Aerospace Science and Technology 7 (2003) 191–200 www.elsevier.com/locate/aescte

Instabilities and pressure oscillations in solid rocket motors ✩ Yves Fabignon a,∗ , Jöel Dupays a , Gérard Avalon a , Francois Vuillot b , Nicolas Lupoglazoff b , Grégoire Casalis c , Michel Prévost c a Office National d’Etudes et de Recherches Aérospatiales (ONERA), Fundamental&Applied Energetics Department, 29 Avenue de la Division Leclerc,

F-92322 Châtillon Cedex, France b Office National d’Etudes et de Recherches Aérospatiales (ONERA), CFD&Aeroacoustics Department, 29 Avenue de la Division Leclerc,

F-92322 Châtillon Cedex, France c Office National d’Etudes et de Recherches Aérospatiales (ONERA), Aerodynamics&Energetics Modeling Department, 29 Avenue de la Division Leclerc,

F-92322 Châtillon Cedex, France Received 2 May 2002; accepted 6 September 2002

Abstract The purpose of this paper is to give an overview of the main results obtained on instabilities and pressure oscillations in segmented solid rocket motors. A major part of this work was carried out in the framework of the ASSM and POP R&T programs supported by the French national space agency CNES during the last decade. ASSM is related to Aerodynamics of Segmented Solid Motors and POP for Pressure Oscillations Program for the Ariane 5 solid booster (P230). Due to the use of segmented technology for the P230 motor and the possible acoustic oscillations inside the motor chamber, anticipated at the beginning of the programs and confirmed later on static firing tests, the main scientific objective of the ASSM program was oriented towards the comprehension and the modeling of the vortex shedding phenomena that are supposed to be responsible of the pressure and thrust oscillations observed in the P230. POP program was started in order to obtain an experimental and numerical data base using subscale tests of 1/15th of the P230. After the description of the instabilities observed in the P230 solid rocket booster, the scientific approach of the ASSM program is detailed insisting on the validation of numerical tools in order to predict oscillation frequencies and amplitudes. The logic of work regarding POP program is also presented. The main section of this paper provides an overview of different results obtained in ASSM and POP programs to understand the mechanisms driving to the instabilities in solid rocket chamber. The most important recent result, inside ASSM and POP programs, was the discovery of the parietal vortex shedding and the role of aluminum combustion on instabilities. Together, these two mechanisms seem to be an important potential source of instabilities and provide a new vision of the P230 stability.  2002 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. Keywords: Solid propulsion; Instabilities; Aeroacoustics; CFD; Cold flow experiments; Propellant combustion

1. Introduction In propulsion systems, the coupling of sound waves and combustion has been recognized, since a long time, as a specific field of interest. Serious works were carried out on combustion instabilities because persistent developmental troubles in rocket engines could degrade the global motor performance. In the frame of solid propulsion, an historical perspective of combustion instabilities in USA can be found in reference [3].

✩

This article was presented at ODAS 2002.

* Corresponding author.

E-mail address: [email protected] (Y. Fabignon).

Instabilities are generally observed as oscillations having a time scale in the range of natural acoustic oscillations. Stability prevision of solid propellant rocket motors has been an active subject, both in the USA and in Europe. Synthesis reviews, on this subject, can be found in reference [8] and [16]. For technology reasons, large solid propellant space boosters, such as the RSRM (Redesigned Solid Rocket Motor) US Space Shuttle and Titan SRMs or European Ariane 5 P230 are made from segmented propellant grains from three to seven segments according to the motor versions. Earlier works in the USA [4,20] had shown that such grain segmentation conducted to low amplitude, but sustained pressure and thrust oscillations, on the first longitudinal acoustic mode frequencies. Although such oscillations do not jeop-

1270-9638/02/$ – see front matter  2002 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. doi:10.1016/S1270-9638(02)01194-X

