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

net/publication/369925084

Understanding Interaction between Reactive Jets in Pre-Chamber Ignition of

Gaseous Fuel

Conference Paper in SAE Technical Papers · April 2023

DOI: 10.4271/2023-01-0225

CITATIONS READS
4 265

6 authors, including:

Yixiao Zhang Xiao Ma

Tsinghua University Northwestern Polytechnical University
31 PUBLICATIONS 101 CITATIONS 276 PUBLICATIONS 4,411 CITATIONS

SEE PROFILE SEE PROFILE

Yuwen Fang Yanfei Li

Tsinghua University Tsinghua University
17 PUBLICATIONS 153 CITATIONS 144 PUBLICATIONS 2,686 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Yixiao Zhang on 20 November 2024.

The user has requested enhancement of the downloaded file.

 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

2023-01-0225 Published 11 Apr 2023

Understanding Interaction between Reactive Jets

in Pre-Chamber Ignition of Gaseous Fuel
Yixiao Zhang, Xiao Ma, Jianshu Mao, Yuwen Fang, Yanfei Li, and Shijin Shuai Tsinghua University

Citation: Zhang, Y., Ma, X., Mao, J., Fang, Y. et al., “Understanding Interaction between Reactive Jets in Pre-Chamber Ignition of
Gaseous Fuel,” SAE Technical Paper 2023-01-0225, 2023, doi:10.4271/2023-01-0225.

Received: 31 Oct 2022 Revised: 09 Jan 2023 Accepted: 17 Jan 2023

Abstract
that ignition can be controlled in the jet-crossing zones under

I
n order to improve the ignition capacity and burning rate different initial conditions, with more repeatable and stable
for spark-ignited engines, pre-chamber jet ignition is a developments than the parallel structure case. The cross angles
promising technique to achieve fast premixed combustion in the range of 20°-30° with respect to the pre-chamber axis
and low pollutant emissions. However, few studies focus on show similar performances and more favorable for ignition
the interaction between multiple reacting (i.e. flamelet) or control overall, while the cross region too closed to the nozzles
reacted (i.e. radical) jets, its effect on ignition, exotherm and reduce the overall burning rate. The simulation results show
flow behaviors also remain to be revealed. This paper inves- that local Da number around ignition zone is in the range of
tigated two types of jet interaction under different pre- 0.2-0.4, and with the addition of NH3, the flame regime is
chamber structures, including the jet-crossing and unequal located in thickened and broken reaction zone during early
nozzle designs. Optical experiments under different condi- jet evolution. For the unequal nozzle design with the diameter
tions were conducted in a constant volume combustion of 2 mm and 4 mm, the ignition delay time of the radical jet
chamber with CH4 as fuel, using simultaneous high speed will be reduced by 60 % than the two equal nozzles with the
schlieren and OH* chemiluminescence method. Meanwhile, diameter of 2 mm, because of the suppression of quenching
computational fluid dynamics (CFD) simulations with CH4 in presence of the flame jet. And the flame expansion in hori-
and NH3/CH4 blend fuels were carried out using Converge zontal direction is also improved. In addition, the ignition
software to provide further insights of turbulent flow and patterns are also influenced by different equivalence ratios
ignition process. For the jet-crossing structure, it was found due to the changes of temporal sequence of two jet evolution.

Introduction
application of some carbon-free fuels such as ammonia and

D
ecarbonization of transport is necessary in order to methanol. As an ideal hydrogen carrier, ammonia (NH3) has
keep the global mean temperature rise within 1.5 °C the potential for huge applications but has very low laminar
by 2100’s [1]. For road, maritime and unmanned flame speed and high ignition energy threshold. Jet ignition
aerial power, internal combustion engine (ICE) will still play for pure NH3 and dual fuel can be used to improve the low
an important role in achieving the “carbon neutrality”. It is reactivity, as some recent studies have made successful
the permanent target for ICE to achieve higher thermal effi- attempts [5, 6]. As a “liquid sunshine”, methanol (CH3OH)
ciency. On the other hand, low-/zero-carbon fuels must can be produced from green electricity and it can easily work
be used to reduce CO2 and pollutant emissions. Among at spark ignition mode when used as mono-fuel. To improve
many advanced ignition and combustion modes, pre- the lean burn stability, pre-chamber jet ignition has also been
chamber jet ignition is a promising technology to improve used in large bore methanol engine [7, 8].
the ignition capacity and burning rate for spark-ignited With respect to the inherent mechanisms of pre-chamber
engines, with multiple ignition sites and highly turbulence jet ignition, it is usually acknowledged that the ignition
mixing generated by pre-chamber flame jets in cylinder [2]. process is dominated by the combination of turbulent, thermal
Therefore, knocking can be significantly mitigated for higher and chemical effect, and the competition of mixing scale and
compression ratio, while lean, stratified mixture can chemical reaction scale determines the ignition success prob-
be operated with “active” or “scavenged” pre-chamber, both ability [9, 10, 11]. In 1950’s, Semenov and Gussak firstly intro-
contributing to the thermal efficiency [3, 4]. Meanwhile, it duced the jet ignition concept into engine, called “LAG”
could become one of effective schemes for the engine process [12], which demonstrated that active radicals or
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

