Initial Thesis Report 2009, ACME, UNSW@ADFA

1
Numerical investigation of Fluidic Injection as a means of
thrust control
Brendan A. Blake
1

Interest in fluidic thrust vectoring has promoted many numerical and experimental
investigations to explore its validity in real world applications. The benefits of such
technology are numerous making investigations into their effectiveness a valid and
worthwhile exercise. A comprehensive literature review of previous work was conducted
from which an appropriate numerical simulation criterion was devised. This was carried
out to replicate experimental results and in-turn validate two dimensional computational
fluid dynamic simulations using FLUENT

as an appropriate method of investigation.

Specifically the method of fluidic thrust vectoring using the technique of shock vector
control will be used. Numerical simulations conducted aim to replicate previous
experimental work by Stagg (2008), using the Australian Defence Force Academy fluidic
thrust vectoring experimental apparatus, and Neely et al (2007). Following initial CFD
validations a numerical exploration of fluidic thrust spoiling as a means of thrust
magnitude control in solid fuel rocket motors will be conducted. Results obtained from
computational fluid dynamics simulations will be then used to develop a suitable control
system for both thrust magnitude and direction will be carried out.
Nomenclature

Abbreviations Symbols
FTV : fluidic thrust vectoring q : thrust vector efficiency (/% inject flow)
CFD : computational fluid dynamics : density (kg/m
3
)
2D : two dimensional p
0
: inlet pressure (kPa)
MTV : mechanical thrust vectoring C
fg,sys
: system thrust ratio
SVC : shock vector control o : thrust vector angle ()
TS

: throat shifting a : speed of sound (m/s)
NPR : nozzle pressure ratio M : Mach number
SPR : secondary pressure ratio p : ambient pressure (kPa)
MFR : mass flow ratio (secondary flow)
PID : proportionate differential integral
2DCD : 2 dimensional converging diverging
MATV : multi-axis thrust-vectoring
HARV : High Angle of Attack Research Vehicle
LaRC : NASA Langley Research Center
I. Background and development of thrust vectoring technologies.
ECHNOLOGY has a very mutual relationship with performance. Aircraft performance can be measured in
a number of ways from the range a plane can travel, to the amount of payload it can carry and the amount of
inherent maneuverability it has. Thrust vectoring has particular relevance for the latter in that it is a maneuver
effector which expands the flight envelope beyond that of aircraft using conventional engine nozzles. Using
thrust vectoring extends the controllability of aircraft at post stall high angles of attack beyond the point where
conventional flight control surfaces lose their effectiveness and in turn provides super maneuverability (Mason
& Crowther 2002). Increases in landing and takeoff performance characteristics are also significantly improved
(Neely et al 2007). The quest for increased aircraft performance has also been motivated by weigh reduction
which is why fluidic thrust vectoring (FTV) is of significant interest for aviation applications. Using fluidic
techniques for thrust vectoring can lead to weight reductions of as much as 80% over their mechanical
counterparts (Deere 2003). Indeed 30% of the F-22 Raptor engine weight is attributed to the mechanical thrust

1
Thesis Project and Work Experience, ZACM4050

T

Initial Thesis Report 2009, ACME, UNSW@ADFA

2
vectoring mechanism utilized to achieve its specific flight performance (Chalmers 1999). The foreseeable
associated weight reduction and performance potential provides significant justification for the future
development and investigation of FTV as a means of thrust control. Despite FTV being investigated since the
1950s when first utilized on the Polaris missile (Wing & Giuliano 1997) much work is still to be done both
experimentally and numerically to first select the most appropriate vectoring technique and then physically
optimize it for real word applications.
A. Mechanical Thrust Vectoring Overview
Mechanical thrust vectoring (MTV) typically uses one of two means to achieve a changing thrust direction.
Firstly paddles, vanes or spoilers can be used within the jet stream of the main flow to physically alter the
direction of generated trust. This method of mechanical thrust vectoring is used on NASAs highly modified
F/A-18 High Angle of Attack Research Vehicle (HARV) and is also incorporated into the Pratt and Whitney
F119-PW-100 engine as used in the F-22 Raptor fighter aircraft (Kowal 2002). Mechanical paddles have also
been used on rockets though out history from the World War Two V-2 cruise missile (URL 5)to the modern day
Nulka active missile decoy system in use with the US, Canadian and Australian Navies (URL 4).


