Aircraft Engine Diagnostics

- -- -

NASA Conference Publication 2190

Aircraft

Engine

Diagnostics

A conference held at Lewis Research Center Cleveland, Ohio May 6-7, 1981

NASA Conference Publication 2190

Aircraft

Engine

Diagnostics

A conference held at Lewis Research Center Cleveland, Ohio May 6-7, 1981

NJ\S/\

National Aeronautics and Space Administration Scientific and Technical Information Branch PREFACE During the past four decades, the Lewis Research Center has been providing advances iri aeronautical propulsion from the research activities of its staff and its university and industrial grantees and contractors. These advances have helped create the preeminence in aeronautics that has con- rributed to our national defense, has provided swift and reliable transportation for our people and their goods, and has greatly aided our position in international trade. In recent years substantial resources have also been directed a t improving our nation's utilization of energy.

NASA is well aware that the aviation industry is an important segment of our national economy. In 1979 aircraft sales led all U.S. manufacturing industries with a trade surplus of over $10 billion - without which the country would have experienced a one-third greater trade deficit. This favorable balance attributable to the aircraft industry is largely a result of being able to provide a superior product and to continue to upgrade the product. Efforts at improvin g the performance retention of today's and future engines which will power commercial and military aircraft represent a pos- itive step toward this end.

To provide to representatives from government, industry, and universities the latest findings directly related to improved aircraft engine performance retention, a two-day conference was held in May 1981. This publication contains the papers presented at that conference.

John F. McCarthy, Jr.

Director iii CONTENTS Page iii PREFACE ENGINE DIAGNOSTICS FOR PERFORMANCE RETENTION CF6 JET ENGINE DIAGNOSTICS

/'

Ron Stricklin, General Electric Company . . . . . . . '.' . 1 CF6 HIGH PRESSURE COMPRESSOR AND TURBINE CLEARANCE EVALUATIONS M. A. Radomski and L. D. Cline, General Electric Company. .

JT9D ENGINE DIAGNOSTICS PROGRAM W. J. Olsson and W. J. Stromberg, Pratt & Whitney Aircraft Group ...

JT8D ENGINE PERFORMANCE RETENTION Albert D. James and David R. Weisel, Pratt & Whitney

63v/

Aircraft Group.. . ............... .

PERFORMANCE RETENTION OF THE RB2ll POWER PLANT IN SERVICE B. L. Astridge, Rolls-Royce Ltd., and J. T. Pinder, Ro lls- Roy ce Inc. . . . . . . . . . . . . . . . . . . . . . . .

PERFORMANCE DETERIORATION - AIRLINE PERSPECTIVE Niels B. Andersen, Pan American World Airways, Inc ......... .

ENGINE DURABILITY AND PERFORMANCE RETENTION CONCEPTS IMPROVING TURBINE ENGINE COMPRESSOR PERFORMANCE RETENTION THROUGH AIRFOIL COATINGS 109/ L. A. Friedrich, Pratt & Whitney Aircraft Group ADVANCED OXIDE DISPERSION STRENGTHENED SHEET ALLOYS FOR . IMPROVED COMBUSTOR DURABILITY

119~

R. J. Henricks, Pratt & Whitney Aircraft Group.

ADVANCED TURBINE BLADE TIP SEAL SYSTEM

137 V'

J. W. Zelahy, General Electric Company .......... .

AN INTRODUCTION TO NASA'S TURBINE ENGINE HOT SECTION TECHNOLOGY (HOST) PROJECT Daniel J. Gauntner and C. Robert Ensign, NASA Lewis

153 V

Research Center . .

THE NATURE OF OPERATING FLIGHT LOADS AND THEIR EFFECT ON PROPULSION SYSTEM STRUCTURES Kenneth H. Dickenson and Richard L. Martin, Boeing 175 /' Commercial Airplane Company . . . .

v CONSERVATION OF STRATEGIC AEROSPACE MATERIALS (C OS AM) Joseph R. Stephens, NASA Lewis R esearc h Ce nte r .. 189/ ENGINE CONDITION MONITORING SYSTEMS WHO NEEDS ENGINE MONITORING?

James L. Pettigrew, Wright-Patterson Air Fo r ce B ase . . . . . . . . . 2 09t/ F100 ENGINE DIAGNOSTICS SYSTEM STATUS TO DATE

James A. Boyless, Wright-Patterson Air Fo rc e Ba s e 22 5~

TURBINE ENGINE PERFORMANCE ESTIMATION AND ITS ROLE I N FUTURE SYSTEMS Ronald L. DeHoff, Systems Control, Inc . , and Charles A. Skira, Wright-Patterson Air Forc e Ba s e . . . .

IMPACT OF AUTOMATED ENGINE MONITORING ON RELIAB I LI TY CE NT ER ED MAINTENANCE AND LOGISTICS SUPPORT Laura E. Baker and W . Earl Hall, Jr . , Sys t em s Co n t r ol, I nc. 263

A-10/TF34 TURBINE ENGINE MONITORING SYSTEM (TEM S) - vi'

Robert G. Christophel, San Antonio Air Logi stic s Cent er . . . . . .. 2 71 REVIEW OF AIDS DEVELOPMENT Henk C. Vermeulen, KLM Royal Dutch Airlines , an d 2 85t/ Sven G. Danielsson, SAS Scandinavian Airli nes Syst e m.

HELICOPTER PROPULSION SYSTEM RELIABILITY AND ENGI NE MONITORING ASSESSMENTS 3 11/ John A. Murphy, Bell Helicopte r Textron .. . .

ENGINE HEALTH MONTIORING SYSTEMS - TOOLS FOR IMP R OVED MAINTENANCE MANAGEMENT IN THE 1980 's 323/ Jonathan C. Kimball, Pratt & Whitney Air c raf t Group ENGINE "ON CONDITION" MONITORING - CF6 FAMILY 60's THROUGH THE 80's Harry J. Kent, General Electric Company, an d 341 / Gerwin Dienger, Lufthansa German Airlines ENGINE HEALTH MONITORING - AN ADVANCED SYSTEM

357 ~

R. J. E. Dyson, General Electric Company.

AN OVERVIEW OF SAE ARP 1587 "AIRCRAFT GAS TURBINE ENGI NE MONITORING SYSTEM GUIDE"

377/

John A. Murphy, Bell Helicopter Textron . . .

vi CF6 ENGINE DIAGNOSTICS Ron Stricklin General Electric Company INTRODUCTION The energy demands of the United States far exceed domestic fuel supplies which creates a severe dependence on foreign oil. This dependence was accentu- ated by the OPEC embargo in the winter of 1973/1974 which triggered a rapid rise in fuel prices. This price rise (Figure 1) further compounded by other inflation factors has brought about a set of changing economic circumstances with regard to the use of energy. As a result, our government, with the sup- port of the Aviation Industry, initiated programs aimed at reducing fuel demands. One such program sponsored by NASA is the Aircraft Energy Efficiency Program which is directed toward reducing fuel consumption for commercial air transports. An integral portion of this program is the Engine Component Im- provement (ECI) Program aimed at improving fuel efficiency Of current engines.

This ECI Program consists of two parts, 1) Performance Improvement and 2) Engine Diagnostics.

General Electric is participating in both parts of the Engine Component Improvement Program. As part of the program, performance deterioration studies for the CF6-6D and the CF6-S0 Engine Models have been conducted. The basic objectives of the latter effort were: 1) to determine the specific causes for engine deterioration which increase engine fuel consumption rates, 2,) to isolate short term losses from longer term losses and 3) identify potential means to minimize the deterioration effects. The deterioration studies have been completed and final NASA reports published.

To quantify the effect of engine performance deterioration, the fleet statistics for the CF6 family of engines in 1981 were projected. It is anti- cipated that the CF6-S0 family of engines will amass approximately 3.4 million flight hours and the CF6-6 family over one million flight hours in 1981.

An average deterioration in cruise specific fuel consumption of 1 percent over new engine levels will result in excess fuel consumption of approximately 36 million gallons for the CF6 fleet alone. The effects of small amounts of deterioration throughout the fleet are obviously substantial.

This paper presents a summary of the activities which led to defining deterioration rates of the CF6 family of engines, a description of what was learned and an identification of means of conserving fuel based upon the pro- gram findings.

HOW DID WE DO THE JOB The program to define the deterioration levels and modes for the CF6 famil y of engines involved four distinct phases: analysis of inbound engine test results, analysis of airline cruise data, analysis of airline test cell data resulting from testing of refurbished engines and inspection of engine hardware.

INBOUND ENGINE TESTS Testing of engines, removed from aircraft after extensive revenue service, was conducted in order to define, on a specific engine basis, how much specific fuel consumption had increased and provide some insight into which components were the prime contributors to the observed deterioration.

Through the CF6-6 and CF6-50 phases of the program, 15 inbourld engine tests were conducted. One of these tests conducted as part of the CF6-6 Pro- gram, was specifically accomplished to identify short term losses.

For . each of the inbound engine tests, sufficient instrumentation was installed to measur~ overall engine deterioration and to indicate the magnitude of deterioration of each major component. After the inbound tests had " been conducted, three of the engines were subjected to a detailed teardown inspec- tion by design en g ineers to relate hardware cbndition to inbound test results.

CRUISE DATA ANALYSIS Inbound engine tests, however, which are conducted on specific engines, yield only limited information concerning the degradation in performance of the average fleet. Recognizing that the intent of the Diagnostics Program was defined to determine the deterioration characteristics of the typical CF6 engine in revenue service, it was concluded that analysis of fleet performance data accumulated during flight was the best means of accomplishing this objec- tive.

Data from many airlines are supplied to General Electric on a periodic basis. These data are supplied in many forms from logs recorded by flight engineers in the cockpit to data recorded via automatic data acquisition sys- tems. These data, which are in general recorded during every flight of an aircraft, were used to define the deterioration characteristics of individual engines during the life of the engines during a given installation period. The process to define the performance trend was to compare the performance indicat- ing parameters (fuel flow level and exhaust gas temperature level) to a reference engine parameter level at the flight condition and power setting.

Data from five airlines using CF6-6D engines and from 9 airlines using CF6-50 engines were reviewed. In all, data from 239 CF6-6D engines and 263 CF6-S0 engines were analyzed in defining deterioration rates of initial insta1~ 1ation and multiple installation engines in revenue service.

General Electric obtained and analyzed data recorded at cruise during initial ~ircraft checkout flights conducted by ~he aircraft manufacturer to determine if performance degradation occurred within an engine prior to initial revenue service. Data from 82 CF6-6D engines and data from 111 CF6-S0 engines were analyzed in order to determine the magnitude of any "Short-Term" deter- ioration of engine performance prior to airline receipt of the aircraft and engines. As will be discussed in more detail later, it was concluded after this analysis that significant deterioration did occur during these aircraft checkout activities.

The CF6-6D engine removed from a DC-IO aircraft and subjected to an in- bound performance run verified that the indicated loss based upon cruise data analysis was indeed real and non-reversible. As mentioned. this engine was disassembled and critically inspected by a team of General Electric engineers to define the area of performance degradation. Another engine, removed early after entrance into revenue service due to vibration problems, was also tested inbound and similarly confirmed that the short-term loss of performance was real.

AIRLINE CELL DATA ANALYSIS An important part of the analysis effort to understand airline fleet engine performance levels centered around the definition of basic engine per- formance levels after overhaul in the airline shops.

Performance levels were reviewed for engines outbound after overhaul at a major airline overhaul facility during the CF6-6D Program and at one consortium central agency and 5 other overhaul facilities during the CF6-S0 Program.

Performance levels from these facilities were compared to new engine per- formance levels from the General Electric Production Facilities in order to define the effectiveness of typical engine works copes in restoring performance by refurbishment to new engine levels.

COMPONENT DETERIORATION MECHANISMS The actual modes of deterioration were identified by hardware observation by General Electric teams. Teams of Mechanical and Aerodynamic Design person- nel visited various maintenance facilities and conducted detailed inspections of the various engine modules in the disassembled stage to assess the condition of component parts relative to the condition of new hardware. Observations of rotor clearances, surface finishes of the airfoils, cleanliness and smoothness of various static structures and potential air leakage paths were reviewed and yielded estimates of component performance relative to a non-deteriorated com- ponent.

Hardware from each major module at various stages of engine life was ob- served, thus allowing estimation of the deterioration associated with any module degradation mechanism as a function of time and cycles.

Combination of the trends established for each module degradation mechan- ism yielded module performance deterioration trends. Combination of the module deterioration characteristics using appropriate knowledge of the engine cycle then led to establishing overall engine deterioration characteristics. In all cases throughout both the CF6-6 and CF6-50 Programs, the estimates of engine deterioration established based upon hardware examinations showed excellent agreement with the overall deterioration rates established by cruise and cell data analysis.

WHAT WAS LEARNED Figure 2 shows the resulting assessment of CF6-6D performance deteriora- tion characteristics for the typical engine thru its initial installation and experience in review service and for the same typical engine after several multiple installations. Each of the elements of deterioration is presented ip Figure 2 for the CF6-6D engine. This Figure shows equivalent cruise specific fuel consumption increases relative to a production new engine. The initial installation is shown on the left. Engines incur an average Short-Term loss of 0.9 percent prior to revenue service. During their initial installation, SFC increases an average of 1.7 percent based on the 4000-hour family of engines. The total increased SFC of the deteriorated engine is thus 2.6 per- cent from production new. Insufficient data is available to determine the amount of performance restoration during the first shop visit.

During the "nth" installation, the serviceable engine re~enters revenue service after a shop visit with an average unrestored cruise SFC loss of 2.1 percent. During revenue service, the cruise SFC of this multiple-build engine increases 0.9'percent for the 3000-hour engine. The total increased SFC of this deteriorated en " gine at 3000 hours was 3.0 percent from new. On the aver- age, 0.9 percent cruise SFC is restored during the shop visit. During the next installation, an average revenue service deterioration of 0.9 percent is incurred. This amount is restored on the average during maintenance and the cycle is repeated.

Though engine-to-engine variations within this cycle are significant, the data presented reflects the typical or average engine deterioration character- istic for the CF6~6D engine.

Figure 3 shows the deterioration characteristics resulting from cruise data analysis of data obtained for the CF6-50 engine on various aircraft.

General findings of the program were that the Short-Term losses which occurred dur " ing the airframer checkout of the aircraft tended to be the same for opera- tion on all three aircraft. Also, the unrestored performance level of the multiple-build engine as refurbished by the various airlines was essentially the same. It can be noted from Figure 3 that the deterioration rate shown during typical 747 operation was lower than observed with DC-10 and A300 operations. The same relationship holds for both the initial and mUltiple installations. The unrestored SFC of the typical engine re-entering revenue service after airline shop visits is 1.8 percent poorer than the new engine baseline for the CF6-50 engine as compared with the 2.1 percent determined in CF6-6D analysis.

The deterioration rates shown on Figures 2 and 3 are presented as a func- tion of flight hours since installation. An analysis was conducted as part of the CF6-S0 program, to understand the variability in deterioration rates which resulted from analysis of DC-10, 747 and A300B data. The conclusion was that deterioration rates for the data surveyed was most strongly influenced by average flight length per cycle and the amount of derate or reduced power being used by the individual operators. Table 1 shows the data from Figure 3 trans- lated into the deterioration rate per 1000 cycles basis. The conclusion is that while the DC-10 and 747 data are reasonably consistent and show approxi- mately the same deterioration rate per 1000 cycles, the A300B data shows a much lower deterioration rate per 1000 cycles. Since the A300B data studied as part of this program were consistent with flight cycle lengths of approximately 1 . 9 hours, the lower deterioration rate per 1000 cycles suggests that deterioration rates are not only influenced by numbers of cycles but also time at tempera- ture.

Figure 4 illustrates the results of the hardware inspection analyses and the resulting deterioration model compared to the performance-data - derived deterioration level for the CF6-6D initial installation. It shows that the largest portion of the 0.9 percent Short-Term SFC loss resulted from High Pressure Turbine (HPT) performance losses. This loss was due largely to HPT clearance increases during the initial checkout phases of the airplane. During initial operation of the aircraft by the aircraft manufacturer, there is little attendant loss in fan, high pressure compressor and low pressure turbine. It is also to be noted that the combined performance deterioration level created by the stackup of the individual component deterioration losses at 4000 ho u rs shows 2.3 percent total performance degradation from the "'as new" condition compared to the 2.6 percent level which resulted from performance data analy- sis.

Figure 5 shows the deterioration mechanisms as assessed by hardware in- spection for the CF6-6D multiple-build engines. The major deterioration of a multiple-build engine is within the HPT module. Typically, HPT performance is restored during every shop visit while fan, HP compressor and LPT performance levels are not. Therefore, each engine as it re-enters revenue service after an overhaul shop visit has new HPT hardware and somewhat deteriorated fan, HPC, and LPT performance levels. It can be noted again from Figure 5 that the results of the hardware inspection show 3.3 percent performance loss at 3000 hours on multiple-build engines compared to the performance data analysis level which indicated 3.0 percent. Again, agreement is good. Similar findings for losses associated with the initial installation and the mUltiple installations of CF6-S0 engines resulted.

Of prime importance to the program was the finding that the unrestored loss f or the typical engine out of the overhaul shop, based on hardware in- spections, was 2.08 percent in terms of cruise SFC compared to the 2.1 percent unrestored performance level as identified by performance data analysis.

Figure 6 describes the component breakdown for both CF6-6D and CF6-S0 engine models as shipped from the airline overhaul facilities compared to the new engine performance levels. It shows that, of the 2.1 percent unrestored per- formance for the CF6-6D engine and 1.8 percent unrestored performance for the C F6-S0 engine, a large portion of these performance losses are due to lack of performance restoration in the fan area with lesser amounts of the performance loss associated with the high pressure compressor and the LP turbines. Note that there is very little performance left to restore in the HPT area for typ- ical outbound engines, again, this is due to the fact that HP turbines are typically completely refurbished during shop visit. Again, the hardware inspection data and the performance data show excellent agreement.

The unrestored performance identified in Figure 6 represents a potential gold mine in terms of f uel and dollars savings to the airlines, if it can be reduced on a cost effective basis . The presence of large amounts of unrestored pe rformance associated with performance degradation of the fan module, HPC module and the LPT module, relative to new modular performance levels, is due to early workscope definitions. These airline shop overhaul work scope defi- nitions were primarily aimed at maintaining reduced EGT levels and at restoring the condition of the hardware primarily from a reliability standpoint. The engine modules, which have the most direct impact on EGT margin and direct impact on reliability, are primarily associated with the hot s" ection of the engine~ the combustor and hig h pressure turbine area. Larger efforts (dollars and manhours) are required to achieve the same amount of EGT margin restoration in the LP system components than in the HP system components . In the early 1970's, it was concluded that it was not cost effective to do significant ver- formance restoration in the fan and LPT areas with fuel prices at a 30 ce~ts per gallon level. With current and projected fuel prices, the cost l effective- ness of doing performance restoration work in all of the engines' major com- ponents must be re-examined.

HARDWARE INSPECTION DETAILS The prime modes of deterioration within each module were established primarily by design team inspections at two major CF6-S0 overhaul facilities and at one major CF6-6 overhaul facility. The details of the findings of these inspections are identified in references 1 and 2, including identifica- tion of the amounts of cruise SFC increase associated with each deterioration mechanism. However, some general statements concerning the more significant deterioration mechanisms are in order .

FAN DETERIORATION The major areas of performance degradation within the fan section for both engine models were: 1) increases in tip clearance due to shroud erosion and the current maintenance philosophy which requires controlling only minimum clearance; this can result in local grinding and, in turn, results in increased shroud out-of-roundness and increased average clearance, 2) fan blade leading edge bluntness due to erosion and 3) fan bypass OGV erosion and leading edge bluntness due to loss of the polyurethene protective coating. During typical shop visits, the leading edge contours of the stage one blades are typicall y restored (with approximately 75 percent frequency). However, the "on-condi- tion" maintenance philosophy, requiring only durability repairs, generally results in very little refurbishment to restore to new engine average clear- ance and to restore the OGV surfaces to the as new condition.

HPC DETERIORATION The major deterioration modes of the CF6 engine high pressure compressors ar e : 1) increases in airfoil tip clearances, 2) degradation of airfoil surface finishes and leading edges and 3) creation of airflow leakage paths primarily through the variable stator vane bushings. With increased time in revenue service, the assembly of engine compressor stator cases (as engine parts are interchanged during shop visits) develop significant tendencies toward out-of- roundness. The maintenance philosophy in matching rotors and stators is to establish a minimum clearance. Thus, any tendency of the stator case to dis- tort inward creates the requirement for short rotor blades (with resulting increased average clearance) and locally short stator vanes. The eventual result is increased airfoil tip clearances and associated deteriorated per- formance. Casing distortion and design changes which will result in a reduc- tion in casing distortion is the subject of a separate paper at this confer- ence.

HP TURBINE DETERIORATION The primary mode of deterioration noted in the high pressure turbine during revenue service is the increase in blade tip-to-shroud clearances, resulting from rubs with some losses in performance due to increased airflow leakage and airfoil surface finish degradation.

It has been found that tip clearances for both stages of the CF6-50 high pressure turbine typically increase during the first 1500 hours of operation and continue to increase, but a a lower rate, thereafter. At 4000 hours, the average increase in tip clearances is 0.013 inch on stage 1 and 0.011 inch on stage 2, which accounts for 0.55 percent increase in cruise specific fuel · consumption. These rubs and resulting clearance changes are primarily due to shroud support distortion, shroud swelling and bowing, shrinkage of the shroud supports and thermal mismatch between rotating and static structures during engine transients. Again, shroud distortion is the subject of another paper at this conference.

LP TURBINE DETERIORATION As in the case of the high pressure turbine, increases in blade tip clearances and interstage seal clearances result in the major portion of deterioration occurring within the low pressure turbine in service. Degrada - tion of airfoil surface finish is another contributor but results in very little performance loss. The increase in clearances was found to be primarily due to wear of the stationary surfaces which result from engine axial mis- matches during different phases of engine operation. While there is little loss of material from the rotating components, the wear of the tip shrouds and interstage seals results in approximately 0.4 percent loss in cruise SFC after 4000 hours of operation with both the CF6-6 and the CF6 - 50 turbines.

SUMMARY OF DETERIORATED ENGINE As is evident, a large part of degradation of engine performance in revenue service results from rubs and subsequent increases in clearance in the high pressure compressor, the high pressure turbine and the low pressure tur - bine.

Considerable effort is being expended by General Electric and the other engine manufacturers to create functional clearance control systems designed _ to eliminate rubs in these components and to maintain optimum clearances at the required cruise condition to maintain peak engine performance.

USE OF WHAT HAS BEEN LEARNED The objectives of this part of the Diagnostic Program were to: 1) deter - mine the specific causes for engine deterioration, 2) to isolate Short - Term losses from the longer term losses and 3) to identify potential ways to mini- mize the deterioration effects. Two potential means are available for mini- mizing deterioration effects on the current fleet. First is identification of product improvements which will provide better performance retention character - istics in the current engine, and the second is to identify improved engine work scopes which can be used by the airlines to improve performance restora- tion and, therefore, absolute performance levels of the engines coming out of the airline overhaul shops.

PERFORMANCE RETENTION As a result of knowledge gained from the Diagnostic Pr o grams, a Perform- ance Improvement and Performance Retention Improvement Program has been identified for the CF6 family of engines. The complications of introducing new performance retention features into an existing engine arises from limitations on changes to aircraft power management and functional interchangeability.

However, some features are currently planned by General Electric for introduc - tion into the CF6-50 engine production models and will be retrofitable within the current fleet. Items being considered which can be included into the cur - rent fleet of engines include: smooth solid shrouds in booster stages 1, 2 , and 3, which result in reduction in shroud erosion and better clearance con- trol; a modified front engine mount, and steel front compressor casing which reduce bending deflections and locally reduce rub potential; improved surface finishes on high pressure compressor blades and vanes; replacement of three stages of titanium stator vanes and four stages of HPC rotor blades with steel which increases erosion resistance; and incorporation of new VSV bushings in the compressor stator case to increase durability and reduce leakage.

Other performance retention features are also being incorporated into the production configuration of the CF6-80 family of engines in addition to the performance retention items just mentioned. The HPC casing is a stiffer, two piece case with insulated rear stages which provides reduced deflections and better roundness thereby reducing rubs. The HPC rotor is cooled by introducing fan air into the bore, resulting in better matching of rotors and stators which again reduces rubs during transients. There will be an improved HPT shroud support system and improved HPT shrouds which reduce distortion and blade tip rubs. A passive cooling system for the HPT stator is being utilized which will provide a better match with the rotor and reduce blade rubs. Also to be included is an active clearance control system in the LPT which will produce close clearances at cruise and larger clearances at takeoff to reduce shroud rubs and prevent deterioration. The deterioration portion of the Diagnostics Program also verified that the performance retention features being designed into the Energy Efficient Engine (E3) Program will have a definite payoff. These performance retention features include: a low tip speed, wide-chord, rugged fan blade; a short stiff compressor case; ruggedized fan OGV's; and active clearance controls on the high pressure compressor, the high pressure turbine and the low pressure turbine. Current estimates are that deterioration rate on the E3 engine should be reduced by 51 percent relative to deterioration rates established for the CF6-50 engine as part of the NASA Diag- nostic Program.

IMPROVED ENGINE WORK SCOPES The most immediate reduction in fuel usage by todays CF6 fleet, which can be achieved as a result of information gained during the NASA Engine Diagnos- tics Program, lies in the definition of improved engine work scopes during engine shop visits by individual airlines. An integral part of the Engine Diagnostics Program with both the CF6-6D and CF6-50 engines were studies con- ducted to define how much of the unrestored performance losses associated with the typical engine as currently shipped from the overhaul test cells could be restored on a cost effective basis. The results of these studies were intended to be used as guidelines for improved definition of modular work scopes at the overhaul facilities.

As part of these studies, assumptions were made which included: material cost as defined by either repair cost or replacement hardware cost established in the General Electric catalogues; estimates of the cost of doing work based upon General Electric experience; performance gains and the life of the gain consistent with the deterioration rates established as part of the Engine Diagnostics Program; and typical missions assumed consistent with DC-lO-lO and DC-IO-30 operation. Fuel price for these studies was assumed to be a dollar a gallon.

It was concluded, based upon these studies, that approximately 60 percent of the unrestored performance currently existing on engines being shipped from the various overhaul test sites could be restored on a cost effective basis for the typical engine. Table 2 shows the results of the cost effectiveness feasibility study conducted by General Electric for the CF6-50 engine based upon typical overhaul test cell performance levels. It is noted that the greatest potential for cost effective refurbishment exists in restoring fan performance. This restoration includes surface finishes, leading edges, and maintainin g clearances . Cost effective performance restoration is also achiev- able on the HP compressor, and slight additional cost effective gains are achievable on the HPT . General Electric has concluded to date based upon the studies for both engine models that performance restoration resulting from tearing down the LPT module and restoring performance is not cost effective.

A word of caution, however. These studies are based upon a typical or average engine as it is shipped from the various overhaul facilities. Some of the restoration work used in these cost effective studies is currently being done by some airlines on a part - time basis. Not all engines that are shipped from the overhaul facilities are equivalent (low) in performance as the typical engine identified and used as part of this study. The cost effectiveness study for an average engine can be misleading on an individual engine basis. The key point to emphasize is that each airline should conduct its own cost effective- ness studies based on individual practices, labor rates and work scopes to define the actual fuel and dollars savings available. General Electric's con- clusions concerning the actual deterioration mechanisms within each module which contribute to the overall module deterioration are established and docu- mented in extreme detail within the referenced NASA reports. These deteriora- tion mechanisms can be used as a basis for each airline to conduct its own cost effectiveness refurbishment study. The implications of these studies are over - whelming. General Electric believes that potential savings of between 50 and 60 millions gallons of fuel could be realized in one-year's time period based upon the current CF6-6 and CF6 - 50 fleet of engines .

WHAT ELSE CAN BE DONE TO SAVE FUEL?

Discussions to this point have dealt with what is known about engine deterioration and refurbishment practices in todays operation and what can and s being done to further f u el conservation. There are other factors which must be considered in order not to use excess fuel . Careful attention to operation- al practices and use of derate power ratings are two such areas .

ENGINE ABUSE Any turbofan engine can be operated in a manner which could produce excessive deterioration. For example, an engine which has been stabilized at high power, then subjected to a reduction in power and subsequently subjected '\ to another accel is exposed to a condition where engine static cases have cool- ed faster than the rotor during the down time and could interfere with the hot rotor blades as they stretch during the accel thereby resulting in rubs and performance losses. This is known as "hot rotor reburst".

Every engine manufacturer publishes guidelines for engine operation which, if heeded, should result in avoiding the "hot rotor reburst" situation and any other similar situation. Proper discipline by all personnel responsible for any phase of engine operation from line maintenance personnel through flight crews is required in order not to abuse the engine.

USE DERATE POWER It is common knowledge throughout the industry that use of reduced power settings has a strong influence on parts life and maintenance cost.

CF6-S0 data analyzed as part of the Engine Diagnostic Program substanti- ated the fact that a larger amount of derate (reduced power) results in a lower deterioration rate. Figure 7 shows the average deterioration rates of the data from the 9 airlines studied. Shown are the average deterioration rates ex- pressed in teums of EGT (at fan speed) and a percent fuel flow increase (at fan speed) for 1000 hours of operation as a function of average flight cycle length (hours/cycle) for each airline studied. The average of the A300B data, the average of the DC-IO-30 data and the average of the 747 data are used to define a "composite characteristic". The numbers enclosed in parentheses indicate the average percentage thrust derate typically used by the indicated airline. While this summary is not sufficie~tly accurate to define an exact relationship between deterioration rate and average percentage derate, it does show a correlation between derate usage and reduced deterioration rates .

Although not implicitly suggested by these data, it is most probably a fact that continued usage of a given percentage of derate power will result in lower deterioration rates than alternately operating above and below that same percentage of derate. The same is true for maintenance cost. Maximum de- rate usage is encouraged.

SUMMARY To summarize, the portion of the NASA Engine Diagnostics Program aimed at defining CF6 deterioration characteristics was highly successful. Deter- ioration rates and modes were identified as were areas of design improvement which can and will result in improved performance retention characteristics.

Also defined were potential means of fuel conservation today with improved cost effective engine performance restoration practices during engine shop visits.

The potential for additional fuel conservation is there if we make maximum / use of this information. The engine manufacturer must desi gn more performance retention into his product; the airlines must analyze and modify engine (and aircraft) maintenance practices.

REFERENCES 1. NASA CR-1S9786, "CF6-6D Engine Performance Deterioration", RH Wulf, January 1980 .

2. NASA CR-1S9330, "C F6 -6D Engine Short Term Performance Deterioration", WH Kramer, JE Paas, JJ Smith, RH Wulf, April 1980.

3. NASA CR-1S9367, "CF6-S0 Engine Performance Deterioration", RH Wulf, November 1980.

4. "CF6 High Pressure Compressor and Turbine Clearance Evaluations", M. Radomski, and L. Cline, Paper Presented at NASA Aircraft Engine Diagnostics Conference, May 6, 1981 .

TABLE 1. CF6 · 50 DETERIORATION IN 1000 CYCLES AlC TYPE I NSTA L LATION DC 10 B747 AJOO I NITIAL INSTALLATION 1. 71 % 2. 07 % . 83 % 6 SFC M UL TlPLE BUILD INSTALLATION 1. 03% 1. 36 % . 51 % 6 SFC TABLE 2. COST EFFECTIVE PERFORMANCE REFURBISHMENT CF6-50 ENGINE % CRUISE SFC UNRESTORED COST EFFECTIVE PERFORMANCE REFURB I SHMENT FAN SECTION FAN BLADE TIP CLEARANCE .38 .38 FAN BLADE L.E . CONTOUR .12 . 12 F A N BLADE SURFACE FINISH .01 .01 SPLITTER LEADING EDGE .07 .07 BYPASS OGV - L.E . . 06 .06 BYPASS OGV - SURFACE FINISH . 18 .18 BOOSTER TIP CLEARANCE 0 . 03 BOOSTER AIRFOIL ROUGHNES S .01 0 HP COMPRESSOR BLADE & VANE TIP CLEARANCE .16 .16 AIRFOIL LEADING EDGE BLUNTNE SS . 05 0 AIRFOIL SURFACE FINISH .03 0 CASING/SPOOL SURFACE FINISH . 01 0 HP TURBINE STAGE 1 NOZZLE DISTORTION . 05 0 AIRFOIL SURFACE FINISH .10 .10 LP TURBINE BLADE TIP CLEARANCE . 30 0 INTERSTAGE SEAL CL E ARANCE . 22 AIRFOIL SURFACE FINISH .04 1.82 1.08 TOTAL 60 % OF UNRESTORED PERFORMANCE CAN BE RESTORED ON A COST EFFECTIVE BASIS

Fuel Cost History

u.s. Ai rline Jet Fuel Price

Monthly Averages CAB Data 100~------ -- ------------------------------~ 'C90 ...

o

-

Q)

= 80

ftS - :::; , (!) 70 , ...

...

Q) Q) I ~ , ~ 60 en , en - , c 'E 50 Q) , Q) ~ 10

~~.J--------~~~::-~-~-' ~ ---

Q)

-

Q) 30 _------ Trunks o .~

... -- __ , -International

~ 5 ~ 20 " ---Domestic Q)

..-

Q) 10 ::s ::s LL LL O~----~--~----~----~----~----~--~-----J 1973 1974 1975 1976 1977 1978 1979 1980 Year

FIGURE 1

CFS-SD

Performance Deterioration

Characteristics

+3.0 +2.0 6. Cruise SFC (%) +1.0 _ Short Term 0.9% t o Production New Baseline o 1000 2000 3000 4000 0 1000 2000 3000 Hours Since Installation

FIGURE 2

CF6-50 Deterioration Characteristics

DC - 10 A30e Inilial Multiple Inslallation Ins lallation Short- Term SFC lOll (%) 0.7 0.7 0.7 Typical Time to Removal Initial Installation (Hrs) 3000 4000 Initial Installation SFC 6Cruise loss ('10) 1.5 1.6 1.1 SFC Typical Time to Removal Deterioration > Unrestored ~ -:' Performance /,: Multiple Installations (HAS) 3000 3850 2000 Sh I T L or elm oss //// IIII Wtttt'jl Multiple Installation SFC 'Produclion New Baseline loss ('10) 0. 79 0.95 0. 79 Time Since Installation Unreslored SFC (%) 1.8 1.8 1.8

FIGURE 3

CF6-6D Initial Installation

Performance Deterioration

Performance Data Hardware Data Total Lo ..

al 4000 Hr. 2. 6% 6Cruise

T

Long SF C 2.0 1---t--t-t-.."..'--':7.::"F'-::;,oo-"'1--l Term (1 .7 '10 ) (%) 1.0 ...--+.,.....,,..s.""'-+---If---+-+--+--I o 2000 4000 6000 Time Since Installation, Hours 6 Cruise SFC at 4000 Hours Hardware 2.3% Performance 2.6% Good Agreement!

FIGURE 4

CF6-6D Multiple Build Engine

Performance Deterioration

4.0 Performance Data Hardware Data 3.0 On Wing O ••• rlor.llon (0 .9'!.1 6Crulse 2.0 SFC (%) Unr •• lored lOll Unr •• tored Lo ..

2. 1'" 1.0 2.01'1.

Time Since Installation, Hours Ll Cruise SFC at 3000 Hours Hardware 3.3% Performance 3.0% Good Agreement!

FIGURE 5

Unrestored Performance

Hardware Inspection Summary Compared to Performance Summary CF6-6D CF6-50 Hardware Performance Hardware Performance Data Data Data Data 2.08 2.1 HPT .06 1.82 1.8 LPT .36 HPT .15

6 Cruise

LPT .56 HPC .73 SFC (%) HPC .25 Fan .93 Fan . 86 o • Hardware Inspections and Performance Data Analysis Show Excellent Agreement

FIGURE 6

1 6

Effect of Flight Cycle Length

and Derate on CFS-50 Performance

Deterioration

, /2105"" Derale O(9.4).J" (5 . 6)

~ ----fJJd [J Compo.lle Characlerl.llc

LlEGT (OC) (1Y.2)~7.S) (7.5) , 1000 Hrs 4

11 10 1:.>----- -~;.6)

Derale (12.1) 0 A300-B U .. r.

o DC-10-30 U.er.

01--- - -- ---- ----- 6 B747 U.er • .6 /2 to 5% Derate (Typical Oera'8) -(94)-1 0(2.1) 4 . --- - --(7.5) Compo.lle Characlerl.Uc LlWF(%) . (S.6)Q] 0 --- __ __ _ 0 (7.5)- 1000 Hrs .2 (11.2)---___ (8.6)

11 10 12./!7-~~~1)Er(7.6L

Derale o~ ________________ __ o 2 3 4 5 6 7 Average Hours/Cycle FIGURE 7 I

__ J

CF6 HIGH PRESSURE COMPRESSOR AND TURBINE CLEARANCE EVALUATIONS M.A. Radomski and L.D. Cline General Electric Company SUMMARY In the CF6 Jet Engine Diagnostics Program the causes of performance degradation were determined for each component of revenue service engines.

It was found that a significant contribution to performance degradation was caused by increased airfoil tip radial clearances in the high pressure (HP) compressor and turbine areas.

Since the influence of these clearances on engine performance and fuel consumption is significant, it is important to accurately establish these relationships, especially now when fuel prices are rapidly escalating. It is equally important to understand the causes of clearance deterioration so that they can be reduced or eliminated.

This paper describes the results of factory engine t e sts run to enhance the understanding of the high pressure compressor and turbine clearance effects on performance. It also indicates the causes of clearance deteri- oration and discusses potential improvements in clearance control.

INTRODUCTION The CF6 Jet Engine Diagnostics Program showed measurable degradation of compressor and turbine airfoil tip clearances in revenue service engines.

The degradations of the compressor tip clearances were caused by spalling of abradable coatings from the stator casings and rotor . spools, by blade and vane tip rubs and by field assembly procedures involving out-of-round stator casings. Degradation of clearances in the CF6 compressor was estimated to produce, on the average, a 0.4 percent increase in the specific fuel con- sumption at cruise. This is estimated to amount to 15 million gallons per year for the 1981 CF6 engine fleet.

The effect of compressor clearances on compressor efficiency has been studied at General Electric for many years using a low speed research compres- sor. The data obtained from this research vehicle have been applied to the compressors of the CF6 engine family. Attempts have been made to verify these data by power calibrations of the CF6 production engines, but they have not been entirely successful because, in these tests, compressor efficiency was affected by several factors, and the effects of the airfoil tip clear- ances alone on efficiency could not be isolated with accuracy. There was, therefore, a need to conduct instrumented CF6 factory engine tests to accu- rately verify the influence of compressor clearances on engine performance.

The tests were conducted as part of the CF6 Jet Engine Diagnostics Program using an instrumented core engine.

The high pressure turbine tip clearance degradation was caused by blade tip rubs on the stator casing. The major cause of rubs was out-of-roundness.

The principal causes of HP turbine out-of-roundness are the thermal gradients and transient responses of adjacent structures, such as the compressor rear frame, turbine mid - frame , and the low pressure turbine case. The increase in HP turbine clearances was estimated to contribute, on the average, a 0.6 per- cent increase in cruise specific fuel consumption. This equates to approxi- mately 22 million gallons per year for the 1981 CF6 engine fleet.

Accurate measurements of HP turbine clearances and out-of-roundness have been difficult to achieve until now. Rub pins and High Energy X-Ray (HEX) tests have yielded reasonable approximations to date. The current performance sensitivity factors are based on turbine efficiency sensitivities established during air turbine testing and SFC effects determined from the engine cycle deck. Direct, on-engine analysis of the sensitivity factors related to turbine clearance were required to determine the turbine's contribution to overall engine performance deterioration. A test program was conducted as part of the CF6 Jet Engine Diagnostics Program to evaluate the effects of stator out-of-roundness and stage 1 blade tip clearance upon performance.

This paper outlines the scope of the liP compressor and HP turbine clear- ance evaluation programs, describes the unique instrumentation and the test procedures used, discusses the data reduction and presents some preliminary results.

HIGH PRESSURE COMPRESSOR CLEARANCE EVALUATION Program Scope The program was designed to determine the influence of compressor clear- ances on engine performance and also to evaluate potential improvements in clearance control. The approach used was to run instrumented core engine tests in which airfoil tip clearances were varied by varying quantities of rotor bore cooling air . The greater the cooling air flow, at any power set- ting, the lower was the bulk temperature of the rotor structure and, hence, the greater were the tip clearances. The running clearances were calculated from measured temperatures of the stator and rotor structures. The calcu- lations were verified using measured blade tip clearances in stages 10, 12 and 13. The clearance changes and the corresponding measured performance changes were then correlated.

Rotor Bore Cooling - . Externally supplied shop air was used to cool the rotor and thus vary the clearances. The total cooling air flow was measured by an instrumented orifice and remotely controlled by a valve upstream of the orifice shown in Figure 1. From the orifice, the air flowed to the manifold around the slave front frame, from there through flexible hoses into the frame struts and then into the HP compressor inlet inner cavity shown in Figure 2. From this cavity, some of the air leaked out through the air/oil and air seals bounding the cavity (only the rotating seals are shown in the figure) and the remainder, the net cooling air flow, entered the rotor main cavity through the holes in the forward shaft. The cooling air exited the rotor cavity through holes in the rear shaft and from there was discharged through the compressor rear frame struts into the test facility exhaust. The net cooling air flow was calculated for each test point by subtracting the seal leakages from the total measured flow.

Instrumentation - In addition to t he standard factory test engine in- strumentation, there were a i r temperature and pressure rakes at the com- pressor inlet and discharge to measure compressor efficiency, and also temper- ature and pressure probes at the rotor nain cavity inlet and discharge to monitor the cooling air f low.

The compressor mechanical instrumentation shown in Figure 2 included the stator casing and rotor structure thermocouples, the clearanceometers and the touch probes. The touch probes were used to doublecheck the clearanceo- meters at the steady-state engine running conditions.

The clearanceometers used were electrical capacitance probes whose output voltage varies with the distance between the clearanceometer and the passing blade tips. A touch probe is a traverse probe with an open electrical cir- cuit which closes when the probe touches the passing blade tip. When contact is made, the probe automatically backs off. The distance traversed by the probe to touch the blade tip is measured by a linear potentiometer.

Engine Tests The tests were run in the General Electric Altitude Test Facility. They included steady-state power calibration tests, rapid accels and decels and a simulated typical flight cycle. The steady-state power calibration tests were made with three different engine inlet conditions, which were the core engine ambient, the simulated fan engine sea level static and cruise inlet conditions, and at a number of different engine speeds. At each speed point, at least three different sets of clearances were produced and their effects on engine performance were measured.

The test procedure was as follows. After the engine inlet condition and speed were stabilized, the rotor bore cooling air flow was set at the desired level, and three minutes later all instrumentation sensors were scanned and their outputs recorded. Instrumentation scanning and recording of data was repeated approximately every three minutes until the rotor disk temperatures became stable, which took fifteen to twenty minutes. At this point, the cooling air flow was changed and scanning of instrumentation sensors and data recording was repeated.

The simulated flight cycle and the rapid accels and decels, which included hot rotor rebursts, were made to evaluate potential improvements in clearance control. The tests were run with ambient engine inlet conditions and two different sets of clearances. The rapid accels were made from ground idle to take-off power setting which was then held constant until rotor disk temperatures became stable after which the engine was rapidly deceled to ground idle.

A hot rotor reburst is said to occur when an engine is rapidly deceled from a high to a low power setting and after a short time at the low power is reburst back to the high power setting. This type of power throttle maneu- ver may result in the most adverse tip clearances in the aft end of the com- pressor, particularly if the engine dwells at the low power setting for such a time period so as to produce the maximum temperature difference between the rotor and stator structures. In the hot rotor reburst tests, the metal temperatures were stabilized at take-off, then the engine was rapidly deceled to ground idle and a short time later, it was reburst back to take-off where the metal temperatures were again permitted to stabilize. Time at ground idle varied from approximately one minute to thirty minutes. In these tests, both the transient and the steady-state data were recorded.

Results and Discussion Results discussed here are the steady-state differences in rotor temper- ature, compressor clearances, compressor efficiency and engine fuel flow produced by different quantities of the rotor bore cooling air flow. Typical results are shown for one particular speed point, the simulated sea level static take-off point. The compressor efficiency changes and the engine fuel flow changes are then shown as functions of normalized average compressor airfoil tip clearance changes. Analysis of the transient test data has not yet been completed, and, therefore, these data could not be included. Finally, the most significant cause of compressor clearance degradation in the CF6-50 field engines is briefly discussed and some data are presented to underscore the relevance of this compressor clearance evaluation program.

Rotor Temperatures - The axial temperature profiles in the rotor main cavity are shown in Figure 3 for different rates of cooling air flow. The data indicate a greater rate of change in the aft end than in the forward end of the rotor, which is primarily due to the geometrical differences of these two areas of the rotor structure and, to some small extent, due to the higher conductivity of Inca 718 as compared to titanium. The data also show, as would be expected, that the higher the cooling air flow the lower the air temperature of the rotor cavity and of the disk bores as shown in Figure 4.

At stage 14, the maximum cooling air flow reduced the air and metal temper- atures by at least 300 F, which is a very significant reduction and some- what greater than the pre-test predictions indicated.

Radial temperature profiles in the stage 14 disk are shown in Figure 5.

The cooling air was most effective at the disk bore. The rim of the disk was much less affected where the maximum temperature reduction was about on e quarter of that at the bore. In other stages, this was even a smaller fraction. Similar temperature profiles were generated for all other power calibration points, and they were used to calculate the airfoil tip clear- ance changes which are discussed in the next paragraph. As will be noted, the pre-test predictions Significantly underestimated the effectiveness of rotor bore cooling.

Airfoil Tip Clearance Changes - At the power calibration points, the stator casing temperatures remained constant, and only the rotor temperatures were affected by the cooling air flow. Therefore, to calculate the clear- ance changes, only the rotor temperature changes needed to be considered.

Clearance changes calculated in this manner, for the sjmulated sea level static take-off power calibration point, are shown in Figure 6 for different cooling air flow rates. Similar calculations were made for all other power calibration points. These data were then used to calculate the normalized average clearance changes which were later correlated with the corresponding measured compressor efficiency and engine fuel flow changes. The correla- tions are discussed in a later paragraph where the normalized average clear- ance is also defined.

The measured and calculated clearances for stages la, 12 and 13 are shown in Figures 7-9. There is an excellent agreement for stage 10. For the other two stages, there are small discrepancies between measured and calculated data. The causes of these discrepancies have not yet been determined.

Efficiency and Fuel Flow Changes ~ Compr ps sor efficiency changes as a function of the cooling air flow are shown in Figure 10. The data were measured at the simtllated sea level static take-off conditions. Indicated and corrected values are shown in the figure. The corrections were made t o obtain the net effect of clearance changes by allowing for 1) leakage of cooling air into the compressor inlet and 2) heat removed from the gas path by the cooling air.

The magnitude of the latter correction was about ten times as large as t hat of the former. Colder cooling air, leaking into the compressor inlet, slightly reduced the actual air temperature downstream of the inlet temper- ature rakes and, therefore, made the actual efficiency reduction, caused by increased clearances, slightly larger than the indicated reduction calcu- lated from the measured temperatures and pressures at compressor inlet and discharge. Heat removed from the gas path by the cooling air reduced the indicated compressor discharge temperature and, hence, made the indicated efficiency reduction, due to the increased clearances, somewhat smaller than the actual reduction.

Engine fuel flow changes as a function of the cooling air flow are shown in Figure 11. The corrections made to the fuel flow were in the oppo- site direction to that of the efficiency corrections, i.e., the fuel flow corrections resulted in a smaller actual fuel flow change than that indicated , because the measured fuel flow included the energy removed from the cycle by the cooling air which was vented overboard. If the clearance changes were produced mechanically, there would be no heat loss from the cycle, and there - fore, the total fuel required would be less.

The compressor efficiency and fuel flow changes were corrected in this manner for all power calibration speed points and then correlated with the average normalized clearance changes which are discussed in the next paragraph .

Correlation of Efficiency and Fuel Flow Changes with Clearances - Com- pressor efficiency changes are shown as a function of the normalized average clearance changes in Figure 12 for four different speed points with three different engine inlet conditions. There is a good correlation of measured efficiency changes with the calculated normalized average clearance changes .

The normalized average clearance change is defined as follows: 66CL/L, where: 6CL clearance change in a given stage and L airfoil length in the same stage.

The line shown in the figure is the best line drawn through the data points .

The engine fuel flow changes versus the normalized average clearance changes are shown in Figure 13. Only the data obtained at the simulated sea level static take-off and cruise conditions are presented. The data for the other two speed points were inaccurate because of a fuel flow meter malfunction.

The line shown in the figure is derived from the efficiency line in Figure 12 through the derivative of fuel flow as a function of efficiency for the test engine. Because the line correlates well with the fuel flow data, it , there - fore, indicates consistency of the efficiency and fuel flow changes.

Compressor Clearance Degradation Compressor stator casing out-of-roundness has been found to be the most significant cause of airfoil tip clearance degradation in revenue service engines. The effect of casing out-of-roundness on compressor blad~ and vane clearances is shown in Figures 14 and 15, respectively. To meet the mlnlmum clearance at build-up, all blade tips have to be machined shorter by the amount equal to at least the inward distortion. Vanes, on the other hand, only need to be machined shorter in the area of the inward casing distortion.

The effect of out-of-roundness on clearances is magnified for two reasons.

First, the permitted interchangeability of modules and, second, because of the field shop practices at engine build-up. At every shop visit, the dis- torted casing may be installed in a different compressor module, thus causing short blades and hence increasing the clearances in all of them.

The field shop practices involving out-of-round casings at engine build - up require that the individual compressor rotor and stator casi n g modules are machined to meet a target minimum clearance. However, verifica t ion of the actual minimum clearance is required,and this is accomplished by applyi n g wax strips of known thicknesses to the stator casing and rotor spool lands and then installing the casing halves around the rotor. The rotor is then rotated through 360 after which the stator casing halves, upper and lower , are removed and the wax strip thicknesses are measured. If the wax strips were rubbed, indicating below minimum clearance, then the airfoil tips are hand grou nd to correct this condition. How the magnifying effect on clearances is produced will be illustrated by a specific field engine incident. An engine that failed to meet the minimum performance standards was disassembled and inspected .

Inspection of the compressor, which had been refurbished prior to the test cell run, indicated out-of-roundness in the stator casing which is shown in Figure 16. The rotor blade tips in the aft end of this compressor were rounded off by hand grinding at assembly, because of below minimum clearance due to the out-of-round stator casing. A rounded-off blade tip from this compressor is compared to a machine ground blade tip from another compressor in Figure 17. The casing out-of-roundness in the aft end was a maximum of 20 mils, but the blade tips were up to 40 mils shorter at the leading and trailing edges. This is a good example of how the effects of an out - of-round casing on the average clearances are magnified. Hand grinding is not accept - able for this purpose. A procedural change has been specified to require remachining rather than hand grinding in similar cases.

Out-of-roundness data from a survey of twenty-six stator casings are summarized in Figure 18, where three sets of values are shown, i.e., the average of all measured and the largest measured in modules Band C. Although the average values were only about 10 mils, the maximum values were as much as 30 mils. Out-of-roundness of the module B stator casing, in particular , had a very significant impact on performance, since the clearances in the aft end of this module had to be increased by at least 20 to 30 mils. To avoid this large performance penalty, field procedures have been specified for the repair of casing out-of-roundness.

Concluding Remarks The data presented are preliminary and are still being analyzed. b ut they indicate that the test technique was effective. By means of rotor bore cooling, appreciable changes were produced in rotor temperatures, compressor airfoil tip clearances, compressor efficiency and core engine fuel flow. A good correlation was obtained of measured and calculated clearances . Compres- sor efficiency changes correlated well with the normalized average clearance changes. and. furthermore, they were consistent with the corresponding fuel flow changes. The clearance changes produced in this engine test were comparable to the maximum clearance degradation observed in the revenue service engines. Significant fuel savings can be achieved if clearance degradations in the revenue service engines are reduced or eliminated. The results of this test are being applied to the development of new General Electric commercial engines.

HIGH PRESSURE TURBINE ROUNDNESS/CLEAruu~CE EVALUATION Program Scope As an engine accumulates operating time in revenue service, its per- formance deteriorates as a function of time and operating cycles. A signifi- cant part of the CF6 engine performance deterioration is chargeable to the high pressure turbine. This deterioration is primarily due to increased blade tip-to - shroud clearances caused by rubbing of the blade tips on the shrouds. The objective of the HP turbine clearance and roundness diagnostics program was to provide test data to improve the understanding of the effect of blade tip clearance on performance as affected by transient engine oper- ating conditions and of stator out-of - roundness.

Engine Tests The tests were conducted in Test Cell 2 at the General Electric Company Plant in Evendale, Ohio. The test vehicle was a CF6-50C engine. The engine was mounted in an overhead frame as shown in Figure 19.

Blade tip clearance data were obtained from clearanceometer probes especially designed for this purpose . Eight of these clearanceometer probes were installed in the engine which had been modified to accept the probes as shown in Figures 20 and 21. The probes were located circumferentially around the engine over the stage 1 blade tips as identified in Figure 22. Clearance data was then recorded for the type of engine operations shown typically in Figures 23 and 24. These operations included numerous steady-state and transient operating conditions.

Test Results Measured data from each of the eight clearanceometer probes were aver- aged to obtain the "round engine" clearance. This clearance, plotted against time. defines the round engine clearance response. The "round engine" data were then used for tuning axisymmetric analytical models of the modified test engine configuration. This knowledge was then utilized to upgrade the pro- duction configuration engine analytical model.

Representative plots of a throttle burst (steady-state idle to steady- state take-off) and throttle chop (steady-state take - off to steady-state idle) are presented in Figures 25 and 26.

Evaluating the results of each of the individual clearanceometer probes relative to the averaged data for any given time yields a measured "out-of- roundness." These data were then used in conjunction with calculated mechanical loads and calculated thermal distortions to identify deficiencies, and they formed the basis for correction of the calculated results. Out-of- roundness results are shown in Fi.gures 27 and 28 for throttle bursts and chops respectively.

Discussion of Results The "round engine" clearance transient response results matched pre-test calculated predictions closely as did clearance at steady-state operating points. This clearance match was also confirmed by thermocouple data from the shroud support, and it verified that the anlytical modeling previously used was representative of the engine structure. Some relatively minor model parameter adjustments were necessary to force calculated results to more perfectly match measured clearance responses. The test verified the calcu- lated importance of hot or warm rotor throttle rebursts on minimum blade tip clearance experienced in engine operation.

Calculations predicted that steady-state engine operating clearances are set at engine transient operating conditions. The worst case or minimum clearance condition, occurs during a hot rotor reburst which has idle dwell times less than 3 to 4 minutes.

The turbine shroud support member is considerably less massive than the turbine disk and, consequently, cools more quickly than the disk. The rapid reacceleration of the engine adds rotational stress growth and blade thermal growth to the still existent disk thermal growth. The net result is a hot blade tip radius greater than that of the supporting structure, which results in rubs.

A significant part of engine deterioration may be caused by "warm rotor rebursts" (idle times over 4 minutes) for which little data was available.

Testing included several runs to provide data relative to this type engine operation. Rebursts with idle dwell times of 8, 6, 4, 2, 1 and 0.5 minutes were completed to quantitatively evaluate these significant round engine and out-of-roundness responses for which good analytical predictions were not available.

Out-of-roundness was evaluated for all the steady-state and transient operations tested. Pre-test predictions of out-of-roundness at steady-state conditions were compared to measured out-of-roundness. While the results compared well in magnitude, the shapes differed sufficiently to indicate that out-of-roundness driving phenomena existed which were not accounted for analytically.

However, test results did clearly indicate that the phenomena not accounted for in the analysis were thermal and not mechanical load induced.

Testing and analysis accounted for mechanical distortions caused by engine thrust, redundant torque, and mount horizontal and vertical loads as well as manufacturing tolerances. These distortions agreed very well with analytical predictions. Test data also included temperature measurements for portions of the engine structure which were analytically determined to affect turbine stator roundness. They included combustor exit temperature profile, turbine mid-frame temperature distributions, shroud support temperature distributions and low pressure turbine case temperature distributions. These temperature measurements lead to improvements in out-of-roundness data correlation .

. 27 Compariso n of t est results, field - observed turbine deterioration rates, and rub l o ca t i ons ag r ee well and verify that most turbine blade tip clearance deteriora t ion is t he result of combined stator out - of - roundness and warm rotor reb u rsts .

Concluding Remarks As a r esul t of t his program accurate measurements of HP turbine clear- ances and o ut- of -roun dness were obtained for steady - state operations as well as thro t tle burs t s , ch o ps and rebursts from various dwell times at ground idle.

The measured aver a ge clearances agree well with the anal y tical predictions.

The lear n i n g ga ined from this testing and from subsequent anal y tical refinemen ts is b ei ng a pplied to identify potential improvements in turbine efficiency for t he CF6- 50 and several other engine models . This is especially true in t he deve lo pme n t of enhanced ability to optimize HP turbine rotor-to - stator tra nsi e nt re spon se rate and stator out - of - roundness calculations.

