404 J. AIRCRAFT VOL. 20, NO.

18:1 Pressure Ratio Axial/Centrifugal

Compressor Demonstration Program
Jeffrey K. Schweitzer*
Pratt& Whitney Aircraftt West Palm Beach, Florida
and
John W. Fairbanks!
U.S. Department of Energy, Washington, D.C.

The results of a component technology demonstration program to design and test an advanced axial/cen-
trifugal compressor for industrial gas turbine applications are presented. Objectives were to demonstrate 18:1
Downloaded by PENNSYLVANIA STATE UNIVERSITY on March 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.44885

pressure ratio at 90% polytropic efficiency with 80% fewer parts as compared to current industrial gas turbine
compressors. The compressor design approach utilizes low-aspect-ratio/highly loaded axial compressor blading
combined with a centrifugal backend stage to achieve the 18:1 design pressure ratio on a single spool in only
seven stages. Demonstrated design point performance was 91.5% polytropic efficiency at 14% stall margin and
70 Ibm/s flow. This represents the highest known demonstrated performance in this pressure ratio and flow
class. The results are particularly significant in that they were accomplished at axial compressor aerodynamic
loading levels approximately 15% above current production engine design practice.

Nomenclature high-temperature (2600-3000°F), high-performance, nominal

Df = diffusion factor 100 MW size gas turbines combined with efficient waste heat
h — enthalpy, Btu/lbm boilers and steam turbines. The primary fuel for the central
N — rotor speed, rpm power system is coal gasification. Dispersed power system
P = static pressure, psia support for the 5-50 MW range gas turbine, either simple or
PT = total pressure, psia combined cycle, includes the development of advanced gas
PR = pressure ratio turbine component technology and multifuel capability.
T = total temperature, °R In January 1976, a contract was awarded to Pratt &
w = relative velocity, ft/s Whitney Aircraft Government Products Division
W = flow rate, Ibm/s (P&WA/GPD) to investigate the utilization of advanced
7 = ratio of specific heats aircraft gas turbine technology to improve the performance
6 =PT!/\4.696 and durability of future industrial gas turbines in either the
7] = efficiency central or dispersed power system applications. Additional
e =r 7 /518.688- emphasis was placed on significantly reducing the number of
gas turbine parts to improve reliability, lower initial cost, and
Subscripts thus enhance commercialization whether through retrofit of
1 — inlet plenum instrumentation station existing systems or through incorporation in a new centerline
2 = axial compressor discharge instrumentation station engine.
3 = pipe diffuser exit plane instrumentation station In phase I of the program, 1 the latest aircraft gas turbine
ad — adiabatic technology was applied in the definition of a conceptual
bl = interstage bleed flow advanced gas turbine in the 15,000 shp/10 MW class (Fig. 1).
cor = corrected to standard day inlet conditions The design approach included designing to higher cycle
net = corrected for interstage bleed pressure ratio (18:1) and combustor exit temperature (2500°F)
p = polytropic for increased cycle efficiency and specific power. Predicted
simple-cycle performance potential of the conceptual engine
Introduction was 13,856 shp at 36.3% thermal efficiency. Performance
potential in combined cycle operation approaches 50%
A CONCERTED national effort is currently in progress to
conserve energy and increase utilization of our abundant
domestic energy resources. As a part of this effort the U.S.
thermal efficiency depending on the type of waste heat
recovery system utilized.
Department of Energy's (DOE) Office of Coal Utilization has The compressor design for this conceptual engine in-
assigned high priority to supporting the development of corporated highly loaded low-aspect-ratio axial compressor
advanced industrial/utility gas turbine engines for operation blading combined with a centrifugal back end stage to attain a
in combined cycle systems on coal-derived fuels. The strategy polytropic efficiency goal of 90% with minimum number of
for large central power systems is to assist the development of stages. The reduced number of stages in combination with the
inherent stiffness of the single-piece axial compressor drum
rotor configuration additionally enabled consideration of a
single-shaft gas generator without requiring a third bearing
Presented as Paper 81-1479 at the AIAA/SAE/ASME 17th Joint compartment in the hot section area. It is of interest to note
Propulsion Conference, Colorado Springs, Colo., July 27-29, 1981; that while the compressor design techniques employed are at
submitted Aug. 12, 1981; revision received Sept. 20, 1982. This paper the forefront of current compressor technology, the concept
is declared a work of the U.S. Government and therefore is in the
public domain.
of axial stages followed by a centrifugal stage was included in
*Design Technology Specialist, Goverment Products Division. Sir Frank Whittle's original aircraft gas turbine engine patent
Member AIAA. disclosure in 1930.
tProgram Manager, Heat Engines Branch, Office of Coal The present paper describes the results of the compressor
Utilization. component demonstration testing conducted under phase II
 MAY 1983 AXIAL/CENTRIFUGAL COMPRESSOR 405

