Академический Документы
Профессиональный Документы
Культура Документы
oxidizer pump impeller, with the following dimen- axial-flow pump is selected. For each stage of
sions: an axial-flow pump, head rises of 5000 to 9000
Outside diameter of the radial ribs, dr= 14.8 feet can be obtained.
in (equal to d 2) The capacity of an axial-flow hydrogen pump
Inside diameter of the radial ribs, ds = 4.8 in is usually limited to about 5500 gpm as a mini-
Height of the radial ribs, t=0.21 in mum. This is due to the minimum practical
Width of the radial ribs, w = 0.25 in (not height h v of the vanes (fig. 6-49). For heights
critical) below 0.5 inch, the tip clearance required for
Average distance between the casing wall efficient performance becomes critical, causing
and impeller back shroud, s = 0.25 in manufacturing problems. A reduction in rotor
Estimate the reduction of the axial forces diameter below certain values is not practical
acting on the back shroud of the impeller, due to either, because of the high rpm required for
the radial ribs. proper blade speed.
Figure 6-50 presents typical operating regions
Solution of various liquid hydrogen pump types, These
include centrifugal pumps of (Ns)l = 500 per
The peripheral speed at diameter dr
_rN . tr×7000
ur:'7ff'd°r= 7"2-0 ×14.8=452 ft/sec STATOR ROTOR
=N = x 7000
× 4.8= 147 ft/sec
Us=_ds= 720
dt2+dh 2 dt2(l+rd 2)
stage (1 and 2 stages); centrifugal pumps of - (6-79)
dm2 - 2 2
(Ns) 1 = 1000 per stage (1 to 6 stages; not recom-
mended for rocket engine use); and axial-flow
pumps of (Ns)l -=3000 per stage (1 to 12 stages).
where rd = impeller hub ratio or dh/dt.
For any given operating region, there is usually
For simplicity, vane characteristics and flow
a best-suited design configuration. However,
conditions are discussed here only with respect
overlapping regions occur which could be ful-
to the mean effective diameter din. The vanes
filled by either a multistage axial-flow pump or
are equally spaced at a circumferential distance
by a single-stage centrifugal pump. The best
Pr
solution then is dictated by other considerations,
such as space envelope, mounting and ducting
ri_ m
During operation of an axial flow pump, it is The chord to pitch ratio generally increases
assumed that the meridional or axial component from rotor tip diameter dt to hub diameter dh for
of the absolute flow velocity Cm is constant structural reasons. The profile of the vane can
throughout all stages of the impeller rotor and be represented by the vane mean line (fig. 6-52)
the stator. To satisfy the flow continuity equa- which determines most of the important hydraulic
tion, the cross-sectional areas of the various properties of the vane. The thickness of the
flow passages at right angles to cm must also vane varies along the mean line for better per-
remain constant. This assumption is reasonable, formance and for structural strength. To impart
except for the effects of frictional drag at the effectively the driving action to the fluid, the
casing walls and the vanes. angle of the vane mean line, or rotor vane angle,
pUmP.
o',/\/ \__
228 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
where
/3c - 2 (6-82) i = angle of attack, deg
ii = angle allowed for circulatory flow at
the outlet, deg
Cr=2Rr sin -sin f_c (6-83) /32', /33' = relative flow angles at the rotor
inlet and outlet, deg
/3c = chord angle of the rotor vane, deg inlet and outlet, deg
cm = meridional or axial component of the
/32 =vane angle at the rotor inlet, deg
absolute flow velocities, ft/sec
/33 =vane angle at the rotor outlet, deg
Cr = chord length of the rotor vane, in Um = rotor peripheral velocity at mean
Lr = axial length of the rotor vane, in c2', cj =design absolute flow velocities at
the rotor inlet and outlet, ft/sec
(All parameters refer to the mean effective diam-
eter, din) Cu2', cu3' = tangential components of the design
An angle of attack or incidence angle "i" be- absolute rotor inlet and outlet flow
velocities, ft/sec
tween rotor inlet vane angle /32 and the direction
of the relative velocity of the flow entering the v2', v 3' : design relative flow velocities at
the rotor inlet and outlet, ft/sec
rotor, /32', is allowed for more effective driving
of the fluid. Also, an angle "ii" is allowed for Qimp =required impeller flow rate at the
circulatory flow between the rotor outlet vane rated design point, gpm
Q = rated design pump flow rate, gpm
angle /33 and the direction of the relative veloc-
ity of the flow leaving the rotor, fls'- For the Qe = impeller leakage loss rate, gpm (2 to
the flows at the inlet and outlet of rotor vanes = contraction factor of vane passage
(0.85 to 0.95)
can be constructed (fig. 6-52) with the following
correlations: AHimp = required developed head per impeller
stage, ft
dm tan fla :d_ tan fi3t Referring to figure 6-52, the velocity diagrams
at the stator inlet and outlet are constructed with
=rib tan /_3h=dx tan /_3x (6-93)
the assumption that the absolute flow velocities
um ut Uh t/x and angles at stator inlets and outlets are equal
(6-94) to the corresponding ones at the rotor outlets
dm dt dh dx
and inlets. This facilitates the design of multi-
stage axial-flow pumps using uniform rotor and
: ca2,__'
Cur__' : (6-95) stator stages.
rlm dt dh
To deflect the fluid effectively, the stator
inlet vane angles a 3 should be greater by a few
C u.___3i Cu3t' _ Cu3h ,
(6-96) degrees than the inlet absolute flow angles aa';
dm dt dh i.e., an angle of attack "i" should be allowed.
Also, an angle "ii" should be allowed between
where outlet vane angle a4 and outlet absolute flow
= rotor inlet vane angles at tip and angle a 4' for the circulatory flow (boundary con-
hub diameter, deg dition). The following correlations can be estab-
= rotor outlet vane angles at tip and lished for the vane and flow velocity diagrams of
_3t, _3h
eter, ft/sec
• -
Function of the Statot Cs=2R s sm _---_---) - si-_-_ac (6-100)
a3, a4 :vane angles at stator inlet and A higher pump specific speed is linked with
outlet, deg lower solidity. Typical design values for vane
Rs : radius of the stator vane curva- solidities for the rotor and stator at the mean
ture, in effective diameter dm range from 1 to 1.3, and
Ls = axial length of the stator vane, 1.5 to 1.8, respectively.
in 3. Number of vanes Zr, zs.-A lower pump
i : angle of attack, deg specific speed generally results in a larger num-
ii : angle allowed for circulatory ber of vanes. Design values of Zr range from 14
flow at the outlet, deg to 20. Design values of Zs vary between 35 and
a3', a 4' = absolute flow angles at stator 45. Their number should have no common factor
inlet and outlet, deg with z r.
Cm = axial component of the absolute 4. Vane curvature and vane setting.-Experi-
flow velocities, ft/sec ments indicate that the head developed by an
%', c,' = design absolute flow velocities impeller rotor is essentially determined by the
of stator inlet and outlet, ft/sec
vane curvature; i.e., fis-/32- Changes in vane
cu3', Cu,' =tangential components of the de- settings, i.e., outlet vane angle /33 and inlet
sign absolute velocities at vane angle/32, by the same amount (/3a- f12
stator inlet and outlet, ft/sec = const) will not affect head rise and efficiency
ast, ash, asx:stator inlet vane angles at tip, materially.
hub and any intermediate diam- The design procedure for the impeller rotors
eter, deg and stators of a multistage axial-flow pump is
a4t, a4h, a4x : stator outlet vane angles at tip, essentially the same as that for a single-stage
hub and any intermediate diam- centrifugal pump, except for the determination of
eter, deg the number of pump stages. Design parameters
(All parameters refer to the mean effective diam- and coefficients established experimentally with
eter din, unless specified differently.) earlier successful designs should be utilized to
the fullest. Special development tests are still
Design of Impeller Rotors and Stators required to verify the characteristics of the new
design. The design procedure includes the
A number of design factors directly affect the following steps:
performance and characteristics of an axial flow 1. To meet a given set of engine system re-
pump. Evaluation of test information, on the quirements such as rated design pump developed
basis of specific speed per stage (Ns)l, shows head H, flow rate Q, and rated pump (NPSH)c,
definitely the following correlations: the pump rot.tting speed N is deternfined first
I. Impeller bob ratio, rd.-The ratio of impel- through selection of a suitable inducer of a given
ler hub diameter dh to tip diameter dt (fig. 6-51) suction specific speed (Nss)ind.
has a direct bearing on the specific speed per 2. With N established, selection of impeller
stage (Ns)_. Higher specific speed pumps have rotor and stator of a given specific speed per
smaller hubs or hub ratios which results in stage (Ns)z, combined with the determination of
greater free flow area, and thus greater capacity, the number of pump stages, can now proceed
but lower head (H/Q characteristics). On the with the aid of the following correlations.
other hand, a higher hub ratio tends to yield a
higher head coefficient per stage (¢)1. Typical (Ns) _ - N(Q)°s (6-106)
values of r d in rocket engine hydrogen pump (H)I o-Ts
designs range from 0.76 to 0.86. Typical design
values for (Ns)_ and (¢), range from 3000 to AH=hHind:Hee+n(hH)l (6-107)
5000, and from 0.25 to 0.35, respectively.
2. Vane solidities St, Ss.-The vane solidi- where
ties or chord-spacing ratios of the rotor and (Ns), = specific speed per axial-flow pump
stator are important design parameters. They stage
are selected on the basis of previous experience. N = rated design pump rotating speed, rpm
DESIGN OF TURBOPUMP PROPELLANT-FEED SYSTEMS 231
Q = rated design pump flow rate, gpm impeller tip diameter. Design values for tip
(AH) 1 = rated design pump developed head per clearances, c, between rotor and stator (fig.
axial-flow pump stage, ft 6-49) range from 0.005 to 0.010 inch.
AH = rated design pump overall developed The vane thickness along the mean line is
head, ft mainly determined by structural considerations,
AHin d = inducer rated head rise, ft since experiments indicate that there is very
Hee = hydraulic head loss at the inducer little effect on Ferforma_Lce from variation of
stator, ft vane thickness. The problems with stressing
n =number of axial-flow pump stages pump rotor vanes are similar to those with tur-
3. For the specific speed per stage (Ns), bine blades. The methods given in section 6-5
thus obtained, various design factors and coeffi- for turbines can be applied here also. Vane
cients such as impeller hub ratio rd, vane solidi- stresses include centrifugal stresses, bendit_g
ties Sr and Ss, number of vanes zr and Zs, head stresses due to lift and drag loadings on the
coefficient per stage (_),, etc., are selected vanes, and vibrational stresses.
based on past designs with comparable (Ns)l
values.
Design of Cavitating Inducers for Axial-Flow
4. The required impeller rotor and stator
Pumps
diameters, velocity diagrams and vane profiles
can now be derived from equations (6-79) through The design procedures and parameters for
(6-105). eavitating inducers in axial-flow pumps are
Impeller rotor and stator vanes are generally essentially the same as those for a centrifugal
machined from forgings (fig. 6-51 and 6-53) using pump (fig. 6-51). Usually, the inducer has a
aluminum alloys or nickel-base alloys such as cylindrical tip contour and the same tip diameter
K-Monel. In view of the relatively low head dt as the impeller. The contour of the inducer
produced by an individual axial-flow impeller hub is highly tapered fl'om a relatively small
stage, reduction of skin friction and flow turbu- diameter at the inlet to a diameter close to that
lence losses are more important than with cen- of the impeller at the outlet.
trifugal pumps. A high degree of vane stream- An inducer stator, which also serves as the
lining and polishing is required for high efficiency. front bearing support, is positioned behind the
The axial distance da between impeller rotor inducer rotor. It is designed to convert into
vanes and stator vanes (fig. 6-51) has some pressure, a portion of the tangential component
bearing on performance. Typical design values of the absolute flow velocity leaving the inducer,
of d a range from 0.02 to 0.05dr, where dr=the and to discharge the fluid at an absolute flow
velocity and angle equal to that at the outlet of
an impeller stator (c 2' = Cu' ). The inducer stator
has the same db and dt as the impeller, and thus
['.." "._I i ROTOR
_rdm
Pis = (6-i08)
Zis
Cis
Sis -- _ (6-109)
Figure 6-53.-Inducer, impeller rotor, and turbine
rotor assembly of a typical multi-stage axial- al + a2
Cm=Cul ' tan UII=Cu2 ) tan a 2' Design of Casings for Axial-Flow Pumps
=c 1' sinai'=C2' sina 2' (6-114) As shown in figures 6-6 and 6-54, the casing
of an axial-flow pump consists of a cylindrical
dm tan al=dt tan al_
section which houses the inducer stage and the
=dh tan a_h:d x tan alx (6-115) impeller stages. It also includes a volute sec-
tion with radial guide vanes which is located
d m tan a 2 =d t tan a2$ behind the last 'impeller rotor stage. In addition
=dh tan a2h=dx tan a2x (6-116) to converting the tangential flow velocity com-
ponent into pressure, the volute section also
where serves to reduce the axial velocity component by
Pis = pitch or inducer stator vane spac- gradually increasing the flow area toward the
ing, in volute discharge.
Zis =number of inducer stator vanes The radial guide vanes of the volute section
Sis =inducer stator vane solidity are designed such that the fluid enters them with
Cis =inducer stator vane chord length, in minimum losses and that it leaves them in a
aic = chord angle of the inducer stator radial plane, analogous to a centrifugal pump
vane, deg (fig. 6-46). The number of radial guide vanes
a I , a 2 = vane angles at inducer stator inlet usually ranges from 17 to 23. They should have
and outlet, deg. no common factor with the number of impeller
Ris = radius of the inducer stator vane rotor vanes. Vane angle av can be determined
curvature, in by constructing the flow velocity diagram for
Lis =axial length of the inducer stator that section.
vane, in The calculations of the required areas at the
various sections of an axial flow pump volute
i = angle of attack, deg
are essentially identical to those for a centrifu-
ii = angle allowed for circulatory flow
gal pump (eqs. 6-69 and 6-70). For liquid hydro-
at the outlet, deg
gen, design values for volute flow velocities
al' , a 2' = absolute flow angles at inducer
range from 100 to 150 ft/sec. The section of the
stator inlet and outlet, deg
volute is generally circular in shape to accom-
Cm = axial component of the absolute
modate the high pressures. First-class thermal
flow velocity, ft/sec
insulation should be applied to the pump outside
CI' , C 2' = design absolute flow velocities at
surfaces. This will prevent excessive hydrogen
inducer stator inlet and outlet,
boiloff.
ft/sec
c i u', c2 u' = tangential components of the design
absolute velocities at inducer
Balancing the Axial Thrust of Multistage
stator inlet and outlet, ft/sec
Axial-Flow Pumps
alt, a_h =inducer stator inlet vane angles at
tip and hub diameters, deg Balancing of the combined axial thrust of a
a2t, a2h =inducer stator outlet vane angles at multistage axial-flow pump is an important func-
tip and hub diameters, deg tion, in view of the high pressures involved.
alx, a2x =vane angles at any diameter dx Special balancing devices, such as automatic
(All parameters refer to the mean effective diam- balance pistons, are frequently used. The bal-
eter din, unless specified differently.) ance piston is secured to the rotor assembly, as
DESIGN OF TURBOPUMP PROPELLANT-FEED SYSTEMS 233
movement (toward the inlet) of the rotor assem- ATa = variation of the forward hydraulic axial
bly, and thus the balance piston, reduces the thrust, lb
clearance at the front seal rub, simultaneously Apc =variation of the fluid pressure in the con-
increasing it at the rear seal rub. As a result, trol chamber, psi
the pressure in the control chamber between front dt =diameter of the front seal rub, in
and rear seal rubs is reduced. This effect coun- dr = diameter of the rear seal rub, in
teracts the forward hydraulic axial thrust of the
rotor assembly and restrains its forward motion.
Sample Calculation (6-10)
Similarly, a rearward movement (reverse thrust)
of the rotor assembly is counteracted by a pres- The following design data, based on engine
sure increase in the control chambers. The system requirements and on experimental model
volume to the rear of the control chamber, through test results, are given for the alternative A-2
cavities in the rotor, communicates with the low- stage engine fuel (liquid hydrogen) pump, based
pressure region of the pump inlet. The variation on an axial-flow, multistage configuration (simi-
in forward axial thrust can be expressed as lar to figs. 6-51 and 6-54).
o t "- vo,
--Jr -- ---- dr df
Vind=Q+Qee +½Qe
= 6080 (1 + 0.03 + 0.03) = 6450 gpm Nrrd _ _ 27 000 × ,7 × 6.57 : 774 ft/sec
"1- 720 720
We use a hub diameter doh=2.9 inches at the
inducer inlet. Considering that we are using a From equation (6-66), the tangential compo-
cylindrical tip contour, the absolute inlet veloc- nent of the inducer design absolute outlet flow
ity and its meridional component of the inducer velocity,
flow result from equation (6-59) as:
, g 6500×32.2
: 270 ft/sec
Qind cu, = AHind _ - 774
Co' _- Cmo :
3.12 ×4(49- 8.41) cl, =VfCul, 2 + Cm 2 : V/72 900+ 52 900 : 354.7 ft/sec
Check for inducer inlet flow coefficient: The inducer design absolute outlet flow angle
From equation (6-86) stators (i.e., dr=7 in, dh=6 in, dm=6.52 in, and
e=0.88). We also assume that the absolute flow
conditions at the inducer stator inlet are identi-
Qimp
C/B-- cal to those at the inducer outlet. From equation
2
3.12x_(dt - dh 2) e (6-112), the stator inlet vane angle at dm
a2 =a2'+ii=65 ° +5 ° =70 °
cm 230
a, +a2_44°26 ' +70 °
=57o13 ' tan /32 ' - um-cu2 -1 - 768-107.2 = 0.344; f12' = 19°
aic - 2 2
Cis 1.844
/32 =/32' +i=19 ° +4° :23 °
Ris=2 sin(-_) -2xsin(12°47')=4"16 in
From equation (6-81), the chord length of the The radius of the stator vane curvatures at dm
rotor vanes at dm
Cs 0,805
Rs: = 1.41 in
Cr = SrPr = 1.05 x 1.281 = 1.346 in
4
238 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
A-2 St:age Engine Fuel Pump Impeller Rotor and able energy of the turbine working fluid and thus
Smtor Design Summary the gas spouting velocity Co is relatively low, a
(Unless otherwise specified, data are all at higher turbine velocity ratio U/C o may be achieved
the mean effective diameter din.) with a moderate turbine rotor blade speed U. As
Rotor inlet flow velocity diagram (fig. 6-52), shown in figure 6-27, this suggests the use of a
a2' =65°; fi2' : 19°; Um=768 ft/sec; v2' relatively simple single-stage single-rotor im-
=699,6 ft/sec; c2'= 253.8 ft/sec; ca2' =107.2 pulse turbine. We have selected this type for
cm = 230 ft/sec the A-2 stage oxidizer tnrbopump, at the same
Rotor outlet flow velocity diagram (fig. 6-52), time taking advantage of its overall simplicity.
aj=32°40'; /33':29°26'; um=768; v 3' In most direct-drive turbopump configurations,
=468.7 ft/sec; cj =426.9 ft/sec; cuj=359.6 such as the A-1 stage engine turbopump (fig.
ft/sec; cm : 230 ft/sec 6-63), where turbine rotating speed N and conse-
Stator inlet flow velocity diagram (fig. 6-52), quently turbine velocity ratio U/C o tends to be
a3' =32°40'; c3' =426.9 ft/sec lower than ideal, a single-stage two-rotor
Stator outlet flow velocity diagram (fig. 6-52), velocity-compounded impulse turbine (figs. 6-9
a4' = 65°; c4' = 253.8 ft/sec; ca4' = 107.2 and 6-55) is selected for best results. Figure
ft/sec; cm = 230 ft/see 6-27 indicates that the optimum efficiency of a
Nominal rotor and stator tip diameter, dt= 7 in velocity-compounded turbine can be achieved at
Nominal rotor and stator hub diameter, dh = 6 in a relatively low U/Co value.
Nominal rotor and stator vane height, hv = 0.5 On the other hand, if a reduction gear train is
in provided between pumps and turbine, such as in
Nominal mean rotor and stator effective diam- the turbopump shown in figure 6-14, the turbine
eter, dm=6.52 in can be operated at a much higher rotating speed
Rotor vane elements (fig. 6-52), d2 =23°; (over 25 000 rpm). A higher value of U/Co can be
fl3=34026'; /_c=28°43'; Sr=l.05; zr=16; achieved with reasonable turbine wheel size.
Pr=l.281in; Cr=l.346in; Lr=0.645in; Then a higher performance, two-stage, two-rotor,
Rr=6.95 in pressure-compounded impulse turbine (fig. 6-10)
Stator vane elements (fig. 6-52), a 3 =36°40'; may be used.
a4=70°; ac=53°20'; Ss=l.61; Zs=41; 2. After the type of impulse turbine has been
Ps=0.5; Cs=0.S05in; Ls=0.645in; selected, the next step is the determination of
Rs = 1.41 in the turbine rotor size. Once the characteristics
of the turbine working-fluid (i.e., inlet temper-
ature To, specific heat ratio y, etc.), the turbine
6.5 DESIGN OF TURBINES
pressure ratio R_, and the pump or turbine roSa-
For rocket engine applications, impulse tire speed N have been set forth, a larger diam-
turbines are preferred, for their simplicity and eter for the turbine rotor tends to result in a
light weight. Our discussion will be confined to higher velocity ratio U/C o , or higher efficiency.
these turbines only. Figure 6-55 shows the gen- However, it also results in higher assembly
eral arrangement of a typical single-stage two- weight, larger envelope, and higher working
rotor velocity-compounded impulse turbine, stresses. Thus, the final selection of the tur-
bine rotorsize, and consequentlythe U/C o ratio,
is often a design compromise.
General Design Procedure
3. The requiredpower output from the turbine
The following steps are essential in the de- shaft must be equal to the net input to the pro-
sign of a rocket engine impulse turbine: pellant pumps, plus the mechanical losses in the
1. The first item of importance is the selec- gear train (if any), plus the net power required
tion of the proper type. A single-stage single- for auxiliary drives. The required flow rate of
rotor turbine (fig. 6-8) is used if the required the turbine working fluid can then be calculated
turbine power is low, since in this case the by equation (6-19) after required turbine power,
efficiency of the turbine has less effect on over- available energy of the working fluid (eq. 6-18),
all engine systems performance. When the avail- and overall turbine efficiency (estimated from
DESIGNOF TURBOPUMP PROPELLANT-FEED SYSTEMS 239
FLOW
TURBINE --
TURBINE INLET INLET
GAS MANIFOLD SHAFT FLANGE
ROTATING
SEALS
FtRST
BLADES
ROTATION
SECOND ROTOR .......
BLADES
SECTION A-A
DRIVING
/-'- ROTOR SPL/NES
DISC
CONTINUOUS RING
TYPE ROTOR BLADE
SHROUD
-_ dt
Figure 6-55.-Typical single-stage, two-rotor velocity compounded impulse turbine.
figure 6-27 for a given U/C o ratio and turbine above that of an isentropie process. This effect
type), have been established. is known as reheat. As a result of the above
4. Now the dimensions of the stationary noz- effects, the actual gas spouting velocity at the
zles, as well as those of the rotor blades, can turbine nozzle exit tends to be less than the
be calculated based on the characteristics and ideal velocity calculated for isentropic expan-
the flow rate of the turbine working fluid. sion (from stagnation state at the nozzle inlet to
the static pressure at the rotor blade inlet).
Furthermore, the effective flow area of a nozzle
is usually less than the actual one, because of
Design of Turbine Nozzles
circulatory flow and boundary layer effects. The
The nozzles of most rocket engine turbines following correlations are established for the
are basically similar to those of rocket thrust design calculations of turbine nozzles:
chambers. They are of the conventional
converging-diverging De Laval type. The main
function of the nozzles of an impulse-type tur- Nozzle velocity coefficient kn
bine is to convert efficiently the major portion of
available energy of the working fluid into kinetic
energy or high gas spouting velocity. The gas- Actual gas spoutingvelocity at the nozzle
flow processes in the thrust chamber nozzles are exit,ft/sec
directly applicable to turbine nozzles. However, Idealgas velocitycalculated forisentropic
the gas flow in an actual nozzle deviates from expansion from stagnation state at the
ideal conditions because of fluid viscosity, fric- nozzle inlet to static pressure at the rotor
tion, boundary layer effects, etc. In addition, blade inlet, ft/sec
the energy consumed by friction forces and flow
C1
turbulence will cause an increase in the temper- (6-118)
ature of the gases flowing through a nozzle, Co
240 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
-°'-1 ( -°.
NOZZLE
t'b BLADE
Jl TURBINE
II SYMMETRICAL
ROTOR.__"___/, _,"__ "'-__ ,__ ROTOR
IIBLADES .....
',_( ROTATION,,_ U \-
-F--i-_f--'--,'_---
----I- Y -7('2
e _ I..-P=--4
_--b.o
".--SHROUD _"_t b b2 " ] U "'-7""--
VELOCITY DIAGRAMS OF A
TYPICAL SINGLE-STAGE
IMPULSE TURBINE
F--"=,---I _.___ __ .
_".._._ l I_ UNSYMMETRICAL _',,._ _'.."_
_X,,_",,_ ) /'_ ROTOR BLADES _-'X."_ "_'_
I hb2> hbl
F b2 7 _--SHROUD b2
Figure 6-5G.-Nozzles, rotor blades, and velocity diagrams o[ a typical single-stage impulse turbine.
,TdmN
U- (6-129)
Design of Impulse Turbine Rotor Blades 720
calculated as
Amount of reheat in the rotor blades, Btu/lb of Design values of kb vary from 0.80 to 0.90. De-
gas flow: sign values of _b range from 0.7 to 0.92.
Referring to figure 6-56, the radial height at
qbr=(1 v?, the rotor inlet, hb, is usually slightly larger (5
-kb )2g j-r(1- r]n)A/-/l-2' (6-136)
to 10 percent) than the nozzle radial height hn.
where This height, together with the blade peripheral
al, a2 =absolute gas flow angles at the inlet speed U, will determine the centrifugal stress in
and outlet of the rotor blades, dog the blades. The mean diameter of the rotor
ill, fi2 = relative gas flow angles at the inlet blades is defined as dm= d_- hb, where d t is the
and outlet of the rotor blades, dog rotor tip diameter. Pitch or blade spacing, Pb,
C,, C: =absolute gas flow velocities at tile is measured at the mean diameter dm. There is
inlet and outlet of the rotor blades, no critical relationship between blade pitch Pb
ft/sec and nozzle pitch Pn. There just should be a
V,, V 2 =relative gas flow velocity at the inlet sufficient number of blades in the rotor to direct
and outlet of the rotor blades, ft/sec the gas flow. The number of blades zb to be
U = peripheral speed of the rotor, ft/sec employed is established by the blade aspect
dm = mean diameter of the rotor, in ratio, hb/Cb and the solidity Cb/Pb, where Cbis
_?n = equivalent nozzle efficiency appli- the chord length of the rotor blades. The magni-
cable to the expansion process in the tude of the blade aspect ratio ranges from 1.3 to
blades 2.5. Design values of blade solidity vary from
AH,_2,=isentropic enthalpy drop of the gases 1.4 to 2. Best results will be determined by
flowing through the rotor blades due experiment. The number of rotor blades should
to expansion or reaction, Btu/lb; have no cdmmon factor with the number of noz-
AH1-2' = 0 if only impulse is ex- zles or of stator blades.
changed The blade face is concave, with radius rt.
All parameters refer to the mean diameter d m, The back is convex, with a circular arc of small
unless specified otherwise, The turbine overall
radius rr concentric with the face of the adjoin-
efficiency Ut defined by equation (6-19) can be ing blade ahead. Two tangents to this arc to
established for a single-stage, single-rotor im-
form the inlet and outlet blade angles 0b_ and
pulse turbine as
Oh2 complete the blade back. The leading and
trailing edges may have a small thickness tb.
y_: _?n_b _rn (6-137)
The inlet blade angle 0b_ should be slightly
where larger than the inlet relative flow angle fl_. If
rl_ =nozzle efficiency Ob_ <fl_, the gas stream will strike the backs of
)7b =rotor blade efficiency the blades at the inlet, exerting a retarding effect
fir. =machine efficiency indicating the mechan- on the blades and causing losses. If Ob_ >fi_,
ical, leakage, and disk-friction losses in the stream will strike the concave faces of the
the machine. blades and tend to increase the impulse. The
Equation (6-134) shows that the blade effi- outlet blade angle Oh2 is generally made equal
ciency ;?b improves when/32 becomes much to the outlet relative flow angle /32.
smaller than I3L. Reduction of/32 without de- The mass flow rate _'t through the various
creasing the flow area at the blade exit can be nozzle and blade sections of a turbine is as-
achieved through an unsymmetrical blade design sumed constant. The required blade flow areas
(fig. 6-56), where the radial blade height in- can be calculated by the following correlations.
creases toward the exit. In actual designs, the Note that the temperature values used in calcu-
amount of decrease of fi2, or the increase of lating the gas densities at various sections must
radial height, is limited considering incipient be corrected for reheating effects from friction
flow separation and centrifugal stresses. Gener- and turbulence.
ally, the _2 of an unsymmetrical blade will be
approximately f3_-(5 ° to 15°). Equation (6-134) p1VlAblebl _p2V2Ab2eb2
also indicates that _b improves as a_ is reduced. Wt = 144 144 (6-138)
1L
244 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
Total blade inlet area, in2: of various blade sections at different radii gen-
erally do not fall on a true radial line. Thus the
Abi =Zbbbzhbz =Zbhbi (Pb sin Obl -tb) (6-139) centrifugal forces acting upon the offset cen-
troids will produce bending stresses which also
Total blade exitarea,in2: are a maximum at the root section.
2. Bending due to gas loading.-The tangen-
Ab2=Zbbb2hb2=Zbhb2(Pb sin Ob2-tb) (6-140) tial driving force and the axial thrust produced
by the momentum change of the gases passing
where over the blades may be treated as acting at the
Pb = pitch or rotor blade spacing midheight of the blade to determine the amount
= rrdm/zb, in (6-140a) of bending induced.
P,, P2 =density of the gases at the inlet and 3. Bending due to vibration loads.-The gas
outlet of the rotor blades, lb/ft _ flow in the blade passages is not a uniform flow
V z, V 2 =relative gas flow velocities at the as assumed in theory, but varies cyclically from
inlet and outlet of the rotor blades, minimum to maximum. The resultant loads repre-
ft/sec sent a dynamic force on the blades, having a
ebl, (b2 = area coefficients at inlet and outlet corresponding cyclic variation. If the frequency
of the rotor blades of this force should become equal to the natural
Zb --number of blades frequency of the blades, deflections may result
hb_, bb2 :radial height at the inlet and outlet which will induce bending stresses of consider-
of the rotor blades, in able magnitude.
bbl, bb2 : passage widths (normal to flow) at Detail stress analyses for rotor blades can be
the inlet and outlet of the rotor rather complex. A basic approach is to counter-
blades, in act a major portion of the bending moments from
0b_, 062 :rotor blade angles at inlet and out- gas loading with the bending moments induced
let, deg by the centrifugal forces at nominal operating
tb = thickness of blade edge at inletand speeds. This can be accomplished by careful
outlet,in
Typical constructionsof rocketturbinerotor
blades and disks are shown in figures6-53,6-55,
i- INTEGRAL TYPE SHROUD
6-56,and 6-57. Usually,blades are designed
with a shroud, to preventleakage over the blade
tips and to reduce turbulenceand thus improve
efficiency.Frequentlythe shroud forms an
integralportionof the blade, the shroud sections
fittingcloselytogetherwhen assembled. In _ gLADE
_ ROOT
otherdesigns the shroud may form a continuous
A-A
ring (fig.6-55)which is attachedto the blades
by means of tongues at the blade tip,by rivets, S.HROUDED BLADES FABRICATED BY PRECISION
blade design. Thus the centrifugal tensile The stresses in a turbine rotor disk are in-
stresses become a first consideration in blade duced by (1) the blades, and (2) the centrifugal
design, while other details such as centroid forces acting on the disk material itself. In
location and root configuration are established addition, there will be shear stresses resulting
later to fulfill design requirements. The follow- from the torque. As seen in figure 6-55, turbine
ing correlations are established at the blade root disks are generally held quite thick at the axis,
section where stresses are most critical. but taper off to a thinner disk rim to which the
Centrifugal tensile stress at the root section blades are attached. In single-rotor applica-
of blade of uniform cross section, psi: tions, it is possible to design a disk so that
both radial and tangential stresses are uniform
Sc = 0.0004572 1pbhbdmN2 (6-141) at all points, shear being neglected. In multi-
g
rotor applications, it is difficult to do this be-
cause of the greatly increased axial length and
Centrifugal tensile stress at the root section of
the resulting large gaps between rotor and stator
a tapered blade, psi:
disks.
Equation (6-144) may be used to estimate the
Ft = tangential force acting on the blades, Sd = centrifugal tensile stress of the turbine
lb/lb/sec (eq. (6-127)) disk, psi
Fa = axial thrust acting on the blades, lb/lb/ Wd = weight of the disk, lb
sec (eq. (6-131)) ri =distance of the center of gravity of tile half
The bending stresses at the root can be cal- disk from the axis, in
culated from the resultant bending moment. The ad =disk cross-sectional area, in 2
vibration stresses can be estimated from past N =turbine speed, rpm
design data. If the blade is fitted with a sepa- For good turbine design, it is recommended
rate shroud, its centrifugal force produces addi- that at maximum allowable design rotating speed,
tional stresses at the root. The total stress at the S d calculated by equation (6-144a) should be
the root section is obtained by adding these about 0.75 to 0.8 material yield strength.
stresses to those caused by the centrifugal Turbine rotor blades and disks are made of
forces acting on the blades. high-temperature alloys of three different base
246 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
materials: iron, nickel, and cobalt, with chro- blades. Angles ]3,, f12, /33, and /34 represent
mium forming one of the major alloying elements. the flow directions of V,, V 2, V 3, and V 4.
Tensile yield strength of 30 000 psi minimum at As with single-rotor turbines, the exit veloc-
a working temperature of 1800°F is an important ity from any row of blades (rotary or stationary)
criterion for selection. Other required properties is less than the inlet velocity, because of fric-
include low creep rate, oxidation and erosion tion losses. It can be assumed that the blade
resistance, and endurance under fluctuating velocity coefficient k b has the same value for
loads. Haynes Stellite, Vascojet, and Inconel X any row of blades:
are alloys frequently used. The rotor blades are
fabricated either by precision casting or by pre- V2 _Ca _V4
cision forging methods. Rotor disks are best kb- v 1 C: V 3 (6-145)
made of forgings for optimum strength.
In a multirotor turbine, the total work trans-
ferred is the sum of that of the individual rotors:
Design of Single-Stage, Two-Rotor Velocity-
Compounded Impulse Turbines (figs. 6-9, 6-55,
and 6-58)
_lI
occurs during gas flow through the moving ROTA
axial. U
Combined nozzle and blade efficiencyof a two- Design of Two-Stage, Two-Rotor Pressure-
rotorturbine: Compounded Impulse Turbines (figs. 6-10, 6-14
and 6-59)
E2b
T]nb= JAH (6-147) An operational schematic of a typical two-
stage, two-rotor, pressure-compounded impulse
where turbine and its velocity diagrams at the mean
diameter are shown in figures 6-10 and 6-59.
AH =overall isentropicenthalpydrop of the
Each stage of a pressure-compounded impulse
turbinegases, Btu/Ib
turbine may be regarded as a single-stage im-
= totalavailableenergy contentof the tur-
pulse turbine rotating in its own individual hous-
bine gases (eq.6-17)
ing. Most of the design characteristics of a
Equation (6-137)can be rewrittenforthe tur-
single-stage turbine are applicable to the indi-
bine overallefficiencyqt of a two-rotorturbine
as vidual stages. The gas-spouting velocities Cz
and C3, at flow angles a, and a3, of the first-
and second-stage nozzles, are designed to be
7It
= T/nbT/m (6-148)
approximately the same. Vz, V 2, V 3, and V 4
represent the relative flow velocities at inlets
Ideally,tin
b is a maximum forthe single-
and outlets of the rotor blades, fiz, fi2, f13, and
stage,two-rotor,velocity-compoundedimpulse
f14 are the corresponding flow angles for Vz, V2,
turbinevelocityratio
V 3, and V 4. The second-stage nozzles are de-
signed to receive the gas flow discharged from
U cos a I
the first-stage rotor blades at an absolute veloc-
C, 4
ity C2, and to turn it efficiently to a desired
angle a 3. Simultaneously, the gases are accel-
i.e.,when U= ¼C,t. The workload forthe sec-
erated to a desired velocity C3, through expan-
ond rotorof a two-rotor,velocity-compounded sion to a lower pressure. The flow at the outlet
turbineis designed at about one-fourthof the
of the second rotor has an absolute velocity C4
totalwork.
and a flow angle a4. U is the rotor peripheral
speed at the mean effective diameter din.
The totalwork performedin the turbineis the
The design procedures for the gas flow pas-
sum of thatof the separate stages. These may
sages of the rotor and stationary blades of a be designed to divide the load equally (i.e.,the
single-stage, two-rotor turbine are exactly the
same as those for a single-rotor turbine. How-
ever, velocities and angles of flow change with
each row of blades. As a result, the radial FIRST STAGE
ROTOR
NOZZLE
volume) in the flow passages must be taken into
account when calculating the gas densities at _ _' '= 2 SECOND STAGE
L-
DESIGN OF TURBOPUMP PROPELLANT-FEED SYSTEMS 249
a_ PO
heating in the nozzles, first rotor blades,
stator blades, and second rotor blades
0 '_k_- CONSTANT
To
A, /_PR_LFE UNES
C1- 4
From this, the gas static pressure at the noz-
zle exit
From this, the peripheral speed at the mean diam-
eter of the rotor
COS a I
Pl =Po CpToJ U=C1_=3940xCOS 425 °
tanfll=Cl C,cosal-U
sinai _ 3940×0.906-
3940x0.42390=0.622
We will use an angle a_ of 25 ° for the Using equation (6-135), the relative gas flow
spouting-gas-flow direction at the nozzle exit. velocity at the exit of the first rotor blades
DESIGN OF TURBOPUMPPROPELLANT-FEED SYSTEMS 251
The static gas pressure at the first rotor AH2_ 3, = AH 1-2' = 7.18 Btu/lb
blade exit
Analogous to equation (6-135), the absolute
gas flow velocity at the stator blade inlets
Y
F A., C a = x/kb2C22 +2gJ_nAH2_ a,
, 2_ 622
The gas static temperature at the exit of the qbs =(l-Kb )2--_+ (1- rln)
AH2_ 2
first rotor blade row following an isentropic
(Analogous to eq. (6-136))
expansion
= 18.53 Btu/lb
The actual static gas temperature at the first
rotor blade row exit The static gas pressure at the stator blade
exits
qbr2 41.975
--- 1374-+ - 1438 ° R
T2 = T2' + Cp - 0.653 Y
7.18 _9.06
Gas density at the first rotor blade exit p3 =p2 [1 -CDT2
_H2-21
j Y-1 =31.6x E1 0.653×1438J
=29.42 psia
144p2 144x31.6
P2 = RT 2 - _ 1--_-_-8
= 0.059 lb/ft 3
Gas static temperature at the stator blade
exits following an isentropic expansion
We use an angle/32 of 25 ° for the relative gas
flow direction at the first rotor blade exits (un- T3, = T2 -AH2_2/Cp= 1438-7.18/0.653= 1427 ° R
symmetrical blades). The absolute flow angle
a 2 at the first rotor blade exits can be calculated Actual staticgas temperatureat the stator
from blade exits
252 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
qbs 18.53 }r
T3 = T3'+-_--= 1427 + x-_--_.
= 1456° R
r =v
= 29.42 x 1 - 0.6537.1s
x 1456] -1_°_
144 P3 _ 144 x 29.42
.= 0.0544 lb/ft 3
P3 = R T 3 53.6x1456 = 27.46 psia > 27 psia (Pc)
We use an angle a3 of 35 ° for the absolute P4 is slightly higher than the turbine exit pres-
gas flow direction at the stator blade exit (a s sure (underexpansion), because of the reheating
Ta2). The relative flow angle t33 at the stator effects.
blade exit can be calculated from
7.18
The relative flow velocity at the stator blade T4' --T3 -AH3__,/C p = 1456--- - 1445 Btu/lb
0.653 -
exit
Point "4"-Second Rotor Blade Exit T4 -- T4 ' qbr2 1445 ' 7.73
+ C--_---= +_= 1457 ° R
The isentropic enthalpy drop in the second
rotor blades
Gas density at the second rotor blade exits
The amount of reheat in the second rotor V 4 sin /34 1306 × 0.695
blades tan a 4 =y4 cos /34- U=1306x0-719-890 -18"5
a 4 =86°55 '
,,y 2
qbr2 = (1 - kb2)-_gj+ (1- 9n) AH_-4' The absolute flow velocity at the second rotor
blade exits
(1312) 2 , ,.
= [1 - (0.89):] x 64--A x 778" Ll - 0.92) x 7.18
C4 =V 4 sin/34_ 1306x0'695=908 fps
= 7.73 Btu/sec sin a 4 0.9985
Nozzle Dimensions
Gas static pressure at the second rotor blade From equation (6-123), the required total noz-
exit zle throat area
DESIGN OF TURBOPUMP PROPELLANT-FEED SYSTEMS 253
92
From equation (6-140a), the number of blades
0.97 x 640 ll/32.2 x 1.124(0.94) lTns
53.6 x 1860
,,dm ,Tx 29.1
= 13.22 in2 Zbrl = Pbr 1 0.769 -119
144 w_ 144 x 92
Ane plClene 0.0658×3940×0.95 53'75in2 Combining equations (1-139) and (1-140a), we
obtain the blade radial height at the exit
Combining equations (6-125) and (6-126), we
obtain radial height and width at the nozzle exit: Ab2rl
hb2rl-_dm sin Ob2rl-Zbtb
Ane 53.75
82.5
hne = rrd m sin On - Zntn - rrx29.1x 0.391- 57 x 0.05 = 2.52 in
x 29.1 x 0.423 - 119 x 0.05
= 1.64 in
The blade passage width at the exit
Ane
The mean blade radial height Using equation (6-138), we obtain the required
total blade exit area
1.77+2.52
hbr, - 2 - 2.145 in 144 w_ 144 x 92
Ab2 s = P3 C3eb2 0.0544 x 1938 x 0.95 = 132.5 in s
Assume a tapered blade with shroud, and that
it is subject to approximately the same tensile Combining equations (6-139) and (6-140a), we
stresses from centrifugal forces, as would be a calculate the blade radial height at the exit
uniform blade without shroud. The blades shall
be made of Timken alloy, with a density pb =0.3 Ab2s
lb/in _. Cheek the centrifugal tensile stresses hb2s=rtdm sin Ob2s-Zbstb
at the root section using equation (6-141).
