CF6 Jet Engine Performance Improvement Program: High Pressure Turbine Aerodynamic Performance Improvement

j l (, NASA CR-159832

I

NI\SI\

I

National Aeronautics and Space Administration

I

CF6 JET ENGIN E PERFORMANCE

I

N o

IMPROVEMENT PROGRAM

1'1"\ o N I

I

c

HIGH PRESSURE TUR81NE AERO-

rX) z

DYNAM,IC PERFORMANCE IMPROVEMENT

I

by

I

I

W.A. Fasching I GENERAL ELECTRIC COMPANY

J

July 1980

I

Prepared For

I

Notional Aeronautics and Space Administration

I

I

I

NASA Lewis Research Center NAS3-20629

I

I

1. R.port No. 3. Re<:ipient's Cat.IOlI No.

I 2. GO\ItI'I'Imtnt AcelUion No.

NASA CR-159832 .t. l'lll. and Gubtltf. 6. Repon Olle CP'6 Jot Engine Performance 'Improvement Program July 1980 High Pro.sure Turbine Aerodynamic Performance Improvement 6. Performing Oroanlzatlon Code O. Performing O'llinization Report No 7. Author'" W.A. Fasching 10. Work Unit No.

G. ""'fMmlng Organlutlon Ntme tnd Mdt ...

General Electric Company 11. Contrlct or Grant No.

Aircraft Engine Group Ci~cinnati, Ohio 45215 NAS3-20629 13. Type of Report and Period Collored ~.,------------------,---------------------------------------~ 12. Spoll$Ol'lng Alltncv Name Ind Addrep Contract or Report NatiollBl Acronau,ticB and Space Administration 14. Spon$Of'lno Agency Code Washington, DC 20546 16. SUppl.mtrltMV Notes Project Manager, J.A. Ziemienski, Engine Component Project Engineer - R.J. Antl Improvement Office -- NASA-Lewis Research Center 21000 Brookpark Road, Cleveland, Ohio 44135 16. Abstract The improved single shank high pressure turbine design was evaluated in component tests consisting of performance, heat transfer and mechanical tests, and in core engine tests.

The instrumented core engine test verified the thermal, mechanical, and aeromechanical characteristics of the improved turbine design. An endurance test subjected the :I.!1lproved single shank turbine to 1000 simulated flight cycles, the equivalent of approximately 3000 hours of typical airline service.

Initial back-to-back engine tests demonstrated an improvement in cruise sfc of 1.3% and a reduction in exhaust gas temperature (EGT) of 10 C. An additional improvement of 0.3% in cruise sfc and 6° C in EGT is projected for long service engines.

17. Key Words (SlJggQItecI by Author(I') 1 B. Ol,trlbution Statement CF6~6 Unclassified - Unlimited HP Turbine Performance Improvement ~--------------------------r------------L--------------~----------~----~----~ 22. Price' 19. SecuritY o.nlf. (of thl' ... port) 20. Security Cllul'. (01 this pagel 21. No. 0\' Pages Unclassified Unclassified • For sale by the National Technical Informal ion Service, Springfield. Virginia 22161 NASA-C·l68 (Rev. 1()'7S) "

I

FOREWARD

~

Th~ work was performed by the CF6 Engineering Department of i~eneral Electric's Aircraft Engine Group, Aircraft Engine Engineering Division, Cincinnati, Ohio. The Design Managers were J.B. Sidenstick, F.C. Herzner /,1 and C.A. Freck. The program was conducted for the National Aeronautics and Space Administration, Lewis Research Center, Cleveland, Ohio, under Subtask 2.3 of the CF6 Jet Engine Performance Improvement Program, Contract Number ~AS3-20629. This report was prepared by W.A. Fasching, General Electric

U

Program Manager, with the assistance of F.C. Henner, C.A. Freck, and C. Feh1. The NASA Projp~t Engineer for this program was R.J. Ant!. The

'I

program was initiated it .. February 1978 and was completed in July 1979 •

. '

I ~

J

J

]

I

1.

f

J

iii

I

I

SU~ftofARY 1 1.0 2.0 INTRODUCTION 2 '3.0 ])ESCRIPTION OF IIIGH PRESSmm TURBINg Ar~ROOYNAf.lIC PERFORHANCg 4 IHPROVEHENT CONCEPT 3.1 Description of Performance Improvement Items 4 3.2 Pl~rrOl'mnncc 12 1.1 Ovarnll Dosign Appronch 12 3.4 Design Description 13 CONPONlm'l' PERFORNANCI~ TBSTS 17 4.0 4.1 Stage 1 Vane Cascade Test 17 11.2 Stago 2 Vmle Leading gdge Flow gvaluotion 27 5.0 COHPONENT mwr 'l'RANSFER Tr~S'l'S 32 5.1 Stuge 1 Vane TrailinR Edge Test 32

,.3

5.2 Stage 2 Vone Leuding Edge Test 5.3 Stago 2 Vone Trailing Edge Pin Fin Tast 49 CONPONEN'I' MEGHANICAL TBSTS 61 6.1 Illade Dynmnie and Stondy-Stotl' Strain Distribution 61 6.2 Blode Frequol1cy ond Amplitude as a Function of Damper 74 Force 6.3 Disk Rim Stress Distribution Test 79 7.0 INSTRUHEN'l'En ENGINg TEST 90 7.1 Test Setup 90 7.2 Instrumentation 90 7.3 Test Procedure 98 7.4 Test Results 98 B.O ENGINE ENDURANCE TEST 117 8.1 Test Setup 117 8.2 Test Procedure 117 8.3 Test Results 119 9.0 ECONOHIC ASSgSSMENT 141 SUMHARY OF RESULTS 143 10.0 APPENDIX A - Quality Assurance 1.50 APPENDIX B - References APPENDIX C - Symbols and Dp.finit:Lons iv

I

J

SUMMARY The purpose nf the lIigh Pressure 'J.'urbine Aerodynamic Performance Improv\1- ment Progral'll was t.o develop the technology and to demonstrate the technical feasibility of the improved single shank turbine design on a CF6-6 core engine.

J

Perfol.."mance. heat transfer. and mechanical component tests were conducted to evaluate the new turbine design. The thermal, mechanical, and aerothermal characteristics of the single shank turbine were verified in an instrumented

•

core engine test. The turbine hardware successfully completed 1000 simulated flight cycles of endurance testing, equivalent to approximately 3000 hours of typical airline service.

I ..

Initial sea level static back-to-back engine tests carried out in March 1980 demonstrated an improvement in sfc of 0.85% and e reduction in exhaust gas temperature (EGT) of 10° C for the high pressure turbine aerodynamic im- provements. A reduction in turbine inlet area will provide an additional sfc reduction of 0.72%. This amounts to a total improvement of 1.57% sfc at sea level which is equiv6lent to 1.06% at cruise. In addition, there is also an ."

improvement in shroud roundness which was assessed to amount to an additir',al improvement of 0.24% sfc at cruise. The total cruise sfc improvement for a new engine amounts, therefore, to 1.3%.

An additional reduction of 0.3% in cruise sfc and about 6° C in exhaust gas temperature is projected for long service engines due to the elimination of the mateface shank cooling air leakage associated with the original twin shank blades. This sums to an improvement of 1.6% in cruise sfc and 16° C ..

in EGT for long service engines • An economic assessment of the improvement was made for new engines on the Douglas DC-lO-lO aircraft. This assessment indicated that the CF6-6 ...

I improved high pressure turbine has a payback period of about 0.2 year. The acceptability of this concept is enhanced with the trend to higher fuel prices.

,t

[1

~

,

I

I

I

2.0 INTRODUCTION NationAL energy demand has outpaced domestic supply creati"g an increased U.S. dependence on foreign oil. This increased dependence was dramatized by the OPJ!;C oil embargo in the winter of 1973 to 1974. In addition, the embargo triggered a rapid rise in r,ht! cost of fuel which, along with the potential of further increases, brought about a changing economic circumstllnce with regard to the use of energy. These events, of course, were felt in the air transport industry oilS well as other forms of transpol"tation. As a result of these ex- periences, the Government, with the support of the aviation industry, has ini- tiated programs aimed at both the supply and demand aspects of the problem.

The supply problem is being investigated by looking at increasing fuel avail- ability from such sources as coal and oil shale. Efforts are curl"ently under- way to develop engine combustor and fuel systems that will accept fueh with broader specifications.

Reduced fuel consumption is the other approach to deal with the overall problem. A long-range eff~rt to reduce consumption is to evolve new technol- ogy which will permit development of a more energy efficient turbofan Ol" the use of a different propUlsive cycle such as a turboprop. Although studies have indicated la1:ge reductions in fuel usage are possible (e.g., 15 to 40%), the impact of this approach in any significant way would be 15 or more years away. In the short tenn, the only practical propulsion approach is to improve the fuel efficiency of current engines. Examination of this approach has indi- cated that a 5X fuel reduction goal starting in the 1980 to 1982 time period is feasible for the CF6 engine. This engine is, and will continue to be, a significant fuel user for the next 15 to 20 years.

Accordingly, NASA is sponsoring an overall Engine Component Improvement (ECl) Program to reduce the CF6 fuel consumption. This ECI program consists of two parts: EnGine Diagnostics and Performance Improvement. The Engine Diagnostics effort ic to provide information to identify the sources and causes of engine deterioration. The Performance Improvement effort is di- rected at developing engine performance improvement. and retention components for new production and retrofit engines. The initial Performance Improvement effort consisted of a Task 1 Feasibility Analysis which was conducted in coop- eration with the Boeing and Douglas aircraft companies and American and United airlines, and is reported in Reference 1. The study identified engine and component modifications which exhibited a fuel savings potential over current practice in CF6 engines; proyided a technical and economic assessment of the modifications, including the impact on airline acceptability and the probabil- ity of production introduction of the concepts by the 1980 to 1982 time period as well as their retrofit potential; and assessed the fuel savings potential for the DC-10-10, DC-10-30, and the B747-200 aircraft.

The high pressure turbine (1IPT) aerodynamic performance improvement con- cept was studied in the Performance Improvement Task 1 Feasibility Study and was selected for development and evaluation because of its fuel savings poten- tial and high payback for the above-mentioned aircraft. The results of the i Feasibility Analysis are r~ported in Reference 1. This report presents the relulta of the developaent work on the concept.

The high pressure turbine Qf the OF6-6 engine is essentially the ... e turbine that was developed for 1..le UpS. Air Force TF39 engine in the mid- 1960's. 'lbe core of the Tr39 engine sublequentlybecme the core engine for the OF6-6 commercial turbofan engine and the LM2500 marine and indultrial turboshaft engine. 'Ibis turbine, which has been in airline ~ervice Mince 1971, has Icc\laulated about 6.000,000 operating hours. During the en.uing period lince initial development. design improvements hive been limited to relatively modest changes directed at improved life and durability.

Design techniques and bl.ic technology have made considerable advance- ments since the original turbine was conceived. In 1976, it wal decided to initiate redesign studies directed at improved life and dqrability incorpo- rating many features employed in the second generation OF6-50 turbine. Much of thil initial work haa been sponsored by the U.S. Navy for the Uf2500 el7,gine which powers the Spruance chili deltroyers. Since significant mechanical de- eign chlngel were contemplated for this turbine, steps were aho taken to .- prove basic perfo~ance and reduce operational perfo~ance deterioration.

Perfo~lnce Unprovement studies indicated that up to 1.3% engine lpecific fuel conlumption (dc) improvement/retention could be achieved through reduced cooling air requirements, f10wpath anQ basic aerodynamic refinement., and re- duced prelllure 10s8es in the turbine rear frame.

,.

I I The objective of the NASA effort, which wa" initiated in early 1978, was ~ to develop the technology of the improved high pressure turbine in component and engine tests. The HPT Aerodynamic Performance Improvement Ooncept was a 17-month effort. Testing included component performance, heat transfer, and mechanical tests along with an instrumented engine and engine endurance tests.

r

~

J

L 3.0 DESCRIPTION OF HIGH PRESSURE 'rUR81ME AERODYNAMIC PERFORMANCE - - IMPROVEMENT CQ!f!EPT, The program wa. initiateo to provide aerodynamic and mechanical improve~ ments for the CF6-6/LK2500 high preuure tutbine. Objective. were aimed at significant reductionl in I~ecific fuel con.umption, increaled ruggedne .. , longer life, and reduced rleterioration in iervice. The ptogrAlll haa concen- tret.-:Jo on: • Improved aerodynamics • Adoption of the single shank blade concept utilized on the CF6-50 • Improved cooling deligns • Improved clearance control.

'l'he improved tllrbine has fewer, more rugged blades, and longer chord Stage 2 nozzle vanes, both of which provide better aerodynamic efficiency and reduced turbine exit losses.

3.1 DESCRIPTION OF PERFORMANCE IMPROVEMENTS ITEMS The new turbine has a number of features which enhance overall engine performance. These features are: • Modified Stage 2 vane aerodynamics • Reduced Stage 2 exit swirl • Better blade cooling effectiveness • More effective wheel space seals • Tighter tip clearance • No mating face shank leakage • Improved airfoil surface finish.

The following paragraphs describe these performance features: Stage 2 Vane Aerodynamics - The new Stage 2 vane has increased solidity.

This was accomplished by increasing the chord length as shown in Figure I while maintaining the same number of vanes. The inner and outer band shape was redefined. and the leading edge was reshaped to optimize the local pres- sure distribution.

Stage 2 Exit Swirl - The original design has a larger than desired exit swirl resulting in significant turbine midframe pressure losses. The new de- sign reduces the exit swirl by 9- which reduces the turbine midframe pressure loss. This increases the efficiency by 0.57% as shown in Figure 2.

,f I.

i i I.

l

I.

•

l

•

•

Increased Solidity

•

r New Band Shape I

~ •

I' New Leading Edge Shape

•

I Improved

\

-

---_ Original ,

•

,\

•

• I

I

Figure 1. Improved Stage 2 Vane Aerodynamics.

l

I

I

I,

I

en

~ 0.6 -J ~I

~ .'~

&. CF6-S0\

.;. CF6-6 Improvea1

tJ e ~ 0.4 tJ .,.; ....

.....

~ Q) e .,.; .0 ~ 0.2 E-<

"'" e

Q) ....

d > .,.; :: C" ::=l 0 _5° 0° 5° 1(11 15° 20° 2S· Turbine Exit Swirl Figure 2. Reduced Stage 2 Exit Swirl.

I

I Bl.de Coolinl Effectiven ••• - Hodern c •• tin, technoloay .llowed .ore

flexibility in th. a •• ian of the St.,. 1 .nd 2 turbine bl.d... Th. ori,in.l

bl.de •• tart with •• 0Ud cutin, which h chf!Jl drilled to fOnl thl! v.riou.

coolina p •••• ae.. By introducina preci.ion ~81t rQtftd bLrf~il., .i.LI.r to

J

the 0'6-50 •• on effectiv6 coolins c.n be .chi4ved. Fe.ture •• uch ., ,h.ped p .... ae ••• idewall turbulence proaotera. pin fin trail ina odie p .... 'e., .nd •• ooth turn.round. in the .erpeRtine p •••• ae •• re .chieved. A co.p.ri.on of the new detian to the oriain.l blade in Fi,ure 3 .how. tignific.nt lIetal t .....

perature r.duction for the ... e .. aunt of cooling flow. The b.tter coolinl effectivene •• could have been u.ed to reduce coolins flow while •• int.inina current te.peratur. levela. To incre .. e Staae 7 bl.de lif., it Wal decided • to lower the lIet.1 t.mperature in.tead. St.ae 2 •• intain. current metal te ....

per.ture •• nd reduce. coolina flow, r.elultina in • p.rfol'1l.nce improve.ent.

• Wh.el.p.ce Seal. - The new turbine incorpor.te. better wheel •• all a • • hown in ,ilure 4. By introducing .eah over .nd under the blade "angel wingl", .dequate cavity purginl is achieved with l'educed air.flow. Thele i.- • proved .eal. h.ve been proven on the OF6-50 turbine. An improved Stage 1 blade ret.iner which reducel flow leakage .cro •• lhe Stage 1 wheel rim il another C'6-50 feature which il being incorporated to help .chieve reductionl in cavity temper.ture and purge flow. The reduction in purle flow relult. in improved over.ll performance.

Tip Clearance - The UtW turbine incorporate. feature. in the Itator SYI- tem which improve cleaur.lce (:ontrol in several way. as indicated in FigUre 5.

8y increa.ing the cro'~-Iectional mass of the clearance control ring and by better ilolating it f~om the ~hroud cooling air, a slQwer thermal re'ponle i.

• achieved. This will re.ult in a better thermal match with the rotor blade tip during .teady-.tat~ and transient operation. Also, the .tiffer clearance con- trol ring. in its controlled envi~onment, will have a greater influence in re- .i.ting engine case distortion • ...

The shroud supports. with a stronger cro'l section. are less flexible.

This makes the supports deflect less inwardly due to the high radially inward .; pre.sure at taKeoff conditions. In turn, this reduces the tip rub at takeoff, resulting {n improved tip clearance at the cruise conditions.

"

I .- Hateface Sh~Leakage - A source of turbine deterioration has been elim- inated with the introduction of the single shank blade delign (lee Figure 6).

With the twin shank blades the two blades have a mating face which must be r brazed together to seal in the cooling air. After a fairly short time. the braze cracks and cooling air leakage results in performance deterioration.

