Document
N A S A -TM-83627
DOE / NASA / 50194-39 I q ; _ qoo I q S ¢ D
• : NASA TM-83627
TR-84-C-7
Cold-Ai r Per f o r manceo f Compressor-Drive
Tur bine o f Department o f Energy Upgraded
Automobile Gas Turbine Engine
Ill--Perf ormance o f RedesignedTurbine
R i chard J. Roelke
National Aeronautics and Space Adminis t ration
Lewis Research Center
and
J effrey E . Haas
Propulsion Laboratory
AVSCOM Research and Technology Laboratories
Lewis Research Center
[Ii}FtARY COPY
May 1984
;'_ t t 6 1984
LANGLEYR ESEA R C H CE NT E R LI B R A RY, NA SA HAM P ]'ON _ VIR GI NIA
: Prepared for
U.S. DEPARTMENTOF ENERGY
Conservation and Renewable Energy
Off ice o f Vehicle and Engine R&D
.-; nf"":T4,:: r= ... f"" \ -rr : L
E-2G44 OOE/NASA/58194-39 NAS 1.15~83627 U5AAVSCOM-TR-84-C-7
UNCLASSIFIED DOCUMENT 4Ci PAt~ES
performance of. compressor-drive turbine of
upgraded automobile gas turbine eng1ne.
tiJi-=!) i rif: .... _r"""l. ...... r--: : ... ..- !
i r0- H/·f-::iJt.LF,t.~ o.J. o..}. t.. ..
CORP~ Nat1ona1 Aeronaut1cs a~d Space Admln stratton~
Research and Technology Labs.~ SAP~
A03/r~lF AEi·!
.. ~ ----~~----------------------------------------~~----
-----
DOE / NASA / 50194-39
NASA TM-83627
TR-84-C-7
Cold-Ai r Performanceof Comp r esso r -Drive
Turbine of Departmentof EnergyUpgraded
Automobile Gas Turbine Engine
Ill-- Performance of Redesigned Turbine
Richard J. Roelke
National Aeronautics and Space Administration
Lewis Research Center
Cleveland, Ohio 44135
and
Jeffrey E . Haas
Propulsion Laboratory
AVSCOM Research and Technology Laboratories
Lewis Research Center
Cleveland, Ohio 44135
May 1984
Work performed for
U.S. DEPARTMENTOF ENERGY
Conservation and Renewable Energy
Office of Vehicle and Engine R&D
Washington, D.C. 20545
Under Interagency Agreement DE-AI01-80C550194
COLD-AIR PERFORMANCEOF COMPRESSOR-DRIVE TURBINE OF DEPARIMENI OF ENERGY UPGRADED AUIOMOBILEGAS 1URBINE ENGINE Ill Pe r fo r man c e o f Redesignedlu r blne by Richa r d J. Roelke National Aeronauticsand Space Admin%stration Lewis Research Cente r Cleveland,Ohio 44135 and Jeff r ey E. Haas U. S. Army Researchand TechnologyLabo r atory Lewis ResearchCenter Cleveland,Ohio SUMMARY lhe aerodynamicperformanceof a redesignedcompressor-drlve turbine of the Departmentof Energy Upgraded Gas Turbine engine was determinedin air at nominal inlet conditionsof 325 K and 0.8 bar. Compa r ed to the fi r st turbine design the subject turbine had a lowe r flow factor, highe r rotor reaction,and a redesignedinlet m anifold. Two versions of the same rotor were tested: an o as- c ast rotor and the s ame rotorwith redu c ed surface roughness. Tests were , also made to determinethe effect of Reynolds number on the turbine performance.
lhe measured turbine efflc%encyvalues at design speed and work were 0.854 and 0.859 for the as-cast and reduced roughnessrotors, respectively.
lhese efflclencieswere obtained with a rotor tip clearanceof 1.2 percent.
At the design clearanceof 2.0 percent a decrease in turbine effic%encyof 0.018 was calculated,resultingin efficlenclesof 0.836 and 0.841 for the two rotors. The design efficiencygoal was 0.85.
At equal rotor tip clearancesand design point operationthe efflc%ency of the redesignedturbine increased0.023 compared to the original design. An analysis of the two turbines indicatedthat the primary reason for the perfor- mance improvementof the redesignedturbinewas lower rotor losses, lhere was no change in efficiencyof the redesignedturbine for the range of Reynolds number covered.
INIRODUCIION lhe Departmentof Energy (DOE) sponsoredan engine research program to design, build, and test an Upgraded Gas Turbine (UG1) automotiveengine, lhe objectivewas to demonstratean updated technologygas turbineengine with fuel economy equal to or better than a conventionalreciprocating engine and having low emissions. The Chrysler Corporationwas awarded the DOE engine contractand the NASA Lewis Research Center agreed to technicallymanage the cont r act. A gene r al descriptionof the UGI engine is given in referenceI.
The Lewis Resear c h Center also agreed to provide the initialaerodynami c designs of the compressor,compressor - drive t u rbine,an d power turbine,as well as conduct the performancetests of these components. This report is the last in a series pertainingto the aerodynamicperformanceof the compressor- drive turbine.
Two compressor-drive turbineswere designed and built for the UGI engine. The first was designed at Lewis and is described in reference2. For reasons of engine packaging,casting Fabrication,and engine acceleration time, several design constraintswere imposedby Chrysler on this first design. The constraintsthat most affected the aerodynamicperformancewere the mandating of a single-stageaxlal--flow deslgn havlng a work factor (i.e., Ah / Um2) of 2.1, to minimize the polar moment of inertia; a relativelythick blade trailing edge for casting purposes;a higher than optimum flow factor (i.e., Vx / Um) to lower blade stress; and a lower than optimum rotor reac- tion to minimize the exit swirl.
Stator inlet and exit surveys of thi s initial design (ref. 3 ), indicated that the flow characteristics deviated significantlyfrom the design intent.
In particular,there were thick inlet boundary layers and high incidence angles at the endwalls resultingin large losses at the stator hub and tip.
Concurrentengine tests made at the Chrysler Corporation(ref. 4), indicated that the compressor-drlve turbinewas not meeting its performancegoals at all engine speeds tested. Later engine tests made at Lewis (ref. 5), indicated that the compressor-drlve turbine approachedits performancegoal at 95 per- cent design speed but fell short at lower speeds. It was left to the compo- nent stage tests to obtain a more definitiveassessmentof the turbine performance.
The initial stage test of this turbine, re f eren c e6, showed an efficiency at design speed and work of only 0.78 compared to the design efficiencyof 0.85. Although subsequentcomponenttesting (ref. 7). demonstrateda stage efficiencyof 0.825 after reworkingthe blade profl]es to correct casting inaccuracies,the initial indicationof poor performanceand considerationof project schedulesresulted in a decision to design a second turbine. This turbinewas designed by Chrysler utilizingthe initial test results of the first design and with Pratt& Whitney / Canadaacting as a consultant.
For this design some of the constraintsplaced on the first design were relaxed. The new design had increasedrotor reaction,a lower flow coeffi- cient, at the expense of increasedblade stress,and a redesignedInlet mani- fold. The turbinewas also designed with a nonuniformradial work distribu- tion and a contouredstator shroud, lhe aerodynamicdesign of this turbine is briefly described in reference4 and additionaldetails are Included herein.
The experimentalcold-air evaluationand analysis of the resultsof this tur- bine is the subject of this report.
The t u rbine bladlng used in t h e componentperformancetests consisted of as-cast hardware representative of the stator and rotor castings used in the engines. Because of the relativelyrough surfacefinish of the as-cast blad- %ng, a second test was made with reduced rotor blade surface roughness. The stator was not modified. The as-cast turbinewas also tested over a range of inlet total pressuresto evaluate Reynolds number effects.
The perfor m an c eof the turbine at its englne Reyn o lds numbe r was dete r - mined with air at a nomlnal-inlettemperatureof 325 K and an inlet pre s sure of 0.8 bar absolute. Performan c edata were taken at total-to-total pressure ratios from 1.4 to 2.4 and rotatlve speeds f r om 50 to llO percent of equlva- lent design speed. Stato r Inlet su r veysof total pres s ure and flow angle w e r e taken at three stator pressure ratios and rotor exit radial surveys of total pressu r e,total temperature,and flow angle were made at equivalentdesign speed and design work factor. For the Reynoldsnumber tests the inlet pres- sure was varied from 0.4 to 1.6 ba r s absolute, resultingin Reynolds numbe r s, based on mean blade r adius f r om 1.2xlO 5 to 4.8xi05.
The aerodynamicperformanceof the compressor-drive turbine is presented in terms of equivalentmass flow, torque, spe c ificwork, efficiency,and flow surveys. A comparisonis made between the performanceof thls turbineand the initiallydesigned turbine.
