You are on page 1of 18

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/309545609

Ignition Transition of Coaxial Kerosene/Gaseous Oxygen Jet

Article  in  Combustion Science and Technology · December 2016


DOI: 10.1080/00102202.2016.1211865

CITATIONS READS

5 757

3 authors:

Dohun Kim Min Son


ADD Universität der Bundeswehr München
33 PUBLICATIONS   64 CITATIONS    72 PUBLICATIONS   341 CITATIONS   

SEE PROFILE SEE PROFILE

Jaye Koo
Korea Aviation University
141 PUBLICATIONS   786 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

MaST (dtec.bw) View project

An Experimental Research on Internal Flow of Solid Rocket Motor View project

All content following this page was uploaded by Jaye Koo on 01 October 2019.

The user has requested enhancement of the downloaded file.


Combustion Science and Technology

ISSN: 0010-2202 (Print) 1563-521X (Online) Journal homepage: http://www.tandfonline.com/loi/gcst20

Ignition Transition of Coaxial Kerosene/Gaseous


Oxygen Jet

Dohun Kim, Min Son & Jaye Koo

To cite this article: Dohun Kim, Min Son & Jaye Koo (2016) Ignition Transition of Coaxial
Kerosene/Gaseous Oxygen Jet, Combustion Science and Technology, 188:11-12, 1799-1814,
DOI: 10.1080/00102202.2016.1211865

To link to this article: http://dx.doi.org/10.1080/00102202.2016.1211865

Published online: 28 Oct 2016.

Submit your article to this journal

Article views: 36

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=gcst20

Download by: [Korea Aerospace University] Date: 07 December 2016, At: 04:27
COMBUSTION SCIENCE AND TECHNOLOGY
2016, VOL. 188, NOS. 11–12, 1799–1814
http://dx.doi.org/10.1080/00102202.2016.1211865

Ignition Transition of Coaxial Kerosene/Gaseous Oxygen Jet


Dohun Kim , Min Son, and Jaye Koo
Graduate School of Korea Aerospace University, Goyang, Gyeonggi, Republic of Korea

ABSTRACT ARTICLE HISTORY


The pressure and temperature during liquid rocket engine ignition Received 28 October 2015
drastically increase because of energy generation from the severe com- Revised 12 April 2016
bustion reaction of pure oxidizer and fuel. Studies on the ignition of Accepted 13 April 2016
oxygen/kerosene combustion are scarce. This research observes ignition KEYWORDS
transition of a gaseous oxygen/kerosene spray by directly visualizing the Combustion visualization;
combustion flow field using a windowed combustor and high-speed Ignition timing; Ignition
shadowgraph imaging technique. The hydrodynamic characteristics of a transition; Kerosene; Oxygen
propellant feed system are investigated before the ignition experiments
by analyzing the time responses of propellant injection pressures and a
high-speed movie of developing spray during injection. The experi-
ments are performed with a 22.5 ms fuel pre-injection sequence. The
high-speed shadowgraph imaging and dynamic pressure measurement
results are analyzed. The effects of ignition timing on the ignition
transition are focused on. The early ignition timing results in the
smoothest ignition and longest ignition delay time from the propellant
injection related to an ignitable transient spray condition. The combus-
tion pressure rapidly increases with ignition timing delay. A peak pres-
sure caused by propellant mixture accumulation is also observed. The
ignition delay is less than 5 ms, and slightly decreases as the ignition
timing is further delayed.

Introduction
Large liquid rocket engine propellants are mixed and burned under supercritical pressure and
temperature conditions. The critical temperature and pressure of these liquid rocket engine
propellants are generally much lower than the propellant feed, injection, and combustion
conditions (Manski et al., 1998; Sutton, 2003a, 2003b). In the supercritical condition, the
distinction between the liquid and gas phases becomes obscure. The thermodynamic and
fluid dynamic characteristics of the liquid (high density) and gas (low viscosity, zero surface
tension, and high thermal diffusivity) phases coexist (Cengel and Boles, 2005; Smith et al., 2004;
Zhong et al., 2012). Propellants are supplied and injected under a subcritical condition before
ignition. Both pressure and temperature dramatically increase at the moment of ignition;
hence, a thermodynamic state transition from subcritical to supercritical occurs. Many
researchers have studied the ignition and high-pressure combustion of the liquid rocket engine
by visualizing the combustion flow field and the free radical distribution.
Gurliat et al. (2003) studied the flame propagation velocity at an ignition transition of liquid
oxygen (LOx)/gaseous hydrogen (GH2) combustion in various spraying conditions. They