192

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

ardize the launcher operation, they induce some penalties to the overall performance. The suppression or the control of such oscillations is then a meaningful objective for the program manager. In the case of Ariane 5 P230 solid rocket motor, early studies concluded that the risk of instabilities existed [25, 29] and that works had to be carried out to understand the physical mechanisms at the origin of this oscillatory behavior. These recent works pointed out the role of the vortex shedding as a source of acoustic energy inside solid rocket motors. The first suggestion of acoustic mode exitation by vortex shedding in solid rocket motors was made by Flandro and Jacobs [12] who emphasized the motor instability risk linked to the hydrodynamic instability of the mean flow sheared regions. It seems that the first evidence of vortex-shedding driven oscillations were encountered in motors with complex grain geometry, used for ballistic missiles upper stages [11] as unanticipated and unexplained oscillatory behavior. An other important conclusion of these recent works [25, 29] is, in such complex situations as the P230 motor, simplified methods (as the acoustic balance) cannot give reliable results in terms of stability previsions. Full numerical approaches must be used, in providing unprecedented insight into oscillatory flow fields and must become irreplaceable tools to predict motor stability, especially in geometrically complex situations. Of course, this tool needed to be adapted and then validated. This required some dedicated experiments and scientific program to accompany the numerical developments. It was the goal of the ASSM program. In parallel of ASSM program, POP program was started to give experimental and numerical data base in order to improve the knowledge of vortex shedding driven pressure oscillations phenomena in segmented solid rocket motors. The LP6 set-up was designed as an 1/15th representation of the

P230 solid rocket motor. Different configurations were studied, mainly centered, on the role of the inhibitors and intersegment cavity. After more than ten years of sustained research works, significant progresses have been made and results are available to build a detailed understanding of the mechanisms at work, with good hope to improve the operation of the P230 motors [28] and to accompany the evolution of this motor.

2. Pressure oscillations in the P230 solid rocket booster The Ariane 5 P230 solid rocket booster was designed, for technology reasons, in three segments. The propellant of the P230 is an AP/Al/HTPB propellant (Ammonium Perchlorate, Aluminium, Hydroxyl Termenated PolyButadiene) containing 68% of AP, 18% of Al and 14% of HTPB. Al is used as a solid-propellant additive to increase specific impulse. Fig. 1 shows a general view of the motor. The first segment (star-shaped grain segment S1 ), at the head end of the motor, was designed to give a maximum of thrust during the first 20 seconds of the total P230 burning time. Segment two (S2 ) is separated to segment three (S3 ) by an inhibitor ring. These two segments provide a pressure chamber between 4 MPa and 5 MPa until 110 seconds of the burning time (Fig. 2(a)). The first specimen, heavy case B1, was successfully tested on a static Bench (BEAP) in French Guyana on February 16, 1993 and the second one, M1, using a nominal lightweight case was also successfully fired on June 25, 1993. These tests were followed by three other development tests (M3, M4, M5). These development tests were then followed by two qualification tests Q1 and Q2. The second one was a qualification test for the complete booster. Three Ariane 5 qualification launches occurred on ELA3 launch

Fig. 1. General view of the Ariane 5 solid rocket motor (P230).

Fig. 2. Typical results for the P230.

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

193

behind pressure oscillations. The ASSM program objectives were thus to develop the necessary numerical tools with the necessary steps to allow their validation in situations relevant to the final application, which was the full scale of P230. Four milestones were initially identified as: (1) (2) (3) (4) Fig. 3. Waterfall plots of dsp contours for the P230.

site between 1996 and 1998. The successful completion of two of them leaded to begin Ariane 5 commercial operations during the second half of 1999 (504 flight). Figs. 2(a) and 2(b) illustrate one of the P230 static firing (Q2) main unsteady results. The black curve (Fig. 2(a)) represents the mean head end pressure chamber during the burning time. The red curve shows the head end pressure oscillations versus time. The zero-peak relative amplitudes are typically less than 0.5% for pressure oscillations and less than 5% for thrust oscillations [22]. Fig. 2(b) exhibits the pressure oscillations versus time for the first acoustic axial mode (20 Hz < f < 22 Hz) and for the second acoustic axial mode (40 Hz < f < 42 Hz) of the P230 chamber. It is also possible to plot the density spectral power (dsp) for the pressure oscillations versus frequency and e/emax (e/emax means the percentage of web time). This particular figure, called waterfall plots of dsp contours (Fig. 3), clearly shows that pressure oscillations mainly occur in the second part of the burning time (or web time) and on the first axial mode of the chamber.