Instrumented Core Engine In Altitude Test Facility FIGURE 1

Compressor I nstrumentation/Flow

Schematic

f Clearanceometer, 3 Per Stage

HPC Inlet Legend Inner Cavity o Static Pressure Probe ... Thermocouple

o Touch Probe

- Cooling Flow FIGURE 2

Rotor Cavity Air Temperatures at Various

Cooling Air Flow Rates /.,-

".

".

800 ./ 0 Gas Path "."

Temp . ~ . / ".

".

".

" .

".

". 100

".

".

".

400 ,.

t

C ooling Air Flow [% of Max.]

o

4 6 8 10 12 14 Rotor Stage FIGU RE 3

Disk Bore Temperatu re s

at Va rious Cooling Air Flow Rates

_ ..

Gas Path Temp ...... _ . -· ._,- 0

.-

.-

.-

.," 30 Bore Temp. [OF] 4 00 Co o ling Air Flow [% of Max.]

10 11 12 13 14 Rotor Stage FIGU RE 4 Radial Temperature Profiles in Stage 14 Di sk at Various Cooling Flow Rates Disk Radius (Inches) FIGURE 5

Airfoil Tip Clearance Changes

at Various Cooling Air Flow Rates

Clearance Change [Mils ] Cooling A ir Flow Incr ease [% of Ma x.]

Rotor Stage FIGURE 6

Airfoil Tip Clearances Measured vs

Calculated

Stage 10 Measured [Mils] Legend: • Ambient Inlet X Simulated Cruise • ~ Simulated S.L.S.

Calculated [Mils] FIGU RE 7

Airfoil Tip Clearances Measured vs

Calculated

30 Stage 12 , Measured [Mils] Legend: • Ambient Inlet X Si mulated Cruise • ~ Simulated S.L.S.

o------~----~------~~

o 10 20 30

Calculated [Mils] F IG UR E 8 -- -_._--

Airfoil Tip Clearances Measured vs

Calculated

30 Stage 13 A.

x

Measured [Mils] Legend : • Ambient Inlet X Simulated Cruise • .:I Simulated S. l.S.

10 20 Calculated [Mils] FIGURE 9

HPC Efficiency vs Cooling Air Flow

HPC Efficiency Change (%)

o 30 60 100

Cooling Air Flow - % of Max.

FIGURE 10

Fuel vs Cooling Air Flow

2.0 6. Fuel Flow Air Flow [%] 1.0

o

o 30 60 100

Cooling Air Flow - % of Max .

FIG URE 11

Compressor Efficiency Change

vs

Airfoil Tip Clearance Change

• Ambient Inlet HPC Efficiency X Simulated Cruise Change[%] ~ •• Simulated S.L.S .

. 2%

~

FIG U RE 12

Fuel Flow Change

vs

Airfoil Tip Clearance Change

Based on HPC X Simulated Cruise Efficiency Derivative • Simulated S.L.S.

1.5 A Fuel Flow 1 0 L.l Air Flow .

[%] .5 Clearance Change % FI GU RE 13

Effect of Casing Distortion

Blade Effect Stator Casing Distortion Rotor Blade All Shor t R otor Structure Large Clearance FIGURE 14

Effect of Casing Distortion

Vane Effec t Stator Cas in g Di stortion \--t---- ~ _Sho rt Vanes Rotor Structure Good Clearanc e FI GURE 15

Casing Out of Roundness vs Stage

Mod ul e A ~R Mils ~ ~ ~

-

~ Stage FIGURE 16

Instrumented CFS-50 Engine In Test Cell

FIGURE 19

Current CFS-50 HPT Cross Section

FIGURE ~

Hand Grinding of Blade Tips at Assy

Hand Ground Machine Ground Blade Tip From Module A FI GURE 17

Casing Out of Roundness vs Stage

Module B llR

~

Mils 30 ~f26c","g.

O~~--~--~~~--~- Stage FIG URE 18

Turbine Modification to Receive

Clearanceometer Probes

I/ Clearanceometer Probe. I r at (8) Location.

FIGURE 21

Probe Angular Position (ALF)

104'19 ' C #3 197'51 ' FIGURE 22

Test Sequence

Th rottle Bursts and Chops 8 Min 8Mln 8Mln IMln T / O 20e.c Decol 1080c Accol Gi l 30Mln I Min Cold Motor ~---------------- ------------------ ----------------~- FIGURE 23

Test Sequence

Power Calibration T / O Gi l FIGURE 24 Blade Tip Clearance Transient Response - Throttle Burst From Idle To Take-off Blade Tip Clearance

EMIIS

10 1000 Time (Sec.)

FIGURE 2S Blade Tip Clearance Transient Response - Throttle Chop From Take-off To Idle Measured Blade Tip Clearance 100 1000 Time (Sec.)

FIGURE 26

Stator Out-Ot-Roundness - Throttle Burst

Blade Tip Clearance ~M I1' C;; Cal culated 10 100 1000 T ime (Sec.)

FIGURE 11

Stator Out-Ot-Roundness - Throttle Chop

Blade Tip Clearance A Cal cuated 10 100 10 00 Time (Sec .)

FI GURE 28 JT9D JET ENGINE DIAGNOSTICS PROGRAM* W. J. Olsson and W. J. Stromberg Pratt & Whitney Aircraft Group SUMMARY The NASA JT9D Engine Diagnos ti cs Program has been a four-year effort to identify and quantify the various engine deterioration phenomena that affect JT9D performance retention and identify approaches to improve performance retention of current and future engines. The program has included surveys of historical data, monitoring of in-service engines, testing of instrumented eng1nes, analysis, and analytical modeling. The Boeing Commercial Airplane Company, Douglas Aircraft Company, Trans World Airlines, Pan American World Airways, and Northwest Airlines participated as subcontractors in various phases of the program. Historical data was provided also by American Airlines.

The initial studies established that performance deterioration is made up of short- and long-term modes, both of which are flight cycle related phenomena. The later efforts provided additional data and refined and expanded on the initial conclusions.

The short-term deterioration occurs primarily during airplane acceptance testing prior to delivery to the airline. Therefore, it has small effect on revenue service performance retention. The long-term deterioration continues throughout engine life with a negative effect on performance retention.

The short- and long-term deterioration modes for combined effect of the 1. An increase of 2 percent in cruise thrust the JT9D-7 1S shown on figure typical after 2000 flight cycles of revenue specific fuel consumption 1S in unrepaired engines. serV1ce due to performance loss Short-term deterioration results from an increase in gas-path running clearances with resultant decreases in engine module efficiencies. This short-term effect is caused by flight-load induced engine deflections with resul ting rubbing of airfoils and seals. Wearing of blades and seals occurs for the most part pr10r to revenue service during the var10US airplane maneuvers associated with the production acceptance testing of the airplane.

This flight-load induced wear occurs in all modules. The results show a 0.8 percent increase in flight thrust specific fuel consumption during the predelivery airplane acceptance testing and an additional 0.3 percent increase during early revenue service.

Long-term performance deterioration is also a flight cycle related phenomenon. It 1S caused by erosion of airfoils and gas-path seals during

* This work was conducted by Pratt & Whitney Aircraft for the National

Aeronautics and Space Administration under Contract NAS3-20632.

ground operation and take-off and by cyclic induced thermal distortion of the high-pressure turbine airfoils. Erosion primarily affects cold section efficiencies by blunting the blade leading edges, reducing airfoil chord, and further opening running clearances. Thermal distortion of airfoils results from high-temperature cycling of the airfoils with resultant gas-path leakage and loss of optimum airfoil shape.

The diagnostics program has shown that performance retention within 1 to 2 percent of initial revenue performance can be maintained with a proper program of hot section and cold section maintenance as shown on figure 2.

INTRODUCTION The NASA JT9D Engine Diagnostics Program is a part of the NASA sponsored Engine Component Improvement (ECI) Project which is directed toward improving the fuel consumption of selected current high bypass ratio turbofan engines and their derivatives by 5 percent over the life of these engines. The ECI project is divided into two subprojects: Performance Improvement and Engine Diagnostics . Performance Improvement is directed toward developing fuel saving component technology which may be applied to current engines and their derivatives. Engine Diagnostics is directed toward identifying and quant i fying engine performance losses that occur during the engine's serVice life and developing criteria for minimizing these losses, as shown on figure 3. The JT9D Jet Engine Diagnostics Program, which is now nearing completion, has successfully identi fied and quanti fied the various causes of JT9D performance deterioration and the pos sible approaches toward improved performance retention .

This paper will briefly describe the various approaches used in this project and the results and conclusions that have been reported to date.

APPROACH The ideal approach for determining the cause and extent of engine deterioration would consist of tracking a large number of individual engines from production testing through extended revenue service, monitoring their flight performance with expanded instrumentation, closely tracking their maintenance histories, then correlating specific maintenance events, performance shi fts, and operational his tory. For numerous reasons this procedure was not feasible . Therefore, we took the following approach.

The JT9D-7A engine was selected for the study since various models had been operating for a long time and some of these models ' were still 1n production; thus, both ample high-time and new engine data were available.

The first task was the collection of available historical data. These data inc luded: o Pratt & Whitney Aircraft production performance records to establish a base level .

o Airframe manufacturers certi fication records to show early changes 1n performance.

o Airline and Pratt & Whitney Aircraft prerepair and postrepair calibration test results and hardware inspection results to explain long-term changes.

o In-flight engine monitoring data to establish the relation between ground performance and cruise performance changes.

Based on the analysis of these data, some preliminary conclusions were drawn: o There are four generic causes of engine performance deterioration, namely: 1) flight-load induced clearance changes; 2) erosion of fan and compressor airfoils and seals; 3) thermal distortion of hot section parts; and 4) variations in airline repair standards.

o Performance deterioration trends may be divided into two distinct time periods: short-term and long-term deterioration. The pr1me cause of short-term deterioration is flight-load induced rubs which open gas-path clearances, thus reducing module efficiencies and influencing airflow. The analysis of the historical data as seen on figure 4 showed a 1 percent increase in thrust specific fuel consumption at sea level in the first few flights conducted by the airframe manufacturer prior to delivery of the airplane to the airlines.

o Performance deterioration then occurs at a slower rate dominated by erosion of cold section airfoils and seals, with resulting blunting of airfoils and further opening of running clearances. This erosion and thermal cycle induced distortion of hot section airfoils results in loss of airfoil efficiency and increased secondary flow leakage. The historical da'ta showed the sea level thrust speci fic fuel consumption to have increased to about 4 percent above production levels after acceptance testing and 3000 revenue flight cycles of an unrepaired engine operation as seen on figure 5.

o The deterioration of the turbine airfoils and seals results from changes in their environment. These changing temperature and flow patterns are caused by deterioration in the compressor and combustor modules. Thus, more frequent cold section maintenance pays off 1n reduced deterioration in the higher priced hot sections.

0 A compar1son of the fleet historical prerepair and postrepair of percent 1n calibration data showed an average performance recovery 1 with potential sea level take-off thrust specific fuel consumption a for percent recovery with increased cold section and hot section refurbishment.

The first phase of the program provided an abundance of information but left numerous gaps in the data. The second phase, or in-service engine performance study, conducted jointly with Pan American World Airways, expanded the data base significantly by allowing us to monitor a controlled sample of 28 JT9D-7A engines in the Pan Am 747SP airplane fleet from preflight testing of the engines at Boeing through 2100 flight cycles of operation. The data collection included: installed engine ground calibrations before the first airplane flight and periodically during subsequent revenue service; in - flight eng~ne calibrations during the flights immediately following the ground calibrations; a complete set of crew-collected eng~ne flight condition monitoring data from the fleet; prerepair and postrepair calibrations and repair histories on each of these eng~nes that came into the shop; and an expanded instrumentation calibration and complete analytical teardown of one of the engines after 141 flight cycles (see figure 6).

The results of this effort firmly established that the flight-loads induced short-term deterioration . occurs ~n the first few flights prior to revenue service. It provided ample data for the refinement of the var~ous engine module deterioration prediction models which were first developed on the basis of the historical data. Finally, it provided a correlation between performance retention at flight cruise conditions and performance change as measured by ground calibrations. The quality of the flight performance data was less than that of the ground tests due to the available instrumentation systems. However, the data sample was large enough that statistical trends could be drawn. One such set of data is 747SP engine condi tioning monitoring (ECM) fuel flow data shown on figure 7. The data were recorded at cruise altitudes between 32,000 and 40,000 feet and corrected to 35,000 feet and constant eng~ne pressure ratio (EPR). A trend line through the 1398 data points shows a 1.7 percent increase ~n fuel flow rate after 1500 revenue flight cycles from the start of airline service on engines with no repa~rs.

The short-term flight-load induced performance loss, though not significantly contributing to revenue service wear, does present a challenge.

If it can be eliminated or significantly reduced, the new airplane could be delivered to the airline with up to 1 percent improved sea level thrust specific fuel consumption which is equivalent to 0.8 percent improved cruise thrust specific fuel consumption. Previous studies have estimated that more than 80 percent of the flight-load induced damage is caused by aerodynamic (pressure) loads applied to the fan cowl, and the remaining damage is caused by inertia loads from gusts and hard landings plus maneuver-induced gyroscopic loads.

The final two data gathering tasks of the JT9D diagnostics program were test programs directed toward a better understanding of this flight-load induced wear. The first of these was the Simulated Aerodynamic Loads Test conducted in a Pratt & Whitney Aircraft test stand. The objectj.ves of this test program were to determine the changes in engine operating clearances and performance under (1) thrust and thermal loads; (2) static simulated aerodynamic flight loads, figure 8; and (3) the combination of thrust, thermal, and static aerodynamic loads during eng~ne operation to permit validation of the levels, module distribution, and causes for short-term performance losses. In addi tion, the test program would validate or permi t refinement of previous analytical study results on the impact of aerodynamic flight loads on performance losses. To accomplish these objectives, an engine was analytically built with average production clearances and new seals as well as extensive instrumentation to monitor performance, case temperatures, and clearance changes. A special loading device was designed and constructed to permit application of known moments and shear forces to the engine by the use of cables placed around the flight inlet. These loads simulated the estimated aerodynamic pressure distributions that occur on the inlet 1n var10US important segments of a typical airplane flight.

The test engine and loading device were installed in the Pratt & Whitney Aircraft X-Ray Test Facility, shown on figure 9, to permit the use of X-ray techniques in conjunction with laser probe clearance measuring instrumentation to monitor important engine clearance changes under both steady state and transient engine operating conditions. Upon completion of the simulated flight-load test program, the test engine was analytically disassembled and the condition of gas-path parts and final clearances was extensively documented.

The performance monitoring calibrations between tests indicated that the engine lost 1.1 percent in sea level take-off thrust specific fuel consumption due to permanent clearance changes caused by the application of these inlet loads. Another 0.2 percent change in thrust specific fue l consumption was produced by an 1ncrease 1n airfoil surface roughness in the low-pressure compressor and thermal distortion in the high-pressure turbine. This additional 0.2 percent was a result of the experimental nature of this test program and does not occur in early revenue serV1ce. Prior to the test program, the change in sea level performance due to clearance changes was predicted to be 0.9 percent. Therefore, the agreement between measured and predicted performance is considered to be satisfactory.

The overall eng1ne performance loss was distributed among all modules; however, the low-pressure compressor and high-pressure turbine contributed the major portion of the loss. The major permanent clearance changes (seal rubs) occurred in the f an, high-pressure compres sor, and high-pressure turbine and were found to be the direct result of the loads imposed. Table I compares these results with previous comparisons of module contribut i on to sea level performance changes with early usage.

Transient testing, conducted after completion of the simulated aerodynam i c loading, indicated no additional performance losses associated with transient engine operation.

The flight loads test was the final phase of the JT9D Diagnostics Program.

It was conducted as a J01nt effort with the Boeing Commercial Airplane Company. Boeing, under contract with NASA Langley, provided the test airplane and measured the flight loads on the instrumented engines. Pratt & Whitney Aircraft, under contract with NASA Lewis, provided the instrumented engines and measured the effects of the flight loads on the engines. The flight loads test was conducted to verify the simulated aerodynamic loads used in the X-ray load test program and to further expand on the flight conditions and flight load effects measured in that program. Specifically, the flight loads test objectives were as listed on figure 10.

The test approach was to install an analytically built and instrumented engine in position No. 3 on the Boeing test 747 (RAOOl) airplane and an analytically built and instrumented fan case on the position No. 4 engine, figure 11. The analytically built engine was calibrated at Pratt & Whitney Aircraft before delivery and then again after installation in the aircraft, pr10r to flight testing.

A ser1es of flight tests was conducted with progressively increasing flight loads. Continuous, simultaneous measurements with all data systems were recorded to accurately document the cause and effect relationship of flight loads to engine deterioriation. Performance and fan clearances were documented after each flight by calibrations and rub measurements to determine the effect of increasing loads.

A final postflight calibration at Pratt & Whitney Aircraft and an analytical teardown of the analytically built engine was conducted following the flight program to quantify the effects of the flight loads, figure 12.

Instrumentation included pressure taps on the positions No. 3 and 4 nacelles to measure the aerodynamic loads, accelerometers on both engines to measure inertia loads, and rate gyros on both engines to measure gyroscopic loads. Clearance closures were monitored by laser probes on the high-pressure turbine of the position No. 3 engine and on both fans. Thermocouples on the high-pressure turbine of the position No. 3 engine measured the transient and steady state thermal effects on the running clearances. Finally, expanded performance instrumentation on position No. 3 engine permitted closer performance monitoring since it was the prime data source. The position No. 4 engine was instrumented sufficiently to identify clearance and load differences due to its position on the wing.

The flight loads testing was successfully completed, and the test data analysis is now in process. The clearance closures at actual flight conditions generally repeated the measured closures in the X-ray load test program under simulated flight loads, see figure 13. This program also confirmed that aerodynamic loads occurring during high power operation, that is, take-off rotation, airplane power-on stalls, and high G maneuvers, are the prime cause of fan performance deterioration. In addition, the preliminary analysis indicates that the combination of mechanical loads and transient thermal expansion during the extended high power climb 1S the pr1me cause of short-term deterioration in the high-pressure turbine.

The final report of the flight test program will be issued this fall and will include a final refinement of the performance deterioration models, including analytical results of the X-ray load test and the flight load test.

ENGINE PERFORMANCE RETENTION PREDICTION MODELS One of the major objectives of this program has been the development and refinement of analytic models for predicting the deterioration with eng1ne usage of both the complete JT9D engine and the individual modules. These models consist of families of curves which define the changes 1n the performance parameters (efficiency, flow capaci ty) wi th usage for each of the engine modules. These var10US parameter changes are applied to the JT9D performance analysis program to determine the predicted performance change with usage of an average engine. The preliminary models were prepared based on analysis of the performance, engine usage, and replaced parts condi tion data collected in the first phase of the program. All the in-service data collected on the Pan American 747SP fleet was used for the first refinement of the models. This effort was followed by a just-completed second refinement of the short-term deterioration predictions based on the X-ray load tes t results.

Table I compares representative results from the different data sources.

Figures 14 and 15 show the thrust specific fuel consumption changes at sea level for an average engine versus usage as predicted by the latest model. For that model, it was assumed that high-pressure turbine performance had been stabilized at a constant level after 1000 flight cycles and the low-pressure turbine after 2000 cycles by a hot section maintenance program. Figure 14 subdivides the predicted deterioration by module. As seen, the low-pressure compressor and high-pressure turbine are most sensitive to early flight-load induced deterioration. Erosion of airfoils and seals is the prime contributor to long-term deterioration in the cold section as shown on fi gure 15, while thermal distortion 1S the prime contributor in the hot section. One more refinement of these models will be made after completion of the flight loads test data analysis.

To validate the models at crU1se condi tions, it was fir s t necessary to establish actual in-flight average performance. The engine condition monitoring and in-flight calibration data collected on unrepaired Pan American 747SP/JT9D-7A eng1nes from start of revenue serV1ce to 1500 flight cycles provided this performance data. Performance at cruise condi tions was determined to be less sensitive to component deterioration than at sea level.

This reduced sensitivity results from the fact that the ram pressure ratio increases the nozzle pressure ratio at cruise and, thus, makes performance less sensitive to gas generator losses. This effect has been demonstrated in the Pratt & Whitney" Aircraft (Willgoos) altitude test facility. The result is that the increase in cruise thrust specific fuel consumption due to component deterioration is about 75 percent of the increase at sea level. The JT9D performance retention model supports the results and was used to develop the curves on figures 1 and 2. Evaluation of cruise performance data from the flight loads tes twill permi t a further refinement of the cruise performance retention model.

CONCLUSIONS Performance deterioration in the JT9D-7 is a flight sensitive phenomenon caused by a short-term and two long-term wear modes. The short-term deterioration occurs primarily during airplane acceptance testing and, therefore, does not affect airline operation. The long-term wear takes place continuously over the eng1ne life so that the performance loss can be minimized by a sound maintenance program. Short-term deterioration is primarily due to flight-load induced blade and gas-path seal wear which result in increased gas-path running clearances. The wear occurs in all engine modules but has the most deleterious effect on the low-pres sure compressor and high-pressure turbine performance. The wear occurs during conditions that combine minimum axisymmetric runn~ng clearances and maximum engine distortion or asymmetric closure.

Minimum axisymmetric clearance occurs during high power operation due to the combined effect of centrifugal forces and high metal temperatures. Maximum asymmetric closure is caused by airplane maneuver induced aerodynamic loads and thrust induced engine bending loads. Thus, short-term deterioration is a cyclic effect in that it occurs during take-off and cl i mb and other maneuvers which combine high aerodynamic loads and high engine power. Cruise, approach, and landing do not contribute to short-term deterioration.

Long-term deterioration is also flight cycle dependent. It is caused by erosion of the airfoils and seals, which cause airfoil roughness, bluntness, chord loss, and increased gas-path clearances, and by thermal distortion of turbine airfoils which reduces their efficiency and increases leakage.

Ingestion of foreign matter during taxi, take-off, and landing operations is the primary cause of erosion. Changing gas-flow patterns caused by erosion plus thermal cycling of the engine are the prime causes of thermal distortion of the turbine airfoils. The split of deterioration by module on figure 14 shows the increasing importance of high-pressure compressor and low-pressure turbine in long-term deterioration.

The refined performance deterioration prediction model shows a 2.1 percent increase in cruise thrust specific fuel consumption in the first 1500 revenue flight cycles. The program results also show that a good program of both hot section and cold section maintenance can maintain cruise performance between 1 and 2 percent of that at start of revenue service.

RECOMMENDATIONS Based on the results of the completed phases of the JT9D Diagnostics Program and the preliminary results of the Flight Loads Test Program, the following recommendations, summarized on figure 16, are made toward improved performance retention in current and future propulsion systems.

Performance retention in current engines can best be maintained by following improved maintenance practices which have been developed jointly with the airlines, based on the early findings of this and i ndustry-sponsored programs. These improved practices provide both cold-section and hot-section refurbishment, and the potential results are summarized on figure 17.

Performance retention in future propulsion systems will benefit from the following: o Performance deterioration caused by flight loads and thrust-induced loads will be minimized by integrated engine and nacelle designs that consider the effects of both flight loads and thrust loads.

o Further development of gas-path clearance control systems and abradable rub strips will provide closer runnipg clearance control 1n the high-pressure turbine.

o Erosion effects on cold section airfoils and seals will be minimized by improved coatings and materials and the consideration of refined gas-path designs for reducing the ingestion of erosive material.

o Erosion effects on hot section airfoils and seals will be minimized by improved high temperature materials.

o Thermal distortion effects will be minimized by refined gas-path designs and improved maintenance programs that reduce temperature profile shifts.

TABlE I COMPARISON OF MODUlE CONTRIBUTION TO SEA lEVEL TSFC DETERIORATION Historical Data In-Service Engine P&WA Testing S imul ated Aero- Analysis Analysis of P-695743 dynamic Loads 049 Cycl es ) 050 Cycl es) (141 Cycles) Test of P-662211 Change in TSFC (%) at Sea Level Static Take-off Thrust Ln N Fan +0.1 +0.2 +0.1 +0.2 Low-P ress ure Compress or +0.2 +0.4 +0.4 +0.3 High-Pressure Compressor +0.3 +0.3 +0.2 +0.2 High-Pressure Turbi ne +0.4 +0.4 +0.6 +0.5 Low-Pres s ure Tur bi ne +0.5 +0 .1 +0.1 +0.1 TOTAL +1.5 +1.3 +1.5 +1.3 ~

} E~oo

ATSFC

---------

Thermal

- }

Causes

at

distortion

cruise

o

losses prior to Blade and seal rubs aircraft delivery 1 1000

Revenue flight cycles

Figure 1 JT9D-7A In-Service Engine Performance Deterioration at Altitude Cruise Conditions. (J24873-2)

% increase

No repair

in cruise 2

TSFC

Hot section " " " + cold section refurb 1000 2000

Flight cycles

Figure 2 Effect of Repair on JT9D-7A Engine Cruise Thrust Specific Fuel Consumption. (J 24603- 24)

1. Define scope of J19D performance

deterioration

2. Identify and quantify sources of J19D

performance deterioration

3. Determine sensitivity of component

performance to engine parts deterioration

4. Develop analytical model of J19D

performance deterioration

5. Recommend performance retention

techniques for current and future engines

Figure 3 JT9D Diagnostics Program Objectives.

(J23878-4) o D-7A o D-7A (SP) Sea level static take-off conditions 6 D-7F

o D-20

Acceptance testing + 3t--.c:.....-.-t------Reveooe service -----

o _---------0- +2(7'

% ~TSFC +2

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

relative to o

production + 1

__ -0-'-2(7' O~~-~-~---~-~--~-~ 50 100 150 250 300

Flight cycles

Figure 4 Historical Short-Term Deter i oration Data.

(J24873-4) AIRLINE BID-3A ..... _ AIRLINE A/O-7A __ -- ..... "'" AIRLINE C/O-7 , -7A JT90-3A / 7/ 7A .......... ~--- _ ::: - - ___ - - AIRLINE E/O-3A

--

-

...".~---- -----

~ --

% 11 TSFC

/" --

/. --

relative to --;,?'

--

/'

start of

revenue

Sea level static conditions

service

2 3 4

Flight cycles - 1000

Figure 5 Historical Long-Term Deterioration Data for Unrepaired Engines.

(J24603-8) 5.0 HistlJricaI 4.0 Brine avnee , !J.TSFC 3.0 I\.mllm(sp) relative I JIWII*" to 2.0 production

~t~/

InstJIed 1T9D- 7 Sp 1.0 HistlJricaI

o

1~~~--~~~15~00~~2~00=0--~~~=-~ Fight cycles

• Installed ground test from 0 - 1100 flight cycles

• Expanded testing and analytic teardown at 141 cycles

• Pre and post repair calibrations

Figure 6 Pan American 747SP/JT9D-7A In-Service Engine Performance Data.

(J24873-6)

Percent

change in

fuel flow

-4 28 engines - 1398 points

o 500 1000 1500

Flight cyc l es

Figure 7 Cruise Fuel Flow Trend with Usage for Pan American 747SP/JT9D-7A Unrepaired Engines. (J24873-7) Predicted aero loads Simulation 102571b NS NS 5648 Ib L.H. SIDE VIEW ITYPI FRONT VIEW ITYPI

Maximum resultant at II A" flange

Simulated Predicted

Moment 356,288 356,116

Figure 8 Inlet Air Loads at Take-Off Rotation. (J21704-193) Figure 9 X-Ray Facility with Test Engine Installed .

(J24603 -15 )

• Measure typical aerodynamic and

in~rtia loads during acceptance

test and revenue service

• Explore effects of gross weight,

sink rate, pitch angle, and

maneuvers on nacelle loads

• Measure engine clearance closures

and engine performance resulting

from the airplane maneuvers

• Provide data for improved propulsion

system designs

Figure 10 Flight Loads Test Program Objectives.

(324603-3 ) Figur e 11 I nstrument ed JT9 D- 7A E ng ine Installed on RA 001 Airplane for Fligh t Lo ads Te st . (J24603 - 22)

• Analytical build of instrumented fan

and high pressure turbine

• Initial engine calibration in test cell

• Installed engine ground calibration

• Production acceptance test flight

at 550,000 Ib takeoff gross weight

• Installed engine ground calibration

• Wind-up turns to 26 s

• Installed engine ground calibration

• Heavy gross weight takeoff , maximum

dynamic pressure and maximum Mach number

• Installed engine ground calibration

• Final engine calibration in test cell

• Analytical teardown of fan and high pressure turbine

F i gur e 12 T est a nd In spe ction Se qu ence for Flight Loads Test Program '.

(J240l 8- 6) Probe - Locatm 60 : Probe - locatm 330 : LI Sin.irtId SO o ActuII CIennai - SO dlqe .

~

n*I - 100 -1 SO Pr obe - lDc::atm 1 SO : - lDc::atm 2411 : Probe SO / " - so CIwn:e lIkecff dlqe.

- 1 00 ~ n*I -1 SO ,

- 200 ~lJ1Dff

- 25(J ltnJ Rf'III Figure 13 Comparisons of Simulated and Actual Flight Test Fan Closures; Acceleration to Take-Off. (J 24873-10) DFan CJLPC fZ::J HPC ~HPT DLPT

% ~TSFC

re!ative to

PfoOOction

Figure 14 JT9D-7 Sea Level Perfomance Deterioration Distribution by Engine Module. (J24603-21)

o 6 CIwn;e - fiirt Dads

rzzl AifoIl anl seal erosioo ~ Thermal cistnrtioo + 2 1!XXl ~

% flTSFC

~ +1

relative to

production

Total % TSfe + 1.0 + 2.0 + 2.9 +3.8

+4.4 detericntion Figure 15 JT9D-7 Sea Level Performance Deterioration Distribution by Cause.

(J 24603- 23) Current engines:

• Cold section refurbishment

Future engines:

• Flight-load resistan t propulsion systems

• Erosion resistant design and materials

• Thermal distortion resistant design and materials

Figure 16 Flight Lo ads Test Pr o gr am Recommendations. (J 24 765-17)

% increase

No repair

in cruise 2

TSFC

Hot section '· + coil section return 1000 2000 3000

Flight cycles

Figure 17 Effect of Repair on JT9D-7A Engine Cruise Thrust Specific Fuel Consumption.

(J24603-24)

Page intentionally left blank

JT8D ENGINE PERFORMANCE RETENTION Albert D. James and David R. Weisel Pratt & Whitney Aircraft Group SUMMARY The attractive performance retention characteristics of the JT8D engine are described. Because of its moderate bypass ratio and turbi . ne temperature, and stiff structural design, the performance retention versus flight cycles of the JT8D engine sets a standard that is difficult for other eng~nes to equal.

In addition, the significant benefits of refurbishment of the JT8D engine are presented. Cold section refurbishment offers thrust specific fuel consump- tion improvements of up to 2 percent and payback in less than a year, making a very attractive investment option for the airlines.

INTRODUCTION The ability of an aircraft powerplant to retain its marketed performance is one of the primary considerations in its development or selection. Escalat- ing price of fuel during the last decade has placed greater emphasis on defin- ing the performance retention characteristics of current, as well as new powerplants. Any performance loss can be directly related to increased operat- ~ng costs, and fuel is an ever increasing portion of those operating costs.

The Pratt & Whitney Aircraft JT8D enpne ~s a first generation turbofan engine, i.e., a low bypass ratio, dual spool, axial flow turbofan that enjoys the distinction of being the most widely used engine in commercial service. To date, over 10,000 units have been delivered and power more than 3,000 aircraft for 175 airline operators worldwide. To assist these operators in minimizing fuel costs and maximizing time on-wing, Pratt & Whitney Aircraft has conducted an extensive analysis of the JT8D engine in order to define the industry po- tential for performance recovery, the modes of performance loss within the component modules and cost effective means of recovering that lost performance.

PERFORMANCE CHARACTERISTICS Quantifying performance losses as a function of engine service time has been an industry problem since the initiation of commercial serv~ce. Manufac- turer development processes include many tests with cyclic power adjustments that provide advanced notice of problems ar~s1ng from thermal distress and fatigue. They do not have the means of accelerating the normal erosion process that the gas path hardware is subjected to daily in commercial service.

However, operators have all these conditions occurring continuously, but usually do not have the facilities, manpower, financial commitment or time to do the extra testing that would provide this basic knowledge.

Reams of inflight cruise monitoring data could be reviewed. Almost all airlines require the flight crew to record one data point during cruise on each engine during each flight leg. Five performance parameters are recorded: engine pressure ratio, high and low spool rotor speeds, fuel flow and exhaust gas temperature. The data quali ty suffers due to the use of aircraft instru- mentation. The principal use of the data is to monitor the engine parameters for abrupt changes that would signal a potential incidence of part failure.

Data scatter is significant. Considerable massaging is required to smooth out the data enough to generate trending information that will identi fy a severe deviation thereby making the identi fication of small performance losses very difficult. It was, therefore, decided to base the on-wing (pre-repair) deteri- oration characteristic on ground engine cell data acquired from engines run prior to repair.

Figure 1 provides the average performance retention characteristics of the JT8D engine and the individual sea level static data points used as the basis for this curve. The average curve is described by a second order curve fit through all the data, and is considered representative for all JT8D-l through JT8D-l7 models.

Figure 2 shows this characteristic in relation to the industry average post repair performance levels. Three levels of performance are shown that exist in commercial service today. The uppermost curve represents the average on-wing (pre-repair) deterioration. Individual operators may experience dete- rioration rates significantly different from this curve, either better or worse, since there are a multitude of factors that can influence the deterior- ation rate. These factors will be discussed in the later paragraphs.

Average industry performance levels after repair are represented by the middle curve. Industry repair practices have reflec ted the philosophy of on- condition maintenance, which were directed at minlmlzing engine maintenance cost. Cheap fuel prices and known operating limits have allowed operation un- der this maintenance philosophy. Engines are operating today that have accumu- lated operating times in the 30,000 hour or cycle region with periodic hot section repair and only occasional minor cold section work to repair airfoil damage from Bill of Material or foreign objects. Refurbishment of the compres- sors to recover performance loss (particularly specific fuel consumption) was dismissed by operators as not being cost effective. These components were generally refurbished only when incidences of compressor surge were en- countered.

This curve then reflects the results of on-condition maintenance philoso- phy with an implied penalty because of limited cold section repair. The curve flattens to a relatively constant loss resulting from the imposition of the operating temperature limit. Gas path deterioration and increasing exhaust gas temperature coexist. Generally, the JT8D engine has not had a problem with the exhaust gas temperature measuring system indicating temperatures that are falsely high or low. Therefore, as the gas path deteriorates, the exhaust gas temperature limit restricts takeoff power setting, forcing the engine back to the repair shop. The practice of minimum repair then manifests itself in ever shorter periods of operation between repair cycles and exhaust gas temperature limiting operation.

------------------------------------------------------------------------------------.-- ------ Fuel shortages and the continuing escalation of fuel prices have refocused the industry emphasis on fuel economy, both in operation and maintenance.

Practices that were not CO&t effective a few years ago, now have reasonable payback periods.

In the interes t of identi fying effective refurbishment, Pratt & Whitney Aircraft funded a JT8D Engine Maintenance Technology study to define the pri- mary modes of deterioration in the engine, identify the modules where maximum performance restoration could be attained and develop cost effective methods of recovering that lost performance. Cooperation of four major airlines (three domestic and one foreign), two years of studying hardware, scrap records and available data, and an in-house testing program using a loaned high time ser- vice engine, have identified and demonstrated that significant performance could be recovered through fan and high pressure compressor refurbishment.

The lower curve (Fi gure 2) represents the average level of performance attainable through a revised maintenance philosophy that recognizes periodic cold section refurbishment as a major part of performance recovery.

It would appear that a significant improvement is still available, since the average remains approximately 2.0 percent above the new engine baseline.

The curve should not be interpreted to conclude that refurbishment to the "as new" performance level cannot be achieved. A few of the engines came within 0.5 percent of their production acceptance levels. Current production perform- ance levels were used as a consistent base for all the post repair tests and the average engine performance has improved over the years. This would tend to make the difference appear larger than it actually was. Realistically, achievement of "as new" performance can be accomplished through refurbishment.

Although material lost to erosion cannot be restored, airfoil shape and sur- face finish, which are the prime performance factors, can be restored with only a residual reduction in performance life.

FACTORS INFLUENCING PERFORMANCE RETENTION Assuming large amounts of usable data were available to define the per- formance retention characteristics for each operator, a wide variation in operator average retention and engine to engine variations about each average could be expected. Engine deterioration rates are dependent on many factors other , than engine to engine differences of gas path geometry and cycle match- ing. Of major influence are operational environment, operational philosophy and maintenance philosophy, which are particular to each individual airline.

Gaspath Geometry The JT8D eng1ne 1S a first generation turbofan engine, 1. e. a dual spool, axial flow, low bypass turbofan. It is constructed with a full length annular fan duct that directs fan air rearward to mix with primary air in a connnon convergent exhaust nozzle. This results in rigid case construction so the installed engine is not adversely affected by the axial bending forces exerted by inlet air loads during aircraft rotation and maneuvering.

In comparison to more current state of the art engines, the gas path is built with relatively loose clearances between rotating and stationary parts, therefore, compressor and turbine airfoil tip rub is not a factor. The primary forms of performance loss are hot section thermal distress, compressor airfoil roughness and compressor airfoil erosion.

Airfoil roughness is the result of environmental contamination and gener- ally occurs in the first 1,000 hours of operation after installation or c l ean- ing. Contamination is the minor mode of compressor performance loss that ap- pears to reach its maximum and then remains relatively constant throughout the operational phase.

High gas velocities and thinner airfoils 1n the high pressure compressor account for a large measure of the performance impact. In addi tion to the leading edge erosion exhibited in the fan and low pressure compressor, blade pressure side erosion also occurs. Both leading edge and pressure side erosion occur concurrently until approximately 4.0 percent chord is lost near the blade tip. At this point, the blade trailing edge has become excessively thin and begins to tear away. Figure 3 graphically shows the relative performance loss of each deterioration mode and the accelerated rate for performance loss that takes places when the trailing edge shreds.

Operational Environment Operational route structure and geographical area have a defini te impact on the deterioration rate. The JT8D engine is used 1n the short to medium range segment of the industry. Aircraft cyclic times, the time between ta ke off and landing, vary between 20 minutes to two hours per cycle. The in d ustry average is about one hour per cycle, which is considerably shorter than the typical long range aircraft which run four to seven hours per cycle.

Hot section distress, the primary mode of engine performance loss in all engines, is cyclic dependent. Time at high temperature is the key parameter in determining turbine hardware lives. Since maximum temperature is achieved dur- ing takeoff, deterioration as a function of engine cycles becomes most signi- ficant. On this basis, the JT8D engine performance retention characterist i c is very attractive.

Cold section airfoil erosion, the principal mode of performance loss in the compressor, is cyclic dependent, and also dependent on geographical area.

A Middle East operator typically exhibits more airfoil erosion due to sand ingestion, in 2,000 cycles than a large domestic operator may see in 30,000 cycles. Figure 4 illustrates the variation in erosion rates by assuming that the rates are linear.

Other geographical areas have their own particular problems, although not as severe as the desert operations. Operators in Alaska also have an erOS10n problem with volcanic ash and operators near salt water or highly industrial- ized areas have problems with corrosion and sulphidation.

Operational Philosophy Power setting procedures that reduce the impact of hot section thermal shock during takeoff have been defined by engine and airframe manufacturers and are available to the operators. The procedures simply require the flight crew to compute the required power necessary to get airborne based on the am- bient temperature, aircraft gross weight and runway length. Dependent on these variables, reduced power may be used instead of the full rating. This results in the engine operating at a lower turbine inlet temperature. A signi ficant reduction in turbine distress and, consequently, a reduced deterioration rate can be realized if reduced power takeoffs are used whenever possible.

Figure 5 shows an analytical assessment of maintenance material cost re- sulting from reduced power takeoffs. The relative cost is shown as a function of engine time per cycle with lines of constant percentage of power reduction.

Maintenance material savings are directly related to a reduction in turbine distress. Similar savings can be realized by using reduced climb power when- ever aircraft loading and routing permit.

Maintenance Philosophy Maintenance policy has been an evolving process as illustrated ~n Figure 6. During the early years of operation, maintenance was on a hard time basis.

Engines were removed, inspected and repaired at specific intervals. As confi- dence in the hardware integrity became proven, the time intervals were extend- ed based on operator experience.

Logical progression led to on-condition maintenance. This was considered at one time to be the ultimate policy. Engines stay on-wing until removal is forced by a fault or the inability to set power because of reaching an oper- ational limit. Some parts still retain hard time limits based on maximum cycles for useful life, but generally these are long term limits. All interim maintenance is based on a problem developing or the engine reaching the ex- haust gas temperature limit pr~or to meeting the power setting requirement.

This was an acceptable maintenance policy during an era where fuel cost was a small portion of total airline operating cost.

Minimal maintenance results in increased fuel consumption and turbine tem- perature levels. Turbine distress occurs with increasing regularity and ~s readily evident during disassembly. Compressor deterioration is not as ap- parent, therefore, the extent of cold section maintenance was to blend repair obvious Bill of Material and foreign object damage.

The tightened world supply of fuels and strategic materials highlighted the need to change philosophies. Fuel became a major part of the airline oper- ating cos t and the maintenance emphasis was redirected toward reducing fleet fuel consumption. The impact of the cold section, fan and high pressure com- pressor particularly, became more obvious, forcing maintenance philosophy to I swing back towards scheduled cold section refurbishment.

PERFORMANCE OPTIMIZATION With the maintenance philosophy shift towards scheduled compressor refur- bishment, the next logical progression is towards complete engine maintenance management. The deterioration rates of each module, and the potential to re- cover lost performance are recognized in this concept. This utilizes a cost effective mix of repair techniques and new parts to attain a service goal.

This also must address on-wing maintenance to retain the performance re- covered through refurbishment for as long as possible. Periodic engine water wash and fuel nozzle cleaning should be included in any engine management pro- gram.

Compressor airfoil roughness has been found to occur in the first 1,000 hours of operation. Experience has shown that long term water washing is not very effective in removing compressor contamination. However, short term per- iodic washing (250 hour intervals) has demonstrated that performance loss due to contamination can be held off for 4,000 to 5,000 cycles. Intervals recom- mended for compressor water wash have been in 1,000 hour increments or less, but as more data is accumulated, the most significant results are associated with the shorter time interval wash procedures.

Considerable data from hot section inspections have correlated burned tur- bine nozzle vanes and streaked combustion chambers with coked fuel nozzles.

Non-uni form fue 1 flow from plugged nozzles resul ts in excessive temperature and burned vanes. On-wing fuel nozzle cleaning procedures have been proven ef- fective in eliminating moderate coke deposits. Therefore, periodic cleaning in the recommended 1,000 hour intervals will maintain coke free operation and op- timize turbine hardware performance.

PERFORMANCE RETENTION MODELING In order to determine more cost and fuel efficient maintenance practices, an accurate model of engine performance loss must be constructed. This model can be used to quanti fy performance loss for a particular engine model as a function of usage. Additionally, a complete model will identify losses both by module and cause, such as erosion of airfoils, thermal distortion of hot sec- tion parts, and clearance increases between rotating and stationary parts due to erosion or rubs.

To begin the process of model construction, appropriate data sources must be selected. Inflight monitoring data is available, but its usefulness as a primary source H limited. The quality limitations of flight data have been discussed previously; large amounts of data must be trended in order to be meaningful. In addition, flight data furnishes far fewer parameters than the number of module losses to be defined. In particular, the lack of thrust mea- surement makes determination of low spool losses very difficult. Finally, flight data is affected by aircraft systems, particularly the pneumatic system, which makes analysis of the data even more difficult.

Two data sources were used to construct the JT8D eng1ne performance reten- tion model. The first of these was pre-repair data obtained from airline test cell runs. Because of the nature of airline operations, particularly with a well-proven engine like the JT8D and prevailing on-condition maintenance prac- tices, pre-repair data is not normally obtained. However, there is a limited amount of first run data available. This data 1S relatively accurate and furnishes pressure ratios and thrust, in addition to more accurate definition of those parameters measured inflight. By comparing the pre-repair data for a particular engine to its production run data, it is possible to use a computer simulation of the engine to calculate efficiency and flow capacity change for every module. This analysis depends on use of known relationships between cold sections efficiency and flow capacity changes, referred to as "coupling" rela- tionships.

The other data source used for the model was teardown inspection data.

Used parts with known service times were collected and analyzed to determine performance changes from the new part configuration. Such data furnishes know- ledge of the causes, as well as the magnitude of module performance loss. From the estimates of individual module performance changes from new, total engine performance change from new can be synthesized with the engine computer simul- ation. The results of this analysis can then be compared to the pre-repair analysis and then both analyses can be iterated until reasonable closure is obtained, as shown in Figure 7. Since each analytical approach has its limita- tions, such iterat i on enhances the validity of the solution.

This process was performed for the JT8D-9 engine. Figure 8 shows a compar- ison of increase in sea level takeoff exhaust gas temperature from production as predicted by the model, compared to the pre-repair data. Figure 9 shows the same comparison for thrust specific fuel consumption. Reasonably good correla- tion between the average of the data and the model is shown. The used part analysis showed that the component deterioration for a typical operator was most strongly related to flight cycles for reasons previously discussed. How- ever, there can be considerable operator-to-operator variation associated with operating environment differences.

Modules losses at 4,DOO and 8,000 cycles that result from the analysis are sh o wn in Figure 10. Cold section losses are dominated by erosion and roughness damage, while hot section losses occur primarily in the high pressure turbine and are largely the result of thermal distortion (vane bow). There are no module losses due to clearance increases. The low bypass ratio of the JT8D en- gine minimizes thrust bending loads, and the long, stiff one piece fan duct effectively isolates the internal engine cases from nacelle aerodynamic loads.

These features, plus the moderate hot section temperatures, result in a stand- ard of performance retention that is difficult for other engines to equal.

Figures 11 and 12 show the overall impact of cold section versus hot sec- tion losses on exhaust gas temperature and thrust specific fuel consumption, measured at sea level takeoff condi tions. The hot section dominates exhaust gas temperature increase, while cold section losses have greater thrust speci- fic fuel consumption impact. Historically, engine overhaul has been directed primarily toward restoring exhaust gas temperature margin, and airline efforts were accordingly concentrated on hot-section repair. However, in the current era of constantly escalating fuel prices, Figure 12 shows the importance of periodic cold , sec tion refurbishment to m1n1m1ze fuel burned.

In order to analyze the potential benefits of performance restoration on fuel burned, fuel consumption at altitude conditions must be evaluated. This can be done with the engine computer simulation. Analysis shows that there is an effect of both flight condition and power setting on thrust specific fuel consumption change due to component performance changes. A weighted thrust speci fic fuel consumption change has been defined which combines both climb and cruise thrust specific fuel consumption at typical altitude conditions, in proportion to the fuel consumed during a typical mission. For the JT8D engine in a typical 727 application, the weighted thrust specific fuel consumption change is approximately the average of climb and cruise changes.

Takeoff exhaust gas temperature and weighted thrust specific fuel consump- tion increases are shown by module and cause in Figures 13 and 14. Exhaust gas temperature increase is primarily controlled by the high pressure turbine; however, high pressure compressor losses also contribute significantly. Thrust specific fuel consumption increases are dominated by the fan and high pressure turbine; the high pressure compressor again contributing significantly.

The performance retention model can also be used to predict the impact of hot section repair, and project performance losses for multiple run engines.

Figures 15 and 16 show the impact of hot section repair only. Typical first removal is shown at about 6,000 cycles, which would probably be for foreign object damage or hot section inspection. At this point, a change in high pres- sure turbine outer airseal from bill-of-material knife edge to honeycomb (typical airline practice) is shown. Erosion of the replacement outer airseal plus worse bow for repaired vanes resul ts in more rapid exhaust gas tempera- ture and thrust specific fuel consumption increase with flight cycles, so that an exhaust gas temperature-limited condition may be encountered at about 10,000 cycles. The effect of hot section only repair ~s that the interval between the exhaust gas temperature-limited shop visits decreases with succes- sive shop visits. This is because of the underlying cold section damage which has not been corrected.

IMPACT OF COLD SECTION REFURBISHMENT The assumed works cope for cold section restoration is shown in Figure 17.

If cold section refurbishment ~s accomplished at "soft" time intervals of 12,000 to 17,000 cycles (nearest convenient time when the engine is in the shop), the model predicts the results of Figure 18. Fan and low pressure com- pressor restoration are accomplished at 13,500 cycles, along with high pres- sure turbine repair. At 17,000 cycles, the high pressure compressor and tur- bine are repaired. The benefits in improved thrust specific fuel consumption are readily apparent.

A number of studies have illustrated the cost-effectiveness of cold sec- tion refurbishment for the JTBD engine. Figure 19 illustrates summary results from one study. The study showed fan and low pressure compressor refurbishment to be most cost effective (earliest payback). Refurbishment of fan, low pres- sure compressor and high pressure compressor combined, resulted in 1.6 percent o weighted thrust specific fuel consumption recovery at refurbishment, 2l F exhaust gas temperature recovery and payback in less than a year for a typical operator, based on 50 cents per gallon of fuel. Current fuel costs would further enhance cost effectiveness of cold section refurbishment, notwith- standing increased labor and parts costs since this study was conducted.

REFERENCES 1. Sallee, G. P., JT8D Maintenance Technology Study, United Technologies Cor- poration, Prat t & Whitney Aircraft Group, Commercial Products Division, June 1980.

2. 1980 JT9D Engine Regional Engineering and Maintenance Conference, Pratt & Whitney Aircraft Group, Commercial Products Division.

October 1980, Zurich, Switzerland and Washington, D.C.

November 1980, San Diego, California and Singapore Prerepair/Takeoff thrust

•

+7

•

o +6 o o

+5 o

•

Percent

change in

+4

thrust

o H8D - 1, -7 o JT8D-9

specific

6 JT8D-15 +3 o

fuel o JT8D-17

o Open syrmols - first run

consumption

+2 (no previous repairs) Closed syrmols - fTl.I~e run

o (have had at least one

+1 previous repcW) 4 6 8 10 12 14 16 18 20

Flight cycles, 1000

Figure 1 Sea Level Static Test Cell Data JT8D Engine Performance Retention AppUcable to all H8D models +6

Percent

+4

change in

thrust

-2%

specific

fuel

+2

t

consumption

-

Demonstrated post refurbistment

o~ ____ ~~ ____ ~ ______ ~ ______ ~ __ __

o 4 8 12

Flight cycles, 1000

Fi gure 2 Industry Average Post Repair Level Shows Significant Potential for Performance Recovery I /

/

Tip section / Rapid chord loss / / I I

/ /

II I /

Change in

/ II /

TSFC and

// / I EGT" , ./ / I _ ..... / I I

loss of

/ I Tramng edge

stall

./ / angle _// /

margin

---~ / / Leading edge angle

-'l-- T~ ~r:e_

- --

o 1 2 3 6 8 10

Average chord loss (percent at 0.100 inches from tip)

Figure 3 High Pressure Compressor Performance Retention Directly Related to Chord 10ss Middle East desert operator loss of trailing edge

""""""""""" ~~ •

Chord

loss

large North American operator 2 4 6 8 10 12 14 16

a

Flight cycles, 1 000

Figure 4 High Pressure Compressor Blade Chord 10ss Rate Varies Greatly Between Operators 1.5 A ver ag e industry ratings (4% derated)

Relative

1 0% derate

maintenance

mater i al

1.0

c ost

1.0 1.5 2.0

Hour s /flight cycle

R ed uced tak eo ff thrust can reduce maintenance material cost substantially Most c os t effective benefit realized within 10% derate Figure 5 Maintenance Cost Reduced by Thrust Derate Procedures 100% ~- ---.c- --......-------------.----- __ _ Module threshold refurbislvnent Percen t On-condition maintenance of CondiOOn lTIOIIitoring air l ine m aintenance programs Ha rd time engile overhaul 1963 1969 1980 HaD Widebody introd u ction introdu c tion Figure 6 Engine Maintenance Concepts are Changing

-I

Iterate to analyze __ _ module losses Iterate to synthesize engine performance Figure 7 JT8D Engine Performance Retention Model-Approach First run JTSO-9 engine Sea level static/Std day/Takeoff thrust

so

•

Change in

exhaust gas

Adjusted temperature, for

•

of

profile shift -0 2 4 6 S 10 12

Flight cycles, 1 000

Model Agrees with Test Data Figure 8 First run JT8D-9 engine Sea level static/Std day/Takeoff thrust +6

•

•

Percent

change

+4

•

in thrust

specific

fuel +2

consumption

Flight cycles, 1 000

Figure 9 Model Agrees with Test Data First run JT8D-9 engine Damage mechanism 8,000 cycles 4, 000 cycles -2

o Erosion

Fan ~ Aifoil rQUihness Change ~ Vane bow Fan in -1 • Vane leakage efficiency,

III Clearance

points HPT LPT 8,000 cycles 4, 000 cycles +6 HPT +4 Percent change +2 in flow capacity -2 Model Losses Derived from Teardown Data Figure 10 First run JTSO-9 engine model

Sea level static/ Std day IT akeoff thrust

Change in

exhaust gas

temperature,

of

Flight cycles, 1000

Figure 11 Exhaust Gas Temperature Increase Dominated by Hot Section First run JT80-9 engine model Sea level static/Std day/Takeoff thrust

Percent

change in

Hot section

specific

fuel

consumption

Cold section 2 4 6

Flight cycles, 1000

Figure 12 Thrust Speci fic Fuel Consumption Increase Controlled by Cold Section First run JT8D-9 engine model Sea level static/Std day/Takeoff thrust Damage mechanism

o Erosion

8, 000 cycles WL1 Airfoil roughness

HPT

~ Vane bow • Vane leakage

Change in

m Clearance

exhaust gas 20

temperature, 4,000 cycles

of

HPT

Total + 48°F

Figure 13 High Pressure Turbine Losses Major Cause of Exhaust Gas Tempera- ture Increase First run JT8D-9 engine model 30,000 ft alt/0.8 Mn/Std day/Constant thrust Damage mechanism o Erosion ~ Airfoil roughness + 1.5 ~ Vane bow 8,000 cycles • Vane l eakage

Percent

&II! Clearance

HPT

change in

4,000 cycles

weighted

thrust

Fan

specific

Fan

HPT

fuel +

consumption HPC

lPT

Total = + 1.2% Total = + 2.35%

Figure 14 Cold Section Erosion and High Pressure Turbine Vane Bow Major Causes of Thrust Specific Fuel Consumption Loss

Hot section repair only

+ 70 Average EGT margin installed in aircraft + 60~--

Change +50

in

exhaust +40

gas

temperature, + 30

of

+ 20 + 10 2 4 6 8 10 12 14 16 18

Flight cycles, 1000

Figure 15 JT8D-9 Engine Multi-Run Model Projects Realistic Exhaust Gas Temperature-Limited Removals

Hot section repair only

30,000 ftJO.80 MnlStrl. day/Constant thrust

Percent

+3

change

in

+2

weighted

thrust

+1

specific

fuel

0~-~-~-~--~~~~~~--~--~---7

o 6 8 10 12

consumption

Flight cycles, 1000

Figu r e 16 JTBD-9 Engine Multi-Run Mode l Shows Importance o f Cold Section Losses

Module

Workscope

Fan - All blades - SWECO clean, chamfer cut - All blades - SWECO clean, chord check, leading edge radius Low pressure restoration, replace as necessary compressor - All stators - vapor clean, check vane angle, re-angle as necessary High - All blades - SWECO clean, chord check, leading edge radius pressure restoration, replace as necessary compressor - All stators - vapor clean, check vane angle, re-angle as necessary Figure 17 Assumed Workscope for Compressor Restoration 30,000 ft/0.8DMn/Std. day/Constant thrust Hot section repair and fan/LPC refurbishment Without compressor refurbishment

Percent

+3

change

In +2

weighted

thrust

specific +1

fuel

consumption

Flight cycles, 1000

Figure 18 JT8D-9 Engine Multi-Run Model Shows Benefits of Cold Section Refurbishment

- - I

• Weighted TSFC recovery at refurbishment = 1.6%

I

Fan, lPC, HPC only

I

• EGT recovery at refurbishment = 21°F

\

• Payback period less than one year (typical operator)

Fi gure 19 Cold Section Refurbishment Cost Benefit Analysis I

\

I

I

\

I

I

I

Page intentionally left blank

PERFORMANCE RETENTION OF THE RB211 POWERPLANT IN SERVICE B. L. Astridge Rolls-Royce Limited J. T. Pinder Rolls-Royce Incorporated INTRODUCTION It is perhaps a statement of the obvious, but an understanding of the mechanisms of deterioration is essential in order that features to counteract performance degradation can be built into the basic design of an engine and nacelle. Furthermore, the interpretation must be continued in service for effective feedback to provide modifications which may be necessary in main- taining a satisfactory performance retention program.

The in-service assessment must, therefore, be accurate as to magnitude and causes and this requires consideration of: 1. The powerplant as a complete entity, i.e. the engine components and nacelle including the thrust reverser.

2. Measurement of performance in flight rather than by sole reliance on the scaling of test cell data to flight conditions (although some correlation should be possible).

3. The relationship of engine parts condition to overhaul performance and in-flight deterioration level of that engine.

Hence a performance retention program covers design, feedback to design, measurement and analysis of in-service experience and continuous review of the condition of engine components. These aspects are addressed by consid- eration of th e RB211 engine in service in both the Lockheed L101l Tristar and Boeing 747 aircraft.

