1、NASA TN D-3095EXPERIMENTAL HEAT-TRANSFER RESULTS FOR CRYOGENICHYDROGEN FLOWING IN TUBES AT SUBCRITICAL ANDSUPERCRITICAL PRESSURES TO 800 POUNDSPER SQUARE INCH ABSOLUTEBy Robert C. I-endricks, Robert W. Graham, Yih Y. Hsu,and Robert FriedmanLewis Research CenterCleveland, OhioNATIONAL AERON,_UTICS AN
2、D SPACE ADNINISTRATIONFor sale by the Clearinghouse for Federal Scientific and Technical InformationSpringfield, Virginia 22151 - Price $1.80Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Provided by IHSNot for ResaleNo reproduction or networking pe
3、rmitted without license from IHS-,-,-EXPERYMENT_ REAT-TRANSFER RESULTS FOR CRYOGENIC HYDROGEN FLOWINGIN TUBES AT SUBCRITICAL AND SUPERCRITICAL PRESSURESTO 800 POUNDS PER SQUARE INCH ABSOLUTEby Robert C. Hendricks, Robert W. Graham, Yih Y. Hsu, and Robert FriedmanLewis Research CenterSUMMARYHeat tran
4、sfer to cryogenic para-hydrogen was experimentally determined inelectrically heated vertical tubes. The investigation covered subcritical andsupercritical pressures from S0 to 800 pounds per square inch absolute, massfluxes from I00 to I000 pounds per square foot per second, and heat fluxes to3 Btu
5、per square inch per second. The data were accumulated for a variety oftube test sections ranging from O.!SS to 0.507 inch inside diameter with aheated length from 16 to 24 inches. Similarities in the behavior of the near-critical to two-phase data were noted, including a minimum in the heat-transfer
6、coefficient near the saturation or transposed critical temperature (temperaturecorresponding to a maximum in the specific heat). Flow oscillations were notedprimarily when inlet conditions were below the transposed critical (or satura-tion) temperature. Preliminary results are also presented for a t
7、est sectionwith axial heat-flux gradients.Techniques for correlating Nusselt numbers are discussed. The correla-tions show promise for states where the hydrogen bulk temperatures are abovethe transposed critical temperature, although suppression of important param-eters in the correlations can somet
8、imes lead to large errors in calculating thebasic parameters such as heat flux and wall temperature. Extensive tables ofthe experimental heat-transfer data are included to aid rocket-cooling-passagedesigners.INTRODUCTIONHydrogen, because of its high specific impulse, is designated as a propel-lant f
9、or advanced chemical and nuclear propulsion systems. Its thermal proper-ties also make it an attractive regenerative coolant for these systems. As acoolant, hydrogen in its para state will be introduced into the cooling passagesat a fluid temperature near the critical value and at pressures either a
10、bove orbelow the critical value. It is well known that the properties of any fluid inthis near-critical state vary appreciably; for this reason it is difficult toProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-correlate forced-convection heat-transfe
11、r data. Indeed, literature can becited (refs. i to 8) to show opposed results: the heat-transfer coefficient ofa fluid near its critical point is a maximum in some cases and a minimum inothers. In a highly controlled, heated-tube experiment reported in reference 9,this contradiction was clarified by
12、 demonstrating that a minimum and a maximumin the heat-transfer coefficient could occur within a very short axial distanceof one another.The phenomenon was apparently associated with the proximity of the fluidto its critical temperature and was not significantly related to the wall tem-perature or h
13、eat-flux distribution. It should be noted, however, that the heatfluxes and temperature differences investigated in reference 9 were very lowcompared with those associated with a rocket engine.Reported herein are convective heat-transfer data taken in electricallyheated, vertical tubes for para-hydr
14、ogen at near-critical temperatures and pres-sures above and below the critical pressure. Experimental conditions were thosegenerally applicable to design situations in propulsion systems. Some data forthe gaseous region far removed from the critical temperature are included forcomparative purposes.
