RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

RDDM
Recherche Design Developpement Management

Turbomachinary Lecture Series

Module 04 – Axial Flow Compressors [x]

Presented to - Présenté à

Ecole Polytechnique

AER4270 Propulsion Aeronautique

01 January 2022 – 01 Janvier 2022

Copyright Statement - Déclaration de copyright

This course handbook is mainly used as Ce manuel de cours est surtout utilisé
reference material for lectures given at comme la matière de référence pour les
Recherche Developpement Design conférences données à Recherche
Management (RDDM) and at Developpement Design Management
educational institutions in Canada and (RDDM) et aux institutions éducatives
U.S.A. Figures and text used in this au Canada et à U.S.A. Figures et textes
volume are from the reference listed in utilisées dans ce volume sont de la
the module. référence énumérée dans le module.

Page 1 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
REFERENCES

[1] textbook, David Gordon Wilson, sixth printing 1993

SUBJECT: The design of high-efficiency turbomachinery and gas turbines

[2] textbook, HIH Saravanamuttoo, GFC Rogers, H Cohen, 5th edition

SUBJECT: Gas Turbine Theory

[3] textbook, S.L. Dixon, 4th Edition

SUBJECT: Fluid Mechanics and Thermodynamics of Turbomachinery

[4] master thesis, Lund Univeristy, Niclas Falck, 2008

SUBJECT: Axial flow compressor mean line design

[5] NACA report 1368, James C. Emery, L. Joseph Herrig, John R. Erwin, A. Richard Felix
SUBJECT: Systematic two-dimensional cascade tests of NACA 65-series compressor blades at low
speeds

[6] NASA, AIAA-2009-1641, Joseph P. Veres

SUBJECT: Axial and centrifugal compressor mean line flow analysis method

[7] Presentation, B&B-AGEMA, 23 March 2011

SUBJECT: Axial compressor design with AixCompFlow 2D and STAR-CCM+

[8] AGARD, Report No. 745, November 1987

SUBJECT: Application of modified loss and deviation correlations to transonic axial compressors

[9] ASME paper GT2009-60141, Chris Drummond, Craig R. Davison

SUBJECT: Capturing the shape variance in gas turbine compressor maps

[10] Rolls-Royce PowerPoint Presentation, Michael Cervenka, Cambridge University 2000

SUBJECT: The Rolls-Royce Trent Engine

[11] NASA technical paper 1337, Lonnie Reid and Royce D. Moore, 1978
SUBJECT: Design and Overall Performance of Four Highly Loaded, High-Speed Inlet Stages for an
Advanced High-Pressure-Ratio Core Compressor

[12] master thesis, Virginia Polytechnic Institute and State University, Joseph E. Cahill, 1997
SUBJECT: Identification and Evaluation of Loss and Deviation Models for use in Transonic Compressor
Stage Performance Prediction

[13] article (original source unknown), Meherwan P. Boyce

SUBJECT: Axial-Flow Compressors

[14] master thesis, Politecnico Di Milano, Matteo Resmini, 2011/2012

SUBJECT: Numerical investigation of the solidity effect on a linear compressor cascade performance and
stability

[15] master thesis, Lund University, Daniele Perrotti, date not mentioned
SUBJECT: Two Dimensional Design of Axial Compressor; An Enhanced Version of LUAX-C

[16] Trans. of the ASME, Journal of Eng. for Power, volume 98, p.411-424, 1976
by Koch C.C and Smith L.H. (File: 411_1-Koch-Smith)
SUBJECT: Loss sources and magnitudes in axial-flow compressors

[17] General Electric Youtube video image captured on 2017

https://www.youtube.com/watch?v=wN36mHhD8-Q

Page 2 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
[18] Brazilian Congress of Thermal Sciences and Engineering ENCIT. 2010
By Barbosa, J.R. and J.S. Figueiredo.
SUBJECT: Surging to choking: a study of the axial channel of axial compressors

[18] NASA technical paper 2597, Ronald J. Steinke, May 1986

SUBJECT: Design of 9.271 pressure-ratio five-stage core compressor and overall performance for first
three stages

[x] RDDM internal memo

SUBJECT: Literature research - axial flow compressors

Page 3 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

Table of Contents

REFERENCES .............................................................................................................................................. 2
INTRODUCTION........................................................................................................................................... 5
DISCUSSION – Axial-flow compressor theory & flow physics ..................................................................... 7
o Rothalpy and Velocity Triangles ................................................................................................ 11
o Compressor Reaction ................................................................................................................ 12
o  - Stage loading coefficient [1, 3] ............................................................................................. 14
o  - Stage flow coefficient ............................................................................................................ 14
DISCUSSION – Compressor Surge & Stall ................................................................................................ 15
o COMPRESSOR SURGE ........................................................................................................... 15
o COMPRESSOR STALL ............................................................................................................. 16
o ROTATING STALL .................................................................................................................... 17
DISCUSSION – Compressor Stage Design ............................................................................................... 18
o Estimating the number of compressor stages for a given pressure-ratio .................................. 21
o Work done factor [2, 3] .............................................................................................................. 22
DISCUSSION – Various Compressor Maps ............................................................................................... 22
DISCUSSION – Compressor blade design................................................................................................. 23
DISCUSSION - Compressor Loss Modeling............................................................................................... 25
o “de Haller” number ..................................................................................................................... 27
o Diffusion Factor (DF) and Diffusion Ratio (DR) ......................................................................... 28
o The solidity in axial flow compressor design ............................................................................. 30
o Incidence Angles, and Reference Incidence Correlation [4] ..................................................... 31
o Deviation Angle, Reference Deviation Correlation and Slip Factors ......................................... 33
o Total loss.................................................................................................................................... 36
o Profile loss, based on Wright & Miller (1991) [5] ....................................................................... 38
o Secondary, or End Wall Loss, based on Wright & Miller (1991) [5] .......................................... 39
o Shock Losses ............................................................................................................................ 40
o Incidence Losses ....................................................................................................................... 40
o Off-Design Compressor Map ..................................................................................................... 40
DISCUSSION – Surge Prediction ............................................................................................................... 41
DISCUSSION – Compressor mean-line or reduced order throughflow methodology ................................ 43
DISCUSSION – Compressor Design Rules of Thumb (ROT) .................................................................... 45
DISCUSSION – Compressor Stall Flutter ................................................................................................... 45
DISCUSSION – Benchmarking Compressor Software ............................................................................... 45
APPENDIX – LOSS MODELLING .............................................................................................................. 47
o Solidity ....................................................................................................................................... 47
o Shock Losses ............................................................................................................................ 48

Page 4 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

INTRODUCTION

“Since the implementation of gas turbine engines for the purpose of aircraft propulsion, continuous
progress has been made in their performance, reliability, and efficiency. However, there remains a
continuing need for continued improvement with regard to noise, cost, efficiency, power, safety, and
weight. Unfortunately, an improvement in any one of these areas will often lead to detrimental effects in
another area. Therefore, different types of aircraft have specific engine design criteria which are
optimized for the particular application. In response to the demand for improved aircraft gas turbine
engine performance, a great deal of research has focused on enhancing the design of axial-flow
compressors.” [12]

The axial flow compressor has the potential for both higher-pressure ratio and higher efficiency than the
centrifugal (radial-flow) compressor. The centrifugal compressor often consists of a single stage and it is
possible to obtain a higher-pressure rise over a single centrifugal compressor, as compared to a single
axial compressor stage. Another major advantage that axial-flow compressors have, especially for turbo-
jet and turbo-fan engines, is that a much larger flow rate is possible for a given frontal area. The axial-
flow machine dominates the field for large powers, and centrifugal compressors are restricted to the lower
end of the power requirements where the flow is too small to be handled efficiently by axial blade rows.

Through the decades, compressor aerodynamic research and development has resulted in a steady
increase in axial-flow compressor pressure ratio, with the effect of reducing the number of stages required
to produce those high pressure ratios. However, even though reducing the number of stages has the
benefit of decreasing overall engine weight for a specific performance, the higher pressure ratios result in
high Mach numbers and large gas deflections in compressor stage blade rows.

“Their operating principle were established [at the start of the 1950’s] more than sixty years ago, and
during the last decades [and since then] all the efforts have been concentrated on how to improve and
develop better machine, and on the study of the behaviour of the elaborated flow.” [15]

Figure 1: Pressure ratio increase along the years [15]

“The difficulties associated with the development of axial-flow compressors stems mainly from the
fundamentally different nature of the flow process compared to that in axial-flow turbines. Whereas in the
axial-flow turbine the flow relative to each blade row is accelerated, in the axial compressor the flow is
decelerated. It is widely known that although a fluid can be rapidly accelerated through a passage and
sustain a small or moderate loss in stagnation (total) pressure, the same is not true for a rapid
deceleration. In the latter case large pressure losses would arise as a result of severe stall caused by a
large adverse pressure gradient. So as to limit the total pressure losses during flow diffusion it is
necessary for the rate of deceleration and turning in the compressor blade passage to be severely
restricted. It is mainly because of these restrictions that axial-flow compressors need to have many
stages for a given pressure ratio compared to axial-flow turbines which need only a few.” [3]

Careful designs of axial-flow compressor blade rows are based both on aerodynamic theory and
experimental data. This is required not only to reduce the various losses that impact overall compressor

Page 5 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
stage efficiency, but also to ensure a minimum of stalling issues that are prevalent in axial-flow
compressor designs.

“The two available means of evaluating new compressor designs are experimental and computational
methods. The advantages of experimental testing are more accurate and provide reliable results.
Therefore, physical experimental testing is required prior to implementation of a new design. The
advantages of computational simulation are speed, reduced cost, more data, and rapid design
modifications. Due to time and cost considerations involved in experimental testing of new compressors,
computational simulations have been used more extensively [since 2000] in recent years. Through the
use of advanced and accurate computer simulations, much of the preliminary experimental testing may
be eliminated.” [12]

“Successful compressor designs are very much an art, and all the major engine manufacturers have
developed a body of knowledge that is kept proprietary for competitive reasons.” [2] However, once in a
while, after a certain time period, one is able to gather information through publications, university thesis,
patents, even news clipping about the various design aspects for axial compressors. A simple Internet
search can uncover valuable information that was once unavailable decades ago.

Page 6 of 50
 RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

DISCUSSION – Axial-flow compressor theory & flow physics

A compressor stage is defined as a rotor blade row (rotating circumferentially about a central access at a specified
RPM) followed by a stator blade row (which remains stationary). The working fluid is initially accelerated by the
rotor blades thus gaining energy, and then decelerated in the stator blade passages wherein the kinetic
energy is converted to static pressure. The flow is always subject to an adverse pressure gradient. This
adverse pressure gradient, developed during compression, limits the amount of pressure rise that a single
axial-flow compressor stage can provide. The higher the pressure ratio the more difficult it becomes to
design the axial-flow compressor.