CONTACT Jaye Koo jykoo@kau.ac.kr Graduate School of Korea Aerospace University, 76, Hanggongdaehak-ro,
Deogyang-gu, Goyang-si, Gyeonggi-do 10540, Republic of Korea.
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/gcst.
© 2016 Taylor & Francis
1800 D. KIM ET AL.

explained the ignition process as four phases: (i) primary ignition, (ii) flame lift-off, (iii) flame
propagation, and (iv) flame anchoring. Cuoco et al. (2004) focused on the transient behavior
of the LOx/GCH4 ignition. They observed three different ignition scenarios depending on a
propellant spraying condition. However, their experiments were conducted under low com-
bustion pressure (i.e., below 2 bar). Mayer et al. (1998, 2001) observed a reacting LOx/GH2
spray under the sub- to supercritical condition. They also morphologically observed LOx/GH2
spray during an ignition process and reported that the rapid vaporization of cryogenic oxygen
affected the transient ignition characteristic. The combustion pressure in the ignition experi-
ment (i.e., 1.8 bar) was much below the critical pressures of the oxidizer and fuel. Kim et al.
(2011) simulated coaxial LOx/GH2 spray under supercritical pressure using flamelet model.
They focused on radial and axial distributions of thermodynamic properties and free radicals
at steady state. Höglauer et al. (2015) investigated a combined numerical method for a heat
load prediction of GOx/kerosene rocket combustor. They validated the simulation codes by
comparing with experimental results.
Most studies on ignition of liquid rocket engine combustor have focused on cryogenic
propellants, such as oxygen, hydrogen, and methane. Experimental studies on the ignition
transition of the GOx/kerosene spray are scarce. The present research visualizes a transient
ignition behavior of a GOx/kerosene spray using a windowed liquid rocket combustor. This
combustor is operated at a pressure around the critical pressure of kerosene. The ignition
transitions of the GOx/kerosene spray are morphologically analyzed. In addition, effects of
ignition timing on the ignition process are observed.

Experimental setup
Combustion experiment facilities
A shear coaxial injector was designed for the high-pressure combustion experiment. As
shown in Figure 1a, the injector assembly comprised two parts, namely, the center post
and the outer body. The pressure drops at the gas and liquid injectors were controlled
using orifices inserted in the propellant inlet ports of the injector. Table 1 shows the
geometrical dimensions and operating conditions. A windowed combustor was designed
to directly observe the high-pressure combustion flow field of the gaseous oxygen
(GOx)/kerosene spray (Figure 1b). The inner diameter and overall length from the
injector face plate to the nozzle throat are 22 mm and 182 mm, respectively. The
combustor body comprised the three following modules: (i) an upstream module for
the flame visualization and an igniter attachment; (ii) a middle module for adjusting the
combustor length; and (iii) a downstream module with an exhaust nozzle and a water
cooling channel. The propellants were ignited using a motorcycle spark plug (NGK
CR8EIX) and a high voltage induction coil (11 kV, 198 J).
Figure 2 presents a schematic of the propellant supplying system. The steady-state mass
flow rate of GOx was calculated by multiplying the volume flow rate measured using a
turbine flow meter and the oxidizer density estimated from the static pressure and
temperature of the oxidizer at the flow meter section. The steady-state flow rate of
kerosene was calculated by multiplying the volume flow rate measured using a displace-
ment-type flow meter and the kerosene density of 798 kg/m3.
COMBUSTION SCIENCE AND TECHNOLOGY 1801

Figure 1. Schematic of the (a) shear coaxial-type injector and (b) subscale windowed combustor.

Table 1. Time delays during the initiation period of the propellant injection.
Mean value
Time delays From To (ms)
τsq_OF Transmission of ox. valve opening signal Transmission of fuel valve opening signal 70
τop_F Fuel valve opening signal transmission Beginning of fuel injection pressure 38.3
increase
τop_O Ox. valve opening signal transmission Beginning of ox. injection pressure increase 120.58
τinj_F Beginning of fuel injection pressure increase Fuel injection pressure reaches peak value 8.52
τinj_O Beginning of ox. injection pressure increase Ox. injection pressure reaches peak value 89.16
τop_FO Beginning of fuel injection pressure increase Beginning of ox. injection pressure increase 12.28
τact_FO Beginning of fuel injection in shadowgraph Beginning of ox. injection in shadowgraph 22.53
movie movie
τign_activ Beginning of ox. injection pressure increase Transmission of igniter operation signal 0–35a
τign Transmission of igniter operation signal Beginning of combustion pressure increase —
a
Experimental variable.

