Search

形態: カラー図版あり, Physical characteristics: Original contains color illustrations, Accepted: 2021-12-24, PA2210037000

We present a straightforward formulation predicting the characteristic velocity and specific impulse for bipropellant thrusters as a direct function of injection conditions, propellant combination,...

形態: カラー図版あり, Physical characteristics: Original contains color illustrations, Accepted: 2021-07-18, 資料番号: PA2210010000

形態: カラー図版あり, Physical characteristics: Original contains color illustrations, Accepted: 2020-06-24, 資料番号: PA2110026000

形態: カラー図版あり, Physical characteristics: Original contains color illustrations, Accepted: 2020-11-21, 資料番号: PA2110018000

Accepted: 2019-12-13, 資料番号: SA1190199000

形態: カラー図版あり, Physical characteristics: Original contains color illustrations, Accepted: 2019-05-28, 資料番号: PA2010047000

形態: カラー図版あり, Physical characteristics: Original contains color illustrations, Accepted: 2019-02-08, 資料番号: PA1910045000

Accepted: 2018-07-23, 資料番号: SA1180173000

51st AIAA/SAE/ASEE Joint Propulsion Conference (July 27-29, 2015.), Orlando, Florida, USA, 形態: カラー図版あり, Physical characteristics: Original contains color illustrations, Accepted: 2017-02-27, 資料番号: PA1710043000

In the present research, we experimentally verified the partial-fill effect in a multicycle pulse detonation rocket engine. The intermittent thrust of a pulse detonation rocket engine was measured by using a spring-damper mechanism that smoothed this intermittent thrust in the time direction. The intermittent mass flow rates were assessed by gas cylinder pressure or mass difference measurement. The maximum specific impulse was 305 ± 9 s at an ethylene and oxygen propellant fill fraction of 0.130 ± 0.004. When the fill fraction was greater than 0.130, the specific impulse was increased as the partial-fill fraction was decreased. When the fill fraction was less than 0.130, the specific impulse was sharply decreased as the partial-fill fraction was decreased. This decrease was due to diffusion between propellant and purge gases and the short length of the transition from deflagration to detonation. The multicycle pulse detonation rocket engine had a partial-fill effect that may have been mainly due to the suctioned air and was consistent with the single-cycle partial-fill model of Endo et al. [Endo, T., Yatsufusa, T., Taki, S., Matsuo, A., Inaba, K., and Kasahara, J., "Homogeneous-Dilution Model of Partially-Fueled Simplified Pulse Detonation Engines," Journal of Propulsion and Power, Vol. 235, 2007, pp. 1033-1041.].

ISBN: 978-1-56700-260-7, International Journal of Energetic Materials and Chemical Propulsion (ISSN: 2150-766X) Vol.8 (6), pp.531-539, 資料番号: SA1001011000

The oscillation mechanism and reignition process of one-dimensional unsteady detonations are numerically studied using a one-step chemical reaction model governed by Arrhenius kinetics. A series of simulations, without perturbations from the outflow boundary to the detonation front, are carried out while the degree of overdrive, f, is varied between 1.10 and 1.74 (f=D2/DCJ2; where D is detonation velocity). Shock pressure histories and x–t diagrams are utilized in order to attain precise understanding of the one-dimensional unsteady detonations. At higher degrees of overdrive, f=1.40–1.74, shock pressure histories agree with those of previous studies. The oscillation mechanism is the same as that of the large-disturbance regime of unsteady shock-induced combustion around a projectile. At lower degrees of overdrive, f

In Japan, a feasibility study of the new “LE-X” booster engine has been underway since 2005. One of the key technologies of the LE-X engine development is a regenerative cooling design to produce enough power to drive the turbopumps. The LE-X engine employs an elongated chamber design to pick up enough heat energy in the regenerative cooling channels. In the current study, a fully conjugated combustion and heat transfer simulation was performed to investigate the flow field and heat transfer characteristics for the regeneratively cooled combustion chamber of the LE-X engine. A three-dimensional Reynolds-averaged Navier–Stokes simulation was used to consider the injection and combustion processes of propellants on the hot-gas side, heat conduction in the chamber wall, and cooling channel flows. Details on the three-dimensional flow and thermal characteristics of the combustion chamber were clarified in the hot-gas and coolant side domains. In particular, significant thermal stratification was found to form in the radial direction of the cooling channel in the elongated cylindrical part of the chamber, which degraded the heat transfer and resulted in the highest wall temperature there.