Figure 1: F/A-18 HARV Nozzles
(URL 1).
Figure 2: Pratt & Whitney F119-PW-100 as found on F-22 Raptor (URL 2).


Secondly, adjustable nozzles can be used as a means of mechanically ducting airflow and as such thrust to
alter its direction. The nozzle on the Industria de Turbo Propulsores (ITP) of Spains EJ200 engine as found on
the Euro fighter Typhoon has such a nozzle.

Figure 3: CAD of ducted nozzle from
EJ200 Euro fighter Typhoon (URL 3).
Figure 4: EJ200 Showing vectored thrust (URL 3).

Both of these options while effective have their significant shortcomings. Mechanical thrust vectoring is
heavy due to the actuators and other hardware necessary to achieve noteworthy thrust angle deflection. The
complexity associated with mechanical vectoring also adds to the maintenance requirements of the aircraft, can
increase the nozzles radar signature reducing survivability (Kenrik & Deere 2003) and can be structurally
difficult to integrate (Mason & Crowther 2002).
B. Introduction to Fluidic Thrust Vectoring
The alternate to mechanical thrust vectoring is to vector thrust fluidically. FTV is the directional alteration of
a main thrust generating flow by the use of a secondary flow. This flow can be used in a number of different
ways to alter the direction of thrust by up to 18 degrees (Wing & Giuliano 1997). FTV has numerous
advantages over mechanical vectoring. It is lighter, less complex, economical, easier to integrate and has

Initial Thesis Report 2009, ACME, UNSW@ADFA

3
potential to reduce the aircraft radar signature which in turn increases survivability (Mason & Crowther 2002).
It is for these reasons that fluidic vectoring is considered superior to MTV however it is not without its short
comings. FTV has less directional changing capability when compared to mechanical vectoring. It is considered
untried and untested in real world life size applications generating some skepticism of its usefulness. Fluidic
vectoring also requires a source of secondary flow and from where this originates may reduce performance if
tapped from the main thrust generating flow. Regardless of these apparent imperfect consequences of its use,
FTV still seems an appealing option when attempting to improve performance in ways previously listed. Later
investigations to be included in this thesis will include an assessment on the potential of fluidic injection for
throttling solid fuel rocket motors. The fact that current use of FTV is not employed in any operating aircraft is
indicative of the need to conduct more research into performance and potentially diversifying its applications.
II. Types of Fluidic Thrust Vectoring
A variety of different methods of FTV have been developed since initial investigations of fluidic vectoring
techniques began in the 1950s. Each method has been investigated both experimentally and numerically with
differing levels of success. All methods are fairly unique in the manner in which secondary flow is used to
vector thrust.
A. Co-flow and Counterflow Fluidic Thrust Vectoring
Co-flow thrust vectoring relies on the Coanda effect to generate
a differential thrust angle. The Coanda effect is the tendency of a
flow to adhere to a curved surface due to the low pressure
generated at the surface as the flow velocity over it increases
(Mason & Crowther 2002). This low pressure causes not only the
injected co-flow to shift off the normal thrust axis toward the wall
but also the primary flow. This phenomenon was named after the
Romanian Henri-Marie Coanda in 1930 who was the first
aerodynamicist to appreciate the importance of this effect for
aircraft applications. (Banazadeh & Saghafi 2008).

Counter Flow FTV was first reported by Strykowski and
Krothapali (Flamm 1998) and involves applying a vacuum to a
slot shrouded by a suction collar adjacent to the primary flow
source. A reverse flow layer is established at the wall of the
suction collar and as in the co-flow situation the Coanda effect
impacts on the main flow. If co-flow is activated asymmetrically
the main flow is shifted towards the side on which the reversed
flow is applied due to the lower pressure causing vectored thrust.


