Francois Vuillot

Instabilities and pressure oscillations in solid rocket motors

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.

Numerical simulation of vortex shedding phenomenon in 2D test case solid rocket motors

Segmented Solid Rocket Motors tend to develop unpredicted thrust and uressure oscillations. linked to a criodic vortex shcdding phcnomenon. Aspart of Arianc

Comprehensive 3D Unsteady Simulations of Subsonic and Supersonic Hot Jet Flow-Fields: Part 1: Aerodynamic Analysis

PUO MPS solid motor stability assessmcnt. a research effort was initiated and aimed at il full numerical simulation of the unsteady, compressible internal flow. The objective of the present work is to demonstrate the feasibility of a direct numerical simulation of the vortex shedding henomenon on test case motors. The computer equations, by means of an expicit predictor/corrector Mc Cormack scheme with an improved (Jameson type) artificial viscosity algorithm. A r p e r test ,case motor was devised by means of a hydro ynamic stability analy+ of mean flow shear layer. Computations done wth optimized artificial viscosity terms gave rise to a marked self excited vortex shedding phenomenon of constant am litude close to the second axial mode frequency, in v excited phenomenon was observefwith both the Euler and Navler-Stokes codes. Moreover planar com utations artificial viscosity added, and gave rise to identical vortex shedding henomenon. First comparisons with Flandro’s codes, so P ve the 2D unstead , Euler or Navier-Stokes bot R axisymmetrie and planar confi urations. Similar self have been performed with the Euler code, w ~ t R out any linearize 2 approach are also presented. INTRODUCTION-OBJECTIVES This work is part of the overall combustion stability assessment of the Ariane 5 P230 MPS solid ro ellant motor and has been supported b CNES and E P J . Due to the segmented design of the P

Numerical Investigation of the Micro-Jets Efficiency for Jet Noise Reduction

motor, it is believed that there exists a potentially severe risk for motor instabiMy, although the moFor is predicted stable by conventional linear acoustic balance computations. Indeed, a number of US works, related to segmented motors (Space Shuttle and Titan solid boosters) 11-51 have riported significant and un redicted chamber frequencies. Further studies [4,6] have shown that a periodic vortex sheddin phenomenon is likely to be at the source of these oscifations. First suggestion of acoustic mode excitation by vortex shedding in solid rocket motors was made by Flandro and ressure and thrust oscillations at t f e first axial mode 1) Jacobs [7l. The eriodic shedding of vortices is the result of a strong cou Ping between the instabilit of mean flow sense, vortex shedding can be viewed as a, by-product of grain segmentation which produces regons of highly sheared mean flow, due to port area discontinuities and protrusions of inhibitor rings. Several related cold flow e eriments [4, 6, 8-12 have these experiments assume that a pair of diaphragms, at least, are required to give rise to vortex sheddin vortices are enerated at the first diaphragm and provife However, this is not the ony,mechanism for vortex shedding acoustic driving: one diaphragm or one region of sheared flow can be sufficient, the feedback mechanism being provided by the exhaust nozzle. This was the case in the ex eriment of reference [7], and could be the case in the & motor (Fig.1). Among, possible nozzle feedback mechanisms are direct vortex impingement on the nozzle wall or modulation of !he exiting mass flow rate by inhomogeneous flow, associated with one vortex. Flandro [13], extending the work of reference [7] has pro osed a linearized approach to evaluate the vortex shesding risk and to incorporate a corresponding driving term in linear acoustic balance codes. A lication of Flandro’s linearized model to Ariane 5 P a m o t o r has been carried out 141 and concluded of a possible vortex Recognizing the possible failure of classical linear acoustic balance computations even improved by Flandro’s mqdel, to achieve reliable stability predictions in complex internal eometries, such as in the P230 numerical simulation of the unsteady, comprcssiblc, internal flow field. Such a simulation would naturally couple mean flow shear layer instabilities and acoustic motions, including nonlinear effects and nozzle res once to vortices. The present work is the first part, of that effort and is devoted to demonstrating the easibility of a direct numerical capture of the periodic vortex shcdding phenomenon in test case solid propellant motors. This work was made possible by recent rogresses at unsteady state flow fields [15,16]. shear layers an B acoustic motions in the clamber. In this documented the vortex shed 3 ing phenomenon. Mlost of the acoustic B eedback when im inging on the second onc. P shedding driving 1 or Ariane 5 P230. motor,, a research ef B ort was initiated, aiming at a full ONERA in numerical simulations of bot g steady and * Research scientist .. Research scientist, Member AIAA Copyright