RPM

th
ro
SS

t a
PS
s


Rotor Stator
Blades Vanes

Figure 2: Rotor-Stator configuration of a compressor stage

Table 1: Compressor blade row nomenclature
SS Suction Surface camber line the symmetric line of the blade
chord line or the line which connect directly the leading and trailing
PS Pressure Surface
true chord edge
S Pitch camber angle The turning angle of the camber line
C True Chord
 Stagger angle

Figure 3: axial compressor blade [nomenclature not included] [15]

Page 7 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

All the power is absorbed in the compressor rotor. The stator merely transforms the kinetic energy to an
increase in static pressure with the stagnation (total) temperature remaining constant if we assume the
compression process to be adiabatic. The increase in stagnation (total) pressure is accomplished wholly
in the rotor. The rotor and stator will create stagnation (total) pressure losses due to fluid friction and
other loss creating flow mechanisms. There is a strong link and interdependence among the pressure
rise, the temperature rise, and the compressor blade shapes. To understand this link one must first
understand the impact of compressor velocity triangles. A typical axial-flow compressor velocity triangle,
and its accompanying nomenclature, is shown below.

The fluid enters the rotor with an absolute velocity Vabs1 with a flow angle from the axial of 1. Combining
the absolute velocity with the blade speed U will give the relative velocity Vrel1 and the relative inlet flow
angle  1 with respect to the rotor. The mechanical energy from the rotating rotors will be transferred to
the working fluid, increasing the stagnation (total) temperature and stagnation (total) pressure. The
energy absorption will also increase the absolute velocity of the fluid. At the exit of the rotor, the fluid will
have a relative velocity of Vrel2 and a relative exit flow angle of  2. If there are no radial changes between
rotor exit and stator inlet, then it may be assumed that what exits the rotor remains the same as it enters
the stator. The fluid leaving the rotor is consequently the same fluid entering the stator rows ( station 3
velocity vectors equal station 2 velocity vectors). Combining the relative exit velocity with the blade speed U will
give the absolute velocity Vabs3 and absolute flow angle 3 that enters the stator. The fluid entering the
stator will be further diffused and will leave the stator with an absolute velocity of Vabs4 and an absolute
flow angle of 4.

V rel2
VrelT2

2
2 VabsX2
2
U
VabsT2

1
V ab
s2

Rotor
Stator
l1
3 VabsX4
VrelT1

V re

VabsT4
1 V 4
RPM
U

absX1
V
abs
1 4
VabsT1

V
ab
s 1 4

Figure 4: Typical axial-flow compressor velocity triangles

NOTE: a so-called “normal” compressor stage is one where the absolute velocities and flow directions at
compressor stage outlet (station 4) are the same as at the compressor stage inlet (station 1) [3]. This
type of configuration is a simplified design of the more generic design where the outlet and inlet
conditions do not equal to each other.

If there are radial differences between the rotor exit (station 2) and stator inlet (station 3) then the
conservation of angular momentum is used to adjust the velocity vectors at station 3 from the velocity
vectors at station 2, as follows:
m 2 rm 2
m 3rm3VabsT 3  m 2 rm 2VabsT 2 VabsT 3  VabsT 2
m 3rm 3

Page 8 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
The compression process consists of a series of diffusions, both in the rotor and the stator blade
passages; although the absolute velocity of the fluid is increased in the rotor, the relative velocity in the
rotor is decreased. “Using the above velocity triangle nomenclature for a compressor stage, the power
input of a compressor is given by (assuming constant mass flow and gas properties)” [2]

Work  m C p TO 2  TO1 

W
Specific Work   ho  C p TO 2  TO1 
m

hO  hO 2  hO1  C p T02  C p T01

 2
Vabs   2
Vabs 
hO   hS 2  2
   hS 1 
 
1

 2   2 

The change of momentum between the flow entering and leaving the rotor can be used to calculate the
force on the rotor, and the equal and opposite forces on the fluid. There are three principal force
components: Axial, Radial, and Tangential

The axial and radial force components cannot contribute to the work transfer between the working fluid
and the rotor. Only the tangential force can produce a change in enthalpy through work transfer.

o The tangential forces on the rotor due to the entering fluid and exiting fluid are
TangentialForce @ Inlet  m VabsT 1 TangentialForce @ Exit  m VabsT 2

o The torque produced is calculated by the distance from the central axis to radial position of the
rotor
Torque @ Inlet  r1 m VabsT 1  Torque @ Exit  r2 m VabsT 2 

o The net torque on the rotor will be the difference between the torque produced by the entering
fluid and the torque produced by the exiting fluid
NetTorque  r1 m VabsT 1   r2 m VabsT 2 

o The energy transfer will be the net torque and the speed at which the rotor is spinning
EnergyTransfer  Work   r1 m VabsT 1   r2 m VabsT 2 

o We know that r = the rotor peripheral speed at the radius of the fluid entering or exiting the rotor.
If we let Um be equal to the rotor mean radius speed rm then we’ll get
EnergyTransfer  Work  U m1 m VabsT 1   U m 2 m VabsT 2 

Work
SpecificWork   hO  U m1VabsT 1  U m 2VabsT 2
m
The above equation is known as Euler’s equation for turbomachines. Positive work means that work is
delivered from the turbomachine shaft (the turbine delivers the work rather than the compressor). We also use this
equation to calculate the energy transferred to the fluid by the compressor. A simple swap of the velocity
triangle terms handles this case.
Work
SpecificWork   hO  U 2VabsT 2  U 1VabsT 1
m
Expanding VabsT, by using basic trigonometry, we’ll get the above equation in terms of fluid flow angles

Page 9 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

hO  U m 2 VabsX 2 tan  2   U m1 VabsX 1 tan  1 

We would like to cast ho in terms of the relative flow angles, because these angles will define the rotor
inlet and exit metal angles (if we use the simplifying assumption that the design point solution equals the rotor metal angles).
Again, using basic trigonometry and the velocity triangles, we will find that

U m1  VabsX 1 tan  1  tan  1  U m 2  VabsX 2 tan  2  tan  2 

U m1 U m2
 tan  1   tan  1  tan  2   tan  2
Vabsx1 Vabsx 2
 U   U 
 hO  U m 2VabsX 2  m 2  tan  2   U m1VabsX 1  m1  tan  1 
 VabsX 2   VabsX 1 

 hO  U m2 2  U m 2VabsX 2 tan  2  U m21  U m1VabsX 1 tan  1

 
 hO  U m2 2  U m21  U m 2VabsX 2 tan  2  U m1VabsX 1 tan  1

For a “purely” axial compressor, rm1=rm2 then Um2=Um1, and if the inlet and exit area of the rotor are
equivalent then Vabsx1 = Vabsx2. Therefore hO for a “purely” axial compressor becomes

hO  U mVabsX tan  1  tan  2 

Combining the following equations together we’ll get

hO  hO 2  hO1  C p T02  C p T01 and hO  U mVabsX tan  1  tan  2 

 C p T02  C p T01  U mVabsX tan  1  tan  2 

U mVabsX
 T02  T01  tan 1  tan  2 
Cp
It can be seen from the above equations that to obtain a high temperature rise in a stage, which is
desirable to minimize the number of compressor stages for a given overall pressure ratio, the compressor
must combine the following characteristics:
1. High blade speed, either by RPM or blade mean radial location, such that the mid-height
compressor blade speed Um is high.
2. High axial velocity, which may be related to high mass flow or high Mach numbers
3. High fluid deflection, or camber, in the rotor blades

However, blade stresses will limit the allowable maximum blade speed, plus aerodynamic considerations
and the developed adverse pressure gradient will limit the mass flow rate and the camber (fluid
deflection) of the compressor blades.

The velocity triangles can be joined together in several different ways to help visualize the changes in
velocity entering and exiting the compressor rotor and stator. The two common methods are:
 Superimposed using the axial velocity as the datum. In doing so we’re assuming that the axial
velocity is constant throughout the compressor stage
 Or superimposed using the blade speed (Um) as the common datum. In doing so we’re assuming
that the inlet and exit blade speeds are equal, i.e. they are at the same radial location.

With this type of visualization, the trained-eye can immediately see how the compressor camber may vary
due to the various design decisions.

Page 10 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

2 1
1 1 1
2

Vabsx

Vabsx
2 2

Vabsx
Um = constant Um = constant

Figure 5: Velocity Triangle Visualization

So far we’ve shown the impact of velocity triangles on temperature rise. But what about pressure rise?
Do we just assume the compression leads to a pressure rise because we just use the word
“compression”? Of course not, that’s too simple. We do know however that temperatures and pressures
are directly related to each other. Using the adiabatic isentropic relationship PS/S together with
PS=SRTS we’ll get the expression
 1
TO 2  PO 2  
 
TO1  PO1 
So if we arrange the above equation as follows
T  U V
TO1  02  1  m absX tan  1  tan  2 
 T01  Cp

 1
 
P  
U V
 TO1  O 2   1  m absX tan  1  tan  2 
 PO1   Cp
 
Knowing the compressor temperature rise will directly lead itself to knowing the pressure rise.

o Rothalpy and Velocity Triangles

One of the more obscure assumptions employed (usually not properly defined in gas turbine engine theory) is the
conservation of rothalpy. The conservation of rothalpy sates that the relative enthalpy through a rotor
blade row remains constant even for a change of radius. Rothalpy, I, is defined as follows:
1 1
Rothalpy , I  horel  U 2  c p Torel  U 2
2 2
Since the inlet rothalpy must equal the exit rothalpy for a rotor

 1 2  1 2
 horel  U    horel  U 
 2  IN  2  EX

 1 2  1 2
 c pTorel  U    c p Torel  U 
 2  IN  2  EX

  U 2    U 2   c pTorel EX  c pTorel IN

1  1 
2  EX  2  IN

Page 11 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz


2

1 2
U EX  U IN
2

 c p Torel  EX  Torel  IN 

COMMENT: Another aspect, which is rarely defined in literature, is the impact of the change in radii on
pressure. If the mid radii of the rotor changes, then there will be an additional stagnation (total) pressure
rise in the rotor, leading to the assessment of a much more efficient compressor. To compensate for the
impact on compressor stagnation (total) pressure rise, due to a change of mid radii, rothalpy may be
further derived to involve pressure terms.


 P02     1 r2 2   r1   
2  1

   1  1      [ref 8, Agard]
 P01   2 R g T01   r2   
 


    1 r2 2   r1   
2  1
 P02,ideal
   1  1      [ref 12, Cahill]
 P01   2 R g T01   r2   
 

The above equation is equivalent to calculating the change in pressure using the free-forced vortex
equations, assuming free-vortex. The equation for the conservation of rothalpy is the same for both
compressors and turbines.

WARNING: literature research has found 2 different formulation representations for the “pressure
correction due to radii change”. Reference [8] is considered the correct derivation, whereas reference
[12] is considered to be a typographical error. NASA SP-36 presents a slightly different formulation, to be
investigated.

o Compressor Reaction
Following the velocity triangle nomenclature of this module, for the case of incompressible and reversible
flow it is permissible to define the reaction of a compressor stage as the ratio of static pressure rise in the
rotor to the static pressure rise across the stage, or the static enthalpy rise in the rotor to the stagnation
(total) enthalpy rise through the compressor stage

Rotor Pstatic PS 2  PS1

Reaction  R  
Stage Pstatic PS 4  PS1

Rotor h S hS 2  hS 1
Reaction  R  
Stage h O hO 4  hO1

If we assume that we have a “normal” compressor stage and that the axial velocity remains constant
across the compressor stage, we can write the reaction in terms of velocity-triangles as follows (derivation
not included):
1 VabsX
Reaction  tan 1  tan  2 
2 Um
1
Reaction   tan  1  tan  2 
2
“The reaction ratio is a design parameter that has an important influence on stage efficiency. Stages
having 50% reaction are widely used as the adverse pressure (retarding) gradient through the rotor and
stator rows is equally shared. This choice of reaction minimizes the tendency of the blade boundary
layers to separate from the solid surfaces, thus avoiding large stagnation (total) pressure losses.” [3]

o If reaction > 50%, then  2>1 and the velocity diagram is skewed towards the right. The static
enthalpy rise in the rotor exceeds that in the stator.

