Prediction of Pressurant Mass Requirements For Axisymmetric Liquid Hydrogen Tanks

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

NASA Technical Memorandum 106973

AIAA-95--2964

Prediction of Pressurant Mass Requirements


for Axisymmetric Liquid Hydrogen Tanks

N.T. Van Dresar


Lewis Research Center
Cleveland, Ohio

Prepared for the


31st Joint Propulsion Conference and Exhibit
cosponsored by AIAA, ASME, SAE, ASEE
San Diego, California, July 10-12, 1995

(NASA-TM-I06973) PREOICTION OF
N95-29642
PRESSURANT MASS REQUIREMENTS FOR
AXISYMMETRIC LIQUTO HYDRCGEN TANKS
(NASA. Lewis Research Center) 8 p Unclas

Nalional Aeronautics and


Space Adminisb'ation G3/20 0055468
PREDICTION OF PRESSURANT MASS REQ REM S
FOR AXISYMMETRIC LIQUID HYDROGEN TANKS

N. T. Van Dresar*

Lewis Research Center


National Aeronautics and Space Administration
Cleveland, Ohio

A_tract Wp total pressurant mass


o
Experimental data from several test series are compared Wp total pressurant mass under conditions of
to an existing correlation that predicts the amount of zeroheatand mass transfer
pressurant gas mass required to expel liquid hydrogen / o
from axisymmetric tanks. It was necessary to use an Wp/W cotlapse
factor
alternate definition of the tank equivalent diameter to V volume
accommodate thermal mass in the tank wall that is AV expelled liquid volume
initially warm and to accommodate liquid residuals in OT total liquid outflow time
the tank after expulsion is stopped. With this
p density
modification, the existing correlation predicted mass
requirements to within 14 percent of experimental
results. Revision of the correlation constants using a subscripts
nonlinear least-squares fit of the current experimental exp experimental
data has a minor effect, thus supporting the validity of G gas
the original correlation's form, its fitted constants, and lid tank lid only
the alternate definition of the tank equivalent diameter, lred
SW swept by the liquid free surface during
Nomenclature expulsion
tank tank wall excluding lid
A surface area wall
w
C ratio of wall-to-gas effective thermal
_ty superscripts
CF collapse factor (= Wp /W ° ) 0 at pressurant inlet temperature and tank
expulsion pressure
Cp specific heat at constant pressure
Deq equivalent tank diameter ov_
hc gas-to-tank wall free convection heat denotes computed value that
transfer coefficient accommodates variable wall thickness or
m mass material
Pl...P8 fitted constants
heat flux from ambient to tank wall Inuxxluction
Q ratio of total ambient heat input to
effective thermal capacitance of gas The pressurized expulsion of cryogenic fluids from
S modified Stanton number propellant tanks was an active research area during the
t thickness 1960's and early 1970's as is evident from the large
TO pressurant inlet temperature number of publications on this subject. Of interest
Ts saturation
temperatureof ll'opelIant
at herein is the cryogenic pressurant requirements corre-
initial
tankpressure lation developed by Epstein 1 in 1965 and subsequently
revised by Epstein and Anderson 2 in 1968. The
correlation predicts the collapse factor, a dimensionless
pressurant mass, given the following dimensionless
*memb_,AIAA

Copyright _ 1995 by the American Institute of Aeronautics and Astronautics, Inc. No copyright is asserted in the United States under Title
17,U.S. Code. The government has a royalty-free licenseto exerciseallrightsunder the copyrightchimed herein forGovernmental purlmSes.
All other rightsare reserved by the copyright owner.
Pl P2 P3 P4 P5 P6 P7 P8

0.330 0.281 4.26 0.857 1.50 0.312 0.160 0.986

Table 1 -- Fitted constants for the Epstein and Anderson 2 correlalion for hydrogen propellant.

groupsmpressurant-to-saturation temperature ratio,


wall-to-gas effective thermal capacity ratio, ratio of C= t"'ptlw Ts (3)
ambient heat input to effective thermal capacitance of 0
pressuranL and a modified Stanton Number for gas-to-
tank wall heat transfer. The original correlation was
developed for cylindrical liquid hydrogen and oxygen
tanks pressurized by the propellant vapor or helium.
s= h 0r r, (4)
The form of the correlation has a thooretical basis and (pCp);Oeq TO
contains eight constants determined by nonlinear least-
squares fitting 1,2. In the later paper, the correlation was
and
revised with updated constants to include axisymmetric
tanks through the use of an equivalent tank diameter.
The revised correlation was compared to experimental Q= Oor (5)
data from numerous sources and reported to agree to
(pCp ); DeqTo
within + 12 percent, provided the data variables are
within specified ranges.

