V ol. 19 No.

1 CH INES E JO U RN A L O F AER ON A U TICS February 2006

Studies on Aerodynamic Behavior and Performance of

Aerospike Nozzles

WANG Chang hui, L IU Yu, L IAO Yun fei

( School of A st ronaut ics, Beij ing Uni versity of A eronautics and A str onautics ,
Beij ing 100083 , Chi na)

Abstract: Ex perimental and numerical studies are car ried out on a 6 cell tile shaped aerospike nozzle, a
6 cell aerospike nozzle with round to r ectang le primary nozzles and a 1 cell linear aer ospike nozzle. Good
alt itude compensation capacities and high efficiencies are obtained in t he tests. T he efficiencies of 6 cell
t ile shaped aerospike nozzle and 1 cell linear aerospike nozzle at design altitude approach to 100% , and
t hat of 6 cell aerospike nozzle with round to rectang le primary nozzles in the same condition is about
95% due to the imperfect cell contour and manufacturing defects. Numerical r esults are in good ag ree
ments w ith t est data. T he effects of ambient pressure on ex haust and then on base behavio r are ana
ly zed. T he effects o f variation in the amount of base bleed on per for mance are also ex amined in the tests.
Key words: ro cket engine; aer ospike nozzle; perfo rmance; ex periment; numerical simulatio n
塞式喷管气动特性与性能研究. 王长辉, 刘宇, 廖云飞. 中国航空学报( 英文版) , 2006, 19( 1) : 1-
9.
摘 要: 对 6 单元瓦状塞式喷管、
6 单元圆转方内喷管塞式喷 管和 1 单 元直排塞式喷管进行了 实验
和数值模拟研究。实验表明, 3 个塞式喷管具有良好的高度补 偿特性和较 高的效率, 6 单元瓦 状塞
式喷管和 1 单元直排塞式喷管在设计高度的效率接 近 100% ; 由 于内喷 管型面 不够理想 和加工 时
存在的缺陷, 6 单元圆转方内喷管塞式喷管的设计点效率在 95% 左右。 数值模拟结 果与实验 数据
吻合较好 , 结合实验和数值模拟分 析了外 界反压 对气流, 进而 对底部 气动特 性的影 响。实验研 究
了底部二次流流量变化对性能的影响。
关键词: 火箭发动机; 塞式喷管; 性能; 实验; 数值模拟
文章编号: 1000 9361( 2006) 01 0001 09 中图分类号: V43 文献标识码: A

Aerospike nozzle is considered to have bet ter rope, Russia[ 1] , Japan and China also devoted
performance at off design alt itudes com pared w ith much att ention to this nozzle concept .
that of the convent ional bell shaped nozzle since it s T he aerospike nozzle eng ine was once a candi
plume is open to t he at mosphere outside and f ree t o dat e applied to t he SSME ( Space Shut tle M ain En
adjust, allow ing the eng ine to operat e at it s opt i g ine) and had been studied from t he early 1960s
[ 2 5]
mum ex pansion at all altitudes. T he concept of unt il the end of the 1970s . How ever, techno
aerospike nozzle was f irst proposed in the 1940s logical dif ficult ies in t he development made it f all
and its performance and f easibility w ere v igorously behind t he development of conventional nozzle and
invest igated in t he Unit ed St at es f rom the 1950s t o the research on the aerospike nozzle st ag nat ed. In
the 1970s. In the 1990s, w ith the increasing inter 1996, L ockheed Martin s Venture Star vehicle
ests in the SST O RL V ( Single St age to Orbit powered by aerospike nozzles was select ed as t he
Reusable L aunch Vehicle ) , renew ed int ensive candidat e of SST O RLV and a subscale demonstra
st udies on the aerospike nozzle w ere carried out a t or X 33 vehicle w as planned to prove t he perf or
g ain all over t he w orld. Mot ivated by the develop mance of t he Venture St ar. Wit hin the prog ram of
ment of X 33 vehicle in t he U nited St at es, Eu X 33, a full scale linear aerospike nozzle engine
R eceived dat e: 2005 03 19; R evision received dat e: 2005 11 01
Foundation it em: 863 Plan of China ( 863 2 3 4 10)
 2 WAN G Chang hui, L IU Yu, L IAO Yun fei CJA

