energies
Article
Design Optimization of a Dual-Bleeding Recirculation Channel
to Enhance Operating Stability of a Transonic Axial Compressor
Tien-Dung Vuong and Kwang-Yong Kim *






Citation: Vuong, T.-D.; Kim, K.-Y.
Design Optimization of a
Dual-Bleeding Recirculation Channel
to Enhance Operating Stability of a
Transonic Axial Compressor. Energies
2022, 15, 159. https://doi.org/
10.3390/en15010159
Academic Editors: Davide Astolfi
and Abdessattar Abdelkefi
Received: 4 November 2021
Accepted: 22 December 2021
Published: 27 December 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Energies 2022, 15, 159. https://doi.org/10.3390/en15010159
Department of Mechanical Engineering, Inha University, Incheon 22212, Korea; vuongtien97bkhn@gmail.com
* Correspondence: kykim@inha.ac.kr; Tel.: +82-32-872-3096
Abstract: The present work performed a comprehensive investigation to find the effects of a dual-
bleeding port recirculation channel on the aerodynamic performance of a single-stage transonic axial
compressor, NASA Stage 37, and optimized the channel’s configuration to enhance the operating
stability of the compressor. The compressor’s performance was examined using three parameters:
The stall margin, adiabatic efficiency, and pressure ratio. Steady-state three-dimensional Reynolds-
averaged Navier–Stokes analyses were performed to find the flow field and aerodynamic performance.
The results showed that the addition of a bleeding channel increased the recirculation channel’s
stabilizing effect compared to the single-bleeding channel. Three design variables were selected
for optimization through a parametric study, which was carried out to examine the influences of
six geometric parameters on the channel’s effectiveness. Surrogate-based design optimization was
performed using the particle swarm optimization algorithm coupled with a surrogate model based
on the radial basis neural network. The optimal design was found to increase the stall margin by
51.36% compared to the case without the recirculation channel with only 0.55% and 0.28% reductions
in the peak adiabatic efficiency and maximum pressure ratio, respectively.
Keywords: axial compressor; recirculation channel; RANS analysis; optimization; stall margin;
genetic algorithm
1. Introduction
An axial compressor is an integral component of gas turbines, which powers the major-
ity of today’s aircrafts. When compressors operate at the maximum pressure ratio, they are
prone to a phenomenon called a stall. A stall is a detrimental occurrence that destabilizes
the flow field and greatly degrades performance and safety. In axial compressors, a primary
cause of stall is the tip leakage vortex occurring in the rotors’ tip region. At a high pressure
ratio, the tip leakage vortex breaks down due to its interaction with the passage shock
wave. As a result, a stagnation zone containing low-momentum airflow is formed, which
acts as a blockage in the flow field. As the pressure continues to increase, the blockage
grows until it destabilizes the whole flow path, and the compressor falls into a stall. With a
rising demand for higher-pressure-ratio axial compressors, the problem associated with the
tip leakage vortex becomes more concerning and thus, increasing efforts are being spent to
solve this issue.
A promising solution to improve the stability of axial compressors is the use of a
recirculation channel. A design schematic of a basic recirculation channel can be found
in the work of Hathaway [1]. When using a recirculation channel, air flow is bled from
a downstream location on the compressor shroud and then re-injected to an upstream
location. Due to the pressure difference between the bleeding and injecting positions,
the accelerated recirculating air re-energizes the low-momentum air inside the blockage.
Consequently, the expansion of the blockage is delayed, and stall inception is prevented.
Several designs of recirculation channels have been proposed and patented so far [2–4].
A basic concept of self-recirculating casing treatment was presented by Hathaway [1], who
https://www.mdpi.com/journal/energies
Energies 2022, 15, 159 2 of 17

combined air bleeding and injection in the rotors’ tip region. The stable range extension
of the tested transonic fan, NASA Rotor 67, was increased by 225% with inlet distortion
while injection location and angle were found to be critical design parameters. A stage-
recirculation design from Strazisar et al. [5] extracted air from a stator casing and re-
injected it to the upstream of a preceding rotor, which extended the stable operating
range of the transonic compressor NASA Stage 35. They concluded that increasing the
circumferential mass-averaged axial velocity at the rotor tip is the key to improving the
compressor’s stability. An end-wall recirculation scheme with an anti-swirl vaned passage
was tested by Khaleghi et al. [6] on the high-speed fan, NASA Rotor 67. The scheme
was able to increase the stable range extension with a very small penalty in efficiency.
In a design of flow recirculation, Weichert et al. [7] adapted the recirculating flow rate
to the compressor’s working condition. The recirculating air flow was minimized at the
compressor’s design condition while maximized at the near-stall condition, which led to a
stall margin improvement in the range of 2.2–6.0%. A recirculation scheme of Khaleghi [8]
enhanced the performance of the high-speed fan, NASA Rotor 67, by using 22 skewed
channels. The increased stability was found to be the result of recirculating flow pushing
the tip leakage vortex and passage shock downstream at the near-stall condition.
The self-recirculation concepts suggested by Hathaway [1] and Strazisar [5] were
parameterized and tested by Wang et al. [9] focusing on the circumferential coverage and
injector’s throat height. The results indicated that optimizing recirculation channel designs
could improve not only compressors’ stability, but also other performance parameters. A
feedback channel design proposed by Dinh et al. [10] enhanced the stall margin of the single-
stage transonic compressor, NASA Stage 37 by 26.8% with only a 0.14% decrease in the
efficiency. Unlike other recirculation schemes, their channel was located between the rotors’
edges, which facilitated its integration into multistage configurations. The effectiveness of
flow recirculation channels on both single- and multi-stage compressors was experimentally
investigated by Li et al. [11]. Application of the recirculation channels increased the stall
margins of the single- and three-stage compressors by 13.7% and 13.0%, respectively,
without a reduction in the efficiency. A recirculation casing treatment introduced by
Khaleghi [12] extracted air from downstream and circumferentially re-injected it into the
flow path in the opposite direction of the blade rotation. The results showed a considerable
extension in the compressor’s stable working range at the expense of a moderate efficiency
loss. Recently, a casing treatment combining flow recirculation and circumferential groove
was proposed by Vuong et al. [13] to enhance the stability of the transonic compressor,
NASA Stage 37. The design was able to provide a maximum stall margin increase of 42.5%
with a less than 1% decrease in the adiabatic efficiency.
In the last couple of decades, surrogate-based optimization techniques have shown
great potential in the design of turbomachinery. Considering the complexity of turboma-
chine design, optimization techniques have contributed greatly to speeding up the design
process and discovering unexpected designs. The blade shape of a transonic axial compres-
sor was optimized by Samad and Kim [14] using the elitist non-dominated sorting genetic
algorithm (NSGA-II) [15] combined with response surface approximation [16]. With the
two extreme-end designs of the Pareto front, two objective functions, i.e., the total pressure
ratio and adiabatic efficiency, were increased by 1.76% and 0.41%, respectively. Using the
hybrid multi-objective evolutionary algorithm (MOEA) [17], Kim et al. [18] attempted to
enhance the efficiency and mitigate the noise level of an axial fan. The NSGA-II algorithm
and RBNN model were utilized by Khalfallah et al. [19] to find an optimal design of a
centrifugal compressor to maximize the efficiency and stall margin. Various optimization
algorithms coupled with the RBNN model were implemented by Ma et al. [20] to find an op-
timal ring cavity design for stall margin improvement of a centrifugal compressor. Among
four popular algorithms, i.e., the genetic algorithm (GA) [21], particle swarm optimization
(PSO) [22], simulated annealing [23], and SQP, PSO was found to be the most effective
one in terms of the computing time and stall margin gain. Cheng et al. [24] successfully
applied an improved artificial bee colony algorithm [25] to find an optimal blade shape
Energies 2022, 15, 159 3 of 17

for a four-stage, low-pressure axial compressor. The optimized blade raised the adiabatic
efficiency at the design point from 82.07% to 82.55% and stall margin from 10.2% to 15.0%.
In the present work, the performance of a recirculation channel design with two
bleeding channels (RE2B) was analyzed and optimized to enhance the stability of the single-
stage transonic axial compressor, NASA Stage 37. In comparison with the conventional
single-bleeding design, the addition of one more bleeding channel was expected to increase
the amount of recirculating airflow as well as the energy it contains, which would improve
the stabilizing function of the channel. A previous study on a similar recirculation casing
treatment design [26] has also confirmed its positive influence on a transonic compressor’s
stability. However, no comprehensive investigation of the casing treatment design has been
presented in the literature. Therefore, in this study, a detailed parametric study with six
geometric parameters of the channel was conducted to assess their levels of influence on the
channels’ effectiveness. Moreover, based on this parametric study, a surrogate-based design
optimization was performed to find an optimal channel design with maximized stability
improvement. Three-dimensional Reynolds-averaged Navier-Stokes (RANS) equations
were solved numerically for the aerodynamic analysis.

