Encit 2016
Encit 2016
Encit 2016
Abstract. The combustion processes in liquid rocket engines are heavily dependent on quality of atomization. Pressure
swirl injectors are widely used in liquid rocket engines due to their high efficiency of atomization in a reduced volume.
The characteristic geometric parameters of these injectors can be adjusted to control the spray cone angle and
improve spray fineness, thus reducing the chamber length and eventual instabilities. A dual pressure swirl injector is
characterized by two independent concentric chambers which can provide independent rotational levels to a single or
two different liquids to improve mixing and combustion of the propellants within liquid rocket engines. This paper
compares theoretical, semiempirical and experimental results concerning the spray cone angles formed by injection of
ethanol (C2H6O, 95% m/m) and hydrogen peroxide (H2O2, 90% m/m) through a dual pressure swirl injector for
application in a 100 N bipropellant rocket engine. Data were obtained for injection of water or ethanol through both
the primary and secondary chambers and for injection of ethanol in the primary chamber and water through the
secondary chamber of the injector. Experimental data are obtained using photographic techniques and are analyzed by
image processing software developed in Matlab language.
Keywords: Spray Cone Angles, Dual Pressure Swirl Injector, Atomization, Ethanol, Hydrogen Peroxide
1. INTRODUCTION
The injector is responsible for the atomization of liquid propellants into the combustion chamber of a rocket engine.
An efficient atomization allows to significantly increasing the surface area of liquid propellants, ensuring high rates of
evaporation, mixing and burning. Small droplets are required to achieve rapid ignition and establish a flame front of the
next injection head. The larger droplets take longer to burn and thus define the length of the combustion chamber
(Khavkin, 2004).
Ethanol and hydrogen peroxide are green propellants that are being investigated in various research centers and
universities (Wernimont, 2006; Maia, 2012; Matos, 2013) due, among other factors, the low environmental impact of
combustion, ease of handling and storage, low cost, and good availability on the market. Ethanol and hydrogen
peroxide are not hypergolic and therefore require a suitable ignition system or process, for example, auxiliary flames,
additives, catalysts, dischargers or heaters.
A pressure swirl injector consists of one or more inlets into a central vortex chamber, the inlets generally being
tangential, thus providing the spin in the vortex chamber. Finally the fluid emerges from an orifice in the form of a film
around the periphery of the orifice; this film then breaks into a cone of spray droplets. The spray of the liquid produced
at the output of this type of injector has the approximate shape of a hollow-cone (Lefebvre, 1989).
A dual pressure swirl injector is characterized by two independent concentric vortex chambers which can provide
independent rotational levels to a single liquid or two different liquids. In the case of a dual pressure swirl injector a
hollow-cone is formed for each chamber and if the two cones collide there is formation of a single cone. Generally
when the cone angle is increased there is also an increase in the contact of droplets of liquid injected with ambient air,
which improves the atomization process, the heat and mass transfer. However, the reduction in the cone angle improves
the performance of the ignition and extends the limits of stability (Ortmann et al., 1985).
The cone angle is an important external feature of a spray. Due to the interactions of the liquid fuel with air, the
spray curve actually has an approximate bell shape, as shown in Fig. 1, thus presenting the difficulty of measuring the
cone angle.
The spray cone angle is the angle formed by two straight lines in a plane projected from the orifice discharge of
injector, at a specified distance (Vásquez, 2011).
This paper compares theoretical, semi-empirical and experimental results concerning the spray cone angles formed
by injection of ethanol (C2H6O, 95% m/m) and hydrogen peroxide (H2O2, 90% m/m) through a dual pressure swirl
injector for application in a 100 N bipropellant rocket engine. Data are obtained for injection of water or ethanol
through both the primary and secondary chambers and for injection of ethanol in the primary chamber and water
through the secondary chamber of the injector. Experimental data are obtained using photographic techniques and are
analyzed by image processing software developed in Matlab language by Vásquez (2011).
