ZESZYTY NAUKOWE POLITECHNIKI RZESZOWSKIEJ 290, Mechanika 86

RUTMech, t. XXXI, z. 86 (1/14), styczeń-marzec 2014, s. 143-150

Bartosz ZIEGLER1

TRANSONIC AXIAL COMPRESSOR ACTIVE

BOUNDARY LAYER CONTROL CONCEPT AND
ITS EVALUATION
The article presents a concept of boundary layer control in transonic axial com-
pressors. Two methods for obtaining boundary layer control are presented. It is
proposed that by combining those two methods, an efficient way of preventing
turbulent boundary layer separation in transonic compressor passages can be ob-
tained without the loss of the compressor flux. The concept, is based on the idea
of removing boundary layer flows at one stage, and using this removed flux as a
blowing type boundary layer control, at previous stage where lower pressures oc-
cur. This creates a loop of secondary flow between those stages preventing the
boundary layer separation (especially the shock induced separation) on both
stages. Two dimensional numerical analyses are performed to assess the possible
passage performance (especially the diffusion factor). On the basis of two dimen-
sional analyses, the possible performance of such compressor stage is estimated.
The aerodynamic possibility of greatly increasing the axial compressing stage
loadings and pressure ratios without increasing blade velocities and passage flow
Mach numbers is shown. This possibility has a growing importance as the overall
pressure ratio of turbojet engines and tends to rise with every next engine genera-
tion. Main conclusions are drawn, showing the possibilities for further research in
the scope of this concept.

Keywords: flow separation, transonic diffusers, turbojet engines,

1. State of the art and motivation

1.1. State of the art transonic compressors in modern jet engines
Vast majority of modern jet engines, generates their overall pressure ratio,
by axial compressor operation. Because the pressure ratio generated on a single
compressor stage is insufficient for efficient engine operation, compressors in
use are multistage compressors. Since first turbojet engines, the pursuance of
stage pressure ratio enhancement takes place.

1
Autor do korespondencji/corresponding author: Bartosz Ziegler, Poznań University of Technol-
ogy, ul. Piotrowo 3, 60-695 Poznań, e- mail: [email protected]
144 B. Ziegler

Achieving high pressure ratios makes it possible to build an engine with high
overall pressure ratio utilizing compressor of reasonable number of stages. The
Changes of average pressure ratio per compressor stage is shown on figure 1,
separately for high bypass turbofans typical for commercial airliners and low
bypass military turbofans typical for multirole fighter aircrafts. From third order
polynomial trend lines on the plot, one can see, that in both engine types there’s
a continuous trend of increasing the pressure ratio per stage. This pursuance led
to the development of transonic compressors, in which, relative flow velocities
in blade-to-blade passages are supersonic at passage inlet and decrease to sub-
sonic values before passage outlet. Higher the inlet relative Mach number, and
velocity drop in the passage, higher the pressure rise in the channel. In passages
of such compressors the highest pressure rise occurs on a normal passage shock
positioned between the compressor blades (see figure 2). This shock, being an
effective way of flow compression, creates a region of very high adverse pres-
sure gradient. Near passage walls, such pressure gradient causes the boundary
layer flow to turbulently separate (also shown on figure 2).

Fig. 1. Changes of average jet engine stage pressure ratio in past years. Black lines are third order
polynomial trend lines for both engine types
Transonic axial compressor active boundary layer control concept … 145

Fig. 2.. Typical flow topology of a transonic

compressor cascade: 1 – passage shock (a nor-no
mal shockwave slowing the flow down from fr
supersonic to subsonic range), 2 – shock foot on
the blade suction side, the place most prone to
flow separation (so called shock induced sepa-
sep
ration), 3 – shock created by an airfoil leading
edge

1.2. Work motivation

Transonic compressors are in wide use now for more than a decade. Efforts
done in their development allow to steadily increase their performance. The
potential pressure ratio, that can be achieved depends mainly on the kinematics
of the working channel flow,
flow, since the potential (isentropic) total pressure ratio
of a stage can be written as:
κ κ
κ ! κ
(1)