The dynamic pressure transducers (PCB 113B24) were installed upstream of the
oxidizer, fuel injectors, and combustion chamber to precisely measure the hydrodynamic
characteristics of the propellant supplying system and the combustion pressure. The
sampling rate was 80 kHz. The dead volume between the diaphragm of the dynamic
pressure transducer and the measuring port was minimized to reduce the damping effect
of compressible flow.
Meanwhile, orifices measuring 3.0 mm and 0.5 mm in diameter were inserted at the
inlet ports of the oxidizer and the fuel injectors, respectively, to prevent reverse flow by a
pressure peak at the moment of ignition.
1802 D. KIM ET AL.

Figure 2. Piping and instrument diagram of the combustion experiment facility.

Figure 3. Plane schematic view of the shadowgraph visualization setup.

Optical observation method


The reacting spray field at the ignition process was observed using the shadowgraph visualiza-
tion technique. The luminosity of the oxygen/kerosene flame was very high because of a black
body emission from carbon particles. Therefore, a light source with higher luminosity was
required to exclude the light emission from the shadowgraph images. The light source used in
the experiments was a 27 Watt high-power light emitting diode, which emitted a luminous
flux of 2100 lumens. The light was collimated to a parallel light using a plano-convex lens with
a focal length of 300 mm as shown in Figure 3. The parallel light passed by the combustion
flow field through two quartz windows. This light was focused on the CCD arrays of a Photron
APX-RS high-speed CCD camera through a Nikkor 105 mm microlens.

Flow velocity measurement at the ignitor location


The transient characteristic of the gaseous oxidizer jet at the ignition process was observed
using a hot wire anemometer by measuring the velocity magnitude at the igniter electrode,
where the ignition energy was produced. The Dantec Streamline constant temperature
anemometry system and the 55P11 single wire probe were used. The sampling rate was 80
kHz. Gaseous nitrogen was used as a fluid because oxygen oxidized the hot wire probe.
COMBUSTION SCIENCE AND TECHNOLOGY 1803

Experimental conditions
Hydraulic characteristics of the propellant feed system
An interval is observed between the valve opening signal transmission time and the time
the injection was initiated at the injector tip. The injection delays are measured by
analyzing the time traces of the valve opening signal and injection pressure and the
high-speed visualization movie. Figure 4 shows some considerable time delays when the
propellant valves were opened with a certain interval (τsq). Table 1 explains the time
delays. Subscript “OF” denotes the time delay between something of oxidizer and fuel.
Subscript “FO” denotes the opposite meaning.
The time delay from the valve opening signal transmission to the initiation of the
injection pressure increase (τop) is dominantly determined by a mechanical delay of the
pneumatically actuated ball valve. The time delay from the initiation of the injection
pressure increase to the transition time to the steady state (τinj) includes the required times
for the full opening of the valve and for charging the volume between the valve and the
injector. Therefore, the hydrodynamic characteristics of the propellant feed system are
affected by whether the fluid is compressible or incompressible. Consequently, the interval
between the transmission time of the gaseous oxidizer and the liquid fuel valve opening
signals (τsq) is different from that between the increasing times of the injection pressure
(τop) and the actual injection times (τact). The ignition timing is one of the important
variables affecting the ignition process when the ignition energy is supplied during the
developing period of the propellant spray.
This study programs the ignition sequence to inject the fuel jet first. The ignitor is activated
during the transient period of the oxidizer jet. The effects of the time delay from the oxidizer
injection pressure increase to the ignitor activation (τign_activ) are investigated. Mechanical and
hydrodynamic delays do not exist in the igniter drive system. Therefore, the ignition energy is
assumed to be simultaneously supplied with the transmission of the ignitor activation signal.

Figure 4. Time delays in the practical liquid rocket engine system.


1804 D. KIM ET AL.

Figure 5. Time traces of the oxidizer and fuel injection pressure in the oxidizer valve 70-ms pre-opened
sequence. Time zero indicates the time when the fuel valve opening signal is transmitted.

Figure 6. Signal timing chart of the operation sequence.

