172

23. PLASMA-INDUCED IGNITION AND PLASMA-ASSISTED COMBUSTION

IN HIGH-SPEED FLOW

Sergey B. Leonov, Valentin A. Bityurin, Konstantin V. Savelkin, Dmitry A Yarantsev Institute of High Temperature RAS, Moscow

Abstract. The paper are dedicated to the experimental demonstration of plasma technology abilities in the field of high-

speed combustion. It is doing in three principal directions: control of the structure and the parameters of the duct-driven

flows; the ignition of air-fuel composition at low mean gas temperature; and the mixing intensification inflow. The work

has been fulfilled in Institute of High Temperature RAS (IVTAN).

1. Introduction

The analysis of supersonic combustors

performance shows that several principal problems

related to the supersonic combustion and the flame

stabilization, especially in the case of hydrocarbon

fuels are to be solved for the practical

implementation of such a technology. The plasma-

based methods of the combustion management

under scramjet conditions are considered now as

one of the most promising technologies in this field

[1-3].

An electrical discharge’s properties

strongly depend on the conditions of excitation,

flow parameters and characteristics of supplying

electromagnetic power. The analysis of applicable

discharge types can be done from the viewpoint of

plasma-assisted combustion concept, which

consists of three important items: duct-driven flow

control, plasma-induced ignition/plasma-assisted

combustion due to combustion chemistry

enhancement, and inflow mixing intensification.

The electrical discharges, which are

generated under the conditions of high-speed flow,

possess several specific properties. These features

might be important for the discharges’ applications

in a field of flow control and plasma-assisted

combustion. There can be different kinds of plasma

instabilities, for example, longitudinal-transversal

instability of plasma filament, which has been

found out recently (see section 4). It leads to

intensive small scale mixing inflow. Extra method

of combustion intensification is plasma jets

blowing out to main flow.

It is clear that to manage the combustion

process fully under any conditions a large level of

additional energy deposition is required, in a range

10% from flow enthalpy. The combustor must

operate properly under the conditions, which has

been designed for. So the idea is not in a strong

effect of energy release but in a gentle control of

chemical reactions rate and local multi-ignition.

The second direction is to give the gear to force

combustor to work under the off-design conditions.

It can be a temporal mode and the level of required

electric energy is not vitally important. Such off-

design conditions are: low temperature (probably,

due to undesirably high speed of flow), relatively

low pressure, lean composition, bad mixing, etc.

Our experiments are going to simulate off-design

regimes of the model combustor.

Unfortunately, specific information

available now is not quite sufficient for proper

choice of the discharge type. Our understanding

now is that there is no universal decision in plasma

assistance design and the method of application.

Presence of even a small amount of free radicals

(for example O, OH, H, ON) or vibrationally

excited molecules can effectively improve ignition

conditions but require a not small amount of the

electric power. Each specific situation has to be

considered separately. Under these conditions the

experimental tests and verification of some

analytical predictions are needed urgently. This

work is one of such efforts.

2. Duct-Driven Flow Control

The unconventional methods to improve a

supersonic/hypersonic combustor performance

using electric discharge’s plasma are discussed

widely [1-11]. Two main ideas stimulate efforts in

this field: the control of the inlet/diffuser

parameters and a control of the combustion

chemistry under supersonic flow. In both cases the

electrical discharge changes the structure of flow,

and the thermo-chemical and electro-magnetic

properties of the medium. The analysis shows that

the influence of the plasma generation in high-

enthalpy flows leads to consequences that are not

immediately evident. It is clear that addition of

large amount of the thermal energy might lead to

modification of the wave structure in duct-driven

flows. From the other side such an addition can

change the parameters of whole flow significantly

and not to the desired direction. Thus, the

efficiency of the plasma influence on flow structure

is very important at the diminishing of the possible

penalties.

173

Duct-driven flow control.

The scheme of the flow modification by

plasma method is shown in Fig.1. An idea of the

method is that the surface-generated plasma

provides the energy release under the predefined

location and creates a “smooth” plasma layer near

the duct surface. In dependence on the input power

such a layer generation an lead to boundary layer

(BL) modification [12-14], local BL separation or

extensive (global) separation. Shock waves

structure modification accompanies these

processes. Flow parameters such as Mach number,

pressure value and shocks position can be changed

controllably, as well as their distribution in cross-

section. The effect of instabilities damping and

obstacles’ screening has been described in our

papers [6-9]. Moreover the scheme of the model

experiments is corresponded with Fig.1 directly

(see the next sections).

Inlet’s shocks control.

The scheme of the flow structure

modification in hypersonic inlet has been proposed

in [7]. The idea is described in Fig.2. Later the

Princeton University team has proposed the idea of

“virtual lip” [10]. In the first case it is suggested

that the speed of a vehicle is less than the inlet

design Mach number Mdes. The most dangerous

here is the spillage regime when the third shock

from the cowl lip falls upstream the edge of inlet.