15、The range of experimental conditions is given in table I.To aid in comparing the fluid regimes, the critical properties of hydrogenare (from ref. i0)Critical pressure, 187.7 psiaCritical temperature, 59.57 RCritical density, 1.921 ib mass/cu ftThis report cannot claim the presentation of a theory or
16、 correlation thatcompletely resolves the problem of prediction of heat transfer for para-hydrogenat its near-critical state. Nevertheless, within the scope of the experimentalinvestigation, the following results are contributed herein:(i) A large quantity of convective heat-transfer data from unifor
17、mly heatedtubes is presented in tabular and graphical form that can be used in designanalysis.(2) Some of the trends in the data are discussed, especially those associ-TABY,_,I. - EXPERIMENTAL CONDITIONSFluid Inletstate tempera-ture,oRLiquid 45 to 60Fluid 60Liquid 55 R) were obtained by passingnorma
18、l gaseous hydrogen through the tank of liquid hydrogen at the desired pres-sure level prior to any run (see fig. i, p. 4 for gaseous connection). Thetank liquid-level thermometers indicated rapid uniform mixing of the para-hydrogen and normal hydrogen resulting in a fluid of unknown composition. Thi
19、spoint will be discussed later under the possible error source in item (4), Off-para composition (see p. i0). The gaseous data were run to check out the exper-imental system, and selected data are included herein for comparison purposes.Data RecordingDuring a run, the local surface temperatures, pre
20、ssures (static and differ-ential), and fluid temperatures were recorded on magnetic tape by an automaticvoltage digitizing system (ref. ii), and were available for write-back on anelectric typewriter. The electrical power inputs were recorded manually andmonitored by the digitizer and oscillograph r
21、ecordings. Checks of the digitizeddata were made on conventional self-balancing potentiometers and on a multi-channel oscillograph.Source of ErrorsA full discussion of the accuracy of the data is presented in appendix D.It is worth noting here the effects of the power-supply waveform on the dataaccu
22、racy and the sources of discrepancies in the heat balances.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Manydifficulties were encountered recording the power input to the testsection. Runswith reference numberslower than XXX-400 (see appendix E) h
23、avecorrected power Q (appendix A) values due to pulsating waveforms from the rec-tified direct-current power supply (150 kW). Subsequent adjustment of the powersupply improved the waveform significantly_ and installation of dynamic instru-ment movementsgave more satisfactory power measurements. The
24、saturable-reactor control element could still yield a peculiar waveform not amenabletoconventional metering. However_the computedheat balance (comparison of elec-trical input to enthalpy rise in hydrogen) on most runs was within i0 percentindicating that the volt-ampere measurementswere accurate.Som
25、eruns, however_ have heat balances off by 20 percent or more, yet re-jection of these data would not be justified on the basis of heat balance only,as other portions of the data were quite consistent. The conduction and radi-ation losses were estimated to be of the order of 1.0 percent or less (see
26、ap-pendix B of ref. 12).Of all the other possible sources of experimental error the authors couldadvance_ only a few appear p!ausib!% namely_(i) Surface phenomena(a) para-ortho conversion, and/or (b) surface adsorp-tion of foreign gases that would poison the surface and perhaps alterthe surface-to-f
27、luid transport mechanismThe high nickel content in test sections could act as a good catalyst for con-version of para- to ortho-hydrogen. Tests (ref. 12) have indicated this conver-sion to be small at moderate wall temperatures; however_ the effect may not benegligible at high wall temperatures_ res
28、ulting in erroneous heat balances. Attemperatures above 600 R_ there is little difference in the enthalpy level ofpara- and ortho-hydrogen_ so that the enthalpy at the wall would be uninfluencedby the fluid state. However_conversion at the wall might have a significenteffect on the thermal profile i
29、n the away-from-wall or wake region of the ther-mal boundary layer. The presence of the ortho species in these cooler regionscould influence the bulk or mixing temperature of the fluid mixture (para plusortho) and thus affect the heat-balance computations.(2) Stratification of fluid in mixing chambe
30、rs_causing an error in bulktemperature measurement(S) Nitrogen in the system that would change the chemical composition ofthe heat-transfer fluidThere were several experiences where appreciable quantities of nitrogen wereevident. This contaminant came from the pressurizing gas equipment. In atleast
31、two instances it was present in sufficient concentrations to accumulatesolid nitrogen in the venturi of the dip tube and stop the flow.(4) 0ff-para composition due to condensation of the pressurization gas intothe cryogenic liquidFor the warm liquid runs, normal hydrogen gas was bubbled through the
32、liquid toi0Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-raise its temperature. CondensationI of the gas would changethe para-orthocomposition for these particular runs. For example, to warm40 pounds of para-hydrogen to 70 R at 300 poundsper square
