Professional Documents
Culture Documents
Influence of Thermal Inhibitor Position and Temper
Influence of Thermal Inhibitor Position and Temper
net/publication/257440387
CITATIONS READS
7 351
6 authors, including:
Junwei Li Qingqing Ye
Beijing Institute of Technology Zhejiang University
120 PUBLICATIONS 872 CITATIONS 21 PUBLICATIONS 145 CITATIONS
Some of the authors of this publication are also working on these related projects:
Combustion instability of solid rocket with Aluminized Solid Propellants View project
All content following this page was uploaded by Junwei Li on 19 August 2014.
a
School of Aerospace Engineering, Beijing Institute of Technology, Beijing 100081, China
b
China Academy of Space Technology, Beijing 100094, China
KEYWORDS Abstract Vortex-acoustic coupling is one of the most important potential sources of combustion
Combustion instability; instability in solid rocket motors (SRMs). Based on the Von Karman Institute for Fluid Dynamics
Pressure oscillation; (VKI) experimental motor, the influence of the thermal inhibitor position and temperature on vor-
Solid rocket motor; tex-shedding-driven pressure oscillations is numerically studied via the large eddy simulation (LES)
Vortex-acoustic coupling; method. The simulation results demonstrate that vortex shedding is a periodic process and its accu-
Vortex shedding rate frequency can be numerically obtained. Acoustic modes could be easily excited by vortex shed-
ding. The vortex shedding frequency and second acoustic frequency dominate the pressure
oscillation characteristics in the chamber. Thermal inhibitor position and gas temperature have little
effect on vortex shedding frequency, but have great impact on pressure oscillation amplitude. Pres-
sure amplitude is much higher when the thermal inhibitor locates at the acoustic velocity anti-nodes.
The farther the thermal inhibitor is to the nozzle head, the more vortex energy would be dissipated
by the turbulence. Therefore, the vortex shedding amplitude at the second acoustic velocity anti-
node near 3/4L (L is chamber length) is larger than those of others. Besides, the natural acoustic
frequencies increase with the gas temperature. As the vortex shedding frequency departs from
the natural acoustic frequency, the vortex-acoustic feedback loop is decoupled. Consequently, both
the vortex shedding and acoustic amplitudes decrease rapidly.
ª 2013 Production and hosting by Elsevier Ltd. on behalf of CSAA & BUAA.
1. Introduction
ding frequency is close to one of the intrinsic acoustic frequen- rocket motor designers to minimize the pressure oscillation.
cies. According to the induced factors, three kinds of vortex On the other hand, numerous studies have focused on vor-
shedding are defined by researchers, namely, obstacle vortex tex-shedding-driven pressure oscillations under the condition
shedding (OVS), corner vortex shedding (AVS), and parietal that vortex shedding frequency is close to acoustic frequency.
vortex shedding (PVS).6 Titan 34D SRM,7 Titan IV SRMU Little work has been mentioned when the two frequencies
(SRM Upgrade),8 Space Shuttle RSRM (Re-designed depart from each other. In this paper, the acoustic frequency
SRM),9 and Ariane 5 SRM10 have all experienced low ampli- is changed by altering the gas temperature, because the natural
tude, but sustained periodic pressure and thrust oscillations. acoustic frequency highly depends on the gas temperature
For technical concerns, these boosters were equipped with when the motor structure is fixed. The present work is based
large segmented motors and the grains were separated by ther- on a VKI subscale motor, which is a 1/30-scale axisymmetric
mal inhibitors. This kind of structure can easily result in flow cold flow model of Ariane 5 booster. Numerical simulations
separation and thus obstacle vortex shedding. Both the are firstly carried out to explore the effects of thermal inhibitor
European and US researchers believed that this kind of com- position on vortex-acoustic coupling characteristics (including
bustion instability was attributed to the obstacle vortex flow field characteristics, pressure oscillation frequency and
shedding. amplitude). Then, the vortex-shedding-driven pressure oscilla-
The basic mechanism of vortex-acoustic coupling was de- tion characteristic is studied when the vortex shedding fre-
fined in the early investigation by Powell11 for the simple case quency deviates from the natural acoustic frequency, by
of a planar jet impinging upon a leading edge. Howe12 pro- changing the gas temperature.