Acoustic losses inside a simulated liquid rocket chamber are investigated by numerical simulation coupled with theoretical calculation. In this study, the losses by injector, resonator and chamber are considered and compared. The oscillation amplitude is assumed to be small (within linear range). For injector and chamber, the loss mechanisms, such as the radiation & convection from the inlet or outlet, and viscous & thermal loss at the wall are considered. For a resonator, the viscous & thermal loss would be the major loss factor. A simulated liquid rocket chamber configuration with an injector installed off-center is investigated especially for tangential oscillation modes. It is found that with well-tuned resonator the resonant frequencies and modes would change from those without the resonator. Therefore, the coupled simulation is indispensable for resonator design. Also an order estimation of each acoustic loss factors is conducted. It is found that the viscous & thermal loss of the chamber and resonator dominate the total acoustic loss in the present configuration. However, if several hundreds of injectors is equipped as in actual rocket chamber, the loss related to injectors would become comparable to that related to resonator on the total loss.

5th European Conference for Aeronautics and Space Sciences, EUCASS 2013 (July 1-5, 2013.), Munich, Germany, The characteristics of film cooling in a CH4/O2 sub-scale chamber with multi-injector elements and two kinds of film cooling slot dimensions are investigated using a calorimeter chamber in experiments and simulations, in which the finite rate chemistry with a reduced CH4/O2 reaction mechanism is taken into account. The computed wall heat flux and pressure distributions are compared to the experimental results, which overall show good agreement. A large slot dimension is shown to induce mixing with a core flow. This mixing causes a low heat flux distribution near a face plate along with high combustion efficiency., 資料番号: PA1510092000

In recent years, methane has attracted attention as a propellant for liquid rocket engines because of its various advantages compared to typical propellants such as hydrogen. When methane is used as a coolant for a regenerative cooling system, its near-critical thermodynamic and transport properties experience large variations because its critical pressure is higher than that of typical propellants; this significantly influences the flowfield and heat transfer characteristics. Therefore, adequate understanding of the flowfield and heat transfer characteristics of methane in regenerative cooling channels is a prerequisite for future engine development. In this study, conjugated coolant and heat transfer simulations were performed to investigate the flowfield and heat transfer characteristics of transcritical methane flows in a sub-scale methane-cooled thrust chamber. The computed results were validated against experimental data measured in hot firing tests. They compared well with the measured pressures and temperatures in cooling channels, and wall temperatures were within the permitted levels. Detailed flow analysis revealed peculiar flow structures in the cooling channel: a strong secondary flow induced in the concave-heated part in the channel throat section and the coexistence of two different gas phases—ideal and real—in a single cross-section in the cylindrical region. A high wall temperature appeared in the cylindrical region of the thrust chamber under the considered conditions; this was due to the heat transfer deterioration induced by an M-shaped velocity profile and a turbulent heat flux reduction.