The quantity w p/w 0 is known as the collapse factor


The form of Epstein and Anderson" s correlation is:
and represents the ratio of actual pressurant
consumption to an ideal amount assuming no heat or
--6"= T-_-O-IqFI- exp(-PlCP2 II mass transfer from the ptessurant. The heat transfer
Wp {(Zs jr. _ ,J coefficient in Eq. 4 is obtained from a Nusselt Number
Wp
correlation for turbulent free convection for vertical
(1) planes and cylinders 3. Table 1 lists Epstein and
x [l-exp(-p3S p')]+ I}
Anderson's values of fitted constants for liquid hydrogen
pressurized by either hydrogen or helium gas.

x eXpL-.st,1--c-
[ (1"_ p6 (1-z-g)
J P7
S ]
Since publication of the revised correlation, additional
experimental data was obtained at the NASA Lewis
where Research Center for the pressurized expulsion of liquid
hydrogen from spherical and nearly spherical tanks 4-8.
The data series and references are listed in Table 2.
Wpo = pOAv (2)

Tank
Data Series Reference Tank Diameter Pressurant Gas
shape
Vail Dresar
I & Stochl 4 2.2 m Oblate _oid GH2

1I Stochl, et al5 4.0 m Sphere GH2


m Stochl, et al6 1.5 m GH2
1V Stochl, et al7 4.0 m GHe
V Stochl, et al8 1.5 m Sphere GHe

Table 2 -- Liquid hydrogen expulsion dam obtainedatNASA.


Variable Epstein & Anderson NASA data

Spherical tank diameter [m] 1.5-9.1 1.5-4.0

Wall thickness [cm] 0.25-2.5 0.21-1.3

Pressurant inlet mmpemture-


to-propellant saturation 2-15 8-17
tempetatme ratio
Total outflow time (see) 200-500 132-1980

Ambient heat flow (Wtm 2) 0-32,000 2.3-100

Table 3 -- Range of variables for correlation.

A total of 60 data points are available from the sources ackieved by omitting the liquid residual volume and the
in Table 2. These data points were obtained using a mass of the corresponding tank wall from the
variety of pressurant gas diffusers. Data obtained with analysis--i.e., the appropriate tank volume does not
straight-pipe gas injectors were not included as this include the liquid residual volume and the appropriate
tank mass does not include the mass of the tank wall
injector configuration results in high heat and mass
that remained wetted at the conclusion of the
transfer rates at the liquid surface ? . With a few
exceptions, the data variables fall within the ranges experiments.
specified for the Epstein and Anderson correlation as
shown in Table 3. The most significant differences are Epstein and Anderson state that their correlation may be
used when the initial ullage volume does not exceed 20
some data points having longer total outflow time and
the low ambient heat flux for the NASA data. percent of the total tank volume. For the present data
set, initial ullage volumes were from 5 to 14 percent of
In this work, the Epstein and Anderson correlation is the tank volume after correcting for the liquid residual
volume.
compared to the NASA data and a revision of the
correlation is provided.
The correlation further assumes a uniform wall
Comparison of Data to E_ostein and Anderson thickness and material. All of the NASA data was
Correlation obtained in tanks fitted with lids that were thick
compared to the tank walls, and in the case of dam from
references 5-8, the lid material differed from that of the
Although not stated, the Epstein and Anderson
correlation assumes that the tank is completely expelled tank. The adjusted tank walldensity, wall thickness and
(i.e., liquid residuals are zero). In the NASA wall specific heat capacity were obtained as follows:
experiments, expulsions were stopped at approximately
five percent liquid fill level. Therefore, when comparing t3.,
= mmnk + ml/a (6)
the predictions to the data, adjustments were made to Vta_ + Vna
correct for the liquid residuals. This correction was

?w = _ank + mlid (7)


