compressor bu~des at~gi · "~1umbers in parenthesis rerer to references. 1 . viii literature...
TRANSCRIPT
![Page 1: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/1.jpg)
CASCADE PERFOR.'1.~NC£ OF DOUBLE CIRCULAR ARC
COMPRESSOR BU~DES . .11 Hl:GH At~GI.ES OF ATTACK
by
Peter T. Tkacik
Thesis submitted to the Graduate Faculty of the
Virginia Polytechnic Institute and State University
in partial fulfillment of the requirements for the degree of
APPROVED:
Master of Science
in
Mechanical Engineering
W. F. O'Brien, Jr .. Cha~rman
H. L. Moses S. 3. Thomason
May, 1982
Blacksburg, Virginia
![Page 2: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/2.jpg)
I I ACK~OWLEDGE:vrE:ns
The author expresses sinc2re ~ppreciation to the men-
bers of his advisory co::nnictee: Professors H. L. Xoses, S.
B. Thomason, and W. F. O'Brien, Jr., Chairman. Dr. O'Brien
was especially helpful throughout the investigation.
Thanks are extended to ?rofessor J. B. Jones and the
Mechanical Engineering Department for providing financial
support during the author's graduate study. He is also
grateful for the assistance of the :I.E. Workshop, S.
Reimers, and D. Bruce during the construction of the wind
tunnel. The author is also indebted to Neta Byerly for her
patient and professional typing of the thesis.
Finally, the author is very indebted to his parents for
their constant encouragement and support.
![Page 3: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/3.jpg)
III TABLE CF CONTE:~TS
I. TITLE
II. ACIGWw"'LEDGEMENTS
III. TABLE OF CONTENTS
IV. LIST OF FIGURES
V. LIST OF TABLES
VI. LIST OF STI1BOLS
VII. INTRODUCTION
VIII. LITERATDRE REVIEW
IX. EXPERIMENTAL EQUIPMENT
A. The Wind Tunnel .
B. The Cascade Test Section
C. The Blade Cascade .
D. The Measuring Equipment
X. RESULTS
XI. DISCUSSION
XII. CONCLUSIONS
XIII. RECOMMENDATIONS
XIV. REFERENCES
XV. APPE~DICES
A. Uncertainty Analysis
B. Inlet Velocity Measurements
c. Blade Production
iii
i
ii
iii
v
.viii
ix
1
7 .., /
13
16
17
21
51
. 58
. 59
. 61
. 63
. 64
. 7 3
. 76
![Page 4: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/4.jpg)
XVI. VITA
ABSTRACT
III TABLE OF CONTENTS (continued)
iv
78
![Page 5: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/5.jpg)
Figure
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
LIST OF FIGURES
Title
Tunnel Schematic
Inlet Filtration System
Inflated Tunnel Blower Junction with Pressure Gage . . . . . . . .
Blade Row Angle Guide, Probe Traverse and Stiffening Box
Blade Profile
The Blade Cascade Showing Close-Up of Instrumented Blade Passageway
Mean Deflection Angle
Mean Total Head Loss .
Stagnation Pressure Loss and Deflection Angle for ).l = 0° .......... .
Stagnation Pressure Loss and Deflection Angle for ).l = 8° .......... .
Stagnation Pressure Loss and Deflection Angle for ).l = 14° .......... .
Stagnation Pressure Loss and Deflection Angle for :i~ = 18° .......... .
J..
Stagnation Pressure Loss and Deflection Angle for :ii= 19° ..... .
Stagnation Pressure Loss and Deflection Angle for -,i = 20° . . . . .....
Stagnation Pressure Loss and Deflection Angle for ').i = 21° . . . . . . . . .
Stagnation Pressure Loss and Deplection Angle for :i. i = 22 ° . . . . . . . . .
v
8
9
11
14
18
20
22
23
25
26
? .., _,
23
29
30
31
. . 32
![Page 6: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/6.jpg)
Figure
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Al
LIST OF FIGURES (continued)
Title
Stagnation Pressure Loss and Deflection Angle for ::~ = 25° . . . . ...
.I.
Stagnation Pressure Loss and Deflection Angle for '.li = 28° ...
Stagnation Pressure Loss and Deflection Angle for :xi = 30 ° . . . . .
Stagnation Pressure Loss and Deflection Angle for '.li = 35° ...
Velocity Distribution for '.li = 0°
Velocity Distribution for '.ll = 8°
Velocity Distribution for '.li = 14°
Velocity Distribution for '.li = 18°
Velocity Distribution for :i.i = 19°
Velocity Distribution for '.li = 20°
Velocity Distribution for :i.i = 21'
Velocity Distribution for :i.i = 22°
Velocity Distribution for :i.i = 25°
Velocity Distribution for ~i = 28°
Velocity Distribution for
Velocity Distribution for
). I = 300 1
). I = 350 1
Flow Visualization Spanwise Across the Blade at :i.: = 14° . . . ....... . ... Flow Visualization Spanwise Across the Blade at ·:li = 28° . .
Cascade Notation
vi
33
34
35
36
38
39
40
41
42
43
44
45
46
47
48
49
53
55
65
![Page 7: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/7.jpg)
LIST OF FIGURES (continued)
Figure Title Page
A2 Nozzle Velocity Variation Above 65 m/s . . . 75
vii
![Page 8: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/8.jpg)
Table
Al
A2
V. LIST OF TABLES
Title
Instrument Readability and Accuracy
Nozzle Velocity Distribution
vi.ii
72
74
![Page 9: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/9.jpg)
i
L
p
s u
u v
v
x
y
:.1 I (ll
I :.12
'(
VI. LIST OF S'T.-'!BOLS
Incidence angle
Blade Height
Chord
Pressure
Pitch
Velocity component in the x-direction
Freestream velocity at the edge of the boundary layer
Velocity component in the y-direction
Freestream velocity
Air inlet velocity
Air outlet velocity 2 Total pressure loss (P01 - P02 /~oV1 )
~ean total pressure loss
Chordwise direction
Direction perpendicular to blade chord line
Air inlet angle
Flow in:et angle, angle of attack, relative to chcrd line
Air outlet angle
Flow outlet angle relative to chord line
Stagger angle
Deflection angle, ~ 1 - ~2
Mean deflection angle
Blade camber
ix
![Page 10: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/10.jpg)
VI. LIST OF SYMBOLS (continued)
Subscripts
o Signifies stagnation state
x
![Page 11: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/11.jpg)
VII INTRODUCTION
Because of the wide-range performance demands in cur-
rent turbomachines, there is a need for more complete infor-
mation on blade flow characteristics. Presently available
information on axial-flow compressor cascade performance
emphasizes design-point data. Recent interest in pre- and
post-stall behavior of axial-flow compressors has disclosed
a need for detailed experimental investigations of cascade
behavior at high angles of attack. This inf~rmation is
useful for improved understanding of the stall phenomenon,
and for input to numerical models of stalling behavior (1) "
To meet this need, the present investigation was con-
ducted with the goals of designing and developing a cascade
wind tunnel especially for high-angle-of-attack investi-
gations, and conducting an initial evaluation of the per-
formance of the tunnel and a representative set cf com-
pressor blade airfoils. Special attention was given to
measurenents evaluating inlet flow uniformity and turbulence
level, as ~vell as departures from tho-dimensional flow over
the blades at high angles of attack. An initial set of data
for a single stagger angle is presented.
_,_ "~1umbers in parenthesis rerer to references.
1
![Page 12: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/12.jpg)
VIII LITERATURE RE'!IEW
Cascade Wind Tunnels
Due to the complexity of the flow through an axial flow
ccmpressor, most research has been based on models which
simplify the study by investigating specific interactions.
A cormnon simplification is to ass'l.lllle two-dimensional flows
through the blade rows. Cascade tests are of this type,
and, while not addressing many of the phenomena in a com-
pressor such as tip clearance leakage, centrifugal effects,
etc., these tests have proved invaluable in the deter-
mination of turning angles and losses for different blade
profiles. Near design-point operation, compressor cascade
flows are nearly two-dimensional, and cascade tests have
served well as a basis for design.
