Jtam.dvi


JOURNAL OF THEORETICAL

AND APPLIED MECHANICS

50, 2, pp. 455-472, Warsaw 2012

50th Anniversary of JTAM

NON-SYMMETRICAL BOUNDARY LAYER SUCTION IN

A COMPRESSOR CASCADE

Karsten Liesner, Robert Meyer, Sebastian Schulz

German Aerospace Center DLR, Institute of Propulsion Technology – Department of Engine

Acoustics, Berlin, Germany; e-mail: karsten.liesner@dlr.de; robert.meyer@dlr.de

In order to simulate non-symmetrical boundary layer suction in anannu-
lar compressor, a cascade investigation has been performed with single-
sided suction slots only. A preceding investigation had revealed a high
potential for loss reduction by two different types of boundary layer
suction. The experimental investigationwas performed with five NACA
65-k48 stator blades at the design Mach number of 0.67 and Reynolds
number of 560.000.The two investigated suction geometries are a narrow
slot following the design of Peacock and awider slot of own origin, both
slots are positioned on one side of the passage only. The Peacock slot is
placed in the corner between suction side of the vane and the side wall,
the wider slot is positioned from suction side to pressure side following
the side wall’s flow detachment line. With half the suction rate of the
preceding investigation the efficiency of the cascade could still be enhan-
ced. In the casewith 2.5% suction rate the total pressure loss coefficient
of the full passage was decreased by 13%, in the case with 1% suction
rate the loss coefficient was decreased by even 10%. The outflow of the
cascade is as expected no more symmetrical and the one sided suction
has no large impact on the flow of the opposite side of the passage.

Key words: compressor cascade, secondary flow, active flow control, flow
suction

Nomenclature

Geometric and flow quantities

M,Re – Mach and Reynolds number, [–]
c,h – vane chord length and vane height, [m]
m – mass flow, [kg/s]
p – pressure, [Pa]
t – cascade pitch, [m]



456 K. Liesner et al.

u,x – cascade pitch and axial coordinate, [m]
z – vane height coordinate, [m]
β – cascade flow angle, [◦]
ζ – total pressure loss coefficient, [–]

Subscripts

1 – cascade inlet
2 – cascade outlet

1. Introduction

Efficient compressors provide a high pressure ratio in combination with a low
number of stages. Not only aerodynamic aspects have to be taken into consi-
deration, also theweight plays a significant role in engine design. In the recent
years,much effort has beendone to improve compressor aerodynamics in order
to enhance the turbo-machine efficiency.

Aerodynamic losses in axial compressors are dominated by the secondary
flow losses (Scholz, 1956). The higher the pressure ratio in the compressor
stage, the more likely the flow will detach in the corner of the suction side of
the blade and endwall (Cumpsty, 2004). This secondary loss is triggered by
movement of thewall boundary layer fluid to the suction sidedue to the strong
pressure gradient. The flow on the suction side, facing an adverse pressure
gradient, can not overcome the pressure and detaches from the surface. In
the detached area in the corner, a vortex system develops and leads to strong
pressure losses (Kang and Hirsch, 1991).

Fig. 1. Vortex system in a compressor cascade (Kang and Hirsch, 1991)



Non-symmetrical boundary layer suction... 457

Three-dimensional bladedesignandcasing treatment are themethodsused
most often for passive secondary flow control in the machine. For compressor
cascades, there are different passive techniques available (Hergt et al., 2006;
Lin and Howard, 1989). The problem of secondary flow between the pressure
and suction side of the passage remains, but decreases.

There is, in fact, a way to remove the secondary flow on the whole out
of the compressor. Suction of the secondary flow has been proven by Peacock
(1965) already in the 1960’s to be able tomake secondary flow losses disappear
from a compressor with very lowMach numbers. An optimized suction slot is
used to draw 0.1% of flow out of the passage and the secondary flow effect in
thewake is nomore found.That effect in lowMach number cascades has been
used ina recent investigation at theGermanAerospaceCenter (DLR) inBerlin
to remove the secondary flow from a high speed cascade effectively. Different
geometries have been used in order to decrease the pressure loss coefficients
of such a compressor cascade. The full results of suctionmeasurements can be
found in a recent publication (Liesner andMeyer, 2010).

The solution found in this investigation is only theoretically a good so-
lution. An annular compressor is a rotating machine. The stator vanes are
connected as cantilevers to the casing, the rotor blades rotate with the hub at
several thousand revolutions perminute. Therefore, it is very unlikely to suck
both the hub and casing boundary layer symmetrically on fixed positions. In
order to get insight into the behaviour of the flow in such non-symmetrical
conditions, this investigation has been performed.