The ignition delay time (τign) is defined as the time interval from the transmission of the
igniter activation signal to the increase of the combustion chamber pressure.
The values in the rightmost column of Table 1 are the mean values measured in the
τsq_OF = 70 ms condition. The interval between the actual injection time of the fuel and
the oxidizer (τact_FO) is measured from the high-speed movie. As shown in Figure 5, the
injection pressures of the fuel and the oxidizer are increased to 38.3 ms and 120.58 ms,
respectively, after the valve opening signal transmission. Accordingly, 8.52 ms and 89.16
ms are needed to reach the peak values. The oxidizer injection pressure begins to increase
to 12.28 ms, faster than that of the fuel (i.e., τop_FO = 12.28 ms). τact_FO is 1.83 times longer
than τop_FO. The difference between τop_FO and τact_FO is caused by the differences in the
internal geometries of the injector and in the location of the dynamic pressure transducer.
COMBUSTION SCIENCE AND TECHNOLOGY 1805

Ignition sequence
The signal timing chart of the ignition sequence is illustrated in Figure 6. The 22.5 ms fuel
preinjection sequence is applied for the entire ignition experiment by transmitting the
oxidizer valve opening signal 70 ms faster than that of fuel. This ignition sequence is used
to eliminate the effect of pressure fluctuation with a frequency of 740 Hz and a maximum
amplitude of 50 bar. The fluctuation is induced by a water-hammering phenomenon. The
ignition time is varied in 5 ms steps to observe the effects of τign_activ on the ignition
characteristic. Moreover, the effect of the prefilled gas composition (Kim et al., 2016b) is
removed by charging the combustion chamber with gaseous nitrogen.

Results and discussions


Transient characteristics
Figure 7 shows the temperature and static pressure measurements during the ignition
experiments. Time zero indicates an initiation of the fuel injection pressure increase. The
fuel static pressure is measured upstream of the fuel valve. Therefore, the values before the
fuel valve opening indicate a stagnation pressure. Both the fuel and the oxidizer tempera-
tures are about 290 K. Table 2 summarizes the mean values of the other measurements.
The upper graph of Figure 8 plots the dynamic pressure traces under three representative
τign_activ conditions. The lower graph of Figure 8 illustrates the instantaneous oxidizer to the
fuel mass flow rate ratio estimated from the differential pressures at the propellant injectors.
The oxidizer injection pressure increases at ~20 ms. The actual injection time is between 25 ms
and 30 ms. The combustion chamber pressure in τign_activ = 0 ms begins to gradually increase

Figure 7. Time traces of the propellant temperature, injection static pressure, and combustion chamber
static pressure.
1806 D. KIM ET AL.

Table 2. Mean steady-state parameters in 5000 averaged samples.


Parameter Value Parameter Value
dPO,inj 32.5 bar dPF,inj 3.50 bar
TO,inj 291.4 K TF,inj 290.5 K
m_O 30.2 g/s m_F 10.8 g/s
CD,O 0.494 CD,F 0.847
Pcc 25.8 bar (abs.) O/F ratio 2.80

Figure 8. Time traces of the combustion chamber dynamic pressure depending on the igniter activa-
tion time (upper graph) and instantaneous O/F ratio transition of cold spray in equal period (lower
graph). Time zero corresponds to the time when the fuel valve opening signal is transmitted.

10‒15 ms after the actual oxidizer injection. The combustion pressures rapidly increase and
show pressure peaks caused by the combustion of the precharged mixture when τign_activ
increases to 20 ms and 30 ms. The effects of τign_activ on τign will be quantitatively discussed in
the “Ignition Delay” section.

Ignition flow-field visualization


Table 2 shows the flow and combustion conditions. The steady-state combustion pressure
is 25.8 bar, which is similar with the critical pressure of kerosene (~25 bar). The total mass
flow rate is 41.0 g/s, while the oxidizer-to-fuel mass ratio is 2.80. The plots in Figures 9
and 10 show dynamic pressure traces during the ignition processes in τign,activ = 0 ms and
35 ms conditions, respectively. The oxidizer injection static pressure, PO,inj, is measured
downstream of the oxidizer control valve. Therefore, PO,inj is increased from 0 bar after
the oxidizer valve opening. By contrast, the fuel injection static pressure, PF,inj, is
COMBUSTION SCIENCE AND TECHNOLOGY 1807

Figure 9. Propellant injection-ignition transition-flame stabilization process in τign,activ = 0 ms condition.