2. THEORETICAL MODEL
The theoretical cone angles were calculated using the approach by Abramovich, as described in detail by Fischer
(2014). Figure 2 shows a scheme with a pressure swirl injector.
The characteristic geometric parameter (K) of a pressure swirl injector is defined by:
As rcv re As R Rrs
K (1)
Ae rs Ae rs Ae
where As is the area and rs the inner radius of the discharge orifice, Ae is the total area and re the radius of tangential
inlet channels, rcv is radius and R the effective radius of the vortex chamber.
Other important parameters of the pressure swirl injectors are the liquid fraction area (ε), dimensionless radius of the
gas vortex (S), discharge coefficient (μ) and spray cone angle (α). All these parameters are directly related to the K by
equations (Bayvel and Orzechowski, 1993):
K
1 2
(2)
(3)
2
2 K
tan (4)
2 1 S 2 4 2 K 2
1 1 2K 2
1 2 K 2 S S 2 2 K 2 2 K 2 ln (5)
S S 2 2K 2
A graphical solution of Eqs. (2), (3), (4) and (5) is shown in Figure 2.
Proceedings of ENCIT 2016 16th Brazilian Congress of Thermal Sciences and Engineering
Copyright © 2012 by ABCM November 07-10th, 2016, Vitória, ES, Brazil
Figure 2. Behavior of the geometric characteristic parameter, discharge coefficient, liquid fraction area and spray
cone angle of the pressure swirl injector.
The liquid fraction area (ε) and dimensionless radius of the gas vortex (S) are defined by:
Al rs2 rsna
2
1
2
rsna
1 S2 (6)
As rs2 rs2
rsna
S 1 (7)
rs
where Al is the area occupied by liquid and rsna the gas core radius in the discharge orifice.
The spray cone angle in the secondary chamber is derived similarly to the primary chamber, but using the following
relation for the characteristic geometrical parameter of the secondary chamber:
Rsec rsec
2
rprim
2
K sec
Ae,sec rsec rprim (8)
where rprim r is the external radius and δ is the wall thickness of the discharge orifice of the primary chamber.
The theoretical model used here for the flow of viscous liquid into a pressure swirl injector is based on Kliachko
approach (Borodin et al., 1967). The characteristic geometric parameter considering the viscous effects (Kλ) must be
corrected by equation:
Rrs K
K (9)
Ae
RR rs 1 B 2 K
2
2
where B R Ae and the characteristic geometrical parameter for secondary chamber must be corrected by:
K ,sec
2
Rsec rsec rprim
2
Ae,sec r r 2
sec
2
prim R r
rprim Rsec rsec rprim
(10)
2
sec sec
where the coefficient of friction (λ) is a function of Reynolds number (Re) of the tangential inlet channels obtained from
the following equation:
25,8
log 2 (11)
log Re 2,58
Proceedings of ENCIT 2016 16th Brazilian Congress of Thermal Sciences and Engineering
Copyright © 2012 by ABCM November, 07-10th, 2016, Vitória, ES, Brazil
Equation (8) for pressure swirl injector is valid for Re 10 3 10 5 . The values determined from this equation are
significantly higher than from other equations commonly used in hydraulic systems (Bazarov et al., 2004).
Thus, it is possible to recalculate the liquid fraction area (ελ) and the dimensionless radius of the gas vortex (Sλ)
considering the viscous effects by Eqs. (2) and (7), respectively.
The friction of the liquid on the vortex chamber wall causes a decrease of the angular motion and also energy
losses. For a more precise calculation of the pressure drop it is necessary to consider the energy losses in the pressure
swirl injector. In the vortex chamber of an injector these losses can be considered as the work of the frictional force on
the trajectory of the liquid. Thus, the corrected discharge coefficient (μλ) is calculated by:
1
(12)
2
K 1 K2
2 i 2
1 C
where C R rs , 1 K C 2 and:
The friction coefficient through the tangential inlet channels depends on the Reynolds number (Re):
4m
Re (14)
l n d e
where m is the mass flow rate, de is the diameter and n the number of tangential inlet channels and μl is the dynamic
viscosity of the fluid.