2. Experimental setup

All experimentswere carried out at the high-speed stator cascadewind tunnel
of the German Aerospace Center (DLR) in Berlin. The test rig, shown in
Fig.2, has a rectangular cross section of 40mm width and 90mm height at
the cascade inlet. The nozzle with a contraction ratio of 1:218 accelerates the
flow up toMach number 0.7. AReynolds number of 0.6 ·106 can be obtained.

The boundary layer height of each inlet wall can be adjusted independen-
tly (see Fig.3). The upper and lower boundary layers are sucked to ensure a
periodic flow to the compressor cascade. The periodicity of the flow is moni-
tored by a row of static pressure probes in the cascade inlet plane. The side
wall boundary layers can be sucked in order to vary the inlet boundary layer
height, but this suctionwas not used in this investigation. The inlet boundary
layer height has been set to 3mm or 7.5% of the chord length.



458 K. Liesner et al.

Fig. 2. Experimental high speed cascade wind tunnel at DLRBERLIN

Fig. 3. Test section and cascade coordinates

The inlet angle of the cascade, β1, can be adjusted geometrically from
122◦ to 146◦. In the aerodynamic design condition measurement, β1 is set to
132◦. For flow data acquisition, a rake with 26 pitot tubes for total pressure
measurement and four Conrad angle probes for outflow angle determination
are used. The horizontal rake is traversed in the circumferential direction in
themeasurement plane 16mm(40%chord)downstreamof the stator exit.The
rake Pitot tubes are made of 0.7mm diameter tubes, their total area covers
only 0.26% of the full outflow area. Aprevious investigation showed that there
is no influence on the cascade flowmeasurable.

The total pressure loss coefficient is calculated for every measurement as
a measure for the outflow quality. The pressure data is recorded in a multi-
channel pressure transducer of high accuracy; temperatures are measured via
Pt100 resistance thermometers. The systematic measurement error has been



Non-symmetrical boundary layer suction... 459

estimated according to DIN 1319 [3]. The measurement uncertainty of the
total pressure loss coefficient is less than 1.2%. The repeatibility accuracy of
the total pressuremeasurements is better than 0.2%.

To be able to compare the results of the measurements to other cascade
investigations, a compressor stage efficiency has been introduced and adapted
to the wind tunnel measurements. It uses a hypothetical rotor of known total
pressure ratio and efficiency. With these constant data, the stator efficiency
can be calculated using the total pressure in the measurement plane. Added
by a term for the vacuum pump energy consumption, the calculated stator
efficiency is a measure for the efficiency of the stator and the suction system.
An increased efficiency corresponds to a net fuel saving.

Amore detailed description of the wind tunnel and its measurement pro-
cedures can be found in the publication of Liesner andMeyer (2008).

3. Results of preliminary symmetrical investigations and

reference cascade

In order to understand the working principles of the flow control, a good
knowledge based on the uncontrolled cascade is essential. There have been
many investigations on the NACA65-k48 cascade at this test facility (Hage
et al., 2007; Hergt et al., 2006, 2007, 2008; Meyer et al., 2007) and in other
scientific institutes (e.g.Hübner, 1996; Scheugenpflug, 1989,Watzlawik, 1991).
This is the primary reasonwhy these investigations have been carried outwith
this kindof profile.All themeasurements havebeenperformedwith equal inlet
conditions of β1 =132

◦, M= 0.67 and Re= 5.6 ·105 based on 40mm chord
length.

The 2-dimensional plot of the total pressure loss coefficient versus the vane
height and the cascade pitch is used to describe the local distribution of losses
in the wake. The base flow is characterised by an area of high pressure loss
above the suction surface of the vane. This is represented by the red area in
the illustration. In this area, a recirculation and vortex system described in
the introduction, is responsible for the defect of total pressure. It is caused by
the rolling up of secondary flow in the rear part of the cascade.

This secondary flow loss is very large in the cascade examined. A state of
the art cascade or axial compressorwithmodern3-dimensionally shapedvanes
shows a less distinct secondary flow loss.Nevertheless the problem remains the
same. In the corner between the vane and sidewall the flow detaches from the
wall and rolls up into secondary flow vortices.



460 K. Liesner et al.

Fig. 4. Total pressure loss coefficient, reference cascade,M=0.67, Re=5.6 ·105

One possible way to eliminate this rolling up of secondary flow in a vortex
and the separation is to suck the boundary layer out of the cascade. There
have been studies that plane that the complete disappearance of the secondary
flow is possible.