B. Throat Shifting
Throat shifting (TS) uses the injection of secondary flow at or just upstream to the throat of a converging
diverging nozzle. Injecting the fluid here prevents the reactive formation of shockwaves and as such the flow is
effectively isentropic (Stagg 2008). This method of thrust vectoring alters or skews the shape of the post throat
shock which in turn moves the thrust line of action. This is achieved prior to the flow going supersonic and
hence thrust performance is not significantly affected by the vectored flow slowing as it travels over a shock.
Figure 7 below shows how the thrust axis is moved using TS as used in the 2005 undergraduate thesis by
Chittleborough. An undergraduate thesis is currently being done by Lim at ADFA on TS compromising of
numerical investigations. Work conducted at LaRC by Deere (2003) suggests that TS is one of the most
promising forms of FTV because of its thrust efficiency.

Figure 7: Throat Shifting FTV (Kowal 2002).
Figure 5: Co-flow FTV (Banazadeh &
Saghafi, 2008).
Figure 6: Counter Flow FTV (Flamm 1998).

Initial Thesis Report 2009, ACME, UNSW@ADFA

4
C. Shock Vector Control
Shock vector control (SVC) uses the injection of secondary flow
downstream of the throat in the diverging section of the nozzle.
Injection of flow here induces the formation of a shock wave at
some angle. The injected flow acts like a virtual pressure ramp
when seen by the main flow. As the main thrust producing flow
passes through this shock it turns away from the injection port and
by doing such vectors the force produced. Maximum vector angle is
achieved just prior to the shock impinging on the opposite side of
the nozzle. If the shock does impinge on the opposite side of the
nozzle the shock is reflected and when the flow passes through the
reflected shock it is turned in the direction of the unaffected flow.
Knowing this enables optimization of pressure or volumetric flow rate ratios. It is obvious that there are two
things that can alter the point at which the main shock first impinges on the nozzle, the specified flow out of the
injector, and the position of the injector relative to the nozzle exit.
D. Combined methods
Several attempts of FTV have been made using a combination of
techniques with differing levels of success. It is possible to
accentuate and compliment one method by either partially or fully
incorporating a second method in order to obtain the greatest
performance possible. Berrier et al (2003) used recessed cavities to
complement the throat shifting method as shown in figure 9.
Variable recessed cavities enable thrust vectoring through the
establishment of vortices within the cavities themselves. Although
the presence of recessed cavities does cause thrust vectoring it is the
combination of injected secondary flow at the throat in combination
with the recessed cavities that causes vectoring. Although achieving
significant results combined methods are not widely investigated
due to their inefficiency and difficulty to incorporate (Berrier et al
2003).
E. Comparison of Methods
Given the advantages of FTV it is a worthwhile exercise to discuss the advantages and disadvantages of each
different method to assess the method most appropriate for incorporation in real world applications. Co-flow
and counter flow methods have by some been assessed to be the most suited to practical use (Deere 2003).
Although the performance of these nozzle are some of the best reported they do have their disadvantages. The
physical complexity of co-flow nozzles can make them difficult to incorporate into the surrounding structure.
Counter flow techniques require an adequate supply of suction which may add both complexity and weight to
the system (Hunter & Deere 1999). The TS method is relatively simple in its design and apparent ease of
integration makes this method of vectoring appealing however vector angle limitations exist that make it suited
to applications only requiring moderate maneuver adjustments (Berrier et al 2003). SVC methods also are
appealing upon first inspection due to their achievable thrust angles however this comes at a cost. Thrust ratio is
reduced by between 6-14% depending on the operational configuration including values of nozzle pressure ratio
(NPR) and secondary pressure ratio (SPR). Weighing up the advantages of each method suggests further
investigation into the performance of SVC is an appropriate and valued exercise.

Method Advantages and Disadvantages
Throat Shifting Method - Most thrust efficient method
- Throat area control possible
Counterflow Method - Requires a suction source
- Jet attachment problems
Coflow Method - Least thrust efficient method
- Control reversal problems at low injection mass flows.
Shock Vector Control
Method
- Simple geometry
- Largest deflection angle (excluding pulsed)
- Shock impingement problems
Table 1: Comparison of FTV techniques (adapted from Chittleborough 2005)

Figure 8: SVC FTV Nozzle (Deere 2003).
Figure 9: Combined TS/Recessed Cavity
FTV Nozzle (Berrier et al 2003).

