Proceedings of ASME TURBO EXPO 2003

Power for Land, Sea & Air

June 16-19, 2003, Atlanta, Georgia USA

GT2003-38392

PERFORMANCE OF STRONGLY BOWED STATORS IN A 4-STAGE HIGH SPEED

COMPRESSOR

Axel Fischer
Walter Riess
Joerg R. Seume
Turbomachinery Laboratory
University of Hannover
Germany

ABSTRACT and swept vanes are the subject in many numerical studies and wind
The FVV-sponsored-Project „Bow Blading“ (c.f. tunnel experiments (e.g. [1-3]).
acknowledgments) at the Turbomachinery Laboratory of the
University of Hannover addresses the effect of strongly bowed stator The characteristic features of a bowed vane are a positive lean
vanes on the flow field in an 4-stage high speed axial compressor angle at the hub and a negative lean angle at the shroud in
with controlled diffusion airfoil (CDA) blading. The compressor is comparison with a normal, radially stacked vane. The leaned vane
equipped with more strongly bowed vanes than have previously been induces radial forces on the fluid, so that at the acceleration part of
reported in the literature. The performance map of the present the vane the streamlines are moved to the mid-span section. At the
compressor is being investigated experimentally and numerically. deceleration part of the vane the streamlines are moved from the mid-
span section towards the end walls. The loading of the flow in the end
The results show that the pressure ratio and the efficiency at the wall regions is reduced which can lead to lower end wall losses and
design point and at the choke limit are reduced by the increase in reduce the tendency toward corner stall.
friction losses on the surface of the bowed vanes, whose surface area
is greater than that of the reference (CDA) vanes. The mass flow at Joslyn [1] showed, that corner stall is one of the most important
the choke limit decreases for the same reason. Because of the change secondary flow-field effects in compressor stator vanes and that
in the radial distribution of axial velocity, pressure rise shifts from corner stall losses have great influence on overall losses.
stage 3 to stage 4 between the choke limit and maximum pressure Breugelmans [2] and Shang [3] showed with linear-cascade wind
ratio. Beyond the point of maximum pressure ratio, this effect is not tunnel experiments, that corner stall can be reduced by leaned and
distinguishable from the reduction of separation by the bow of the bowed stator vanes. Because of the shift of mass flow towards the
vanes. mid-span section of the vane, the losses at mid span increased more
than they decreased at the hub. Weingold et al. [4] demonstrated that
Experimental results show that in cases of high aerodynamic a 3-stage high speed compressor, equipped with bowed stator vanes
loading, i.e. between maximum pressure ratio and the stall limit, can increase pressure ratio and efficiency from the choke limit
separation is reduced in the bowed stator vanes so that the stagnation through surge conditions.
pressure ratio and efficiency are increased by the change to bowed
stators. It is shown that the reduction of separation with bowed vanes To study the effects of strongly bowed stator vanes on
leads to a increase of static pressure rise towards lower mass flow so compressor performance and on the radial distribution of
that the present bow bladed compressor achieves higher static aerodynamic loading, the last two stages of the present test rig were
pressure ratios at the stall limit. equipped with strongly bowed stators, without change of the rotor
blading and the vane count. The flow fields were measured by
INTRODUCTION pneumatic four-hole probes. On the suction and pressure side of the
To optimize state of the art turbomachinery, it is necessary to vane of the third stage, static surface pressures were measured to
control the flow in the end-wall regions. Therefore, the current determine the local pressure distribution and to identify zones of
development of new turbine and axial compressor blading is separation. The performance map of the compressor with bowed
characterized by three-dimensional blading concepts. Leaned, bowed, vanes was determined and conditions of stall, choke, and best

1 Copyright © 2003 by ASME

Downloaded From: [Link] on 02/03/2016 Terms of Use: [Link]

 efficiency were identified at design speed. Detection of separation the compressor is located close to the last stage, so that the volume
and of rotating stall were supported by unsteady pressure between outlet and throttle is small. This allows throttling the
measurements. The effect of bow was predicted using commercial compressor until full-span rotating stall occurs without facing the risk
CFD codes. Pressure distributions, pressure rise and efficiency trends of a low frequency surge.
were predicted at design point and choke.
The mass flow is metered by an orifice. The axial inlet flow
shroud conditions are measured by radially traversing with a Prandtl tube
and a temperature probe. The outlet flow conditions and the flow
30°
bow field between the blade rows are measured by radially and
circumferentially traversing with a pneumatic four-hole probe
suction including a temperature sensor. Efficiency and overall total pressure
reference ratio are then computed from the flow field data. Static surface
side
pressure distributions on the surface of the third stator are determined
0° with instrumented vanes Figure 3.
pressure
side

35°

hub

Figure 1: Stacking line of bow vane and drawing of bowed stator 3.

