Content uploaded by Rajavasanth Rajasegar
Author content
All content in this area was uploaded by Rajavasanth Rajasegar on Jun 16, 2017
Content may be subject to copyright.
American Institute of Aeronautics and Astronautics
1
Progress Towards Simulations of Plasma-Assisted
Combustion in a Swirled Flow Reactor
Joseph W. Zimmerman1,a, Rajavasanth Rajasegar 2,b, Constandinos M. Mitsingas 3,b, Andrew D. Palla 4,a,
Darren M. King 5,a, David L. Carroll 6,a, and Tonghun Lee 7,b
aCU Aerospace, Champaign, IL, 61820
bUniversity of Illinois, Urbana, IL, 61801
The capability of extending lean operational limits of methane-air flames has been
demonstrated using the combination of annular swirled flow and plasma power deposition.
The inclusion of plasma power deposition to the flame region initiates rapid decomposition
of flow constituents driven by electron-impact processes as well as interactions with
electronically-excited neutral species produced in the plasma. The objective of the current
work is to leverage progressing experimental efforts to provide validation for simulations of
plasma-assisted combustion (PAC). The multiphysics simulation capability brings together
modules for various important aspects: fluid dynamics, turbulence modeling, electric field
coupling, transport models for neutrals and charged species, as well as a detailed reaction
mechanism for air-plasma and combustion chemistry. The corresponding experimental
work provides various validation data from plasma-assisted combustion flames in air:CH4
mixtures including OH planar laser-induced fluorescence (PLIF) measurements, Raleigh
scattering thermometry (RST), particle image velocimetry (PIV) of the reactor flow field,
and spectroscopy of nitrogen emissions in the flame. Recent work has identified swirl-
stabilized and PAC cases to be used for validation of simulation work. The present paper
describes the experimental approach, reviews key experimental results, discusses
development of neutral combustion simulations based on the PAC experimental geometry,
and overviews next steps in development in PAC simulations.
Nomenclature
EEDF = electron energy distribution function
E/N = electric field to gas density ratio (reduced electric field)
ne = electron density
SLPM = standard liters per minute
Te = electron temperature
Tg = gas temperature
u = flow velocity
t
= turbulent kinematic viscosity
p
= primary fuel-to-oxidizer equivalence ratio
1 Senior Scientist, 301 N. Neil St., Suite 502, AIAA Member.
2 Research Assistant, Mechanical Science and Engineering, 3001 Mech. Eng. Lab
3 Research Assistant, Mechanical Science and Engineering, 3001 Mech. Eng. Lab
4 Senior Physicist, 301 N. Neil St., Suite 502, AIAA Member.
5 Senior Engineer, 301 N. Neil St., Suite 502
6 President, 301 N. Neil St., Suite 502, AIAA Fellow.
7 Assoc. Prof., Mechanical Science and Engineering, 3001 Mech. Eng. Lab
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
48th AIAA Plasmadynamics and Lasers Conference
5-9 June 2017, Denver, Colorado
AIAA 2017-3674
Copyright © 2017 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
AIAA AVIATION Forum
American Institute of Aeronautics and Astronautics
2
I. Introduction
lasma discharge techniques are being investigated to address problems in ignition and combustion [1, 2]. Areas
of research range from automotive engine technologies for improving fuel efficiency to aerospace propulsion
where researchers are developing technologies for stabilizing combustion in hypersonic engines, improving
performance envelopes and emissions characteristics of subsonic jet engines, and stabilizing flames in lean-fuel
conditions [2]. There is also interest in applying similar nonequilibrium plasma devices to fuel synthesis and
chemical production technologies that use carbon-dioxide or methane captured from industrial and other
anthropogenic sources [3, 4]. Various configurations have been investigated for coupling plasma power with flames:
longitudinal DC field coupling [5], RF plasma-ignition [6], nanosecond-pulsed dielectric barrier discharges (ns-
DBDs) [7], laser-induced breakdown with microwave power deposition [8], and waveguide coupled microwave
systems [9-11].
Recent experiments were devised to investigate microwave power coupling into methane-air mixtures jetting
from coaxial microwave cavities [9-11]. The plasma-assisted combustion flames generated in these microwave
torch-type configurations can be characterized with Rayleigh scattering for temperature determination, 2-D planar
laser-induced fluorescence (PLIF) for characterization of radical distributions, particle image velocimetry (PIV) to
characterize flow-fields and flame-fronts, as well as probe measurements and spectroscopic techniques to establish
electronic characteristics, excited species contents, and non-equilibrium behaviors. Rao et al. [7] investigated plasma
re-ignition of premixed CH4/air in a coaxial reentrant type cavity, and demonstrated the capability of increasing
microwave power deposition to extend flow rate blow off limits (for lean, stoichiometric, and rich conditions).
