See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/263808170

Flow Visualization of a Scramjet Inlet - Isolator Model in Supersonic Flow

Thesis · May 2012

CITATIONS READS

0 4,324

1 author:

Serdar Seckin
University of Texas at Austin
14 PUBLICATIONS   95 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Ses Üstü Hava Alıklarında Şok Dalgası Sınır Tabaka Etkileşiminin Deneysel İncelenmesi View project

All content following this page was uploaded by Serdar Seckin on 10 July 2014.

The user has requested enhancement of the downloaded file.

 ISTANBUL TECHNICAL UNIVERSITY

FACULTY OF AERONAUTICS AND ASTRONAUTICS

GRADUATION PROJECT

Major : Astronautical Engineering

Subject : Flow Visualization of a Scramjet Inlet -

Isolator Model in Supersonic Flow

Name Surname : Serdar Seçkin

Student ID : 110060132

Advisor : Asst. Prof. Dr. K. Bülent Yüceil

Due Date : 21/05/2012

TABLE OF CONTENTS

LIST OF FIGURES .......................................................................................................................................ii

LIST OF SYMBOLS.....................................................................................................................................iii
PREFACE .................................................................................................................................................. iv
SUMMARY ................................................................................................................................................v
1. INTRODUCTION ................................................................................................................................... 1
2. EXPERIMENTAL SETUP ......................................................................................................................... 5
2.1. WIND TUNNEL .............................................................................................................................. 5
2.2. SCHLIEREN SYSTEM .................................................................................................................... 10
2.3. MOUNTING APPARATUS SETUP AND MODIFICATIONS ............................................................. 11
2.4. INLET/ISOLATOR DESIGN AND CALCULATIONS .......................................................................... 15
3. RESULTS ............................................................................................................................................. 20
3.1. CFD ANALYSES ............................................................................................................................ 20
3.2. EXPERIMENTAL RESULTS ............................................................................................................ 28
4. CONCLUSION ..................................................................................................................................... 31
5. REFERENCES ...................................................................................................................................... 32

i
LIST OF FIGURES

Figure 1 Schematic of a scramjet [6] ....................................................................................................... 1

Figure 2 Scramjet components [3] .......................................................................................................... 2
Figure 3 Specific Impulse and Mach Number for Various Engine Types [5] ........................................... 3
Figure 4 Ramjet/Scramjet modes change from Mach 3 to Mach 8 [3] ................................................... 4
Figure 6 Supersonic wind tunnel test section and other components ................................................... 7
Figure 7 Axial Control Valve from Mokveld Company ............................................................................ 7
Figure 8 Nickel membrane ...................................................................................................................... 8
Figure 9 Test section module for Mach 2 and the designed inlet/isolator model assembly .................. 9
Figure 10 Schematic of schlieren system [7] ......................................................................................... 10
Figure 11 Schlieren system with mirrors [8] ......................................................................................... 11
Figure 12 Mounting apparatus and inlet/isolator model assembly ...................................................... 12
Figure 13 3D scanner from GOM Company .......................................................................................... 12
Figure 14 Scanned perforated plate ...................................................................................................... 13
Figure 15 Designed first mounting apparatus ....................................................................................... 13
Figure 16 Scanned and manufactured bottom cover of the wind tunnel ............................................. 14
Figure 17 Oblique Shock [9] .................................................................................................................. 15
Figure 18 Relation between θ,  and Mach number [9] ....................................................................... 16
Figure 19 Comparison between Mach numbers and  angles [9] ........................................................ 16
Figure 20 side view of the inlet/isolator model .................................................................................... 17
Figure 21 top view of the inlet/isolator model ..................................................................................... 18
Figure 22 Exploded view of the model assembly .................................................................................. 19
Figure 23 Inlet/isolator model assembly ............................................................................................... 19
Figure 24 Mach number distribution of the 2D inlet/isolator model ................................................... 20
Figure 25 CAD model of the fluid structure .......................................................................................... 21
Figure 26 Mesh structure from different aspects ................................................................................. 22
Figure 27 Shock waves and Mach number distribution from side view ............................................... 24
Figure 28 Boundary layer separation vectors ....................................................................................... 25
Figure 29 Mach number distribution from the top view ...................................................................... 25
Figure 30 Density distributions from the side view .............................................................................. 27
Figure 31 Average density distribution along z-direction ..................................................................... 28
Figure 32 Experimental result of the inlet/isolator model at Mach 2 ................................................... 29
Figure 33 Shock wave structure of the experimental result ................................................................. 29
Figure 34 Shock wave structure of the CFD results............................................................................... 30

ii
LIST OF SYMBOLS

 : Shock angle

 : Specific heat ratio

 : Mach wave angle

θ : Ramp angle

a : Speed of sound

Isp : Specific impulse

M : Mach number

M1 : Mach number before the shock

M2 : Mach number after the shock

Mn1 : Normal vector of M1 for the shock line

Mn2 : Normal vector of M2 for the shock line

P : Pressure

P0 : Total pressure

R : Gas constant

T : Temperature

T0 : Total temperature

iii
PREFACE

This thesis covers the basis of ramjet and scramjet inlet/isolator model design, boundary layer and
shock wave interactions and Schlieren flow visualization method. This graduation project is the first
step which carries the role of preparing the system for investigating inlet unstart conditions.

