You are on page 1of 39

Compressible flow

The course of compressible flow / gas dynamics is concerned with the

causes and the effects arising from the motion of compressible fluids
particularly gases.

It is branch of more general subject of fluid dynamics.

Compressible flow involves significant changes in density. It is

encountered in devices that involve the flow of gases at very high speeds.

Compressible flow combines fluid dynamics and thermodynamics

in that both are necessary to the development of the required theoretical

The analysis of flow problems is based on the fundamental principles

given below:

1. Conservation of mass
2. Newtons second law of motion
3. Conservation of energy

The continuity equation for a control volume is

For steady flow, any partial derivative with respect to time is zero and the
equation becomes:

For one-dimensional flow any fluid property will be constant over an entire
cross section.
Thus both the density and the velocity can be brought out from under the
integral sign.

If the surface is chosen perpendicular to V, the integral is very simple to


For steady, one-dimensional flow, the continuity equation for a control

volume becomes

If there is only one section where fluid enters and one section where fluid
leaves the control volume, continuity equation becomes

An alternative form of the continuity equation can be obtained by

differentiating equation. For steady one-dimensional flow this means that

Dividing by AV yields

Momentum Equation
The time rate of change of momentum of a fluid mass equals the net
force exerted on it.

The integral form of equation is

If there is only one section where fluid enters and one section where fluid
leaves the control volume steady one-dimensional flow, the momentum
equation for a control volume becomes:

Energy Equation

The first law of thermodynamics is a statement of conservation of energy.

For a system composed of a given quantity of mass that undergoes a
process, we say that

The transformed equation that is applicable to a control volume is

With enthalpy, the one-dimensional energy equation for steady-in-the-
mean flow is

where q and ws represent quantities of heat and shaft work crossing the
control surface per unit mass of fluid flowing.

Sonic Velocity

A disturbance at a given point creates a region of compressed molecules

that is passed along to its neighboring molecules and in so doing creates
a traveling wave.

The speed at which this disturbance is propagated through the medium is

called the wave speed.

This speed not only depends on the type of medium and its
thermodynamic state but is also a function of the strength of the wave

The speed of waves of very small amplitude is characteristic only of the

medium and its state.

Sound waves are infinitesimal waves (or weak pressure pulses) which
propagate at the characteristic sonic velocity.

Consider a long constant-area tube filled with fluid and having a piston at
one end.

The fluid is initially at rest. At a certain instant the piston is given an

incremental velocity dV to the left.

The fluid particles immediately next to the piston are compressed a very
small amount as they acquire the velocity of the piston.

As the piston (and these compressed particles) continue to move, the

next group of fluid particles is compressed.

The wave front is observed to propagate through the fluid at the

characteristic sonic velocity of magnitude a.

All particles between the wave front and the piston are moving with
velocity dV to the left and have been compressed from to + d and
have increased their pressure from p to p + dp.
For the analysis we choose the wave region as a control volume and
assume the wave front as a stationary wave.

For an observer moving with this control volume, the fluid appears to
enter the control volume through surface area A with speed a at pressure
p and density .

The fluid leaves the control volume through surface area A with speed a
dV, pressure p + dp and density + d.
When the continuity equation is applied to the flow through this control
volume, the result is


Since the control volume has infinitesimal thickness, the shear stresses
along the walls can be neglected.
We shall write the x-component of the momentum equation, taking
forces and velocity as positive if to the right.

For steady one-dimensional flow:


Equations (1) and (2) are now be combined to eliminate dV,

The derivative dp/d is not unique. It depends entirely on the process.

Thus it should really be written as a partial derivative with the appropriate


Weak shock waves (i.e., small disturbances) approach an isentropic

process in the limit.

Therefore, equation of sound can properly be written as

Sound velocity can be expressed in terms of bulk or volume modulus of
elasticity Ev.

Since air is more easily compressed than water, the speed of sound in air
is much less than it is in water.

