently more efficient in regards to weight tnan the inline type either air or lisguid cooled, In both cases it has tne advantage of a considerably lighter crank shaft , tne heaviest individual component of a power plant, plus the radiator system in tne case of the liquid cooled. With the develonment of tne larger twin rows, it appears that but little more can be Gone with tue type due to cooling difficuities. In service conditions the co:ling of this type is barely satisfactory, necessitating in generai the use of a considerabiy richer mixture than is dictated by either maximum power or econony in order tnat the incoming charge may absorb a portion of the excess heat. Thus the increese in power per unit for the radial engine depends on auxiliary cooling to aliiow an increase in rows of cylinders or an increase in com eression ratio, with, of course, the common possibility of ail engines of increased R.P.ii. The applications discussed in tnis paper are c.nsidered only in cennection with the radial tyne, alth ugh in general applicable to the inline types, The three possibilities of water injection are: To (1)Perward the end of the p.wer stroke tnus decreasing tie gas temperature during the exhaust stroke. (2)Go to a six stroke cycle, two of which are made on a mixture of air and water oarticles. (3)Injection of water with the charge of fuel. The first two methods will be discussed as to resulting cocling, and the third in cmjunction with the recuired ccndensing system, The writer hastens to advise that his experience with airplane power plants is limited to use under service conditions and the study of books and reports on the subject. In this connection the two voliwms of D.R. Pye are strongly recommended to those interested as being unique among technical books in regards to organization and presentation, The Units Fmploved Standard Cubic Foot - SCF - the quantity of gas which occupies 1 cubic foot at 1 atmospheric pressure and O° Cc. Pound Holecule - OL -— is a weight in pounds of a gas eaual to its molecular weight and occupying 455 Cu.ft. at standard conaitions. Specific heats will be expressed in foot pounds per SCF or CHU per MOL. CHU - heat reauired te raise 1 1b. of water from 14°C to 15°C, Fuel-Air Relation Employed in Calculations Since the term gasoline covers a mixture of hydrocarbons varying greatly in number and relative proportion, for the purpose of calculations the fuel in this paper is considered to be composed of 50% each by volumn of Benzene (CgHg) and Heptane (CoHig)- Their complete combustion by air is as follows, eCghg + 150. + 56.4Ny = 1200, + 60,0 + 56.2No + 1.52 x 10° CHU weights 155 2060 2216 volumns 2 714 74.4 Roh g + £205 + 62.8N., = 140. + 16H. + 82.6N., + 2,155 x 10° CHU wei-ents 20 3020 5220 volumns 2 154,8 112.8 Total veight of fuel is 356, water formed = 22 x 18 = 536. Thus the ratio of water formed to fuel burned is ab:ut 1.20. Average aviation gasoline runs ab.ut 1.55 libs. of water to 1b. of fuel, Actual Cyele Temperatures To compute either the maximum temperature of the combustion streke or the temperature of the gas at the time the exhaust velve crvens for an actual csse is almost impossible. The total heat energy liberated with a given fuel air ratio is weli enough «nown but the effects of dissociation, iee. the fact that at high temperatures carbon, carbon monoxide, hyurvgen and oxygen may exist in certain proportions in eouilibrium, the effects of heat transfer to the walls and the values of the specific heats at high temp- eratures all combine to prevent an analytical solution, By messurenent, however, the maximum temperatures of the working substances in a modern engine vary between 2500°C. and 3000°C. and at the time the exhaust valve ovens, normally 50° to 60° before the end of the combustion stroke, the gas is between 1200°C and 1400°C, When this has been expanded to about 15 ibs. absolute pressure its temperature varies from 660°C to 800°C. If a lean mixture is used the temperatures are lower during the combustion and higher during the exhaust stroke since the rate of combusticn is reduced, I. Injection of Water Towards End of Combustion Stroke. Assuming the gas is at a temperature of 1400°C when water is injected, we can calculate with sufficient accuracy the temperature resuiting for various amounts injected. Since the specific heats very c.ntinously, strict accuracy would only result frm a series of successive approximstions. In this case mean values as showm are. taken. Referring to the combustion equation; Gas No. of MOL Cp(CHU/20L) CO., 26 x 14,08 = 866 No 135.2 x 7.45 = lo<4 H.0 22 x v.88 = 217, Heat to be absorbed per degree drop in icO?7 CHU tenperature —4— Assume the water added is at 100°C. but has lost none of its latent heat, also that the water is converted to steam at 15 lbs.Abs., a conservative assumption since the heat energy absorbed increases with pressure. The latent heat in CHU per MOL is 9712 at 15 lbs. Abs., and for the range involved for the injected water the specific neat is taken as 9.05 CHU/MOL. Let Tr be resulting temperature and W = no. of MOL cf water added. Then 1607(1400-Tr) = 97LOW + F(Tr-100)9.05 or 177.5Tr + Tri = 248,500 - v72W W i 4 . iG 2U 45 Tr 1385 1347 1273 1156 oy Wt. of water COL 202 2005 1,01 2.02 Wt. of fuel Effect of this Type of Injection on Component Cylinder Parts Since by definition any successful engine will be designed up to the limits allowed by tne cooling aspects, it is a coroliary that any cooling will increase nerformance. However, for an airnisne,oerformance is judged only on a weight bssis. In an attempt to evaluate the cooling resulting from such an injection, ress:nable assumptions @ould not be found