Full Flight Envelope Direct Thrust Measurement on a ... · FULL FLIGHT ENVELOPE DIRECT THRUST MEASUREMENT ON A SUPERSONIC AIRCRAFT Timothy R. Conners * and Robert L. Sims † NASA

NASA/TM-1998-206560

Full Flight Envelope Direct Thrust Measurement on a Supersonic Aircraft

Timothy R. Conners and Robert L. SimsDryden Flight Research CenterEdwards, California

July 1998

The NASA STI Program Office . . . in Profile

Since its founding, NASA has been dedicatedto the advancement of aeronautics and space science. The NASA Scientific and Technical Information (STI) Program Office plays a keypart in helping NASA maintain thisimportant role.

The NASA STI Program Office is operated byLangley Research Center, the lead center forNASA’s scientific and technical information.The NASA STI Program Office provides access to the NASA STI Database, the largest collectionof aeronautical and space science STI in theworld. The Program Office is also NASA’s institutional mechanism for disseminating theresults of its research and development activities. These results are published by NASA in theNASA STI Report Series, which includes the following report types:

• TECHNICAL PUBLICATION. Reports of completed research or a major significantphase of research that present the results of NASA programs and include extensive dataor theoretical analysis. Includes compilations of significant scientific and technical data and information deemed to be of continuing reference value. NASA’s counterpart of peer-reviewed formal professional papers but has less stringent limitations on manuscriptlength and extent of graphic presentations.

• TECHNICAL MEMORANDUM. Scientificand technical findings that are preliminary orof specialized interest, e.g., quick releasereports, working papers, and bibliographiesthat contain minimal annotation. Does notcontain extensive analysis.

• CONTRACTOR REPORT. Scientific and technical findings by NASA-sponsored contractors and grantees.

• CONFERENCE PUBLICATION. Collected papers from scientific andtechnical conferences, symposia, seminars,or other meetings sponsored or cosponsoredby NASA.

• SPECIAL PUBLICATION. Scientific,technical, or historical information fromNASA programs, projects, and mission,often concerned with subjects havingsubstantial public interest.

• TECHNICAL TRANSLATION. English- language translations of foreign scientific and technical material pertinent toNASA’s mission.

Specialized services that complement the STIProgram Office’s diverse offerings include creating custom thesauri, building customizeddatabases, organizing and publishing researchresults . . . even providing videos.

For more information about the NASA STIProgram Office, see the following:

• Access the NASA STI Program Home Pageat

http://www.sti.nasa.gov

• E-mail your question via the Internet to [email protected]

• Fax your question to the NASA Access HelpDesk at (301) 621-0134

• Telephone the NASA Access Help Desk at(301) 621-0390

• Write to:NASA Access Help DeskNASA Center for AeroSpace Information7121 Standard DriveHanover, MD 21076-1320

NASA/TM-1998-206560


Timothy R. Conners and Robert L. SimsDryden Flight Research CenterEdwards, California

July 1998

National Aeronautics andSpace Administration

Dryden Flight Research CenterEdwards, California 93523-0273

NOTICE

Use of trade names or names of manufacturers in this document does not constitute an official endorsementof such products or manufacturers, either expressed or implied, by the National Aeronautics andSpace Administration.

Available from the following:

NASA Center for AeroSpace Information (CASI) National Technical Information Service (NTIS)7121 Standard Drive 5285 Port Royal RoadHanover, MD 21076-1320 Springfield, VA 22161-2171(301) 621-0390 (703) 487-4650

FULL FLIGHT ENVELOPE DIRECT THRUST MEASUREMENTON A SUPERSONIC AIRCRAFT

Timothy R. Conners* and Robert L. Sims†

NASA Dryden Flight Research CenterEdwards, California

Abstract

Direct thrust measurement using strain gages offersadvantages over analytically-based thrust calculationmethods. For flight test applications, the directmeasurement method typically uses a simpler sensorarrangement and minimal data processing compared toanalytical techniques, which normally require costlyengine modeling and multisensor arrangementsthroughout the engine. Conversely, direct thrustmeasurement has historically produced less thandesirable accuracy because of difficulty in mounting andcalibrating the strain gages and the inability to accountfor secondary forces that influence the thrust reading atthe engine mounts. Consequently, the strain-gagetechnique has normally been used for simple enginearrangements and primarily in the subsonic speed range.This paper presents the results of a strain gage–baseddirect thrust-measurement technique developed by theNASA Dryden Flight Research Center and successfullyapplied to the full flight envelope of an F-15 aircraftpowered by two F100-PW-229 turbofan engines.Measurements have been obtained at quasi-steady-stateoperating conditions at maximum nonaugmented andmaximum augmented power throughout the altituderange of the vehicle and to a maximum speed ofMach 2.0, and are compared against results from twoanalytically-based thrust calculation methods. Thestrain-gage installation and calibration processes are alsodescribed.

Nomenclature

A cross-sectional area, in2

ACTIVE Advanced Control Technology for Integrated Vehicles

1American Institute of Aero

*Aerospace Engineer, Senior Member AIAA.†Aerospace Engineer.Copyright 1998 by the American Institute of Aeronautics and

Astronautics, Inc. No copyright is asserted in the United States underTitle 17, U.S. Code. The U.S. Government has a royalty-free license toexercise all rights under the copyright claimed herein for Governmen-tal purposes. All other rights are reserved by the copyright owner.

F force, lbf

g gravitation acceleration constant, 32.2 ft/sec2

IDEEC improved digital electronic engine controller

M Mach number

P static pressure, lbf/in2 absolute

total pressure, lbf/in2 absolute

S/MTD Short Takeoff and Landing/Maneuver Technology Demonstrator

total temperature, °F

USAF United States Air Force

V velocity, ft/sec

WACC station-corrected engine mass flow, lbm/sec

WAT true engine mass flow, lbm/sec

standard deviation

Engine Stations

0 free stream (ambient)

2 engine-inlet plane

Introduction

For flight-testing applications, direct thrustmeasurement using strain gages offers advantages overtraditional, model-based analytical thrust calculationmethods.1 Depending on the objectives of the flight testprogram and resources available to the test facility, theseadvantages may permit in-flight thrust measurement thatwould not be feasible otherwise.

Depending upon the application, the directthrust-measurement method can be less complex andcostly to implement compared to model-basedtechniques. The strain-gage sensor arrangement istypically less cumbersome and costly to procure, install,and calibrate than the multisensor package required tocollect the input data for traditional analytically-based

PT

T T

σ

nautics and Astronautics

thrust models, particularly if a wind-tunnel calibration ofthe analytical model to the specific test engine wouldotherwise be required prior to test programcommencement. The computer models are an additionalrequirement to which the direct measurement method isnot subject. These models can be very costly for the enduser to procure and maintain, if the models even exist. Ifthe models do not exist and therefore need to bedeveloped for the engine in question, then the cost forprocurement can easily be prohibitive.

Although the typical in-flight thrust model has limitedself-tuning capability, if the engine has strayed far froman average baseline state (for instance, because ofsignificant deterioration or damage), then the calculationaccuracy will suffer. Because the strain gage–basedtechnique measures thrust directly, the technique is notsubject to this type of error.

An in-flight thrust computer program is normallycapable of modeling only steady-state or quasi-steady-state engine operation. A computer programtypically assumes thermal equilibrium andstoichiometric fuel-to-air ratios that do not account forengine acceleration or deceleration schedules, and canbe further limited by the responsiveness of the inputparameters.2 The strain-gage technique, however, is nothindered by modeling and input measurement ratelimitation to the same extent. Because of the inherenthigh dynamic response of strain gages, the directmeasurement technique is superior to model-basedtechniques when dynamic thrust measurement is arequirement. Not having the computational burden of theanalytical methods, the direct thrust-measurementtechnique is better suited for real-time monitoringapplications as well.

Although the direct thrust-measurement technique hasseveral advantages, careful attention must be paid to thedesign and calibration of the system in order to reducemeasurement error and calibration drift. Misconceptionsand misapplications have hampered the widespread useof the direct thrust-measurement technique, and themethod has historically produced less than desirableaccuracy. Without careful installation and calibration ofthe strain gages, the accuracy of the direct measurementtechnique will be inferior to the model-based method.

Proper design of a direct thrust-measurement systemis required to ensure that secondary load paths arenegligible or can be accounted. Such load paths canresult from external engine connections and interfaces,inlet seals, inlet pressure forces, engine-body pressure

force differential, and nozzle drag, all of which combineto increase the measurement uncertainty. Propersecondary load-path bookkeeping has been a primaryproblem of past attempts at direct thrust measurement.As a result, the strain-gage technique has normally beenused for simple engine arrangements such as podinstallations under the wing. However, as this reportshows, the technique can be properly applied tocomplex, buried-engine arrangements, provided care isused to ensure that nonnegligible secondary forces areunderstood and included in the analysis.