During my project, I am very grateful and thankful to my advisor, Asst. Prof. Dr. Bülent Yüceil, for
encouraging, guiding and supporting me.

I would like to like to thank Aliosman Tabanlı for his help during my experiments.

Also very special thanks to my family for always giving countenance to me.

iv
SUMMARY

This thesis is based on ramjet/scramjet inlet/isolator models and visualization of the shock waves
with Schlieren method. The supersonic wind tunnel is modified to be able to mount the designed
inlet/isolator model. After modification setup was completed, an inlet model was designed that had
only one ramp and the angle of the ramp was 12. CFD analysis, that represented the test section of
the supersonic wind tunnel and the inlet/isolator model in it in three dimensions, were performed to
investigate shock wave structure in the inlet along the isolator. Results about shock wave – boundary
layer interactions and the effects of flow separations with increasing boundary layer thickness are
discussed. As the next step, inlet unstart is mentioned for future work.

v
1. INTRODUCTION

Scramjet and ramjet propulsion are prior systems, which is today’s developing technologies, for
supersonic and hypersonic flight. Ramjet/Scramjet engines use the surrounding air as a working fluid
and known as a kind of air breathing engines. Air first passes through inlet/isolator than mix with fuel
and ignited in the combustor. Hot exhaust gases pass through a nozzle and produces thrust.
Ramjet/Scramjet uses forward motion of the vehicle to compress incoming air at the inlet instead of
a rotary compressor in conventional jet engines. Therefore they have less moving parts and they are
much lighter than turbine engines.

In ramjet, air is slowed down to a subsonic speed by multiple shock waves in the inlet/isolator. Fuel is
mixed with the subsonic flow and ignited then hot exhaust is accelerated through a throat to
supersonic speeds. Inefficiencies regarding performance loses occur due to shock waves in the
inlet/isolator. Also thermal and mechanical loads increase on the combustion chamber walls.
Because of these performance limitations, ramjet propulsion is not suitable above Mach 5 or five
times the speed of sound [1].

In scramjet (supersonic combustion ramjet) combustion process occurs supersonically. Therefore a

choking mechanism in the throat is not required to accelerate the airflow. Since flow does not slow
very much in the combustor, it is more efficient for hypersonic flights thus for higher Mach Numbers.
Also compared to rockets, which must carry its own oxidizer, these air breathing engines are more
efficient for a flight in the atmosphere [2]. As a hypersonic craft, scramjets will have different
categories: weapons, such as cruise missiles; aircraft, such as for global strike and reconnaissance
missions; and space-access vehicles taking off and landing like airliners [3].

In Figure 1 a schematic of a scramjet is shown. Forebody provides the first compression of airflow.
Shock waves in the inlet/isolator convert the dynamic pressure due to the velocity of incoming air
into higher static pressure. The shock train can be shown clearly in Figure 1 and Mach number
reduces while static pressure and temperature increase along the isolator through shock waves.

Figure 1 Schematic of a scramjet [6]

1
 Figure 2 Scramjet components [3]

For ramjet/scramjet, isolator is a critical component that adjusts the static back-pressure. At the
beginning of the combustion, boundary layer starts to separate and a precombustion shock appears
in the isolator [3]. In Figure 2 boundary layer separation and precombustion shock structure for a
scramjet isolator can be observed clearly. Because of the combustion process, static back-pressure
changes in the isolator. The pressure rise can force the shock waves to be pushed out of the isolator
and inlet. This condition is called inlet unstart in which enough airflow cannot enter the isolator.

Specific impulse (Isp) can be used to compare efficiencies between various engine types and
performance levels. In Figure 3, different engine types are shown on a graph Specific Impulse versus
Mach Number and it is indicated that hydrogen is more efficient than hydrocarbon fuel. Also, it is
clearly shown that at Mach numbers higher than 6-7, only rockets and scramjets are available and
scramjets are more advantageous with having higher specific impulse than rockets [4]. On the other
hand ramjet/scramjet propulsion does not work at subsonic speed which means that it is needed a
vehicle to carry them at a suitable speed and altitude in order to start the flight on their own or gas
turbine engines or rockets can provide the acceleration form takeoff to speeds about Mach 3. At
Mach 3–4, a scramjet transitions from low-speed propulsion to a situation in which the shock system
has sufficient strength to create a region(s) of subsonic flow at the entrance to the combustor [3].