Wi th the new design. the cooling air pauages are cast in with the cooling air

-1 being fed from underneath the dovetail. This entirely eliminates deterioration

dlle to leakage of cooling air.

Airfoil Finish - The blade and vane surface finish requirements for the CF6-50 are presently more stringent than those for the CF6-6 twin shank design.

Surface finish requirements for the single shank design have been brought in line with those used on the CF6-50 •

.1

]

I

'.pro,,"

II.r • • .... 1· C C·IJ· r, Or&.,.u I.,roved 6T Avera.a +).3' C (+6" ') Figure 3. Comparison of Original and Improved Blades Showing Improved Cooling.

I

I

I

I

I

I

I

I

J

f .1'

J

Lower Cavity Temperatures

I

I

Figure 4. Improved Wheel Space Seals.

I

I

I

I

I

I

....

o

:=1

'" (!)Improved Stator to Rotor Thermal Match ~ Basic Structure More Isolated from Gas Path ~ Stiffer Hoop • Roundness Improvement • Improved Cruise Match Figure 5. Tighter Tip Clearance.

~fateface Braze to Seal in Cooling Ail."

Airflow Airflow Original - Twin Shank Improved - Single Shank Figure 6. Single Shank Turbine Deterioration Reduction - No Mateface Shank Leakage.

3.2 PERFORMANCE The predicted I>orformance improvements of the individual items with re- spect to cooling air and turbine efficiency are summarized in Table I below!.

'r~b1e I. Estimated PeriomHmce Improvements.

AnT Improvement Item % A We % Stage 2 Vane Aero +0.14 Stage 2 Exit Swirl +0.57 2) Blade Cooltng (Stage -0.20 -+() • 03 Space Seals -0.55 +0.10 Wheel Tip Clearance +0.45 Air foil Finish +0.20 Li fe Features ('l'h ic ker +0.19 -0.40 blade edges, longer .~ chord, better cooling) Net 'fotal -0.56 +1.09 '1'he predicted net effect of these chllnges is estimated to reduce exhaust gas temperature (EGT) by 20.6° C (37· F) and sfc by approximately 1.3%.

111e above estimates apply to new engines. For long service engines, the single shank turbine should provide an additional 0.3% sfc decrease and 5.6 C (10· F) reducti.on in EGT due to the elimination of the mateface shank leakage discussed above. 'Jhis sums to a sfc benefit of 1.6% and an EGT reduction of 0 0 26.1 F (47 F) for long service engines.

3.3 OVERALL DESIGN APPROACH 11\e design appr:oach chosen for the sing 1e shank turbine is based upon the proven sliccessful CI:<'6-50 turbine design. 111is design incorporates relatively large size airfoils in reduced numbers relative to the CF6-6 for improved ruggedness and lower cost. In addition, numerous design features of the CF6-50 turbine were incorporated such as:

j

Lightweight blade dampers

•

• Over-under rotor stator cavity baffle.

• Intrarotor centrifugal pump for cooling air delivery to Stage 1 blades.

These features offer thp. demoller-rated reliability of the CF6-50 and aho offe r the potential for further r(it' for'"ance improvements by reducing the cavi- - ty cooling air requirements.

!

Another aspect of the chosen design approach was to util ize technology to further enhance performance and reliability. This is particularly evident in the approach taken to blade cooling where highly sophisticated cast coring was

J

used in both stages. This technology pe~itted the judicious reduction in cooling flows without the expected compromise$ in metal temperature levels.

In addition, significant amounts of weight were saved in spite of overall re- ductions in component stress levels for enhanced life.

Further improvements in overall turbine performance were achieved by utilization of an improved shroud support system as described in Section 3.1.

This support system was an integral part of the redesign and was executed in such a fashion 8S to achieve the desired clearance benefits along with improved cooling air screening capability and better containment capability. These fea- • ture. were incorporated without adversely affecting the good maintainability feature of the current design.

3.4 DESIGN DESCRIPTION 3.4.1 Rotor In keeping with the overall design approach, the rotor design was done in 8 fashion which blended the proven features of the CF6-50 with current design and manufacturing technology. This design approach utilized the following key design features: ...

Large blades/fewer in number

•

• E£fective blade damping

I

• State-of-the-art cooling technology • Conservative design thickness/stress • Proven rotor structural concept.

J The turbine blade design is cast-cored and has most of the cooling system

cast-in as an integral feature. Wall thicknesses are relatively thick having been sized to meet the requirements of the LH2500 marine environment. However,

J

1 13

the key dealgn feature it the ule of the lightweight damper sYltell which haa proven 80 lucces.ful on the CF6-50 on the core engine inltrumented te.t which meuured blade vibratory reapon.e. Ruggedne •• it aha enhanr.ed by increaaing airfoil lize while reducing numbers of bladea. Thia ha. the .ide benefit of reducing hardware cOlta. Blade luction side. are p"li.hed tc. achieve an addi- tional benefit in overall turbine efficiency.

The rotor structure bear. a atrong family resemblence to the current CF6-6D turbine and to the CF6-50 turbine. All materials are the same in these rotan as are the methods used to manufacture them. The CF6-S0 and the improved single shank rotors both have a apacer. between the Stage land 2 disks which have tadial vanes and a cover over these vanes. This is done to deliver cooling air to the Stage 1 blade at a higher pressure and thus improve coaling efficiency. Interfaces ~ith the compressor .haft and the aft bearing are unchanged. Only small modifications to the current midframe liner are re- quired to make the aingle shank turbine interchangeable with the original de- sign. Thi. modification involves machining approximately 3.5 mm (0.14 in.)

from the inner flowpath liner and adding a simple ring seal.

Design and analysis of the rotor structure have been ongoing throughout the program. This effort involved extensive computer analysis which investi- gated steady-state and transient effects. In addition, operational severity, commonly called "mission mix", has been included in the life analysis. In areas where it was appropriate. model and/or photoelastic testing has been conducted to verify analytical results. Results of this work will provide the basi. for Shop Manual life limits for all rotating turbine hardware 8S required by Government regulation.

In summary. the turbine rotor design is based on the successful features of the CF6-50 and blend ing in proven manufacturing technology. The design is interchangeable ~hen minor modifications are made to the turbine midframe.

3.4.2 Stator The improved single shank turbine stator design was carried out with the objective of improved life and incr.eased performance using components which used current state-of-the-art capabilities in both the design and fabrication of parts.

The design of the stator includes: • An improved cooling scheme for the Stage 1 vane • An aerodynamically redesigned Stage 2 vane • Impr:oved wheelspace cavity baffling • Stator structure designed for improved clearance control.

The Stage 1 vane has been designed using the same cast-cor~d cooling features normally used in blade design. Leading and trailing edge cooling

J

J

circuits have increa.ed area for reduced lo •• e. in coolant flow pre •• ure drop.

The trailing edge hal cut-in cooling dot. with turbulence prOliotel'l in the .lot wall.. A ca.t pin fin array precede. the trailing edge .lot., The vane wall thiekne •• ea have been inerea.ed and ca.t-in rib. added to prevent balloon-

J

ing of vane walls and to preei.ely locate impingement insert.. The iarle radiu. of the vane leading edge reduces local heat flux and allow. the u.e of i a greater concentration of film cooling hole ••

J

The Stage 2 vane has an increased chord and minor flowpath changes re- quired for better rotor-to-stator matching and reduced exit swirl. This vane

j

alao U.eI cast-in cool ing featuru, along with casting change. naade to provide a more rugged design • ..

~ Wheellpace baffles and intentage sealS are designed as fabricated sheet ~ metal cOlDponents. The intentage seals have honeyconab surface. for sealing at the rotor seal teeth and are designed so as to be positioned and retained

r

without the u.e of a bolted joint. These baffles and seals provide the im- , proved wheelapace cavity baffling through the use of a cylindrical surface

"

which maintains a cl08e clearance both radially and axially with the mating cylindrical Sllr faces on the blades.

j ~ The Stage 2 nozzle support aho carries the shrouds for the Stage land 2 blade tips. This support has been designed to provide stiffening rings over the Stage 1 shroud. These rings make the stator more resistant to both axi- symmetric and asymmetric distortions induced by engine operation. The Stage 2 nozzle support has an improved cooling design and provides a better rotor-to- stator thermal match. The support i8 machined from a ring for,ging in much ..

the 8ame fashion as the original support • ,.,.

Overall, the turbine 8tator design incorporates design features based on current manufacturing capability which are aimed at improved life and increased ...

performance. The design is in keeping with the CF6-6 lind CF6-50 design philosophy and closely resumbles both engines. l'he stator is interchangeable as an assembly when used with the twin shank turbine rotor.

J

J

3.4.3 Maintainability Maintainability features of the improved turbine are very similar to the original production design. The biggest differences are in two areas:

I

1. The cooling air screen for the Stage 1 vanes is attached to the Stage 2 nozzle support in the original production design. On the

J

single shank design, the screen is attached to the Stage 1 vane assembly. This allows the Stage 2 nozzle support to be removed without disturbing the screen assembly. This feature was built I in at the request of the airlines.

I

I

I

2. The turbine rotor Stage I disk l'im is bolted to the forward shaft and thermal shield by a bolt which is torquerl from the head into a

captive nut. This is 4S opposed to a. bolt/.tud arrangement in the

original design. The overall rotor 8uembly is not grouly changed by this feature.

No other significant changes in maintainability features have been made.

3.4.4 Safety Safety of the improved turbine is 88 good or better than the original design. Overall reliability should be significantly enhanced 8S previously discussed. Larger blades are used but the strength of the shroud support, and, therefore, cQntainment capability is signi ficant ly increased for the Stage 1 blades. From an overall point of view, conservative design features throughout the turbine will provide a safe design to at least currently achieved levl>ls.

""""'''-

[

I

4.0 COMPONENT PERFORMANCE TESTS

[

4.1 STAGE 1 VANE CASCADE TEST High prellure Stage 1 vane de. ign. are required to un a large qUf)nt, icy of cooling air to control the _tal temperature. in a .evere thermal envirola- Mento This coolinl air i. used to convection cool the inside of the airfoil

r

and is then expelled throulh ai:foil and band hole. to film cool the ga.-.ide metal surfaces. The addition of thi. cooling air to the gas .tream cau.es a turbine performance penalty in the form of a .flowpath total pressure loss.

1.

The objectives of th ie test program were to compare the aerodynamic efHciency of the original CF6-6 production Stage 1 vane and a new CF6-6 Stage 1 vane (.ingle shank) design with full cooling air. and to determine the aerodynamic efficiency effect of various row. of airfoil and band cooling injection.

r

4.1.1 Test Setup

r

The cascade test. were set up using eight vanes (four vane pairs) welded toge ther as showl. in Figure 7. A 8ch ema tic of the tes t setup is .hown in

J

Figure 8 while Figure 9 shows the. test cell with the vane cascade and supply systems. Engine conditions in terms of flowpath and cooling air pressure and temperature ratios were simulated. Shop air at its normal supply temperature of 43- C (110- F) was used for co01in8 air, requiring the gas str~am to be

r

heated to 354- C (670· F) to attain the proper engine gas stream to coolant temperature ratio. The cascade exhausted to atmospheric pressure, requiring 2 ) a gas stream pressure of 17.72 N/cm (25.17 Ib/in. to match the design

r

engine pressure ratio. To obtain the correct design cooling flow, the cool- 2 (26.17 ing air feed pressures were set at 18.04 N/cm Ib/in.2) at the outer 2) band and 18.13 N/cm (26.3 lb/in. at the inner band. The data were re-

r

corded on two X-Y plotters which recorded the circumferential traces of up- stream total pressure minus downstream total pressure and downstream total temperature.

r

4.1.2 Instrumentation

[

The instrumentation used in setting the above conditions consisted of a thermocouple and three total pressure probes installed ahead of the vane cas- cade and a thermocouple and static pressure probe in each of the inner and

r outer coolant feed plenums. The downstream data were taken with a total pres-

sure probe and thermocouple which were attached to an actuator system that provided control of the angle, radial position, and circumferential travel speed of the probe. The actuator systam is shown in Figure 10.

r

I

I

I

, klngi\1 ) .

F i g u t" 7 . V.ln' '. lS ' au . ' -' ) rwCl rd 1.0 ORIGINAL PAGE IS OF POOR QUALITY

~ ~ ~ t...... ~ .... _~ __ AI,:."

~.~ i~.

-- ..............

--.--..~ . ,

-

'lbenoeouple and Total Pressure Probe Tra.enable To~l Te8perature and Pressure ~robe Inlet Air (354° C) e

--

&; Em.ust

----

CoolinC Plenua J79 Burner Can Shop Air (43· C) Figure 8. Vane Cascade Test Schematic Draving • ....

~

~

c r [ • \II .' 1 () • I >t. II I' l't. III \. III I .t I I mOl IHI 1 P G ,

L ,TY

4.1.3 TOlt Procedure Telt COlldition. - A total of 19 te.t point. WAlfUn for the two vane d ....

JiSn' wh:' variation. in the cooling flow. and inlet condition. A .umnun:y of the tilt cClndition. for each point it given in Table II. Run. 2-11 were made with the orisinal producti~n vine. Run 2 wa. at de.ign condition. and cooling flow ~.ile Run 3 Was at a lower cABc.de pre.8ure ratio and Run 4 wa. at a higher ca.cade pral.ure ratio, De.ign cA.cade condition. were u.ed for Runa 6-9 with deerellina amount. of eoolin8 air. Ttlb wa. accolapU.hed by filling the coolina hol .. with a palte-like compound which hardenl into a ceullic- like material and then landing to a .mooth contour. The cooling hoi .. were proares.iv.J~ lealed as follow.: band hole. (6), .uction .ide gill hole.

(7), leadins edge and pre •• ure aide gill holos (8), and trailing edge holes (9). With all the cooling hole. lea led, Run 10 Wal run at a low calcade pI'''- lure ratio and Run 11 wa. run at a high calcade prea,ur.e ratio.

The ca.cade hardware was then changed to the improved design, and the teat point. were repeated. Runs 12-14 had full cooling air with Run 12 at de.ign ca.cade conditions. Run 13 at a lower ca.cade preslure ratio, and Run 14 at a higher cascade pressure ratio. Runl 15-19 were run at cucade design condition. progreslively sealing these holes: band holes (15), auetion side gill hole. (16), leading edg~ and forward pressure .ide gill hole. (17), aft pre.sure side gill holes (18), and trailing edge hole~ (19), With all cool- ing holea sealed, Run 20 was run at low cascade pressure ratio and Run 21 was run at high e •• cade ~ressure ratio.

For each test point, a total of 15 circumferential transverses was made at the.e radial locations: 2,,7, 10, 15, 20, 30, 40, 50, 60, 70,80,85, 90, 93, and 98% span8.

Data AnalX,sh - Efficiency for a vane cascade is defined as:

n • actual exit Kinetic enersx

ideal exit Kinetic energy For the particular case where PT coolant ~ PT mainstream' the effi- ciency reduces to the classical casca~e efficiti~cy , y - y

«S2)

_ PT2 11 • 1 y - y

CS2)

PTO

Where: PTO = inlet total pressure

PT2 = exit average total pressure

PS2 • exit static pressure p' .,._..., ~ " ..... ,.,( ~ ..

.i- ~.-.,. .. ~

........ .... ...-.+

t--+ .",,- ... ~

.-:::"~

~ ~ - ----,. ........... ~

-

Table II. Summary of Vane Cascade Test Conditions.

eal TUh 354- C; T • 43 C Cooliu 1Io1H PIGllli.

r pc. outer pc. inner 'ru •• c .. SuctlOil SUe l.ucIiB& ldae ... Af"tPra8Ure 2) lun Delip (N/c.

(N/cal ) (N/ca2) .. nil Gill PrellUre Side cill Si.e Cill Traili .. U • 2 Orilinal 17.72 18.04 18.13 X l X X

--

3 16.20 )[ 16.50 16.58 X X X

-

20.27 20.64 20.75 X X )[ X

-

17.72 18.04 11.13 X X I.

-

7 )[ )[ 17.72 18.04 11.13

~ -

11.72 18.04 18.13 I.

-

9 17.72 18.0;'" 18.13

-

16.20 16.50 16.58

-

20.27 10.64 20.75

-

12 bproved )[ )[ 17.72 18.04 18.13 X X X I

~

13 )[ )[ 16.20 1~.50 16.58 X X X 20.17 20.64 20.75 X X

. X .X X

)[ 15 17.72 18.04 18.13 X X X )[ 16 )[ 17.72 18.04 18.13 X ..

I 17 )[ 17.71 18.04 18.13 X 17.71 18.04 18.13 X 19 17.72 18.04 18.13 ~n 16.20 16.50 16.58 .

20.27 20.64 20.75

J

t>:I Col)

!« ., '1 nt'.... . t ""'''-.-.#' -

l'lIe computer program "CADREX" was used to reduce the tes t data. This pro~ram uses a constant static pressure mixing assumption to integrate the exit pressure and temperature wake surveys.

4.1.4 'fest Results l'ypical pressure and temperature wake traverses are shown in Figure 11 for the original production airfoil and Figure 12 for the improved Vane de- sign. The pressure loss peaks correspond to the vane trailing edges. This total pressure .oss is caused by airfoil boundary layer drag. trailing edge blockage. and cooling air injection. A total of 285 wake traverses was made for this test program.

A summary of the results of all tests is shown on Table ttl. The single value used to characterize each configuration was obtained by mass flow aver- aging the resl~lts at each of the radial traverse locations.

