You are on page 1of 19

Shock Tube

(Brief introduction on its theory and applications)


by Grazia Lamanna
Institut fr Thermodynamik der Luft- und Raumfahrt
Universitt Stuttgart

1.0 Introduction
A shock tube is a device for generating, in a relative simple and low-cost manner, gas flows or gas
conditions that are difficult to achieve in other test devices. By its nature, the shock tube produces these
conditions for a very short duration. In its simplest configuration, it takes the form of a long smooth wall
steel pipe, of either circular or rectangular cross-section divided into two compartments separated by a
diaphragm of thin material, as shown in Fig. 1.1. The shorter section of the tube is at higher pressure and
is termed the driver section. The longer part of the tube is at lower pressure and is called the driven
section. The gas in the driver and driven section need not to be the same, and they can also be at different
temperatures.
Diaphragm
Driver Section (High Pressure)

Driven Section (Low Pressure)

Figure 1.1. Schematic representation of a shock tube


When the diaphragm is removed rapidly, for example by bursting it, a flow of short duration is established
in the tube and a compression wave travels into the low-pressure (driven) section. The front face or
leading edge of this pressure wave acts as the head of a piston of gas, which drives a pressure pulse
ahead and rapidly develops into a shock wave, as shown in Fig. 1.2.

Figure 1.2. Transition from a pressure pulse to a shock wave


At the same time, a train of rarefaction (expansion) waves travels into the high-pressure (driver) section
(see Fig. 1.3). The flow regions induced by these two waves are separated by an interface, also named
contact surface (or discontinuity), across which the pressure and speed are equal, but the density and

temperature are quite different. This difference in density and temperature is due to the opposite actions
performed by the compression and expansion waves, respectively. The first one heats and compresses the
shocked driven gas, the latter cools down and decompress the expanding driver gas. Incidentally, we
observe that the contact surface is also the interface between the driver and driven gas, hence different
gases may be present on either side of the contact surface as well. Provided the shock tube is of constant
cross section, the shock wave will remain unattenuated with distance, and the pressure and particle
velocity will be uniform along the cross section and constant over a certain region behind the shock. It is
this property of the shock tube to provide both a controlled thermodynamic state and a one-dimensional
gas flow, which makes it an invaluable tool in many investigations.

Figure 1.3. Wave system occurring in a shock tube upon bursting of the diaphragm: a phenomenological
representation.
The objective of this report is to provide an understanding of the principles of operation of a shock tube
and an account as to some of the difficulties encountered during the design, realization, and operation of
such devices. To that aim, the report starts with a brief survey on the propagation of finite disturbances in
a quiescent gas, which is an essential preliminary to the understanding of the mode of operation and usage
of shock tubes. The difference in the mechanism of propagation between finite disturbances and sound
waves (i.e. small disturbances) is also emphasised. Subsequently, this report illustrates the calculation of
the experimental conditions as functions of the initial thermodynamic state in the driver and driven
section, focusing in particular on the theoretical interdependence among all parameters of a shock tube.
Special attention is also paid to the generation of the tailored-condition, which is of relevance to many
practical applications.
To conclude this section, we will provide a brief overview of some of the most important investigations,
which avail themselves of the shock tube technique. Thanks to its general simplicity, versatility, and
relative cheapness, it has been used to study a wide range of physical problems. Although an attempt has
been made to give a balanced account, the survey presented here is by far not complete and reflects the
particular interest of the author. For more details, the reader is referred to the copious literature available
on the subject [1, 2, 3, 4]. One of the first applications of the shock tube is in the development of
instrumentation for measuring some physical properties of gases. For example, already in 1943, Reynolds
employed a shock tube to calibrate piezoelectric pressure gauges. Other typical investigations, which have
been traditionally conducted in a shock tube, are the reflection, refraction, and diffraction of shock waves,
and studies on fundamental flow instabilities. Figure 1.4 shows a shock wave propagating through a
layered medium. The interface is perturbed by a (baroclinic) instability, which rapidly develops into a
region of turbulent flow.

