Design of Liquid Propellant Rocket Engines Dieter K Huze and David H Huang - Part2

DESIGN OF TURBOPUMP PROPELLANT-FEED SYSTEMS 225
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
VANES --_ _ VANES

The peripheral speed at the diameter ds \ /TiP
=N = x 7000
× 4.8= 147 ft/sec
Us=_ds= 720
From equation (6-77), the reduction of the

axial forces
F 3zr rolaaA oa aAx (204 304 - 21609)

"= 4-g-6g_...... - ..... ' _ _-5g_.2 PU_P AXIS " 1
x 71.38 × (0.25 +0.21)_ 74 680 lb OF ROTATION

2x0.25
6.4 DESIGN OF AXIAL-FLOW PUMPS
Except when used as inducers, application

of axial-flow pumps in rocket engines is essen-
tially limited to liquid hydrogen systems in a
multistage configuration. Thus, the following I.U
k3
Z
discussions are applicable to axial-flow hydro-
gen pumps only. Multistage axial-flow hydrogen n-
O
I_. _!:¸¸_
pumps are applied in regions which are beyond e_
i,i
the capability of a single-stage centrifugal pump,
O_
since their construction is comparatively simple =E
(fig. 6-4). As can be seen in figure 6-6, the fluid

in an axial-flow pump flows from one stage to
.S
the next with a minimum of connecting passages. VANE HEIGHT hv, INCH
The head rise of a typical single-stage cen-
trifugal hydrogen pump is limited to about 65 000 Figure 6-49.-E[fect of vane height on the per-
ft (2000 psi). Beyond this point, a multistage formance of an axial-flow pump.
226 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
I C[_'rmFU_,L I Operation of the Impeller Rotor

! t_,,,, I
The main function of the impeller rotor of an
c E_TR_FUGAL ;
_,u=l,
axial-flow pump is to impart kinetic energy to
zo_
_ TIP U'!IED
the fluid by increasing the tangential component
i u_ • 16_r_¢'f_C of the absolute flow velocity, This is accom-
plished by the action of airfoil-shaped rotor
vanes (figs. 6-51 and 6-52). It is convenient to
,i
] AXlIL FLOe PUlP i
describe the vanes on several developed cylin-
drical sections. Three sections are of particular
..... I, , I,J
interest: at the impeller tip diameter dr, at the
impeller hub dh, and at its mean effective diam-
Figure 6-50.-Typical operating regions of vari- eter dm (inches). The mean effective diameter
ous pump types for liquid-hydrogen-fueled is defined by
rocket engine applications.
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
arrangement, and others. For instance, the re- Pr -=- (6-80)

Zr
quirements for the region above 5500 gpm, and
head rises from 30000 to 65 000 feet, could be
where
met by either a single-stage centrifugal pump or
Pr--pitch or rotor vane spacing at the mean
a multistage axial-flow pump. The centrifugal
effective diameter d m, in
pump also has its dimensional limitations. A
value of less than 0.2 inch for the impeller dis- z r -=number of rotor vanes
The ratio of the rotor vane chord length Cr to
charge width, b 2 (fig. 6-34), would complicate
manufacture of shrouded impellers and make the pitch Pr is called rotor vane solidity Sr
critical the tip clearance of an open-faced impel-

Cr
ler. This establishes the lower capacity limit
for centrifugal hydrogen pumps at about 250 gpm.
s,:_ (6-s_)
where S_-= rotor vane solidity at the mean effec-

Basic Assumptions for Axial-Flow Pumps
tive diameter din.
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.
impeller st_tor of f_n _xi_l-[loW
impeller rotor, _ncl
o',/\/ \__
is gradually increased from fi2 to f13. The dif- (AH)I

ference between the two, /33-/32, is a measure (d,), - 2 (6-90)
Um
of the vane curvature along any particular vane
g
section. Generally, in axial-flow pump designs,
all vane mean lines can be approximated by a
Cm =Cu2' tan a2' =cu3' tan aj =c 2' sin a2'
circular arc. The following correlations can be
established: = c 3' sin a 3' = v 2' sin /32' = vj sin/33' (6-91)
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
where a2', a3' =absolute flow angles at the rotor
/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
Rr = radius of the rotor vane curvature, in effective diameter din, ft/sec
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
design of impeller rotors, velocity diagrams of 10 percent of Q)
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
/32 =/32' + i (6-84) (AH) 1 = rated design developed head per

axial-flow pump stage, ft
He = hydraulic head losses per stage of
[93=/33' + ii (6-85)
impeller stator, ft
(d,), =head coefficient per axial flow pump
Qimp
Cm - (6-86) stage
3.12 x;(dl 2 - d h 2) (All applicable parameters refer to the mean
effective diameter, din)
Qimp = Q + Qe (6-87) At various cylindrical sections between vane
tip diameter dL and hub diameter dh, the follow-
.N ing correlations between vane angles and flow
um= _-_dm (6-88) velocities are established:
UmCuj - UmCu2 f dm tan fi2 --dr tan/32_

AHirap = (AH)1 + He - (6-89)
=dh tanfl2h=dxtan /32x (6-92)
DESIGNOF TURBOPUMPPROPELLANT. FEED SYSTEMS 229
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
hub diameter, deg the stator (fig. 6-52):

Ut, Uh = rotor peripheral velocities at tip
and hub diameter, ft/sec rrdm
Ps :-- (6-97)
Cu2t', Cu2 h' :tangential components of the de- gs
sign absolute rotor inlet flow
velocities at tip and hub diam- Cs
eter, ft/sec Ss :P_-s (6-98)

Cu3t', Cu3h' =tangential components of the de-
sign absolute rotor outlet flow (23 + (24
(6-99)
velocities at tip and hub diam- {2C-- 2
eter, ft/sec
• -
Function of the Statot Cs=2R s sm _---_---) - si-_-_ac (6-100)
The purpose of the stator of an axial-flow

aa :as'+ i (6-101)
pump is to convert a major portion of the tangen-
tial component of the absolute flow velocity
a 4 :a 4' +ii (6-102)
leaving the rotor into static pressure. This is
accomplished by "straightening" the flow as it C m : CU31 tan _3': CU4' tan a 4'
leaves the rotor. The stator vane curvature is
= c a' sin a a' = c 4' sin a 4' (6-103)
designed so that the fluid enters the vanes with
minimum loss, and leaves the stator with a re-
dm tan an=dr tan a3t
duced tangential component of the absolute flow
velocity. The cross-sectional areas of the stator :dh tan aah=dx tan a3x (6-104)
flow passages normal to the axial direction are
dm tan a 4 =dt tan a 4 t
equal to those of the rotor. Thus, the axial com-
ponent of the absolute flow velocity is main- =dh tan a4h=dx tan a4x (6-105)
tained. The dimensions dt and dh of the stator
can be treated as equal to the tip and hub diam- whe:e
eters of the rotor. The chord-pitch ratio of the Ps : pitch or stator vane spacing, in
stator vanes generally increases from hub diam- zs = number of stator vanes
eter d h to tip diameter dr. The axial length Ls Ss : stator vane solidity
of the stator vane at the mean effective diameter Cs = stator vane chord length, in
is usually made equal to that of the rotor, Lr. _C = stator vane chord angle, deg
230 DESIGNOF LIQUID PROPELLANT ROCKET ENGINES
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
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
the same effective passage cross-sectional area

normal to the axial velocity component cm. The
] inlet and outlet velocity diagrams
stator are shown in figure 6-52. The following
for the inducer
correlations can be established for the design of

an inducer stator:
_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
flow pump. aic= 2 (6-110)

Inducer stator vane solidity design values

Cis=2Ris sin -sin aic range from 1.5 to 1.8; the number of vanes Zis,
ranges from 15 to 20. The number Zis should
al =al' +i (6-112) have no common factor with the number of impel-
ler rotor vanes, zr.
a2 = a2' + ii (6-113)
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
shown in figure 6-54. It consists of a disk hav- r7 2

ATa=APc--c(df -dr 2) (6-117)
ing small clearances with a pair of seal rubs,
located on either side of the disk. A forward where
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,
(:Iv, - ..... _,:\_ _ _ :\

--4---.. ....... \\
--Jr -- ---- dr df
•"_"_!\_4, ',"-,"R.P.M. _..-_,_

.\",",\
/.L_.__\'< :_<\\ : \:',c', _ _ , ,,\', >_ Zf I
/-- ......... K_;,_' I _?/ I
/ _ _,_ , /V" I
........ _/ jl
_.Y/>
FRONT SEAL RUB'--" X-BALANCE PISTON

CONTROL CHAMBER
Figure 6-54.-Axial-(low pump volute casing and balance piston arrangement.

Rated design pump developed head, AH

[N(q)°s] _.33_
= 44 800 ft (AH), = L(--N_), J
Rated design pump flow rate, Q = 6080 gpm
Rated design pump (NPSH)c = 135 ft _[27000×(60s0)°713
Inducer general configuration = cylindrical tip - [ 325-O ] = 5580 ft
contour; tapered hub contour
Inducer suction specific speed, (Nss)in d
= 53 400
From equation (6-90), the peripheral speed at
Inducer inlet flow coefficient, ¢i=0.09 max
the impeller mean effective diameter dm
Inducer head coefficient, ¢ind = 0.307
Inducer leakage loss rate, Qee = 0.03 Q
132.2 × 5580
Inducer stator head loss, Hee =0.08AHind
Inducer stator vane solidity, Sis--1.53
Number of inducer stator vanes, Zis = 17
Specific speed per axial-flow pump stage, From equation (6-88), the impeller mean effec-
tive diameter
(Ns) 1 = 3200
Head coefficient per axial-flow pump stage,
(,/,), : 0.304 720 Um_ 720 x 768
Impeller hub ratio, rd : 0.857 din- 7rN 7r×270_=6.52 in
Impeller leakage loss rate, Qe = 0.06 Q

Head loss per stage of the impeller stator, From equation (6-79), the impeller rotor tip
He = 0.08 (AH)I diameter (cylindrical tip contour)
Impeller rotor vane solidity Sr at the mean
effective diameter = 1.05 2 _ 2
---7 in
Number of impeller rotor vanes, z_ = 16 dt--dm (1 +rd2 ) 6.52 1+0.734
Impeller stator vane solidity 8s at the mean
effective diameter : 1.61 The impeller rotor hub diameter
Stator and rotor vane passage contraction
factor, e = 0.88 dh--rd×dt:O.S57x7=6 in
Angle of attack at the vane inlet, i = 4°
Angle allowed for circulatory flow at the vane The rotor vane height
outlet, ii : 5°
Calculate and design basic pump dimensions
and vane detail of: (a) inducer stator, and (b) hv =_=_-_= 0.5 in
impeller rotor and stator.

Seven main pump stages and one inducer
stage are used. From eq. (6-107), the required
inducer head rise
Solution (refer to fig. 6-52)
(a) Inducer stator AHind (1 - 0.08) : AH- n(hH),

From equation (6-10), the rated pump rotating
speed can be determined: _ (44 800 - 7 × 5580)
AHind
0.92
Nss (NPSH)c o._ = 6240 ft

N: (N ss : (N ss)ind : 53 400)
QO_
Use an inducer tip diameter of 7 inches, equal
to the impeller rotor tip diameter. Inducer head
N - 53 400 × (135) °Ts = 27 000 rpm
rise:
(6080) _s
From equation (6-106), the pump developed Nrrdt _ 27 000 × 7r× 7

= 826 ft/sec
head per stage ut = 720 720
Substitute this into equation (6-66):
3.12 × _-× 230

ut2 (826)2 = 6500 ft
AHind = _'ind _ = 0.307 × 32.----_
The mean effective diameter at the inducer
This is slightly more than the minimum of 6240 outlet
feet required.
The required impeller and inducer flows are
obtained from equations (6-35) and (6-63): d_ -- dt2 +d,h
2 2 = _49+
' _37.58 =6.57 in
Qimp -- Q + Qe = 6080 (1 + 0.06) = 6450 gpm Tile peripheral speed at d,
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 x-_(dt 2 -doh 2)

The inducer design absolute outlet flow
velocity
6450
- = 64.8 ft/sec
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
Cmo 64.8 , cm 230

_ind ..... 0.0784 (< 0.09 max. specified)
ut 826 tan a I =--cul' :_-0= 0.852; a 1 =40°26 '
Use identical values cm for the meridional

We use an inducer stator with a meridional
component of the absolute flow velocities through
the inducer outlet, the stators and the rotors. flow area equal to that of the impeller rotors and
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
6450 al =a,'+i=40°26 ' +4 ° =44°26 '

= 230 ft/sec
3.12 x4× (49- 36) x 0.88

For a design absolute flow angle at the stator
outlet, a2' of 65 °, equation (6-113) yields a vane
From equation (6-60), the required hub diam- angle at the stator outlet
eter at the inducer outlet
a2 =a2'+ii=65 ° +5 ° =70 °
dlh=_dt2 Qind The tangential component of the stator abso-

3.12×4×cm lute outlet flow velocity
cm 230 230 Nominal mean effective diameter, d m =6.52 in

c, 2' = tan a 2' = tan 65 ° - 2.145 = 107.2 ft/sec
Vane elements (fig. 6-52), al =44°26'; a2
=70°; aic=57°13'; Sis=l.53; Zis=17;
The stator absolute outlet flow velocity Pis=l.205in; Cis=l.844in; Lis=l.55in;

Ris =4.16 in
c 2' = \/c u 2' 2 + cm 2 = X/11 490 + 52 900
-- ,,:'65 350 = 253.8 ft/sec (b) Impeller rotor and stator

Assume that the design absolute flow condi-
From equation (6-108), the pitch of the in- tions at the impeller rotor inlets and at the im-
ducer stator vanes at dm peller stator outlets are identical to those at the
inducer stator outlet. Thus
Pis - rrdm _ ,7×6.52 1.205 in

Zis 17 a2' =a4' =65 °
c 2' =c 4' =253.8 ft/sec

From equation (6-109), the chord length of the
inducer stator at dm cm =230 ft/sec
cu2' = cu4' = 107.2 ft/sec

Cis = Sis Pis = 1.53 x 1.205 = 1.844 in
The design relative flow angle at the impeller

From equation (6-110), the chord angle of the
rotor inlets
inducer stator vanes at dm
cm 230
a, +a2_44°26 ' +70 °
=57o13 ' tan /32 ' - um-cu2 -1 - 768-107.2 = 0.344; f12' = 19°
aic - 2 2
The relative flow velocity at the impeller

From equation (6-111), the axial length of the
rotor inlets
inducer stator vanes at d m
v2' =\/ium-Cu2') 2 +cm 2

Lis = Cis sin aic -- 1.844 × sin 57°13 ' = 1.55 in
= ,_/436 700 + 52 900 = 699.6 ft/sec
The radius of the inducer stator vane curva-

ture at dm From equation (6-84), the rotor inlet vane
angle at d m
Cis 1.844
/32 =/32' +i=19 ° +4° :23 °
Ris=2 sin(-_) -2xsin(12°47')=4"16 in
From equation (6-89), the required developed

A-2 Stage Engine Fuel Pump Inducer Stator
head for the impeller rotor
Design Summary
(Unless specified otherwise, all data are at
AHimp = (AH), + He = 5580 (1 + 0.08)
the mean effective diameter din.)
= 6026 ft per stage
Inlet flow velocity diagram (fig. 6-52), aa'
=40°26'; c 1' =354.7 ft/sec; c m =230 ft/sec;
From equation (6-89), the tangential compo-
Cu,' = 270 ft/sec
nent of the design absolute flow velocity at the
Outlet flow velocity diagram (fig. 6-52),
impeller rotor outlet
a s' = 65°; c 2' = 253.8 ft/sec; Cm = 230 ft/sec;
cu2' = 107.2 ft/sec
ghHimp 32.2 × 6026
Nominal tip diameter, dt = 7 in '- _ Cu2' - +107.2
Cu_ um 768
Nominal hub diameter, dh = 6 in
Nominal vane height, hv=0.5 in = 359.6 ft/sec
The impeller rotor design absolute outlet flow Cr 1.346

velocity Rr= /33-/32 2sin5°33 '=6'95in
2 sin-
2
C3' = V/Cu3' 2 + Cm2 _- V/129 300+ 52 900 = 426.9 ft/sec
Assume that the design absolute flow condi-

The impeller rotor design absolute outlet flow tions at the impeller stator inlet are identical to
angle
those at the impeller rotor outlet. From equation
(6-101), the impeller stator inlet vane angle at
Cm 230 dm
tan a3'=Cu 3,-359.6=0.641; a3'=32°40'
a3 =a3' +i = 32°40 ' +4 ° -=36o40 ,
The impeller rotor design relative outlet flow

From equation (6-102), the impeller stator
velocity
outlet vane angle at dm
v 3' :V/(Um - Cu3') 2 + cm 2

(Z4 -- a 4' + ii : 65 ° + 5 ° -- 70 °
= X/166 800 + 52 900 : 468.7 ft/sec
The axial length of the stator vanes at d m is

The impeller rotor design relative outlet flow
equal to that of the rotor vanes; thus
angle
Ls-- Lr: 0.645 in

_ cm 230
tanfl3' (um-cu3') 408.4 -0.564; fij=20°26'
stator vanes at dm
From equation (6-85), the rotor outlet vane
angle at d m a3+a4_36°40'+70°=53o20,
ac- 2 2
f13 : fi3' + ii : 29026 ' + 5 ° = 34026 '

From equation (6-100), the chord length of the
From equation (6-80), the rotor vane pitch stator vanes at d m
rfdm 77×6.52 Ls 0.645

Pr= -= - 1.281 in Cs:_n ac -sin- (53o20,)=0.805 in
Zr 16
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

rotor vanes at dr, From equation (6-98), the stator vane pitch
at dm
/92 +/33 _23 ° +34026 '
/3c - 2 2 = 28o43'
Cs 0.805 _
Ps-=_-;: 1-y-AT:0._ in
From equation (6-83), the axial length of the
rotor vanes From equation (6-97), the number of the stator
vanes
Lr = Cr sin tic = 1.346 × sin 28°43 ' = 0.645 in

Z_
m rrx6.a2 r- ..
Z s =--:--"-x--c---. = ql
The radius of the rotor vane curvature at dm Ps u.5
4
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
Nozzle efficiency _?n AH0__" = isentropic enthalpy drop of the gases

flowing through the nozzles, due to
Actual gas kinetic energy at the nozzle exit expansion, Btu/lb
Ideal gas kinetic energy (isentropic The performance of a turbine nozzle, as ex-
expansion) pressed by its efficiency or velocity coefficient,
is affected by a number of design factors, such
C12
as
2g _ka2 (6-119) (1) Exit velocity of the gas flow
(2) Properties of the turbine gases
2g (3) Angles and curvatures at nozzle inlet and
exit
Nozzle throat area coefficient ent (4) Radial height and width at the throat
(5) Pitch or spacing, and number of nozzles
Effective area of the nozzle throat
Z (6-120) Design values for the efficiency and velocity
Actual area
coefficients of a given turbine nozzle may be
determined experimentally, or estimated from
Actual gas spouting velocity at the nozzle exit,
past designs. Design values of nozzle effi-
ft/sec:
ciency 77_ range from 0.80 to 0.96. Design val-
ues of nozzle velocity coefficient kn vary from
0.89 to 0.98. The nozzle throat area coefficient
ent generally will increase with nozzle radial
C_ = k n C O = kn gJCpTo -\Po/ J
height, with design values ranging from 0.95 to
=kn_/2gJAHo-i" (6-121) 0.99.
The cross-sectional shape (fig. 6-56) of rocket
Amount of nozzle reheat: turbine nozzles is square, or, more frequently,
rectangular. They are closely spaced on a cir-
(i - kn 2)C 2 _ (1 - _n) C 2 cular arc extending over a part of (partial admis-
(6-122)
qnr= ka22gJ 7?n2gJ sion), or all (full admission), the circumference.
Most high-power turbines use full admission for
better performance.
Required total nozzle throat area, in2: While the gases are passing through a nozzle
and expanding, the direction of flow is changing
wt (6-123) from an approximately axial direction to one
Ant- ),+___A_l forming the angle al (fig. 6-56) with the plane of
r 2 7Y-' rotation, at the nozzle exit. Thus the turning
angle is 90 ° - a,. The angle 0a of the nozzle
_ntPo_ RT ° centerline at the exit usually is the result of a
design compromise. Theoretically, better effi-
where
ciency is obtained through the use of a smaller
Cp = turbine gas (working fluid) specific
nozzle exit angle, since the rotor blading work
heat at constant pressure, Btu/lb-
is larger and the absolute flow velocity at the
deg F rotor blade exit is smaller. However, a smaller
Y = turbine gas specific heat ratio
nozzle exit angle means a larger angle of flow
R =turbine gas constant (1544/molecular
deflection within the nozzle, which causes higher
weight), ft/°R
friction losses. Design values of 9n range from
To = turbine gas total (stagnation) temper-
15 ° to 30 ° . The actual effective discharge angle
ature at the nozzle inlet, °R
al of the gas jet leaving the nozzle tends to be
fvt = turbine gas mass flow rate, lb/sec
greater than On, because of the unsymmetrical
Po = turbine gas total pressure at the noz-
nozzle shape at the exit.
zle inlet, psia
= turbine gas static pressure at the rotor A sufficiently large nozzle passage aspect
Pl
blade inlet, °R ratio, hnt/bnt, is desirable for better nozzle
DESIGN OF TURBOPUMP PROPELLANT.FEED SYSTEMS 24l
-°'-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.
efficiency. For a given nozzle height, an in- Pitch or nozzle spacing:

crease in aspect ratio can be secured by de-
creasing the nozzle pitch, Pw However, a small
dm
pitch, and consequently a large number of noz- Pa = rr-- (6-126)
Zn
zles, zn, w_th attendant increase in wall surface,
tends to increase friction losses. The determi-
where
nation of nozzle pitch thus also requires a de-
#/t = turbine gas mass flow rate, lb/sec
sign compromise. The following correlations are
established for the calculation of nozzle flow pt =density of the gases at nozzle exit, lb/ft 3
areas: C 1 =gas spouting velocity at nozzle exit,
ft/sec
_ne =nozzle exit area coefficient
Total nozzle throat area, in2: hat =radial height at nozzle throat, in
hne =radial height at nozzle exit, in
Ant = znbnthnt (6"124) bnt =width normal to flow at nozzle throat, in

bne=width normal to flow at nozzle exit, in
Total nozzle exit area, in2: zn =number of nozzles
0n =angle between nozzle exit centerline and
144 _i,t plane of rotation, deg
Ane-pICtene-Znbnehne tn =thickness of nozzle partition at exit, in
dm= mean diameter of nozzles and rotor blades,
=znhae (Pn sin 0n-tn) (6-125) in
Turbine nozzle block and inlet gas manifold

cos al +C2 cos a2)
assembly can be made of, for instance, welded
sections of forged Hastelloy C. However, the
airfoil surfaces should be blended smoothly =_(VI cos E1 +V2 cos/_2) (6-128)
between the defined contour and the sections.
,TdmN
U- (6-129)
Design of Impulse Turbine Rotor Blades 720
The function of the rotor blades in an impulse

For subsequent calculations, the following
turbine (figs. 6-55 and 6-56) is to transform a
relation will be useful:
maximum of the kinetic energy of the gases
ejected from the nozzles into useful work. Theo-
C 1 sin al
retically, there should be no change of gas pres-
tan/_1-C1 cos al-U (6-130)
sure, temperature, or enthalpy in the rotor blades.
In actual operation however, some gas expan-
sion, i.e., reaction, usually occurs. Furthermore, Axial thrust at blades, lb/lb of gas flow/sec
the actual gas flow through the rotor blades
deviates from ideal flow conditions because of C 1 sin al - V2 sin f12
Fa = (6-131)
friction, eddy currents, boundary layers, and g
reheating.
Blade velocity coefficient:
The velocity vector diagram shown in figure
6-56 describes graphically the flow conditions at V2
the rotor blades of a single-stage, single-rotor kb:V_-, (6-122)

turbine, based on tile mean diameter din. The
gases enter the rotor blades with an absolute Blade efficiency:
velocity C_, and at an angle a, with the plane of
rotation. The tangential or peripheral speed of Work transferred to blades Eb
the rotor blades at the mean diameter is U. V 1 ,?b : Kinetic energy input =C---5 (6-133)
and V2, the relative velocities at the blade inlet 2g
and outlet, differ, i.e., V, > V2, due to friction
losses. Ideally, the gas should leave the blades
Ideally, 71b is a maximum for a single-rotor
at very low absolute velocity C 2 and in a direc- impulse turbine, when the turbine velocity ratio:
tion close to axial for optimum energy conver-
sion in the blades. The forces generated at the
U cos al
rotor blades are a function of the change of
C1 2
momentum of the flowing gases. The following
correlations may be established for design cal-
i.e., when U=½C,t
culations of the rotor blades of a single-stage,
where C_ t is the tangential component of C_.
single-rotor turbine.
Tangential force acting on the blades (lb/lb of

Max. ideal _b = cos22a_(1 ' cos_
+ _b co--_-_-_fll)(6-134)
gas flow/sec):
Ft=I(c, COS a 1 + C 2 cos a2)

If there is some reaction or expansion of the
gas flowing through the blades, the relative gas
=l(v, cos _, + V2 cos f12) (6-127) flow velocity at the rotor blade outlet can be
5
calculated as
Work transferred to the blades (ft-lb/lb of gas

V 2=\/kb2Vl2 + 2gJ_?nAHi-2' (6-135)
flow/sec):
DESIGN OF TURBOPU_P PROPELLANT-FEED SYSTEMS 243
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
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
or is welded to the shrouds. The blades may be CASTING PROCESS
eitherwelded to the disk, or attached to itusing

"fir-tree"or otherdovetailshapes.
The main loads to which a rotorblade is
exposed can be dividedintothreetypes:
_ BLADE !_ z
1. Tension and bending due to centri[ugal
[orces.-The radial component of the centrifugal JOINT L.,_ DISK
forces acting on the blade body produces a cen-

BLADES WELDED TO 8LADES ATTACHED TYPICAL "FIR TREE"
trifugal tensile stress which is a maximum at the THE DISK TO THE DiSK BY TYPE TANG
"FIR TREE"TYPE
root section. As a remedy, blades are often TANG
tapered, with the thinner section at the tip, for

lower centrifugal root stresses. The centroids Figure 6-57.-Typical rotor blade constructions.
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
Sct= O.O0045721pbhbdmN 2 stresses in a uniform stress turbine disk, neglect-

ing rotor blade effects:
(6-142) Sd=O.OOOll41pd dd 2N2

log ___o_ (6-144)
Bending moment due to gas loading at the root where
section, in-lb: Sd =centrifugal tensile stress of a constant

stress turbine disk, psi
Pd= density of the disk material, lb/in s
M _ hbcct /F,_+Fa 2 (6-143) d d = diameter of the disk, in
S-2Zb _
N =turbine speed, rpm
to =thickness of the disk at the axis, in
where tr = thickness of the disk rim at d d, in
pb =density of the blade material, lb/in 3 Equation (6-144a) permits estimation of the
hb =average blade height, in stresses in any turbine disk, neglecting effects
din =mean diameter of the rotor, in of the rotor blades:
N =turbine speed, rpm

ar = sectional area at the blade root, in 2 Sd = O.O0044251W dri N2 (6-144a)
at'= sectional area at the blade tip, in 2 g ad
_i,t = turbine gas flow rate, lb/sec
zb =number of blades where
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
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)
In most impulse turbines, the number of rotors

is limited to two. It is assumed that in a single-
stage, two-rotor, velocity-compounded impulse
C]
turbine, expansion of the gases is completed in
Vl
the nozzle, and that no further pressure change NOZZLE
_lI
occurs during gas flow through the moving ROTA
blades. As mentioned earlier, the two-rotor, -- -_ .... FIRST ROTOR
velocity-compounded arrangement is best suited

C2
for low-speed turbines. In this ease, the gases
ejected from the first rotor blades still possess
considerable kinetic energy. They are, there-
fore, redirected by a row of stationary blades
into a second row of rotor blades, where addi-
tional work is extracted from the gases, which
ROT.T,ONOX
"_'_ _/,,_,= SECOND ROTOR
usually leave the second rotor blade row at a

moderate velocity and in a direction close to the S_a4 _¥4 4
axial. U
The velocity diagrams of a single-stage, two- Figure 6-58.-Velocity diagrams o{ a typical

rotor, velocity-compounded impulse turbine are single-stage, two-rotor, velocity-compounded
shown in figure 6-58, based on the mean rotor impulse turbine.
diameter. The peripheral speed of the rotor
blades at this diameter is represented by U. The
gases leave the nozzles and enter the first rotor
Total work transferred to the blades of a two-
blades with an absolute velocity Cx, at an angle
a_ with the plane of rotation. V_ and V2 are the rotor turbine, ft-lb/lb of gas flow/see
relative flow velocities in ft/sec at the inlet and
outlet of the first rotor blades, The gases leave U
E2b=-_(Cl cos al +C2 cos a2
the first rotor blades and enter the stationary
blades at an absolute flow velocity C 2, and at +C a COS a3+C 4 C0S a4)
an angle a2. After passing over the stationary

blades, the gases depart and enter the second H
rotor blades at an absolute flow velocity C 3, and =g(V, cos/31 +V2 cos/32
at an angle as. V3 and V4 are the relative inlet
and outlet flow velocities at the second rotor +V 3 cos /3a +V4 cos /34) (6-146)
DESIGN OF TURBOPUMPPROPELLANT-FEED SYSTEMS 247
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
height of symmetrical blades increases with each C1
row, roughly as shown in figure 6-55. The V1 01
effects of reheating (increase of gas specific "_l FIRST STAGE
ROTOR
NOZZLE
volume) in the flow passages must be taken into
account when calculating the gas densities at _ _' '= 2 SECOND STAGE
various sections. Equation (6-136) may be used U _ NOZZLE
to estimate the amount of reheat at each row of

V3 a 3
blades. Also see sample calculation (6-11) and
C3 SECOND STAGE
figure 6-60 for additional detail. ROTOR
In the calculations for multirow unsymmetrical

blades, the radial heights at the exit side of
each row are determined first by equation (6-140).
The radial heights at the blade inlets are then Figure 6-59.-Velocity diagrams of a typical two-
made slightly larger, approximately 8 percent, stage, two-rotor, pressure-compounded impulse
than those at the exit of the preceding row. turbine.
velocity diagrams of each stage are identical or where

Cl=C 3, C2=C 4, al=a 3, a2=a 4, etc.). The T2t = turbine gas total (stagnation) temper-
friction losses occurring in the first stage is ature at second-stage nozzle inlet,
passed on in the gas stream as additional en- °R
thalpy and increases the available energy for the T 2 = turbine gas static temperature at
second stage. Also, the kinetic energy of the second-stage nozzle inlet, °R
gases leaving the first stage is largely used and P2 t = turbine gas total pressure at second-
not entirely lost as with a single-stage turbine. stage nozzle inlet, psia
The carryover ratio rc, i.e., the ratio of the P2 = turbine gas static pressure at second-
kinetic energy actually utilized as inlet energy stage nozzle inlet, psia
by the second-stage nozzles to the total kinetic C2 =absolute gas flow velocity at first-
energy of the gases leaving the first stage, can stage rotor blade outlet, ft/sec
vary from 0.4 to close to unity. The axial dis- C3 =gas-spouting velocity at second-stage
tance between the first-stage rotor and the nozzle exit, ft/sec
second-stage nozzle, as well as the leakages rc = second-stage carryover ratio of kinetic
through the sealing diaphragm between stages, energy
should be minimized for optimum carryover. Cp = turbine gas specific heat at constant
The determination of the right enthalpy drop pressure, Btu/lb-deg F
resulting in equal work for each stage may re- y =turbine gas specific heat ratio
quire a trial-and-error approach, in view of the AH2_ 3, =isentropic enthalpy drop of the gases
effects of reheating. Or, the proper enthalpy flowing through the second-stage
drop may be estimated from previous designs and nozzles due to expansion, Btu/lb
test data. With the velocity coefficients for noz- (Ant) 2 =required total second-stage nozzle
zles and blades given by past or concurrent area, in 2
experiments, equations (6-122) and (6-136) can k_ : nozzle velocity coefficient
be used to estimate the amount of reheating. ent :nozzle throat area coefficient
Most equations established for the single-
stage turbines may be employed in the design
calculations for two-stage turbines. The follow-
ing additional correlations are available for the From sample calculation (6-5), the following
design of second stage nozzles: data have been obtained for the turbine of the
A-1 stage engine turbopump.
C2 2
Turbine gas mixture ratio, LO2/RP-1 : 0.408
T2 t = T2 + rc _ gJCp (6-149) Turbine gas specific heat at constant pres-
sure, Cp=0.653 Btu/lb-deg F
Y Turbine gas specific heat ratio, y= 1.124
Turbine gas constant, R = 53.6 ft/°R
p2( T2t_ )'-1 (6-150)
P2t: k_] Gas total temperature at turbine inlet, T o
= 1860OR
Gas total pressure at turbine inlet, Po = 640
psia
C3=kn gJCpT2t _ P3 Y Gas static pressure at turbine exhaust, pc=27
psia
= kn ,(_-cC22 + 2 gJAH2_ s, (6-151) Total available energy content of the turbine

gases, AH = 359 Btu/lb
Turbine gas flow rate, _i,_= 92 lb/sec
(Ant)2 = (6-152) Turbine shaft speed, N= 7000 rpm
l Y+l
Overall turbine efficiency (when using velocity-
/ r27 _-zT
compounded wheels), _t= 58.2 percent
4/ ,,'LT-qj In addition, the following design data are set
forth:
L-
Nozzle aspect ratio = 9.7 Solution

Nozzle velocity coefficient, kn = 0.96
(a._) Single-stage, two-rotor, velocity-
Nozzle throat area coefficient, ent: 0.97
compounded impulse turbine.
Nozzle exit area coefficient, ene = 0.95
A representative velocity diagram for this
Rotor and stator blade velocity coefficient,
turbine is shown in figure 6-58. Figure 6-60
kb=0.89
represents the temperature-entropy-enthalpy dia-
Rotor and stator blade exit area coefficient,
gram for the gas processes involved in the oper-
_b2 =0.95
ation of this turbine. The following subscripts
Chord length of rotor and stator blades,
denote the various points and processes listed:
Cb=l.4 in
Partition thickness at the exit of nozzles and
blades, tn = tb = 0.05 in
0, 1, 2, 3, 4 = Points representing inlet condi-
Solidity of first rotor blades = 1.82 tions at the nozzles; first rotor blades;
Solidity of stator blades = 1.94
stator blades; second rotor blades; and the
Solidity of second rotor blades = 1.67 exit conditions of the second rotor blades.
1', 2', 3', 4' =Points representing exit condi-
(a_) Determine the velocity diagrams and prin- tions at the nozzles; first rotor blades;
cipal dimensions of the single-stage, two-rotor,
stator blades; and second rotor blades, for
velocity-compounded, impulse-type turbine for
an ideal isentropic expansion process.
the A-1 stage engine turbopump, with about 6 0-1', 1-2', 2-3', 3-4' =Path of an ideal
percent reaction in rotor and stator blades down-
isentropic expansion process in the noz-
stream of the nozzles.
zles; first rotor blades; stator blades; and
(b_) Determine the velocity diagrams of an second rotor blades.
alternate two-stage, two-rotor, pressure- 0-1, 1-2, 2-3, 3-4=Path of actual processes
compounded, impulse-type turbine for the A-1 in the nozzles; first rotor blades; stator
stage engine turbopump, with equal work in each blades; and second rotor blades.
stage and about 3 percent reaction in the rotor 1'-1, 2'-2, 3'-3, 4'-4 = Differences along con-
blades downstream of the nozzles of each stage.
stant pressure lines, between ideal isen-
tropic expansion processes and actual
processes, due to friction losses and re-
a_ PO
heating in the nozzles, first rotor blades,
stator blades, and second rotor blades
0 '_k_- CONSTANT
To
A, /_PR_LFE UNES
s I Point " O"-Nozzle Inlet

:2: T O = nozzle inlet total temperature : turbine inlet
>_- total temperature = 1S60°R
/" II _..,"_,'_.,% / p4 o.
Po = nozzle inlet total pressure = turbine inlet

-r
........... -._-_/ b== total pressure = 640 psia
AH = overall isentropic enthalpy drop of the tur-
bine gases = total available energy content
of the turbine gases = 359 Btu/lb
T/n = nozzle efficiency = kn 2 : (0.96) 2 = 0.92
Point "1 "-Nozzle exit= First Rotor Blade Inlet

ENTROPY, S
Since about 6 percent of the overall isentropic
Figure 6-60.-Temperature-entropy-enthalpy dia- enthalpy drop AH is assumed to occur in the
gram o[ the gas processes in a single-stage, rotor and stator blades, the isentropic enthalpy
two-rotor, velocity-compounded impulse turbine drop in the nozzles
with small amount of reactions downstream o[
the nozzles. 'AH o-1' : A H (1 - 0.06) -- 359 x 0.94 : 337.5 Btu/lb
We can write: Ideally, the efficiency r/n b of a two-rotor,

velocity-compounded impulse turbine is a maxi-
mum when the turbine velocity ratio
AHo_p = CpT o I 1-(P-21U y_,]

\Po] U COS a 1
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 °
1.124 =3940 × 0.226 = 890 fps

337 5 ]o.124
=640× E1 0.65-3x-_s60.J From equation (1-129), the turbine rotor mean
diameter
= 640 x (0.722) 9._ = 640 x 0.053 = 33.94 psia
dm =720 U_0_720 × 89___- 29.1 in

From equation (6-121), the gas spouting 7r N 7rx7000
velocity at the nozzle exit
From equation (1-130), the relative gas flow
CI = kn V2 gJAHo_l, = 0.96 _/16.9 × 106 = 3940 fps angle/3, at the inlet to the first rotor blade can
be calculated:
From equation (6-122), the amount of reheat
in the nozzles
tanfll=Cl C,cosal-U
sinai _ 3940×0.906-
3940x0.42390=0.622
(1 - kn 2) C12 _ 0.08 × 15 524 000 /31 =31053 '

qnr = = 27 Btu/lb
kn22gJ 0.92× 64.4 x 778
Referring to figure 6-58, the relative gas flow
Referring to figure 6-60, the gas temperature velocity at the first rotor blade inlet
at the nozzle exit, following an isentropic ex-
pansion C 1 sin a1._3940× sin 25 °
sin fll sin 31°53 '
AHo_,, 337.5
_ 3940 x 0.423
TI, ZTo Cp -1860 0.653-1344°R 3156 fps
O. 528
The actual gas static temperature at the

Point "2"-First Rotor Blade Exit= Stator Blade
nozzle exit
Inlet
Assume that the given 6 percent reaction
27 downstream of the nozzles is equally divided
T1 = TI' + Cp -- 1344 +0.---_
qn---sr = 1385° R
between the two rotors and the stator. Then the
isentropic enthalpy drop in the first rotor blade
The gas density at the nozzle exit can be approximated as
P, 144 33.94x 144

Pl =-_-T x--_ -1385.4 × 53.6 = 0.0658 lb/ft 3 AH,_2, =P-_ x 359 = 7.18 Btu/lb
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
V 2 = V'kb 2V l2 + 2 gJTlnAH l_2, V 2 sin/32 2866×sin 25 °

tan a 2 = V cos /32 - U 2866 × cos 25 °- 890 = 0.707
= V/(0.89 × 3156) 2 + 64.4 × 778 × 0.92 × 7.18
a 2 =35°15 '
= 2866 fps
The absolute flow velocity at the first rotor

From equation (6-136), the amount of reheat blade exit
in the first rotor blades,
V2 c2-V 2 sin /32 _ 2866 × sin 252 1210 2080fps

2 1 sin a 2 sin 35o15 ' =0.57---_=
qbrl =(1 - kb )_-_+ (1 - r/n) AHI_ 2,
Point "3"-Stator Blade Exit : Second Rotor
(3156)2
= [1- (0-89)2] x 64.4 x 778 _(1-099)×7.18
• , Blade Inlet
The isentropic enthalpy drop in the stator
= 41.975 Btu/lb blades
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,
= X/(0.89 × 2080) 2 + 64.4 × 778 × 0.92 × 7.18
=33.94x 1 0.653×1385 = 1938 fps
= 33.94 x 0.93 = 31.6 psia Reheat in the stator blades
, 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
= [i- (o.89) 2] × (2080)2 + (1 - 0.92) × 7.18

T 2, = T 1 - AH,_2,/C p = 1385 - 7.18/0.653= 1374 ° R 64.4 × 778
= 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
qbs 18.53 }r
T3 = T3'+-_--= 1427 + x-_--_.
= 1456° R
r =v
Gas density at the stator blade exit
= 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
C 3 sin a s 1938x 0.574 The gas static temperature at the second

= 1.596
tan/33-C3 cos a3- U=1938x0.819-890 rotor blade exits following an isentropic expan-
sion
f13 = 57°56'
7.18
The relative flow velocity at the stator blade T4' --T3 -AH3__,/C p = 1456--- - 1445 Btu/lb
0.653 -
exit
The actual gas static temperature at the sec-

V3 = C s sin a 3 _ 1938 × 0.574 - 1312 fps ond rotor blade exit
sin /33 0,847
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
AH___,= AHa_f --7.18 Btu/Ib

144 P4 _ 144 x 27.46
= 0.0506 lb/ft a
P4 - RT 4 53.6 x 1457
The relative gas flow velocity at the second
rotor blade exit
We use an angle/34 of 44 ° for the relative gas
flow direction at the second rotor blade exits
g 4 = \/kb 2 V3 2 + 2 gJ77nAHa.. 4,
(unsymmetrical blades). The absolute flow angle
= V'(0.89 × 1312) 2 + 64.4 x 778x 0.92 x 7.18 u 4 at the second rotor blade exits can be calcu-
lated from
= 1306 fps
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
wt First Rotor Blade Dimensions (at dm)

Ant =
/ y+l The pitch or blade spacing
r 2 i?-_
gyL iJ Blade chord length Cb 1.4
= 0.769 in
Pbrl = Blade solidity - 1.82
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
We use a radial height hnt of 1.5 inches atthe

Allow 2°7 ' between inlet blade angle _)b_ri and
nozzle throat. Thus the nozzle width at the
inlet relative flow angle fl_; thus
throat
Obxrl --_1 +2°7' = 31°53' +2°7' =3't°

hnt 1.5
bnt =Nozzle aspect ratio - 9.7 - 0.1548 in
Make exit bladeangle Ob2rl equal toexit rela-
tive flow angle /_2
The number of nozzles
Ob2rl =_2 =25°

Ant 13.22
Z/l= - _
bnthnt 0.1548× 1.5 57
We select a blade radial height at the inlet
Pitch or nozzle spacing hb,rl =hne (1 x 0.08)= 1.64 x 1.08= 1.77 in
Pn - ,din _ _ x 29.1 = 1.604 in The blade passage width at the inlet

zn 57
bblrl = Pbrl sin Oblrl - tb = 0.769 x 0.559- 0.05
We allow 2 ° between nozzle exit angle On and
= 0.379 in
nozzle spouting-gas flow angle al; thus
On=a1 -2=25-2=23 ° From equation (6-138), the required total

blade exit area
From equation (6-125), the required total noz-

zle exit area, 144 w_ 144 x 92
= --
Ab2rl p2V2Eb2 0.059 x 2866 x 0.95 82.5 in 2
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
zn 53.75 bb2rl = Pbrl sin Ob2rl - to = 0.769 x 0,443 - 0.05
bne = hne - 57 x 1.6_- 0.576 in = 0.291 in

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
Stator Blade Dimensions Second Rotor Blade Dimensions
Pitch or blade spacing Pitch or blade spacing
Blade chord length Cb 1.4 Blade chord length Cb 1.4
Pbs - Blade solidity - 1.94 - 0.721 in Pbr2 = Blade solidity 1.67 0.888 in
From equation (6-140a), the number of the

From equation (6-140a), the number of blades blades
ndm_ _× 29.1 vdm _ v×29.1=i09

- 127
Zbs = Pbs 0.721 Zbr2 - Pbr2 0.838
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
8b2s=a3 =35 ° Ob2r2 =/_4 =44°
From equation (6-149), blade radial height at From equation (6-149), the blade radial height
the inlet at the inlet is
hb,s= l.08×2.52=2.72 in hbir 2 = 1.08× 2.87= 3.10 in
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
From equation (6-138), the required total _: 0.582

blade exit area
A-1 Stage Engine Turbine (Single-Stage, Two-

144 #t 144 × 92
Rotor, Velocity-Compounded Impulse Type)
Ab2r2 -P4 V4eb2 - 0.0506 x 1306 x 0.95 = 211 in 2
Design Summary
For velocity diagrams at mean diameter din,
Combining equations (1-139) and (1-140a), we
see figure 6-58.
obtain the blade radial height at the exit
U = 890:
a 1=25°; fl_=31°53'; C_=3940fps; V1
Absr2
=3156fps; a:=35°15'; fl_=25:; C 2=2080
hbsr2-_d m sin Obsr2-Zbtb
fps; V 2=2866 fps; a3=35°; fl3=57°56';
211 C 3 =1938 fps; V 3 =1312 fps; a4 =86°55';
=_x29.1 x 0.695- 119× 0.05 =3"66 in _4=44_; C4:908fps; V4=1306 fps
Isentropic enthalpy drops:
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
(_b) Two-stage, two-rotor, equal-work, pressure-

890 (3940 × 0.906 + 2080 x 0.817
compounded impulse turbine. (For velocity dia-
+ 1938 x 0.819 + 908 × 0.055)
grams, see fig. 6-59.)
32.2 x 778x 359
From prior trial-and-error calculations, the
= 0. 683 following isentropic enthalpy drops resulting in
(approximately) equal work for each stage were
From equation (6-148), the turbine machine obtained. We assume a stage carryover ratio
efficiency rc = 0.91.
First-stage nozzles: Point "2"-First-Stage Rotor Blade Exit=Second-

Stage Nozzle Inlet
AHo_I, = 50%; AH = 0.5 × 359= 179.5 Btu/lb From equation (6-135), the relative gas flow
velocity at the first-stage rotor blade exit
First-stage rotor blades:
V 2 : \/kb 2V12 + 2 gJT?n'_H1_ 2,
AH__ 2, = 3%; AH : 0.03 × 359 = 10.75 Btu/lb
: V_(0.89 × 1784) 2 + 64.4 × 778 × 0.92 × 10.75
Second-stage nozzles: = 1736 fps
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
C 1 = kn \/2 gJAHo_ f = 0.96 × 223.8 × vJl-7-9_5.5