MODIFICATIONS OF BOWED VANES

Figure 1 shows the stacking line of the bowed vane. The shift of
stacking line is purely tangential. The lean angle at the hub is 35°, at
the shroud it is 30°. At mid span, the vane is stacked radially. The
vanes of the reference design are radially stacked over the entire
span. Stagger angle and chord length of the bow vanes do not vary
relative to the reference blades except of a small change at hub and
shroud section which were twisted by 1° towards flow direction. The
vane count of the reference design and the design with bowed vanes
is equal.

throttle

Figure 3: Third stator instrumented reference vane

Static pressures at the casing are measured to determine the pressure

ratios of the single blade rows and the stages. For investigations of
stall inception and the initialization of full-span rotating stall,
unsteady pressure transducers are installed in the casing and in an
additional unsteady four-hole probe.

Figure 2: 4-stage axial compressor 3D-NAVIER-STOKES COMPUTATIONS

To investigate the effects of bowed blading numerically, the
finite-volume based Navier-Stokes solver FINETM Turbo by
TEST RIG AND MEASUREMENT TECHNIQUES NUMECA International was applied. The complete compressor,
The test rig is driven by a 1300 kW DC motor. The compressor including the inlet duct and the exhaust diffuser was meshed, Figure
itself is a 4-stage high speed machine fully equipped with CDA 4, and the constant speed lines for the reference design and the
blading. The maximum speed is 18000 rpm, the design speed of the configuration with bowed vanes were computed. The rotor/stator
present tests is 17100 rpm and the compressor generates an overall interaction was simulated by using quasi-steady mixing planes which
total pressure ratio of 2.65:1 [5]. Figure 2 shows that the throttle of average the flow properties at the rotor/stator interface [6].

2 Copyright © 2003 by ASME

Downloaded From: [Link] on 02/03/2016 Terms of Use: [Link]

 For symmetry reasons, only one pitch of every blade row needs up to a maximum and then decrease as the outlet throttle is closed.
to be computed. The inlet boundary is set to standard ambient Excessive aerodynamic loading of an airfoil generates such strong
conditions. At the compressor discharge, the static wall pressure is flow separation that turning is reduced, leading to a reduction of
set. If this wall pressure is raised it is like closing the outlet throttle pressure ratio. Therefore, flow separation exists at the points of
of the machine. For flow field analysis, the same planes are used as maximum static pressure ratio of single blade rows and beyond
the probe traversing planes of the test compressor. towards the stall limit. Since the overall static pressure ratio is the
product of the single pressure ratios, its maximum indicates one or
A portion of the mesh is shown in Figure 5. Every single row is more blade rows with separated flow.
resolved with about 160000 cells, including nine radial cell layers for
the rotor tip clearance. The total compressor is meshed with 1.8 Depending on the system downstream of the compressor,
million cell volumes. The turbulence model is the algebraic Baldwin- rotating stall and/or surge occur at the left limit of the operation line
Lomax model, because in comparison with multiple-equation models [7]. If the volume between compressor discharge and throttle can
it is unsusceptible to numerical instability and accelerates the store enough compressed gas, the compressor will tend to surge at
computation . The computation of one operating point takes half a day low frequency. If there is effectively no volume between the discharge
on a single-processor workstation equipped with a 750 MHz and the throttle as in the present research compressor, a closing of the
processor and 2.5 Gb memory. outlet throttle does not result in surge but in rotating stall. This
property qualifies the present compressor test rig for investigations on
flow separation and rotating stall without running the risk of
damaging the rig.
measurement planes

„near stall“ „maximum static

pressure ratio“
rotating
static pressure ratio

Inlet boundary: outlet boundary: stall

total pressure=101300 Pa static pressure + radial equilibrium
total temperature=288,15 K
axial flow
„design point“
Figure 4: Meridional view of the computational domain

rotor 4 stator 4
traversing points
stator 3

„choke“
Wandruck
radiales Gleichgewicht massflow
Figure 6: Characteristic of the present axial compressor

