JOURNAL OF MECHANICAL AND TRANSPORT ENGINEERING

Vol. 67, no. 3 2015

DOI 10.21008/j.2449-920X.2015.67.3.06

Bartosz ZIEGLER*

BOUNDARY LAYER CONTROL

IN TRANSONIC COMPRESSOR FLOWS

This article presents a novel concept for coupled boundary layer control in transonic ax-
ial compressors. Proposed control scheme utilizes two steady active methods and couples
them creating compact and in principle passive boundary layer control method. Both pre-
sented methods (steady suction and steady tangential blowing) are known since many years.
In this article they are presented in the context of coupling them on two neighbouring
stages. Presented analysis shows that the proposed solution allows for very high stage load-
ings and pressure ratios. The concept is presented on the set of transonic cascades, where it
shows to prevent separation in regions of high adverse pressure gradients (associated both
with occurrence of shockwaves and high aerodynamic loading).

Keywords: transonic flows, compressors, boundary layer control

1. SHOCK INDUCED SEPARATION AND LIMITATIONS IT

INTRODUCESTO COMPRESSOR FLOWS

1.1 Shock boundary layer interaction and its effects

Shocks are almost dicontinous changes in fluid parameters (velocity, pressure,

temperature etc.). Shocks occur in flows with relative velocities greater than the
speed of sound (M>1) and are associated with compression of the fluid flowing
through a shock (in gasses, expansion cannot occur through shocks, because it would
violate the second law of thermodynamics). In most internal and external flows (es-
pecially those of engineering application), shockwaves interact with boundary layers
near channel walls. The nature of this interaction was recognized and described dec-

* Chair of Thermal Engineering, Poznan University of Technology.

54 Bartosz Ziegler

ades ago, but is still of major interest in the aerospace and turbomachinery engineer-
ing. Of special importance is the behaviour of boundary layers downstream of tran-
sonic shocks. Such phenomena occure in passages of compressors and turbines, on
surfaces of transonic aircrafts (they for example cause substantial part of drag of
modern passenger aircraft wings) and other aerospace applications. Experimental
studies show the complexity and nature of transonic shock – boundary layer interac-
tions [9]. Whenever a shock interacts with the boundary layer (thin region of flow
near the wall where the flow velocity gradually decreases to zero at the wall surface),
it creates a complex flow pattern (Fig. 1 – shows SBLI in the case of flow reattaching
immediately donstream of shock).In the upstream of shock boundary layer, total
pressure is lower than outside boundary layer (because of lower velocity), hence
increase in static pressure imposed by the shock causes thickening of the boundary
layer streamtubes. Some part of the upstream boundary layer, has stagnation pressure
lower than the static pressure downstream of shock. That part of the stream without
shear forces from outer layers, would not be able to pass behind the shock. It there-
fore creates a separation bubble (region of recirculation, marked SB on Fig. 1).

Fig. 1. Simplified flow pattern for SBLI with reattachment downstream of shock, MS –
main normal shock, OS – oblique shock, WS – weaker normal shock, CF – compression
fan, SB – separation bubble, SL – slip line, d – boundary layer edge
The relation between rate of momentum added to the slowest stream tubes from
outer layers, and rate of pressure rise, determines whether the flow downstream of
shock stays separated or reattaches. That is why, for the same shock strength, laminar
boundary layers separate easier than turbulent ones (high level of turbulence increase
the rate of shear forces between outer and inner layers of BL). If the pressure rise is
sufficient, than flow stays separated, especially on convex walls.In cases where the
separation reattaches, the boundary layer behind the interaction is much thicker, and
has high deficit of momentum which makes it prone also to flow separation caused
by adverse pressure gradients downstream of the shock. Those phenomena are typi-
cal for all high Mach number diffusers (transonic or supersonic channels where flow
 Boundary layer control in transonic compressor flows 55

is decelerated and compressed). Those are especially supersonic inlets, but also
channels of dynamic compressors experiencing globally or locally supersonic flows.
Flow separation (especially occurrence of big separated regions) is a very undesir-
able phenomenon in compressor and turbine flows.
It is reducing the efficiency (due to excessive amounts of kinetic energy being
cascaded to turbulence kinetic energy and then dissipated to heat by viscous shear)
and limiting the blade loading, thus limiting the amount of work that can be ex-
changed with the fluid.Shock boundary layer interaction with separation also intro-
duces additional unsteadiness to the flow field [4]. Separated regions influence shock
positions and its oscillations introduce periodic changes in aerodynamic loading of
the blade forcing undesirable vibrations in wide range of frequencies.

