Stator Endwall Leading-Edge Sweep and Hub Shroud Influence on Compressor Performance

Transcription

1 THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS 345 E. 47 St., New Yrk, N.Y GT-197 The Sciety shall nt be respnsible fr statements r pinins advanced in papers r in discussin at meetings f the Sciety r f its Divisins r Sectins, r printed in its publicatins. Discussin is printed nly it the paper is published in an ASME Jurnal. Papers are available frm ASME fr fifteen mnths after the meeting. Printed in USA. Cpyright 1986 by ASME Statr Endwall Leading-Edge Sweep and Hub Shrud Influence n Cmpressr Perfrmance D. L. TWEEDT T. H. OKIISHI Department f Mechanical Engineering Engineering Research Institute Iwa State University, Ames, Iwa M. D. HATHAWAY U.S. Army Research and Technlgy Labratries, Cleveland, Ohi ABSTRACT The use f statr endwall leading-edge sweep t imprve axial-flw cmpressr statr rw perfrmance was examined experimentally. The aerdynamics f three statr hub (inner diameter) cnditins, namely, a running clearance, a statinary clearance, and a shrud were head rise cefficient lss cefficient als investigated. Leading-edge sweep in the endwall regins f a statr blade can be beneficial in terms f lss reductin n the casing (uter diameter) end f a statr blade. It can als help at the hub end f a statr blade when either a statinary hub clearance r a hub shrud is used. A leading-edge sweep is detrimental (lcal lss increase) n the hub end f a statr blade when a running hub clearance is used. A running clear- verall ance is aerdynamically preferable t a statinary clear- efficiency ance. NOMENCLATURE aspect rati rati f blade span length t blade aerdynamic chrd length Q v rati f head rise frm inlet measurement statin t exit measurement statin acrss a blade rw t rtr tip speed squared rati f ttal-head lss relative t and acrss a blade element frm inlet measurement statin t exit measurement statin t dynamic head relative t blade at inlet measurement statin cmpressr drive shaft speed, rpm rati f tb 'verall t (13mechanical pressure side f blade venturi vlumetric flw rate, m 3 /s radius frm cmpressr axis, m blade angle angle between tangent t blade sectin camber line and axial directin, degrees suctin side f blade cmpressr drive shaft trque, Nm blade setting angle deviatin angle angle between blade aerdynamic chrd and axial directin, degrees ttal head rati f ttal pressure t air density, Nm/kg U rtr blade speed, m/s blade aerdynamic chrd length, distance between blade leading and V fluid velcity, m/s trailing edge circle centers, m p air density, kg/m 3 angle between circumferentiallyaveraged exit flw relative t blade 4) flw cefficient, rati f crssand tangent t blade camber line at sectin area average axial velcity trailing edge circle center, degrees t rtr tip velcity diffusin factr 1 V' r V' - r V' ut in 0,in ut 0,ut Vin ' (r in + rut) a Vin slidity, rati f blade aerdynamic chrd length t blade-t-blade distance between tw adjacent blades Presented at the Internatinal Gas Turbine Cnference and Exhibit Dusseldrf, West Germany June 8-12, 1986

2 mechanical verall ritn 30pQU v rati f crss-sectin area average ttal head at the secnd stage statr exit measurement statin t rtr blade tip speed squared crner fluid tward midspan. Earlier experience [1,21 1 suggested that this kind f gemetry mdificatin culd be beneficial in terms f reducing the amunt f lwmmentum fluid in the endwall, suctin-side crner. Statr hub shruding is cmmnly used in prductin cmpressrs. Three different statr hub cnfiguratins were examined fr merit: a running hub clearance, a statinary hub clearance, and a shruded hub. subscripts h in ut t 0 superscripts INTRODUCTION hub (inner diameter) value value at blade-rw inlet measurement statin value at blade-rw exit measurement statin tip (uter diameter) value tangential cmpnent relative t blade value Further reductin f the fluid