Aircraft Fan Noise Absorption: DNS of the Acoustic Dissipation of Resonant Liners

This paper presents the acoustic analysis of the hot subsonic and supersonic jets aerodynamic simulations described in Part I. Acoustic predictions are performed using the flow-fields provided by the 3D unsteady Navier-Stokes computations, for both supersonic and subsonic hot jets. Two surface integral formulations are used: the Kirchhoff method and the Ffowcs Williams and Hawkings (FW-H) equation on a porous surface. Integrations are performed on conical control surfaces as much as possible enclosing the jet and its mixing layers. The influence of the formulation and of parameters, such as control surface location and extent, open or closed nature of the control surfaces, on acoustic far field predictions is first investigated. The results show that the chosen control surfaces are far enough from the jets for FW-H and Kirchhoff calculations starting from pressure but not well suited for Kirchhoff calculations starting from density, in the present cases of hot jets. Comparisons with direct CFD results in the near field confirm this analysis and validate FW-H and Kirchhoff (pressure) radiated noise predictions. The results also show that an open control surface at its upstream and downstream ends can be used in practice. Closed control surface benefits appear only for acoustic predictions at low angle from the jet axis. However such surfaces intersect the jet and are not compatible with Kirchhoff calculations. Far field predictions are then compared with experiments for both supersonic and subsonic jets. The sound directivity and the Strouhal number of the maximum level in the spectral density as a function of the angle of observation are rather well predicted. However, the maximum noise level is overestimated. A deeper analysis of these aeroacoustic predictions is in progress. An example of improvement of CFD/acoustic jet computations is presented.

2D NAVIER-STOKES STABILITY COMPUTATIONS FOR SOLID ROCKET MOTORS : ROTATIONAL, COMBUSTION AND TWO-PHASE FLOW EFFECTS.

in this paper, we present aerodynamic and acoustic simulated results from a 50 mm single stream nozzle with continuous and pulsed micro-jets. It follows a former paper presenting the main results observed with continuous control and gives more detailed aerodynamic analyses on these configurations, as well as new results with pulsed micro-jets. Isothermal and hot main jet configurations are considered with an acoustic Mach number Ma = Vj/a∞ set to 0.9. Control is made through 12 micro-jets located around the nozzle exit with an impinging angle of 45 degrees to the jet axis; their temperature and velocity are identical to the main isothermal jet configuration. Instantaneous aerodynamic fields are obtained with a LES model using the MILES approach and the acoustic far field is computed with the Ffowcs-Williams and Hawkings surface integral formulation. When the control is active, jet flow investigations show a reduction of the turbulence in the shear layer and a decrease of the far field broadband noise, for all micro-jets configurations. Broadband noise reduction is more effective with pulsed control, which nevertheless generates penalizing tonal noises. Azimuthal analysis demonstrates that continuous control decreases all pressure modes in the near field. Pulsed control lead to a more complex behavior but also results in a global modes reduction for x/D ≥ 3. To end, computation of the thrust highlights the great interest of pulsed control for jet noise reduction, as it requires the minimum mass flow injection to work and gives the minimum thrust loss.

Applications of the CEDRE unstructured flow solver to landing gear unsteady flow and noise predictions

The acoustic dissipation mechanisms of resonant liners are strongly nonlinear. Usually set in the inlet nacelles to reduce fan noi se, perforated honeycomb sandwich panels are excited by grazing-incidence high sound pressure level (SPL) acoustic waves and air flows. The purpose of this paper is to stud y and quantify the acoustic absorption of a resonant liner, through direct numerical simul ations (DNS) performed with the Onera CEDRE code. The liner is modeled by a cylindrical Helmholtz resonator, with a 0.8 mm diameter opening. Its response to acoustic w aves is computed for different frequencies, SPL and incidences (normal and grazing). For the lowest SPL, the viscous wall friction and the acoustic radiation are identi fied as the “linear” dissipation mechanisms; for the highest SPL, the additional “no nlinear” vortex shedding mechanism is highlighted. Independently on the SPL, the flowf ields around the resonator hole are found to be mostly axisymmetric (along the resonator axis). As for the usual acoustic properties, the impact of a rise of the SPL is not as obvious as anticipated: for the frequencies of low linear absorption, the absorptio n gets better when the SPL increases, but for the frequencies of high linear absorption, this evolution is reversed beyond a given threshold. In the latter case, the absorption spectrum of the liner is flattened but wider. A simplified aeroacoustics excitation is fin ally approached: a grazing low Mach number inviscid air flow is coupled with the acoust ic waves. Several effects are pointed out, the main of which being the convection of the vortices downstream the resonator.

