Analysis of Averaging Methods

for Nonuniform Total Pressure

Fields
A common objective in the analysis of turbomachinery components (nozzle guide vanes
(NGVs) or rotor blades, for example) is to calculate performance parameters, such as
total pressure or kinetic energy (KE) loss coefficients, from measurements in a nonuniform
flow-field. These performance parameters can be represented in a range of ways. For
example, line-averages used to compare performance between different radial sections of
a 3D component; plane-averages used to assess flow (perhaps loss coefficient) development

Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022

Daniel Burdett
Department of Engineering Science, between different axial planes; and fully mixed-out values used to determine the total loss
University of Oxford, associated with a component. In the literature, the weighting method used for line- and
Oxford OX2 0ES, UK plane-averaging (e.g., area, volume flow, mass flow, or entropy-flux) is sometimes regarded
e-mail: [email protected] as an unimportant issue. Indeed, many authors neglect to even state which weighting
method was used in their work. In certain low-speed test cases, or where measurements
Thomas Povey1 are made a long distance from the component, the nonuniformity in the flow will be rela-
Department of Engineering Science, tively small and the practical difference between different weighting methods may be neg-
University of Oxford, ligible. However, in high-speed applications or for measurements close to a component
Oxford OX2 0ES, UK trailing edge, this becomes increasingly unlikely. In this paper, we compare a range of
e-mail: [email protected] methods for calculating aerodynamic performance parameters—for example, the kinetic
energy loss coefficient—including plane-average methods with different weighting
schemes and several mixed-out methods. We analyze the sensitivities of the different
methods to the axial location of the measurement plane, the radial averaging range, and
the exit Mach number. We use high-fidelity experimental data taken in several axial
planes downstream of a cascade of engine parts: high-pressure (HP) turbine NGVs oper-
ating at transonic Mach number. The experimental data are complemented by computa-
tional fluid dynamics (CFD). We discuss the underlying physical mechanisms which give
rise to the observed sensitivities. The objective is to provide guidance on the accuracy of
each method in a relevant, practical application. [DOI: 10.1115/1.4053020]

Keywords: boundary layer development, computational fluid dynamics (CFD), fluid

dynamics and heat transfer phenomena in compressor and turbine components of gas
turbine engines, measurement techniques, turbomachinery blading design, wind turbine
aerodynamic design

Introduction (3) In a mixing process, should the flow: be mixed to an entirely

uniform static pressure defined at a single, swirl-averaged
The flow-field downstream of a turbomachine blade row is—to
radius [1]; be mixed at every radius to a circumferentially
first order—two-dimensional, but often with significant three-
uniform static pressure, producing a radial profile [2] not nec-
dimensional secondary features. For the purpose of basic
essarily satisfying a radial equilibrium condition; or be mixed
performance assessment and comparison, simplified performance
to a pre-specified radial static pressure condition defined by a
parameters are commonly used. Two of the most common perfor-
free or forced vortex (which does satisfy radial equilibrium)?
mance parameters are the total pressure loss coefficient, Y, and
the kinetic energy (KE) loss coefficient, ζ. The relative merits of
these two metrics are considered later. Whichever performance A tree of possible representations of a given performance metric is
parameter is preferred, however, there remain numerous questions represented by Fig. 1. For each performance parameter (e.g., Y or ζ)
about how to represent said parameter. In particular, the following this leads to 13 distinct methods of representing the variable. In this
questions arise paper, we compare and discuss these representations using both
experimental and computational data, in particular considering the
(1) Do we wish to represent the performance parameter as a local sensitivity to the weighting and averaging processes. These differ-
value (at a point in a plane), a line-averaged value (to repre- ent representations of a loss parameter, are, of course, suited to dif-
sent a circumferential or radial variation), a plane-averaged ferent analysis objectives. For example:
value (as a single numerical representation of the efficiency),
or a mixed-out value at a hypothetical downstream plane? • Local distributions allow identification of specific loss struc-
(2) In a line or plane-averaging process, should the integral be tures within a flow-field, and detailed local flow comparisons
weighted by area, volume-flux, mass-flux, or entropy-flux? between different experiments or between experiment and
computational fluid dynamics (CFD).
1
• Line-averages are useful for analyzing wake characteristics
Corresponding author.
Contributed by the Heat Transfer Division of ASME for publication in the JOURNAL
(depth and thickness) and mixing rates.
OF TURBOMACHINERY. Manuscript received December 1, 2020; final manuscript received • Plane-averages are useful for analyzing overall mixing rates
August 11, 2021; published online January 28, 2022. Assoc. Editor: Tom Verstraete. and as a simple method for analyzing total loss.

Journal of Turbomachinery Copyright © 2022 by ASME; reuse license CC-BY 4.0 MAY 2022, Vol. 144 / 051011-1
 Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022
Fig. 1 Tree of possible performance metric calculations

• Mixed-out values are perhaps the most robust method for high spatial resolution in several axial planes, alongside compli-
representing the total (plane-independent) loss associated mentary CFD data.
with a component.

In many experiments in the open literature where averaged quan- History of Total Pressure Performance Parameters,
tities are presented, a simple area-average is used, either because of
the ease of calculation or because of difficulty in measuring the
Averaging, and Mixing Methods
additional flow variables required for more sophisticated weighting It is generally accepted that averaging of total temperature should
methods. While simplicity is clearly desirable, it is important to be done using a mass-flux-weighted average. For a perfect gas (i.e.,
consider the sensitivity (to axial plane, Mach number, etc.) that h0 = cpT0), this conserves total enthalpy flux between the measured
the choice of weighting method might imply. This is particularly nonuniform and equivalent (hypothetical) uniform flows and is
important where comparing between different experiments, or required in order to satisfy the first law of thermodynamics [3–5].
between experiment and CFD. In contrast, there is neither complete consensus regarding the
In this paper, we examine the sensitivity of average loss represen- most appropriate performance parameter to represent total pressure
tations to weighting method for the case of line-averaged and loss, nor the best way to calculate an average of that parameter.
plane-averaged flows, and the sensitivity to mixing method for Some of this divergence may arise from misunderstanding about
the case of mixed-out flows. We explain how differences between best practice, while in other cases it could be argued that it is the
the methods arise (through examination of the nonuniform static result of reasonable compromises having been made (prioritizing
pressure field), consider the sensitivity of the results to exit Mach practicality over rigor).
number, and discuss the physical interpretation of the various repre-
sentations of a performance parameter. We use experimental data Review of Performance Parameters. Arguably the two most
from a transonic annular engine-parts facility, acquired at very commonly used total pressure performance parameters (see, for
example, Denton [6]) are the total pressure loss coefficient, Y, and
the KE loss coefficient, ζ. We define Y for a turbine NGV or
stage as
p01 − p02
Y= (1)
p01 − p2
where p01, p02, and p2 are the row inlet total pressure, row exit total
pressure, and row exit static pressure, respectively. The denomina-
tor is therefore the ideal row exit dynamic pressure. A weakness of
this variable is that in a compressible flow, the denominator is not
directly proportional to the exit kinetic energy (p0 − p ≠ 0.5ρv 2).
If the primary purpose of the nozzle is taken to be the conversion
of potential (pressure) energy to outgoing angular momentum
flux, the nonlinearity of definition (1) at compressible Mach
numbers makes this variable problematic. The KE loss coefficient
does not suffer from this problem because it can be related to the
energy conversion efficiency, η, at all Mach numbers. We define
η as

Actual exit KE v22 h02 − h2

Fig. 2 h-s chart showing the simplified turbine expansion η= = = (2)
process Ideal exit KE v22s h01 − h2s