flw viscus lsses ccurring in prductin axial-flw turbmachines cntinues t challenge designers. Even small gains in aerdynamic efficiency are vigrusly sught by manufacturers. Better management f the cmplicated viscus flws in endwall regins f the blade rws f a turbmachine is ne example f a specific perfrmance imprvement gal. This paper deals with lw-speed cmpressr research initiated t prvide a clearer understanding f the ptential fr better management f the endwall flws in an axial-flw cmpressr. The use f statr leading-edge sweep t imprve statr rw perfrmance was examined. The aerdynamic benefits f statr hub shruding were als investigated. The same rtr rws were used fr all tests. Tw kinds f statr blades were used. A baseline statr, cnventinal in gemetry, prvided baseline data against which t cmpare data fr ther statr gemetries. A mdified statr featuring frward symmetrical sweep f the leading edge frm midspan t the inner and uter annulus walls was als utilized. The primary bjective f symmetrical frward sweeping f the statr leading edge was t draw higher-mmentum fluid int the suctin side endwall crner flws and frce the lwer-mmentum AXIAL-FLOW RESEARCH COMPRESSOR A sketch f the tw-stage, axial-flw research cmpressr rig used fr all f the tests is presented in Fig. 1. Mre details abut the system may be fund in Refs. 3 and 4. Each cmpressr stage was designed t be representative f typical mdern cmpressr practice in terms f high stage reactin, axially discharging statrs and the absence f inlet guide vanes. A unifrm spanwise distributin f ttal pressure was prescribed fr the rtr exit. Duble circular arc blade sectins were cnsidered apprpriate fr the lw-speed testing invlved. Design data are summarized in Tables 1 and 2. The swept statr blades invlved symmetrical frward sweeping f each statr blade leading edge frm midspan t the inner and uter endwalls as specified by the equatin: r - r = t ( r t - r Ct rt - rh r t - rh) 2 ( Swept and baseline statr blades are cmpared in Table 3. Representative rtr and statr blade sectins are sketched in Figs. 2, 3, and 4. Statr hub clearances (running and statinary) were set at in. (0.864 mm; 1.4% span). Other blading details, including fabricatin infrmatin, may be fund in Refs. 3 and 4. As illustrated in Fig. 5, the different gemetry cmbinatins tested were: 1. baseline r swept statr, running hub clearance (first stage) 2. baseline r swept statr, statinary hub clearance (secnd stage) 3. baseline r swept statr, shruded hub (secnd stage). The cmpressr was perated at 2400 rpm fr all tests. Overall (atmsphere t venturi meter) head-rise 6m 15 HP MOTOR ( COMPRESSOR INLET TWO-STAGE COMPRESSOR 1/4 ( ^---1 THROTTLE - -/ AIR STRAIGHTENING TUBES VENTURI FLOW METER m Fig. 1. Schematic f research cmpressr. 1 Numbers in brackets designate References at end f paper. 2

3 Table 1. Summary f cmpressr design data. Number f stages Rtr speed Flw rate Pressure rati Number f blades per rw Rtr Statr Blade material Blade aerdynamic chrd Blade aspect rati Blade sectin prfile Blade stacking axis lcatin Leading and trailing edge radius t t aerdynamic chrd rati Maximum thickness t aerdynamic chrd rati Annulus flw path Hub radius Tip radius Reynlds number based n aerdynamic chrd and entering velcity at midspan Rtr Statr Flw cefficient tw 2400 rpm 5.25 lb m /s (2.38 kg/s) fiberglass with steel trunnin and spine 2.39 in. (6.07 cm) cnstant fr rtr and baseline statr 2.39 in. t 3.03 in. (6.07 t 7.70 cm) fr swept statr 1.0 cnstant fr rtr and baseline statr 1.0 t 0.79 fr swept statr duble circular arc radial line thrugh center f gravity f blade sectins fr rtr and baseline statr blades radial line thrugh blade sectin trailing edge circle centers fr swept statr blade 0.01 cnstant 0.10 t 0.06 linear variatin frm blade rt t ther end f blade span 5.60 in. (14.22 cm) cnstant 8.00 in. (20.32 cm) cnstant 2.0 X X Table 2. Blade design details. Slidity Inlet Blade Angle, Degrees Exit Blade Angle, Degrees Blade Setting Angle, Degrees Design Value f Diffusin Factr Baseline % span frm hub rtr blade midspan % span frm hub Baseline % span frm hub statr blade midspan % span frm hub Swept statr % span frm hub blade midspan % span frm hub