132.5
- _x 29.1 x 0.574 - 127x 0.05 -2.87 in
Scr I = O.O0045721pbhbrldm N2
The blade passage width at the exit
0.0004572 × _-0_-X,
× 2.145 × 29. i X (7000)2
bb2 s = Pbs x sin 062 s - Ib = 0.721 x 0.574- 0.05
= 13 050 psi = 0.364 in
Pbs - Blade solidity - 1.94 - 0.721 in Pbr2 = Blade solidity 1.67 0.888 in
Allowing 2024 ' between inlet blade angle Allow 2o4 ' between the inlet blade angle
0b, s and inlet absolute flow angle a 2 0blr2 and the inlet relative flow angle f13; thus
Obls=a2 +2024 `= 34°36 '+2o24 ' =37 ° Oblr2 =f13 +2_4' = 57_56' +2°4' =60o
We hold exit.blade angle 0b2s equal to exit We make the exit blade angle 062r2 equal to
absolute flow angle a3: the exit relative flow angle /94
From equation (6-149), blade radial height at From equation (6-149), the blade radial height
the inlet at the inlet is
The blade passage width at the inlet The blade passage width at the inlet
bbls= Pbs sin Ob,s-tb=O.721xO.602-O.05 bb 1r2 = Pbr2 sin Ob l r2 - tb = 0.838 x 0.866- 0.05
=0.384 in = 0.677 in
DESIGN OF TURBOPUMP PROPELLANT-FEED SYSTF.j_,S 255
The blade exit passage width Nozzles, AHo_ I,= 337.5 gtu/lb
First rotor blades, AHI_ 2, = 7.18 Btu/lb
bbsr2 = Pbr2 sin Obsr: - tb = 0.838 x 0.695- 0.05 Stator blades, AHs. _, = 7.18 Btu/lb
Second rotor, AH3. _, = 7.18 Btu/lb
= 0.533 in Total 5H = 359 Btu/lb
Working efficiencies:
The mean blade radial height
7/t=58.2%; ;7n=92%; _7nb=68.3%; qm=85.2%
Mean diameter of nozzles and blades:
3.10+ 3.66
dm= 29.1 in
hbr2 - 2 - 3.38 in
Nozzle dimensions (at din):
Aspect ratio = 9.7; z n = 57; Pn = 1.604 in;
Check the centrifugal tensile stress at the 0n=23°; hnt=l.5in; nne=l.64in; bnt
root section using equation (6-141) =0.1548 in; bne=0.576in
First rotor blade dimensions (at din):
Solidity=l.82; Cb=l.4in; Zbr_=llg;
Scr 2 = O.O0045721pbhbrsdmN 2 Pbrl =0.769 in; t)blrl =34°; Obsri =25¢;
bblrl = l.77 in; hb2rl =2.52 in; bblrl=0.379
= 0.0004572 x_x 3.38 × 29.1 x (7000): in; bbsrl =0.291 in
Stator blade dimensions (at din):
= 20 550 psi Solidity = 1.94; Cb = 1.4 in; Zbs = 1.27;
Pbs=0.721 in; 0bls=37°; 0b28=35°;
Turbine Efficiencies hbls=2.72in; bb:s=2.87in; bbls=0.384
From equations (6-146) and (6-147), the com- in; bb28=0.364 in
bined nozzle and blade efficiency Second rotor blade dimensions (at din):
Solidity = 1.67; Cb = 1.4 in; Zbr2 = 109;
U(C 1 cos a 1 +C: cos a2 Pbr2 =0.838 in; Oblr2 =600; 0b2t2 =44°;
÷C 3 cos a 3 +C 4 cos a4) hblr 2 =3.10 in; hbsb2 =3.66 in; bblrs=0.677
tinb -- in; bbsr2 =0.533 in
gJAH
AH2_ 3, = 44%; AH = 0.44 × 359 = 158 Btu/lb We chose a relative exit gas flow angle
/3 2 --38 ° for the first-stage rotor blades. The
Second-stage rotor blades: absolute gas flow angle, a2, can then be calcu-
lated as
AH3_4, = 3%; AH = 0.03 × 359 = 10.75 Btu/lb
V 2 sin f_2 1736 ×0.616
tan a s- i,,2 cos /32- U 1736 x 0.788- 1308 17.25
Point " O'-First-Stage Nozzle Inlet
T O-- 1860oR a2= 86040 '
Po = 640 psia
The absolute gas flow velocity at the first-
stage rotor blade exits
Point "1 "-Firs$-Stage Nozzle Exit = Rotor Blade
Inlet
V 2 sinfl2 1736×0.616
From equation (6-121), the gas-spoutingveloc-
C2 : sin a2 0.998 - 1070 fps
ity at the first-stage nozzle exit
Yl C1 sinai 2880×0.423 1784 fps From equation (6-147), the combined nozzle
sin 131 0.683 and blade efficiency
DESIGN OF TURBOPUMP PROPELLANT-FEED SYSTEMS 257
U(C x cos a_ +C 2 cos a 2 result of forces which act on the shaft or on the
+C 3 cos a3+C 4 cos a4) parts supported by the shaft. These forces may
r/nb = g JAil be divided into two classes: those which act at
right angles to the shaft axis (radial forces), and
1308 (2880 × 0.906 + 1070 x 0.058
those which act parallel to the shaft axis (thrus:
+ 2880 × 0.906 + 1070 × 0.058)
loads).
32.2 × 778 × 359
- lira _
258 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
_.:1• [
I
I
I I II
I [ I [
JBALL BEA
=,
NG C
. TRt
N IFUGAL LOAD
1LIMIT
1
r
¢ol I J I I ] L,_,TSE=,ESS,ZEI !
(EXTRA
&o i
I _.ou_-
e
II - I I
It
l [
II IOUTE_
../---TST'0,
=.cE
_*FO*=,T.
T--SPEC, .58
.55
o $O MRS
i J I '
J I
i i \-T-_ _ --_-\-q----
Io ZO 3o 40 50 60 70 8090 _
o. 200 _0 400 =X>o
fluid pressure surges, vibration, expansion and p = density of the fluid lb/in 3
contraction of sealing components, contact pres- Cs = seal leakage coefficient, established
sure between sealing surfaces, rubbing veloci- experimentally
ties of the sealing surfaces, smooth and friction- Labyrinth seals are used for the wearing rings
free operation of internal sealing parts, and on of pump impellers, as well as for the rotating
squareness of the sealing surfaces. Any influ- seals attached to tim sealing diaphragm between
ence which directly or indirectly subsequently two turbine stages.
alters these factors can cause improper operation In a face-riding-type seal the sealing is
of the seal.
accomplished through rubbing contact between
As shown in figure 6-62, the labyrinth seal is the precision-lapped faces of a floating seal
a clearance-type seal. The fluid tending to pass washer and a shoulder ring. The mating faces
through the sealing interface is throttled many are at a right angle to the axis of rotation. As
times and is forced to follow a devious path.
shown in figure 6-62, the floatin£ seal washer is
The function of a labyrinth seal is not to prevent attached to a metal bellows. The bellows, in
fluid leakage entirely, but rather to reduce leak- turn, is welded to a stationary seal housing
age to a reasonable level at a minimum of fric- which is secured and statically sealed to the
tion and wear. The amount of leakage through a casing. The bellows provides flexibility and
labyrinth seal can be estimated by the correla- spring force to the contact face, permitting it to
tion follow axial and angular movement without leak-
age. Sometimes a lip seal is used in conjunction
Qe=CsAcV'24g Aps/p (6-155)
with a spring-loaded floating washer, instead of
where a bellows.
-PUMP CASING
HOUSING
___ _SHAFT
-_-_- CC SHAFT
LABYRINTH SEALS
/-_ TURBOPUMP
_-TURBOPUMP CASING /CASING
FLOATING .\'..x>,_" P- SE;TL HO_SING
SEAL WASHER_\_\_/ST ICSE L ____GARTER SPR,NGS
_HER_/SEAL HOUSING
....._/,/_
_ ..... SEGMENTS
_ ___cu-_L_L - -- _ SHAFT
Thus self-adjusted dynamic sealing is accom- low-speed inducer. Gear arrangement and geom-
plished between shaft outside diameter and seg- etry depend upon power transmitted, propellants,
ment inside diameter. Axially, the segments are speed ratio, and other factors. During operation,
forced against a flat surface of the seal housing the gears are cooled and lubricated with oil, or
by a retaining plate and a spring washer, thus with the propellant being pumped.
providing a static seal. Shaft-riding seals tend The gears are usually housed in an aluminum
to occupy less space than the face-riding seals. casing. To minimize weight further, webs are
A wide variety of materials is available for held as thin as possible as are cross sections at
floating seal washers and seal ring segments. the rim and hub. The hubs are often internally
Carbon is used most frequently. The rubbing splined for best results. Spur gears are most
seal faces on shoulder ring or shaft must be widely used, since they minimize thrust on bear-
hardened or plated, and lapped to a very smooth ings. Tooth loads and speeds in turbopump
finish. The seal face rubbing speed should not gears are very high. The designer, therefore,
exceed 300 fps. Frequently, vent or purge lines must achieve high tooth strength and high re-
are connected to the cavities between two or sistance to wear. Turbopump gears are usually
more dynamic seals installed in series. This made of high-alloy steel, with the tooth surfaces
assures positive sealing for critical applications hardened by either case carburizing or induction
such as interpropellant seals. hardening. If possible, the tooth surface should
be accurately finished by a grinding process.
Materials and dimensional tolerances of turbo-
Turbopump Gear Design
pump gears must be held under very close control
The gear trains used in liquid rocket turbo- during manufacturing.
pumps (fig. 6-16) afford speed differentials be- To improve gear life and load-carrying stabil-
tween turbine, pumps and accessory drives, and ity, certain modifications to standard design
also sometimes between a pump impeller and a practices can be applied. Pinions are frequently
DESIGN OF TURBOPUMP PROPELLANT.FEED SYSTEMS 261
made with long addendum and gears with short considerations in preparing the layout. For
addendum to adjust tip-sliding velocities and to instance, one of the more important criteria which
strengthen the pinion. Furthermore, pinion tooth influences the selection or arranging of the turbo-
thicknesses are often increased, at the expense pump mechanical elements is the ease of devel-
of gear tooth thickness. High pressure angles opment. Standard or proven mechanical detail
as high as 22_ °,25 _j,or 271/:
° may be applied to should be extensively adopted in the layouts.
reduce contact stresses on filetooth surface and The following is a list of important turbopump
to increase the width of the tooth at the base, design layout considerations:
Involute-profile modifications are often also (1) Compatibility with engine systems pack-
made to compensate for bending and to keep the aging and plumbing
tips from cutting the mating part. (2) Structural integrity
(3) Positive interpropellant sealing
(4) Compensation for thermal expansion and
6.7 DESIGN LAYOUT OF TURBOPUMP
contraction
ASSEMBLIES
(5) Ease of development
Figure 6-63 presents the design layout of (6) Ease of assembly
the A-1 stage engine turbopump assembly. Logi- (7) Ease of manufacturing
cal packaging and arranging of the basic mechan- Considerable experience and skill are required
ical elements of tim turbopump are among the in turbopump design layout work for best results.
iNLET
PUMP i _-FUEL PUMP
PUMP VOLUTE AND
VOLUTE AND DISCHARGE
DISCHARGE MANIFOLD
i
FUEL PUMP -TURBINE NOZZLES
ROW ROTATING
BLADES
IMPELLER-_
OXIDIZER
PUMP
ROTATION /
OXIDIZER
PUMP
ROTATING
OXIDIZER
BLADES
PUMP INDUCER
SEAL
IMPELLER
HYDRAULIC
3TATIONARY
PUMP IMPELLER
BLAOES
AUXILIARY POWER
P1CKUP SHAFT
Figure 6-63.-Assembly design layout of the hypothetical A-1 stage engine turbopump.
Chapter VII
7.1 CONTROL METHODS opening and closing the propellant valves. Minor
deviations from the design mixture ratio or pro-
The foremost design requirements for any
pellant flow rates, such as from fabrication
control system are accuracy, stability, and reli-
tolerances of engine components, are corrected
ability. Two basic control methods are avail-
beforehand by insertion of accurately sized
able: open-loop (no feedback) and closed-loop
orifices into the propellant flow lines to effect
(feedback) control systems. Both have found
the desired pressure drops (also see ch. II). The
wide application in liquid propellant rocket pro-
extent of correction is determined from systems
pulsion systems. Open-loop control is confined
preflight calibration test data. Open-loop control
to those systems which are designed to operate
has the advantage of simplicity. However, it is
at a fixed, steady-state level over a narrow range
limited to a specific set of operating parameters,
of environmental conditions. Most other applica-
and is unable to compensate for variable condi-
tions require one of the many forms of closed-
tions during operation.
loop control. For these, mathematical models
Accurate sequencing of an open-loop control
can be constructed with which the functions and
system such as is used for engine start and stop
dynamics, such as gain factors and stability of a
is usually accomplished with the aid of inter-
proposed system, can be analyzed.
locks. Mechanical interlocks are preferred for
The selection of the best-suited method then
their high reliability. For instance, the propel-
is an important firststep in control systems
lant valves of many small engines or gas gener-
design. It will be influenced by the required
ators are mechanically linked and are operated
accuracy, the dynamic characteristics of the
by a single actuator. Proper sequencing between
system being controlled, and by allowable time-
fuel and oxidizer valves is achieved by adjusting
lags. Once the method is determined, the basic
the relative positions of the valve gates or pop-
elements for the proposed system must be se-
pets, with respect to the mechanical linkage.
lected, such as type of components of the power
Interlock requirements can also be furnished by
supply or working fluid (electric,hydraulic, or
other means (electric, hydraulic, or pneumatic).
pneumatic), and of the operating mechanism for
In high-thrust engines, sequencing between main
the specific control, which all depend on the
propellant valves and ignition system is often
specific application. Ideally, the basic theories
accomplished by the combination of various
and past experience should permit design without
interlock designs. A typical example is the A-1
experiment or development work. However, some
stage engine, for which the start and stop se-
development work with attendant redesign will
quence and their interlocks were described in
always be required in the process of perfecting a
detail in chapter III.
new system.
With this system, control is accomplished by Closed-loop control is also called automatic
preset control means, such as orifices, and on- or feedback control. This system usually in-
off command devices. A typical example of cludes sensing means, computing means to detect
open-loop control is an engine propellant flow errors, and control means to correct them. An
system, calibrated to a fixed set of conditions. accurately sensed feedback is compared with a
The propellant flows are controlled simply by fixed or variable reference by a computer, which
263
_mm I= mw _
elements will be presented in section 7.10. lators, and with a minimum of calibration firings.
Following sizing, orifices must be properly However, "thrust" regulators or "controllers" are
identified, by stamping or banding, and their employed in vehicle systems which require a
actual sizes recorded in the engine logbook. higher degree of precision and repeatability,
Control and calibration orifice locations and such as in single-stage vehicles starting at sea
sizes of a typical system are shown in figure 7-2. level, or in final stages of a multistage system.
Thrust regulators are actually chamber pressure
regulators. At altitude (vacuum), their effect is
Engine Systems Checkout and Test Controls
identical to thrust regulation, since at altitude
To verify operational readiness of the engine thrust for a given engine and mixture ratio is
system and its subsystems, suitable control solely a function of chamber pressure. The same
means are required for postassembly and pre- is essentially true for systems starting at sea
firing checkouts. These permit simulation of the level, because the relationship of thrust to cham-
operation of the engine and its critical control ber pressure as a function of altitude is predict-
components, without actually firing the engine able with high accuracy.
system. Utilizing suitable ground-support equip- Occasionally, vehicle missions will require
ment (GSE), an engine checkout control system in-flight thrust control over a wider range.
should include- Usually, in such cases, the need is for a planned
(l) Provisions to conduct leak checks and reduction of thrust, or "throttling," during the
electrical-continuity checks of the entire last portion of propelled flight. Two basic pro-
engine system. cedures are possible:
(2) Provisions for verifying proper operation (a._) Stepwise reduction of chamber pressure,
of all instrumentation pickups, such as Pc
dc bus voltage and spark plug firing (b) Continuous reduction of Pc
monitors; open, closed, and continuous Each of them can be accomplished by control
position signals for valves; propellant of-
flowmeters; and pressure transducers. (1) Turbine power (in the case of turbopump
(3) Provisions for verifying the proper func- fed systems), through regulation of gas
tion and operating range of all control generator propellant flow rate or hot gas
devices and subsystems, such as flow flow rate (preferred method).
control valves, pressure regulators, and (2) Main propellant flow rate
thrust and mixture ratio control devices. (3) Variation of main tank pressures (in the
(4) Provisions to simulate vehicle signals for case of pressure-fed systems).
"cold" checkout of the engine system Additionally, in multiple (clustered) engine sys-
operating sequence, such as for start tems, stepwise thrust reduction can be effected
and cutoff. by shutoff of one or more engines of the sub-
In addition to tile checkout equipment, the engine systems.
ground-support system must include equipment to The example chosen in section 7.1 to illus-
permit control of static test firings. This often trate a closed-loop control system is typical for
requires additional instrumentation. a system effecting thrust control through turbine'
power variation. Figure 7-3 shows the thrust
7.3 ENGINE THRUST LEVEL CONTROL control system proposed for our A-4 stage engine,
which relies on main propellant flow variation.
The significance of the thrust level of a Here, the closed-loop control system operates on
liquid rocket engine (sea level or vacuum) has the principle of variable fluid resistances in the
been explained in section 2.1. It is usually main oxidizer and fuel feed lines to achieve
specified with a tolerance; for instance, "_+3%." propellant flow-rate modulation. Engine reaction
It is possible, with modern "fixed thrust level" is determined by sensing chamber pressure, the
engines, to guarantee this band with simple parameter most indicative of thrust level, and by
orifice calibrations in the various propellant comparing the feedback b with the command
subsystems of the engine, without resort to regu- reference pressure input r. Any resultant error e,
268 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
following amplification and compensation as The simplest form of engine mixture ratio
required, is used to drive the thrust throttle con- control is obtained by the installation of properly
trol actuator of the main propellant control valves sized calibration orifices in the main propellant
in a direction which reduces the error. Ideally, lines. Acceleration effects during flight are
the system operates over the entire thrust throt- usually accurately predictable as a function of
tle range with minimal disturbances to other trajectory and flight time. Thus, simple averag-
critical engine parameters; in particular, the ing of flight mixture ratio and selection of the
propellant mixture ratio. In practice, these dis- corresponding orifice size reduces mixture ratio
turbances are not entirely avoidable, but can be deviations over the duration of flight to a level
minimized by maintaining a given resistance acceptable for optimum total propellant utiliza-
ratio between the two main propellant control tion in many missions.
valves throughout the control range. A most Open-loop mixture-ratio control can often be
reliable method toward this objective would be further refined by the following procedures:
mechanical coupling of the two propellant valves 1. Weighing ot the propellants loaded; i.e.,
(fig. 7-4). accurate determination oI the tanked propellant
Orifices, propellant valves, and servovalves mixture ratio.-The vehicle to be launched rests
required for thrust control will be described in on load cells, thus permitting weighing of the
subsequent chapters. propellants actually loaded. In mixed systems,
the noncryogenic component is loaded and
weighed first. The cryogenic component follows
ENGINE ST_.RT, CUTOFF {I TPiRUST
with the open-loop method, approaching that of a oxidizer flow in a similar manner. For instance,
closed-loop system (single stages; first stages). a propellant-utilization servo control valve,
In certain applications, however, the varia- which regulates the pneumatic pressure to the
tion of mixture ratio as a function of increasing main oxidizer valve actuator, may control the
acceleration may exceed tolerable limits. Accel- oxidizer flow by adjusting the angular position
eration in most vehicle tank arrangements affects of the oxidizer valve gate during engine main-
predominantly the propellant in the forward tank. stage operation. In certain applications it may
Because of the long supply line, acceleration be desirable to integrate the propellant flow
continues to act upon a relatively large fluid rates and to compare the masses consumed to
column, even near the end of powered flight (tank one another and to those tanked for optimum
depletion). By comparison, the effect on the propellant utilization.
fluid in the rear tank is often nearly completely It is readily seen that control systems, based
offset by the simultaneous decrease in fluid head on propellant flow-rate measurements, are a
(short liquid column). refinement of open-loop systems using fixed
To offset excessive acceleration effects on orifices. They are basically still mixture-ratio
the fluid from the forward tank and thus on mix- controls and thus merely "assume," but do not
ture ratio, head-suppression valves are some- measure directly, the amount of propellants actu-
times used at the pump inlet of turbopump fed ally remaining in the tanks and their unbalance.
systems. Here, pump inlet pressure increase is To accomplish this function, usually referred to
sensed as a function of acceleration. Corre- as "propellant utilization" (PU), additional con-
sponding signals are fed through a logic device trol elements must be employed in the form of
to the head-suppression valve which will gradu- vehicle tank-level sensors. Numerous principles
ally close, thus acting as a throttling device. are known: point sensing, sonar, acoustic, radi-
This valve also protects the pump structurally. ation sensing, differential pressure, and capaci-
tance probes.
Figure 7-6 presents the propellant utilization
Closed-Loop Mixture Ratio Control control system for the A-4 stage propulsion sys-
tem. The residual propellant quantities in the
In certain cases, such as in last stages, or in
missions requiring engine restart following ex- main tanks are continuously monitored, summed,
and compared with a PU control reference in the
tensive cruising periods involving propellant
boiloff, a closed-loop system may be required. propellant utilization control computer. Any
error detected is used to modify the command
In figure 7-5 we see the A-4 stage engine mixture
ratio control loop which operates on the basis of reference mixture ratio input, (MR)r, to tl_e
mixture-ratio control computer. This method
continuous propellant mass flow sensing. Both
fuel and oxidizer mass flow rates are monitored isolates the mixture ratio control from the pro-
and integrated to establish the ratio of either the pellant utilization control, and thus prevents
interaction between them. The bandwidth of the
propellants consumed or the propellants remain-
ing. The mixture ratio feedback, (MR)b, is then
compared with a command reference mixture ratio
input, (MR)r, in the propellant utilization control ilXTU_E _*TIO
(6) Acceleration effects during flight always a possibility and must be prevented by
(7) Propellant tank pressure deviations proper design and handling procedures. Many of
(8) Turbopump speed deviations these, such as cleaning procedures, will be
(9) Differences between fuel and oxidizer called out in the shop drawings. Furthermore,
pump characteristics as a function of the design, where applicable, will have to in-
speed clude filters, check valves, and suitable line
(10) Line resistance changes as a function of routing in order to prevent contamination and/or
temperature and for miscellaneous contact with incompatible materials. Note that
mechanical reasons some propellants may change their properties
(11) Temperature effects in rotating machinery merely as a function of time, such as hydrogen-
In the following we will discuss important peroxide, which loses its concentration due to
steps toward maintenance of high quality, and (very slow) decomposition (with attendant gas
toward further improvement in the listed areas, development), even if absolute cleanliness has
for highest accuracy of mixture-ratio (and been maintained. This affects design conditions
propellant-utilization) control. in addition to contamination considerations since
First, continued improvement of propellant proper venting devices must be provided. The
flow-metering devices is imperative. Here, latter, in turn, have to be designed in such a
turbine-type flowmeters have achieved a high way that no contaminants, including moisture,
degree of accuracy (conformance with truth) and can enter the propellant system.
precision (repeatability). Since mass flow rates delivered by pumps
The accurate calibration of these meters to and/or regulated by orifices will be a function of
most reliable standards is important. Since the fluid densities, mixture ratio may be affected
engine inlet pressures also affect the mixture accordingly. The densities, in turn, aside from
ratio, pressure measurements of the highest conformance with specifications, will be affected
reliability are equally necessary. Wherever pos- by temperature (noncryogenic fluids) or ambient
sible, the rocket engine design should include pressure; i.e., boiling point (cryogenic fluids).
vital metering and measuring elements from the To overcome these effects, it may be necessary
outset. Dynamic sensing devices, in particular to temperature-condition the propellants. This
flow meters, are drastically influenced by their may be done by heating or cooling. Or, it may
installation configuration. If these end organs, be accomplished by suitable storage, such as
following accurate calibration, remain with the shielding against solar radiation. For cryogenic
engine through its entire life cycle, including propellants, it is usually sufficient to keep the
flight, a maximum degree of accuracy is obtained. containers vented to atmosphere until immedi-
The design and machining of all calibration ately prior to use, since the possible changes of
orifices should closely follow accepted stand- atmospheric pressure at a given altitude can
ards (see section 7.10). Selection of suitable only introduce relatively minor temperature
materials to eliminate or at least to reduce to a changes. The designer, through a suitable oper-
minimum, temperature influences and corrosion, ating sequence (engine schematic) and through
is important. The design of orifice holders must provision of vent valves, recirculators, heaters,
prevent the possibility of incorrect (upside down) and other components, can minimize temperature
installation and of distortion of the orifices. effects.
Regulators, if any are used, must be designed The actuation of mixture ratio control devices
for highest accuracy and precision with particular affects the nominal engine performance param-
consideration of the medium to be controlled. eters. Depending on the type of engine, in par-
More detail will be presented in section 7.12. ticular its turbopump characteristics, these
The purity and composition of the better effects may be significant. In an actual case,
known propellants are regulated by official gov- the effects shown in table 7-1 were observed.
ernment specifications. The designer can expect It is clear that the vehicle thrust structure
that approved sources will deliver the propel- must be capable of absorbing the higher thrust
lants in conformance with these. However, sub- loads. Also, the vehicle tanks and their oper-
sequent contamination, dilution or alteration is ating pressures must be capable of meeting the
272 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
NPSH requirements for extreme mixture ratio ary coils by a passing magnet, or others? (The
excursions. Furthermore, chamber cooling may engine designer will be involved in this selec-
be affected. During sea-level testing, nozzles tion only if the generated signals affect engine
with high expansion area ratios may experience components.)
jet separation at the lower thrust levels (low Selection of the best-suimd electronic control
Pc), resulting in vibration, destructive to engine system.-This will be largely influenced by sen-
as well as vehicle structure. sors and control-valve selections.
Since vehicles are tanked for their nominal A propellant utilization system is not a mal-
mixture ratio, and since engines are calibrated function prevention system. It does not add to
to this ratio, mixture-ratio valve excursions vehicle reliability, possibly subtracts from it.
should be small for vehicles which are expended Rather, it is a system required to live with a
within a few minutes after takeoff. For stages, marginal preliminary vehicle design. It is a safe
with long cruising periods prior to operation or assumption, however, that the first flights of a
reignition, and which use one or two cryogenic new vehicle will not be for its ultimate mission.
propellants, boiloff may have altered the ratio of PU, therefore, will not be a vital necessity for
the propellants in the tanks to such a degree these flights.
that the PU system may be called upon to oper- Thus, enough time is available to thoroughly
ate at or near its maximum excursion. It is, investigate, analyze, select, and develop the PU
therefore, vital that the engine designer appraise system. This time should be utilized. Both
the vehicle builder of all performance variations engine and vehicle builder have facilities and
as a result of mixture-ratio adjustments, beyond test programs to permit mutual exposure of their
the standard tolerances of the nominal perform- selected systems to flight and simulated-flight
ance values. Also, engine turbopumps must be environment.
capable of operating for extended periods with
the valve in either extreme position.
7.5 THRUST-VECTOR CONTROL
A propellant utilization system is a complex
system. If required, it must be of the highest To steer a vehicle over its trajectory, thrust-
quality. Otherwise, it will do more harm than vector control is applied. The following methods
good. Only closest cooperation between vehicle have found application:
and engine designer will assure optimum quality. (1) Gimbaled thrust chamber or engine assem-
Areas of particular significance to teamwork are: bly (widely used)
Selection of the mixture ratiocontrol method.- (2) Jet vanes (obsolescent)
For instance, should the PU system be active (3) Jetevator
during the entire flight duration, or only for the (4) Gimbaled thrust chamber nozzle (rare with
last, say, 30 percent. (Both methods have been liquid propellants)
successfully used.) (5) Secondary injection (into the thrust cham-
Selection of the mixture ratio control valve ber)
specilications.-Should it be a variable orifice, (6) Auxiliary jets
or a bypass valve? What should be the permissi- The first method is used most frequently, due
ble pressure drops, required response rates, and to its inherent reliability and performance. The
accuracies? In case of sensor failure, should first four systems require actuators which may
the valve return to the neutral position or remain be operated by hydraulic, pneumatic, or electric
in its last working position? (Self-locking.) means. The remaining systems are controlled by
Selection of the sensors.-Should it be one of flow regulation.
several available continously reading types,
such as capacitance gages or differential pres-
Thrust Vector Control Systems Using Actuators
sure (tank top to bottom) gages? Or should point
sensors be employed, such as hot wires (change Figure 7-7 presents a simplified schematic
of heat loss as a function of being immersed in for a thrust vector control system, employing
fluid or exposed); switches triggered magneti- hydraulic or pneumatic actuators. It may serve
cally by floats; voltage pips induced in station- to explain the fundamentals of closed-loop thrust
DESIGNOF CONTROLSAND VALVES 273
II =° ered by a continuously
metal bidirectional
operating,
clutches.
constant-
speed, 28 volt de motor, fitted with dry-powder
The control com-
puter consists of summing junctions and an am-
(_ ELeCTRiC
#U_N
9,JMMPNG
plifier as in the case of hydraulic actuators.
_,.ECT RIC _PLIFJER
The dc motor drives the actuator through the
bidirectional clutches which are controlled by
Figure 7-7.-Typical schematic o{ a thrust vector
the error signal generated through comparing
control system using hydraulic or pneumatic
guidance command reference input with systems
actuators.
position feedback. To provide adequate systems
damping, the actuating speed is sensed by a rate
vector control, even though the systems used in generator or through differentiation of the posi-
practice may differ significantly in detail. The tion signal.
actuators are controlled by commands, originating Apart from electrical feedback and compensa-
in the vehicle guidance system, which are a tion systems, mechanical feedback systems
function of the vehicle's deviations from a pre- coupled with hydromechanical compensation
scribed path and of its response to corrective "networks" are coming into increased usage.
steering action. These signals are fed through They are inherently simpler and thus offer higher
an electronic thrust vector control logic to servo reliability. Two basic types of hydromechanical
valves. In the system shown in figure 7-7, each compensating devices may be distinguished:
servovalve modulates the fluid flow to its re- piston-bypass devices and load-pressure-sensing
spective actuator assembly in response to an devices.
electrical error signal which is proportional to Piston-bypass devices utilize leakages past
the difference between desired actuator position the actuator piston to introduce system damping
and its actual position. Feedback of the actual and may make use of dynamic relationships to
position is obtained through a transducer attached control time constants (a hole drilled through the
to the actuator. Additionally, tim actuating piston is an example).
speed is sensed by a rate transducer and applied Load-pressure-sensing devices, commonly
tO the control computer to stabilize the closed- called "pressure feedback" (PQ) valves or
loop control through adequate damping. Instead "derivative pressure feedback" (DPQ) valves,
of a rate transducer, electronic differentiation of are widely used.
the position transducer output may be applied
Figure 7-9 shows a typical servovalve and
actuator schematic with derivative pressure
feedback (DPQ) and mechanical feedback. The
only electrical signal required is the input to the
"torque motor" (an electromagnetic actuator)
resulting in deflection of the flapper of a differ-
ential valve, which drains to the sump. If the
flapper is deflected, as indicated in figure 7-9
by the arrow, nozzle flow on side B decreases,
r.lc _m,¢ lU_W,.G
with an attendant pressure rise. The reverse is
true for side A. The resulting pressure differen-
Figure 7-8.-Typical schematic for a thrust vec- tial forces the power-stage spool to the left,
tor control system using electromechanical blocking the return line on side B, and opening
actuators. it on side A. As a result, pressure Pb increases,
274 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
_ I, IEc:,I _&
FEEDB, ACI(
_WER ST_ _'OOL S_4mNG
Demonstration Example
RETURN -- m[TUml
Two basic types of electrical compensation
O_IFICE ORF_.J[ networks exist: current output for voltage input,
_'N.TER
tial equations of the two network types often Figure 7-10.-Current output for voltage input
permit the use of existing electrical networks network.
DESIGN OF: CONTROLS AND VALVES 275
V, voltage drop, volts ......... ' V = V A - VB AP, pressure drop, psi AP: PA - PB V _AP
dv
_I_------- V -------_ .... [_---- A p ----_ Q (Ap)2d(,_p) Ap 2 ! Massless piston as-
i=C_- = -g -_ C _ --_ sumed
--.._Q
Capacitor, farads
Ap : piston area, in 2
K :spring constant, lb/in
CSIIE3 _CxA P CX
operation point
Q ---...,lllb
t
------D- Q
Source: D. h. Engels, "A Method of Synthesizing Electro-Mechanical Compensation Networks for Hydraulic Servo-
mechanisms," Proceedings of the IEEE, PTGAC, October 1964.
The correct hydromechanical network, which The significance of good thrust alinement can
is of the piston-bypass type, is shown in figure be seen from the fact that in an engine cluster,
7-11. at the randomly distributed maximum of these
tolerances, a trim deflection of close to 0.5 °
Engine Installation and Alinement if the trim deflections seem to reduce effective
For minimum demands on the vehicle guid- thrust and guidance capability only slightly, the
ance and engine actuation systems, it is required need to apply them for the full duration of pow-
that the engine thrust vector be properly pre- ered flight results in appreciable payload re-
point in all three planes. Typical specified It is customary to aline the engine thrust
tolerances are: t0.25 inch laterally, _*0.5 ° verti- vector to the upper face of the gimbal bearing
cally. prior to shipment• Both optical and dynamic
276 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
Q
_C-_ /--VECHICLE THRUST MOUNT CENTERLINE
//,4i7/// /
_7 ._// / /-_..,_ cE.TE_ ,
-/- ,'.t-_.._q
EAANCDH
yIA_/ Pt:CAHE ) F_ __ _ GIMBAL "ARING [
\ \ \ Actu_ L.el_gl.h
:A • B * COITe_I_
_tCCD
\ \ \
),4 Cx
\.\\ I
-, \
" "
AP
,,c,o,
_--,.,UST /l\
I,=,
Figure 7-12.-Engine alignment.
"o
S
specifications. Figure 7-12 shows installation
methods of a prealined engine into the vehicle.
For the first vehicles of a new type produced, it
size and lightest weight of hydraulic pump, Resolution is low (gain, expressed in degrees
actuators, and associated equipment. Recogni- deflection per inch actuator stroke, is high).
tion of this need and careful design can do much The long-stroke design results in low effective
toward this goal. The system must be capable gimbaled mass and reduces actuator size and
of stable, well-damped response when cold gim- hydraulic system dimensions. However, the
baled, such as during prelaunch checkout, even spring rate is low and so is the stiffness. More-
though the loads encountered here may be quite over, the arrangement is bulky and requires extra
different from those occurring during engine fir- structural members. Only a detailed design
ing. This dual-load situation may pose serious analysis conducted jointly by engine and vehicle
problems. designer will determine which configuration is
During startup of the engine, brief peak side best for the flight system.
Hydraulic pump
Reservoir (low pressure, or "sump")
Accumulator (high pressure)
Servovalve
Actuators
Feedback (electric or mechanical)
Lines, check valves, filters, connectors,
instrumentation
11
PIEr $$ TRA NSD_JCEJl
TEIIIIp TI l N _'_DU_EEII
PUMP SUCTION
RETRACT
17 SERVO-ACTUATOR
pump. Only electrical wires will then cross the 22. SERVO VALVE
Secondary Injection
(C) LIQUID INJECI'O'R
The K factor determines the quantity of second- Liquid injection systems (inert fluid or pro-
ary injectant fluid required (for a known duty pellants) offer the simplest arrangement. This
cycle); the maximum flow rate; the additional is offset by their low performance, K-factors
tankage, pressurization fluid, and secondary being in the order of unity, at flow rates from 5
injection hardware weight; and the effect of the to 6 percent of the primary flow. However, in
added inert weight on vehicle trajectory. The systems with low-duty cycles, they may still be
K 1 factor evaluates the Is penalty on the propul- very attractive.
sion system. If K_ is equal to 1, the specific As a rule, four elements are required for a
impulse of the secondary fluid is equal to that of given system, equally spaced on the main cham-
the primary fluid and, therefore, the propulsion ber circumference, of which no more than two
system suffers no specific impulse penalty due adjacent ones would be operating at a given
to the SITVC system. time. The control of the required valves is ac-
Both the amplification factor K and the thrust complished through a logic and a servosystem
augmentation factor K1 are influenced by the analogous to that of a hydraulic gimbal actuator
secondary injection orientation. For each appli- system.
cation, a tradeoff must be made between the two
factors to determine the optimum injection orien-
7.6 DESIGN CONSIDERATIONS FOR FLUID-
tation for maximum propulsion efficiency. Let
FLOW CONTROL COMPONENTS
the force of an external jet of comparable geom-
etry at right angles to the primary nozzle be By theirvery nature,liquidpropellantrocket
unity. Then amplification factors greater than 2 engines use many control elements for regulating
are obtained if secondary injection is made with and measuring of fluid flows, such as valves,
the nozzles pointing upstream, rather than in a pressure regulators, and flowmeters. Some of
normal or downstream direction. Side forces for the design considerations governing these cqm-
a given _Ps are further increased if injection is ponents are discussed below.
made through a series of holes arranged on a
horizontal arc, rather than through a single ori-
Basic Flow Characteristics of an Ideal Fluid
fice. Note that the manifolds required in this
case may adversely affect response, however. Fluids, by definition, include both liquids and
Test experience suggests that overall pressure gases. A liquid is an incompressible fluid which
ratio and injector size appear to have little is characterized by a tendency to retain a fixed
effect, while gas temperature does, optimum val- density or volume; but not shape. A gas is a
ues being a function of propellant combination. compressible fluid which has no tendency to
For an oxygen/hydrogen tapoff system, the either a definite shape or volume. Its density or
range between 3000 ° and 4000 ° F appears most volume will vary according to the basic gas laws
favorable; however, as with turbines drives, (eqs. 1-9, 1-12, and 1-13). In general, the same
material strength and cooling problems will dic- fundamental laws of force, mass, and velocity
tate values substantially lower, say 1800 ° F. In apply to matter in all forms, and thus are also
a typical tapoff SITVC system, the gas flow rate applicable to the flow of fluids.
may be 1.5 to 2.5 percent of the primary flow The analysis of fluid-flow controls may be
rate, the upper value indicating the situation of simplified by initially assuming ideal conditions.
maximum force required between two injection For the calculation of physical dimensions and
stations (two jets operating). The tapoff system functional characteristics of specific control
offers simplicity and good performance. How- components, the results can then be modified by
ever, with low-duty cycles, a continuous bleed additional assumptions and empirical factors,
may be necessary to maintain temperatures at which often are the result of extensive testing.
the valves. A frictionless (zero viscosity), incompressible
The performance of a gas generator SITVC fluid which is nonturbulent and loses no mechan-
system is comparable to that of a tapoff system, ical energy as heat is referred to as an ideal
probably slightly better. This is offset by higher fluid. For steady, ideal-fluid flow in a closed
complexity (valves, injectors, ignition, cooling). conduit, Bernoulli's energy equation applies:
DESIGN OF CONTROLS AND VALVES
Assuming z t =z 2, and rearranging the expres- compensated for by the velocity coefficient Cv.
sions, we obtain For gaseous flows, however, pressure and tem-
perature have a significant influence on the den-
sity of the fluid and must be taken into account
144 (Pl - P2)_ v22 - v,_
(7-3) for calculations.
p 2g
PrLOT _¢ALVE
and
vl A2
(7-5)
v2 A_
_ENTURI
where _ATE
/ /
affects friction. The basic correlation is given / I
by Newton's law of viscosity (see fig. 7-19): / I
/ I
/ I
gU / I
r =-- (7-6)
gt l I
/ I
where t
r = shear stress = F/A, lb/ft: -- _-veuoc_tv OF FWtO PArTtcue_s zeao
F = shear or friction force of the fluid tangent
to the surface in consideration, lb Figure 7-I9.-Angular deformation of a real Iluid.
DESIGN OF CONTROLS AND VALVES 283
(R e : Dvp/tz, where D = equivalent diameter of the ciably. Also, there will be a slight change in
conduit, ft; v=flow velocity, fps; p=fluid den- the friction factor. Consequently, itis recom-
sity, lb/ft3; and/_=fluid viscosity, lb/ft-sec.) mended that equation (7-7) be used with com-
For most calculations, it is assumed that the pressible fluids only where the pressure drop Ap
flow is laminar for Reynolds numbers less than is less than 10 percent of the fluid static pres-
1200, and turbulent for Reynolds numbers greater sure at the outlet point. To calculate higher
than 1200. pressure drops of compressible fluids, other
Real fluid flows always involve friction methods should be used.
caused by rubbing of the fluid particles against If the flow is laminar (R e < 1200), the friction
one another and against the conduit wall. Con- factor is a function of the Reynolds number, and
sequently, there is a loss of energy; i.e., drop can be arrived at by Poiseuille's equation for
in pressure in the direction of the flow. This laminar flow
energy loss is converted into heat energy. The
heat thus produced may be entirely absorbed by
f__6_4_4 (7-8)
the fluid, in one extreme case (adiabatic flow), Re
or it may be entirely dissipated through the
conduit wall, in the other extreme (constant- When the flow is turbulent (R e > 1200), the
temperature or isothermal flow). Generally-at friction factor depends not only upon the Reynolds
ambient temperature-the flow of liquids and number but also upon the roughness of the duct
gases through pipes is assumed to be isothermal. or tube walls. Tile friction factors of turbulent
However, adiabatic flow is assumed to take flows may be found by means of the Moody dia-
place in nozzles, orifices, short tubes, and gram' shown in figure 7-20. The dimensionless
valves through which the fluid is moving at high term, "relative roughness" (e/D), is a measure of
velocities. the size of the surface roughness projections
The pressure drop Ap (psi) of a fluid flowing relative to the duct diameter. Average values of
in straight conduits (ducts or tubes) in a hori- surface roughness projections e for rocket engine
zontal position can be estimated by equation flow control components are given in table 7-3.