1.1 Flow separation in transonic axial compressor flows

Modern axial compressors of turbine engines (especially designated for aerospace

applications) work very often in transonic flow regime. This is due to the fact that the
relative flow Mach numbers determine the relative flow dynamic pressure (which is
the theoretical limit of pressure rise that can occur in a single channel). The relation
between the total pressure (which is the pressure the gas would have if it was isen-
tropicaly decelerated to zero velocity) and flow Mach number is given by relation (1)

(1)
where p0 – is the total pressure, p – is the value of static pressure, g– is the specific
heat capacity ratio, M – is the flow Mach number. This relation shows that the limit
of pressure rise that can be achieved in single diffuser is quickly rising with the Mach
number. Hence to achieve high stage pressure ratios it is essential to operate in tran-
sonic flow regime (with passages having supersonic inlets and subsonic outlets).
Such change in flow parameters in practice cannot be achieved without a shock or
a series of shocks. In fact, most of the compression occurring in passages of modern
transonics axial compressors and fans occurs on transonic shocks. Presence of those
shocks however, limits the amount of subsonic compression that can be achieved
downstream of the shock due to the phenomenon of SBLI and associated flow sepa-
rations. That’s why airfoils of modern transonic compressors are thin and only
slightly cambered. Downstream of SBLI, any significant adverse pressure gradient
would cuase flow separation, so effective pressure increase is impossible.In the case
of separated flow, pressure rise downstream of separation point is very small and
associated with high total pressure losses.
56 Bartosz Ziegler

2. STEADY METHODS OF ACTIVE BOUNDARY LAYER CONTROL

FOR TRANSONIC COMPRESSORS

Boundary layer control, is a term used to a wide range of methods used to ma-
nipulate the flow in the near wall region. In the case of transonic diffusers (so also
the passages of transonic axial compressors) methods in scope of interest will be
those that can weaken the shock boundary layer interaction, and delay flow separa-
tion. In order to achieve those, the layer of the flow with highest momentum deficit
(closest to the wall) has to be removed, or reaccelerated. Steady suction, isamethod
that has been used since decades in controlling transonic diffuser flows, mainly in
case of supersonic aircraft engine inlets. It relies on removing the boundary layer
flows through, slots, holes or porous surfaces. In case of a compressor the removed
flow has to be carried out through cavities in blades, hubs and casing. In few re-
searches [3, 6, 8] it was investigated as a boundary layer control method for highly
loaded axial compressors. Especially in research by Kerrebrock [6] from MIT
GTL, it was shown that in the case of transonic design, by removing boundary
layer flows the specific work of a compressor can be nearly doubled in respect to
designs without boundary layer control. The loading achieved by aspirated com-
pressor in study [6] was around 0.7 which for the tested design (blade tip speed 457
m/s) allowed for total pressure ratio of 3.17 [6].The second method investigated in
the scope of this article is the steady tangential blowing. This method was investi-
gated in the past mainly in the context of wing circulation control but has the ad-
vantage of being potentially able to utilize the fluid removed through steady suc-
tion.Previous research on the same concept [11] showed that it ispossible to
achieve loadings around 0.8 for 2D flows in the range of relative inlet Mach num-
bers around 1.4 [11]. In this study it will be shown that it is also possible at lower
relative mach numbers (around 1.2) and less complex blowing slot arrangements
(with only one blowing slot downstream of the shock). The concept (which was
also investigated in [11]) assumes use of air removed from compressor stage by
steady suctoion as a blowing medium on a preceeding stage (Fig. 2). Altough both
methods (steady suction and blowing) are considered active methods, when they
are coupled they create a passive method, which means that no external source of
energy is required to operate the boundary layer control loop. Such configuration
has the advantage of not requiring any additional apparatus to generate suction and
carry the removed flux out of the compressor. Another advantage is that it is a
zero-net mass flux method, which unlike in the case of steady suction boundary
layer control makes the compressor outflow flux equal to the inflow flux.
 Boundary layer control in transonic compressor flows 57