4 Table 3. Cmparisn f statr blade gemetries. Similarities between Baseline and Swept Statr Blades Number f blades per rw Blade surface finish Midspan chrd length Spanwise distributin f maximum thickness t chrd ratis Differences Between Baseline and Swept Statr Blades Baseline Stacking pint at center f gravity N leading-edge sweep Cnstant spanwise distributin f chrd length Swept Stacking pint at trailing edge circle center Symmetrical leading-edge frward sweep Varying spanwise distributin f chrd length a MID SE FLOW -Iww in AXIAL DISTANCE, cm Fig. 3. Representative baseline statr blade sectins AXIAL DISTANCE,cm Fig. 2. Representative cmpressr rtr blade sectins (same fr baseline and swept statr builds). 2.00^ 0.80 I;3' perfrmance data were recrded fr a range f flw rates. Detailed data were taken at a flw cefficient, (I), f nly. DATA ACQUISITION All measurements were made with cnventinal slwrespnse instruments and a cmputer cntrlled data acquisitin system as explained in Refs. 3 and 4. Ttal pressures were measured with a Kiel prbe. Static pressures were btained thrugh casing taps. Abslute flw angles were acquired with a "cbra" yaw prbe. Axial lcatins f the measurement statins used are indicated M 0.00 AXIAL DISTANCE, cm Fig. 4. Representative swept statr blade sectins. 4

5 1 1 1 in Fig. 6, and the circumferential lcatin and extent f the midspan measurement windws at each axial statin are cnveyed in Fig. 7. EXPERIMENTAL RESULTS As explained in Ref. 4, an uncertainty analysis f the experimental data was accmplished using the prcedures f Refs. 5 and 6. Uncertainty estimates fr data presented in this paper are summarized in Table 4. Overall (atmsphere t venturi meter) head-rise perfrmance curves fr tw distinct cmpressr builds are shwn in Fig. 8. The head-rise and flw cefficient data used t cnstruct these curves were btained at the venturi flw meter where reliable casing tap data culd be quickly acquired ver a large range f flw rates fr the tw builds indicated. The difference in head rise between the tw curves, while small, is cnsidered significant in that it was repeatedly btained. As als indicated in Fig. 8, the stall limit fr the swept unshruded first statr, swept shruded secnd statr cmpressr build was at a lwer flw than fr the baseline unshruded first and secnd statrs cnfiguratin. Actually, the stall limit imprvement shwn was als btained by a baseline unshruded first statr, baseline shruded secnd statr build and a swept unshruded first and secnd statrs versin. The prbable reasn fr the stall limit imprvement will be discussed later when detailed secnd stage data are presented and discussed. Overall efficiency data fr the baseline unshruded first and secnd statrs build are prvided in Fig. 9. These results were btained with shaft trque and secnd stage statr rw exit casing tap pressure measurements. Nte that the perating flw cefficient selected fr all f the detailed data runs, namely, (I) = 0.500, is well within the peak efficiency range fr this cnfiguratin. Measured ttal-head data btained at statin 2.0 (between the first stage rtr and statr) fr all baseline and swept statr builds were very similar [4]. Fr this reasn, the results fr the baseline unshruded statrs build nly are shwn in Fig. 10 t represent all statin 2 data sets. The ptential flw field influence f the first stage statr rw was nt discernible at statin 2.0 even when the swept leading edges were used. Representative rtr blade element diffusin factrs calculated frm blade rw slidity values and circumferentially-averaged velcity data are summarized in Table 5 and indicate that the rtr was