To build a framework of a prediction tool for injection-coupled combustion instability with coaxial-type injectors, Hutt and Rocker’s methodology including Crocco’s n-τ model was applied. This linear stability analysis considers a coupling among LOX/fuel flow-path acoustics, chamber responses, and Crocco’s combustion characteristics. To validate the tool, LOX/methane subscale firing tests, which were performed and reported by NASA, were analyzed. The injection-coupled oscillating combustion, which was occurred at 5 kHz, was selected as unstable case. A stable combustion case was also selected for comparison purposes. Injection and combustion characteristics, which amplify the oscillation, were found to be dominated by the LOX-post acoustic characteristics. Chamber responses, which decay the oscillation, were estimated with two approaches: (a) onedimensional acoustic analysis with Natanzon’s methodology with a short-nozzle approximation, and (b) threedimensional acoustic analysis with a commercial software, ACTRAN. The stability was evaluated with AFC (amplitude-frequency characteristics) diagram, in which the injection and combustion characteristics and chamber responses are compared with regard to amplitudes. As a result, significant differences were not seen in the AFC diagrams between the unstable and stable cases (both analyses showed “unstable”). Further investigations for the chamber responses are needed to evaluate the potential instabilities of the system, correctly. In addition, we need to introduce a phase relationship into the tool to understand the underlying physical phenomena.

For LOX/LH2 shear coaxial injectors, it is well-known that high-frequency combustion instabilities may occur when the injection temperature of hydrogen decreases below a certain value, but the mechanism of the initiation of combustion instability with a coaxial injector is still not clear. In the present study, firing tests were conducted with five types of single shear coaxial injector elements by using LOX and LH2 as propellants to further investigate the mechanism of the initiation of combustion instability during temperature-ramping changes during hydrogen injection. Results showed that unstable combustion was initiated when the hydrogen injection temperature decreased to less than a certain cryogenic temperature. The combustion instabilities observed in the present firing tests are discussed and classified into three different types.

Flowfield and heat transfer characteristics of supercritical parahydrogen flows in a cooling channl of a sub-scale hydrogen-cooled thrust chamber are investigated using Reynolds-Averaged Navier-Stokes simulation, in which conjugated heat transfer between coolant flow and chamber wall is taken into account directly. The considered system pressure ranges from 3.8 to 4.6 MPa, and temperature from 43 to 324 K at Reynolds number more than 1× 10 5 . The computed results are validated against the experimental data measured in the hot firing testings, which compare well with measured pressures and temperatures in a cooling channel, and wall temperature in a hot firing testings. Detailed flow structure and heat transfer characteristics in a cooling channel are clarified, showing strong thermal and density stratification, secondary flow effect, and particular asymmetric heat transfer characteristics.

A practical design method of slit resonator for rocket engines is being developed. First a weak linear acoustic solution inside the resonator is obtained by finite element method. Using this result the viscous and thermal damping at the wall of resonator is calculated as a post process. Then, a multi-objective evolutionary computation is applied for finding the slit shapes with high linear damping characteristics. The obtained three slit shapes are investigated intensively. In the actual design, however the resonator is exposed to large amplitude of pressure oscillation. Therefore a non-linearity appears and the damping mechanism changes from that in the linear range. Recent numerical study has revealed the non-linear damping mechanism, that is, the generation of the vortex at the inlet of resonator. In order to estimate the resonators with complex structure, we are now developing a numerical tool to solve Navier-Stokes equations on unstructured grid. The non-linear damping characteristics of the three shapes are investigated and compared with the linear ones. Some problems of estimating the resonator damping in the non-linear range are also discussed.

The physical phenomena in the liquid rocket combustion chamber are very complicated such as turbulence, reaction, and real-gas effect. The prediction of heat flux on liquid rocket chamber wall is a challenging problem due to these complex physics This work is aimed particularly at the heat flux validation for the turbulent boundary layer in three physical situations; turbulent boundary layer on heated flat plate, turbulent boundary layer with pressure gradient, and recirculation zone of heated expansion tube. Each physical situation corresponds to the straight part of the combustion chamber, the rocket nozzle, and the exit of injectors. For the Reynolds-Averaged Navier-Stokes simulations, an adequate turbulent model should be selected to suit the flow feature. In this paper, three turbulent models have been tested for the above three validation problems. There is no perfect turbulent model for the prediction among the three turbulent models for all validation cases. The two-layer turbulent model works reasonably well for GH2/GO2 single injector conducted at Penn State University among the three turbulent models. Furthermore, turbulent models affect on not only the heat flux but also the turbulent intensity in the contraction tube of the GO2 injector and the flame shape.