051011-2 / Vol. 144, MAY 2022 Transactions of the ASME

where v2 and v2s are the actual (measured) and isentropic exit velo- Prasad [9] conducts a similar analysis to Main et al. [8], but
cities, respectively. For an adiabatic nozzle, h02 = h01 = h0. The sim- enforces a forced vortex—rather than free vortex—at the hypothet-
plified expansion process is shown in the h-s chart in Fig. 2. ical mixed-out plane. The loss coefficient calculated in this manner
Defining the corresponding KE loss coefficient as ζ = 1 − η, and is found to be insensitive to the location of the plane in which the
using the isentropic relation T2/T1 = (p2/p1)χ where χ = (γ − 1)/γ, the nonuniform data are taken. There are advantages and drawbacks
loss coefficient becomes of both the mixed-out conditions (free or forced vortex). A forced

χ vortex is the only possible swirling, annular flow that can be in com-
h2 p2 plete kinematic (all elements of the strain rate tensor are zero) and
1− 1−
h0 p02 thermal (flow supports no temperature gradients) equilibrium.
ζ=1− =1−
χ (3)
h2s p2 However, the resulting mixed-out flow has a radial gradient of
1− 1− total pressure, p02(r), with higher total pressure at larger radii.
h0 p01
This means it is not possible to define a reference isentropic forced-
In this paper, we primarily use ζ as the measure of NGV loss, and vortex condition (required to define a loss coefficient), in cases
reference Y only to compare the sensitivities of the two parameters. where the inlet total pressure, p01, is radially uniform. In contrast,
For completeness, we note here that in the incompressible case, a free vortex has a nonzero component of the strain rate tensor

Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022

where p02 − p2 = 0.5ρv22, ζ, and Y converge to the same numerical and hence is not entirely self-sustaining, but mixes the flow to a
value. uniform total pressure, p02. The free-vortex method therefore does
not suffer from the issue of defining a reference isentropic
condition.
Review of Averaging Methods. Pianko and Wazelt [3] pre-
sented a comprehensive review of the methods in common use
for averaging nonuniform internal flows in turbomachines. They
concluded that it is not possible to replace a real, nonuniform Definitions for Averaging and Mixing Methods
flow with an equivalent uniform (i.e., averaged) flow matching all
In this section, we define the calculation process for each of the
significant fluxes and parameters. The significant literature is in
possible performance metric representations in Fig. 1. The KE
agreement on this point [5,7].
loss coefficient is used in these examples.
Oates [4] discussed common methods for averaging nonuniform
total pressure. For the simple case of two jets of different total pres-
sure mixing adiabatically in a frictionless, constant-area duct, he
Local Loss Parameters. The KE loss coefficient can be pre-
showed that three different averaging methods produced average
sented as a function of r and θ as follows:
total pressure results with three different sensitivities to axial
mixing distance: mass-averaged total pressure decreased with

p2 (r, θ) χ
mixing distance (due to mixing out of secondary kinetic energy 1−
(SKE)); stream-thrust-averaged total pressure (in which the flow p (r, θ)
ζ(r, θ) = 1 −
02  (4)
is fully mixed to a uniform condition) was insensitive to mixing dis- p2 (r, θ) χ
tance; and continuity-averaged total pressure (calculated from a 1−
p01
measured duct area, mass flowrate, static pressure, and total
temperature by assuming the flow to be uniform over the whole where p2(r, θ) and p02(r, θ) are the static and total pressure distribu-
area—sometimes used in situations where the conditions of individ- tions in the plane, and χ = (γ − 1)/γ.
ual constituent flows are known, but direct traverse probe access to
the mixed flow is not possible) increased with mixing distance.
Cumpsty and Horlock [5] propose that the “correct” method for
averaging total pressure in a nonuniform field is that which ensures Line and Plane-Averaging Methods. There are three principal
that the property of interest in a particular application (e.g., work steps in a line-averaging or plane-averaging process:
input/output, thrust, capacity, etc.) has the same value in the aver- (i) Calculation of area-average local static pressure. The
aged flow and the real (nonuniform) flow. In particular, they average static pressure is defined from the net force acting
discuss some common applications in turbomachinery flows: non- on the reference area [5]
uniform flow entering a compressor or turbine (in which the work

input or output is the important parameter); nonuniform flow enter-
ing a propelling nozzle (in which the gross thrust is the important F = p dA =
pA (5)
parameter); and nonuniform flow entering a choked nozzle row
(in which the flow capacity is the important parameter). They also where
p is the area-average static pressure. That area-
note that, while there is no good physical basis for mass-averaging averaging is the appropriate method for averaging static
total pressure (flux-based averages such as the mass-flow-weighted pressure is generally accepted and is not debated in this
average have no physical meaning when applied to intensive rather paper. The plane-average of p2, defined over a specific
than extensive quantities), in many typical experimental situations radial (r1 ≤ r ≤ r2) and circumferential (θ1 ≤ θ ≤ θ2) range is
the absolute variation in total pressure is sufficiently small that
the difference between mass-averaging and more physically rigor- θ2 r2
θ1 r1 2
p (r, θ)rdrdθ
ous methods is small. p2,A = θ2 r2 (6)
θ 1 r1
rdrdθ
Review of Mixing Methods. The most commonly used methods
for calculating the fully mixed-out total pressure in a nonuniform The corresponding line-averages (e.g., at constant r or θ)
field are: mixing to a uniform static pressure defined at a single are calculated by integrating only in one dimension.
swirl-averaged radius (developed by Dzung [1]); uniform static (ii) Calculation of average local total pressure. Total pressure
pressure, proposed by Amecke [2]; and the free-vortex method of can be averaged using the chosen weighting method
Main et al. [8]. The history of these three methods is reviewed by (methods 2a–d in Fig. 1), by integrating the product of
Main et al. [8]. For the case of an annular NGV cascade, a local total pressure and the local value of the weighting
maximum disagreement in the evaluated efficiency between the variable, and dividing by the integral of the weighting
three methods of 0.12% was demonstrated, across a range of pres- variable. For each of area-weighting, volume-flow-
sure ratio conditions and downstream traverse plane locations. weighting, mass-flow-weighting, and entropy-flux-

Journal of Turbomachinery MAY 2022, Vol. 144 / 051011-3

 weighting, we get mixed-out plane. To close the equations, it is necessary to
θ2 r2 specify the static pressure distribution in the mixed-out flow.
θ1
p (r, θ)rdrdθ
r1 02 The way this is specified is the primary difference between dif-
p02,A =  θ 2  r2 (7) ferent methods. The mixed-out averaging processes can be
θ 1 r1
rdrdθ thought of as effectively including the residual SKE present in
θ2 r2 the measured nonuniform flow in the loss evaluation. The
θ1
p (r, θ)vx (r, θ)rdrdθ
r1 02 result is therefore less sensitive to the axial plane in which
p02,V =  θ 2  r2 (8) the data are measured. It could also be argued that the result
θ 1 r1 x
v (r, θ)rdrdθ
is more relevant in situations where the SKE cannot be usefully
where vx is the magnitude of the velocity component normal recovered by the downstream machinery (as in most practical
to the integration plane situations in compressors and turbines).
There are four mixed-out averaging methods in the literature:
θ2 r2
p02 (r, θ)ρ(r, θ)vx (r, θ)rdrdθ
p02,M = θ1 r1θ2 r2 (9) (1) Swirl-averaged mixing (method of Dzung [1]; see also
θ1 r1
ρ(r, θ)vx (r, θ)rdrdθ Ref. [3]). Average quantities of the mixed-out flow (includ-

Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022

ing total and static pressures) are computed at a single,
and swirl-averaged radius (that is, a nominal radius at which
⎧  

 ⎫ swirl velocity has its average value). This method is recom-
⎪ p02 (r, θ) γ T02 (r, θ)
dṁ⎪
θ 2 r2
⎪
⎨ θ 1 r1 ln − ln ⎪
⎬ mended for linear cascade flows, but in annular swirling
p02,S pref γ −1 T01 flows the implied mixed-out flow cannot satisfy radial
= exp  θ 2  r2
pref ⎪
⎪ dṁ ⎪
⎪ equilibrium.
⎩ θ 1 r1 ⎭
(2) Two-dimensional mixing (method of Amecke [2]; see also
(10) Ref. [10]). The flow is conceptually split up into a series of
thin, radial layers of 2D flow, which are individually
where T01 and T02(r, θ) are the mass-averaged inlet mixed out without interaction with neighboring layers. This
total temperature and the local distribution of total tempera- gives a radial distribution of mixed-out flow parameters
ture in the measurement plane, respectively, pref is (including total and static pressures). The radial static pres-
an arbitrary nonzero reference pressure, and dṁ = sure distribution would not necessarily satisfy a radial equi-
ρ(r, θ) vx (r, θ) r dr dθ represents a local element of mass librium condition due to the complete absence of radial
flowrate (normal to the plane). The entropy-flux-weighted mixing. An overall loss coefficient is evaluated from single
method is a line-averaging or plane-averaging method in total and static pressures obtained by integrating the radial
which we choose to set the total entropy-flux of the averaged distributions (usually the area-integral of static pressure and
flow to the same value as the measured (nonuniform) flow at the mass-flow-integral of total pressure).
the same location (e.g., in the same axial plane). This might (3) Free-vortex mixing (method of Main et al. [8]). The flow is
be considered desirable because the change in entropy-flux mixed to a free-vortex condition; that is, the circumferential
is the most direct measure of lost efficiency through the velocity component is constrained to be inversely propor-
vane row. The method is applicable only to local line or tional to radius. The mixed-out flow therefore has zero
plane-averaging processes and is not a mixed-out method: vorticity. The resultant mixed-out flow has uniform total
entropy-flux should not be conserved in a mixing process. pressure and a radial static pressure distribution
In the case of uniform total temperature (T02 (r, θ) = which, unlike the other methods, does satisfy a radial equilib-
T01 = T0 ), Eq. (10) simplifies to rium condition and is therefore closer to a physically realiz-
⎧ 
 ⎫ able condition. This method has the drawback that in
⎪ p02 (r, θ) experiments it is problematic to include coolant flows in
dṁ⎪
θ 2 r2
⎪
⎨ θ 1 r1 ln ⎪
⎬
p02,S pref the definition of the ideal exit flow (for the purpose of defin-
= exp  θ 2  r2 (11) ing a loss coefficient) unless the radial distribution of coolant
pref ⎪
⎪ dṁ ⎪
⎪
⎩ θ 1 r1 ⎭ input is known.
(4) Forced-vortex mixing (method of Prasad [9]). The flow is
That is, in the case of negligible total temperature varia- mixed to a forced-vortex condition; that is, the circumferen-
tion, entropy-flux-weighted averaging is the same as tial velocity component is constrained to be linearly propor-
mass-flow-averaging the natural logarithm of total pressure. tional to radius. The resultant mixed-out flow has nonzero
(iii) Calculation of average value of the performance parameter vorticity and radial profiles of both static and total pressure
based on area-averaged static pressure and weighted- (higher total pressure at higher radii). Unlike the free-vortex
average total pressure (whichever weighting method used). condition, this is a state of complete kinematic and thermal
The average KE loss coefficient is expressed as equilibrium (all shear stress components are zero). Like the
free-vortex method, the inclusion of coolant flows in the


p2,A χ loss coefficient definition is problematic.
1−

ζ = 1 − p

02 χ (12)
p2,A
1−
p01 Note on Order of Operations. In these calculations, it is the
where p02 is the weighted average total pressure, calculated underlying physical variables (pressure, temperature, mass, etc.)
by any one of the four methods outlined above. A similar which are averaged. The performance parameter is then defined
process is followed for the average total pressure loss coef- using those average quantities. The question could arise: should


ficient Y. we instead calculate the local distribution of a performance param-
eter, e.g., ζ (r,θ), then perform a weighted average of the parameter.
Although in practice—despite nonlinearity of the equations, e.g.,
Mixed-Out Averaging Methods. In mixed-out averaging (12)—the difference between the two process orders has minimal
methods, equations for conservation of mass, linear and impact on the result, there is greater physical basis (control
angular momentum, and energy are solved between the nonuni- volume conservation arguments) for averaging the physical vari-
form flow in the measurement plane, and a hypothetical fully ables and this is therefore strongly preferred.

051011-4 / Vol. 144, MAY 2022 Transactions of the ASME

 Fig. 4 Locations of the three experimental axial traverse planes

Traverse measurements were made downstream of the NGV

cascade using an automated radial-circumferential traverse mecha-

Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022

nism. Measurements were performed in three axial planes located
0.25, 0.50, and 0.75 axial chord lengths downstream of the
cascade trailing edge plane. These axial planes are indicated in
Fig. 4. Traverse measurements covered a circumferential range of
Fig. 3 Simplified diagram of an axisymmetric cross section 2.5 vane pitches. The traverse system was clocked in the circumfer-
through the ECAT facility working section
ential direction between the axial planes so that wakes from the
same vanes were measured at each axial position. A schematic is
shown in Fig. 4. The radial traverse range was approximately
Table 1 Summary of nominal experimental test conditions 95% of vane span. In this range approximately 60 discrete circum-
ferential passes of the traverse probe were performed at fixed radius.
Variable Nominal value
The average radial step height was approximately 1.4 mm. A typical
Inlet total pressure, p01 2.00 bar complete area-survey traverse path is shown in Fig. 5. The complete
Inlet total temperature, T01 290 K path was made up of six separate experiments, represented by
Exit static pressure, p2 1.10 bar blocks in the figure. Adjacent blocks overlapped by at least one
Mass flowrate, ṁ 14.0 kg/s probe-pass to ensure consistency. Dashed black lines show the
Inlet turbulence intensity, Tu ∼8% approximate locations of vane wakes.
Coolant-to-mainstream pressure ratio, p0c/p01 1.025 The traverse probe used in this study was a five-hole pneumatic
Vane pressure ratio, p2/p01 0.550 pressure probe with tip diameter 2.8 mm. The probe was calibrated
Vane exit Mach number, M2 0.965 over a range of flow Mach number 0.3 ≤ M ≤ 1.4, and a range of
Reynolds number, Re 1.3 × 106
incident flow angles (in two axes) of ±40 deg. The traverse speed
was set based on the estimated dynamic response of the probe-
transducer system, such that the effective circumferential resolution
was equivalent to half of the probe tip diameter (1.4 mm). This,
in combination with the traverse pattern, gives an effective resolu-
Experimental Measurement Process tion of approximately 1.4 mm in both the radial and circumferential
Experimental traverse measurements were performed in the directions.
Engine Component Aerothermal (ECAT) facility at the University
of Oxford [11]. The ECAT facility is a high technology readiness
level short-duration (approximately 60 s run time), blowdown facil- Numerical Method
ity currently built as an annular cascade of NGVs. A simplified axi-
3D Reynolds-averaged Navier–Stokes CFD simulations were
symmetric schematic diagram is shown in Fig. 3. The NGVs are
performed using the commercial solver ANSYS CFX. The shear
fully cooled parts from a modern civil engine, operating at condi-
stress transport (SST) k-ω model was used for turbulence closure.
tions (Mach number, Reynolds number, coolant-to-mainstream
The simulated geometry was an annular sector containing two
pressure ratio) nondimensionally matched to the nominal engine
vane pitches with rotational periodicity. The geometry is shown
operating point. The nominal operating conditions in ECAT tests
in Fig. 6(a). The vane geometries were taken from the design
are summarized in Table 1. The quoted Reynolds number is
intent CAD drawing of the component tested in experiments,
based on the vane tangential chord length and vane exit flow condi-
tions. Inlet conditions of total pressure and temperature are mea-
sured at rakes located three axial chord lengths upstream of the
cascade leading edge plane. A pressure regulator upstream of
the working section maintains the inlet total pressure at steady
state for the duration of the test. Vane exit static pressure is mea-
sured at a large number (approximately 100) of surface tappings
located on the vane exit platforms at hub and case. The ratio of
vane exit-static to inlet-total pressures defines the operating condi-
tion of the facility. Further details of these measurements and the
associated uncertainties can be found in Refs. [12,13].
The vane cooling system includes approximately ten rows of
surface films and coolant ejection from a trailing edge (TE) slot.
Coolant is fed from cavities at both hub and case (see Fig. 3).
The coolant feed mass flowrate is carefully controlled to achieve
the target coolant-to-mainstream pressure ratio. The ECAT
facility has previously been used to perform high-accuracy mea- Fig. 5 Example traverse blocking arrangement for a single area-
surements of vane capacity [12–14], aerodynamic loss, and metal survey. Six separate tests are represented by the six blocks.
effectiveness [11]. Dashed lines show the wake locations.