w as t ested successf ully on a sea level t est f acilit y in diagonal decomposition algorit hm [ 22] is employed
the U nited St at es. LASRE ( L inear Aerospike SR and only t he steady st at e solut ions are considered.
71 Ex periment) w as carried out by NASA to define T he external flow is st ill air and t he det ails dealing
the aerospike engine performance under realist ic wit h main boundary conditions and the t urbulence
flig ht condit ions[ 6] . Because of overw eight com pos model can be found in Ref. [ 19] .
it e cryogenic f uel tank and high cost, t he X 33 plan In numerical simulation, the cell chamber
w as called off event ually, how ever, leaving m any pressure P c is g iven according to its design value,
dat a that can be useful in t he development of such gas tem perature in t he cell chamber is T c= 300 K,
nozzles. T he studies on aerospike nozzle in Japan T he gas const ant and specific heat rat io of air are
st art ed in the m id 1990s at Japanese Nat ional R g= 287 J/ kg K and = 1. 4, respect ively. P a is
Aerospace L aborat ory. Initially t he annular t ype the ambient pressure and the ambient temperature
aerospike nozzles were invest ig at ed[ 7, 8] , and t he is designated as T a = 300 K. T he nozzle pressure
research then progressed to t he investigat ion of lin rat io NP R= P c / P a is t he indicator of altitude. T he
ear aerospike nozzles[ 9, 10] . Similar research on such thrust ef ficiency is adopt ed to evaluate t he per
[ 11]
nozzles w as also performed in Europe . T he in f orm ance of aerospike nozzle w hich is the rat io of
vestigation on aerospike nozzle has been carried out experimental or comput at ional thrust coeff icient CF
in China since 1997. T he act ivit ies on invest ig at ion to t he ideal one C F0 at a cert ain altitude. = CF/
[ 12]
include cont our design and it s opt imization , CF 0 , here C F = F / ( P c A t ) and F and A t denote
[ 13, 14]
both cold f low and hot f iring t ests and nu the thrust and the t hroat area, respect ively. C F 0 is
[ 15 18]
merical simulat ions . T he tile shaped aeros comput ed as follow ing
pike nozzle w as proposed and its primary per +1 -1

CF0 = 2 2( - 1) 2 1- 1
formance v alidat ion w as done by hot f iring + 1 - 1 NPR
test s[ 13, 19] . In present paper, st udies on aerody For symmetry, only a half domain flowf ield of a
namic behavior and performance of aerospike noz
cell and its corresponding plug are com put ed. T he
zles are carried out by cold f low tests and numerical comput at ional domain lies bet ween t w o planes,
simulations. which are called cent er plane t hat cuts through t he
1 Numerical Method ax is of t he cell and side plane t hat parallels t he cen
t er plane and cuts t hroug h the sy mmet ry line be
1. 1 Governing equation
t w een tw o adjacent cells. T ypical plug grid genera
Because t he cell ex haust flow is supersonic,
t ion result is show n in Fig. 1.
the eff ect s of ambient pressure on t he f low in cell
can be neglect ed; theref ore, calculat ions of int ernal
flow for t he cell and ex ternal flow f or t he plug are
done separat ely. T he flow parameters at the cell
exit are used as an inflow boundary condit ion for
the plug flow . T he governing equations are com
pressible three dimensional Reynolds averaged
Navier Stokes ( N S) equations. T he values of t he
Reynold stresses and t urbulent heat fluxes for N S
equations are obtained by t he L am Brem host k F ig. 1 T y pical plug gr id generation result
[ 20] [ 21]
turbulence model . M USCL scheme is em
2 Experim ental Aerospike Nozzles
ployed for t he discrete convect ive terms and t he
second order central differencing scheme is used for 2. 1 6 Cell tile shaped aerospike nozzle
the v iscous t erms. An implicit L ower Upper ( LU) T he cells and t he plug are manuf actured sepa
 February 2006 St udies on A erodynamic Behavior and Perf ormance of A erospik e N ozzles 3

