Rock 1

EXPERIMENTAL STUDY OF A STRUT INJECTOR FOR

CIRCULAR SCRAMJET COMBUSTORS

Christopher Rock

Graduate Research Assistant

and

Joseph A. Schetz

Advisor, Holder of the Fred D. Durham Chair

Department of Aerospace and Ocean Engineering

Virginia Tech, Blacksburg, VA, 24061

Supersonic combustion is a major challenge in scramjet engine design. Supersonic fuel injection

and mixing research contributes to the effort to make the scramjet a viable option to power

hypersonic aircraft, economical and reusable launch vehicles, and hypersonic missiles. An

experimental study of a strut injector configuration was performed for application to high-Mach-

number scramjets with circular combustion chambers. The strut injector has sixteen inclined, round,

sonic injectors distributed across four struts within a circular duct. The struts are slender, inclined at

a low angle to minimize drag, and have two injectors on each side. The strut injector was

experimentally studied under Mach 4, cold-flow conditions using two different molecular weight

injectants, helium (molecular weight = 4) and air (molecular weight = 28.97). The primary goal of

this study is the refinement of turbulence models for these complex mixing flows. Furthermore,

injectant molecular weight has been identified as a parameter of critical importance in the

development of the turbulence model upgrades. Experimental data such as presented here will be

used to guide the continuing upgrade of turbulence modeling in a closely integrated program.

Nomenclature

A* = plume area of stoichiometric mixture

Cd = discharge coefficient

d = jet diameter

deq = equivalent diameter

Gj = injectant mass flow rate

M = Mach number

P = static pressure

P0 = stagnation pressure

q = jet-to-free-stream momentum flux ratio

R = resistance

U = velocity

w* = plume width

y = vertical distance from the duct centerline

y* = vertical distance to injectant center of

mass

α = mass fraction of injectant

ρ = density

γ = ratio of specific heats

Subscripts

j = jet exit property

∞ = freestream property

I. Introduction

In view of the very high freestream velocity of

scramjets reaching Mach 10, fuel residence time is

on the order of milliseconds1 and supersonic

combustion presents an interesting challenge in

scramjet engines. It is, therefore, desirable to

enhance penetration and mixing of the fuel plume

in order to accomplish rapid combustion leading to

a reduction of the required combustor length,

reducing the skin-friction drag and heat transfer

and increasing the net thrust. To improve the

overall engine efficiency, the injection process

must also induce low total pressure losses. Jet

injector mixing enhancement in high-speed flows

also has applications in other fields such as thermal

protection systems and vehicle control by jet

thrusters.

Many injector configurations have been

studied by various groups in an attempt to produce

enhanced mixing and penetration. Some of these

configurations can be seen in Figure 1 including

Rock 2

wall jets, struts, and swept ramps. Extensive

reviews of injector mixing characteristics are given

in Schetz et al2 and Kutschenreuter3. Flush-walled

injectors are often preferred over in-stream

injectors because they minimize total pressure

losses and heating, but some configurations can

require the use of in-stream injectors in order to

obtain adequate distribution of the fuel across the

combustor. A circular combustor cross-section is

one example where struts might be attractive. Of

course, one has to reckon with the drag of the struts

in assessing engine performance3.

Figure 1: Examples of various injector configurations (from Kutschenreuter3)

Very few of the detailed, high-speed mixing studies

available in the literature concern injection and

mixing in confined ducts representative of

combustors, and one can expect that the effects of

such confinement are very large. This is especially

true for struts protruding into the flow. There are

also bow shocks from the injection process itself.

The purpose of the present research is to

investigate the effectiveness of a four-strut injector

configuration with multiple round, sonic injectors

on each strut in a circular duct for application to

high-Mach-number scramjets with circular

combustion chambers. The nominal Mach 4 air

flow simulates conditions a scramjet combustor

would encounter in Mach 10 flight. The general

goals of cold-flow studies of injection and mixing

in simulated scramjet combustors are first to

determine if the penetration and mixing patterns

observed are in agreement with those used for the

injector design. Second, the experimental data can

be used to gauge the uncertainty in computational

predictions of such flows. The computational tools

can then be used to design and analyze for hot-flow

conditions with known uncertainty. The third and

primary goal of this study is the refinement of

turbulence models for these complex mixing flows.