Turbulent Flow

Shock

Figure 1.4. Example of a flow instability studied in a shock tube. Courtesy of the Fraunhofer Institute for
High Speed Dynamics.
Figure 1.5 illustrates a special case of shock wave diffraction, where a very complex pressure distribution
can be observed as consequence of an interior detonation. This colour schlieren-photograph is part of an
extensive study on typical elements of a building structure.

Figure 1.5. Side view of a shock wave propagating along a staircase. Courtesy of the Fraunhofer
Institute for High Speed Dynamics.
Shock tube applications include also studies of break-up of liquid droplets by a shock wave, determination
of ignition delay times and rate of chemical reactions, nucleation processes, and medical applications,
such as lithotripsy. Shock tubes have also been used to study the properties of gases and extend their
equation of state at high-pressure and temperature conditions. Furthermore, by a suitable tuning of the
initial conditions, it is possible to achieve temperatures as high as 20000 K behind the shock front. In this
case, the distance behind the shock in which steady-state conditions are attained gives valuable
information about relaxation effects and ionization properties of gases. Last but not least, the shock tube
can be also employed as a low-cost variant of an intermittent wind tunnel for the analysis of hypervelocity
superorbital flows, aerothermodynamic processes, and scramjet propulsion systems, just to name a few.
Figure 1.6 shows schematically the operating principle of such a shock tunnel. The gas in the driver
section is compressed gradually by the downstream moving piston, until the bursting pressure of the
diaphragm is exceeded. At this point, a shock wave propagates into the driven section followed by a
particle flow. As the shock wave reaches the end wall, it is reflected again as a shock wave, leaving behind
a high-pressure high-temperature zone, which may act as a reservoir for the nozzle flow. The apparatus is
arranged so that the flow from the driven tube passes directly into the nozzle, where it can be accelerated
to hypersonic speed. The duration of the flow is equal to the time between the arrival of the shock and the
contact surface at the nozzle entry.

Figure 1.6. Schematic illustration of the operating principle of a shock tunnel. Courtesy of Frank
Houwing The Australian National University.

2.0 Theory of compressible flows


This section discusses the basic principles of the theory of compressible flows, with particular emphasis
on the propagation of finite disturbances and shock waves. This is an essential preliminary to the
understanding of the mode of operation and usage of shock tubes. We start off from the equations of
motion of a compressible fluid, using the Eulerian approach. In the hypothesis of one-dimensional,
inviscid flow and perfect gas behaviour, they read:
u

0
t
x
u
u
1 p
u

t
x
x

where

p RT .

(2.1)

Assuming there is no heat exchange and all changes in the flow occur in a sufficiently slow manner that
all irreversible processes can be neglected, the flow is also isentropic. The fundamental thermodynamic
relation can then be written as:
dU dQ pdV .

For a perfect gas and in specific terms, it becomes cv dT Tds pdv, hence
1
,
ds cv dT pd

(2.2)

Since, for a perfect gas, with ratio of specific heat ,


c p cv R

we may rewrite the equation of state as:

and c p c v ,

p
1
1
pd
0
d
1

which simplifies to the normal adiabatic expression that, for a given fluid element, reads p =
constant. The set of equations (2.1) and (2.2) can be used to determine the properties and speed of
propagations of waves of finite amplitude.

2.1 The propagation of finite disturbances


The terminology finite disturbance denotes a perturbation in the thermodynamic state of a quiescent gas,
where the associated pressure and density variations are of the same order of magnitude as the local values
of pressure and density, that is p = O(p) and = O(). When the strength of the disturbance is much
lower than p (i.e. p/p << 1), then it is a well-known result in acoustic theory that small disturbances (i.e.
sound waves) propagate at a speed of sound c:
p