In the cascade, the flow needs to be sucked out by vacuum pump energy.
This does, of course, energy consumption. In the turbo-machine, the ambient
pressure ismuch lower than the compressor pressure.The boundary layer flow
can be driven out of the machine by just opening a valve through the casing
wall.

Fig. 5. Geometry of suction slots (red) in the compressor cascade walls, Type A
(left) and type B (right)

In the preliminary investigation, the cascade has been influenced symme-
trically. There have been orifices for boundary layer suction on both cascade
side walls. The results were impressive, such as 38% of total pressure loss re-
duction with 5% of mass-flow was drawn off. In this case, the stage efficiency



Non-symmetrical boundary layer suction... 461

was increased by 2%, including the suction energy consumption. In another
type of orifice, a total pressure loss reduction of 22% could be realised by
drawing off 2% of the main mass flow only. This corresponds to a net stator
efficiency increase of 1%. Keeping in mind that the secondary air system in
an airplane uses compressed air for pneumatic systems and air condition, a
useful implementation of the flow suction is possible.
The geometries have been derived from a literature study of the active

flow control for type B, while type A geometry has been developed after a
numerical study of the base flow cascade performed by the TU Berlins ISTA
(Meyer and Thiele, 2010).

Fig. 6. Total pressure loss coefficient, symmetrical boundary layer suction, type A,
M=0.67, Re=5.6 ·105, suction rate=5%

Figure 6 depicts the pressure loss distribution of the cascade with 5% suc-
tion, cascade type A. The vortex system has almost completely disappeared;
the loss distribution of the high values seems almost like the wake of a flat
plate. The secondary flowhas been removed from the sidewalls in great parts.
Figure 7 shows a comparison between the base flow and the cascade with

the symmetrical boundary layer suction of 5% of the main mass flow. It is
obvious that in the region with high pressure loss in the base flow, there is a
huge gain in pressure recovery. The separated region is very small compared
to the base flow. The lower part of Fig.7 shows the distribution of the flow
deflection versus the vane height. Since there is a symmetrical distribution of
the flow parameters, there is only a limited number of angle probes present.
Figure 8 shows the known downflow overturning distribution of the

NACA 65 cascade according to Watzlawick. It illustrates the presence of the
corner vortex and the passage vortex and their influence on the outflow angle.
The corner vortex on the wall with its sense of rotation counterwise to the



462 K. Liesner et al.

Fig. 7. Comparison of reference cascade and symmetrical suction type A

passage vortex turns the flowupwards in the area of the first angle probe.The
other side of the strong passage vortex with rotation downwards makes the
flow turn less in the area of the second pressure probe. As the influence of the
vortex system diminishes towards mid-span of the cascade, the overturning
decreases until it disappears near the centerline.

The deflection of the cascade is measured by subtraction of the exit an-
gle β2 from the inlet angle β1. The lower the outflow angle, the higher the
deflection.

The typical low deflection in the vortex area of the vane wake is no more
to be seen. With the flow longer attached to the suction surface, it can be
turned to a higher degree. This effect is mostly visible at the region between



Non-symmetrical boundary layer suction... 463

Fig. 8. Flow overturning near the wall according to Watzlawick (1991) and position
of the angle probes in the channel

z/h= 0.1 and 0.4 as well as on the other side. It reaches to the mid-span of
the passage, where it becomes small of course. The deflection is at the same
timemore balanced through the span and higher on the whole.
The lower part of the comparison (Fig.7) shows the total loss coefficient

pitchwise integrated. The cascade with flow control shows a wide plateau of
moderate loss, from 0.1 to 0.9 of the vane height. The sidewall loss is also
reduced significantly. In explicit numbers, the total pressure loss coefficient is
reduced from 0.9 to 0.55, which corresponds to a reduction of 38%.

Fig. 9. Total pressure loss coefficient, symmetrical boundary layer suction, type B,
M=0.67, Re=5.6 ·105, suction rate=2%

The next picture (Fig.9) shows the local total pressure loss distribution of
the cascade with the boundary layer suction type B employed. A mass flow
rate of 2% has been drawn off the cascade. Although the local distribution



464 K. Liesner et al.

does not look too impressive, the comparison with the base flow brings more
light into that investigation.