Initial Thesis Report 2009, ACME, UNSW@ADFA

5
III. Background theory
The method of FTV using SVC relies on the inherent physics involved in both converging diverging nozzles
and supersonic flow. Thrust vectoring is achieved within these nozzles by passing the flow through an oblique
shock which is generated in the divergent section of the subject nozzle.
A. Nozzle theory


Figure 10: Standard Convergent/Divergent Nozzle Nomenclature (Anderson 2007)
To gain an understanding of how supersonic flow is established in a converging diverging nozzle it is
necessary to examine the driving relationship, namely the pressure ratio. The ratio of pressures is
p
1
p
2
or using
the above nomenclature which is
p
0
p
c
, the pressure upstream divided by the exhaust pressure which can be found
using the following:
1
2
2
1
2
1
1

|
.
|

\
|
+ =

M
p
p
(1)

With gamma (y) being the ratio of specific heats (
C
p
C

) and M the Mach number. If the upstream specific density

and temperature are know the ideal gas equation can be used to solve for p
1
.

RT p =
(2)
Where R is the gas constant, (R
air
=287.05Nm/kg.K).
Knowing the pressure ratio, the temperature (T) and density () ratio can be found using isentropic
relationships.
1
2
1
2
1
1
2
T
T
p
p
= =
|
|
.
|

\
|

(3)
If the pressure ratio is sufficiently great enough, at the critical pressure ratio,
|
|
.
|

\
|
*
0
p
p
where p* is the throat static
pressure and
0
p the inlet static pressure, the local flow at the nozzle throat will be exactly Mach 1 and the
nozzle is said to have choked. The critical pressure ratio can be found using the pressure ratio relationship of
equation (1) using a Mach number of 1 and the ratio of specific heat for air, 1.4.
893 . 1 ) 1 (
2
1 4 . 1
1
*
1 4 . 1
4 . 1
2 0
=
|
.
|

\
|
+ =

p
p

2: Equations from Anderson 2007 and Anderson 2003
If the pressure ratio is further increased beyond the critical pressure ratio the velocity at the throat remains
constant at Mach 1 and hence mass flow rate is fixed.
B. Supersonic Compressible Shock Theory
The primary mechanism for thrust vectoring using the SVC method is the presence of an oblique shock.
From a conventional converging diverging nozzle there is a normal shock that propagates from the throat once
the critical pressure ratio is reached as discussed above. Applying a secondary injection of fluid from a slot in
the divergent nozzle wall section an oblique shock is established immediately upstream of where the secondary

Initial Thesis Report 2009, ACME, UNSW@ADFA

6
flow is injected as if there is a physical compression ramp within the nozzle. As flow passes through this shock
there is an instantaneous change in pressure, temperature, density and velocity all of which increase accept the
latter, Mach number. It is the reduction in velocity over the shock (Anderson 2003) that adversely affects the
thrust efficiency when using SVC as a means of FTV (Deere & Hunter 1999). When using SVC the thrust is
vectored in the direction of the shock due to a change in flow direction as it passes through the induced oblique
shock. Supersonic compressible flow behavior is a function of both the Mach number and hence the local speed
of sound using these relationships.

The local speed of sound (a) is a function of the ratio of specific heats, the ideal gas constant and the
temperature of the gas by the following.

RT a =
(4)


For an oblique shock the appropriate numerical relationships for ramp/shock angle, density ratio and
pressure ratio are:

( ) 2 ) 2 cos(
1 sin
tan
2
tan
2
1
2 2
1
+ +

=

M
M
(5)
( )
( )

2 2
1
2 2
1
2
1
1
2
sin 1 2
sin 1
M
M
u
u
+
+
= =
(6)
( ) 1 sin
1
2
1
2 2
1
1
2

+
+ =

M
p
p
(7)
1: Equations from Anderson 2003)
In equations 5 to 7 above, theta () is the angle
of the compression ramp that corresponds to the
flow directional change and beta () is the angle of
the induced oblique shock as per figure 11, diagram
of a 10 degree wedge in a Mach 2 flow inducing an
oblique shock of nearly 40 degrees.