Point "3"-Second-Stage Nozzle Exit = Second
= 2880 fps Rotor Blade Inlet
From equation (6-151), the second-stage noz-
We use a value of 25 ° for the spouting-gas zle gas-spouting velocity
flow angle a l. For optimum efficiency, tile

peripheral speed at the rotor mean diameter C a = k n \/rcC22 + 2 gJAH2_ 3'
= 0.96 V'0.91 × (1070) 2 + 64.4 × 778 x 158
-- cosalC1 -- 0.906×2880 _ 1308fps

2 2 : 2880 fps
Since C 3 = C_, the remainder of the second-

Using equation (6-130), the relative gas flow
stage velocity diagram is the same as that of the
/3_ at the first-stage rotor blade inlet can be
calculated as first stage, i.e., a3:a I =25°; J33 =]31 =43°8';
V3=Vl=1784 fps; a4:a2=86°40'; /_4=/_2=38°;
C 4:C 2=1070fps; V 4=V_=1736fps.
C, sin al 2880×0.423
tan /31 = C1 cos al- U - 2880 × 0.906 - 1308 = 0.936
From equation (6-129), the turbine rotor mean
131 =43°8 ' diameter
The relative gas flow velocity at first-stage

rotor blade inlet dm=720U-720x1308 42.7 in
_N _r × 7000
Yl C1 sinai 2880×0.423 1784 fps From equation (6-147), the combined nozzle
sin 131 0.683 and blade efficiency
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
=0.78 Radial loads on turbopump bearings may re-

sult from one or more of the following sources:
The turbine machine efficiency is assumed to (1) Weights of parts such as shafts, pump
be the same as that used in design (a._): impellers, turbinerotors, gears
(2) Centrifugal forces du_:, to unbalance of
qm = 0.852 these rotating parts
(3) Forces due to inertia, resulting from rapid
From equation (6-148), the overall turbine acceleration
efficiency (4) Resultant radial forces on the impeller
due to nonuniform pressure distribution
r/¢ = rlnbr/m = 0.78 x 0.852 = 0.664
in the discharge volute of the pump
(5) Tangential or torque forces induced by the
A-I Stage Engine Alternate Turbine Design
gears
Summary (Two-Stage, Two-Rotor, Pressure-
Compounded, Impulse Type)
Thrust loads on turbopump bearings may re-
For velocity diagrams at mean diameter din,
sult from one or more of the following sources:
see figure 6-59.
(1) Weight of rotating parts mounted on a
U = 1308 fps:
vertical shaft
a1=25°; _1=43°8 '' C1=2880fps; V 1=1784
(2) Unbalanced axial thrust of the pumps
fps; a_=86°40'; _2=38°; C2=1070 fps;
(3) Resultant axial thrust on the turbine rotor
V 2 = 1736 fps; a3 =25°; f13 =43°8'; C3=2880
blades
fps; Va=1784 fps; a4=86°40'; _4=38°;
C4=1070fps; V 4=1736fps
Isentropic enthalpy drops: For the turbopumps of liquid rocket engines,
First-stage nozzles, AH__,, = 179.5 Btu/lb high-speed ball and roller bearings are used
First-stage rotor blades, AH,_ 2, = 10.75 Btu/lb almost exclusively. A typical two-bearing de-
Second-stage nozzles, AH2_2 -- 158 Btu/lb sign is shown in figure 6-7. A ball bearing
Second-stage rotor blades, AHz._4, = 10.75 carries both radial and thrust loads. It is paired
Btu/lb with a roller bearing which carries only radial
Working efficiencies: loads, however, of a higher magnitude. A typi-
r/t= 66.4%; r/n = 92%; rlnb= 78%; _?m= 85.2% cal three-bearing arrangement is shown in figure
Mean diameter of nozzles and blades: 6-63. The shaft radial loads are carried by a
dm= 42.7 in single roller bearing at the turbine end and by a
Comment: The overall efficiency of the roller and a ball bearing on the pump side. The
pressure compounded turbine is higher than that ball bearing also absorbs the thrust loads. As a
of design (a). However, a relatively large dm is rule, the shaft thrust loads in a turbopump are
required (weight, size). carried by a single or dual bearing located at
one end of the shaft. Thus loads from thermal
expansion or contraction of the shaft are avoided.
6.6 DESIGN OF TURBOPUMP BEARINGS,
Bearing design data with regard to load-
SEALS, AND GEARS
carrying capacity, operating speed, and service
life are usually furnished by the manufacturers.
Turbopump Bearing Design
The useful life of a bearing is dependent upon
A turbopump shaft is supported by two or more its speed and load, and may be expressed by the
bearings. The loads on the bearings are the correlation:
- lira _
proportionally with the increase of shaft design

{ Rated speed, rpm ._
Life, hours = gb \Actu-a--1o--p-er_p-_'-d, rpm/ speed. Thus the required bearing DN value
rapidly increases for high-speed turbopumps. As
[ Rated capacity, ib ,_4 a result, especially for liquid hydrogen applica-
×_Actual working load, lt)] (6-153)
tion, the turbopump rpm is often determined by
the DN limits of the bearings.
where Kb = design factor usually furnished by
A most important bearing design consideration
manufacturer.
is the expected operating life of the rocket en-
If a bearing is subjected to both thrust and
gine. The bearings must have adequate statisti-
radial loads, the two can be combined into a
cal probability of conforming with this require-
single equivalent radial load:
ment. A generally accepted life rating for ball
and roller bearings is the "B-10 life." The term
P = R + xA (6-154)
denotes the operating life (hours) of a _popula-
tion" of bearings at a given load and speed, at
where the expiration of which statistically 10 percent
P = equivalent radial load used for bearing of them will have failed. Of course, in actual
selection, lb rocket engine operation, component reliability
R = actual radial load, lb must be much higher. Bearing life at a given
A = actual thrust load, lb load and speed varies inversely with reliability.
x = design coefficient usually furnished by For instance, the B-1 life (99 percent reliability)
manufacturer is one-tenth of the B-10 life (90 percent relia-
Rocket turbopump bearings quite commonly bility), or one-fiftieth of the B-50 life (50 percent
are cooled and lubricated by the propellants reliability). Therefore, turbopump bearings are
pumped. They are usually operated at very high generally designed for a B-10 life of at least 100
"DN" values, a parameter which is the product hours. This corresponds to a B-1 life of 10
of the bearing bore D (millimeters), and the bear- hours, or a B-0.1 life (99.9 percent reliability) of
ing rotative speed N (rpm). 1 hour, the latter by order of magnitude being the
Propellant lubrication has the advantage of life the bearing most likely will actually see.
eliminating an additional lubricant supply sys- For critical applications, an even higher life
tem, and of simplifying bearing sealing problems. rating may be selected.
The following are important design considera- Figure 6-61 presents the centrifugal load DN
tions for propellant-lubricated bearings: limits in terms of 10, 100, and 1000 hours of
(1) Characteristics of the propellants, such B-10 life for a typical ball bearing design (extra-
as thermal stability, operating temper- light series).
ature, chemical inertness, viscosity. The stress-limiting DN values of roller bear-
(2) Compatibility of the bearing materials ings are much higher than for ball bearings;
with the propellants. The application however, it is extremely difficult to control the
of certain high-strength alloys is some- temperature rise in a roller bearing, if the DN
times limited by the propellants used. value is above 1.5 × 106, due to excessive cage
The "DN" rating is convenient when selecting slip. Generally, rocket turbopump bearings have
high-speed ball or roller bearings. As the rota- been successfully operated at DN values up to
tire speed of a bearing increases, contact fatigue 1.5 × 106. Limited test information indicates
of the outer race caused by centrifugal loads of possible satisfactory operation at 2.0 x 106 DN.
the balls or rollers may cause failures. In addi-
tion, bearing contact speeds will result in non- Dynamic Seal Design
rolling phenomena with attendant failures caused
by overheating. Through proper selection of the The principal dynamic, i.e., rotating seal
bearing geometry, these problems can be mini- types used in liquid rocket turbopumps are the
mized, and the DN rating increased. Note that labyrinth, face-riding, and shaft-riding seals.
for a given horsepower rating, the shaft size Satisfactory seal operation depends upon good
based on allowable stress does not decrease design which considers many factors, including
_.: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----
_.o#-------------J_/--\--4_,_. _ ._ I I I I -_ _ --- ' _.3u-=_.
",., .c__..c- 2:.':. ... 't, , I .....
Io ZO 3o 40 50 60 70 8090 _
o. 200 _0 400 =X>o
BEARING BORE SIZE, D mrn
Figure 6-61.-Bali bearing centrifugal load DN limits.
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.
Qe =leakage rate, in3/sec The shaft-riding seal (fig. 6-62) consists of a

Ac =seal clearance area, in 2 seal housing, a seal retaining plate, and several
Aps = pressure differential across the seal, seal segments. The segments form a ring around
lb/in 2 the shaft and are held against it by garter springs.
-PUMP CASING
HOUSING
WEARING RING ....,..\\_ /-- PUMP IMPELLER
___ _SHAFT
-_-_- CC SHAFT
LABYRINTH SEALS
_ ,.AL RETAINING PLATE
/-_ TURBOPUMP
_-TURBOPUMP CASING /CASING
FLOATING .\'..x>,_" P- SE;TL HO_SING
SEAL WASHER_\_\_/ST ICSE L ____GARTER SPR,NGS
_HER_/SEAL HOUSING
SHOULDER RING_ BELLOWS
....._/,/_
_ ..... SEGMENTS
_ ___cu-_L_L - -- _ SHAFT
FACE-RIDING SEAL SHAFT- RIDING SEAL
Figure 6-62.-Principal turbopump dynamic seal types.
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
Design of Controls and Valves
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.
Open-Loop Control Closed-Loop Control
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 _
then generates signals to correct for any devia-

tions. The main system thus does not require
precise calibration for a specific set of condi-
tions. Unlike open-loop control, closed-loop
control depends on sensing absence or presence
of an error to maintain a desired condition or to
Figure 7-1.-Schematic of a typical closed-loop
bring about a correction. In general, the objec-
control system.
tive of closed-loop control is to minimize errors
during operation and reduce system sensitivity (electric summing junction and amplifier), and a
to environmental changes and changes in compo- controller (gas generator flow control valve), the
nent characteristics. It is applied to areas such command reference input r is compared with the
as engine-thrust control and/or throttling, pro- sensor feedback b. The controller then manipu-
pellant mixture-ratio control, and thrust-vector lates • g in response to an error signal e from
control. the computer. Ideally, r should be in linear pro-
For rocket engine application, closed-loop portion to Pr and b to Pc, save for constants
control systems usually employ one or a combi- required to convert one physical quantity into the
nation of the following modes of operation: other. However, this ideal condition is difficult
1. Simple "on" and "off" type.-(Example: to attain because of the dynamic characteristics
pressure switch/valve combination for tank pres- of the pressure transducers. These characteris-
sure control.) tics are influenced by physical properties such
2. Proportional type.-Employs a continuous as mass inertia, fluid compressibility and vis-
control signal which is proportional to the error. cosity, and frictional resistance. Instead of r
(Example: transducer output for chamber pres- being directly proportional to Pr, the two param-
sure control.) eters are actually related through a differential
3. Derivative type.-Employs a continuous equation which represents the dynamic behavior
control signal which is a function of the error of the elements involved. The same is true for
and its time derivative(s) (rate of change). This the feedback b and the controlled variable Pc. It
is principally used when systems stability is is also applicable to other systems components.
critical. (Example: thrust vector control system Hence, the analysis of a closed-loop control
with phase lead.) system usually involves the solution of sets of
4. Integral type.-Employs a continuous sig- often complicated differential equations.
nal which is proportional to the cumulative inte- Refer again to figure 7-1, where Pc is the
gral of one or more errors. (Example: two flow- controlled variable, _¢g the manipulated variable,
meter outputs for mixture-ratio control.) e the error signal, b the feedback, r the reference
Closed-loop or feedback control systems are input, and Pr the desired value. A, B, C, and D
essentially dynamic systems. Their design symbolically represent the dynamic relation
characteristics may be analyzed according to the between input and output of the respective com-
basic laws of physics. Figure 7-1 shows a typi- ponents. The following terms representative of
cal example. Its function is to maintain the the differential equations for this closed-loop
controlled variable Pc equal to the desired value control system can be written:
Pr, by manipulating the variable wg. Maintaining
Pc equal to Pr is assumed to maintain the in- r =Apt e =r- b Pc =C_vg
directly controlled quantity F. In a typical (7-1)
turbopump fed engine control system, Pc would b = Dpc ¢¢g = Be
be the combustion chamber pressure, maintained
equal to a fixed reference pressure Pr by means The solution of these equations in combina-
of a valve controlling the gas generator propel- tion with a systematic experimental program will
lant flow wg. F then would be engine thrust, suffice to analyze the dynamic performance of
which is indirectly maintained at a desired value. the system.
In this control system which consists of a The continuous corrective action of a closed-
sensor (chamber pressure transducer), a computer loop control system may promote dangerously
DESIGN OF CONTROLS AND VALVES 265
unstable operation when control elements or Engine System Cutoff Control

components are employed having high gain and
Rapid and safe engine shutdown, during nor-
significant response lags. An unstable control
mal operation as well as in an emergency, is
system is one that is no longer effective in main-
desirable for minimum and repeatable cutoff
taining a variable at its desired value. Instead,
impulse, and to enhance reliable systems opera-
large divergent oscillations may set in. The
tion. The cutoff sequence usually consists of
requirements for control accuracy and for stabil-
shutoff of subsystems power (gas generator,
ity are often difficult to combine. Higher accu-
etc.); shutoff of main chamber power; and, in
racy requires high amplification; i.e., high gain.
case of test firings, postfiring securing (purges,
The high amplification results in overshoot dur-
flushes). As a rule, the propellant valve-closing
ing corrective action, thus promoting various
sequence is adjusted to provide a fuel-rich cut-
degrees of system instability. However, through
off in the main combustion chamber. This pre-
appropriate means of compensation, such as
vents damaging temperature spikes and results
"anticipatory" phase leads (time derivatives), it
in smooth and rapid thrust termination.
is possible to obtain a high gain control system
with satisfactory stability. Additional informa-
tion on compensation will be presented in con-
nection with thrust-vector control. Engine Main Stage Duration Control
Important consideratious governing engine

duration have been discussed in section 2.i.
7.2 BASIC LIQUID PROPELLANT ROCKET The signal for engine in-flightcutoff, unless it
ENGINE CONTROL SYSTEMS is the result of a malfunction, will be supplied
by the vehicle and fed directly into the cutoff
Most engine systems require several or all
control system discussed in a preceding para-
of the basic control systems summarized in the
following paragraphs. Typical applications are graph. For lower stages, where optimum utiliza-
tion of the propellants is desired, a tank low-
found in chapter Ill, for the A-I, A-2, A-3, and
A-4 propulsion systems (figs. 2-10, 3-3, 3-6, and level sensor is often employed. In final stages,
3-9). where precise cutoff velocity is essential, an

integrating aceelerometer or equivalent device
will signal cutoff.
Engine System Start Control
The prime objective of a start-sequence con-

Engine System Safety Controls
trol is to bring the engine system safely from
start signal to main-stage operation. A typical Special monitoring devices, such as combus-
sequence may consist of systems preconditioning tion stability monitors for detecting combustion
(purging, chilldown); application of start energy, instability, gas generator overtemperature sen--
if required (start tanks, turbine spinner); and sors, or turbopump overspeed trips, are frequently
introduction and ignition of the propellants in the employed to prevent undesired or unsafe condi-
main combustion chamber. Secondary sequences tions by effecting prompt, automatic, nonhazard-
may be required for certain subsystems such as ous, fail-safe system shutdown during all phases
the gas generator system. A reliable engine- of engine operation. In addition, most engine
start sequence is maintained through interlocks control systems are designed so that an interrup-
and by monitoring each functional step of engine tion of electrical power supply will cause the
operation during the start transient. The system to shutdown safely. (For certain mis-
propellant-valve opening sequence is set to sions it may be desirable to switch to an emer-
effect either an oxidizer-lead or a fuel-lead start. gency power source, or prevent shutdown by
This is usually dictated by propellant type and mec_Janical latching, for continued operation.)
chamber ignition and cooling methods. Figures Mechanical and electrical interlock devices are
2-11, 3-5, 3-8, and 3-11 present typical engine extensively used in the control system to assure
system start and cutoff sequences. the reliability of the safety control systems.
Propellant Tank Pressurization Control Most propellant tank pressurization control

systems are of the closed-loop type.
Various propellant tank pressurization sys-
tems have been discussed in chapter V. The
Engine System Control Calibration
design requirements for the control of these sys-
tems must consider- The systems described in the preceding para-
(1) Means to maintain the required tank pres- graphs require proper adjustment and calibration
sure level within an allowable range for desired engine operating characteristics and
during all phases of vehicle and engine performance. This includes the setting of timing
systems operation, including steady- devices, pressure switches, position switches,
state engine mainstage; start or dynamic and the sizing of orifices. The correct values
throttle transients; and vehicle coasting for each of these are verified during engine cali-
periods between restarts. bration and checkout firings. Of the orifices,
(2) Effective safety devices such as pressure some are placed in propellant lines for perform-
relief valves to prevent overpressuriza- ance parameter calibration. Others are used in
tion and rupture of the propellant tanks. pneumatic or hydraulic lines as timing and re-
(3) Compatibility with other subsystem con- stricting devices. Specific orifice applications
trols, such as propellant-utilization- for thrust and mixture ratio control will be dis-
control and thrust-control systems. cussed in sections 7.3 and 7.4; orifice design
Figure 7-2.-Control ori[ice locations and sizes o[ a typical engine system.

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,
7.4 PROPELLANT-MIXTURE-RATIO AND

PROPELLANT-UTILIZATION CONTROL
The significance of propellant mixture ratio

C_lu_m and its control have been discussed in section
_¢_E
,NPUT
, _ o _mu T
2.1. The principal reasons for mixture-ratio
control are recalled:
Optimum engine performance (important)
Complete propellant utilization; i.e., mini-
mum residuals (most important)
Both goals are closely interrelated and essen-
[_ C_EC_m_ i,,_._,_ tially inseparable.
Figure 7-3.-Main-stage thrust throttle control
loop Ior the ,4-4 stage engine.
Open-Loop Mixture Ratio Control
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
and is subsequently maintained at level through

FROM THROTTLE CONTROL _II£TL_TO_ 7 FI_:_I
OXIDIZER a topping line. The mass of both propellants is
TANK / FUEL
_'ANK
._, _ / i determined from on-the-spot temperature and

N OXIDIZER
T _ MECHAN ICdI, L
CO_ ROL LVE COOPLING ambient pressure readings while the tanking
procedure is progressing.
2. Use o[ adjustable, rather than [ixed, ori-
[ices in one or both propellant lines.-As close
to vehicle takeoff as possible, and as a function
of tanked weight and temperature readings, a
hand or remotely ground-controlled prestart-ori-
flee adjustment is made. This method is usually
TO NO. I TO NO Z TO NO. I TO NO Z confined to noncryogenic fluids.
THRUST CHAMBER THRUST CH*_*M_R THCtUST CHC_MBER THRUST CHAMBER
For systems where engine operation closely

Figure 7-4.-Schematic of the propellant control follows that obtained during final calibration,
system for A-4 stage engine start, cutoff, remarkable accuracy of targeted mixture ratio
throttle and mixture ratio control. and thus propellant utilization can be obtained
DESIGN OF CONTROLSAND VALVES _9
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
computer. The resulting error signal, (MR)e, is

fed to the mixture ratio control oxidizer valve mXTU_f _A_
vernier position actuator, which forms a link in

the mechanical coupling between the two main
propellant control valves, as shown in figure 7-4.
The oxidizer flow rate is thus modified to elimi-
nate the error. In high-thrust turbopump-fed
[CEC_e,( SUVltNG
engine systems such as the A-2 stage engine, _UV.C'nON
_' eLecrllc lm_t.,1,1Ei
where the propellant valves are independently

actuated, the system propellant mixture ratio Figure 7-5.-Propellant mixture ratio control
control can be accomplished by varying the main loop for the A-4 stage engine.
mixture ratio to assure simultaneous propellant

depletion. However, by programing a mixture
ratio in favor of the heavier component during the
early portion of flight, and then switching it in
favor of the lighter one, the accelerated vehicle
mass is reduced faster. Also, mixture ratio may
°° ] be programed to provide a higher thrust level
during the steeper portion of a trajectory. This
provides a better thrust-to-weight ratio in the
ELECT't; AMPL=FI(R
presence of gravitation, with attendant velocity

Figure 7-6.-Propellant utilization control system increase benefits. These methods, possibly in
[or the A-4 stage propulsion system. combination, may substantially increase stage
payload capacity, since the effects of mixture
propellant utilization control system is made ratio on performance (Is) are usually small within
a reasonable range (see table 7-1). Optimization
narrow as compared to that of the mixture-ratio
can readily be made with the aid of an electronic
control system, because propellant residual
computer program. In a number of applications,
errors may be expected to develop slowly; i.e.,
initial tanking errors can be corrected over the programed mixture ratio control without PU con-
trol, i.e., open loop mixture ratio control with
entire duration of engine operation.
The sensors used in the vehicle tanks may PMR, may give best results, simultaneously
serve additional purposes. In combination with reducing complexity.
Valves suitable for mixture ratio control will
suitable ground equipment, they may permit an
be discussed in section 7.8.
automatically controlled loading, high-level limit-
ing and topping procedure. In static firings and
flight, they may serve as redundant low-level TABLE 7-1
sensors to initiate engine cutoff. For such a
Flow rates
complete system, the term "propellant manage- Mixture i
ment system" has come into increased usage. ratio, Thrust NPSH Is Oxi-
Apart from throttle valves placed in the main O/F Fuel dizer
propellant lines, bypass lines have been suc-
cessfully applied to vary mixture ratio. Here, a 3hange, percent +10 +11 +12 -1.3 +4 +14
-10 -11 -12 +I.3 -4 -14
line is tapped off the pump outlet and ducted
back to the pump inlet. A servo valve, possibly
supported by an orifice, can be varied so that
the bypass flow is adjusted from no flow to full
bypass flow.
General Design Considerations
The implementation of closed-loop propellant-
utilization control through mixture-ratio control The precision with which a desired mixture
is a major vehicle-to-engine interface area. The ratio is obtained or maintained is affected con-
requirements or criteria will usually be estab- siderably in open-loop systems, and to some
lished by the vehicle builder and/or user. Close extent in closed-loop systems, by the following:
coordination between engine and vehicle designer (1) Instrumentation accuracies (in particular,
is essential. flow and tank-level metering)
A closed-loop mixture-ratio- and propellant- (2) Machining tolerances of orifices
utilization-control system may not only be used (3) Operating tolerances of regulators
for accurate maintenance of a fixed mixture ratio (4) Temperature influences on orifices and
but it also has the potential for programed mix- regulators
ture ratio control (PMR). Here, the mixture ratio (5) Density tolerances of the propellants, as
is varied during flight, either continuously or in a function of temperature and of purity
steps. It must be kept in mind that the average (composition according to specifica-
mixture ratio still must be equal to the tanked tions; contamination and dilution)
DESIGN OF CONTROLSAND VALVES 271
(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
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
toward the same end. Malfunction safety circuits

are included to effect engine cutoff in the event
V_HOJ
C_OANCE
CO_MANO of erratic operation.
_FE_ENC_
A typical schematic for a thrust vector con-

trol system using electromechanical actuators is
shown in figure 7-8. Here, the actuator is pow-
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,
N"I TO_OU[ MOTOe

and transfer functions by substituting the equiv-
alent hydromechanical time constants.
Table 7-2 may be found valuable by those
NOZZLE • B
D4ERJVATIV_ PlES$_RE _TE_ - _IECH

who wish to familiarize themselves with some
FEEDBACK LEAF SPRING FEEDBACX
fundamentals in this field. Detail on the design
DERIV&TIVE LO_D DIRECTNDN O$"
PE'E$ SURE P_STON FEEDBACK of servovalves will be found in section 7.11.
IS,
OL AT _0_ P'_ST_ON i_OT _',J
_ 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
and voltage output for voltage input networks.

Figure 7-10 shows a simple form of a current
output for voltage input network. Find the anal-
ogous hydromechanical network.
Figure 7-9.-DPQ valve wiLh actuator.
and Pa decreases, forcing the actuator piston to Solution

the left to apply the desired load force. Attached The transfer function for the electrical net-
to the actuator piston rod is a tapered extension work is
which acts upon the mechanical feedback link-
age, including a roller and a spring. The me- i I+RCS
chanical feedback attaches to the torque motor. --:-- Amp/volt
V R
The pivot point of the valve flapper becomes the
where
error torque summing junction. Note that the
i = electrical current (amps)
nozzle jets also have a feedback effect. The
time derivative of the actuator motion, i.e., the V = voltage (volts)
R : resistance (ohms)
hydromechanical compensation, is obtained
C : capacitance (farads)
through action of a derivative load pressure pis-
S : LaPlace transform operator (= j_o for sinus-
ton. This piston is affected by the same pres-
oidal forcing functions)
sure differential that acts upon the actuator
From table 7-2, we obtain the equivalent
piston; i.e., by the load pressure. However, by
hydromechanical parameters for i, V, R, and C.
inserting an isolation piston and permitting flow
The new transfer function then is
through an orifice bypassing the derivative pres-
sure piston, the pressures affecting the latter
A 2
can equalize. The degree of this effect is a
function of the actuator pressure differential and Q _1+ C_xPK)S
its rate of change and of the bypass orifice size AP 1
(shock absorber principle). As seen in figure Cx
7-9, the derivative load pressure piston acts
upon the valve flapper when displaced. Thus it
provides the required time derivative of the
actuator motion for compensation.
As has been seen, it is possible to provide
compensation in thrust-vector-control systems by o
either electrical or hydromechanical means, the V
latter now being often preferred for actuators.
Conceivably,
converted
networks.
other control systems could be
from electrical
The analogies
to hydromechanical
between the differen-
1 G
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
TABLE 7-2.-Electrical-Hydromechanical Component Analogies
Electrical quantity Describing Hydromechanical Describing

Analogy Remarks
or component equation component or quantity equation
V, voltage drop, volts ......... ' V = V A - VB AP, pressure drop, psi AP: PA - PB V _AP
i, current coulombs/sec ....... dq Q, flow, in3/sec dV i_Q

_ =-_ Q = d--t
V: volume
q = coulomb
charge
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
V Parabolic flow curve

_=_-
Ap Q_Cx_,A-'_-_ R _ 1._
linearized about
CSIIE3 _CxA P CX
operation point
Q ---...,lllb
Resistor. ohms inS/sec

Orifice, --
Piston mass not

V -----_ .... ,, p AP2 y(AP)dt b
Q:"M- "Ap" negligible
t
------D- Q
Inductor. henries Ap= piston area, in 2

M =piston mass
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 °
would be required from all engines to offset the

Engine-to-Vehicle Interfaces With Actuator
misalinement.
Systems
For larger (looser) alinement tolerances, the
trim deflection would be further increased. Even
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-
alined with respect to the vehicle attachment ductions.
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
Q
_C-_ /--VECHICLE THRUST MOUNT CENTERLINE
//,4i7/// /
_7 ._// / /-_..,_ cE.TE_ ,
-/- ,'.t-_.._q
EAANCDH
yIA_/ Pt:CAHE ) F_ __ _ GIMBAL "ARING [
P_llel to T_IJ_ v_14m
\ \ \ 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
"_ x'x- installation of the engine then simply consists

of attaching it, observing the engine logbook
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
_L_ is advisable to specify verification of engine
Figure 7-11.-Piston bypass hydromechanical alinement following transportation to the launch-

network. ing site.
methods (load cells) are used. Tile optical or Actuators, Loads

cold alinement establishes the geometrical loca- Actuators are usually of the hydraulic-piston
tion of the thrust vector in the shop, through type. Hydraulic-rotary, electromechanical, and
finding the centers of nozzle throat and nozzle pneumatic turbine-driven types have also been
exit, and alining their connecting line perpen- investigated.
dicular to the gimbal plane. A simple plumb Engine gimbal actuators are attached to the
attached to the injector center may be used in engine at one end, and to the vehicle at the
support of this operation. Subsequently, during other (fig. 7-13). They may be procured by the
engine firing, this measurement may be confirmed vehicle builder or by the engine builder, and
dynamically using side load cells in lieu of gim- must be properly dimensioned. The attachment
bal actuators. As a rule, after a few engines points at either end must be capable of absorb-
have been alined in this manner, experience will ing the forces encountered with an adequate
permit meeting specification by optical means reserve. As a rule, two actuators are required
alone. The vertical alinement can be simply for each engine. Together, they permit deflec-
documented as the eye-to-eye distance of the tion of the engine in all directions. It is impor-
actuator attach points, or as the line through tant to note that if the maximum deflection
two index points (fig. 7-12). Lateral dimensions effected by an individual actuator, for instance,
can be marked in a suitable manner at or on the is 7 ° , the combined maximum angle ("corner
gimbal bearing face. If the mating vehicle face deflection") of a pair is approximately 10 °. Inlet
had been properly alined to the vehicle axes, ducts, flex lines, gimbal bearing, and possibly
where filling of the nozzle takes a longer time

,_-_ ACTUATOR_
-- -- ..... /-" LONG STROKE during buildup than with shorter ones. Unless
\
AOTUA.OR
--\
SHORT
".CON,,OORAT,O
STROKE
these loads _-j_ ACTUATOR
can be eliminated or at least re-
duced, they must be considered for the design of
R ,\,.\\ actuators and attach points at both ends (notifi-
cation of vehicle builder).
Crosstalk and Spring Rate

Since engine and vehicle designer are not
- - I I entirely free regarding actuator installation, a

OlMBAL CENTER_ I I --I [i 1 I situation may exist wherein motion of the actu-
ator in one plane affects the other actuator in its
Figure 7-13.-Engine actuator installations.
plane. This is referred to as "crosstalk." If it
is excessive, control instabilities may develop.
other components affected, must be able to Therefore, close coordination between engine
"take" this deflection. If their capability is designer and vehicle builder to minimize cross-
limited, proper gimbal restriction (stops or snub- talk is essential.
bers) must be provided (circular gimbal pattern The actuator must be able to translate its
instead of square). motion without delay into engine deflection. If
Selection and design of the actuators is based the control loop formed by actuator, engine struc-
on the gimbal forces required. In a typical case, ture, and vehicle thrust structure is "soft," i.e.,
the actuator force may be 25 percent of the en- if it has a low spring rate (lb/in), the engine
gine thrust level. The force is determined by does not react promptly to an actuator motion as
considering the following: called for by the guidance system. A delay with
Inlet duct reactions subsequent overshoot and continued oscillation
Flexible service line reactions may result. The natural frequency of this oscil-
Gimbal bearing friction lation is a function of actuator stroke per degree
Heat shield reaction (if any) engine deflection, feedback gains, and compen-
Correction for misalinements sation network parameters.
Aerodynamic loading (if any) Figure 7-13 shows two typical engine actuator
Vehicle acceleration effects installations. The short-stroke configuration has
Inertia of gimbaled mass the advantage of high spring rate and high stiff-
Miscellaneous minor effects ness, and of compactness. However, the effec-
It is stressed that the engine design should tive gimbaled mass is high, requiring larger
reduce these forces to a minimum for smallest actuators and a more powerful hydraulic system.
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.
loads in excess of those occurring during normal

gimbaling can be generated by the thrust chamber Hydraulic System
itself. This is especially true for high expan- Until other means are available, a hydraulic
sion area nozzles being developed at sea level system is probably required to power the engine
(for engines designed for altitude operation), actuators. Its basic elements are:
Hydraulic pump
Reservoir (low pressure, or "sump")
Accumulator (high pressure)
Servovalve
Actuators
Feedback (electric or mechanical)
Lines, check valves, filters, connectors,
instrumentation
If continuous hydraulic power is required

prior to engine start, such as for recirculation of
the hydraulic fluid or for gimbal tests, an elec-
trically driven auxiliary pump is also provided. Figure 7-I4.-Engine actuation system schematic
In most instances, the auxiliary pump will be (hydraulic).
operated until vehicle liftoff only, and can,
therefore, be ground powered. For upper stages,
the accumulator will then provide, for a limited
time, the hydraulic power required during staging
1 1 ......
PulIP SuCI_N '--'=-'
and turbopump buildup following its unlocking.
Since some of these components will be part
of the engine system, while others are stage
...................o....
mounted, an important vehicle/engine interface , II OUI_-'K Lq$CONN|C_._It|$ $ SUPPI.y
exists. Through an auxiliary drive shaft, the

main hydraulic pump may be driven from the ;._
j ll] : .......................
HIGH-PrESS RELIEF VALVE
13 LON-F'IF|S IIELIEF VALVE

engine turbopump. It is connected to the other _
i
12
14
OUICK D_SCONNECI-[LOW
NITIINEN CHA|GING
PI|SSI
VALVE
hydraulic equipment and to the actuator through 15

I_
I_IIMAL
ILEED VALVE
SWITCH
high-pressure lines, several of which must be _10
11
PIEr $$ TRA NSD_JCEJl
TEIIIIp TI l N _'_DU_EEII
flexible. These other elements may be mounted xt 4 33 IESEIIVOII Po£JrlON TRANSDUCEI
on the vehicle at the expense of longer lines

Figure 7"15.-Accumulator-reservoir schematic.
which also must cross the gimbal plane and must
therefore be flexible. Or, they may be engine
mounted. This, however, increases the engine PRESS. SUPPLY
gimbaled mass and may pose space and envelope RETURN
PUMP SUCTION
problems. To compensate for misalinements and EXTEND _1
RETRACT
thermal expansion and contraction, a certain 9. FILEEI
amount of flexibility must be provided for the 16 BLEED VALVE
17 SERVO-ACTUATOR
lines even in this case. It is possible to con- II. PREFILTRAIION VALVE
nect an electric generator to the main turbopump, 19 SAMPLING VALVE
20 DIFFERENTIAL PRESS. INDICATOR

and drive electrically a stage-mounted hydraulic 21 . CYLINDER BYPASS VALVE
pump. Only electrical wires will then cross the 22. SERVO VALVE
gimbal plane, with the exception of the hydraulic 23 SNUBBER
24. PISTON BYPASS VALVE
lines to the actuators which always must be 2S. LOCKING MECHANI_d4
flexible. Another possible simplification is to 26 MECHANICAL LOCK
26 30. PRESS. TRANSDUCER

combine servo valve and actuator into one single
31. [_:.W. TRANSDUCER
unit. Figure 7-14 shows a typical hydraulic 32. FEEDBACK TRANSDUCER
engine actuation system. Figures 7-15 and 7-16

may serve to identify the major components of Figure 7-16.-Servoactuator schematic.
this system.
From the above it becomes apparent that
numerous hydraulic connections will have to be must be long enough and of the proper pressure
made when installing the engine into the vehicle. rating, and, above all, must have a mating part
All of them must fit, and permit adequate flexure, on the vehicle.
Furthermore, the designer must know and

consider the environment in which the hydraulic
system will perform. The narrow-tolerance com-
ponents and the hydraulic fluid are very temper-
ature and contamination sensitive. Thus, the
designer will have to specify extreme cleanli-
ness requirements and adequate temperature
conditioning for maximum reliability of this vital
system. On cryogenic engine systems, heating
of the more sensitive components of the hydraulic
system is often required. Groin. supplied and
thermostat-controlled electric heaters, which
disconnect at liftoff, are a favored temperature-
conditioning method. Continuous recirculation
of the hydraulic fluid by means of the aforemen-
tioned ground electrically driven auxiliary pump
is another.
Secondary Injection
(C) LIQUID INJECI'O'R
Thrust vector control through secondary injec-

tion of matter into the thrust chamber nozzle Figure 7-17.-Secondary injection systems.
(SITVC) has been successfully applied to solid

motors. It has found only limited, predominantly factors: amplification (K) and axial thrust aug-
experimental application in liquid propulsion mentation (K,). These factors are defined as
systems, where it appears especially promising follows:
for upper stage engines, in which the lateral
forces required are smaller than with boosters.
K=ISs Fs/fvs_Fs/Fp (7-1a)
The principal methods of secondary injection
lsp Fp/_Vp fVs/_V p
are-
(l) Gas injection, using- lsa_AFa/ws_AFa/Fp

(a) Inert stored gas K, (7-15)
=lsp Fp/_'p fVs/fV p
(b) Thrust chamber tapoff (fig. 7-17A)
(c) Gas generator (fig. 7-17B) where
(2) Liquid injection (fig. 7-17C), using- ¢Is =secondary flow rate, lb/sec
(a) Inert fluid ¢¢p =primary flow rate, lb/see
(b) Propellants Fs =side force, pounds
Other methods, such as injection of preheated Fp =undisturbed axial primary thrust, pounds
hydrogen, have been investigated but proven AFa = axial thrust increase, pounds
uneconomical.
Isp = undisturbed axial specific impulse of the
In a gimbaled thrust chamber, the side force primary chamber (seconds) = Fp/fyp
is located approximately at the injector end. Is s = side specific impulse (seconds)= Fs/_'s
With an SITVC system, the applied side force is Is a = secondary axial specific impulse (sec-
located downstream of the nozzle throat and onds) = AF a/fV s
approximately at the point of injection, resulting Essentially, the K factor determines the quan-
in an increased moment arm which decreases the tity of fluid required to obtain the side force,
required side force. and the K_ factor determines the penalty on the
overa_A system Is to obtain the required side
Performance Evaluations force. If both of these factors are known, the
The performance of any type of secondary total effect of a given secondary injection sys-
injection system is based upon two performance tem on a propulsion system may be determined.
28O DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
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
144pl V2"I V2 to at least 10 times its diameter for repeatable

z[ +-- + = z2 + __
144 P2 +_-=2 constant (%2)
p 2g p results. For liquid flows, this flow-measuring
method is fairly accurate, if frictionlosses are
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
In conformance with the continuity law of

fluid flow OtFFE_ENTIAL _IAPH_ C,kA
PrLOT _¢ALVE
# = Cvpv,A 1 _ Cvpv2A2 (7-4)

144 144
and
vl A2
(7-5)
v2 A_
_ENTURI
where _ATE
z z, z 2 =elevations at sections 1 and 2, ft

P,, P2 = static pressures of the fluid at sec- Figure 7-18.-Schematic of a typical closed-loop,
tions 1 and 2, psia fluid-flow control system.
v,, v 2 =velocities of the fluid flow at sections
1 and 2, fps
Figure 7-18 is the schematic of a typical
p = density of the fluid, lb/ft 3
closed-loop fluid-flow control system, in which
g =gravitational constant, 32.2 ft/sec 2
the static pressure differential across a venturi
IP = weight flow rate of the fluid-flow,
is sensed and used to control fluid flow. The
ib/sec
flow is regulated by means of a butterfly gate
Cv =venturi or orifice flow velocity coeffi-
which is positioned by a fluid-powered actuator
cient. This is a function of the de-
diaphragm. The working fluid pressure to the
sign configuration and of the fluid-flow
diaphragm is controlled by a pilot valve. Its
characteristics, and is determined by
position in turn is controlled by the pressure
tests.
differential between venturi inlet (sec. 1) and
A_, A 2 =cross-sectional areas of the fluid-flow
throat (sec. 2). Because of the dynamic charac-
conduit at sections I and 2, in 2
teristics of the venturi sensing ports, diaphragms,
The above basic fluid-flow characteristics springs, butterflies, and other parts, the rela-
can be used to measure or sense the flow rate in tionship between flow rate w and venturi pres-
flow control systems. An accurately sized re- sure differential (p_ -P2) is not exactly linear.
striction, such as an orifice, nozzle, or venturi, However, theoretical analyses usually permit
is inserted in the conduit. Pressure taps are good approximations of these dynamic functions.
provided for reading the static pressure p, and The empirical factors thus obtained will permit
P2 at the inlet (sec. I) and at the minimum area design calculations resulting in a reasonable
of the restriction (sec, 2). If the flow areas A I, degree of control accuracy. Flow-bench calibra-
A 2 and the fluid density p are known, the flow tions, including adjustments of, for instance,
velocities v, and v 2, and the flow rate _¢ can be spring forces, serve to further increase this
calculated with the aid of equations (7-3), (7-4), accuracy. The control of fluid flow and pressure
and (7-5). The venturi or orifice meter should be by means of orifices and regulators will be further
preceded by a straight length of pipe equivalent discussed in subsequent sections.
Sample Calculation (7-1) A =area of the surface in consideration, ft 2

# =viscosity of the fluid, lb/ft-sec
The followingdata are given fora horizontal
(viscosity conversions:
venturimeter,measuring liquidoxygen flow:
Venturiinletdiameter,dj = 6 in
1 lb/ft-sec = 3@.21b-sec/ft2
Venturithroatdiameter,d2 = 3 in
Venturiflow velocitycoefficient,Cv = 0.92
Pressure differential
between inletand throat - 4616.81b-see/in 2 = 14.84 poise)
(Pl - P2) = 22.5 psi
Density of LOX, p=71.38 lb/ft 3 U =velocity of a fluid particle at the surface
Determine flow rate w. in consideration, ft/sec
t =distance from the point where the velocity
of a fluid particle is zero, to the surface
Solution
in consideration, ft
From equation (7-5): U/t =rate of angular deformation of the fluid
When a fluid is forced to flow through a closed

VZ A2 (d2_ 2 1
conduit, its flow is laminar or nonturbulent below
v2-A , \d,/ =-4
certain "critical" velocities. In a laminar flow,
V 1 = ¼V 2 the fluid moves in layers, or laminae, one layer
gliding smoothly over an adjacent layer, with
Substitute this into equation (7-3): only a molecular interchange of momentum. The
velocity of the fluid is greatest at the center of
144 (p_- P2)_ v22_ (,_v2)
2 the conduit and decreases sharply to zero at the
p 2g conduit wall. As the flow velocity is increased
above the "critical" point, the flow becomes
turbulent. In turbulent flow an irregular random
V2 1 motion of the fluid exists, in directions trans-
=_ 2× 32'2 ×i44 × 2i'5-55.9 fps
71._s1- verse to the direction of the main flow. The
turbulent flow velocity distribution is more uni-
Substitute this into equation (7-4): form across the conduit than with laminar flow.
Even in turbulent flow there is always a thin
77
layer at the conduit wall, the boundary layer,
Cvpv2A 2 0.92 x 71.38 × 55.9 ×-_-x 9 which moves as a laminar flow.
Flow rate ¢¢=--
144 144 Experiments and theoretical considerations
have shown that the Reynolds number, R e of a
= 180.2 lb/sec
given fluid flow can be used as a criteria to
indicate whether a flow is laminar or turbulent.
Real Fluid Flows Involving Pressure Drops
All real fluids possess the physical property

mA
of viscosity; i.e., they offer resistance to shear
F u
stresses. The viscosity of the fluid directly ira,
/ /
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.
(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
used when compressible fluids are being han-

dled. The density of compressible fluids changes
considerably as a function of pressure; there-
fore, if the pressure drop between two points is 1Moody, L. F., Friction Factors for Pipe Flows,
great, density and velocity will change appre- Trans. ASME, Nov. 1944.
Figure 7-20.-Moody diagram.
of the design configuration. Sometimes the in-

creased resistance of a specific flow passage
can be accounted for by assigning to it a ficti-
tious or equivalent length Le, of straight duct
which is arrived at empirically. The sum of this
equivalent length and the actual passage length,
(L e + L), is then used in equation (7-7) for the
calculation of turbulent flow. Figure 7-21 pre-
sents typical resistance characteristics of 90 °
bends.
/
Because flow-control components
valves and fittings disturb the flow pattern, they
such as
/
produce an additional pressure drop in a duct or /
line of tubing. The loss of pressure produced by
a flow-control component cQnsists of the pres-
sure drop within the component itself, as well as
the pressure drop in the upstream and down-
stream ducting or tubing in excess of that which
would normally occur if there were no component Figure 7-21.-Typical resistance characteristics
in the line. With certain exceptions, the fluid o! 90 _ bends.
flows through rocket engine control components where

are usually treated as being turbulent. The true Ap = pressure drop chargeable to the com-
pressure drops chargeable to the components can ponent as defined by the test method
only be evaluated accurately through actual flow shown in figure 7-22, psi
tests. p =density of the fluid, lb/ft 3
Figure 7-22 shows a typical test setup for v= 144 _= characteristic flow velocity of the
fluid-flow-control components. Pressure taps pA* component, fps
are located 4 diameters upstream and 10 diam-
= flow rate of fluid passing through the
eters downstream of the component to be evalu-
component, lb/see
ated. This minimizes the flow disturbances at
A* = characteristic flow area of the com-
the pressure tapoff points caused by the compo-
ponent. This is the minimum cross-
nents. The combination of pressure gages,
sectional area in the flow path of
U-tube manometer, and weighting tank produces
the component when fully open, in 2.
quite accurate and repeatable test data. The net
Usually this area is designed to
pressure drop caused by the component is ob- have about the same cross-sectional
tained by subtracting from the measured Ap that
area as the duct leading to the
pressure drop which is caused by an uninter-
component.
rupted straight pipe of the same size and length
The coefficient K is essentially constant for
(a+b)=14 diameters, at the same flow condi-
any given component over a large range of Reyn-
tions.
olds numbers, providing the flow is turbulent.
Because of the large number of fluid-flow-
For a given type of component configuration, K
control components used in rocket engines and
may vary with its size. A smaller size tends to
the great variety of service conditions, it is
have a higher K value. In general, the higher
virtually impossible to obtain individual test the flow resistance of the component, the more
data for every type and size of a component for
nearly independent of size is the resistance
the determination of pressure drop. It is desir- coefficient g. If a series of flow-control compo-
able instead to extrapolate from test data which
nents of different sizes were geometrically simi-
may be already available. This can be done by lar (constant ratio in all the linear dimensions),
employing a component resistance coefficient K
the resistance coefficient K would then depend
when calculating pressure drop using the corre- upon the Reynolds number only and would not be
lation
influenced by component size. However, the
design of a component is influenced by design
pv 2 standards, economy of material, structural
Ap = K288----
_ (7-9)
strength, available space, etc. None of these
considerations necessarily require geometric
similarity of the various sizes of a given design.
/.--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
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
for the duct. From figure 7-20, friction factor,

f: 0.0112. Substitute the equivalent total length
The characteristics of fluids flowing through (L e + L) into equation (7-7). The pressure drop
orifices will be further discussed in section
chargeable to the oxidizer flexible duct then is
7.10.
f(Le + b)pv 2
Sample Calculation (7-2) Ap =- 288 gd
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
(b_._)gain oxidizer valve surfaces by the fluid stream. This is especially