c 2

c RT

Two other important conclusions from the theory of propagation of sound waves are: 1) there is no particle
flow across an acoustic wave; 2) the thermodynamic state of the gas is not altered by the passage of a
small disturbance. The same conclusions do not hold in the case of finite disturbances. In particular, we
will show that, when the strength of the disturbance becomes of the order of p (i.e. waves of large
amplitude), the following effects take place and affect the speed of propagation: 1) a particle flow is
generated in the wave; 2) the gas is compressed /expanded by the passage of a compression/rarefaction
wave, respectively; 3) the subsequent temperature increase/decrease produces locally an increase/decrease
of the sound speed. Hence, the peak of the disturbance will travel faster/slower than the speed of sound
in the undisturbed gas, thus resulting in a steepening/broadening of the propagating pressure pulse. For the
mathematical analysis of these effects, it is convenient to introduce the quantity f:

d
.

f c
o

(2.3)

The quantity f has the dimension of a velocity [m/s] and, in the hypothesis of isentropic flow, represents
the total variation of sound speed due to the propagation of the disturbance /. This assertion can be
easily verified in the case of a perfect gas and isentropic flow assumption. In fact, using the isentropic
relation in the form p = const and remembering c = ( p/), we can integrate equation (2.3) to
f

2co

p
p
o

(2.4)

Combining the relations p = const and c = ( p/), the sound speed may similarly be written as
p
po

c co

Substituting relation (2.5) in Eq. (2.4), we obtain

1 2

(2.5)

2
c co ,
1

(2.6)

which confirms our initial interpretation of the quantity f. It is noteworthy noticing that f 0 when p
po. That is, in the limit of small disturbances, there is no variation of speed of sound across the wave (i.e.
acoustic waves propagate at uniform speed).
By introducing the definition of f in Eqs. (2.1) and remembering that
f
c

x
x

and

f
c

,
t
t

we can rewrite the equation of motion (2.1) as

and

f
u
f
c
u
0
t
x
x

(2.7)

u
u
f
u
c
0.
t
x
x

(2.8)

In deriving Eqs. (2.7) and (2.8), we assumed that the flow is isentropic and the gas is initially at uniform
temperature and pressure, so that the energy equation takes the form p = const. In this case it holds
p p

c2
c
.
x
x
x
x

By adding and subtracting Eqs. (2.7) and (2.8), we obtain

and

f u
f u
u c
0
t
x
f u
f u
u c
0.
t
x

(2.9)

(2.10)

These relations represent the equations of propagation of waves in the direction + x with speed u + c and
u - c, respectively. Specifically, Eq. (2.9) indicates that for a particle traveling with a velocity u + c, the
value of the quantity f + u will be constant. Similar considerations hold for Eq. (2.10) as well. The
quantities f + u and f u are known as Riemann invariants. The curves in the x,t space, along which these
invariants are constants, are known as characteristics. From Eqs. (2.9) and (2.10), we can draw another
important conclusion: during the propagation of a finite disturbance, a particle flow with velocity u is
generated in such a way that a reversible conversion from internal to kinetic energy takes place along a
characteristic curve.
In order to examine in more details the propagation of a pressure disturbance and determine an expression
for the particle velocity u, let us consider a uniform stationary gas where at the instant to a finite pressure
pulse is applied, as shown in Fig. 2.1a. In time (Fig. 2.1b) the pulse will have separated into two parts,
specifically that traveling with speed u c for which f u is constant and that traveling with speed u + c
for which f + u is constant. A particle B, which is initially in the undisturbed region, will eventually be

overtaken by the disturbance traveling with speed u + c (see Fig. 2.1c). At this point, the following two
relations must be simultaneously verified for the particle B: f + u = const and f - u = 0. This follows
from the fact that the particle B belonged originally to the undisturbed gas region, where f = 0 and u = 0,
and is overtaken by the whole disturbance. Therefore, we obtain:
2co
u f

p
o

1 .

(2.11)

(a)

x
(b)

(c)

Figure 2.1. The propagation of a finite pressure pulse.