2 UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL

incomplete combustion products from quenched flames with more uniformly distribution of flames, where the middle jet
small orifices are critical factor for ignition. Later, Yamaguchi is for quick axial flame propagation and the two weaker side
et al. [13] summarized four ignition patterns by optical experi- jets are for supporting the flame near the jet hole. It was found
ments: well dispersed burning for smallest nozzle diameter, that at later times, its flame projected area develops faster than
flame kernel or front torch ignition for larger nozzle diameter, that in the single-hole and equal three-hole designs, and it
and composite ignition with both mechanisms, and they performs well in engine fuel consumption and
concluded that the intermediate pattern i.e. combination of emission reduction.
chain and thermal reaction is most favorable for lean limit The previous studies above have demonstrated that it is
extension. Recently, with the help of various optical diagnos- necessary to figure out the interaction between multiple
tics and simulation tools, the kinetic and flow characteristics reactive jets with different composition, especially the role of
during jet evolution are resolved more clearly. Biswas et al. active species in the initial ignition zone and subsequent flame
[14] experimentally studied the CH4/H2 jet ignition under development. However, little attention has been paid in two
wide conditions, and found that there is almost no OH* aspects: firstly, the effect of jet crossing or impinging with
radical in a certain distance downstream for “jet ignition” different nozzle angles has been not discussed before; secondly,
mode. Sidey and Mastorakos [15] focused on the auto-ignition the effect of interplay with co-existing of reacting (i.e. flamelet)
behavior in the jet mixing layer and a zero-dimensional jet and reacted (i.e. radical) jet from unequal nozzle diameters
numerical simulation was conducted; it was found that remains to be studied.
ignition still happens with a reaction progress of close to unity Motivated by the research gap above, this study aims to
under extremely high mixing rate. Chi et al. [16] performed investigate the effects on ignition and combustion character-
a parametric direct numerical simulation (DNS) study of istics of jet-to-jet interactions under different pre-chamber
ignition by a pre-chamber hot jet, and it showed that the jet- nozzle structures, including the jet-crossing and unequal
induced local enrichment for radical and heat is essential for nozzle designs. Optical experiments of jet ignition fueled by
successful ignition, while too intense turbulence in the main CH4 were conducted on constant volume combustion chamber
chamber will cause global quenching. These studies show that device, using simultaneous schlieren and OH* chemilumi-
the characteristics of reacting jet (i.e., containing flamelets) nescence method. What’s more, three-dimensional compu-
and reacted jet (i.e., containing radicals) [17] need to be care- tational fluid dynamics (CFD) simulations were also performed
fully considered in pre-chamber design and its matching with to provide further insights of turbulent flow and ignition
engine combustor. Besides, the ignition and subsequent process and auxiliary reference for the experiments. It helps
combustion are strongly correlated to the distribution of better understanding the physicochemical mechanisms in
active radicals. multiple jet interaction and principles for ignition control,
In addition, multiple orifice pre-chamber design is gener- providing the guidelines and data support for the pre-
ally used to generate evenly distributed jets in the main chamber designs.
chamber, however, the interaction between the reactive jets
may affect the effectiveness, which should be taken into
consideration. Some studies have discussed it and provided
fundamental insights. Maxson et al. [18] and Hensinger et al.
[19] both investigated the performance of pulsed jet combus-
Methodology
tion (PJC) and jet plume injection and combustion (JPIC)
system, which composed of single or triple orifices, and single Experiment Apparatus and
or double generators. It was found that with multiple jet
plumes, the flames expand to entire space faster and pressure
Conditions
traces are steeper, benefited from longer residence time for The experiments were conducted on a full-visible constant
reacting species in large scale vortex structures. Biswas et al. volume combustion chamber (CVCC) and the system diagram
[20] compared the ignition patterns of a single jet, triple is shown in Figure 1(a). The assembly of CVCC is shown in
straight and angled jets under different spark positions and Figure 1(b). It has a main body of regular octagon shape with
equivalence ratios. It shows that ignition probability can 80 mm in thickness, and the actual main chamber is cylin-
be improved by multiple jets, and ignition of individual jet is drical with 100 mm in diameter and 24.5 mm in length. Two
influenced by its interaction with other jets, due to various quartz glasses are fitted to provide optical access. Four heating
initial flame shapes in the pre-chamber. Chinnathambi et al. plates (4×400 W) are symmetrically mounted on the side
[21, 22] assessed the performance of three orifice configura- walls. Two K-type thermocouples are inserted into main
tions: a single jet, dual diverging and converging jets for chamber and on the wall respectively, connected with a
passive or active jet igniters. It was found that the diverging controller and solid state relays to reach the needed
design is overall superior, while converging design results in initial temperature.
an ignition location away from the wall, and they noted that The structures of pre-chambers designed in this work are
interference between jets should be avoided. Sun et al. [23] shown in Figure 1(c). It is un-scavenged and built-in, i.e.
used a single- and double-deck pre-chamber designs from the installed in the bottom of main chamber, stuck tightly by the
perspective of promoting the active radicals gathering, and it glasses, and fixed with the spark plug. The sides are visible to
shows that the lean limit can be extended and the ignition observe inner flame and flow. There are two types of pre-
zone is more stable under double-deck structures. Zhou et al. chamber nozzle designs in this study, Type 1# with crossing
[24] proposed an unequal three-hole aperture structure for angles and Type 2# with unequal nozzle diameters.
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL 3

FIGURE 1 Schematics of (a) experiment system; (b) full- TABLE 1 Pre-chamber geometries and conditions.
visible constant volume combustion chamber (CVCC); (c) Parameter (Unit) Value or Range
pre-chamber structures.
Geometrical parameters
Nozzle axis distance, L (mm) 5.5
Nozzle diameter, D (mm) 1, 2 (base), 4
Nozzle inclined angle, θ (°) 0 (parallel), 15, 20, 30, 45
Volume ratio, VPC / VMC (%) ~ 3.5
Mixture and initial conditions
Fuel Experiment 100 % CH4
Simulation 100 % CH4,
50 % NH3 + 50
% CH4
Oxidizer 79% N2 + 21% O2 (Synthetic air)
Equivalence ratio, ϕ 1.0 (base), 0.8
Initial pressure, P (MPa) 0.5 (base), 0.3
Initial temperature, T (K) 300 (base), 373