The fully cooled improved single shank vane has an aerodynamic efficiency equal to the fully cooled original production vane. Elimination of cooling flow caused an increase in efficiency in all cases. except when the trailing edge holes were plugged. Apparently, the roughness of the aerodynamic contour where the trailing edge s lots were plugged caused more loss than was ga:f.ned with elimina- tion of the trailing edge flow.

Table Ill. Summary of Cascade Efficiency Results.

Cooling Holes Flowing Aft Leading Edge Vane Suction and Pressure Pressure Trailing Percentage Design Band Side Gill Side Gill Side Gill Edge Points Original X X 0 (Base) X X

---

Original X X +0.31

X ---

Original X X +0.70

---

Original X +0.91 original +0.80 Improved X X X X X +0.04 Improved X X X X +0.55 Improv<"!d X X X +0.95 Improved X +1. 26 X Improved +0.96 X

I

,- ....

24.

l

Clrouaf.r.ntl.l Plltlne., ."C~ ••

II :t 0 :t , I ..

•

~

...

.

--_ ..... _- .. , .. _- :t

a

- :t

~ ~ -

-- -

•

~ ,.: • . ' ..

a

~ -

I ..

_\ l _ • ., - .~- -.. ... '**' ....

-'

l' -- ..'"

--

--

It ~ \ 1 ....

..

..

II ...

..

; 'il ..

~ J

f-\ -,...... .• ~.;!,., ,._.-

....

~ Nl 7 II \I 10 tI ,; ..

II ..

• ..

t

..

S

ao

;J .. \I 10

1\ T

"

Figure 11. CF6-6 Stage 1 Vone - Original Production Design, Full Coolir,g Air I 50% Sptll1, Pl'OSSUl'e ond Tempel'otlu'O Travel'SO.

:I :to ..

.1

_M

'" :I II

1--- .- - :t

"e ~-

-

...

~ e.

.

•

..

-:: II ...

:a _ .. , -.-~.

:t f-

-- - --

--~ .. -----

~

I E

.. ~

.

• ..,

J __

a II

...

0 0 t ...

-"

-I

-

-1 o 7 II 10 Clrcuafe"Dnthl IHllhnao, inch •• .. uo~0 ______ ~ _____ -r __ -Tl __ ~ ____ ~ _______ ~3 _____ ~ ______ ~ __ ~3r-~ ____ ~ ____ --, , , I '700 3110 f-'

600 g

300 ~-.-+-.-_I-~ --f---

---

...

(I ...

• 1100 I t----~~ ..

:1110 . ~--- i ~ ..

il

• g

..

300 1----+------1 400 ,.

!

..

..

M :----+---i-~--+-----t_'--_t 300 I"l f-- 0 a :t

.. 6 7 I

9 10 FigUl'e 12. CF6-6 Stnge 1 Vane - Improved Design, Full Cooling Ail', 50% Span, Pressure and Temperature Traverse.

J

, 4.2 STAGE 2 VANE LEADING EDGE FLOW EVALUATION I I· The objective of this program was to obtain and compare flow character- istics between the vane shell and impingement:lnsert, and radially within the insert for the original production and improved Stage 2 vane designs. This was accomplished by measuring the static pressure distribution between the vane shell and the insert, and radially within the insert cavity with con- trolled supply and discharge pressures.

4.2.1 Teat Setup The test models were actual vane segments and inserts having pressure taps attached to both the pressure and suction side walls. l~e basic differ- ence between the two designs as related to this test was the enlarged flow entrance area into the improved insert. The objective of this increase in inlet flow area was to reduce the static pressure drop through the inlet by lowering the velocity of the air in this region.

The test configuration shown in Figure 13 COliS is ted of the vane to be tes ted having the insert of the vane at tached to a supply plenum at the ou ter band and another plenum attached to the inner band disr.:harge. The inner band plenum WliS designed to allow .lack pressUt"ing of the fluid circuit to simulate engine pressure ratios.

4.2.2 Instrumentation In order to deter.mine the vane flow distribution between the vane shell and the impingement insert, pressure taps were installed through the shell on both the pres sure and suction sides. Pres sure tap locations used for both the original and improved designs are Rhown in Figure 14. The locations were selected to allow measuring the pr~ssure between the insert holes to avoid sensing total pressure contributions from an iffipinging jet. A total of 58 pressure taps was installed in the original design vane and 63 pressure taps were installed in the improved vane.

Pressures were recorded using both mercury and water-filled manometers.

Airflows were measured with a rotometer and orifice measuring devices.

The radial pressure distribution inside the inset was determined by tra- versing the insert cavity with a static pressure probe at five selected radial positions. These radial positions are shown by the dashed lines in Figure 14.

4.2.3 Test Procedure The test procedure consisted of supplying air to the model and then back- pressuring the plenum attached to the inner band to the desired pressure ratio. The vane trailing edge holes discharged to ambient pressure. The ...

. . . . .

.

. .

· .

. . .

• . ..

. .

. .. .

. .

. .

..

· .. .

..

. . .

. ..

. . .. · ..

.

H c:: o ...

OJ Vl I..

o , Vl .., CIl f- e" .....

O RIGIN L OF POOR

r

I Inner Band

r

•

•

• II

•

•

•

til) II

•

• til)

•

I

•

is

I

is

--- - --- til)

•

•

• • til)

• • • •

•

• •

~ ~ I

r ...... ----

....

• •

• .,..

• •

•

~ • • •

• •

III ...

~

•

~

•

•

•

•

• •

• • •

•

I ---t---

,

• •

• •

• •

• •

• •

•

I.

•

Pressure Side Suction Side ~.

(a) Original Design il

il

1.

•

•

•

•

•

•

•

•

•

• Q)

•

•

btl

r •

-a

•

•

'" •

• fI,:I

• •

•

•

I· •

•

til)

•

•

---\--- . •

•

• Flow .~

• •

•

• •

....

r

•

'1"4

•

•

• ."

• •

•

• • ...

•

•

•

~

•

. ---~--.

•

•

•

•

•

• •

f

•

• •

I

• ---

Suction Side

r

~ 1

Outer Band Improved Design (b)

r

_____ Denotes Probe Locations • Denotes Pressure Tap J.ocation

r

....

Figure 14. Schematics Showing Surface Pressure Tap Locations on Original and Improved Design Vanes.

I

I

pre .. ure ratio. from the .upply plenUII to the inner band plen,* and to .-bient at the trailing edge were .et to match engine de.ian point condition.. Once a te.t point was e.tabli.hed. pressure and airflow reading. were recorded.

Each prellure tap was monitored individually using either water or mercury .ana.eters as the pre.sure level. dictated. Coolant flow rates were measured at the inlet plenum (outer band) and at the inner band exit. The trailing edge flow was determined to be the difference of the measured flows. Pressure measurements at the inlet and exit plenums were recorded as well a. at the pres.ure tap location.. Five radial location. in the insert cavity were Ilea- .ured u.ing the pre •• ure probe.

4.2.4 Telt Ruultl At duign point conditions, pressure drop measurement through the insert impingement holes showed a 20% variation from the average for the original de- lign. Similar data for the improved design resulted in a 6% variation from the average. These percentages are ind icative of local flow ':,:"riations for the respective deaig~ls. Spanwise pressure measurements within the original design inserts showed a 4%, loss at the insert inlet (outer band) with an average lQs8 in the insert of U of the inlet pressure. Correspondingly, the improved de- sign had a 1% pressure loss through the insert inlet (outer band) and no de- tectable loss radially within the insert. Enlargement of the improved design insert inlet area has reduced the static pressure loss resulting in a reduction of the radial pressure gradient within the insert. This has eliminated the low impingement lip cond ition at the outer band.

The pressure and flow data revealed that when extrapolated to engine con- ditions, the flow split between the trailing edge holes and the inner band exit is different than the calculated design intent as illustrated in Figure 15. The difference is due to the trailing edge holes being undersized in both the current and improved vanes. The trailing edge holes would have to be en- larged in order to achieve the design flow split.

I

I

Improved Original

J Engine De.ign Point Conditions

1.85% 2.05%

J

r

........ 1.15%

r

0.60% 0.90%

r

Bench Test Results

.-

2.05% 1.86% I , ~ ..

. -

--4 ....... 0.75% 0.71% .,.

1.34% 1.11% Note: All Flows are Percent of Compressor Inlet Flow Figure 15. Oesign Point Comparison of Flow Splits at Engine Design Point and Bench Test Conditions for Original and Improved Stage 2 Vanes.

:n -; 5.0 COMPONENT HEAT TRANSFER TESTS 5.1 STAGE 1 VANE tRAILING EDGE TEST The objective of thi. te.t wa. to experiaentally determine the heat tran.fer and pres.ure 10.s characteristics of the Stage 1 vane trailing edge flow pas.age.. In particular, the effect of rib-type turbulence promotor.

wa. inve.tigated. Compari.on. were made between a .. ooth duct and three turbulence promoting configuration •• 5.1.1 Test Setup A lOX scale model simulating a single trailing edge flow pa •• age of the improved turbine de.ign .hown in Figure 16 was utilized. Four configurations of pas.ag~ geometry and rib type were investigated and are shown in Figure 17.

The fir.t configuration was a smooth duct to e.tabli.h baseline heat transfer coefficient levels. The second configuration had 2.5 mm square rib. directly opposite on opposing endwalls. The third configuration had the 2.5 mm square ribs, a. de.cribed previouely, with the addition of 2.S ram radiu8ed ribs mounted on the remaining partition walls spaced between the square ribs. The fourth configuration was the de.ign configuration which had 2.5 ram square ribs on the two Qpposing endwalls; however, the ribs were not directly opposite one another but were arranged in a staggered array. The test model is shown .chematically in Figure 1&.

Photographs ~f the te.t configurations and the assembled model are pre- .ented in Figures 19 through 22. Configuration 2 is not shown because it is the same a. Configuration 3 without the partition wall ribs.

5.1. 2 In8trument.ation The turbulence promotor section (Figure 18) contained the data heaters.

Eight individual data heaters were located on the four sides of the passage and split the passage into two axial segments. Data heaters are heated copper plates whose heat loss due to assigned mass flow rates was used to determine corresponding heat transfer coefficients. The endwall ribs were simulated as 2.5 mm square cross sections. Guard heaters to minimize axial heat losses were located on each end of the data heater .ection. At the exit end of the turbulence promotor section, the passage converged and discharged to ambient pressure.

The data heaters were fabricated from copper plate. Electrical heaters and two thermocouples were silver-soldered to the plate backs. This allowed soft soldering of ribs to the test ·surfaces of the plates. In assembly, asbestos insulation was placed on the heater end8 to minimize contact between heaters. The heaters were bonded to the Textolite housing sides with RTV adhe.ive.

*' .,~. ., ... ~.

-<., -. -, .. , • ---. ................

- - ......... ~ ~ ~ 4._.

t

1: .

• Turbu1ated Test Section • I Pr:essure Side Airflow Insert \ Suction Side \ Square Endwall Ribs Figure 16. Pitchline - Stage 1 Vane Trailing Edge Cross Section.

w w au • " Configuration 1 Smooth Configuration 2 2 Endwalls Promoted Configuration :3 2 Endwalls and 2 Partition Walls Promoted Configuration 4 2 Endwalls Promoted Staggered Figure 17. Test Configuration Geometries.

j' !

-

\'

-

....... --- - -

Passage 16.2 .. x 14.7 ..

Square Endwall llibs 2.5 rom ,Discharge Passage 6.4 mm x 14.7 ..

( Insert • I' t To Ambient Discharge I

•

Ai.r - P .

Partition tat1.C s Ribs Data 2.5 1IID Heaters

1___ __ Radius

r Section 1 .. I .. Section 2 • I

Turbulence Promoter Sections Figure 18. Stage.1 Vane Trailing iEdge Heat Transfer !Iode1..

c:..J CIt

-

c c Vl

-

c

-

c: ....

:: ...

t.&.

c: o

-

-

M c: o ....

...

c: o u o I ....

t.t.

w • "' ;0 ;::Vi Figure 21. Contiguration 4, Endwalls Promoted - Staggered Ribs.

..

...

, , • t .0 ."

'r' ~

l

5.1.3 T~8t Procedure I)owtlr input to the tcst plates waH controlled by individual vl1rhbh transfonuet:'fl. 'l'he inputs to each test I)late were measured by use of II switching circuit which connected all lIl1meter and voltmetet' into each of the power supply circuita, '1'0 electrical circuit is shown in Figure 23 where one variable trans.former circuit is illustrated. A correction was made for lead wire lind ammeter resistances when the measuring systenl was in the cir- cuit. nle con8tant wall temperature test condition WliS obtained by use of a control panel. nlis permitted individual control of power input to each of the 10 heateu.

UlE! flow rate to the 8YRtcm Wlla measured by II three-bank assembly of roto- meters also shown in Figure 23. The flow capability could be increased bv operating above atmospheric pressure at the rotometer by throttling between rotameter and test section. Heat loss from the systcllI was determined by fil Ling the interior of the duct with insulation and heating to steady- state conditions at severlll heater plate temperature Levels to define he4t Loss from each heater in the test SIl'\C tion. 'rhe heat loss for the bulk of the test points was less than 10%. CalibrlJtion curvea for the electrical measuring equipment and the flow measuring equipment were obtained before and after the teat to make aClIrate corrections for test reading •• 'DIe test o~ration consisted of setting a flow level llnd then adjusting each of the 10 individual variable trunaformers to obtain a uniform operating plate temperature of 96· C (205· F). Aft(~r assuring that equilibrium had been achieved, the data for .11L L6 thernlocouples in the teot sec tion two supply air thermOcOul)1es t and the 10 ind iv idual voltage and ammeter readings were recorded. In addition, plenum pressures were recorded so that friction pressure drop could be determined.

5.1.4 Test Results Heat Transfer - The data were correlated using the actual dUct inlet dimenaions in the Nusselt lind Rt.'ynolds nllmber calculati.ons. Results of the Section 2 data for the four configurations tested nre !llunmarized in Table IV. Section 1 datl\ are influenced by entrance effec ts and were, therefore, presented for comparison only.

l1\e two-wall promoted data for both the staggered lind nonstllggered ribs are abdut equal with the staggered being less than 10% lower. For staggered ribs on two opposing walla t the average pl'omotion is about 2 times that in the same duct with no ribs. The presence of the ribs 011 cwo adjacent walls promoted heat transfer 011 the two altern~'te smooth walls by a factol" of 1.4.

The data for the selected design, two walls promoted staggered (Configlll'ation 4) t are presented in Figure 24.

In a separate thermul analysis of the vnne, it was cnlculutec1 thut the selection of the two walls promoted in the staggered IIIlltriX rather than the four walls promoted as tIle design configuration results in temperature increases of only 5.6 C in the prollloted region and reduces the trailing edge temperature by 16.7° C.

J

-~--...,.. ....... --

--~~--~~~--~

AC Power

I

Supply

I

lIeater

j

I

Constant Primary No. 1 Voltage Voltage

I

Voltage Transformer Adjustment Adjustment i

I

i ,; ,

I

J

I

Heater J No. 1

I

Resistance Electrical Circuit Load

I

I

I

I

Ammeter

I To Other

Ji-- .... ~---------""----- Heasurement , Circuits Voltmeter I ii 1 ___ _ " I H

--

-

Rotometer Pressure " , Calibrated Rotameter Bank ~ ~ Airflow Supply ~ Test Section Measuring System ' ....

,l

J

Air~~ Supply PreSBure Regulator

.1

Figure 23. Schematic of Electrical and Airflow Supply Measuring Systems.

I

I

" I

I

p Iwct.on 1 'quar. alba o bcHon 1 a.oo,IIw.U 10' Hyd iDll • 4A/Wr • I ••• a ea

•

Are ••••• • 1.111 ea

•

8/

;, II'" I

./~

...

.:

)1//

I

~-/--

..

...

•

,

I

,~

I.

TUrbulent Pull), Developed ;, Duct par.

lIu • ° o:n ... iprO.4CT IT )0.'

D' D b t ;, lOS o! 4 10 2 2 3

• • • •

WD hynold. lkI.ber, a· • A;i D Ry~ Di •• 1.0'2 c.

Ar, ••• X H • 2.012 c.

•

II" 2

•

f

10 ~~~~~~--~--~~~~~~--~~ ...

I ;, •• 10 2 4 •• 10 Z I ",.old. 1111.,. "D • :: Figure 24. Stage 1 Vane Turbulence Promoter - Two \o/alls Promoted Staggered.

'12

-

:(

J

Table IV. Heat Transfer Test Results.

r J Turbulence No. Configuration Promotion l All Wall. 1.0 (Ba.e) Smooth 2 Two Walls Promoted Smooth Partition Wall. 1.67

•

Ribbed Endwalls 2.79

•

3 Fo~r Walls Promoted Ribbed Partition Walls 2.47

•

Ribbed Endwalls 3.21 '1i

•

4 Two Walls Promoted Staggered Smooth Partition Walls 1. 70

•

Ribbed Endwalls 2.44

•

Pressure Drop - Friction factors were determined from the pressure data.

These calculations were based on maximum velocity and equivalent diameter dictated by the individual geometries. The data were correlated from experi- ments having the inlet and exit geometries removed. When the inlet and exit geometries were in place, realistic friction factors could not be calculated due to the indeterminable geometrical influence on the pressure measurements.

A comparison of the friction factor data correlations is shown in Figure 25. The friction factor of the four-wall promoted data is significantly

r

higher (by approximately a factor of 2) than the two-wall nonstaggered data.