From Equation, we can conclude that if a fluid is truly incompressible, its

bulk modulus would be infinitely large

Equation can be simplified for the case of a gas that obeys the perfect
gas law:

For perfect gases, sonic velocity is a function of the individual gas and
temperature only. Sonic velocity is a property of the fluid and varies with
the state of the fluid.

We define the Mach number as

If the velocity is less than the local speed of sound, M is less than 1 and
the flow is called subsonic.

If the velocity is greater than the local speed of sound, M is greater

than 1 and the flow is called supersonic.

Wave Propagation

Consider a point disturbance that is at rest in a fluid. Infinitesimal

pressure pulses are continually being emitted and they travel through the
medium at sonic velocity in the form of spherical wave fronts.

To simplify matters we keep track of only those pulses that are emitted
every second.

Now consider a similar problem in which the disturbance is no longer

Assume that it is moving at a speed less than sonic velocity, say a/2.
Figure shows such a situation at the end of 3 seconds.

Note that the wave fronts are no longer concentric. Furthermore, the
wave that was emitted at t = 0 is always in front of the disturbance itself.

Therefore, any person, object, or fluid particle located ahead will feel the
wave fronts pass by and know that the disturbance is coming.

Next, let the disturbance move at exactly sonic velocity. Figure shows this
case in which all wave fronts coalesce on the left side and move along
with the disturbance.

In this case, no region upstream is forewarned of the disturbance as the

disturbance arrives at the same time as the wave front
Now suppose the disturbance is moving at velocity V > a. The wave fronts
coalesce to form a cone with the disturbance at the apex.

This is called a Mach cone. The region inside the cone is called the zone
of action since it feels the presence of the waves.

The outer region is called the zone of silence, as this entire region is
unaware of the disturbance.

The half-angle at the apex is called the Mach angle and is given the
symbol . It should be easy to see that
In the subsonic case the fluid can sense the presence of an object and
smoothly adjust its flow around the object.

In supersonic flow this is not possible, and thus flow adjustments occur
rather abruptly in the form of shock or expansion waves.

Since the supersonic and subsonic flows have different characteristics, it

is suitable to use Mach number as a parameter in our basic equations.

Flow Regimes

It is useful to illustrate different regimes of compressible flow by

considering an aerodynamic body in a flowing gas.

Far upstream of the body, the flow is uniform with a free stream velocity
of V

Now consider an arbitrary point in the flow field, where p, T, , and V are
the local pressure, temperature, density, and velocity at that point.
All of these quantities are point properties and vary from one point to
another in the flow. The speed of sound a is a thermodynamic property
of the gas and varies from point to point in the flow.

If a is the speed of sound in the uniform free stream, then the ratio V/ a
defines the free-stream Mach number M.

Similarly, the local Mach number, M is defined as M = V/a, and varies

from point to point in the flow field.

Consider the flow over an airfoil section as sketched in Figure. Here, the
local Mach number is everywhere less than unity.

Such a flow where M < I at every point, and hence the flow velocity is
everywhere less than the speed of sound is defined as subsonic flow.

This flow is characterized by smooth streamlines and continuously

varying properties.

Note that the initially straight and parallel streamlines in the free stream
begin to deflect far upstream of the body i.e. the flow is forewarned of the
presence of the body.

Also, as the flow passes over the airfoil, the local velocity and Mach
number on the top surface increase above their free-stream values.

However, if M is sufficiently less than 1, the local Mach number

everywhere will remain subsonic.

For airfoils in common use, if M < 0.8, the flow field is generally
completely subsonic.

Therefore to the airplane aerodynamicist, the subsonic regime is loosely

identified with a free stream where M < 0.8.
If M is subsonic, but is sufficiently near 1, the flow expansion over the
top surface of the airfoil may result in locally supersonic regions, as
sketched in Figure.

Such a mixed region flow is defined as transonic.

M is less than 1 but high enough to produce a pocket of locally

supersonic flow.