For installation environments subject to significantthermal cycling (such as with high-speed aircraft usingaugmented engines), the strain-gage measurementmust be temperature-compensated, an importantconsideration for the calibration process. When a testprogram commences, depending on the programobjectives and duration, in-flight tare readings againstsome reference (such as output from a simple engineperformance specification model) may be prudent tohighlight a need for sensor recalibration.

Published reports of successful applications of thestrain gage–based direct thrust-measurement techniqueare not numerous. Two reported examples are a pod-mounted J85 (General Electric, Lynn, Massachusetts)engine installation tested on an F-106 airplane at speedsfrom Mach 0.6 to Mach 1.3 throughout the power rangeof the engine,3 and an F-14A TF30 (Pratt & Whitney,West Palm Beach, Florida) engine applicationdemonstrated at speeds from Mach 0.4 to Mach 1.6 attwo power settings, including maximum augmentation.4

The analysis discussed in this paper was performedusing the NASA Advanced Control Technology forIntegrated Vehicles (ACTIVE) F-15 aircraft. TheACTIVE test program5 is a joint effort between NASADryden Flight Research Center (Edwards, California);the United States Air Force (USAF) Materiel Command(Dayton, Ohio); Pratt & Whitney, a division of UnitedTechnologies (West Palm Beach, Florida); and TheBoeing Company (formerly McDonnell DouglasAerospace) (St. Louis, Missouri). A major goal of theACTIVE program is to demonstrate the benefits of aproduction-like thrust-vectoring system applied to thefull flight envelope of a supersonic vehicle.6 Thesebenefits include enhanced aircraft performance,maneuverability, and controllability.

Engine-mount strain gages were originally installedand calibrated for the primary purpose of monitoring netstructural loads, but the extraction of usable thrust- and

2American Institute of Aeronautics and Astronautics

vector-force data was also deemed possible. The straingage–based nozzle-vectoring force-measurement systemhas produced reasonably precise results and has becomean element crucial to the success of the nozzle-vectoringprogram.

This paper presents the results of a strain gage–baseddirect thrust-measurement technique developed byNASA Dryden and successfully applied to the full flightenvelope of the F-15 ACTIVE aircraft, which is poweredby two F100-PW-229 (Pratt & Whitney) turbofanengines. More than 5200 time cuts of data wereprocessed for quasi-steady-state operating conditions toproduce the results in this report, which represent aMach range from 0.0 to 2.0 and altitudes from near sealevel to higher than 45,000 ft. Measurement-based gross-thrust values were obtained at maximum nonaugmentedand maximum augmented power and are compared tothe results from two analytically-based thrust calculationmethods. The direct thrust-measurement technique and adescription of the significant secondary force termsrequired to compute gross thrust are presented. Inaddition, the strain-gage installation and calibrationprocesses for the axial-thrust measurement are describedin detail, as are the important lessons learned. Use oftrade names or names of manufacturers in this documentdoes not constitute an official endorsement of suchproducts or manufacturers, either expressed or implied,by the National Aeronautics and Space Administration.

Aircraft Description

The ACTIVE test aircraft (fig. 1) is a highly modified,preproduction, two-seat F-15B airplane on loan toNASA Dryden from the USAF. The aircraft waspreviously used for the USAF F-15 Short Takeoff andLanding/Maneuver Technology Demonstrator (S/MTD)program.

Airframe

Two large canards mounted on the upper inlet area,forward of the wing, comprise the dominant externalfeature of the vehicle differing from a standard F-15airplane. To accommodate the off-axis force producedby the vectoring nozzles, certain engine-mount andsupporting structure has been strengthened. The skincontour and structure in the aft fuselage area were alsomodified to accommodate the vectoring actuationsystem and to provide clearance for full nozzle-vectoringmovement. The modified engine-mount structure isdescribed in detail in the “Direct Thrust-MeasurementSystem Description” section.

The aircraft is controlled using a quadruply redundant,digital fly-by-wire flight control system. All mechanicallinkages between the control stick, rudder pedals, andcontrol surfaces have been removed from the aircraft.The throttles digitally control the engines through the


EC96 43485-3

Figure 1. The NASA F-15 ACTIVE test vehicle.

flight controller, and no mechanical linkages existbetween the throttles and the engines. Ten separatecomputers using 1553 multiplex data buses form thehighly integrated flight and propulsion control system.These computers are used for digital flight, engine,nozzle, and inlet control; cockpit informationmanagement; pilot-vehicle interface; and auxiliaryresearch computation capability.

The flight envelope of the F-15 ACTIVE vehicle isreduced compared to the production F-15 airplane. Thecomposite material used to manufacture the canardslimits the maximum Mach number to 2.0. The flightceiling is an altitude of 50,000 ft because the digitalflight control system has not been cleared to higheraltitudes.

Propulsion System

The ACTIVE propulsion system consists of twoNASA-owned F100-PW-229 engines, each of which isequipped with a new-generation axisymmetric thrust-vectoring nozzle. An engine-mounted improved digitalelectronic engine controller (IDEEC) and avionicsbay–mounted nozzle controller provide closed-loopcontrol of each respective component.

F100-PW-229 Engine

The F100-PW-229 engine is an augmented 29,000-lbfthrust-class motor that features a three-stage fan andten-stage compressor, each driven by a two-stageturbine. An eleven-segment fuel delivery system is usedwithin the augmentor. The full-authority IDEECprovides the pilot with unrestricted throttle movementthroughout the flight envelope while maintaining engineoperation within limits. A hydromechanical secondaryengine control provides “get-home” capability if theIDEEC becomes unable to adequately control theengine. The addition of the vectoring system does notcause any loss of engine functionality, operationalcapability, or modification of fault accommodation.

Small modifications to the main body of the enginewere required to support the addition of the vectoringnozzle. These modifications included strengthening theaugmentor duct and front and rear fan ducts toaccommodate the off-axial loads generated duringnozzle vectoring.

Nozzle-Vectoring System

Each axisymmetric vectoring nozzle provides amaximum of 20° of mechanical vector angle in anycircumferential direction. Independent control of thenozzle exit area allows the nozzle exit–to–throat area

ratio to be optimized for performance, unlike thestandard F100 divergent nozzle section, which uses amechanical linkage system that tracks throat area andpressure loads but is not capable of independent exit areamodulation. Torsional freedom built into the divergentseal design permits the nozzle to maintain a tightgas-path seal even while vectored. The tight seal allowsthrust coefficient performance nearly identical to astandard F100 nozzle while preventing vectoring flowinstabilities. The majority of hardware forward of thethroat is common to the standard nozzle. The convergentsection maintains its mechanical and controlindependence from the divergent section despite theaddition of the divergent section–vectoring system.

Instrumentation System

The aircraft has been extensively modified with aflight test instrumentation system for recording digitaland analog sensor data. Approximately 3400performance, flying qualities, structural loads, data bus,and propulsion parameters are recorded onboard theaircraft and are simultaneously downlinked to the NASADryden mission control center.

A description of the strain-gage thrust sensorarrangement, calibration, and data processing procedureis provided in the “Direct Thrust Measurement SystemDescription” section. A discussion of the inputparameters used in the analytical thrust calculationmodels as well as for converting the direct engine-mountforce measurements into gross thrust can be found in the“In-Flight Thrust Analysis” section.

Direct Thrust-Measurement System Description

Figure 2 shows the basic engine-mount reactions.These longitudinal, vertical, and lateral reactionstransfer the combined thrust, inertia, gyroscopic, andany nozzle airload forces into the fuselage structure. In

Inboard main-mountthrust and vertical load

Foward-linkvertical

load

Side-linklateral load

Outboardmain-mountthrust and

vertical load

980308

Figure 2. Left engine-mount reactions.


the case of the ACTIVE program, pitch and yaw thrustvectoring superimpose additional forces on the mounts.

The mount arrangement is fairly typical for fighteraircraft designs with engines buried in the fuselage. Thelayout and hardware are designed such that the reactionsare essentially statically determinate. For the frontmount, a dog-bone hanger link with monoball bearingsensures that only vertical loads can be reacted andsimultaneously accommodates significant enginethermal expansion in the longitudinal direction.