2
 Figure 3 Specific Impulse and Mach Number for Various Engine Types [5]

Ramjet/Scramjet modes and properties from Mach 3 to Mach 8 are shown in detail in Figure 4.
Isolator pressure ratio peaks at Mach 6 and between Mach 5 and Mach 7, transition forms which is a
mixture of ramjet and scramjet combustion. The total rise in temperature and pressure across the
combustor begins to decrease and the precombustion shock is pulled back from the inlet throat
toward the entrance of the combustor. Beyond Mach 7, the combustion process can no longer
separate the airflow, and the engine operates in scramjet mode without a precombustion shock. The
inlet shocks propagate through the entire engine. Operations at Mach 5-15 bring some problems for
efficiency. Supersonic mixing of the air and fuel at speeds around and above Mach 12 is difficult and
rising temperature by increasing Mach numbers cause dissociation and ionization in the combustor
[3].

Hypersonic vehicles subject to several extraordinary challenge both for structure and material. These
are [3]:

- Very high temperatures

- Heating of the whole vehicle
- Steady-state and transient localized heating from shock waves
- High aerodynamic loads
- High fluctuating pressure loads
- The potential for severe flutter, vibration, fluctuating and thermally induced stresses
- Erosion from air flow over the vehicle and through the engine

3
 Figure 4 Ramjet/Scramjet modes change from Mach 3 to Mach 8 [3]

By developing technology, new research on materials and more experiments are done on scramjet
propulsion, hypersonic flights will be available both for space access and long distance transportation
in the future.

4
2. EXPERIMENTAL SETUP

2.1. WIND TUNNEL

Experiments were performed in the supersonic wind tunnel in the Trisonic Research Laboratory at
the Faculty of Aeronautics and Astronautics of Istanbul Technical University. This supersonic wind
tunnel has a test section size of 15x15 cm and can operate at Mach number ranges of 0.4 – 2.2 and
2.4 – 4.0 by using two different test-section / nozzle-block modules. However, since the lower Mach
number test section has larger glass side walls for imaging, and it was possible to modify the test
section floor for model mounting this test section was used for the experiments. A schematic CAD
drawing of the supersonic wind tunnel compressors and tanks are shown in Figure 5 with numbering
components in order:

1. 10-Bar Compressor
This is the starting point of the system. Air first is compressed to about 10 bars and is delivered to the
next station by pipes as shown in blue in Figure 5.

2. 10-Bar Pressured Tank

Compressed air is stored in this tank up to 10 bars and passes through a filter at the exit of the tank.

3. Gas Dryer
Gas dryer provides first step in decreasing the humidity of the compressed air.

4. 40-Bar Compressor
This screw-type compressor receives the compressed air at 10 bars and compresses it up to 40 bars.
High pressured air is delivered to the next station by pipes as shown in red in Figure 5.

5. Cooler
Because of the compression process, the temperature of the compressed air increases very much.
Since the working range of the chemical dryer at the next station is between -10 and 60 °C, cooler is
used to drop the temperature of the air. It has a simple radiator mechanism with a fan cooling the air
as it passes through capillary tubes.

6. Filter
Compressed and cooled air passes through a filter by which small particles and oil are hold.

7. Chemical (Adsorption) Dryer

Chemical dryer, which is the type 0260-1200 form Donaldson Ultrafilter Company, has two cylinders
placed vertically, which operate in order. While one of the cylinders is active, other becomes inactive.
The humidity of the air is hold by adsorption using suitable chemicals at high pressure. When a
maximum humidity keeping rate is reached in the operating cylinder it becomes inactive and goes
into regeneration phase in which low pressure air is passed through to dry the moist chemicals. In
the inactive cylinder, the air is first released by decreasing the pressure to about atmospheric value,
then it is kept at this regeneration state for about 10 minutes until it becomes active while the other
cylinder goes into the regeneration phase.

5
8. High Pressure Tank for Dryer
This tank is used to maintain high pressure around 35-40 bars at the exit of the chemical dryer and
works with a check-valve.

9. Check-Valve
This valve opens above 30-35 bars. This valve is used the keep the pressure in the High Pressure Tank
for Dryer at or above 30 bars in order for the chemical dryer to work efficiently.

10 - 11. High-Pressure Tanks

Compressed air is stored up to 40 bars in this long horizontal tanks. Their volumes are 40 m3 each and
they are tested at 60 bars by Turkish LLoyd. The diameter and length of the tanks are 2 meters and
15 meters, respectively. Air is supplied to the wind tunnel from these tanks during the experiments.