For this study, two different injectants were used,

helium (molecular weight = 4) and air (molecular

weight = 28.97), since injectant molecular weight

has been identified as a parameter of critical

importance in the development of the turbulence

model upgrades.

II. Experimental Methods

A. Test Facility

These experiments were conducted in the Virginia

Tech blow-down type high-speed wind tunnel

shown in Figure 2, which operates at speeds

ranging from Mach 2 to 7. The blow-down type

wind tunnel offers run times on the order of a few

seconds at high Mach numbers with relatively

steady flow conditions. This facility was obtained

through our close and long-term collaborations

with the Institute of Theoretical and Applied

Mechanics of the Russian Academy of Sciences in

Novosibirsk, Russia. Air (or other working gas) is

supplied from a compressor to charge the storage

bottles visible within the frame at the bottom. A

special fast-acting control valve initiates flow into

the plenum chamber. The flow then passes through

Rock 3

a contoured, converging-diverging nozzle and out

through the diffuser. Due to the working principle

of the tunnel and the fast-acting control valve, there

is a gradual decrease in total pressure during the

run. The variation of the total pressure during the

run is in the range of approximately 10%. For

Mach numbers above 4, an electric heater raises the

total temperature up to 800 K to prevent

liquefaction. The nozzle exit diameter is 100 mm.

The test cabin arrangement permits the use of

relatively large in-stream models, especially at the

higher Mach numbers.

The wind tunnel setup for these experiments

used a converging-diverging nozzle to achieve a

nominal Mach 4 flow in the test section. Nominal

flow conditions involve total pressure and

temperature in the plenum chamber of 1317 kPa

and 295 K. However, there is a weak oblique shock

observed at the end of the nozzle, where the

injector model attaches resulting in actual inflow

conditions of Mach number, M∞ = 3.9, total

pressure 1311 kPa and total temperature 295 K.

Figure 2: Layout of the Virginia Tech high-speed wind tunnel

B. Injection System

The injector setup investigated in this project

resembles the combustion chamber of a scramjet

engine. In a real scramjet, the combustor is situated

downstream of the inlet and an isolator which

compress the ingested air. In the experimental

setup, the ducted strut injector model is mounted

downstream of the convergent-divergent Laval

nozzle of the high-speed wind tunnel. The injector

model consists of a total of 16 injectors distributed

over four struts within a circular duct and the

necessary connections for the injectant supply.

Downstream of the injection position, a cylindrical

flange connects the injector duct to the test cabin.

A traversing system is installed on top of the test

cabin, which positions flow measurement probes

within the cabin. The injectant is supplied from a

group of commercial gas bottles. The mass flow

rate of the injectant is controlled using a system of

two Teledyne-Hastings model HFC-D-307 digital

mass flow controllers. Each mass flow controller

uses a proportional–integral–derivative (PID)

control valve.

C. Ducted Strut Injector Model

The ducted strut injector model is based on a

circular duct extension of the tunnel nozzle. It

contains four struts with 16 circular injection

nozzles. Figure 3 shows a picture of the strut

injector model. The struts have a width of 8.2 mm,

they start at the front of the extension duct (i.e. the

end of the tunnel nozzle), and they extend 148 mm

in the flow direction with an inclination of 10°.

Two 1.52 mm circular nozzles on each lateral side

of each strut create jets that penetrate into the

tunnel crossflow at an angle of 30° relative to the

streamwise axis of the duct. The centers of the

injectors are located 92 mm from the leading edge

of each strut. The number, shape, and size of the

struts were based on drag considerations and

previous experience. The number, size, and

location of the injectors were based on CFD

studies.

Due to the physical obstacle created by the

struts, the formation of shocks at the edges and an

expansion at the rearward facing edge of the struts

can be expected. As a result, high total pressure

losses in such a configuration are unavoidable. As

pointed out before, an optimized injection system

for a scramjet engine should combine good mixing

efficiencies with low total pressure losses. A

system including fixed flow obstacles has to

compensate for this disadvantage by enabling

Rock 4

better mixing in order to remain competitive with

other geometries.

