nasa sp-19 - chemical rocket propulsion - olsen

Upload: john-clayton

Post on 06-Apr-2018

224 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    1/58

    s_"

    o-s_/

    SESSION M

    Chairman, WALTER T. OLSON

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    2/58

    Do. WAL____ T. OLSO_ is Chief o/ the Chemistry and Energy Conversion Divi-slon of the NASA Lewis Research Uenter. Appointed to the Lewis staff in19/_2_ he has conduvted_ supervised and reported research an /uels and com-bustlan program, s ]or turbojet_ ramjet_ and rocket engines. Dr. Olsan has a_soserved as a consultant to the Department of Defense and the U.S. Air Force.He wa_ graduated/rom DePauw University in 1939 with a B.S. degree in Uhem-stry, earning his M.S. and Ph. D. degrees in UhemL_try and Uhem;cal Engi-eering from Case Institute of Technology in. 1930 and 19/2_ respectively. Dr.Olsan is an Associate Fellow o/the Institute of the Aerospace Scienees_ a Direc-tor of the Combustion Institute and a member of the American Chemical ,_ciety and the American Rocket Society.

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    3/58

    IntroductionBy WALTER T. OLSON

    The status of some of the key technology ofthe chemical rocket system is reviewed in thisSession. In this review it is hoped that (1) thecurrent problems in this technology will berevealed, (2) the fundamental physical prin-ciples involved m these problems will be madeapparent, and (3) the directions in which solu-tions probably will fie will be indicated.A sonde, satellite, space probe, or manned

    spacecraft requires for its mission a specifiedamount of kinetic energy, more often expressedas a velocity. With regard to rocket-pro-pulsion, the speed imparted to the payload de-pends on just two items: the energy per poundof propellant and the amount of propellantthat is carried. The former is specific impulse,is expressed as seconds, and is dependent onpropellant chemistry; the latter is mass ratioand is dependent on vehicle design. Figure 1shows the relations. For example, an RP-1(kerosene)--liquid-oxygen engine in a vehiclein which 92 percent of the initial weight is con-sumable could provide a velocity to the payloadof 27,000 feet per second (92 percent is a num-ber that is commensurate with several large ve-hicles using this propellant). If the thrustingis done near the earth, 32.2 feet per second mustbe deducted for each second of burning. About10,000 feet per second must be deducted for a5-minute launch because of work againstgravity; therefore, 17,000 feet per second is thenet velocity for this example. Other correc-tions can be ignored for the present purposes.Of course, one rocket vehicle can be made

    the payload on another, larger vehicle---bystaging--to build up the ultimate flight velocityneeded.The inherent simplicity and instant readiness

    of solid propellants makes them useful formany applications: launching of smaller satel-

    ]ire payloads, as in Scout; orbital changes orinjection, as in Echo, Tiros, or Surveyor; stageseparation and emergency escape, as in Mer-cury, Gemini, and Apollo. Their great valuein the military mission need not be detailed.Over the years since the first asphalt--po-

    tassium perchlorate and double-base powders,the specific-impulse performance of solids hasrisen from about 180 pounds per pound persecond to values for today's composite propel-lants of about 245 seconds (on a 1000-1b/sq in.chamber pressure basis). Further gains, how-ever, are difficult. Expenditures of many mil-lions of dollars in the last half dozen years inan attempt to find new ingredients for higherspecific impulse have made specific-impulsevalues of the order of 300 seconds appear pos-sible, but probably at the expense of the physi-cal properties of the grain. Best gains in theperformance of solid rockets have come fromimproving the mass ratio. Innovations such ashigh-tensile steels, cases wound from high-tensile glass filaments, and composite structuresfor nozzles have boosted propellant loadingsto as much as 0.87 to 0.91 of the vehicle grossweight.

    FIGURE 1.--Rocket vehicle performance.

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    4/58

    CHEMICAL ROCKET PROPUL$10N

    Not shown on figure 1, but performing in therange between solids and RP-l--liquid oxygen,are propellants like nitric acid and alkyl-hydra-zine mixtures. Having the advantage of beingliquids at room temperatures, they are useful inspace rockets like Agena, the Mariner course-correction engine, and the lunar excursionmodule of Apollo.

    The higher end of performance includingeven reasonably stable chemicals, is representedby hydrogen-oxygen and hydrogen-flourine.Addition of metals like beryllium or lithium tothese systems improves the specific impulse, butthis approach requ.ires a very high weight frac-tion of hydrogen, which makes a poor densityfor the system, and also adds the complexity offeeding and burning metal.

    For reference, the coming first generation ofnuclear heat-transfer rockets is shown, with per-formance at 700 seconds for hydrogen at 2672 F to 800 seconds at 3500 F. With u propellantdensity about one-fifth to one-twentieth that ofchemical rockets, with an inherently heavierpowerplant, and with the necessity of shielding,nuclear heat-transfer rockets as presently con-ceived have a long, rough, and expensive roadbefore they will be able to compete in perform-ance with chemical rockets. Propellant frac-tions will certainly have to exceed 0.75 of grossweight. Therefore, chemical rocket systems willbe used for a long time, and it is vital to thespace program that they be made as efficientand reliable as possible.The particular propellants to which the dis-

    cussion of this Session are directed are RP-1--

    ....... RP j._-:_.0XY_GF__ HYDROGEN-OXYGEN I:-_C:-!

    J-2

    188,000 200,000

    ,,5oo,ooo i:-2oo,ooo

    FI(_VRE 2.--NASA chemical rocket engines and poundsof thrust produced.

    COio.YsicsDYNAMICS ANDCONTROLS

    PUMPS AND TURBINES

    COMBUSTION

    THRUST CHAMBER

    FIOU_ 3.--Chemlcal rocket.

    liquid oxygen and hydrogen-oxygen, althoughmuch of what is presented is broadly applicableto more than these propellants alone. Time doesnot permit detailed discussion of other liquidsor of solids. Furthermore, these propellantsare prominent in the several principal engineprograms of NASA as seen in figure _. Theseengines are vital to many of our foreseeablespace flights, particularly manned flights.The H-1 engine is an outgrowth of the Thor

    and Atlas engines; a cluster of eight of theseengines drives the Saturn C-1 booster. TheF-1 engine is currently in development and willpower launch vehicles for manned lunar flight.The three hydrogen-oxygen engines are forvarious upper stages of lunar vehicles. Theysignify a heavy reliance of the NASA spaceprogram on hydrogen-oxygen. Consequently, alarge technology effort is progreasing on thiscombination, which should have its first suc-cessful flight in lhe Centaur vehicle, poweredby two of the A-3 chambers. The J-2 engine,in development, is for upper stages of Saturnfor manned lunar flight. The M-i engine is forstill larger launch vehicles.Figure 3 shows the five topics of the chemical

    rocket discussed in this Session: fluid systems,pumps and turbines, combustion, thrust cham-bers, and dynamics and controls. The physicalprinciples, the understanding and mastery ofwhich are vital to successful engineering ofrocket-propulsion systems, are discussed.

    4

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    5/58

    INTRODUCTIONBIBLIOGRAPHY

    GRELACIII, C. J. and TANNENBAUM, S. : Survey of Current Storable Propellant. ARSJour., vol. 32, no. 8, Aug. 1962, pp. 1189-1195.

    OLSON, _VALTER T. : Problems of High Energy Propellants for Rockets. Rocket and Miss ileTechnology, Chem. Eng. Prog. Symposium (Am. Inst. Chem. Eng.), set. 33, voh 57, 1961,pp. 28-37.

    SEIV_T, H. S. : Twenty-Five Years of Rocket Development. Jet Prop., vol. 25, no. 11,Nov. 1955.

    St_rw0N, GEORGE P. : Rocket Propulsion Elements. Second ed., John Wiley & Sons, Inc.,1956.The Chemistry of Propellants. Combustion and PropuLsion Researches and Reviews.AGARD, Pergamon Press, 1960.

    Jet Propulsion Engines. Vol. XII of High Speed Aerodynamics and Jet Propulsion, Prince-ton Univ. Press, 1959.

    Liquid Propellants Handbook, vols. 1 to 3, and LPIA Abstracts, 1958 to Present. LiquidPropellant Info. Agency, Appl. Phys. Lab., Johns Hopkins Univ.

    Propellant Manual. SPIA/M-2, Feb. 1961, and SPIA Abstracts. Solid Propellant Info.Agency, Appl. Phys. Lab., Johns Hopkins Univ.

    Rocket Motor Manual. SPIA/M-1, vols. 1 and 2, Sept. 1962, SPIA, Appl. Phys. Lab., JohnsHopkins Univ.

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    6/58

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    7/58

    37.FluidhysicsfLiquidropellantsBy Donald L. Nored and Edward W. Otto

    DONALD L. NORED, an Aerospace Scientist at the Lewis Research Center, isengaged in program planning and design studies for pressurization systemsand fluid components in the Apollo Propulsion Office. He has wor]ced onadvanced propulsion facility design and operation, high-energy-propellantrocket engine development, and vehicle systems studies. Mr. Noted receivedhis B.S. degree in Chemical Engineering from Texas Technological. Collegein 1955. He is a member o/the American Rocket Society and the AmericanInstitute o/ Chemical Engineers and i_ a registered Pro/essional Engineer o/the State of Ohio.EDWARD W. OTTO iS Head o/ the Fluid Systems Section o/the Lewis ResearchCenter. He has specialized in propulsion system_ dynamics and controls andiS currently engaged in the st_My o/the behavior o) liquids in a zero-gravityenvironment. Mr. Otto is a graduate o/ [owa State University, where hereceived his B.S. degree in Mechanical Engineering in 1944. He is a memberof Pi Tau Sigma and Phi Kappa Phi.

    INTRODUCTIONThe use of liquid propellants in chemical

    rocket engine propulsion systems, while quitecommon, has necessitated the solution of a num-ber of problems peculiar to rocket vehicles.Two relatively new areas of application, how-ever, have presented new fluid physics problems.These areas are (1) the use of cryogenic propel-lants in general and (2) the use of liquid pro-pellants in a space vehicle. This discussiontreats these two new areas of application by re-viewing the various attendant problems andindicating the extent of present knowledge.Figure 37-1, showing a typical liquid bipro-

    pellant rocket vehicle, indicates those areas di-rectly concerned with the physics of the fluidsystems. These fluid systems are composed ofthe tankage for the liquid propellants, asso-ciated internal baffles and diffusers, insulationsurrounding the tanks, and the pressurizationsystem. Also included are the necessary lines,control components, and heat exchangers. Theproblems existing in such systems may be di-

    vided into two categories: those pertaining tothe powered phases of a mission and those per-taining to the coast phases. Pressnrization andsloshing of propellants are the two problem

    CONTROL _J-PRESSURIZATION

    []

    |"F_ouRE 37-1.--Problem areas in typical liquid propel-

    lant rocket fluid systems.