- 4 0 4 8 12 16 20 24 28 32 36 40 44 48
Axial Length - in.
Fig. 2 Axial/centrifugal compressor flowpath.

0.50
Fig. 1 Advanced technology 10 MW industrial gas turbine.

DOE Axial Compressor

of the above program. A full-scale compressor test rig was 0.46 Design Point
Development
designed, fabricated, and tested in the P&WA/GPD com- d Test Rigs
pressor test facility located in West Palm Beach, Fla. Per-
Downloaded by PENNSYLVANIA STATE UNIVERSITY on March 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.44885

formance characteristics of both the axial and centrifugal S

| 0.42
compressor stages as well as that of the overall compressor
were documented over the entire compressor operating range.
The results of this compressor development have subsequently © 6
been included in the coordinated efforts of the Department of I 0.38 o
Current Engine
Energy and the Electric Power Research Institute (EPRI) in I Compressors
the reliable advanced liquid fuel engine preliminary design
studies.2 0.34 — O O
Aerodynamic Design
The inherent benefits of low-aspect-ratio blading have been
investigated by P&WA and others in recent years, beginning 0.26 0.30 0.34 0.38 0.42 0.46 0.50
with single-stage rig test programs3'4 and proceeding through Average Static Pressure Rise Coefficient - AP/(PT-P)
full-scale multistage rig development programs. The test Fig. 3 Axial compressor loading comparison.
results have consistently demonstrated the capability of low-
aspect-ratio blading to achieve increased efficiency and
stability at high axial compressor loading levels. The low- Table 1 Axial compressor airfoil series selection
aspect-ratio/high loading combination results in fewer Blade row No. airfoils Airfoil seriesa
required stages and fewer airfoils per stage for a given overall
compressor pressure ratio. The larger geometry low-aspect- IGV 13 63
ratio blades can also provide improved durability and in- Rl 18 MCA
creased performance retention in extended operation. SI 20 CA
Significant advances have also been made in the past decade R2 21 MCA
S2 22 CA
in centrifugal compressor technology such that performance R3 25 CA
approaching that of axial compressors can be obtained S3 24 65/CA
especially in low-flow, low-pressure-ratio applications. This R4 27 CA
makes the centrifugal compressor stage attractive for use as a S4 26 65/CA
back end stage, where a single rugged certifugal stage can R5 27 65/CA
replace a number of small axial stages. The utilization of the S5 26 65/CA
centrifugal compressor stage thus can result in a further R6 25 65/CA
reduction in the number of required stages where frontal area S6 20 400
is not a critical constraint (such as in stationary gas turbines). a
MCA = multiple circular arc. CA = circular arc. 65/CA = 65 series thickness
The combination of the centrifugal stage with low-aspect- distribution superimposed on a circular arc meanline.
ratio axial compressor aerodynamics in the current con-
figuration allows the achievement of 18:1 pressure ratio on a
single spool with high efficiency and 80% fewer parts as
compared to current industrial gas turbine compressors. loading limit was therefore set (Fig. 3) which, although
The selection of the final compressor flowpath geometry providing an approximate 15% increase in loading levels over,
(Fig. 2) was the result of a comprehensive parametric and current production engine compressors, represented a logical
optimization study. 1 The study results indicated that the peak next step advancement in the state-of-the-art relative to
compressor performance potential could be attained with a previous multistage compressor development test rig
six-stage axial/single-stage centrifugal compressor con- programs. The 0.4 inlet hub/tip ratio, bulged i.d. flowpath
figuration with a 6.67/2.7 pressure ratio split. An inlet air- evolved as the flowpath which provided the best ef-
flow rate of 70 Ibm/s was selected to be compatible with an ficiency/loading trade while maintaining a low-loss transition
advanced industrial gas turbine application in the 15,000 to the centrifugal stage.
shp/10 MW class. Rotor speed was set at 15,000 rpm based on The airfoil series selection for each of the axial compressor
turbine stress limitation considerations. blade rows, along with the number of airfoils in each row, are
given in Table 1. The selection of airfoil blading geometry for
Axial Compressor each blade row was made on the basis of optimizing airfoil
The objective of the axial compressor aerodynamic design performance for the particular Mach number regime for that
was to exploit to the maximum extent possible, without row. Multiple circular arc airfoils were used for the first-stage
jeopardizing near-term applicability, the potential of low- rotor, which had an inlet tip relative Mach number of 1.20 at
aspect-ratio blading to achieve high aerodynamic loading the design 1376 ft/s tip speed, and the second-stage rotor,
levels with high efficiency and good stability. An aggressive which had an inlet tip relative Mach number of 1.05. Airfoil
 406 J.K. SCHWEITZER AND J.W. FAIRBANKS J. AIRCRAFT