1808 D. KIM ET AL.

Figure 10. Propellant injection-ignition transition-flame stabilization process in τign,activ = 35 ms


condition.
COMBUSTION SCIENCE AND TECHNOLOGY 1809

measured upstream of the fuel control valve. Therefore, PF,inj is dropped from the fuel
supply pressure when the valve opened.
As shown in the combustion chamber dynamic pressure (Pcc,dyn) plots in Figures 9 and
10, the propellant spray ignites at about 41 ms and 62 ms, respectively, after the fuel
injection pressure increase is initiated. The series of shadowgraph images in Figures 9 and
10 recorded in sync with the dynamic pressure measurements shows representative events
during the ignition processes. In the ignition sequence setup of this study, the ignition
process is categorized into five steps as follows: (1) fuel injection; (2) oxidizer injection and
initial atomization; (3) ignition of developed (or developing) spray; (4) transition; and (5)
flame stabilization. The fuel/oxidizer injection and the initial atomization steps are
independent of τign,activ. However, the ignition and transition process is significantly
different according to the ignition timing. The ignition in the τign,activ = 0 ms condition
particularly shows a moderated igniting spray behavior. Finding the exact scene of ignition
by observing the original shadowgraph images is difficult because the light source for the
shadowgraph imaging is much brighter than the flame-emitting light at the moment of
ignition. Hence, the shadowgraph still images are converted into x–y gradient magnitude
images, which allows a clear detection of the phase-change phenomenon (Kim et al.,
2016a) (Figures 11 and 12). The shadowgraph images J, K, and L in Figure 9 correspond to
the converted images B, C, and D in Figure 11. The dark region denoted by a dashed circle
in Figure 11A is caused by the dense fuel mist, which covers the background light. The
temperature increases when it ignites from the igniter electrode gap. It then preferentially
vaporizes fuel droplets with small diameter (arrow-pointed region, Figures 11B and 11C).
Consequently, the finely atomized portion of the fuel spray disappears. Only the un-

Figure 11. Rapid vaporization of atomized fuel at the moment of ignition in τign,activ = 0 ms condition.
The original images are converted to x–y gradient magnitude images.

Figure 12. Reacting coaxial kerosene/oxygen spray during the flame stabilization process in τign,activ = 0
ms condition.
1810 D. KIM ET AL.

disintegrated core of the fuel spray remains as the flame propagates (dashed circle,
Figure 11D). In contrast, the ignition moment in the case of τign,activ = 35 ms is
significantly distinguishable by its severe luminous flame generation (Figures 10J, 10K,
and 10L). The luminous flame with continuous spectrum is an index of a fuel-rich
combustion reaction of hydrocarbon fuel (Hayashi et al., 2011; Lee and Seo, 2015; Linck
and Gupta, 2003). Therefore, it could be deduced that the ignition of the τign,active = 35 ms
case occurs under a lower O/F ratio condition because the oxidizer mass flow rate is
higher than that in the τign,activ = 0 ms case, accelerates the disintegration of the liquid fuel
jet, and increases the global O/F ratio in the combustion chamber. In a similar study on
the LOx/GH2 spray ignition, Mayer et al. (2001) surmised that the ignition occurred in a
fuel-rich condition because a gaseous hydrogen jet, which encloses a center LOx jet, blew
away a large portion of locally vaporized oxygen downstream and the combustion
chamber was filled with plenty of gaseous fuel. The propellant combination is different
from the experiments in the present study. However, the results of both studies could be
interpreted that the initial interaction between the gaseous and liquid jets sensitively
affects the ignition process (e.g., equivalence ratio).
The transition processes are strongly affected by the global O/F ratio of the precharged
mixture (sequential images, Figures 12 and 13). The gaseous oxidizer injection in the τign,
activ = 0 ms case is temporarily discontinued for about 3 ms (Figures 12A and 12B). The
oxidizer injection is concurrently restarted with atomizing fuel jet (Figures 12C and 12D).
Thereafter, the density gradient patterns mainly caused by the temperature gradient of the
annular oxidizer jet appear in the shadowgraph images (Figures 12E and 12F). Moreover,
the combustion chamber pressure increases as the reacting oxygen/kerosene spray gen-
erates heat from 3 bar (Figure 12D) to 25.28 bar (Figure 12L). No particular behavior is
observed during the flame stabilization process (Figures 12G to 12L), except for the
intensity of the periodically generated luminous flame at the lower half region of the
reacting spray. No flame is detectable at the upper half region of the reacting spray
because of the much higher intensity of the shadowgraph light source than the anchored
flame. However, the image recorded without the light source in Figure 14 provides an
evidence of anchored flame.
The transition process of the τign,activ = 35 ms case could not be clearly observed
because of the intense luminous flame covering the entire visualization window (Figures

