4 Efficiency Definitions

for Compressors

4.1 Overview

4.1.1 Introduction
Two examples should suffice to whet the reader’s appetite for the importance of clarity in
efficiency definition in radial compressors. First, the commonly used isentropic effi-
ciency – which is sometimes called adiabatic efficiency – compares the actual work
transfer to that which would take place in an ideal isentropic adiabatic process with no
losses and no heat transfer. Unfortunately, the isentropic efficiency does not actually
represent the real quality of a machine at all well. For example, consider a two-stage
turbocharger with a pressure ratio of 2 in both stages. If each stage achieves the same
isentropic efficiency of 80%, then on combining them to a two-stage compressor with a
pressure ratio of 4, the isentropic efficiency is then lower than that of the individual stages
(78.1%). How strange and disorientating. The so-called polytropic efficiency overcomes
this problem and, in this case, if both stages have 80% polytropic efficiency, the two-stage
compressor would have the same polytropic efficiency as its individual stages.
Second, a radial compressor impeller may have, at the same time, a total–total
polytropic impeller efficiency of over 90%, a static–static isentropic efficiency of well
below 60% and an impeller wheel efficiency of 30%. You can guess which definition
a sales engineer would prefer to use to sell his product. It turns out that one can misuse
efficiencies just as one can misuse statistics.
In this chapter, the background to the systematic definition of the isentropic,
polytropic and isothermal efficiencies, with and without kinetic energy, is considered.
This chapter concerns only the different definitions of the efficiencies, and Chapter 10
gives more detail on the source of losses and expected efficiency levels.

4.1.2 Learning Objectives

Be aware of the many different forms of efficiency definition and the need to
decide which one is appropriate for use in different situations.

Understand the different reversible reference processes used for definition of
isentropic, polytropic and isothermal efficiencies.

Understand the implications of the kinetic energy of the fluid for total–total,
total–static and static–static efficiencies.

Downloaded106
from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.2 Compressor Efficiency 107

Understand why the aerodynamic work is integrated along the path of static states
and not along the path of total states in the definition of total–total polytropic
efficiency and why this is acceptable for the total–total isentropic efficiency.

Be able to distinguish between the different terms; reversible work, aerodynamic
work and the polytropic head.

4.2 Compressor Efficiency

4.2.1 Issues with the Definition of Efficiency

The efficiency is the key parameter on which the assessment of the process in a
compressor is based. It is used to compare the quality of different compressors, or to
identify the variation in performance of a particular machine with a change in
operating conditions or the deterioration over time. The efficiency is important as it
is directly related to the power consumption, the environmental impact and the overall
operating costs of a compressor.
Compressor efficiency is best defined as the ratio of the minimum effort that would
be required to compress the flow in an ideal reversible compression process between
two pressure levels relative to the effort actually expended to achieve this pressure
rise. It is expressed in such a way that the efficiency has the value of unity (or 100%)
for a perfect machine. This statement looks very straightforward but is actually highly
complex. It is necessary to consider slowly and methodically all the issues that may
cause difficulties and require clarification, and this discussion is started in this section
and continued throughout the chapter.
The ideal thermodynamic process in the reversible machine which is being used as
a basis to define the minimum work has to be defined and, in practice, this reference
process could be isentropic, polytropic or isothermal. The isentropic adiabatic process
might naturally be considered to be the most useful definition of the reference process
for this perfect machine as it has no losses and no heat transfer. This leads to the
definition of the isentropic efficiency. But as shown in this chapter, it is sometimes
more appropriate to consider a reversible machine with a reference process that is the
same as the polytropic process which actually occurs in the real machine. This leads to
the definition of the polytropic efficiency, which in compressors is larger than the
isentropic efficiency. In intercooled compressors, a reversible isothermal process may
be more useful as reference, and this leads to the definition of the isothermal
efficiency. Because of the divergence of the constant pressure lines in an h-s diagram
towards high temperature, an isothermal process defines the minimum possible work
between two pressure levels. The reversible isothermal process leads to the lowest
possible work for a perfect machine, and so the isothermal efficiency is lower than the
isentropic and polytropic efficiencies
The effort in the compression process is best considered to be the sum of the useful
aerodynamic work together with any change in kinetic energy of the fluid. The
aerodynamic work is the integral of vdp along the change of static conditions in the

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 108 Efficiency Definitions for Compressors

reversible process from the inlet to outlet conditions. In pumps and water turbines, the
change in potential energy also needs to be included. Chapter 2 shows that the
aerodynamic work in a thermodynamic process is defined in terms of an integration
along the process path of vdp, with static conditions, and not as an integration of vtdpt,
along a process path of total conditions. This is an important distinction to be
examined in more detail in this chapter.
Connected with the difference between total and static states is the fact that clarity
is needed about the value of any kinetic energy present at the inlet and outlet planes. Is
the kinetic energy in these planes negligible, and if not, is it useful or not? In some
cases, the changes in kinetic energy are sufficiently small that they can be neglected, in
some they must be included and in others including the outlet kinetic energy can lead
to a distortion of the quality of the compression process. The locations of the planes
used for the definition of the inlet and outlet of the process also need to be clearly
defined. These planes may be the inlet and outlet of a single rotor, a single component,
a single stage, multiple stages, or the whole machine from flange to flange. In the
special case of a stationary component with no work input, such as a diffuser, there is
no shaft work so the definition of efficiency for such a component needs to be able to
include this possibility. Another issue, not discussed in detail here, is how to deter-
mine the appropriate 1D mean value of the thermodynamic properties at the inlet and
outlet planes. For this, the reader is referred to Cumpsty and Horlock (2006), who
recommend area averaging of the pressures and mass-averaging of the velocity
components and enthalpy for the applications considered here.

4.2.2 The Ideal Reversible Reference Process

The choice of the reversible process used as an ideal reference is a matter of
convention. It depends on the application of the machine and results in different
efficiency definitions. Common practice makes use of isentropic processes, but poly-
tropic or isothermal processes can also be used to define efficiencies, as shown
schematically in Figure 4.1. The isentropic efficiency is the easiest to understand
and considers the relevant reference process to be an ideal reversible process at
constant entropy from state 1 to 2s between the inlet and outlet pressure levels. In
this reference process, there is no heat transfer, so the isentropic efficiency is also
known as the adiabatic efficiency.
The polytropic efficiency considers the ideal reference process to be a reversible
process whose static states coincide with the actual change of state of the gas in the
compressor from state 1 to 2. This ideal reversible process has no losses so that
heat has to be added within the process to bring the change of state along this
reversible path into line with the change in entropy due to the dissipation in the
real process. In this way, the reversible reference polytropic path is actually
not adiabatic.
The isothermal efficiency considers the ideal reference process to be a reversible
process between the inlet and outlet pressure levels but which have the same

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.2 Compressor Efficiency 109

h p2 p1
2
h2
2s Polytropic
h2s Isentropic
2i Isothermal
h1
1
s1 s2 s
Figure 4.1 Reference reversible processes process for different efficiency definitions.

temperature of the gas as at the inlet, from state 1 to 2i. This is important because
the lowest possible work between the two pressure levels would be in an isother-
mal process. The ideal isothermal process is also not adiabatic. Heat has to be
removed along the reference isothermal compression path to account for the
reduction in entropy.