mderately laded. In Fig. 10 and ther similar figures that fllw, data were FIRST ROTOR FIRST STATOR SECOND ROTOR MEASUREMENT STATION 1.0 MEASUREMENT STATION 2.0 MEASUREMENT STATION 3.0 SWEPT STATOR LEADING EDGE SWEPT STATOR LEADING EDGE MEASUREMENT STATION 4.0 BASELINE STATOR LEADING EDGE CL ARANCE SEALED FOR SHROUDED STATOR SECOND STATOR IDENTICAL CLEARANCE FOR BASELINE AND IDENTICAL. CLEARANCE SWEPT STATOR BUILDS FOR BASELINE AND SWEPT STATOR BUILDS Fig. 5. Meridinal plane view f cmpressr blading Y/S, MEASUREMENT STATION 5.0 Fig. 7. Blade cascade shwing circumferential measurement windws at midspan. PROBE MEASUREMENT STATIONS TIP TIP FIRST FIRST ROTOR STATOR SECOND ROTOR SECOND STATOR HUB Ha Fig. 6. Schematic shwing axial lcatins f prbe measurement statins relative t adjacent blade rws fr baseline statr cnfiguratin (dimensins in mm). 5

6 Table 4. Uncertainty estimates (20:1 dds). Overall head-rise cefficient ± Venturi flw cefficient ± at 0 = 0.4 ± at 0 = 0.5 ± at 0 = 0.6 Overall efficiency ±0.021 Ttal head ±5.0 N-m/kg Statr lss cefficient ±0.008 Statr deviatin angle Diffusin factr ±0.02 ±0.7 degree suppressing viscus flw regin thickening there. Fr the pressure side crner, there is a little larger regin f lwer ttal-head fluid fr the swept statr rw than fr the baseline statr rw. These trends are cnsistent with thse bserved by Sen et al. [1] in their plane cascade tests. The swept leading edges alter the static pressure field in the endwall crner regin f flw t such an extent that the abve described behavir results. On a circumferential-average basis, there is less ttalhead lss in the casing endwall regin (70 t 95% span frm hub) f the swept statr rw than fr the baseline statr rw (see Fig. 13). The deviatin angle results in Fig. 14 indicate that sweeping the leading edge als leads t generally better flw turning in the casing endwall regin. The blade-element diffusin factrs calculated frm circumferentially-averaged flw field measurements (Table 5) suggest that the swept statr 0.0 acquired ver ne statr pitch nly and were repeated peridically ver a secnd pitch distance t enhance visualizatin f the flw patterns invlved. Measured ttal-head data acquired at statin 3.0 (between the first stage statr and the secnd stage rtr) fr the baseline and swept unshruded statrs are presented in Figs. 11 and 12. The statr pressure and suctin surface sides are designated with P and S. Between statins 2.0 and 3.0, ttal-head lsses ccurred alng all pathlines, even thse assciated with the cre flw. In Ref. 7 the authrs suggest that this cre flw lss prbably ccurred mainly in the axial gap between rtr trailing and statr leading edges and is attributable t rtr wake mix ut. It appears as if the mix-ut lss rate is unifrm acrss the cre span; the trend in the spanwise distributin f the cre flw ttal head remained unchanged between the measurement statins. Near the casing (statr uter diameter), a substantially smaller regin f lwer ttal-head fluid is assciated with the suctin side crner f the swept statr rw than with that same crner f the baseline statr. The swept leading edge appears t have succeeded in drawing higher-mmentum fluid int the suctin crner and SECOND STATOR EXIT SHROUD STATIC PRESSURE/SHAFT TORQUE BASED VALUES 0-0 BASELINE UNSHROUDED STATORS FLOW COEFFICIENT Fig. 9. Overall efficiency fr the baseline unshruded statrs build rip U = VENTURI STATIC PRESSURE BASED VALUES BASELINE UNSHROUDED FIRST AND SECOND STATORS SWEPT UNSHROUDED FIRST STATOR SWEPT SHROUDED SECOND STATOR FLOW COEFFICIENT CURVE CURVE VALUE LABEL OF TOTAL HEAD Fig. 8. Overall head-rise fr tw cmpressr Fig. 10. Representative first stage rtr exit builds. ttal-head cntur map (0 = 0.500). 6