Equation (2.11) shows that the propagation of a finite disturbance induces a particle flow, with velocity u.
In particular, the particle speed is directed in the direction of propagation in a compression wave (p > p0)
and against the direction of propagation in a rarefaction wave (p < p0).
Furthermore, the speed of propagation of the wave f + u can be deduced from Eqs. (2.5) and (2.11) and it
may be written as:
1 p

u c co
1 po

2
.
1

(2.12)

From Eq. (2.12), it follows that when the wave is weak (i.e. p => p0) then the velocity of propagation tends
to c0, that is the speed of sound of the undisturbed gas. For strong disturbances, it follows that waves may
be propagated at speeds considerably greater than the speed of sound. Furthermore, waves may also
change their shapes as they propagate. This effect can be inferred directly from Eq. (2.12), since it follows
that the lower the value of p, the lower the velocity of propagation. This implies that different portion of a
wave will necessarily travel at different speeds, specifically rarefaction wave will become less steep as
they propagate; while compression waves will steepen as they propagate (Fig. 2.2a and 2.2b).
P
(a) Rarefaction wave
P0

time
x
P
(b) Compression wave
time

Shock formation
P0
x
Figure 2.2. The propagation of a (a) rarefaction and (b) compression wave, respectively.

2.2 Theory of shock waves


As shown in the previous section, the steepening process, occurring during the propagation of a
compression wave, may lead to the formation of a shock wave. This means that, at a given time, the
velocity and temperature of the gas change so rapidly with distance that the hypothesis of isentropic flow
no longer holds and one has to revert to a separate treatment to describe the behaviour of shock wave. It is
not the purpose of this section to provide a detailed treatment of shock theory, for a more comprehensive
account the reader is referred to any of the standard gas dynamics textbook [5]. A detailed theory shows
that a shock front is a stable system in which temperature, pressure, density and flow velocity change their
values in a few mean free path lengths. Let us consider, without loss of generality, a shock front moving
into still air with velocity U. The gas properties behind and ahead of the shock will be denoted by the

subscripts 0 and 1, respectively, as shown in Fig. 2.3. In the evaluation of the conditions across the
shock it is convenient to represent the flow conditions as seen by an observer travelling at the same speed
as the shock front (Fig. 2.3).
Shock Front
P1

P0

U-u

Figure 2.3. The flow conditions in a shocked gas as seen by an observer travelling at the same speed as
the shock front.
By applying the conservation laws at the fluid volume EF, which contains the shock wave, we obtain the
so-called Rankine Hugoniot relations:

0U 1 U u m
momentum mu p1 p0
1
1
2
energy
me0 mU 2 p0U me1 mU u p0 U u
2
2
mass

(2.13)

These relations allow to determine the properties of the gas behind the shock wave, provided that the state
of the gas ahead of the shock, the fundamental thermodynamic relation (i.e. the function e = e(p, )) and
one parameter of the gas behind the shock are specified. In the hypothesis of a perfect gas, the
fundamental thermodynamic relation takes the form:
e

p
.
1

(2.14)

By rearranging Eqs. (2.13) and (2.14) in terms of the gas properties behind the shock, we obtain:

c0

p10
;
1

T1
1 p10
p10
;
T0
p10
u

c0

1 p10 1
1 p10

p10

p1
p0

- 1
1

(2.15)

It is noteworthy noticing that as p10 tends to one, the shock wave tends to behave as an acoustic wave, i.e.
M 1 the shock wave travels at the speed of sound, and u 0 no particle flows is generated. Some
properties of shock waves are shown in Fig. 2.4 a and b.

monoatomic gas

monoatomic gas

P10

P10

(a)

(b)

Figure 2.4. (a) Ideal gas Mach number and (b) temperature ratio as function of shock strength. Courtesy
of J.K. Wright, Shock Tubes, John Wiley &Sons, 1961.
As it can be inferred from Fig. 2.4, in a polytropic gas, the higher number of internal degree of freedom
results, when compared to a monoatomic gas, in a further increase of kinetic energy in the shocked gas.