Understanding and predicting the flowfield and heat transfer characteristics in cooling channels are prerequisite to improve design and performance of regeneratively cooled rocket thrust chambers. In order to realize them, a CFD code, able to predict such characteristics, is developed based on a pressure-based solver and the cubic-type equation of state to take into account the real gas effect. As a preliminary study, simulations of transcritical parahydrogen flows in a uniformely heated circular tube are performed in order to validate the developed code against the reference experiment and investigate the flowfield and the heat transfer characteristics under transcritical conditions. The computed results agree well with the experimental data with regard to the wall temperature, the heat transfer coefficient, the bulk pressure and temperature. Also, the peculiar behavior, called “heat transfer deterioration”, under transcritical condition with high heat flux, is successfully predicted. The simulated flowfield reveal the mechanism of it. The parametric studies with different heat flux levels clarify the condition in which the heat transfer deterioration takes places.

The performance of liquid rocket of expander bleed cycle engine system strongly depends on a regenerative cooling performance to provide the required heat in the Main Combustion Chamber (MCC). The prediction of the MCC wall heat transfer characteristics and wall temperature distributions is very important for designing a new engine. Developing a tool for a combined analysis among the combustion gas region, chamber structure, and the cooling channel is our ultimate goal. As a part of the development, we carried out numerical simulations of combustion flow fields in an injector, in order to understand the flow characteristics, which may influence the heat transfer and temperature on the MCC. This report shows the numerical simulation results of GH2/GO2 coaxial flow. This problem is adequate for validation of mixing, diffusion, and chemical reaction. Numerical simulations are carried out using the commercial code FLUENT, Advance/FrontFlow/red, and CRUNCH CFD. Computed velocity and mole fraction are compared with the firing tests data conducted at the Pennsylvania State University, using Eddy Dissipation model, Laminar Finite Rate model, and Eddy Dissipation Concept (EDC) model. The results of steady state simulations have a tendency to estimate a longer flame than the test data, since the vortex mixing around the GO2 post was underestimated and moreover the flow field is basically unsteady. The time-averaged values of unsteady simulations of FLUENT with EDC model simulate turned out to be good agreement with the test data relative to the steady state solution, indicating the importance of capturing the mixing process accurately. Influence of several important numerical factors such as grid resolution, combustion model, and turbulence model on the unsteady combustion flow field will be discussed in feature.

§A numerical framework is established to predict the erosion amount of nozzle wall in solid rocket motor. Understanding of the flow field and erosion mechanism is very important for the performance of solid rocket motor and the success of the launch. A coupled analysis of fluid dynamics and surface recession simulates total ablation amount. The analysis consists of two-dimensional axisymmetric fluid simulation and estimation of ablation amount using a one-dimensional heat conduction equation with thermal decomposition. The model includes the effect of the mechanical erosion. Three kinds of simulations are carried out to reveal the features of ablation and evaluate the erosion model. Features of nozzle shape can explain heat flux distributions and surface recession amount for each nozzle. The simulation results of total surface recession amount qualitatively recreate the experimental results. The numerical simulations estimate the erosion rate on the safe-side for all nozzles.