The characteristic velocity of the valve acute with the higher density liquids. To obtain
reliable control performance characteristics with
79.4 liquids at high velocities, the control compo-
V:o.--._18= 101.6 fps nents subject to impact loading must be designed
to withstand the stresses involved. Also, they
From table 7-4, the resistance coefficient for should be contoured so as to maintain small
butterfly valves K=0.31. Substitute this into impingement angles with the fluid stream and to
equation (7-9) to obtain the pressure drop charge- keep inpact forces to a minimum.
able to the main oxidizer valve:
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 control 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
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-
SPRING fied into those for medium-temperature service

(-60 ° to 400 a F), low-temperature service (-60 °
to -425 ° F), and high-temperature service (400 _ F
0 FI BODY
and up). The selection of the configurations and
the materials for these seals is based to a large
extent on service conditions and type of fluid
involved. Generally, soft nonmetallic or elas-
VSNT _-__!I!! ///, tomeric seals are used wherever possible. The
outstanding advantage of these seals is that
p,sTo;
N"I
A. they
fections
function
in the
satisfactorily
seal or the
despite
mating part.
minor imper-
Figure 7-23.-Schematic of a typical piston-type

actuator. Design of Dynamic Seals for Medium-Temperature
Services
ap = acceleration of the piston, ft/sec 2
A_ --area of the piston actuating side, in s Elastomeric O-rings have been widely applied
A s =area of the piston vent side, in 2 as dynamic seals for moving cylindrical parts as
p_ =actuating pressure, psia. This is tile well as for valve seats. However, proper design
pressure at the actuating fluid source, techniques must be observed to assure success.
less the supply-line pressure drop which Figure 7-24 and table 7-5 represent recommended
depends on the flow rate (a function of design practices for diametral-squeeze-type
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
tion ap clearance.-A frequent cause of seal failure is

Ff = friction forces (seals, sliding surfaces, the extrusion of diametral-squeeze-type O-rings
etc.), lb into the clearance gap adjacent to the O-ring
Fs = initial spring force, lb (at x = 0) groove (when under pressure) (fig. 7-25), leaving
C =spring rate, lb/in a permanent deformation after the pressure is
x =distance traveled by the piston from its reduced. Fluid pressure influences O-ring seal
initial position, in design because it affects the choice of compound
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 #:
STATIC
SEALING\
_ ¢,-_ I'-- "1 D-GROOVE
\BREAKCORNERS, : _b -,J U:NGTH
?o APPROX.
o.oos' D l .-f
"=" . _L__L._4___ a_-_--L "- CENTERLINE OF

THE PISTON ACTUATOR
Figure 7-24.-Diametral-squeeze-type O-ring seals in typical dynamic and static applications.
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
TABLE 7-5.-Recommended Design Practice for Diametral-Squeeze-Type O-Ring Seals
[See fig. 7-15 for explanation of dimensions; all dimensions in inches]
O-ring Diametral squeeze, 2 x E-

O-ring C-glandwidth D- R-
diametral
nominal rain
section groove radius,
section clearance
diameter length rain
diameter Dynamic Static aDynamic bstatic max
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
Fluid pressure O-ring compound hardness
(3-1000 psi 70 Shore "A" Durometer

1000-2000 psi 80 Shore "A" Durometer
2000 psi and higher 90 Shore "A" Durometer
aTolerance = +0.000, -0.001.

bTolerance : +0.000, -0.005.
_AOIAL -,-x _" _PRESSU_E __ P_ESSURE

7-27, 7-28, and 7-29) are used effectively as
C k F.A,_I* N C _ __]_ _"
dynamic seals for moving cylindrical parts and
valve seats. This type of seal has been applied
EXTRUSION OF (_RING RJNGS TO PREVENT
UNDER PRESSURE EXTRUSION
successfully at sealing pressures over 2000 psi,
and at temperatures as low as -425 ° F. They
t"igure 7-25.-Extrusion of the diamelraI-squeeze-
are also reasonably effective when sealing low-
type O-ring under pressure and the application
molecular-weight gases such as helium and
o[ backup rings.
hydrogen.
properties (furnished by the producer) is based

on operating conditions such as type of fluid or
°"' _° ...........
'Z:,;_ I 1710 t 010
propellant, pressure, temperature, and type of
seal (dynamic or static).
5. Installation o[ diametral-squeeze-type
O-ring seals.-Proper installation of O-rings dur- ....,oo,I 1, ...../ \ 1
ing component assembly is extremely important
to assure an effective seal. Generous chamfers L k t, "-/->' 1 ;
or radii should be provided on all edges and
corners in contact with O-rings to minimize the
possibility of cutting or scratching during the
/ k
assembly process. L _IPSEALS _. --A- _A_ _TE UNLESS aT_E_SE

121 _(_QLIIRIEO "_ _ _ °PIER SPECJFP[I]I F_NISH O_l i'LL
SUtteES _ T_E ._AkS

6. O-ring seals for valve seazs.-O-rings can ANC} t_ CONTACT WiTH THE
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).
Design of Dynamic Seals for Low-Temperature

Services
For cryogenic or low-temperature services,

lip-type seals made of elastomer sheets (figs.
VALVE STEM
Figure 7-28.-Typical valve seat lip-seal design

VALVE PRESSURE ' PRESSURE
POPPET I VALVE
[or butterfly-type valves.
_ OUStNG
The basic design principle of lip seals is to

employ the fluid pressure to increase the contact
,_AL RETAINING pressure at the sealing surfaces. Due to their
RING
lip configuration, the resilience of these seals

Figure 7-26.-Typical valve seat O-ring seal is maintained even at very low temperatures.
design for poppet-type valves. Design considerations for lip seals are similar
-Acrvz. ,_ _%ulO
to those for O-ring seals. The design approaches

can best be illustrated by examples. Figure 7-27
shows a typical valve shaft rotating lip seal
.-T_P,Ca_ WEL._EO .N_N_
arrangement, including dimensions and surface /' _ S,{'rUJXOm I,LA'II
_o_ ( CLOS_N& ) }_ KLLDW$ ASEhqI._

finishes, for liquid oxygen and hydrogen service.
Figure 7-28 presents the valve seat lip seal of a
butterfly valve for use with the same liquids.
Valve seat O-ring seals (fig. 7-26) made of Kel-F
have also been successfully applied in poppet-
W: i _.,: . ',_ ../'
type valves for liquid oxygen. The design of

lip seals for piston-type actuators using low- ;b"_ . }
temperature helium gas as the actuating fluid is
shown in figure 7-29.
In liquid hydrogen service, metallic bellows
(as shown in fig. 7-30) have been used to a great
extent to achieve positive dynamic sealing. _L_tE_ _ 001NV_
However, pressure levels and available space _.&LIN* 5t._f_3t5 LBO_H _LVE AN0 ,_v_t ) L_PPED N 8-10
impose limitations on their application. WELI)ED TYPE B£LLOWS USEO AS

RIEC_PRI_C,ATING Ty_P'_ D'yINiAIMI_;
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:
1. Applicalion of pressure.-When a bellows

is subjected to a differential pressure between
interior and exterior, it is preferable to apply the
higher pressure to the exterior. This reduces
stress, and permits higher pressures and longer
VALVE [_ __-t__ _ _j;!
_ i'_.>_AC T UATOR ROD life for a given design.
2. Provision of mechanical stops.-These
should always be provided to prevent extension
ACTUATOR PISTON
of the bellows beyond its permissible extended
Figure 7-29.-Lip seals for piston-type actuators. length and compression beyond its "bottomed"
Double lip seals seal pressures both ways. height.
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
can he approximated by contact, as shown in figures 7-30 and 7-32. This

seal design has two basic requirements. Firstly,
Bellow effective area = 0.1963 a finish of 10 nns or better is required for the
× (inside diameter +outside diameter) 2 (7-11) sealing surfaces. Secondly, a high-enough unit
loading must be applied to create a compensating
5. End attachment.-Typical welded joints for deformation of the sealing surfaces and to
the end attachment of bellows are illustrated in achieve the intimate contact required to over-
figure 7-30. Silver brazing and soft soldering come manufacturing tolerances, distortion of the
can also be employed for low-temperature serv-

ices.
A typical design of a rotating-type dynamic
seal for high-temperature services is illustrated PRESSURE
HOT GAS _ HOUSING
VALVE
in figure 7-31. The dynamic sealing is achieved

through the spherical mating surfaces between
the graphite seal ring and the steel shaft collar.
The contact force of the sealing surfaces is
maintained by the shaft thrust spring. Any mis-
alinement between the thrust bearing and the
SEAL "X O
shaft is compensated by the spherical seal face
SURFACES ROTATE
and side movement of the seal ring. This seal
FINISH tO RMS TO OPEN \ BUTTERFLY VALVE GATE
OR BETTER VALVE
arrangement has been applied to a turbine hot
VALVESHAFT
{a) CONVOLUTED R_NG TYPE VALVE SEAT SEAL USED _N A
HOT GAS SHUTOFF BUTTERFLY VALVE
_ •. _ /- _.. _ _1_?,_,_
_ _,,_
HOT GAS
PRESSURE VALVE POPPET
SPRING DISK | /
VALVE
/
VALV¢_ ...... _, _ sto,,
/ HOUSING
POSITIVE SEAL SURFACES FINISH

STOP IO RMS OR BETTER
_b} SPRING DISK TYPE VALVE SEAT SEAL USED IN A
S(AL _s, aN HOT GAS SHUTOFF POPPET VALVE
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
The degree of sealing (or the allowable leak

Butterfly-Type Propellant Valves
rate) is a very important specification which will
dictate the type of seal to be selected for a The butterfly valve is one of the most widely
specific fluid-flow-control component design. used propellant valve types in large liquid pro-
The basic reference for leak rates is Specifica- pellant rocket engines. It has established a
tion MIL-S-8484. It states that a Grade A seal, reliable operational record in LO2/RP-1, LO2/
the highest quality seal, shall have a leakage LH 2, storable, and other liquid propellant serv-
rate not to exceed 1 standard cubic centimeter of ices. Existing butterfly valve designs range
air/year/inch of seal at a pressure differential from 2 to 17 inches nominal diameter, for use at
of 1 atmosphere. This corresponds to a leakage propellant pressures from 20 to over 1500 psia.
rate of 3.171 × 10 -s cc/sec/inch of seal. It is a With improvements in sealing and structural de-
design assumption that any seal leak rate below tails, successful designs for higher capacities
or equal to this value is considered zero leak- and propellant pressures are certain to be
age. For many applications, higher leak rates achieved.
are permissible. For instance, a check valve Fignre 7-33 presents a typical butterfly valve
may be specified with a leak rate of 5 scim's design. Sealing is provided by a lip seal, which
(standard cubic inches of gas per minute). This engages a spherical surface on the valve gate,
is still a relatively tight specification. similar to figure 7-28. The valve gate pivots on
the valve shaft, the axis of which passes through
the geometric center of the spherical sealing
7.8 DESIGN OF PROPELLANT VALVES
surface. In most designs, the valve gate rotates
Propellant valves are used to initiate and 90 ° from the closed to the fully opened position.
terminate propellant flows to main thrust cham- The valve is operated by a piston-type actuator,
bers and gas generators. They are usually open- through a connecting link and shaft crank arm.
closed, two-position, normally-closed valves. Lip seals are used as dynamic seals for the
To meet specific sequencing requirements, other rotating valve shaft (fig. 7-27). The actuating
designs may provide for an intermediate opening power is furnished either by noncryogenic pro-
position. For thrust-throttle or mixture-ratio- pellant pressure, or by an inert gas supply, and
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"
//. r --:-_ ....

---_ _ ,...... Figure 7-34.-Four-inch butterfly-type main liquid
oxygen valve used on Rocketdyne Atlas ICBM
booster engines.
Figure 7-33.-Typical butterfly-type propellant
valve design (shown in the closed position). valve. A butterfly valve maintains a relatively
smooth fluid-flow stream over a wide range of
controlled by a pilot valve. The valve shown is valve-gate angular positions. Thus, when used
designed to be normally closed by a spring which as a throttle valve, it has little tendency toward
is installed on the closing side of the actuator turbulence with attendant adverse effects such
piston. Except for shaft and pins which are as local propellant cavitation. Typical fluid-
made of stainless steels, most of the other parts flow resistance coefficients K at various opening
are made of aluminum-alloy forgings. Figure positions of a butterfly valve are listed in table
7-34 shows a 4-inch, butterfly-type, liquid oxy- 7-4.
gen valve used on the Rocketdyne Atlas ICBM When RP-1 is used as the actuating fluid for
booster engine. the liquid oxygen valve, in a LO2/RP-1 system
Butterfly valves have relatively low resist- such as the A-1 stage engine, a heater may be
ance to fluid flow. They are compact, light, and required at the actuator to keep the RP-1 from
easy to service. They have a high characteris- freezing. The actuator-valve arrangement of the
tic area which can be expressed as butterfly valve shown in figure 7-33 provides
flexibility for specific engine control system
needs: the valve may be normally open or nor-
A*-_-d _-Ag (%12)
-4 s mally closed; position indicators may be added;
closing of the valve may be accomplished by
where (see fig. 7-33 for dimension references) means of a pyrotechnic squib, rather than by
A* = characteristic area of the valve, in _ pneumatic pressure. Figure 7-35 illustrates the
ds =inside diameter of the valve seat lip seal, linkage between the main oxidizer valve and the
in igniter fuel sequence valve of the A-1 stage
Ag=projected valve gate area at the fully open engine. During the opening stroke, the cam
position, in _ attached to the main oxidizer valve shaft actu-
Values for A* range from 65 percent of the ates _he igniter fuel sequence valve to open.
duct area (duct area = v/4 dn 2, where d_ = valve Frequently, a potentiometer is also attached to
nominal diameter, in) on a 2-inch size valve, to the shaft for continuous indication of the angular
about 87 percent of the duct area on a 12-inch position of the valve gate.
The amount of torque required to turn valve

_._tN OXJDtZEk VALV_ shaft and gate is determined by the summation of
OPE*WtNG SIDE ADTIJATING.
P=E_J_E APPUED hydraulic and friction torques. Hydraulic torque
M,&pN OXIDIZ_I VALVE
iN FULLY C_=ENED POSITION
is the unbalance of forces on the valve gate
AC IIJAIIN G FWID caused by the flow of fluid around it. If the axis
_CLOSlNG) ° of the valve shaft is located as shown in figure
FtUIP INLET PORT
7-36(a), the fluid striking the gate portion pro-
AC_UMING
(C_ENI*WG)
truding farthest upstream is deflected more than
FUEL PIIS_U_! r_ow _,
E_G_NE CO_WIItOL VALVI that at a point near the other end of the gate.
POIATE TO J
O_EN VALVt
This produces an unbalanced force which tends
VALVl SH_.= T --_ IGNiIEI FUEL to close the gate. Offsetting the valve gate as
SEOU_NC_ VALV_
shown in figure 7-36(b) would further increase

the closing torque, because the fluid velocity
rises as it approaches the downstream side open-
IGNI_EI C_mrDC_
/,ND IGNmON ing. Consequently, the resulting low-pressure,
MONIIO_ VAtVt
area tends to increase the unbalance in the clos-
ing direction. For this reason, butterfly valves
Figure 7-35.-Mechanically linked arrangement
are usually designed offset as shown in figure
between the main oxidizer valve and the igniter
7-36(c) (also see fig. 7-33). This produces a
fuel sequence valve of the A-I stage engine.
fluid velocity effect tending to ease opening of
the gate, because of the lower net closing hy-
BUTTERFLYVALVE GATE draulic torque acting on the valve gate.
Nevertheless, the net hydraulic torque will
AXISOF VALVE
still be acting in the closing direction for most
./ / SHAFT
angular gate positions (9°-80°), unless the valve
FLOW gate is further offset. Friction torque always
(a)
opposes rotation. For most operational valve
DIRECTION
designs
FARTHESTUPSTREAM
PROTRUDING To = Tt+ Th (7-13)
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)
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,
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
Valve gate angular Ap, psi K[ Kh

REQUIRED OPENING TORQUE position, deg
(To=T f +T h )
5 .............. 1058 0.78 1.11 × 10 -a
15 .............. 769 0.78 2.55x 10 -_
/FRICTION TORQUE (Tf) 40 .............. 87.5 1.57 12.50×10 -_

W 85 .............. 25 3.61 -11.64×10 -3
D
0
re"
0
I-- Determine the required opening and closing
torques at the 5° , 15 ° , 40 ° , and 85 ° angular posi-
T, ORQUE (Th) tions of the valve gate.
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--
at85°: T[=3.61×0.Sx0.05×(7.7) 2×25

= 214 in-lb
GATE ANGULAR POSITION / From equation (7-16), the hydraulic torques
0= _ 90 °
Th = KhdsSAp
CLOSING
at 5°: Th = 1.11 x 10-3 x (7.7)3 x 1058
Figure 7-37.-Typical required opening and clos- = 535 in-lb
ing torques versus gate angular position fora at 15°: Th;2.55x10 -3x(7.7) 3×769
butterfly valve. = 895 in-lb
at 40°: Th = 12.50x 10 -3 × (7.7) 3 x 87.5

= 500 in-lb
aCT_TOa
at 85°: Th=-11.64×10-3×(7.7)3x25 J_L
=-133 in-lb
From equation (7-13), the required opening
torques
To = TI+ Th
at 5°: To=1960+535 =2495in-lb
at15°: To=1425+895 =2320in-lb
at 40¢: To=326+500 = 826in-lb
at 85°: To=214+(-133)= 81in-lb
From equation (7-14), the required closing
torques
Tc = (Tt- Th)
Figure 7-38.-Typical ball-type propellant valve
at 5°: Tc=1960-535 =1425in-lb
design.
at l5°: Tc=1425-895 = 530in-lb
at40°: Tc=326-500 =-174in-lb
at 85°: Tc=214-(-133) = 347in-lb
assembly consists of a seal ring and an attached
metallic bellows. The area enclosed within the
Ball-Type Propellant Valves

effective seal diameter, (4ds2) , is designed to
The major advantage of a ball valve is its be less than the effective area of the bellows.
low-pressure drop, since it permits in-line, unre-
During valve closing, this creates an unbalanced
stricted fluid flow. Its use also enhances struc-
force acting on the sealing surface, as affected
tural soundness for high-pressure service. It has
by the fluid pressure within the bellows.
a reliable record in cryogenic and some storable
Each valve ball is trunnion mounted using
propellant applications, for high-capacity gas
two integral axles on antifriction bearings. The
generators as well as for lower thrust main cham-
axis of valve rotation passes through the geo-
bers (up to about 50000-1b thrust). Up to a nom-
metric center of the spherical sealing surface.
inal diameter of 3 inches, ball valves are com-
In most designs, the valve ball rotates 90 ° from
parable to other valve types with respect to
the closed to the fully open position. Typical
space envelope and weight. For larger diam-
fluid-flow resistance coefficients K for ball-type
eters, ball valves are used only infrequently,
valves, as determined at the fully open position,
because it becomes increasingly difficult to meet
are listed in table 7-4. Ball-type valves can
flight weight and envelope requirements. How-
readily be used as flow-regulating devices, such
ever, for ground service applications where as for propellant throttling. Flow characteristics
weight and size are not critical, the ball valves,
for constant pressure drops of a typical ball
in all sizes, are used quite frequently. Many
valve at various angular positions are presented
ball-type propellant valves are designed in a
in figure 7-39.
mechanically linked, dual-valve arrangement,
The activation of the ball valves shown in
operated by a single actuator.
figure 7-38 is provided by a piston-type actuator,
Figure 7-38 illustrates the design of such a which could be powered by either fuel pressure
valve. Here, the valve elements are mechani- or inert gas pressure. The reciprocating motion
cally linked, controlling both oxidizer and fuel of the actuator is translated to a rotary motion of
flows. The valves can be sized either individ- the balls by means of a connecting link and crank
ually as shown, and according to the specific arm arrangement. The opening sequence of the
volumetric propellant flows, or be designed to two valves can be adjusted by varying the rela-
have a common size. The sealing of a ballvalve tive angular positions between the valve axles
is accomplished by lip- or O-ring-type seals, and the crank arm. The actuator shown has two
riding on the spherical sealing surface of the stages. First, the valves are opened to an inter-
valve ball. In our specific case, the valve seal mediate position (partial opening), then to the
*,_l_ vAtvl oo_Lt7
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
Figure 7-39.-Flow characteristics for constant

in the balance chamber are dimensioned so as to
pressure drops ol a typical ball-type valve at
various angular positions. result in the proper counteracting force which
varies as a function of the unbalanced force at
various positions of the valve. A small sequence
fully open position using separate pistons for valve is mechanically attached to the main valve.
each stage. The closing of the valve is effected This type of poppet valve is suitable for high-
by venting both opening ports and pressurizing flow attd high-pressure, storable as well as
the closing port. Dual seals with a drain be- cryogenic, propellant services•
tween them are provided for all dynamic seals
sealing to ambient.
Venturi-Type Propellant Valves
Figure 7-41 presents a typical design for a

Poppet-Type Propellant Valves
venturi-type propellant valve. In certain instal-
Figure 7-30 shows a typical poppet valve lations, it may be desirable, for various reasons,
with metal-to-metal seats. This valve is de- to use a valve of a nominal size smaller than
signed to be pneumatically operated and nor- that of the main duct. A valve installed in the
mally closed. All sealings are achieved without throat of a venturi is a possible solution. The
the use of elastomers. Because of the nonwiping smooth contours of the venturi limit pressure
characteristics of all dynamic seals, this design drop penalties to a few psi. Adjacent ducting
is particularly suitable for use with fluorine and permitting, it is conceivable that the venturi may
other highly reactive propellants. A main ad- simultaneously be used for flow measurements.
vantage of poppet valves is their relative sim- Typical resistance coefficients K for venturi
plicity. This is largely due to the reciprocating valves are given in table 7-4.
operation which permits the direct, in-line con- The venturi may be designed to operate as a
nection of an actuator. However, this arrange- cavitating venturi. Based on Bernoulli's energy
ment requires turning of the flow inside the pas- equation (eq. (7-2)), the minimum pressure of a
sage. and consequently results in relatively liquid is made to fall below its vapor pressure.
high-pressure drops. Typical fluid-flow resist- As a result, a gaseous region forms at the throat.
ance coefficients K for poppet valves are given If this gas moves through the throat at the veloc-
in table 7-4. ity of sound, downstream pressure variations and
Figure 7-40 presents the design of a typical, disturbances cannot advance beyond the throat.
large size (6 to 10 inches nominal diameter), Up to minimum pressure differentials across the
poppet-type propellant valve. To reduce the venturi (say _0 percent of upstream pressure),
unbalanced hydrodynamic forces, and thus the flow rate is dependent on upstream pressure
size of the actuator, a balance chamber is pro- only. When used as a throttling device, the
vided. The effective area and the fluid pressure cavitating venturi affords smaller pressure drops,
3O0 DESIGN OF LIQUID PROPELLANT ROCKET ENGINES
loaded poppet seated at the throat. Propellant

-- VALVE POPPET
line pressure controlled by a solenoid pilot valve
is used to actuate the valve open or closed. As
shown in the valve schematic (fig. 7-41), the
normally closed pilot valve is inserted in a pas-
,._j_ LET sageway interconnecting the poppet cavity and
an opening in venturi throat. Normally, upstream
propellant pressure fills the poppet cavity and
provides additional seating force on the poppet
in the same direction as the valve spring to as-
sure valve closure. When the pilot valve is ener-
SPRING gized to open, propellant pressure behind the
tNLET
valve poppet is vented out at a greater rate than

_---SOLENOID PILOT VALVE
it can be replaced by leakage through the poppet
clearance area. This results in a reduced pres-
.,,_F--- SOLENO1D PILOT VALVE sure overcoming the valve spring and causes the
main valve poppet to open. The venturi valve
contains no dynamic seals. Since the valvebody
is not pierced by a shaft or actuator rod and
_ --------=_ FLOW
there are no dynamic seals, no pathways exist
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
Gate-Type Propellant Valves

Figure ?-41.-Typical venturi-type propellant
valve designed and manufactured by Fox Valve Figure 7-42 shows a typical design of a pro-
Development Co. pellant gate valve. Its major advantage is unre-
stricted fluid flow, resulting in low-pressure
drop. It also provides a relatively short distance
since the gaseous characteristics at the throat
between the valve inlet and outlet in the direc-
effect a near linear relationship between flow
tion of the flow. The design shown in figure
rate and supply pressure, rather than according
7-42 uses elastomer O-rings as the valve-seat
to a square law. The fluid-flow venturi valves
have been applied successfully in cryogenic and
storable propellant services.
In fluid-flow systems which require flow limi- ACTLIATrP,_G F'LUI_ INLET
tation as well as a shutoff control valve, the

venturi valve with a cavitating diffuser will pro-
vide both, at a weight of only the valve and at a f PISTO_ TYPE ACTUATOR
pressure drop of only the venturi.

INLET PORT
Venturi valves occupy a relatively long space (O_ENING) f ACTUATOrl _OD DYP4,_,I_C O-PiNG
in the direction of flow, between 4 to 6 times of

the line nominal diameter. This length imposes ST_TtC O.R G_,TE
VALVE SEAT t _ _SEAT SPRING
limitations on size and application in engine
systems. However, venturi valves up to 10
IhlLE T /_u= _ OCTLET
inches nominal diameter have been successfully
built for rocket vehicle systems.
Figure 7-41 presents a typical venturi-type
V_LVE G_TIE GUIOE _" 0 - RINI_ RET_INEA
propellant valve designed and manufactured by PLATE SEAL \L
_R_ VALVE GATE GUIDE PLATE
Fox Valve Development Co. It is a pilot-operated
shutoff valve which consists of a convergent- Figure 7-42.-Typical gate-type propellant valve
divergent venturi section with a simple, spring- design.
seal. These are suitable only for medium- actuate other fluid-flow-control components, such
temperature services. In cryogenic application as propellant valves, or to control engine se-
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-
ground-support services. portional type. We confine our discussion here

to the on-off type. Since the proportional-type
pilot valves are used widely in closed-loop con-
Needle-Type Propellant Valves
trol systems, they will be treated as regulating
A typical needle-type propellant valve is devices, and discussed in section 7.11.
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
. (3) Four-way types
4;:: i]___' ,;::" . :: "'.- ACTUATOR Two-Way-Type Pilot Valves

?', .><::.: ":r " " .,,
The term "two-way" refers to the number of
ports. A two-way pilot valve is basically a two-
__ \'\'\_ WETALLIC BELLOWS
port, open-close-type shutoff valve, similar to a
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.
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
solenoid for a high-capacity valve. The pilot-

operated solenoid valve, necessarily, requires a NORMALLY CLOSED VALVE NORMALLY CLOSED VALVE
IN NORMAL POSITION IN OPEN POSITION
certain minimum actuating fluid pressure to over-

come friction and spring loads before it will open
or close.
Figure 7-44 presents a typical direct-acting CYL VENT CYL VENT
solenoid, three-way normally closed pilot valve RET RET
NORMALL'f O_[N VALVE NC.RMALLY GREN VALVE
IN NORMAL POSITION IN CLOSED PO$fflON
CONNECTION
- ELECR_CAL
5PRgNG
PffS_dRE
DIAPH Rt, GM -_.f
-- SOLENOID CYUNOER RETURN
CLOSED _J AS_E_,_LY
• OffN ELECTRIC
POPPET _ ; VENT OR
OR
CYLINOER • RETURN _-- ELECTRO

PEES_GRE
MAGNET
$OLENO_b
ENE_GIZEbOIt
Figure 7-44.-Typical direct-acting solenoid Figure 7-45.-Typical design ol a pilo_-operalefl

three-way normally closed pilot valve devel- solenoid, _hree-way pilot valve developed by
oped by General Controls. Skinner.
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
has the same diameter ds as the poppet. The

From equation (7-18), the required valve
valve diaphragm is made of multilayer, thin (0.01
spring preload
inch) Mylar sheets which are pressure formed
with heat added. The effective diaphragm area
can be determined experimentally. The required _r
Sp:-_dd 2 Ps _Ft=4(2.1)2×20_14:69.2_14
preload of the valve spring may then be esti-
mated by : 55.2 lb
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
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):
pd=350- 146.5=203.5 psig

DYN_,_C O-RJN_ SaA_
_rUtJ_ _,LrH _Ue_ _tES_UmE INlay p'Olr Substitute into equation (7-17), the valve
rO V_Nr eOl_
output
¢ONT_Ot _lT _ VAL_! FC_r W: Pcl_,: 203.5 × 200 = 40 700 in-lb/s ec

¢ON_EC_[0 rO rH_US_
v[
; ¢tos_D Four-Way-Type Pilot Valves
A four-way-type pilot valve can replace two

three-way valves, for control of double-acting
*ctu*ro* --
CO_ErEo ro M_ _u_ _*LV_ *C_UMO_

(two-directional) actuators, as shown in figures
7-33 and 7-40. Figure 7-47 presents the basic
Figure 7-46.-Typical [luid pressure actuated schematic of a four-way pilot valve. The ports
three-way pilo_ valve used as the ignition mon- are arranged so that one side is venting while
itor valve in the .4-1 stage engine. the other is pressurizing.
304 DESIGN OF LIQUID PROPELLANT RI_CKET ENGINES
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.

NORMAL POSITION ENERGIZED POSITRON
The following data are given for the self-

Figure 7q7.-Basic schematic oi a {our-way
locking-type, four-way pilot valve shown in figure
pilot valve.
7-49.
$1 spring: Preload=35 lb, rate=210 ib/in

The typical design of a pilot-operated four-
S 2 spring: Preload = 25 lb, rate = 250 Ib/in
way solenoid pilot valve, developed by Valvair,
Static friction of the valve poppet, F{=24 lb
is shown in figure 7-48. This valve type is used
Pressure port pressure, Pi = 400 psig
to control pneumatic actuators.
Actuating fluid pressure, Pa = 250 psig
Figure 7-49 presents the schematic of a typi-
l_eturn port pressure = ambient
cal self-locking-type, four-way pilot valve. It is
Poppet guide diameter, d, = 0.5 in
held normally closed (cylinder 1 port) by springs
Valve poppet total travel = 0.05 in
S_ and S 2. When an actuating pressure Pa is
Determine diameters d 2, d 3, and d 4 (using an
applied to the opening port, the pintle moves and
actuating-force contingency factor of 1.5).
connects the pressure port to cylinder 1 port, and
cylinder 2 port to the return port. An unbalanced
Solution
self-locking force, 77/4 (all -dl 2) Pi, acting in the
opening direction, causes the valve to stay open, Calculate diaphragm diameter d 2 for the re-
even after the actuating pressure Pa is removed quired actuating force to open the valve:
A'_
PILOT PASSAGE i TERMINAL BOX /-- COIL
LOACNL%INLGoTvYEPRE'D-_, _ _ _L" _/y' P,LOT
, /\
BODY _ _'. Z_II_ ! III _=-_i__I
// /' ^ ] O-RING _---_ L-SEA L
PLUNGER ASSEMBLY _/ / _ SECTION A-A

SPACER _ PO2T EXHAUST PO;T
SECTION B-B
INLET
Figure 7-48.-TypicaI design of a pilot-operated, lout-way solenoid pilot valve, developed by Valvair.
Design of Solenoid Actuators
........ \ / _, ........... r ............... Solenoid actuators are usually applied to the

....... '\ \ / // r" / .........,_o,,:o, sliding stem of pilot valve poppets, for on-off,
two-position operation (figs. 7-44, 7-45, and
7-48). Solenoids are electromagnets, containing
an armature or phmger which moves in a coil of
wire. When the coil is energized, a magnetic
force is exerted on the plunger. The magnitude
_- AC"UAnNO OEA_,_O_ 'S-- C"n.t_ER _ _O_t _- ,EAt OlUL4N of the force is proportional to the cross-sectional
area of the plunger, the number of the coil turns,
Figure 7-49.-Schematic of a typical self-locking- the electric current applied to the coil, and the
type, four-way pilot valve. gap G between plunger and core. It increases as
the gap narrows.
Figure 7-50 shows a typical direct-current
t/" 2
solenoid actuator as frequently used in two-
-_d2 Pa = (S _ preload + S_ preload + F/)
position-type, pilot valve services. Its magnetic
x contingency f_tctor
current consists of a stationary core, the case, a
base, and the plunger. When the coil is ener-
d22_(35+25+24)× 1.5_0.64
gized, the plunger, along with the valve stem
rrx250
4 connected to it by a bolt, is pulled toward the
core, against a compression spring. This spring
d_ = 0.8 in
pushes the plunger back to the normal position
when the coil is deenergized.
Determine piston diameter d s, based on the
The following are general correlations for
required force to lock the valve in the open posi-
flat-faced, plunger-type magnets of the type
tion, in the event the actuating pressure Pa is
shown in figure 7-50.
lost:
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
c m[sso_ _ss SE_L
Calculate diameter d, for the required force to

close the valve:
G
SOtE_¢*_ east, i_o_ ./_
4 d,,2pa=(4(d3 _ -d,2)pi+ F.e-spring forces)
x contingency factor
0.222 x 400 + 24- (35 + 25
d2= + 210 × 0"05 + 250 x 0"05) x 1.5 = 0.22
_× 250 Figure 7-50.-Typical direct-current solenoid

actuator design frequently used in two-position-
d 4 = 0.47 in type pilot valves.
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.
The required electric resistance of the sole-

noid coil, at operating temperature
The following data are given for the solenoid
actuator of the A-2 stage engine four-way pilot R E 28 _^
control valve: =_-= 1.--:_=zu ohms
Required actuating force at start of stroke,
F=27 lb The wire can now be selected for the required
Nominal electric supply: 28 volts de: 2.0 resistance, For instance, we may choose a No.
amperes maximum 26 AWG copper wire (0.017-inch diameter, enam-
Valve poppet travel = Air gap between solenoid eled) and wind it to an average coil diameter of
core and plunger, G=0.05 in 1.05 inches.
Plunger diameter =0.56 in
Resistance = 51 ohm/1000 ft, warm.
Assumed flux leakage factor, f= 0.7
Total length = 1.05 x ,-r× 1425/12 = 391 ft.
Determine the amphere-tum requirements of
R =391 × 51/1000=20 ohms.
the solenoid.
Lay wires 22 deep, 65 high.
Thickness of coil: 22 × 0.017 = 0.374 in.
Solution Inside diameter = 1.05- 0.374 =0.676 in.
Plunger area: This leaves 0.676-0.56= 0.116 for insulation,

sleeves, and plunger clearance. Height of coil:
77 65 x 0.017 = 1.1 inches. Outside diameter = 1.424
A =_-x (0.56) 2 = 0.246 in 2 inches.
7.10 DESIGN OF FIXED-AREA-TYPE number and can only be evaluated by experi-

REGULATING DEVICES ments. Typical designs and flow coefficients of
frequently used orifices and flow nozzles are
Fixed-area-type regulating devices, such presented in figures 7-51(a), 7-51(b). and7-51(c).
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
The basic correlations for liquid flow through

orifices and nozzles are (assuming uniform flow
distribution):
, AIAIAIAIAIAIAIAIAI_Ap
(7-21)
_'= 0.0438 Cdo2v/_"L'_= 0.525 Cdo-_p
_i, = 0.0438 CdoZp ,_ = 0.525 Cdo 2 _ (7-22)
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.

do/d t (ORIFICE)
The following data are given for the fuel pump
0.2 0.4 0.6 Q78
discharge duct of the A-1 engine: 1.0
Required fuel weight flow rate, w = 892 lb/sec
Fuel pump discharge duct diameter, d 1 = 7 in
0.9
kl I I
Selected design value of total pressure

for orifice or flow nozzle installed
fuel duct, APo = 100 psi
drop
in the 0.8 "l _
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
P4(dl) 2 50-45 x_x (_2) 0 Q2 0.4 0.6
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 equations (7-22) and (7-23)

Cdo 2 0.67 x (4.47) 2 = 23.9
\/_-- \/0:5:
w= 0.525 Cdo2_/-_ = 0.525 Cdo2V_oop/k
(b_) For the VDI-type flow nozzle, the same

Cd 02 w 892
approach can be used to determine do. A value
•,'_ - 0.525 \/_p 0.525 x \/100 x 50.45 of 3.55 inches is found for do; thus
892
- 0.525 x 71 =23.9 (a) do_3.55
- 0.506
dl 7
(a_) For the VDI-type orifice, we determine by
trial and error that d o , which will simultaneously From figure 7-51(c), a value of C= 1.005 is
satisfy equation (a), figure 7-51(a), and figure obtained, for Re-- 1.306 × i0 6. From figure 7-52,
7-52. A value of 4.47 inches is found for do. the corresponding value for k is 0.28. Substitute
Confirm: this into equation (a):
do 4.47 Cdo 2 1.005_(3.55) 2

=0.638
d1 7 = V"_/-.-.-.-.-.-.-.-_8
:23.9
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:

CApIZ
w: V_-- (7-24) The following data are given for an orifice:
Orifice diameter, d o = 0.06 in
If (p2/p,)<_ Critical ratio, i.e., Flow coefficient, C = 0.6
Helium gas pressure = 500 psig
y [ y+l Helium gas temperature = 100 ° F

</ -\y-1 Ill / 2 \_-=-i" Calculate the orifice flow rates for down-
= , z: Igy ,yr0 (7-253 stream pressure, P2 = 14.7 (ambient) and P2 = 350
psia.
If (P2/Pl) > Critical ratio,
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):
R =gas constant, R/°R

T = gas temperature upstream of the orifice, °R 77 ,_
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)
C = flow coefficient, a function of design con-

_ 0.6 x 0.00283 x 514.7 x 4.11
figuration and flow Reynolds number• = 0.00774 lb/sec
V'386 x 560
May be approximated from figure 7-51(a)
and 7-51(b)
For a downstream pressure of 350 psia, the
orifice pressure ratio
0.3
P{ 2'__ 350 _
514.7 0.682
From figure 7-53, we obtain a compressibility

0.4_
0.6 factor Z=3.88; thus
o.7 _/_o' I I ! /_/I__

_'iZ0.5 • 3.88
x 0.0077
w =-:-:-v_. = 0.00726 lb/sec
4.11
0.9
7.11 DESIGN OF SERVOVALVES

o__ • 1.0 1.5 2.0 2.5 3.0 3,5 4.0 1.5
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
other conditions. By contrast, variable-area- (4) Minimum bleed of control fluid

type devices will function under both dynamic (5) Simplicity of construction and line con-
and static conditions. Classified according to nections
their function, the most frequently used variable- (6) Compatibility with environmental condi-
area-type pressure and flow regulating devices tions: temperature, humidity, accelera-
for rocket engines can be grouped into- tion, and vibration.
(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.')
on-off pilot valve, except that it can continu- Flapper-Nozzle-Type Servovalves

ously vary the pressure or flow rate of the actu-
ating fluid to control the desired actuator posi- The flapper-nozzle-type servovalve is essen-
tion. In a pneumatic actuating system, the tially a variable orifice or nozzle. Figure 7-55
servovalve operates as a pressure control de- illustrates schematically the operation of a typi-
vice. It functions as a flow control device in cal double-bleed unit. Actuating fluid is sup-
hydraulic systems. In rocket engine applica- plied at a constant pressure or flow rate. _'aen
tions, where usually only low-power electrical an input signal is fed to the torque motor from a
signals or sensing pressure forces are available servoamplifier, the electric signal is converted
to operate the valves, servovalves with very into a mechanical force. The resultant deflec-
high gain characteristic are required. Three tion of the flapper is proportional to the input
basic servovalve types are most frequently used: signal and causes the flow of actuating fluid to
the flapper-nozzle type (fig. 7-55), the spool type increase at one nozzle, and to decrease at the
(fig. 7-56), and the poppet type (fig. 7-57). They
are used independently, or jointly, to form a two-
stage servovalve in which the flapper-nozzle
4o t1_,8o
valve acts as a pilot valve for the spool valve.
A typical example was presented in figure 7-9. ,_ '-_o_
28, _ phose/I I
E" "- /I I"/5o
General Design Considerations for Servovalves:
(1) Type of control fluid (gas or liquid) and

its conditions (pressure and temperature)
::,,,! '
",>/
. \ I I 2oo
(2) Systems gain: In some applications the " I e
ratio between electrical quiescent input
power to the valve coils and the maxi-
mum valve output (as defined by eq.
I I , , -,oo
0608Z0 2 3 4 56 8 /0 20 30-90
(7-17)) is as high as 1600-2000 Frequency in cycles per second
(3) Bandwidth and frequency response; Figure 7-54.-System characteristics of a typical

dynamic stability servovalve and driving amplifier combination.
.r- ACTUATOR the throatarea of the nozzle. Most designs are

PISTON based on the firstcondition. The effective (ring
FIXED /
ORWICE--_ et A / shaped) flow area then is
i.__.N_ I, I / , [ I SIGNAL
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).
"
Solution Since A 3 = _d3X = nd32/8
(a) Since the gas flow through the nozzles is

discharged to ambient, it is assumed that the d3 = d4 :_ 8 x 0.000279 = \r/7 .095 X 10 -4 = 0.0266 in
77
pressure ratio across the nozzles will always be
less than critical (sonic flow).
d3 t=0"0266+0.004=0.01065 in
From figure 7-58, Z 3 =Z 4 =4.11. T=_-+ 4
From equation (7-24), the following correla-
tions are established: (b_ When the flapper is deflected downward
0.001431 inch from its neutral position, the dis-
CAtPsZI 0.7×500 . _ placement of the flapper from the upper nozzle
#' - \/RT -_/_ AIzI =O'753AIZ'
T-t
X =2
CA2PsZ2 + 0.001431 = 0.003325 + 0.001431
#2- V_-_- -0.753A2Z2
= 0.004756 in
CA3PoZ3 0.7x4.11
#3 =
VR T \/_ A3Po=O'OO62A3P° With the flapper deflected
CA4PcZ4
A 3 = ,,-rd3X = ,7× 0.0266 × 0.004756 = 0.000397 in s
#4- V_R-T
- -O'O062A4Pc
The flow rate through the upper nozzle, #3, is

The flow areas of fixed orifices and flapper now equal to the fixed orifice flow rate #x, plus
nozzles (eq. (7-27)) are: the flow rate from the actuator #o; thus
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
When the valve is at neutral position, the 0.753 × 0.000427 Z, +0.00021

actuator is at rest, and P°= 0.0062×0.000397 = 130.6 Z 1 +85.4
#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
Since Pc = Po = 450 psia under these condi- Po=436.5 psia

tions,the pressure ratioacross the fixedorifices
is 450/500=0.9. Z, = 2.687
Checking the results for a pressure ratio of

From figure 7-53, Z a =Z 2 =2.418
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
a Z_ value of 2.687 is derived from figure7-53;

d /o oo0427
thus
dr= 2=_ 4 -\/5"44×10-4=0"0233in
Po = 130.6 × 2.687 + 85.4 = 436.5 psia

#3 0.000778
= --
As =A4 =0.0062p o 0.0062x450 0.000279 in s With the flapper deflected
A4 = _4 (T- t - X) ._ ..... _::_:_::_;_i;!!i;iiiili iii!i!iiii!i!ii:!;i!i!i:ii

= . ×0.0266 (0.01065 - 0.004 - 0.004756)
iiiiiiiiiii::"_::::::::::::::':'""'"'_i '- ==========================_ _ _r_RN
=0.0001583 in 2
The flow rate through the lower nozzle, w4, is

equal to the fixed orifice flow rate w2, minus the
flow rate to the actuator, We:
0.0062 A4P c = 0.753 A 2 Z 2 - 0.00021 C'_LINC(R I CYLINaER 2 '

PORT PORT \_ V,_LVE 8CgY
(A) SPO(O. TYPE $ERV_ -VALVE SCF_E_ATIC

0.753 ×0.000427 Z2 - 0.00021
Pc= 0.0062×0.0001583 =328Z 2 -214
We use again a trial-and-error method to ob-

tain the values for Pc and Z 2 which will satisfy
the above correlations and figure 7-53. We have x ,I
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.
Sample Calculation (7-1 O)

ACTUATED BY REGULATOR
The following design data are given for a CONTROLLER
spool-type hydraulic servovalve (shown sche-

matically in fig. 7-56): VALVE STEM--,_w
BLEED
RP-1 is supplied at pressure Ps=2000 psia PORT
/_____,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
following design constants:

PRESSURE _._____%
a =254 _ED
ACTUATOR
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
BLEED PORTI - _ PRESSURE Pc TO

Substitute a, b, c, X, and p into equation
OUTPUT
REGULATORCONTROL
VALVE
(7-28) SUPPLY -.,-..._]'_ h'-_ ACTUATOR
GAS ,_
Ap=[_-_+(12_+ 1"79×10-7''2]
(0._9-_ ]v ] ×(50.45) 2 PRESSURE P=----_"_VALVE POPPET
BALANCE
Ap= 10_,+7.62_ "2 PISTON
(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.
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
METALLIC BELLOWS _ _ CALIBRATING SPRING

INTERNALLY EVA C b___
Elements of Gas Pressure Regulators
Most pressure regulators include two basic SENSING PRESSURE _'_ "_
elements: the regulator controller and the regu-

lator valve. The former controls the regulator
valve opening, which in turn regulates the gas
SENSING INPUT
flow through the regulator.