Additional work with this type of system [10,11] characterized the influence of power deposition on OH content and
temperature of PAC flames from a direct-coupled microwave applicator and characterized the plasma-associated
heating of stoichiometric premixed CH4/air discharge compared to that in an air discharge. Similar characterizations
of PAC devices have been made with pulsed DBDs [12] and with microwave coupling to flat flames within
waveguides [13]. Development of combustion technologies based on these research concepts would benefit from
robust multiphysics simulation techniques which capture the associated combustion chemistry, plasma excitation
mechanisms, and fluid dynamics.
The current paper summarizes the experimental apparatus being modeled for validation and overviews some of
the data generated using the swirl-flow PAC system. Recent progress on development of the reactor simulations are
discussed. The setup of the neutral combustion simulations in BLAZE MultiphysicsTM is described. Initial neutral
combustion simulations with a detailed reduced combustion mechanism are reported, and the next steps toward
adding plasma kinetics to simulate the PAC validation case are discussed.
.
II. Experimental Configuration and Key Results
A. Swirl-Stabilized PAC Reactor Experiment
A microwave waveguide plasma system (developed in collaboration with Amarante Technologies) was used to
directly couple microwave energy into flames with various geometries; the chamber and swirl injector is pictured in
Figure 1(a). The system allows for complete access of the plasma-enhanced flame for laser and optical diagnostics.
Microwave radiation can be generated in excess of 1 kW but also as low as 30 W. Microwave energy is generated
by a research-grade 2.45-GHz magnetron directly mounted to a WR284 waveguide and powered by an SM840
power supply. The microwave transmission inside the WR284 waveguide is in TE 10 mode. A directional coupler
mounted next to the magnetron head is used to measure the incident power. A circulator directs reflected energy into
a dummy load with a coupler, so that reflected energy can also be measured. Both the magnetron and dummy load
are water cooled. A three-stub tuner and sliding short are used to adjust the electric field and impedance to match the
load of the plasma applicator and the nozzle. A typical arrangement of the optical diagnostics is shown in Figure
1(b).
The coaxial nozzle has a solid center electrode that protrudes inside the WR284 waveguide. Microwave energy
travels through the nozzle, acting as a coaxial guide, in transverse electromagnetic mode. The internal nozzle
geometry is optimized for plasma breakdown by delivering a high electric field at the nozzle tip. A plasma
discharge is initiated by adjusting the maximum E-field position of the standing wave inside the waveguide using
the sliding short. As the standing wave is repositioned, the E-field at the nozzle tip also varies. When the E-field
reaches the breakdown threshold, gas around the electrode tip will be ionized, generating a plasma discharge that sits
entirely above the surface of the nozzle, producing a direct-coupled plasma-enhanced flame accessible for direct
measurements by optical diagnostics. The torch would likely be more efficient with a truncated inner electrode,
P
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
3
where the outer electrode (nozzle body) extended further than the inner solid electrode; however, a heavy interest in
complete access to the plasma ignition and coupling point directed the decision to use a “flush” configuration, where
the electrodes end at the same length.
(a) (b)
Figure 1. Swirl-stabilized PAC experimental setup: (a) Photos of quartz combustion chamber and microwave coax with
tangential swirl injector, PAC flame images at upper right show 40 SLPM swirl air added to 10:1 SLPM Air:CH4
through the main annular jet; (b) Laser diagnostics setup for PAC with microwave torch system.
The swirl chamber is constructed around the existing 2.45 GHz waveguide-coupled coaxial microwave
applicator, and consists of a 1.5”-diameter circular cylindrical swirl injector section which is 0.5” high, with four 2.5
mm diameter tangential injectors spaced at 90 intervals about the circumference of the outer wall. Above the
circular section, the combined annular jet and swirling flow enters a 1.5” x 1.5” rectangular cross-section fused-
silica tube with rounded corners measuring 12” long. An alternative aluminum rectangular tube of the same internal
dimensions was also created with 0.125” holes drilled at various positions, allowing access for a high-temperature
probe-type microphone in the chamber. These configurations mimic realistic combustion geometries and prevent
entrainment of outside air into the jet. In the present work, only air is injected through the tangential swirl inlet ports.