Active Axial
Drum Rotor
Split Stator Clearance
Flapped IGV Case _^ Control-

Variable
Cantilevered
Stators
All Stages-
Fig. 4 Axial/centrifugal compressor rig cross section.

Fig. 5 Compressor rotor assembly.

aspect ratios varied from 1.35 for the first rotor to 0.8 for the
sixth-stage stator.
Both the centrifugal impeller and axial compressor drum
Centrifugal Stage rotor were machined from solid Inconel 718 forgings.
Downloaded by PENNSYLVANIA STATE UNIVERSITY on March 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.44885

The centrifugal stage design incorporates a 25 deg Titanium 6A1-4V was selected for the axial compressor rotor
backswept impeller and a 2:1 area ratio conical pipe diffuser blades.
designed to produce a stage pressure ratio of 2.7:1 (total-to- The rotor assembly is supported by a roller bearing on the
total). The design point inlet corrected speed and flow are front end, while a dual-bearing configuration is used in the
11,182 rpm and 14.1 Ibm/s, respectively, which results in a rear in which a roller bearing carries all the radial load and a
specific speed of 64. ball bearing carries all the thrust load. The ball bearing
The centrifugal impeller has 1.8 full blades with one set of retainer can be moved axially to adjust the position of the
splitter blades. The inducer region of the impeller was compressor rotor relative to the cases and provide active
modified slightly to accommodate the leading-edge geometry control of the impeller-to-shroud tip clearance. All three
of a 400-series axial compressor airfoil and incidence was set bearings are under-race cooled with holes provided through
to obtain minimum loss based on axial compressor cascade the inner race to direct oil to the cage riding surfaces and
data. The axially defined leading-edge geometry was then rolling elements. Hydraulically damped bearing hairsprings
smoothly transitioned into the conventional impeller blading. are utilized for both roller bearings with spring rates selected
The mean impeller diffusion ratio (w max /w min ) of 1.57 is to achieve a 20% rotor critical speed margin.
within previously demonstrated P&WA design experience.
Impeller discharge Mach number is 0.88. Stationary Cases
The impeller discharge flow is further diffused to a Mach All of the front bearing compartment and slip ring services,
number of 0.83 in a short vaneless/semivaneless space region including instrumentation routing, are provided through 13
prior to entering the diffuser passages. Each of the 26 conical inlet struts. Flapped inlet guide vanes are mounted im-
diffuser passages consists of a 3 deg cone angle segment to a mediately adjacent to each strut and are hydraulically driven
1.2:1 area ratio followed by a 5 deg cone angle segment to 2:1 through a sync ring to provide the desired rig inlet air angle.
area ratio (which results in a design exit Mach number of The six stages of variable cantilevered stators are mounted
0.35). A constant diameter throat length of 0.100 in. was in two horizontally split case halves. To minimize endwall
included to reduce the sensitivity of the throat area to leakage losses the leading edge of each stator, where the
manufacturing deviations and leading-edge wear. maximum pressure differentials occur, are entirely supported
by the stator support button with the airfoil trailing edge
Test Rig cantilevered off the back of the button. A nominal 0.016 in.
Based on the compressor aerodynamic design defined for thick abradable coating of ekonol-aluminum was applied to
the conceptual engine, a compressor test rig (Fig. 4) was the stator airfoil tips to work in conjunction with the drum
configured for use in the component validation testing phase rotor abrasive coating to provide rub protection. A silicone
of the program. Where practical from time and budget rubber rub strip material was included in the case o.d.