The generation of the plasma layer just upstream

the inlet edge can prevent undesired shock

reflection (dashed lines are old shocks). At the

opposite case when M>Mdes the surface plasma

generation could be useful to change the position of

the second shock upstream and stabilize the

location of the third shock reflection near the edge

of the inlet. The scheme of the model experiments

is shown in Fig.3. The change of the shock position

and angle on the artificial wedge and the duct bulk

parameters are studied under the surface plasma

effect.

Physical model of surface plasma – flow interaction.

The model of the phenomena includes

three important items: (1) discharge structure and

parameters in airflow near the surface; (2) plasma

layer generation process; (3) interaction of the

plasma layer under the supersonic flow.

The discharge structure in airflow depends

on the type of discharge sufficiently. The

transversal quasi-DC discharge has been described

in recent publications [2,7-9,11] on

phenomenological manner. Plasma generation

occurs by means of surface multi-electrode

distributive electric discharge at two different

modes: longitudinal and transversal.

The experimental data on plasma

generation near the body’s surface and influence on

parameters and volume of stabilized separation

zone downstream of wall step have been reported

recently [5, 7, 11]. Some data on the surface

discharge in free stream was presented in paper [7].

Fig.1. Scheme of the flow modification by the surface

plasma.

Fig.2. Idea of the inlet’s shock structure adjustment.

Fig.3. Model experiment arrangement on inlet control.

174

In the case of transversal discharge a relaxation

type of the plasma generation process took place.

An initial plasma filament is being blow down,

breaking and starting again in about 10-20us. In

our case the transversal discharge is characterized

by large level of modulation of the main

parameters, including gap voltage, resistance,

radiation and the position of downstream visible

edge of the plasma. The appropriate oscillograms

are shown in Fig.4.

Fig.4. Detailed correlation between voltage and plasma

radiation at instability.

Two regimes of the discharge renovation

are possible: when the plasma is distinguished (low

voltage mode) and repetitive mode with sequential

runs of individual plasma filaments. The frequency

of such a relaxation process is defined by the flow

local velocity and can be tuned. In our specific

conditions the maximal amplitude of the Fourier’s

spectrum occurs near the frequency 30-50kHz.

Under the conditions of the experiment such a

value is well correlated with the characteristic

length of the plasma channel (xmax) about 2cm. The

process is drawn in Fig.5. An initial breakdown

between the flush-mounted electrodes takes place

under the flow conditions (Pst=100Torr) at the

electric field strength about E=3kV/cm. The first

and sequential breakdowns happen, as a rule, inside

the boundary layer due to a lower value of the gas

density there. The gas temperature inside the

plasma channel starts to rise and the channel swells

up itself volumetrically. This time the plasma

filament comes into the main flow-field and blew

down with the flow speed. The filament’s shape

looks like an increasing loop. The gap’s voltage

increases in accordance with the filament’s length

up to the level when the new breakdown is realized

in the position much closer to the electrodes. Such

a relaxation type of process is repeated with the

frequency, which can be referred to a gas-dynamic

time (see chart in Figs.4, 5). The external shape of

the generated on such manner plasma layer looks

like a near-surface wedge, which is streamlined by

the main flow, as it can be seen in photo and

Schlieren photo.

For the process consideration three

assumptions are done: (1) the gas is weakly

ionized, equilibrium and ideal; (2) Plasma

channel expands isobarically (2-D volumetric

expansion); (3) the local energy input is

proportional to the length of the plasma channel at

the constant electrical current. Energy release to the

plasma filament increase the temperature of the

gas:

TRGTctmW p

11 ,

where W- is the local power input, m- is the gas mass in the plasma channel, G1- is the mass flow rate through the plasma layer, T- is the temperature increase. The isobaric conditions give

the relation between the temperature and the

filament’s volume:

VpTRm ,

where

yyV ×z/2,

Fig.5. Scheme of the model consideration.

175

where y- is the plasma layer thickness (filament’s diameter). Simple transformations give the

following expression for the plasma wedge angle :

yW

pzxytg 111)( ,

where -is the flow velocity, z-is the plasma layer depth. The relation between local power input and

the coordinate y can be found from the expression

on the average power Wav , where - is the period of the plasma filaments oscillations and W(x)- is the local power input, L0- is the inter-electrodes gap.

max

0

)(1

x

avxxWW ,

)2()( 0 xLIExW .

Utilizing these expressions it is possible to

find out the plasma wedge expansion angle through

the integral (bulk) characteristics:

pW

xzxytg av

max

22 211

.

Remarkably, that the angle of the plasma

wedge doesn’t depend on the initial mass flow rate

through the plasma area (thickness of an initial

plasma channel y0). In frames of the model the

temperature of the gas could be estimated as

following:

22

0

max0max tgy

xTT st .

Using the typical values of the

experimental parameters (namely: Wav=1kW, z=20mm, xmax=20mm, =550m/s, Tst0=200K,y0=ybl=0.5mm, p=100Torr) the estimations give the following results: 14 and T 2000K.