Later investigations into the suitability of 2D
symmetrical fluidic injection for throttle control will
be carried out at which time normal shock theory
will be included. This is important as the pressure
loss over the shock will have a large degree of influence on how effective throttle using fluidic injection will be.
IV. Summary of previous thrust vectoring research
There has been considerable research done in the field of FTV and there are several main research facilities
that specializes in the field of fluidics such as the NASA Langley Research Centre (LaRC) in Hampton
Virginia, the United States Air Force (USAF) research laboratory as well as various industry and academic
partners globally. The collaboration of all results from these research venues compromise a considerable
database of fluidic techniques that enable further development for practical applications. The ground work for
FTV began in the 1950s when secondary injection systems were investigated for incorporation into rocket
propulsion systems (Wing & Giuliano 1997). Later, during the 1960s, gimbal and swivel mechanisms attached
to the tailpipe of jet engines were investigated for their suitability for mechanical thrust vectoring. Also during
this decade liquid injection thrust vector control (LITVC) was developed for solid fuel rockets (Anon 1974).
Since then further refinement and investigation of both mechanical and fluidic vectoring technologies has been
carried out. Throughout the 1970s non-axisymmetric nozzle research was carried out on 2DCD nozzles, wedge
and single expansion ramp nozzles (SERN) for their ability to vector thrust (Deere 2000). These nozzles offered
performance benefits with regard to integration and installation above axisymmetric nozzles previously
investigated. In the 1980s research into mechanical multi-axis thrust vectoring (MATV) was carried out for
pitch and yaw control for super maneuverability and for the first time its potential for post stall controllability to
further expand current flight envelope limitations. It was discovered that sub-sonic vectoring techniques obtain
Figure 11: Oblique Shock Nomenclature
(Anderson 2007)

Initial Thesis Report 2009, ACME, UNSW@ADFA

7
the best effectiveness and thrust efficiency and that performance was independent of both sideslip angle and
angle of attack (Deere 2000).

Initial Fluidic injection research at LaRC began in an attempt to reduce nozzle weights and the radar
signature of low observable aircraft during the 1990s and has continued since (Deere 2003). Activities at
LaRC have included experimental work using the Jet Exit Test Facility (JETF) on a variety of different nozzle
configurations as well as numerical simulations conducted using software developed in house called PAB3D.
PAB3D is a structured, multiblock, parallel, implicit, finite volume solver of the three dimensional reynolds
averaged navier stokes (RANS) equations, and advanced turbulence models are available in the code. With the
vast amount of experimental and numerical simulation carried out at the LaRC and making comparisons
between them, PAB3D has proven its self well in the replication of high Reynolds number shear flows (Abdol-
Hamid et al 2006).

FTV research at ADFA dates back to 2004 with a paper from ORourke titled Fluidic Thrust Vectoring.
Since then several works have been completed by both academic staff and students. FTV research at ADFA has
focused on the method of SVC following the construction of a test rig in 2005 by Chia. Previous in depth
numerical and computational investigations have been carried out using the commercially available CFD
program CFD++by Chittleborough (2005) and with FLUENTby Neely, Gesto and Young (2007). Stagg
carried out experimental work using the test rig after modifications he carried out in 2008.
V. Aim and Scope
The potential exists to use FTV for not only for thrust vectoring but also throttle control. This would be of
significant benefit especially for solid fueled rocket motors which are inherently difficult to throttle through
conventional means. Solid fuel rockets have fixed or very limited variability in combustion rate of the propellant
material which makes thrust variation much more difficult than their liquid fueled counter parts (Sutton &
Biblarz, 2001). The practical applications of a system that not only controls thrust direction but also magnitude
would therefore be of considerable benefit for missile applications such as the Nulka hovering missile decoy
system.
A. Aim
Initially it is hoped that FLUENT

will be validated as an appropriate CFD program for obtaining a

satisfactory solution to previously experimental work on FTV using the method of SVC. Once this assessment
has been made simulations will be carried out on a modified nozzle geometry that allows the symmetric
injection of secondary flow from both sides of the divergent nozzle duct. Various symmetrical SPRs will be
investigated for their ability to throttle thrust. Later simulations will include the injection of unsymmetrical
SPRs to investigate the potential for both vector and throttle of the thrust produced. Once all CFD simulations
have been completed data will be used for the design of a control system potentially using proportionate,
integral, differential (PID) controllers. Annex E formalizes the aims associated with this report in the way of a
client brief. Progress is to be tracked through the use of the Milestone Chart in Annex F.
B. Scope
Many numerical research projects have been done before on FTV using SVC however most have been done
using software developed in house by organizations such as NASA (PAB3D) and various aircraft
manufacturers. It would be of significant use to ensure that commercially available software such as FLUENT

