Document
JT9D JET ENGINE PERFORMANCE DETERIORATION* A. Jay, E. S. Todd and G. P. Sallee Commercial Products Division PRATT & WHITNEY AIRCRAFT GROUP SUMMARY Escalating fuel costs and the need’ to meet national energy conservation goals have led to a new industry awareness of the importance of maintaining good fuel consumption through- out the life-cycle of an engine. However, higher fuel consumption is only part of the overall engine deterioration picture which consists of reduced surge margin, higher exit gas tempera- ture, and other hot section distress. Engine deterioration characteristics cq in general, be divided into two time periods. The first, called short-term deterioration, occurs in less than 250 flights on a new engine and in the first few flights following engine repair. Engine de- terioration in the second time period, characterized as long-term, involves primarily hot section distress and compression system losses which occur at a somewhat slower rate than short-term deterioration.
It is generally accepted that the causes for short-term deterioration are associated with clearance changes which occur in the flight environment. In this paper, the analytical tech- niques utilized to examine the effects of flight loads and engine operating conditions on per- formance deterioration are presented. The role of gyroscopic, gravitational and aerodynamic loads are shown along with the effect of variations in engine build clearances. These analyti- cal results are compared to engine test data along with the correlation between analytically predicted and measured clearances and rub patterns. Conclusions are drawn and important issues are discussed.
INTRODUCTION The current and projected high cost of fuel for gas turbine engines places a premium on incorporation of design features which increase the operating efficiency of aircraft propul- sion systems. One such feature, universally recognized to be of major importance, is the maintenance of tight operhting clearances between static and rotating components of flow- path seals. In practice, this is rather difficult to accomplish since the individual seal compon- ents and their supporting structures experience wide excursions in temperatures, rotational speeds, and other loadings at different points in the flight cycle which give rise to relative deflections that can lead to contact, wear, and increased clearances between seal parts. Early gas turbine designs (turbojets and low bypass ratio turbofans) accounted for these time vary- ing loads and associated deflections as part of the standard process and attempted to tune rotor and case growths such that tight clearances (maximum efficiency) would occur during steady-state operation (climb, cruise) without introducing damaging rubs during transient conditions (takeoff, landing). In view of the ready availability of inexpensive fuel and other * Workperformed.under NASA contract NAS3-20632 factors, designs which dropped a few percent in efficiency after a year or two in service were considered adequate at that time.
Today’s situation is quite different as a consequence of two factors, First, fuel costs have more than doubled in the past few years and are expected to continue to rise, and sec- ond, higher bypass ratio engines are more susceptible to structural deformations which can cause tight seal clearances to be degraded by rubs. The second factor follows from the larger size (increased thrust and air loads) and increased thrust-to-weight ratio (lightweight, flexible structures) characteristics of the turbofan engines which power current commercial ;transports. In order to defme powerplant configurations which will meet the more stringent ‘performance retention requirements of tomorrow’s marketplace, today’s designer must have at his disposal a more advanced set of analytical tools with which to anticipate the response (deflections) of an engine to its flight environment (loads) than was previously necessary.
This paper describes progress that has been made toward development of a comprehen- sive analytical procedure for predicting the effects of flight loads on short-term gas turbine performance deterioration (fig. 1). The damage mechanism considered is wear of flowpath outer airseals due to interference of rotating (blade tip) and stationary (rubstrip) seal com- ponents. Wear behavior of inner airseals is more complex and has been omitted from the current study. Other mechanisms such as erosion and contamination which decrease engine efficiency more gradually than seal rubs are deemed to be of secondary importance in short- term deterioration and have been excluded.
SYMBOLS Values are given in SI units. The measurements and calculations were made in U.S.
Customary Units.
Engine inlet Mach number Mn % A TSFC Percent change in engine thrust specific fuel consumption Airplane stall velocity VS Performance influence coefficient for stage j, condition i ij T* Average clearance change for stage j J ANALYTICAL MODEL This section provides a general description of the analytical procedure employed to assessthe effects of steady flight loads on short-term performance deterioration of the JT9D-7/747 propulsion system. In essence, the model provides a vehicle for predicting blade tip rub damage caused by structural deformations which occur during flight opera- tion-and relates the corresponding enlarged seal clearances to increases in engine thrust specific fuel consumption (TSFC).