Where: T01, T02 – total temperatures at the working channel inlet and outlet,
κ – heat capacity ratio of the gas, h01, h02 – total
otal enthalpies at the working chan-
cha
nel inlet and outlet respectively, lst – specific work input to the gas.
As can be seen from equation 1, the only way to increase the pressure ratio
attainable from a single compressor stage (at given inlet parameters) is to in-
crease its work.. Since real flows are never isentropic, real pressure ratios are
always somewhat smaller than those resulting from equation 1. The specific
work input, can be determined by the specific power of the rotor, being the
product of working cascade velocity and specific force (being the rate of spe-
sp
cific momentum change). For a simplified 2D flow like presented on figure 2,
this work can be written as:

· &' ()* + & ()* +& ' (2)

Where: u – working cascade velocity, w1, w2 – flow relative velocities at

working channel inlet and outlet respectively, β1, β2 – relative flow angles at
working channel inlet and outlet respectively (see figure 2).
146 B. Ziegler

Usually, in transonic compressors the cascade relative flow inlet parame-

ters are limited by structural strength of the rotor and reasonable flow Mach
numbers (high Mach numbers cause inadmissible total pressure losses on
shocks). Hence the possibility of increasing the stage work is focused mainly on
the reduction of relative outlet velocity and flow angle. Ability of the channel to
slow the flow can be somehow characterized by simplified form of a diffusion
factor – DF:
/012 3/
-. (3)
/

Where: wmax – maximum relative velocity in the passage, w1, w2 – flow

relative velocities at working channel inlet and outlet respectively.
In transonic axial compressor increasing DF is very hard to do, because the
high adverse pressure gradients make flow prone to separate from the blade.
Separated flow cannot be effectively slowed down, nor can it effectively change
direction. A strong separation in a high turning passage flow, leads to compres-
sor stall preventing it from operating at all. To allow for higher flow turning
angles and more diffusion (compressing the gas by slowing it down) in the
channel, the boundary layer separation phenomenon has to be eliminated or at
least controlled. A study taken by MIT Gas Turbine Laboratory [1] showed that,
in fact preventing the flow separation by active boundary layer control allows to
roughly double the compressor stage work. The designed, built and tested in
this study single stage compressor unit, demonstrated total pressure ratio of 3.17
at around 78% efficiency at 100% rotational speed and partial loading. At lower
speeds, higher efficiencies (up to 89%) were recorded. The tested compressor
stage utilized a set of slots in the flow channel that removed the boundary layer
flows (so called boundary layer bleeding). It was also shown, that the compres-
sor designed for such high loadings is unable to operate without the boundary
layer bleeding. After closing the bleeds the compressor stalled immediately, and
was unable to recover from stall without bleeding restored (regardless the value
of applied back pressure) [1]. Some other studies [2-4] showed also the possibil-
ity of boosting axial compressor performance by the removal of boundary layer
flows in regions prone to flow separations. The problem with controlling
boundary layer flows by bleeding is the removal of few percent of flux on each
stage. Doing this on a multistage compressor would significantly reduce the
compressor output flux, hence also the engine power and efficiency. The con-
cept being an object of this article, aims into overcoming that issue.

2. Proposed concept
Boundary layer bleeding seems to be the most effective known method for
preventing boundary layer separation (especially the shock induced separation).
It was utilized since decades in supersonic air inlets (which similarly as com-
Transonic axial compressor active boundary layer control concept … 147

pressor passages are examples

exam of transonic diffusers). The mechanisms
mechanism of bleed
interaction with compressible flows are therefore well known.

Fig. 3. Scheme of proposed boundary layer flow

control arrangement on two adjacent compressor
stages. The working channel of subsequent stage
(R2) is controlled by the boundary layer removal,
the bled air is then used as blowing medium on
preceding stage working channel
hannel R1. Same with
stationary channels (S2 and S1).