The total friction loss in the tangential channels (ξi) is computed by:
le
i e (15)
de
where the initial loss coefficient, ξe, is determined from Fig. 3, and the inclination of the tangential inlet channels, αe, is
obtained from:
rcv
e 90 o tan 1
(16)
le
Figure 3. Initial viscous loss coefficient versus inclination of tangential inlet channels.
The corrected spray cone angle (αλ) is calculated considering the effects of geometry and friction losses:
2 K
tan (17)
2 1 S 2 42 K2
In the case of a dual pressure swirl injector, the external collision of the two spray cones formed generates another
spray cone whose final angle (αfinal) can be obtained using the momentum conservation equation. Assuming steady flow
Proceedings of ENCIT 2016 16th Brazilian Congress of Thermal Sciences and Engineering
Copyright © 2012 by ABCM November 07-10th, 2016, Vitória, ES, Brazil
conditions, uniform pressure, uniform velocity at the exit orifices and nobody forces give the equilibrium equations in
the radial (w) and axial (u) velocity components are:
primwprim m
m sec wsec m
prim m
sec w final (18a)
primu prim m
m secusec m
prim m
sec u final (18b)
and it can be written as function of cone semi-angles of inner chamber, αprim, and external chamber, αsec:
prim sec
final w final m primV primsen m secVsec sen
tan 2 2 (19)
2 u final prim sec
m primV prim cos m secVsec cos
2 2
where V 2P is the total injection velocity, ∆P the pressure drop in the injector and ρ the density of the fluid.
Rizk and Lefebvre (Lefebvre, 1989) considered the effects of liquid properties, geometrical parameters, and
injection pressure on the liquid film thickness, and derived the following equation for the spray angle:
0 ,11
Pd s2
6 K 0,15
(20)
l
2
2
Benjamin (1998) validated an equation using a database and modified the coefficients indicated by Rizk and
Lefebvre (Lefebvre, 1989) for large size injectors and obtained:
0 , 067
Pd s2
9.75 K 0, 287
(21)
l
2
2
3. EXPERIMENTAL METODOLOGY
Table 1 shows a general overview of the dual pressure swirl injector dimensions.
Tables 2 and 3 present a general overview of the operational and geometric parameters for design of the primary
and secondary chambers of the dual pressure swirl injector, respectively.
Figure 4 show a picture of the manufactured injector.
Table 2. General overview of the characteristics of the primary chamber of the dual pressure swirl injector.
Input data
Pressure drop – ∆P (MPa) 0,2533
Mass flow rate – m (g/s) 10,0538
Spray cone angle – α (o) 90
Work fluid Ethanol Water
Dynamic viscosity – μl (cP) 1,2 1
Density – ρ (kg/m³) 809,3 1000
Injector parameters
No viscosity With viscosity
Work fluid Ethanol Water Ethanol Water
Friction coefficient (λ) 0 0 0,073 0,0638
Reynolds number (Re) 0 0 5391,8 6821,6
Spray cone angle – α (o) 90 90 82,2085 83,1008
Discharge coefficient (μ) 0,2226 0,2226 0,264 0,2591
Liquid fraction area (ε) 0,4271 0,4271 0,4808 0,4747
Characteristic geometrical parameter (K) 2,9028 2,9028 2,2028 2,2714
Dimensionless radius of the gas vortex (S) 0,8272 0,8272 0,7689 0,7747
Total injection velocity – V (m/s) 5,5683 5,0093 6,6042 5,8321
Mass flow rate – m (g/s) 10,0538 10,0538 11,9241 11,7052
Table 3. General overview of the characteristics of the secondary chamber of the dual pressure swirl injector.
