CFD Simulation of Regression Rate in Hybrid Rockets

M. Lazzarin1, F. Barato2
1,2

Center of Studies and Activities for Space,(CISAS) G. Colombo, University of Padova, Padova, Italy, 35131

and A. Bettella3, D. Pavarin4

3,4

Department of Mechanical Engineering, Center of Studies and Activities for Space (CISAS) G. Colombo, University of Padova, Padova, Italy, 35131

Abstract
In this work, a CFD code has been used to simulate hybrid rocket motors using O2 as the oxidizer and HTPB or HDPE as the fuel. Two different kinds of simulation have been performed: a) the first, using pre-defined fuel and oxidiser mass flow rates, and b) the second calculating the fuel mass flow rate as a function of the wall heat flux. For this second type of simulations, regression rate has been determined and compared to the average value derived from the reference experiments; no tuning coefficients nor any other correction parameters have been introduced. The numerical models applied are discussed and a complete description of the set up is given. First of all, a mesh convergence study is presented, for one of the geometries treated. Then, CFD results are compared with the corresponding experiments described in the literature. In the test cases calculating regression rate as a function of the wall heat flux, the fuel grain has been divided into segments, and the average regression rate resulting from the simulation has been plotted as a function of the average Gox. Concerning pressure at the head-end of the combustion chamber and C*, the simulations show a good agreement with the experimental results for the different rocket configurations analyzed. Regression rate is determined by the CFD with an underestimation around 30% for HDPE and 50% for HTPB, if only convective heat exchange is accounted for. It has been proved for the HTPB case, that for relatively low Gox, the radiative heat transfer contribution to the total heat flux is not negligible and this is confirmed by CFD results: in fact, when the radiative heat flux contribution is added, the error on the regression rate calculated is reduced to around 15% and in this case the experimental value is overestimated. Moreover, it is true that CFD does not predict correctly regression rate absolute value, due to a

wrong estimation of the wall heat flux, but it calculates correctly the n parameter of the regression rate formulation: r = aGoxn.

I. Introduction
Some of the efforts made since 2001 to create and use numerical tools predicting hybrid rocket performance or analysing the flow field are reported. In 2003, Serin and Gogus [1] studied the HTPB/O2 reacting flow field inside the hybrid rocket motors and the corresponding regression rate. They used a commercial Navier-Stokes code, CFD-ACE to understand the mechanisms affecting regression rate. In 2005, Antoniou and Akyuzl [2] published a mathematical model predicting the entire hybrid rocket performance. Guobiao and Hui wrote a paper [3] about their theoretical analysis of propellant performance, solid fuel regression rate, combustion and flow in hybrid rockets. Recently, a paper has been presented by Astrium about CFD simulations of GOX/HTPB lab-scale rockets [4], where axisymmetric motors are analysed using a steadystate approach. In 2010 and 2011, two papers have been presented by Lazzarin et al. [5], [6] concerning CFD simulations of hybrid rockets employing diaphragms inside the combustion chamber and using liquid oxidiser: the main interest was to prove that CFD tools are able to predict the efficiency of different motor configurations. For this paper however, not only numerical work and experience is important, but also the experiments conducted on hybrid rockets. In fact, the study carried out is validated against some experimental results obtained for specific lab-scale rockets. Concerning experimental work and results, Carmicino and Russo Sorge [7] published a paper about the effects of different kinds of injection on hybrid rockets performance. In 2009, Carmicino et al.