posed a generalization to internal flows. According to the Pow-
ell–Howe theory of vortex sound, acoustic is generated by the
Coriolis force due to the vortex trajectory intersecting with 2. Theoretical basis
acoustic streamlines. The vortex-acoustic coupling research
work has been carried out for nearly 40 years since researchers
The chamber can be regarded as a self-excited acoustic oscil-
realized that it was one of the potential sources of combustion
lation system. The natural acoustic mode would be excited
instability in solid rocket motors (SRMs). Flandro and Ja-
when the chamber suffers from some weak perturbation.1
cobs13 firstly emphasized that vortex-acoustic coupling was
The collision of vortex to nozzle head is the right small per-
the potential cause of the instability in rocket motors. Besides,
turbation that can lead to pressure oscillations in rocket
they linked pressure oscillations to hydrodynamic instability of
chamber. When the vortex shedding frequency matches the
the sheared regions. Culick and Magiawala14 as well as Dunlap
fundamental acoustic frequency, serious combustion instabil-
and Brown15 experimentally demonstrated that the coupling of
ity would occur. From this point of view, vortex-acoustic
vortex shedding frequency with those of acoustic modes in the
coupling is one kind of non-linear acoustic combustion
chamber could lead to pressure oscillations. Anthoine and his
instability. In order to explore the relationship between vor-
coworkers carried out a series of theoretical,16 experimental,17
tices and acoustic field, both the acoustic characteristics in
and numerical18 studies based on Von Karman Institute for
the chamber and the vortex theory are needed to understand
Fluid Dynamics (VKI) motor. The results demonstrated that
vortex-acoustic coupling.
pressure oscillations could be excited when the vortices shed
by the thermal inhibitor passing in front of the submerged cav-
ity, and the oscillation amplitude increased linearly with the 2.1. Acoustic characteristic
submerged cavity volume. In addition, Brown et al.,19 Flatau
and Moorhem,20 and Scippa et al.10 conducted cold gas exper-
The rocket chamber is closed at the forward end. On the oppo-
iments on the subscale motors of Titan 34D, Space Shuttle,
site, the sonic condition at the nozzle throat prevents any
and Ariane 5 SRM, respectively. Similar conclusions, such as
acoustic perturbations to propagate through. Therefore, the
pressure oscillations were due to vortex shedding caused by
chamber can be assimilated to a closed–closed tube from
the thermal inhibitor, were obtained.
the acoustic point of view. As a sound wave propagates in
For the purpose of assessing the combustion instability of
the chamber, it disturbs the fluid from its mean state, which
Ariane 5 solid booster, both experimental and numerical work
can be expressed as: p ¼ p þ p0 , q ¼ q
þ q0 , where p, q are pres-
was carried out in the framework of the aerodynamics of seg-
sure and density, ‘‘-’’ and ‘‘ 0 ’’ represent Reynolds-averaged
mented solid motors (ASSM) program and the pressure oscil-
quantity and fluctuating component, respectively, p0 = p 1,
lation program (POP) supported by French National Space
q 1. According to the classical acoustic theory,21 the
q0 =
Agency CENS during the last decade.6 Vuillot3 illustrated
acoustic wave equation is expressed as:
the vortex impinging mechanism and numerically simulated
the coupling between the vortex and acoustic waves in a sub-
1 @ 2 p0
scale motor of Ariane 5 P230 with a thermal inhibitor pro- $2 p0 ¼ 0 ð1Þ
truded in the flow. Kourta4 numerically studied the vortex a2 @t2
shedding phenomenon in segmented solid rocket motors and where a is the sound speed. To simplify the solution of Eq. (1),
depicted the vortex paring progress. In fact, the thermal inhib- the three-dimensional sound wave is reduced to a one-dimen-
itor plays an important role in the vortex-shedding-driven sional standing plane wave. Then, the general solution is writ-
pressure oscillations. Researchers have comprehensively stud- ten as:
ied the characteristics of the thermal inhibitor, including mate- 8 0
< p ¼ Aei2pfx=a þ Bei2pfx=a
rial, shape, and size. However, the influence of the thermal
1 ð2Þ
inhibitor position has not taken into consideration. It is very : u0 ¼ ðAei2pfx=a Bei2pfx=a Þ