has the capability to simulate SVC in 2DCD nozzles. Having the results of previous experimental work the
numerical solutions obtained will be compared and assessed as valid or otherwise. Once the validation of the
CFD solution has been achieved further research can be conducted on symmetrical injection for throttle control
using FLUENT

with the aim to investigate the suitability of fluidic injection for thrust magnitude control.
Later a basic control system will be designed using proven software such as MATLAB SIMULINK.
VI. Development of CFD Solutions
The selection and development of computational solutions has been driven by the availability of both
hardware and software. For this reason preprocessing including geometric modeling and meshing has been
carried out using GAMBIT

and all flow simulations have been conducted using FLUENT

. The requirement
may develop to later use T-Grid for further preprocessing and Tech Plot for further post processing.
A. Introduction to CFD and FLUENT

FLUENT

is a commercially available CFD program able to accurately simulate fluid flow by solving the
appropriate conservation equations in 2D or 3D. Conservation partial differential equations of mass, momentum

Initial Thesis Report 2009, ACME, UNSW@ADFA

8
and energy are solved simultaneously for each discretised section of the specified grid (Anderson 1995). This
highlights the importance of correct gridding as an insufficiently defined grid will not enable the accurate
calculation and simulation of flow interaction with its self or other physical objects such as walls.

To monitor the accuracy of the solution it is necessary to observe the difference between the left and right
hand side of the partial differential conservation equations of mass, momentum and energy. The difference
between them is known as the underlying residual of the solution and it is necessary to get this as low as
possible in order to obtain a solution to sufficient accuracy. Typically solutions with residuals below 1e
-3
are
sufficiently accurate.
B. Solver selection using FLUENT

The solver selection is rather important when using any CFD as it dictates to a degree the accuracy of the
solution obtained at the rate at which the solution will converge. For the simulations carried out thus far a
density based explicit solver has been used with reasonable success. Due to the complexity of the flow and the
fact that it does not align well with any possible grid structure it is insufficient to used a first order discretisation
scheme as the convective discretisation increases the numerical discretisation error and as such an erroneous
solution or divergence. For this reason all useable solutions have and will be obtained using a second order
discretisation scheme. There are several different turbulence models available within FLUENT

including the
Spalart-Allmaras model e models, and models. As the turbulence model obtained has a dramatic
impact on the rate and level of convergence (Hamed & Vogiatzis 1997) it was deemed appropriate to begin with
the simplest and least computationally demanding model available, namely the one equation Spalart-Allmaras
model. Although the e model has been assessed to suit fluidic injection problems well and is widely used
(Waithe & Deere 2003) resource limitations have prevented its use to date with the option of later exploration of
different turbulence models.
C. Geometry and justification of 2D solution
Initial geometry was constructed to replicate as close as practical the ADFA FTV experimental test rig using
GAMBIT

geometry and gridding software. Physically this apparatus has a throat width of 3 mm, a maximum
converging opening width and length of 13mm and 15mm, a maximum divergence width and length of 9 mm
and 14mm at the nozzle exit. From the exhaust of the 2DCD nozzle, which is boarded on each side with a solid
wall, the flow exhausts into a simulated atmosphere tiered in two steps. The secondary flow is injected through
a 1 mm opening in the upper wall of the diverging section of the nozzle. Above this span wise slot a plenum
was created to facilitate the injection of the secondary flow. Initial attempts to obtain a converged solution were
hindered even when using a first order discretisation scheme due to geometrically induced x and y axis velocity
residuals. This was rectified by altering the shape of the slot to remove a sharp corner at its joint with the
plenum. With the current geometry there also seems to be a concentration of velocity residuals at the corner of
the nozzle exit however these are within acceptable limits to obtain convergence. A picture of the used domain
and grid is included in Annex A.
VII. Preliminary Results
Initial solutions were obtained using a 1
st
order discretisation scheme and later convergence was achieved
using 2
nd
order schemes. The difference between to two is considerable which further emphasizes the
importance to use the appropriate discretisation scheme when attempting to model complex flows.
A. 1
st
Order oversized solution
After solving some initial geometry issues which caused
elevated residuals it was possible to obtain a converged 1
st
order
solution after some 2100 iterations at a SPR of 3.33. Initial
solutions were obtained using an oversized model and as such this
data is useful only for comparison. Residuals below 10e
-4
were
obtained when the previously described solver was used. Using
the figures directly reported form FLUENT