PERFORMANCE RETENTION DESIGN FEATURES Basic Design Features Performance deficiency will arise wh~n main gas path airflow, either past blade tip seals or through internal cooling air passages or leakage overboard, is ex cessive. Further major sources of performance loss will arise in the event that significant blade damage or erosion or aero foil contamination arising because of dirt deposition on oil contaminated - components takes place.

The RB2ll was designed with features which address these problems, as indicated in Figure 1. Blade tip clearance control and internal cooling air passage seal clearance are considered in the structural design, thermal matching and the use of shrouded blades and stators. Overboard leakage is considered in both main engine casing design and powerplant sealing features such as the reverser seal. Limitation of core engine contamination and erosion is featured in the core engine intake config uration where the generous spacing of the core splitter relative to the fan, together with the wide chord fan, allows the majority of the ingested particles to be centrifuged through the bypass duct.

Details of the particular aspects relating to these features in the RB211 engine are shown in the following illustrations. Blade tip clearance control at all conditions is enhanced by the features illustrated in Figure 2. The three shaft configuration allows a short engine with stiff shafts and casings.

Furthermore, the features which determine the gas path are separated from the structural casings and mounted with radial flexibility so that inertia and thrust forces in flight have minimum effect on tip clearances. The effective- ness of these unique features is reflected in the lack of initial performance lost in flight relative to the test bed passoff performance, the RB211 having demonstrated a deterioration including any losses associated with the first flight, of less than 0.3% SFC (s pecific fuel consumption) in 100 flights from new.

Thermal matching of static and rotating engine components is illustrated in Figures 3 and 4. Engines are, even in normal operation, subject to rapid changes in operating temperatures. Serious mismatch in expansion of the static and rotating components will result in large compressor and turbine blade tip and cooling air seal component excursions and, hence, wear such that, at stablilized conditions, large clearances will occur with a resulting serious performance deficiency. Engine axial matching is illustrated in Figure 3 and shows the HP (high pressure) turbine static and rotating seal members are located to a common datum, the HP location bearing and with load path structures exposed to substantially the same temperatures thus ensuring common expansion. The equally important radial matching feature in the turbines is shown in Figure 4. The HP turbine features a thermal control ring to restrict thermal growth of the static seal member to match the growth of the rotating member, the expansion of which is restricted to a low rate due to the mass of the turbine disc. The IP (intermediate pressure) and LP (low pressure) turbine static seal members are located in the turbine casings and their growth is matched to that of the discs and blades by a combination of casing insulation and external cooling with undercowl ventilation air. X-ray data obtained at various conditions has been extensively used to optimize the design of the features which fix the relative position of the seals and the blade tip fins and this has been achieved without requiring introduction of a scheduled turbine casing cooling air control system.

The air used to cool the turbine casing is also used to ventilate under- cowl zones and is no more than the amount required for that purpose. An inherent performance loss does not, therefore, require debiting to the system.

However, this powerplant air is taken from the fan stream and excessive leakage of this air, either through the ventilation system or directly overboard via reverser seals, etc., has a powerful effect on cruise SFC and a mU0h smaller, and hence, more difficult to detect effect at static conditions. Consideration of the RB211 nacelle as a complete powerplant has included development of a powerplant leakage check for use in development programs and as a check of service experience. The efficiency of powerplant sealing with time has, therefore, been checked against new engine behavior as shown in Figure 5.

Feedback From Service No resume of design features would be complete without consideration of the essential feedback of experience from service. There are many examples arising from RB211 experience and a few have been chosen as typical of the aspects discussed hitherto. The first is in the area of tip clearance control where experience with the RB211-22B dictated extensive changes to the HP turbine sealing arrangements for the RB211-524. The second relates to the control of compressor contamination resulting from oil leakage and the third to elimination of excessive variation in powerplant ventilation air, both of which have resulted in more minor but nevertheless, significant changes.

Figure 6 shows the changes in HP turbine shroud segment design for the RB211-524 version as a result of experience gained with the RB211-22B. In the earlier design the leakage paths shown arise in time from component distortion and wear, the NGV (nozzle guide vane) support ring distorting under axial load and the shroud segment leading edge fretting and allowing leakage and loss of axial location. In the RB211-524 design the offending joint in the NGV support ring is eliminated and the shroud segment location is changed from axial to radial under which conditions the sealing is reinforced by aerodynamic loading. It should be further noted that the RB211-524 shroud segment is of such a form that the segment base protects the leading edge of the honeycomb from erosion.

A modification which was designed to prevent oil contamination is sketched in Figure 7. Experience has demonstrated that excessive front bearing housing oil leakage was arising from an over generous hydraulic oil seal chamber volume.

On shutdown the excess oil occupying this chamber leaks into the IP compressor and acts as a base for contamination. A simple modification which has alleviated the situation was the introduction of a liner to reduce the oil chamber volume.

The third category of modification which has been introduced because of service experience is associated with the sealing of the powerplant undercowl ventilation air. Excess airflow was arising from variation in the birdmouth of the seal between turbine section undercowl ventilation zones. The addition of a more positive seal, as illustrated in Figure 8, has cured this problem.

Integrated Exhaust Pressure Ratio (EPR) System In order to monitor a performance retention program it is necessary to be in a position to measure, accurately, the in-flight performance of the engine .

Fundamental to this problem is the ability to measure the thrust setting of the engine. This is achieved in the RB211 by means of the integrated EPR system.

Thrust is fundamentally related to nozzle pressure ratio and it, therefore follows that , provided it can be measured reliably, exhaust pressure ratio must be the most accurate means of power setting. However, this accuracy can only be achieved in a two nozzle engine by sampling the pressure in both streams and then deriving a mean of the two pressures which is weighted in proportion to the areas of the two streams. The RB211 integrated EPR system fits this requirement and its basic nature relating thrust and pressure ratio means that this relationship is virtually unaffected by deterioration. Hence, measurement of in flight performance is possible . The total system is shown diagrammatically in Figure 9 and shows the fan air to be sampled with three five point rakes and the hot nozzle by five four point rakes. Integration of the two nozzle pressures is aerodynamic by the simple block shown in Figure 10 , the integrated pressure being sampled between two appropriately sized orifices.

It should be noted that the fan nozzle pressure is always higher than the turbine exhaust pressure and the system flow is, therefore, always from fan to turbine exhaus t. Carbon or other contamination risk is, therefore, minimal.

As a further refinement, the RB211 integrated EPR system is trimmed, by electrical means , in order to produce a relationship between integrated EPR and thrust which is common to all engines . This is required to take into account the change in the integrated EPR/thrust relationship brought about by variation between units within build tolerances, of the factors which influence this characteristic. Figure 11 illustrates the principle of trimming. The test integrated EPR/thrust relationship is checked to be within quality limits determined from experience, and the trimmer is selected to adjust the relationship , within trimmer steps , to a common characteristic for all engines. The pilot uses the control lever to select an integrated EPR and the trimming procedure previously described ensures that each engine produces the required thrust at the selected power lever setting. The assurance of minimum thr u st with alternative setting procedures leads to some engines producing more thrust than required with consequent deleterious effect on component lives .

MEASUREMENT OF DETERIORATION In Flight The integra t ed EPR system described above , with the inherent insensitivity of its relationship to thrust with engine deterioration, allows consistent meas u rement of i n flight performance. However, the reliabili t y of in flight data and its method of interpretation can only be understood with some knowledge of the potential inaccuracies which can occur. At the risk of some over simplification Table 1 lists inherent inaccuracies arising from measuring instrument inaccuracies, both aircraft and engine, and, another prime source of variability, the normal aircraft Environmental Co ntrol System (ECS) . The latter source of variability comes about because engine air is supplied t o a common manifold such that engine to engine and duct loss variations can result in differing amounts of bleed being extracted from each engine in the installation. If the integrated EPR/thrust relationship were not substantially insensitive to deterioration a further variable for power setting at the basic parameter, EPR, would be necessary in attempting to interpret fuel flow changes as changes in fuel flow at a thrust, i.e., SFC.

Nevertheless it can be seen that up to ±2.2 % variation in fuel flow at an integrated EPR could arise between engines due to the measurement inaccu -· racies. In a fleet of aircraft it would be expected that the incidence of inaccuracy would, in most cases, be random and that measurement of fleet performance, as an average, be possible, but that individual engine performance would be somewhat less reliable. Mean fleet data for the RB211-524 engine is shown on Figure 12. The flight monitor line is that established as represent- ative of RB211-524 engines using both averages of larger quantities of flight crew recorded data and some specific aircraft audits. Variation in deterior- ation rates will arise from differences in operation procedures and rates as low as that shown in Figure 12 for specific audits have been recorded with confidence.

Having made these points, it should be noted that man y of the instrument inaccuracies of Table 1 may remain consistent during a particular installed life and, hence, become systematic for the engine in question. This will become further reinforced if data can be recorded with the subject engines either operating without ECS bleed or at least isolated t o a potentially reliable source, i.e., engine isolated to one ECS pack. If, indeed, the errors are systematic, it is possible to measure individual performance changes during an installed life, i.e., individual engine deterioration, and reasonable success has been achieved in that individually assessed engines have demon- strated characteristics similar to that established by fleet mean evaluation of flight data and comparison of pre and post-installation test bed data.

Test Cell Confirmation It has been noted that the relationship between test cell and flight performance can be distorted in the event of large variations in powerplant leakage overboard, the effect at cruise being 2.5 times that at sea level static. In addition, the effect of component efficiency changes at cruise are smaller than at sea level and, therefore, test cell measurements of deteriora- tion must be scaled down by the order of 25 % for comparison with flight. This reduction has been derived analytically from engine models and proven in sea level/altitude test cell comparisons.

However, control of powerplant leakage has been a consistent feature of the RB211 design, development and operation and the relationship of measured flight deterioration to that indicated by pre and post service test cell checks could be expected to show consistency only with application of the latter scale.

This is confirmed by Figure 13 showing the current summary of RB211 - 524 experience. The curve is the current mean of considerable flight data and represents the difference between first flight performance and the performance at any point during the installation . The test points shown represent the difference in SFC, on the test bed, between the pre - installation pass-off test and post-installation performance checks scaled by 25 % to represent component efficiency changes only.

Overhaul Deterioration The total deterioration picture for a fleet is a composite of the rate of deterioration of installed engines and the amount of performance recovered by overhaul. This is usually illustrated by the familiar "saw tooth" plot . The relationship of overhaul performance to new engine performance is a function of the degree of rework applied . The modular design of the RB211 engine per- mits overhaul of individual modules to be carried out without stripping the whole engine and, therefore, it is important to understand the sources of deterioration such that the appropriate emphasis can be applied as the opportunity arises during shop visits. Examples of the type of overhaul practices pursued for the RB211 follow but as introduction Figure 14 can be shown as illustration of the increased scatter, relative to new, which is inevitable in modular overhaul. The mean level associated with this is, however, a not unreasonable 0 . 5 to 1.0% above that of new engines.

OVERHAUL PRACTICES No discussion on performance retention would be complete without some reference to overhaul practices . The amount of overhaul work carried out at each maintenance shop visit for purely performance reasons needs to be judged against the economic return which can be expected from the work carried out.

This is a continually changing picture. As the price of fuel continues to rise it now becomes feasible to carry out work which would not previously have been considered to be worthwhile .

This final section of the paper illustrates the effect which various levels of overhaul can have on the performance of the RB211.

RB211 Fan Leading Edge Restoration The RB211 fan, in cornmon with all other fans, suffers from leading edge erosion which, if allowed to carryon unchecked, eventually leads to the fan having a 'square' leading edge with resulting loss in performance. Tests carried out with ex-service fans which have this amount of leading edge erosion indicate that if the leading edge is restored correctly, by mean~ of careful rounding of the leading edge, as shown in the upper illustration of Figure 15, and improvement in SFC of 0.4% can be obtained. If, however, the restoration is applied to a fan which has an eroded leading edge which is greater than 1.40 mm (.055 inches) thick it is essential that the blade is thinned and blended over the first inch of chord in order that the correct radius of lead- in g edge can be applied. If this procedure is not carried out and the restora- tion is done as shown in the lower half of Figure 15, the benefits obtained will be minimal.

RB211 IP Turbine Shroud Segments The maintenance of the minimum tip clearances at vital performance conditions, such as climb and cruise ratings, is essential if the best economy of operation is to be attained. A particular area where tip clearance control has a very powerful effect on performance is the turbine. As previously explained, the RB211 utilizes a particular structure and turbine casing cooling air system which ensures that the axial and radial growth of the rotors and cases are closely matched during the important parts, from the performance viewpoint, of the flight.

All RB211 turbines utilize a honeycomb static tip seal segment and many of these honeycomb seals are pre-profiled. The shape of this pre-profile is designed to be closely similar to, but slightly smaller than, the shape obtained from service wear patterns and a careful running in procedure is then carried out prior to the engine carrying out its performance acceptance test in order that the turbine can machine out the excess honeycomb in a controlled way, thus ensuring a precise matching of the turbine to the honeycomb seals. The benefits of this careful pre-profiling and running in procedure are that the turbine knife edge wear is minimized, thus allowing the turbine to be used again without the necessity for regular knife ~dge rework and that the performance obtained after this is that which will be attainable by the customer.

In overhauling these seal segments, however, it is vitally important that the honeycomb is correctly positioned on the carrier plate since, if this is not done, the correct relationship between the turbine rotor and its tip seals cannot be maintained. While this may appear to be a statement of the obvious on several occasions the quality of some of these reworked seals has not always been perfect. Figure 16 shows on outline of an IP turbine shroud segment and many examples have been observed where the honeycomb has been too short or out of position by significant amounts and tests carried out using new IP turbine shroud segments made with honeycomb mis-positioned by 1 cell (0.068 inches) width have indicated that a loss of SFC of approximately 1.5% can be expected. This loss in performance is directly attributable to the turbine becoming disengaged from its seal segment.

RB211 HP System Tolerances In the days when fuel was relatively inexpensive many tolerances were written into overhaul manuals which, although satisfactory from the mechanical standpoint, were not always optimized for fuel economy. Examples which are shown here concern the amount and type of damage to compressor blades, rotating air seal and rotor tip clearances. The following examples indicate the amount of improvement which can be achieved when the acceptable limits are modified to take into account the need for economy rather than considering mechanical integrity alone.

RB211 Rear HP Turbine Stepped Seal. On introduction of the HP feed standard of HP turbine, it was necessary to raise the chamber pressure to the rear of the HP turbine to maintain the correct bearing load. This was achieved by introduction of a balance piston seal and removal of HP3 flow restriction upstream. The flow of the cooling airflow is then controlled by the balance piston seal and the stepped seal and is critically dependent upon the qualit y of these seals. The cooling air to the rear of the HP turbine is illustrated in Figure 17.

In investigating the poor standard of overhaul of one operator it was discovered that the operator was consistently working to the maximum allowable radial clearance allowed by the manual for the stepped seal. The maximum limit had been set at a time when fuel was much cheaper and it presented no mechanical hazards. The limit was reduced to an amount which is more appropri- ate to fuel economy and, as shown on the CUSUM trend plot, Figure 18, the average level of TGT (Turbine Gas Temperature) has been reduced by an average of 3.6°C, equivalent to a reduction in SFC of approximately 0.35%.

Blade Dressing. In investigating the effect of overhaul manual limits on performance, with the intention of making economical modifications to the overhaul manual limits, the HP system illustrated on Figure 19 was removed from an ex-service engine which had just undergone a performance evaluation.

It was decided to arbitrarily reduce the amount of blade dressing associated with the HP compressor to one-half that allowed in the overhaul manual and to tighten the HP compressor tip clearance to within book minimum + 0.005 inches. In addition to this all HP turbine seals were checked for compliance with the book. The in~pection revealed that, with the exception of excessive dressing on one stage, always a subjective judgment, the machine was within book limits but not the stated goal . In order to achieve these goals two stages were rebladed to reduce tip clearance and dressing to the new limits and a few blades were changed for similar reasons on 3 of the other 4 stages.

In addition to this, although within limits, it was decided to fit new static air seals to the HP turbi n e. In the event the assessment was that the amount of dressing was reduced to a little over one-third of that deemed acceptable by the overhaul manual . The net effect of these changes was to reduce the level of TGT by 12°C and the SFC by 1.1%, an amount which is consistent with an improvement in HP system efficiency of 1.7% together with a reduction in cooling airflow of 0.8%.

The economics of these changes are now being assessed.

The design of successive members of the RB211 family of engines has been continually modified to take into account experience gained from service engines, but each has retained the original unique features of separation of structural and gas path casings and modularity together with the ability, by use of integrated EPR, to have its performance monitored accurately in flight. With the current price of fuel, a 1 % sustained reduction in SFC is worth approximately $30,OOO/year/engine and it is easy to see that fairly major changes in overhaul practice, which were once uneconomical, are now becoming eminently desirable and will become more so as the price of fuel continues to advance. The structural features of the RB211 ensure that the rate of deterioration in performance is low and that the cost of overhauling the engine is minimized.

TABLE I RB211 ENGINES ACCURACY OF IN FLIGHT MEASUREMENT OF FUEL FLOW AT AN EPR SOURCE OF POTENTIAL EFFECT OF VARIABILITY INACCURACY OF FUEL FLOW MEASUREMENT AT AN EPR FLIGHT CONDITION MACH NUMBER ±O . Ol ±l.O6 % TAT ±2.0°C ±O.46 % ALTITUDE ±30 METRES ±O . 46 % ENGINE INSTRUMENTATION EPR TRANSMITTER ±O.4S% ±1.7 % AND GAUGE FUEL FLOW ±O.S% ±O.S% ENVIRONMENTAL CONTROL SYSTEM FLOW VARIATION ±O.2S% ±O.S% BETWEEN ENGINES RMS ACCURACY ±2.2 %

RB 211 Performance retention features

Blade tip clearance Ingestion and control by erosion limiting --------- • stuctural design

features ~

• thermal matching

-

/ \ Low sensitivity to tip

~b===" q~ ~~~ti~IS~~~~ ~;~:: ~-+clearance changes by

:1 use of shrouded • turbine blades .NGVs • compressor stators FIGURE 1

RB211 Load carrying structure

Rear bearing housing Front bearing housing ----:-_ jlI~rf:it~~~ ncase Core engine 'A frames

~

View on rear of engine FIGURE 2

R B 211 H P tu rbine tip seal matching

STATIC MEMBER LOCATED VIA INNER CASING ...- /" I HP LOCATION I

BEARING ~ _ ___ \

BLADE FIN I MATCHED AXIAL LOCATED VIA I EXPANSION HP SHAFT I L __ _ __ __ FIGURE 3

RB211524 Turbine tip seals

~ Sel f-clearing Strip seals between se gments honeycomb seals and between N G V platforms Thermal control ring I I I LPl LP2 LP3 IP - All seals segmented FIGURE 4

l

RB 211 powerplant leakage

Leakage test rig

A ir pressu .r {B lan ~

Leakage Ex-service engine area - after 10 695 hours sq . cms.

/ 710 7 cycles - 0 -- 0 -- 0--0* -- 0 Checks on typical . production powerplant C rUise pressure difference

O~----~----~------~ __

1.1 1.2 est pressure ambient pressure FI GURE 5

RB211 Revision to HP Turbine shroud

segment design

Solid NGV support ring Segment Leakage sealing re inforced by paths aerodynamic loads RB 211 - 524 RB 211-22 FI GU RE 6

RB211 I P compressor rotor front stub shaft

hydraulic oil seal

IP 2 _____ ------1 ___ -------- L P Shaft Liner to reduce seal volume FI GURE 7

RB 211 Control of pod ventilation

Zone 4A/4B bulkhead

Seal .~ / segmen~ - -tt+-H+ttH-ftHt- Attaching parts unchanged FI GURE 8

RB211-22 IEPR system general arrangement

Cockpit Instrument Note : PF & P for production pass-off & E A IDS if required

j--- _/

r--_~ FIGURE 9

RB211 E PR System integrator block

Test/A.I. D.S

t

Pf/ P8 INTEGRATOR BLOCK ASSEMBLY FIGURE 10 --------

Selection of EPR trimmer on pass off

I

Max EPR __ _

r

EPR Typical test result I I Acceptable E P R test range based on mea su red pressures I min thrust Thr ust ..

FIGURE 11

RB211-524 In flight deterioration

Increase in fuel flow % Source : +2.0 Fli ght monitoring Flight audit +1 .0 1000 2000 3000 4000 5000 6000 7000

I Flight hours I

I I I I

600 800 1000 1200 1400 200 400 Equivalent flight cycles (approx) FIGURE 12

RB211-524 Comparison flight monitor and

test cell derived deterioration

Increase in fuel Derived from ti me flow % expired cell testing Fl i ght monitor +2.0 /

o 0 0

o

+1 .0

o

1000 2000 3000 4000 5000 6000 Flight hou rs

I

I I I I

I

200 400 600 800 1000 1200 Equ ivalent fligh t cy cl es FIGU RE 13

RB 211-524 Overhaul engine pass off

performance relative to new

S.FC.

relative to new •

•

% • • • • • • • + 2.0 •• • • •• • • • • • •• • • • • •

• • •

New engine range •• ~. ..~. -_ +1. 0 '--M ax -- -- -.-.~.- .

. .. - ..

••• • • .. .. ., ... .

• •

o +----- -----: .=-------:.:---4>-----"----------A ve

•

• • •

•

-

• •• • ---- ----~----M in -1 .0 • •

- 2.0 L...------ooor------.,.------ ...... -

1978 1979 FIGURE 14

RB 211 Fan leading edge restoration

Restoration of 25. 4 leading edge profile

Rad - __...tI+-"

Unacceptable Ma x 1.40 minimum blend of M in 1.27 leading edge wear 0. 70 Rad 0.64 FIGURE 15

RB211 I.P. Turbine shroud segment

FIGURE 16

RB211 Turbine internal cooling airflow

Stepped s eal FIGURE 17

RB211-22B Test bed takeoff

T.G.T. cusum plot

o o o o o C) C) o C) cu s um T . G .T . ~ + (datum 670 C) o o o ..

o + " ~ " " O'~ -'-.--'-.--r-.--'-.--r-'--r- ~-.~~ 10.00 20.00 JO.OO .0.00 50.00 60.00 70 .0 0 observat i on number FI GURE 18 RB211 H.P. System Foreign Object Damage control FIGURE 19 PERFORMANCE DETERIORATION - AN AIRLINE PERSPECTIVE Niels B. Andersen Pan American World Airways, Inc.

SUMMARY Due to the drastic energy cost increases which have been with us since the 1973 oil embargo, Pan Am, along with most of the world's airlines, has become painfully aware of the impact on costs and operations caused by the steadily deteriorating fuel efficiencies that characterize the aircraft we operate.

We estimate that the fuel efficiency of our original 747 fleet is down 6 to 6 1/2% from when it was new, of which about 1 1/2 to 2% is airframe deterioration and the balance of 4 to 5% is in the engines.

Although the NASA engine diagnostics program recommends periodic refurbishment as a technique for reducing deterioration, our own experience with this approach (which we adopted for improved reliability rather than performance restoration) has been disappointing.

Pan Am has consistently held that efforts for improvement of existing engines to achieve reduced fuel consumption should be in the direction of retaining the performance already in the engines rather than developing sophisticated design changes to reduce fuel burned.

Furthermore, we have always stressed the necessity for retrofitability in a practical and cost effective sense of any fuel savings feature .

Additional on-board engine instrumentation to allow component performance analysis using data from actual flight conditions is a most desirable feature for new transport aircraft designs. This will allow us to define for our shops which parts of the engine need attention to restore excessive performance losses. Effective measurement of one very important engine parameter continues to elude us, namely thrust. We need a thrust meter.

Main engine bearing configuration (number and location), cowl load sharing, inlet reactive loading along with practically any other design and/or installation feature that will stiffen the engine will, in our view, have significant payoffs in retaining engine performance efficiencies and reducing fuel consumption.

INTRODUCTION Like most of the world's airlines today, Pan American World Airways has become painfully aware of the impact on costs and operations caused by the steadily deteriorating fuel efficiencies that characterize the kinds of transport aircraft we all operate.

This awareness was spawned by the 1973 oil embargo, which precipitated the sharp, relentless energy cost escalations that have been with us since that time and seems destined to continue with no relief in sight.

Fuel prices have increased nine-fold since pre-embargo days. Pan Am's fuel costs for 1973 for a fleet of 142 aircraft (30 747 ' s and 112 narrow-body aircraft) was $170 million (1.2 billion gallons) for an average price of approximately 14¢/gallon . For 1981 Pan Am ' s fuel budget is $1.36 billion (1.1 billion gallons) for a fleet of 112 aircraft (64 wide-body and 57 narrow - body aircraft) for an average price of $1.24/gallon. Fuel costs have risen from 25% of direct operating cost in 1973 to nearly 50% currently.

After just over 11 years of operation, we find that our 747-100 aircraft have deteriorated to the point where fuel efficiency is down 6 to 6 1/2% from when they were new. Our fleet of 747SP airc r aft which entered service in the period 1976 to 1979 has deteriorated to a point where on the average fuel efficiency is down about 5 to 5 1/2% from when they were new .

Based on our own performance monitoring effort along with what has been learned from the NASA deterioration studies, we feel pretty confident that we can isolate about 1 1/2 to 2% of that deterioration to the airframe.

The remaining 4 to 5% performance loss is attributable to the engines - this loss appears in spite of a number performance improvement · modifications which we have incorporated in the JT9D over the years .

At current fuel prices, recovering or retaining just 1% fuel efficiency amounts to $7.6 million (6.9 million gallons of fuel) saved for the year just for our 747 fleet alone. Across the entire fleet, savings for a 1% improvement would exceed $10 million.

With this basic background information, it is not difficult to understand why we are so concerned about performance deterioration .

Pan Am's commitment to finding the causes of and cures for these punishing performance losses is reflected at least in part by its enthusiastic support for and extensive participation in the NASA Engine Component Improvement Program which commenced in 1977. We served as reviewers for both the Performance Improvement Program and for the Engine Diagnostics Program. In addition, Pan Am was under subcontract to Pratt & Whitney Aircraft to provide extensive historical engine performance data as well as making available certain JT9D engines in a program of special test cell, on - wing and in - flight tests to determine the mechanisms of deterioration in the JT9D- 7A engine .

While our experience and effort to date have focused on the JT9D engine, we are closely monitoring the performance of our newly-acqu i red LIOll-SOO aircraft powered by the Rolls Royce RB211 - S24B engines. Further - more, our efforts will now broaden to incorporate the CF6 - 6 and CF6 - S0 along with a sizable JT8D contingent.

As a result of our efforts and concerns, we have formed certain views and ideas about various aspects of engine performance deterioration and retention. These thoughts and ideas are set out hereunder.

NASA ENGINE DIAGNOSTICS AND PERFORMANCE IMPROVEMENT PROGRAM At the outset of the NASA Engine Diagnostics and Performance Improvement Programs we strongly urged, in our capacity as program reviewers, that emphasis be placed on finding ways to retain performance with particular stress on retrofitability. We have consistently maintained that from the standpoint of reducing overall fuel consumption, the potential payoff is greater if we are can retain performance that is already in the engines we operate rather than to develop sophisticated design changes to reduce fuel burned.

We have always stressed the importance of retrofitability of any modification, whether for performance improvement or for performance retention. Unless the now more than 3,000 CF6 and JT9D engines in service can playa part, it is doubtful in our minds at least that significant fuel savings will be realized for these model engines.

As it has turned out so far, very few of the concepts developed in the Performance Improvement Program are retrofitable in any practical sense.

ENGINE REFURBISHMENT One of the principal recommendations to come from the Engine Diagnostics Program - and one that was somewhat disappointing to us - was that operators should periodically refurbish the compressor section as well as the turbine section as an effective means of partially overcoming deterioration.

We have always known that new parts will improve engine performance.

However, this is a very costly way to gain performance, and as long as the basic design of the parts is unchanged the deterioration characteristics are fundamentally unchanged. At best this technique restores some performance for a limited period of time, but performance LOS retention is really not improved.

Quite coincidentally, in 1978 as the Eng ine Di agnos ti cs Program was well under way, Pan Am initiated a major cha nge in its engine maintenance philosophy, changing f rom the long-popular on-condition maintenance concept to a periodic re furbishme nt program of the kind recommended b y the engine diagnostics study for its JT9D engines.

This refurbishment program was adopted at Pan Am specifically to achieve improved engine reliabilit y , with reduced fuel consumption as an anticipated secondary benefit.

The new maintenance program has been quite successful from the standpoint of improving reliabilit y of the JT9D. In addition, TSFC of completely refurbished engines are on the average 1 to 1 1/2% lower than all other engines when measured in the test ce ll after repair.

However, over the 3 year period the program has been in effect, we have been unable to see any ieal impro ve ment in fuel consumption attributable to refurbishment based on our routine aircraft and en g ine performance monitoring procedures. The most we might be able to sa y is that further deterioration may have been somewhat arrested. This has been an unexpected and disappointing result, for which we have no good explanation at this time.

ENGINE INSTRUMENTATION The day is approaching, at least at Pan Am, where engines with high fuel consumption may occasionally be removed for that reason.

Hitherto unscheduled removals have largely been associated with high EGT, mechanical failure or boroscope inspection revealing incipient failure.

As we approach an economic environment where high fuel consumption becomes a cause for engine removal, it is becoming increasingl y apparent to us that the current variety of on-board engine instrumentation, which has changed little during the some twenty year that jet transport aircraft have been operating is inadequate.

We believe that new generations of transport aircraft should incorporate expanded on-board engine instrumentation to allow comprehensive engine component analysis using data from actual flight conditions rather than having to rely on sea level, static test cell data. When an engine is removed we must be able to specify to the shops with confidence which parts of the engine require attention to recover valuable fuel efficiency.

We believe this is feasible.

Specifically, additional instrumentation should probably include at least pressure and temperature between engine stages. Where variable vanes are featured, vane angle should also be displayed in the cockpit .

Such add it ional engine parameters probably need not be displayed continuously. One approach would be to have one set of gages installed on the engineer's panel with a selector switch to display one engin e at a time. With the advent of sophisticated performance management and flight management systems, there should be all sorts of possibilit ies for automatic recording of data on command from the flight engi neer.

A discussion of engine instrumentation would be incomplete without mentioning thrust meters . Such an instrument has been the dream of people lik e us for many years. Thrust remains one of the two o r three most import ant performance parameters for jet aircraft, yet its accurate and reliable measurement in flight continues to elude us.

Until a good thrust meter is developed, we feel that improvements ca n and must be made in the two most popular thrust-setting parameters: engine pressure ratio, as on the JT9D, and low-spool RPM, as on the CF6.

As a reli able, accurate measure of actual thrust, especially under cruise conditions, we believe both systems have some serious fla ws. In b o th systems there are what appear to us to be unexplained shifts in their relationships to net thrust so that we are not necessarily getting the thrust we think we are getting when we set EPR or Nl. At this point common sense tells us that the integrated engine pressure ratio system used on the RB2ll engine is probably superior to either of the other two systems. However, since the RB2ll is quite new to us, we will have to withhold judgment for a while.

ENGINE DESIGN AND INSTALLATION FEATURES Certain features of engine design and installation are clearly demonstrat ing important advantages in engine performance retention.

Bearing arrangement no doubt has an important role in performance retention. Four bearings seem to be insufficient while six are probably more than are required. A well-designed 5-bearing system would seem to b e an optimum configuration.

Bearin gs with over-hung components such as fans, should be designed to minimiz e such over -h ang to limit associated wobble, which in turn leads to shroud rub, or to allow closer running clearances.

We are convinced that almost any effort to improve stiffness and ge ner ally reduce flexing of the engine structure will payoff significantly in perform ance retention - even at a weight penalty. For this reason we favor cowl load sharing to provide additional rigidity at a relatively small cost in additional weight and complexity. Studies by the manufacturers are presentl y under way in this area for the JT9D -7 installation on the 747s. The approach under development is particularly attractive in that 1 07 it looks very promlslng and cost effective for retrofit. Pan Am has indicated a strong interest in this program and we have offered to participate in any meaningful way, such as perhaps a service test program.

Along these same lines, Pan Am is planning to participate with the Boeing company this year in a service test program of a device designed to react against flight loads on the engine inlet of the 747, thereby reducing fan and low compressor shroud rub. This too is very attractive to uS because of the retrofit potential, which is indeed what will be done for the service test program.

These three areas, bearing location and number, cowl load sharing, and inlet reactive loading are, in our view, key areas in the battle to retain engine performance efficiencies - particularly since the performance which is ordinarily lost when an engine flexes is lost during the first flight or two and has been largely considered unrecoverable.

CONCLUSIONS The foregoing points up Pan Am's great concern about maintaining the fuel efficiencies of its fleet of aircraft and engines. We have actively supported past programs to determine causes of and cures for engine performance deterioration and will continue to pursue efforts to apply this valuable knowledge effectively to current as well as future engine designs.

There are over 3,000 JT9D and CF6 engines in service at this time with the number growing slowly but steadily. These engines can he expected to remain in service for a good many years to come. The challenge therefore remains to develop practical, retrofitable performance retention features that can save significant quantities of fuel on this verv large body of engines in the 1980's and no doubt the 1990's.

IMPROVING TURBINE ENGINE COMPRESSOR PERFORMANCE RETENTION THROUGH AIRFOIL COATINGS* L.A. Friedrich Pratt & Whitney Aircraft Group Introduction: Alteration of compressor airfoils by the erosive action of engine ingested particulate matter is a cause of performance deterioration in commercial aircraft turbine engines. A NASA sponsored JT9D Engine Diagnostics pro- gram quantified the problem for the commercial aircraft engine fleet indicating that the performance deterioration of the compressor - and erosion of the compressor airfoils - was related to total engine cycles rather than total engine operating hours. Thus the erosion problem becomes more severe when considering short mission applications wBere ~ tbe number of engine operating cycles builds rapidly in relation to total engine operating hours. The appearance of a set of high compressor air- foils operated for approximately 10,000 cycles is shown in Figure 1.

Erosion of turbine engine compressor components has been a serious problem for military helicopter operations. In this application the erosion pro- blem is so severe that factors of ten improvement in erosion resistance are required for any material or coating developed to alleviate the erosion problem. The titanium carbide and titanium diboride coatings that offer this degree of protection also compromise blade fatigue strength to a level not tolerable in commercial turbine engine applications. However, since the erosion problem in commercial engine service is considerably less severe than in military helicopter operations, coating solutions are available that may provide adequate erosion resistance without critically compromising the fatigue strength margin of the airfoils.

In order to evaluate the potential effectiveness of coatings in limiting erosive damage to compressor airfoils, an effort was initiated to evaluate candidate coatings for substrate alloys typically used in commercial engine high compressor blades. Laboratory and rig erosion testing of plasma deposited and diffusion coatings described in this paper has shown the potential of a two-to four-fold improvement in erosion life. The selective application of these coatings to approximately the outer third of the airfoil - the area that is subject to erosion degradation - avoids coating the fatigue critical region of the blade, thus providing erosion resistance potentially without compromising the fatigue strength of the blade. Both the plasma and the diffusion coatings also offer the advantage of low initial cost and a multi-source production base.

*The reported work has been performed under NASA Lewis Research Center Contract titled Materials for Advanced Turbine Engine (MATE)(Contract NAS3-20072) P&WA Project 4, Erosion Resistant Compressor Airfoil Coatings.

Coating Selection : A useful first order classification system for potential erosion resistant coatings identifies three major types of coatings. Specific coatings selected from each class for this study include : 1. Multiphase Overlay Coatings: tungsten carbide - cobalt 2 . Diffusion Coatings: chromium-boron 3. Single Phase Hard Compound Overlay Coating: titanium-diboride The tungsten carbide - cobalt composition is applied by modern plasma spray- ing. This type of coating has been widely used in the aircraft engine industry principally to minimize contact wear involving galling, fretting and impact. High energy thermal spray processes, the most important of which are plasma spray and detonation gun, have been developed for the application of high integrity coatings. These processes are highly com- mercialized and supplier facilities capable of producing these coatings exist world wide.

Representing the diffusion coating class is a chromium - boron composition .

This type of coating is formed by diffusional interaction of chemical elements with substrate alloys to form erosion resistant phases at the alloy surface. An intensive commercialized technology base exists for the fabrication of diffusion coatings for the turbine engine industry.

Single phase, hard compound overlay coatings such as TiB2, TiCN and TiC have been demonstrated to provide the greatest degree of erosion resis - tance, particularly at low particle impingement angles. The two most widely investigated processes for fabrication of these coatings are chemical vapor deposition and fused salt electrolysis . TiB2 produced by fused salt elec t rolysis is representative of this type of coating.

Coating Evaluation : Laboratory Erosion Testing Coatings were produced on three alloys representing typical materials used in commercial turbine engine compressor airfoils. These alloys are the titanium base alloy Ti-6Al - 4V (AMS 4928), a stainless steel alloy (AMS 5616) and a nickel base alloy (IN90l). In this paper the laboratory erosion test results are reported for the coatings on stainless steel (AMS 5616) . The alloy specimens were coated to a nominal thickness of 50 microns (2 mils).

The laboratory erosion testing was performed using an S.S. White Airbrasive Unit . Aluminum oxide with a nominal 27 micron particle size was used as the abrasive material . The abrasive particles are accelerated to approximately 300m/sec and impinge on the test specimen approximately 1. 5c m from the nozzle . Three abrasive impingement angles were tested - 20, 45 and 90 degrees. The erosion resistance was measured by weight and volume change as a function of time, and by the time to erode 25 microns (1 mil) of coating .

Erosion data for the three types of coatings on AMS 5616 at the 20° impingement angle shows considerable improvement in terms of volume loss compared to the uncoated stainless steel alloy (Figure - 2).

Erosion at this angle is typical of airfoil trailing surfaces.

A comparison of test results at all three abrasive impingement angles is presented as time to erode 25 microns of material. The coatings are particularly effective at the low impingement angles (Figure 3).

These data are in general agreement with the literature, with the hard coatings demonstrating greater resistance to erosion at low impinge- ment angles than the baseline uncoated alloys. At the test condition used, the hard compound TiB2 coating demonstrated improved resistance at a 90° abrasive impingement angle, which is not typical of this type of material in field service engine testing. These laboratory erosion tests are valuable tests to quickly and inexpensively rank coating compositional and processing variations. However, they are inadequate to provide an assessment of the potential life improvement coatings can provide on compressor airfoils.

Rig Erosion Testing To address the challenge of establishing a test procedure that would simulate relative compressor airfoil life when subject to erosive conditions, a facility was constructed to erosion test actual com- pressor airfoils. A combustor system was modified to include a particle injection system (Figures 4,5). A holder was designed to place the test airfoil at controlled downstream locations with the airfoil positioned at controlled angles to the particle stream.

Airfoil temperatures are monitored using an optical pyrometer.

Typically, nominal twenty micron aluminum oxide is used as the erosive agent. The Laser Doppler velocimetry technique was used to determine particle velocity and particle flux in planes at a number of locations from the combustion exit nozzle. These measurements were made as a function of test rig control variables: fuel pressure, air pressure, and particle feedrate. Thus the test rig was cali- brated to produce known particle velocities and test airfoil temperatures by varying the rig controls and the airfoil distance from the exit nozzle, providing the capability of simulating the temperature and velocity conditions at each stage of high compressor in gas turbine engines.

To determine the ability of this rig to reproduce erosion patterns seen in field service operated hardware, a group of blades were rig tested. Visual appearance of field service and rig tested blades was similar (Figure 6). Profiles taken at standard planes indicated similar erosion patterns with both types of testing resulting in significant reduction in blade leading and trailing edges as well as thinning of the concave airfoil.

- ---- - -- -~-- In addition to duplicating the erosion pattern seen on field service operated compressor blades, this rig test has been able to demonstrate the blade leading edge chipping phenomenon seen in field service with titanium diboride coated blades (Figure 7). The blade leading edge blunting is an important effect to determine in screening candidate coatings as the blunt leading edge results in unacceptable aerodynamic penalties and would preclude the use of erosion resistant coatings exhibiting this effect.

Initial rig testing of AMS 5616 compressor blades with approximately 30 micron (1 mil) thick coating of plasma applied tungsten carbide- cobalt and diffusion coated chromium-boron exhibited a three fold improvement in erosion resistance measured by volume loss compared to the uncoated blades (Figure 8). The test conditions used in these tests were a blade temperature of 390°C (730°F) and a particle velocity of 290 m/sec (950 ft/sec). In these tests neither the plasma applied tungsten carbide-cobalt coating nor the diffusion chromium-boron coating eroded in a manner to produce the aerodynami- cally unacceptable blunted leading edge appearance seen with the titanium diboride coated blades (Figure 9).

These initial results indicate that the plasma applied coatings and the diffusion coatings offer the potential of limiting the erosive damage to high compressor airfoils.

Continuing Activity: In the next phase of this erosion resistant coating development activity rig erosion resistance data will be generated for selected coatings on a number of airfoil stages chosen to be representative of all stages - of modern turbine engine high compressors. In addition fatigue testing of coated blades has been initiated as well as surface treatments to produce blade surface finishes on the order of 20 micron AA.

FIGURE 1 COMMERCIAL ENGINE COMPRESSOR AIRFOilS AFTER SERVICE OPERATION 200 abrasive impingement angle TiB2 90 0 o TiB2 o Diffusion applied Cr-B Time 50 <) Plasma applied (sec) 40 WC/Co o Uncoated AMS 5616 o 0 . 0001 0 . 0003 0.0005 0 . 0007 0 . 0009 0.0011 Volume loss (eel FIGURE 2 LABORATORY EROSION TEST RESULTS ON COATED AMS 5616 STAINLESS STEEL ------------- -- -- -- -- ---- AMS 5616 substrate >500

t t t

50t

rrn

Time to

100 t-

erode r- r- 25 microns 80 r- r- 60 r- ~ o r- f0- 40 r- r-- r- 45 °

rn a

rn

o WC / Co Cr-B Uncoated FIGURE 3 EROSION AS A FUNCTION OF ABRASIVE IMPINGEMENT ANGLE Primary Exit Transition duct

se~ner

d duct I

Secon.dary section

Instrument ~

collar '" l /

'"

I I Liner

~Fuel 1------ ==:'-=:::-:-=-= . n

r nozzle

\ ,

~ y

-~-- Primary Particle Secondary air air injection FIGURE 4 SCHEMATIC OF AIRFOIL EROSION FACILITY FIGURE 5 FACILITY TO TEST EROSION RESISTANCE OF COMPRESSOR AIRFOILS Engine service eroded Rig eroded FIGURE 6 COMPONENT RIG TEST SIMULATES ENGINE SERVICE EROSION Eng i ne serv ice eroded Rig eroded FIGURE 7 COMPONENT RIG TEST SIMULATES ENGINE SERVICE LEADING EDGE CHIPPING EROSION OF TITANIUM DIBORIDE COATED AIRFOILS o TiB2 o TiB2 o Plasma WC /Co 6 Cr-B di ffusion o Uncoa t ed o Uncoated 0.01 Volume loss 0.02 Icc) 0.03

0 . 04~ __ ~ ___ + __ ~

0.04L-- __ --' ___ -:- ___ -: o o FIGURE 8 EROSION TEST RESULTS OF PLASMA AND D IFFUSION COATINGS ON AMS 5616 ......

......

-....J WC/Co Cr-B TiB2

FIGURE 9 PLASMA CARBIDE/METAL AND DIFFUSION Cr-B COATINGS

WITHSTAND LEADING EDGE CHIPPING EROSION

Page intentionally left blank

ADVANCED OXIDE DISPERSION STRENGTHENED SHEET ALLOYS FOR IMPROVED COMBUSTOR DURABILITY* R. J. Henricks Pratt & Whitney Aircraft Group Introduc tion Burner durability has become a serious problem in many current generation aircraft gas turbine engines. Advances in structural metal temperature capability and in burner hardware cooling technology have not kept pace with demands for more efficient (higher gas temperature) engine performance.

Hastelloy X burners designed for around 87loC (1600°F) metal temperature opera tion are experiencing hotter streak cond itions with heavy penal ties to opera ting life. Both improved burner materials and designs are required to provide the large durability increase essential to future aircraft turbine engine operation and maintenance. A decrease in engine maintenance costs can result both from increased burner life and from reduced turbine section damage caused by burner distortion.

The substitution of advanced oxide dispersion strengthened (ODS) alloy sheet materials with improved creep strength and oxidation resistance compared to Hastelloy X can produce a significant increase in burner durability. Pro- perties of two advanced ODS alloys, Incoloy MA 956 and Haynes Developmental Alloy 8077, compared to Hastelloy X indicate that they exhibit a 167°C (300°F) advantage in creep strength and in cyclic ox idation resistance (Figure 1). However, these ODS materials exhibit low cycle fatigue pro- perties that show no improvement over Hastelloy X.

It is the objective of a NASA/P&WA MATE (Materials for Advanced Turbine Engines) program to evaluate burner design modifications that will take ad- vantage of the improved creep and cyclic oxidation resistance of ODS alloys while accomodating the reduced fatigue properties of these mater ials. This program will culminate in a JT9D experimental engine test of the selected combustor design and ODS alloy. A status report of this MATE program is the subject of this paper.

Burner Em irornnent The principal failure modes for louvered combustor liners are creep bucluing of the louver lips, oxidation and low cycle fatigue cracking, examples of which are shown in Figure 2. Creep buckling and oxidation are the dominant failure modes in long missions; low cycle fatigue is the dominant fai _ lure mode in short missions.

* Work performed under NASA~TE Contract NAS3-20072 with NASA-Lewis

Research Center.

Creep buckling of the louver lip results from the strain imposed by the con- straint between the thermal growth of the hot louver lip and the relatively cooler weld (knuckle) area over an extended period of time (Figure 3) . The louver lip develops a high stress that yields the material in compression.

Continued cycling produces circumferential distortions which are sufficiently large to close off the louver gap in local areas (Figure 3). These local closures reduce the cooling air flow for the downstream louver, increasing the local tempera tur e, and accelera ting the buckling process on the downstream louver lip. Eventually the severity of the process reaches such a magnitude that rapid ox idation produces a burn-through of the cooled liner. When this point is reached, repairs are necessary.

Low cycle fatigue failures of conventional louver liners are the result of high radial temperature gradients through the liner producing excessive thermal strains. The hot side of the liner is subjected to average tempera- tures of about 871°C (1600°F) in the area of the welds. The cold side of the liner, in the area of the cooling holes, is subjected to temperatures of abo ut 900°F. The severe thermal strain resulting from this radial temperature gradient is aggrava ted by unavoidable circumf erential variations in hot side temperatures of at least 56°C (100°F).

Burner Design and Structural Analysis for ODS Alloys Structural analysis of combustor liner cooling geometries was conducted to determine the stress and strain distribution produced by thermal and mechan i- cal loads acting on the liner during an engine flight cycle. Using a finite element analysis, the elastic stress state can be defined with elastic modul u s and the coeff icient of thermal expansion as a func tion of temperature. While the thermal analysis and elastic stress are aff ec ted by the physical properties of the materials, burner life prediction is dependent upon mechanical pro- per ties.

The design system (Figure 4) incorporates the technique of "exhaustion of ductility" for calculating life predictions (ref. 1). The interaction of creep and fatigue modes in a cumulative damage model becomes the failure criterion determined for the design system. For ODS sheet alloys, the ductility determined from tensile data and implied from LCF data is about 5% , whereas the creep ductility measured in creep testing can be 0.1 - 0.2%. An available ductility of 0.1 % for ODS alloys was assumed throughout the design phase. The structural analysis and life predictions are based on the engine operating conditions applicable to an advanced energy efficient engine .

In the design phase of this program, the operating strains for the ODS alloys were minimized in a series of five candidate advanced combustor designs . This was accomplished by designing a series of segments in the circumferential direction to significantly reduce the hoop (circumferential) strains and by eliminating fixity between the hot wall and cold wall shell to reduce radial constraint (strain). Based on the thermal and structural analysis for these five combustor designs the predicted lives were calculated for both ODS sheet alloys.

In addition to the predicted lives for these combustor designs, other factors for assessing the relative benefits of the designs and for selecting the final two were considered; these factors included liner fabricability and engine maintenance and operating costs. The maintenance cost (MC) is based on predicted lifetimes, on initial construction cost and on the type of repair procedure employed for the design. The direct operating cost (DOC) is derived by using the initial fabrication cost, the overall weight and the maintenance cost. Equally important in the determination of the particular designs worth pursuing is the consideration of risk for the construction and repairability of a given design. Based on significant life improvement, lower maintenance and direct operating costs and estimated moderate risk factors for fabrica- tion and repairability, two designs were selected for continued evaluation: 1) a mechanically attached, film cooled segmented louver and 2) a mechanically attached, transpiration cooled segmented twin wall. Schematic diagrams of these two designs are presented in Figures 5 and 6. These two designs were assessed relative to a current commercial engine JT9D-59/70 using the same design selection criteria. Comparison of these designs using ODS alloys to the film cooled JT9D combustor of Hastelloy X in Table I shows a four times improvement in life reflecting the high temperature strength of ODS alloys and the reduced strain range present in the two segmented designs. The lower cooling air levels in the transpiration cooled, twin wall reflects the in- creased effectiveness in that design. While the initial construction cost and the combustor weight are somewhat higher than for the standard combustor, the maintenance cost and direct operating costs of the advanced designs are significantly lower. It must be pointed out that the relative changes in DOC represent decreases in overall engine operating cost and are not limited solely to combustor cost.

Mechanically Attached, Film Cooled Segmented Louver A mechanically attached, segmented louver using current film cooling techni- ques (Figure 5) is attractive and can accomodate the low strain capability of ODS alloys. By mechanically attaching with rivets each ODS segment to a Hastelloy X shell (cold wall) rigidly at only one location and providing room at the other rivet and bushing locations for differential thermal expansion of the hot segments relative to the cold shell, circumferential and radial constraint does not occur under these conditions. The only thermal strains present are those generated from the non-linear temperature variation within each segment. The failure mode in this design is established as initiation of a 0.79mm (1/32") crack at the louver lip.

Calculating the combustor liner life based on the number of cycles of exhaust 0.1% strain as the criterion, the predicted life for MA 956 is in excess of 10,000 cycles, while the predicted life for HDA 8077 is 2000 cycles (Table 1:0.

Mechanically Attached, Transpiration Cooled, Segmented Twin Wall The "twin wall" transpiration cooled design is an advanced cooling technique with the capability of significantly reducing metal temperature and/or cooling flow and thermal strain. A schematic cutaway view of this trans- piration cooled, segmented twin wall combustor is shown in Figure 6. The transpiration cooled panel is attached to the impingement plate with a series of studs so that leakage around the edge of the panel at the operating tem- perature of 1010°C (1850°F) is less than 10 % of the panel cooling air. How- ever, to reduce the strain at maximum temperature, the panel is pre-stressed at room temperature to a predetermined contour duplicating the shape it will assume at the operating temperature of 1010°C (1850°F). An impingement plate which serves as a mandrel employs a contoured edge radius and a centeral posi- tioning stud to impart the desired deflected panel shape.

Analysis shows that during simulated engine operation as the average tempera- ture and the through-thickness gradient increase, the mechanically induced pre-stresses are reduced and become essentially nonexistent at the operating temperature. The only stress at operating temperature is a small cooling air pressure load. The nature of this design is such that the largest stress occurs at 20°C (68°F) where the material strength is the highest. The loca- tion of the high stresses is in the center of the edges of the panel at room temperature; during heat-up although some stress redistribution occurs, the maximum stress remains at the center of each edge. Throughout the thermal loading, the maximum stress level remains below the proportional limit so that no plastic damage occurs.

The current method for predicting the effect of hole arrays in a transpiration cooled geometry on the thermal-mechanical fatigue life employs the concept of linear elastic, isotropic fracture mechanics (ref. 2). For transpiration cooled designs, failure is defined as linkup of cracks emanating from adjacent cooling holes. As a result of prestress, operating stress, and thermal c ~ cle, the predicted service lives for the ODS alloys are in excess of 10,000 cycles (Table II). Since there is no thermal interaction between the panel and the studs due to prestressing, at the operating temperatures the cooled studs support only a small radial pressure load.

Alloy Evaluation Mechanical property testing of the two candidate advanced ODS alloys, Incoloy MA 956 and HDA 8077, was directed towards the selection of one alloy for evaluation in the remainder of the program. The three main criteria for this alloy selection were creep, oxidation and thermal fatigue resistance. Three separate measures of thermal fatigue resistance were defined: creep ductil- ity, isothermal LCF life, and hot spot blister (thermal cycling) cracking.

Incoloy MA 956 alloy sheet of 1.1-1.3mm (0.043-0.051") thickness was supplied by Wiggin, Ltd. of the International Nickel Company and HDA 8077 alloy sheet of 1.1-1.4mm (0.043-0.0551~ thickness supplied by Cabot Corporation. The nominal chemistry of each alloy is shown in Table III; both alloys are strengthened by a fine dispersion of yttrium oxide (Y203)' The sheet materllUs were produced by mechanical alloying of powder, powder consolidation and a series of hot and/or cold rolling procedures. These processing techniques result in coarse "pancake" grains in the plane of the sheet and elongated grains through the thickness in both alloys (Figure 7) with fine yttria particles dispersed throughout the structure.

Creep evaluation of MA 956 and HDA 8077 ODS alloys and Haste110y X, bil1-of- material in JT9D combustor liners, in the 871-1093°C (1600-2000°F) range demon- strates significant creep strength advantage for both ODS alloys over Haste110y X qt the higher temperatures (Figure 8); approximately 167°C (300°F) for MA 956 compared to Hastel10y X. These Larson-Miller curves represent an average of time to 0.1 % creep strain data generated on these sheet alloys within the present program. Comparison of the ODS alloys shows that HDA 8077 sheet p0ssesses time to 0.1 % creep strain and final creep ductility superior to those of MA 956 sheet. Figure 8 compares the high creep ductility of Haste1loy X to the limited ductility of the ODS alloys. The final creep ductility is defined as the last creep extension measurement within two hours of specimen failure. The MA 956 sheet, which is not cross rolled, is aniso- tropic as exhibitedby the difference in creep ductility between the longi- tudinal (parallel to rolling direction) and transverse (perpendicular to rolling direction) orientations. The average transverse creep ductility of MA 956 is lower than the longitudinal ductility, although the minimum values are similar. There is no difference in time to 0.1 % creep strain (creep strength) for these orientations; however, the creep-rupture life for the transverse orientation is lower reflecting the decreased ductility.

Cyclic oxidation testing was conducted at 982°e (1800°F) using a six minute cycle with cooling to 316°C (600°F) using a four minute hold at maximum tem- perature. Specimens of the the three alloys were tested in a rotating fixture subjected to a JP4-R fuel gas flame for heating and forced air for cooling.

Surface attack was determined metallographically on tested specimens. The results of this oxidation testing (Figure 9) show that there was insignificant surface attack <.013mm (0.0005") in MA 956 after 1000 hours and that it is superior to HDA 8077 which had .025-0.05Imm (.001-.002") of surface oxidation.

Both ODS sheet alloys possess excellent oxidation resistance compared to Hastelloy X, which exhibited 0.36mm (.014") of surface oxidation and spalla- tion after 1000 hours. The relative oxidation resistance of MA 956 and Haste110y X were verified in a duplicate 1000 hour oxidation test using different heats of material.

Isothermal low cycle fatigue (LCF) testing utilizing strip specimens in a fully reversed bending mode was conducted for the three alloys at 760°C (1400°F) and 871°C (1600°F) with a + 0.25 % strain range (Figure 10). Specimen fatigue life is defined as failure into two pieces. At 760°C (1400°F), the Hastelloy X and HDA 8077 showed similar average failure lives, although the latter exhibited a high degree of data scatter, and both were somewhat higher than the MA 956 life. At 871 e (1600°F), Hastelloy X LCF life was lower than MA 956; however, MA 956 demonstrated approximately a five-fold higher life than HDA 8077 sheet. Limited testing at 982°C (1800°F) of the ODS alloys showed this same 5:1 life advantage of MA 956 over HDA 8077.

A Tlhot spot blister test" was designed to produce localized thermal fatigue cracking and/or deflection similar to that produced by flame impingement on a burner louver in an engine. Seventy-six millimeter (3.0 inch) diameter disks were subjected to a thermal cycle (5 cycles/min.) from a Tmin of 538°C (lOOO°F) to a Tmax of 982°e (1800°F) by use of an alternating oxy-acetylene flame and a cool air blast (Figure 11). The Tmin temperature of 538°C (lOOO°F) was main- tained on the edge of the disk using a propane gas burner. Strain analysis of this hot spot blister test shows that compressive strain peaks at maximum tem- perature and tensile strain at minimum temperature; this type of strain-temper- ature cycle is typical for current engine combustors. At pre-determined cycle intervals the degree of surface cracking was recorded visually and specimen deflection height was measured. Actual crack depths were determined metallo- graphically on discontinued test samples. As strain range on the cold side of the "hot spot blister" specimen is calculated to be much higher than that of the hot side, initial cracking was observed on the cold side of the sheet.

Metallographic determination of cold side cracking shows that MA 956 exper- iences substantially earlier crack initiation and more severe cracking than HDA 8077 which, in turn, displays more cold side cracking than Hastelloy X as shown in Figure 12. Hastelloy X deflects substantially more than the two ODS alloys, while MA 956 and HDA 8077 exhibit similar deflection heights. The specimen deflections reflect the relative creep strengths of each of the three alloys. Additional testing to a Tmax of 1093°C (2000°F) showed an identical ranking of the alloys in cracking and deflection.