4.2.3 The Minimum Work of Compression

The minimum shaft work required to compress the flow in a compression process
along the flow path is given by the reversible steady flow shaft work, which is the
sum of the aerodynamic work and the change in kinetic energy as defined in
(2.43). The reversible work along the flow path from state 1 to state 2rev is then
defined as
ð 2rev ð 2rev


ws12, rev ¼ y12, rev þ ½ c2rev  c1 ¼
2 2
vdp þ cdc: (4.1)
1 1

The first term in these expressions is the aerodynamic work, and the second term is
the change in kinetic energy. It is important to note that in this definition, the
aerodynamic work is defined by the integration of vdp along the path of static
states and the kinetic energy terms at inlet and outlet are added to this to give the
total reversible work. The bookkeeping of the change in kinetic energy is done
separately from the bookkeeping of the aerodynamic work. The internal work does
not include losses external to the flow path, such as the disc friction losses or
leakage losses, and these need to be considered separately, as described in Section
4.6. Furthermore, the actual effort expended is taken to be work done by the shaft,

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 110 Efficiency Definitions for Compressors

ws12 on the fluid. Therefore, the general form of efficiency for compressors can be
written as
ws12, rev
ηc ¼ : (4.2)
ws12
On this basis, several different definitions of compressor efficiency can all be
expressed as
Ð 2rev
vdp þ ½ðc2rev 2  c1 2 Þ
ηc ¼ 1 : (4.3)
ws12
where the integrations are carried out along the chosen reversible compression path.
For the case of the polytropic efficiency, the reversible reference path coincides
with the static states of the actual compression path, giving
Ð2
vdp þ ½ðc2 2  c1 2 Þ
ηp ¼ 1 , (4.4)
ws12
where state 2 is the actual state of the gas at the outlet condition. Furthermore, for this
process the shaft work is the sum of the reversible work and the dissipation energy that
occurs in the real process and can be written as
ð2 ð2

2 
2 
ws12 ¼ y12 þ ½ c2  c1 þ j12 ¼ vdp þ ½ c2  c1 þ Tds:
2 2
(4.5)
1 1

On this basis, the polytropic efficiency is defined as

Ð2 Ð2
1 vdp þ ½ðc2  c1 Þ
2 2
Tds j
ηp ¼ ¼1 1 ¼ 1  12 (4.6)
ws12 ws12 ws12
as already given in (2.44), but there it was derived without considering the kinetic
energy terms. An ideal reversible machine with no dissipated energy would have an
efficiency of unity. For the polytropic efficiency, the deviation from unity is directly
related to the dissipation incurred along the real process path as a fraction of the shaft
work. The common engineering use of the value 1-η as a convenient statement of the
losses is then more or less self-evident.
The integration of Tds along the actual compression path is the only correct
measure of the internal dissipation or lost work that has occurred in the process. By
the same token, from (4.6), the polytropic efficiency might be considered to be the true
efficiency of a compressor. However, this definition requires detailed knowledge of
the thermodynamic properties along the actual compression path. It is shown in this
chapter that both the isothermal and isentropic efficiencies have the advantage that the
reference reversible compression path is fully defined by the end states alone. This
detailed knowledge of the true compression path is generally not available, and the
integration of vdp along it is then not possible. In practice, the polytropic efficiency
has to be defined by reference processes based on a good approximation to the real
compression path.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.3 Isentropic Efficiency 111

4.3 Isentropic Efficiency

4.3.1 The Static–Static Isentropic Efficiency

In the first instance, the isentropic efficiency is considered here in terms of static states
assuming the kinetic energy is negligible. The extension to include total states and the
changes in kinetic energy is given in Sections 4.3.2 and 4.3.3. If the kinetic energy is
negligible, there is no relevant difference between the total and static conditions and
the shaft work is then the same as the change in the static enthalpy. The isentropic
efficiency defines an isentropic (constant entropy) compression path as the reference
process to determine the reversible work. This isentropic path lies between the initial
state and the final pressure that is achieved in the actual process, as represented by a
vertical line in the h-s diagram shown in Figure 4.1. The final condition of the
reference process, 2s. is the state at the same outlet pressure as the real process, p2,
but with the entropy of the initial state, s1.
For a small isentropic step with ds ¼ 0 along the compression path, as shown in
Figure 4.2, the Gibbs equation reduces to (vdp)s ¼ dhs, demonstrating that the
reversible work along the isentropic path is equal to the isentropic enthalpy change
in the process. The small-scale isentropic efficiency is therefore simply defined as
the ratio of the isentropic to the actual enthalpy rise. This can be written differen-
tially as
ðvdpÞs dhs
ηs ¼ ¼ : (4.7)
dh dh
The overall static–static isentropic efficiency of the machine is defined as the ratio of
the overall isentropic enthalpy rise to the actual enthalpy rise in the compressor based
on static conditions:
Ð 2s Ð 2s
ðvdpÞs dhs h2s  h1
ηs ¼ 1 Ð 2 ¼ Ð1 2 ¼ : (4.8)
dh dh h2  h1
1 1

p + dp
h
p
dh dhs

s
Figure 4.2 Reference process for the small-scale isentropic efficiency.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 112 Efficiency Definitions for Compressors

The reference isentropic process is both reversible and adiabatic. The denominator in
the equation is the change in static enthalpy, but if the change in kinetic energy is
negligible, then this is the same as the change in the total enthalpy, which is the proper
definition of the shaft work. The isentropic efficiency is also referred to as the
adiabatic efficiency, although this terminology could be misleading as an adiabatic
process is not necessarily isentropic.
Isentropic efficiency is easy to understand and can be straightforwardly applied to
real gases as well as ideal gases. Furthermore, the reversible work of the reference
isentropic process is a function of the inlet temperature and the pressure ratio; it is not
affected by the actual performance of the machine. This means that knowledge of the
end state of the real process is not required in order to calculate the reversible
aerodynamic work. In connection with thermodynamic cycle calculations, the isen-
tropic efficiency leads to simple expressions, and is therefore generally used in relation
to processes in single-stage turbochargers, refrigeration cycles and gas turbine cycles,
as shown in Sections 18.9.1 and 18.10.1.
Making use of the isentropic relations for a perfect gas from Section 2.7.1, the
static–static isentropic stage efficiency can be written as
γ1 γ1
T 2s  T 1 ðT 2s =T 1 Þ  1 ðp2s =p1 Þ γ  1 ðp2 =p1 Þ γ  1
ηss ¼ ¼ ¼ ¼ : (4.9)
s
T2  T1 ðT 2 =T 1 Þ  1 ðT 2 =T 1 Þ  1 ðT 2 =T 1 Þ  1
The difference in enthalpy between states 2s and 2, shown in Figure 4.2, at the
pressure level p2, can be obtained by integration of the Gibbs equation for a constant
pressure process, dp ¼ 0, (dh)p ¼ Tds, between states 2s and 2, which shows that
ð2 ð2
h2  h2s ¼ dhp ¼ Tds: (4.10)
2s 2s

This indicates that the measure of the detrimental effect of real irreversibilities on the
isentropic efficiency is the integral of Tds from 2 to 2s. Thus, an entropy-based
isentropic efficiency can be defined as
Ð2
h2s  h1 ðh2  h1 Þ  ðh2  h2s Þ Tds
ηs ¼
ss
¼ ¼ 1  2s : (4.11)
h2  h1 h2  h 1 h2  h1
Equation (4.11) shows that the effective lost work in the definition of isentropic
efficiency is related to integration of the dissipation along an isobar from 2s to 2. This
is in contrast to the integration of the dissipation along the actual compression path from
state 1 to state 2, which is the true measure of the losses in the compressor as is used in
the definition of polytropic efficiency described in the next section. This points to a
major disadvantage of the isentropic efficiency and is discussed in Section 4.3.6.
Following Denton (1993) and Greitzer et al. (2004), a useful engineering approxima-
tion to (4.11), with only a small error when the efficiency is high, may be obtained by
assuming that the temperature is nearly constant along the path from 2s to 2 such that
T 2 ðs2  s1 Þ
ηs  1  : (4.12)
h 2  h1

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.3 Isentropic Efficiency 113

This indicates that the isentropic efficiency is related to both the entropy change in the
process as well as the temperature at the compressor outlet. In practice, the isentropic
efficiency of the compressor is normally presented in the form of total–total and total–
static isentropic efficiencies, where these are defined in terms of the thermodynamic
state of the gas (total or static) that is used to represent the inlet and the outlet conditions.