12. Isolation Valve

Isolation valve is used to separate the high-pressure air stored in the main tanks from the rest of the
system. Until the high-pressure tanks are filled with air up to required pressure, isolation valve is kept
closed. At the beginning of the experiment, isolation valve is opened by the operator. For emergency
or ending the experimental process, this valve is closed remotely.

Figure 5 Supersonic wind tunnel compressors and tanks

6
 Figure 6 Supersonic wind tunnel test section and other components

13. Regulation Valve

After opening of the isolation valve, axial control regulation valve starts to operate. Since the exit of
the wind tunnel opens to the atmosphere, at the beginning of the experiment, air flow fluctuations
occur due to high pressure difference. Regulation valve adjusts the required air flow pressure. The
axial control regulation valve of the Mokveld Company is shown in Figure 7 and also it is numbered
by 13 in Figure 6. It takes about 6 seconds to be fully opened or to be fully closed of this valve.
Through a feedback mechanism, a stable pressure in the settling chamber is provided by controlling
the valve.

Figure 7 Axial Control Valve from Mokveld Company

7
14. Settling Chamber

After regulation valve, air flow goes into the settling chamber. This long settling tank provides to
reduce fluctuations and turbulence in the air flow and to obtain a uniform flow in the test section.
Settling chamber is the place where the static pressure is the highest at all along the downstream
points of the tunnel. Therefore two nickel membranes, which are shown in Figure 6 numbered by 14,
are used to be burst in case of exceeding the maximum allowable pressure in the settling chamber
for safety. These semispherical nickel membranes, shown in Figure 8, were designed for different
test section modules and according to the test section module, a 5-bar membrane and a 14-bar
membrane are used for Mach 2 and Mach 4 test section modules, respectively. Also a pressure
transducer in the settling chamber is used to close the isolation valve by sending signal to the control
computer when the allowable pressure is exceeded.

Figure 8 Nickel membrane

15. Test Section

Test section of the supersonic wind tunnel is modular and two different test section modules are
available for the Trisonic Wind Tunnel. One of these modules is designed for speeds up to Mach 4
and the other one was designed for speeds up to about Mach 2. For the present experiments, test
section for Mach number 2 is used and it has a cross section area of 15 x 15 cm. With the help of an
electric motor and a gear mechanism, the throat cross section area can be adjusted for different
Mach numbers. In the ideal condition, subsonic air flow, coming from the settling chamber, speeds
up to Mach 1 in the converging section of the Converging - Diverging nozzle in the upstream portion
of the test section module. In the diverging section of the C-D nozzle, air flow speeds up and the
desired Mach number is reached. In order to be able to visualize the shock waves with Schlieren
method and to view what is happening during the experiment, quartz windows are used in the test
section as shown in Figure 9 with the designed inlet/isolator model assembly for the experiment.
These quartz windows have great resistance for thermal shocks and have extreme flat surfaces with a
homogeneous structure without impurities such as seeds and bubbles inside. Also to decrease the

8
boundary layer thickness, a mechanism that sucks the air flow through the perforated upper and
lower walls can be used in the test section module., However, lower perforated wall was modified to
be able to mount the designed inlet/isolator model as shown in Figure 9 and no boundary layer
suction mechanism was used.

Figure 9 Test section module for Mach 2 and the designed inlet/isolator model assembly

16. Diffuser

Air flow enters the diffuser after test section portion of the wind tunnel. In the diffuser, flow
decelerates and a weak normal shock forms near the second throat. From this point on flow
becomes subsonic. As shown in Figure 6 numbered by 16, with the increasing cross section area,
subsonic air flow is slowed down and exhausted to the atmosphere.

9
2.2. SCHLIEREN SYSTEM

Schlieren system is used as a visualization technique which is based on the first derivative of the fluid
density. This method indicates the density gradients in the flow which is suitable for the flow that
sudden density changes occur across the shock waves. As shown in the schematics in Figure 10, light
from a slit source ab is collimated by the lens L1 and goes through the plane 1 in the test section.
After the light passes through the lens L2, inverted image of the source at the focal plane 2 is
produced. In this focal plane a portion of the light is cut by the knife edge and L3 lens generates the
image of the test section on the screen at plane 3. By the knife edge, contrast is increased which is
directly proportional to the density gradient in the flow.

In practical applications most schlieren systems use mirrors instead of lenses due to economic
reasons. An example of a schlieren system with mirrors is shown in Figure 11. The light after passing
through the slit, goes onto the mirror and reflected toward the test section. The light deflected by
the test section goes onto another mirror and reflected again. A knife edge, that is placed in the focal
point, blocks some of the light and the remaning schlieren image is sent to a photographic recording
device. These systems with mirrors are also used to shorten the total length of the schlieren system.