A useful parameter for correlating transverse

jet injection results is called the jet-to-freestream

momentum flux ratio, q , defined as follows

∞∞

=≡)(

)(

)(

)(2

2

2

2

Mp

Mp

U

Uq

jj

γ

γ

ρ

ρ (1)

For this study, two experimental cases were

run using different injectants. One case was run

using helium injection, which safely simulates

hydrogen fuel in a cold-flow, non-combusting

environment. A second experimental case was then

run using air injection. Each case was run with the

same jet-to-freestream momentum flux ratio ( q )

to obtain a similar amount of fuel plume

penetration. For the helium injection case, the total

mass flow rate was set to Gtotal = 22.5 g/s, which

corresponds to q = 3.49 for this injector geometry

and operating conditions. For the air injection case,

the total mass flow rate was set to Gtotal = 62.66 g/s

to match the value of q = 3.49. These values of

q are representative of good practice in strutted

scramjet combustors. For both cases, the injectant

jets are at sonic conditions and are highly under-

expanded. Table 1 summarizes the injection

parameters for the two experimental cases.

Figure 3: Strut injector model Table 1: Injection parameters for the two experimental cases

(for a single injector)

D. Concentration Sampling Probe

In order to analyze the mixing of the helium with

the air freestream, it is crucial to acquire accurate

gas composition measurements. The concentration

is measured in terms of the mass fraction of helium

in the overall gas mixture. To determine this mass

fraction, a special probe is used to simultaneously

sample and analyze the gas mixture at a given

position accurately. The fundamental concept of

the gas analyzer used for this work was developed

at Virginia Tech by Professor Ng4. The

concentration sampling probe is an aspirating type

probe that is attached to a vacuum pump. A picture

and diagram of the concentration probe are shown

in Figure 4. The unit consists of a constant

temperature hot-film sensor operating in a channel

with a choked exit. The housing was designed to fit

around the body of a TSI 1210-20 platinum sensor.

The hot-film has a diameter of 50.8 µm and an

active sensor length of 1.02 mm which is used in

conjunction with a Dantec model 56C17 constant

temperature anemometer (CTA) fitted with a

Dantec model 56C01 CTA bridge. The overheat

ratio of a hot-film sensor is defined as (Rop - R0) /

R0, where Rop is the heated sensor resistance at

operating temperature and R0 is the cold sensor

resistance at ambient temperature. An overheat

ratio of 1.0 was used for the hot-film sensor of the

concentration probe.

The inlet hole at the tip of the probe has the

same diameter as the choked orifice, 0.63 mm.

These diameters were designed to preclude the

occurrence of a standoff shock at the probe tip for

supersonic flow due to the suction of the vacuum

pump through the choked orifice. Schlieren flow

visualization confirmed the absence of a standing

normal shock. The internal probe diameter diverges

from 0.63 mm at the inlet to 3.8 mm at the sensor

plane, causing a normal shock to occur inside the

probe in the diverging channel. By swallowing the

shock into the internal diverging section of the

probe, a stream tube equal in area to the probe

capture area can enter the probe undisturbed and

undistorted. Thus, an isokinetic sampling of the

stream tube is accomplished. Through the

Parameter Unit Helium injection

case,

Gtotal = 22.5 g/s He

Air injection case,

Gtotal = 62.66 g/s

air

Gj [g/s] 1.41 3.92

q [-] 3.49 3.49

Cd [-] 0.70 0.70

Uj / U∞ [-] 1.30 0.47

P0j / P0∞ [-] 0.63 0.70

d [mm] 1.52 1.52

deq = (Cd)1/2d [mm] 1.27 1.27

Rock 5

diverging section and the normal shock inside the

probe, the flow is decelerated to very low

velocities. At Mach numbers around M = 0.05, the

pressure and temperature measured inside the

probe can be approximated to represent the total

pressure and temperature of the fluid.

The concentration sampling probe was

calibrated to measure the helium molar fraction

uniquely related to a given pressure, temperature

and rate of heat transfer sensed at the hot-film. The

hot-film responds to local mass flux variations.

From the known helium molar fraction of the

sample, the mass fraction can be calculated. The

measurement uncertainty of the probe was found to

be approximately +/- 0.005 for helium mass

fraction measurements.