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    8/58

    CHEMICAL ROCKET PROPULSION

    areas of interest during the powered phases,while thermal control and effects of zero gravityare of interest during coast.

    PRESSURIZATIONThe first major problem area, pressurization,

    exists whether the vehicle is pressure-fed orpump-fed since a pressurization system of somekind will always be required. Such a pressuri-zation system can become a large part of vehicledry weight for those space vehicles utilizingcryogenic propellants for the following reasons:

    (1) Liquid hydrogen has a low density, hencethe tanks wi]] be large, and a resultant largeamount of pressurant will be required.

    (..9.) There will be a drastic drop in the tem-perature of the incoming pressurizing gas;thus, the final gas density will be high, and evenmore gas will be required.The largest problem encountered in the field

    of pressurization is the accurate prediction ofpressurant requirements. Such prediction isdifficult because of the complex transient phys-ical and thermodynamic processes occurringwithin the propellant tank. These processesare shown in figure 37-2.The heat-transfer processes of primary im-

    portance are the heat transferred to the walls,heat transferred to the liquid, and heat. trans-ferred to the ullage space vapor and gas by theincoming pressurizing gas. The mode of inter-nal heat transfer to the walls and liquid is someform of free convection; thus, the geometry ofthe tank becomes important. The heat trans-

    EXTERNAL /HEAT /TRANSFER-_ \

    ! HEAT i'R'ANSFER : _ MASS TRANSFER"_'-TO WALLS: : BY CONDENSATION

    OR OR EVAPORATION

    CONDUCT,ON_vlOVING INTERFAC__HEAT

    TRANSFER____TO LIOUID_mLIOUID c$-_,_

    FIOUY.E 37-2.--Heat- and mass-transfer processes affect-ing pressurization during expulsion.

    ferred to the ullage gas and vapor is dependent on the mixing occurring and hence is a func-tion of diffuser design (i.e., inlet gas distribu-tion), propellant vapor density changes, andpressurant density changes. Heat transferfrom the external environment, from the wallto the liquid, and by axial conduction is gen-erally quite small during the expulsion periodbut of extreme importance during coastingperiods.The mass transfer within the tank occurs by

    condensation and evaporation at the liquidinterface and wall interface. Generally, it isassumed that mass transfer, while significant insmall tanks, becomes quite small as tank sizeincreases. Recent data taken on Saturn tanks,however, indicate that such mass transfer maybe larger than is _enera]ly assumed. In anyevent, the analytical prediction of mass transferis perhaps the most difficult task in determiningpressurant requirements. The transient natureof the processes occurring, as well as the un-known pressurant and vapor distributionthroughout the ullage space, are only two of themasons for the task being difficult. Futuremathematical models based on analyticalstudies must of necessity include considerationsof mass transfer, with experimental tests beingrequired to establish the actual magnitude andmode of such mass transfer, especially for largetanks.Because of the transient nature of the various

    processes, with system parameters varying inboth time and space, temperature stratificationoccurs during expulsion. The stratifica.tionwithin the liquid causes a warm layer of liquidto form at the interface, although the tempera-ture of the bulk of the liquid remains relativelyconstant. Temperature stratifica.trion that oc-curs within the ullage space varies from thewarm liquid temperature at the inCerface to thegas inlet temperature at the top of the tank.There are also concentration gradients through-out the ullage space of the inlet pressurant, pres-surant used in previous expulsions, and propel-lant vapor. These gradients are dependentupon diffusion effects and degree of mixing,variables not subject to ready analysis, as wellas history eff_ts (i.e., time of pressurizationprior to expulsion, heat transfer during coast,

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    9/58

    FLUID PHYSICS OF LIQUID PROPELLANTS

    venting during coast, pressurant added in pre-vious firings of the engine system, etc.).

    :ks can be appreciated, an analytical approachto the problem of predicting pressurant require-ments is extremely difficult. In general, thevariables affecting the heat- and mass-transferprocesses, and thus influencing pressurant re-quirements, are (1) pressurant type and state,(9) propellant type and state, (3) propellanttank characteristics, (4) expulsion character-istics, (5) history of the tank and contents, (6)slosh, vortexing, and diffuser design, and (7)gravity field. Several methods have been ad-vanced for predicting pressurant requirements;however, they are limited in application, partic-ularly for space vehicles, either because of sim-plifying assumptions or because many of thevariables are neglected. Better methods, of amore generalized nature and verified by repre-sentative experimental data_ will be required inthe fu.ture in order that realistic space vehiclesmay be designed.Another problem concerning pressurization

    systems is the selection of a particular system.The general requirements for a pressurizationsystem for a space vehicle always include: (1)lightweight and (2) high reliability, and de-pending upon the mission, may include (3) re-start capability, (4) instant readiness, (5) long-term compatibility with the propellants, and (6)variable flow capability. Since a multitude ofsystems exist that potentially meet these re-quirements, some type of classification is re-quired for the process of selection. If systemsare classified according to source of pressurizingfluid, there are only four major classes.Roughly in order of decreasing system weightand increasing system complexity, the classes ofsystems are those that use

    (1) Stored gases(9) Evaporated propel]ants(3) Evaporated nonpropellants(4) Products of a chemical reactionThe first class is one of the most common,

    being used in some form in most of the present-day liquid-propellant rocket vehicles. Storedhelium system, s typical of this class have highreliability; compatibility is not a problem; andthere is _ backlog of experience on components.

    Some of the parameters influencing the weightof the system are: (1) storage temperature, (2)storage pressure, (3) type of g_ expansion, aad(4) material used for the storage tank.The second class, using evaporated propel-

    lants, is the next most common type of system.This type of system has been used on severalpump-fed vehicles where the liquid, usuallyliquid oxygen, is tapped off downstream of thepump, vaporized in some convenient heat ex-changer, and then used to pressurize the propel-lant tank from which it was obtained. Systemsfor pressure-fed vehicles have not been devel-oped to any extent, but they offer great poten-tial in weight savings (especially for liquidhydrogen) and should be studied further.Evaporated-nonpropellant systems_ the third

    class, have not been tested extensively. Indeed,only two types of systems could be contemplatedas being useful for cryogenics, those usingliquid helium and liquid nitrogen. Each ofthese has certain disadvantages, however, thatwould generally preclude their use.The last class, systems using products of a

    chemical reaction, perhaps offers the greatestpotential in future weight savings. Gas gen-erators to generate gaseous products for use inpressurizing storable propellant tanks havebeen developed and flown. Another method,having undergone a small amount of testingrecently, involves the injection of a hypergolic(spontaneously ignitible) fluid into one of thepropellant tanks. This permits the reaction tooccur within the tank. The method is attrac-tive for the liquid-hydrogen tank in particu-lar, since vaporized liquid hydrogen, if obtain-able, has a lower gas density than the reactionproducts. Although the method is basicallysimple, it does present problems of control, es-pecial ly of temperature.The selection of a system from the four

    classes for a particular space vehicle dependsnot only on the requirements just. discussed, butalso on the tradeoffs involved with other sys-tems. The development of the more advancedsystems, particularly in the areas of controls,heat exchangers, components, and materials,will be desirable in order to give a more validbasis for the tradeoffs involved.

    9

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    10/58

    CHEMICAL ROCKET PROPULSION

    SLOSHING OF PROPELLANTSThe second major problem area influencing

    the use of liquid propellants in rocket vehiclesis propellant sloshing. Oscillations may result,for example, from the attitude control systemand can exert forces and moments on the vehiclethat cause a shift in the center of gravity.Figure 37-3 shows a schematic of a vehicleundergoing sloshing. The propellant sloshingexerts side forces which, at a maximum, mayreach a force approximately equal to one-quar-ter of the apparent weight of the propellant.These side forces not only increase structuralloads on the tankage and supports, but also re-quire thrust-vector control f_om the rocketengine. In extreme cases, critical disturbancesto t3id stability of the vehicle may occur. Fur-thermore, when sloshing occurs, the pressuriza-tion gas entering the tank may be cooled morerapidly, and thus more gas would be required.Still another effect might be the exposure ofthe outlet, to vapor. This would result in vaporentering the flow systems, where it could havedisastrous effects. The alleviation or elimina-tion of sloshing is, therefore, extremely de-sirable.Fig_lre 37-4 shows the slosh-force parameter

    as a function of the frequency parameter for ahalf-full spherical tank. These parameters arenondimensionalized and are independent oftank size or density of the liquid. The solidcurve represents the theoretical results for un-

    iFiGum_ 37-3.--Forces exterted on vehicle during

    sloshing.

    PARAMETER

    THEORY,(UNRESTRICTEDNONVISCOUS FLUID}_

    t0-- "_ VISCOSITY.8 _-_1.CENTIPOtSE

    I I"-_ I UNRESTRICTEDI I "_! BAFFLES

    FORCE 6 Ia._/'_f I EXPULSION BAG

    0 .4 .8 1.2 1.6 2.0 2.4 2.8FREOUENCY PARAMETER cs-nu.,i

    FIGURE 37-4.--Effects of control devices on slosh forcesin half-full spherical tank,

    restricted nonviscous sloshing. The effects ofvarious methods of slosh control are shown, andthey are best compared at the first natural fre-quency at which the largest slosh forces arise.The first natural frequency is the point at whichthe excitation and fundamental frequencies ofthe contained liquid are equal.The most common method to date of con-

    trolling sloshing is by placing baffles within thetank. _Vhile effective, these are seen to be theworst of three available methods. The secondmethod utilizes a continuous surface film, suchas obtained by expulsion bags or diaphragms.This method is more effective than baffles, butinvolves problems of compatibility, leakage,venting, fabrication, and other mechanicaldifficulties, particularly for large tanks. Sinceexpulsion bags also permit displacement ofliquids under zero gravity, however, such anapproach has been used for relatively small-scale propellant, tanks--slosh control beinggained as a secondary feature. The thirdmethod is the most effective and involves in-creasing the viscosity of the liquid ; for the caseshown in figure 37-4, the viscosity was increasedfrom 1 to 920 centipoise. Successful attemptshave been made at increasing the viscosity of so-called Earth storable propellants by gelling thepropellants and making a thixotropic materialout of the liquid; however, very little progn*esshas been made in such gelling for cryogenicpropellants. Indeed, the gelling of cryogenicsremains as one of the few fields in which largeadvances may still be made.

    10

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    11/58

    FLUID PHYSICS OF LIQUID PROPELLANTSTHERMAL CONTROL OF PROPELLANTS

    The third major problem area is thermal con-trol of the propellants. For cryogenic liquidsthe incoming thermal energy must be mini-mized since such heat input will cause propel-lant heating, vaporization, and a resulting lossof propellant by venting. Unless these lossesare small, the potential advantages of usingcryogenic propellants will vanish.The sources of propellant heating may be

    either internal or external with respect to thespace vehicle, as indicated by figure 37-5. Thespace vehicle is assumed to comprise four sec-tions. The payload section contains the instru-mentation and guidance equipment and mustbe maintained at a temperature on the order of520 R. The liquid-oxygen section requires atemperature of approximately 163 R, whilethe liquid-hydrogen tank must be maintainedat approximately 37 R; the engine tempera-ture will depend on the type of engine system.These temperatures may vary somewhat, de-pending upon the pressure of the propellant,the type of instrumentation, and the specificdesign of the vehicle, but the relative order ofmagnitude will remain the same.