Mechanism of the bowed vanes

Figure 5: portion of the mesh
At the front (acceleration) part of the suction side compressive
normal blade forces act on the flow. In the case of positively leaned
RESULTS or bowed vanes, a radial component is superimposed on these forces
which drives the flow towards the mid-span section of the blade.
Characteristics of the present axial compressor Therefore, the flow near mid span of a bowed vane is faster than in
Figure 6 shows a typical constant-speed operating line of the the reference case while the velocity in the regions near the hub or
present research compressor in terms of static pressure ratio. The shroud is reduced. At the aft (deceleration) part of the blade the fluid
static pressure ratio is the static outlet pressure divided by the static is driven from the mid span section towards the end walls by the
inlet pressure. Closing of the outlet throttle of the compressor leads to suction normal blade forces. The higher acceleration in the front part
a rising static outlet pressure and a decreasing mass flow. Because the at mid span implies a higher pressure rise and thus tends to cause
inlet dynamic pressure decreases with decreasing mass flow, the separation in the deceleration part. Near the end walls, the effect is
static pressure at the compressor inlet rises, if the inlet stagnation vice versa: The adverse pressure gradient in the deceleration part of
pressure is held constant as is the case for each constant speed the suction side is reduced and so is the tendency towards separation.
operating line. While the static pressures behind each row of rotor In particular near the hub, the tendency toward corner stall is
blades and stator vanes, and at the outlet of the compressor rises up to reduced.
the point at which the compressor stalls, the static pressure ratios of
the individual stages and their product, the overall static pressure rise

3 Copyright © 2003 by ASME

Downloaded From: [Link] on 02/03/2016 Terms of Use: [Link]

 reference and the bowed vane of the third stator at the design point.
The dark grey and black zones indicate low pressure due to high
acceleration. The flow is thus deflected from the end wall toward
reference bow mid-span in the front portion of the blade. The dotted lines in Figure
7 represent 15%, 50% and 85% span. The relevant operating point is
indicated in the pressure ratio vs. mass flow diagram (from Fig. 6) in
Figures 7 through 11. In Figures 8 and 9, the computed and
measured (normalized with compressor discharge pressure) surface
pressure distributions of the reference and the bow blading are shown
for these three spanwise locations. In comparison with the reference
vane, the acceleration at the front portion of the bowed vane mid-span
section increases while the acceleration is reduced in near wall zones
of the bowed vanes when compared to the reference vane. This effect
of reduced pressure decrease is more pronounced near the hub
because of the lesser bow angle and the greater pitch to chord ratio at
the shroud. Downstream from the pressure minimum on the suction
Figure 7: Pressure distribution on suction side of stator 3 side, the flow is decelerated and the pressure rises. In this
deceleration area, where flow separation occurs if the gradient of
pressure rise becomes too large, the gradient at mid span is increased
by the bow while the gradient near the hub zone is decreased by bow.
computation 85% blade hight
computation 50% blade hight
computation 15% blade hight
0.9
measurement 85% blade hight
Figure 10 shows the effect of bow in stator 3 on the axial
measurement 50% blade hight velocity in the axial gap between stator 3 and rotor 4 at the design
measurement 15% blade hight
point. The values of velocities are circumferentially averaged from a
standardized static pressure

0.8
15x15 point traversing grid covering one pitch by a pneumatic four-
hole probe which includes a thermocouple. Due to the higher
0.7 accelaration at the mid span of the bowed vane, which causes a shift
of mass flow towards mid span, the axial velocity at mid section of
the duct increases, while the axial velocity near the hub decreases as
0.6 a result of the lower acceleration here. Because of the higher axial
velocities at mid span of the bowed vane the incidence angle at the
inlet of the following rotor blade will decrease and the aerodynamic
0.5
0.0 0.2 0.4 0.6 0.8 1.0 loading of the rotor will be reduced
distance from the leading edge / chord length

1.0
Figure 8: Normalized surface pressure on the reference vane
normalized blade height

0.8

0.6 reference
computation 85% blade hight
bow
computation 50% blade hight
0.9
computation 15% blade hight
0.4
measurement 85% blade hight
measurement 50% blade hight
measurement 15% blade hight
standardized static pressure

0.2
0.8

0.0
0.90 0.95 1.00 1.05 1.10 1.15
0.7 circumferential averaged axial velocity /
pitch averaged axial velocity (design point, reference compressor)