2. Compressor Model and Casing Treatment Design

The single-stage transonic axial compressor tested in the present work was NASA
Stage 37. The report of Reid and Moore [27] provided detailed information regarding
the compressor’s geometry and aerodynamic performance. This compressor exhibits a
spike-type rotating stall that originated from the rotors’ tip [28,29], which makes it well-
suited for the investigation of casing treatments. The compressor comprises 36 rotor blades
rotating at a design speed of 17,185.7 rpm and 46 stationary stator blades. The tip clearance
values are 0.0400 cm under the rotor shroud and 0.0762 cm over the stator hub, which
correspond to 1.44% of the axial chord length at the rotor tip and 2.11% of the axial chord
length at the stator tip, respectively. From the experiment by Reid and Moore [27], when
running at 100% design speed and the peak efficiency condition, the compressor’s total
pressure ratio and adiabatic efficiency were 2.00% and 84.00%, respectively, with a mass
flow rate of 20.74 kg/s. At the near-stall condition, the compressor’s total pressure ratio
was 2.093 and the mass flow rate was 19.60 kg/s. The mass flow rate reached 20.93 kg/s
at the choking condition. The reference pressure and temperature were 101,325 Pa and
288.15 K, respectively.
The schematic of RE2B is illustrated in Figure 1. The recirculation channels are located
on the shroud between the leading and trailing edges of the rotor blades. Compared to the
previously mentioned similar design [26] that positioned its injection and bleeding ports
outside the rotor’s edges, the present design does not intrude into other blade row regions
and is more lightweight. The whole system consists of 36 channels uniformly distributed
around the compressor’s annulus, each of which covers 8◦ in the circumferential direction.
The spacing between individual channels (2◦ ) is necessary in practice for the electrical
wiring around the compressor’s casing. The design of the RE2B channel is parameterized
using six geometric parameters: The injection location (LF ), injection angle (α), middle bleed
location (LM ), middle bleed angle (β), rear bleed location (LB ), and rear bleed angle (γ).
Their values for the reference design are listed in Table 1, with all the location parameters
non-dimensionalized by the rotor tip’s axial chord length (CR ).
ies 2022, 14, x FOR PEER REVIEW 4 of 17
Energies 2022, 15, 159 4 of 17

(a)

(b)
Figure 1. Geometric
Figuredescription of NASA
1. Geometric Stageof37NASA
description compressor and
Stage 37 RE2B: (a) Geometric
compressor and RE2B: parameters
(a) Geometric parameters
of RE2B; (b) meridional view
of RE2B; (b) of NASAview
meridional Stageof37 with RE2B.
NASA Stage 37 with RE2B.

Table 1. Reference design of RE2B.

Table 1. Reference design of RE2B.
Parameters LF/CR (%) LM/CR (%) LB/CR (%) α (°) β (°) γ (°)
Parameters LF /CR (%) LM /CR (%) LB /CR (%) α (◦ ) β (◦ ) γ (◦ )
Values 45 55 70 45 45 45
Values 45 55 70 45 45 45

3. Numerical Analysis
3. Numerical Analysis
3.1. Numerical Methods
3.1. Numerical Methods
The aerodynamic analysis was conducted using the commercial code ANSYS CFX
The aerodynamic analysis was conducted using the commercial code ANSYS CFX
15.0® [30]. To reduce the computing time, only one compressor’s passage comprising one
15.0® [30]. To reduce the computing time, only one compressor’s passage comprising one
rotor and stator pair was considered in the computation. As illustrated in Figure 2, the
rotor and stator pair was considered in the computation. As illustrated in Figure 2, the
computational domain comprises three sub-domains: The rotor, stator, and RE2B channel.
computational domain comprises three sub-domains: The rotor, stator, and RE2B channel.
A hexahedral mesh system for the computational domain was made using TurboGrid®
A hexahedral mesh system for the computational domain was made using TurboGrid®
and ICEM-CFD®. TurboGrid®® was used to® generate the mesh for the rotor and stator
and ICEM-CFD . TurboGrid was used to generate the mesh for the rotor and stator
blocks (more information can be found in ref. [10]). The mesh of RE2B was created using
blocks (more information can be found in ref. [10]). The mesh of RE2B was created using
ICEM-CFD® with the determinant quality kept larger than 0.7. Assigning the boundary
ICEM-CFD® with the determinant quality kept larger than 0.7. Assigning the boundary
conditions, solving the governing equations, and post-processing the numerical results
conditions, solving the governing equations, and post-processing the numerical results
were performed by ANSYS CFX-Pre, CFX-Solver, and CFD-Post, respectively.
were performed by ANSYS CFX-Pre, CFX-Solver, and CFD-Post, respectively.
All simulations were performed using steady RANS analysis in the present work.
For the steady analysis, the rotating and stationary blocks were linked by a general grid
interface (GGI) method. The frozen-rotor method was applied at the interface between
the rotor and stator domains (360◦ /36 = 10◦ for the rotor, 360◦ /46 = 7.826◦ for the stator)
and also between RE2B and rotor domains (8◦ for the channel and 10◦ for the rotor). The
influences of the frozen-rotor method on the numerical results in comparison with those of
the stage method were discussed by Shim and Kim [31].
Energies 2022,
Energies 15, x159
2022, 14, FOR PEER REVIEW 5 5ofof17
17

Figure
Figure 2.
2. Computational
Computational domain
domain and
and grid
grid system.
system.