Page 12 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

“The advantage of a stage with greater than 50% reaction is the low exit loss resulting from lower
axial velocity and blade speeds. Because of the small static pressure rise in the stationary blades,
certain simplifications can be introduced such as constant section stationary blades and the
elimination of inter-stage seals. Higher actual efficiencies have been achieved in this stage type
than with the symmetrical stage - primarily because of the reduced exit loss. The disadvantages
result from a low static pressure rise in the stationary blades that necessitates a greater number
of stages to achieve a given pressure ratio and thus creates a heavy compressor. The lower axial
velocities and blade speed, necessary to keep within inlet Mach number limitations, result in large
diameters. In stationary applications where the increased weight and frontal area are not of great
importance, this type is frequently used to take advantage of the higher efficiency.” [13]

o If reaction = 50%, then  2=1 (or the rotor inlet angle equals the rotor outlet angle) and the velocity triangle
is symmetrical. The stage enthalpy rise is equally distributed between the rotor and stator rows.

“The reaction for a symmetrical stage is fifty percent (50%). The fifty percent (50%) reaction stage
is widely used, since an adverse pressure rise on either the rotor or stator blade surfaces is
minimized for a given stage pressure rise. When designing a compressor with this type of
blading, the first stage must be preceded by inlet guide vanes to provide pre-whirl, and the correct
velocity entrance angle to the first-stage rotor. With a high tangential velocity component
maintained by each succeeding stationary row, the magnitude of W1 is decreased. Thus, higher
blade speeds and axial-velocity components are possible without exceeding the limiting value of
0.70-0.75 for the inlet Mach number. Higher blade speeds result in compressors of smaller
diameter and less weight.” [13]

“Another advantage of the symmetrical stage comes from the equality of static pressure rises in
the stationary and moving blades, resulting in a maximum static pressure rise for the stage.
Therefore, a given pressure ratio can be achieved with a minimum number of stages, a factor in
the lightness of this type of compressor. The serious disadvantage of the symmetrical stage is the
high exit loss resulting from the high axial velocity component. However, the advantages are of
such importance in aircraft applications that the symmetrical compressor is normally used. In
stationary applications, the symmetrical compressor is normally not used. In stationary
applications, where weight and frontal area are of lesser importance, one of the other stage types
is used.” [13]

o If reaction < 50%, then  2<1 and the velocity diagram is skewed towards the left. The stator
enthalpy rise exceeds that in the rotor.

“While a reaction of less than 50% is possible, such a design results in high inlet Mach numbers
to the stator row, causing high losses. The maximum total divergence of the stators should be
limited to approximately 20o to avoid excessive turbulence. Combining the high inlet for the
limiting divergence angles produces a long stator, thereby producing a longer compressor.” [13]

Reaction > 50% Reaction = 50% Reaction < 50%

Figure 6: Impact of Reaction on Velocity Triangles

In axial turbines the limitation on stage work output is imposed by the rotor blade stresses. In axial
compressors, stage performance is limited by Mach number consideration. If Mach number effects could
be ignored, the permissible temperature rise (based on incompressible flow cascade limits) increases with the
amount of reaction. “With a limit of 0.7 on the allowable Mach number, the temperature rise and
efficiency for axial-flow compressors are at a maximum with a reaction of 50 %.” [3]

Page 13 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

But in practice a higher stage reaction is preferred. Increasing the stage reaction results in a decrease in
tangential velocity prior to the rotor. “A smaller whirl will create a larger relative velocity to the rotor row
making it easier for the rotor to increase the static pressure.” [4]

o  - Stage loading coefficient [1, 3]

The total enthalpy rise through a rotor blade is expressed by the Euler equation. It is often useful to
introduce dimensionless stage performance parameters. One of those parameters is the “stage loading”
coefficient (or factor), also known as the “work” or “temperature coefficient”. This value is positive for
turbines and negative for compressors.

The “stage loading” coefficient is an important design parameter for a compressor stage rotor, and is one
that strongly affects the off-design performance characteristics. Its value can be determined from the
stage velocity diagram. In compressor stages a highly and lowly loaded stage would be as follows:
o Highly loaded  > 0.5
o Lowly loaded  < 0.3
hO  U m 2VabsT 2  U m1VabsT 1

hO U m 2VabsT 2  U m1VabsT 1

 
U m2 U m2

hO c p TO
  
U m2 U m2

Assuming that Um is constant at stations 1 and 2

hO VabsT 2  VabsT 1

  
U m2 Um

Data from compressor cascade tests show that a design-point “stage loading”  should be in the range of
0.3 to 0.4 for the most efficient operation, however substantial variations of  can be expected at off-
design conditions.

“The “stage loading” coefficient alone is insufficient to specify the aerodynamic or boundary layer loading
on the compressor blade rows and for the inner & outer walls. With the same value of  and a high or
low value of axial velocity, the required flow deflection through any blade row could be low or high,
respectively. We therefore need to define a second velocity-diagram dimensionless parameter known as
the “stage flow” coefficient.” [1]

o  - Stage flow coefficient

The “stage flow” coefficient, also known as the “velocity” coefficient, expresses the ratio between the
meridional (or axial velocity) and the blade rotational velocity. A high “stage flow” coefficient indicates a
high flow through the stage relative to the blade rotational velocity. Assuming a “purely” axial compressor

Vabs  m VabsX
 
Um Um

NOTE: VabsX is directly linked the compressor stage mass flow rate and the stage area.

It can be shown that the “stage loading” is related to the “stage flow” coefficient by

  1   tan  3  tan  2  or    tan  1  tan  2 

Page 14 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
For a “normal” compressor, where 4=1, we’ll have
  1   tan  1  tan  2 
Where
 2 is the rotor relative flow exit angle
 4 is the stator absolute flow exit angle

Idea
l
(wh
e re 
1 , an
Actual d
2 rem
ain
con
s
Stage Loading  =h/U 2

tan
t)

Stall as  

Choke as 

Flow Coefficient  =CxU

Figure 7: Interaction of Flow Coefficient and Stage Loading

DISCUSSION – Compressor Surge & Stall

o COMPRESSOR SURGE
“An important part of compressor performance, as shown in a compressor map, is the limit to stable
operation called the surge line. This limit can be reached by reducing the mass flow while the rotational
speed is maintained. When a compressor goes into surge the effects are usually quite dramatic.
Generally an increase in noise level is experienced, indicative of a pulsation of the airflow and mechanical
vibration. Surge involves an axial oscillation of the total mass flow, a condition highly detrimental to
efficient compressor design.” [3]
Pressure Ratio

RPM / TT

Islands of
COMP

m TT / P T

Figure 8: Typical compressor map

“In a typical compressor it is normal that if the mass flow is reduced the pressure ratio increases. At a
certain point in an operating range the pressure rise is at its maximum. A further reduction in mass flow

Page 15 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
will lead to an abrupt and definite change in flow pattern in the compressor. This change in flow pattern is
known as surge and can cause the flow to start oscillating backwards and forwards, and after a while the
compressor will break down. A mild version of surge causes the operating point to orbit around the point
of maximum pressure rise. An audible burble is a clear indicator when the compressor is on the limit of a
more severe surge.” [4]

Inlet @ Tip

Exit @ Mid

Figure 9: NASA's Surge Calculation

“Surge has been determined empirically, from the NASA Stage 37 and the 74-A compressor data,
correlated to the values of rotor diffusion factor and relative velocity ratios at the tested surge point. The
relative velocity ratio through the rotor from the inlet tip to the exit mean value is a measure of the amount
of diffusion through the rotor and thus can be correlated to the onset of separation. The rotor relative
velocity ratio that is calculated by NASA is based on the relative velocity at the inlet tip and the exit mean-
line relative velocity. This parameter may be used as a first order guide to determine the compressor
surge point. Based on NASA’s tests and mean-line analysis of Stage-37 and Stage-74A, it was found
that most if not all of the compressors would be in surge when the rotor diffusion factor is approximately
0.60 and the relative velocity ratio is approximately 1.9.” [6]

o COMPRESSOR STALL
Stalling arises when the difference between the flow direction and the blade angle becomes too
excessive. The fact that the pressure gradient is acting against the flow direction is always a danger to
compressor flow stability. Flow reversals may occur at off-design conditions of mass flow rates and
rotational speeds that are different from the design-point condition to which the compressor was
developed for.

“If the mass flow is reduced the axial velocity will, according to the continuity equation, also decrease.
This will increase the air inlet angle and, due to the difference in air inlet angle and blade inlet angle, will
create flow incidence. With an increasing incidence angle the flow will eventually separate from the
surface at the trailing edge. The separation will grow with a further increase in incidence angle, and
finally cover the whole blade. Stall changes the performance characteristics of an axial-flow compressor.”
[4]

Two types of stall could be found in a compressor blade row:

 Blade stall
It is a two-dimensional type of stall where a significant part of the blade has a large wake leaving
the blade suction surface.

 Wall stall
It is a type of stall related to the boundary layer growth of the flow over the gas path surfaces.

Page 16 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

Figure 10: Wall and Blade stall [3]

o ROTATING STALL
Rotating stall is a particular phenomenon of axial-flow compressors. Rotating stall is a mechanism that
allows the compressor to adapt to a mass flow that is too small. Instead of trying to share the limited flow
over the whole annulus, the flow is shared unequally such that some areas have a larger mass flow than
other areas.

When a rotor blade row reaches the “stall point”, the blades do not stall all together, but instead stall in
patches and these patches travel around the compressor annulus. These stall patches propagate from
blade to blade in the direction of rotation. The stall patch causes a partial obstruction to the flow that is
deflected on both sides of it. The incidence of the flow on to the blades on the right of the “stall cell” is
reduced. The incidence of the flow on the blades to the left of the “stall cell” is increased. As these blades
(those to the left and to the right) are already close to stalling, the effect is for the “stall patch” to move to
the left, and the motion is self-sustaining. “Stall patches” travelling around the blade rows load and
unload each blade at some frequency related to the speed and number of the “stall patches”. This
frequency may be close to a natural frequency of blade vibration, which will induce undue vibratory
stresses, and possible blade failure due to frequency resonance may occur.

Figure 11: stall cell propagation [3]

The cells always rotate in the direction of the rotor. Full-span cells extend axially through the whole
compressor while part-span cells can exist in a single blade row [4].
o Part-span cells very often rotate at close to 50% of the rotor speed
o Full-span cells usually rotate more slowly in the range of 20 to 40% of the rotor speed.

NOTE: Reference [15]

states that rotating stall
“moves in the opposite
direction of the rotation
of the shaft”

Figure 12: Different types of rotating stall [4]

There is considerable freedom to choose the design value of “stage flow” coefficient, particularly in those
cases where inlet Mach numbers is not a problem to be avoided. Compressor blades with the same “de
Haller” ratio but different stagger (setting) angles will have different design point characteristics and
different off-design characteristics.
 The high-work, high-flow, low stagger compressor blades tend to have a sharp stall, with a large
fall is pressure rise and possibly a hysteresis effect that makes recovery from stall more difficult

Page 17 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

 The low-work, low-flow, high-stagger compressor blades tend to have an almost imperceptible
stall, and a stalled pressure rise higher than at design-point condition, with no drop off in pressure
differential at the stall itself.