SYMBOLS AR blade aspect ratio based on actual mean Chord length and exit blade height c a c tual c ho r d , c m c heat capacit y at constant pressu r e,J / (kg)(K) eR rotor kineti c energy loss c oefficient,I - W . 3 / W .3,id
stator kinetic energy loss coefficient,i - V2.5 / V2.5,i d_ _ Vs
Ah specificwork, J / kg m mass flow rate, kg / sec Am incrementalmass flow rate, kg / sec p absolute pressure,bars Re Reynolds number, m / _r m R rotor reaction,(P5.5- P6.3) / (P4.5- P6.3 ) X r radius, cm s blade spacing, cm l absolute temperature,K U blade velocity,m / sec V absolute gas velocity,m / sec AV change in abs o lute tangentialvelocity,m / sec u W relative gas velocity,m / sec WF work factor, Ah / U 2 m , _ abs o lute gas fl o w angle measured from axial direction,deg B relativegas flow angle measured from axlal direction,deg x ratio of spe c ifi c h eats a ratio of inlet total pressu r e to U. S. standard sea- l evel pressure, P'4.5 / p * c fun c tion of y used in relating parametersto those using alr inlet conditionsat U. S. standard sea-levelconditions, (0 .740 / y ) [(y+l) / 2] Y / (Y-I) I I n' efficiencybased on total pressure ratio P4.5 / P6.3 I anstag e loss in stage total ef f iciency e s quared ratio of c riti c al velo c ity at tu r blne-lnlettemperatureto cr critical velocity at U. S. standard sea-leveltemperature, • 2 (Vcr / V c r) viscosity,kg / m sec flow factor, Vx / U m T t o r qu e, N m mass flow parameterused in equation (4) Subscripts: av average cr conditioncorrespondingto Mach l Id ideal local local cond_tlon m mean meas meas u red s ur s urvey T total x axial dire c tion 4.5 station at manifold inlet (fig. l) 5 station at stator inlet (fig. l) 5.5 station at stator exit (fig. l) 6.3 stationat rotor exlt.(flg,l) Superscripts: ' absolute total state " relativetotal state • U.S. standard sea - levelconditions(temperatu r e,288.15 K; pressure 1.013 bars) IURBINE DESIGN lhe second design of the UGI compressor-dr_ve turbine was a s_ngle stage axial-flow machine w_th a rotor t_p d_ameter of 11.46 cm. The _nlet man_fold was a sp_ral-shaped volute w_th a s_ngle entr y , lhe stator had a contoured outer wall w_th a vane height of 1.65 cm at the leading edge and 1.28 cm at the tra_l_ng edge. lhe blades of th_s turbine were 15 percent longer than the f_rst tu r bine design. A cross-sect_on of the turbine as _t appeared _n the test r_g _s shown _n f_gure I. lhe _nstrumentat_on stations shown _n f_gure 1 are further defined _n the section RESEARCHEQUIPMENI AND PROCEDURES. T he dupl_cate engine parts used _n the test r_g were the _nlet volute, the stator r_ng, and the rotor. Photos of these parts are shown _n f_gure 2.
lhe hot engine, equivalent design and nominal component test conditions are l_sted _n table I. T he turbine hardware was fabricated sl_ghtl y undersize so that the flow passage would expand to the design area when the engine was operating at the design _nlet temperature. T hus, _t _s necessar y to show the equivalent flow conditions for both hot and cold hardware. The _nlet tempera- ture during the component test (table I), was selected to avoid exhaust duct _c_ng and the _nlet pressure was set to repl_cate the hot turbine Reynolds number.
lhe turbine design velocit y d_agrams are shown _n f_gure 3. T he radial gradients of pressure and tangential momentum used to generate these d_ag r ams were based on the surve y data obtained w_th the f_rst _nlet volute. Compare- son of these d_agrams to those of the f_rst turbine design (ref. 2), shows that the second design had lower stator velocities, h_gher rotor reaction, and _ncreased exit sw_rl. Spec_f_call y , at the mean radius, the second turbine design had a stator exit velocity ratio of 0.847 versus 0.929, a rotor reac- tion, Rx, of 0.357 versus 0.258, and an exit swirl of 30.9 ° versus 21.1 ° T he design radial variation of work _s shown _n f_gure 4. As the f_gure shows the specific work was greatest at the m_d-span and reduced at the endwalls.
lhe redesigned _nlet man_fold _s a s_ngle ent ry volute, as was the f_rst man_fold, but w_th much larger volume and w_thout the ax_s y mmetr_c chute at the stator _nlet. The calculated velocities _n the redesigned volute were about 30 percent lower than the original volute, lhe ax_s y mmetr_c chute of the f_rst design was replaced by a h_ghly converging section at the stator _nlet. These changes _n the volute design were made to reduce the turbine _nlet wall boundar y la y ers which were as thick as 20 percent of the passage _n the f_rst design.
lhe stator and rotor profiles are shown _n f_gure 5. lhe stator had 15 vanes, an aspect ratio of 0.43 (based on the exit blade height), and a con toured shroud wall. Table II l_sts further design parameters. Major geomet r_c d_fferences between th_s stator and the f_rst stator design are: nom_nal- ly 30 percent h_gher sol_d_ty, exit angles 6 ° to 7 ° nearer tangential _n the mean and t_p regions, and design _nc_dence angles 9 to 15 degrees h_gher at the hub and t_p, respectivel y .
lhe design parameters for the rotor are l_sted _n table llI. The so|_d_- t y of th_s design was nom_nall y 8 percent h_gher than the f_rst design and had 8 ° to 14° more turning at the hub and mean sections, respectivel y , lhe design rotor t_p clearance was the same for both turbines (0.25 mm) but due to the longer blade of the se c ond design, the blade clearan c e,as a per c entageof the blade height, was slightly reduced, from 2.2 to 2.0. In the c omponentper- formance tests, the tip clearanceswere nominallyl.? and 1.2 percent of the rotor blade height for the first and second designs, respectively. Compari- sons of design parametersof the two turbinesare summarizedin table IV.
lhe design stator and rotor blade surfa c evelo c ltlesare shown in figures 6 and 7 . The c alculationwas made by Chrysler using the computer code described in reference8. A second calculationof the blade surfacevelo c i- ties was recentlymade to assist in analyzingthe measured turbine perfor- mance. The results of that computationare given in the section labeled AnalyticalResults.
RESEARCH EQUIPMENIAND PROCEDURE The a pp aratusu s ed in thi s inve s tigation co nsi s ted o f the r esea rc h t u r- bine, an alrbrake dynamometerused to control the speed and absorb and measure the power output of the turbine, an inlet and exhaust piping system in c luding flow controls,and appropriate instrumentation. Figure 8 shows a s c hemati c of the facility and a photographof the test installation. The rotationalspeed of the turbine was measured with an electroniccounter in c onjunctionwith a magnetic pi c kup and a shaft-mounted gear. Mass flow was measured with a ca11- brated venturi. Turbine torque was determinedby measuring the reaction torque of the alrbrake,which was mounted on air trunnion bearings,and adding correctionsfor the turbine bearings and seal losses and the coupllng and rotor disk windage loss. These tare losses were previouslymeasured and c or- respondedto about ?.5 percent of the measured torque obtained at design equivalentspeed and work factor, lhe torque was measuredwith a c ommer c ial straln-gageload cell.
The turbine instrumentation stations are s hown in figure I. Figure 9 shows the instrumentation at ea c h station. Stations 4.5 and 5 were c hosen because they correspondedto the station locationsin the UGT test engine (ref. 4). Stations 5.5 and 6.3 were added for c omponenttesting. Instrumen- tatlon at the manifold inlet (station 4.5) measured wall stati c pressure, total pressure,and total temperature. At the stator inlet (station 5), located approximately0.60 centimeterupstreamof the stator,the static pres- sure, total pressure,andflow angle were measured. Static pressureswere obtained from six taps, with three each on the inner and outer walls. The inner and outer wall taps were locatedopposite each other at different inter- vals around the circumference. Two radial traversingprobes, locatedmidway between adja c ent stator vanes, were used to determinethe radial variation in total pressure and flow angle. These probes were positionedat a fixed angle, and the total pressure and flow angle were determinedFrom calibration curves. At the stator exit (station 5.5) located l mm downstreamof the stator trailing edge, static pressureswere measured with six taps, with three each on the inner and outer walls, locatedopposite each other at different intervalsaround the c ircumference.
At the rotor exit (station 6.3), lo c ated about three axial chord lengths downstreamof the rotor, static pressure,total pressure,total temperature, and flow angle were measured. The stati c pressurewas measured with six taps, with thr e e each on the inner and outer walls, lhree self-alignlngradial- traversingprobes locatedaround the circumference were used for measurement of total pressure, total temperature, and flow angle. Station 6.3 was located downstreamof the rotor where the rotor blade wakes were uniformlym_xed-out.
lhe stage test program consistedof three parts: Part I determinedthe turbine performancewith the as-cast bladlngover a range of equivalenttotal pressure ratios and rotatlve speeds. The manifold-lnlet-total to rotor-exlt- total pressure ratio was varied from 1.4 to 2.4 and the speed from 50 to llO percent of equivalentdesign speed. Part II was a Reynolds number evaluation of the as-cast bladlng. Reynolds number was varied from 1.2xlO 5 to 4.8xi05 over a range of turbine pressure ratios at design equlvalentspeed.