Departures from two-dimensional flows in a cascade ar2
more severe at high angles of attack. Ikui, Inoue, and
Kuromaru ran high-stagger-angle tests in 1970 (2) which
required a large number of high-aspect-ratio blades in the
cascade in order to obtain two-dimensional flows in the
meas~rement region. As with high-stagger-angle tests, high
angle-of-attack tests also encounter problems with two-
dimensionality. Keller at VPI&SU in 1978 (3) noted endwall
effects in a ca3cade of double circular arc blades with an
aspect ~atio of 2.33 as the angle of attack increased. :n
2
![Page 13: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/13.jpg)
3
order tc extend the Keller tests at VPI&SU, a new wind
tunnel was built with modif icaticns similar to those of
Ikui, et al.
The design of the :~nnel (see Fig. 1) described in this
report was based on standard wind tunnel practices. As
described by R. D. Mehta in 1977 (4), a centrifugal blower
is used because of the cost, efficiency, and non-stalling
characteristics when heavily loaded. This blower is con-
nected to the tunnel with a flexible coupling to reduce
vibration transfer as proposed by R. D. Mehta and P. Brad-
shaw in 1970 (5). A tongue on the blower exit was also
modified according to reconnnendations of Mehta and Bradshaw
(5) .
From the blower, the flow enters a diffuser which has
screens to retard separation and the inherent tendency for
the flow to attach to one wall. These are arranged as
described by L. F. East in 1972 (6) and Mehta, et al. (5).
From the diffuser, the flow passes to a settling chamber.
Three screens are placed in series here to further smooth
the velocity profile and reduce boundary layers. These
scr-=ens are placed in accordance with recormnendations by :1.
A. Jackson in 1972 (7), although Mehta and Bradshaw later
recommended ir.creasing the distance between screens (5).
With a large settling chamber, a uniform velocity
prof~le can be obtained with a minimum of head loss. To
![Page 14: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/14.jpg)
4
obtain high velocities in the cascade test section, a six-
teen-to-one area ratio contraction was employed. This
contraction is characterized by an almost circular-arc
contour, matched to a short inverse circular arc at the exit
region. This was done in accordance with recommendations of
M. N. Mikhail (8).
The design of the tunnel provides for honeycomb flow
conditioner within the settling chamber. These honeycomb
sections were not available at the time of the initial tests.
Even without honeycombs in the settling chamber, this com-
bination resulted in a test-section entrance turbulence
level of only 1.5% and a maximum variation in the nozzle
exit velocity profile of only 1.3%. Furthermore, the flow
is filtered to allow hot-wire anemometry tests. Flow
velocities greater than sixty-five meters per second are
provided.
Other Related Topics
In order to understand some other phenomena that arise
in cascade testing, a library investigation was conducted
on three subjects. The first was blade flutter, which was
anticipated as a problem due to the blade's being pinned at
the centerline, which was not the center of aerodynamic
lift. This caused an extreme flutter problem in tests
described by L. Bcnciani, P. L. Ferrara, and A. Timori at
![Page 15: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/15.jpg)
5
Florence, Italy in 1980 (9), and J. Sparks at VPI&SU (10) i~
1981. In the present cascade, rubber dampers and extra
blade supports were installed to reduce the vibration to a
minimal level.
The second phenomenon investigated was endwall stall as
described by J. H. Horlock in 1973 (11). In the present
cascade, this problem arose at the base of the blades at
angles of attack approaching full stall. A gap at the full
span end of the blades as suggested by Dring prevented
endwall stall in the tip region.
Propagating stall was the third problem area of in-
vestigation. Much was learned about its cascade behavior in
a film by G. Sovran at the General :1otors Research Center
(12). Further investigations on propagating stall in double
circular arc blades were done by F. Cheers and L. E.
Macartney in 1959 (13) and simultaneously with the present
investigation by D. Mathioulakis of the ESM department at
VPI&SU in 1982 (14) .
For measurement of local static pressure, the blades
were manufactured with chordwise static pressure taps as
described by U. Keller (3). Downstream total and static
pressures were measured with a pitot-static probe as recom-
mended by D. J. Kettle in 1954 (15). Calculations of the
mean total-head-loss coefficier.t, :, and blade velocity
curves were done as proposed by S. Lieblein in NASA SP-36 in
![Page 16: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/16.jpg)
6
1965 (16). Lastly, the :low tu~ning angle was based on mass
flow and momentum measured in the blade row direction as
defined by G. C. Oates (17).
![Page 17: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/17.jpg)
IX EXPERI:~ENTAL EQL'IPMENT
A. Wind Tunnel
In order to provide the cascade with sufficient mass
flow and the related velocity, a wind-tunnel facility was
developed. This facility was constructed in the Turbo-
machinery Research Laboratory of the Mechanical Engineering
Department, and is shown in Fig. 1.
The tunnel inlet consists of a horizontal cylindrical
frame four-feet long and up to 45 inches in diameter. It is
mounted between the blower inlet and the laboratory wall and
is covered by a chicken wire mesh. The inlet is covered by
blue media filter material and provides a clean air flow to
the blower with no distortion due to the surroundings.
This filtration system is shown in detail in Fig. 2.
The blower is an Aerovent Model 630 BIA and is of the
centrifugal type. It is equipped with a 15-horsepower motor
and is rated at 12,000 CFM at a delivery pressure of eight-
inches of water. It is of steel construction and is bolted
through a wooden 4" x 4" frame into the concrete floor with
expansion bolts. In order to provide a more uniform exit
velocity profile, the exit tongue was modified as suggested
by Bradshaw (5). Power is controlled by an on-off switch
mounted on the laboratory wall and is of the 220 volt, three-
phase type. An emerger.cy power cuc-o~f switch pushbutton
.., I
![Page 18: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/18.jpg)
Blower
Inlet Filter
Motor
Diffuser
,. ~Screens ~ ..
I ' I I I
I I '
Air I Flow•• I I I
I I I
' I '
PLAN VIEW Settling Chamber
t I I I I I I I I I I
I I I
I I I
I I I
, I I
I I I
' I I
ELEVATION 'JIEW
Nozzle
Figure l. Tunnel Schematic
Cascade co
,, " ... .........
![Page 19: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/19.jpg)
9
.OP
Figure 2. Inlet Filtration System
![Page 20: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/20.jpg)
10
is mounted on the cas~ade test section. The blower is also
equipped with an adjustable pulley drive system which allows
for speed variation.
Between the blower and the tunnel, a vibration ab-
sorption and sealing junction was fabricated. This junction
makes use of an inflated 27" x l.\" bicycle innertube infla-
ted to six P. S. I. :-1ounted on the blower case is a moni-
toring pressure gage and a Schraeder valve for refilling the
tube. These are shown in Fig. 3. A straightening screen is
also mounted at this junction.
As with most tunnels of this type, the flow frcm the
blower is slowed in a diffuser. This diffuser nominally
has an eight-degree half angle and expands for a distance of
six feet to an outlet area of sixteen square feet. To
reduce the possibilities of separation in the diffuser,
another screen was placed midway down the length at the
flange located at that point.
The wind tunnel walls were fabricated from three-
quarter inch A-C plywood, as were the flanges. The tunnel
sections are supported by two-inch steel angle legs with
adjustment bolts as feet to keep the tunnel centerline
horizontal as shown in Fig. 1. The exterior of the tunnel
is painted Nhite. During fabrication, the smooth side of
the plywood was assembled facing inwards and initially
prepared with a coat 0f sander sealer and wood putty. Final
![Page 21: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/21.jpg)
Figure 3.
11
Inflated Tunnel Blower Junction with Pressure Gage
![Page 22: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/22.jpg)
12
preparation included several coats of glossy polyurethane.
From the diffuser, the flow passes through a settling
chamber. The 12,000 CFM flow passes through this chamber at
nominally only sixteen feet per second, so that straight-
ening and smoothing can be done without incurring a large
pressure drop. Provision is made for the mounting of a
five-eights inch cell honeycomb for turbulence reduction.