A regulator controller essentially is a sensing- Figure 7-58.-Schematic of a typical gas pressure
computing unit which measures the difference regulator controller.
• . --- .i:, I
CONTROL PRESSUFE Pc FROM A SINGLE- BLEED TYPE
ou . T LO O . .soo O
SERVO - VALVE ATTACHED TO THE REGULATOR CONTROLLER.
////--- REFERENCE SPRING LOAD / S DOME
_,_11_.\\\\_ ---T----_OuTLET Pr
(A ) DIRECT SPRING - LOADED (B) DOME - LOADED
CONTROL PRESSURE PC FROM

A SIGNAL-BLEED TYPE SERVO-VALVE F- CONNECTIONS TO A FOUR-WAY SERVO-VALVE
ATTACHED TO THE REGULATOR CONTROLLER _ ATTACHED TO THE REGULATOR CONTROLLER
!
._-\\_ DIA PH RAG M _ ._>.\_-_/j_--- ACT UAT OR PIST ON
VENT TO _ CALIBRATING SPRING I _ DYNAMIC SEAL

ATMO. pa-.F-- DYNAM,C SEAL _-i [
_'_----_OUTLET Pr _----_"__--OUTLET Pr
,NLET _-----__ ,NLET pi ___.__ REGULATOR VALVE
REGULATOR VALVE
_,C]CONTROL PRESSURE OPERATED (O) FOUR-WAY SERVO-VALVE OPERATED
Figure 7-59.-Schematics oi various gas pressure regulator designs.
in figure 7-59(b). A large-capacity, introducedby the oftenextreme variationof Pi.

dome-loaded pressure regulator can also A simple,directlyspring-loaded, low-capacity
be loaded by a separate, small-capacity regulatorsimilarto figure7-59(a)is requiredto
precision pressure regulator (precision reduce Pi to a reasonably constant supply pres-
loader). sure Ps for the control servo circuit. This regu-
(2) Control pressure Pc is used as the input lator is commonly known as the bleed regulator
to a mechanism that positions the regu- and can be a part of the main regulator controller.
lator valve as a linear function of con-
trol pressure. In its simplest form, this
Design ConsiderationsforGas Pressure
mechanism consistsof a spring and a
Regulators
diaphragm (fig.7-59(e)).
In some designs,a four-way servovalve (as in The followingare basic considerationsforthe
figs.7-55 and 7-56)operates the double-acting design of gas pressure regulators.
piston-typeactuatorof the regulatorcontroller to 1. The principal causes of error in a pressure
controlthe positionof the regulatorvalve (fig. regulator are-
7-59(d)).This design is known as an integrating- a. Variation in inlet pressure
type pressureregulator. b. Variation in flow demand
When a single-bleed servovalve is used with c. Variation in temperature
the regulator controller, the supply pressure Ps d. Mechanical hysteresis and friction
to the control servo circuit must be isolated from e. Creep of stress members
the inlet pressure Pi to the main regulator valve, f. Variation in effective length of mem-
because of the steady-state error that would be bers because of angular displacement
GAS INLET PRESSURE GAS INLET PRESSURE

g. Variation in vehicle acceleration 3O00 PSIA 12(_3 PS;A
h. Vibration
i. The fact that pressures are not sensed
.
at the point at which control of pres- ci c_t_ ............
sure is desired (influence of pressure

drops caused by flow through down-
wet.
stream systems) o 2
The basic test of a good regulator design is TIME, SEC TIME, SEC
whether these errors have been held within

Figure 7-60.-Dynamic response characteristics
allowable tolerances, without mechanisms (in-
o[ a typical pneumatic pressure regulator.
cluding diaphragms and springs) of unreasonable
complexity and size.
relief valves are provided for the main propellant
2. Where feasible, proven concepts and con-
trol mechanisms of previous designs should be tanks to prevent possible intermixing of propel-
lant vapors.
utilized. This is especially important for dy-
namic characteristics, stability and transient 6. Other design considerations for gas pres-
sure regulators are-
response, and for mechanical details such as
poppets, seals, diaphragms, and springs. a. Type of regulated gas
b. Gas inlet-pressure and temperature
3. Balance of the regulator valve poppet
range
should be provided to minimize the forces im-
c. Regulated gas outlet pressure level
parted to the valve by inlet pressure Pi. This
with respect to ambient or vacuum, its
may be achieved by attaching a balance piston,
tolerance or accuracy of regulation,
similar to that of the poppet servovalves (fig.
and range of adjustment
7-57). However, this will require some type of
d. Required maximum flow capacity
dynamic seal, which are known trouble sources.
e. Required response time and allowable
In certain applications, therefore, an unbalanced overshoot
valve may be preferred. f. Type of line connections
4. In most cases, the regulator must lend
g. Environmental conditions (temperature,
itself to the alternate mode of operation as a
vibration, humidity, etc.)
shutoff valve, until a start pilot valve is opened.
Figure 7-60 presents the dynamic response
5. The various gas pressure regulator designs
characteristics of a typical pneumatic pressure
shown in figure 7-59 regulate pressure by ad-
regulator upon initiation of a demand. The regu-
mitting additional gas, as required to the regu-
lator was set at a regulated outlet pressure of
lated pressure region. If the flow demand on the
400 psia. The gas used (helium) is discharged
regulator is reduced to zero (dead-end condi-
into a system of approximately 60-cubic-inch
tions), the regulator valve shuts off. However, volume.
the pressure in the regulated region may increase
above the design point, even with the regulator
Sizing of the Gas Pressure Regulator
valve closed, due to-
a. A timelag in closing, following reduc- The required flow area across the regulator
tion of demand to zero valve when fully open. i.e.,the characteristic
b. Inevitable regulator valve leakage flow area of a pressure regulator, may be calcu-
c. Thermodynamic effects, such as in- lated for required flow capacity and regulated
crease in temperature of the gas in the outlet pressure, at minimum allowable gas inlet
regulated pressure region, or from pressure.
vaporization of a liquid. The following correlation derived from equa-
This condition is known as a lockup, and often tion (7-24) for gas flow orifices is applicable to
poses problems. Lockup can be eliminated by gas pressure regulator design:
incorporating a relief valve downstream of the
regulator valve as part of the regulator assembly. w,]'R-T
A* - (7-29)
In the A-4 stage propulsion system, individual CpiZ
where employs a metallic diaphragm for low friction.

A* :characteristic area of the pressure regu- The dome pressure Pc, which determines the
lator, in 2 force to the actuator diaphragm, and thus regu-
: design weight flow rate of the pressure lates valve position and outlet pressure, can be
regulator, lb/sec regulated either internally by a pressure con-
R =gas constant, ft/°R troller, or externally by a low-capacity pressure
T :gas inlet temperature at minimum inlet regulator or loader. There are many variations
pressure condition, °R of dome-loaded regulators. Design approaches
Z :compressibility factor, fto's/sec, a func- may best be illustrated by typical examples.
tion of pressure ratio pr/Pi and specific Fig-ure 7-61 presents the schematic of a typi-
heat ratio (fig. 7-53) cal dome-loaded gas pressure regulator which
C : flow coefficient, a function of design con- has an alternate mode of operation as a shutoff
figuration. Design values range from 0.6 valve. The prime purpose of the regulator is to
to 0.7 maintain a regulated outlet pressure at a pre-
Pi :minimum allowable gas inlet pressure, psia selected value called the set point. The regu-
Pr =required regulated outlet pressure, psia lator is composed of four elements: the start
pilot valve, the bleed regulator, the sensing-
Sample Calculation (7-11) control unit, and the main regulator valve. The
combination of bleed regulator and sensing-
The following design data are given for the
control unit forms the function of main regulator
helium gas pressure regulator of the A-4 stage
controller.
propulsion system:
Design weight flow rate, w: 0.048 lb/sec The start pilot valve is a solenoid-actuated
Minimum allowable gas inlet pressure, pi=245 poppet valve which is normally held closed by a
psia spring. In the closed position, it locks up the
Gas inlet temperature at minimum inlet pres- outlet pressure of the bleed regulator. Thelatter
sure condition, T= 1030 ° R is a normally closed, directly spring-loaded-
Required regulated outlet pressure, Pr: 168 type, pressure-reducing valve and supplies an
psia approximately constant preset pressure Ps to the
Regulator flow coefficient, C =0.65 sensing-control unit. This pressure is greater
Determine the characteristic area A* of the pres- than the main regulator outlet pressure Pr. It is
sure regulator. the sensing-control unit. then, which establishes
the main regulator outlet pressure. It detects
Solution small deviations from the set point and magnifies
the error signals by means of a single-bleed,
At minimum inlet pressure conditions, the
flapper-nozzle servovalve linked to a sensor
pressure ratio across the regulator
Pr_ 168 - 0.685

Pi 245
From figure 7-53, the compressibility factor Z

is 3.9 for helium. Substitute this and other data
into equation (7-29). The characteristic area
A* = _'Vr/_'_- 0.048 × \/386 x 1030 : 0.0487 in s

CpiZ 0.65 x 245x 3.9
Dome-Loaded Pressure Regulators

Figure 7-61.-Schematic of a typical dome-loaded,
This regulator type (fig. 7-57(b)) is one of the negative-gain-type gas regulator with an alter-
most frequently used. The main actuator usually nate mode o[ operation as a shutoff valve.
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
TABLE 7-6.-Principal Design Parameters of a Typical Dome-Loaded, Negative-Gain-Type Gas

Pressure Regulator (Fig. 7-61)
Parameter Design data
Regulated gas ................. Helium

Inlet gas pressure, Pi ........... 4500 psia nominal. 5800 psia maximum, 500 psia minimum
Inlet gas temperature, T ........ 110 ° to 360 ° R
Regulated outlet pressure, Pr .... 400 psig t 25 psi
Flow demand .................. a. Maximum: 0.1 lb/sec at 500 psia inlet pressure and 160 ° R inlet temperature
b. Minimum: zero
Modes of operation ............. a. Start pilot valve solenoid deenergized: regulator remains shut off. Outlet pres-
sure : 0
b. Start pilot valve solenoid energized: regulator functions normally
Controller bleed ............... a. None when start pilot valve solenoid deenergized
b. 0.0002 lb/sec at 530 _ R when solenoid energized
Main regulator valve ............ Seating diameter, 0.205 in; characteristic area, 0.0281 in 2
Main regulator valve actuator .... Effective area, 7.55 in-_; bias spring preload, 75.5 lb; rate, 500 lb/m
Start pilot valve ............... Seating diameter, 0.025 in; characteristic area, 0.000437 in 2
Bleed regulator valve ........... Seating diameter, 0.028 in; characteristic area, 0.000301 in 2
Bleed regulator sensor .......... Effective area, 0.185 in2; reference spring preload, 93.5 lb; rate, 1500 lb/in
Bleed flow restrictor ........... Diameter, 0.0116 in; area, 0.000105 in 2
Main regulator controller sensor.. Effective area, 0.315 in2: reference spring preload, 126 lb; rate, 702 lb/in
Flapper-nozzle servovalve ...... Seating diameter, 0.043 in (2 holes); maximum stroke, 0.011 in; maximum flow area,
0.003 in_; bias spring preload, 1.0 lb; rate, 5.0 lb/in
REGULATOR CONTROLLER
SENSO_ ASSEMBLY
SENSOR OUTEB BELLOWS
ORb=ICED _ARTITION
___ SENSOR &AMRER

$ERVO- VALVE
VENT
/ r-- SPOOL TYRE FOUR-WAY

\ \ % & r SEmvO-
VALVE
SENSED PRESSURE
FROM OXIDIZER
TANK INLET
SENSOR INNER ' REGULATOR VALVE

BELLOWS ACTUATOR
DYNAMIC SE_L
_ -..._,
OXIDIZER TANK OX+D,ZER lANK
_r
' [ _ BALANCE TYPE
_ REGLJLATOR INLET,
Pi
Figure 7-63.-integrating-type, gas pressure

regulator with spool-type, lout-way servovalve.
Figure 2-62.-Typical dome-loaded, zero-gain-

type gas pressure regulator loaded by a bleed tank pressure is to be sensed through a rela-
regulator. tively short external line. Fuel tank pressure is

maintained at I0 psi below the oxidizer tank
pressure by different pressure settings of line
bleed from the control dome to the regulator out- check valves and tank relief valves. A spool-
let. The control dome pressure Pc is maintained type, four-way servovalve, which is directly con-
at a constant level by the bleed regulator. This nected to the sensor of the regulator controller,
is known as a zero-gain-type control circuit. controls the regulator valve actuator. Supply
Here, too, control pressure Pc is higher than the pressure Ps of the servovalve is taken from the
regulated outlet pressure Pr by an amount deter- regulator outlet pressure Pr; i.e.,Ps = Pr. The
mined by the main regulator valve actuator spring sensor contains two bellows with an oil-filled
preload. The regulator shown is designed to damper in between. The damper consists of a
maintain an outlet pressure Pr of 2S2 psig + 5 psi, partition having a properly sized orifice. The
at inlet pressures ranging from 5000 to 375 psig. outer bellows, which is exposed to the propellant
The actuator diaphragms are made of a 0.0025- vapors, is hydroformed from thin 321 stainless-
inch 17-7 PH high-strength stainless-steel sheet. steel tubing, because the pressure differential
440C stainless steel is used for all sliding mem- across it is small and because it contains fluid.
bers, such as regulator valve poppets and guides, The inner bellows must withstand a pressure
to eliminate the surface galling problem. Regu- differentialequal to fulltank pressure, since the
lator main body and housing are made of aluminum- interior of this bellows is vented directly to
alloy forgings. atmosphere. For this reason, the bellows is
machined from 17-4 PH stainless-steel stock.
Integrating-Type Pressure Regulators
This regulator is called an integrating type be-
Figure 7-63 presents the design of a typical cause a constant tank pressure error will produce
integrating-type gas pressure regulator, designed a constant actuator piston velocity (neglecting
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)
gas flow rate. Tank pressure is then the integral

of tank pressure error versus time. The integrating-
type pressure regulator provides the constantly
increasing regulator valve flow area required for
decreasing supply pressures, with minimum error
in tank pressure, but may be unstable if the gain
of the servovalve-actuator combination is too
high. A suitable gain value should be deter-
mined by computer studies of the regulator dy-
namic characteristics.
7.13 DESIGN OF LIQUID FLOW AND

PRESSURE REGULATORS
Figure 7-64.-Schematic of a typical sliding-
Liquid flow and pressure regulators basic- piston-type liquid flow regulator.
ally are variable-area-type, pressure-reducing
valves. A liquid flow regulator maintains a con-
stant rate of flow, or limits the maximum rate of that a constant pressure drop is maintained
a flow. A liquid pressure regulator, like a gas across the orifice disk.
pressure regulator, maintains a constant fluid As the fluid inlet pressure increases, or the
pressure at its outlet under variable flow condi- back pressure decreases, the flow rate tends to
tions. Design considerations for liquid flow and increase also. However, the attendant increase
pressure regulators are similar to those for gas is pressure differential, i.e., pressure force on
regulators. the piston, moves it against the reference springs,
simultaneously increasing tlle flow restriction.
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.
Figure 7-64 presents the principle of another where

type flow regulator which is frequently used in Fs =regulator reference spring preload, for
rocket engine systems, because of its simplicity. a given flow rate, Ib
It consists of two restrictors, one of fixed and Ap_ =pressure drop across the fixed orifice
the other of variable area. This combination disk at that flow rate, psi
automatically maintains a constant pressure drop AP2 = pressure drop across the piston throt-
across the fixed restrictor, and thus a constant tle ports at that flow rate, psi
flow. The fixed restrictor consists of a disk Ap = effective piston area, in 2
containing a number of orifices and mounted in a Ks =regulator reference spring rate, ib/in
piston. The variable restrictor is formed by (Ap/A}) = change of pressure drop across the
variable throttle ports located around the piston piston throttle ports per inch of
periphery. The pressure differential across tim movement of piston, lb/in_-in
fixed orifices causes the piston to move against Adjustment of the regulated flow rate is made by
the regulator reference springs. Piston throttle adjusting the reference spring force, using a
ports and reference spring rate are designed so graduated vernier mechanism.
' j
? " !
_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
direct-operated relief valves are used for large-

tolerance, low-capacity services. For large flow
BUNA-N PACKING
requirements, pilot-operated relief valves are
used for quick response and to avoid excessive
a'NEUMATIC REFEPENEE
size.
PRESSURE INLET_ 4_ DIAPHRAGM
qSEE INSET A) Most pressure relief valves are used in gas
pressures systems. Important design considera-
tions are-
(1) Type of gas and its conditions
(2) Pressure relief level, its tolerance and
range of adjustment
(3) Required response time, dead band (dif-
ferential between actuation or opening,
....
., and deactuation or closing pressure),
and other dynamic characteristics
(4) Required maximum flow capacity
Figure 7-65.-Typical liquid pressure regulator (5) Simplicity of construction and line con-
nections
design [or liquid oxygen service.
(6) Environmental conditions, such as tem-
perature, humidity, acceleration, and
vibration
Design of Liquid Pressure Regulators
The required flow area for a gas pressure
Figure 7-65 presents a liquid pressure regu- relief valve at its maximum opening position,
lator design frequently used in rocket engine i.e., its characteristic flow area, can be calcu-
systems. The actuator diaphragm of the regulated from equation (7-24) for gas flow orifices.
lator valve (plunger) can be loaded either by a
pneumatic reference pressure (as shown), or by a
reference spring. Direct-OperatedGas Pressure Relief Valves
Regulated outlet pressure is sensed directly. Figure 7-66(a)shows the design of a typical
An increase in outlet pressure tends to reduce
low-capacity,direct-operated gas pressurerelief
the flow area of the regulator valve until an
valve. The valve poppet is loaded directlyby a
equilibrium is reached between outlet pressure coilspring. This is one of the simplestdesigns.
and reference pressure. The actuator diaphragm Itmay, however, become dynamically unstable
is made of Mylar laminations which are compati- under externallyintroducedvibration.
ble with cryogenic propellants such as liquid Figure 7-66(b)is the schematic of an im-
oxygen. proved design of a direct-operatedreliefvalve.
Itassures a predictabledead band and elimi-
nates chatterunder vibration.This design em-
7.14 DESIGN OF PRESSURE RELIEF
ploys coned-disk-type(Belleville) springs,oper-
VALVES
ating between positivestops,in the negative
The prime functionof pressurereliefvalves rateportionof the force-deflection curve of the
is to protectfluidsystems and pressurevessels spring. As shown in figures7-66(b)and 7-67,
from being overpressurized.There are two basic the springremains againstpretravelstop C,
types of pressure reliefvalves; the directoper- untilthe appliedforce (valveinletpressure
ated and the pilotoperated. In the first,the Pi× valve poppet seat area nd2/4) reaches a
valve poppet is loaded directlyby a properly level A. The springthen snaps from C to the
CAP APPLIED FORCE, LB.

(VALVE INLET PRESSURE X
VALVE POPPET SEAT AREA)
SETTI NG __
SPRING _L-------_-_23_ /--8ODY SEAL SPRING
_ O- RING
(o)
__ LOCK NUT
pOPPET_ RELIEF A -ACT'UATION

FO_ /
B 'DE-ACTUATION/-- " -I-- _ I----at_
SOFT o- ,NOo,
DISC
INLET
POST-TRAVEL C D
(b) _ ADJUSTMENT SPRING DEFLECTIC, N, IN.
_ 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
Figure 7-66.-Low-capacity, direct-operated gas

pressure relief valves. (a) Coil-spring-loaded,
direct-operated relief valve; (b) Snap-action,
coned-disk, spring-loaded, direct-operated
relief valve.
PILOT VALVE CONTROLLEF
posttravel stop D, with no further increase in

applied force. Reduction of the applied force to
B will cause the spring to snap back from D to
C. The positions of the pre- and posttravel
stops can be adjusted for constant actuation and
deactuation forces, independent of spring manu- VENT Ldp,_'!=
e_
facturing tolerances. The coned-disk spring

washers are usually made of beryllium copper or Figure 7-68.-Schematic o[ a typical high-capac-
17-7 PH stainless steel. This type of relief ity, pilot-operated tank gas pressure relief
valve is suitable for high-pressure helium stor- valve.
age bottle services, such as in the A-4 stage
propulsion system.
controller also senses the tank pressure p:.
When the latter reaches or exceeds the preset
Pilot-Operated Gas Pressure Relief Valves
level, the pilot valve is actuated to crack. This
Figure 7-68 presents the schematic of a typi- vents the main relief valve actuator control pres-
cal high-capacity, pilot-operated tank gus pressure Pc and, in turn, permits opening of the main
sure relief valve. Normally, both main and pilot valve. Main valve poppet position is a function
valves, are held closed by valve spring forces of the control pressure, which in turn is con-
Fsm, Fs3, and pressure Pc. The pilot valve trolled by the position of the pilot valve poppet
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:
Fsm+pcA2- Pt(As-A1)- paA3 = Fseat

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
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
(Pc - Pa) Ap + Fs3 = Fp

__ _ UPSTREAM
(2) Actuator:
_PET
(Pt- pa)Ab- Fb- Fsl - Fs 2 = Fa z FLOW -- RETURN _ O-RiNGS

SLEEVE SPR_NG
where
Fp = pilot valve poppet seating
force, lb
Fa =pilot valve actuating force, lb
Fb = sensor bellows force, lb OPEN CLOSED
Fsl, Fs2, Fs3 =forces of the various springs

When F a > Fp, pilot valve starts to open. Figure 7-69.-Typical poppet-type check valve.
" ii.....
UPSTREAM
WING-GATE BURST FLUID P_RESSURE
HINGE
DOWNSTREAM _ UPSTREAM
V-GROOVE COINED IN
A NORSESHOE PATTERN
OUTLET BOD_I __ R_ _ R_T S_ G A
Figure 7-70.-Typical swing-gate-type check

valve.
(A} FLAT DISC BURST DIAPHRAGM
orifice holes are drilled in poppets or swing

gates to allow a controlled backflow for specific
control purpose. O_'EN CLOSE I:- ELASTOI/£R
I _ _,L_E=,L
Burst Diaphragms
When positive, hermetic sealing is essential,

in fluid-flow systems burst diaphragms are used. pROPELLANT MACHINED GROOVE
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
serve as safety valves to prevent excessive

pressures, or to initiate flow at a predetermined Figure 7-71.-Typical burst diaphragm designs.
pressure. Burst diaphragms can be ruptured
either by the upstream fluid pressure or mechani-
cal means. General design considerations for
jointed to the duct by welding, brazing, or bolt-
burst diaphragms are-
ing. This design has been satisfactory for many
(l) Type of fluid and its corrosive character- applications.
istics Figure 7-71(b) presents a poppet-type, stor-
(2) Duration of storage (especially with cor- able propellant valve with built-in burst dia-
rosive storable liquid propellants) phragm. The latter assures a positive, hermetic
(3) Method of diaphragm installation in duct seal of zero leakage during long storage periods;
or valve body it also protects the elastomer seal of the valve
(4) Method of rupture (upstream fluid pressure seat. Rupture of the diaphragm occurs during
or other means) valve opening, either from valve actuation or
(5) Burst pressure level (if upstream fluid from upstream propellant pressure. When clos-
pressure is used) and its tolerance; ing, satisfactory valve sealing is provided by
environmental temperature effects the elastomer seal only.
(6) Retention of the diaphragm after rupture Burst diaphragms are made from a wide variety
(no metal particles must be ejected) of metals. Annealed aluminum alloys such as
(7) Allowable pressure drop across the burst 1100-0 and 6061-0 are among the most easily
diaphragm controlled. Burst diaphragms rupture in a com-
Figure 7-71(a) presents a typical flat-disk bination of shearing, bending, and tearing. The
type burst diaphragm. A V-groove has been exact burst pressure of a specific diaphragm
coined into it in a horseshoe pattern. The depth design can only be evaluated experimentally.
of the groove will determine the burst pressure. Variations of material ultimate strength have a
The uncoined section serves as a hinge during pronounced effect. Consistency of the desired
rupture and as a retainer. The diaphragm can be diaphragm burst pressure level will be greatly
enhanced by the following precautions in manu- --'A _ EXPLOSIVE CHARGE
facturing and quality control: --=B ° RAM
"C _ METAL I"O _4ETAL SE_L C

(1) Grinding of raw sheet or diaphragm to
insure uniform and accurate thickness
(2) Close control of groove depth, by preci- b,,d_t ,-'_" _,,o_L_,O_ _ Oo_", _ F
sion coining or machining, and by stress
relieving before and after coining
(3) Continuous testing of the diaphragm mate-
rial for hardness INLET
(4) Close control of the clamping pressure

with bolted diaphragms {o) BEFORE FIRli'qG (b) AFTER FIRING
(5) Close control of the welding or brazing

processes to minimize heat effects Figure 7-72.-Typical explosive-actuated pilot
valve.
(6) Proper dimensioning of diaphragm thick-
ness and groove depth to allow for cor-
rosive effects of the propellant during "D." The seal of the valve is provided by a
storage solid metal diaphragm "E," machined as an inte-
It is possible to hold burst pressure varia- gral part of the valve body. Upon an electrical
tions of a specific diaphragm design to within signal the explosive charge detonates and drives
_+2percent for diaphragms of over 1 inch diam- the ram forward to cut out the seal diaphragm as
eter and design burst pressures greater than 300 a single piece of metal and hold it firmly against
psi. For diaphragms of smaller size and lower the end plug "F." Ram "B" has a tapered head,
burst pressure, this tolerance may increase up to which rams into the guide-hole edge and causes
±5 percent. a perfect metal-to-metal seal at point "C." This
prevents contamination of the working fluid with
gases from the explosive charge. The input cur-
Explosive-Actuated Valves
rent required for this valve is about 0.5 ampere;
In certain applications, a valve may only be actuation of the valve takes about 0.002 second.
required to operate once (to either open or close). Explosive cartridges may also be used to
In these cases, explosive-actuated valves pro- actuate valves of larger size, such as the main
vide the smallest possible size and weight. propellant valves. In the form of threaded plugs,
Since the power of the explosive actuation is they can be attached to the valve actuator ports
virtually unlimited, a rigid, hermetic zero- in lieu of pneumatic lines. For maximum relia-
leakage-type seal can be used with this type of bility, two separate cartridges are often installed.
valve. The pressures produced by the cartridges range
Figure 7-72 presents a typical explosive- from approximately 2000 to 10 000 psi. Tile valve
actuated pilot valve developed and produced by designer will have to determine the required
the Conax Corp. The actuator of this normally pressure and volume of the gases at the end of
closed valve consists of an explosive charge the actuator stroke, which then serve as the
"A," and a ram "B" including a shearing head specification to tile cartridge supplier.
Chapter VIII
Design of Propellant Tanks
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
STRUCTURAL CONNECTION TO PAYLOAD (INTERFACE OR FWD SKIRT)
GRAIN
DIAPHRAGM
fUEL TANK
COOLING
BURST DIAPHRAGM TYPE MAIN FUEL VALVE

REMOVABLE
IGNITER
BURST DIAPHRAGM TYPE

MAIN OXIDIZER VALVE
THRUST CHAMBER
ASSEMBLLES
PRESSURE
RELIEF TYPE
REGULATOR
VENT
SOLID PROPELLANT GAS GENERATOR
Figure 8-1.-Propellant tank design configuration of a typical prepackaged storable liquid propulsion
system.
_-6.0 TYP-_ --_ _---I.25 TYP
,.',--TYPICAL SKIN-STRINGER SECTION
l ,, ,,
[ :
, i!
_ ___ _ __ '_/'_ _ iT _'_ _
IZ_TNTER-STAGE _MAIN TANK WALL LN"' _ /FUEL IOXIDI_Z_T_--THRUST S STEM
D_;TwR
UC T URET } FUEL L__ ER __2111E22D I_IC_2_ON D UCT DUCT SKIRT MOUNTS ! M'PEFE
(INS. OPTIONAL FOR CRYOGENIC PROPELLANTS)
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
tanks may be insulated independently, for cryo-

genic propellant service.
Propellant Tanks for Upper Stage Systems
Many upper-stage vehicles employ a pressur-
Figure 8-3 presents the propellant tank de- ized gas feed propulsion system. Tank pres-
sign configuration of a typical pressure-fed, sures range from 100 to 400 psia. The system
upper stage propulsion system. Here, the tanks shown in figure 8-3 uses stored helium gas for
HIGH PRESSURE HELL{UM

STORAGE 8OTTLES MAiN OXIDIZER VALVE
MAiN FUEL DUCT
MAIN FUEL VALVE
TO FUEL TANK SUPPORT
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.
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
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
tions at ambient pressure are used for the cryo-

genic propellants. The tank ullage-volume cal- The fuel tank ullage volume
culations should allow for propellant volume
changes due to temperature change of the tanked Ut= (Vi + Ti) × 0.02,5= 140 × 0.025 = 3.5 ft3
propellant, and for tank deformation when pres-
surized. This is especially important for pre- From equation (8-i), the required design vol-
packaged storable liquid systems to prevent ume of the fuel tank
excessive tank ullage pressures when the system
is exposed to a specified upper temperature limit Vii = VI+ TI+ U[= 13S.2 + 1.8 +3.5= 143.5 ft3
during storage. Adequate ullage volume is also
required to maintain tank pressure at starting
Propellant Tank Arrangement
when the relative ullage volume increase is large
and may tax the response of the pressurization In most vehicle systems, the propellant tanks
system. are arranged in tandem. Other arrangements are
used for specific design reasons. Figure 8-4
Sample Calculation (8-i)
The following data are given for the A-4 stage

propulsion system, including two engines:
Oxidizer (N204) density, 90.12 Ib/ft 3
,,L
-_DRYC.G./( )_
Oxidizer weight flow rates, 12.78 Ib/sec/
engine
(ol TANDEM PROPELLANT TANKS WtTH COMMON BULKHEAD
Fuel (N2H4) density, 63.17 Ib/ft s AND SEPA.R-STE HEUUM BOTTLES -

TANKAGE WEIGHT, IOO PERCENT
Fuel weight flow rate, 10.65 Ib/sec/engine
Nominal engine firing duration at full thrust,
410 sec
Trapped oxidizer volume, T o = 0.9 ft 3
Trapped fuel volume, Tt= 1.8 ft 3 .... / , _ DRY CG. /II I_
Tank ullage volume, U = 2.5% of propellant

I
volume [b) TANDEM PROPELLANT TANKS WITH INTEGRATED
HELIUM BOTTLE IN THE MIDDLE -
Determine the volume of the propellant tanks. TANKAGE WEIGHT, 93 PERCENT
Solution ooo4aO_,5 f 0,o86 o isoJ _
The required usable oxidizer volume

.Tp,
YCG\
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
The oxidizer tank ullage volume

i-o 3TB o,o2s--_ F_
.... DRY CG.-_O.tBB-, _/_'_ J
Uo=(Vo + To) x 0.025= 117.I ×0.025=2.9 ft3

I _'--0.130
.E.... E- O, OBO
to. I-.,,;
0.020-_
FUEL i
OXI'OIZ E R _ _ ._ "_'-'_ I"_
From equation (8-i), the required design vol-

(d) MULTIPLE PROPELLANT TANKS AND HELIUM BOTTLE
ume of the oxidizer tank IN CLUSTER- TANKAGE WEIGHT, 16; _ P_RCENT
Figure 8-4.-Various propellant tank arrange-

Vto=Vo+To+Uo=l16.2+0.9+2.9=120 ft3 merits of a typical vehicle syste_n.
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
(1) Internal pressure loads and their dynamic Sw = 1.65 (8-7)

effects
(2) Axial thrust loads and their dynamic where
effects Sw =maximum allowable working stress, psi;
(3) Bending moments due to vehicle trans- i.e., the stress due to maximum tank
verse accelerations, wind loads, and working pressure under normal transient
shifting of the center of gravity and steady operating conditions
(4)Aerodynamic forces Fy = yield strength, psi, of the tank construc-
(5)Thrust vector control forces tion material, at operating temperature
(6)Vibration loads conditions
(7)Loads produced by mounting arrangement F u = ultimate strength, psi, of the tank con-
(8)Loads caused by thermal transients and struction material under operating tem-
gradients perature conditions
(9) Loads produced during ground handling Sw is calculated for both Fy and F,. The
In most vehicle systems, internal tank pres- lower value should then be used. All propellant
sure loads and axial thrust loads are the princi- tanks are subjected to hydrostatic pressure tests
pal ones. Other loads require careful evaluation, prior to acceptance. For case (1), the proof test
including model tests or full-size experiments. pressure equals maximum tank working pressure.
For case (2) and (3), the proof test pressures or

should be II0 percent of the maximum tank work- Sw=IF._5 45000
= 1.-']-_--5-o: 33 360 psi
ing pressure. For high-pressure inert gas, long-
duration storage bottles, the following correla- The maximum allowable working sttess for
tions are recommended: mainstage = 31 S00 psi.
No hazard to personnel: (b._) Tank made of aluminum alloy 6066-T6.-
During start transient
Fy (8-8)
Sw - 1.33
Fy _ 50 000_ 37 600 psi
Sw= 1.33 1.33
Hazard to personnel: or
Fu _57000 34 560 psi

Sw-1.33 1.65 :
Fy (S-9)
Sw - 1.67
The maximum allowable working stress during

start transient = 34 560 psi.
During mainstage operation
The start transient of a prepackaged storable
liquid propulsion system for an aircrafblaunched
missile is programed not to reach main-stage s., _FY
- i = 50000 45 400
i.---/-: psi
level until the missile is at a specific distance or

from the aircraft.
Fu 57 000
Calculate the maximum allowable working Sw=-f_ - 1.35 - 42200 psi
stresses for the propellant tanks, if they are
made of-
The maximum allowable working stress during
(a) Aluminum alloy 6061-T6, Fy = 35 000 psi,
mainstage = 42 200 psi.
Fu = 45 000 psi
(b) Aluminum alloy 6066-T6, Fy = 50 000 psi, Material and Fabrication Considerations in
Fu = 57 000 psi
Propellant Tank Design
In addition to considerations of propellant

Solution
compatibility and operational temperature ranges,
(a) Tank made of aluminum alloy 6061-T6.- selection of construction materials for propellant
Since the system involves personnel safety during tanks is based on their strength-to-density ratio
start transient, equations (8-6) and (8-7) will be at a given temperature and on their ductility.
applied. For a given working pressure, the lightest tank
structure will be the one made of the material
Fy 35 000
with the highest ratio of ultimate strength to
Sw-1.33 1.33 =26300 psi
density. Most frequently used construction ma-
or terials for propellant tanks are:
Fu 45 000 (1) Aluminum alloys, such as 6061-T6,
Sw=1.65- 1.65 =27260 psi 6066-T6, and 2014-T6. Room temper-
ature properties: Average density p
Thus, the maximum allowable working stress =0.1 lb/in 3. Fy up to 60000 psi, Fu up
during start transient = 26 300 psi. to 70000 psi, Fu/p=70× 104
During mainstage operation, personnel are (2) Stainless steels, such as AISI 347 (for
considered safe and equations (8-4) and (8-5)can low-pressure tanks only), 17-7 PH and
be used: PH 15-7 Mo. Room temperature proper-
ties: Average density p =0.285 lb/in 3,
Fyupto200000psi. Fu up to 220000
Fy _ 35 000 = 31 S00 psi
Sw=l.l I.i psi. Fu/p=77.2×lO 4
(3) Fiber glass, filament wound with an

aluminum-alloy liner. Room temperature
GORE SEGMENTS_
properties (fiber glass only): Average
density p= 0.0S Ib/in 3, Fu = 120000 psi.
Fu/p = 150 x 104
CIRCUMFERENTIAL
Aluminum alloys are compatible with most
storable and cryogenic propellants, and may be
used for working temperatures up to 350 ° F.
Stainless steels are suitable for storable (lim-
ited duration) and cryogenic propellants, and are
suitable for higher temperatures (800 ° F maxi- 4...
mum). Fiber glass is limited to moderate tem-
LONGITUDINAL
......
.............
perature conditions (-60 ° to 160 ° F).
Fabrication methods for propellant tanks
depend largely upon the type of material used. WELDS_..: .......................
Most important considerations for tank fabrica-

tion are dimension control, heat treating, and
welding. The lowest recommended wall-stock
size for propellant tanks is about 0.010 inch for
stainless steels and 0.020 for aluminum alloys.
While tank stress calculations must consider the
lower limit of wall-thickness variation, the upper
Figure 8-5.-Typical welded propellant tank
limit is used for tank-weight calculation. The construcLion.
strength of a metal may fall into a band, too, the
width of which depends on the heat-treating subject to aerodynamic heating, the filled por-
process. Stress calculation will be based on the
tion may be at a very low temperature. Addi-
minimum expected strength. The quality of the
tional thermal problems may arise in outer space,
welding process or the efficiency of a welded
from solar heating of one tank side and radiation
joint may require extra stock added to the wall cooling of the other.
thickness as calculated from other working
Also partly the tank designer's responsibility
stresses. An assumed weld efficiency of 85 to
is the solution of certain problems associated
95 percent is a reasonable value for steels (50
with the management of the propellants within
to 65 percent for aluminum). To minimize weight the tanks, such as:
penalties, build-up lands may be used at the (1) Uniform dispersion of the entering tank
welds (fig. 8-9) for an equivalent 100 percent
pressurant
weld efficiency. Figure 8-5 presents the con-
(2) Sensing of propellant quantities (PU)
struction of a welded propellant tank. Note the
(3) Prevention of propellant sloshing
segmented tank end which is typical for large- (4) Expulsion of the propellants under ad-
size tanks.
verse conditions
(5) Fill, drain, vent and pressure relief of the
Other Propellant Tank Design Problems propellant tanks
Many other design and analysis problems will

8.3 STRUCTURAL DESIGN OF PROPELLANT
have to be carefully considered before a suc-
TANKS
cessful propellant tankage can be produced. The
relatively thin, highly stressed shells make it As a rule, the wall thickness of propellant
difficult to attach concentrated loads. The loads tanks is first calculated from stresses caused by
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
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:
Spherical Tanks _._a2b

Ve - 3 (8-15)
(1) Volume, in3:
Spherical tank end volume, in3:

4 7ra3
Ys - 3 (8-10)
2 _a a (8-16)
Vs=---ff--
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
used to calculate the weight of the tank

,2° ...... _ . :;. ...... _L_ ....3
ends. K
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)
ioo _2_ ESo 175 200
(8-21) =
ts (g+_) (LkiPS[ RATIO
"K'NOT[ FOR SPH[RICAL T&NK$ USE K* 050 &T k it 0
FOR SPH£RICJL MEADS USE K-0 $7 &T i ,_ 0
Figure 8-7.-E11ipse ratio k versus knuckle stress

(8-22)
tc 2 /actor K, compression stress-K, and parameter
E'. (From ARS paper, "Design Criteria and
where Analyses for Thin-Walled Pressurized Vessels
k = tank end ellipse ratio = a/b; k = 1 for a and lnterstage Structures," by T. J. Hart.)
spherical end
R =tank end crown radius, in=ka; R=a for a
spherical end where
K = stress factor, a function of the ellipse
ratio k. Figure 8-7 presents a K versus e =eccentricitv. - a - k2
k curve for combined membrane, discon-
tinuity, and local bending stresses (4) Weight:
tk = wall thickness at the knuckle, in Ellipsoidal tank end weight, lb:
tcr = wall thickness at the crown, in
te : equivalent wall thickness of an ellipsoidal 2 i
_,a teE p
tank end, in We - 2k (8-25)
ts =equivalent wall thickness of a spherical
tank end, in
Spherical tank end weight, lb:
tc :wall thickness of a cylindrical tank sec-
tion, in
Ws = 2 _a _tsp (8-26)
(3) Wall surface area:
Ellipsoidal tank end surface area, in2: where
k+ 1
r(1 + e)-] E' --design factor =2 k + 1 in
\/k - 1 k- \,/-k_ - 1
.b 2 In [f_-_j (8-23)
Ae=a_÷ 2e " (see fig. 8-7)
(5) Critical pressure due to external loading,
Spherical tank end surface area, in2: psi:
For an ellipsoidal tank end, it may be
As=2 rta 2 (8-24) approximated as
Cb2Ete 2 (5) Weight, lb:

Pcre: R2 (8-27)
Wc = 2 =alctcp (8-32)
For a sphericaltank end
(6) External loading critical pressure, psi:
0.342 Ets _ For a short tank (i.e., lc<4.gav/a,-_):
Pcrs -- a2 (approximately) (8-28)
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
(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
Solution quired wall thickness at the knuckle of the oxi-

dizer tank end
(a__) Since the oxidizer tank consists of two
ellipsoidal ends without a cylindrical section,
Kpta _ 0.80 × 180 x 41
equation (8-15) may be applied: =0.171 in
tk° = S w 34 560
2 × 2 _ra2b
Vt° - 3 From equation (8-18), the required wall thick-
ness at the crown of the oxidizer tank end
We rearrange to obtain the minor elliptical

half-diameter of the tank ends: ptR 180× 1.395×41
tcro = =-=--= -0.149 in
z3w 2×34560
3xVto 3x120x1728 29. 4

b=2x2_a2- 4 x ,_x (41)2 in From equation (8-19), the equivalent wall
thickness of the oxidizer tank end
The tank end ellipse ratio:

(tko+tero) (0.171+0149)_0.16in
a 41 re°= 2 - 2
k =-_= 2---_-_.4
= 1.395
NOTE: In some designs, weight has been saved
by taking advantage of the fact that the bulkhead com-
mon to both tanks is subject to a relativelysmall dif-
ferentialpressure in operation. Such systems, how-
Since we use an ellipsoidal end of the same
ever, requiremore elaboratepressurizationand loading
proportion at the fuel tank top, the fuel tank vol-
systems including interlocks. In case of malfunction,
ume may be treated as the volume of a cylindri- the common bulkhead may suffer seriousdamage.
cal tank section with the length lc. From equa- The required wall thickness at the knuckle of
tion (8-29), the volume of the fuel tank: the fuel 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:
equations (8-6) and (8-7), the maximum allowable

working stresses are derived:
(tkf + tcrf) _ (0.162 + 0.141) = 0.152
tel= 2 2
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:
Fu _ 57 000_ 34 560 psi

Sw=1.65 1.65
ptfa_ 170x 41 _ 0.202 in
tc - Sw 34 560
We use the lower value of 34560 psi.
From figure 8-6, the tank end stress factor K Provide a buildup of 0.4 t c on the cylindrical
of the combined stresses is 0.80 for an ellipse tank section wall near the juncture to allow for
ratio k of 1.395. From equation (8-17), the re- discontinuity stresses:
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
tko=0.171 in, tcro=0.149in, teo=0.16in

tkf=0.162in, tcrf=0.141, ter=0.152in We find that Ic<4.9av'a/tc. Thus equation
(8-33) applies, yielding an external loading criti-
tc=0.202 in, tcj =0.283 in
cal pressure for the cylindrical tank section
(c_) From equation (8-25), the weight of the

oxidizer tank end: 0.807 × 10.4 x 106 × (0.20.,)-
Pcrc- 46.9× 41
_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
Weotef_ 139.5 x 0.152 withstand large axial compressive loads during

_/ef--- te ° 0.16 - 132.8 lb vehicle handling and operation. It' the tank is
not pressurized, i.e., if tank pressure = ambient
From equation (8-32), the v)eight of the cylin- pressure, the critical axial compressive stress
drical tank section: for an unstiffened cylindrical tank may be calcu-
lated as
14Ic= 2 rralctcp = 2 x 41 x 46.9 x 0.202 x0.101= 246.4 lb

tc 13
Add 4 percent of overall tankage weight for

local wall thickness buildups, to allow for
welded joints, discontinuity stresses, etc., and where S c = critical axial compressive stress, psi.
for tolerances during fabrication. This is the axial compressive stress
that will cause the tank to buckle
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)
Cb2Eteo 2_0.10x 2x 10.4x 106×(0.16) 2

Pressurization will also reduce tank failures
Pcreo- R 2 (1.395 x 41)2
from very large bending loads. However, if the
= 16.3 psi pressure is ever permitted to drop below a value
necessary to carry the axial and bending loads,
The critical external loading pressure for the the tank will collapse and the vehicle will prob-
fuel tank end: ably be damaged beyond repair.
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)
Sample Calculation (8-4) c= / 1 + [-25 EPalEtc

(1- 0"8 v)
For the A-4 stage tankage, calculate: where

(a.) Critical axial compressive load of the Ps =pressure surge due to the impact, psi
cylindrical tank section with no internal
= equivalent flow rate of the propellant due
pressure. to the impact, lb/sec
(b__)Required internal tank pressure to offset a =radius of the cylindrical tank, in
an axial compressive load of 100 000 tc = wall thickness of the cylindrical tank, in
pounds with no compressive stress on g =gravitational constant, 32.2 ft/sec 2
the cylindrical tank section. c = acoustic velocity of the restrained propel-
lant, in/sec
Solution c' = free acoustic velocity of the propellant,
in/sec
(a_) From equation (8-35), the critical com-
Ep = compressive modulus of elasticity of the
pressive stress of the cylindrical tank section: propellant, psi
E = modulus of elasticity of the tank construc-
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
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
8.4 DESIGN OF STORABLE LIQUID

Solution PROPELLANT TANKS
(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.
From equation (8-38), the acoustic velocity of Tank Material Compatibility

the restrained fuel
Prior to final selection, a detail study is re-
80 100 quired to determine the compatibility of potential
C = ,,.
tank construction materials with the propellant
combination involved. In this study, two major
considerations should be kept in mind: (1) the
=43100 in/see expected storage life of the propulsion system,
and (2) the credibility of the available data.
Impact time delay in tile tank: For a required storage period of less than 3
years, the compatibility data for many materials
lc 50 with various propellants are generally reliable.
--= --=l.16x10 -3 sec(i.e., <1.2×10 -a sec)
c 43 100 However, for longer periods, such as a 10-year
storage requirement, the design data are usually
Using equation (8-37), we obtain the tank extrapolated from the results of relatively short-
pressure surge due to impact after a 6-foot free term testing. Figure 8-8 illustrates the typical
drop: extrapolated corrosion and passivity rate of
aluminum alloy 6066-T6 exposed to tile liquid
c#,, 43100×434.5 phase of inhibited red fuming nitric acid (IRFNA).
Ps = _a-_-g= 7r×(4) 2 x386 =964 psi Much of the available compatibility test data
tend to contradict each other. This is the result
(b__)After a 20-foot free drop, the final veloc- of differing test methods, conditions, and criteria.
ity at impact is It involves the risk that although a material may
be acceptable for a relatively short exposure, it
V=_=,_2×32.2×20=35.9fps may fail when the exposure time to a particular
A full-penetration "100 percent" weld is one

where complete fusion has taken place to the
root of the joint. Unfused edges, oxide notch, or
I any concavity at the root indicate lack of full
0.32
penetration. All full-penetration welds which
cannot be visually examined on the reverse side
must be X-rayed to assure full penetration. Fig-
ure 8-9 presents the design detail of a typical
full-penetration, single-welded butt joint.
I0 If possible, interpropellant welding (i.e., the
fuel and oxidizer are separated by only one
Figure 8-8.-Typical extrapolated corrosion and welded joint) should be avoided in tank designs.
passivity rate of aluminum alloy 6066-T6 ex- A double-weld design will prevent the mixing of
propellants due to a single failure of welded
posed to the liquid phase of IRFNA.
joints. Figure 8-10 presents the design of a
typical storable propellant tank with a forged
environment is increased. Whenever a discrep-
one-piece common bulkhead. A double-weld de-
ancy of materialcompatibilityis observed, thor-
sign is used on the center-hold joint for the
ough verification of the testconditionsis in-
oxidizer duct.
dicated. For most of the long-termstorable
propellanttankage,the aluminum alloys are used
forbest results.Austenitieand semiaustenitic Tank Wall Surface Requirements
stainlesssteelsshould be used only forrather
The inner surfaces of the storable propellant
short-term storage.
tank wall must be relatively free of rust deposits
Tank Construction
/
One of the prime design objectives of storable

propellant tankage is the elimination of mainte- _° T,. ,L ----
nance during storage and prefiring activities.
This may be accomplished by an all-welded,
hermetically sealed construction. To achieve fft_P orr_L
hermetic sealing, leakproof burst diaphragms are

installed throughout the inlets and outlets of the Figure 8-9.-Design detail of a typical full-
tankage. Propellant fill ports are welded shut penetration, single-welded butt joint for pro-
after tanking. Tandem tanks are favored over pellant tanks.
concentric tanks, since propellants are more
positively separated during storage, transporta- FO_3-F_D ONE - PIECE
tion, and other adverse conditions. In some ap- COMMON BULKHEAD
plications, the common bulkhead between two N_R