3.0 Theory of the Shock Tube


Driver

3.1 Elementary shock-tube theory

Driven
(a)

Section 1.0 presents a phenomenological description of the processes occurring in a shock tube. In this
section, we re-examine these processes from a more rigorous mathematical perspective. Figures 3.1 (a)
Fanshockttube assembly and its x-t diagram, respectively.
Reflected Shock
and (b) show theExpansion
most simple
When the diaphragm
is shattered, a shock wave travels into the driven section while a rarefaction
wave
travels
back into the
5
Contact
Surface
compression chamber. The pressure and temperature distributions along the tube, at an instant t1, are
3 line in Fig. 3.1 (c), which
displayed in Fig. 3.1 (c) and (d). The dotted
2 separates region oft1gas at pressures
P3 and P2, denotes the position occupied by that gas which was originally at the diaphragm. The(b)
gas to the
right has been compressed and heated by the shock wave; the gas to the left of this line has been expanded
Shock
4
1 general be a change of gas
and cooled by the rarefaction
wave. At this position there will, therefore,
in
x
type, temperature and density [see Fig. 3.1 (d)], while velocities and pressures are the same
on both sides.
Such point is known as contact discontinuity (or surface). When the shock wave reaches the end of the
P4 a reflection. The theory of shock reflection is discussed in section 3.2.
shock tube, it undergoes
P3

P2
P1

(c)
x

T2
T4

T3

T1

(d)

x
Figure 3.1. Wave diagrams in a shock tube.
Let us denote pressures by p, densities by , sound speed by c and flow velocities by u. Referring to Fig.
3.1 (a), the region ahead of the shock is indicated with 1, the region comprised between the shock and
the contact surface is denoted with 2, ahead of the rarefaction wave travelling in to the driver section is
region 4, and between the tail of the rarefaction wave and the contact surface is region 3. Assuming
that the diaphragm plays no part in the process, once it has shattered, and perfect gas behaviour, the flow
conditions in the shock tube can be determined as follows:
1. the initial conditions are assigned: p4 , T4 , 4 ; p1 , T1 , 1 ;
2. the flow velocity u2 behind the shock can be determined via the shock relation (2.15)

u2

c1

1 1 p21 1
1 1 p21 1 ;

3. the flow velocity u3, behind the rarefaction wave, is given by Eq. (2.11), taking into account that
the flow is in the opposite direction of propagation of the expansion wave
4 1 2 4

u3
2 p3

1 ;

c4 4 1 p4

4. since there must be no discontinuity of pressure or flow speed across the contact surface, we have
p2 = p3 and u2 = u3, these equations now give
1 2
c1
1 1 p21 1 2 1 p1 p 4 4 .

(3.1)
c4
1 1 p21 1 4 1 p4 21

Equation (3.1) relates the shock strength p21 to the pressure ratio p4/p1 in the compression and expansion
chambers, as shown in Fig. 3.2. It is noteworthy noticing that, for any practical purposes, the pressure
ratio p21 is never much higher than 10 when we employ the same gas mixture in the driver and driven
section.

p2/p1
10
c1/c4 = 1

p4/p1
1

10

50

100

500

Figure 3.2. Shock strength as function of initial pressure ratio, using air in both sections.
By solving relation (3.1), we can derive p2 and subsequently all other flow properties in the shock tube,
using the following relations:

MS

p21 1
;
1 1

MR MS

1 1 2 p21

1 p21 1
p p
1
21 21

1 1 p21

T2 p21 p21 1

T1 1 p21 1

(3.2)

where MS and MR are the Mach numbers of the incident and reflected shock wave, respectively. The shock
Mach number MS is defined as US/c1, where US is the wave speed of the shock, and c1 is the speed of
sound in the undisturbed gas in the driven tube, i.e. MS is the Mach number relative to the stationary
driven gas. The shock Mach number MR is defined as MR = (UR+ u2)/c2, where UR is the speed of the
reflected shock (relative to the shock tube), and u2 is the speed of the flow induced by the incident shock
(relative to the laboratory reference frame). Thus, MR is the wave speed relative to the oncoming gas
heated by the incident shock (with sound speed c2). The dependence of MS and MR as function of the
pressure ratio p41 is shown in Fig. 3.3.