Reynolds number was changed by varying the turbine inlet pressure from 0.4 to 1.6 bars absolute. The third part of the st a ge test program determinedthe turbine performancewith reduced rotor blade surface roughness. The blade surface finish was smoothed in the same manner as the first turbine. The suc-- tlon surface was hand-pollshedand a coat of lacquerwas applied to the pres- sure surface. Table V lists the rotor surfacefinishes of both turbines before and after smoothing, lhe particulargeometry of the Integrallycast stator preventedmeasuringand smoothingthe vane surfacefinish. The appear- ance of the surface finish of the as-cast stator was slm|lar to the as-cast rotor.
In each part of the test progam, a rotor-exitradial survey was first conductedat equivalentdesign values of speed and specificwork. Radlally mass-averagedvalues of flow angle, total temperature, and total pressurewere obtained for each of the three circumferential survey locationsat station 6.3. These mass-averagedvalues were then arithmetically averaged to obtain overall values. The survey probes were then positionedwith one each near the tip, near mldspan, and near the hub so that the average flow angle from these three positionswould correspondclosely to the overallmass-averagedvalue obtained from the survey. Performancedata were then obtained at other opera- ting conditions.
lhe stage evaluationwas conducted in air at nominal inlet conditionsof 326 K and a range of turbine-lnletpressuresfrom 0.4 to 1.6 bars absolute.
The turbinewas rated on the basis of total efficiency. The actual work was calculatedfrom torque, speed, and mass flow measurements. The ideal work was based on the manifold-lnlet-to-rotor-exlt total pressure ratio. The manifold- inlet (station 4.5) and rotor-exlt(station 6.3) total pressureswere caIcu- lated from mass flow, static pressure,total temperature,and flow angle. For the calculationof manlfold-inlet total pressure the flow angle was assumed L o be zero.
ANALYSIS MEIHOD In order to make a more d e tailed assessment of the efficiency Improvement of the second turbine compared to the first turbine, detailed loss analyses were made for both turbines. The procedure, described fully In reference 9, analytically calculates the turbine losses for a given turbine geometry, tur- bine operating condition, and flow field at the stator Inlet. For each turbine the stator and rotor coordinates were adjusted based on throat measurements to reflect the actual test hardware, lhe analyses were made at two o pe r atingc on diti o ns: at design equivalentspeed and work, and at 70 per- cent of design equivalentspeed and a stage pressure ratio of 1.45. This lat- te r conditio n correspondsto a pa r t power conditionwhere the turbineoperates a major part of the time when installedin the engine. The stator inlet flow characterlst_cs o f endwall d_spla c ementthicknesses,flow angle radial distri- butionsand volute total pressure losseswere obtained using expe r imental results from the two turbine tests. The measured stator pressure ratio was also specifiedfor the calculation. The procedurefollowed used the MERIDL and 1SONIC computer codes (refs. 8 and lO), to compute flow conditionsin the blade channels includingblade surfaceand endwall velocities. The BLAYER compute r code (ref. ll), was t hen used to calculatestator and rotor d_splace - ment and momentum thlcknessess, which were, in turn, used to calculateprofile friction losses (includingthe mixing loss) and endwall frlct_on losses.
Additionalpublishedcorrelationswere used to calculate losses due to inci- dence, secondaryflow, rotor tip clearance,and the exhaust duct friction.
Losses were calculatedas kineticenergy loss coefficients(_) for the stator and rotor. For each blade row, the total kineticenergy loss coefficientwas converted into a stage efficiencyloss using the followingequations: , \P6.3 / - - eS'T . eS'T 1 - eR,1 + eR, T \P6.3 ] n =
y-l
For the stator: !
Anstag e = l - n', when eR,T = 0 For the r otor: A n stag e l n' - ' Anstage ' = - (stator),where eS,T, and eR,T, _ 0 The stage efflc l encylosses for the stator and rotorwere added to those for the manifold and exhaust duct to calculatean overall efficiency.
Since the boundary layer calculationused in the analysis implicitly applies to smoo t h surfaces only, the calculatedprofile and endwall Friction losses were assumed those for smooth rotor blades. For this reason compar_- sons of the analyticallycalculatedturbine flow characteristics were made only with the experimentalresults obtained for the turbineswith smooth rotor blades.
RESULTSAND DISCUSSION The perf o r m an c e o f the redesignedt u rbine i s presented in five parts.
The results of the stator inlet survey are presentedfirst Followedby the overall stage performancefor the as- c ast bladlng. The effect of Reynolds number on the turbine performanceis then presentedfollowed by the change in efficiencyfor the polished rotor blades. Finally, the performanceof this turbine is compa r ed , b o th analytlcally and experimentally, with the first tu r- blne design.
Stator Inlet Surveys Radial surveys of flow angle and total pressurewere made at the stator inlet (station 5) at two circumferential locations(fig. 9) for manlfold- Inlet-totalto stator-exlt-statlc pressure ratios of nominally1.38, 1.56, and 1.70. The results were similarfor all three pressure ratios. The survey results obtained at a pressure ratio of 1. 7 0 are shown in figure lO. The angle measurementsare plotted in figure lO(a). The dashed llne is the design radial variation in flow angle and the three filled-lnsymbols are the stator- inlet blade angles. The flow angle measuredwas generallybetween 40 and 50 degrees over most of the passage height but decreasednear the endwalls. The largest deviationbetween design and measured flow angle occurred in the outer half of the passagehelght. Although the flow directiondeviated appreciably from the design prediction,the net result was a slgnlflcantreductionin stator incidencewhich benefitedthe stage performance. The shape of the radial variation in flow angle at locationA suggestsa CCW vortex being shed from the shroud llp (fig. l). It is not known how far around the perlphery flow was being shed from the shroud llp in thls manner since much of the area CCW to locationA was not accessablefor surveying.
lhe radial variationsin manlfold-exlttotal pressureat the two survey locationsare plotted in figure lO(b), lhe total pressure variationsare slightlydifferent between the two locations. There appears to be a slight decrease in pressureas the flow moves around the volute. Wall boundary layers were about 12 percent thick at the hub and 5 percent thlck at the tip.
By comparisonthe boundary layers for the first volute were about 20 percent thick at the hub and 15 percent thick at the tip. The sheddlngof the flow from the shroud llp at locationA resulted in a local pressure loss from about 60 percent passage height to the shroud. The calculatedmass average pressure loss in the manifold was 1 / 3 percent.
lhe radial variationsin calculatedmass flow at the two survey locations are shown in figure II. The mass flows were calculatedfrom the survey resultsand are expressed in terms of the mass flow parameter,_. The mass flow parameter is defined as: (Am)iota I l - (am)av and was used to provide nondlmenslonallzed numbers because of the difference between the design and measured mass flows. The (am)ioca I was the calcu- lated mass flow through each of 40 equal Incremental flow areas. The (am)av was the summationof the calculatedlocalmass flows divided by 40. The cal- culationswere made assuming a linear variationin static pressure from hub to tip. The mass flow parameter,_, indicatesthe percentagedeviation in the local mass flow from the averagemass flow at a given radial and clrcumfer- entlal location. The dashed curve in the figure indicatesthe radial dlstrl- butlon in mass flow calculatedfrom the design diagrams.
The t o tal mass fl o w c al c ulatedfor ea c h of the two su r vey loc ati o nswas within 3 percent of each other but diffe r ed markedly from design. The mea- sured mass flow distribution c ompared to design had substantiallyless f low per unit area between the hub and about 60 per c ent span and substantla l l y more flow per unit area from there to the tip. The change in the radial dlstrlbu- tion of mass flow from that designed to the measured distributionwas probably not detrimentalto the turbine performance, lhe measured distributionindi- cated a shifting of mass flow awa y from the hub where losses were high. The general trend of more mass flow per unit area near the walls and less mass flow per unit area from about lO per c ent to 60 percent span is similar to that found in the first volute design (ref. 3). Perhaps the ideal mass flow dis- tributionwould be to reduce the mass flow near the walls where the losses are the highest and increase it in mldspan where the losses a r e low. This ideal- ized mass distributionwas not achieved in either volute design.
Performan c eof As-Cast Bladlng Mass flow. - The variation in equivalentmass flow rate with equivalent total pressure ratio and speed is shown in figure 12. The equivalentdesign mass flow given I n the figure is the design mass flow value for c old har d ware that is listed In table I. The data show that the rotor choked before the stator for all speeds above 70 percent and thus controlledthe turbinemass flow. The measured mass flow at design speed and the equivalentdesign pres- sure ratio of 2.048 was 0.311 kg / se c whi c h is 1.2 per c ent less than the design flow. The rotor throat area was measured and found to be l percent too large. Therefore,the mass flow per unit area was 2.2 percent less than design.
Torq ue and specificwork. - The variationin equivalentt o rque with equivalentpressure ratio and speed is presentedin figure 13. At the equiva- lent design speed and pressure ratio, the measured equivalenttorque was 4.91 N-m. Th l s Is 0.8 per c ent less than design. Since the mass flow was 1.2 per- cent less than design, the turbine specificwork was 0.4 percentgreater than design. This is illustratedin figure 14 which shows that at the equivalent design total pressure ratio of 2.048 and design rotor speed the work output of the turbine was 45,800 3 / kg.