7 Table 5. Representative blade-element diffusin factr values based n experimental data. 11?TIP 9p % 10% Frm Hub Midspan Frm Hub T Rtr za 50 Baseline statr with running clearance ,0 Baseline statr with statinary clearance 0.32 Shruded baseline statr 0.32 Swept statr with running clearance Swept statr with statinary clearance CU LA E CURVE VALUE I OF TOTAL HEAD HUB MOTION Fig. 12. Swept unshruded (running clearance) statr exit ttal-head cntur map (M = 0.500) Shruded swept statr blade was less laded near the casing than was the baseline blade. Near the mving hub (right t left mtin as indicated by arrws n Figs. 11 and 12), where a 1.4% span clearance existed, the ttal-head cnturs f Figs. 11 and 12 suggest the presence f a lw-mmentum regin f flw frm the pressure crner nly. This regin is an accumulatin f lwer-mmentum fluid scraped ff frm the rtating hub and is larger fr the swept statr blade than fr the baseline statr blade. Mre fluid is scraped w BASELINE UNSHROUDED (RUNNING CLEARANCE) STATOR A SWEPT UNSHROUDED (RUNNING CLEARANCE) STATOR itt V Li p UNCERTAINTY PERCENT SPAN FROM HUB Fig. 13. Cnstant span ttal-head lss values fr baseline and swept unshruded (running clearance) statrs (M = 0.500) CURVE LABEL CURVE VALUE OF TOTAL HEAD I. MOTION Fig. 11. Baseline unshruded (running clearance) statr exit ttal-head cntur map (I = 0.500) O BASELINE UNSHROUDED (RUNNING CLEARANCE) STATOR A SWEPT UNSHROUDED (RUNNING CLEARANCE) STATOR UNCERTAINTY PERCENT SPAN FROM HUB Fig. 14. Deviatin angles fr baseline and swept unshruded (running clearance) statrs (M = 0.500). 7

8 Table 6. Cmparisn f passage-average statr lss cefficients fr the first stage statr builds. Build Statr Lss Cefficient Baseline unshruded (running 0.0 clearance) statr Swept unshruded (running clearance) statr ff the hub by the lnger chrd swept statr blade sectin near the hub. The suctin side cnturs reflect the drawing away f lw-mmentum fluid frm the suctin crner by the mving hub surface. On a circumferentialaverage basis, there is less ttal-head lss at 5% span frm the hub fr the swept statr rw than fr the baseline statr rw; the swept leading edge resulted in a small benefit there. Hwever, at 10% span frm the hub, the ppsite is true and the swept leading edge resulted in a detriment. The deviatin angles are slightly smaller in this regin f flw fr the swept statr than fr the baseline statr. The swept and baseline statr blades were abut equally laded near the running hub (Table 5). Frm 10% t just shrt f 70% span frm the hub, the ttal-head lss f the baseline statr blade was appreciably less than that fr the swept statr blade (see Fig. 13). The cntur plts (Figs. 11 and 12) indicate a smaller prtin f lwer-mmentum fluid n the suctin surface f the baseline statr blade in the 10% t 70% span regin than fr the swept statr. The swept leading edge resulted in mre lwer-mmentum, casing suctin crner fluid being frced t the midspan regin f the blade. The baseline and mdified statr wake cnturs are similar n the pressure side. The deviatin angles fr the tw kinds f blades are similar in the midspan regin althugh slightly better fr the baseline statrs. The swept and baseline statr blades were equally laded at midspan (Table 5). On a passage-average basis, the lss values in Table 6 reveal that sweeping an unshruded (running clearance) statr blade des nt result in less lss fr the entire blade. The baseline cnfiguratin is superir in the lwer tw-thirds f the span in terms f lss. Sweeping des result in imprved turning in the endwall regins and less lading in the vicinity f the casing, hwever. Measured ttal-head data btained at statin 4.0 (between the secnd stage rtr and statr) were very similar fr the baseline and mdified statr builds [4] and are represented by the cnturs f Fig. 15. Between measurement statins 3.0 and 4.0, the range f ttal-head values ver the passage is reduced substantially. At statin 3.0, the range is 350 Nm/kg; at statin 4.0, the range is 160 t 180 Nm/kg. There is ample evidence f statr wake chpping and dispersin as described by Smith [8]. A casing wall "bundary layer" is apparent. (a) BASELINE UNSHROU ON OF STATOR PITCH CU VE CURVE VALUE LA EL OF TOTAL HEAD (N-N/KG) ED (STATIONARY CLEARANCE) STATOR Fig. 16. Secnd stage statr exit ttal-head cntur maps fr each build (q) = 0.500). 0, / CURVE CURVE VALUE LABEL OF TOTAL HEAD Fig. 15. Representative secnd stage rtr exit ttal-head cntur map (4) = 0.500). 5 0,5 CURVE CURVE VALUE LABEL OF TOTAL HEAD (b) BASELINE SHROUDED STATOR Fig. 16. Cntinued. 8