Figure 3 shows the inboard main-mount and side-linkarrangement. An adapter pin on each side of the engine isclamped in the main mounts, which are attached to amajor bulkhead and support table. During installation,the engine is slid on rails into the bay, and the pin isinserted into the main bearing and then rolled into themount. When seated, the pin is secured by theswing-down clamping jaw. To accommodate thermalexpansion in the radial direction, the engine monoballbearings ride on the end of the pins that are coated with ahigh-temperature dry-film lubrication to minimizefriction.

The pins react all of the thrust load and approximately75 percent of the vertical inertia forces. Nozzle pitchvector generates additional vertical forces, and all of themoment produced by yaw vector is reacted through

differential thrust on the pins. The net force applied tothe pin can thus be in any direction, depending on enginethrust, aircraft maneuvering kinematics, and vectoring.Ideally, all of the lateral inertia and yaw vector force isreacted by the side link attached to the lower part of theengine frame. This link keeps the engine approximatelycentered on the pins and prevents excessive sliding. Inreality, some lateral load can be reacted by the pinsthrough friction or binding.

For the F-15 ACTIVE program, certain engine-mountand supporting structure was strengthened to handle theadditional vectoring forces. Measurement of theengine-mount forces thus originated as a structural loadsobjective. The most viable approach for monitoring theengine-mount loads was thought to be to directlyintercept and measure the six load reactions identified infigure 2. Direct and reasonably accurate measurement ofthese components would satisfy safety-of-flightobjectives during envelope expansion with the vectoringnozzles. At the same time, the realization was made thatwith high measurement accuracy and a good inertia andgyroscopic model, extracting thrust and vector forces tosupport research objectives, including correlation toconventional engine thrust models, would be possible.To extract thrust and vector forces, the assumption wasmade that some secondary load paths through enginesystem interfaces are not significant. Although desirable,instrumenting and directly measuring these secondaryload paths is not generally viable. Potential sources forsystem interfaces that rob load from the engine mountswill be discussed in a later section.

As discussed in the introduction, few examples ofprevious successful engine-mount load measurementexist upon which to base or guide the ACTIVE programapproach. Engine-mount load measurement has not beenattempted in any past programs at NASA Dryden.During the preceding S/MTD program,7 the forward linkwas instrumented and 12 strain gages were installed onvarious parts of the engine-bay and main-mount tablestructure. The primary objective was safety-of-flightmonitoring during pitch vectoring. Because of scheduleand budget constraints, a ground calibration was notperformed. Instead, calculated applied loads andflight-measured strains for nonvectoring maneuverswere used in an attempt to derive thrust and vertical loadequations for the main mounts. Usable equations thatcould support safety-of-flight monitoring or researchobjectives were not able to be derived.

Even with an extensive ground calibration, thisapproach invites several problems. Foremost is that theengine-bay strain gages could respond not only to the

Major bulkhead

Reference engine structure

Reference engine structure

Main bearing

Adapter pin

Main mount with clamping jaw

Support table

Side link

980309

Figure 3. Inboard main-mount and side-link structure.


main-mount loads, but also to other fuselage strains thathave no relationship to engine-induced loads (includingfuselage inertia loads and even horizontal and verticalstabilizer loads). Separating out these combined effectsso that engine induced loads could be calculated wouldbe nearly impossible. The engine-bay strain gages werealso susceptible to thermal contamination. To generatethe necessary thermal correction data, an engine-bayground heating test could be conducted to simulate thethermal operating environment. For schedule, cost, andpractical considerations, this type of test is generally notan option and was not performed for the S/MTDprogram.

The following sections discuss the strain-gageinstallation and calibration process, status, and lessonslearned that have evolved in support of the ACTIVEprogram. The approach is founded on the direct andseparate measurement of the thrust and vertical forcespassing through the main-mount pins. As this paper isonly concerned with main-mount thrust measurementand its conversion to gross thrust, details related to thelink loads or vectoring forces will not be addressed.

Strain-Gage Installation

In order to install and calibrate strain gages on themain-mount pins, a fundamental starting requirement isto fix the angle at which the pin is clamped in the mount.

For this purpose, a V-shaped alignment pointer wasinstalled in the mount end of each pin (fig. 4). Thispointer mates with a matching plate secured to the backside of the mount, which forces the pin to the requiredangular orientation during engine installation. Thepointer is a one-piece steel machining with a plugextension that is inserted into the hollow bore of the pinand secured with a machine-thread fastener adhesive.After a recent engine removal, inspection of the pinsindicated that the bond on one of the pins wascompromised. To preclude future problems, steel dowelswere installed through the pointer face into the thick wallof the pins. This modification (not shown in the photo)provides a secure mechanical attachment that preventsany possible rotation of the pointer relative to the pinduring engine installation or removal.

The line through the pointer apex is coincident withthe thrust axis, and the vertical load axis is orthogonal.With these reference axes established, the strain-gageinstallation follows the layout defined in figure 5. Thecola bottle–shaped pin is D6-AC high-strength steel, isvery thick-walled, and tapers between the mount clampzone and the small diameter where the engine bearingrides. Four-active-arm shear bridges are installed in theopen area between these zones. Separate thrust andvertical load bridges are oriented to sense the shear loadalong their respective axes. A prime and spare bridge is


EC94 42762-3

Figure 4. Main mount, adapter pin, and alignment pointer.

7

Am

erican Institute of Aeronautics and A

stronautics

Bridge type: four-active-arm shear

( ) in leg callout indicates back side surface. Back side gages inverted to form X pattern

Strain gages: Micro-Measurements WK-06-125TH-350 (Orthogonal two-element torque gage)

Thermocouple: RdF 20112 (type K chromel)

Terminal strips and interbridge wiring not shown

Section areas: 1.86 in2 prime station

2.06 in2 spare station

All dimensions in inches

Wear zone

AA

Main-mount clamp zone

Top view

Forward

Rear view

Exposed area

1.25

2.96

1.952.28

Spare PrimeVertical-load bridges

Thrust-load bridges

Section A-Aat prime bridge station

Thermocouple

Wire bundle towall-mounted

disconnect plug

Forward Leg 3 and 4

Leg1 and 2

Leg1 and 2

Leg3 and 4

Thrust-load bridges

Vertical-load

bridges

2.00

Alignmentpointer

Centerline

1.750.90

Leg 1(3)

Leg 2(4)

2.70

1.75

Calibration loads

0.600.80

Taper

Inboard

2.5˚ maximum rotation

Thermocouple

Spare Prime

0.9345

bearing

Engine1.7500

980310

Figure 5. Strain-gage and thermocouple instrumentation on outboard main-mount pin.

provided for each axis, totalling four for each pin. Thegages are Micro-Measurements (Measurements GroupInc., Raleigh, North Carolina) WK-06 series forincreased toughness and temperature capability in thepotentially harsh environment, temperature-matched forsteel. The specific TH model gage used has a gridconfiguration with two orthogonal legs precision-aligned in a V shape. The other two legs are installed onthe back side of the pin, but are rotated 180° to form asuperimposed X shape. This bridge configuration isconsidered the optimum for measuring shear and shouldbe inherently insensitive to any lateral or torque loads onthe pin.

To accommodate potential corrections for thermalerrors in the bridge output, a type K thermocouple (RdFCorporation, Hudson, New Hampshire) is installed inthe space between and close to the gages. Figure 6 showsa closeup photograph of the completed installation priorto application of a heavy protective coating of sealingcompound. Interbridge wiring is routed to solder tabslocated in the areas between the orthogonal axes, andfrom there, a single wire bundle with built-in strain reliefis routed to a wall-mounted disconnect plug. With the16 individual strain-gage legs, interbridge wiring, andthermocouple, the pin circumference is congested butmanageable.

Prior to committing to this approach, potentialproblems and disadvantages with both the installationand obtaining a usable calibration were identified asdiscussed below:

• The usable space between the edge of the mountand the engine bearing is only approximately 1 in.The wear zones evident on the original pins fromthe previous program were surveyed to determinethe maximum range of lateral bearing movement onthe pin. The survey results dictated the safe area forstrain-gage installation (figure 5 shows the worst-case bearing position and angle). New pins wereobtained for the ACTIVE program so that the pinsstarted with no wear.

• The close proximity of the gages to both the bearingand mount clamping zones generated concernsregarding localized loading and reaction effectsthat could produce a strain field that was notwell-behaved. Such a strain field could lead toerratic, nonlinear response; hysteresis; and otherundesirable effects. The strain gages areapproximately centered between these zones, andan approximately 0.5-in. gap exists from each straingage to the edge of the nearest gage grid.

• The shear bridges had to be installed on a taperingcross section for this application, which a


EC94 42893-4

Figure 6. Installed strain-gage instrumentation.

sensitivity analysis indicated should not alonepresent any problems.