In the series of tests described, the tunnel was operated
without the turbulence reduction honeyco~b, but with flow
straightening screens installed at twelve, sixteen, and
twenty inches downstream of the entrance. The full length
of the settling chamber is twenty-four inches.
Following the settling chamber is the first nozzle
section. In a distance of two feet, this section reduces
the cross section from four feet square to three feet
square. At three feet square, the dimensions became small
enough for the second nozzle section to be fabricated of
formed sheet metal.
The second nozzle section matches the slope of the
first at the flange connection. Furthermore, to optimize
the flow and reduce the chance of boundary-layer separation,
the first section was brought into the shop and the junction
dimensions were mated by hand, which provided a smooth wall
shape. The profile of the second section is essentially a
combination of two arcs matched together with the radius of
![Page 23: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/23.jpg)
13
the second arc being :he starter as proposed by M. ~.
Mikhail(8). At the outlet, flow area is maintained constant
for two inches to insure the ccrrect exit slope of zero
degrees.
In order to reduce the effect of boundary-layer corner
flows, the cross-section down the tunnel is maintained
square and in the second nozzle section the corners were
rounded with body putty.
B. Cascade Test Section
The cascade test section has a one foot square hori-
zontal inlet with up to eighteen test blades which turn the
flow upwards. To provide both variable stagger and in-
cidence angle the blade row is mounted between two rotatable
discs of plexiglas each four feet in diameter and ~ inch
thick. This is shown in Fig. 4.
!he discs are mounted against three-quarter inch thick
wooden backboards. These are five feet square and have four
foot diameter ports as shown to allow viewing of the blades
for possible laser doppler anemometry tests. These back-
boards are then held rigid by a perimeter frame. At the
outer edge of this frame are three-quarter inch panels
connecti~g both backboards and providing overall cascade
strength. L~rgc openings are provided for the exit flow and
at the inle: to allow for access to the adj~stable top
![Page 24: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/24.jpg)
14
Figure 4. Blade Row Angle Guide, Probe Traverse and Stiffening Box
![Page 25: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/25.jpg)
15
and bottom flow panels as well as che inlet itself.
As shown, the cascade is s 1.ipported at each corner by
one and one-half by three inch steel channel legs. These
are fourty-four inches long and each has an adjustment bolt
as a foot. The inlet design allows this adjustment to be
made independently of any tunnel adjustments and helps to
maintain a smooth transition throughout the inlet region.
This smooth transition proved to be one of the more
difficult jobs of design and construction but was finally
executed with the help of a high-speed router. The router
allowed very clean, straight, and perpendicular cuts.
~1ounted on an arm, this router also allowed the blade mount-
ing discs as well as the disc carriers to be machined to an
accurate radius.
The two blade discs provide a smooth inlet transition
and allow the angle of the blade row to be variable. Since
the tolerances are close around the circumference and the
radius is large, the contact friction is also large. In
order to avoid turning the discs independently and thus
possibly breaking the blades, torsional rigidity between the
two discs is required. This rigidity was provided in the
form of a 16" x 11~" x 12" plexiglas box which is fastened
to both discs and is located below the blades and away from
the tes~ section range. On the front disc is a pointer
![Page 26: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/26.jpg)
i6
which references to marks on the cascade body and identifies
the blade row angle. This angle-oeasurement device, as well
as the box for torsional rigidity, can be seen in Fig. 4.
Since the inlet must remain horizontal and still allow
the blade row angle to be variable, the top and bottom walls
of the inlet are moveable. These walls are completely
removable half-inch plexiglas panels with adjustable feet
mounted at four points. These feet are on tall blocks and
push outward, thus bracing the panels in any position be-
tween the two discs. Since this is a compressor cascade the
inlet section operates below atmospheric pressure; there-
fore, all possible leaks must be sealed (here with duct
tape) in order to avoid feeding boundary layers and dis-
turbing the flow.
Xore of these panels were fabricated to be used as
tailboards downstream of the blades. These were judged to
be unnecessary for the present tests and were not used.
C. Blade Cascade
Eighteen blades constitute the blade cascade. With a
solidity of one, this number of blades allows a blade row
angle as high as seventy-five degrees. The blade row angle
is defined as the angle of attack plus the stagger angle.
For corrnnonly-used stagger angles, very high angles of attack
can ~e reached.
![Page 27: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/27.jpg)
17
At these high angles the t~in blades are heavily loaded
and support becomes a problem. Fer this reason, each blade
and its base were cast as a unit. A description of the
blade-forming process is given in the Appendix. Since the
base can only give support at one end, pins are fitted to
the opposite end of the blade. When assembled, these pins
protrude into shallow holes in the near plexiglas disk and
while giving support, still allow the stagger angle to be
changed. An aluminum bolt is epoxied to the blade base and
is bolted to the far disc. A steel rack mounted in that
disc assures accurate alignment.
The blade cross-section employed is shown in Fig. 5.
This blade was chosen because it represents the midspan
profile of the first stage of a compressor on a common
aviation gas turbine.· The engine referred to is the General
Electric T64-CE-6B. All cascade blades are cast without
twist and have a chord length of 2.58". Since endwall
effects are increased at high angles of attack the blades
were developed with the high aspect ratio of 4.65, i.e. the
blade length is 12.00". The stagger angle selected for the
initial tests is 36.s~.
D. Measuring Equipment
The primary flow measurement device in the cascade is a
pitot-static probe model P~A-12, built by United Sensors &
![Page 28: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/28.jpg)
i.8
-co 0
DIA= 0.055"
0.20"
I •
/~ II
p::;
Figure 5. Blade Profile
![Page 29: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/29.jpg)
19
Control Corp., Massachusetts. This probe is mounted o~ a
traverse 2 mm downst=eam of the b:ades and measures the
total and static pressures across the blade wake and between
the blades. The probe was aligned with the help of a rod
mounted telltale which limited the accuracy to one degree
of yaw angle; however, with this system errors associated
with steep velocity gradients were eliminated. Unfortunate-
ly even this system has low accuracy in a fully stalled
blade wake region but this was felt to be insignificant due
to the low relative mass flow. The traverse is readable to
1/5 of a degree in angular measurements and to 1/4 of a
millimeter in rectilinear measurements.
Static taps were cast into the pressure and suction
sides of opposing blades at the center of the cascade.
Twelve taps on each blade were located at midspan. This
positioning was chosen so as to avoid non-two-dimensional
flows and is shown in Fig. 6.
A multitube inclined manometer is used to display the
pressures from the static blade taps as well as the pitot
static probe. It has thirty-one channels, uses red gage oil
(S.G. = 0.826), and is inclined by 60°. Readings can be
ascertained down to 0.20 mm of water.
A Thermo-System Inc. Hotwire Anemometr:r System was used
to determine the freestream turbulence levels at the in~et
of the cascade.
![Page 30: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/30.jpg)
Figure 6.
20
The Blade Cascade Showing Close-Up of Instrumented Blade Passageway
![Page 31: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/31.jpg)
X RESULTS
The circular arc compressor blade cascade described was
tested at a single stagger angle of 36.5°, and the angle of
attack was varied across twelve angles ranging from 0° up to
35°. The chosen angles of attack were biased towards higher
values for purposes of more fully mapping that region.
The measured mean deflection angle versus angle of
attack is shown in Fig. 7. This was calculated from mea-
surements taken downstream across the blade span. The
relationship uses continuity and conservation of momentum in
the blade row direction. The definition is from G. C. Oates
and describes mean turning angle as follows (17):
s f JUVdS
-1 0 = ~l - tan - s 2 OU ds 0
The mean total head loss calculated versus angle of
attack is shown in Fig. 8. This was calculated from
measurements do·..vnstream across the blade span. The mean
total head loss relation used was as defined in NASA special
publication SP-36 as follows (16):
21
![Page 32: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/32.jpg)
22
25 en w w a:: C) w 20 0 ...