Flight Profile Definition The starting point for all deterioration predictions to be discussed in this paper is a de- scription of the sequence of operating conditions or events which comprise an engine mis- sion or flight profile. The cycle may be relatively simple in terms of power level changes and exposure to external loads, as is the case for standard “green runs” and engine acceptance tests on the ground, or may encompass load spectra from runway roughness to clear air tur- bulence which are commonly encountered by commercial airlines. Each such cycle is con- structed from a series of time segments (start-up, taxi, takeoff, climb, cruise, descent, ap- proach, landing, shutdown), the end points of which can be characterized by unique com- binations of aircraft and engine operating parameters (gross weight, altitude, attitude, Mach number (Mn), rotor speeds, temperatures, pressures, flows) that serve to define boundary conditions for subsequent aerodynamic, thermodynamic and structural analyses. For this paper, attention has been primarily focused on the airplane acceptance test flight (figs. 2, 3) chiefly because the profile is well defined and secondly, test data defining the magnitude of the change in TSFC is available for a number of engines covering the flight cycle range of interest. Ground tests and fleet service which precede and follow the flight acceptance test, respectively, have also been included but in somewhat less rigorous fashion. Quantitative in- formation on idealization of these cycles for the JT9D-7 engine is discussed later.
Loads and Structural Deflections Temperature, pressure and centrifugal force fields play an important role in determin- ing internal seal clearances. These are always present during engine operation. Perhaps, the most convenient feature of this set is the common assumption that circumferential variations in these fields are small and, for the purposes of deflection analysis, can be neglected. The second important characteristic is that each field varies appreciably in response to changes in power level and requires a transient analysis for proper representation.. Specialized com- putational procedures have evolved to perform the secondary flow, heat transfer, and other analyses that define temperature, pressure, and rotor speed time histories for desired flight profiles. These loads are input to axisymmetric structural analysis programs which generate histories of relative deflections (gap closures) between static and rotating components. Com- bination of the axisymmetric closures with values for the initial build clearances (cold gaps, also assumed to be uniform) then provides the sought after hot clearances as functions of time. Since they essentially indicate the gaps available for accommodation of additional de- flections due to external flight loads, plots of these data will hereafter be referred to as base- line clearance curves.
A second set of structural deformations is related to loads which are not uniformly distributed with respect to the engine centerline. Generally, this set arises from external motions or restraints imposed by the flight environment and is composed of airloads (inlet lift), maneuver loads (g’s, gyros), and thrust (including thrust reverse). As would be expec- ted, consideration of these loads and their contribution to the performance deterioration A NASTRAN problem presents a greater challenge than was the case for the previous group.
finite element model of the JT9D-7/747 was required to simulate the engine’s response to external loads (figs. 4, 5.).
The burden of defining cowl pressure distributions (au-loads) and maneuver load factors for candidate flight missions has traditionally been borne by the airframe manufac- turers. Since these data are usually supplied to Pratt & Whitney Aircraft (P&WA) only in gross form (force/moment resultants, design limits/envelopes), provision was made for Boeing Commercial Airplane Company (BCAC) to generate detailed pressure load descrip- tions. Conversion of internal and external pressure distributions into appropriate descrip- tions of thrust and thrust reverse loads was also performed by P&WA and BCAC, respectively.
Nodal forces consistent with inlet cowl pressure distributions, internal thrust build-up, man- euvers, and thrust reverse loadings were applied to the NASTRAN model and corresponding rotor/case displacement solutions obtained.
Blade-Tip/Rub-Strip Damage Calculation and Performance Deterioration The process whereby structural deflections are translated into blade-tip/rubstrip dam- age involves calculations for a sequence of time points selected from a given flight profile.
For each time point, the effects of axisymmetric loads (baseline clearances), engine offset grinds, and rub damage from previous time points are combined to establish the circumfer- ential variation of clearance that is available for accommodation of non-axisymmetric struc- tural deformations. Asymmetric rotor/case deflections are then introduced and when the relative closures exceed the available gap, the extent of local interference is recorded.