The direct measurements give very close

result for the plasma wedge angle and a bit more

value of the gas temperature that can be explained

by the specific method of the measurements and by

plasma non-homogeneity. Some difference in value

of the angle (12 experimental and 14 calculated)

could be explained by the energy loses. Well

known that the plasma of molecular gases has got a

large vibration excitation and relaxation time is

rather large in comparison with gasdynamic time.

The interaction of the plasma near-surface

layer with the main flow occurs on the following

manner (Fig.6). We will consider that the plasma

effective “wedge” acts as a solid “wedge”, although

the actual process of the interaction is much more

complex. Nevertheless the effective angle of

oblique shock wave ( ) associated with the plasma

generation can be related to the effective angle of

the plasma wedge ( ), i.e., finally, to input power.

Fig.6.Scheme of oblique shock generation.

In accordance with such an approach the

angle of the oblique shock can be expressed in

terms of:

cossin)1(

sin)1(22

22

MMtg .

If the effective angle of the plasma wedge

is quite small, the expression can be simplified and

increase of the static pressure can be calculated

through the following formulae:

1

1)(

2Mtg ,

2

2 1

1)(

MtgPst .

As the result we can bind the angle of the

oblique shock with the input power to the surface

plasma. It has to be considered that such a simple

model can be applied successfully for the

prediction of the surface plasma effect, including

the thermal chocking.

Experimental setup description.

The experiments have been conducted at

supersonic speeds in short duration blow-down test

installation PWT-10 of IVTAN. The air used in the

blowdown operation is provided to a high-pressure

reservoir, which has a working pressure of 1 to

5Bar. A fast-acting electromagnetic valve with

diameter 50mm and response time less than 10ms

is connected to the nozzle block and the test

section. Following passage through the test section,

the flow is ducted to either the atmosphere or to a

vacuum tank. Two different test sections are used.

176

Both of them have a circular window for

spectroscopic and natural observations. The two-

dimensional test sections are equipped by a special

flush-mounted insert with electrodes. The operating

mode can be characterized by the following

parameters: Mach number M = 1.1-1.99, static pressure 50-300Torr, Reynolds number of

undisturbed flow Re=(4-10) 106/m, boundary layer

thickness =0.5-2mm, bulk enthalpy of the flow at

100Torr is about 20kW, duration of steady-stage

operation 0.2-0.8sec, and typical air mass flow rate

through the duct about G 0.1kg/sec. The surface discharge can be characterized roughly (see below)

by the following parameters: type of plasma –

quasi-DC multi-electrode surface discharge, typical

electric current through an individual electrode Id1-3A, electric field up to E/n=40Td, typical total

input electric power W = 1-10kW, duration of plasma pulse was between 50 and 100ms.

The first test section was designed for the

experiments on plasma ignition at condition of

fixed separation zone. It has a rectangular cross-

section with dimensions 20x100mm, depth of

rearward-facing step 15mm. The electrodes

insertion is installed at the TS brink, upstream the

wall step. The second TS is made from the

dielectric materials. It has rectangular cross-section

with dimensions 20(h) 50(v)mm and two windows

for the Schlieren observations. The photographs of

the test sections and appropriate schemes of

experiments are presented in Fig.7.

The test sections are equipped by the

following diagnostics: 16 channels pressure

measurement system with response time 0.5ms, fast

CCD camera up to 240fr/sec frame rate and

electronic shutter, Schlieren device with the frame

duration down to 1us, spectroscopic CCD camera,

sensors of radiation, voltage, current and magnetic

field. Plasma temperatures within the discharge

were measured using optic spectroscopy, although

it should be noted that the measurements of plasma

parameters is difficult and presents uncertainties

due to the strong non-homogeneity of the discharge

structure.

Fig.8. Typical Voltage-Current record (bandwidth of the

voltage channel is limited).

Fig.7. Test section 1 with fixed separation zone (a). Test section 2 for duct-driven flow control (b).

177

Typical record of gap voltage and

discharge current is shown in Fig.8. The current-

power characteristic of the transversal surface

discharge is presented in Fig.9. The edges in power

input, when the BL separation without

reattachment took place, are pictured in the last

figure for two values of static pressure (see below).

Fig.9. Transversal surface discharge characteristics at

two anodes, M=1.9, inter-electrodes gap 7mm.

Plasma temperatures within the discharge

were measured using optic spectroscopy by second

positive system of molecular nitrogen and ion of

molecular nitrogen, violet system of cyan and

molecular band of CH. A small addition of CO2was used for the CN generation. The method of processing is found on accurate fitting of

experimental and calculated spectra at rotational

and vibration temperatures variation. The plasma

temperature was measured in the region of the

discharge cords using spectroscopic techniques and

was: 2-3.5kK by second positive system of

nitrogen, 3-4.5kK by CH and up to 6.5kK by CN in

depending on the experimental conditions.

