Note: Descriptions are shown in the official language in which they were submitted.
CA 02778722 2012-06-04
FLOW CONTROL OF COMBUSTIBLE MIXTURE INTO COMBUSTION CHAMBER
Field of the Invention
The present invention relates in general to fluid control of combustible
mixtures of fuel
and air from a supply into a combustion chamber, where the combustion chamber
presents an enlarged cross-section in comparison with the flow.
Background of the Invention
Controlled burning of combustible mixture of oxidizer and fuel, like in a gas
fueled
combustor, or a prevaporized premixed liquid fuel combustor, requires a
balance of
several conditions concurrently. One balance is between a leaner (relatively
higher
oxidizer content) and a richer (relatively higher fuel content) mixture.
Another is the
balance between higher and lower mass flow rates of the mixture. The mass flow
rate
effectively determines a velocity of the mixture in most operational
combustors. Richer
mixtures, with lower mixture velocities tend to flashback. Flashback is a
dangerous
condition where the flame begins to travel back through the supply tube:
obviously a
condition to be avoided because the required thermal protection to safely
contain a flame
is only provided in the combustion chamber. Too lean a flame, and too high a
mixture
velocity tends to result in a blow-out, in which the flame is extinguished.
Too high a
velocity in a rich gas supply can also lead to flame lift-off, which is
another potential
problem with flame delocalization. Within these ranges are the operable limits
of a
combustor, and within the operable limits are the limits for stable
combustion. There are
many features that affect stable operating limits of flames that cross over
disciplines of
chemistry, thermodynamics and fluid dynamics.
Generally, the highest combustion efficiency is provided by the hottest, most
concentrated flame, the richest premix (up to stoichiometric balance). It is
also desirable
for combustors to exhibit various properties, such as: low emissions, high
flame stability,
and stability under changing conditions such as fluctuations in the fuel
supply
composition, temperature, moisture content, thermal demand, etc. There are
applications
for which a leanest safe burning regime is particularly important, and there
are
applications for which a widest operating range (between coolest and hottest
operation)
are particularly important.
There are many designs for combustors that are well suited to particular
applications, but
there are generally only two levers available to control operating conditions
of
combustors, that allow for the varying combustion conditions. Typically
combustors
1
CA 02778722 2012-06-04
control the supplies of oxidizer gas, and fuel, and therefore their ratio.
While various
sensors may be applied to detect different operating conditions, the feedback
typically
only controls a mass flow rate of the oxidizer and the mass flow rate of the
fuel gas.
Dielectric Barrier Discharge devices (DBDs), also known as plasma actuators,
non-
equilibrium plasmas, non-thermal plasmas, or corona discharge devices are
actuable
devices that have known applications. To date, experimental applications of
DBDs have
been mainly in the area of flow-control / aerodynamics. For example, an ionic
jet that is
generated from DBDs has been shown to be effective in controlling the stall of
stationary
airfoils [5], provoking the stall of wind turbine blades [6], damping the
vortex shedding
induced oscillations of bluff bodies [7], controlling the boundary layer
transitions [8], etc.
Typically these are for ambient temperature, ambient pressure flows of non-
combustible
gasses.
Recently, DBDs have also been applied in chemically reacting flows, to enhance
combustion kinetics. In these applications, DBDs improve flame stability
and
consumption efficiency by cracking the fuel and/or air particles into smaller
stable
molecules and active radicals, thus assisting in the initiation of chain
branching reactions,
to better control a supply of reactants for combustion. Such cracking has been
provided
by DBDs in the fuel supply and/or the oxidizer supply. In some applications,
the DBD has
been placed in a premix chamber [9]. In these combustion experiments, which
were
conducted at atmospheric conditions, interactions between the fuel and/or air
particle flow
and the cold plasma was encouraged by requiring the flow to pass through the
plasma
region. To this end, the mixing chamber was configured with the DBD covering
the
complete combustor cross-section, using a high voltage electrode needle
inserted axially
through, and concentric with, the premix chamber. The electric power necessary
to
obtain a significant effect on combustion kinetics has been found to be less
than 1% of
the combustor thermal power. It has also been shown that plasma actuation
improves
the flame blowout limit and reduces ignition delay time by an order of
magnitude. Other
chemical kinetics studies show that application of DBD leads to a more
complete
combustion [10] and helps reducing soot production in diffusion flames [11].
All of these
papers focus on improving stability of a flame by extending a blowout limit of
the burner.
Accordingly there is a need for improved control over a premix supply when the
flow
transits from a premix chamber to an enlarged combustion chamber.