for the resolved X

and Y forces on the nozzles upper and lower walls the below
vector angle was achieved and figure 12 shows the obtained Mach
number contours. Full domain images are included in Annex B.
Total reported forces:
F
x
=20403793
F
y
=268580.7
0 =sin
-1
P
j
P
x
=sin
-1
268580.7
20403793
=0.75
Figure 12: 1st Order Mach Contours

Initial Thesis Report 2009, ACME, UNSW@ADFA

9
B. 2
nd
Order solutions
There was considerable trouble encountered when
trying to obtain a converged second order solution to
the same level of residual as the first order solution.
Residuals would reduce and then stabilize at below 10e
-
3
however they would not reduce further. After several
dozen runs and initiating monitors for mass flow into
and out of the nozzle as well as pressure fluxes on the
top and bottom of the nozzle walls (which were found
to be the same) it was clear that the solution had
stabilized despite residuals not reaching the default
definition for convergence of 10e
-4
. Figure 13 is nearly
identical to physical schlerian images obtained by
Neely, Gesto and Young from 2007 using the same SPR
(Figures 14 & 15). Large images for a more detailed comparison are included in Annex D.


Oversized Correctly scaled
F
x
(N) 20163993 20178.6
F
y
-572015.72 548.03
0 =sin
-1
F

F
x
sin
-1
572015.72
20163993
sin
-1
548.03
20178.6

0 =

1.62 1.55
Table 3: Comparison of oversized and correctly scaled model vector angle.
Interestingly there is minimal difference of 0.07 degrees or 4% in the vector angle between the oversized and
correctly scaled 2
nd
order solutions. Comparing the two, correctly and incorrectly scaled models, it seems to
indicate good similarity in shock geometry and Mach number magnitude. This is not clearly shown due to the
default magnitude scaling done by FLUENT

which is one area where Tech Plot may be used later. This
however is an accidental discovery and was not initially intended within the scope of this thesis but is note
worthy should a similar situation such as this occur in the future.



Figure 15: Schlieran from Neely et al 2007.
VIII. Conclusions and Summary of Future Work
From initial and very preliminary results it is apparent that FLUENT

is capable of developing acceptable

solutions to simulations of FTV using the method of SVC. This will enable the program to be used for future
simulations utilizing secondary injection for thrust control both in magnitude and direction. Convergence was
achievable however not to the default level specified by the program. This was not an issue as the solution
residuals and mass flow into and out of the nozzle had stabilized as recorded on various monitors. The ability to
carry out these simulations accurately will potentially reduce the amount of experimental work required when
physically exploring the potential of fluidic injection for throttle control and thrust vectoring. These simulations
will also aid in the design of the basic control elements required for a system such as this.

Figure 13: Oversized 2nd order Mach contours.
Figure 14: Scaled 2nd order solution of Mach Contours.