Thesimulations
All k-ε turbulence weremodel was used
performed usingwith
steadya scalable wall function,
RANS analysis while work.
in the present y+ values
For
of the first nodes near the walls were maintained in a range of
the steady analysis, the rotating and stationary blocks were linked by a general grid 20–100. A fully implicit
inter-
element-based
face (GGI) method. finiteThevolume method
frozen-rotor was implemented
method was applied to discretize
at the interface three-dimensional
between the ro-
(3D) governing differential equations. For the advection terms, a
tor and stator domains (360°/36 = 10° for the rotor, 360°/46 = 7.826° for the stator) andhigh-resolution scheme
also
using theRE2B
between principles
and rotorof Barth and Jesperson
domains (8° for the[32] was used,
channel which
and 10° for has
the second-order accuracy
rotor). The influences
in the
of space.
frozen-rotor method on the numerical results in comparison with those of the stage
method Thewere
simulations
discussedused idealand
by Shim gasKim(air)[31].
as the working fluid. At the stator’s outlet,
the average static pressure condition was
The k-ε turbulence model was used with a scalable used, while atwall
the function,
rotor’s inlet,whilethey+turbulence
values of
intensity and total temperature were specified as 5% and 288.15
the first nodes near the walls were maintained in a range of 20–100. A fully implicit K, respectively. Smooth
ele-
ment-based finite volume method was implemented to discretize three-dimensionalof(3D)
and adiabatic conditions were applied to all wall boundaries. The side boundaries the
rotor and stator domains were specified with periodic conditions.
governing differential equations. For the advection terms, a high-resolution scheme using
In this work,
the principles of Barththe numerical
and Jespersoncalculations
[32] was were
used, performed until the near-stall
which has second-order accuracypoint,
in
which shows the highest pressure ratio. Due to the intrinsic unsteady characteristics of
space.
the stall/surge phenomenon, strict criteria for the near-stall point must be implemented to
The simulations used ideal gas (air) as the working fluid. At the stator’s outlet, the
achieve a reasonably accurate assessment with steady simulation. The convergence criteria
average static pressure condition was used, while at the rotor’s inlet, the turbulence inten-
used in this work were those suggested by Chen et al. [33]: The variation between the inlet
sity and total temperature were specified as 5% and 288.15 K, respectively. Smooth and
and outlet mass flow rates is less than 0.3%, the inlet mass flow rate fluctuation is less
adiabatic conditions were applied to all wall boundaries. The side boundaries of the rotor
than 0.001 kg/s for 300 steps, and the change of adiabatic efficiency is less than 0.3% per
and stator domains were specified with periodic conditions.
100 steps.
In this work, the numerical calculations were performed until the near-stall point,
To obtain the performance curve of the compressor, at the compressor’s outlet, the
which shows the highest pressure ratio. Due to the intrinsic unsteady characteristics of
average static pressure was varied from the choking condition (0 Pa) to the near-stall
the stall/surge phenomenon, strict criteria for the near-stall point must be implemented to
condition with a step size of 100 Pa around the peak efficiency condition and a step size of
achieve a reasonably accurate assessment with steady simulation. The convergence crite-
10 Pa (which corresponds to about 0.0001 kg/s of mass flow rate) at the near-stall condition.
ria used in this work
The convergence were those
criterion for thesuggested by Chen et
root-mean-square al. [33]:ofThe
residual each variation
governing between the
equation
inlet and outlet mass
− 6 flow rates is less than 0.3%, the inlet mass flow
was limited to 10 . The calculations were carried out by a computer with an Intel i7-4930K rate fluctuation is
less than 0.001 kg/s for 300 steps, and the change of adiabatic efficiency
3.4 GHz CPU. Each operating point was obtained after 2–3 h simulation time, and it took is less than 0.3%
per
about10070steps.
h to find a performance curve on average.
To obtain the performance curve of the compressor, at the compressor’s outlet, the
average static pressure
3.2. Performance Parameterswas varied from the choking condition (0 Pa) to the near-stall con-
ditionThree
with parameters,
a step size ofi.e.,100the
Pa total
around the peak
pressure efficiency
ratio condition
(PR), adiabatic and a step
efficiency (η),size
andofstall
10
Pa (which corresponds to about 0.0001 kg/s of mass flow rate) at
margin (SM), were used to assess the aerodynamic performance and operational stabilitythe near-stall condition.
The convergence
of the criterion forwith
transonic compressor the and
root-mean-square
without the casing residual of each
treatment governing equation
[10,34,35]:
was limited to 10−6. The calculations were carried out by a computer with an Intel i7-4930K
3.4 GHz CPU. Each operating point was pt, outafter 2–3 h simulation time, and it took
PRobtained
= (1)
about 70 h to find a performance curve on average. pt, in
  γ −1
3.2. Performance Parameters pt, out γ
pt, in −1
= pressure
Three parameters, i.e., the ηtotal Tt, out ratio × 100%
(PR), (2)
adiabatic efficiency (η), and stall
T − 1
margin (SM), were used to assess the aerodynamic
t, in performance and operational stability
of the transonic compressor with and without .
mthe casing
! treatment [10,34,35]:
PR NS peak
SM = × . − 1 × 100% (3)
PR peak𝑃𝑅 =m𝑝NS,
(1)
𝑝,
 𝑝,
1
𝑝, (2)
𝜂= 100%
𝑇,
1
𝑇,
Energies 2022, 15, 159 𝑃𝑅 𝑚 6 of 17
𝑆𝑀 = 1 100% (3)
𝑃𝑅 𝑚
where 𝑝 , 𝑇 , and. 𝑚 are the total pressure, total temperature, and normalized mass flow
where pt , Tt , and m are the total pressure, total temperature, and normalized mass flow rate,
rate, respectively. The subscripts in and out indicate the values measured at the compres-
respectively. The subscripts in and out indicate the values measured at the compressor’s
sor’s inlet and outlet, respectively. The subscripts peak and NS refer to the peak adiabatic
inlet and outlet, respectively. The subscripts peak and NS refer to the peak adiabatic
efficiency and near-stall condition, respectively. The subscripts SC and CT refer to the
efficiency and near-stall condition, respectively. The subscripts SC and CT refer to the
smooth casing and casing treatment, respectively.
smooth casing and casing treatment, respectively.
3.3. Validation
3.3. Validation and
and Grid-Dependency
Grid-Dependency Test
Test
The validation
The validation ofof the
the numerical
numerical results
results for
for the
the NASA
NASA Stage
Stage 37
37 compressor
compressor waswas pre-
pre-
sented in detail in the study of Dinh et al. [10]. Their work showed that
sented in detail in the study of Dinh et al. [10]. Their work showed that the numerical the numerical
results closely
results closely reproduced
reproduced the the performance
performance characteristic
characteristic compared
compared toto the
the experimental
experimental
data [28] as shown in Figure 3. Since the same compressor model and numerical methods
were used
were used in
in this
this work,
work, this
this validates
validates the
the accuracy
accuracy ofof the
thesubsequent
subsequentnumerical
numericalresults.
results.

Figure 3. Validation of numerical results for NASA Stage 37 compressor model [10] against
against experi-
experi-
mental measurement
mental measurement [27].
[27].

To evaluate
To evaluate the
the grid-dependency
grid-dependency of of the
the numerical
numerical results
results with
with RE2B,
RE2B, analysis
analysis using
convergence index (CGI) based
the grid convergence based on
on the
the Richardson
Richardson extrapolation
extrapolation technique
technique was
was
performed.
performed. This analysis was proposed
proposed by Celik
Celik and Karatekin
Karatekin [36]
[36] and is widely used to
assess the grid dependency
dependency of of numerical
numerical solutions.
solutions.
Details of the CGI analysis of RE2B’s meshes are presented in Table Table 2.
2. As shown
shown in
this table, the stall margin value converges monotonically
monotonically as as the number of mesh elements
increases. From these results, mesh G1 was was selected
selected for
for all
all proceeding
proceeding calculations.
calculations. The
grid system in the compressor without RE2B is the same as that used by Dinh et al. [10]
(340,556
(340,556 nodes
nodes for
for rotor
rotor block
block and
and 249,524
249,524 nodes
nodes for
for stator
stator block).
block).

3.4.
TableOptimization Methods
2. CGI analysis results for the grid in RE2B.
The single-objective optimization problem is defined as follows:
Parameters Symbols Values
Number of mesh elements G1/G2/G3
Maximize F(x) 293,319/148,665/63,632(4)

xiL ≤ xi ≤ xiU
where F is an objective function and x (={xi }) is a vector of n design variables. An optimiza-
tion algorithm performs a search procedure to find an optimal solution within the design
space specified by the lower limit xiL and upper limit xiU of each design variable. The stall
margin was selected as the objective function for optimization in the present study:

F = SM (5)
Energies 2022, 15, 159 7 of 17

Table 2. CGI analysis results for the grid in RE2B.

Parameters Symbols Values

Number of mesh elements
G1/G2/G3 293,319/148,665/63,632
in RE2B
Grid refinement factors r21/r32 1.97/2.34
Stall margin corresponding to
SM1/SM2/SM3 13.40/13.38/13.24
G1/G2/G3
Apparent order p 3.13
Extrapolated value 21
φext 13.40
Approximated relative error e21
a 0.0746%
Extrapolated relative error e21
ext 0.0101%
Grid convergence index CGI 21
f ine
0.0126%

The design of experiment (DOE) was performed using the Latin hypercube sampling
(LHS) method [37]. Sampling points (or design points) are generated by LHS and are
composed of an i × j matrix, where i and j represent the numbers of sampling points and
design variables, respectively. Each column j is filled with a randomly paired permutation
level of 1, 2, . . . , i to form a Latin hypercube. The method is provided as the function
lhsdesign in MATLAB [38].
The objective function was approximated by surrogate modeling using RBNN. RBNN
is a two-layer neural network that comprises a hidden layer of the radial basis function
and a linear output layer. The two hyper parameters of this surrogate model are an error
goal and a spread constant. The former’s value can be chosen by the user based on the
permissible error from the mean input response. MATLAB provides a built-in function,
newrb, to apply RBNN modeling [38].
PSO [25] is a meta-heuristic optimization algorithm that replicates the “swarm be-
havior” of a bird flock or fish school. The algorithm starts with a population, called a
swarm, of particles, each of which is a possible solution. These particles are then moved
around the search space following some defined formulae. The movement of each particle
is under the influence of its own best-known position (local best objective function value)
and the whole swarm’s best-known position (global best objective function value). Fol-
lowing this principle, the particles are guided toward a better position after each iteration.
The process is repeated until a solution that satisfies all conditions is found. In MATLAB,
the single-objective PSO is provided in the optimization toolbox and can be utilized by
using function pso [38]. In the current study, the swarm size is set at 100 particles and the
objective function tolerance is set at 10−6 . The optimization procedure can be summarized
as in Figure 4.
Energies 2022,
Energies 14,15,x 159
2022, FOR PEER REVIEW 8 of 17 8

Energies 2022, 14, x FOR PEER REVIEW 8 of 17

Figure 4. Single-objective optimization process.