3. Logic of activities and scientific program Obviously, previous experimental results, for the P230 solid rocket motor, were unknown before starting ASSM and POP programs. But such unstable behavior was anticipated due to the US experience [11–13]. Inhibitor rings or intersegment cavities were then identified as the possible sources of instabilities in a segmented motor. It was thus necessary to develop tools to understand the physical mechanisms and to predict the frequency and the amplitude of the pressure oscillations [27]. This was a quite challenging task, which was clearly beyond the current state of the art. The numerical tool was deliberately placed at the center of the research work due to the hostile character of the environment inside a solid rocket motor where only pressure measurements could be performed. Numerical simulations were the best way to acquire knowledge of details of the internal flowfield and of the physical mechanisms at work. In parallel, cold flow experiments [14] were also developed to study the type of flow encountered in solid propellant motors and to acquire fine measurements in order to analyze mechanisms

simulation of vortex-shedding driven flows, two-phase flow phenomenon, propellant combustion, turbulence development.

Obviously, the proper numerical tools should pass these milestones before they could be used in the full scale motor. This progressive approach, based on the parallel developments of the numerical models and the corresponding validation experiments, was one of the reasons of the final success of the enterprise, as described below. At the beginning of the ASSM program, two phases were identified [27]: Phase 1: demonstration of the capability of the numerical tools to reproduce vortex driven instabilities. This implied proper acoustic capabilities for the numerical schemes used to solve the Navier Stokes equations and proper treatment of boundary conditions that should behave correctly in the presence of acoustic and vorticity waves. This milestone was judged the most important one since it was the foundation for future developments. Several test-cases were built to validate the codes from simplest acoustically forced test case to the complex test case which was issued from an actual subscale firing of the LP6 motor. The progressive approach led to the designing of validation experiments, ranking from cold flow set-ups to actual subscale motors including LP6 motor. Only single phase flow was considered and propellant combustion response was excluded from this first phase. Phase 2: development and validation of the necessary physical models to include two-phase flow, propellant combustion response and turbulence. This phase was viewed as a decisive step enabling actual validations against actual firing experiments. These experimental subscale motors were developed in parallel and cross checking was the rule. This phase, once completed, would lead to the capability to apply the numerical tools to the full scale motor. However, at this stage, the progressive approach permitted to detect flaws in the initial reasoning and conducted to reassess the overall scientific program [26]. This will be presented in the following section. POP program was elaborated in the frame of the Ariane 5 program [21]. The goal was to look for an answer to the questions for which the P230 development test number was considered too small to be satisfactory such, as: (1) identification of parameters influencing oscillation levels, (2) dispersion evaluation,

194

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

Fig. 4. Sketch of the LP6 and C1x experimental set-ups.

Fig. 5. Illustration of the acoustic coupling phenomenon.

(3) passive control solution assessment. This program included both an experimental part, based on LP6 1/15th test bench (Fig. 4) and a numerical one, mainly devoted to full-scale transposition. The first part of this program was to find LP6 configurations that are representative enough to exhibit the elementary phenomena that are supposed to occur in a three-segment motor, but basic enough to permit confrontation with numerical simulations. For this last point a very good estimate of the internal geometry of the motor at the computed time must be known. Different configurations were thus defined based on the initial scenario leading to the internal flow oscillation driven by vortices shed from the second inhibitor ring at the forward end of the third segment S3 (Fig. 1). An important point of this program was to find pressure measurements very close to full-scale observations (Figs. 2, 3). A second part of this program was devoted to the use of this experimental tool to the analysis of dispersion causes and their prediction as well as the study of passive control solutions. This was completed by numerical activities for test understanding and full-scale transposition.

tex shedding phenomenon in test cases including simplified 2D planar conditions with chamfered edge propellant grain and more complex geometry as the LP6 configuration. All test cases have been very fruitful and provide a well defined and well-documented basis for the code evaluation. The physical mechanism behind the vortex shedding, and proposed by researchers in USA [4,12] can be described as a flow-acoustic coupling with a feedback loop (Fig. 5). Hydrodynamic instability is generated by shear layers in the gas flow due to the presence of annular restrictors and the acoustic feedback results from impingement of the vortices on the nozzle or other restrictors. The better understanding of the vortex shedding phenomenon [25] acquired during this first phase allowed to design a subscale rocket motor (Fig. 4) exhibiting natural vortex shedding instabilities (the “whistling motor or C1x in ONERA momenclature). Several firing tests were realized with excellent results. Such motor was the base of validation for turbulence models, propellant response and two-phase flow effects. Main results on pressure oscillations are reported in the three following subsections. 4.1. Effects of two-phase flow on pressure oscillations