• To handle the design temperature of 350 °F and befatigue resistant, the pin has a very large marginrelative to normal operating loads. This largemargin produces fairly low shear bridge response,predicted to be approximately 10–12 mV at loadlimit. For typical maximum thrust operating levels,the output would only be approximately 5–6 mV.The low response could lead to noise problems andincrease susceptibility to thermal contamination.

• Shear bridges can show some sensitivity tobending moment. The thermal expansion andlateral sliding of the engine produces a range ofbearing positions that vary approximately 0.2 in.and are unpredictable. For the prime strain-gagestation, this thermal expansion and lateral slidingproduces a bending arm variation of approximately±10 percent.

• One of the biggest concerns was the expectation forcross-axis loading effects. When a pure thrust loadis applied, the vertical axis gages ideally producelittle or no response. Conversely, the thrust-orientedgages ideally respond only to the component alongthe thrust axis. In actuality, the magnitude andbehavior of these cross-axis effects would depend inpart on the precision alignment and placement ofthe strain-gage installation. A good deal of care wastaken in marking the axes and positioning the gagesto minimize these effects.

• A level of concern existed regarding the potentialfor damage to the gages, wiring bundle, and pointer

during engine installation, removal, and normaloperation. The alignment pointer, in particular,requires some extra care and effort during engineinstallation, as the pointer must slide, withoutrolling, during the final seating in the mount. After95 flights and several engine removals andinstallations, concerns in this area have diminished.The only problem to date has been that related to thepointer discussed earlier.

• Application of the calibration loads does require acustom fixture, as described in the next section.

The potential issues listed above were cumulativelyimposing; however, the realization was made that if goodpin calibrations could be obtained, then nonengine-generated forces could not contaminate the strain-gageoutputs. Some additional advantages inherent to thisoverall approach result from being able to remove theinstrumented pins from the aircraft. One advantage is anoff-aircraft calibration loading should be simpler,quicker, less costly, and less hazardous than anequivalent on-aircraft test. Another advantage is thatwith their relatively small size, the pins can beoven-tested to establish thermal corrections to the gageoutputs, as discussed in a following section. Also, thepins could be installed in comparable mounts to conducta test program on a different aircraft. These options donot exist if strain gages are installed on engine-bay ormount-supporting structure.

Calibration Process

The calibration process has evolved through fourdistinct steps summarized in table 1. The philosophy is


Table 1. Summary of calibration steps.

StepTest

descriptionPrimary

objectivesKey

elementsTest

complexityTest

time span

1 Off-aircraft fixture loading

Derive strain-to-load calibration equations

Shear loading fixture

Instrumented pins in main mounts

Simple 1 day for each pin

2 On-aircraftlaboratory test

“Cold loads”

Verify accuracy of strain-gage equations

Establish baseline engine deflections

Dummy engines with calibrated pins

Simulate full thrust and vectoring

Moderate to complex

~3 weeks(10 test days)

3 Off-aircraft heating test

Derive strain-gage thermal corrections

Instrumented pins in laboratory oven

Simple 2 days

4 On-aircraft combined systems test

“Hot loads”

Validate operation of all integrated systems

Correlate thrust models to measurements

Aircraft restrained on thrust stand

Full thrust and vectoring

Moderate to complex

~2 weeks(5 test days)

common to many flight tests: start with a relativelysimple basic step, and with each new step, addcomplexity, understand and control errors or problems,and incorporate lessons learned from preceding steps.

Off-Aircraft Fixture Loading

The first step established the basic calibrationequations to convert the measured strain-gage outputs(measured in mV) to the desired force quantities(measured in lbf). For this purpose, a shear-loadingfixture (fig. 7) was constructed. A main mount from theairplane was used to hold the pins, and loads wereapplied through a hydraulic jack, a load cell, and“lollypop” hardware containing a bearing obtained fromthe engine manufacturer. All of this hardware wasmounted on a large I beam. The calibration processrequired loading the pin at various angles to simulate thecombinations of thrust and vertical load expected inflight. To simplify the fixture setup, the followingprocedure was devised: the pointer mating plate wasremoved, and instead, a custom protractor templatemade from a viewgraph transparency was taped to theback side of the mount face. So, rather than rotating thewhole jack structure to apply loads at varying angles, thepin or pointer was simply rotated to the desired angleand reclamped in the mount. The mount itself wasalways loaded in the thrust direction.

Each pin was loaded at six different angles thatcovered the normal operating portions of themain-mount strength envelope. These loading conditions(fig. 8) are superimposed on the strength envelope

EC94 42896-5

Figure 7. Pin calibration shear-loading fixture.


Figure 8. Pin calibration loading conditions.

Down,+ g

Verticalload,lbf

Thrust load, lbf Forward– 24 – 16 – 8 0

90° 67.5°

45°

22.5°

0°

– 45°

8 16 24 x 103

24 x 103

Main-mount strength envelope

16

8

0

– 8

– 16

980311

defined in terms of vertical load limit as a function ofthrust load. Positive vertical load is a down load on themount, corresponding to a positive aircraft load factor.Initial loadings ranged from 27,000 lbf forward to–20,000 lbf aft (approximately 115 percent of the flightoperating limits) using a slow ramping profile. Becausethe gage outputs were fairly linear, the profile wasreduced for most runs to the range from 24,000 to–12,000 lbf shown in the figure. The first quadrant,representing forward thrust and positive g, is well-covered, as this quadrant is the normal operating regionwhere calibration accuracy is most important. Negative gor aft thrust conditions were given less emphasis. Ateach angle, the pin was loaded at the most inboard andoutboard bearing positions to assess the bending andlocalized loading effects discussed previously. Figure 5shows these positions in the lower rear view.

The calibration data for the 8 bridges (prime and sparefor thrust and vertical for each of the two pins) wereevaluated for each of the 12 loading cases (6 angles for2 lateral positions each). Figure 9 shows an example ofthe bridge output for the prime thrust bridge on theoutboard pin (measured in mV) plotted as a function ofapplied load for all 12 loading cases. This example istypical of all the prime bridges. Looking first at theresponse for the primary axis loading (0°, thrust axiswith no vertical component), the output is quite linearover the positive load range. The difference between the

two lateral loading positions (open and solid symbols) isminor, less than 1 percent in slope. A least-squares slope,based on the positive data, is extended through thenegative region to show that the response is slightlygreater for aft thrust- axis loading.

Bridge output for loading in the vertical axis at 90°would ideally be 0 mV. The measured response isapproximately 3 percent of that for the thrust axis. Forloading angles representing various combinations ofthrust and vertical load, the response is also greater thanwould be expected from multiplying the 0° slope by thecosine of the angle, as indicated by the lines labeled onthe right border. For example, data for the 45° and 68°loading angles are 6 and 12 percent greater, respectively,than would be expected. Repeat loading cycles showedthe data to be very repeatable and well-behaved, butthese cross-axis effects indicate that a single-bridgethrust equation with high accuracy over any loadingangle is not feasible.

Data from all 12 loading conditions were processedthrough a regression analysis to evaluate two-bridgeequations that combine a thrust bridge with a verticalbridge. Figure 10 shows the resulting equation accuracyfor the prime bridges on the outboard pin. The equation,listed at the top of the figure, uses the vertical bridgeoutput with a relatively small coefficient to correct forthe cross-axis effects. To obtain the best accuracy in the


Figure 9. Typical bridge output for shear fixture loadings.

Bridgeoutput,

mV

Cos 21°* slope

Cos – 44°* slope

Cos 45°* slope

Cos 68°* slope

Cos 90°* slope

Applied load, lbf

– 10 – 5 0 5 10 15 25 x 10320

21°

45°

– 44°

68°

Verticalaxis90°

– 15– 8

– 4

0

4

8

12

16

980312

Thrust axis0°

Outboard pinPrime thrust bridge10 V excitation

Symbols

OpenSolid

Position

OutboardInboard

Figure 10. Thrust equation accuracy for shear fixture loadings.

Strain-gagemount thrust,

lbf

Applied mount thrust, lbf– 10 – 5 0 5 10 15 20 25 x103

25 x 103

– 10

– 5

0

5

10

20

15

980313

Outboard pinPrime two-bridge equationAll 12 loading runs6 angles * 2 positions

Correlation coefficient (R2) = 0.99985Standard error = 81 lbf

Mount thrust (lbf) = 1625* Thrust bridge (mV) + 65.7* Vertical bridge (mV)

positive thrust direction, aft thrust data were excludedfrom the regression analysis; then the increased error inthis region was accepted. The regression equation does aremarkable job for any loading angle or lateral position.The correlation coefficient and standard error for this fitare shown on the figure. The 3- error divided by themaximum load is 1 percent.