Alloy Selection Both ODS alloys demo nstrated the 167°C (300°F) advantage in creep and oxida- tion resistance over Hastelloy X. Comparing the ODS alloys, MA 956 is superior in oxidation resistance and isothermal LCF life and HDA 8077 is slightly better in creep strength and ductility and is superior in hot spot blister cracking resistance. Under the MATE program, concurrent to the materials evaluation phase , the two alloy manufacturers were engaged in a sheet reproducibility program. Wiggin, Ltd. of INCa successfully demon- strated MA 956 sheet product reproducibility for both sheet quality and mechanical properties in a second heat of material; Cabot Corporation was unable to reproduce the intial excellent formability and creep properties of HDA 8077 in subsequent shee t product. While neither ODS alloy exhibited an obvious overall superiority in properties, on the basis of product repro- ducibility and excellent life predictions in both combustor designs, Incoloy MA 956 alloy was selected for evaluation throughout the balance of the MATE program, including component and experimental engine testing.

Low Cycle Fatigue Structural Tests LCF rig testing of components of the two candidate designs using MA 956 and Hastelloy X alloys is directed toward selection of one design for experi- mental test in a JT9D engine. Such component rig evaluation allows for testing of the structure of each design under simulated combustor conditions.

Feasibility studies were conducted to define the best approaches for LCF structural assessment of the two ODS combustor designs.

For the mechanically attached, film cooled, segmented louver design (riveted louver) a single louver segment of MA 956 or Hastelloy X attached to the in- side of a Hastelloy X shell in conjunction with a double return pie-wound induction coil and external cooling air successfully simulated the temperature profile along the louver as shown in Figure 13. The test consisted of a 45 second heating cycle to the desired temperature profile with a maximum tempera- ture of 1010°C (1850°F) at the lip, a two minute hold at this condition and a 30 second cooling cycle to a louver lip temperature of 538°C (lOOO°F).

A total of six (three each) of MA 956 and Hastelloy X riveted louver segments were tested in this induction heated rig; the results are reported in Table IV and Figure 14. MA 956 segments exhibited considerably more dimensional stability (less distortion) than did the Hastelloy X segments. While there was a signLficant degree of test scatter in both materials, a comparison on the basis of cycles per millimeter points up the greater resistance to bucklfug for the MA 956 alloy. Typical bow of the panels removed from the Hastelloy X shell are shown in the photograph in Figure 14.

No crack indtcations were evident by Post Emulsion Fluorescent Penetrant inspec tion on any of the test segments. The Hastelloy X tests were discon- tinued when the bow in the segments became ex cessive and preluded main- taining the axial temperature profile on the bow or distortion; specifically after 66, 242 and 1,500 cycles. The MA 956 tests were discontinued after 4000, 5000 and 6000 cycles. These test results demonstrate the excellent creep resistance of MA 956 compared to Hastelloy X in this component test and the low strain ranges achieved as evident by no cracking in the seg- mented louvers.

A second component test of the riveted louver design was defined and conducted in a thermal cycle rig (Figure 15). The axial temperature profile established in the louver at transient conditions approximated the steady state profile of the induction heated rig test. Rotating gas burners impinge on the lip of the segmented louvers during heat-up and an air manifold directs cooling air onto the louver lips during the cool-down for a total cycle time of 60 seconds.

The lip is cycled between 954°C (1750°F) and 593°C (llOO°F), while the Hastelloy X shell is cycled between 538°C (lOOO°F) and 399°C (750°F). In this test, alternating segments of MA 956 and Hastelloy X were installed around the ID circumference of the Hastelloy X shell. To date, the first test com- ponent has achieved 4000 cycles with no distress evident in the attachment rivets or bushings of any of the louver segments. Additional thermal cycle testing of this component and of a duplicate is scheduled.

For LCF structural testing of the transpiration cooled, twin wall design the te~t rig is srown in Figure 15. A statiQnary gas burner heats the hot side of the pre-stressed panel/impingement plate assembly which rests on a box providing a plenum of cooling air. The hot side of the panel reaches a maximum tempera- ture of 927°C (1700°F) and an average through-thickness gradient of 22Co (40FO) and is lowered from the flame, applying increased cooling air flow until it cools to 649°C (1200°F) at which time it is raised back into the gas flame and the airflow is reduced. The entire test cycle is 30 seconds in length (22 sec. heating and 8 sec. cooling). An initial MA 956 panel tested for 2000 cycles contained laser drilled cooling holes; all subsequent panels were electrochemical machined (ECM) because of improved cooling hole integrity in the MA 956 panels. The film from the transpiration holes is an effective cooling mechanism and useful in obtaining maximum life of the segment . Sin c e the holes are ECM drilled at an acute angle, one edge of the panel in the r ig is void of film . (In actual engine use, a layer of film would be establish ed to provide insulation until the transpiration film became established.) Th is region of low film in the rig tested panel exhibited numerous hot and cold side cracks between the edge and the middle rivet (Figure 16). The cracks extended from acute corners of the cooling corners of the cooling holes in this high strain region of the panel (edge center); several cracks linked up to form a larger cracks . The laser holes contained pre-existing . 08 - .1Omm (.003-.004") cracks resulting in earlier crack growth than would be expecte d with ECM holes. MA 956 and Hastelloy X pre-stressed panels have each achieve d 10,000 cycles without any evidence of crack initiation at the ECM holes .

Summary A NASA-sponsored MATE project for ODS alloy combustor liners is in progress; a summary of the program to date follows: 1 . Five advanced combustor designs were evaluated based on preliminary anal - ysis and life predictions, on construction and repair feasibility and on maintenance and direct operating costs. Two designs - the film cooled , segmented louver and the transpiration cooled, segmented twin wall - were selected for LCF component testing.

2. Detailed thermal and structural analysis of these designs established th e strain range and temperature at critical locations resulting in predicte d lives of 10,000 cycles for MA 956 alloy.

3. ODS alloys, MA 956 and HDA 8077, creep strength and oxidation resistance demonstrated a 167°C (300°F) temperature advantage over Hastelloy X alloy .

MA 956 alloy was selected for mechanical property and component test evalua tions • 4. MA 956 was superior to Hastelloy X in LCF component testing of the film cooled, segmented louver design.

5. Thermal cycle testing of the riveted louver design and LCF structural testing of the twin wall design are in progress.

Ref erences 1. Polhemus, J.F., Spaeth, C.E .. , and Vogel, W.H., "Ductility Exhaustion M od el for Prediction of Thermal Fatigue and Creep Interaction", Fatigue at Ele - vated Temperatures, ASTM STP 20, American Society for Testing Materials, 1973, Pages 625 - 636.

2. Gemma, A.E. and Phillips, J.S., "The Application of Fracture Mechanics t o Life Predictions of Cooling Hole Configurations in Thermal-Mechanical Fatigue", Engineering Fracture Mechanics, 1977, Vol. 9, Pergammon Press, Great Britain.

_ .- _. -- - -------- -

LIFE/COST COMPARISON

OF DESIGNS

(MA 956) Total Cooling st rai n Life Co s t Weight MC DOC air % W range (%1 c y c les/ hr s S K Ibs S/ hr % AB 45 - 0.40 1.0 1.0 1.0 1.0 Base JT9D base Film cooled, 45 0.145 4X 1.26x 1.06x 0.63x - 0. 21 segmented louver Segmented 33 0.225 4x 1.48x 1.03x 0.65x - 0. 21 twin wall Table I

ODS ALLOY COMPARISON

MA 956 HDA 8077 Total Total strain Life strain Life Design range (%) (cycles) range (%) ( cycles) --- Film cooled, segmented 0.145 > 10,000 0.185 2,000 louver Transpiration cooled, 0.225 > 10,000 0.245 > 10,000 segmented twin wall • Strain range differences are associated with thermal expansion character i stics J2.fo82711 ,,300< Table II

CANDIDATE ODS ALLOYS

Incoloy MA 956 Sal 20.0 4.5 0.5 0.5 HDA 8077 Sal 16.0 4.0 0.8 Ta b le III

PROPERTIES OF CANDIDATE

ODS ALLOYS

Isothermal low cycle Creep strength Oxidation r esistance fatigue resistance 167 (300) r-- 167 (300) ;-- 1. 0 Temp Relative advant 0 - o - life 0.5 ° C (OF) o L-J.------'. ----'---'-- _ Hast Adv Has t Adv Hast Adv X ODS X ODS X ODS alloys alloys alloys J2<&827 ·2 <1 .,300< Fig u re 1 12 8

COMBUSTOR FAILURE MODES

. . . . . . . . . . ~

Creep buckling • Oxidation lCF cracking Figure 2

CONVENTIONAL FILM

COOLED LOUVER COMBUSTOR

Gas flow W eld Louver gap

====:::::::::: ~~o~~~ -Z..:;r~-- Cool i ng holes

Knu c kle/ ~_~~_ -- ~---~=================v J2 4827 "

.. """

Figure 3

DESIGN SYSTEM

Aerothermal boundary conditions Figure 4

MECHANICALLY ATTACHED, FILM

COOLED SEGMENTED LOUVER

ODS alloy Hastelloy X Figure 5 l30 TRANSPIRATION COOLED,

SEGMENTED TWIN-WALL DESIGN

Gas flow Cooling air Hastelloy X J2"6186 811303 Figure 6

TYPICAL MICROSTRUCTURES

OF CANDIDATE ODS SHEET ALLOYS

HDA son

MA956 J 2.a2 7· 13 ., .....

Figure 7

CREEP PROPERTIES

Creep strength 982°C (1800°F) creep ductility 20 r- 690 (30) HDA 8077 ----__ (0 .1% d Prior 1.0 207 (10)

' ~ ----- --

creep . 0.8 ',, (~~ O ~) MA ~ ' , Stress.

Ma x- % " (0.1% d , 0.6 MPa (ksi) , , 69 .0 (3)

-

\ Hastelloy X 0.4 " (0 .5% d

0.2 Min- R

o ~~--L-L-~J-_L-~ 20.7 (1) L-~::----:'---:L:--~:---:':-_..J 32 34 36 38 40 42 Hastelloy Long Trans HDA ~ X Larson - Mille r Parameter 8077 MA 956 P = T (15 + log t) x 10 .3 J24327 ·25 .,""" Figure 8

982°C (1800°F) CYCLIC OXIDATION

Six minute cycle 0.4 (16) Has telloy X 0.3 (12) Total 0.2 (8) depth of attack, mm (mils) 0.1 (4) ° HDA 8077 ° MA 956 ° 500 750 1000 J24827· 26 Hours """" Figure 9

LOW CYCLE FATIGUE

± 0.25% st rain , 40 cpm 760 ° C (1400 ° F) 871 ° C (1600 ° F) 10,000 M ax Av e Cycles M in to failure H as t ell oy X MA 956 HDA H as t ell o y X MA 956 H DA 80 77 F i g ure 1 0

THERMAL FATIGUE

Hot spot blister test Test rig Test specimen F i g ure 11

HOT SPOT BLISTER TEST RESULTS

TMA X = 982 " C (1800 " F) 0.5 (20) TM IN = 538 " C (1 000 " F) 0.4 (16) MA 956 0.3 (12) 0 Crack penetration, mm (mils) 0.2 (8 ) 0.1 (4 ) 0 100 200 300 400 500 J2 4 82728 Cycles .,""" Figure 12

RIVETED LOUVER LCF

STRUCTURAL TEST

Test rig Test specimen J2 461 8- 15 . ,, ""'" Figure l3

RIVETED LOUVER LCF TEST RESULTS

• No edge cracking • Dimensional stability of MA 956 superior to Hastelloy X Conclusion : MA 956 demonstrates excellent creep resistance with no lCF cracking 4 x 1()4(10') 4 x 10' (1()21

-

Bow resistance, cycles/mm 4 x 1()2 (10') t- (cycles/mill 4 x 10'11001

-

Post - test louver segments ' em 4 X 100 110-'1 L-....L.._L-----'_-'- __ Ha.talloy X MA 966 86-1500 4000-5000 cycle. eye", J:t..v - 2SI

.. -

Figure 14

COMPONENT RIG TESTS

Twin wall Riveted louver LCF test thermal cycle test J2.ta18-17 RO'_ Figure 15

TESTED TWIN - WALL PANEL

• Numerous cooling nole cracks from laser drilled holes • Cr acks in high strain area

o

2000 cycles Figure 16 ADVANCED TURBINE BLADE TIP SEAL SYSTEM J. W. Zelah y General Electric Company In axial-flow gas turbines, the turbine is designed to mlnlmlze the radial clearance between the blade tips and mating shroud segments. This helps to ma x imize aerodynamic efficiency. In spite of the designers' best intentions, the shroud assembly may go out-of-round, and/or the rotor and shroud may be slightly eccentric resulting in potential interference between the blade tips and the shrouds. An y interference which occurs generally removes material from the blade tips (Figure la) in preference to the stationar y shroud, creating a larger annular clearance between the rotor and stator than if the blade tips had remained unaffected and the shroud material had been removed. Furthermore, the blade tip may be damaged, reducing useful blade life, and/or requiring expensive repair operations. At best, an y rub on the bucket tip removes the environmental coatin g , thus making the blade vulnerable to both oxidation and hot corrosion (Figure lb).

A NASA-sponsored (MATE Project 3) program is being conducted to establish and demonstrate the payoff of an advanced blade/shroud system designed to maintain close clearance between blade tips and turbine shrouds and at the same time, be resistant to environmental effects including high-temperature oxida- tion, hot corrosion and thermal c y cling.

The target g o al of this project is to demonstrate the increased efficiency and increased blade life attainable b y using the advanced blade tip. seal sys- tem. Increased efficienc y results from the improved clearance control when blade tips preferentiall y wear the shrouds. Increased blade life results from the superior single-crystal superallo y tip.

The project will establish tip design, joint location, characterize the single-crystal tip alloy, finalize the abrasive tip treatment, fabricate blades, component test and engine test. The project will also establish quality control plans and define the total manufacturing c y cle required to fully process the blades.

The turbine blade tip is of a multicomponent construction consisting of an Activated Diffusion Bonded (ADB) oxidation/hot corrosion resistant single- crystal superallo y squealer capable of withstanding thermal c y cling, combined with a thin layer of alumina (A1 0 ) abrasive particles held in place by an oxidation/corrosion resistant matrlx (Figure 2). Th e shroud materials investi- gated included the current CF6 shroud (Bradelloy) and two advanced shroud materials, Genaseal and Vacuum Plasma Deposited (VPD) CoNiCrA1Y.

The project is structured toward the successful engine demonstration of an improved efficiency, long life turbine blade tip system. The technical effort is divided into nine principal tasks.

1 --

Initial blade tip design work established the joint design and location, optimum squealer thickness and single-crystal orientation (Figures 3 and 4).

The design that was established allows the single-crystal tip-to-blade bonding to be accomplished very early in the manufacturing cycle (possibly at the casting vendor) thereby not appreciably altering the standard manufacturing sequence. The tip design eliminated inside contour mismatch, located the joint in a low stress region and had total manufacturing acceptance. Using property data of both the single-crystal tip material and the bond joint, an economic benefit analysis (payoff) was subsequently performed by CF6-50 engin- eering on the single - crystal/abrasive tip system. The analysis predicted a minimum 2X increase in blade life via the superior tip material and a 0 . 013" tip clearance improvement (0.43% Specific Fuel Consumption (SFC) reduction) as the result of the abrasive tip treatment.

Since the 2X blade life goal was totally dependent upon both the increased environmental resistance of the single-crystal blade tip and the strength of the activated diffusion bonding (ADB) tip attachment process, a comprehensive evaluation of the mecahnical and physical properties of both the Normalloy (single-crystal tip material) and the Normalloy-to-Rene'80(blade material) was conducted. The evaluation included elevated temperatur~ tensile, rupture, oxidation, corrosion and simulated engine thermal shock (SETS) testing. The results of the testing (Figures 5-8) confirmed that the properties exceeded those req u ired for safe engine operation and would be expected to achieve the goal of 2X tip life.

The SFC reduction attainable with the advanced tip system is the direct result of the capability of the abraSive-tipped turbine blade (Figure 9) to resist wear during rub interactions with the shroud material. Several factors including particle size, particle type, particle relief, incursion rate, tip speed, test temperature and to a large degree shroud material have been shown to affect the wear characteristics of the abrasive system. Variations in particle size and type, degree of particle relief and rub incursion rate were evaluated . Test temperature (2000F) and tip speed (1400 ft/sec) were held constant. Three shroud materials: Bradelloy, Genaseal, and VPD CoNiCrA1Y were evaluated. The particle types included various grades of aluminum oxide (A1 0 ) and Borazon (Cubic Boronitride). In all cases, the method of abrasive application was the electroplate encapsulation process.

All wear testing was conducted at the Solar Research Laboratory (division of International Harvester) in San Diego, CA. Solar's facility has the cap- ability of 1400 ft/sec. tip speed, 2000F shroud temperature, and direct read- out/record of all vital functions including chamber temperature, shroud temperature, rotor speed and incursion rate. Measurements of both the blade specimens and shroud specimens were made before and after wear testing to establish the total wear of each. In addition, thermocouples were placed at the surface and 0.050" into the shroud specimens to record surface temperature and shroud temperature rise (and rate) as the result of the incursion. After each test, the blade specimens were evaluated visually, dimensionally, micro- struc.turally and in some cases, by SEM analysis to establish both the total amount of blade and shroud wear and the wear mechanism (i.e., machining, com - paction, melting, etc.) of each (Figure 10). Throughout the program over 50 wear tests were conducted.

The results of the testing showed that in all cases the abrasive tips resisted wear when rubbed into the Genaseal (both new and preoxidized) and the VPD CoNiCrA1Y (Figures 11 and 12). The new Bradello y was shown to be moder- ately abradable. The oxidized Bradelloy, however, was extremely difficult to "cut" and in most cases, after a small incursion into oxidized Bradelloy, the abrasive tips were rapidly consumed (Figure 13). The results of all of the wear testing are summarized in Figure 14. With respect to particle type, with the exception of Borazon, all particles behaved similarly. The Borazon system, in virtually all instances, abraded the shroud materials to a greater degree; even the oxidized Bradelloy was abraded more effectively by the Borazon particles. In addition, neither increased size nor relief signifi- cantl y affected the abrasive characteristics of any particular system. The onl y test variable shown to appreciably effect abrasiveness was incursion rate. Slow incursions, i.e., 0.001 inch per sec or less, were shown to generate higher shroud temperatures and resulted in greater tip wear than at the 0.002 and 0.004 inch/sec tests (typical incursion rates in engines have been estimated at 0.002 inches/second or greater).

The results of the above wear testing have tentativel y indicated that: 1. A large allowable latitude in abrasive system variables exists, i.e., particle type, particle size, relief, and environmental coating can be varied considerably without decreasing the abrasive characteristics of the tip treatment.

2. An oxidation resistant shroud material (e.g. Genaseal or VPD CoNiCrA1Y) should be used to achieve full benefit of the abrasive s y stem.

3. Alundum 38X, 100 grit aluminum oxide/NiCr electroplate with a Codep aluminide coating is the best all-around tip system.

4. Slower incursion rates (i.e., 2 0.001 in./sec.) are more detrimental to the abrasive system than faster incursion rates (0.002 to 0.004 inch/sec).

The abrasive tip system designated for component and engine testing is defined below.

• particle type: 38X alundum (A1 0 ) • particle size: 0.005" - 0.007" diameter • matrix: 0.006"Ni, O.OOl"Cr Diffusion H.T. with aluminide coating • relief: matrix plated "flush" with particles • shroud: either Genaseal or VPD CoNiCrA1Y Using both simulated and actual hardware, the environmental resistance and abrasive capability of the environmental/abrasive tip/shroud system was veri - fied . Wear testing was conducted on Solar wear specimens that were modified with single - crystal/abrasive tips (Figure 15). The wear testing of the simu - lated tip system specimens indicated the tip system was capable of withstanding the rigors of severe shroud rub with no deleterious affects on either the single-crystal tip material or the ADB joint. The single-crystal-to - Rene'80 joint sustained ve . ry severe rub loading, particularly in the case of one bare bladed rub where ~ 0 . 050" of tip was removed and no j oint degradation was evi- dent. Although minimal success was achieved in rubs of abrasive tipped blades into Bradelloy, successful rubs were made into Genaseal and CoNiCrA1Y shrouds without loss of abrasives.

The environmental testing (i. e . , oxidation, corrosion), impact and thermal shock testing will be conducted on actual hardware (scrap "fall-out" from fabrication task) . This testing is currently in progress.

An integrated quality control plan including control over the tip material , the attachment process , the abrasive treatment and all related blade processing operations is currently being prepared. Temporary specifications have been issued and will be revised and updated as needed . Drawings for the single - crystal tip have been issued defining crystallographic orientation and tip configuration . Tooling for inspection of joint thickness has shown dimensional accuracy of + 0.0005 " and has been used to inspect all fabricated blades to da t e.

Each of the separate processing steps established in earlier tasks were formulated into an integrated processing sequence for the manufacture of tur- bine blades with the advanced tip system. The sequence of operations allowed the single-crystal tip bonding to be accomplished without any appreciable changes in normal blade processing (Figure 16). The blades were removed from the production airfoil operation immediately prior to tip cap cavity EDM opera- tio n and ground to a specified length. The single-crystal tips were bonded to the blades and the blades were re - introduced to the airfoil operation for the tip cap cavity EDM operation. The EDM operation provided a smooth tip squeal- er/blade internal wall surface and eliminated any need for internal tip/blade "blending" operations. This task is also still in progress and when completed will fully define the blade casting configuration, tip preparation and heat treatment. the single-crystal tip configuration. orientation and processing, the bonding process operations, fixturing and inspection, the abrasive tip treatment and all nonstandard operations associated with the blade manufacture.

A total processing plan, including step-by - step sequence, will be provided .

A total of 171 blades were subsequently fabricated using the man u facturing sequence defined earlier . Tips were bonded in "dead weight" load fixtures in a cold-wall high vacuum furnace. The activated diffusion bonding (ADB) alloy was D15 (Rene- 80-BASBD Chemistry) and was applied as 0.003" foil. Of the 171 parts that were bonded only 2 failed inspection (joint thickness measurement).

Approximately 150 blades are fully manufactured (Figures 17-20) and are either undergoing or awaiting factory engine test evaluation (Fig u re 20). The remain- ing blades will undergo exhaustive destructive evaluation to further assess process reliability and reprod u cibility .

Two engine tests are planned to fully evaluate the pa y off of the advanced tip s y stem. The first engine test will evaluate the benefits of the single- crystal tip via "e-cycle" (simulated flight cy cle) endurance test ing (1000 cycles minimum). The second engine test will evaluate the abrasive capabilit y of the s y stem via performance testin g under closel y controlled clearance and engine operating conditions. The second test will be of short duration and is designed to "push" the abrasive tip s y stem to the "limit" to full y establish maximum abrasive capability.

The results of the engine tests will be evaluated and anal y zed to assess the effectiveness of the entire system to achieve the program goals.

Successful completion of the program can provide engine manufacturers a viable approach to increase blade life and reduce fuel consumption.

(b) OXIDATION/CORROSION/ (8) TIP WEAR CRACKING FIGURE 1. TYPES OF TIP DETERIORATION Abra ve TIp Treatment Normalloy FI G U R E 2. C F 8 - 50S TAG E 1 H P T B LAD E WIT HAD V A N C E D B LAD E TIP SYSTEM Task I - Blade Tip Seal System Design

Joint Design Location

/ SQUealer Tip Bond Tip Cap

• Low Stress/High Reliability

• Ease of Manufacture

• Consistent With Current Blade Processing

FI G U R E 3. J 0 I N TOE 5 I G N L 0 CAT ION Grain Orientation <.001)

/

FIGURE 4. MONOCRYSTAL ORIENTATION 13 . 0r- ------------------ ------, 12.5 12.0 11 .0 10 .7 Stress (KSI) 10 .0 9.0 Normalloy Normalloy Parent To Rene 80 Material ADS Joint (Avg) (Avg) FI G U R E 5 . 2 0 0 0 F TEN S I L E PRO PER TIE S , , , ' ......

......

......

....

....

....

....

.....

.. ... ...

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

Stress (KSI) ...... ... , .. , ............... ........... ...

...... .. , .. ..... ............... ---- ....

.... ...... ,................... - ...........

Normalloy - No Coating ................ ~ .. -..:-- ..

Normalloy - Coated ........... .... ... : . ::--- Rene 80 - 015 Normalloy - No Coating .........

Rene 80 - 015 Normalloy & Coated

/*

CF6·50 Engine Requirement 1 10 100 1000 Time (Hours) FIGURE 6. 2000F STRESS RUPTURE (AVERAGE) Static Oxidation Hot Corrosion De pt h 01 Depth of 30 Penetration Penetration (in. x .001) (in . x .001) 20 20 Rene ' 80 10 10 No r malloy 0 0 1750 F/ 5ppm SALT @ 700 Hours 2000· F @ 500 Hours FIGURE 7. ELEV A TED TEMPERATURE OXIDATION & CORROSION TESTING

Normalloy

RBO

FI G U R E 8 . THE R MAL FAT I G U ERE R 1ST A N C E (2 0 0 0 THE R MAL C Y C L E S )

r -

Abrasive Particles NICr Electro- Plated Matrix Turbine Blade Tip Material -1 005 In FIGU RE 9_ ABRASIVE TIPPED TURBINE BLADES FI G U R E 1 0 _ SOL A R W EAR T EST S PEe I MEN S - AFT E R T EST :1,.46 Shroud Blade Tip Shroud New Genaseal Shroud Blade Tip Shroud Preoxidized Genaseal FI G U R E 1 1 • W EAR T EST I N G - A BRA S I VET I PIN TON E W & PRE 0 X I 0 I ZED GENASEAL

Blade Tip Shroud

FIGURE 12. WEAR TESTING - ABRASIVE TIP INTO NEW VPD CoNICrAIY Shroud Blade Tip New Bradelloy ..

Shroud Blade Tip Preoxidized Bradelloy

FIGURE 13. WEAR TESTING - ABRASIVE TIP INTO NEW a PREOXIDIZED

BRADELLOY

D Tip loss

- 50 ~ Shroud Wear - 40 VI VI o ...J - 30 "0: :l,....

.r:.-.:,. - 20 en Ci i= - 10 - 5 O~~~~~~~UL-L-L~L+~~~~~~~-L-L~~~~~~--~~~ Blade + 5 Abras i ve Abrasive None None Abrasive None Abrasive Abrasive Treatment Shroud Oxidized Oxidized Material Bradelloy Genaseal CoNiCrAIY FIGURE 14. WEAR TEST RESULTS (2000F, 1400 ft/.ee) Abrasive Tip .VMonocrvstal Tip Material --MarM421 FIGURE 15. WEAR TEST SPECIMEN EDM Braze Drill Receive Machine Codep Finish ~ Tip Cap Airfo il ~ Tip ~ Casting Dovetail Coat Process Cavity Cap Holes Rad i us ADB Bench Grind

Mono U

... ... I-

L 0 .0 . r-'

and Tip Contou r Abrasive To Blade Treat

II n 1'-----)1 H H

Grind ADB Tip and EDM Tip Braze Bench Tip Blade Tip Diffusion H.T. Cap Cavity Tip Cap Outside Contour Finishing Steps: • Finish Process as Standard Blade • Radius Grind Tip • Apply Abrasive and Diffusion H.T.

• Coded Coat and Age FIGURE 18. MANUFACTURING PROCESS PLANS Monocrystal Tip 015 ADS Alloy Rene 80 Slade J FI G U R E 1 7 • B LAD E / TIP / A 0 B ALL 0 Y ASS E M B L Y FI G U R E 1 8. B LAD E SUB COM P 0 N E N T S & F I X T U R E FI G U R E 1 9 . B LAD E WIT HMO N 0 CRY S TAL TIP AFT E R BON DIN G

FI G U R E 2 0 • A BRA SlY E TIP A P P LIE D a F U L L Y PRO C E SSE D

FIGURE 21. STAGE 1 HPT BLADES WITH ADVANCED TIP SYSTEM ASSEMBLED IN TURBINE ROTOR PRIOR TO ENGINE TEST AN INTRODUCTIO~ TO NASA'S TURBINE ENGINE HOT SECTION TECHNOLOGY (HOST) PROJECT Daniel J . Gauntner and C. Robert Ensign NASA Lewis Research Center INTRODUCTION Today's modern gas turbine engines with their high thrust to weight ratio and low specific fuel consumption are comprised of many sophisticated components utilizing the latest high strength materials and technology. This is especially true in the hot section components of the combustor and turbine where high temperature superalloys and protective coatings are necessary in an environment where gas temperatures are well above the melting point of the materials. Current hot section components must endure higher temperatures, higher stresses, and more severe thermal transients than ever before. The durability and efficiency goals of the hot section components operating in this adverse environment will be difficult to achieve. Any shortfalls in achieving these goals could have significant effects on the overall operating cost of the modern gas turbine engine. Early in 1978, NASA began to plan a major project of turbine engine hot section research. Plans called for in-house and contract research to develop and improve the accuracy of current analysis methods so that increased durability could be designed into future engines. This paper is an overview of the new NASA Turbine Engine Hot Section Technology (HOST) project that began officially in January, 1981.

The HOST project was formulated around a simple, yet basic premise.

Specifically, present analysis methods for designing combustor and turbine components need improvements in accuracy and applicability before increases ~n life can be attained during the initial design process of advanced turbine engines. The improved accuracy in life prediction can be attained by conducting focused and directed research efforts in each of the areas involved in cocponent design, including description of the thermal and aerodynamic environments, the material's mechanical response, and the interactions between enviro=ental and structural response . Verification of the more accurate predictions will be a necessary element of the HOST project and it will require high temperature instrumentation capable of measuring near-engine environment effects. The achievement of these improvements will require a rigorous and systematic research effort, beginning with evaluations of current predictive methods by comparing their predictions to benchmark data from special component tests, followed by supporting research to improve the modeling of the physical phenomena, and concluding with tests to verify the improved models in each of the pertinent discipline areas. These areas include structural analysis, surface protection, combustion, turbine heat transfer, and instrumentation .

TURBINE ENGINE HOT SECTION The hot section components of an advanced turbine engine include the combustor and the turbine. A schematic of a typical turbine engine hot section is shown in figure 1. The contoured shaded areas represent an annular flow combustor connected to an axial flow turbine. The combustor liner and turbine airfoil outlines are represented in the figure. The arrows on the schematic represent the flow of the hot section cooling air around the components and through the turbine disk cavities. Because the liner of the combustor and the airfoils of the turbine are the hot section parts exposed to the highest temperatures and consequently suffer a large degree of damage, the research efforts in HOST will be concentrated on improving the analysis methods used to design these three parts. Typically the hot section has twenty percent of the engine weight but accounts for almost sixty percent of the maintenance costs. The consequences of combustor liner failures are generally more economic than operational and result only in a slow, general deterioration of the engine. It is included with the turbine airfoils as part of the HOST project primarily because of the combustor's close coupling and direct effect on the turbine durability.

A knowledge of the basic functions of the combustor and the turbine is necessary if the impact and importance of hot section durability problems is to be understood. In the combustor, the basic release of energy to the core airflow takes place with the burning of the turbine engine fuel. Involved in this energy release are many phenomena, including flow mixing, combustion kinetics, turbulence, flame radiation, soot formation and consumption, liner heat transfer, and gradual acceleration of the high temperature combustor airflow into the turbine. The control of these phenomena by suitable design factors will determine the temperature distribution in the combustor liner and the exit temperature profiles of the airflow leaving the combustor and entering the turbine. In the turbine, this entering airflow is channeled through a set of inlet guide vanes to properly align the flow vectors for optimum and efficient transfer of momentum and energy to the rotating blades of the turbine. The efficiency of the turbine, which contributes greatly to the overall performance and fuel efficiency of the engine, is directly related to the gas flow and temperature distributions. Besides the gas temperature distribution, the gas flow behavior is also needed. Any flow disturbance that inhibits uniform circumferential temperatures or proper radial temperature distribution imposes a penalty on engine performance. This is particularly true near a turbine hub, where large secondary flow vortices are often generated. The extent of these vortices depends upon the quality of the flow entering the turbine. They can be partially controlled by the radial gradients of the energy extracted from a turbine. The temperature and flow phenomena must be better understood and predicted with greater accuracy if life prediction methods are to be improved.

HOT SECTION DURABILITY PROBLEMS The durability of hot section components is highly dependent on such factors as the type of aircraft mission flown, the geographical location of the operating base, and pilot operation. All of these factors affect the temperature and pressure environment and the cyclic load history of the parts in the hot section. During a typical turbine engine design, the type of aircraft mission expected to be flown is expressed in terms of engine cycle information. Design life predictions are made for accumulating levels of repetitive, and somewhat simplified engine cycles. The engine's hot section temperatures and pressures from the engine cycle information are used in these predictions. Variation in conditions due to geographical locations and weather conditions are accounted for in non-standard day test conditions.

Variations due to individual pilots, however, can not be treated deterministically. The design assumes that the engine operates along a worse case cycle.

The incorporation of these real-life variants is beyond the scope of the HOST project. What is possible is to look at factors which affect the durability of the individual components in the gas path of the hot section.

Other programs have gathered experimental and field service data regarding the actual and probable modes of failure for combustor liners and turbine airfoils. Examples of durability problems in components are shown in figure 2. Typically, air-cooled combustion chambers experience large, thermally induced strains that exceed elastic limits of materials at points of maximum stress and/or temperature. Creep-low cycle fatigue interactions and louver lip collapse have been established as primary burner liner failure modes. Oxidation/erosion modes tend to be secondary failures, usually caused by some other damage mechanism. For turbine airfoils (vanes and blades) creep-fatigue cracking and oxidation/corrosion tend to be dominant failure modes. But for the airfoils, these modes are of more importance and usually necessitate engine removal when detected to prevent further damage such as blade or vane rupture. These modes of failure for the combustor liner and the turbine airfoils have been selected as pertinent examples, but they are not all inclusive.

RESEARCH EFFORTS Approach The HOST project will support research to improve the accuracy of analysis methods, which can be used during engine design to increase component durability levels, thereby reducing maintenance and operating costs of the turbine engines. Research will be funded in the areas of structural analysis, surface protection, combustion, turbine heat transfer, and instrumentation.

The overall approach of HOST in each of these areas will be to: (1) evaluate existing models; (2) quantify their strengths and weaknesses; (3) conduct new experimental and analytical research to more accurately model the physical phenomena; (4) use the new models in predictive a~alyses and verify their improvements in accuracy; and, (5) conduct a sensitivity study to assess the improvements in overall hot section durability to be achieved by use of a combination of these new methods.

The HOST project emphasizes the coordination of the research activities (fig. 3) to provide a system of more accurate analysis methods. The use of these improved methods will lead not only to enhanced durability; but also to lower maintenance costs for the hot section, freedom for more innovative design and checkout of new ideas, the ability to perform more accurate trade-off studies between performance and durability plus high reliability 1n future engines.

The specific elements of research that will be supported and coordinated in HOST are shown in the work breakdown structure in figure 4. The technical aspects of the activities in the six columns of figure 4 will be managed by staff members from four different divisions at the Lewis Research Center.

This delegation of technical responsibility is illustrative of the matrix management concept that will be used for the HOST project. Descriptions of the planned research are presented in the following paragraphs.

Structural Analysis Some typical damage observed on one component from the hot section of a turbine engine is shown in figure 5. This section of a combustor liner shows thermal fatigue cracking. To approach such a problem, the structural analyst must have sophisticated tools for accurate analysis. These include a knowledge of the thermal and mechanical loads, inelastic methods of analysis such as nonlinear finite element computer codes, cyclic constitutive (stress-strain) relationships, and the capability to determine the effects of the creep-fatigue interactions on crack initiation.

The structural analysis efforts under HOST will pursue areas such as thermal loading prediction methods, specialized vane and liner geometric and structural analysis models, methods and procedures to determine time and temperature structural response characteristics, and improved methods to describe time dependent and time independent inelastic material behavior. In addition, material constitutive relationships will be improved for predicting material behavior response to cyclic variations in stress, strain, and temperature with time. Also, life prediction methods will be developed for crack initiation models. The existing methods for such problems will be improved and automated to reduce the required manpower and computer time. For instance, the thermal analysis methods will be integrated with the structural analysis codes, so that the relatively coarse thermal map of a component becomes the input to the more detailed finite element program. Also, the methods will include self-adaptive solution strategies that use substructuring to examine the inelastic regions of a component with an overall elastic behavior.

The specific elements under HOST in the area of structural analysis are listed in figures 6(a) and 6(b) along with the expected results. The bars show the expected starting times and durations of each effort, in terms of fiscal years, which run from October through September.

The first element listed under structural analysis in figure 6(a) represents a planned effort to develop a computerized method to transfer the thermal loads that a burner liner might experience to a structural analysis model. The method will automatically integrate information from a thermal analysis computer code with an advanced nonlinear structural analysis code.

The next element, shown by the bar extending from FY82 to 86, extends the application of the first method to prediction of loads that are component related, and time dependent for other hot section components for various engine mission cycles. It also will include effects of local hot streaks, cooling holes, and thermal anisotropy, as found in turbine blades and vanes.

The structural analysis methods of the future will require improved versions of today's computer capabilities such as 3-D nonlinear finite element methods that can handle plasticity, creep, strain concentration, and unsymmetrical thermal effects found in hot section components. For effective structural analysis of the hot section, the codes will handle all these interacting inelastic effects. HOST will develop these capabilities and determine strategies and self-adaptive algorithms for solution of such complex analysis problems. After these automated modeling and solution strategies are completed, they will be verified by comparison with data from tests of specific engine components subjected to typical thermal and mechanical forces from appropriate mission cycles. The final program element in figure 6(a) will include component specific models and verification of the above efforts.

Within the computerized structural analysis methods are equations which model the behavior of the material when it is subjected to various loads and temperatures. These equations represent different theories and engineering models that attempt to describe the physical phenomena taking place. The theories, and hence these constitutive equations, must describe the response of a material subjected to both mechanical and thermally induced stresses and strains. For low temperatures, when the material is in the elastic , range, the theories are quite adequate. But in the hot section of the turbine, most parts are well into inelastic behavior, and the modeling becomes very complex. Many theories have been proposed to describe this behavior. Several elements of HOST in figure 6(b) will evaluate the various theories and models to understand and improve upon the constitutive equations.

The first of these elements in figure 6(b) will determine the best model to represent the cyclic behavior of isotropic materials. The model will include the complexities of creep-plasticity, multi-axial stress and strain, plus the effects of long-time exposure of surfaces. The second element will develop and verify similar constitutive models for anisotropic materials, such as those used in the manufacturing of directionally solidified vanes or blades. The final product of these efforts will be sets of equations that represent the inelastic response of hot section components with greater accuracy than today's methods.

The second aspect of the HOST project in figure 6(b) is the prediction of the life of hot section component parts using an understanding of the synergistic effects of creep, fatigue, and environment on crack initiation behavior. Existing models that are explicit in the primary variables of stress, strain, temperature, environment, and time will be screened. The first element under Life Prediction Methods will select and develop a model for a specific isotropic material/coating combination that is typical for a liner or vane. The effects of mission loading, multi-axial stress and thermal cycling will be included. A second and parallel program element of HOST will develop a similar life prediction method for an anisotropic material/coating combination for a liner or blade. Both of these efforts will consist of a concentrated effort of laboratory testing resulting in modifications to life prediction methods. After sufficient validation, a second material/component combination will be examined in each of these efforts.

Surface Protection Significant work has been done to further the science and technology of coatings. Figure 7 shows micrographs of a NiCrA1Y cQating before and after soaking for a long time at 1366 K. The coating is degraded not only by the hostile environment, but also by its diffusion at the substrate boundary.

HOST will concentrate on analytical methods to account for each of the effects of environment, corrosion/erosion, oxidation/diffusion, and metallic coatings to be able to predict the time to crack initiation of coatings and coated hot section parts.

The HOST effort will concentrate On modeling the effects of environmental attack and coatings On crack initiation, the location and rate of erosive particle impact and corrosive salt depositions On airfoils, and also the coating degradation On blades, vanes, and combustors to provide coating life predictions. The various surface protection elements under HOST will study the phenomenological effects and interactions, and will produce analytical models for different types of components (i.e., turbine blades, vanes, and combustor liners) .. All of these models will be evaluated and verified using either real data from engine field failure experience or laboratory data from erosion/corrosion burner rigs.

The first surface protection element of HOST in figure 8(a) will model the effects of environment and coatings On the creep-fatigue crack initiation of isotropic materials used for liners and vanes. Later, another element will produce similar models, but for anisotropic materials such as directionally solidified vanes and blades. These two elements will be combined with the Life Prediction Methods of figure 6(b), as represented by the dashed lines of figure 8(a). As another element, the behavior of sheet materials coated for use as combustor liners will also be obtained during cyclic testing in a suitable test rig.

Other research efforts in the area of Surface Protection are shown in figure 8(b). The effects of corrosion and erosion, and their interaction, will be modeled and then used as part of a more comprehensive coating life model. To assist in developing a corrosion/erosion model, research elements investigating mass deposition on airfoils and the location and rates of erosion on airfoils will be conducted. The third element of the corrosion/erosion model effort will include rig burner tests of the combined corrosion/erosion mode to verify the deposition ~nd erosion models.

Under the coating life model of figure 8(b), the first research element will collect engine field failure experience data for coatings to provide a real environment data base. The effects of oxidation and corrosion will be investigated and then modeled in inhouse tests to verify the effects of this dual cycle mode of attack. Next, the corrosion/erosion and dual cycle models will be combined. Life predictions will be verified in test rigs. Finally, a test program will obtain correlations of rig test effects and engine test effects on coating life.

Combustion Present turbine engine combustors exhibit very complex flow conditions and high levels of heat transfer by radiation (fig. 9). These conditions make the prediction of gas and metal temperatures very difficult. To aid in this task, the combustion research will be conducted in the areas of aerothermal modeling and liner cyclic testing. To support this analytical work, a test program will be developed to study gas flow and mixing phenomena and flame radiation effects. Also, plans call for the design of a test rig that can obtain accelerated low life data for liner segments subjected to thermal cycling. The design will be difficult to obtain, since the thermal loads on the liner segment must simulate the real loads on a full circular liner, if the accelerated life data is to be useful.

The first combustion element in figure 10 will assess the existing aerodynamic and thermodynamic models to determine their capabilities, deficiencies, and the priority of areas requiring improvement. Research and model refinements will then be made in areas such as internal flow and exit temperature pattern factor, as well as in the mathematical routines used in computer solutions (e.g., faster solutions of the Navier Stokes equations). An experimental study of the penetration and thermal mixing characteristics that result when secondary (dilution) jets are used in combustors will be conducted. This work will add the empirical relations to explain the effect of dilution jet parameters on the exit temperature profile. Another experimental program will provide comprehensive luminous flame radiation and liner heat flux data for varying gas flow conditions. The effects of pressures up to 40 atmospheres on the luminous flame radiation will be included in the test program, and the results modeled.

The final aspect of the aerothermal modeling activities under HOST will begin in 1984. This "integration" phase will put together all of the submodels and routines that will be developed in the model refinement and testing activities. It will also include an assessment of the improvements made.

The low cycle fatigue life of combustor liners will be studied by running thermal cyclic tests on segments of liners. This. data will be used with the life prediction methods described earlier under structural analysis. The effects of hot streaking on t he life of combustor liners will also be investigated.

Turbine Heat Transfer The turbine heat transfer research to be conducted ~n this area includes research into gas path analysis, gas side heat transfer, coolant side heat transfer, and metal temperature prediction. Advanced turbine engine design requires accurate predictions or knowledge of the local metal temperatures of the various static and rotating parts. For the turbine, as exemplified by the schematic cutaway in figure 11, these analyses must consider the characteristics of the gas flow at the entry, including its temperature, pressure and turbulence levels. The extremely complex flow field around the blades and along the walls must be understood and modeled, before the temperature of static and dynamic airfoils can be calculated. If the gas temperature and heat flux conditions for each row through the turbine are known, the heat transfer coefficients for blades, vanes and endwalls can be calculated. The coefficients can then be used to calculate the operating temperatures of these parts for transient as well as steady-state conditions.

Finally, the information can be used to analytically optimize the design (and durability) of the components for various materials and geometries.

The efforts under HOST will be both experimental and analytical. They will establish benchmark quality data, model the complex heat transfer mechanisms, and, finally, provide the methodology for determining temperatures and heat transfer coefficients, which can then be input to structural analyses routines. The first two turbine elements in figure 12 will evaluate the effect on flow transition of variables such as Reynold's number, turbulence, geometry, and temperature ratios for vanes with and without the effects of film cooling. Viscous 3-D analyses to predict heat transfer and gas flow for stator and rotor cascades, including side and endwall effects, will be undertaken as part of HOST. In the next element, the heat transfer and flow characteristics will be determined for various geometries of multiple jet impingement arrays. The influence of rotational (Coriolis) forces and entrance geometry on the prediction of coolant-side heat transfer coefficients will be studied in another element of HOST. Steady-state and transient metal temperature prediction codes will be improved, and interfaced with structural analysis codes, by using the improved flow and heat transfer models above.

Research to measure local heat transfer coefficients over a stator vane, and a rotating blade will be included to verify the development of the above models. Measurements using improved instrumentation will be made to help evaluate the accuracy of codes for predicting gas -side heat transfer coefficients, metal temperatures and static strains in the materials .

J

Instrumentation Development Crucial to the experimental effort of HOST are accurate measurements of the temperature, pressure, strain, and heat flux in the hot gas flow stream of the turbine engine. These measurements will be made to provide the benchmark quality data required for verification of the models developed in the other areas. Many of the measurements will require instruments that extend the present state-of-the-art. Fortunately, new techniques and computerized instrumentation (fig. 13) offer promising solutions and exciting extensions of current technology.

The first element in figure 14 will make use of new thin film sputtering techniques to develop a miniature heat flux sensor that is applied directly on blades and vanes. Also to be developed is a method for measuring the radiation portion of the total heat flux to sections of the combustor.

Current static strain gages can operate at temperatures of 650 K (or 920 K for a few hours). By using thin film or powder metallurgy techniques, HOST will develop new static strain gages and installation methods for temperatures up to 1250 K. The third element of figure 14 will be the development of a viewing system that is needed for observation of the hot section components during operation at near engine condition temperatures and pressures. For instance, inside the combustor, the edges of the liner could be viewed to see if they are buckling or closing. Also, the interactions of the swirling flow of gases and fuel spray could be carefully studied. The increasing of the clarity of the view of these phenomena within the hot section will be a major part of this effort.

An automated laser anemometer system to measure the three components of average and fluctuating velocities will be developed. The final effort ~n instrumentation under HOST will produce a probe to measure dynamic gas temperatures up to 1000 Hz. Present temperature probes with fine wire thermocouples having electronic compensation are limited to a frequency response of about 30 Hz. The compensation depends on the gas stream flow properties (Mach number, density, etc.) but these vary during a test, so the compensation must also be dynamic, following these parameters in real time.

CONCLUDING REMARKS The Turbine Engine Hot Section Technology (HOST) project, discussed above, will utilize current models and conduct new research to develop improved and more accurate analysis methods for the design of advanced turbine engine components. The research in the five areas of structural analysis, surface protection, combustion, turbine heat transfer, and instrumentation will be focused so that problems in hot section component durability can be understood and overcome. Current plans for the research call for eighty percent of the work to be done by engine manufacturers and other competent research institutions. The remaining twenty percent of the HOST effort will be accomplished inhouse by NASA Lewis Research Center technical personnel.

Although the HOST project includes research efforts in a number of separate technical disciplines, its organization is that of a systems technology project. As such, it has identifiable schedules with intermediate milestones and project end dates. The specific project products, as defined above in the text and in figures 6 to 14, are the key part of a systems tech - nology project. While current plans and thinking are presented, it must be recognized that a certain amount of risk exists that some of the project's products may prove to be too far beyond the state of the art or not achievable by the end of the HOST project in fiscal year 1986. Individual research efforts will be monitored and appropriate plan changes made, if required, to ensure that the HOST project attains its objective .

The products of all of the HOST-supported research, excluding the instrumentation development, will be presented in the form of individual models, or in some cases, as computer modules, that can be acquired separately and utilized by engine manufacturers in analyzing designs of advanced turbine engine components. No attempt will be made during the HOST project to integrate the individual models into one overall model. All improved models, benchmark data bases, and any programmed computer modules will be disseminated to the U. S. domestic aerospace industry through formal reports and at suitable workshops and meetings. Thus, the U.S. engine manufacturers will be able to improve the durability of hot section components in their advanced turbine engine designs. This enhanced ability will enable the U.S. aerospace industry to maintain its favorable position in an increasingly competitive world aerospace market .

\ I -< >-

"/ " /

I \ "" .

----

- --- ______ w ____ --~· ......

a- .c-

-

Figure 2. - Hot section components.

HEAT TRANSFER TEMPERATURE

~ow MODELlN~

ACCURATE MATERIAL ANALYSIS COMPETITIVE INNOVATIVE DESIGN COSTS EVALUATIO Figure 3. - Coordinated research activities.

Figure 4. - Work breakdown structure.

THERMAL FATIGUE CRACKS IN COMBUSTOR LINER t-' 0\ 0\ Figure 5. - Structural analysis research.

- -- ------ ----------------------- ANALYSIS METHODS EXPECTED RESULT FY81 82 83 84 85 86 PROGRAM ELEMENT COMPONENT-RELATED, TIME- INTEGRATED THERMAL-MECHANIC DEPENDENT, THERMAL - MECHANICAL LOAD-MISSION MODEL LOAD HISTORY SELF-ADAPTIVE ALGORITHMS 3-D INELASTIC ANALYSIS FOR OPTIMUM 3-D, NONLINEAR, METHODS WITH 2-D!3-D TIME-DEPENDENT FINITE LINEA R/NONLINEA R ELEMENT STRESS-STRAIN SOLUTION STRATEGIES ANAL YSES CONSISTENT AUTOMATED THERMAL- COMPONENT SPECIFIC MODELS STRUCTURAL MODELS WITH WITH VERIFICATION TESTS COMPONMENT BENCH VERI FICA TESTS (a) Analysis methods.

MATERIALS CONSTITUTIVE RELATIONS PROGRAM ELEMENT FY81 82 83 84 85 86 EXPECTED RESULT MODEL FOR ISOTROPIC ENGINE CYCLIC MODELS FOR MATERIALS COMPONENT SPECIFIC (LINERS, VANES) MATERIALS DEMON- STRATED MODEL FOR ANISOTROPIC ENGINE MATERIALS (D. S. BLADES) LIFE PREDICTION METHODS PROGRAM ELEMENT EXPECTED RES ULT CREEP-FATIGUE CRACK INITIATION MODELING FOR ISOTROPIC ENGINE MATRL'S (LINERS, VANES) LIFE PREDICTION MODELS FOR COMPONENT SPEC I FIC MATERIALS DEMONSTRATED CREEP-FATIGUE CRACK INITIATION MODELING FOR ANISOTROPIC ENGINE MATRL'S BLADES & VANES (b) Materials constitutive relations and life prediction methods.

Figure 6. - Structural analysis.

SURFACE PROTECTION

ENVIRONMENTAL AND SUBSTRATE REACTIONS

CORROSION EROSION DEGRADE COATINGS

~ L--~ ____ ~ __________________________________________ ~ DIFFUS IO N

METALLIC COATINGS

~ REACTION

ENVIRONMENT

X250 • X250 AFTER 200 h r AT 2000° F AS-DEPOS ITED NiCrAIY COATING Figure 7. - Surface protection research.

~ ENVI RONMENTALIMECHANICAL PROPERTY INTERACTIONS RAM ELEMENT CIF INITIATION ISOTROPIC (LINERS, VANES) ENV./COATING EFFECTS TASKS MODELS FOR EFFECTS OF ENVIRONMENTAL ATIACK & COATINGS ON CRACK INITIATION CIF INITIATION ANISOTROPIC (D. S. VANES, BLADES) ENV./COATING EFFECTS TAS SHEET MATERIALS IH EFFECTS OF ENV.

ATIACK & COATINGS ON CRACK INITIATION IN COMBUSTOR LINERS (a) Environmental/mechanical property interactions.

- CORROSION/EROSION MODEL - PROGRAM ELEMENT FY8! 82 83 EXPECTED RESULT 84 85 86

I I

MODEL TO PREDICT CORROD- AIRFOIL DEPOSITION MODEL - G ANT DEPOSITION ON TUR- BINE AI RFOILS

I I

AIRFOIL EROS ION MODEL MODEL TO PREDICT LOCA- TlON AND SEVERITY OF EROSION ATIACK OF TUR- BINE AIRFOILS

I I

CORROSIONIEROSION MODEL - IH VERIFICATION OF MODELS I IN RBT I I I - COATING LIFE MODEL - PROGRAM ELEMENT FY8! EXPECTED RES ULT 82 S3 84 85 86

I

ENGINE DATA BASE OXIDATION LIFE PREDICTION - IH

I

CAPABILITY TO CORROS ION LIFE PREDICTION PREDICT COATING DEGRADATION ON DUAL CYCLE ATIACK - IH BLADES, VANES, HOST-CORROSIONIERQSION MODEL COMBUSTORS LIFE PREDICTION VERIFICATION RIGIENGINE CORR - IH (b) Corrosion/erosion and coating life models.

Figure 8. - Surface protection.

COMBUSTION

AEROTHERMAL MODELING

f-' " o

LINER PANEL SEG M ENT

THERMAL CYCLIC DATA

Figure 9. - Combustion research.

PROG RAM ELEMENT EX P ECTED RES ULT FY 81 82 83 84 85 86 AEROTHERMAL MODE LI NG A.

MODEL AND DATA 1. AS SESSMENT DE FICI ENC I ES IDENT IFI ED 2. MO DEL REFIN E MENT & NEW PHYS I CAL MODELS, SUPPORT I NG R E SEARCH IMPROVED COMPUTING METHODS a. DILUT I ON -J ET EXIT TEMPERATURE PROF I LES MIXING PRED I CTED HIGH PRESSURE FLAME RAD I- b. FLAME RAD I ATION (i - H) ATIO N A ND H EA T FLUX DATA 3. I NTEG RA nON VER IFI ED MODEL I MPROVE- MENTS B. LI NER CYCLIC LIFE 1. LI NER PANEL LINER TEST PANELS DE- .,..

DESIGN SIGNED 2. CYC LI C TEST I NG IMP ROVED AB I LITY TO ACCU- II - H) R ATEL Y PR ED I CT LINER CYCLIC LIFE Fi g ure 10. - Combustion analysis.

• ENTRY GAS FLOW CONDITIONS -TEMPERATURE , PRESSURE, & TURBULENCE CHARACTERISTICS • GAS CONDITIONS FOR EACH ROW • HEAT TRANSFER COEFFICIENTS -BLADES, VANES, END WALLS • TEMPERATURE OF COMPONENTS -STEADY STATE & TRANSIENT • OPTIMIZATION TO IMPROVE DURABILITY -METAL TEMPERATURES & GEOMETRY F igure 11. - Turbine heat transfer research.

PROGRAM ElfMENT ,Y81 82 83 84 85 86 EXPECTED RESULT GAS-S ID E HEAT TRANSFER, NON - ROTATING, DETERMINE INFLUE NCE OF

•

2-D VARIABLES ON FLOW TRANS-

-

ITiON & DURATION & IMPROVED GAS-S IDE HEAT TRANSFER, NON-ROTAT IN G, MODELS F IL M

•

•

GAS FLOW ENVIRONMENT AND HEAT TRANSFER, IMPROVE PREDICTION CODES NON-ROTAT I NG FOR ENV I RONMENT & GAS-SIDE HEAT TRANSFER, NO ROTATION GAS FLOW ENV I RONMENT AND HEAT TRANSFER, SAME AS ABOVE WITH ROTATION ROTATING, 3-0 MUL TIPLE JET ARRAY IMPINGEMENT IMPROVED HEAT TRANSFER CORRELAT I ON FOR IMPINGEMENT

•

COOLING COOLANT S IDE HEAT TRANSFER W IT H HEAT TRANSFER CORRELATIONS , ROTATION AND ENTRANCE GEOMETRY I NCLUD I NG EFFECTS OF ROT A- TlONS AND ENTRANCE GEOMETRY METAL TEMPERATURE PREDICTION CODES METAL TEMPERATURE PRED- IC TION CODES WITH IMPROVED HEAT TRANSFER MODELS I CORRELATIONS IN -HO USE RESEARCH AND VERIFICATIONS VERIFICATIONS OF FLOW, HEAT TRANS FER, AND STRAIN PREDICTION AT NEAR REAL TU RBINE ENV I RONMENT Figure 1 2. - Turbine heat transfer.

PHOTOMUL n PLiER • SENSORS: HEAT FLUX, (PM) TUBE STRAIN & TEMPERATURE • LASER ANEMOMETER \ ~ VELOCITY '~ COMPRESSOR ROTOR MEASURED Figure 13. - Instrumentation development.

PROGRAM ELEMENT FY81 85 86 EXPECTED RESULT TOTAL AND RADIATIVE HEAT · FLUX SENSORS HEAT FLUX MEAS UREMENTS STRAIN SENSORS MINIATURE STATIC STRAIN GAGES FOR 1800F APPLICATIONS HOT SECTION SYSTEM FOR VIEWING VIEWING SYSTEM HOT SECTION COMPONENTS DURING OPERATION LASER ANEMOMETER FLOW MEASUREMENTS

H

••••• _iI )

SYSTEM DESIGN AND DEVELOPMENT FOR A LeRC FACILITY DYNAMIC GAS TEMPERA- GAS TEMPERATURE TURE MEASUREMENT SENSOR PROBE WITH 1 kHz RESPONSE Figure 14. - Instrumentation.