4. Main results The first phase of ASSM program allowed to develop and validate CFD codes having the capability to capture the vor-

The use of aluminized propellant for the P230 solid rocket booster explains the interest to study the possible interaction of aluminum combustion and alumina oxide on pressure oscillations. An important effort was focused on

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

two-phase flow effects on acoustic oscillations in the framework of the ASSM program. A first step was to introduce a two-phase flow model in computer codes based on classical eulerian approach and to start validations on basic, but unsteady, test cases [9]. The second step was to study the effect of an inert particulate phase on the instabilities. A logic of experimental and numerical works was developed around the C1x subscale motor (Fig. 4) by using four composite propellants with different mass fractions and particle diameters. As expected, vortex shedding was identified on all of the firings. The experimental data, essentially pressure recordings, were used to validate aforementioned CFD codes. The main results deduced from this study are the following [10]: • The influence of the dispersed phase must be taken into account to achieve a reliable stability prediction; • The condensed and gaseous phase interactions are extremely complicated; some threshold effects can be identified; • The increase in the loading does not necessarily accentuate the attenuation of instabilities; • Particle size is a sensitive parameter that may act on the oscillation levels and on the selected instability mode. Following this progressive approach, a third step was also scheduled. The objective was to include, in CFD codes, an aluminum combustion model in order to study the impact of the heat released on pressure oscillations. This particular point will be discussed further in Section 4.5. 4.2. Effects of propellant response on pressure oscillations The term of propellant response is usually taken to mean the linearized (small amplitude) frequency response function for burning rate to harmonic pressure oscillations. Models are generally developed using a same analytical form for the linear response function [7]. This kind of model was introduced in CFD codes and tested on LP6 configurations. In that case, a very limited effect was reported (for the frequency range of interest) on pressure oscillations. But, on C1x subscale motor, reference [18] illustrates the rather good agreement obtained between simulations including measured propellant pressure coupled response function and firing tests. In conclusion, it turns out that the propellant response can have an influence on pressure oscillations if the frequency range of propellant response is close to the acoustic oscillations (depending on the motor size). In the P230 solid rocket motor, the frequency of pressure oscillations is low in comparison with the frequency range of propellant response. 4.3. Effects of turbulence on pressure oscillations The study of interactions between vortex shedding and turbulence is very complicated due to the difficulty to have a good physical representation between coherent structures

195

(from vortices) and small turbulent structures. Nevertheless, 2D calculations were carried out by using a semideterministic model [15] on C1x test case. In parallel, a 3D LES calculation was also performed [23]. For each case, results have shown small effects of turbulence on vortex shedding propagation. More recently, Reynolds Stress Model [6] was tested in an oscillatory channel flow with wall injection. Results confirmed that the level of turbulence remained low and did not seriously modify the natural instabilities. POP program [21] allowed to bring fruitful informations on pressure oscillation phenomena in segmented subscale motor. A complete data base for both experimental and numerical results was generated. During the experimental program, the propellant was changed for non-metallized HTPB composite propellant. Indeed, the use of Ariane 5 propellant (with aluminum) in the LP6 subscale motor has shown that it was not possible to obtain sufficient level of pressure oscillations. In the case of HTPB composite propellant, without aluminum, pressure measurements were very close to fullscale observations. The justification for transposition from 1/15th scale to full-scale remained an open point. Nevertheless, excellent reproductibility of results have been exhibited even in the case of successive propellant mixings. Dispersions are especially high near the end of the firing test which is a problem with respect of the analysis of the third oscillation peak encountered at full scale. Results also showed the great sensitivity to certain parameters such as the temperature or geometrical configuration. Different configurations were also tested, focusing the effort on the effect of the second inhibitor ring at the forward end of the third segment [24]. This second inhibitor ring was simulated in the LP6 subscale motor by using a metallic restrictor and different protrusions in the flow. It was found that second or third longitudinal modes were excited depending on the restrictor configuration. But, a motor configuration, without restrictor between the second and the third segment, was also tested. This configuration having no restrictors at all is clearly different from the other configurations and gave meanwhile the largest pressure oscillation amplitude on the first axial mode (Fig. 6). This important, at first sight, disconcerting result was also confirmed by numerical simulations [17]. Although satisfactory results were obtained from model development and validation tasks, it became apparent that the initial reasoning was flawed and needed reassessment. Indeed, the initial scenario implied an internal flow oscillation driven by vortices shed from the second inhibitor ring at the forward end of the third segment (Fig. 1). If this was found to actually occur, it was also found that it drove mostly higher (second or third) longitudinal modes, early into the firing and blocked the appearance of the first mode, late into the firing. This observation was the opposite of the experimental evidence obtained from full scale firings where first mode oscillations appeared in the second half of the firing (Figs. 2, 3). At the same time, the progressive research works in ASSM program permitted to propose a classifica-