Input data
Pressure drop – ∆P (MPa) 0,2533
Mass flow rate – m (g/s) 40,2152
Spray cone angle – α (o) 114,4947
Work fluid H2O2 Water
Dynamic viscosity – μl (cP) 1,13 1
Density (kg/m³) 1405 1000
Injector parameters
No viscosity With viscosity
Work fluid H 2O 2 Water H2O2 Water
Friction coefficient (λ) 0 0 0,0509 0,0482
Reynolds number (Re) 0 0 10784 12191
Spray cone angle – α (o) 100,3327 100,3327 95,6415 95,8026
Discharge coefficient (μ) 0,155 0,155 0,1649 0,1643
Liquid fraction area (ε) 0,3415 0,3415 0,3595 0,3585
Characteristic geometrical parameter (K) 4,6667 4,6667 4,2033 4,2254
Dimensionless radius of the gas vortex (S) 0,8834 0,8834 0,8708 0,8714
Total injection velocity – V (m/s) 2,9425 3,4878 3,1315 3,6985
Mass flow rate – m (g/s) 40,2152 40,2152 48,0543 48,8002
The experimental setup used for measuring the spray cone angle by photographic techniques consists of a support
to fix the injector on the test bench and a digital camera. The pictures were obtained by a Sony DSC-F828 digital
camera, with 8 megapixels of effective resolution, or 3264×2448 pixels. The support has a mark that serves to indicate a
known length to be used as a reference to relate the number of pixels and the true length of the image, allowing
determine the experimental values of the spray cone angles from the respective images.
Figure 5 shows the GUI (Graphical User Interface) developed in Matlab language by Vásquez (2011), especially
written for this work to process spray images. The use of this GUI is relatively simple and the images can be treated in
JPEG, TIFF or BMP formats.
After taking and selecting the appropriate images, the image processing is done with the GUI developed for this
purpose. Finally, the experimental values of the spray cone angles of these images are registered. After data collection
and treatment the experimental curves are obtained and compared to the theoretical data.
Figure 6 shows theoretical and experimental values of the mass flow rate in the primary and secondary chambers of
the dual pressure swirl injector versus pressure injection.
Figure 7 shows theoretical and experimental values of the discharge coefficient in the primary and secondary
chambers of the dual pressure swirl injector versus mass flow rate.
Figure 8 shows theoretical, semi-empirical and experimental values of the spray cone angle in the primary and
secondary chambers of the dual pressure swirl injector versus mass flow rate.
The test fluids used in all experiments for the primary and secondary chambers were ethanol and water,
respectively.
Figure 9 shows theoretical and experimental values of the mass flow rate in the primary chamber keeping fixed the
mass flow rate of 40 g/s for the secondary chamber of the dual pressure swirl injector versus spray cone angle for
simultaneous injection of ethanol/water or water/water.
As seen in Fig. 6, the mass flow rate of both chambers coincide with the analytical solution considering the viscous
effects in the project conditions equals to 10 g/s and 40 g/s for primary and secondary chambers, respectively. In both
chambers, note that smaller mass flow rates results in greater loss of angular momentum causing a slight discrepancy
between the analytical solution and experimental data. The reverse also occurs.
As seen in Figs. 7, the discharge coefficients of both chambers of the dual pressure swirl injector had good
agreement with the analytical solution considering the viscous effects for the entire operating range, except at low mass
flow rates as mentioned above.
As seen in Fig. 8a, the Benjamin semi-empirical formulation provided the best estimated of the spray cone angle.
The analytical solution considering the viscous effects has good agreement with experimental data, but was 16% lower
in the project condition (10 g/s). This is expected, since the theoretical solution is a function only of the injector
parameters and does not take into account the fluid properties and operating conditions. Another relevant fact is that
analytical solution does not take into account the length of the vortex chamber and discharge orifice.
As seen in Fig. 8b, the analytical solution considering the viscous effects coincide with the experimental value in
the project condition (40 g/s), but the experimental results did not follow the trend of the analytical solution, i.e., the
spray cone angle should increase along with increasing mass flow rate. This behavior is due to the fact that secondary
chamber presented was projected for a low pressure drop and therefore a fully developed spray was obtained only after
Proceedings of ENCIT 2016 16th Brazilian Congress of Thermal Sciences and Engineering
Copyright © 2012 by ABCM November, 07-10th, 2016, Vitória, ES, Brazil
the project condition. The semi-empirical models could not be used for the secondary chamber, because they do not
take into account the effect of the internal geometry of the flow of external injector.