I

I

I THE NATURE OF OPERATING FLIGHT LOADS AND THEIR EFFECT ON PROPULSION SYSTEM STRUCTURES Kenneth H. Dickenson Richard L. Martin Boeing Commercial Airplane Company ABSTRACT Past di agnosti cs studi es revealed the primary causes of performance deterioration of high by-pass turbofan engines to be flight loads, erosion and thermal distortion. This paper examines the various types of airplane loads that are imposed on the engine throughout the lifetime of an airplane. These include flight loads from gusts and maneuvers and ground loads from take-off, landing and taxi conditions. Clarification is made in definitions of the airframer's limit and ultimate design loads and the engine manufacturer's operating design loads.

Finally, the influence of these loads on the propulsion system structures is discussed.

INTRODUCTION The traditional transport airplane structures analyst's treatment of an engine is very simple: The engine is a "concrete block" whose properties are entirely inertial, and the main concerns are that it should not falloff the strut and that it is located properly from a wing flutter viewpoint (figure 1). In the era of the turbojet and the low bypass ratio turbofan, such treatment was acceptable because it was nearly correct. Engines and their inlets were so compact and rigid (figure 2) that the internal structural problems could be left to the engine manufacturers who needed only to be advised of the accelerations to be applied at the mount locations.

In the late 1960's the advent of the high bypass ratio turbofan engine with it's large fan case relative to its core brought change. Early in the Boeing 747 program, for example, it was found that thrust forces caused "ovalization" of the engine case because the engine's combination of large diameter (figure 3) and high thrust imposed a substantial couple at the engine mounts. This problem was alleviated by adding a "thrust yoke" that transferred thrust directly to the strut and reduced distortion of the case due to thrust.

The high bypass ratio turbofan engines have large inlet airflows relative to the engine core size. Thus, a large momentum change is required to align the airflow with the engine at high angles of attack. Since inlets are usually bolted to the front flange of the fan case, the inlet aerodynamic loads associated with this momentum change induce bending and distortion into the smaller diameter engine core case. These case distortions may cause rubbing between the rotors and the static case structure while the desire for higher overall pressure ratio requires better control of tip clearances. The aerodynamically induced operating loads that act on the inlet are modest in comparison to overall airframe design loads.

It remained for the 1973 oil embargo and the ensuing dramatic rise in fuel prices to motivate a deeper investigation of airframe and engine structural interaction.

Attention was focused on the causes of engine fuel consumption deterioration in

the NASA sponsored Pratt & Whitney Aircraft JT9D Diagnostics Program. In th i s

effort, several different probable deterioration mechanisms were identified and evaluated analytically. Prominent among them was rubbing between the rotor tips (of the fan, compressors, and turbines) and the engine case caused by flexing of the engine under operating loads. The result of rubbing was increased clearance between the rotor and case since material was worn from the "rub strips" and the blade tips. Increased clearance reduced component efficiency and increased specific fuel consumption.

AIRPLANE LOADS Design Loads Due to the overriding importance of safety, airplane design loads have been studied intensively for many years and are the subject of a large body of doctrine and practice developed by airframe manufacturers and enforced by the Fed era 1 A vi at ion Adm i ni s t rat ion (F AA) . A II 1 i m it des i gn loa d II i s de term i ned for airframe structure as the maximum load that the structure can be expected to encounter during the entire life of the airplane fleet. At the limit design load, the structure is not permitted to suffer permanent deformation; i . e., the maximum stress may not exceed the elastic limit. To provide an added degree of safety, an "ultimate design load" is also specified, usually as 1.5 times the limit load value. Up to the ultimate load, the structure is permitted to suffer "permanent set", but it must not fail.

The design loads are determined by analyzing the airplane in a variety of load conditions that are contained within the envelope of the "V - n diagram" (a plot of the accelerations that the airplane must withstand versus airspeed). Three main types of load conditions are considered. The first is maneuver. Transport category airplane limit loads are determined in 2.5 g turns or pull-ups with flaps retracted and in 2.0 g turns with flaps down. The second type of load condition is due to atmospheric turbulence. It is assumed that gu s ts of a defined shape and velocity will be encountered by the airplane at speeds specified in relation to the design operating speed limits chosen by the manufacturer. The airplane's response is determined by analyzing the aerodynamic, elastic, and inertial characteristics in detail. The third loading category is associated with the ground. These loading conditions include take- off, landing and taxi. Analytically determined design loads are corroborated by extensive fl ight load surveys using accelerometers, strain gages and pressure transducers.

The oustanding structural safety records of today's commercial air fleets demonstrate that the design loads issues are very well understood. However, the loads that cause day-to-day TSFC deterioration are less severe than design loads and are not so well understood. They may be termed "operat i ng" loads, and a different approach is needed to understand them.

Operating Loads The parameters for determining operating loads are the same as those for design loads; i.e., airplane aerodynamic, elastic, and inertial properties on the one hand, and fl ight conditions (maneuvers, turbulence, etc.) on the other. To a considerable degree, the problem resembles that of analyzing structural fatigue.

Service life, maintenance, and economy are the main considerations, with statistical descriptions of the operating environment being the scenario, as opposed to a set of extreme conditions.

Fatigue damage is assessed from the cumulative occurences of different stress levels. The part of TSFC degradation due to clearance change, on the other hand, depends on the probable time period (or number of flights) until any given load level is exceeded once. Figure 4 shows the analysis sequence. The starting point is the set of mission profiles that typify the airplane's utilization (upper left corner). Mission length is important because it determines how many "ground-air-ground" (GAG) cycles are flown per hour of operation. The altitude and speed profiles determine the frequency and severity of gusts.

Load exceedance probability per flight can be inferred from airplane character- istics and the mission profile. The sketch at the upper right of figure 4 refers to inertia load exceedances, but a similar plot can relate to airloads. When the probable loads are known, the probable tip clearance changes can be inferred from the elastic properties of the engine itself. These may be obtained by analyses varying from simple beam representations to finite-element models containing thousands of elements. Recent experience supports the need for the more complex finite-element approach. When tip clearances become negative, rubs are indicated, and TSFC deterioration can be expected.

In addition to revenue service missions, other flight profiles must be considered, such as crew training. A significant mission that occurs only once on each airplane is the "acceptance flight" (figure 5). All transport airplanes are checked for satisfactory flight characteristics and functioning of warning systems before delivery to the customer airline. In such flights, the airplane is not tested to the limit loads of the flight design envelope but to more normal operating conditions, such as maximum airspeed (dynamic pressure), maximum Mach number, and minimum airspeed (stall warning) where warning devices such as stick shakers automatically alert the pilot to the situation. Since such a flight always occurs first in an airplane's history, it establishes a starting set of rubs and clearances for subsequent exceedance studies.

Statistical descriptions of the inertia load environment have been obtained by accumul at i on of a great many speed/acce 1 erati on/a lt itude ("VGH") recordi ngs made in actual airline service (figure 6). Airspeed (V), normal acceleration (g), and altitude (H) are recorded continuously. The recordings are later analyzed by counting acceleration peaks. The number of "occurrences" of a particular acceleration level is defined as the number of peaks found over some time period that fall between the upper and lower bounds of that 1 eve 1. The number of "exceedances" of that level is the sum of the occurrences of that level and all higher levels as shown in figure 7.

Histograms (figure 8) can then be constructed which show occurrences and exceedances per flight hour or per flight versus load level, and plots such as shown in figure 9, depicting probable nacelle inertia load exceedances, can be drawn. This figure, incidentally, shows a characteristic feature of the environ- ment of wing-mounted engines. The motions and accelerations of the nacelles are 1arger than those at the a i rp 1 ane center of gravity because of the wi ng aeroelastic response to dynamic loads such as gusts and landing impacts. In additon to accelerations, gyroscopic loads caused by airplane angular motions must also be considered.

Under normal conditions, the most severe engine aerodynamic loading occurs at takeoff when the maximum engine thrust produces a high mass flow rate through the inlet combined with a high angle of attack. These effects are illustrated in figure 10 which shows low pressure caused by suction on the lower inside lip of the inlet. This is associated with high local velocity as the flow turns sharply. Supersonic flow usually occurs in this region, followed by shocks and sometimes by local flow separation.

Of lesser importance, but significant because it is a condition that creates a load reversal, is the maximum dynamic pressure condition shown in figure 1l.

This condition involves a negative local inlet angle of attack and an inlet pitching moment acting downward. This moment is, however, of much smaller magnitude than the nose-up moment at takeoff.

Neither the maximum dynamic pressure condition nor the one shown in figure 12 -- stall warning at 10 flaps -- are normal revenue service flight conditions. Both of these, however, are currently flown in the flight acceptance test of every new airplane.

One of the more uncertain assumptions regarding inlet pressures has been the circumferential distribution. A simple, one wave cosine distribution, illustrated in figure 13, rotated to account for non-symmetric effects has been customary. This assumption awaits validation by the results of the Nacelle Aerodynamic and Inertia Loads (NAIL) flight test program.

To illustrate the joint efforts in the JT9D Diagnostics Program and ,the interde- pendence of the engine and airframe manufacturer in the propulsion interface, Figure 14 shows the mathematical model used to analyze the 747 propulsion system.

Government and industry foresight several years ago provided the NASTRAN finite element program giving wide availability to this technology. Air breathing propulsion structures are a relatively late application of this technology, and there is currently a large effort being made toward test and analysis correlations to enhance this application.

Currently, Boeing and the three major jet engine manufacturers are utilizing the type of models illustrated here to calculate deflections, clearance changes, internal loads, and vibration behavior on the 767, 757 and 737-300 new airplane programs and on future powerplant installations for the 747 airplane program.

The airframe and engine manufacturer must each conduct their own analyses for their specific needs. An exchange of data files provides each with this capability and the integrated model as illustrated in figure 15. Recent trends in nacelle design have resulted in much closer structural coupling between the engine, nacelle, and strut. For example, in the 767 design the front mounting system has been placed to minimize thrust bending moment, and the thrust reverser and fan exhaust cowling are hinged from the strut and clamp onto the engine through circumferential V-grooves. This not only simplifies engine removal and maintenance but also serves as a dual load path with the mounts which provides the redundancy required for fail safe design in case of mount failure. An extra benefit from this dual load path is a reduction of engine loading which enhances engine performance through reduced clearance changes under fl ight operating loads. The close coupling inherent in this type of design necessitates use of detailed nacelle-engine-strut finite element models to define interface loads accurately.

A characteristic of engine structure relative to conventional airframe structure is its inherently greater stiffness and complexity. This obviously must be the case in order to maintain the dimensional constraints so important to engine performance. The maximum engine bending deformations are typically one to two orders of magnitude less than maximum strut deflection as exhibited in figure 16 for a II gil loading condition. The attainment of accuracy in the engine and nacelle math model comparable to conventional airframe structures, therefore, requires a great deal of experience and effort and should rely heavi lyon accurately measured data when available.

To illustrate this point, figure 17 shows typical calculated clearance change contour lines for a normal takeoff condition. This plot is for the inboard side of the number three engine on the 747 airplane. Clearance closure is denoted by the shaded regions. The information shown here is used by the engine manufacturer in a separate post processor program that calculates blade rubs, stage-by-stage clearance increases, and TSFC deterioration.

The ultimate goal in the diagnostics effort is to provide adequate data for taking actions toward eliminating performance deterioration. Much of the required data has been generated in the early tasks. The flight loads portion of the JT9D Diagnostics Program in which engine clearance changes are measured in flight and the concurrent NAIL flight loads program will complete the data. The task ahead is the application of this data and the appropriate use of design too 1 sin concerted efforts between the engi ne and ai rframe manufacturers to reduce performance deterioration. ·A considerable effort has evolved in the area of integrated engine-nacelle design studies aimed at stiffening current power plant installations. More important is the use of the diagnostics data in systems currently under design and development that recognize and build deterioration prevention into the initial designs.

DE~GN LOAD ~LOSOPHY EARLY TURBOJET INSTALLATION FIGURE 2 ~ - ~"~ ~--~--------------- MODERN ENGINE INSTALLATION FIGUREJ RELATION OF AIRPLANE FLIGHT LOAD EXCEEDANCES TO BLADE TIP CLEARANCE IIBU.DE TIP CLEARANCE· 6. • II "NACELLE ~LEARANCI XCEEDANC CO "nCO ~A"NACELLI II TIP CLEARANCE FIGURE 4

ACCEPTANCE FUGHT PROFILE

CRUIS E ~ MAXIMUM MAOi NO .

SHUTDOWN/RESTART ALTITUDE Toum AND G LAND, REVERSE TIME THRUST FIGURE 5 ILLUSTRATIVE VGH RECORDING fliGHT PHASE -- CLIMB ENROLm DESCENT ACCEL.

TYPE -- TAXI AND TAK E -o FF FIGURE 6 ANALYSIS OF LOAD FACTOR TIME HISTORY FROM VGH RECORDER EXCEEOANCES • OCCURRENCES t OCCURRENC6S 0.4 0.3 I I I J 4 0.2 3 POSI .TlVE A -:A- • 6 8 0.1

t

8 16 ton

A 1 \I 717 I" I lW--Wt\

\I I v -11 ,1\ I rY ' e 15

-0 .1 6 8 -V 1/ II . - - 0.2 3 4

l NEGATIVE

1 1 -0 .3 Y " -----, , -0 .4 I UNIT TIME I • . , ., I HOURI FIGURE 7 HISTOGRAM AND PlOT OF OCCURRANCES OCCURRENCES PER UNI T TIME - 0.4 - 0.2 0 0,2 0.4 6n FIGURE 8

NACELLE LOAD EXCEEDANCE DATA

3 ~cr~ 10- l---I- -- -f--I----- - +-+ -- -->r -- 1000 Fli GHTS 10 -4 L.:..._.L---:~~...J..._~ _ ___i......;.---'~_...I..__:...J -3. 0 - 2. 0 -1.0 0 1.0 2. 0 3. 0 4. 0 5. 0 LOAD FACTOR I G'S I FIGURE 9 AI RLOADS TAKEOFF ROTATION

"" , L __ --Ll--

" "

'- '-. '\.

"' "

'- "-

CI

t2 ......... "-

-.... "- RESULTANT ---....

LOADS PR ESSURE DISTR IBUTI ON FI GURE 10 18 4 AIRLOADS MAXIMUM DYNAMIC PRESSURE ./ / / /

V£/

I '- \ \ \ \ '\ \

"-. L--------'-"'-

"

\

Cl

'" \

RESULTANT LOADS

.......... ""

-............... '"

~

--- ___ .......... PRESSURE - DISTRIBUTION FI GURE Jl AIRLOADS STALL WARNING, 10° FLAPS /"(~

/'.. " \

(1

/-.- ....... " '\

--::: ......... " RESULTANT

......

LOADS PRESSURE DISTRIBUTION FIGURE 12 INLET CIRCUMFERENTIAL PRESSURE DISTRIBUTION --+- OUTBOARD FIGURE 13 PROPULSION SYSTEM SUBSTRUCnJRES AND RESPONSIBILITIES AIRFRAME r-------------------~ ~ I THRUn YOK! ~ . '

.~~~~;;-. I

4~1 r~~:

'>~ ~'

'. ~ ~TURUINe I . I =<J""~LO'\ HPC 1IIIIIIUnc ,,,eeooYI I 3 .. OYH"UlC 'II£EOO ...

I 3 .. fREQUENCIES

'- I 60 MODE 8HAPEII

I I "evIRII"I, COWLING, lTe 101\ : INCLUDED IN ANAlY".

I." I\Q I

1.. ________ ~ .. .2I~E _______ .J FIGURE 14 FINITE ELEMENT MODEL FI GURE 15 STRUCnJRAL DEFLECTION ORIGINAl\ --'"----,

--

-~~

----

"--- -- - -

DEFlECTED/ FIGURE 16

CALCULATED CLEARANCE CHANGES (INCHES)

LPT FAN TOP a U) w . 030 -;080 w a: -. 060 C!l .100 I-' w -. 0 20 .020-l1 00 a .020 .020 .000 .050 .050 .060 ~

,n" i ,niO

.000 .000 -.005 .000 .:::: . 000 .040

~

90 -i -.050 ..J

.. 000 -; a I 0 ~m: -; 0 I a

-;050 ..J . 000 . 0001~ '::: -; 020

« . 020

-.100 .005 . 005 ~ 120 .025 :020 w ~ IOO . 010 a: .000 .000 w - .1 SO ~ ISO .020 .005 .030 :::l U a: -.1 SO .025 U 180 0IS -; OJU -;020 .005· .005 BOTTOM r ~:~~:~~{ 1 DENOTES CLOSURE .. .. .. ...... .

FI GURE 17 CONSERVATION OF STRATEGIC AEROSPACE MATERIALS (COSAM) Joseph R. Stephens NASA Lewis Research Center SUMMARY NASA has undertaken several projects directed at conserving strategic materials used in the aerospace industry. Research efforts involving univer- sities and industry as well as in-house activities at the NASA Lewis Research Center comprise the current "Conservation of Strate g ic Aerospace Materials" COSAM effort. The primary objective of COSAM is to help reduce the dependence of the United States aerospace industr y on strategic metals, such as cobalt (Co), columbium (Cb), tantalum (Ta) , and chromium (Cr ) , by providing the mate- rials technology needed to minimize the strategic metal content of critical aerospace components for gas turbine engines. Thrusts in three technology areas are appropriate for COSAM. These include near-term activities in the area of strategic element substitution; intermediate-range activities in the area of materials processing; and long-term, high-risk activities in the area of "new classes" of high temperature metallic materials. This paper describes in some detail th e efforts currently underway and the initial results gener- ated to date. In i tial emphasis has been placed in the area of strategic ele- ment substitution . Specifically, the role of cobalt in nickel-base and cobalt- base superalloys vital to the aerospace industr y is being examined in great detail by means of cooperative university-industry-government research efforts.

Investigations are also underway in the area of "new classes" of alloys.

Specificall y, a study. has been undertaken to investigate the mechanical and physical properti es of intermetallics that will contain a minimum of the stra- tegic metals. Current plans for COSAM are presented in this paper also.

INTRODUCTION The United States relies heavilY upon foreign sources for the supply of most strategic metals required by our aerospace industry. With the exception of molybdenum, iron, magnesium, and the rarp. earths, the United States imports from 50 to 100 percent of such aerospace metals as Co, Cb, Ta, Cr, And Mn (ref 1). However, the potential for foreign cartels, political unrest, and production limitation is great and is intensified by steadily declining known reserves. Thus, the United States can expect to be faced with supply shortages and price escalation for many strategic metals. Since these metals are vital to the welfare of the nation'S economy, their continued availability at a rea- sonable cost is a national issue which requires cooperative action between the aerospace industry and appropriate government agencies.

The aerospace industry is currently a major factor in the posltlve inflow of funds from U. S. exports (ref. 2). This industry, and within it the air- craft engine industry in particular, relies heavily upon imports for several key strategic metals including cobalt, columbium, tantalum, and chromium. In order to offset or minimize future disruptions in supply, efforts to develop viable options must begin now, since a new material can take from 5 to 10 years of research and development efforts before qualifying for aerospace service.

NASA currently has plans to address the aerospace industry's needs to minimize the use of strategic metals for advanced aerospace systems. COSAM has as its broad objective the reduction of the dependence of the u.S. aero- space industry on strategic metals. This objective can be accomplished by providing the materials technology options needed to allow individual com- panies to trade-off the material properties of critical components versus cos t and availability of their strategic metal content. This paper summarizes NASA's current activities in this area and broadly outlines the plans for COSAM.

STRATEGIC MATERIALS A definition of strategic materials as used in this paper is given in figure 1. Strategic materials are those predominantly or wholly imported elements contained in the metallic alloys used in aerospace components which are essential to the strategic economical health of the U.S . aerospace indus- try . As the basis for what are considered strategic metals, we will focus on the aircraft engine industry's needs. Based on a survey of the ASHE Gas Turbine Panel and a subsequent survey of a number of aerospace companies, the elements listed in figure I were considered to be the most strategic with respect to the aerospace industry. As a result of prioritizing by NASA's COSAM planning team supplemented by further discussions with several aircraft engine manufacturers, four elements emerged that were of particular concern.

The alloys used to build the critical high temperature components for aircraft propulsion systems require the use of the four metals - cobalt, columbium, tantalum, and chromi um. These metals are contained in steels, stainless steel~ and superalloys that are used in engine manufacturing. Figure 2 lists these four ele ments in the high priority category with a brief rationale for th is ranking. The remaining five strategic elements evolving from our surveys were given a lower priority and figure 2 also contains a short explanation for this ranking.

The location of these metals in aircraft engine compressors, turbines, and combustors is illustrated in figure 3. The need for such metals has i n- creased as the demands have grown for higher durability plus high performance, fuel efficient aircraft turbine engines. Based on the essential nature of these metals and for the u.s. aircraft industry to maintain its competitive position, it is necessary that supplies be readily available at a reasonably stable cost. To achieve these requirements, domestic sources of key metals are desirable .

Today, we are almost totally dependent on foreign sources for these metals as shown in figure 4. In several of the countries listed in figure 4, recent political disturbances have led to supply interruptions. Therefore, the U.S.

aircraft engine industry can be seen to be highly vulnerable to supply insta- bilities of the essential metals for engine manufacturing. Accompanying supply disruptions or increased demand is an accelerated price increase. Escalated prices during the recent few years are evident for tantalum, columbium, cobalt, and to a lesser degree for chromium, as shown in figure 5. These rapid price increases illustrate the additional vulnerability of the U.S. aircraft engine industry to cost fluctuations. The essential nature of cobalt, columbium, tantalum, and chromium along with their vulnerability to supply instabilities and cost fluctuations combine to cause these metals to be classified as strate- gic aerospace metals.

The portion of these four metals used in superalloys for the aerospace industry compared to all other U.S. uses is shown in figures 6 to 9. The 'he use of these metals in superalloys as compared to total U.S. consumption in 1979 was: cobalt - 30 percent, columbium - 28 percent, tantalum - 5 percent, and chromium - 3.4 percent. These data reveal that superalloys comprise the largest single use of both cobalt and columbium.

OVERVIEW OF COSAM COSAM has as its primary objective the reduction of the dependence of the U.S. aerospace industry on strategic metals. COSAM can also provide the indus- try with some options for making their own property versus availability/cost trade-offs when selecting aerospace alloys. These objectives will be achieved by providing the technology needed to minimize the strategic metal content of critical components in aerospace structures. Initial emphasis will be placed on the aircraft engine industry. COSAM initially is focused on conservation of the strategic metals cobalt, columbium, tantalum, and chromium. Strategic metals such as titanium, the precious metals, tungsten, and others may be brought into COSAM as it progresses.

Along with prioritizing the strategic elements that were identified, the role that the NASA's COSAM effort should encompass was also evaluated. Options that were considered are listed in figure 10. All of these options could con- tribute to the conservation of strategic materials and minimization of U.S.

aerospace industry vulnerability. However, within the scope of our program a decision was reached based on Lewis' traditional roles and expertise to focus on the three areas noted in figure 10. These areas consist of strategic element substitution, process technology, and alternate materials. Contribu- tions to the other areas may benefit from COSAM through cooperative programs with other governmental agencies such as in the area of scrap reclamation or through cooperation with technical societies in establishing a critical material index. Having selected a list of four high priority strategic ele- ments and having defined the areas of emphasis for COSAM and specific objec- tives, a technology approach was adopted as shown in figure 11. Conservation, as well as reduced dependence on strategic metals, will be achieved in the area of strategic element substitution by systematically examlnlng the effects of replacing cobalt, columbium, and tantalum with less strategic elements in current, high use engine alloys . This will help guide future material speci- fications if one or more of these metals becomes in short supply. Conservation through process technology will be achieved by advancements in those net-shape and tailored-structure processes that minimize strategic material input re- quirements. This will lower total usage. And in the longer term, development of alternate materials that replace most strategic metals with those highly available in the u.s. could lead to a substantial reduction in the U.S. depen- dence on foreign sources. Both of the later two technology areas will help conserve the four strategic metals Co, Cb, Ta, and Cr.

EARLY COSAM ACTIVITIES COSAM efforts began in FY'80. Efforts on planning and organlzlng are still underway. In addition to the planning activities, several small re- search activities have been initiated. These research activities focus on two of the three major thrusts of COSAM - strategic element substitution and devel- opment of alternate materials. Special emphasis of these initial efforts is on developing a fundamental understanding of the role of strategic elements in current aircraft engine alloys so that effective alloying element substitution can be conducted. Similarly, in the development of alternate materials, a basic understanding of materials properties and alloying concepts is being emphasized. Consequently, university grants playa major part in COSAM. In addition, cooperative programs with industry augmented by in-house reseahch at the NASA Lewis Research Center comprise the approach used in these initial projects. This cooperative approach will continue to be followed in COSAM and industry, universities, and government in-house research will each playa key role. The subsequent paragraphs will describe in some detail early COSAM research efforts.

Strategic Element Substitution Four metals were mentioned previously as being classified as high priority strategic metals. Cobalt was selected from these four metals for the early COSAM strategic element substitution research. The basis for selecting cobalt was twofold. First, the largest single use of cobalt in the U.S. is in superalloys for jet engine applications as was shown in figure 9 (ref. 3) .

Many of the other applications indicated in figure 9 are also important to the nation's economy and security as well. Secondly, the specific roles that cobalt plays in nickel-base superalloy fabrication and performance has not been clearly established. Most superalloys currently in use were developed at a time when cobalt was plentiful and inexpensive. Literature results (Ref. 4) are conflicting as to the role that cobalt plays in nickel-base superalloys in important areas such as phase stability, y ' partitioning, strength, fabricabil- ity, and oxidation and hot corrosion resistance. Because of these uncertain- ties, there exists a strong possibility that the strategic element cobalt can be substantially reduced or possibly eliminated from several superalloys without sacrifice of the key properties for which these alloys were selected for engine service.

Four nickel-base and one cobalt-base superalloys were selected for this investigation. The five alloys are listed in figure 12 along with the i r typical applications in the aircraft engine industry, the forms in which the alloys are used, and remarks as to why they were selected for this activity.

Applications include turbine disks, turbine blades, and combustors. A variety of product forms are represented by the applications of the five alloys as noted in figure 12 . The selection of the five alloys was based primarily upon the considerations given in this figure. Waspaloy* was selected because it represents the highest tonnage of cobalt now in commercial aircraft engines.

Selection of Udimet-700* was based on the fact that this alloy is used in the as-cast, as - wrought ingot, as-wrought powder, and as-HIP powder metallurgy fabricated conditions. The potential for determining the impact of cobalt on both conventionally cast as well as on single crystal turbine blades was the reason for ' selecting Mar-M247*. Rene' 150* was chosen because it is one of the most advanced directionally solidified alloys. The wrought, sheet alloy HA-188* was selected because it represents one of the largest uses of a cobalt- base alloy in aircraft engines.

The primary purpose of the cobalt strategic element substitution research is to determine the fundamental role of cobalt in a wide variety of nickel- base superalloys and in a high-use cobalt-base superalloy. A secondary purpose is to develop the methodology to explore the roles of other strategic elements in similarly chosen alloys so as to have maximum impact on a wide range of users.

Figure 13 shows the participants in this COSAM effort on cobalt strategic element substitution. These initial research efforts are planned for a three- year period and consist of cooperative programs involving universities, indus- try, and NASA Lewis Research Center. Nominal compositions of the five alloys given in figure 13 indicate that cobalt content ranges from 10 percent in Mar- M247 to 39 percent in HA-188. In addition, the y ' phase ranges from 20 per- cent in Waspaloy to 65 percent in Rene' 150. The first phase in each research effort will involve substituting the less strategic element, nickel, for co balt in incremental steps to a zero cobalt content. The effect of this sub- stitution on properties and phases present, such as y ', will make up the major portion of the research effort in the first year of each program element.

Efforts in subsequent years will be directed at identifying and optimizing alloying elements as substitutes for cobalt in the five alloys so as to main- tain the key properties of these alloys.

The cooperative nature of the research being conducted on Waspaloy and Udimet-700 is illustrated in figure 14. The role of industry as represented by Special Metals Corporation is outlined. Their primary role is to charac- terize and optimize fabrication and heat treating procedures for the reduced *Trademarks Waspaloy United Technologies Corporation Udimet Special Metals Corporation Mar-M Martin Marietta Corporation Rene' General Electric Corporation HA Cabot C o ~p o ration cobalt Waspaloy and Udimet-700 alloys. The university role in this effort is . also shown in figure 14. Columbia University will be involved with mechanical property characterization, structural stability, microstructur al features, and theoretical formulations to identify future alloy modifications if required for the second phase of the project. Purdue University will be primarily responsible for microstructural and microchemistry characterization of the reduced cobalt content alloys. To round out the program, NASA Lewis Research Center will be involved in further mechanical and physical metallurgy charac- terization of the alloys as shown in figure 14. The output of this coopera- tive effort is expected to be a clearer understanding of the role of cobalt in nickel-base superalloys.

Some preliminary results on th~ effects of reducing cobalt in Waspaloy, a 13 percent cobalt alloy, are shown in figure IS (ref. 5). Tensile strength appears to be insensitive to the amounts of cobalt in the alloy. However, rupture life decreased with decreasing amount of cobalt in Waspaloy. Further testing will be required to better characterize this apparent effect.

Similar effects of cobalt on the rupture life of Mar-M247 have been deter- mined as shown in figure 16 (ref. 6). A possible contributing cause to this reduction in rupture life is the decrease in amount of y ' in this alloy with decreasing cobalt content as shown in figure 17. Also shown in figure 17 is the change in y ' composition. As cobalt is removed from the alloy, the largest change in the composition of y ' is the increase in tungsten content.

Further studies are underway to clarify the role of cobalt in this alloy.

The research efforts on Udimet-700 and Rene' ISO parallel the previously described efforts on Waspaloy and Mar-M247. It is anticipated that these studies will lead to an understanding of the fundamental role of cobalt in a variety of conventional and directional nickel-base superalloys. These results should provide an improved technical base to develop modified super- alloys in later stages of COSAM, as illustrated in figure 18.

Alternate Materials Research in this area must be considered to be high risk and long range, but it has the potential of a high payoff in terms of significantly reducing the nation's dependence on strategic materials. As an example of alternate materials, intermetallic compounds are currently being investigated for possible structural applications. Initial efforts are centered on nickel and iron aluminides. Successful development of this type of alternate nlaterial offers the possibility of partially or totally replacing all the strategic materials in components where intermetallic compounds can be utilized.

Intermetallic compounds are of interest because of their potential high temperature strength as shown in figure 19 (ref. 7). It can be seen in this figure that nickel aluminides have the strength capability of competing with current nickel-base alloys. However, a possible disadvantage of this type of material is that simple binary aluminide compounds have shown a lack of room temperature ductility (fig. 20). The factors which influence the high ductile- to-brittle transition temJ?erature of nickel aluminide C~600 + °C) are currently being investigated. A NASA grant with Dartmouth University is aimed at understanding the fundamental deformation mechanisms in nickel aluminide. From these inves · tigations, methods of improving the low temper- ature ductility of nickel aluminide may be suggested. An accompanying in- house research project at NASA Lewis Research Center is focusing on the high temperature mechanical properties of aluminides. These studies can provide a fundamental basis for more extensive research to develop these nonstrategic, alternate materials as shown in figure 21.

COSAM PLANS Future COSAM efforts can build on the fundamental understanding from the early res-earch for cobalt substitution, as was- shown in figure 18. Major efforts will be devoted to developing, and if warranted, to scaling-up low or no-cobalt nickel-base superalloys for fabrication into various components.

Demonstration of continued promise could also lead to verification in engine tests by major engine producers. Similar efforts will also be conducted for other strategic metals such as columbium and tantalum .

In the area of alternate materials, much more work will be required to develop materials such as intermetallic compounds. As was shown in figure 21, initial efforts will focus on fundamental studies aimed at improving low temperature ducitility and high temperature strength of FeAl and NiAl inter- metallics. Complete property characterization will follow on more promising compositions. Reiterations of these basic steps will be required to further optimize the alternate materials and make them viable candidates as structural materials for aircraft engines. Scale-up and rig testing of promising compo- sitions for blades and vanes will follow. The development of alternate materials will help conserve the strategic metals Co, Cb, Ta, and Cr.

A third area of the COSAM consideration involves conservation through improved materials processing technology. Although none of these activities have been initiated, plans have been made for investigating processing technology in such areas as advanced melting techniques, tailored fabrication, advanced coatings, joining techniques, and fabrication efficient processes. A reduction in strategic material usage should result from these processing technologies. For example, early efforts on near-net-shape fabrication of a turBine disk (ref. 8) have been S' hown to be able to reduce input material weight compared to conventional casting/forging practice and further gains appear possible. Improved processing technology will also help conserve the s· trategic metals Co, Cb, Ta, and Cr.

CONCLUDING REMARKS This paper has presented NASA's COSAM efforts and planning. The primary points are summarized below: 1. Advancements in materials ' technologies are needed to provide the aerospace industry with alternative materials options in the event of future strategic metal shortages or excessive price increases.

2 . The primary role of NASA's COSAM efforts will be to address strategic material problems within the aerospace industry. COSAM should make contribu- tions to a national data base that will benefit many other domestic industries as well.

3. COSAM was designed to involve cooperative research efforts with industry (alloy producers, component fabricators, and engine manufacturers), with universities, and with government research facilities (primarily the Lewis Research Center) .

REFERENCES 1. Mineral Co mm od ity Summaries 1979. Bureau of Mines, U.S. Department of the In t eri o r, 1979 2. "Aerospace Review and Forecast 1979/80", Aerospace, vol. 18, no .1, Winter, 1980, pp. 2- 4.

3. Mineral Industry Surveys. Bureau of Mines, U.S. Department of the Interior, Dec. 1979 .

4. Tien, J. K.; et al.: Cobalt Availability and Superalloys. J. Metals, vol. 32, no. 10, 1980, pp. 12-20.

5. Maurer, G. E.; Jackman, L. A.; and Domingue, J. A.: Role of Cobalt in Waspaloy. Superalloys 1980, J. K. Tien, ed., American Society for Me t als, 1980, pp . 43-52.

6 . Nathal, M.: The ~le of Cobalt in a Nickel Base Superalloy. Presented at the AIME 110 Annual Meeting, Chicago, II., Feb. 22-26, 1981.

7 . Schulson, E. M.: The Ductility of Polycrystalline NiAl. Oral Presentation at NASA Le w is R esearch Center, Sept. 1979.

8 .

Dreshfie1d, R. L. ; and Miner, R. V., Jr.; Application of Superalloy Powder Metallurgy for Aircraft Engines. NASA TM-81466, 1980.

STRATEGIC MATERIALS • DEFINITION: THOSE PREDOMINANTLY OR WHOLLY IMPORTED ELEMENTS CONTAINED IN THE fl£TALUC ALLOYS USED IN AEROSPACE COMPONENTS WHICH ARE ESSENTIAL TO THE STRATEGIC ECONOMIC HEALTH OF THE U. S.

AEROSPACE INDUSTRY • SURVEY RESULTS A. S. M. E. GAS TURBINE PANEL SURVEY STRATEGIC ELEMENTS IDENTIFIED Cb, Co, Cr, Ta, AND W NASA AEROSPACE COMPANY SURVEY (ADDITIONAL ELEMENTS) Mn, Pd, P~ Sn Figure 1. - COSAM background.

HIGH PRIORITY Co SUPERALLOYS LARGEST SINGLE USER (3m> OF TOTAl)

Cb SUPERALLOYS LARGEST SINGLE USER (m OF TOTAl)

Cr CRITICAL TO ENVIRONW£NTAL RESISTANCE OF ENGINE COMPONENTS Ta CRITICAL TO ADVANCED ENGINE ALLOYS LOWER PRIORITY W r£w U. S. MINES ON STREAM, PRO.£CTED SELF-SUFFICIENCY BY 1985 Mn WIDEL Y USED IN STEEL INDUSTRY, HOWEVER POTENTIAL LOW COST ALTERNATIVE TO Ni Pd, Pt USED FOR ELECTRICAliHECTRONIC APPLICATIONS Sn USED FOR AI AIRFRAME ALLOYS AND IN SOME Ti ENGINE ALLOYS Figure 2. - Strategic element focus.

CHROMIUM CHROMI UM COBALT COBALT CHROMIUM COLUMBIUM COLUMBIUM TANTALUM COBALT NEEDED FOR PERFORMANCE AND LONG LIFE COBALT - HIGH TEMPERATURE STRENGTHENER COLUMBIUM - INTERMEDIATE TEMPERATURE STRENGTHENER TANTALUM - OX IDATION RESISTANCE CHROMIUM - CORROSION RESISTANCE Figure 3. - Strategic metals are critical to turbine engi n es.

% IMPORTED MAJOR FOREIGN SOURCE METAL COBALT 97 ZAIRE COLUMBIUM 1 00 BRAZIL TANTALUM 97 THAILAND CHROMIUM 91 SOUTH AFRICA, ZIMBABWE Fig u re 4. - U. S. aerospace is vulnerable to supply in st abili t ies .

19 8 $150 TANTALUM O J PRODUCER PRICE, H) COLUMBIUM $lIb 1972 1973 1974 1975 1976 1977 1978 1979 1980 Figure 5. - U. S. aerospace is vulnerable to cost fluctuations.

MAGNETIC

MATERIALS

2~ Figure 6. - U. S. consumption of cobalt in 1979 (Total pounds, 20.3 million).

CARBON STEELS 2JJ.

Figure 7. - u. S. consumption of columbium in 1979 (Total pounds, 6.3 million).

CUTIING TOOLS li"t CAPACITORS ~ Figure 8. - U. S. consumption of tantalum ln 1979 (Total pounds, 1.7 million).

STAINLESS STEELS 7fJJo -- OTHER - - CARBON - - HSLA - - - SUPERALLOYS - --- TOOL STEE1 ~ , ' CAST IRON Figure 9. - U. S . consumption of chromi um i n 1979 (Total p ou nds ch rom i um ferroalloys, 1.OxlO ).

OPTIONS CONSIDERED IN PREPARA T ION FOR STRATEGIC MATERIALS SHORTAGE • EXPAND EXPLORATION • IMPROVED RECOVERY • STRATEGIC MATERIAL SUBSTITUTION ,; • SCRAP RECLAMATION • REDUCED WASTE IN PROCESSING ,; • AlTERNATE MATERIALS ,; • STOCKPILING • CRITICAL MATERIAL INDEX F ig ur e 10.- Options consid e red in pre pa r a t i on for strategic materials short - age).

OBJECTIVE: • PROVIDE TECHNOLOGY OPTIONS WHICH WILL SUPPORT THE AEROSPACE IND . USTRY IN MAKING STRATEGIC ECONOMIC DECISIONS AIft£D AT SIGNIFICANTLY REDUCING S TRA TEGIC METAL CONS UMPTION - Co, Cb, Ta, Cr, AND OTHERS AS IDENTIFIED APPROACH: • DEVELOP UNDERSTANDING OF ROLES OF Co, Cb, Ta, AND Cr IN CURRENT S UPERALLOYS • IDENTIFY SUBSTITUTES AND LOW STRATEGIC ft£TAL CONTENT ALLOYS • DEVELOP PROCESS TECHNOLOGY THAT WILL MINIMIZE STRATEGIC METAL INPUT AND WASTE • IDENTIFY ALTERNATE MATERIALS AND PROCESSES THAT HAVE HIGH LONG TERM POTENTIAL IN REDUCING STRATEGIC ft£TAL USAGE (HIGHER RISK APPROACH) Figure 11. - COSAM program objective and approach.

REMARKS ALLOY TYPICAL ENGINE FORM APPLICATION FORGED HIGHEST USE TURBINE DISK WASPALOY WROUGHT ALLOY IN CURRENT ENGINES SIMILAR ALLOYS USED UDIMET-700 TURBINE DISK fORGED } AS-HIP- TURBINE DISK IN VARIOUS FORMS (lCI ASTROLOY POWDER AND APPLICA nONS , (RENE 771 LP BLADES CAST MAR-M247 TURBINE CAST CONVENTIONALL Y-CAST, O . S.

BLADES AND SINGLE CRYSTAL , RENE 150 TURBINE OS-CAST HIGHLY COMPLEX BLADES DIRECTlONAll Y-CAST ALLOY HA-188 COMBUSTORS WROUGHT HIGH USE COBALT-BASE SHEET ALLOY Figure 12. - Superalloys selected for initial COSAM activities.

NOMINAL COMPOS ITION PARTICIPANTS ALLOY y' Ni Cr Cc Mo W Ta Re AI Ti HI COLUMBIA UNIV WASPALOY 58 70 13 4 -- 1. 3 "3 20% PURDUE UNIV SPECIAL METALS NASA-LEWIS - - - - COLUMBIA UNIV UDIMET-700 53 15 19 5 -- 4. 3 3. 5 -- 40% PURDUE UNIV SPECIAL METALS NASA- LE WIS CASE-WESTERN MAR-M247 8 -- 5.5 60 10 .6 10 3 1 1. 4 55% RESERVE UNIV TELEDYNE NASA -L EWIS I NASA-LEWIS RENE ISO 59 5 12 5 6 -- L5 3 5.5 65 % (TBD) HA-l88 22 22 39 14 Figure 13. - Elements of initial COSAM a ct ivities .

MATERIAL PHYSICAL ft£TALLURGY AND EXPE C TED PROP£RTY CHARACTERIZATION RESULTS FABRICATION OLUMBIA UNIVERSITY PURDUE UNIVERSITY $P£CIAL METAL CLEARER UNDERSTANDING • '( AND CARBIDE EXTRACTIONS OF ROLE OF COBALT • y AND '( CHEMICAL ANALYSIS IN NICKEL-BASE • X-RAY DIFFRACTION SUP£RALLOYS [L£CTRON MICROSCOPY NASA - LERC • LOW CYCLE FA TIGUE • THERMAL FATIGUE • CYCLIC OXIDATION • HOT CORROSION Figure 14 . - Coo perat ive program to determine fundamental role of cobalt in Waspaloy and U -700.

TENSILE STRENGTH, 535 C RUPTURE LIFE, 7'IfJ C; 550 MPa " COBALT

I

I

(WASPALOY)

I

I I I I I 1.1 l2 0 25 50 75 1.0 LIFE, hr STRENGTH, GPa Figure 15. - Preliminary results of reducing cobalt in Waspaloy .

o

J % COBALT

I

(MAR-M247l 10

I

I

I

I

o

100 200 300 LIFE, hr Figure 16. - Preliminary results of reducing cobalt in Mar-M247.

(Rupture life, 870°C; 360 MPa.)

70 t="-------Ni ATOMIC 12 PERCENT 10 ~-------AI ELEMENT WEK;HT 40 PERCENT GAMMA 20 PRIME

o 5 10 o 5 10

WEK;HT" COBALT WEK;HT " COBALT Figure 18. - Planned flow of COSAM strategic element substitution research.

FY8D FY81 FY82 FY83 FY84 FY85 FY86 UNDERSTANDING ROLE Of COBALT IN SUPERALLOYS IDENTIFY ALLOY DEVELOPMENT APPROACH ALLOY CHARACTERIZA TlON COMPONENT FABRICATION ENGINE VERIFICA liON : E. G •• LOW-COBALVHIGH STRENGTH DISK LOW-COBALT/HIGH PERFORMANCE BLADE (ENGINE MANUFACTURERS) Figure 17. - Amount and composition of gamma prime as a function of cobalt content in Mar-M247.

J

MAR-M246 1000 h RUPTURE LIFE STRESS, MPa ~ "-

"

"- ...

.... Ni2AlTi/NiAI o~--~ -- ~~~=-~ 700 800 900 1000 1100

TEMPERATURE, °c

Figure 19, - Typical strengths of alurninides and superalloys.

TENSILE ELONGATION, % 20 600 700 800 900 1000 TEMPERATURE, DC Figure 20. - Typical ductility values for alurninides and superalloys.

FY80 FY81 FY82 FY83 FY84 FY85 FY86 UNDERSTANDING DUCTILITY AND STRENGTHENING MECHANISMS IN FeAI AND NiAI INTERMETAlLl CS PROPERTY CHA RACTERIZA TlON MECHANICAL - PHYS ICAl- ENVIRONMENTAl IDENTIFY AND SCALE-UP INTERMETALLICS FOR BLADES AND VANES RIG TESTING Figure 21. - Planned flow of COS AM alternate materials research.

Page intentionally left blank

WHO NEEDS ENGINE MONITORING?

James L. Pettigrew, P.E., Lt Col, USAF Wright-Patterson Air Force Base ABSTRACT The requirement for Engine Monitoring Systems (EMS) is elusive even for its advocates. Decisions not to invest large sums of up front money in equip- ment which will be of uncertain value are easily made by conscientious program managers. Even as on-condition maintenance (OeM) is being established as the desired approach in the Air Force, many people in the decision chain doubt the potential value of on-board engine monitoring equipment.

EMS advocates have not provided convincing answers to many hard questions, some of which are: "Should the EMS capabil ity provide on-board GO- N O-GO infor- mation? How much engine monitoring is enough? What parameters are required?

How will the EMS capabil ity be used to direct maintenance actions? Does the environment require only engine parts tracking, engine usage, or performance trending data?" Answers may not have uncontested technical support but may require judgement based on something 1 ike Pareto's 80-20 law appl ied to operational data.

The true EMS values are certain only in the future operational environments.

The EMS advocates I problems are to find for the system managers acceptable up front rational ization for the added EMS cost. Past operational evaluations of a few EMS units for short periods have not all produced convincing results.

This presentation will discuss these evaluations and their lessons learned, then review the options for each required EMS phase, and close w ith a review of the guidance being provided for EMS on new systems.

COST TRENDS Table I shows the maintenance cost of flying various Air Force engines for thousand engine fl ight hours. The maintenance cost are in 1980 dollars.

In most cases, these costs equal or exceed the acquisition cost for that engine. The acquisition costs shown are first production contract costs in then year dollars. For a true comparison, the earl ier engine cost would be corrected for inflation. From these numbers, the throwaway engine might not be such a bad concept, especially when you remember that with increasing engine age performance deteriorates and engine service 1 ife between repair shortens. Maintenance manhour per fl ight hour on the newer engines is also increasing to some very high numbers.

FAILURES Failures result from wear, leaks, structural damage and human error.

There are many things which influence operating ti m e before engine failure.

An Engine M onitoring System (EMS) provides a data base from which failures can be predicted, detected, and diagnosed early, before there is a loss in mission capabil ity.

IMPORTANCE OF VIEWPOINT The viewpoint from which an individual looks on things has a large influence on what he is able to see. This is illustrated by the old saying that "A jackass on a hill can see more than a genius in a valley."

An EMS is more than black boxes full of electronic circuits . The peop l e who look at an engine monitoring system as the blac k boxes might be considered the genius in the valley. In the total system view, EMS is the executive control system w hich tells the maintenance supervisor that an individual engine requires diagnostic work to find out why it is abnormal. The EMS data function is similar to the blood pressure check performed by the doctor . If he finds any abnormal ities in blood pressure he runs other diagnostic test s to determine what is causing you to be abnormal.

REQU I RED TASKS Figure I illustrates the data flow in an EMS. Data can be obtained in ways ranging from a manual recording of cockpit instrument readings to sophisticated complete electronic systems which automatically records, stores, and transfers the data to ground computers. Airborn engine monitoring system electronics often have decision logic to determine engine status as soon as the aircraft lands. Airplanes with two pilots and mission requirements for a cruise leg are generally able to use manual recording. On single pilot air- craft work load generally prevents the use of manual recording. The abil ity to get in-fl ight engine performance is the missing piece for single pilot fighter aircraft. Therefore, current thrust in developing EMS capabil ity is improvement of in-fl ight data acquisition abil ity.

Before the in-fl ight data can be used to predict, detect and diagnose failures, it must be val idated, corrected, compressed, displayed, and then interpreted. There are a number of ways of interpreting . The status of engines can be obtained from the data by limit exceedence or by observing trends. The important results from an EMS is the effect of the information on the maintenance system. If we only gather the data, and look at the data, and do not use it to direct maintenance, EMS is of 1 ittle value to the total system.

WHY OCM In February 1974, the Department of Defense gave the following logistics and material support guidance: I. Establ ish engine maintenance pol icies to el iminate maximum operating time.

2. Exploit modular designs in new engines.

3. Use on-condition maintenance techniques.

4. Apply to existing engine types wherever practicable.

With the on-condition maintenance you need a methodology to tell you what the existing condition is within the engine so you can schedule it for maintenance. Figure 2 illustrates why on-condition maintenance has an advantage. It can reduce risk and save dollars.

Engine usage varies by the mission being flown. For example, a fighter aircraft on a low-level mission flying at 600 knots, t ~ach .95, would have its inlet pressure increased by a factor of 1.8. On the low-level mission, the engine with a 20 to 1 compression ratio would have a combustor case pressure of 36 atmospheres. On the other hand, an intercept mission which cruises out at 30,000 feet MSL would only see 12 atmospheres combustor case pressure. The cruise engine obviously is capable of operating more hours before failure because of the less stressful usage. If maintenance is driven by maximum operating time, the additional operating capabil ity of an engine used at the lower rate will not be util ized. If the condition of each engine determines when it must be repaired, then the full engine capabil ity can safely be used.

AIR FORCE EMS PROGRAMS Air Force EMS programs are divided into three categories: (1) developed with aircraft, (2) contract maintenance, and (3) add-ons to operational aircraft.

See Table I I for a 1 isting of EMS developed with aircraft.

M RS is a Maintenance Recording System that is appl ied on the SR7l with a J58 engine. It is an analog recorder that gives a continuous trace of the engine operating parameters throughout that mission. It has an approximately 1100 hours meantime between failure (MTBF), and is considered a successful EMS system. Its data is automatically acquired and formatted with manual interpretation by a technician roll ing the strip chart and looking at total trace for each sortie. That strip chart may be 8 to 10 feet long for a sortie. The interpretation of the analog traces is a disadvantage on this system.

The Malfunction Detection Analysis and Recording System (MADARS) was built and developed with the C-5 aircraft. It automatically acquires and formats the data. Interpretation is both manual and automatic. L.o9is is proved to print out the maintenance action required in many cases. A shortcoming of MADARS is an overall system MTBF. MADARS monitors all aircraft systems. The MTBF for the engine portion of the MADARS which provides engine data is approximately 100 hours.

The Central Integrated Test System (CITS) is another system designed to monitor the total airplane as well as the engine; it has been tested on the four B-l1s during their Category I & II fl ight tests. It is rather complex and there are some differences of opinion on its real potential benefit to the operational weapons system.

The two systems at the bottom of the figure, Events History Recorder (EHR) and Engine Time Temperature Record (ETTR), are different in that they record usage more than they record the traditional performance monitoring parameters.

The ETTR infers engine health from counters that pick up the amount of time above a certain temperature and the speed cycles on the engine in terms of core engine speed. This information allows low cycle fatigue tracking. The operational units have some problem of short meantimes between failure. The EHR runs about 600 hours and the ETTR runs about 2500 hours.

Contract maintenance is used on systems with only a few aircraft. Under this approach, the Air Force uses the aircraft and asks the aircraft company to provide all of the support away from the fl ight I ine. The maintenance approach used by the Air Force is a threefold approach: fl ight 1 ine, intermediate, and depot maintenance.

Fl ight 1 ine maintenance does remove and replace activities, as well as servicing. The intermediate maintenance shop located at the base does minor overhaul work. The major overhaul facil ities does the complete overhaul. In the contracted approach, the contractor provides the intermediate and the depot maintenance.

Contract maintenance systems have a Contractor Operated and Managed Based Supply system (COMBS) at each base operating the type aircraft. Blue suit, fl ight I ine maintenance personnel go to the COMBS facil ity which provides a replacement part over-the-counter. See Table I I I for a summary of USAF contract maintenance programs.

The T-43 aircraft uses a fl ight log engine monitoring program with manual data acquisition, automatic computer formatting and both automatic and manual interpretation. The C-9 uses ground trim data from routine ground runs as a basis for determining engine conditions. From the Air Force standpoint, both of these programs are still fairly new. The T-43 is just now reaching the first overhaul on the engines. The KC-10, also contract maintenance, wil l use fl ight deck monitoring with manual acquisition, automatic formatting and manual interpreting of the data. The E-4, which is the SAC Command and Control airplane, also uses flight deck monitoring, with manual recording , automatic formatting and manual interpretation.

EMS EXPERIENCE These appl ications show the wide range of choices available to accompl ish each of the required EMS tasks. Each of the systems discussed currently fulfills the engine monitoring requirements for its weapon system. However, cost benefits from the EMS appl ication are difficult to accurately quantify.

The benefits are real, but normal system data has not been defined to break out the results. These systems give insight into how the next monitoring system should be designed and built. A sell ing point often used to justify an EMS is el imination of all ground support equipment. These programs generally show that ground equipment may even see additional use. Monitoring EMS data does give us additional insight into engine health, and is capable of controll ing on-condition maintenance.

ADD ON EMS Several operational aircraft have added EMS for service test in an attempt to demonstrate the value of the engine monitoring. See Table IV for a summary of EMS add-ons to operating aircraft. The Engine Health Monitoring System (EHMS) was tested on the T-38. It automatically acquired and formatted the data for semiautomatic and manual interpretation. The results of the T-38 test indicates that EMS probably would not be cost effective. The operational use for the airplane is important. The Air Training Command wants to assure highly reI iable engines. Therefore, its overhaul interval is shorter. The test was run within the ATC standard operational framework; therefore, there were few failures. If the engines do not fail, the monitoring system cannot prove its capabil ity and benefits.

Engine Condition Monitoring Program (ECMP) employed in SAC is fl ight deck monitoring. ECMP is being credited with secondary damage savings of $2 mill ion dol lars a month and reducing the in-fl ight shutdown rate on the SAC fleet by better than 50%. The 50 % is based on the three year, in-fl ight shutdown rate average prior to implementation of the program, compared against the three years since the program has been in use.

Again, look at the concept of operation. ECMP is used on a multi-engine aircraft. With multi-engine aircraft, in-fl ight shutdowns do not have a strong safety indication. Therefore, the overhaul interval is much longer than on a single engine aircraft. Failures do, therefore, occur within the maximum operating time. The ECMP was able to detect these failures before occurrence, allowing repair when the deterioration was in the earl ier stages.

More than 2000 engines have been repaired solely because of ECMP indications.

Only six have been disassembled during this period where no problem could be identified.

The A-IO Turbine Engine Monitoring System (TEMS) has been service evaluated with positive results, and is following on with a squadron level evaluation planned to determine how well that system functions to drive maintenance in the operation scenario.

The electrostatic probe is new technology that came out of the Air Force Institute of Technology (AFIT) about ten years ago. The theory is that rub or errosion in the engine gas path produces an electrostatic charge in the exhaust stream. The quantity measure of electrostatic charge per unit time infers the rate of deterioration within the gas path. The phenomena has been verified but it has not been operationally employed as a monitoring system.

The engine diagnostic system EMS is a monitoring system for the F100 engine in the F- 15 aircraft. It is a service test to val idate state of the art EMS capabil ity against thirty-two goals. The results proved the system would get the data with accuracy equ a l to the test stand.

GENERAL RESULTS FROM ADD-ON TE STS Experience does not show optimistic near term expectations for add - on monitoring systems. EMS generally drives the maintenance cost higher. Start- up problems show that a successful new system takes time to mature. Software problem solutions have taken longer than expected before the EMS successfully records in-fl ight data. Test plans often are written to conduct the evaluation within normal operating scenario w hich prevents the test yielding conclusive evidence on EMS value. The test aircraft are used to meet mission requirements in the normal manner. Maintenance is done by the TOs with 1 i ttle flexibil ity allowed to meet test objectives. Therefore, the test articles may not obtain sufficient fl ight hours or get appropriate focus.

Many valuable benefits come from a monitoring system. You get design feedback, correlation between the testing and operation usage, and verification of repair effectiveness. Verification of repair is often overlooked in the benefits analysis of the program. Maintenance replaces the wrong part, puts the aircraft back in service and it fl ies without a squawk, so it is concluded that the repair fixed the original squawk. Data from the monitoring system allows one actual l y to see the performance trace change providing a powerful qual ity control capabil ity on maintenance and repair. EMS certainly provides improved knowledge of failure modes.

Technical orders are based on a number of A PRIORI assumptions. These assumptions are presupposed by experience, and are not subject to further examination or analysis. Based on A PRIORI assumptions, technical orders are written as if the A PRIORI knowledge illustrates the true behavior of an engine.

A m onitoring system may provide data which causes one to question A PRIORI assumptions . EGT margin is bel ieved to have full capab i l ity to effectively identify an engine as good or bad. EMS data shows that the EGT does go through the red 1 ine just before the engine is torn asunder. However, experience with a SAC monitoring program showed that severely deteriorated engines with basket case turbines often run cooler with a greater EGT margin.

The cooler operating engine can be explained by the facts that the EGT probes are not covering the total exhaust stream and that the turbine nozzle areas change with deterioration. Engines were found by the SAC program with missing first stage nozzle and burner center cones broken off and laying back against the first stage nozzle. These engines passed EGT tests. In fact, two-thirds of them passed complete test runS and were certified for fl ight.