2
CA 02778722 2012-06-04
Summary of the Invention
The advantages of a DBD plasma actuator over other known flow control devices
are at
least that it has a very small profile (minimum protrusion into the premix
supply), simplicity
of use and design, robustness (no moving parts), and yet can have significant
impact on
the flow. In addition the setup needs low-power and low-weight generators for
operation.
Accordingly a premix supply is provided for supplying a fuel and oxidizer gas
into a
combustion chamber for burning, where the premix supply has a smaller diameter
flow
than the combustion chamber. The premix supply has at least one dielectric
barrier
discharge device (DBD) comprising two electrodes separated by a dielectric.
In
accordance with the invention, both electrodes are provided in a single wall
of the premix
supply, or the electrodes are arranged to generate an ionic wind
preferentially directed in
a direction of flow through the premix supply, or the electrodes are arranged
substantially
upstream / downstream of each other.
The at least one DBD may be disposed symmetrically around the premix supply,
for
example adjacent a dump plane defined by the interface of the premix supply
and the
combustion chamber. The electrodes may be arranged to accelerate the fuel and
oxidizer gas in a direction along the wall. The DBD may provide greater
acceleration
occurring closer to the wall of the premix supply, whereby a pipe-flow
velocity profile of
the fuel and oxidizer gas is modified to increase a velocity gradient at the
wall. The at
least one DBD may include a plurality of DBDs, each which being in a same
wall, the wall
defining an outer diameter of a flow of the fuel and oxidizer gas.
Also provided is a burner comprising a premix supply in communication with the
combustion chamber. Furthermore, a method is provided for reducing flashback
within a
burner having a premix supply in fluid communication with a combustion chamber
where
the premix supply has a smaller diameter flow than the combustion chamber. The
method involves: providing in a wall of the premix supply at least one
dielectric barrier
discharge device having two electrodes separated by a dielectric; and applying
current to
the dielectric barrier discharge device to generate an ionic wind
preferentially in a same
direction as a flow through the premix supply, at least when there is an
elevated risk of
flashback along the wall.
Further features of the invention will be described or will become apparent in
the course
of the following detailed description.
3
CA 02778722 2012-06-04
Brief Description of the Drawings
In order that the invention may be more clearly understood, embodiments
thereof will now
be described in detail by way of example, with reference to the accompanying
drawings,
in which:
FIG. 1 is a schematic illustration of a dielectric barrier discharge (DBD),
showing an
induced plasma, and ionic wind;
FIG. 2 is a schematic illustration of flow from a premixer into a combustor,
schematically
showing how flow is altered by DBD actuation;
FIG. 3 is a schematic illustration of the stability range improvements
provided by reducing
flashback, as one application of the present invention;
FIG. 4 is a schematic illustration of a commercial burner in accordance with
the prior art,
in which DBDs may be positioned in accordance with the present invention;
FIG. 5 is a schematic illustration of an experimental burner used to
demonstrate the
present invention;
FIG. 6 is a graph showing typical current and voltage load of the DBD in the
experimental
burner in use;
FIGs. 7 and 8 are panels of images of flames produced with the experimental
burner with
and without DBD actuation;
FIGs. 9 and 10 are graphs showing velocity profiles with and without DBD
actuation; and
FIG. 11 is a table showing equivalence ratio and fuel mass flow rates showing
stability
ranges that are extended by the selective actuation of DBDs, in accordance
with the
present invention.
Description of Preferred Embodiments
Applicant has demonstrated that a stability range of a burner can be extended
with the
modification to flow profile offered by a dielectric barrier discharge (DBD)
(non-thermal
plasma) located within a premix supply leading to a combustion chamber that
has a wider
flow cross-section than the premix supply. The DBD can be provided in a pair
of
electrodes separated by a dielectric, and both electrodes may be in a common
wall of the
premix supply. The electrodes may be arranged to generate the ionic wind in a
direction
of flow through the premix supply. If a first electrode is substantially
upstream of the
second electrode, the ionic wind may include a substantial component that is
parallel to
4
CA 02778722 2012-06-04
the wall in the direction of flow through the premix supply. As DBDs exert
greater field
strengths closer to the dielectric, the DBD naturally accelerates the flow
more closer to a
wall of the premix supply where they are embedded, and accordingly it is
natural to
deploy the DBDs to impart a velocity profile of the fuel and oxidizer gas that
is modified to
increase a velocity gradient near the wall. Regardless of how the ionic wind
is specifically
arranged with respect to the flow, the DBD operates to impart an acceleration
to the
boundary layer flow in the presence of abrupt changes in geometry, such as at
the mouth
of the combustion chamber.