Figure 10 Schematic of schlieren system [7]

10
 Figure 11 Schlieren system with mirrors [8]

2.3. MOUNTING APPARATUS SETUP AND MODIFICATIONS

Due to the geometrical shape of inlet/isolator models, difficulties occur while mounting the models
to the test section of the wind tunnel. Inlet/isolator models are a kind of duct that has an inlet
providing the entrance of incoming air flow and has an outlet providing the connection to the
combustion chamber. Therefore, they should be mounted from top or bottom along the air flow
direction. The designed inlet/isolator model represents a ramjet/scramjet inlet that is mounted
under the body of an aircraft. As shown in Figure 9, lower wall of the wind tunnel acts as the fore
body of the vehicle.

In the supersonic wind tunnel at the Trisonic Research Laboratory of ITU, only a stick is available in
the center of the test section along the air flow direction as a mounting apparatus. Since this stick is
not suitable to mount a ramjet/scramjet inlet/isolator model, some modifications had to be done on
the test section module. Supersonic wind tunnel has a boundary layer suction mechanism
underneath both the top and the bottom walls to reduce the thickness of the boundary layer
developing along the tunnel nozzle and test section walls. Since this mechanism is not used during
the experiments, it is decided to remove the one in the bottom wall and a new mounting apparatus
is designed as seen in Figure 12 numbered by 1.

11
 Figure 12 Mounting apparatus and inlet/isolator model assembly

The original walls of the tunnel test section has a perforated plate both on the floor and the ceiling
for boundary layer suction. In the present study, a new stainless steel floor plate with a plug for
model mounting is designed and manufactured to replace this perforated plate. Since no CAD data of
the original tunnel is available, some problems rose while matching the 40 of 3 mm diameter holes in
the correct position in the new design. First, a large graph paper is used to specify the centers of the
holes of the original 2.9 mm thick perforated plate which is the main element consisting of the
bottom wall of the test section boundary layer suction mechanism. After scanning the paper, the
scanned image is transferred to CATIA, which is a CAD program, and the 3D model was created. In
order to verify whether the holes fit with the test section structure, 3 mm thick Plexiglas were cut
with a laser cutter by using CorelDRAW software. But, it is observed that holes did not suit well
enough. By placing the Plexiglas to the wind tunnel, even though wrong centered holes were tried to
slide to the correct position, after several more new Plexiglasses were cut, it is understood that this
method is not suitable. To specify the coordinates of the holes better, a 3D scanner is used from
GOM Company as shown in Figure 13.

Figure 13 3D scanner from GOM Company

12
3D scanner generates geometries in a STL format as shown in Figure 14. After this STL format was
imported into CATIA, main geometry and correct positions of the holes were drawn on the top plane
as a sketch and the 3D CAD model was created. The result was verified again with a Plexiglas model.

Figure 14 Scanned perforated plate

By considering future work, another floor plug, which can be replaceable with another one when
needed, is decided to be placed into the center of this scanned model. This second plate is shown in
Figure 12 numbered by 2. After determining the dimensions, the final CAD model was manufactured
from stainless steel with a CNC machine. The resultant model is shown from top in the upper side
and from bottom in the lower side of Figure 15.

Figure 15 Designed first mounting apparatus

13
Boundary layer suction mechanism also contains the bottom cover that provides a sealing between
the atmosphere and the environment inside the test section. Therefore, after removing the
mechanism, a new bottom cover should have been required to be designed. The same 3D scanner
was used to determine the holes and the sealing slot. Then the model was manufactured from an
aluminum block with a CNC machine as shown in Figure 16.

Figure 16 Scanned and manufactured bottom cover of the wind tunnel

14
2.4. INLET/ISOLATOR DESIGN AND CALCULATIONS

Inlet design of a ramjet/scramjet is critical due to the first shock wave development and as a result
compression of the incoming air after the shock wave. The mechanism that creates shock wave is a
ramp which deflects incoming air flow. As shown in Figure 17, incoming air flow at M1 goes through
an oblique shock wave with an angle of  due to the ramp and the flow deflects by an angle of θ with
a speed M2 where Mn1 and Mn2 are the Mach number components normal to the shock for M1 and
M2, respectively.

Figure 17 Oblique Shock [9]

According to the flow velocity, maximum θ angle is shown in Figure 18 which can be determined with
the equation (1):

(1)

If the maximum θ angle is exceeded, instead of an attached oblique shock wave, detached bow
shock wave is formed which can be observed in Figure 19. Detached bow shock wave is an undesired
condition for a ramjet/scramjet because this cause the satiation called inlet unstart and also the flow
behind the shock wave becomes subsonic.

