Effect of Turning Angle On Flow Field Performance of Linear Bowed Stator in Compressor at Low Mach Number
Effect of Turning Angle On Flow Field Performance of Linear Bowed Stator in Compressor at Low Mach Number
= = a R .
2 2
sin / ) (
a b R =
1
=0.6
Fig.2 Aerofoil configuration and blade stacking line
November 2006 Effect of Turning Angles on Flow Field Performance of Linear Bowed Stator in Compressor at Low Mach 273
will be gained; and then an aerofoil line can be got
through adding NACA65 aerofoil thickness to the
middle arc line. Finally, a computational cascade
can also be obtained by using stagger angles for-
mula
2 T s 2
= , where = 75 . 9
2 T
according
to the experiment
[8]
.
As shown in Fig.2 (b), the bowed blade stack-
ing line is composed of three straight lines, among
which a line is connected to its neighbor line with
smooth curvature. The bowed blade is given by
moving the end-wall sections perpendicular to the
axis. The bow angle is defined by the angle
between the pressure surface and the normal of
end-wall.
2 Comparison of Results of
Computation and Experiment
In the following text, a comparison of computa-
tional and experimental
[7]
results is achieved to vali-
date the reliability of computation which is carried out
with the same geometry, aerodynamic parameters and
boundary conditions as the experiment. The com-
parison parameters are listed in Table 1 and other
geometric parameters are shown in Table 2.
Table 1 Geometry and aerodynamic parameters
in comparison
Experiment Computation
Inlet geometry angle ) /(
2 T
49.75
Turning angle ) /( 59.5
Stagger ) /(
s
20
Incidence ) /( i 0
Renault number Re
5
10 50 . 6
5
10 53 . 6
Inlet Boundary layer thickness /mm 20.50 20.55
Displacement thickness /mm
*
2.50 2.25
Momentumloss thickness /mm
* *
1.75 1.79
Bow angle ) ( / 0 25
A straight rectangular cascade and a bowed
cascade with bow angle of 25 were studied in the
experiment. The total energy loss coefficients are
compared in Table 3. And energy loss coefficient
distributions along blade height are compared in
Fig.3. Here the coefficient is defined as
7 258 . 0 *
0 s
7 258 . 0 *
0 s
7 258 . 0 *
s
) ( 1
) ( ) (
p p
p p p p
= 3
Table 2 Geometry parameters of computation cascade
Chord length b/cm 16
Aspect ratio h/b 1.00
Axial chord length B/cm 15
Exit geometry angle
) /(
2 T
-9.75
Turning angle ) /( 59.5 (54.5 49.5 44.5)
Bow angle ) /( 5 (10 15 20 25 30)
Pitch length t/cm 8
Blade height h/cm 16
Solidity b/t 2.00
Table 3 Comparison of total mass-average energy loss
coefficients of computation and experiment
Straight
blade
Bowed
blade
Relative
reduction / %
Experiment
0.137 0.119 13
Computation 0.110 0.098 11
where p
s
is the static pressure; p
*
is the local total
pressure and
*
0
p is the inlet total pressure. The
variation trends of the coefficient for both re-
sults are fairly similar except for some differences in
the absolute value, where exist possible reasons that
there are some errors in the experiment measure-
ments and computation codes and there are some
differences between the post-process methods of the
two research techniques. The variations of energy
loss coefficient of bowed cascade show the bowed
blade improves the cascade performance to high
extent.
Fig.3 Comparison of energy loss coefficient
distributions along blade height
274 ZHANG Yong-jun, CHEN Fu, FENG Guo-tai, SU J ie-xian CJ A
The contours of energy loss coefficients at exit
plane are given in Fig.4. The distributions obtained
from experiment and those obtained from computa-
tion are similar to each other. There are two high
loss regions near the lower corner and the upper
corner of the straight blade, while the two loss re-
gions of the bowed blade become much smaller.
Besides, the energy loss values and the positions of
the loss cores are changed: the values are reduced
by 9% and the positions moved to the middle span.
The contours of axial velocities near suction side
are shown in Fig 5. In incompressible flow motion,
it is known that the axial velocity decides the mag-
nitude of the flow capacity and the low velocity
region corresponds to the high loss region. As shown
in Fig.5, there are two back flow regions near the
suction/end-wall corners in the passage of straight
cascade. While the back flow regions are removed
after the bowed blade technique is applied. The com-
parison of experiment with computation indicates that
the computational results are believable and ac-
cepted.
