Hypersonic Term Paper
Hypersonic Term Paper
Hypersonic Term Paper
Hypersonic Aerodynamics
Term Paper
submitted by
Avijit Saha
Roll No. 13AE30025
1 Introduction 4
1.1 Different situations under which shock-shock interactions occur in practical configu-
rations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1
List of Figures
[2]
1.1 Location of six shock-wave interference patterns on leading edge . . . . . . . . . . . 4
[2]
1.2 Edney catalogued six types of shock interactions . (IS = Incident Shock, BS = Bow
Shock, RS = Reflected Shock, EF = Expansion Fan, TP = Triple Point, SL = Shear
Layer, M∞ = Free stream Mach number ) . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Shock interaction in turbine blade row . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Shock interaction in engine cowl (upper) and space shuttle (lower) . . . . . . . . . . 6
2.1 Shock Polar for a Mach number of 1.8 and a specific heat ratio 1.4. . . . . . . . . . 7
2.2 The dependence of the (P,θ)-shock polar on the incident M for γ= 1.4. . . . . . . . 8
2.3 The dependence of the (p, θ) - shock polar on real gas effects for nitrogen at Mo =10,
Po =15 torr and To =300K. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Schematic diagram of Type I interaction. . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5 Shock Polar diagram of Type I interaction. . . . . . . . . . . . . . . . . . . . . . . . 9
2.6 Schematic diagram of Type II interaction. . . . . . . . . . . . . . . . . . . . . . . . 10
2.7 Shock Polar diagram of Type II interaction. . . . . . . . . . . . . . . . . . . . . . . 11
2.8 Schematic diagram of Type III interaction. . . . . . . . . . . . . . . . . . . . . . . . 12
2.9 Shock Polar diagram of Type III interaction. . . . . . . . . . . . . . . . . . . . . . . 12
2.10 Schematic diagram of Type IV interaction. . . . . . . . . . . . . . . . . . . . . . . . 13
2.11 Detail of supersonic jet structure of Type IV interference: situation a. . . . . . . . . 14
2.12 Detail of supersonic jet structure of Type IV interference: situation b. . . . . . . . . 14
2.13 Shock Polar diagram of Type IV interaction. . . . . . . . . . . . . . . . . . . . . . . 14
2.14 Schematic diagram of Type V interaction. . . . . . . . . . . . . . . . . . . . . . . . 15
2.15 Shock Polar diagram of Type V interaction. . . . . . . . . . . . . . . . . . . . . . . 16
2.16 Schematic diagram of Type VI interaction. . . . . . . . . . . . . . . . . . . . . . . . 16
2.17 Shock Polar diagram of Type VI interaction. . . . . . . . . . . . . . . . . . . . . . . 17
3.1 Sequential Schlieren photographs of the flow development for a CR of 8.4 at Mach 8. 18
3.2 Comparison between shock pattern obtained both experimentally and numerically
for CR 8.4 at Mach 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3 Flow morphology observed for the 8.4 CR case at Mach 8. . . . . . . . . . . . . . . 19
3.4 Sequential Schlieren photographs of the flow development for a CR of 5.0 at Mach 8. 20
3.5 Sequential Schlieren photographs of the flow development for a CR of 4.3 at Mach 8. 20
3.6 Edney type II shock interference observed for CR 4.3. . . . . . . . . . . . . . . . . . 21
3.7 Schlieren image of Edney IV interference, using a 20 shock generator and a 25 mm
hemisphere cylinder. Flow is from left to right. . . . . . . . . . . . . . . . . . . . . . 21
3.8 Shadowgraph of Edney Type IV interactions on a hemisphere cylinder at different
angles of attack. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.9 a) Cell Removal, b) local re-meshing. . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.10 a) treating the shock as internal boundary, b) interpolation of the phantom nodes. . 23
3.11 Mesh points jumped by the shock and their re-computation. . . . . . . . . . . . . . 23
2
3.12 a) Shock-capturing mode:all shocks and contact discontinuities are computed by the
S-C solver, b) Hybrid mode 1: the bow shocks and oblique shock connecting the triple
point are fitted; the remaining shocks and contact discontinuities are computed by
the SC solver. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.13 a) Hybrid mode 2: the impinging shock, the bow shocks and the oblique shock con-
necting the triple points are fitted; the remaining shocks and contact discontinuities
are computed by the S-C solver, b) All fitted mode: all shocks and contact disconti-
nuities are fitted together with the two triple points. . . . . . . . . . . . . . . . . . . 24
3.14 Diagram of Mach number vs. second wedge angle showing the regimes of the different
interactions, a) high Mach no. flows, b) low Mach no. flows, physical parameters are
0
θ=15 , γ=1.4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.15 a) Nondimensional heating-rate distributions, b) Nondimensional pressure distribu-
tions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.16 a) Mach contours and streamlines, b) Density contours in interaction region. . . . . 26