This dilutes the main jet boundary and thereby lowers the overall operating equivalence ratio of the burner. The
swirl number of the flow entering the combustor is varied by changing the proportion of the tangential to the axial
flow rates. The axial (fuel and air) and tangential (air) flow rates are metered individually by three laminar-flow,
differential pressure mass flow controllers (MKS mass controller). This configuration allowed for generation of
stable lifted flames at a height of 5 to 25 mm from the nozzle exit.
B. Review of Key Experimental Results
The experimental effort using the apparatus described above has recorded data for a variety of conditions from
the swirled-flow PAC reactor, with power depositions up to 130 W, outer air swirl flows up to 50 SLPM, and varied
equivalence ratio with core flows around 10 SLPM. The data and analyses collected in recent efforts includes:
OH PLIF and RST of free jets from coaxial microwave torch flames [10]
OH PLIF measurements of swirl-stabilized chamber flows (no plasma case) [11]
OH PLIF measurements illustrating PAC in typical LBO conditions [11, 14]
Spectroscopic measurements of N
2
vibrational states in the PAC flame [11]
PIV measurements of the cold flow (air only) in the swirl-geometry [15]
Measurements of chamber acoustics in swirl-stabilized and PAC flames [14, 15]
POD analysis of OH PLIF with comparison to chamber acoustics [16]
The work by Rajasegar et al. [11] has shown the influence of continuous microwave power on swirl-stabilized
flames, characterizing the impact of swirl flow on flame stand-off distance and lean blow-off (LBO) limit, and also
demonstrating the impact of microwave power coupling to lower the LBO limit (by >40%) and increase the OH
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
4
radical content (~150%) with addition of microwave power [11]; this work also illustrated significant vibrational
excitation of N
2
by the plasma (T
vib
~ 6100 K, T
rot
~ 2800 K).
A key observation in this configuration is the transition of a lifted flame to a body-anchored flame with the
introduction of microwave excitation [11, 14]. The OH PLIF results have shown that the PAC improves stability
near LBO, and significantly impacts the character of the flame in the wake of the annular injector compared to the
swirl-only case. In the swirl-only case, the OH flame is lifted, and stands off the annular injector. The lift-off
distance decreases as fuel-to-air equivalence ratio increases, and also decreases as the tangential swirl flow rate
increases [11]. For the baseline conditions of OH experiments (10 SLPM in main jet, 30 SLPM air in the swirler,
and
= 1), the lift off is ~18 mm measured at the base of the OH flame. The OH concentration tends to maximize
above this region, resulting in two lobes in the 2D PLIF plane (a toroid shape in 3D). With the introduction of the
coaxial microwave plasma, the OH flame “anchors” in the center-body wake, Figure 2. As power is increased, the
OH content at the centerline jet increases, with one maximum at the centerline (no dual lobe in 2D PLIF as in the
lifted flame).
Figure 2. 2D OH PLIF showing influence of plasma on flame structure. The swirl air flow rate is 30 SLPM, and the air
flow through the coax is 10 SLPM.
Study of the acoustics in the swirl-stabilized PAC chamber results in some key observations [14, 15]: (i)
decrease in the fundamental mode intensity when plasma is introduced, (ii) shifting of the resonant modes to higher
frequency with increased plasma power deposition, and (iii) redistribution of energy to higher frequency modes as
fuel-to-air equivalence ratio is increased. Additionally, the magnitude of the pressure fluctuations in the air plasma
case were found to be significantly lower than in the combustion cases, and the shifts in frequencies of the acoustic
modes are well correlated to increases in plasma power deposition, suggesting the positioning of the modes as an
indicator of average chamber temperature. This shift in the resonant mode within the chamber is in good agreement
with the gas heating provides in other microwave discharge experiments (e.g. Hammack et al. [10]).
Further work by Rajasegar et al. [16] applied Proper Orthogonal Decomposition (POD) of OH PLIF to study the
impact of plasma on flame stability, with comparison of these results to chamber acoustics measurements. The POD
results confirmed that the introduction of plasma improved stability, decreasing fluctuation in heat-release and
pressure. With plasma power deposition corresponding to <5% of the thermal power output, heat-release
fluctuations (determined by POD) were significantly reduced, establishing improved mean energy content up to
~23%. Similarly, RMS pressure fluctuations were reduced by 47%.
III. Swirl-stabilized PAC Chamber Simulation Development
The advanced BLAZE Multiphysics™ Simulation Suite (BLAZE-7, http://blazemultiphysics.com) [19], which
was developed by CU Aerospace, is being applied for high-fidelity, multi-dimensional modeling of a plasma-
assisted combustion test chamber. The initial application of BLAZE-7 to simulate plasma assisted combustion in a
free-jet microwave torch configuration was reported by Palla et al. [14]. Cold-flow simulations of the swirl-flow
chamber were presented in previous work along with experimental efforts [14, 15]. Additional applications of
BLAZE to plasmas and laser systems have been reported [19-23].