[8] published their study about hybrid engine operation, where the results of the tests conducted for different fuel compositions were reported. In the year 2000, Chiaverini et al. published a paper about the regression rate characteristics of HTPB [9], which is important to understand the relative importance of radiative and convective heat flux for regression rate determination. This paper has then been collected in a book [10] which can be considered as the first complete survey on hybrid rocket motors.

been simulated creating a dedicated circular surface at the inlet section of the pre-combustion chamber. This surface patch has the same diameter of the injector nozzle, in order to represent correctly the flow parameters at its exit section. The only problem which can occur in this case is an instability at the beginning of the simulation, induced by the high injection velocity of the oxidiser gas. Anyway, it is sufficient to apply a lower oxidiser mass flow rate at the beginning of the simulation and to increase that value progressively, until the nominal value has been reached, to obtain a stable oxidiser flow. HDPE and O2 Rocket Geometries For the HDPE study as well, some experimental results have been used, which are illustrated in [7], in order to validate the CFD simulations. Once again, only three of the tested configurations have been reproduced numerically, choosing different Gox values and mass flow rates to analyse a wider working range for the rocket. For this reason, three simulation domains have been prepared for HDPE; each referring to a specific lab-scale configuration tested and described in [7].
Test 5 Lfuel [m] 0.560 0.560 0.560 dinj [m] 0.008 0.008 0.008 dport [m] 0.04138 0.05473 0.04233 dt 0.016 0.016 0.016 Ae/At 2.44 2.44 2.44

II. Geometry Description

HTPB and O2 Rocket Geometry In order to validate the CFD results obtained from this HTPB analysis, some experimental findings have been used, which are illustrated in [8]. Only three configurations have been reproduced numerically, choosing different Gox values and mass flow rates to analyse a wider working range for the hybrid rocket. For the HTPB test cases into consideration, the simulation domains have been created from the information available in the literature and concerning related experiments [8]. Tab.1 resumes the main geometry parameters used for the three analyses performed. The test numbers reported correspond to those used in [8] for the sake of clarity.
Test 2 10 12 Lfuel [m] 0.574 0.570 0.572 dinj [m] 0.008 0.008 0.008 dport [m] 0.04240 0.05675 0.05793 dt 0.016 0.016 0.016 Ae/At 2.44 2.44 2.44

7 10

Tab.2: Main geometry parameters of the HDPE fluid volumes simulated.

Tab.1: Main geometry parameters of the HTPB fluid volumes simulated.

In this case, the average port diameters are already given as a result of the experiments performed, so that no other calculation has been necessary.

The average port diameter has been calculated using the average L/D ratio. For example, for test n.2, L/D is 13.53, which accounts for an average diameter equal to D=13.53Lfuel.

Fig.2: Geometry of one of the HDPE test cases, determined from experimental data [7]. Fig.1: Geometry of one of the HTPB test cases, determined from experimental data [4].

The injector fluid dynamics has been reproduced in this analysis, in order for the CFD simulations to be as close as possible to the real conditions. Therefore, the axial injector used during the experiments has

In all the HDPE experimental tests as well as for HTPB, the injector used was axial. Therefore, the CFD geometry reproducing this configuration is a circular surface, having the same diameter of the injector nozzle, as already seen for the HTPB case.

III.

Simulation Approach and Numerical Models

Simulations Set up This section illustrates the main models applied and the boundary conditions used in the CFD studies performed. The simulations have been run in the steady state mode, using only gaseous species, as in [5]. Gaseous O2 is injected in the pre-chamber at 230 K, through a circular patch having the same diameter as the injector nozzle. The fuel is C2H4 for the HDPE test cases and C 4H6 when simulating HTPB. The eddy dissipation model has been used to describe combustion and the chemical reactions applied are single-phase, involving only gaseous species. More details about the reasons of these choices in terms of physical models are given in [5]. For HTPB, the reaction applied is: 103O2+22.5C4H6=48H2O+8O+20OH+7H+50CO+40 CO2+6H2, whereas for HDPE: 101O2+38C2H4=58H2O+8O+20OH+4H+36CO+40 CO2+6H2. The turbulence model activated in the simulations is the standard k-. It uses some standard coefficients, which have not been specifically validated for hybrid rocket fluid dynamics. This is the reason why using this numerical model probably accounts for a simplification of the real physics involved. The fuel grain temperature is maintained fixed during the whole simulation, for both HTPB and HDPE; this temperature value can obtained from the following equation:

the experiments. Due to the high activation energy, eventual regression rate variations can produce only small Ts changes. A sensitivity analysis has also been performed, which showed that, if Ts varies within the expected range, this does not sensibly change the wall heat flux calculated by the CFD. The following list resumes the boundary conditions applied to the various geometry patches: Injector patch: inlet, with pre-defined mass flow rate and temperature; Fuel grain: wall + source; Nozzle and injector walls: adiabatic walls; Nozzle exit section: outlet with a defined average static pressure. In those cases where only a 90-wedge of the full 3D rocket geometry has been represented and simulated, another boundary condition has been applied to the internal surfaces. They have in fact been treated as symmetry boundaries, which is appropriate due to the presence of an axial-type injection. Since the injector to be simulated is axial, it has been sufficient to apply a specified mass flow rate to the patch representing the injector nozzle exit section. The oxidiser flow has then been defined as normal to that patch. Three different solutions have been created for the source simulating the mass flow rate due to the pyrolysis of the solid fuel: a) Fixed total mass flux in kg/s, imposed directly as an average value derived from the experiments; b) Total mass flux derived from the convective heat flux calculated by the CFD as

r Ae

Ea nTs R

m fuel

qconv A fuel . In this case, the value Hv

where A is a constant in mm/s, R is the universal gas constant in J/(kg mol), Ea is the activation energy of the process in J/mol, T s is the fuel grain surface temperature in K and r is the average regression rate, in mm/s.

used is not imposed by the user, but directly solved by the software; c) Total mass flux derived from both convective and radiative heat fluxes, as

m fuel

(qconv qrad ) A fuel . Hv

Ea A [mm/s] n [kcal/mol] HTPB 4.91 11.04 1 HDPE 60.00 4.78e+6 2 Tab.3: Main parameters for grain surface temperature determination. Fuel

IV. Simulation Results Hereafter, the results obtained for the mesh convergence study are presented. For each of the geometries analysed, various meshes have been created and the results in terms of combustion chamber pressure and wall heat flux have been

The regression rate values used for the grain surface temperature determination are those obtained from

compared to ensure their convergence as a function of the mesh refinement. Mesh Parameters and Convergence Study Different meshes have been created to investigate the effect of the mesh size on the fluid dynamic parameters characterizing the rocket combustion chamber. In order to decide if the mesh refinement level was enough to guarantee the best possible results, both pressure and wall heat flux stability have been monitored in the different cases. As an example, the pressures obtained with three meshes created using different element numbers and characteristic sizes are compared in Fig.3, whereas the corresponding mesh parameters are resumed in Tab.4.
Fuel Body size Inflation surface Elements [m] layers size [m] 1 1.50e-3 1.75e-3 724,957 10 2 1.25e-3 1.00e-3 2,008,598 10 3 1.15e-3 7.50e-4 3,762,494 20 Tab.4: Mesh parameters for the convergence study. Mesh

Optimise Computational Time Simulating a Wedge Using the selected mesh n.3, a comparison has been made to verify the possibility to optimise the computational time simulating only a quarter of the entire rocket geometry. A specific experimental case has been reproduced using both the entire rocket geometry and a quarter of the total rocket (90 wedge). The results obtained with these two different configurations have been compared in terms of the average combustion chamber pressure and the maximum temperature, and they are reported in Tab.5.
Pcc Tmax Tmax [%] [K] [%] Entire 21.44 < 0.2 4062 < 0.25 Wedge 21.4 < 0.2 4053 < 0.25 Tab.5: Differences in temperature and pressure for the two geometries tested. Geom. Pcc [bar]

As showed in Tab.5, the differences in the results obtained allow using the wedge geometry, which eliminates 3/4 of the cells used for the simulations presented in this study. This clearly accounts for a reduction in the CPU time required for the whole analysis.

Fig.3: Convergence curve for the meshes in Tab.4. On the x axis the mesh number, on the vertical axis the chamber pressure obtained. Fig.4: Temperature plot comparison for the entire and wedge rocket geometries.

