ESDU 90025

Issued November 1990

Reprinted at Amendment A
March 2007

Performance of conical diffusers in

subsonic compressible flow
ESDU Copyright material. For current status contact ESDU.

Endorsed by
The Institution of Mechanical Engineers
The Royal Aeronautical Society
 ESDU 90025
ESDU DATA ITEMS

Data Items provide validated information in engineering design and analysis for use by, or under the supervision
of, professionally qualified engineers. The data are founded on an evaluation of all the relevant information, both
published and unpublished, and are invariably supported by original work of ESDU staff engineers or consultants.
The whole process is subject to independent review for which crucial support is provided by industrial companies,
government research laboratories, universities and others from around the world through the participation of some
of their leading experts on ESDU Technical Committees. This process ensures that the results of much valuable
work (theoretical, experimental and operational), which may not be widely available or in a readily usable form, can
be communicated concisely and accurately to the engineering community.

We are constantly striving to develop new work and review data already issued. Any comments arising out of your
use of our data, or any suggestions for new topics or information that might lead to improvements, will help us to
provide a better service.

THE PREPARATION OF THIS DATA ITEM

ESDU Copyright material. For current status contact ESDU.

The work on this particular Data Item was monitored and guided by the Internal Flow Panel, which first met in 1979
and had the following membership:

Chairman
Dr J.A. Eaton — University College, Galway, Eire

Members
Dr T.W. Broyd — W.S. Atkins Engineering Services
Mr D.A. Campbell — Rolls-Royce Ltd, Derby
Mr J. Campbell — Ove Arup and Partners
Dr C.J. Clark — BP International Ltd
Dr D.J. Cockrell — University of Leicester
Prof. D.H. Freeston* — Auckland University, New Zealand
Mr A.J. Green — British Hydromechanics Research Association
Dr A. Hunt — Schlumberger Cambridge Research Ltd
Prof. J.L. Livesey — University of Salford
Mr J.A. Ward — AEA Technology, Harwell.

development of the Data Item was undertaken (under contract to ESDU) by

Prof. W.A Kamal — University of Qatar

Prof. J.L. Livesey — University of Salford

The person with overall responsibility for the work in this subject area was Mr G.H. Walter, Head of Heat Transfer,
Internal Flow and Physical Data.
 ESDU 90025
PERFORMANCE OF CONICAL DIFFUSERS IN SUBSONIC COMPRESSIBLE FLOW

CONTENTS
Page

1. NOTATION 1

2. INTRODUCTION 4
2.1 Purpose and Scope of this Data Item 4
2.2 Diffuser Geometry 4
2.3 Background Information 5
ESDU Copyright material. For current status contact ESDU.

2.4 Performance Parameters 5

2.5 How to Use this Data Item 9

3. PERFORMANCE DEPENDENCE ON INLET GEOMETRY AND MACH NUMBER 9

3.1 The Diffuser Inlet Condition 9
3.2 Outlet Flow Specification 10
3.3 Performance Maps 11
3.4 How to Use the Maps 12
3.5 Worked Examples 13
3.6 Estimation of the Diffuser Total-pressure Loss Coefficient, Kt 16

4. PERFORMANCE DEPENDENCE ON INLET CONDITIONS 18

4.1 Effects of Variation in Reynolds Numbers 18
4.2 Effects of Variation in Inlet Turbulence Level 19
4.3 Effects of Variation in Inlet Profile Shape 19
4.4 Mach Number Approaching Choking 20
4.5 Diffuser Preceded by a Shock Wave/Boundary Layer Interaction 20

5. PERFORMANCE DEPENDENCE ON SOME OTHER GEOMETRICAL PARAMETERS 23

5.1 Effects of Fairing the Junction with the Inlet Duct 23
5.2 Effects of Using a Tailpipe 23
5.3 Other Design Improvements 24

6. DERIVATION AND REFERENCES 25

6.1 Derivation 25
6.2 References 27

FIGURES 28 to 44

i
 ESDU 90025
PERFORMANCE OF CONICAL DIFFUSERS IN SUBSONIC COMPRESSIBLE FLOW

1. NOTATION

SI Units

A flow area m2
ρv dA
AE effective flow area, ∫A ------------
ρ v
- m2
0 0
B blockage ratio, ( A – A E )/A (approximately 2δ*/R for thin boundary –
layers)

b ratio of one-dimensional static pressure, p , to entropy-mean static pressure, –

ESDU Copyright material. For current status contact ESDU.

p , see Section 2.4

cp specific heat capacity at constant pressure J/kg K

cv specific heat capacity at constant volume J/kg K

C pr static pressure recovery coefficient, see Section 2.4 and Table 2.1 –

C pr* optimal value of C pr for given length ratio L/R1 –

C pr** optimal value of C pr for given area ratio A 2 /A 1 –

D diameter m

H boundary layer shape parameter, δ*/θ –

I local turbulence intensity, q/v –

R ½
∫ 0 ρvq
2
r dr
1
IG diffuser inlet turbulence parameter, ------ ----------------------------- –
V0 R
∫0
ρvr dr

Kt total-pressure loss coefficient, see Section 2.4 –

K t* minimum value of K t for given length ratio L/R 1 –

K t** minimum value of K t for given area ratio A 2 /A 1 –

L axial length of diffuser m

Ld length of straight, parallel duct downstream of diffuser (tailpipe) m

Lu length of straight, parallel duct upstream of a point 1.5 D 1 from diffuser, see m
Sketch 2.1

Issued November 1990 – 44 pages

Reprinted at Amendment A – March 2007
1
 ESDU 90025
M Mach number –

m total mass flow rate kg/s

p static pressure N/m2 (Pa)

pt total pressure N/m2 (Pa)

q turbulence velocity, [ v′ x2 + v′y2 + v′ z2 ] ½ , where v x′, vy′, v z′ are orthogonal m/s

components; often only component in x-direction, vx′ , is known, in which
case, assuming isotropy, q ≈ [ 3v x′2 ] ½ .

R radius m

r local radius m
ESDU Copyright material. For current status contact ESDU.

Rg gas constant (= 287.1 for air) J/kg K

Re Reynolds number, V1D 1 /ν1 –

T static temperature K

Tt total temperature K

V velocity in one-dimensional flow m/s

V area averaged mean velocity, m/ρA m/s

v local velocity m/s

v time mean local velocity m/s

x axial coordinate m

y coordinate normal to duct wall m

α kinetic energy coefficient, see Section 2.4 –

β momentum coefficient, see Section 2.4 and Table 2.1 –

γ ratio of specific heat capacities, c p /c v –

R ρv r
δ* boundary-layer displacement thickness, ∫0 1 – ------------ --- dr
ρ0 v0 R
m

R
ρv v r
θ boundary-layer momentum thickness, ∫ 0 ρ-----------
0 0v
- 1 – ----- --- dr
v0 R
m

η diffuser efficiency (see Section 2.4 and Table 2.1); sometimes called –
effectiveness or conversion efficiency

µ dynamic viscosity N s/m2

2
 ESDU 90025
ν kinematic viscosity m2/s

ρ fluid density kg/m3

φ angle between diffuser wall and diffuser axis (total angle = 2φ ) degree

Subscripts

e exit plane of diffuser and tailpipe, see Sketch 2.1

0 mainstream or outside boundary layer
opt optimum value
t total or stagnation conditions
ESDU Copyright material. For current status contact ESDU.

1 inlet plane of diffuser (1.5 D 1 upstream of pipe-cone junction, see

Sketch 2.1)
2 exit plane of diffuser, see Sketch 2.1

Superscript

′ time dependent variable

Mean variables are denoted by a bar over a symbol, for example, p t ; the type of mean is important and is
identified in the text.

3
 ESDU 90025
2. INTRODUCTION

2.1 Purpose and Scope of this Data Item

This Data Item enables static pressure recovery to be predicted and total-pressure loss and flow pattern*
estimated for subsonic compressible flow through straight-axis conical diffusers (right cones). The effects
of various geometrical and flow parameters on diffuser performance are discussed. By selection of a
geometry it enables a subsonic diffuser of required performance to be designed or, alternatively, enables
the performance of a diffuser of specified geometry to be assessed.

Performance is considered primarily in terms of static pressure rise for a range of entry mean Mach numbers,
0.2 ≤ M1 ≤ 0.8 , and a range of entry duct lengths, 0 ≤ L u /D 1 ≤ 36 .