(7-7). This is essentially the same as equation For a curved-flow passage or for other shapes,
(4-32), except for the dimensions. the friction factor f obtained from figure 7-20 has
to be modified by an empirical correction factor,
which is a function of the Reynolds number and
Ap:f L PV2 (7-7)
288 g
where
TABLE 7-3.-Average Values of Surface Rough-
L = length of the conduit, in
ness Projections for Rocket Engine Fluid-Flow
p = density of the fluid, lb/ft 3
Control Component Designs
v = flow-velocity of the fluid, fps
d = equivalent diameter of the duct or tube, in
IRoughness
4 × duct cross-sectional area (any shape) Surface description projection,
_, ft
Wetted perimeter
I : friction factor, determined experimentally Drawn tubing with very clean surface ........ 0.000005
Smooth machined and clean surface .......... .00001
Machined or commercial cold-rolled surface... •00005
Equation (7-7) is valid for laminar or turbulent .0001
Rough machined surface ....................
flow of any incompressible fluid in ducts or Smooth cast or forged surface ............... .0003
tubes. With suitable restrictions it may also be Commercial cast, forged and welded surface .. .0008
/.--MANOMETER
Average resistance coefficients for various fluid-
flow-control components of liquid propellant
i-I rocket engines are presented in table 7-4.
For minimum pressure drop across a flow-
control component, the following flow-passage
CONTROL VALVE __ / I /CONTROL VALVE design considerations should he observed:
(1)-Allow sufficient characteristic flow area
for the component
(2) Avoid abrupt changes of flow area
(3) Avoid abrupt changes of wall contour, and
sharp turns in the flow path
(4) Minimize the length of the flow path within
the component
Figure 7-22.-Typical test setup for fluid-How (5) Provide a smooth surface finish for the
control components. flow passages
286 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
TABLE 7-4.-Typical Resistance Coefficients for main oxidizer valve (butterfly type) of the A-I
Various Fluid-Flow-Control Components of the stage engine.
Liquid Propellant Rocket Engines
Liquid oxygen flow rate, Q = 12 420 gpm
Resistance Liquid oxygen density, p = 71.38 lb/ft 3
Component description coefficient K Flexible duct inside diameter, d=8 in
Flexible duct actual length, L= 16 in
Butterfly-type valves (fig, 7-33):
0.31 Flexible duct equivalent length considering
90 ° open .............................
80 ° open............................. .41 resistance due to flow passage contour devi-
70 ° open ............................. ,77 ation, Le = 6 d
60 ° open ............................. 198 Main oxidizer valve characteristic flow area
50 ° open ............................. 5.68
= 78 percent of duct area
40 ° open ............................. 15.45
30° open ............................. 44.7
20 ° open ............................. 124.2 Estimate:
Ball-type valves (fig. 7-38): (a_.) The pressure drop chargeable to the duct
90 ° open ............................. 0.81
The pressure drop chargeable to the valve
70 ° open ............................. 1.58
50 ° open ............................. 3.6
18.2 Solution
30 ° open .............................
20 ° open ............................ 63
10 ° open ............................ 362 (a) Oxidizer flexible duct
2.5-3.5 The average flow velocity in the flexible duct
Poppet-type valves (fig. 7-40): Full open
Venturi-type valves (noncavitation)
(fig. 7-41) ............................ 0.8-I.5 Q 12 420
Gate-type valve (fig. 7-42): - 79.4 fps
v= d2 =3.12xTrx16
Full open ............................ 0.18
3.12x _-
_Aopen .............................. 120
½ open ............................. 5.6
_Aopen .............................. 24 From table 6-3, the viscosity of liquid oxygen
Poppet-type cheek valve (fig. 7-60) ....... 2to 4 is 0.277 × 10 -_ lb-sec/in2; thus/1 = 0.277 × 10-7
Swing-gate-type check valve (fig. 7-61) .... 1 to 2.5
x4636.8-0.1282x 10 -3 lb/ft-sec (see eq. 7-6).
Standard tee ........................... 18
.90 The Reynolds number of the flow in the duet
Standard elbow (90 °) ....................
Medium sweep elbow .................... .75
Long sweep elbow ..................... .60
45° elbow ............................. .42 Dvp 8x 79"4 x 71"38
Sudden enlargement: Re --_
/z
_
0.1282 x 10 -s =2"94x 107
d_/d 2= ¼ ........................... 0.92
d_/d2='_ ........................... .56
dl/d2=_A ........................... .50 Use a surface roughness projection size • of
Ordinary entrance ...................... .50 0.00005 or a relative roughness
Sudden contraction:
d_/d I --tA ........................... 0.42
d:/d 1:½ ........................... .33 0.00005 0 000075
.19 D 8
d2/d I :¾ ...........................
12
f(Le + b)pv 2
The following design data are given for the _ 0.0112 (6 x 8 + 16) x 71.38 x (79.4) 2 = 4.34 psi
oxidizer pump discharge flexible duct and the 288 x 32.2 x 8
DESIGN OF CONTROLSAND VALVES 287
Fluid-Flow Temperature
Pv2 0"31×71'38×(101'6)2 = 24.65 psi
Ap = g-_8_g - 288 × 32.2 Temperature is an import_tnt consideration for
the design of fluid-flow controls. This is espe-
Control Fluid Pressure Level cially true if the controls are for fluids at tem-
peratures in excess of, or far below, norn_al
The working pressure level and the temper- ambient.
ature of compressible fluid-flow-control system In liquid propellant rocket engines, fluid-flow
are important factors, since both govern the den- controls may have to handle hot gases at temper-
sity of the fluid. Means of compensation for atures up to about 1700 ° F. Example: the con-
changes of pressure in a compressible fluid con- trol of a turbine working fluid. Hot liquids need
trol system must always be provided. With an not be considered, since none of the liquid pro-
incompressible fluid, the pressure has relatively pellants have sufficiently low vapor pressures to
little influence on density. permit handling at high temperatures. Ability to
The working pressure level of the fluid deter- operate at elevated temperatures without any
mines the selection of the structural design of form of lubrication is a prime objective in the
the components as well as of the sealing meth- mechanical design of fluid-flow control. This
ods, especially for dynamic seals. Special pro- can be accomplished by using bearings of either
visions are often made to meet the stringent extremely hard, wear-resistant alloys, such as
requirements in high-pressure applications. For stellite and sintered carbides (high loading
example, the cutoff events in a high-pressure condition), or relatively soft materials such as
turbopump-feed engine system may be sequenced graphite (low loading condition). Bearings are
so that turbine power is cut first; thus the main usually subject to compression loads only and
propellant valves are not required to shut off are therefore not subject to failure if the mate-
against the high main-stage discharge pressures. rials used are of low ductility. For structural
members not subject to wear or bearing loads,
conventional high-temperature alloys such as
Fluid-Flow Velocity
stainless steels and other nickel-base alloys
The requirements for smooth component-flow- may be used. For static and dynamic seals,
passage contours are more critical with controls metal gaskets and bellows, carbon or graphite
for compressible, or low-density, fluids such as face seals, and labyrinth-type seals are suitable
hydrogen than for incompressible fluids, because at high temperatures.
their design flow velocities usually are much At the other end of the scale, liquid propel-
higher than those of the denser liquids. Also, in lant rocket engine controls may see extremely
general the design trend for high-thrust, high- low-temperature levels, such as in liquid hydro-
pressure engine systems is toward smaller pro- gen service (-425 ° F). Hero, two principal con-
pellant duct and valve sizes, and consequently ditions must be considered: (1) The physical
toward higher flow velocities (over 100 fps). characteristics of the fluids which at these low
An important consideration in the design of temperatures may affect control performance; and
high velocity flow-control components is the (2) the physical characteristics of the materials
high-impact loading imposed upon the control from which the control components are made and
• • [ .
which may affect the operation and, thus, the limiting factors governing response rate are (1)
performance characteristics of the control the speed with which signals can be transmitted,
devices. and (2) the mass/force ratio or its function, the
inertia/force ratio of the main control organ.
Many of the cryogenic fluids, i.e., liquefied
In many fluid-control systems the controlled
gases, experience somewhat unpredictable phase
fluid is used to transmit the sensed signal, In
changes (two-phase conditions) for relatively
others, part of the sensing link employs electri-
small temperature changes. No serious difficul-
cal or mechanical means. However, in most
ties need to be expected, however, if the heat-
cases, part or all of the sensing loop utilizes an
transfer rate from components in critical control
impulse generated by a pressure change. This
areas is low enough to prevent vaporization of
impulse is transmitted at the speed of sound in
the liquid. This is particularly important in
the fluid. As a typical example, the velocity of
liquid hydrogen service, where insulation may
sound in water is five times that in air; accord-
pose difficult design problems. At any rate,
ingly, a control signal would be transmitted five
except for viscosity changes, nearly all liquids
times faster in water.
exhibit more stable physical characteristics with
The actuators for most fluid-flow-control
large temperature variations, within the range
mechanisms use pistons or diaphragms, powered
between their freezing and boiling points, than
by fluid pressure which, in turn, is regulated by
do gases if the temperature range reaches to
some form of pilot valve. If suitable, the con-
their liquefaction temperatures.
trolled fluid may be used as the actuating fluid.
The construction materials for fluid-control
The response and flow capacity of the pilot
components for low-temperature applications
valve, the effective area of actuator piston or
must be especially carefully selected. Practi-
diaphragm, and the actuating fluid pressure level
cally every metal undergoes irregular phase
influence directly the response rate of the con-
changes at low temperatures which may seri-
trol mechanism for given mass inertia and fric-
ously affect its physical properties. While the
tional or other resistances.
strength of metals generally increases with a
decrease in temperature, further temperature To satisfy certain operating conditions and to
decrease beyond certain limits may result in a attain stable control it sometimes becomes nec-
decrease in strength. Many metals also become essary to introduce simple damping devices. In
brittle at very low temperatures. Most of the most control systems, stability is inversely pro-
aluminum alloys and the 300-series stainless portional to sensitivity or response rate. Thus,
steels exhibit much better stability at tempera- the design of a fluid-flow-control system should
tures in the cryogenic range than do others. reflect a realistic balance between sensitivity or
Elastomers such as Teflon, Kel-F, and Mylar, response rate, control accuracy, and system
when used for sealing purposes, exhibit satis- stability.
factory mechanical characteristics at extremely Figure 7-23 illustrates the schematic of a
low temperatures. Teflon-coated surfaces addi- typical piston-type actuator for fluid-flow-control
tionally have good anti-icing characteristics. devices. The piston when actuated moves
For further detail on materials, see chapter II. against the spring in the direction of the arrow.
Fluid-flow-control components for operation The basic correlation between the response rate
at cryogenic temperatures should be designed to or acceleration of the piston, and other opera-
be free of external icing effects. In addition to tional parameters, can be expressed by
insulation, moisture-preventing purges should be
provided internally in critical areas such as Map
-Alp 1-A2p2- Fr- Ft- Fs- Cx (7-10)
bearing interfaces. Also, actuators and/or bear- g
ings may require heating. where
M :effective mass accelerated by the actuator
piston, lb. It consists of piston mass,
Rate of Response in Fluid-Flow Controls
that of moving parts mechanically con-
Response rate is an important design consid- nected to the piston, and of the mass of
eration in any control system. Basically, the all the fluid columns in the system
DESIGN OF CONTROLS AND VALVES 289
ACTUATING FLUID
SUPPLY LINE
tons, shafts, and rods; and seals for valve seats.
i Here, too, temperature is one of the most impor-
tant design considerations. Seals can be classi-
acceleration ap) O-ring seals for typical dynamic and static appli-
P2 =vent pressure, psia. This is the ambient cations. Figure 7-26 shows a typical O-ring seal
pressure, plus the vent-line pressure drop used for a valve seat. Important design consid-
which also depends on flow rate (again a erations for dynamic O-ring seals are summarized
as follows:
function of acceleration ap)
Fr = resistance force of the control function, lb, 1. Design correlations between fluid pres-
which also may be a function of accelera- sure, O-ring compound hardness and diametral
Since the relations between p_, P2, Fr, and ap hardness. The combination of fluid pressure and
are not linear, equation (7-10) may become com- chosen hardness will determine the maximum
plex and require a high-speed computer for clearance E that can be tolerated safely. A
solution. proper combination of clearance and O-ring hard-
ness may prevent O-ring extrusion (table 7-5). In
general, the O-ring nominal section diameter is
7.7 DESIGN OF DYNAMIC SEALS FOR
chosen as large as space and installation con-
FLUID-FLOW-CONTROL COMPONENTS
siderations will permit. Past designs usually
Apart from the static seals, which will be will provide a useful guide. In addition, Teflon
treated in chapter IX, two basic types of dynamic backup rings as shown in figure 7-25 are recom-
seals are required for fluid-flow-control compo- mended for dynamic O-ring seals at sealing pres-
nents: seals for moving (reciprocating and rotat- sures over S00 psi and for static O-ring seals at
ing) cylindrical elements such as actuator pis- pressures above 1500 psi.
illtBl #:
29O DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
STATIC
SEALING\
_ ¢,-_ I'-- "1 D-GROOVE
\BREAKCORNERS, : _b -,J U:NGTH
?o APPROX.
o.oos' D l .-f
2. Surface finish requirements.-The finish of ter is recommended for surfaces in contact with
the sliding surfaces in contact with dynamic static diametral-squeeze-type O-ring seals.
O-ring seals should be as smooth as possible. 3. Friction of dynamic O-ring seals.-The
They should be ground, honed, or polished to a breakaway friction of a dynamic O-ring seal is
microinch finish of 8 to 10 rms. It has been usually about three times the running friction.
found that a finish within this range yields a Breakaway and running frictions increase with
longer life than either rougher or smoother fin- fluid pressure, O-ring hardness, diametral
ishes. Codirectional patterns, as produced by squeeze, and decrease of temperature. Accurate
honing, have been proven to be the best surface values of O-ring frictions can only be obtained
finish for any type of dynamic sliding seal. For experimentally for a given design.
still better results, after an initial finishing, the 4. Selection of O-ring compounds,-A great
surface could be hard chrome or nickel plated variety of O-ring elastomer compounds is avail-
and again finished. The plating provides a hard, able, with trade names such as Silicone rubber,
slippery surface that resists corrosion, wear, and Buna N, Butyl, Viton, Teflon, and Kel-F. The
scratching. A microinch finish of 60 rms or bet- selection of an O-ring compound and its physical
1/16 0.070± 0.003 0.010 0.015 0.057 0.052 3/32 1/64 0.005
3/32 0.103 *_0.003 .010 .017 .090 .083 9/64 1/64 .005
1/8 0.139 *_0.004 .012 .022 .123 .113 3/16 1/32 .006
3/16 0.210 *_0.005 .017 .032 .188 .173 9,f32 3/64 .007
1/4 0.275 *_0.006 .029 .049 .240 .220 3/8 1/16 .008
SEALS TO BE 12 RM$ OR
be applied effectively as seals for valve seats OETTER. ALL NICENSIOhS
to _ _H _NC_ES
(fig. 7-26). The resiliency of the O-ring absorbs
shock loads and seals tightly at all pressures,
Figure 7-27.-Typical rotating lip-seal design [or
even when some dirt and grit are present in the valve actuator shalt.
system. One design problem is to prevent the
O-ring from being blown out of the groove. This
can be prevented by providing a dovetail O-ring
groove in a two-piece valve poppet (see fig.
7-26).
VALVE STEM
-Acrvz. ,_ _%ulO
However, pressure levels and available space _.&LIN* 5t._f_3t5 LBO_H _LVE AN0 ,_v_t ) L_PPED N 8-10
Design of Dynamic Seals for High-Temperature Figure 7-30.-Metallic bellows used as recipro-
Services cating-type dynamic seals in a typical poppet
valve for high- and low-temperature services.
The metallic bellows (fig. 7-30) is most fre-
quently used as reciprocating-type dynamic seals
elastic modulus, and of approximately the square
for high-temperature services. Two types of
of the thickness of the material. It is also a
metallic bellows are distinguished: the hydraulic-
function of the outside-inside diameters and of
formed and the multidisk welded type. The
former is made of one to three plys of sheet the number of convolutions and their curvature.
metal and is designed for all pressure ranges. For maximum flexibility (inches of stroke/lb of
The latter is for relatively low-pressure services load), a minimum inside diameter combined with
a maximum outside diameter should be used.
and for high flexibility.
A metallic bellows of any type behaves, in Also, material thickness (within stress limita-
part, like a helical spring. The spring rate tions) and modulus of elasticity should be mini-
mum.
(lb/in of.movement) is a direct function of the
Generally, bellows design data, such as stock
DOUBLE LIP SEALS
size, allowable working pressure, spring rate,
FOR ACTUATOR PISTON
materials and service temperature, are supplied
DOUBLE LIP SEALS / by the manufacturers. Important design consid-
erations are discussed in the following:
3. Selection or materials.-Selecting of bel- gas throttle valve which was operated success-
lows material should be governed by fluid com- fully at temperatures ranging from 1200 ° to
patibility or corrosion considerations, operating 1S00 ° F at pressures up to 700 psia. An alter-
temperature ranges, and spring characteristics. nate design is to attach a flat-face graphite seal
Some high-temperature alloys such as stainless ring to the end of a metallic bellows which is
steels, Monel, Inconel, and Hastelloy B have welded to the shaft (fig. 7-31). Here, the shaft
proven suitable. misalinement is compensated by the flexibility
4. EHective area.-This is that area which, of the bellows.
when multiplied by a change in bellows length, The sealing of valve seats for high-temperature
yields the actual displaced volume. This area services is usually achieved by metal-to-metal
_ •. _ /- _.. _ _1_?,_,_
_ _,,_
HOT GAS
SPRING DISK | /
VALVE
/
VALV¢_ ...... _, _ sto,,
/ HOUSING
Figure 7-3I.-Turbine hot gas throttle valve with Figure 7-32.-Valve seat seals for high-tempera-
typical rotating-type dynamic seals. ture services.
_1 mi _rx
valve parts due to temperature, internal stress, control purposes, ability for continu,_usly vari-
and mechanical loading of the mating parts. able opening position may be required.
Figure 7-32(a) shows a convoluted-ring-type In addition to propellant compatibility and
valve seat seal used in a hot gas shutoff butter- structural integrity, prime design considerations
fly valve. Depending upon the specific applica- for propellant valves are:
tion, the convoluted ring may be made of high- (1) No leakage of propellant through the valve
when closed
temperature alloys such as Inconel-718. The
rings effect a leakproof seal in the closed posi- (2) Proper actuating time during opening and
tion, since the upstream fluid pressure tends to closing in accordance with the require-
expand the convolute and produces a high con- ments of the control system
tact unit force at the sealing surfaces. The (3) Minimum pressure drop
curvature of the convolute ring tends to maintain A great variety of propellant valve types is
a continuous contact with the valve seat. Figure available. Each design has certain characteris-
tics which make it suitable for a specific appli-
7-32(b) presents a metallic-spring-disk-type
valve seat seal used in a hot gas shutoff poppet cation. Frequently used propellant valves, clas-
valve. Again, the upstream gas pressure pro- sified according to their design configurations,
duces a high contact unit load on the sealing are:
surfaces. The valve seat has a curved contour (1) Butterfly valves
which effects a continuous contact with the flat (2) Ball valves
face of the seal disk. (3) Poppet valves
(4) Venturi valves
(5) Gate valves
Sealing Specifications (6) Needle valves
plsroN _ A
O"Rr
_'_pE _- -_ : i' _-AT FL;LLYOP_NEO PO_ITIO"+, _ i
+ ...... ; .... • •
/ -- \ \ I,o+,,,o.......... + ......,.+o
/ ++ ....... -
/ SEe:tON A--A\ u_ + I
L_.......o X__ ....... r---,_ %A
z,
+, I+.'-C +,.<
+- + "-:"+-._ : , _ + i"
T c = Tf- Th (7-14)
where
To = required opening torque, in-lb
FLOW
Tc = required closing torque, in-lb
(b)
Tf = friction torque, in-lb
DIRECTION
Th = hydraulic torque, in-lb (assumed to act in
the closing direction)
BUTTERFLY
VALVEGATE
The friction torque Tt varies with the pres-
sure differential across the valve gate, and with
AXISOF VALVESHAFT the valve gate projected area which is a function
./ of gate angular position. Friction torque can be
estimated by
(c)
FLOW
DIRECTION _ ____ T! = Ktrsfmds2hp (7-15)
\
where
BUTTERFLYVALVEGATE
K[ = friction torque coefficient, which is a func-
Figure ?-36.-Various locations of valve shaft tion of gate angular position (to be de-
axis with respect to butterfly valve gate. termined experimentally)
DESIGN OF CONTROLS AND VALVES 297
rs =radius of valve shaft at the bearing sec- estimated opening and closing torques. In addi-
tion, in tion, at the start of the opening stroke, the actu-
fm = coefficient of friction between shaft and ator has to overcome the static friction forces of
bearing (0.20 for aluminum journal and all seals. Butterfly-type propellant valves are
steel shaft; 0.05 for needle bearing and relatively fast acting. Opening and closing
steel shaft) times range from 20 to 200 milliseconds.
ds = inside diameter of valve seat lip seal, in
Ap=pressure differential across the valve gate,
Sample Calculation (7-3)
psi
Hydraulic torque Th may be estimated by The following design and experimental data
are given for the main oxidizer valve (butterfly
Th = Khds 3Ap (7-16) type) of the A-1 stage engine.
Design Data
where Kh ---hydraulic torque coefficient, which is
Radius of valve shaft at bearing section,
a function of gate angular position
rs=0.8 in
(to be determined experimentally)
Inside diameter of the valve seat lip seal,
Figure 7-37 shows plots of required opening
ds -- 7.7 in
and closing torques versus gate angular posi-
Coefficient of friction between shaft and
tions for a typical butterfly valve. In actual
needle bearing, f m =0.05
design practice, the actuator of a butterfly valve Test Data
will provide two to three times the maximum
0° 90 ° Solution
OPENINg"
From equation (7-15), the friction torques
TI= Klrsfmds2.Ap
at 5°: T[=O.78xO.8×O.O5×(7.7)2x1058
= 1960 in-lb
REQUIRED CLOSING TORQUE
at 15°: T[=0.78×0.8×0.05×(7.7) 2×769
' = -- )
= 1425 in-lb
o at 40°: T[=l.57x0.8×0.05×(7.7) 2×87.5
o
-'- 326 in-lb
I--
t
tooor_ I _ svtL
700_
i [ ,2"- ' P'
I - X-_-__ .._--n-Jh_'_i_-_,,--_ ' _-\1"'× ....
400---
=J
2_ 30" .t_"
ANGULAR
._" 60"
POSITION
70" gO" _+
= \!; ......
OPEN
Figure 7-40.-Typical large-size poppet-type
NOTE: Zero f_ from 0"- S'due to lea1 deslgn propellant valve design.
_P2 a5 -'%" PI
POPPET C/_VITY I \ _ _---'- VENTUR) for leakage to ambient. The small number of
I \ "----VALVE
POPP_ moving parts further enhances reliability.
CLEARANCE AREA LVALV E SPRING
VALVE SCHEMATIC
seal. These are suitable only for medium- actuate other fluid-flow-control components, such
other seal types are required. Gate valves are quence events, such as the admission of igniter
designed for propellant line pressures up to 3000 fuel. There is a great variety of control pilot
psi. Because of their relative bulkiness, gate valves available for liquid propellant rocket
valves are limited to low propellant-flow appli- engine services. Basically, they can be grouped
cations such as for gas generator control and into two categories: the on-off type and the pro-
shown in figure 7-43. This valve type is used Proportional-type pilot valves are also known as
for extremely low flow applications such as for servo valves.
attitude-control thrust chambers. The assembly
A pilot valve may be operated eleetrieally or
shown is a dual-valve arrangement, positively by fluid pressure, or through a mechanical con-
linked by a mechanical yoke. The valve body is nection with other control components. Impor-
an integral part of the thrust chamber injector tant design considerations for pilot valves are:
assembly. Both valves are normally closed. (1) Fast response
Their actuation is provided by a quick-response (2) No leakage of control fluid through the
electric solenoid. valve when closed
Sealing at the valve seat is achieved by the (3) Required actuating power souree compati-
elastomer tip of the valve needle. Dynamic seal- ble with systems design
ing at the actuator rods is achieved by means of (4) Sufficient output at the design point
metallic bellows. This seal design is compatible The output of a pilot valve can be defined as
with cryogenic as well as storable propellants.
The pintle vanes provide a guide for the recipro- W = Pd_' (7-17)
cating motion of the valve needle.
In chapter XI we will discuss other special where
valve types, as they are needed for very low W = pilot valve output at the design point,
propellant flow service in miniature-size space in-lb/sec
engines. Pd = valve control fluid discharge pressure at
the design point, psig
7.9 DESIGN OF CONTROL PILOT VALVES {/ =valve volumetric flow rate at the design
point, in3/sec
The main function of a pilot valve is to con- The most frequently used on-off pilot valves
trol a fluid which in turn is used to control or
may be classified according to their design con-
figurations into
(I) Two-way types
THR_T CHAMSI[R A F iNLET (2) Three-way types
iNJECTOR _lOOY
\ ..........
U(L
L. ELASTOMER TiP OF
propellant valve. The sequence valves shown in
T_E VALVE NEEDLE
figures 7-35 and 7-40 are typical examples of
Figure 7-43.-Typical needle-type propellant two-way-type pilot valves. Both examples use a
mechanical link, actuated from the main valves.
valve design.
302 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
Solenoid or fluid pressure operated two-way pilot developed by General Controls. Figure 7-45
valves are also frequently used. shows a typical design of a pilot-operated sole-
noid, three-way pilot valve developed by Skinner.
It uses a solenoid to control fluid flow to a dia-
Three-Way-Type Pilot Valves
phragm which opens or closes the valve. The
A three-way pilot valve (fig. 7-44) has three valve may be normally closed or normally open.
ports: inlet or pressure port, outlet or cylinder The selection of standard pilot valves is based
(actuator) port, and vent or return port. If the on the design data furnished by their manufac-
valve is designed normally closed (N.C.), the turer. For specific applications, modifications
fluid path between pressure and cylinder ports is can be incorporated into standard designs.
closed, while the path between cylinder and A typical fhid-pressure-actuated, three-way
return ports is open. Actuation of the valve pilot valve design is shown in figure 7-46. This
effects closing of one fluid path and opening of valve may be used as the ignition monitor valve
the other. The reverse is true for a normally in the A-1 stage engine control system. The
open (N.O.) valve; i.e., the fluid path between valve is held normally closed by a spring. The
pressure and cylinder ports is normally open. valve diaphragm is designed as a combined sens-
Most of the standard pilot valves furnished by ing and actuating diaphragm. During engine
specializing manufacturers are solenoid valves, start and when satisfactory main thrust chamber
actuated by electrical energy, A direct-acting ignition has been achieved, the pressure buildup
solenoid valve (fig. 7-44) is one in which open- sensed at the thrust chamber injector fuel mani-
ing and closing is controlled by solenoid only. fold will cause the ignition monitor valve to open
A pilot-operated solenoid valve (fig. 7-45) is one
in which the solenoid controls the flow of a
small portion of the pressure fluid, which in turn
operates the valve. This results in a smaller
CYL. VENT CYL. VENT
electrical current required to operate a smaller
PRESS PRESS
CONNECTION
- ELECR_CAL
5PRgNG
PffS_dRE
DIAPH Rt, GM -_.f
CLOSED _J AS_E_,_LY
• OffN ELECTRIC
POPPET _ ; VENT OR
OR
MAGNET
$OLENO_b
ENE_GIZEbOIt
by pressurizing its diaphragm. The opening of Valve characteristic flow area in the fully
the ignition monitor valve, in turn, directs the open position, A* : 0.19 in2; d d = 2.1 in;
fuel pressure to the main fuel valve actuator Ps=20psig; F/:141b
opening port. The valve spring can be calibrated Valve resistance coefficient at the fully open
corresponding to the effective diaphragm area, position, K= 3.5
so that the valve will open at a predetermined Required valve volumetric flow rate, _,: 200
sensed pressure. During engine cutoff, decreas- in_/sec
ing fuel pressure allows the ignition monitor Inlet port fuel pressure :350 psig
valve to close. This, in turn, vents the opening Determine the required preload of the valve
side of the main fuel valve actuator, closing the spring, Sp. Also, calculate the valve output W
valve. in the fully open position.
The valve poppet is balanced by internal fluid
pressure acting on a dynamic O-ring seal which Solution
2
The characteristic flow velocity of the valve
-_dd Ps = Ft+ Sp (7-18)
where
- 1052 in/see = $7.6 fps
dd: effective diaphragm diameter, in
Ps : rated sensed threshold pressure to open the
The density of the fuel (RP-1) is 50.451b/ft 3.
valve, psig
Substitute this and other data into equation (7-9).
Ft= static friction of the valve poppet, ib
The pressure drop through the valve at the de-
Sp = required preload of the valve spring, Ib
sign point
7 2
Sample Calculation (7-4)
Ap = K_p _- _ 3.5 x 50.45 × (87.6) 2
288g 288×32.2 = 146.5 psi
The following design data are given for the
ignition monitor valve of the A-1 stage engine The fuel pressure at the valve discharge
(fig. 7-46):
_rUtJ_ _,LrH _Ue_ _tES_UmE INlay p'Olr Substitute into equation (7-17), the valve
rO V_Nr eOl_
output
v[
PRESS PRESS
from the opening port. The valve can be closed
only by pressurizing the closing port, and vent-
ing the opening port.
CYL. I. RET. CYL.2. CYLI RET CYL2.
A'_
PILOT PASSAGE i TERMINAL BOX /-- COIL
, /\
SECTION B-B
INLET
Figure 7-48.-TypicaI design of a pilot-operated, lout-way solenoid pilot valve, developed by Valvair.
DESIGN OF CONTROLS AND VALVES 305
F= B_'A (7-19)
4 (ds _ - dt _) Pi = (spring forces - Ft) C
x contingency factor
B fPNI
(35+25+210x 0.05 :--C- (7-20)
+250x 0.05-24) × 1.5
ds_ : +(0.5) _
where
-_x 400 F : pull force on the plunger in its normal posi-
tion, lb
=0.281 + 0.25=0.531
d s = 0,73 in
G
SOtE_¢*_ east, i_o_ ./_
x contingency factor
B = magnetic flux density in the air gap, kilo- Substitute this and other values into equation
maxwells/in 2 (7-19):
A = plunger cross-sectional area, in 2
C = a factor comprising constants and allow- F=B2A. 27 B2x0.246
ances for stray flux. A value of 72 is C' 72
applicable to round, flat-faced, plunger-
type magnets The required flux density:
P=a factor comprising constants and the
permeability of the fluid in gap G between B = V27 × 72/0.246 = 88.8 kilomaxwells/in 2
core and plunger; a value of 0.00319 ap-
plies if an airgap is assumed
Substitute this into equation (7-20):
f = flux leakage factor. It is determined by
analyzing the magnetic circuit
N = number of coil turns
B=f-_; 88.8= 0"7 x 0"00319 x NI
I =electric current applied to the coil, amperes 0.05
G = gap between core and plunger, in
Solenoid actuators, particularly if energized The required ampere-turns for the solenoid
coil
for extended periods, must be designed with
sufficient radiating surface to prevent the tem-
perature from becoming excessive. To give the 88.8 × 0.05
required current, the resistance of the coil should NI = 0.7 × 0.00319 = 1990 ampere-turns
be based on its maximum temperature. Suitable
protection, such as seals, should be provided to Assume a current of 1.4 amperes for the sole-
prevent the solenoid from becoming contaminated noid; then the required number of turns
with propellants. If the plunger is designed to
bottom-out against the core when energized, it is
advisable to provide for a thin wafer of nonmag- N 1990 1990_ 1425 turns
I 1.4
netic material at its face to prevent sticking.
as orifices, nozzles, and venturis, are frequently Orifices of the VDI type are preferred because
used in rocket engines for fluid system calibra- their behavior is more predictable. This is at-
tributable to the bevel at the backside, which
tion and for control purposes. Although the
energy loss caused by an orifice is high, as prevents erratic wall reattachment of the con-
compared to that of nozzles or venturis, orifices tracted jet within the orifice. It is important
are extensively used because of their compact- that the leading edge be absolutely sharp. The
value of k is a function of the diameter ratio
ness and simplicity (fig. 7-2).
(dod,). Tile (Apo/Ap) versus (do/d 1) relation-
ship is plotted in figure 7-52.
Orifices and Nozzles for Liquid Flow
, AIAIAIAIAIAIAIAIAI_Ap
(7-21)
_'= 0.0438 Cdo2v/_"L'_= 0.525 Cdo-_p
APo
k: Ap (7-23)
where
:volumetric flow rate, fta/sec
_i, =weight flow rate, lb/sec
p = density of the fluid, lb/ft a
Ah = static pressure head drop across the ori-
fice or nozzle, ft
Ap = static pressure drop across the orifice or
nozzle, psi
FLOW
do =diameter of the orifice or nozzle throat, in
dI =diameter of the duct leading to the orifice
or nozzle, in
C =flow coefficient for orifices or nozzles.
Considers the effects of discharge jet
contraction in orifices, velocity of ap-
proach, diameter ratio do/d z, friction
and flow profile
Apo = total (or permanent) pressure drop charge-
able to the orifice or nozzle, psi
k = factor, correcting nozzle or orifice work-
ing pressure to permanent pressure loss
The correlations are still reasonably accurate
for gas flows, if the pressure drop across the Figure 7-51.-Typical designs and flow coeffi-
control orifice is small (<10 percent). cients C of flow regulating orifices. A, VDI
The value of flow coefficients C is a function type; B, square edged; C, VDI type flow
of the design configurations and flow Reynolds nozzles.
3O8 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
07 \1'%1
Estimate the required diameter d o, (a__)for a
VDI-type orifice, and (b_) for a VDI-type flow
nozzle (figs. 7-51(a) and 7-51(c)).
<3 0.6 \
I!
0.5 \
Solution
0.4 ,\
The average flow velocity in the fuel duct
892
0.3 , \
V--
- _r 7 2=66.3fps i i
do/d I ( NOZZLE )
From table 6-3, we obtain the viscosity of K
I-do/d I FOR ORIFICES
RP-1 =3.22× 10 -7 lb-sec/in 2, or #=3.22× 10 -7
× 4636.8= 1.494 x 10 -3 lb/ft-sec (see eq. 7-6).
Figure 7-52.-(Apo/Ap) versus (dod1) [or ori-
The Reynolds number, based on duet diameter
fices and nozzles.
dl
Re _ Dvp _7 x 66.3 × 50.45 value of 0.32 for k is obtained from figure 7-52.
/1 1.494 × 10 -3 = 1.306 × 10 6
Substitute this into equation (a):
From figure 7-51(a), a value of C = 0.67 is Thus, for the VDI orifice, do=4.47 in; for the
derived for R e = 1.306 × 106. A corresponding VDI flow nozzle, d0=3.55 in.
DESIGN OF CONTROLS AND VALVES 3O9
Orifices for Gas Flow Values for Z applicable to equation (7-24) for
various pressure ratios (p2/pl) are presented in
The basic correlations for an isentropie, com-
figure 7-53.
pressible gas flow through an orifice are:
Solution
I_ [ £ Y+'I
For a downstream pressure of 14.7 psia, the
(7-26) orifice pressure ratio
where
P f) :500+14.7
14.7 =0.0286
=gas weight flow rate, lb/sec
A =orifice area, in 2
Pl =gas pressure upstream of the orifice, psia From figure 7-53, the compressibility factor Z
P2 =gas pressure downstream of the orifice, is derived as 4.11 for Y--1.66. Substitute this
psia and other parameters into equation (7-24):
CAp_Z 0.6x_(O.O6)-x(500+14.7)x4.11
Z = compressibility factor, ft aS/sec
y =gas specific heat ratio %/-R-T x/386 × (460+ 100)
0.3
P{ 2'__ 350 _
514.7 0.682
CO_P_6$SI_ILITY FACTOR Z, F, 0'5 APPUCASLE TO EQUATION ;7-221 Fixed-area-type regulating devices have
sec
definite limitations. For instance, an orifice
Figure 7-53.-Compressibility factors for an isen- regulates flow and pressure only under specific
tropic compressible flow through an orifice. conditions of flow, but does not function under
310 DESIGNOF LIQUID PROPELLANT ROCKET ENGINES
(1) Throttle valves (including valves for The open-loop gain and phase shift versus
frequency characteristics of a typical servovalve
thrust and PU control)
and driving amplifier combination are shown in
(2) Gas pressure regulators
figure 7-54. These characteristics are obtained
(3) Liquid flow regulators
by applying an input signal to the servoamplifier
Throttle valves have been discussed in sec-
from an oscillator. The amplifier drives the
tion 7.8. Detail on gas pressure and liquid flow
valve by means of a current input to the valve
regulators will be found in sections 7.12 and
transducer coil (torque motor). In turn, the valve
7.13. Many of these devices use some form of
controls the flow of working fluid to the actuator
fluid-pressure-operated actuator. The position
which produces the desired load force. The
of the actuator, and thus the area of the con-
voltage output from a potentiometer attached to
trolling valve opening, is effected by applying a
the actuator is then compared to the amplifier
pressure differential across the actuator piston
input. Instead of electrical feedback, mechani-
or diaphragm, by means of various types of servo
cal feedback may be employed. (See sec. 7.5,
pilot valves, which will be discussed. The
function of a servovalve is similar to that of an "Engine Thrust Vector Control.')
IIF,.
i!INPUT An = 7rdnX (7-27)
sGu_PPLy .____fl (_I _-_-[__
where
PRESSORE
---7 F;/:i-F-q A n =effective flow area, in 2
dn =diameter of the nozzle, in
X =displacement of the flapper from the noz-
FIXED / NOZZLES--" zle, in. (Maximum value should be less
ORIFICE J TRANSDUCER J
than dn/5.)
Figure 7-55.-Schematic of a typical flapper-
nozzle-type pneumatic servovalve.
Sample Calcul'ation (7-9)
other. An increased flow reduces the fluid pres- The following dimensions and data are de-
sure (compressible fluid) or fluid volume (incom- fined for a flapper-nozzle pneumatic servovalve
pressible fluid) on the corresponding side of the (schematically shown in fig. 7-55), which is used
actuator piston. Correspondingly, the fluid pres- as a pilot valve of the servo PU control valve
sure or fluid volume on the other side of the attached to the main oxidizer valve of the A-2
piston increases. The resultant pressure differ- stage engine:
ential across the actuator piston causes it to Helium supply pressure, Ps = 500 psia, and
move in the desired direction. The flapper- temperature T = 560 ° R
nozzle valve is also applicable to servo systems d z,A,,d 2'A 2, d 3, A 3, d 4,A 4=corresponding
with single-control nozzle bleed. Here, the actu- diameters and flow areas of fixed orifices
ator position is controlled by regulating the and nozzles
actuating fluid on one side of the piston or dia- d, =d 2 ds =d 4
phragm only. This is analogous to the single- #¢,, 0¢2, _i'3, w4, _i'o, VCc=fl°w rates through
bleed pneumatic poppet servovalve (fig. 7-57). fixed orifices and nozzles, and to and from
Flapper-nozzle valves are particularly suit- the actuator, lb/sec
able as pilot valves for larger servovalves (see Z,, Z 2, Z3, Z 4 =compressibility factors of
fig. 7-9). Because the transducers or torque the flows through the orifices and nozzles
motors for these valves require rather low power Flow coefficient of the orifices and nozzles,
levels, they usually consist of coil relays exert- C=0.7
ing forces of only a few ounces. The effect of Distance between the two nozzles,
the flapper spring rate is often counterbalanced T=(d3/4)+ t
by the gradient of the magnetic force developed t=thickness of the flapper=0.004 in
in a properly designed transducer or by mechani- At the neutral or equilibrium position of the
cal means. To prevent spreading of the jets servovalve:
leaving the nozzles and to ease the rate balanc- The pressures in the actuators, Pc = PD= 450
ing between flapper spring and transducer mag- psia
netic forces, the travel of the flapper should be The bleed through nozzles, w3 = w4 =0.000778
kept reasonable small. lb/sec
Equations (7-22) and (7-24), which describe Determine:
the flow of liquids and gases through orifices (a__)The dimensions of fixed orifices and noz-
and nozzles, are applicable to the design calcu- zles, and of distance T.
lations of flapper-nozzle valves. Two conditions (b) The pressure differential across the actu-
may exist for the flow through the nozzles of a ator piston when the flapper is deflected down-
flapper-nozzle valve: first, when the restriction ward 0.001431 inch from its neutral position, and
is determined by the position of the flapper; and the flow rates _i_o and Wc to and from the actuator
secondly, when the flapper has moved far enough are 0.00021 lb/sec (as governed by the speed of
from the nozzle for the flow to be restricted by the piston).