Efficien cy. - The turbine efficiencyis shown in figure 15. The two sym- bols on the figure indicatethe design (filled-ln)and a part power (open) condition. The estimatedpart powe r turbine operatingparametersof speed and pressure ratio were calculatedfrom data given in reference4. The predicted part power effi c ien c ywas 0.84. At this part power conditionthe measured efficiencywas 0.83 7 whi c h is in very good agreementwith the predicted value. At equivalentdesign speed and pressure ratio, the turbine efficien c y was 0.854 which is slightly higher than the design goal of 0.85. Although the test effl c lenciescompare well with the goals, it should be noted that the e xperimentaleffl c len c ieswere obtainedwith a tip c learan c eof 1.2 per c ent rather than the design tip clearance of 2 percent. The effect on efficien c y of this tip c learancechange is discussed in the sectionentitled Effect of rotor tip clearance.
Rotor-exltsurvey. - The resultsof the radial surveys at station6.3 of flow angle, total pressure,and total temperatureare shown in figures 16(a) lO to ( c) . T he measurementswere taken with t h e tu r bin e op er atingat eq u i v alent design speed and spe c ifi c work. The data shown are the mea s urementsof the threesurvey probes. With these measurements,the radial variation in stage effi c iencywas calculatedand is shown in figure 17. The dashed curves in the figures representthe c alculateddesign radial variationsat the turbine exlt taken from the deslgnreport.
The measured flow angle (fig. 16(a)), shows the flow was turned more tangentiallynear the hub and shroud and less in mldspan. Differencesin magnitudeand trends existed between the measured flow angle and the cal c u- lated design angle. Similar differencesbetween design c a lc ulationsand mea- surementsare also evident for the pressure and temperatureshown in figures 16(b) and (c). It appears that the blade shapes generatedduring the design were inadequateto establishthe design flow condltlonsL However, a re c ent anal y ti c alrecalculation of the turbine flow conditionsfor the blade shapes tested gave resultsthat agreed much better with the measurements. That an a l y s is I s d isc u s sed l a te r I n this r ep or t.
The measured overall pressure ratio ( fig. 16(b)), i nd i cateda la r ge r gradient from hub to mldspan to tip than the design. The lowest pressure ratio o cc urred at mldspan and the highest pressure rat i os occurred at the h u b and tip. The temperaturemeasurements(fig. 16( c )), indi c atedvery uniform work extra c tionradially and differed from the design intent.
The radial var i ation i n effi c iency(fig. l?), a grees very well with design. This result of achievingdesign effi c iencywhile not establishing design flow gradientsm a y be unexpected. However, the devi a tion in flow angle from design at the stator inlet redu c ed the inlet inciden c e. Also, the more uniform radial m a ss flow distribution(f i g. ll), was probabl y benefi c ial.
Finally, the design values of efficienc y in figure 1 7 were calculatedfor a 2 percent rotor tip clearancewhereas the test was m a de with a 1.2 per c ent tip clearance.
Effe c t o f R ey n o ldsNumber A Reyn o lds number test was made with the as- c ast bladlng. As stated earlier the Reynolds number was varied by varying the turbine inlet pressu r e while holding the speed and pressure ratio constant. For ea c h inlet pressure the Reynolds number and turbine efficiencywere cal c ulated,from smooth-curve data, at the design work factor of 2.1. The test results are shown in figure 18. The Reynolds number at the hot-englnedes i gn conditionwas 2.36xi05. As the data indicate,no effe c t of Reynolds number was measured over the range covered.
Effect of Blade Surface Roughness During the performan c eevaluationof the first turbine design (ref. 6), it was found that smoothingthe as- c ast rotor blade surfa c e finish resulted in a one po i nt gain in efficiency. The resultsof retestlngthe second turb l ne after smoothingthe rotor blades are shown in figure 19. Overall performance was measured over a range of pressure ratios at equivalentdesign speed. Over the entire range of pressure ratios,the insertionof the polishe d rotor II resulted I n no m inallyI / 2 point improvementin efflc_ency, lhls Is one-half the improvementin efficiencymeasured with the first turbine design. Radial flow surveystaken at the exit of the polished rotor (station 6.3) showed resultsalmost identicalto those for the as-cast rotor.
AnalyticalResults The objectiveof the loss analysis was to determine the reasonsthe second turbine performedbetter than the first turbine. This was done by cal- culatlng the individualloss componentsfor the two turbines at the two opera- ting conditions. Since the loss analysis procedurewas largelytheoretical, c onfidencein the final calculatedlosses is increasedby experimentally verifyingthe calculationswhere possible. This was done by comparingthe stage exit flow conditionsand overall efflclenclescalculatedfrom the loss analysis with the exit flow conditionsand efflclenclesobtained from experi- mental data. Also, to assist in explainingthe reason a calculatedloss Increasedfrom one turbine to the other, blade surfacevelocity d_strlbutlons are presentedfor both turbines for both operatingconditions. The results are first presentedfor the design speed and work conditionand then for the part power condition. The section ends with a discussionof the effect on turbine efficiencydue to the differencein rotor tip clearancethat existed between the two turbines.
Analysis at design speed and work. - The calculatedvelocitydiagrams for the second turbine at the design work and speed conditionare shown in figure 20. The design velocity diagrams from figure 3 are superimposed for compari- son. A comparisonof the calculatedvelocity diagramswith the design dia- grams shows significantdeviations, lhe calculateddiagrams show larger velocit i esand flow angles at the stator exit (station 5.5) for all three radial sections. At the rotor exit (station 6.3) the results show lower rela- tive velocitiesat the hub and mean sections and lower absolute velocitiesat all three sect i ons.
A comparisonof the analyticallycalculatedrotor exit flow angles and velocitieswith those obtained from the rotor exit radial survey measurements is shown in figure 21. The experimentalradial variationsin velocitiesand flow angles were computed from the rotor exit survey for the smoothedblade.
The dashed curves are the analyticallycalculatedvalues. The agreement between the experimentaland analyticalresultsover most of the blade height is excellent. Some deviationsexist at the hub and tip sections but the agreement is still consideredgood. Since MERIDL is an Invlscld program that calculatesflow propertiesalong a hub to shroud mldchannelstream surface, the best agreementwould be expectedaway from the endwalls were viscousand tip clearanceeffects are not a factor. Since stator exit radial surveyswere not made, a similar comparisonbetween the analyticaland experimentalresults at that location could not be made.
The surface velocitiescalculatedin the loss analysis for the first and second stator designs are shown in figures 22(a) and 23(a), respectively. The design surface velocitiesare superimposedfor comparison. For the first stator (fig. 22(a)), there were large velocitypeaks near the leading edge on the suction surfaces of all three vane sections,followed by relativelycon- stant velocitiesto the trailing edge. These velocity distributionsdiffered from the design dist r ibution s , be c au s e (1 ) the cur rent r esultswe r e obtained with a quasl-three-dlmenslonal computationinstead of the two-dlmenslonal c om- putation used in the design, (2) the stator geomet r ywas changed to reflect the actual hardware,and (3) the stator inlet flow conditionswere obtained from test measurementsrather than design estimates.
The vane surfacevelocitlescalculatedin the loss analysis for the rede- signed stator, figure 23(a) showed moderate accelerationon the suction su r - face of the hub and mean sections. However, at the contouredtip section, the vane was unloaded over the first 25 percent of the chord. This was followed by a rapid a cc elerationto the 75 percentchord locationand then large dif- fusion to the trailing edge. The crossingof the surfacevelocit y c urves near the trailing edge indicatesexcessivevane camber for the exit velocitydia- gram. Differencesthat exist between the surfa c e velocitiesfrom the loss analysis and the design distributionsare attributedto differencesin the streamsheetthicknessesused for the two calculations. The calculations made for the analysis used the streamsheetthicknessobtained from the MERIDL pro- gram which was not used when this stator was designed.
Figures 22(b) and 23(b) show the surfacevelocitiescalculatedfor the loss analysis for the first and second rotor designs, respectively, with the design variationssuperimposed for comparison. The first turbine rotor shows moderate diffusionon the suctionand pressure surfaces of all three sections and good agreementwith the design loading. The second rotor shows min%mal diffusionat all three se c tions. Compa r ed to the design variations,there was good agreementat the hub section, but only fair agreementat the mean and tip s e c tions.
The results of the analyticallycalculatedlosses for the two turbines are shown in table VI. The stator losses include profile friction ( including m ixing), endwall friction,secondaryflow, and incidence. The losses ar e tabulated in the sa m e m anner in the rotorwith the inclusionof tip clearance loss. Two main conclusionscan be drawn from this table. First , there was good agree m ent between the anal y ticall y calculate d an d experimentalef f l- ciencies for both turbines. Second, the analysis shows that most of the effl- cienc y differencebetween the two turbineswas due to reduced rotor losses f or the second turbine, k discussionof the individuallosses in the s tator and rotor for the two turbines follow s .