9 Measured ttal-head data acquired at statin 5.0 (dwnstream f the secnd stage statr) fr the baseline and swept statr builds are displayed in Fig. 16. Gemetrical details assciated with each f the builds are easily seen in Fig. 5. As in the first stage, ttal-head lsses ccur alng all pathlines between statins 4.0 and 5.0, even within the cre flw. The ttal-head patterns in the casing and uter half f the annulus regins are similar fr the tw baseline statr builds and fr the tw swept statr builds. They differ, hwever, between baseline and swept statr blades. Hub shruding did nt significantly influence the flw in the uter half f the flw passage. The swept leading edge resulted in a substantial reductin f lw-mmentum fluid build up in the suctin side casing (uter diameter) crner and a sizeable increase f lwmmentum fluid in the pressure side casing crner. On a circumferential-average basis (see Fig. 17), there is much less ttal-head lss in the casing endwall regin (70% t 95% span frm hub) with the swept statr builds in cmparisn t the baseline nes. The leading-edge sweep als resulted in lwer deviatin angles in the casing endwall regin (see Fig. 18). The secnd stage swept statr blade was laded less than the baseline blade in the casing regin as in the first stage (Table 5). In the statinary hub regin f flw, all f the builds behaved differently. The ttal-head cntur data f Figs. 16(a) and 16(b) indicate that hub clearance leakage flw can result in a large accumulatin f lwmmentum fluid in the suctin hub crner. This phenmenn has been bserved by thers (see, fr example, Ref. 9). The extent f this "leakage vrtex" is substantially suppressed by the swept leading edge. The circumferential-average lss data f Fig. 17 demnstrate the crrespnding large reductins in lss in the hub regin (5% t 30% span) pssible with sweep. Shruding appears t be beneficial in terms f reduced hub lsses fr baseline and swept statrs but mre extensively s fr the baseline cnfiguratin (Fig. 17). Sweeping a shruded statr leads t better crner flw management. Data frm Fig. 18 suggest that sweeping and shruding imprve blade turning perfrmance in the hub endwall regin. The imprvement in stall margin nted earlier CURVE CURVE VALUE LABEL OF TOTAL HEAD (c) SWEPT UNSHROUDED (STATIONARY CLEARANCE) STATOR Fig. 16. Cntinued ' BASELINE UNSHROUDED (STATIONARY CLEARANCE) STATOR BASELINE SHROUDED STATOR SWEPT UNSHROUDED (STATIONARY CLEARANCE) STATOR 4- SWEPT SHROUDED STATOR UNCERTAINTY I LE TOP PERCENT SPAN FROM HUB a 50 Fig. 17. Cnstant span ttal-head lss values fr baseline and swept unshruded (statinary clearance) and shruded statrs (ci) = 0.500) O. (La l. 0 H UB CURVE CURVE VALUE LABEL OF TOTAL HEAD (H-M/KG) (d) SWEPT SHROUDED STATOR Fig. 16. Cncluded. 0 0 UNCERTAINTY al) BASELINE UNSHROUDED (STATIONARY CLEARANCE) STATOR SWEPT UNSHROUDED (STATIONARY CLEARANCE) STATOR SWEPT SHROUDED STATOR PERCENT SPAN FROM HUB Fig. 18. Deviatin angles fr baseline and swept unshruded (statinary clearance) and shruded statrs 01) = 0.500). 9