LeCroy 3054z) to collect the transient combustion pressure

signal. The fuel and air are charged and premixed in a mixing
tank using Dalton’s law, then the well-mixed gas is charged
in the chamber. For elevated temperature conditions, there is
a time interval to ensure the temperature uniformity. The
signals of spark ignition, two cameras, and pressure collection
are triggered synchronously by a digital delay signal generator
(SRS DG535), and the images are transmitted and saved by
PFV software. Once the combustion is completed, the exhaust
is immediately exited and then evacuated by a vacuum pump.
The values or ranges of pre-chamber geometrical param-
eters, mixture and initial conditions used in this work are
presented in Table 1. For pre-chamber design, the basic nozzle
diameter is selected as 2 mm, considering that the wall
quenching distance of CH4 under conditions in Table 1. The
flame will almost quench when passing the jet hole as “jet
auto-ignition” mode and the following results also confirm
it. To study the effect of jet crossing, four nozzle inclined
angles with respect to axis are used, where the nozzle diameter
Simultaneous high speed schlieren and OH* chemilumi- is 2 mm, as illustrated in “Type 1#” in Figure 1(c). To study
nescence imaging are used in this study, to capture the evolu- the effect of interaction between reacting jet and reacted jet,
tion of hot gas jet and determine the ignition and flame zone. two unequal nozzle diameter combinations are used, i.e. the
For schlieren imaging, the “Z”-type light path is applied, two holes have different diameters: D1 = 4 mm, D2 = 2 mm
consisting of a green LED lamp (5 W) with an aperture, two (D1/ D2 = 2); D1 = 4 mm, D2 = 1 mm (D1/ D2 = 4), as illustrated
flat and parabolic mirrors, a knife, and recorded by high speed in “Type 2#” in Figure 1(c). The nozzle axis distance is kept
camera 1# (Photron, FASTCAM SA-X2) with a frame rate of unchanged. Pure methane is used as fuel in experiment. The
10000 fps and a resolution of 1024×1024 pixels. For OH* basic condition is (P = 5 bar, T = 300 K, ϕ = 1), if not specially
chemiluminescence imaging, high speed camera 2# (Photron, noted in the following discussion, for the reason that the focus
FASTCAM SA-Z) attached with an image intensifier (LaVision, of this study is to explore the fundamentals of re-ignition
HSIRO) is oriented to the same view of schlieren but slightly behavior, while higher initial condition is not necessary. The
inclined placed, with a UV filter (308±10 nm) in front of lens change of conditions is used for the validation of repeatability.
to detected OH* signal. A setting of 20000 fps, 1024×1024 In addition, NH3/CH4/air mixture is also used in simulations,
pixels, 60 % gain, 35 μs gate width is used. The total image while the condition is kept all the same.
recording time is 50 ms for schlieren and 40 ms for
OH*, respectively.
A spark plug (IGK) with ignition coil (IGK) and a 12V
Image Processing
DC power is used to ignite the mixture in the pre-chamber. The images from experiments are processed by a series of
Meanwhile, a pressure sensor (Kistler 6115C) is mounted on programs, to improve the quality and extract the features of
the top wall and cooled by circulating water, connected with jet flame. For schlieren images, un-sharp filtering operation
a charge amplifier (Kistler 5011) and an oscilloscope (Teledyne is used to enhance the contrast. For OH* chemiluminescence
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

4 UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL

FIGURE 2 Image processing procedures for FIGURE 3 Computational domain and meshing.
OH* chemiluminescence.

TABLE 2 Numerical model selection.

Model Type
Combustion model G-Equation coupled with MZ-SAGE
Reaction mechanism CH4 GRI-Mech 3.0 [27]
NH3/CH4 Okafor et al. [28]
Flame speed Laminar Tabulated
Turbulent Peters’ correlation
[29]
Turbulence model RNG k-ε
Spark model Spherical flame kernel with two stage

and the reduced reaction mechanism from Okafor et al. [28]

images, the procedures are demonstrated in Figure 2. Firstly, is used for NH3/CH4, which performs well in predicting the
the effective window is cut out and corrected to regular laminar flame speed under relative high pressures. A table of
circular shape. Then, on the one hand, to show the burned SL is generated by Converge 1-D tool as model inputs. Local
zone and its luminous intensity more clearly, gray image is turbulent flame speed is modeled by Peters’ correlation [29]
converted to pseudo-color image and enhanced. On the other for predicting the flame propagation, while MZ-SAGE model
hand, to compare the flame development represented by OH* is only for computing the intermediate species and reaction
distribution, the image is binarized by threshold treatment, heat production. For turbulence modeling, RNG k-ε model is
then the flame edge is recognized and the total area is calcu- used for compressible flows. For spark modeling, the process
lated at different times. is simplified to two stage (breakdown and glow discharge) in
the form of energy source, with a diameter of 1 mm spherical
CFD Simulation Setup flame kernel around the electrode gap. The detailed model
descriptions and governing equations can be found in Ref.
To further reveal the mechanisms of jet interaction during [30] and not be presented here. The density-based, transient
the ignition and combustion process, CFD simulations were solver for flow and CVODE solver for chemistry are utilized,
conducted based on Converge software. The geometry is the and the computations are conducted on a server with
prototype of the CVCC above, and the spark plug is omitted maximum 96 Intel CPU cores.
for simplification. The computational domain and meshing The model validation results of pressures for CH4 are
at one intermediate time is shown in Figure 3. A basic mesh presented in Figure 4. It is found that the IDT predicted by
size of 4×4×4 mm is used. For mesh refinement, the pre- CFD simulation is in good agreement with the experiment.
chamber and one half of jet ejection in main chamber are However, there is a deviation of the burning rate and the
embedded, where 3 scale of adaptive mesh refinement (AMR) maximum pressure, mainly originated from the uncertainty
is used all along, and a radius of 5 mm spherical zone around in the reaction mechanism, inaccurate calculation of turbulent
the spark kernel is refined by 5 scale (i.e. 0.125 mm) at initial flame speeds, and unavoidable gas leak in our experiments.
laminar flame stage. Similar conclusions are also drawn in previous study [31],
The numerical models selected are presented in Table 2. mentioning that the relative error is within acceptable range.
For combustion modeling, G-Equation coupled with Multi On the whole, the numerical model has a good ability to
Zone (MZ)-SAGE model is applied, which is a promising predict the ignition process, which is main focus in this work,
approach to resolve the turbulence-chemistry interaction for although it cannot calculate well in the later burning rates.
multiple turbulent flame regimes in Borghi-Peters diagram Therefore, the simulation results can be used for the
[25, 26]. The GRI-Mech 3.0 [27] mechanism is used for CH4, following analysis.
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL 5