3.17 a) Profiles at y = -2 mm (above the upper shock triple point), b) Profiles at y = -4
mm (crossing the transmitted shock and intersecting the surface just below the jet
impingement) c) Profiles at y = -5 mm (just below the second triple point). . . . . . 27
3.18 Test model: a) X = 60-70 mm, Y = 19.5 mm; and b) wedge (label 1), cylinder with
cavity (label 2), pressure transducers (label 3), and end plate (label 4). . . . . . . . 28
3.19 a) Experimental shadowgraph image, b) Experimental pressure-time history. . . . . 28
3.20 a) Numerical shadowgraph image, b) Numerical pressure-time history. . . . . . . . . 29
3.21 Type IVa shock interaction: the jet grazes the upper surface. . . . . . . . . . . . . . 29
3.22 Type IV shock interaction: the jet impinges on the lower surface. . . . . . . . . . . 29
3.23 Pressure histories in the cavity (left: Experimental, right: Numerical): forward-
backward oscillation mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.24 Set of time-sequential shadowgraph images of unsteady Type IVa shock interaction:
forward-backward oscillation mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.25 Temperature contours and streamlines: up-down oscillation mode. . . . . . . . . . . 31
3.26 Pressure histories in the cavity: upâĂŞdown oscillation mode. . . . . . . . . . . . . 31
3.27 Set of time-sequential shadowgraph images of unsteady type-IV shock interaction:
up-down oscillation mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.28 Temperature contours and streamlines: up-down oscillation mode. . . . . . . . . . . 32
3
Chapter 1
Introduction
Shock-shock interactions is nothing but interaction between shock waves. Hypersonic vehicle such
as Space Shuttle, missile with integrated ramjet or scramjet engine is subjected to this phenomenon.
Shock interactions causes significant increase in local surface temperature and pressure which if not
predicted accurately may lead to damage the whole vehicle. Six general type of interaction between
bow shocks around blunt bodies and incident shocks have been identified by Edney. Consequences of
the interaction depends on the strength as well as relative angle of the bow shock and the incident
shock. The above stated factors determines how will be the reflected shock, whether it will be
curved or not, how much angle it will have, how many number of shear layer will form, will there
[1]
be any supersonic jet present or not all etc .
Approximate angular regions and the type of interaction patterns that occurs when a incident shock
[2]
Figure 1.1: Location of six shock-wave interference patterns on leading edge .
wave impinges on the curved bow shock formed by the leading edge of the body are depicted in the
Fig 1.1. Among these Type III interaction causes huge localized pressure and heating to the vehicle
due to direct impingement of shear layer on the surface. In case of Type IV pressure and heating
rate will be maximum if supersonic jet behind the intersection region impinges perpendicular to the
[3]
leading edge surface .
4
[2]
Figure 1.2: Edney catalogued six types of shock interactions . (IS = Incident Shock,
BS = Bow Shock, RS = Reflected Shock, EF = Expansion Fan, TP = Triple Point,
SL = Shear Layer, M∞ = Free stream Mach number )
5
Figure 1.3: Shock interaction in turbine blade row
Figure 1.4: Shock interaction in engine cowl (upper) and space shuttle (lower)
6
Chapter 2
Figure 2.1: Shock Polar for a Mach number of 1.8 and a specific heat ratio 1.4.
state behind the oblique shock wave is identical to the flow state ahead of it. In this case shock
wave angle is β = µ = sin (1/M0 ) and flow deflection angle is θ = 0. Also pressure jump across the
−1
oblique shock wave is minimum (i.e., p1 /p0 = 1). Point “b” represents the state which is obtained
by passing the flow through the strongest oblique shock wave i.e., the normal shock wave (shock
wave angle β = 90 ). Pressure jump across the normal shock wave p1 /p0 will be maximum and flow
0
deflection (θ) will be zero. Point “s” separates the shock polar into weak and strong shock portions.
7
Above “s” Mach number behind the oblique shock wave is less than 1 (i.e, flow is subsonic) and
below “s” Mach number is greater than 1 (i.e., flow is supersonic). Section “a” - “s” and “s” - “b” are
known as weak shock and strong shock portions respectively. Maximum deflection or detachment
point is denoted by point “m”. The difference between θs and θm is very small. In general both of
these points are treated as a single point.
As Mach number increases maximum deflection angle and pressure jump across the shock also
increases, thus shock polar becomes larger (shown in Fig 2.2). Another important fact about shock
polar is that if we take into account the real gas effects shock polar will change (Fig 2.3).
Figure 2.2: The dependence of the (P,θ)-shock polar on the incident M for γ= 1.4.
Figure 2.3: The dependence of the (p, θ) - shock polar on real gas effects for nitrogen
at Mo =10, Po =15 torr and To =300K.