4.3.2 The Total–Total Isentropic Efficiency

The total–total isentropic efficiency extends the definition of static–static isentropic
efficiency, given in the preceding section, to consider changes in the kinetic energy in
the flow and is visualised in Figure 4.3. The total energy available at the inlet of the
compressor is given by the total or stagnation conditions. If the outlet kinetic energy can
be used in a downstream component, the total energy at the outlet is also of interest. This
is the case at the outlet of an intermediate stage in a multistage compressor. It also
occurs when a further increase in static pressure can be achieved by downstream
stationary components. In such instances, the outlet total condition may be specified
as the useful end state of the process. The actual work required in the machine is then
compared with the work in an ideal machine which compresses the flow isentropically
from the inlet total condition to the same total pressure as the actual process.
Following the procedure given by (4.3) for the efficiency, in which the changes in
the kinetic energy are added to the reversible aerodynamic work to give the reversible
work, this results in the total–total isentropic stage efficiency defined as

Figure 4.3 Reference processes for the static–static, total–total and total–static efficiencies.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 114 Efficiency Definitions for Compressors

Ð 2s Ð 2s Ð 2s Ð t2s
ðvdpÞs þ 1 ðcdcÞs 1 dhs þ ½ðc2s  c1 Þ
2 2
dhts ht2s  ht1
ηtts ¼ 1
Ð t2 ¼ Ð t2 ¼ Ðt1t2 ¼ :
ht2  ht1
t1 dht t1 dht t1 dht
(4.13)
An important aspect of the total–total isentropic efficiency is that the integration of the
aerodynamic work vdp along the isentropic path of static states and subsequently
adding the change in kinetic energy leads to the same result for the reversible work as
the integration of vtdpt along the isentropic path of total states because of the following
identity from the Gibbs equation (vtdpt)s ¼ dhts:
ð 2s ð t2s ð t2s


wrev ¼ dhs þ ½ c2s 2  c1 2 ¼ dhts ¼ ðvt dpt Þs : (4.14)
1 t1 t1

This identity only holds for an isentropic process and cannot be used for other
processes. It does not hold for a polytropic or isothermal process, but is often
mistakenly used in the efficiency definition for these processes.
Equation (4.13) can be expressed as

ht2s  ht1 ðh2s  h1 Þ þ ½ðc2s 2  c1 2 Þ

ηtts ¼ ¼ : (4.15)
ht2  ht1 ð h 2  h1 Þ þ ½ ð c 2 2  c 1 2 Þ

In practical applications, it is not easily possible to determine the isentropic outlet

velocity at the virtual state 2s which would occur in a reversible isentropic process,
and so c2s is taken to be the same as c2. Where the inlet and outlet kinetic energy are
the same, or where they are negligible in comparison with the stage enthalpy rise,
(4.15) reverts to (4.8), and then the total–total efficiency is the same as the static–static
efficiency. For a perfect gas, the total–total isentropic stage efficiency can then be
written with the help of the isentropic relations as
γ1 γ1
T t2s  T t1 ðT t2s =T t1 Þ  1 ðpt2s =pt1 Þ γ  1 ðpt2 =pt1 Þ γ  1
ηtts ¼ ¼ ¼ ¼ : (4.16)
T t2  T t1 ðT t2 =T t1 Þ  1 ðT t2 =T t1 Þ  1 ðT t2 =T t1 Þ  1

4.3.3 The Total–Static Isentropic Efficiency

In many applications of ventilation fans and single-stage centrifugal compressors, the
kinetic energy at the stage outlet is dissipated further downstream. Here, the resulting
losses can be included in the efficiency of the compressor by subtracting the outlet
kinetic energy from the reversible work. Following the same procedure given by (4.3),
this leads to the total–static isentropic efficiency as
hÐ
2 i Ð
2S
ð vdp Þ þ ½ c  c 2 t2s
t1 dhs  ½c1 ðh2s  h1 Þ  ½c21 h2s  ht1
s 2s 1 2
1
ηtss ¼ Ð t2 ¼ ¼ ¼ :
dht Δht ht2  ht1 ht2  ht1
t1
(4.17)

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.3 Isentropic Efficiency 115

In other words, the total–static isentropic efficiency compares the actual work
required in the real machine with the work in an isentropic machine which com-
presses the flow from the inlet total condition to the same outlet static pressure. The
integration of the aerodynamic work is still considered to be along the reversible
path of static conditions. For a perfect gas, the total–static isentropic stage effi-
ciency can be written as
γ1 γ1
T 2s  T t1 ðT t2s =T t1 Þ  1 ðp2s =pt1 Þ γ  1 ðp2 =pt1 Þ γ  1
ηtss ¼ ¼ ¼ ¼ : (4.18)
T t2  T t1 ðT t2 =T t1 Þ  1 ðT t2 =T t1 Þ  1 ðT t2 =T t1 Þ  1

4.3.4 Comparison of Total–Total and Total–Static Isentropic Efficiencies

Equations (4.13) and (4.17) can also be used to analyse the development of the
isentropic efficiencies through a compressor. By placing the end point, state 2, at
any location along the compression path, such as the impeller outlet or the diffuser
outlet, the isentropic efficiency from the inlet up to that point is obtained. The
difference between the total–total and total–static isentropic efficiencies at any plane
in the compressor, ξ, is the ratio of the isentropic kinetic energy available at that plane
to the overall shaft work as

ht2s  h2s ½c22s c2 ht2  ht1

ξ ¼ ηtts  ηtss ¼ ¼ ¼ 2s2 , λ¼ , (4.19)
ht2  ht1 ht2  ht1 2λu2 u22

where λ is the nondimensional work coefficient. The total–total isentropic efficiency at

a specific location in the compressor shows the theoretical upper limit for the total–
total stage isentropic efficiency that can be achieved at the outlet of the downstream
components if no further losses are incurred. The total–static isentropic efficiency at
that location shows the theoretical lower limit for the total–total stage isentropic
efficiency if none of the kinetic energy at that location is converted into pressure rise
and there are no further losses. In this way, the ratio of the local kinetic energy at a
specific location to the shaft work, ξ, is a measure of the maximum potential for
improvement in isentropic efficiency that can be gained downstream of that plane.
The difference between these two efficiencies, ξ, is the ratio of the kinetic energy at
2
that plane to the shaft work input. The isentropic kinetic energy, ½c2s , is the maximum
available kinetic energy that can be transformed to static pressure rise in downstream
components in an isentropic process (without incurring any losses). In practical
applications, it is assumed that the isentropic velocity, c2s, is the same as the actual
local velocity, c2. Considering state 2 to be at the impeller outlet, the kinetic energy at
this location is commonly accounted by the so-called kinematic degree of reaction as
described in Section 2.2.5, defined as the ratio of the total–static to the total–total
enthalpy rise across the impeller:

h2  ht1 ht2  ht1  ½c22

rk ¼ ¼ ¼ 1  ξ 2: (4.20)
ht2  ht1 ht2  ht1

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 116 Efficiency Definitions for Compressors

For typical stages with a degree of reaction of close to 0.6, the difference in the
total–total and the total–static efficiencies at the impeller outlet is then 0.4. The
total–total efficiency is then a highly misleading measure of the impeller efficiency
as it can be increased simply by reducing the degree of reaction of the impeller
through increasing the kinetic energy at impeller outlet, but as its value is often
above 90% many authors like to use it.

4.3.5 Evolution of the Isentropic Efficiency in the Compressor

The evolution of the total–total and total–static isentropic efficiencies at different
locations downstream of a typical centrifugal compressor impeller operating at its
design point is shown schematically in Figure 4.4. The total–total isentropic effi-
ciency at the impeller outlet is the maximum isentropic efficiency that can be
achieved by the compressor if the impeller exit flow is brought to rest isentropically.
Its value is obviously less than unity due to the losses already incurred in the
impeller and will further reduce as more losses are incurred in the downstream
stationary components. The total–static isentropic efficiency at the impeller outlet
gives the minimum isentropic efficiency that can be achieved by the compressor
stage if no further static pressure rise is achieved downstream of the impeller,
whereby all the dynamic head at the impeller outlet is lost in the machine. This is
the most pessimistic measure of compressor efficiency. As the flow proceeds
through the compressor to the diffuser and the volute outlet, the total pressure
decreases due to losses, while the static pressure rises depending on the pressure
recovery in the diffuser and volute sections. Therefore, the value of ξ reduces,
meaning that there is less potential for further gains in efficiency. In other words,
the impact of downstream components on efficiency reduces as the flow decelerates
downstream of the impeller outlet. The value of ξ, given in (4.19), identifies the
sensitivity of the stage efficiency to the losses in the downstream components. At the
diffuser inlet, the typical ratio of the local kinetic energy to the stage work is 0.4. At
the inlet to a return channel, it may be 0.08, which indicates that in terms of

h x=
c2
2lu22 K stt
x 2 = 1 - rk
Ksts
Impeller Diffuser Stage
outlet outlet outlet m
Figure 4.4 Evolution of the isentropic efficiency downstream of an impeller.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.3 Isentropic Efficiency 117

Figure 4.5 Calculated performance of a stage and impeller.

efficiency an improvement in the loss coefficient of the diffuser will be about five
times more effective than a similar improvement to the return channel.
The trend in the variation of the total–total efficiency and the associated total–total
pressure ratio through a turbocharger stage from a CFD calculation is shown in
Figure 4.5. The total–total efficiency and pressure ratio drops from the impeller outlet,
2, via the diffuser outlet, 3, the volute outlet, 4, to the flange, 5, as the kinetic energy at
impeller outlet is not completely converted into pressure rise.