standpoints, materials and fabrication techniques used for the flowpath wall over each rotor blade tip region for rotor rub
major rig hardware items were selected to be compatible with protection.
those that would be used in an engine application. Design life The impeller shroud is bolted to its outer support case at the
goal was 20,000 h/10,000 cycles. impeller exit with the inlet end of the shroud allowed free
axial movement. Piston rings are used to seal the shroud inlet.
Rotor Assembly The shroud has a 0.003-0.005 in. thick silver coating to
minimize the effects of a minor rub should one occur. An
The low-aspect-ratio axial compressor blades are mounted interstage compressor bleed capability of up to 30% of the
in a single-piece drum rotor. Blades for stages 2 through 6 are inlet flow rate is provided just forward of the impeller shroud
loaded circumferentially into annular grooves in the disk for starting and off-design compressor matching.
through loading slots provided on diametrically opposite sides The compressor rig hardware was designed to be com-
of the rotor. Blade locks on the side of each loading slot patible with two diffuser configurations: 1) a vaneless dif-
secure the final blade inserted. The first-stage blade at- fuser for use during the axial compressor documentation
tachment is a conventional axial lug configuration in order to portion of the test program, and 2) a conical pipe diffuser for
obtain sufficient lug shear area at the relatively small first- use during the overall compressor performance documen-
stage bore diameter. Rubber seals applied to the under side of tation testing. Capability was provided to change diffuser
the blade platforms are utilized to minimize leakage into the configurations on the test stand without dismounting the rig.
attachment cavities for all stages. An abrasive coating con- Flow from the diffuser is discharged into an annular collector
sisting of a mixture of 97% aluminum oxide and 3% titanium which has six exhaust flanges connected to the facility
dioxide was applied to the drum rotor flowpath between discharge ductwork.
blades to provide stator tip rub protection.
The centrifugal impeller is attached to the axial compressor Instrumentation
drum rotor by means of a Curvic ® coupling and 12 tiebolts. A The compressor aerodynamic instrumentation was selected
photograph of the complete rotor assembly is shown in Fig. 5. to accomplish three primary objectives: 1) define the overall
 MAY 1983 AXIAL/CENTRIFUGAL COMPRESSOR 407

The facility atmospheric inlet system includes a 132 ft .long,

60 in. diam duct which is connected to a 14 ft long, 144 in.
diam plenum chamber. The flow rate is measured by means of
a standard ASME thin-plate orifice which is located in the 60
in. diam duct section. Both 15 and 24 in. diam orifices were
used during the test program depending on the anticipated
flow range of the particular test sequence scheduled.
The compressor test rig was installed in the facility by
means of a free-standing mount aligned to the rotating
centerline of the drive turbine. The compressor rotor was
coupled to the drive turbine rotor utilizing a tight-fitting
splined drive coupling. The compressor discharge air was
ducted from the rig discharge collector through six 4 in. lines.
Four of these lines were manifolded together into an 8 in. duct
containing both an 8 in. main discharge control valve and a 2
in. vernier control valve, which were regulated as required
during the test program to vary the compressor back pressure.
The remaining two lines each contained a manually operated
Downloaded by PENNSYLVANIA STATE UNIVERSITY on March 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.44885

Fig. 6 Compressor test rig.

gate valve and a surge pop-bleed relief valve.