from the general collection of test data. Thus, tne following is the only method the writ- er could visualize, If we assume that a soecific temperature is known, woici wuld give the same resulting wall temperatures under c-nditions of a steady state as are actually realized, the problem becomes one of merely equating heat transfers. This "potential" gas temperature will be a continuous variable throughout the length of the cylinder. The critical points are at the trp of the cylinder wall, the head, the exhaust valve and the center of the piston, A consideration of the cycle indicates that excluding the exhaust valve, these parts would reflect approximately the same "steady state" temperature of the working substance. Let Tp be potential gus temperature lor tne twp portion of tae cylinder wall, Tsi inside cyiinder wall tempersture at the top, Tso corresponding outer wall tempersture. Tm tenverature of the cop] ing air stream. Q the rate of hest transfer per unit ares, assuming that all heat from the inner wall gces directiy to the outer, i.e. no longitudinal transfer. For transfer from gas to inner wall Q@= C p Vm Cp (Tp-Tsi) where C is a function of the Reynolds number, and Vm the mean velocity. From inner wall to outer wall Q= & (Tsi-Tso) where K is a conductivity coefficient and t the wall thickness. From outer wall to air stream Q = CG p Vm Cp (Ts0-Tm) with constant defined as above, Thus, the three equations result for any given physical setup in linear equations for the temperature differences multiplied by some constant. These constants in the cases of transfer between solids and gases, above, are to all purposes inpossible of thevretical determinati.n, Thus we have Q= K, (Tp-Tse) q = Ko (Tsi-Tsc) &> | = kK, (Tso—Tm) Thus & ii (it Kotz ) x (Tp-Tm) ~6— and Tsi = Qj-L+i:+i + Tp + Ta e/ kK, K, kK, 2 ' Tsi = Test K+, 2h | roll | Fa | L K, Kk, i Yo i +1 + 2 | + Tp +Tm 2 Calling the resultant constant K Tsi = Tp (K+1) + Tm (1-K) 2 2 With this relation if Tp can be determined from other ccnsideraticn, the other temperatures being known, an idea as t» tne value of K may be obtained . Solution for Tp. As stated the heat transfer between the gas and the wall is given by the relation: Q = CeVm Cy(Tp-Tse) where C is a functicn of the Reynolds number, tot « If the flow is considered as being completel, turbulent C will be proportional 7% ra . Neglecting the variations of p and Vm to .i65 power we have Qu pVm AT which checks well with experiment, Figure 1 shows a plot ot for the four stroke cycle in red. The assumptions for 47 are that the maximum temperature is 2800° C., the exhaust stroke 1400°C, at the top of the wall or head temperature is running being 300°C, Vm is assumed to be that cf the piston, and V is plotted in black. At the time tne exhaust valve cpens, V is assumed tol nerease rapidiy to Vmax and then fall off gradually to zero by the time the piston is again at top dead center and is shown in dashed black. The ratio of p is shown in blue. pilex Thus as stated above the product of these three variables should be pro- portional to the rate of heat transfer from the gas tc the wall. In tnois case only the effects of the exhaust and combustion stroxe: were ccnsidered, the other two being negligible in comparison. The green curves give this product uf the three variables, and the areas A and B will thus give the reiative magnitude of the heat transfer from gas to wall for the two strekes. It is interesting to note that A:E::2:3. This indicates that the total heat trans- fer during the two strokes varies approximately as the mean temperature differences between the gas and the wall during the strokes, since these mean temperatures are themselves approximately of the ratio 4:2, In other words, the variations of density end velocity are such as to compensate each other. We can thus assume our stesdy state conditions such that Tp will be the arithmetical mean of the temperatures of tne working substance. throughout the cycle. If the effects of the intake and compression stroke were considered they would cause a reduction in Tp since,except in the latter portion of the compression strcke, there is a flow of heat from the well to the working substance. Neglecting this effect we may consider Tp as ocefined as Tp = Tmax + To + nex | 1/4 where Ty is the exhaust temperature with no water added. We mav thus say that a one degree drop in the exhaust gas due to cooling shculd give a 1/4 of a degree dr.p in Tp. Determination of K. Taxing es assumed Tmax es 2600°C. and 1400°C. as the exhaust gas temper- ature. Since both of the final ecuations sre expressed on temperature differences, it is immaterial whether temperatures are expressed on the absolute scale or net. Tp = | zeae +1400 + 1400] 1/4 2 = 875°C, And, assuming we have a hot cylinder head of 300° or actually at the tcp of the cylinder wall with an outside air temperature of 15°C., we obtain for K 300 = 875 (K+1) + 15 (1-K) 2 2 K = —.337 or Tsi = Tp .33 + Tm .