rately. Each cell is an ax isymmet ric bell shaped adopt ed for v isualization in t he tests. F our pieces of
nozzle w ith sm all area rat io and t he cell contour is transparent org anic glass are f ix ed on t he plug as
g uarant eed t o connect w ith t he plug contour per sidewalls to prevent air f rom expanding sidew ard of
f ect ly by precisions in manufact ure and fabricat ion the plug surface. Derived from the cell chamber,
( in F ig . 2) . T he throat diameter is D t = 14 ! 0. 01 about 1. 6% of t he tot al mass flow rate is injected
mm and t he cell inclinat ion is != 30∀. T he cont our into the base region t hrough 158 holes w it h diame
of the aerospike nozzle is designed by t he met hod of t er of 1 mm.
charact erist ics. T he area rat ios of the cell and t he
overall aerospike nozzle are c = 4. 0 and all = 40,
respectively. Opt imum expansion is achieved at
NP Rd = 1046.

F ig. 3 6 cell RT R aerospike nozzle

When t he engine was fabricat ed, some defect s

F ig. 2 6 cell tile shaped aerospike nozzle
were found. T he cells holding st ruct ure of t he
2. 2 6 Cell round to rectangle ( RTR) aerospike plug, w hich is made of aluminium alloy, is not
nozzle rigid enough. For t he reasons in manuf act uring,
T he cells and t he plug are manuf actured sepa there are small discont inuous st eps betw een cell
rately. Each cell is a round t hroat to rect angle exit contours and plug cont our and t he plug surface is
nozzle w hose t ransit ion f rom t he round cross sec not as smoot h as t he desig n demanded also. All
t ion t o t he rectangle ex it cross sect ion st arts at t he above defects result in some perf orm ance losses.
area rat io of 1. 78 ( in F ig. 3) . T he cell cont our in 2. 3 1 Cell linear aerospike nozzle
the horizontal plane that cut s t hrough t he axis of T he 1 cell linear aerospike nozzle is show n in
the cell is smoot h, but t he one in t he vert ical plane Fig. 4. T he cell and the plug are manufact ured t o
that cut s through t he ax is of t he cell is desig ned get her from a block of stainless steel. T he engine
w it h t ransit ion t o form the rectangle exit. T he mainly consists of a cell plug unit, a base com po
throat diameter is D t= 10. 025 mm and t he cell in nent, tw o plug sidew alls, t w o base sidew alls, a di
clinat ion is != 31. 45∀. T he area rat ios of t he cell mension fixing component and some support ing
and t he overall aerospike nozzle are c= 4. 47 and struct ures. T he dimension f ix ing component is em
all = 41. 01, respect ively. Opt imum expansion is ployed to keep t hroat dimension const ant during
achieved at NP Rd = 1085. T o observe the flow f ield test s. Symmetrical condit ion of t he base flow is
of t his aerospike nozzle, t he schlieren t echnique is achieved by the ext ended base plane whose extend
 4 WAN G Chang hui, L IU Yu, L IAO Yun fei CJA

ed leng th is more than three t imes of t he base test ed aerospike nozzles. T he cold f low air is pro
height. T he engine supported by ball bearings un vided by eighty nine tanks in parallel w ith tot al
der t he base plane can move f reely in t he direct ion volume of more t han 3. 6 m 3 . T he air pressure can
of horizont al thrust. be risen up to 13 M Pa, supplying a m ax imum op
erat ional w orking mass f low rate of about 16 kg/ s.
T he vacuum syst em, w it h a tot al vacuum volume
3
of about 120 m , includes a vacuum pump, t hree
large cylindrical vacuum ex haust tanks and a cubic
test vacuum vessel. T he t est data are acquired by a
thirt y t wo channels comput er dat a acquisit ion sys
t em .
T o obt ain t he performance and base behavior
of t he ex periment al aerospike nozzles, t hrust F,
cell cham ber pressure P c, ambient pressure P a and
base pressure P b are measured. T he test stand is
mount ed horizont ally in t he cubic test vessel and
the t est ed aerospike nozzle engines are f ix ed by
steel wires or bearing supporting bracket . T he
root mean square error analysis reveals that t he un
certaint y of the ent ire t hrust acquisit ion is less than
Fig . 4 1 cell linear aerospike nozzle
1% and t he acquisit ion of pressure is more accurate
T he w idth of the aerospike nozzle is W d = 60 than t hat of t hrust.
mm, the height of t he throat is D t= 4. 09 mm and
4 Results and Discussion
the inclinat ion of the cell is != 35. 9∀. T he area ra
t ios of t he cell and t he overall aerospike nozzle are 4. 1 6 Cell tile shaped aerospike nozzle
c= 4. 0 and all = 79. 98, respect ively. Opt imum F ig. 5 show s t he pressure dist ribut ions along
expansion is achieved at NP Rd = 2840. Derived the tile shaped plugs at t he cent er and side planes
from the cell chamber, some amount of t he total by numerical simulat ion. At low altit udes ( NPR=