Figure 11 shows the discrete error for each data pointplotted. The final equation coefficients were manuallytuned to minimize and balance the errors for the positivethrust and positive g loading conditions. This tuningproduced a maximum error band of ±200 lbf for theseconditions. For the low-angle loadings most pertinent tothis paper, figure 11(a) shows that the maximumhysteresis spread is approximately ±50 lbf, and thelargest difference between the lateral loading positions is200 lbf. Thrust errors for the more acute angles,representing significant g loadings (fig. 11(b)), are alsominimal except for the less important –44° condition.Prime equation accuracy for the inboard pin thrust wascomparable to that discussed here. Accuracies for theprime vertical equations or those using the spare bridgeswere not quite as good, with 3- errors varying from2 to 4 percent. The overall assessment, however, is thatthe calibration equations resulted in as near optimum aset of shear sensors as could be hoped for, especiallyconsidering all the potential issues discussed earlier.

One negative aspect that was not anticipated, however,was uncovered in the shear fixture loadings. Theclamping jaw exhibited what could be consideredexcessive flexibility. This flexibility allowed the pin toyaw, causing substantial gapping between the pin andthe bore face. The amount of gapping was dependent onthe magnitude and direction of the applied load. Adeflection transducer was added to the test setup tomeasure the fore and aft yawing of the pin because thisaddition produces an additional source for enginemovement relative to the bay. At maximum thrust, theengine is expected to move forward 0.05 in. because ofthe pin-and-mount assembly flexibility in addition toairframe flexibility and thermal expansion effects.

This gapping also generated additional concern thatthe cross-axis effects mentioned above would not beproperly simulated because the pin was rotated insteadof the jack. The clamping flexibility, and thus thegapping characteristics, could be different for eachmount and loading angle. Cross-axis bridge output was2–4 percent of the primary axis output for the primebridges, but was 10–15 percent for the spares, which arecloser to the mount, indicating that the clamping andgapping characteristics can affect the bridge response.The simplification in the test setup thus produced somecompromise in the fidelity of the simulated loads thatcould influence the validity of the excellent calibration

σ

σ


13

American Institute of Aeronautics and Astronautics

(a) Low angle loadings.

(b) High angle loadings.

Figure 11. Thrust equation error for shear fixture loadings.

Strain-gagemount-thrust

error, lbf

Applied mount thrust, lbf– 10 0 5– 5 10 15 20 25 x 103

– 400

– 300

– 200

– 100

0

100

200

300

400

980314


21°

21°

21°

0° 0°

0°

Symbols

OpenSolid

Position

OutboardInboard


error, lbf

Applied mount thrust, lbf– 10 0 5– 5 10 15

– 44°

– 44°

45°

45°

68°

68°

90°

20 25 x 103– 400

– 300

– 200

– 100

0

100

200

300

400

980315


Symbols

OpenSolid

Position

OutboardInboard

equations just discussed. The ground test schedule didnot allow consideration for redesigning the fixture setupand repeating the loadings. Instead, the on-aircraftlaboratory test, described next, would be used to checkthe calibration validity.

On-Aircraft Laboratory Test (“Cold Loads”)

In early planning of the overall ground test program,the necessity and importance of conducting anon-aircraft laboratory test was debated, especially inlight of the effort and time required and that a combinedsystems test with engines installed and operating (the“hot loads” test) was already planned. In reality, theon-aircraft laboratory test without engines installed (the“cold loads” test) supported the multiple objectiveslisted below:

1. Validate the aircraft restraint system to be used inthe “hot loads” test.

2. Satisfy proof test requirements for the strengthenedaircraft structure.

3. Establish baseline engine deflections relative to thebay.

4. Verify side-link and main-mount frictioncharacteristics.

5. Establish the accuracy of baseline strain-gageequations.

In lieu of “cold loads” testing, objectives 1, 2, and 3could be accomplished during “hot loads” testing. The“hot loads” test could be used to validate the straingage–derived thrust equation, but could not addresselevated g maneuvering or vector-force calculations.With the concerns expressed in the previous sectionregarding the main-mount flexibility, the desirability ofconducting the “cold loads” test to support objectives 3and 5 was considerably elevated.

Figure 12 shows a rear view of the test setup in theNASA Dryden Flight Loads Laboratory. Two“sewer-pipe” engine fixtures were installed using thecalibrated pins and links. These dummy, or “cold,”engines were available from early F-15 structural testingand modified to take vector loadings. The enginesrealistically simulated the engine mount points but didnot simulate any of the engine system interfaces. Ballastwas added to match the dead weight on the pins for theactual engines. The thrust loading jacks can be seenextending horizontally to the engine aft lug plates.Additional jacks for pitch and yaw vector loadings alsoattach to the same lug plate. Just forward of the mainmounts, a jack to apply simulated vertical g loadings wasattached at the engine center of gravity by cutting a smallhole in the bay door.

The airplane was restrained using the same hardwaredeveloped for the “hot loads” test. This hardware


Figure 12. “Cold loads” test setup in the flight loads laboratory.

980344

Thrust jack

Right dummy engine

Restraint truss

Yaw-vector lever arm

Yaw-vector jack

Pitch-vector jack

consisted of spare main landing gear, bottomed out andsecured in massive cradles; and the attachment of a largetruss structure to the arresting hook trunnions. Testmeasurements consisted of 16 mount strain gages,7 hydraulic jack loads, and 14 displacement transducersto establish engine and aircraft deformations. Note that,because of the vector-force requirements, the overall testsetup was far more complex than would be required fortesting just thrust and g loadings. For a typicalnonvectoring fighter configuration with a muchsimplified restraint and loading system, the test couldhave been conducted within a one-week span. The “coldloads” test required approximately three weeksdedicated aircraft time, including setup, testing, andteardown.

The test was considered very successful and producedvaluable data in support of all five objectives. Maximumloadings simulated were as follows:

• Mount thrust of 20,000 lbf (for each engine)

• Vertical load factor equal to 5 g

• Pitch and yaw vector force greater than 110 percentof design limit

• Combinations of the above including simulation ofpitch-yaw vectoring at maximum thrust andelevated g

Significant results pertinent to this paper related to“cold loads” test objectives 3 and 5 follow below. Enginedisplacements relative to the bay were measured in allthree axes at the inlet seal on the left engine, along withlateral movement at the main-mount pins. At maximumthrust, the longitudinal displacement was approximately0.1 in. For the unrestrained airframe in flight, someadditional displacement combined with thermalexpansion is likely. The test matrix provided extensivedata for evaluating the baseline strain-gage equationsderived from the fixture loadings. Because the builduptest matrix progressed from individual jack loadings tomany realistic combination loadings, assessing theequation errors for each loading scenario was possible.

Figure 13 shows the accuracy of the combinedmount–thrust calculation for two loading cycles. Thecorrelation is excellent to a maximum level of 5000 lbf,but then tracks 2–3 percent high from that point tomaximum load. The hysteresis band is approximately±100 lbf. Examination of the individual pin thrust dataindicates most of the error is caused by the inboard pin,with the outboard pin showing nearly zero error over theentire thrust range. The overall correlation is stillconsidered very good for a pure thrust loading, butpotential for improvement does exist.

Large errors occur for the off-axis loadings,particularly for vertical g loading, as shown in figure 14.


Figure 13. Thrust accuracy during “cold loads” test.

Strain-gagemount thrust,

lbf

Applied mount thrust, lbf4 8 12 16 20 24 x 103

4

8

12

16

20

24 x 103

0

980316

Sum of outboard andinboard mounts

Cycle12

Figure 14. Thrust error caused by vertical

g

loading.


error, lbf

Applied vertical g load, lbf

1 g 2 g 3 g 4 g 5 g0 5 10 15 20 x 103

– 600

– 800

– 400

– 200

0

200

400

600

980317

Sum of outboard andinboard mounts

Thrust

0 lbf12,000 lbf20,000 lbf

The 20,000-lbf vertical load endpoint equates to anequivalent 5-g condition. For the two cases where thrustwas applied first, the error starts high, as shown byfigure 13. As the vertical loading is applied, the errorreverses with a gradient of approximately –300 lbf/g.The total change of –1200 lbf from 1 to 5 g wouldrepresent a significant and artificial thrust loss duringturning performance or drag polar work. Smaller but stillsignificant errors in thrust occurred for pitch vectoring.Until additional regression analysis is performed usingthe “cold loads” data, the current mount-thrustcalculation is considered inadequate for useful researchat elevated g or during vectoring. Mostly for thesereasons, the thrust comparisons in this paper excludeaggressive maneuvering and vectoring.