@7 Since K is only a function of the physical setup and independent of the temperatures, it will remain constant with any internal cooling done, Resultant Cooling of Head and Exhaust Valves. It may thus be assumed that. a 1° drop in the exhaust temperature will give .33 x .25 = .0825° drop in the wall or tc what amounts to the same Exhaust valve temperatures run the cylinder, being on the order of are used. Since the temperature of is due solely to the exhaust gases, erature will decrease linesrly with On the basis of tne previcusly following table gives the resultant Wt of Water Exhaust temp. We of fuel 0 1400°C, 2051 1385 2202 1347 2005 1273 1,01 1158 2.01 967 temperature of the top of the cylinder thing, the head. much higher than any other portions of 800°G., unless elabcrate cooling devices the valve above that of adjacent portions it is reasonable tc assume tnat its temp- the exhaust temperature, determined exhaust temperatures, the cooling. Head temp. Exhaust Valve temp. 500°C. goo°c, 298.8 792 295.6 770 289.5 728 280.0 661 264.2 553 The drop in other portions of the cylinder would be proportionately a little greater than at the head from the same analysis. Also, it is believed that the drcp in head temperature would be somewhat greater than the above since the presence of the very hot exhaust valve has an appreciable effect on tne rest of the cylinder due both to conduction and radiation As to what effeet this cocling would have on power and fuel consumption no an et ytical analysis will apply since, if the materials would stand the temperseture and the fuel not detonate, the hotter the better. As tc test results, along this line, the following is quoted in substance from Pye. (Ref.1) "In scme experiments by Gébson the singleexhaust valve was artificaliy cocled by including a fluid within the hollow stem and designing a tiny redietor at the end of it, with tne following results. Cyiinder and Compression Max Petre] pint Exhaust valve Vaive valve size RP. Ratic BMEP ver BHP hour Temp 5 1/2" x 6 1/2"1650 5 117 .60 40°C, Cooled Valve 2.5"in dia. 113.4 262 760°C ~~ Unecoled Thus the meximum power was increased about 3% with a simultanecus re- duction in fuel consumption of 3.3%. Furthermere, the maximum power obtainable uncooled could be maintained with a mixture so much weaker as to reduce the fuel consumption by 8%. The mean temperature of the cylinder head was lowered 26°C. and the hottest point on the piston by 25°C," Conclusions Relative to This Type of Injection. An analysis of obtaining water from exhaust gas condensation will be undertaken later in the paper, but even without this it would seem that this method would not te applicable to airgraft engines. Even though the cooling effects are desirable it seems thet the required amount of water would mae this method of cbtaining cocling uneconomical, If tne water content could -10- be separated from the rest of the exhaust products and so condensed the weight of the required instaliation would be greatly reduced, Separation of the Water Content from the Rest of the Exhaust Products. The possibility of such a separation exists in the high velccity centrifuge. If the mixture could be rotated at velocities of the order of the mean molecular velccities which are themselves cf the order of the speed of sound, an appreciable separaticn could be obtained. The situation is ideal for this application since the molecular weight of water is 18 whereas the next lightest molecule is hitrogen whose weight is 28, In other fields separations are obtained in this manner, and the rotor of the present turbo-super charger reaches velocities of the required order. It would seem that the development of such a process may be accomplished if the value to be obtained warrants tne effort. II. Cooling By a Six Stroke Cycle. By follewing up the exhaust stroke with en air intake to which a finely divided water spray was added would be a most effective method of absorbing the heat from the cylinder wall, especially if the air intake were so located that the incoming air were directed perpendicuiar to the cyiinder axis, thus giving a high swirl, The same intake under the above cenditions would be satisfactory for the fuel-air mixture, The reason that such wculd be undesirable in present type engines is that it is desired to have the charge pick up as little of the wall heet as possible. Aside from the reducing of the wall temperatures this method would have ancther important advantage in that the temperature of the residual gases left to mix with the fresh charge very much lower giving -ll- a much lower pre—compression temperature as sume later calculations will show. Aside from the effect on the detonaticn, this would mean that the whcle temperature cycle is reduced which gives two distinct advantages, First, the Carnot * eycle efficiency is automatically increased by tne drop in T, if the numerator remains constant. Seeendly, dve to the very ravid increase of both dissociation and the specific heats at the higher temperatures this shift of the cycle tends to increase the srresd between the waximum and minimum temperatures. The above gain in efficiency is borne out in the "hit ene miss" type of engine which gives the highest thermal efficiencies of any type, In this method of speed centrol a governor closes the gas iniet valve for 4 stroke or two, the cylinder working on air cnly. Being unable to find any test data applicable upon which to base calculations it is possible to make only scme general observations as tu the value of such a scheme, In general you would be dissipating the same amount of heat as the conventional water jacket ano radiator system. Best efficiency would be obtained if the steam air mixture were recaptured without mixing with exhaust products. This cculd be dome by keeping the exhaust valve ciosed forcing the mixture out through the same valve it entered and on to a condensor. If it were to be ejected through tne exhaust valve a "Y" valve system weula be used. Either method wuuld sdd to the complication of the cam and valve mechanism, but does not present any greater technical difficulties than wany other features, The advantages of this would be that: (1) the same total amount of heat dissipation w.uld result in a much lower cylinder temgerature. (2) The power olant c wld be completely cowled, reducing considerably tne total drag os . , . way | wae while, by