Fig. 2. Layout of coupled boundary layer control on two neighbouring compressor stages,
R1 and R2 indicates rotor cascades and S1 and S2 indicate stator cascades

The inlet flow of blowing controlled stage and outflow of suction controlled
stage are equal for each set of stages with coupled boundary layer control. Between
those stages, mass flow rate is higher by the amount recirculating in the boundary
layer control loop. In case of pure boundary layer suction, few percent of the flow
is removed from each stage making the outflow flux significantly lower than at
compressor inlet, especially if boundary layer suction was to be applied on multiple
stages. The biggest challenge in realising the boundary layer control in axial com-
pressor in a coupled manner according to the proposed concept, is that pressure
ratio between both stages has to be sufficiently high to allow for spontaneous flow
of the boundary layer control flow.

3. NUMERICAL INVESTIGATION ON TRANSONIC CASCADE FLOWS

WITH STEADY BOUNDARY LAYER CONTROL

3.1. 2D test cases and numerical setup

In order to present the influence of both control methods on the transonic cas-
cade flowfield, two two-dimensional cascades were designed and analysed numeri-
cally. Those two cascades were designed to be able to work in series on two
neighbouring stages. Geometries are tailored to work at the same tangential veloci-
ties of 360 m/s (like if theye were two succeeding stages of a multistage compres-
sor). First cascade controlled by steady tangential blowing, is designed for inlet
with standard atmospheric temperature (288K). At such conditions, and inflow
58 Bartosz Ziegler

angle of a1=30° (angle measured from the tangential direction) it is calculated to

have relative inlet Mach number of around 1.25 (this number will be lower because
of the leading edge shock, not taken into account). It is assumed that the cascade
will be able to slow the flow down to relative Mach number of around 0.45 (local
diffusion factor DF= 0.6 for one dimensional simplification). From Mach number-
Area relation, an outflow anglea2 can be determined:

(2)

where is the isentropic ratio of flow area to critical area:

(3)
and is a dimensionless channel pressure recovery ratio (ratio of outlet rela-
tive total pressure to inlet relative total pressure, assumed 0.96 for first cascade and
0.98 for second). Knowing the relative total temperature, outflow angle and relative
Mach number, specific work of a cascade can be determined (through standard
gasdynamic relations, hence omitted in this paper). The resulting total enthalpy
temperature and pressure are taken as inlet parameters for second cascade. The
anticipated flow parameters at this level of simplification are shown in Table
1.Because of higher temperature (resulting in lower Mach numbers for the same
velocity) second cascade is given higher inflow angle. For both geometries, mul-
tiblock structured meshes were created using ICEM CFD software. A high stagger
angle topology was used for both geometries.

Table 1
Selected design parameters of both cascade flows
Blowing controlled cascade Suction controlled cascade
Parameter
inlet outlet inlet outlet
Total pressure [kPa] pamb= 101,325 239,1 239,1 535,0
Total temperature [K] 288 371 371 467
Relative Mach number 1.25 0.45 1.15 0.4
Relative flow angle [°] 28 42 33 60
Absolute flow angle [°] 90 25 90 27
 Boundary layer control in transonic compressor flows 59