FIG. 1 is a schematic illustration of a DBD 10 that may be used in accordance
with the
present invention, and a plasma 12 and ionic wind 14 it tends to produce. The
DBD 10
consists of two electrodes 16,18 separated by a dielectric barrier layer 20.
Specifically, a
buried ground electrode 16 is shown embedded in a wall 22 of a premix supply,
and a
high voltage electrode 18 is shown schematically coupled to a high voltage,
high
frequency waveform generator 24.
The use of DBDs as ionic jet inductors, is well known in the art. In use, a
high-frequency
AC electric signal of, typically, several kilovolts is applied to the
electrodes. Gas in the
vicinity of the electrodes gets partially ionized, generating a plasma 26. The
electric field
accelerates the charged particles, effectively rarifying the vicinity, and
producing an ionic
wind of a few meters per second in the principal direction of the electric
field. The
rarefaction naturally draws more neutral gas into the vicinity by the positive
gas pressure.
The ions subsequently transfer their kinetic energy to the neutral gas
molecules around
them via collisions, distributing the momentum throughout the gas flow. Thus
the sum
effect on a flow is shown schematically as a wind 28, which draws the gas into
the
vicinity, and ejects it substantially along the wall 22. The wind 28 is
generally needed on
just one side of the dielectric barrier layer 20 and therefore the electrode
16 is embedded
within the wall 22 under a layer of insulation, for example. This avoids the
formation of
unexploited plasma and improves the efficiency of the DBD without changing the
working
principle [1].
It is worth noting that in generating the ionic wind, DBDs do not introduce
any exogenous
material into the gas flow. The DBDs redirect some part of the gas flow
towards the
location where plasma is [2], and accelerate the ions generally parallel to
the dielectric
barrier layer 20. Other investigations [3] have shown that for a fixed
actuator input
frequency, the equivalent body force (DBD's strength) induces flow
acceleration (ionic
wind) that generally increases with increasing input voltage.
5
CA 02778722 2012-06-04
Based on the knowledge in the art, it is expected that for a tubular flow and
fixed mass
flow rate, a DBD will accelerate the flow at the wall by directing flow away
from the
centerline axis towards the wall 22. This is schematically illustrated in FIG.
2, where
tubular flow into a sudden expansion is shown. FIG. 2 shows a premix supply
(i.e. a part
of a premix chamber that is coupled to a combustion chamber for supplying the
premixture of oxidizer and fuel). The bottom and the top parts of the figure
depict the flow
profiles without and with DBD application respectively. With DBD actuation,
the velocity
profile changes from a typical pipe-flow profile to one that shows an increase
in flow
velocity near the wall and a decrease in velocity at the centerline. This is
more
conveniently illustrated with an assumed laminar flow, as shown in FIG. 2,
however it will
be appreciated that flows feeding combustors are typically turbulent. While
the laminar
velocity profile is shown modified, with the highest flow rates closest to the
centre of the
flow without the DBD, and the highest flow rates closest to the wall with the
DBD, it will be
appreciated that this illustration is schematic, and the extent of the profile
modification will
depend on numerous parameters, including the velocity and mass flow rate of
the
mixture, the power supplied to the DBD, the Reynolds number of the flow, and
the
geometry. FIG. 2 also shows the expected influence of this change in velocity
profile in
the tube on the flow characteristics after the sudden expansion, i.e. within
the combustion
chamber. In particular, the location of the reattachment point is expected to
be closer to
the expansion (dump) plane [4], resulting in a stronger outer recirculation
zone. As the
turbulent shear layer is minimized, and separated from the bulk of the flow,
it is expected
that fluctuations and vortices at a periphery of the gas flow entering the
combustion
chamber will be minimized, and that this can improve a stability limits of a
flame.
FIG. 3 schematically illustrates how the present invention provides for a
broadening of the
region of stable operation, specifically with reference to variation of a mass
flow rate and
equivalence ratio. The flashback limit precludes low mass flow rates, for a
given
equivalence ratio. By providing greater velocity of the flow along its
periphery, the
present invention reduces the initiation of flashback, even when operating
with lower
mass flow rates or richer fuel mixtures, which would otherwise tend to
flashback.
Advantageously, the DBD may be disposed on a single wall of the premix supply,
and
therefore preferentially affects a peripheral flow of the fuel and oxidizer
gas flowing
therethrough. Advantageously, the DBD may be provided near a dump plane to
have a
most direct effect on the flow as it approaches the flame. Advantageously the
DBD may
consist in an arrangement of one or more DBDs that are arrayed symmetrically
about a
periphery of the flow to uniformly and preferentially affect the periphery of
the flow. The
DBD may emit a wind that is substantially along a single wall in which both
the electrodes
6
CA 02778722 2012-06-04
are embedded. The ionic wind may be in a direction that follows the flow, for
example to
increase a flow gradient along the wall in the direction of the flow.