important to select an optimized inhibitor position for solid q0 a
546 W. Su et al.
where u0 is the oscillation components of velocity, f is the oscil- The vortex shedding frequency would be deduced from Eq.
lation frequency, q0 is gas density. (5):
As mentioned before, that solution has to satisfy the
U ma
boundary conditions of the closed-closed tube at both ends fvortex ¼ ð6Þ
of the chamber: lloop Ma=ð1 MaÞ þ 1=k
u0 ¼ 0; jp0 j ¼ p0max ð3Þ where Ma is the Mach number, and a is the correction factor
of the delay time Dt, which is chosen between 0 and 0.25.
Then, the longitudinal acoustic mode frequencies of the
chamber can be yielded from Eqs. (2) and (3): 3. Numerical scheme and computational model
pffiffiffiffiffiffiffiffiffi
na n cRT
f¼ ¼ ð4Þ 3.1. Governing equations and discrete format
2L 2L
where n = 1,2, . . . is the acoustic mode number, L is the cham-
ber length, T is chamber temperature, R is gas constant, c is Large eddy simulation (LES) predictions have less modeling
specific heat ratio. When the motor structure is fixed, the nat- errors compared with the Reynolds-averaged Navier–Stokes
ural acoustic frequency mainly depends on gas temperature. In (RANS) and require less computational time compared with
engineering, the rocket chamber is also simplified as a cylindri- the direct numerical simulation (DNS). Therefore, LES has
cal tube. Therefore, Eq. (4) provides an estimate of the longi- been widely used to simulate detailed flows and combustion
tudinal acoustic mode frequencies of the rocket chamber. in solid rocket motors. The main idea of LES is that large ed-
dies are computed directly, while the effect of small scale mo-
2.2. Vortex-acoustic coupling theory tions on the resolved scales is modeled. A spatial filter G is
used to decompose the flow variables into resolved and unre-
solved scales23:
The flow field in the chamber is prone to hydrodynamic insta- (
bility. A sudden change of structure or thermal inhibitor could ~ tÞ ¼ f~r ðx; tÞ þ f~s ðx; tÞ
fðx;
cause flow separation, hereafter, vortex shedding would occur. R ð7Þ
f~r ðx; tÞ ¼ D Gðx x0 ; DÞfðx ~ 0 ; tÞd3 x0
Serious pressure oscillation would be excited when the vortex
shedding frequency is close to one of the acoustic frequency, where ‘‘’’refers to the Favre average, D is the computational
and a closed loop of vortex-acoustic coupling would be domain, D is the computational mesh size that determines the
formed. The mechanism of acoustic resonance was identified size and structure of the subgrid scales, and f stands for any
by Rossiter22 for airflow over rectangular cavities in 1964, flow property, say, q, p, u, T. The superscripts ‘‘r’’ and ‘‘s’’ re-
who first developed an empirical formula for predicting cavity fer to the resolved and subgrid scales of flow properties.
flow resonance frequencies. Dotson, et al.,8 studied the vortex The Favre-averaging is used to simplify the governing equa-
shedding pressure oscillation phenomenon in Titan IV SRMU tions for the compressible flows, which is described as:
by using the Rossiter’s mechanism for reference. He described
qf
the vortex-acoustic coupling mechanism as a feedback loop f~ ¼ ð8Þ
consisting of four steps, which were the generation of vortices, q
vortex shedding, impingement of the vortex, and acoustic feed- where ‘‘-’’ refers to the Reynolds average.