proper design, the condensor could be made negligible. (3) Ali of * cornet aff 7-1 a” 42% ‘eee ‘ ~Le— the auxiliary engine ccmponents such as magnetos, generetors, pumos, etc. would operate without present temperature difficulties. (4) A considerabie increase in economy plus an increase of tie maximum power per stroke. (5) This method if practical would allow an indefinite incresse in rows of cylinders. The obvious disadvantages would be a decrease in maximum power cf somewhat less than one third on the basis of present types,and the condensor system. It is possible that other factors such as displacement and RPM cculd be increased to improve the power cutout, A test cf this principal would not be difficult and might prove fruitful. It has been suggested that either this methca or the previous wight result in the formation of cracks in the walls and pistons. While no test data is available it Goes not seem that this would be likely, since the temperature variaticns would be no greater than those which are ex- perienced in starting. This is because uncer throttled oneration only the maximum pressure not the maximum temperature is appreciably recuced, III. Water Injection for Detoneticn Superession at Higher Compressivun Ratics. The thermal efficiency of an internal combustion engine is a direct function of the compression ratio, or actually the expansion ratio at which it may be operated. Thus for a gasoline engine tnis is definitely limited by the fect that even the best grades of fuel with an optimum amount of tetra-ethyl leas, detonate at CRs between 6 and 7. This results in practical engines of indicated thermal efficiencies of from 41% to 35%. A curve of indicated thermal efficiencies against CR is shown in Fig. 2, both on the basis of the so called air standard efficiencies which consider the working substance te be a perfect gas with no¥ heat loss and also for the best actual test engines, The latter is extropolated by the generally accepted ~15- experimentally obtained relation, efficiency = 1- (a) **¢ as given by CR Pye (Ref. 1) for volatile fuel engines, It will be noted that the theoretical and exoerimentel curves are of the same tyoe differing by a constant amount. There are also shown exnerimental points for two test diesels overeating at two different load conditions. At this point it is believed desirable te explain the funcamentel differences in the problem between the volatile fuel engine and the diesel type. While with the latter, indicated thermal efficiencies well above the experimental curve K-M may be obtained at high air to fuel ratios and low RPM, it has been found impossible to bern more than 75% of the air in a high speed diesel. This is because a more complete mixing of the fuel and air cannot be obtesined in the time availeble, Any further injection cf fuel in an attempt to burn more air results oniy in e vroduction of black sioke with no appreciable increase in power, This Jimiteticon would obviously not apoly to the volatile fuel type which experiences no’ oarticular difficuity on that score, Another limiteticn which would aoply eoualiy to beth types of engines at high CRs is thet of maximum pressure. At a CR of 12:1 the diesel on 2 basis of 75% of the air burned at close te constant volum would give maximum vressures of about 1200-1300 los, ver sa, inch, while the volatile fuel engine would develop pressures of the order of 1500 to 1600 lbs. per sa. inch. The diesel operating at nighbRe can evoid the high oressures with a consequent reduction in efficiency by delaying the fuel injection until later in the stroke. The volatile fuel engine would be efforded no such relief, Theory of Detonation As indicated the major orcblem in increasing the thermal efficiency of a gasoline engine by increase of CR is the shenomene of detonetion. Since this first beceme a problem a vast amount of exoerimentel] work hes given rise to several thecries as te the actual cause. Basically there are the same but the: yary widely in attempting to exnlsin sssocisted phenomena in connection with "dones", The theory which annears ta the writer as offering the best physical eynlanation in the light cf the various tests is that offered by Egerton end reviewed by Pye in Vol.l(Ref.1) and of which the following is a brief summery. When no detonaticn taxes place the flame started by the spark spreads ot steadily until the fuel-air mixture is consumed with 2 speed whicn is a function of turbulence, toe fuel itself and the pre-ccombustion tempereture, Vhen the vre-combustion temperature is such that onJy 2 siight increese in temnerature ig recrired to reach the snontanesus ignition temperature of the mixture, the increase in pressure of that sertion near the spark Ding may be enfficient to cavee the rest of the mixture tn ignite’ spentanecusiv _giving very high locel pressures, Pistons heave actualiv been broken from thet cause. Tt was goon discovered that verious substences would deley the meet of detonstion when mixed with the fuel in very smell oromertions, and that there were definite aptimim mixtures in esch case. By far the most s whet is commonly called "tetra-ethyl lead", waich is 2 mixture de effective fate of leed ethicle, Pb (CoHs ) gq» end ethylene dibromide, C,H,Br., toe latter being added for the murvose of preventing ercessive lerd deposits. Jn the nuclear drop theory the value of this dope tarcugh inck of more complete mowledge is referred to es an "inhibitor of chemics? reaction", By various tests it has been determined that ah temperatures considerably helow thet of tenition most fuel-sir mixtures undergo lor order oxidations with the formetion of highiv wnsteble peroxides, These peroxides have