If the premix supply defines a pipe flow, there may be one continuous wall
defining the
premix supply, and each DBD may be embedded in this wall. The DBD may encircle
the
premix supply with a pair of conductive rings that are separated by a
cylindrical dielectric
layer, with one of the rings being upstream of the other so as to generate a
field having
electric field lines that are at least somewhat directed parallel to the
surface in a direction
of the flow. Substantially equivalently, there may be several DBDs arranged on
the wall
in a rotationally symmetric group, for example with each DBD being a same
distance from
the dump plane. These DBDs may be independently actuable or may be on a common
bus for concurrent actuation. There may further be several rings or groups at
different
distances from the dump plane.
If the premix supply defines an annular flow, or has a mandrel partially
inserted into it
from the supply side, it may be desirable to provide one or more DBDs on the
mandrel.
The velocity gradient provided by flows in the neighbourhood of walls, as
shown in FIG. 2
with the pipe flow example, leads to a relatively wide variety of velocities
within the
laminar flow.
It has been observed that flashback is known to preferentially occur along
walls. So
increasing a gradient by directing ionic wind in a direction of the flow along
the walls is
likely to reduce this kind of flashback. Similarly, one or more DBDs may be
disposed on
a mandrel or inner wall of the premix supply to direct ionic wind to
additionally preclude
flashback along a boundary layer of the mandrel. Alternatively there may be
DBD
groups/rings on both outer and inner walls of the premix supply. The mandrel
actuation
may be have different frequency, amplitude and/or phase depending than that of
the
outer wall, to optimize desired flow conditions.
It may further prove useful to direct flow in an azimuthal direction, by
orienting one or
more DBDs to emit ionic wind azimuthally, or by actuating the aforementioned
group of
symmetrically arrayed DBDs for actuation at different times as a function of
azimuth. For
example, the azimuthal flow may improve mixing of the mixture, or flame
anchoring. By
using a plurality of DBDs in the premix supply for flow control, a variety of
actuation
regimes may be defined to impede flashback along a plurality of flashback
paths.
It will be noted that the previous reacting-flow studies used DBDs to enhance
of chemical
kinetics to improve flame lean blowout limits. The effect may additionally be
provided to
7
CA 02778722 2012-06-04
some degree using the present invention, however the prior art DBD
arrangements did
not favourably use the fluid dynamic effects of DBDs.
FIG. 4 is a schematic illustration of a commercially available burner having a
plurality of
combustor manifolds with independent fuel supplies, and a diffuser supplying
an oxidizer.
As an illustration of how the DBDs may be deployed in a premix supply of a
combustor,
the enlarged section shows one possible arrangement. It will be appreciated
that
implementation in a wide variety of burners is possible.
Examples
The following examples demonstrate the reduction of the flashback limit
provided with
actuation of DBDs in the premix supply, on one wall of the premix supply, with
one
electrode upstream of the other, for which a resulting ionic wind is
substantially in the
direction of flow (even with the DBD not actuated).
FIG. 5 is a schematic illustration of an atmospheric combustion rig used in
present
demonstrations. A gaseous combustible mixture enters the vertical rig through
four inlets,
equally spaced circumferentially at the bottom of the rig. The flow then
passes through a
settling/flow-conditioning section comprising a diffuser, a stainless steel
honeycomb, a
fine-meshed screen and a contraction, before entering the premixer. The
premixer
includes a first part made from stainless steel, which provides
instrumentation ports and
acts as an interface between the contraction section and the second part of
the premixer.
The second part of the premixer is made out of quartz (a good dielectric) to
facilitate the
integration of DBD. The two parts of the premixer are separated by a fine
stainless steel
wire mesh, which is installed as a security precaution, to prevent the flame
from travelling
upstream, into the contraction section. The wire mesh was grounded through the
metal
section of the premixer as part of the DBD setup. The premixer has an inner
diameter of
0.050 m and a length of 0.152 m.
The flow exiting the premixer enters a quartz combustion chamber that is 0.419
m long
and has an inner diameter of 0.103 m. Two type-K thermocouples are installed,
one in
the stainless steel premixer section and the other at the combustor exit to
measure the
temperatures of the combustible mixture and the exhaust gases, respectively.
The rig is also provided with a tubular central lance (i.e. center body)
through which fuel
or fuel-air mixture may be introduced for diffusion and partially-premixed
flame studies.
For the present work, the centre body was not used and thus the lance was
pulled
upstream to sit in line with the exit of the contraction section, as shown in
FIG. 4.