back. The mechanism is schematically illustrated in Fig. 1. Then, the filtered mass, momentum, and energy equations
The feedback loop mechanism could be mathematically de- are written as:
scribed as22:
q @
@ qu~i
mt ¼ lloop =ðkUÞ þ lloop =ða UÞ þ Dt ð5Þ þ ¼0 ð9Þ
@t @xi
where t is the vortex shedding period, U the upstream velocity, @ @ @ p @
ð
qu~i Þ þ ð
qu~i u~j Þ ¼ þ ~ij ssgs
r ij ð10Þ
lloop the distance between the shedding point and the impinge- @t @xj @xi @xj
ment point, m the vortex number within lloop, and Dt a small ~ij and ssgs
where r ij represent the molecular viscosity stress tensor
time increment. The value k is a dimensionless empirical con- and total subgrid stress tensor, respectively.
stant between 0.57 and 0.63.
@ @ @
qe~Þ þ
ð qu~i e~Þ ¼
ð ~ij Hsgs
pu~i qi þ u~i r sgs
i Hi ð11Þ
@t @xi @xi
where e is total specific energy, Hsgsi and Hsgsi are subgrid en-
thalpy flux and subgrid viscous work, respectively.
The molecular viscosity stress tensor ~
dij and the heat flux qi
are given as:
@ u~i @ u~j 2 @ u~k
~ij ¼ l
r þ l dij ð12Þ
@xj @xi 3 @xk
cp l @ T~
qi ¼ ð13Þ
Pr @xi
where l is dynamic viscosity, cp is specific heat at constant
Fig. 1 Vortex-acoustic coupling mechanism. pressure, Pr is Prandtl number.
Influence of thermal inhibitor position and temperature on vortex-shedding-driven pressure oscillations 547
The governing equations are finally supplemented with the and a static pressure of 0.178 MPa. The nozzle exit is defined
ideal gas state equation: as a pressure outlet, and the cross-section parameters can be
extrapolated from the internal ones. The wall boundary is
p ¼ qRT ð14Þ characterized by a no-slip condition. Finally, the computa-
The wall-adapting local eddy-viscosity (WALE) model is tional domain is initialized with an entrance velocity of
selected for the closure of subgrid stress tensor ssgs
ij , subgrid en-
30 m/s. For another hand, nine cases with the inlet tempera-
thalpy flux Hsgs sgs
i ; and subgrid viscous work Hi ; respectively.
ture Ti varying from 200 K to 3000 K are studied to explore
To avoid the unphysical oscillations in the solution fields, the temperature effect under the thermal inhibitor position
the bounded central differencing (BCD) scheme is chosen for of Case 13. Upstream velocity is the key factor to determine
mass and momentum equations, while the power law scheme the vortex shedding frequency. Therefore, the entrance velocity
for energy equation. The implicit second order method is used is kept constant under different temperatures. In order to
in the unsteady solver, and the time step is 5 · 106 s based on maintain the same static pressure and entrance velocity, the
a Courant–Friedrichs–Lewy (CFL) number of 0.6. throat radius should be changed by ri = r0(T0/Ti)1/4. Further-
more, the boundary condition of inlet mass flux is replaced by
3.2. Computational domain and boundary conditions m_ i ¼ m_ 0 T0 =Ti , accordingly.
The numerical study is based on the VKI experimental mo- 4. Validation of numerical methods
tor,18 which is a 1/30-scale axisymmetric cold flow model of
Ariane 5 booster. The computational domain is simplified into The numerical simulation of the internal solid rocket motor
an axisymmetry as shown in Fig. 2. A thermal inhibitor is lo- has been performed with a commercial computational fluid
cated at 0.29 m in the axial position, with a total length of dynamics (CFD) code, Fluent. To ensure the accuracy of
0.393 m and a throat radius r0 of 0.015 m. The thickness of numerical methods, mesh sensitivity analysis is firstly studied
the thermal inhibitor is 3 mm. The distance between the ther- to choose a moderate mesh number. Then, a comparison with
mal inhibitor and the nozzle head (AB in Fig. 2) is defined Anthoine’s experimental and numerical results is carried out to
as the vortex-acoustic feedback loop, which is also the lloop assess the reliability of the numerical scheme.
in Eq. (6).