4.3.6 Divergence of Constant Pressure Lines

Despite its simplicity, a major disadvantage of isentropic efficiency is that it does not
accurately reflect the aerodynamic quality of the machine. In practice, the isentropic
efficiency falls as the pressure ratio increases due to a purely thermodynamic effect. The
isentropic efficiency of a compressor with two stages, each having the same isentropic
efficiency, is lower than that of the individual stages due to the displacement of the constant
pressure lines. The vertical gap between the enthalpy of two isobars in an h-s diagram
increases with the entropy such that a larger enthalpy change is required to produce the
same pressure ratio at a higher temperature. The isentropic efficiency considers the
enthalpy rise from state 1 to state 2s as the reversible work for comparison with the actual
work input in its definition. The actual increase in the isentropic enthalpy from state 1 to the
outlet pressure of the two-stage machine is then less than that of the two stages considered
individually. This can be seen in the following equations and in Figure 4.6.
Δh1s Δh2s Δh1s þ Δh2s
ηsð1Þ ¼ ¼ ηsð2Þ ¼ ¼
Δh1 Δh2 Δh1 þ Δh2
Δh1s þ Δh2ss Δh1s þ Δh2s þ ðΔh2ss  Δh2s Þ (4.21)
ηsð1þ2Þ ¼ ¼
Δh1 þ Δh2 Δh1 þ Δh2
Δh2s > Δh2ss , ) ηsð1þ2Þ < ηsð1Þ , ηsð2Þ :

The drop in isentropic efficiency for the two-stage machine is related to integration of
the dissipation energy along an isobar from 2s to 2, as opposed to the integration along
Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 118 Efficiency Definitions for Compressors

p3
h 3
p2
3s
Dh2
Dh2s
3ss
Stage 2
Dh2ss
2 p1
2s
Dh1
Dh1s
Stage 1

s
Figure 4.6 The isentropic efficiency of a two-stage compressor.

the actual compression path from state 1 to 2, as shown in (4.11). The average
temperature increases from 2s to 2 with increasing the pressure ratio, so the second
stage with the same dissipation losses has a lower efficiency! The increase in the work
required in the second stage accounts for the fact that the inlet to the second stage is
preheated by the first stage, and as a consequence requires more work for the same
pressure ratio, as shown in Section 2.7.2 and in Figure 2.16. This problem can be
avoided by using polytropic efficiency.

4.4 Polytropic Efficiency

4.4.1 The Polytropic Process

The key difference between the polytropic and the isentropic efficiency is that the
reference compression process in the definition of the reversible work is taken as a
reversible compression path from the inlet state 1 to outlet state 2. Of course, the real
process from state 1 to state 2 is not reversible, but in the determination of the
reversible work the changes of state that are used are those that occur in the real
process. It is shown in this section that the intermediate states can be achieved in a
reversible diabatic process from state 1 to state 2 if the reversible heat added in the
reversible process is the same as the dissipation occurring at each step along this path.
Looking at a differential compression step shown in Figure 4.2, the small-scale
polytropic efficiency is defined as the ratio of the reversible enthalpy rise, vdp, to the
actual enthalpy rise, vdp/dh. This leads to the definition of the small-scale polytropic
efficiency as
vdp dh  Tds Tds
ηp ¼ ¼ ¼1 : (4.22)
dh dh dh

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.4 Polytropic Efficiency 119

In a real compression process, this small-scale efficiency varies along the actual
compression path, as can be seen in the h-s diagram of the compression path of a
typical compressor in Figure 2.17. To calculate the aerodynamic work along this path,
the integration of the aerodynamic work, vdp, must be carried out as
Ð2 Ð2
vdp vdp y12
ηp ¼ Ð12 ¼ 1 ¼ : (4.23)
dh h 2  h 1 h 2  h1
1

This integration requires a detailed knowledge

. of the thermodynamic properties along
the actual path, which is not available in practice. Instead, the aerodynamic work in
this process can be calculated along a so-called polytropic path, hypothesised as a path
between the actual initial and final states, along which the small-scale efficiency
remains constant. It has been shown in Section 2.5.7 that, for an ideal gas, this
corresponds to a process described with pvn ¼ constant. The exponent n is the
polytropic exponent and from Section 2.5.8 is related to the polytropic efficiency in
an adiabatic process as
n=ðn  1Þ n  1 ð γ  1Þ γηp
ηp ¼ , ¼ , n¼ : (4.24)
γ=ðγ  1Þ n γηp γηp  ðγ  1Þ

The equation, pvn ¼ constant does not suffice to define the polytropic process for real
gases, and the calculation of the aerodynamic work and the polytropic efficiency
becomes more complex, as outlined in Section 3.6. This is the main weakness of the
polytropic efficiency compared with the isentropic efficiency.
The assumption of a constant efficiency along the compression path is consistent
with the view that two compressor stages with the same efficiencies should result in
the same overall efficiency if they are combined to give a two-stage compressor. The
overall reversible work in a polytropic process is referred to as ‘polytropic head’ or
aerodynamic work and is denoted in (4.23) by y12. The polytropic process is discussed
in detail in Section 2.7.3 for ideal gases and in Section 3.5.1 for real gases.
The polytropic efficiency is also referred to as the ‘small-scale isentropic effi-
ciency’. For a small compression step along a polytropic path, the Gibbs equation
shows that the isentropic enthalpy rise, dhs, is equal to the polytropic work, vdp, and
therefore the polytropic and isentropic efficiencies of a small compression step are
identical. This can be understood by considering that a small compression step from p
to p þ dp along the polytropic path can be replaced by an isentropic work input, dhs,
followed by an isobaric reversible heat addition that generates the same entropy
increase as the irreversible loss generates in the real process. The work done during
the heat addition is zero, and so the reversible (polytropic) work from p to p þ dp is
equal to the isentropic work, dhs . Thus, for a small compression step, the isentropic
and polytropic efficiencies are identical.
The overall polytropic work is then calculated by integrating the small-scale
isentropic enthalpy rises along the polytropic path. Given that the overall isentropic
enthalpy rise is obtained by summation of the small-scale isentropic enthalpy rises
along an isentropic path from p1 to p2s, it is always smaller than the polytropic head.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 120 Efficiency Definitions for Compressors

The polytropic efficiency is, therefore, always larger than the isentropic efficiency for
a compressor.
The choice of the thermodynamic conditions, either total or static conditions, to
define the end states of the polytropic process results in different definitions for
polytropic efficiency. These are discussed in Sections 4.4.4–4.4.6.