Increasing the input power leads to slow rise of the

maximal temperature. The proper interpretation of

spectroscopic data is still a large problem.

Results of observation.

The plasma generation near the surface of

the duct creates the layer of hot air downstream the

electrodes area. In dependence on conditions such a

hot layer appears in different configurations and is

a cause of different sequences.

Mainly two methods have been exploited

for the observations of the flow structure: namely,

the Schlieren shadow method and measuring of the

pressure distribution. The typical images of the

surface plasma-airflow interaction are shown in

Fig.10. The plasma overlayer and an appropriate

oblique shock wave are seen well. The essentially

subsonic area or huge separation zone is generated

near the surface downstream the plasma area.

Fig.10. Standard Schlieren photo of surface plasma

interaction with duct-driven supersonic flow. Exposure

30us.

Three different situations can be

described: low-power plasma (input power below

1kW for pressure about 100Torr) leads to subsonic

layer generation; medium-power plasma (input

power 1-2kW) leads to subsonic layer generation

with local separation; high-power plasma

deposition (more then 2kW or 10% of the flow

enthalpy) leads to global separation processes,

which is presented in Fig.10. All intermediate

regimes of the interaction have been obtained

experimentally. At input power more than 4-5kW

the chocking of the duct takes place under the

standard conditions of the experiment. The weak

shock waves are generated due to wall irregularities

and local separation areas between nozzle and the

duct. A new oblique shock wave generation is

observed in all cases. New shock has practically the

Mach angle at low power of the plasma and

changes the Mach number in whole duct negligibly.

There is possible to change the Mach number of the

flow downstream and the angle of the oblique

shock by means increasing the input power. The

position of the shock wave can be accurate

localized by the place of plasma generation. The

stagnation pressure downstream the plasma area

decreases that is become apparent in increasing of

Po` downstream, which is measured by the Pitot

gage.

At higher level of the power input to the

plasma the result can be characterized by two

important features. The amplitude of the plasma-

induced shock wave increases sufficiently, up to

direct shock generation and the thermal chocking

of the duct. The second peculiarity is that the

boundary layer is separated without sequential

reattachment. Mach number in whole duct can be

178

changed significantly. Accurate analysis of the

pressure measurements data leads to conclusion

that it is a typical situation when the full pressure

near the wall is decrease and the static pressure is

constant practically or increases, when the global

separation takes place. At the same time the

stagnation pressure at the axes increases slightly.

The difference between the cases is well

recognized: in case of large input power the

“stagnation” pressure near the bottom wall occurs

lower than “static” pressure. It means that a

circulative flow near the wall with reverse direction

of the flow velocity vector is observed. The

situations with and without the global separation

are presented in Fig.11. In the first case the mean

power input was W=1kW, in the second case it was

W=1.7kW correspondingly.

The angle of the oblique shock wave due

to plasma generation depends on the level of the

energy release. We can compare this angle with the

calculated data obtained on the base of proposed

above model. The result of such a comparison is

presented in Table 1. The angles should be referred

to Fig.7, “exp” means experimental, “calc” means

calculated.

Table 1. Dependence of oblique shock angle on the input

power of the electric discharge (in degrees).

Power

, kW

Solid

Wedge

,

0 1 1,7 2,4

, exp 14 0 12

, calc 14 0 14 18 21

, exp 45 31 43 46 50

, calc 45 31 45 49 52

Well seen that the oblique shock angle

grows with the plasma power and that the

experimental values a bit less than calculated ones.

Now we are explaining this difference by the

power loses due to vibration reservoir of molecular

gas.

Fig.12. Mach number modification under the surface

plasma generation.

It is clear that the presence of shock wave

with non-Mach angle has to be reflected in Mach

number modification in whole duct. The result of

measurements is shown in Fig.12. The Mach

number has been recalculated on base of pressure

measurements in three sequential sections of the

duct: upstream and downstream the plasma

generator place. It has to be noted that the plasma

effect doesn’t entail any other harmful sequences

like a turbulence or instabilities generation. Quite

the contrary, the gasdynamic instabilities damping

occurs [9].

Fig.11. Pressure redistribution at surface plasma generation.

179

Shocks position control.

The experimental scheme presented in

Fig.3 has been applied to verify the plasma layer

ability for the oblique shocks control (change the

position and angle). Small model obstacles with

10% and 17% of thickness (3 and 5mm

correspondingly) have been installed into the duct

to simulate the inlet’s configuration. The second

model was blunter. They were positioned 16mm

downstream the electrodes area on the bottom wall

of the duct. The oblique shock falls on the top wall

of the duct in about 28 and 52mm downstream the

obstacle fore-edge zone correspondingly. The

typical Schlieren photos of the fore-generated

plasma effect on the shocks position are shown in

Fig.13 for the different level of the power input.

Fig.13. Schlieren images of plasma –flow interaction

under the different power input.

The Fig.14 presents the result of the shock

position control experiment for two different

obstacles. As it could be noted the plasma effect is

relatively more intensive for a small obstacle.