Fig. 10. Comparison of the reference cascade and symmetrical suction type B

While big parts of thewake remain unchanged, there are two large regions
with great pressure loss reduction. This is the region where the vortex system
joins themain flow in the central region of the passage. The vortices have been
reduced in area and value.
The deflection of the passage is enhanced as well; the outflow angle distri-

bution is smoother than in the cascade without flow control.
The integrated pressure loss versus vane height shows a drastic reduction

of pressure loss in the region of the vortexmainflowborder. In absolute values,
the total pressure loss coefficient has been decreased to 0.7, which corresponds
to a reduction by 22%.



Non-symmetrical boundary layer suction... 465

4. Results of non-symmetrical boundary layer suction

Knowing the results of the preliminary investigations, the question for possible
implementation arises. The turbo-machine itself is a rotating hub inside a
casing.This automatically leads to parts that are translated versus each other.
Amethodofplacing suction slots in the corners of bothhuband tip is therefore
virtually impossible. Possible insteadwould be the implementation of only one
sided flow suction out of a machine. In order to figure out the effects of such
a non-symmetric behaviour, the suction slots have only been realised on one
side of the cascade. The suction flow rate has been bisected compared to the
previous investigation, so the amount of flow drawn through the suction slots
remains unchanged.
Unfortunately, thedownflowangle determination is onlypresent on the left

side of the cascade. In the usual investigation, there is a proven symmetry in
the passage, and it can be assumed that the channel deflection be symmetrical
as well. In this setup it is not the case. Since the side with suction has been
examined in the previous Section, the angle determination has been made at
the side of the cascade without control. The effect is described later.

Fig. 11. Total pressure loss coefficient, non-symmetrical boundary layer suction
type A, M=0.67, Re=5.6 ·105, suction rate=2.5%

On the local total pressure loss distribution in Fig.11, it is obvious that
the suction slots have been placed on the right hand side of the cascade. The
wake of the cascade is as non-symmetric as expected. On the right hand side,
the secondary flow loss is little. There is only a small vortex formation in the
corner between the vane and side wall. The effect is not as strong as in the
symmetrical case. On the opposite side, the vortex is huge. It has been even
extended a little in its dimension compared to the base flow.



466 K. Liesner et al.

Fig. 12. Comparison of the reference cascade and non-symmetrical suction type A

In comparison with the base flow cascade (Fig.12), the total pressure loss
has still been reduced by 13%. The loss reduction is obviously taking place on
the right hand side of the cascade where the suction is placed. On the left side
only, a small loss increase is obvious. The big changes in the colour map are
due to an offset of the high loss area from down upwards. The outflow angle
of the cascade is higher than the one in the base flow cascade, which means
a lower flow turning in the left part without suction. This can be explained
by higher vorticity of the passage vortex, which induces higher speeds in the
cascade pitch direction (see Fig.8). In the comparison picture, this is also
visible in the offset between the reference and controlled cascade.

The next investigation has been made with the cascade of type B and a
suction rate of 1% of the mass flow (Fig.13). Again, the right side has been



Non-symmetrical boundary layer suction... 467

Fig. 13. Total pressure loss coefficient, non-symmetrical boundary layer suction
type B, M=0.67, Re=5.6 ·105, suction rate=1%

equipped with suction slots and the left one remains unchanged. This setup
shows about the same behaviour as the previous one. The side with suction is
showing less vortex structure, the region of high total pressure loss has been
decreased in its dimension. On the left side, there is little change visible. The
vortex on that side seems to be slightly bigger.

In the comparison layout (Fig.14), the region where the boundary layer
suction takes effect is obviously the border between the secondary vortices and
themain flow. In the integral loss coefficient distribution, the step on the right
side at z/h= 0.7 to 0.8 disappears. This is due to the fact that the passage
vortex is weakened and therefore causes less total pressure loss. The pressure
loss in this case is decreased byapprox. 10%.Thedeflection of the passage side
without suction is plotted, as in the previous example there is a small negative
effect in the deflection obvious, this is due to the slightly bigger dimensions of
the passage vortex in this case. The up-facing velocity of the vortex forces the
flow to follow.

The offset of the reference cascade and the controlled one is obvious again;
the lower deflection is responsible for this effect.

5. Discussion

The effect of the secondary flow in compressor cascades is known very well.
The flow along the sidewall, following the pressure gradient from the pressure
to suction sideof thepassage, causes theflowon the suction surface to separate



468 K. Liesner et al.

Fig. 14. Comparison of the reference cascade and non-symmetrical suction type B

fromthevanes. Suppressing the secondaryflowtherefore, leads to adecrease of
the flow detachment on the stator surfaces. Different techniques for secondary
control have been employed.