BLAZE-7 is comprised of a number of modular inter-operable and highly scalable MPI-based parallel finite-
volume models for the analysis of complex physical systems dependent upon laminar and turbulent fluid-dynamic
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
5
(incompressible and compressible subsonic through hypersonic regimes), molecular transport with gas kinetics, non-
equilibrium gas- and plasma-dynamic, electrodynamic, thermal, optical (radiation transport and wave optics), and
fluid/solid interaction physics using any modern computational platform (Windows, Mac, Unix/Linux). BLAZE is
compatible with a number of free, open-source, yet commercial quality grid generation and post-processing software
packages, which greatly reduces training and operating costs. The cases shown here applied Gmsh [23] for grid
generation and ParaView [24] for post-processing of solutions. Details of some typical meshes for the swirl-
stabilized PAC chamber were discussed in previous work [14,15,18]. BLAZE is also compatible with state of the art
commercial grid generation and post-processing solutions. Additionally, BLAZE users can create, compile, and
include their own complete models into BLAZE without any knowledge of parallel programming, input/output,
grid/mesh formats, sparse linear system solution schemes, etc. as this functionality is provided to the user by the
simulation engine via an easy to use application programming interface (API).
A number of industry standard, yet challenging example cases have been run with the BLAZE Multiphysics
software package for validation, including the driven cavity, the backwards facing step, the rotating cylinder,
unsteady von Karman vortex shedding, NASA turbulent airfoils, gas- and plasma-kinetic simulations, coupled gas-
kinetic / optical field simulations, supersonic shock and hypersonic flow simulations, heat exchanges, neutral
chemistry and plasma-driven hydrocarbon combustion simulations, and wave-optics simulations among others (see
[19] for sample results).
A. Combustion Kinetics Models
Current work on neutral combustion calculations is taking a modified approach from previous work [18], which
applied portions of the GRI-Mech 3.0 model for methane oxidation. The new approach conducts initial simulations
with a simple 8-step mechanism [25], followed by simulation using the DRM-19 model (Detailed Reduced
Mechanism) [26]. These mechanisms are reviewed briefly here.
1. Simplified 8-step Model
For preliminary combustion simulations discussed here, a simple 8-step mechanism was applied. The simplified
combustion mechanism, used primarily for development of combustion modeling schemes, e.g. meshes and fluids
modeling approaches, is based on a highly-simplified methane molecule cracking mechanism along with limited
treatment of intermediate species.
Table 1. Simplified 8-step methane combustion mechanism [Hyer, 1991]
# REACTION
1 CH4+0.5O2(3X)→CO+2H2
2 CH4+H2O→CO+3H2
3 H2+0.5O2(3X)→H2O
4 H2O→0.5O2(3X)+H2
5 CO+H2O→CO2+H2
6 CO2+H2→CO+H2O
7 C2H6+O2(3X)→2CO+3H2
8 C2H6+2H2O→2CO+5H2
2. Detailed Reduced Model
Reduced methane reaction sets (based on GRI-Mech [27]) were reviewed during this period to determine the
most practical detailed mechanism to apply in combustion simulations. The DRM-19 mechanism [26], a 19-species
“detailed reduced model”, was selected as the best option as it provides a significant reduction in total reactions
compared to GRI-Mech 3.0, without requiring the use of any algebraic relations or quasi steady-state species. The
reduction technique for producing the reduced model was described by [28]. The method applies two criteria to
identify non-contributing reactions:
Chain Branching: irref
ReR
Heat Release: maxii q
RH eQ , with
max maximum ii
QRH
i
R and i
H
are respectively the reaction rate and enthalpy change of the ith reaction. The reference reaction
rate ref
Ris that of H + O2 OH + O. The coefficients are chosen to be much smaller than unity, 0.02
rq
ee .
A custom MATLAB script was used to convert DRM-19 reactions in CHEMKIN II format to the XML format
used for BLAZE input. The pressure-dependent reaction coefficients were evaluated at 1 atm for use in the BLAZE
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
6
input (which does not currently support pressure dependence, or interpolation between high pressure and low
pressure rates).
B. Grids and Solver Setup for Validation Cases
Two- and three-dimensional neutral methane combustion simulations presented in previous reports have been
updated to include meshes which are more representative of the experimental geometry and to include the
previously described 20-specie, 168-reaction methane combustion mechanism. Two-dimensional calculations were
performed using the BLAZE Pressure Coupled Navier-Stokes, Standard k-
, and Molecular Transport models on an
18,436-cell finite volume multi-block mesh. Structured rectilinear mesh cells with a characteristic mesh spacing of
2.5x10-4 m were used in the regions of peak combustion activity and unstructured quadrilateral cells were used
downstream of this region. An illustration of this mesh is presented in Figure 3.
