International Journal of Scientific Engineering and Applied Science (IJSEAS) – Volume-3, Issue-12, Dec 2017

ISSN: 2395-3470
www.ijseas.com

Experimental Study of the Flow in a Linear Cascade of Axial

Compressor Blades
Miguel Toledo-Velázquez, Guilibaldo Tolentino-Eslava, Miguel Leonardo Cervera-Morales, Juan
Abugaber-Francis, Luis René Rangel-López
Instituto Politécnico Nacional, Escuela Superior de Ingeniería Mecánica y Eléctrica. Sección de Estudios de Posgrado e
Investigación. Laboratorio de Ingeniería Térmica e Hidráulica Aplicada (LABINTHAP). Unidad Profesional “Adolfo
López Mateos”, Edif. 5, 3er piso, Col. Lindavista, C.P 07738, Ciudad de México, México.

ABSTRACT blades that it agrees to the compressor it is preferable

This article presents an experimental study of the that the Layer Limit is as small as possible, so that
performance of the boundary layer in the meridional there is a major effective area in the step of flow
line, measured on different planes of the suction and without it is disturbed, causing a major alignment in
pressure surface of the blade. Besides the tip leakage the lines of current [2].
of the blade, with tip clearances of 0% and 1% of the In the environment of real compressors,
blade span were measured. The experiment was measurements detailed in blades of stator they show
carried out in a linear cascade of 5 axial compressor layers limits to laminate extended on 30-50 % of the
blades. With blade chord of 0.2 m and blade span of rope of the blade. To numbers of Reynolds of
0.4 m, placed into the suction test section of a low 0.8x106, a separation to laminate with turbulent
speed wind tunnel, with a Reynold`s number of 3.62 transition is visible in the side of sucking, which does
5
x 10 . It were tested two conditions, corresponding at not disappear when the turbulence of the free current
inlet angle of 29.44º y 34.44º. The measurements is increased to 3 % [3]. On the other hand, in the
were made with hot wire anemometry, with the majority of the rotors of the turbo machinery, the
general purpose probe 55P11and the boundary layer space between the blade and the wall of the casing
probe 55P15, besides static pressure tappings in the induces a flow of escape across the space. This flow
wall. From the experiments were obtained the also happens in stages of the stator with the space
velocity vectors and pressure contour in one close to the rotor. The dissipation and miscellany of
dimension, that show the performance of the the flow of escape and whirlpool introduce
boundary layer and downstream flow in the aerodynamic losses and inefficiency. The distribution
meriodional region of the blade, so the vortex of the of losses indicates that the flow of escape includes
tip leakage close to 16 18 % of the height of the blade, from the
Keywords: Blade of axial compressor; Boundary top [4]. Also, the flow close to the surface of the top,
layer; Tip leakage; Flow linear cascade. diverges from the half of the profile of the blade and
produces a whirlpool of separation of the top in the
1. INTRODUCTION entry of the spread. This whirlpool increases in size
The flow in the region of the southern line and the top and forces along the rope, and due to the area of
of the blade is very complex, due to the interaction of contraction in the spread, the fluid moving across the
several such viscous phenomena as layer limit of the spread, one accelerates [5]. One of the principal skills
blade and flows of loss in the top. This flow is in the development of design of compressors is the
responsible for a significant portion of losses and study of cascades across wind tunnels, since it allows
reduction of load of blades of compressors, turbines realizing a detailed study of the mechanism of loss in
and bombs. In case of compressors, it produces these areas [6]. For this reason, a cascade of blades
blockade of the step of flow and distorts the will settle, in the wind tunnel of the Laboratory of
distribution of the angle of current flow below of the Thermal Engineering and Applied Hydraulics, to
rotor [1]. The losses that they increase with these realize a series of experiments and to determine the
phenomena are the source of major inefficiency in the characteristics of flow and losses related in the
rotor and of the increase of noise. In case of the southern line and top of the blade.