The two-wall staggered rib data are at approximately the same level as the two-wall nonstaggered data. Smooth wall friction shows a descending trend as Reynolds number increases and is higher than the predicted level shown by the dashed line.

'I

Friction factor data for the smooth passage of this model configuration were 2.1 times that for a straight smooth duct without entrance effects.

This difference is attributed to entrance effects in the model data. The effects of two or four walls having turbulence promotor ribs increased the

I

friction factor levels respectively to about 7.9 and 17.5 times that for a straight smooth duct.

I

5.2 STAGE 2 VANE LEADING EDGE TEST The objective of this test was to investigate the Stage 2 vane leading

I

edge region impingement heat transfer characteristics. Scale models of both the original production and the improved vane inserts were used.

I

I

o Smoothwall 04 Walls Pl'omoted

o 2 Wnlls Promoted ~ 2 Walls Promoted StJlggered

o,zoo

0 0 0,100 0

-

• 0.080

·

• ..

0.060 IH ~

.. 0 0

4J

104 0 0 6 6 6 ~

0.040 .., 6

u til ra..

I- c .,..

.., u ~ 0.020 ....

ra..

0.010 o -

'"

0.008 Sntoothwall

· - _ ["_dieted

• -- ......

...............

0.006 • ......

...... ~

• 0.004 • • I • • ••

• • • • • • •

• •

4 5 3 4 6 8 10 2 4 6 8 10 2 3 WD Reynolds Number Re =- • D Aj..t (A is Bused On Duct Entrance Area) Figure 25. Comparison of Friction Factor Results.

f

5.2.1 :r.:..est Setup

J

The model scalinl criteria used for this test allowed testing of both the original and improved insert configurations by using the new vane inner contour.

A geometric camparilon re.ulted in a 7X scale model of the improved insert and vane inner contour as the base. Using the Unproved vane inner contour to evalu- ate the original insert resulted in a 11.5X scale model. Figure 26 shows the superpo.itioning of the original and improved vane scale contours noting the region of intere.t.

CrolS-sectional profiles of the two model configurations tested lire shown ..

in Figure. 27 and 28. Thele figures show the insert position relative to the heated inner vane contour and impingement flow axes for Rows 1, 2, and 17.

The improved insert differed from the original insert in that it was closer to the vane inner surface. The design intention of the Unproved insert is to pa.s the same coolant flow through the three rows of leading edge holes (1, 2, and 17) while using a larger number of ~maller diameter metering holes in the Itagnation region (Row 1).

5.2.2 Instrumentation Hodel instrumentation consisted of plenum pressure gages on both inserts.

The simulated inner vane contour walls were composed of copper heater plates.

Each heater plate was instrumented wi th two thermocouples to monitor the plate temperature.

5.2.3 Test Procedure j The test conditions were determined by calculating the Reynolds numbers for both inserts at takeoff engine operations conditions. The diameter of the leading edge impingement holes was used as the characteristic dimension 1n the Reynolds equation. ~fodel airflow, pressure, temperature, and power input were measured. The Mach numbers were not matched, since in the compressible regime the effects were negligible.

I

The test points were selected by using the Reynolds number criterion: pV Dl ReDl •

I

where

I

Dl K Diameter of the Row 1 impingement holes ~ • Viscosity evaluated at the coolant plenum temperature

I

W • Coolant mass flow

I

I

A = Total area of impingement holes present

;1

I

~~~~~~~ - -- - -I

I

I

I

I

I

I

c ....

I

CIl I I c:: I

_____ ..J

"'0 > o ...

0.

....

.....

~ "" C ...

....

c:: ~ ....

...

o c:: o ~ CO C QI ....

0:: C o ....

....

....

,....

CIl ,.., o QI 0.

.-i CI 1-1 .-i CJ 0 111 0.

til ::l / J') Ul >< >: l/') N I"- ......

'C > 0-

""

S

Ii"

Perforated Plate Original Insert Heaters Hole Axis Denotes Guard Heater G Improved Vane J) Denotes Data Heater Inner Contour Heater Width Used L in Defining Reynolds and Nusselt Numbers Figure 27. Original Stage 2 Vane Insert Cross Section.

Perforated Plate Iaproved In.ert Heatel'!J Hole Ad.

Denotes Guard Heater Denotes Data Heater L • Heater Width U.ed in Defining Reynolds and Nusselt Numbers Figure 28. Improved Stage 2 Vane Insert Hodel Cross Section.

I

I

Te.t point. were .et by fir.t .etting the flow rate to the model. Elec- trical input wa. then adju.ted to each heater plate to give a con.tant temPftra·

r

ture 8urface. Time WI' allowed for tran.ient re.pon.e of the .y.tem and for adju.tin. the power .upplie. to provide a conltant temperature: then airflow.

temperature, and electrical input. were recorded. To determine backBide heat

j

lo •• e., an additional telt wa. run with the in.ert removed and the heater plate. heavily inBulated on the front .urface.

t

5.2.4 Te.t Relultl A canparhon of test reault" at engine conditil)ns for the odginal and im- proved delign. is Ihown in Figure 29. The characteristic dimenaion used in defining Nu •• elt number and Reynold. number in this figure is the equivalent heated width for each model. This is necessitated by the fact that the model.

1 had different scale factor ••

For a given Reynolds number based on the width of the impinged .urfdce, the average Nu.selt number over that surface is about 20% higher for the im-

t

proved new in.ert configuration than for the original insert configuration.

j Compared at the respective deeign Reynolds numbers. the improved deBign ~ ,.

vane in.ert show. a 10% increase in Nusselt number in the leading edge region.

J

5.3 STAGE 2 VANE TRAILING EDGE PIN FIN TEST

-

Pin fin turbulence generators are commonly used to augment convective cooling in high heat transfer applications. The objective of this program

i

was to determine the benefits of pin fins in the unimpinged region ahead of the trailing edge holes of the Stage 2 vane trailing edge. The testing t utilized a lOX acale model of the vane trailing edge and configurations

i

having no pins, pins of low conductivity (Textolite), and pins of high con- ductivity (copper).

;

1-

5.3.1 Test Setup A cross-sectional view of the vane area modeled is shown in Figure 30.

r

Based on this configuration, a lOX engine size model was designed and fabri- cated which simulated a radial height of 15.2 mm at the pitchline. This pro- vided six 12.7 mm diameter pins per row. Two pin materials were utilized as

' -

mentioned previously: copper and Textolite. The Textolite pins were used to determine the effect of the pin created turbulence on the sidewall heat trans- fer by the l.ow conductivity of Textolite. Correspondingly, the copper pins having a much higher conductivity than Textolite were designed to determine

I

the contribution of the pins on the sidewall heat transfer. These pins were hand-fitted at each location and attach~d to the plates mechanically by machine screws. Highly conductive grease was used at the pin and plate inter-

I

faces when copper pins were used to reduce heat losses at this l~iation.

Pin locations are shown in Figure 31.

!1

I

I.proved D08ign Point II •

10 ...... ----+---

...

•

I

Engine Design Points ReOrlg1nal • 207,854 Re • 179,545 ImprovtiJd Original Design Point

•

2 4 WI.

Reynolds Number, Re

.-

L AIJ.

(L • Heater Width for Each Model) Figure 29. Comparison of Heat Transfer Results for Original and Improved Design Impingement Inserts, • ~ .,,., ,., .,~ ,~" 1

......... --

"

56 Holes • 26 Pins/Row ..

Figure 30. Stage 2 Vane Trailing Edge.

en ....

"

I •• 01 • • 'In.

C V not • G ua d H at r o 0 not • O .t a H at r Inlet Plenua nan Fi ur> 31 .

g~ V n Tr i1ing Edg Pin Fin H at Tr n f r Mod 1.

I

I

The model was fabricated such that it could be split axially to allow for in.tallation/removal of the pins as required by the test configuration.

I

The configuration. te.ted were: Number Configuration

I

1 No Pins Textolite Pins

j

Copper Pins Photographs of the assembled model and the test setup are shown in Figures 32 and 33.

5.3.2 Instrumentation Three data heaters were located on each side of the passage dde<.rall as shown on Figure 31 to obtain data for each pin row as well as immediately ..

ahead of the pin rows. Guard heaters were located upstream and downstream of the data heaters to reduce heat losses and to p~ovide upstream boundary layer heating. These heater plates were nominally 25.4 mm wide and 152.4 mm in ..

length utilizing 1.6 mm thick copper. Heating elements were soldered to the back of the copper plates along the full length of the plates. Three thermo- couples were attached to the plates to record the plate temperatures during ..

test • < ..

5.3.3 Test Procedure oil The test procedure consisted of first setting the desired airflow and then electrical power was supplied to the heater plates. Vari8<ble trans-

J

formers controlling each heater were adjusted such that each heater plate was at approximately 93° C (200· F). When the heater plate temperatures were stabilized, volts, ampere, flow rates, plate temperatures, and plenum pressures

]

were recorded.

A time-share computer program was used to reduce the data. The program is divided into two parts: the first is handling the heat transer data while the second part converts the results to dimensionless parameters such as Stanton number, Reynolds number, and Nusselt number. As previously mentioned, three model configurations were tested. The geometry without pins was used

]

as a basis to determine the increase in heat transfer resulting form the addi- tion of pins. The heat transfer part of the program input consist s of volts, ampere, plate temperatures, and heater resistance.

Total power input to each heater plate is calculated as

I

p = I2R - Losses

I'

I

Fi'lIl'V L), 1'110 1l1\I-dpll (I Pin Fin II d , I, 5 4 • •

~ •

• • •

•

•

•

• • • • • The heat transfer coefficien~ is calculated using the following relation- .hip: The first configuration te.ted had no pins in the trailing edge passage.

The 8econd configuration tested used the low conductivity Textolite pins.

These low conductivity pins were used to determine the influence of the pins on the wall heat transfer. In this analysis, the first step was to calculate the heat loss due to the pins.

A heat transfer coefficient for each pin row was calculated using the re- lation.hip for turbulent flow over banks of tubes or pipes (Reference 2).

~ (0.33) Prl/3 Re 0.6 Dpin D The pin heat loss was calculated as a fin calculation attributing half the pin length to each plate having pins.

The pin fin heat loss was then accumulated for each heater plate according to the humber of pins touching the heater. The pins were known to be in good contact with the heaters since a dab of silicone cOlnpound was placed on the end of each pin prior to assembly. The final step in the data reduction process was to calculate the average heat transfer coefficient for each heater as follows: P - Q p h ..

The data from the third configuration were analyzed by assuming tha~ the wall heat transfer was the same as that resulting from Configuration 2. The remaining heat transfer ~as then attributed to the copper pins. The relation- ships used for the previous configurations were used for the Configuration 3 data reduction.

The reduced heat transfer data were correlated using the following dimen- sionless parameters: hD NUD - Nusselt number based on pin diameter a ~ ReD • Reynolds number based on pin diameter ..

~.-

[

r

5.3.4 Test Re.ult.

[

The data were correlated based on pin diameter for all three configura- tions. It .hould be noted that the Reynolds numbers are based on local velocit)' whh no inerea.e in velocity due to presence of pins which allows

r

a direct compArison of three configurations.

A review of Figure 34 shows the average of the smoothwall data to be

r

about 1.40 thnes the turbulent duct prediction and is attributed to the inlet effect. resulting from the model configuration. Shown also in this figure are the wall heat transfer data from Configuration 2 showing an increase of 3.0 times the smoothwall data.

I.

A comparison of the Configuration 3 data as an average of the walls and pins is shown in Figure 35 with the smoothwall results and the turbulent

[

duct predi~tion. These data .howa 2.8 times increase in heat transfer over the soothwall data which is slightly leBS than the increase shown by Configura- tion 2. This slight reduc tion is due to the averaging of the Configuration 3

r

pin and plate data. A comparhon of heat transfer coefficient of the wall and of the pin of Configuration 3 is shown in Figure 36. This figure shows the pin heat transfer coefficient to be 25% to 35% lower than the walls.

II

!I'.'"

[

[

I

I

~r .,

T

'I

t •

I

I

I'

() Smoothwall Data - Config. 1 lJ.

Fiber Pin Data - Config. 2 6~

~~

H

2 o

o

o 0

0)8 0

~--------~-- 0

Turbulent Duct Flow

o~ CO

Correlation

a l/3

NU = 0.021 Re o. Pr

8 0 0:

D o

o

6 8 4 6 824 W D

Reynolds Number, Re =---

o A lJ

Figure 34. Comparison of Fiber Pin Wall Data with Smoothwall Data.

I

I

J

r

Smoothwalt Data - Config. 1

I

Copper Pin Data 3 <> - Contig.

Note: Copper Pin Data is Average $

of PlatoR + Pins

I

•

f

I

:al~ II

-

I

I ,

•

'" ~ tV

I

~

@

Q

~ • ~

..,

$

.-j

S

r

tV til til =' z - Tur.bulent Duct F10\~

]' 10 Cb8

0 Correlation

O S 1 3

= 0.021 Re ' Pr /

CO

D

O~

I

000'

I

I

J

2 4 Reynolds Number, Re =.!:LQ.

D A \1

I

l

Figure 35. Comparison of Average Copper Pin Data with Smoothwall Data.

I

I

I

Wall (Only) - Config. 3 Pin (Onl.y) - Config. 3

e; t

~

i

C9.

a!)

~

'#

fal~ 00

~

II ,

e

~ z ..

t

~ °0

e

III

~

~ ~ .-t III C/l

•

C/l ~ - Turbulent Duct Flow ~ ~ Correlation

o 021 R 0.8 P 1/3

NUn =. en r

4 6 8 10 2 4 W T1

Rt:'ynolds Number, ReI) = AI"7

Figure 36. Separation of Wall and Copper Pin Data for Configuration 3.

I

:r

6.0 COHroN!N'l' MECHANICAL TESTS

I

r 6.1 SLAD! DYNAMIC AND STEADY-STATE STRAIN DISTRIBUTION

!

l In order to adequately determine the dynamic characteristics of a turbine blade operating in a gaff turbine engine, ,it is necessary to conduct an instru- mented engine test in which vibratory stress data can be measured. However.

to properly collect and analyze such data, it is first necessary to perform a IIeries of tests to determine the resonant frequencies of the blade nnd the relative distribution of strain throughout the blade while it is vibrating in each resonant mode. These data permit the construction of a Campbell dia- gram for the blade operating in the engine. Additionally J the data are needed to select optimum locations for engine test instrumentation (Le., locations which will provide a large relative response in modes of interest) and to de- termine safe operational limits in each mode of vibration. The steady-state strain distributions are desirable on new blade designs in order to insure that the load transition through the airfoil root is relatively uniform and free from "hard spota". 'rhe objectives of these tests are to provide the re- quired dynamic and steady-state strain distribution.

6.1.1 Test Description The test hardware consisted of the Stage 1 artd Stage 2 turbine blades of the improved (singLe shank) turbine. The test consisted of three parts: the determination of resonant frquencies and nodal patterns, the relative strain distribution tests, and the steady-state strain distribution test.

Resonant Frequencies and Nodal Patterns - In order to determine resonant frequencies and nodal patterns, the blades were cut off at the middle of the dovetail top tang pressure face and then brazed to blocks. The blades were driven electromagnetically at resonance and a crystal pickup was used to de- termine nodal patterns. Resonant frequencies from 0-24 ~lz were recorded.

The test setup for this test is shown in Figure 37.

Relative Strain Distribution - Relative strain distributions were ob- tained by clamping highly strain-gaged blades in dovetail coupons and driving the blades at each natural frequency with pressure pulses created by a siren, The effect of the damper loads on the relative distribution was detennined by using a fulcrum-type fixture to apply a load at each damper pad which can be seen in Figures 38 and 39. Distributions were determined for three different conditions: (1) clamped at dovetail only, (2) clamped at dovetail with damper load equal to that experienced in the engine, and (3) clamped at the dovetail with a load approximately four times the engine load at: the damper pads to simulate a limiting condition. An example of the strain gage instrumentation is shown in Figure 40.

., 40J C C o Vl I':l ex

-

III ......

\.0

-

-

~~~~ .. ~~~~~~:-:-:-:- ~ ..

..

- . . . . .

. . .

. .

. .

F f 'll • 1 10. I. " 111 P I l' () t S 1 I .. i II ("I' t r liS tIll I II t "t lUll (' t I 'I' 2 CIHH . I V· Sid · ).

'r: Stead~-St(lte Strain Distributions - 'rhe teat St'ltup for these tes;l:s an' shown in l-'lgure 41. In order to conduct stclIdy .... stllt(> strain distributions, an ~poxy block was Cant on the tip of each instrumented airfoil to providv " mf'<lnS of loading the llirf.oil ~ithout causing distortion during testin~. The blade dovetail was cLamped in coupons, as in the vibratory atrain distribuLion testing. Hath positive and nl~p;fltive moments were applied about all three nxrs, and 8,900 N (2,000 lb) and 17 800 N (4,000 Ib) tensile loads were applied.

t /{(>sults Wl!rC! recorded for lweh loadinlt: condition. The coordinate gYRt!'m lllH'!d was Londing X in the direction of rotation, Y axially forward, and Z radial.

(,.1. 2 Tes t Results

. -

Hesonant Frequencies and Nodal Patterns - 'rho resonant frequencies .and nodal patterns obtained for the Stage 1 and Stage 2 bladE'S are presented in Figurt!s 42 and 43. These data are used to constru~,;t the Camphel' ctiagrams shown in Figures 44 and 45 to determine whicll, if a~y, resonances with known per revs could be potenti~ll problems. Also indicated are known enl;;ine excita- tions and engine operating range.