Run Program
compressor performance, 2) allow definition of the axial The compressor test program was conducted in three
compressor and centrifugal stage component performance, phases. The initial phase of the compressor test program was
and 3) provide interstage measurements for diagnostic structured to verify satisfactory operation of the compressor
purposes to aid in performance optimization. Overall per- rig and all of its auxiliary support systems. During this test
formance instrumentation consisted of measurements of total sequence, vibration and strain gage data were monitored and
pressure and temperature at the compressor inlet (station 1.0) recorded continuously.
and discharge (station 3.0). Compressor inlet conditions were In the second test phase, axial compressor steady-state data
measured in the test facility inlet plenum chamber utilizing were acquired along a series of speedlines from wide-open
three Rosemount precision resistance temperature sensors and discharge to peak efficiency at 50, 70, 80, 90, 95, and 100%
four standard total pressure sensors. Compressor exit con- design speed. In order to obtain an early estimate of the axial
ditions were measured at the centrifugal stage pipe diffuser compressor stability, additional data were obtained at 50 and
exit plane. Four total pressure rakes and four total tem- 70% speed above the nominal operating line to approximately
perature rakes were utilized, each containing five individual 20% stall margin.
Kiel-head sensors. Each rake was located at the exit of a The final phase of the test program was conducted to
different diffuser pipe with the probe sensors placed so that document the overall compressor performance over the
the superimposed array for each sensor type would cover 17 complete operating range. For this phase, the vaneless dif-
equal flow area segments with the center sensors on each rake fuser, which had been installed for the prior two phases, was
providing redundant measurements of the center area. replaced with a pipe diffuser. The test sequence was
Measurements of total pressure and temperature were also specifically conducted to obtain the following data:
provided at the axial compressor discharge (station 2.0) just 1) Starting and low speed aerodynamics.
upstream of the interstage bleed to allow calculation of the 2) Interstage bleed requirement definition.
axial and centrifugal compressor component performance. 3) Aerodynamic design point performance.
This instrumentation consisted of five circumferential wake 4) Axial, centrifugal, and overall compressor performance
rakes distributed radially across the flowpath. Each rake had (pressure ratio, speed flow, and efficiency characteristics).
eight pressure and eight temperature sensors alternated across 5) Surge line definition.
the probe head to define the pressure and temperature gapwise
distributions. The pressure sensors were impact tubes while
the temperature sensors were Kiel-head stagnation tubes. Performance Calculations
Structural instrumentation was provided primarily to Overall compressor performance was defined from the
monitor and evaluate the vibratory characteristics of the axial compressor rig inlet plenum (station 1) to the pipe diffuser
compressor airfoils. A total of 60 strain gages were utilized exit plane (station 3). The redundant low-velocity inlet
for this purpose, including 34 dynamic gages located on the measurements were arithmetically averaged, while the pipe
blades and 26 dynamic gages located on the vanes. diffuser exit temperature and pressure arrays were mass
A photograph of the fully assembled and instrumented averaged to obtain representative mean values for use in the
compressor rig ready for shipment to the test stand is shown performance calculations. All calculations were performed on
in Fig. 6. a total-to-total basis. Overall adiabatic efficiency was
calculated based upon the overall pressure ratio and dry air
Test Program values of enthalpy
Test Facility
Demonstration testing of the axial/centrifugal compressor A/T =f(PR,jT,)
rig was conducted in the full-scale compressor test stand (1)
AA f(T,T3)
located in P&WA/GPD's Turbojet Engine Altitude Test
Facility. The drive system for this facility consists of a single-
stage, direct-drive TF30-P414A high-turbine module which A net adiabatic efficiency was also defined for the part-
has been adapted to run on high-pressure steam. The stand speed overall compressor performance which accounts for the
has the capability of providing 26,000 hp at 16,200 rpm over a work done on the interstage bleed flow. The net efficiency is
wide range of pressurized and heated inlet conditions. defined as
Operation of the current program, however, required a
throttled inlet (approximately 10 psia) due to the
unavailability of one of the three facility boilers at the time 1) net ~ (2)
the test program was being conducted.
 408 J.K. SCHWEITZER AND J.W. FAIRBANKS J. AIRCRAFT