Figure 3. Updated two-dimensional methane combustion simulation mesh.
Three-dimensional calculations will be performed using the BLAZE Pressure Coupled Navier-Stokes, Standard
k-, and Molecular Transport models on a 594,928-cell finite volume cell single-block mesh. Unstructured
hexahedral cells were used throughout with finer cells near the core and swirler injectors and coarser cells
downstream of the regions of peak combustion. An illustration of this mesh is presented in Figure 4.
Figure 4. Updated three-dimensional methane combustion simulation (left) geometry and (right) mesh.
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
7
IV. Neutral Combustion Simulations
A. Neutral Combustion with the 8-step Mechanism
The time-evolution of gas temperature and CH4, H2, and H2O mass fractions along a central stream line
beginning at the core flow injector as a function of core flow rate assuming no swirler flow is shown in Figure 5.
Calculations predict more rapid combustion in cases with higher core flow rates and velocities.
Figure 5. BLAZE calculated (top-left) gas temperature, (top-right) CH4 mass fraction, (bottom-left) H2 mass fraction, and
(bottom-right) H2O mass fraction as a function of residence time along central stream line leaving the core injector for 10,
30, and 50 SLPM core flow rates with no swirler flow for the three-dimensional simplified 8-step methane combustion
case.
As a consistency check, one-, two-, and three-dimensional simulation results were compared along predicted
central streamlines for a 10 SLPM core flow rate, 0 SLPM swirler flow rate case. Calculations indicate relative
consistency between the predicted time-evolutions of gas temperature and CH4, H2, and H2O mass fractions in one-,
two-, and three-dimensional simulations, Figure 6.
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
8
Figure 6. BLAZE calculated (top-left) gas temperature, (top-right) CH4 mass fraction, (bottom-left) H2 mass fraction, and
(bottom-right) H2O mass fraction as a function of residence time along central stream line leaving the core injector for the
10 SLPM core flow rate with no swirler flow for the one-, two-, and three-dimensional simplified 8-step methane
combustion cases.
Three-dimensional turbulent flow combustion simulations were constructed using the simplified 8-step methane
combustion mechanism described above. Calculations were performed using the BLAZE Pressure Coupled Navier-
Stokes, Standard k-, and Molecular Transport models on 78,800 cell, rectilinear finite volume mesh. Calculations
were run in parallel on 96 cores of the high-performance computing cluster Taub at the University of Illinois. A 10
SLPM core, 30 SLPM swirler flow case was used as the three-dimensional baseline simulation. Predicted gas
temperatures and CO2 mass fractions are shown in Figure 7. The core injector is apparent as the bottom of the
vertical jet and the swirler inlets are shown at the lower edges.
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
9
Figure 7. BLAZE calculated gas temperature and CO2 concentration for the 10 SLPM core flow, 30 SLPM swirler flow
baseline three-dimensional simplified 8-step methane combustion calculation.
B. Neutral Combustion with the Detailed Reduced Mechanism
1. 1-D Neutral combustion with DRM-19
1-D Neutral combustion (no plasma) cases were simulated using a modified version of DRM-19 as described
above. Calculated gas temperatures as a function of stream-wise (axial) location are presented in Figure 8. Similar
temperatures to the 8-step mechanism were achieved, and the results are in relatively good agreement with
experimental results [10].
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
10
Figure 8. BLAZE calculated gas temperature as a function of axial location for the baseline one-dimensional 168-reaction
methane combustion calculation.
The predicted increase in gas temperature downstream of the combustion region is consistent with
thermalization of the methane chemical energy. The predicted peak OH mass fraction, Figure 9, of 4.55×10-3
corresponds to a concentration of 1.42×1017 m
-3 which is marginally larger than the 3 to 5×10
16 m
-3 peak OH
concentrations observed experimentally.
Figure 9. BLAZE calculated relevant species mass fractions as a function of axial location for the baseline one-
dimensional 168-reaction methane combustion calculation
2. 2-D and 3-D Results with DRM-19
The baseline case selected for development of these two-dimensional neutral methane combustion simulations
included a core injector flow rate of 10 SLPM and a total swirler flow rate of 30 SLPM. Two-dimensional
simulations are presented here for 1:1 air/methane stoichiometric mixture. Representative aspects of the fluids
solution for these two-dimensional simulations, i.e. fluid velocity and turbulent kinetic energy, are presented in
Figure 10.