Other influences such as Reynolds number, velocity profile shape, turbulence level, shock-wave/boundary
layer interaction upstream of the diffuser and the approach to choking are considered as a guide to
ESDU Copyright material. For current status contact ESDU.

understanding the behaviour and performance of such diffusers and as an aid to design or performance
assessment.

Guidance on the estimation of total-pressure loss from static pressure rise is given.

2.2 Diffuser Geometry

The geometry of a typical conical diffuser and its associated ductwork is shown in Sketch 2.1. For the case
of a diffuser with free discharge the diffuser exit plane 2 forms the end of the system.

Ld D2

e
1.5D1

φ
Lu
2

1
D1
Flow

Sketch 2.1 Typical conical diffuser geometry

*
Flow pattern refers to flow regime and velocity profile.

4
 ESDU 90025
The principal geometrical parameters are related by

A L 2
-----2- = 1 + ------ tan φ (2.1)
A1 R1

and the performance maps in this Item are presented in the form of contours of C pr on an orthogonal grid
of area ratio, A2 /A 1 , and the ratio of diffuser length to inlet duct radius, L/R1 , with lines of constant diffuser
total angle, 2φ , as an auxiliary scale.

2.3 Background Information

ESDU 760271 gives a comprehensive introduction to the design and performance of diffusers generally
and ESDU 730242, 7401530 and 7502631 give detailed performance data for conical, plane-walled and
annular diffusers. ESDU 870153 considers methods of performance improvement. However, these Items
ESDU Copyright material. For current status contact ESDU.

are only applicable to incompressible flow.

A user coming to this Data Item without previous knowledge of diffuser behaviour and performance is
directed to the first two Data Items above (and to the final one if seeking performance improvement). The
user should first become familiar with diffuser function, regime behaviour, performance parameters (static
pressure recovery coefficient, C pr , total-pressure loss coefficient, K t , efficiency or effectiveness, η , and
kinetic energy coefficient, α ), optimality defined by C pr* , C pr** , K t* and K t** , insufficient and
inefficient diffusion and the influence of inlet flow conditions. A good understanding and familiarity with
one-dimensional gas dynamics is also required.

In this Data Item the influence of compressibility (implied by inlet Mach numbers greater than about 0.2)
and the spatial non-uniformity of the flows involved necessitates a formal reconsideration of the definitions
of performance parameters with particular reference to total-pressure losses. Also required is the
establishment, and understanding of the implications, of suitable mean-variable systems used in the
one-dimensional description of the spatially non-uniform compressible flows involved.

2.4 Performance Parameters

Table 2.1 defines the performance parameters, static pressure recovery coefficient, C pr , total-pressure loss
coefficient, K t , and diffuser effectiveness, η , in terms of mean flow variables.

The first column of the Table shows the parameter definitions for one-dimensional incompressible flow.

The second column shows the definitions for incompressible, spatially non-uniform flow. The density is
constant and the static pressure is assumed uniform across the duct at measurement stations 1 and e.
However the velocity varies across the duct and a simple mean variable set is required. The incompressible
flow model is ess+entially isothermal and of constant density. The set of mean variables is

ρ = ρ = ρt , p = p , T = T ≡ Tt , V , α , β .

The mean velocity, V , is defined as

1
A A ∫
V = --- v dA . (2.2)

5
 ESDU 90025
The mean static pressure, p , is given by

p = p,

the measured static pressure for the one-dimensional parallel flow.

The mean total pressure, p t , is mass weighted, giving

1
p t = ---
∫
- p dm
m A t
(2.3)

1
or
m A ∫
p t = ---- ( p + ½ρv 2 ) dm ,

which can be also expressed as

ESDU Copyright material. For current status contact ESDU.

p t = p + α½ρ V 2 , (2.4)

where the kinetic energy coefficient, α , is given by

1
½mV 2 A
∫
α = ------------------ 1--2- v 2 dm . (2.5)

6
 ESDU 90025
TABLE 2.1 Diffuser Performance Parameters

Incompressible
Incompressible Compressible
One-Dimensional
Spatially Non-Uniform Spatially Non-Uniform
Spatially Uniform

ρ constant ρ constant ρ varies

p ≡ p ; C pr ≤ 1 ; α ≡ 1 p ≡ p ; C pr ≤ 1 ; α ≠ 1 p ≠ p ; C pr ≤ 1 ; α ≠ 1
1
p t = p + ½ρV 2
p t = ---- p t dm
m A ∫ 1
loge p t = ---- loge p t dm ∫
m A

pe – p1 pe – p1 pe – p1
ESDU Copyright material. For current status contact ESDU.

C pr ------------------ ------------------------------ -----------------------

½ρV 1
2
½ρα 1 ( V 1 ) 2 p t1 – p 1

*
p t1 – p te p t1 – p te R g T t loge ( p t1 ⁄ p te )
Kt ---------------------- -----------------------------
- -----------------------------------------------------
½ρV 12 ½ρα 1 ( V 1 )2 ½α 1 ( V 1 )2

Irreversibilities
≡ ---------------------------------------------------------
Inlet kinetic energy flux

pe – p1 pe – p1 R g T t loge ( p t1 ⁄ p te )
η ---------------------------------------------------------
- ---------------------------------------------------------------
- 1 – ------------------------------------------------------------
½ρV 2 [ 1 – ( A 1 ⁄ A 2 ) 2 ] ½ρ [ α 1 ( V 1 ) 2 – α e ( V e ) 2 ] ½ [ α1 ( V1 )2 – αe ( Ve )2 ]

Actual ∆p Actual ∆p Irreversibilities

≡ ------------------------------- ≡ ------------------------------- ≡ 1 – ---------------------------------------------------------------
Ideal ∆p Ideal ∆p Kinetic energy conversion

1 1
m A∫
---- 1--2- v 2 dm = α 1--2- ( V ) 2 ;
∫
---- v dm = β V ;
m A
K t = ( 1 – η ) [ 1 – α e ( V e ) 2 ⁄ α 1 ( V 1 ) 2 ].

* See Section 3.6 for more conventional K t definition.

The kinetic energy coefficient, α , is a convenient lumped parameter enabling the kinetic energy flux to be
described in terms of the other mean variables without the use of integral forms. A coefficient, β , describing
momentum flux, is written similarly as

1
β = --------- v d m .
mV A
∫ (2.6)

7
 ESDU 90025
The third column of Table 2.1 shows the definitions of diffuser performance parameters for compressible,
spatially non-uniform flow; here a more complicated mean-variable system is required, see Derivations 4, 5
and 6. Note that now density varies so that ρt > ρ and also p ≠ p . The mean total- pressure, p t , is defined
as a mass weighted logarithmic mean,

1
loge ρ t = ---
∫
- log p dm ,
m A e t
(2.7)

which ensures conservation of entropy flux (see Derivation 4).

The compressible flow model is adiabatic, of constant total temperature Tt = Tt and specific heat capacities
( γ is constant), and the flow is assumed parallel and of constant static pressure at a cross section. These
assumptions are consistent with an adiabatic gas flow of constant Prandtl number* equal to unity and are
considered valid for Prandtl numbers near to unity, for example, air. The set of mean variables is
ESDU Copyright material. For current status contact ESDU.

ρ , ρ t, p , p t, T , T t, V , M , α, β, b .

The Mach number, M , is introduced as a derived fundamental variable, important in recognising the
influence of compressibility effects. The quantity b recognises that p ≠ p , the assumed constant static
·
pressure, by writing p = bp . Consideration of flows with variable T t, p and γ is found in Derivations 4
and 7.

In the third column of Table 2.1, the term R g T t loge [ p t1 ⁄ p te ] represents the specific energy equivalent
of the irreversibilities arising in the diffuser flow. The definitions of C pr , K t and η reduce
systematically from right to left to the definitions in the second and first columns, as firstly the density is
assumed constant and as secondly the flow is assumed to be spatially uniform (rigorously one-dimensional).

Note also the definitions of C pr in columns two and three include in the denominator the difference between
the total pressure p t1 and the static pressure p 1 . It is thus ensured that C pr ≤ 1 and these definitions differ
in this respect from those in earlier Data Items.

In the lowest compartment of Table 2.1 will be found the relationship between K t and η ,

αe ( Ve )2
Kt = ( 1 – η ) 1 – --------------------
- . (2.8)
α1 ( V1 )2

This is a complicated relationship in general terms, depending on the ratio of the kinetic energy fluxes at
exit and inlet. The approximate relationship,

Kt = ( 1 – η ) ,

applies only when V e → 0 (very large A e ) or if the exit kinetic energy is lost by free discharge. See also
the discussion in Sections 2.3.2 and 2.3.4 of ESDU 760271 on the implications of high α e in Equation (2.8)
implying low K t (note that high α e also implies low C pr ) and on the distinction between insufficient and
inefficient diffusion.