The total kineticenergy loss for the second stator,table VI, was only 0.00 3 less than that calculatedfor the first stator. In fact, the profile friction loss for the second stator was larger by 0.00 7 co m pared to the first stator, indicatingthat a benefit d u e to use of a contoured stator was appar- entl y not realized. The reason for this is attributedto the surface velocity d iscussed in figures 2 2 (a) and 23 (a). For the first stator , the leadin g edge velocity peaks, although un d e s irable,did not substantiallyincrease the cal- culated profile friction loss. For the second s tator the hub and m ean section pro f ile losses were about the same as the first stator. However , there was a much higher profile loss calculatedfor the tip section due to the unfavorable velocitydistribution. The other three stator losses were lower for the s econd stator, particularlythe incidenceloss. Due to the lower stator pres- sure ratio of the sec o nd turbine co m pared to the first turbine the small reductionin the stator total kinetic energy loss for the redesignedturbine resulted in a 0.008 increase in stage efficiency.
fable VI s how s that for the s e c ond rotor design t h ere was a 0.050 lowe r total kinetic energy loss c oefficient. This 0.050 differen c etranslatedinto a 0.034 in c rease in stage effi c iency. All five loss categoriesfor this rotor were lower than the first rotor. The lower profile and endwall fri c tion losses are consistentwith the lower diffusionof the second rotor c ompared to the first rotor shown in figures 22(b) and 23(b).
Analysis at part power. - The stator and rotor surfa c evelocitiesare shown in figure 24 for the first turbine at the part powe r condition. The surfa c e velo c ities c alculatedin the loss analysis at the design speed and work condition (fig. 22), are includedfo r reference. The stator suction sur- fa c e velocitiesdid not show peaks near the leading edge as large as those at the design speed and work condition. Also, the stator exit velocity ratios were less, resultingin slightlymore diffusion. The rotor blade surface velo c itiesshowed larger differencesbetween the two flow c onditions. The lower level of rotor reactionat the off-deslgncondition c aused larger fric- tion losses.
The s tat o r and rot o r s urface vel o c itiesfor the se c ond turbineat the pa r t power conditionare shown in figure 25. The stator surfa c e velo c ities showed slight differencesbetween the two flow c onditions. The rotor surfa c e velocitiesshowed much larger diffe r enceswith the suction surfacevelo c ities for the part power condition showingmuch less ac c eleration. This, in turn, caused higher profile and endwall friction losses.
The tabulated re s ults of the lo s s analyses for the two turbines operating at the part power c onditionare given in table VII. For this operating c on- dition, the experimentalstage efflclenclesfor both turbineswere obtained from performan c edata for the as- c ast versionsof both turbines with adjust- ments made to account for efficiencyimprovements due to reworklngand / or polishingthe rotor blades. The cal c ulatedstage effi c iencyfor the se c ond turbine agrees well with the experimentalvalue. However, for the first tur- blne the calculatedefficien cy was nearl y two points lowe r than the experi- mental value.
A comparisonof the losses for the first turbine at the two operating conditions,tables VI and VII, showed increasedstator losses at the part power c onditiond u e mo s tly to i n c reasedsurfa c e fr ic t i on and incidence losses. The increasedstator losses caused a reductionin stage effi c ien c yof about two points. All of the rotor losses except for the rotor tip clearan c e loss were also significantlyhigher at the part power condition. The primary cause for the increasedlosses was the rotor in c idenceloss. The rotor inlet relative flow angles for the part power c onditionranged from 4° to 9° larger than at the design speed and work condition.
Similar conclusionscan be drawn fo r the second turbine in compa r ing the losses for the two operatingconditions. The stator loss increasedslightl y for the part power c ondition,which resulted in about a one point decrease in stage efficiency. The rotor total loss increasedby about 0.04, also causing about a one point decrease in stage efficiency. As with the first turbine, the primary cause for the increasedrotor losseswas due to rotor incidence.
Effect of rotor tip clearance.- The larger tip c learance losses for the first rotor listed in tables VI and VII were due to a larger tip clearan c e (1.7 per c ent of t h e rotor b l ade height c ompar e d t o 1. 2 per c ent for the s e co nd rotor) . Two additiona l c a s e s were ana l yzed to predl c t the in c rea s e In tip cl earan c e l o ss for the rede s igned turbine for rotor tip cl earan c e s equa l to 1 . 7 and 2.0 percent of the rotor b l ade height . The ca lc u l ation s were m ade for the de s ign s peed and work c ondition. The 1 .7 per c ent c a s e wa s done to mat c h the r o tor tlp cl earan c e f o r the first turbine . T h ¢ 2 . 0 per c ent c a s e wa s done to mat c h the de s ign cl earan c e va l ue . For the 1 .7 per c ent c a s e, the kineti c energ y l o ss due to tip cl earan c e increa s ed to 0.056 re s u l ting In an additional redu c tion In overa ll s tage effi c ien cy of 0.0 11 . F or the 2.0 per c ent c a s e, the kineti c energ y l o ss in c rea s ed to 0.066 re s u l ting In an addltlona l reduction in overal l s tage effi c ien c y of 0.0 1 8 .
Ba s ed on the re s u l t s di sc u ss ed above, If the 0.0 1 1 redu c tion tn s tage e f fi c ien cy were app l ied to the experimenta l re s u l ts, a mea s ured s tage effi- c ien cy of 0.848 wou l d be predi c ted f or the s e c ond turbine at a ttp c learan c e of 1 .7 per c ent. Thi s proje c ted ef f i c ien c y wou l d be 2.3 point s higher than the initia l turbine ef f i c ien cy at the same cl earan c e. Simi l ar ly , at the de s ign l eve l of rotor tlp cl earan c e (2.0 per c ent) a measured s tage effi c ien cy o f 0.84 1 wou l d be predi c ted, whi c h I s s l tght ly l e ss than the de s ign goa l of 0.8 5 .
CONCLUDINGREMARKS The results of the compressor-drlve turbine test series showed that good efficiencycan be obtainedwlth a small hlghly-loaded axial turbine. A poten- tial improvementIn efficiency,above that so far measured, exists as improved computer codes are developedand wlth better control In bladlng fabrication.
For example, neither inlet manlfold design generatedflow conditionsat the stator inlet that the design analysis predicted. However, the deviationsfrom design at the manifold exlt (stator inlet) for the second turbinedeslgn likely benefitedthe turbine performance. Also, the series of tests made wlth only small hardware changes (i.e., smoothingthe blade surfaces and, In the case of the first design, reworklngthe blade profiles)illustratesthe crlt- Icallty of having very accuratelymade airfoils In small machines.
Analysis of the test resultsof the two turbines indicatedthat most of the improvementIn performanceof the redesignedturbine occurred In the rotor. The calculatedstator kinetic energy loss for the two turbineswas virtuallythe same. Rotor kinetic energy losses due to profile and endwall frlctlon,mixing, and secondaryflow were lower In the redesignedturbine. It appears then that the higher efficiencyof the second design, at the same tlp clearance,was primarilydue to higher rotor reactionand reduced rotor blade diffusion. Benefits due to the contoured statorwall, larger volume inlet man}fold and the parabolicwork distributionwere not apparent.
SUMMARY OF RESULTS The aerodynamicperformanceof a redeslgnedcompressor-drlve turblne of the Departmentof Energy Upgraded Gas Turbine engine was determinedIn alr at a nominal inlet temperatureof 325 K. Two versions of the rotorwere tested: an as-cast rotor and the same rotor wlth reduced surface roughness. Reynolds number tests were made for the as-cast rotor by varying the inlet pressure from 0.4 to 1.6 bars absolute. The resultsof the investigation were as follows: 1. T h e tur b ine effi c ien c y v al u es at d e s ign s pee d and work we r e 0.854 and 0.859 f or the as-cast and reduced roughnessrotors , respectivel y . An anal y sis of the smooth rotor at a tip clearanceof 1. 7 percent indicatedan efficiencyof 0.848. This compares to an e f ficiencyof 0.825 for the initial- ly designed turbine at the same tip clearance. At the design tip clearanceof 2.0 percent a turbine efficiencyof 0.841 was indicatedfor the reducedrough- ne s s rotor con f iguration. There was no change in efficienc y with Re y nolds nu m ber.
2. An analysis of the two turbines indicatedthat the primary cause of the performanceimprovementof the redesignedturbinewas lower surfacefric- tion, mixing, and secondaryflow losses in the rotor.
3. The measured efficiencyat part power decreasedonly moderatelyand agreed well with the part power prediction.
4. The effect of several features incorporated in the redesignedtur- bine to reduce the aerodynamiclosses were inconclusive. These includea larger volume inlet volute,a contouredstator shroud, and parabolicwork distribution.
REFERENCES l. Ball, G. A.; Gumaer, J. I.; and Sebestyen,T. M.: The ERDA / Chrysler Upgraded Gas Turbine Engine Objectivesand Design. SAE Paper 760279, Aug. 1976.