The thermal inhibitor position and gas temperature are the 4.1. Mesh sensitivity analysis
main objects in our research work. Both of them have different
effects on vortex-shedding-driven pressure oscillation charac- Resolving only the large eddies allows one to use much coarser
teristics. For one hand, 16 cases defined in Table 1 are numer- mesh size in LES than in DNS. However, LES still requires
ically studied to explore the inhibitor position effects. The substantially finer meshes than those typically used for RANS
thermal inhibitor position in Case 13 is rightly shown in calculations. Therefore, selecting a moderate mesh size is very
Fig. 2. All the cases have the same boundary conditions for important for both saving the computing time and obtaining a
comparison. The cold gas is simplified as ideal gas whose relative accurate result.
parameters are: T0 = 285 K, specific heat ratio c = 1.4, Seven different meshes have been employed for the analysis,
dynamic viscosity l = 1.7894 · 105 kg/(mÆs), and specific ranging from 20000 to 140000 by an increment of 20000. The
heat at constant pressure cp = 1006.43 J/(kgÆK). Cold gas is in- mesh number of 20000, 80000, and 140000 represents coarse,
jected axially with a constant mass flux m_ 0 of 66:3 kg=ðm2 sÞ medium, and fine meshes, respectively. A virtual pressure mon-
itor (Point 1) is set at the head end in the computational do-
main to record the pressure changing history, as shown in
Fig. 2. After a fast Fourier transform (FFT) on the pressure
oscillation results, pressure oscillation frequencies and ampli-
tudes are then obtained. The most excited frequency and its
amplitude are selected for the mesh sensitivity analysis. The
changing of frequency and non-dimensional pressure oscilla-
(a) T=800 K
6. Conclusions
(4) As the vortex shedding frequency departs from the nat- 13. Flandro GA, Jacobs HR. Vortex-generated sound in cavities. In:
ural acoustic frequency, the vortex-acoustic feedback AIAA aero-acoustics conference; 1973 Oct 15–17; Seattle, Wash-
loop is decoupled. Thereby, both the vortex shedding ington; 1973.
and acoustic amplitudes decrease rapidly. 14. Culick FEC, Magiawala K. Excitation of acoustic modes in a
chamber by vortex shedding. J Sound Vib 1979;64(3):455–7.
15. Dunlap R, Brown RS. Exploratory experiments on acoustic
oscillations driven by periodic vortex shedding. AIAA J
Acknowledgement
1981;19(3):408–9.
16. Anthoine J, Buchlin JM, Hirschberg A. Effect of nozzle cavity on
This study was supported by the National Natural Science resonance in large SRM: theoretical modeling. J Propul Power
Foundation of China (Grant No. 51076015). 2002;18(2):304–11.
17. Anthoine J, Lema MR. Passive control of pressure oscillations in
References solid rocket motors: cold-flow experiments. J Propul Power
2009;25(3):792–800.