*
The Prandtl number of a fluid is defined as (dynamic viscosity × specific heat capacity at constant pressure) / thermal conductivity.

8
 ESDU 90025
2.5 How to Use this Data Item

(i) Users unfamiliar with diffuser performance should refer to the Data Items mentioned in Section 2.3
for a general discussion of diffuser behaviour.

(ii) Proceed to the maps in Section 3.3 to obtain values of C pr and other unknown parameters.

(iii) Follow the steps in either of the two worked examples given, as appropriate.

(iv) Proceed to Sections 4 and 5 for consideration of other possible influences.

(v) Assess resultant design or performance.

(vi) Re-assess the design or performance in terms of static pressure rise and total-pressure loss.
ESDU Copyright material. For current status contact ESDU.

3. PERFORMANCE DEPENDENCE ON INLET GEOMETRY AND MACH NUMBER

3.1 The Diffuser Inlet Condition

Conical diffuser design involves determining one of the two main geometrical parameters (length or area
ratio) if the other is given. The optimum value of that parameter is considered to be the one yielding the
maximum pressure recovery, minimum internal losses, and least non-uniform velocity profile at the diffuser
exit. Compromise may be necessary.

For the proper design of a diffuser, its inlet flow condition needs to be known or estimated. This is defined
in terms of five flow parameters: the inlet mean Mach number, M 1 , Reynolds number, Re , boundary
layer blockage ratio, B 1 , shape parameter, H 1 , and turbulence level14,26. A minimum of three input
parameters are needed. The inlet Mach number must always be known but the inlet Reynolds number is
found to have little influence on diffuser performance so long as the boundary layer at inlet is turbulent9,25
(but see Section 4.1). Of the other parameters, B 1 and I G are the more important. Typical values of B 1
and I G for developing pipe flow are given in Figures 1 and 2 respectively. Figure 1 indicates that
compressibility has no significant effect on the value of the blockage ratio, B 1 , with deviations from the
incompressible correlation showing no definite trend. However, Figure 2 shows clearly that bigger values
of I G are expected for compressible flow. For naturally-developing pipe flow, the design situation is simply
defined in terms of L u /D 1 and the entry Mach number.

The performance of a diffuser is normally evaluated from measurements in its inlet plane, which is strictly
the plane of geometrical transition from the parallel pipe to diffuser, and its exit plane. However, the
geometrical transition, especially when it is sharp, creates a non-uniform static pressure distribution across
the plane of geometrical transition which, in addition to the strong streamline curvature effects there, makes
measurements difficult to analyse. To avoid this difficulty, it is usual to take the diffuser inlet section some
distance upstream of the geometrical transition, thus avoiding the effects of flow instabilities and
unsteadiness in the transition plane and, more importantly, attributing the transition losses to the diffuser.
In the majority of the sources of data, Derivations 10 to 14, this distance was 1.5 pipe diameters and
accordingly it was selected as the basis for the performance maps presented in this Data Item. Data from
other sources, where different distances were used, were corrected to allow for notional frictional pressure
drop in the entry region.

9
 ESDU 90025
3.2 Outlet Flow Specification

The state of the outlet flow depends on the geometry. Sketch 3.1, taken from ESDU 730242, shows the
location of the flow regime boundaries as a function of geometry with corresponding descriptions of the
flow given in Table 3.1. No experimental data are available on the influence of Mach number on the regime
boundaries. Theoretically, the expectation would be that the lower end of aa would move lower (leaving
the upper end of aa little affected) and a'a' would move nearer to aa for inlet Mach numbers significantly
greater than 0.2.

φ°
90 40 2 12 8 4
20
w b
15
flo

n
tio
t
Je

ara
10

sep
ESDU Copyright material. For current status contact ESDU.

ry
a′ 2

ito
ns
6

tra
ing
as
4
cre

a 1
In

3
A2/A1
b
w
2
flo
ed
1.8 a′
ch
tta

1.6
A

1.4
1.3
1.2
a
1.1 6 8 2 3 4 6 8 2 3 4 6 8
1 10 102
L/R1

Sketch 3.1 Flow regime boundaries for conical diffusers

10
 ESDU 90025
TABLE 3.1 Description of Flow Regimes indicated on Sketch 3.1

Typical features of outlet

Flow regime Description of flow patterns Sketch
conditions

Attached flow Steady flow, attached throughout Similar to inlet conditions but
(below aa , diffuser. thicker boundary layer.
Sketch 3.1).

Transitory Separated regions form at walls Unsteady flow condition,

separation and move downstream as intermittently asymmetric
(between aa bubbles, separation intermittent velocity profile. Outlet
and b b, and scattered. Below the conditions deteriorate as
Sketch 3.1). boundary of appreciable geometry moves upwards from
ESDU Copyright material. For current status contact ESDU.

separation ( a'a' , Sketch 3.1), the boundary a'a' but below this
separation is on a small scale and boundary unsteadiness has little
is confined mainly to the area significant effect on overall
near the outlet. flow.
Jet flow Flow separates all round a Essentially steady conditions,
(above bb , circumference. central core surrounded by
Sketch 3.1). separated flow annulus.

3.3 Performance Maps

Figures 3 to 18 give the pressure recovery coefficient contours for a wide range of subsonic inlet Mach
numbers and inlet pipe lengths for steady, non-swirling subsonic flow in conical diffusers with
naturally-developing boundary-layer type inlet flow and with tailpipes but with the exit measuring station
at the end of the divergence. The effect on performance of different conditions is discussed in Section 4.

The maps were constructed using experimental data from Derivations 8 to 17, together with the correlated
data from Derivation 2. The data were corrected as necessary to allow for variations in the definition of
C pr and the system used to obtain mean values, and for the differences in the distance upstream of the
entrance corner where inlet conditions were measured. If a reported C pr value for a specific geometry and
inlet condition was found to be more than 30% in deviation from the bulk of the data for that geometry
under comparable inlet conditions, that data point was excluded.

The maps show a general deterioration in diffuser performance as the inlet Mach number increases,
especially for the case of longer inlet pipes (larger inlet boundary layer blockage). For example, Figure 18
shows that, with M 1 = 0.8 and L u /D 1 > 30 , it is impossible to obtain C pr = 0.75 with a diffuser of
any length or area ratio.

Moreover, the maps indicate that whatever the diffuser inlet conditions ( M 1 and B 1 ) might be, the best
overall pressure recovery is always in the region of area ratio A 2 /A 1 = 7 to 10 and L/R 1 = 38 to 44,
which corresponds to a diffuser total angle of 5 to 6°. In practice, the designer can seldom design for the
best overall performance because of constraints such as a fixed area ratio or a fixed length.

On each of the sixteen maps presented, two important design lines are shown as broken lines. The C pr*
line is the locus of points that define the diffuser area ratio producing maximum pressure recovery in a
diffuser of prescribed non-dimensional length. If this area ratio is not attained, the length of the diffuser

11
 ESDU 90025
will be excessive and the surface friction unnecessarily high, whereas if it is exceeded premature separation
will result. The C pr** line is the locus of points that define the diffuser non-dimensional length producing
maximum pressure recovery in a diffuser of prescribed area ratio.

In almost all the maps, the C pr* line lies above the line of first appreciable stall, while with moderate area
ratios near to 2 both the C pr* and C pr** lines extend into the transitory stall region. This means that
optimum performance is obtained with unsteady stall within the diffuser, and the designer must be aware
of possible implications at other points in the flow system.

The designer may thus need to choose a performance that is not optimum in the interests of achieving flow
steadiness and less sensitivity to inlet conditions.

3.4 How to Use the Maps

Three design cases are considered as follows.

ESDU Copyright material. For current status contact ESDU.

Case I. To determine diffuser performance for known inlet conditions and known geometry.

Select the appropriate map for the given inlet conditions, interpolating, if necessary, for L u /D
and M .

Read off the expected pressure recovery coefficient, C pr , for the geometrical parameters and
review how far the design is off the optimum geometry for these inlet conditions.

Note: Diffuser performance may, of course, be predicted over a range of inlet Mach numbers, if
desired, by using the appropriate performance maps for the given geometry.

Case II.To determine optimum diffuser for known inlet conditions with constrained length.

Select the appropriate map for the given inlet conditions, interpolating, if necessary, for L u /D
and M .