0.6
Figure 10: Circumferentially averaged axial velocity at design point

0.5 In summary, stator 3 shows two exemplary main effects of

0.0 0.2 0.4 0.6 0.8 1.0
bowed vanes. Because the pressure rise from minimum pressure to
distance from the leading edge / chord length
trailing edge pressure of the bow-bladed near-hub section is less, the
tendency towards separation is reduced. Near mid span, conversly the
Figure 9: Normalized surface pressure on the bowed vane tendency towards flow separation increases. The second effect of bow
on stator 3 is a displacement of mass flow from the near-wall zones
The mechanisms of bow in a vane are demonstrated by stator 3 towards the core flow zone in the deceleration (aft) portion of the
because the vanes of stator 3 have pressure taps. Figure 7 shows the vane, resulting in a lower axial velocity of the near wall zones and a
computed static pressure distribution on the suction side of the

4 Copyright © 2003 by ASME

Downloaded From: [Link] on 02/03/2016 Terms of Use: [Link]

 higher axial velocity in the mid section in the following axial gap at operation point big separation zones can be seen in both, the
the design operating point. reference and the bowed vane designs.

Effects of the present bow on the flowfield In the reference blading, corner stall extends from the hub through
Figure 11 and Figure 12 show the results of flow field traverses mid span. In the bowed blading, the loading of the vane and therefore
on a 12x12 point grid for the reference and the bowed vane the separation is shifted from the hub towards mid span. This
downstream of the fourth stator at different operation points. Figure separation zone is removed from the hub so that there is no merging
11 is a plot of total pressure levels and velocity vectors at the design of corner stall separation with the separation on the suction side of
point of the compressor. Figure 12 shows the same diagram at an the blade. The pitch-averaged value of total pressure of the bowed
operating point near stall. High values of total pressure show zones of vane becomes higher, as indicated by the performance map in Figure
undisturbed flow, low values of total pressure show zones of high 16. The effect of moving the loading and separation towards mid span
losses like boundary layers, wake regions, or zones of separation. is smaller in the third stator. Flow field data of the third stator are not
shown in this paper. In the third row of vanes, the loading and the
shape of separation zones is also moved towards mid span but in
contrast to stator 4, the separation at the hub and the corner stall is
reference bow not suppressed but merely reduced.

Effects of the present bow on stage and blade-row

performance

normalized static pressure ratio of stages 1.02

1.00

0.98

0.96 reference stage 3

bow stage 3
low total pressure high
0.94 reference stage 4
bow stage 4
Figure 11: Flow field downstream of stator 4, “design point”
0.92

0.90
0.80 0.85 0.90 0.95 1.00 1.05
reference bow
normalized mass flow

Figure 13: Stage static pressure ratio of the reference and the bowed
design
ss ps ss ps
normalized static pressure ratio of blade row

maximum pressure ratio

1.02

1.00

low total pressure high reference stator 3

bow stator 3
0.98
reference stator 4
Figure 12: Flow field downstream of stator 4, “near stall” bow stator 4

At the design point, zones of losses are small. At the hub and the 0.96
shroud of the reference design, small regions of thick boundary layers 0.80 0.85 0.90 0.95 1.00 1.05

or small separation zones occur. These zones are reduced in the normalized mass flow
bowed stator vanes. The distinction of momentum deficit due to
boundary-layer losses, separation losses, and wake losses was made Figure 14: Static pressure ratio of reference and bowed stator vanes 3
in the present study by repeated flow field measurements along the and 4
complete constant-speed operating line from choke to stall. The plots
are not shown here due to space constraints. In the near stall Figure 13 shows the static pressure rise of each stage in the
reference and the bowed vane configuration at design speed. All

5 Copyright © 2003 by ASME

Downloaded From: [Link] on 02/03/2016 Terms of Use: [Link]

 values in Figures 13 to 17 were normalized with the design point affect this limiting mass flow. At the stall limit, the pressure rise of
conditions of the reference compressor. For the bowed configuration the bow-bladed compressor exceeds the pressure rise of the reference
the pressure rise shifts such that the pressure rise of the stage 4 with compressor by 1.4%.
bowed vanes increases while the pressure rise of stage 3 with bowed
vanes decreases when compared with the reference design. The loss
of pressure rise in bowed stator 3 and stage 3 is affected by the shift 1.10
of mass flow to mid span and by the resulting higher velocity level in
the mid span section which in turn results in lower deceleration. The