"
312 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
CA4PcZ4
A 3 = ,,-rd3X = ,7× 0.0266 × 0.004756 = 0.000397 in s
#4- V_R-T
- -O'O062A4Pc
2 _ 2 #3 =#I +#o
A, =-_d, , A 2 =-_d2 ,
0.0062 Asp o = 0.753 A,Z 1 +0.00021
A s = vd_X, (T - t - X)
A4 = ,'zd4 0.0062 ×0.000397 p o =0.753 × 0.000427 Z 1+ 0.00021
#c = #o = O; #1 = #2 = #3 = #4 = 0.000778 lb/sec;
We use a trial-and-error method to find these
T-t ds values for Po and Z,, which will satisfy the
X-
2 8 above correlations and figure 7-53. We find that
Po 436.5
fi' 1 0.000778 = 0.875,
Ps 500
A, = A 2 = 0.753 Z, = 0.753 x 2.418 - 0.000427 in 2
=0.0001583 in 2
found that
Pc --464 psia (8)_E3,GN CO*W_ENSaT;ON _0_ _ _XrALLY
UNBALANCED FC)RCEO
Z 2 = 2.0808
Figure 7-56.-$pool-typeservovalve.
Thus the pressure differential across the Although the spool valve theoretically is
actuator piston force balanced, hydrodynamic and friction forces
cause relatively large loads which must be over-
P=Pc- Po =464-436.5=27.5 psi come. Refer to figure 7-56(a). Due to the dif-
ference in flow velocities, the static pressure at
face A will be less than that at face B. Simi-
Spool-Type Servovalves
larly, the pressure is less at face C than at face
The spool-type servovalve (schematically D. This results in two approximately equalaxial
shown in fig. 7-56) is basically a four-way valve. forces, both of which tend to move the valve
A cylindrical valve spool is accurately fitted spool to the right so as to close the valve.
into valve inserts, which in turn are shrunk into These unbalanced axial forces can be compen-
the valve body. Both inserts and spool are made sated by design remedies. One is to increase
of hardened alloy steels. The thickness of the diameter d2 (as shown in fig. 7-56(b)). It is rec-
inserts in the axial direction, and thus the loca- ommended that the maximum control port flow
tion of the ports, is held to very close tolerances area rrd,X (where X= spool displacement, in) be
by lapping their faces. The outside diameter of just less than twice the annular area between
the inserts is accurately ground for a tight seal spool diameters d_ and d2. As a result, the flow
with the valve body. The surfaces of their axial velocity along the spool is substantially in-
passages are also lapped. The diametral clear- creased and the axial forces on faces B and D
ance between insert and spool is of the order of are considerably reduced.
0.0002 inch, at which the spool must still slide The difference between minimum flow rate
freely. The axial location of the spool lands (leak flow in neutral position) and maximum flow
must also be closely controlled. To minimize rate (actuator in motion) is substantial. Various
leak flow in the neutral position, the spool lands means of adjustment may be employed, such as
may be designed for slight overlap. As a rule, simple relief bypass valves, or the pump output
leak flows are less when the spool is displaced, may be adjusted. For instance, in a piston pump
due to better isolation of the drain lines. A operated by a wobble plate, the pitch of the plate
typical leak flow rate in neutral position is 0.5 may be adjusted as a function of pressure to
gpm. give strokes varying from maximum to zero.
. • • ~
The correlation between pressure drop and To find the maximum valve output design
flow in a spool valve is not as predictable as point, this expression is differentiated and set
one might expect from a sharp-edged orifice. equal to zero
Experimental.data are required to verify a de-
sign. Equation (7-24) for gas flow orifices may
_-_=2000- 20 - 22.86 ¢,2= 0
be used to approximate the flow through a pneu-
matic spool valve, where orifice area A= _d,X or
and flow coefficient C = 0.8. For a spool valve 22.86/r2+20 P- 2000= 0
using hydraulic oil or RP-1 as the actuating
fluid, the following empirical equation applies: -20 + X/202 - 4 × 22.86 × (-2000)
2 × 22.86
] p2
Ap = I_ + (I + ___21/z2 (7-28) -20 + V400 +183 000
= 8.95 in3/sec
where 45.72
Ap = valve pressure drop, psi
Ap= 10×8.95+7.62× (8.95) 2 = 89.5+610= 699.5 psi
= valve volumetric flow rate, in3/sec
X = spool displacement, or valve opening, W = (2000-699.5)×8.95=11 620 in-lb/sec=l.764 hp
in
p = density of the liquid, lb/ft 3 Itshould be noted thatstabilizedflow condi-
a, b, c = empirical constants depending on the tions rarelyexist,because of mass inertiaslim-
design itingaccelerationand deceleration,and feedback
effects.
/_____,VALVE POPPET
Valve pressure drop and flow characteristics GAS SUPPLY ___ OUTPUT CONTROL
PRESSURE _ TO
may be obtained from equation (7-28), for the REGULATOR VALVE
b = 1270 ORIFICE l
/-BALANCE PISTON
c=1.79× 10 -7
Maximum spool displacement, X = 0.009 in Io)
Determine the flow rate _,, pressure drop Ap,
and output W of the valve at maximum displace- ACTU,_TEO BY REGULATOR
CONTROLLER
ment.
VALVE ST EM---__
Solution
DYNAMIC SEAL
Ap=[_-_+(12_+ 1"79×10-7''2]
(0._9-_ ]v ] ×(50.45) 2 PRESSURE P=----_"_VALVE POPPET
BALANCE
(b}
From equation (7-17), the valve output Figure 7-57.-Schematics o[ typical single-bleed,
poppet-type, pneumatic servovalves used in
W = pdi' =(Ps - Ap)/_: 2000/_ - 10 _2 _ 7.62/_3 gas pressure regulators.
DESIGN OF CONTROLS AND VALVES 315
Single-Bleed, Poppet-Type Servovalves between actual and desired pressure. Its output,
called the error signal, can be directly applied
The single-bleed, poppet-type servovalve
mechanical force, or control pressure output to a
operates as a variable orifice like the flapper-
servovalve. Figure 7-58 presents the schematic
nozzle valve. Figure 7-57 presents, schemat-
of a typical gas pressure regulator controller.
ically, the principle of operation of typical
Here, the pressure being regulated is sensed
single-bleed, poppet-type, pneumatic servovalves
externally by a bellows which is internally evac-
as used in gas pressure regulating services.
uated, or vented to atmosphere. The vacuum
Two basic configurations are in use. The first
establishes an absolute pressure reference, while
(fig. 7-57(a)) effects output control pressure Pc
a vented bellows uses ambient pressure for ref-
regulation through variation of bleed port flow
erence ("gage pressure"). As the regulated pres-
area. In the second (fig. 7-57(b)), the bleed port
sure (the sensing pressure) changes, the bellows
area is fixed. Here, Pc is regulated by varying
deflects against the calibrated internal spring
the supply gas flow rate.
load, simultaneously positioning a directly con-
The selection of configuration depends on
nected servovalve (fig. 7-57), which in turn regu-
application. To minimize unbalance forces, a
lates a control pressure output.
balance piston is usually attached to the valve
A regulator valve consists of the control
poppet. The area of the balance piston is made
valve and an actuator. If only limited accuracy
equal to the projected area of the poppet seat
is required, or very small capacities are involved,
diameter, less that of the valve stem.
it suffices if the regulator controller develops
the error signal directly as a mechanical force to
7.12 DESIGN OF GAS PRESSURE position the regulator valve. Such a directly
REGULATORS spring-loaded pressure regulator is represented
schematically in figure 7-59(a).
Basically, gas pressure regulators are
Where greater accuracy is required, the regu-
variable-area-type pressure-reducing valves.
lator valve actuator is positioned by controlled
Their prime function is to maintain constant
pressure from a servovalve connected to the
pressure at their outlet, or in a downstream
regulator controller.
region, even though the pressure at their inlet
Since the servovalve amplifies the regulated
may vary (decrease). Gas pressure regulators
pressure errors, small errors in regulated pres-
may be operated independently, or in conjunction
sure will cause large changes in its control
with pressure relief devices, under either dy-
pressure output. This control pressure Pc can
namic or static conditions. For example, the
then be applied to control the regulator valve
gas pressure regulator for the A-4 stage propul-
position in the following ways:
sion system (fig. 3-9) is designed to maintain a
(1) Control pressure Pc is used as the loading
constant main oxidizer tank pressure of 165 psia,
pressure for a simple dome-loaded pres-
while the helium gas pressure at the regulator
sure regulator, as shown schematically
inlet varies from 4500 to 245 psia. Additional
protection is provided by a tank relief valve,
should the tank pressure continue to rise with
/-- SCREW ADJUSTMENT FOR
the regulator completely closed, because of such
effects as aerodynamic heating. j_// SPRING CALIBRATION
Most pressure regulators include two basic SENSING PRESSURE _'_ "_
ou . T LO O . .soo O
SERVO - VALVE ATTACHED TO THE REGULATOR CONTROLLER.
_,_11_.\\\\_ ---T----_OuTLET Pr
_'_----_OUTLET Pr _----_"__--OUTLET Pr
REGULATOR VALVE
stream systems) o 2
The basic test of a good regulator design is TIME, SEC TIME, SEC
diaphragm by increasing or decreasing the con- spring preload. Under these steady-state condi-
trol pressure Pc which operates the actuator dia- tions, gas continues to flow from the bleed regu-
phragm of the main regulator valve. This valve lator through the servovalve at a rate determined
is a normally closed, dome-loaded pressure re- by the restrictions and out to the outlet manifold•
ducing valve. The regulator controller circuitry has what is
In operation, the following sequence of events called a negative gain. A sensed increase in
occurs. Gas enters the regulator through a filter pressure Pr causes an amplified decrease in con-
located at the regulator inlet port, but is pre- trol pressure Pc, with attendant decrease of the
vented from entering the main regulator control main regulator valve opening. Similarly, a de-
pressure dome by the closed start pilot valve.
crease in Pr causes an increase of valve opening.
Upon opening of the latter by energizing its The controller gain can be defined as
solenoid, gas flows through the bleed regulator
and is reduced to pressure Ps, as determined by Ape
the preset reference spring force. The gas then G= - Ap---_- (7-30)
flows through a fixed restrictor and passes into
the control dome. This tends to open the main where
regulator valve against the actuator spring. G =controller gain
Dome pressure Pc is controlled by varying the Ape = change in control pressure, psi
flow area of the flapper-nozzle servovalve which Apt = change in sensed pressure, psi
bleeds the loading gas into the main regulator The principal design parameters of the gas
outlet manifold. When the sensed regulated out- pressure regulator shown in figure 7-61 are listed
let pressure Pr and the sensor spring force are in in table 7-6.
equilibrium, the servovalve flapper is positioned Fig-ure 7-62 presents the design layout of a
a sufficient distance off the nozzle seat and
typical dome-loaded, zero-gain-type gas pressure
maintains a steady-state control pressure Pc. regulator. The main regulator controller con-
This pressure is always greater than Pr, as de- sists of a spring-loaded bleed regulator and a
termined by the main regulator valve actuator fixed-area bleed orifice, which gives a constant
REGULATOR CONTROLLER
SENSO_ ASSEMBLY
ORb=ICED _ARTITION
DYNAMIC SE_L
_ -..._,
_r
_ REGLJLATOR INLET,
Pi
to regulate the propellant tank pressures of a extraneous forces). This, in turn, gives a con-
pressurized gas propellant feed system such as stant regulator valve opening (or closing) veloc-
the A-4 stage propulsion system. Main oxidizer ity and a constantly increasing (or decreasing)
DESIGN OF CONTROLS AND VALVES 321
Design of Liquid Flow Regulators A new equilibrium is achieved as the inlet pres-
sure increase is absorbed by an increase in
Liquid flow control can be obtained with a pressure drop across the piston throttle port.
venturi. The pressure at the throat is propor- The following design correlations are estab-
tional to tile velocity of the (incompressible) lished for this flow regulator:
fluid and thus the flow rate. The pressure dif-
ferential between throat and inlet can be used to
Fs =(Api +Ap2)Ap (7-31)
control the position of a butterfly valve, and
thus the flow rate, by means of a servocontrol Ap
circuit, A regulator of this type is shown sche- K s = -_ Ap (7-32)
matically in figure 7-18.
? " !
_RING
calibrated reference spring. The valve opens as
the pressure force acts against the spring force.
In the second, the main relief valve actuator is
MYLAR GASKI[TS
controlled by a pilot valve which is calibrated
[N_ A for the desired relief pressure setting. The
DIIP_N I_ ASSEMSLV
_ O- RING
(o)
__ LOCK NUT
SOFT o- ,NOo,
DISC
INLET
POST-TRAVEL C D
_ PRE-TRAVEL
Figure 7-G7.-Coned-disk-spring, force-deflection
__ ADJUSTMENT
"; r/Z_.--_',.,%-_./r_ _ STACKOFCONEO-O_SC Curve.
I _',._ _('_/JL_'%('T/_"_ ]/'/J / ." SPRINGS IN SERIES OR
c t PRE,SU E
_'_ CC_NT RO(_ ORIFICE
do, %
P[STONflNG SE_-L"_ \
JRELIEF_ . d
VALVESEAT
---/ Zl _
m
and by the tank pressure. This correlation can Force balance equations of the main relief
best be illustrated by the following sample cal- valve poppet at various conditions:
culation. (1) Basic equation:
The following design data are given for the (2) Condition at cracking:
A-4 stage propulsion system main oxidizer tank
relief valve (schematically shown in fig. 7-68): Fsm+ PccA2- Ptc (A2 - A1)- paA3 = 0
Tank pressurant temperature, T--700 ° R (he-
lium gas)
(3) Condition at any intermediate valve posi-
Relief pressure set point, Pt= 165 psia
tion:
Required maximum flow capacity, w = 3 lb/sec
Main valve flow coefficient, C = 0.75
Fs m + K s mXm + Pc A 2 - Pt (A 2 - A 1) - Pa A 3 = 0
Estimated leakage past the main valve actu-
ator piston seal, _i's = 0.003 lb/sec
(4) Fully open condition:
Control orifice diameter, d o =0.080 in
Flow coefficient for control orifice and pilot
Fsm + KsmXmo+ PcoA2 - Pro(As -A,)- PaAa =0
valve, Co=0.6
d, A = Diameters and areas of various ele-
(5) Condition at start to reseat:
ments, in; in s
Pc = Control pressure, psia
Fsm+KsmXmo+PcrA2- Ptr (A2- A1)- paA3 =0
Pa = Ambient pressure (14.7 psia maximum)
Determine:
(6) Fully reseated:
(a) Dimensions of the main relief valve and
force balance equations for various
Fsm + PcrA2 - Ptr(A2-A1)-PaA3 =0
conditions
(_b) Dimensions of pilot valve poppet and where
actuator, and force balance equations
Fseat = main valve seating force, lb
Fsm, Ksm : preload and rate of main
Solution valve spring, lb; lb/in
Xm, Xmo = main valve poppet travel in-
(a__)Main relief valve.-The valve is sized for termediate and fully open
an isentropic compressible flow through an ori-
position
fice. From equation (7-24), the main relief valve
Pcc, Pco, and Pcr:Control pressures at crack-
port area
ing, full open, and reseat
of the main valve, psia
A3- CptZ
Ptc, Pto, and P_r =tank pressures at cracking,
full open and reseat of the
Since Pa/Pt= 14.7/165 = 0.089, Z = 4.11 (from main valve, psia
fig. 7-53) (b_.) Pilot valve.-The pilot valve flow capac-
3 × V;386 x 700 ity must be larger than the combined flow rate of
A3 =0.75× 165× 4.11=3'07 in 2 control orifice and leakage past the main valve
actuator piston seal for adequate venting of Pc.
d 3 : 1.98 in'_2 in From equation (7-24), the flow rate of the control
orifice
Required minimum travel of the main valve to
the fully opened position CozrdoSPt z 0.6x 0.00503x 165×4.11
_i,o = 4 _/-R-T V386 x 70-----'-O
A3 3.07
Xmo =-----=
-rzd3=---_= 0.487 in_0.5 in = 0.00394 lb/sec
DESIGN OF CONTROLS AND VALVES 325
The total flow into the control cavity, _i,0 + Ws 7.15 DESIGN OF MISCELLANEOUS VALVES
= 0.00394 + 0.003 = 0.00694 Ib/sec.
Setting the flow capacity of the pilot valve 50 Check Valves
percent higher, its flow capacity results as
The prime function of check valves is to
allow fluid flow in only one direction. Tilere are
_i,p= I. 5 × 0.00694 = 0.0104 Ib/sec
two basic check valve types: the poppet alid the
swing-gate type. The selection depends to a
It is desirable that this flow be maintained
great extent on application. General design con-
freely at all points, independent of back pres-
siderations for check valves are-
sures. This is achieved by maintaining sonic
(l) Type of fluid and its pressure and tem-
velocity, i.e., critical or supercritical pressure
perature
ratios, at the restrictions. Thus, tile maximmn
(2) Required flow capacity
allowable control pressure
(3) Allowable pressure drop
(4) Allowable rate of leakback (including zero
_/ 1.67
leakage)
ky----r (. 2 = 80.6 psia (5) Space envelope and line connection
= t oT/ (6) Simplicity of construction
Figure 7-69 presents the design of a typical
Based on this pressure, and using equation poppet-type check valve. A light, compression
(7-24), the required minimum pilot valve port area return spring normally holds the poppet in the
results as
closing position. This prevents any possibility
of fluid leaking back. When fluid pressure is
A wp RV_--T 0.0104 _,3/'3_ × 700 introduced upstream, the poppet will open against
P= C--_cZ = ff-.6_8O.-.-.6x4.11 = 0.0272 in 2 the return spring. Because of tlleir relatively
high-pressure drop, check valves of this type are
dp=0.186 in_0.2 in
used only in low-capacity services. Either elas-
tomer O-rings or metal to metal seals are used.
Required travel of the pilot valve to the fully
open position Figure 7-70 presents a swing-gate-type check
valve. The valve consists of two elements: the
inlet and the outlet body. The swing gate is
Xp o Ap 0.0272 secured to the inlet body. A torsional-type re-
- _dp- _x 0.2 =0.0433 in
turn spring holds the gate in the closing posi-
tion. Swing-gate check valves minimize pressure
drop. However, positive sealing against back-
Force balance equations of the pilot valve flow is more difficult. In some applications,
poppet and actuator:
(1) Poppet:
S_ CLOSURE PLUG
(2) Actuator:
_PET
where
Fp = pilot valve poppet seating
force, lb
Fa =pilot valve actuating force, lb
Fb = sensor bellows force, lb OPEN CLOSED
" ii.....
326 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
UPSTREAM
WING-GATE BURST FLUID P_RESSURE
HINGE
DOWNSTREAM _ UPSTREAM
V-GROOVE COINED IN
A NORSESHOE PATTERN
I _ _,L_E=,L
Burst Diaphragms
They are especially useful in storable liquid _VALVE BODY _'-- BURS T D_APHR_GM
(S) BURST _IAPHRAGM ATTACHED TO _l,
propellant engine system applications; they also PROPELLANT VALVE POPPET
Liquid propellant rocket engines and the dynamic loads. In some designs the tanks are
tanks feeding them, together with certain auxil- further stabilized by the internal pressure against
iary systems such as pressurization, form the buckling; i.e., the wails are always kept under
propulsion system. For reasons set forth in tension loads by a specified pressure level main-
section 1.5, a discussion of tank design is in- rained during storage and handling. In smaller
cluded here. The need for close coordination units, the wails are usually capable of taking
betweeri engit_e and tank designer is reempha- external loads without being pressurized in_er-
sized, particularly if they are members of dif- nally. Operational tank pressures for these range
ferent contractors. from 4OO to 2000 psia.
Prepackaged storable liquid systems are usu-
ally employed to relatively short-duration, low-
8.1 BASIC DESIGN CONFIGURATIONS OF thrust applications. Since the tanks form an
PROPELLANT TANKS integral part of the propulsion system, they are
The configurationof propelhmt tanks de- designed and furnished by the engine builder. In
pends largelyon vehiclemission and size. In one design, the thrust chamber is located inside
most modern designs, the tanks form an integral the aft tank and welded directly to the tank
bulkhead.
part of the vehicle structure. Propellant tanks
can be categorized, according to vehicle appli-
cation, as follows: Propellant Tanks fur Booster Stage Systems
(1) Prepackaged storable liquid systems Figure 8-2 presents the propellant tank design
(2) Booster stage systems configuration of a typical propulsion system as
(3) Upper stage systems used in the booster stages of a large launch
vehicle such as the Saturn V. The system shown
can be used for either storable or cryogenic pro-
Propellant Tanks for Prepackaged Storable
pellants. The tanks are arranged in tandem;
Liquid Systems
their wails form an integral part of the vehicle
Figure 8-1 presents the configuration of a structure. For booster application, overall ve-
typical prepackaged storable liquid propulsion hicle systems optimization usually dictates use
system. The tanks are arranged in tandem, with of a turbopump-fed engine system. This permits
a common bulkhead between. This system is relatively low operational tank pressures, rang-
designed for long storage periods, perhaps 5 to ing from 30 to 100 psia.
10 years. A main characteristic of these sys- Since the tanks represent a large percentage
tems is that the propellants are factory loaded of the vehicle structural (inert) weight, advan-
and are hermetically sealed in the tanks by burst tage is taken of the low-pressure levels by con-
diaphragms. Both tank and diaphragm construc- structing the tanks with extremely thin-wall
tion materials must be compatible with the pro- thicknesses. However, the often huge tank
pellants for the storage duration. In the exam- structures thus become sensitive to external
ple, the propellants are expelled by pressurized buckling loads. To stabilize the tank structures
gases produced by a solid propellant gas gener- of large booster stage systems, two basic design
ator. The tank walls form an integral part of the avenues are open: pressure-stabilization and
vehicle structure and are designed to withstand self-supporting structures. In the pressure-
the internal pressure loads as well as the vehicle stabilized systems, such as the Atlas ICBM, the
329
330 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
GRAIN
DIAPHRAGM
fUEL TANK
COOLING
THRUST CHAMBER
ASSEMBLLES
PRESSURE
RELIEF TYPE
REGULATOR
VENT
Figure 8-1.-Propellant tank design configuration of a typical prepackaged storable liquid propulsion
system.
l ,, ,,
[ :
, i!
D_;TwR
UC T URET } FUEL L__ ER __2111E22D I_IC_2_ON D UCT DUCT SKIRT MOUNTS ! M'PEFE
Figure 8-2.-Propellant tank design configuration of a typical booster stage propulsion system.
DESIGN OF PROPELLANT TANKS 331
tank pressures must be constantly maintained are contained within an outer cylindrical shell,
above a specified minimum by elaborate con- through which thrust is transmitted to the pay-
trols. This tank structure basically is a thin- load. The shell is designed to withstand all
wall monocoque, requiring special handling anticipated boost and flight loads. The propel-
procedures. In most booster-stage systems, the lant tankage consists of two individual units.
propellant tanks are self-supporting types, the The main fuel tank is located forward, and the
tank walls being reinforced by skin stringers main oxidizer tank near the aft end. The two
(fig. 8-2), or by other structural means, such as welded aluminum-alloy tanks are modified
waffle grid patterns. spheres, faired into conical sections at the bot-
When cryogenic propellants are used, tank tom, for propellant discharge. The tanks are
insulation may be required. Insulation is manda- bolted to the shell structure around their support
tory in liquid hydrogen system to prevent ambient ring. The thrust chamber assembly is located
air liquefaction which causes high heat transfer just below, and closely coupled to, tile oxidizer
rates with attendant high boiloff rates, and safety tank. Thrust is transmitted to the outer shell
hazards. through the aft half of the oxidizer tank. }5oth
P£YLOAO LN2VEN T
OXID*ZER TANK
VENT
SECTION A-A
Figure 8-3.-Propellant tank design coMiguration o[ a typical upper stage propulsion system.
332 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
pressurization. The gas is stored at an initial vessel for a given volume is a spherical shell,
pressure level from 4500 to 5500 psia at -300 ° F since it has the smallest surface to volume ratio.
in two liquid-nitrogen-jacketed, high-pressure It also has the smallest shell stress for a given
spherical bottles located between the two main internal pressure. While a sphere may be the
propellant tanks. lightest pressure vessel, the combination of
It is for upper stage propulsion systems that several spheres into the generally cylindrical
the engine designer may most likely also design envelope typical for most rocket vehicles would
the tankage. The following discussions, there- cause a sizable volume penalty. Furthermore, a
fore, will confine themselves to these systems. sphere would preclude the use of the tank wall
By contrast, the design of booster tanks will as a load-carrying member of the vehicle struc-
probably always be made by an independent ture (figs. 8-1 and 8-2), resulting in further
group. However, many of the design principles weight and volume penaIties.
presented here are equally applicable to booster Thus both vehicle configuration and tank
tanks and may thus further the understanding of pressure level will determine the shape of pro-
their design problems. pellant tanks. For vehicles of relatively large
length-to-diameter ratios and of limited space
8.2 GENERAL DESIGN CONSIDERATIONS envelopes, cylindrically shaped tanks are used.
FOR PROPELLANT TANKS For relatively high tank pressures and less
stringent space conditions, spherical tanks may
The tank design is greatly influenced by be empIoyed to best advantage (fig. 8-3). The
systems optimization within the overall vehicle ends of cylindrical tanks can have either spheri-
design. A principal design objective is for the cal or ellipsoidal shapes. The basic cylindrical
vehicle to yield the highest payload and/or ve- tank with spherical ends is lighter than one with
locity increment with maximum possible relia- ellipsoidal ends. However, the overall weight of
bility. Design details depend largely upon type an ellipsoidally ended tank may be less when
of propellants, vehicle mission requirement and the shorter interstage structure required is con-
configuration, propulsion system design, and sidered. In some designs the propellant tank aft
available construction materials and fabrication ends are faired into conical or other special
techniques. Some of the most important consid- shapes to accommodate the thrust loads from
erations follow. engine assemblies as well as to minimize trapped
propellants.
The required size or volume of a propellant
Propellant Properties
tank is the sum of usable propellant volume and
Propellants affect tank design mainly by their other volume requirements:
physical and chemical characteristics. The boil-
ing point or storage temperature of a propellant Vt:V+ T+B+U (8-1)
determines the operating temperature range of where
the tank assembly. Cryogenic propellants cause Vt = propellant tank design volume, ft3
tank design problems from thermal gradients, V = usable propellant volume calculated from
from the need for insulation and from the need propulsion system requirements, ft$ (may
for construction materials capable of remaining include a "usable residual" term repre-
ductile at very low temperatures. The very low senting design reserves, mixture ratio
density of some propellants, such as liquid hy- shift effects, etc.)
drogen, requires tanks of considerable volume.
T =trapped propellant volume. This is a func-
The highly corrosive and reactive properties of
tion of system design configuration and
other propellants severely limit the selection of
tank materials. includes propellants trapped in tank,
ducts, thrust chamber cooling jacket,
etc., ft 3
Shape and Size of Propellant Tanks
B =boiled-off propellant volume (applicable
Propellant tanks are pressure vessels. Dis- only to cryogenic propellants), ft 3
regarding other factors, the lightest pressure U =tank ullage volume, ft 3
DESIGN OF PROPELLANT TANKS 333
The calculation of propellant volume is based The required usable fuel volume
on the propellant density at specific tempera-
tures. A standard temperature of 68° F is used 10.65 × 2 x 410_ 13S.2 ft3
for storable propellants. Boiling point condi- Vi= 63.17
Vo = 12.78x 2 × 410_ 116.2 ft3 (c) CONCENTRIC PROPELLANT TANKS WITH INTEGRATED
90.12 HELIUM BOTTLE AT AFT END-
TANKAGE WEIGHT, He PERCENT
presents various propellant tank configurations Safety Factors for Propellant Tank Designs
for a typical vehicle system using helium for
The recommended criteria for working loads
tank pressurization. A design analysis will de-
presented in chapter II (eqs. 2-8 through 2-11) are
termine the best solution for a given propellant
generally applicable to propellant tank designs.
storagevolume and vehiclespace envelope.
However, when calculating allowable working
General considerationsare:
stresses from tank internal pressure, the follow-
(I) Minimum overallweight
ing correlations are recommended for various
(2) Maximum storagevolume in a given en- situations:
velope
(1) No hazard to personnel or vital equipment:
(3) Least possibilityof propellantmixing
(4) Clean aerodynamic vehicleshape
Sw = Fy (8-2)
(5) Ease of installation of ducts and lines
or
(6) Ease of insulation
Eli
(7) Ease of fabricationand handling Sw - 1.25 (8-3)
(S) Minimum trapped(unusable)propellants
Arrangement (a)is taken as standard. Ar- (2) Special safety devices are provided for
rangement (b),combining tandem propellant personnel (example: the booster for a
tanks with an integratedhelium bottlein be- manned upper stage which has an ejec-
tween, resultsin lowest weight; however, it tion device with an exceptionally high
poses design problems in the routingof pneu- degree of reliability):
matic lines and controls.Arrangement (c)with
concentrictanks eases the installation of pro- Fy
pellantducts,but has the possibilityof simul- Sw =]-_- (8-4)
taneous punctureof both tanks (by bulletsor or
from other causes),and subsequent mixing of the F.
propellants.Arrangement (d)with multipletanks Sw = _ .35 (S-5)
has the highest weight, but is easier to fabricate
and handle especially for very large vehicles.
(3) ttazard to personnel or vital equipment:
Working Loads
f-'y
The propellant tanks are structural members Sw = 1.33 (8-6)
which must be designed to withstand a combina- or
tion of the following probable working loads: Vtt
Fy (8-8)
Sw - 1.33
Fy _ 50 000_ 37 600 psi
Sw= 1.33 1.33
Hazard to personnel: or
LONGITUDINAL
......
.............
perature conditions (-60 ° to 160 ° F).
Fabrication methods for propellant tanks
depend largely upon the type of material used. WELDS_..: .......................
must be spread out in a suitable way to prevent internal pressure loads and discontinuities. Then
localized overstresses. Cryogenic propellants the design is checked for other loads. If the
may create thermal transient and gradient prob- wall thickness of a pressure vessel is small com-
lems. While the empty portion of a tank may be pared to the radii of wall curvature (t/r< 1/15),
I • • III
DESIGN OF PROPELLANT TANKS 337
and offers no resistance to bending, the wall where a = nominal radius of the tank, in
is subjected only to direct or hoop-membrane (2) Wall thickness, in, required to withstand
stresses, assumed to be uniformly distributed membrane stresses from internal task
over the thickness. However, if any disconti- pressure:
nuity is present along the wall, such as an abrupt
change in radius of curvature or wall thickness, pta
discontinuity and bending stresses are added. ts-2Swe w (8-11)
At a sufficient distance from the juncture be-
tween tank ends (sphere or ellipsoid) and cylin- (3) Wall surface area, in2:
drical shell (where interaction does not occur),
the maximum stress in the tank wall due to in- As =4 =a2 (8-12)
ternal pressure is calculated using the hoop-
membrane stress fornmla only. (4) Weight, lb:
Figure 8-6 identifies the major tank elements.
In an optimum tank design, the wall thickness Ws = 4 ;ra _ tsp (8-13)
varies according to a combination of local mem-
brane, bending, and discontinuity stresses. This (5) Criticalpressuredue to externalloading,
is especially true for the spherical and ellipsoi- psi. When the external pressure is
dal tank ends. The structural calculation meth- higher than the internal tank pressure,
ods for volume, wall thickness, wall surface the pressure differential across the tank
area, and weight of various tank shapes will now wall may cause the tank to buckle.
be discussed. The following general terminology
is used: 2 Ets2 ,
pcr=_V3(1-v 2) (s-14)
Pt = maximum tank working pressure, psig
Sw =maximum allowable wo,'king stress of the
tank construction material, psi Ellipsoidal and Spherical Tank Ends (fig. 8-6)
p =density of tile tank construction material,
Ib/in3 Note that the spherical end is a special case
of ellipsoidal end, in which the major half-
E =modulus of elasticity, psi
diameter, a, equals the minor half-diameter b.
v =Poisson's ratio
(1) Volume:
e w = weld efficiency
Ellipsoidal tank end volume, in3:
where
a = elliptical tank end major half-diameter, in
= radius of the cylindrical tank section
b = elliptical tank end minor half-diameter, in
(2) Wall thicknesses consideringcombined
membrane, discontinuity, and localbend-
ing stressescaused by internaltank
pressurePt. An equivalentwall thick-
Figure 8-6.-Nomenclature or principal tank ness, which is an average value of
elements. knuckle-and-crownthickness,may be
338 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
gpta
tk : Swew (8-17)
ptR
tcr=2 Swe W (8-18)
6 !'
o /" _"_/ i /
(tk+ tcr) pta(K+_) .......I/--/- -4-.<-
te = 2 2 Sw (8-19)
p_a (g + _-)
ts-- 28w (8-20)
(8-21) =
where
Etc 2 _f 1 _3 tc 2
Cb = buckling coefficient, a function of R/te, (8-33)
Pcrc=0"807 Ica _\1_V2/ a2
ranging from 0.05 to 0.10
For long tanks (i.e.,Ic_ 4.9 a V/a/---7_c):
Cylindrical Tank Section
Etc 3
(1) Volume, in3:
Pcrc-4(l_ v2)a a (8-34)
Vc = .a21c (8-29)
where E = modulus of elasticity.
where
a =radius, in
I c = length, in
Sample Calculation (8-3)
(2) Wall thickness, in, required to withstand
membrane stresses due to internal tank The following design data are specified for
pressure: the A-4 stage propulsion system, which employs
a cylindrical propellant tank section with ellip-
Pta soidal ends (preliminary layout shown in fig.
tc = Swew (8-30) 3-10).
Required design volume of the oxidizer tank,
(3) Discontinuity stresses.-The discontinuity Vto = 120 ft 3
at the juncture between the cylindrical Maximum oxidizer tank working pressure,
tank section and the tank ends will Pro = 180 psia
cause bending and shear loads along the Required design volume of the fuel tank,
cylindrical circumference at the junc- Vtf= 143.5 ft 3
ture, and the adjacent areas. These Maximum fuel tank working pressure, ptf= 170
discontinuity stresses are superimposed psia
upon the membrane stresses and com- Internal radius of the cylindrical tank section,
a=41 in
prise: (a._) axial bending stress; (b__)hoop
Tank construction material, aluminum alloy
bending stress; (2) additional hoop
stress due to the shear load; (__)shear 6066-T6: Fy = 50 000 psi; F,, = 57 000 psi;
stress. Discontinuity stresses fade out p=0.1011b/in3; E=10.4×106psi; v=0.36
rapidly,so that they become negligibly Weld efficiency, ew = 100 percent.
small a shortdistance from the juncture. Determine the following:
Detailanalyses ofdiscontinuitystresses _a) Required internal tank dimensions
can be found in standardtextbooks. In (b_._)Required thickness of the tank walls at
general,buildupof wall thickness of various sections, considering internal
less than 0.5 tc, near the juncture, will pressure loads, discontinuity, and local
suffice for most designs, with only small bending stresses
weight penalty. (,_C_)Approximate weight of the tankage
(4) Wall surface area, in 2" (d__)Critical external loading pressures, using
a buckling coefficient, Cb = 0.10 for the
Ac = 2 _alc (8-31) tank ends
340 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
2 × 2 _ra2b
Vt° - 3 From equation (8-18), the required wall thick-
ness at the crown of the oxidizer tank end
Vtf= ;ra2 Ic
tkoPtf 0.171x170
tkf .... 0.162 in
Pro 1S0
Ic Vtf 143.5x1728
= a 2- rrx(41)2 =46.9in
The required wall thickness at the crown of
the fuel tank end:
To summarize the internal dimensions of the
tankage:
tcroPtf 0.149× 170
tctf: - - 0.141 in
a=41in, b=29.4in, k=1.395, Ic=46.9in Pto 180
(b_.) We assume that certain missions of the The equivalent wall thickness of the fuel tank
A-4 vehicle require it to be man rated. From end:
Fy _ 50 000 _ 37 600 psi From equation (8-30), the required wall thick-
Sw = 1.33 1.33
ness of the cylindrical tank section:
tcj = tc + 0.4 t = 0.202 + 0.4 × 0.202 -- 0.283 in Cb2 Etef 2 _ 0.10x 2 × 10.4 × 106x(0.152)-*
Pcre[ = R2 (1.395x 41) 2
To summarize:
= 14.7 psi
_a2teoE'p
We°- 21: x {(1 - (_-.36_2)
1 3 \_--/
(0.202_ "_
=13.9 psi
From figure 8-7, E' is 4.56, for k=1.395:
Weo- rr(41) 2 x 0.16 x 4.56 x 0.101 _ 139.5 lb Axial Compressive Loading on the Cylindrical
2 x 1.395
Tank Section
The weight of the fuel tank end: In integrated propellant tank designs (figs,
8-2 and 8-3), the cylindrical tank section must
Approximate overall weight of the tankage One method of increasing the axial load-
(less accessories): carrying ability of a cylindrical tank section with
minimum weight penalty is to pressurize the tank.
Wt = 1.04 (2 x Weo + Wef + We) This is known as pressure stabilization. Inter-
hal pressure will raise the critical buckling
= 1.04 (2 x 139.5 + 132.8 + 246.4) = 685 lb
stress of a tank: or it may be used to counter-
balance an axial compressive load Fa, lb, where
(d__) From equation (8-27), the critical external
loading pressure for the oxidizer ends
F a = za=pt (8-36)
An alternative method of increasing the exter- the tank. This effect produces a su:ge of the
nal load-carrying ability of a cylindrical tank is tank internal pressure. For very short impacting
to make it self-supporting. This revolves stiff- times (less than 1.2 x 10 .3 see), the following
ening the cylindrical skin by means of longitudi- correlations are established for cylindrically
nal and circumferential members, or honeycomb shaped propellant tanks:
structures. The members may be either separate cw
(8-37)
stiffeners welded to the tank wall, or may be Ps = ,Ta2g
made integral with the wall by machining or
chemically milling a thicker sheet. C I
(s-38)
sc: [+(-°2°2Y
°+01+
• (0202
/ j,+7
×10.4×106
tion material,
v = Poisson's
psi
ratio of the tank construction
= 20 360 psi material
In many prepackaged liquid applications, the
The critical axial compressive load of the propellant tankage is required to withstand cer-
cylindrical tank section: tain impact loads, as specified by the height of
the drop tests. The details of estimating tank
Fc=S c x2_ratc=20360×2,_× 41 ×0.202 pressure surges from a free-fall impact are illus-
trated by sample calculation (8-5).
= 1 060 000 lb
From the results it is obvious that the A-4 Sample Calculation (8-5)
stage tankage is capable of withstanding a sub- The following data are given for the cylindri-
stantial axial compressive load without internal cal fuel tank of a prepackaged storable liquid
pressurization. propulsion system:
(b__)From equation (8-36), the required internal Fuel, N2H 4
tank pressure:
Fuel density, pp=63.17 lb/ft 3
Compressive modulus of elasticity of the fuel,
Fa 100 000 Ep= 6.06x l0 s
Pt= _-_a = _x (41) 2"= 18.95 psi Free acoustic velocity of the fuel, c' = 80 100
in/sec
Tank construction material, aluminum alloy
Water Hammer Effects Due to Impact 6066-T6
When a loaded propellant tank is subject to an Modulus of elasticity of the tank construction
impact force, a water-hammer effect occurs within material, E= 10.4x 106
DESIGN OF PROPELLANT TANKS 343
Poisson's ratio of the tank construction mate- Tile equivalent propellant flow rate due to
rial, v=0.36 impact:
Radius of the cylindrical tank, a = 4 in
Length of the cylindrical tank, lc = 50 in
Wall thickness of the cylindrical tank, tc _¢=pp,va2V=63.25x;rx(t-_)2x35.9 = 792.5 lb/see
=0.167 in
For a tank falling in direction of its longitu- The tank pressure surge due to impact after a
dinal axis, estimate: 20-foot free drop:
(_a) Tank pressure surge due to the impact
after a 6-foot free drop cw 43100×792.5
psi
(b__)Tank pressure surge due to the impact Ps = _a_- _,x (4) 2 x386
after a 20-foot free drop
(a.) For a 6-foot free drop the final velocity at The design of tanks for storable liquid pro-
impact: pellants uses the s'tme general design practices
applied to other propellant tanks, except in the
V = _2gfi = _¢'2× 32.2 × 6 = 19.65 fps area of compatibility. Most storable propellants
will remain stable for long periods of time if
stored in tanks constructed of materials compati-
This yields an equivalentpropellantflow rate
ble with the propellants. To minimize propellant
due to impact of
decomposition and tank material corrosion, the
surface of tank walls in contact with the propel-
9¢= ppTra2V = 63.17x ,7× × 19.65 = 434.51b/see lants must be smooth and clean.
Tank Construction
/
Full-penetration welds should be used in the Figure 8-IO.-Design o[ a typical storable pro-
construction of all propellant tanks. This is pellant tank with a [orged one-piece common
especially important for storable propellants. bulkhead.