Flight data for cases where the mount loads are notsignificantly changing in both axes simultaneously arestill considered quite valid. But the “cold loads” testprovided important evidence that improvement in theprimary force calculations and minimization of thecross-axis effects is desired. Most of the errors identifiedduring the “cold loads” test are attributed to the pin andmount gapping previously discussed. The flight resultsin this paper use the baseline equations from the shearfixture calibrations with no corrections from the “coldloads” test.

Off-Aircraft Heating Test

With strain gages matched to the material, afour-active-arm bridge is normally expected to be self–temperature compensating. Because of the relativelylow output on the pins and the encouraging loadcalibration results, the decision was made to runapparent-strain tests to determine if thermalcontamination errors would be worth correcting. Theinstrumented pins and links were placed in a thermaltesting oven, and the bridge output was measured as thetemperature was slowly ramped over the operatingrange. Because of their relative bulk, the pinthermocouples showed large lags compared to the oventemperature. Worst-case bridge outputs varied greatlyfrom bridge to bridge, from as low as 0.02 mV to asmuch as 0.65 mV. These thermal errors are not large inmV terms, but because of the relatively low outputcaused by mechanical load, these errors can representnoteworthy errors in the load calculation.

Figure 15 shows the equivalent load errors for theprime thrust bridges and the load errors summation thatequates to the mount-thrust calculation. The actual mVoutput error is shown in the right margin. Normalin-flight operating temperatures at high power settingsare 200 to 250 °F, based upon flight data. Thetemperature environment is thus considered warm, not


Figure 15. Thrust error from heating test.

Mount-thrustthermal error,

lbf

Temperature, °F

Reference temperature = 70 °F 10 V excitation

50 100 150 200 250 300 350– 1200

– 1000

– 800

– 600

– 400

– 200 – 0.12 mV

– 0.53 mV

0

980318

Typical flightoperating rangeat high power

Inboard mount

Outboard mount

Combined-mount thrust

hot, but the combined error in the mount-thrust equationwould be approximately 700 lbf. At typical maximumaugmented (maximum) power conditions, this valuerepresents a mount-thrust error of approximately–3 percent, increasing to –6 percent at maximumnonaugmented (military) power levels. Theapparent-strain corrections for some bridges are thusconsidered important, not only for research quality data,but also for safety-of-flight monitoring.

Many of the bridge outputs showed minor nonlinearitywith temperature, but simple linear fits were derived foreach bridge and were thought to be adequate. Outputcorrection terms (measured in mV) took the form of aconstant multiplied by the difference in temperaturefrom a reference state, which simplified the real-timeand postflight data processing. Some effort has beenmade to validate the individual thermal corrections usingdata from the “hot loads” test, but the results are not fullyconclusive.

Aircraft Combined Systems Test (“Hot Loads”)

The combined systems ground test was the definitiveclearance check performed prior to committing theaircraft to flight. Major objectives included operationalvalidation of all integrated systems, both hardware andsoftware. In particular, validation of the onboard thrustmodel and the vector-force limiting logic was key. The

aircraft, restrained on the thrust stand facility at EdwardsAir Force Base (California), was capable ofdemonstrating full vectoring with both engines atmaximum power. Relative to “cold loads,” this testingadded the elements of thermal effects and engine systeminterfaces, but did not have inertial g forces adding to themeasured engine-mount data. Prior to installing theengines, a survey of the engine bay was conducted toassess the most likely sources for secondary load paths.The highest potential candidates were identified as thefollowing:

• Engine-face inlet seal: Under worst-case highpower conditions, combined thermal growth anddeformations could have fully compressed therubber seal, allowing some thrust transfer throughthe supporting structure.

• Airframe-to-nozzle fairings: The fairings arecinched down onto a nozzle rub strip. If the cinchstraps were overly tight, a potential for reactingthrust existed.

• Engine-bleed air duct: The design has built-in flexelbows that accommodate engine pitch and yawmovements. How the design would handle forwardengine movement caused by the pin and mountflexibility was not obvious from visual inspection.Forward engine movement could have affected theoutboard mount–thrust measurement.


None of the above issues alone created major concern,but their combined effect could have some potential forrobbing thrust from the main mounts. All other systemconnections (hydraulic, fuel, and electrical) were not aconcern because of built-in flexibility.

Correlating the strain gage–derived thrust to themeasured readings obtained using the thrust stand andoutput from two analytical thrust models (discussed inthe following section) had very encouraging results.Figure 16 shows this comparison, which plots thepercent error, relative to the thrust stand, as a function ofthrottle setting. Data are shown for the baselineequations and a modified result using corrections fromthe “cold loads” test. Both sets of data use the baselinethermal corrections. The strain-gage data incorporatedstandard tare corrections, which forced the thrustcalculation to read 0 lbf prior to engine start. Relative toa ±2-percent error band (a targeted goal), the postflightmodel is outstanding over the entire power range. Theonboard model and the modified strain-gage results skirtthe ±2-percent band, except at idle power where thethrust is extremely low. To compare well, the thermalcorrections had to be reasonable. The secondary loadpaths caused by the system interfaces appeared to beminimal, which verifies initial assumptions.

Flight Data Extraction of Mount Thrust

The mount strain gages respond to the net structuralloads, which represent a summation of the appliedforces: thrust, vector, inertia, and gyroscopic loads, andany external airloads. If these latter forces can beaccurately computed and subtracted from themeasurements, the thrust, pitch, and yaw vector forcescan be determined from the statically determinatereaction equations. Fortunately for the mount-thrustcalculation, the only inertia force is that caused bylongitudinal acceleration. This term can be a significantbut simple correction for large accelerations. A fairamount of effort has gone into careful engineweight-and-balance checks to establish confidence in theinertia properties. The longitudinal inertial load is theonly force term accounted for in the baseline dataprocessing before the mount thrust is converted to grossthrust (described in the next section).

The baseline strain-gage flight data processing isstraightforward and consists of the following steps:

1. Correct the raw bridge outputs for variation inexcitation voltage.

2. Correct bridge outputs for thermal error.


Figure 16. Thrust accuracy from “hot loads” test.

Percent errorrelative to

thrust stand

Throttle position, deg

Modified strain-gageresults use correctionsfrom "cold loads" test

Military power Maximum power

8040 50 60 70 90 100 110 120 130– 4

– 3

– 2

– 1

0

1

2

3

4

5

6

980319

Postflight modelOnboard modelBaseline strain gageModified strain gage

3. Process corrected outputs through the loadcalibration equations.

4. Subtract inertia forces from the pin thrust loads.

5. Sum the outboard and inboard pin loads.

6. Apply pre–engine start tare corrections.

The strain-gage data are measured at 100 samples/sec,but unfortunately for the thrust calculation, thesignal conditioning is ranged to read beyond normalflight operating limits in both the forward and aftdirections. This range, required for structural loadsmonitoring, produces much less than optimumresolution for the mount-thrust measurement. Using a10-bit instrumentation system, current data resolutionfor the prime thrust bridges is 58 lbf/count. For a systemoptimized for positive thrust measurement, the dataresolution is estimated to be as low as 15 lbf/count.Ambient system noise prior to engine start isapproximately 1 count. Combined with accelerometernoise feeding the inertia calculation, the mount-thrustdata, even with the excessive ranging, has a relativelylow root-mean-square noise band (less than ±100 lbf).

In-Flight Thrust Analysis

The strain-gage sensors installed on the engine mountsmeasure the engine force applied directly to the airframe(the mount thrust) and, therefore, measure a load that ismore similar to net thrust than to gross thrust.Unfortunately, because of the manner in whichpropulsive forces are accounted for, net thrust, as it is

defined, does not always resemble the applied force atthe engine mounts, particularly at high speeds. To deriveeither net thrust or gross thrust from the direct thrustmeasurement, considering and accounting for allsignificant forces acting upon the engine directly andsumming these forces as required with the direct valuetherefore becomes necessary. Figure 17 shows thetypical array of forces acting upon an installed engine, asin the F-15 ACTIVE vehicle.

Assumptions

With the goal of this exercise being to compute grossthrust, several assumptions would obviously be requiredregarding the forces shown in figure 17, because manyforces are immeasurable or were not instrumented forthe ACTIVE program. For example, connector drag andaircraft interface rub are both unknown; however, stepswere taken to minimize their effect so that they could beassumed equal to 0 lbf. In particular, connector drag wasminimized by providing freedom of movement in thecabling and plumbing linking the airframe with themotor. The ACTIVE F-15 aircraft has a unique cinchstrap system, used to tighten the aircraft titanium sealingflaps against the aft static structure of the nozzle. Thissystem is believed to provide a balanced distributionbetween bay and ambient pressure which simplifies thegross thrust calculation as discussed later in this section.