In most cases, this pocket terminates with a shock wave across

which there is a discontinuous and sometimes rather severe change in
flow properties.

If M is increased to slightly above unity, this shock pattern will move to

the trailing edge of the airfoil, and a second shock wave appears
upstream of the leading edge.

This second shock wave is called the bow shock, and is sketched in

In front of the bow shock, the streamlines are straight and parallel, with a
uniform supersonic free-stream Mach number.

In passing through that part of the bow shock that is nearly normal to the
free stream, the flow becomes subsonic.

However, an extensive supersonic region again forms as the flow

expands over the airfoil surface, and again terminates with a trailing-edge
Both flow patterns sketched in Fig. b and c are characterized by mixed
regions of locally subsonic and supersonic flow.
Such mixed flows are defined as transonic flows, and 0.8 < M < 1.2 is
defined as the transonic regime.

A flow field where M > 1 everywhere is defined as supersonic. Consider

the supersonic flow over the wedge-shaped body in Fig. 1.

A straight, oblique shock wave is attached to the sharp nose of the

wedge. Across this shock wave, the streamline direction changes

Ahead of the shock, the streamlines are straight, parallel, and horizontal;
behind the shock they remain straight and parallel but in the direction of
the wedge surface.

Unlike the subsonic flow in Fig. a, the supersonic uniform free stream is
not forewarned of the presence of the body until the shock wave is

The flow is supersonic both upstream and (usually, but not always)
downstream of the oblique shock wave.
The temperature, pressure, and density of the flow increase almost
explosively across the shock wave shown in Fig. d.

As M, is increased to higher supersonic speeds, these increases

become more severe. At the same time, the oblique shock
wave moves closer to the surface, as sketched in Fig. e.

The incompressible flow is a special case of subsonic flow; namely, it is

the limiting case where M 0.

Since M = V/a we have two possibilities:

The former corresponds to no flow and is trivial. The latter states that the
speed of sound in a truly incompressible flow would have to be infinitely

Use of Mach Number in governing equations

Since supersonic and subsonic flows have different characteristics, it

would be instructive to use Mach number as a parameter in our basic

This can be done easily for the flow of a perfect gas as in this case we
have a simple equation of state
Stagnation Relations

The internal energy and the flow energy of a fluid are frequently combined
into a single term, enthalpy.

Whenever the kinetic and potential energies of the fluid are negligible, as
is often the case, the enthalpy represents the total energy of a fluid.

For high-speed flows, such as those encountered in jet engines, the

potential energy of the fluid is still negligible, but the kinetic energy is not.

In such cases, it is convenient to combine the enthalpy and the kinetic

energy of the fluid into a single term called stagnation (or total) enthalpy.

It is defined per unit mass as

When the potential energy of the fluid is negligible, the stagnation

enthalpy represents the total energy of a flowing fluid stream per unit

Thus the stagnation enthalpy indicates the enthalpy of a fluid when it is

brought to rest adiabatically.

The properties of a fluid at the stagnation state are called stagnation

During stagnation process, since the kinetic energy of a fluid is converted
to enthalpy (internal energy or flow energy), the fluid temperature and
pressure is increased.

Stagnation (or total) temperature is the temperature the gas attains when
it is brought to rest adiabatically.

The term V2/2Cp corresponds to the temperature rise during such a

process and is called the dynamic temperature.

For high-speed flows, the stagnation temperature is higher than the static
(or ordinary) temperature.

For example, the dynamic temperature of air flowing at 100 m/s is about 5
Therefore, when air at 300 K and 100 m/s is brought to rest adiabatically
(at the tip of a temperature probe, for example), its temperature rises to
the stagnation value of 305 K.

The pressure a fluid attains when brought to rest isentropically is called

the stagnation pressure.

The stagnation and static properties can be related in terms of Mach

The stagnation process is considered isentropic and following relations
can be used for pressure.