    Heating due to internal sources is caused bythermal radiation between the various compo-nents of the vehicle and by conduction throughpropellant lines and structural supports. Thepossible external sources of heat are the sun,the planets and their moons, and the galaxies.Heat is transferred between the vehicle andthese sources solely by thermal radiation. Thelargest external heat flux encountered by a space

    GALACTICADIATION SOLAR'_ RADIATIONPAYLOAD_ _'

    _[__--ENGINEPLANETARYRADIATION

    C$-2Sk&O

    FIGURE. 37-5.--Sources of heat input to cryogenic vehiclein space.

    vehicle within our solar system is usually thatwhich originates from the sun. The heat fluxreceived from a planet results partly fromplanetary radiation and partly from reflectedsolar radiation. This planetary heat flux be-comes greater as the planet, is approached, and,in some cases, could become of the same order ofmagnitude as the solar heat flux. Galacticheating is negligibly small and is usuallyignored.In general, interstage conduction is controlled

    by using minimum areas and low-conductivitymaterials for all components attached to thetank. Thus, the major problem of heat inputis one of proper control of thermal radiationdue both to internal and external sources. Vari-ous methods of control of thermal radiation areavailable, and they may be classified as orienta-tion, coatings, and insulations.

    By proper orientation of the vehicle certainsections can shade other sections. For example,the payload could be pointed toward the sun,thus acting as an effective external sun shadowshield for the rest of the vehicle. Also, ar-rangement is important since the oxygen tankcould act as a shadow shield for the hydrogentank, protecting it against internal thermalradiation from the payload. In any case, how-ever, some other method of thermal radia-tion control must still be employed since sucharrangement or orientation will leave at leastone section at too high a temperature for effec-tive use.Various inorganic or organic coatings are

    available which, by the proper balance of emis-sivity and absorptivity, permit some equilib-rium temperature to be attained by the surfaceon which the coating is applied. The lowesttemperature obtainable, however, by presentlyavailable coatings, is still much higher thanthe temperatures of the cryogenic liquids.Hence, while useful, coatings will not aloneprovide the proper temperature environmentand must be supplemented by other methods ofthermal radiation control.The best means of control is the use of insula-

    tions. Figure 37-6 shows a comparison ofvarious insulations that are useful for spacevehicles. The product of apparent thermal con-ductivity and density is used as the means of

    _ess2 o--e2_2 1 1

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    12/58

    CHEMICAL ROCKET PROPULSION

    i0-1

    iO'_APPARENT THERMAL

    CONDUCTIVITYx DENSITY, 10-3

    I0-5

    _EVACUATED_POWDERS

    _'_MULTI LAYER-"- _RADIATION SHIELDS

    $-Z5_.50

    FIGURE 37-6.--Comparison of insulations for spaceapplications.

    comparison since this product arises when thebasic heat-transfer equation for conduction ismodified to reflect insulation weight. Optimuminsulation occurs when this product is a mini-mum. The apparent thermal conductivity isdetermined from measurements of the net heatflux across a thickness of insulation, and thusincludes conduction, radiation, and convectionif applicable. Foams are typical of those in-sulations that restrict the flow of heat primarilyby means of low thermal conduction alone.Evacuated powder-type insulations not onlyhave low conduction and very little convection,but also reduce tile radiant energy transfer andthus result in a much lower apparent thermalconductivity than foams. At present, however,the best of the insulations consists of parallel,thermally isolated, reflected shields. Suchshields are more effective than powders in re-ducing radiant heat transfer, and thus are betterinsulations for thermal control of propellantsunder space conditions. If close together, theseshields are commonly called foils. The so-calledsuperinsu]ations consist of reflective foils sepa-rated by very low-conduction materials.In use, foams are relatively easy to apply to

    the vehicle tanks but give high weights of insu-lation required. The powders require a rigidjacket around the tank for correct placement,which increases the weight and causes problemsof settling and the occurrence of voids. Foilsalso have an installation problem, as compres-sive loads reduce their efficiency. Trade-offs,therefore, between the various insulations mustbe made for each particular mission, as well asfor the proper use of each of the methods ofthermal control. At present, however_ the ef-fects of compressive loads, and techniques ofdesign, such _ fastening, supporting, and evac-

    uation, require more work before valid trade-offs may be performed.

    EFFECTS OF ZERO GRAVITYThe final problem area to be discussed con-

    cerns the effects of weightlessness (i.e., zerogravity) on the liquid propellants. Weightless-ness will be encountered by space vehicles dur-ing orbital or space missions, and a knowledgeof the final equilibrium liquid-vapor interfaceconfiguration will be needed to solve the prob-lems of effective tank venting and proper pumpinlet design. Furthermore, a knowledge of themodes of heat transfer to stored propellant.qduring periods of weightlessness will also boneeded.

    Control of InterfaceFor effective control of the liquid-vapor in-

    terface the various forces predominating in azero-gravity environment must be understood.Such forces are generally intermolecular innature; that is, they are relatively small, areexerted over a short range, and are independentof the gravitational field. One such force thatshould persist in zero gravity is the surfacetension of the liquid, or, a property associatedwith surface tension, the contact angle at whichthe liquid meets the solid surface (measured inthe liquid). Another consideration for a zero-gravity environment actually is based on theobserved characteristic of systems to assume astate of minimum energy when external forcesare removed. This would suggest that the in-terfaces, if free to move, would assume some

    SOLIDN0 o

    WN

    _5 o

    HIG

    CONTACT155 ANGLE

    FI(_URE ,_7-7.--Interface configurations in cylindricalcontainer as function of gravity field and contactangle.

    12

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    13/58

    FLUID PHYSICS OF

    spherical shape, this being the shape representa-tive of minimum energy.Figure 37-7 indicates the effect of combining

    these two conditions for liquid in a cylindricalcontainer. Both the 1-g and zero-g configura-tions are shown, for a range of contact angles.For the 1-g case the contact angle of the liquidis observed in the meniscus at the liquid-solidinterface. Upon entering the zero-gravitystate, the contact angle will be preserved, andthe liquid-vapor interface s being free to move,will adjust itself until a constant curvature sur-face is reached. In an enclosed container_ ofcourse, the actual shape of the liquid surfacewould not only be determined by contact anglepreservation and considerations of minimiza-tion of energy, but also by the shape of the con-tainer and the amount of liquid in the container.To verify the reasoning used in this analysis

    many experimental tests have been conducted atthe Lewis Research Center, primarily in a 2.25-second-free-fall drop tower. These tests haveverified the predictions in every detail.Selected photographs taken from motion-pic-ture data (using high-speed photography) fora typical test drop are shown in figure 37-8.This figure shows a 100-milliliter glass spherewith a liquid- to tank-volume ratio of 40 per-cent. The liquid in this ease is ethyl alcohol,which has a contact angle of zero and is thusrepresentative of the cryogenic fluids. Thephotographs represent the conditions at 1 gand the equilibrium conditions for zero g. Itcan be seen that, at 1 g, the alcohol rests in apool at the bottom of the sphere. The liquidsurface is essentially fiat. The meniscus is con-cave, and the liquid contacts the wall at, an angleof 0% During the zero-g period, tile liquid be-gins to rise along the wall, and the liquid sur-face becomes curved. At the equilibrium condi-tion the result is a completely wetted tank wallwith a spherical vapor bubble in the interior ofthe liquid.The results of the experimental studies, as

    l)reviously indicated_ have tended to verify theprediction that the total surface energy isminimized in zero gravity. In order to providefurther verification, a definitive analyticalmodel was studied. This model used a geom-etry consisting of a hollow tube suspended

    LIQUID PROPELLANTS

    1 G

    13

    C-62254FI6URE 37-8. --Inter face configurations of wettingliquid in spherical container at normal gravity andduring weightlessness. Tank 40 percent full.

    within a cylindrical tank, aligned with thevertical axis, and being open to the liquid atboth top and bottom. An analysis of this con-figuration_ solving for the minimization of sur-face energy, indicated that for liquids with con-tact angles less than 90 the liquid will rise inthe tube if the tube radius is less than one-halfthe tank radius, will remain level if the tuberadius is equal to one-half the tank radius, and

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    14/58

    CHEMICAL ROCKET PROPULSION

    :iii_!i_!!ii!!i!iiii!_RT.FIGURE 37-9.--Effect of tube radius on interface con-figuration for wetting liquid in cylindrical container.

    will fall if the tube radius is greater than one-half the tank radius.Experimental tests have been conducted with

    wetting liquids in a similai" geometry, in whichthe tube radius was varied and the direction ofliquid motion was observed as the experiment,was subjected to zero gravity. Results of theseexperiments are summarized schematically infigure 37-9 for a liquid with a contact angle of0. Sketches of the interface configurations areshown after equilibrium conditions in zerogravity have been reached. Three tube radii(Ro) were used, being respectively less than,equal to, and greater than one-half the ta_kradius (RT). These experiments verified theprevious analysis in every detail and providedproof that the solid-liquid-vapor system willtend to assume a configuration of minimumfree surface energy when placed in a weightlessenvironment. APPLICATIONSAs a result of these fundamental studies, it

    appears that the tank systems can be designedby using suitable internal baffling to positionthe liquid-vapor interface properly. The prin-cipal requirements of an interface control sys-tem in a propellant tank are that it maintain aquantity of liquid over the pump inlet, and, ifventing is necessary, that it maintain the mainullage volume over the vent. The capillarytube geometry employed in the minimum-sur-face-energy studies appears to provide a satis-factory interfa_ configuration in that it locatesliquid over the pump inlet if such a pump inlet,is placed at the bottom of the capillary tube and

    vapor over the vent if the vent is located abovethe tube.The configuration of the liquid-vapor inter-

    face in zero gravity, obtained with a standpipegeometry for two orientations of the container,is shown in figure 37-10. This figure shows a100-milliliter glass sphere with a liquid- totank-volume ratio of 30 percent. Again ethylalcohol (containing a dye for ease of photog-raphy) is used in order to duplicate cryogenicliquids having a contact angle of zero. Thecapillary tube (or standpipe) has a tube diam-eter of 33 percent of sphere diameter and arelative height of 60 percent. The photo-graphs in figure 37-10(a) indicate the abilityof the liquid to fill the standpipe and thevapor to form over the vent (if located abovethe standpipe) for the case in which the sphereenters zero gravity at an angle of 0 betweenthe gravity vector and the axis of the stand-pipe. Figure 37-10(b) is similar, except thatthe sphere enters zero gravity at an angle of90 , and again the ability to position the liquidand vapor bubble is shown. Such ability, how-ever, is very time dependent, and for low fill-ings a relatively long period of time is re-quired to properly position the vapor bubbleafter zero gravity has been entered at a highangle to the initial gravity vector.The standpipe ge_)metry encounters difficul-

    ties in positioning the vapor bubble over thevent at high fillings, however, when the vaporbubble is small. If the standpipe is made rela-tively high, there is danger that the bubblecould becometrapped in the annular space fol-lowing an adverse acceleration. A geometrythat avoids entrapment of the vapor bubble andurges it to the vicinity of the vent or, conversely,pmnps the liquid to the bottom of the tankwould be desirable. Liquid can be pumped bycapillary forces from one end of a tank to theother if tapered geometry is used and suitablereturn paths are provided. This principle maybe employed to overcome some of the difficultiesencountered with standpipes at high fillings.The application of this principle to a sphericaltank is shown in figure 37-il. The taperedsection is formed in this case by offsetting asecond sphere within the first. The geometryis now such that the vapor bubble obtains its

    14

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    15/58

    FLUID PHYSICS OF LIQUID PROPELLANTS

    I G, 0 SEC

    0 G; 0.5 SEC 0 G; 0.5 SEC_

    I0) 0 G; 1.5 SEC 0 G; 1.5 SEC C-62255(a) Initial position, 0 between gravity vector and standpipe.(b) Initial position, 90 between gravity vector and standpipe.