FIGURE 4 Comparison of the pressures between FIGURE 5 Images of hot gas jet shown by schlieren and
experiment and CFD simulation results (θ = 30°, CH4/air, ϕ = 1, flame shown by OH* chemiluminescence at ignition time, for
P = 5 bar, T = 373 K). different nozzle inclined angles.

Results and Discussion

Effect of Jet-Crossing Nozzle
Structure
In the observations of our experiments, the entire combustion
process can be divided into four phases under jet auto-
ignition mode:
1. Phase 1: the mixture in the pre-chamber (PC) is spark
ignited and a laminar flame is established;
meanwhile, the unburnt cold jet is ejected to the
main-chamber (MC).
2. Phase 2: the burnt hot jet is ejected in a high speed
and quenched to some extent, entraining the
surrounding unburnt mixture and moving closed to
or against the opposite wall.
3. Phase 3: after an ignition delay time (IDT), the second
ignition occurs in MC and the wrinkled flame fast
sweeps across the entire space; as the flame develops
in MC, the pressure in MC will exceed that in PC,
resulting in back-flow of burnt gas, until the end of ignition starts away from the jet-crossing zone, because the
combustion (EoC) in MC. two jets merge into a single jet too early. Meanwhile, the
4. Phase 4: the exotherm is completed and pressure ignition of parallel structure only occurs on the one side,
decreases after the EoC in MC. showing more dispersed and asymmetric distribution.
The flame (represented by OH*) area development of jet-
The findings above are in accordance with previous crossing, parallel and conventional spark plug (CSP) struc-
studies [10, 13]. Obviously, phase 2 and 3 are the main concerns tures at 0-40 ms is shown in Figure 6, where the flame area is
in this study. The hot gas jet and flame at initial ignition in normalized by the total area of main chamber. It can be found
MC can be shown by schlieren image and OH* chemilumi- that the IDT is shortest for θ = 45° case, followed by θ = 15°,
nescence image, respectively. As presented in Figure 5, the then similar for θ =30° and parallel case, and longest for θ
ignition zones at ignition time (start of phase 3) under different =20° case. Besides, the slope of the curves i.e. flame expanding
jet directions are marked, and the color-bar of normalized velocity does not deviate much for all pre-chamber cases. For
OH* luminous intensity is in the last row. It can be seen that CSP case, the flame is basically laminar and propagates about
for θ = 15°, 20°, and 30° cases, ignition all occurs around the 10 times slower than the pre-chamber cases.
jet-crossing zone. The flames exhibit a triangle shape initially, Figure 7 shows the pressure in main chamber and its
expanding to the upstream in the jet hole direction and to the rising rate (dP/dt) traces obtained from experiments, where
overlapping zone downstream together. For θ = 45° case, the pressure trace is filtered and averaged by repeating three
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

6 UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL

FIGURE 6 Flame area development for different nozzle FIGURE 8 Combustion duration for different nozzle
inclined angles and CSP (dashed line means the end of inclined angles and CSP from experiments.
combustion in pre-chamber).

combustion heat release. Besides, it can be also found that

about 80 % reduction of combustion duration is reached
FIGURE 7 Pressure and its rising rate for different nozzle
compared with CSP.
inclined angles and CSP from experiments.
The reasons for different ignition and combustion char-
acteristics under the jet-crossing and no-cross (parallel) struc-
tures can be further explained from simulation results. Here,
θ = 30° and parallel case are taken as examples. Figure 9 shows
the contours of temperature and Damköhler number at
ignition time and the flame front represented by G = 0 at one
subsequent time, where T* denotes the normalized time by
the times at maximum pressure, considering that the absolute
time is not the focus in the simulation. The Da number is a
critical dimensionless number for determining the turbulent
flame regime and judging the ignition success or failure [20],
and its related variables are calculated as written in Eq. (1) -
Eq. (4):