Figure
Figure Single-objective
4. 4. optimization
Single-objective process.
optimization process.
4. Results and Discussion
4. Results and Discussion
4.1.
4. Results
4.1. Results with
Results with and the Reference
theDiscussion
Design of RE2B
Reference Design of RE2B
4.1.Figure
Results
Figure 55compares three
threeperformance
with the Reference
compares Design of
performance curves
RE2B
curves of
ofthe
thecompressor
compressorfor forthe
thecases
caseswithout
without
any
any casing treatment (SC), with the conventional recirculation channel (RE)and
casing treatment (SC), with the conventional recirculation channel (RE) andwith
withthe the
referenceFigure
RE2B 5channel
compares (Table three
1). Asperformance
shown in the curvesthe
figure, of working
the compressor
mass flow for theofcases wit
range
reference RE2B channel (Table 1). As shown in the figure, the working mass flow range of
any
the
the casing treatment
compressor
compressor isisextended
extended (SC),
to withmass
tolower
lower the flow
mass conventional
flowrates
rateswhen
when recirculation
using
usingboth
bothcasingchannel
casing (RE) and wit
treatments.
treatments.
However,
reference RE2B
RE2B lowers
channel the near-stall
(Table 1).mass
As flow
shown rate in to
thethe
However, RE2B lowers the near-stall mass flow rate to the normalized value of 0.9179, normalized
figure, the value
working of 0.9179,
mass flow ran
which is lower
the compressor
which than
is lower than is that
that of
extendedRE (0.9198).
to lower
of RE (0.9198). In exchange,
mass flow
In exchange, the losses in
ratesinwhen
the losses efficiency
using
efficiency and
andboth the pres-
casing treatm
the pressure
sure
However, RE2B lowers the near-stall mass flow rate to the normalizedthese
ratioratio
are are higher
higher withwith
RE2B RE2B compared
compared to to those
those of ofREREatathigh
highmass
mass flow
flow rates,
rates, but
but these
value of 0.
differences
differences diminish
diminish for
for lower
lower mass
mass flow
flow rates.
rates. Consequently,
Consequently, the
the difference
difference between
between the
theand the
which
pressure
is lower than that of RE (0.9198). In exchange, the losses in efficiency
pressure ratios
ratios at at peak
peakadiabatic
adiabaticefficiency
efficiency andand the
thenear-stall
near-stallcondition
conditionincreases
increaseswith withthe the
suredesign.
RE2B ratio are In higher
Equation with
(3), RE2B
the stallcompared
margin to those
value is of REbyatthis
affected high mass flow
difference in rates, but
the
RE2B design. In Equation (3), the stall margin value is affected by this difference in the
differences
pressure
pressure ratio diminish
ratio and
and the for lower mass
thecorresponding
corresponding mass flow
mass flow rates.Therefore,
flow range.
range. Consequently,
Therefore, RE2B the difference
RE2Bincreases
increases the stall betwee
thestall
pressure
margin from ratios
9.95% at
of peak
SC toadiabatic
13.40%, efficiency
while RE only and the
yields anear-stall
stall
margin from 9.95% of SC to 13.40%, while RE only yields a stall margin of 12.37%. This stall margin condition
of 12.37%. increases
This wit
stall
RE2B
marginmargin increase
design.
increase with
RE2BRE2B
In Equation
with is significant
(3), the (about
is significant (about
34.7% 34.7%
stall margin value
compared compared that to
istoaffectedof that
by ofthis
smooth smooth
difference i
casing),
casing),
atpressureat the
the expense expense
of only of only
0.52% 0.52%
and 0.99%and 0.99%
reductions reductions
in in
adiabatic
ratio and the corresponding mass flow range. Therefore, RE2B increases adiabatic
efficiency efficiency
and the and the
pressure the
pressure
ratio at ratio
the peakat the peak efficiency
efficiency condition, condition, respectively.
respectively. The The following
following discussion discussion
explains ex-
the12.37%.
margin from 9.95% of SC to 13.40%, while RE only yields a stall margin of
plains
effectsthe effectson
of RE2B ofthe
RE2B on the compressor’s
compressor’s performance performance
by analyzing by analyzing the flow field.
the flow field.
stall margin increase with RE2B is significant (about 34.7% compared to that of sm
casing), at the expense of only 0.52% and 0.99% reductions in adiabatic efficiency an
pressure ratio at the peak efficiency condition, respectively. The following discussio
plains the effects of RE2B on the compressor’s performance by analyzing the flow fie

Figure
Figure5.5.Performance
Performancecurves
curvesfor
forthree
threeconfigurations.
configurations.

Figure 6 illustrates streamlines of the tip leakage vortex at the near-stall condition.
When either RE or RE2B is used, it can be seen that the tip leakage vortex is cut off from a
Energies 2022, 15, 159 9 of 17

Energies
Energies 2022,
2022, 14,
14, xx FOR
FOR PEER
PEER REVIEW
REVIEW 99 of
of 17
17

Figure 6 illustrates streamlines of the tip leakage vortex at the near-stall condition.
When either RE or RE2B is used, it can be seen that the tip leakage vortex is cut off from a
portion
portion
portionof of
ofthethe leakage
theleakage
leakageflow flow
flowatat the
at the injection
injection
the injection position.
position.
position. According
According
According to
to the
to the numerical
numerical
the numerical results,
results, the
results,
the leakage
leakage
the leakagemassmass mass flow
flow flow rate
rate through through
rate through the
the rotor’s rotor’s tip
tip region
the rotor’s region
is reduced
tip region is reduced from
from 0.00569
is reduced 0.00569
kg/s ofkg/s
from 0.00569 SC to
kg/s of
of
SC
SC to
to 0.00539
0.00539 kg/s and
0.00539 kg/s
kg/s and
and 0.00534
0.00534 kg/s ofkg/s
0.00534 of
RE and
kg/s of RE
RE and
RE2B, RE2B,
RE2B, respectively.
and respectively. This
This reduction
This reduction
respectively. of leakage
reduction of
of leak-
flow
leak-
age
age flow
indicates
flow thatindicates that
that the
the expansion
indicates the expansion
of the blockage
expansion of
of the
the blockage
blockagebyformed
formed the vortex
formed by
by the vortex
vortex breakdown
breakdown
the is delayed,
breakdown is
is
delayed,
which
delayed, which
contributes contributes
to the delay
which contributes to the
to in
thethedelay in the stall
stallininception.
delay inception.
There are three
the stall inception. There are
Theremechanismsthree mechanisms
related to
are three mechanisms
related
RE
related to
to RE
and RE2B: RE and(1) The
and RE2B:
RE2B: (1)
(1) The
injected flow
The injected
pushesflow
injected the pushes
flow pushes the
leakage flow
the leakage flow
flow at
at the injection
leakage at the injection
location
the injection to thelo-
lo-
cation
main
cationflowto the
to the main
direction,
main flowflow direction,
(2) the leakage(2)
direction, (2)
flowthe leakage
theisleakage flow
ingestedflow is
intoisthe ingested into
channelinto
ingested the channel
by the bleeding
channel by by the
ports,
the
bleeding
and (3) a flow
bleeding ports,
ports, and
and (3)
barrier aa flow
flow barrier
is created
(3) in the tip
barrier is
is created
gap by the
created in
in the tip
tip gap
gap by
recirculating
the the
byflow recirculating
the from the injection
recirculating flow
flow
from
to the
bleeding injection
positions. to bleeding
As a result,positions.
the flow As a result,
supply to the
the
from the injection to bleeding positions. As a result, the flow supply to the tip leakage flow
tip supply
leakage to
vortex the tip
reduces. leakage
The
vortex
vortex reduces. The decrease in the leakage flow with RE2B is higher than that with RE
decrease reduces.
in the The
leakage decrease
flow in
with the
RE2Bleakage
is flow
higher with
than RE2B
that withis higher
RE due than
to that
the with
additional RE
due
due toto the
bleeding additional
thechannel,
additional which bleeding channel,
increases
bleeding which
which increases
the amount
channel, of bled and
increases the amount
amount of
the re-injected of bled and
and re-injected
air flow.
bled It is noted
re-injected
air
air flow.
that a new
flow. It is
is noted
It vortexnoted is that
formed
that aa new
new vortex
in the caseis
vortex isofformed
RE2B. in
formed the
the case
However,
in of
of RE2B.
case since RE2B.it isHowever,
located behind
However, since
since it it is
theis
located
located behind
passage shock, itthe
behind does
the passage
passage shock,
not suffer
shock, it
it does
vortex does not
not suffer
breakdown vortex
suffer andvortex thusbreakdown
does not have
breakdown and
and thus
thus does
does not
a substantial not
destabilizing
have
have aa substantial
substantialeffect destabilizing
to induce stalleffect
destabilizing inception.
effect to
to induce
induce stall
stall inception.
inception.

Figure 6.
Figure6. Streamlines
6. Streamlines of
Streamlinesof tip
oftip leakage
tipleakage vortex
leakagevortex at
vortexat near-stall
atnear-stall condition.
near-stallcondition.
condition.
Figure

Figure
Figure777shows
Figure shows streamlines
showsstreamlines near
streamlinesnear the
nearthe suction
thesuction surface
suctionsurface
surfaceof of the
ofthe rotor
the rotor blade
rotorblade at
bladeat the
at the near-stall
the near-stall
near-stall
condition.
condition.The
condition. The red
The red dashed
red dashed line
dashed line indicates
line indicates the
indicates the lines
the lines of
lines of flow
of flow separation
flow separation
separation on on
on the suction surface,
the suction surface,
which
whichisis
which pushed
ispushed downward
pusheddownward
downwarddue due
duetoto the
the
to the recirculation
recirculation channels.
channels.
recirculation channels. Consequently,
Consequently,
Consequently,the the interac-
interaction
the interac-
tion
tion between the tip leakage vortex and flow separation is reduced, which providesaaamore
between
between the the
tip tip leakage
leakage vortex
vortex andand
flowflow separation
separation is is reduced,
reduced, which
which provides
provides more
more
stable
stable flow
stableflow field
flowfield
fieldinin the
inthe tip
thetip region.
tipregion.
region.