Journal of Turbomachinery MAY 2022, Vol. 144 / 051011-5

 introduce the nonuniform experimental and CFD flow fields.
Figure 7 shows local distributions of KE loss coefficient, ζ
(Eq. (4)), in three different axial planes, from experimental and
numerical data, respectively.
Most importantly for our purpose, we have high spatial resolution
in the experimental measurements, and similar general trends
between CFD and experiment with well resolved wakes and sec-
ondary flow features. The under-predicted mixing rates in CFD
simulations are a well-known problem and this is reflected in
higher values of ζ in the wakes with respect to experiment (at a par-
ticular axial plane), and narrower wake profiles. The absolute KE
(or total pressure) deficit associated with film cooling is also over-
estimated by the source terms model. Improved agreement with
experiment would likely be achieved by fully resolving the geome-
try of all film cooling holes in the CFD model.

Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022

The radial pressure gradient, a characteristic feature of annular
cascade flows, promotes migration of low-momentum flow
toward the hub and away from the case. This generates intense sec-
ondary loss cores in the wake near the hub—also exacerbated by
higher Mach numbers in this region—and lower intensity of sec-
ondary flows near the case. The radial profile of total pressure
loss in the wake center over the vane span is also relatively well
matched. The wake loss profile is broken up into a number of dis-
tinct sections separated by local minima in between. This effect is
the result of internal cooling features within the trailing edge
coolant slot. The wake regions decay in depth and broaden in
width between the axial planes. The nonuniformity in the radial
direction also gradually decays.
Although we describe the flow features for reference in later anal-
ysis, our primary purpose here is not good agreement between CFD
and experiment, but rather a comprehensive dataset that shows
typical differences between such results.
Fig. 6 (a) CFD geometry and (b) midspan mesh cross section

defeatured to remove film cooling holes from the vane surfaces. The Plane-Average Weighting Functions. We now compare
domain was discretized with an Octree-based mesh using Boxer- area-weighted, volume-flow-weighted, and mass-flow-weighted
Mesh. A cross section through the mesh at midspan is shown in methods for plane-averaging total pressure. The weighting func-
Fig. 6(b). The mesh was refined close to the vane surfaces, in the tions are the multipliers for total pressure in the numerators of
vicinity of the trailing edges, and in the paths of the wakes down- Eqs. (7)–(9). We do not treat the entropy-flux-average separately
stream of the cascade. The final mesh size was approximately 57 because in the isothermal simplification (Eq. (11)) the method
million cells. reduces to weighting the natural logarithm of total pressure using
Boundary conditions were set to match the operating conditions the mass-flow-weighting function. The weighting functions for
achieved in experiments (Table 1). Uniform total pressure and tem- the axial plane 1 are shown in Fig. 8, for both experimental data
perature were imposed at the inlet plane. A target value of area- and CFD, normalized with respect to the maximum value in each
average static pressure was set on the exit boundary. The radial dis- case. Horizontal dashed lines indicate the radial limits used for
tribution of exit static pressure was additionally required to satisfy a the evaluation of the profile KE loss coefficient. We now consider
radial equilibrium condition, while the circumferential static pres- the weighting functions in turn. To most clearly demonstrate the
sure distribution at any radial height was unconstrained. Static pres- trends, the underlying (high-resolution) data are re-sampled onto
sures were evaluated at the locations of the experimental pressure a uniform grid in r and θ. The results presented are independent
tappings on the hub and case vane exit platforms, and an iterative of this processing step, and raw data following the traverse
process (tuning the target value for area-average static pressure at pattern of Fig. 5 could equally be used (with associated radial dis-
the exit boundary) was used to match the exit-platform tortions in the weighting functions).
static-pressure-to-inlet-total-pressure ratio to the experimental The area-weighting functions (frames (a) in Fig. 8), have—trivi-
value. ally—a gradient in the radial direction on account of the uniform
Film cooling flows were modeled using the source term method. grid of data in r and θ.
A coolant mass flowrate (specified as a fraction of the mainstream The volume-flow-weighting functions (frames (b) in Fig. 8) have
mass flowrate) was applied at each film exit location. Details of comparatively low weighting in the wake and film-mixing regions,
the model for calculating mass flowrates for individual films can associated with the low momentum in these areas. A strong radial
be found in Ref. [14]. The trailing edge coolant flow was set at pressure gradient from case to hub leads to slightly higher weight-
an inlet boundary some distance (approximately eight slot widths) ing at lower radii. The same effect can be seen with the cross-
inside the trailing edge slot (to ensure proper flow development passage pressure gradient, with highest weighting near the vane
before issuing into the mainstream). A mass flow boundary condi- suction side (SS), and lowest near the pressure side (PS).
tion was used. Further details of the trailing edge slot modeling can The mass-flow-weighting functions (frames (c) in Fig. 8) are very
be found in Ref. [14]. similar to the volume-flow-weighting functions, but with small
deviations associated with density variation.
To compare the weighting functions quantitatively, the values at
Comparison of Averaging Methods midspan in planes 1, 2, and 3 are plotted as line functions in Fig. 9.
In this section, we compare the results of the various averaging The approximate locations of the wake centers are marked with
methods using both experimental and numerical data. We first dashed lines. Each trend is normalized with respect to its own