Figure 13: low vs high stagger compressor blades and stall characteristics [1]

DISCUSSION – Compressor Stage Design

“When selecting an axial-flow compressor configuration for a specific application, there are usually
several design features that are of importance. The list of important features always includes
performance, initial cost, and maintenance cost, and often includes such features as weight, volume, and
resistance to inlet flow distortions. In all cases the compressor must be surge-free and free of significant
aerodynamically excited vibrations throughout its domain of operation.” [16]

“When a new compressor design is contemplated, it is customary to conduct preliminary design studies
aimed at providing quantitative assessments of these several features. At this point in the design process
it is not necessary to specify detailed blade shapes, but rather it is desired to quantify factors that relate to
cost, weight and volume, and to identify the efficiency potential of each design being studied. The term
efficiency potential means the efficiency that can be expected if the detailed design (i.e., the specification
of airfoil shapes, to be done later) is carried out using the best state-of-the-art design practices.” [16]

A method for assembly purposes is to have the rotor blades fixed to the rotor drum (or shaft) and the
stator blades fixed to the outer casing. The stator rows may or may not be fixed to an inner ring, such as
that found in turbine stator rows. If the stators are only affixed from the outer rings, then this will create
stator hub clearances. In doing so, each compressor stage rotor and stator can be assembled by
stacking one-by-one the individual blade rows.

Rotor
Hub clearance

Stator
Hub clearance

Figure 14: Assembly stacking of compressor blade rows

As shown below, the disks are held together with tie bolts and nuts that run through the rotor disks [17].

Page 18 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

“On some first stage axial-flow compressors, the first rotor row has an upstream row of inlet guide vanes
(IGV). These are not considered to be a part of the compressor stage and are treated differently. Their
function is quite different form the other compressor rotor or stator blade rows. The inlet guide vanes
direct flow away from the axial direction, and they act to accelerate the flow rather than diffuse it.
Functionally, inlet guide vanes are the same as turbine nozzles; they increase the kinetic energy of the
flow at the expense of pressure energy.” [3]

“Varying the IGV setting angle is the first step in aerodynamically matching a multistage compressor at
part-speed operating conditions. The variable guide vane has a large influence on the flow rate into the
first rotor. However, variable inlet guide vanes produce a loss of total pressure, which is a function of the
setting angle.” [6]
Inlet
Guide Vane

Figure 15: Compressor stage with IGV

It is desirable to keep the axial velocity approximately constant throughout a multi-stage compressor.
With the flow density increasing as it progresses through the stages, it will be necessary to reduce the
flow area and hence the airfoil height. When the compressor is running at a lower speed than the design-
point speed, the density in the rear stages will be far from the design value, resulting in incorrect axial
velocities which will cause blade stalling and compressor surge. There are various design methodologies
that can be used (individually or in combination) to overcome these problems, all of which result in
mechanical complexity:
 Use a multi-spool configuration (Rolls-Royce method)

 Use of variable stator blades (GE method)

Varying the stator reset angles in matching multistage compressors can be effective in controlling
the incidence of the downstream rotor. The variable geometry stators enable the alignment of the
flow with the downstream rotor to achieve good rotor incidence and hence good stage matching.
During the process of resetting the variable stator in order to improve the incidence on the rotor,
the stator leading edge itself becomes subjected to incidence and additional losses.

 Use of blow-off valves (Pratt & Whitney method)

“In the field of aircraft gas turbines the engine designer is more concerned with maximizing the work done
per stage while retaining an acceptable level of overall performance. Increased “stage loading” almost
inevitably leads to some aerodynamic constraint. This constraint will be increased Mach number,

Page 19 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
possibly giving rise to shock-induced boundary layer separation or increased losses arising from poor
diffusion of the flow.” [3]

There are three various compressor blades to rotor attachment arrangements as shown below
o Conventional blade row and rotor
o Blisk (Blade & disk)
The Blisk is a compressor blade row and rotor that is forged and machined together as one
component. This type of arrangement can reduce weight up to 30%.
o Bling (Blisk & ring)
The Bling is an evolution of the Blisk, where a high strength ring replaces the disk. This type of
arrangement can reduce weight up to 70%

Conventional disk & blades Blisk Bling

Figure 16: Various Compressor Blade and Rotor Attachments
“Today, due to material limitations of what compressors are manufactured from, the outer diameter
peripheral speed (U) is limited to approximately about 450 m/s. As several rotors of a compressor are on
the same shaft with the same rotational speed, it is usually preferred to work with the maximum value of U
possible. From this choice, the outer diameter of the several rotors is constant, resulting on the type of
axial channel shown below” [18]

Figure 17: Axial channel shape [Saravanamuttoo, Rogers and Cohen 2001]

“n some cases, to simplify the design of the disks on which the rotor blades are fixed, the inner diameter
is fixed at a constant value even if this causes a loss in capacity to transfer energy to the fluid due to the
lower value of U.” [18]

Page 20 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

“There are also designs that require that the rotor blade mean height be constant, and other designs
where the rotor blade mean height may be variable. Clearly, the compressor with axial channel with
constant outer diameter is the most compact, because the stages can develop the greatest pressure
ratios. Whatever the type of axial channel chosen, one must ensure that the shape of the channel does
not impose barriers to air flow under conditions of unfavorable gradients, favoring the boundary layer
separation and the consequent loss of performance.” [18]

o Estimating the number of compressor stages for a given pressure-ratio

It is possible to apply the various equations, described in this module, to determine a preliminary number
of compressor stages for a given pressure-ratio. The procedure requires the calculation of pressure and
temperature changes for a single stage, and the stage exit conditions are then used for the next
compressor stage as the inlet conditions. This calculation is repeated for each stage in turn until the final
conditions are satisfied.

However, for compressors having identical stages it is more convenient to resort to a simple compressible
flow analysis [3]. The procedure is as follows:

1. Gather the list of requirements for

a. Required overall pressure-ratio
b. The compressor inlet temperature
c. Stage RPM
d. Stage reaction (assume as the same for all the stages)
e. Um, mean blade speed (assume as the same for all the stages)
f. , Flow coefficient (assume as the same for all the stages)
g. , Stage loading factor (assume as the same for all the stages)

2. Use the following equations to obtain the values for tan1 and tan 2 as follows

   tan  1  tan  2 

1
Reaction   tan  1  tan  2 
2
 Reaction     Reaction   
2 2
 tan  1   and tan  2  
 
3. Use the following equation to obtain the individual stage stagnation (total) temperature rise as
follows

hO c p TO
  
U m2 U m2

4. Denoting the overall compressor spool inlet and outlet conditions as I and II use the following
equation to assess the number of compressor stages

 1
TOII N T P   c   1

 POII   c
 1  STAGES O   OII   N STAGES
T
 OI   1
TOI TOI  POI  TO  POI  
 

It should be reminded that the above procedure is an overly simplified method to obtain a quick estimate
for the number of compressor stages required to obtain a particular overall pressure-ratio. Assuming
constant reaction, mean blade speed, stage loading, and flow coefficient across all the stages of a
compressor is an idealized design assumption.

Page 21 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
o Work done factor [2, 3]
“Because of the adverse pressure gradient in compressors, the boundary layers along the annulus walls
thicken as the flow progresses. The main effect is to reduce the area available for flow below the
geometric area of the annulus. This will have a considerable effect on the axial velocity through the
compressor and must be allowed for in the design process. The flow is extremely complex, with
successive accelerations and decelerations combined with changes in tangential flow direction.” [2]

“The radial distribution of axial velocity is not constant across the annulus, but becomes peaky as the flow
proceeds, settling down to a fixed profile at about the fourth stage of a multi-stage compressor. Over the
central region of the blade, the axial velocity is higher than the average value. The mean blade section
(and most of the blade span) will, therefore, do less work than is estimated from the velocity triangles
based on the mean axial velocity.” [3]

Due to the influence of both end wall boundary layers and clearance vortices, the total work capacity of a
stage is decreased. This effect becomes more pronounced as the number of compressor stages is
increased. The reduction in work capacity can be accounted for empirically, as shown in the figure below,
with the “work done factor” , which is always a number less than 1.

0.95
Work done factor

0.9

0.85

0.8
0 4 8 12 16 20
number of compressor stages

Figure 18: approximated work done factor

The “work done” factor can be applied to the “stage loading” coefficient as shown below, and the
estimation process mentioned above can be adjusted to reflect the change.

     tan  1  tan  2 

DISCUSSION – Various Compressor Maps

“The figure below shows the design points for 27 compressors whose maps have been collected from
reference [9], which is based on the literature review of the literature Kurzke and Riegler. The y-axis is the
pressure ratio across the compressor; the x-axis is the mass flow through it. The axes use a log scale to
more easily see the differences between the design points for smaller values. If the series of mass flows
falls on the same PR then the machine could just be a scaled version of the same geometry.” [9]

“Not only are there different types of compressor, they also vary in the number and complexity of the rows
of blades depending on the size and thrust of the engine. The letters in the figure below indicates the
type:” [9]
o F for fan
Fans move a lot of air but produce only a small pressure rise, so are close to the x-axis.

Page 22 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

o R for radial compressor

Radial compressors move little air but can produce large pressure rises, so are close to the y-
axis.

o A for axial compressor

Axial compressors are much more flexible, although they have other disadvantages. Depending
on the number of stages, they are capable of a large range of pressure ratios over a large range
of mass flows. They are scattered across the figure, although mainly on the diagonal.

“Thus, there is a large range of values covered in both dimensions by the different compressors.” [9]

Figure 19: Compressor Design Points [9]

DISCUSSION – Compressor blade design

Three types of compressor airfoil design can be used for different ranges of Mach numbers:
 Subsonic
For flows that are entirely subsonic, it has been found that the use of aerofoil section type blading
could be used to obtain high stage efficiencies. The need to pass higher mass flow rates at high
pressures increases the Mach numbers, which become especially critical at the tip of the first row
of rotor blades.

 Transonic
For transonic compressors, where the flow over part of the blade is supersonic, it’s been found
that the most effective blade shape consists of sections of circular arcs, often referred to
“biconvex” blading.

 Supersonic
For much higher Mach numbers, it’s been found that blade profiles based on parabolas become
more effective.

The choice of blade aspect ratio also has an impact on compressor design characteristics:

Page 23 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
 Usage of high aspect ratio airfoil makes engines more compact in axial length.
 Usage of low aspect ratio airfoil in axial-flow compressors has high loading capabilities, high
efficiency, and a good performance range.

There are many types of compressor blade spawning from commercial and academic institutions, each
one having a different profile shape and associated performance characteristics.