normalized total pressure ratio

1.05
higher axial velocity in the core flow at the outlet plane of stator 3
reduces the incidence inlet flow angle of rotor 4 at mid span and so
1.00
effects a lesser loading of rotor 4. The incidence flow angle of stator 4 reference measurement
increases which leads to an increased pressure ratio for the bowed reference computation
bow measurement
stator 4 when compared with the reference. This effect decreases for 0.95 bow computation
decreasing mass flow, presumably because the acceleration zone of
stator vanes becomes smaller for lower mass flow due to increasing 0.90
inlet incidence angle.
0.85
Figure 14 shows the pressure ratio of the individual vanes of 0.85 0.90 0.95 1.00 1.05
stator 3 and stator 4. The relative changes of the constant speed lines normalized mass flow
show the effects of the bowed stator vanes on the stator characteristic
which are caused by the reduced separation. After passing the point
of maximum pressure ratio towards lower mass flow, the loss of Figure 16: Overall total pressure ratio
pressure ratio is reduced for the bowed stator vane 3 compared to the
reference design. This is achieved by a reduction in flow separation.
The loss of pressure ratio of stator 4, in which the corner stall was not reference measurement
just reduced but removed, is not just reduced but the point of 1.01 reference computation
maximum pressure ratio is also moved towards lower mass flow. bow measurement
normalized isentropic efficiency

bow computation
1.00

Effects of bow on compressor performance

0.99

0.98

0.97
rotating stall
normalized static pressure ratio

1.0
0.96

reference 0.95
bow
0.9 0.94
0.85 0.90 0.95 1.00 1.05

normalized mass flow

0.8
0.80 0.85 0.90 0.95 1.00 1.05
Figure 17: Isentropic efficiency
normalized mass flow
Figure 16 shows the measured and computed overall total
pressure ratio characteristic of the compressor at design speed with
Figure 15: Overall static pressure ratio the reference and the bowed vane design. Figure 17 shows the
corresponding isentropic efficiencies. The left-most measurement
Figure 15 shows the static pressure ratio of the reference and points (at normalized mass flow = 0.9) in Figure 16 and Figure 17
the bowed compressor blading. The pressure ratio at constant speed are the last operating points where the flow probe could be traversed
of the bowed stator is reflected in the characteristics of the overall without the risk of initiating compressor instability by the blockage,
compressor. Beyond the maximum pressure ratio towards lower mass caused by the probe itself. The relative error (in terms of 95%-
flow, the decrease of static pressure ratio is reduced by bow. From a confidence intervals) for mass flow is ±0.5%. The relative error for
normalized mass flow of 0.975 to 0.925, the constant speed line of total pressure is ±0.32% and for efficiency it is ±0.4%-age points.
the compressor with bowed vanes runs almost horizontal, indicating From choke (normalized mass flow=1.025) to the point of maximum
that the reduction of flow separation by the bowed vanes results in a static pressure ratio (normalized mass flow=0.975), the total pressure
higher static pressure rise if the machine is highly loaded. At a mass ratio and isentropic efficiency of the reference compressor is higher.
flow of 0.83, stable operation of the compressor is limited. Exceeding At the point of maximum static pressure ratio, the measured lines
this limit leads to rotating stall, initialized in the first stage [8]. Since overlap. From the point of maximum static pressure to the highly
the first stage is initializing the rotating stall, the bowed vanes do not loaded point near stall (normalized mass flow=0.9), the total pressure

6 Copyright © 2003 by ASME

Downloaded From: [Link] on 02/03/2016 Terms of Use: [Link]