Mesh n.3 has been considered as refined enough to be used for the simulations comparing experimental and CFD results, due to both the difference in the pressure obtained with respect to mesh n.2, which is only 0.21%, and the difference in the average wall heat flux, which is 2%. To establish that each single simulation has reached its steady state result, three criteria are applied at the same time: a) all the rms residuals shall be below 1e-4, b) the average pressure within the combustion chamber shall be stable, c) the average wall heat flux shall be stable.

Comparison between Experimental Results and CFD for the HTPB-O2 rocket The references for the comparisons proposed in this section are tests n.2, 10 and 12 described in [8]. For each test case, two simulations have been prepared: the first applies a constant fuel mass flow rate (as determined from the experiment), the second uses a fuel mass flow rate depending on the convective wall

heat flux calculated by the CFD. When regression rate is calculated as a function of the heat flux, the grain is divided into different segments, in order to allow a more precise evaluation of the wall heat flux and regression rate and to permit a better numerical convergence and stabilisation. Tab.6A and 6B show the results of the simulations carried out: nd means not determined and F indicates the fixed fuel mass flow rate tests.
Test Exp. P[bar] P[bar] P[%] Exp. r[mm/s] r [mm/s]

due to the fact that the fuel mass flow is only a small percentage of the total mass flow within the combustion chamber, so that a regression rate underestimation around 50% causes a smaller pressure underestimation, being the oxidiser mass flow rate imposed at the inlet section. Considering test n.10, the experimental O/F ratio is near 1, which means that an error on the fuel mass flow rate implies a greater error in the average combustion chamber pressure evaluation, compared to the other two test cases. Even if the predicted regression rate is always around 50% lower than the experimental one, the slope of the regression rate curve as a function of the average Gox is very similar between CFD and experiments; this is due to the constancy of the CFD error in terms of regression rate determination. This result accounts for a good CFD prediction of the n coefficient in the regression rate formula: r = a Goxn, as proved by Fig.6. It is true that the CFD points obtained for the n coefficient evaluation are too few to draw definitive conclusions, but it is worth to remark that when a specific experiment shows a certain deviation from the average result, the same happens to its corresponding CFD solution.

2 16.42 16.31 -0.70 0.931 0.505 2F 16.42 16.61 1.19 0.931 nd 10 6.99 6.11 -12.59 0.473 0.247 10 F 6.99 6.44 -7.92 0.473 nd 12 24.67 24.07 -2.44 0.961 0.517 12 F 24.67 24.94 1.08 0.961 nd Tab.6A: HTPB: comparison between experiments and CFD.

Test 2 2F 10 10 F 12

r[%] -45.73 nd -47.78 nd -46.20

Exp. C*[m/s] 1614.80 1614.80 1479.50 1479.50 1689.00

C*[m/s] 1713.90 1659.20 1370.09 1405.47 1738.02

C*[%] 6.14 2.75 -7.39 -5.00 2.91

nd 1689.00 1720.83 1.88 12 F Tab.6B: HTPB: comparison between experiments and CFD.

Fig.6: Comparison between CFD and experimental regression rate in a log-log scale for HTPB tests.

Fig.5: Temperature comparison: fixed fuel flow (above) and wall heat flux-dependent fuel flow (below).

As can be seen from the results reported for tests n.2 and 12, the pressure predicted by the CFD at the head end of the combustion chamber is closer to the measured value compared to regression rate. This is

The good correspondence between the CFD results obtained using a fixed fuel mass flow rate and the experiments is due to the CFD ability to correctly predict the species mixing, but even in this case, a difference is present between experiments and numerical test cases, due to some approximations in the models used. In fact, the geometries created for the CFD studies are cyclindrical and the port diameter simulated is the average value obtained from the experiments. Experimental tests highlighted

instead a non-uniform grain consumption, with a higher regression rate near the injection section. Moreover, the combustion process is simulated with a very simple model and the turbulence model used is created with standard coefficients, which have not been adapted to this specific kind of problems. Another reason of the differences between CFD and experiments is due to the wall functions applied, which have not been validated nor tested for hybrid rockets boundary layer. It should be assessed which wall functions can be used to describe chemically reactive boundary layers, with high gradients of the fluid dynamic parameters; dedicated wall functions shall be developed to describe these conditions.