DESIGN OF PROPELLANT TANKS 345
and tool marks. All welded joints should be acute with the hydrogen tank insulation. It often
ground to achieve a smooth dent-free contour. If becomes one of the most critical design factors in
an aluminum alloy is used for tank construction, a hydrogen-fueled vehicle system. The difficul-
anodizing of the inner surfaces is recommended. ties arising in hydrogen systems in connection
The tank should be cleaned and treated before with heat transfer may be dramatically illustrated
use. The process rendering the surfaces inac- as follows:
tive is known as passivation. Basically, passi- Assume two tanks of equal size, subject to
vation involves the cleaning of tank interior with the same heat influx per unit time. One is filled
solvents and, in some cases, the treatment with with liquid oxygen, the other with liquid hydro-
alkaline or acid solution. Passivation differs gen. The ratio of heat of vaporization per unit
from conditioning or stability testing of a tank. weight 0/H=0.48, but the density ratio is
Stability testing of a tank involves the applica- O/H= 14.3. Thus the volume rate of vaporization
tion of a propellant rinse prior to propellant in the hydrogen tank is 6.85 times faster than in
tanking. oxygen. In reality, the heat influxes would not
be equal for two uninsulated tanks, because of
the higher temperature differential across the
8.5 DESIGN OF CRYOGENIC LIQUID wall, and especially because of the greatly in-
PROPELLANT TANKS AND THEIR
creased heat-transfer rate from air liquefaction
INSULATION
on the hydrogen tank surface. This may further
accelerate the hydrogen volume boiloff rate to
In the design of cryogenic propellant tanks,
approximately 70 times that of oxygen. In an
there are several potential problem areas which
may affect proper functioning and reliability: actual oxygen/hydrogen system, operating at a
(1) Properties of the tank construction mate- weight mixture ratio O/H=5, the tanked mixture
rials at the cryogenic propellant service ratio by volume is H/O--3.23. Depending on the
shape of the tanks (surface ratio), this may again
temperature range
(2) Thermal stresses induced in the tank double or triple the relative boiloff rate of hydro-
gen. The absolute necessity for insulation to
structure by temperature gradients
(3) The relief of tank pressure caused by drastically reduce heat influx into a hydrogen
(4) Thermal insulation of the tank walls Boiloff rates are not the only problem caused
The knowledge of the precise strength char- by the physical properties of hydrogen. Near the
acteristics, degree of brittleness, and notch sen- ambient boiling temperature, the gradient of
sitivity of the tank construction materials at vapor pressure is 2.4 psi/°R, as compared to
cryogenic temperatures (as low as -423 ° F for 0.78 psi/°R for liquid oxygen. Moreover, this
liquid hydrogen service) is a prerequisite for gradient increases rapidly with increasing tem-
their selection. In general, most of the aluminum perature, which would be experienced following
alloys, the austenitic and semiaustenitic stain- tank pressurization. Figures 8-11 and 8-12 show
less steels, possess good mechanical properties the trend for both liquid oxygen and hydrogen.
at cryogenic temperatures (also see ch. II). The The data have great significance to pump NPSH.
thermal stresses can be analyzed by determining As may be seen, just one degree of liquid-
the temperature profile at various regions of the hydrogen-temperature rise requires a 3-psi in-
tank and may be minimized by discrete design crease in tank pressure to maintain proper NPSH.
approaches. The capacity of the tank relief In a large vehicle, the required increase in tank-
valve should be based on the maximum antici- wall thickness may affect payload capability
pated propellant boiloff rate during ground hold noticeably. The situation is further aggravated
and actual operation of the vehicle systems. by the high heat influx into hydrogen, for reasons
Among the cryogenic propellants, liquid hy- mentioned in connection with boiloff. Even for
drogen imposes the most serious tank design relatively short boost periods, rapid warmup may
problems. This is mainly due to its very low create a problem more severe than boiloff, and
service temperature and its relatively large spe- places further emphasis on adequate insulation.
cific volume. Design problems are especially The following discussion of tank insulations,
346 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
io tures
For
are
the latter,
finding
honeycomb-supp0rted
wide application. Figure
struc-
8-13
,='l ,s _c =z i
shows the typical example of an externally ap-
SATU_ATICI_I TEle_RATUPE AND PRESSURE OF LIOUID OXYGEN
pressure to temperature [or liquid oxygen. installed between an inner and outer facing
DESIGN OF PROPELLANT TANKS 347
v,
ON[_C_U_ CC_[ wao_ O_ Solution
@_S _ • Z 2 Le _FT 1 -- ATMOSPH£RE
LIQUIO HYDROGg N
for which lateral passage ways must be provided. tank insulation design. The insulation may be
The purge also serves as a leak-detection de- located internal or external to the tank wall; it
vice, in conjunction with gas analyzers, to de- can be integral, or disposable during boost; i.e.,
tect contamination of the helium from leaks. In it can be bonded in place or mechanically re-
figure 8-13, a separate gap is purged with helium, tained. Basically, any insulation applied to a
rather than the honeycomb ceils. tank must be justified with respect to advantages
The thermal conductivity of the insulation of performance and/or economy.
shown is about 3.8× 10 -_ Btu-in/in_-sec-°F (0.2 Locating the insulation inside of the propel-
Btu-in/ft2-hr -° F). lant tank has the obvious adwmtage of protecting
Thermal conductivities of various types of the insulation from handling damage. The tank
insulation vary from 0.05× 10 -? to 6.0× 10 -7 structure is isolated from the severest low-
Btu-in/in_-sec-°F. Their densities range from temperature effects of the propellant and is thus
2.0 to 20 lb/ft a. Since the quality of an insula- subjected to only moderate thermal cycling from
tion will affect cost and weight, an optimization its source. Internal insulation also minimizes
study will have to be made, based on mission propellant loss when chilling the tank during
characteristics. filling. However, if a crack or leak should occur
in internal insulation of a hydrogen tank, gaseous
Sample Calculation (8-6)
hydrogen would enter the crack and gradually in-
Determine the heat-transfer rates in Btu/in _- crease the heat transfer. Other undesirable fea-
see across the tank insulation shown in figure tures of internal insulation are apparent, such
8-13 during- as: difficulties in installation; in locating and
(a__) Ground hold.-The temperature of the in- repairing of leaks; and in cleaning the tank.
sulation surface near the tank wall is Internal insulation is also subject to higher
around -360 ° F and the outer surface is pressures and more severe temperature effects
70 ° F. which tend to impair the insulation sealing.
(b_.)
Boost phase.-The inner insulation sur- External insulation has the advantage of iso-
face temperature reaches -210 ° F and lating the tank structure from the extreme tem-
the outer insulation temperature reaches perature of aerodynamic heating during boost.
800 ° F. Installation, repair, and sealing of the external
348 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
\\"1
--_ /' :z_ 4 .'
/ )"4'-,. ./
8.7 DESIGN OF PROPELLANT TANK ", ' , d_-< _ l],
PRESSURANT DIFFUSERS
_SSVRAmr GAS
INLEt
Positive Expulsion _O_C[_ R_¢ _RCU_t_ ¢VmvoLwr_
Metal expulsion diaphragms made of 1100-0 Figure 8-18.-Corrugated metal positive expulsion
aluminum (0.010 to 0.020 inch thick) are desir- diaphragm used in a cylindrically shaped pro-
able for long-term storage contact with storable pellant tank.
DESIGN OF PROPELLANT TANKS 351
tank volume is less efficient due to dimensional disadvantage of pure elastomers in storage con-
limitations (i.e., size of the center post com- tact with many storable propellants is tensile-
pared to tank diameter). This design, too, mini- strength degradation as a function of time. They
mizes shift of center of gravity. are also incompatible with hot pressurant gases
One disadvantage of the metal diaphragms is produced by gas generators.
that only one complete expulsion cycle is possi-
ble, because the metal probably will be cold
Positive Expulsion By Movable Pistons
worked. The cold working not only changes
metal elastic properties and the Ap required for An alternate method of obtaining positive ex-
expulsion, but may cause fatigue failure. pulsion in cylindrical tanks is a movable piston
actuated by pressurant gas. Positive attachment
is required during storage to keep the piston in
Elastomer Expulsion Diaphragms
the proper position. To prevent leakage during
Elastomer-type diaphragms are applicable to operation, seals will be required. The seals
most tank configurations and often offer a more may be piston-type rings or some type of metal-
efficient utilization of tank volume. One design lic wiper seal. In either case, the dimension and
approach takes advantage of the stretching prop- surface finish of the tank inside diameter should
erties of pure elastomer and uses the diaphragm be maintained relatively accurate and smooth. In
as a bladder. Figure 8-19 presents the bladder- some designs a concentric center post is utilized
type elastomer positive expulsion diaphragm to guide the piston, requiring an additional seal.
used in a spherical tank. As the bladder is Figure 8-20 presents the design of a movable
pressurized and inflates, the propellant is dis- piston guided by a center post.
placed and positively expelled and uniformly The pressure differential across the movable
confined during all phases of expulsion. By piston required to overcome friction during oper-
positioning the bladder in the geometric center ation increases the required pressurant pressure
of the tank, center-of-gravity shift during expul- and the tank structural pressure loads for a given
sion is virtually eliminated. propellant pressure at the tank outlet. It may be
The initial unpressurized volume within the estimated by
bladder can be utilized as the ullage volume.
The elastomer diaphragm will produce an expul- ApAp = [sFnLs (8-39)
sion efficiency of up to 99 percent in many tank
configurations. In addition, it is capable of where
many expulsion and refill cycles. An inherent Ap=pressure differential across the piston, psi
TAN_ _D
pR_SS_
I_L_r
I, _ p_pe_LA_r
_r_eT
Ap-- cross-sectional area of the piston, in: Estimate the pressure differential across the
is = coefficient of friction of the piston seals movable piston.
Fn =unit normal force reacted by the piston
seals on tank wall and guide post, lb/in
Solution
of seal
Ls = total length of all piston seals, in The cross-sectional area of the piston
9.1 THE PRINCIPAL INTERCONNECTING increases the duct envelope and may cause in-
COMPONENTS AND MOUNTS terferencewith other vehicle systems. Figure
9-2 shows a typical flexible propellant supply
In section1.4,we brieflyintroducedthe
duct which has restraining linkages for stabiliz-
principalinterconnectingcomponents and struc-
ing the bellows. At the vehicle end, these ducts
tures. Figure 9-I shows these fora typical
will connect to longer or shorter vehicle ducts,
LH2/LO 2 turbopump-fedengine system. A dis-
the length of which depends on whether the for-
cussion of the most frequentlyused components
ward or the rear tank is being connected. It is
follows. Design detailis presented in subse-
important that the engine designer not only m-
quent subehapters.
form the vehicle builder of connecting flange
dimensions and types of gasket being used but
also of the forces transmitted by the engine duct
Propellant Supply Ducts
to the vehicle during gimbaling as well. In most
Among the principal interconnections between pump-feed systems, the working pressure of pro-
engine and vehicle are the propellant supply pellant supply ducts usually does not exceed 50
ducts. Since every psi of propellant tank pres- psig. In upper stages, during lower stage boost,
sure above the minimum required for proper en- however, pressures may temporarily be substan-
gine performance results in additional weight of tially higher as a result of a combination of high
the tank walls and of the gas pressurants, it is accelerations and full tanks (100 psig and over).
desirable to keep the pressure losses between Figure 9-3 shows the propellant ducting and
tank outlets and engine inlets to an absolute mounting for a cluster of four storable propellant
minimum. pressure-feed engines. The thrust mounts are of
The task of designing these ducts, or at least a box-type, beam construction. The propelhmt
the flexible portion thereof immediately upstream duets, which consist of restrained bellows and
of the engine, frequently falls to the engine de- rigid sections including flowmeters, connect the
signer. The designer must find an optimum bal- thrust chambers to the main propellant valves.
ance between low pressure drop, by making the The main valves, in turn, connect directly to the
duct diameter as large as possible, and flexibil- propellant tank ducts. Care has been taken in
ity and structural integrity, which in general the design to keep the flow path and ducting
becomes more difficult with increasing diam- volume constant between main valves and indi-
eters. The designer must further consider the vidual chambers. This arrangement also assures
fact that the ducts, because of their location off uniform pressure drops to all thrust chambers.
the engine gimbal center, are subjected to tor- Furthermore, all interconnecting components on
sional loads, in addition to bending. Furthermore, the fuel side as well as on the oxidizer side are
the ducts are subjected to internal pressure, fre- designed to be interchangeable. The valves
quently in a stringent cryogenic and vibration used in this design include burst diaphragms.
environment. Because of the many forces acting Thus, exposure to the propellants of all ducting
upon the ducts, restrainers against buckling are downstream of the valves will occur only during
frequently required. These may be located in- engine firing. In most pressure-feed systems,
side the ducts, thus adding to undesired pressure the working pressure of the propellant ducts is
drop; or they may be applied externally, which less than 500 psia.
353
354 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
TURBOPUMP
ASSEMBLY
G.G. OXIDIZER
LI NE
NE GIMBAL
ACTUATING LUGS
OXIDIZER
TURBOPUMP ASSEMBLY
FUEL
TURBOPUMP
OXIDIZER MOUNT
TURBOPUMP MOUNT
PUMP DISCHARGE
HIGH PRESSURE
MAIN OXIDIZER
DUCT
TURBINE
EXHAUST
HOT GAS DUCT
PUMP DISCHARGE
TURBINE
HIGH PRESSURE
DRIVE HIGH MAIN FUEL DUCT
PRESSURE HOT
GAS DUCT
CHAMBER
ASSEMBLY
OXIDIZER
TURBINE
EXHAUST HOT
GAS DUCT
Figure 9-1 .-Various interconnecting components and mounts in a typical LH2/LO 2 pump {eed engine
system.
Pump-Discharge, High-Pressure Propellant Ducts rating pressure loads acting on the two compo-
nents connected by a flexible member must be
In turbopump-fed engines the pump-discharge absorbed by restraining links attached to the
high-pressure propellant ducts are designed to bellows, or by other compensating means.
connect the oxidizer and fuel pump discharges to Figure 9-4 presents a typical pump-discharge,
the main oxidizer and main fuel valves attached high-pressure propellant duct with external re-
to the thrust chamber. The ducts contain bel- straining links. It is used for the main oxidizer
]ows sections which permit the degree of move- (LO2) duct of the engine system shown in figure
ment required between thrust chamber and turbo- 9-1. Another typical pump-discharge, high-
pump to accommodate tolerance buildups, pressure propellant duct, used as the main fuel
misalinements, and motion due to temperature (LH2) duct for the same engine system, is shown
change and acceleration loads. However, in in figure 9-5. This duct has a unique end-load
some engine designs rigid, in place welded ducts compensator which incorporates two bellows tied
have been successfully applied. The working together by a restraining rod to limit bellows
levels have been proposed. Therefore, the sepa- movements• One bellows opposes the other in
DESIGN OF INTERCONNECTING COMPONENTS AND MOUNTS 355
BRU- J_TS J
CON_IECTbNG TO
PUMP DISC_ARC-[--_r
E_O LOAO
COMPENSAT_ --
BEU.
, 1 J
O_JTCET FL._,_IG_E
CONN£C_r EO TQ
Propellant-Tank-Pressurization Lines
Figure 9-3.-Propellant ducting and mounting These lines are employed to connect the main
arrangements for a cluster o[ [our storable- propellant tanks to the pressurant sources such
propellant pressure-Iced engines. as stored pressurant gas systems (including
356 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
storage bottles, heat exchangers, pressure regu- Cryogenic Propellant Bleed Lines
lators), gas generators, cryogenic propellant heat
Cryogenic prope!!ant engine systems with
exchangers, etc. High-pressure hoses and tub-
ings are used. turbopump feed may experience difficultiesdur-
ing start if the metal parts containing cryogenic
fluid are at ambient or insufficiently low temper-
Seal Drain Lines
atures, and if the fluids in the volumes below
It is difficult to achieve perfect dynamic seal- the tank outlet are superheated and form gas
ing at the shafts or rods of turbopumps and other pockets. Since the pressure upon opening of the
components. Therefore, seal drains are often main valves and start of the turbopumps will
provided between two dynamic seals placed in further reduce the static pressure at the pump
series. The required seal drain lines including inlet, accelerated gas production will result.
flexible hoses and tubings are routed away and This, in turn, may lead to pump cavitation and
overboard. For LO2/RP-I systems, this can be starving of the gas generator. To prevent this, a
done by routing the drain lines along the thrust continuous bleed from a point farthest down-
chamber wall to the chamber exit. For propellant stream from the pump inlet is applied until
shortly before engine start. In this manner, fresh
combinations which can form highly explosive
mixtures, routing to sufficiently spaced vent liquid at tank bulk temperature will continuously
replace the warming fluid, and cool the contain-
ports at the vehicle periphery is required. Fig-
ing metal parts. To avoid hazardous conditions
ure 9-6 presents a typical pump seal drain line
schematic of an upper stage system. The seal at the launch site, the bleeds, particularly if
drain lines are routed to the vehicle periphery they can form combustible or explosive mixtures,
during boost flight and to chamber exits during are ducted away. To make this possible, a line
Liquid rocket engines usually are equipped engines in upper stages which will not start
until some time after the bleeds have been closed
with one or more pressure vessels to supply
pneumatic pressure for valve actuation, for tur- at liftoff,or which have to start after prolonged
bine start, for sequenced purges, and possibly cruising periods. Here, a recirculation system
which returns the fluids to the tanks rather than
for other purposes. The vessels must be charged
dumping them overboard is preferred. Lines
prior to test run or flight,requiring high-pressure
flexible line connections to the vehicle, and across the interface between engine and vehicle
are required. Bleed and recirculation flows can
disconnects at the vehicle periphery. Their
design not only must consider the mating coun- be minimized if the vehicle builder provides
terpart on the vehicle side but also the type of means, such as subcooling, insulation, avoid-
ance of temperature stratification, maximum eco-
fluid and its temperature and pressure.
nomic tank pressure to keep the temperature of
the bulk sufficiently below the boiling tempera-
\ ture at operating tank pressures. In some cases,
Z
.... : J_ prechilling of metal parts exposed to cryogenic
\,
fluids may be advantageously accomplished by
\ cooling media supplied from ground through a
vehicle disconnect to the engine. Routing of all
lines mentioned will be similar to drain lines
,,(Lr _'_':{ -_ x (fig. 9-6).
_
_
r
Purge Lines
9.2 DESIGN OF TUBINGS AND TUBE decreases tubing outside diameter less than 6
FITTINGS percent. Recommended minimum bend radii for
stainless-steel and aluminum-alloy tubing are
To interconnect fluid lines in rocket en-
presented in table 9-3.
gines, high-quality, cold-finished seamless
Tubing assemblies must be properly supported
tubing of either aluminum alloys or austenitic
to prevent stresses and consequent weakening of
stainless steels are frequently used. In rocket
the system under vibrating conditions. In addi-
engine design practice, the term "tube" refers to
tion, proper support minimizes the danger of
lines up to 2 inches. Stated nominal tubing
recoil and live whip in the event of tube failure.
sizes refer to their outside diameters. For sizes
Where tube fittings are employed, support spac-
of more than 3/4 inch, flanged joints should be
ing should be reduced by 20 percent to account
used.
for the added weight. Supports should be placed
Two principal classes of tubing are distin-
as close to each side of fittings, valves, and
guished: common pressure tubing and mechan-
other components as practical. Overhang should
ical tubing. The latter is designed and manufac-
be minimized by placing supports as close to
tured with closer tolerances then the common
bends as conditions allow. Table 9-4 presents
tubing. It is more expensive, but may save much
the recommended maximum bracket spacing for
time when machining and chucking in uutomatic
common-size tubing assemblies in engine sys-
production machines is involved. Unless speci-
tems. Where tubes of different diameters are
fied otherwise, pressure tubing is used in rocket
connected, average spacing may be used.
engines for all purposes.
General design practicerequiresno detail
drawings fortubingassemblies less than 3/8-
Tubing Design Working Pressures inch size. They are to be shop fittedaccording
to an engine mockup duringthe assembly of the
Pressure tubes are supplied fully annealed
engine system.
and pressure tested. They fall into two major
classifications: average wall and minimum wall.
Average-wall pressure tubes are made with a Tube-Fitting Designs
wall thickness which may vary from the nominal
size by plus or minus 10 percent. Minimum-wall The three-piece AN flared tube fitting per
tubes are made with a wall thickness that may AND10056 or MS33656 (fig. 9-7) is widely used
not be less than that specified, but may be heav- as a standard in rocket engine designs.
ier by 20 to 40 percent, depending on the type of The angle of flare of the fitting was estab-
tubing. For lines carrying pressurized fluids, lished at 37 ° to give maximum flare contact and
minimum-wall tubes should be used. provide a nose sufficiently strong to resist crush-
A factor of safety (i.e., the ratio of the ulti- ing. The AN three-piece fitting consists of a
mate strength of the tubing material to the maxi- coupling nut (AN818), a sleeve (ANS19), and a
mum allowable working stress) of 4 may be used connector. The sealing occurs between the nose
for general rocket engine applications. Higher of the fitting connector and the inside of the
tube flare. The AN819 sleeve has an external
values (6 to 8) should be used in applications
involving hazard and excessive vibration. For shoulder against which the AN818 nut acts.
high-pressure application, aluminum tubing This produces a locking effect on the nut, as the
should be avoided where possible. Allowable sleeve is slightly sprung upon proper torque.
working pressures for a factor of safety of 4, for The sleeve is free to turn during initial assembly
various stainless-steel and aluminum tubings, to allow for slight eccentricity of the tubing or
are presented in tables 9-1 and 9-2. the nose of the fitting connector.
Figure 9-8 illustrates a typical flareless tube
fitting per MS33514, the Ermeto, manufactured by
Tubing Installations in Engine Systems
the Weatherhead Co. It is designed for use with
Installations requiring bends must be accom- flareless, heavy-walled tubing in high-pressure
plished with minimum distortion and constriction applications. The assembly of this fitting in-
of the tubing. A satisfactory bend is one which volves a preassembly to check for contact of the
DESIGN OF INTERCONNECTING COMPONENTS AND MOUNTS 359
TABLE 9-2.-Aluminum Alloy, 5052 Round Seamless Drawn WW-T-7Sa Temper H34
[Allowable working pressures in psi at 100 ° F; safety factor of 4]
[ 0.042
3/8
I
1/2
0.049 i
-[--1
518
0.05S [
3/4
0.072
1
J 0.095
1-1/4
I 0.120 I
]
Maximum working pressure, 750 psi
_beOO,.
.............................. I lJ41 3_81lJ2I 5_8 3'4 _ 111_11'21_
w_ll<.,ek.ess,
in ........................
I _ 1_0 020 0 02S
I 0 032
°k22_L___
0 035 0.0-t2 0.049
I 0 065 10.072 l °°95
SPECIAL
sE
OF FLARE
MUST PASS NUT
FLARE PER INNER EDGE
AND t0061 C OF SLEEVE
MS 33584 BUT NOT
EXCEEO
OUTER EDGE
*'AN THREAD
SEALING SURFACE
STRAIGHT THREAD
MALE CONNECTOR MALE _
CONNECTOR
Figure 9-7.-3-piece AN flared-tube fitting per Figure 9-S.-Flareless tube fitting (Ermeto) per
AND 10056 or MS3365G. MS33514.
36O DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
Figure 9-I O.-Typical flange joints of a propellant ferentially along the serrations, and will seal
fluid.-These greatly affect the type of seal de- 5. Gasket toading.-The minimum requirement
for good sealing of a flange joint is sufficient
sign and material. The problems of maintaining
a leaktight flange joint are greatly compounded gasket precompression to close tip all paths
through which fluid flow could occur. Further-
by temperature effects. Temperature differentials
more, the sealing load must be maintained so
at the flange joints of rocket engines are apt to
that a specific level of gasket compressive
be large, because of steep heating and flow
stress is induced to resist the internal fluid
transients. They introduce thermal stresses and
pressure.
strains which may disturb the sealing. When
designing an elevated temperature (or subzero) 6. Sea/drain.-In sonie applications, positive
flange joint, tim temperatures of the various sealing at the flange joint is required. Dual
(series) seals and overboard drain line are then
parts should be analyzed to assess their effect
on the sealing load. Ideally, thermal effects provided (fig. 9-15).
, ;k_ _
longitudinal tension of the duct, Ib/in
F 3 = force per unit length of the flange ring from
/ _ !' internal pressure p, lb/in
F 4 = force per unit length of the flange rag from
....... )2 A_Kg'r
!
F s = force
gasket
per
loading,
unit length
lb/in
of the flange rag from
compression load at the flange outside
diameter, lb/in
Figure 9-11.-Structural design configuration o[ a
Sg = required average gasket compressive
typical flange joint.
stress, for proper seating against an in-
ternal fluid pressure p, psi
so that a gasket compressive stress is main- m =gasket factor, a function of gasket design
tained to seal effectively against fluid leakage, and to be determined experimentally;
under maximum working pressure and other loads. design values range from 0.8 to 10
Defining a flange ring segment of unit length n = flange factor, a function of flange config-
(i.e., 1 inch along the arc of a circle passing uration and its rigidity; design values
through the flange cross-section centroid) as a range from 0.1 to 0.8
free body, the correlations between all forces We = end loads on the duct due to inertia and
acting on that segment, and the minimum required thermal effects (tension or compression),
design flange bolt loading, are established as lb
follows: Wb = minimum required design flange bolt load-
ing, lb
F1=F2+F3+F4+Fs (94)
Sample Calculation (9-1)
From equation (9-5), the required average A = design factor, ranging from 4 to 6
Sg = mp= 0.8 x 1750 : 1400 psi hub portion of the flange ring.
The stress and strain analysis of the flange
Combine equations (9-1) and (9-6): ring may be treated as the twisting of a thick
circular ring of uniform cross section under the
(F2 + F 3 + F 4) influence of turning couples which are uniformly
Combine this and equations (9-2), (9-3), (9-4), correlations approximate the maximum working
stress and strain of a flange ring (refer to fig.
and (9-7) to obtain the minimum required design
bolt leading of the flange joint: 9-11):
[prtDi 2 + 4 We + p,'r(D22 - D12) + Sgzr(D32 - D22)] Mtz FlYi + F2Y2 + F3Y3 + F4Y4- FsYs (9-11)
Wb: 4(l-n)
(D 4 - D)
[1750 × zr x 82 + 4 x 2400 + 1400 × r;(S.5 _ - 8:)] (9-12)
Y_: 2
Wb : 4 (1 - 0.3)
= 141 000 lb (D - D _ - t)
(9-13)
Y2-" 2
After the required design bolt loading has
been determined, the number, size, and torque
(2 D - D 2 - D1)
value of the bolts needed to produce that load (9-14)
Y3- 4
must be chosen. Since tensile elasticity is ad-
vantageous, the use of small-diameter, high-
(2 D- D 3 - D2)
strength bolts is desirable. Bolt spacing should (9-15)
Y4-
be sufficiently close to insure a reasonably even
distribution of the load around the gasket cir-
cumference. The following empirical correlation (D5 - D)
(9-16)
is recommended for maximum spacing to produce Ys:- 2
a tight joint:
MtD 2
Ps : 2 d + T (9-8) 0 : -- (9-17)
4 Ely_y
where
MtD2(L1 - L)
Ps = maximum bolt spacing, in
S z- 2Dlly_y (9-18)
d =nominal bolt dia, in
T = flange thickness, in
where
The general proportion of the flange ring may
be determined by the following empirical equa- D4 : diameter of the bolt circle,
in
tion (fig. 9-11):
D s =outside diameter of the
T=At (9-9) flange ring, in
Y_, Y2, Y3, Y4, Y5 :distances between the ring
cross-section centroid and
L_ = B_ (9-10)
forces F,, F 2, F 3, F 4 and
where F s , in
t =thickness of the duct wall, in, as deter- Mt = resultant twisting couple
mined by a hoop stress calculation per unit length of flange
F4 = Sg(D32 - D22) _ 1400 (8.52 - 82) _ 336 lb/in From equations (2-8) and (2-9), the yield
4D 4x8.6
load stress = 121 200 x 1.1 x 1.1 = 146 600 psi.
This is smaller than the minimum yield strength
From equation (9-6):
Fy = 170 000 psi of the material.
From equations (2-8) and (2-10), tile ultimate
F s = aF 1 = 0.3 x 5200= 1560 lb/in
load stress= 121 200x 1.1 x 1.5=200 000 psi.
Thus, the proposed flange configuration is
From equation (9-12):
satisfactory. We will now determine the number
of bolts and, from equation (9-8), the maximum
(D 4 - D) _ (9 - 8.6) _ 0.2 in bolt spacing
Yl=- 2 2
(2 D- D 3 - D2) _ (2 × 8.6 -8.5 - 8) _ 0.175 in sample calculation (9-1) was based on the maxi-
Y4- 4 4 mum transient pressure. Using equations (2-8)
and (2-10), the required ultimate bolt loading:
From equation (9-16):
W b x 1.1 × 1.5= 140 400× 1.65=231800 lb
(D s - D) _ (9.62 - 8.6)_ 0 51 in
Ys- 2 2 " Therefore, the ultimate loading on each bolt
=231 800,/26=8900 lb. This is smaller than the
From equation (9-11): allowable ultimate bolt load of 10262 lb.
The required preload on each bolt:
Mr= Flyl + F2y 2+F_y 3 + F4y 4-Fsy s
Wb_ 140 400
= 5200 x 0.2 + 3347 x 0.26 + 336 x 0.175-1560 × 0.51 - 5400 Ib
26 26
= 1174 in-lb/in
900 ° F.
Elastomer and Metal O-Rings for Flange Joints BRIAR COR_ERS 005"
_--i
D3 --
D2
_ G_OOVE SIDE FINISH f2
aSee fig. 9-12 for flange design. Duct diameter D, may be equal to D,..
be used. These gaskets are made to various centric. Typical design data for spiral-wound
stiffnesses (spring rates) by varying the tension gasket flange joints are presented in table 9-7.
of the wrapping. The harder gaskets are recom- Gaskets of this type require high compressive
mended for the higher pressures. loads. The values range from 6000 to 25 000psi.
The gaskets are used in groove-type flanges, The amount of gasket compression is concrolled
as shown in figure 9-12. For optimum results, by metal-to-metal contact of the flanges, with
tool marks on sealing surfaces should be con- allowance for maximum tolerance buildup of
Approximate re-
Gasket dimensions Flange groove dimensions
quired compressive
load, lb at deflection t
A diameter B diameter Sealing area, I D2 I D: r [,
_'0.016 t 0.032 sq. in. 0.025 0.035
tO.OlO t ±0010
i ]
aSee fig. 9-13 for seal design. See fig. 9-12 for flange design. Duct dlameter D, may be equal to D2.
368 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
A_*BESTO$
Fs, = 8 (H - 0.02)_C, (9-19)
FILLEII
p(H - 0.02)
ENL _t_GED •
Fs2 = Fsl + 2 (9-20)
where
II/ T----T/V- II '_" "="
Fs_ = seal contact load per inch of seal circum-
ference without internal fluid pressure,
lb/in
Fs2 = seal contact load per inch of seal circum-
gasket thickness and groove depth. One diffi- (i.e., the difference between initial seal
culty with spiral-wound gaskets is the presence width M and flange groove depth)
seal
An ingenious
deflection
approach
capabilities
toward
to accommodate
attaining better
de-
_
,,
K : 7 ' -oo3
,)i oo._ooo.
'' -:oz
section. When such a seal is compressed be- _-_)FLAT I¢'I%"rllN _3(]_ INCH P£R INCH CIKCUM} £R_t4C£
tween the surfaces of a flange joint, the high @ PARALLEL TO SURFACE X IHTHI_ _J:_ INCH PER INCH CIRCU=FER[;N_CE.
TABLE 9-8.-TypicaI Design Data o[ NaHex Pressure-Actuated Seals and Flange Joinzs
[All dimensions in inches]
A diameter B diameter!
H D2 D3 a
+0 I *0 +0 +0.004+0 + O. 002 r [,
K + 0J - ooa.M03 R ac
- 0.005 (REF) [ - 0.005 i- 0.005 -0 - 0,003 +00.005f+0 _-0.010 -0
- 0.010 -0
0.934
1.496
2,058
2.621
0.464
1.026
1.548
2.111
235_
I ,o.,o,[ " l
.155 [
' 0,190 0.433
190 '
l
.995
1.507
2.070
0.948
1.510
2.072i 0.14S
2.6351
i
3.080
3.620
4.618
2.570
3.110
4.108 |
I3
o .210 2529
3.070
4.070
3.0941
3.6351
4.6351
1
0.020t0.005 32 rms
4
I 1
5.616 5.116
6.614 6.114
I
5.070
6.070 6.635,
5.635' I
8.610 8.110 8.070 8.635i
10.606 10.106 .250 .017 . .1 4 .205 10070 10.635 .173
15.596
20.870
15.096
20.370 { 15.070
20.320
15.635 i
20.$85 i
L
aDimension increases 0.009-meh maximum after mekel plating and Teflon coating.
NOTES: 1. See fig. 9-14 for Naflex seal design.
2. See fig. 9-1_ for flange design.
3. Duct diameter D_ may be equal to D2.
370 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
582
Sc = 1 x 0.019 - 30600 psi
G AREA
767
Sc = 1 × 0.01---_
= 40 400 psi
_ The allowable
flange sealing
depth of imperfections
surfaces should be calculated
on the
The following data are given for a Naflex 9.4 DESIGN OF BRAZED JOINTS FOR
seal represented by section A-A of figure 9-14: ROCKET ENGINES
Design factor, C 1 =0.07; H=0.205 in; J=0.017
in; M=0.187 in. The seal material modulus of The best tube fitting or gasketed flange
elasticity, E=30× l0 6 psi. The flange seal joint falls short of a good welded or brazed joint
groove depth, a=0.173 in. The thickness of the in reliability. Hence, every sealed joint on each
Teflon coating, tc = 0.002 in, with a modulus of rocket engine should be studied aiming at its
elasticity, Ec = 5 x 104 psi. Estimate the seal elimination, or its replacement by a hermetically
coating compressive stresses with and without sealed joint, for increased overall system relia-
an internal pressure, p, of 2000 psi. Also, de- bility. Specifically, certain sealed joints origi-
termine the allowable depth of imperfections on nally provided for convenience during the devel-
the flange sealing surfaces. opment phase of an engine system may no longer
be needed as development aDproaches completion.
The brazed joint design shown in figure 9-16
Solution
may be utilized effectively to replace tube fit-
The seal deflection tings and flange joints. A union sleeve with
internal recesses for preplaced braze alloy foils
d = M - a = 0.187 - 0.173 = 0.014 connects the tube ends. The brazing can be
accomplished by induction-heating coils, either
From equation (9-19), the seal contact load in-place on the engine, or on the bench. Coax
without internal pressure cables with low impedance losses have been
developed, permitting a brazing fixture to be
80 × 106 × (0.017) 3 x 0.014 used in-place, even at great distances from the
= 582 Ib/in
Fsl- 8 (0.205- 0.02) a ×0.07 generator.
__-CLEAP, ANCE FOR BRAZING FIXTURE successfully induction brazed. Tubes and
sleeves for brazed joints can be made of alumi-
1 A /------'---INTERNAL RECESSES FOR PRE-PLACEO i num alloys, austenitic and semiaustenitic steels,
and nickel-base alloys such as Monel.
A preplaced braze alloy in the form of 0.001-
to 0.003-inch-thick foil is snapped into recesses
inside of the union sleeve. Diametrical clear-
ances of 0.002-0.008 inch between sleeve and
tubing are found to give satisfactory results.
I :
Holes are provided on the sleeve to permit visual
joint inspection after brazing. The quality of
the brazed joint can be also judged by the de-
gree of external braze-alloy filleting at the outer
edges of the sleeve. The recesses for the braze
alloy are located midway between tube and
sleeve ends. The length of the union sleeve can
be determined by the braze bond shear strength
and the effective bond area. Sleeve lengths
_---- UNION SLEEVE l
range from 0.5 to 2 times the tube outside diam-
J
eter.
Many types of brazing alloys can be used. A
Figure 9-16.-Typical brazed ]aint design. eutectic alloy of composition 71.8 percent silver,
28 percent copper, and 0.2 percent lithium, with
a melting point of 1435 ° F, can braze without
Pertinent design consideration and details of flux and permit in-place brazing.
brazed joints for rocket engines are as follows:
System Applications of Brazed Joints
Working Fluids, Pressures, and Temperatures
After an engine system has reached its pro-
Applicable to Brazed Joints
duction phase, the majority of the small- or
There are virtually no design limitations on medium-size sealed fittings and flange joints
the types of fluids that can be handled by brazed should be replaced by well-located brazed joints.
joints. Chemically active propellants such as These should be designed to facilitate in-place
fluorine, hydrazine, etc., have posed no problems brazing, if possible. This requires a specific
to brazed joints. Similar satisfactory results clearance around the joint for the brazing fixture.
have been obtained with brazed joints used in The minimum clearance A for the brazing fixture
liquid hydrogen, liquid oxygen, hydraulic, and (fig. 9-16) is 5/8 inch for tube diameters of 1/4
pneumatic lines. to 5/8 inch; 3/4 inch for a 3/4-inch diameter;
Structurally, brazed joints are designed to be 7/8 inch for a 7/8-inch diameter; and 1 inch for
as strong as the strongest tubing or duct of like 1- to 1-3/4-inch diameters. For repair of lines
material. Brazed joints have demonstrated relia- or servicing of components with brazed joints, a
bility at working fluid pressures as high as 4000 length of line may be cut out and a new length
to 5000 psi. Generally, brazed joints are recom- inserted and brazed in-place, with two union
mended to be limited to service temperatures of sleeves.
less than 1200 ° F. Allowance must be made for Figure 9-17 presents the in-place brazed
strength degradation of brazed joints used at joints applied to an upper stage propulsion sys-
elevated temperatures. tem. About S0 percent of the original fittings
A typicalbrazed jointis illustrated in figure and flanged joints were replaced by brazed
9-16. The union sleeve can be used to join tube joints, which eliminate considerable weight. In
to tubes, tube to bellows, or tube to components. addition, it also permits the use of high-strength,
Line sizes as large as 6 inches (insidediam- thin-wall tubing, instead of the older heavy-wall
eter)and walls as thinas 0.005 inch have been tubing.
372 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
Figure 9-17.-In-place brazed joints applied in Techniques for fluid-flow pressure drop deter-
an upper stage propulsion system. mination have been discussed in chapter VII.
Equation (7-7) may be used to estimate the pres-
sure drop of a straight duct section. Either
9.5 DESIGN OF DUCTS FOR ROCKET
equation (7-7) or (7-9) can be utilized to predict
ENGINES
the pressure drops of other shapes, for which
some experimental data are available.
Basic Design Considerations Besides avoiding excessive flow velocities
Primary design considerationsforvarious at various sections of a duct system, basic re-
ducts used on rocketengines are: quirements for minimum pressure drop are: con-
1. Fluid flow rate and system pressure drop.- stant flow area, smooth flow surfaces and path,
The size of the duct is determined largely by the minimum length, and few turns. In most duct
required flow rate and permissible pressure drop designs, a certain number of turns is unavoid-
of the flow system. An optimization must be able. The design of these turns affects consid-
made considering duct weight, pressure drop, and erably the overall pressure drop of a duct system.
space, before finalizing the duct size. A turn may offer a large resistance to the flow
2. Working temperature and pressure of the if not carefully designed. Investigations show
fluid.-The selection of construction materials that the flow resistance in a bend with constant
and structural design for ducts depends mainly cross section is affected directly by its turning
on the working temperature and pressure level of radius ratio R/D (where R = radius of curvature
the fluid, but also on chemical compatibility. of the axis of the duct, D=diameter or width of
Temperature also influences line flexibility and the duct) (see fig. 7-21). By increasing the turn-
duct geometry. ing radius ratio, large reductions in pressure
3. Duct/oints.-The number and type of joints drop are possible. Bends with a circular section
in a duct system are usually determined by sys- are better than those with a square section.
tem assembly and servicing requirements. All However, the square section may be much im-
mechanically sealed joints should be reviewed, proved by changing it to a rectangle, so that the
at various engine design and development phases, turn is made on the short side; that is, the loss
for their possible elimination or replacement by can be decreased by increasing the rectangle
in-place, induction-brazed or welded joints. (See aspect ratio W/D (where W= width of the long
secs. 9.8 and 9.4.) side, D = width of the short side). For a duct
4. Line flexibility and geometry.-The re- with a rectangular section, therefore, a very
quired degree of flexibility in a duet system is efficient corner may be produced if the values of
dictated by component tolerance buildups, mis- both the radius ratio and the aspect ratio are
alinements, and required freedom of movement to kept large.
DESIGN OF INTERCONNECTING COMPONENTS AND MOUNTS 373
,XIAL
"qF--- OEFLECTION
1.2
I,I
(diPARALLEL 1.0
OFFSET DF .--L°---_
ARTICULATED p.--- L--_I
BELLOWS 0.9
0.8
07
0.6
Figure 9-21.-Elements and various motions of
bellows. 0.5
0.4
0 JO 20 30 40 5O 6O
degree
Sb :bellows bulging stress, psi Figure 9-22.-Bellows wall thinning correction
Sa =bellows hoop stress, psi [actor Ct versus percent thinning.
Sm = bellows motion stress, psi
Ss :bellows shear stress, psi the minimum thicknessat the convolution outside
St =bellows torsion stress, psi
diameter(fig.9-22). Amounts of thinningrange
T =torsional moment, in-lb
from I0 to 40 percent. Effects of thinning,
are
Tcr : critical stability torque of the bellows, psi
considered by applying the thinningcorrection
v = Poisson's ratio of the bellows material, in
factorCt (fig.9-22) to the bellows design.
y = transverse deflection of the bellows, in
2. Bellows axial spring rate:
0 =bending angle of rotation, degrees
6 =torsional angle of rotation, degrees
1.49 CtCpNpEdit _ lb/in
1. Thinning of the bellows wall.-Hydraulic- Ra - Nch_
formed bellows are usually made by starting with
(steel and nickel alloys) (9-22)
tubes of the same diameter as the bellows diam-
eter at the root of convolution, d i. The typical
thinning profile of a bellows wall starts with the 1.23 CtCpNpEdit 3 , lb/in
Ra- Nch _
original material thickness at the root of the
convolution, and tapers approximately linearly to (aluminum alloys) (9-23)
376 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
229.2 FsLa
6. Motion stresses due to axial de[lections of (9-32)
0=' Radm _ , degrees
the bellows.-These are due to bending of the
bellows side wails. Allowable motion stresses
3dmY
for bellows materials, with respect to design --, in (9-33)
es-- 2L a
cycle life, are given in table 9-9:
TABLE 9-9.-Yield Strength, Limiting Bulging Stresses, and Allowable Motion Stresses of Frequently
Used Bellows Materials
321 and 347 stainless steels ........................ 39 000 140000 120000 208000 150000' 92000
19-9DL ............................................ 88000 140000 120 000 208000 1500001 92000
A-286 ............................................ 180000 190000 150000 160000 150000 138000
lnco 718 .......................................... 170000 190000 150000 160000 150000 138000
lnconel X ......................................... 98 000 190 000 150000 160000 150000 138000
6061-T6 aluminum alloy ............................. 40 000 65000 106000 68000i 28000
DESIGN OF INTERCONNECTING COMPONENTSAND MOUNTS 377
3dmy
14. Bellows squirm due to torsion.-When a
ep----L--_a,
in (9-36)
bellows is loaded by pure torsion, it tends to
buckle in some manner as with internal pressure
Motion stresses due to paralleloffsetcan be
squirm:
calculatedby substitutingep forea in equations
(9-27) and (9-28).
rrdm2Ra
10. Parallel offset of articulated bellows Tcr- 2 , in-lb (9-44)
(fig. 9-21d):
The values for Pcr and Tcr of bellows under
Fs 3dm2Ra angular and offset deflections will be reduced
Rp=-7=- 2 considerably. A correction factor which ranges
from 0.2 to 0.9, as determined by experiments,
[4La 2+6LaLb+3Lb2],Ib/in (9-37)
should be applied.