    FzGU_ 37-10.--Movement of wetting liquid in capillary tube ullage control geometry.

    15

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    16/58

    CHEMICAL ROCKET PROPULSION

    LIO I G; 0 SEC_

    C$-Z5$2_(a) 1-g configuration.

    (b) Zero-g configuration.FZOVaE 37-11.--Application of tapered section principle

    t_) spherical tank.

    most nearly spherical shape under the vent butcannot become completely spherical even forfillings of 95 percent. The result is the move- 0 G; 0.8 SECment of the bubble to the vent from any start-ing position. The ability of this geometry tomove the vapor bubble is shown in figure 37-12;the vapor bubble is initially 90 away from thevent but is pumped toward the vent by capillaryaction when zero gravity is entered. In thisexperiment, the outer sphere has a volume of100 milliliters. The inner sphere is 60 percentof the diameter of the outer sphere, whichforms a tapered annular section around theinner sphere. The liquid is again ethyl alco-hol, and the liquid- to tank-volume ratio is 95percent.While the experiments to date have indicated

    that the liquid-vapor interface can be ade-quately controlled, two questions remain to be 0 G; I./,answered. The scaling parameter of time mustbe further studied, and the amount of accelera-tion to disrupt a surface under zero gravitymust be known. Acceleration disturbances(crew movement, orientation maneuver_ etc.)may cause the liquid and vapor to be improp-erly oriented, and thus an unknown length oftime may be required for proper positioning.

    Heat TransferAnother problem on which further research

    is needed is that of heat transfer under weight-less conditions. The mechanism of heat inputto a tank under space conditions correspondsto that of a classical pool-boiling situation, thevarious regions of which are shown in figure37-13 for a 1-g field. The coordinates are heat

    C-62255FIGURE 37-12.--Movement of vapor bubble in tapered-section ullage control geometry. Initial position, 90 from vent; wetting liquid.

    16

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    17/58

    FLUID PHYSICS OF LIQUID PROPELLANTS

    flux plotted as a function of temperature dif-ference.The heat-transfer regions start with conduc-

    tion at very low flux, and, as flux is increased,progress through convection and nucleation,and finally, at very high flux, to the mechanismof film boiling. The regions of convection andnucleation may be substantially affected by theloss of gravity. Convection currents shouldcertainly be absent in zero-gravity environ-ment, although residual currents might persistfor some time; however, the nucleation pictureis less clear. The velocity of bubbles throughtile liquid depends on the gravity field. If thevelocity becomes zero because of loss of buoy-ancy, the nucleation region may be replaced bya film mechanism of boiling. This situationwould be predicted by a proposed nucleationconcept, that is gravity dependent. A nuclea-tion model has been proposed, however, whichis independent of gravity in that bubbles areremoved from the surface by a flow field set uparound tile bubble because of its rate of expan-sion. If this model is correct, the nucleationregion should be relatively unaffected by a zero-gravity environment, and this region wouldprobably extend down to cover the convectionregion.A series of experiments using liquid hydro-

    gen has been performed by General Dynamics/Astronautics in which the heat transfer in thenucleate range as a fmlction of temperaturedifference was measured at 1-g and zero-g con-ditions. Results of this study are shown infigure 37-14. These results indicate that zero

    HEAT FLUX,q/A

    (LOG SCALE)

    r-CONDUCTION-CONVECTION_NUCLEATE

    fTEMP DIFFERENCE, THEATER--TBuLK (LOG SCALE)

    FIG{TEE 37-13.--Typical pool boiling charcteristics for1-g field.

    HEATFLUX.BTUHR/SQ FT

    20,C0tO,O00

    400C200CIOOC

    40C20C0

    O/o H TESTS

    1_ t i t I l2 4 6 8 rO t2TEMP DIFFERENCE, R cs-ts_o5

    FIOURE 37-14.--Comparison of 1-g and zero-g liquid-hydrogen heat-transfer characteristics for nucleateboiling range. Small, directly heated metal speci-mens in pool of liquid hydrogen.

    gravity has little effect on heat transfer in thenucleate region, which supports the nucleationmodel that is independent of gravity. Thescaling parameter of time, however, amongothers, is probably quite important, in theseexperiments, and much more work in this areais needed in order to extend our knowledge ofheat transfer processes to the zero-gravitysituation.The primary effect of the heat-transfer mech-

    anism for cryogenic liquids is the pressure risewithin the container. This problem could beserious in a long term coasting situation. In anormal gravity field, convection currents andhigh bubble velocities cause most of the heat to13o delivered to the gas space, which causes anabnormal pressure rise and thus excessive vent-ing. The heat sink capacity of the liquid,which is considerable in the case of cryogenicpropellants, is lost. In a zero-gravity field, ifthe conduction and nucleation regions are re-placed by a film process, again most of the heatwill be delivered to the gas, and the heat sinkcapacity of the liquid will be lost. If the grav-ity-independent nucleation concept is correct,however, the nucleation continues, but at re-duced bubble velocity. Thus more time isavailable for bubbles to condense as they movethrough the liquid, more heat is delivered tothe liquid, and a lower pressure rise results.Since the available data tend to su,pport thegravity-independent nucleation concept, itwould appear that less pressure rise will beexperienced in zero gravity than under normal

    17

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    18/58

    CHEMICAL ROCKET PROPULSION

    gravity. The magnitude of this effect is stillrelatively unknown, however, and warrantsfurther study.

    CONCLUDING REMARKSIn summary, four problem areas primarilyinvolving the use of cryogenic liquids in space

    vehicles have been briefly discussed. Some an-swers exist for these areas, but further work,of both a fundamental and an applied nature, isrequired. In particular, more detailed in-

    formation is needed on the physical and thermo-dynamic properties for the liquids of interest.Furthermore, the areas of heat and mass trans-fer need better definition under conditions pe-culiar to a space vehicle.The problem areas of fluid physics as outlined

    are not the only areas of interest today, and,as space flights become more ambitious, evenmore areas requiring study will appear. In-deed, the successful operation of space vehiclesmust depend upon future solutions to these andother problems.

    BIBLIOGRAPHYPressurization

    AEPACI, V. S., and CLARK, J. A. : Dynamic Response of Fluid and Wall Temperatures Dur-ing Pressurized Discharge for Simultaneous, Time-Dependent Inlet Gas Temperature,Ambient Temperature, and/or Ambient Heat Flux. Vol. 7 of Advances in Cryogenic Eng.,K. D. Timmerhaus, ed., 1961, p. 419.

    ARI'ACI, V. S., CLARK, J. A., and WIriER, W. O. : Dynamic Response of Fluid and Wall Tem-peratures During Pressurized Discharge of a Liquid from a Container. Vol. 6 of Advancesin Cryogenic Eng., K. D. Timmerhaus, ed., Plenum Press, 1960, p. 310.

    B0WERSOCK, D. C., GARDNER, R. W., and REID, R. C. : Pressurized Transfer of CryogenicLiquids. Vol. 4 of Advances in Cryogenic Eng., K. D. Timmerhaus, ed., Plenum Press,1958, p. 342.

    BOWE_S0CK, D. C., and RF_ID, R. C. : An Analytical Method for Estimating Gas Requirementsin the Pressurization and Transfer of Cryogenic Fluids. Vol. 6 of Advances in CryogenicEng., K. D. Timmerhaus, ed., Plenum Press, 1960, p. 261.

    CLARK, J. A., VAN WYLEN, G. J., and FEASTER, S. K. : Transient Phenomena Associated withthe Pressurized Discharge of a Cryogenic Liquid from a Closed Vessel. Vol. 5 of Advancesin Cryogenic Eng., K. D. Timmerhaus, ed., Plenum Press, 1959, p. 467.

    CoxE, E. F., and TATOM, J. W. : Analysis of the Pressurizing Gas Requirements for anEvaporated Propellant Pressurization System. Vo]. 7 of Advances in Cryogenic Eng.,K. D. Timmerhaus, ed., Plenum Press, 1961, p. 234.

    FENSTER, S. K., VAN WYLEN, G. J., add CLARK, J. A. : Transient Phenomena Associated withthe Pressurization of Liquid Nitrogen Boiling at Cmmtant Heat Flux. Vol. 5 of Advancesin Cryogenic Eng., K. D. Timmerhaus, ed., Plenum Press, 1959, p. 226.

    GLVCK, D. F., and KLINE, J. F. : Gas Requirements in Pressurized Transfer of Liquid Hydr(_gen. Vol. 7 of Advances in Cryogenic Eng., K. D. Timmerhaus, ed., Plenum Press, 1961,p. 219.

    KAPLAN, C. : Selection of Pressurization System for Storable Liquid Propellant Rocket En-gines. ARS Jour., vol. 31, no. 6, June 1961, pp. 786-793.

    NEIN, M. E., and HEAD, R. n. : Experiences with Pressurized Discharge of Liquid Oxygenfrom Large Flight Vehicle Propellant Tanks. Vol. 7 of Advances, in Cryogenic Eng.,K. D. Timmerhaus, ed., 1961, i). 244.

    SCH.MmT, A. F., PURCELL, J. R., WILSON, W. A., and SMITh, R. V. : An Experimental StudyConcerning the Pressurization and Stratification of Liquid Hydrogen. Vol. 5 of Advancesin Cryogenic Eng., K. D. Timmerhaus, ed., Plenum Press, 1959, p. 487.