One-dimensional H2-Air detonation is numerically studied. The simulations are carried out with various parameters, such as degree of overdrive, initial pressure, and equivalence ratio. Steady state, high-frequency mode, and low-frequency mode are observed in this order, when the degree of overdrive reduces. Although the order of the transition of oscillation mode does not depend on the initial pressure or the equivalence ratio change, the degree of overdrive of transition mode does. When the initial pressure is higher or the equivalence ratio is not stoichiometric, the degree of overdrive is higher. Introduction The supersonic combustion and the detonation in the premixed combustible gases have been studied to develop the supersonic propulsion devices. A number of ballistic range experiments and numerical studies [1-7] have reported periodic oscillations of the shock-induced combustion around projectiles flying at hypersonic velocity into detonable gases. The mechanisms are described with the wave interaction in an x-t diagram on the stagnation streamline in front of the projectiles. The presence of two typical oscillation modes were reported in previous studies [1-7]: one is referred to as regular regime (RR) , whose oscillations are highly regular and low in amplitude, and the other is referred to as large-disturbance regime (LDR) , whose oscillations are less regular and low in frequency. Figure 1 shows the x-t diagram of (a) RR and (b) LDR. In RR, compression waves from the new reaction region propagate upstream and downstream, and the wave propagating upstream interacts with the bow shock. After the interaction, a contact discontinuity is generated. A new reaction region is formed again, because higher temperature behind the contact discontinuity has a shorter induction time. In LDR, a detonation wave is instantaneously generated at the reaction boundary. The detonation propagates upstream, and then penetrates the bow shock. This direct interaction between the bow shock and the reaction front is characteristic of the largedisturbance regime. One-dimensional detonation has been compared with the shock-induced combustion around projectiles flying at hypersonic velocity in a few previous studies [8, 9] because the oscillation mechanisms are similar both examples. Matsuo and Fujii [8] using two-step chemical reaction model have reported that the oscillation type does not depend on the intensity of the concentration of the heat release due to the combustion, which is trigger of the unsteadiness of oscillation type of the shock-induced combustion around projectiles. Sussman [9] using the detail chemical reaction model has presented that the frequencies for the one-dimensional high-frequency mode are approximately the same as the values computed for similar shock velocities for the blunt body flows. Many previous works of one-dimensional detonation [1020] have used the one-step irreversible chemical reaction governed by Arrhenius kinetics. Linear stability analysis [10-13] has been studied so as to investigate onedimensional longitudinal unsteadiness. Using linear stability analysis, Bourlioux et al. [10] have numerically studied the spatio-temporal structure of unsteady detonations in one-dimension. In their work, degree of overdrive (f=D/DCJ , D is the detonation velocity) is chosen as a parameter, and the values of other dimensionless parameters (activation energy, heat release, and ratio of specific heats) are fixed. As the degree of overdrive is reduced, the oscillation mode changes unsteadily. Sharpe and Falle [11] have chosen an activation energy as a parameter, and the values of other dimensionless parameters are fixed. As the activation energy is reduced, the oscillation mode changes. In our previous study [14], a series of simulations using a one-step Graduate Student, Student Member AIAA. Associate Professor, Member AIAA. Copyright © 2003 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved. 42nd AIAA Aerospace Sciences Meeting and Exhibit 5 8 January 2004, Reno, Nevada AIAA 2004-794 Copyright © 2004 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved. 2 American Institute of Aeronautics and Astronautics chemical reaction model have been carried out in an attempt to understand the oscillation. The oscillation mechanism was the same as that of LDR of unsteady shock-induced combustion around a projectile, but RR oscillation was not observed. There are a few reports of one-dimensional detonations using a detailed chemical reaction model. In the present study, the oscillation behavior using the detailed chemical reaction model in the one-dimensional detonation is focused on. The degree of over drive, the initial pressure, and the equivalence ratio are chosen as the parameters of the present work. The simulation results are compared in the previous works of one-dimensional detonation and the shock-induced combustion. Computational Setup The governing equations are the one-dimensional Euler equations with finite rate chemistry. In the current study, a 9-species, 19-reaction mechanism for hydrogen-oxygen combustion [21] is used. The algorithm used for solving theses equations is Yee s non-MUSCL-type TVD upwind explicit scheme [22]. The equations are integrated explicitly and the chemical reaction source term is treated in a linearly point-implicit manner. The detonable mixtures are hydrogen-air mixtures with varying equivalence ratio from 0.50 to 2.00. The initial pressure set to be 0.2, 0.421, 1.0 atm. The initial temperature is fixed as 293K. The initial state is given by steady detonation in which the unburnt mixture enters through the left boundary at the overdriven detonation velocity and the burnt mixture exits through the right boundary. The inflow and outflow boundaries are fixed with the initial conditions so as to maintain mass conservation in the computational domain. We choose a shock tracking method for capturing the reignition process, because the leading shock speed is supposed to vary in a wide range at lower degrees of overdrive. The shock is tracked on the grid point where 5% of total grid points are left upstream of the shock. Therefore, the grid always moves at the shock speed, which is derived from the shock pressure jump and the RankineHugoniot relations. The long downstream length behind the shock is prepared, because under weaker shock the chemical reaction takes a longer time to complete. We confirmed that no perturbations reflected from the outflow boundary reach the detonation front during the time used calculation, based on the observation of the wave propagation of the perturbations on the x-t diagram. This idea is conceptually the same as the previous study [14, 19]. The grid spacing of 240 points/L1/2 in the fine area and subsequently the stretched grid between the end of fine area and outflow boundary are prepared. The half-reaction length, L1/2, is defined as the distance behind the shock at which the mass fraction of hydrogen is equal to the average W av e Co m pr es sio n Wave Cpression Front Reaction Cycle One Period of (ath lne) D isctinuity