Aeroacoustic Numerical Method Assessment for a Double Stream Nozzle

D Navier-Stokes stability computations are performed on a simple cylindrical port motor. The 2D results, in terms of motor frequency and damping, as well as in terms of full acoustic field, are compared to classical ID linear acoustic balance. The 2D computations are performed with models representing the propellant combustion response and two-phase flow behavior. The models are used separately and then together. The results show that the overall tendencies are correctly obtained over a large range of model parameter settings. Surprisingly it is found that the 2D results do not depend on the grid spatial resolution and that the details of the so-called acoustic boundary layer (ABL) need not to be resolved. This finding is also true for the two-phase flow damping. Analysis of the computed 2D acoustic field show that the ABL displacement effect is an effective damping source (even for an ABL penetrating into the core of the flow) and that an extra damping exists which is not incorporated into the classical acoustic balance.

Numerical Simulations Of A Model Solid Rocket Motor Ignition Overpressure Wave

Landing gear noise computations require the calculation of unsteady flow fields around complex geometries. Recent experience was gained at Onera in CFD computations on LG simplified geometries, based on an aerodynamics solver which uses block structured grids ( els A). However, simulations of more complicated configurations are now also envisaged, and there are questions tha t a structured approach may not succeed in handling complex geometries, without requiring excessive grid works. The present study describes first attempts to assess th e DDES (Delayed Detached Eddy Simulation) method of Onera’s unstructured flow solver CEDRE, with respect to the simulation of landing gear aeroacoustics. This work is part of an Onera internal effort that is conducted in the framework of the Benchmark for Airframe Noise Computations (BANC). The paper firstly presents results obtained on the supercritical tandem cylinders configuration, an academic test case prop osed in the BANC framework. As a reference, computations are first run with a classi cal structured grid, which is easy to construct for such a configuration. This computatio n is followed by a calculation performed on an unstructured grid, which objective is to assess the sensitivity of the results to unstructured grids. For both calculation s, the farfield noise is then extrapolated from the local CFD results, using an i ntegral FW-H (Ffowcs WilliamsHawkings) method. Finally, the solver CEDRE is applied to a more complex test case representing a 1/4 th model of a partially dressed nose landing gear, al so proposed in the

On LAGOON Nose Landing Gear CFD/CAA Computation over Unstructured Mesh using a ZDES approach

Industrial results obtained from different steady and unsteady numerical and analytical methods for a double stream jet in confluent flow configuration are presented. These results are compared with measurements from anechoic wind tunnel tests of a model nozzle. The test campaign was conducted in the ONERA CEPRA19 facility, and exhaust parameters are representative of take-off, with and without simulated flight effects. During this campaign, in addition to acoustic measurements, LDV (Laser Doppler Velocimetry) technique was used in order to measure steady and unsteady aerodynamic data in several planes downstream the exhaust plane. Those data are used to assess several RANS methods (Reynolds Averaged Navier Stokes - steady calculation) on structured and unstructured 2D meshes using k-e turbulence model. Particularly, the effects of the refinement and of the structure of the mesh, as well as flight effects are investigated. Those data are also used to assess LES method (Large Eddy Simulations - 3D unsteady calculation) performed for the same model nozzle in confluent flow configuration, but without plug. Because results of both steady and unsteady aerodynamic calculations are usually used as inputs of acoustic prediction methods, an assessment of two different aeroacoustic prediction processes is also achieved, with different RANS calculations as input of an improvement of MGB (Mani Gliebe Balsa), and LES calculations as input of an acoustic integral method in the time domain (FfowcsWilliams and Hawkings surface integration). Particularly, flight effects and the ability of the whole aeroacoustic chain to predict them are studied.