The concept of boundary layer suction has been proven to be a powerful
tool to get rid of the secondary flow separation in the corners between the
stator vane and wall. With this investigation, the effect of non-symmetric
boundary layer suction has been examined.

In a turbo-machine, however, it is impossible to have both hub and casing
of a compressor equipped with fix position suction slots. It is necessary to
decide whether one or the other will be implemented.



Non-symmetrical boundary layer suction... 469

Todecide how theflowbehind the cascade does look like, not only pressure
measurements are useful. It is interesting to see if the flow does a movement
from the sidewithout to the onewith suction, following the pressure gradient.
Itwould be preferable to obtain a two-dimensional flowbehind the stator. The
pressure distribution behind the cascade is not a real measure for the two-
dimensionality. Laser measurements must show, whether the flow structure is
sufficiently symmetric or not.

Themeasurements showtheability to employone-sided suction intoa com-
pressor cascade.Their effect lies in the removal of the low speed secondary flow
along the side wall. This secondary flow is likely to produce large detachment
areas on the surface of stator vanes.With the secondary flow not reaching the
suction surface, this detachment occurs differently. The 3-dimensional detach-
ment on the surface is no more to be found.

In addition to the total pressure loss coefficient, there is determination of
the cascade efficiency for every casemeasured (Bräunling, 2004). The efficien-
cy is calculatedwith a reference rotor with constant pressure and temperature
ratio. In the reference case, the cascade has an efficiency of 89.21%.The casca-
dewith suction slots of typeA has an increased efficiency of 89.92%, in which
the vacuum pump energy consumption has already been taken into account.
The efficiency of the cascade type B is 89.68%. In the annular compressor
case these values can easily be improved, since the air can be drawn off the
compressor without external pump energy.

Another interesting fact about this investigation is that the amount of the
preferred flow suction determines the best shape of the suction slots. With a
desired suction rateof only1%, theconfigurationof typeB is the recommended
form. It has already been developed early and still did not lose its efficiency.
For higher suction rates, different geometries are preferable, e.g. the one of
cascade type B with a bigger cross section and different positioning on the
endwalls.

6. Conclusion

The paper presents experimental measurement data on the non-symmetrical
boundary layer suction in a compressor cascade with a Mach number of 0.67
andReynolds number of 5.6 ·105. The cascade consisted of fiveNACA65-K48
vanes with a height of 40mm and an aspect ratio of z/c=1. The previously
conducted investigations with boundary layer suction had shown a possible
total pressure loss decrease of 38% (and 22%) using mass flow rates of 5%
(and 2% respectively) in the symmetrical suction case.



470 K. Liesner et al.

In the non-symmetrical measurements, the flow through the suction orifi-
ces was kept constant, the resulting mass flow rate was 2.5% and 1%. With
those settings, a total pressure loss enhancement of 13% and 10%, respecti-
vely, was reached. There is a positive effect on the efficiency of the cascade
in pressure loss as well as in deflection of the flow on the side with suction
orifices. The passage side without suction is slightly degraded, but there is no
bigger influence than an obvious small increase of the vortex formation behind
the cascade. The total pressure loss remains almost unchanged.
The efficiency of the cascade on the whole is improved; a further enhance-

ment in an annular compressor is also expected.

Acknowledgement

The results presented in this paper have been obtained within the project “Le-

istungssteigerung von Strömungsmaschinen durch aktive Sekundärströmungsbeein-

flussung in einer Verdichterstufe”. Funding by the German Research Foundation

(DFG) is gratefully acknowledged.

References

1. Bräunling W., 2004,Flugzeugtriebwerke, Springer Verlag, 2. Auflage

2. Cumpsty N.A., 2004,Compressor Aerodynamics, Krieger Publishing Compa-
ny, 2nd edition

3. DIN 1319, Auswertung von Messungen einer einzelnen Messgrösse, Messunsi-
cherheit, Latest version: 05/1996, Deutsches Institut für Normung

4. Hage W., Meyer R., Paschereit O., 2007, Control of secondary flow in a
high loaded compressor stage by means of a groove structure on the sidewalls,
AIAA Applied Aerodynamics Conference, AIAA 2007, 4278-489

5. Hergt A., Meyer R., Engel K., 2006, Experimental investigation of flow
control in compressor cascades,ASME TurboExpo 2006, GT2006-90415

6. Hergt A., Meyer R., Engel K., 2007, The capability of influencing secon-
dary flow in compressor cascades by means of passive and active flow control,
Council of European Aerospace Societies CEAS, 2007-216