Since the experiments are carried out at about Mach 2, the maximum determined θ ramp angle is
about 22.97° for standard air with specific heat ratio () of 1.4. Therefore, all the ramps of the
inlet model, that face with the free stream, should be less than this maximum angle.

15
 Figure 18 Relation between θ,  and Mach number [9]

Figure 19 Comparison between Mach numbers and  angles [9]

16
In order to keep the velocity behind the first shock in supersonic range, ramp angle of the inlet model
was chosen as 12° as shown in Figure 20 where the air flow direction is from left to the right. To be
able to observe shock wave reflections, the length of the isolator is kept as long as possible and the
total length became 160 mm. To observe at least two reflections, and to prevent boundary layer
interactions and inlet unstart, the height is set to 50 mm with the help of CFD (Computational Fluid
Dynamics) analyses that will be discussed in further sections.

Figure 20 side view of the inlet/isolator model

For the first shock wave, Mach number calculations can be done for M1 = 2 and θ = 12° ramp angle:

From equation (1),  angle is calculated as 41.6°.

= 1.3272 (2)

= 0.7725 (3)

= 1.5651 (4)

From equation (4) for a free stream Mach number of 2 and ramp angle of 12°, the velocity behind
the oblique shock wave becomes 1.5651 which is still supersonic.

17
When viewed from the top, the ramp angles were chosen 20° which is again below the maximum of
θ as shown in Figure 21. To be able to start the wind tunnel, the blockage ratio, which is the total
projection area of the model divided by test section area, should be below 15%. The ratio for the
designed inlet/isolator model is about 5.7% which satisfies the required condition.

Figure 21 top view of the inlet/isolator model

The inlet/isolator model was designed as consisting of 5 pieces for the production simplicity. The
parts numbered by 4 and 5 in Figure 12 were made of brass due to strength considerations and the
number 3 was made of Plexiglas since it had to be transparent to observe shock waves with schlieren
method. The part numbered by 2 is designed to connect the inlet/isolator model to the part
numbered by 1. An exploded view of the assembly is shown in Figure 22. For the fasteners M3, M4
and M5 type of cap screws were used in the appropriate positions.

18
 Figure 22 Exploded view of the model assembly

Also to minimize turbulence formations in the flow, all the fasteners were embedded. Due to the
pressure differences on the walls of the inlet/isolator, as can be observed from Figure 23, liquid
sealant was used between the connections of the parts to prevent any leakage of air. Since any
geometrical defect on the surface causes turbulence or even flow separations, a kind of wax was
used to fill the small cavities on the head of bolts.

Figure 23 Inlet/isolator model assembly

19
3. RESULTS

3.1. CFD ANALYSES

Computational Fluid Dynamics, usually called CFD for short, uses numerical methods and algorithms
to analyze and solve the problems related with fluid flows. For the analysis of the inlet/isolator model
and the test section cell, Star-CCM+ software was used. First, a 2D coarse solution of the problem
was investigated to understand the characteristics of shock waves and expansion waves. Afterwards
a 3D fine solution was employed. Details of these studies will be explained in further sections.

The problem were run with 43293 cells and 86198 faces. For Mach 2, the physical model was
selected as:

- Two Dimensional
- Inviscid
- Gas
- Coupled Flow
- Ideal Gas
- Coupled Energy
- Steady

This fast solution with a coarse mesh was done to decide the dimensions of the experimental model.
As shown in Figure 24 at least two reflections were observed. Since y-momentum of the residuals
was not converged enough, the profile at the back side area of the isolator does not represent the
correct situation.

Figure 24 Mach number distribution of the 2D inlet/isolator model

20
Since the actual flow is 3D, and the flow is viscous and turbulent, a much better defined problem was
run. First, the volume that fluid flows is modeled as solid and the CAD model was imported into Star-
CCM+ as shown in Figure 25.

Figure 25 CAD model of the fluid structure

The next step was generating a suitable mesh. Hexahedral mesh type was selected since it was
suitable for this kind of problems. Since the flow is viscous, to observe the boundary layer
interactions and turbulent flow separations, Prism Layer Mesher was selected in the Meshing Models
option. Also Trimmer and Surface Remesher were selected for a good volume mesh.

Fine mesh is not required in everywhere, therefore in order to decrease the number of cells which
saves time, five regions were created with mesh sizes getting bigger in every step. Maximum Cell
Sizes of the regions were set to 1, 4, 16, 32 and 64 mm in order. Prism Layer Thicknesses on the walls
were set to 2 mm and Prism Layer Stretching was set as 1.1. Number of Prism Layers was defined as
20, which created twenty layers in the defined boundary layer thickness. Also for a slow transition of
surface cell size between different regions, Surface Grow Rate was set to 1.1 and Template Grow
Rate was set as slow.