8
2.2 Type I Interaction
Type I interaction involves intersection between two weak oblique shock waves of opposite fam-
ilies. The left running incident shock (IS) impinges on right running bow shock (BS) at a location
sufficiently downstream of the sonic point behind the bow shock leads to the formation of a two
transmitted oblique shock and a slip line that trails downstream from the intersection point. In
this case flow is locally supersonic everywhere. While passing through these weak shock waves (IS
and BS) free stream flow field is turned into itself resulting the flow field in regions 2 and 3. Flow
fields behind the two refracted shock waves are denoted by 4 and 5. These two regions are having
same pressure and flow direction and they are separated by a shear layer or popularly known as
slip line since velocities in region 4 and 5 may differ in magnitude. Transmitted shock wave that
separates region 2 and 4 impinges on the surface which in turn leads to shock-wave-boundary-layer
interaction. Now in order to become tangent to the upper wall boundary in region 6 the flow field
in region 4 is again turned through an oblique shock wave.
Now let us explain the above interaction using shock polar diagram. In region 1 we have θ = 0. At
this location we can draw a pressure deflection diagram for free stream Mach number M1 . The free
9
stream flow is turned downward (say by θ2 ) through a right running BS. Thus in region 2 flow is
already deflected by θ2 . We can draw another pressure deflection diagram for region 2 and 3 whose
vertex is denoted by 2 and 3 respectively (and these vertexes will lie on the pressure-deflection curve
of region 1). Flow direction (say φ) and pressure in region 4 and 5 will be same and in shock polar
this is represented by the intersection point (4, 5) of the pressure-deflection curve of region 2 with the
curve of region 3 (φ would be zero if flow field in these region is parallel to the free stream). Apart
from this intersection the curves of region 2 and 3 also intersects the strong shock portion of the
[3]
curve of region 1 at a pressure above the intersection with each other. As reported by Glass, C. E.
if such situation exists then the flow filed resulting from interference will occur at lower pressure.
The flow field in region 6 is denoted by point 6 which lies on the pressure deflection curve of region 4.
10
Figure 2.7: Shock Polar diagram of Type II interaction.
stream flow is turned downward (say by θ2 ) through a right running BS. Thus in region 2 flow is
already deflected by θ2 . Similarly in region 3 flow is already deflected by θ3 . Another pressure
deflection diagram is drawn for region 2 and 3 whose vertex is denoted by 2 and 3 respectively
(and these vertexes will lie on the pressure-deflection curve of region 1). Flow direction (say φ) and
pressure in region 4 and 5 will be same and in shock polar this is represented by the intersection
point (4, 5) of the pressure-deflection curve of region 2 with the curve of region 1. Similarly flow
direction and pressure in region 7 and 8 will be same and in shock polar this is represented by the
intersection point (7, 8) of the pressure-deflection curve of region 3 with the curve of region 1. Due
to weak compression flow in region 4 and 7 will be supersonic. However in region 5 flow will be
subsonic because it results from normal shock compression. Here the intersection of curve of region
2 and 3 is above their intersection with the strong shock portion of the curve of region 1. Movement
of the intersection point of region 2 and region 3 curve to a pressure above the strong shock solution
of the region 1 curve will determine the transition point between Type I and Type II.
11
Figure 2.8: Schematic diagram of Type III interaction.
The pressure-deflection diagram for Type III interaction is shown in Fig 2.7. As we have dis-
cussed earlier pressure in region 3 will be located on the pressure-deflection curve of region 1. The
intersection point of p − θ curve of region 1 and 3 will give the turning angle and pressure of the
region 2 and 4. Shear layer separating region 2 and 4 will attach to the wall boundary layer which
in turn cause shear layer-boundary layer interaction. Aerothermodynamic load faced by the vehi-
cle surface will depend upon the impingement angle θ5 , pressure rise between region 4 and 5 and
whether the shear layer is laminar or turbulent. Pressure in region 5 will lie on p − θ curve of region
4 and region 6 will be located on p − θ curve of region 5 (though it is shown in the figure). The
intersection point of p − θ curve of region 3 and 5 will give the turning angle and pressure of the
region 6 and 7.
12
2.5 Type IV Interaction
Type IV interference pattern results from either intersection of shocks of opposite families or
the shocks of same family. Intersection point can be above or below the normal shock portion of
the bow shock. Here Incident shock impinges on the bow shock at a location far upstream of the
sonic point behind the bow shock. In this case turning angle of the flow by the body is too large for
the supersonic flow to be deflected downstream through an attached oblique shock. the flow field
is same as Type III pattern until region 4. So we can treat it as a special case of this interference.
However in downstream region it differs significantly from Type III pattern. Like Type III here also
we got second triple point J2, but the shock emanating from J2 is much more intense which causes
the flow field behind it to become subsonic. This means there won’t be much variation of pressure
compared to the supersonic condition. Downstream of the left running shock that originates from
the J2 is supersonic. It is clearly visible that supersonic jet formed behind the impingement point
is bounded on both sides by the slip lines that separates the subsonic and supersonic region. Along
these two slip lines pressure is nearly constant. A centered expansion fan originates from J3 in order
to compensate for the pressure jump across the right running shock emanating from J2. Deflection
of the flow increases due to formation of these expansion wave. Thus the supersonic flow penetrates
toward the body and terminates in a normal shock wave close to the body surface, followed by
a subsonic region with a stagnation point at the impact on the body. Pressure and heat fluxes
experienced in this region of the body surface is much higher than in case of Type III interference.