4.4.2 Preheat Effect

Following (4.11), the static–static isentropic efficiency can be related to the static–
static polytropic efficiency as
Ð2
h2s  h1 ðh2s  h1 Þ 1 vdp ðh2s  h1 Þ ss
ηs ¼
ss
¼ Ð2 ¼ Ð2 ηp ¼ Ph ηss
p, (4.25)
h2  h1 vdp h2  h1 vdp
1 1

where Ph is known as the preheat factor, which directly relates the polytropic
efficiency to the isentropic efficiency of the stage. The physical background to the
preheat effect is given in the discussion in Section 2.7.2. From the Gibbs equation, the
preheat factor can be rearranged as
Ð2 Ð2 Ð2 Ð2 Ð2
ðh2  h1 Þ  ðh2  h2s Þ 1 vdp  1 Tds  2s Tds 2s Tds  1 Tds
Ph ¼ Ð2 ¼ Ð2 ¼1 Ð2 :
1 vdp 1 vdp 1 vdp
(4.26)
Considering that the entropy changes from 1 to 2 and from 2s to 2 are the same, the
integral of Tds from state 2s to state 2 is always larger than that from 1 to 2 as the
integral is carried out at a higher average temperature. This indicates that the preheat
factor in a compressor is always less than unity and that the isentropic efficiency is
always smaller than the polytropic efficiency. The difference between the isentropic and
polytropic efficiencies with increasing pressure ratio, as shown in Figure 4.7, is given by

ðγ1Þ=γ 
π 1
ηs ¼
ðγ1Þ=η γ : (4.27)
π p  1

4.4.3 Polytropic Efficiency in Terms of Entropy Changes

The Gibbs equation for a thermally perfect gas can be shown to be
   
dT dp
ds ¼ cp R : (4.28)
T p

For a small-scale process at constant pressure, that is with dp ¼ 0, the Gibbs equation
leads to (ds)p ¼ cp(dT/T) and for a process at constant temperature, with dT ¼ 0,

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.4 Polytropic Efficiency 121

Figure 4.7 Isentropic and polytropic efficiency for compressors with γ ¼ 1.4.

it leads to (ds)T ¼ –R(dp/p). Hence, the small-scale polytropic efficiency in (4.22) can
be derived to be
0 1
dp
R
B p C ðdsÞT
ηp ¼ B C
@ dT A ¼  ðdsÞ , (4.29)
cp p
T
which, on integration, for a large-scale process leads to the entropy-based polytropic
efficiency
ðΔsÞT s1h  s2
ηp ¼  ¼ : (4.30)
ðΔsÞp s1h  s1

The denominator is the change in entropy along a line of constant pressure (the
inlet pressure) between the inlet and outlet temperature, and the abscissa is the
change in entropy along a line of constant temperature between the inlet and outlet
pressure. This formulation allows the polytropic efficiency to be represented in an
h-s diagram as the ratio of two horizontal lengths (c/d in Figure 4.8) on the entropy
scale (abscissa) just as the isentropic efficiency can be identified as the ratio of two
vertical lengths (a/b) on the enthalpy scale (ordinate) (Barbarin and Mikirtichan,
1982; Casey, 2007).
In practice, the polytropic efficiency may, like the isentropic efficiency, take
different forms depending on the choice of the thermodynamic conditions (total or
static conditions) to define the end states of the process. These forms are discussed in
the following.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 122 Efficiency Definitions for Compressors

Figure 4.8 The isentropic and polytropic efficiencies in an h-s diagram.

4.4.4 Static–Static Polytropic Efficiency

If it is only the quality of the static enthalpy rise in the machine that is of interest, a so-
called static-to-static polytropic efficiency is used. It is defined as the ratio of the
polytropic enthalpy rise along a polytropic path between inlet static and outlet static
conditions to the actual enthalpy rise along the same path; see Figure 4.9. The
differential and integral forms of the static-to-static polytropic efficiency can be
written as
vdp
ηss
p ¼ (4.31)
dh
Ð2 Ð2
vdp 1 vdp y12
ηss ¼ Ð2
1
¼ ¼ : (4.32)
p
dh h2  h1 ws12  ½ðc2 2  c1 2 Þ
1

where ½(c22  c12) is the change in kinetic energy from state 1 to state 2. If the change
in kinetic energy across the machine is negligible, the static–static polytropic effi-
ciency is the true efficiency of the machine.
For ideal gases, where the polytropic process can be represented by the relation
pvn ¼ constant, the aerodynamic work in (4.32) can be derived from integration of vdp,
ð2 " n1 # " n1 #
n p2 n n p2 n
y12 ¼ vdp ¼ p v1 1 ¼ RT 1 1 :
1 ð n  1Þ 1 p1 ð n  1Þ p1
(4.33)
For a perfect gas, the static–static polytropic efficiency can be derived by direct
integration of (4.31) as

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.4 Polytropic Efficiency 123

h pt 2
t2
ht 2 p2
2
½c 2 True polytropic process
h2 2 represented by static-to-static
polytropic path
pt 2
t1
ht1 2
p1
½c 1
Nonphysical polytropic process
h1
1 represented by total-to-total
polytropic path

s1 s2 s
Figure 4.9 Polytropic efficiency defined on two process paths.

      c dT    
vdp RTdp dp p γ dT
ηss
p ¼ ¼ ) ¼ ηss ¼ η ss
dh cp pdT p p
R T p
γ1 T
     
p γ T2 γ  1 ln ðp2 =p1 Þ
ln 2 ¼ ηss ln ) ηss
p ¼ :
p1 p
γ1 T1 γ ln ðT 2 =T 1 Þ
(4.34)
These equations are generally used to determine the aerodynamic work and the
efficiency from measurements of temperature and pressure.

4.4.5 Total–Total Polytropic Efficiency

As with the isentropic efficiency, a polytropic efficiency between the total conditions
at the inlet and outlet can be defined. This definition assumes that the kinetic energy at
the outlet is useful and can be included in the useful output of the compressor with the
aerodynamic work that is achieved. From (4.3), the reversible work is the sum of the
aerodynamic work and the change in kinetic energy, so the definition of the total–total
polytropic efficiency is then given as
Ð2
y þ ½ð c 2 2  c1 2 Þ vdp þ ½ðc2 2  c1 2 Þ
ηttp ¼ 12 ¼ 1 : (4.35)
ws12 ws12
This can be considered to be the true definition of the total–total polytropic efficiency
as it integrates the aerodynamic work with the static states and considers the kinetic
energy separately.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 124 Efficiency Definitions for Compressors

The total–total polytropic efficiency is, however, commonly defined in a less

cumbersome but inexact way. In a similar approach as given in (4.14) for the
isentropic efficiency, this definition posits a polytropic path traversing all the total
conditions between the inlet total conditions, t1, and outlet total conditions, t2. The
polytropic efficiency is then taken as the ratio of the aerodynamic work along this
polytropic path of total conditions to the actual enthalpy rise; see Figure 4.9. This
(incorrect) definition of the total–total polytropic efficiency can be written in differen-
tial and integral forms as
vt dpt
ηttpt ¼ (4.36)
dht
Ð t2
vt dpt ytt
ηttpt ¼ t1
¼ 12 : (4.37)
ht2  ht1 ws12
In line with the derivations of (2.80) and (4.34), the total–total aerodynamic work can
then be written as
" n1 # " n1 #
n p n
n p n
ytt12 ¼ p vt1 t2
1 ¼ RT t1 t2
1 (4.38)
ðn  1Þ t1 pt1 ð n  1Þ pt1

and efficiency as
γ  1 ln ðpt2 =pt1 Þ
ηttp ¼ : (4.39)
γ ln ðT t2 =T t1 Þ

The total–total polytropic efficiency defined in this way is, of course, not the true
polytropic efficiency. The definition is based on the reversible work calculated along a
path of total conditions from t1 to t2 rather than the path of static conditions from 1 to
2 with separate consideration of the kinetic energy, as shown in Figure 4.9. The
difference between the two equations for the polytropic efficiency arises because of
the following inequality:
ð2 ð 2t


vdp þ ½ c2 2  c1 2 6¼ vt dpt (4.40)
1 1t

such that
ð2 ð2 ð 2t ð 2t


wshaft ¼ vdp þ ½ c2 2  c1 2 þ Tds 6¼ vt dpt þ T t ds: (4.41)
1 1 1t 1t

The difference between the two formulations for the reversible work is the difference
in dissipation energy along the total and the static paths, which is larger for the total
path as the total temperature is larger than the static temperature. This difference
reduces to zero for an isentropic process with ds ¼ 0. This implies that the total–total
polytropic efficiency, as defined in (4.39), is always slightly less than the true
polytropic efficiency of the compressor, (4.35). In practice, the difference is small,
of the order of 1% or less depending on the level of kinetic energy change in the

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.4 Polytropic Efficiency 125

machine compared to the work input. This difference may be smaller than the
difference due to the fact that the integration of vdp is generally carried out along a
virtual polytropic path rather than the path of real states between the inlet and
the outlet.