At the specified value of L/R 1 , the optimum area ratio can be obtained from the C pr* line. The
optimum diffuser angle can be interpolated or obtained more accurately from the relation

( A2 ⁄ A1 ) – 1
φ = tan-1 -------------------------------------- . (3.1)
L ⁄ R1

The maximum pressure recovery coefficient, C pr , to be expected can be interpolated from the
values on the contour lines.

Case III.To determine optimum diffuser for known inlet conditions with fixed area ratio.

Select the appropriate map for the given inlet conditions, interpolating, if necessary, for L u /D
and M .

At the specified value of A 2 /A 1 , the optimum diffuser length can be obtained from the C pr** line.
The optimum diffuser angle can be interpolated or obtained from Equation (3.1).

The maximum pressure recovery coefficient, C pr , to be expected can be interpolated from the
values on the contour lines.

12
 ESDU 90025
3.5 Worked Examples

Example 1. Cases I and II

An 8° (total angle) conical diffuser, 280 mm in length, connects a 2.15 m long pipe 100 mm in diameter to
a large tank. If the static pressure at the diffuser inlet station is 100 kPa and the measured temperature is
45°C, what should be the tank pressure required to maintain a steady air flow of 2.5 kg/s ? What would be
the diffuser angle giving the highest achievable static pressure at the diffuser exit ?

Since the diffuser discharges into a large space, the kinetic energy remaining in the flow at the exit plane
is not recovered. Thus, it is safe to assume that the tank pressure will very approximately equal diffuser
exit static pressure.

To determine the diffuser pressure recovery coefficient, C pr , the diffuser inlet mean Mach number must
first be found. The mass flow rate of air is given by the relation:
ESDU Copyright material. For current status contact ESDU.

p
- M γR g TA 1 .
m = ---------
Rg T

Assuming the measured temperature is the total (stagnation) temperature, that relation becomes

γ–1
p 1  1 + ------------ ( M 1 ) 2 γR g T t ½
 2  π 2
m = ---------------------------------------------------- M 1 ---------------------------------------- --- D 1
Rg Tt γ–1 4
1 + ------------ ( M 1 ) 2
2

or, substituting the given values,

3
100 ×10 ( 1 + 0.2 ( M 1 ) 2 ) 1.4 × 287.1 × 318 ½ π
2.5 = -------------------------------------------------------------- M 1 --------------------------------------------- --- ( 0.1 ) 2 .
287.1 × 318 1 + 0.2 ( M 1 ) 2 4

This equation is solved iteratively to give M 1 = 0.77 , and, with Lu /D1 = 2 ⁄ 0.1 = 20 , the closest
performance map is Figure 17 for M1 = 0.8 and Lu /D1 = 20 . From the map, for the 8° diffuser at
L/R 1 = 280 ⁄ 50 = 5.6 , the pressure recovery coefficient, C pr , is 0.405.

But p t1 = p 1 [ 1 + 0.2 ( M 1 ) 2 ] 3.5 = 100 [ 1 + 0.2 ( 0.77 ) 2 ] 3.5 = 148 kPa,

so that p 2 can now be determined from the definition of C pr :

p2 – p1
C pr = ----------------------- ,
p t1 – p 1
which gives p 2 = 119.4 kPa.

Now, to obtain the optimum diffuser angle for the prescribed diffuser length ratio L/R1 = 5.6 , the value
of C pr* at Lu /D1 = 20 is required. Then the optimum area ratio is found to be 2.8 and, from
Equation (3.1), the optimum angle, 2φ , is found to be 13.7°. In this case the pressure recovery coefficient,
C pr, opt , increases to a value of 0.45.

13
 ESDU 90025
Example 2. Case III

It is required to design a conical diffuser to connect a 200 mm diameter pipe, 2.3 m long with a short
compressor suction pipe, 400 mm in diameter. The inlet mean Mach number is 0.6 and the static pressure
at the diffuser inlet station is 150 kPa. What will be the static pressure in the exit pipe ?

First determine which performance map to use. With M1 = 0.6 and Lu /D1 = ( 2.3 – 0.3 )/0.2 = 10 , the
performance map of Figure 12 is the closest, with Lu /D1 = 9 – 10.5 .

For a given area ratio, the optimum geometry is given by the C pr** line. In this case, with A2 /A1 = 4 , the
C pr** line gives ( L/R1 )opt = 29 . Using Equation (3.1), φ opt ≈ 2° and the optimum diffuser cone angle,
2φ , will be 4°. The map shows that for this optimum geometry, Cpr = 0.78 .

γ
γ–1 ------------
2 γ–1
Using p t1 = p 1 1 + ------------ ( M 1 ) ,
ESDU Copyright material. For current status contact ESDU.

2
p t1 = 150 [ 1 + 0.2 ( 0.6 ) 2 ] 3.5 = 191.3 kPa.
Thus, p 2 = 150 + 0.78 ( 191.3 – 150 ) = 182.2 kPa.

This is the static pressure at the diffuser exit section. As shown in Section 5.2, a tail pipe, however short,
can improve the diffuser pressure recovery coefficient, but in this example, with the high value of C pr of
the diffuser itself and a short tail pipe, only a relatively small improvement in C pr is to be expected.

14
 ESDU 90025

0.80
A2 /A1
0.16
0.78 4 A2 /A1
0.14 1.7
0.76 4 A2 /A1
0.12 1.6
0.74 3 4
3
0.10 1.5
0.72
0.10 1.4
0.70 3
0.08 1.3
2
0.68
ESDU Copyright material. For current status contact ESDU.

2
0.06 1.2 2
0.58
0.57 0.04 1.1
1.5 1.5
0.56
0.55 0.02 1.0
10.5 11.0 11.5 12.0 12.5 10.5 11.0 11.5 12.0 12.5 10.5 11.0 11.5 12.0 12.5
IG × 100 IG × 100 IG × 100

(a) 2φ = 5˚, M1 ~
– 0.87, B1 ~
– 0.055

0.32
A2 /A1
0.30
4
0.60
A2 /A1
0.28 2.0
4 A2 /A1
0.56 0.26 1.9
4
0.24 1.8
0.52 3
3 0.22 1.7

0.48 0.20 1.6

2
3
0.18 1.5
0.44
0.16 1.4
2
2
0.40 0.14 1.3
8 9 10 11 8 9 10 11 8 9 10 11
IG × 100 IG × 100 IG × 100

(b) 2φ = 12˚, M1 ~
– 0.84, B1 ~
– 0.07

Sketch 3.2 Examples of variation of diffuser performance parameters with inlet turbulence level14

15
 ESDU 90025
3.6 Estimation of the Diffuser Total-pressure Loss Coefficient, Kt

The number of experimental data available on K t (Derivations 11 to 14) is not sufficient to construct contour
maps similar to those obtained for C pr . Most of the data available are for four diffuser angles, 2φ , of 5°,
8°, 12° and 20°.

Yet the designer may wish to estimate the total pressure loss in the conical diffuser, in order to evaluate the
losses in the whole fluid system. For this purpose, a crude estimate of the total-pressure loss coefficient
can be calculated from the value of C pr (obtained from the performance maps) using the following
approximate relations:

1
 αe ρ1 ⁄ ρ2 2  γ–1 ------------
2 γ–1
K t =  1 – ------ ------------------ – C pr  1 + ------------ ( M 1 ) , (3.2)
 α1 A2 ⁄ A1  2
ESDU Copyright material. For current status contact ESDU.

1
γ–1 M 2 2 ------------ γ – 1
1 + ------------ ------------------
ρ1 2 A2 ⁄ A1
where ------ = ---------------------------------------------------- .
ρ2 γ–1
1 + ------------ ( M 1 ) 2
2

The relations are extensions of those given in Derivation 1. The energy coefficients, α 1 and α e , are
required. If measurements are not available the magnitudes can be estimated; α 1 can be estimated from
pipe flow data whilst typical α e magnitudes are given in Sketch 3.2 (Derivation 14).

From Equation (3.2), the total pressure loss can be determined by inverting the definition of K t in Table 2.1
to give

p t1 K t [ --12- α 1 ( V 1 ) 2 ]
--------- = exp ----------------------------------- . (3.3)
p t, e Rg Tt

Sketch 3.3 from Derivation 11 shows the variation of K t with inlet Mach number for three cone angles and
three entry pipe lengths. As expected from Equation (3.2) and the behaviour of C pr with Mach number,
K t increases rapidly when choking is approached. Values of K t in the Sketch can be predicted using
Equation (3.2) and the contour maps for C pr to within a maximum error of 25% and often much more
precisely.