Fig 2.9 shows the expanded view of the supersonic jet. Compression and expansion waves will be
simple waves (except the region where they intersect) and flow in region 5, 7 & 8 will be uniform, if
the flow downstream of the shock is uniform and pressure in the subsonic region 3 & 6 are constant.
As we see from the figure that compression waves are converging but not intersecting with each
other. There is a possibility that they might intersect. Then the supersonic jet structure will
be like Fig 3.0. Intersection of compression waves result a formation of focalization shock within
the supersonic jet. The origin of the centered expansion fans located at a point on the boundary
[8]
(slip line 1) where the focalization shock impinges on it. The jet will contain rotational region
rotational region as the focalization shock becomes curved. These structure will continue and finally
end in a normal shock wave close to the surface. In order to calculate the flow properties inside
this supersonic jet we need to use rotational method of characteristics (since the flow is rotational
simple MOC won’t give accurate result).
13
Figure 2.11: Detail of supersonic jet structure of Type IV interference: situation a.
14
Fig 2.11 depicts the p − θ diagram for Type IV interference pattern described in Fig 2.8. Point 3
on the p − θ curve of region 1 represents the flow in region region 3. Turning angle and pressure of
the region 2 and 4 will be given by the intersection point of p − θ curve of region 1 and 3. Similarly
the intersection point of p − θ curve of region 4 and 3 will give the turning angle and pressure of the
region 5 and 6. The images of the successive expansion and compressions are part of the isentropic
polars (Prandtl-Mayer Expansion). Point 7 on the strong shock portion of the p − θ curve of region
1 denotes the flow condition in region 7. The point at which p − θ curve of region 7 intersects with
p − θ of region 3 gives the flow direction and pressure of the region 8.
Type V interference pattern occurs when the incident shock impinges on the bow shock at a
location just above the upper sonic point behind the bow shock. Thus the Type V interaction is
somewhat like Type II interaction. In this case both shocks are of same family, namely left running
shock waves. The free-stream flow in region 1 is initially compressed to region 2 through a weak
left-running incident shock wave. Flow in region 2 is then turned towards itself through another
shock wave in order to match the leading-edge wall turning angle. Pressure and flow turning angle
will have to be same across the shear layer that separates region 4 and 5 . The flow in region 3 is
turned through another right-running wave to match condition in region 5. Shock-wave-boundary
layer interaction is caused due to impingement of this righ-running shock on the wall boundary
layer. This phenomenon in turn leads to increase the pressure and heat transfer at the body surface
like Type I and Type II. The two intersecting shocks (IS & BS) are connected by a normal shock
(nearly), which is essentially a Mach refection like in Type II. An embedded core of subsonic flow
trails downstream from the intersection region. This subsonic flow field is bounded on side by a slip
line and on the other side by a very thin supersonic jet. This jet is so thin (much thinner than the
Type IV supersonic jet) that it actually serves as another slip line. Type V jets does not impinge
on the surface, it turns away from the surface and dissipates.
15
Figure 2.15: Shock Polar diagram of Type V interaction.
Shock polar diagram for Type V interaction is shown in the above figure. Region 1 we have
θ = 0. Pressure-deflection diagram is drawn at point 1 for free stream Mach number M1 . The free
stream flow is turned upward (say by θ2 ) through a left-running IS. Thus in region 2 flow is already
deflected by θ2 . Point 2 in p − θ curve of region represents the flow condition in region 2. Similarly
point 3 also denotes the flow in region 3. The intersection point of p − θ curve of region 3 and 2
will give the pressure and flow turning angle in region 4 and 5. Point 6 that represents the flow
condition in region 6 will lie on the p − θ curve of region 5 drawn for Mach no. M5 at the location
of the intersection of p − θ curve of region 2 and 3.
16
Type VI interference pattern occurs from the intersection of the two shock waves of similar
families, namely left-running shock similar to the Type V interaction. Incident shock impinges on
the bow shock at a location far downstream of the sonic point behind the bow shock. Like Type
I here also intersection of two same family shock results two transmitted shock and a trailing slip
stream. The free-stream flow is initially compressed through a weak left-running shock to region
3. Then it is compressed again to region 4 through another left-running shock wave in order to
match the local turning angle at the wall. Shock wave that generates from coalescing of IS and BS
compresses the flow from region 1 to region 2. Pressure and flow direction in region 2 and 5 must
be same. But in general it is not possible to find a single shock wave which will produce the flow
field with pressure and velocity same as the flow field passing through two shock wave of different
strength. Thus in order to match the condition of the flow field in region 2, flow in region 4 must
be compressed or expanded (depend on how much intense the coalescing shock is) through a shock
wave or expansion wave respectively. This compression wave or expansion wave impinges on the
wall and causes shock-boundary layer interaction.