Finally, wear characteristics of the contacting materials are considered to determine the trade-off between blade-tip/rub-strip damage due to the interference. Gap changes caused by shortened blades and the worn rubstrip are in turn carried forward to appear as increased initial clearances for the next time point. At the end of the cycle, the accumulated damage for each rub-strip is circumferentially integrated and added to blade-tip wear to provide the average clearance change for the stage.
The final step to be taken involves conversion of permanent clearance changes for the total cycle to increases in TSFC under standard performance conditions. This is accomplished by simply summing the contributions from each stage, or, (% A TSFC)i = ~ ~ij ~ all stages The influence coefficients (5) are unique to a particular engine model.
ENGINE DETERIORATION SIMULATION Airplane Flight Acceptance Test The production flight acceptance test was selected for simulation because every 747 off the assembly line is tested this way, according to a routine that is kept as standard as possible. The purpose of the flight acceptance test is to check out airplane internal systems.
Airplane takeoff gross weight, airspeed, and throttle setting vary somewhat from one test to another because pilot instructions are given in terms of obtaining a signal from a warning or control instrument rather than in terms of achieving a specified flight.condition.
Operating conditions to be considered as part of the flight. test are defined in terms of rotor speeds, pressures, and temperatures from engine performance tables along with tlight- related parameters such as attitude, altitude, inlet Mach number, airplane weight, and fuel distribution in the wing. Airloads present on the inlet are described for the flight acceptance profile. Thermal, pressure and centrifugal loadings are accounted for by the use of baseline clearance curves. These curves describe axisymmetric clearances between rotating and stationary seal components as functions of time for the flight acceptance profne.’ Inertia (g’s) and gyroscopic effects as a function of time are also characterized for the acceptance profile.
The computer simulation of the flight acceptance test incorporates the proper com- bination of nacelle loadings, engine thrust, inertia and gyroscopic effects, baseline clearances, and engine airseal/blade abradabihty factors. Exposure to thrust and maneuver loads results in deformation of propulsion system structural members and leads to relative motion between static and rotating components of flowpath seals. If the motions are larger than can be ac- commodated by the available clearances, rubs and wear will occur and hence a loss in per- formance. This simulation covers 16 conditions along the mission profile as shown in figure 2. A summary of relevant flight parameters is given in figure 3.
JT9D-7/747 Service Experience In the simulation of the 747 service experience, the previously defined flight acceptance test was refined to include only those maneuvers which are typical of a revenue flight. For the simulation, relative values of the loads remain unchanged, but the absolute values are increased to account for the probability of encountering larger loads during the life of the airplane.
The simulation of 747 service experience, as well as airplane acceptance, has been ac- complished at several discrete times (after 500, 1000, etc., flights) in the lifetime of the air- plane. In the lifetime of the airplane an ever-increasing chance of exposure to increasing load levels causes engine deterioration. Results from the model strongly reinforce the conclusion that flight induced seal rubs are the primary cause.of short-term (250 flights or less) engine performance deterioration and they are a significant contributor to the additional deterioration accumulated over the long term. As part of this simulation, it has been shown that certain engine components are particularly sensitive to certain types of loads. In figure 6, it can be seen that the fan stage is very sensitive to gyroscopic loadings and relatively insensitive to varying gravitational (g) loading levels. It can also be seen that the high pressure turbine (HPT) stages are relatively insensitive to gyros but sensitive to g loadings.
Effects of Engine Build Clearance Tolerances Ultimately an engine’s rate of deterioration is a function of its design. An engine built with open clearances deteriorates at a slower rate than an engine with tight clearances.
An engine with loose clearances has a high initial fuel consumption, but shows little deteriora- tion from load effects with time. An engine with tight clearances has a low initial fuel con- sumption, but exhibits a much greater initial short-term deterioration rate. ‘Modeling studies have shown that although an engine built with tight clearances deteriorates at a greater rate, it still exhibits better fuel consumption than a nominally built engine over its life cycle. This is true, in part, because rubs are local and the stage is still tighter on the average.