66
 International Journal of Scientific Engineering and Applied Science (IJSEAS) – Volume-3, Issue-12, Dec 2017
ISSN: 2395-3470
www.ijseas.com

2. THEORETICAL FOUNDATIONS 3. SELECTION AND DESIGN OF THE

The flows in turbomachinery are the most complex CASCADE
found in the practice of fluid dynamics and are The blade that was used is a stator vane of the
always three-dimensional, viscous and unstable [7]. compressor´s first stage of Ruston gas turbine. The
For example, the secondary flow increases are due to compressor blade was digitalized by optical means, to
the viscous boundary layer, but in the analysis of their obtain the coordinates and have the exact
effects are treated as non-viscous. Some three- correspondence between the actual blade and model.
dimensional not viscous effects are due to: This we obtained the coordinates of the middle
section of the blade. Once the coordinates are
1. Compressibility, radial density and pressure obtained we proceeded to build the aluminum blade
gradients. for experimentation, through CNC. It was took the
2. Radial change in thickness and geometry of the plane that places to 43 mm from the base of the blade.
blade. This profile is scaled to 7 times the actual size for the
3. Presence of finite walls of shafts and casings, shaft installation blade. The blades have a chord of 199.94
area change, curvature and rotation. mm, a height of 400 mm and thickness is 25.03 mm.
4. Working entering or leaving that varies radially. So the parameters for the cascade of 5 blades are
5. Presence of two-phase flow, and refrigerant shown in Table 1.
injection.
6. Leakage flow due to the clearing at the tip and 4. EXPERIMENTAL DEVELOPMENT
axial spaces. The wind tunnel that is used which is of the suction
7. Regions of mixed flow. type is located in the LABINTHAP. The test section
8. Secondary flow caused by input velocity gradients, for the experiment was modified to obtain the
stagnation pressure and angle of rotation. different configurations of the blades cascades. At the
bottom section, the acrylic is formed by a circular
Studies by Prandtl showed that a in a body immersed cross section to the rotation of the cascade, that
in a fluid in motion all the frictional energy losses section has 5 holes for the rotatory base of the blade.
occur in a thin layer adjacent to the surface in contact To achieve the different spacings on the tip of the
with the fluid. This layer is known as boundary layer. blade, the module has an intermedium wall of acrylic,
The flow outside the boundary layer can be with 121 holes for static pressure taps arranged in an
considered as lacking in viscosity. As small errors in array of 11 x 11. To measure the effect of leakage
the alignment of the blades results in a boundary flow on the tip were measured the velocity and
layer separation earlier than expected causing the turbulence profiles downstream, as the static
thickness increases very quickly resulting in pressures on the intermedium wall. The conditions to
widespread inefficiencies due to aerodynamic effects. measure are spacings of 0 and 1 of de blade height (0
Moreover, the stage blades rotate near the stationary and 4 mm of spacing), and angles of entry flow of
outer wall, or casing seal. Furthermore, the flow 29.44 ° and 34.44 °. For the boundary layer, the
condition when pass the crown of blades creates a measurement planes are composed of two sets of
pressure difference between pressure and suction planes. The first series consists of the entry plane
sides of blade, which induces a leakage flow through located at a blade chord of the attack edge;
the spacing or clearance between the blade tip and downstream five planes are located at 1, 1.25, 1.5,
seal. A vortex is generated between the jet and the 1.75 and 2 cords of the attack edge of the blade. The
main flow and is wound with its induced velocity to second set of measurement planes is composed by the
form a “tip leakage vortex”, as shown in Figure 1. measurement planes of boundary layer, distributed
Although the effects of other flows are appreciable, along the suction side and pressure. These planes are
the flow of tip plays an important role in the flow located with a distribution of 5, 10, 20, 30, 40, 50, 60,
phenomenon in this region, where raises an important 70, 80, 90 and 100% of the blade string starting at the
part of the losses. attack edge. In the case of the tip it was choosed to
measure five transversal planes downstream of the
central blade of the cascade, that is, at 1.125c, 1.25C,
1.5c, 1.75cy 2c, in corresponding to the measuring