Above figure depicts the shock polar diagram for Type VI interaction. Since free-stream flow is
turned upward through BS θ3 will be positive. Flow field in region 3 is represented by point 3 on
p − θ curve of region1. Similarly if we draw another p − θ curve on point 3, point 4 that denotes the
flow field in region 4 will lie on that curve. Now depending on whether p4 is less than p2 or not,
shock wave or expansion wave will emanate from triple point T. In the above figure second case
is shown. Since here p4 > p2 a centered expansion must emanate from point T in order to make
pressure and flow direction compatible with region 2. In case of p4 < p2 the intensity of the shock
wave originates from triple point T will be smaller as p − θ curve of region 3 and 1 come close to
each other. Intensity will increase as these above mentioned curve move far away from each other.
17
Chapter 3
As we mentioned earlier that in worst case shock-shock interaction may lead to complete damage
of the vehicle. To avoid this type of interaction we need to design the vehicle carefully. For this
we have to know first how exactly they are occurring, at which location shock are intersecting with
each other so that we can design our vehicle accordingly. Several researcher performed various
experiments to understand the physics behind shock interference. Some of the results are presented
in this chapter.
[5]
Mahapatra, D. & Jagadeesh, G. performed an experiment to show the effect of contraction ratio
on shock interference phenomenon for 2d, planar scramjet inlet model. Scramjet engine inlet should
be capable of operating in a wide range of Mach numbers. Also inlet should have enough pressure
recovery capability while providing correct mass flow rate to the combustor. However with fixed
geometry inlet it is not possible to operate in wide range of Mach numbers. For this we need variable
geometry inlet. In order to get optimum performance from any variable geometry inlet it is very
important to adjust contraction ratio (CR) to the flight conditions. Different forebody-cowl shock
interactions will occur due to change in CR. which in turn lead to change in inlet performance
characteristics. In this work time-resolved Schlieren technique was used to visualize the shock
structure and shock reflection pattern for different CRs.
Figure 3.1: Sequential Schlieren photographs of the flow development for a CR of 8.4
at Mach 8.
18
Numerical simulations were also performed to compliment the experiments. For numerical sim-
ulation a commercial 3d compressible NS code was used and flow was assumed to be Laminar.
Hexahedral mesh was used and it was generated using a commercial mesh-generating code, Ansys
ICEM computational fluid dynamics 10. Comparison between shock pattern obtained from exper-
iments and CFD are shown below. For CR 8.4 flow feature is shown Fig 3.2. Regular reflection
Figure 3.2: Comparison between shock pattern obtained both experimentally and
numerically for CR 8.4 at Mach 8.
Figure 3.3: Flow morphology observed for the 8.4 CR case at Mach 8.
(Type I) was observed for this CR. Separation bubble and shock were seen to the lower part of the
ramp and a strong pressure gradient was observed near shoulder region. While CR changes to 4.3,
shock reflection becomes irregular. Schlieren photographs show the development of growing Mach
stem during run time. For CR 5.0 there exists Mach Reflection. Upstream movement of Mach stem
(it also grows with time) was observed in this case.
19
Figure 3.4: Sequential Schlieren photographs of the flow development for a CR of 5.0
at Mach 8.
Flow morphology for CR 4.3 is completely different from earlier cases. Here Edney’s Type II
pattern was formed. C1 and C2 are ramp shock and cowl shock respectively. C5 represents the Mach
stem and two triple points T1 and T2 are separated by it. Reflected shock C3 from triple point T1
moves towards the chamber. Reflected shock from triple point T2 is not clearly visible since it is
very weak in nature. Two slip stream originating from two triple points are also visible from Fig
36.
Figure 3.5: Sequential Schlieren photographs of the flow development for a CR of 4.3
at Mach 8.
20
Figure 3.6: Edney type II shock interference observed for CR 4.3.
[6]
Chettle et. al. experimentally investigated Edney’s Type IV shock-shock interaction at Mach
5. Experiment was performed for a hemisphere cylinder at different angles of attack like 0, + −10
−5, +
and −15. Incident shock was created by a shock generator. Schlieren and shadowgraph technique
+
were used to visualize the shock interference pattern.
From shadowgraph images Edney’s Type IV interaction is clearly visible. However shock structure
Figure 3.7: Schlieren image of Edney IV interference, using a 20 shock generator and
a 25 mm hemisphere cylinder. Flow is from left to right.
changes with angle of attack. It is found that the length of the straight shock connecting the
impinging oblique shock and bow shock increases with increase in angle of attack.
21
Figure 3.8: Shadowgraph of Edney Type IV interactions on a hemisphere cylinder at
different angles of attack.
[7]
Paciorri, et. al. have shown different computing approaches and assessed their performances
through the computation of a type IV shock-shock interaction. There will be multiple shock present
in the slow field around a hypersonic vehicle with complex shape which in turn lead to originate
shock-shock interaction. Due to these interactions vehicle surface will face huge dynamic and
thermal load. So accurate computation is an important issue. In their work they applied shock-
fitting technique to the computation of the type IV shock-shock interaction.