a lower Tenition temnereture and when reached give a very renid release of ote "9 energy, The formation of these neroxides emeidered te be the resvit po fe) of a resctive chain set up in the higher energy roleculee of fuel and oxvgen, each perovide molecule formed incneing the formaticn of more such molecules, Without going throuch the chemical line of ress ening the presence of the lead ethide comletely sunereeses the formation of there low temnereture partic] ovidetions br bresking um the reective chains, This it cen be seen in theoory why leoded fuel sen orevent detemetion only up to 2 ecertein paint, 2 fact every one ho hes criven e hot motor realized in * We thus errive at the point, that te increase indefinitel; the onset of detonation the temmereture, at the time of tne spark xescage of the mixture must be reduced to the ooint that the tempersture of snontaneour ignition will not 3 be renched hefore the flame front hes covered tne mejor portion of the combustion chamber. The most oreeticel method of doine this Is br tne addition of water te the fouel-sir mixture, its very hich latent heat of veporizetion will aovreciably deeresse the temnerature rise during comoress-— fon, Thus the mater to be really effertive mist not he sdded as vaver but ae Finely dividend water carticles In c water injection tect whieh ds not evetlchle et oresent to the writer, the arm fdr Goros Fngtneering Seation, emersted = test enzine at BMEP of 579 The, ner sc, inch, an almost inemmeeivable figure, Aircraft engines nt nrasent onerate at 2 mevimim of the order of 140 lbs, BMP, oF | Water to Permit Operation at Varin Gomoression Ratios. On the basis just discussed is now oresented 2 series of comoutetiones to determine the amount of water necessary to orevent detonetion to suecess— ~16- ively increased CRs. The first phase of this has to do with a determination of the maximun temperature of compression, teking into aceomt the resioual exhaust gas left present when the fresh charge is dravm in. A conservative note is sadd- ed by teking the temoerature of the residual sas at 800°C, each time, since this temperature would be continuously lower, Assumptions: Ll. A cylinder with 2 swent volumn of 1 cubie foot, 2. "N" as recommended by Pye as 1.3%. OQ e Kv (volumetric heat of svecifie of a given volum) Kv = 28.5 £6, los. ver SCF for products of eombustion for the renge up to 800°C, Kv = 20,5 ft. los. ner SOF for fuel-air mixture up to 100°, 4, 60°C. es fuel-air temperature voon entering the cylinder but before mixing with the residual exhaust gases, CR Vol. of Exhenst Vol.of Exhaust Vol. fresh Temp. before Max. Temp of gas at 800°C gas at O°¢ enarge-°C Gomoression Gomoression ® o aR 220 O51 28? 199 ¢ 419 ¢ 10 eill 2928 202 88 50 11 «100 025 282 85 518 12 Rory) 2028 .82 82,5 BES Now if we sesame 2 given fuel operating at 2 CR of 831 without detonation, then by sdding weter to the fuel-air mixture et hicher ORs sufficient to bring the meximim temoerature of compression to that of @/ tnere should be no detonation, This sassames that the injected rater is completely mixed and vanorized in the time aliowable. Whether this will be true or not, theory cannot reasonably state. However, vaporization in other tyoes of apparatus namely, the cloud chamber, would seem t indicate that fhe vaporization would be complete. However this ehemld be one of the first experiments in case the theory edvanced in this paner is given an experimental test, It is necessary to make some simplifvinge assumotions for this calenlation as an exact method would be mite tedious. The basis used is that the water is added after the compression is comnleted and that the water all changes state st 9 oressure equal to that at the end of the com- oression stroke, Any reasonable veriation from the above vill alter the results only br 9 vert small amount. The fuel-air mixture used is that given at the first of the naner. It being easier to work in MOL, the change is mede, anc the comnutetion for a CR of 7:1 is as follows, From steam tables it is found thet one MOL of water will sbsorb 10157 CHU in going from water at 100°C, to steam under 9 oressure of 160 lbs. the tempereture of the change nf state being anrrovimetely 184°C, The volumetric heat of steam in that range is about 6.62 CHU per MOL. Thus 10157 + 6,62 (420-184) # = 105 where WY = number of MOL of water renuired, if It 200895 pe mm The moumber of MOL of fuel in the fuel-air mixture used 4 = ,O02e2,. 180.2? The molecular weight of CgH, = 78 and of C7H|, = 100, thus the mean ~18— molecular weight of the assumed fuel, 50% of each by volumn is 89 and the total weight of fuel in a MOL of the fuel-air mixture is 1.v@ lbs. Also 200895 MOL of water is .1461 lbs, Thus the ratio of the weight of vater recuired to the weight of fuel used is .0817, Following is the result for the verios assumed CRis. CR CR Comm. Wax. temo. of CHU/MOL Wt, of water Pressure Como.— 420° to be absorbed required te fuel used 6 10,84 159 0? s) 7 18.8 196 20° 105 .O817 8 15.85 233 41? 