Figure 20: Various compressor Blade Designs [7]

There are four common compressor blade designs that have been investigated, with data recorded as
performance maps, losses, or correlations. These blade types are the:
 British C-series, C4, C5, and C7 compressor blades (National Gas-Turbine Establishment)

 NACA 65-series compressor blades (National Advisory Committee for Aeronautics)

“These blade forms were used the most for commercial axial-flow compressors in gas turbines
built before 1990.” [13]

 Double Circular Arcs (DCA)

 Multiple Circular Arcs (MCA)

“These types of blades provide limited control of span wise and stream wise work distribution
during the design process” [6]

20

0
0 20 40 60 80 100

-20

Figure 21: British C4 (solid) and C7 (dashed) compressor-blade base profiles

Page 24 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

20

0
0 20 40 60 80 100

-20

Figure 22: NACA 65-series compressor-blade base profile

20

0
0 20 40 60 80 100

-20

Figure 23: Double Circular Arc

20

0
0 20 40 60 80 100

-20

Figure 24: comparison of compressor-blade base profiles of

British C4 (external solid), NACA 65-series (dashed), and DCA (middle solid)

DISCUSSION - Compressor Loss Modeling

“Transonic decelerating flow in the turbo-machine is one of the most complex flows which can be found in
fluid mechanics. This is because a large number of factors affect the flow regime and therefore influence
the magnitude of losses. The complexity is overwhelming because of the fact that the effects of several
loss producing factors are not independent, but are cumulative and interrelated. A well-known example is
the shock-boundary layer interaction. In the presence of such an interaction the order of magnitude of
profile loss is closely related to the strength of the shock wave as well as its location on the blade surface.
The losses at the end walls are due to the interactive phenomena between blade boundary layers, end-
wall boundary layers, passage shock-waves, leakage flow and secondary flow effects. It may be quite
unrealistic to single out a loss source, finds a loss coefficient for it from experimental work, and then add
these linearly for sections along a blade span where the particular loss under consideration is of
importance. The equations which govern the entropy accumulation are highly nonlinear, thus loss
assessment using superposition is not likely to be realistic.” [8]

“The design of an axial-flow compressor of high performance involves three-dimensional high-speed flow
of compressible viscous gases through successive rows of closely spaced blades.” [5]

Page 25 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
“A throughflow calculation program is probably the most important tool of the compressor aerodynamic
designer. The main objective of a throughflow calculation is to provide a span wise prediction of
thermodynamic and other flow variables such that suitable blade profiles can be selected to cope with the
variations of inlet angle, turning, Mach number, etc. Throughflow programs may be either of analysis [off-
design] or design type [design point]. In the former, which is most extensively used, blade geometry is
specified and solutions are sought for the resulting flow pattern. The design method, which is less
commonly available, requires enthalpy variations and obtains solutions for meridional velocity and hence
flow angles. To determine realistic solutions with both techniques it is necessary to provide, reasonably
well, predicted entropy gradients resulting from viscous effects. Similarly, the solution obtained from an
analysis program is very dependent upon the predicted values of blade exit flow angles. Once loss and
deviation methods have been incorporated into a program it can be used to predict overall machine
performance.” [8]

“The ability to model compressor off-design performance is necessary for system evaluation of
compressors within gas turbine engines. The off-design performance of the compressor is required to
enable engine system analyses at several operating conditions throughout the envelope. The mean line
code can be used to do an initial estimate of the variable geometry reset schedule of a multistage
compressor that will result in aerodynamically matched stages at off-design compressor speeds. This
procedure results in the generation of a compressor characteristic map that can be utilized in a
thermodynamic system model of an engine. The compressor map characteristics can often influence the
final design of the compressor and therefore compressor sizing, compressor map generation and engine
thermodynamic system modeling is an iterative procedure. An example where an engine can have
multiple design points are supersonic vehicles, where at the takeoff condition the engine operates at 100
percent design corrected speed and flow rate, while at the cruise condition the engine operates at
reduced corrected speed and flow. By knowledge of the performance at off-design conditions, the
compressor designer can optimize the configuration to provide acceptable performance in an engine
system at all operating conditions.” [6]

“Various aspects of the problem have been treated theoretically, and the results of those studies are quite
useful in design calculations. All such studies, however, have been on idealized flow, with the effects of
one or more such physical realities as compressibility, finite blade spacing, and viscosity neglected.
Consideration of viscosity effects has been particularly difficult.” [5]

“Some of the information required can be obtained only by experiment in single-stage and multi-stage
compressors. Much of the information, however, can be obtained more easily by isolating the effects of
each parameter for detailed measurement. The effects of inlet angle, blade shape, angle of attack, and
solidity on the turning angle and drag produced can be studied by tests of compressor blades in two-
dimensional cascades tunnels. Cascade tests can provide many basic data concerning the performance
of compressors under widely varying conditions of operation with relative ease, rapidly and at low cost. A
more refined procedure, however, would use cascade data not as the final answer, but as a broad base
from which to work out the three-dimensional relations.” [5]

The flow in a compressor is highly three-dimensional with many different loss (or entropy) creating flow
mechanisms that impact the overall efficiency of a compressor stage. It is these loss mechanisms that
the compressor designer will try to reduce to obtain both design-point and off-design performance
characteristics that will be beneficial to the overall SFC of the gas turbine engine. The image below
shows the various flow mechanisms that can be found in a compressor blade row, valid for both the rotor
and stator.

Page 26 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

Figure 25: Compressor flow mechanisms [4, original ref NA]

“It is hypothesized that the correlations which best represent the data at the meanline will also perform
adequately for the majority of the remainder of the flowfield. This is due to the fact that loss and deviation
at the meanline are functions of the behavior of the flowfield along the entire span of the blade.” [12]

Since 1956, compressor loss models have been developed, modified, and improved upon. In general,
the three dimensional flow field, which creates the compressor losses, have been categorized in the
following loss components:

o Profile losses, or blade surface boundary layer loss

o Secondary losses; these could be considered to be general vortex flow losses or end wall losses
o Clearance losses; tip clearance loss for rotors and hub clearance loss for stators
o Inlet shock losses
o Off-design losses, which include losses due to variation in incidence

Compressor
Loss Model

Reference Reference
Incidence Deviation

Off-Design
Incidence Loss
Total
Loss

Profile Secondary Shock

Loss Loss Loss

Subsonic Critical Transonic Casing Tip clearance Critical Transonic

Mach Mach Mach (end wall) Loss Loss Mach Mach

o “de Haller” number

As mentioned in this module, in an axial-flow compressor stage both the rotor and stator are designed to
diffuse the flow. The flow kinetic energy is converted into an increase in the stage static enthalpy and the
stage static pressure. The more the fluid is decelerated the bigger is the pressure rise, and with the rise in
pressure is a rise in the adverse pressure gradient. Blade boundary layer growth and stall are the limiting
flow mechanisms on a compressor blade design.

Page 27 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
A simple dimensionless parameter is used to choose an appropriate diffusion factor, first suggested by
Mr. de Haller a Swiss engineer. This dimensionless parameter compares the rotor relative velocity ratio
and the stator absolute velocity ratio as follows

V  V 
de Haller No.   rel 2  ,  abs 4 
 Vrel1  rotor  Vabs 3  stator

“Historically this value has been set to be more than 0.72 as an appropriate target. Lower values lead to
excessive losses. Because of its extreme simplicity the “de Haller” number is still used in preliminary
design work, but for final design calculations a criterion called the “diffusion factor” is preferred.” [2]

o Diffusion Factor (DF) and Diffusion Ratio (DR)

The air passing over a compressor blade will accelerate to a higher velocity on the suctions surface. In a
stationary row this will give rise to a drop in static pressure. On the pressure surface, the fluid will be
decelerated. The velocity distribution through the blade passage will be as shown below. The maximum
velocity on the suction surface will occur at around 10 to 15% of the blade chord from the leading edge,
and will then fall steadily until the outlet velocity is reached.
VSS-m ax
Compressor Blade surface velocity

Suction Surface

V LE

VTE
SS Pressure Surface

PS

0% 100%
Percent Chord

Figure 26: compressor cascade blade surface velocity distribution

“The losses in a compressor blade row arise primarily from the growth of boundary layers on the suction
and pressure surfaces of the blade. These surface boundary layers come together at the blade trailing
edge to form a wake, giving rise to a local drop in the stagnation (total) pressure. Relatively thick surface
boundary layers, resulting in high losses, have been found to occur in regions where rapid changes of
velocity are occurring.” [2]

The blade loading is usually assessed by the “diffusion factor”, DF. This relates the peak velocity on the
suction surface of the blade to the velocity at the trailing edge. Values of diffusion factors in excess of 0.6
are thought to indicate blade stall and values of 0.45 might be taken as a typical design choice.

V SS max  VTE
DF 
V LE

The “diffusion ratio” DR is that ratio between the maximum velocity and the outlet velocity

VSS  max
DR 
VTE

Page 28 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
NACA Diffusion Factor (DF) [2, 4]
The derivation of the NACA diffusion factor is based on the establishment of the velocity gradient on the
suction surface in terms of VLE, VTE, and VSS-max in conjunction with results from cascade tests. From the
cascade tests it was deduced that the maximum velocity on the suction surface is

VrelT s
Vrel , SS  max  Vrel , LE 
2 c
The above diffusion factor then becomes
VrelT s
Vrel , SS  max  Vrel ,TE Vrel , LE   Vrel ,TE
2 c VrelT s Vrel ,TE
DF    1 
Vrel , LE Vrel , LE 2  Vrel , LE c Vrel , LE

NASA Diffusion Factor [18]

NASA reported the use of a slightly modified diffusion factor in various axial compressor testing reports,
for example NASA Stage 74A testing [18]. As shown below, the diffusion factor has been adjusted for the
changes in leading and trailing mid radius location, the factor of ½ disappears, and the tangential velocity
components become absolute values, as compared to the NACA diffusion factor.

rTE VT ,TE  rLE VT , LE s Vrel ,TE

DF  1  
rTE  rLE Vrel , LE c Vrel , LE
Lieblein’s “equivalent” Diffusion Ratio (DReq) [3]
“Lieblein’s correlation is based on the experimental observation that a large amount of velocity diffusion
on blade surfaces tends to produce thick boundary layers and eventual flow separations. Lieblein states
the general hypothesis that in the region of minimum loss, the wake thickness and consequently the
magnitude of the loss in total pressure is proportional to the diffusion in velocity on the suction surface of
the blade in that region. The hypothesis is based on the consideration that the boundary layer on the
suction surface of conventional compressor blades contributes the largest share of the blade wake.
Therefore, the suction surface velocity distribution becomes the main factor in determining the pressure
loss.” [3]

Knowledge of the suction surface velocities before an actual design is difficult to predict. As this data
may not be available it is necessary to establish an “equivalent” diffusion ratio that approximates V SS-
max VTE.
/

V SS  max cos  1  s 

DR eq   1.12  0.61 c  cos  1 tan  1  tan  2 
2
[ref Wilson]
VTE cos  2    
V SS  max cos  2  s 
DR eq   1.12  0.61  cos 2  1 tan  1  tan  2  [ref Dixon]
VTE cos  1  c 
V SS  max cos  2  s 
DR eq   1.12  0.61 c  cos 1 tan  2  tan  1 
2
[ref Falck]
VTE cos  1    
WARNING: literature research has found 3 different representations for Lieblein’s DReq equation.
Currently there is no consensus as to which is the correct equation to use. The equations are to be used
with caution.

Koch and Smith “equivalent” Diffusion Ratio (DReq)

Koch & Smith further modified Lieblein’s “equivalent” diffusion ratio by including the impact due to
maximum thickness vs chord ratio (tmax/c) and a parameter called the Axial Velocity Density ratio (AVDR).

NOTE: Their derivation has not yet been included in this literature research

Page 29 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
o The solidity in axial flow compressor design
It should be noted that in an axial flow compressor, the solidity (c/s) represents the ratio of the
aerodynamic chord over the pitch and it is directly related to the blade number and to engine weight.