C0 w _J C) z ~
z 0 t-(.) 10 w _J u_ L&J 0 z <t 5 IJ.J ~
0 ... ----------------------------------------. -5 0 5 10 15 20 25 30 35
ANGLE OF ATTACK o;', DEGREES
Figure 7. ~ean Deflection Angle
![Page 33: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/33.jpg)
23
1.0
0.8 13
(('} 0.6 (('} 0 _J
0 <t L&.J 0!4 ~
_J
~ 0 .... z 0.2 <t l.IJ ~
oL----=::::;:;:~;::::::~---.---.,.......--....--..., -5 0 5 10 15 20 25 30 35
ANGLE OF ATTACK ~ , DEGREES
Figure 8. ~ean Total Head Loss
![Page 34: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/34.jpg)
24
1 s Pa1 - Pa2 ~ = - f ds
s o ~ o v2 1
Downstream stagnation pressure loss and deflection
angle were found by the use of a pitot-static probe mounted
on a traverse. The resulting data are plotted with respect
to fraction of chord parallel to the cascade blade row.
Again, in order to observe trends, this information is
plotted for all tested angles of attack, and may be found in
Figs. 9 through 20.
The loss curves show the blade wake levels as cal-
culated versus position down the blade row. Except for the
zero-angle-of-attack situation (negative incidence angle)
the suction side of the blade produces the larger wake.
This can be seen by the fact that the plotted wake loss
maxima are nearly vertical down the left side (pressure
side) and slope more gently down the right side (suction
side). This is true even into the highly-stalled blade
range.
For all unstalled angles of attack, part of the flow
downstream of the blades retains the upstream total pressure
value. At twenty-five degrees where the fully stalled
condition originates, none of the measured downstream total
pressure values are as high as the upstream values. As the
![Page 35: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/35.jpg)
35
30 CJ)
~ 25 a: <!>
~ 20 <..U z 0 .... (.) w _J u.. w 0
15
10
5
0
~ 0.8 g w ~ 0.6 CJ) w a: Cl. 0.4 z 0
~ 2 0.2 <!> ~ Cf)
0
Figare 9.
.,_
25
I
--- Ji - - ~ -0 - - ~ - - ~ r-o
1'7
r R
f I I !
j I \ \ )
J - ~I - - -1.0 1.5 2.0
FRACTION OF CHORD PARALLEL TO CASCADE Stagnation Press~re Loss and ueflection A..~gle for :ii = 0°
![Page 36: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/36.jpg)
35
30 (/)
t!:j 25 a:: (!)
~ 20 .. <-0 z 0 1-u w _J Lr-w 0
15
10
5
0
~ 0.8 g w ~ 0.6 ~ 0::: a.. 0.4 2 0 ~ z 0.2 (!)
~ CJ)
0
~~
26
I
,. _fA. ~ "-' ~ "" ~ -
'7
) )
..,
\ ~
ol
I'"'\ ,... ,... r ,... ,... .'°\ ,... ,... ,... \_ I I
1.0 1.5 2.0 FRACTION OF CHORD PARALLEL TO CASCADE
Figure 10. Stagnation Pressure Loss and Deflection Angle for :ti = 8°
![Page 37: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/37.jpg)
35
3 (/)
~ 2 a:: <!)
5
?. 7
~ 2 0 ~'fl p-o-.___ -.. <.0 z 0 .... u w _J LL.. w 0
5
0
5
0
0
~ o. g 8
w ~ 0. (/) w a::
6
a.. 0 .4 z 0
~ z 0. <!)
~ CJ)
2
0
~,,
Figure 11.
~I ~ l----0- v -"' I'f'·
~~ P--
I~
T\ \ R
\ -J ~ ,.. ('\ ,..., I"' "' ~,..., ,..J, '"'
1.0 1.5 2.0 FRACTION OF CHORD PARALLEL TO CASCADE Stagnation Pressure Loss and Deflection Angle for :tl = 14°
![Page 38: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/38.jpg)
35
3 (/)
~ 2 er: (!)
5
~ 2 0 ... <.u z 0 ..... (.) w _J LL. w 0
5
0
5
0
~ 0. g w
0
8
~ 0. 6 (/) UJ
~ o. 4 z 0 ~ z 0. (!)
~ en
2
0
~ ,.. ~
' ~
28
A~ J \ "\,j - ~ -....._ r ~
IS
("I
\ (" l1
I I
I I c \
j \ "" "' "' j ~ ("I
I
1.0 1.5 2.0 FRACTION OF CHORD PARALLEL TO CASCADE
Figure 12. Stagnation PYessure Loss and Deflection Angle for :ti = 18°
![Page 39: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/39.jpg)
35
3 (/'J
t±1 2 0: (!) w 0 2 ..
tU z 0 ..... (.) w ...J LL. w 0
5
0
5
0
5
0
~~I. 0
~ 0. g w
8
~ 0. 6 (/'J IJJ 0: Q.. o. z 0
~ z 0. (!)
~ en
4
2
0
~-
I<
Figure 13.
v
29
I I I T
I ~r~
,_..__,~.:r
~ °'\ I -v-- b-' ~
\0 'CU
~
\ I \
~ ~
I ' \ I \ I~ I
I ~ \ (' l ~ - ~ ~
1.0 1.5 2.0 FRACTION OF CHORD PARALLEL TO CASCADE Stagnation Pressure Loss and Deflection Angle for .ti = 19°
![Page 40: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/40.jpg)
35
3 Cf)
~ 2 a: C!> w 0 2 ..
<.U z 0 ..... (.) w ....J LL w 0
5
0
5
0
5
0
0
~ 0 g .8
w ~ 0. Cf) w
6
ff 0 .4 z 0 ~ z 0. C!> ~ CJ)
2
0
30
~~ v· v~ ~ 0 -~ ~ f-o
0 r---._"' J")
~--~ '7
f\ ~
\ \ \,J
\ \
\ \ I ' \ 0 l ~ \
" ' ' " '
1.0 1.5 2.0 FRACTION OF CHORD PARALLEL TO CASCADE
Figure 14. Stagnation Pressure Loss and Deflection Angle for :ii = 20°
![Page 41: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/41.jpg)
35
3
.. t<l z Q .... (.) w _J LL w 0
5
0
5
0
5
0
~ 0. g w
0
8
~ 0. 6 ~ b: 0 z 0
.4
~ z 0. <!>
2 ~ CJ)
0
31
(~ ~
/ '"'""- Q..
~~ ~ ,..__
o/. ~
-~ I~
r\ T\
\ \ \ (l
\ \ \
J ' """ ~ J ~~ n ~
1.0 1.5 2.0 FRACTION OF CHORD PARALLEL TO CASCADE
Figure 15. Stagnation Pressure Loss and Deflection Angle for ~i = 21°
![Page 42: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/42.jpg)
(/)
35
30
ttJ 25 a: (.!)
~ 20 .. tU z 0 .... u w ....J LL. w 0
15
10
5
0 C\J 0"" a._ > ~ ~ 1.0 o _,C\I a._
~ 0.8 g w ~ 0.6 (/) w ~ z 0 ~ 2 (.!)
~ Cf)
0.4
0.2
0
~
~u
~.,,~
r\ \
J p'n; 1.0
32
I .--.<"\
~ jJ r-
~ " _[] /
147
r. \
\ \ \ \ \
\
\ j ~.~ 1.5 2.0
FRACTION OF CHORD PARALLEL TO CASCADE Figure 16. Stagnation Pressure Loss and Deflection Angle
for =ii = 22°
![Page 43: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/43.jpg)
(/) LL.I LL.I a: (!) LL.I c .. ~ z 0 ~ u LL.I ...J LL LL.I c
35
25
20
15
10
5
0
1.0
~ 0 g .8
LL.I
~ 0 .6 ffl ~ 0 .4 2 0 ~ z 0 (!)
~ (/)
.2
0
Figure 17.
33
,G ' -I \ I ¢ \ "' ('
~ I ~ I ~-
~ J "'-0
- 0 'Q ~
r\ \ \ b
\
\ \
\ \
J \ I '~
7 \ J 1.0 1.5 2.0
FRACTION OF CHORD PARALLEL TO CASCADE Stagnation Pressure Loss and Deflection Angle for :ii = 25°
![Page 44: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/44.jpg)
Cf)
35
30
~ 25 a: (!)