67
 International Journal of Scientific Engineering and Applied Science (IJSEAS) – Volume-3, Issue-12, Dec 2017
ISSN: 2395-3470
www.ijseas.com

transversal planes 1, 2 , 3, 4 and 5, respectively, downstream flow similar to previous condition. The
shown in Figure 2. Each transversal measurement no-symmetry profile of the downstream flow 1c for
aligned with the cascade angle, covered about 90 mm 29.440 is because of the boundary layer of the
transversely of the flow channel of the pressure side, pressure side, while downstream flow for 34.440 the
and 110 mm for the suction side, taking deviation is in opposite direction because the biggest
measurements every 0.008 m. For the two cases of thickness of the boundary layer is in suction side. The
spacing, in addition to the transversal measurements difference of suction side separation point of 58% at
it was measured in 2 planes. Considering the blade 34.44º, in relation with 77% at 29.44º, is because the
height of 0.4 m, the foreground is located at difference between the blade outlet angle and the
0.97625% of the height (z = 0.97625) from the base, flow in the wind tunnel. For the pressure side,
the second at z = 0.96125. separation point is longer at 34.440 because the
difference between blade inlet angle and the flow in
5. INSTRUMENTS the wind tunnel is smaller.
Hot wire Anemometer brand Dantec model 90C10. Figures 5-6 show the difference in turbulence level
General purpose probe 55P11 for velocity and and boundary layer separation point giving higher
turbulence up and downstream from blade cascade. turbulence level in pressure side than the suction side.
Boundary layer probe 55P15. For velocity and For inlet point figures 6 and 7 show static pressure
turbulence mapping it was use positional system coefficient behavior at the wall. In Fig 6a it can be
brand Dantec model 41T50, measurement range x = observed that vortex tip leakage for 29.440 do not
0.590 m, y = 0.590 m y z = 0.690 m, resolution 12.5 match with next blade. But wall influence is bigger as
mm (x, y) and 6.25 mm (z). “L” Pitot tube, brand shown in 6b, where leakages stand from 30% to 81%
Airflow, 8 mm diameter, ellipsoidal pointer, one chord, extending downstream. Fig 7a shows for
orifice for total pressure and 7 orifices at 25mm each 34.440 an increase vortex because de incidence angle,
for static pressure, connected to two inclined an almost no influence in the others blades. Fig 7b
manometers brand Airflow type 4. vortex is moved toward inlet point, standing almost
For total pressure it was used manometer from 0 to half chord. For 29.440 wall influence is up to z=
2.5 kPa, for static pressure it was used manometer 0.96125. Fig 8c show vortex thickness gets 23mm
from 0 to 5.0 kPa. For static pressure it was used an between 1.5c and 1.75c and dissipates in fig 8d. For
intermediate acrylic wall with 1.75 mm orifices, 34.440 fig. 9a shows a bigger wall influence
where put 1.5875 mm external diameter brass tubes z=0.97625. or 1% space, fig 9c and d, vortex moves
(0.8763 mm internal diameter) perpendicular to wall. greatly toward suction side, getting influence of
These tubes connected through plastic hoses to a bigger thickness in every plane up to 40 mm.
specially constructed differential manometer.
Differential Manometer scale from 0 to 250 mm H2O,
with 121 independent U probes (one end open to
atmospheric pressure).

6. RESULTS
Figures 3-4 show velocity profiles of boundary layer
suction side and downstream flow in meridian zone
of the blade for 29.44° and 34.40 respectively. For
29.440 the maximum thick of boundary layer suction
side is δ11 = 5.799 mm, in pressure side δ11 = 10.564
mm, with separation at 77% (suction side) and less
than 20% (pressure side) of the chord with a
downstream flow of 0.025m (15% of the step). For
34.440 the maximum thick of boundary layer suction
side is δ11 = 9.096 mm, in pressure side δ10-11 =
7.83 mm, with separation at 58.14% (suction side)
and 27.94% (pressure side) of the chord. With a Fig. 1 Compressor blade three-dimensional flow