Figure
Figure 7. Streamlines
7. Streamlines
Figure7. on
Streamlineson suction
onsuction surface
suctionsurface of
surfaceof rotor
ofrotor blade
rotorblade at
bladeat near-stall
atnear-stall condition.
near-stallcondition.
condition.

In
In Figure 8, another advantage of RE2B over RE is observed. When using RE, aa region
In Figure
Figure 8,
8, another
another advantage
advantage ofof RE2B
RE2B over
over RERE isis observed.
observed. When
When using
using RE,RE, a region
region
of
of low speed flow, which might be aaresult of flow separation, is found over the rear-half
of low speed flow, which might be a result of flow separation, is found over the rear-half
low speed flow, which might be result of flow separation, is found over the rear-half
of
of the rotor’s suction surface. However, this stagnation zone disappears in the case with
of the
the rotor’s
rotor’s suction
suction surface.
surface. However,
However, this this stagnation
stagnation zone zone disappears
disappears in in the
the case
case with
with
RE2B,
RE2B, which reduces the blockage in the flow passage. While the stagnation zones in the
RE2B, which
which reduces
reduces thethe blockage
blockage in
in the
the flow
flow passage.
passage. While
While the
the stagnation
stagnation zones
zones in in the
the
stator
stator region are larger with RE2B, their negative effects on the stability are expected to be
stator region are larger with RE2B, their negative effects on the stability are expected to be
region are larger with RE2B, their negative effects on the stability are expected to be
outweighed
outweighed by the positive changes in the rotor region.
outweighed by by the
the positive
positive changes
changes inin the
the rotor
rotor region.
region.
Energies 2022,15,
Energies2022, 14,159
x FOR PEER REVIEW 10 of
10 of 17
17

Figure 8.
Figure 8. Relative
Relative Mach
Mach number
number contours
contours at
at 98%
98% rotor
rotorblade
bladespan
spanat
atnear-stall
near-stallcondition.
condition.

Figure
Figure 99 shows
shows contours
contours of of circumferential-averaged
circumferential-averaged static static entropy
entropy at
at the
the peak
peak adia-
batic
batic efficiency
efficiencycondition.
condition. The
The reason
reason for
for the
the increased
increased loss
loss in
in efficiency
efficiency is
is found
found toto be
be the
the
increase in entropy generation. In Figure 9, a region of high entropy emerges
increase in entropy generation. In Figure 9, a region of high entropy emerges at the rotor at the rotor
tip
tip region
region downstream
downstreamof of the
the recirculation
recirculationchannel.
channel. This
This is
is probably
probably due
due to
to the
the formation
formation
of
of the
the secondary
secondary leakage
leakage vortex,
vortex,which
whichunavoidably
unavoidablyinduces
inducesefficiency
efficiencyloss.
loss.

Figure 9.
Figure 9. Circumferential-averaged
Circumferential-averagedstatic
staticentropy
entropycontours
contoursat
atpeak
peakadiabatic
adiabaticefficiency
efficiencycondition.
condition.

4.2.
4.2. Parametric
Parametric Study
Study
To
Todetermine
determinethe most
the influential
most variables
influential in the
variables in channel’s design,
the channel’s a parametric
design, study
a parametric
was
studyperformed with six
was performed geometric
with parameters
six geometric of RE2B.
parameters TheirTheir
of RE2B. tested ranges
tested are specified
ranges are spec-
in Table
ified 3. The
in Table reference
3. The values
reference in Table
values 1 are
in Table maintained
1 are maintained for
forthe
theparameters
parametersthat
that are
are
not changed.
not changed.

Table3.3.Tested
Table Testedvalues
valuesofofsix
sixgeometric
geometricparameters.
parameters.

Parameters
Parameters LLFF/C
F /CRR(%)
R (%) LLMM /CRRR(%)
M/C (%) LB /C
LBBR/C(%)
R (%)
R α (α◦ )(°) (◦ )
ββ(°) (◦ )
γγ(°)
Lower
Lowerlimit
limit 40
40 50
50 65 65 3030 15
15 15
15
Upperlimit
Upper limit 5050 60
60 80 80 7575 60
60 60
60

Figure10
Figure 10presents
presentsthe theeffects
effectsofofinjection
injectionlocation
locationonon
the the performance
performance of of
thethe compres-
compressor.
sor. The highest stall margin of 13.40% is achieved at LFF/CR = 45% while the effect on the
The highest stall margin of 13.40% is achieved at LF /CR = 45% while the effect on the peak
peak adiabatic
adiabatic efficiency
efficiency is relatively
is relatively small.small. The influence
The influence of these
of these parameters
parameters on pressure
on the the pres-
sure shows
ratio ratio shows different
different trendstrends for near-stall
for near-stall and peakandefficiency
peak efficiency conditions.
conditions. In FigureIn 11,
Figure
the
11, theofeffects
effects of injection
injection angle on angle on the performance
the performance are illustrated.
are illustrated. The stall
The stall margin margin
reaches its
peak at α = 60 while the best angle for efficiency is 45 . Increasing the angle from 45◦
reaches its peak◦ at α = 60° while the best angle for ◦
efficiency is 45°. Increasing the angle
from 45°reduces
slightly slightlythereduces
pressurethe ratio
pressure ratio
at peak at peak efficiency
efficiency conditionsconditions but increases
but increases the pressurethe
pressure ratio at the near-stall
ratio at the near-stall condition. condition.
Energies 2022,
Energies 2022, 15,
2022, 14, 159
14, x FOR
FOR PEER
PEER REVIEW
REVIEW 1111of
of 17
of 17
Energies 2022, 14, xx FOR PEER REVIEW 11
11 of 17

Figure 10.
Figure Effects of
10. Effects
Effects of injection
injection location on
location on performance.
on performance.
performance.
Figure 10. of injection location

Figure 11.
Figure 11. Effects
Effects of
of injection
injection angle
angle on
on performance.
performance.
Figure 11. Effects
Effects of injection angle on performance.

Figure 12
Figure 12 shows
shows that
shows that the
thatthe best
thebest middle
bestmiddle bleed
middlebleed location
bleedlocation for
locationfor stall
forstall margin
margin
stall margin improvement
improvement
improvement is
is
L
L M
isMML/C R =
/CMRR/C 55%. Moving
R = 55%.
= 55%. the
Moving
Moving middle bleed
the middle
the middle port
bleedbleed downstream
port downstream
port downstream slightly
slightly reduces
slightly the
reduces
reduces peak
the peak adia-
the adia-
peak
batic
adiabatic
batic efficiency.
efficiency. For For
efficiency.
For the pressure
the pressure
the ratio,
pressure aa similar
similar
ratio,
ratio, trend
a similar to that
trend
trend to that
to thatof the
of thethe
of injection
injection
injection location is
location
location is
observed.
is observed.
observed. Figure 13
Figure
Figure 1313presents
presents
presents the effects
theeffects
the of the
effectsofofthe middle
themiddle bleed
middlebleed angle
bleedangle
angleonon the
on the performance.
the performance.
performance.
Overall, this
Overall, this angle
this angle has
angle has aaa small
has small impact
small impact on
impact on all
on all performance
all performance parameters.
performance parameters. In
parameters. In terms
In terms ofof the
of the stall
the stall
stall
margin, γ
margin, γ ==
γ 15◦ provides
= 15°
15° provides
provides thethe best value
the best value of
of 13.66%.
13.66%.

Figure 12.
Figure 12. Effects
12. Effects of
Effects of middle
of middle bleed
middle bleed location
bleed location on
location on performance.
on performance.
performance.
Figure
Energies2022,
Energies
Energies
Energies 2022,14,
2022,
2022, 14,xx159
15,
14, xFOR
FORPEER
FOR PEER REVIEW
PEER REVIEW
REVIEW 1212of
12
12 of
of 17
of17
17
17

Figure
Figure
Figure 13. Effects
Figure 13.
13.
13. Effects
Effects of
Effects of
of middle
of middle bleed angle
middle bleed
middle bleed
bleed angle on
angle
angle on performance.
on performance.
performance.