8
CA 02778722 2012-06-04
The gas supply system used for the studies had five feed lines: one for
compressed air,
and the other four for fuels and inert gases. The gases were fed to a static
mixer where
various fuel compositions and air were mixed uniformly online, before the
introduction of
the combustible mixture to the combustion rig. The fuel lines were provided
with solenoid
valves and check valves for safety. The only fuel used for the present study
was natural
gas, supplied by a commercial fuel line. Typical composition of the fuel is:
Methane:
96.49 vol.%; Ethane: 1.41 vol.%; Nitrogen: 1.31 vol.%; Carbon dioxide: 0.68
vol.%;
Propane: 0.09 vol.%; Normal-Butane: 0.01 vol.%; and lso-Butane 0.01 vol.%.
The air flow rate was metered via an electronic flow controller. Fuel supply
was
controlled using manual needle valve but monitored using an electronic flow
meter. The
controller and the flow meter were calibrated for the correct range of supply
with full scale
accuracy of 1%.
Digital images and videos of the flame were captured using a 12.3 megapixel
Nikon D300
camera with a shutter speed of 1/8000s-30s and repetition rate of 8 frames per
second.
Two different camera lenses were used: a 50mm and an 85mm lens. The camera was
mounted with its image plane parallel to the combustion chamber's centerline
axis.
A Constant Temperature Anemometer (CTA) from Dantec Dynamics was used to
characterize the flow velocity profiles, with and without DBD application.
The
measurements were made under non-reacting iso-thermal conditions,
approximately
1 mm downstream of the combustor dump plane along two orthogonal axes. The
mass
flow rate of air through the combustor during these measurements was set to
match the
average flow velocity under combustion experiments. For each measured point,
10,000
samples were recorded at a rate of 3 kHz. To avoid electromagnetic
interference from
the DBD actuator (operated at 4 kHz) on the velocity measurements, a low-pass
filter of
the CTA system was set at 3 kHz. For these measurements, the combustion
chamber
length was shortened to allow the introduction of the CTA probe.
A photograph of the rig setup is given in HG. 5, which shows the location of
DBD actuator
with respect to the combustor inlet (dump plane). For the purpose of the
experiments the
DBD was installed on the 2.5 mm thick wall of the quartz premixer. A thin
stainless steel
sleeve was inserted concentric to the inner diameter of the quartz premixer
and attached
to the grounded stainless steel mesh (which primarily served as a flame
arrestor). The
stainless steel sleeve thus served as the ground electrode for the DBD
actuator setup.
The length of the sleeve was adjusted to end 0.044 m short of the combustor
dump
plane. This length was chosen to improve the influence of ionic wind from
actuated DBD
9
CA 02778722 2012-06-04
on the flow/flame dynamics and to prevent electric arcing with the stainless
steel dump
plane. A copper foil tape 0.013 m wide, and 0.074 mm thick (0.038 mm of
adhesive and
0.036 mm of metal) was placed on the outer wall of the premixer and covered
with
insulation. The copper foil, which served as the encapsulated electrode, was
connected
to a high-voltage generator from Electrofluids Systems (Minipuls 6). When
supplied with
the required voltage, this DBD actuator configuration provided its ionic wind
in the
downstream flow direction.
A typical AC signal generated by the high-voltage (HV) generator and applied
across the
electrodes for all the experiments reported here, unless otherwise explicitly
stated, is
shown in FIG. 6. It consists of a 4 kHz triangular wave, with a peak-to-peak
voltage of
19.2 kV. The signal was continuously monitored via a probe mounted on the HV
generator and connected to an oscilloscope.
FIG. 6 also shows the time resolved current consumption of the system measured
with a
PEARSON model 4100 current monitor, also connected to the oscilloscope. The
current
curve shows streaks during the rise and fall of the voltage in the cycle.
These streaks
indicate the generation of plasma and thus plasma was only generated for a
fraction of
the cycle. Since the period of the AC signal is generally much smaller that
the response
time of the flow, it is common to consider the plasma generation and the
associated
induced (ionic wind) velocity as a continuous process.
From the typical voltage and current signatures shown in FIG. 6, one may
calculate the
total power consumption for DBD to be 156.6 W, which corresponds to only 2.4 ¨
4.4 cYc,
of the thermal power of the combustion system. However, it may be noted that
all of this
electrical power is not consumed by the DBD actuator because of the losses in
the
system [15].
During the experiments, the combustor was ignited at a given fuel flow rate
using a spark
igniter and the flame was stabilized at the dump plane by adjusting the air
flow rate.