As can be seen in the figures, several crossi.ngs occur in thE> range of operation of tile engine. However data taken from instrumented core engine testing (see Section 7) demonstrat~s that all resonant vibratory responsos wer~ within acceptable limits. The peak resonant response levels 'vere lOWE,'):" for both stages than the corresponding levels on the original blade design.

Helative Strain Distributions - Relative strain distributions were (h'- termined for all modes which give the detailed distribution of stress in tile blades for these modes. 'rhese data are then used to insurH that I:h(> 11I0S l limiti.ng ar(~a oE the blad(l is knr)Wll {lnd that ::hi.s critlcfll point strain C,Hl bE! related to strains m0.flfw[,f'd during engine testing.

Steady-State Strain Distributioq - For the Stage 1 blade, positive and negative moments or 79,1. Nlm (700 in. -ib) were applied about the X and Y axes, 56.5 H/m (500 in.-lb) about the Z axis and tensile loads of 8,896 N (2,000 lb)

'as well as 17,793 N (4,000 lb) in the z axis. For the Stage 2 blade, positive

and negative moments of 56.5 N/m were applied about the x, Y, and Z axes and tensile loads of 8,896 Nand 17,793 N. Results of beam theory analyses were compared with the measoreo values for both Stage I ~lnd Stage 2 blades at each loading. ExampLes of the aIllllytical results and the comparison are shown in FigurRs 46 and 47.

'l'he results fnr both blades are similar in that they tend to unload on the leading and tr,ailing edges and load up over the r.elatively stiff rail areas. The Stage 1 blade also has a tendency to unload at the maximum curva- ture on the convex side where the blnde overhangs the shank.

The steady-state results did not show any unusual end effects for l'lither Stllg(~ 1 01." 2. The magnitude of the end (!ffects is within the realm of current exper:i.(~I1(!!' for blading of this type. Testing revealed no adv('lrse steady-state end effects/stresses.

F i 'lll'l' I, I • 11.1 ill ()isl, illll l iPIl 1,",1 , ,( III' ' >7 If" ~- N"-- Cl Q) .

.

(Concave Side Show~) } _. ~'~::x~,Y'< ::::==-='-=~ 4,320 Hz 2,394 Hz 6,150 Hz 8,102 Hz 11,,918 Hz 9,630 Hz ' C

•

"'I

: ;~ -i_ l' ffi3 I' , -~'-..-:-:l

G-~ JRff:: t¥4~'.'1 --'.-'-::~ ~ 1&'?"~\1

'1

! -1 f.\ '" -dj -£-c.m ~ ~~

C , 12,590 Hz 13,434 Hz 14,012 Hz 15,858 Hz 17,450 Hz 23,834 Hz :l ~ t , Figure 42. Stage 1 Blade Frequencies and Nodal Patterns.

...

• '~"".- -.......... •• "'-'- .. ~~ • .'lIr.. ......... b, .... "'.~ •• £ • ........... t_· .......... ~ ....... _,.0 ~ _ _ ~ JIIiiiiI ---- jIIiiiiiiiJ' ~ iiiiiiiiiiiiiOI ~ ~ ~ -~ ~ '-=y ~ ---. -===="" ~ ...-r (Concave Side Sho~Ttl)

IF

::--=-~~ 10,030 Hz 5,230 Hz 9 286 Hz 1,392 Hz 4,286 Hz 7,460 Hz 2,826 Hz 9 2S , i=-::-::":·~!.~ '~£:= r-"-i

-------

20,314 Hz 23,814 Hz 17,800 Hz 18,356 Hz 15,654 Hz 10,830 Hz 13,306 Rz Figure 43. Stage 2 Blade Frequeucies and Nodal Patterns.

en t.O 18,000 2-SCI~ip~ 16.000 I

12,000 2 Tension ---t---

til 10,000 t= 2 Axial ..

>.

s:: Q) Flex ::I C' Q) ~ ~ 1.

I I

Rnginc Operating Range

o

2000 4000 6000 8000 10 ,000 12,000

°

Rotor Spoed, rpm Figure 44.

Campbell Diagram for Stage 1 Blade.

I

J

18,000 ~------~------~------~----~------~------~

I

128/Rev ,.

16,000 2-Stt'il)e ---4- I

J

" 14,000 ~------+---.----r-----~

I

12,000 \, N p:: ..

» () ~ III 10,000 =' 0" , III , ....

I r....

J

j

..

M

·

I

, 33/Rev

* -+-----t---~r 32/Rev I

·

30/Rev I I ..

· r, 4000

J!

l

", .

...... ' J ',,;

I I

Engine Operating Range o ...

o 4000 6000 8000 10,000 12,000

I

~ Rotor Speed, rpm Figure 45. Campbell Diagram for Stage 2 Blade.

,,",.

.....

t-j (-3000)

• }fx: = +79.1 Nm (700 in.-1b)

G - MPa (lb/in. ) -115.9 (-16,800) 21.4 (3,100) (7,600) 113.1 (16,400) 84.2 -77.9 (-11,000) (12,200)

~

-45.5 (-8,600)

~

-J6.6 (-2,,400) Figure 46. Typical Results of Beam Theory Analytical Steady-State Strain Distribution (Stage 1 Root Section Shown).

r-

• .- ...... ~, ....

~ • .....

...... -

~ --

'"

----

--- -- -

-- --

Distance fro. I.eadine Ed,e" in.

150 -2 -1 0 ~ I II 1 I 1 20,000 [) No End Effects.

:,

o Measured

~ ,~~ 100l-t -----i---+----+-------1r---.::

:1

n 'I 10,000 , ' ; 50t~---~----+-----~----_r_

•

..

..

~

•

• - • ~ OJ ~ '» ------4-----~~----~----~ ~O

.. fr

.. (700 in.-lb)

....

rn

"

= ~ -5000 -50 t-I-- -10,000 -100 t +--, ,I • Bea1ll Theory Analysis -15,000 Convex Concave

• •

-150 I ,- 20,000 -6 -5 -4 -3 -2 -1 o 1 2 3 4 Distance fro1ll Leading Edge, em Figure 47. Comparison of Analytical and Measured Steady-State Strain Distribution (Stage I·Root Section Shown).

-.J Co)

....." 7b _ all :!! .'O:::::~::-:::"_, __ ~,_ . ____ _~ ___ ., __ ~~~ __ ~~~~~

6.2 BLADE FREQUENCY AND AMPLITUDE AS A FUNCTION OF D»1PER FORCE Frictional dllml)erS have been used to reduce turbine blado vibl'lltOt'Y stresses for a number of yoal's and have been the subject of 8tllVt'rlil ilWl'flL i- gatLons in the past yetlra. Similllr teating was uRad in the development of lightweight dampers for the CF6-50 Ill''!' bhde. 'fha purpose of bench dlUnl)ina teats L8to determine if the engine damper load will provide die de6ired com- bination of first flex frequency incrense and stress reduction.

6.2.1 Test Description 'rhe improved Stnge I and Stnf~e 2 bhdes to be tested where modified as shown in Figure 48. 'l'he blades Wl'f(' Cllt off lit the middle of the dovetail top tlllig pressure hce and brazed to 11 block. In nddition, the damper pnds were mach in~d perpend icular to the clovetni 1 c(\nterline in nn ul fort to maintain friction loading by the dampers. A strain gage wAs also applied to the air- foil root leading edge.

The blado was clamped between two large blocks a.nd loading was applied to the damper pads widl a fulcrum-type fixture (Figure 38). A load of 6 to 7 times the engine load was applied to the damper pads and the blade was then driven in the first flex vibratory mode with a siren. Holding the siren pres- sure excitation constant, the damper load was removed incrementally down to a minimum of 329 N (74 Ib). At each load, the first flex frequency, tip deflec- tion, and root leading edge stress were measured and recorded.

6.2.2 Test Results Results are presented in terms of blade frequency, tip deflection, and root leading edge stress as a function of total damper load. The Stage l and Stage 2 frequency results are shown in Figure 49. The first flex fre- quency increases of about 11% for botll blades were as expected.

The tip deflection And root leading edge stress results are shown for the Stage 2 blade in Figures 50 and 51. Tip deflection and stress data Eor the Stage I blade could not be obtained because of the extreme difficulty in driving this blade. The inability to drive the Stage 1 blade to high ampli- tudes has been observed in the past on other low aspect ratio turbine blading.

\Hth maximum damper load and full siren pl'eSSlIre, the tip deflection was less thlln 25 microns (0.001 tn.) and the strt.'ss at the root leading edge was only 14 MPa (2000 Ib/in. ). Tht" magnitude of the stress reduction on the Stage 2 blade is similar to that seen on other blades of this type, such as the CF6-50 Stage 2 blade.

Uecause the damper effectiveness appeared to be very large, further test- ing attempts were not conducted. III conclusion, the results indicate that the dampers will provide the desired effects relating to frequency gain and stress reduction.

I

f

~

J

t

I ,

i'

i i ~

f

I . ,

c==='::::t .......... _----- Pads Machined

~ ./ Blade B'razed

1 ", to Block

~----~\~,-----(~,I----~~

,. "

r

"t " , , , .

- --

r

r

II, 'I,

I

Figure 48. Blade Modifications.

I

I

I

I ntal Mimpor howl, f4momin 700 I1N1 ON AfNL HOW) ]wM N v N t+, 0 1000 MOO 1000 ron0 5000 6000 Total nIimp or Londe Nowtvi m Total 111IMpor 1,000, NUUndu N mY.

a^ 1475 N o^ a M (w 0 1000 2000 1000 4000 $000 T001 Damper Load, Newtons Figure 49.. Frequency Versus Damper Load for the Stage 1 and Stage 2 Blades.

t.

...... -,. v--,- "T' -;. ---. ~ ~ ~ ~"'" ~ ~~ ...... Iiiiiiii • I .--.. - Total Damper Load. pounds 400 600 800 1000 rn ~ ~ ....

;...

'C CJ ..-{ f:

i'

..

...

'"

c:: tiD (') ..-{ .., 10 :: CJ g QJ :- ..-I .....

-

ClJ II ....

Engine Damper Load --tl------------~~------------~----------~ ...

Co .,..;

•

E-I

o Run No.1

o Run No.2

Blade Brazed to Block -- Pads Machined Flat CF6-6 HPT Stage 2 Improved Blade

I '0

o 5000 6000

a 2000 3000 4000

Toal Damper Load, Newtons Figure 50. Tip Deflection Versus Damper Load for Stage 2 Blade.

-.J -.J

---

-.J lID Total Daaper Load, pounds 200 400 600 800 1000 1200 I I I I • I.-.D' 30,000

-

~

i

l

oJ,~

- 15,000 !

Va

•

1Pt-

I

..

f--- -I ttl

•

j

•

~

•

'" 10,000 :;

til -

til .......

~ aI ..

).I ~Engine Damper Load Run No. 1 " ~ P til N Run No. 2 I

-

Blade Brazed to Block Pads Machined Flat

-

CF6-6 HPT Stage 2 .

Improved Blade

t

.J. o

o

o 1000 2000 3000 4000 5000 6000

Total Damper Load, Newtons Figure 51. Root Leading Edge Stress Versus Daaper Load for Stage 2 Blade.

~

- t: ed"-" 'k "'

I . • 1 h d i 1 e ll in ' .

ci \I III till (I II p o it l U ll.

J.3, 2 ' (11 lin II p ( d \I t () f t II i II lr lit i bll i ( 1\ d - o r.

< HI i tn 1.

III n -q lll n il ,- t n 'v ... ll - II In . rum n l l( n w I'i 'ur - ')2 . ln~lrur.llnt: -0 S La gl' J 1110 2 Disk Rim S lr 'SH T's f d -1. .

8 0 F[!'lll-' I. IHI 'C l Rddiill 1.( adin g 101 Disk Rim S ll 'l'Sh ' 1\'~lS (I. Ho ' 1- 'I'; tn g 1.0 lin g) , I'!:III ' " / .. CI,I!;"\11'1I1 Ilill', ' l R.ldi.11 I.o,ldillg 101' J)i~k Rim Stf''-;S I"'HI~. (1,11\1'1' r,lll 1..),ld!lI~), '''; ~ o I • til ....

e:: .¥ U) ....

Q ~ o ....

=

.0 .0 0: ....

o o :l v: C l.: 8 t H5 Figur 57 . Rim Ncvl 1 InsLrum 'nL.l li 'Jn .

Loadi ng aymm~try from s ide-to-s ide was checked by plot t ing the uniaxial .

strain ver~us its radial position nlong the rim face. Idejl loading would produce eq'aivulent strain plots fOl' each symmetrical gage strip along the torward and aft fillets. Each curve plots either the forward or aft fillet strain for the corner, middle, inner, and side gage strips. Location of these strain gages is delineated as shown in Figure 57.

'l'ypical data plots are shown in Figures 58 and !.i9 for the Stage 1 upper tang load case. These curves reveal that the measured strain in some cases shows a rather significant side-to-s ide mismatch while for the same load case other gages show a very close agreement. In the cases where a side-to-side imbalance is measured, the strain amplitudes are averaged to determine the value of KT' The results of this test are illustrated in Figure 60 which summarizes the KT values determined from this test and compared with the 2D photoelas- tic test. The 20 photoelastic test wllS completed prior to the design of the rim fillets and was used to optimil'.e the fillets. The Stage 2 results show good correlation while the Stage I results show some difference. The Stage I differences are probably due to the inability to accurately compute the nomi- n~l bending stress. This reasoning is probably justified; because, if the forward and aft KT's for Stage I were avenged, which would negate the un- known bending stress, the KT would be about 2 which would then compare favor- ably with the 2D photoelastic test result of 1.9.

An additional output from this stress test reveals that the upper tang loading shows a lower concentrated stress in the rim area than an equivalent tang loading at the middle or lower tang pair. The middle and lower tang load- ing produces approximately the same rim concentrated stress. This result is believed to be a function of the dovetail geometry and cannot be concluded as a general result for all parts.

o IIlght 'rQnl lnnor OM~U.

C \I'n lIadlal DUlancll 11,,10* lIallbot, In.

-O.OM o 0.08 0.16 o.:U 1.\11/ HIOO 1/1 pO ~ I ?

I ...

l' J tlClO ..... "' ... ,:~'". *""

---

~ rr .. ....

I:) e I ..

1/1 ~ ...

.

" 11100 III ,.

~ ..

I ."

...

I

P ., ~ ...

.S Q I C ..

~ !II ..

l' \I III ,.

...

...

'" I :' ....

I

l' ::l ~ ....

Q " I

e

...

~ III ..

!'

0 0 Fronl Sid" G31:oS , I' en

!'

...

il

I

F I ...

I !'

HOO ::l 1100 ~ ~ ....