4.4.6 Total–Static Polytropic Efficiency

As with isentropic efficiency, a total–static polytropic efficiency can be defined to
measure the efficiency of the compression from the inlet total condition to the outlet
static condition to consider those cases where the outlet kinetic energy is not a useful
output from the compressor. Different definitions for total–static polytropic efficiency
have been used in the literature, a number of which are discussed here.
Considering that the exit kinetic energy is lost, the useful total–static polytropic
work from t1 to 2 can be defined as the reversible work minus the kinetic energy at the
outlet. Following the same procedure as given in (4.3) yields
hÐ i
 ½c2 2 Ð 2 vdp  ½c1 2
2
1 vdp þ ½ ð c 2
2
 c 1
2
Þ
ηtsp ¼ ¼ 1 : (4.42)
ws12 ht2  ht1
This is analogous to the definition of the total–static isentropic efficiency, where the
isentropic work required to compress the flow from state t1 to state 2 is compared with
the actual shaft work consumed. If this were to be used for the definition of the total–
static efficiency of an impeller it would lead to values near to 0.5 (50%), given that the
kinetic energy at impeller outlet is about 40% of the work input.
It is therefore more useful to define the total–static polytropic efficiency based on
the actual shaft work minus the kinetic energy at the outlet. This represents the fraction
of the actual work expended to compress the flow from the inlet total to the outlet
static conditions:
hÐ i
2 Ð2
vdp þ ½ðc2 2  c1 2 Þ  ½c2 2
1 vdp  ½c2
2
1
ηp ¼
ts
¼ : (4.43)
ws12  ½c2 2 ht2  ht1  ½c2 2
A slightly different approach may be adopted, as presented by Mehldahl (1941) and
used by Dalbert et al. (1988), whereby a polytropic path is assumed between inlet total
and outlet static conditions, and the total–static polytropic work is calculated by
integration of vdp along this path from t1 to 2. This is usually an acceptable
approximation as the inlet and outlet kinetic energy in the inlet and outlet planes are
small. The total–static polytropic efficiency is then derived as
Ð2
t1 vdp
ηp ¼
ts
: (4.44)
ht2  ht1  ½c2 2
A popular definition for total–static polytropic efficiency is similar to that given
in (4.39) except that the total pressure at the outlet is replaced with the static
pressure:

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 126 Efficiency Definitions for Compressors

γ  1 ln ðp2 =pt1 Þ
ηtsp ¼ : (4.45)
γ ln ðT t2 =T t1 Þ

The popularity of this equation is due to the fact that it is directly related to the
measurable thermodynamic properties at the inlet and the outlet and so obviates
the need for any knowledge of the exit dynamic head. However, it cannot be
derived from direct integration of vdp from static states along a polytropic path to
give the aerodynamic work and, unlike (4.38), has a less physical basis for its
definition. It is an acceptable approximation for situations where the exit kinetic
energy is small.

4.4.7 Stage and Component Polytropic Efficiencies

In Section 4.4.4, the static–static polytropic efficiency of a compression process was
defined as the ratio of the polytropic work along the polytropic path to the static
enthalpy change between the initial and the final states. By dropping the suffix p (for
polytropic) for convenience, the static–static efficiencies for the stage, impeller and the
diffuser can be written as
Ð3 Ð3
1 vdp 1 vdp
ηstage ¼
ss
¼ (4.46)
h3  h1 Δht  ½ðc3 2  c1 2 Þ
Ð2 Ð2
1 vdp 1 vdp
ηimp ¼
ss
¼ (4.47)
h2  h1 Δht  ½ðc2 2  c1 2 Þ
Ð3 Ð3
2 vdp 2 vdp
ηdiff ¼
ss
¼ , (4.48)
h3  h2 ½ ð c 2 2  c3 2 Þ

where 1, 2 and 3 denote the stage inlet, impeller outlet and the stage outlet. In this
extension of the earlier derivations, the polytropic efficiency for the diffuser is defined
in terms of the aerodynamic work achieved as a fraction of the change in kinetic
energy across the diffuser. If it is assumed that the polytropic head for the stage is the
sum of the polytropic head in each of its components, which is only exactly the case if
state 2 lies on the polytropic path from state 1 to 3, then
     2 
½c23 ½c21 ½c22 ½c21 ½c2 ½c23
ηstage 1 
ss
þ ¼ ηimp 1 
ss
þ þ ηdiff
ss
 (4.49)
Δht Δht Δht Δht Δht Δht

or
ð1  ξ 3 þ ξ 1 Þ ηss
stage ¼ ð1  ξ 2 þ ξ 1 Þ ηimp þ ðξ 2  ξ 3 Þ ηdiff ,
ss ss
(4.50)

where ξ1, ξ2, and ξ3, are the ratios of the local kinetic energy to the shaft work at planes
1, 2 and 3. It is possible to extend the analysis to include additional locations within
the stage, for example a location between the diffuser and the downstream volute or
return channel could be added. This would lead to the following extension to (4.50),

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.4 Polytropic Efficiency 127

where state 1 is the stage inlet, state 2 the impeller outlet, state 3 the diffuser outlet and
state 4 the stage outlet:

ð1  ξ 4 þ ξ 1 Þ ηss
14 ¼ ð1  ξ 2 þ ξ 1 Þ η12 þ ðξ 2  ξ 3 Þ η23 þ ðξ 3  ξ 4 Þ η34 :
ss ss ss
(4.51)

With knowledge of which contribution each component makes to the overall effi-
ciency, it is relatively easy to determine whether it is worthwhile expending develop-
ment effort on a particular part of the machine. In common practice, the total–total
efficiency is used to measure the overall performance of the stage giving ξ1 ¼ ξ3 ¼ 0
in (4.50). Substituting the degree of reaction for ξ2 from (4.19) into 4.50 and putting
ξ1 ¼ ξ3 ¼ 0 for inlet and outlet total conditions, the total–total efficiency of the stage
can be written in terms of the degree of reaction, impeller total–static and diffuser
static–total efficiencies:

ηttstage ¼ r k ηtsimp þ ð1  rk Þηstdiff : (4.52)

An exact derivation leading to the same expression is provided by Mehldahl (1941).

This equation allows the performance of the components of the impeller and the
diffuser to be investigated separately once the stage total–total polytropic efficiency
and the static pressure at impeller outlet are known. A procedure for this is given in
Section 20.4.4. The first derivation of these equations in this form goes back to the
work of Mehldahl (1941) and subsequent relevant publications are the papers of
Hausenblas (1965), Ambuehl and Bachmann (1980) and Dalbert et. al. (1988).
An example showing the stage efficiency and the separate impeller and diffuser
efficiencies for a backswept industrial process compressor stage tested with a range
of different vaned diffusers with different vane inlet angles is given in Figure 4.10,
adapted from the data given by Dalbert et al. (1988). This diagram does not include a
scale, as the original data was published in a normalised form, but absolute values
are not necessary to show how the impeller efficiency and the diffuser efficiency
combine with the degree of reaction to determine the stage efficiency. In this
example, at a tip speed Mach number of 0.8 the impeller has a wide range of high
efficiency from 80% to 120% of the design flow. Each diffuser tested has a much
narrower operating range covering only part of the full operating range of the
impeller. The degree of reaction remains nearly constant across the operating range
and does not change when the diffuser changes.
In this case, the research stage was tested with many diffusers, but in many test
campaigns, this procedure is not possible. The best stage efficiency and widest
operating range is obtained at the tip speed Mach number of 0.8 with the diffuser
setting angle of 64 . If the information of the impeller and the diffuser efficiency were
not available, then the stage test data would not immediately identify why the
efficiency drops at high flow on this characteristic. The separation of the stage
efficiency into separate diffuser and impeller efficiencies identifies in this case that
the impeller efficiency remains high over the operating range of this diffuser and that
the shape of diffuser efficiency characteristic essentially determines the stage effi-
ciency and hence the shape of the pressure rise characteristic. The shift in the stage