Fig. 3. Blowing controlled cascade mesh and boundary conditions

In both cases the blade was surrounded by a C-grid sub-topology allowing for
precise control of the boundary layer node distribution in normal to wall direction.
In order to keep the y+ parameter close to unity, the C-grid boundary layer mesh
has element sizes in normal to wall direction decreasing exponentially towards the
wall to values in the order of 10-6 m. It was later shown that the y+ was less than
1 for almost whole blade surface.In both cases, additional mesh blocks were added
for blowing/suction slots plenums, transforming smoothly into the boundary layer
regions. The computational setup with exemplary mesh and boundary conditions is
shown on Fig. 3 for the blowing controlled cascade. Nodes on periodic faces are
conformal to avoid interface interpolations. Beside pronounced mesh density in-
crease in the boundary layer region, two denser (in streamwise direction) regions
can be observed. First one (in the direction from the inlet to outlet) is created to
possibly accurately capture the transonic shock, second corresponds to the edge of
the blowing slot. Inlet boundary condition is a “total pressure inlet” condition, with
specified total pressure, total temperature, flow direction (axial in stationary refer-
ence frame), and negligible (0.5%) turbulence level.At cascade outlet “pressure
outlet” boundary condition is used with specified static pressure value. Same type
of boundary conditions (total pressure inlets and pressure outlets) are used for suc-
tion controlled case and blowing/suction slots. Calculations were made using
ANSYS Fluent density based (coupled) implicit solver. The gas was modelled as
calorically perfect, with temperature dependent molecular viscosity (Sutherlands
formula). Layout of used settings is included in Table 2. Turbulence model chosen
for this calculation was a 2-equation SST model [7] with boundary layer resolving
approach (as mentioned earlier first node distance from the wall corresponds to
y+»1 which should place it well in the viscous sublayer). It was chosen for its abil-
60 Bartosz Ziegler

ity to predict flow separation and reattachment phenomena, as shown by many

computational studies (for example [1, 10]). Study shown in ref. [2] is performed
on the case of a Sajben transonic diffuser [10], which is a transonic diffuser, with
normal shock occurring at similar as in the investigated flows Mach numbers
(around 1.2 to 1.3), and similar Reynolds number range (few millions).In study
[10], Menter’s SST model showed to be most accurate (from investigated) at pre-
dicting separation position. SST has also the advantage of having the tendency to
overpredictseparated region length and boundary layer thicknessdownstream of
shock. SST should therefore predict performance worse than in real flow and it is
therefore chosen as a safe assumption model. Calculated flowfields exhibit many
instabilities (both physical and numerical)so the first order upwind discretization
scheme was applied for its stability and convergence capabilities.

Table 2
Selected Fluent solver settingsused in calculation
Setting Chosen model/algorithm
Time dependence steady state
Turbulence model SST with compressibility effects and curvature correction
Compressibility ideal gas law
Viscosity Sutherlands formula
Frame motion moving frame, vy= -360 m/s
Spatial discretisation 1st order upwind
Convective fluxe type Roe flux difference splitting scheme
Initialization FMG (fast multigrid) after standard uniform initialization
Flow variables (pressure,
coupled block Gauss-Seidel method with algebraic multigrid
velocity, density) solving
(AMG)
algorithm
Other scalars (turbulence
segregated Gauss-Seidel with AMG
parameters)
scaled residuals below 10-4 + stabilisation of shock position
Convergence criteria
+ mass imbalance less than 0.5%

A multiple convergence criteria was used. Scaled residuals– (according to [5])

in density based solvers, for a conservation variable – W are defined as:

(4)
where:
 Boundary layer control in transonic compressor flows 61

(5)
From equations (4) and (5) it can be seen that for oscilating flows (and real tran-
sonic cascades flows, are usually characterised by shock position oscillations caus-
ing other downstream oscillations) scaled residuals can have local minima near
extreme schock positions (just like a velocity of a pendulum is zero at its maximum
deflection). Hence mass conservation over inlets/outlets and stabilisation of shock
position were also monitored.

3.2 Blowing controlled cascade results and discussion

The flowfield obtained for first cascade was found by iteratively rising the back-
pressure till shock stabilised near the passage inlet (overlapping of the leading edge
shock with passage normal shock). Analysis of wall shear stress shows that the
flow is attached on the whole blade surface (positive shear stress in flow direction
on both pressure and suction surface). Lack of shock boundary layer interaction
separation bubble indicates that the pressure gradient at shock foot is reduced. A
low momentum region can be observed near suction surface trailing edge (relative
Mach number profiles at Fig. 4) This low momentum region is not adjacent to
a blade surface, it appears between still attached flow from the blowing slot and
passage core flow.The leading edge shock reaches blade surface upstream of the
blowing slot. Because of the nonuniform pressure distribution downstream of the
shock, shock bends in streamwise direction and weakens, thus weakening the inter-
action with suction surface boundary layer. Even though the interaction is weak,
a significant increase in boundary layer thickness occurs. To ilustrate the mecha-
nism of boundary layer enhancement by tangential blowing, velocity profiles near
suction surface wall are taken from upstream the slot (just after shock position),
slot position and downstream position (roughly half way between slot and trailing
edge).
Profiles of relative velocity – w (normalized by inlet critical velocity in moving
frame)are shown on Fig. 5 together with their position on the blade surface. Dis-
tance from the wall – h, on Fig.5 is normalised by blade chord – C. Profiles down-
stream of the slot (stations 2 and 3) are shifted downwards by the value of slot
height to match the mixing layer height to h = 0 (which is the wall height upstream
of the slot).Its worth to mention, that blowing slot operates at supercritical pressure
ratio so its flow is choked, and it operates as an uderexpanded nozzle so a little bit
more acceleration takes place just after station 2 (to about 1.3 of inlet critical veloc-
ity normalised value).
62 Bartosz Ziegler