The stagnation quantities Tt, pt etc. can be calculated from the

actual conditions of M, V, T, p, and at a given point in a general

The actual flowfield itself may not have to be adiabatic or isentropic from
one point to the next.

The isentropic process is only for definition of total conditions at a point.

One dimensional isentropic flow

Area changes, friction, and heat transfer are the most important factors
that affect the properties in a flow system.

Some situations may involve the simultaneous effects of two or more of

these factors.
In this topic the general problem of varying-area flow under the
assumptions of no heat transfer (adiabatic) and no friction is considered.

To develop relations for the variation of fluid properties with area changes
and Mach number we begin with energy equation.


Tds (Gibbs relation) based on First law of Thermodynamics equation and

definition of W (flow work) and Q (Tds) is given as:

Since the flow situation is assumed to be adiabatic and without losses ds

= 0. The equation becomes


From (1) and (2)

We introduce this into differential form of the continuity equation

For the isentropic case, the subscripts is dropped and the partial
derivative is changed to an ordinary derivative in the velocity of sound


Substituting (4) for dp into (3) yields




dA dp
A V2
( 1M 2 ) (8)

From equation (5) we see that at low Mach numbers, density variations
will be quite small, whereas at high Mach numbers the density changes
very rapidly.

This means that the density is nearly constant in the low subsonic regime
(d 0) and the velocity changes compensate for area changes.

If the pressure is assumed to be decreasing (i.e. dp is negative) following

results can be obtained from equations 5-7.
Equation (8) describes the variation of pressure with flow area. For
subsonic flow (M < 1), dA and dp must have the same sign.

That is, the pressure of the fluid must increase as the flow area of the
duct increases and must decrease as the flow area of the duct decreases.

Thus, at subsonic velocities, the pressure decreases in converging ducts

(subsonic nozzles) and increases in diverging duct (subsonic diffusers).
In supersonic flow (Ma >1), dA and dp have opposite signs.

Two common devices involving area change are nozzle and diffuser.

The same piece of equipment can operate as either a nozzle or a diffuser,

depending on the flow regime

Thus a device is called a nozzle or a diffuser because of what it does, not

what it looks like.

A nozzle is a device that converts enthalpy (or pressure energy for the
case of an incompressible fluid) into kinetic energy.

From Figure we see that an increase in velocity is accompanied by either

an increase or decrease in area, depending on the Mach number.
A diffuser is a device that converts kinetic energy into enthalpy (or
pressure energy for the case of incompressible fluids).

The highest velocity we can achieve by a converging nozzle is the sonic

velocity, which occurs at the exit of the nozzle.

To accelerate a fluid, we must use a converging nozzle at subsonic

velocities and a diverging nozzle at supersonic velocities.

Based on Equation (6), which is an expression of the conservation of

mass and energy principles, a diverging section must be added to a
converging nozzle to accelerate a fluid to supersonic velocities.

The result is a converging diverging nozzle.

The fluid continues to accelerate as it passes through a supersonic

(diverging) section. A large decrease in density makes acceleration in the
diverging section possible.

Effect of Back Pressure

Consider the subsonic flow through a converging nozzle as shown in
The nozzle inlet is attached to a reservoir at pressure pr and temperature
T r.
The reservoir is sufficiently large so that the nozzle inlet velocity is

The fluid velocity in the reservoir is zero and the flow through the nozzle
is approximated as isentropic,

The stagnation pressure and stagnation temperature of the fluid at any

cross section through the nozzle are equal to the reservoir pressure and
temperature, respectively.
Now we begin to reduce the back pressure and observe the resulting
effects on the pressure distribution along the length of the nozzle, as
If the back pressure pb is equal to pt, which is equal to pr, there is no flow
and the pressure distribution is uniform along the nozzle.

When the back pressure is reduced to p2, the exit plane pressure pe also
drops to p2. This causes the pressure along the nozzle to decrease in the
flow direction.