Figure 4: Picture of the concentration sampling probe with an integrated cone-static probe (left) and

diagram of the concentration probe (right)

E. Cone-Static Pressure Probe

The concentration probe has an integrated cone-

static probe (see Figure 4). The cone-static probe is

not required to determine the mixture composition,

but the use of the cone-static probe allows for the

determination of other quantities of interest for a

given flow field using a multiple probe survey

method. The cone-static probe was attached to the

concentration probe to allow simultaneous

concentration and cone-static measurements to be

made. The cone-static probe consists of a 1.59 mm

outer diameter pipe capped with a 10° half-angle

cone. Four small pressure ports are located at 90°

spacing around the surface of the cone. The cone-

static probe is positioned in a location that is

always outside of the oblique shock generated by

the tip of the concentration probe.

F. Miniature Five-Hole Probe

A miniature, fast-response, conical, five-hole

pressure probe is used to measure local values of

Mach number, total pressure, and flow angularity.

A picture and diagram of the five-hole probe are

shown in Figure 5. The probe uses five miniature

piezoresistive pressure transducers, which are

located directly in its tip. The tip of the probe is a

45° half-angle cone with an outer diameter of 1.65

mm. Each pressure port has a diameter of 0.25 mm.

The response time of the probe to a step input is

about 11 ms. Two separate calibrations were

performed to allow for the determination of Mach

number and flow angularity. First, the five-hole

probe was calibrated to determine the Mach

number of an airstream as a function of its port

pressures. The calibration of the five-hole probe to

determine Mach number is necessary due to the

geometry of the probe. The probe has a blunt tip

where the center (Pitot) port is located, which is

surrounded by four peripheral ports. Downstream

of the blunt tip, flow expansion occurs resulting in

a region of lower pressure behind the tip in

comparison to a sharp cone with the same half-

angle. Beyond the region affected by flow

expansion, the pressure distribution will quickly

recover to that of a sharp cone. However, the

peripheral ports for the probe used in the current

study are located in the region affected by flow

expansion. Sharp cone theory cannot be used to

predict the readings for these ports and therefore, it

is necessary to calibrate the probe to determine

Mach number. The calibration was performed over

a Mach number range of 1.6 – 3.9 and includes a

total of 37 data points. This Mach number

calibration is only valid for use of the probe in air,

since mixture composition influences this

calibration curve.

The five-hole probe was also calibrated to

determine the flow angularity of an airstream as a

function of its port pressures. The calibration was

performed at Mach 3.1 with an angularity range of

+18° to -18° of pitch and 0-360° of roll. The

angular calibration of the five-hole probe includes

a total of 795 data points. The angular calibration is

valid over a wide range of Mach numbers and it is

also valid for use in both air and air-helium

Cone-Static

Probe

Concentration Probe

Rock 6

mixtures according to the work of Swalley5. Using

a 40° half-angle cone in experiments run at a Mach

number of 3.55 in air and 21 in helium, Swalley

confirmed the theory that only one calibration

curve is required to determine flow angularity over

a wide Mach number range in either air or helium

for this type of instrument. In addition, Centolanzi

calibrated a 20° half-angle cone to determine the

flow angularity of an airstream at Mach numbers of

1.72, 1.95, and 2.46 and also concluded that the

effects of Mach number on the calibration map are

either negligible or small6.

Figure 5: Picture of the five-hole probe (left) and drawing of the probe tip (right). All dimensions are

in millimeters.

G. Multiple Probe Survey Method and Data

Analysis Procedure

1. Outside of the Region of the Fuel Jet

Outside of the region of the fuel plume, the gas

composition is known to be entirely air, so the

concentration probe is not used in this region. The

five-hole probe Mach number calibration is valid

and the angular calibration is most accurate in this

region. Therefore, the five-hole probe alone can be

used outside of the fuel plume to determine local

values of Mach number, total pressure, and flow

angularity. The data reduction process needed to

convert the port pressures into incoming flow

properties follows that of Centolanzi6. The

measurement uncertainty of the five-hole probe

was found to be approximately +/- 1.5% for Mach

number, +/- 3% for total pressure, and +/- 1° for

flow angularity in this region.