    SUTTON, J. F., et al. : Main Propellant Tank Pressurization System Study and Test Program.Rep. ER4728, Lockheed Aircraft Corp., Marietta, Georgia, Feb. 1961.VAN WYLEN, G. J., FENSTER, S. K., M_aTE H., Jm, and WARREN, W. A. " Pressurized Dis-charge of Liquid Nitrogen from an Uninsulated Tank. Vol. 4 of Advances in CryogenicEng., K. D. Timmerhaus, ed., Plenum Press, 1958, p. 395.

    WALSH, THOMAS Y., ttlDBARD, R. R., and ORDIX, PAt_L M. : Dillution of Liquid Oxygen whenNitrogen is Used for Pressu ri zati on. NACA RM E58AO3a, 1958.

    18

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    19/58

    FLUID PHYSICS OF LIQUID PROPELLANTSAnon. : Main Propellant Tank Pressurization System Study and Test Program.5238, Lockheed Aircraft Corp., Marietta, Georgia, Aug. 1961.

    Anon. : Main Propellant Tank Pressurization System Study and Test Program.5296, vol. III. Lockheed Aircraft Corp., Marietta, Georgia, Dec. 1961.

    Rep. ER-

    Ret). ER-

    SloshingARRAMSON, H. NOR-_IAN, and RANSLEBEN, GUIDO E., JR. : A Note on the Effectiveness of Two

    Types of Slosh Suppression Devices. Tech. Rep. 6, Southwest Res. Inst., June 15, 1959.AaRAMSON, H. NORMAN, and RANSLE,EN, GUZDO E., JR. : A Note on Wall Pressure Distribu-

    tions During Sloshing in Rigid Tanks. Tech. Rep. 5, Southwest Res. Inst., June 15, 1959 .ARRAMSON, H. NORMAN, and RANSLERE_, GUZDO E., JR. : Simulation of Fuel Sloshing Char-

    acteristics in Missile Tanks by Use of Small Models. Tech. Rep. 7, Southwest Res. Inst.,Apr. 25, 1960.

    BAUER, IIELMUT F. : Fluid Oscillations in a Circular Cylindrical Tank. Rep. DA-TR-1-58,Army Ballistic Missile Agency, Apr. 18, 1958 .

    BAUER, HELMUT F. : Fluid Oscillation in a Cylindrical Tank with Damping. Rep. DA-TR-4-58, Army Ballistic Missile Agency, Apr. 23, 1958.

    BAUER, HELMUT F. : Propellant Sloshing. Rep . DA-TR-18-58, Army Ballistic MissileAgency, Nov. 5, 1958.

    BUDrANSKY, BERNARD : Sloshing of Liquids in Circular Canals and Spherical Tanks. Jour.Aero/Space Scl., vol. 27, no. 3, Mar. 1960, pp. 161-173.

    COLE, tIENRY A., JR., and GAMBVCCZ, BRYN0 J. : Tests of an Asymmetrical Baffle for Fuel-Sloshing Suppression. NASA TN D-1036, 1961.

    LAMB, HORACE : Hydrodynamics. Sixth ed., Dover Pub., 1945.LEONAaD, H. WAYNE, and WAIJrON, WILLIAM C., JR. : An Investigation of the Natural Fre-quencies and Mode Shapes of Liquids in Oblate Spheroidal Tanks. NASA TN D-904,1961.

    5IcCARTY, JOHN LOCKE, LE01_ARD, H. WAYNE, and WALTON, WILLIAM C., JR. : ExperimentalInvestigation of the Natural Frequencies of Liquids in Toroidal Tanks. NASA TN D-531,1960.

    McCARTY, JOHN LOCKE, and STEPHENS, DAVID G. : Investigation of the Natural Frequenciesof Fluids in Spherical and Cylindrical Tanks. NASA TN D-252, 1960.

    O'NEILL, J. P. : An Experimental Investigation of Sloshing. STL-TR_5_, SpaceTech. Labs., Inc., Mar. 4, 1960.

    STEPHENS, DAVID G., LEONARD, 1:_. WAYNE, and SILVERIA, MILTON A. : An Experimental In-vestigation of the Damping of Liquid Oscillations in an Oblate Spheroidal Tank Withand Without Baffles. NASA TN D-808, 1961.

    STOFAN, ANDREW _., and ARMSTEAD, ALFRED L. : Analytical and Experimental Investigationof Forces and Frequencies Resulting from Liquid Sloshing in a Spherical Tank. NASATN D-1281, 1962.

    STOFAN, ANDREW J., and PAUL L ALi_ERrr J. : Experimental Damping of Liquid Oscillations ina Spherical Tank by Positive-Expulsion Bags and Diaphragms. NASA TN D-1311, 1962.

    Thermal ControlBALLIB;OER, J. C. : Environmental Control Study of Space Vehicles, Pt. 1. Rep. ERR-AN-004, General Dynamics/Astronautics, Mar. 9, 1960.

    BALLINOEa, J. C. : Environmental Control Study of Space Vehicles, Pt. 2. Rep. ERR-AN-004, General Dynamics/Astronautics, Nov. 1, 1960.

    BALLINOER, J. C. : Environmental Control Study of Space Vehicles, Pt. 2. Supp. A, Rep.ERR-AN-O16, General Dynamics/Astronautics, Dec. 31, 1960.

    BALLINOER, J. C. : Environment Control Study of Space Vehicles, Pt. 2, Supp. B, Rep.ERR-AN-016, General Dynamics/Astronautics, Jan. 20, 1961.

    BALUNOER, J. C.: Environmental Control Study of Space Vehicles, Pt. 3. Rep. ERR-AN-051, General Dynamics/Astronautics, Dec. 31, 1960.

    BURRY, R. V., and DE_NER, V. R. : Liquid-Propellant Storage Evaluation for Space Vehicles.Fourth Symposium on Ballistic Missile and Space Tech., Aug. 1959.

    CHRISTIANSEN, R. M., HOLLINGSWORT]_, M., JI_., and MARSH, H. N., JR. : Low TemperatureInsulating Materials. 1959 Cryogenic Eng. Confer., Sept. 2-4, 1959.

    CBAMEIL KENNETII R. : Orbital Storage of Cryogenic Fluids. TN 58-282, WADC, Oct. 1958.

    19

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    20/58

    CHEMICAL ROCKET PROPULSIONCORNOG, ROBERT : Temperature Equilibria in Space Vehicles. Proc. Am. Astronautical Soc.,Western Regional Meeting, Aug. 18-19, 1958.

    CUNNINGnAM, F. G. : Ear th Reflected Solar Radiation Input to Spherical Satellites. NASATN I)=1099, 1962.

    CUNN_NGHA_, F. G. : Power Input to a Small Flat Plate from a Diffusely Radiating Sphere,with Application to Earth Satellites. NASA TN D-710, 1961.

    DaISCOLL, D. G. : Cryogenic Tankage for Space Flight Applications. 1959 Cryogenic Eng.Conf., Sept. 2-4, 1959.

    GA2.LEY, CARL, JR., KELLOGG, W. W., and VESTINE, E. H. : Space Vehicle Environment. Jour.Aero/Space Sci., vol. 26, no. 12, Dec. 1959.

    GLASER, PETER E., ed. : Aerodynamically Heated Structures. Prentice-Hall, Inc., 1962.GOLDMAN, D. T., and SXNGER, S. F. : Radiation Equilibrium and Temperature. Proc. of VII

    Int. Astronautical Cong. (Rome), Sept. 17-22, 19.56.IIEI.LAR, GERHA_m : Thermal Control of the Explorer Satellites. Paper 834-59, ARS, 1959.T[NILICKA, _ILO P..* Engineering Aspects of Heat Transfer in Multilayer Reflective Insula-

    tion and Performance of NRC Insulation. 1959 Cryogenic Eng. Conf., Sept. 2-4, 1959.JOHNSON, FRANCIS S.: Satellite Environment Handbook. Stanford Univ. Press, 1961.JOHNSON, FRANCIS S. : The Solar Constant. Jour. of Meteorology, vol. 11, no. 6, Dec. 1954,pp. 431-439.

    KaErrH, F. : Radiation Heat Transfer and Thermal Control of Spacecraft. Pub. 112, Eng.Exp. Stat ion, Oklahoma State Univ., Apr. 1960.

    OLl_a, J. R., and DEMPSTZa, W. E. : Orbital Storage of Liquid Hydrogen. NASA TND-559, 1961.

    S(:OTT, RUSSELL B. : Cryogenic Engineering. D. Van Nostrand Co., Inc., 1959.S_fOLAK, G. R., and KNOLL, R. H. : Cryogenic Propellant Storage for Round Trips to Marsand Venus. Paper 60-23. Inst. Aero. Sci ., 1960.S_(OLAK, GEORGE R., KNOLL, RICHAaD H., and WALLNER, LEWIS E. : Analysis of Thermal-Protection Systems for Space-Vehicle Cryogenic-Propellant Tanks. NASA TR R-130,1962.

    WooD, GEOROE P., and CARTER, ARLEN F. : The Design Characteristics of Inflatable Alumi-nized-Plastic Spherical Earth Satellites with Respect to Ultraviolet, Visible, Infrared andRadar Radiation. Paper 59-AV-38, ASME, 1959.

    Zero GravityArAb, NEIL KENSINOTON: The PhysiCs and Chemistry of Surfaces. Oxford Univ. Press,1941.

    AI)A_fSON, ARTHUR W. : Phy_cal Chemistry of Surfaces. Intersci. Pub., 1960. AerophysicsGroup : June-August Progress report for the Combined Laboratory and KC-135 AircraftZero-G Test Program. Rep. AE61_0871, Convair Astronautics, Sept. 5, 1961.

    ALEXANDERI A. E. : Surface Chemistry. Longmans, Green, and Co., 1951.BARTWELL, F. E. : Wetting of Solids by Liquids. Colloid Chem., voI. III, Chem. Catalog Co.,Inc., 1931.

    BENEDIKT, E. T. : Dynamics of a Liquid Spheroid Under the Action of Cohesive Forces.Rep. ASRL-TM-60-10, Norair Div., Northrop Corp., Sept. 1960.

    BENEnIKT, E. T. : General Behavior of a IAqutd in a Zero or Near Zero Gravity Environment.Rep. ASG-T_I_()-9Z6, Norair Div., Northrop Corp., May 1960.

    B_NEDIKT, E. T. : Scale of Separation Phenomena in Liquids Under Conditions of NearlyFree Fall. ARS Jour., vol. 29, no. 2, Feb. 1959, pp. 150-151.

    BERNETT, ),_. K., and ZXSMA_, W. A. : Relation of Wet tabi li ty by Aqueous Solutions to theSurface Constitution of Low-Energy Solids.' Jour. Phys. Chem., vol. 63, no. 8, Aug. 1959,pp. 1241-1246.

    BIKERMAN, J. J. : Surface Roughness and Contact Angle. Jour. Phys. and Colloid Chem.,vol. 54, no. 5, May 1950, pp. 653--658.