Teardowns later found the bent and broken hardware within the engine. See reference I for an example from the SAC ECMP, The ECMP showed that the beginning failure in the majority of the J57/ TF33 engines started with fuel nozzles. Some fuel nozzles in a couple of burner cans would plug with the engine continuing to meet performance specs.

The good burner cans got more fuel causing hot spots which resulted in burning and bowing of the vanes. Hours later, a vane would eventually burn through. The piece of broken vane would have about an 80% probabil ity of making it through the turbine without engaging in the stationary and rotating vane rows. That is hard to bel ieve, but under the monitoring system, many engines were missing a half first stage turbine nozzle vane on tear down. The missing piece had marked the turbine stages as it passed through. In other cases, the piece would engage between the rotating turbine wheel and stationary nozzle with sufficient force to break a blade. The engines are amazingly tough.

The SAC ECMP uncovered a change in depot maintenance procedures. Fuel nozzles were designed to be repaired in matched sets. It was decided that overhaul of the fuel nozzles in matched sets was too costly. So, I ike parts were worked in batches. Nozzles were assembled randomly from the batches.

Toleranc~ control was gone from the batch repaired fuel nozzles. The result was a very short service 1 ife on badly mismatched sets.

Within six months after depot changed overhaul procedure, fleetwide ECMP monitoring on SAC engines identified the problem. The fuel nozzle overhaul problem potential will never be known because it was not allowed to exist long enough to have its full impact on the fleet. ECMP identified engines with bad nozzles for repair before other parts were damaged. How do you value something that is responsible for turning a problem around before its impact is documented?

MANAGEMENT LESSONS LEARNED Responsibil ities should be defined at the outset of an EMS program. Keep on board equipment simple which may be aided by 1 imiting the in-fl ight task to data acquisition. Do the formating and interpretation of the da~a in the ground system. Remember that every pound of weight on a fighting aircraft costs performance. The mission of the Air Force is to fly and fight. Man should be in the loop so he is able to understand what the output from the monitoring system means. Provide real istic time and training, support equip- ment, and EMS spares. Organize a real istic, timely base monitoring team to use the in-fl ight data to drive maintenance actions. Effectiveness is improved if the EMS system is built-in versus retrofit. One should not wish to monitor everything.

____ J TECHNICAL LESSONS LEARNED If the necessary parameters can be defined, it is possible to ' minimize sensor requirements. Insure that the output of the in-fl ight equipment is compatible with the existing test equipment . Provide flexibil ity so the necessary data can be obtained to track a new failure mode. Provide self- check to isolate the bad data. Trending does allow you to determine deterioration within the engine. Increasing fuel prices are emphasizing the need to obtain the engine data while the engine is in the revenue service, to use the airl ine term, rather than do a ground run. If the engines don't have problems, you don't need monitoring. Good engines receive no benefit from being monitored . If you know what the engine's performance parameters are doing, you can determine its reI iabil ity potential and therefore enhance fl ight safety.

THE ADVOCATES PROBLEM Why is it such a problem to get EMS on AF equipment? (See Figure 3) The figure shows the time I ine for a weapons system versus accumulative or I ife cycle costs. Air Force System Command (AFSC) is responsible for the acquisition process until Program Management Responsibil ity Transfer (PMRT).

then, Air Force Logistics Command (AFLC) takes over for logistic support.

The process begins with an approved operational requirement for a specific weapons system to do a job. The System Project Office (SPO) director is assigned the responsibil ity for the acquisition. He is given a certain budget and has to acquire the required capabil ity within that budget. An engine monitoring system adds an immediate cost increase to the system which is apparent. EMS benefits accrue in system operation after PMRT. Several years of operation may pass before the meantime between failure for the major items of the system is reached. During the acquisition phase there is no way of knowing the correct slope on the operations cost curve. Therefore, the SPO director on his watch sees only the impact of EMS cost on his system. EMS potentially available benefits accrue in service after PMRT when AFLC has the watch.

ELEMENTS OF SYSTEMS EFFECTIVENESS Earl ier the importance of viewpoint was discussed with the idea that the "jackass on the hill could see further than the genius in the valley."

Analyze that idea from a standpoint of system effectiveness. (See Figure 4) Three people are involved in the Weapon System Effectiveness Problem: The overall field commander decides what weapon will be employed on what target at what time, and the branch on the right of figure 4 represents his interest; The wing commander has to implement the field commander's orders as the center branch represents his interest. He wants X equipment on the I ine and ready to meet the mission requirement. The Deputy Commander Maintenance (DCM) is charged with the responsibil ity of making that equipment available. His interest is in the branch on the left. The EMS system in order to be judged cost effective and worthy of purchase by the SPO director must clearly improve each of these elements for total system effectiveness. That is the heart of current EMS development guidance that is being given to industry for the new weapon system starts.

EMS DEVELOPMENT GUIDANCE The fol lowing general guidance for the development of an Engine Monitoring System (EMS) was provided by the Propulsion Director of Engineering, 30 January 1981, to maximize system effectiveness of our new weapon systems.

The EMS will be dedicated primarily to the performance of "Engine H onitoring," i.e., capture of in-fl ight engine operating data. The EMS function operating within the planned logistics/maintenance concept will not be compromised by over sophistication of tasks and multiple roles for the EMS hardware. Where airframe monitoring systems are to be used, the EMS must be compatible and compl iment that system. However, an option for independent operation of the EMS should be planned in the event that an integrated airframe/engine monitoring system is not included. The current development guidel ines for on-board EMS capabil ity are: I. Simpl ify on-board equipment by limiting in-fl ight requirements to data capture with data interpretation on the ground.

2. Limit EMS design goals to evaluation of engine suitabil ity for continued service rather than fault isolation to an individual module/ component.

3. Plan use of ground test equipment, e.g., borescope, chip detector, to confirm EMS indications and enhance diagnostics prior to engine removal.

4. Integrate the EMS output from an operational weapons syste m viewpoint by use of ground station data processing with the man in the loop for interpretation and direction of maintenance.

PLANNED APPROACH During the early phases of each EMS program the engine contractor will be tasked by the Air Force to prepare a detailed feasibil ity analyses covering the following areas: I. A I ist of aircraft/engine parameters to be monitored/recorded i n-fligh t by the EMS.

2. Feasibil ity of performing the following engine monitoring/diagnostic functions using the parameters recorded in-fl ight: Engine Documentary Data - . Parts Life Tracking Parameter Tracking/Trending Engine Suitabil ity for Fl ight Warranty Val idation (if required) Suitabil ity for Fl ight In addressing the feasibil ity of performing each of the above functions with the EMS, the contractor must direct his analysis to answering the following questions: a. Does the technical expertise exist currently to adequately perform each function without causing a negative impact on the planned maintenance/logistics concept for the appl ication?

b. How would each function's data product interface with the planned maintenance/logistics concept?

Where and by what means would the EMS data product be converted c.

into useful information?

d. Who would eventually use the information?

What will an EMS do for system effectiveness?

e.

Once this feasibil ity analysis is provided, a complete review will be conducted by the Air"Force. The direction for the development of the EMS for the engine will then be establ ished.

TOTAL SYSTEM VIEWPOINT The engine contractor should be tasked with the responsibil ity for developing all aspects of the EMS system with Air Force assistance. This includes all hardware required on-board, on the fl ight I ine and in the ground station, plus all software required for the EMS to function satisfactorily.

The EMS must work hand in hand with the planned engine logistics/maintenance concept. As such, both systems or programs must be developed concurrently to insure optimum utility of the EMS. It is essential that a total system perspective (airframe, maintenance, logistics) be the overriding consideration in the development process and that the EMS and the maintenance concept be concurrently developed. The overriding question is: "How much EMS is enough for system effectiveness optimization while remaining affordable?" Pareto's criteria can help zero in on the answers during the acquisition phase.

Us i ng the weapon system approach as an evaluation criteria early in the acquisition phase will hopefully help get the genius out of the valley onto the back of the jackass on the hill so that they can together gallop toward real ization of po t ential EMS capabil ities.

REFERENCE 1. McCord, Robert M . ; Engine In-Flight Monitoring (Part II). Maintenance Magaz i ne, Ai r Force Inspection and Safety Center, Norton AFB, Cal ifornia, 1977, pp. 31-38.

-- - - ----

USAF TURBINE ENGINES

MAINTENANCE VS ACQUISITION COST TRENDS 1880 MAINT COST A COUISITION REMOVALS MMH /l000 EFH ENGINE NOT INITIAL • DOLLARSI I OOOEFH COST · THEN, 1000EFH MISSION OPEIIATING CAPABLE '1\ CAPABILITY NON AB 1116K 1. 28 .181K lUI (TJ) .88K .210K 100 0 . S6 1l1li0 (TF) $1281( $888K 1080 0 .38 1181 (TF) 13 .510K $681K 1.2 28 820 1876 (TF) A8 '18OK

,. 13201(

,.4 3. 1 1868 (T J) 153)( .86K 4.2 22 87 1l1li2 (TJ) $130K 1.4 U18K lttl (IF) 11 '2 ,0001( 8.2 1171 (TF) 17 1' ,8S0K TABLE I - M6.INTENANCE COST TRENDS FOR USAF T~INE ENGINES, AS A BASIS FOR CCMPARING ACQUISlTlOO COST, 1rE COST Ffm 1lE INITIAL Aa;}UISITIOO COOTRACT IS PROVllEl IN _ l1'f,;N YEAR ro~~,Iv, THE COST DATA IS EXTRACTED FRO'! 1rE 1900 ASD ENGINE AlNI SORY ffiOUP (EAG) MINUTES Oo\TED 25-24 SEPTE/"6ER ,C:jdI,

ENGINE MONITORING SYSTEMS

DEVELOPED WITH AIRCRAFT AIR CR Af T.[ NGINE SIATUS SYSTEM ~ ACOUIRE / fORMA T li NTE RPRE T OPERATIONAL MT8f 1100 HOURS SR71 /J58 M MRS OPERAIIONA L IoITBf 'DO HOURS M/A C5Allf39 MAOARS TOIAL MAOAR STSHM a HOURS OPERAIIONALIN CAl 161 IoI /A B: ' f'O I ClTS TESIS COMPLEX NOI READY fOR OPERA TIONAL OEPLOYIoIENT OPERATIONAL MI8R 600 HOURS FH /F 16 /FIO O EHA 101 101 OPERATIONAL MT8R 2500 HOURS M A 10 / lfl' EITR M TABLE II - SlH'WlY OF E~It£ M)NITORI~ SYSTEMS I:EVELOPED WI1H ~ AIRCRAFT,

ENGINE MONITORING SYSTEMS

ADD-ONS TO OPERATIONAL AIRCRAFT OA TA ACQUIftEJfORIiA I ' I~ IERPRET AIRCRAfT / E~GIIIE ~ EHIIS iliA TJIIJH JUOlOEO ~Ol COST EffECTIVE

.. KC1J~/lfJJ OPERAIIONAL . SAVES

ECIIP II II B~70 . G ' 51 S7 " llllO ~ I~ SECONOARY OAIIAGE EACH IIONTH REOUCEO B~ 71H1TfJJ C 141 If SO RA IE BY NEAR ~O· I •.

TEllS II /A A 10nFJ4 SERVICE EVALUATION O~ ~ AIRCRAfT WARRANTS fOllOW O~ SOUAORON EVALUATION

ElECTROSTATIC .. NUMI.ROUS PHENOIIENA VERIfItO

P1IOBES ENCINES EDS iliA f 15 /f 100 SERVICE lESION 5 A IRCRAfT IIHIS IIANT Of THE J7 OESIGN GOALS ACCURACY EQUAlS TES I STANDS .

TABlE I I I - SlM"ARY OF ENGINE ~ITORING SYSTEMS TESTED /lS ADD-OOS TO OPERATIONAL AIRCRAFT,

ENGINE MONITORING SYSTEMS

CONTRACT MAINTENANCE AIftCRAfTIENGIllE ~ ~ OATA ACOIMEJFORIlAT/lIHERPRET UIIE Fl I LINE BLUE SUIT A/ II T4J/ JUD 9 FL"HT LIllO II COIIBS' •• UNllEOiSFO ClEIiAS' oell (TCLE LIIoIITS Fl I LINE BLUE SUIT GRDUIIO TRIIoI II II II C· !l / JUO 9 CDIIBS DA TA TREIIOtD HUO TIIoIE · CYCLE LIIoIITS II KC · l0 / CF ' · ~OCl Fl I LIIIE · BLUE SUIT EHI .. II COIIBS oell · CYClE LIIIITS II E4A1CF' 50 Fl T LIIIE · BLUE SUIT EHI II COll8S oell · CYCLE LIIIITS • CENTRAL INFORIIATION E,",*E IIO.TORING AND AIRCRAfT SYSTEIIS .. ENGIIIE PEfORIlANCE 1I0000TORiIIG . GE ••• COIHIACTOR OPfU TEO AMO IIAUGEO lASE SUPI'L Y . INCLUDES IN TERIIEOlA IE AIIO DEPOT ACTIVITIES TABlE IV - stH'ARY OF ENGlt£ ~ITORING SYSTEMS USED WITH USAF AIRCRAFT OPERATED THE COOTRACT fo'Al NT£NANCE cnuPT.

ENGINE MONITORING SYSTEMS

REQUIRED TASKS SYSTEM IM PACT • IMPROVED MI SSIO N CAPABILITY ?

o·~

• LOWER LCC ?

o vO

ACQUIR E DATA TAKE ACTION ------- --1--------- • DIRECT SPE CIF IC REPAIRS WITH • G AS PATH : • MECHANI CAL WOR K ORDERS .

PARAME TERS I PARAMETERS I FORMAT DATA INTERPRET DA TA • VALIDATE • C HECK L IMIT S • COR RECT • CHECK TRE ND S DIAGNO SE • COMPRESS • PR OGNOSICATE • DISP LA Y FIGrnE 1. n-tE TASKS I'IH ICH M.JST BE ACCO"t'L1SHED FOR DJRING THE OPERATIoo OF AN ENGINE I"OOITORING SYS TEM.

n-tERE ARE OPTIooS AT EACH LEVEL ~ICH RANGE FRO-I PENCIL AND PAPER TO CAPABLE ELECTRooICS. THE I"OST IMPORTANT LINK IS INTERACTIoo WIn-t n£ OPERATIOOAL SYSTEM.

WHY?

ON - CONDITION MAINTENANCE (OeM) REDUCES

RISK .. SAVES H

o w ~ ::I fn Z o U FIELD LIMIT w MAXIMUM u..

OPERATING ::::i TIME ENGINE OPERATING TIME FIGURE 2 - ~ USAF WANTS ON""CooDITIoo to'AINTENANCE (QCM). IF IT IS NOT BROKEN, ~ FI X IT? to'AXIMJ'I O PERATING TIt-E MA. Y F IX ONE GXlD ooE I'IHILE ANOTHER FLIES TO FAlWRE. 0C1'1 Ia:NTIFIES TI£ EXTREt-ES AND WlKES TI£ REQUIRED REPAIRS AT TI£ AP PROPRIATE Tlt-E.

ELEMENTS OF SYSTEM EFFECTIYENESS

ISYSTEM EFFECTI V ENESsl AV AILABILITY DEPENDABILITY CAPABILITY ME ASURE OF SYSTEM MEASURE OF SYSTEM MEASURE OF CO NDIT I ON AT CONDITION DURING RESULTS OF ST ART OF MISSION PERFORMANCE OF MISSION MISSION RE LIABILITY REPAIRABILI TY RANGE MA INTAINABILITY SAFETY ACCURACY LO GISTICS flEXIBILITY POWER HU MAN FACTORS SURVIVABILITY LETHALI TY FI (lJRE 3 - TI£ ELE1'£NT S OF SYS TEM EF FE CTIVENESS, POSITI<l'l IN STACK EXERTS GR EA T IN FL LE NCE <l'l INDI VIIlJAl V AlLE PU.CED <l'l EACH ELE1'£NT OF SYS TEM EFFE CT IVE N E SS, TI£ ULTI t-'A TE VI EWPO IN T IS SYS TEM EFFECTI'v£NES S,

THE ADVOCATE'S PROBLEM

PROVE EMS VALUE FROM SYSTEM VIEWPOINT SPO AFL C I WITH OUT EMS ?

PMRT I ' I I ____ I I .... 1 W HI CH ~_- UNKNOWN TO l- I .... _--- THE SPD V) o I /'_-S LOPp l u I /_ - I WITH EMS ?

> -: ~ ~ ____ 7':::'-::' ~I~~E~ ~N~T':"A.!: .!...N;:~S2~~N2 .

I- <l I _--- CE RTA IN TO THE SPD -' ::> ~...:- S _ _ _ ____ ,=-O_W~~ ~~~,=-I~'!.-E.?!..M...!'~!..

::;; I ::> I Q . " WHY BUY EMS ?

u u I <l I A " IT WILL BE GOOD FOR YOU I NEED PROOF · NOT SPECULATION I I TI M E (YE A RS) MT B Fs · GENERALLY GREATER THAN EXPERIENCE AT PMR T FI~ 4 - n£ AIMlCA TE'S P R:l BI..EM, roNVINCE THE SPO DIRECTOR TO SPEND n£ LP -FRONT I'm EY REW IRED TO BUY AN ENG I ~ I'mITORING SYSTEM, 22 3

TOTAL SYSTEM VIEWPOINT

WHO DOES WHAT?

--- ACCUMULATED R / RESULT HOW MUCH IS ENOUGH?

E S U GOAL L T S MAXIMUM SYSTEM EFFECTIVEN ESS VITAL TRIVIAL FEW MANY F IGUR E 5 - THE TOTAL SYSTE M S VIEWP OINT AND PARETO'S LAW MAY IN COMBINATION BE THE KEY TO AN S WERING CRI TICAL QUEST ION S NEEDED TO DEFINE T HE REQUIRE D ENGINE M O NITORING S YSTEM .

FIOO ENGINE DIAGNOSTIC SYSTEM STATUS TO DATE James A. Boyless Wright-Patterson Air Force Base SUMt-1ARY An engine diagnostic system, proposed for the FIOO engine, is being tested in five specially modified Tactical Ai r Command F-15 ai rcraft during a 16-month flight evaluation at Langley AFB, Vi rginia. After more than 3300 engine operat- ing hours encompassing almost 900 fl ights during the flight evaluation, these aircraft provided a data base, stil I being analyzed, that has shown successful demonstration of the original functional characteristics. Table IA presents the general system evaluation in six areas whi Ie Table IB presents a more detailed look at these functional characteristics through March 81. Those areas listed as partially demonstrated are now being further tested at Langley AFB. Four general design requirements; recording engine operatin g time/low cycle fatigue event detection, engine trim and trend and performance data collection were demonstrated. It also successfully demonstrated validation of maintenance actions taken and indicated needed maintenance.

I NTRODUCT I ON The U.S. Air Force On-Condition Maintenance (OCM) concept, defined in AF Regulation 66-14, directs maintenance on the basis that the condition of the equipment dictates the need for maintenance. To adequately perform OCM, inputs from engine maintenance management tools such as oil analysis, borescope inspection, parts tracking, periodic and phase inspection, monitoring, and diagnostics are requi red. Of these, monitoring and diagnostics are, by far, the most difficult to achieve. Monitoring and diagnostics development activi- ties have encompassed aircraft/engine s y stems from the F100/J57 to the recent F- 15/FI00. On each system, various parameters, both airframe and engine, have been used to provide a sumrlary of information for maintenance personnel, logistic support and provide a feedback loop for future engine development. A review of the F-15/FI00 Engine Diagnostic System (EDS) through a system descrip- tion and status to date is presented.

BACKGROUND HISTORY 1. As gas turbine engine technology increased in complexity, so too, did the need to assist maintenance personnel to perform and diagnose problems for maintenance. In addition to assisting on-base maintenance through increased emphasis on On-Condition Maintenance, logistic requirements for improved engine I ife usage data dictated a means of acquiring that data be developed.

2. Preliminary studies by the Air Force Propulsion Laboratory indicated that an Advanced Fighter Diagnostic System (AFDS) could prove feasible in an application for an advanced design jet engine. The A FDS results led to defini- tion of both hardware and software requirements as well as researching existing capabi lities. Additional studies were then conducted to evaluate the signifi- cant areas of on-board processor, engine sensors, and use of existing equipment for system development. This system (renamed FIOO Engine Diagnostic System (EDS)) was targeted for the FIOO engine in both the F-15 and F-16 aircraft.

These two aircraft powe red by the same bas i c Pratt & \1h i tney Fl 00, we re chosen for the complexity and the operational envi ronment envisioned for the engine.

SYSTEM DESCRIPTION The FIOO engine is well suited for the complex task of engine diagnostics. It is a modular engine designed for operational readiness and maintainabi lity. It is also complex and requi res knowledgeable maintenance personnel for repair.

To ably assist the maintenance personnel, the FIOO EDS went through an exten- sive review of Fai lure Modes and Effects Analysis (FEMA) and cost effective analysis. Thi rty eight engine and airframe parameters were included in the EDS.

Once these parameters were selected, hardware was developed to monitor the requi red information. The FIOO EDS has eight primary elements that visually present cues of engine status, and/or provides a means to collect and diagnose engine anomal ies. These elements shown in Figure I are: Ground: Data Collection Onboard: EDS Eng i ne Sensors Unit (DCU) Engine Multiplexors (EMUX) Diagnostic Display Data Processor Un it (DPU) Un i t (DDU) Status Panel Cockpit Advisory Lights Cockpit Pi lot Option Switch ENGINE MULTIPLEXER (EMUX) The engine multiplexer unit was developed under Air Force contract to collect, condition and multiplex sensor signals serially to the onboard Data Processor Unit (DPU). The EMUX replaces both the present FIOO Event History Recorder (EHR) and the junction box (J-Box) for engine aircraft electrical connections. The unit is fuel cooled using existing EHR cooling lines and is hard mounted in the area vacated by the J-Box. EMUX reliability and durability to perform its functions is achieved through internal vibration isolation.

DATA PROCESSOR UNIT (DPU) The DPU is an airframe mounted, air cooled unit consisting of a central processor, Intel 8080, core memory, and interface ci rcuits. Both cool ing and electrical power requirements are provided by the aircraft. This unit is the nerve center of the inflight monitoring system. It is programmed through software logic to detect a limit exceedance, declare an event, and store that event for later collection/diagnosis.

AIRCRAFT COMPONENTS I. There are three aircraft components that are integral parts of the EDS; the cockpit warning lights and pilot option switch, status panel, and transfer receptacle. These components are also depicted in Figure 1. The DPU can, on command, store a data record by means of the pi lot option switch located to the left of the pilot. In addition, a cockpit warning lights indicate Fan Turbine Inlet Temperature (FTIT) Overtemperature or excessive temperature occurrences.

2. To aid Flight Line personnel to quickly determine if an aircraft can be turned around, an EDS Status Panel is located in the existing maintenance access door, 48L. The s tatus panel has latching indicators that can be set by either the DPU or EMUX. Either built-in-test for DPU and EMUX will set these latches as well as Hot Start detection for either engine.

3. The tran s fer receptacle, located in the same access door as the status panel quickly connects the DPU to either the DCU or DDU for extraction of stored data. Average transfer time ' is six seconds. Either collection or diagnostic operation can be performed under the Ilwing."

DATA DIAG~OSTIC UNIT (DDU) Just as the DPU is the nerve center of the onboard system, the DDU serves that function on the ground. It is a portable ground unit with an alphanumeric display screen and keyboard for interfacing the maintenance personnel. For storage of fl ight data the DDU has the capability to maintain five records. The unit shares common components with the DPU for increased maintainability. The il interface of maintenance personnel and the under the Ilwing is accomplished by providing power through batteries. The unit can also be used w ith 115 volts AC in test areas or engine shop.

DATA COLLECTION UNIT (DCU) The DCU is small, 1 ight weight, and portable unit that uses internal battery power. It is used to collect and transfer data stored by the DPU. There are indicators for successful transfers of data from DPU and the presence of any maintenance advisory information. The DCU is designed to collect data from 10 to 15 aircraft and shares common modules with the DPU and DDU.

SYSTEM CAPABILITIES The entire Engine Diagnostic System functional characteristics are designed to perform in five specific areas. These areas include Time and Cycle recording, Event Detection, Diagnostic and Troubleshooting, Engine Trim, and Trend & Per- formance data collection. A capsulized view of these capabilities vs either installed or uninstalled engine is shown in the following table: FUNCTIONAL CAPABILITIES VS ENGINE INSTALLATION INSTALLED UNINSTALLED Time and Cycle DPU DPU Even t De tect ion DPU DDU Diagnostic & Troubleshooting DDU DDU Engine Trim DPu/DDU DDU Trend & Performance DPU A schematic view of how EOS data was collected is shown in Figure 2. The components shown detai I the units involved in the . airborne and ground portions of the system description.

FLIGHT EVALUATION TEST ENVIRONMENT 1. The test environment was a Tactical Air Command operational base, where the EOS was an adjunct to the existing base level maintenan ce organization.

The base lev el maintenance organizations were involved throughout the Fl ight Evaluation Program (FEP) but the impact of EOS on maintenan ce was to be on a non-interference basis. This basis was justified in the fact that EOS was in a validation phase rather than actually being incorporated into the entire fleet.

2. Both test equipment, and Auxil iary Ground Proce ssor (AGP) we re procured and installed in the EOS Laboratory. Nine permanent party individuals were on-site during the FEP.

TEST AIRCRAFT Five Tactical Ai r Command (TAC) F-15 (10 FIOO Engines + one spare) were specially modif ied with EOS equipment. A control group of 12 non-EOS FIOO engines were identified for comparison.

TEST METHOD 1. The objective of the test can be summed by Figure 3. The functional capabi lities of the FIOO EOS we re to be val idated through actual inflight collection, ground transfer, and on-site evaluation of data. As a basis for valida tion, 3000 engine operating hours was set as a goal. Furthermore, a de- tailed Flight Evaluation Plan (FEP) was used as a tool in evaluating the inflight data. Every diagnostic find was verified and val idated by performance of a resulting maintenance action.

2. Time and cycle recording functional capabil ity was to be accompli shed automatically by the EOS. Transfer of the recorded data would take place from the OOU to a teletype in the proper format of the present AFTO form 93.

3. To accompl ish the test method the present Maintenance Action Cycle used at Langley was to be integrated with the EOS. Figure 4 shows schematically how this occurred. During the test the cre wchief would check the EDS status panel to determine aircraft availabi I ity. If any wo rk would be needed the fl ight dispatcher would send a technician with the DDU or DCU. The DDU provided information would be revi ewe d by the propulsion maintenance unit with advice and/or assistance provided by the EDS team.

4. To gather the data, there were five F-15 aircraft and eleven engines specially modified for the test (See Figure 5). These aircraft were to average between 45 and 60 engine operating hours per month (See Figure 6).

- , 5. Val idation of the data included an indepth critique of the infl ight data. There were five categories in which the data were grouped. These groups include Hits, Goods, False (I & II) and Misses. Hits would be scored as follows; An event was declared only by EDS and/or the event was confirmed by the present reporting system, a pilot or maintenance write up. A good is an event not declared by EDS nor reported by the present system. On the other hand, False I is an event declared by EDS and not by the system while False II is false but a known "fix" is in work to remedy the cause. Finally, a Miss is a pilot or maintenance write up not detected by EDS when it should have been.

6. Engine trim both instal led and uninstal led would be performed in place of the present trim procedures using the present M-37 test stand. Careful monitoring of the time used to trim, and fuel used were recorded for comparison against non-EDS engines.

7. Diagnostics and troubleshooting was to be evaluated by careful review of actual usage of the equipment by the Maintenance personnel. If maintenance was declared once an event detection occurred, the procedure cal led for the repairman to use the DDU. Once validation of the event occurred, the DDU was to be used to diagnose or "troubleshoot" the malfunction. Maintenance records would be screened and data kept that expressed the amount of usage the DDU attained, time to troubleshoot, and diagnose malfunctions and compared to the control group.

8. Once the diagnostic and troubleshooting scenario ended and an Engine trim was requi red, the fourth capability was to be evaluated. In fact, this evaluation included all engines to be trimmed. Records were to be kept for manhours required to trim and fuel used.

9. Finally, the abi 1 ity of the system to collect trend and performance data was to be tested.

RESULTS AND DISCUSSION 1. The Fl ight Evaluation Program (FEP) test results are presented in the succeeding figures. These results are based upon the period 1 Apr - 12 Dec 80 and the data gathered at Langley AFB, Virginia. The FEP, because of software complexity, was divided into a debug and actual validation period. Figure 7 gives a comparison of the actual vs projected engine operating hours. There were 2577 hours attained by 12 Dec 80 and an additional 738 hours through 26 Mar 81. Time and Cycle data was automatically printed by a printer to correspond with the actual AFTO Form 93 required by the engine management information system (see Figure 8). From the beginning of the program there were 13 events that were continuously monitored from start-up to shut-down of the engine. As the program progressed, however, there were lessons learned that deleted one event and switched four from No-Go to Maintenance Advisory.

These four events are depicted in Table I I I. The system's ability to detect events accurately was extremely important. Figure 9, Event Detection Accuracy, shows how the accuracy of the system progressed. The check sum on the abcissa is an identity associated with software improvements of the basic event detection logic. As can be seen accuracy increased from a point of 88.7 per- cent to 99.7 percent at 12 Dec 80. The 99.7 percent assessment is based upon 77 transfers of data (See Figure 10) where 63 Hits were recorded. These Hits inturn were used to recommend maintenance action. There were 1006 goods reports with onl y 3 False I events detected.

2. These results demonstrate the successful capabi I ity of EDS to detect events. B y using the la st operable checksum, 0119, the enti re evaluation period was reviewed and summarized as shown in Table V. Of the twelve events, continuously monitored by EDS there were five events that we re detected on numerous flights during the evaluation period. Three events clearly stand out.

Stal Is were very prevalent during the evaluation period. How extensive they are can only be speculated at the present time. Detailed analysis of this event cont i nues.

3. Fan Turbine Inlet Temperatures (FTIT) Spread events were also numerous during the evaluation. These EDS modified engines have reported numerous FTIT Spread Events. Investigation continues to determine cause and effect on the engine. One attempt to seek infor mation on effects has been to change borescope (Visual) inspection of the three engines to a 50 hour interval rather than the 100 hour interval presently used.

4. Just as Stalls and FTIT events have been prevalent, Rear Compressor

Variable Vane (RCVV) events have also been numerous. These events have been tracked throughout the evaluation period. Investigation as to caUSe continues.

Diagnostics and Troubleshooting capabi lity have been demonstrated by EDS personnel in the laboratory and to a 1 imited extent by maintenance personnel.

Diagnostics and Troubleshooting by the EDS personnel throughout the evaluation period occurred on a daily basis to confi rm detected events and recommend maintenance. Maintenance personnel used EDS in a limited manner as a mainte- nance tool as well as a training aid.

5. Tables V, VIA and B show the potential engine and maintenance saVes credited to EDS durin~ fhe flight evaluation. Engine saves included a high scavenge pressure event that the pilot was unaware of. Had the discrepancy continued the engine could have reached the catastrophic state of complete loss.

The most obvious save for maintenance is preventing mis-di rected maintenance.

Four of the eight pilot assessments included dual engine anomal ies whereas EDS confi rmed only one engine had the anomaly not both.

6. EDS engine trim capability was demonstrated. Both uninstalled and instal led engine trim was performed using EDS. After five partially successful attempts at uninstalled trim, identified software changes have been made and testing continues. Installed engine trim has been successfully demonstrated after several attempts. The enti re trim procedure w ith exception of Engine Pressure Ratio (EPR) check was performed. EPR check could not be accomplished due to a false sensor reading.

7. Finally, trend and performance data was collected. Accumulated data indicates that 183 data points we re captured EDS and the resulting analysis revealed that 64 % of the data points lay in the lower PLA range (See Figure 11).

This range, 30 -40°, is the area of idle reset area where the augmentor is wide open. The conditions for data to accurately reflect trending and performance required a stabilization time of 180 seconds and was often reached in a landing approach. But due to the reprogramming abi lity of EDS a change of constants for o PLA was approved and capture of data wi1 1 be in the desired PLA range, 40-80 .

SYSTEM EVALUATION

SUMMARY • OVER 650 FLIGHT SETS OF DATA ANALYZED • DEVISED/IMPLEMENTED SOFTWARE DIAGNOSTIC TECHNIQUES • EVALUATED 8 SETS OF SOFTWARE· 3 OFPs, 2 OOPs, 20CPs • VALIDATED SYSTEMS CONTRIBUTION IN AREA OF • ENGINE ANOMALIES • MAINTENANCE ACTIONS • VALIDATED SYSTEMS ABILITY TO COLLECT DATA • TIME/CYCLE • PERFORMANCEITREND • DEMONSTRATED SYSTEM FLEXIBILITY THRU REPROGRAMMABILITY TABLE IA

CAPABILITIES

SUMMARY OF RESULTS (I APR 80 . 20 MAR 81) • DATA COLLECTION · 87 % OF FLIGHTS RECORDED • EVENT DETECTION· 99 % ACCURATE • DATA ANALYSIS · DEMONSTRATED • TRIM CAPABILITY· PARTIALLY DEMONSTRATED • GROUND DIAGNOSTICS · DEMONSTRATED • USER EVALUATION · PARTIALLY DEMONSTRATED • TREND AND PERFORMANCE· DATA COLLECTED TABLE IB

EDS TOTAL SYSTEM DESIGN

FUNcnONAl CHARACTERISTICS • DETECT EVENTS AND LIMIT EXCEEDANCES • COllECT IN·FLlGHT TREND DATA • PROVIDE IN·FlIGHT PERFORMANCE CHECK CAPABILITY • RECORD OPERATING TIME AND LCF COUNTS • PROVIDE "NO·GO" INDICATION AT THE FLIGHT LINE • CONDUCT FAULT ISOLATION AND DEFINE MAINTENANCE ACTIONS • PROVIDE HARD COPY OF ENGINE RECORDS FOR INPUT INTO CENTRAL DATA SYSTEMS • PROVIDE CAPABILITY TO PERFORMANCE ENGINE TRIM TABLE I I EVENT MENU AT START OF ATEND OF FLIGHT PROGRAM fliGHT PROGRAM 13 EVENTS (11 NO · GOI 12 EVENTS (6 NO · GOI MAINTENANCE MAINTENANCE EVENT TYPES NO · GO NO · GO AOVISORY ADVISORY - X X - HOT START - - N2 OVERSPEED X X - X X - fliT OVERTEYP - - fliT SPREAD OUT OF LIMITS X I X - X - OIL f'RESSURE OUT OF LIMITS - SCAVENGE I'RESSURE OVER LIMITS X - I - X - VIBRATION OVER LIMITS I EEC fAULT X X

/&

- ENGINE STALL X I

:~

X - AUGMENTOR BlOW·OUT -MISLIGHT X.&. I - RCVV OUT OF LIMITS X - I - MAIN FUEL f'UM/, DETERIORATION - X - - - MAIN FUEL f'UMf' FAILURE X X

& If deared by pilot .&. If out of envelope

TABLE II I EVENTS* (HITS) DETECTED PER ENGINE 10 JULY THRU 18 DECEMBER ENGINE SIN TOTALS 330 415 470 528 639 694 160 311 722 801 907 HOT START 2 O'SPEED ENGINE O'TEMP 1 1 2 NO -GO OIL PRESS . 1 2 1 4 EEC 4 7 4 8 1 24 MFP FAIL 0 STALL 1 1 11 2 3 3 21 SPREAD 23 22 38 1 1 85 MAINTENANCE SCAV PRESS. 9 9 ADVISORY AUG aO/ML 1 1 2 RCVV 5 19 2 3 1 16 46 3 4 vias 4 13 4 7 6 1 42 9 9 24 18 6 183 TREND 16 15 29 17 27 13 RECORDS PERF 3 3 17 3 2 6 10 12 4 3 67 "Hits - as determined by checksum 0119 in use at the end 01 the program TABLE IV POTENTIAL ENGINE "SAVES" BY EDS ENGINE PILOT EVENT . CORRECTIVE ACTION SIN REPORTED YES SERVICED OIL TANK OIL PRESS, LOW 311 VAC CHECK NO, 4 COMPARTMENT.

NO FOUND FOREIGN MATERIAL IN SCAV PRESS., HI 470 ENGINE OIL SYSTEM.

EDS DETECTED O'TEMP LEVEL O'TEMP 330 YES HIGHER THAN REPORTED BY THE PILOT. EEC CHANGED.

BORESCOPE EVERY 50 FLIGHT 160 YES· HOURS AS A PRECAUTION UNTIL FTITSPREAD 311 YES· PHENOMENON AND CONSEQUENCES YES· CAN BE QUANTIFIED.

311 N.A . REPLACED NO. 4 PROBE, VERIFIED FAILED 694 N.A. REPLACED NO. 4 PROBE, VERIFIED FTITPROBE 694 N.A. REPLACED NO.5 PROBE, VERIFIED NON BILL OF MATERIAL RELAY FAILED COCKPIT NO PANEL BLOCKED AN EEC FAILURE WARNING LIGHT WARNING TO COCKPIT.

"FTIT Spread is not monitored in aircraft other than EDS equipped aircraft.

TABLE V

MAINTENANCE "SAVES" BY EDS

ENGINE PILOT ASSESSMENT EDS RECORD SIN HAVE TO MISMATCH PILOT OPTION DATA RECORD THROTTlES TO MATCH CONFIRMED MISMATCH IN RPM RPM, fliT. PLA RIGGING.

470 NO COMPLAINTS REPEATED EEC LEVEl 1 FAUL TS. ODU CABLE SHORTED.

639 NOT APPLICABLE SEVERAL FALSE RCVV EVENTS ON RECENT fliGHTS, TT2.5 ERROR. MISSING AP2 PLUG.

LOW OUT OF TRIM, 722 LOW THRUST RPM IN STABILITY PILOT OPTION RECORD.

801 AlB BLOWOUT ON BOTH NOTHING WRONG WITH 80t.

(528) ENGINES (528/801) ENGINE SIN 528 HAD A "HARD LIGHT/BLOWOUT" FOllOWED BY A STALl. EPR HIGH 0.11.

TABLE VIA

MAINTENANCE "SAVES" BY EDS

ENGINE PILOT ASSESSMENT EDS RECORD SIN 311 AlB BLOWOUT ON 311, TOOK STALL FOLLOWING AN PILOT OPTION. AUGMENTOR "HARD LlGHT/ BLOWOUT." RCVV's OUT OF BAND, AXIAL ON STALL, RCVV AND PILOT OPTION EVENTS.

907 AlB BLOWOUT ON BOTH NOTHING WRONG WITH 907.

(528) ENGINES (907/528). ENGINE SIN 528 HAD A "HARD LIGHT/BLOWOUT" FOLLOWED BY A STALl.

PILOT REPORTED AUGMENTOR EDS DETECTED STALLS IN ANOMALIES ON THREE FLIGHTS . AUGMENTATION ON EACH OF ON SECOND FLIGHT DOUBLE THE THREE FLIGHTS FOR HARO LITE ON BURNERS. ENGINE S! ' ~ 722 ONLY.

TABLE VIB - -- - -- - -

S1STEM DESCRIPTION

EDS STATUS PANEL ADVISORY l (NQ (0' IIIun$TUT EDS ENGINE SENSORS SWITCH 10.

..-ANU!,

$0

I~A"'i1 ~A"'. J ®.

!:'ll!~

$$0

~ "'.MUII "'ING orv HOT n,,"'T

00$

".MUII "'NG 'HG ENGINE MULTIPLEXER UNITS (EMUX) HOT I'"''''

/ ~(\

/'V W ~ \.."J

'"

NO DATA PROCESSOR UNIT (DPU) DIAGNOSTIC DISPLAY UNIT (DDU) DATA COLLECTION UNIT (DCU)

FIGURE 1

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

EDS DATA COLLECTION AIRBORNE COMPONENTS LEFT ENGINE RIGHT ENGINE SENSORS/ EMUX SENSOR/EMUX STATUS PANEL NO· GO ~ PILOT FLAGS OPTION SWITCH FIGURE 2

OBJECTIVE

• OVERALL • VALIDATE SYSTEM CONCEPT , DESIGN, AND CAPABILITIES FIGURE 3 2 37 -- _ _ __ __ __ IILVll ., wf Aa OJlOll U D JltOUUT Srt C IALISf • MAINTENANCE ACTION CYCLE r--------- · I ( as I I O ... · Slll CO ... T AA C TOJIS I ADVISE I TU .. AS SIST I " OS ) 0 " AU"l AC U LN O 'N l W Il H ~ ..... kf ,-.o w CIIS 1 _ _____ __ ...J L-_-',-_-.J O CUI O OU TO l ....

fOA PH IN T or .n o u "'''-U OI " ONUS IICS FI GUR E 4

METHOD

AIRCRAFT/ENGINE ASSIGNMENT AIRCRAFT (5) 74·099 74·103 74·105 74·107 74·108 Fl00 ENGINES (11) PW E680160 PW E680639 PW E68031 1 PW E680694 PW E680330 PW E680722 PW E680415 PW E680801 PW E680470 PW E680907 PW E680528 FIGURE 5

EDS EXPECTED FLIGHT EVAL

ENGINE OPERATING HOURS e.

w :E j::: SAle 60 HRS/MO···· ~ II: w z ·· · ·sA/e 45 HRS/MO (; z III ~ I- I- 2 3 4 5 a 7 • • 10 (MONTHS) FIGUR E 6 GOAL SAle 4S HRS/MO o 2 3 4 S 6 7 8 9 10 11 12 13 14 MO . NAMES

\_ 0E8UO----\

FIGURE 7

ENGINE TIME AND CYCLE DATA RECORDS

AUTOMATED AFTO 93 " v l ·'.11..,," ["'IHE 11"f ,'"fCU A ({ ',.,'OlIII I ON REe O "' '' ~ ~ - - - . -- - - ---- . --- - - - .

TH ·t [.;,vIP Ofi D I 'lAL [VENT HISIOfll ''''I T II<1 t 10 I~A N 10 DATE liP!( £OT H S, H; ' l(.f " II """ .... ,"" ---- _ .. -- .. .. _ .. - .. .. _ .... - _ .. ...... .. .... -_ .. .. ...... .. ........ _ .... - .. .. _ .............. .. ................ .. .. .. .. _ ... - . . . . ..

I 11111 111 1 lllll llli:.l Illl) lllJJ 4 , III ./ &9 . lii4' 'I'f .l ll 4 '&IG 9 'I ~J4 'b/~I • EMf! II u [1." till114 164' tI~ '1o ; t tt ~ I ' ;j . .. ,,"') . ZI)5 [HII MH fL'l> e TC 4444 4 I ;: 14 , N .f:- a NAM E GA .. UE _____ _ 0 0 0 EN G IN E MODE HS OT8 OTC ENGINE HOTnCTIOH F T, ... £ TIMI LCf G T

o o o

(tffi)

G EJ G

~ ~ ~

LEVEL' L£Villl .... NU"L CVC L£SO

<IllD CIIID ClIID ([J]])

SEHSOR HOT OVER OVER fAil START TE .. , TEM'

o H, 0 ··8 " ··C·· 0

COMPLETE BACK SlOE FDA ALL REMOVAL ANO INSTALLATION ACTIONS PAEVIOuS foerlON IS 08S01 HE AFro FOAM ,.. "A Ii 93 ST ANDARD AFT O 93 orl 1 01." 43

FI GUR E 8

EVENT DETECTION ACCURACY

N~y

EVENTS

~

~ ./

~LEVENTS

INO- GO & M. A.I

/

V EVENT DETECTION

/

ACCURACY PERCENT

/

/ PERCENT 100 • IHITS + GOODS) ACCURACY" IHITS + GOODS + MISSES + FALSE ALARMS)

I

1978 194A 0428 OJAD OOEl CHECKSUM F IGUR E 9

SUMMARY OF

EVENTS DETECTED (TlIRIJ DEC 10) HITS FAlSE II FAlSE I MISSES ACCURACY % GOODS e 3 0 89.7 63 100&

I TOTAL

FI G URE 10 TREND/PERFORMANCE RECORD SUMMARY

I

o Trend rec.ord I> Performance r~rd

L PERFORMANCE

WINDOW r-- f-, ---f------tlI ~trtHf:n-l"t I> I> P 1>1> ~ f>.AA ~ /10 I , ALTITUDE I ~ A"h A M 1,000 FT

: LA

.o~ 4.l , / / A , ~ loA I A , A I ._

13_

---- ----

--

p

~

'- TREND WI1DOW

o.,r0(7 nJ

~ n>~O ~- o

o ~* 0.2 0.4 0 .6 0.8 1.0 1.2

MACH NUMBER IMoI FIGURE 11 TURBINE ENGINE PERFORMANCE ESTIMATION AND ITS ROLE IN FUTURE SYSTEMS Ronald L. DeHoff Systems Control, Inc.

Charles A. Skira Wright-Patterson Air Force Base (Text was unavailable at the time of printing.)

OBJECTIVES EVALUATE DATA ACCURACY, REPEATABILITY AND SENSOR VARIATIONS DEVELOP ALGORITHMS TO REDUCE PERFORMANCE DATA TO USABLE PARAMETERS ASSESS APPLICABILITY OF RESULTS TO THE ENGINE MAINTENANCE PROCESS

FIGURE 1

MIMS INFORMATION ARCHITECTURE OAT.

""",.... S MIMS CENTI" j ~

~ .... ~

FIGURE 2

PERFORMANCE DATA PROCES S IN G FLOW PATH - - --I FLIGHT DATA I TEl'IS S CM S I (DAILY) I L ________ .J

OFFLlM£ PROGRN\ 1 -

Pli /IS DA TA PROOUC TS I SE hSOR D I AGl() S TI CS I

I

SE N SOR NO I SEiS I A S I I P£RFORI'IA I4 CE TRE I4 D S I I I I SH I frS ! JLt\P DElE C II ON I I

L ___ _ _ _ __ J

---I

I ON - L1HE

I PROGRAII

I PARMETER I TRE HO IIIG L ___ ______ J FIGURE 3 ENG IN E MO DEL PERTURBED FFSET OPERATI N G POINT O 0y SET POINT YARIABLE - x y =, f O(x) , + BASELINE PERTURBATIO~ N OISE FIGURE 4 MODEL DEVELOPMENT TECHNIQUES BASELINE (fO(x)) FIT POPULATION OPERATING DATA WITH A GROUP OF CORRELATED EXPLANATORY VARIABLES PERTURBATION (fg( x) g) FIT PERTURBATION DATA GENERATED AT REPRESENTATIVE FLIGHTPOINTS BY VARYIN G ENGINE PARAMETERS SINGLY FIGURE 5 SUBSET REGRESSION - A MODEL GENERATION ALGORITHM P REGRE SS ION DATA SET /'()DELS MI~IMIZE rDDIFIED FIT ERROR P f---- P N P 2

-

L (v.- L: o · x ·· ) + ). Lw . o.2

(XiY\),i=LN

L OJ xi j

. I J I J J J j= l i=i j=i j=i MODEL FOR P = 3: ETC , FIGURE 6 TYPICAL BASELINE MODEL FIGURE 7 DATA SCREENING ELIMINATE FAILED CHANNELS • VERIFY DATA STATIONARITY N ORMALIZE POPULATION FIGURE 8 --- -- -- -- -- FLIGHT SERVICE EVALUATION DATA PROCESSING ALL TEMS DATA SCANS - 12000

I

I

STABILIZED DATA - 7000

~-----

FIRST GROUP - 5400 SECOND GROUP - 3500

L

9% REJECT 17% REJECT SCREENIN G - 3200 SCREENING - 3800 4: REJECT VALIDATION FIGURE 9 RAW lEI'IS SlABILl Z ED SC AN S ...... ... ......

... ,", . ... . .... ..... ..

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

" .. ..... .. ..

.. .

C> ...

DO ~ NON REPRfSf Nl 'ATI vr ..

.

PO IN TS ..

.... ,'" S406 POI N T~ fAN S PEED - %RPI1 FIGURE 10 TF34ITDIS SCREENED DATA SC ANS ......................

.. :::::~:::::!::::: ::u::" .~ .............. " ........................ .

.. .. . .................. ......... , .. ..

•••••••••••••••••••••••••••••••••••• I .....

.. •••••• u.u .• ,l •• "." ..... , ............ , ....... ..

•• a •••• " ................................... ..

.. " .............. " .............. , .. " ..... ... .

•••••••• " ............................. 1 ... ,. •••• 0> '" .............................................. . ...

, ••• " •••••••• , .. . ............................... i ...... ... .. ............. , .............. " ... , . ... . ....

.............. , ................. " .......... ... .. ..

··."··" •• •••• ........... ,", •• u .• " ••••• , .. ,,, .... ..

it . , .......................................... , ..... .

'" ......... ,."' .. , ........ , ........ "'''' .. , """'''',., ..

""",U, •• ""l""""''''''''''''''''''' ""''''.,''' '''' ,., ...... , ............................... .

••• • ••••• 11 • • ••••••••••••••••••••••••• ,'" ,. "'.'" ., •••••••••••••• a ••••••• ,"', •• "' •• G'l '" '''' ................................ '" .

w ........................................

"- ••• au ... "'''' •• u •••• ''' ...... ''' ••••• ,. , VI

.............................. '" .. '"

......... , ... , ........ , .... .

'" <:> ." •• " •• "' ••••••••••• ".l ••• "' ••• Vl ................ , ..... , ..... '" '" , ..

Vl "'. , ......... , , ........... , ...... .

l:!

• • •• ,"' •••••••• u.,"''' ••••••• , .I..

"- •••• , •• • ••••••••••• u .••• , :E: , ................. , .... .

<:> , , •••• , Aa ,........ "l' U <:> ••••••••••• I ., •• • 0> , " '" ".I. au., • u I 'I.... •. I.'

.............. , ...

. . "., ......

.. ....

~~8~ POINTS r),ll SPEED % RPM FIGURE 11 RE S IDUAL PLOT FOR PS3 roO£L-ALL ENGINES I· · ~------,--- ----,-------,-------I I ------ -r! -------r i------~ l ~----~ , ~----~ r --- --~ I I i I ! 0 I . 0 d ~ , 1 0 • . o . • ~ 6 ~ ~:I>\~ t ! ~_ ~ ~ . :~ ;::O ".~~ ,i ;: J'I~.<$-., ~ .. ,/..iI(!r ·° O><-o_O +- __

~---~~ " ~~~ _ ~~r:'i.J~fWfIl~r;:..t'~ ~~ . 1)Ot ""{<-.(.t . 0.6 _ 0

I i I ~ ~ 0) c. i 00 ,,~) , IS) 0 00 I,) fP;) ~ ... 0

! 0 I l 0 ~ I_ O __ ~ . j _ : ____ ! I

- 18 .

-----1- ---I I . I I

M ~

.. I ! I : i

-.28.

- - .- j ---- ; -- I -i io --- j -O J

oJ : I i I ! ! I , ~ ...

i ! i i +0 ! ! i

- l9.

'"

--T, --I- -- i -- 1 - i 1 ---- .1 ---- 1

I ! IIi l

i ! I ,I Iii i

-48·1·hr----~ i ~-----r.hr. I -----onnr. ! -- --"F<nr. -- ---r~ ~--"F<~ · -----ri. ' ,.-----~ · .._ ---- "".._----li .-: 14 49. _ t. ee . )d~ . Le.

PS3 - PS IA FIGURE 12 fAN SPEED RESIDUAl SHOI1ING SJIORT TEII1 TREND

I

, . 8 --- , ...

z ~ « ~ - . I " '" ..

TWI WATER WA SH FIGURE 13 PS3 IllDEL RE S IIlUAL SHOIj I KG LOtIG TERII TP.EHD ...

2 . 8 if ~ - Z .8 0< '" - 4.8 FIGURE 14 SENSOR VALIDATION ALGORITHM • USES ENGINE MODEL DERIVED FROM OPERATING DATA INCLUDING CHANNEL ERROR VARIANCE • USES RANGE CHECKS FOR HARD FAILURE S • USES CHANNEL ERROR STATISTICS FOR SOFT FAILURES DETECTS MULTIPLE FAILURES WITHOUT FAULT TREE ESTIMATES FAILED CHANNEL READING FOR SUBSEQUENT DIAGNOSTIC UTILIZATION

FIGURE 15

SENSOR VAll DATI ON FLmlPATH N O VALI D A T ION N O T PO SS ISL E I-- -~ FLA G OIi T - OF- RAN G E CURRENT ENGINE IIODEL 0 y NO S ET FAILE D y TO EN D f (y , g )

FIGURE 16

HERLTH PARRMETER SHOWING SOFT ITT FRULT < •• I I I I I

I~\

.. ~

... ! I - - - -- - -- ITT S f:NSOa OFFSr!T (_ .. on e ) i ?

Z . O - - - - r -- - -- - •

)

O .

o 0 • 0 -- 0 ;:'-2 .• -- - -- • 0 ~: f o~.o

. ·l·w... - --

°:8

0 , ftl0 o., ., 0 ' ~~1.&8 ~ ... ~ t .. .

o~ ., ,0 o : ~. ~o:> 0 .... 0 _ 6'00 (' ('60 0 ., !- -- - 0 0 0 o 0 ?,.r.; o 0

~ "]

:~, t ~~.J Jj~

- .. ..

ri~Ir.. .J ~ o . . ....... . ., • 101 IIW ~ I <" ' ... -r...;. H..:c:E=rR..:c:L=-T=rH -=-...: P '-r R c:.. R c:.. R:..rI c:..· IE ~ T ::. E :..: R -.=-S..:c: H -T 0 ..:: ~J ..:. I .:.r ~I ...:: G -----TS...:: O ..:.. F-oT~I '-T T ....: T -r F --'- H ~ U =-r L =-- T :....: S=.---,

I I I I I v-::p ... "

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

ITTSF. HSO RO nSnr- 4n " C) '/ °v ~ -

"°of ---+- - -+-- - _._ -+- - - -1 '- --j- - ____ +_ 1 __ ~~~> VI

o E-z .• -. - I . i 1f.'}1\ o f .

'b 0 ., 8., :.0 rf'.~ ~ , ~ ~ _ Ita <Po 0 o ~ , '0: • ~ 0 0. fI 0 ~ <>;0' 0 0, 0 ., -- .

. 1:: ... ... . .. J:;<,"( . ..

- • .• 7 :h .•• - --.1- - ., . - rl- .... . -.-- ...-!- .• ..--..,l,.- . .... .. . -.Jh .• ...-.-l .• ~ ;':~~ . o .. . ..... . . 18 .n ..

FIGURE 17 "._---y--.:... P;::.S..::3 - - ..:.. P --'- T T 5 :.......:: C:...:Oc:..R :.; R :..: E:.:L :.:. R .:... T :....: I ;..: O :..: ~ ..:... ' ....: F.....:O ;: R ..:..' .....: l=..5=..2 ::.,. 3=..8=-- _,-- __ .., .. .

.. . -- - - - ---

" . - - ----- ,. . ------ <S.

<0.

~ ~-' ---

rTS 01 r ~ a : r ... 0 :N J' __ .. _ - -

" .I

n ~,~.r.-- rtr --oJ ll-r- . - - Ti;~~i) .- ---, .... -- -. ~r .-- -.h-- - ... L

" S ,lr S IAI FIGURE 18

J

.. . __ .--..:.. H '-T 0 :.:. R .:.: N .:..: A T l =--- P :...,=. S -=. 3 ---T P ....: T ~ 5 :..,..: C :..: O :..: R .:; R .:..: , E =.: l =-;.:. A..:..T ..:. I .:; O ..:: H:....- .--,..---"'T- -' -

I . ,LF ·1

... - -l - --i---t---1 I -' - -- i- -i---- l- - ji7 0': .- - -

,,-

- - J.- - .-- ·- 1 ,-- - - .. _- -- - - it_ A --- -- '.

0" ", ' -- -\- -- -1-- - .. -. ----

... -- "","" ·-- 1- 1 - 1---1- '-

<e • . - - - -I-- - , . !!. - - -- -- -.-- --- - -- - - - -. -- >0 .' - - -f- -" I ~ - - - - - -- 1--- -- ---- ..

-1 -- 1-- - -- ,- - - >e . --= + - ~ --- -- .-- .. .

, ~-.mr-.ir -.~r-.~ -,ir-. tr; - " - . ,~,, r i - . ,Th ' - ~-~- Yh -~- ~ h -, ~ •.

~-,~~~~~ C ~ 0 '-T R .:.: R~ E ~ l ~ A ..:.. T ..:. l r O ~ H WI TH SOfT fAilURES .. .

.. .

.. . ,.~ .~ h-~ ~ -rlr- ~ r- ~-dr-. ~~ I r--rir- ~ r-~r- ~ -.dr~ . r.-z •.

,. ~, ' U'SI'III ' FIGURE 19 SOFT PLR FRILURES ..

I

., I

.

• .

.

.

.

3Jo P LA O IANNEL FA ULTS • 1 - 0 ---

.. 0'-

- n ---

• II ~ ---'"

· --~ f -- -- ---

.

.~ '" .

. o·

"1 :~

_.- ..

a. . -(J

. i - - - .

~ .

";' 0 . .

oS> 70 .e --- - ---- - --.0

-

. .

.#

.. _- . ~

._ - -- -- . ~ .

0 '10

- +- 1 -

7 < • . , ~ . ., ..

6 • • 61 •• .a .a . • , .! .. .. .. I 5.

, ." .