mass flow rat e is inject ed into the base region 55, 75) , t he pressure is raised along t he rear part

throug h 366 holes w it h diameter of 1 mm. T hree of t he plug and t he higher ambient pressure level
pressure transducer ports are placed along t he base leads to t he larger pressure amplitude. As the am

surface to measure t he base pressure P b . T ests are bient pressure decreasing, t he pressure amplit udes

conducted in t he condit ions of 0. 0% , 1. 0% , 1. become smaller and t he pressure rising posit ions

5% , 2. 0% , 3. 0% and 4. 0% base bleed rat ios,

and the base bleed ratio means t he rat io of mass
flow rate inject ed into the base region t o the total
mass f low rate.

3 Experimental Apparatus
H ighly pressurized cold f low air is employed as
the propellant . T he experiment al apparatus consist
of an air supply syst em, tw o air compressors, a
v acuum syst em , a cont rolling and monitoring sys
Fig . 5 Pressur e distributions along tile shaped plug
t em, a comput er data acquisit ion system and t he
 February 2006 St udies on A erodynamic Behavior and Perf ormance of A erospik e N ozzles 5

move rearward unt il out of the plug surface eventu markable perf orm ance gain of the aerospike nozzle
ally. T he pressure dist ribut ion along the plug sur is 27% compared w ith t he bell shaped nozzle at sea
f ace is no longer changed at NPR more t han 400 level. By com parison of the present t est results and
and t he pressure dist ribut ion prof ile at NPR= 400 that obtained w it h cylindrical tile shaped plug in
coincides wit h that at NPR= 1000. Ref . [ 19] , it is concluded t hat t he simplif ied plug
It is also found t hat there is dist inct dif ference contour ( cylindrical plug ) may cause several per
bet ween t he flow phenomena at t he center and side cent of performance losses.
planes, f or t he pressure prof iles at t he cent er plane
are different f rom that at t he side plane where X is
less t han 0. 06 m. H owever, the pressure differ
ence at the t wo planes only occurs w it hin some dis
t ance f rom t he cell ex it, w it h expanding dow n
st ream , t he exhaust mixes together unif orm ly and
the pressure profiles at t hese t wo planes become
similar. T he unification of exhaust is much depen
dent on t he ex tent of ex pansion.
F ig . 6 show s t he measured ( EXP) and numer
ical simulat ed ( CF D) t hrust efficiencies of 6 cell Fig . 6 Efficiencies as a function of NPR
tile shaped aerospike nozzle and theoret ical t hrust 4. 2 6 Cell RTR aerospike nozzle
eff iciency of 1 dimensional bell shaped nozzle w ith Numerical simulated Mach number cont our in
the same area rat io as a f unction of NPR. T he nu the RT R cell is presented in F ig . 7. A shock w ave
merical result of t he aerospike nozzle is in good a is generat ed at t he beginning of the RT R cont our
g reement w ith t he t est dat a. When NP R is more transition and t he numerical simulat ed nozzle eff i
than 400, the alt itude charact eristics of t he ciency at cell nozzle full ex pansion is 95. 65% .
aerospike nozzle coincide w it h those of the bell Fig. 8 show s the measured t hrust ef ficiencies of 6
shaped nozzle and 400 is also the m inim um NPR, cell RT R aerospike nozzle as a funct ion of NPR.
from w hich t he pressure distribut ion along the plug T he experimental ef ficiencies are not high enough
is not changed any more. It is show n t hat t he ex and t he measured values are from 0. 9 to 0. 95.
pansion characterist ics of t he aerospike nozzle ap
pear analogous t o those of t he bell shaped nozzle af
t er NPR = 400 and it is concluded t hat f rom
NP R= 400 on, t he pressure dist ribut ion along t he
plug is similar t o t hat along the inner w all of t he
bell shaped nozzle. At t he desig n NPR ( near
NP R= 1000) , the thrust eff iciencies of bot h t he
aerospike nozzle and the bell shaped nozzle approx i
mate t o 100% . However, at NPRs low er t han de F ig. 7 M ach number contours in RT R cell
sign value, t he aerospike nozzle has cont inuous ca
pacit y of alt itude compensation compared wit h t he
bell shaped nozzle and st ill show s hig h performance
at off desig n altitudes. From sea level t o desig n al
t itude, t he eff iciencies of t he aerospike nozzle and
bell shaped nozzle vary w ithin t he range of 93% t o
100% and 66% to 100% , respectively. T he re Fig . 8 Efficiencies as a function of NPR
 6 WAN G Chang hui, L IU Yu, L IAO Yun fei CJA