Now suppose the back pressure is reduced to p3 (= p*), which is the

pressure required to increase the fluid velocity to the speed of sound at
the exit plane or throat).

The mass flow reaches a maximum value and the flow is said to
be choked.

Further reduction of the back pressure to level p4 or below does not result
in additional changes in the pressure distribution, or anything else along
the nozzle length.

The effect of back pressure on the nozzle exit pressure pe is.

(i) pe = pb for pb p*
(ii) pe = p* for pb < p*

Under steady-flow conditions, the mass flow rate through the nozzle is
constant and is expressed as:

m=AV =
( )
A ( M RT )= pAM


In terms of stagnation properties


Applying this equation to the outlet and considering choked flow, M = 1

and A = A* we get


For a given gas from equation (9) we can write

We now look at four distinct possibilities:

1. For a fixed Tt, pt, and A* mmax constant.

2. For only pt increasing mmax increases.
3. For only Tt increasing mmax decreases.
4. For only A* increasing mmax increases.

A relation for the variation of flow area A through the nozzle relative to
throat area A* can be obtained by combining Eqs. (9) and (10) for the
same mass flow rate and stagnation properties of a particular fluid.

This yields:
( +1) /[ 2 ( 1 ) ]
A =
M [( )(
1 2
M )]
Another parameter sometimes used in the analysis of one-dimensional
isentropic flow of ideal gases is M*, which is the ratio of the local velocity
to the speed of sound at the throat
V V a RT
M = M =
a a a R T

Using definition of stagnation temperatures

M =M
2+ ( 1 ) M 2

ConvergingDiverging Nozzles

Accelerating a fluid to supersonic velocities (Ma >1) is possible only by

attaching a diverging flow section to the subsonic nozzle at the throat.

Forcing a fluid through a convergingdiverging nozzle is no guarantee

that the fluid will be accelerated to a supersonic velocity.

In fact, the fluid may find itself decelerating in the diverging section
instead of accelerating if the back pressure is not in the right range. The
various cases are given below:

(A) pb = pt, there is no flow through the nozzle

(B) The flow remains subsonic throughout the nozzle, and the mass
flow is less than that for choked flow.
The fluid velocity increases in the first (converging) section and
reaches a maximum at the throat (but Ma < 1).
The velocity reduces in the second (diverging) section of the
nozzle, which acts as a diffuser.
(C) The throat pressure becomes p* and the fluid achieves sonic
velocity at the throat and maximum mass flow takes place.

But the diverging section of the nozzle still acts as a diffuser,

slowing the fluid to subsonic velocities.
(D) The fluid that achieved a sonic velocity at the throat continues
accelerating to supersonic velocities in the diverging section as
the pressure decreases.
This acceleration comes to a sudden stop, however, as a normal
shock develops at a section between the throat and the exit.
(E) The normal shock moves downstream away from the throat as pb
is decreased. The shock forms (normal or oblique) at the exit
plane of the nozzle at point E (in Figure).
(F) The flow is supersonic through the entire diverging section in this
case, and it can be approximated as isentropic.
(G) The flow in the diverging section is supersonic, and the fluid
expands to pF at the nozzle exit with no shock forming within (or
outside) the nozzle.
Irreversible mixing and expansion waves occur downstream of the
exit plane of the nozzle.

Performance of Real nozzles

The performance of real nozzle differs slightly due to frictional effects.

Since the departure from isentropic flow is usually small isentropic

formulas can be used and then modified using coefficients such as nozzle
efficiency, velocity coefficient and discharge coefficient.
Suppose the nozzle is supplied with a gas at stagnation pressure and
temperature of pt1 and Tt1 respectively.

The gas expands adiabatically but with increasing entropy. The nozzle
efficiency is calculated as:

For nozzles that involve negligible heat transfer (per unit mass of fluid
flowing), we have

Since inlet velocity is very small when compared to exit we can write
V 2/ 2
n = 2 (1)
V 2s/ 2
Velocity coefficient is defined as the square root of efficiency

C v=
V 2s

Sometimes a discharge coefficient Cd is used and is defined as the ratio

of actual mass flow rate to ideal mass flow rate which is obtained by
expanding the gas isentropically to same final pressure.