Thereafter, the fuel flow rate was held constant while the air flow rate was
metered via the
DAQ system in predetermined steps toward either the rich (flashback) or lean
(blowout)
conditions. For every measurement point, the flame was first stabilized at the
dump
plane for two minutes to allow for the combustion chamber and the premixer to
achieve
thermal equilibrium. For the flashback points, seven flashbacks were induced.
The
measured fuel and air flow rates for the last five were averaged and used to
produce the
data, while the first two flashbacks were not considered and were aimed mainly
at heating
up the premixer. To ensure repeatability of the data, the experiments were
conducted at
CA 02778722 2012-06-04
an almost constant pace such that the time between every flashback occurrence
was
nearly constant (-4 minutes). Hence, around 30 minutes of burning were
necessary to
record a single data point. A similar procedure was adopted to produce other
data points
corresponding to flame liftoff and blowout.
Controlled flashback occurrences were generated by metering the air flow rate
at a fixed
fuel flow rate. It was found that in the experimental combustor the flashback
happens
through the core flow when DBD actuation is applied. Thus the flashback along
the
periphery of the flow was effectively prevented using the particular DBD
arrangement
used.
To further verify the repeatability of the results, the flame flashback and
liftoff limits at a
fixed fuel flow rate of 0.102 g/s, with and without DBD actuation, were
measured twice on
two consecutive days. The variations were found to be less than 0.65%, thus
giving
confidence in the data accuracy. In addition, standard deviation for all data
points was
also calculated. The maximum uncertainty of the fuel flow rate normalized by
its
corresponding mean value was found to be 0.23%, while the uncertainty in
equivalence
ratio at flashback, liftoff and blowout conditions were found to be 0.27%,
0.18% and
0.51% respectively.
FIGs. 7 and 8 show images of the flame, without and with the application of
DBD,
respectively. For the frames (a) to (d) of each figure, the 85mm lens was
used, while the
50mm lens was used for the rest. These images were recorded at a fuel flow
rate of
0.102 g/s and various air flow rates i.e., various equivalence ratios. On each
image, the
bold lines represent the position of the dump plane and the walls of the
premixer and
combustor sections. It was found that the application of DBD actuation had a
pronounced
effect on the flame, including the flame anchoring mechanism in the combustor,
and
flashback and liftoff limits.
FIGs. 7 and 8, image (a), represent stable flames at the same fuel and air
flow rate
conditions. A comparison between these images shows that the flame
stabilization
mechanisms encountered without and with DBD application are quite different.
Without
DBD actuation, the flame stabilizes on the rim of the premixer, whereas with
DBD
actuation i.e., under the influence of the ionic wind generated by the DBD,
the flame
anchoring is provided by the outer recirculation zones. This behavior is
supported by the
observations available in the literature [4]. The acceleration of flow in the
vicinity of the
wall of the premix supply, by the application of DBD, helps in strengthening
the outer
11
CA 02778722 2012-06-04
recirculation zone in the combustor, which in turn helps in anchoring the
flame at the
dump plane.
Comparison between the images of FIGs. 7 and 8 further shows that the
application of
DBD modifies the flame characteristics at any given operating condition. This
is true both
for increase in equivalence ratio (leading to flame flashback) as well as for
decrease in
equivalence ratio (leading to flame liftoff and blowout). Also, when comparing
FIG. 7 with
FIG. 8, it may be noted that at an equivalence ratio of 0.696 (Image (e) on
both figures),
the flame is attached to the dump plane for the case without actuation, while
it is lifted-off
for the actuated case. Hence, the flame detaches at lower air mass flow rate
with the
DBD actuation. Controlling actuation rates may therefore be useful in
extending a range
of stable flame conditions.
Concerning the flashback process, FIG. 7, images (b) through (d), show that
for the non-
actuated case, as the equivalence ratio is increased through reduction in air
mass flow
rate, the stabilization location of the flame front gradually moves upstream
along the axis
of the combustor until it fully detaches from the rim. Immediately thereafter
the flame
starts to propagate into the premixer. However, for the actuated case of FIG.
8, it will be
noted that although the flame protrudes upstream of the dump plane with the
increase in
equivalence ratio, it stays stabilized until a much higher equivalence ratio.
Thereafter the
flame flashes back in the premixer.
In order to verify that the observed differences in the flame characteristics
were indeed
due to the modification of flow field caused by the actuation of DBD, flow
velocity
measurements were made using the CTA. These measurements were conducted at 1-
mm downstream of the dump plane and across the combustor cross section under
non-
reacting flow conditions at various flow rates. Sample results are shown in
FIG. 9, where
the operating conditions correspond to those of stable flames of FIGs. 7 and
8. The
mean velocity profiles show a dip at the centerline axis of the combustor.