F low fields at NPR= 500 obtained by numeri losses notably.

cal simulat ion and schlieren technique are cont rast 4. 3 1 Cell linear aerospike nozzle
ed w ith each ot her in Fig. 9. T hough unfortunate F ig. 10 ( a) and ( b) show the measured t hrust
ly, t he scope of t he schlieren device does not cover eff iciencies of t he 1 cell linear aerospike nozzle w ith
the f low field of t he base region, t he numerical re the base bleed rat io vary ing from 0. 0% to 4. 0% as
sult com pares w ell w ith t he schlieren image w it hin a f unction of NPR. It is shown t hat t he ef ficiencies
the domain above the plug. An ex haust around t he wit h dif ferent base bleed rat ios are almost t he same
exit of the cell and an envelope shock w ave are ob when the base is open. In open w ake condit ion,
served. T he schlieren image is similar to what was the base bleed inject ion does not exert an obvious
observed in Ref. [ 23] . influence on t he aerospike nozzle s performance.
Wit h the base bleed rat io increasing, three differ
ences in alt itude characterist ics are observed. First,
the ext ents to w hich t he eff iciencies are deg raded at
the base t ransit ions are dif ferent . Second, t he
NP Rs at w hich t he base t ransit ions occur are dif
f erent. T hird, t he performances are dif ferent in
close wake condit ion. In the case of no base bleed
( 0. 0% ) , t he notable eff iciency degradat ion is ob
served at the base t ransit ion ( near NPR= 140) and
the ef ficiency reduces f rom 89% ( in open wake
condit ion) t o 82. 3% ( in close wake condit ion) .
T he aerodynamic transition of the base behavior re
sult s in a remarkable change of the thrust and con
t ribut es t o the perf orm ance degradat ion. When
some base bleed is injected, a sharp decrease of per
f orm ance reduct ion at base t ransit ion is observed.
T he ef ficiency reduct ion at base transit ion decreases
from 3. 2% t o 0. 12% w it h t he base bleed ratio
from 1. 0% to 4. 0% . When more t han 2. 0% of
base bleed rat io is inject es, no remarkable perf or
mance reduction occurres at base transit ion.