    BRAZINSKY, IRVING, and WEIss, SOLOMON: X Photographic Study of IAqutd Hydrogen atZero Gravity. NASA TN D-1107, 1961.

    B_TRnON, R. S. : Surface Tension and the Spreading of Liquids. Cambridge Univ. Press(London), 1940.

    CALLAOHAN, EDMUND E. : Weightlessness. Machine Design, vol. 34, no. 24, Oct, 11, 1962,pp. 156-161.

    CASSIE, A. B. D. : Contact Angles. Discussions of Faraday Sot., no. 3, 1948, pp. 11-16.

    20

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    21/58

    FLUID PHYSICS OF LIQUID PROPELLANTSDUPRE, ATHANASE: Memoirs of the Mechanical Theory of Heat. Nos. 5, 6, and 7, Am_.

    ChJm. et Phys., 1866-1868.FORSTER, H. K., and ZUBEa, N. : Growth of a Vapor Bt_bble in a Superheated Liquid. four.

    Appl. Phys., vol. 25, Apr. 1954, pp. 474-478.FOWKES, FREDERICK M., and HARKINS, WILLIAM D. : The State of Monolayers Adsorbed at

    the Interface Sol,id-Aqueous Solution. four. Am. Chem. Soc., vol. 62, no. 12, Dec. 1940,pp. 3377-3386.

    GAuss, CARL FRIEDERICH: Principta Generalia Theorise Figurae Fluidorum in StatuAequilibrii. Gottingen, 1830.

    GERATHEWOHL, SIEGFKIED J. : Zero-G Devices and Weightlessness Simulators. Pub. 781,Nat. Acad. Sci.-Nat. Res. Council, 1961.

    GIBBS, J. WILLARD: Collected Works. Voh 2. Longmans, Green, and Co., 1928.GLASSTONE, SAMUEL: Textbook of Physical Chemistry. Second ed., D. Van Nostrand Co.,

    Inc., 1946, pp. 481--483.GOOD, ROBERT J., and GIalFALCO, L. A. : A Theory for Estimation of Surface and InterracialEnergies. III. Estimation of Surface Energies of Solids from Contact Angle Data.four. Phys. Chem., vol. 64, no. 5, May 1960, pp. 561-565.

    GOODRICH, f. C. : The Mathematical Theory of Capillarity, I; II; III. Proc. Roy. Soc.(London), ser. A, vol. 260, no. 1303, Mar. 21, 1961, pp. 481-509.

    GnEOG, S. fl. : Hystersis of the Oontact Angle. four. Chem. Phys., vol. 16, no. 5, May 1948,pp. 549-550.

    GURNmL C.: Surface Forces in IAquids and Solids. Proc. Phys. Soc. (London), sec. A,no. 358, vol. 62, Oct. 1, 1949, pp. 639-648.

    HAnKINS, WILLIAM D. : The Physical Chemistry of Surface Films. Reinhold Pub. Corp.,1952.

    HARKINS, WILLIAM D., and LOESER, E. n. : Surfaces of Solids. XIX Molecular InteractionsBetween Metals and Hydrocarbons. flour. Chem. Phys., vol. 18, no. 4, Apr. 1950, pp.556-560.

    JONES, EDWARD W. : What Does "Weightlessness" Really Mean? Space/Aero., vol. 38, na5, Oct. 1962, pp. 65-67.

    KOENIO, FREDERICK O. : On the Thermodynamic Relation Between Surface Tension andCurvature. flour. Chem. Phys., vol. 18, no. 4, Apr. 1950, pp. 449-459.

    LANOMUIa, IRVINO : The Constitution and Fundamental Properties of Solids and Liquids, I.flour. Am. Chem. Soc., vol. 38, no. 11, Nov. 1916, pp. 2221-2295.

    LAPLACE, PIERRE SIMON : Meeanique Celeste, Supplement to Tenth Book, 1806.LI, T. : Cryogenic IAqnids in the Absence of Gravity. Vol. 7 of Advances in Cryogenic Eng.,

    K. D. Timmerhaus, ed., Plenum Press, 1962, pp. 16-23.LI, TA : Liquid Behavior in a Zero-G Field. Rep. AE 60-0682, Convair-Astronautics, Aug.1960.

    MCNUTT, f lAMES E., and ANDES, GEOROE M. : Relationship of the Contact Angle to InterfacialEnergies. flour. Chem. Phys., vol. 30, no. 5, May 1959, pp. 1300-1303.MERTE, H., and CLARK, J. A. : Boiling Heat Transfer Data for Liquid Nitrogen at Standardand Near-Zero-Gravity. Vol. 7 of Advances in Cryogenic Eng., K. D. Tlmemrhaus, ed.,1961, p. 546.

    MILLER, N. F. : The Wetting of Steel Surfaces by Esters of Unsaturated Fatty Acids.four. Phys. Chem., voh 50, no. 4, July 1946, pp. 300-319.

    NEINER, J. J. : The Effect of Zero Gravity on Fluid Behavior and System Design. TN 59-149, WADC, Apr. 1949.

    PARTINOTON, J. R. : An Advanced Treatise on Physical Chemistry. Vol. II. Longmans,Green, and Co., 1951, pp. 134-207.

    PETRASH, DONALD A., ZAPPA, ROBERT F., and OTTO, EDWARD W. : Experimental Study ofthe Effects of Weightlessness on the Configuration of Mercury and Alcohol in SphericalTanks. NASA TN D-1197, 1962.

    PLATEAU, J. A. F. : Statique Experimentale et Theorlque des Liquides. 1873.POISSON, SIMEON DENIS : Nouvelle Theorie de L'Action. 1831.RAYLEIOH : Collected Scientific Papers. Vol. VI. Cambridge Univ. Press (London), 1920.REYNOLDS, W. C. : Hydrodynamic Consideration for the Design of Systems for Very LowGravity Environments. Tech. Rep. LG-1, Dept. Mech. Eng., Stanford Univ., Sept. 1961.

    REYNOLDS, WILLIAM C. : Behavior of Liquids in Free Fall. Jour. Aero/Space Sci., vol. 26,no. 12, Dec. 1959, pp. 847-848.

    21

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    22/58

    CHEMICAL ROCKET PROPULSIONROHSENOW, W. M.: A Method of Correlating Heat-Transfer Data for Surface Boiling

    Liquids. Trans. ASME, vol. 74, 1952, p. 969.SCHOLI_ERa, HAROLD _[., and WETZEL, W. W. : S(_Ine Observations on the Theory of ContactAngles. Jour. Chem. Phys., vol. 13, no. 10, Oct. 1945, p. 448.

    SHAFRIN, ELAINE G., and ZIS_IA_, WrLLIA_[ A. : Constitutive Relations in the Wetting ofLow Energy Surfaces and the Theory of the Retraction .Method of Preparing Monolayers.Jour. Phys. Chem., vol. 64, no. 5, May 1960, pp. 519-524.

    SIIERLEY, J. E., and MERINO, F. : The Final Report for the General Dynamics/AstronauticsZero-G Program Covering the Period from May 1960 Through March 1962. Rep. AY62-0031, General Dynamics/Astronautics, Aug. 15, 1962.

    SHERLEY, JOAN: Evaluation of Liquid Behavior in a Zero-G Field Using Free FloatingTechnique. Convair-Astronautics, 1960.

    SIIUTTLEWORTH, R. : The Surface Tension of Solids. Prec. Phys. Soc. (London), sec. A,vol. 63, no. 3_, May 1, 1950, pp. 444457.

    SHIJTTLEWORTtt, n., and BARLEY, G. L. J. : The Spreading of a Liquid Over a Rough Solid.Discussions of Faraday Soc., no; 3, 1948, pp. 16-21.

    SIEOEL, ROI_ERT : Transient Capillary Ri_ in Reduced and Zero-Gravity Fields. Paper 60-WA-201, ASME, 1960.

    STEINLE, HANS: Heat Transfer in a Zero-G Field. Convair-Astronautics, 1960.USlSKIN, C. M., and SIEGEL, R. : An Experimental Study of Boiling in Reduced and ZeroGravity Fields. Paper 60-HT-10, ASME-AIChE, 1960.

    WASHBURN, E. ROGER, and ANDERSON, ELMER A. : The Pressure Against Which Oils WillSpread on Solids. Jour. Phys. Chem., vol. 50, no. 5, Sept. 1946, pp. 401-406.

    YARNOLD, G. D. : Understanding Surface Tension. Contemporary Phys., vol. 1, no. 4, Apr.!960, pp. 267-275.

    YARNOLD, G. D., and MASON, B. J. ; The Angle of Contact Between Water and Wax. Prec.Phys. Soc:i Sec. B, vol. 62, pt. 2, Feb. 1949, pp. 125-128.

    YARNOLD, G. D., and MASON, B. J.: A Theory of the Angle of Contact. Prec. Phys. See.,sec. B, vol. 62, pt. 2, Feb. 1949, pp. 121-125.

    YouN(L TgO_IAS: Essay on the Cohesion of Fluids. Roy. Phil. Trans., Dec. 20, 1805, pp.65-87.

    YUST, WALTER, ed. : Surface Tension. Encyclopaedia Britannica, Inc., vol. 21, 1947, pp.593-602.

    22

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    23/58

    38.NewProblemsncounteredithPumpsndTurbinesBy Melvin J. Hartmann and Calvin L. Ball

    MELVIN J. HART:bIAN'N" i8 Head o/ the Pump Section of the Fluid SystemComponents Division o/the NASA Lewis Research Center. He has conductedand supervised research in the field o/turbomachinery. Mr. ttartmann earnedhis B.S. degree in Mechanical Engineering from the University o/Nebraskain 19/t4. He is a member o/the Institute of the Aerospace Sciences.CALVIN L. BALL, Aerospa/ce Scientist at the NASA Lewis Research Center, isconducting research on turbomavhinery; primarily involving pumps/or cryo-genic )quids that are applicable to rocket engines. Mr. Ball attended MichiganCollege of Mining and Technology and received his B.S. degree in MechanicalEngineering in 1956. He is a member o/the American Society of MechanicalEngineers.

    INTRODUCTIONAs the missions in the space program become

    more difficult, a wide range of new requirementsare placed on all components of both chemicaland nuclear rocket engine systems. The adventof manned space flight has resulted in stringentrequirements on the reliability of all rocket com-ponents in an effort to eliminate catastrophicengine failure. The trend toward larger boosterengines and higher engine operating pressureshas required increased size and pressure produc-ing capability of the turbopump units. Thesmaller thrust engines used to provide thrust inspace as well as in landing and rendezvousmaneuvers require that the turbopumps per-form satisfactorily over a wide range of thrustlevels (engine throttlability) and, in addition,provide multiple restart capability. The pump-ing systems for these engines must respondnearly instantaneously to demands of the con-trol system even though tim propellants mayhave been stored for long periods in the spaceenvironment. For some of the new enginesthere is the requirement that the pump acceptthe propellants from the tank at or near theirboiling points.