051011-6 / Vol. 144, MAY 2022 Transactions of the ASME

 Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022
Fig. 7 Experimentally measured and CFD-predicted local distributions of kinetic energy loss coefficient, ζ, in the axial planes:
(a)–(b) x/Cx = 0.25, (c)–(d) x/Cx = 0.50, and (e)–(f) x/Cx = 0.75

maximum value. In all cases, the area-weighting functions are cons- experiment, and by only 2.9 deg in CFD. The under-prediction of
tant at unity. whirl angle to the SS of the wake by CFD is thought to arise due
We now compare the volume flow and mass-flow-weighting func- to an exaggerated shock-boundary layer interaction, which causes
tions. There is only a very slight difference moving between the SS boundary layer to separate upstream of the trailing edge.
volume-flow-weighting and mass-flow-weighting, which results in The larger whirl angle variation in the experimental data leads to
biasing the average further in favor of the freestream regions (e.g., greater circumferential variation in the axial velocity component,
0.4 ≤ θ ≤ 0.6) and against the wake regions (e.g., −0.3 ≤ θ ≤ 0.1). and therefore greater variation in the local volume flowrate and
The general shape of the trends is very similar between CFD and mass flowrate normal to the axial plane. This explains the deeper
experiment, but with significantly larger variation (from unity) for troughs in the normalized experimental weighting functions—see
the experimental data compared with CFD. In the wake regions the Fig. 9—which align with peaks (and associated lower local axial
experimental weighting functions have much deeper troughs than velocity) in the whirl angle distributions of Fig. 10.
those for CFD (minimum values of normalized mass flow-weighting To understand the difference in whirl angle distribution between
differ between experiment and CFD by 36.9%, 15.0%, and 16.2% in the three planes, we now consider the total and static pressure fields
planes 1, 2, and 3, respectively), and the troughs are shifted toward in the exit flow. CFD-predicted midspan profiles of total and static
the SS. This is surprising: given the more mixed-out local loss distri- pressure are presented in Figs. 11(a) and 11(b), respectively. Planes
butions observed in the experimental data compared with CFD in the 1, 2, and 3 are marked in the figures by dashed black lines.
equivalent plane (Fig. 7), we might have expected the opposite to be Looking first at the total pressure field (Fig. 11(a)), there are char-
true. The differences can be explained by differences in exit whirl acteristic wakes which broaden in width and reduce in depth with
angle profiles, and total and static pressure fields between experiment distance in the streamwise direction. Even at plane 3, the wakes
and CFD. A comparison of whirl angle distributions at midspan remain well-defined.
between experiment and CFD is shown in Fig. 10, while static and The static pressure (Fig. 11(b)) reduces through the covered part
total pressure fields are presented in Fig. 11. of the vane passage as the flow is accelerated. In the downstream
Consider first the midspan whirl angle distributions in Fig. 10, flow between planes 1 and 3 (in the axial direction), the static pres-
where whirl angle, β, is defined as the angle between the local velo- sure is relatively uniform over most of the domain, but with a sig-
city vector and the axial direction, in the axial-circumferential nificantly elevated region of pressure directly axially downstream of
plane. The profiles differ significantly between the three planes, each trailing edge.
becoming generally more mixed out with successive movements There are two interesting consequences of the interaction of the
downstream. At a particular plane, the variation in angle in the total and static pressure fields, which we now examine.
experimental data is greater than the CFD predictions. For Firstly, we consider the development with axial distance of the
example, in plane 1, β varies circumferentially by 8.8 deg in the whirl angle profiles shown in Fig. 10. In Fig. 12(a), we plot the

Journal of Turbomachinery MAY 2022, Vol. 144 / 051011-7

 Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022
Fig. 8 Experimental and CFD weighting functions in axial plane 1 for (a) area-weighting, (b) volume-flow-weighting, and
(c) mass-flow-weighting

normalized static pressure along the wake centerline (red line)—i.e., They also explain the apparently contradictory result that the exper-
following one of the dashed white lines in Fig. 11(b)—and in the iment has greater variation in weighting functions across the plane
center of the freestream region (blue line). The trends are periodic than CFD (Fig. 9). While CFD is under-mixed with respect to
and approximately in anti-phase with each other. This effect the experiment, and therefore has narrower and deeper wakes, the
arises because of the periodic intersection of the wake centerlines more extreme variation in axial velocity in experiment—caused
(and the freestream centerlines) with the elevated static pressure by greater whirl angle variation—is the dominant factor in deter-
regions extending axially downstream of each TE. Analysis of the mining the weighting function distributions.
entire flow-field is complicated by the fact that the flow is transonic, The second interesting effect is that the streamwise pressure gra-
but in the region between planes 1 and 3, the flow is almost entirely dient, which causes diffusion and acceleration of the wakes (red
high-subsonic (predominantly in the range 0.75 < M < 1.00, but with line in Fig. 12(a)), significantly influences the decay process of
a small region up to M = 1.05 at plane 1). the wake profiles. This in turn affects the sensitivities of different
In subsonic flow, as streamlines enter the elevated pressure weighting methods to axial plane. We characterise the wakes in
region (which extends in the downstream direction, almost perpen- terms of the maximum value of ζ, which we refer to as ζ̂, and
dicular to the streamlines) they are diffused, and turned toward the the width of the wake (over which ζ exceeds half of ζ̂), which
axial. As streamlines leave the high-pressure region they are accel- we refer to as w. These two parameters are plotted in Figs.
erated and turned away from the axial, toward higher whirl angle. 12(a) and 12(b), respectively, extracted from both experimental
The alignment between the wakes and the high-pressure region and CFD data at midspan.
differs between the three axial planes. In planes 1 and 3, streamlines Peak loss (Fig. 12(b)) follows a quasi-exponential decay in both
approximately in the center of the freestream region pass through experiment and CFD. An offset between the two data arises from
the high-pressure region and are thus turned toward the axial. In the familiar under-prediction of mixing by CFD, leading to higher
plane 2, streamlines approximately in the center of the wake peak loss along the wake centerline.
region pass through the high-pressure region. Therefore between Axial trends in the wake width, w (Fig. 12(c)), show a highly vari-
planes 1 and 2, the effect is manifest as a relative increase in able rate of increase. There is a reasonable correlation between
whirl angle in the freestream regions, and a relative decrease in regions of streamwise acceleration identified in Fig. 12(a) (0.00 <
whirl angle in the wake regions (compare Figs. 10(a) and 10(b)), x/Cx < 0.15, 0.45 < x/Cx < 0.60, 0.95 < x/Cx < 1.00) and regions of
and vice versa between planes 2 and 3. This effect, in combination essentially constant w. That is, acceleration due to the local stream-
with the gradual mixing out of circumferential nonuniformity, wise pressure gradient thins the wake (disproportional influence on
accounts for the complex development with axial distance of the low-momentum fluid) but the effect is in competition with mixing
whirl angle distributions in Fig. 10. processes which always act to thicken the wake. The corollary is
These distributions are important because they significantly that diffusion acts to thicken the wake, and in regions where dp/dz
impact the weighting functions for different averaging methods. is positive (where z is streamwise distance), there is significant

051011-8 / Vol. 144, MAY 2022 Transactions of the ASME

 Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022
Fig. 9 Normalized midspan weighting functions for each of the Fig. 10 Circumferential distributions of whirl angle at midspan
three in-plane-averaging methods for experimental and CFD data for experiment and CFD in (a) Plane 1, (b) Plane 2, and
at (a) Plane 1, (b) Plane 2, and (c) Plane 3. Dashed lines show (c) Plane 3. Dashed lines show approximate wake center loca-
approximate wake center locations in each plane. tions in each plane.

Comparison of Plane-Average and Mixed-Out Methods.

growth in w. In these regions, the effects of diffusion and mixing are Plane-average and mixed-out-average results for KE loss coeffi-
additive. A sympathetic impact on the gradient of peak loss might be cient, ζ, are presented for each of the methods discussed, using
expected (increase in decay rate where the wake is accelerated, and data from a series of 2D axial planes. The relationships between
vice versa), but is hard to discern in the quasi-exponential decay. the trends for total pressure loss coefficient, Y, are almost identical
These results are in accord with Ref. [15], a study of wake flows to those for ζ—though they differ in absolute level—and are not
exposed to different streamwise pressure gradients. presented.