Under the large level of the input power the shocks

position and configuration are defined only by the

plasma. At such power input the global separation

in the duct occurs.

Fig.14. Shock wave position on the top wall of the duct

vs power input.

Fig.15. Mach number modification under the plasma

effect.

At the same time the plasma generation

influence on the flow Mach number in whole duct

is not strong in comparison with the case when the

obstacle is installed. It is correct if the power level

is much less then the chocking level. The Mach

number dependence on the input power is

demonstrated in Fig.15. The difference with a free

discharge is well visible (see Fig.10 for the

comparison). Actually in this conditions the

pressure loses increase with the power rising.

Detail measurements of pressure

distribution downstream the plasma show that in

180

some important cases instead of braking of the flow

we can observe the flow enhancement in terms of

total pressure recovery factor. The shock changes

its position and the visible amplitude is decreased.

It is easy to see how the plasma excitation

transforms the airflow structure not only near the

surface but also in whole duct. The result of the

measured total pressure recovery factor in

dependence on conditions is shown in Table 2. The

measurements have been done in 200mm

downstream the plasma generation place in

comparison with the pressure just upstream the

plasma generation place. The conditions were

chosen when the Mach number fell from about 1.96

down to about 1.44. Seen that the result at plasma

is better than in case of 17% solid profile.

Table 2. Total pressure recovery factor. Po2/Po

(Mo=1.94-1.98, M2=1,42-1.45).

Power,

kW

0 1 1.4 1.7 2.4 2.7

Wedge 0.77

Plasma 0.86

M2=1.7

0.8 0.81 0.8

Wedge+

Plasma

0.71 0.71 0.8 0.81

3. Plasma-induced ignition of hydrocarbon fuel

at low temperature

The mechanisms of the plasma of

electrical discharges influence on chemical

processes in high-speed flow can be considered and

listed as following:

o Fast local heating of the medium.

o Active radicals and particles deposition.

o Shock waves generation.

o Photo-dissociation and ionization.

Local heating of the medium leads to

intensification of the chemical reactions in these

areas. Besides of this the modification of flow

structure can be done by means of controlled

energy deposition. At enough large level of the

input power the artificial separation of the flow can

be realized. It is a method to increase a local

residence time to provide a zone of local

combustion and the real mechanism of the mixing

intensification. Active radicals’ deposition occurs

due to molecules’ dissociation and excitation by

electrons in electric field and by more complex

processes. If the chain chemical reactions are

realized, the deposition of active particles can lead

to large (synergetic) benefit in reactions’ rate as

well as in required power. Very often the first two

mechanisms are inseparable and the active radicals’

generation is equal to hidden heating. Local shock

waves generation promotes the mixing processes in

heterogeneous medium and initiates chemical

reactions due to heating in shock’s front.

Experiments on non-premixed ignition at

low mean temperature have been done in IVTAN

during last 2 years [2,3]. The scheme of the

discharge excitation in case of backward wall-step

is shown in Fig.15.

Fig.15. Test arrangement under the fixed separation

zone.

Two different modes of the discharge

operation have been found in case of separation

zone. At the first mode the discharge is excited

between electrodes. At the second mode the

discharge current has connected to a metallic wall

in separation zone. We are talking about the second

mode in this paper as the most prospective for an

application. Electric energy input to the plasma

volume was up to 10kW at transversal direction of

plate about 0.1m and axial distance between the

electrodes about 10mm. The mean input power and

the gap voltage in dependence on the static

pressure are presented in Fig.16. The discharge

voltage is not proportional to the discharge gap.

The fast video shows that the length of plasma

“cords” is much more in the case of the second

discharge mode, which is reflected in the gap

voltage increasing. Other difference is that the

discharge in separation zone is generally more

stable. The discharge stability in separation zone is

Fig.16. Discharge characteristics vs Pst.

181

deranged at an external influence like magnetic

field or fuel injection. The photo of the discharge

appearance in the separation zone is presented in

Fig.17.

Fig.17. Surface discharge appearance in separation zone.

Fig.18. Pressure redistribution due to plasma effect.

Plasma excitation near and inside the

separation zone effects on static pressure

distribution in this zone. The Fig.18 shows the

diagram of the pressure redistribution in all points

at the different conditions. It can be considered that

the result is the same for explored conditions:

plasma excitation leads to increase the static

pressure near wall step approximately on 15-20%.

Generally, this statement is correct for the standard

and the second mode of the discharge spacing both.

It can be noted that the pressure distribution in a

separation zone is change significantly. The

pressure gradient occurs much less than in the case

without plasma. Such an effect can lead to

increasing of gas exchange between a separation

zone and a main flow.

The test sequence was the following: wind

tunnel start (duration of steady stage 150ms

approximately), 50ms pause, discharge start

(duration 50-70ms), 20ms pause, fuel injection

during 2-20 ms. Pulse type of the fuel injector with

electromagnetic control was used during the test.