C I

~~~

'" ...

" !'

47~ I~I

() o Ij -2 o 6 Figure 58. Strain Versus Radial Distance from Rabbet - Front Gages.

---~~----·----'-+------'-'--------·-1

.,a~ IftA,r GI ••• •• dill Dl.'afto, .. )~ Klb~l, 1ft.

-0.01 o 0.0' 0,1. 0,24 1/1 :: ..

..

I I zti

!'

...

~

400 r± ~

,;

•

...

0,

e

..

.....

1/1 ..

!"

-400 .eM~ Middle Oa".

200 1200 III ..

0 iI ..

:s I ..

II

I

.

400 400 :l

•

...

Ii ..

I

e

-!'

..

..

III !"

-400 -400 1aoo

, aoo

III ..

..

I J ..

...

€I= :s

I

..

:1 !"

400 O(l Ii ...

...

"

e t::i ~

..

'l III ~ 0- !"

-400 -400 Roar Side Oalle.

1aoo laOO III ..

..

..

o.

I J' ....

l

!'

;:I Ii " ...

..

I • ..

..

~ III ..

!"

2 4 6 8 .adla1 Diatanco Below Klb~t, ..

Figure 59. Strain Versus RaOial Distance from Rabbet - Rear Gages.

..

-.

'

Stage 1 Kr:s Stage 2 KI'S 2D Photoelastic Test 3D Test 3D Test 2D Photoe1astic Test 1.74 (inner) 1.91 (inner) 1.70 1.74 (middle) 1.64 ___ ___, 1.89 (_iddle) (corner) 1.78 1.67 (Corner) (side) 1.47 1.85 (side) ~ " 2.28 (side) (side) 1.92 \ 2.02 (corner) 1.90 (corner) 2.29 , 2.16 (.tddle) 2.25 (middle) 2.49 , 2.20 (inner) (inner) 2.25 Figure 60. Summary of Rim Test Results.

O'J u:> 7.0 INSTRUMENTED ENGINE TEST The objective. of the instrumented engine te.t. were to evaluate the thermal, mechanical, and aeromechanical characteri.tics of the improved delign turbine rotor and stator hardware. Preuure, temperature, cooling flow, and .tre811 data obtained were aho used to conduct additional life and de. ian analy.e.. Testing at .ea level included transient and steady-.tate condition. from minimum to maximum engine power level.

7. 1 TEST SETUP The telt vehicle used was a TF39 core engine, which is the high preuure sy.tem, fitted with an exhaust gas recirculation ",ystem which provided heated air to the inlet as shown in Figure 61. this system was desi&ned to deliver the inlet air at whatever temperature is required to set "redline" cycle parameter.. Exhaust gaa is "scooped" out of the exhaust stream and suppl ied to the front of the engine by an insulated pipe. At the front of the engine, the air splits into a Y-shaped pipe"and is delivered to a manifold which distributes the air uniformly to the inlet of the engine.

the TF39 core ia the same as a CF6-6 core with the exception that selec- ted rotor spool joints are of doweled deaign versus the rabbets used on CF6-6.

It. standard turbine mid frame wa.s used to support the No. 5 bearing. The Ilave exhaust nozzle which simulated the low pressure turbine effective area was variable via bolt-on tabs to produce the proper turbine preuure ratio.

Rotor spool in.trumentation was read out via a 100 point slip ring mounted to the hub of the turbine mid frame. Provision was made during rotor buildup and inatrumentation application to facilitate rapid reprogramming of the slip ring so as to permit three slip ring 0 full of rotor instrumentation to be read out during one build of the ~ain engine. Slip ring changeover was accomplished by removing the turbine mid frame, reprogramming the leads, and reinstalling the turbine mid frame. Figures 62 through 65 show the actual test hardware alonR with typical instrumentation applied.

7.2 INSTRUMENTATION The test inatrumentation used to measure engine performance, turbine stress and temperatures and to monitor engine operation is broken down into three groups: general instrumentation, aerodynamic instrumentation, and turbine instrumentation. Instumentation in each group is given below and is indicated on Figure 66.

~i '1I1l ,1 . 11 jI) (\" J 1\'1n III t.lll u In It l ell .

P GE IS UA llry c c · 3 to ..,.

~ -< en

r

\

~

Figure 64 . Instrumented Stage 2 Rotor (Aft View) Showing Thermocouple and ~train Gage Leadouts~ I'i 'I II " 1\ I • 11l :; II\l l h ' III, ' " .· t . l ~~ " c.p (j) A~-:-:J ttl .. 44 t:.-rb1ae t"J.:!,!l'.-e :;bd~.ric --"'\ 11 =e.p.tT&t;n-e Prnk$ ) S '~e: •• U'e" h~.

~ ~~ ~~."i"" !;t.ton~ P!.a.!)e 3 CorIiFres80r t)'1s;J-tlr,e ~ c:.

... all Stat1::= 'FrusliU.s 14" ~tal !::C'S ."

1 5-Eleaeat '!'e:apelOat~re Prbhe- \ ~i ;:!ti~:;~Stwl'es ~~,'~; t .... erl) Floo.n '.; ._, ~~ ~r-"i;.....

to ~ 13th ~;:t:.ate t E? ltec:.up

9t;~ge ""I:tV \ T' ~ = == 0:-

·-r:;...:;:~~~ .. rti .. "a~ F+.\ .~ ~. ~,;.~ .. ,>~: .. ;; i'!~ ~i d' ~.;~ r:;;::;Il'rt,:cfflJrUi:\~ Turb:ll~ General _II iIe.1.'t'Uag IIoud~ ~ric Press:lre -4 lhnlSt J:ddcea Eclditv =~:b!"-~ Cell Sutlc i'resS"..:re ~J,rnO'E' Cpre S";:ee.:!

~Ute 1 IlMes 21 SIC":' .-.L 1'9 ':':':"5 If"...t1!l :r ..;el Flow St««e 2 U.es 2) $""5 _ 2) TIC's Verification Fuel Flou Jour Stnocr .... es :s S/r.'S _ 4l TIc's Fuel 'I'e:::;:erar.:Ire Ft:el 5a:::yle Spe'Ci!ie Gravity Fuel Sa..,Ie :~...rarure Var1061e: Sutor 'i"a!le Pos1t:icn Figure 66, Engine Test Instrumentation.

General Instrumentation

-

Barometer Pressure

•

Hum:l.dity

•

Cell Static Pressure

•

Core Spe(;llJ

•

Mdn Fuel Slow

•

Verification Fuel Flow

•

F'~el Temperature

.'

Fuel Sample Specific Gravity

•

Fuel Sample Temperature

•

Variable Stator Vane Position.

•

Aerodynamic Instrumentation Inl~t Temperature

•

Core Inlet Static Pressure

•

Core Inlet Temperature

•

Compressor Discharge Pressure

•

Compressor Discharge Temperature

•

Turbine Mid frame Discharge Temperature

•

Turbine Midframe Discharge Pressure

•

Instrumentation was installed and flow calibrated to measure

•

compressor ninth stage flow to turbine mid frame, thirteenth stage compressor flow to Stage 2 UP turbine nozzle, high pressure recoup to turbine mid frame and cooling flow to high pressure turbine rotor.

Turcine Instrumentation Turbine Rotor Strain Gages - Twenty-three dynamic strain

•

gages on Stage 1 blades, 23 dynamic gages on Stage 2 blades, and 8 dynamic gages on nonairfoil components.

Turbine Rotor Thermocouples - Twenty-nine chromel-alumel

•

thermocouples on Stage 1 blades, 23 on the Stage 2 blades, and 42 on nona1rf011 components.

, • Turbine Stator Thermocouples - A total of 148 skin thermocouples and 51 air thermocouples (all chromel-alumel).

• 'rlirbine Pressure Instrumentation - Twenty-three static pressures.

o 4d Dearing nlruat Bridges - Four static strain gain bridges on the 4» bearing housing.

7.3 TEST PROCEDURE Instrumented testing of the improved turbine was divided into three dis- tinct testa. This was required due to the quantity of sensors on the rotor and the capacity of the slip ring.

Test 1 - This test concentrated on accumulati.ng the blade vibratory data throughout the engine speed range. A few "safety" thermocouples we~e read on the Stage 1 turbine blades. Simultaneously, turbine stator temperatures and pressures were recorded.

test 2 - Reprogramming the slip ring provided data on Stage 1 blade tem- peratures, rotor spool strai.n gages, and selected spool temperatures. Numer- ous transients were run to determine heating/cooling rates. Steady-state data were obtained at a variety of power settings and inlet conditions.

Test 3 - Ret"outing of rotor instrumentation was done with the rotor remain- ing in the engine. All remaining rotor spool thermocouples were read on this test along with Stage 2 blade metal temperatures. Essentially, the same test plan liS Test 2 was followed to gather steady-state and transient data.

7.4 TEST RESULTS The instrumented core ennine tests were designed to provide vibratory response, temperature, pres~ure, and cooling flow data. 'rests included a range of steady-state and transient engine operating conditions. Results presented are divided into three groups: rotor, stator, and cooling.

7.4.1 Rotor Stage 1 Blade Vibratory Response - Blade strain gages were applied to the blades at three locations: shank with radial orientation, root trailing edge pressure side with radial orientation, and pitchline with axial orienta- tion. The threE' locations chosen were enough to provide sensitive gages for all expected modes of vibration. Typical results of Stage 1 blade vibratory response are shown in Figure 67. Examination of this figure shows that pri- mary Stage 1 blade response was in the first flex mode as expected. However, it is significant to note that no resonant response was observed. The blade responded in a force-driven fashion to very low l.evels. Maximum blade responst~ 'n ' L

.-

.-

e ~

i

Z e !

~ i .. '-'

• •

tU • ..

• ...

• ...

• ... ... ...

N • A .Hl.l1Ib. 'l .1 1\ was 19% of allowable stress, extrapolated to fan engine conditions and account- ing for the +30 blade sttess. These stresses wer~ very low and well within design limite. No '.iignificant responses were measured in any other mode.

Stase 2 Blade Vibratory Response -Blade strain gages were applied to the blades in two locations: root pressure side trailing edge with radial orientation, and root suction side trailing edge with radial orientation. The two locations chosen were enough to respond well to all expected modes of vibration. Typical re~JUlts of Stage 2 blade vibratory response are shown in Figure 68. E'xamination of this figure shows that primary Stage 2 blade re- sponse is in first flex as expected. fl~aoni1n\: res ponsi'! was observed wi th 10/rev crossing at ~600 rpm and with 8it$.v at ~10,400 rpm. The speed of 10,400 rpen is well above CF6-6D core speed of 9827 rpm and, therefore, the blade will not see S/rev or respond to it during engine operation. The test vehicle was run to this speed for the sole purpose of thoroughly investigating the S/rev blade response. Blade response was to 11% of allowable stress at the 10/rev crossing, 19% of allowable stress at CF6-6D max speed, and 38% of allow- able stress at the 8/rev resonance, adjusted for fan engine condition~, and +3 blade response. Again, these stresses were considered low and well within design limits.

Rotor Vibratory Response - Strain gages were applied to the rotor spool.

Response was to levels comparable to those e~perienced on the original (twin 2 2 shank) CF6-6 rotor. The maximum observed level was 4826 N/cm (7000 Ib/in. ) double amplitude. The mode shape was rotor bending responding to 2/rev. This 2/rev excitation is due to a frame feedback to the rotor of a l/rev unbalance load applied by the rotor to the frame. This effect was expected and is con- sistent with response of the current production rotor. Therefore, no rotor vibratory problems are anticipated.

Stage 1 Blade Metal Temperatures - Numerous thermocouples were applied to the Stage 1 blades to measure metal temperatures at pitchline and below. Re- sults varip.d depending on the specific region as can be seen on Figure 69, but the following general conclusions can be drawn: • Bulk temperature was slightly lower than predicted.

• Most local temperatures agrees with the heat transfer model.

• An area of higher-than-expected temperatures was found on the pressure side at pitchline.

Based on the lower bulk temperature, it is IX>ssible that further perfor- mance benefits can be achieved with reduced cooling. This will be studied in conjunction with cooling pattern changes to reduce local areas of higher- than-projected temperatures.

I .• •• 1 ..•• ,. I' .1 .•.. I •• I .1., .• t t • I fl. t ' .1 ••• , I .,. ,.

en c: o ~ !

~ !'O u en .

. .

.

I I

I I

fit ..

I I

.. I I

•

• ... ..

... ..

711 ' " lU.ln!l J.i 1 1 ,...

o ~ • • • • • • • • (960)

·

• ·,1 • • • (1022) • (l018) (983) • (1032) · '.

• • • • (1018)

!

I

f (999) (952) ~: 1088 (1052) (1021) Pitch Line Temperatures -: 'i xxxx: Ileasured (Tbulk = 916)

T • = ~401

4 1 (XXXX) Design T3 = 543 Figure 69. Stage 1 Blade Neta1 Temperature (0 C).

J

Stage 2 lHade Hetd 'femeeraturell - As shown on Figure 70, mCllsurl"d Stng\\ 2 blade metal temperatures were closc to the expect.ed values with Home d ia- crepancies in the midchord and trailing edge regions of the blade. 'l1te tern" perature of the midchord region of the blade pressure side pitchline was Home- what higher than predicted, while the trailing edge was substantially cooler.

This results in a nct reduction in the blade bulk metal temperature. Posttest analysis has confirmed that there is a life/perfo~ance benefit potential due to this lower bulk temperature. ~~prrpriate design changes for cooling flow reduction have been incorporated to achieve further perfo~ance benefits.

Rotor SpOOl Metal Temperatures - Numerous rotor spool temperatures were measured at steady-state and transient conditions throughout the engine opera- ting range. The instrumentation was applied to key areas of the rotor to de- termine an overall rotor temperature distribution. Controlled transient tests were run to carefully determinp- the response rates of the rotor spool. Re- sults for steady-state design conditions are shown in Figure 71.

With one exception, all rotor temperatures were in good agreement with pretest predictions. The exception was in the area of the front ahaft which was about 28· C (50· F) warmer than expected at design conditions. The effect on stress and life has been analyzed and found to be acceptable, i.e., the life of the part meets requirements at the higher temperature.

7.4.2 Stator Stator Temperatures - The Stage 2 nozzle support and surrounding struc- ture were heavily instrumented to measure metal and air temperatures and static pressures. Figure 72 shows the difference between measured and pre- dicted temperatures on the shroud support structure, extrapolated to red line turbine inlet temperature (T4) conditions. Note that predictions were in good agreement with measured data except in two areas. The first of these areas is in the region near the Stage 1 shroud leading edge whe're the support

I

temperature was meaJured to be about 83° C (150· F) cooler than predicted.

It is suspected that the boundary conditions in the analytical model were in error, causing somewhat excessive predicted temperatures. The second area of

I

disagreement between predicted and tr,easured temperatures is in the area of the compressor rear frame and turbine mid frame flange. In this test, this flange ran substantially cooler than predicted. It is believed that this difference is due to the difference between a fan engine and a core engine.

In the core engine configuration, cold ambient air is free to flow onto this flange. In a fan engine, the flange is surrounded by the core cowl and air inside the cowl space is warmer and has little velocity. Predictions were

I

made for the fan engine and measurements made on a core engine; hence, the measured data is colder than predicted. This is apparently reflected in the structure immediately inboard of this flange, which also ran cooler than

I

predicted.

Figure 73 indicates the degree of success which has been achieved in

I terms of isolating the shroud hooks fom the hot gas stream. Note that the

Stage 1 hooks are 16.7° to 41.7° C (25° to 75° F) cooler, even at a relatively

I

f

:rj 952 (SS9) 939 (943) 961 (929) 903 (896) 934 (936) 960 (967) 948 (999) 951 (1C'J05) (878)

~(1009)

889 (879) 973 (921) (902) -

_/

(898) (920) Tbulk ;: pitch MC;Hlsurod XXXX Design (XXXX) Figure 70. stage 2 Blado Metal Temperatures (0 C).

I

1.-

... -

(554)

XXX MEASURED

(XXX)

PREDICTION

~

Figure 71. Steady-St~te Rotor Spool Temperatures (0 C) for Design Conditions.

....

o C) ...1151-20, -83.3/-150 -37.8/-68 ~

•

-50.61-91 -6.1/-11 -83.9/-151 +3.9/+7 I *16.1/+29 -55/-9'9 I +16.7/+30

I

+13.3/+24

I

.\ _.--

"'\~ II. _ --=-'t O Figure 72. }Ieasured Stator Temperature Deviatio!ls from Predictions (A CiAO F).

J

I

J

!

J ~ : I

J

i

'4 i I ., Figure 73, Isolation of Improved L Shroud Hook f ong-Time Oi s rom Hot G mensional St b,as Ingestion for a 11ity,

J

I

low compreesor d Lscharge temperature (1'3) of 398.9· C (750· F). 'nte Stage 2, nozzle vane flange is 41.1· C (75" F) cooler. ntis ahould improve the long term d~nen8ional stability of the support and reduce engine deterioration clHlsed by ollt ... of-roundneas of the nozzle support.

Shroud support hook radial gradients have also been reduced, as shown in Figure 74. In this figure, the steeper sloped curves iod icate a slower rate of change of temperature as n function of rad ius. On the aft-most flange, a lower overall grad lent had been achieved by scalloping the flange.

These lower gradients should improve both long term dimensional stability and reduce the potential cracking of the Stage 2, vane flange.

Figure 15 shows the circumferential thermal grad ients at t\oiO selected points on the nozde support. Note that the gradient is smatl in the forward dng, while a 45 C (811) F) grad ient exists in the afl: ring. nle cause of the higher gradient in the aft ring is the impingement of 13th stage CaTlpreSSc,r air, which provides cooling air for the Stage 2 nozzle directly on the Stage 2 nozzle support. The thermocouples which are indicating the lower tempera- tures are dtrectly under the 13th stage air pads.

Stator Transient Response - 'rhe response of the improved turbine rotor Ilnd stator during an acce1 is shown in Figure 76 based On rad ial growth calcl,l- lations. '!1te data shown reflect fall engine conditi.on predictions which have been made based Oil the correlation of the analytical mode.ls using the core engine test data. The improved turbine was calculated not to rub. 'lhe se- quence of testing on the engine was such that the anount of blade tip rub exper'ienced dur ing slow rol1ups to maximum speed dur ing blade vibratory stress determination exceeded the amount of tip rub norntally seen, and no additional ti.p rub was measured duri.ng subsequent rapid accels and decels. However, during testi.ng on subsequent improved turbines, it has been established that the blade tips do not rub during a rapid accel.

TIle transient response during a deceli s shown in Figure 77. '!'he com- bi.ned rotorl stator response lead s to the hot rotor rebur st predic tions for tip rubs shown in Figure 78. Note that the improved turbine has substantially lower tip rubs incurred as a result of hot rotor reburst than the current twin shank turbine.

In summary, stator temperature measurements made dur ing engine testing of the improved turbine show several trends relative to the current production design. In general, bulk tempe.:atures have been reduced 13.9° to 47.2° C 0 0 (25 to 85° F). Radial temperature gradients are at least 11.1° C (20 F) lower, and circumferential gradients are very low. The transient response of the stator shows that it responds slower than the current stator on both accels and decels. This indicates potential improvement in the lI110unt of blade tip rub during both an accd :md a hot rotor reburst.

--y-

I

J

0- --(] Improv(·d 0--0 Original l'roductJUJI AT - • II fiT - • F liT - • F o 25 50 o 25 50 75 100 o 25 50 75 100 I I I

1- t~!..' 1 I ,-

I -f]

'-

4S

1--

0..

~

B 44 '1\

9 ....

Ef~ ---

~

...... ~

~

f--.~I~ .- ,~--- ".-

....

::l .

.

o 10 20 30 40 50 60 0 .10 20 30 1,0 50 60 o 10 20 30 40 50 GO N!' - 0 C 1\1' _ 0 C fiT - 0 C Figure 74. Shroud Support Radial Temperature Gradients.

]

]

:l

l ....

....

o ~ -...--, , ,J------ :

==!l....- ___ -

=

---'~ '- ./

~ = --.-/"\

rllFL 'J,S} 't;-~~~rre:~

~ta' ~ . J (F" ~t:: f Jr, ti-t~!0-=

~~~ r ,~ " d) - "==:'

j

i ,,--- T ......J ~ i "\--'-' L

\

~====~ .. , ;:' ~ '----

t;~.J ! r

rr'; i i __ ;_, i ! '" : ; .. -l \ 1 • T l: emperature in 0 C (0 F) 513.9 511.1 (957) , (1024) 488.9 482.2 ,(900) (912) " ilT :: 45 (81) 527.2 ilT = 17.8 (32) 544.4 563.9 (968) (981) (1030) (1047) , I 522.8 Out (973) Temperature 0 F Figure 75. Stage 2 Nozzle Suppor~ Circumferential Temperature Distribution.