A brief description of shock fitting algorithm is discussed below. Here two-dimensional domain is
considered and the shock front crossing the domain at time level t. At time t solution is known
in all grid and shock points. The steps associated with the computation of the subsequent time
level t + δt are as follows: 1. Cell removal around the shock front 2. Local re-meshing around the
22
shock front 3. Computation of the tangent and normal unit vectors 4. Solution update using the
capturing code 5. Shock calculation 6. Interpolation of the phantom nodes 7. Shock displacement
8. Interpolation of the jumped nodes. Fig 3.12b shows the solution which is computed with the
Figure 3.10: a) treating the shock as internal boundary, b) interpolation of the phantom
nodes.
Figure 3.11: Mesh points jumped by the shock and their re-computation.
contribution of the S-F technique. Here bow shock and oblique shock are modeled as a unique
fitted shock. S-C code is used to capture the remaining shocks and the contact discontinuities.
Modification of the background mesh which is done by shock-fitting algorithm can be identified by
the comparison between this mesh and the one displayed in Fig 3.12a.
Fig 3.13a depicts a little improvement with respect to the previous solution. Direction changing
between the bow shock and the oblique shock in the first triple point is sharper here. Fig 3.13b
shows significant improvements with respect to all the previous solutions. In this case all the
shocks and contact discontinuities are fitted in the interaction region. And two triple points are
explicitly computed. Here all the elements of the flow structure in the interaction zone are perfectly
distinguishable.
23
Figure 3.12: a) Shock-capturing mode:all shocks and contact discontinuities are com-
puted by the S-C solver, b) Hybrid mode 1: the bow shocks and oblique shock con-
necting the triple point are fitted; the remaining shocks and contact discontinuities are
computed by the SC solver.
Figure 3.13: a) Hybrid mode 2: the impinging shock, the bow shocks and the oblique
shock connecting the triple points are fitted; the remaining shocks and contact discon-
tinuities are computed by the S-C solver, b) All fitted mode: all shocks and contact
discontinuities are fitted together with the two triple points.
24
[9]
Olejniczak et al. computationally investigated the inviscid shock interactions on double-wedge
geometries in order to understand the fundamental gas dynamics of these interactions. They also
studied the physical mechanisms for transition and identified the transition criteria. Finite volume
method is used to solve the 2d compressible Euler equations. In order to calculate the fluxes across
the cell surfaces second-order-accurate upwind flux-vector splitting method is used. Boundary
conditions are: Supersonic inflow and outflow, normal velocity & normal-direction pressure gradient
at the wedge surface is zero. To ensure that the smallest-scale structures are resolved grid sizes of
1024 × 1024 is used for computations. Typical results from their studies are discussed here. As
reported by Olejniczak et al. detailed simulations of steady inviscid shock interactions showed that
except TypeVI, Type V, Type IV, Type I interactions (these are well known as Edney’s classification
scheme) one other interaction Type IVr also occurs. It is more or less similar to Type IV only
difference is that the shock wave impinging on the wedge surface undergoes a regular reflection
instead of a Mach reflection like Type IV and Type IVr interaction occurs due to the geometrical
constraints of the double wedge and it won’t occur for the flows around blunt body investigated
by Edney. According to their calculation there exists a critical Mach number and the inboard flow
along the wedge surface above this Mach no is always underexpanded with respect to the outward
flow. In case of Type V, Type IV and Type Ivr interactions, the above phenomenon results an
underexpanded supersonic jet along the surface which further interacts with the contiguous subsonic
region resulting a high-amplitude steady pressure variations on the wedge surface. While Mach
no. is below critical either underexpanded Type VI or overexpanded (Surface pressure decreases
smoothly from a maximum at the corner of the wedges in this case) Type I, Type V and Type IV
interactions occur.
They have also identified the transition criteria between the various interactions. Below the critical
Figure 3.14: Diagram of Mach number vs. second wedge angle showing the regimes
of the different interactions, a) high Mach no. flows, b) low Mach no. flows, physical
0
parameters are θ=15 , γ=1.4.
Mach no., in the downstream of the interaction region when inboard and outboard flows are perfectly
expanded relative to each other a Type VI-Type I transition occurs. For given free stream condition
when the second wedge angle reaches the maximum deflection angle a Type I-Type V transition
occurs. When Mach no is above the critical and outward flow reaches the maximum deflection angle
for given free stream a Type VI-Type V transition occurs. Analytically it is possible to determine
these transition criteria using Method of Characteristics (MOC).
25
[10]
Moss et al. numerically studied the shock-shock interactions that include both Edney Type IV
and Type III. For computations Direct Simulation Monte Carlo (DSMC) method of Bird (M =10)
is used. A shock generator is used in order to create the incident shock. Results are compared with
properties (like surface heating and pressure and flow field values of density etc.) obtained using
Dual-Line Coherent Anti Stokes Scattering (DL-CARS) technique. Results of the investigation
of grid sensitivity emphasizing the effect on surface heating and pressure distributions are shown
below for three different body-fitted grids. Computational time step is reduced in order to refine
the grid. This refinement shows noticeable effect on location and magnitude of pressure and peak
heating. Experimental data is also included in Fig 3.15. Consistent trends are obtained for both
experimental a d computational data. However the magnitude and location of peak values differs
slightly.