    Further complexities exist in the developmentof pumping machinery for future rocket en-gines. Statistics have indicated that about one-third of the effort in a rocket engine develop-ment program is devoted to the turbopump.Because the development of the turbopump isa long process, it is often necessary to establishmany of the engine parameters long beforeturbopump test data are available. If the tur-bopump does not achieve the performanceassumed in establishing engine performanceparameters, a complete redesign involving ex-pensive and critical time delays may be neces-sary. In the past this problem has been handledby overdesigning the turbopump to meet pos-sible deficiencies. Because of the new require-ments placed on the turbopump, desigllparameters are being pushed closer and closer tolimitations set by the present state of technol-ogy. Therefore, there no longer exists sufficientmargin for overdesigning to satisfy possibledeficiencies showing up in the developmentphase. As a result, it is necessary to improvepump design techniques to meet the increasedrequirements with assurance that a suitable tur-bopump will result.

    23

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    24/58

    |

    FIGURE 38-1.--Chemical-rocket pumps and turbine.

    From figure 38-1, some of the general func-tions and requirements of the turbopump in achemical rocket system may be inferred. Ingeneral, the propellant must be moved from thelow-pressure tank to the high-pressure enginesystem. The propellants are usually stored atlow pressures for minimum tank weight, andboiling or cavitation m,_y occur in the fluid sys-tem wherever the pressure is further reduced byvelocity increase. The pump must be capableof ingesting such vapor and still increasing thepressure of the propellants. Furthermore, thepropellants must be delivered to the enginewithout pressure or flow fluctuations that couldresult in a thrust fluctuation or possible destruc-tive engine instabilities. The power for pump-ing the propellant must be developed by a tur-bine. The exhaust from the turbine develops arelatively small thrust, and thus the flowthrough the turbine must be minimized to avoida sizable reduction in engine specific impulse_the energy release in pound-seconds per poundof propellant.This paper covers three topics pertinent to

    the turbopump :(1) Pumping under cavitating conditions(9) Flow stability in high-pressure pumps(3) Turbine drives with high-energy propel-

    lantsThe discussion mainly concerns hydrogen orhydrogen-oxygen turbines and pumps. It is ap-plicable_ in general, to modern turbomachinery

    ,eMICAL eOCKEr eeOeULSIONfor all rocket propellants and to both chemicaland nuclear rocket systems.PUMPING UNDER CAVITATING CONDITIONS

    In general, when the pump must operatewith the inlet fluid very near the boiling con-dition_ vapor forms in the pump inlet. Theformation of vapor bubbles and their subse-quent collapse as the fluid is exposed to boilingor vapor pressure in low-pressure regions of theflow is referred to as cavitation. The inlet por-tion of the pump must be capable of increasingthe pressure to suppress the boiling even whenthe flow contains some vapor. Several methodsof approaching the cavitation problem areillustrated in figure 38-2. In the upper leftsketch of figure 38-2 an inducer stage is addedto the pump. This stage is designed to operateunder cavitating conditions and increases thepressure sufficiently above the vapor pressure sothat the main-stage pump is not affected bycavitation. This can also be thought of as aprepumping stage raising the pressure in theline ahead of the main pump rather than in-creasing the pressure in the entire tank to sup-press boiling. The lower ]eft sketch illustratesan inducer stage ahead of an axial-flow pump.The upper right sketch illustrates a pump inwhich the inducer is integral with the mainstage of the pump. In this type of pump thedesign must avoid interactions between the

    SEPARATEINDUCER_CENTRIFUGAL

    AXIAL

    ' D ER

    SEPARATELY DRIVENBOOST PUMP

    C$-ZS&65

    Fi(_vm_ 38-2.--Pump and inducer configurations.

    24

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    25/58

    NEW PROBLEMS ENCOUNTERED WITH PUMPS AND TURBINES

    cavitation and the blade loading necessary toobtain high pump pressure rises. In each ofthese three configurations either a separate in-ducer or an inducer portion of the pump isutilized to obtain energy addition or pressurerise even under cavitating conditions. In eachcase the inducer is operated at pump speed. Ifthe inlet conditions are too close to fluid vaporpressure to obtain the necessary suction perfor-mance in this manner, it is necessary to add alow-speed booster pump ahead of the high-speed pump. Such a configuration is shown inthe lower right sketch of figure 38-2. Thebooster pump may be driven by a variety of de-vices, including gears, an electric motor, andgas or hydraulic turbines. _Vhether a high-speed or low-speed cavitating pump inducer isutilized, it is necessary to understand the vaporformation within the pump rotor and to findways to minimize the formation of vapor andthe effect of these vapor formations on thepressure production capability of the pump.Vapor forms when the low-pressure regions

    of a pump rotor are at or below fluid vaporpressure. The low-pressure regions exist invortices formed with tip or corner flows and onthe low-pressure portion of the pump blades.Figure 38-3 shows two pump blades rotatingin the downward direction with the relative flowentering from the left. The high- and low-pressure portions of the blade are indicated byplus and minus signs. The integrated differ-ence in these pressures reflects the level of en-ergy addition to the pumped fluid. As a resultof this pressure difference there is a flow in the

    T ZOTAT,0NL0W/ lROTATIONCS-25_60FIGURE 38-3.--Vortex formed because of flow throughtip clearance space.

    ,c-BUBBLE,_TREAy

    -_-CAVITY REGION

    \\\ /

    \\

    FIGURE 38--4.--Vapor formation at blade surface.

    clearance space, as indicated in the lower leftsketch of figure 38-3. At the intersection ofthe crossflow and the through flow, a vortex isformed. The center of this tip vortex formsa line at an angle to the inducer blade dependingon the relative velocity of the through flow andthe crossflow. The center of this vortex is a low-pressure region in which vapor first forms. Asthe inlet pressure is lowered further, vaporforms in the region of the crossflow and closesin the region between the blade and the tip vor-tex and thus completely obscures the view ofthe vaporous regions on the low-pressure sur-face of the inducer blade.The vapor formation on the blade surface

    can only be observed by viewing the inducerfrom the upstream direction. In this mannera view under the tip vortex can be obtained.The blade surface cavitation forms in one oftwo configurations shown in figure 38--4. Theright sketch illustrates a cavity region formingon the blade leading edge and closing on thelow-pressure surface of the blade. This typeof vapor formation occurs when vapor pressureoccurs at the blade leading edge. When theleading-edge shape is optimized so that fluidvapor pressure first occurs some small distanceback from the blade leading edge, streamers ofbubbles occur along the low-pressure surface ofthe blade. This latter type of vapor configura-tion is illustrated in the left sketch of figure38--4. Under normal operating conditions verylittle difference in performance exists as a resultof the type of blade surface cavitation, butrather is determined by the degree of cavitation.

    25

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    26/58

    CHEMICAL ROCKET PROPULSION

    (o

    ( C-62099

    26

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    27/58

    NEW PROBLEMS ENCOUNTERED WiTH PUMPS AND TURBINES

    In either case, the vapor region is very close tothe blade surface, and the blade surface pressureis very nearly fluid vapor pressure.Basic research on cavitation can be done most

    easily and inexpensively in water with verifi-cation tests using the actual propellants. Aseries of photographs illustrating a pump in-ducer and the various types of vapor formationthat are observed in water is shown in figure38-5. The helical inducer that was operatedin the Lewis water tunnel with a transparenthousing is shown in figure 38-5(a). The di-rections of rotation and flow are indicated inthe photograph. The orientation of the in-ducer blades to the flow is similar in the follow-ing five photographs, selected to illustrate thecavitation configuration discussed in connectionwith the sketches in figures 38-3 and 384.

    The photograph shown in figure 38-5(b)illustrates the formation of cavitation vaporin the center of the tip vortex. The side of theinducer blade being viewed is the low-pressuresurface. Thus, the cross-flow is normal to theblade toward the lower right direction, and itcauses the tip vortex to form as shown. Overthe forward portion of the blade some vaporforms in the crossflow, and the tip vortexappears nearly attached to the blade. Vaporis then confined to a narrow region of the vor-tex. When this process is viewed as a movie,there is a slight oscillatory motion of the tipvortex. Figure 38-5 (c) illustrates the tip vor-tex and crossflow vapor formation as the inletpressure is lowered toward fluid vapor pressure.The blade tip and the tip vortex form the edgesof the vaporous region. Under these condi-tions the vapor extends to only about 1/_ inchfrom the outer housing. Any surface cavita-tion that may exist on the low-pressure surfaceof the blade is not visible unless the camera isplaced in a more forward position so that a viewunder this tip cavitation can be obtained.

    (a) Orientation of helical inducer.(b) Formation of tip vortex cavitation.

    (c) Cavitation in tip vortex and crossflow region.(d) Cavitation appearing as streamers on blade low-pressure surface.

    (e) Blade surface cavity at moderate inlet pressure.(f) Blade surface cavity at lower inlet pressure.

    FIGURE 38-5.--Visual studies of cavitation on pumpinducer in Lewis water tunnel.

    Figure 38-5(d) was selected to show the for-mation of streamers of vapor bubbles on thelow-pressure surface of the pump inducer blade.(_Vhite tufts of yarn used to indicate tip eddyflows are shown on the inside of the transparenthousing. These yarn tufts are not significant tothis discussion and should not be of concern.)The streamers of bubbles appear to start atvarious points on the blade surface somedistance behind the blade leading edge and existat all radii. The number of streamers and thelength of the streamers increase as the inletpressure is lowered toward fluid vapor pressure.When streamer cavitation is viewed as a movie,streamers seem to appear and disappear at ran-dom_ even though overall cavitation conditionsappear to be constant.Figures 38-5(e) and (f) were selected to

    illustrate a cavity region on the low-pressuresurface of the pump inducer blade. The cavitybegins at, the blade leading edge and extendssome distance along the blade. Lowering theinlet pressure toward fluid vapor pressure in-creases the length of the surface cavity. Thesurface of the cavity is "dimpled" in such amanner as to reflect light as bright points. Theclosure of the cavity also appears very brightbecause of the light reflection. When surfacecavitation is viewed as a movie, the cavityclosure point is noted to oscillate slightly for-ward and backward along the blade. Witheither type of surface cavitation the vapor for-mation is confined to a relatively narrow regionfrom the blade suction surface with an esti-mated thickness comparable to blade thickness.It is necessary to relate the vapor formation

    as inlet pressure is lowered toward fluid vaporpressure _o the blade pressure distributions thatcontrol the energy addition to the pumped fluid.The sketch in figure 38-6 shows two bladesmoving downward with the relative flow fromthe lower left. The torque force being exertedon the fluid is related to 'the summation of thepressure difference on the high- and low-pres-sure blade surfaces. In this sketch a vaporousregion is shown to exist on the low-pressure sur-face of the blade. Tip vortex cavitation is notshown since this constitu'tes a blockage effect onthe through flow rather than a major effect onthe blade pressure distribution (especially if the