Design a bellows, as shown in figure 9-4, for din= _/(di2+ 2 d°2) - _ (82+ _ '822)
-- - 8.42 in
Comparing this with the maximum working pdd 2tr pdm 2rr_ 1750× (8.42) 2× ,7
pressure of 1750 psi, a safety factor of F=FP+ 4 - 4 4
3080/1750= 1.76 remains to allow for bellows
= 97 500 Ib
stability under conditions of angulation.
From sample calculation (9-2), the yield
pressure =2117 psi, and the ultimate pressure 9.6 DESIGN OF GIMBAL MOUNTS
= 2887 psi. Substitute these into equation (9-26),
to obtain the yield hoop stress of tile bellows: Gimbal Design Considerations
.'
380 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
THRUST CHAMBER
383
384 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
dlmmma
ENGINE SYSTEMS DESIGN INTEGRATION 385
processes. By describing these components and Some physical processes, such as thrust
processes in detail, as well as their interaction, chamber combustion dynamics, are not always
the system can be described analytically with as quantitatively fully understood. Rate of combus-
much detail as is necessary. The complete set tion is known to be a function of pressure, pro-
of equations then represents a mathematical pellant type, mixture ratio, and combustion
model of tile engine system. chamber geometry, but a specific quantitative
Through the solution of the equations repre- expression for reliable use with rocket engine
senting the mathematical model, the important combustion chambers is not available. Certain
characteristics of an engine system are studied, system-start-transient analyses made with the
problem areas are defined, and improved compo- aid of an engine model (set of equations like eq.
nent designs for the solution of these problems (10-2)) have ideally assumed instantaneous com-
may be evaluated. Also, transient and steady- bustion. Thus, combustion instability or thrust-
state engine systems operation, as affected by chamber-feed-line, system-coupled instability is
various component characteristics, may be simu- not described: This deficiency can be rectified
lated and checked by these dynamic analyses. by experimental systems evaluation.
If necessary, modifications will be incorporated Dynamic analyses can also be effectively
prior to hardware testing. used during the development-redesign phase of
Dynamic analyses, however, have their limita- an engine system. Once test information is
tions, because not all of the physical processes available, the predicted characteristics (with
involved in a given rocket engine system are idealized assumptions) and the actual system
immediately and/or thoroughly understood. As operating characteristics can be compared. Dif-
each of the processes becomes better defined ferences can be noted and evaluated. This test-
functionally and quantitatively, confidence in the analysis cycle defines the limits of component
mathematical analyses increases. performance and thus serves as a guide for the
Let us look at an example. The hydraulic redesign of the components to be integrated into
head developed by a propellant pump is, as we an optimum, final engine system. Similarly, from
know, a function of pump speed, flow rate, and analyses of the engine-vehicle operating dynam-
geometry. We can write: ics, and in conjunction with test results, the
engine system and its components can be modi-
H=[(N, Q, r, A) (10-1) fied and improved.
where
H = pump developed head, ft
Criteria for the Mathematical Model of an Engine
N = pump speed, rpm
System
Q = pump flow rate, gpm
t = pump impeller radius, in The mathematical model of an engine system
A=area normal to the meridional flow, in 2 generally consists of a group of lumped param-
Correlations, such as equation (10-1), can be eters, and of linear or nonlinear algebraic and
used to determine the interdependence of the differential equations, which are formulated and
many processes within an engine system. then programed for an analog or digital computer.
Furthermore, equation (10-1) may be expressed Careful examination of an engine system sche-
as a specific form of function, as shown by equa- matic will be sufficient to determine whether a
tion (10-2), which is valid for a particular pump mathematical model will be possible for the sys-
design only. tem. This simply amounts to an observation of
the many significant physical processes involved
H = aN _ + bNQ + cQ (10-2) in the entire system which may be expressed
mathematically. Some idealized assumptions are
When the numerical values of a, b, c are usually required to obtain a quantitative expres-
known, equation (10-2) becomes a quantitative sion of the various equations. The physical
description of a given pump design and a means significance of these assumptions must be under-
to obtain the numerical solution of the operation stood before the mathematical model can become
These can usually be indicated by the Laplace Vc =volume of combustion chamber from
transformation operators p-ros and p-rfs. If we injector to throat, in 3
assume a homogeneous combustion gas, we can R' = universal gas constant, 18 528 in-lb,/°R
define the following correlations: mole
_Ii =molecular weight of the combustion
gas, lb/mole
pc=R,(\T/kV3
(wc (lo-8) Wc = weight of the gas stored in the com-
bustion volume, lb
Wo
Ro =weight fraction of oxidizer stored in
Ro = (w ° + wt) (10-10) the combustion chamber
At =throat area, in 2
g =gravitational constant, 22.2 ft/sec 2
C* = ctiaracteristic velocity, ft/sec
wo=_o+ (p-r°Sfvo-Rofvc)dt (10.11)
These equations for the combustion process can
be applied to engine main thrust chambers, gas
a =pump pressure rise design factor, lb- fer and combustion characteristics
Rg = speed ratio of the turbopump gear train perforalance variations and engine int2u-
ence coefficients. (See see. 10.4.)
Tp =torque at pump shaft, in-lb
Tt = torque at turbine shaft, in-lb (5) Evaluation of various engine control prob-
_Pt =turbine gas flow rate, lb/sec lems during main-stage operation, such
AH = available energy content of the turbine as thrust and mixture ratio controls
Cp = turbine gas specific heat at constant pres- tions and their effects on mainstage
sure, Btu/lb-°F operation
To =turbine gas total temperature at inlet, °R Once the basic mathematical model for the
I =inertia of the gear train, referred to the problems with additional inputs. Vor example, a
main pump shaft, in-lb-sec 2 basic mathematical description of the 150 000-
r/t =turbine overall efficiency control subsystem, its electronics, main valves,
and servovalve drive system were tied to an
analog computer by suitable transducers to allow
Dynamic Analysis of Engine System Mainstage transient performance checkout and controller
Operation gain adjustment. An updated mathematical en-
gine model, including the nonlinear perturbation,
In general, engine design requirements at the
was then used for more detailed investigations
mainstage level, as well as initial component
of the thrust-control-loop dynamics.
and system specifications, can be determined
with the aid of a mathematical model consisting
of linearized descriptions of the complete engine
Dynamic Analysis of Engine System Start and
system and a computer. Based on the static
Cutoff Transients
operating values at main stage (such as given in
tables 3-2 to 3-5), tile static design factors (such The main objectives of dynamic analyses of
as a and b in eqs. 10-2, 10-3, 10-16, and 10-22) engine system start transients are:
may be obtained. The primary dynamic analysis (1) Investigation of the systems schematic
objectives for mainstage operation are: for needed start-transient controls, snch
as type and quantity of control compo-
(1) Evaluation of the engine system sche- nents, and sequencing and timing of
matic, with respect to mainstage opera- their operation
tion (2) Determination of thrust chamber ignition
(2) Evaluation of the dynamic characteristics, requirements
interactions, and the performance of (3) Estimation of start energy, time, and
various components at mainstage level. thrust buildup characteristics
ENGINE SYSTEMS DESIGN INTEGRATION 389
800 _---_- f--_. --- "_--_ --_- ....... ; F._ --./_ .... -OPEN
ture surges, and propellant pump stalling r "{l I / ' MA,N ////"_"--OX,D,ZER
' FUEL , , PUMP I
(5) Evaluation of system dynamic stability
7°°!1/ r /M_IN [ I / i II .....
during the start transient. (The aim is I ,/ '/OXiDiZe'..
I .....
I/ I t/ GASGENERATOR
0 600 --_--'_-" ..... _ t1 --- PPESSbRE-'_ . -CLOSED
to avoid prolonged operation at levels
exhibiting system or thrust chamber
instability.)
IC-NITER FUEL IGNITION--,( ,/ _,_"/
(6) Evaluation of various potential perturba-
tions and their effects on start transients,
such as a start where the propellant-
settling effects of gravitation are absent r PRPRESSURE
200 --
For some engine systems, such as the
LO2/LH2 turbopump-feed A-2 stage engine, IO0
_____._.__._>
ENGINE SYSTEMS DESIGN INTEGRATION 391
OE SIGN POINT--\ /
\ / / _ SYSTEM DESIGN PRESSURE VS FLOW CURVE
/ _ OR SYSTEM RESISTANCE CURVE WITHOUT
PRESSURANT LINE PRESSURE DROP_._ _'x= _
CALIBRATION ORIFICE "_1 -- _=/-_/ _ ORIFICE
DESIGN PRESSURE DROP-_._ /I __--*_---'-_PRESSURANT L_NE PRESSURE VS FLOW CURVE
DESIGN PRESSURE DROPS
/ _ {AT ENTRANCE TO TANK)
iN VALVES AND LINES_ _
// _ / _ DESIGN CHAMBER INLET PRESSURE VS FLOW CURVE
DESIGN PRESSURE DROP !N THE // /i /-_OESIGN CHAMBER INJECTOR END PRESSURE VS
CHAMBER ( MANIFOLD COOLING_ /f _ ./ PLOW OURVE
PASSAGE AND INJECTOR -_
DESIGN CHAMBER
_N,JECTOR END PRESSURE J
Figure 10-5.-Propellant flow design characteristics of a typical pressure feed engine system
(oxidizer or fuel).
---.... ,,,
,---FUEL PUMP PRESSUI_:
CALIBRATION
"_, _ VS FLOW CURVE AT NZ
ORIFICE DESIGN
t
Pfd
°°°
J" OX,Q, ZER
PUMP
PRESSURE
i i
FUEL FLOW LB/SEC
OXIDIZER FLOW LB/SEC
Wfd wfo Wfb %a Wod
Figure IO-E-Propellant flow design characteristics of the A-I stage turbopump feed engine system.
of its flow characteristics from its de- ratio, and Is (as verified by actual thrust cham-
sign value will be kept within a reason- ber test firings). Based on these flow rates, the
able limit, in order to facilitate systems pressure drops of the various components at the
calibration, and to keep other system design operating point can be estimated from
components in their design operating previous design data, or as obtained from actual
region. testing. Certain components may have to be
(3) Sufficient pressure head should be set newly designed for the specific design pressure
aside in each engine propellant feed drops allowed by the system.
system to compensate for contingencies The design pressure versus flow curve of
due to component flow resistance devia- either propellant flow system can be obtained by
tions. The propellant feed system can the summation of design chamber pressure versus
then be calibrated by means of orifices flow and component design pressure drop charac-
or other adjusting means. teristics, as shown in figure 10-5. In addition,
an orifice is introduced in each propellant flow
system for calibration. The minimum required
Design for Calibration of a Pressure Feed System
tank _ressure versus flow curve for each propel-
The first design step is the determination of lant is thus derived. In most pressure feed sys-
the design flow rate of each propellant, as calcu- tems, the design orifice pressure drop for
lated from rated systems thrust, design mixture systems calibration determines the maximum
..... _ ----_:::.:=?:
392 DESIGNOF LIQUID PROPELLANT ROCKET ENGINES
allowable cumulative pressure drop increase of feed system are similar to those of a pressure
the components above their nominal values. A feed system. However, the difference in turbo-
suitable tank pressurization system can then be pump pressure or head versus flow characteris-
designed, compatible with minimum required tank tics from those of a pressurized system dictates
pressure versus flow characteristics. a somewhat different approach to systems cali-
bration. For mechanically coupled turbopump
Sam ple Calcu la tion (I0-I)
feed systems, such as the A-1 stage engine,
The following data are available from analyses systems calibration generally involves adjust-
and component tests for the A-4 stage propulsion ment of the turbopump speed as well as the in-
system, at rated thrust conditions: stallation of an orifice in one of the propellant
Thrust chamber injector end pressure range lines. For turbopump feed systems with dual
required to maintain rated thrust = 110 ¢ 3 psia turbine drive, such as the A-2 stage engine, the
Thrust chamber injector pressure drop range calibration can be accomplished by adjusting the
(both oxidizer and fuel) = 25 ¢ 2 psi speeds of both turbopumps.
Thrust chamber oxidizer dome pressure drop The design principles for the calibration of
=3_ + 1 psi mechanically coupled turbopump feed engine
Oxidizer line pressure drop = 5-+ 1 psi systems are best illustrated by a typical exam-
Main oxidizer valve pressure drop (at the fully ple, as shown in figure 10-6. Here, the propel-
open position) = 4 ¢ 1 psi lant system resistance curves without orifices
Thrust chamber fuel manifold pressure drop (representing conditions downstream of the pump
=4_+1 psi discharges) are constructed based on the designs
Fuel line pressure drop = 4 ¢ 1 psi and test results of the components for the A-1
Main fuel valve pressure drop (at the fully stage engine system. Next, the discharge pres-
open position) = 4 ¢ 1 psi sure versus flow curves of both pumps are con-
Pressure allowance required for mixture ratio structed from test data obtained with the A-1
control by oxidizer valve vernier positioning stage engine turbopump, operated at speed N1.
(fig. 7-4)=¢ 10 psi These pump curves intersect the corresponding
Determine the design pressure drops of the cali- system resistance curves at point A. At this
bration orifices, and the minimum required tank speed, fuel flow rate _/fa is above, and oxidizer
pressures for design flow rates. flow rate _#oa is below the required design flow
rates, Wfd and _i,od.
Solution
To achieve the design oxidizer pump flow
The design pressure drop of a calibration _Pod, at a desired discharge pressure Pod, the
orifice must be equal to the sum of the maximum design operating speed of the turbopump assem-
pressure drop increases of components above bly mr, st be raised to a required level N 2 by in-
their design values. Thus: creasing the turbine gas flow. However, at this
The design pressure drop of the oxidizer cali- speed, the fuel pump, which is mounted on the
bration orifice = 3 + 2 + 1 + 1 + 1 = 8 psi. same shaft as the oxidizer pump, would be de-
The minimum required oxidizer tank pressure livering a flow rate d'fb considerably above the
at the design flow rate = 110 + 25 + 3 + 5 + 4 + 8 + 10 required design flow rate 1//fd (point B in fig.
= 165 psia. 10-6). To reduce the fuel flow to _fd, a calibra-
The design pressure drop of fuel calibration tion orifice is placed in the fuel line. This
orifice=3+2+l + 1 +1 =8 psi. amounts to increasing the fuel pump discharge
The minimum required fuel tank pressure at the pressure at constant speed hr2 to Pfc, where _'fd
design flow rate = 110+ 25 + 4 + 4 + 4 + 8 = 155 psia. is reached at point C. The pressure drop across
the calibrating orifice is represented by Pfc-Pfd,
where Pfd is the desired fuel pressure.
Design for Calibration of a Turbopump Feed
If fuel flow rate _/fa is below and oxidizer
System
flow rate _/o._ is above the required design flow
The propellant flow characteristics down- rates, the calibrating process would be to speed
stream of the pump discharges of a turbopump up the turbopump to obtain the desired fuel flow,
ENGINE SYSTEMSDESIGN INTEGRATION 393
and to place an orifice in the oxidizer line. Determine the location of the calibration orifice,
However, it is generally desirable to place the its nominal design pressure drop, and its ex-
orifice in the system of the propellant with the pected range of adjustment.
higher boiling point. In this situation, therefore,
and also when the pressure drop across a cali- Solution(see sample calculation(6-2))
brating orifice tends to become excessive, it is
customary to trim the pump impeller so as to The required oxidizer pressure head at the
reduce the effective speed, and thus attain the design point = 1095 + 200+ 150 + 25 + 35 = 1505
required flow and pressure levels. In view of psia.
pump efficiency effects, it is desirable to trim The required fuel pressure head at the design
the pump drawing the smaller horsepower, usually point = 1095 + 200 + 290 + 10 + 15 = 1610 psia.
the one with the lower mass flow rate, except in Since the LOX pump discharge pressure is
cases of extreme density differences. The ad- 1505 psia, but the fuel pump discharge pressure
justment of the turbine gas flow rate, and thus is 1720 psia, the calibration orifice must be
the turbopump operating speed, can also be made located in the fuel system.
by means of orifices in the turbine inlet gas line, The nominal orifice design pressure drop
or in the gas generator propellant lines. = 1720- 1610= 110 psi.
In general, turbopump feed systems afford From a detail analysis, we have found that
less stringent requirements for the various com- the change of the fuel pump discharge pressure,
ponents regarding deviations from their design as a function of turbopump speed increase or
steady-state flow values, because the system is decrease, is a fraction of that of the oxidizer
inherently more flexible. However, systems pump discharge pressure. Due to the effects of
dynamic characteristics under transient condi- chamber pressure deviations, therefore, the max-
tions may restrict these deviations. imum value of fuel calibration orifice pressure
drop is required when the following conditions
Sample Calculation (I0-2) exist:
(a._) Thrust chamber injector end pressure is at
The following design values and allowable
its lower limit (1065 psia)
deviations are given for the A-1 stage LOX/RP-1
(_b) All pressure drops in oxidizer passages
engine system components, at rated thrust:
are at their higher limits
Thrust chamber injector end pressure= 1095
(c_) All pressure drops in fuel passages are at
_+30 psia their lower limits
Thrust chamber injector pressure drop (both
(d.._)Oxidizer pump discharge pressure is 25
oxidizer and fuel) = 200 ± 20 psi
psi below its nominal value at the turbo-
Thrust chamber oxidizer dome pressure drop
pump speed commensurate with the
=150+10 psi
stated specific speed
Oxidizer line pressure drop = 25 -+2 psi
(e) Fuel pump discharge pressure is 25 psi
Main oxidizer valve pressure drop = 35 ± 3 psi above its nominal value, at the same
Oxidizer pump specific speed, Ns = 1980 rpm
speed
Oxidizer pump suction pressure = 55 psia rain
Oxidizer pump discharge pressure at 7000 rpm The equivalent required oxidizer pump dis-
and a design flow rate of 1971 lb/sec = 1505 charge pressure under these conditions = 1065
-+25 psia +220+160+27+38+25= 1535 psia.
Thrust chamber fuel jacket and manifold pres- Required oxidizer pump developed head
sure drop = 290 _+20 psi
Fuel line pressure drop = 10 + 2 psi H = 144 × (1535 - 55) _ 2990 ft
Main fuel valve pressure drop = 15 +-2 psi 71.38
Fuel pump specific speed, Ns = 1090 rpm
Fuel pump suction pressure = 45 psia rain Oxidizer pump volumetric flow rate
Fuel pump discharge pressure at 7000 rpmand
a design flow rate of 892 lb/sec = 1720-+ 25 1971 ×449 12 420 gpm
Q= 71.38 =
psi
......LL ::---
394 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
or
N- Ns H°'Ts _ 1980 × (2990) °'Ts = 7190 rpm
QOS (12 420) °-s 4730 × 50.45
= 1666 psi
144
Fuel pump volumetric flow
or
10.4 ENGINE SYSTEM INTEGRATED
PERFORMANCE CHARACTERISTICS
4900×50.45
-1715 psi
144 In the process of engine system design
integration, an importanttask is the integration
The equivalent fuel pump discharge pressure of engine system perfonnance characteristics.
under these conditions would be 1715 +45 + 25 These data are preparedand compiled by the
= 1785 psia. rocketengine designer to provide the vehicle
The required pressure drop for the fuel line systems engineer with intbrmationnecessary to
calibration orifice thus would be 1785-1065-180 integratethe propulsionsystem with the vehicle
- 270 - 8 - 13 = 249 psi. system. _ere possible,a briefexplanationof
Similarly, a minimum fuel calibration orifice the data and itsapplicationshould be included
pressure drop is required when the following to provide clearerunderstandingand greateruse-
conditions exist: fulness. The followingare importantaspects of
(a_a_)Thrust chamber injector end pressure is at integratedengine performance characteristics.
its higher limit (1125 psia), and condi-
tions (b), (c), (d), and (2.) above are re-
Nominal Engine Performance Values at Rated
versed
Conditions
The equivalent required oxidizer pump dis-
charge pressure under these conditions = 1125 These are usually prescribed by the engine
+ 180 + 140 + 23 + 32 - 25 = 1475 psia. model specification. These data are for engine
Required oxidizer pump developed head system nondnal steady-state operation, at rated
conditions. Tables 3-2 to 3-5 are typical exam-
H=144(1475-55)=2870 ft ples of nominal engine operating and performance
71.38 parameters, which include nominal thrust, spe-
cific impulse, propellant combination, flow rates,
Substitute this into equation (6-7); the re- mixture ratio, and various component operating
quired pump speed data. Allowable deviations are specified for
important parameters such as: thrust, -+3 per-
N =1980×(2870)°_ =6970 rpm cent, and mixture ratio, *-2 percent. Engine sys-
(12420) °.s tem specific impulse is usually specified at its
minimum value. The performance of all deliver-
From equation (6-7), the fuel pump nominal able engine systems must be above this minimum
developed head during acceptance tests.
ENGINE SYSTEMSDESIGN INTEGRATION 395
In addition to tables for nominal engine per- Required oxidizer flow for vehicle tank pres-
formance parameters, nominal engine performance surization-- 3 lb/sec
graphs such as chamber pressure versus engine Determine the following nominal performance
thrust, and engine specific impulse versus engine values at rated conditions:
thrust, are often included as additional monitor- _.) Thrust generated by the turbine exhaust
ing aid. Figure 10-7 presents a typical perform- gas
ance graph for the A-1 stage engine system, of (b) Thrust generated by the main thrust cham-
chamber pressure versus engine thrust at sea ber
level. _) Engine system propellant flow rates
Engine system mixture ratio
(e_) Engine system specific impulse
Sample CalcuIation (10-3)
Fuel pump developed head = 4790 ft used to approximate the corresponding oxidizer
and fuel flow rates:
Fuel pump overall efficiency =65.9 percent
Gas generator O/F mixture ratio = 0.408
Turbine gas available energy content = 359 Engine oxidizer flow rate
Btu/lb 2778x2.35
Turbine overall efficiency = 58.2 percent _i'° = (2.35+ 1) - 1948 lb/sec
Required auxiliary drive shaft power = 500 bhp
Oxidizer flow rate = 1948 + 3 = 1951 lb/sec
1951 × 2930
-r'
_' ] IOO
Oxidizer pump horsepower - 550 × 0.706 = 14720hp
u
::)W -
(_ 1050 --/-- -- Fuel pump horsepower - 830x 4790 = 10 96Q hp
-r / CONFIDENCE
550 x 0.659
I.- / LIMITS [
I000 --_ _
The corresponding turbine shaft horsepower:
Figure lO-7.-Chamber pressure versus engine From equation (6-19), the corresponding tur-
thrust at sea level [or the A-1 stage engine. bine gas flow rate:
396 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
Thrust chamber flow rate This value closely confirms the assumptions
for the last trial. Thus:
-,qmNlm
ENGINE SYSTEMS DESIGN INTEGRATION 397
Engine Influence Coefficients example, the change of engine thrust for the A-1
stage engine system (without C* correction) can
These are used to convert or correct steady-
be expressed as:
state, main-stage engine system performance
parameters (dependent variables) from one condi-
tion to another of parameters (independent vari- ( F - F n) _ C l(pa- Pan)+ C 2(po - pen)+ C 3([ [- pfn)
Fn Pan pon Pfn
ables) such as atmospheric pressure, fuel tem-
perature, oxidizer density, etc. This may be a C4(Poi- Poin) Cs(Pfi- Pfin)
correction to standard sea-leveI conditions (first- + + (10-26)
Poin Pfin
stage booster engine), or a conversion to other where
specified conditions. The coefficients are de- F, F, : engine system thrust and its nominal
rived from the linearized solution of a set of value, lb
steady-state differential equations which describe
Pa, Pan : atmospheric pressure and its nominal
the performance of an engine system. These •value, psia
equations are solved by a digital computer and
Po, Pon : oxidizer density and its nominal
presented in tabular form, as shown in table 10-1 value, lb/ft 3
for the A-1 stage engine system. Each influence
P[, Pfn = fuel density and its nominal value,
coefficient is expressed as a percentage and lb/ft s
represents the change of a dependent engine Poi, Poin :oxidizer pump inlet suction pressure
variable, such as thrust, as produced by a and its nominal value, psia
1-percent change in an independent variable, Pfi, Pfin = fuel pump inlet suction pressure and
such as atmospheric pressure. A coefficient its nominal value, psia
preceded by a positive sign (+) indicates that an
C_, C 2, C a, C 4, C s :influence coefficients
increase of an independent variable produces an
increase in the dependent variable. Conversely,
Sample Calculation (10-4)
a coefficient with a negative sign indicates a
decrease in the dependent variable, as a result Estimate the thrust of the A-1 stage engine
of independent variable increase. These influ- system operated at the following conditions,
ence coefficients are usually sufficiently accu- without considering the effects of C* correction:
rate over the entire design operation range of an Atmospheric pressure, Pa = 10.2 psia
engine system. Oxidizer density, Po = 71.00 lb/ft 3
Because the influence coefficients are linear, Fuel density, p[= 50.90 lb/ft 3
the total effects of several influences acting Oxidizer pump inlet suction pressure, Poi = 65
simultaneously on an engine system can be de- psia
termined by summing the individual effects. For Fuel pump inlet suction pressure, Pfi--49 psia
[Value of C* correction to be obtained from the C* correction versus mixture ratio curve shown in fig. 10-8]
Solution 10-1, which are for the A-1 engine system. The
change of engine mixture ratio is computed for
From equation (10-26) and table 10-1:
changes in atmospheric pressure, propellant
densities, etc., assuming the C* correction first
(F - Fn)_ (-0.178) x (10.2 - 14.696)
to be zero. For the resultant change in engine
F_ 14.696
mixture ratio, the C* correction is read from the
graph. The value of C* correction found is then
+ 1.875 x (71.0 - 71.38) _ (-0.742) x (50.90 - 50.45)
71.38 50.45 used with other independent variables to compute
the changes in the remaining dependent variables.
0.044 × (65- 55) + (-00066) × (49 - 45)
55 45
Sample Calculation (10-5)
=0.04531 or 4.531 percent
Estimate the thrust of the A-1 stage engine
Engine system thrust of an altitude, where system operated at the conditions listed for
Pa = 10.2 psia: sample calculation (10-4), adding the effects of
C* correction, Also, for the same conditions,
F = 0.04531 x Fn + Fn = 0.04531 x 750 000 + 750 000 estimate the thrust assuming an additional mix-
: 784 000 lb ture ratio error of +10 percent, due to faulty
calibration.
Nonlinear Corrections
Solution
When the linear approximation is not suffi-
ciently accurate, the usefulness of the engine By analogy with equation (10-26) and using
influence coefficients can be extended by a table 10-1, the engine system mixture ratio
technique which allows nonlinear corrections for change due to the conditions of sample calcula-
certain parameters. An example of this method tion (10-4) are determined as
is the C* correction. For instance, a plot of C*
correction versus engine mixture ratio change (MR- MRn) _ 1.642 x (71.0- 71.38)
may be used in conjunction with a table of influ- MR n 71.38
ence coefficients such as figure 10-8 and table
-_ (-1.365) x (50.90 - 50.45) ÷ 0.027 × (65- 55)
50.45 55
_ -0.4
u (Percent change in F) = 4.531 + (-0.02) x 1.103
PERCENT
F= 750000 × (1 + 0.04509) = 783 820 lb
Figure 10-8.-C* correction versus change in
engine mixture ratio curve for the A-1 stage If the mixture ratio error of 10 percent is
engine. added, the total mixture ratio change = 10-1.62
ENGINE SYSTEMSDESIGN iNTEGRATiON 399
= 8.38 percent. From figure 10-8, the C* correc- in the vehicle in the field should not require
tion then is approximately -0.11 percent. assembly of additional major components. Integ-
Thus rity of the propellant feed and hot-gas systems,
(Percent changein F)=4.531 +(-0.ii)× 1.103 once verified in a complete system during ac-
ceptance test, is not necessarily mdlified by the
=0.441, or 4.41 percent need to temporarily disassemble the engine for
Engine system thrust: shipment. The integrated engine package con-
cept provides added assurance that static test-
F = 750 000 × (i + 0.0441) = 783 080 Ib
stand firing results have verified structural
soundness of the package to a substantially
greater degree, than is the case for a system
10.5 MECHANICAL INTEGRATION OF where the vehicle provides portions of the engine
ENGINE SYSTEMS structure.
An example of a special case of mechanical
Basic Considerations integration of a liquid propellant rocket engine
is the prepackaged storable liquid rocket propul-
Besides combining allcomponents and sub-
sion system shown in figure 8-1. This system is
systems functionallyand physically,the design
a completely integrated assembly of all-welded
formechanical integrationof an engine system
construction, consisting of thrust chamber as-
must considerti_eoverallenvelope of the system
semblies, propellant tanks, pressurization sys-
and itsweight. This includes the locationof
tem, and necessary controls. This provides
the system's centerof gravity. Also, itshould
maximum assurance of system integrity from the
permit simplifiedmaintenance and checkout
time of manufacture, which includes loading of
practices. Judiciouspackaging design tech-
the propellants, through delivery, vehicle assem-
niques should be applied to minimize the number
bly, and launch. Complete propellant separation
of interconnectinghydraulic,pneumatic, and
until systems start is achieved by hermetically
electrical lines, with their attendant fittings,
sealed burst diaphragms for maximum safety.
connectors, joints, and other potential trouble
Acceptance tests are conducted by taking sample
spots. Welded and brazed joints should be used
units at random from the production line, and hot
as much as possible. Problems introduced by
firing them. In addition, destructive tests of
vibration, high temperatures and pressures, leak-
various types are performed.
age and space restrictions are thus more easily
handled. Engine mechanical integration is a
Packaging of Rocket Engine Components
vital part of the system design concept; there-
fore, all factors related to integration and pack- Most major rocket engine components, such
aging of components and subsystems must as thrust chamber (fig. 4-1) and turbopump (fig.
receive careful consideration early in the prelim- 6-14) assemblies, readily form a logical, inde-
inary design stage. pendent mechanical unit by virtue of their func-
In general, a modular engine packaging ap- tion and their physical shape. However, in the
proach should be selected such as used for the case of minor components such as control valves,
A-1 (fig. 3-2) and A-2 (fig, 3-4) stage engine gas generators and igniters, packaging design
systems, as well as for many advanced opera- principles can best be served by making them an
tional engine systems. This assures engine integral part of a major component assembly, or
integrity from time of manufacture through vehi- to integrate them by grouping. A typical example
cle launch. It also provides a compact package is a gas generator assembly externally attached
for ease of handling, transportation, and installa- to a turbine inlet flange (fig. 3-2). Similarly, gas
tion in the vehicle. Ease of checkout and com- generator propellant valves and combustor can
ponent accessibility is also afforded by the be integrated into one unit (fig. 4-51).
packaging concept. Certain types of hydraulic and pneumatic
The engine should be completely assembled rocket engine control components lend them-
in the manufacturer's plant. Subsequent accept- selves most conveniently to the packaging de-
ance testing, air transportation, and installation sign. Here, one of the main objectives is to
4O0 DESIGNOF LIQUID PROPELLANT ROCKET ENGINES
reduce line runs, by combining all parts and sure regulator to the control system. The helium
passages into one housing. Such a housing (or is routed internally to the main control valves
mounting plate) is relatively leakproof, trouble through a fail-safe check valve. This insures
areas now being limited to external line connec- that the various engine propellant valves remain
tions to other components. Furthermore, if com- pressurized and thus open, should the helium
ponents are packaged in this manner, reductions gas supply system fail.
of weight and size are achieved through the use
of common walls and through the elimination of
extra mounting platforms, clamps, and fasteners. Packaging of Turbopump Feed Engine Systems
Since relatively few packages are required as In earlierhigh-thrustrocket propulsionsys-
compared to the usually large number of individ- tems, some of which may still be in operational
ual components, maintenance of such a system use, allmajor engine components were mounted
is greatly simplified. Integrated packages are into a cage-shaped thrust mount, which was
about as easily removed and replaced as are the bolted to the vehicle thrust frame by way of lugs.
separate components making up each package. Figure 2-4 allows several typical examples. With
However, the packaged design is not necessarily these systems, vehicle steering was accom-
desirable for every control system. Each case plished by means of carbon jet vanes protruding
must be carefully studied. into the jet (V-2 and Redstone), or by swiveling
As a rule, one or a combination of the follow- the thrust chamber (Thor, Jupiter). In the latter
ing methods is used for packaging engine control case, the high-pressure feed lines between
corn ponents: pumps and injector had to be much more flexible
(1) Bank packaging: A group of similar flat- than for misalinements and thermal expansion/
sided component assemblies are bolted contraction alone.
together in a bank or stack, with common Most advanced liquid rocket engines are
porting through the mating surfaces from tightly packaged. All major components are
one unit to the next. attached to the main thrust chamber, directly or
(2) Subplate packaging: Attachment of two or by means of mounting structures, as shown in
more individually housed components to figures 3-2, 3-4, and 9-1. Here, the thrust cham-
a subplate, so that all ports of the in- ber serves as the principal structural member of
dividual component housings lead into the entire engine system. For steering, the com-
the subplate manifold, through their plete engine package is gimbaled from a gimbal
mating surfaces with the subplate, and bearing which attaches directly to the thrust
on to the systems plumbing. chamber dome. The other half of the bearing is
(3) Cartridge packaging: Two or more compo- attached to the vehicle thrust structure. The
nents housed individually in cylindrical low-pressure propellant supply duets must be
cartridges are in turn assembled in a sufficiently flexible to accommodate the gimbal
common body with suitable manifolding motions. It is noted that vehicle steering through
to the systems plumbing. gimbaling of a single engine or chamber is effec-
(4) Multiple-component packaging: Detail tive only for the pitch and yaw planes. For roll
parts for two or more components are control, at least two engines are required. For
assembled in a normal fashion in a com- vehicles with a duster of engines, therefore,
mon housing or body. this poses no difficulties. For single-engine
Figure 10-9 presents a typical pneumatic con- vehicles, special roll-control devices are needed.
trol package for a large liquid propellant rocket These may be small auxiliary nozzles, possibly
engine. This package combines two pressure- simultaneously used as vernier engines after
regulator assemblies, two relief valves, a series main-engine cutoff. The use of the turbine ex-
of solenoid valves, filter units, and check valves. haust for roll control has also been proposed.
It controls the flow of helium gas to various Whether the engine attaches to the vehicle
engine components. When engine start is initi- thrust structure by means of a thrust frame or a
ated, the helium control solenoid is energized gimbal bearing, either device must be designed
allowing helium to flow through the main pres- to be capable of transmitting the full thrust
ENGINE SYSTEMS DESIGN INTEGRATION 401
IGNITION PHASE
HELIUM SOLENOID CONT.
IN LET VALVE
VENT
CHECK VALVE
INSTAGE SOLENOID
CONTROL VALVE
HELIUM
CONTROL
SOLENOID
VA LVE
BLEED PRESSURE
REGULATOR G.G BLEED VALVE CONTROL SOLENOID VALVE
Figure lO-9.-Typical pneumatic control package design used in liquid propellant rocket engine
systems.
J
FUEL
work of this book, to describe the physical laws
and the general fundamentals of electrical cir-
cuitry. They are covered abundantly in the
_CLOSURE COVER literature. Moreover, in contrast with most other
basic liquid engine subsystems, the rocket en-
gine designer will try to use commercially avail-
able "off the shelf" components for his electrical
system. However, other cognizant members of
the design team will have to provide the basic
circuit diagram (schematic) and other data in
support of the installation of the required elec-
trical components. Characteristically, the elec-
trical system of a rocket engine is one of the
"_'__THRUST CHA,_SE/_ EXIT
last subsystems to be "frozen" before produc-
CLOSURE COVE_ tion. This is because sequencing for start and
stop represents one of the major engine develop-
Figure 10-12.-Various protective closure covers
ment activities, often resulting in repeated modi-
for the engine shown in [igure 10-10.
fication of the electrical system as development
progresses. Emphasis is therefore placed on the
flexibilityof electrical design. More recently,
("breathing"). In this case, the closures may be
this process has been greatly aided by dynamic
equipped with desiccant filters to permit access
analyses (see sec. 10.2).
of dry air only. Some of the covers must be re-
moved for installation of engine to vehicle.
Others will be left in place until the engine is
Electrical Schematic
actually operated. These must be readily acces-
sible and clearly marked, such as with bright The complete electrical schematic of a typi-
colors, to prevent their being left in place inad- cal liquid rocket engine system, including its
vertently. Figure 10-12 shows the location of ground- and vehicle-based elements, fills a
various protective closures for the Rocketdyne sizable drawing. With the aid of figure i0-13,
LR79-NA- 11 engine. which presents a portion of an earlier engine
static-firingschematic, the basic features are
discussed as follows. In ordinary wiring dia-
10.6 ELECTRICAL SYSTEM
grams, such as that of a radio receiver, all the
All rocket engines depend on some type of contacts of, for instance, a multiple switch or a
electrical system for their operation. This is tube are drawn to appear in the same location, as
true for solid systems, where at least ignition is they do in reality. This requires numerous wire
initiated electrically, as well as for liquid sys- crossovers in the diagram. The number of cross-
tems, in which the electrical system assumes overs would become prohibitive in a typical
numerous additional tasks. As with any common engine electrical schematic and may lead to
household device, electrical circuits in rocket confusion and errors. For rocket engine pur-
engines have caused troubles, due to poor de- poses it has long been found preferable to draw
sign, misapplication, abuse, poor maintenance, the basic diagram so as to show each circuit
human errors, and wear. Properly applied, how- separately. In this "functional flow diagram,"
ever, electrical circuits can" substantially sim- the various contacts of a relay, for instance,
plify the operation of a rocket engine, and will appear in different places, and often away from
increase its usefulness and reliability. There the circuit for the corresponding relay coil. The
4O4 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
drawing system also will materially simplify K28C, "Hypergol Cartridge Installed," as well
later troubleshooting. The diagram, of which as the normally closed contacts of cutoff relay
figure 10-13 shows a portion, was drawn sequen- KglC. If all contacts are properly closed, the
tially from left to right; i.e., circuits which are "C" contacts of K34C will cause signal light
energized during test preparation and start are DS34C to light up. Following selection of the
shown in the left portion, while those associated ig-nition power source by means of switch S16C,
with the cutoff sequence appear on the right. ignition can now be initiated by means of push-
In the schematic, connectors are shown as button $51C, since the "D" contacts of K34C are
continuous double lines, or portions thereof, now closed, and provided ignition disconnect
running horizontally through the diagram (J16, timer K54C has not picked up (TDPU e time de-
P16-- receptacle 16 and plug 16, etc.). layed pickup : 0.1 seo). In the diagram, several
Each of the contacts is called out by a letter circuit elements appear which are part of other
(T, G, K, etc.). All wires are numbered, as in- circuits not discussed. Note that in places two
dicated. Power buses, like connectors, are relays are used in parallel (e.g., K16C), if the
shown as horizontal lines, or portions thereof number of contacts required is too large for one
(heavy, single line = positive buses, usually relay. The numbers shown in hexagonal frames
shown near the top of the drawing; and double refer to the channels of an inking sequence strip
lines = negative or ground bus). The meaning of chart recorder or equivalent instrument. A spe-
the remaining symbols becomes clear by follow- cial test bus K615 is provided which when ener-
ing the circuit at the left of figure 10-13. Plug gized makes all signal lights go on and thus
P5 is shown connected to the main power bus permits spotting burnt-out bulbs.
K101 at terminal TB1-8. If certain facility sig- In earlier engine designs, many of the ele-
nal contacts are properly closed, such as those ments shown in figure 10-13 were installed in an
verifying "'Cooling water OK," "Firex armed," engine-mounted relay box. The trend has been
"Observer on Station," and many more, power to place as many parts of the electrical system
returns through plug P5, contact "Z," and is on ground as possible. This is easier with first
applied to relay coil K31C. The "B" contact of stages, which start while still connected to
this relay closes a circuit to lamp DS31C, which ground, or even held down mechanically until
lights up. The "A" contacts of the same relay released, for brief periods following start, than
are in the chain to the coil of K34C, together it is with upper stages which must start and
with the normally open contacts of several other stop, and sometimes restart, some time after
relays, such as K29C, "Heater Power On," and. takeoff. Beca_se of individual approach and of
ENGINE SYSTEMS DESIGN INTEGRATION 40,5
preferences for the types of interlocks, safe- must be at a lower potential than the emitter.
guards required, and type of component used, This is accomplished by inserting a bias volt-
two designers of a comparable engine may arrive age, V_, in the emitter leg, which raises the
at substantially different electrical diagrams. emitter potential V 1 volts above ground. Again,
Specifically with respect to the number of inter- as long as no appreciable current flows through
locks and monitoring circuits applied, caution is R I and R 2, the base will essentially be at ground
advised, since these circuit elements by them- potential, and Q1 will remain off.
selves are subject to malfunction and may do When input voltage Vin is applied, current Ix
more harm than good. will flow. This raises the Q, base potential,
The diagram discussed above employed re- until the combined Q, base-emitter voltage,
lays, well-developed types of which continue to (Vbel), plus bias voltage (V_), is overcome. At
be used in several of today's rocket engines. In this point, current will flow into the base (1b,)
others, solid-state (transistorized) switches are causing the transistor to turn on. This occurs
being applied which fulfill a similar function. when
Solid-state switches have the advantage of re-
R3(Icbol)>(Vbel +VI) max (10-27)
quiring no moving parts and thus are much less
sensitive to vibration effecLs. A circuit of this
When Q, turns on, resistor R 4 is switched to
type is shown in figure 10-14. It functions as
follows: ground, and current Ic, begins to flow. As Ic
increases, the potential at the base of Q2 is
Transistor Q2 is held in a "turned off" mode
lowered until it reaches the combined potential
by maintaining the base at a higher voltage
of V 2 and the Q2 base-emitter voltage, Vbe2. At
potential than the emitter. This is achieved by
this point, current Ib2 flows out of the base of
inserting a bias voltage, V 2. As can be seen,
Q2 and the transistor "turns on," thereby supply-
as long as there is no appreciable current flow
ing current to the load. This occurs when
in R,, the base potential will be essentially Vo,
while that at the emitter is Vo-V 2. Thus, Q2
Rs(Icbol+lcbo2)>(V2+Vbe2) max (10-28)
will remain off.