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
11
Figure 10. BLAZE calculated two-dimensional (left) fluid velocity for 1:1 and 1:1.3 air/methane stoichiometric mixtures,
and (right) fluid turbulent kinetic energy for 1:1 air/methane stoichiometric mixture.
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
12
Gas temperature and concentrations of OH, O
2
and CO
2
are presented in Figure 11. Additional simulations (not
shown for brevity) predicted a slight asymmetry is predicted in the flame for the 1:1.3 air/methane stoichiometric
mixture. However, peak OH concentrations are in reasonable agreement with data for cases with these flow rates.
These two-dimensional simulations predict flame dimensions and OH concentrations consistent with experimental
data, however the lifted flame character is not predicted in this 2-D case.
(a) T
gas
(b) [OH] (c) [O
2
] (d) [CO
2
]
Figure 11. BLAZE calculated two-dimensional (a) gas temperature, (b) OH concentration, (c) O
2
concentration, (d) CO
2
concentration for 1:1 air/methane stoichiometric mixture.
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
13
As presented previously, two-dimensional simulations predict flame dimensions and OH concentrations
consistent with experimental data. Analogous three-dimensional simulation results for gas temperature, OH
concentration, O2 concentration, and CO2 concentration for a 1:1 air/methane stoichiometric mixture are presented in
Figure 12 through Figure 15 and are consistent with the results obtained from the two-dimensional simulations.
For visual clarity each three-dimensional simulation result is presented in both isosurface and streamline form and
only the region of the simulation containing combustion activity is shown.
Figure 12. BLAZE calculated three-dimensional gas temperatures for a 1:1 air/methane stoichiometric mixture as (left)
isosurfaces, and (right) streamlines.
Figure 13. BLAZE calculated three-dimensional OH concentrations for a 1:1 air/methane stoichiometric mixture as (left)
isosurfaces, and (right) streamlines.
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
14
Figure 14. BLAZE calculated three-dimensional O2 concentrations for a 1:1 air/methane stoichiometric mixture as (left)
isosurfaces, and (right) streamlines.
Figure 15. BLAZE calculated three-dimensional CO2 concentrations for a 1:1 air/methane stoichiometric mixture as (left)
isosurfaces, and (right) streamlines.
V. Next Steps for PAC Simulation
Development of an augmented version of the previously described 20-specie, 168-volume reaction atmospheric
pressure air-methane combustion mechanism capable of representing the inclusion of air plasma effects has been
completed. The augmented methane plasma combustion mechanism is based on the following 47 species: N2,
O2(3X), CO2, CH4, O(3P2), H, H2, CO, H2O, C2H6, OH, HO2, CH2, CH2O, CH3O, CH3, HCO, CH2(S), C2H4, C2H5,
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
15
Ar, Ar+, Ar(1S5), e-, N(4S3/2), N+, N2+, N2(A1), N2(A1), N2(A3), N2(B3), N2(B3), N2(v=1), N2(v=2), N2(v=3),
N2(v=4), N2(v=5), N2(v=6), N2(v=7), N2(v=8), N2(W3), NO, NO2, O2+(X2g), O2(a1g), O2(b1g), O3. The methane
plasma combustion mechanism includes 25 surfaces reactions and 606 volume reactions (not presented here for
brevity). Preliminary two-dimensional simulations based on this mechanism, the neutral combustion simulations
previously presented, and various plasma power levels have been completed. Parametric studies based on these
simulations, including comparisons to data are in progress.
VI. Conclusion and Future Work
The neutral combustion simulations with simplified mechanisms described here are a preliminary step toward
the simulation of PAC in a swirl-flow chamber, with the next step being the addition of plasma kinetics (EEDF-
dependent electron-impact mechanisms) and power deposition to the simulations. The results obtained are as
expected for oxidation of methane, however there are some discrepancies compared to experimental values (e.g.
higher [OH] with the DRM-19 mechanism compared to experimental values). This may be a consequence of the
reduction procedure. The temperature obtained with both mechanisms are similar to experimental results. Results
with varied core flow (no-swirl) showed the impact of jet velocity on the gas heating, with the flame developing
over a shorter residence time as the flow rate increases. Future work should demonstrate whether or not this
simulation approach can predict the dependence of the lifted flame height on the ratio of swirl flow to core flow
(decrease in flame height with increased swirl). Good agreement was found comparing 1-D, 2-D, and 3-D profiles in
simulations of various flame species.