The preliminary design process of an axial flow compressor, according to Obrecht (ref [1] of ref [14]), is
represented in the figure below. This phase leads to the ultimate definition of the main geometrical and
aerodynamic characteristics, which should not be changed during the remaining parts of the process.

Figure 27: Axial Compressor preliminary design process [14]

“Since the very beginning, three main issues must be taken into account: the aerodynamic performance,
the weight and the cost of the system; the solidity, namely the ratio of the airfoil chord to the peripheral
spacing between two adjacent blades of the cascade, appears to be deeply involved in each of these
aspects, making its choice a crucial issue for the designer. The very first step, the thermodynamic cycle
definition, makes the designer translate the customer requirements into constraints: the thrust
specification determines the pressure ratio, while the fuel consumption constraint requires a definite level
of efficiency and weight; the acceleration time and operability concerns are addressed by the stall margin
specification and the weight of the compressor is mainly a consequence of the number of stages and
compressor overall size; in the end the maintenance costs and the mean time between overhaul are
strongly dependent on the number of parts and thus on the number of stages and blades. Considering the
following definition of solidity () explicitly expressed in function of the number of blades or airfoils (NOA),
it could be understood how much this parameter has a central role when linked to all the aspects
mentioned above:” [14]
c c  NOA c NOA
   ax
s 2    r cos  2    r

As a matter of fact, one can directly see the connection of to the number of blades, implying
consequences on the weight of the engine-and then on its global efficiency and on manufacturing and
maintenance costs; at the same time, the proportionality to the chord over the radius can drop a hint on
the influence on the engine dimensions, as the axial extent of a compressor depends on the spacing
between two different rows and on the axial chord of each single row, defined as follows:

cax  c  cos 

where  is the angle between the chord and the axial direction, named "stagger". The aerodynamic effects
can be summarized as follows:

Flow turning
Considering the stagger fixed, it is possible to see that the turning effectiveness of the row strongly
depends on the driving capability of the vane, determined by its blade to blade aspect ratio and then by

Page 30 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
the solidity. It can be clearly inferred that an increase in the number of blade, keeping the same chord and
radius, leads to smaller passage sections and then allows to achieve a higher stream deflection through
proper blade profiles.

Blade loading and losses

The inlet and outlet angles of the flux are assigned according to a single dimensional analysis based on
the performance required and the rotation speed limit and it could be said that the same deflection of the
flux can be obtained with different configurations of the row particularly regarding the solidity. In theory,
however, an optimum value for solidity can be found, laying between two opposite conditions:

 high solidity: this situation corresponds to a low blade load and then to a smooth pressure
profile along the airfoil, leading to a low friction coefficient and low pressure losses per area
unit. However the lapped surface is more extensive and the losses are high.
 low solidity: the blade loading is high and the pressure gradient is rough and characterized by
high diffusion and then high friction coefficient and consequent high losses.

Unfortunately the minimum loss condition usually corresponds at very low solidity value(Figure 1.5) and
really close to flow separation.

Stability
As blade loading increases, stall condition can be easily reached.

o Incidence Angles, and Reference Incidence Correlation [4]

The “incidence” angle is the difference between the inlet blade angle and the inlet flow angle. As the fluid
approaches the leading edge of a compressor airfoil it will experience an “induced” incidence The
variation of incidence has been correlated to both the airfoil flow coefficient and corrected speed.
Incidence is one of the pressure loss creating mechanisms found in axial-flow compressors.

When testing a given compressor cascade, at different inlet flow angles with fixed inlet blade angles, the
overall loss will vary with the flow incidence. There will be an increase in both positive and negative
incidence angles with a range of low values of loss. The “reference incidence” angle is that angle defined
to be within the half-range of incidence where the loss is twice the minimum loss value. Outside of this
range blade stall is said to occur.
“reference”
incidence
Loss

i / 2 i / 2

2X Minimum
loss Minimum
loss

incidence

Figure 28: definition of reference incidence angle

As mentioned above, three types of compressor cascade airfoils have been analyzed: NACA 65-series,
British C-series, and DCA. A majority of the test data was based on a compressor blade with a maximum
thickness of 10% of the chord length. “The “reference” incidences for these three types have been
correlated as follows” [4]
iref  K sh  K it  i010  n  
Where
 Ksh is a compressor blade shape correction factor
 Kit is a compressor blade thickness correction factor
 i010 is the incidence angle based on a compressor blade with 10% thickness and zero camber

Page 31 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
 n represents the incidence slope factor, in degrees
  is the compressor blade camber angle, in degrees
 α1 is the relative inlet flow angle for the rotor and the absolute inlet flow angle for stators, in degrees

 Ksh is defined to be
 0.7 for DCA
 1.0 for NACA 65-series
 1.1 for British C-series

2 3
t  t  t 
K it  0.0214  19.17 max   122.3 max   312.5 max 
 c   c   c 
 c  c  c
i010   0.0325  0.0674     0.002364  0.0913  1  1.64 x10 5  2.38 x10  4  12
 s  s  s
 c  c  c
n    0.063  0.02274     0.0035  0.0029  1   3.79 x10 5  1.11x10 5  12
 s  s  s

WARNING: the equations for i010 and n from reference [4, Falck’s thesis] do not coincide with the graphs
presented in the same reference. The equations are to be used with caution. The graphs have been
reproduced based on the above equations.

12

10 0.4
0.6
8
0.8

6 1.0
1.2
4 1.4
1.6
2
1.8
0 2.0
0 10 20 30 40 50 60 70 80
-2

Figure 29: i010 at zero camber and 10% thickness, for various solidity
0.5

0.4
0.4
0.6

0.3 0.8
1.0
0.2 1.2
1.4
0.1 1.6
1.8
0
2.0
0 10 20 30 40 50 60 70 80
-0.1

Figure 30: slope factor "n", for various solidity

Page 32 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

14

12

10

0
0 0.1 0.2 0.3 0.4 0.5

Figure 31: Correction factor for maximum thickness, Kit

o Deviation Angle, Reference Deviation Correlation and Slip Factors

“The “deviation“ angle is the difference between the exit blade angle and the exit flow angle. It arises
from a combination of two effects. Firstly the flow is decelerating on the suction surface and accelerating
on the pressure surface as it approaches the blade trailing edge. The result of this is that the streamlines
are diverging from the suction surface and converging towards the pressure surface so that the mean
(average) flow angle is less than the blade angle. This is an inviscid effect that increases in magnitude
with the rate of diffusion and acceleration towards the trailing edge. Secondly, the rapid boundary layer
growth on the suction surface towards the trailing edge “pushes” the streamline away from the surface,
contributing to the deviation.” [4]

The deviation angle is dependent on blade camber and stagger angle, making it difficult to predict.

Page 33 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
Slip Factor
The deviation angle is dependent on blade camber and stagger angle, making it difficult to predict. Using
a deviation correlation or a “slip factor” can account for the deviation angle, however both methods are
not to be used at the same time, for it’ll be considered as “double accounting”.

“The slip factor,, is the difference between the theoretical and absolute fluid tangential velocities
normalized with blade peripheral speed. The slip factor is usually an input item to a compressor model
and is typically from 0.93 to 0.96 for axial compressors. For centrifugal compressors the slip factor can be
on the order of from 0.80 to 0.90, depending on number of blades and exit blade angle.” [6]

Vabs  T  Vabs  X tan  2 


U m2

Deviation Methodology 1 – Carter’s Rule

One of the earliest deviation rules, suggested Mr. Carter, is defined as follows

m c
 x
c
s
Where
 mc is an empirical function of stagger angle
 x is an experimental factor typically equal to 2

Deviation Methodology 2
The reference deviation angle is defined as follows; similar to the above mentioned reference incidence
angle correlation:
 ref  K sh  K δt   010  m  
Where
 Ksh is a compressor blade shape correction factor
 Kt is a compressor blade thickness correction factor
 010 is the deviation angle based on a compressor blade with 10% thickness and zero camber
 m represents the deviation slope factor, in degrees
  is the compressor blade camber angle, in degrees
 α1 is the relative inlet flow angle for the rotor and the absolute inlet flow angle for stators, in degrees

 Ksh is defined to be (same as for the reference incidence angle correlation)

 0.7 for DCA
 1.0 for NACA 65-series
 1.1 for British C-series

2
t  t 
K δt  0.0142  6.172 max   36.61 max 
 c   c 
 c  c  c
 010    0.0443  0.1057    0.0209  0.0186  1    0.0004  0.00076  12
 s  s  s
mBlade Type
m b
c
 
s
b  0.9655  2.538 x10 3  1  4.221x10 5  12  1.3 x10 6  13

Where mBlade Type is for the different compressor blade types mentioned

Page 34 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

m NACA 65series  0.17  3.33 x10 4 1.0  0.1 1  1

m DCA  m British C series  0.249  7.4 x10 4  1  1.32 x10 5  12  3.16 x10 7  13

WARNING: the equations for 010 and m from reference [4, Falck’s thesis] do not coincide with the graphs
presented in the same reference. The equations are to be used with caution. The graphs have been
reproduced based on the above equations.

5
0.4
0.6
4 0.8
1.0
3 1.2
1.4
2 1.6
1.8
2.0
1

0
0 10 20 30 40 50 60 70 80

Figure 32: 010 at zero camber and 10% thickness, for various solidity

0.8 0.9

0.7 0.8
0.4 0.7 0.4
0.6
0.6 0.6
0.8 0.6 0.8
0.5
1.0 0.5 1.0
0.4 1.2 1.2
1.4 0.4 1.4
0.3
1.6 0.3 1.6
1.8 1.8
0.2
2.0 0.2 2.0
0.1 0.1

0 0
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80

Figure 33: slope factor "m", for various solidity

NACA 65-seried (left), DCA & British C-series (right)

12

10

0
0 0.1 0.2 0.3 0.4 0.5

Figure 34: Correction factor for maximum thickness, Kt

Page 35 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
o Total loss
“Transonic decelerating flow in the turbo-machine is one of the most complex flows that can be found in
fluid mechanics. This is because a large number of factors affect the flow regime and therefore influence
the magnitude of losses. The complexity is overwhelming because of the fact the effects of several loss
producing factors are not independent, but are cumulative and interrelated. A well-known example is the
shock-boundary layer interaction. In the presence of such an interaction the order of magnitude of profile
loss is closely related to the strength of the shock waves as well as its location on the blade surface. The
losses at the end walls are due to the interactive phenomena between blade boundary layers, end wall
boundary layers, passage shock waves, leakage flow and secondary flow effects. It may be quite
unrealistic to single out a loss source, find a loss coefficient for it from experimental work, and then add
these linearly for sections along the blade span where the particular loss under consideration is of
importance. The equations, which govern entropy accumulation, are highly non-linear, thus loss
assessment using superposition is not likely to be realistic.” [8]

The total loss for a compressor cascade has been defined to be a combination of the following blade loss
mechanisms: profile, secondary (end wall), clearance, shock, and incidence. The combinations of these
losses are dependent on the various authors’ research scope and design limitations. The table below
gives a chronological view of the various forms encountered in literature.