P41
Figure 3.3. Variation of incident and reflected shock Mach number over a small range of p41, for a shock tube
with air as both driver and driven gas and T4/T1 = 1. Courtesy of J.K. Wright, Shock Tubes, J. Wiley, 1961.

In order to examine the effect of different gas combinations, let us assume a large initial compression
ratio:

p4
1 1 p21
p1
p4

4 1

2 4

0.

Under this assumption, Eq. (3.1) reduces to:

c1

c4

1 1 p21(max) 1
1 1 p21(max) 1

2
,
4 1

where p21(max) is the maximum theoretically achievable shock strength. If p21(max) is large, this expression
further simplifies to

2 1 c
1 1 2 4 .
4 1 c1
2

p21(max)

(3.3)

It follows that, in order to obtain strong shocks in the shock tube (i.e. high values for p21 and MS), it is
necessary to make the sound ratio c4/c1 as high as possible and also to choose a driver gas with (4 - 1) as
MR
small as possible.
The same effect can also be recognised in Fig. 3.4, where the dependence ofMthe
S shock
Mach number as function of the pressure ratio p41 is shown for different driver/driven gas combinations.

P41
Figure 3.4. Variation of incident MS and reflected MR shock Mach number for various
driver/driven gas combinations as function of initial driver to driven gas pressure ratio p41.
Courtesy of H. McMahon et al. Georgia Tech School of Aerospace Engineering.

Figure 3.4 shows that, for a given temperature and pressure ratio, it is possible to increase the shock Mach
number by choosing increasingly lighter driver gases. For a given driver/driven gas combination, the same
effect can be obtained by increasing the initial temperature ratio (T4/T1), i.e. the initial sound speed ratio.

3.2 Shock reflection against a wall


In this section, we discuss the reflection of shock waves against a wall, which is fundamental to the
understanding of the operation of a shock tube in reflected mode. In Fig. 3.5 we have a shock wave of
strength p21 traveling from left to right. As it hits the rigid wall, it is reflected as a shock wave traveling in
the opposite direction and of strength p52, where p5 is the total pressure behind the reflected shock. We
may immediately write down the shock relations for the two waves. With reference to Eqs. (2.15), we
have, for the incident and reflected wave:

u2
p 1
1 21 1 2 ;
c1 1 p21 1

2
- 1

(3.4)

u5
p 1
1 52 1 2 ;
c2 1 p52 1

(3.5)

c2
T2
1 1 p21

p21
c1
T1
p21

and

12

(3.6)

Shock
5

Shock
2

2
1
x
Figure 3.5. The normal reflection of a shock wave at a rigid wall.

Since the shock is being reflected at a rigid wall, the magnitude of the flow velocities in the two waves
must be equal since gas particles must be at rest behind the reflected shock, i.e. u2 = u5. Combining the
latter condition with Eqs. (3.4), (3.5), and (3.6) we get

p21 1
1 p21 11 2

1 1 p21
p21
p21

12

p52 1 .
1 p52 11 2

This relation, after rearrangement, yields

p52

p21 1 2 1
.
p21 1

(3.7)

We can now easily deduce an expression for the reflected pressure in terms of the incident overpressure

p5 p1 2 1 1 p21 1

.
p2 p1
1 p21 1

(3.8)

In the limit of very weak shock (p21 1), this ratio tends to two, which is a classical result of elementary
acoustic, i.e. weak shocks behave like sound waves. For stronger shocks, the reflected overpressure to
incident overpressure ratio increase steadily with shock strength tending asymptotically to the value 2 +
1/ for very strong shocks. For air ( = 1.4), the asymptotic overpressure tends to 6.