No simple steps could be taken to ensure theelimination of inlet-seal reaction, but the seal designshould accommodate a minimum of 0.25 in. oflongitudinal deformation and thermal expansion of the


Figure 17. Forces acting upon an installed engine.

Aircraftinterface

rub

Nozzledrag

Grossthrust,Fgross

Engine-bodypressure force

Connectordrag

Engine-mount

reaction,Fmount

Engine-face pressure drag, Fface pressure

Inlet-seal reaction

Engine-face ram drag, Fram

Engine-body pressure force, Fbody pressure

Engine-bodypressure force

980320

engine. With no other recourse, inlet-seal reaction wastherefore assumed equal to 0 lbf. As will be discussed,some evidence exists that indicates very highgross-thrust conditions may cause nonnegligibleinlet-seal reaction, but only in a small portion of theflight envelope at maximum (full augmentation) power.

Nozzle drag is essentially an immeasurable force thatis normally computed, as in the case of the F-15 airplane,using the airframe manufacturer’s aircraft-engineinstallation effects model. The installation effects modelalso computes such variables as engine-bleed air andhorsepower extraction. However, the model presents alarge computational burden itself, and its use was notconsistent with the goal of a reduced-complexity thrustcalculation method using as input only measured data ordata computed in real time onboard the aircraft.

Conveniently, nozzle drag is a negligible factorthroughout much of the flight envelope at augmentedpower. However, with the smaller nozzle configurationsat military (maximum nonaugmented) power and less,nozzle drag rises dramatically as speed approachesMach 1. But gross thrust also rises rapidly as speedincreases, so the effect on thrust of neglecting nozzledrag is minimized in percent terms. With this effect inmind, nozzle drag is assumed equal to 0 lbf for thisanalysis. The effects of this assumption will be discussedin detail in the “Results and Discussion” section.

Gross-Thrust Calculation

The remainder of the force terms are nonnegligible;however, these terms can all be computed from theinformation available from the F-15 ACTIVEinstrumentation data stream. As a result, gross thrust cannow be calculated based upon the strain gage–basedaxial-force measurement at the engine mounts.

Summarizing the force terms shown in figure 17 andignoring those terms assumed to equal zero in the abovesection, the following equation results:

(1)

where is gross thrust, is the straingage–based axial-force measurement at the enginemounts, is the engine inlet–plane ram-drag termcaused by engine intake flow momentum,

is the drag caused by flow pressure onthe engine face, and is the resultingforce acting on the aft-looking-forward projected area of

the engine and includes the effects of ambient andengine-bay pressure. All forces are measured in unitsof lbf.

The projected area over which the inlet-face andnozzle-body pressure forces counteract each other issimplified by the assumption explained earlier that theengine-bay pressure is equal to ambient. As a result ofthis assumption, the face and body (ambient) pressurescounteract each other over the exposed engine-inlet facecross-sectional area, , and which is equal to951.0 in2 for the F100-series engine. Ambient pressure,

, is a position-corrected measurement available fromthe aircraft data stream. Engine inlet–face staticpressure, , is available as a production output of theIDEEC computer and is a spatially averagedmeasurement from the engine nosecone probe.

With and now known, the following equationcan be constructed:

(2)

where the inlet cross-sectional area is measured in unitsof in2 and the pressures are measured in units of lbf/in2

absolute.

The engine inlet–plane ram-drag term, , is nowthe remaining unknown variable on the right side ofequation 1. This variable is computed (as shown in thefollowing equation) by multiplying true engine-inletmass flow, , in units of lbm/sec with engineinlet–plane velocity, , in units of ft/sec, both of whichcan be derived as shown below using output parametersfrom the IDEEC.

(3)

Engine mass flow is computed within the IDEECbecause mass flow is an important parameter for controlscheduling and stability management. However, theengine mass flow is output from the IDEEC in an enginestation–corrected format and must be converted into trueinlet mass flow to be usable for this application. Thisconversion is shown in the following equation:

(4)

Fgross Fmount= Fram F face pressure + +

F body pressure –

Fgross Fmount

Fram

F face pressureFbody pressure

Ainlet

P0

P2

P0 P2

F face pressure F body pressure – =

A

inlet

P

2

P

0

–

( )⋅

Fram

WAT 2V 2

Fram

WAT 2 V 2⋅32.17

---------------------------=

WAT 2 WACCPT 2

14.70-------------

T T 2459.67+

518.67-------------------------------

0.5–

⋅ ⋅=


where

WACC

is the IDEEC-estimated corrected mass

flow measured in lbm/sec, is the IDEEC estimate

of engine-inlet total pressure measured in lbf/in

2

absolute, and is the measured engine-inlet total

temperature, also available from the IDEEC, measured

in °F.

Finally, is computed according to standard gasdynamic relationships as shown in the following twoequations. First,

(5)

where is the sonic velocity in ft/sec at the engine

inlet, and is the Mach number of the incoming flow.

By expanding these terms and using the appropriate gas

constant and specific heat ratio for air, the following

equation can be derived:

(6)

where the engine-inlet plane velocity is now completelyexpressed in previously defined terms. Note that whenequation 6 and equation 4 are used in equation 3, the

terms cancel.

All terms in equation 1 are now known, and grossthrust based upon the direct measurement at the enginemounts can be computed. Note that net thrust can easilybe derived from the gross-thrust value by subtractingaircraft inlet–plane ram drag from equation 1. Aircraftinlet–plane ram drag is computed by multiplying trueengine mass flow by the velocity of the aircraft.

Computing the engine inlet–plane ram-drag andpressure-force terms is not a complex matter providedthat the required input parameters are readily availableon the aircraft data stream, as in this case. Using theproduction data stream of a digital engine controller cantherefore prevent the costly requirement to instrumentthe engine face in order to measure the temperatures andpressures necessary to calculate these force terms.

Benchmark Analytical Model Descriptions

Two analytically-based models supplied by Pratt &Whitney were used to compute gross thrust in order toserve as benchmarks for the strain gage–based directthrust-measurement technique. These models were thepostflight aerothermodynamic thrust model and theonboard nozzle controller thrust model.

Postflight Aerothermodynamic Thrust Model

The postflight model

8

is a high-fidelity

aerothermodynamic simulation of the F100-PW-229

engine and is designed for customer use. A combination

of engine performance modeling, engine component

ground test data, and measured engine and aircraft flight

data permit the model to accurately calculate thrust

calculation using the mass flow–temperature

method.9, 10 The use of measured engine parameters

allows the model to partially compensate for engine-to-

engine performance variations. Measured values of

free-stream altitude and Mach number, fan speed, fan

guide-vane angle, turbine discharge total pressure, core

and afterburner fuel flows, , and were used as

inputs to the model for this analysis. Free-stream altitude

and Mach number were obtained from the flight

controller, and the remainder of the parameters were

obtained from the IDEEC. On some missions,

high-accuracy, flight test, volumetric fuel flow meters

were available and used in place of the IDEEC values.

An F-15 installation effects subroutine, developed byThe Boeing Company (formerly McDonnell DouglasAerospace), is used to account for airframe detriments toengine thrust performance, including horsepower andbleed air extraction. The combined engine-thrust andinstallation-effects modeling process is verycomputationally intensive and can require several hourson a dedicated computer workstation to process a singlemission. The uncertainty band11 for calculated grossthrust from this model is estimated to range from 2 to4 percent, depending on flight condition and powersetting.

Onboard Nozzle Controller Thrust Model

A gross-thrust model residing within the nozzlecontroller onboard the F-15 ACTIVE aircraft is used toprevent the production of unsafe vector forces by thenozzle. This model was derived from the postflightmodel described in the above section and also usesmeasured data as input. However, this model usessimplified routines to speed the execution of the code,permitting its use in the real-time application but alsoincreasing the uncertainty of the calculation.

The area-pressure thrust calculation method is used bythe onboard model, which reduces the number of inputparameters required to compute gross thrust. Nozzle

PT 2

T T 2

V 2

V 2 V sonic M2⋅=

V sonicM2

V 2 12014 T T 2459.67+( ) 1

PT 2

P2--------

0.286–

–

⋅ ⋅=

T T 2

PT 2T T 2


throat area and turbine-discharge total pressure, bothderived from the IDEEC, are used as input, along with

from the flight controller. The reduced number ofinput parameters further increases the gross-thrustuncertainty relative to the postflight model because thearea-pressure thrust calculation method tends to be moresensitive to input measurement error than the massflow–temperature method.