 m l T / uT 
Da   (1)
 c  L / SL

times under same conditions. Four combustion phases are k 3 /2

lT  C  3 / 4 (2)
marked in the dP/dt trace, defined by the description above. 
It can be seen that the tendency is almost same as the flame
area development in Figure 6, among which θ = 45° case
performs most differently from the others, with lower heat 2
uT   k (3)
release rate. The combustion duration is defined as the period 3
from spark timing to 10 % and 90 % of maximum MC pressure
since there is no volume variation, corresponding to MFB
(Mass Fraction Burned) 0-10 and MFB 10-90, and the results  / c p
L  (4)
are presented in Figure 8. The relative tendency of MFB 0-10 SL
is the same as that of IDT. MFB 10-90 is similar for θ = 15°,
20° and 30°, but longer for θ = 45°, mainly due to its advanced where τm, τc denotes mixing flow and chemical reaction
ignition and smaller jet coverage area. Overall, the cross angles time scale, respectively. lT is turbulence integral length scale,
in the range of 20°-30° shows similar performances and more uT’ is turbulence fluctuating velocity, which are computed
favorable for ignition zone control, while the cross region too from the turbulence model. δL is laminar flame thickness and
closed to the nozzles will reduce the effectiveness in initial tabulated in prior. It can be found from Figure 9 that a rela-
ignition but may be beneficial for the suppression of violent tively high temperature zone exists at the jet tip, where ignition
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL 7

FIGURE 9 Contours of temperature, Da, and flame front represented by G = 0 at typical times, for “θ = 30°” and “parallel”
structures, with 100 % CH4 as fuel, under ϕ = 1, P = 5 bar, T = 300 K condition.

occurs. Nevertheless, it can be easily broken by the effect of be seen that the effect of NH3 addition is similar to the role of
hydrodynamic quenching and be re-constructed by supply of EGR or air dilution [34].
hot products from pre-chamber continuously. Local ignition
and extinction will happen during the IDT, until global
ignition occurs and a self-sustaining flame is established. The Repeatability and Stability of
high temperature zone for jet-crossing structure is larger and Ignition Zone
more stable than that for parallel structure, and easier to
maintain with concentrated heat. On the other hand, the jet To further study the repeatability and stability of ignition zone
impinging can cause the turbulence dissipation and stagna- for different jet hole directions, the initial conditions are
tion somewhat, promoting the transport of active species. changed for θ = 30° and parallel case, as shown in Figure 11.
Therefore, the accumulation of heat and transport of active It is found that as the IDT is shortened with a higher initial
species under jet overlap or impinging condition is beneficial temperature from 300 K to 373 K, the ignition zone is still
for more stable ignition. While for parallel structure, more located in the jet-crossing zone with evenly distributed flame
intense turbulence shear at jet lateral sides with higher flame
strain rate and heat loss could lead to lower ignition proba-
bility. After the ignition, it shows that the flame front for jet- FIGURE 10 Turbulent flame regimes during ignition and
crossing case is more wrinkled owing to smaller turbulence combustion process in Borgi-Peters diagram, for “parallel”
length scale, and it develops slower in jet radial direction than structure, with 100 % CH4 and 50 % NH3 + 50 % CH4 as fuels,
the parallel case. In addition, at re-ignition time in main under ϕ = 1, P = 5 bar, T = 300 K condition.
chamber, the range of local Da number around ignition zone
is in 0.2-0.4, which is in agreement with some previous numer-
ical studies [32, 33], and the values show not much difference
for the two cases.
Besides, the turbulent flame regimes in the main chamber
ignition and combustion process are plotted in classical Borgi-
Peters diagram [29], for pure CH4 and NH3 blends cases, as
shown in Figure 10. It can be found that re-ignition starts at
thickened flame zone for both two cases. With the addition
of NH3, the flame regime firstly moves to broken reaction zone
and re-ignition starts later than 100% CH4 case, indicating
that during early jet evolution, turbulent mixing and stretch
are very intense owing to lower SL and higher δL . Re-ignition
only occurs until turbulence level reduces to some extent,
while the criteria is beyond the scope of this work. It can also
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

8 UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL

FIGURE 11 Images of hot gas jet shown by schlieren and the red arrow, while the radical jet also evolves into an inde-
flame shown by OH* chemiluminescence at ignition time, for pendent flame, as marked by the red circle. For the “D1 = 4
different conditions under “θ = 30°” and “parallel” structures. mm, D2 = 1 mm” case, the flame jet penetration length is
higher at ignition time, because of the reduction in total nozzle
cross-section area. Meanwhile, the jet from 1 mm nozzle is
too late to form an independent flame. As the ignition is
retarded with a lower equivalence ratio, the flame jet from
4 mm nozzle is also significantly quenched, failing to start an
immediate ignition zone after its impingement on the wall.
Therefore, both are jet auto-ignition modes under
this condition.
The phenomena above can be explained from the view of
jet dynamics. An ignition criterion from the phenomeno-
logical model of jet ignition proposed by Malé et al. [10] can
be used. The maximum mixing rate α0 is defined as follows:

1 U
0  K  K inj (5)
F D

where τF is jet flow time scale, Uinj is jet injection velocity,

and K is a constant coefficient. Ignition is only possibly
successful when α0 is lower than its critical value αc. To
compare the ignition probability of two jets and analyze the
mechanisms, the simulation results for contours of tempera-
ture with streamlines and velocity magnitude at typical inter-
mediate times are presented in Figure 13. In general, it can
be deduced that the maximum velocity at centerline of two
unequal jets is almost the same for incompressible, homog-
enous and subsonic flows. However, for the flame jets with
high density gradient, the mean velocity for the larger jet hole
is higher according to the velocity distribution shown in
Figure 13, which is different from the conditions of single jet
approximately. While for parallel case, as ignition is delayed cases [35]. Therefore, the ignition probability of the smaller
with a lower initial pressure from 5 bar to 3 bar, the initial diameter nozzle is increased by the reduction of velocity
flame becomes symmetrical distributed. It is also noticed that magnitude. For the simulation results, maximum Uinj is 579
there are larger variations in IDT and ignition locations. In m/s for nozzle diameter ratio D1/D2 = 2 and 726 m/s for D1/D2
summary, the ignition is more repeatable and stable with the = 4, and the ratio of Uinj for two holes is 1.2 for D1/D2 = 2 and
jet-crossing structure. 1.5 for D1/D2 = 4. It can be calculated that α0 of larger nozzle
is still lower than that of small nozzle. Therefore, ignition is
still more probable for the larger nozzle and it is in agreement
Effect of Unequal Nozzle with the experiment results. For the D1/D2 = 2 case, the velocity
Diameter Structure difference of two jets is relatively small, leaving sufficient time
for the development of radical jet, and finally ignited by the
To study another interaction between the flame jet and radical flame jet. For the D1/D2 = 4 case, it can be found from Figure
jet, two groups of unequal nozzle diameter with the ratio of 12 that larger scale vortexes are generated in the 4 mm flame
2 and 4 are applied, respectively. Figure 12 shows the schlieren jet, which leaves insufficient spaces for the 1 mm radical jet
and OH* images at initial ignition times under two different to entrain unburnt combustible mixture. Meanwhile, the
equivalence ratios, ϕ = 1 and ϕ = 0.8. A flame jet from 4 mm higher Uinj for 4 mm flame jet leads to insufficient residence
nozzle shows significantly shorter IDT than the 2 mm nozzle, time for chemical chain reaction initiation. These two aspects
due to little wall quenching. For the “D1 = 4 mm, D2 = 2 mm” lead to ignition failure of the 1 mm radical jet. In addition,
case, the images reveal that there is a few amount of OH* in the IDTs and turbulent flame speeds are changed at different
the 2 mm radical jet throughout the IDT, and it develops into equivalence ratios, which affects the temporal sequence
a flame with the support of the flame jet, thus the flame expan- between the evolution of flame jet and radical jet. Therefore,
sion in horizontal direction (to the jet radical side) is improved, it leads to different ignition zones and timings in Figure 12
as marked by the red circle. This is different from the equal under different equivalence ratios.
2mm nozzles, “parallel” case in Figure 5, and IDT of 2 mm The two types of jet interaction could play an important
radical jet is reduced by about 60 %. Meanwhile, as the equiva- role in the pre-chamber designs. The former is superior in the
lence ratio gets leaner, the ignition induced by the flame jet is control of jet auto-ignition zone for more stable ignition under
retarded and moves towards the jet radical side, as marked by jet auto-ignition mode, corresponding to low load, cold start
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL 9

FIGURE 12 Images of hot gas jet shown by schlieren and flame shown by OH* chemiluminescence at two typical times, for
different nozzle diameter combinations under different equivalence ratios.

FIGURE 13 Contours of temperature with streamlines and condition. The latter is beneficial to improve the uniformity
velocity magnitude at typical times, for different nozzle of flame propagation in various directions, and better match
diameter combinations. the combustion chamber wall shapes.

Summary/Conclusions
In this work, optical experiments and CFD simulations were
conducted to investigate the fundamentals of interaction
between two reactive jets and its effects on ignition and
combustion characteristics. The main conclusions are listed
as follows:
1. For the effect of jet-crossing structures, ignition
occurs around the jet-crossing region at θ = 15°, 20°,
30° and above it at θ = 45°, which is contributed by the
accumulation of heat and transport of active species.
The cross angles in the range of 20°-30° with respect
to the pre-chamber axis show similar performances
and more favorable for ignition control overall, while
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