Fig . 9 Flo wfields by numerical simulation and schlieren

technique ( NPR= 500)

T he imperfect cell cont our results in t he shock

w ave generating and makes the most contribut ion
to t he eff iciency losses. As is mentioned above, de
f ect s of the t est ed aerospike nozzle in m anufactur
ing f urther reduce t he performance. Invest ig at ion
about the design of RT R cell is carrying out and
some design principia are concluded. For example,
the RT R contour w it h start ing circular cross sec ( a) 0. 0% 2. 0% base bleed r atio

t ion from throat seems t o reduce t he performance

 February 2006 St udies on A erodynamic Behavior and Perf ormance of A erospik e N ozzles 7

4. 4 Ambient pressure on base behavior

F ig. 11 show s t he measured P b / P c and P a/ P c
in 1 cell linear aerospike nozzle test s as a funct ion
of NPR wit h base bleed rat io f rom 0. 0% t o
2. 0% , here, P b is the average of t he t hree ac
quired base pressures. Fig. 12 shows the pressure
dist ribut ion along t he full plug ( only X < 1 m is
given) of 1 cell linear aerospike nozzle by numerical
simulation at some NPRs dist ribut ing from sea level
to high alt it udes.
( b) 2. 0% 4. 0% base bleed ratio

( c) 0. 0% base bleed ratio

F ig. 10 Efficiencies as a function of NPR
F ig. 11 P b/ P c and P a / P c as a funct ion of N PR
Wit h the base bleed rat io increasing, t he base
transition NPR increases and t he base t ransiton
NP R has t he low est value when no base bleed is in
jected. T he tested aerospike nozzle has the highest
eff iciency at desig n NPR in case of 1. 0% and 1.
5% base bleed ratios. T he performances in case of
0. 0% and 2. 0% base bleed rat ios are similar, but
a lit t le low er t han those in 1. 0% and 1. 5% base
bleed rat io cases. T he performance w ith 3. 0% and
4. 0% base bleed ratios f urther low eres compared
w it h former cases, especially in case of 4. 0% base
Fig . 12 Numerical r esults of full plug pressure distribu
bleed ratio. T he base pressure increases and t he
tion
performance degradat ion at base t ransit ion is avoid
ed due to some base bleed inject ion. How ever, too In t he range of NP R < 140 ( NPR = 100,
much bleed inject ion resultes in perf ormance degra 110) , t he exhaust is compressed for the second
dat ion and 1. 5% to 2. 0% base bleed inject ion time by ambient pressure in base region ( in F ig.
seems to be t he best in the test. F ig. 10( c) show s 12) and t he open w ake is observed, for the base
the measured and numerical simulated t hrust eff i pressure changes w it h the ambient pressure ( in
ciencies of t he 1 cell linear aerospike nozzle w it hout Fig. 11) . With t he ambient pressure decreasing,
base bleed as a f unction of NPR. T he predict ed the compressed posit ions of t he exhaust move
performance is a lit tle higher t han t he ex perimental downstream and the pressure amplit udes by com
result in open wake condit ion but has a good agree pression become lower. In the range of 140 <
ment w ith t he t est dat a in close wake condition. NP R< 200 ( NPR= 160, 180) , the base region is
 8 WAN G Chang hui, L IU Yu, L IAO Yun fei CJA