3.3 Interaction between a shock wave and contact surface


This subsection discusses the refraction of a shock wave as it crosses the interface between two gases (i.e.
a) to study is that of normal refraction, where the direction
b)
contact discontinuity). The simplest case
of
propagation of the incident shock is normal to the plane of the boundary (as in the case of a simple shock
tube). It is possible to distinguish two distinct types of normal refraction that when there is a reflected
shock
Transmitted
and that when there is a reflected expansion wave. Figures 3.6 a) and b) show the two cases
Transmitted
Shock Figure 3.6 c) shows a third
schematically.
option, the so-called
tailored interface, where no disturbance is
Reflected
Reflected
Shock
reflected from the contact surface back towards
contact
Shocks the rear wall of the shock tube. The tailoredExpansion
surface configuration offers a number of advantages when applied to the operations of shock tubes,
namely it permits
an increase
Contact
Surface in the testing-time and an improvement in the homogeneity of the working
gas parameters (i.e. it decreases possible contamination effects in the test section caused by the driver
gas).
Shock
Shock

c3 c2

c3 c2

c)
Tailored Interface
Transmitted
Shock

Shock

c3 c2

Figure 3.6. Wave diagrams for three different


shock/contact surface interactions: a) refracted
shock; b) refracted rarefaction wave; c) tailored
interface. Air is used as gas both in the driver
and driven sections.

The general theory of normal refraction consists in writing down expressions for the particle velocity on
either side of the gas interface in terms of the unknown pressures behind the transmitted shock and
reflected shock or rarefaction wave. The boundary conditions to be satisfied at the interface are that
pressure and particle velocity should be the same on either side. This gives enough information for a
unique solution to be obtained. The tailored condition is determined by further imposing the equality in
pressure between region 6 and 5 (see Fig. 3.7 for reference).
Distinctive methods for computing the tailored interface region, which require successive determination
of the gas parameters in different flow domains (see Fig. 3.7 (b)), are discussed in [1, 6, 7]. This section
presents a simplified treatment for the calculation of the tailored interface condition, under the
assumption that the driver and driver gas are the same (i.e. 1 = 4 = ). This simplifying assumption is
easier to understand and gives less lengthy algebraic expressions without any loss of generality.

a)

b)

US4

Transmitted
Shock
3

Reflected
Shocks

Contact Surface
Shock

US3

3
2

US2
1

US1

Figure 3.7. Wave diagram for the refracted shock configuration with indication of the different (a)
regions and (b) wave speeds.

Recalling Eq. (3.4), we have for the shock transmitted through the contact surface

u3 u7 p73 1

.
c3
p73 11 2

(3.9)

Since u5 is zero, we have, if the incident shock is reflected from the contact surface as a shock

u6
1 2
p65 1 p65 1 ,
c5

(3.10)

and, if it is reflected as an expansion wave

u6
1

;
1 p65
c5

-1
.
2

(3.11)

Since u6 is also equal to the velocity of the contact surface after the interaction with the normal shock,
Eqs. (3.10) and (3.11) show that the contact surface moves towards the shock tube rear-wall if the
reflected disturbance is a shock, and away from it if the reflected disturbance is an expansion wave [see
Fig. 3.6 a) and b) for clarity]. Furthermore, since u6 = u7 (contact surface condition), Eqs. (3.9), (3.10) and
(3.11) yield

p73 1 p73 1

u3
1 2
c53 p65 1 p65 1
c3

u3
1

c53
1 p65
c3

1 2

for a reflected shock, and

p73 1 p73 1

1 2

(3.12)

(3.13)

for a reflected expansion. Equality of static pressure across the contact discontinuity requires that for
reflection as a shock or as an expansion
p73 p65 p53 ,
(3.14)
p53 can be calculated from relation (3.7), being p53 = p52. By substituting Eq. (3.14) in either Eq. (3.12) and
(3.13), it is possible to determine a unique (i.e. physically realistic) value for p73 as function of the incident
shock strength.
To determine the tailored condition, we have further to impose that the disturbance reflected from the
contact surface is a Mach wave, i.e. p65 = 1. Since p2 = p3, Eq. (3.14) shows that p73 = p52. Substituting the
latter relation in Eqs. (3.12) or (3.13) yields

u3 p52 1 u2

.
c3 p52 11 2 c2

(3.15)