26
Fig 3.16 represents the flow-field structure for the Type IV interaction. Two triple points,
transmitted shock, supersonic jet associated with shear layer are clearly visible. Supersonic jet
terminates very close to the surface. The point at which jet impingement occurs, overall kinetic
temperature is 1050 K and the maximum density is 183 times the free-stream value. In Fig 3.17
horizontal profiles for density and temperature are shown for both DSMC calculation and ONERA
DL-CARS measurements. Good agreement is observed between experimental and computational
results at the location of bow shock, transmitted shock and extended bow shock locations. For each
of the cases temperature and density adjacent to surface is very high (for case b it is 178 times free
stream).
Figure 3.17: a) Profiles at y = -2 mm (above the upper shock triple point), b) Profiles
at y = -4 mm (crossing the transmitted shock and intersecting the surface just below
the jet impingement) c) Profiles at y = -5 mm (just below the second triple point).
27
[11]
Xiao et al. experimentally and numerically investigated te hypersonic Type IV shock interac-
tion over a cylinder with a forward-facing cavity. The objective was to identify the unsteadiness of
the coupled interaction flow and to clarify how presence of cavity affects the flow parameters. NS-2-
type piezoresistive pressure transducers with sensing area of 3 mm in diameter were used to measure
the pressure. Unsteady interaction patterns were visualized using high-speed schlieren photography.
A direct shadowgraph method was also used due to high sensitivity while introducing the knife edge
into the optical system. 2D axisymmetric vectorized adaptive solver VAS2D was used to solve the
Figure 3.18: Test model: a) X = 60-70 mm, Y = 19.5 mm; and b) wedge (label 1),
cylinder with cavity (label 2), pressure transducers (label 3), and end plate (label 4).
laminar compressible Navier-Stokes equations. This solver was second-order accuracy in both time
and space. It uses explicit Finite Volume Scheme (FVS). In order to solve the parabolic convection
equations a monotone upwind schemes for conservation laws were used. A hybrid Harten-Lax-van
Leer/Harten-Lax-van Leer contact Riemann solver was employed to evaluate the numerical flux.
To identify the effect of cavity on the flow over the cylinder without shock interaction a baseline
measurement was examined with the impinging shock removed. Fig 3.19 & 3.20 represents this re-
sults. Typical behaviour of pitot pressure in a shock tunnel was observed with no large oscillations
(Fig 3.19b). Fig 3.20b also shows that the base pressure approaches constant (this is obtained in
numerical simulation). This implies that in their study flow over the cylinder with forward-facing
cavity was steady in nature.
28
Figure 3.20: a) Numerical shadowgraph image, b) Numerical pressure-time history.
Quasi-steady Type IV interaction was recognized from experimental as well as numerical shad-
owgraph. Fig 3.21a shows that rather than impinging on the surface, the supersonic jet deflected
upward. This is known as Type IVa shock interaction. Fig 3.21b shows that supersonic jet im-
pinges on the lower surface which makes this interaction different from the earlier one. However
the disturbances created by the shear layer is not enough strong to cause large oscillations in the
flow field structure.
Figure 3.21: Type IVa shock interaction: the jet grazes the upper surface.
Figure 3.22: Type IV shock interaction: the jet impinges on the lower surface.
29
Forward-backward mode was observed by moving the shock-intersection point down (Type IVa).
Instantaneous moments when the flow surface pressures are comparable with experimental shad-
owgraph images and the numerical temperature contours are indicated by the dots on the plot.
Average value of pressure and the oscillation frequency from numerical simulation shows good ar-
gument with experimental data. Flow morphology near the interaction region with instantaneous
Figure 3.23: Pressure histories in the cavity (left: Experimental, right: Numerical):
forward-backward oscillation mode.
Figure 3.24: Set of time-sequential shadowgraph images of unsteady Type IVa shock
interaction: forward-backward oscillation mode.
temperature contours and fluid streamlines are shown in Fig 3.25. Arrow is used here in order
to identify the direction of motion of the bow shock and the jet. Fig 3.25a and 3.25c depicts the
situation when pressure at the base of the cavity reaches its peak and valley respectively. Fig 3.25a -
3.25c correspond to the experimental results of labels t1, t2 and t4 given in Fig 3.24. Supersonic jet
comes closer to the upper wall of the cylinder compared to the quasi-steady Type IVa interaction,
while forward-backward mode occurs. Fig 3.25a shows that only a small part of the fluid in the
supersonic jet enters into the cavity which in turn leads to rise of pressure in the cavity. Pressure
rise causes disturbance in the flow field and this disturbances travel upstream until it encounters the
bow shock wave. In Fig 3.25b it can be seen that bow shock is moved upstream by small amount
of amplitude, at the same time jet moves upstream by curving upward away from the cavity like
30
Type IVa configuration (Fig 3.25c). As a result the high pressure fluid gets piled in the cavity in
order to flow away rapidly. Just after this situation there is a drop in pressure inside the cavity
that makes the supersonic flow and bow shocks to move downstream back to their original position.