16
 ESDU 90025

0.5

0.4

0.3
Kt
2φ
0.2
12
0.1
8
5
ESDU Copyright material. For current status contact ESDU.

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
M1
(a) Lu/D1 = 0, A2/A1 = 16

0.6
0.5
2φ
0.4
12
Kt 0.3 8
0.2
5
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
M1
(b) Lu/D1 = 23.7, A2/A1 = 16

0.7
0.6
0.5
2φ
0.4
Kt 12
0.3
8
0.2
0.1 5

0
0 0.1 0.2 0.3 0. 0.5 0.6 0.7 0.8 0.9 1.0
M1

(c) Lu/D1 = 63.7, A2/A1 = 16

Sketch 3.3 Examples of effect of inlet Mach number on K t 11

17
 ESDU 90025
4. PERFORMANCE DEPENDENCE ON INLET CONDITIONS

The most important parameters influencing compressible diffuser flow are M 1 and B 1 .

The performance maps, Figures 3 to 18, demonstrate the variation of conical diffuser pressure recovery as
either the inlet Mach number, M 1 , or the inlet pipe length ratio, L u /D 1 (and consequently the inlet boundary
layer blockage, B 1 , and turbulence level, I G ), vary.

The figures relate to the case of naturally-developing flow in the diffuser inlet pipe. However, in many
instances the designer is faced with perturbed inlet flows (Sketch 4.3, Section 4.5); for example, the case
when a diffuser flow is preceded by a shock wave/boundary layer interaction in the inlet pipe, as may occur
in supersonic jet intakes or pipe-diffusers of centrifugal compressors. In such cases, some of the other
diffuser inlet parameters become more influential on diffuser performance. The effects of these parameters
are examined in this Section.
ESDU Copyright material. For current status contact ESDU.

4.1 Effects of Variation in Reynolds Numbers

6
The contours shown in Figures 3 to 18 are derived for an inlet Reynolds number, Re , of about 10 .The
bulk of experimental data suggests that, for incompressible diffuser flow, so long as the inlet flow remains
turbulent, the pressure recovery is substantially independent of Re 9,20,25.

For compressible flow, the effects of Reynolds number are certainly of secondary importance to those of
the inlet Mach number. However, an attempt to decouple the effects of Reynolds number and M 1 for a
subsonic diffuser flow, preceded by a shock wave-boundary layer interaction in the inlet pipe14, reveals
that the effects of increasing Reynolds number on diffuser performance are as follows.

(a) For unstalled geometries, that is below line aa on Sketch 3.1, performance improves ( C pr
increases and K t decreases) as Reynolds number increases.

(b) For geometries near the line of first appreciable stall (line aa in Sketch 3.1), performance is
independent of Reynolds number.

(c) For stalled geometries, (above line aa , Sketch 3.1) performance deteriorates as Reynolds number
increases.

Combined Re-effect
Cpr Cpr
Re-effect predominates, M-effect and M-effect
M-effect
no significant M-effect predominates predominant

M1 0.8 1.0 M1 1.0

Unstalled Stalled

Sketch 4.1 Cpr versus M for unstalled and stalled geometries

1

18
 ESDU 90025
Thus the observed improvement in pressure recovery of unstalled diffusers, with the increase in Mach
number (at Mach numbers well below choking conditions), is related to the accompanying increase in
Reynolds number, while for stalled geometries an increase of intake velocity ( M 1 and Reynolds number)
yields a lower C pr , as illustrated in Sketch 4.1. However, it has been reported (Derivation 14) that, for all
diffusers, outlet profile uniformity improves ( α e decreases) with increase in Reynolds number.

4.2 Effects of Variation in Inlet Turbulence Level

For a given inlet velocity profile, increases in turbulence intensity result in improved diffuser pressure
recovery and a more uniform outlet velocity profile (lower α e ). This was shown for incompressible flow
(for example, Derivations 17, 20 and 23) using a variety of artificial means to generate the turbulence such
as profile generators, grids, screens, rods, trips and roughened inlet pipes. Work on compressible diffuser
flow reported in Derivation 14 (see Sketch 3.2), where I G was varied at constant M 1 and B 1 , revealed
similar trends to those experienced for incompressible flow. Furthermore, it is reported that the
improvement in performance is more significant at the bigger area ratios and for cases of unfavourable inlet
ESDU Copyright material. For current status contact ESDU.

conditions; that is, for cases with higher possibility of early stall. Yet the variations in C pr and α e seem
to level out as I G is further increased, indicating that diffuser sensitivity to variations in inlet turbulence
intensity is high in the low I G range and decreases considerably at higher values.

The variation of K t with I G (Sketch 3.2) depends on the flow regime prevailing at the area ratio considered.
In the unstalled region, an increase of the turbulence level across the diffuser inlet section (free stream and
boundary layer) increases the turbulent transport within the boundary layer inside the diffuser, resulting in
a fuller velocity profile and increased wall shear ( K t increases as I G is increased).

In the stalled flow regimes, an increase of I G delays the actual separation, thus reducing separation losses
at the bigger area ratios, and ultimately reducing K t . For the bigger diffuser angles and higher area ratios
the decrease in K t with I G is nearly linear (rather than levelling out), indicating a continuous downstream
shift of the stall inception.

The values of I G in Sketch 3.2 are those pertaining to the case of a diffuser preceded by a shock
wave/boundary layer interaction, and are much higher than in the case of diffusers preceded by wholly
subsonic flow where the values of I G would not exceed 8%, as shown in Figure 2. The higher I G values
are responsible for the improved performance of diffusers in the former case, as shown in Sketch 3.2.
However, enhancing the inlet turbulence level artificially (using grids, gauzes, rods, etc.) causes an
additional pressure loss in the system that negates any possible improvement in the diffuser pressure
recovery and thus gives no overall improvement in system performance.

4.3 Effects of Variation in Inlet Profile Shape

V V V V

Uniform shear Wake Peaked Asymmetric

Sketch 4.2 Velocity profiles

19
 ESDU 90025
The performance maps presented in Figures 3 to 18 relate to diffusers with naturally-developing boundary
layer inlet flow, where flow parameters at a given section can be characterised approximately by the
non-dimensional length of duct upstream, L u /D 1 . However, for non-naturally developing inlet flows, a
detailed velocity profile must be specified or, at least, both B 1 and H 1 have to be known. Distortions in
velocity profile have different effects when momentum defects occur in the main stream, as compared with
near the diffuser wall. Sketch 4.2 shows some of the possible velocity profiles.

Work reported in Derivations 20, 26 and 27 indicates clearly that the effects of turbulence and inlet shape
factor are interconnected because profile distortion depends on both. Indeed, the value of I G determines
whether the velocity profile will subsequently grow or decay in a pipe. Higher inlet turbulence levels cause
profiles to decay (that is, become more uniform) in the inlet pipe, and retard their growth in the diffuser,
yielding better performance (higher C pr and lower α e ). Wake-type inlet velocity profiles yield better
performance than peaked profiles or even near-uniform flows. This is attributed to the higher energies near
the wall leading to a delayed separation in the diffuser. In general, diffusers with decaying inlet profiles
give an improved performance compared with those with growing inlet profiles27. Such decaying profiles
ESDU Copyright material. For current status contact ESDU.

are common in the cases of diffusers situated downstream of turbo-machines.

4.4 Mach Number Approaching Choking

Compressibility of the diffuser inlet flow affects both the core flow and the boundary layer, especially when
the incoming flow approaches choking.

Effects are very strong for a one-dimensional flow near to M = 1 . For example, at M = 0.9 choking occurs
in a pipe length of two diameters, while the flow curvature implied by the inlet pipe and diffuser junction
and the corner flow reduces this to a pipe length of less than one diameter.

On the other hand, the effects of compressibility on the boundary layers growing in the diffuser are mainly
attributed to changes in fluid properties (density and viscosity) near the wall29, where there is typically a
large temperature rise. Work on compressible turbulent boundary layers in the last three decades has shown
that the turbulence structure is similar to the case of incompressible flow. The increase of viscosity at the
wall (some 33% across the boundary layer if M = 1 ) and the reduction in density result in a lower sublayer
Reynolds number. Thus the boundary layer behaves like an incompressible boundary layer at a lower
free-stream velocity, with a lower skin friction and a greater possibility of separation. Moreover, the lower
density of the near-wall fluid results in a smaller momentum thickness, θ , and a larger shape parameter,
H (30% greater compared with incompressible flow at M = 1 ) and hence earlier separation.