The effects of build clearance tolerances were investigated by using company standard engine build tolerance values in conjunction with the previously mentioned baseline clear- ance curves. Figure 7 depicts the results obtained from this investigation. From the figure, it is obvious that an engine built with open clearances shows less deterioration with time than an engine built ivith tight clearances. In this figure, however,-no attempt has been made to bias the curves due to initially higher fuel consumption or initially lower fuel consump- tion. Correlation of predicted % A TSFC ranges with short-term engine deterioration data is shown in figure 8. The predicted values bracket the engine data quite well.
The trend in predicted changes in % A TSFC of the JT9D-7 from loads versus time was found to be in good agreement with 747 fleet experience trends as shown in figure 9. Figure 10 suggests that beneficial effects of module reoperation or replacement (where only build clearances are restored) are only temporary and, as the engine re-enters service and encounters flight loadings, the airseals in the restored module (HPT or LPT) once again experience rub damage and deterioration within a short-term time frame.
Analysis of the results from the simulation reveals that calculated average clearance changes for the individual stages also correlate well with the JT9D-7 experience. Except for the fan stage (fig. 1 l), however, predicted and observed circumferential rub damage distributions do not compare satisfactorily. Figure 12 depicts lst-stage HPT outer airseal damage as a function of angular location for both a NASTRAF prediction and measured data on an engine that was torn down and measured m-house. A significant point to make with respect to this correlation is that the rub pattern exhibited by this engine is not typical of data measured on several other engines. In each case, the NASTRAN predicted values of average damage are .in agreement with measured engine data. Figure 13 reinforces the con- clusion that the trend in change of turbine tip clearance is increasing with time.
CONCLUSIONS An analytical procedure for assessingthe effects of flight loads on engine performance deterioration has been developed and applied to predict short- and intermediate-term changes in TSFC for the JT9D-7/747 installation. Good correlation between predicted and observed values for A TSFC serves to confirm the basic assumption that load-induced seal rubs have a significant effect on short-term performance deterioration.
The correlation between analytical prediction and measured clearance change is accept- able, but further refinements are desirable to explain specific rub pattern variations.
The analytical procedure provides for the detailed description of loads and deflections as they vary with time for arbitrary flight profiles and thereby permits the effects of individual loads to be isolated for evaluation. For the JT9D-7, studies of this kind indicate that airloads (inlet lift) are the dominant factor, maneuver loads (g’s, gyros) are of secondary importance, and thrust loads do not contribute to performance losses after the engine acceptance test.
Usefulness of this procedure as a diagnostics tool for understanding the major causes of early performance deterioration (rub-induced clearance changes) has been demonstrated.
At the same time, potential usefulness of the procedure as a design tool which can be used to minimize these effects in the future through use of load-sharing nacelles, active clearance control, optimum bearing placement, etc., has also been inferred. Better in-flight load predictions will be needed for future engine design clearance optimization for active clear- ance control systems.
CURRENTLY
ATSFC
7%
increase ,..
AVERAGE REPAIRED ENGINE TREND I I :; :. ,.,. . . . .
TIME
General characteristic of TSFC performance deterioration trends.
Figure l.- * FLIGHTCONDITION GROUND RUNS T.O.
APPROACH *SEE FIGURE 3 Airplane acceptance test flight profile.
Figure 2.- FLIGHT ALTITUDE AIRSPEED CONDlTlON DESCRIPTION (Ml MN 100 TAKEOFF ROLL 0.186 101 TAKEOFFROTATION 0 0.214 102 EARLYCLIMB 914 0.401 5,330 103 MID CLIMB 0.617 104 10,670 0.860 HIGHMACHCRUISE 105 LOW MACHCRUISE 10,670 0.770 106 MAXIMUMMACH 9,750 0.920 107 IN-FLIGHTSHUTDOWN 6,380 0.720 106 MAXIMUMSPEED 6,100 0.830 1.3VS.O" FLAPS 109 5,180 0.340 1.3VS. 1O"FLAPS 110 5,180 0.340 1.3VS.30" FLAPS 5,180 0.340 112 EARLYDESCENT 0.440 5,180 113 APPROACH,ZO"FLAPS 914 0.240 114 TOUCHDOWN 0 0.271 115 THRUSTREVERSE 0 Figure 3.- Airplane acceptance test flight profile parameters.