Polytropic efficiency was calculated from the adiabatic acquired with the variable axial compressor vane rows in their
efficiency and pressure ratio assuming a 7 of 1.4 nominal positions per the design vane angle schedule. Part-
speed interstage bleed flow rates were selected to insure
adequate flow range capability at each of the speeds
(3) documented. No attempt was made to optimize or minimize
the bleed flows. The approximate percent bleed at peak ef-
A constant-flow stall margin definition was used as ficiency for each speed line is indicated in Fig. 7. Since the
follows: overall performance data were acquired at slightly reduced
inlet pressure, the value of the test Reynolds number index
PR sur e — PR (5/01-24) is also included on the figure for each speed line.
* °P xlOO (4) Part-speed efficiency values were also calculated with a
correction to account for the work done on the interstage
Performance for the axial and centrifugal stage com- bleed flow. The resultant "net" overall performance map,
ponents was determined in a manner similar to the overall which is useful for simplified cycle analyses which cannot
performance except for using the total pressures and tem- separately book-keep the bleed flow, is presented in Fig. 8. It
peratures measured at the axial compressor discharge plane should be iterated that no attempt was made to minimize the
(station 2). part-speed bleed flow rates, hence improvement in the part-
speed "net" performance relative to that shown is feasible.
Downloaded by PENNSYLVANIA STATE UNIVERSITY on March 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.44885

Uncertainty Estimates
Data acquired from the compressor rig were examined for
consistency and validity and a detailed uncertainty analysis
performed. The uncertainty analysis included an evaluation &*
a *&
i
U-fi

- ^-1'
c|8-.

of potential sources of error due to the measurement sensors,

the data acquisition and recording equipment, calibration rI* * i
data curve fits, data reouction procedures, sampling rates, ->j <o
Adiabatic Efficiency

Reynolds
and run-to-run measurement variations. The predicted un- , 1 x
% N/v^6 % Bleed Index
JR?
- % (Gross)

certainties in the compressor rig inlet and discharge Surg 3 Line!

en
0

50 28 0.90
measurements used in the determination of the overall =W ^c
10

70 24 0.85
80 23 0.78
compressor performance are shown in Table 2. The resultant s^
-±
000

90 13 0.71
estimated uncertainty of the overall compressor efficiency is ^\^ ^> — Peak
—k

100 0 0.64
Efficiency
1.31%.
Results ^
0 0 t\3 0 )

Vo

^0% -
Overall Compressor Performance —
Pressure Ratio

I ~*>~-~3t
An overall compressor performance map is presented in |^500/0
Fig. 7. The peak design speed overall performance measured ) 10 20 30 40 50 60 70 80 90 100 110
.U

was 87.8% adiabatic efficiency at 18.45:1 pressure ratio and % Design Inlet Corrected Flow
^0

approximately 10% stall margin. Performance for the data Fig. 7 Overall compressor performance map.
point closest to the 18:1 pressure ratio design point was 87.6%
adiabatic efficiency and 14% stall margin. The demonstrated
oo <o
0 0 0 0

performance for these two points is compared to the contract *

Efficiency - %
Net Adiabatic

1 G
- ——L
goals in Table 3. Performance for the data point closest to the ^^ 4
^J

design pressure ratio exceeds the contract efficiency goal by ^

§0>

^
2.0% and exceeds the contract minimum stall margin goal by <*
4.0%. The highest efficiency measured at any speed
(neglecting bleed effects) was 88.76% adiabatic efficiency at 20 Reynolds
12:1 pressure ratio and 90% rotor speed. All data were % N/v/0 /o Bleeci Index Surge
50 28 0.90 Line —v ?
16 70 24 0.85 ^
o 80 23 0.78 100% N/V~0
08 90 13 0.71

^'if
Table 2 Uncertainty estimate CC12 \
0) 100 0 0.64 ^^ ^p<3ak Eff ciency
No. of Bias, Precision, Uncertainty, 58 ^ 90%

Variable sensors ± ± ± I 80%

J* ^
Plenum Pp psia 4 0.047 0.014 0.061 4
70%
Plenum T, °F 3 0.37 0.47 0.84 J®^
Discharge Pt, psia 20 3.23 0.294 3.52 50%
0
3 10 20 30 40 50 60 70 80 90 100 11
Discharge T, °F 20 7.60 0.82 8.42 % Design Net Corrected Flow
Efficiency, % — 1.13 0.18 1.31
Fig. 8 Net overall compressor performance.