This is
expected for flows exiting a contraction section and not having enough length
to adopt a
fully developed tubular flow, which was the scenario in the current work, and
is expected
in combustion chambers. FIG. 9 also shows that the velocity gradient in the
near wall
region increases with the application of DBD. Because the mass flow rates are
held
constant, the increase in velocity closer to the wall is substantially
compensated by the
decrease in velocity magnitude at the axis. The net increase in velocity
caused by the
acceleration of the ionic wind is marginal.
12
CA 02778722 2012-06-04
FIG. 10 shows mean velocity profiles for operating condition where average
flow velocity
was matched to that at flashback condition (FIGs. 7,8 Image (c)). Because the
equivalence ratio under the DBD case is significantly higher compared to that
of non-
actuated DBD case, the total flow rate for the DBD case is thus lower.
Nevertheless,
comparison of velocity profiles shows that one of the effects of DBD actuation
is to
increase the velocity gradient at the wall. For the conditions of FIG. 10,
this increase is
from 250 s-1 (as measured from without-DBD profile) to 346 s-1. The increase
in velocity
gradient shown in both FIGs. 9 and 10 explains flame liftoff with DBD
actuation (for
example, compare Image (e) of FIG. 7 with Image (e) of FIG. 8, recorded under
the same
flow rate and equivalence ratio).
The combustor performance as encountered during the present work is mapped on
the
stability diagram of FIG. 11. The flashback, liftoff and blowout limits
separate combustor
operation into four different zones. The mapping was conducted by keeping the
fuel
mass flow rate constant and varying the air flow rate in discrete steps either
towards
flame flashback or towards flame liftoff and subsequent blowout. From the
flashback
point of view, decrease in air mass flow rate translates into lower flow
velocities as well as
higher equivalence ratios, which in lean conditions leads to higher flame
speeds [11].
Hence, in the present work, reduction of air flow rate promotes flame
flashback in two
different ways. This behaviour corresponds to the upper region of the
stability diagram.
On the other hand, as the air flow rate is increased, the flow velocity rises,
the
equivalence reduces and so does the flame speed. At sufficiently high air flow
rate, the
flame starts to liftoff and the combustor operation is characterized by an
unstable and
lifted flame. This point corresponds to the liftoff limit defined by the
triangles on the
stability diagram. Ultimately, the flame is completely lifted off and this
behaviour is
observed for operating conditions in the detached flame region of the
stability diagram.
When the air flow rate is increased further, the flame is pushed out of the
combustion
chamber. This condition is indicated by the blowout points (circles) on the
stability
diagram of FIG. 11.
As shown in FIG. 11, stability limits were extended by the DBD actuation for
the three
higher fuel flow rates. In terms of equivalence ratio, a maximum improvement
of 4.6%
was achieved at a fuel flow rate around 0.102 g/s and with a 19.2 kV peak-to-
peak
voltage at 4 kHz frequency applied to the DBD actuator. The fact that the
equivalence
ratio at which flashback occurs increases with the application of DBD, does
not only
mean that it takes place at a lower air flow rate (lower by 4.2% at this fuel
flow rate), it
also signifies that flashback happens at a much higher flame speed. Flame
speeds (SO
13
CA 02778722 2012-06-04
calculated using data from [12] for methane are shown in FIG. 11 for some
selected
flashback points. For the operating point where maximum control is achieved
through
DBD use, the flashback occurs at a flame speed 9.8% higher compared to a flame
speed
of 0.298 m/s under no actuation case. This implies that if the experiments
were to be
conducted at fixed equivalence ratio (thus at fixed flame speed) by varying
the air and
fuel flow rates simultaneously (as is commonly performed in commercial
burners), the
improvement in flashback control via DBD application would have been even
higher.
For the two lower fuel flow rates of FIG. 11 it may be noted that the
application of DBD at
a voltage of 19.2 kV, favours the occurrence of flashback, thus causing a
shrinkage in the
stable operation regime. This is due to the reduction of the axial velocity in
the core flow
that occurs to compensate the velocity rise near the wall when the DBD is
turned on. At
an already reduced air flow rate this additional velocity reduction at the
axis supports
flashback initiated at the core under these flow conditions. Therefore non-
actuation of the
DBD may be an important aspect for maximizing stability conditions within a
burner.
Specifically, Applicant has found that it is possible to improve the flashback
not only back
to the non-actuated limits but even beyond, by tuning the strength of the
induced ionic
wind through the adjustment of the voltage applied to the DBD actuator. As
shown in
FIG. 11, reducing the excitation voltage from 19.2 kVp_p to 13.2 kVp_p,
increased the
equivalence ratio at which flashback occurs by 3.9%.