3 Study on Losses of Bowed Blade in
Different Turning Angle Cascades
As learned from Fig.6, the effect of the bowed
blade in lowering loss of high turning angle cascades
is more apparent than the effects in low turning an-
gle cascades. In particular, the cascade loss coeffi-
cient of 30 bowed blade in 59.5 turning angle
cascade is less than that of straight blade in 54.5
turning angle cascade.
In the following discussions on the effects of
aerofoil turning angle, the abbreviations of STR595
and BOW595STR395, BOW395 stand for straight
blade with =59.5and bowed blade with =59.5
, traight blade with=39.5, bowed blade with
=39.5, respectively.
The contours of axial velocities and vectors of
secondary flow at exit plane in the two turning an-
Fig. 4 Contours of energy loss coefficients at exit plane
Fig.5 contours of main flow velocity at exit plane
Fig.6 Total energy loss coefficient of computational cascades
November 2006 Effect of Turning Angles on Flow Field Performance of Linear Bowed Stator in Compressor at Low Mach 275
gle cascades are compared in Fig.7. In Fig.7(a) and
(b)the areas of low speed flow and back flow in the
cascade BOW595 are evidently reduced and the
trend of high speed flows immigration is obviously
strengthened.
Accordingly, the effects of different aerofoil
turning angle cascades on corner separation are dif-
ferent. As shown in Fig.8(a), the separation line on
suction surface(S.S) of the cascade BOW595 is
much nearer to trailing edge than that of the cascade
STR595 and the separation region of BOW595 is
much smaller than that of STR595. As seen in
Fig.8(b), the variations of separation line and sepa-
ration region of BOW395 are comparatively
negligible to STR395.
4 Effects of Bowed Blade on Secondary
Flow in High Speed Fluid
As proved in present paper, a better explana-
tion for bowed blade to be effective in controlling
the end-wall flow is the rise of static pressure on the
end-walls together with the decrease of low energy
fluid mass near the end-walls caused by high speed
fluid immigration from the middle span. A sche-
matic model of bowed blades controlling secondary
flow is given in Fig.9. As shown in Fig.9(a), when
high speed fluid flows through the cascade passage,
it is subjected to a centrifugal force pointing to the
pressure surface(P.S) due to cascades turning. If the
cascade is straight, the secondary flow in high speed
fluid does nearly not immigrate to the end-walls (as
Fig.7 Contours of axial velocities and vectors of secondary
exit plane
Fig.8 Separation lines on suction surfaces
Fig.9 Effect model of bowed blade on secondary flow
276 ZHANG Yong-jun, CHEN Fu, FENG Guo-tai, SU J ie-xian CJ A
seen in Fig.9(b)). While if the cascade is bowed, the
composition of a bowed blade force and the cen-
trifugal force impels some high speed fluid to im-
migrate to the end-walls(as seen in Fig.9(c)) and the
pitched mass-averaged streamlines turn to the
end-walls in the rear part of passage (as seen in
Fig.10). These phenomena will result in: Qthe total
mass of high speed fluid is increased and the flow
speed is raised so as to effectively control the ac-
cumulation or separation of low energy fluid near
the end-walls; the rise of static pressure on the
end-walls strengthens the radial immigration and
weakens the crosswise immigration of low energy
fluid in boundary layer, thus the accumulation or
separation of low energy fluid is further controlled.
Theoretically, if the bow angle, bow height of
the bowed blade and the centrifugal force are
great, the high energy fluid easily immigrates to the
end-walls. But if the bow angle and height are ex-
tremely great, the loss of the surface friction obvi-
ously increases. Thereby in bowed cascade passage,
the immigration capacity of high speed fluid mainly
relies on the centrifugal force.
5 Effects of Bowed Blade on Different
Turning Angle Cascade
In Fig.11, a parameter mass-flow density ,
local density multiplied by local axial velocity
v
z
, is defined to describe the flow capacity, i.e., the
mass-flow in a unit area., As learned from Fig.11,
the mass-flow density along a radial line A-A for the
cascade BOW595 is much higher than that for the
cascade STR595. While for the cascade BOW395 the
mass-flow density is increased a little and the trend of
high speed fluids immigration capacity is weak.
The bowed blade can evidently control the
separation, and consequently, obviously lower the
losses at the suction and end-wall corners in high
turning angle cascade. The contours of total pres-
sure loss coefficients at outlets section are shown in
Fig.12. A high loss region with a loss coefficient of
0.700 in its core exists at the hub/suction corner and
expands to 25% span-wise position in the cascade
STR595. While in the cascade BOW595, there ap-
pear two high loss regions: the one is located at the
hub/suction corner with a loss coefficient of 0.550
in its core, the other is located at the small zone be-
tween 30% and 40% span-wises also with a loss
coefficient of 0.550 in its core. But unlike the high
Fig.10 Pitch-wise mass-averaged streamlines in the
meridian planes of straight and bowed blades
Fig.11 Distributions of mass-flow densities along the line
Fig.12 Contours of energy loss coefficients close to suction
at exit section
November 2006 Effect of Turning Angles on Flow Field Performance of Linear Bowed Stator in Compressor at Low Mach 277
turning angle cascade, the cascade BOW395 hardly
improves the performance, and the total pressure loss
coefficient is kept as a value of 0.650 in the loss core
when compared to STR395. Actually the overall
losses of low turning angle bowed blade are increased
due to the increment of surface friction losses.