Transistor QI is held in a "turned off" mode
The switch is turned off by either removing
in a similar fashion, except that here the base
the input signal, Vin , or by using another static
switch to ground the base of Q_.
Switch lock-in (to maintain output after input
, _ vo
signal is removed) is accomplished by feeding
V2
Electrical Components
module be[ore and after potting. ture sensors, spark plug monitors, voltage sen-
sors, timers, and other devices, they form an
engine-contained logic which will execute a
P51 in figure 10-16, with the aid of an auxiliary sometimes elaborate starting sequence in re-
relay, could be used to assure engine readiness sponse to only two external signals: start and
in the chain leading to the coil of relay K34C in stop. In practice many more signals are ex-
figure 10-13. changed between engine and vehicle and/or
R * ( GIN[ ST*AT
•5, I <D
. I s c
I' _
r,_;'(_ ........
, _ e÷ CLOS*J_I[ • _ _Z{)$-II,,.,
3 r •
I_NITION
L. ................. J
ground control center, for checkout, monitoring, houses 4 exciter coils for the spark plugs of
confirmation, instrumentation, telemetry and thrust chamber and gas generator.
emergency reaction systems. Also, power supply
connections are required. Position Indicators
The relays and switches, together with sup- To verify the position of a valve, or the proper
porting circuitry elements such as resistors, installation of an igniter and similar mechanical
capacitors, diodes, terminal strips, and connect- conditions, position indicators are used. Most
ing wiring, are best housed in a common box. common types are:
This sequence controller is hermetically sealed Switches.-These are mainly used for "black
and often includes temperature-conditioning pro- and white" indications, such as "open" or
visions by means of electrical heaters (typical "closed," "installed," "connection OK," etc.
power requirement: 200 watts) or inert gas Numerous commercial products are qualified for
purges. The engine systems designer very likely rocket engine application and are available as
will receive from the cognizant department merely compact, miniaturized, sealed units. They are
a "black box" description of the sequence con- part of the individual component designs and are
troller, giving external and mounting bracket described in other chapters. It will be the burden
dimensions, connector descriptions, and installa- of the engine systems designer, however, to as-
tion specifications, with all internal detail sure uniform standards, specifically for connec-
omitted. Because of miniaturization, relay boxes tors. Position indicators of the switch type are
or sequence controllers have become very com- instrumental for sequencing the start and stop of
pact units. Figure 10-17 shows a typical can. a liquid rocket engine.
Note brackets for engine attachment. In addition Continuous indicators.-To know any inter-
to approximately 30 modules, numerous diodes, mediate position of a component, such as the
resistors, timers, etc., the container shown
angular position of a valve gate or the linear
displacement of an actuator, continuously read-
ing devices are required. Most widely used
types are potentiometers and variable reluctance
pickups. Potentiometers require adc power
source (typical: 5 volts), while inductive-type
sensors require an ac power source (such as 400
cps, 28 volts). The output of both, often in com-
bination with a bridge circuit, is fed to a telem-
etry and/or ground recording system. Indicators
of the continuous type are predominantly used
for instrumentation. It should be a goal to de-
crease their number as the development of the
engine progresses.
Timers
During engine start and stop, timing devices
are required for two principal reasons: correct
sequencing of valve actuations and other events,
and for monitoring the correctness of critical
sequence times. In the latter case, they will
inhibit progression of the sequence or initiate
cutoff in case of malfunctions. A degree of tim-
ing can be accomplished by means of orifices
and by sizing of volumes in the pneumatic or
hydraulic activation system. All other timing
Figure lO-17.-Liquid rocket engine sequence must be provided by timers as part of the elec-
controller. trical system. For engine systems starting on
408 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
the ground, these timers are mostly located in main battery. Because of the relatively high
the ground control system. Several types of power requirements for heaters airborne heating
timers are commercially available for this appli- should be kept to a minimum. This can be ac-
cation. They may be motor driven or may use complished by judicious placement of compo-
dashpots, springs, or other delaying devices for nents away from areas of extreme cold, and by
their function, usually with the capability for insulation and isolation. For instance, a hydrau-
external adjustment within a specified range. If lic pump attached to the auxiliary drive of a LOX
ground mounted, the timers' weight and size are pump, or the actuator of a cryogenic valve (see
of minor importance. fig. 10-18, near center) may require drastically
For systems starting at altitude, all timers reduced heating, or none at all, if a heat barrier
required must be engine mounted (or at least (gasketlike wafer of suitable material) is placed
stage provided). Here, weight and size do be- between connecting flanges. Furthermore, in the
come important. Modifications of the solid-state hydraulic system of an engine, heating require-
switches described above, through addition of ments may be completely avoided if the hydraulic
suitable capacitance/resistance circuits, have fluid is continuously circulated during standby.
been successfully applied. As a rule, these For cases where electrical heaters must be
timers are not adjustable once assembled, but applied, a number of types are available. A
require replacement of the entire module in case common one is the blanket heater, which is ap-
of timing changes. plied externally as a sleeve or cover, formfitted
for best efficiency, and equipped with a ther-
mostat.
Heaters
In other applications, an immersed heater,
Ideally, all components of an engine system
i.e., a Calrod-type heating element cast or em-
should be capable of operating reliably and un-
bedded into the component metal, is used, for
aided at all temperatures occurring during opera-
instance, to protect a cryogenic pump bearing.
tion. In many applications, however, extreme
temperature variations occur, due to the environ-
ment at high altitudes, or within an engine sys-
tem, particularly within those using cryogenics.
The development of components able to with-
stand these extremes without assistance would
often be very costly or altogether impossible.
Thus it is much simpler to reduce or eliminate
the temperature extremes by suitable means. For
some areas, mere insulation may be adequate. j
Others require heating (external means for cool-
ing are rarely required for rocket engines, but are
used for other vehicle systems). Bulk tempera-
ture conditioning is usually provided on static !II11
firing stands and within vehicles during standby
by blowing preheated inert gases or air into the
engine compartment. However, certain strategic r-::. -_- _:_ ._: __'_.:_;.:,_ig
Power Sources; Batteries defines the branch points and their rela-
tive distance from the main plug or
For its operation, a rocket engine electrical
receptacle.
system requires power which is almost always
It is usually difficult to determine the exact
supplied by the vehicle or from the ground. The
engine designer need not concern himself with length of individual wires and of some of the
branches on the drawing board. It is customary
the power supply, except for the specification of
to finalize these dimensions on a mockup engine.
requirements and provisions for connection.
For this reason, certain dimensions in the physi-
Frequently used voltages are:
cal routing diagram are left blank and the total
28-volt dc for heaters, control solenoids,
cable length is specified "as required."
relays, switches, igniter spark exciters,
For proper design of a rocket engine wire
certain instrumentation (typical peakload:
harness, the following must be considered:
2000 watts per engine)
5-volt dc for instrumentation (potentiometers) Routing (minimum distance, weight, electrical
resistancel interference)
8-volt dc for spark monitors (see ch. IV, spark
Secure clamping (safety, avoidance of wire
plugs)
chafing)
ll5-volt, 400-cps ac for instrumentation
Selection of attachment points (avoidance of
The required power is supplied directly from
special brackets)
batteries for dc needs, or through converters for
Adequate support (harness weight, vehicle
ac consumptions.
acceleration)
Miscellaneous Components Moisture protection (potting, sheathing)
Heat protection (routing, wrapping)
To complete an engine electrical system, a
Arcing protection (component selection, dry-
number of standard or special components are
heSS)
required: resistors, capacitors, diodes, terminal
Flexibility (installation, stiffness versus
strips, connectors, and receptacles. All are
gimbal load)
commercially available, qualified for rocket
Connectors must have means to secure them
engine use.
to prevent accidental disconnect. This has been
Connecting Wiring done with the connectors shown in figure 10-18
The various subsystems of the engine elec- by means of threaded sleeves, further secured by
trical system must be connected by suitable wir- safety wires. To prevent incorrect connections
ing, which for proper mounting and clean routing and/or damage to the connector pins, "clocking"
is combined into one or several wire harnesses. by means of key and slot, of connectors and
Figure 10-18 shows portion of a harness in- receptacles, is highly recommended.
stalled on a liquid rocket engine. For manufac- In addition to the cabling required to inter-
ture, the engine systems designer must supply connect the various engine electrical subsys-
certain information. This is best combined into tems, other cables are required to link these
a single drawing, which contains the following systems to the vehicle systems. This includes
information (fig. 10-19): wiring for power supply, controls (start and stop,
(1) Wire list.-This list calls out each harness PU, throttling, etc.), instrumentation and check-
wire by number and lists its length. It out circuits. To facilitate engine installation
also specifies the wire routing between and line connections, these wires are combined
plugs and/or receptacles. into trunks, each terminating in connectors which
(2) Harness schematic.-For clarity, the sche- must have, at the proper location, a mating coun-
matic repeats in pictorial form most of terpart on the vehicle. For the stage, a "J-Box"
the information supplied in the wire list, (junction box) is recommended, into which are
together with additional information. It plugged both the engine and the stage systems
also calls out the total length of wire connectors. On the inside, the box permits easy
required for material procurement pur- and environment protected redistribution of
poses. incoming and outgoing wires to assigned con-
(3) Physical routing diagram.-This diagram nectors. Figure 10-20 shows a typical block
410 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
WIRE LIST*
__PlO
WIRENO 1LENGTH I FROM I TO
EIOBtB{RD)
EIOAI8 12OO JIIO-A PIO-A
EIOBIe I 1200 I JIIO-B I PiO-e
EIIAI6 19SO I JIIO-C PII-A
i IOAIB(WH]
EIIBIe l 19.50 JIIO-D PII-B
El2 A H_I ! 12,50 JIIO-E PIZ-A
: EIZBI8 ' 2,50 J O-F PI2-B
_- EIIBI8tRD)
EI3AI8 12.00 JIIO-G PI3-A
JIlO
EI3BIS 12.OO J IIO-H PI3-B
EI4A 18 23OO JIIO-J P}4-A
EI4BI8 23.00 g IIO-K PI4-B
-'_ EIOAtBtWH) -_
EIOBIB(RD) -_ EISAle 23,75 JIIO-L PIS-A
EISBI8 23,75 Jll0-M PIS-B
" F'I6AI8 2775 JIIO-N ] PI6-A
_1_ El II AIB(wH) "_
El BI8(RD) "_
EIYAIB 16.00 JIIO-R JIII-A
• EI6BI8 27.75 JIIO-P t PI6-B
EIYBI8 I6.O0 JllO-$ JHI-B
EI4BI8(RD} •
EI5BIB(RD) J
;_____ j/IEIyBIB_RD)
PHYSICA-ROUTING DIAGRAM
-_ RECEPTACLE ._PI6
::::::((:;:'-_--J RECEPTACLE
[
GSE STATIC / VEHICLE
TO G ROUNO
CONTAJNE!R
(BATTERIES AND UMBIUCAL
ELECTRICAL 1 CONNECTIONS PATCH pANEL I VEHELE 1
CHECKOUT [ INSTRUMENTATION CONNECT{ONS
CONTROL DEVICES
I
VEHICLE MOUNTE_ JUNCTION BOX
(ONE PER ENGINE]
[] [] ]
MEASUREMENTS
!li
c_ < c_ < (SENSORs)
prior to transmission. In other cases, it may be TABLE lO-2.-Basic Liquid Propellant Rocket
desirable to convert a digital signal into an Engine Measurement List
analog one, prior to transmission (tachometers, (A-1 Stage Engine)
flowmeters). Small though the electronic (solid
Parameter Range
state) signal condition packages may be, room in
Thrust ......................... 0-1 x 106 Ibs
a compatible environment must be provided for Thrust chamberpressure
them on the engine. Under adverse conditions, (injector end) .................. 0--1200 psig
it may be necessary to mount the conditioners on Fuel pump inlet pressure ......... 0-50 psig
the vehicle near the engine. Oxidizer pump inlet pressure ...... 0-75 psig
Some transducers generate their own electric Fuel pump outlet pressure ........ 0-2000 psig
Oxidizer pump outlet pressure ..... 0-2000 psig
signal, such as the thermoelements, piezoelec-
Fuel injection pressure .......... 0-1500 psig
tric elements, and many of the flowmeters and Oxidizer injection pressure ....... 0-1500 psig
tachometers. Most others, however, require an Heat exchanger outlet pressure .... 0-1000 psig
external electrical input which they modify as a Gas generator chamber pressure... 0-1000 psig
function of the quantity sensed. Potentiometers, Turbine inlet pressure ........... 0-800 psig
Turbine discharge pressure ....... 0--30 psig
capacitors, inductances, and resistance tempera-
Stored gas container pressure ..... 0-3500 psig
ture bulbs are examples. Resistance types can
Fuel pump inlet temperature ...... Ambient-100 ° F
use ac or de; capacitance and inductance types Oxidizer pump inlet temperature... Ambient to -300 _ F
require ac. Typical instrumentation power sup- Cooling jacket inlet temperature... Ambient-100 ° F
plies are 88 volts and 5 volts de, and 115 volts Cooling jacket outlet temperature.. Ambient-150 ° F
Turbine inlet temperature......... Ambient-l,500 ° F
ac, 400 cps. Certain transducers with built-in
Stored gas container temperature .. -I00°-+100 ° F
electronics receive 28 volts de, chop, use and Fuel flow ...................... 0-10000 gpm
rectify it, and then return 5 volts (maximum) Oxidizer flow ................... 0-15000 gpm
signals (de-to-de transducers). Strain gages and Turbopump speed ................ 0-10000 rpm
variable reluctance gages are typical examples. Electric bus ,,Jltage ............. 20--35 volts dc
Spark igniter OK ................ On/off
Table 10-2 presents the basic static-firing
Miscellaneous valve positions .... On/off
measurement list for an engine of an A-1 stage.
In practice, particularly during early develop-
ment, this list may be expected to be augmented
by special measurements, requiring additional
thermocouple well proposed by B. N. Bose (ISA
instrumentation such as flowmeters, temperature
Journal. Sept. 1962) designed to offset this
gages, high-accuracy pressure gages, vibration effect.
pickups, and strain gages.
Carbon deposits.-In systems using kerosene-
based fuels, carbon deposits may cause erratic
Instrumentation Installation thermoelectric and grounding effects. While
Some basic design considerations for correct certain electrical connections are possible to
installation of instrumentation end organs (pick- minimize this, regular cleaning and sometimes
ups, sensors) will be briefly discussed. replacement of the probe is required. The engine
designer must consider this need for easy acces-
1. Temperature Sensors sibility of the instruments.
The most widely used temperature sensors are Installation of resistance thermometers or
thermocouples. In rocket engine application, "bulbs" is governed by similar consideration.
thermocouples are subject to certain measure-
ment errors. Among these are: 2. Pressure Gages
Velocity errors.-Fluid friction may cause an Two principal types of pressure measurements
excessive temperature reading, particularly at are usually conducted during rocket engine test-
high fluid velocities. Suitable protection which ing: static and dynamic. As a rule static meas-
does not affect the true measurement may have urements are employed to record, with high accu-
to be provided. racy, steady-state conditions or parameters
Conductive errors.-A heat leak may exist varying only very slowly. Dynamic measure-
from probe to support. Figure 10-21 shows a ments are used where rapidly changing conditions
ENGINE SYSTEMS DESIGN INTEGRATION 413
Flowmeters
Turbine-type flowmeters appear best suited to
rocket engine installations. They are well de-
Figure lO-23.-Typical installation of a thrust veloped and, as a rule, can be mounted in a sim-
chamber pressure transducer. ple fashionbetween availableflangesin existing
F 3 I I1 !l
L/.---DIAPHRAGM
BOSS_
ADAPTER -_/
_.--ENGINE
COMPONENT
_--- DIAPHRAGM
ENGINE COMPONENT WAll
WALL
JL-s _ ;
fn =
2"/r
WHERE :
WHERE: C = ACOUSTIC VELOCITY
n -" ANY WHOLE NUMBER (I,2,5...etc)
S = CROSS-SECTIONAL AREA OF TUBE IN SQUARE IN.
n = I FOR THE FUNDAMENTAL FREQUENCY
V = VOLUME OF CAVITY AT THE END OF THE TUBE
C = ACOUSTIC VELOCITY OFTHE MEDIUM IN CUBED IN.
03,550 IN./SEC FOR AIR AT 68 ° F}
IJ= EFFECTIVE LENGTH OFTUBE IN INCHES
L= LENGTH QFTUBE iN INCHES
I_= L +l.7a
WHERE: L: ACTUAL LENGTH
o=RADIUS OF TUBE
(A) {B)
"
416 DESIGNOF LIQUID PROPELLANT ROCKET ENGINES
Still another approach for ballistic missiles zation of thrust and total impulse (propellant
is the combination (clustering) of a pair of booster load).
engines with a large-expansion-area sustainer To examine some of the considerations which
engine. During flight, the booster engines may govern engine cluster arrangements, we assume
be jettisoned following the boost period, while that payload and mission of a multistage vehicle
the sustainer continues, fed from the same tanks. are defined and that stage propellant loads (total
These configurations are also referred to as impulse) and tank geometry (diameter) have been
"One-and-one-half- stage vehicles." optimized. We further assume that engines are
available and will not be redesigned, and that
the optimum thrust for a typical stage suggests a
Recent Cluster Design Trends
range of from four to six engines. An analysis
The first clustered engine vehicle, signifying has shown that payload performance probably
the modern trend toward multipurpose engines, is increases with the larger number of engines (fig.
NASA's S-1 first-stage booster for the Saturn I. 10-28); however, there are other considerations.
It is powered by 8 LOX/RP-1 engines, the basic One of the most influential considerations is
elements of which were transplanted with rela- "Engine Out Capability" for increased vehicle
tively minor modifications from earlier engines. reliability (also see ch. II, "Thrust Level").
The S-1 is noteworthy for its tank arrangement. Specific engine provisions for engine-out (EO)
Eight tanks from an earlier missile are clustered are summarized in section 10.9. Capability of a
around a central larger tank. This "multicellular" vehicle to complete its mission with one engine
design, which in the case of the S-1 permitted inoperative always entails some performance
early availability of large tank capacity without losses. However, these losses decrease with
major retooling, has been recommended by some increasing total number of engines, because less
vehicle designers for still larger vehicles. The performance reserve is needed, and because less
eight engines of the S-1 are grouped into four correcting thrust vector trim is required from the
fixed inner and four gimbaled outer engines. operating engines. Trim from engine-out is
Except for the outer-engine actuators, all eight further affected by the diameter of the engine-
are identical, independent units, built and tested mounting circle. The contribution to mission
singly, and combined for the first time on the reliability by EO capability is substantial, with
vehicle proper. cluster failure potential reduced by more than
In the S-1, the number of engines was, no half. In addition, absolute reliability values are
doubt, almost entirely governed by the availa- a function of the number of engines in the cluster.
bility of existing major engine components and This is most evident with no engine-out.
tankage. For subsequent vehicles, such as the Figure 10-29 illustrates this at two points of
Saturn V three-stage vehicle, greater freedom time of engine overall development status. As
existed, particularly with respect to the optimi- the latter progresses, the difference between
ENGINE SYSTEMSDESIGN INTEGRATION 417
F-
08
YAW -- --0_
ACTUAT
I
a. b.
E I i
4 fi 6
NUMBER OF ENGINES
Adequate mechanical design data, vital for When installing an engine in a vehicle, it is
the physical integration of the engine into a not only necessary that it can be properly bolted
vehicle system, must be properly documented by to the vehicle thrust mount, but it is equally
the engine systems designer. The following data important that no other vehicle parts interfere.
are considered minimum requirements: Space in the vehicle engine compartment must
(1) Engine system general arrangement draw- not only accommodate the engine envelope when
ing.-This drawing defines the engine in the neutral position, i.e., pointing straight aft,
space envelope and the locations and but when fully deflected in all directions as
detail of various agreed-upon vehicle well. Typical maximum engine deflections range
connect points, such as from 4 ° to 7 ° . For clusters, moreover, the possi-
Gimbal actuator attach points taken into account. Since space in the engine
Fuel and oxidizer inlet flanges compartment of vehicles is usually limited, the
Hydraulic and pneumatic system con- need for the closest cooperation between engine
nections and vehicle designer becomes apparent. This is
Electrical and instrumentation con- especially true for upper stage clusters where
nections large expansion area nozzles must be housed in
(2) Mechanical dimensions, tolerances, seals minimum interstage structures.
TEST STAND
,ul
STRUCTURE TA,
ENGINE CONNECT
Lax IEL_O_--
PANEL .o= pu_ -- - c, _, _,.o_
Lox ,_c,c_,-
manifested themselves as longitudinal vehicle These systems have been in use since the
oscillations (also referred to as "Pogo Stick early beginnings of liqmd rocket engine applica-
Effect," "Accordion Motion," or "Spring-Mass tion. They are specially important during engine
Effect") at frequencies from approximately 10 to
20 cps. They have led, in at least one instance,
to vehicle destruction. The relationship of the BURNOUT
<..j..1
contributing parameters is illustrated in figure
10-33 for a typical single-engine vehicle. The
analyses and correction of these oscillations are
extremely complex, particularly since they can-
not be reliably reproduced during captive firings.
FLIGHT TIME
It is believed, however, that any one of the con-
tributing factors can modulate thrust at a fre- Figure IO-34.-Typical pump inle$ pressure vari-
quency at which the vehicle is resonant at cer- ation of a vehicle affected by longitudinal
tain tank levels (flight time; fig. 10-34). Through oscilla$ions.
ENGINE SYSTEMS DESIGN INTEGRATION 421
the others.
Disconnection of electric power to the defunct
engine only. SECTION
A-A
Resetting or disarming of backup cutoff tim-
FULl. CLOSE-OUT PARTIAL CLOSE-OUT
ers, since the reduced number of engines (CURTAINS OPTIONAL)
80
80
4C _ _- SKIN BELOW THRUST MOUNT /--SKIN BELOW THRUST MOUNT
_"
-40
(_ _ /_-SKiN ABOVE THRUST MOUNT
g o \
_:_ j_,,_.NITROGEN AMBIENT E'XTERN_L TO MOUNT
-40
_.-ao
I'-
-120
/_ N_TROGEN AMBIENT EXTERNAL TO MOUNT I.U_''1l_1 _%,_ _--ENGINE TEMPERATURE
F- -80
i(
-200 _ I_THRUST MOUNT TEMPERATURE _NITROGEN AMBENT INTERNAL TO MOUNT
HOLD TIME FROM PROPELLANT LOADING(HRS) HOLD TIME FROM PROPELLANT LOAI_NG (HOURS)
Figure 10-37.-TypicaI interstage temperature en- Figure 10-38.-Typical interstage environment for
vironment for an uppei space vehicle stage an upper space vehicle stage using cryogenic
using cryogenic propellants. 565 SCFM, propellants. 4170 SCFM, (260 ° F) nitrogen
(-I00 ° F) nitrogen purge. purge.
additional advantage of counteracting the accu- be expected to maintain their temperature within
mulation of combustible gas mixtures from minor an allowable band after vehicle liftoff during the
leaks. However, the vehicle builder's possibili- relatively short boost periods. Only stages with
ties of heating through use of compartment gas prolonged cruising times may require an onboard
purges are definitely limited. Some of the limita- purge supply system which imposes payload
tions stem from the engine designer's own speci- penalties.
fications, which require avoidance of elevated
temperatures around components containing cryo-
genic fluids to minimize boiloff and to prevent
formation of gas bubbles. Also, certain struc-
tural members must be kept below maximum tem- UPPER STAGE, LOWER TANK BOTTOM
Literally a few inches away from the compo- conditions, and remoteness from ground stations
nents requiring protection against low tempera- can only crudely be simulated in ground tests.
tures are others which must be protected against Therefore, until first flights of a new vehicle
warmup due to influx of heat. Here again non- have been accomplished, the engine in-flight
cryogenic propellants pose few problems. By start sequence, which is closely interwoven with
contrast, cryogenic systems, in particular those the stage separation sequence, must be planned
including liquid hydrogen, are very sensitive to based largely on analytical work. Neither ve-
heat influx. The effects on pump NPSH, which hicle builder nor engine designer can do this
includes the vapor pressure of the fluid pumped, without consideration of, and consultation with
were discussed in detail in section 8.5 in con- the other. The following principal relationships
nection with tank insulation. For engine start, must be considered:
which involves the cryogenic fluids stored in the Starting of the upper stage engines while the
ducts immediately upstream of the engine and lower stage is still at full-thrust level is diffi-
within the engine itself, insulation may become cult, mainly because of the problems of flame
prohibitively complex. Continuous removal of impingement and exhaust gas dissipation from
superheated cryogenic fluids from the engine the interstage. Therefore, engine start following
system and substitution of fresh liquid, there- stage separation is preferred. However, in this
fore, has been widely used. This can be accom- case propellant settling from acceleration effects
plished through continuous overboard bleeding or of the lower stage is no longer assured. Auxiliary
through recirculation back to the tank. In both propellant settling (or "ullage") rockets (usually
cases, the liquids are tapped off at a point far- solid propellant type) are required instead during
thest away from the pump inlets and routed to the engine starts. These add weight and drag and
stage for overboard dumping or return to the reduce payload. The size of the weight penalty,
tanks. A small auxiliary pump is often required in turn, depends on engine-start characteristics
to obtain the minimum flow rates required for (thrust-buildup time). For engines with relatively
adequate bleeding. protracted buildup times, advanced initiation
During engine development, and even during (with respect to lower stage thrust decay and
vehicle static firing programs, relatively favor- separation) of upper stage engine start may sub-
able prestart conditions exist: the enclosing stantially reduce the penalties. A secondlopped
interstage is not in place, hold periods are under off the ullage rocket burning times may save
better control, cold lower stage tank surfaces hundreds of pounds of stage weight. The same
are absent. There is thus the danger that what amount of time added may not only reduce pay-
worked faultlessly under these conditions will load weight but pose serious control problems to
fail in vehicle flight. A realistic recognition of the space vehicle which floats essentially "rud-
this situation and close coordination between derless" after separation and prior to the start of
the two design teams in the early phases of its own engines. From the foregoing the need
engine and vehicle design and development is for thorough mutual understanding of this impor-
therefore vital. tant vehicle-to-engine interface becomes apparent.
For first- stage applications, the engine builder
is relatively independent regarding start charac- Vibration Environment
teristics and sequence. The vehicle will not
take off until thrust exceeds vehicle weight, Even a normally or "smoothly" operating
gravity holds the propellants at the tank outlets, rocket engine generates a vibration pattern which
and all parameters are closely monitored from the is transmitted mechanically through the thrust
ground control center. If a holddown period prior mount, or acoustically through the ambient at-
to vehicle release and takeoff is included, an mosphere, if present, to the vehicle structure.
additional safety margin exists. Neglect of these vibrations by the vehicle builder
For upper stage engine start, conditions be- may lead to serious structural weakness or mal-
come considerably more complicated. Absence function of vital control organs mounted to the
of the propellant-settling effects of gravitation, vehicle structure. Means of reducing the normal
the surrounding vacuum, altitude temperature engine vibrations are limited and hard to analyze.
ENGINE SYSTEMS DESIGN INTEGRATION 425
Rather, itis important that the engine developer interstage length will cost approximately 40
establish as accurately as he can the existing pounds. Thus, here too, the added interstage
acoustic and vibration pattern and inform the weight required to accommodate a longer, better
vehicle designer of his findings. Since the tie- performing nozzle may void the theoretical gains.
down firing measurements cannot reliably predict
free-flightconditions, a serious handicap exists. Stage Separation
However, through his thorough understanding of When separating the stages of a space vehi-
the potential problems, and through specification cle, three basic possibilities exist:
in cooperation with the vehicle designer of static 1. Leaving the interstage with the lower
test and flight measurements, the engine designer stage.-This requires the engines to pull out of
can materially assist in reducing the incidence the interstage cylinder during separation (fig.
I 1
Upper stage ullage rockets ignited
I I
Engine start sequence initiated
I
Separation timer started
1
Lower stage retro rockets f_e
• Upper
I
stage engine nozzles clear interstate
I
• Tuner allows engines to gimbal
• Engines i_nlte
• Engines reach
full thrust
t
Tm_e (Seconds)
greatly affect the separation sequence and the stands and for various servicing functions, the
clearances. To optimize these, the vehicle engine builder requires numerous fixtures com-
builder needs from the engine designer accurate monly referred to as engine GSE (ground support
information on thrust-decay characteristics and equipment). Some of these fixtures can be used
tolerances. This may be in the form of a graph, by the vehicle contractor as designed, others
as shown in figure 10-41. could be used if only a few minor changes or
additions were made. The funding required for
GSE in a typical vehicle program is substantial
Flame Deflector Size
and approximates that of the flight hardware.
Although this is not entirely an engine-to- Common use of as many of these fixtures as pos-
vehicle interface, it is pointed out that engine- sible by engine and vehicle builders is strongly
nozzle size and arrangement on the vehicle, advised. To assure this the two must work to-
together with maximum gimbal angles, strongly gether from the earliest design inceptions.
influence the design of required flame deflectors Standardization
on engine and stage static firing test stands, and
of launcher flame deflectors. Here the problem The task of mating engine and vehicle will be
is not so much one of optimizing vehicle per- substantially facilitated if the designers of both
formance, but of minimizing cost and of assuring work to the same standards. This applies to
the adequacy of these indispensable development national standards, Government standards, com-
tools. pany standards, to terminology, mathematical
symbols, and to the measuring system (metric or
English, decimals or fractions, tolerances).
Engine Handling, Installation, and Servicing
Since several avenues are open, it is necessary,
Fixtures
at the very outset, to agree on which one shall
For handling engine parts and assemblies be pursued. This requires close cooperation and
during engine installation into development test full documentation.
ENGINE SYSTEMS DESIGN INTEGRATION 427
too
\
\
\\ \
\ \
\ \
7s \
\ \
\ \
ILl \
I&.
so
o [ t I
0.S 0.6 0.7
SIGNAL
For the vehicle builder to deliver his stage Until reliable and economic methods are de-
on time, complete and equipped with engines, the veloped to recover and refurbish rocket vehicles,
latter must be available to him with sufficient rocket engines will be used only once in flight,
leadtime. If the engine builder is directly con- which will be preceded by a reasonable number
tracted by the vehicle contractor, this can be of checkout and acceptance firings. The engine
negotiated and followed up in a straightforward model specification will include the total pre-
manner. In most cases, however, particularly scribed run capability (lifetime) of the engine,
with large liquid engines, these will be supplied usually expressed in multiples of the rated flight
to the vehicle contractor as Government-furnished duration. A factor of, for instance, 6, which
equipment (GFE). The engine supplier, there- includes a reserve for repetition of checkout
fore, may not necessarily be fully cognizant of runs, is typical. Because of the stringent weight
the detail vehicle needs. In either case, the considerations in rocket vehicles, engine design
engine designer and developer, to avoid sudden attempts to assure this life expectancy, but no
unexpected compression of his schedules, must more. Although reasonable safety factors make
make sure that his schedules for design, drawing it most unlikely that the engine would completely
release, manufacture, development, and engine collapse shortly after the allowable maximum run
delivery are in accord with the vehicle needs. time is exceeded, the statistical probability of
428 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
some engine part failing does increase. For this ing blocks, such as turbopumps, injectors, com-
reason most engine specifications prescribe bustion chambers, gas generators, valves, gimbal
overhauls of a stated scope whenever this limit blocks, and high-pressure ducts. Peripheral
is exceeded. equipment, such as instrumentation lines, servic-
Similarly, upper limits are set for storage ing lines, inIet ducts, and wire connections,
times. If these are exceeded, routine replace- should be left flexible enough to adapt them to
ment of certain engine parts will be made, nota- each vehicle without compromise. It is tempting
bly of gaskets and seals. Rocket engine preser- to procuring agencies to warehouse just one
vation and packaging methods are well developed, engine model and ship "from the shelf _ to wher-
permitting storage periods from 3 to 5 years ever the need arises. This convenience, how-
without overhaul. ever, may cost substantially more than the ex-
During shipping, handling, installation, servic- pected savings: In mission compromises, reduced
ing, checkout, static firing, and launch prepara- reliability, increased coordination effort between
tion, damage of various types may be incurred by engine builder and vehicle builders, substantially
the engine because of parts failures, mishandling, increased possibility of oversights and communi-
oversights, or accidents. All incidents must be cation gaps, and cost- and time-consuming retro-
corrected by repair and/or parts replacements. fits. As always, a joint thorough and unbiased
Any or all of the described maintenance action analysis considering all aspects, including that
may be required while the engine is still with of the long-range future, will readily yield in-
the engine builder, or following _ts delivery to formation about the point of diminishing returns.
the vehicle contractor. An accurate and effec-
tive maintenance or logistics plan must be Reserves and Safety Margins
worked out between the two contractors, or with
their contracting agency, considering all han- In his negotiations with the engine user, i.e.,
dling needs at all test stations. This includes "the customer," which quite likely is a Govern-
consideration of under what conditions engine ment agency, the engine designer, like the vehicle
return to the factory for overhauls, repairs, and builder, will frequently find himself under pres-
parts replacements should be made. If return is sure to compromise. This may be to cut weight,
not required, the correct handling fixtures and to accelerate schedules, or to maintain relia-
tools must be provided for each location. Above bility with dwindling funds. It is then that he
all, an adequate stock of spare parts must be will be most in need of his top management's
planned. The engine designer will frequently be understanding and support. But it is here also
consulted for his advice in an effort to avoid where he will be most criticized if it is found
both time losses due to lack of parts and costly that his analyses were incomplete, superficial,
overstocking. not optimized, or heavily biased by safety factor
upon safety factor. To strike an optimum bal-
ance between high performance and adequate
Multiple-Engine Use
safety factors and reserves is one of the finest
In view of the cost of rocket engine design, arts of engineering and is directly translatable
development, and procurement, common usage of into the degree of success. Once the designer
a given engine for several vehicles, or for sev- is certain that he has achieved this balance, or
eral stages of the same vehicle, is very desir- if he is concerned that he may lose it, he should
able. Also logistics, handling, checkout, launch so go on record. If he does, he should also
preparation, and instrumentation requirements remember Edmund Berkeley's observation:
can be substantially reduced or simplified. Cau- "Thoughtful and tolerant disagreement is the
tion must be exercised however, not to go "over- finest climate for scientific progress. "_
board," lest engine-to-vehicle interfaces become
considerably more complicated.
Effort should be concentrated on the common _Edmund C, Berkeley: Preface to "Giant Brains,"
usage of the major cost items and engine build- John Wiley, 1949.
Chapter XI
Liquid propellant space engines embody the Spacecraft Main Propulsion Systems
same operating principles and general character- The main propulsion systems for most space-
istics as the liquid propellant rocket systems craft are pressure-feed, storable hypergolic pro-
previously discussed. However, their specific pellant systems, such as the A-4 stage propul-
missions for the use in spacecraft require special sion system, with thrust levels upto about 25 000
design considerations, which will now be dis- pounds. These systems include propellant tanks
cussed. and their pressurization system, control valves,
main thrust chamber assembly or assemblies,
and a gimbal mechanism or some other type of
II.I PRINCIPAL SPACE ENGINE
thrust vector control. Most main thrust chambers
APPLICATIONS
are ablatively cooled, while attached nozzle
By definition, space engines supply all extension skirts are radiation cooled.
those forms of rocket propulsionwhich a space- In a few cases, turbopump-feed systems have
craft requires for various maneuvers in space. been used, such as the 50 000-pound thrust (sea
This may include attitude control and stabiliza- level) turbopump-feed liquid oxygen/ammonia
tion, coplanar and interplanar orbit changes, rocket engine produced by the Reaction Motors
trajectory corrections, rendezvous maneuvers, Division of the Thiokol Corp. for the North
lunar and planetary landings and takeoffs, and American X-I5 research plane. As space mis-
retrofiring (reversed thrust for deceleration) dur- sions become more ambitious, requiring increased
ing reentry into the Earth's atmosphere. The total impulse and higher-energy cryogenic pro-
thrust of a space engine may be a fraction of a pellants, turbopump-feed systems will undoubtedly
pound or many thousands of pounds. Besides a play an increasingly important role in spacecraft
few solid propellant systems (used mainly for propulsion. Most space missions require multiple
single-start, retrofiring rockets) and stored gas starts and a certain degree of thrust throttling.
systems (used only in applications for thrust These requirements usually account for the prin-
levels less than 1 pound, and for a total impulse cipal differences between the main propulsion
of less than 5000 lb-sec), the great majority of systems for spacecraft and those for booster
the space engines are of the liquid propellant vehicles.
type. Because of their inherent operational
advantages, liquid systems most likely will con- Reaction Control System
tinue to dominate the space engine field, even These systems deviate more drastically from
when advanced nuclear and electrical propulsion other liquid propellant rockets, because of their
systems become available. design arrangement and their thrust levels, which
Liquid propellant space engine systems may run from 1 pound up to 500 pounds, with most
be divided into two basic groups: vehicle main systems probably failing within the 1- to 100-
propulsion systems and reaction control systems. pound range. The thrusts of these systems may
These differ not only in function and thrust level be used to provide attitude control to properly
but also in the type of propellant they use, the position a spacecraft, to aline a spacecraft for a
degree of required controllability and thrust midcourse corrective or terminal maneuver, and
variation, and system components. Common to to stabilize the vehicle after separation from
virtually all of them is the requirement that they another stage or during Earth reentry.
be able to start and operate reliably in the cold As a rule, attitude-control engines are used
vacuum conditions of space. in opposing pairs to produce pure couples about
429
x._. ii--
1
The reaction control engine systems are
usually pressure fed, using monopropellants or
storable hypergolic propellants. In all cases
thrust level and duration must be very closely
%t71
......
G[ RO_E1 $ :
basic engine parameters to determine their in- mass (limited by the payload capability of the
fluence on the spacecraft's ability to perform its launch vehicle) and required mission velocity
intended mission. Here, again, an overriding increment, the mission payload capability is a
consideration is reliability. Typical space mis- function of the main engine system specific im-
sions include the establishment of an orbit pulse and the spacecraft inert weight, as indi-
around a planet following a transfer from Earth cated by equation (2-1). Similarly, for a given
which may last several months or years. The spacecraft initial mass, inert weight, and pay-
ability of a propulsion system to coast in space load, the mission velocity increment capability
for these prolonged periods and then fire reliably is directly proportional to the main engine sys-
is typical for many contemplated missions. It is tem specific impulse. The specific impulse of a
one of the foremost considerations in the design reaction control engine system affects directly
of space engines. Based on mission, trajectory, its system package weight. A higher specific
and payload requirements, the following major impulse will result in a lower system weight for
design areas must be examined for the optimiza- a given total impulse requirement.
tion of one or a group of propulsion systems: 2. Operating temperature.-In an overall
(1) Propellant type vehicle optimization study, the criterion of spe-
(2) Vehicle system operational requirements cific impulse must not be used as the only hard-
(3) Mission environmental effects and-fast rule for propellant selection, because
Several different designs (assuming various the problem of maintaining the propellants at
propellants, feed systems, chamber cooling temperatures that permit effective use of the
methods, etc.) may be conducted simultaneously propulsion system after coast is an additional,
for a given space mission. These designs must major consideration in spacecraft design. The
be carried far enough to evaluate their relative requirement for thermal control of the propellants
merits such as reliability and performance, and during coast affects the selection of the propel-
to establish the advantage of one design over lants and the vehicle configuration. Heat trans-
the other. An optimum system design should fer between propellants, and between propellants
thus emerge. and other vehicle components, must be consid-
ered, as well as heat radiation out to space, in
Selection of Propellants
from the Sun, and between vehicle components.
Various aspects of liquid rocket propellants Studies have indicated the feasibility of insu-
have been discussed in section 1.4. Here we lating the propellants against excessive temper-
will discuss the characteristics which are perti- ature changes, even during long coast periods in
nent to their selection for space engine applica- deep space missions. However, for the cryo-
tion. The reliability and performance of a space- genic propellants more insulation weight is
craft will depend greatly on the propellants used. required than for Earth-storable propellants.
This is of paramount importance in establishing This difference becomes greater the longer the
the optimum design criteria and final operational mission coast periods.
capabilities of a vehicle. For any planned space 3. Density.-The bulk density of a propellant
mission, the projected program time periods, as combination has a pronounced effect on vehicle
related to the anticipated state of the art and payload. This can most easily be seen in the
operational reliability of the propellants to be case of the LF2/N2H _ combination, the specific
used, should be the first criterion for their selec- impulse of which is comparatively low, as com-
tion. Propellants with undesirable characteris- pared to LF2/LH 2 or LO2/LH 2. However, its
tics and those that have no advantages over bulk density advantage may result in a higher
similar, more desirable types should be elimi- ideal velocity increment for certain space
nated early in the studies. In addition, propel- missions.
lants are selected on the basis of the following: 4. Ignition cflaraczeristics.-Hypergolicity of
1. Speci[ic impulse.-The criterion of spe- the propellant combination used in space mis-
cific impulse is, perhaps, the most commonly sions is always considered very desirable to
used basis for comparison in nearly all propel- effect a simpler and more reliable engine sys-
lant evaluations. For a given initial spacecraft tem, particularly for multiple starts.