Figure 4.- JT9D-7/747 integrated NASTEUN structural model.
ym STRUT RCAC I _ THRUSTYOKE / -e -TURBINE 11.000 STATIC FREEDOMS I Figure 5.- JT9D-7/747 propulsion system substructures.
AVERAGE DAMAGE [MM] l/150 FLIGHTS l/150 FLIGHTS l/150 FLIGHTS THRUST, AIR AND THRUST, AIR, G THRUST ANDAIR STAGE LOADS G LOADS ANDGYRO LOADS 0.508 0.787 FAN 0.508 2LPC 0.025 0.051 0.102 3LPC 0.381 0.381 0.381 4LPC 0.584 0.584 0.584 3 LPT 0.025 0.025 0.025 0.076 0.076 4 LPT 5LPT 0.127 0.127 t-i?; 6 LPT 0.254 0.254 0:254 - - - ALL HPC 1 HPT 0.127 0.178 0.178 0.305 2HPT 0.356 0.356 1.5 1.61 1.65 Figure 6.- JT9D engine average damage.
0 JT9D-7/747 MIH. BUILD CLEARANCES •I JT9D-7/747 NOMINAL BUILDCLEARANCES 0 JT9D-7/747 MAX. BUILD CLEARANCES
% ATSFC
RELATIVE
TO
INDIVIDUAL
ENGINE
PERFORMANCE
WHEN NEW
FLIGHT CYCLES*
‘BASEDON A 3 HOUR FLIGHT Figure 7.- Predicted effect of build clearance on short and long term deterioration.
+6.0- o JT9D-7A +5-D- q JTSD-7AISPI D JT9D-7F o JT9D-20 +4.D- n P-695743 ASRCVD.
x NASTRAN PREDICTIONS
%I ATSFC
+3.0 -
RELATIVE
MINIMUMBUILD CLEARANCE
Y 1
Y
TONEW
”
+2.0 - x DATA AVERAGE Ia o n $ FiepL- +l.O - a A = " " A MAXhJM BUlLDCLEARhE I I I I 5; liD 0 150 200 250 300
FLIGHT CYCLES
Short term deterioration predictions bracket engine data.
Figure 8.- ONASTRAN PREDICTIONS (NOMINALENGINE BUILD) 0500 FLIGHTS* @lo00 FLfGHTS* 03000 FllGHTS* . PRE-REPAIR DATA •I POST REPAIRDATA I 9000 IOllOO 1l000 ‘EASED ON3 HR FLIGHT
FLIGHT HOURS
Figure 9.- Long term deterioration predictions are in good agreement with fleet experience.
0 lT9D-7/747 I I I I 1000 2000 3000 4ooo ‘EASEDONA 3 HOUR FLIGHT
FLIGHT CYCLES*
Figure lO.- Effects of turbine module replacement and subsequent deterioration.
r --
ANALYTICAL/EXPERIMENTALCORRELATlDN 2.54MY OUEASURED EN6lNEP695743 1.27 MY 1141 FLIGHTS1 PREDICTED 115DFllGHTS) 1150FllGHTS) I FRONT VIEW Figure ll.- Predicted fan rub patterns compared to measured data.
PREDICTED FOR 150 FLIGHTSVERSUSMEASURED DATA AT 141 FLIGHTS MEASURED ENGINE P695743 I AVERAGE CLEARANCE PREDICTED: + 0.178 MM MEASURED:+D.203MM PREDICTED t Figure 12.- Predicted first HPT outer airseal rub damage compared to measured data.
(TEARDOWN COMPARED TO ORIGINAL BUILD-1 AIRLINE) o MEASURED D LOCAL( PREDICTED) 0 AVERAGE WITHMINIMUMBUILDCLEARANCE (PREDICTEDI l.Or 0.75c
CHANGE IN
CLEARANCE cm-
MM
I
0. I I I
I
0 1 TIME!HRS 11k3) 4 5
Figure 13.- JTgD-7 high pressure turbine tip clearance change with time.