68
 International Journal of Scientific Engineering and Applied Science (IJSEAS) – Volume-3, Issue-12, Dec 2017
ISSN: 2395-3470
www.ijseas.com

Fig. 2 Measurements planes

Fig.5 Turbulence pattern 29.440

Fig. 3 Velocity profile, boundary layer and

downstream, 29.440

Fig. 6 Turbulence pattern 34.440

Fig. 4 Velocity profile, boundary layer and

downstream, 34.440

69
International Journal of Scientific Engineering and Applied Science (IJSEAS) – Volume-3, Issue-12, Dec 2017
ISSN: 2395-3470
www.ijseas.com

Fig.7 Static pressure coefficient behavior 29.440

Fig. 8 Static pressure coefficient behavior 34.440

Fig.9 Velocity Behavior 29.440

70
 International Journal of Scientific Engineering and Applied Science (IJSEAS) – Volume-3, Issue-12, Dec 2017
ISSN: 2395-3470
www.ijseas.com

Fig.10 Velocity Behavior 34.440

Table 1 Cascade parameters. chord was observed; this gets influence in the
Scale factor 7 distance of downstream flow affected for the high
Chord length 20 cm turbulence level caused by boundary layer. Which is
Step 16.66 cm for 29.440 conditions, bigger of 50% chord,
Chord / step 1.2 downstream from blade outlet point? For 34.440
Blade Length / chord 2 conditions, this length is 25% chord. Vortex tip
Re 3.62 x 10 5 leakage is figure out through pressure and velocity
Inlet outlet blade angle 150 mapping, vortex formation length for 29.440
Blade Position angle 29.589790 condition, 30% chord, and vortex thickness 40mm.
U ref 32.4 m/s This shows losses tip region, between blade tip and
flow middle Inlet angle 29.4490 wall.
Incidence angle 0.140790
APPENDIX: NOMENCLATURE
CONCLUSIONS
C = Rope
Measurements made in blade meridian region show
S= Spend between blades
importance of alignment between flow and blade inlet
U∞ = Speed of entry of the flow (m/s)
angle. For 29.440 conditions, a thickness of 5.79 mm
at separation point at 77% chord was observed. For α = Angle of entry and exit of the blade
34.440 conditions, at suction side, a maximum β = Angle of entry and exit of the flow
thickness of 9.09 mm at separation point at 58% 1 = Blade of number
δ = thickness of limit layer (mm)

71
 International Journal of Scientific Engineering and Applied Science (IJSEAS) – Volume-3, Issue-12, Dec 2017
ISSN: 2395-3470
www.ijseas.com

x = Axial direction Compressor Rotor, Journal of Engineering for

y = Tangential direction Gas Turbines and Power, ASME.21-30.
z= Blade height (m) 5. Kang, S. Hirsch, C. Octubre, 1994. Tip Leakage
Re = Reynolds number Flow in Linear Compressor Cascade. Journal of
Tu = Turbulence intensity (%) Turbomachinery, ASME. 657-664.
6. Hodson, H. Abril, 1984. Boundary Layer and
REFERENCES: Loss Measurement on the rotor of an Axial-Flow
1. Lakshminarayana, B. 1996. Fluid Dynamics and Turbine. Journal of Engineering for Gas Turbines
Heat Transfer of Turbomachinery. John Wiley & and Power, ASME. 391
Sons, Inc. New York. 505-510. 7. Lakshminarayana, B. Abril, 2001. Numerical
2. Hawthorne, W. 1964. Aerodynamics of Turbines Simulation of Tip Leakage Flows in Axial Flow
and Compressors. Princeton. New Jersey. 70-79. Turbines. with Emphasis on Flow Physics: Part I-
Effect of Tip Clearance Height. Transactions of
3. Badran, O. Bruun, H. Julio, 2002. Experimental the ASME. 314-323.
Investigation of Turbulent Flor with Separation
Over a NACA 4412 Airfoil at Angle of Attack 20
Degree. Proceedings of ASME. 1-8.
4. Lakshminarayana, B. Enero,1986. End-Wall and
Profile Losses in a Low-Speed Axial Flow

72