FIGURE 20 PERTURBATION MODEL GENERATION • SELECT REPRESENTATIVE FLIGHTPOINTS • RUN SIMULATION IN NOMINAL CONFIGURATION • PERTURB COMPONENT DESCRIPTORS THROUGH EXPECTED RANGE OF VARIATION • CALCULATE PERTURBATION FROM BASELINE • FIT SIGNIFICANT VARIATIONS AS (POSSIBLY NONLINEAR) FUNCTIONS OF SETTING PARAMETERS

FIGURE 21

PERTURBATION ENGINE MODELS -l UJ z: z: « :::r:: 6_ ENGINE PARAMETER U a:: a U) z: UJ U) SET POINT VARIABLE - X PERTURBATIONS ARE LINEAR ABOUT NOMINAL • PERTURBATIONS ARE DEPENDENT ON SET POINT • PERTURBATIONS REFLECT COMPONENT CONDITION AND ARE ADDITIVE

FIGURE 22

TYPICAL PERTURBATION MODEL

FIGURE 23

ENGINE PERFORMANCE MONITORING • DIRECTIONAL VS. DIRECT INVERSION • MODULE DIRECTED RATING PARAMETERS • TF34 PERFORMANCE ESTIMATOR

FIGURE 24

APPROACH ES TO PL R roRMNCE ffiH ITOR I HG t'ODEL IHVH1 SI0It • S ItIGl£ SCAN ! ~ (SHAPSIIOTl

c=:)

-v .y ~ V - Yo • = n( v) Ay • MULT I PLE SC AN S (FILTER)

}

D

HARDWARE DIRECT 1000~L INV ERSION DIRECTED "FAN - t COPE t

•

- - DCORE fAIl

•

•

- - HPT + HPT

• - -

- H 2 S (N

•

TRIM t + +

•

FIGURE 25 PERFORMANCE ESTIMATOR • USES VALIDATED DATA SCANS • AVERAGES OVER FIXED WINDOW CALCULATES PARAMETER ESTIMATES AND UNCERTAINTY FIGURE 26 PERFORMANCE ESTIMATION FLOWPATH TEllS ST ABILI ZED SC AN S ( VALIDAT ED) NO Of/VERGED ? >--- ---'

FIGURE 27

PERFORMANCE ESTIMATOR MQD.El : y = He + v A A ESTIMATOR : + PI e 1+1 = e I "I -I -I + HIT R- DISPERSION : = (01_1 HI) PI I HtjOVllTiON : = Y I - H e I "I

FIGURE 28

MODULE DIRECTED PARAMETERS • EFFICIENCY AND FLOW AREA REPRESENT INDEPENDENT CONSTITUTIVE (I.e.} MASS AND ENERGY) PARAMETERS • THEY DO NOT REPRESENT INDEPENDENT FAILURE MODE S • ALL FAILURE/DETERIORATION MODES ARE COMBINATIONS OF INDEPENDENT PARAMETERS • PARAMETER ESTIMATION CAPABILITY IS LIMITED BY SENSOR SET

FIGURE 29

COMPONENT FAILURE MODES EROSION/CORROSION >- LJ z: W LJ u...

u...

w FOD

REBUILD ____ . ~::::::==----- --------- •

.- - -;:.. .;#'"'// ... -// I FLOW AREA // 4{'/ I /

I WATER WASH

FIGURE 30

MODULE DIRECTED RATING PARAMETERS • SELECT TYPICAL FAILURE DIRECTIONS • CALCULATE MOST EASILY ESTIMATED PARAMETER EXAMPLE; DECREASES AS LOW SPOOL COMPONENT S DEGRADE FIGURE 31 SIMULATION RESULTS • 350 HOURS OF E5186 TEMS OPERATING POINTS • SIMULATE SENSOR NOISE LEVELS • SIMULATE HARDWARE FAULTS AND TRENDS FIGURE 32 S !t · 1ULATED FAULT TREND : .. , :.

I

- ..

~~ I _ .

if'--

- ,Ie . ~ ._ - ._--

I -. -.- 1- . /" '-./

'"

~N ~"'F

1\

- ...

--I- r- ' ~ ,,~ ........

I

. . 1:-- - .11 -- - -- -

r -/ -"- "'--I"--- - -

, \

/ t- "'-

_ Oli~ ___ I --- t- __

._ ,- "'- .82 I~ r :I ..

. \ .. - ~ - olZ I ' - - . '- v o o ..

~f

- LI --

"'- . 83 e - ... -- - 'C ..:::

'-- T-

- =t

'-, -1 - . 81 5 -

--- 1 '- 1 '-

-- l- -,

- ., .. I

. ., ., .. , ..

• • <40 • ~8. 646. \4 , 8. 666 . ~B. 62111 . ,. 18 . 66B • rz • . ... . .

. . '

• TOT tHRS) FIGURE 33 SIMULATED FAULT EVEtlTS .01

I

-...

.I' ..

'- /

l \

~COM RES ~OR - .1.

- - - I- --- -' - \ ---- -

1-

~ -" ~- \ - . 12 ..J. I.

E- r--

fAN '"

~K

~f-.-/ I--

-

""- --

-. '3 - -

- 1

~ -

i" 1 "-

_ V-: _ - .84 l:IPL ,- -- • -~

~ r --

I

I I I

I , ... . . , .

I •• 4211 • ... ~8. 648 . 61 ;8. 598. ~8. 648. 61 . 6U8 • . .. 6 e.

fOT (HltSJ FIGURE 34 WA TER WA SH WAT ER Wn SH FIGURE 35 FIGURE 36

J

' .0 -- l ---

2. !h-. --rlr- . -do. c-. -.±..------.ri:.~ h--..+..__ . -.h- , .---.4, .----.:h--rk.--,-h,~ , = ,. ---; ,,!.,:o:.---ol l .;eo e rOT IHP. ll

FIGURE 37

30.01 --+- --1 - --+ - 25 '8-'h--- . -.rh- . ----z+. • .-- . .....h-~..-"ob---xh- . ~ . -. ....-h- ..........br---.-J~ ~--,.,h. , ----d" ~ I 13 8 TOT (HP. l'1

FIGURE 38

_J

SUMMARY • PERFORMANCE r· 1ODEL DERIVED FRO~1 OPERATING DATA AND STATUS DECK • MODULE DIRECTED PARAMETERS SELECTED FOR LOH AND HIGH SPOOLS SEN S OR DIAGN OS TIC S VALIDATE S DATA PRIOR TO • PARAMETER E ST WATI ON E S TIMATIOII ALGORITHM U SE S FIXED TI~lE WIrm ow FIGURE 39 IMPACT OF AUTOMATED ENGINE MONITORING ON RELIABILITY CENTERED MAINTENANCE AND LOGISTICS SUPPORT Laura E. Baker and W. Earl Hall. Jr.

Systems Control, Inc.

(Text was unavailable at the time of printing.)

DEFINITION OF INTEGRATED SYSTEM • MAINTENANCE SUPPORT SCENARIOS UNDER OCM - TROUBLESHOOTING REOUIREMENTS - PREVENTIVE SPECifiCATION REOUIREMENTS • SYSTEM LEVEL REQUIREMENTS • INfORMATION INTEGRATION REQUIREMENTS FIGURE 1 DECISION PROCESSES FOR ENGINE MAINTENANCE FUGHT UNElJDM FIGURE 2 L_ ENGINE HEALTH INDICATORS/INTEGRATI.0N REQUIREMENTS USAGE FAC T ORS

•

- TIME - CYCLES EVENT I E XCEEDANCE DETECTION

•

VIBRATION

•

PERFORMANCE MONITORING ITRENDING

•

OIL ANA L YSIS

•

MAINTENANCE HISTORY

•

CONFIGURATION STATUS • FIGURE 3 IDENTIFY FAILURE /(JOE CHARACTERISTIC FRO M MA INTE N ANCE EXPERIENCE FEASIBL IDE~\\ fY >-"" NO '-- ___ ---, INCIPIENT ?

Y ES LIFE USAGE • ESTABLISH LI M ITS PRACTICAL ' & > -""-- --i. IMPLEMENT ACCOUNTING COST PROCEDURES EFFECTIVE ?

YES SATISFIES CO~~~~iNT >-...::. NO"-- ___ _ -' ?

YES ON ·CONO ITI ON RELIABILITY CENTERED • INSPECTION PROCEDURES MA INTENANCE • MONITORI N G TECHNIQUES P LAN • D A TA IN TEGRATION FIGURE 4 26 4 - - - - - - - - - . -- - -- - - - - -- - - ---- ---- - MIMS DEVELOPMENT BACKGROUND

tJ

TlJIlOlHf ENGINE fAUl.T DEJECtION AHO ~ ISOLAT ION PIIOGRAU ~~I

FIGURE 5

TEFDI-A FOUNDATION FOR AN ENGINE MAINTENANCE INFORMATION MANAGEMENT SYSTEM (MIMS) 8~\Il~n • _ORMAtION/fORMAT • DAtA fLOW • ACCESS CAPAlllLIJY ~ !~ l!~~ n !!§T!!;~ • EHGIHE GENERIC • MOOULAR SOFTWARE: • IHIEGllAl ro 6YSTEM ~~ffi~ • GHAPttIC 0ISPl.A Y • IoIlJI.lI LEVEL ACCESS • MAHAGEMENT BY UCEI'TION

FIGURE 6

MIMS INFORMATION ARCHITECTURE D AtA • MAlNTIEJotAMCE S YS I EN S • SOAP MIMS CE N TE R • P AR" TRAC .r. 1HO FI GURE 7 REQUIREMENTS • REDUCTION OF PERFORMANCE DATA TO CONCISE, USABLE PARAMETERS • SYSTEM OUTPUTS CONSISTENT WITH THE DECISION PROCESSES OF MAINTENANCE PERSONNEL AND LOGISTICS ANALYSTS • INCORPORATION OF STANDARD ENGINE USAGE FACTORS (E .G., LCF, TIME, ETC.)

• WELL-DEVELOPED PlAN FOR INTEGRATION INTO MAINTENANCE/LOGISTICS OPERATIONS FIGURE 8 __________ ~~ __ ~ _________ L.._.__ ,.)

MIMS SOFTWARE DEVELOPMENT • STRUCTURED AUTOMATIC DATA ACQUISITION, PROCESSING AND TRANSFER TO PERFORM GAS PATH ANALYSIS OIL ANALYSIS VIBRATION ANALYSIS COMPONENT LIFE USAGE • ADP ARCHITECTURES, DATA BASE INTERACTIONS COMPATIBLE TO CURRENT AIR FORCE LOGISTICS ORGANIZATION • SOFTWARE MODULES TO PROCESS/ANALYZE DATA AND SUPPORT MANAGEMENT INFORMATION SYSTEM FIGURE 9 MIMS DATA BASE SOFlW ARE • MODULAR DESIGN • RANDOM ACCESS FILE STRUCTURE • CURRENT DISPLAY/SOFTWARE IMPLEMENTATION BASE STATUS SUMMARY USER INTERROGATION - FIND/RANK WATCH STATUS - USER CONTROLLED ENGINE PROFILE HISTORY -LINEAR/POINT PLOTS AND TRENDS TABULAR MAINTENANCE EVENTS MMICS INTERFACE • SOFTWARE IN CONCEPT DESIGN/DEVELOPMENT ALARM - GPA ALGORITHM (SNAPSHOT/TREND) FIGURE 10 MIMS DATA BASE \-4---\ SUUloVJ1Y _TEIASDATA DATA FILE USER ~~ ___ _ INDEXED DATA BASE t-.--- MMICS FILE ACCESS IooIANAGER DATA ~---SOAP FILE ~---Io4DCS FILE FIGURE 11 MIMS IMPLEMENTATION PROGRAMS PROGRAM OBJECTIVES SCOPE MIMS/EDS • EVALUATION OF MIMS liIIRDWARE • ACQUIRE DATA FROM FIVE EOS EQUIPPED AIRCIIAFT AND SOFTWAflE INTEIIFACES AT LANGLEY AlA FORCE BASE - F 100 EDS • VAUDATION OF 1 EFUI GAS - SOAP PATH ANALVSIS ALGORlTIIM - MAINTENANCE EVEN TS ON EOS GE NERAl EO DATA - PARTS TRACKING • TRANSMIT DATA TO MIMS AND ANALVZE RESULTS WITIIIN TEFOI ODJECTIV ES • BARKSDALE AIR FORCE OASE INSTAllATION OF MIMS MIMS/ITP • OEMONSTRAliON OF MIMS DATA PRODUCTS DATA STATIOUS, DATA AQUISITION PROCEDURES, AND COMPUTER INTEIIFACE LINKS • PIIELILIINARV EVALUATION OF .. I .... THE-FIEU) " INSTAll.ATlON • TWO A-IO/ TF- J" TIP AIRCRAFT • OEMONSnlATION OF TIIREE LEVELS OF DATA PRODUCTS AT BASE AND flEM01E SITES • DATA ASSIMILATION FROM 24 A- IO AFAES STEMS/MIMS • DEMONSTRATION OF SQUADRON SQUAOAON lEVEL STEMS/MIMS AIACnAFT AT BARKSDALE AlB FORCE DASE EVAlUATION OPERATIONAL EFFECTIVENESS • AIR FORCE OPERATION OF LlIMS • BASE LEVEL COMPUTER UTILIZATION FIGURE 12

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

FUTURE IMPACTS / CONCLUSIONS • PILOTlFLlGHTCREW • FLIGHT LINE I AMU • INTERMEDIATE SHOP • DEPOT • COMMAND FIGURE 13 OVERVIEW

I

I

I

• DEFINITION OF INTEGRATED CONCEPT

I

• EVALUATION CAPABILITIES

I

• IMPLEMENTATION REQUIREMENTS I

• FUTURE IMPACTS I CONCLUSIONS I

I

I

I

I

FIGURE 14

I

I

I

I

I

I

I

I _ _________ _ __ __ J

Page intentionally left blank

------- ------ - - - - ---------------- - A-10/TF34 TURBINE ENGINE MONITORING SYSTEM (TEMS) Robert G. Christophel San Antonio Air Logistics Center SUMMARY The A-10/TF34 Turbine Engine Monitoring System (TEMS) integrates inflight and ground hardware to sense, signal condition, perform computations and analysis, and record various engine and aircraft information and parametric data for the purpose of fault detection, isolation and trending. Basically, the data are collected, processed and stored by the airborne Electronic Pro- cessor Unit (EPU) then transferred through the GO/NO-GO indicating Umbilical Disconnect Unit (UDU) to the Diagnostic Display Unit (DDU) for flight line maintenance use before final transfer to the TEMS ground station peripheral equipment for Jet Engine Intermediate Maintenance (JEIM) shop use, actuarial processing and permanent storage. If flight line display and use of the data is not required, transfer to the ground station may be done with the Data Collection Unit (DCU). TEMS data will be used at the flight line to assess engine GO/NO-GO status, aid in troubleshooting and fault isolation and to per- form engine trim. Potential JEIM and depot TEMS information uses include engine troubleshooting and fault isolation, test cell trim and data collection, maintenance programming, parts tracking, spare parts forecasting, and actuarial analysis.

INTRODUCTION In 1974 the Department of Defense (DOD) adopted the Reliability Centered Maintenance (RCM) concept for all military aircraft systems, consequently requiring restructure of existing aircraft scheduled maintenance programs and establishment of RCM programs for all new aircraft. The DOD RCM concept is based on the commercial airline maintenance decision logic called MSG-2 developed by a committee known as Maintenance Steering Group 2 composed of representatives from the commercial airlines, Air Transportation Association, and Federal Aviation Administration (ref. 1). Basically, RCM is a decision logic process which divides scheduled maintenance requirements into the three basic categories of hard limits, on condition, and condition monitoring, followed by a Maintenance Requirements Analysis that translates the maintenance requirements into specific inspections, limits, tasks, and work packages and produces technical data and instructions for the maintenance of a specific system (ref. 2). The United States Air Force (USAF) incorporated the DOD RCM philosophy into an expanded On Condition Maintenance (OCM) concept, defined as maintenance that allows the condition of the equipment to dictate the need for maintenance or the extent of repair/overhaul required (ref. 3). Successful conversion to full OCM for a complex turbine engine requires the use of monitoring systems such as the A-IO/TF34 Turbine Engine Monitoring System (TEMS) (ref. 4) .

The fundamenta l success of OeM is directly dependent on the ability to adequately perform the tasks dictated by the nature of the three ReM cate- gories, continually assessing the OeM data and updating the RCM analysis by transferring items from one category to any other as necessary. The hard limits category requires parts time and cyclic tracking; on condition generates the need for repeti tive inspections or tests; and condition monitoring is greatly enhanced by diagnostic and trending capability. The USAF has always practiced OCM to a certain extent with these functions satisfied by a variety of manual and automatic data acquisition systems. However, recent radical developments in microprocessor technology and data processing have made possi- ble completely automated systems capable of acquiring and processing the vast amounts of data needed to support the oeM of a modern, complex turbine engine (ref. 5).

The TEMS being incorporated into the A-IO/TF34 system is design ed and built by Northrop Electronics Division and was originally flown on the T-38/ J85 combination before being upgraded for the A-IO/TF34 application. This paper discusses the operation and interfaces of the A-IO/TF34 TEMS hardware focusing primarily upon function, capabilities and limitations . The TEMS data types are defined and the various data acquisition modes are explained.

Potential data products are also discussed.

SYMBOLS AND ABBREVIATIONS ITT - inter turbine temperature NF fan speed NG core speed PLA - power lever angle P - compressor discharge static pressure S3 RPM - revolutions per minute T - compressor inlet total temperature 2C VG variable geometry W fuel flow rate F HARDWARE AND INTERFACES System The hardware used in the A-10/TF34 TEMS, (Fig. 1), is comprised of in- flight and ground equipment to sense, dignal condition, compute and analyze, I

I

___ J

record a nd st or e va r i o us aircraft and eng i ne information for fault detection, isola t ion, diagno stics, trending, and parts t ra cking (Fig. 2). T ne basi c co m po nen ts ar e the airborne Electronic Proc e ssor Unit (EPU) and th e grou nd used Diagnos t i c D ispl ay Unit (DDU) and Data C ollection Unit (DCU). These u ni ts are m i crocomputers that share common compo n ents and are based on 8 080 micropr ocessor ar ch itecture. The EPU, Umbilical Disconnect Unit (U DU) , senso r s and sig nal conditioners, and associa t ed wiring make up the airbo rne hardw are . The ground equipment consis t s of the DDU , DCU, a printer, Intel li - gent Dis k U nit (IOU), and telephone modem .

In ope ra tion, th e EP U (Fig . 3) continuously receives and monitors sensor and t ra ns duc er di gnals (Fig. 4) and records and stores a data frame autom a tic- ally for presel ec ted flight conditions or whenever the pre-establi shed normal lim i t s of a critical parameter are exceeded. Data frames are manu al l y ta k en and st or ed u po n pilot command through a cockpit data switch or for mai ntenance reco rd pu rpos es through the DDU (Fig. 5). Data sto r ed in the EPU i s re t ri eved on t he gr oun d by either the DDU or DCU through the UDU (Fig. 6), whic h also provides GO /NO-GO and limit exceedance event indicators. The DDU has r eal- time di splay and o peration capability and provides maintenance pers o nn el with a di s play of engine perf o rmance parameter s , operating conditions, and othe r infor m a ti on permitting review of routine data and troubleshooting/ d ia gnost i c capab i l it y at the flight line. Engine trim functions can also be done using th e DDU i n dep end e nt of other test equipment. The DCU is essentia l l y the same as th e O DU wit hout di s play capability and both un i ts transfer data to the pr int er and IDU intthe Jet Engine Intermediate Maintenance (JEIM) shop for perma nent storage and further troubleshooting, fault isolation and di ag nostic activit y as r equired. Reference 6 contains a complete, detailed des cr i p ti on of the ha rdw are and its op eration .

Electronic Processor Unit Th e EPU provides central administration, execution and regulat io n of the TEMS. It continuously receives and moni t ors inputs from aircraft an d engine transducers and sensors and performs various functions relating to the signal co nditio nin g, proce ssing a nd storage of the data. The signal cond itioning func t io n conver ts the sen sor si g nals into scaled direct current valu es. High impedence iso l a tion between the sensors and conditioners protect s on-boa r d i nstrumentat i on, all o wing the use of existing aircraft instrumentat ion without affecti ng the c oc kpit indica tors. After conditioning and multipl ex ing, the signals ar e di git ized by th e Analog to Digital Converter and inp ut to the processor . The process or is the computer portion of the EPU and uses both Random Access Memory ( RAM) and Programmable Read-Only Memory (PROM ) . It constantly m oni t o rs and processes the data and, when a maintena nce actio n item has been con firmed, transmi ts the appropriate information to data stor age for ground recove r y. Th e PROM stores the executive routine, equati on su brout i nes, diagno stic log i c , si gn al ave raging and instructions. Program co nstant s , cali- brat i on data , engine signat u res, threshold levels and logic option s are stored in the RA M . The RA M also is the working memory and provides temporary data storage f or ground r e trieval. These memories can be progr ammed through the DDU to ac count for engine changes or limit changes without r emoval of the TEMS hardware . T he pr oc essor also provides interface control for EPU communications through the UDU to the DDU or DCU.

Umbilical Disconnect Unit The UDU is mounted in the A-IO nose gear storage compartment for easy access and provides the capability to retrieve data from the EPU, to display aircraft and TEMS status and event mode indicators, and to enter mission con- figuration information for structural tracking use. EPU data is transferred automatically by connecting the DDU or DCU umbilical to the UDU and depressing the Data Transfer button. Aircraft and TEMS status indicators include red/ yellow/green light indicators for NO-GO, Caution, and no limit exceedance events stored in the EPU, respectively. If the NO-GO or Caution indicators are lit, additional information is available in the form of a four digit alpha- numeric code, displayed upon command by depressing the status button. This display also indicates TEMS malfunctions.

Diagnostic Display Unit The DDU is a one-man portable microcomputer unit that communicates with the EPU through the UDU to transfer EPU stored data to the DDU for flightline maintenance use and/or further transfer to the peripheral ground equipment for printout, permanent storage, and processing. The data transfer is simply and expeditiously done and includes automatic data validity checks. The Light Emitting Diode display capability of the DDU provides for flightline review of routine data as well as plane-side troubleshooting and fault isolation when desired, and the performance of engine trim functions independent of other test equipment. The keyboard is used to re-initialize and calibrate the EPU follow- ing an engine change or as required by other maintenance action. The ' DDU micro- computer is the same as that in the EPU and various modules are interchangeable.

Data Collection Unit The DCU performs the same data transfer function as the DDU but does not have the display capability for flightline data review. The computer and data transfer elements are identical to those in the DDU but the elimination of the display section and part of the power supply results in a much smaller, lighter unit weighing approximately eight pounds that can be easily haRdcarried whereas the DDU is usually bicycle transported.

Peripheral Ground Equipment This equipment consists of a Tally T-1612 Printer, a Northrop 094020-301 Intelligent Disk Unit (IDU), and a Vadic VA 3451 Telephone Modem. This equip- ment provides for permanent hard copy printout for file records and analysis, permanent magnetic floppy disk storage, and transmission of the TEMS data to a central site or more encompassing data system such as the Comprehensive Engine Management System. The IOU has computer logic and programming capability and can provide a variety of printed and plotted data for diagnostics, trending, I

I

I

-- - ~---~--~- __________ J

life usage, and maintenance planning purposes.

DATA ACQUISITION Automatic Data Collection During normal flight and ground operation, the various sensor and trans- ducer signals are continuously monitored by the EPU. However, the EPU only records, for ground retrieval, a data frame whenever specific, preprogrammed conditions are satisfied or when commanded by the air or ground crew through the cockpit switch or DDU. There are two classes of preprogrammed, or auto- matic, data frame recordings: trend data frames and limit exceedance data frames (refs. 7 and 8).

The purpose of the trending data is to obtain operational flight data for comparison with previous records to detect changes in engine performance, collect parts life tracking and usage information, and provide actuarial documentation data. A large supply of data points usually enhances trending accuracy and confidence, but recovery, storage, and analytic capacity limitations restrict the amount of data that can be processed. This has re- sulted i n the present procedure of two trendinq data categories, each of which may be taken a maximum of once per flight (Fig. 7). The "Liftoff" frame is taken once each flight and consists of the last data scan monitored just before the weight-on-wheels switch indicates liftoff . The "Cruise" frame is taken later in the flight, after satisfying the stability conditions necessary to ensure repeatable data, valid for comparison purposes. Data is taken the first time the stability conditions are met, with no repeats during the sortie. The l l stability parameters associated with the "Cruise data frame are elapsed flight time, PLA, NG: T2C' gun firing, airspeed, altitude, angle of attack and vertical acceleration.

The purpose of the limit exceedance data frames (Fig. 8) is to report ab- normal engine operation and to provide supporting data for troubleshooting and fault isolation. The parameters triggering a limit exceedance data frame are ITT overtemp, NG/NF overspeed, oil pressure, vibrations, variable geometry schedule, NG RP~, compressor stall, slow starting, fuel filter by-pass indica- tion, over-g, and fluctuations in oil pressure, N , N : W and P ' S3 G F F These parameters were selected from studies of historical failure records, maintenance impact, and detection reliability. When possible, existing USAF Technical Order (T.O.) limits are used for the detection criteria but, in those cases where no T.O. limits exist, reasonable values were determined and assigned through consultation with General Electric Company, the designer and manufacturer of the TF34 engine. In operation, a limit exceedance data frame is recorded upon the initial detection of an out-of-limit parametric value.

Data frames are not recorded for a succeeding limit exceedance of that particular parameter but the number of occurrences and total duration of the limit exceedance for that parameter are accumulated and stored for retrieval.

Of course, an out-of-limits event by any other parameter will produce a recorded limit exceedance data frame.

Manual Data Collection Data frames can be manually taken for record purposes by depressing the cockpit switch or upon command through the DDU. The primary purpose of the cockpit switch is to allow the pilot to record data at his discretion to document abnormal or unusual circumstances. A one second depression of the switch produces a data frame. Continuous activation of the switch results in a new data frame every two seconds. Data will be taken by maintenance person- nel using the DDU for record purposes during engine trim, engine maintenance or EPU calibration.

DATA Diagnostic Display Unit The DDU displayed data is categorized as documentary, measured or com- puted . This data provides maintenance personnel at the flightline and JEIM shop with engine trim data and troubleshooting, fault isolation, and trending information for performing engine maintenance.

The documentary data includes aircraft and engine serial numbers, fl i ght and record numbers, Julian date, record time, elapsed flight time and flight condition information. The data is primarily for actuarial, record and classification uses.

The measured data consists of the output from each engine sensor. This includes the detected event limit exceedance data and special diagnostic indicators which provide spool differentiation for vibration data, aircraft modes such as slat deployment and out of envelope conditions, and instability information.

The computed data is composed of the results of calculations concerning trim and performance verification. The trim relationships verify airborne and ground fan speed trim, trim margin, variable geo m etry schedule , and idle trim.

The relationships are corrected for bleed air, power extraction, Mach number and droop and, although the airborne checks are valid at part or full power, there are engine minimum speed, maximum pressure, and altitude li~itations.

The performance relationships have been identified as being effective in measuring specific characteristics through sensitivity analysis pertaining to engine degradation and performance changes.

JEIM Printed Data All data displayed on the DDU is available by printout in addition to backup data including corrected parameters, calculations, cumulative times, aircraft parameters and fluctuations. Also, information including ITT time above 790 C, ITT time above 810 C, and temperature, . fan speed, core speed and compressor static pressure cycles are presented for special parts tracking, life usage, and actuarial functions.

DATA PRODUCTS Over the period of the past few years, the USAFAero-Propulsion Laboratory and Systems Control, Inc. (Vt) have been investigating the integration of various data sources, including TEMS, into the USAF maintenance/logistic process (ref. 9) with the objective of developing procedures for reducing and processing raw data elements to provide maintenance decision information to the flight line, JEIM shop, depot, and major command level. The raw data includes maintenance action records, oil analysis results, configuration tracking, and TEMS data. These data are processed into a data file, ranked and sorted, and stored for subsequent access.

A preliminary set of data products have been identified for various user levels. These include summary reports of the operational status of the engine population by base location including such pertinent information as Time Compliance Technical Order completion, spare engine availability, engine and component life data and usage trends. Also, reports for individual engines could be generated with the same type information in addition to maintenance history, oil analysis data and trim and performance trends. Documents pre- pared specifically for depot and command level use could include a wide variety of actuarial information, parts tracking and forecast usage, fleetwide distribution of maintenance manhours expended for specific failure modes and general fleetwide engine health trends (Fig. 9).

CONCLUDING REMARKS The A-IO/TF34 TEMS hardware and software development is virtually complete and, from the viewpoint of a qualified system, the TEMS is now ready for in- corporation to the A-IO force. However, before total retrofit is done, it i s necessary to fully develop, validate and establish engine maintenance and management procedures based on TEMS data and to integrate the TEMS data into the mainstream of the USAF maintenance and logistics process . A pilot program is now being initiated with that objective. The program will consist of one full squadron of A-IO aircraft equipped with TEMS and will be done in con- junction with the Comprehensive Engine Management System Increment IV proto- type. This will provide for both the development and evaluation of new or modified A-IO/TF34 maintenance procedures, capitalizing on TEMS technology, and the engine management data products necessary to provide the basis for a composite, total On Condition Maintenance system for a modern, complex turbine engine.

REFERENCES 1. Reliability Centered Maintenance Analysis Course, User's Guide. Air Force Institute of Technology, Wright-Patterson Air Force Base, Ohio, Undated.

2. Reliability Centered Maintenance, On Condition Maintenance - What Do They Mean. Staff Briefing by HQ AFLC/LOP, Wright-Patterson Air Force Base, Ohio, Undated.

3. Equipment Maintenance Policies, Objectives, and Responsibilities. AFR 66- 14, United States Air Force, Washington D.C., 15 November 1978.

4. Report. USAF Scientific Advisory Board Ad Hoc Committee on Turbine Engine Monitoring Systems, 28 November 1980.

5. DeHoff, R.L.; Baker, L.E.; and Hall, W.E., Jr.: Impact of Automated Monitoring on Engine Operations and Support. AlAA/SAE/ASME 15th Joint Propulsion Conference, June 18-20, 1979, Las Vegas, Nevada. Paper No.

79-1276.

6. A-I0/TF34 Turbine Engine Monitor System (TEMS) Phase I Final Report.

NORT 80-244. Northrop Corporation, Electronics Division. May 1980 .

7. Software Definitions Turbine Engine Monitor System (A-I0 Aircraft). NORT 79-14B. Northrop Corporation, Electronics Division. October 1979.

8. Software Definitions Turbine Engine Monitor System (A-I0 Aircraft). Update of NORT 79-14B. Northrop Corporation, Electronics Division. To be published.

9. Baker, L.E.; DeHoff, R.L.; and Hall, W.E., Jr.: Turbine Engine Fault Detection and Isolation Program - Phase I Requirements Definition for An Integrated Engine Monitoring System, AFWAL-TR-80-2053, Volume I. Air Force Wright Aeronautical Laboratories, Wright-Patterson Air Force Base, Ohio. April 1980.

USAFTEMS HARDWARE APPROACH TEMS STATUS PANEL ENGINE SENSORS' CONTROLS

'~~~, ~~r . i 41] DIAGNOSTIC

~--===--~~ '~~, ELECTRONIC PROCESSOR DISPLAY "'" UNIT (EPU) UNIT '~"~'....""'" O ,,~ :. :h . (DDUI '~, tl~

DATA COLLECTION '~~::: -=.::.::. -=- ~ -_-_ --~ . t 'I ~ . 'l _ . . 11 l~

UNIT (DCU) , _:) GROUND STATION FHlI( 1

A-l0/TF34 TEMS

OVERVIEW OF ENGINE MONITORING GATHER DATA ACTION :

------T - ------T------

,.&AS PATH •• ROTATIONAl

• OPERATIONAl AND ACCESSORJES I COMPONENTS I MECHANiCAl MAINTENANCE , I EQUIPMENT MESSAGES DATA REDUCTION : DATA ANAlYSIS: • VAlIDATION • UMIT CHECKS • CORRECTION • DIAGNOSTIC DECISION • COMPRESSION • PROGNOSTICATION FIGURE 2 FIGURE 4 FIG U RE 6

AUTOMATIC DATA FRAMES

LIFT OFF CRUISE WEIGHT OFF WHEELS TAKEN NOT LESS THAN 15 MINUTES A fTER LIFTOFf AIRSPEED \ 100 KCAS NG CORR > 85 . 4% FO R BO TH ENGINES NG > 56% FOR ONE OR TWO ENGINES PLA ST ABL E .:. 1' /2 SEC ) 16 SEC T2C STABLE.:. 1" / 2 SEC > 16 SEC NO GUNFIRE PRECEEOING 16 SEC A IRSPEED 200 - 300 KCAS AL TlTUOE \ 10 . 000 FT AN GLE OF An ACK ( 15 OEG VERTICAL G 's 1. 5:" 1. 0Q ) 16 SEC F IGURE

DETECTED EVENT FRAMES

In OVERTEMP ENGINE STALL NG OVERSPEEO SLOW START NF OVERSPEED FUEL FILTER OIL PRESSURE OVER G VIBRATIONS MAXIMUM In SHIFT flUCTUATIONS NF VS In ERROR VG SCHEDULE NG SPEED ERROR FI GU RE 8

A-l0/TF34 TEMS

ANTICIPATED BENEFITS WHAT CAN ENGINE DIAGNOSTICS SYSTEMS DO FOR USAF?

• REDU C E UNWARRANTED MAINTENANCE R . • REDU C E PARTS AHD FUEL E FIXED M I SSION CONSUMPTION A SCENARIO D • INCREASE A I RCRAFT I AVAILABILITY N • PROVIDE AUTOMATED E I ENG I NE D A TA

S - --l----

S I • FEED BACK REAL I OP E RATIONAL DATA fOR I FU T URE DEVELOPMENTS No . OF AIRCRAfT REOUIRED FIGURE 9

- - - - - - - _. _. - - - - - __ ~ - - - - - - - - - _I

Page intentionally left blank

- - - - - - - - - - REVIEW OF AIDS DEVELOPMENT Henk C. Vermeulen KLM Royal Dutch Airlines Sven G. Danielsson SAS Scandinavian Airlines System SUMMARY Since the introduction of the wide-body aircraft KLM, SAS and Swissair have been able to collect a mass of experience, meager as will as excellent and in total profitable. All three are very determined to continue with AIDS on AJOO/310 Airbus (Swissair also on the DC-9-80).

The AJOO/310 AIDS as selected by KSS (KLM, SAS and Swissair) and Lufthansa has been developed into a very powerful Engine Monitoring System (EMS) and engineering tool capable to enhance aircraft regularity, reliability and economic operation.

INTRODUCTION KLM, SAS and Swissair started with AIDS at the introduction of the wide- body aircraft in 1970. The AIDS hardware specification for the Boeing 747 and McDonnel Douglas DC-IO aircraft was based on the experience obtained from: A digital recording experiment by KLM on a Douglas DC-8 in 1963 and 1964 (3) and DC-9 trials with prototype equipment in 1969.

The development of the ARINC 573 specification for a Flight Data Acquisition Unit (FDAU) to satisfy the new FAA requirements.

The initial objectives of KSSU with respect to the AIDS were primarily direc- ted to the monitoring of parameters related to: the safety of the flight the performance of the aircraft the performance of the flight guidance system the performance and condition of the engines.

The AIDS-EMS function was and still is considered supplemental to the existing monitoring tools.

For AIDS-EMS practically the same parameters were selected as already provided for display on the cockpit instrument panels with a few exceptions. Table I provides a list of EMS parameters monitored on KSSU aircraft. The total number of parameters monitored on KSSU 747 and DC - IO aircraft amounts to 380 and 280 respectively of which more than 50% are discretes (on/off signals).

In order to accomodate all these parameters and enable sampling at reasonable rates the system was configured around 3 Data Acquisition Units and a Data Management Unit (DMU) with limited data acquisition capability. Figure I de- picts the system block diagram.

AIDS OPERATION In KSS the AIDS is primarily applied as an engineering tool with a strong emphasis on analysis of recorded information. A printer was added to the air- borne system because it was recognized rather early that hard copies of excee- dance reporting could provide a very effective aid in trouble reporting. The application of on-board processing for limit exceedance monitoring and recor- ding control allowed to add a printer which could provide hard copy reports on request by the crew or automatically.

The tape-cassette is removed every landing made at the home-base and is subsequently transcribed to IBM compatible tape and processed. The routine programs applied comprise o.a.: flights logging, AIDS status reporting, EMS programs, autoland verification, etc.

For a good understanding of the function of AIDS in an EMS, it is essential to give some details on the functions of two specific software programs: the AIDS flight logs and the plot/list program. The first program provides a listing of all recorded flights per aircraft registration and the second pro- gram allows users to request a time history of a set of 8 analog plus 8 dis- crete type parameters either in table format or plotted. The user can call the AIDS flight log, select the airplane, the flightleg and the parameterset of interest and request a listing either for a specified flight mode or a GMT time span using the VDU terminals of KLM's data handling system. The very successful use of this program proved that AIDS is an invaluable engineering tool and fully met the set objectives.

CURRENT AIDS EMS APPLICATIONS ON BOARD For short-term trend analysis KLM relies on a trendchart that is up- dated by the flight engineer on every flight that lasts more than 4 flight hours. The flight engineer then selects a stabilized flight condition to re- quest and engine data print (table 2). The flight engineer uses this print to calculate the trend delta's with the aid of an engine performance calculator provided by the engine manufacturer and enters the delta values in his trend chart. Engine bleed and engine indicating problems, serious compressor/turbine problems and EGT-margin losses can be detected by the flight engineer using these short term trends. In addition these trends are checked by powerplant engineers on a regular basis and in case of crew complaints.

The engine data prints are also automatically presented during take-off and in case of limit exceedance e.g. the print of table 2 shows an automatical- ly reported impending hotstart on engine nr. 1.

The take-off prints are used to monitor the hot-day EGT margin, thrust settings etc. Crew complaints are supported by prints selected by the flight engineer and/or limit exceedance prints. In case of critical engine problems the flight engineer will contact the main-base via a single-sideband company channel for expert advise. With aid of the print he is able to provide exact information on the characteristics of the trouble or exceedance, the exact durations and the peak values of exceedances.

Strict adherence to the manufacturers engine operating limits could in- crease the number of engine removals because of the capability of EMS to very accurately report exceedances of operating limits that are based on experience, which include the human factor. It is obvious that these limits need to be adapted when an advanced EMS is used to prevent increased removals or inspec- tion rates. KLM was able to obtain the approval to extend the limit nn the allowable EGT exceedance time-limit on a particular engine when using the AIDS printer. This printer function has proven an invaluable tool for short-term engine monitoring, incident reporting and trouble-shooting to the extent that powerplant engineers consider this feature alone was worth the investment in AIDS.

GROUND-BASED LONG TERM ANALYSIS For long term trend analysis the AIDS provides weekly trend reports on KLM's JT9D and CF6 engines. These trend reports consist of 3 parts viz. an engine start trend (fig. 2a), an engine take-off trend (fig. 2b) and an engine cruise trend (fig. 2c). The trends shown apply to the Pratt&Whitney JT9D engine as installed in KSS 747 airplanes.

Engine start trend Of the engine start trend one important feature should be adressed. After careful analysis using the AIDS plot/list program KLM engineering decided to trend the initial fuel flow (IFF) at the moment of "fuel-on". This analysis showed that by monitoring this value it was possible to relate a positive or negative deviation from the required value to a fuel control adjustment. Al- though the fuel control adjustment screw was by design meant for shop use only, it is now used for on-wing adjustment. This meant that by monitoring the IFF trend, hot and hung starts and consequent fuel control removals are reduced considerably and related unnecessary engine test runs avoided. Since its in- troduction more airlines became interested and have requested the engine manu- facturer to provide proper means for on-wing adjustment.

Take-off and crU1se trends The take-off and cruise trends are both used for monitoring of engine deterioriation, primarily by checking EGT rise and EGT margin, blade failures, compressor and turbine problems and indicating system errors. The take-off trends is also used to monitor powerlever-alignment which avoids valid crew complaints and allows to neglect invalid crew complaints and save on otherwise

Page intentionally left blank

necessary follow-on actions.

Gas path analysis SAS first started using GPA based on testcell data to analyze modular deficiencies on the JT9D engine. After showing positive results KLM installed the same program but to analyze the CF6 engine. After an extensive evaluation and calibration program to determine sensitivity, performance levels and abili- ty to find degraded modules, the program also at KLM is deemed useable to iden- tify problems on preshop tested engines.

The function of the test-cell however has always been primarily to verify that an overhauled or repaired engine meets performance requirements. The iden- tification of modules degraded below limits should therefore preferably occur on-wing such that an engine's work scope can be predicted prior to shop entry.

SAS and KLM has therefore decided to develop the necessary procedures and know how with GPA based on AIDS recorded data.

SAS has started with JT9D-7 engines installed in Boeing 747 while KLM will analyze CF6 engines installed in Dc-lO. The JT9D program will use only partly instrumented engines while the CF6-DC10 will be fully instrumented.

Future aircrafts within KL~I, SAS and SWR will always be fully instrumented and in particular the A310 with also a PMUX, presently the only one specified with PMUX, will already from the beginning be monitored by programs capable of modu- lar performance analysis.

Table 5 shows the parameters used in the presently ongoing GPA programs.

As can be seen the SAS JT9D-7 GPA uses the least number of parameters and there fore also has the least capability. This program however has advanced the most and a discussion showing some results follows. Because of the small number of parameters that are available to describe the engine operating characteristics, some assumptions have been necessary to do on the modular deterioration that is analyzed. The assumptions are a fixed ratio between change in efficiency and air pumping capacity on the compressor modules and also both the FAN and the LPC are treated as one module. The main disadvantage with the hard coupling between efficiency and air pumping capacity is expected to be seen in the HPC where the front stator vane stages are variable and changes in pumping capacity might be induced this way.

Figure 12 shows the variation in ambient conditions under which data is collected for GPA. Each datapoint being used is a stable frame that has been recorded by the airborne system. The measurements are reduced to sea level static and corrected for effects of Reynolds number, engine service bleed and the offset initially found in the actual installation position. This the cor- rected value is compared to a fleet average baseline and the percentage diffe- rence is calculated. This difference known as "gross delta" is first used to look for apparent sensor errors. If the datapoint is deemed erroneous an appro- priate message will be issued and no further analysis is done. The accepted datapoints are used for further analysis.

Figure 13 shows one month worth of data in terms of gross deltas. The result when using each individual gross delta point for analysis is shown in figure 14. As can be seen the scatter is significant and therefore with this combination of sensors and the this way obtained accuracy of gross deltas the result is not accurate enough to correctly analyze module performance based on the data point. The way to get around the problem that is presently used, is that gross-deltas from individual flights are calculated, checked for apparent sensor errors and if found within sensor check limits the gross deltas are passed on to a ten flight average calculation. The gross deltas again in the average calculation will be checked for outliers in a simple correlation ana- lysis. The result from using averaged gross deltas as input to GPA is seen on figure 15.

This particular engine had been installed for several months already at the beginning of this trending, why very limited deterioration is to be ex- pected over the trend period. The 13 trend points now corresponds to 130 flights with stable data, or on this particular route-net approx. 200 cycles.

GPA based on the very limited parameter set used in this trial can not replace pre-shop tests of engines. It gives however additional valuable infor- mation on top of the normal trending and will be further studied for use in our preventive removal concept.

LAP, Life Accounting Program This program is used to bookkeep the amount of damage on critical parts based on actual engine performance and routes flown.

The High Pressure Turbine airfoils on the JT9D-7 are amongst the most critical parts in that engine and are deemed possible to be modelled accurate enough for an analysis. The basic program, written bij PWS, uses precalculated severity factors for each mission and actual AIDS data from each flight to de- fine engine performance levels and routes flown.

The actual life consumption depends upon the mission and the performance status of the engine. The mission is described by a sverity factor that has been precalculated by a Mission Analysis Program using statistical data. Each citypair has several severity factors that varies with respect to season of the year and actual failure mode accounted for, see table 3.

The reason for useage of precalculated statistical severity factors is, that the KSSU airborne 747 AIDS program at the time of specification was not defined to collect mission analysis/life accounting data.

The following failure modes are referred to in the program.

- NGV Nozzle Guide Vane crack 1st blade creep-fatigue - 1 BCF - 1 BOC 1st blade oxidation corrosion 2nd vane deflection - 2V - 2B 2nd b lade creep The IBOC is the ·only life useage that can be reset by repair e g recoating. All other failure modes refer to life useage that is incremented at each flight throughout the service life of the part.

A typical output generated on a monthly basis is shown on table 4 where the percentage life used is shown for each failure mode. 100 means that ex- pected service life is completely used up.

Experience The most critical part ~s 1st blade and therefore, the experience on LAP-

J

life compared to actual failures non-failures is shown in figure II. A curve ending with F means failure and FO means failure due to overtemperature or where also overtemperature has been confirmed. As shown by the graph there is a good correlation between actual failures and 100% life used for lBCF. The weak point with this program is that statistical missions are used instead of actual . Missions significantly deviating from nominal as well as engine excee- dances for example, higher temperatures than normal during take-off, startup or reverse if encountered is not accounted for.

Autoland verification program This program produces per aircraft a two-monthly review of Autoland performance. Per line the Autopilot disconnect heights, ILS tracking quality, wind at 100 ft, touch-down dispersion and touch-down maximum g-loads are pre - sented. For the total fleet a two-monthly statistical performance review is presented. The program is used to demonstrate an acceptable Autoland success rate to the authorities and keep control over the maintenance of the Autoland system.

747 APU monitoring The AIDS system allows to apply a very effective means of health monito- ring to the Auxiliary Power Unit. The groundbased computer monitors accelera- tion time, EGT peak and rotor speed at peak EGT, airduct pressure during air- conditioning system operation with 3 packs and EGT at no-load condition. This program provides indication of and/or clues to mechanical problems, starting problems, airleaks, compressor and turbine inefficiencies.

747/DC-IO aircraft structure lifecycle programs Since ma ny years KLM collects AIDS recorded data for assessment of ser- vice load experience. Results allow comparison to Boeing's fatigue integrity program with the objective to compare the severity and schedule structural in- spections on this basis.

Studies on the recorded data also revealed that changes in pressurisation procedures would extend the life of the pressurised structure.

TROUBLE AND INCIDENT ANALYSIS In reference (2) the analysis and monitoring of JT9D starting problems an~ the analysis of JT9D auto-accelerations was presented with the cures.

Since that presentation the CF6 compressor stall problem has been solved.

The total story is as follows: CF6 compressor stalls Although the KSSU CF6 compressor stall-rate reached a low level of 0.06 per 1000 engine hours, the nature of this problem urged to aim for elimination.

The major reason is that stalls can occur in critical phases of the flight and thus might endanger the safety of the flight or comfort of the passengers.

Several of the experienced CF6 compressor stalls in the winter period of 1979/ 1980 occurred at the moment the airplane entered a rainshower (fig. 3). Appar- ently the compressor inlet temperature (cit), measured at the high pressure compressor inlet, drops at constant N2 rotor speed and constant total air tem- perature, causing the variable stator vane (vsv) to move with the ultimate re- sult that the Nt rotor speed increases at constant N2, causing some individual engines to stall.

First action was to monitor Nl vs N2 to avoid an N1/N2 matching critical to stall. At the same time the study and analysis of all AIDS recorded compres- sor stalls by a group of specialists was performed with the objective to ulti- mately eliminate in-flight stalls.

The studies resulted in two actions: First Protective rainshields around the CIT sensors were installed figure 4.

Second In the testcell the Variable Stator Vane (VSV) schedule was adjusted to the more closed position figure 5.

By these actions the engine stall-rate started to drop significantly as can be seen in figure 6.

Today CF6-50 aero stalls are practically eliminated. The cost of aero stalls to the airline were negligable for CF6's with steel compressor casing but amounted to over $ 200 000 each for CF6's with titanium casing. The savings resulting from elimination of aero stalls alone paid for a very substantial part of the AIDS investments if not complete. .

It should be noted that all CF6 operators benefit through the AIDS EMS system as used by KLM and the engineering efforts that KLM and other AIDS users put into the task of solving problems like the CF6 compressor stalls.

Fligh t Technical Special Flight Technical analysis programs provide means to analyze crew complaints more thoroughly. Programs developed for this purpose are: - an ILS beam quality check program - a runway surface analysis program and - a windshear analysis program Fuel Consumption Management In order to ensure accurate flightplan fuel determination KSS uses a computerprogram to monitor the consumption levels per airplane type and per individual airplane. Accounti~g for consumption levels of individual airplanes allow tighter fuel reserves. In order to more effectively analyze individual high consumers KLM developed a program based on AIDS data that enables engineering to verify effects of maintenance actions on consumption on a short notice.

Figure 8 shows CF6 gas generator curves developed from AIDS recorded data on one flight u sing a least squares approximation.

It demonstrates the high degree of repeatability achieveable with AIDS recorded data. The maximum deviations of individual data points from the cur- ves are 0,4% corrected fuel flow, 0 . 13% corrected N2 RPM and 0.05 corrected EPR.

GENERAL ASPECTS From the examples of the previous paragraphs AIDS appears as a reliable and useful source. Because of its accurate observations, lessons can be learn- ed fast, proper actions can be subsequently applied and cockpit procedures op- timized. AIDS also demonstrated on various occasions the ability to observe and report problems outside the observation capability of the crew, it does not conceal human imperfections nor human excellence in performance. Careful treatment of problems where the human factor is involved is a must when dis- turbance of human relations is to be avoided.

A300/310 AIDS Development Ten years of experience with expanded AIDS systems have demonstrated to KLM, SAS and Swissair that the AIDS has matured into an effective engineering tool as predicted. With increasing positive experience it became evident that more effective airborne software was desirable but prohibited by the capacity of the system and the extreme costs of fleet modifications. With the advance of digitalisation of aircraft systems ARINC started to develop new characteris- tics for these systems known as their 700 series characteristics. In the Boeing 757 and 767 and the Airbus 310 these 700 series digital systems are ex- tensively used . The impact of this development can best be illustrated by com- paring the types of KSS AIDS inputs for the 747 and A310: 747 210 discretes, 170 analog and 4 digital data busses A310 39 discretes, 49 analog and 38 digital data busses.

Taking advantage of this progress in the application of digital technolo- gy, KSSU and the ATLAS European group of airlines, with full cooperation of Air- bus Industrie,started to develop an Expanded AIDS for the A310 (ref. 8) based on ARINC 717 (ref. 9). Early 1980 this effort was successfully completed. A block diagram of this system is shown in fig. Subsequently KSSU developed a specification which in more detail ' specified the desired software functions and specific features derived from KSSU AIDS experience over the years and contracted the system to a major U. S. supplier. In general the increased ca- pacity of the A310, both in terms of parameter inputs and installed software, is used to enhance the AIDS a s , an engineering tool and expand its trouble- shooting capabilities. The relative expansion on parameter inputs partially comes by itself on a digital airplane where most parameters can be sampled fro m ARINC 429 data busses.

Justification As previously stated AIDS provecl to be an effective engineering tool and provided to KSS an ample return on investments.

Of course the 747 AIDS was not an optimum system compared to the possi- bilities and the available experience of today. Therefore applying the lessons learned to the A310 AIDS application will result in a still more cost effective system. The conditions have also changed, c.q. the amount of sensor wiring in the airplane has been decreased compared to the 747 and DC-IO AIDS and the ca- pacity of the electronics increased drastically with the result that the costs of a complete installed system is less than of its predecessors .

Translating the investments to costs per flighthour and assuming a reasonable aircraft utilisation these costs will rougly amount to $ 10.- per flighthour. The total costs will double when the AIDS operating costs are added.

A KSS return on investment study for the A310 AIDS did not produce a homogeneous result between KSS partners because there were several differences in estimated savings per individual program or different emphasis on values of benefits.

The ultimate conclusion as derived from past experience can therefore best be presented schematically. Figure 8 shows that the level of quantifyable savings will more than balance the AIDS operating costs.

Characteristics of the KSS A310 AIDS The control of the recording, printing and display functions is per- formed by the Data Management Unit (fig. 9 ) . In this unit an Intel 8086 16 bit microprocessor is installed to perform the required functions. The memory comprises 58 K bytes PROM, 18 K bytes RAM, 29 K bytes protected RAM and 39 K bytes EAROM.

Depending on the flight mode the DMU commands continuous recording or selective recording.

All data that is to be recorded passes a 20 second delay buffer suc h that at detection of specific events always 20 seconds pre-event data is av ' ailable on the tape-cassette. 75 % of the cassette-tape capacity is used for routine recording and the remaining 25 % is programmable via the Control Display Unit (CDU). This feature provides to engineering a tool to analyse and solve persistent problems.

The on-board printer can be used by the crew for hard copy engine data but i ts primary purpose is to provide maintenance with all necessary informa- tion. For this reason all maintenance prints are stored during flight and a light on the CDU will inform the maintenance crew that exceedance reports are stored in the DMU memory. In case of specific events as defined in the pro- gram, 20 seconds of data prior and 20 seconds of data after the event will be stored and can be recalled by the maintenance engineer using the printer. The memory section used for this feature is called the replay buffer.

From the replay but fer the maintenance engineer can select "canned" sets of parameters prividing a second by second listing of the event occurrance.

For the objectives set for the EMS part of the program it became necessary to provide additional sensors on the engines. KSS and ATLAS in close coopera - tion with Airbus trndustrie and the engine manufacturers succeeded to specify a Powerplant Multiplexer (PMUX) as standard part of an A310 Expanded AIDS.

This PMUX multiplexes temperature and pressure signals, combines these with the output of the Electronic Engine Control (EEC) or Power Management Control (PMC) and sends all this information with the engine serial number via an ARINC 429 dataline to the DMU. This improvement ensures a tighter mainte- nance control on the quality level of those inputs not monitored by the flight- crew .

A310 ENGINE MONITORING KSSU formed a team of engine monitoring experts to part~c~pate in the specification work on the A310 AIDS. This group being able to take advantage of experience with already existing programs has defined a system incorpora - ting several new features that have never before been available.

Program functions New functions Two of the new program functions namely the "history buffer" and the "replay memory" are briefly described already above. One other is known as "stored prints" and works such that the printer does not print in real time, with some very few exceptions, but data that shall be printed is stored in a "print buffer" until a special request is made. There is capacity for storeage of 11 prints with the distribution as defined in table ~ Exceedance Control The software of course ~s capable of defining flight modes and to com- pare selected engine parameters against flight mode related limits throughout the m~ss~on.

In case of an exceedance, different actions will be taken. An exceedance can thus cause: I. Update of exceedance print buffer 2. Recording wi thou t 20 sec. pre-event data 3 . Recording with 20 sec pre-event data 4~ Update of replay memory.

Stable Condition search The search for stable conditions is, done with a new logic. The method can be described as a window sliding in time over actual and previous data, that are remembered in the computer, looking for stability. Old data is com- pressed in such a way that it is represented by its average over a 16 sec.

period. A maximum of 8 periods will be kept in memory this way. Previously, normally was used a method that stored a reference sample to which actual data was compared over a predefined stable time period or until out of stability occurred leading to a new sample being set.

If a stable period occurs that is not considerably longer, than the pre- defined stable time period, it is very likely so, that the old logic should not be able to find it. The A300/310 is a short haul aircraft and the time during cruise is so short that an improved search method for stable data was needed.

Divergence monitoring Another new feature will be the "EGT divergence monitoring" (EDM). The purpose of EDM is to have a method that immediately can recognize a sudden gas path damage through its effect on monitored engine parameters. The influence coefficients for a typical twinspool engine show that regardles of performance deterioration, except for fan flow capacity any gas path damage will affect EGT in an increased direction. Therefore during specified conditions EGT of the two . engines are compared.

Both engines are operating in the same ambient conditions so that no cor- rection for that has to be done. One engine is corrected for its offset in thrust setting to the reference engine and the resulting difference in EGT be- : tween the two engines is compared to a reference difference established first flight every day. If the difference between actual EGT delta and reference EGT delta exceeds a certain value, the one engine with the increased EGT is automa- tically pointed out as the unhealthy one by the AIDS system.

Recording Control General Continuous recording takes place in the following flight modes, engine start, take-off and approach landing. Selective recording is performed in the other flight modes, which means one frame every 100 sec. will be recorded.

Exceedance recording 20 sec. of pre exceedance data and additional data m~n 20 sec, max 120 sec. or until the exceedance is passed is recorded.

Corner points Corner points are recorded to better define the mission. The corner points are used in groundbased programs for mission analysis with respect to life accounting on critical parts and refined cycle counting on life limited parts. These points are defined as end of climb or start of descent.

Recording of selected frames for oil consumption monitoring Either during taxi before and after the flight or prior to engine start and after shut downJdata can be recorded for the purpose of oil consumption monitoring.

Recording of stable data Above a certain altitude, mach number and in cruise, stable conditions are continuously searched for. When stable conditions are found data is auto- matically recorded.

Recording of APU Every time the APU is started on external or engine power, 16 frames worth of data will be recorded.