Figure 13. Reacting coaxial kerosene/oxygen spray during the flame stabilization process in τign,activ =
35 ms condition.
COMBUSTION SCIENCE AND TECHNOLOGY 1811

Figure 14. Luminous flame image during the steady-state combustion: pictured in 1 μs exposure time
without illumination.

13A to 13D). However, the combustion chamber pressure clearly begins to rapidly
increase after the transition process, which seems to be a fuel-rich combustion. The
ignition in the τign,activ = 35 ms case approximately occurs 21 ms later than that in the
τign,activ = 0 ms case. The combustion pressure in the τign,activ = 35 ms case reaches the
mean steady-state value (25.8 bar) 4.7 ms earlier than that in the τign,activ = 0 ms case.

Ignition delay
Figure 15 shows the effect of the ignition timing after oxidizer injection (τign,activ) on the
ignition delay time (τign). The red circle denotes an averaged ignition delay time measured in
the τign,activ = 0 ms case. The other symbols represent the ignition delay times in case of τign,activ
> 0 ms (i.e., cases of delayed igniter activation). The ignition timing is not equal to the oxidizer
injection timing because of its periodic arc discharging characteristic (320 Hz) and because the
ignitor is operated prior to the oxidizer injection at the experiments in the τign,activ = 0 ms case.
Hence, a bias error of ±1.6 ms should be considered in the ignition delay time measurement so
the mean data point (red circle symbol, Figure 15) has a longer error bar. τign in the τign,activ = 0
ms case is thought to be a minimum time when the mixture is in its ignitable condition and is
about 11.3 ± 1.6 ms in average value. The velocity measurement of the transient gaseous jet
provides some clues on this. Figure 16 plots the velocity magnitude at the igniter location. The
injection conditions, such as feeding pressure, pipe line configurations, and flow control
devices, are the same with those in the ignition experiment, except that the injection fluid is
nitrogen. The turbulence intensity increases from 0.109 for the 5.6‒6.3 ms period to 0.441 for
the 10‒10.7 ms period as the injection velocity increases. From this, it could be deduced that
the small-scale turbulent mixing between the oxidizer and the fuel dramatically enhances at
around 10 ms and causes the ignition.
Theoretically, the sum of τign,activ and τign should be 11.3 ± 1.6 ms. Moreover, the
decreasing trend of τign complies τign,activ + τign = 11.3 ± 1.6 in the range of τign,activ
1812 D. KIM ET AL.

Figure 15. Ignition delay times depending on the igniter activation time.

Figure 16. Time evolution of the gas flow velocity at the spark igniter electrode.

between 0 and 8. However, τign should not be zero because the ignition process necessarily
accompanies the delay time (minimum τign). τign is below 5 ms for the entire experimental
cases when τign,activ increases to over 8 ms. τign then gradually decreases according to the
increasing τign,activ. The changes of τign in these conditions should be induced by changes
in the spraying and mixing conditions.
COMBUSTION SCIENCE AND TECHNOLOGY 1813

Conclusions
The ignition transition of a GO2/kerosene spray was observed by directly visualizing the
igniting spray using the shadowgraph imaging technique. The effects of the ignition
timing were investigated by analyzing the high-speed shadowgraph movie and the time
responses of the injection and combustion pressures at the moment of ignition. The initial
interaction between the gaseous and liquid jet influenced from varying the ignition timing
within 30 ms significantly affected the equivalence ratio of the ignition flame. The
simultaneous execution of an igniter activation and an oxidizer injection resulted in the
gentlest ignition behavior and the longest ignition delay time from the propellant injec-
tion, which were related with the fulfillment of the developing spray’s ignitable condition.
The combustion pressure rapidly increased and showed peak pressure at the moment of
ignition as the ignition timing retarded over 10 ms from the oxidizer injection. This result
was attributed to the ignition of the accumulated propellant mixture. In addition, the
ignition delay was below 5 ms and slightly decreased as the ignition timing further
retarded. The importance of ignition timing was identified although the correlation
between diverse ignition variables and ignition characteristic parameters could not be
derived because of insufficient data. The thermal durability of the windowed combustor
was under improvement. Therefore, it was expected to present better results in further
research.