-WELD DE95N
-PROPELL3_gr
APPLJED
JOINT
propellant tanks is designed to withstand a dif-

ferential pressure in either direction that may
, L T_NK t_
result from inconsistent burst diaphragms. How-
ever, the burst diaphragm at the pressurant inlet ---_-_--_
to the oxidizer tank of the A-4 system is de-
signed to a lower burst pressure
pressurant
This insures
always the higher.
than that of the
inlet diaphragm for the fuel tank.
that the oxidizer tank pressure is
FULL PF'N ETR£rlON
WELDED JOINT
OX$1Z_l
/'2
D_T
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.
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
boiloff of the cryogenic propellants system becomes apparent.
(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,
to provide an insulating blanket on the tanks,

which is remotely removed just prior to liftoff.
During the boost phase, high temperatures from
aerodynamic heating, and large aerodynamic
forces are encountered. Although of short dura-
tion, this phase dictates the structural elements
of the insulation. During coast flight in space,
8 the principal source of thermal energy is radia-

tion from the Sun and the planets. By the use of
6 radiation shields surrounding the basic tank in-
w sulation, the heat flux across the tank wall can
be effectively controlled. The properties of the
S,&'TURA'r_DN TEM_'EF_TURE AND PRESSLIR£ O_ UOUI'D H_fDROGEN_

materials used in a solar shield are very impor-
tant, such as the absorptivity and emissivity of
Figure 8-I1.-Rate of change of saturation vapor
the surface when subjected to various types of
pressure to temperature [or liquid hydrogen.
radiation and body temperatures. Magnesium
oxide and silver are two materials showing prom-
therefore, is specially slanted to hydrogen ise for use in solar shields. These materials
service. may be applied as coatings onto a lighter base
material such as aluminum. The amount of
shielding required depends on the duration of the
Insulation Requirements for the Cryogenic
coast flight and on optimization of shield weight
Propellant Tanks
versus propellant boiloff weight.
Most vehicle missions include three distinct
phases during which effective insulation for the
cryogenic propellant tanks is essential: ground Basic Insulation Types
hold period, the boost phase, and coast flight in
Major desirable design features of an insula-
space. The propellant tank insulation design
tion include: light weight, uniform and repeat-
will primarily be based on flight performance
able insulation characteristics, ease of applica-
criteria. However, propellant evaporative losses
tion, low cost, low hazard, reasonable ruggedness,
during long hold periods on the ground requiring
ease of repair, good reliability, and above all,
continuous topping may also become a signifi-
low heat conductivity.
cant cost item. If long hold periods with a filled
Excellent results can be achieved with a
tank are anticipated, it may be indeed economical
laminated-type insulation. This employs an
aluminum foil and fiber-glass structure, often in
multiple layers. Tim aluminum foils act as re-
flectors, effectively rejecting radiative heat,
while the evacuated space in between prevents
conductive heat transfer. This insulation can be
applied to single-curved and to large-diameter,
double-curved surfaces. The laminar insulation
is sensitive to damage, however, possibly re-
sulting in loss of vacuum due to cracks and to
o=t-
infiltration of leak gases. Application to air-
borne system thus has been infrequent.
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
plied hydrogen tank insulation of this type. A

Figure 8-12.-Rate of change of saturation vapor _- to _-inch nominal-size plastic honeycomb is
pressure to temperature [or liquid oxygen. installed between an inner and outer facing
v,
ON[_C_U_ CC_[ wao_ O_ Solution
@_S _ • Z 2 Le _FT 1 -- ATMOSPH£RE
\\ (a) During ground hold, [he temperature dif-

ferential across the insulation XT= 70-(-360)
= 430 _ F. The overall thickness of the insulation
.......... _k. o;,. /// t=2×0.03_0.75=0.81in. From equation (4-19),
the heat transfer rate
_¢o @Y _Euu_ 5AS
LIQUIO HYDROGg N
= 2.02 × 10-4 Btu/in_-sec

Figure 8-13.-Construction elements of a typical
liquid hydrogen tank insulation design (exter-
(b) During the boost phase, AT = 800- (-210)
nal type).
= 1010 ° P. The heat-transfer rate
sheet. The cells may or may not (weight!) be

[k'_,_ /3.8 × 10-?_
filled with an isocyanate-type foam. The foam
bubbles or the properly sealed ceils will form
individual vacuum spaces when cold (cryopump- :4.74× 10 -_ Btu/in_-sec
ing). However, because of the possibility of
vacuum degradation by infiltration of air (outer
Selection of Tank Insulation Designs
insulation), or hydrogen (inside insulation), it is
often preferred to purge the ceils with helium, Many factors will influence the selection of a
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
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
be produced with mechanical precision and ver-

satility of construction and materials. A strength-
to-density ratio, Fu/p, can be obtained by fila-
II i ............... 14' i (-1 ment winding and impregnating the fiber-glass
// M co,_ • c,.os.,-ouT I _il; '_" __:"_R _'''L
roving with a suitable resin, which is higher than
can be achieved by any other method or material.
For instance, the strength-to-density ratio of a
fiber-glass filament-would resin-bonded structure
(average p =0.08 lb/in 3, Fu = 120 000, F,/p
= 150 x 104) is about twice that of high-strength
i_ _ECr,O_
_-. alloy steels.
SECTION A-A
Good reliability and close dimensional control
Fig.re 8-14.-Design o[ a typical insulated com- of filament-wound pressure vessels is assured
through the use of calibrated winding machines
mon bulkhead separating LH 2 and LO 2 tanks.
which orient the reinforcing fibers precisely.
However, most of the liquid propellants are not
insulation can all be done without special con- compatible with resin-impregnated, fiber-glass
sideration of access, ventilation, and curing structures. Therefore, an aluminum tank liner
techniques in closed areas. However, special which is compatible with both propellants, must
consideration will be required for tank leak de- be used to separate the fiber glass from the liq-
tection, especially with integrally insulated uid. It also provides a positive sealing barrier,
areas. If a crack does occur in the external in- since the resin-bonded fiber-glass material is
sulation, liquefaction of air and cryopumping pervious to both liquid and vapor over extended
will occur resulting in a significant rise in heat periods. Figure 8-15 presents a typicalaluminum-
transfer. lined, fiber-glass filament-wound liquid propel-
lant tank.
Design and fabrication of fiber-glass filament-
Insulation for Common Bulkheads
wound tanks with a thin aluminum liner is
When a common bulkhead is used between basically simple. However, a key problem arises
propellant tanks and cryogenics are involved, from the fact that the modulus of elasticity of
insulation is required to prevent freezing of the fiber-glass resin-bonded materials is about 5×106
propellant with the higher boiling point. Figure psi, while that of aluminum is about 10x 106 psi,
8-14 shows the equatorial area of a typical in- and that the strength of the aluminum is usually
sulated common bulkhead. A fiber-glass honey- no more than a third of that of the fiber glass.
comb is located between a forward and an aft Since the geometry of the tanks usually imposes
facing sheet. In the equatorial area, where the equal strain on aluminum liner and fiber glass as
facing sheets have to be faired into the cylindri- the tank is pressurized, the aluminum becomes
cal tank portion, they are reinforced by waffle loaded to its elastic limit long before the fiber
grid ribs. This design also requires special glass reaches the level of its high-strength capa-
insulation on the inside of the LOX tank in the bility. Consequently, with a plain liner config-
equatorial region. No foam filler was used in uration, the aluminum is stretched far beyond its
tkis particular example. elastic limit and forced back to its original shape
each time the tank is pressure cycled. Thus, the
liner may experience fatigue failure after a small
8.6 DESIGN OF FIBER-GLASS FILAMENT- number of pressure cycles. The pressure-cycle
life of a plain liner depends on amount of stretch
WOUND LIQUID PROPELLANT TANKS
beyond the elastic limit, type of aluminum, bond-
Most of the pressure vessels which possess ing between lines and fiber glass, weld joints,
surfaces of revolution can be fabricated conven- variation in thickness and contours, etc. Awell-
iently by the technique of filament winding. By designed, plain-type tank liner should have a
this method a lightweight structural member can life of about 10 to 20 pressure cycles.
tural members. In these cases, a separation tube

(Hilsch tube) may be combined with the diffuser.
Without moving parts, it operates on the principle
of separating the higher energy molecules from
those with lower energy. Temperature spreads
of 100 ° F or more can be obtained, depending on
construction and available pressure drop. Ad-
vanced configurations have been developed by
Figure 8-15.-A typical aluminum-lined, tiber- the AiResearch Manufacturing Co. In a LOX/
glass, filament-wound liquid propellant tank. hydrogen system with common bulkhead, the de-
vice may be used to direct tim cold stream toward
the bulkhead, thus lowering the temperature dif-
ferential and resultant heat transfer across it.
One design approach to extend cycle life is
to use a corrugated liner. The corrugations have
an effect equivalent to reducing the modulus of i
elasticity of the aluminum liner to a value less

than that of the fiber glass. For instance, if the
fiber glass is stressed to 100 000 psi and has a
modulus of elasticity of 5× 106, its extension
will be 2 percent. The corrugations of the liner
then should be designed to permit the 2-percent
extension so that the aluminum will not be /' ' M _"c._._-_ _," -,"_:--_f_"
stressed beyond its elastic limit.
\\"1
--_ /' :z_ 4 .'
/ )"4'-,. ./
8.7 DESIGN OF PROPELLANT TANK ", ' , d_-< _ l],
PRESSURANT DIFFUSERS
The main function of a pressurant diffuser

is to introduce the pressurant gases evenly into
the propellant tank at a desired direction and *
velocity. The gases are usually injected in a

Figure 8-16.-Typical designs o[ propellant tank
plane near the tank forward end, at a right angle
pressurant diHusers.
to the tank and vehicle axes. This will minimize
disturbances at the pressurant-propellant inter-
face. Figure 8-16 presents some typical propel-
8.8 PROPELLANT EXPULSION UNDER
lant tank pressurant diffuser designs. A radial-
ZERO GRAVITY OR OSCILLATORY
type diffuser, located at the tank axis, is used
G-LOADING CONDITIONS
for the fuel tank. This arrangement permits a
simple, lightweight diffuser design. However, Under zero or oscillatory g-loading condi-
the reversed ellipsoidal common bulkhead at the tions typical for many vehicle trajectories, fluid
forward end of the oxidizer tank requires a ring- location in a propellant tank becomes uncertain
type diffuser, consisting of many individual dif- and requires means to prevent gas pockets from
fusing nozzles located at the circumference of interfering with propellant expulsion. The pro-
the tank. pellants must be either subjected to an accelera-
The pressurant enters the tank at a tempera- tion force for settling immediately prior to usage,
ture level depending on the source, such as a or they must be continuously confined in the
LOX heat exchanger. In some cases, this may tank. The two approaches can be defined as
not be the optimum temperature for critical struc- impulse settling and positive expulsion.
Impulse Settling propellants and higher temperature pressurant

gases. Figure 8-17 presents a concentric-
This method employs a small propulsive force
circular, convoluted metal positive expulsion
directed axially, parallel to the vehicle center-
diaphragm used in a spherical tank. This sys-
line of thrust. This acceleration forces the pro-
tem is unique for bipropellant systems because
pellant to cover the tank outlet prior to initiation
it allows both propellants to be stored in the
of main engine operation. This approach is es-
same tank, as long as the volumetric mixture
pecially suitable for most space vehicles, because
their low-thrust reaction control units can be ratio does not deviate appreciably from unity.
An expulsion efficiency above 97 percent (i.e.,
used to perform the auxiliary function of provid-
the percentage of the propellant which can be
ing the thrust required for propellant settling.
expelled from the tank) may be expected. This
Although this method would eliminate the
design also minimizes shift of center of gravity
need for positive expulsion devices for the main
during the entire process of expulsion.
propellant tanks, it would necessitate separate
Figure 8-18 presents the corrugated metal
positive-expulsion propellant tank or tanks for
positive expulsion diaphragm applied to a cylin-
the sole use of the reaction control system. The
drical tank. A center post is provided to support
possible disadvantages of this impulse-settling
the diaphragm. The diaphragm is formed into a
method are (I) no control of vehicle center-of-
lobed corrugation capable of performing the re-
gravity shifts, and (2) low thrust-to-weight ratio
quired mechanical expansion. Transition sec-
under these conditions, which may increase
tions are provided at both ends of the diaphragm
response times beyond tolerable limits.
for attachment to the tank ends. An expulsion
For single-start upper stages, jettisonable
efficiency of 97 to 98 percent may be achieved
short-duration solid settling rockets may also be
with this design. However, the utilization of
applied to advantage.
_SSVRAmr GAS
INLEt
Positive Expulsion _O_C[_ R_¢ _RCU_t_ ¢VmvoLwr_
The other method of achieving proper fluid

orientation within the propellant tanks is by con-
tinuously confining the propellant to the vicinity <
; \
of the tank outlet. A positive expulsion propel-
lant tank assembly usually consists of an outer
tank shell and an inner movable expulsion de-
vice. Frequently used positive expulsion de-
vices are: (1) metallic diaphragms, (2) elastomer
diaphragms, and (3) moveable pistons. The
design of the expulsion devices depends upon a
number of considerations. Primary among these Figure 8-17.-Concentric-circular convoluted
is the design configuration of the propellant metal positive expulsion diaphragms used in a
tank. Another is the allowable shift of vehicle spherical-shape propellant tank.
center of gravity during expulsion, which in many
vehicle applications must be minimized to avoid
instabilities induced by a pitch or yaw maneuver.
A choice of construction materials capable of
sustained contact with the pressurants and pro-
pellants must also be made.
Metal Expulsion Diaphragms
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.
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
ANTI - EXT_u S4ON _TE
Cai.ON _t_eAo_ ¢_e_rm_
I, _ p_pe_LA_r
_r_eT
_'-LAOO_[R _ £LASTOMgR POSITIVE

_*_G SelL kl CY_Na_C_L 11_
Figure 8-19.-Bladder-type elastomer positive ex-

pulsion diaphragms used in a spherically Figure 8-20.-Movable piston used in a cylindrical
shaped propellant tank. propellant tank [or positive expulsion.
r
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
Ap - 7r× (20) _ _ _× (1.5) 2 _ 314.16 - 1.76 = 312.4 in s

4 4
The total length of piston seals:
The following data are given for a cylindrical
positive expulsion tank with a movable piston as
Ls= _× (20+ 1.5)=67.54 in
shown in figure 8-20:
Inside diameter of the tank, 20 in
Substitute this into equation (8-39); the pres-
Diameter of the guide post, 1.5 in
sure differential across the piston
Coefficient of friction of the piston seals,
is =0.1
Unit normal force reacted by the seals, Fn _p- fsFnLs - 0.1>:500×67.54 10.8psi
= 500 lb/in Ap 312.4
Chapter IX
Design of Interconnecting Con ponents
and Mounts
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
ENGINE GIMBAL MOUNT ASSEMBLY
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
Figure 9-4.-Typical pump-discharge, high-

pressure propellant duct with restraining links.
CON_IECTbNG TO
PUMP DISC_ARC-[--_r
Figure 9-2.-Typical (lexible propellant supply

duct for a large turbopump-[eed engine system.
(Note: Duct is mounted in test fixture with
fluid pressure connections.)
8ELA._fS
RI[ST RAtNING RO0
E_O LOAO
COMPENSAT_ --
BEU.
, 1 J
O_JTCET FL._,_IG_E
CONN£C_r EO TQ
Figure 9-5.-Typical pump-discharge, high-

pressure propellant duct with end-load
compensator.
balancing the duct, thus reducing duct-separating

loads by 90 percent.
The high-pressure propellan_ lines between
pump discharge and gas generator usm_lly can be
made of wire-braided-type flexible hoses or tub-
ing, because of their relatively small size (often
less than 1 inch).
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
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
stage operation. connection from the engine to the vehicle and

beyond is required. The lines are generally
made of wire-braided flexible hoses and tubings.
Pneumatic Supply Lines
A special ease is presented by cryogenic
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
During the start and shutdown sequences, and

Figure 9-6.-Typical pump seal drain schematic. for prerun and postrun servicing, inert gas purges
DESIGN OF INTERCONNECTING COMPONENTSAND MOUNTS 357
are frequently required to keep a system dry, to Turbine-Exhaust, Hot-Gas Ducts

prevent combustible mixtures from forming, or to
The hot gases from the turbine exhaust may
expel residual propellants. If they are contin-
be ducted to near the main chamber nozzle exit
uous during ground hold, it is advantageous to or into a thrust chamber exhaust manifold. Most
feed them from ground to avoid increased flight
turbine-exhaust duets are of welded stainless
weight. This requires connecting lines from the
steel or nickel-alloy, sheet-metal construction.
englne to the vehicle. It is strongly recom-
Bellows sections permit the degree of movement
mended that entirely separate purge systems for
required by the system for thermal expansion,
each propellant be employed to avoid backup of
misalinements, and tolerances. In most cryo-
one fluid into the system of the other. Check
genic propellant systems, a heat exchanger is
valves are not considered as reliable as strict
incorporated in the exhaust duct assembly to
separation. For engine design we are concerned
vaporize one of the propellants, usually LOX,
here only with those purges which connect di-
for tank pressurization (fig. 5-7).
rectly to an inlet port of an engine part, such as
the thrust chamber cooling jacket or injector
Thrust Mounts
manifold. From the viewpoint of the environment
in which the engine must operate, the engine The thrust mount transmits the thrust loads
designer should additionally specify the purges between engine and vehicle thrust frame. In
to be provided by the vehicle builder to condi- some designs (fig. 2-4), the thrust mount also
tion the engine compartment as needed. All serves as the structural mount for many other
purge lines can be made of wire-braided flexible major components of the engine system. Thrust
hoses and tubings and may be routed as are the mounts usually consist of truss members of tubu-
drain lines.
lar and box shape, made of aluminum alloy or
alloy steels. In engine systems using fixed
thrust mounts, thrust vector control is accom-
Hydraulic Lines plished by swiveIing the thrust chamber, or by
other means such as jet vanes, secondary injec-
Various high-pressure hydraulic lines are
tion, or jetavators.
used in a liquid rocket engine for hydraulic actu-
ating, lubrication, etc. All lines must permit
flexure and must be of the proper size and pres- Gimbal Mount Assemblies
sure rating. Most of the hydraulic lines are made
of wire-braided flexible hoses and tubings. In most large turbopump-fed engine systems
(figs. 3-2, 3-4, and 9-i), the compIete engine
assembly is gimbaied at a universal-joint-type
gimbal mount assembly which connects the thrust
Turbine-Drive, High-Pressure, Hot-Gas Ducts chamber to the vehicle thrust structure.
In most turbopump-feed engines, the gas gen-
erator is connected directly to the turbine inlet;
Turbopump and Other Component Mounts
thus, there is no need for ducting. In some de-
signs, however, high-pressure hot-gas ducts are Turbopump mounts (fig. 9-1) are used to se-
required to connect between gas generator and cure turbopumps to thrust chamber or other struc-
turbines (in systems with two individual turbo- tural members, such as the thrust mount. They
pumps), or between main thrust chamber bleedoff often consist of welded tubular trusses, forged
ports and turbines (tapoff gas-turbine-drive sys- structural members, and flexible end joints. In
tems). These ducts usually consist of rigid and some designs, rigid-type, pump-discharge propel-
flexible sections made of stainless steels and lant ducts are used as members of the turbopump
nickel-based alloys, for service temperatures up mounts. Most propellant valves are fastened
to approximately 1700 ° F. The ducts must be directly to the thrust chamber or gas generator.
capable of absorbing considerable deflections Other control components and small storage ves-
from thermal expansion, in addition to deflec- sels may be secured by aluminum-alloy sheet-
tions due to misalinements and dynamic loads. metal brackets.
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
TABLE 9-1.-Corrosion-Resistant Steel (18-8) Annealed (MIL-T-8504 ASG) Tubing

[Allowable working pressures in psi at 100 ° F; safety factor of 4]
Maximum working pressure, 3000 psi
--1/2 5/8 3/4 1
Wall thickness, in ......................... 0.022

_0.035
• _°JXX_LL_:__I
0 042
• I 0.109
0 058
0 083 0 065
_e oD.in.............................. 1/4 _ .2-1---_-1---_"_---I

1-1/4
I 1-1J2
I
Wa.<.io_ne_s,
in......................... 10.020
1°.028
10.035
100_9|0.0_8
10.0_2
10.095
10.10._
1
Tube OD, in .............................. --

1,/4
0 020
318
0 028
1/2
0 032 0 035 0 0-t2
1
0 049
-1,t4
0.058
1-1f2
0.065
2
0.095
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]
Tube OD, in ..............................
Wall ,hickness. in ......................... i 0.025

1/4
[ 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
]
_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
SLEEVE SUPPORTS COUPLING

TUBE AND PREVENTS OR B-NUT
WIPING OF NUT
ON FLARE
SPECIAL
OUTER EDGE COUPLING
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.
TABLE 9-3.-Minimum Bend Radii [or Stainless-

Steel and Aluminum-Alloy Tubing
[All measurements in inches]
Inside bend Radii to center

Tube OD Wallthickness Figure 9-9.-Typical installations o[ $ube fitting
radii of tube
ends into AND10050 boss.
3/16 Any 5/8 23/32
1/4 Any 3/4 7/8
5/16 Any 3/4 29/32 The tube fittings should be of the same mate-
3/8 Through 0.022 1-1/2 1-11/16 rial as the tubing. Fittings are designed to be
Over 0.022 1 1-3/16 as strong as the strongest tubing of like material
1/2 Through 0.028 1-3/4 2 used with the fitting. Design details and dimen-
Over 0.028 1-1/2 1-3/4
sions of tube fittings can be found from AND and
5/8 Through 0.028 2-1/2 2-13/16
Over 0.028 I-3/4 2-1/16 MS specifications. In some cases, AN and MS
parts are interchangeable.
3/4 Through 0.028 3 3-3/8
The boss design for engine components, to
Over 0.028 2-1/2 2-7/8
7/8 Through 0.035 3-1/4 3-11/16 connect AN tube fittings, is specified by
Over 0.035 2-3/4 3-3/16 AND10050. Figure 9-9 illustrates typical in-
1 Through 0.035 3-1/2 4 stallation of tube fitting ends into the AND10050
Over 0.035 3 3-1/2 boss.
1-1/8 Through 0.035 4 4-9/16
Over 0.035 3-I/4 3-13/16
1-1/4 Through 0.035 4-1/2 5-1/8 9.3 DESIGN OF FLANGE JOINTS
Over 0.035 3-1/2 4-1/8
1-1/2 Through 0.035 6 6-3/4 The interconnecting lines of a rocket engine
Over 0.035 4 4-3/4
system must retain fluids at various pressures
1-3/4 Through 0.035 7 7-7/8 and temperatures, Each of these lines may have
Over 0.035 5 5-7/8 one or more mechanical joints requiring static
2 Through 0.035 7 8
seals to prevent leakage. The joint as an as-
Over 0.035 6 7
10-1/4 sembly must be capable of maintaining intimate
2-1/2 Through 0.049 9
Over 0.049 7 8-1/4 conformance of seal and sealing surfaces through-
out its operating life, regardless of all strains,
3 Through 0.049 11 12-1/2
Over 0.049 9 10-1/2 loads, and thermal gradients. For line sizes of
4 Through 0.065 12 14 3/4 inch or larger, bolted flange joints are gen-
Over 0.065 I0 12 erally used in rocket engine design. The design
of these flange joints is also applicable to other
components, such as thrust chamber injector
TABLE 9-4.-Recommended Support Bracket Spac- dome flanges and turbopump and valve housing
ing [or Tubing Assemblies in Engine Systems flanges.
Maximum support spacing in inches

Flange Joint Design Considerations
Tube size(OD), in
Aluminum ahoy Stainlesssteel
Figure 9-10 illustrates typical flange joints
12 14 used in a propellant duct assembly. Most flange

1/4 through 1/2 ....
5/8 through 3/4 .... 17 20 joints consist of three elements: the flanges,
1 and over ......... 21 24 the fasteners (bolts and nuts), and the gasket.
In addition to an effective gasket, a leaktight
flange joint must have proper flanges and bolt-
special hardened sleeve, which cuts into the ing. In flange joint structural design, the goal
tube wall. Flareless tube fittings may be used of optimum weight within safe stresses must
optionally in engine designs. consider the elastic behavior of the part.
a material so that when properly loaded, a barrier

'A/ASHER BOLT
will be formed in depressed surface regions, pre-

UI _BOLr
venting leakage of the contained fluid. Gaskets
which damage the flange sealing surfaces are
usually unsatisfactory, if the joint has to be
repeatedly disassembled and reassembled. For
these applications, gasket surfaces should
always be softer than those of the flanges.
4. Sealing surface condition.-As a rule, leak-
age will occur when the internal fluid pressure
begins tO exceed the compressive stress which
\r_ \ _-[. x.x_) _ S ,_LLO,.._ $ A$$,MB LY /
holds the gasket in contact with the sealing

surface. Radial scratches on the flange sealing
surface tend to induce a lower gasket local com-
pressive stress along that radial path, resulting
in leakage. Conversely, if the flange sealing
surface consists of concentric serrations, higher
_WASHER
local compressive stress will be induced circum-
Figure 9-I O.-Typical flange joints of a propellant ferentially along the serrations, and will seal
duct assembly. higher internal pressures for a given nominal

compressive stress of the gasket. The required
sealing-surface finish is a function of gasket
design. Generally, it ranges from a mieroinch
In addition to structural needs, other design
finish of 32 to 125 rms, with concentric tool
considerations for flange joints are:
1. Working temperature and pressure of the markings.
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).
should neither loosen nor tighten the joint. In

actual design the flange joint must be sufficiently
Flange Joint Structural Design
elastic, so that sealing loads are at least main-
tained within reasonable limits. Figure 9-11 presents the structural design
2. Degree of sealing.-For most applications configuration of a typical flange joint as fie-
it is sufficient to prevent mass flow as evidenced quently used in rocket engines. The flange ring,
by liquid or gas tightness under soap bubble or under various working loads, is subject to bend-
immersion tests. When sealing the lighter gases ing as shown exaggeratedly by the dotted lines.
(H 2 and He), diffusive flow through openings of The bending moments may become quite large;
molecular size may become significant. the resulting stresses reach their maximum at
3. Gaskets.-The function of a gasket is to corner Z, where the flange joins the wall of the
seal effectively between two flange surfaces. It duct.
is usually made of a material which will readily The basic approach to the design of a flange
conform to the surfaces, or which is coated with joint is to prestress the flange bolts in tension
_II I II I I _
D s =outside diameter of the gasket, in

1
p =internal fluid pressure, psi (the mammum

J %
working pressure should be used)
F, =force per unit length of the flange rag from
bolt loading, lb/in
F 2 = force per unit length of the flange rag from
, ;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)
2._ We The following design data are given for the

F2: \-7-5-] + ;-5 (9-2)
flange joint of the oxidizer-pump-discharge flex-
ible duct of the A-1 stage engine:
p(D22 - Dt 2)
Working pressure under normal steady oper-
F3 - 4 D (9-3)
ating conditions, 1505 psia
Maximum fluid working pressure under occa-
Sg(D32 - D22) (9-4) sional transient conditions, 1750 psia

F4 - 4 D
Inside diameter of the duet, D 1 =8 in
Inside diameter of the gasket, O 2 = 8 in
Sg =mp (9-5) Outside diameter of the gasket, D 3 = 8.5 in
End loads on the duct, due to thermal con-
F s = nF1 (9-6) traction, W e = 2400 lb
Gasket factor, m = 0.8
W b = FtzrD (9-7) Flange factor, n = 0.3
Determine the minimum required design bolt
where (referring to fig. 9-11)
loading of the flange joint.
D =diameter of the circle passing through the
eentroid of the flange ring cross section,
Solution
in
D_ =inside diameter of the flange and duct, in We will use a maximum working pressure
D2 = inside diameter of the gasket, in p= 1750 psia
From equation (9-5), the required average A = design factor, ranging from 4 to 6
gasket compressive stress: B =design factor, ranging from 10 to 14

A taper angle of 25 ° is generally used for the
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
Fi - (1 - n) distributed along a circle passing through the

eentroid of the ring cross section. The following
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
L1 = overall axial length of the flange ring, in ring, in-lb/in

ly_y = the moment of inertiaof the Yield pres sure × D 2117×8

t=
ring cross sectionabout 2 x 170 000 x 0.75 = 0.066 in
2xFy×ew
the Y-Y axis, in4
E = modulus of elasticityof the From equation (2-10), ultimate pressure = 1.5
flangematerial,psi x 1925 = 2887 psi.
0 = angular displacement of the Based on ultimate strength, the thickness of
flange ringunder maximum the duct wall
working pressureand
loads, rad Ultimate pressure x D: _ 2887 x 8
S x = flange ringmaximum tensile t=
2 × Fu × ew 2 × 200 000 x 0.75
stress, which occurs at
the corner Z and is normal = 0.077 in
to the plane of figure 9-11, We will use the higher value and round it to
i.e., in the hoop direction t= 0.08 inch, as the selected wall thickness of
the duct.
Sample Calculation (9-2) Generally, based on given bolt size, gasket
diameters, and duct dimensions, a proposed
With the data given in sample calculation
flange configuration may be derived simply from
(9-1), design the flange for the oxidizer pump
good design layout practice. Then the proposed
discharge flexibleduct of the A-I stage engine,
design is checked for working stresses and
with the followingrecommended materialand
strains. The flange thickness T may be adjusted
boltdata:
accordingly. Assume flange design factors A = 6
Flange and duct material,Inco 718 and B = 12.2.
Minimum yieldstrength,Fy = 170 000 psi From equation (9-9). the flange thickness:
Minimum ultimatestrength,F u = 200 000 psi
Modulus of elasticity,E = 29.6× 106 psi T=At= 6 x 0.08=0.48 in
Duct weld efficiency,ew = 0.75
Bolt size and material,5/16-24 A-286 (200000
From equation (9-10), the flange axial length:
psi),stainless-steelbolt,0.526 head dia
Allowable ultimateboltload, 10262 lb
L_ =Bt=12.2xO.08=0.976 in, say 0.98
Solution Also, from a design layout, we obtained the

Use the flange design and nomenclature shown following additional values:
in figure 9-11. From equation (2-8):
D4 = 9 in
Design limit pressure Ds =9.62 in
= 1.2 × steady operating pressure = 1.2 x 1505 L = 0.37 in (approximately)

=1806 psi D = 8.6 in (approximately)
or ly_y = 0.0273 in 4 (approximately)
Design limitpressure From equation (9-7):
= i.i× maximum transientpressure= I.I× 1750 Wb 140400

= 1925 psi F, =_-_= 8.-------ff
r,× "= 5200 lb/in
From equation (2-9), yield pressure= 1.1 From equation (9-2):

× 1925=2117 psi.
In accordance with standard practice, the fpo,2xwo(1750×82
/ 2400
thickness of the duct wall is determined as F 2 = _--_D-]+,_= _ -_x-83g _/+--_-___.6 = 3347 lb/in
DEMON OF INTERCONNECTING COMPONENTS AND MOUNTS 365
From equation (9-3): stress under maximum transient pressure con-

ditions:
p(D 22 - D L2)
F3 - 4D -0 MtD2(L1 - L) _ 1174 x (8.6) 2 × (0.98 - 0.37)
Sz- 2Dlly_y 2 ×8x 0.0273
From equation (9-,1):
= 121 200 psi
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
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
Ps=2d+ T=2×0.312 +0.48= l.104 in

The required number of bolts:
(D - D_ - t) _ (8.6 - 8- 0.08) = 0.26 in

Y2- 2 2 r;D4 n"× 9
P--T- 1.104=256' say 26
The required bolt loading W b calculated in
(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:
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
Sheet Gaskets for Flange Joints

From equation (9-17), the angular displace-
ment of the flange under maximum transient pres- Gasket materials consisting of asbestos fibers
sure condition: bonded with rubber or neoprene are readily avail-
able in flat sheets, from which tim gasket may
be cut to suit most requirements. Frequently
MtD 2 1174 x(8.6) 2
8=
used sheet gaskets range from 1/32 to 1/8 inch
4 Ely_y 4 × 29.6 x 106 ×0.0273
in thickness. At fluid pressures less than 50
:0.0268 rad or 1.54 ° psi, sheet gaskets can be used with flat faces
at both flanges. For higher pressure applica-
From equation (9-18), the maximum working tions, sheet gaskets can be secured in a flat,
concentric groove machined on one of the flanges,

with the groove depth sized for proper preloading
of the gasket. Preferably, the flange facing
should be serrated concentrically. In rocket
_SEAL SURFACE FINISH fl'
engine applications, sheet gaskets should be IT $1"(OULD EE FREE OE
NICKS, SCR,ATCHES AND
limited to pressure levels of less than 200 psi, OTHER Ih_ERFECTION$
EXCEPT ANNUblk:_ TOOL
and to temperatures in the range from -100 ° F to MARKS
900 ° F.
Elastomer and Metal O-Rings for Flange Joints BRIAR COR_ERS 005"
_--i
D3 --
D2
_ G_OOVE SIDE FINISH f2
Elastomer O-rings have been used with good

results for rocket engine flange joints in medium
temperature service (-60 ° to 400 ° F). Satisfac-
tory results have also been experienced at pres-
sure levels as high as 3000 psi. Typical elas-
NIC*:S, SC_,TC._ES J,t'4D
tomer O-ring flange joint design data are OT_E_ IMPER_E£TIO_S
EXCEPT Ar_,'_UL_R TOOL
presented in table 9-5 for use in groove-type

flanges, as shown in figure 9-12. Table 7-3 may
Figure 9-I 2.-Flange seal groove design.
be used to select the hardness of the O-ring
materials with respect to working pressures.
Elastomer O-ring seals for various flange size
may easily be made by cutting straight O-ring sn_t._ Metal O-rings can be used in a groove-
stock to a specific length, and cementing the type flange, as shown in figure 9-12. Table 9-6
ends together. presents typical design data for metal (3-ring
For low- and high-temperature services (-430 ° flange joints,
to 2000 ° F), and for sealing against very high
pressures (up to 4000 psi), metal O-rings may be
used in rocket engine flange joints, with proper
Spiral-Wound Gaskets for Flange Joints
detail design. Hollow stainless-steel O-rings,
such as manufactured under the trade name Spiral-wound gaskets manufactured by the
"Toruseal" by the D.S.D. Co., have been used Flexitallic Gasket Co. have been used exten-
extensively in various services. They can be sively in rocket engine flange joints for liquid
coated with Kel-F and Teflon for cryogenic serv- oxygen and hot gas (up to 1600 ° F) services.
ice. One of the advantages of metal O-rings is Pressures are generally kept below 1000 psi.
that they require considerably less space than The gaskets (detail shown in fig. 9-13) are made
other seal types. However, the O-rings must of spiral-wound 304 stainless-steel ribbon, with
have a good seal surface finish, and high com- a Canadian asbestos filler. It is not recom-
pressive seal loads, to assure satisfactory re- mended that gasket widths of less than 3/16 inch
TABLE 9-5.-Typical Elastomer O-Ring Flange Joint Design Data a

[All dimensions in inches]
Nominal Actual Bottom

Minimum Depth of groove, Seal surface
O-ring O-ring Width of groove, radius, Seal surface
squeeze a b finish, fl finish, f2
crosssection cross section i"
1/16 0.070 *_0.003 0.017 0.045_ 0.005 3/32 1/64

3/'33 •103 * .003 .020 075± .005 9/64 1/64
1/8 •139 -* .004 .025 •1052 .005 3/16 1/32 64 rms 64 rms
3/16 .210 t .005 .030 .170± .005 9/32 1/32
1/4 .275 _* .006 .039 .225± .005 3/8 1/16
TABLE 9-6.-Typical Metal O-Ring Flange Joint Design Data a

Actual O-ring size Flange groove dimensions
Nominal O-ring, OD rube OD [ Tube OD I

Wall D2 [ Ds + 0.005 a+0.005
+0.005 I +0.003
thickness - 0 [l
-0 maximum ] - 0
,
1 .................................. 1.000 ] 0.061 0.009 0.867 I 1.005 0.042

1-1/2 .............................. 1.500 ] .061 .009 1.367 f 1.505 .042
2 .................................. 2.000 ] .061 .009 1.867 ] 2.005 .042
2-1/2 .............................. 2.500 ] .061 .011 2.367 ] 2.505 .042
3 .................................. 3.000 I .061 .011 2.867 ] 3.005 .042
4 .................................. 4.000 I .094 .011 3.801 [ 4.005 .075
5 .................................. 5.000 ] .094 .011 4.801 I 5.005 .075 32 rms
6 .................................. 5.000 I .094 .011 5.801 i 6.005 .075
8 .................................. 8.000 1 .124 .011 7.741 ] 8.005 .105
10 .................................. 10.000 1 .124 .011 9.741 I 10.005 .105
15 .................................. 15.000 ! .124 .011 14.741 I 15.005 ,i05
20 .................................. 20.000 ] .124 .011 19.741 I 20.005 .105
30 .................................. 29.741 I 30.005 .105
30.000 1 .124 .011
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
TABLE 9-7.-Typical Spiral-Wound Gasket Flange Joint Design Data a

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 ]
1.000 1.375 0.442 82O0 10 600 0.9001 1.425 I

e_ e,
1.500 1.875 .638 8300 11 500 1.400 1.925 I
2.000 2.375 .835 9000 12 700 1.900 i 2.425 l
2.500 2875 1.031 9800 14 100 2.400 2.925 ]
3.000 3.375 1.227 I0 800 15600 2.900 3.425 ]
4.000 4.375 1.620 12950 18 600 3.900 4.425 I
5.000 5.500 3.056 23000 32 400 4.875 [ 5.560 ] 0.010 125 rms
6.000 6.500 3.645 26400 37 000 5.875 6.560 } +0
8.000 8500 4.832 33 600 43 000 7.875
8.560[ _
Ii .000 11.750 11.110 73 600 97 000 10.875 11.810 I c_
c5 c_ c5 c_
15.000 15.750 15.040 98 600 122 500 14.860 15.820 I + , -'-
20.125 21.000 24.150 153000 178500 19.985 _i.005
25.000 25.875 29.900 186 500 206000 24,840 25.955 j _ d
aSee fig. 9-13 for seal design. See fig. 9-12 for flange design. Duct dlameter D, may be equal to D2.
._PIRAL WOUND EJSd

r STAINLESS STEEL RIBBON
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-
_¢ .... C-C ference with internal fluid pressure,

lb/in
E =modulus of elasticity of the sealmaterial,
Figure 9-I3.-Spiral-wound gasket for llange
psi
joints.
J =end thickness of the seal leg, in
H =length of the seal leg, in
d = seal deflection in the flange joint, in
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)
of erratic compression loads, due to thickness p =internal fluid pressure, psi
tolerances and variations in wrapping. Another

is their tendency to indent the flange sealing
surfaces, particularly when the flanges are of
M.*,TERIAL ¢340 STEEL H'g#.T TREATED TO _SO 0'_ - 18OOOO
aluminum. FINISH - _L_CTROLI;SS NICRFA. PLATE. 00_0 - 0012 INCH £NBRITTLE.M_;_"i" _ELIL',rED
TEFLON COAT Co1 - 0,_ INCH
Pressure-Actuated Seals for Flange Joints
seal
An ingenious
deflection
approach
capabilities
toward
to accommodate
attaining better
de-
_
,,
K : 7 ' -oo3
,)i oo._ooo.
'' -:oz
formations of the flange joint under working

loads is the pressure-actuated type seal, shown
....... - ....... J. I,.L_o? sEc'r_. A -A
in figure 9-14. The spring effect is achieved by
a metal, circular seal having a U-shaped cross THESE SURFACES TO BE FRE:E OF NICKS B_RS _ND TOOL MARKS EXCEPT ANNULAR TOOL
_tARCS
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.
/_J THICKNE&S OF TEFLON COATING ON THESE SURFACES OPTIONAL

points at its open ends make contact with the
flange surfaces. If properly designed, the spring-
like legs flex to follow flange deflections.
The open end of the seal must always be
oriented toward the inside of the flange joint.
Thus the internal fluid pressure is utilized to
generate a considerable portion of the seal con-
tact load. Pressure-actuated seals are installed
in a concentric groove cut into the flange joint.
This limits the amount of total flange preload
OPEN GRO0_I_ CLO_EI) GROOVE
transmitted through the contact points of the L_T #,LLA TION INSTALLS'lION
seal. The following correlations are established

for the seal design shown in figure 9-14 for the Figure 9-I 4.-R o c k e td y n e Naflex pressure-
approximation of seal contact loads: actuated seal design [or [lange joints.
C I =design factor. This is a function of seal tc :thickness of tllecoating, in

leg configuration, its value ranging from Many versions of pressure-actuated seals
0.04 to 0.8 have been applied to rocket engines. Rocketdyne
Generally, the contact surfaces of pressure- has developed the Naflex-type pressure-actuated
actuated seals are coated with a material of a seals, made of alloy steel and coated with a thin
lower elastic modulus than that of the material layer of Teflon. They are most effective for
from which the seal is made. Thus the seal cryogenic (-430 ° F) and high-pressure (up to
material can be utilized to achieve the spring- 4000 psi) services. Figure 4-32 illustrates a
like p,'opertiesrequired for maximum deflection typical Naflex seal installation at tileflange
capability, while the coating material provides joint between thrust chamber and liquid oxygen
the properties needed to establish tileintimate dome. Design details and data for typicaINaflex
contact required to prevent leakage between seal flange joints may be seen in figure 9-14. Also
and flange surfaces. To prevent leakage, suffi- see table 9-S. For hot gas applications, up to
cient compressive stress must be developed in approximately 1700 ° F, the seals may be coated
the coating to make it flow into imperfections on with soft metals such as copper, aluminum, or
the flange surfaces. The required coating com- _old.
pressive stress may be determined by For certain stringent applications, all propel-
lant and turbine gas flanges in a rocket engine
Eoh can be equipped with a Naflex-type dual static
S c : -- (9-2i)
tc seal incorporating an intermediate cavity be-
tween high-pressure and low-pressure sealing
where areas. A seal drain line may be connected to the
Sc = seal coating compressive stress, psi intermediate cavity, as shown in figure 9-15.
Ec = modulus of elasticity of the coating mate- This line can also be used to measure the seal
rial, psi leakage rate, ifany. This arrangement may also
h =depth of the imperfections on the flange be applied to pnenmatic system flange joints, to
sealing surfaces, in monitor leakage.
TABLE 9-8.-TypicaI Design Data o[ NaHex Pressure-Actuated Seals and Flange Joinzs
Seal dimensions I Flange groove dimensions

J
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.
From figure 9-14, the seal land of contact

=0.019. Thus, the seal coating compressive
stress without imernal pressure
582
Sc = 1 x 0.019 - 30600 psi
From equation (9-20), the seal contact load

with internal pressure
2000 (0.205- 0.02)_ 767 lb/in

Fs2 =582+ 2 -
"--- NAFLEX TYPE DUAL STATIC SEAL

'-- SEAL DRAIN LINE
The seal coating compressive stress with an
internal pressure of 2000 psi
G AREA
767
Sc = 1 × 0.01---_
= 40 400 psi
_ The allowable
flange sealing
depth of imperfections
surfaces should be calculated
on the
from the initial coating compressive stress,

Figure 9-]5.-Typical design of a flange joint using equation (9-21):
with Naflex-type dual static seal and seal
drain line.
h- Sctc 30600x0.002=0.0012 in
Ec 5×104
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.
.__.- -" .... _-- _:_,S,

__-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.
allow fortemperaturechanges, dynamic loads,

and engine gimbal effects,ifany. This often
determines the linegeometry. In general,how-
ever, lines should be as shortand straightas
possible.
5. Structural integrity.-The structural design
of a duct system may involve a number of com-
plex problems. This is especially true with
ducts for elevated or very low temperatures, and
requiring a high degree of flexibility. The de-
sign and fabrication of metal bellows for applica-
tion in ducts is a highly specialized technology.
Duct Design for Minimum Pressure Drop
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.
This consideration should be applied in all

cases where a sharp turning corner is unavoid- SMOOTHED
DIAGONAL
able, such as the elbow-type turbopump inlet WALL
JUNCTION
ducts shown in figure 6-15. Here, the loss
around the sharp corner is due largely to the
zero radius ratio. To reduce this pressure drop,
guide vanes can be very effectively used. If the
junctions of the duet walls on the diagonal are
smoothed out and vanes added as shown in figure
9-18, the compartments so formed may be madeto
have a good radius ratio and a high aspect ratio.
r * 1¼
Since in most cases only a few vanes need to be
added for high aspect ratios, the remaining prob-
Figure 9-18.-Elbow-type sharp turning duct with
lem is one of securing a good radius ratio.
guide vanes.
The radius and aspect ratios of each compart-
ment are increased as the gap/chord ratio S/C is
reduced (fig.9-18). Theoretically, then, the I
corner pressure drop will decrease as tileratio