~ 20 .. t0 15 z
0 ..... (.) w _J LL. w 0
10
5
0 C\J 0 ..
9- ~ l.O o _,C\I Cl.
~ 0.8 g w ~ 0.6 (/) Cf) w ~ z 0
0.4
~ 2 0.2 (!)
~ Cf)
0
34
r ' '"" I o'~
R ~ I c
\ ~ '\ l
/ I
'b---0" 'c-~
r ~ ~~ \ ~ Y7
\ \ \
\ J \ ~ \ \
~ / '\J' )
1.0 1.5 2.0 FRACTION OF CHORD PARALLEL TO CASCADE
Figure 18. Stagnation Pressure Loss and Deflection Angle for 1i = 28°
![Page 45: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/45.jpg)
.. <.U z 0 ~ (.) lJJ _J ~ lJJ c
3
3
5
5 v 0
5
0
5
0 "' -
.0
~ 0 g .8
LLJ a:: 0 ~
.6 CJ) LU
~ 0 .4 z 0
ti z 0 (!)
~ Cf)
.2
0
Figure 19.
\
35
/°\, / ~
\ I \ v\\ v \
1 ,,
\ I \ ~ \_ Id
?"- ~
I \ ifl \
\ I \ I \ ' (
\
\ \ ( .
\ \ r ) b
\J r'-.J
1.0 1.5 2.0 FRACTION OF CHORD PARALLEL TO CASCADE Stagnation Pressure Loss and Deflection Angle for :ii = 30°
![Page 46: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/46.jpg)
.. tU z Q ~ (.) LLJ ...J LL LLJ 0
35
3
5
0
5
0
5
I
0 ~-
0
~ 0 g .8
LLJ ~ 0. (/) LU
6
~ 0 .4 2 0 ~ z 0. (!)
t! Cf)
2
0
't;". ?Q ~ i.gure - .
I\
36
/"R 1:> .......--0,
v \ 1~/ v- \ 0
p c
\ / \ I \ I \ \ v 'o_
I-'
1'7
\ r ~ \ I \ I
\ 1/ \ ~
1D
\ \ )
\ J v \ J 1.0 1.5 2.0
FRACTION OF CHORD PARALLEL TO CASCADE Stagnation Pressure Less and Deflection Anele for :ii = 35°
![Page 47: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/47.jpg)
37
angle of attack is increased further, the regions of high
loss increase in extent, but nee in level.
From zero degrees up to sixteen degrees angle of at-
tack, the curve reveals a mini~UI!l in losses. The actual
minima appears to be at about eight degrees, which cor-
resonds to zero incidence angle. At eighteen degrees angle
of attack the mean total head loss starts to increase rap-
idly. This region corresponds to the first maxima on the
deflection curve, Fig. 7. Following this region the de-
flection curve falls off, and simultaneously the slope of
the loss curve reduces. In the high-angle-of-attack region,
where the mean deflection curve begins to rise again, the
mean total head loss curve again further reduces in slope.
For purposes of showing trends, the calculated velocity
distribu~ion on the blades is shown for all tested angles of
attack. (See Figs. 21 through 32.) On the ordinate of each
graph is the fraction cf inlet velocity related to pressure
by the relation:
Fraction cf Inlet Velocity = , ~ v2 ~ - 1
In each case, the suction side blade results are the upper
curve and the pressure side blade results are the lower
curve.
A point cf interest is at an angle of attack equal to
![Page 48: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/48.jpg)
38
1.6
1.4 SUCTION SIDE
>-.... CJ 1.2 0 ...J w \ > \ .... \ UJ ...J /
~ 1.0 LL. PRESSURE SIDE 0 z 0 ~ 0.8 u <l a; LL.
0.6 I
04 .... ________________________________ ___
0 0.2 0.4 0.6 0.8 1.0 FRACTION OF CHORD
Figure 21. Velocity Distribution for :..{ = o~
![Page 49: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/49.jpg)
39
1.6
1.4
> SUCTION SIDE f--u 1.2 0 ..J IJJ > f-IJJ ..J ~ 1.0 lJ.. ' ' 0 -z 0 ~ 0.8 u PRESSURE SIDE ~ a::: lJ..
0.6
04 ...... -----..-------~--------------------o 0.2 0.4 0.6 0.8 1.0 FRACTION OF CHORD
Figure 22. Velocity Distribution for ~i = 8°
![Page 50: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/50.jpg)
.:.o
1.6
04LU-----~.------..-------,-------,.------r 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
Figure 23. Velocity Distribution for ~i = 14°
![Page 51: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/51.jpg)
~l
1.6
1.4
>- SUCTION SIDE .... u 1.2 0 _, w > .... w _, ~ 1.0 LL ..... 0 .... ., z 0 .-= 0.8 (.) ~ a::
PRESSURE SIDE LL.
0.6
04w-----------------------------------o 0.2 0.4 0.6 0.8 1.0 FRACTION OF CHORD
Figure 24. Velocity Distribution for :xi = 18 °
![Page 52: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/52.jpg)
42
1.6
1.4
> SUCTION SIDE .... u 1.2 0 ...J LU > .... LU ...J ~ 1.0
' ' LL ,,.. 0 z 0 ~ 0.8 u PRESSURE SIDE ct a:: LL
0.6
0.4 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
F'igure 25. Velocity Distribution for ' 19° ::il =
![Page 53: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/53.jpg)
43
1.6
1.4
> .... u 1.2 SUCTION SIDE 0 ...J UJ > .... UJ ...J z 1.0
..... LL. ' /
0 z 0 t:= 0.8 u
PRESSURE SIDE ct a: LL.
0.6
04u...----------........ ------.,..------.------.., 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
Figure 26. Velocity Distribution for ~{ = 20°
![Page 54: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/54.jpg)
44
1.6
1.4
> t-u 1.2
SUCTION SIDE 0 ...J UJ > t-UJ ...J ~ 1.0 ..... .....
/ u.. 0 z 0 ~ 0.8 PRESSURE SIDE (.) ex a:: LI..
0.6
0.4 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
Figure 27. Velocity Distribution for I
::il = 21°
![Page 55: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/55.jpg)
1.6
1.4
>- SUCTION SIDE I-u 1.2 0 ...J l&J > I-LLJ ...J ... ~ 1.0 ' .,, LL 0 z 0 t= 0.8 PRESSURE SIDE u ct a:: LL
0.6
0.4 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
Figure 28. Velocity Distribution for I
:t 1 = 22°
![Page 56: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/56.jpg)
46
1.6
1.4
> ..... -u 1.2 SUCTION SIDE 0 ...J IJ.J > ..... .....
' IJ.J / ...J ~ 1.0 u.. 0 z PRESSURE SIDE 0 ~ 0.8 u <t a:: u..
0.6
0.4 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
Figure 29. Velocity Distribution for ' :i = 1 25°
![Page 57: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/57.jpg)
47
1.6
1.4
> ._ u 1.2 0 SUCTION SIDE _J UJ >
' ._ ' UJ \ _J /
~ 1.0 ~ LL. 0 z PRESSURE SIDE 0 j:: 0.8 (.) <t a:: . LL.
0.6
0.4 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
Figure 30. Velocity Distribution for ' = 28° :, 1
![Page 58: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/58.jpg)
48
1.6
1.4
>-._ u 1.2 0 SUCTION SIDE ..J UJ > ._ ' ' LLI / ..J 2 1.0 u.. 0 z PRESSURE SIDE 0 t== 0.8 u ct a: u..
0.6
04 ..... __________ ......,. ______ ,...... ____ _,_ ____ --t 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
Figure 31. Velocity Distribution for ~i = 30°
![Page 59: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/59.jpg)
49
1.6
1.4
> ... u 1.2 0 SUCTION SIDE ..J LIJ > ...
' LIJ ' ..J ,, z 1.0 LL. 0 z PRESSURE SIDE 0 ~ 0.8 u <t a: LL.