As seen in Fig. 9, the analytical solution agreed with the experimental values. In tests, it was noticed that the
interaction of both sprays leaves occurred when the primary chamber reached 8 g/s for both cases and it can be
evidenced by the identical experimental values.
Figure 9. Theoretical and experimental values of the spray cone angle of ethanol versus mass flow rate in the
primary chamber for simultaneous injection of ethanol/water or water/water.
5. CONCLUSIONS
Theoretical and semi-empirical spray cone angles for injection of ethanol and water in a dual pressure swirl injector
were calculated and compared to experimental data. A graphical user interface was used to determine spray cone angles
from spray images.
The discharge coefficients of both chambers of the dual pressure swirl injector had good agreement with the
analytical solution considering the viscous effects for the entire operating range.
The semi-empirical models could not be used for the secondary chamber, because they do not take into account the
effect of the internal geometry on the flow of external injector.
The Benjamin semi-empirical formulation provided the best estimates of spray cone angles for injection of ethanol
in the injector primary chamber, thus indicating that both liquid physical properties and injector geometry are important
to determination this parameter.
6. REFERENCES
Bayvel, L. and Orzechowski, Z., 1993. Liquid Atomization, Taylor & Francis, Washington, DC.
Bazarov, V.; Vigor, Y. and Puri, P., 2004. Design and dynamics of jet and swirl injectors, liquid rocket thrust chambers:
aspects of modeling, analysis, and design. United States of America: American Institute of Aeronautics and
Astronautics.
Borodin, V.A.; Dityakin, Yu.F.; Klyachko, L.A. and Yagodkin, V.L., 1967. Liquid Atomization, Mashinostroenie,
Moscow.
Fischer, G.A.A., 2014. Injetores centrífugos duais e jato-centrífugos para aplicação em propulsão de foguetes, Master
thesis, Instituto Nacional de Pesquisas Espaciais, São José dos Campos, SP, Brazil.
Giffen, E. and Muraszew, A., 1953. Atomization of Liquid Fuels, Chapman and Hall, London.
Khavkin, Y.I., 2004. Theory and practice swirl atomizers, Taylor & Francis, New York.
Kulagin, L.V. and Okhotnikov S.S., 1970. Combustion of Heavy Liquid Fuels, Mashinostroenie, Moscow.
Lefebvre, A.H., 1989. Atomization and Sprays, Hemisphere, New York.
Maia, F.F., 2012. Novo catalisador para decomposição de peróxido de hidrogênio em micropropulsores de satélites,
Master Thesis, Instituto Nacional de Pesquisas Espaciais, São José dos Campos, SP, Brazil.
Matos, S.C., 2013. Ensaios em câmara de combustão com injetor centrífugo bipropelente líquido criogênico. Master
thesis, Instituto Tecnológico de Aeronáutica, São José dos Campos, SP, Brazil.
Ortmann, J., and Lefebvre, A.H., 1985. Fuel Distributions from Pressure-swirl Atomizers, Journal Propulsion and
Power, Vol. 1, No. 1, pp. 11-15.
Vásquez, R.A., 2011. Desenvolvimento de um injetor centrífugo dual para biocombustíveis líquidos, Master Thesis,
Instituto Nacional de Pesquisas Espaciais, São José dos Campos, SP, Brazil.
Wernimont, E.J., 2006. Monopropellant hydrogen peroxide rocket system: optimum for small scale In:
AIAA/ASME/SAE/JOINT PROPULSION CONFERENCE & EXHIBIT, n. 42, 9 – 12 de Jun de 2006, Sacramento
CA. Proceedings... Sacramento: AIAA.
7. RESPONSIBILITY NOTICE
The authors are the only responsible for the printed material included in this paper.