dominat ed by expansion absolut ely ( near NPR = aerospike nozzle w ith round to rect angle primary
140) or principally ( near NPR = 200) ( in F ig. nozzles m ainly resultes from t he imperfect cell con
12) , and as a result , the ex haust has enough t our. T he effect s of base bleed rat ios on base be
st rength t o close the base reg ion and t he close wake havior and performance of the aerospike nozzle are
is observed, for t he base pressure is independent of studied and t he 1. 5% t o 2. 0% base bleed rat io is
the ambient pressure and st ays const ant ( in F ig. the recommended value range for t he tested
11) . In the rang e of 200< NPR < 525 ( NPR= aerospike nozzle. T he base behaviour largely de
300, 400) , t he first compressed positions of t he ex pends on t he aerodynamic status of t he ex haust in
haust move downstream t o the base region ( in F ig. base reg ion. If compressed by ambient pressure,
12) , t he w ake turns open again and the base pres the open w ake occurs and t he close w ake occurs if
sure comes under t he inf luence of ambient pressure expanded.
once more ( in F ig . 11) . In t he rang e of NPR> 525
6 Acknowledgment
( NPR= 1000) , t he first compressed position of t he
exhaust moves further dow nst ream and out of t he Ex perimental w ork has been performed by all
base reg ion ( in F ig. 12) , and t he wake is closed by members of aerospike nozzle research t eam. T he
the int ensively expanded ex haust again and t he aut hors would like t o acknow ledg e professor
base pressure is not inf luenced by the ambient pres ZHANG Guo zhou, eng ineer MA Bin for t heir ad
sure any more( in Fig. 11) . vice and suggest ions in experiment al setup con
T hough the base separat ing f low is complex, st ruct ion and professor CHENG Xian chen for his
g enerally speaking, the base behav ior ( open or direct ion in data acquisit ion system and control sys
close) depends on w hether t he ex haust in base re t em. Special t hanks also go t o Ph. D. WANG Yi
g ion is dominant ly compressed or ex panded by am bai, QIN Li zi and L I Jun wei for t heir cont ribu
bient pressure. In case of some base bleed injec t ions in ex periments.
t ion, due t o moment um and pressure of base bleed,
References
a higher expansion int ensit y in base reg ion is neces
[ 1] Baf talovskii S V , K raiko A N, T illyayeva N I. O pt imal de
sary t o close the base flow compared wit h that in sign of self cont rolled spike nozzles and t heir t hrust determ i
case of no base bleed injection. As show n in F ig. nat ion at st art [ R] . AIA A 99 4955, 1999.
11, tw o open to close transition NPRs increase [ 2] R ao G . Spike nozzle cont our f or opt imum t hrust [ J] . Ballis
t ic M issile and Space T echnology, 1961, 2: 92- 101.
w it h base bleed rat io increasing from 1. 0% to 2. [ 3] Lee C , Thompson D. Fortran program f or plug nozzle design
0% and t he range of NPR in w hich the f irst close [ R ] . N ASA T M X 53019, 1964.
w ake condition occurs becomes narrow er as t he [ 4] Fuller P. Linear rocket engine design fabrication t est ing[ R ] .
A IA A 73 1179, 1973.
base bleed increasing.
[5 ] Huang D. A erospike engine t echnology demonst rat ion for
space propulsion[ R] . AIA A 74 1080, 1974.
5 Conclusions
[ 6] St ephen A W , T imot hy R M . A base drag reduct ion experi
ment on th e X 33 linear aerospike SR 71 experiment
Ex perimental and numerical st udies are per
( LAS RE) flight program[ R] . A IAA 99 0277, 1999.
f ormed on a 6 cell t ile shaped aerospike nozzle, a 6 [ 7] T omit a T , Tamura H, Takahshi M . An experiment al evalua
cell aerospike nozzle w ith round to rect angle prima t ion of plug nozzle flow f ield[ R] . AIA A 96 2632, 1996.
ry nozzles and a 1 cell linear aerospike nozzle. Nu [ 8] T omit a T, Takahashi M , T amura H. Flow f ield of clust ered
plug nozzles[ R] . AIA A 97 3219, 1997.
merical altitude characterist ics of t hrust ef ficiencies [ 9] Onodera T , Tomit a T , Tamura H . Numerical investigat ion
and f low field st ruct ure are in good agreements w ith of t he f low f ield around linear aerospike nozzles [ R ] . A IAA
those obt ained in t he t est s. Good altit ude compen 99 2588, 1999.
[ 10] S akamot o H, T akahashi M , Sasaki M , e t al . A n experimen
sation capacity and hig h ef ficiency at design alt it ude
t al st udy on a 14 kN linear aerospike nozzle combustor [ R ] .
are achieved in the tests. T he numerical analysis A IA A 99 2761, 1999.
indicat es that t he inf erior performance of 6 cell [ 11] O nofri M . Plug nozzles: summery of f low f eat ures and engine
 February 2006 St udies on A erodynamic Behavior and Perf ormance of A erospik e N ozzles 9