Initial Thesis Report 2009, ACME, UNSW@ADFA

10
References
Anon, 1974, Solid rocket thrust vector control, NASA SP-8114.
Abdol-Hamid, KS, Pao, SP, Hunter, CA, Deere, KA, Massey, SJ, and Elmiligui, A, 2006, PAB3D: Its
History in the Use of Turbulence Models in the Simulation of J et and Nozzle Flows, AIAA-2006-489, 44th
AIAA Aerospace Sciences Meeting, Reno, Nevada.
Anderson, JD Jnr 2007, Fundamentals of Aerodynamics, 4
th
ed., McGraw Hill.
Anderson, JD Jnr 1995, Computational Fluid Dynamics: The basics with applications, McGraw Hill.
Anderson, JD 2003, Modern compressible flows with historic perspective. 3
rd
ed., McGraw Hill.
Charmers, P 1999, Return Flight. ASME, http://www.memagazine.org/backissues/membersonly/dec99
/features /flight/flight.html, [accessed on 15th April 2009].
Chittleborough, S 2005, A Computational Investigation of the Shock Vector Control Method of Fluidic
Thrust Vectoring, BE Thesis, SACME, Canberra, UNSW@ADFA.
Deere, KA 2003, Summary of Fluidic Thrust Vectoring Research Conducted at NASA Langley Research
Centre, AIAA-2003-3800, 21st AIAA Applied Aerodynamics Conference.
Deere, KA 2000, Computational Investigation of the Aerodynamic Effects on Fluidic Thrust Vectoring,
AIAA-2003-3802, 36th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Huntsville, Alabama, July 17-19.
Deere, KA, and Hunter, CA 1999, Computational Investigation of Fluidic Counterflow Thrust Vectoring,
AIAA-99-2669, 35th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Los Angeles,
California, June 20-24.
Flamm, JD 1998, Experemental Study of a Nozzle Using Fluidic Counterflow for Thrust Vectoring,
AIAA-98-3255, 34
th
AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, Cleveland, OH, July 13-
15.
FLUENT, 2009, Version 6.3.26 ANSYS Canonsburg, PA United States.
GAMBIT, 2007, Version 2.4.6 ANSYS Canonsburg, PA United States.
Hamed, A, Vogiatzis C 1997, Overexpanded Two-Dimensional Convergent-Divergent Nozzle flow
simulations, Assessment of Turbulence Models, Journal of Propulsion and Power, Vol. 13, No. 3, May-June.
Mason, MS, Crowther, WJ 2002, Fluidic Thrust Vectoring of Low Observable Aircraft, CEAS Aerospace
Aerodynamics Research Conference. University of Manchester, School of Engineering, Cambridge, UK.
Neely, AJ, Gesto, FN and Young, J 2007, Performance Studies of Shock Vector Control Fluidic Thrust
Vectoring, AIAA-2007-5086, 43rd AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit
8 - 11 July 2007, Cincinnati, OH.
Stagg, AJ 2008, Design of a Parametric Nozzle for Fluidic Thrust Vectoring Experiments, BE Thesis,
SACME, Canberra, UNSW@ADFA.
Sutton, GP, and Biblarz, O 2001, Rocket Propulsion Elements, 7
th
Edition, John Wiley and Sons INC.
Brisbane, Queensland, Australia.
URL 1: http://www.dfrc.nasa.gov/Education/Educator/SupportMaterials/Careers/Images/harv.jpg [Accessed
31st March 2009]
URL 2: http://www.pratt-whitney.com/vgn-ext-templating/v/index.jsp?vgnextoid=
2c9058bf8a6eb010VgnVCM1000000881000aRCRD [Accessed 31st March 2009]
URL 3: http://typhoon.starstreak.net/Eurofighter/engines.html [Accessed 31st March 2009]
URL 4: Janes Radar and Electronic Warfare systems,
http://search.janes.com.ezproxy.libadfa.adfa.edu.au:2048/Search/documentView.do?docId=/content1/janesdata/
yb/jrew/jrew1010.htm@current&pageSelected=allJanes&keyword=nulka&backPath=http://search.janes.com.ez
proxy.libadfa.adfa.edu.au:2048/Search&Prod_Name=JREW& [Accessed 26
th
April 2009]
URL 5: http://search.janes.com.ezproxy.libadfa.adfa.edu.au:2048/Search/documentView.do?docId
=/content1/janesdata/binder/jsws/jsws0565.htm@current&pageSelected=allJanes&keyword=V-2&back
Path=http://search.janes.com.ezproxy.libadfa.adfa.edu.au:2048/Search&Prod_Name=JSWS& [Accessed 27
th

April 2009].
Waithe, KA, and Deere, KA 2003, Experimental and Computational Investigation of Multiple Injection
Ports in a Convergent-Divergent Nozzle for Fluidic Thrust Vectoring, AIAA-2003-3802, 21st AIAA Applied
Aerodynamics Conference, Orlando, Florida, June 23-26.
Wing, DJ, Giuliano, VJ 1997, Fluidic Thrust Vectoring of an Axisymmetric exhaust nozzle at static
conditions, ASME Fluids Engineering Division Summer Meeting, June 22 26.