10 UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL

the cross region too closed to the nozzles reduce the International Journal of Hydrogen Energy, doi:10.1016/j.
overall burning rate. ijhydene.2022.04.283.
2. For parallel nozzle structure, the distribution of 7. Leng, X., Deng, Y., He, D., Wei, S. et al., “A Preliminary
ignition zone is more dispersed and asymmetric. The Numerical Study on the Use of Methanol as a Mono-Fuel for
ignition becomes less repeatable and stable than that a Large Bore Marine Engine,” Fuel 310 (2022): 122309,
in the jet-crossing structure case. doi:10.1016/j.fuel.2021.122309.
3. Local Da number around the ignition zone is in the 8. Vacca, A., Rossi, E., Cupo, F., Chiodi, M. et al., “Virtual
range of 0.2-0.4, and with the addition of NH3, the Development of a Single-Cylinder Engine for High Efficiency
flame regime is located in thickened and broken by the Adoption of Efuels, Methanol, Pre-Chamber and
reaction zone during early jet evolution. Millerization,” SAE Technical Paper 2022-37-0018, 2022,
4. For the effect of unequal nozzle diameter structures, 10.4271/2022-37-0018.
under the two nozzle diameter of 4mm and 2mm, the 9. Attard, W.P., Fraser, N., Parsons, P., and Toulson, E., “A
ignition delay time of the radical jet will be reduced Turbulent Jet Ignition Pre-Chamber Combustion System for
by 60 % than the equal nozzle design with the Large Fuel Economy Improvements in a Modern Vehicle
diameter of 2mm, because of the suppression of Powertrain,” SAE Technical Paper 2010-01-1457, 2010,
quenching and increased ignition probability with the 10.4271/2010-01-1457.
reduction of velocity magnitude in presence of the 10. Malé, Q., Vermorel, O., Ravet, F., and Poinsot, T., “Direct
flame jet. And the flame expansion in horizontal Numerical Simulations and Models for Hot Burnt Gases Jet
direction is also improved. While for the two nozzles Ignition,” Combustion and Flame 223 (2020): 407-422,
with the diameter of 4mm and 1mm, the 1mm radical doi:10.1016/j.combustflame.2020.09.017.
jet is not ignited due to larger velocity difference and
11. Ghorbani, A., Steinhilber, G., Markus, D., and Maas, U.,
insufficient residence time. In addition, the ignition
“Ignition by Transient Hot Turbulent Jets: An Investigation
zones and timings are also influenced by different
of Ignition Mechanisms by Means of a PDF/REDIM
equivalence ratios due to the changes of temporal
Method,” Proceedings of the Combustion Institute 35, no. 2
sequence of two jet evolution.
(2015): 2191-2198, doi:10.1016/j.proci.2014.06.104.
Future work will be conducted on developing novel pre- 12. Gussak, L.A., Karpov, V.P., and Tikhonov, Y.V., “The
chamber geometries with the concepts in this study and vali- Application of Lag-Process in Prechamber Engines,” SAE
dating their performances in optical or metal engines. Technical Paper 790692, 1979, 10.4271/790692.
13. Yamaguchi, S., Ohiwa, N., and Hasegawa, T., “Ignition and
Burning Process in a Divided Chamber Bomb,” Combustion
and Flame 59 (1985): 177-187, doi:10.1016/0010-
References 2180(85)90023-9.
14. Biswas, S., Tanvir, S., Wang, H., and Qiao, L., “On Ignition
1. International Energy Agency, World Energy Outlook, 2021, Mechanisms of Premixed CH4/Air and H2/Air Using a Hot
https://www.iea.org/reports/world-energy-outlook-2021. Turbulent Jet Generated by Pre-chamber Combustion,”
2. Toulson, E., Schock, H.J., Attard, W.P., “A Review of Pre- Applied Thermal Engineering 106 (2016): 925-937,
Chamber Initiated Jet Ignition Combustion Systems,” SAE doi:10.1016/j.applthermaleng.2016.06.070.
Technical Paper 2010-01-2263, 2010, 10.4271/2010-01-2263. 15. Sidey, J.A.M. and Mastorakos, E., “Pre-Chamber Ignition
3. Stadler, A., Sauerland, H., Härtl, M., and Wachtmeister, G., Mechanism: Simulations of Transient Autoignition in a
“The Potential of Gasoline Fueled Pre Chamber Ignition Mixing Layer between Reactants and Partially-Burnt
Combined with Elevated Compression Ratio,” SAE Technical Products,” Flow, Turbulence and Combustion 101, no. 4
Paper 2020-01-0279, 2020, 10.4271/2020-01-0279. (2018): 1093-1102, doi:10.1007/s10494-018-9960-0.
4. Atis, C. and Schock, H., “Comparison of Excess Air (Lean) 16. Chi, C., Abdelsamie, A., and Thévenin, D., “Transient
vs EGR Diluted Operation in a Pre-Chamber Air/Fuel Ignition of Premixed Methane/Air Mixtures by a Pre-
Scavenged Dual M ode, Turbulent Jet Ignition Engine at chamber Hot Jet: A DNS Study,” Flow, Turbulence and
High Dilution Rate (~40%),” SAE Technical Paper 2021-01- Combustion 108 (2022): 775-795, doi:10.1007/s10494-021-
0455, 2021, 10.4271/2021-01-0455. 00292-9.
5. Liu, Z., Zhou, L., and Wei, H., “Experimental Investigation 17. Biswas, S., “Dimensionless Quantities in Pre-chamber
on the Performance of Pure Ammonia Engine Based on Turbulent Jet Ignition of Premixed Methane/Air,” Alternative
Reactivity Controlled Turbulent Jet Ignition,” Fuel 335 Fuels and Advanced Combustion Techniques as Sustainable
(2023): 127116, doi:10.1016/j.fuel.2022.127116. Solutions for Internal Combustion Engines (Singapore:
Springer Singapore, 2021), 257-279, doi:10.1007/978-981-16-
6. Wu, X., Feng, Y., Xu, G., Zhu, Y. et al., “Numerical
1513-9_11.
Investigations on Charge Motion and Combustion of
Natural Gas-Enhanced Ammonia in Marine Pre-Chamber 18. Maxson, J.A., Hensinger, D.M., Hom, K., and Oppenheim,
Lean-Burn Engine with Dual-Fuel Combustion System,” A.K., “Performance of Multiple Stream Pulsed Jet
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL 11