Fig. 4. Contours of relative Mach numer and static pressure of blowing controlled cascade

Fig. 5. Velocity profiles at three stations on the blade suction surface

3.3 Suction controlled cascade results and discussion

Second cascade was designed to have two slots removing boundary layer flows,
It was observed in first calculations with only one slot, that after removing bound-
ary layer after its interaction with shock, resulting boundary layer would still sepa-
rate too quickly if exposed to high adverse pressure gradient (which is necesseray
downstream of the shock to achieve high diffusion on a limited channel length).
That’s why a second smaller slot was added roughly halfway the distance between
 Boundary layer control in transonic compressor flows 63

first slot and trailing edge. Dual slot approach allowed for maintaining high pres-
sure gradients at whole distance between slots. The design of second slot however
is far from optimal because it introduces too big pressure gradaients in its vicinity
(what can be observed as pressure contour lines focusing on second slot position on
Fig. 6) and removes too small mass flux. In effect the flow separates not far down-
stream of the second slot. When the flow separates, the pressure rise stops. Fig.7
shows plots of wall shear stress and pressure at near trailing edge part of blade
suction surface. The x position is normalized by the x position of trailing edge –xTE.

Fig. 6. Relative Mach number and pressure contours for suction controlled case

Fig. 7. Pressure and wall shear stress near trailing edge at blade suction surface
64 Bartosz Ziegler

Around x/xTE= 0.79 the wall shear stress reaches zero, which determines the flow
separation point, negative values of wall shear stress shows the adjacent to wall
backflow region of a separation bubble. It can be observed how the flow separation
stops the pressure rise, and the static pressure levels atp/p0,1,rel around 0.83. This
separation limits the diffusion of the suction controlled cascade.
Because the suction controlled cascade has inlet relative Mach numbers so close
to unity (the mass flow averaged relative Mach number at inlet is 1.13 which corre-
sponds to flow area only 1.3% larger than critical) the suction surface in front of
blade to blade passage entrance is highly convex causing strong supersonic expan-
sion in front of a passage shock. Such design was intentional to prevent the flow-
field from unstarts (in internal flows normal shock position is stable only in accel-
erating flows). This however causes the flow upstream of normal shock to have
high Mach numbers of around 1.35 to 1.4 near suction surface. Hence stronger
shock than in the case of blowing controlled cascade.

3.4 Cascades performance

To summarize the results, and compare them with design assumptions, the flow
parameters are averaged over inlet and outlet planes. All parameters mentioned
further are averaged over mass flow rate which means that for a variable f its mass
weighted average is defined as [5]:
(6)
where: r– is density, v– is velocity vector, dA– is a surface normal vector element,
– is a mass flow rate.Values of selected parameters dimensionless values are
included in Table 3. In Table 3, it may appear odd that for blowing controlled cas-
cade outlet relative total temperature is higher than inlet value. This difference is
caused by addition of much hotter stream by the tangential blowing jet. Another
table (Table 4) shows parameters of boundary layer controlling flows.
Last table (Table 5) shows cascades selected performance. Those are total pres-
sure ratio–p, local diffusion factor (in simplified form) defined as:

(7)
where wmax – is a maximum relative velocity in the flow field, w2– is cascade outlet
average relative velocity. Another parameter is cascade loading which is defined as:
(8)
where is a total enthalpy rise and u – is cascade velocity. Last parameter is
efficiency. In this case, because inlet and outlet flow rates of cascades are not iden-
tical, general definition of efficiency must be applied:
 Boundary layer control in transonic compressor flows 65

(9)
where: – is the total enthalpy flux rise across all inlets-outlets flow, – mass
flow without boundary layer control (throughflow flux, that isinlet for cascade with
blowing, outlet for cascade with suction), Cp – specific heat capacity at constant
pressure, p– total pressure ratio, g– specific heat capacity ratio.