The effects of compressibility on both the core flow and boundary layer, especially at inlet Mach numbers
close to unity, lead to a sharp decline in diffuser performance as the inlet flow approaches choking. With a
boundary layer present at the diffuser throat, the mean inlet Mach number at which the diffuser performance
begins to deteriorate (the critical Mach number) will be less than unity, and the thicker the inlet boundary
layer the lower the critical Mach number. Derivation 28 gives a correlation for the critical Mach number
versus throat blockage.

The performance maps in Figures 3 to 18 give data on diffuser performance up to a mean inlet Mach number
of 0.8. A further increase of the inlet Mach number, even for the case of the thinnest inlet boundary layer
(shortest entry pipe length) considered, will certainly yield a lower pressure recovery.

4.5 Diffuser Preceded by a Shock Wave/Boundary Layer Interaction

The case of diffuser flow affected by a shock wave/boundary layer interaction in the entry pipe has its
application in the design of supersonic aircraft intakes, 'pipe' diffusers for modern high speed centrifugal
compressors and supersonic wind tunnels. In such cases the incoming supersonic flow is transformed into

20
 ESDU 90025
subsonic flow through a shock wave in the diffuser entry duct and then flows through the subsonic diffuser
before entering the next component. Thus, both pressure recovery and flow distortion produced by the
upstream inlet system affect the performance of the following component (the compressor in the case of
intakes) as well as the entire system.

Derivation 14 examines the fundamental differences between the case of a diffuser preceded by such an
interaction and a diffuser preceded by wholly subsonic flow. Basically, the inlet velocity profile in the
former case is in a state of decay with a substantially different turbulence structure from that experienced
in the latter case in which the profile is naturally growing at the diffuser inlet. Sketch 4.326 shows some
typical variations of the diffuser inlet turbulence intensity distribution with shock wave location and entry
pipe length. The resulting I G values are much higher than those for fully-developed pipe flow. This is
evident when comparing I G values in Figure 2 and Sketch 3.2 from Derivation 14. The increased turbulence
levels at the diffuser inlet limit the growth of the velocity profile shape parameter, H , in the diffuser, delay
separation and result in higher pressure recoveries and less distorted outlet profiles from the diffuser. For
practical diffuser geometries the improvement in pressure recovery varies from some 10% for a 5° total
ESDU Copyright material. For current status contact ESDU.

angle diffuser to more than 20% for a 20° diffuser, with the improvement being more significant in cases
of thicker inlet boundary layer and higher inlet Mach numbers.

The main conclusion of Derivation 14 is that for the case of a diffuser preceded by interaction, the design
should be such that the shock wave occurs as early as possible in the entry duct in order to obtain maximum
pressure recovery.

21
 ESDU 90025

0.28 13 12 11

0.24

Pipe centre-line
0.20 Symbol
0.16 Position 11 12 13
I
0.12

0.08

0.04
1/8 2/8 3/8 4/8 5/8
ESDU Copyright material. For current status contact ESDU.

y/R1
(a) Lu/D1 = 5.5

0.24
Symbol
0.20
Position 16 17 18

Pipe centre-line
0.16
I
0.12

0.08

0.04
1/8 2/8 3/8 4/8 5/8
y/R1
(b) Lu/D1 = 6.7

28 27 26
0.20
Pipe centre-line

0.16
Symbol
I 0.12 Position 26 27 28

0.08

0.04
1/8 2/8 3/8 4/8 5/8
y/R1
(c) Lu/D1 = 7.7

Sketch 4.3 Variation of diffuser inlet turbulence intensity distribution with shock location and
entry length for a diffuser with 2φ = 12° and A 2 /A 1 = 10 (Derivation 4)

22
 ESDU 90025
5. PERFORMANCE DEPENDENCE ON SOME OTHER GEOMETRICAL PARAMETERS

5.1 Effects of Fairing the Junction with the Inlet Duct

The performance maps presented in Figures 3 to 18 are for diffusers with a sharp geometrical transition
from their inlet pipes. However, it is always desirable to have a smooth faired entry to the diffuser, especially
at higher included angles and with compressible flow. With a smooth transition from the inlet pipe to the
diffuser cone, the possibility of premature separation in the diffuser is reduced as well as that of early
choking of its inlet flow. Tests reported in Derivation 12 show that increasing the length of the fairing
reduces the static pressure depression at the junction and subsequently yields a higher recovery in the
diffuser. As can be seen from Sketch 5.1 (Derivation 12) the effects of fairing are more significant for the
larger diffuser angles and at higher inlet Mach numbers. A thicker inlet boundary layer, however, tends to
reduce the intensity of the local depression in static pressure, thus slightly improving the performance of
diffusers with sharp transitions. Nevertheless, sharp corners should be avoided whenever possible.
ESDU Copyright material. For current status contact ESDU.

0.3 0.3
Sharp junction
Faired junction
0.2 0.2

x 0.1 0.1
Faired
Cpr Cpr

Sharp 0 0
R1 2 4 6 1 2 3 4
x/R1 x/R1
3.2R1
–0.1 –0.1

–0.2 –0.2

(a) 2φ = 5˚, M1 = 0.16 (b) 2φ = 12˚, M1 = 0.7

Sketch 5.1 Effect of fairing the inlet junction on C pr 12

5.2 Effects of Using a Tailpipe

The addition of a section of parallel duct downstream of a diffuser allows the non-uniform velocity profile
leaving the diffuser to re-distribute, leading to a decrease in α e . In the process a further increase in static
pressure is usually achieved at the expense of an increase in total pressure loss, due to mixing. If the exit
pipe is long enough, fully-developed flow conditions are attained, yielding a more complete pressure
recovery and considerably reducing the effect of variation in inlet conditions.

23
 ESDU 90025

0.7

0.6

Cpr 0.5

0.4

0.3

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

M1
ESDU Copyright material. For current status contact ESDU.

(a) 2φ = 12˚, A2/A1 = 2

At diffuser exit plane (with tailpipe)

At 1.75D2 downstream in tailpipe

0.6

0.5

0.4
Cpr
0.3

0.2

0.1
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
M1
(b) 2φ = 20˚, A2/A1 = 2

Sketch 5.2 Variation of C pr with M 1 showing effect of tailpipe exit duct

Sketch 5.2, from Derivation 12, shows that, even with a tailpipe less than two diameters long, an increase
of C pr of up to 35% is achieved for a 12° diffuser of area ratio of 2. However, experience with
incompressible diffuser flow indicates that, for diffusers with naturally-developing inlet flow when the
geometry lies within the region of transitory separation (Sketch 3.1), a tail pipe length, L d , of at least 4D 2
is required for full pressure recovery. With less favourable inlet conditions and/or at higher area ratios (that
is, for geometries in the jet flow or fully-stalled region), a longer tail pipe is needed, up to 8D 2 . Thus, in
order to take advantage of this recovery, for most practical applications, 4 < L d /D 2 < 8 .

5.3 Other Design Improvements

ESDU 87015, Derivation 3, presents design recommendations and guidance on performance improvement
by various techniques for incompressible flow. Few data are available on the worthiness of these design
modifications for compressible flow.

24
 ESDU 90025
6. DERIVATION AND REFERENCES

6.1 Derivation

The derivation lists selected sources that have assisted in the preparation of this Item.

1. ESDU Introduction to design and performance data for diffusers.

ESDU 76027, ESDU International, London, 1976.
2. ESDU Performance of conical diffusers in incompressible flow.
ESDU 73024, ESDU International, London, 1980.
3. ESDU Performance improvement of axial diffusers for incompressible
flow. ESDU 87015, ESDU International, London, 1987.
ESDU Copyright material. For current status contact ESDU.