215.5 168 9 18.6 274 61° 220.3 251 10 21.4 #14 RO? 420 0350 ll 24,2 256 98° 515 401 12 27.2 400 115° 6U4 471 Gooling Problem with Increased CRs. | The increase in CR antomatically solves the cooling oprobiem since tne mean temperatures of both the combustion and exhaust strokes are lower. This is exemplified by the following test quoted in substance from Pye (Ref. 1) tA "As regards the effect of CR, Gibson's results shor a cuite regular droo in the fraction of Horse Power to cooling over rated BHP of from .9% to .69 for a series of CRs of 4.5, 5.0, 5.5 end 6 to 1, all tests being at 1600 RPM. The reduction of heat to cooling water at higher retios will be chiefly in that part lost during the exhaust stroke. The higher the ratio of expansion, the mowe efficient will be the cycle, and therefore the covler the working fluid when the exhaust valve opens." The "notential temperature" analysis might be apolied to this case, but ~19- the determinati.n of the exhaust gas temperetures analytically is a difficult and dubious process, This tempersture in tne vrevious case vas assumed from actual engine operation. The real problem in this case is the weight of the eendensor necessary, so we will now try to make some reasonable estinates on that score, IV. The Condensor System. As every one knows the later dirigibles end blimrs used condensors to obtain the water content of the exheust exses in order to obviate tne valving of its exnensive helium, For these craft the c:ndenscors were extremely heevy, those on the Akron weighing about 2000 lbs. ver engine including attaching structure and connections, As compared to this tne oroblem for the airplene can be solved with a small fraction of this weight. In the first place the boundury lever around the dirigible is very thick and it wes found on tne Akron that the condensors were operating na field of greatly reduced velocity as compered te tne craft's speed. Le Secondly, in the case of the dirigible it was necessary to act on tne entire exhaust products if recovery of 100% on the basis of fuel c@nsumed was to be expected. In any of the methods herein contemplated the water added to the working substance makes the recovery that much more economical for any given required amount, since the total amount of heat in the exhaust products is a constant for e given set uv, though of course at a lower mean temperature differential. Thirdly, the sattschment to the structure in the case of the airplane is much more simple. Since any such system if aprliceble st all would be oarticulerly advantageous in long range operation such as that carried out by olanes of the patrol boat, trans-oceanic cimmercisl boat or long ranged bombardment -20- types, the ideal. ccndensor location would be in the wing section, In the larger c.ntemplsted tyves there will be a depta at the region between the two motors of a fiar motored ship of armind five feet at the meximum point, tous allowing amole room, The ccndenscer structure could be designed to support the stiffeners and to transmit the torsional load. Thus the design problem from a structural view point while difficult should not be critically so, In Fig. 3 is shown a schematic drawing of what the writer believes would be the most efficient type of condenscr, The theory of the ducted radiator has received mich attention of late and indicate@ that if the velocity through the radiator is sufficiently reduced below that of the free stream, the heat energy added may be made to equal or even exceed the energy removed from the stream thus reducing the drag of the system to a negligible figure and offering the theoretical possibility of e thrust. The basic theory is covered in Rand M 1684 and 1702. In experimental test the to drag has been reduced to a negligible figure butathe writer's knowledge no thrust has yet been developed, In order to get some idea of just what an outlay a condensor system would be for various reouired recoveries the writer used as a basis the work reported in ACIC No. 44 of Mey 1, 1924, "Condensation of Water from Engine Exhaust for Airship Ballasting". This was used ruther than some of the more recent types of installation since the velocity was known, thus no poundary layer effects of the airship had to be considered. This appears to have been the first systematic tests made in regards to this problem and extensive service tests were made. The Model I which) used, 25 a basis of extropolation was nothing more then a series of 1" pipes 6U feet in length and arranged in three benks thus giving an over all length of 20 feet, with the necessary headers and separator. The totai weivht of tne pipes alone was about 40U01bs,. and with headers and separator about 450ibs. The power plant was a Liberty engine which with the prcpellor used wes calibrated at 507 HP at 160URP¥. The result of a 90 hour test is summarized as follows; Average alr speed entering condensor 48 i.PH n "temperature entering condensor 15°C. Total weight of fuel used 15,075 lbs. " n " water collected 13,943 lbs, Water collected on basis of fuel used 92.5% It was further stated that due to various leaks considerable water was lost, the amount not being estimated. The condensor was suspended at a slight angle to the horizontal, the longer dimension being in tae direction of flight. With that tryoe of condensor there must have been a very considerable loss of cooling efficiency by the very geometry of the system, also the unsupported tube lengths were 10 feet, making the tube strength ana consequently the weight considerably greater then would otherwise be necessarv, From tois and boundary layer considerations along the tubes it is believed a much more effective system could be worked ont and a very conservative assumption would seem to be that weight of tne conducting surfaced could be reduced from 1.3 lbs. per BHPecndensed to 1 1b. for identical performance, Since long ranged aircraft operations is tending towards higher altitudes, the condensor system should be preparec to operate effectively at 20,000 feet at least. Thus we will sttempt to extropolate these test results to that condition, knowing that if successful their satisfactory operation is automatically assured at lower altitudes, Effect of Velocity Variation In the matter of velocity through the condensor the following is assumed. 