432 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
are frequently used in space engines. r WEEK _ kq:_TM $ blONT_tS :AR I0 YEA_
craft should use the same propellant combina- COAST I_UI_TIO_, _*OUl_
TABLE ll-1.-Comparison of Various Liquid Propellant Combinations [or Space Engine Applications
.,a======
DESIGN OF LIQUID PROPELLANT SPACE ENGINES 433
I. Engine system total impulse, design thrust tion. This may be expected, considering that
level and run time.-These are determined and continuous operation involves a minimum of con-
optimized by analyses of the mission trajectory, trol components and actuations. Multiple starts,
considering spacecraft operating limitations such on the other hand, require that each component
as acceleration loads. Typical liquid propellant responds perfectly each time it is called upon.
rocket engine total impulse, thrust, and run times In addition, multiple starts require much longer
for various space missions are presented in fig- periods of absolutely minimum propellant and
ure 11-3. Required mission total impulse and pressurant leakage. Therefore, more rigorous
engine thrust level influence, to a large extent, approaches toward improved reliability, such as
the choice between a pressure feed and a pump component or subsystem redundancies, must be
feed system. considered in the design.
2. Engine system operating characteristics.- 4. Engine system component design.-Besides
This includes engine thrust throttling range, mission environmental effects (to be discussed
thrust as a function of time, number of starts and below), the detail design of the engine system
repetition rate, cutoff impulse and accuracy, and components are directly affected by the space-
thrust vector control requirements. Most of these craft design configurations. For instance, the
characteristics are determined by various pro- location of the system (internal in or external to
jected spacecraft maneuvers. It may also be the spacecraft) determines the feasibility of
desirable for the same engine system to fulfill radiation cooling for the thrust chamber. Engine
more than one type of maneuver, or to be reused space envelope limitations affect the choice of
on subsequent missions. Typical thrust-time nozzle type, allowable expansion area ratio, or
histories for various spacecraft maneuvers are even the feed system type. (In some applica-
presented in figure 11-4. tions a pump feed system affords a much higher
3. Engine system design.-Experience has
shown that multiple-start operation as required
by most space engine systems has a more severe
effect on a rocket engine than has continuous
steady-state operation of comparable firing dura- TIME
CONSTANT THRUST, VARIABLE IMPULSE
(LUNAR LANDING)
10
0.i l
THRUST - THOUSAND
|
I0
LB";
100
FI/ T IME
(LUNAR TAKEOFF)
Figure 11-3.-Typical liquid propellant rocket Figure ll-4.-Typical liquid propellant rocket
engine total impulse, thrust, and run time for engine thrust-time histories for various space-
various space missions. (Coulbert, C.D., crait maneuvers. (Coulbert, C. D., "Selecting
"Selecting Cooling Techniques [or Liquid Cooling Techniques [or Liquid Rockets [or
Rockets lot Space Cra/t," J. Spacecra[t and Space CraIt," J. Spacecralt and Rockets 1,
Rockets I, 129-139, 1964.) 129-139, 1964. )
434 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
expansion area ratio for a given envelope.) The planetary emitted radiation. Direct solar radia-
structural design of the system components is tion is the largest external heat source. It is
influenced by the maximum vehicle acceleration therefore advantageous to orient the spacecraft
and by vibration loads. for maximum protection from direct solar radia-
tion to the vulnerable subsystems such as pro-
pellant tanks. The quantitative evaluation of
Mission Environmental Effects
thermal radiation in space indicates that damage
Afterthe propellantshave been selected and to some components may result from the exces-
the basic propulsionsystem design established, sive absorption of radiant energy, unless protec-
with preliminaryconsiderationsof mission envi- tive means are applied such as reflective sur-
ronmental effects,itis possible to examine the faces or coatings. A very real problem with the
various engine subsystems in detailto determine prediction of the thermal properties of materials
which components or phase of system operation in the vacuum is the lack of knowledge about the
is stillaffectedby the projectedmission envi- effect of the ever-present oxide-metal surface
ronment in space. The characteristics of this film on heat transfer mechanisms.
environment include the following: 3. Nuclear radiation.-The two most important
effects of nuclear radiation in space on metals
1. High vacuum.-The low ambient pressures
are the production of internal heat, and disloca-
experienced in space missions may cause vapor-
tion of the atoms within their crystalline struc-
izationof metals and partialor complete removal
ture. The nuclear particles of interest are fast
of filmor adsorbed gas layers at the surfaceof
neutrons, lower energy protons, alpha particles,
the material. Principal problems associated
electrons, and gamma rays. The fast neutrons
with high-vacuum conditions are:
present the most severe problem. It is known,
(a) Reduced ability of a component to perform
however, that they do not exist to a great extent
its function, due to mass loss through
in space. Generally, the total radiative flux
material vaporization or bulk property
from all nuclear radiations is not sufficient to
changes.
(b) Changes in the radiative heat transfer cause any appreciable damage to metals over a
period less than 2 years.
properties of the material, which could
increase, to a destructive level, the 4. Meteoric bombardment.-Spacecraft will be
exposed to the impact of meteoroids ranging in
bulk temperature of a component such as
a radiation-cooled nozzle skirt. size from microscopic dust particles to bodies of
asteroid dimensions with a wide distribution of
(c) Possibility of condensation of metallic
kinetic energies• Collision of the spacecraft
vapor on electrical components, result-
with these materials would result in surface
ing in short circuiting.
erosion,punctures,or totaldestructionof the
(d) Changes in tile fatigue, frictional, and
vehicle. The idealprotectionforpropulsion
creep properties of materials, as well as
system components would be to locatethem
the possibility of self-welding of metals.
entirelywithin the vehicle. This, however, is
Design remedies against high-vacuum effects
usually not practical for parts such as the thrust
include material selection and proper location of
the components within the vehicle to take advan- chamber nozzle skirt. Therefore, adequate con-
sideration must be given to the design of these
tage of nonvacuum environments.
2. Thermal environment.-The thermal envi- parts with respect to the cited effects.
ronment encountered in space missions, such as 5. EHects ol gravity.-The absence of accel-
in Earth orbit or during the transfer phase in eration forces, as well as the presence of large
interplanetary flights, is an important considera- acceleration forces, should cause no mechanical
tion in the design of space propulsion systems, design problems if proper provisions are made in
especially for the storage of the propellants. the design phase. The main problem will be
Three external sources of radiant energy must be with the propellant feed system. This can be
considered when evaluating the thermal environ- overcome by positive expulsion or by providing
ment of a spacecraft. These are direct solar, propellant settling rockets (which themselves
albedo (solar energy reflected from a planet), and would require a positive expulsion feed system).
DESIGN OF LIQUID PROPELLANT SPACE ENGINES 435
Such aeration requires a surprisingly small pro- propellants for deep throttling is provided by
portion of gas by weight. The alternative ap- separate helium lines and redundant valves. If
proach is to vary combustion chamber injection the system throttle requirements were removed,
area, as shown in figure 4-47(a). Other potential the main propellant throttling valves could be
space engine throttling methods include use of replaced by simple on-off types. Propellant
throttlable propellant gasifiers or precombustors, filters are provided in both main propellant lines
or the subdivision of the injector manifolds into to minimize system line contamination with
segments, each of which can be shut off by foreign particles.
simple on-off valves. The package design layout of the pressure-
Figure 11-5 presents the schematic of a typi- feed spacecraft main propulsion system sche-
cal pressure-feed spacecraft main propulsion matically illustrated in figure 11-5 is shown in
system using hypergolic Earth-storable propel- figure 11-6. The integrated main valve package
lants. A smaller and a larger pressurant tank is mounted directly to the main thrust chamber
satisfy pressurization requirements before and assembly, immediately upstream of the injector.
after a long coast period. The propellant tanks The integrated engine assembly can be either
are equipped with positive expulsion diaphragms rigidly mounted, or gimbal mounted, depending
and remain pressurized to mission completion. on requirements. Lightweight spherical propel-
An integrated main valve package contains main lant and pressurant tanks are secured by simple
propellant throttling and isolation valves, in- frames. An all-welded and brazed construction
cluding redundant features. This redundancy is is employed for system assembly to prevent
obtained by series-parallel connection of the external leakage of pressurant gas or propellant.
mechanically linked main propellant valves and In a complete space vehicle system, perform-
increases system reliability, virtually eliminating ance gains of the spacecraft stages will effect
critical failures (see also fig. 2-8). Aeration of increasing weight savings for each succeeding
lower booster stage, as indicated by equations
(2-1) through (2-5). When planning future space-
craft missions, high-performance liquid propel-
lants, such as LO2/LH 2 and LF2/LH_, thus
definitely should be considered. For high mis-
sion total impulse requirements, these propel-
lants, when used in turbopump-feed systems,
tend to provide a considerable performance edge
over pressure-feed systems (fig. 11-2). A
turbopump-feed system for these applications
must reliably supply high-pressure propellants to
the combustion chamber, under vacuum and zero
gravity conditions, and following extended coast
periods. All pumps must be fully primed prior to LFz/LH 2 propellant combination. The main
starting of the engine system to prevent delays thrust chamber is regeneratively cooled by the
in pump buildup and overspeeding. This may fuel to a nominal nozzle expansion area ratio of
require ullage settling rockets, before and during 44:1. A radiation-cooled nozzle extension skirt
main engine start, or positive propellant expul- may be added for certain missions to yield an
sion devices. overall expansion area ratio of 75:1. The pro-
The well-proven turbopump subsystem designs pellants are supplied to the main chamber by two
and their operating concepts as established for separate turbopump assemblies which are mounted
large booster engine systems, and as discussed to the thrust chamber body, resulting in an inte-
in preceding chapters, should be effectively gral, compact engine. The turbines are driven
applied to the design of turbopump-feed space- by gases bled from the main combustion chamber
craft main propulsion systems. In addition, the in a parallel arrangement (fig. 6-13). A hot-gas
systems must be able to start an unlimited num- throttle valve (such as shown in fig. 7-31)located
ber of times (if possible from available tank upstream of the'turbine inlets controls tile tur-
pressure energy, without auxiliary starting de- bine power, and thus the thrust level. A Y-type
vices) and to vary the thrust level (for some hot-gas duct is used to route the turbine ex-
applications over a wide range). Major problem hausts into the main chamber nozzle. The entire
areas associated with turbopumps under space engine package can be gimbaled at the gimbal
environmental conditions other than material mount by linear gimbal actuators.
considerations, are as follows:
(1) Vacuum, temperature, and radiations
Main Thrust Chamber Design
effects on exposed high-speed bearings
and dynamic seals In addition to the design elements presented
(2) Micrometeorite penetration of fluid pas- in chapter IV, primary requirements for thrust
sages chambers of spacecraft main propulsion systems
(3) Absence of the effects of gravitation on are:
pump operation
Figure 11-7 presents the package design lay- (1) Optimum steady-state cooling methods,
One of the problems associated with ablative coated with molybdenum disilicide (MoSi2) to
cooling is the dimensional change of the chamber prevent oxidation, are frequently used for nozzle
due to erosion. Nozzle throat erosion, if con- skirt extensions (for skirts starting at _ = 8 to 12)
trolled and pre2ictable, is acceptable for some which operate at sustained temperatures around
engine systems and becomes proportionally less 2600o-2800 ° F. For nozzle skirts of higher expan-
significant in larger thrust chamber units. For sion ratios (starting at _ = 40 to 45)° Hastelloy-X
fixed-area injectors and fixed propellant supply is sometimes used at operating temperatures
pressures, propellant flow and engine thrust. around 2000°-2200 ° F. One potential disadvan-
would increase with time, while specific impulse tage of radiation-cooled devices is their require-
would decrease, due to throat enlargement. How- ment to "see" space. A radiation-cooled skirt
ever, for a 40:I expansion nozzle operated in should face outboard and should not radiate
vacuum, the specific impulse loss would be only undesirable heat to vehicle components. Radia-
0.5 percent, for as much as i0 percent increase tion cooling appears unsuitable for clustered
in throat area, ifthe aerodynamic characteristics engines.
of the nozzle contour did not deteriorate. For exclusively ablative-cooled chambers
One remedy against excessive throat erosion used in LO2/LH 2 or LF2/LH z systems, analyses
is the use of a refractory throat insert. A ceramic, and tests have proven liquid hydrogen to be a
silicon carbide, has been used extensively for very effective throat film coolant in low chamber
throat inserts in space engine applications. It pressure applications (less than 100 psia). As
has a high melting point (4400 ° F), excellent low as 0.1 percent of the total propellant flow
thermal shock characteristics, relatively high used as the throat film coolant greatly reduces
thermal conductivity (115 Btu/hr-ft2-°F/in), low throat erosion for extended firing durations. The
coefficient of thermal expansion (2.4 x 10 -6 effect of such a small film coolant flow on sys-
in/in-°F), excellent oxidation resistance, and tem performance is so slight that it can be ne-
high abrasion resistance. Sometimes, a molyb- glected. Ablative chambers with their throat
denum backup sleeve is used when the silicon film-cooled by liquid hydrogen, therefore, may be
carbide insert cannot conduct heat sufficiently. considered excellent prospects forfuture pressure-
Another potential problem is the danger that feed spacecraft main propulsion systems using
the high-silica glass in the ablative material of LO2/LH 2 or LF2/LH _.
the combustion chamber section becomes suffi- Following the firing of an ablative thrust
ciently fluid to be swept downstream and to be chamber, the heat stored in the charred phenolic
deposited in the throat section. This causes and silica reinforcement or in the throat insert
thrust variances and promotes an unsymmetrical refractory material soaks into the unburned virgin
velocity profil6 in this area. This phenomenon material. This postrun soakback propagates
may be prevented by a liner of JTA (45 percent further thermal degradation, which might also be
graphite, 45 percent zirconium diboride, and I0 affected by the vacuum conditions, for 100 sec
percent silicon) inserted in the combustion zone. or more, until the mean temperature of the char
The liner is usually segmented, to provide a path is reduced to about 500 ° F. The weight of gas
for the gases from the pyrolyzed ablative (and generated and expelled by soakback charring is
reduce the Ap across the liner),and to prevent about 15 percent of the weight of ablative mate-
cracks that would develop in an expanding un- rial charred. It could cause a postrun residual
seg[;ented liner. impulse undesirably exceeding the desired mini-
Exclusively radiation-cooled thrust chambers mum cutoff impulse. However, this effect is
would be subject to large stresses caused by the small for larger systems. For long-duration
high temperatures in combustion zone and throat space missions, the temperature effects from
by the thrust transmission, and at the injector solar radiation may cause vaporization of the
attachment points. By contrast, the radiation- ablative chamber material during coasting and
cooled nozzle skirt is designed to accept and should be prevented ifpossible.
emit only the heat flux transmitted by the ex- Figure 11-8 shows the design of a typical
panding gases, and some of the loads imposed thrust chamber for a pressure-feed spacecraft
by thrust transmission. Molybdenum alloys, main propulsion system using hypergolic
DESIGN OF LIQUID PROPELLANT SPACE ENGINES 439
propellants is absorbed by allowing a small per- divided into a number of intervals. Equilibrium
centage of boiloff. Since conditions will vary skin temperatures and net heat fluxes (consid-
widely during travel, the net heat input to each ering both internal and external sources) to the
propellant tank may be computed by a numerical propellant tanks are then determined for each
integration process for the specific vehicle path, orbit location. By algebraically adding the net
such as a complete orbit. This path can be heat inputs for each, the net heat input rate for
OUT
i E- PISTON SEALS
ox,o,z
ACTUATING
o.-O FV W
FLOW
VENT
f SERVO
MOTOR TORQUE-- 7 _
. A
SERVO I,_--':1 /_ I _
I',,, VALVE / v
AS A PROPELLANT THROTTLING VALVE
1 _ WELD THERMAL RELIEF (TYPICAL)
---- A THROTTUNGVALVEONLY
SECTION A- A
Figure 1 l-9.-Typical mechanically linked, poppet-type, dual-propellant on-oil or throttling valve, [or
spacecra[t main propulsion systems using hypergolic Earth-storable propellants.
442 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
one complete orbit can then be obtained and the jectory correction maneuvers, for which it must
amount of propellant boiloff per orbit calculated. be oriented properly.
In a system without venting, any heat input Another example is spacecraft separation from
to the propellant results in a propellant tempera- the launch vehicle, during which momentums may
ture rise and a corresponding increase in vapor be imparted which require correction for proper
pressure. This system can be applied to both orientation.
cryogenic and Earth-storable propellants. The Ideally, when a spacecraft is oriented at some
allowable pressure rise, and consequently the attitude, it should remain there indefinitely.
maximum heat input, is dictated by the allowable This, however, is not the case, because the
tank pressure. In this case, the storage analysis spacecraft is continuously subjected to small
assumes an allowable vapor pressure rise during external and internal forces which will cause it
the mission, based on initialconditions. This to driftoff the desired position. An attitude
permits determination of the final propellant control system must function to counteract all
density and ullage volume requirements, and tinuous, or of an on-off type. Practical consid-
pressure-dependent propellant tank weights. For erations tend, in general, to favor the latter.
cryogenics, it is usually assumed that the pro- The attitude control system itself contains cer-
pellant experiences a uniform temperature rise. tain nonlinearities and nonideal conditions. In
The maximum storage time is then computed by actual operation these inherent nonlinearities
dividing the propellant heat capacity by the cause the vehicle to settle in a periodic motion
average rate of heat input. In this system, the about a reference point. This motion is repre-
storage weight penalties are composed of extra sented by a closed curve in the phase-plane and
tank weight required to accommodate the pres- is termed the "limit cycle of operation."
sure increase, the extra tankage required to The problem of attitude control, then, be-
allow for thermal expansion of the propellants, comes twofold: (i) what is required to rotate the
and the time-dependent insulation weight. For vehicle through a given angle to some new atti-
Earth storables, sufficient insulation is also tude, and (2) what is required to maintain the
required to prevent freezing. vehicle in this attitude. The energy requirement
to accomplish these two basic maneuvers can be
calculated, and total attitude control require-
11.4 DESIGN OF REACTION CONTROL
ments for any space mission can be computed,
ENGINE SYSTEMS
by simply determining the total number of times
that these two maneuvers occur.
Spacecraft Attitude Control Requirements
Of the various attitude control systems, the
Spacecraft attitude control requires the appli- reaction systems are the most versatile and pro-
cation of torques about the three axes (yaw, vide a wide range of torque, up to high levels.
pitch, and roll) passing through the vehicle's They are universally applied to manned space-
center of mass. The problem of attitude control craft and in large unmanned vehicles.
(1) Rotation of the vehicle over a given angle, thrust delivered on command must be very pre-
(2) Stabilization of the vehicle in a required hunting are to be avoided. Three basic opera-
tional modes are available:
position, within acceptable tolerance
limits 1. On-and-off, or "bang-bang" control.-This
"multiple start" system operates intermittently
The position of a vehicle, if left uncontrolled, as long as necessary at its rated thrust level. If
may not be the one required for efficient use of the on-and-off command signals are given as a
its main propulsion system. For example, a function of sensed position error only, the sys-
spacecraft may have to perform a series of tra- tem would tend to be unstable because of the
DESIGN OF LIQUID PROPELLANT SPACE ENGINES 443
timelags that are present in all real systems. tolerances, maximum impulse per cycle, space
This situation can be corrected through the use storage and environment, systems integration,
of control systems which utilize both position and logistics.
error and rate of error change to time the on and For monopropellant reaction control systems,
off command signals. the system vacuum specific impulse varies from
2. Proportional control.-The operating thrust 140 to 165 seconds for 90 percent H202, and
of th(sreaction control system is varied accord- from 200 to 230 seconds for N2H4. For bipropel-
ing to the sensed error signals. lant systems using Earth storables or cryogen-
3. Repetitive pulse control.-This type of ics, the system vacuum specific impulse may be
system needs only position error sensing and estimated by applying an efficiency factor rang-
uses a system which delivers thrust in a con- ing from 90 to 94 percent for steady state, and
tinuous series of accurately reproducible impulse from 80 to 86 percent for transient operation
bits to assure orientation and stabilization of the (less than 25 milliseconds pulse width), of the
vehicle. The optimum thrust pulse outline should ideal specific impulse values given in table
be a square-wave. Control may be achieved by 11-1. The 90 percent H202 systems are applied
modulating pulse width, or the frequency of a up to a maximum total impulse of about 50 000
fixed pulse width, or by combining both methods. lb-sec and the N2H4 systems to the same or
somewhat higher values. Beyond this level, the
Selection of Reaction Control Engine Systems hypergolic Earth-storable bipropellant systems,
For reaction control engine systems, both which are in the majority, take over.
liquid monopropellants and bipropellants have
Optimization of Reaction Control System
been used. The selection of an optimum system
Operating Parameters
for a given mission is based primarily on total
systems weight (including propellants) versus For most reaction control engine systems, a
total impulse. As mission time increases, so pressure-feed system using stored inert gas as
does the propellant and tankage portion of the pressurant is employed because of its simplicity
total systems weight, while the weight of noz- and reliability. After the propellants and their
zles, valves, and plumbing remains fixed. A feed system have been selected, parametric
comparison of two reaction control systems of analyses of all engine system design variables
different fixed weight and specific impulse may are performed to establish the minimum weight
show the lower fixed-weight system to have an system, within mission requirements, which still
overall weight advantage, despite a considerably satisfies other important considerations. Major
lower specific impulse. However, sometimes the operating parameters affecting systems weight
lower fixed-weight system cannot be used, simply are: chamber pressure, pressurant storage pres-
because it has not enough thrust or cannot be sure, mixture ratio, and nozzle expansion area
operated more than a specific number of cycles. ratio.
If total impulse is the variable and total sys- Chamber pressure and nozzle expansion area
tems weight the evaluating criteria, a plot of ratio are foremost optimization parameters, since
total systems weight versus total impulse can be the weight of major system components is directly
used to compare competitive systems. If thrust related to and defined by them. Component
and number of restarts (operating cycles) are weights considered in the analyses should in-
also factors, added dimensions in presenting the clude the pressurant and its storage tank, the
comparison are required. Based on the selection fuel and oxidizer tanks, and the thrust chamber
criteria of minimum total systems weight, the assemblies. Optimum values for chamber pres-
three major factors to be considered as inde- sure range from 50 to 200 psia, and for nozzle
pendent variables for different space missions expansion area ratios from 25:1 to 60:1. The
are total impulse, thrust level, and number of optimum mixture ratio depends on propellant type
cycles. Selection of a reaction control engine and thrust chamber cooling method used. Opti-
system is largely dependent on these three re- mum design values for inert gas pressurant stor-
quirements, in addition to reliability considera- age pressure range from 3000 to 5000 psia (using
tions. Other important factors are performance 6 A1-4V titanium alloy forgings as tank material).
i ....
Basic System Design for Reaction Control helium at the required pressure level pressurizes
Engines the positive expulsion bladders in the oxidizer
and fuel tanks. Opening of propellant isolation
For N2H 4 and 90 percent H202 monopropel- solenoid valves allows the pressurized propel-
lant systems, the propellant tanks can be pres-
lants to flow through micronic filters, distribu-
surized by a stored inert gas system, as shown tion lines, and line isolation solenoid valves to
in figure 5-1. The propellant valves are located
the normally closed injector solenoid valves
downstream of the propellant tanks and are used
which control the operation of the individual
to control the propellant flow to the thrust cham-
reaction control thrust chambers. The system is
ber. In an H202 system, each thrust chamber shut down by removal of the start signal. This
has its own silver-mesh catalyst, similar to the
closes the propellant-isolation valves and dis-
one in the H202 gas generator shown in figure connects the reaction-control injector-valve
4-50. When H202 comes in contact with the
solenoids from the command system.
catalyst, a hot-gas mixture of steam and oxygen A relief valve sealed with a burst diaphragm
is produced, the temperature of which is a func-
protects the pressurant tank against overpres-
tion of H_O 2 concentration and its liquid temper-
surization. A high-pressure transducer monitors
ature. For N2H 4 monopropellant systems, some helium-tank pressure. In a hermetically sealed
kind of ignition source is usually provided be-
system, it could also be utilized to indicate the
cause the catalyst life with N2H4 is limited.
amount of propellants remaining in the tanks. A
The ignition source could be either an electrical
low-pressure transducer is located downstream
heating element, or a continuously operated
of the helium pressure regulators to detect regu-
nuclear energy device. Frequently, to simplify lator malfunctions. Relief valves sealed with
ig'nition requirements in a N2H 4 system, a com-
burst diaphragms are also provided for the pro-
mon gas generator feeds several nozzles.
pellant tanks to protect them against over-
Since the liquid H202 is cool until it decom-
pressurization.
poses, the components upstream of the catalyst
do not require high-temperature materials. How-
ever, feedback of heat from the catalyst chamber Systems Redundancy in Reaction Control Systems
and the nozzle to other components must be pre- For reliable reaction control engine operation,
vented, since it would not only damage these
redundancy may be provided in the following
components but could also initiate vaporization three areas:
or even decomposition of the stored propellant.
The H20 2 system has a temperature advantage 1. Redundancy within a subsystem.-Several
(maximum temperature 1500 ° F) over a system typical examples of redundancy within a sub-
using N2H 4, although the latter has a higher system are shown m figure 11-10. Two explosive-
specific impulse. In the N2H4 system, the high actuated pressurant start valves are used in
operating temperature of 1S00 ° F requires suit- parallel. Similarly, two pressure regulators are
able materials all the way from the generator to used in parallel, pilot selected by a three-way
the nozzles. solenoid valve. In each propellant pressurant
Figx_re 11-10 illustrates the basic schematic line, a "quad _ check valve assembly is used to
of a typical reaction control system using hyper- insure that propellants which may have permeated
golic Earth-storable bipropellants and helium gas through the positive-expulsion bladders will not
pressurization. Two redundant subsystems are reverse-flow into the common pressurant line
provided, each of which is physically and func- downstream of the pressure regulators. A similar
tionally independent. Each subsystem is acti- "quad" check valve assembly is installed down-
vated independently by an electrical command stream of each propellant tank to prevent reverse
signal to the explosive-actuated pressurant-start flow of propellants from the other propellant sub-
valves located in the helium pressurization line, system, or from the main propulsion system.
and to the various solenoid isolation valves. 2. Redundancy between subsystems.-Refer
After passing through a micronic filter, different again to figure 11-10. Here, redundancy between
pressurant-isolation valves, pressure regulators, two independent subsystems is provided by nor-
and the "quad" check-valve assemblies, the mally closed solenoid intersubsystem-propellant
-- - -- -- -- --- - I III I ii
DESIGN OF LIQUID PROPELLANT SPACE ENGINES 445
_oo__O_T_O_
_,oo_
---_ --_U_NT,_?L_O_
NoR SSU NT,SOLAT,ON
, /T NS OCE
ACTUAT_O
P_ESSURANT
\ F----" _"
N.c.
EXPLOSIVE-..
START VALVE \J
_ \
_ _"
_ , _ I_°_
°'o3 WAY
i _ LOW
MODULAR CONTROL__
._ FUEL TANK
PACKAGE _ _ __._
VALVE I__ I
I
PACKAGE
ouc_
ECHEC_
_ : _r _ I
l_ _ CONTROL
"----- SOLENOID
_SOLATION
J VALV E
OUAD
CHECK
IDENTICAL --_" VALVE
ISOLATION VALVE
f FUEL SOLENOID
OXIDIZE SOLENOID
SUBSYSTEM
--Z_Z iSOLATION VALVE
SUBSYSTEM PROPELLANT
ASSEMBLY (iNCLUDING
r- THRUST CHAMBER
VALVE INJECTOR VALVES) TYP
m
@ SERVICE CONNECTION
valves. If one propellant subsystem should fail, systems are pressure fed and use a N204/MMH
these valves open and connect to the propellant propellant combination. The tasks of the Gemini
lines of the other subsystems. The distribution reaction control engine systems built by Rocket-
line solenoid isolation valves can isolate a dyne are high-altitude abort trim, attitude control
group of thrust chambers (pitch, roll, or yaw) in orbit and during reentry, and rendezvous and
shouid an individual thrust chamber in the group docking control. In addition, four solid propel-
malfunction. lant engines built by Thiokol are used for high-
3. Redundancy outside the system.-Redun- altitude abort and reentry retrofiring.
dancy outside the system may be provided by The reaction control engine systems for the
connecting to the pressurant and propellant sys- Gemini spacecraft (fig. 11-11) consist of-
tem of the spacecraft main propulsion system, (1) Sixteen 25-pound thrust chambers for
which must be capable of supplying one or both reentry control (also see fig. 11-14),
subsystems with pressurant and/or propellants. consisting of two complete propulsion
systems, with a pulsing requirement of 6
cycles/see and a minimum impulse
Packaging and Installation of Reaction Control bit= 0.25 lb-sec
Engine Systems
(2) Eight 25-pound thrust chambers for orbital
As with spacecraftmain propulsionsystems, attitude control-pulsing requirement: 6
the controlcomponents of a reactioncontrol cyclessee, minimum impulse bit = 0.25
engine system can be modular packaged. The lb-sec
modular packages will incorporate related con- (3) Four 100-pound thrust chambers for trans-
trol components within the same housing (forging lation control in rendezvous and docking
or casting), as shown in figure 10-9. Typical maneuvers-pulsing requirement: 2
modular control packages for reaction control cycles/see; minimum impulse bit = 26
systems are indicated schematically in figure lb-sec
11-10. The thrust chamber injector propellant (4) Two 100-pound thrust chambers and two
valves are frequently designed as an integral S5-pound thrust chambers for longitudinal
part of the thrust chamber assembly (fig. 11-13). propulsion-pulsing requirement: 2
All-welded and brazed construction is preferred cycles/see; minimum impulse bit = 25
to prevent pressurant and propellant external lb-sec
leakage. The Gemini control thrust chambers are all
Figure 11-11 presents the installations of the located inside the vehicle, with the nozzles
reaction control engine systems used on a typi- trimmed flush with the outer skin. Thus, no
cal manned spacecraft, the Gemini capsule. All radiation cooling can be used. All thrust cham-
10('_._ IM_'LI5 _ _NAMB[I_ 12
bers are ablatively cooled.
with ablative or radiation cooling in overall sys- equations (4-36), (4-37), (11-1), and (11-2) may
tem simplicity, minimum system pressure drop, be used. However, in the case of reaction con-
and minimum performance loss (film or transpira- trol thrust chambers designed for repetitive
tion cooling requires extra propellant flow). pulse, o; for intennittent operation, the heat-
Also, the advancement of the state of the art of transfer conditions may be quite different. Many
materials, and of analytical and design tech- studies and experiments have been conducted in
niques, has made both ablative and radiation this area, and theoretical and empirical correla-
cooling, rather reliable and practical cooling tions have been generated.
methods for reaction control thrust chambers. A study by Lee and Hahn' indicates that for
The design and construction principles for repetitive short-pulse operation, the char-front
ablative-cooled reaction control thrust chambers region in an ablative-cooled thrust chamber does
are not basically different from those for main not experience appreciable temperature fluctua-
propulsion systems. However, the small physi- tions, because of the attenuating effect of the
cal sizes (from 1- to 100-pound thrust), and the low thermal diffusivity of the char layer. The
operational modes (such as the pulse mode) of char front region therefore stays at the pyrolyzing
reaction control thrust chambers, require some temperature during the entire period of cycling.
special considerations. The char regression under pulse operation can,
For example, a typical roll control thrust therefore, be treated as a case of continuous
chamber is designed to produce thrust ranging firing with effectively reduced gas-side heat
from 1.6 to 2.5 pounds at chamber pressures from transfer. The results show that the char depth
S0 to 130 psia. The design throat area is 0.0120 based on equal cumulative firing time is a func-
in 2 (0.124-inch diameter). This throat diameter tion of percentage of the firing time over the
should be produced accurately and remain un- elapsed time. It increases with the decrease of
affected by erosion during firing. Generally, in the percentage of firiug, to a maximum value
small ablative rocket motors, the internal geom- several times that obtained with continuous fir-
etry remains essentially unchanged as ablation ing. At low percentages of firing, i.e., below
progresses. The thermal protection in this case approximately 5 percent burn (pulse width/pulse
is provided by the internal change from the cycle), the char depth drops again, because of
pyrolysis or decomposition of the plastic resin. the increasing p_oportion of radiation or convec-
This yields a porous char layer without any sig- tion losses from the outer skin surface of the
nificant dimensional changes. The transition chamber, to heat influx from the combustion
zone between the virgin ablative and the char is gases. In fact, at some critical pulse mode, the
referred to as the _char front." The ablation char regression ceases after an equilibrium char
rate, in this case, is not governed by a surface depth is attained.
regression, but rather by a regression of the During intermittent, on/off, i.e., multiple-
interface between the virgin ablative and the start firings, with off periods in between fiLings
char. However, if highest accuracy is required, of relatively long durations, the temperature at
throat inserts made of refractory ceramics, such the char front does not remain at the pyrolyzing
as silicon carbide, and combustion chamber temperature all the time. The char-depth pro-
liners made of graphite or ceramics, will further gression for a typical multiple-start system is
reduce throat-size changes due to erosion and shown in figure 11-12. The ascending straight
will prevent ablative fluid from being swept line represents char-depth progression for a con-
downstream from the combustion chamber and tinuous firing (100 percent burn). Assume that
deposited in the throat section. Silicon carbide the first firing cycle was terminated at "a."
can be machined to rather close tolerances by Since the char layer at this time is at a higher
grinding. The diameter of a throat can generally temperature than the pyrolyzing temperature,
be maintained with a tolerance of _+0.001 inch.
The methods of determining the char depth of
ablativereactioncontrolthrustchambers are LLee, J. C., and Hahn, J. R., "Regression Rate of
similarto those of the largersize chambers. Char Front in Ablative-Cooled Rocket Motor Under
Semiempiricalor empiricalequationssuch as Pulse Operation," AIAA Preprint 64-262.
448 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
,.:
",Z PARALLEL WRAP Lgo"
...... \ ....NSER,
ORIENTED
assembly.
ber andCERAMIC LINER
nozzle,
It has
with
an integral
a thick
combustion
throat section.
cham-
The
ABLATIVE _ L A90/L;TIVE (SEGMENTED)
designed and applied, is probably the simplest, structure and is thermally insulated by ceramic
lightest, and relatively most reliable. The re- spacers.
quired working temperature of a radiation-cooled The injectors for reaction control thrust cham-
thrust chamber wall could be as high as 3400 °- bers using hypergolie Earth-storable propellants
3700 ° F for Earth-storable propellants operated are usually designed with a conventional fixed-
at or near their optimum mixture ratio, and at orifice, single-ring, unlike-impinging doublet
chamber pressures of 100 psia. Advanced refrac- pattern. A splash plate is often utilized to
tory materials such as the 90 percent tantalum- improve performance.
INJECTOR
WALL THICKNESS
FUEL INLET
COMB. CHAMBER
DIA
THROAT DIAMETER
OXIDIZER'-_-
INLET
-- FITTING
MOUNTI NG
STRUCTURE
A-1 stage engine, 64, 65, 66, 67, 68, 86, 96, 102, A-2 stage engine (continued)
109, 110, 129, 130, 193, 201, 217, 218, 219, 222, starting sequence, 72
224, 249, 250, 261, 295, 296, 303, 364, 378,391, steady state operation, data for, 166
392, 393, 395, 397, 398, 399, 4I_ system operation, 70
thermal conductance, 103
alternate turbine design, 257
assembly design layout, 261 thrust chamber configuration, 95, 96
bellows, 378 turbopump, 199, 389, 392
C* correction in performance, 398-399 venting, 166
centrifugal pump, 224 A-3 stage engine, 72, 73, 74, 75, 76. 86, 97, 156, 167
chamber pressure vs. engine thrust, 395 chamber product gases, values for, 86
chamber product gases of, values, 85, 86 engine and propulsion system operational sequence,
cutoff sequence, 68 75
injection momentum ratio, 129 A-4 stage engine, 74, 75, 76. 77, 78, 95, 97, 121, 144,
injection velocity, 130 154, 156, 158, 159, 163, 268, 269, 270, 324, 325,
main oxidizer valve, 295-297 338, 342, 344, 392
oxidizer pump, 217-219, 222, 364 burst diaphragm on oxidizer tank, 344
chamber nozzle extension, 121
packaging, 399
preliminary layout for 750K, 67 chamber product gases, values for, 86
propellant flow design characteristics, 391 comparison of various tank pressurization systems,
injection velocity, 131 relief valve, design data on, 324, 325
injector data, 130 storage pressure requirement, 159
nozzle expansion area ratio, 76 system operation, 76
operating parameters (150K), 69 tankage, 342
packaging, 399 thrust chamber configuration, 95
451
452 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
Propellant utilization (PU), 268, 269, 270, 272, 311, Reliability, 42-49
336, 393 definition, 42
Propellant vaporization, 152 design reviews, 49
Propellants, 18-20, 22-23, 188, 203, 431-432 implements for obtaining highest, 43
additives, 19, 20 pointers for obtaining, 42, 43
bi-, 19, 26, 27 Repetitive pulse control, 443
cooling and other characteristics, 432 Resins, pyrolysis, 118
cryogenic, 19, 24 Resistance coefficients for fluid-flow control com-
density, 431 ponents, 286
flow of, engine start, 203 R-F (reverse-flow) nozzle, 92
fluid properties, 188 RF signals, 68, 138
ignition characteristics, 431 Ring slot injector, 127
mono-, 18, 25 Rocket thrust, general, equation and model for, 2, 3
operating temperature, 431 Rocketdyne, 448, 449, 450
optimum mixture ratio for additives, 19, 20 Rocketdyne AR-1, 141
performance and physical properties, 20 Rocketdyne Atlas ]CBM, 295
selection, 18-20 Rocketdyne Division of North America, 37
specific impulse in evaluation, 431 Rocketdyne LR79-NA-11 engine, 403
storable, 19, 22-23 Rocketdyne Naflex, 368-370
Propeller type impeller, specific speeds, 191 Rotor blades of turbine, main loads, 244
Proportional-type control, 264, 443 Rotor-stator assembly, centrifugal pumps, 176, 177
Propulsion system, definition, 28 RP fuel, 141
Pulses, 150 RP-1 fuel, 64, 109
Pump, cavitation, 192 Run duration, 32-34
Pump developed head, 189
Pump head coefficient, 189
Pyrolytic graphite, 119
Pyrometers, 142 Safety features, disadvantages, 43
Pyrophoric fluid, 19 Safety margins in design, 334-335, 428
Pyrotechnic igniters, 137, 138 Sample calculations. See Calculations for 4 stages
of Alpha vehicle, samples of
Saturn I first stage (S-I) booster for, 416
Saturn V, 329, 416, 425, 430
Scheduling, availability of design in, 50
Qualification tests, 32
Schmetterling AA rocket, 140
Quality control, areas, 48
Sealing, 361
Quintuplet injector, 125, 127
Secondary injection thrust vector control (SITVC),
279, 280
Self-impinging injector, 125, 127
Servovalve flapper, 319
Radial-flow impeller, minimum basic design elements, Servovalves, 309-315
208 Shepard, Cdr., 141
Radial in-flow nozzle, 92 Showerhead injector, 125, 127
Radial out-flow nozzle, 92 Sieder-Tate equation. 106
Radial-type impeller, specific speeds, 190 Simplicity of engine design, 32. See also yon Karman
Radiation cooling of thrust chamber, schematic, 120, Single gas generator-helium system, 172
121
Single gas generator system with injection cooling,
Rannie equation, 118 171
Rao, G. V. R., 91 Single-geared turbopump, 182
Rating stability, 150 Single-stage vehicle, growth factor equation, 40
Rayleigh flow process, 6 SITVC. See Secondary injection thrust vector control
Reaction control system, 429-430 Skinner pilot valve, 302
Reaction Motors Division of the Thiokol Corp., 429 Solid propellant gas generators, 132, 133
Reaction turbine, 180 Solubility of pressurizing gas, 152
Redstone rocket, 33, 34, 64, 141, 142, 400 Space engines, design, 429-450
Redundancy, 46, 142, 444-446 general applications, 430-434
Refrasil-filled chambers, 119 main propulsion systems, 435-442
Regenerative cooling, 105-114 principal applications, 429-430
Regulating devices, fixed-area, 307-309 reaction control, 442-450
Regulators, gas pressure, 315-321 Space system operational requirements, 433
Regulators, liquid flow, 321-322 Spark plugs, 139, 140
Relays and switches, 406-407 Special alloys, 61, 62
INDEX 459
Turbine gas flow rate, 186 Valv_ir four-way solenoid pilot valve, 3q4
Turbines, 178-181 Valves, 294-306
Turbopump systems, 175, 181-189, 191- 199, 200-261 Vapor condensation, 152
axial-flow, design, 225-238 Vasco_et, 246
balancing axial thrust, 223-225 VDt orifices, 307, 308
booster engine. 182. 183 Vehicle, definition, 21
calibration and off-design characteristics. 202, 203 Vehicle, design, cost, 50
cavitating inducers, 210-219 Vehicle diameter, 425
cavitation, 192 Vehicle performance, deviations, 35
centrifugal, 204-225 Vehicle range, effect, 35
cycle efficiency, 199, 200. 201 Vehicle reliability and flight safety, 44
description of developed, 182-186 Vehicle system, three-stage, weight increase calcula-
design layout, 261 tions, 41
design parameters, 187-202 Velocity, characteristic, 12
drive arrangements, 181- 182 Velocity-compounded impulse turbine, 180
elements, 176 Venting, 166
envelope, 175 Venturi-type valves, 299-300
gear design, 260 Vernier cutoff system, 34
heads and flow rates developed, 189 Von Braun, Wernher, 63. 415
inlet suction pressure head. 186 Von K_rm_n, 82
operating ranges, 202 V-2 rocket, 400
overall performance and operating efficiency, 196,
197
pump flow coefficient, 189
\%'asserfall antiaircraft rocket, 140
range of operation, 175
Weatherhead Co., 358
resistance and pump characteristic curves, 189
Weight of payload, effect, 35-37
specific speeds, 190
Wet gimbaled weight of engine, 37
trimming effects, 202
Wet weight of engine, 87
turbines, 238-257
variation of efficiency with speed, 195 Wiring, connective, 409
variation of engine weight and flow ratios with Working loads. 56-58
WorldWar II guided missiles, 140, 141
chamber pressure change, 175
NASA-l.,angley, 1968 _ $8