4. LIVESEY, J.L. Suitable mean values in one-dimensional gas dynamics. J. mech.

HUGH, T. engng Sci., Vol. 8, No. 4, pp. 374-383, 1966. (See also Derivation 7.)
5. AGARD Suitable averaging techniques in non-uniform internal flows.
AGARD Advisory Group 14, Report AGARD-AR-182, 1983.
6. LIVESEY, J.L. Duct performance parameters considering spatially non-uniform
flow. AIAA Paper No. 72-85, AIAA 10th Aerospace Sciences
Meeting, San Diego, California, 1972.
7. LIVESEY, J.L. Corrigendum to J. mech. engng Sci., Vol. 8, No. 4, pp. 374-383,
1966. J. mech. engng Sci., Vol. 9, No. 3, pp. 243-245, 1967.
8. VAN DEWOESTINE, R.V. An experimental investigation on the effect of subsonic inlet Mach
FOX, R.W number on the performance of conical diffusers. Int. J. mech. engng
Sci., Vol. 8, No. 12, pp. 759-769, 1966. Also available as Rep.
FMTR-66-1, Fluid Mechanics Group, Purdue Univ., USA, 1966.
(See also Derivation 9.)
9. FOX, R.W. Subsonic flow in conical diffusers. Tech. Rep. FMTR-67-1, Fluid
Mechanics Group, Purdue Univ., USA, 1967. (Contains a corrected
version of the data presented in Derivation 8.)
10. NORTH, P. The performance of conical diffusers with compressible flow. Ph.D.
Thesis, Nottingham Univ., 1967.
11. HUGH, T. Performance of conical diffusers with high subsonic entry Mach
numbers. Ph.D. Thesis, Manchester Univ., 1969.
12. ASHCROFT, P.A. Performance of conical diffusers with high entry Mach numbers.
Ph.D. Thesis, Manchester Univ., 1968.
13. ODUKWE, A.O. Performance of subsonic conical diffusers including the effects of
preceding normal shock boundary layer interaction. Ph.D. Thesis,
Salford Univ., 1971.
14. KAMAL, W.A. Inlet flow structure effects on compressible conical diffuser flow.
Ph.D. Thesis, Salford Univ., 1976.

25
 ESDU 90025
15. LITTLE, B. Performance and boundary layer data from 12° and 23° conical
WILBUR, S. diffusers of area ratio 2.0 at Mach numbers up to choking and Re up
6
to 7.5 ×10 . Nat. Adv. Comm. Aero. Rep. No. 1201, 1954.
16. SAJBEN, M. Conical diffuser flow with natural and screen-simulated inlet
KROUTIL, J. conditions. Am. Inst. Aero. Astro. J., Vol. 14, No. 12, pp. 1723-1730,
SEDRICK, A. 1976.
17. POZZORINI, R. Das turbulent Strömungsfeld in einem langen Kreiskegel-Diffusor.
Ph.D. Dissertation, ETH, Zürich, published by Eduard Truninger
AG, Zürich, 1976.
18. SOVRAN, G. Experimentally determined optimum geometries for rectilinear
KLOMP, E.D. diffusers with rectangular, conical or annular cross-section. In Fluid
Mechanics of Internal Flow, Sovran, G., Editor, Elsevier Publishing
Co., pp. 270-319, 1967.
ESDU Copyright material. For current status contact ESDU.

19. ISHIKAWA, K. Performance charts and optimum geometries of conical diffusers

NAKAMURA, I. with uniform inlet flow and free discharge, Jap. Soc. mech. Engrs J.,
Series II, Vol. 32, No. 4, pp. 559-567, 1989.
20. KLEIN, A. Review: effects of inlet conditions on conical diffuser performance,
J. fluids Engng, Vol. 103, pp. 250-257, June 1981.
21. KLEIN, A. Review: turbulent developing pipe flow, J. fluids Engng, Vol. 103,
pp. 243-249, June 1981.
22. JAPIKSE, D. A new diffuser mapping technique, Part 1: studies in component
performance, J. fluids Engng, Vol. 108, pp. 148-156, June 1986.
23. BRADLEY, I. The response of diffusers to flow conditions at their inlet.
COCKRELL, D.J. Symposium on Internal Flows, Salford Univ., Paper 5, pp. A32-A41,
1971.
24. MILLER, D.S. Internal flow systems. Brit. Hydromech. Res. Assoc., Cranfield, UK,
p. 168, 1978.
25. McDONALD, A.T. Behaviour of diffusers with disturbed and unsteady inlet conditions.
RUNSTADLER, P.W. Symposium on Fluid Dynamics of Unsteady Three-dimensional and
Separated Flows, Project SQUID, Office of Naval Research, School
of Aerospace Engng, Georgia Inst. Tech., Atlanta, June 1971.
26. KAMAL, W.A. Prediction of diffuser flow and performance following a normal
LIVESEY, J.L. shockwave-turbulent boundary layer interaction. Proc. First
Symposium on Turbulent Shear Flows, pp. 12.1-12.10. Penn. State
Univ., USA, 1977.
27. LIVESEY, J.L. The effect of velocity profile decay on shear flow in diffusers. Int. J.
TURNER, J. mech. Sci., Vol. 6, No. 5, pp. 371-379, 1964.
28. HALLEEN, R.W. Inlet Mach number effects on subsonic diffuser performance. Creare
JOHNSTON, J.P. Tech. Note TN-56, Creare Inc., Hanover, N.H., USA, February
1966.
29. CHILDS, R. A computational method for subsonic compressible flow in
FERZIGER, R. diffusers. Rep. PD-24, Thermosciences Division, Mech. Engng
KLINE, S. Dept., Stanford Univ., USA, 1981.

26
 ESDU 90025
30. ESDU Performance in incompressible flow of plane-walled diffusers with
single-plane expansion. ESDU 74015, ESDU International, London,
1974.
31. ESDU Performance of circular annular diffusers in incompressible flow.
ESDU 75026, ESDU International, London, 1975.

6.2 References

The references given are recommended sources of information supplementary to that in this Item

.
32. RUNSTADLER, P.W. Straight channel diffuser performance at high inlet Mach numbers. J.
ESDU Copyright material. For current status contact ESDU.

DEAN, R.C. basic Engng, Vol.91, pp.397-422, September 1969.

33. DOLAN, F.X. Pressure recovery performance of conical diffusers at high subsonic
RUNSTADLER, P.W. Mach numbers. NASA CR-2299, August 1973.
34. RUNSTADLER, P.W. Diffuser data book. Creare Tech. Note TN-186, Creare Inc., Hanover,
DOLAN, F.X. NH, USA, May 1975.
DEAN, R.C.

27
 ESDU 90025

0.20

0.16

B1 0.12

X
13
0.08 Odukwe
X 8
Van Dewoestine and Fox
X X North
10
12
ESDU Copyright material. For current status contact ESDU.

X Ashcroft
0.04 20
X Klein (Incomp. pipe flow)

0.00
0 10 20 30 40 50 60 70 80 90 100 110 120

Lu / D1

FIGURE 1 VARIATION OF DIFFUSER INLET BLOCKAGE RATIO WITH UPSTREAM PIPE LENGTH

10

max

8
Compressible

max

6
lGH100

4
Incompressible
M1 Re
LOW 2.4 H 105
2 0.6 8.6 H 105
0.6 10.4 H 105

0
0 10 20 30 40 50
Lu/D1
FIGURE 2 VARIATION OF DIFFUSER INLET TURBULENCE PARAMETER WITH UPSTREAM PIPE
LENGTH (DERIVATION 14)

28
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

Cpr 0.45 0.55 0.65 0.75

6

10
Cpr
9 0.85 4

5
5
0 .2
0 .3
8
7
0.8
6

A2 / A1 5
0.7
4
0.6
ESDU Copyright material. For current status contact ESDU.

3
0.5

0.4
Cpr*
0.3
0.2
Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 3 PERFORMANCE MAP FOR M 1 = 0.2 AND Lu / D1 = 0 - 1.0

29
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

6
Cpr 0.25 0.35 0.45 0.55 0.65

0.75
10
9 Cpr
4
8 0.8
7

6 0.7
0.6
A2 / A1 5 0.5

4
0.4
ESDU Copyright material. For current status contact ESDU.

2
0.3
Cpr* 0.2

Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 4 PERFORMANCE MAP FOR M = 0.2 AND Lu / D1 = 9 - 10.5

1

30
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

6
Cpr 0.2 0.3 0.4 0.5 0.6 0.65 0.75

0.85
10
9 4
8
7

6 0.8

A2 / A1 5
4
ESDU Copyright material. For current status contact ESDU.

0.7

0.6
2 Cpr* 0.5
0.4
0.3
0.2

Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 5 PERFORMANCE MAP FOR M = 0.2 AND Lu / D1 = 20

1

31
 ESDU 90025

2φ = 30° 20 15 12 10 8
20
Cpr

65
55
45
0.2
0.3

4
6

0.
0.
0.
0.
5
0.7

10
9 4
8 0.85
7

6
0.8
5
A2 / A1
4
ESDU Copyright material. For current status contact ESDU.