2. Within the Region of the Fuel Jet

Inside the region of the fuel plume, the

properties of the mixture must be accounted for and

the data analysis procedure is more complex. First,

a concentration probe survey is used to determine

the local gas composition. The measurement

uncertainty for the concentration probe is

approximately +/- 0.005 for helium mass fraction

measurements. Next, a method is needed to solve

for the Mach number in the region of the fuel

plume, since the five-hole probe Mach number

calibration is not valid for gas mixtures. To

determine the Mach number in this region, a

multiple probe survey method is used.

Corresponding data points are taken with the

concentration probe, the cone-static probe, and the

five-hole probe. The helium concentration data is

then used with a combination of the Rayleigh-Pitot

formula and a numerical solution of the Taylor-

Maccoll equation for the local gas composition to

determine the local Mach number at each

measurement location. Once the Mach number at

each measurement location is known, the total

pressure and flow angularity can be solved for

using the method of Centolanzi6 as the five-hole

probe flow angularity calibration is valid for air-

helium mixtures. The measurement uncertainty is

approximately +/- 2% for Mach number, +/- 5% for

total pressure, and +/- 1° for flow angularity in the

region of the fuel plume.

III. Experimental Results

A plane 178.3 mm (1.8 duct diameters)

downstream of the circular injector centers was

selected for data measurement purposes. At this

measurement plane, which is about 2 mm beyond

the end of the duct where the flow enters the test

cabin, the flow field downstream of one half of one

strut was surveyed. For the helium injection case,

helium concentration, Mach number, and total

pressure values were measured at the data

measurement plane. For the air injection case,

Mach number and total pressure values were

measured. To check for symmetry, data points

were also measured on the opposite side of the

strut. The symmetry plane for the fuel plume was

found to be shifted approximately 1 to 2 mm

laterally relative to the centerline of the strut. This

slight 1-2 mm shift of the fuel plume over a length

of 178.3 mm is most likely attributed to a small,

but undetected misalignment of the experimental

hardware.

Rock 7

A. Helium Concentration Results

Results for the helium distribution presented as

mass fraction contours across a section of the duct

are shown in Figure 6. These mass fraction

contours were determined from 127 experimental

data points distributed across the fuel plume in the

radial and peripheral directions. The projected

outlines of the strut and the circular injectors are

shown for reference. An examination of Figure 6

shows that at 1.8 duct diameters downstream, the

injectant achieved good penetration across the

combustor cross-section. However, the individual

jets merged into a single large plume and the rate

of mixing was somewhat slow.

Calculations over the measurement grid

provide parameters that characterize the plume and

the mixing behavior, which are summarized in

Table 2. Here, y* is the location of the center of

mass of the injectant in the plume below the duct

wall, w* is the maximum width of the equivalent

stoichiometric hydrogen/air concentration contour,

A* is the plume area within that contour, αmax is the

maximum injectant mass fraction, and yα,max is the

distance from the duct wall to the location of αmax.

B. Mach Number and Total Pressure Results

Figures 7 and 8 show contour plots of Mach

number and total pressure at the measurement

plane for the helium injection case compared to the

air injection case. The projected outlines of the

strut and the circular injectors are shown for

reference. For the helium injection case, these plots

were generated using the concentration data

combined with 109 experimental data points taken

with the cone-static probe and 186 data points

taken with the five-hole probe according to the data

analysis procedure described in Section II-G. For

the air injection case, the contour plots were

generated using 176 experimental data points taken

with the five-hole probe.

Figure 6: Contour plot of helium concentration at a plane 1.8 duct diameters downstream of the

circular injector centers

Parameter Unit Value

y* [mm] 16.9

w* [mm] 24

A* [mm2] 408

αmax [-] 0.101

yα,max [mm] 16

Table 2: Strut injector mixing parameters (for an entire plume created by 1 strut with 4 injectors)

Rock 8

Helium Injection Air Injection

Figure 7: Contour plots of Mach number for helium injection vs. air injection at a plane 1.8 duct

diameters downstream of the circular injector centers

Helium Injection Air Injection

Figure 8: Contour plots of total pressure for helium injection vs. air injection at a plane 1.8 duct

diameters downstream of the circular injector centers

A complex shock system forms downstream of the

injector array, which includes oblique shocks from

the struts and bow shocks from the injectors. This

phenomenon reduces the Mach number of the flow

downstream of the strut as shown in Figure 7. Also,

the injectant jets are at sonic conditions, whereas

the freestream is nominally at Mach 4 conditions.