33、 inch absolute, the required heat in-put becomes_H70o - ) = 7576 Btu40 pounds R H36oR P=300psiaThe incoming gas to condensemust acquire this energy or- = 1720 Btu/ibHambient HTo R)p=300 psiaConsequently, about 4.3 pounds could condense, resulting in about 9 percentortho-hydrogen in the fluid mixture
34、. This estimated figure is based on the re-quired change in enthalpy if perfect mixing is assumed. As a check, somedatawere recalculated by assuminga maximumlO-percent ortho-hydrogen - 90-percentpara-hydrogen mixture. The results were not substantially modified; however_someof the heat balances impr
35、oved. The analysis of the data herein, exceptgas, assumeslO0-percent para-hydrogen conditions.Reduction and Analysis of DataThe symbols used herein are listed in appendix A, and the equations andassumptions employed in reducing the experimental data appear in appendix B.The thermodynamic and transpo
36、rt data of gaseous and liquid hydrogen usedin the supercritical pressure calculations are obtained from subroutine STATE(ref. iS). This program represents a single-source hydrogen-property programhighly amenableto heat-transfer and fluid-flow calculations. The calculationsin the subcritical pressure
37、 regime employ a method analogous to that presentedin reference 12 along with somesaturation properties and subroutine STATE.Specific heat, viscosity, thermal conductivity, density, and a fluid-propertiesparameter, computedfrom reference 12, are presented in plots in appendix C:figure 20 is the inte
38、rim viscosity computedfrom reference 14. The ne_er vis-cosity data have not been used because of the desirability of having a consist-ent relation between thermal-conductivity and viscosity data. As new thermal-conductivity data are unavailable, the use of the new viscosity data mayintro-duce errone
39、ous Z and k ratios, as exist for example in the Prandtl number,and thus the figures are included to aid the reader in manual calculations.DATAPRESENTATIONThe experimental data and somecomputedparsmleters are tabulated in tablesIV to VII (appendix F). The hydrogen used in these experiments was greate
40、r than95 percent para-hydrogen except for the gaseous data which were for normal hy-drogen. The colm_n headings and notations are explained in appendix E; the cal-culation procedures are given in appendix B and a directory to locate runsIAs used here, the cooling of gas to the bulk fluid state.iiPro
41、vided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-corresponding to specified test sections and operating conditions (table Iii)appears in appendix E.The supercritical-pressure data are segregated into two inlet temperatureregimes: Tin _ Tcrit and Tin _ Tc
42、rit. The division of Tin = Tcrit wastaken to assure that the low-temperature regime would be liquid-like. For thelow inlet temperature, the data are grouped first with respect to pressurelevel, then they are ordered with decreasing heat input_ and finally, within agroup of nearly equal heat inputs,
43、with increasing weight flow.DISCUSSIONOFDATAAs a result of this experimental program_ a number of observations havebeen made that are relevant to an appreciation of the results. The followingsubheadings will deal with these observations.Similitude Between Two-Phase and SupercriticalFluid-Forced-Conv
44、ective ResultsIn reference 15, it was pointed out that an apparent similitude betweenthe two-phase and supercritical forced-convection heat-transfer results existed.Subsequent examination of many of the runs reported herein did verify a similar-ity in the heated tube results between the two-fluid re
45、gimes, as long as thebulk fluid temperature at the tube entrance was less than or equal to the trans-posed critical temperature for the supercritical-pressure state. (The trans-posed critical temperature T* is the temperature where the value of specificheat maximizes.) The similarity in the characte
46、r of the data can be observedfrom an inspection of the temperature profiles shown in figure 4. The solidand dashed curves are axial wall-temperature distributions for a two-phase runand a supercritical run, respectively. The experimental conditions are tabu-lated in the figure. Note in both cases th
47、at the wall temperature increases toa maximum and then falls off. This trend is in contrast to a gaseous hydrogenrun_ shown as a dot-dash curve, in which the wall temperature increases through-out the length of the tube. Note that the profile for fluid hydrogen_ heatedwell above T*; resembles the ga
48、seous profile. Since the two-phase and super-critical runs presented in figure 4 involved a constant heat flux throughout thelength of the tube, the heat-transfer-coefficient distributions would be approx-imately the inverted images of these wall-temperature curves.Evidence from pool-heating studies
49、 of liquid hydrogen (ref. 16) furthersupports the thesis of a similitude between film boiling and heating of a super-critical fluid. Shadowgraph motion pictures available as a film supplement toreference 16 give a graphic portrayal of this similarity. The associated pool-heating data for both fluid regimes in reference 16 exhibited remarkable simi-larity even to the extent of near-identical magnitudes. While the