The following injector parameters are typical: type

of fuel – liquid hydrocarbon mixture on base of

isooctane’s; type of injection – high pressure

spraying to separation zone; fuel expense -

Gf=10 100mg/pulse=2 5g/sec. Such a portion requires stoichiometric amount of air in a range

100-1500cm3 at the atmospheric pressure. At the

parameters of the test duct this air portion flows

through the separation zone during several

milliseconds in dependence on the conditions.

Fig.19. Sequential images of fuel ignition by plasma

filaments in high-speed airflow.

182

The dynamic of the ignition process was

analyzed on base of temporal behavior of the flame

luminescence and pressure redistribution in the area

of interaction. Typical frames sequence of the fuel

ignition process is shown in Fig.19. It has been

done at 30us of each frame exposure, 4ms of inter-

frames pause and near-IR optical filter application.

The camera sensitivity was decreased in respect of

non-fuel case due to large level of the luminosity.

Fig.20. Image of the flame from the diffuser side.

To define the exact place of the ignition

the video-record from the side of the diffuser

(exhaust direction) has been fulfilled. Appropriate

image is presented in Fig.20. An upper image

shows how the duct is visible without a flame. The

image below shows the duct at the flame presence.

It is easy to recognize that the flame locates inside

of the separation zone. Sometimes a small amount

of the reacting components penetrates to free-

stream. It is should be considered that at the

condition of the test the location of the flame is just

close to the wall step and along the separation

surface due to fuel-free conditions in a main

stream.

The effect of the proper ignition exists at

quite narrow range of time delays between the

plasma and the fuel injection. The good result takes

place only if the discharge plasma and the fuel

injection occur simultaneously. Moreover, the fuel

pulse pre-injection gives the better result in ignition

than standard mode.

The fuel combustion essentially effects on

pressure distribution in the duct, especially in the

separation zone. The chart in Fig.21 shows the

result of pressure measurements. Note, that the

level of the pressure inside of the separation zone at

combustion is equal the static pressure upstream

the backstep. Tree important things should be

considered. The first one is that the flow regime

upstream of the energy release area is not modified

under the ignition (Mach number about M=1.2).

The second statement is that the large additional

energy release to the area took place, much more

than due the plasma generation. The comparison

with the results of CDF simulation allows us to

estimate the power release in about Wcomb=25kW. It means that the combustion completeness lies in a

range =0.2-0.6 in this operation mode. The third

statement is that the combustion process was

stopped at the discharge switching off. It could be

explained by the low gas temperature in flow

without plasma generation.

Graphs in Figs.22 show the plasma/flame

luminosity in wide near-IR and visible spectrum

and in atomic oxygen resonant line. The charts

Fig.21. Pressure redistribution in separation zone due to fuel ignition by plasma.

183

present a case when a double fuel injection took

place. Well seen that the fuel injection and

sequential ignition lead to extreme increase of the

total luminosity as well as in the OH band and O-line.

Fig.22. Integral spectrum plasma luminosity and atomic

oxygen (778nm) radiation from the discharge-flame

occupied area at double fuel injection.

Several important remarks might be posed

in relation with these observations. Analysis of

plasma-flame luminosity spectral distribution

allows marking out two the most intensive zones:

near-IR molecular radiation (not detailed yet) and

very intensive CN band (389nm). At the same time the continual part of the short-wavelength spectra is

increases too at the fuel injection. The integral

radiation reflects the presence of carbons and

carbon-contented radicals in plasma. The radiation

of atomic oxygen is increase dramatically at the

fuel injection (in a range 2-3 orders of magnitude).

The same behavior can be recognized for the OH-band with the hydrogen continuum. The difference

is that the OH radiation maximum occurs in couple milliseconds later. Such an effect is understandable

if taking into account that the OH complex is a product of chemical reactions. The atomic oxygen

behavior is not so clear. From the one side the

atomic oxygen is important radical, which increases

the rate of chemical reactions. From the other side

it is the product of reactions. We recognize this

effect as an important feature of plasma assisted

ignition method.

The reviewing of the experimental data

and results of kinetic calculations allows

considering that the plasma ignition is very

effective in spite of strong non-uniformity of initial

temperature distribution. Dynamics of the ignition

is determined mostly by the maximal gas

temperature in plasma cords. The combustion

process is nonlinear, so the simple averaging

procedure in regards on the temperature is not

applicable. When the combustion is started in

number of points, the gas dynamic mixing takes

place due to large local gradients of temperature

and pressure. Thus the induction time in non-

uniform system might be much less than in

accordance with mean temperature (averaged as

(mi Ti)/ (mi), where “i”- is some part of reacting volume). That is an effect, which is observed.