:w_
_"For the present data set, the ratio of total mass transferred
across the liquid-vapor interface-to-total pressurant mass mad
ranged from -0.26 to 0.19, where a positive value repre-
sents condensation. Although this information is not
generally known, one should be careful to apply the corre- _p = ratankC p'tank + mlidC p'tid (8)
lation in its present form only to conditions where mass mtank +mtid
transferred across the liquid-vapor interface is no greater
than +25 percent of the pressurant mass. Some cases where The adjusted values obtained from Eqs. 6-8 were then
this condition is known not to hold are expulsion during
entered into Eq. 3 to calculate the "C" parameter.
liquid sloshing and expulsion of slush hydrogen.
Comparison of predicted and experimental results are In their paper, Epstein and Anderson defined the
shown in Fig. 1. The data points generally fall above equivalent diameter as the "diameter of a eylin&ieal tank
the diagonal line representing perfect agreement. having the same wall surface area and total volume as
Specifically, the Epstein and Anderson correlation the tank under investigation." Here an alternate
predicts a greater collapse factor than the experimentally definition for the equivalent diameter is suggested:
determined value for all but two points. Errors ranged
from -4 percent to +27 percent with a mean errorof + 15 4AV
percenL The root mean squared error is 5.6 percent. Oeq Asw (9)

4.5
.... I .... I .... I .... I .... I ' ' '_. where AV is the volume of expelled liquid and Asw is
% ! the area of wall surface swept by the liquid free surface
4.0
v." during the expulsion process (i.e., the wall surface area
initially wetted by the propellant that is exposed to gas
3.5 at the end of the expulsion). For an initially full tank
that is completely expelled, this definition is equivalent
to Epstein and Anderson's definition. Otherwise, this
a,_-3.0 definition removes the influence of both liquid residuals
L_
and warm tank walls above the initial liquid level.
2.5 When Epstein and Anderson's correlation form and
constants are used with the alternate equivalent diameter,
_Vda/ . 11I much improved results are obtained as seen in Fig. 2.
2.0 Errors range from -10 percent to +14 percent with a
v mean error of +0.6 percent. The root mean squared error
1.5 ,,,1_,_,1_,,,I,,,,1_,,,I,,,," is 4.7 percent.
1.5 2.0 2.5 3.0 3.5 4.0 4.5
CF 4.0
.... I .... I .... I .... i .... ,r' ' ''/.
exp

Figure 1 -- Correlation results using Epstein and


Anderson's definilion for equivalent diameter.
3.5

3.0
• _+14oovo,o
\ ........-"
• .e,_ S° #°s. _S !

Modification of the Correlation


_2.5
All of the NASA data were obtained with tank hardware
having initially warm thermal mass conconWaled at the
2.0
top of the test tanks. It is suspected the major cause of • " -'" • H
the discrepancy in the above comparison is the ..'" _ IU
inclusion of the warm thermal mass of the upper tank 1.5 - ." .-" ""....*"*" V-i-l::
A IV .:
wall tank neck and lid. This mass is not initially at the "" -" V V "
** o e ....... "

cold saturation temperature of the propellant, but at


1.0
elevated temperatures approaching that of the ambient 1.0 1.5 2.0 2.5 3.0 3.5 4.0
temperature of the surroundings or of the pressurant gas CF
exp
inlet temperature. In Test Series II to V, the ullage was
exposed to warm pressurant gas flow prior to the test
Figure 2 -- Correlation with alternate definition for
during a gas temperature conditioning procedure. In Test
equivalent diameter.
Series I, there was no conditioning of the pressurant gas
temperature, however, initial lid temperatures were near Revision of Fitted Constants
ambient temperature. Since this upper wall thermal
mass is initially warm, it is not expected to absorb
The constants Pl through P4 were updated using a
much thermal energy from the pressurant gas. Thus, it
nonlinear least-squares fit of the NASA data. Since the
is reasonable to attempt to modify the correlation by maximum ambient heat flow of the NASA data was
excluding the initially warm thermal mass.
less than one half of one percent of the maximum from
Pi P2 P3 P4 P5 P6 P7 P8

Epstein & Anderson 0.330 0.281 4.26 0.857 1.50 0.312 0.160 0.986

Revised Constants 0.300 0.291 5.71 0.906 1.50 0.312 0.160 0.986

Table 4 _ Comparison of original and revised constants for the Epstein and Anderson 2 correlation
for hydrogen propellant.