The features of one-dimensional unsteady detonations are studied numerically using a hydrogen-air detailed chemical reaction model. A series of simulations are carried out while degree of overdrive, initial pressure, initial temperature, and equivalence ratio are varied. The oscillation modes and mechanisms of the one-dimensional detonations are discussed with reference to shock pressure histories and x-t diagrams of density distributions. As the degree of overdrive is reduced with a stoichiometric mixture of hydrogen-air at P0=0.421atm and T0=293K, a steady state appears, along with a high-frequency mode and a low-frequency mode. The oscillation mechanism of the high-frequency mode is the same as that of the regular regime of unsteady shock-induced combustion observed around a spherical projectile flying at hypersonic velocity in detonable gases. The degree of overdrive threshold between the steady and unsteady region increases monotonically with initial pressure and decreases monotonically with initial temperature. When the equivalence ratio is changed, the threshold has a minimum value around ϕ=1. We focus attention on a nondimensional effective activation energy, which is generally used for linear stability analysis. The oscillation mode depends highly on the nondimensional effective activation energy. The oscillation of the detonation front appears as the nondimensional effective activation energy goes past a threshold value of 5.2. Furthermore, we investigate the failed regime and possible reignition in this regime. In the failed regime, a detonation wave breaks up into a leading shock, a contact discontinuity, and a rarefaction wave. When the shock is weak, reignition time becomes very long. Therefore, the reignition after the failed regime is difficult to reproduce in the restricted computational domain. The reignition process in the failed regime is investigated by means of analysis consisting of integration along the point of intersection between a Rayleigh line for weak leading shock and a partially burnt Hugoniot curve. The reignition time increases dramatically with decreasing temperature behind the shock wave, when the gas condition goes past the second explosion limit. The second explosion limit is one of the characteristics of the hydrogen-air detailed chemical reaction model and does not exist in the one-step chemical reaction model. Lastly, the reignition time obtained by the analysis is compared with that obtained by the simulation results. The simulation results agree well with the analytical results.

Regenerative cooling is still one of key technologies to develop high performance liquid rocket engines. To achieve high efficiency and reliability, understanding and accurate prediction of flowfield and heat transfer characteristics in regeneratively cooled thrust chambers are prerequisite. In the current study, a fully conjugated combustion and heat transfer simulation for full-scale regeneratively cooled thrust chambers was proposed and demonstrated for the LE-5B thrust chamber. In the proposed strategy, the injection and combustion processes in the hot-gas side, heat conduction in the chamber wall, and cooling channel flows are taken into account based on three-dimensional Reynolds-Averaged Navier-Stokes simulation. The computed results were validated against measured data from a hot firing test, showing reasonable agreement except for chamber outer wall temperatures. Detailed three-dimensional flow and thermal characteristics in the thrust chamber were clarified in the hot-gas side and the coolant side domains. Although the proposed numerical approach needs to be further improved quantitatively, it was confirmed that the present methodology is promising to understand and precisely predict flowfield and heat transfer characteristics in regeneratively cooled thrust chambers.