The resulting mesh consists of 7307598 cells and 21864559 faces as shown in Figure 26 from
different aspects. The back side of the model was kept long in order to get better convergence for
better results and this also prevented the reverse flow warnings during iteration.

21
 Figure 26 Mesh structure from different aspects

For the Physical Models selection, the following were chosen to define the physical conditions in the
program:

- Three Dimensional
- Gas
- Coupled Flow
- Ideal Gas
- Coupled Energy
- Turbulent
- Reynolds-Averaged Navier-Stokes
- Spalart-Allmaras Turbulence
- Standard Spalart-Allmaras
- All y+ Wall Treatment
- Steady

22
This problem is analyzed under some assumptions. First of all it was accepted that the flow was
steady and the fluid was standard air which was assumed as ideal gas. Also the flow is turbulent and
Spalart-Allmaras turbulence model was used. For the turbulent boundary layer y+ Wall Treatment
was selected.

In the CFD model the entrance of the flow to the tunnel is defined as inlet and the exit from the
tunnel is defined as outlet. In order to define the boundary conditions for inlet and outlet, isentropic
flow equations are used for a free stream Mach number of 2.

(5)

(6)

(7)

For the free stream Mach number of 2, the static temperature of the airflow at the inlet was set to
160 K derived from equation (5) and Reference Pressure was set to 33670 Pa derived from equation
(6). Also outlet static temperature was defined as 288 K. For a fast solution of the problem, an initial
500 m/s velocity in the x-direction, which is very close to the two times the speed of sound derived
from equation (7), is defined as initial condition.

Even if it takes more time to solve, for a better convergence, Courant Number, which is also known
as CFL number, is taken as 0.1 for this problem. After 1270 steps, it was observed that shock waves
settled enough.

23
When viewed from the side as shown in Figure 27 Mach number distribution clearly shows the shock
waves. After the first shock wave Mach number decreases as normal for an oblique shock wave but
with the help of the expansion wave, flow speeds up again where it reaches nearly to the free stream
Mach number. But, in the regions where shock wave interacts with boundary layer, the flow slows
down. Because of the pressure difference between upstream and the downstream of the shock
wave, boundary layer thickness increases significantly especially in the area where the interaction
with the first shock wave since it is the strongest of all shock waves. This thick boundary layer acts as
a bump on the wall and for this reason another shock wave appears. Shock waves continue by
reflecting along x-direction and their strength decrease. Also as shown in Figure 28 boundary layer
separation can be observed clearly for the first shock wave – boundary layer interaction. The velocity
vectors that point in the negative x-direction is the proof for the boundary layer separation. The red
region in Figure 27 can be misleading because due to the expansion waves at the back and the
reflection from the side walls of the wind tunnel, Mach number is shown as the maximum in this area
which does not represent the real situation in the air rather than the wind tunnel.

Figure 27 Shock waves and Mach number distribution from side view

24
 Figure 28 Boundary layer separation vectors

When viewed from the top, similar conditions are observed as shown in Figure 29. Since the tip angle
from the top is higher than the angle from the side, the shock wave appears to be stronger as shown
clearly in Figure 29.

Figure 29 Mach number distribution from the top view

25
Using density distribution to visualize the results of the CFD analysis has been considered appropriate
since the Schlieren method is related with the density gradients. As shown in Figure 30, the density
distributions from the side view help understand the shock wave structure better due to the sudden
density change across the shock wave. By starting from the centerline, side view cross sections of the
density distribution are shown in Figure 30 numbered from 1 to 8. Number 1 represents center and
number 2 represents the cross section 10 mm away from the center in the z-direction. This 10 mm
spacing goes until number 8 which represents 70 mm away from the center and the wall of the wind
tunnel is at 75 mm away. Since the solution is symmetric with respect to the center plane, only one
side is enough to analyze the problem. From number 1 to number 3, shock waves and reflections can
be clearly observed in the inlet/isolator. Number 4 is the cross section where outer side of the
inlet/isolator vertical wall. This shows the shock wave profile of the 20 angle tip ramp of the side
walls in the front. Numbers from 5 to 8 show the shock wave structure between outside of the
inlet/isolator and the wind tunnel side wall. Since the light rays in the Schlieren method go through
the z-direction, the resultant image is affected by all cross sectional regions. Therefore, the average
of all the solution of the cells from one side to the other side of the wind tunnel side walls were
calculated by exporting results to Tecplot and creating the avarege final image by using Matlab. The
expected schlieren image is shown in Figure 31 and it will be compared with experimental results in
the further sections.