Fig.7: Temperature comparison: fixed fuel flow (above) and wall heat flux-dependent fuel flow (below).

Comparison between Experimental Results and CFD for the HDPE-O2 rockets The references for these simulations are: tests n.5, 7 and 10 presented in [7]. For each of the geometries studied, two test cases have been prepared, one with a fixed fuel mass flow rate and the other with a heatdependent regression rate. As for the HTPB, when treating regression rate, the fuel grain has been divided into segments. Tab.7A and 7B resume the results obtained from the CFD analysis.

The CFD error in the determination of pressure at the head end of the combustion chamber is lower than 8% whereas, even in the HDPE case, CFD regression rates result to be lower than the measured value by a greater amount (around 30%) compared to pressure. This is due to the fact that an error in the fuel mass flow rate determination causes a smaller error in pressure. In fact, the fuel mass flow rate in the HDPE tests used for this study only corresponds to the 2528% of the total mass flow rate within the port volume and the oxidiser mass flow is fixed.

Test

Exp. P[bar]

P[bar] P[%]

Exp. r[mm/s]

r [mm/s]

5 15.665 14.49 5F 15.665 14.62 7 19.211 18.96 7F 19.211 19.10 10 20.518 19.510 10 F 20.518 21.440 Tab.7A: HDPE: comparison CFD.

-7.5 -6.67 -1.31 -0.58 -4.91 4.49 between

0.63 0.43 0.63 nd 0.59 0.41 0.59 nd 0.82 0.55 0.82 nd experiments and

Test 5 5F 7 7F 10

r[%] -32.06 nd -30.34 nd -32.92

Exp. C*[m/s] 1920.49 1920.49 1783.83 1783.83

C [m/s] 1797.06 1796.78 1793.44 1784.19

C [%] -6.43 -6.44 0.54 0.02 -0,075

Fig.8: Comparison between CFD and experimental regression rate in a log-log scale for HDPE tests.

1721,793 1720,494

nd 1721,793 1804,277 4,791 10 F Tab.7B: HDPE: comparison between experiments and CFD.

The interesting result is that, considering the three tests presented, CFD replicates the slope of the experimentally-determined regression rate curve as a function of the average Gox. This is due to the same fact explained for the HTPB and can be seen from Fig.8, where a log-log plot is presented of regression rate versus the average Gox. As discussed for the HTPB case, the CFD points used for the n coefficient evaluation are too few to draw complete conclusions. Anyway, also for the HDPE, when a specific experiment shows a deviation from the average

result, the same happens to its corresponding CFD solution. Even if there is a very good agreement between the CFD results obtained using a fixed fuel mass flow and the experiments, there still are some discrepancies due to some approximations introduced: the CFD geometries have been simplified, the numerical models applied have not been specifically created for hybrid rockets internal fluid dynamics description, the wall functions have not been validated for the hybrid rockets reactive boundary layer.

Two HTPB test cases have been run considering the radiative heat flux contribution: one corresponding to test n.2 and the other replicating test n.12 [8]. The Gox and O/F values of these two cases are reported in Tab.8, together with the flame temperature Tflame used for the radiative heat determination. Tflame has been calculated as the 95% of the value determined by the thermo-chemical code for each O/F ratio, as illustrated in Tab.8.
Test 2 12 O/F 1.942 2.560 Tflame [K] 3408.69 3430.92 Gox [kg/m2s] 94.0 74.6

Introduction of Radiative Heat Flux for HTPB Fuel Mass Flow Rate Determination The results obtained showed that CFD calculates HTPB regression rate with a greater error than for HDPE. A possible explanation for this is the importance of the radiative heat flux for HTPB regression rate, at relatively low Gox values. In fact, Chiaverini proved that HTPB tends to produce a high quantity of soot particles [9], which cause a significant radiative heat transfer. At low G ox, this contribution can be of the same order compared to the convective heat flux [10]. In order to verify this conclusion, another contribution has been added to the wall heat flux of the HTPB CFD simulations: radiation. However, instead of adding another theoretical model to those already applied for the CFD calculation, the amount of the wall heat flux due to radiation has been estimated using the data available from the literature [10]. This permitted to avoid: Consuming CPU time to solve other equations for the radiative heat flux; Applying an inappropriate radiative model for the underlying physics. The formulation of the radiative heat flux applied to the CFD calculations is:

Tab.8: O/F, Gox and flame temperatures for HTPB radiative heat flux.