10 in this paper culd be related t the extent f the secnd stage statr hub leakage vrtex. Sweeping and/r shruding suppressed the frmatin f this vrtex. Shruding did nt affect statr blade-element lading near the hub; sweep resulted in nly a small reductin in hub lading (Table 5). Near midspan, the swept statr blades suffered larger lsses (see Fig. 17) and mre deviatin in flw turning (see Fig. 18) than the baseline statr blades, suggesting that lw-mmentum fluid frm the endwall regins f the swept statrs was mved tward midspan. Lading at mid-span was similar (within uncertainty limits) fr all secnd stage statr cnfiguratins (Table 5). Passage-average statr lss cefficients fr the different secnd stage statr builds are listed in Table 7. The imprvements pssible with statr leadingedge sweep are significant. A cmparisn f the passage-average data f Tables 6 and 7 suggests that the baseline unshruded (running clear- REFERENCES ante) statr perfrmed best f all thse tested in terms f minimum lss. Inspectin f mre detailed results (see Figs. 11, 12, and 13) culd lead t a prpsal that in a running clearance situatin, sweeping the statr leading edge near the casing nly, wuld be beneficial. Cmparisn f the lss data f Figs. 13 and 17 pints ut that the running hub clearance resulted in a significant reductin f lss ver mst f the lwer half f the baseline statr span. Table 7. Cmparisn f passage-average statr lss cefficients fr the secnd stage statr builds. Build Statr Lss Cefficient Baseline unshruded (statinary clearance) statr Baseline shruded statr Swept unshruded (statinary clearance) statr Swept shruded statr CONCLUSIONS Leading-edge sweep in the endwall regins f a cmpressr statr blade can be beneficial n the casing (uter diameter) end f a statr blade. It can als help at the hub end f a statr blade when either a statinary hub clearance r a hub shrud is used. Sweep can be detrimental n the hub end f a statr blade when a running hub clearance is used. If structural cnsideratins allw a statr hub clearance, a running clearance is aerdynamically preferable t a statinary ne. ACKNOWLEDGMENTS We are grateful t Gerge Servy and Art Wennerstrm fr suggesting that a study f statr leading-edge sweep as a pssible means f imprving axial-flw cmpressr endwall flws might be useful. Gerge Servy and Elmer Hansen were principals in the blade design prcess f this prject. We appreciate their expertise. The Air Frce Office f Scientific Research (Dr. James D. Wilsn, Prgram Manager) is acknwledged fr its financial supprt. The Iwa State University Engineering Research Institute Office f Research Services staff was very helpful in the preparatin f this paper. We thank them fr their cmpetent assistance. 1. Sen, Y., Taylr, E. S., Batra, S. K., and Hinck, E., "Cntrl f Wall Bundary Layer in an Axial Cmpressr," Massachusetts Institute f Technlgy Gas Turbine Labratry Reprt Number 59, Cambridge, Massachusetts, June Wennerstrm, A. J., Law, C. H., Buzzell, W. A., and Derse, R. D., "Investigatin f a 1500 ft/sec, Transnic, High-Thrugh-Flw, Single Stage Axial Flw Cmpressr With Lw Hub/Tip Rati," AFAPL-TR , Air Frce Aer Prpulsin Labratry, Wright Pattersn AFB, Ohi, Octber Hathaway, M. D. and Okiishi, T. H., "Aerdynamic Design and Perfrmance f a Tw-Stage, Axial-Flw Cmpressr (Baseline)," TCRL-24, Iwa State University Turbmachinery Cmpnents Research Labratry, Ames, Iwa, December Tweedt, D. L. and Okiishi, T. H., "Statr Blade Rw Gemetry Mdificatin Influence n Tw-Stage, Axial- Flw Cmpressr Aerdynamic Perfrmance," TCRL-25, Iwa State University Turbmachinery Cmpnents Research Labratry, Ames, Iwa, December Kline, S. J. and McClintck, F. A., "Describing Uncertainties in Single Sample Experiments," Mechanical Engineering, Vl. 75, 1953, pp Mffat, R. J., "Cntributins t the Thery f Single-Sample Uncertainty Analysis," Transactins f the ASME, Jurnal f Fluids Engineering, Vl. 104, 1982, pp Hansen, J. L. and Okiishi, T. H., "Axial-Flw Cmpressr Rtr-Statr Interactin," TCRL-31, Iwa State University Turbmachinery Cmpnents Research Labratry, Ames, Iwa, December Smith, L. H., Jr., "Wake Dispersin in Turbmachines," Transactins f the ASME, Jurnal f Basic Engineering, Vl. 88D, 1966, pp Lebeuf, F., Bari, F., Bris, G., and Papailiu, K. D., "Experimental Study and Theretical Predictin f Secndary Flws in a Transnic Axial Flw Cmpressr," ASME Paper N. 82-GT-14,