7. Hergt A., Meyer R., Müller M.W., Engel K., 2008, Loss reduction in
compressor cascades bymeans of passive flow control,ASMETurboExpo 2008,
GT2008-50357

8. Hübner J., 1996, Experimentelle Untersuchung der wesentlichen Einflussfak-
toren auf die Spalt- und Sekundärströmung in Verdichtergittern, Ph.D. thesis,
Uni BW,München



Non-symmetrical boundary layer suction... 471

9. Kang S., Hirsch C., 1991, Three dimensional flow in compressor cascades,
ASME TurboExpo 1991, GT-1991-114

10. Liesner K., Meyer R., 2010, On the efficiency of secondary flow suction in
a compressor cascade,ASME TurboExpo 2010, GT-2010-22336

11. Liesner K.,MeyerR., 2008,Experimental setup for detailed secondary flow
investigation by two-dimensionalmeasurement o total pressure loss coefficients
in compressor cascades, VKI XIX Symposium on Measurement Techniques in
Turbomachinery

12. Lin J.C., Howard F.G., 1989, Turbulent flow separation control through
passive techniques,AIAA-1989-0976

13. MeyerR., Bechert D.W.,HageW., 2007, Secondary flow control on com-
pressor blades to improve the performance of axial turbomachines, 5th Euro-
pean Conference on Turbomachinery – Fluid Dynamics and Thermodynamics,
Prague

14. Meyer R., Thiele F., 2010, Leistungssteigerung von Strömungsmaschinen
durch aktive Sekundärströmungsbeeinflussung in einer Verdichterstufe, DFG
Projekt,TUBerlin, ISTAandDLR, Institute ofPropulsionTechnology,Engine
Acoustics Department; Ergebnisberichte

15. Peacock R.E., 1965, Boundary layer suction to eliminate corner separation
in cascades of airfoils,NACA-RM-3663

16. Scheugenpflug H., 1989, Theoretische und experimentelle Untersuchungen
zur Reduzierung der Randzonenverluste hochbelasteter Axialverdichter durch
Grenzschichtbeeinflussung, Ph.D. thesis, Uni BW,München

17. ScholzN., 1956, Über dieDurchführung systematischerMessungen anSchau-
felgittern, ZFW-1956

18. WatzlawickR., 1991,UntersuchungderwesentlichenEinflussfaktorenauf die
Sekundärverluste in Verdichter- und Turbinengittern bei Variation des Schau-
felseitenverhältnisses, Ph.D. thesis, Uni BW,München

Problem zasysania w niesymetrycznej warstwie przyściennej kaskady

sprężarki

Streszczenie

Wpracyprzeprowadzonosymulacjęproblemuzasysaniaczynnikawniesymetrycz-
nej warstwie przyściennej sprężarki, analizując kaskadę z jednostronnymi wlotami
ssącymi. Badania poprzedzające ujawniływysoki potencjał w ograniczaniu strat przy
zastosowaniu dwóch różnych rozwiązań sposobu zasysania w warstwie przyściennej.



472 K. Liesner et al.

Doświadczenia przeprowadzono, używając pięciu łopatek kierownicy sprężarki NA-
CA 65-K48 projektowanej do przepływu przy prędkości 0.67Ma i liczby Reynoldsa
560000.Geometria obydwu rozwiązań zasysaniawykorzystuje projektwąskiej szczeli-
ny wlotowej R.E. Peacocka oraz projekt własny szczeliny szerszej, przy czym obydwa
typy umiejscowiono wyłącznie po jednej stronie kanału przelotowego. Szczelinę Pe-
acocka umieszczono w rogu pomiędzy stroną ssącą łopatki oraz ściany, natomiast
szczelina szeroka pokrywa obszar od strony ssącej do sprężającej wzdłuż linii odry-
wania przepływu. Przy tempie zasysania stanowiącym połowę wydatku stosowanego
w poprzedzających badaniach nadal odnotowano możliwość zwiększenia sprawności
kaskady.Przy spadku tempa ssania do 2.5%,współczynnik całkowitych strat ciśnienia
zmalał o 13%, natomiast dla dalszego spadku do 1%współczynnik ten zmniejszył się
o 10%. Zgodnie z oczekiwaniami wypływ z kaskady utracił symetrię, a zasysanie jed-
nostronne nie zmieniło znacząco charakteru przepływu na przeciwnej stronie kanału
przelotowego.

Manuscript received September 13, 2010; accepted for print July 25, 2011