Epstein and Anderson's work, no attempt was made to


update their constants for the environmental heat input
(constants P5 through p8)§ . The comparison of The correlation of Epstein and Anderson is considered
reliable for axisymmetric liquid hydrogen tanks provided
predicted collapse factor with the experimental data is
one remains within the specified range of variables, the
shown in Fig. 3 and the revised constants are listed in
Table 4 along with Epstein and Anderson's values. The initial ullage space is not more than 20 percent, the
liquid is completely expelled, and the tank wall initially
revised constants give a slightly smaller error envelope
above the liquid level is near the saturation temperature.
and root mean squared error. Errors range from -8 to +13
percent with a mean error of +0.5 percent. The root
If the liquid is not completely expelled, or if the upper
mean squared error is 4.1 percent. Note that the revised
constants reduce the error envelope, but only by an tank wall temperatures axe significantly above the
incremental amount. This indicates that the use of the propellant saturation temperature, then the alternate
definition of equivalent diameter presented within should
alternate definition of equivalent diameter with the
be employed. The portion of thermal mass initially at
original Epstein and Anderson correlation has merit and
can be used with confidence. elevated temperatures with respect to the saturation
temperature or thermal mass below the final liquid level
should be excluded when calculating the "C" parameter
4.0 .... I .... I .... I .... I .... ,1''''_ (Eq. 3) and equivalent diameter (Eq. 9).

The correlations do not contain dimensionless groups


+13% error envelN_ -"'"'"'4"'-':ff'i that quantify heat and mass exchange between the
pressurant and the propellant. Therefore, the correlations
should not be used to predict pressurant mass
requirements in systems where these effects axe
relatively large--e.g., systems with liquid sloshing or
_'2.53.5 __
r..) slush hydrogen systems.

2.0 • I. The +13 to +14 percent error envelope of the present


work compares favorably with the +12 percent error
1.5 - .'" "'" A IV 2 envelope reported by Epstein and Anderson. The
.." ..'" V V : correlations are useful tools for estimating pressurant
mass requirements in axisymmetric liquid hydrogen
tanks. The reliability of these correlations may be as
1.0 1.5 2.0 2.5 3.0 3.5 4.0
good as, if not better than, current computer codes used
CF
exp to predict pressurant mass requirements.
Figure 3 -- Correlation with alternate definition for
equivalent diameter and revised constants. References

. Epstein, M., "Prediction of Liquid Hydrogen and


Oxygen Pressurant Requirements," International
§ The exponential multiplier in Eq. 1 containing the "Q"
Advances in Cryogenic Engineering, Plenum Press,
parameter has values ranging from 0.949 to 0.999 for the
New York, 1965, pp. 303-307.
present data set. Therefore, its impact on the correlation is
relatively small.
.
Epstein,M., and Anderson, R.E.,"An EquationFor
thePrediction of Oyogenic Pressurant Requirements
Fox Axisymmetric Propellant Tanks," Advances in
Cryogenic Engineering, Vol. 13, Plenum Press,
New York, 1968, pp. 207-214.

. McAdams, W.H., Heat Transmission, third edition,


McGraw-Hill, New York, 1954, p. 172.

. Van Dresar, N.T., and Stochl, R.J., "Pressurization


and Expulsion of a Flightweight Liquid Hydrogen
Tank," AIAA paper 93-1966 (NASA TM 106427),
1993.

. Stochl, R.J., Masters, P.A., DeWitt, R.L, and


Maloy, J.E., "Gaseous Hydrogen Pressurant
Requirements for the Discharge of Liquid Hydrogen
from a 3.96 Meter (13 ft) Diameter Spherical Tank,"
NASA TN D-5387, 1969.

. Stochl, R.J., Masters, P.A., DeWitt, R.L., and


Maloy, J.E., "Gaseous Hydrogen Requirements for
the Discharge of Liquid Hydrogen from a 1.52 Meter
(5 ft) Diameter Spherical Tank," NASA TN D-5336,
1969.