196

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

Fig. 6. Typical set of experimental results obtained with the LP6 set-up (test 96-14 without restrictor).

Fig. 7. Three main situations of pressure oscillations obtained by using numerical simulations.

tion of vortex shedding situations [26]. Three main situations (Fig. 7) were described as: – Corner vortex-shedding (VSA) corresponding to simplest case where shedding was produced by an obstacleness shear layer created by a chamfered edge propellant. – Obstacle vortex-shedding (VSO) where the shear layer responsible for the shedding of vortices is created by a protruding obstacle (such an inhibitor ring). – Surface or parietal vortex-shedding (VSP) where the shedding results from an intrinsic instability of the internal flow, corresponding to the results observed in the LP6 experimental set-up. This last situation corresponds to a new mechanism [17], which needed further works to be completly understood. For this reason cold flow experiments were used in order to analyze flow stability induced by injection at walls. Moreover, due to the presence of vortices emitted from the burning surface, coupling with distributed aluminum combustion had to be incorporated in the scientific program as a probable mechanism that had to be modeled. Up to that point, aluminum combustion was thought to have a little impact on the oscillatory behavior, since it occured close to the burning surface, away from vortices shed by the protruding inhibitor ring. The next two subsections present the main results obtained on these specific fields of interest.

4.4. Cold flow experiments As the origin of the occurrence of acoustic motions without inhibitor ring inside was unknown, it was decided to carry out experimental investigations on the flow stability induced by injection at walls with a cold gas simulation. For lack of evident cause, the natural flow instability was suspected to initiate flow oscillations resulting in the generation of vortices at some distance of the head end region of the motor. Experiments were carried out with a two dimensional setup (Fig. 8), called VECLA, composed by a rectangular channel at the basis of which air is uniformly injected through a porous plate. During these experiments, the fluctuating velocity field inside the channel was explored using a hot wire installed on the top wall. These explorations were performed at different distances from the head end for different channel heights and injection velocities. Experimental results were compared to theoretical results based on computations of the amplification factors of natural frequencies obtained by an analytical analysis of the flow stability [5]. Fig. 8 shows a comparison of the experimental and theoretical velocity spectra obtained for a channel height of 30 mm at three longitudinal positions. On this figure, the amplitude peaks denote an amplification of the flow oscillations in a relative short range of frequencies. The frequency domain experimentally found coincides with that deduced from stability analysis results. Moreover, once the fit of a theoretical spectrum on an experimental one done, coincident amplitudes are

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

obtained for the spectra of the others longitudinal positions. These results confirmed the existence of flow oscillations induced by the natural flow instability as was anticipated at the beginning of the experimental study. For particular channel heights, it was observed that the previous amplification could be at the origin of a coupling mechanism between a parietal vortex shedding and the acoustic of the channel. Nevertheless, no evidence of the presence of vortices was proved since the measurements were not able to distinguish between velocity oscillations due to acoustic motions or vortices. A visualization of the flow inside the channel setup was then carried out by means of a PLIF (Planar Laser Induced Fluorescence) method applied on gaseous acetone that was mixed to air into two nonconsecutive prechambers of the channel. Fig. 9 shows a picture obtained in resonance conditions where it is clear that vortices, formed at the downstream part of the channel, are convected towards the exit of the channel. The picture is composed of several instantaneous pictures taken at different times and positions of the channel. For this reason, the fluorescent zones do not coincide between them from a position to the other. The

Fig. 8. VECLA set-up and experimental results obtained.