26
Figure 30 Density distributions from the side view

27
 Figure 31 Average density distribution along z-direction

3.2. EXPERIMENTAL RESULTS

Experiments were performed in the supersonic wind tunnel in the Trisonic Research Laboratory at
the Faculty of Aeronautics and Astronautics of Istanbul Technical University. Mach 2 was reached in
the repeating experiments and nearly the same results were obtained as shown in Figure 32.

As calculated in the previous sections for Mach 2 and 12 ramp angle, from equation (1), β angle is
41.6°. Also for Mach 2 the Mach wave angle is calculated as:

(8)

The structure of the shock waves is drawn approximately in Figure 33 and the angles do not
represent exact values. According to the approximate angles, first line coming from the tip of the 12
ramp as numbered by 1 is a shock wave because the angle is closer to the calculated β angle than the
Mach angle, , and this shock wave interacts with boundary layer of the tunnel. Because of this
boundary layer - shock wave interaction, a lambda-shock forms in this region. It is clearly shown that
shock structure of the CFD results in Figure 34 is very similar to the experimental results in Figure 33.
A similar but smaller lambda-shock is formed in the CFD results but since the boundary layer in CFD
analysis is much thinner compared to that of the wind tunnel walls. Lambda shocks are the indicator
of the possibility of boundary layer separations. As indicated in the previous sections, boundary layer
separations were observed in Figure 28 with the reverse velocity vector profiles. Therefore, in the
experiments an extensive boundary layer separation takes place as indicated by such a big lambda
shock formation.
28
Figure 32 Experimental result of the inlet/isolator model at Mach 2

Figure 33 Shock wave structure of the experimental result

29
The shock wave numbered by 1 reflects and the shock wave number 2 is formed. This reflection goes
on until the shock number 4 is formed. In fact, the shock waves numbered by 4 and 5 are the same
but it looks as if they are different from each other because this is caused by the refraction of the
light by Plexiglas side walls of the inlet/isolator. The angle of the wave number 6 is very close to that
of a Mach wave and it is probably a Mach wave or a weak shock. Number 7 is formed due to the 20
tip angle of the side walls of the inlet/isolator and it is between inlet/isolator walls and wind tunnel
inner walls. This shock wave can be observed better in Figure 30 numbered from 5 to 8.

Figure 34 Shock wave structure of the CFD results

In summary, when shock waves interact with boundary layer, the thickness of the boundary layer
increases and a lambda-shock occurs as a result of flow separations. A second shock wave forms at
the back of the separated boundary layer. In order to understand and prevent unstart, shock-induced
boundary layer separations should be investigated carefully.

30
4. CONCLUSION

In this thesis, a supersonic inlet/isolator model is investigated at Mach 2 both computationally and
experimentally. With the help of CFD analysis the shock wave structure of the inlet/isolator is
observed and the interaction between shock waves and boundary layer shows the possibility of flow
separation. Also shock waves have a great effect on increasing the boundary layer thickness. This
thick boundary layer can cause the formation of a second shock wave. All of these effects have a
great role on inlet unstart which may occur at supersonic Mach numbers. For the future work, by
placing a flap in back of the isolator and adjusting the flap angle remotely, unstart conditions can be
induced and investigated better. By seeding titanium dioxide into the flow, PIV (Particle Image
Velocimetry) method can be used to visualize the flow with the help of laser light and high-speed
cameras. Also pressure taps can be used for measuring static pressure at different locations during
the unstart process.

31
 5. REFERENCES

1. Ramjet Propulsion. (n.d.). Retrieved April 2012, from NASA:

http://www.grc.nasa.gov/WWW/k-12/airplane/ramjet.html

2. Scramjet Propulsion. (n.d.). Retrieved April 2012, from NASA:

http://www.grc.nasa.gov/WWW/k-12/airplane/scramjet.html

3. Dean, A. (2004, August/September). Scramjets Integrated Air and Space. The Industrial
Physicist , 24-27.

4. Roberts, K. N. (August 2008). Analysis and Design of A Hypersonic Scramjet Engine With A
Starting Mach Number of 4. University of Texas.

5. Fry, S. R. (2004). A Century of Ramjet Propulsion Technology Evolutşon. Journal of Propulsion

and Power , 20 (1), 27-58.

6. How Scramjets Work. (n.d.). Retrieved April 2012, from NASA:

http://www.nasa.gov/centers/langley/news/factsheets/X43A_2006_5.html

7. Holman, J. P. (2001). Experimental methods for engineers. McGraw-Hill.

8. Schlieren System. (n.d.). Retrieved April 2012, from NASA:

http://www.grc.nasa.gov/WWW/k-12/airplane/tunvschlrn.htmll

9. Yükselen, M. A. (2006, March). Compressible Aerodynamics Lecture Notes. Istanbul.

32

View publication stats