The region of reduced Mach number in the vicinity

of the fuel plume is substantially larger for the

helium injection case compared to the air injection

case. This is due to the properties of the gases such

as molecular weight and specific heat ratio. These

properties influence the sound speed of a gas,

therefore helium has a higher sound speed than air

at the same temperature conditions. Mach number

is inversely proportional to sound speed, so the

Mach number in the region of the fuel plume is

lower for the helium injection case. Both cases

showed good penetration of the injectant across the

combustor cross section, but there is a substantial

total pressure loss downstream of the strut as

shown in Figure 8. The total pressure loss is larger

for the helium injection case than for the air

injection case, which is largely due to Mach

number effects. Another factor that contributes to

the lower total pressure for the helium injection

case is the total pressure in the injector manifold,

Rock 9

which is lower for the helium injection case than

for the air injection case (see Table 1).

IV. Application of Experimental Research to CFD Turbulence Model Upgrades

The experiments presented here are part of an

integrated experimental and computational study

being conducted by a team of researchers from

Virginia Tech and a small business, CRAFT Tech

(Combustion Research and Flow Technology,

Inc.). The experimental research is being conducted

at Virginia Tech, whereas the computational

research is being conducted at CRAFT Tech. The

primary goal of this study is to upgrade the

turbulence models that are used for CFD

predictions of the flow inside a scramjet

combustor. There are two primary upgrades that

are currently being developed for the k-ε turbulence

model: (1) scalar fluctuation modeling and (2) a

baroclinic turbulence source term correction. Scalar

fluctuation modeling predicts local variations in

turbulent Prandtl and Schmidt numbers. CFD

predictions for the flow inside a scramjet

combustor currently generally use turbulence

models that utilize global estimates for the

turbulent Prandtl and Schmidt numbers. In this type

of flow, the local values of turbulent Prandtl and

Schmidt numbers can vary significantly and this is

believed to be a source of error. The baroclinic

torque term is a modification to the turbulence

model to account for the effects of strong density

gradients, which occur in high-speed mixing flows.

In high-speed mixing flows, the flow conditions

deviate significantly from those used in the

standard k-ε model derivation. Therefore,

refinement of the turbulence model is necessary to

accurately predict these complex mixing flows.

To support the turbulence model upgrades,

experiments are required that parametrically vary

the jet molecular weight and freestream Mach

number, while keeping other basic mixing

parameters the same. The experiments provide a

database of fuel injection and mixing data that is

being used for turbulence model refinement and

CFD code validation. Figure 9 shows an example

of how the experimental data is being used to

upgrade the turbulence modeling. In this figure, the

experimental helium concentration results along

the centerline of the strut injector are compared to

CFD predictions with and without the additional

baroclinic torque term. The baroclinic torque term

was tested and validated for a wide range of

experimental cases consisting of different flush-

wall and in-stream injector designs, varying

molecular weight fuels, and varying freestream

Mach numbers7. From examining Figure 9, it is

evident that the baroclinic torque term significantly

improves the CFD predictions.

Figure 9: Comparison of the mixing results predicted by the CRUNCH CFD code to the

experimental data along the centerline of the strut for the helium injection case

V. Conclusion

This paper presented the results of an experimental

study of a four-strut injector configuration with

multiple round, sonic nozzles on each strut in a 100

mm diameter circular duct under cold-flow

conditions for application to high-Mach-number

scramjets with circular combustion chambers. The

freestream Mach number was nominally 4, which

simulated the conditions a scramjet combustor

would encounter in nominal Mach 10 flight. The

primary goal of this study is the refinement of

turbulence models for these complex mixing flows.

Injectant molecular weight has been identified as a

parameter of critical importance in the

development of the turbulence model upgrades and

the use of two different injectants, helium and air,

Rock 10

allowed the effects of injectant molecular weight to

be studied. For comparison purposes, a constant

jet-to-freestream momentum flux ratio was

maintained to achieve a similar amount of fuel

plume penetration for the two experimental cases.