22. If the velocity is kept at from 1/2 to 1/% of that of the free stream, theory indicates (Rand M 168%) that the drag will be negligible with flight velocities of the order of 200MPH. Thus the expanded velocity is teken as 72 siPH. Since the rate of heat dissipation for this type of flow varies experimentally with (v) 27 approximately tne increase of effectiveness due to this & 78 .738. If we hed assumed — velocity essumpticn on a weight basis is (48 (72 ) ) a velocity of LOOMPH as we probably well could with an almost negligible increase in drag this factor would be .622. Effect of Temperature Veriation,. The extropclation for temperature differance in the cooling airstream is a bit more difficult in this case, In the various tests of the apparatus the temperature variation along the length of the tubes was not operating measured, thus it is impossible to give a mean, temperature Ser-eperating as a function of the totai dissipating surface. If this mean temperature were given the correction for outside air temperature would be very simple. It seems quite apparent thet this mean temperature would be very much smaller tham the arithmetical mean, since after the water has started to condense toe amount of heat transfer will be greater than previously for a given temperature drop, In other words, a considereble linear proportion of the c.ndensor will be dissapating heat at temperatures of less than 100°C. This fact makes the effect of the drow of the outside air temperature of 40°C. between the surface and 20,000 feet result in a very large increase in the effectiveness of the condensor, For example, that portion where the temp- erature is 100°C. wuld have its rate of dissipation increased by the follow— ing factor 378-248 = 135 = 1.59 and in increasing amounts at the lower 372-288 288 temperatures, While it can thus be seen that the temperature decrease of the outside air would be very effective to the condensor operation, not knowing the linear variation another method must be used to obtain an: analytical expression. A study of the variation of the final exhaust temperature with outside air temperature in the report above mentioned indicates that the mean effect of a 1° drop in outside air is to give a little greater than a 1° drop in the final exhaust temperature. In the test cited the mean ex- haust temperature was 35°C. and the cooling air 15°C., the drup furthermore seemed to be accentuated at the lower air temperatures. Thus if it is assumed that: the 40° drop in outside air will give a decrease in final exhaust temperature of 20° we may consider this to also inciude the variation of mixing retio with temperature, The Effect of Velocity and Density Variation with Altitude. If the velocity through the condensor is ellowed to increase in pro- portion te the plane's velocity the effect on the drag of the condensor will be slightly negative. In a plane equipped with turbo super charger and constant speed orovellors the variation of velocity with altitude may be 5 Ve / . a . , ae expressed as Tr? 7% up to the critical altitude. The effect of velocity o vey - mate’ in regard to heat dissipation is expressed by (4) approximately. o The heat trensfer will vary inversely with the density, tnus the factor for the combustion of these two will be Ce) = O(F% 2f Since g~ at 20,000 feet is .532 the factor of effectiveness wili be .514. Combining the various factors with the originelly assumed weight of 1 lb. per BHP c.ndensed _1 x .748 = 1.2 lbs. ver BHP condensed at 20,000 feet 2014 ~24- with a final exhaust gas temperature of 15°C. Thus et 15°C. with mixture ratio of 15, .149 lbs. of water ver pound of fuel burned will be carried out with the saturated exhaust gas. This is a function of the partial pressure of water vapor, the caiculations being shown in ACIC Ng.44. As previously assumed cembustion of each pound of the average aviation type gasoline gives 1.35 lbs. of water, considering tne air coning through the carburetor as dry, which makes the amount of water recovered about 1,20 lbs. plus any water injected ber pound of fuel burned. At this time it shouid be stated that the water collected will contain ecnsiderable carbon or soot ang alse small emounts of sulphuric and hydro- bromic acids. This will necessitate the instaliation of mecnanical and chemical filter but no\ particuiar problem should result from this score, Another practical difficulty wuld be the protecting of the water lines, etc. from freezing, Evaluation of the Complete System. From the previous analysis it is possible now to determine the weight added by the condensor system and balance this egainst the reduction in fuel consumption. No attempt is made to estimate the maximum power which might be develoved from an individual unit as tnis will be a function of too many other factors. In each case it is assumed that a 1000 HP engine is developed, the following assumptions tv cover any increase in weight with tre increase in Maximum pressure. It seems probable that wito sufficient development tais increase in weight would disappear or even go negative. While the increase in maximum pressure will require an increase in bearing sizes and wall strengths with constant pcwer, the cubic dimensions will decrease. To be sure of remaining conservative on this score, however, .C5 ibs. ver EHP is added for each unit increase CR above 6 at which the specific weight is taken as 2 lbs. ver EHP developed. The still born Packard Diesel is covered by this allowance since it weighed 2.27 lbs.per BHr and operated at a CR of 12. Also an aliowance is made for the weight of the injectors, connections and separator, etc. separately as shown. CR of Exhaust total st. of Increase in injectors Total to be cooled cooling