0.6
OAw....-----------.------..-----....... ----__, 0 0.2 0.4 0.6 0.8 1.0
FRACTION OF CHORD
Figure 32. Velocity Distribution for ~i = 35°
![Page 60: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/60.jpg)
50
zero degrees. Here the in~idence ~ngle equals minus eight
degrees, and for the first thirty-five per cent of the chord
there is a lower pressure (higher velocity) on the pressure
side than on the sucticn side. This is due to the negative
angle of incidence.
At an angle of attack of eight degrees, (zero-degree
incidence) the suction side of the blade shows higher
velocity values across the entire chord. This behavior is
maintained for all subsequent values of angle of attack.
As can be seen on the figures, the velocity distri-
bution shows the expected characteristics of airfoils for
angles of attack up to twenty-two degrees. At twenty-five
degrees the suction side of the blade curve becomes level
revealing a fully-stalled flow. This fully-stalled curve
shape is maintained to the highest values of the tested
angles of attack.
![Page 61: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/61.jpg)
XI DISCUSSION
It is shown in Fig. 7 that as angle of attack in-
creases, the turning angle cf the flow also climbs. This
process deteriorates as the blades approach the stalled
condition. As this occurs the mean-turning-angle curve
actually becomes negative in slope. This slope is continued
until about twenty-five degrees angle of attack. At this
point the blades are fully stalled and the shape of the
suction side is thought to be inconsequential with respect
to the flow; however, the flow is forced to pass through the
blade passages and is guided primarily by the pressure side.
For this reason, even though the total losses still climb
with increasing incidence angle, the turning angle is forced
again into a rising condition with rising incidence angle.
The extremes of this phenomenon are uncertain due to
reduced flow in each passageway at the higher angles of
attack. Furthermore, the mean total head loss at thirty-
fi7e degrees angle of attack, is already over eighty-five
per cent and therefore cannot possibly maintain such a high
s}.ope :or much longer; much less to the theoretical limit of
ninety degrees. For this reason very little extrapolation
from the measured data may be done. Reducing the s~agger
angle would allow the necessary high angles of attack but
t~at requires more tests and will be addressed in the
reco~errd2tion sectic~ to follow.
51
![Page 62: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/62.jpg)
The two-dicensicnality of :he flow in the blade passage-
way was of interest. This is a complex question to address,
but some observations have been viewed in the test section
which allow some generalizations. Fig. 33 is a view looking
downwards into the blade row at fourteen degrees angle of
attack. The flow is from bottom to top and strings are
mounted across the leading edge of one blade at 0.5", 1.0",
1.5", 5.0", 7.0", 10.5", 11.0", and 11.5" across the span.
It was observed that under these conditions the strings
(telltales) on the pressure side of the blade showed very
little deflection in the spanwise direction, the exception
being at the extremes where the flow apparently moves toward
the edges feeding the boundary layer below. Due to the gap
on the right hand side the telltale has deflected to the
extent of being blown around the edge and down the suction
side. The gaps are 0.12" on average.·
Cn the suction side of the blade, greater deflection of
the strings is observed. Unlike the pressure side, however,
the direction is toward midspan. In general the greatest
deviation angle is at the outer positions, but at one inch
from the blade ends the deviation is only on the order cf
five degrees. For both the suction and the pressure side at
the five and seven inch locations, no measurable deflection
was noted.
![Page 63: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/63.jpg)
Figure 33.
53
Flow Visualization Spanwise Across the Blade at ai = 14°
![Page 64: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/64.jpg)
54
Figure 34 shows a similar view in the fully stalled·
range. In this photo, t:he angle of attack is twenty-eight
degrees, but similar flow patte=ns csn be observed. On the
pressure side the flow direction is in the downstream di-
rection with deviation only being at the extremes. Again,
deflection is away from midspan. Furthermore, on the right
hand side the telltale is again blown around the edge. For
the most part, the midspan flow appears two-dimensional,
being off-axis by only about five degrees on the left side
and at two inches from the right side showing no deflection
at all.
The suction side of the blade has similar flow patterns
with respect to the unstalled condition but the turbulence
in the stall-generated wake makes this phenomenon difficult
to photograph. With these observations, the error due to
departure from two-dimensional flow at the midspan position
may be considered minimal.
Another question arises concerning the smooth drop in
turning angle at angle of attack of 22° and the smooth
return again to a positive slope at 35° angle of attack.
The ~ain reason for the curve shape however has much to do
with a phenomenon known in compressor testing as "rotating
st:all".
In axial flow compressors rotating stall occurs in the
transi~ion region bet~een normal operation and the fully-
![Page 65: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/65.jpg)
SS
t I
I . .,.,.
- ·.• \ ., f J , I
·- ---· · · c4 f ;
,1111---~-~~R=~- : : . .. I
I • • ;
~ I I
Figure 34.
_-""lulW~---- - .... -
- ~ t
Flow Visualization Spanwise Across the Blade at al = 28°
![Page 66: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/66.jpg)
stalled condition.
stalled passageway
56
This phenomenon consists of a single
interacting w~th the flow such that the
fol lowing passageway stalls ar.d i.mloads the first. This
phenomenon may continue, grow, and extend over several blade
passages. In full rotacing stall, a stall cell propagation
speed will be en the order of half of the engine speed and
can be maintained continuously.
In a linear compressor cascade this phenomenon can also
occur and is called "propagating stall". Since the stall
cell propagates along the blade row it may eventually pass
the instrumented passageway and thus alter the reading.
Fortunately, the question of this effect has been addressed
on similar blades (circular arc) and found that natural
propagating stall initially appears at about twenty-two
degrees. In the present tests, the propagation was found to
~ove at around 0.8 times the flow speed but to travel only
two or three passageways before dying out. For this reason,
tests on propagating stall in cascades usually have an
initiatir.g device for stall cell generation.
As the angle of attack is increased towards the fully-
stalled condition, the strength of the stall cells increases
and propagation will continue along the full blade row. Un-
like the compressor case, the stall cell stops due to the
end of the blade row, this being a discontinuity. As found
in the cascade in question and in the referenced tests, the
![Page 67: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/67.jpg)
57
blades are fully stalled at approxirrately twenty-five
degrees angle of attack.
Based on the above, i~ the :ested cascade at twenty-two
degrees angle of attack propagating stall is likely.
However, due to the velocity of the flow, the time of pro-
pagation past the instrumented passageway is less than a
~illisecond. Considering this and the frequency response cf
the multitube manometer, it is evident that the readings are
time averaged and as the angle of attack goes up from 22°,
the percentage of time spent in a stalled condition in-
creases. This explains the gradual drop on the turning-
angle curve which then smoothly begins to increase at an
angle of attack of 25° where the blades initially become
fully stalled.
Finally, in the transition region, observations were
made which showed endwall stall at the base end of the
blades. This phenomenon did not occur at the pinned ends of
t~e blades due to the gap. The effect of this single end-
wall stall is therefore reduced, and causes less effect with
~espect to stream tube contraction and a related lessening
of total pressure losses. As with propagating stall, this
effect is observed only in the transition region between
normal operation and full stall.
![Page 68: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/68.jpg)
XII CONCLCSIONS
The results have s~o·..m good repeatability with earlier
data (3). Variations exist in the values of mean total head
loss. These variations are minimal and correspond to in-
provements in an endwall contraction effect. It is con-
cluded that these modifications did, in fact, improve the
measurements.
Available data in the high-angle-of-attack regions is
rare. From the experimental results obtained it was found
that as the angle of attack is increased the turning angle
of the flow is increased. For double circular arc com-
pressor blades this situation changes at 22° angle of attack
where the turning angle starts to go down. As the angle of
attack is increased above 25° the blades become fully
stalled and the turning angle begins to climb again.
7he mean total head losses are minimal in the unstalled
region with an actual mimimum corresponding to a zero-
degree incidence angle. As the angle of attack increases
the slope of the mean total head less curve increases. This
=ise in slope reaches a maximum at 22° angle of attack but
t~e values of mean total head loss increase up to the the
tested limit of 35° angle of attack.