_ ,_ 1

I

.. II 6719 :t .. 116789 2 :I .. 56789 a Figure 76. CF6-6 Fan Engine Accel Response for Stage 1 HPT.

u " I I ~,..; t '1

]

:I 41167119 2 ;J 4 ~ II 7 II II 2 :t 4 5 6 7 "II 100 1000 Thill, HuconclH Figure 77. CF6-6 Fan Engine Decel Response for Stage 1 HPT.

..---- ....... -+--Orllllin .. Turbtne ..

Q, ...

...

...

.c ..

Q, ~ 0 50 100 150 200 2~0 100 TIIM! Oe.lny ReCore ReburRt, "('cond,, Figure 78. CF6-6 Average Stage 1 HPT Rub During Hot Rotor Rebursts.

I

7.4.3 Coolins Rotor Cooling Flow - Measured cooling flows were about 1% higher than designed. This WllS probably due to leakage throught the multiplicity of slots in the rotor for instrumentation leadout. Subsequent noninstrumented tests have indicated flows to be about 0.2% above intent. possibly still due to rotor leakage. Mechanical design improvements are being incorporated into the production rotor design in the areas of suspected leakage.

Stage 2 Nozzle Cooling Flow - The thirteenth stage cooling flow rate to the Stage? vane was intended to be reduced by 0.2% for the improved single shank design. The test results showed flow increases of 0.2% to 0.4% above design intetlt. This increase in flow appeared to be caused by leakage around the insert where it rests against the nozzle as indicated in Figure 79. This potential leakag~ area will be sealed and evaluated in a subsequent engine test. A design change will be implemented to reduce this leakage.

Interstage Cavity Temperatures - Design changes were made in the rotor- stator cavity areas to improve the sealing from the hot gas flowpath. thus reducing hot gas inflow and permitting a reduction in cooling flow. As shown on Figure 80, measured temperatues were lower than predicted on both cavities aft of the rotor, but somewhat higher than predicted on the rotor forward cavities. In both regions, the cavity temperatures were still lower than those currently experienced on the OF6-50 engine. E££ee:s of these tempera- tures are being evaluated and, if necessary, design changes will be imple- mented.

Rotor and Stator Transient Temperatures - Measured temperature transients on the rotor showed good agreement with predictions in the bore region for both Stages 1 and 2. The Stage 1 rotor responded slower than predicted in the web and rim region. This is due to the change in the Stage 1 blade cool- ing supply circuit to bring the air through the impeller attached to the rotor spacer rather than allowing it to flow along the disk as on the original production design.

Transient response of the stator agreed well with predictions and con- firmed the desired slower response of the new Stage 1 shroud support.

c: ..

o % .....

Prediction 61l.

Test (1133) Fi:!ure 80 . Ieproved Turbine Cavity 1'emperatu r e C rison, e C (e F).

:1

8.0 ENGINE ENDURANCE TEST , The objective of the endurance testing was to subject the improved high pressure turbine to an t1Jnvironment representative of approximately 3000 hours of typical airline send.ee. llardware conditlon at the end of the testing was evaluated and compared to current production hardware having comparab 10 on the wing eerv ice.

8.1 TEST SETUP 'The test vehicle used was the same as for the instrumented engine test, a TF39 core engine which is the high pressure system. A slave air supply was used 4S required to pressurize the sumps and a bellmouth and centerbody, designed and fabrictaed specifically for a TF39 core engine, was installed.

In place of the low pressure turbine, a slave exhaust no~zle was installed to give an effective area similar to the inlet of I:he low pressure turbine.

Figure 61 shows the engine installation and the exhaust recirculation system used to simulate fan discharge temperatures at the compressor inlet.

Instrum~ntation used for endurance testing included that required to measu~e engine performance and to monitor engine operation.

8.2 TEST PROCEDURE The engine endurance test consisted of running 1000 cycles of the type shown in Figure 81. This cycle includes a maximum rate transient to the red- line exhaust gas temperature (EGT). The engine was held at this maximum tem- perature with periodic throttle adjustments to maintain temperatures for 5 minutes. At the end of 5 minutes, the throttle was chopped to flight i,d le (7700 rpm) for 30 seconds and then to ground idle (5630 rpm) for 30 seconds to cool the rotor spool components. After t,he 30 seconds at ground idle, the engine was shut down to maximize the stress range on the rotor spool parts.

The rotor was c~ught by the starter at between 1000 and 500 rpm and restarted.

Eight minutes was then spent at ground idle to complete the rotor cooling. At the end of the 8-minute ground idle period, a throttle burst was again made to begin the next cycle.

As previously stated, the intent of the test was to subject the improved turbine to an environment which simulates approximately 3000 hours of typical airline service. Judgment as to a given test cycle severity varies with each component and, in fact, each point on each component. However, an indicator of the test cycle severity can be gained by consideration of the rupture life consumed by the Stage 2 turbine blade. This calculation indicates that 1000 endurance cycles simulate approximately 9600 hours of airline service when consideration is given to derated takeoffs. Other components (like the HP tur- bine shrouds) J when inspected after the test, appeared to be in a condition representative of 2500 flight hours in airline service. This judgment was confirmed by airline representatives when they made a posttest inspection of the hardware.

f

t

11"1

o 5

1 , lTILnu F 1 liT 1. ' ndu

I

I

'l'he "oxtr~l 'f!Vt~rity" cycle Wllfl choun in ordor to quickly determine de- sign weaknuilcs. I.t W4l8 allo done to permit tImely solution to rav(ulhd weak- nespel prior to field lorvico evaluation of the turbine. In order to do this,

I

the tut may hllve been overly severe in Bome rupects.

t

8.3 TEST RESULTS The test engine accumulated 255.37 hours of total running time with 84:14 hours at EOT's greater than 857' C (1575' F). During the test, 1000 endurance cycles were run with 1022 transients between speeds below 7700 rpm and above 8800 rpm. A total of 347 steady-state data points was obtained.

After completion of the 1000 cycle tost, tho engine was disassembled and the turbine torn down for inspection. 'l'his procedufa included a "dirty" lay- out and visual inspection, cleaning, clean inspection, zyglo inspection, and appropriate dimensional checks. In addition, flow checks were made on sample airfoils to determine flow reduction due to l1urface debris buildup. The indi- vidual rotor and stator parts discussed are indicated in the improved turbine cross !Jection shown in Figure 82.

8.3.1 Rotor Hardwa~ Posttest condition of the turbine rotor was excellent. Photographs of the rotor after the 1000 cycle endurance test are shown in Figurf;s 83 and 84.

Stage 1 and 2 Blades - The Stage 1 blades were in excellent condition after 1000 cycles. Typical leading edge condition can be seen in Figure 84, and. the pressure s ide and trailing edge condition are shown ;.n Figure 85.

Each blade in the set was inspected under a microscope (up to 60X) in the dirty condition. The findings were as follows: • Almost all blades had suction-side gill hole cracks up to 0.63 mm (0.025 in.) long. These cracks were predominantly around the pitch- line but were observed to be as low as the fourth hole up from the root.

Squealer tips were in good condition with no cracking. Some oxida-

•

tion/erosion of i:he base Rene 80 wa~ noted on the rubbed surface at the extreme tip of the blades.

All other holes, including the trailing edge slots, appear to be un-

•

cracked.

U9

f

Sellill.

Imp (vld IlIrhln 1'1 \II \ H • • .....

...

-< c:: .~ .x o

.s

I/; .

~ u > u o o o ~ l-o -< l-o o ...

o ex: t"'l

(

(

-----

-

1 1 l'iflll' HI, S (d'i I .llld ' 1I1,Idl' ill 1',1(1 A~HI'lIIhl I It ' I I(H)() (\'1'11'" VI <Ii ~ U ~ U o o o ~ u Vl o c o '.1 U '.1 "0 C o U ......

('J U ......

'.1 <Ii 0."0 :> , '.1 f-< tIl 1/'\ CO ORIGI l PA ' f: S OF POO R Q U, l Y • The pressure sides of the blades were carefully examined (an aren wlH~re other blades have shown distress) and no cracking was found.

• Dovetails; platforms, fillets, ctc., appear normal with no unusual wear or cracking.

In addition to the ahove, two blades were vapor-blast cleaned and rein- spected. No additional distress was secn.

All blades were judged serviceable, and nll but three were returned for additional endurance running. The three blades removed were submitted for detailed metallurgical cutup and evalUAtion. '1'llis evaluation basically con- firmt'd the visual findings. The only h()l(~s that were Lound cracked were the sucti.on sidl1 gilt holes and one lending edge hole at about 80% span. No other distrl~ss Wi\S observed in the airfoil. Som(' minor cracking was observed in the tip cap to squealer tip fillet at midchOL-d pressure side. llased on these re- Bults, the Stage 1 blade is judged to be in excellent condition considering the severity of the 1000 cycle tes t.

All the Stage 2 blades were inspected under a microscope up to 60X. The blades were in excellent condition with no unusual distress or wear. Typical condition of the Stage 2 blades are shown in Figure 86.

Stage 1 and 2 Blade Dampers- Lhe Stage 1 blade dampers (Figure 82) were in excellent· condition after the 1000 cycle test. Four Stllge 2 dampers were found cracked at the 500 cycle engine inspection. These cracked dampers were replaced for the second 500 cycles of running. After the 1000 cycle test, 33 dampers were found cracked. 'l'he:;e damper cracks did not cause any .functional problems. All the Stage 2 dampel;'s which were teste'.! were judged unservice- able and were replaced With new modified dampers f<lr additional cndurunce testing. The damper design has been modified to eliminate this cracking problem.

Jotor Seool and Disks - All rotor spool hardware appeared to be in excel- lent condition. Posttest inspection included visual dirty examination, clean- ing, and zygia. No cracks were found nor was any other unusual or unservice- able distress found. In addition, dimensional checks were made at key loca- tions on the rotor. Of particular interest were the disk outside diameters, blade retainer axial drop dimens ions, and bolt circle dif.lm/~ters of the com- ponents dowelled at the Stage 2 disk inner bolted joint. No significant per- manent distortions were disclosed by the dimensional inspections.

In SUlmllary, rotor spool components were all found to be in good service- able condition. These parts were reassemb led and reins talled in the engine for additional endurance testing.

8.3.2 Stator Hardware The condition of the turbine stntor following the test was very good.

Figure 87 shows the HtnAc 1 nozzll' aHHl~mhly as rcmoved from the engine.

I

I

\.0 ::I III III \.0 0-

-

III ~ ....

u :» u ....

"'" <1; III QJ "0 ....

CO N !O ..., Vl .....

c: ....

..., ....

"0 c: u ....

!O U .......

'M 1.1 0.-0 >- ' M ...... Vl c ....

o u o o ...-!

o . , ..

1_.-' .... . , ..

II

Stage 1 Vanes - Typical St41{W L vane scgmentR .u,) HhoWIl in I·'igllrl's 88 olld 89. Twenty-two vane segments exh ill tted v41rying degr{~(>s of cracking. 'rM H(\g- ments showed no cracking When inspected in the as-run condition. Vane suction sides were clean while the pressure sides exhibited buildup of iron oKide scale and were extremely rough. Even though leading edge film holes showed considerable plugging and clogging, no leading edge cracking was observed.

iJ

1- Elimination of leading edge cracks was a major objective of this redesigned vane. All parts were judged serviceable.

I The cracking on the pressure side of the Stage 1 vanes is thought to be I.

caused by pressure side temperatures which are too high. Design modifications have been made and subsequently tested. The vane cooling system has been modi- fied to shift air from the suction side to the pressure side by adding film cooling holes to the pressure side.

Subsequent testing consisting of 1000 endurallce cycles was conducted with the Stage 1 vane configuration as follows: 7 vane segments - Best from previous test (as-is) 7 vane segments - Worst from previous tes t (as-is) 6 vane segments - New vanes wi.th inserts modified to put more cooling air to pressure side 6 vane segments - New vanes with modified inserts plus pressure side film cooling holes 6 vane se~'11Ients - New vanes unmodified (to act as a control) Results of this additional testing have shown that the addition of pres- sure side film holes provides substantial reduction in the pressure side cracking.

Stage 2 Vanes - Prior to disassembly of the Stage 2 nozzle assembly, an inspect ion of the vane inserts through the spoolie ports revealed significant deposits of aluminum, plugging the impingement holes of the inserts. 'rhis was particularly heavy in a 90° sector centered about the top vertical center- line. The source of this aluminum was the gap in the thirteenth stage seal located at the top vertical centerline. Plugging of Vane 2 resulted in a burned and bulged leading edge which is shown in Figure 90.

Examination of individual vane segments revealed that the two axial cracks found in the outer-band-to-suction-side fillet radius of the trailing vanes of Segments 4 and 18 during the 500 cycle inspection has grown to a ''f length exceeding 2.54 cm (1 in.). A closeup view of the vane is shown in \ j Figure 91. These two segments were directly behind areas of Stage 1 shroud distress (missing BradelLoy). Nine other vane segments also had axial cracks in this area, but they were much smaller and tighter. Lm" cycle fatigue is suspected as the cause of this problem.

'l

I

'tl ...

Vl u U o c:

-

o c: o .....

.,; .~ '"0 c: o , .

'0 ....

Vl C o c o .. J .....

c o u u ....

0- :., ~ \...

-

-

.....

ORIG'N~\.

Of P OOR o .-4 CJ > U o ~ u "0 .~ Vl c: o .~ u ;:) Vl 'l'ypic:lIl StllR~ 2 Vlln(l segments art' shnwn in Pigur(l 92. Elpven vanp tJpg- m(lntn had outrr balld crncks betw(!(>n VlInNl. 1hp6(, crllckH run rouJ~hly perpt'ln- diculn!." to the gns flow dir"ction brtween VUI\(lS. Sixteen vane segments hllt!

axiol cracks in the inn~r band trailing odgc. Sixteen vanes had Srnll 11 llxial cracks in ~,c trailing edge of tho airfoil.

During machining of the shrouds. the loading edges of the Stage 2. vanes were notched inadvertently. These notches orc indicated in Figure 92. This is an nrea where cracking occurs in the original production vane. None of these notches in the vane leading edges progressed into cracks, indicating that the improved single shank Stage 2 vane design has successfully solved the leading edge cracking problem. All of the Stage 2 vane segments are con- sidered serviceable and will be returned to test.

Stage 2 Nozzle Assembly nnd Shrouds - Overall views of the Stage 2. nozzle assembly (Figure 82) showing the posttest condition of the Stage 1 and 2 shrouds are presented in Figures 93 and 94. While neither shroud was rede- signed for the improved (single shonk) turbine, the condition of the shrouds was significant.

The shroud distress observed after 1000 accelerated endurance cycles in the core engine, with the single shank turbine, indicates the severity of the test. These shrouds are identical to original production parts. According to Service Engineering personnel and to airline powerplanc engineers who examined the alngle shank turbine piirts following the test, original production shrouds would be in a similar condition after 2500 to 3000 hours of typical airline operation.

Stage 2 Nozzle Support - All cooling holes in the Stage 2 nozzle support (shown in Figure 82) were open and clean. The thirteenth stage air seal showed some inward distortion at 12 o'clock. No hot spots were observed. The thirteenth stage seal, which fits around the nozzle support, is not a full hoop, since it mUst be slit to be assembled to the support. This slit, even though it has a seal underneath it, allowed large amounts of aluminum to get through to the second !:ltage vanes. This slit in the thirteenth stage seal and the aluminum particles in the Stage 2 vane cooling air inlet (spoolie port) are shown in Figure 95. On future builds, this slit will be welded shut, ~nd improved design will be incorporated on production parts.

Inner/Outer Fishmouth Seal - The location of the inner and outer fish- mouth seals can be seen in Figure 82. Aluminum deposits were found along the aft face of the inner fishmouth seal. Wear patterns on the cylindrical area and bowing of the vertical flange indicated good contact with both the com- bustor and the nozzle forward flange. Wear patterns on the outer fishmouth seal ap~eared normal. Both ~\e cylindrical and the radial portions of the seal Ilre bowed.

....

VI .- u u c:: > N o c:: o ....

...

"C c:: o u ...

~

- u

U ....

...

'tl o ~I .r:.

VI .....

VI M Fi '\\ 1 1' q) .