There are however similar techniques of active boundary layer control, one
of which is the blowing boundary layer control. The main argument for the in- i
troduction of this method, is that it could theoretically
theoretically utilize as a blowing me-
m
dium the air removed in passages controlled by boundary layer bleeding. Con- Co
necting those two methods for boundary layer control, on two stages of a com- co
pressor would not only eliminate the flux loss, but also eliminate the needne of
providing auxiliary channels and aperture carrying the bled medium out of the
compressor.. Scheme for proposed arrangement of boundary layer control on
two adjacent compressor stages is shown on figure 3. A challenging task in such
arrangement, is to establish an effective way of preventing flow separation by
blowing, with limited pressure
pressure of used medium. The method has to be suffi- suff
ciently effective to manipulate boundary layer with only the overpressure cre- cr
ated between stages and decreased by the lossess in the auxiliary channel con- co
necting the stages.

3. Preliminary 2D working passages

passage design
Two
wo dimensional geometries for both working passages were designed and
steady state analysis with ANSYS Fluent code was performed.. Preliminary
designs, showed that the most effective arrangement of blowing slots is to use
two slots, oriented streamwise, one before the intended position of passage
shock and one after that position. First one, having smaller throat is used to
reaccelerate boundary layer flow before it meets the passage shock, thus pre- pr
venting shock induced separation. The second one, was used to allow for possi-
poss
bly high further deceleration of the flow. The turning angle of the cascade is not
so big (roughly 18°)) this is due to the fact that in two dimensional
ional cascade
flows, the outflow relative Mach number is directly dependent on the outflow
angle. So to slow the flow down to around M=0.4 in two dimensions the turning
angles are quite small (in 3D rotor cascade, the channel span is decreasing so
for the same area ratio a much higher turning angle is obtainable).
obtainable). The later
stage has two bleeding slots.
148 B. Ziegler

Fig. 4. Geometries and anticipated flow topologies for two dimensional cascades. The blowing
controlled working cascade – a) and bleeding controlled working cascade of later stage – b)

Main, with a sharp leading edge, positioned in the place of intended passage
shock foot. This kind of bleed geometry is based on the geometry of so called
shock trap bleed used in supersonic air inlets of turbojet engines.
engines. As the name
suggests, this kind of bleed aside of removing boundary layer has also the abil-
abi
ity to stabilize passage shock position. Both geometries are shown with antici-
antic
pated supersonic flow topologies at figure 4. The plenum for blowing and
bleeding mediumdium inside the blades is not included in the two dimensional
model. Both passages were meshed using ANSYS ICEM software what allowed
for creation of high quality, quadrilateral meshes matched to the anticipated
flow topology (mainly alignment with the flow
flo and higher resolutions in bound-
boun
ary layer regions and near anticipated shock position). Analysis were made for
total pressure – inlet boundary condition and static pressure – outlet boundary
condition. In the case of blowing controlled channel, the k–ω-SST T turbulence
model was used as believed to be the most accurate 2 equation model for flow
separation prediction. In the case of bleeding controlled channel, with k–ω-SST
the solution didn’t tend to converge when thethe shock foot was near the bleed.
bleed
Convergence ce for second channel was finally achieved with the use of Spalart-
Spalart
Allmaras turbulence model. On figure 5, contours of relative Mach number can
be seen for both cascades. On the figure 5 also dimensionless (in relation to the
total inlet pressure of the first
fi cascade) relative total pressures of boundary con-
co
trol flows are given.
Transonic axial compressor active boundary layer control concept … 149

Fig. 5. Mach number

umber contours in moving frame for blowing controlled stage working
channel – a) and bleeding (suction) controlled stage working channel – b).

Although the stationary

stationary channel connecting those two was not modelled,
the inlet parameters of the second cascade are taken from the outlet of the first
one (with 3% loss of total pressure to approximate for the pressure loss in stator
passage). This was done to show, that the relation in pressures between cooper-
coope
ating stages can be sufficient for proposed boundary layer control arrangement.
It can be noted, that the medium injected into the blowing controlled stage can
have significantly lower total pressure than the medium removed from later
stage. This gives some margin for the losses in auxiliary channel connecting the
stages. Performance
erformance of both channels can be seen in table 1.