The main reason for considering an intrusive

injector design for application in such a

challenging thermal environment as a Mach 10

scramjet combustor is the goal of minimizing the

combustor length in the low aspect ratio

combustors currently being considered for such

applications. An in-stream injector inherently

yields better penetration and airstream coverage in

short axial distances than a flush-wall injector. If

adequate mixing can also be achieved in a short

distance, then the drag and thermal load penalties

of the in-stream injector can be overcome.

For the case with helium injection, the use of

helium safely simulates hydrogen fuel in a non-

combusting environment. The experimental results

for the helium injection case obtained at

approximately 1.8 duct diameters downstream

showed good penetration of the injectant across the

combustor cross-section, but the individual jets had

merged into a single large plume and the rate of

mixing was somewhat slow. In addition, a

substantial total pressure loss occurred in the flow

downstream of the strut. There were substantial

regions of fuel-rich (using the stoichiometric ratio

of hydrogen to air, 0.0292, as a metric)

concentrations in the plumes, even though the

overall helium-air mixture in the duct was lean, on

the same basis. The maximum concentration of

helium detected by the experiment was 10.1%. One

might have expected better mixing based on

simple, isolated injector correlations. The ratio of

the measured injectant mass flow to the calculated

air flow (equivalent fuel/air ratio) in the duct was

0.0098 for the helium injection case, which

corresponds to an equivalence ratio of 0.34 in

terms of hydrogen. The area of the helium plume

created by one strut with four circular injectors was

calculated to be 408 mm2 based on the

experimental data. Thus, the overall helium plume

area in the duct covered by a total of 16 jets

distributed across four struts is 1632 mm2, which

accounts for 20.8% of the total duct cross section.

The air injection case exhibited similar features

to the helium injection case including good

penetration of the injectant across the combustor

cross-section, the individual jets merging into a

single large plume, and substantial total pressure

loss in the flow downstream of the strut. However,

the helium injection case had a larger overall total

pressure loss than the air injection case. The

equivalent fuel/air ratio for the air injection case

was 0.0273 in comparison to 0.0098 for the helium

injection case. The increase in the fuel/air ratio for

the helium injection case vs. the air injection case

is similar to the fuel/air ratio increase that would

occur for a scramjet engine operating on a low

molecular weight fuel vs. a high molecular weight

fuel.

The experiments presented here are part of an

integrated experimental and computational study

that is being conducted to improve the turbulence

models that are used for CFD predictions of the

flow inside a scramjet combustor. A database of

fuel injection and mixing data is being built for

turbulence model refinement and CFD code

validation. Certain improvements to the turbulence

modeling have already been achieved as part of this

integrated experimental and computational study.

Nevertheless, it will likely be possible to further

improve the turbulence modeling by obtaining

additional experimental data.

References

1. Maddalena, L., Campioli, T.L., Schetz, J.A., “Experimental and Computational Investigation of an

Aeroramp Injector in a Mach Four Cross Flow,” AIAA/CIRA 13th International Space Planes and

Hypersonics Systems and Technologies, AIAA 2005-3235, June 2005.

2. Schetz, J.A., Thomas, R.H., and Billig, F.S., “Mixing of Transverse Jets and Wall Jets in Supersonic

Flow,” IUTAM Symposium on Separated Flows and Jets, Novosibirsk, July 1990.

3. Kutschenreuter, P., “Supersonic Flow Combustors,” in Scramjet Propulsion, (E.T. Curran and S.N.B.

Murthy, Editors), AIAA, New York, 2000.

4. Ng, W.F., Kwok, F.T., and Ninnemann, T.A., “A Concentration Probe for the Study of Mixing in

Supersonic Shear Flows,” AIAA paper 89-2459, July 1989.

5. Swalley, F.E., “Measurement of Flow Angularity at Supersonic and Hypersonic Speeds with the Use

of a Conical Probe”, NACA TN D-959, September 1961.

6. Centolanzi, F.J., “Characteristics of a 40 degree Cone for Measuring Mach Number, Total Pressure,

and Flow Angles at Supersonic Speeds”, NACA TN 3967, May 1957.

7. Ungewitter, R.J., Brinckman, K., and Dash, S.M., “Advanced Modeling of New Fuel/Air Mixing Data

Sets for Scramjet Applications,” 45th AIAA/ASME/SAE/ASEE Joint Propulsion Conference &

Exhibit, Denver, CO, AIAA-2009-4940, August 2009.