4. Plasma/MHD Inflow Mixing

The next item of the plasma assisted

combustion concept is the advanced mixing

[16,17]. The test on pulse discharge influence on

mixing processes under supersonic flow has been

carried out in special dielectric test section of

PWT-10 facility. The scheme of the experiment is

shown in Fig.23. Main flow with Mach number

M=2 and static pressure of air about Pst=100Torrcontained a plain jet of another gas (CO2 in this case) with a close actual velocity of co-flow. The

pulse transversal discharge was excited by means

of two flush-mounted electrodes. The discharge

filament crossed the gap (50mm) and plain jet. The

facility has been equipped by Schlieren device with

spatial resolution not worse than 0.2mm and

temporal resolution about 1us.

Fig.23. Draft layout of the experiment on mixing

intensification in high-speed flow.

The discharge can be characterized by the

following parameters: pulse duration more than

50us (actually it has been limited by discharge

channel breakage due to blowout of plasma

filament by flow), maximal discharge current 100-

150A, steady-stage gap voltage U=500-800V,inter-gap distance 50mm, mean E/n parameter

value 30Td, averaged input power up to W=50kW, spectroscopically defined rotational temperature

Tg 3kK. The dynamics of such a discharge plasma filament in ambient air can be considered by means

of observation of Schlieren photos. Sample of such

photos is presented in Fig.24 for different time

delay in respect of discharge breakdown.

Well seen the process of the discharge

channel expansion as well as the discharge-induced

shock wave propagation. The result looks a quite

184

different under the conditions of inflow-generated

plasma filament. Appropriate Schlieren photos are

presented in Fig.25.

Fig.24. Schlieren photos of the discharge dynamics in

ambient air. Pst=100Torr, gap d=50mm.

Fig.25. Schlieren photos of the discharge dynamics in

airflow. M=2, Pst=100Torr.

We suppose to arrow your extra attention

to the second frame. A zone of longitudinal-

transversal instability of the discharge channel is

well seen as an area with strong density

irregularity. This is a zone of intensive mixing

combined with non-homogeneous heating. The

next frames show the process of central jet mixing

with a main flow due to plasma filament

generation. It is important that while the plasma

channel is excited inflow it moves downstream

with a main flow at the same velocity. The mixing

occurs in a gas portion and the energy release

doesn’t lead to a dramatic change of the bulk

parameters of the flow in duct and to thermal

chocking.

5. Preliminary Conclusion.

It is clear that the problem of supersonic

flow control and supersonic combustion

intensification is rather far from the final solving.

But now the extra mechanisms (plasma technology)

are described as the promising contenders of the

mechanical methods. It is quite possible that the

plasma technology can be applied effectively for a

flow/flight control under non-optimal conditions or

off-design regimes. It can give a possibility of fast,

inertia-less control of external and internal flows,

guiding of separation processes, and control of the

high-speed combustion. This paper is presenting

some last results of small-scale experiments in a

field of duct-driven flow control by inflow

generated electrical discharges, plasma-induced

ignition, and plasma-intensified mixing. The last

important understandings could be reviewed as

following:

The structured plasma changes the flow

parameters in duct on a controllable manner. It

occurs due to local heating, shocks generation and

plasma induced separation.

1. Global separation and unsteady local separation

in duct-driven flow due to the surface plasma

generation have been demonstrated

experimentally. It can be used, probably, as the

agent of the flame stabilization.

2. The conception of plasma-assisted combustion

has been formulated. There were considered

that plasma-induced ignition, plasma-intensified

mixing, and flame-holding by plasma

generation are the methods for high-speed

combustion control. Main physical mechanisms

of the plasma effect are the local heating, active

particle deposition, shock waves generation,

local separation, plasma instabilities generation,

and MHD forces.

3. The effect of plasma-induced ignition in non-

premixed high-speed flow has been

demonstrated under the conditions of fixed zone

of separation. The energetic criteria have been

found out.

4. The plasma intensified mixing effects have

been demonstrated by pulse discharge and due

to MHD interaction. Specific plasma instability

in high-speed flow has been considered as

positive agent of the mixing enhancement.

5. Possible penalties are analyzed under an

application of plasma-assisted processes.

185

6. Acknowledgements.

The results, reviewed here, have been

obtained in frame of the work has been performed

partly under the support of EOARD/AFRL/ISTC

(Dr. John Schmisseur), APL/JHU (Dr. David M.

VanWie), Russian Academy of Science (Cor.

Member of RAS Vyacheslav Batenin).

Also the thankfulness should be passed to

Dr. Alexey Bocharov of IVTAN, Prof. Anatoly

Yuriev of Mozhaysky MESA, Dr. Vladimir

Skvortsov and Dr. Yury Kuznetsov of TsAGI for

the useful discussions. Obviously, that the results

would not be possible without excellent work of

IVTAN’s laboratory personnel who participate in

experimental efforts.

References

1. V. Bityurin, A. Klimov, S. Leonov “Assessment

of a Concept of Advanced Flow/Flight Control

for Hypersonic Flights in Atmosphere.”

Presented to 3rd

Workshop on WIG. November

1-5, 1999 / Norfolk, Virginia, AIAA 99-4820.