2. Roelke, R. J.; and McLallln, K. L.: The AerodynamicDesign of a Compressor-Drive Turbine for Use in a 75 kW AutomotiveEngine. NASA TM X-?l?l?, 1975.
3. Roelke, R. J.; and Haas, J. E.: Cold-Air Performanceof Compressor-Drive Turbine of Departmentof Energy Upgraded AutomobileGas Turbine Engine.
I--Volute-Manlfold and Stator Performance. NASA TM-82682,1981.
4. Wagner, C. E.; and Pampreen,R. C.: Upgraded AutomotiveGas Turbine Engine Design and DevelopmentProgram, Vol. 2: Final Report.
(CO0-2749-43-VOL-2, Chrysler Corp.; EY-76-C-02-2749.) NASA CR-159671, 19?g.
5. Horvath, D.; et al: Test Results of the Chrysler Upgraded AutomatlveGas Turbine Engine - Initial Design. NASA TM-81660,1981.
6. Roelke, R. J.; and Haas, J. E.: Cold-Air Performanceof Compressor-Drive Turbine of Departmentof Energy Upgraded AutomobileGas Turbine Engine.
II--Stage Performance. NASA TM-82818,1982.
7. Roelke, R. J.; and Haas, J. E.: The Effect of Rotor Blade Thicknessand Surface Finish on the Performanceof a Small Axial Flow Turbine. ASME Paper 82-GT-222,Apr. 1982.
8. Katsanls,T.: FORTRAN Program for CalculatingTransonic Velocitieson a Blade-to-Blade Stream surface of a Turbomachlne. NASA TN D-5427, 1969.
9. Boyle, R. J.; and Haas, J. E.; Katsanls,T.: ComparisonBetween Measured and PredictedTurbine Stage Performanceand PredictedPerformanceUsing Quasl-3D Flow and Boundary Layer Analyses. NASA TM-83640, 1984.
lO. Katsanis,T.; and McNally, W. D.: Revised FORTRAN Program for Calcula- ting Velocitiesand Streamlineson the Hub-ShroudMidchannelStream Sur- face of an Axial-, Radial-, or Mixed-FlowTurbomachlneor Annular Duct.
I--User'sManual, NASA TN D-8430, Ig?7.
If. McNally, W. D.: FORTRAN Programfor CalculatingCompressibleLaminarand Turbulent Boundary Layers in ArbitraryPressure Gradients. NASA TN D-5681, 1970.
TABLE I. - TURBINE DESIGN PARAMETERS Equivalent Parameter Hot Hot Cold Component engine hardware hardware test Turblne-lnlettemperature,K 1325 ?88.2 288.2 325 Turblne-lnletpressure,bars 4.04 l.Ol l.Ol 0.83 Mass flow rate, kg / sec 0.588 0.323 0.315 0.242 Rotatlve speed, rpm 58 500 2? 673 27 673 29 386 Specific work, J / kg 203 7 00 45 600 45 600 51 400 Torque, N m 19.5 5.09 4.95 4.03 Power, kW I19. 7 14.7 14.3 12.4 ! I Total pressure ratio, P4.5 / P6.3 1.982 2.048 2.048 2.048 Total effi c ien c y,n' 0.85 0.85 0.85 0.85 Work factor, Ah / U_ 2.10 2.10 2.16 2.16 Reynolds number, m / pr m 236 000 355 930 351 000 244 000 Mean blade speed, m / se c 311.7 14 7 .4 145.4 154.4 1 7 TABLE II. - STATOR DESIGN PARAMEIERS Parameter Hub Mean T_p P r ofile radius at T.E., cm 4.445 5.107 5.730 Actual chord, cm 2.591 3.023 3.810 Axial chord, cm 1.449 1.269 1.506 Leading edge radius,cm 0.060 0.050 0.090 Tra i ling edge radius,cm 0.019 0.019 0.019 Tra_llng edge blockage,percent 5.0 5.9 5.8 Inlet blade angle, deg 29.2 45.0 38.7 Incidence,deg --ll.2 -4.1 -17.2 Exit blade angle, deg 64.3 71.6 7 3.2 Solidity,c / s 1.39 1.41 1.59 Blade number ...... 15 .....
Blade height at T.E., cm ...... 1.285 .....
Aspect ratio, AR ...... 0.43 .....
T AB L E Ill. - R O TOR DESIGN PARAMEIERS P ara meter Hub Mean Tip Profile radius,cm 4.445 5.088 5.?30 Actual •chord, cm 1.085 ].058 ] .035 Axial chord, cm 1.022 ].002 .852 Le a ding edge radi u s, ¢m 0.028 0.028 0.020 Trail l ng edge radius, cm 0.020 0.0 1 9 0.015 lra|l|ng edge blockage,percent 16.2 14.2 I].9 Inlet blade angle, deg 47.0 5].6 27.8 In c idence,deg -6.8 -6.1 -7.2 Exlt blade angle, deg 58.0 59.8 65. 1 Sol_dlty,c / s 2.33 1.98 1.725 Blade number ..... 60 .....
Blade height, cm ...... 1.285 ......
Aspect ratio, AR 1.215 ......
• 19 T ABLE I V . - COMPARISONOF COMPRESSOR-DRIVE TURBINE DESIGNS Parameter First design Second design STAGE: Mass flow rate, kg / sec 0.598 0.588 Specificw o rk, 3 / kg 1 98 lO0 203 700 Work factor, Ah / U_ 2.1 2.1 Flow factor, Vx / Um 0.88 0. 7 3 Exit swirl at mean, deg 21.0 30.9 Reynolds number,m / _r m 2.44xi05 2.44x105 MANIFOLD: Type volute volute Inlet velocity ratio, (V / Vcr)4,5 0.150 0.080 E x it v el oc i ty r a t i o , ( V / V c r ) 5 0. 40 5 0 .2 77 STATOR: Shroud type c ylindri c al c ontoured Average reaction,Rx 0.593 0.566 Average solidity, c / s l.lO 1.46 Aspect ratio, AR 0.484 0.430 Trailing edge thickness, c m 0.038 0.038 Average t. e. blo c kage,per c ent 4.4 5.6 ROTOR: Tip diameter, c m I t .1 II.46 Blade length, c m 1.13 1.28 Average rea c tion,Rx 0.246 0.3?3 Average solidity, c / s 1.86 2.01 Aspe c t ratio, AR 1.219 1.215 Tip c learance,per c ent 2.2 2.0 Tr a iling edge thlckness, c m 0.38 0.036 Average t. e. blockage,per c ent 12.0 14.0 TABLE V. - ROTOR SURFACE FINISH COMPARISON First design Se c ond design Blade Surfa c e Suction Pressure Su c tion Pressure As-Cast Finish, 1.4 1.4 2.0 2.0 m x lO-6 Reduced roughness 0.3 l.O 0.2 0.9 finish, m x lO-6 T A BLE V l. - C OMP A RISONOF ANALYTICALLYCALCULATED LOSSES AT DESIGN EQUIVALENTSPEED AND WOR K Fir s t de s i gn Se co n d d e si g n S T ATO R L O SSES ( es ): P rofile Fri c tion O.Ol? 0 .024 Endwall Friction .02 1 .019 Se c ondary .008 .006 Inciden c e .OlO .004 Total .056 .05 3 A n'stage O.047 0.038 P5 / Ps.5 1.86 1.69 ROTOR LOSSES (-eR): Profile Fr_ctlon 0.061 0.048 Hub Endwall Fri c tion .014 .008 Secondary .039 .026 In c iden c e .016 .005 Tip Clearance _.04__ 7 .040 Total .1 7 7 .12 7 A n'stage 0.116 0.082 I I P5.5 / P6.3 1 .61 1.5 7 MANIFOLD LOSS, An'stage 0.004 0.004 EXHAUST DUCT LOSS, An'stage 0.005 0.006 CAL C ULATEDOVERALL EFFICIENCY 0.828 0.8 7 0 MEASURED OVERALL EFFICIENCY 0.825 0.859 TABLE VII. - COMPARISONOF ANALYlICALLYCALCULAIED LOSSES A T 7 0 PERCEN T EQUIVALENIDESIGN SPEED AND A STAGE PRESSURE RATIO OF 1.45 First desig n Se c ond desi g n STATOR LOSSES (_S): Profile Fri c tion 0.019 0.023 Endwall Friction .026 .02 1 Secondary .009 006 In c iden c e .013 005 Total .067 055 An'stage .068 050 l P5 / Ps.5 1.4 7 l.42 ROTOR LOSSES (_R) : Profile Fri c tion 0.0 7 1 0.059 Hub Endwall Fri c tion .021 .012 Secondary .05 1 .02 7 Inciden c e .044 .032 Tip Clearance .04 7 .040 Total .234 .1 69 An'stage .128 .091 II P5.5 / P6.3 I.23 l.22 MANIFOLD LOSS, An'stage O.OlO 0.008 EXHAUST DUCT LOSS, An'stage 0.004 0.004 CALCULATEDOVERALL EFFI C IENCY 0.790 0.847 EXPERIMENTALOVERALL EFFICIENCY 0.807 0.843 Station 4.5 i s atthevolute inlet Inlet manifold I n st ru me n tatio n stat ion s 5 . 5 5 6.3 Fl o w
\
Fi gu re 1. - Cr o s s s ecti on o f redesi g n ed co m p r esso r- drivet urb i ne .