Sensitivity to Radial Averaging Range. To characterize the

profile loss, it is necessary to understand the sensitivity to radial
averaging range. Mass-flow-weighted ζ is plotted in Fig. 13 as a
function of the radial averaging range (expressed as a percentage
of vane span, centered on midspan) for both experiment and
CFD, and in three axial planes.
For radial averaging ranges between 2 and 10% of span, the
returned value of ζ is sensitive to localized spanwise nonuniformi-
ties caused by, for example, individual cooling flows or TE slot
features. For radial averaging ranges between 10% and approxi-
mately 80% of span, the value of ζ is insensitive to local variation,
but slowly varying due to nonlinear radial variation in loss coeffi-
cient. Near the upper end of this range—say, around 60% of span
—may be taken as a sensible radial range for characterizing profile
loss. For radial averaging ranges between 80 and 100% of span,
the endwall loss becomes more significant, resulting in significant
increases in ζ . The depth and breadth of the endwall boundary
layers increases with increasing x/Cx. In evaluating ζ for the
entire plane (both profile and endwall losses), it is possible to
crudely infer the sensitivity to the closeness of approach to the
wall (for experimental data) from the near-wall gradient of the Fig. 11 CFD-predicted midspan contours of (a) normalized total
trends in Fig. 13. pressure and (b) normalized static pressure

Journal of Turbomachinery MAY 2022, Vol. 144 / 051011-9

 Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022
Fig. 14 Axial trends in
ζ based on average total pressure
according to a number of methods, for experimentally measured
and CFD-predicted flows

(i) There is generally good agreement between the experimen-

tal and CFD data, both in terms of absolute level (average
difference of +10.0% across all data) and in terms of
trends, where the experimental data are in good accord
with low-resolution sampling of the corresponding numeri-
cal trends. The relative ranking of the in-plane-averaging
methods (area-average, volume-flow-average, mass-flow-
average, and entropy-flux-average) between experiment
and CFD in each of planes 1–3 is similar. In both the exper-
Fig. 12 Axial trends at midspan of (a) static pressure along the
wake centerline and in the freestream, (b) peak value of ζ in the imental and CFD data, the area-average is notably higher
wake center, and (c) wake width than the other three in-plane methods in plane 1, but in
closer agreement in planes 2 and 3.
(ii) The differences in absolute loss between experiment and
Average values of ζ are presented in Fig. 14 as functions of axial CFD can generally be attributed to a combination of the
location for CFD (16 evenly spaced planes in the range 0 ≤ x/Cx ≤ CFD being under-mixed; CFD over-predicting the loss asso-
1) and experiment (traverse planes 1–3 at x/Cx = 0.25, 0.50, and ciated with surface film cooling; CFD incorrectly predicting
0.75, respectively). We first compare the trends between experiment the complex TE shock structure and hence the losses asso-
and CFD, and conclude the following: ciated with shock-boundary layer interaction on the vane
suction surface; and inadequate turbulence modeling in the
CFD. The differences in the deltas between different averag-
ing methods between experiment and CFD can generally be
attributed to the under-prediction of mixing by the CFD. For
example, the differences between area-averaged and mass
flow-averaged
ζ in plane 1 were 10.2% and 18.2% in exper-
iment and CFD, respectively, being smaller in the experi-
mental case because the flow is more fully mixed out in
the equivalent axial plane, with less remaining nonunifor-
mity in the wake profile.
(iii) The average discrepancy in the value of mixed-out KE
loss coefficient,
ζ, between CFD and experiment (4.3%) is
significantly smaller than that in mass-flow-weighted
plane-averaged
ζ (11.7%), when averaged across all data.
On the basis that the CFD is under-mixed with respect to
the experiment, this result is initially surprising: one
would expect the SKE conversion to loss (difference
between in-plane and mixed-out calculations) to be greater
for the CFD than the experiment. A possible explanation
is provided by Fig. 10, in which we see that the whirl
angle variation in the experimental data is great than the
CFD data. This provides the experimental data with an addi-
tional mechanism for loss generation with axial distance,
and may be part of the explanation for this apparently sur-
Fig. 13 Mass-flow-weighted average ζ as a function of radial prising result. Because the agreement is better for mixed-out
averaging range for experiment and CFD, in three axial planes

051011-10 / Vol. 144, MAY 2022 Transactions of the ASME

 coefficients, where absolute comparisons are to be made sensitivity to axial plane. Between planes 1 and 3 (0.25 ≤
between experiment and CFD, and where inaccurate predic- x/Cx ≤ 0.75) the average increases in mixed-out
ζ in CFD
tion of mixing rate is expected (most methods), mixed-out and experimental data were +1.8% and +3.1%, respec-
comparisons should be used. tively. Looking more closely at the CFD trend, we see
that 85% of the rise in
ζ occurs within the range 0.44 <
Comparing the different averaging methods, we draw the follow- x/Cx < 0.56, while all of the rise in experiment occurs down-
ing key conclusions: stream of plane 2. That is, the wakes passing through an
adverse streamwise pressure gradient—i.e., positive dp/dz,
(i) Trends of in-plane KE loss coefficient,
ζ, based on
where z is streamwise distance (see Fig. 12(a))—causes an
volume-flow-averaged, mass-flow-averaged, and entropy-
increase even in the KE loss evaluated from the mixed-out
flux-averaged total pressure are for the most part relatively
methods. An adverse pressure gradient increases the velo-
similar to each other: well-behaved, and slowly and mono-
city gradients in the free shear layers at the wake edges
tonically increasing with x/Cx due to the gradual conversion
(see discussion related to Fig. 12), and so increases the
of SKE to total pressure loss. The average (across
entropy generation associated with mixing out of the wake
volume-flow-weighted, mass-flow-weighted, and entropy-
profile (rate of entropy generation is proportional to dv/dy,
flux-weighted trends) change in
ζ between x/Cx = 0.00 and

Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022

where y is streamwise-normal distance). Conversely, a
x/Cx = 1.00 was +23.4% of the latter value. This sensitivity favorable streamwise pressure gradient (negative dp/dz,
should be considered when choosing an axial traverse plane where z is streamwise distance) causes a slight reduction
for flow-averaging or comparing between systems. All of in the mixed-out loss. These effects are very small at the
the in-plane methods (including area-weighted average) operating conditions used in these tests.
appear to reach a stable value by x/Cx = 1.00 and are in
agreement in terms of absolute value (to within 2.0%).
The absolute convergence of
ζ based on the four in-plane
total pressure weighting methods is expected, because the Effect of Mach Number. In this section, we use CFD to con-
variations in the underlying flow-field which give rise to sider the impact of vane exit Mach number, M2, on the applicability
the differences between the weighting functions for the dif- of the general conclusions drawn above. We consider five distinct
ferent methods decay rapidly with axial distance (see Fig. 9 M2 conditions in the range 0.63 ≤ M2 ≤ 1.11, including the target
and associated analysis). Mach number (M2 = 0.97). Absolute values of plane-averaged and
(ii) The in-plane KE loss coefficient,
ζ, based on area-averaged mixed-out
ζ computed from data at x/Cx = 1.00 are plotted as func-
total pressure is highly sensitive to axial plane, departing tions of M2 in Fig. 15. The data are normalized with respect to the
from the other in-plane methods by up to +72.2% (at x/ mean of the four plane-average methods at M2 = 0.97.
Cx = 0.00) and varying across the range 0.00 < x/Cx < 1.00 We established previously that at the nominal value M2 = 0.97,
by between −5.7% and +31.9% of the settled value (at x/ the difference between the four plane-average methods was very
Cx = 1.00). There are numerous problems with the account- small at x/Cx = 1.00. Figure 15 shows that this result is robust
ing in this method. For example, the significant over- across a wide range of M2. In terms of absolute level, all of the
weighting of low-momentum wake regions with respect to mixed-out methods give higher loss coefficient than the in-plane
other methods (which produces the large peak at x/Cx = methods. This is due to the mixing out of SKE in the mixed-out
0.00 where the wake is deepest) and the strong sensitivity methods. The Dzung and free-vortex methods agree almost
to local fluctuations in the wake characteristics (Fig. 12). exactly. The absolute levels of all trends of
ζ fall approximately lin-
The simplicity of area-weighted averaging is attractive, early with M2 in the range 0.63 ≤ M2 ≤ 0.97. We note that this trend
but very high sensitivity to axial plane makes it an extremely is specific to average KE loss coefficient,
ζ. The equivalent trends
poor choice for comparisons between CFD and experiment, (with Mach number) for other performance parameters, e.g.,
or between different experiments. In the range 0.00 ≤ x/Cx ≤ average total pressure loss coefficient, will be different due to
0.50, the method should be regarded as unacceptably sensi- their different Mach number sensitivity.
tive to axial location.
(iii) Volume-flow-weighted averaging of total pressure is more
sensitive to axial plane than the mass-flow-average and
entropy-flux-average methods, but much less sensitive
than the area-average method. The maximum deviation of
the volume-flow-average
ζ from the mass-flow-average
ζ
was +8.2%, at x/Cx = 0.19. Volume-flow-averaging total
pressure therefore offers a compromise between achieving
a relatively low sensitivity to axial location and practical
measurement requirements (does not require measurement
of local density variation which may be problematic in
some experimental environments).
(iv) The Dzung and free-vortex mixed-out methods give essen-
tially indistinguishable results for both experimental and
CFD data, in accord with the results of Main et al. [8].
The forced-vortex method gives a consistently higher loss
with respect to all of the other methods, by an average of
+5.8% across all of the data. This difference is attributed
to the additional entropy generation associated with produc-
ing the required linear radial total pressure distribution. The
Amecke method gives the highest loss coefficient of all of
the mixed-out methods: and average of +8.8% greater
than the Dzung method. The cause of this higher loss is
not fully understood. Fig. 15 CFD-predicted plane-average and mixed-out KE loss
(v) All of the mixed-out methods reference a numerically coefficient,
ζ, at x/Cx = 1.00, as functions of vane exit Mach
mixed-out flow so—as expected—there is relatively little number

Journal of Turbomachinery MAY 2022, Vol. 144 / 051011-11

 In the range 0.97 ≤ M2 ≤ 1.11, plane-average
ζ still falls with M2, density gradients in the flow at higher Mach numbers also cause a
but at a slightly lower rate than at lower M2. In contrast, mixed-out
ζ slightly greater difference between the volume-flow-weighted and
increases slightly with increasing M2. This suggests a rise in mixing mass-flow-weighted methods.
loss at higher M2. The additional loss is attributed to the presence of
a SS trailing edge shock at M2 = 1.11. There is a small amount of
“direct” loss across the shock (total pressure immediately down- Summary and Conclusions
stream of the shock is slightly lower than that immediately
In this paper, we have compared a range of methods for calcu-
upstream)—this is the cause of the very slight shallowing of the gra-
lating aerodynamic performance parameters—e.g., KE loss
dient of the in-plane trends in Fig. 15. In poorly resolved CFD, there
coefficient—including plane-average methods with various weight-
is smearing-out of the shock, producing an adverse static pressure
ing schemes (area, volume flow, mass flow, entropy-flux) and
gradient, which is capable of increasing mixing loss by increasing
mixed-out methods in which the flow is artificially mixed to a
the rate of shear in regions of velocity gradient. This is a possible
more uniform state. We do this for the case of a transonic high-
cause of the rise in mixed-out KE loss coefficient between 0.97 ≤
pressure (HP) NGV using high-fidelity experimental data from a
M2 ≤ 1.11 in Fig. 15.
transonic engine-parts facility, and using complementary CFD.
We now consider the impact of Mach number on the variation of

ζ with axial distance. To do this, we normalize the trends (of the We quantify the sensitivity of the various methods to the axial loca-

Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022

tion of the averaging plane, radial averaging range, and exit Mach
type presented in Fig. 14) by the average value of the four
number. The underlying physical mechanisms are discussed. The
in-plane methods at x/Cx = 1.00. The resulting trends with axial dis-
purpose is to provide guidance for a common practical situation
tance are shown in Fig. 16 for the two Mach number cases M2 =
on the accuracy of each method when used for comparisons
0.76 and M2 = 1.11. These data can be compared with those for
between two or more experiments, or between experiment and CFD.
the design Mach number (M2 = 0.97) shown in Fig. 14. The
The key conclusions of this study are as follows:
primary conclusion is that, for a given averaging method, the varia-
tion of
ζ with axial distance is relatively independent of Mach (1) The weighting functions for area-weighted, volume-
number. This result is true for all averaging methods. That is, the flow-weighted, and mass-flow-weighted plane-averaging
conclusions drawn from the data of Fig. 14 remain valid across methods are strongly influenced by static pressure gradients
the entire range of typical operating Mach numbers. in the flow. Diffusion and acceleration of the flow cause
For completeness, we note an interesting second-order influence. development of the whirl angle distributions between axial
The axial variation of both the area-weighted and volume-flow- planes and therefore cause changes in the volume flow
weighted in-plane methods, and mixed-out methods, increases as and mass flow distributions at different planes. The sensitiv-
Mach number is increased. This likely arises because of larger ities reported in this paper will be typical of transonic
streamwise static pressure gradients (both adverse and favorable) turbine cascades, but—more generally—the sensitivity to
in the flow-field, which amplify the changes in velocity gradient static pressure gradient would be expected to be
(and therefore changes in entropy generation rate) in the shear application-specific.
layers at the edges of the wakes. The magnitude of the effect dis- (2) The plane-averaged KE loss coefficient based on an area-
cussed in the context of Fig. 12 is therefore increased. Larger averaged total pressure is highly sensitive to axial plane,
showing variations over the range 0.0 ≤ x/Cx ≤ 1.0 of up to
−5.7% and +31.9% of the settled value (at x/Cx = 1.0). The
axial sensitivity is considered to be unacceptable for mea-
surements made within 0.5 Cx of the trailing edge. Despite
the simplicity of this method (which makes it a common
choice in the literature), the results show that area-averaging
is an extremely poor choice for comparing CFD and experi-
ment or comparing different experiments. This conclusion is
likely to be valid in most applications with significant wake
flow (characterized by low total pressure but also low volume
flowrate).
(3) The plane-average KE loss coefficients based on mass-flow-
averaged and entropy-flux-averaged total pressure are very
closely matched and have the characteristic of slowly and
monotonically increasing with axial distance due to the
gradual conversion of SKE to total pressure loss. Plane-
average KE loss coefficient based on volume-flow-averaged
total pressure is closely matched to the mass-flow-averaged
and entropy-flux-averaged methods in terms of both absolute
values and axial sensitivity, but shows small differences
close to the vane trailing edge due to local density variation.
Volume-flow-averaging could be considered to offer a com-
promise between low axial sensitivity, and practical mea-
surement requirements (does not require measurement of
local density variation).
(4) As expected, mixed-out KE loss coefficients calculated using
four different methods are almost independent of axial plane.
This is likely to hold true in most applications. Small rises in
mixed-out loss coefficient arise when the wakes pass through
local streamwise pressure gradients, which increase the velo-
city gradients (and therefore the rate of entropy generation) in
the shear layers at the wake edges. All four mixed-out
methods provide closer comparisons between experiment
Fig. 16 Variation of
ζ with axial distance for two exit Mach and CFD than the plane-average methods. This highlights
numbers: M2 = 0.76 and 1.11 the importance of using mixed-out methods for comparing