The three parameters n C v Cd are related. From energy equation

V 22 s
ht 1 =h2 s +

For perfect gas

T 2s
V 22 s=2 C p T t 1 1 ( Tt 1 )
For isentropic process
1 /
T 2 s P2 s
T t1
= ( )
Pt 1

( ( ) )
V 22 s=2 C p T t 1 1
P2 s
Pt 1

Assuming imaginary isentropic process between the actual exit state 2

and stagnation state t2 we get
V =2 C p T t 2
( ( ) )
1 2

The stagnation temperature remains same since process is adiabatic

T t 1=T t 2

Substituting (2) and (3) in (1)

/ 1

{ [ ]}
1 /
P2 P
Pt 2
= 1n 1 2 s
Pt 1( ) (4)

From previously obtained equations

By eliminating M in above two equations

{ [( ) ( ) ]}
2/ ( +1 ) /
m 2 p p
=p t (5)
A R Tt 1 pt pt

Using this equation for isentropic ideal flow between 1 and 2s we get

{ [( ) ( ) ]}
2/ ( +1 ) /
ms 2 p2s p
= pt1 2s (6)
A RT t 1 1 pt 1 pt 1

Assuming imaginary isentropic process between p2 and pt2

{ [( ) ( ) ]}
2 / ( + 1) /
m 2 p2 p
=p t 2 2 (7)
A R Tt 1 1 pt2 pt 2

Dividing equations (6) and (7) and noting that p2= p2 s we get Cd

( )
( 1 ) / ( 1 ) /
p2 s p2

C d=
( )
pt 1
1 ( ) pt2
( 1 ) / ( 1) /
p2 p2 s
( )
pt 2
( ) pt 1

From equations 4 and 8 and eliminating pt 2

( 1) /

C d=
n (( ) )p2 s
pt 1

[ ( ) ]
( 1 ) /
p2 s
1 n 1
pt 1

Multi-dimensional flow effects

When one-dimensional flow is assumed, the flow properties are uniform

across a plane which is perpendicular to flow direction. In the actual
nozzle this assumption is not valid particularly for large angles.

Thus if the nozzle size and weight is reduced by increasing its angle, the
assumption of 1D flow becomes incorrect.

In converging nozzle, the inward radial momentum of the fluid also results
in vena contracta region. The cross-sectional area of vena contracta is
much smaller than nozzle exit area.
The combination of non-uniform flow and vena contracta reduces the
mass flow rate. The non-uniform flow is shown in a nozzle of 40 at pt/pe =
4 as an example. It is observed that the Mach No. is not constant in the
vertical direction.

The sonic lines at various pressure ratios indicate that due to increase in
pressure ratio the flow profile becomes flatter.

Force / Thrust Calculations

Consider a control volume that surrounds all the fluid inside the device /
Velocities are shown relative to the device, which is used as a frame of
reference in order to make a steady-flow picture. The x-component of the
momentum equation for steady flow is:

We define an enclosure force as the vector sum of the friction forces and
the pressure forces of the wall on the fluid within the control volume.

Let Fenc is the x-component of this enclosure force on the fluid inside the
control volume.

If the enclosure is pushing on the fluid with a force of magnitude Fenc to

the right, the fluid must be pushing on the enclosure with a force of equal
magnitude to the left
The combination of variables found in the equation is called the thrust or
impulse function.

The external forces are the ambient pressure over the entire enclosure.

Let Fext is the positive thrust that arises from the external forces pushing
on the enclosure.

Since this has been represented as a constant pressure

It is customary in the field of propulsion to work with the free-stream
conditions (p0 and V0) that exist far ahead of the actual inlet.