. Stochl, RJ., Maloy, J.E., Masters, P.A., and


DeWitt, R.L., "Gaseous Helium Requirements for
the Discharge of Liquid Hydroge n from a 3.96 Meter
(13 ft) Diameter Spherical Tank," NASA TN D-
7019, 1970.

° Stochl, RJ., Maloy, J.E., Masters, P.A., and


DeWitt, R.L., "Gaseous Helium Requirements for
the Discharge of Liquid Hydrogen from a 1.52 Meter
(5 ft) Diameter Spherical Tank," NASA TN D-5621,
1970.
REPORT DOCUMENTATION PAGE I OMBNo.0704-0188
Form Approved
put)ll¢ _ buf¢km lot tlda ¢xdlection el Infegmallon II mtlmamd to avefaQo I hotv lO*r mq0_xme, Ineludlng 11m Iknl tof mvlm., no .[mtnx_lonl., .lullcNng lKlltlng dill _
_ andr_nWn_ emdatan_d,d, andcemp_noandrev_nO.rap
..o_eCe_..d
_ .S_d_ e_,. _ee_r_.,_. __ m,
0_ II1_. ]11C_ IN_ lot redudfl_ this buIdert, to WilMIl_orl Heg0qualtofl _ DfllClOfall tot Ifll0n;lllt0lt (_ B _ lZ1=) _
Dar_ Hkihm'/. Su_ 1204./,q_to_ VA 222_-430_ and to _he Offkn o_ Mana_n_t and Budg_ Pq_mx_ Reduct_n Project (o7o4-olee). waeh_gt=_ 13(3 20e0_

1. AGENCY USE ONLY (Leave b/ank) 2. REPORT DATE 3. REPORT TYPE AND DATES COViV--KE_

June 1995 TechnicalMemorandum


4. Tm.S Am) SUBTrrLE S. FUNmNGNUMBERS
Prediction of Pressurant Mass Requirements for Axisymmetric
Liquid Hydrogen Tanks

WU-242-70-04
e. SUJTHOR(S)
N.T. Van Dresar

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGJU_ilZATION


REPORT NUMBER

National Aetonantics and Space Administration


Lewis Research Center E-9734
Cleveland, Ohio 44135-3191

9. SPONSORING/MONITORING
AGENCYNAME(S)AND ADDRESS(ES) 10. SPO NOORING,MONIT(N_NG
AGENCY REPORT NUMBER

National Ae=onanfics and Space Adminiswation NASA TM- 106973


Washington, D.C. 20546-0001 AIAA-95-2964

11. ,_IPPLE]IENi"ARY NOTES

Prepared for the 31st Joint Propulsion Conference and Exhibit COSlX3nsmed by AIAA, ASME, SAE, and ASEE, San
Diego, Califmnia, July 10-12, 1995. Responsible person, N.T. Van Dresar, organization code 5340, (303) 947-7553,
Boulder, Cokrado.
_2.. BSTPJBUT_WAWUUU_:LrrY
STATEMENT 12b. DISTRIBUTION CODE

Unclassified
-Uulimited
Subject
Categoa3,
20

publication is available from the NASA Center for Aerospaze Information, (301) 621-0390.

13. ABSTRACT (Max/mum 200 words)

Experimental data from several test sexies are compared to an existing correlation that predicu the amount of pressurant
gas mass required to expel liquid hydrogen from axisymme_c tanks. It was necessary to use an alternate definition of the
tank equivalent diameter to acconunodate thermal mass in the tank wall that is initially warm and to accommodate liquid
residuals in the tank after expulsion is stopped. With this modifw.ation, the existing correlation predicted mass require-
merits to within 14 percent of experimental results. Revision of the correlation constants using a nonlinear least-squares
fit of lhe current experimental data has a minor effect, thus supporting the vafidity of the original correlation's form, its
fitted constants, and the a]_te definition of the tank equivalent diameter.

14. SIJBJECT TERMS 1&NUMBER OF PAGES

O8
Pressurization; Fluid transfer
16. PRICE CODE

A02
17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSiFICATiON 20. uMrrATION OF ABSTRACT

OF REPORT OF THIS PAGE OF ABSTRACT

Unclassified Unclassified Unclassified

NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)


Pre_crlt_d by ANSI SKi. Z39-18
298-102

You might also like