It will also be noted in FIG. 11 that DBD actuation causes flame liftoff at a
much higher
equivalence ratio as compared to no actuation cases as shown for the three
higher fuel
flow rate conditions. As pointed out earlier, this is due to the rise in the
velocity gradient
at the wall induced by the DBD. However, it was found that DBD actuation does
not
affect the blowout limit, which is a surprising result.
To further verify that DBD actuation was indeed preventing flashback, a flame
was
stabilized at a given fuel flow rate and the DBD actuator was turned on. The
air flow rate
was then reduced to attain an equivalence ratio mid-way between the actuated
and non-
actuated flashback limits on the stability diagram of FIG. 11. In the absence
of DBD
actuation, the flame immediately started to propagate upstream in the
premixer. The
process was repeated a few times and the same result was obtained every time.
In conclusion, a premix supply having a DBD for flow control was demonstrated.
Operation of the DBD improved stability limits of a flame in the adjacent
combustion
14
CA 02778722 2012-06-04
chamber as was successfully demonstrated over a range of operating conditions
in a
premixed atmospheric dump combustor.
References: The contents of the entirety of each of which are incorporated by
this
reference:
[1] Laurentie, J.-C., Jolibois, J. and Moreau, E., 2009, "Surface dielectric
barrier
discharge: Effect of encapsulation of the grounded electrode on the
electromechanical
characteristics of the plasma actuator", Journal of Electrostatics, 67, pp. 93-
98.
[2] Jacob, J., Rivir, R., Carter, C., and Estevadeordal, J., 2004, "Boundary
Layer Flow
Control Using AC Discharge Plasma Actuators", AIAA 2004-2128, AIAA 2nd Flow
Control
Meeting, Portland, Oregon.
[3] Thomas, F. 0., Corke, T. C., lqbal, M., Kozlov, A., and Schatzman, D.,
2009,
"Optimization of Dielectric Barrier Discharge Plasma Actuators for Active
Aerodynamic
Flow Control", AIAA Journal, 47(9), pp. 2169-2178.
[4] De Zilwaa, S. R. N., Khezzarb, L. and Whitelaw, J.H., 2000, "Flows through
plane
sudden-expansions", Internation Journal for Numerical Methods in Fluids, 32,
pp. 313-
329.
[5] Roth, J. R. and Dai, X., 2006, "Optimization of the Aerodynamic Plasma
Actuator as a
Electrohydrodynamic (EHD) Electrical Device", 44th AIAA Aerospace Sciences
Meeting
and Exhibit, Reno, Nevada, pp. 1-28.
[6] Versailles, P., Ghosh, S., Vo, H. D., and Masson, C., 2010, "Preliminary
Assessment
of Wind Turbine Blade Lift Control via Plasma Actuation", Submitted for
publication in
Wind Engineering.
[7] Mureithi, N.W., Rodriguez, M., Versailles, P., Pham, A., and Vo, H. D.,
2008, "A POD
based analysis of the 20 cylinder wake mode interactions", 9th Internal
Conference on
Fly, Prague, Czech Republic, pp. 1-6.
[8] Grundmann, S. and Tropea, C., 2007, "Experimental transition delay using
glow-
discharge plasma actuators", Experiments in Fluids, 42(4), pp. 653-657.
[9] Vincent-Randonnier, A et al., April 2007, "Experimental Study of a Methane
Diffusion
Flame Under Dielectric Barrier Discharge Assistance" IEEE TRANSACTIONS ON
PLASMA SCIENCE, 35(2) pp. 223-232.
[10] Rosocha, L. A., Coates, D. M., Platts, D., and Stange, S., 2004, "Plasma-
enhanced
combustion of propane using a silent discharge", Physics of Plasma, 11, 7
pages.
[11] Cha, M.S., Lee, S.M., Kim, K.T., and Chung, S.H., 2005, "Soot suppression
by
nonthernnal plasma in coflow jet diffusion flames using a dielectric barrier
discharge",
Combustion and Flame, 141, pp. 438-447.
CA 02778722 2012-06-04
,
[12] Vagelopoulos, C.M. and Egolfopoulos, F.N., 1998, "Direct Experimental
Determination of Laminar Flame Speeds", Twenty-Seventh Symposium
(International) on
Combustion, 27(1), pp. 513-519.
Other advantages that are inherent to the structure are obvious to one skilled
in the art.
The embodiments are described herein illustratively and are not meant to limit
the scope
of the invention as claimed. Variations of the foregoing embodiments will be
evident to a
person of ordinary skill and are intended by the inventor to be encompassed by
the
following claims.
16