    8o65s2 0---62--3 27

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    28/58

    CHEMICAL ROCKET PROPULSION

    HEAD{PRESSURE)

    b

    T VAPORr // HEAD IPRESSURE]

    BLADE LENGTH cs-2s_61G C

    I_IOUBE 38-6.--Effect of cavitation on blade pressuredistribution.

    flow some small distance away from the bladetip is considered). The pressure distributionalong 'the length of an inducer blade is sketchedin figure 38-6(a). The mean pressure line atthe inlet represents the inlet static pressure.The large pressure difference (high- to low-pressure blade surface) near the inlet is due tosome angle of incidence between the inlet flowand blade mean angle. The pressure differencegradually decreases as the mean pressure ap-proaches the discharge pressure. The cross-hatched region is the region below fluid vaporpressure. Pressure distributions of this gen-eral nature were utilized for pump inducerblades, and performance was not affected whenthe lowest blade surface pressure was at vaporpressure. In fact, in operation the inlet pres-sure can be lowered so that the indicated low-pressure point will be well below vapor pres-sure (assuming that the pressure distributionremains the same) before changes in perform-ance are experienced. As the inlet pressure islowered, the portion of the pressure distributioncurve that would fall below the fluid vapor pres-sure line (as indicated by dashed lines in fig.38-6(b) ) probably does not exist. If the pumpinducer is to continue to deliver the fluid to thedischarge pressure, it is necessary for the pres-sure difference over the blade to increase in therearward portion of the blade (solid lines), andfor _he total pressure loading to be sufficient to'tchieve the required head rise. Thus the in-ducer can continue to maintain the dischargepressure or head rise as long as the overall pres-

    sure loading can be maintained by redistribu-tion rearward as the inlet pressure is loweredtoward vapor pressure. This process can con-tinue as long as the pressure gradient at the rearof the low-pressure blade surface can beachieved. _Then the fluid can no longer be de-livered at the desired discharge pressure, theenergy addition (area within the pressure load-ing diagram) decreases. This condition is illus-trated in figure 38-6(c), and under these con-ditions the inducer will produce only a smallhead rise. The original or noncavitating bladepressure distribution is shown as a dotted linein figures 38-6 (b) and (c). The area of thepressure loading distributions represented bythe solid line of figure 38-6(b) is meant to beessentially the same as that shown by the dottedcurve. It should be noted that vapor formationjust, begins at, the inlet pressure shown in figure38-6 (a) and that the length of vapor formationwill increase from figure 38-6 (b) to 38-6 (c), asnoted in figure 38-5.The pressure loading diagrams can be re-

    lated to pump indu_r performance by com-parison with figure 38-7. In this figure thehead developed is plotted against the head abovevapor pressure at the suction side of the pump.Condition A, where vapor first occurs on theinducer blade, on the solid curve exists at asizable pressure above vapor pressure. It canbe noted that tip vortex cavitation exists atinlet pressures well above point A. The vaporcavity continues to grow without affecting theinducer head produced as the inlet pressure is

    I, HYDRO .... --

    _ HEAD _"'- --RISE I__( EssuRE

    !........._ WATEr,i

    FIGURE 38-7.--Cavitation performance of pump inducerin water and in liquid hydrogen.

    28

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    29/58

    NEW PROBLEMS ENCOUNTERED

    lowered to B. At lower inlet pressures corre-sponding to point C the pump inducer head risefalls off very rapidly. From the comparison itis concluded that the flow model description issuitable for inducers operating in water (solidline in fig. 38-7).The cavitation performance of a given in-

    ducer in hydrogen may be vastly different fromthat obtained in water, as indicated by figure38-7. As the inlet pressure was lowered at somepoint, the inducer head developed in waterdropped off rather rapidly. In hydrogen, how-ever, the head produced did not decrease dras-tically, even though the inlet pressure was re-duced to vapor pressure (NPSH=0). An ad-ditional concept must be introduced into theblade loading diagram considerations to ac-count for this large fluid property effect en-countered with liquid hydrogen. These con-cel)tS are illustrated in figure 38-8. The sketchon the left shows the pressure distribution pre-viously discussed at an inlet pressure just abovethat at which a performance loss would be expe-rienced. It is believed that this is the type ofperformance encountered with water. The in-let static pressure is shown slightly higher thanthe inlet fluid vapor pressure. The inlet head,or total head, must be higher to account for thefluid velocity head at the pump inlet. (Theterm NPSH, inlet head above inlet fluid vaporpressure, is usually used to describe the inletperformance and is indicated in fig. 38-8.) Onthe other hand, other fluids, such as hydrogen,indicate a marked tendency to maintain someof the pressure loading below the inlet vaporpressure level. Several possible effects can oc-cur to allow this low-pressure region to exist:

    WATER HYDROGENINLET TOTAL INLET TOTALHEAD INLET HEAD(PRESSURE)- VAPOR (PRESSURE)_/ HEAD /

    [ (PRESSURE}-_ /HEAD I t / "_/ I /'P"E ORE)/'/ //,, //,' , ,l//

    BLADE LENGTH BLADE LENGTHCS-25._56

    t_o_ 38-8.--Fluid property effects on blade-loadingdistribution.

    WITH PUMPS AND TURBINES

    (1) lower local fluid vapor pressure as a re-sult of temperature drop due to fluid evapo-ration, (2) vapor and liquid with relativelysimilar densities and thus a relatively largepressure gradient that can be supported by avaporous region, (3) fluid tension effects thatretard the tendency to cavitate, and (4) rates ofphase change relatively different between fluids.One or more of these effects or others not

    indicated may occur to allow the blade pressureto drop below the fluid inlet vapor pressure.The net result is that the inlet pressure can belowered further without experiencing a loss inhead producing capability. The right sketchfigure 38-8 illustrates a case where the tank headand vapor pressure are coincidental. The_mount that the inlet pressure can be loweredwithout reaching the limiting pressure gradientor the effectiveness of the pressure diagram thatmust exist below the i_let vapor pressure is aresul.t of the thermodynamic properties of thefluid and thus is termed thermodynamic sup-pressi.on head (TSH), _ illustrated in figure38-8. It is likely that the TSH effects arepresent in all fluids, but they are relatively smallin water and relatively large in a fluid such ashydrogen. It may be feasible to incorporatethe TSH effec_ into the equations for suctionspecific speed S_ and blade operating cavitationnumber k in the following manner:Suction specific speed :

    N_fQ _. NvrQS,= (NPSH)3/4 (NPSH+TSH)3/4

    where N is rotational speed in rpm, and Q isthe flow rate in gallons per minute.Effecti_ve cavitation number:

    k h,--h, h,--h,+TSH=l/2p(Vi) 2_ 1/2p(V/)*

    where h, and h,p are the fluid static head andfluid vapor head at the inlet, respectively, and1/2p(Y[) 2 is the inlet relative velocity head.It can be noted that for the performance pro-

    duced in hydrogen as shown in figure 38-7 thesuotion specific speed term is infinity and thecavitation number becomes zero or negative,none of which are meaningful unless appro-priate fluid property effects (TSH) areincluded.

    29

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    30/58

    CHEMICAL ROCKET PROPULSION

    FRACTION __

    i REDUCTION OF FLUID VAPOR H_AD --FIOURE 38--9. Reduction of fluid vapor head due to

    hydrogen thermodynamic properties.

    A thermodynamic chart that illustrates thebasis for the difference in cavitation perform-ance between liquid hydrogen and water isshown in figure 38-9. This theoretical plotshows the thermodynamic changes that occurwhen evaporation of part of the flu, id reducesthe vapor pressure and temperature of the liq-uid-vapor mixture. The ratio of vapor volumeto liquid volume is plotted against the reductionin fluid vapor heat that occurs for evaporationfrom several initial liquid-hydrogen tempera-tures. Contours of the corresponding constantvapor weight fractions are also indicated. Theinteresting conclusion from this chart concernsthe large reduction in fluid vapor pressure thatcan be achieved with a relatively small amountof evaporation. For example, if only 0.016 ofthe weight fraction of a quantity of hydrogenadjacent to the low-pressure surface of theblade were evaporated from an initial tempera-ture of 36 R, equal volumes of liquid and vaporwould result in the local region, and the localvapor pressure would be reduced about 100 feetbelow that of the fluid generally. This local re-duction in vapor head corresponds approxi-mately to the low-pressure curve shown in fig-ure 38-8, the average of which corresponds ap-proximately to the value of TSH: This analyti-cal approach appears to be at least qualitativelyuseful in explaining the phenomenon. The re-quired vaporization can occur locally in theregion where vapor head reduction is specifi-cally needed without subjecting the entire flowpassage to a large volume of vapor. Theamount of TSH _hieved from this process de-pends on how effectively the various factors are

    utilized on the pressure loading diagram of agiven pump blade design.Better understanding of the hydrodynamic

    principles and the internal flow conditions dur-ing cavitation is necessary. Further researchto determine definitive values of the hydro-dynamic and thermodynamic interactions dur-ing cavitation will allow the designer to takefull advantage of these principles.

    HIGH-PRESSURE PUMP STABILITYThe problem of pump flow stability will be

    approached by noting the effects illustrated infigure 38-10. The variation in head rise withflow is shown in the upper left sketch of figure38-10. At some low flow value the pump andsystem operated with violent pressure fluctua-tions and are said to have been stalled. Even inwhat appears to be a good region of flow, how-ever, the pump head rise varied with time, asshown in the upper right sketch. The ratio ofpeak-to-peak variation in pressure head to thehead rise in general Hp.JAH varies w_th boththe flow rate and the suction heart above vaporpressure in the manner indicated by the lowersketches in figure 38-10. These pressurefluctuations are a result of unsteady flow con-figurations such as flow eddies. If the floweddy exists near the pump discharge, and thusaffects energy addition, the variation with Qcan be expected to be a major factor. On theother hand, if the eddy condition exists near theinlet and can interact with cavitation, thevariation shown with inlet head above vaporpressure can be expected to be large.

    HEAD I_SL

    \ RISE ]_FLOW,

    V V.-_Hp_p

    TIME

    FLOW, _ NPSH _ CS-ZSL.6,T,FIGURE 38--10.--Pump discharge pressure fluctuations

    due to unsteady flow conditions.

    30

  • 8/3/2019 Nasa SP-19 - Chemical Rocket Propulsion - Olsen

    31/58

    NEW PROBLEMS ENCOUNTERED WITH PUMPS AND TURBINES

    Before discussing the internal flow condi-tions which may affect pump stability, it wouldbe well to describe the overall characteristicsof hydrogen pumps. The pump shown infigure 38-11 is a research hydrogen pump rotorunder investigation at the Lewis ResearchCenter. When operated at 40,000 rpm (tipspeed of about 1100 ft./sec) the power requiredto drive this rotor is o