S/C is decreased. However, each additional
THRUST CHAMBER/-_p_i -, _Q_
vane adds more surface area and blockage, and
cop sequently tends to increase losses. The S/C
ratio can be optimized in tests and ranges from ---÷--- I _-
0.25 to 0.45. Well-designed elbow-type sharp [l + ") x'x i , 1 ', ] UNIVERSAL
turns with guide vanes may have resistance

coefficients K as low as 0.i to 0.2 (see eq. 7-9).
Duct Design for Flexibility

. _
For most interconnecting ducts between two

BEILvL2226! .Inlet ]INLET CONNECTED
engine components, the required flexibilitysuch
................. _ TO VEHICLE DUCT
as for misalinements and temperature changes
can be achieved by placing one or two bellows
sections in the line (figs. 9-4 and 9-5). These
may be either restrained or unrestrained, depend-
ing on structural loading considerations. In
ducts connecting two components involving large
relative movements, a minimum of three flexible
OUTLET CONN CTED _--r-,k}..__._ CHAMBER
sections is required. The longitudinal axis of at
least two of the bellows sections should be posi-
tioned at a right angle toone another. Restrained
ENGINE GIMBAL
bellows are preferred.
The duct systems shown in figure 9-3 illus-
Figure 9-19.-Typical propellant supply duct de-
trate a basic three-bellows arrangement as ap-
signed for the ilexibility required ior engine
plied to gimbaled clustered engines. Figure 9-19
gimbaling.
presents one of the frequently used design
approaches for a relatively short duct system to
facilitate large movements during gimbaling of iently incorporated into this duct system by
an engine system. The three flexible sections welding. These joints are made of stainless
are kept in the plane of the engine gimbal point. steel and will operate at cryogenic and elevated
This results in minimum displacements of the temperatures (up to 1000 ° F). The internal re-
sections for a given engine movement. Bellows- straining mechanism is designed to give a smooth
type universal joints (fig.9-20) may be convert- flow passage for low-pressure drop.
ally expanding thin-wall tubing, at spaced inter-

vals along its axis, coincident with an axial
--PLAIN END FOR WELDED compression, to a configuration as shown in fig-
INSTALLATION
ure 9-21. Our discussion will be confined to
this type of bellows. The various structural
FORK ASSEMBLY characteristics and design correlations of bel-
lows are presented using the following nomen-
clature (see figs. 9-21 and 9-22):
--.._ GIMB/kL RING
_ ---'_"--_ ASSEMBLY Ct :bellows wall-thinning correction factor

RIVET Cp = ply interreaction factor (1.00 for 1-ply bel-
lows, 0.90 for 2-ply bellows, and 0.85 for
3-ply-or-more bellows)
d o : outside diameter of the bellows, in
di = outside diameter of the convolution root of
the bellows, in
d m = V'(di 2 + do2)/2 = root-mean-square diameter
of the bellows, in
d d =mean duct diameter, in
E --modulus of elasticity of the bellows mate-
rial, psi
e a = axial deflection of the bellows, in
eb = equivalent axial deflection of the bellows
,?+o + due to pure bending, in

ep = equivalent axial deflection
due to parallel offset, in
of the bellows
e s = equivalent axial deflection of the bellows

due to pure shear, in
F s = shear load, lb
Fp = pressure separating load, lb
G = shear modulus of elasticity of the bellows
material, psi
b : (d o - di)/2 : mean convolution height, in
L --pitch of the bellows (axial length of a con-
volution), in
Figure 9-20.-Typical gimbal-ring-type bellows
La = (Nc - 0.5)L + Npt = free axial length of the
universal joint designed and manufactured by
bellows, in
Marman Division of Aeroquip Corp.
Nc = number of bellows convolutions
Np =number of bellows plys
For higher pressure applications (over 200 t = thickness of the bellows wall, in
psi), external wire braids or links may be used Lb = axial length of the rigid duct section, in
to restrain the bellows section. The braids can M =bending moment, in-lb
be made of stranded stainless-steel wires, about p =internal (or external) fluid pressure, psi
0.012-inch diameter. In extreme cases, more Pcr = critical stability pressure of the bellows,
than one layer of braid should be used with re- psi
gard to the high separation loads. R a = axial spring rate of the bellows, lb/in
R b = bending spring rate of the bellows, lb-in/
degree
Structural Design of Bellows For Flexible Ducts
Rp : parallel offset spring rate of the bellows,
Most of the bellows used for flexible ducts in lb/in
rocket engines are hydraulically formed by radi- Rs -- shear spring rate of the bellows, lb/in
,XIAL
"qF--- OEFLECTION
(a)8EL.LO'_$ UNDER PURE

SENDING
{hi _ELLOW$ UNO£R L

PURE SHEAR
i
(¢} BELLOWS UNDER

PARALLEL OFFSET
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
Rt = torsional spring rate of the bellows, lb-in/ PERCENT THINNING
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)
3. Bulging stresses.-These are the radial 1.78 CtEtea

bending stresses induced in the side walls of the Sm : Nch2 , psi
bellows, by internal or external pressure. Bulg-
(aluminum alloys) (9-28)
ing stresses should be kept below those listed
in table 9-9:
All other types of bellows motions, such as
angular and parallel offset, can be converted into
S Ctph 2
an equivalent bellows axial deflection, and ap-
b : 2--'ffp-_'pt2.
psi (9-24)
plied to equations (9-27) and (9-28) to calculate
the corresponding motion stresses.
4. Separating loads.-Bellows are pressure
7. Bellows under pure bending (fig. 9-21a):
loaded so that they experience a separatingload,
in addition to the normal duct axial force: M dm2Ra
Rb =_- 458.4 ' lb-in/degree (9-29)
P(dm 2 - dd2)_
Fp- 4 , lb (9-25)
dm sin 0
, in (9-30)
eb- 2
5. Hoop stresses.-The bellows hoop stresses
are calculated by using the total area per inch of Motion stresses due to pure bending can be cal-
bellows axial length of a tube of equivalent wall culated by substituting eb for ea in equations
thickness. Hoop stresses should always be kept (9-27) and (9-28).
lower than the yield and nltimate strength of the 8. Bellows under pure shear (fig. 9-21b):
bellows materials by a specified margin.
Fs 3dmSRa
Rs : lb/in (9-31)
(9-26) y 8 La 2 '
Sh : 2--'_p_
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:
Motion stresses due to pure shear can be calcu-

1.40 CtEtea
lated by substituting es for ea in equations
Sm: Nch2 , psi
(9-27) and (9-28).
(steel and nickel alloys) (9-27) 9. Bellows under parallel oHset (fig. 9-21c):
TABLE 9-9.-Yield Strength, Limiting Bulging Stresses, and Allowable Motion Stresses of Frequently
Used Bellows Materials
Limiting bulging Allowable motion stresses,

Yield stresses,psi psi
Material strength,
psi 1000 10000 100000
t_0.012in t_0.013 in
cycles cycles cycles
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
Fs 3dm2Ra 13. Bellows buckling due to external pres-

Rp - y - 2 La 2 ' Ib/in (9-34) sure.-When a bellows is pressurized externally,
it buckles in same manner as a thin cylinder:
FsLa 3 dm2YRa
M =_--=-+ + in-lb (9-35) 4 EtNph 3
2 4 La 2 '
Per-(1 _ v2)dmSPb, psi (9-43)
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.
Fs(2 La + Lb) 15. Allowable working stresses for bellows

M--±
2 materials.-Bellows generally operate in the
plastic range. However, it has been the practice
3dm2(2 La + Lb)yRa
in-lb (9-38) to calculate stresses on the basis of elastic
= 4(4 La 2 +6 LaLb+3 Lb2) '
deformation, and correlate the data on that basis.
In reality, the calculated stresses are only index
3 din(2La + Lb)Y
(9-39) stresses, which define the plastic conditions of
ep_(4 L 2 + 6 LaLb + 3 Lb 2),in the bellows, under the influence of pressure and
motion loading. The yield strength, limiting
ep can be substitutedforea in equations(9-27) bulging stresses, and allowable stresses of fre-
and (9-28)to calculatemotion stresses. quently used bellows materials are presented in
table 9-9.
Ii. Bellows torsion.-The stress due to bel-
lows torsion is given by the twist of a thin tube: 16. Bellows used at elevated temperatures.-
For applications at elevated temperatures allow-
2T
able working stresses for bellows materials must
St- _"prrdi2 t , psi (9-40) be adjusted accordingly. Generally, an internal
liner should be provided to protect the bellows
T 1.37 × i0 -2 GditNp, (9-41) against high-velocity, hot-gas flows.
Rt lb-in/degree
(_ (2 h +0.57 L)Nc
17. Bellows forming lin_,its.-Bellows designs
12. Bellows squirm due to internal pressure.- are limited by how severely the material can be
When a restrained bellows is pressurized inter- worked during forming. Generally, the following
nally beyond a critical level, it experiences a geometric limits should be used for bellows up
stability failure of the same type as a buckling to three plys:
column:
do
5.02 Ra Maximum: _ = 1.35 (9-45)

Per= d/, \' psi (9-42)
La
Minimum: L = (8 + 2 Np)t, in (9-46)
Design a bellows, as shown in figure 9-4, for din= _/(di2+ 2 d°2) - _ (82+ _ '822)
-- - 8.42 in
the oxidizer pump discharge flexible duct of the

A-1 stage engine, with the following data and From equation (9-30), the equivalent axial
requirements, in addition to those given in deflection of a bellows, due to pure bending or
sample calculations (9-1) and (9-2). angulation:
Bellows material, Inco 718
Thinning of bellows wall, 20 percent tim sin 0_ 8.42× sin 3°
Outside diameter of the convolution root of eb = 2 2 = 0.22 in
the bellows, di= 8 in
Free axial length of the bellows, La = 7 in Table 9-9 lists, for Inco 718 and a life
maximum
of 10 000 cycles, an allowable motion stress of
Required angular movement, 0= ±3 ° 150000 psi. It is good practice, however, to use
Life, 10 000 cycles a lower value for high-pressure bellows designed
In addition, determine the following:
for improved stability. In our design, we use a
Bellows axial spring rate, Ra value of 0.36 times the limiting bulging stress.
Bending or angular spring rate of the bellows, Thus, the motion stress = 0.36 × 150 000 = 54 000
Rb psi. Substitute this and eb into equation (9-27):
Bending moments of the duct at 3 ° angular
motion, M I AO CtEteb
Required restraining link load at maximum
Sm- Nch2
working pressure
Solution 1.40 CtEteb

(from fig. 9-22, Ct=l for 20 per-
Nc - Smh 2
We will first use the limiting bulging stress cent thinning)
of the bellows material to establish the ratio of
convolution height h to wall thickness t. Prom _ 1.40× 1× 29.6× 106× 0.022x 0.22 = 22.1,
- 54 000 × (0.410) 2 say 22
figure 9-22, wall-thinning correction factor
Ct= 1.36, for 20 percent thinning. From table
9-9, the limiting bulging stress for Inco 718 at From equation (9-46), the pitch of the bellows
t>0.013 inch is 150000 psi. From sample cal-

culation (9-2) the design limit pressure of the L=(8+2 Np)t=(8+2× 3)×0.022=0.308, say 0.310in
duct is 1925 psi. Substitute all these into equa-

tion (9-24) for a three-ply bellows: Free axial length of the bellows:
La = (Nc - 0.5)L + Npt
= (22 - 0.5) × 0.310+ 3 × 0.022
=6.66+0.066=6.726 in (i.e., lessthan 7in)

h 2N/_-_ 4/2×3x150000 18.5
-CTT :Ip 1.36x1925 From figure 9-22, the correction factor C t for
the axial spring rate at 20 percent thinning is
For a reasonable value of h, in a bellows of this
0.72. Substitute this into equation (9-22) to
size, we arrive at a wall thickness t of 0.022 in
obtain the axial spring rate of the bellows:
(after several tries):
h=18.5×0.022=0.407 in, say 0.410 in 1.49 CtC pN pEdit 3

Ra= Ncha
The bellows root-mean-square diameter and
the equivalent axial deflection due to 3 ° angula- _ 1.49 × 0.72 × 0.85 × 3 × 29.6 × 106 × 8 × (0.022) 3
tion can now be found: 22 × (0.410) 3
do=di+2h=8+2×0.410=8.820 in = 4550 lb/in

From equation (9-42), the critical internal p(dm 2 - dd2)r_

stability pressure for bellows without angulation: Fp= 4
5.02×Ra 502×4550 The required restraining link load at maximum

Per= do working pressure, considering the normal axial
La 6.726 force
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
Primary design considerations for gimbal

mounts are:
Sh =-2 _pt I. Required pivotal movement o[ engine as-
sembly or thrust chamber for thrust vector con-
_ 22117
x 3 xx0.022
8.42 × I0.57 + (2
k_ x 0"41)1J trol.-Geuerally ranging from ±4 ° to +__10 °.
2. Required adjustment for thrust alinement
= 42 000 psi (less than 170000 psi)
and positioning.
3. Thrust level.-This determines the struc-
Tile ultimate hoop stress of the bellows: tural and bearing design of the gimbal mounts.
4. Required operational li[e.-Generally 1000
42 000×2887 =57 300 psi (less than 200 000 psi) cycles minimum.
Sh = 2117 5. Minimum deformation of the bearing sur-
face, prevention of bearing surface galling.
6. Propellant duct installation.-Some designs
The bellows design configuration is now es- require flow of one of the propellants through the
tablished: center of the gimbal mount.
7. Maintenance o[ the gimbal rnount.-This is
di =8 in d o =8.82 in dm =8.42 in largely affected by the lubrication requirement of
t =0.022 in Np=3 h =0.41 in the bearing surfaces. If possible, a gimbal boot
Nc=22 L =0.31 in La=6.726in should be provided to protect the bearing sur-
R a = 4550 lb/in faces from dust, water, and foreign materials.
8. Lightweight.-High strength-to-weight
From equation (9-29), the angular spring rate materials should be used.
of the bellows
Ring-Type Gimbal Mounts

dm2Ra _ (8.42) _ x 4550_ 705 in-lb/degree
Rb = 458.4 458.4 Figure 9-23 presents a typical gimbal mount
designed for low-thrust, upper stage engine ap-
The bending moment on the duct at 3 ° angula- plications up to 20 000 pounds. The design is a
tion closed-yoke, flowthrough-ring-type gimbal, utiliz-
ing plain bearing pivot joints. This configura-
M = RbO = 705 × 3 = 2115 in-lb tion provides for the main oxidizer duct to pass
through the assembly in the longitudinal axis to
From equation (9-25), the bellows pressure the thrust chamber dome. The gimbal mount is
designed to connect the vehicle main oxidizer
separating load
.'
THRUST CHAMBER
Figure 9-24.-Typical cross-type gimbaI mount

designed for medium-thrust engine applications.
incorporating bearing surfaces, upper and lower

gimbal bearing blocks, upper and lower retainers,
and thrust vector alining slides. All parts of
this design are made of 4340 alloy steel. The
bearing surfaces of the cross unit are chromium
or electroless nickel plated. The bearing sur-
Figure 9-23.-Typical ring-type gimbal mount de-
races are phosphate treated. Solid-film dry-
signed [or low-thrust, upper stage engine
lubricant coatings are applied to all bearing
applications.
surfaces. The design bearing pressure of this
type gimbal mount ranges from 15 000 to 20 000
duct and the top of the thrust chamber dome. The psi of bearing projected area. The bearing coef-
lower support of the gimbal mount provides an ficient of friction varies between 0.06 and 0.1.
adjustment mechanism for thrust vector aline-
ment. Except for the steel pivot pins and the
alinement bolts, all other parts can be made of
either aluminum or titanium alloys for minimum
weight. The bearing surfaces of the pivot pins
should be chromium or electroless nickel plated.
Solid film dry lubricant coatings or grease may
be applied to all bearing surfaces. The design
bearing pressure of this gimbal type is about
10000 psi of bearing projected area. The bear-
ing coefficient of friction is around 0.08.
Cross-Type Gimbal Mounts
Figure 9-24 presents the design of a typical

cross-type gimbal mount used on medium thrust
engines (up to 200000 Ib), such as shown in fig-
ure 9-I. Here, the gimbal mount secures the
engine assembly to the vehicle thrust frame, and
is mounted to thrust chamber dome and elbow Figure 9-25.-Typical socket-type gimbal mount
assembly. It consists of a cross-shaped unit designed tot high-thrust engine applications.
Socket-Type Gimbal Mounts position coating, which provides a dry, low-

Figure 9-25 presents a typical heavy-duty friction bearing surface.
gimbal mount designed for high-thrust engine The bearing surface on the sphere is electro-
applications (up to several million pounds). This less nickel plated. No lubrication or maintenance
mount is also designed to secure the engine is required. For these gimbal mounts, design
assembly to the vehicle thrust structure. The bearing pressures range from 24 000 to 2S 000
unit consists of integral upper and lower bearing pounds per square inch of bearing projected area.
blocks, and their retaining bar. All parts are The bearing coefficient of friction is about 0.15.
made of alloy steel. The thrust load is carried Lateral adjustments for thrust alinement are
by a socket-type spherical bearing located be- accomplished using a tongue and groove arrange-
tween two bearing blocks. The surface of the ment between gimbal lower bearing block and
spherical socket has a Teflon-fiber-glass corn- thrust chamber dome.
Chapter X
Engine Systen s Design Integration
10.1 SYSTEMS ENGINEERING superiors he needs adequate authority, commen-

In the foregoing chapters the design of the surate with his responsibilities. He must have a
major liquid rocket engine subsystems and their broad and thorough understanding of the various
components has been discussed. By themselves, disciplines governing the design and develop-
these subsystems cannot fulfill a useful purpose. ment of the subsystems. In view of the latter's
Only when integrated into a whole will they func- high specialization, he cannot be expected to
tion as a system and produce for what they were have a complete education in all these disci-
designed: thrust. While the engine designer plines. However, in the words of Du Pont's
thus must view the complete rocket engine as the Crawford H. Greenewalt, he must have "the abil-
system, he should not lose sight of the fact that, ity to create a harmonious whole out of what the
to the stage builder, the engine is again a sub- academic world calls dissimilar disciplines,"
system. (In fact, to the launch vehicle systems and that in this respect his job may be compared
engineer, the stage in turn becomes a subsystem, to that of a "symphony conductor under whose
in addition to other stages, ground-support equip- hand a hundred or so highly specialized and very
merit, launch facilities, downrange stations, etc.) different talents become a single effort of great
In an earlier chapter the major engine param- effectiveness."
eters requiring optimization during implementa- The job of systems integration does not
tion of rocket engine design were defined. It is merely assure that all parts physically fit to-
apparent that the subsystems discussed will not gether, although this certainly is a first basic
form an optimized complete system unless they requirement. Wherever parts and subsystems
are designed for one another witil this goal in join, an interface exists, a demarcation line be-
mind from the beginning. The subsystems pre- tween these systems. This may be a bolt hole
sented in the foregoing chapters have become pattern; a flow of fluid, heat, or electrical cur-
highly specialized fields during recent years. rent; an exchange of forces, loads, or torques;
Their designers and developers often have only or a complex dynamic interaction.
a rather general feeling for the functions and The importance of systems engineering is
peculiarities of the other subsystems. Their being recognized to an increasing degree. Figure
concentrated effort will not come about unless it 10-1 shows the final configuration of the engine
is guided by an important function: systems for an early postwar ballistic missile. Compar-
engineering. ing this to figure 2-4 we realize the progress
In most rocket engine projects this function made through good systems engineering. How-
is assigned to the "systems engineer," commonly ever, to have designed and developed a good-
called project engineer or project manager. In looking, perfectly functioning rocket engine is
conformance with user requirements, he has the not enough.
job of establishing all ground rules and perform- It may appear superfluous to state that the
ance parameters (ch. II) and to subsequently design of rocket engines must also consider
optimize them through complete integration of all vehicle application. Yet in the daily grind of
subsystems during the entire design and devel- problem solving, engine detail design and devel-
opment period. opment, this is easily forgotten. The rocket
Because of the previously discussed interac- engine and component designer and developer
tion of funding, time, and reliability, the project who can think and converse in terms of vehicle
engineer must be a good manager, too, in addi- application will be sought after indeed, and so
tion to his engineering qualification. From his will be the company he represents.
383
We can easily realize the importance of con-

tinuous and close communication between engine
designer and vehicle builder. Complete and free
exchange of correspondence, documentation and
progress reports, frequent design reviews, and
above all, prompt notification of changes is
vital. During engine development, closest simu-
lation of flight vehicle configuration, when per-
forming static engine development tests, will
materially reduce the possibility of later "sur-
prises." Each vehicle type, according to its
mission, will have somewhat different interface
problems. Many of these occur, however, in
almost any vehicle type.
10.2 ENGINE SYSTEM DESIGN INTEGRATION

BY DYNAMIC ANALYSES
The Scope of Dynamic Analyses
Dynamic analyses are essential for optimum

engine systems design integration, commensurate
with vehicle mission requirements. Dynamic
analyses may be grouped into two basic areas:
1. lnternai operating dynamics of the engine
system.-This refers to system schematic and
control optimization, component optimization
from a functional as well as system transient
point of view, optimization of system start tran-
sients, minimization of cutoff impulse, determi-
nation of cutoff surges, etc.
2. Engine-vehicle operating dynamics.-This
refers to vehicle tank pressurization systems
design for adequate transient and steady-state
engine performance, engine-vehicle structure and
engine-guidance operating compatibility and
Figure IO-I.-Ballistic missile liquid rocket stability, overall vehicle performance during
engine showing greatly improved simplicity. special maneuvers, etc.
General Approach to the Analyses

Rocket engine design for vehicle application
is a consideration not only for vehicle flight or The techniques and equations used, and the
for optimum mission performance. Long before general approach taken by industry and Govern-
the engine will fly, and after it has been de- ment agencies toward the solution of the various
signed and developed, it must be installed. The dynamic problems in rocket engine systems, are
facets of engine installation in a vehicle range the result of many years of effort and experience
from more complex areas, such as guidance loop in the areas of analysis, synthesis, and correla-
compliance and prestart conditioning to simple tion of rocket engine operation. The philosophy
"nuts and bolts" problems, such as matching governing this approach postulates that the rele-
connecting fitting dimensions and attachment vant characteristics of any system depend on the
bolt sizes. characteristics of their components and physical
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
of an engine system. meaningful.

There are many ways to describe a rocket K (10-4)

engine system mathematically. The choice will a=2gpA2Ra2
be based on the answers expected from the
model. A model used to predict engine systems
orificing requirements may call for high steady- b=_g (10-5)
where
state accuracy over a stated operating range,
Pl =pressure at the inlet, psi
while the dynamic characteristics are of little
P2 =pressure at the outlet, psi
consequence. A model used for the design inte-
a =design factor for turbulent flow pressure
gration of control components and subsystems
drop, sec2/lb-in 2
may require good dynamic accuracy for inter-
b = design factor for fluid inertia to changes
mediate frequencies (0.01-100 cycles/sec)and
in flow rate, sec2/in 2
good static accuracy. A model needed to de-
_z = weight flow rate, lb/sec
scribe high-frequency effects (100-10000 cycles/
it = change in weight flow rate, lb/sec 2
sec) may require little in the way of static accu-
h1_2 = gravity head of the fluid in the duct, in.
racy, since the phenomena under study occur so
If the system is subject to an accelera-
rapidly that the system as a whole has no time
tion other than I g, this term must be
to shift to new operating levels.
modified
The mathematical model most frequently used
p = fluid density, lb/in s
is of the type employed for system main-stage
K = duct or component resistance coefficient,
operation and control design. This usually
to be obtained from flow tests
serves as the basic mathematical model of an
A = flow area of the duct, or flow area of the
engine system, with some modifications incorpo-
component at the design operating
rated for other, special applications. In general,
point, in 2
all mathematical descriptions of a basic model
Ra = component area ratio=component flow
are for conditions around the systems design
area/A (at the design operating point,
point, with the following assumptions:
Ra = 1)
(1) All liquid propellant flows are incompres-
L :length of the duct or component, in
sible and at constant temperature
g =gravitational constant, 386.4 in/sec 2
(2) All gas equations are based on perfect
2. Combustion process and operating dy-
gas performance, where the gas proper-
namics:
ties are functions of composition
Refer to figure 10-2, which describes sche-
(3) All vehicle-supplied parameters, such as
matically the combustion process. A transporta-
main propellant tank discharge pres-
tion time delay is assumed, representing the time
sures, are constant.
required for the propellants to pass through the
injector and enter into the combustion process,
and expressed as
Examples of Equations for a Mathematical
Engine Model poAo( Lo + Lo')
(10-6)
r° = g'o
The mathematical expressions and equations
for the various physical processes and operating
p[At(L[+ LI')
dynamics of a rocket engine may be derived from (10-7)
equations given in chapter I and in other chap- r[= Wf
ters for the design of the various components.
Here, we will present several typical examples
to illustrate the application of these equations
, _ -, w o
in a mathematical model.
,.=: AI
1. Pressure drop of fluid flow in a duct or

component:
Figure I O-2.-Schematic description of the com-
P, - P2 =aw2 +bit-Ph,-2 (10-3) bustion process.
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
Wc = wo + wt (10-9) g'c = flow rate of the gas emerging from the

combustion chamber, lb/sec
Wo
Ro =weight fraction of oxidizer stored in
Ro = (w ° + wt) (10-10) the combustion chamber
At =throat area, in 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
t generators, and other types of combustor.
3. Turbopump operating dynamics:

wf= wf÷ fo [P-rfSwl-(1 - R°)fVc]dt(lO'12)
P2 - Pl =aNp 2_y (10-16)

A tPcg
wc- C* (10-13)
4 rr2rp2p
a= (10.17)
g
T¢
=t(Ro) (10.14) ¢ = [(4,) (10-1s)
C*=t(Ro) (10-15) 4,¢

x= (10-19)
r/p
= time required for oxidizer and fuel to
pass through the injector and enter
Tp = bNp2x (10-20)
into the combustion process, sec
Po, Pt = density of oxidizer and fuel, lb/in 3
Ao, At =injector area, oxidizer and fuel, in 2 b = 8 z73rp3App (10-21)
Lo, Lt = travel distance of injected oxidizer g
and fuel prior to impingement, in

Lo', Lt' = travel distance of impinged oxidizer Nt = RgNp (10-22)
and fuel prior to combustion, in
t = time period in consideration, sec 1485 _i'tAHot
Tt - (10-23)
¢¢o. fvf =injected flow rate of oxidizer and Nt
fuel, lb/sec
iI_o, _f = weight of oxidizer and fuel stored in
the combustion chamber volume, lb,
(10.24)
at the beginning of time period t AH=CpT° - -_t
Wo, wt = weight of oxidizer and fuel stored in
the combustion chamber volume, lb, lNp = f (RgTt - Tp)dt (10-25)
at the end of time period t
s = complex operator in a Laplace trans- Equation (10-25) is a torque-balanced equa-
formation tion for turbopumps. Any unbalanced torque be-
Pc = pressure of the combustion gas, psia tween turbine and pump will initiate a change of
Tc = temperature of the combustion gas, °R shaft speed causing the integrand to seek zero.
where This includes pressure drops in various

p, =pump inlet pressure, psia flow passages, turbopump operating per-
P2 =pump discharge pressure, psia formance, and thrust chamber heat trans-
a =pump pressure rise design factor, lb- fer and combustion characteristics
sec2/in 2 (3) Optimization of engine system steady-
b =pump torque design factor, in-lb-sec 2 state operation and performance, by

properly calibrating and matching tile
rp =pump impeller radius, in
Ap =pump flow area, in 2 design operating points of various com-
p =density of the pumped fluid, lb/in 3 ponents. (See sec. 10.3.)
Np = pump shaft speed, rev/sec (4) Determination of engine system mainstage

Nt =turbine shaft speed, rev/sec performance characteristics, including
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
gas, Btu/lb (6) Evaluation of various potential perturba-
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
y =turbine gas specific heat ratio main-stage operation of an engine system is
Rt =turbine pressure ratio established, it can be utilized to study special
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-
_, =nondimensional pump head coefficient pound-thrust LOz/RP-1 pump-fed rocket engine
¢ =nondimensional pump flow coefficient for an intermediate-range ballistic missile had

x =nondimensional pump torque coefficient been established on an analog computer. Fol-
r/p = pump overall efficiency lowing design of an additional engine-thrust-
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
(4) Evaluation of comp¢,uent dynamic charac-

, FUEL PUMP OISCHARGE-_%,
teristics and interactions during start ,ooo, _ _i ]
i rI 7-
I "--_--__,,....
! .._,w,
.,,- "- j iVALVE
900 " _,-- - b -_--_ ..... ,..... T " -_
transients, such as combustion chamber
I Ga,S i
I i
' ! f ;
, , RO$.',O,,
ignition delays, gas generator tempera- [ \ VALVE I i ' / "
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
dynamic analyses of its start transient become

rather complex. They may include effects such °o_.l 02 0.3 0.4 o._ 0.6 o7 o8 09 _o
as water hammer (wave equation) in the propel- TIME FROM START SIGNAL, SECONDS
lant feed systems, distribution of heat transfers

Figure lO-3.-Graphic presentation of the .start
and pressure drops throughout the high-pressure
fuel feed system, choking of hydrogen gas in the transient model for a typical turbopump-feed
engine system utilizing a gas generator for
chamber coolant passages, stall characteristics
turbine drive.
of the fuel pump, cavitation at the pump inlets,
changes in fuel density in the pump caused by
enthalpy changes from pumping, and many others. (3) Evaluation of possible temperature surges
Because of this complexity, me equations of in main chamber or gas generator during
these mathematical models are usually programed the cutoff transient
for a digital computer. (4) Evaluation and optimization of total cutoff
Figure 10-3 presents graphically the start impulse, by minimizing cutoff time and
transient model of a typical turbopump-feed en- improving repeatability
gine system utilizing a gas generator for turbine (5) Evaluation of engine thrust decay charac-
drive. Valve timing, as well as pressure buildup teristics
characteristics of gas generator, propellant A series of cutoff sequences with modified
pumps, and main chamber are indicated. Other shutoff timing of the control valves is usually
parameters, such as gas-generator combustion- simulated with the mathematical model. The
gas temperature and propellant flow rates, can various simulated cutoff runs are then analyzed
be included in the model. to determine potential problem areas, and whether
Several alternate engine-start methods are these problems are a function of the particular
usually simulated with the start model, in order sequence used.
to evaluate potential problem areas and to opti- Figure 10-4 presents graphically the cutoff
mize systems start transient operation. transient model of the typical engine of figure
Important objectives of dynamic analyses of 10-3. The engine-thrust-decay characteristics
engine cutoff transients are: are represented by the main chamber pressure
(1) Investigation of the systems schematic decay curve. The integrated area under the
for needed cutoff transient controls, chamber pressure versus time curve may be used
including operational sequencing and to assess the engine cutoff impulse.
timing of various control valves
(2) Evaluation of pressure surges and other
Dynamic Analysis of Engine-Vehicle Interactions
adverse effects in propellant ducts and
feed system components during the cut- These analyses may be performed during the
off transient initialdesign phase of an engine system, as
I000 10.3 DESIGN INTEGRATION FOR ENGINE

I ! ; r I I
SYSTEM CALIBRATION
900 ---
- i • i,.___VAI
|POSITION
......... ' E
"--FUEL PUMP 3_SCHARGE I
i I
I
i
[
z
Design Requirements for the Calibration of an
\1_- [ i' _ .... r T OPE N
_ i b I ",. I J Engine System
_'_ /-_ GAS GENERATOR _ ././././_MAINFUEL
L,_"F VALVE! i x',_i'l VALVE 1 Because of unavoidable mechanical toler-
700
ances, it may be expected that the operating

I' _'_' VALVE I I _ /
L_
--_.- _; - ._ -_::-CLOSEI_ characteristics and performance of the various
6oo-d
II I ',I I ! , engine system components will deviate some-
Ld what from their nominal design value. A certain
E 500----
03 \ \ _ OXiDiZER PUMP DISCHARGE, amount of calibration is always required for

03
hi
at"
r_
400 -- -"L-/I I I_\_ :_ GAS GENERATOR
I L
CHAMBER
these components, as well as the engine system
as a whole, to attain the desired engine perform-
l__, / PRESSURE I [ I !1
ance characteristics within design specification.
[ _ \_lr"_" MAIN CHAMBER PRESSURE [ Therefore, provisions must be made in component
and systems design to permit effective calibra-
tion during system integration.
The specific impulse Is of an engine system
is the ratio of thrust F to propellant weight flow
o i rate &. Thus, any deviations affecting F or &
0 0._ o.z 0.3 0.4 0.5 0.6 o.-t o.e 0.9 Lo
TIME FROM CUTOFF SIGNAL, SECONDS will affect system performance. I s also is a
function of propellant mixture ratio. It is desir-
Figure lO-4.-GrapMc presentationof the cutoff able, therefore, and beneficial to calibrate an
transient model of the typical engine of figure engine system by adjusting its propellant feed
10-3.
system. Prior to complete engine system cali-
bration, the pressure or pressure drop versus
well as following development test firings. flow characteristics of each individual compo-
A_eas of general interest to be analyzed may nent should be calibrated and evaluated.
include, but should not be limited to- Hydraulic and pneumatic components, such as
(l) Engine system operation and performance pressure and flow regulators, valves, flowraeters,
requirements from a vehicle mission ducts, and lines, can all be readily calibrated on
point of view flow benches. However, those components which
(o) Matching of engine propellant supply re- operate at temperature extremes, such as thrust
quirements with the vehicle propellant chamber assemblies, gas generators, and turbo-
system, including dynamic evaluation of pumps, are best calibrated by combining the flow
the vehicle propellant tank pressuriza- tests with actual hot firings. The characteristic
tion system, PU control system, vehicle propellant flow curve of an engine system is
acceleration and sloshing effects, and obtained by summing the pressure or pressure
feed system-combustion coupled in- drop versus flow curves of the various compo-
stabilities nents (figs. 10-5 and 10-6).
(3) Matching of the engine controls with the The general design approaches toward cali-
vehicle guidance system, including re- brating an engine system to attain its design
sponse of engin.o start, cutoff, thrust thrust at design mixture ratio are:
level, and vector controls to vehicle
guidance commands (1) The design operating point of each com-
(4) Simulation of interaction between engine ponent should be kept within the rela-
systems operation and vehicle dynamics. tively flat region of its pressure or
(This may involve closed-loop coupling pressure drop versus flow curve.
of an analog simulation of vehicle guid- (2) The mechanical tolerances and built-in
ance and trajectory characteristics with adjustments of each component should
an engine system during hot firing) be designed so that the random deviation
_____._.__._>
_MINIMUM REQUIRED TANK PRESSURE VS r_.OW

_CURVE OR CALIBRATED SYSTEM RESIST/kNCE
REGULATOR OUTLET PRESSURE SYSTEM _r _ CURVE
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
PROPELLANT FLOW, LB/SEC
Figure 10-5.-Propellant flow design characteristics of a typical pressure feed engine system
(oxidizer or fuel).
v--CALIBRATED FUEL SYSTEM

\RESISTANCE CURVE
x--FUEL SYSTEM RE, STANCE CURVE WITHOUT ORIFICE

_. _- OXIDIZER SYSTEM RESISTANCE
\CURVE WITHOUT ORIFICE
---.... ,,,
,---FUEL PUMP PRESSUI_:
CALIBRATION
"_, _ VS FLOW CURVE AT NZ
ORIFICE DESIGN
PRESSURE DROP [ /--FUEL PUMP PRESSURE
t
Pfd
°°°
J" OX,Q, ZER
PUMP
PRESSURE
i i
FUEL FLOW LB/SEC
OXIDIZER FLOW LB/SEC
Wfd wfo Wfb %a Wod
FUEL SYSTEM FLOW DESIGN CHARACTERISTICS OXIDIZER SYSTE},IFLOW DESIGN CHARACTERISTICS
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
..... _ ----_:::.:=?:
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 ::---
Substitute this into equation (6-7) to obtain

F6970 (7950) °'s11.aaa
the required pump speed H = L: _ ] = 4730 ft
or
N- Ns H°'Ts _ 1980 × (2990) °'Ts = 7190 rpm
QOS (12 420) °-s 4730 × 50.45
= 1666 psi
144
Fuel pump volumetric flow
The equivalent fuel pump discharge pressure

892 × 449 = 7950 gpm under these conditions would be 1666 + 45- 25
Q- 50.45
= 1686 psia.
The required pressure drop of fuel line cali-
From equation (6-7),the fuel pump nominal bration orifice under these conditions would be
developed head at 7190 rpm 1686 - 1125- 220- 310- 12- 17=2 psi.
Therefore the required range of adjustment for
the pressure drop of the fuel line calibration
H = (.NQaS_
\"_-s / I333 = f7170 ×
_ (7950)°s_ ] 1333 = 4900 ft orifice is from 2 to 216 psi.
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.
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)
The following data were obtained from design Solution

analyses and component tests of the A-1 stage
A trial-and-error
method is used to solve this
LOX/RP-1 engine system at nominal rated con-
ditions; i.e., 750000 pounds thrust at sea level: problem. Our firststep is to approximate engine
Thrust chamber sea level specific impulse at system and gas generatorpropellantflow rates.
We substitutethrustchamber Is intoequation
1000-psia nozzle stagnation pressure, and a
mixture ratio of 2.35 O/F=270 sec (1-28):
Turbine exhaust gas specific impulse = 32.6
sec = F _ 750 000 = 2778 lb/sec
Oxidizer pump developed head--2930 ft Is 270
Oxidizer pump overall efficiency = 70.7 per-
cent The thrust chamber mixture ratio can now be
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
1250 l ' "

Engine fuel flow rate _i'f= 2778 - 1948 = 830 lb/sec
n,.,
0 1200
From equations (6-12) and (6-13), the required
oxidizer and fuel pump drive horse power for this
Z .
-- t,t.I 1150 approximation are:
n., rr'
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:
7OO,OO0 7500(30 8OO,OO0

Thp = 14 720 + 10 960 + 500 = 26 180 hp
ENGINE THRUST, POUNDS
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:
0.707 x 26180 Fuel pump flow rate

_i't= 359x0.582 -88.4 lb/sec
=(91.7- 26.6)+(2768- 1941)= 892.1 lb/sec
We use this value to start a new calculation Turbine shaft horsepower

cycle to separate main chamber and gas gener-
ator data. 1970.6x 2930 892.1 x 4790
ThP = 550 x 0.706 + 550 × 0.659 + 500
The thrust generated by a turbine exhaust gas
flow rate of 88.4 lb/sec = 88.4 x 32.6 = 2880 lb. = 14850+ 11 790+ 500= 27 140 hp
Thus the main chamber thrust:
Turbine gas flow rate w_ = 0.707 ×27140 _ 92 lb/sec

Ft = 750 000- 2880 = 747120 lb 359 × 0.582
Thrust chamber flow rate This value closely confirms the assumptions
for the last trial. Thus:
l_,tc = 747270120 2769 ib/sec (a__)Nominal turbine exhaust gas thrust

= 92 x 32.6 = 3000 lb
(b__)Nominal main thrust chamber thrust
Oxidizer pump flow rate = 750 000 - 3000 = 747 000 lb
(c) Nominal thrust chamber propellant flow
_ 88.4 x 0.408 + 2769 x 2.35 + 3
1+0.408 1+2.35 rate = 747 000/270 = 2768 lb/sec
Nominal engine system propellant flow
= 25.6 + 1942+ 3= 1970.6 lb/sec rate = 2768 + 92 = 2860 lb/sec
Nominal engine system oxidizer flow rate
Fuel pump flow rate 92 x 0.408 2768×2.35
+ - 1967.7 lb/sec
= (88.4- 25.6) + (2769- 1942) = 889.8 lb/sec 1+0.408 1+2.35
Nominal engine system fuel flow rate
Turbine shaft horsepower -: (92 - 26.7) + (2768- 1941) = 892.3 lb/sec
(d__)Nominal engine system O/F mixture ratio
1970.6×2930 889.8x4790 = 1967.7/892.3 = 2.20
Thp- 550x0.706 + 550x0.659 +500=27090 hp
(e_) NominaI engine specific impulse
= 750000/2860= 262.4 sec
Turbine gas flow rate
0.707x27090 Engine Performance Variations Resulting From

wt- 359 x 0.582 = 91.7 lb/sec Off-Nominal Conditions
Engine performance characteristics at various

We use this value for another calculation
off-nominal conditions must be available to the
trial:
vehicle system engineer. They can be summa-
rized in the graphic form such as figure 2-1 (en-
The thrust generated by a turbine exhaust gas gine thrust and specific impulse versus altitude
flow rate of 91.7 lb/sec=91.7 x 32.6=2980 lb. curve), or by means of tabulated engine influence
coefficients which will be discussed, The effects
Main chamber thrust Ft = 750 000-2980 = 747 020 lb of off-nominal conditions of the following engine
system performance parameters are considered
Thrust chamber flow rate vital for the design of a vehicle system:
(1) Atmospheric pressure
747 020
_i'tc = 27-----0_
= 2768 lb/sec (2) Propellant densities
(3) Pressures at _he engine propellant inlets
(4) Propellant mixture ratio and vehicle PU
Oxidizer pump flow rate
control
(5) Vehicle acceleration
_ 91.7 × 0.408 _ 2768 x 2.35 + 3= 1970.6 lb/sec
1 + 0.408 1 + 2.35 (6) Throttling of the engine system
-,qmNlm
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,
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
TABLE lO-1.-Influence Coefficient for the A-I Stage Engine System
[Value of C* correction to be obtained from the C* correction versus mixture ratio curve shown in fig. 10-8]
Independent variables and nominal values

A 1-percent increase of
independent variables
Oxidizer pump Fuel pump
causes the following Atmospheric Oxidizer Fuel C _
inlet suction inlet suction
percentage change of pressure, density, density, correction
pressure, pressure,
dependent variables 14.696 psia 71.38 lb/ft a 50.45 Ib,/ft a 1.0000
56 psia 45 psia
Dependent variables and nominal values:

Engine thrust, 750000 lb .......... -0.1780 1.8750 -0.7420 0,0440 -0.0066 1.1030
Engine specific impulse, 262.4 sec . - .1780 ,2650 - .0640 .0072 - .0150 1.1350
Engine mixture ratio, 2.20 ......... .0000 1.6420 -1.3650 .0270 - .0020 - .0260
Engine oxidizer flow, 1967.7 lb/sec. .0000 2.0430 -1.1120 .0465 .0108 - .0632
Engine fuel flow, 892.3 lb/sec ..... .0000 .6530 .3120 .0207 .0045 .0094
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
=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
(-0.002) × (49-45)__0.0162 or -1.62 percent

45
-0,1
_-.
Z
L_ From figure 10-8, the C* correction for a mix-
-02 I ture ratio change of-1.62 percent is approxi-
i mately -0.02 percent. From table 10-1, the in-
Z
9 -o.3 fluence coefficient for engine system thrust is
i.-
1.1030, for a i percent C* correction. Thus
e¢
_ -0.4
u (Percent change in F) = 4.531 + (-0.02) x 1.103
-0.5 = 4.509 percent
-O.6 Therefore, engine system thrust considering

-16 -12 -8 -4 8 12 effects of C* correction:
CHANGE IN ENGINE MIXTURE RATIO,
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
5 MICRON FILTER- fG.G._ BLEED VALVE CONTROL SOLENOID VALVE

BLEED PRESSURE REGULATOR-_ / HELIUM CONTROL SOLENOID VALVE
\ / /---'--'--MAIN PRESSURE REGULATOR
MA,NSTA_F CC1NT _\ tlJ_.._/_ ,/._t_ j.---,GN,T,ON PHASE CONTROL

.......... _ L
-..-1::;-_:_i'_>_!_ __FAIL-SAFE CHECK VALVE
HIGH PRESSURE __ _ 10MICRON FILTER

RELIEF VALVE t _x__LOW PRESSURE RELIEF VALVE
HELIUM
INLET
PNEUMATIC CONTROL PACKAGE SCHEMATIC
VENT PORT CHECK VALVE PRESSURE RELIEF VALVE
IGNITION PHASE
HELIUM SOLENOID CONT.
IN LET VALVE
VENT
CHECK VALVE
VENT PORT 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
forces at full gimbal deflection, including an

adequate reserve for normal and for side loads.
The bolt-hole pattern must permit adjustment for
tolerance deviations. In general, engine and
vehicle attachment halves must be designed for
one another. LU|E OIL TANK
Figures 10-10 and 10-11 illustrate the pack-

aging design details of a typical pump feed sys-
tem. It is a LO2/RP-1, fixed-thrust engine with
constant chamber pressure control. The basic
engine package consists of the following sub-
packages:
(I) Gimbaled main thrust chamber assembly
(thrust chamber, injector, dome, oxi-
dizer elbow, and gimbal mount)
(2) Turbopump assembly (propellant pumps,
turbine, gearbox, lube pump, electric
heater, auxiliary drive)
(3) Gas generator assembly (combustor, con-
trol valves, regulator and turbine inlet
L_N OX_OIZE_ V_LVE
duct)
(4) Main oxidizer duct assembly (including
main oxidizer valve) Figure 10-I I.-System packaging design detail of
(5) Main fuel duet assembly (including main the engine shown in figure 10-10.
fuel valve)
(6) Turbine exhaust duct assembly (includ-
(10) Electrical control package
ing heat exchanger)
(11) Engine thrust frame assembly
(7) Engine start subsystem (oxidizer and
fuel tanks, control valves)
The majority of the major component and sub-
(8) Turbopump lube subsystem (lube oil tank system packages are installed within or at the
and fittings) periphery of the engine thrust frame assembly.
(9) Pneumatic control package The main thrust chamber assembly is attached to
the thrust frame through a gimbal mount.
Mechanical Protection of Engine System

Packages
It is normal that several years may elapse

between the date a liquid propellant rocket en-
gine is completed and accepted by the user, and
its vehicle flight. Design, therefore, must con-
sider means to protect the engine system pack-
age in transit or storage against moisture, dirt,
and shock. These include simple closures, such
as caps, plugs, and cover plates applied to valve
and regulator vent holes, propellant inlets, and
to other openings. Frequently, these closures
contain desiccant bags and indicators which
Figure lO-lO.-Major component and subsystem warn, through change of color, of undesired in-
packages of turbopump-[ed liquid propellant trusion of moisture. Certain lines, however, may
rocket engine. require communication with the ambient air
ENGINE SYSTEMS DESIGN INTEGRATION 4O3
are many tasks, such as sequencing which can

be accomplished nmch more effectively electri-
cally than would be possible by mechanical
means.
It is not possible nor necessary, in the frame-
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
Figure I 0-13.-Typical liquid rocket engine electrical diagram (partial).
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
the output voltage back to the input through a

diode. In this case, shorting the base of Q,
must be used to turn off the switch.
Figure 10-15 shows a typical solid-state