Future work will be guided toward simulating the validation cases described in this paper (i.e. 10 SLPM core
flow, 30 SLPM swirl, and power deposition up to 130 W). A variety of data have been collected near this operating
condition for use in validation work. Work on further parametric studies for both neutral-combustion (swirl-only)
and PAC conditions based on the apparatus described here is currently in progress. In particular, the 3-D simulations
can be validated by comparison to various OH PLIF measurements, as well as operating temperatures determined
from spectroscopic measurements. Prediction of the impact of plasma power deposition on the OH flame shape and
character, as well as the increased gas heating observed with spectroscopic measurements and acoustic
measurements will be a key objective. Another objective will be the prediction of lean blow-off limits, and the
stabilization near these conditions with the introduction of power deposition. Continued work applying non-
equilibrium plasma experiments, with well-defined geometries and comprehensive diagnostic suites, will offer
excellent opportunities to validate multiphysics modeling techniques for PAC applications.
Acknowledgments
This work was supported by Air Force STTR contract #FA8650-15-C-2547 “Non-equilibrium Plasma-Assisted
Combustion Efficiency Control in Vitiated Air.” The authors wish to thank the program manager Paul Litke of Air
Force Research Laboratory for helpful interactions and guidance, and co-investigators Gregory Elliott and Marco
Panesi of University of Illinois Aerospace department for their collaboration.
References
[1] Starikovskaia, S. M., “Plasma assisted ignition and combustion,” J. Phys. D. Appl. Phys., Vol. 39, pp. R265-R299
[2] Starikovskiy, A., and Aleksandrov, N. "Plasma-assisted ignition and combustion," Progress in Energy and Combustion
Science Vol. 39, No. 1, pp. 61-110 (2013).
[3] Zhang, Y., Li, Y., Wang, Y., Liu, C., and Eliasson, B., “Plasma methane conversion in the presence of carbon dioxide using
dielectric-barrier discharges,” Fuel Proc. Tech., vol. 83, pp. 101-109 (2003).
[4] Mitsingas,C. M., R. Rajasegar, S. Hammack, H. Do, T. Lee, “High Energy Efficiency Plasma Conversion of CO2 at
Atmospheric Pressure Using a Direct Coupled Microwave Plasma System,” IEEE Transactions on Plasma Science, Vol. 44,
No. 4, pp. 651-656 (2016).
[5] Noorani, R., and Holmes, R., “Effects of Electric Fields on the Blowoff Limits of a Methane-Air Flame,” AIAA Journal,
Vol. 23, no. 9, pp. 1452-1454 (1984)
[6] Chintala, N., Bao, A., Lou, G., and Adamovich, I., “Measurements of combustion efficiency in nonequilibrium RF plasma-
ignited flows,” Combustion and Flame, Vol. 144, pp. 744-756 (2006).
[7] Uddi, M., Jiang, N., Mintusov, E., Adamovich, I., and Lempert, W.R., "Atomic Oxygen Measurements in Air and Air/Fuel
Nanosecond Pulse Discharges by Two Photon Laser Induced Fluorescence," 39th AIAA Plasmadynamics and Lasers Conf.,
23-26 June 2008, Seattle, Washington, AIAA 2008-1110 (2008).
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
American Institute of Aeronautics and Astronautics
16
[8] Nishiyama, A., Moon, A., Ikeda, Y., Hayashi, J., and Akamatsu, F., “Ignition characteristics of methane/air premixed
mixture by microwave-enhanced, laser-induced breakdown plasma,” Optics Express, Vol. 21, No. S6, pp. A1094-A1101
(2013).
[9] Rao, X., Hammack, S., Carter, C., Grotjohn, T., Asmussen, J., and Lee, T., “Microwave-Plasma-Coupled Re-Ignition of
Methane-and-Oxygen Mixture Under Auto-Ignition Temperature,” IEEE Trans. Plasma Sci., Vol. 39, No. 12, pp. 3307-
3313 (2011).
[10] Hammack, S., Rao, X., Lee, T., and Carter, C., “Direct-coupled Plasma-Assisted Combustion Using a Microwave
Waveguide Torch,” IEEE Trans. Plasma Sci., Vol. 39, No. 12, pp. 3300-3306 (2011).
[11] Rajasegar, R., C. M. Mitsingas, E. K. Mayhew, S. Hammack, H. Do, T. Lee, “Effects of Continuous, Volumetric, Direct
Coupled, Non-Equilibrium, Atmospheric Microwave Plasma Discharge on Swirl Stabilized, Premixed flames,” IEEE
Transactions on Plasma Science, Vol. 44, No. 1, pp. 39-48 (2016).