3
0.7

0.6
2
Cpr* 0.5
0.4
0.2

Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 6 PERFORMANCE MAP FOR M = 0.2 AND Lu / D1 = 30 - 35.5

1

32
 ESDU 90025
2φ = 30° 20 15 12 10 8
20

Cpr 6

5
5
5

5
0.2

0.6
0.4
0 .3

0 .5

0 .7
0.8
10
9 4
8
7

A2 / A1 5
0.7
4
ESDU Copyright material. For current status contact ESDU.

0.6
3

2 0.5

Cpr* 0.4

0.3
Cpr** 0.2
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 7 PERFORMANCE MAP FOR M = 0.4 AND Lu / D1 = 0 - 1.0

1

33
 ESDU 90025
2φ = 30° 20 15 12 10 8
20

0.75
6

5
Cpr

5
5

0 .5
0 .3

0 .4

0 .6
0 .2
10
9 4
8
7
0.7
6

A2 / A1 5
4
0.6
ESDU Copyright material. For current status contact ESDU.

3 0.5

Cpr* 0.4

0.3
Cpr**
0.2
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 8 PERFORMANCE MAP FOR M = 0.4 AND Lu / D1 = 9 - 10.5

1

34
 ESDU 90025

2φ = 30° 20 15 12 10 8
20
Cpr 5

5
5
0.7

0 .2

0 .3

0 .5

65
0.4
6

0.
0.8

10
9 4
8
7
0.75
6

A2 / A1 5
0.7
4
ESDU Copyright material. For current status contact ESDU.

0.6

0.5
2
0.4

Cpr* 0.3
0.2
Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 9 PERFORMANCE MAP FOR M = 0.4 AND Lu / D1 = 20

1

35
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

Cpr

5
5
6

5
5

0 .7
5

0 .6
0.4
0 .3
0.2

0.5
10

0.8
9 4
8
0.7
7

5
A2 / A1 0.6
4
ESDU Copyright material. For current status contact ESDU.

0.5
2
Cpr*
0.4

0.3
Cpr**
0.2
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 10 PERFORMANCE MAP FOR M = 0.4 AND Lu / D1 = 30 - 35.5

1

36
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

Cpr

5
5
5

5
5

0 .6
0 .3
0 .2
0 .7

0 .5
0 .4
6
0.75

0.8 0.7

10
9 4
8
7

A2 / A1 5
4
0.6
ESDU Copyright material. For current status contact ESDU.

0.5
2

0.4
Cpr* 0.3
0.2

Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 11 PERFORMANCE MAP FOR M = 0.6 AND Lu / D1 = 0 - 1.0

1

37
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

Cpr

5
5

5
0 .2

0.3

0 .6
0.4

0 .5
6
0.75

0.8
10
9 4
0.85
8
7

A 2 / A1 5 0.7

4
ESDU Copyright material. For current status contact ESDU.

0.6
3

2 0.5
Cpr*
0.4

0.3
0.2
Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 12 PERFORMANCE MAP FOR M = 0.6 AND Lu / D1 = 9- 10.5

1

38
 ESDU 90025

2φ = 30° 20 15 12 10 8
20
0.75

5
Cpr

0 .2

0.6
0.5
0 .4

0 .6
0 .3
6
0.8

10
9 4
8
7

A2 / A1 5
4
ESDU Copyright material. For current status contact ESDU.

0.7

3
0.6

2 0.5

Cpr* 0.4
0.3

0.2
Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 13 PERFORMANCE MAP FOR M = 0.6 AND Lu / D1 = 20

1

39
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

5
5 6

5
Cpr

0 .3

0 .4
0 .6

0 .1

0 .2

5
0.
0.7

10
9 4
8 0.75
7

5
A2 / A 1 0.6
4
ESDU Copyright material. For current status contact ESDU.

0.5
2
0.4
Cpr*
0.3
Cpr**

0.2
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 14 PERFORMANCE MAP FOR M = 0.6 AND Lu / D1 = 30 - 35.5

1

40
 ESDU 90025

2φ = 30° 20 15 12 10 8
20
5
Cpr 0 .7 6

5
0 .3
0 .2
0 .3

0 .6
0.4

0 .5

7
0.
0.8

0.85
10
9 4
8
7

5
A2 / A 1 0.7
ESDU Copyright material. For current status contact ESDU.

3 0.6

Cpr*
2 0.5

0.4

0.3
0.2
Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 15 PERFORMANCE MAP FOR M = 0.8 AND Lu / D1 = 0 - 1.0

1

41
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

Cpr

55
65

75
5
3

4
6

0.
0.

0.
0.

0.
5

0.
0.
0 .8

0.9
10
9 4
8
7
0.8
6

5
A 2 / A1
ESDU Copyright material. For current status contact ESDU.

4
0.7

0.6

2
0.5
Cpr*
0.4

0.3
Cpr** 0.2
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 16 PERFORMANCE MAP FOR M = 0.8 AND Lu / D1 = 9 - 10.5

1

42
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

5
5

5
5
Cpr

75
5

0. 6
0 .2

0. 4
0. 3

0. 5

0.
10
9 0.85 4
8
7

5 0.8
A 2 / A1
0.7
4
ESDU Copyright material. For current status contact ESDU.

0.6
Cpr* 0.5
2
0.4

0.3
0.2
Cpr**
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 17 PERFORMANCE MAP FOR M = 0.8 AND Lu / D1 = 0 - 20

1

43
 ESDU 90025

2φ = 30° 20 15 12 10 8
20

5
5

5
Cpr 5

0 .1

0 .3
0 .2

0 .4
0. 6

0 .5
0.7
10
9 4
8
7

5
A2 / A1
ESDU Copyright material. For current status contact ESDU.

4 0.6

0.5

2 Cpr* 0.4

0.3
0.2
Cpr**
0.1
1
1 2 3 4 5 6 7 8 9 10 20 30 40 50 60

L / R1

FIGURE 18 PERFORMANCE MAP FOR M = 0.8 AND Lu / D1 = 30 - 35.5

1

44
 ESDU 90025
KEEPING UP TO DATE

Whenever Items are revised, subscribers to the service automatically receive the material required to update
the appropriate Volumes. If you are in any doubt as to whether or not your ESDU holding is up to date, please
contact us.

Please address all technical engineering enquiries and suggestions to:

ESDU International plc Tel: 020 7490 5151 (from the UK)
+44 20 7490 5151 (from outside the UK)
Fax: 020 7490 2701 (from the UK)
+44 20 7490 2701 (from outside the UK)
E-Mail: [email protected]
Website: www.ihsesdu.com
ESDU Copyright material. For current status contact ESDU.

For users in the USA, please address all Customer Service and Support enquiries and suggestions to:

IHS Engineering Products Tel: 1 800 447 3352 (toll free number)
and Fax: 1 303 397 2599
Global Engineering Documents Website: www.ihs.com
www.global.ihs.com
 ESDU 90025
Performance of conical diffusers in subsonic compressible flow
ESDU 90025

ISBN 978 0 85679 751 4, ISSN 0141-4011

Available as part of the ESDU Series on Fluid Mechanics. For information

on all ESDU validated engineering data contact ESDU International plc,
27 Corsham Street, London N1 6UA.

ESDU 90025 gives performance maps for straight axis diffusers with
sharp transition from an inlet pipe with naturally-developing flow. They
apply for an inlet Reynolds number (based on pipe diameter) of one
million. They plot static pressure recovery against diffuser area ratio
(inlet/outlet) and length ratio (length/inlet radius) and each applies to a
specific value of inlet Mach number (from 0.2 to 0.8) and a limited range
of inlet pipe length/diameter ratios (from 0 to 35.5). Shown on each map
are curves of optimum performance for either given length ratio or given
ESDU Copyright material. For current status contact ESDU.

area ratio. The maps, whose use is illustrated by two worked examples,
can be applied to determine the performance of a given design or to
design an optimum diffuser. An approximate method is suggested for
deriving the total head loss from the static pressure recovery. Various
other influences on diffuser performance are discussed and illustrated
with sketches for specific cases. They include Reynolds number variation,
inlet turbulence intensity, inlet velocity profile shape, upstream shock
wave/boundary layer interaction, the approach to choking Mach number,
fairing the inlet/diffuser junction, and fitting a tailpipe.

© ESDU International plc, 2007

All rights are reserved. No part of any Data Item may be reprinted, reproduced, or
transmitted in any form or by any means, optical, electronic or mechanical including
photocopying, recording or by any information storage and retrieval system without
permission from ESDU International plc in writing. Save for such permission all
copyright and other intellectual property rights belong to ESDU International plc.