Table 2: Chronological view of axial compressor mean-line loss models [1]

Author Total Loss Formula
Lieblein
Profile
(1953) [2]
Profile & Trailing Edge
+ End wall & TipClr
Koch & Smith
+ Shock
(1976) [3]
+ Part span shrouds (interpreted formula from reference)

Profile
+ Secondary
Barbosa + Shock
(1987) [4]
Corrected for
 Re
Profile
+ Endwall & TipClr
Wright & Miller + Shock
(1991) [5]
Corrected for (interpreted formula from reference)
 Re
Bloch, Copenhaver, O’Brien Profile
(1997) [6] + Shock
Profile
Cahill
+ Secondary
(1997) [7]
+ Shock
Profile
+ Shock
Lynette Smith
(1999) [8]
Corrected for:
 incidence
Profile
Ramakdawala
+ Endwall & TipClr
(2001) [9]
+ Shock
Profile
+ Shock
Boyer + Secondary (tip & hub)
(2001) [10]
Corrected for:
 incidence
Profile
+ Secondary
van Antwerpen + Annulus , for i>imin
(2007) [11]
Corrected for:
 Incidence

Page 36 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
 Reynolds,
 Ma , for i<imin

Falck Profile
(2008) [12] + End wall
Veres Profile
(2009) [13] + Shock
Basic loss
+ shock
Benini
+ secondary
(2010) [14]
+ tip clearance
+ off design correction

PO P  PO 2
TOTAL   O1
PO1  PS1 PO1  PS1

Once the total loss for the rotor and stator is calculated, then one is able to calculate the compressor
stage efficiency as follows
 1
 
TO1  PO 4  
    1 Stage total-to-total adiabatic efficiency
TO 4  TO1  PO1  
 

Table 3: axial compressor efficiency equations

Configuration Formulas
This is a generic axial compressor configuration that consists of a rotor followed by a
stator. The total-to-total efficiency equation will be based on the rotor inlet
temperature, and the pressure ratio will be based on the rotor inlet and stator exit.
Since we are using the adiabatic assumption, there is no temperature gain due to the
stator. Hence the change in temperature will be mainly due to the rotor.
 1
 
 PO , stator exit  
TO , rotor inlet
   1
TO , rotor exit  TO , rotor inlet  PO , rotor inlet  
  

The generic axial compressor configuration is modified by adding a structural strut

after the stator. that consists of a rotor followed by a stator. The total-to-total
efficiency equation will be similar to the above, where the strut exit conditions will
replace the stator exit conditions.
 1
 
 PO , strut exit  
TO , rotor inlet
   1
TO , rotor exit  TO , rotor inlet  PO , rotor inlet  
  

In this axial compressor configuration, we have added an IGV. The IGV may be a
turbine vane shape or a compressor stator shape. Here, we may assume that the
stage pressure ratio will be based on the IGV inlet condition.
 1
 
 PO ,strut exit  
TO , rotor inlet
   1
TO , rotor exit  TO , rotor inlet  PO , IGV inlet  
  

Page 37 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
“Note that 2D linear cascade data and theory must be modified in order to prepare them for use in
modeling an actual 3D compressor flow field. Some adjustments in converting from a 2D to a 3D flow-field
include:” [12]
1) New definition of loss to account for change in radius
2) Conversion from 2D to 3D loss
3) Converting 2D Axial Velocity Density Ratio (AVDR) to analogous 3D Meridional Velocity Density
Ratio (MVDR)
4) Introduction of new losses such as those due to mid-span dampers, casing and hub boundary
layers, vortices between the casing and the rotor tip, and additional secondary losses which are
not present in strictly 2-D flow-fields.

o Profile loss, based on Wright & Miller (1991) [5]

The compressor blade profile loss correlation is based on a modified version of the two-dimensional low-
speed equation developed by Lieblein as follows.

1 V LE2
p cos  2  functionM LE , DR eq 
2 VTE2

2
1 V 
 p  LE  cos  2  functionM LE , DR eq 
2  VTE 

functionM LE , DR eq 
p  2 2
 V LE 
  cos  2
 VTE 

NOTE 1: From the profile loss correlation for compressor cascade blades, we see that the “de Haller”
number has a role in the derivation.
V  V  V 
de Haller No.   rel 2  ,  abs 4    TE 
 Vrel1  rotor  Vabs 3  stator  V LE 

f M LE , DR eq de Haller No 
2

p  2
cos  2
where
 p is the Profile Loss Coefficient
 VLE inlet velocity to the compressor blade
 VTE exit velocity from the compressor blade
 2 exit flow angle from the compressor blade
 MLE inlet Mach number
 DReq equivalent diffusion factor

Page 38 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

0.07

0.06

Profile Loss Parameter

0.05

0.04

0.03

0.02
MLE = 1.0
0.01 MLE = 0.7
MLE = 0.3
0
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
DReq

Figure 35: Profile Loss Parameter wrt M LE

NOTE 2: The above profile loss correlation is based on compressors that have relative inlet Mach
numbers from 0.3 to 1.0.

o Secondary, or End Wall Loss, based on Wright & Miller (1991) [5]
Based on numerous compressor data, the end wall loss has been correlated to be a function of the
following parameters
 End wall clearance ratio /c = clearance/chord
 Blade aspect ratio h/c = span/chord
 Mean line loading
h V LE2  
 ew 2
 function , DF 
c VTE c 
2
h  V LE   
 ew    function , DF 
c  VTE  c 

 
function , DF 
 ew  c 
2
 h  V LE 
  
 c  VTE 
0.14

0.12
End wall Loss Parameter

0.1

0.08

0.06
e/c = 0.10
0.04 e/c = 0.07
e/c = 0.04
0.02 e/c = 0.02
e/c = 0.00
0
0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65
Diffusion Factor

Figure 36: End wall Loss Parameter with respect to end wall Clearances
“In the last stages of the compressor, in order to reach the desired total pressure ratio and to properly
follow the flow, the annulus height is really small, so the hub to tip ratio increase, it means that the blade
Page 39 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
become shorter, thus the percentage of the tip clearance on the total height of the blade increase; this
affected stall margin, pressure rise and efficiency.” [15]

“In general, for one percent increase in clearance-to-span ratio, there is a one to two percent decrease in
the efficiency, two to four percent decrease in the pressure rise, and three to six percent decrease in stall
margin (Chen, 1991).” [15]

Figure 37: Effects of the increased clearance on the performance (Wisler, 1985) [15]

o Shock Losses
Corrections for transonic relative inlet Mach numbers and corrections for inlet shock losses have not
been added in this section. Refer to the Appendix and to the problem solving for transonic axial
compressors, Modules 04 A01 and A02.

o Incidence Losses
“The efficiency of the rotor as a function of incidence is shown below. This equation was determined
empirically from the NASA Stage 37 and 74-A compressors by Veres.” [6]

  0.0006  iR3  0.0185  iR3  0.1699  iR3  0.5187

o Off-Design Compressor Map

In MDIDS-GT, choke is determined numerically when the compressor stage reaches a Ps based reaction
value of 100%. NASA code COMDES calculates the choke value as the limit of flow where a mean-line
code fails to converge on the rotor exit velocity triangles [6].

NOTE: the MDIDS-GT reaction choke limit is considered incorrect. It will be corrected post the transonic
problem solving research as mentioned above.

“The design point of a compressor lies on the 100% design speed line and indicates maximum
performance with the given surge (stability) margin. Off-design conditions are defined as being at any
point other than the design point on the compressor map, regardless of operational speed. Although the
compressor often spends much of its lifetime at off-design conditions, the design point is still regarded as
the most critical point on the map.” [12]

“The surge (stability) line of a compressor is created by maintaining a constant compressor rotational
speed and closing the exit throttle. This moves the operational condition of the compressor toward a
decreased corrected mass flow rate and increased total pressure ratio along that same speed line. The
performance of the machine improves until it reaches a point of unstable operation and an undesirable
breakdown in machine performance. The surge (stability) line is the locus of these stability points for each
of the constant corrected rotor speeds.” [12]

“Since the design point is defined as being close to the maximum efficiency, minimum-loss for the 100%
speed typically occurs at or near the design point. Many investigators have also applied design loss
theory to minimum-loss points that occur at off-design compressor speeds, with moderate success. In
order to obtain an accurate design point stage characteristic, one must not only calculate an accurate
design loss, but accurate design incidence and deviation angles as well.” [12]

Page 40 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

DISCUSSION – Surge Prediction

A relationship is used, created by Koch, to calculate how close each compressor stage is to surge. By
calculating the static pressure rise coefficient, Cp, based on pitchline dynamic head, and comparing it to
the maximum static pressure rise, will give a good indication of how close the stage is towards stall. The
maximum static pressure rise, Cp,max, is based on a diffuser correlation modified by the influence of
Reynolds number, tip clearance, and the axial spacing between the rotor and stator of the compressor.

For the influences of Reynolds number, tip clearance and axial spacing between blade rows a set of
graphs are used. These graphs were created by Koch on numerous sets of tests in which different
parameters were varied. The tests were performed with a General Electric low-speed multistage
compressor and the blade geometry and clearance was systematically varied. These tests, plus some
additional low speed experimental configurations, also provided data for the correction for Reynolds
number and showed the effect of extreme values of stagger angle, flow coefficient and reaction [4].

 Static pressure rise coefficient Cp

“Since there is a change in radius throughout the rotor, the difference in peripheral speed U will
decrease the static pressure rise. This decrease in static pressure rise will not put any more
stress on the boundary layers because there will not be any further diffusion caused by the
change in radius.” [4]

The equation for the static pressure rise coefficient is calculated as a function of the static
pressure ratio across the stage (from the inlet of the rotor to the exit of the stator) and as a function of
peripheral speeds and velocities across the rotor. For the value of , that of the rotor may be
taken to be equivalent to that across the stage.

 1
 
  P   U 2  U m2 1 
c p ,air Ts1  s4
  1   m 2 
 Ps1    2 
 
Cp 
Vrel2 1  Vabs
2
2

 
 2 

 Maximum static pressure rise coefficient Cp,max

The max static pressure rise coefficient is based on low-speed compressor tests. Caution should
be employed if trying to apply Cp,max for high speed compressor designs.

 Cp   Cp   Cp 
C p ,max  Fef C pD      
C  C  C 
 pD  Re  pD   pD  Z

Except for the Fef parameter, all the values for the (Cp/CpD)x functions are obtained from graphs.
Since Cp,max is impacted by both the rotor and the stator, a weighted average is used to assess
the combined influence of the two airfoil types. Here X represents the parameter of interest.
Vrel2 , IN X rotor  Vabs
2
, IN X stator
X weighted ,ave 
Vrel2 , IN  Vabs
2
, IN

 Calculation of CpD
The value of CpD, which is obtained from the first graph below, is considered to be a
function of the diffusion length, L, versus exit passage width, g2. L/g2 is a function of the
solidity , the exit blade angle  EX, and the total turning or camber  of the airfoil.

Page 41 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

c
L  s
 
g 2 cos EX  cos / 2  cos EX  cos / 2 

L/g2 is calculated for both the rotor and stator, and then the weighted average is used to
find the combined effect.

 Calculation of (Cp/CpD)Z
This parameter is based on the normalized axial spacing Z/s, where Z = axial spacing
between airfoil rows, and s = airfoil pitch. A rule of thumb for the axial spacing in
compressors is Z = 2*chord.

Two different weighted averages may be calculated as follows:

METHOD 1: [Falck]
 the axial spacing Z of the rotor and stator is weighted
 the pitch s of the rotor and stator is weighted
 the an overall Z/s is calculated

METHOD 2: [MDIDS-GT]
 The Z/s of the rotor and stator are calculated individually
 then the weighted average is found.