(a) I nlet volute and stator assembly.
Figure 2. - Test hardware.
( b)Rotor.
Figure2. - Concluded.
(VNcr)S = .278 ~5.S = 45.7 as.S = 68. 6 / (WNJcr)s.s =.456../ A.--- ......
~.3 = -35.3 ~6.3 =-61.4 (WNJ .857 cr)6.3 = (VN er)6.3 =. S08
(VN h.3 = .477
cr (a) 0% Streamline (hub).
(b) SO% Streamline.
(VN cr)S=.183 VNcr)s.s = .697 a6.3 = -22.4 ~6.3 = -64.0 cr (WNJ ) 6. 3 =.764 (VN )6.3 = .373 cr (c) 100% Streamline (lipl.
Figure 3. - Design velocity diagrams .
...
• 17 .& Q.> E '" ri~O- -"'u
... -
o.c ;:<1 ~ .14 ·u Q.> 0- .13 Hub Vl o .4 .6 Fraction of passage height Figure 4. - Radial work distribution.
1--- (a) Hub . (b)Mean . (c)Tip(r = 5 . 7 3 cm) Figure 5. - Blading p r o f il es and f lowpassag e s.
, 9 -- . 8 , 9 -- , 7 , 8 -- .6 . 7 -- .5 ,6-- .4 .5-- .3 •4-- . 2 .3- -- .1 I I I I I I I I I (a) T ip section. .2 -- . 9 -- .1 -- _•s- o I I I I I I I I I " . 7 -- . 9 -- - . 6 i _ (a}Tipsec t io n • _ 7 __ , • .3--- >., .2 I I I I I I I I I _ - (b)Mean sec t ion• > . 4 -- --g 1.C-- _ .3 -- i _
7 o
• 9- Exi t v e loci ty ra t io- / . S I I I I I I I I I (D) Mean s ec t ion• • 8-- Suc t ion s u r fa c e _ •9 -- Exit velocit y rat io -_ • 7-- .8 _ _' •6 . 7 _ _ . sur f ace . _ ,.,, _ / / • 5-- . 6 - -, , , _ Inletvel oc ity ratio • 4 -- Inl e tvelocit y P r e ssure . 5 -- / r,_ essur e /// _ ratio s u r f ace .3 _," . 4 -- ._.__ _ / surface I I I I I I I I I .3 I I I I I I I I I . I . 2 .3 .4 .5 .6 .7 . 8 . 9 1 . 0 . I .2 . 3 .4 . 5 . 6 . 7 . 8 . 9 I.
Fractiono f a xi alc ho rd F ra c tono f axialchord ( c )H ubse ctio n • (c) H u bse cti on • F i g u r e 6 . - Sta t or van e d es i g n sur f a c ev eloc ity di s trib u - F i g ure 7. - Rotor blade d e s i g n surface v e loci ty di s tribu- tion s , tions.
Pressur i zed air heater A ir Flowventur i \ P re s sure co n t r o l va l ve Airbrake Testturbine dynamometer Pressurized air _ Tolo w- pressure exhaustsystem (a)F a c i l i ty schematic.
(b)Turbinetest apparatus.
Figure8.
• Static pr essure o To t al p ressure e Totalpr e ssu r e and f lowa ng l e x Total temperature N a n i f o ld i nl e t St ato t inl e t (st a tion4.5) ( s t a tion 5) S tator ex it Ro tor e xi t (s t a ti on 5 . 5 ) (s tati on 6 . 3) Figure9. - F lowpathin s trument a tion,viewed lookingdown- s tream .
eL
50 _
_ B
u N -_ 4 0 2C (a)Flow angle.
o - A
._ _, 1.oo
_ _ N-_-, .
0 .2 .4 .6 .8 1 . 0 Frac t ion of passage heigh t (b) To t a l pressure ratio F i gure 10.-V a r i at i on of vo l ut e ex it flo w a ngle and t o t a l pressure wit h radia l posi t ion.
20 \ \,, Design \ \ \ lO K A \ _-1 ', \ T- \ -20 \ \ \ \ \ \ \ - 30-- \ \ \ \ \
.4o Hub I I I I Tip"t
0 .2 .4 .6 .8 l.O Fraction ofpassage height Figure11. - Radial variationin mass flowparameter at the voluteexit.
T e s t s pee d , percentof e q u iv a l ent de si gn s pee d /--D esi gn I 50 -x i • 32 -- 60-, \ , o 7 0_ _ . 30 -- _'_ L 80 L. 9 0 "E .28-- L.i00 L.110 26 I I I 1.2 1.4 1.6 1.8 2.0 2.2 2.4 Equivalent total pressure ratio , P'4. 5 1 P'6. 3 Figure 12. - Variation of equivalent massflow wi t h total pressure ratio andspeed• Testspeed, percent of e q uiv a lent design s p eed 7 - _ 60
, j 7 0
x _ o 80 6 _ "' _90 i _ "_- D e sign 1 I I I I I I 1. 2 1 .4 1. 6 1.8 2 . 0 2 .2 2. 4 E q uival e nt t o talpre s sure ratio , p _. 5 1 P ' 6 . 3 Figure 1 3. - Variation of t orquewi t hto t alpressu r e r a t ioands p e ed .
Testsp ee d , p e r c e n t o f e qu i v alent desi gn s pe ed I 110 50x ] O 3 Desi g n 40 7 0 .__
2 o I I I I I
1 .4 1 .6 1 . 8 2 . 0 2.2 2,4 E q uivalent total p ressure ratio , P '4. 5 / P '6. 3 Figure]4. - Tur b ine s pe ci fi c w o rk, ' _ _ _ . 8 0 _ .90 ra
_' . 7 o Hub I I I I r io ' 1
0 20 4 0 6 0 80 IOO P a ss a ge height, percent Figure 17. -Radia l variation intotal e f ficiency.
I A s-c a st blad e s .87I_ _ . 88}- _ C ) 0 O - _ . 86_::::::::_ , L......_ / - P o li s hed blades E .83 .84|_- I I I T e .81 I I I I __ ._ I 10 20 30 40 50x104 1.6 1.8 2.0 2.2 2 . 4 Reynolds number, m / prm Equivalent total pressureratio, P'4.5 1 P' 6 . 3 Figure18. - Effectof Reynolds numberon turbine Figure19 . - Turbinetotal efficiencyfor as-cast efficiency, andpolished rotor blades .
Q5 " U . 5 .3 U D e s ign(f i g . 3 ) _, V cr / 6.3 Ca l c u la t ed b yMER IDL f or loss ana l ysis (V / Vcr ) 5 .279 . 2 5 2 . 22. 5 a5 34. 7 48 . 7 30.5 (V N r)= 5 .9 7 3 . 8 1 6 . 7 86 (Wl_ r i'; , _ .6.51 .490 .3.58 ( U N cr)5_" 5 " .4 10 . 4 67 . .529 aS. 5 66. 4 6 9 . 7 6 9.8 _vS_ _ 51. I 4 9 . 5 37 . 5 r ) 6 3 . 4 5 9 .490 .357 (w1_.l_; 3 . 77.5 . 7 _ ._.5 ( U N cr_ 6 . " _ .445 ..5 0 8 .58 0 a6. 3 - 3 4 . 9 - 19 . 2 - 4 3. I IB6. 3 -6 2 .O -.5.5.4 - 7 2. 4 Hub M e a n Tip F i gure 20. - Compa ri son of des i gn ve l ocity d i agrams wit h those ca l cu l ated for l oss ana l ysis.
Calculated f r o m -5 0 -- experimental data _, ...... Analyticalresults "5 -40 -- from MERIDL ( re f . 10) . / / _ - 3 0 _ - _" - 2o " -_" /" o " Hub I I I I TiP l '10 ( a )Absolut e flow a ngle.
•.50 -- " 5 • _" _, .42 .__ .38 • 34 (b ) Absolu t e ve l oci t yr at i o .
i =' = __-°_ = _ _-_-6o _ - 7o _____ ., _ -d
o_ .5oi I ---i --- I I I (c)Rela t iveflowangle.
o ._ .68 O 20 40 60 80 100 P a ssage h ei gh t, p er ce n t ( d ) R e la ti ve ve l oci t y ratio.
F igur e 2 1 . - Com p a r iso n o f a n a lytical a nd e xp e r i m e n tal r o tor exit f l ow a ngle s a n d v e l oc ity r a t i o s.
1 .6 -- E xi t Ca l c ul a te d f or 1. 4 A v e l o ci ty l os sanalysi s D e s i gn( r e f . 2) , _- r| r a t i o L_l .o_ 2 1 .2 I , _---_ \. Su c tion I. O -- / _ > .8 - - " / .6 // // " _ __\\ surf a ce .4 -- - - - "" ; s ur e ,,_ " - _ .. / .......... s u r f ace _ / ' - - I n le t v e locity 0 . 2 . 4 .6 .8 1 . 0 0 .2 . 4 . 6 . 8 1 . 0 0 . 2 .4 .6 . 8 1.0 Fr a ct i on o f m e r i odi o n a l l eng t h (a)Hub (b)50 %S treamline (c) T ip ( a ) S t ator .