Funding
This work was supported by a National Research Foundation of Korea (NRF) grant funded by the
Korean Government (Ministry of Science, ICT and Future Planning) (Nos. NRF-2012M
1A3A3A02033146 and NRF-2013R1A5A1073861 through the SPRC of Seoul National University).

ORCID
Dohun Kim http://orcid.org/0000-0002-2360-4136

References
Cengel, Y., and Boles, M. 2005. Thermodynamics: An Engineering Approach, 5th ed., McGraw-Hill,
New York.
Cuoco, F., Yang, B., Bruno, C., Haidn, O.J., and Oschwald, M. 2004. Experimental investigation on
LOx-CH4 ignition. AIAA Paper 2004-4005. Presented at the 40th AIAA/ASME/SAE/ASEE Joint
Propulsion Conference and Exhibit, Fort Lauderdale, FL, July 11–14.
Gurliat, O., Schmidt, V., Haidn, O.J., and Oschwald, M. 2003. Ignition of cryogenic H2/LOX sprays.
Aerosp. Sci. Technol., 7, 517–531.
Hayashi, J., Watanabe, H., Kurose, R., and Akamatsu, F. 2011. Effects of fuel droplet size on soot
formation in spray flames formed in a laminar counter flow. Combust. Flame, 158, 2559–2568.
Höglauer, C., Kniesner, B., Knab, O., Schlieben, G., Kirchberger, C., Silvestri, S., and Haidn, O.J.
2015. Modeling and simulation of a GOX/kerosene subscale rocket combustion chamber with
film cooling. CEAS Space J., 7, 419–432.
Kim, D., Lee, K., and Koo, J. 2016a. Effects of wall-injection length on spray and combustion in a
coaxial porous injector. J. Propul. Power., 32, 533–541.
Kim, D., Shin, B., Choi, S., and Koo, J. 2016b. Effects of pre-filled gas and spraying condition on
ignition of oxygen/kerosene spray combustion. Combust. Flame., 167, 1–13.
1814 D. KIM ET AL.

Kim, T., Kim, Y., and Kim, S. 2011. Real-fluid flamelet modeling for gaseous hydrogen/cryogenic
liquid oxygen jet flames at supercritical pressure. J. Supercrit. Fluids, 58, 254–262.
Lee, H., and Seo, S. 2015. Experimental study on spectral characteristics of kerosene swirl combus-
tion. Procedia Eng., 99, 304–312.
Linck, M., and Gupta, A.K. 2003. Effect of swirl and combustion on flow dynamics in luminous
kerosene. AIAA Paper 2003-1345. Presented at the 41st Aerospace Sciences Meeting and Exhibit,
Reno, NV, January 6–9.
Manski, D., Goertz, C., Saβnick, H.D., Hulka, J.R., Goracke, B.D., and Levack, D.J.H. 1998. Cycles
for earth-to-orbit propulsion. J. Propul. Power. 14, 588–604.
Mayer, W., Schik, A., Bruno, V., Chauveau, C., Gökalp, I., Talley, D.G., and Woodward, R.D. 1998.
Atomization and breakup of cryogenic propellants under high-pressure subcritical and super-
critical conditions. J. Propul. Power, 14, 835–842.
Mayer, W.O.H., Blazenko, I., Schik, A., and Hornung, U. 2001. Propellant atomization and ignition
phenomena in liquid oxygen/gaseous hydrogen rocket. J. Propul. Power, 17, 794–799.
Smith, J.J., Bechle, M., Suslov, D., Oschwald, M., and Haidn, O.J., and Schneider, G.M. 2004. High
pressure LOx/H2 combustion and flame dynamics preliminary results. AIAA Paper 2004-3376.
Presented at the 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Fort
Lauderdale, FL, July 11–14.
Sutton, G.P. 2003a. History of liquid propellant rocket engines in the United States. J. Propul.
Power, 19, 978–1007.
Sutton, G.P. 2003b. History of liquid-propellant rocket engines in Russia, formerly the Soviet
Union. J. Propul. Power, 19, 1008–1037.
Zhong, F., Fan, X., Wang, J., Yu, G., and Li, J. 2012. Characteristics of compressible flow of
supercritical kerosene. Acta Mech. Sin. 28, 8–13.

View publication stats

You might also like