 Calculation of (Cp/CpD)Re
This parameter is based on the weighted average of the rotor and stator Reynolds
number. The Reynolds number is defined as follows, where  is the viscosity of the fluid

Vrel ,IN  TrueChord

Re 

Re is calculated for both the rotor and stator, and then the weighted average is used to
find the combined effect.

 Calculation of (Cp/CpD)/g
 is the tip clearance between the casing wall and the rotor and stator.

g is a function of the mean radius rm, the inlet and exit blade angles  of the airfoil, and
the airfoil count Z.

  rm cos IN   cos EX 
g
Z
 and g are calculated separately for both the rotor and stator, and then the weighted
averages of both are used to find the combined effect.

 Effective dynamic pressure factor Fef

The effective dynamic pressure factor Fef has been described by Koch as follows, “A
parameter giving a quantitative measure of this flow coefficient/stagger angle effect,
termed the effective dynamic pressure factor Fef, was defined as the effective dynamic
head divided by the pitchline free stream dynamic head. As indicated by the equation, the
effective dynamic head was represented by a weighted average of the free stream
dynamic head, the minimum possible dynamic head and the dynamic head at zero axial
velocity. By trial and error, the best fit of the data was found to occur if the minimum
dynamic head was weighted 2.5 times as heavily as the free stream head, and the head
at zero axial velocity was weighted one half as heavily as the free stream head. The
minimum dynamic head was set equal to the free stream value for vector triangles where
(+) was greater than 90 deg., because in such a case the minimum possible dynamic
head could only occur at axial velocities higher than the free stream value. Also, for
vector triangles in which the upstream blade row turned the flow past the axial direction,

Page 42 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
the minimum dynamic head was not allowed to become less than the zero through-flow
value, because in this case the mathematical minimum dynamic head could only occur at
negative axial velocities.” [4]
2
Vabs , ef
2
Vabs  2.5  Vabs
2
. min  0.5  U EX
2
Fef  
2
Vabs 4  Vabs2

2
Vabs
If (+)90 and 0 then 2
,min
 sin 2    
Vabs
2
Vabs
If (+)>90 then 2
,min
 1.0
Vabs
2
Vabs 2
U EX
If <0 then 2
,min
 2
Vabs Vabs

Further into the compressor, the surge margin will increase. The limit to surge is strongly
dependent on the diffusion factor. By decreasing the diffusion factor will result in a lower static
pressure rise and hence increase the surge margin. The diffusion factor is itself; it is dependent
on several other factors. The pitch-to-chord ratio, s/c, has the most profound effect on diffusion
factor. If the pitch chord ratio is decreased, then the diffusion factor will also decrease, resulting in
a lower static pressure rise

Figure 38: Graphs used to calculate Cp,max [4]

DISCUSSION – Compressor mean-line or reduced order throughflow methodology

Basic Design-Point (DP) input parameters to an axial compressor stage are

o Stage inlet total pressure
Page 43 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
o Stage inlet total temperature
o Stage inlet mass flow
o Shaft rotational speed
o Stage work or pressure ratio
o Stage reaction
o Inlet and exit blade aerodynamic blockages
o Slip factor, for those methodologies not using a deviation correlation

The fluid properties are determined at each blade leading and trailing edge location, including all of the
components of the inlet and exit velocity diagram. These parameters are initially calculated at the
geometric mean line radius.

Page 44 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

DISCUSSION – Compressor Design Rules of Thumb (ROT)

Throughout this module various axial-flow compressor stage and airfoil design criteria have been
described. Below is a summary of the main rules-of-thumb that can be employed in the preliminary
design of an axial-flow compressor.

Table 4- Some axial-flow compressor design ROT

Parameter Rule of Thumb
Compressor Stage Reaction 50%
Compressor Stage Loading,  0.3    0.4
Work done factor,    0.86
(stage count dependent)
Hub-to-Tip radius ratio1, Hr/Tr 0.45  Hr/Tr  0.75
Pitch to chord ratio1, s/c 0.50  s/c  0.95
Diffusion Factor1, DF DF  0.55
Allowable Mach number 0.7
De Haller Number.  0.72
Rotor tip o  0.9 if subsonic
Relative Mach number o  1.2 if transonic
Stator hub  0.9
Absolute Mach number
Aerodynamic blockage Cd@inlet  0.98
(if used) Cd@exit  0.97
(requires calibration, default Cd=1.0)
Slip factor 0.93 to 0.96
(if used) (requires calibration, default = 1.0)
1: Saravanamuttoo et al.( 2001) and Walsh and Fletcher (1998)

Table 5 - Axial Flow Compressor Characteristics [13]

Type of Inlet Relative Velocity Pressure Ratio Efficiency
Type of Flow
Application Mach Number per Stage per Stage
Industrial Subsonic 0.4 - 0.8 1.05 - 1.2 88% - 92%
Aerospace Transonic 0.7 - 1.1 1.15 - 1.6 80% - 85%
Research Supersonic 1.05 - 2.5 1.80 - 2.2 75% - 85%

DISCUSSION – Compressor Stall Flutter

Add Boyce’s stall flutter

DISCUSSION – Benchmarking Compressor Software

NASA: CDP, Compressor Design Program [11]

This code was used to design the NASA compressor stages 35, 36, 37, 38

NASA: COMDES
This mean-line compressor code provides the capability of sizing single and multistage compressors
quickly during the conceptual design process. Based on the compressible fluid flow equations and the
Euler equation, the code can estimate rotor inlet and exit blade angles when run in the design mode. The
design point rotor efficiency and stator losses are inputs to the code, and are modeled at off design.
When run in the off-design analysis mode, it can be used to generate performance maps based on simple
models for losses due to rotor incidence and inlet guide vane reset angle [6].

Page 45 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

The design point rotor efficiency is modeled at off-design by an empirical correlation with rotor incidence.
The relative velocity ratio and the diffusion factor (Ref. 2), through the rotor are correlated to the onset of
stall [6].

Page 46 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

APPENDIX – LOSS MODELLING

Authors & Equations
o Solidity
The early works published from 1945 to 1955 consisted mainly in theoretical studies analysis and two-
dimensional low-speed experimental campaigns. Howell, Carter, Zweifel, Lieblein laid strong foundations
for all the future researches. In the following years till the late seventies, many experiments were
conducted to investigate the phenomenon though three-dimensional analysis. However, despite the
extended amount of time and resources spent in these projects, no general rule has been published for
choosing optimum solidity neither in terms of loading nor efficiency. As a matter of fact the solidity choice
appears to be based on the experience of the designer and on the common practice of each company, as
a result of a trade-off between weight, blade loading and efficiency requirements. In the past decades, to
fulfil the requirements in terms of stages loading both the solidity and aspect ratio, namely the blade
height over the chord, have been increased, These design trends are mostly due to a better behaviour of
long-chord airfoil but, as consequence of the effects mentioned before, it is clear that an inversion of
these tendencies can really improve the efficiency of the engines allowing to accomplish the new
constraints in terms of fuel consumption and pollutant emissions, and would also likely lead to cut the
manufacturing costs [14].

Figure 39: solidity vs total loss & time [14]

Page 47 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz

Figure 40: Stator (left) and Rotor (right) efficiency of DDOR (Direct Drive Open Rotor) booster
stage for different number of blades [14]
Add figure observations

Solidity analysis by [14]

c c  NOA c NOA
Solidity Equation    ax
s 2    r cos  2    r

Table 6: Optimum Solidity Equations

Author Year Formula
Howell
Zweifel
Lieblein
[14] Hearsey
[14] McKenzie
[14] Liu

o Shock Losses
[Dec 2013: LinkedIn conversation with Magnus Genrup, Associate Professor at Lund University]
Q. which axial compressor shock loss equation is best, and easy to implement, for a mean line analysis?
A. - Miller, Lewis and Hartmann
- Koch and Smith
- Schobeiri

Q. In your opinion which best models closely test data?

A. None is really suited for 1D calculation and is typically used for throughflow (on a streamline basis).
The loss scales with (Ma_x^2-1)^3 and is not significant below about 1.3 (x denotes upstream Mach
number and is typically the average between the inlet and after the P-M expansion on the suction side).

Page 48 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
Table 7: Shock Loss
Author (year) Formula
Miller, Lewis,
& Hartmann ???
(1961)
Koch & Smith M S2  ???
(1976)
 1
   1M S2   1    1   1  1
 2   
 2    1M S   2M S2    1 
 Shock  
   1 2   1
1  2 M 1  1
 
Schobeiri
???

  1  1M W2 1m  23   M W2 1m    9

Rizwan R. 1
Ramakdawala y*    1M W2 1m  3    
(2001) 4M W2 1m
1 1
NOTE: y* not Pte 

  1   1    1M W2 1m y *   1
cross referenced   
Pti  2M W2 1m y *   1   2    1M W 1m y * 
2

P
1  te
Pti
Shock  
   1 2   1
1  1  M W 1m 
 2 

References [underlined]

1. Kidikian, J. and M. Reggio. Off-Design Prediction of Transonic Axial Compressors: Part 1—Mean-
Line Code and Tuning Factors. in ASME Turbo Expo 2018: Turbomachinery Technical
Conference and Exposition. 2018. American Society of Mechanical Engineers.
2. Bullock, R.O. and I.A. Johnsen, Aerodynamic Design of Axial Flow Compressors. NASA report,
1965(SP-36).
3. Koch, C. and L. Smith, Loss sources and magnitudes in axial-flow compressors. Journal of
Engineering for Power, 1976. 98(3): p. 411-424.
4. Barbosa, J., A Streamline Curvature Computer Programme for Performance Prediction of Axial
Compressors. 1987, Ph. D. Thesis, Cranfield Institute of Technology, England.
5. Wright, P. and D. Miller. An improved compressor performance prediction model. in IMechE
Conference Proceedings CP 1991-3. 1991. Rolls-Royce plc.
6. Bloch, G.S., W.W. Copenhaver, and W.F. O’Brien, A shock loss model for supersonic
compressor cascades. Journal of turbomachinery, 1999. 121(1): p. 28-35.
7. Cahill, J.E., Identification and Evaluation of Loss and Deviation Models for use in Transonic
Compressor Stage Performance Prediction. 1997, Virginia Polytechnic Institute and State
University.
8. Smith, S.L., One-Dimensional Mean Line Code Technique to Calculate Stage-by-Stage
Compressor Characteristics. 1999, University of Tennessee - Knoxville.
9. Ramakdawala, R.R., Preliminary design code for an axial stage compressor. 2001, NAVAL
POSTGRADUATE SCHOOL MONTEREY CA.
10. Boyer, K.M., An Improved Streamline Curvature Approach for Off-Design Analysis of Transonic
Compression Systems. 2001, Virginia Polytechnic Institute and State University.
11. Van Antwerpen, W., Multi-quadrant performance simulation for subsonic axial flow
compressors/Werner van Antwerpen. 2007, North-West University.

Page 49 of 50
RDDM Turbomachinary Lecture Series – Série de Conférence du Reacture de Gaz
12. Falck, N., Axial Flow Compressor Mean Line Design. Master's thesis, Lund University, Lund,
Sweden, 2008.
13. Veres, J.P., Axial and centrifugal compressor mean line flow analysis method. NASA Technical
Memorandum, 2009(TM-2009-215585).
14. Benini, E., Advances in Aerodynamic Design of Gas Turbines Compressors. 2010: INTECH Open
Access Publisher.

END OF / FIN DE MODULE

Page 50 of 50