197

flow pattern shown on Fig. 9 was confirmed by numerical computations of the flow inside the VECLA’s channel with similar geometrical and injection velocity conditions than those experimentally chosen [2]. 4.5. Effects of aluminum combustion on acoustic coupling The third step, previously defined in Subsection 4.1, was to take into account the distributed combustion of aluminum particles that may occur in a significant portion of the chamber. This phenomenon can also affect appreciably combustion instabilities by acting as driving or damping mechanisms. The LP6 test-case, proned to develop parietal vortex shedding (VSP) driven instabilities, was chosen to study this effect [19]. To avoid multi-species simulations, droplets were supposed to vaporize and release heat to their vapor, following the classical d2 -law. Several two-phase flow simulations were carried out to study the effects of the droplet diameter on the oscillation behavior. Results from two of these computations are displayed on Fig. 10 and compared with a single-phase computation, serving as reference. To simulate the production of a condensed phase, droplet combustion was arbitrarily stopped at fixed diameters and, to simplify the study, aluminum droplets and alumina residues were supposed to have same properties. The first two-phase flow computation involved 30-µm droplets leading to a residue of 3 µm in size and the second involved 125 µm droplets leading to a residue of 60 µm in size. It is clear on the pressure recording, at the head end, that oscillation levels are amplified with the smaller droplets and damped with the larger droplets. Reactive two-phase flow computations have been also performed in the P230 booster. Details of this study are described in Ref. [19]. As for LP6’s simulations, oscillation levels were drastically increased by adding the droplet combustion. Without droplets, both VSO (obstacle vortex shedding) and VSP (parietal vortex shedding) coexist but the unsteady behavior seems to be dominated by the VSO. Spectral analysis of the pressure shows a rather wideband spectrum. With burning droplets, oscillation levels are drastically increased and the signal bandwidth is reduced. The combustion near the propellant surface reinforces the VSP and its coupling with the VSO. Results are coherent with data recording during static firings at Kourou. That seems to prove the predominant role of the VSP-distributed combustion pair. It remains now to use a validate aluminum combustion model because the classical d2 -law is probably unrepresentative in propellant gas environement.

Fig. 9. Picture of the flow obtained by PLIF technique.

198

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

Fig. 10. Vorticity plots and pressure recordings at the head end in the LP6 motor. From left to right: single-phase computation and two-phase computations with 30-µm droplets (residues of 3 µm in size) and 125-µm droplets (residues of 60 µm in size).

4.6. Effects of cavities on pressure oscillations The effect of cavities on pressure oscillations was not initially a theme of investigation in ASSM and POP programs, but it was decided to start some experiments, at first, by using a cold gas model having inhibitor rings [1]. Experimental results clearly indicate that different nozzle cavities can modify acoustical resonances and thus pressure oscillations. Later, and focusing the effort on the understanding of the parietal vortex shedding (VSP), it was decided to design a new subscale motor, called LP9, having a pure cylindrical port grain and without inhibitor rings (Fig. 11(a)). This motor was also designed so that acoustic coupling could occur based on the experience on LP6 subscale motor, numerical

simulations and analytical analysis of the flow stability. First results were rather encouraging with appearence of the parietal vortex shedding in the expected frequency range and the expected times into the firing (Fig. 12(a)). However, pressure oscillation amplitudes were not as high as expected. After that, it was decided to study the effect of cavities on pressure oscillations. Two different configurations were tested. The first one (Fig. 11(b)) initially contains a small cavity (intersegment cavity) in the propellant grain. Obviously, this initial small cavity leads to a larger cavity at the end of the firing. For the second configuration it was decided to modify the head end geometry of the motor by setting an inert cavity (Fig. 11(c)). Experimental results are very interesting and show, on the head end signal pressure, the role of cavi-

Fig. 11. Schematic of the different LP9 configurations.

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

199

Fig. 12. Experimental results obtained with the LP9 set-up for the different configurations.

ties on acoustic resonance. In the case of the inter-segment cavity, it was found (Fig. 12(b)) that the pressure oscillation amplitudes are lower in comparison with the nominal configuration (Fig. 12(a)). In the case of an inert cavity at the head end of the motor combined with the inter-segment cavity, pressure oscillation levels are higher (Fig. 12(c)) in comparison with the nominal configuration. The reasons of this behavior are not yet clear and need further works to explain the role of cavities on pressure oscillations.