Table 3
Dimensionless values of mass averaged flow parameters in cascade inlets and outlets
First cascade Second cascade
Averaged (blowing) (suction)
Normalised by
parameter
inlet outlet inlet outlet
Static pressure 0.81 1.68 1.81 3.55
First cascade inlet
Total pressure 1.00 2.38 2.38 5.02
total pressure
Relative total pressure 1.98 1.91 4.09 3.98
Static temperature 0.98 1.23 1.25 1.52
First cascade inlet
Total temperature 1.00 1.35 1.35 1.69
total temperature
Relative total temperature 1.27 1.28 1.57 1.57
First cascade inlet
Relative velocity 1.31 0.53 1.40 0.56
critical velocity
Relative Mach number – 1.21 0.44 1.13 0.41

Table 4
Dimensionless values of parameters in boundary layer controlling flows
Averaged
Normalised by Blowing Suction
parameter
Static pressure First cascade inlet 2.31 2.42
Relative total pressure total pressure 2.40 2.98
Static temperature First cascade inlet 1.57 1.60
Relative total temperature total temperature 1.55 1.51
cascade outflow for blowing
Mass flow [%] 3.18 2.70
cascade inlfow for suction
66 Bartosz Ziegler

Table 5
Cascade characteristic dimensionless parameters
Parameter Blowing Suction
Total pressure ratio –p 2.38 2.11
Diffusion factor –DF 0.66 0.66
Cascade loading –y 0.78 0.75
Cascade efficiency – h [%] 95.1 93.7
Chord based Reynolds number –ReC 2.8×106 4.1×106

It has to be mentioned that unusually high values of efficiency in Table 5, are

associated with the fact that those are only moving cascade efficiencies that does
not take into account losses in compressor stationary vanes. Also many phenomena
responsible for major part of losses in real compressors are not accounted in pre-
sented analysis. Phenomena like tip leakage and vortex, hub and casing boundary
layers evolution and secondary flows are absent in two dimensional flows. Because
of that, efficiencies presented in Table 5 are not to be compared with efficiencies of
real compressors.

4.CONCLUSIONS

On the course of numerical analysis, it was shown that both methods steady
tangential blowing and suction are capable of enabling transonic compressor cas-
cades to operate at very high loadings and diffusion factors beyond the possibilities
for uncontrolled flows. It was shown that tangential blowing performs very well in
controlling transonic passage flows which is beyond flow regime common for that
method of boundary layer control.The novel concept of coupled scheme for bound-
ary layer control presented in this paper is shown to be feasible, as pressures re-
quired for suction-blowing boundary layer control are within the pressure ratio
occurring between two subsequent stages. Compressors designed according to the
proposed concept, could achieve required pressure ratios on roughly half the num-
ber of stages required in case of a conventional design which could significantly
improve turbine engine power (or thrust) to weight ratio, and reduce its axial di-
mensions. However, it is essential for the concept of coupled suction-blowing
boundary layer control to further increase achievable aerodynamic loading, and
establish a blowing-suction coupling plenum design of minimal pressure losses, as
the reserve of relative total pressure between neighbouring stages could be insuffi-
cient at off design conditions. Presented coupled method may thus appear to be
most feasible for compressors requiring narrow operating ranges. It is possible
 Boundary layer control in transonic compressor flows 67

thatfor turbine engine application, coupling neighbouring stages would be difficult

and it would be more feasible to couple stages further away in a multistage com-
pressor. The order of mass flow fractions used for boundary layer control and di-
mensionless performance (especially diffusion factors and loadings) are similar to
what was shown by other reserchers investigating boundary layer control in tran-
sonic compressor flows.