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 128 Efficiency Definitions for Compressors

himp
ts

rk
hdiff
st

a 2¢ d = 72° 68° 64 ° 60 ° 56 °

h stage
tt

ft1 / ft1ref

Figure 4.10 Impeller efficiency, degree of reaction and the diffuser efficiency combine to define
the stage efficiency with different diffuser angle settings. (data provided by courtesy of MAN
Energy Solutions)

characteristic with diffuser setting angle is determined by the throat area of the diffuser
vanes; see Section 12.9.
If, as in Figure 4.10, the impeller efficiency remains fairly constant over a wide
range, the diffuser efficiency must drop to very low values to lead to a stage efficiency
of zero, as at zero stage efficiency the diffuser efficiency is given by
rk
ηstdiff ¼  ηts : (4.53)
ð1  r k Þ imp

By examining a virtual case where the diffuser has no loss, ηstdiff ¼ 1, it is possible to
show that the total–total efficiency of the impeller becomes
 
ηttimp ¼ r k ηtsimp þ ð1  r k Þ ¼ 1  r k 1  ηtsimp : (4.54)

In this form, it is easy to see that for a given total–static efficiency of the impeller a
decrease in the degree of reaction (by increasing the kinetic energy at the impeller
outlet) leads to an increase in the total–total impeller efficiency. This feature makes the
use of the total–total impeller efficiency rather awkward as a comparative measure of
the quality of a stage, as the associated losses are only really specified if the degree of
reaction is also specified.
A further advantage of the three efficiency definitions given in (4.46)–(4.48) is that
at peak efficiency, as shown in Figure 2.14, the static state at impeller outlet lies close
to the polytropic path of the stage compression process so that the numerical values of
the three efficiencies are similar.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.5 The Impeller Wheel Efficiency 129

4.5 The Impeller Wheel Efficiency

The static enthalpy rise in the impeller includes a substantial portion which is often
considered to be isentropic as it is due to the lossless centrifugal effect from the
change in the blade speed from inlet to outlet. Vavra (1970a) suggested a definition of
the impeller efficiency, called the wheel efficiency, to assess the efficiency of the
diffusion process in the impeller without the centrifugal effect. Because the work done
by the centrifugal effect was effectively loss-free, he argued that the wheel efficiency
without the centrifugal effect was a better measure of the quality of the rotor
performance than the impeller static–static efficiency. The wheel efficiency is not
often used but has been examined by Moore et al. (1984) and by Larosiliere et al.
(1997) in the interrogation of their CFD simulations.
Vavra defined the wheel efficiency in terms of an isentropic efficiency as
h2s  hu h2s  h1  ½ðu2 2  u1 2 Þ
ηws ¼ ¼ , (4.55)
h2  hu h2  h 1  ½ ð u 2 2  u 1 2 Þ
where hu is the enthalpy at state u as shown in Figure 2.18. Vavra evaluated the inlet
blade speed as that at the casing, but Moore et al. (1984) and Larosiliere et al. (1997)
used the mean wheel speed at inlet. In terms of a polytropic analysis, the Vavra wheel
efficiency can be defined as

y12  ½ðu2 2  u1 2 Þ
ηwp ¼ : (4.56)
h2  h 1  ½ ð u 2 2  u 1 2 Þ

From the Euler equation, the value of the divisor in these equations can be seen to be
the change in the relative kinetic energy in the rotor, ½(w21  w22 ), and in this way the
wheel efficiency relates only to the effectiveness of the diffusion process in the relative
flow of the impeller. Of course, the losses include not only those due to the diffusion
but also all of the other sources, such as tip clearance, secondary flows and viscous
losses. Vavra (1970a) comments that the relatively low value of the wheel efficiency
of radial stages compared to the higher values possible in axial compressors (usually
with no centrifugal effect) suggests that some improvement in aerodynamic design
should be possible.
Rearrangement of (4.56) together with the definition of the static–static polytropic
impeller efficiency and in combination with other dimensionless parameters leads to
 
2
ηss
p λr k  ½ 1  ðr 1 =r 2 Þ
ηwp ¼  , (4.57)
½ðw1 =u2 Þ2 1  ðw2 =w1 Þ2

where λ is the work coefficient and rk is the degree of reaction. This shows that, for a
given diffusion ratio in the impeller, the wheel efficiency falls with a higher inlet
radius ratio as the centrifugal effect is reduced. Larosiliere et al. (1997) examined a
stage with a small inlet casing radius of r1/r2 ¼ 0.486 and with a design de Haller
number of w2/w1 ¼ 0.714. The peak total–total isentropic efficiency of the impeller

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 130 Efficiency Definitions for Compressors

was found to be nearly 94%, but the peak wheel efficiency without the centrifugal
effect was only 20%. Moore et al. (1984) examined a stage with a higher inlet casing
radius of r1/r2 ¼ 0.646 and a design de Haller number of w2/w1 ¼ 0.72. The peak
total–total isentropic efficiency of the impeller was also found to be over 93% but the
peak wheel efficiency without the centrifugal effect was 60%.
The large difference between the wheel efficiency of Larosiliere et al. and that of
Moore et al., despite the total–total efficiency and the de Haller number being nearly the
same, indicates that the wheel efficiency is strongly affected by the impeller radius ratio as
given in (4.57). It turns out that the wheel efficiency is not a very useful parameter to
identify the aerodynamic quality of the diffusion process in the impeller, as it is not really
true that the enthalpy rise due to the centrifugal effect is loss-free. The tip clearance losses
and the friction losses within the impeller can just as well be attributed to the centrifugal
effect as to the change in static enthalpy due to the relative flow deceleration.

4.6 External Losses and Sideloads

In compressors with sideloads, such as the economiser flows in a multistage refriger-

ation compressor, each compressor section has a different mass flow. The overall
polytropic efficiency can then be calculated with the mass average of the changes in
the enthalpy and aerodynamic work through the machine
P
_ 12
my
ηp ¼ P , (4.58)
_ t12
mΔh

where the different mass flows in the different compressor sections are taken into
account. This approach was suggested by Traupel (2000) for the case of feed heating
in steam turbines.
In addition to the fluid dynamic sources of loss in the flow path outlined in the
previous efficiency definitions, parasitic losses and mechanical losses are present, such
as the bearing losses, windage in the bearings, gearbox losses and the disc friction
losses of the impeller backplate and the shroud if present. If the mechanical power
absorbed by the bearings and gears is Pm and the power required internally by the
compressor is P, then the mechanical efficiency is defined as
P
ηm ¼ : (4.59)
P þ Pm
Consideration of the parasitic losses is given in Section 10.6.

4.7 Efficiency in Diabatic Processes

4.7.1 Diabatic Compression Processes

The efficiency of a diabatic compression process, that is, one with heat transfer, is also
defined as the ratio of the reversible work to the actual work expended to achieve the

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.7 Efficiency in Diabatic Processes 131

pressure rise. For a small compression step with heat transfer, the Gibbs equation, the
first law and the second law of thermodynamics can be written respectively as
dh ¼ vdp þ Tds
δwact þ δqact ¼ dh þ dke (4.60)
Tds ¼ δqact þ Tdsirrev :

where δwact and δqact are the actual work and heat transfer per unit mass, dke is the
change in kinetic energy and Tdsirrev is the entropy production caused by the irrever-
sibilities such as friction and heat transfer though a finite temperature difference. The
preceding equations can then be combined to give
δwact ¼ vdp þ dke þ Tdsirrev ¼ δwrev þ Tdsirrev : (4.61)

This demonstrates that, just as with adiabatic processes, the reversible work in diabatic
processes is calculated by the integration of vdp and dke along the reference reversible
compression path. Heat transfer to the compressor does not appear explicitly in this
equation. The amount of heat transfer does, however, influence the integration path
from state 1 to state 2 and so impacts on both the reversible work and the dissipation.
The overall actual work expended is still the shaft work and therefore the compressor
efficiency defined in (4.3) is valid for both diabatic and adiabatic processes. For a
diabatic process, the overall enthalpy change is a function of both the shaft work and
the overall heat transfer. For the case with heat transfer, the general form of the
compressor efficiency can therefore be written as
Ð2 Ð2
1 vdp þ ½ðc2  c1 Þ vdp þ ½ðc2 2  c1 2 Þ
2 2
ηc ¼ ¼ 1 : (4.62)
ws12 ðht1  ht2 Þ  q12