58
![Page 69: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/69.jpg)
XIII RECO~NDATICNS
The results of the experi~ents demonstrate that much
progress has been made in the conscruction of the dedicated,
high-angle-of-attack, cascade faci~ity. Prior research has
alleviated some of the inherent oroblems related to high
angle of attack flows. Nevertheless, many refinements can
still be made and include:
(i) manufacturing two test blades of metal to reduce
blade flutter,
(ii) Applying boundary layer suction just upstream of
the tested blade row,
(iii)increasing the gap at the base of the blades to
reduce the endwall stall phenomenon,
(iv) installing honeycomb in the settling chamber and
controlling turbulence levels with screens placed
in the nozzle.
Besides these modifications the following tests are
recommended.
(i) An investigation into the effects of natural
propagating stall in the transition region between
normal operation and fully-stalled operation.
(ii) A study at low stagger angles probing into the
ultra-high-angle-of-attack region including the
59
![Page 70: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/70.jpg)
60
tunnel designed limit of 75'.
(iii)An investigation with Laser Doppler Anemometry
equipment so as to avoid probe interactions with
the flow.
(iv) An investigation into downstream struts to find
steady upstream interactions and perforra flow
visualization tests.
(v) A correlation of results of tests using various
blade cross-sections and stagger angles.
![Page 71: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/71.jpg)
XIV REFERENCES
1. Sexton, M. R., and W. F. O'Brien, "A Model for Dynamic Loss Response in Axial-Flow Compressors," ASME Paper, 81-GT-154, 1981.
2. Ikui, T., M. Incue, and N. Kuromaru, "Researchers on the Two-Dimensional Retarded Cascade; Part 3, Cascade Performance at High Inlet Angles," Bulletin of the J.S.M.E., Japan, August, 1970.
3. Keller, U. J., "Performance of a Double Arc Blade at Low Reynolds Number," Masters Thesis at VPI&SU, Virginia, October, 1978.
4. Mehta, R. D., "The Aerodynamic Design of Blower Tunnels with Wide-Angle Diffusers," J. Aerospace Sci., Vol. 18, pp. 59-120, Pergamon Press, Great Britian, 1977.
5. Mehta, R. D., and P. Bradshaw, "Design Rules for Small Low Speed Wind Tunnels." Aeronautical Journal, pp. 443-449, November, 1979.
6. East, L. F., "Spatial Variations of the Boundary Layer in a Large Low-Speed Wind Tunnel," Aeronautical Jour-nal, p. 43, January, 1972.
7. Jackson, N. A., "Theoretical Velocity Distributions Downstream of Non-Uniform Single and ~1ultiple Smoothing Screens," Aeronautical Journal, p. 251, April 1972.
8. Mikhail, M. N., "Optimum Design of Wind Tunnel Con-~ractions," AIAA Journal, Article No. 78-819, Vol. 17, No. 5, :fay 19 79.
9. Bonciani, L., P. L. Ferrara, and A. ':'imori, "Aero-Induced Vibrations in Centrifugal Compressors," Cen-trifugal Compressor Design Department, Nuovo Pignone, Florence, Itanly, ~ASA Conference Publication 2133, Xay 12-14, 1980.
10. Spa!:'ks, J. F., "A Modeling Technique for Subsonic Stall Flutte'!:' in Cascades," Ph. D. Dissertation at VPI&SU, February, 1981.
11. Horlock, J. H. , "Axial Flow Compressors-Fluid ~!echanics and T~ermodynamL::s," Robert E. Krieger Publishing Company, ~~e~.; York, 197 3.
61
![Page 72: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/72.jpg)
62
12. Sovran, G., "Propc.gating Stall in Airfoil Cascade," (Fil~) FM-7, General Motors Research Center.
13. Cheers, F., and L. E. ~!aca:-tney, "Propagating Stall in a Water Cascade Tunnel," Internal Laboratorv {-!emo-randum, National Research Council, Division of ~!echanical Engineering, Ottawa, Canada, 1959.
14. Mathiolakis, D., "Steady and Unsteady Cascade Measure-ments," Masters Th es is in SS~I at VPI&SU, 1982.
15. Kettle, D. J., "The Design of Static and Picot Static Tubes for Subsonic Speeds," Journal of the Royal Aero-nautical Society, p. 835, December, 1954.
16. Lieblein, S., "Experimental Flow in Two-Dimensional Cascades," NASA Special Publication-36, Aerodynamic Design of Axial-Flow Compressors, NASA Scientific and Technical Information Division, Washington, D.C., 1965.
17. Oates, G. C., "Cascade Flows," Chapter 12, The Aero-namics of Aircraft Gas Turbine En ines, AFA?L-
0 io, Jan.
18. Kline, S. J., and F. A. :.rcclintock, "Describing C'n-certainties in Single-Sample Experiments," ~Iech. Eng. , p. 3, January 1953.
19. Dzung, !... S. ED., "Flow Research on Blanding," Elsevier Publishing Company, New York, 1970.
![Page 73: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/73.jpg)
XV APPENDICES
63
![Page 74: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/74.jpg)
A. UNCERTAINTY ANALYTIS
The uncertainty in calculating the mean flow-turning
angle is found with the methods of Kline and McClintock
(18). The angle notation is as defined in Fig. Al. The
uncertainty W- is generated from the relation
W- = E
N
i=l
2 k 2
h w 3N N.
l. • 1 1
where Ni are the various primary measurements that make up
s. The relation for~ produces the averaged turning angle
based on mass-averaged momentum downstream of the blade row
s (17) .::uvds tan -1 0
~ = :.tl - s 2 .::u ds 0
A finite-difference method of integration was used for
data reduction and the relation developed was simplified to
s ' ' ~ Chz ho2) sin (2 :.t2) - ,, - ~
'-
-1 0 s :.tl - tan s
~ Chz - ho2) sin (:; 2) 0
where h is the level on the measurement manometer and ~2 is
the data step size. The above equation may be simplified by
assuming :(. = 14 ° . .l..
Unfortunately the partial derivatives to
64
![Page 75: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/75.jpg)
65
s
I Gl
I
I
Figure Al. Cascade Notation
![Page 76: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/76.jpg)
66
be used in the generation of the uncertainty are still a
problem. For this reason the uncertainty for the two sum-
mations will be found first and used in a simplified rela-
tion for uncertainty in mean turning angle. The uncer-
tainties of the various parameters may be found in Table Al.
Wsumtop = ( { ih;n Whz} 2 + { ;h:z Wh02) 2
( ' )2 (' )2)~ + :~um W.J.2 + ?~1:1m Wc:.1 J ..l 2 J..:. ,_
In order to proceed we need the various partial derivative
values.
-:.sum = ~
sin ,:_i,tot
= --2- 5.23 mm )sum = jfio2
~sum = 22.54 mm ' ' ' ,) ...... (.,
Wsum top = ((5.25 x 0.2) 2 + (5.23 x 0.2) 2
2 19102 rr.m
+ (19102 x 0.017) 2 + (22.54 x 0.25) 2 )~
Wsum top = 324.8 ? mm-
![Page 77: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/77.jpg)
67
The lower sunnnation was done i:-i a similar fashion with
the resultant,
) Wsum bottom= 313.2 l.!ffi~.
For simplification, the uncertainty of the ratio of
these two summations was then calculated.
Wratio =({'ratio Wsum top)2 +(~ratio Wsum bot}2)~ )sum top l~ sUrll bot
The partial derivatives are then needed.
:!ratio ~sum top = 1
sum bot = 4.90 x 10- 5 -2 mm
:!ratio ~sum bot
= -sum top (sum bot) 2
= 4.90 x 10- 5 -2 mm
The uncertainty of the ratio is then
Wratio = ((4.90 x 10- 5 x 324.8) 2 + (4.02 x 10- 6
x 313.2) 2 )~ = 0.016.
We r.ow have enough information to proceed directly to
the final value of turning-angle uncertainty.