The role of aluminum combustion on pressure oscillation levels seems to be an important parameter but it turns out that the modeling used for the computations is not validated. Further works are also needed in order to find a physical explanation of this particular phenomenon and to predict reliable level of instabilities. These different main questions should be addressed, using the available tools that were developed during ASSM and POP programs.

5. Conclusions

Acknowledgements

This paper allowed to give an overview of the main results obtained on pressure oscillations in a large solid rocket motor in the frame of ASSM and POP programs. The scientific approach of ASSM program was based on the understanding of the physical phenomena and their modelisation. For us, it is the best way to carry out applied research. Milestones are also important in order to evaluate the progresses, as well as the difficulties, and also to reassess when necessary. ASSM and POP programs allowed to identify three basic mechanisms capable of driving pressure oscillations. But the most important result was the discovery of the parietal vortex shedding (VSP) and the role of aluminum combustion on pressure oscillations. Together, these two mechanisms seem to be an important potential source of instabilities and provide a new vision of the P230 stability. This major result would have not been possible without numerical simulations and experimental data coming from cold-flow simulations and firing tests. However, it turns out that some difficulties remain and need further works. The role of turbulence on the acoustic resonance has not been really identified. One unanswered question is what is the role of the aft segment inhibitor on the overall turbulence developement. Other questions that were set aside concern the role of the combustion noise on the parietal vortex shedding triggering and the role of cavities on the acoustic resonance. An other point should be investigated and concerns the full-scale transposition of experimental results from firing tests with subscale motors.

Much of this work was performed within the ASSM (Aerodynamics of Segmented Solid Motor) research program coordinated by ONERA, and POP (Programme d’Oscillations de Pression) program, both conducted with the financial support of CNES (Centre National d’Etudes Spatiales), Direction des Lanceurs. Authors express their thanks to P. Kuentzmann who supported and encouraged this synthesis, G. Lengellé, I. Dubois for many fruitful discussions and Société Nationale des Poudres et Explosifs (SNPE) for its large cooperative effort on this subject.

References [1] J. Anthoine, J.M. Buchlin, A. Hirschberg, Effect of nozzle cavity on resonance in large SRM: Theoretical modeling, J. Propulsion and Power 18 (2) (2002) 304–311. [2] G. Avalon, B. Ugurtas, F. Grisch, A. Bresson, Numerical computations and visualization tests of the flow inside a cold gas simulation with characterization of a parietal vortex shedding, AIAA Paper 2000-3387, July 2000. [3] F.S. Blomshield, Historical perspective of combustion instability in motors: case studies, AIAA Paper 2001-3875, July 2001. [4] R.S. Brown, R. Dunlap, S.W. Young, R.C. Waugh, Vortex shedding as a source of acoustic energy in segmented solid rockets, J. Spacecraft and Rocket 18 (4) (1981) 312–319. [5] G. Casalis, G. Avalon, J.-Ph. Pineau, Spatial instability of planar channel flow with fluid injection through porous walls, Phys. Fluids 10 (1998). [6] B. Chaouat, R. Schiestel, Reynolds stress transport modelling for steady and unsteady channel flows with wall injection, in: Second

200

[7] [8]

[9]

[10]

[11] [12] [13] [14]

[15] [16] [17]

[18]

[19]