As in the case of adiabatic flows, by choosing different ideal reference processes for
the calculation of the reversible work, different efficiency definitions may be derived
for diabatic flows. However, a reversible diabatic flow (with no dissipation losses) is
not isentropic. Thus, there is no justification to use an isentropic process as a reference
for the ideal work required by a perfect diabatic compressor (Casey and Fesich, 2010).
The isentropic work between the initial and final pressures is essentially independent
of any heat transfer to the system. The application of polytropic and isothermal
efficiencies as a measure of the quality of compression in diabatic processes is
discussed in the following sections

4.7.2 The Effect of Heat Transfer in Turbocharger Compressors

Sirakov and Casey (2013) point out that in the testing of turbocharger compressors in
hot gas stands in close proximity to the turbine, there is a large heat exchange from the
turbine to the compressor. The efficiency of the compressor is usually determined
from the pressure ratio and temperature ratio using (4.16), but the measured tempera-
ture rise includes the heat transferred from the turbine as well as the work done on the
flow. The neglect of this heat transfer causes the apparent work input to be higher than

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 132 Efficiency Definitions for Compressors

in reality. Conventional performance maps then underestimate the efficiency of the

compressor stage by as much as 20 points at low speeds. A correction procedure for
this effect based on (4.62) can be used to convert performance maps obtained with
heat transfer to performance maps for adiabatic conditions. The effect of the heat
transfer on the apparent work input, and hence on the efficiency, is given by a
nondimensional parameter. This is defined as the difference between the apparent
work input and the actual work input expressed nondimensionally with the square of
the blade tip-speed
q Q_ Q_ 1 1
λact  λ ¼ ¼ ¼ ¼ kc , (4.63)
u22 ρt1 ϕt1 u32 D22 ρt1 a3t1 D22 ϕt1 M 3u2 ϕt1 M 3u2
where kc is a dimensionless coefficient that depends on the rate of heat transfer per unit
area into the compressor. This coefficient can be determined by experiment and,
depending on the design of the turbocharger housing, typical values are found to vary
widely between 0.002 < kc < 0.004. When tested in hot gas stands driven by the hot
turbocharger turbine, the apparent efficiencies on the low-speed characteristics are
affected strongly by the heat transfer in different turbocharger configurations. This
makes comparison of the aerodynamic quality of compressors measured in such test
stands very difficult. Equation (4.63) allows this effect to be estimated and the true
efficiency of the performance map to be calculated.

4.7.3 Diabatic Polytropic Efficiency

If the compression path between the initial and final states in a diabatic process is
approximated by a polytropic path, and the kinetic energy terms are negligible, then
the static–static polytropic efficiency is simply defined as the ratio of the reversible
work to the actual shaft work:
Ð2 Ð2
1 vdp 1 vdp
ηss ¼ ¼ : (4.64)
p
ws12 ðht1  ht2 Þ  q12
The aerodynamic work is still determined using (4.38), derived for adiabatic processes.
The differential form of the polytropic efficiency can be derived by considering that for
a polytropic process, the polytropic ratio, which is defined as the ratio of the actual
enthalpy rise to the polytropic enthalpy rise, remains constant. The polytropic ratio has
the properties of the polytropic efficiency in an adiabatic process and can be written as
dht ðn  1Þ=n
υ¼ ¼ , (4.65)
vt dpt ðγ  1Þ=γ
where n and γ are the polytropic and isentropic exponents. The reversible work and the
actual work in the process are given by δwrev ¼ vtdpt and δwact ¼ dht  δq such that
the polytropic efficiency of a diabatic process can be written as
vt dpt 1 1
ηttp ¼ ¼ ¼ , (4.66)
dht  δq ðdht =vt dpt Þ  ðδq=vt dpt Þ υ  ζ q

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 4.7 Efficiency in Diabatic Processes 133

where the heat transfer ratio, ζq, is defined as the ratio of the heat transfer to polytropic
work and is assumed to stay constant along the compression path.

4.7.4 Isothermal Efficiency

In intercooled multistage compressors, where heat is removed between intermedi-
ate compression stages (or during the compression, using external cooling jackets),
the required aerodynamic work is lowered by reducing the temperature at the inlet
of the individual stages. The minimum theoretical compression work is achieved
by a reversible isothermal process which can be viewed as a compression from p1
to p2i using an infinite number of intercoolers which remove all the energy from
the work input by heat transfer (see Figure 4.1). The end state in an isothermal
process can then be denoted as 2i in the h-s diagram, which has the temperature of
the initial state and the pressure of the final state of the actual process. In real
gases, the enthalpy varies with both pressure and temperature and therefore does
not remain constant in an isothermal process. The static–static isothermal
efficiency is defined as the ratio of the reversible work, integrated along the
isothermal compression path from state 1 to state 2i, to the compressor shaft work
and can be written as
Ð 2i Ð 2i
vdp 1 ðdh  TdsÞ ðh2i  h1 Þ  T 1 ðs2i  s1 Þ
ηss ¼ 1
¼ ¼ : (4.67)
i
ws12 ðht2  ht1 Þ  q12 ðht2  ht1 Þ  q12

For ideal gases, the enthalpy is a function of temperature only and therefore remains
constant along the isothermal path. Furthermore, the entropy at state 2i can be
determined from the Gibbs equation, leading to the conventional equation for the
isothermal efficiency for an ideal gas:

R ln ðp2 =p1 Þ
ηttt ¼ : (4.68)
ws12

Intercooling of compressors provides a reduction in the compression shaft work. The

efficiency given in (4.68) is, however, purely theoretical since it requires an infinite
number of intercoolers and takes no account of the pressure losses in the heat
exchangers. As a result of including the pressure losses in the heat exchangers, a
finite number of intercoolers is determined as the theoretical optimum (Simon, 1987);
Vadasz and Weiner, 1992). In a two-stage compression process with intercooling with
no pressure losses, the optimum intermediate pressure is given by p ¼ √(p1p2)
(Haywood, 1980). The optimum location of the intercoolers in terms of the pressure
ratio for the individual stages with multiple intercoolers is given by Vadasz and
Weiner (1992). In reality, the costs of the heat exchangers play a role as the effective-
ness is determined by the surface area, so economic considerations also need to be
taken into account, which leads to a smaller number of intercoolers than the
thermodynamic optimum.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006
 134 Efficiency Definitions for Compressors

4.8 Efficiency Definitions for Real Gases

The basic equations for the efficiency definitions given in this chapter are valid for
both real gases and ideal gases. For the isentropic and the isothermal efficiency, the
intermediate states are fully defined, so an integration of vdp with real gas equations to
obtain the aerodynamic work is possible. With real gases, a difficulty occurs with the
polytropic efficiency as the integration of vdp to calculate the aerodynamic work
requires knowledge of the intermediate states of the process, and these have to be
estimated. There is a large literature on the difficulty of calculating the polytropic head
by integrating vdp for real gases, and on the inaccuracies that this causes. This is
summarised in Section 3.6. Provided that the aerodynamic work is calculated with the
equations for real gases given there, the efficiency definitions in this chapter are valid.
The entropy definition of polytropic efficiency, (4.30), is sometimes considered to
be accurate for real gases; see, for example, Casey (2007). However, this is not correct
for real gases as its derivation makes use of ideal gas relationships. Its accuracy for
real gases can be increased by splitting the overall pressure rise into several steps with
one or more virtual intermediate pressures (Dubbel, 2001). This is shown in
Figure 4.11 with a single intermediate pressure. Using a large number of intermediate
pressures, the following equation can be used for the entropy-based polytropic
efficiency for real gases:
P P
ðΔsÞh ðΔsÞh
ηp ¼ ¼ P : (4.69)
ðΔsÞp Δs þ ðΔsÞh

poutlet
h 3
pintermediate
(D s) h 2
pinlet

(D s) h1

Ds

Figure 4.11 The entropy changes related to the polytropic efficiency with an intermediate
pressure.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 27 Aug 2021 at 14:04:36, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108241663.006