1. Sun WS. Combustion instabilities in solid rocket motor. Bei- 18. Anthoine J, Buchlin JM, Guery JF. Effect of nozzle cavity on
jing: Beijing Institute of Technology Press; 1987 [Chinese]. resonance in large SRM: numerical simulations. J Propul Power
2. Blomshield FS. Lessons learned in solid rocket combustion 2003;19(3):374–84.
instability. In: 43rd AIAA/ASME/SAE/ASEE joint propulsion 19. Brown RS, Dunlap R, Young SW, Waugh RC. Vortex shedding
conference and exhibit; 2007 Jul 8–11; Cincinnati; 2007. as a source of acoustic energy in segmented solid rockets. J
3. Vuillot F. Vortex-shedding phenomena in solid rocket motors. J Spacecraft Rockets 1981;18(4):312–9.
Propul Power 1995;11(4):626–39. 20. Flatau A, van Moorhem W. Prediction of vortex shedding
4. Kourta A. Vortex shedding in segmented solid rocket motors. J responses in segmented solid rocket motors. In: 26th AIAA/
Propul Power 1996;12(2):371–6. ASME/SAE/ASEE joint propulsion conference and exhibit; 1990
5. Zhang Q, Li JW, Wang WC, Wang NF. Numerical analysis on Jul 16–18; Orlando, Florida; 1990.
oscillation characteristics in a tailpipe nozzle solid rocket motor. J 21. Ma DY. Modern acoustic theory foundation. 3rd ed. Beijing: Sci-
Spacecraft Rockets 2011;48(1):103–9. ence Press; 2006 [Chinese].
6. Fabignon Y, Dupays J, Avalon G, Vuillot F, Lupoglazoff N, 22. Rossiter JE. Wind-tunnel experiments on the flow over rectanglar
Casalis G, et al. Instabilities and pressure oscillations in solid cavities at supersonic and transonic speeds. Reports and Memo-
rocket motors. Aerosp Sci Technol 2003;7(3):191–200. Randa. British A.R.C.; 1964. No. 3438.
7. Chang IS. Investigation of space launch vehicle catastrophic 23. FLUENT Version 6.3.26 User’s Guide.
failures. J Spacecraft Rockets 1996;33(2):198–205. 24. Karthik B, Chakravarthy SR, Sujith RI. The effect of orifice
8. Dotson KW, Koshigoe S, Pace KK. Vortex shedding in a large thickness on the vortex-acoustic interaction in a duct. In: Seventh
solid rocket motor without inhibitors at the segment interfaces. J AIAA/CEAS aeroacoustics conference; 2001 May 28–30; Maas-
Propul Power 1997;13(2):197–206. tricht, Netherlands; 2001.
9. Mason DR, Morstadt RA, Cannon SM, Gross EG, Nielsen DB. 25. Flatau A, van Moorhem W. Cold-flow investigation of vortex-
Pressure oscillations and structural vibrations in Space Shuttle shedding induced sound amplitude variations. In: 33rd aerospace
RSRM and ETM-3 Motors. In: 40th AIAA/ASME/SAE/ASEE sciences meeting and exhibit; 1995 Jan 9–12; Reno, NV; 1995.
joint propulsion conference and exhibit; 2004 Jul 11–14; Florida; 26. Zhang Q, Wei ZJ, Su WX, Li JW, Wang NF. Theoretical and
2004. numerical analysis on thrust oscillation characteristics in solid
10. Scippa S, Pascal P, Zainer F. Ariane-5 MPS: chamber pressure rocket motors. J Propul Power 2012;28(2):312–22.
oscillations full scale firing results analysis and further studies. In:
30th AIAA/ASME/SAE/ASEE joint propulsion conference and Su Wanxing is a Ph.D. student in the School of Aerospace Engineering
exhibit; 1994 Jun 27–29; Indianapolis; 1994. at Beijing Institute of Technology. His areas of research include
11. Powell A. Theory of vortex sound. J Acoust Soc Am combustion instability, aero-acoustics, and solid rocket motors.
1964;36(1):177–95.
12. Howe MS. Contributions to the theory of aerodynamic sound,
Li Shipeng is an associate professor and Ph.D. advisor in the School of
with application to excess jet noise and the theory of the flute. J
Aerospace Engineering at Beijing Institute of Technology. His current
Fluid Mech 1975;71(4):625–73.
research interest is solid rocket motor technology.