6 Conclusions
(1) The experimental and computational results
both verify that the bowed blade has obvious effect
in high turning angle cascade. The overall losses are
lowered by 13% and 11% in experiment and com-
putation respectively.
(2) In different turning angle cascades, the ef-
fects of bowed blade are different. In higher turning
angle cascade, the bowed blade can lower the loss
by 17.9% in maximum. While in lower turning an-
gle cascade, the bowed blade can lower the loss by
0.7% in maximum.
(3) In different turning angle cascades, the
bowed blade can strengthen the radial secondary
flow and weaken the traverse secondary flow. Both
of these two effects can lower the end-wall loss but
increase the mid-span loss, so they can not explain
why the effects of bowed blade are different in dif-
ferent turning angle cascades.
(4) An important effect of the bowed blade is to
strengthen the flow capacity near the end-walls, of
which the reasons include: Qradial immigration of
low energy fluid in boundary layer; high speed
fluids turning to the end-wall in meridian plane.
Thus the strong flow capacity near the end-wall
makes the anti-separation capacity stronger and thus
lowers the pressure-difference loss. This is the main
reason for the bowed blade to significantly lower
the losses in high turning angle cascade.
(5) In low turning angle cascade, the bowed blade
increases the mass-flow density comparatively less
and the corner separation is much weaker. So the bene-
fit of bowed blade in reducing losses is negligible.
References
[1] Deich M E, Gubalev A B, Filippov G A, et al. A new method of
profiling the guide vane cascade of stage with small ratios diameter
to length[J ]. Teplienergetika, 1962, 8: 42-46
[2] Weingold H D , Neubert R J , Behlke R F , et al. Reduction of
compressor stator endwall losses through the use of bowed sta-
tors[R]. ASME Paper 952GT2380, 1995.
[3] Zhao G J ,Chen F SONG Y P, et al. Experimental study on the aero-
dynamic performance of swept-Curved blade[J ]. Chinese J ournal
of Aeronautics, 2004, 17(3):136-141.
[4] Breugelmans F A H, Carles Y, Demuth M. Influence of dihedral
on secondary flow in a two dimension compressor cascade[J ].
ASME J ournal of Engineering for Gas Turbine and Power,
1984,106(3):578-584.
[5] Smith L H ,Yeh H. Sweep and dihedral effects in axial-flow tur-
bomachinery[J ]. J ournal of Basic Engineering,1963, 85: 401-416.
[6] Gmmer V, Wenger U.The impact of sweep and dihedral on axial
compressor end-wall aerodynamics[C]//The 8
th
International Sym-
posiumon Transport Phenomena and Dynamics of Rotating Ma-
chinery (ISROMAC). Honolulu, Hawaii, 2000: 26-30.
[7] Chorin A J . A numerical method for solving incompressible vis-
cous flow problems[J ]. J of Computation Phys, 1967,135(2):12-26.
[8] .)_,'|],[[,{[D].
]],,2001.
[8] Wang D. Experimental investigations of blade curving effect on
aerodynamic characteristics of compressor cascades[D]. Harbin:
Harbin Institute of technology, 2001.(in Chinese)
[9] ,,][,.|])|[Z].;:',}[
|_1999:197-198.
[9] Shu S Z, Zhu L, Ke X L, et al. Turbomachinery theory[Z]. Beijing:
Heat Energy Engineering Department of Tsinghua University,1999:
197-198.(in Chinese)
Biographies:
ZHANG Yong-jun Born in 1977, he received B.S. and
M.S. from Harbin Institute of Technology in 2000 and 2002
respectively. Now he is working for his doctoral degree in
Harbin Institute of Technology. Tel:(0451)86412433,
13030086504, E-mail:[email protected]
CHEN Fu Born in 1970 , he received the doctoral degree
fromHarbin Institute of Technology in 1997 and then became
a teacher there. Now he is working as an professor in Energy
Science and Engineering School of Harbin Institute of Tech-
nology. He has published several research papers in various
periodicals. Tel : (0451) 86412368,E-mail:[email protected]