Since u2 = u3, equation (3.15) implies c2 = c3. The condition for a reflected wave of zero strength is,
therefore, that the speeds of sound on both sides of the contact discontinuity are equal. In physical terms,
this may be explained as follows. The reflected shock travels against a gas stream of uniform speed and
pressure, in which the speed of sound changes suddenly at the contact surface. If the speed of sound falls
(c2 > c3), the shock Mach number and pressure ratio increase on crossing the contact surface, so that the
static pressure behind the contact surface will tend to be higher than ahead of it. A reflected shock is,

therefore, required to restore the condition of equal static pressure on either sides of the contact
discontinuity, see Fig. 3.6 a) for clarity. Conversely, if the speed of sound increases at the contact surface
(c2 < c3), the shock Mach number falls as it passes the contact discontinuity, and a reflected expansion
wave is needed to restore the equilibrium of pressures. Only when c2 = c3, can the reflected disturbance be
a Mach wave (i.e. p65 = 1) and the transmitted shock as the same speed as the incident shock. Rigorously
speaking, this result is strictly valid only when we have the same gas in the driver and driven sections. In a
more general case, there is a slight dependence of the ratio c23 on the incident Mach number [7].
The speed of the reflected shock wave is given in terms of its pressure ratio by Eq. (2.15) as
1
US3
p65 1 2 ,
c5

(3.16)

and the head of the reflected expansion travels at the speed of sound c5. The velocity US4 of the transmitted
shock also follows from Eq. (2.15) as
1
US4
u
p73 1 2 - 3 .
c3
c3

(3.17)

The functional dependence of the wave speeds in terms of the Mach number of the primary shock MS1 is
shown in Fig. 3.8. It is noteworthy noticing that for a given driver/driven gas combination, there exists
only one primary shock Mach number in correspondence of which the tailored interface condition can
be attained. In order to established tailored conditions in correspondence of a different primary shock
Mach number, it is necessary either to heat up the driver/driven gas or the modify the of the driver gas
by modifying its mixture-composition.

Figure 3.8. Variation of the wave speed for the incident US1, reflected US3, and transmitted
disturbances US4 as function of the primary shock Mach number MS1. Courtesy of D.H. Holder
& D.L. Schultz ARC Reports & Memoranda, 1960.

References
[1] H. OERTEL (1966), Storohre, Springer-Verlag, Wien.
[2] D. BERSHADER AND R. HANSON (1985), Shock Waves and Shock Tubes, Proceedings of the 15th
International Symposium on Shock Waves and Shock Tubes, Berkeley, California, July 28-August 2,
Stanford University Press
[3] H. GRNIG (1987), Shock Tubes and Waves, Proceedings of the 16th International Symposium on
Shock Tubes and Waves, Aachen, Germany, July 26-31, VCH Publishers.
[4] Y.W. KIM (1989), Current topics in Shock Waves, Proceedings of the 17th International Symposium
on Shock Waves and Shock Tubes, Bethlehem, Pennsylvania, AIP Conference Proceedings 208
[5] P.A. THOMPSON (1972), Compressible Fluid Dynamics, McGraw Hill.
[6] D.W. HOLDER AND D.L. SCHULTZ (1960), On the flow in a reflected-shock tunnel, Aeronautical
Research Council (ARC) Reports and Memoranda, No. 3265.
[7] R.L. PETROV (1976), Analysis of the tailored contact surface region in a shock tube, pp. 11061110, M.L. Kalinin Leningrad Polytechnic Institute. Translated from Inzhenermo-Fizicheskii
Zhurnal, Vol. 31, No.3, pp. 537-542.

You might also like