The effects
The
The effects of
effects of the
of the rear
the rear bleed
rear bleed location on
bleed location
location on the
on the performance
the performance
performance are
performance are
are shown
are shown
shownin
shown in Figure
in Figure
in Figure14.
Figure 14.
14.
14.
L
L B
B /C
/C
/C RR =
= =75%
75%
75% provides
provides
provides the
thethe best
best
best stall
stall margin
stall margin
margin of
of
of 14.69%
14.69%
14.69% and
and
and the
the
the best
best
best
LBB/CR R= 75% provides the best stall margin of 14.69% and the best peak adiabatic efficiencypeak
peak
peak adiabatic
adiabatic
adiabatic efficiency
efficiency
efficiency
of 83.45%.
of
of 83.45%. While
83.45%. Whilethe
While
While thepressure
the
the pressureratio
pressure
pressure ratio
ratio
ratioatatpeak
at
at peakefficiency
peak
peak efficiency
efficiency
efficiency conditions
conditions
conditions
conditions fluctuates,
fluctuates,
fluctuates,
fluctuates, shifting the rear
shifting
shifting
shifting the
the
the
bleed
rear bleed
rear
rear location
bleed location
bleed downstream
location downstream
location downstream reduces
downstream reduces the
reduces the
reduces pressure
the pressure
the ratio
pressure ratio
pressure in the
ratio in
ratio examined
in the
in the examined
the range.
examined range.
examined Figure
range. Figure
range. Figure
Figure15
presents the the
effects of the ◦
15 presents
15
15 presents
presents the
the effects
effects
effects ofrear
of
of the
the bleed
the rear
rear angleangle
rear bleed
bleed
bleed on theon
angle
angle on
onperformance.
the performance.
the
the At β =At
performance.
performance. 45 β
At
At ββ, the stallthe
=== 45°,
45°,
45°, margin
the
the stall
stall
stall
reaches
margin
margin the highest
reaches
reaches the
the value
highest
highest of 13.40%.
value
value of
of Increasing
13.40%.
13.40%. the angle
Increasing
Increasing mitigates
the
the angle
angle
margin reaches the highest value of 13.40%. Increasing the angle mitigates the loss in peak the loss
mitigates
mitigates in peak
the
the adiabatic
loss
loss in
in peak
peak
efficiency,efficiency,
adiabatic
adiabatic
adiabatic and increases
efficiency,
efficiency, and
and the pressure
and increases
increases
increases therise
the
the at therise
pressure
pressure
pressure peak
rise
rise atefficiency
at
at the peak
the
the peak conditioncondition
peak efficiency
efficiency
efficiency while reducing
condition
condition while
while
while
the pressure
reducing
reducing the rise at
pressurethe near-stall
rise at the point.
near-stall point.
reducing the the pressure
pressure riserise atat the
the near-stall
near-stall point.
point.

Figure
Figure
Figure 14. Effects
Figure 14.
14.
14. Effects
Effects of
Effects of rear
of rear
of bleed location
rear bleed
rear bleed location on
location on performance.
on performance.
performance.

Figure 15.
Figure 15. Effects
Effects of
of rear
rear bleed
bleed angle
angle on
on performance.
performance.
Figure 15. Effects
Figure 15. Effects of rear bleed angle
of rear on performance.

Figure
Figure
Figure 16 compares
Figure 16
16
16 compares the
compares
compares the levels
the
the levels of
levels
levels of influence
of influence of
influence of all
of all tested
all tested geometric
tested geometric parameters
geometric parameters on
parameters
parameters on the
on
on the
the
the
compressor’s
compressor’s
compressor’s stall
stall
stall
compressor’s stall margin.
margin.
margin. This
This
This
This comparison
comparison
comparison shows
shows
shows that
that three
three parameters
parameters (L
(L F,,, L
that three parameters FF LBB, α)
F L B,,, α)
B α)
α) have
have
have
the
the
the highest sensitivity
the highest
highest
highest sensitivity on
sensitivity
sensitivity on
on the stabilizing
on the
the
the stabilizing effect
stabilizing
stabilizing effect
effect of the
effect of
of the RE2B
the RE2B channel.
RE2B channel. Therefore,
channel.
channel. Therefore,
Therefore, they
Therefore, they were
they
they were
were
were
selected
selected as
as design
design variables
variables for
for the
the subsequent
subsequent optimization,
optimization, while
while the
the
selected as design variables for the subsequent optimization, while the other geometricother
other geometric
geometric
parameters were
parameters
parameters were fixed.
were fixed.
fixed.
Energies 2022, 14, x FOR PEER REVIEW 13 of 17
Energies 2022, 15, 159 13 of 17

Figure
Figure16.
16.Sensitivity
Sensitivityanalysis
analysisof
ofgeometric
geometricparameters.
parameters.

4.3.Optimization
4.3. Optimization
AAsingle-objective
single-objectiveoptimization
optimizationhas hasbeen
beenperformed
performedwithwiththree
threedesign
designvariables,
variables,LLFF, ,
L , and α, to maximize the stall margin by RE2B. The design space,
LBB, and α, to maximize the stall margin by RE2B. The design space, i.e., the ranges i.e., the ranges of
of the
the design
design variables
variables andvalues
and the the values
of theofother
the other parameters
parameters are specified
are specified in Table
in Table 4. The4.fixed
The
fixed values of the parameters (L M /C R , β, and γ) are their best values for
values of the parameters (LM/CR, β, and γ) are their best values for stall margin improve- stall margin
improvement
ment found from found
thefrom the parametric
parametric study. Thestudy.
DOEThe DOE
has has 20 design
20 design points generated
points generated using
using LHS. RANS simulations were performed to obtain the
LHS. RANS simulations were performed to obtain the stall margin values at stall margin values
these at these
design
design points. Based on those data, an RBNN model was created to approximately
points. Based on those data, an RBNN model was created to approximately relate the stall relate
the stall margin to the design variables. The spread constant and the sum squared error
margin to the design variables. The spread constant and the sum squared error goal for
goal for the RBNN model were 0.1 and 0.00002, respectively.
the RBNN model were 0.1 and 0.00002, respectively.
Table 4.
Table Design space
4. Design space for
for optimization.
optimization.

DesignVariables
Design Variables FixedParameters
Fixed Parameters
LL F /C
F/C RR (%) LL
(%) B /C
B/C (%)
R R(%) (◦ )
αα(°) LLMM/C
/CRR(%)
(%) (◦ )
ββ(°) (◦ )
γγ(°)
Lower
Lowerlimit
limit 4040 7070 45
45
Upperlimit
limit 55
55 4545 15
15
Upper 5050 8080 75
75

Using
UsingthethePSO
PSOalgorithm,
algorithm, the the optimal
optimal design (PSO.opti) was
design (PSO.opti) was found
found at atLLFF/C
/CRR==46.25%,
46.25%,
LLBB/C
/CR = 77.57%,
= 77.57%, and
andα =
α 46.56°.
= 46.56 ◦The
. new
The new design
design provided
provided a astall
stallmargin
margin ofof15.04%,
15.04%, which
which
R
isisaasignificant
significantincrease
increaseofof51.16%
51.16%from fromthat
thatofofthe
thesmooth
smooth casing
casing case. In In
case. Table 5 and
Tables 6, to6,
5 and
assess the the
to assess effectiveness of optimization,
effectiveness of optimization,the optimal
the optimaldesign is compared
design with awith
is compared new arefer-
new
ence design (nREF). For the nREF design, L , L , and α are set at the
reference design (nREF). For the nREF design, LF , LB , and α are set at the reference values
F B reference values in
Table
in Table1, while LM, β,
1, while LMand
, β, γand
are γfixed at the at
are fixed values shownshown
the values in Table in4.Table
The optimum design
4. The optimum
improves the stallthe
design improves margins of SC and
stall margins of SCnREF by 51.16%
and nREF and 9.22%,
by 51.16% and 9.22%,respectively. ThisThis
respectively. re-
markable
remarkable achievement
achievement validates
validates the benefit of performing
the benefit of performingoptimization.
optimization.In terms of accu-
In terms of
racy, the difference
accuracy, in the
the difference stallstall
in the margin
marginbetween
between thethe
prediction
prediction bybyMATLAB
MATLABand andRANS
RANS
simulationisis only
simulation only 0.07%.
0.07%. ThisThissmall
smallerror
errorindicates
indicatesthat thatthe
the combination
combinationof of the
the RBNN
RBNN
model and PSO algorithm is accurate in finding the optimal design
model and PSO algorithm is accurate in finding the optimal design for this problem. for this problem.
 Energies 2022, 15, 159 14 of 17
Energies 2022, 14, x FOR PEER REVIEW 14 of 17

Table 5. Optimization results.