NOMENCLATURE

a – speed of sound a – flow angle, measured from tangential direction

A – area g – specific heat capacity ratio
BL – boundary layer r – density
C – blade chord t – wall shear stress
H – specific enthalpy/distance SUBSCRIPTS AND SUPERSCRIPTS:
normal from wall
M – Mach number 0 – stagnation parameter
p – pressure 1 – cascade inlet
SBLI – shock boundary layer 2 – cascade outlet
interaction amb – ambient conditions (288,15 K; 101325 Pa)
T – temperature rel – in reference frame associated with cascade
v – velocity TE – trailing edge
w – relative flow velocity * – critical parameter
x – axial direction position

REFERENCES

[1] Apsley D.D., Leschziner M.A., Advanced Turbulence Modelling of Separated

Flow in a Diffuser Flow, Turbulence and Combustion, 1999, vol. 63,
DOI:10.1023/A:1009930107544.
[2] Bogar T.J., Sajben M., Kroutil J.C., Characteristic Frequencies of Transonic
Diffuser Flow Oscillations, AIAA Journal, 1983, vol. 21, no. 9, p. 1232–1240.
[3] Colombo E. et al., Investigations on axial compressor cascades with aspiration on
blades and hub, in: European Workshop on New Aero Engine Concepts, Munich
2010.
[4] Dussauge J. et al., Unsteadiness in shock wave boundary layer interactions with
separation, Aerospace Science and Technology, 2006, vol. 10, DOI:
10.1016/j.ast.2005.09.006.
[5] Fluent documentation at http://aerojet.engr.ucdavis.edu/fluenthelp (accesed:
27.02.2015).
68 Bartosz Ziegler

[6] Kerrebrock J., Final Technical Report on AFOSR Grant F49620-98-1-0493

entitled Research and Development on Aspirated Compressors, Cambridge, Gas
Turbine Laboratory De-partment of Aeronautics and Astronautics MIT, 2003.
[7] Menter F. R., Two-equation eddy-viscosity turbulence models for engineering
applications, AIAA Journal, 1994, vol. 32, no. 8, p. 1598–1605. DOI,
10.2514/3.12149.
[8] Merchant A. et al., Experimental Investigation of a High Pressure Ratio Aspirated
Fan Stage, ASME Journal of Turbomachinery, 2005, 127(1), p. 43–51 DOI,
10.1115/1.1812323.
[9] Seddon J., The Flow Produced by Interaction of a Turbulent Boundary Layer with
a Normal Shock Wave of Strength Sufficient to Cause Separation, UK
Aeronautical Research Council Reports and Memoranda No. 3502, London 1967.
[10] Yoder D.,Sajben Transonic Diffuser: Study #2, Ohio, NASA Glenn Research
Centre, 2008,http://www.grc.nasa.gov/WWW/wind/valid/transdif/transdif02/trans
dif02.html (accessed: 27.02.2015).
[11] Ziegler B., Transonic axial compressor active boundary layer control concept and
its evaluation, ZeszytyNaukowePolitechnikiRzeszowskiej, 2014, 290, Mechanika,
86.

KONTROLA WARSTWY PRZYŚCIENNEJ

W SPRĘŻARKACH TRANSONICZNYCH

Streszczenie

W artykule zaprezentowano nowatorską koncepcję sprzężonej kontroli warstwy przy-

ściennej wtransonicznych sprężarkach osiowych. Koncepcja ta jest oparta na dwóch zna-
nych metodach aktywnych (stacjonarnym odsysaniu i stycznym nadmuchu), z których
utworzono pasywną metodę kontroli warstwy przyściennej.Metoda ta polega na wykorzy-
staniu strumienia odsysanego z jednego stopnia sprężarki wielostopniowej jako źródła
nadmuchu na stopniu poprzednim. Analiza wykazała, że proponowane rozwiązanie pozwa-
la osiągać duże obciążenia i duży spręż stopnia sprężarki osiowej. Koncepcjęzaprezento-
wano na przykładzie przepływów przez dwie transoniczne palisady sprężarkowe – kontrola
warstwy przyściennej zapobiega separacji przepływu (zarówno indukowanego faląuderze-
niową,jak i związanego ze znacznym obciążeniem aerodynamicznym).