I

I

f

I

J

I

[n"tn/Outer Scr.~~n - 'l'he lactat Lon of thj!' innor an" OlltClt' IH'I~t?(lnB cnn be ~ l tW('I1\ in Figllrt 82. A 7-inch long ChCUlIIhl't'llt int crltck was (ound along the

J odgtll of the hmer iHll'eel,-to-(llange weld ml showll in }l'igure 96. 'l'his weld was

of ,)001' quality, au noted ,>rior to bui.ldull of the enKine. StlJdies ~aro ul1der- WilY to iml)1'()Ve this joint by l'cdesignin.! the joint configurltt iem, In the i,,- stllHed posit ion, the weld crnck al>peued tight, but llluminum Wit8 found in

f

tho StllgO 1 vane inner plut form holes during teardown. Some waviness of the screon .flange wag evident) as well as depos its of duminum on ita forward side. Thi8 mi~lt also account for 4luminum I>a$sing the screen.

'rwo radial cracks were found in tho Otiter screen lift: sheetmet/al l'ing.

"oth cl'ucks lll'pollrcd to start: in the wold joint betweell the macillned \'in8 lHHI the aheetmetal iml)ingement I)tate vert ica L Wllll. One cl'ack 8tal't8 in the wt:'ld and tC'rminntes in ia cooling hole in the l/ltH row of impingement holes. The radial sed riveted to tho forward mllchined dng and t:'xtending fr01ll t:h~ outer IHlnel lo the :;creen has 19 l'udial crllcks lll,proximlltely 6 111111 (0.250 inch) long.

'1'h(,1'm.\ll grndients in the selllare the SliRpect calise. O(\sign l1lodiHcnt.ion ,,,Hl be> I1Indt' if continued testing indiclltes these crack/;! m'(' II problem. 'rhe sct.'el!ln is serviceable as-ia.

Channel Sl,ring - '1'he clHHUlt'll spl'ing, holding the i.l1llcr fishmoll th seul., can bp sepn in FIgure 82. Li~lt scuffing was noted on the contact surfaces of the springs; but otherwise, the springs Wlare in excellent condition. ~Iating wellr nlllrks on the illlltn' screen flange indicate no circumferentilll motion of the inner screen relative to the spring.

Clllumel Cover - 'rIle channel COWt', shown in Figure 82, provides windage cowrllge for the wheelspace baffle bolt circle nuts. Som(" of the nut plate nuts appeared bound in the untorqued position but couLd be £toe ad with 1I light blow. No other distreas W~la noted.

\~hcelspace Ilaffle - The wheelspac(' baf.fLe, shown in ~~igl1re 82, discoll1'- I.IgPR rl'circulacion of floWI)llth gaaes into the cavity. 'fhis reduces the purge nir required and parasitic losses. No difficulties were found with the basic b~lrfh~. lIowever, two baffles d1'i11('(1 for pressul'l:' tups neal' the outer edgl' hud t'tlClinl cracks extel)cling frllm tlw hl~l(\ Co thl~ l>c1gl·~. These holes will not b~ included in the production design.

'111crmni Shield (Stage 1 VmH1) - 'l'hircct:'11 of 16 clH'l'nml shields (Figurl' HZ) IHld radid cracks through t.heir locating pin holes, tlH shown in l~iglll'e 97, and two shields had circumf('rential Ct'lIC1c.S above the bu1t holes. Thet'e were two shields which rubbed the Stage 1 blade (lItgel wings. tn the first ClUH', one-half of the lip WllR missing from t.he bolt holes outward lllld i.n tht' sec()nd cnse, 1132 IIl1n 0.25 inch) length W~1S rubbed through. A design solution for this problem 11115 been implemented and tested in subsequHnt endurance tests.

1~c deSign modification consisted of reducing tile thermol stresses on the heat shi(,'lds by mnkina them half as lonA and adding damping by IllclkillB the shields nn interference fit with the underside of the Stage 1 vane.

IJl ,....j u > u o o o ,....j u '"' u u o o ~ ."

c:

"'

lnteratage Seals - Disassembly of the intentage seals (shown in Figure 82) from the Stage 2 vane segments required considerable force to overcome sliding friction, indicating that engine operation had not opened up clear- ances between mating slJrfaces. Subsequent examination of mating surfacel!l revealed no fretting or wear. Axial Cr4cks were found in the aft flange at the tangential load stops on 5 of the 11 segments. Three of these also had very small crack$ at the forward load stops. Wear grooves in the honeycomb seals were 6 mm (0.25 in.) wide Py approximately 1.5 mm (0.060 in.) deep.

This indicates an axial excursion of the seal relative to the rotor of at least twice that of a standard OF6-6 engine.

,~urbine Stator Summary - The Stage 1 vane preuure side cracking is a problem which was not expected on Chese parts but which was addreued in sub- sequent testing. Work on alternate cooling configurations is complete and modified parts have been tested and introduced on production parts. The Stage 2 vane redes ign appears to have overcome tile lending edge cracking at the outer band but has generated new problems believed to be related to thermal gradient strell8es. Although this test indicated that some changes will be re- quired to improve interstage seal durability in the load stop areas and to facilitate assembly and disasaembly, the boltless interstage seal concept was proven feasible. The Stage 1 vane thermal shield has been redesigned. An inner screen with an improved quality weld has been subsequently tested with no evidence of cracking after 1000 core engine cycles. Virtually all parts in the stator are judged to be serviceable. Altnough some parts will be re- moved for cutup and rework, the parts in the most advanced state of deteriora- tion will deliberately be used to demonstrate total useful life.

9.0 ECONO~tIC ASSESSME!!f.

!

i ( The UP'l' (lt~r()dynamic performance improvement concept Wlla evaluated by Douglas lllUle[ 'l'ask 1 of this progralll (Reference 1) for a cruise. specific fuet consumpti.on improvement of t.3%. '11,e C .. '6 turbofu.n engine static back-to-back test demonstrllted this I)(~r formllncc> i.mprovement nt cruise equivalent power for new engines.

Irhe 1. 3% rc>dllctton in cl.'u:lse arc rf~sul ts in till' hloc\, fuel snvj,ngs shown i.n Table V for the 111 i.nil1HII1l fuel consumpt ion mission. l1ds is based on the datil of Reftlrence 1.

Table V. High Pressure Turbine Aerodynamic JlerfOtlllllnCe Improvement Block Fuel Snvings.

.1:. ~'ue 1 Rllilge Airc raft (km) kg %

-

l)C-1O-10 -1.3 645 -103.4 -215.0 -1.3 3700 -4/.9.5 -1.'.

The estimated ltnlllllll fu('.l. savings pe,r "lrCrllft for the "bove block fuel SllV Lng s arc shown in Tab le VI.

Table VI. lIigh Pr(>ssure Turbine AC'rodynlllnic Perfotll1ancc Improvement-gstimated Annual Fuel Savings Per Aircraft.

R~l\\ge A Fuel Airc raft (knl) 1/ AC/Year 00-10-10 645 276,000 1690 381,800 3700 426,000 The economic nss(;~ssment for thl' MSlulled medium fucl prLce of 11.89Uliter (45Uglll) LS SUl1Ul1llrized in Tnble VII. '11,c inct"cllsl' in funl cost to about double the IIbove medium price 24Ultter (90Uglll) l"educes the pllybllCk time to 0.105 YURT which mnkos this concept ~v('n more nttroctivo.

141.

Table VII. Economic Assessment of liP Turbine Aerodynamic Performance Improvement Concept.

(Medium Range, ~fedium Fuel Price, Minimum Fuel Analysis) PllybliC k ROl Aircraft Years % DC-lO-10 600 0.17 10.0 SUMMARY OF RESULTS

i1

~ : The high pressure turbine aerodynamic performance improvement concept I has been evaluated in component tests and in engine ground tests. The main

•

result. of the.e te.ts are di.cussed below: Component Tests Stage 1 Vane Cascade Test - This test demonstrated that the fully cooled improved turbine Stage 1 vane-has an aerodynlll\ic efficiency equal to the fully cooled Qriginal production vane.

Stage 2 Vane Leading Edge Flow Evaluation - This test demonstrated that the improved design vane insert inlet area reduced the static pressure loss, resulting in a reduction of the radial pressure gradient within the inlet.

Larger trailing edge holes are required to achieve the design flow split.

Stage 1 Vane Trailing Edge Test - This test demonstrated that the selec- ted improved vane design of two walls promoted in the staggered matrix has 2.44 times the heat transfer promotion than all walls smooth. The selected design reduces trailing edge temperature by 16.7- C (30 F) with a temperature in- crease of only 5.6 C (10· F) in the promoted region.

Stage 2 Vane Leading Edge Test - The test showed that the improved design vane insert results in a 10% increase in Nusselt number in the leading edge regi()O. This indicates an improvement in the cooling of the vane leading edge.

Stage 2 Vane Trailing Edge Pin Fin Test - This test demonstrated that the selected trailing edge pin fin geometry of the improved Stage 2 vane design increased the heat transfer coefficient by a factor of 3 over that of a smooth wall. " Blade Dynamic and Steady=State Strain Distribution - Resonant frequencies and nodal patterns were obtained for the Stage 1 and 2 improved design blades.

Relative strain distributions were determined for all modes, giving the de- tailed distribution of stress in the bl~des. The steady-state strain distri- butions obtained showed no unusual effects for either Stage 1 or Stage 2.

The magnitude of the end effects is wi thin the realm of experience for blading of this type.

Blade Frequency and Amplitude as a Function of Damper Force - The results of the test indicate that the dampers will produce the desired effects re- lating to frequency gain and stress red uc tion. A gain of about 11% in the first-flex frequency was found for both Stage 1 and Stage 2 blades. The Stage 1 blade could not be driven at high amplitudes which was expected. The magni- tude of the Stage 2 stress reduction was similar to that seen for the CF6-50 Stage 2 blade.

Turbine Disk Rim Stress Distribution To!!

The stress concentration factors of the forward and aft rabbet fillets of the Stage 1 and Stage 2 turbine d iak rims wete determined from this teat. The Stage 2 results are in good agreement with the two-dimensional analysis while the Stage 1 results show some differences.

Instrumented Engine Test this test determined the operating characteristics of the improved (single shank) turbine, such as: Stage 1 Blade Vibratory Response

•

Stage 2 Blade Vibratory Response

•

Rotor Vibratory Response

•

Temperatures Stage 1 Blade Metal

•

2 Blade Metal Temperatures Stage

•

Spool Metal Temperatures Rotor

•

Rotor Cooling Flow

•

2 Nozzle Cooling Flow Stase

•

!nterstage Cavity Temperatures

•

Rotor and Stator Transient Temperatures

•

Stator Structure Temperatures

•

Stator Structure Transient Response

•

The results were used to conduct additional life and design analyses.

Endurance Test The 1000 cycle core engine endurance test on engine 441-019 fulfilled the test objectives of subjecting the new turbine design to the equivalent of 2500-3500 hours of airline service. A number of problem areas were uncovered but, in general, the posttest condition of the turbine was excellent. In the areas where problems were uncovered, modifications have been defined and pro- grams were put in place to procure hardware, perform analysis, and evaluate the mod ification effec ti veness via addi tional engine and cOOlponent testing.

No major design flaws were uncovered by the test which would hinder COOl- pletion of the overall program on schedule. The 1000 cycle endurance test was successful in fulfilling all objectives in quickly identifying potential field problems. Basic integrity of the new turbine design was established.

I

Further Development

J

As a result of these tests and additional General Electric funded efforts, the development and certification of the imp~oved (single shank) turbine wue continued. Initial back-to-back engine tests of the original and the irnproved turbine demonstrated an improvement of .1.3% in cruise sfc and a lOCI C reduction in exhaust gas temperature (EGT). An additional improvement of 0.3% in cruise sfe and 6° C on EGT is projected for long service engines.

The Bingle shank turbine will be used in the advanced versions of the CF6-6 engine such as the CF6-6K and the CF6-32.

r

I

J

gUALIT~ ASSURANCE

II

INTRODUCTION

f

The quality progra.m applied to this contract is a documented system throughout the dcsiBn, manufacture, and repair, overhaul, and modification

r

cycle for gas turbilte aircraft engines. The quality system has been con- structed to comply with military specifications MIL-Q-9858A, MIL-I-45208, and HIL··C-45662 and Fed/~ral Aviation Regulations FAR-145 and applicable por- t,ion~ of FAR-21.

~

The quality system and its implementation are defined by a complete set '-' of I)rocedures which has been coordinated with the 000 and FAA, and which has their concurrence. In addition, the quality system as described in the qual- ity program for this contract has been coordinated with NASA-Lewis Research 'i ,. Center. The following is a brief synopsis of the system.

1# QUALITY SYSTEM The quality syotem is documcnted by opcrating procedUres which cOQrdi- nate the quality-related activities in the functional areas of Engineering, Manufacturing, Materials, Purchasi.ng, and Engine Programs. The quality sys- tem is a single-standard system wherein all product lines are controlled by the common quality system. The actions and activities associated with deter- mination of quality are recorded, and documentation is available for review.

Inherent in the system is the assurance of conformance to the quality re- quirement~. This includes the performance of required inspections and tests.

In addition, the system provides change control requirements which nssure that design changes are incorporated into manufacturing, procurement and quality documentation, and into the products.

Heasurihg devices used for product acceptance and instrumentation used to control, record~ monitor, or indicate results of readings during inspec- tion and test are initially inspected and calibrated and periodically are reverified or recalibrated at a prescribed frequency. Such calibration is performed by technicians against standards which are traceable to the National Bureau of Standards. The gages are identified by a control number and are on a recall schedule for reverification and calibration. The calibration func- If tion maintains a record of the location ot each gage and the date it requires recalibration. Instruc tions implement the provisions of MIL-C-45662 and the , appropriate FAR requirements.

Work sent to outside vendors is subject to quality plans which provide for control and appraisal to assure conformance to the technical requirements.

Purchase orders issued to vendors contain a technical description of the work

i

to be 'perfonned and instructions relative to quality requirements.

I

I

Engino parts are inspected to documented quality phnllwhich dofine tho characteristics to be inspected J the gatos and tooLs to bo uscd J tho condi- tions under which the inspection is to be pcrfonncd, the sampling phn, labo- ratory and .pecht procolJI testing, and the identification and 1'ccord require- ments.

Work instruc tions are bsued for compliance by operators, inspectors) testers J and mechanics. Component part manufacture provides for laboratory overview of all special and criticaL processes, including qualification and certification of personnel, equipment, and processes.

When work is performed in accordance with work instructions, the opera- tor/ inspector records that the work has been performed. This i8 accanpl ished by the operator/inspector stamping or si.gning the operation sequence sheet to 8ignify that the operation hllS been performed.

Various designs of stamps are used to ind icate the inspection status of work in process and finished items. Perfomance or acceptance of special pro- cesses is indicated by distinctive stamps assigned specifically to personnel performing the process or iuspection. Administration of the stamr system and the Lssurance of stamps are functions of the Quality Operation. The stamps are applied to the paperwork identi fying or denoting the items 1:equir ing con- trol. When stamping of hardware occurs, only laboratory approved ink is used to assure against damage.

The type and location of other part marking are specified by the design engineer on the drawing to assure effects do not compromise design require- ments and part quality.

Control of part hand ling, sto'rage and delivery is maintained through the entire cycle. Engines and assemblies are stored in special dollies and trans- portation carta. Finished assembled parts are stored so as to preclude damage and contamination, openings are covered, lines capped and protective covers applied as required.

Nonconfoming hardware is controlled by a system of material review at the component source. Both a Quality representative and an Engineering representative provide the accept (use as-is or repair) decision. Noncon- fomances are documented including the disposition and corrective action if applicable to prevent recurrence.

The system provides for storage, retention for specified periods, and retrieval of nonconformance documentation. Documentation for components is filed in the area where the component is manufactured/inspected.

A buildup record and test log are maintained for the assembly, inspection, and test of each major component or engine. Component and engine testing is performed according to documented test instructions, test plans, and instru- mentation plans. Test and instrumentation plans are submitted to NASA for approval prior to the testing.

r

Records (uuQntial to the economical and effective op~ration of tht" quality program ~ra maintained. r~viawcd, and used as a bauis for action.

The.e record. include inspection and tQst results, nonconfonning material findina', laboratory anlly.i., and receiving inspection.

:1

'f ~ f.

APPENIHX n

1. It'asehing, W.A., "CF6 Jet Engine Perfomlance Improvement Program - Task 1 Feasibility Analysis," NASA CR-1S9450, March 1979.

Kreith. F.. P_RINCIJ>LES.J,lF HEAT TRANSFER, International 1.'cxt Book Company, 2.

1958.

APPENDIX C SYMBOLS AND l>I~FINI'1'I0NS A Cross-sectional Arca, m Ac Aircraft All Spec ifie Ilent at Constant PreIHWrQ, W hr C p kg e C I)p Pin Uiaml>I:er J m h Heat Transfer Coefficient, W I Electrical Current, A k Conduc tiv ity, ",2 0 G/m Pin Length/2, 111 Lp Pin Perimeter, m P p Electrical Power Input, W

Pr Prandtl Number = ~

k Tota L PreSSlIr(', N/ cm Por Total Ptn Fl.n IINlt L(lI-,s 1"01" Eneh llCllt(.'1,', Iv Q p Electrical Resistance, 0 R Reynold s Number (bnsed on dimnetcd, P V(> D ReO \I Air 'remperlltllrt~, Q C 'fA Thermocouple T/C

Hent Plate Temperature ,0 C

Tn Pin 'l'cmpcrntur(' J U C T p Loelll Velocity, m/s('c.

Vl We CooLing Airflow, % of COmpressor Flow p 'l'\lrbine l~rficl('n('y, % Villcodty, ~ \1 hL' III