 ratio and isentropic efficiency of the compressor with bowed vanes is ACKNOWLEDGMENTS
higher by comparison. If not highly loaded, the higher losses of the This report is the scientific result of a research project,
bowed vanes are plausible, considering that the surface area of the sponsored by the FVV (“Forschungsvereinigung
bowed vane is about 7.5% larger in comparison with the basic vanes. Verbrennungskraftmaschinen e.V.” = Combustion Machinery
The higher isentropic efficiency and overall total pressure of the Research Association, Frankfurt, Germany) and carried out at the
bowed blading, when it is highly loaded, confirms the flow field Turbomachinery Laboratory, University of Hannover, Germany, under
measurements which showed reduction of the corner stall (Fig. 12). the direction of first Prof. W. Riess and then Prof. J.R. Seume. This
work was partially (50%) financed by the German Ministry for
The constant-speed lines of the Navier-Stokes computations Economics (BMWi) through the Society for Industry Research (AiF),
show, in comparison with the measurements, a displacement towards (AiF-Nr. 067370).
lower mass flow of about 1.5%. The loss of isentropic efficiency and The FVV would like to thank the initiator Prof. Riess, the
overall total pressure of the bowed vane design compared to the advisor Prof. Seume, and the research assistant A. Fischer for the
reference compressor from choke to the point of maximum static completion of the present project.
pressure rise match the measured results. The left-most points of the The authors would like to thank Berhard Kuesters, Siemens
computations represent the respective points of maximum static Power Generation, Muelheim, for giving helpful advice and for
pressure rise for each constant speed line. As described above, the guiding the working group “FVV Bow-Blading” and Uwe Schmidt-
static pressure defines the outlet boundary of the computation. If the Eisenlohr, MTU Aero Engines, Munich, for the aerodynamic design
static pressure at the outlet boundary is raised further, the of the compressor blading and for giving helpful advice.
computation becomes unstable. Figures 16 and 17 show that the
computed overall total pressure rise and the isentropic efficiency to
the left side of the measured point of maximum pressure ratio are too REFERENCES
high, respectively. This failure is caused by the inability of the [1] Joslyn, H. D., 1985, “Axial Compressor Stator
present computation to capture the separated flows in the last stator Aerodynamics”, ASME Journal of Engeneering for Gas
rows. Before the onset of the separation in the vanes, the Turbines and Power, Vol. 107, April 1985
computations become unstable.
[2] Breugelmans, F.A.H., et. al., 1984, “Influence of Dihedral
CONCLUSIONS on the Secondary Flow in a Two-Dimensional Compressor
Numerical and experimental results show that through a 35° bow Cascade”, ASME Journal of Engineering for Gas Turbines
angle at the hub and a 30° bow angle at the shroud, bowed stator and Power, Vol. 106, July 1984
vanes shift the flow from the near-wall regions to the mid-span of the
vanes in the front portion of the vanes. This leads to lower loading [3] Shang, E., et. al., 1993, “The Experimental Investigations
near the wall at the deceleration portion of the vane and reduces or on the Compressor Cascades with Leaned and Curved
eliminates corner stall at the hub, when compared with the reference, Blade”, ASME, Cincinnati, May 1993, 93-GT-50
i.e. the radially stacked blade design. At the shroud, the effect is
insignificant, because for conventional compressor designs, [4] Weingold, H. D. et. al., 1995, “Reduction of compressor
separation first arises at the hub. The higher loading of the mid-span stator endwall losses through the use of bowed stators”,
section of the bowed vane leads to earlier separation in this area. The ASME Gas Turbine and Aeroengine Congress and
increase of mass flow and thus the increase of axial velocity at mid Exposition, Huston, 95-GT-380
span of stator 3 leads to a lesser loading of the following rotor 4 and
to a shift of static pressure rise from stator 3 to the last stator 4. [5] Walkenhorst, J., 2000, “Axialverdichter mit
Wandkonturierten Leitschaufeln – Prüfstandsentwicklung
The reduction of separation leads to an increase in overall static und Strömungsuntersuchung”, “Axial Compressor with
pressure rise, total pressure rise, and overall efficiency, when the end-wall shaped stator vanes”, Ph.D. thesis, University of
compressor is highly loaded. For moderate loading, the static Hannover, Germany and VDI-Bericht Reihe 7, Nr. 390,
pressure rise does not change in comparison with the reference Duesseldorf, Germany
blading design but efficiency is reduced because of higher friction
losses caused by the larger in surface area of the bowed vanes. The [6] NUMECA Int., 2000, “Numeca’s Flow Integrated
constant-speed operating lines of bowed stators change so that in the Environment for Turbomachinery and Internal Flows”, User
case of stator 3, the decrease of static pressure ratio is reduced for Manual, Numeca Int., Brussels, Belgium
high loading. In case of stator 4, the maximum static pressure ratio
moves to lower mass flow for bowed vanes. At the stall limit, the [7] Cumpsty, N. A., 1989,“Compressor Aerodynamics”,
static pressure rise of the bow-bladed compressor exceeds the static Longman Singapore Publishers Ltd., Singapore, ISBN 0-
pressure rise of the reference compressor by 1.4%. In future engine 582-01364-X
designs, this effect can be used move the surge limit towards higher
static pressures. The loss of efficiency may be compensated by [8] Hellmich, B. et. al., 2002 “Estimation of the rotating stall
reducing the number of vanes which then need to achieve a higher frequency from rotor frequency and common blade number
pressure rise, which, as shown above, can be achieved by bowed integer fractions in a four stage axial compressor”, Sixth
vanes. Int. Conf. On Rotor Dynamics, Proceedings Vol. II, Sydney

7 Copyright © 2003 by ASME

Downloaded From: [Link] on 02/03/2016 Terms of Use: [Link]