Table 5. Optimization results.
Objective Function Improvement (%)
Design LF /CR (%) LB /CR (%) α (◦ ) Objective Function
Error (%) Improvement (%)
Design LF/CR (%) LB/CR (%) α (°)
MATLAB RANS Error (%)
vs. SC
MATLAB RANS vs. SC vs.
vs. nREF
nREF
SC - -
SC - - - - - 9.95
- 9.95- - - - --
nREF 45 70
nREF 45 45 70 - 45 13.77
- 13.77- - 38.39 38.39 --
PSO.opti 46.25 77.57 46.56 15.05 15.04 −0.07 51.16 9.22
PSO.opti 46.25 77.57 46.56 15.05 15.04 −0.07 51.16 9.22

Table
Table6.6.Comparison
Comparisonofofpeak
peakefficiency
efficiencyand
andpressure
pressureratio
ratiobetween
betweenoptimal
optimaldesign
designand
andother
othercases.
cases.

Changes
Changesvs.
vs.SC
SC(%)
(%)
Design
Design ηη
peak (%)
peak (%) PR
PRpeak
peak PR
PRNS
NS
ηηpeak PR peak
PRpeak PR
PRNS
peak NS
SC 83.85 2.0045 2.0820 - - -
SC 83.85 2.0045 2.0820 - - -
nREF
nREF 83.41
83.41 1.9778
1.9778 2.0785
2.0785 −0.52
−0.52 −1.33
−1.33 −0.17
−0.17
PSO.opti
PSO.opti 83.39
83.39 1.9562
1.9562 2.0762
2.0762 −0.55
−0.55 −2.41
−2.41 −0.28
−0.28

Figure 17 compares PSO.opti and nREF in terms of the rotor’s tip loading at the peak
Figure 17 compares PSO.opti and nREF in terms of the rotor’s tip loading at the peak
efficiency condition. At this condition, PSO.opti shows that a shockwave is located further
efficiency condition. At this condition, PSO.opti shows that a shockwave is located further
downstream, and the pressure difference at the upstream region is smaller compared to
downstream, and the pressure difference at the upstream region is smaller compared to that
that of nREF.
of nREF. As As a result,
a result, thethe flow
flow becomes
becomes more
more stable
stable andand the
the compressor’sstable
compressor’s stableworking
work-
ing range extends to a higher normalized mass flow rate of 0.9953 (the values for SC
range extends to a higher normalized mass flow rate of 0.9953 (the values for SC and nREF and
nREF are 0.9934
are 0.9934 and 0.9939,
and 0.9939, respectively).
respectively).

Figure
Figure17.
17.Blade
Bladeloading
loadingatat99%
99%rotor
rotorspan
spanatatpeak
peakadiabatic
adiabaticefficiency
efficiencycondition.
condition.

Figure18
Figure 18 shows
shows the
the Mach
Machnumber
numberdistribution
distribution at at
thethe
near-stall condition.
near-stall While
condition. there
While
is noissignificant
there difference
no significant between
difference the two
between the cases in theinrotor
two cases domain,
the rotor the stagnation
domain, zones
the stagnation
in theinstator
zones domain
the stator are smaller
domain than than
are smaller thosethose
in theincase
the of PSO.opti,
case which
of PSO.opti, is related
which to the
is related
toimprovement
the improvementin flowinstability. The combination
flow stability. of this change
The combination of this and the improvement
change and the improve-at the
peak efficiency condition mentioned above seems to be the main reason
ment at the peak efficiency condition mentioned above seems to be the main reason for for the higher stall
margin with the PSO.oti design.
the higher stall margin with the PSO.oti design.
 Energies
gies 2022, 14, x FOR2022,
PEER 15,REVIEW
159 15 of 17 15 of 17

Figure 18. Mach number

Figure contours
18. Mach at 98%
number blade span
contours atblade
at 98% near-stall
spanconditions.
at near-stall conditions.

5. Conclusions5. Conclusions
The effects ofThe
theeffects
RE2B of the RE2B
channel on thechannel on the aerodynamic
aerodynamic performanceperformance
and stabilityand stability of the
of the
single-stage single-stage
transonic axialtransonic axial compressor,
compressor, NASA StageNASA 37Stage
were37investigated
were investigated
using using
3D 3D RANS
RANS analysis.analysis. Numerical
Numerical resultsresults
confirmedconfirmed the RE2B’s
the RE2B’s ability
ability to to improve
improve the stability
the stability of of the
the transonictransonic
compressor,compressor,
with only with
smallonly small reductions
reductions in the efficiency
in the efficiency and pressure
and pressure ratio. ratio. The
The reference reference RE2B design
RE2B design increasedincreased the compressor’s
the compressor’s stall margin
stall margin fromfrom9.95% 9.95% of the smooth
of the
casing to 13.40% with a less than 1% reduction in the peak
smooth casing to 13.40% with a less than 1% reduction in the peak efficiency and pressure efficiency and pressure rise.
rise. As the conventional single-channel design (RE), RE2B showed the stabilizing effects effects on
As the conventional single-channel design (RE), RE2B showed the stabilizing
the compressor’s
on the compressor’s flow field:flow field: stagnation
Delaying Delaying stagnation zones’ and
zones’ expansion expansion
distancingandflow
distancing flow
separation
separation from from the
the leakage leakage
vortex. vortex.with
However, However, with thebleeding
the additional additional bleeding
channel, thechannel, the
RE2B channel removed a sizable stagnation zone near the rotor’s trailing edge shownedge
RE2B channel removed a sizable stagnation zone near the rotor’s trailing in shown in
the case of RE. This improvement might be the main reason the stall margin with RE2B iswith RE2B is
the case of RE. This improvement might be the main reason the stall margin
approximately approximately
8.33% higher8.33% than higher
that with than
RE.that with RE.
From the parametric study,
From the parametric study, the injection location, the injection location,
injection injection
angle, angle,
and rear and rear bleeding
bleeding
location are found to be the most influential parameters on the compressor’s performanceperformance
location are found to be the most influential parameters on the compressor’s
amongsixthe
among the tested tested six
geometric geometricofparameters
parameters of RE2B.those
RE2B. By selecting By selecting those three parame-
three parameters
ters as design
as design variables, variables, optimization
single-objective single-objective optimization
using using the
the PSO algorithm PSO algorithm
combined with combined
with the RBNN surrogate model was conducted to maximize the stall margin improve-
the RBNN surrogate model was conducted to maximize the stall margin improvement.
ment. With only 0.55% and 0.28% reductions, respectively, in the peak adiabatic efficiency
With only 0.55% and 0.28% reductions, respectively, in the peak adiabatic efficiency and
and maximum pressure ratio, the optimal design at LF /CR = 46.25%, LB /CR = 77.57%,
maximum pressure ratio, ◦the optimal design at LF/CR = 46.25%, LB/C R = 77.57%, and α =
and α = 46.56 , provided a large stall margin of 15.04%, i.e., 51.16% and 9.22% increases
46.56°, provided a large stall margin of 15.04%, i.e., 51.16% and 9.22% increases from those
from those of the smooth casing and the reference design, respectively.
of the smooth casing and the reference design, respectively.
In future work, the study of RE2B needs to be expanded to its effects on aerome-
In future work, the study of RE2B needs to be expanded to its effects on aeromechanic
chanic and aero-acoustic characteristics to provide a more comprehensive assessment of its
and aero-acoustic characteristics to provide a more comprehensive assessment of its ap-
applicability to real compressors.
plicability to real compressors.
Author Contributions: T.-D.V. presented the main idea of the dual-bleeding recirculation casing
Author Contributions: T.-D.V. presented the main idea of the dual-bleeding recirculation casing
treatment for the axial transonic compressor, analyzed the proposed casing treatment, and performed
treatment for the axial transonic compressor, analyzed the proposed casing treatment, and per-
numerical analysis. K.-Y.K. suggested improvements for the idea and revised and finalized the
formed numerical analysis. K.-Y.K. suggested improvements for the idea and revised and finalized
manuscript. All authors have read and agreed to the published version of the manuscript.
the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported by the National Research Foundation of Korea (NRF) grant
Funding: This work was supported by the National Research Foundation of Korea (NRF) grant
funded by the Korean government (MSIT) (No. 2019R1A2C1007657).
funded by the Korean government (MSIT) (No. 2019R1A2C1007657).
Institutional Review Board Statement: Not applicable.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not available.
Data Availability Statement: Not available.
Conflicts of Interest: The authors declare no conflict of interest.
Conflicts of Interest: The authors declare no conflict of interest.
Energies 2022, 15, 159 16 of 17