Next oscillation cycle will begin as soon as the part of the fluid in the jet enter the cavity again.
In this way both the shocks and jets move forward and backward in front of the cavity. Though
due to this oscillation mode, amplitude of the cavity pressure oscillation don’t vary significantly.
Backward-forward mode has only small effect on its amplitude. However this forward-backward
oscillation is associated with high frequency.
As shock interaction point is moved to the downward direction, other mode of oscillation occurs.
In this case major part of the supersonic jet enters into the cavity. In this case also experimental
and computational results shows good agreement. Though the amplitude of pressure oscillations
of numerical simulation is larger than the experimental one as observed in forward-backward mode
due to some constraints. In Fig 3.27 labels t1 and t3 depicts the situation when pressure at the
base of the cavity reaches its peak and valley respectively.
The point at which shock wave and supersonic jet interacts with each other, move up and down in
this oscillation mode while in forward-backward mode the bow shocks move forward and backward
in front of the cavity. Fig 3.27 (labels t1-t3) denotes this up-down process and labels t4-t6 shows
that how both shock and supersonic jets move back to their original positions. Like Fig 3.27 (labeled
by t1 and t3) Fig 3.28a and 3.28c also represents the when pressure at the base of the cavity reaches
its peak and valley respectively. Experimental results (labels t1-t3) which is shown in Fig 3.27 are
corresponding to the numerical results of Fig 3.28a - 3.28c. In this oscillation mode the whole jet
impinges on the cavity (Fig 3.28a). Due to this impingement, there is a huge rise in pressure inside
31
the cavity which in turn results strong interference between the supersonic jet the cavity is formed.
Disturbances created by rapid rise of pressure in the cavity moves upstream and causes unsteadiness
in the bow shock in the upstream region. Apart from the unsteadiness, this disturbances forces
the bow shock to move further upstream which increases the width of the supersonic jet and cause
more severe interference between jet and cavity. As soon as the jet impinges on the lower surface,
the pressure in the cavity reduces that results the bow shock and jet to move back to their original
positions. Next oscillation cycle will occur when upward-moving jet impinges on the cavity again.
Here back and forth movement of bow shock is associated with larger amplitude compared to the
earlier mode. However frequency is lower due to this large amplitude.
32
Referenecs
1. Mason, L. M. & Berry, A. S. “Global Aeroheating Measurements of Shock-Shock Interac-
tions on a Swept Cylinder”, NASA, Langley Research Center, Hampton, VA, 236814
2. Fuata, J. “Quantifying the Effects of Convective Aerodynamic Heating over a Double Swept-
Wedge Configuration at Mach 6”, Bachelor of Technology Dissertation, The University of Southern
Queensland, 2004.
3. Albertson, A. W. & Venkat, V. S. “Shock Interaction Control For Scramjet Cowl Leading
Edges”, American Institute of Aeronautics and Astronautics, 2005.
4. Glass, C. E. “Computer Program To Solve Two-Dimensional Shock-Wave Interference Prob-
lems With an Equilibrium Chemically Reacting Air Model”, Langley Research Center, Hampton,
Virginia, NASA Technical Memorandum 4187.
5. Mahapatra, D. & Jagadeesh, G. “Shock tunnel studies on cowl/ramp shock interactions in
a generic scramjet inlet”, JAERO365 Âľ IMechE 2008.
6. Chettle, A., Erdem, E. & Kontis, k. “Edney IV Interaction Studies at Mach 5”, University
of Manchester, Sackville St. M13 9PL, Manchester, United Kingdom.
7. Paciorri, R. & Bonfiglioli, A. “Numerical Simulation of Shock-Shock Interactions with an
Unstructured Shock-Fitting Technique”, Proc. ‘The 6th European Symposium on Aerothermody-
namics for Space Vehicles’, Versailles, France, 3 - 6 November 2008 (ESA SP-659, January 2009).
8. Delery, J. “Handbook of Compressible Aerodynamics”, Wiley, 2010.
9. Olejniczak, J., Wright, M. J. & Candler, G. V. “Edney IV Interaction Studies at Mach
5”, University of Manchester, Sackville St. M13 9PL, Manchester, United Kingdom.
10. Moss, J. N., Pot, T., Chanetz, B. & Lefebvre, M. “DSMC Simulation of Shock/Shock
Interactions: Emphasis on Type IV Interactions”.
11. Xiao, F., Li, Z., Zhu, Y. & Yang, J. “Hypersonic Type-IV Shock/Shock Interactions on a
Blunt Body with Forward-Facing Cavity”, Journal of Spacecraft and Rockets.
33