Tests n.2 and 12 have given results which are resumed in Tab.9.

Test

Exp. P[bar] P[bar]

P [%]

Exp. r[mm/s]

r [mm/s]

2 16.42 16.70 1.705 0.931 1.026 12 24.67 25.04 1.50 0.961 1.100 Tab.9A: HTPB comparison between experiments and CFD with radiative heat flux.

Test r[%] 2 10.18

Exp. C*[m/s] 1614.80

C*[m/s] 1723.69

C*[%] 6.74

14.46 1689.00 1620.51 0.35 12 Tab.9B: HTPB comparison between experiments and CFD with radiative heat flux.

qrad T 4 T 4 (1 e k )
where is the emissivity, T is the flame temperature and is the Stefan-Boltzmann constant. k is another parameter, determined using the following formulation:

In these cases, regression rates and therefore fuel mass flow rates have been overestimated, so that pressures result to be higher than the corresponding experimentally-determined values. Errors on pressures are again smaller than those on regression rate, due to the fact that fuel mass contributes only in a small proportion to the total mass flow rate within the combustion chamber. The error in the regression rate values determined by the CFD results now in an overestimation of 10-15%, which proves the importance of considering the effect of radiative heat flux for HTPB regression rate evaluation, especially at low Gox values. In these cases in fact, radiative heat flux can be as high as convective heat flux, thus contributing to approximately 50% of the total wall heat flux.

k 0.51 0.113 (O / F )
where O/F is the ratio of the oxidiser to fuel mass flow rate. This correlation has been found assuming that the soot particles are in thermal equilibrium with the hot gases surrounding them. The flame temperature has been estimated as the 95% of the equilibrium flame temperature calculated by a thermo-chemical code [10].

Fig.9: Temperature plot: only convective contribution (above) and radiative plus convective heat flux (below).

Injector Modelling All the tests used for these analyses have been conducted using the same axial injector. As previously said, it has been simulated using an inlet boundary condition imposing fixed oxidizer mass flow rate and temperature. This boundary has been applied to a circular planar patch created at the inlet surface of the pre-combustion chamber and having the same diameter of the real injector nozzle. This simulation strategy has proved to be appropriate, in fact CFD results show some recirculation areas at the head-end of the fuel grain, as seen during the experiments, which are fully reported in [11].

Fig.10: Velocity vector plot: axial injection causes two recirculation vortices in the pre-combustion chamber.

As can be seen in Fig.10, two different recirculation areas are present in the pre-combustion chamber, which are responsible for a local turbulence and temperature increase; they also help flame stabilisation at the grain head-end.

V. Conclusions
The purpose of the simulations presented in this paper is to verify if CFD can simulate regression rate in a specific hybrid rocket geometry, as a function of the wall heat flux. CFD results are considered acceptable as long as they are qualitatively coherent