Table 3 Design speed overall performance summary

Demonstrated
Closest to
design
Peak efficiency pressure ratio Program goal
Flow, Ibm/s 70.7 70.5 70
Pressure ratio 18.45 17.81 18
Adiabatic efficiency, % 87.8 87.6 85.6
Polytropic efficiency, % 91.7 91.5 90
Stall margin, % 10. 14 10
 MAY 1983 AXIAL/CENTRIFUGAL COMPRESSOR 409

Table 4 Centrifugal stage design speed performance summary

Demonstrated
Peak 10% stall Program
efficiency margin goal
Pressure ratio 2.76 2.70 2.70
Airflow, Ibm/s 14.19 15.20 14.10
Adiabatic efficiency, % 92.7 88.3 85.5
Exit Mach number 0.32 0.35 0.35
Stall margin, % 2.3 10.0 10.0

40
Table 5 Impeller/diffuser design point performance summary
) Range of Bleed
) Flow Rates as Tested Demonstrated . Program goal
30 Impeller
Pressure ratio 2.82 2.82
Downloaded by PENNSYLVANIA STATE UNIVERSITY on March 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.44885

Adiabatic efficiency, % 94.7 90.0

Diffuser
20
Total pressure loss, % 2.1 4.2

10

50 60 70 80 90 100

Fig. 9 Interstage bleed flow requirement.

Note: All Data Obtained

With -30% Interstage
Bleed Flagged Symbols - Vaneless Diffuser Data^ J
90%
Unflagged Symbols - Pipe Diffuser Data f"

80%
Surge Line

10 20 30 40 50 60 70 80 90 100 110
% Design Corrected Flow
10 15 20 25 30 35
% Design Inlet Corrected Flow
Fig. 11 Axial compressor performance map.
Fig. 10 Overall compressor low-speed performance map.
compressor performance data from both test sequences is
shown in Fig. 11. The highest axial compressor performance
The pipe diffuser throat areas could also be rematched at a was measured during the initial vaneless diffuser test
lower pressure ratio to further improve part-speed per- sequence. Peak design rotor speed performance was 89.47%
formance should that requirement exist for a particular adiabatic efficiency at 6.5:1 pressure ratio (as compared to
application, but at the expense of a probable drop in design design goals of 89.0% adiabatic efficiency at 6.67:1 pressure
speed performance. An axial/centrifugal matching analysis ratio). The design rotor speed performance acquired at the
was undertaken based on the available interstage data to same pressure ratio during the subsequent pipe diffuser
determine the minimum bleed flow requirements at any given testing was approximately 0.7% lower in efficiency. Cause of
rotor speed with the design diffuser throat area. The results of the performance deterioration can be primarily attributed to
the analysis are shown in Fig. 9, which summarizes the actual slightly increased "nominal" airfoil running tip clearances
minimum bleed flows required as compared to the range of for the later test runs due to wearing away of the abradable
bleed flows utilized for the compressor test program. rub materials during transient operating conditions.
Additional low-speed performance data were acquired A cursory analysis of the axial compressor stage charac-
from 20-50% design rotor speed (Fig. 10) to document the teristics indicates the compressor surges recorded at 50 and
compressor starting characteristics. The compressor behavior 70% rotor speeds were initiated in the axial compressor front
in this region was surprisingly stable even during limited end stages. The high-response instrumentation included in the
operation in the rotating stall region to the left of the "stall test rig was not sufficient to distinguish in which component
line" indicated on the figure. No compressor surges occurred (the axial or centrifugal stages) the design speed surge
at any time at these low speeds. Analysis of the low-speed data initiated. Stall margin for the axial compressor at design
acquired indicated no starting problems which would com- speed from the axial compressor peak efficiency point to the
plicate use of this compressor configuration in an engine overall compressor surge point was approximately 14%. The
application. axial compressor stall margins at 50 and 70% rotor speeds
were approximately 30 and 26%, respectively.
Axial Compressor Performance
Axial compressor performance data were acquired from Centrifugal Stage Performance
both the vaneless and pipe diffuser portions of the test All of the centrifugal stage performance goals were either
program. A composite performance map showing the axial met or exceeded as summarized in Table 4. An analysis of the
 410 J.K. SCHWEITZER AND J.W. FAIRBANKS J. AIRCRAFT