Combustion Systems,” SAE Technical Paper 910565, 1991, Chamber,” International Journal of Hydrogen Energy 43, no.
10.4271/910565. 45 (2018): 21102-21113, doi:10.1016/j.ijhydene.2018.09.156.
19. Hensinger, D.M., Maxson, J.A., Hom, K., and Oppenheim, 33. Xu, G., Kotzagianni, M., Kyrtatos, P., Wright, Y.M. et al.,
A.K., “Jet Plume Injection and Combustion,” SAE Technical “Experimental and Numerical Investigations of the
Papers 920414, 1992, 10.4271/920414. Unscavenged Prechamber Combustion in a Rapid
20. Biswas, S. and Qiao, L., “Ignition of Ultra-Lean Premixed Compression and Expansion Machine under Engine-Like
H2/air using Multiple Hot turbulent Jets Generated by Pre- Conditions,” Combustion and Flame 204 (2019): 68-84,
chamber Combustion,” Applied Thermal Engineering 132 doi:10.1016/j.combustflame.2019.01.025.
(2018): 102-114, doi:10.1016/j.applthermaleng.2017.11.073. 34. Benajes, J., Novella, R., Gomez-Soriano, J., Barbery, I. et al.,
21. Chinnathambi, P., Thelen, B., Naylor, M., Cook, D. et al., “Computational Assessment towards Understanding the
“Performance Assessment of a Single Jet, Dual Diverging Energy Conversion and Combustion Process of Lean
Jets, and Dual Converging Jets in an Auxiliary Fueled Mixtures in Passive Pre-Chamber Ignited Engines,” Applied
Turbulent Jet Ignition System,” SAE Technical Paper 2018- Thermal Engineering 178 (2020): 115501, doi:10.1016/j.
01-1135, 2018, 10.4271/2018-01-1135. applthermaleng.2020.115501.
22. Chinnathambi, P., Thelen, B., Cook, D., and Toulson, E., 35. Gentz, G., Thelen, B., Toulson, E., Litke, P. et al.,
“Performance Metrics for Fueled and Unfueled Turbulent Jet “Combustion Visualization, Performance, and CFD
Igniters in a Rapid Compression Machine,” Applied Thermal Modeling of a Pre-Chamber Turbulent Jet Ignition System in
Engineering 182 (2021): 115893, doi:10.1016/j. a Rapid Compression Machine,” SAE Int. J. Engines 8, no. 2
applthermaleng.2020.115893. (2015): 538-546, doi:10.4271/2015-01-0779.
23. Sun, S., Ma, Y., Cui, L., Ma, X. et al., “Effect of Single and
Double-Deck Pre-Chamber Designs to the Combustion
Characteristics of Premixed CH4/Air,” SAE Technical Paper Contact Information
2018-01-1688, 2018, 10.4271/2018-01-1688.
First Author
24. Zhou, L., Song, Y., Hua, J., Liu, F. et al., “Effects of Different
Hole Structures of Pre-Chamber with Turbulent Jet Ignition Yixiao Zhang, Master of Science Student
on the Flame Propagation and Lean Combustion Address: School of Vehicle and Mobility, Tsinghua University,
Performance of a Single-Cylinder Engine,” Fuel 308 (2022): Beijing, 100084, China
121902, doi:10.1016/j.fuel.2021.121902. [email protected]
25. Silva, M., Liu, X., Hlaing, P., Sanal, S. et al., “Computational
Corresponding Author
Assessment of Effects of Throat Diameter on Combustion
and Turbulence Characteristics in a Pre-Chamber Engine,” Xiao Ma, Ph.D., Associate Professor
Applied Thermal Engineering 212 (2022): 118595, Address: School of Vehicle and Mobility, Tsinghua University,
doi:10.1016/j.applthermaleng.2022.118595. Beijing, 100084, China
26. Xu, L., Li, G., Yao, M., Zheng, Z. et al., “Numerical [email protected]
Investigation on the Jet Characteristics and Combustion
Process of an Active Prechamber Combustion System Fueled
with Natural Gas,” Energies 15 (2022): 5356, doi:10.3390/
en15155356.
Acknowledgments
27. Smith, G.P., Golden, D.M., Frenklach, M., et al., “GRI-Mech The authors acknowledge the financial support from the State
3.0,” http://combustion.berkeley.edu/gri-mech/releases.html. Key Laboratory of Automotive Safety and Energy (Grant
ZZ2021-034) and the National Natural Science Foundation
28. Okafor, E.C., Naito, Y., Colson, S., Ichikawa, A. et al.,
“Measurement and Modelling of the Laminar Burning
of China (Grant No. 51976100, Grant No. 52211530098), and
Velocity of Methane-Ammonia-Air Flames at High the software support from Convergent Science Inc.
Pressures Using a Reduced Reaction Mechanism,”
Combustion and Flame 204 (2019): 162-175, doi:10.1016/j.
combustflame.2019.03.008. Definitions/Abbreviations
29. Peters, N., Turbulent Combustion (United Kingdom,
CFD - Computational Fluid Dynamics
Cambridge: Cambridge University Press, 2000)
CSP - Conventional Spark Plug
30. Convergent Science Inc, “CONVERGE v3.0 Manual.”
31. Gholamisheeri, M., Wichman, I.S., and Toulson, E., “A Study
CVCC - Constant Volume Combustion Chamber
of the Turbulent Jet Flow Field in a Methane Fueled EoC - End of Combustion
Turbulent Jet Ignition (TJI) System,” Combustion and Flame IDT - Ignition delay time
183 (2017): 194-206, doi:10.1016/j.combustflame.2017.05.008.
MC - Main Chamber
32. Wang, N., Liu, J., Chang, W.L., and Lee, C.F., “A Numerical
Study of the Combustion and Jet Characteristics of a
MFB - Mass fraction burned
Hydrogen Fueled Turbulent Hot-Jet Ignition (THJI) PC - Pre-Chamber
 Downloaded from SAE International by Yixiao Zhang, Friday, April 07, 2023

12 UNDERSTANDING INTERACTION BETWEEN REACTIVE JETS IN PRE-CHAMBER IGNITION OF GASEOUS FUEL

Nomenclature uT’ - Turbulence fluctuating velocity (m·s-1)

Uinj - Jet injection velocity (m·s-1)
D - Nozzle diameter (mm) α0 - Maximum mixing rate (s-1)
Da - Damköhler number δL - Laminar flame thickness (m)
k - Turbulence kinetic energy (m2·s-2) ε - Turbulence dissipation rate (m2·s-3)
lT - Turbulence integral length scale (m) ϕ - Equivalence ratio
P - Pressure (MPa) θ - Nozzle inclined angle (°)
SL - Laminar flame speed (m·s-1) τc - Chemical reaction time scale (s)
T - Temperature (K) τF - Jet flow time scale (s)
T* - Normalized time τm - Mixing flow time scale (s)

© 2023 SAE International. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE International.

Positions and opinions advanced in this work are those of the author(s) and not necessarily those of SAE International. Responsibility for the content of the work lies
solely with the author(s).

ISSN 0148-7191

View publication stats