with the experiments. The main objective is to prove that CFD can be used to study different rocket configurations, predicting which one is the most promising in terms of performance optimization. Another important goal is to assess if CFD can predict regression rate inside this type of rockets and what are the limitations in this regard. As already found in [5] and [6], when simulating a rocket geometry where both the oxidiser and fuel mass flow rates are imposed and are maintained constant throughout the entire simulation, both combustion chamber pressure and C* are determined with an error lower than 8%, which is considered as a good result. When the fuel mass flow rate is not imposed, but calculated directly from the regression rate derived from the CFD, the errors remain very similar in terms of C* and pressure, but regression rate itself differs from the average experimental value by far more than 8%. In particular, the difference between the experimental and CFD average regression rate results in an underestimation around 50% for the HTPB test cases and around 30% for HDPE. The error in the HTPB regression rate determination can partially be attributed to the fact that the radiative heat flux contribution has been neglected in the first simulations, which calculated regression rate from the convective wall heat flux only. In fact, for the Gox range simulated in this work (below 100 kg/m2s) for HTPB, it is not true that radiation can be neglected and that it does not give a significant contribution to the wall heat flux. This has been proved by two different simulations, carried out for HTPB tests n. 2 and 12. In both of them, adding a contribution to the wall heat flux due to radiation and derived from the empirical data available in the literature, it has been possible to reduce the error in regression rate determination, so that regression rate was finally overestimated by approximately 10-15%. Without considering any radiative heat flux, regression rate is determined with a lower error in the HDPE test cases compared to those of HTPB. This is due to the fact that HDPE produces less soot particles than HTPB. Unfortunately, no accurate data about HDPE soot production is available. Moreover, no clear information exists regarding the radiative heat flux relative importance with respect to convective heat for HDPE. Therefore, no radiation models nor radiative heat flux contributions have been introduced in the HDPE test cases. It is important to highlight that no tuning coefficients have been used to obtain these results for CFD regression rate estimation. Moreover, no appropriate correcting factors have been introduced to make these results fit the experimental findings in a better way.

The reason of the CFD regression rate difference from the experimentally-determined one is connected to the various approximations in terms of the models applied and specifically, it depends on the wall functions and turbulence models, which should be validated for these kind of problems. Finally, the way axial injection has been simulated has proved to produce an internal flow inside the combustion chamber, which is very similar to that found during the corresponding experiments [11].

of Propulsion and Power, Vol. 21, No. 4, July August 2005 [8] Carmicino C.,Orlandi O.,Russo Sorge A. et al.Basic Aspects of the Hybrid Engine Operation. 45th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit. AIAA-2009-4937. Denver,CO, 2009 [9] M. Chiaverini,N. Serin,D. Johnson,Y.C. Lu,K. K. Kuo,G.A. Risha.Regression Rate Behavior of Hybrid Rocket Solid Fuels. Journal of Propulsion and Power. Vol.16, n.1, Jan-Feb 2000 [10] M. Chiaverini,K.K. Kuo.Fundamentals of Hybrid Rocket Combustion and Propulsion, 2006 [11] Carmicino C.,Russo Sorge A.The Effects of Oxidizer Injector Design on Hybrid Rockets Combustion Stability. 42nd AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA-20064677. Sacramento, CA, 2006

VI. References
[1] Serin N.,Gogus Y.A., Navier-Stokes Investigation on Reacting Flow Field Of HTPD/O2 Hybrid Motor and Regression Rate Evaluation, 39th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA-2003-4462. Huntsville, AL, 2003 [2] Antoniou A.,Akyuzlu, A Physics Based Comprehensive Mathematical Model to Predict Motor Performance in Hybrid Rocket Propulsion System, 41th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA-2005-3541. Tucson, AZ, 2005 [3] Guobiao C. and Hui T.. "Numerical Simulation of the Operation Process of a Hybrid Rocket Motor". 42nd AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit. AIAA-2006-4506. Sacramento, CA, 2006 [4] M. Frey, J. Gorgen,G. Fassbender.Modelling of Combustion and Flow in a Hybrid Thrust Chamber. 47th European Conference for Aerospace Sciences EUCASS [5] N. Bellomo,M. Lazzarin,F. Barato and M. Grosse. Numerical Investigation of the Effect of a Diaphragm on the Performance of a Hybrid Rocket Motor. 46th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit. AIAA-2010-7033. Nashville, TN, 2010 [6] M. Lazzarin,M.Faenza,F. Barato,N.Bellomo et al. CFD Simulation of a Hybrid Rocket Motor with Liquid Injection. 47th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit. AIAA-2011-5537. San Diego,CA, 2011 [7] Carmicino C.,Russo Sorge A.Role of Injection in Hybrid Rockets Regression Rate Behavior. Journal