design rotor speed centrifugal stage performance at peak (18:1 pressure ratio and 2500°F combustor exit temperature)
efficiency was conducted to divide the measured stage per- each percent increase in compressor adiabatic efficiency can
formance into impeller and diffuser component per- be equated to an approximate 1.0% improvement in simple
formances. Impeller exit conditions were calculated from the cycle engine specific fuel consumption and an approximate
continuity and momentum equations based on the measured 1.4% increase in shaft horsepower. The results are that much
impeller exit static pressure and stage temperature rise. Design more significant when viewed in light of the fact that they
blockages were assumed. The resultant component per- were achieved at axial compressor aerodynamic loading levels
formances are summarized in Table 5 relative to the design 15% above current engine design practice.
goals. The analysis indicates that the excellent stage per- The test results conclusively verify the readiness of the
formance demonstrated is the result of both higher than advanced low-aspect-ratio axial compressor and centrifugal
predicted impeller efficiency and lower diffuser total pressure backend stage technologies for incorporation in future in-
loss. dustrial gas turbine applications. No mechanical, structural,
or low-speed starting problems were encountered which
Rig Mechanical Performance would preclude use of the compressor configuration in an
The rotor dynamics of the compressor were excellent over engine application. The technologies demonstrated are ad-
the entire operating range. All of the observed rotor responses ditionally potentially applicable to a wide range of advanced
occurred at low speeds, were well within the allowable engine applications including small gas turbines, marine
operating limits of the rig bearings, and could be ac- propulsion systems, and high-bypass-ratio turbofan engines.
Downloaded by PENNSYLVANIA STATE UNIVERSITY on March 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.44885

celerated/decelerated through without impact on the test rig

operation.
The low-aspect-ratio blading vibratory stress levels ob- Acknowledgment
served during the compressor test program were also well The vork presented in this paper was conducted for the
below allowable limits. No rotor blade row vibratory stress U.S. D partment of Energy, Office of Coal Utilization under
level greater than 60% of the allowable was recorded at any Contract DE-AC05-760R05035.
time during the test program. Vibratory stress levels for the
test vanes were an even smaller percentage of their allowables References
than those for the blades. In general, the vibratory data
Schweitzer, J.K. and Brown, B.T., "Advanced Industrial Gas
revealed good correlation between the predicted natural Turbine Technology Readiness Demonstration Program, Final
frequencies and the observed resonance conditions, par- Report, Phase I Design Study," U.S. Department of Energy,
ticularly for first bending and first torsional mode response. Washington, D.C., Rept. HCP/T 5035-0001, Oct. 1977.
2
Boenig, F.H. and Day, D.L., "Compressor Configuration and
Conclusion Design Optimization for the High Reliability Gas Turbine," United
The compressor performance demonstrated not only ex- Technologies Corp., Power Systems Division, South Windsor, Conn.
ceeded the program goals, but even more significantly Rept. GTR-2136, April 1980.
3
Cheatham, J.G., et al., "Single-Stage Experimental Evaluation of
represents the highest known single-spool compressor per- Low Aspect Ratio, Highly Loaded Blading for Compressors, Part IX
formance demonstrated to date in this pressure ratio and flow Final Report, Stage F and G, Volume I," NASA CR-134993, May
class. This performance can be directly translated into 1976.
equivalent benefits in terms of gas turbine engine per- 4
Reid, L. and Moore, R., "Experimental Study of Low Aspect
formance potential. For the conceptual engine cycle defined Ratio Compressor Blading," NASA TM 79280, March 1980.

Reminder: New Procedure for Submission of Manuscripts

Authors please note: If you wish your manuscript or preprint to be considered for publication, it must be
submitted directly to the Editor-in-Chief, not to the AIAA Editorial Department. Read the section entitled
*'Submission of Manuscripts" on the inside front cover of this issue for the correct address. You will find
other pertinent information on the inside back cover, "Information for Contributors to Journals of the
AIAA." Failure to follow this new procedure will only delay consideration of your paper.