Combustion flowfields in GH2/LOX sub-scale calorimeter chambers with multi-injector elements and full-scale thrust chamber are investigated using Reynolds-Averaged NavierStokes simulation, in which the finite rate chemistry with the H2/O2 detailed reaction mechanism is taken into account. The computed wall heat flux distributions are compared to that of the simplified cases to reduce a computational cost. The considered simplifications are a presence of reaction and a number of injector rows. At first, these simplifications are validated in the simulation of sub-scale chambers. The reaction is essential for the prediction of heat flux because it makes change the species distribution in a thermal boundary layer on a thrust chamber wall. A heat flux using a combustion simulation with only outermost injectors shows a good agreement with that with an original configuration near a face plate. On the other hand, it overestimates the heat flux around nozzle and throat parts. It is clarified that this overestimate comes from the shortage of unburned hydrogen near a chamber wall in the simplified method. Next, the simplification of the number of injector rows are applied to the simulation of full-scale thrust chambers. The effectiveness of this simplification for the prediction of wall heat flux is revealed. The optimal solution by using of the simplification is proven to be effective for the prediction of heat flux in a full-scale thrust chamber.

Evaluation of ablation of nozzle wall and natures of three-dimensional flow in solid rocket motor are studied numerically. Three kinds of simulations are carried out to reveal the features of ablation. At first, a coupled analysis of fluid dynamics and surface recession simulates a total ablation amount. The analysis consists of the two-dimensional axisymmetric fluid analysis and the estimation of ablation amount using a correlation equation of surface recession rate. The features of nozzle shape can explain heat flux distributions and surface recession amount for each nozzle. The simulation results of total surface recession amount agree well with experimental results. Most of solid rocket motors use the different materials in the throat and nozzle parts because the throat diameter should be unchanged as possible as it could throughout the whole burning period. Therefore, a backward-facing step is formed at the boundary between those materials because the ablation rates of walls are different. Next, three-dimensional fluid steady analysis in solid rocket motors is carried out. Three-dimensional grid is made based on the results of the axisymmetric coupled analysis. Behind the step, longitudinal vortices and streak of heat flux appear. An effect of shape irregularity at nozzle inlet nose is considered. The shape irregularity makes the fluctuation with low frequency and high amplitude on the heat flux. Lastly, three-dimensional fluid unsteady analyses around step on the nozzle wall are carried out to reveal the relation between longitudinal vortices and streaks of heat flux. The artificial steps with various heights are set on the nozzle wall. The number of longitudinal vortices depends on the step height. The non-uniformity of shear layer in circumferential direction affects the generation of the longitudinal vortices. The collisions of longitudinal vortices to the wall make the high heat flux.

2 Acoustic structure inside a cylindrical rocket chamber is numerically and theoretically investigated, which is the key for determining the damping performance of the whole chamber to measure the combustion instability. When a mean flow is present, the boundary conditions for acoustics are expressed by complex numbers which is called impedance or admittance. It depends on the Mach number of the mean flow, the frequency, the mode of the oscillation and the shape of the inlet (with injectors) and convergent part of a nozzle. In this study the first tangential mode is intensively investigated for some chamber configurations. First the acoustic structure is investigated with and without a mean flow. Then the damping characteristics is estimated numerically and explained from the theoretical point of view. Finally, the intense first tangential pressure oscillation with a mean flow is numerically reproduced and the characteristic features are compared with those without a mean flow.