LOCK-IN METHOD
L'I OUTPUT switch module. Following assembly, a module

DIODE is immersed in a potting compound. Overall
R4
module dimensions after potting are comparable
L to those of a matchbox.
Ichor//
i t IM '4-" "_
Figure 10-16 shows a small portion of an
INPUT engine-mounted sequence-controller diagram.
+( --------_v.
R2 LOAD The modules shown, together with the others
required for engine operation, including connect-
ing wiring and connectors, are housed in a her-
tiN
metically sealed box or can.
Although the two diagrams are not related,
the signal emitted at pin X of receptacle J-5 in
GROUND
figure 10-13 would be suitable to initiate engine
Figure I 0-14.-Typical rocket engine solid-state start at contact A of plug P51 in figure 10-16.
switch circuit or "module" diagram. Similarly, the signal emitted from pin K of plug
Electrical Components
To implement rocket, engine electrical cir-

cuitry such as the one partially shown in figure
10-13, a number of components are required.
These are:
Relays and Switches

Relays and, more recently, the solid state
switches are used to receive usually low current
command signals from the vehicle or ground con-
trol center and to translate them into properly
sequenced actuation signals to igniters, control
valves and other elements. In combination with
interlocking relay contacts, bias voltages, valve
Figure lO-15,-Assembled soJid-state switch position switches, continuity monitors, tempera-
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
Jal FO_ 20m$ _l_

P_r
•5, I <D
. I s c
I' _
T($1 ¢O_'L(?( z Z?-','

..... ,
Cm,T g_BI (_smg s'rAx'r
r,_;'(_ ........
, _ e÷ CLOS*J_I[ • _ _Z{)$-II,,.,
3 r •
I_NITION
L. ................. J
zz_'s_ c_2_ c_SO
Figure 10-16.-Portion of a typical engine-mounted sequence controller diagram using

solid-state switches.
ENGINE SYSTEMSDESIGN INTEGRATION 4O7
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
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
components will require individual heating at all

times after propellant loading. Of these, some
may require this only through takeoff from a 28-
volt dc ground source, if their heat capacity
keeps them sufficiently warm for their relatively
short period of operation following heater dis-
connect. Other components, particularly those
which have to operate during extended periods
following takeoff, will require supply from an Figure IO-18.-Wire harness installation for main
airborne power source, which is usually the 28-V valve position switches.
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
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
El2 BI8(RD) "_

,.- EI3AI8(WH) :?'-_ PP3
_- E]3BI8 (RD)
EI3B18 (RD) "_
EI4BI8(RD} •
EI5BIB(RD) J
_'J _ PI6 8.'25

EI6BI8(RD} -J
RI_0"_ E,ZAIS<WH)_ EISAIB

EI5BIB(RD) (WH) _._% PI5
E_.3 PI4 _____ P[2
;_____ j/IEIyBIB_RD)
PHYSICA-ROUTING DIAGRAM
(CUMULATIVE DISTANCES, INCHES
-_ RECEPTACLE ._PI6
::::::((:;:'-_--J RECEPTACLE
DENOTES BRANCH POINT
, (WH) B (RD} DENOTES COLOR OF WIRE INSULATION
_" DENOTES TWISTED SHIELDED PAIR

'_LEN_H--GT DENOTES TWISTED PAIR
FROM FACE OF MAIfJ RECEPTACLE (J-riO),
Figure l O-lg.-Typical rocket engine wire harness diagram.
[
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]
[] [] ]
=u _ _'IN o c.,....r)_r@ 0]_

TEMPERATURE
MEASUREMENTS
!li
c_ < c_ < (SENSORs)
i E.,NE I i PRIMARY ENGINE

SECONDARY
ENG_
INSTRUMENTATION --E
: CONTROL I INS TRU MENTATI0_N
PACKAGE
PACKAGE
PACKAGE I
Y' LTI PU V_LVE POSITION (POTE.NTIOMETERS AND SWITCHES'I L}.J

MONITOR _ MEASUREMENTS
Figure 10-20.-Typical block diagram, engine to vehicle electrical connections.

diagram for the principalelectrical

connections its significance. The engine designer must be
between engine and vehicle. Flexibility of these fully cognizant of it, and must be sure to consult
trunksusually poses no problem in view of the the development engineer and the instrumentation
relativelysmall engine deflectionduring gimbal- specialist at the very beginning of his design.
ing,but ifheavy sheathing(armor)is used, care
must be taken to avoid excessive loads to the Principal Types of Instrumentation
gimbal actuatorsor damage to the cables. Instrumentation signals to the vehicle are
generated by engine-mounted end organs or
transducers, with typical applications as follows:
10.7 ENGINE INSTRUMENTATION
Temperature gages.-Thermocouples or re-
To evaluate the qualityof his design dur- sistance bulbs are the principally used types.
ing and aftertest,to monitorprogress,and finally Their main application is for fluid-temperature
to "sell"his engine to the customer,the de- measurement at the pump inlets for the determi-
signerneeds recorded evidence of the engine's nation of mass flow rates in combination with
overallperformance. These records may broadly volume-reading flowmeters. Also, temperature
be classifiedby two categories: visual inspec- measurements are made in other critical metal or
tion and measurements. fluid areas, such as at the gas generator, at the
Visual inspectionsare self-explanatory. They turbine inlet, at the gas storage vessels, and at
consist,oftenafterdisassembly of a component, solid propellant gas spinners, if such are em-
of an inspectionforcracks, fractures, deforma- ployed.
tions,signs of overheating,etc. Surfacedis- Pressure switches.-A typical application
colorationsmay oftenbe a sign of fluidleaks. would be a signal from the engine to the vehicle
In fact, minute amounts of a dye have been added when thrust (chamber pressure) reaches a preset
to a number of propellants such as alcohol- and value during decay to initiate start of the next
the kerosene-type fluids, which leave qualitative stage and separation of the spent one.
evidence at the leak spot. Visua}-checks as- Pressure transducers.-A variety of types is
sume a particularly important role in the event of available: strain gages, capacitance types,
serious malfunctions. As a rule, they do not variable reluctance types, potentiometer types,
require special preparations by the designer, piezoelectric types. Table 10-2 lists, among
except perhaps written instructions (drawing others, important pressures which can be ex-
callouts) and forms, pected to require monitoring and transmission to
By contrast, the second category, measure- the vehicle and its telemetry system.
ments, does require consideration during engine Accelerometers.-It is often important to
design. To make the required measurements, measure the vibrations occuring in various loca-
instruments are needed. Some of these may be tions of the engine system, and how these vary
simple devices such as gages, micrometers, and between static (tiedown) firing and flight. Accel-
torque wrenches, which can only be used before erometers are used for this purpose and also,
and/or after test, and which will not be dis- instead of pressure switches, to sense and sig-
cussed here. However, for the majority of the nal thrust decay for staging.
instruments used for engine design and perform- Tachometers and flowmeters.-Small magnets
ance evaluation, particularly during test, the mounted in a suitable manner to rotating parts of
designer must provide adequate provisions for flowmeters or turbine wheels will induce periodic
installation. Furthermore, the correct execution voltage pips in stationary coils. By counting the
of command signals fed to the engine and the pips as a function of time, turbopump speed or
behavior of the major performance parameters volume flow rate can be determined.
must be transmitted back to the vehicle and from Depending on the type of transducer used, a
there to ground, either by hard wire (static tests) certain amount of signal conditioning will have
or by means of a radiating telemetry system to be provided at or near the transducer. Some
(flight). The instrumentation of a liquid rocket of the latter give relatively weak signals (ther-
propulsion system is a large field, and within the moelements, capacitance-type gages, tachom-
framework of this book we can merely point out eters, and flowmeters), which must be amplified
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
cant levels. A transducer thus mounted, in com-

bination with a good inking recorder or precision
gage, may have a full span response time of
about 1 second. It would be pointless to con-
nect such a transducer to an oscillograph.
For certain parameters it may be desirable to
combine accuracy with at least some frequency
response. By shortening the line length to, for
instance, 18 inches, a given setup may have a
flat frequency response up to 25 cps. In these
cases it will be the burden of the engine de-
signer to provide for means to mount and connect
the transducer. It should be noted that it is tube
length which affects response, while bends
appear to have a negligible effect. Occasionally
attempts have been made to improve response by
filling the connecting line with a liquid. Apart
from operational inconvenier;ces, severe un-
damped oscillations may result and the practice
is not recommended. Figure 10-22 shows a pres-
sure transducer installed and connected tn engine
pressure taps.
INSULATOR For frequency responses in the range from 100
to 2000 cps, close-coupled mounting of the trans-
ducer is required. Typical transducers, resem-
bling a spark plug, are directly screwed into the
instrumentation boss (fig. 10-23). It is importan_
that the resonant frequency of the combined tap-
and-transducer cavity should be well above that
Figure l O-2I.-Properly designed thermocouple

well (installed in a 4-inch pipe).
(transients) must be recorded, with rapid response

but at some sacrifice in accuracy.
Because of the higher accuracy usually de-
sired for static measurements, the sensitive
transducers are often mounted in environment-
controlled boxes to protect them against vibra-
tion, temperature (both ambient and of the meas-
ured medium), and other influences. They are
therefore usually mounted some distance away
from the engine on the test-stand structure or on
the vehicle, at the end of several feet of tubing, Figure I O-22.-Pressure transducers installed and
which reduces frequency response below signifi- connected to engine pressure taps (lower right).
of the frequency range of the measured param-

eter. Figure 10-24 shows two typical installa-
BRAIDED CABLE
tions and associated equations to determine in.
Figure 4-28 shows a typical tap arrangement for
injector-end chamber pressure.
For still higher frequencies, above 2000 ops,
transducers are installed so as to leave essen-
tially no cavity at all (flush mounting). This
poses problems with hot-gas measurements.
Special water-cooled transducers have been de-
veloped for these applications. Needless to say,
an effort toward high-frequency response at the
pickup point is lost if the recording system is
i INNER SURFACE
not capable of using it.
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
._-----TRANSDUCER T-T /---TRANSDUCER
F 3 I I1 !l
L/.---DIAPHRAGM
BOSS_
ADAPTER -_/
_.--ENGINE
COMPONENT
_--- DIAPHRAGM
ENGINE COMPONENT WAll
WALL
TRANSDUCER INSTALLATION TRANSDUCER INSTALLATION
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
ACOUSTIC EQUIVALENT ACOUSTIC EQUIVALENT
(A) {B)
Figure 10-24.-Acoustic equivalents of pressure transducer mounting cavities.

For both large or small units, reliability is not a

fixed value, but is a function of development
"_"_ ELECTRIC PICKUP time. A realistic analysis considering all fac-
tors may determine the appropriate choice. The
absolute truth, however, may never be known,
since the cost of going both routes for a given
mission and of comparing final scores would be
prohibitive.
Clustering of liquid rocket engines or, in
other words, subdivision of vehicle thrust into
FL smaller units may be accomplished in a number
of ways. The choice for the most part will be
based on vehicle considerations. Depending on
the method chosen, however, engine design will
be directly affected.
ROTOR
Some of the basic considerations of this topic
Figure I O-25.-Turbine-type [lowmeter. which are specifically related to engine cluster-
ing and as they affect engine systems design and
integration, are briefly reviewed.
ducting (fig. 10-25). Turbine flowmeters are very There are no hard-and-fast rules which can be
sensitive to changes in upstream duet configura- applied to determine optimum cluster configura-
tion. Once calibrated as installed, however, tions. With the aid of liquid engine clusters
they exhibit a high degree of precision. actually used in practice, however, we will pre-
sent a number of considerations which may serve
Accelerometers and Strain Gages as a guide for future designs.
These are almost always applied externally,
in connection with special measurement pro-
grams. At the time of engine systems design, no Earlier Cluster Configurations
special consideration is usually required. Figure 10-26 shows the cluster of two experi-
mental LOX/alcohol engines for a planned but
discontinued ballistic missile. It consisted of
10.8 CLUSTERING OF LIQUID PROPELLANT
two units of 120 000-pound thrust each. Both
ROCKET ENGINES
subunits included a tubular-wall, gimbaled thrust
The idea of obtaining higher thrust levels chamber, and a geared turbopump. The turbines
expeditiously through combination of several were powered m parallel from a single, common
smaller rockets probably is as old as rocketry gas generator. The control system was also
itself. To today's liquid rocket vehicle and common for both units. Thus this propulsion
engine builders the topic of clustering still system was not a true cluster, since it was not
causes spirited debates, which reached a peak possible to develop and fire each unit independ-
in March 1952 when Collier's published an arti- ently. Several successful experimental flights
cle by Dr. yon Braun in which a space vehicle were achieved with this engine.
was described as having a first stage powered Figure 10-27 shows another experimental
by a cluster of 51 turbopump-fed rocket engines. cluster consisting of three units which, although
The debates invariably hinge around the deceiv- never flown, achieved a remarkable reliability
ingly simple question: is higher overall relia- record during static firings at thrust levels up to
bility obtained through combination of a number 500 000 pounds. This cluster, too, used a com-
of well-developed smaller engines, or through mon control and gas generating system for the
application of one large, specially developed three subunits. Most of its components were
unit. The answer is not simple. The failure essentially the same as those used for the two-
probability of the smaller units may be low, but engine cluster, except that propellants were
it increases with the number of units clustered. RP-1 and liquid oxygen.
"
Figure 10-26.-Dual-engine cluster. Figure 10-27.-Three-engine cluster.
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
load for a substantially higher cluster reliability

at any single-engine reliability level.
NO ENGINE D_JT
Another important consideration is cluster
diameter in relation to vehicle diameter. As witl
be discussed in section 10.9, it is mandatory
that all engines fit into the interstages from
which they must separate without interference.
This restriction is not valid for the first stage,
where engines may protrude beyond the tank pro-
J _- 0NE ENG,NE OUT file. The degree of p['otrusion, however, must be
/ balanced against drag losses.
/
Specific considerations for each chosen num-
L I i
4. 5 6 ber of engines follow (fig. 10-30).
NUMBER OF ENGINES
Figure 10-28.-Stage-payload weight as a function &Engine Cluster

of number of engines in cluster. This configuration provides the lowest total
thrust in the established band of 4 to 6 engines,
but it is optimum for engine interehangeability.
..____E_IEONE ENGINE OUT ( ADVANCED ENGINE
All four engines will be gimbaled. Only one
1,0
LOPMENT PHASE ) installation is required: all inlet ducts and other
engine to stage connections can be made id.. _ti-
NO ENGINE OUT _ ADVANCED ENGINE -'-"--
cal. The thrust vector control mode is simplest:
all engines deflect in pitch and yaw, and all
actuators participate equally in roll control (fig.
10-80(a)).
0.9--
F-
08
YAW -- --0_
ACTUAT
I
a. b.
E I i
4 fi 6
NUMBER OF ENGINES
Figure 10-29.-Typical cluster reliability predic-

tion as a lunction ol number ot engines in
cluster and of development time.
engine-out and no engine-out narrows (see also

table 2-2). It may be stated that for a given
number of vehicle flights, the benefits of engine-
out capability are the greater, the lower the
single-engine reliability. Figure 10-28, in com-
¢. _L
bination with figure 10-29, shows that simulta-
neous addition of one engine and inclusion of Figure lO-30.-Typical engine cluster arrange-
engine-out capability retain about the same pay- ments.
5-Engine Cluster and systems thrust vector control response is

This number provides greater thrust, but com- seriously affected, due to roll and yaw coupling,
plicates matters in other respects. Two basic particularly under malfunction conditions. As to
arrangements .are possible, as shown in figure mounting-circle diameter, the two 5-engine clus-
10-30(b) and (c). The arrangement with center ters are about equal.
engine is preferred. The center engine is mounted
fixed; i.e., stiff arms are installed in place of 6-Engine Cluster
gimbal actuators. It is advisable to install the This combination provides the highest thrust,
engine with complete gimbal bearing for stand- at the expense of a larger mounting circle. This
ardization and ease of alinement. With the re- in turn increases the effects of one-engine-out
maining four engines gimbaled and grouped as in on trim required and may compound interstage
a 4-engine cluster, the same benefits are derived and interengine clearance problems. The control
as with the latter. However, because of the mode is comparable to the four engine cluster in
center engine, two installations are required. simplicity. The 6-engine cluster requires two
Furthermore, two sets of inlet ducts are needed, installations, with inlet and other effects similar
their different length affecting pressure drop and to those discussed with the 5-engine configura-
fluid velocity profile (engine turbopump NPSH), tion.
water hammer from closure of valves (valve tim- This cluster offers the potential of removing
ing), trapped propellants, and possibly insula- (not installing) 2 engines and still retaining a
tion requirements (weight penalties and complex- satisfactory 4-engine combination, which pro-
ity). Figure 10-31 shows a typical 5-engine vides added flexibility at minimum scar weight
installation with fixed center engine. (i.e., weight of components that cannot readily
With all five engines arranged on a circle, the
be removed along with the engines).
basic engine package may become more uniform, If at all possible, the engine designer should
since all will be gimbaled. However, three dif- specify that the engine intended for cluster use
ferent installations are required (fig. 10-30(c)), be tested under conditions closely simulating
vehicle installation. It may be expected that the
vehicle builder will conduct a firing program of
his own; however, this must be devoted to clus-
ter behavior and performance evaluation, and
should not deteriorate into continued single en-
gine development. A major difference between
single engine and cluster firing is in the base
conditions, i.e., the heat and pressure environ-
ment of the engines, particularly at altitude.
Flame radiation effects, backwash of combustion
gases, and impingement may create much more
severe conditions than are present during single-
engine firing. The pressure environment pro-
duced by several engines firing together may
create moments on the engines which must be
accounted for in the design of the gimbal sys-
tem. The engine designer should be familiar
with these conditions so that his claim that the
engine can be clustered remains valid after
delivery.
ENGINE Awareness by the engine designer of the con-
{FIXED) siderations governing engine clustering, as pre-
Figure ]O-3].-Typica] five-engine cluster con- sented here, will enable him to complete his
figuration. Center engine fixed, four outer systems integration with a broader view to
engines gimbaled. application.
10.9 ENGINE-TO-VEHICLE INTERFACE vehicle builder is in need of numerous engine

design details in support of his stage design.
Throughout this book, references have been
An excellent source for this infonnation is an
made to the ultimate purpose of rocket engine
"Engine Design Manual." It should be started
design and production: propulsion of a vehicle.
during engine initial design and should be aug-
Some of the principal engine-to-vehicle inter-
mented as the design and, at a later date, the
faces, such as thrust mount, and pneumatic,
development progresses. Necessary revisions
fluid, electrical and propellant lines, have been
must be disseminated promptly.
discussed in preceding chapters. In this section,
Equally important is the early generation of
we will summarize a number of vital engine-to-
an Interface Control Document. This defines
vehicle interfaces, of which the engine designer
each interface, for both mating sides, on one
should also be cognizant.
drawing.
Design Documentation Space Envelope
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-
Thrust or gimbal mount bility of faulty deflection of engines must be
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.
(if any), fasteners, and loads at the

vehicle connect points listed
Connect Panels
(3) Engine system mass properties, as are
shown in figure 2-6, which include en- All engine-to-vehicle lines, which may amount
gine weights, gimbaled mass, center of to a dozen or more per er_ ine, must be reliably
gravity, and moment of inertia for the connected for each engine installation, not only
basic engine, including accessories in the vehicle but in the static development and
(4) Engine performance data (as in fig. 3-1) acceptance firing stands as well. It has been
(5) Engine functional description (as in ch. found beneficial to define this interface clearly
III) by combining all lines in one or several terminal
(6) Engine handling procedures and equip- connecting panels (fig. 10-32). These panels,
ment, needed for installation and main- uniformly designed for all test locations, may be
tenance mounted on the vehicle (or test stand), or on the
(7) Engine servicing needs engine. For certain installations it may be ad-
As a rule, this information is compiled in vantageous to have matching panels on both:
detailed handbooks, which will accompany the vehicle (or stand) and engine. Alined a short
engines when delivered to the vehicle contractor. distance apart, standardized jumper lines be-
However, several years before the engines are tween pairs of panels will permit rapid and reli-
delivered and handbooks become available, the able connections. A possible disadvantage of
TEST STAND
,ul
STRUCTURE TA,
ENGINE CONNECT
Lax IEL_O_--
PANEL .o= pu_ -- - c, _, _,.o_
Lox ,_c,c_,-
Figure 10-33. -Closed-loop coupling of propulsion

system parameters and longitudinal vehicle
*: ,= 'q
mode.
acceleration, as a function of thrust, a powerful

feedback exists. It has been demonstrated that
the system can be detuned or uncoupled, through
manipulation of propellant duct volumes, stiffen-
Figure lO-32.-Typical line connections on an
ing of structural members, and by other means.
experimental liquid rocket engine.
Itis difficultif not impossible for the engine
designer to predict analytically and prevent this
the panel method is that routing a line through problem, particularly since the vehicle configu-
the panels may result in additional line length, ration may not yet be firm. He may be called
as compared to individual routing of each line. upon later for changes, however, if the problem
Through careful design and close coordination shows tipduring flight.
between engine and vehicle builder, considering
optimum location, subdivision and orientation of
Malfunction Reaction Systems
the panels, advantages will, in most cases, far
outweigh disadvantages. In chapter II we have discussed reliability
and failure modes, as well as certain preventive
measures. In the framework of engine-to-vehicle
Dynamic Interactions
interfaces, some additional detail follows.
The close coupling between propellant feed

system, propulsion system, and vehicle structure Engine Failure Sensing and Shutoff System
may lead to undesired interactions. These have (EFSS)
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.
and early cluster development. Part of it may Heat Protection

later become a portion of flight emergency detec-
Rarely in our technology had lowest and high-
tion systems (EDS) and engine-out systems (see
est temperatures to be handled so close together
ch. II). The following list of major engine-
as in a liquid propellant rocket engine. Earlier
oriented malfunctions, which is not necessarily
we discussed the need for insulation, to main-
complete, should be considered:
tain temperatures as low as -423 _ F in certain
ducts, or for heaters to protect sensitive compo-
Pump inlet pressures below safe minimum
nents against these temperatures. At the same
Turbopump overspeed
time, in other areas at or around the engine, pro-
Turbopump bearing overheating
tection must be provided against the very high
Excessive turbopump leaks
temperatures of the combustion process and the
Turbine gas overtemperature
emerging gas jet, such as cooling of the thrust
Combustion instability
chamber. The exhaust jet, at sea level, usually
Abnormal injection pressures
Ignition failure is not a major problem, unless blowback occurs
from the flame deflector. At higher altitudes
Premature propellant depletion
with vacuum or near-vacuum pressures, however,
Electrical power failure
Pneumatic pressure failure which are experienced even by first stages for
the last portion of their flight, a substantial
Improper valve positions
Fires portion of the thrust chamber gas jet expands
sideways from the nozzle exit, forming a plume,
Because of their potential sudden destructive
creating considerable backwash and radiating
effects, many of the malfunctions will be sensed powerfully back into the engine compartment.
and the signals used to initiate immediate auto- This endangers both engine and vehicle compo-
matic cutoff. For others which would not create
nents. Surface temperatures of 1000 ° F or more
an emergency within, say, fractions of a second, may result unless heat protection is provided. In
it is not uncommon to simply mark their allow- some cases it may be too cumbersome to provide
able minimum and/or maximum values on an ink-
individual insulation for each component. Also,
ing recorder chart which is watched by an ob- excessive weight penalties may be incurred. A
server during test. If the recorder needle goes protective heat shield, forming a closeout dia-
outside of the marked margin, the observer initi- phragm, may then be more effective (fig. 10-35).
ates cutoff manually. This shield may be supported from a stationary
For flight, a few selected highly critical (center) engine, if available, or from a supporting
parameters of the EFSS may be retained for auto- structure.
matic engine cutoff or mission abortion. In this
case, if the engine failure sensing and shutoff
system were called upon to shut down an engine,
and the flight is continued with the remaining
engines, a vehicle-mounted electronic logic must
sense the shutdown and take certain actions.
These may include:
A
Closure of emergency shutoff valves in the
FLEX
ducts to the defunct engine, but not those in CURTAIN _
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)
will consume the available propellants over

a longer period of time. Figure l O-35.-Typical base beat protection con-
Locking of the defunct engine in the neutral cepts (center engine [ixed, outer engines
gimbal position. gimbaled).
Design and installation of the base heat

shield requires closest cooperation between
engine and vehicle designer. It must be deter-
mined, through special model tests, whether a CHANh
partial or a full closeout is required. In most

designs the shield connects to the engine noz-
zle. Suitable brackets must therefore be pro-
REFRACTORY
vided. During gimbaling the heat shield will SHI
resist the engine motion. The forcesencountered
must be considered in the power budget for the CHAMBER
hydraulic actuators. Here it must be taken into TUB ES
account that the heat shield flex curtains may be

INSULATI(
quite cold prior to engine start and still remain
relatively cool at the far side, as planned, during
stage operation.
In addition to backwash and radiation, which
are generally not harmful to the internally cooled
engine nozzle itself, mutual gas jet impingement
CHA_
between engines may occur from extreme deflec-
tion during gimbaling. Unless a major control
malfunction occurred, this should affect the noz-
zle only for very brief periods near the exit.
However, the heat may affect chamber structural
elements, such as stiffening bands, which are EXIT RING AND FLANGE
not internally cooled. A heat-protective strip, a
few inches wide, of ablative or other suitable Figure JO-36.-Center-engine flame impingement
material applied to the nozzle should suffice in shield.
most cases (fig. 10-36). Obviously, this type of
protection is not required for single-engine
vehicles. lion mechanisms, can function properly only if
they are within a specified temperature range.
Engine systems not employing cryogenic propel-
Engine Prestart Conditioning and Start
lants may be dependent on heating or cooling
Functional conditions required within the only if prolonged coasting times in space are
engine to assure its readiness to start have been involved. Engine systems which do use cryo-
discussed earlier. When installed in the vehicle, genic propellants almost always need at least
the engine also requires external conditions some heating. The cryogenic propellants within
which must be met by the vehicle builder. Pro- the engine following start of tanking, and the
vided that prelaunch checkouts have ascertained heat absorption of tank surfaces and lines may
readiness with regard to absence of leaks, cor- rapidly lower the air temperature surrounding the
rect valve positions, etc., prestart conditioning engine to several hundred degrees below 0 ° F.
of the engine essentially refers to temperature Since most vehicle systems will specify an
and pressure levels around the engine and at the allowable hold period following tanking to allow
pump inlets. For both, early cooperation be- for adjustments, checkout of other systems, to
tween engine and vehicle designer to arrive at a wait for optimum launch times (rendezvous mis-
mutually feasible solution is essential. Without sions), etc., severe subcooling of engine compo-
it, the optimistic note in engine drawings "to be nents may occur.
supplied by vehicle contractor" will accomplish The temperature environment can be substan-
little. tially improved by the vehicle builder through
Certain engine subsystems, such as hydraulic engine compartment purges with warm gases (fig.
components, control systems, and valve actua- 10-37). If an inert gas is used, this has the
80
80
4C _ _- SKIN BELOW THRUST MOUNT /--SKIN BELOW THRUST MOUNT
40 WL,._--SKIN At3OVE 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
-160 _ ENGINE TEMPERATURE _. "N_'-THRUST MOUNT TEi'IRERA'FURE
i(
-200 _ I_THRUST MOUNT TEMPERATURE _NITROGEN AMBENT INTERNAL TO MOUNT
m_-NITROGEN AMBIENT INTERNAL TO MOUNT

-16( I I I l l I
i I i 2 6 I0 12
-2,o. , 2 , _, , _ , _, , ,_ ,2 4 8
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
peratures because of their strength characteris-

tics. Another limitation is established by the
purge gas requirements and necessary heating _ (LIQUtO OXYGEN ] /
provisions which would become prohibitive if it

were attempted to raise compartment tempera-
tures to above 32 ° F or even to 0 ° F. Figure INTER-
STAGE
10-38 shows the effects of increased purge flow
and temperatures over those of figure 10-37. The
analyses on which the graphs are based have
shown that a further increase of the purge rate to
12 000 SCFM (standard cubic feet per minute)and l ..... _ \ /ENG,
NE\ / ENO,
NE
250 ° F barely raises the interstage temperature
above 0 ° F.
At the same time the propellant boiloff rates
in exposed ducts increased tenfold over those
under conditions of figure 10-37. The locations
quoted are identified in figure 10-39.
For most applications the purges to the vari-
ous areas are supplied from a ground source.
Once preconditioned, affected components can Figure 10-39.-Multistage vehicle interstage.
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.
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.
of serious problems. 10-40). If the clearance between engine skirts

and interstage wall is marginal, complicated
Nozzle Expansion Area Ratio means may be required to avoid collisions be-
tween the separating stages. These may include
From his own analyses, the engine designer
a control system to tuck the engines inward, and
may have selected a nozzle expansion area ratio
then swing them out upon separation and start
which appears to be an optimum compromise be-
them. Any such system adds complexity, lowers
tween engine weight, engine performance, engine
reliability,and may add weight. A shorter noz-
size, and producibility. When analyzed in the
zle skirt, of lesser diameter although somewhat
larger framework of vehicle performance, the
lesser performance, may be better overall.
selection may turn out not to be optimum. In
2. Leaving the interstage with the upper
reviewing the various influences, we will disre-
stage.-To avoid the problem of stage collisions,
gard single-engine firststages which are least
it may appear attractive to leave the interstage
affected. For stages with engine clusters, in
attached to tileupper stage. However, this adds
particular, upper stages, the following must be
inert weight to the upper stage. This penalty
considered:
may be considerably larger than the performance
loss from a somewhat shorter nozzle. Also, a
Gimbal Angle
control problem may be incurred in that the en-
Following determination of the maximum gim-
gine nozzles, in their outward gimbal deflection,
bal angle required for safe vehicle guidance,
stillmust clear the interstage wall. In joint
including consideration of actuation malfunc-
optimization studies between engine and vehicle
tions, the nozzle exit diameter will determine
designer, the possibility of leaving part of the
the attachment point distances (mounting pattern)
interstage with either stage may be considered.
of the engines in a cluster.
3. Leaving the interstagewith the upper stage
but dropping it in a second separation maneuver
Vehicle Diameter
several seconds a/ter first separation.-In this
When mounted for proper gimbal capability,
dual-separation sequence, the second separation
the envelope of the engine cluster must be in
may consist of shedding the interstage as a com-
proper relationship to the vehicle diameter. If
plete ring, possibly in combination with guide
the cluster envelope is too large, the increase in
rails, or of blowing it off to the side in segments.
vehicle dimensions may void all gains from a
For either method, some form of actuation is
larger expansion area ratio. If engines are not
required in addition to accurate timing. Also,
permitted to protrude beyond the projected vehicle
while the interstage is still attached, a serious
plane in neutral and/or in gimbaled position, an
base-heating problem may develop. The means
even more stringent situation exists.
required to overcome any difficulties with this
separation method again may increase complexity
Interstage Length
and reduce reliability and partially neutralize
When mounted to the lower stage, the length
nozzle performance gains.
of the upper stage engines will determine the
length of the required interstage. For a vehicle Whichever method is applied, the thrust-decay
of the Saturn V class, each additional inch of characteristics of the spent lower stage engines
Shutdown of lq_er stage engines (from propellant depletmn signal)
I 1
Upper stage ullage rockets ignited
I I
Engine start sequence initiated
I
Separation timer started
Physical separation begins

(from timer or accelerometel
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)
Figure 10-40.-Typical stage-separation sequence.
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.
too
\
\
\\ \
\ \
\ \
7s \
\ \
\ \
ILl \
I&.
so
o [ t I
0.S 0.6 0.7
CUTOFF TI_E SEC ONDS
SIGNAL
Figure 10-41 .-Typical engine thrust decay deviations.
Delivery Schedules Maintenance and Logistics
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
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
Design of Liquid Propellant Space Engines
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--
an axis. They are mounted in clusters to sim-

plify plumbing and wiring. Parallel pairs of
engines, or individual units, are used for trans-
lational movements along the vehicle axis, as in tNS'_U_NT "9 _
rendezvous and docking maneuvers. -- Off EN,_,kWE _ _EACT,O_
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 $ :
controllable. The thrust chambers may be abla-

tively cooled or radiation cooled, depending on
the application.
L___@_A_
Application Example
Figure 11-1.-The propulsion systems of the
The application of various space engine sys- Saturn V [or the U.S. manned Apollo flight to
tems may best be illustrated by typical exam- _he Moon program.
ples, such as the spacecraft for the U.S. Apollo
Moon-landing program. The space vehicle sys- The command module requires thrust only for
tem to carry the Apollo aloft, the Saturn V, con- three-axis attitude control during reentry into the
sists of the S-1C (five F-1 LO2/RP-1 engines- Earth's atmosphere and to assist the service
total thrust, 7500000 pounds), the S-If (five J-2 module's reaction control system during separa-
LO2/LH 2 engines-total thrust, 1.15 million tion of the two modules. The command module
pounds), and the S-IVB (one J-2 LO2/LH 2 reaction control system is made up of two sepa-
engine-thrust 230000 pounds) stages (fig. 11-1). rate, interconnected pressure feed N20_/MMH
The Apollo spacecraft itself consists of the com- engine systems, either of which is capable of
mand, the service, and the lunar excursion completing the Earth reentry mission. These
modules. engine systems are located in the aft equipment
The service module has a 21 900-pound, fixed- compartments. Each system consists of six
thrust restartable pressure feed N_O4/50 (UDMH)- pulse-modulated, 91-pound, thrust, chamber
50 (N2H4) engine as its main propulsion system. assemblies of a 9-to-1 nozzle expansion area
Its thrust chamber has a nozzle expansion area ratio.
ratio of 60 to 1 and is gimbaled for thrust vector The lunar excursion module (LEM) has two
control by a redundant electromechanical device. main engine systems, one for the descent to the
This engine system provides the thrust for mid- Moon's surface, and the other for takeoff fromthe
course correction as well as for entry into, and Moon and injection into lunar orbit for rendez-
escape from, lunar orbit. It also serves as an vous with the command and service modules.
abort system, if necessary, after the launch The LEM descent engine system has a nominal
escape system has been jettisoned. thrust level of about 10 500 pounds, which can
The service module's reaction control system be gimbaled and throttled over a 10:1 range. The
is composed of four separate and interconnected takeoff engine is a rigidly mounted, 3500-pound,
pressure-feed engine systems utilizing the same fixed-thrust engine, which doubles as an abort
propellants as the main propulsion system. The engine in the early portion of the mission. The
systems are mounted external to the forward skin LEM also carries an attitude control system
of the service module at 90 ° intervals. Each which uses 16 thrust-chamber assemblies. All
engine system consists of four pulse-modulated, are pressure fed, using a N20_/50 (UDMH)-
100-pound thrust chamber assemblies with a 50 (N20 _) propellant combination.
40-to-1 nozzle expansion area ratio. These pro-
vide the three-axis attitude control for the space-
11.2 GENERAL DESIGN CONSIDERATIONS
craft during Earth orbit, lunar injection, lunar
orbit, LEM docking and separation from the Prior to the actual design of a space en-
command module. gine system, it is necessary to examine the
DESIGN OF LIQUID PROPELLANT SPACE ENGINES 431
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.
5. Cooling and other characteristics.-Some

propellants are excellent coolants (for regenera-
tive or film cooling), while others have little
cooling capability. The propellant combination
will determine the combustion temperature and -- FI/_ z PuuP
the gas constituents, which vary widely as to

their compatibility with chamber materials, espe- _ Fem_N , px[ssu_£
cially with ablative or refractory materials which %_1%_4 PUllS
are frequently used in space engines. r WEEK _ kq:_TM $ blONT_tS :AR I0 YEA_
If possible, all propulsion systems in a space- 1.000 i0,O00 t_.0¢¢
craft should use the same propellant combina- COAST I_UI_TIO_, _*OUl_
tion. However, in some applications the reaction

Figure ll-2.-EHec$ of mission coast duration on
control engines are also used for propellant set-
relative payload capabilities o[ various propel-
tling in the main system tanks. Thus they must
1ant engine systems for a given mission veloc-
be supplied by separate tanks equipped with
ity increment.
positive expulsion devices. Different propel-
lants may have to be used in this case. Charac-
teristics of various propellants suitable for Vehicle System Operational Requirements
space engine applications are summarized in
table 11-1. Characteristic trends of relative After the vehicle system operational require-
payload capabilities versus mission coast dura- ments have been analytically established for a
tion, for various propellants and engine systems, given space mission, many design parameters
are shown in figure 11-2. It can be seen that for and special considerations for the engine system
coast periods up to about 4300 hours (6 months), or systems can be derived. Principal areas of
there is no significant change in payload capa- consideration, in addition to propellant selection,
bility. are:
TABLE ll-1.-Comparison of Various Liquid Propellant Combinations [or Space Engine Applications
aSpecific Density Freezing and boiling points C'F, Mixture Combustion

Propellant oxidizer/fuel ratio, temperature, Remarks
impulse, impulse,
combination
seconds sec-gm/cc (FP)(BP)/(FP)(BP) O/F °F
241 (-364)(- 307)/(-435X-423) 9.3 6440 Hypergolic

_2/H2 ............... 474
456 143 (-362)(-298)/(-435)(-423) 4.7 5110 Nonhypergolic
_O2/H 2 ..............
424 7010 Hypergolio
3F2/B2H6 ........... 430 (-371)(-299)/(-_65X- 135) 3.5
357 (- 371X-299)/(-265)(- 135) 2.15 5910 Hypergolic
3F2/B2H 6 ........... 420
419 551 (-364)(-307)/(35)(236) 2.4 7285 Hypergolic
F'2/N2H 4 .............
417 451 (-371)(-299)/(-300)(-260) 5.6 670O Nonhypergolic
3F2/CH 4 ............
408 303 (-362)(-298)/(-265)(-135) 2.0 5960 Nonhypergolic
32/B2H 6 .............
402 254 (35)(9.36)/(- 265)(- 135) 1. 4085 Nonhypergolic
_2H4/B2H_ ..........
375 340 (11)(70)/(-265)(- 135) 2.9 5710 Nonhypergolic
_20,/B2H6 ..........
341 409 (11)(70)/(35)(9.36) 1.23 5513 Hypergolic
_N:O_/N2H _ .........
_MON/EMHF ......... 341 407 (-23)(29)/(-76X144) 2.2 5330 Hypergolic
bMON/MMH ........... 338 401 (-9.3)(29)/(-63)(189) 2.4 5370 Hypergolic
bN_o4/50-50 ......... 339 408 (11)(70)/(18)(170) 2.1 5175 Hypergolic
bN_O4/MMH . 339 407 (11)(70)/(-63)(189) 2.3 5290 Hypergolic
Notes: MON =Mixed oxides of nitrogen, 85% N204-15% NO.

MMH =Monomethyhydrazine, CH_ N_H_, 50-50= 50% UDMH. (H_C)_N_H_-50% N:H_.
EMHF = Eutectic mixture of hydrazine fuels = 87.6% MMH-19..4% N_H,.
abased upon theoretical shifting equilibrium at 150-psia nozzle stagnation chamber pressure and 40:1 nozzle expan-
sion area ratio in the vacuum.
bpropellant technology and application are well established.
.,a======
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)
i 10,000 / VARIABLE THRUST,

TIME
VARIABLE
(RENDEZVOUS)
IMPULSE
._ .CIO X PU_LSE ROCKET--

(ATTITUDE
TIME
DISCRETE
CONTROL)
IMPULSE BITS
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. )
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).
11.3 DESIGN OF SPACECRAFT MAIN tioncould be providedby severalindividual

PROPULSION SYSTEMS pressuranttanks of differentsize,each of which
remains sealed by explosive-actuatedvalves
System Design (fig.7-72) untilneeded. A sealed pressurization
system would not requiregas venting during any
The reliability requirementsforspace mis-
partof the mission, thus eliminatinglosses due
sions have led to simple, storedinertgas pres-
to intentionalventing,or because of leakage
sure feed main propulsion systems (fig.5-1),
past the vent valves.
using conventionalhypergolic Earth-storable
In some applications,the propulsion system
propellantsand redundantcontrolcomponents.
can be operated on a propellanttank ullage gas
For tube connections, the trend is to welded or
blowdown principleduring the latterportionof
brazed joints, with flexibility provided either by
the mission. This could reduce the required
braided metal hoses or metal bellows. Integrated
amount of pressurant (and pressurant tank vol-
engine packages are preferred, consisting of
ume) by as much as one-half of that for a regu-
thrust chamber and injector assembly, integrated
lated, constant pressure system.
propellant control valve package, and flight
instrumentation. Besides meeting thrust and Ignition of the main thrust chamber is critical
performance specifications, principal system with any multiple-start space propulsion system.
design requirements for a spacecraft main engine If hypergolic propellants are used, ignition de-
are: vices such as spark plugs of ASI units are not
needed. However, hypergolics suffer from igni-
(1) High reliability and crew safety (through
tion delays at low temperatures (such as are
redundancy in the control system; use
encountered in space). Under the peculiar con-
of proven design concepts, materials,
ditions of vacuum starting, irregular transients
and fabrication techniques; and ex-
may occur upon ignition, such as violent chamber
tended life and overstress testing)
pressure surges. These may not be destructive,
(2) High combustion stability rating with
but they could place a severe burden on the
respect to perturbations during start
control system.
and throttle transients
For nonhypergolic propellants, spark plugs or
(3) Capability of numerous starts, with re-
augmented spark igniters (ASI) may be used.
peatable cutoff impulse
Spark plugs have the disadvantage of corona-
(4) Capability of deep throttling (optional,
arcing and electrical leakage, even in a partial
not required by all missions), with
vacuum. However, injection of unburned propel-
simple controls lants into the thrust chamber or ASI with attend-
(5) Capability of propellant utilization and
ant vaporization usually builds up sufficient
tank blowdown (ullage expansion)
chamber back pressure to allow the plugs to
(6) One signal each, for start and cutoff
function. In some systems, this process may be
(7) No purge, bleed, or lubrication equipment
augmented by purge gases. Certain propellant
required
combinations tend to ionize the gas around the
(8) Effective isolation of the propellants
plug and thus quench the spark.
from the system during long coast peri- Various schemes have been considered and
ods (use of propellant isolation valves)
developed for the deep throttling often required
(9) System growth potential and flexibility
by spacecraftmain propulsion systems. For
to mission modifications
instance,the Apollo LEM descent engine is
(10) Sufficient protection against the effects
requiredto vary thrustcontinuously over a range
of heat, nuclear radiation, and meteoric
from 1050 to 10500 pounds to permit hovering,
bombardment.
selectionof a landing site,and the landing
For missions involving intermittent propulsion itself.Two alternateapproaches to throttling
system operation, during periods of more than a are being pursued. One calls forvariationof the
few months, the pressurization system design thrustlevel by main propellantline throttle
calls for special provisions, such as a com- valves,combined with aerationof the propellants
pletely sealed pressurant system. Pressuriza- with helium gas, as shown in figure4-47(b).
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
=__-'_'_ CHECK VALVE
=_ FtLTE_ _w{} PRES_E

REGULATOR
._,- PROPELLINT FILTER
Figure ll-5.-$chematic of a typical pressure- Figure ll-6.-Package design layou( ol a $ypical

feed spacecraft main propulsion system using pressure-feed spacecraft main propulsion sys-
hypergolic Earth-storable propellants. tem using hypergolic Ear_h-s_orable propellan$s.
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,
out of a hypothetical turbopump-feed spacecraft with due consideration of application in
main engine system using a LO2/LH2 or space

(2) Ability to withstand postrun soakback,
intermittent operation, and storage in the
space environment
(3) Ability to enter planetary atmospheres at
OX;D;ZERTUR_O_U_
F_:!__/_ _AO_AT'ON-COOLEO high velocities
NOZZLE EXTENSION SKIRT
GI M_L MOUN_ _- "_ Many of the thrust-chamber-cooling techniques

described in chapter IV are applicable to space
engines. Especially the ablatively cooled thrust
chambers, which offer the inherent advantage of
simplicity, ruggedness, and independence of
propellant coolant flows (including their pressure
EXHAUST I
drops), and thus offset their limited total life and
oucr i. 44 • - 75 relatively heavy weight, are suitable for most
/ -- THRUST
/ C HAMIER pressure-feed space engine systems using Earth-

storable propellants at chamber pressures of
/ L HOT-,GAS ,o::'Z'¢:
THROTTLE VALVE
less than 150 psia. In some applications, abla-
/-'--- FUEL TURI, OPUMP
tive cooling is supplemented by one or a com-

Figure ll-7.-Design layout of a hypothetical bination of the following techniques:
turbopump-feed spacecraft main engine system (i) Refractory throat inserts
using a LOJLtt 2 or LFJLtt 2 propellant com- (2) Radiation-cooled nozzle extension skirts
bination. (3) Propellant film cooling
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
ings should not affect total chamber duration.