051011-12 / Vol. 144, MAY 2022 Transactions of the ASME

 absolute KE loss coefficient between experiment and CFD, Greek Symbols
or in situations where inaccurate prediction of mixing rates β = flow turning (or whirl) angle, deg
is expected. γ = ratio of specific heats, dimensionless
(5) Mixed-out KE loss coefficients calculated using the Dzung ζ = kinetic energy loss coefficient, dimensionless
and free-vortex methods were in excellent agreement. Abso- ζ̂ = maximum kinetic energy loss coefficient, dimensionless
lute values of mixed-out loss coefficient calculated from the ζ̂ ref = reference maximum kinetic energy loss coefficient at
Amecke and forced-vortex methods were higher than the first x/Cx = 0.0, dimensionless
two methods. If a single method is to be chosen, the Dzung
ζ = average kinetic energy loss coefficient, dimensionless
method might be preferred due to its ease of implementation
ζ ref = reference average kinetic energy loss coefficient (mean of
and unproblematic inclusion of coolant flows. plane-average methods at x/Cx = 1.0), dimensionless
(6) The axial sensitivity of KE loss coefficient is relatively inde- η = isentropic efficiency, dimensionless
pendent of Mach number—a result that holds true for all θ = circumferential position, deg
plane-average and mixed-out methods. At higher Mach θ′ = normalized circumferential position, vane pitches
number, the associated larger density gradients in the flow θ1 = lower circumferential bound for integration, deg
slightly increase the axial variation in both plane-average θ2 = upper circumferential bound for integration, deg

Downloaded from http://asmedigitalcollection.asme.org/turbomachinery/article-pdf/144/5/051011/6830463/turbo_144_5_051011.pdf by guest on 05 March 2022

and mixed-out loss coefficients. ρ = mass density, kg/m3
(7) For the NGV studied, radially averaging over the central χ = (γ − 1)/γ
60% of vane span is robust for characterizing profile loss.
This range is sufficiently large to avoid sensitivities due to
local radial nonuniformities (particular cooling features, Subscripts
etc.) but not so large as to include sensitivity to endwall 0 = stagnation condition
flows, which is often problematic where comparisons 1 = vane inlet station
between situations (e.g., CFD and experiment) are required. 2 = vane exit station
A = area-weighted average value
Acknowledgment M = mass-flow-weighted average value
s = isentropic condition
The support of Rolls-Royce plc and Innovate UK are gratefully S = entropy-flux-weighted average value
acknowledged. V = volume-flow-weighted average value

Data Availability Statement References

The authors attest that all data for this study are included in the [1] Dzung, L. S., 1971, “Konsistente Mittlewerte in der Theorie der Turbomaschinen
paper. für Kompressible Medien,” BBC-Mitt., 58, pp. 485–492.
[2] Amecke, J., 1970, “Anwendung der Transsonischen Ähnlichkeitsregel auf die
Strömung Durch Ebene Schaufelgitter,” VDI-Forschungsheft, 540, pp. 16–28.
[3] Pianko, M., and Wazelt, F., 1982, “Propulsion and Energetics Panel Working
Nomenclature Group 14 on Suitable Averaging Techniques in Non-Uniform Internal Flows,”
h = static enthalpy, J/kg AGARD Advisory Report No. 182. NATO Advisory Group for Aerospace
p = static pressure, Pa Research and Development, Neuilly-sur-Seine, France.
[4] Oates, G. C., 1997, Aerodynamics of Gas Turbine and Rocket Propulsion, AIAA,
r = radial position, m Reston, VA, pp. 197–203.
s = static entropy, J/(kg K) [5] Cumpsty, N. A., and Horlock, J. H., 2005, “Averaging Non-Uniform Flow for a
v = velocity, m/s Porpoise,” Proceedings of ASME Turbo Expo 2005, Reno-Tahoe, NV, June 6–9.
w = wake width, m [6] Denton, J. D., 1993, “The 1993 IGTI Scholar Lecture: Loss Mechanisms in
x = axial distance or direction, m Turbomachines,” ASME J. Turbomach., 115(4), pp. 621–656.
y = streamwise-normal distance, m [7] Greitzer, E. M., Tan, C. S., and Graf, M. B., 2004, Internal Flow: Concepts and
z = streamwise distance, m Applications, Cambridge University Press, Cambridge, UK.
[8] Main, A. J., Oldfield, M. L. G., Lock, G. D., and Jones, T. V., 1997, “Free Vortex
A = area, m2 Theory for Efficiency Calculations From Annular Cascade Data,” ASME
M = Mach number, dimensionless J. Turbomach., 119(2), pp. 247–255.
Y = stagnation pressure loss coefficient, dimensionless [9] Prasad, A., 2005, “Calculation of the Mixed-Out State in Turbomachine Flows,”
ṁ = mass flowrate, kg/s ASME J. Turbomach., 127(3), pp. 564–572.
q̇ = volume flowrate, m3/s [10] Dominy, R. G., and Harding, S. C., 1989, “An Investigation of Secondary Flows
cp = specific heat at constant pressure, J/(kg K) in Nozzle Guide Vanes,” Proceedings of AGARD PEP 74th Secondary Flows in
Turbomachines, Luxembourg, Aug. 28–Sept. 1, Paper No. AGARD-CP-469.
h0 = stagnation enthalpy, J/kg [11] Kirollos, B., Lubbock, R., Beard, P., Goenaga, F., Rawlinson, A., Janke, E., and
p0 = stagnation pressure, Pa Povey, T., 2017, “ECAT: An Engine Component Aerothermal Facility at the
pref = arbitrary reference pressure, Pa University of Oxford,” Proceedings of ASME Turbo Expo 2017, Charlotte,
r1 = lower radial bound for integration, m NC, June 26–30, Paper No. GT2017-64736.
r2 = upper radial bound for integration, m [12] Povey, T., Sharpe, M., and Rawlinson, A., 2011, “Experimental Measurements of
Gas Turbine Flow Capacity Using a Novel Transient Technique,” ASME
vx = axial velocity component, m/s J. Turbomach., 133(1), p. 011005.
wref = reference wake width at x/Cx = 0.0, m [13] Burdett, D., Hambidge, C., and Povey, T., 2021, “Analysis of Ultra-Low
Cx = axial chord length, m Uncertainty Gas Turbine Flow Capacity Measurement Techniques,” Proc. Inst.
M2 = vane exit Mach number, dimensionless Mech. Eng. Part A: J. Power Energy, 235(5), pp. 1053–1079.
T0 = stagnation temperature, K [14] Burdett, D., Goenaga, F., and Povey, T., 2021, “Understanding Capacity
p02 = mean total pressure in a downstream plane, Pa Sensitivity of Cooled Transonic Nozzle Guide Vanes,” ASME J. Turbomach.,
143(5), p. 051001.
p2 = mean static pressure in a downstream plane, Pa [15] Thomas, F. O., and Liu, X., 2004, “An Experimental Investigation of Symmetric
Re = Reynolds number, dimensionless and Asymmetric Turbulent Wake Development in Pressure Gradient,” Phys.
Tu = turbulence intensity, % Fluids, 16(5), pp. 1725–1745.

Journal of Turbomachinery MAY 2022, Vol. 144 / 051011-13