Table 1. Parameters of both cascades obtained

obtai in the numerical analysis

Parameter Cascade 1 Cascade 2

Chord based Reynolds number – ReC 8,4 × 106 9,0 × 106
total pressure ratio – TPR 3,02 2,53
Channel pressure
ressure recovery ratio – PRR 0,951 0,980
Bleeding/blowing flux fraction 3,24% 3,72%
Diffusion factor – DF 1,15 (0,76)* 0,78
∆89: 0,830 0,862
Loading ψ
* The jets injected into the channel achieve locally very high velocities, hence high wmax, the
value in brackets gives the DF calculated with exclusion of that injection region, which gives
more feasible information about main channel work.

4. Conclusions
Conducted study which elements are adduced in this article, showed the
possibility of creating compact and efficient flow arrangements achieving
boundary layer control on two adjacent highly loaded transonic compressor
stages. The proposed method could allow for significant improvement in tran-
tra
sonic passages diffusion factors, without the need of wasting the compressor
150 B. Ziegler

flux and providing external apertures for bled air removal. It should be also
noted, that channels shaped in three dimensions could achieve higher flow turn-
ings with smaller diffusion factors. This create chance to achieve even higher
cascade loadings than presented, or allow for simpler single point boundary
layer blowing/bleeding arrangements in the blade to blade passage.

Reference
[1] Kerrebrock J. : Final Technical Report on AFOSR Grant F49620-98-1-0493 entitled
RESEARCH AND DEVELOPMENT ON ASPIRATED COMPRESSORS, Gas
Turbine Laboratory Department of Aeronautics and Astronautics MIT, Cambridge
2003.
[2] Epstein A. et al. : Final Technical Report on AFOSR Grant F49620-03-1-0309 enti-
tled ASPIRATED HIGH PRESSURE COMPRESSOR, Gas Turbine Laboratory
Department of Aeronautics and Astronautics MIT, Cambridge 2006.
[3] Knapke R.D., Turner M.G.: Detailed Unsteady Simulation of a Counterrotating
Aspirated Compressor with a Focus on the Aspiration Slot and Plenum, Int. J. Rotat-
ing Machinery, 2013, Article ID 857617.
[4] Colombo E. et al.: Investigations on axial compressor cascades with aspiration on
blades and hub, European Workshop on New Aero Engine Concepts, Munich 2010.

KONCEPCJA KONTROLI WARSTWY PRZYŚCIENNEJ

W SPRĘŻARKACH TRANSONICZNYCH I JEJ EWALUACJA

Streszczenie
Artykuł prezentuje koncepcję dwóch metod kontroli warstwy przyściennej w kanałach tran-
sonicznych sprężarek. Zaproponowano aby przez połączenie tych dwóch metod, skuteczny spo-
sób zapobiegania turbulentnego oddzielenia warstwy granicznej w transonicznych kanałach sprę-
żarek można uzyskać bez straty strumienia sprężarki. Koncepcja opiera się na idei usuwania
przepływów w warstwie brzegowej w jednym etapie oraz wykorzystaniu strumienia jako kontroli
warstwy brzegowej we wcześniejszym etapie, gdy występuje niższe ciśnienie. To tworzy wtórną
pętlę przepływu pomiędzy tymi etapami uniemożliwiających rozdzieleniu warstwy granicznej
(zwłaszcza rozdzieleniu wywołanemu wstrząsem) na obu etapach. Dwuwymiarowa analiza nume-
ryczna została przeprowadzona w celu oceny możliwej wydajności przejściowej (zwłaszcza
współczynnika dyfuzji). Na podstawie dwuwymiarowej analizy określono możliwość wykonania
takiego stopnia sprężarki. Przedstawiono aerodynamiczną możliwość znacznego zwiększenia
obciążeń stopnia sprężania osiowego i wskaźników ciśnienia bez zwiększania prędkości łopat i
liczb Macha podczas przepływu. Możliwość ta ma coraz większe znaczenie jako wskaźnika
całkowitego ciśnienia silników turboodrzutowych i wykazuje tendencję do wzrostu w każdej
następnej generacji silnika. Nakreślono główne wnioski, pokazując możliwości dalszych badań w
tym zakresie.
Słowa kluczowe: separacja przepływu, dyfuzory transoniczne, silniki turboodrzutowe
DOI: 10.7862/rm.2014.16
Otrzymano/received: 15.05.2014
Zaakceptowano/accepted: 27.05.2014