Y. Fabignon et al. / Aerospace Science and Technology 7 (2003) 191–200

International Symposium on Turbulence and Shear Flow Phenomena, Stockholm, 2001. F.E.C. Culick, A review of calculations for unsteady burning of a solid propellant, AIAA J. 6 (12) (1968) 2241–2254. F.E.C. Culick, Combustion instabilities: mating dance of chemical, combustion, and combustor dynamics, AIAA Paper 00-3178, July 2000. J. Dupays, M. Prévost, P. Tarrin, F. Vuillot, Effects of particulate phase on vortex shedding driven oscillations in solid rocket motors, AIAA Paper 96-3248, July 1996. J. Dupays, Y. Fabignon, P. Villedieu, G. Lavergne, J.L. Estivalezes, Some aspects of two-phase flows in solid propellant rocket motors, in: V. Yang, T.B. Brill, W.Z. Ren (Eds.), Progress in Astronautics and Aeronautics in: Solid Propellant Chemistry, Combustion, and Motor Interior Ballistics, Vol. 185, AIAA, 2000, pp. 859–883. G.A. Flandro, AGARD consulting mission to ONERA, September 16– 19, 1986. G.A. Flandro, H.R. Jacobs, Vortex generated sound in cavities, AIAA Paper 73-1014, October 1973. A. Flatau, Vortex driven sound in a cylindrical cavity, Ph.D. Dissertation, The University of Utah, March 1990. J.F. Guéry, G. Avalon, F. Plourde, J. Anthoine, B. Platet, Use of cold flow experiments in the ASSM program, lessons and results, in: 2nd European Conference on Launcher Technology, Space Solid Propulsion, Rome, Italy, 2000. A. Kourta, Acoustic-mean flow interaction and vortex shedding in solid rocket motors, J. Propulsion and Power 12 (1995) 371–376. P. Kuentzmann, Combustion Instabilities, AGARD LS 180, September 1991. N. Lupoglazoff, F. Vuillot, Parietal vortex-shedding as a cause of instability for long solid propellant motors. Numerical simulations and comparisons with firing tests, AIAA Paper 96-0761, January 1996. N. Lupoglazoff, F. Vuillot, Simulations of solid propellant rocket motors instability including propellant combustion response, in: 6th Internat. Congress on Sound and Vibration, Lingby, Denmark, 1999. N. Lupoglazoff, F. Vuillot, J. Dupays, Y. Fabignon, Numerical simulations of the unsteady flow inside Ariane 5 P230 SRM booster

[20]

[21]

[22]

[23]

[24]

[25] [26]

[27]

[28]

[29]

with burning aluminum particles, in: 2nd European Conference on Launcher Technology, Rome, Italy, 2000. D.R. Mason, S.L. Folkman, M.A. Behring, Thrust oscillations of the space shuttle solid rocket booster motor during static tests, AIAA Paper 79-1138, June 1979. M. Prévost, Y. Dommee, J. Maunory, J.C. Traineau, F. Vuillot, P. Marion Duval, On the representativity of small scale motor tests, in: 2nd European Conference on Launcher Technology, Space Solid Propulsion, Rome, Italy, 2000. S. Scippa, P. Pascal, F. Zanier, Ariane 5 MPS – chamber pressure oscillations full scale firings results analysis and further studies, AIAA Paper 94-3068, June 1994. J.H. Silvestrini, P. Comte, M. Lesieur, Simulation des Grandes Echelles; Application aux Moteurs à Propergol Solide Segmentés, in: Proceedings of Journées R&T CNES-ONERA, CNES, Paris, France, 1995. J.C. Traineau, M. Prévost, F. Vuillot, P. Le Breton, J. Cuny, N. Preioni, R. Bec, A subscale test program to assess the vortex shedding driven instabilities in segmented solid rocket motors, AIAA Paper 97-3247, July 1997. F. Vuillot, Vortex-shedding phenomena in solid rocket motors, J. Propulsion and Power 11 (4) (1995) 626–639. F. Vuillot, Point sur les recherches relatives à la stabilité de fonctionnement du MPS-P230 d’Ariane 5, in: 3ème colloque R&T ENSMA/ CNES/ONERA sur les écoulements internes en propulsion solide, Poitiers, France, 1998. F. Vuillot, P. Kuentzmann, Programme de R&T ASSM (Aerodynamics of Segmented Solid Motors), in: Colloque CNES/ONERA/CNRS sur les écoulement propulsifs dans les systèmes de transport spatial, Bordeaux, France, 1995. F. Vuillot, G. Casalis, Recent advances on the stability of large segmented space boosters, in: 2nd European Conference on Launcher Technology, Space Solid Propulsion, Rome, Italy, 2000. F. Vuillot, P.Y. Tissier, R. De Amicis, Prédiction de stabilité des gros moteurs segmentés à propergol solide, in: AAAF, 5ème symposium international sur la propulsion dans les transports spatiaux, Carré des Sciences, Paris, France, 1996, pp. 2.13–2.27.