2. S.Leonov, V.Bityurin “Hypersonic/ Supersonic

Flow Control by Electro-Discharge Plasma

Application.” 11th AIAA/AAAF International

Symposium Space Planes and Hypersonic

Systems and Technologies, Orléans, 29 September – 4 October, 2002, AIAA-2002-5209.

3. S. Leonov, V. Bityurin, A. Bocharov, K.

Savelkin, D. VanWie, D. Yarantsev,

“Hydrocarbon Fuel Ignition in Separation Zone

of High Speed Duct by Discharge Plasma”,

Proceedings of the 4rd Workshop “PA and MHD

in Aerospace Applications”, April, 2002,

Moscow, IVTAN.

4. T. Cain, D. Boyd “Electrodynamics and the

effect of an electric discharge on cone/cylinder

drag at Mach 5”, 37th AIAA Aerospace Sciences

Meeting and Exhibit, January 11-14, 1999/Reno,

NV, AIAA 99-0602.

5. S.Leonov, V.Bityurin, K.Savelkin, D.Yarantsev

“The Features of Electro-Discharge Plasma

Control of High-Speed Gas Flows.” AIAA-2002-

2180, 33-th Plasmadynamic and Laser

Conference, 20-24 May, 2002, Maui, HI.

6. S. Leonov, V. Bityurin, A. Klimov

“Effectiveness of Plasma Method of Flow/Flight

Control.” Proceedings of the Symposium on

Thermal-Chemical Processes, St-Petersburg,

“Leninets”, July, 2002.

7. S. Leonov, V. Bityurin, K. Savelkin, D.

Yarantsev “Effect of Electrical Discharge on

Separation Processes and Shocks Position in

Supersonic Airflow.” 40th AIAA Aerospace

Sciences Meeting & Exhibit, 13-17 January 2002

/ Reno, NV, AIAA 2002-0355.

8. S. Leonov, V. Bityurin, N. Savischenko, A.

Yuriev, “Study of Surface Electrical Discharge

Influence on Friction of Plate in Transonic

Airflow”. AIAA-2001-0640, 39th AIAA

Aerospace Meeting and Exhibit, 8-11 January,

Reno, NV, 2001.

9. S. Leonov, V. Bityurin, A. Klimov, Yu.

Kolesnichenko, A. Yuriev “Influence of

Structural Electric Discharges on Parameters of

Streamlined Bodies in Airflow.” 32th AIAA

Plasmadynamics and Lasers Conference and 4th

Weakly Ionized Gases Workshop 11-14 June

2001 / Anaheim, CA, AIAA-2001-3057.

10. S. Macheret, M. Schneider, R. Miles

“Nonequilibrium Magnetohydrodynamic Control

of Turbojet and Ram/Scramjet Inlets”, AIAA-

2002-2251, 33-th Plasmadynamic and Laser

Conference, 20-24 May, 2002, Maui, HI.

11. S. Leonov, V. Bityurin, A. Bocharov, E.

Gubanov, Yu. Kolesnichenko, K. Savelkin, A.

Yuriev, N. Savischenko “Discharge plasma

influence on flow characteristics near wall step in

a high-speed duct.” The 3-rd Workshop on

Magneto-Plasma Aerodynamics in Aerospace

Applications, Proceedings, Moscow, IVTAN,

24-26 April, 2001.

12. Kazakov A., Kogan M., Kuriachi A., Influence

on the friction of local heat addition to the

turbulent boundary layer. Mech. Of Fluids and

Gases, N1, 1997.

13. Kurjachi A. P., Boundary layer transition by

means of electrodynamics method. Prikl. Math. I

Mech., vol.49, issue 1,1985.

14. A.V.Kazakov, A.P.Kuryachii,

Electrogasdynamic influence on the development

of the small disturbances in a boundary layer in

the thin profile Izv. N USSR, Mekhanika

zhidkosti i gaza, 1, 1986.

15. S.Leonov, V.Bityurin, K.Savelkin, D.Yarantsev,

Plasma-Induced Ignition and Plasma-Assisted

Combustion of Fuel in High Speed Flow.

Proceedings of the 5th Workshop “PA and MHD

in Aerospace Applications”, 7-10 April, 2003,

Moscow, IVTAN.

16. Bocharov A., Bityurin V., Klement’eva I.,

Leonov S. Experimental and Theoretical Study

of MHD Assisted Mixing and Ignition in Co-

Flow Streams // Paper AIAA 2002- 2228, 40th

AIAA Aerospace Sciences Meeting & Exhibit,

14-17 January 2002/ Reno, NV, P.8.

17. A.N.Bocharov, V.A.Bityurin, I.B.Klement’eva,

S.B.Leonov, A Study of MHF Assisted Mixing

and Combustion. // Paper AIAA 2003- 1226,

41th AIAA Aerospace Sciences Meeting &

Exhibit, January 2003/ Reno, NV, P.8.