[12] Vincent-Randonnier, A., Larigaldie, S., Magre, P., and Sabel’nikov, V., “Plasma assisted combustion: effect of a coaxial
DBD on a methane diffusion flame,” Plasma Sources Sci. Technol., Vol. 16, pp. 149-160 (2007).
[13] Stockman, E., Zaidi, S., Miles, R., Carter, C., and Ryan, M., “Measurements of combustion properties in a microwave
enhanced flame,” Combustion and Flame, Vol. 156, pp. 1453-1461 (2009).
[14] Zimmerman, J., Palla, A., King, D., Carroll, D., Mitsingas, C., Hammack, S., and Lee T., “Simulations of Plasma-Assisted
Combustion Flames in Coaxial Microwave Reactors,” 54th AIAA Aerospace Sciences Meeting, 4-8 January 2016, San
Diego, CA (AIAA 2016-0190).
[15] Zimmerman, J., Mitsingas, C., Rajasegar, R., Palla, A., King, D., Carroll, D., and Lee T., “Plasma-Assisted Methane
Combustion in Swirled-Flow Reactors: Experiments and Multiphysics Simulations,” AIAA Aviation, 13-17 June 2016,
Washington, D.C., AIAA-2016-4139 (2016)
[16] Rajavasanth Rajasegar, Constandinos M. Mitsingas, Eric Mayhew, Tonghun Lee, and Jihyung Yoo. "Proper Orthogonal
Decomposition for Flame Dynamics of Microwave Plasma Assisted Swirl Stabilized Premixed flames", 55th AIAA
Aerospace Sciences Meeting, AIAA SciTech Forum, (AIAA 2017-1973)
[17] Kim, W., Snyder, J., and Cohen J., “Plasma assisted combustor dynamics and control,” Proc. Comb. Inst. vol. 35, pp. 3479-
3486 (2015).
[18] Palla, A., Zimmerman, J., Carroll, D., Mitsingas, C., Hammack, S., and Lee T., “Multiphysics Simulation of Plasma-assisted
Combustion Dynamics in a Direct-Coupled Microwave Reactor,” 46th AIAA Plasmadynamics and Laser Conf., 22-26 June
2015, Dallas, TX (AIAA 2015-2661).
[19] BLAZE Multiphysics Simulation Suite, Software Package, Ver. 7.0, http://www.blazemultiphysics.com
[20] Palla, A.D., Carroll, D.L., Verdeyen, J.T., and Solomon, W.C., “Post-Discharge Modeling of Recent ElectricOIL
Experiments,” 36th AIAA Plasmadynamics and Lasers Conf., 6-9 June 2005, Toronto, Ontario Canada, AIAA 2005-4919
(2005).
[21] Palla, A.D., Carroll, D.L., and Solomon, W.C., “Modeling of a Multi-Stream Injection COIL with Enhanced Mixing
Ejectors,” 42nd AIAA Plasmadynamics and Lasers Conf., 27-30 June 2011, Honolulu, Hawaii, AIAA 2011-4006 (2011).
[22] Hewitt, J., Houlahan, T., Gallagher, J., Carroll, D., Palla, A., Verdeyen, J., Perram, G., and Eden, J., “Role of excited state
photoionization in the 852.1 nm Cs laser pumped by Cs-Ar photoionization,” Appl. Phys. Lett., Vol. 102, 111104 (2013).
[23] Gmsh, automatic three-dimensional finite element mesh generator, Software Package, Ver. 2.8.5, copyright C. Geuzaine and
J.-F. Remacle, 2014
[24] ParaView Parallel Visualization Application, Software Package, Ver. 4.0, copyright (creative commons license at
http://creativecommons.org/licenses/by/4.0/)
[25] Hyer, P., Stocker, D., Clar, I.O., Gravitational Effects on Laminar Diffusion Flames, Creare. X Users’ Group Meeting
Proceedings, pp. 345-372, (1991).
[26] Kazakov, A. and Frenklach, M., “Reduced Reaction Sets based on GRI-Mech 1.2”
http://www.me.berkeley.edu/drm/
[27] Smith, G.P., Golden, D.M., Frenklach, M. et al., GRI-Mech 3.0 http://www.me.berkeley.edu/gri_mech/
[28] Wang, H., and Frenklach, M., “Detailed Reduction of Reaction Mechanisms for Flame Modeling,” Combustion and Flame,
vol. 87, pp. 365-370 (1991).
Downloaded by UNIVERSITY OF ILLINOIS on June 16, 2017 | http://arc.aiaa.org | DOI: 10.2514/6.2017-3674