Other planned A310 AIDS applications As for the current KSSU wide -body aircraft the A310 AIDS will also be used for: verification of satisfactory autoland system operation trouble and incident analysis assessment of service load experience and the monitoring of: the condition of the auxiliary power unit the braking and anti-skid system aircraft performance deterioration safety limit exceedances A310 AIDS ground system The system proposed for KLM is depicted in figure 10. It shows again the emphasis KSS lays on the function of AIDS as an engineering tool. The AIDS data as recorded on a cassette continuously will update a data-base in the main EDP center where in paralell also data is stored providing information on crew complaints, maintenance actions, etc. This EMS data base provides perio- dic trend reports, status reports and automatic exceedance reports and is via terminals accessible by line maintenance and engineering REFERENCES I. SAE ARP-1587 "Aircraft Gas Turbine Engine Monitoring System Guide".

2. Vermeulen Henk C. KLM Royal Dutch Airlines "Current and Future Use of an AIDS integrated EMS" SAE ~aper 801219 3. Driessen Ed. A., Vermeulen H.C. and Ledeboer K.H.

KLM Royal Dutch Airlines "Use of recorders in future aircraft operations" AIAA paper no. 64-352 June 1964 and Journal of Aircraft Vol II no. 3 1965 pp.

176-184.

4. Urban L. A. Hamil ton Standard "Parameter Selection for MUltiple Fault Diagnostics of Gas Turbine Engines".

ASME paper 74-GT-62 March 1974.

5. Danielsson Sven G., "Gas Path Analysis applied to pre and post-overhaul testing of JT9D turbofan engine" SAE paper 77-0093 6. Danielsson S.G. and Dienger Dr. G., A. European v~ew on Gas Turbine Engine Monitoring of Present and Future Civil Aircraft".

AIAA/SAE/ASME paper 79-1200 June 1979.

7. De Hoff R.L., Baker L.E. and Hall Jr. W . E ., Systems Control Inc. (Vt) Palo Alto, Ca "Impact of Automated Monitoring on Engine Operations and Sup- port.

AIAA/SEA/ASME paper 79-1276 June 1979.

8. Kalbe H. Messerschmi tt-Bolkow-Blohm GMBH "New Aircraft Integrated Data Systems for Airbus A310" Paper presented at the 10th AIDS symposium of the "Deutsche Studiengruppe Fur Flugdatensysteme (DSF) March 1980. Aachen Germany.

9. ARINC characteristic 717 "Flight Data Acquisition and Recording System" March I, 1979

j

EMS PAR A M ETERS MONITORED ON KSS AIR C RAFT AIRPLANE DC to : AIRBUS A3tO: :PIU JT9~ 'GE CF6-50E :GE CF6-50C :GE CF6-80At : ENGINE --------- -- ----------- -- -_. _----- , , FAN ROTOR SPEED X X X x X X X CORE ROTOR SPEED EXHAUST GAS TE"PERATURE (EGT ) X X X X X X X FUEL FlOU X lPC DISCHARGE PRESSURE X lPC DISCHARGE TE"PERATURE X X X X HPC DISCHARGE PRESSURE X X HPC DISCHARGE TE"PERATURE X ENGINE PRESSURE RATIO X VARIABLE STATOR VANE POS. (VSV) : X x LPT INLET PRESSURE x X x x VIBRATION FAN ROTOR X X x X x VIBRATION CORE ROTOR X x X ENGINE OIL QUANTITY x x ENGINE OIL PRESSURE x ENGINE OIL TE"PERATURE x rOUER LEVER ANGLE x x X IGNITION X X X FUEL SHUT-OFF VALVE X X X x BLEED RELATED PARA"ETERS: START VALVE POSITION X X X x PNEU"A T IC BLEED VALVE POSITION X X x ISOLATION VALVES X x PACK "ODE SELECTOR X X X x ENGINE INLET ANTI-ICE X X X x UING ANTI-ICE X X X X VARIABLE BLEED VALVES x APU SHUT-OFF VALVE X x NACELLE TE"PERATURE x X x BL EED PRESSURE X X X x ALEED T E"PERATURE X X X x FLI GHT I DENT X X x x DATE X X X X X X X X G"T X X FLI GHT "ODE X X PRESSURE ALTITUDE X X X X X TOTAL AIR TE"PERATURE X X X MACH HU"BER X X X STEADY STATE IDENT X X X AUTO THROTTLE ENGAGED X X X TABLE 1 SAMPLE ENGINE DATA PRINT FROM AIDS ON-BOARD PRINTER AlC 747. 1 MODE 2 FlT 688 AlT 6"0 UTE 11.05 CAS 46 1734 "ACH 6"T 'a'" SCAN 1781 TAT HCT SCU 1781 EPR 1.007 1.012 1.017 1.016 Nl 13.1 23.9 28 . 0 28.3 E6T 629.5 451.6 408.8 399.1 63. 7 N2 39.9 60.3 65 . 0 H 153 572 647 635 VII .0 .0 .1 .0 VItT .2.1.3.3 JR.P .2 .2 .3 .9 ~AT£R 0 0 0 P~RL 0 - 2 PS4 12.4 28.3 37.4 38 . 1 APU EGT 460 APU RPft 100.2 DUCT PR L 30 DUCT PR R 29 16N 1 0 0 0 0 16H 2 0 0 0 0 PN.VALVE 0 0 0 0 OIL T 45 57 85 17 "ACT A 1.6 1.7 1.9 2.0 UCT. 2.0 1.7 1.7 1.7 EAI 0 0 0 0

~AI ° EPR LIM 1. 445

EPR "ODE 2 APU EGTL 900 E6T L. 649.9 649.9 649.9 649 . 9 Nl LIM 104 . 2 104.2 104.2 104.2 IR . T 43.6 61.0 88 . 7 85 . 1 NOTEI IMPENDING HOT START NUM'ER 1 ENGINE (AT MEXICO CITT AIRPORT ).

TABLE 2 TABLE 3 SEVERITY FACTORS PER CITY PAIR ~ FAILURE HODE 5o!'&- fIt. 5~&- 5~&- 5o!'&- t&ilu"~ city- p~i ,. tim. son1 son2 son3 son4 mod.

I=lMS FRA 7.9 2169 2371 2898 2€-16 NGV RMS rRR 7.9 1248 1656 34113 2889 Iller RMS 1418

rRR 7.9 '3936 3836 2U37 noc

FlMS FRFI 7.9 0190 0360 1449 9671 2\.'

RMS FRR 7465 7.9 8643 118130 8673 2B TABLE 4 OU iPU T FRO~ ~AP- ~A P ~ROC~ ' 3S I ~jG DATE 27MI18t REFERENCE STOHK -R F~ANI:N SORTED BY A /C REC ~O THE FOLLOWr~ IS A SUMMAR Y OF THE DAMAGE ACCUMULATED ON EACH PAk T

1------------- % LIFE USED --------- ----1

lS0C 2P IHST ENGINE NGV 1BCF :?V PO~ HO DDU bb3074 6 1 "! 3 76 5 3 2 5 2 19 1 16 2 5 .9 1>62~7 18 0 16 0 DOl" ., 1 50 9 DDL3 b6~41 93 .3 17 9 76 6 DOl 4 1>63167 95 .7 37 8 23 .9 31 8 23 2 KHA1 bb2814 9S .9 54 1 8 1 83 9 26 5 (HA2 9 4 49 .7 3 1 6b2999 7.3 7 4 KHA3 t.b275 0 6 .8 63 .7 8 . ~ 47 .9 39 7 IHA4 3 . 9 !:Jb2909 8. 5 9 . 1 11 2 76 . 5

n 4 49 4

lCAl 6b27b2 92 3 b3 b 2 2 b2 4 .5 4b . 4 o 0 lCA2 ~b30"3 50 0 ICA3 6b307~ 110 9 23 9 21 S 9S9 67 1 lCA4 t.6280 3 96 5 S5 S 3t 9 52 .5 66 S 7 9 4b 9 bS 2 lC&1 ~b2S27 47 1 14 0 24 9 27 .2 11 2 Iets2 bSS637 35 2 2S 6 4 1 ICB3 bb2982 9S 3 70 7 7 3 40 .3 9.1 48 2 57 9 IGB4 t.85b36 68 2 44 .S 43 b 36"!

18 J 10 2 BUAl b85616 36 2 4.7 49 .7 1 9 90 2 4.3 BU A 2 6bJetl 90 1 . 1 9S 3 BUA3 !>b2818 90 2 SS2 95 .2 19 .0 16 1 21. 7 BUM 685635 1 9 80 4 71 0 94 1 87 3 BUBt 6627S4 24 3 32 1 .8 b62979 79 1 41 9 BUB:?

54 b 44 5 b62930 91 9 31 5 18 0 BU&3 84 1 92 0 62 1 8 .1 BUll 4 ~2846 97 1 TABLE S GPA PROGRAH PARAHETERS USED IN KLH.SAS SAS JT90-B747 KLH CF'"b - DCi.O Ni.

Ni.

N: _ ~ N?

WF WF CGT L:GT PS4 PS2C TT2C EPR PS3 TT3 F'TS . 4 VSV ~ VSV ON SAHPLE ENGINE ONLY

*

TABLE b P RINT FORHAT~ / NAHES COHPARISON REPORr PRINT/EVENr STARLE 2 EXCCEDANCE 2 TAKE OFF 1 CDU REQUESTL:D i.

TAKE OFF TREND REPORT i.

CRUISE TREND REPORT FAILED ENGINE START REPORT HAINTENANCE REPORf i.

LOAD/FLIGHT CONTROL REPORT i.

~

AIRBORNE COMPUTER

AIRCRAFT

PARAMETER

AIDS-

RECORDER

w o N

•

CASSETTE

AIRBORNE PRINTER

COCKPIT PANEL

FLIGHT DATA

l:Y

ACQUISITION

UNITS

l

,...--

a· ~ , SERIAL TAPE

GROUND CONVERTER

OUTPUTS

....

BASED

a.....--~

~

COMPUTER

(]+---.

FIG. 1 AIDS SYSTEM AS INSTALLED IN KSS 747 AND DC -10 AIRCRAFT

------ -- --- - - ------

INSTALLED 10JAh. 81 VfRSION 30 PRINT DATE 7 APR 8 I PAGE ENGI NE 00,879 7.7B A/C PH-8UB POS, ENGlhE ANALYSIS OUTPUT TYPE: OAllY FLIGHT EXCEEDAN(ES NO EXCEEDANCES START TREND BLV APUN EGTP NIP h2P IFF SFF DEP/ARR T20 TlO T50 BlP BlT O/T PALT TAT D,lY GMT FlT 4e .2 2400 705 0101 91 80 .8 10101 533 14.0 845 eMB/SIN 17 5 10 29.3 155 302 723 74. 0100 90 101.0 458 14.0 .~.o 376 o 10 35.8 100 35 Z3 302 13,9 8.5 SlN/Ht:.8 19 487 13.9 717 5 15 33.7 143 78 483 19 101.0 302 2215 845 ME8/SYO 15 15;g 726 0100 35.7 144 5 0 222 29 100.8 .7. H:~

e. 6 SYO/MEB .. o 15 '''.2

• 02 '55 100:(1:0 0(1)(1 0100 520 21 100.8 14.2 47.2

840 / ...... o o 35.8 151 91 '10

40J 707 101. , o5~ 07.7 415 825 18 I, 30.2 141 1, -333 9 1".5 202 1859 845 SPL/OXB 10 -13 17 10 1. I .60 1 J.9 .. ~ .0 3.9 737 OXS/ o~:o 10 33.0 I.e 85 302 ,17 e45 l ' 14.0 4e.z 2,,00 705 , 723 ,9.3 155 91 80 .8 101.1 533 845 (MB/SIN 17 5 10 .. t.6 0100 35 90 23 101.0 1'.0 376 7.4 845 SIN/MEB 19 o 10 35.8 100 '58 302 1329 .83 19 101.0 487 13.9 47.4 345 717 84 5 MEB/SYO 15 5 15 33.7 143 78 - 302 2215 ,9 100.8 47. 1 ... 2 47.2 15 55 726 SYO/ME~ 14 o 15 35.7 144 50 222 455 8.0 -. 02 47.2 t)(I~.Q 0100 (:I:~.Q 91 520 ,I 100.8 47. 10\.2- e40 / o o 35.e 151 " ......

'03 707 397 0101 10 -HO 13 99.6 409 ,2.1 5f.0 762 SPL/JfX 19 o .... 0 35.7 142 ' JO. 12.1 0.1 -01 12 100·1 435 13.0 4C.0 338 735 0., Jfx/SPl 10 7 15 30.3 143 58 ,307 30'

., 13 99.0 033 13.9 OC .9 .01 768

SPL/ORO 10 5 15 37.5 1. 0 12 10 611 50' ; ~~i 755 0101

30.0 13. 01 o 99.6 435 47.5 329 a

2,41 012 ORD/SPL 15 5 15 1"."

4,. 793 0101 -350 12 99.0 07.1 385 077 SPL/YMX 15 o 10 3'1.7 13. 30 .04 122 7 1'.'

0 1 10 7, 99.6 003 13.5 "0.4 331 689 10 5 15 35.2 13e 835 16 404 2J30 078 VYZ/YMX 13 99.0 469 22. I 5 8 .0 397 702

19 o ...... 35.7 10 -Z90

1,01 0.1 SPl/JfX I', 30' 735 0101 I, 100.1 435 13.0 4t.O 338 10 7 15 30 .3 143 58 -01 ,307 042 JF'VSPL 30' 1 A KE -OF I' TRENO -6 ••• -~ ••• ·2 ••• EGTM • • 2 ••• • ~ -Z ••• -I ••• RPM ••• 1 •••• 2 2 ••• BRP ••• 6 •••• 8 rb~ ... 2 •••• ~ XIOC -5 ••• P~H ••• 5 •••• 10 .... 5 " -6 ••• - •••• -2 ••• tPR ••• 2 PSI Q3~~R·O.5 ••• 1 •• 01.5 ( .

o i e o INITlAL .1 E Z 1 C • 8 o 10 · fLT .1 1 E o C • B o E o .2 I AVERAGE .1 T 0 , C • .5 D 2. I .1 1 E • 0 2 I C • B .1 T E D o 2 1 C • .8 o • 1 1 E DAIL Y fLTS 403 846 02 C • : B 202 845 .1 .2 C • 8 E o I 302 &45 .1 D 2 I C • e. o 302 &45 .1 T E C _ .B o 2. 1 302 8~5 .1 T E o 2 I c. B H 302 8.5 .1 T E o C • B H o 2 I 402 8_6 .1 1 E .03 8.0 o 2 1 C • 5 .0

304 m :1 T E o

o o • 2 1 C • o 042 .1 1 E 30' • B o o o I C • 50_ OIl .I T Z • e

b

o

o .2 I ( . .B o

612 .1 1 E 50' o o 2 I C .B 0: 404 677 01 T o o 40_ o 2 1 C • B.

677 .1 T E o 2 I C • .B H '0_ 018 .1 T E o o 2 I C • .8 o 67& .1 T E o 50' o a

o • 2 I C • :0

30. ~~1 : 1 T E • B OA 2 I C o 642 .1 1 E o 30' • B BEl NORMAL NUMBER Of A/C PACKS IS TAKE-OfF PERfORMANCE ACP I,. A • G 8PR 0/1 I-VI5-T THF PAL T TAl CAS TTO TTA A-NT - B 123 V I I N DAY GMT SCAN fLT oEP/ARR EPR EPRl EGT f/F HI N2 PS4 7 604 90.4 9 •• 1 291.7 91 0.6 2.3 •• 9 09 ·56 II 125 0.0 0 1.0 1.8 III o 0 B 2321 1060 64, JFK/SPl 1. '" 1.42 818 84 0.0 2.1 4.8 7 2 -275 8 129 0.0 0 1.7 1.9 111 lOB 122J 1897 611 SPL/ORO 1.427 1.44 841 81J5 91.0 93.8 298.S 87.7 g2.7 88 0.5 1.7 •• 7 03 491 o 127 0.0 0 1.6 1.9 III o 0 B 2251 3599 012 ORo/SPl 1.378 1.39 781 70J2 216.4 8 2 0.5 203 6. -335 7 121 0.0 0 1.71.9 III lOB 12.3 1717 SPL/YMX 1.378 1.38 80 2 7377 88.0 92 •• 2eO.9 4.7 93 o.s 2.2 4.0 22 13~ 0.0 0 1.0 1.8 III o 0 B 2 039 2242 ~H YHX/YYZ 1.363 1.39 8 2J 7101 89.0 95.2 270.5 71 52' yyl/YMX 206.7 92 0.0 2.9 4.5 06 eso 15 124 0.0 OI.01.8JII o 0 B 2347 1512 67 8 1.36. 1.38 802 0917 88.6 9 •• 1 11 125 0.0 0 1.6 1.8 III 008 1'6 057 018 YMX/SPl 1.371 1.38 813 710. 89.3 9 •• 5 272.9 92 0 •• 2.7 •• 5 07 .611 30. XXX/ NO TO-DATA 5.0 e_ 0.6 2 . ~ 68 ·28 I 10 134 0.0 0 1.6 1.9 III o 0 5 1257 940 041 SPL / JfK 1.410 1.41 818 7710 90.S 9'.1 293.8 4.9 91 0.0 2.3 69 ·S6 II 125 0.0 0 1.6 I.e III o 0 B 2321 1060 p.2 JfK/SPl 1.411 1.42818 76M 90 •• 9' .1 291.7 CRUISt - TREND -~ ••• -2 ••• EGT ••• 2 ••••••••• 6 -Z ••• -l ••• N/l ••• l •• •• Z -6 ••• -_ ••• -2 ••• CF'A ••• 2 Y18 ••• J •••• 2 X10C -2 ••• -J ••• F/F ••• 1 •••• Z ,,-2 ••• -J ••• N/2 • • • 1 •••• 2 -J ••• -2 ••• -1 • •• 8TH ••• 1 IPS XIOOKG/Hil " XIOC E • F • • z B INITIAL .1 10-FL 1 • I E. F 1 2 C B T AVERAGE • I E. f 1 2 C B T E f ol 2 C B • I T E. f • 1 2 C B • I T E. f 1 2 C B .1 T DAILY fLTS 403 a40 202 8.5 B 302 845 645 8 8 o J02 845 302 8.5 402 8.6 .03 840 30.

o 30. E. F 2 C B 041 .1 T 30' 611 • I E. I' .1 2 BC o 50' 50' • O. ~H :1 B o c: 400 677 B • 04 678 504 018 o F C 8 ~~Z : I 2 30' A 304 042 AIR-8REATHf:R-TEMP EBI CRUISE P ERfORMANCE NORMAL NUMBER OF A/C PACKS IS 2 AcP HARGIM ALG DAY GMT SCAN fLT DEP/ARR EPR EPRL EGT ff HI H2 PS. O/T I-VIB-T THP PAlT TA T CAS H.eH A-N1-B 123 ABT EPR TAT 0/1 I~N NO STABiliZED FRAME NO BTM OUA 642 Jf"/SPL 008 27 13 86 87 121 93 309 -19 300 .822 I.~ 1.8 101 94 011 OIl SPL/ORD 1.240 1 •• 3 0.5 1.0 JO _I 148 1.361 2 1538 1600 NO STABILiZED fR,lME 012 ORO/SPL NO BTM DATA 1314 3000 017 SPL/YMX 1.309 1.'3 092 2931 87 88 127 93 o. J 1.1 310 -19 304 .t19 1.7 1.8 III ABN l~l 1.3~O 6 9" GIG 077 YMX/YYZ NO STABILizED FRAME 153 1.303 15 9.

NO 8 TM 0 ... T A o7B YYl/YMX NO STABILI ZED FRAME HO STABILIZED FRAME NO eTH DATA 678 YMX/SPL NO STABILIZED fRAME NO BTM DATA 301t xxx/ l!)~ 1.339 ,_ 310 -15311 .t30 1.0 1.6 101 041 SPL/ JfK 133. J200 711 307' 89 89 131 93 9" OlO 04 2 Jfl(/SPL NO STABILIZEO fRAME NO 8H4 DATA FIG. 2 AIDS LONG TERM TREND JT9D ENGINE ___ NR 2 ENGINE _ NR 3 ENGINE 1~ r-----~--~----~----~----~ COHPRESSOR 102 ~==::~~~~~~~~ SPOQ SPEEDS (% RPM) 100 f-----I-- ----4----f- -I- -+-- - ----< HPspOO:- I -----j"-----:---------, 98 I I EX HAUST GAS TEMP. (oc)

~~ f----fffi,--~ I

HP COMPRESSOR

INLET TEMP. : [-----r ·

(oC) 10r---~----~----~----~--~

TOTAL AIR TEMP. L.-- ___ :--_.......! __ -:-----+--J

(oC) 0 .... l ___ ---'-- ___ -'-- ____ -'-- ____ -'-- __ --J o 10 20 30 40 50 TIME (SEC(N)S) FIG.3 AIDS TRACE OF COMPRESSOR STALL ON A DC ·10 AIRPLANE (CF6-50)

CIT Sensor

With Rainshield Standard Sensor

FIG. 4 CIT SENSOR --------------- ---------------- ~----------------------- -- -- 1

VSV Tracking

Closed Original Tolerance VSV Angle Open

I

FIG. 5 VSV TRACKING

Aero Stalls

Fleet ---- 3 Month Roiling Average ------------- Monthly Events Per Million Hours J F M A M J J A SON DIJ F M A M J J A SON 0 1979 1980 FIG. 6 CF6 AERO STALLS

_____________________ J

oct;,. ENGINE : CF6· S0

KG / HR I I I

8500 ~- +-- +--+- - - -+ -"~--4 CORRECTED FUEL FLOW I 7$00 \----C----,vCL.- + - RPM ' 0' '0' ,00 ...

1 .0 ' .0 r-- i--+--+--t- - -t---If---i " CORRECTED N1 RPM FIG. 7 AI DS CF& GAS GENERATOR CURVES I

,---

APU OPERATING COSTS ,

7'

STUCTURAL INSPECTIONS, ETC.

I

I

, ,

i

I COSTS OF ENGINE TEST RUNS,

I

\

TEST CELL, U/S REMOVALS , SHOP VISITS \

I

I , NON QUANTI FYABLE I

I

\ ENGINE MATERIAL COSTS

I

I

I

I /

I

L ____ ~

I , I AIDS SAVINGS IN USD/ FLT HR.

OPERATING ~ COSTS I FIG. 8 PROJECTED A310 AIDS RETURN ON INVESTMENT I I I \ _ ___ ~ _________________________ ~ _____________________ .--J 1 __ - ____ _ OIAU ....

OFOAU I MlNe"J I OTHE"W.IHTlH .

TAS KS ANDUAilf AIOI COO P:'AHQfO .'1lI FIG . 9 AIRBUS AlIO AIDS SYSTEM DIAGRAM ENGINEERING MenON It lIeTION • lleTiON C ,--.!~ __ - _OICOPI - 10"" - ~ ... OlTllOAl - X· RAV - VlIUAL INWlCTK* - (1'c.

FIG. 10 AlIO AIDS GROUND SYSTEM LIrE USED FOR FIRST TUREINE BLRDE ~ LIrE USED ...

, •• •• 7 • ...

SI 4.

'0 o H D " ... " .J .J' ... 9 S~""":I:1"'BOHDJ'F 1979 1980 FIG 11 RMBIENT CONDITIONS

HH I \~j ~V":- \ ... ~ .. .. /~ .:/ ',\/' \ ' / ' '':---\ '/ ':'~A:~T i MAC .H .86

. ; . < ~ ... "' .. : : V . ,., 3seee .8 4

I, . .. . ' I

• • • : • ~ . • _ • ... • ~ \ po AL T 1 e I .. .. ~ , .l, . • \ .. ,' \" ... .. 7 . ... \ I \ ... , .. , .. : I . . I ...

/ .~ . \1 \ .: Ij , . . \. ' _ . ' ' . . :: \ , e 3eeee .82 . : -le

\ .... .. )V\J=~:~L;~~ .... ·

-2e . . .. . v. . . . . ; .. . . . ......... . ... ~ . . ....... ... . .... ;. .... . .. .. ... . . . . . . : ............ . ..... ; ..... . . . .. . . . . . . .. . . . . ... .. ..... ... .. ( -38

. I

-4e~----~----------~----~----~------~ (1 5 10 15 20 25 30 DATE OF "ONT ..

rIG 12 ~ DEVIATION FROM BASE 'GROSS DELTA ' CALCULATED MODULAR PERFORMANCE DEVIATION "

'1 . - ........ ... ... .... .. .. .. ..... . ........ .. ... .. . ..... . .... - 1

1.0 N1 ....... . ......... . ...... ~ . .. . ........... , );: '7'\: A

e.8 ' , .... "A )Q< I> < . ......

1 .~ 1 Fan lPe -.:;;;.,;,. : ~~~ 1

-1.8 _ ~:: t ~ ~ ~ mm ~ m .'m m m ..... '. .

'0(

'1 - .. .. .. . .. ..... . . . ... .... . 1

1.0 · HPC ........... .. .... 'T .. .. ... .. .. ..... ~ C> ~ ~ I' y

:::lH:. I ~fh~ : . ~ = : m T m I mmm 1

r:;;..., A • , D... •• 0.0. ~ 0' - 1.0 -1.0 '!~

mm

1 • ~t HPT .. ~ .. ...... ....... .. " .... ....... .

l'jW' : tts - tc51V:: m ·· ·· 1

A· . . '

I c ... ." •• ' ''' HI" .f'l( le n! " w 0.0 ~ F"A : ~ i / '- \i , ' , '>.

0.0 o . - . .... : / ...... - -1 - - \ - ~ - ............,. Ch_'.'". " . ( tl., ••

. _...... ._. J

\C . .

-1.8 -1 • 0 ~ .. .... .. ...... ...... ....... .. .

'1& ···· 1

::: A{t . 7S;: ~ : ~y .

-1.0 ~

do. 1 ~ 1 : [ T.m; .mmm. '1

1.0 t EGT , i .

I ~ ~ ~ ; ?oe> <'0> ¢¢ \ J"\ ~

0.0 ,'r , or:; \ r .

- I: f v= ~ ~ ~:~. : <=~~ : =

-1. 0 ' ..

(; 30 DA TE OJ' "ON n.

5 10 15 20 25 3 0 DA TE OF' ...... T ..

o s 10 15 20 25

rIG 1 4 F"IG 13 I I I L _____ ~ _______________________________________________ . ___ . _~_. ___ . ____ ._ . ____ . __ CALCULATED MODULAR PERFORMANCE DEVIATION ...

~::1~:~ .·.m • ... . . m , . ~.~ . .. ... . . , · ·1

'"

.51 HPC 1

0.0~~1--+-~1~~--+-~1~~~-+--~1--~41-~-+--~1 ~~ j -.5 I '5" 1 HPT .. ..

.

-----

~:: .. . '

~fPT

e .0 : : I ' 7" , : ' , , , -.6 . ... ~ . -. - . ,...~ .; . .. ~'-...... . ---.. .. ... / ~ -..J ' --- ~ I • -1.9 .... .. . .

d~O C ... _ .. .. .. .. . : .. ...... . : .. ... ... ~ . ... .. . .. ~ ....... . . ...... ........... . -., .. . ... . .. . ':' . ..... .

TIT . : ./ ~ ... _ .. .. .. , .... .. .. , .......--~ " ~ . ..... ~ .. ... ~ .. .. ... t . ...... ~ ........ ,: ........ , ~ . ~~ . • : .. . .. . :' ; '. " .

: L..----:-- ... / . "'- / ", .: . ~ ... .... . t" .. ... . ; .... .

o 1 2 3 .. 5 6 7 8 9 10 11 12 13 TOTRL RPPROXIMRTELY zee CYCLES Ref no fIG 15 I ------- - -- ------ - -- ----- - - - - -- - -- - ---------------- ------ HELICOPTER PROPULSION SYSTEM RELIABILITY AND ENGINE MONITORING ASSESSMENTS* John A. Murphy Bell Helicopter Textron SUMMARY Bell Helicopter is conducting a study of helicopter propulsion system reliability problems, specific technology solutions, and engine monitoring im- plications. The study is approximately 50 percent complete. Engine monitoring implications thus far include consideration of reciprocating engine monitoring, I I realization of maintenance cost savings due to derated operation, bookkeeping of I three-engine installations using two-engine cruise, monitoring of contingency J rating usage and integration of drive s y stem monitoring functions. Commercial I acceptance will require "up-front" demonstration of cost effectiveness.

I

I

I I i'iTRODUCT IO N I J

I

I Of the various helicopte r subsystems, the propulsion subsystem (i.e. en- I gines and drive train) h as the greatest impact on reliability, maintainability I and safety. A study by Boeing Ve rtol (R ef erence 1) of over 1500 FAA Malfunction I or Defect Reports (FAA Form 9330) for the U .S. civil turbine-powered helicopter I I fleet shows the propulsion subsystem accounts for: J I I 49.2 % of the failures I I 60.1 % of the unscheduled maintenance manhours I I 87.3 % of the repair parts co st I Similarly, an in - house stud y by Bell Helicopter Textron (BHT) shows that the propul si on subsystem is involved in 79% of the major, and 49% of the minor, material failure related accidents in a ~litary light helicopter fleet.

Again s t this background, BHT has been awarded a contract by ~ASA - Ames to conduct a stud y of "P ropulsion Sy stems Reliability and Integrated Engine Mon- itoring Technology As sessments for Civil Helicopters". The objective of the stud y is to increase civil helicopter productivity by improving propulSion sys- tem life, reliability an d maintainability through proper focusing of future re- s earch technology programs. Study tasks and schedule are shown by Figure 1.

Those tasks which are engi ne related such as derated engine characteristics, engine design changes, and engine cycle modifications have been subcontracted to Detroit Diesel Allison.

* Performed under NASA - Ames Contract NAS 2-10722, Dr. John Zuk, Technical r-lonitor.

I PROBLEM IDENTIFICATION The Statement of l"Jork for this task is to "Identif y the major short life, unreliable, and high maintenance engine and power transfer components and sub- s ystems in current civil helicopters. Categories shall include both reciprocat- ing and turbine engines, single and multiple engine configurations, single and tandem rotor vehicles, and light, medium and heavy helicopters".

This task was approached in three ways: - Accident rate data - Maintenance rate data - Direct operator input Accident Rate Approach For this approach, the U .S. civil helicopter population was determined, us- ing the latest available data (Reference 2), with results as shown by Table 1.

This population was broken out by the various study categories as shown by Table 2 . Also shown are the flight hour distributions of the various catego ries, as determined from Reference 3. Accident data for this population and time per- iod, as reported by the ~ational Transportation Safety Board was then examined.

A total of 2302 helicopter accidents occurred, of which 1472 (64 % ) involved no material failure of the helicopter. However, of the 831 which were material failure related, 586 (70 %) had an engine or drive train related cause factor.

Table 3 shows a breakdown of accidents by engine type (reciprocating - or turbine).

Considering the 469 accidents which had a powerplant related cause factor, 354 (75.5 %) involved reciprocating engine models, while the remaining 115 (24.5 % ) involved turbine powered models. These numbers in themselves are meaningless, however when compared to either the population percentages or the flight hour percentages shown in Table 2, it is evident that reciprocating engine powered helicopters are involved in a disproportionately large percentage of powerplant related accidents. Similar results were obtained for drive train related acci- dents, and for accidents due to all causes. This suggests that perhaps some of the automotive technology which is emerging for monitoring reciprocating engine s might well be applied to aircraft.

Similar breakdowns of accident cause factors were made for the other study categories, i.e. single vs twin, light, medium and heavy weight, and single vs tandem rotor. These breakdowns are handicapped by the small twin, heavy, and tandem populations, and did not yield any particularly startling results.

~~intenance Rate Approach For this approach, an analysis was made of U.S. Navy ~~intenance and Material Management (3M) data for the models, systems, and characteristics noted I I

------------ _____________________ J

in Table 4. In this gro up, the twins, he a vies, and tandems are better repre- sented than in the civil population, as shown by Table 5. Typical results are shown by Figure 2, which is for engine related problems . The parameters shown across the top are: MFHBF - Mean Flight Hours Between Failures MFHBMA - Mean Flight Hour s Between (Unscheduled) Maintenance Actions MMH/FH - Maintenance Man Hours per Flight Hou r MMH/MA - Maintenance Man Hours per Maintenance Action EMT/MA - Elapsed Maintenance Time per Maintenance Action The different categories are shown along the left. When categorized by weight, it can be seen that altho u gh failure rates are similar, the light heli - copter is much easier to repair. When number of engines is considered, it is interesting to note that }lliH/FH for the twins is exactly twice that of the singles. Catorizing by type of rotor system has little effect on engine Relia- bility and Maintainability (R&M) characteristics, as would be expected. Sum- maries similar to Figure 2 were also prepared for the other systems listed in Table 4.

Direct Operator Input Direct operator input to the study was obtained by visits to two large gulf coast commercial helicopter operators . The Petroleim Helicopter, Inc. heavy maintenance facility in LaFayette , Lousiana was visited on 1 October 1980. The following day, the Air Logistics facility near New Iberia, Louisiana was visited.

Table 6 lists the major R&M concerns mentioned by maintenance management person- nel at these facilities. It is significant that no one component, subsystem, or issue was of overriding importance . With regard to engine montoring systems, the concern is that a dedicated piece of electronic equipment, with associated wiring and connectors, will be just another maintenance burden. This feeling is based on a long history of electronic equipment problems in the gulf coast environment. There must be an obvious economic payback from the monitoring sys- tem to trade-off, against its possible liabilities.

Problem Identification Conclusion It is concluded that there is no one overriding R&M issue, but rather a broad spectrum of areas i~ need of improvement .

Technology Solutions Severql technology solutions to the overall problem of improving R&M char- acteristics were investig a ted , as shown by Figure 1. These are discussed as follows, with regard to engine monitoring implications .

J Engine Derating Study The objective here was to determine the relative benefits of derating a current technology engine, vs utilizing a non-derated advanced technology engine, to improve R&M characteristics.

"Derating", in this case, means to install an "oversized" engine, while maintaining the same aircraft performance. A "baseline" situation was assumed such that the "zero derate" case consisted of a Bell Model 206B "JetRanger III" with Allison 250-C20B engine, operating on a mission which requires "takeoff" (red-line) power for takeoff and landing, and "maximum continuous" power for cruise. Allison provided data for scaling engine size, weight, cost, specific fuel consumption (SFC) , and maintenance cost to twice baseline rated power.

Engines derated (i.e. oversized) in 10 percent increments were parametrically installed in the aircraft. Aircraft weight, cost, and fuel consumption were allowed to increase as engine size increased. The net result is shown by Figure 3. As expected, the cost of fuel goes steadily upward, due to higher power required (because of increased gross weight), and increased SFC due to part power operation. Insurance and depreciation increase due to increasing aircraft cost caused by increased weight empty and increased engine cost. Maintenance cost decreases due to operating at reduced gas temperatures. The net effect is negligable out to about 20 percent derating. Data for an advanced technology engine, rated at baseline power, were also provided by Allison and used in the same manner. In this case, aircraft weight and power required decreased, so that this engine was in effect derated also. The net effect is a significant saving in operating cost. To put these numbers in perspective, it should be noted that the current leasing rate for this aircraft is around $350/hour.

The engine monitoring implication is that the maintenance saving shown will not be realized without a monitoring system. This is because a large percentage of the saving is due to extended Time Between Overhaul (TBO) and component re- tirement times, as specified by the engine manufacturer. In order to grant these extensions, the manufacturer must be assured that the engine is in fact being operated in a derated mode. A fool proof, continuous on-board monitor will be required for this assurance.

Fuel Control Fuel control improvements are centered around the Full Authoricy Digital Electronic Control (FADEC). This type of control is well suited to performing engine monitoring functions.

Configuration Changes One of the configuration changes investigated is the effect of number of engines. One-Engine-Inoperative (OEI) requirements are increasing in importance, as evidenced by the number of new twin-engined designs entering the market. One problem, of course, is that the initial cost and operating cost of a twin- ------- --------------------------- ~- --- - --~----~-- ~ ---, engined helicopter is considerably greater than that of a comparable single- engined design. This could be aleviated somewhat by special "contingency" ratings, to be used only in OEI situations. This would increase the power avail- able from the remaining engine if one engine were lost, thus permitting a reduc- tion in total installed power, weight, and improved SFC since the engines would normally operate at a higher percentage of rated power. This subject is thor- oughly discussed in Reference 4.

Actual usage of the contingency rating must be closely controlled. Special inspections, parts replacements, or TBO reductions may be required if the rating is used. An engine monitor will be required to record the extent of usage of the contingency rating.

Another solution is to use three engines instead of two. Since loss of an engine would reduce the power available by 33 percent instead of 50 percent, total installed power can be reduced. Also, with three engines, one engine can be shut down for long range cruise, improving the power match and reducing SFC.

The engine monitoring implication is that a system would be required for book- keeping of engine operating time and cycles. Manual bookkeeping would be in- ordinately complex and error prone if the two-engine-cruise mode were frequently used.

Transmission Sy stem Improvements Transmission and drive system monitoring could be incorporated into the engine monitoring system. One low-cost candidate function would be to record actuations of the various "chip detectors". These are magnetic devices in the lubrication system which collect magnetic debris from the oil. Collection of sufficient debris completes an electrical circuit and turns on a cockpit light.

Detectors are now available which will self-clear normal wear particles. Fre- quent self-clearing, however, may indicate an incipient problem. Therefore, an indication of the frequency of clearing operations, whether automatic or pilot initiated, would provide useful diagnostic information. More sophisticated monitoring techniques based on vibration signal analysis are also available and could be incorporated into the monitoring system.

ENGINE MONITORING CONCLUSIONS Engine monitoring implications of the study which have evolved thus far are summarized as follows: Reciprocating Engine Monitoring All aircraft engine monitoring programs of which we are aware are directed at turbine engines. However, most civil helicopters are powered by reciprocat- ing engines. Although this percentage of the population is decreasing, it is estimated by the FAA that "recips' will still represent 40 percent of the popu- L __ _ 1ation in 1992. Accident statistics indicate that reciprocating engines could benefit from a monitoring system. Automotive technology in this area is rapidly emerging, and should be considered for adaption to the general aviation fleet .

Extended TBO and Retirement Lives In establishing TBO's and retirement schedules, engine manufacturers must assume "worst case" operation, since actual engine usage is unknown. Scheduled maintenance based on actual usage would be feasible if such usage were recorded by the monitoring system.

Engine Record Bookkeeping Accurate logging of engine starts, run time, time above certain power levels, and engine cycles is a valuable monitoring function, even for a single - engine aircraft. This function becomes more useful as the number of engines increases, and is a practical necessity if engine run times are unequal, i.e. OEI cruise.

Contingency Rating Usage If a "contingency" or "emergency" power rating is available for infrequent, short-term use, actual usage of this rating must be closely controlled . By re- cording such usage, the engine monitor can trigger any required maintenance actions.

Transmission and Drive System Monitoring If the engine monitor is to be used in a helicopter, integration of trans- mission and drive system monitoring functions should be considered. Candidate functions are chip detector actuations, chip detector clearing operations, bearing temperatures, and vibration signal analysis.

Cost Effectiveness Commercial operators are concerned that an engine monitor might cost more than it saves. The monitoring system must be tied to an extended or "on- condition" maintenance schedule so that cost effectiveness will be evident "up- fron til, not as a hopeful result.

-""---~-- -~--- --- - - -- REFERENCES 1. Dougherty, J.J. and Barrett, L.D.: Research Requirements to Improve Reliability of Civil Helicopters.

NASA CR-145335, April 1978.

2. Census of U.S. Civil Aircraft, Calender Year 1978. U.S. Department of Transportation, Federal Aviation Agency, Office of Management Systems.

3. General Aviation Activity and Avionics Study, U.S. Department of Transportation, Federal Aviation Administration, Office of Management Systems.

4. Sample, R.D.: Emergency-Power Benefits to Multi-Engine Helicopters.

Journal of the American Helicopter Society, July 1977, pg. 27.

U. S. CIVIL HELICOPTER POPULATION

CY 1978 MANUFACTURER NO. OF MODELS NO. REGISTERED AEROSPATIALE 7 239 AGUSTA 1 12 7 3209 BELL BRANTLY 2 153 2 369 ENSTROM HILLER 2 85 HUGHES 2 1728 KAMAN 1 19 MBB 1 61 SIKORSKY 11 383 2 23 VERTOL TOTAL 162811 MISC 173 TOTAL 6454 TABLE 1 HELICOPTER POPULATION SETS PERCENT MODELS REGISTERED POPULATION FLT. HOURS CHARACTER I STI C (NUMBER) (NUMBER) 1978 1970-1978 14 3779 60 .2 46 . 4 RECIP TURBINE 24 2502 39.8 53.6 27 5979 95.2 95.4 SINGLE ENGINE TWIN ENGINE 11 302 4.8 4.6 21 5549 88.3 91:0 LIGHT MEDIUM 11 627 10.0 8.5 HEAVY 6 1.7 0.5 S INGLE ROTOR 35 6239 99.3 99 . 9 42 0.7 TANDEM ROTOR 3 0.1 TABLE 2 I

__ ~ ___ ~ ________________ ~ __________ J

ACCIDENT CAUSE FACTORS BY ENGINE TYPE FLT . HOURS "!o ACCI DENTS ACCI DENTS POPULATION "!o ACCIDENTS TYPE "!o "!o "!o "!o POPULATION "!o FLT. HOU RS POWERPLANT RELATED RECIP 75.5 60.2 46.4 1. 25 1. 63 TURBINE 24 .5 39 . 8 53.6 0. 62 0.46 DRIVETRAIN RELATED RECIP 76.1 60 . 2 46 . 4 1. 26 1.64 TURBINE 23 . 9 39 . 8 53 .6 0.60 0.45 ALL CAUSES RECIP 75.9 60 . 2 46.4 1. 26 1.64 TURBINE 24. 1 39.8 53.6 0.60 0. 45 TABLE 3 3M DATA ANALYSIS T I ME PER I 00 : 1 YEAR 4/79 - 3/80 FLIGHT HOURS : 15000 - 27000 MODELS SYSTEMS CHARACTER I STI CS TH-57 ENGINE MFHBF UH-IE MAIN DRIVE SHAFT MFHBMA UH-IN MAIN TRANSMISSION MMH IFH SH-2F TAIL ROTOR DRIVE SHAFT MMH/MA SH-3H HANGER ASSEMBLY EMT/MA CH-53D INTERMEDIATE GEARBOX CH-46D TAl L ROTOR GEARBOX TABLE 4 3M HELICOPTER CHARACTERISTICS WT. ROTOR NO. CIVIL MFGR MODEL ENGINES CLASS SYSTEM EQUIV.

FLT . HOURS TH-57 1 S 2!XJ 18079 BELL L BEL L UH-IE 1 M S 205 14990 UH -IN 2 M S 212 27536 BELL SH- 2F 2 KAMAN M S NONE 16921 SIKO R SKY SH - 3H 2 H S S-61 25440 SIKORSKY CH -5 3D 2 H S NONE 15505 CH-46D VERTOL 2 H T 107 16890 TABLE 5 OPERATOR CONCERNS • ELECTRON I C FUEL CONTROLS • ENGINE MONITORING SYSTEMS • ENGINE ACCELERATION DEVI CES • DRIVE SHAFT COUPLINGS • OIL LEAKS • ENGINE POWER CHECK PROCEDURES • CORROSION • REPAIRABILITY • MODULAR INTERCHANGEABILITY • TWIN ENGINES • DERATING • MILITARY TECHNOLOGY TRANSFER • INSPECTION REQUIREMENTS TABLE 6 I - - - --- ------------------------------------------------ --- --- - I

I

PROGRAM TASKS AND SCHEDULE 1980 1981 J F M A M J J A S S 0 N D IDENTIFY PROBLEM TASK 1 TASK 2 TECHNOLOGY SOLUTIONS ENGINE DERATING

-: FUEL CONTROL

CONFIGURATION CHG.

=-

ENGINE DES IGN CHG.

XMSN DESIGN CHG .

•

ENGINE CYCLE MOOS

-

--

TASK 3 - DIAGNOSTICS TASK 4 - EMERGENCY POWER TASK 5 - VEHICLE COMPARISON TASK 6 - RESEARCH OPTIONS r- BRIEFING - AMES "~A BRIEFING - LEWIS .6 QUARTERL Y REPORT A .6 A BRIEFING REPORT

.. ~

~ DRAFT FINAL REPORT r-- r-- F INAL REPORT '-- FIGURE 1 3M SUMMARY - ENGINES PARAMETER MFHBMA MMH/FH MMH/MA EMT/MA MFHBF WEIGHT CLASS NUMBER OF@;.6 ENGINES 2 ROTOR SYSTEM FIGURE 2 DERATING STUDY RESULTS + 20 + 10 6. COST $/HOUR

o

-10

r ADVANCED TECHNOLOGY, NET

- 20 +----t=h.=----+----t-----t

o 10 20 30 40

EFFECTIVE DERATING "'" "/.

FIGURE 3 I I I , ~--- - I ENGINE HEALTH MONITORING SYSTEMS l TOOLS FOR IMPROVED MAINTENANCE MANAGEMENT IN THE 1980 s Jonathan C. Kimball

Pratt & Whitney Aircraft Group

ABSTRACT The increased cost of fuel has placed an added importance on the "performance health II of cOl11T1erical aircraft engines and greater emphasis on the performance-monitoring aspect of maintenance. This paper presents an

overview of Engine Health Monitoring activities at Pratt & Whitney

Aircraft. The development of Engine Health Monitoring, a description of systems currently used, and a sUl11T1ary of programs for improved monitoring l in the 1980 s are discussed.

INTRODUCTION The first generation of commercial gas turbine engines (early JT3D and JT8D models) was largely maintained on a "hard time" or lias required" basis. Engine maintenance was performed when certain parts achieved a pre- determined life limit (hard time) or when specific symptoms indicated main- tenance was needed (as required).

Hard-time maintenance employs the same maintenance schedule for all similar engines. Since no two engines perform exactly alike, the schedule 1I is based on an "average engine. Differences in engine performance may be the result of manufacturing variations and variations in engine mission and service experiences. Some engines may become fuel inefficient although safe and capable of continued operation.

There is, therefore, a need for scheduling maintenance on an on- condition, individual basis. Engine Health Monitoring satisfies this need by continually monitoring engine performance and providing the diagnostic tool for interpreting changes in performance in terms of maintenance requirements.

ENGINE HEALTH MONITORING AND FUEL COSTS The repalrlng and refurbishing of engines at fixed intervals or when required has long been a standard practice in the airline industry and will probably persist at many airlines. These procedures have two advantages: 1) they are easily managed and 2) as-required maintenance tends to result in engines remalnlng on the wing for longer periods. However, these practices do not emphasize minimizing fuel costs.

Engine Health Monitoring can provide visibility into the performance levels of each engine, allowing better maintenance planning, as illustrated in figures 1, 2. Figure 1 shows a typical cycle of engine operation and repair, and figure 2 shows a repair schedule customized to the performance level of each engine. Fuel savings is the primary benefit . Since all engines do not deteriorate at equal rates, hard-time maintenance results in engines with high fuel consumption remaining in service too long and engines with low fuel consumption being repaired or refurbished too soon .

Using a system of Engine Health Monitoring to provide a customized maintenance schedule for each engine has benefits in addition to fuel sav- ings. Lower overall maintenance costs are possible because engines with the highest deterioration levels will be repaired rather than left in service where they would be exposed to increasingly high turbine temperatures. Sim- ilarly, lower repair costs on engines with the lower deterioration levels can result by deferring the repair. An additional benefit of engine moni- toring is greater reliability. Many developing engine problems can be fore- seen with an engine monitoring system.

High fuel costs increase the need for determining repair requirements on an engine-by-engine basis with visibility from an Engine Health Monitor- ing system. The effect of fuel cost on repair interval is illustrated in figures 3, 4, 5. The total engine operating cost per flight cycle is the sum of the repair cost and fuel cost (fig. 3). As the time between repair increases within a reasonable range, repair costs per flight hour decrease and fuel costs increase, resulting in an interval of engine repair time over which total operating costs are minimized. Higher fuel costs shorten the optimum repair time and narrow the optimum band (fig. 4). Since the performance of each engine varies, the optimum repair intervals for ind i - vidual engines are not the same (fig. 5).

Engine Health Monitoring can determine which engines require early or deferred -maintenance to provide the lowest total operating cost.

DATA ACQUISITION SYSTEMS FOR AIRBORNE ENGINE HEALTH MONITORING Systems used in flight to acquire data for Engine Health Monitoring vary considerably in both complexity and capability. The four systems most widely used by airlines today are illustrated in figure 6.

The most basic system (System 1) is also the system most commonly used today. The flight crew manually records the data, which can then be processed either manually or by computer.

I

I

I I I

I

____ _________________________________________________ J

Automated data acquisition systems expand on the capabilities of the mandatory flight data acquisition unit and digital flight data recorder. In System 2 (fig. 6) a flight data entry panel allows the crew to input docu- mentary data (date, engine time, etc.) to the system. Data is recorded on cassettes (by means of the quick access recorder), which are easily removed for processing at a ground station and analyzed by computer.

An important addition to the data acquisition system is the on board computer, data management unit, which allows the system to perform many additional functions. ~ith this unit the system can record data selec- tively, looking for appropriate parameter range and stability criteria. It can also , make calculations and provide results to the crew by means of an on-board printer. If desired, the computer can scan the data and notify the crew of limits that have been exceeded. In system 3 (fig. 6), data acquisi- tion is controlled by the data management unit, and the data is recorded with the on-board printer.

System 4 (fig. 6) is the most complex and capable of the in-flight data acquisition systems. Data recording is controlled by the data manage- ment unit. There is also an auxiliary data acquisition unit for more exten- sive analyses. Data can be stored on the quick access recorder (cassette) or on printouts from the on-board printer. The data is normally processed by computer, but can be analyzed manually if desired.

Automated data acquisition systems have, however, experienced a var- iety of problems. Problems have been experienced with unreliable instrumentation and inaccurate data, burdensome calibration and maintenance requirements, and difficulty with data management.

In order to improve the reliability and accuracy of the data acquired

during flight, Pratt & ~hitney Aircraft and Hamilton Standard jointly

developed the Propulsion Multiplexer, an integral engine data acquisition system for acquiring the high quality data required for module performance analysis. The Propulsion Multiplexer (fig. 7) is a compact, durable system, housing pressure transducers, a microprocessor, and all required elec- tronics for acquiring and sending multiplexed engine data to a recording device, such as the Airborne Integrated Data System.

INTERPRETATION OF ENGINE DATA ANALYSIS APPROACHES Techniques for analyzing engine data obtained in flight or in test cells can be catagorized as follows: o Limit Exceedance Checking o Parameter Trend Monitoring o Module Performance Analysis o Turbine Life Accounting I I I I I I I I I I I I I

______ J

1- __ _ Checks of specific engine parameters for operation in excess of limits is a fundamental requirement of post repair engine testing. In the test cell, compliance with engine limits indicates that an engine is suit- able for service. Limits exceeded in flight can be monitored with an Airborne Integrated Data system. The information can provide cautionary warnings to the flight crew (similar to such warnings as low oil pressure) or help to define maintenance actions (as with Exhaust Gas Temperature exceedances).

Graphic display of key engine parameters is the most common method of monitoring the health trends of an engine in service (airborne data) or the performance trends of post-repair engines (test cell data). These plots provide indications of engine performance trends or instrumentation mal- functions.

The Engine Condition Monitoring Computer Program developed by Pratt &

Whitney Aircraft has long been widely used to provide graphical parameter trends to improve visibility of the health of engines during service. The

program can be used with all Pratt & Whitney Aircraft commercial engines

and can operate with flight data acquired either manually or automatically.

Manual data is recorded by the flight crew and then processed at a later date. Automatically recorded data is provided by an Airborne Integrated Data System.

The output of the Engine Condition Monitoring Program of primary importance to the user is the "plot report," which presents chronological trends of engine parameter shifts (fig. 8). Because engine parameter shifts are highly visible on the plot report, timely detection of developing engine problems is possible. The report also provides visibility into the long term deterioration trends of an engine or fleet and allows detection of large errors in measured parameters.

The primary advantage of the program is that a large amount of in- formation about engine condition can be obtained without additional engine or airframe hardware, providing considerable benefits with little cost. A limitation to the program is that although it can recognize that a problem has occurred, it can not diagnose the cause. The user must apply judgment to determine the nature of the problem, and if necessary, request further investigation with other troubleshooting methods. For example, an exper- ienced analyst would be required to distinguish between a bleed valve mal- function, a damaged engine module, or an error in measured engine pressure ratio.

Module Performance Analysis is a technique for using measured engine parameter shifts to determine specific engine module performance changes.

This process can be illustrated using the example in figure 9. Measured parameter changes are first determined (shift in corrected high rotor speed (%AN2) at a constant engine pressure ratio, for example). The analysis is used to calculate the most likely cause of these shifts, such as deterioration in high-pressure turbine efficiency. Finally, the shift in I I

__________________________________________________________ J

key parameters attributable to each module can be calculated. For example, exhaust gas temperature may have increased by 20 C relative to a new engine. The analysis will tell a user how this 20 C can be accounted for (e.g., 10 C due to high-pressure turbine deterioration, SoC to fan performance losses, and SoC to low-pressure compressor performance" losses), thus indicating areas that may need maintenance.

Module Performance Analysis is currently most often used with test cell data as a tool for evaluating the effectiveness of a repair. The an- alysis is also used on prerepair data specifically acquired to help define shop work scope. The JT9D Test Cell Module Analysis Program, developed for analysis of JT9D engine data acquired in the test cell, combines a sea level data reduction system with module performance analysis and data validity screening. The program is very flexible, accomodating varied data input, analysis baselines, and test cell corrections. A sample output is shown on figure 10.

Module Performance Analysis systems have also been developed for use with data acquired in flight with an Airborne Integrated Data System. Data for inflight module performance analysis, including the additional param- eters required for module performance analysis, must be recorded automatic- ally by an airborne data system. The on-board computer of this system selects what data is to be recorded based on predetermined ranges of engine and aircraft parameters (data acquisition windows) and parameter stability criteria. Data from the airborne system may be manually input to the air- borne module performance analysis program using data from an on-board printer or automatically using data transferred from an on-board recorder.

A typical program output, the module analysis plot report, is shown in figure 11. This report presents graphical trends of performance changes of each module in a highly visible format even if a module has been instal- led on a different engine.

The program can provide many benefits to a user. Knowledge of the performance of each module can be helpful in making maintenance decisions.

For example, if an engine has a history of high exit gas temperature, the plot tells whether the high temperature is caused by the high-pressure com- pressor or the high-pressure turbine, or both. Appropriate maintenance can be planned. The In-flight Module Performance Analysis can be a useful tOO"1 for troubleshooting engine problems on the wing and can assist in improving shop scheduling.

The JT9D Airborne Integrated Data System/Module Performance Analysis Program is currently being used and evaluated at four major airlines. The program is emerging from the developmental stage and may soon be considered a developed engine monitoring tool.

Special attention must be paid to instrumentation in order to suc- cessfully perform module performance analysis. Parameters not normally measured are needed, as shown in Table I. In addition to the parameters normally acquired, temperatures at the discharge of the low-pressure com- pressor, high-pressure compressor, and low-pressure turbine are measured.

Special emphasis on data accuracy is also required for reliable Module Performance Analysis results. For example, if a fuel flow measure- ment is used only to determine engine suitability for service, a measure- ment error of 2% may go unnoticed. If the data is to be used for module performance analysis, a 2% error in fuel flow may be misinterpreted as an engine performance shift. Although the analysis systems now available have provisions for detection of erroneous data, a greater emphasis on data quality is necessary.

Module performance analysis capability, therefore, is an extremely useful engine maintenance tool. The use of module performance analysis with data from either a test cell or an Airborne Integrated Data System requires a commitment to additional instrumentation, closer instrumentation accuracy monitoring, and personnel trained in module performance analysis interpre- tation and use.

I

A Life Accounting Program can be used to calculate the fraction of

I

life consumed for any set of critical high-pressure turbine airfoils. Since all routes are not equally severe on high-pressure turbine airfoils, large variations in part lives can exist. The life accounting program calculates I the amount of life consumed for each critical airfoil, using analytical models of airfoil deterioration. The program can run as a subroutine of the

I

Airborne Integrated Data System/Module Performance Analysis Program, using the accumulated time exposure of the parts to temperature, pressure, and

I

rotor speed. The program can also be run by itself without airborne data, using the specific route structure and engine derate experience as input. A typical output from the life accounting program is shown in figure 12.

I

The primary purpose of the program is to maximize airfoil service

I

life while minimizing the possibility of turbine damage. The program can also assist in efficiently scheduling hot-section maintenance, controlling inventories of airfoils, and in better planning of the hot-section assembly

I

(e.g., a turbine could be assembled with airfoils having similar amounts of life remaining).

FUTURE TRENDS Increasing airline fuel and maintenance costs have resulted in great- er airline interest in Engine Health Monitoring. Since many aspects of this process can be addressed most efficiently by the engine manufacturer, Pratt

& Whitney Aircraft is committed to providing superior Engine Health Moni-

toring system support.

A special emphasis will be placed upon data quality in future engine health monitoring systems. A common shortcoming of engine monitoring sys- terns today is that analysis algorithms, although accurate, are unacceptably sensitive to sensor errors. The approach to data validity must be three fold: 1) encourage the development and proper use of accurate data measurement systems 2) develop software routines that recognize and report probable data errors for follow-up maintenance actions 3) design algorithms to be as insensitive to data errors as pos- sible.

Present plans are to continue to develop and refine analysis software routines and to monitor current systems, making improvements where re-

quired. We will work with customers to define ways of monitoring different