References
1. Hathaway, M.D. Self-recirculating casing treatment concept for enhanced compressor performance. In American Society of
Mechanical Engineers; International Gas Turbine Institute, Turbo Expo (Publication) IGTI: Amsterdam, The Netherlands, 2002;
Volume 5, pp. 411–420.
2. Koff, S.; Mazzawy, R. Case treatment for compressor blades. US005282718A, 1 February 1994.
3. Hobbs, D.E.; Windsor, S. Active Vaned Passage Casing Treatment. US005431533A, 11 July 1995.
4. Nolcheff, N.A. Flow Aligned Plenum Endwall Treatment for Compressor Blades. US005586859A, 24 December 1996.
5. Strazisar, A.J.; Bright, M.M.; Thorp, S.; Culley, D.E.; Suder, K.L. Compressor stall control through endwall recirculation. In
Proceedings of the ASME Turbo Expo 2004, Vienna, Austria, 14–17 June 2004; Voluem 5, pp. 655–667.
6. Khaleghi, H.; Tousi, A.M.; Boroomand, M.; Teixeira, J.A. Recirculation casing treatment by using a vaned passage for a transonic
axial-flow compressor. Proc. Inst. Mech. Eng. Part A J. Power Energy 2007, 221, 1153–1162. [CrossRef]
7. Weichert, S.; Day, I.; Freeman, C. Self-regulating casing treatment for axial compressor stability enhancement. In Proceedings of
the ASME Turbo Expo 2011, Vancouver, BC, Canada, 6–10 June 2011; Volume 7, pp. 225–238.
8. Khaleghi, H. Effect of discrete endwall recirculation on the stability of a high-speed compressor rotor. Aerosp. Sci. Technol. 2014,
37, 130–137. [CrossRef]
9. Wang, W.; Chu, W.; Zhang, H.; Wu, Y. The effects on stability, performance, and tip leakage flow of recirculating casing treatment
in a subsonic axial flow compressor. In Proceedings of the ASME Turbo Expo, Seoul, Korea, 13–17 June 2016; Volume 2A-2016.
10. Dinh, C.T.; Ma, S.B.; Kim, K.Y. Effects of a circumferential feed-back channel on aerodynamic performance of a single-stage
transonic axial compressor. Proc. ASME Turbo Expo 2017, 2A-2017, 57–65. [CrossRef]
11. Li, J.; Du, J.; Li, Z.; Lin, F. Stability enhancement with self- recirculating injection in axial flow compressor. J. Turbomach. 2018, 140,
071001. [CrossRef]
12. Khaleghi, H. A new approach of endwall recirculation in axial compressors. Aerosp. Sci. Technol. 2020, 98, 105704. [CrossRef]
13. Vuong, T.; Kim, K.; Dinh, C. Recirculation-groove coupled casing treatment for a transonic axial compressor. Aerosp. Sci. Technol.
2021, 111, 106556. [CrossRef]
14. Samad, A.; Kim, K.Y. Shape optimization of an axial compressor blade by multi-objective genetic algorithm. Proc. Inst. Mech. Eng.
Part A J. Power Energy 2008, 222, 599–611. [CrossRef]
15. Deb, K.; Pratap, A.; Agarwal, S.; Meyarivan, T. A Fast and Elitist Multiobjective Genetic Algorithm: NSGA-II. IEEE Trans. Evol.
Comput. 2002, 6, 182–197. [CrossRef]
16. Lucas, J.M. Response Surface Methodology: Process and Product Optimization Using Designed Experiments, 3rd edition. J. Qual.
Technol. 2010, 42, 228–230. [CrossRef]
17. Deb, K. Multi-objective Optimisation Using Evolutionary Algorithms: An Introduction. In Multi-Objective Evolutionary Optimi-
sation for Product Design and Manufacturing; Wang, L., Ng, A.H.C., Deb, K., Eds.; Springer: London, UK, 2011; pp. 3–34. ISBN
978-0-85729-652-8.
18. Kim, J.-H.; Ovgor, B.; Cha, K.H.; Kim, J.H.; Lee, S.; Kim, K.Y. Optimization of the aerodynamic and aeroacoustic performance of
an axial-flow fan. AIAA J. 2014, 52, 2032–2044. [CrossRef]
19. Khalfallah, S.; Ghenaiet, A.; Benini, E.; Bedon, G. Surrogate-Based Shape Optimization of Stall Margin and Efficiency of a
Centrifugal Compressor. J. Propuls. Power 2015, 31, 1607–1620. [CrossRef]
20. Ma, S.B.; Afzal, A.; Kim, K.Y. Optimization of ring cavity in a centrifugal compressor based on comparative analysis of
optimization algorithms. Appl. Therm. Eng. 2018, 138, 633–647. [CrossRef]
21. Goldberg, D.E. Genetic algorithms in search, optimization, and machine learning. Choice Rev. Online 1989, 27, 27-0936. [CrossRef]
22. Kennedy, J.; Eberhart, R. Particle swarm optimization. In Proceedings of the ICNN’95-International Conference on Neural
Networks, Perth, Australia, 27 November–1 December 1995; Volume 4, pp. 1942–1948.
23. Van Laarhoven, P.J.M.; Aarts, E.H.L. Simulated Annealing BT-Simulated Annealing: Theory and Applications; van Laarhoven, P.J.M.,
Aarts, E.H.L., Eds.; Springer: Dordrecht, The Netherlands, 1987; pp. 7–15. ISBN 978-94-015-7744-1.
24. Cheng, J.; Zhao, S.; Dong, Z.; Yang, C. Design Optimization of a Multistage Axial Flow Compressor Based on Full-Blade Surface
Parameterization and Phased Strategy. Int. J. Aerosp. Eng. 2021, 2021. [CrossRef]
25. Jinxin, C.; Chen, J.; Xiang, H. Aerodynamic Optimization Design of Compressor Blades Based on Improved Artificial Bee Colony
Algorithm. In Proceedings of the 2018 Joint Propulsion Conference, Cincinnati, OH, USA, 9–11 July 2018; AIAA Propulsion and
Energy Forum. American Institute of Aeronautics and Astronautics: Reston, VA, USA, 2018.
26. Yan, S.; Chu, W.L. Influence of self-circulating casing treatment with double-bleed ports structure on compressor performance.
Proc. Inst. Mech. Eng. Part G J. Aerosp. Eng. 2020, 234, 1743–1756. [CrossRef]
27. Lonnie, R.; Moore, R.D. Design and Overall Performance of Four Highly Loaded, High-Speed Inlet Stages for an Advanced
High-Pressure-Ratio Core Compressor. In NASA Technical Paper 1337; NASA Lewis Research Center: Cleverland, OH, USA, 1978.
28. Wilke, I.; Kau, H.P.; Brignole, G. Numerically aided design of a high-efficient casing treatment for a transonic compressor. In
Proceedings of the ASME Turbo Expo, Reno-Tahoe, NV, USA, 6–9 June 2005; Volume 6, pp. 353–364.
29. Hofmann, W.; Ballmann, J. Tip clearance vortex development and shock-vortex-interaction in a transonic axial compressor rotor.
In Proceedings of the 40th AIAA Aerosp. Sci. Meet. Exhib, Reno, NV, USA, 14–17 January 2002. [CrossRef]
30. ANSYS. ANSYS CFX-Solver Theory Guide Release 15.0; ANSYS: Canonsburg, PA, USA, 2014.
Energies 2022, 15, 159 17 of 17

31. Shim, H.-S.; Kim, S.-H.; Kim, K.-Y. Analysis and Optimization of Staggered Partial Diffuser Vanes in a Centrifugal Pump. J. Fluids
Eng. 2020, 142, 051207. [CrossRef]
32. Barth, T.; Jespersen, D. The design and application of upwind schemes on unstructured meshes. In Proceedings of the 27th
Aerospace Sciences Meeting; Aerospace Sciences Meetings, Reno, NV, USA, 9–12 January 1989; American Institute of Aeronautics
and Astronautics: Reston, VA, USA, 1989.
33. Chen, H.; Huang, X.; Fu, S. CFD investigation on stall mechanisms and casing treatment of a transonic compressor. In Proceedings
of the Collection of Technical Papers-AIAA/ASME/SAE/ASEE 42nd Joint Propulsion Conference, Sacramento, CA, USA, 9–12
July 2006; Volume 6, pp. 4789–4798.
34. Kim, J.H.; Choi, K.J.; Kim, K.Y. Performance evaluation of a transonic axial compressor with circumferential casing grooves. In
Proceedings of the Institution of Mechanical Engineers. Part A J. Power Energy 2012, 226, 218–230. [CrossRef]
35. Dinh, C.T.; Heo, M.W.; Kim, K.Y. Aerodynamic performance of transonic axial compressor with a casing groove combined with
blade tip injection and ejection. Aerosp. Sci. Technol. 2015, 46, 176–187. [CrossRef]
36. Celik, I.; Karatekin, O. Numerical experiments on application of richardson extrapolation with nonuniform grids. J. Fluids Eng.
Trans. ASME 1997, 119, 584–590. [CrossRef]
37. McKay, M.D.; Beckman, R.J.; Conover, W.J. A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis
of Output from a Computer Code. Technometrics 1979, 21, 239–245. [CrossRef]
38. MATLAB. The Language of Technical Computing-Release 14; Mathworks: Natick, MA, USA, 2004.