surface wt, of engine Connections, increase required etc, 7 6.3 76 50 198 226 8 12.3 148 100 120 568 9 17.3 208 150 140 498 10 21.5 256 200 160 616 ll 25,0 300 250 186 730 12 28,2 338 300 200 838 From this the folldwing table indicates the effect on range. The thermal efficiencies are based on an extropolated experimental curve previously discussed. The brake thermal efficiencies would all be scme- what lower but since the ratios shogid all be the same it makes no difference in the result as the comparisons are reduced to ratios with the specific fuel comsumption for the 6:1 engine taken as .40 lbs. ser BHP hour. The type of “peration assumed is maximum patrol boat range which at present would give a fuel load of 6000 lbs. per 1LOCOHP engine. <Any other specific type of operation could be worked for given conditions. CR Indicated ¢%increase Consumption Flight hrs Tctul wt. Total 4 thermal bver that per RHP hour to pay for of fuel flight increase efficiency at CR=6:1 installation carried hours P12 at .f5 pwr, per engine .75 pwr. fo) 0885 240 6000 20,00 7 0568 9.85 561 772 5774 £1.20 6. 8 2589 16.1 2836 7.66 oog2 22d 13.7 9 2405 20.2 2517 8.0C 5502 25.20 16.0 LO 2420 20 4 23g 8,15 5484 24,00 20,0 il 4353 29.28 2283 B52 5270 24,85 24.2 Le 0445 52,8 2ecg 6,04 5162 20.60 28,9 Conclusion. As anyone who takes the time to follow through this analysis must very forcibly realize many somewhat dubious assumptions have had to be made, While in such cases an attempt hes been made tc keep well te the conservative side the writer has a very high respect for the laws of perversity and consequently feels rather pessimistic es to the ultimate vaiue of the methods considered. It is believed, however, that a sufficient theoretical value exists to warrant the experimental check of the indicated possibiiities., It is not anticipated that the required preliminary experimental work woulda be particularly costly. As regard to detcnetion suppression by water injection, it would seem that test engine meesurements could be reedils taxen, if not alread, in existence. The preliminar: condensor work could all be done in a wind tunnel with an eesily constructed scale secticn, and the heat trensfer determined by a steo by step integration after the pressure drop and velocity distribution has been obtained, Following is a brief summary cf salient c.nsideraticons. -27- 1. Water injection for cocling of exhaust gas only does not seem to be feasible unless the water content could be separsted from the rest of the exhaust products before condensation, 2. No evaluetiin can be made of the six stroke cycle without more experimental data, 8. Water injection for detcmaticn suppression seems to offer reasoneble possibilities of ultimate value, 4, In amy condensatiin method at high altitude the gas to be opereted on must be kevt at approximately 15 lbs. absolute vressure as the exhaust gas will carry ut too high a water content at reduced pressures in the saturated state. This means that the fraction condensed must by pass the turbo suver charger. 5. The ancrease in opereting compressicn retios automatically solves the cooling problem of the engine itself. 6. The condensor system can be operated with a negligible increase in total drag. i Tit I mt j us : u =] i ee imeem! i a tt + . i" tH Seon BD i | To Coot i I t aa = CeCCCeeeree i =| = e rt 4 t or it j 1 =m 5 5 Ba at I Ee a t == ogn8 i 4 \ orb awe ; i Trt cP Tt. cot T 7 ‘ rr H z ssue cry i : + - he — t t tt : Ter i t t = t t = I ‘ = n | a gj om HY “HAE EHS i" ] | i! i | 7 Crett adi i oe is { : ri n 1 T T ' See nee 2 = T s toot | Biel = T T | LJ rs TT TT an i t - t | i HHH 3 i hit a8 ist tie ss NS 5 3 i = = = T + “HERI T uns = : oe TT Girt t t wae = I Co i i T t = t ae : EERE T | i t tt SBNE SERBS SRY i i ; a5 T be B +4 i Aas aa ooo =" eee oe t M 4 — P +t pt t T | BEGSEnR 4 THe ee eee ee Sees! +. i a z t i t aaanm = + 1 se we = < 20 eae i. me) sen tJ J t oH t HPs HE h t is aan Saw : a Pe ne bs t : Sct een 4 = =e bem i 4 ia 5 BE usm t : ASH 3 t | ol C F Sore 4 t + t Ay i u | ean I ah ; T =n ™ ‘ 1 2 S| Sat Pl CAE RT th F == ne an ai = : Ce ; { == oT : y itt T rd t T T . i an I BS , : = : 5 ‘ - _ 7 j L TTT = -. jase | t - + SR C | i. 2 SSS Cece z = t = I i T if n 4 4 it nt a Lat cTry i is a8 + T t t t ta = | imi T T " T t i t ¥ an t t po Et phy : i. aa t ¢ s 2 as ar + i ena TI t + TTT Le T a 5 —t a : coe f i os - - — — ~ +f aa r T "a i 5 : t : + ar. 7 4 Eee = : SCORE an 2 a r omme int oe } i Secor sae a mu Bene w tt T Rect mt | t t = tt : a mera fe <= “ 1 v or t 7 t " 5 i “VS "MN Ni SOV : L ' dec aide io yo3 peo ' i , , 4 rf Ld { . , |. . + ao ae 4 } j 6 4 ty a8? ds eseys A v4 GOT PrESS ih" _ + aa earn + eeeber ge peoce a re + 4 4 i [ey ede. EASE | et | ara Ee, bea Z hol The a a i 1 | L i es Sea + ! T | 1 i | P| ' 5 . | Gaen.s cee. : 7 oN i : r A t . 7 ‘wf i Tote Sop Pe cote : ’ TTT x bor ope hee Ll — et - opto a wr 7 Pande pe oo pe on Oneere belie oie t 1 f fr : Paani cen f pe bef oo Re f *>GLPINOPLI FE AS THOPPUsE OU synosedd BINGONAPS 944 4VU4 Ya_QoU oq Tin BI "SeSe@sOUL [eppUsleL jp sunjeusdiiey ay} 3B pssvetosD ST Jys Buyfooo oy yo AgfTooyTsa auyz snyy *go4j40q 944 64 do} ey woay pus ucot7oeu Sty. oF TeuZoU vt MOLTZ. sus qonooke syL *ZO8UGPUOQ pesodoig Jo UoPposu-sso01g op peuMayoy The Internal Combustion Fngine by D.R. Pye. The Internel Combustion Engine, Vol. ITI, The Aero-Fngine By D.R. Pye. Cooling of Aircraft Fngines, Pith Svecial Reference to rthylene Glycol Radiators Mnclosed in Ducts br FAW. Meredith, R and M 1582, Condensation of Water from Fneine Pxneust for Airshin Ballesting by Robert F, Kohr, ASIC No. 44, pay International Critical Tables, Vol. 5, The Heat Canecity of Gases anc Vaoors.