Equations (4-36) and (4-37) may be used to cal-
culate the char depth. For Refrasil phenolic
;,, i',.'_ ,_,

.............. _ iS.- j ablative thrust chambers operated at pressures
around 100-125 psia, and using N204 and hydra-
zine type fuels, the following empirical equa-
tions may also be used to estimate char depth,
including the effects of charring due to soakback:
(1) For the combustion zone and the throat

section (with or without liner and/or
insert):
I
a=0.0415t °-s (11-1)

Figure l l-8.-Typical thrust chamber design de-
tails [or a pressure-[eed spacecraIt main pro-
(2) For nozzle sections downstream of the
pulsion system using bypergolic Earth-storable
throat:
propellants.
a = 0.0335 t as e -°024;e (11-2)
Earth-storable propellants at a chamber pressure

of 125psia. Refrasil-filled, phenolic-impregnated where
ablative materials are tape wrapped on a man- a = char depth, in
drel, at the combustion chamber and nozzle sec- t --thrust chamber cumulative firing duration,
tions, up to an expansion ratio of 42. A JTA sec
liner and a silicon carbide insert are provided e = base of natural logarithms. 2.718
for combustion chamber section and throat, =nozzle expansion area ratio at the section
respectively. The chamber outer wall is made of under consideration
fiber-glass filament, wound to accept hoop Once the char depth based on design duration,
stresses, and then wrapped longitudinally (heli- including effects of soakback, is determined, the
cally) to contend with meridional stresses. An thickness of the insulating layer required to keep
insulating layer, consisting of the basic or the outer wall at the required temperature can be
another ablative material such as asbestos or calculated.
alumina silica, is wrapped around the basic For turbopump-feed space engine systems
ablative layer and cured to reduce the rate of operated at relatively high chamber pressures,
heat flux to the structural fiber-glass outer wall. regeneratively cooled tubular-wall thrust cham-
A radiation-cooled skirt made of 0.030-inch bers appear more suitable from heat transfer
Hastelloy-X sheet is tape wrapped to the thrust considerations. A possible disadvantage for
chamber nozzle to provide the nozzle extension space engine application is the response time
from an expansion ratio of 42:1 to 80:1. An inte- from start signal to full thrust, which may be
grated injector is attached to the end of the com- substantially increased if the propellant valve is
bustion chamber and is made of an aluminum- located upstream of the coolant jacket. Relocat-
alloy forging. It is a conventional, fixed-orifice, ing it to a point downstream of the jacket would
single-ring, unlike-impinging doublet type. reduce the response problem, but would introduce
Ablative thrust chambers for space engines, heat-_ransfer problems after shutdown, with the
such as shown in figure 11-8, should be designed propellants trapped in the jacket.
to permit many restarts. At the end of a speci- Other potential problems are the heat-transfer
fied total firing duration (design values: 400- characteristics during throttling. Also, tubular
2000 sec), sufficient insulation material must walls are more susceptible to meteoritic damage
remain between char and fiber-glass shell to than are solid walls. In some applications, an
limit the outer wall temperature to a specified optimum overall design may result from combin-
maximum value (design values: 400°-700 ° F). ing regenerative cooling with film cooling, and
Any heat soakback during ofttimes between fir- adding a radiation-cooled nozzle extension skirt.
Design of Control Components seals, and hermetically sealed solenoids or

torque motors.
The various aspects of control component
The valve is spring loaded normally closed,
design have been discussed in chapter VII.
and additionally pressure unbalanced closed
Pertinent considerations for the design of space
when inlet pressure is applied. Actuation to
main engine control components, in general and
open is achieved by allowing the oxidizer fluid
for propellant valves in particular, are:
to enter the actuation chamber at nominal oxi-
(1) Maximum reliability
dizer valve inlet pressure which, in turn, over-
(2) Use of bellows for dynamic sealing in
comes the closing forces on the poppets. For
view of the need for long-term operating
on-off and isolation valve applications, the full
lifein a vacuum environment
open or closed valve positions are effected by a
(3) Combination hard and softvalve seats,
three-way solenoid pilot valve which, when ener-
forminimum leakage
gized, pressurizes the valve actuation chamber
(4) Provisionof a minimum of two seals and
to open, and which, when deenergized, vents the
a vent between different propellants
same chamber allowing the pressure unbalance
(5) Avoidance of slidingsurfacesin compo-
and the springs to close the poppets.
nents operatingin a vacuum environment
In the applications as a dual-propellant throt-
(6) Where possible,mechanical linkage of the
tling valve, the position of the linked main pop-
propellantvalve actuatorsby a mecha-
pets is controlled by a servo pilot spool valve
nism thatensures positive,consistent,
which meters the control fluid flow (oxidizer) to
and synchronized opening and closing
the main valve actuation chamber. The pilot
forsmooth, repeatableignitionand
spool valve, in turn, is proportionally positioned
thrusttermination
by a servo torque motor, as a function of the
(7) Minimum electrical
energy requirementfor
command signalcurrentand itspolaritysupplied
actuations
from a servo amplifier.By means of properly
The basic design detailsforthe various con- shaped valve poppet contours,the main throt-
tlingvalves controlthe flow of oxidizerand fuel
trolcomponents forspace main engine systems,
to the main thrustchamber at constantmixture
such as pilotvalves,regulators,and ventvalves,
ratio,and thus regulatethe main chamber pres-
are quite similarto those forbooster-stageappli-
sure (thrust).
cations. The mechanically linked,poppet-type
dual-propellantvalve arrangement shown in
figureIi-9 is typicalformain-propellantor
System Design of Propellant Storage for Space
propellant-isolationvalves of space main engines
Missions
using hypergolicEarth-storablepropellants.In
the valve shown, linked actuationis accom- As previously mentioned, the storage of pro-
plished by locatingthe valve poppets in series, pellants for space-system propulsion represents
thus causing the oxidizerpoppet motion to simul- an important design area. Cryogenic propellants
taneously move the fuelpoppet. Complete sepa- require protection against high temperatures,
rationof fuel and oxidizeris achieved by welded particularly for extended storage periods. Earth-
bellows seals locatedat both ends of the fuel storable propellants, by contrast, require protec-
poppet and at the connection end of the oxidizer tion against low temperatures, because of the
poppet. Since valve actuationis provided by danger of propellant freezing. There are two
oxidizerfluidpressure,no bellows is requiredat basic design approaches to propellant storage
the actuationchamber. The center vent chamber for space missions, the vented and the nonvented
between valves sectionsis the criticalarea systems. The selection of the optimum system
where propellantmixing must be prevented. design depends on mission, propellants type,
However, both bellows would have to failto and type of engine feed system.
cause this failuremode (reliability throughre- In a yen-ted storage system, propellant pres-
dundancy). The design also incorporatesall- sure and temperature are maintained at a con-
welded static sealing joints, all-metal construc- stant value. This system is mainly used for
tion except for the Teflon dynamic poppet valve cryogenic propellants. Any heat input to the
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
SOLENOID_ II] /\ _OXIDIZER OUTLET /_ FUEL OUTLET
I_ _-t_" _Y__NNDSEE_L_ VE %NT_/ TEF(LBONDSEEA_

/--'---WELD
ox,o,z
ACTUATING
o.-O FV W
FLOW
VENT
f SERVO
MOTOR TORQUE-- 7 _
. A
WELD'_ ACTUATINGPISTON_ /__?-_
_',_ I I/ //// t \\\X'I

-- A_..:._ --%'j_ONTOURED(TYPICAL) POPPET
T__ I_ _j \ "IROTTLING VALVE ONLY
OXID
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.
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
temperature, expected changes in propellant disturbing torques. It could be of either a con-
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.
implies rotation and/or stabilization of the vehi-

cle about these axes. Its operation consists of Operational Modes of Reaction Control Systems
two main phases:
If any reaction control system, the amount of
(1) Rotation of the vehicle over a given angle, thrust delivered on command must be very pre-
within a specified time cisely controlled, if overshoot or undershoot and
(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
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
_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------X ! _ _ / ' PRESSURE

i, TRANSDUCER
VALVE ',_ I¢I I

l _ I CONTROL
MODULAR CONTROL__
._ FUEL TANK
PACKAGE _ _ __._
SOLENOID ISOLATION_----_ •.._-----_-- ,,mI.---MODULAR
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
PITCH YAW ROLL
[] MICRONIC FILTER _ BURST DIAPHRAGM ( RELIEF VALVE
@ SERVICE CONNECTION
Figure ll-lO.-Basic schematic ofa typical reaction control system.

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.
Design of Reaction Control Thrust Chambers
For systems using hypergolic Earth-storable

bipropellants, the reaction control thrust cham-
bers may be effectively cooled by one of the
following methods:
(1) Ablative cooling
(2) Radiation cooling
(3) Ablative cooling with radiation-cooled
nozzle skirt
(4) Ablative cooling supplemented by regen-
erative or film cooling
Other methods, such as regenerative cooling and.
Figure 11-11.-Reaction control engine systems film cooling, can probably be used successfully
for the Gemini manned spacecraft. to a certain extent. However, none can compete
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.
eratively cooled jacket is provided for extra-long

durations (up to 30 minutes). Fuel is used as
the regenerative coolant to supplement the abla-
tive cooling. A segmented 3TA graphite liner is
used in the combustion chamber section; the
throat insert is made of silicon carbide. Abla-
tive materials are oriented at 90 ° and 0 °,respec-
tively, at various regions according to structural
requirements. Asbestos-phenolic is used for the
outer insulation. The injector propellant valves
TIME, SECONDS are brazed to the injector, which forms an inte-
gral part of the thrust chamber assembly.
Figure II-I2.-Combustion chamber char depth Figure 11-14 presents the 25-pound thrust
versus cumulative firing time of a typical chamber assembly for a reentry control system,
ablative-cooled reaction control thrust chamber.
also designed and manufactured by Rocketdyne.
Its design details are similar to the thrust cham-
ablation continuous until the temperature at the ber shown in figure 11-13. However, no auxiliary
char front drops below the pyrolyzing temperature regenerative cooling jacket is provided. Rather,
(point "b"). When the next cycle starts at _e," a ceramic liner (silicon carbide) is used in the
the char depth remains constant during warmup, combustion chamber section. The nozzle exit is
until pyrolyzing resumes through shutoff at "e. _ trimmed flush with the vehicle outer skin.
The various events may be projected onto the Total radiation cooling applied to reaction
100-percent burn line for convenience. The line control thrust chambers may pose the following
d'-e' in this case represents the soakback char problems:
depth following the second cycle. Subsequent
(1) Difficulties when inboard-installation of
cycles can be treated similarly. Note that in the
the thrust chamber is required
short second cycle, soakback charring is a mul-
(2) Thermal shielding requirements for com-
tiple of the charring during burning. In general,
ponents surrounding the thrust chamber
it can be seen from figure 11-12 that soakback
(3) Recrystallization of the chamber construc-
usually consumes a large percentage of virgin
tion metals in multiple-start applications
ablative and does not contribute to the useful
(4) Need for a larger thrust chamber, when
firing time. Even though some char preheating
operated at a relatively low chamber
time is gained during the subsequent firing, it
pressure
could hardly compensate for the loss of burn
time which would have been available had the
chamber been fired continuously. In this re-
spect, it is desirable to minimize the number of
soakbacks, particularly toward the last portion
of the virgin ablative.
Combinations of repetitive pulsings and mul-

tiple starts can be treated in the same manner as
described above, except that the "on" cycles
composed of repetitive pulsings will effect a
different gradient of the ablation curve, based on
/' /
the particular percent-bum value, instead of the t
INJIfCTOll J SILICON CARBIOE --f
continuous firing or 100 percent curve.

Figure 11-13 presents the design details of a Figure II-13. -Typical I O0-pound-thrust, ablative-
typical 100-pound-thrust, ablative-cooled, reac- cooled, reaction control thrust chamber assem-
tion control thrust chamber assembly for hyper- bly for hypergolic Earth-storable bipropellants,
golic Earth-storable bipropellants designed and designed and manufactured by Rocketdyne (re-
manufactured by Rocketdyne. An optional regen- generative cooling jacket optional).
INJECTOR _-_GLASS WRAP 10 percent tungsten alloy, electroplated with a

PROPELLANT X (STRUCTURAL) _ PARALLEL
cermet coating of chromium and zirconium
VALVE$--?_ MOUNTING _ ,_---ASBESTOS \WRAP
diboride, could be used for the chamber. An
/_ CAN __WRAP "_BLATIVE
alloy of molybdenum and 5 percent titanium,
coated with MoSi 2, may also be used.
Figure 11-15 presents a typical, totally
radiation-cooled reaction control thrust chamber
,.:
",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)
injector plate can be either bolted as shown, or

Figure ll-14.-25-pound thrust reentry control welded to the chamber. A heat shield is mounted
system thrust chamber assembly designed and aft of the injector to reduce radiant heat transfer
manufactured by Rocketdyne. to the injector propellant valves. The shield is
shaped to minimize reflection back to the thrust
A radiation-cooled thrust chamber, if properly chamber. The shield is attached to the mount
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.
FUEL INJECTOR VALVE
OXIDIZER INJECTOR VALVE
INJECTOR
WALL THICKNESS
FUEL INLET
COMB. CHAMBER
DIA
THROAT DIAMETER
OXIDIZER'-_-
INLET
POLISHED REFLECTING INTEGRATED RADIATION-

TYPE HEAT SHIELD COOLED THRUST CHAMBER
-- FITTING
MOUNTI NG
STRUCTURE
Figure 11-15.-Typical total radiation-cooled reaction control thrust chamber assembly.

Design of Control Components for Reaction

Control Engine Systems
The various aspects of control component

design for reaction control engine systems are
similar to those for spacecraft main propulsion
systems, except that propellant and pressurant
flow rates are at much lower levels. For in-
stance, the propellant flow rates of a typical roll
control thrust chamber may be 0.0041 to 0.007
lb/sec (oxidizer), and 0.0021 to 0.0032 lb/sec
(fuel), to produce a thrust range from 1.6 to 2.5
--lr"--*.o o,,.
pounds. Besides the propellant injector valves,
all other control components, such as pressure Figure 11-16.- Typical injector propellant valve
regulators, check valves, and vent valves, are used with reaction control thrust chambers in
similar to those of other systems. pulsing applications, designed and manufac-
In addition to high reliability and suitability tured by Rocketdyne.
for the space environment, pertinent design re-
quirements for the injector propellant valves of thrust chamber assemblies shown in figures
reaction control engine systems are: 11-13 and 11-14 for pulsing applications. The
(1) Fast (as short as 2 milliseconds) and pre- valve is a fast-acting, solenoid-operated, direct-
cise response when opening or closing actuated type. The propellants flow through the
(2) Tight shutoff, with no afterdribble solenoid core at nearly constant velocity, result-
(3) Low power consumption ing iu low-pressure drop. The valve incorporates
Most injector valves are solenoid operated. a hermetically seaied electrical coil, and a
The needle-type propellant valve design shown metal-ball-type seal. The latter is crimped
in figure 7-43 is suitable for integral injector soiidly into tl_e nose of the armature and engages
valve assemblies, with both valves actuated by a conical metal seating surface. Positive seat-
a common solenoid.
ing is obtained by lapped fits. The inlet port
Figure 11-16 shows another typical injector contains a filter to protect against contaminants.
propellant valve, designed and manufactured by The valve is capable of operating reliably in the
Rocketdyne. It is used with the reaction control temperature environment of space.
INDEX
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
design values, 393 general description, 72, 73

engine performance diagram, 66 helium system, 156
general description, 64, 67 preliminary design layout for 48K, 75
ignition monitor valve, 303 system operation, 73, 74
impulse turbine, 249, 250 thrust chamber configuration, 95, 97
influence coefficient, 397-398 thrust load, 76
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,
starting sequence, 67 158, 159

static-firing measurement list, 412 engine and propulsion schematic, 78
system operation, 67 engine mixture ratio control loop, 269
thermal conductance, 102 estimated pressure drops in oxidizer tank, 158
thrust chamber, 95, 96, 109, 110 feed system, 320
turbopump, 193, 199, 201, 249, 392 general description, 76
A-2 stage engine, 37, 38, 39, 68-72, 76, 96, 111, 112, helium system, 156
129, 131, 154, 165, 166, 199, 233, 236, 238, 269, internal configuration for thrust chamber, 97
389, 392, 399 main-stage thrust throttle control loop, 268
accessory weight history, 37, 38, 39 omission, 64
coaxial shell chamber design, 111, 112 operating parameters for 7.5K, 77
control of propellant, 269 operational sequence, 79
cutoff sequence, 72 oxidizer tank, 154
fuel pump, 233, 236, 238 propellant control system, 268

general description, 68-70 propellant mixture ratio control loop, 269, 270
injection momentum ratio, 129, 131 propulsion system, 78, 338, 392
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
payload weight of, 38 Ablative cooling, 118, 119, 448

Acceleration, isentropic and nonisentropic, 13
preliminary layout for 150K, 71
schematic for tank pressurization of propellant, 165 Accelerometers, 411, 415
Accordion motion, 420
sequence diagram, 71
451
Acoustic chamber liners, 149, 150 C* correction, 398-399

Actuators for thrust vector control, 272.279 Calculations for 4 stages of Alpha vehicle, samples
Additives for propellants, 19 of, 63-79
Aerodynamic spike nozzle, 93 Catalysts, 141
AirReseaxch Manufacturing Co., 349 Catalytic bed of monopropellant gas generator, 133,
ALBM, 26 134
Alloys, 61-62 Catastrophic failure, definition of, 44
aluminum, 61, 76, I18 Cavitating inducers, 210
cobalt base, 62 Center of gravity, data sheet for typical, 39
copper base. 62 Centrifugal pumps, 176, 177, 196, 220-225
magnesium, 61 Chamber, 4, 13, 14, 15, 149, 200
nickel-base, high temperature, 61 Chamber wall cooling system, demands on, 34
special, 61, 62 Change control, 51
Alpha vehicle, 4-stage, 63, 95 Char front, 447-448
Altitude control requirements, 442 Char-rate analysis, 120
Altitude thrust coefficient, 17 Chemical reaction products, systems using, 167-171
Alumina, palladium-impregnated, 141 Chugging, 146
Aluminum-alloy tubing, minimum bend radii for, 360 Closed-loop mixture ratio control, 269-270
Aluminum alloys, 61, 76 Clustering of liquid propellant, 415-419
Aluminum on ablative liner, 118 Coaxial injector, 127
Ambient air pressure, effect on performance of, 1, 14 Coaxial shell thrust chamber, 109
Apollo LEM, 435 Cobalt base alloys, 62
Apollo moon-landing program, U.S., 430 Colburn calculation of gas-side heat-transfer coef-
AR systems, 141 ficient, 100
Asbestos on ablative liner, I18 Collier's magazine, 415
ASI (augmented spark igniters), 140, 435 CoIumbian, 62
Assurance of reliability, 43 Combustion chambers, 4. 5, 123, 148, 149, 150
Astronautics magazine, 127 baffles, 149
Atlas ICBM, 329 design as factor in preventing triggering, 148
Atlas Mercury, 64 disturbances, 150
Atomization, evaluation of, 131 divergent wall gap, 149
Augmented spark igniters (ASI), 140, 435 function, 5
Axial pumps, multistage, 177-178 gas flow processes, 4
gas flow through, 5
heat protection, 123
shape, 5
velocity of liquid propellants at injection plane, 5
B- 10 life, 258 Combustion devices, failures, definition, 47
Baffled injector, 122 Combustion instability, 144-147
Baffles, combustion-chamber, 149 design approaches toward control, 147
Ball bearing centrifugal load DN limits, 259 German type, 144
Ball-type valves, 298, 299 high-frequency, 145
Bung-bang control. See On-off type control intermediate-frequency, 146
Bartz calculation, 100:101 low-frequency, 146
Bell nozzle, 90, 91 types, 144
Bellows, 292, 293, 375-379 Combustion product gases, 12, 197
Berkeley, Edmund C., 428 Combustion stability, 122, 147, 150
Bernoulli's energy equation, 280 injector, 122
Beta angle, definition of, 128, 129 limits, 147
Bipropellant combinations, 25-27 rating, 150
Bipropellant gas generators, 134, 171, 181 Combustion timelags, 146
Bipropellant injector, 126 Complex redundancy, 46
Bipropellants, 19, 132, 138 Components and mounts, interconnecting, 353-381.
Bleed lines, cryogemc propellant, 356 See also Interconnecting components and mounts
Bootstrap, 67 Conax Corp., 327
Breadboard engines, 64 Confidence level, relation to reliability, 42
British systems, 141 Conical nozzle, 90
Buddy plan, 45 Connectors, terminology for, 404, 405
Burnout weight of engine, 35 Conservation of energy,, principle of, 4, 5
Burst diaphragms, 326 Conservation of matter, principle of, 5
Butterfly valves, 294-298 Control of manufacturing and material defects, 48
INDEX 453
Control systems, 263-327 De Laval nozzle, 6, 8, 89, 239

calibration, 266 Density impulse of propellant additives, 20
checkout and test, 267 Derivative pressure feedback (DPQ) valves, 273-274
closed-loop, 263-265 Derivative type control, 264
cutoff, 265 Design change checkoff sheets, 51
dynamic seals for fluid-flow components, design Design Check-Off Sheet, purpose of, 50
for, 289- 294 Design integration of engine systems, dynamic analy-
engine thrust level, 267-268 sis for, 384-390
fixed-area-type regulators, design of, 307-309 Design quality, importance of, 50, 51
fluid flow components, design for, 280-294 Design reviews, 49-50
gas pressure regulators, designof, 315-321 Diametral-squeeze-type O-ring seals, 290
liquid flow and pressure regulators, design of, Directed nonexplosive gas flows, 150
321-322 Discontinuity stresses, 339
main stage duration, 265 Divergent wall gap, chamber, 149
miscellaneous valves, 325-327 DN value, 258
open-loop, 263 Dome-loaded gas pressure regulator, 3 19-320
orifice locations and sizes, 266 Doublet injector. 125, 127
pilot valves, design for, 301-307 Dry weight of engine, 35
pressure relief valve, design for, 322-827 D.S.D. Co., 367
propellant-mixture ratio and utilization, 268-272 Du Pont, 383
propellant valves, design of, 294-301 Dual bipropellant gas generator system, 172
safety, 265 Dual combustion, 181
servovalves, design of, 309-315 Dual shaft turbopump, 182
start, 265 Duct design, 373, 374
thrust vector actuators, 272-279 Dump cooling, 114
Coolant side heat transfer, 105, 106, 107 Dynamic analysis in engine system design integra-
Cooling system for thrust chamber, 98-99, 100-102, tion, 384-390
104- 121 Dynamic seal, design, 258
ablative, 119 Dynamic stability, 148
circular tube wall of regenerative, 108
coaxial shell, 109
considerations in selecting, 99 E-D (expansion-deflection) nozzle, 92, 93
coolant side-heat transfer, 105
EDS. See Emergency detection system
correction factor, variation across boundary layer,
Eiastomer rings, 366
101 Electrical failures, definition of, 47, 48
dump, 114
Electrical hydromechanical component analogies, 275
film, 115
Emergency detection systems (EDS), 45, 421
gaseous film, 116, 117
Energy losses during pump operation, 194-195
gas-side heat transfer, i00
Engelhard Industries, Newark, N.J., 141
heat transfer schematic for regenerative, 104
Engels, D. A., 275
liquid film, 115, 116
Engine components, failure modes, 45
radiation, 120, 121
Engine design, 1, 3-4, 16, 28, 31-62, 96-97, 181, 416,
regenerative, 104- 114
419-421
techniques and selection, 98
A-I, A-2, A-3, A-4 stage, thrust chamber configura-
thermal resistance of carbon wall deposit, 102
tions for, 96-97
transpiration, 117, 118
areas for increase in quality, 51
Copper base alloys, 62
availability, 50
Cost of vehicle design, 50
burnout weight, 35
Coulbert, C. D., 433
departures from basic simplicity, 28
Critical failure, definition, 44
detailed procedures, 51-55
Critical pressure ratio between throat and chamber, 6
dry weight, 35
Cryogenic propellants, 19, 68, 165, 166, 167, 187,
duration of run, 32-34
271, 279, 288, 291, 331, 332, 336, 345-348, 356,
envelope (size), 4_
422-424, 429, 440
flight run duration, 33
Curtis turbine, 180
gas exhaust velocities, sea level and in space,
Cutoff impulse, 34
sample calculation, 3-4
manual for, 419
parameters, major, 31-62
Damping devices, 149 parameters, ranges of, 16
Decay deviations, 427 performance as function of altitude, 32
Deferred failure, definition of, 44 performance parameters, 32
Engine design (continued) Error, human. 49

quality, importance of, 50, 51 EWF (equivalent-weight factor), 187
shape, most used laws and formulae for, 1 Expansion of gas, isentropic and nonisentropic, 13
simplicity, 32. 48
stress analysis, 56
thrust in space, sample calculation to determine, 3-4
Failure, 44-48, 51, 58, 59
thrust level, 31-32
catastrophic, 44
tools for improving quality. 51
combustion devices, 47
typical sequence diagram, 55
criteria, 59
wet gimbaled weight, 87
critical. 44
wet weight, 37
deferred, 44
Engine failure sensing and shutoff system (EFSS).
electrical, 47
420-421
fatigue, 46
Engine Out (EO) capability, 416
functional, 45, 46
Engine out, definition, 44
mission, 44
Engine out source, 181
modes of, 44, 58
Engine system, 1-30, 42, 43, 55, 59, 60. 63-79, 144,
order of, 44, 45
148
reports, 51
A-l, A-2, A-g, A-4 stages for Alpha, 63-79
unexplained, classification of, 48
basic elements, 21
Failure modes of engine components, definition, 45
components, 9.8
Fatigue failure, definition, 46
function, 1
FBM, 26
introduction, 1-30
Feed system hydraulic capacitor, 149
low alloy steels. 60
Feed system perturbations, 150
model of thrust equation, 3
Fiber-glass filament-wound propellant tanks, 348-349
operating characteristics in prevention of triggering,
148 Film cooling of thrust chamber, 115
Flame holding characteristic, 146
oscillations, effects of, 144
Flame propagation velocity, 146
pressure environment as function of altitude, 2
Flange joints, design, 360-370
reliability of design, 42, 43
Flexitallic Gasket Co., 367
selection of materials, 59
Flow (of propellant), 9- I0
size, effects of, 42
areas, computations for, 10
thrust chamber nozzle, specific function of, 2
computations, 10
thrust equation (general), 2
specific volumes. 9
thrust generation, momentum theorem in basic equa-
static pressures, computation for, 9
tion for, 1
temperatnres, computation for, 9
typical pressurized gas feed, 28
velocities, computation for, 19
typical schematic, 55
Flowmeters, 271. 411, 414-415
typical turbopump feed, 29
Fluid, ideal, basic flow characteristics of, 280
Engine systems design integration, 883-428
Fluorine, in A-3 stage engine, 72
calibration, 390-394
Formula(s), 1, 7, 8, 18
clustering of hquid propellant, 415-419
design of shape and prediction of engine behavior, 1
dynamic analysis, 384-390
effect of pressure on thrust, 13
electrical, 403-411
ideal gas flow, 7, 8
engine-to-vehicle interfaces, 383-384
Fox Valve Development Co., 300
instrumentation, 411-415
Franms-type impeller, specific speeds, 190
mechanical development, 399-403
Frozen vs. shifting composition of gas flow, 21
performance characteristics, 394-399
Fuel additive blender unit. 183, 184
simplicity, 383-384
Fuel additive subsystem, 182, 184, 185
Engine-to-vehicle interfaces, actuator systems. 276-
Fuel characteristics. 34, 35
278
Functional failures, definition, 45
crosstalk and spring rate, 277
Fusible wire links, 142
engine installation and alinement, 276
hydraulic system, 277, 278
loads, 276-277
performance evaluations, 279 Gas constant, effect on performance of engine, 14,
secondary injection of matter, 279 15, 34
Envelope, definition, 42 Gas feed systems, 151, 167
Environmental effects on stress, 56 Gas flow, 4-8
Enzian missile, 140 combustion chamber, 4
Ermeto flareless tube fitting, 358 formulas for ideal, 7, 8
INDEX 455
Gas flow (continued) Hydrogen (continued)

parameters and terms applicable in thrust chamber, 4 drawbacks, 68
terms used in calculation, 5 gas tapoff systems, 136
thrust chamber, ideal conditions, 4 Hydrogen peroxide, 133, 138, 140. See also Catalysts
Gas flow orifices, 309 Hypergolic fluid, 167
Gas-generating devices, 131-136 Hypergolic ignition, 136, 138, 139, 140
classification, 132 Hypergolic slug, 139
design objectives, 131-132 Hypergolic vs. nonhypergolic, 432
solid propellant, 132 Hypergolics, definition of, 19. See also Noeggerath
Gas generator igniter, 138 and hypergolics, 138
Gas generators, 132-136, 171, 172
Gas generators, liquid bipropellant, operating charac-
teristics, 136
Iacobellis, S. F., 93
Gas law, perfect, 4
ICBM, 26, 27, 33
Gas pressure relief valves, 32.2-323
Igniters, 136, 137
Gaseous film cooling of thrust chamber, 116, 117
Ignition detection, 141
Gaskets, 361
Ignition devices, 136-143
Gate-type valves, 300-301
augmented spark, 140
Gears for turbines of turbopump, 260-261
hypergolic, 138- 140
Gemini capsule, 446
hypergolic slugs, 139
General Controls pilot valve, 302
igniters, 137
German A-4 engine (V-2), 138, 141, 142, 143, 156, 181
indirect, ].43
German efforts, Peenemunde, 139
optical, 142
German systems, 141. See also Catalysts
pressure- sensing. 143
German V-2 (1942), ratio of engine weight to thrust, 36
pyrotechnic. 137
GFE. See Government-furnished equipment
resistance wires. 143
Gimbal angle, 425
spark plugs, 139
Gimbal mounts, 379-381
special designs, 140
Government-furnished equipment (GFE), 427
visual, 142
Gravity, effects of, 4.14
Impeller, centrifugal pump, definition, 177
Greenewalt, Crawford H., 383
Impeller rotor and stator, 236
Ground support equipment (GSE), 267, 426
Impeller rotor, ope[ation, 226
Growth factor of single stage or multistage vehicle, 40
Impellers for centrifugal turbopumps, 204-207, 230-231
GSE. See Ground support equipment
Impellers, specific speeds, 190- 191
Impulse, specific, 10
Inducer, centrifugal pump, definition, 176
Hahn, J. R., 447 Inducer, typic_l, 211
Hart, T. J., 338 Influence coefficients for engine. 397-398
Hatch and Papell equation, 116 Inhibited red fuming nitric acid (IRFNA). 343, 344
Heat combustion chamber, protection, 123 Injecting methods, throttleable, 128
Heat exchanger propellant evaporation system, 165 Injection pattern, 131
Heaters, 408 Injection pressure drop. 128
Helium in A-3 stage engine, 73 Injectors. 83, 113. 121-131. 123
Helium in stored gas systems, 156-158 baffled-type, 122
Helium thrust chamber systems, 162 bipropellant. 126
H-F (horizontal-flow) nozzle, 92 coaxial-type, 127
High-frequency instabilities, 145, 146 concentric ring. 124
High vacuum, characteristics of, 434 configurations, 124- 127
Hilsch separation tube, 349 doublet. 125. 127
H-Q curve, 189 experimental evaluation, 131
H-Q, efficiency, and required power curves of cen- heat protection in combustion chamber, 123
trifugal pump, 196 hydraulic qualities, 123
Human error, 49 impingement patterns, 125. 127
Hydraulic capacitor, feed system for, 149 integral face plate, 125
Hydraulic capacitor, schematic of isolation type, 149 manifold, 113
Hydraulic qualities of injector, 123 objectives in design, 121-123
Hydrazine, as a monopropellant, 133, 171 parameters, 127- 131
Hydrazine-hydrate, 188, 141 premixing-type, 127, 128
Hydrogen, 68, 107, 136, 167 quintuplet-type, 125, 127
coolant properties, 107 ting-slot-typ e. 127
critical pressure considerations, 167 self-impingingotype, 125, 127
Injectors (continued) Liquid-hydrogen-fueled engines, operating regions of

showerhead-type, 125, 127 pumps, 226
special requirements, 123 Liquid hydrogen tanks, 345-348
splash plates, 127 Liquid monorepellant gas generators, 133
throttleable-type, 127 Liquid propellant, definition, 18
thrust chamber, 83 Liquid propellant rocket combustion chambers, gas
triplet-type, 125, 127 flow through, 5, 9, 10
Instability, 146, 147 Liquid Rocket Engine Design Check-Off Sheet, 52-54
Instability of combustion. See Combustion instability Lip seals, 291, 292
Instrumentation of engine, additional elements, 48 Loads on rotor blades of turbine, 244
Insulation for cryogenic propellant tanks, 345-348 Low-frequency instabilities, 146
Integral-type control, 264 L-star(L*), 86, 87, 131, 134
Interconnecting components and mounts, 353-381
assemblies, 357
brazed joints, 370-371
Magnesium alloys, 61
cryogenic propellant bleed lines, 356
Main propellant tank direct injection pressurization
ducts, 372-379
system, 172, 173
flange joint, 360-370
Man Rating, definition, 44
gimbal, 379-381
Manufacturing and material engine defects, control
hydraulic lines, 357
areas defined, 48
pneumatic supply lines, 356
Mark III turbopump gears and bearings, 184
propellant supply ducts, 353
Mark V turbopump, 186
propellant-tank-pressurization lines, 355-356
Marman Division of Aeroquip Corp., 374
purge lines, 356-357
Materials for engine part, selection of. 59, 60
seal drain lines, 356
Messerschmitt Me- 163, 141
thrust, 357
Messerschmitt rocket fighters, 140
tubings and tube fittings, 358-360
Metals, criteria for use with liquid hydrogen, 60
turbine-drive, high-pressure, hot-gas ducts, 357
Meteoric bombardment, 434
turbine-exhaust, hot gas ducts, 357
Mission coast duration, effect, 432
turbopump and other, 357
Mission failure, definition, 44
Interstage length, 425
Mission success, definition, 44
IRBM, 26, 27
Mixed-flow type impeller, specific speeds, 190, 210
IRFNA (Inhibited red fuming nitric acid), 343, 344
Mixture ratio of fuel, 34
Isentropic flow process, relationship between any two
Moment of inertia, data sheet for typical, 39
points in, 5
Momentum theorem, 1, 2
Monergols. See Propellants, mono-
Monopropellant gas generator schematic, 133, 181
J-Box, 409 Monopropellants, 18, 19, 132, 133, 171
J. Spacecraft and Rockets, 433 Moody, L. F., diagram, 283, 284
JTA liner. 438, 439, 448 Multiceilular design, 416
Jupiter rocket, 400 Multiple-engine use, 428
Multiple start, 442
Multistage vehicle, growth factor equations, 40
Kaplan method of error triangles, 210

Kinetic energy, conversion from thermal energy, 1
NASA, S-1 first stage booster for Saturn I, 416
Needle-type valves, 301
Newton' s law of viscosity, 282
L*, (characteristic length of combustion table), 87. Nickel-base alloys, 61, 76
See also L-star Noeggerath and hypergolics, 138
L*, effect on c* value of thrust chamber, 87 Noncomplex redundancy, 46
Laws, design and predict rocket behavior, 1-4 Noncryogenics, 167, 271
Lee, J. C., 447 Nonmetallic materials, uses, 62
LEM (lunar excursion module), 430 Nonpropellants, systems evaporating, 167
Li quid bipropellant gas generator, 134-136 North American Aviation, Rocketdyne Division, 136
operating characteristics, 136 North American X-15, 429
valve assembly, 134 Nozzles, 2, 4, 6, 7, 8, 13, 14, 15, 89, 90, 92, 93, 94,
Liquid film cooling of thrust chamber, 115, 116 95, 97, 98. 239, 307
Liquid flow regulators, 321-322 aerodynamic spike, 93, 94
Liquid hydrogen, selection of metals for use with, 60 annular, 92, 94, 95
INDEX 457
Nozzles (continued) Performance. injector, 122

area expansion ratio, effect on performance, 15 Performance parameters for engine, 10, 15-18
bell, 90, 91, 93, 98 correction factors and magnitudes, 15. 16
conical, 89. 93 measurement and dimensions, 10
convergent section. 13, 14 sample calculation, 16, 17, 18
De Laval type. 6, 7, 8, 89 Perturbations, feed system, 150
divergent section, 14 PFRT. See Preliminary flight rating tests
E-D (expansion-deflection), 93 Pilot valves, control, design. 301-306
expansion area ratio, 89 Pneumatic pressure regulator, 317
gas flow through, 4. 6 Pogo stick effect, 420
incorrect length and jet separation, effect on thrust, Poiseuille's equation, 284
7 Poisson's ratio, 108, 109, 110, 163, 209, 343, 375
liquid flow througb, 307 Poppet-type valves, 299
overexpanded, 6, 7 Position indicators. 407
performance comparison, 94 Postwax engine, 36
pressure losses at inlet, 6 Potassium permanganate, 141. See also Catalysts
prime functions, 6 Power sources, turbine, 181
radial in-flow and out-flow, 92 Prandtl-Meyer turning angle, 94
reverse flow (R-F), 92 Prandtl number, 100, 106, 118
specific function in thrust chamber, 2 Preliminary flight rating tests, 32
turbine, 239 Pressurant requirements. 152- 156
NPSH (net-positive suction head). 166, 176, 188, 192- Pressurant use factor, 158
194, 210, 346 Pressure-compounded impulse turbine, 180
Nuclear radiation, 434 Pressure drop, effective injector, 131
Nusselt number, 100, 106 Pressure environment as function of altitude, 2
Pressure feedback (PQ) valves. 273
Pressure, formula demonstrating effect on thrust of, 13
Pressure gages, principal types of. 412-414
Off-design characteristics of turbopump, 203 Pressure losses at nozzle inlet, 6
Off-nominal conditions, engine performance variations, Pressure-sensing devices, 143
396 Pressure switches and transducers, 411
Offset turbopump, 182 Pressurized-gas propellant-feed systems, 166-167,
One-and-one-half stage vehicles, 416 170, 171, 173-174
On-off type control, 264, 442 A-4 stage, 174
Open-loop mixture-ratio control, procedures, 268-269 applications, 166-167
Optical detection, 142 criteria for selecting, 173, 174
Optoline. 140 evaporating nonpropellants, 167
Order of failure, definition, 44, 45 liquid propelled, 170. 171
Orifices, liquid flow through, 307 using products of chemical reactions, 167
O-rings, 45, 83, 289-291, 292, 298, 300, 303 Pressurizing gas turbulence, 152
correlations in design, 289, 290 Pro rained mixture ratio control (PMR), 270
flange joints, 366 Pro Jellant combination, 129, 148
friction, 290 Pro)ellant evaporation systems, 165-167
poppet-type, 291 Pro)ellant feed system, triggering prevention, 148
seal failure in. 45 Pro)ellant feed system, centrifugal pumps, 176. 177
selection, 290-291 Pro)ellant feed system perturbations, 150
squeeze-type, 291 Pro)ellant flow design, pressure feed engine, 391
surface finish, 290 Pro _ellant fraction, definition of, 35
valve seat, seals for, 291 Pro)ellant management system, 270
Over-stress and over-strain, definition, 47 Pro)ellant mass fraction of complete vehicle, defi-
Oxidizer pump inducer, 215 nition of, 11
Propellant storage, system design for, 440-442
Propellant tanks, 329-352
cryogenic, 345-348
Packaging, 399-403 design, basic, 329-331
Pancake turbopump, 182 design, general, 332-336
Parts list, rocket system, symbols, 28-29 expulsion under zero gravity or oscillatory g load-
Payload, specific impulse effect on, 11 ing, 349- 352
Payload weight vs. growth factor, 40 fiber-glass filament-wound, 348-349
Peenemunde, 139, 140 pressurant diffusers. 349
Perchlorate-type fuel, 138 storable liquid, 343-345
Perfect gas law, 4 structure, 336, 343
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
Specific heat ratio, 9, 15 Thiokol, 446

effect on performance of, 15 Thoma parameters, 193, 194, 211
values for functions of, 9 Thor rocket, 64, 400
Specific impulse, 10, 11, 12 Three-stage vehicle system, weight increase calcula-
Specifications improving engine design quality, 51 tions, 41
SPGG. See Solid propellant gas generator Throttleable injector, 127
Spike nozzle, 92 Thrust, 7, 12, 15, 31, 33. 34, 81-83, 1S1
Splash plates injector, 127 coefficient, 12
Spool-type servovalves, 313, 314, 315 decay, 34
Spring mass effect, 420 effect of incorrect nozzle length and jet separation,
Squeeze-type O-rings, 291 7
Stage burnout velocity of single-stage vehicle, equa- optimum, ambient pressure, 15
tions, 39, 40 total requirement, 31
Stage separation, basic possibilities, 425-4°6 Thrust chamber
Stainless steel in pressurant storage vessels, 162 assembly, 82
Stainless steel on ablative liner, 119 basic cylindrical combustion in. 88
Start (thrust buildup), characteristics, 33 basic elements. 81-83
Stator, primary functions of, 177 bleed, 181

Stay time of burning gas in combustion chamber, 34 characteristic velocity, 83
Steels, stainless, 60, 61, 76, 119, 132, 162 coaxial shell, 107, 109, 110
Stefan-Boltzmann radiation heat transfer constant, 121 coefficient, 84
Stoichiometric mixture ratio, 34 combustion data (theoretical), 84, 85
Storable liquid propellants, 19 configuration layout, 86
Storedgas systems, 156- 165 cooling, 98- 121
commonly used configurations, 156-157 deposits on walls, 101. 102., 103, 104
design of components, 162 designs, 13, 14
requirements of calculations, 157-162 film-cooled, I17
Strain gages, 415 gas tapoff systems, 136
Stress analysis, 56-59 geometrical shapes, 88
experimental, 58, 59 heat exchanger, 162
failure modes and criteria, 59 ideal, data, 8
selection of materials and dimensions, 58 injector, 82
typical steps, 56 injector design, 121-131
working loads and environmental effects, 56, 57 L*,effect on c* value, 87
Structural integrity of injector, 123 model of, 3
Suction speclhc speed, 192 nozzle expansion area ratio, 89
Sweeteners, 141 nozzle, specific function of, 2
Swing-gate-type check valve, 326 nozzle shape, 89
Symbols for system parts, 28, 29 performance calculation, 84
System parts, symbols, 28, 29 performance parameters, £,3
Systems using chemical reaction products, 167-171 pressure, 144
recommended combustion characteristic length (L*)
for propellants, 87
selecting cooling method, 99
Tachometers, 411 shape, 87, 131
Taifun rocket, 140 specific impulse, 11, 83
Tank wall temperature, 152 thrust generated, 13, 14
Tanks, liquid hydrogen, 345-348 tubular wall, 107
Tanks, propellant. See Propellant tanks Thrust efficiency vs. bell nozzle length, 91
Tantulum, 62, 121 Thrust vector control, 274-275
Tapoff system, 136, 180 Timers, 407-408
Telemetry of engine, additional elements required, 48 Timosheuko, S., 209
Temperature Titanium alloys, 121, 163
effect on low-alloy steels, 60 Titanium-base alloys, 62, 12I, 162
chamber, calculation, 21 Topping flow cycle, 200
gages, 411 Topping systems, 181
sensors, 412 Transpiration cooling of thrust chamber, 117-118
steels, low-alloy, effect, 60 Triethyl aluminum and oxygen, 139
Thermal energy, conversion from chemical energy, 1 Triggering processes, prevention of, 148-149
Thermal energy, momentum caused by release of, 1 Triplet injector, 125, 127
Thermal environment, 434 Tubing, 358-360
Thermocouples, errors in measuring by 412, 413 Tungsten, 119, 121
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
Ullage gas compression, 152 Zucrow and Sellers equation, 115
NASA-l.,angley, 1968 _ $8

Design of Liquid Propellant Rocket Engines Dieter K Huze and David H Huang - Part2

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Design of Liquid Propellant Rocket Engines Dieter K Huze and David H Huang - Part2

Caricato da

Copyright:

Formati disponibili

DESIGN OF TURBOPUMP PROPELLANT-FEED SYSTEMS 225

VANES --_ _ VANES

From equation (6-77), the reduction of the

F 3zr rolaaA oa aAx (204 304 - 21609)

x 71.38 × (0.25 +0.21)_ 74 680 lb OF ROTATION

6.4 DESIGN OF AXIAL-FLOW PUMPS

Except when used as inducers, application

(fig. 6-4). As can be seen in figure 6-6, the fluid

I C[_'rmFU_,L I Operation of the Impeller Rotor

arrangement, and others. For instance, the re- Pr -=- (6-80)

critical the tip clearance of an open-faced impel-

where S_-= rotor vane solidity at the mean effec-

impeller st_tor of f_n _xi_l-[loW

impeller rotor, _ncl

is gradually increased from fi2 to f13. The dif- (AH)I

where a2', a3' =absolute flow angles at the rotor

Rr = radius of the rotor vane curvature, in effective diameter din, ft/sec

design of impeller rotors, velocity diagrams of 10 percent of Q)

/32 =/32' + i (6-84) (AH) 1 = rated design developed head per

UmCuj - UmCu2 f dm tan fi2 --dr tan/32_

hub diameter, deg the stator (fig. 6-52):

eter, ft/sec Ss :P_-s (6-98)

The purpose of the stator of an axial-flow

the same effective passage cross-sectional area

correlations can be established for the design of

flow pump. aic= 2 (6-110)

Inducer stator vane solidity design values

shown in figure 6-54. It consists of a disk hav- r7 2

(:Iv, - ..... _,:\_ _ _ :\

•"_"_!\_4, ',"-,"R.P.M. _..-_,_

FRONT SEAL RUB'--" X-BALANCE PISTON

Figure 6-54.-Axial-(low pump volute casing and balance piston arrangement.

Rated design pump developed head, AH

Impeller leakage loss rate, Qe = 0.06 Q

impeller rotor and stator.

(a) Inducer stator AHind (1 - 0.08) : AH- n(hH),

Nss (NPSH)c o._ = 6240 ft

From equation (6-106), the pump developed Nrrdt _ 27 000 × 7r× 7

Substitute this into equation (6-66):

3.12 × _-× 230

Qimp -- Q + Qe = 6080 (1 + 0.06) = 6450 gpm Tile peripheral speed at d,

3.12 x-_(dt 2 -doh 2)

Cmo 64.8 , cm 230

Use identical values cm for the meridional

6450 al =a,'+i=40°26 ' +4 ° =44°26 '

3.12 x4× (49- 36) x 0.88

dlh=_dt2 Qind The tangential component of the stator abso-

cm 230 230 Nominal mean effective diameter, d m =6.52 in

The stator absolute outlet flow velocity Pis=l.205in; Cis=l.844in; Lis=l.55in;

c 2' = \/c u 2' 2 + cm 2 = X/11 490 + 52 900

-- ,,:'65 350 = 253.8 ft/sec (b) Impeller rotor and stator

Pis - rrdm _ ,7×6.52 1.205 in

c 2' =c 4' =253.8 ft/sec

cu2' = cu4' = 107.2 ft/sec

The design relative flow angle at the impeller

The relative flow velocity at the impeller

v2' =\/ium-Cu2') 2 +cm 2

The radius of the inducer stator vane curva-

From equation (6-89), the required developed

The impeller rotor design absolute outlet flow Cr 1.346

Assume that the design absolute flow condi-

The impeller rotor design relative outlet flow