Figure22. - Compari s on of de s ign a ndrecalcula t ed stator a ndrotor b l a de surf a c e velocitydi s tribu- tions for t he fir s t turbine at equival e nt d esign s pee d a n dw ork.
Cal cu l ated f o r l oss anal ysis Des ign ( r ef . 6) 1 .2 -- E x it v e loc it y r a tio = -- S uction -- / _ o" \
• . 8 -
= _'5 _ 1. o -_ • 4 -- _" P ress ure _ ./ Z ' , - In l e t v eloci ty r a ti o su rf a c e V -- _ / / t I I I i I i I I I I I I I I .2 .4 .6 . 8 1.0 0 .2 . 4 .6 . 8 1.0 0 .2 .4 .6 .8 1.0 F ra c ti on of mer i od i onal l ength (a) Hub (b) 50%Streaml i ne (c ) T i p (b) Rotor.
F i gure 22.- Conc l uded.
C alc ul a te d fo r los s an al y s i s Des i g n (f i g . 6 ) Exit vel o city ratio -_ | o- 1" 2F \ 1 . 0I -- _ I_ Su cti o n " F_ .4 . . -" I _ .. / P r es s u r e j,_j,/ . 2 I-: > ' - _ inlet velocity r at io _ sur f ace _' I I I I I I I I I I I I 0 .2 . 4 .6 .8 l.O O . 2 .4 . 6 .8 1 . 0 0 .2 .4 .6 .8 1 . 0 F r act i on of meridionallength ( a )H ub ( b ) 50%S t r e aml i n e (c )Ti p (a ) Sta t or .
Figure2 3 . - Comparison of designandrecal c ulated s tatorandrotor bladesurfa c evelocitydistri- butionsfor the s e c on d turbine at equivalent design s peed an d work.
Ca l culated for l o s sanalys is D es i gn (f i g. 7) 1,2 _ m o- z- Inle t velocityra ti o 1.0- / -- Suction .--_>, / surfa c e _ _ , .
/ / /
.4 _ " I , /
Exit v elocityra t i o J surface - " " I I I I I 0 .2 .4 . 6 .8 1 . 0 O . 2 .4 .6 .8 1 .0 0 .2 . 4 . 6 . 8 1 .0 Fract i on of me ri d i onal l ength (a) Hub (b) 50"Io Streamline (c) T i p (b) Rotor.
Figure23. - Concluded.
_ . 70% Sp eed andP'4 . 5 / P' 6.3 = 1 . 45 Des i gn s peed a n d wo rk(f i g . 2 3 ) 1 .6-- -- -- .o 1 . 2 ' q'_-"-_ Suction , =
_ -- '
I !
1 . 0 ......... -J ] surface ,' -- - . ; r - Inlet / " / " J/ 7 ratio /, __.. [ -- t.
L f_xit !
.4- . " "_ / " .2 ; ::_i_ _ __ surface
r ; I I I I I I I I I I I I I I I
0 .2 .4 ,6 .8 1, 0 ,2 .4 .6 ,8 1 . 0 0 .2 .4 .6 .8 1 . 0 Frac ti on of m e r i odional l e ngth (a)Hub (b)50%Streamline (c) T ip (a)S t a t o r.
F igure2 4 . - Com p arison of t hedesign a ndr e calculated statorandro t or surfa c e v e loci t ydi st ribu- tions f o r the first turbine at 70 pe r ce ntequivalent designs p eed anda s t a ge p r e ssureratio of 1 . 45.
7 0% Sp e ed a n d P' 4. 5 / P'63 " 1.45 D esign speed and w ork(fig. 23) 1 . 2 -- -- Suc t ion -- o" Exi t v e locityrat i o _ surface sl "_, I . 0 -- . . .. . --.,, \ I ,, .8 - ; X - __ /. --..,__ - , ,, _- "\ I" "_ t / b .4 L_ Inl et v e loci t y r a t i o .Z-- surface I I I I I I I I I I I I I I I 0 ,2 .4 . 6 .8 1.0 0 .2 .4 . 6 . 8 1 . 0 . 2 . 4 .6 . 8 I . O F rac t ion of meriodiona l leng t h ( a ) Hub (b)Me an ( c)Ti p ( b)Rotor Fi g u r e 2 4 . - Co n c l ud e d .
70% S peed a ndP'4 5 1 P' 6 3 =1 . 45 D e si gn speed andwork(iig. 24) 1.2-- -- / , \\ o- Exi t velocityratio -,\ .--_>" _ s u rfac e j / '-'"1 _- -_ .6 - , " b .4 e ssur e ._ , _' .2' _ su rf ace , _ , s / I I I I I I I I I I } 0 .2 .4 .6 .8 1.0 . 2 .4 .6 .8 1.0 0 .2 .4 .6 .8 1.0 Fraction of meriodional length (a) Hub (b} 50_Io Streamline (c) Tip (a) Stater.
Figure 25.-Comparison of thedesign andrecalculated stator androtor surface velocity distribu- t ions for t he second turbine at70percent equivalent design speed anda stage pressure ratio of I. 45.
70%Speed andP'4.5 1 P'6. 3 _ i. 45 Designspeed andwork(fig . 24) o " 1.2 -- / - Inlet velocity -- I 1.0 -- // / ra t io >, / _,_, Suc t ion . , , , _ ,, " ' - - ""
L
/ "_ " _ surface_ / s _ _ / S/ '_' / .4 e . 2_ y ratio ! I I I I I I I I I .2 .4 . 6 .8 1.0 0 .2 .4 .6 .8 1.0 0 .2 .4 .6 .8 1.0 F r action of meriodional l ength (a)Hub (b)50%Streamline (c)Tip (b)Rotor.
Figure2.5.- Concluded.
1 . Report No. NASATM-83627 2. Government Accession No. 3 . Reciplent's C a t a log No .
U SAA V S COM TR - 84 -C-7 4. Title and Subtitle 5 . Report Date i Cold-Air Performance of Compressor-Drive Turbine of May 1984 Department of Energy Upgraded Automobile Gas Turbine Engi ne 6. Performing Organization Code III - Performance of Redesigned Turbine 778-32-01 7 . Author(s) 8 . Performing Organization Repo rt No .
Richard J. Roelke and Jeffrey E. Haas E-2044 10 . Work U n it No .
9 . P e rforming Organization Name and Address NASA Lewis Research Center and 11 . Contract orGrantNo .
U.S. Army Research and TechnologyLaboratories (AVSCOM) Cleveland, Ohio 44135 13 . Type ofRe p o rt a nd Perio d C overed 12. S po nsoring Agency Name and Address TechnicalMemorandum U.S. Department o f Energy 1 4 Spo nsorlng A ge n cy _ , e _ e -Rep o rt N O .
Office of Vehicle and Engine R&D Washington, D.C. 20545 DOE / NASA / 50194-39 15. Su pp lementary Notes Richard J. Roelke, NASALewis Research Center; Jeffrey E. Haas, Propulsion Laboratory, U.S. Army Research and Tec h nology Laboratories (AVSCOM). Prepared under Interagency Agreement DE-AIOI-80CS50194.
16. Abstract The aerodynamic performance of a redesigned compressor-drive turbine of the Department of Energy Upgraded Gas Turbine engine was determined in air at nominal inlet conditions of 325 K and 0.8 bar absolute. The turbine was designed with a lower flow factor, higher rotor reaction and a redesigned inlet volute compared to the first turbine. Comparisons between this turbine and the originally designed turbine showed about 2.3 percentage points improvement in efficiency at the same rotor tip clearance. Two versions of the same rotor were tested: an as-cast rotor and the same rotor with reduced surface roughness. The effect of reducing surface roughness was about one' half percentage point improvement in efficiency. Tests made to determine the effect of Reynolds number on the turbine performance showed no effect for the range from 100 000 to 500 000.
17. Key Words (Su g geste d by Autho r( s)) 18 . Distribution Statement Turbine engines Unclassified - unlimited Axial flow turbines STARCategory 02 , Automobile engines DOECategory UC-96 19 . Security Classlf . (of this report) 20 . Security Classif . (of this page) 21 . No, of pages 22 . Price " Unclassified Unclassified A02 *Fo r s ale by t he Na t ional Tec hnical i nfo rmation Service, Spring f i e ld. Vi r gin i a 22 1 61 SpaceAdministration eooK Washingt on , D.C.
2O546 \
Ae ron . u,, o. -- L -- .H I LI III
O f f ic _ l Busines s Penalty for Priv a te Ulm . $300 Po s tlge _ Feet_Piid Nitionll Aeronaut i c s a nd Splice Administrltion PO S TMASTER: If Undelivetahle (5 ¢c t i . _ n I S : _ Postal Manual) I _ ( J N u l R et urn I I I