= ~ 1 - tan-l (ratio)
W-
![Page 78: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/78.jpg)
68
1
= -1 = -.993 ~ratio 1 + ratio-
W~ = ((1 x 1°) 2 + (-.993 x 0.016) 2 )~ = + 1.0°.
It may be seen that the accuracy is then limited by the yaw-
probe capabilities and further reductions in uncertainty of
mean turning angle would have to come from a higher pre-
cision yaw-angle measurements.
The uncertainty of mean total head loss is also cal-
culated using the methods of Kline and ~cClintock. As
before, the relation of mean total head loss to its primary
measurements is needed. This is found in ~ASA SP-36 (16).
s JU(POl - P02 )ds
1 0 = 2 s _lz.::Vl ::uds 0
This relation was simplified and then integrated using
finite-difference methods in a manner similar to the mean-
turning-angle calculation. The resulting relation is as
follows.
![Page 79: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/79.jpg)
69
s - (h2 1 0
,I = (hl ho1) - s ~ (h') ... 0
All ter:ns have been previously been defined and as with
the mean turning angle, the uncertainty of the two sum-
mations will be calculated first. The development will be
done in the same order as before.
(( 3sum ) 2
w bottom = 3(h2 ho2) W(h - ho2) + sum - 2
(;~~ woj2 r 3sum = ~ I-# A, = 2.093 ;; Cn 2 - ho2) 2(h2 - h ) ,,.2
02
;sum 16.08
Wsum bottom= ((2.093 x 0.4) 2 + (16.08 x 0.25) 2 )~ = 4.1
The upper summation was done in a similar fashion with
the resultant,
w = 431.6 sum top
![Page 80: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/80.jpg)
70
We now have enough information to proceed directly to
the final value of mean-total-head-loss uncertainty.
1 surr, ton ~ = (h1 - n01 ) sum bot
w:::-=((;(h1 '~ n01J W(hl -hail) 2+(;s~:;top wsum top)
+ ( ; 5~ \ot W sum bot ) 2 ) ~
~sum top
1 = sum top = -8.806 x 10- 5 sum bot
= 1 = -1.523 x 10- 5 (h1 - h01 ) sum bot
--1 -:;.,--.-- = -l sum top = -1.981 x 10-4 3sum bot 2 Ch1 - h01 ) sum bot
10 -5 x ? -5 2 w: = ((8.806 x 0.4)~ + (1.523 x 10 x 431.6)
w:- = + 0.0065
Although the uncertainty band associated with this is
greater than the minimum value of :, most measurement vaiues
exhibit fairly good accuracy. ~uch of this is due to the
dimensionless character of : in which factors such as
![Page 81: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/81.jpg)
..., 1 '~
density and temperature are cance:led out and not included
in the analysis. Nevertheless, L. S. Dzung (19) noted a
strong relationship between turbulence level and mean total
head loss. For this reason, we may conclude that the
results are accurate but should not correlate exactly with
similar tests run with different levels of upstream tur-
bulence.
![Page 82: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/82.jpg)
72
TABLE Al. INSTRUXENT READABILITY AND ACCURACY
Multitube Manometer+ 0.2 mm H20
Rectilinear Traverse Scale+ 0.25 mm
Traverse Yaw Scale + 1/5 Degrees
Probe Yaw Capability± 1.0 Degrees
Blade Row Angle± 0.5 Degrees
Stagger Angle± 0.5 Degrees
Tunnel Air Temperature + l.0°F
Relative Humidity ± Si~
Atmospheric Pressure± 0.1 mm Hg
Turbulence Level = 1.6%
Velocity Variation at Inlet = 1.5%
![Page 83: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/83.jpg)
B. INLET VELOCITY ~1EASTJREMENTS
The following are the results of a velocity profile
measurement at the inlet, four inc~es from the nozzle.
Twenty-five measurements were made with a pitot-static
probe. This did not include measurements on or near the
walls but were spaced evenly two-inches apart. Table A2
lists the velocity values as a function of position. Values
of i are at 2", 4", 6", 8", and 10" from the north wall and
values of j correspond to 2", 4", 6", 8", and 10" from the
top of the section. In an exaggerated format, Fig. Al
reveals the variation in velocity across the inlet.
73
![Page 84: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/84.jpg)
74
TABLE A2. NOZZLE VELOCITY DISTRIBCTION (m/s)
~ 1 2 3 4 5
1 65. 65 65.54 65.75 65.75 66.05
2 65.33 65.43 65.33 65.33 65.54
3 65.22 65.22 65.22 65.22 65.43
4 65.86 65.86 65.86 65.86 65.33
5 66.07 65.75 65.22 65.22 65.22
![Page 85: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/85.jpg)
75
i 1 2 3 4 5
j =
1
2
3
4
5
Figure A2. Nozzle Velocity Variation Above 65 m/s
Vertical Height/0.5'' + 65 = Velocity (m/s)
![Page 86: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/86.jpg)
C. BLADE PRODCCTION
The information on the shape of the blade and the
forming processes used was generated by U. Keller. This
included measurements on an actual blade which was cut at
midspan and also experiments on blade forming techniques.
As with U. Keller's experiments, these tests used blades
molded of fiberglas.
In the final form, the blade manufacturing process was
similar to the Keller method. His notes are so extensive
that this subject will be addressed with details concerning
only the differences.
The basic mold is made of aluminum and has three main
parts for generating the pressure side, the suction side,
and the base of the blade. When manufacturing a blade, the
surfaces which are to come in contact with the resin must be
prepared with a mold release. After much testing it was
found that this coating must be verv thin, or large scale
imperfections appear on the blade surfaces.
The next step is the cutting of the fiberglas clot~
such that three sheets are obtained which will lay in the
suction side of the mold. Due to the size of the mold and
to allow for overhang, these sheets were approximately 2~ x
12 inches. Small two by three-quarter inch strips were
placed at the base end such that fiberglas in the final
blade will add to the strength of the blade-to-base
76
![Page 87: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/87.jpg)
77
connection. When the meld is assembled and sealed with
putty, resin is poured into the mold forming the blade and
the base simultaneously. Curing takes about a day.
When cured, the blade was forced out of the mold.
Care must be taken to avoid damaging the blade. The next
step in blade preparation was trimming the blade to twelve
inches in length. This was followed by drilling a 0.095
inch diameter hole, 0.25 inches in depth, into the center of
the trirruned end. A 0.090 inch diameter dowel pin was then
epoxied into this hole. When trimmed, this dowel fitted
into the cascade wall and supported the tip end of the
blade. At the base of the blade an aluminum mount was
attached with epoxy. This mount was then bolted through the
cascade wall and was the support which maintained the blade
angle.
![Page 88: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/88.jpg)
The vita has been removed from the scanned document
![Page 89: COMPRESSOR BU~DES At~GI · "~1umbers in parenthesis rerer to references. 1 . VIII LITERATURE RE'!IEW Cascade Wind Tunnels Due to the complexity of the flow through an axial flow ccmpressor,](https://reader033.vdocuments.net/reader033/viewer/2022041906/5e636b732b9d2563f46cade4/html5/thumbnails/89.jpg)
CASCADE PERFOR!'-!.ANCE OF DOUBLE CIRCUL\R AH.C
COMPRESSOR BL\DES AT HIGH ANGLES OF ATTACK
by
Peter T. Tkacik
(ABSTRACT)
The design of a cascade wind tunnel for testing of com-
pressor blades at high angle of attack is described. ~eth
ods to insure uniform velocity profiles and control of inlet
turbulence are discussed. The problem of maintaining two-
dimensional flows at high angle of attack was addressed.
A tunnel capable of testing cascades of compressor
blades at angles of attack up to seventy-five degrees was
constructed. Performance of the tunnel was evaluated and
data were acquired for flow over double-circular-arc blades
with angles of attack extending into the fully-stalled
region. Comparisons were made with available data in the
installed flow regime. Results showed that the tunnel had
adequately uniform inlet velocities and low turbulence
levels, and that two-dimensional flow was maintained over
the center two-thirds of the high-aspect ratio ~lades.