perf ormance[ R] . A IA A 2002 0584, 2002. scheme V: a second order sequel t o Godunov s met hod [ J ] .
[ 12] D ai W Y , Liu Y , Cheng X C, et al . A erospike nozzle per Journal of Comput at ional Physics, 1979, 32: 101- 136.
formance st udy and it s cont our opt imizat ion [ R ] . A IAA [ 22] Shuen J S , Yoon S. N umerical st udy of chemically react ing
2001 3237, 2001. f low s using a low er upper symm etric successive overrelaxat ion
[ 13] D ai W Y , Liu Y , C heng X C, et al . Numerical and experi scheme [ J] . A IAA Journal, 1989, 27( 12) : 1752- 1760.
ment al study of tile shaped aerospike[ R] . A IAA 2002 4035, [ 23] T omit a T, T akahashi M , Onodera T , et al . Thrust loss due
2002. t o design of linear aerospike nozzles [ R] . A IA A 2000 3290,
[ 14] 王长辉, 刘宇, 廖云飞. 塞式喷管底部特性研究[ J] . 航空 2000.
学报, 2005, 26( 1) : 18- 22. Biographies:
W ang C H, Liu Y , Liao Y F. Invest igat ion on aerospike WANG Chang hui Born in 1977, he re
nozzle base behaviour [ J ] . A ct a A eronautica et Ast ronaut ica
ceived B. S. and Ph. D. degr ee in propul
S inica, 2005, 26( 1) : 18- 22. ( in Ch inese)
sion theory and engineer ing in aeronautics
[ 15] D ai W Y, Liu Y, Zhang Z K , et al . N umerical investigat ion
and astronautics from Beijing U niversity of
on linear aerospike nozzles[ R] . AIAA 2001 3568, 2001.
[ 16] 郭正, 刘君, 瞿章华. 塞式喷管流场的数值模拟[ J] . 推进
Aeronautics and Astro nautics ( BU AA ) in
技术, 2000, 21( 4) : 24- 26. 2000 and 2005 respectively and has done
G uo Z, Liu J, Q u Z H. N umerical simulat ion of plug nozzle postdoctor al r esearch work since then in
f low field [ J] . Journal of Propulsion Technology, 2000, 21 BU AA. His main resear ch inter est is in advanced r ocket en
( 4) : 24- 26. ( in Chinese) g ine technology. T el: ( 010 ) 82339444, E mail:
[ 17] 宋明德, 张涵信, 刘金合. 火箭发动机塞式喷管流场的 数 wang chang hui@ buaa. edu. cn
值模拟[ J] . 推进技术, 2000, 21( 5) : 19- 21. LIU Yu Bor n in 1963. He received the Ph. D. degree in
Song M D , Zhang H X , Liu J H. N um erical st udy on plug propulsion theory and engineer ing in aeronautics and astro
nozzle flow field [ J] . Journal of Propulsion Technology,
nautics in N orthwestern Polytechnical U niversity and now he
2000, 21( 5) : 19- 21. ( in Chinese)
is a professor of Beijing U niversit y of aeronautics and astro
[ 18] 李江, 郭唐稳, 马家欢. 气动塞式喷管底部二次流特性 的
nautics. His r esearch interests include fluid mechanics, aero
数值模拟[ J] . 推进技术, 2001, 22( 1) : 15- 18.
dynamics and advanced aer ospace propulsion. F rom 1998 to
Li J, Guo T W, M a J H . N um erical analysis of base bleed ef
f ect on aerospik e nozzle performance [ J] . Journal of Propul
1999, he w as a senior visiting scholar to Baltic T echnology
sion T echnology, 2001, 22( 1) : 15- 18. ( in Chinese) University. T el: ( 010) 82316540, E mail: liuyu @ buaa.
[ 19] D ai W Y , Liu Y , C heng X C, et al . Analyt ical and experi edu. cn
ment al studies of t ile shaped aerospike nozzl es [ J] . Journal of LIAO Yun fei Born in 1979, now a Ph. D. candidate stu
Propulsion and Pow er, 2003, 19( 4) : 640- 645. dent in Beijing U niversity of Aeronautics and Astronautics
[ 20] Lam C K G , Bremhorst K A . M odified f orm of t he k mod ( BU AA) , he received B. S. from BU AA in 2001. His main
el for predicting w all t urbulence [ J] . Journal of Fluids Engi research interest is in advanced nozzle of rocket engine. T el:
neering, 1981, 103( 3) : 456 460.
( 010) 82339444, E mail: liao yunfei@ sa. buaa. edu. cn
[ 21 ] Van Leer B. T ow ards t he ultimate conservat ive diff erence