The uncertainty band for this model is greater than forthe postflight model; and the error increment grows asthe engine departs from an average health state becausethe onboard model has less ability than the postflightmodel to accommodate off-nominal engine operation.Because of the critical role the onboard model plays inlimiting the magnitude of the nozzle-vectoring forces inreal time, an important objective of the ACTIVEprogram has been to validate the onboard gross-thrustmodel against the postflight model.

Analysis Constraints and Scope

For the purposes of this analysis, directthrust-measurement data were processed only forquasi-steady-state aircraft and engine operation. Theobjective in doing so was to minimize the uncertainty incomparing the direct thrust-measurement techniqueagainst the reference models, which were designed forquasi-steady-state engine operation only. And althoughthe strain-gage signal conditioning process has beendesigned to accommodate aircraft maneuvering andnozzle-vector forces, the effects of these forces on theengine-mount axial-force readings are still beingquantified throughout the flight envelope and are not yetfully understood, as discussed previously.

As a result, the data were filtered to eliminate any timecuts with normal aircraft acceleration greater than 2 gand less than 0 g, lateral acceleration greater than 0.1 g,aircraft pitch rate greater than 2.5 deg/sec, yaw rategreater than 2 deg/sec, roll rate greater than 6 deg/sec,pitch angle greater than 10°, roll angle greater than 20°,and climb rate greater than 50 ft/sec. The data were alsofiltered for 6 sec following any rapid throttle positionchange of more than 5°. All thrust data collected duringnozzle vectoring or off-nominal exit area schedulingwere also excluded by filtering.

This analysis focused on two throttle settings, militaryand maximum power, primarily because of the largequantity of data available at these two power settingsthroughout the ACTIVE flight envelope and because thereference models are considered most accurate atmilitary and maximum power. Thirteen missions, spacedover an eight-month period in 1996, were selected

because those missions most fully covered the flightenvelope and offered a wealth of stabilized engine datafrom low to high Mach number. These missions wereanalyzed in their entirety, and their data were filteredaccording to the above constraints. The result was a database of 3822 time cuts at military power and 1420 timecuts at maximum power.

Differences were computed between the directthrust-measurement method values and the output fromthe reference models for each time cut of data. The datafrom all flights were combined for each power settingand then sorted based on Mach number and altitude intoblocks, each spanning 5000 ft and 0.1 Mach. The datafor each parameter within each block were thenaveraged. At military power, 52 Mach-altitude blocksresulted; at maximum power, 58 blocks resulted. Thelarge spread of the data throughout the flight envelopecan be seen on the Mach-altitude cross plots in the“Results and Discussion” section. Scale references havebeen removed on absolute data to protect the proprietarynature of the information.

Results and Discussion

Figure 18 shows the percent difference betweenmilitary-power gross thrust computed using the directthrust-measurement technique and gross thrustcomputed by the postflight model. The difference isplotted as a function of Mach number for arepresentative range of altitudes: 10,000 ft, 20,000 ft,30,000 ft, and 45,000 ft. The percent differences rangefrom approximately 1 to 12 percent, depending on flightcondition. At a given altitude, the difference tends todecrease as subsonic speed increases. This trend reversesprior to Mach 1.0, with the difference peaking nearMach 1.0 and then decreasing again as the speedincreases. The average percent difference for allmilitary-power points is 4.2 percent.

Figure 19 shows a similar arrangement of data asshown in figure 18, except that the power setting is nowat maximum. In this case, percent differences rangebetween approximately 2 and 13 percent. The local peakin the percent differences seen in the vicinity of Mach 1at military power is less clear at maximum power;however, a strongly decreasing trend is evident at analtitude of 45,000 ft to a maximum speed of Mach 2.0.At an altitude of 30,000 ft, the percent difference isrelatively stable as speed increases to a maximum ofMach 1.7, beyond which the difference abruptly risesand continues to do so until Mach 2.0 is reached. Theaverage percent difference for all maximum-powerpoints is 3.8 percent.

P0



Figure 18. Percent difference in gross thrust between direct thrust-measurement technique and postflight model as afunction of flight condition; military power.

Figure 19. Percent difference in gross thrust between direct thrust-measurement technique and postflight model as afunction of flight condition; maximum power.

Difference,percent

Mach number

Average

.2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.00

2

4

6

8

10

12

14

980321

Altitude10,000 ft20,000 ft30,000 ft45,000 ft


Difference,percent

Mach number

Average

.2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.00

2

4

6

8

10

12

14

980322

Figure 20 shows absolute differences in gross thrust atmilitary power between the direct thrust-measurementtechnique and the postflight model. A sharp change inthe difference occurs at Mach 0.8 and faster at altitudesof 10,000 ft and 30,000 ft (no data exist for speedsgreater than Mach 0.8 at an altitude of 20,000 ft).

At a given flight condition, nozzle drag is typicallymuch higher at military power than it is at maximumpower. This difference grows dramatically in the lowerright-hand side of the flight envelope, and computermodels indicate that nozzle drag can exceed 1000 lbf atthe highest dynamic pressures. At these flight conditions,nozzle drag at maximum power actually reversesthrough 0 lbf as the large, outboard, divergent-sectionboat-tail angles combined with the large throat areascreate a forward-acting force component. If thesemodeled nozzle-drag values are included in equation 1for military power, the supersonic curves tend to matchmore closely with their subsonic counterparts at a givenaltitude. Regardless, the additional error caused by theexclusion of nozzle drag in equation 1 does not result ina correspondingly large percentage difference (as can beseen in figure 18) because as speed increases, so doesgross thrust.

Figure 21 shows the absolute difference in gross thrustat maximum power between the direct thrust-measurement technique and the postflight model. In this

case, the differences tend to be somewhat stable withMach number, with the exception that at Mach 1.8 and analtitude of 30,000 ft, a dramatic change occurs as Machnumber increases. This event can also be seen infigure 19 in the same portion of the flight envelope. Thissudden departure in the data occurs in the highestdynamic-pressure portion of the F-15 ACTIVE flightenvelope where gross thrust is highest and enginethermal conditions are most severe. One possible causeis that the high gross-thrust values were combining withmaximized thermal lengthening of the engine and thusflattened the engine inlet J flange against the K-sealinterface. With a new unquantifiable load pathdeveloping, the engine mounts experienced less load,and thus the difference between the directthrust-measurement technique and the postflight modelgrew with speed.

As shown in figures 18 and 19, the percent differencesbetween the two techniques are somewhat high ataltitudes of 30,000 ft and 45,000 ft and low Machnumber conditions. But as figures 20 and 21 show, theabsolute differences at these conditions do not stand outwhen compared with other flight conditions. Grossthrust rapidly decreases as the upper left-hand corner ofthe flight envelope is approached, and so the response inpercentage difference is amplified there.


Figure 20. Absolute difference in gross thrust between direct thrust-measurement technique and postflight model as afunction of flight condition; military power.


400 lbfDifference,

lbf

Mach number

.20 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0

980323

Figure 21. Absolute difference in gross thrust between direct thrust-measurement technique and postflight model as afunction of flight condition; maximum power.


400 lbfDifference,

lbf

Mach number

.20 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0

980324

Bubble plots are displayed on a Mach-altitude crossplot in figures 22 and 23 for the percent differences atmilitary and maximum power, respectively, and infigures 24 and 25 for the absolute differences at militaryand maximum power, respectively. An outline of theflight envelope of the F-15 ACTIVE aircraft is alsoshown on each figure for reference. Whereas figures 18to 21 show data trends at selected altitudes for the sakeof clarity, figures 22 to 25 show all of the Mach-altitudeaveraged data and qualitatively show the trends in thedata on the full flight envelope.

The size of each bubble is directly proportional to itsvalue, and this proportionality is maintained from plot toplot. Of particular note are the high percentagedifferences seen in the upper left-hand corner of theflight envelope and the large absolute differences seenalong the lower right-hand edge of the envelope, bothpreviously discussed. Note the very small differences,both absolute and percent at each power setting, for thetakeoff roll near sea level; some of the data point bubblescannot be seen on the plots because the bubbles aresmaller than the width of the flight envelope boundary.

Figure 26 shows gross thrust as a function of Machnumber at an altitude of approximately 30,000 ft for amaximum-power acceleration from Mach 0.75 toMach 1.95 and the subsequent deceleration at military

power. Gross thrust is shown for the directthrust-measurement technique, the postflight model, andthe onboard model. At speeds less than Mach 1.3 duringdeceleration, the engine is closer to idle power than it isto military power. This particular example illustrates thelargest absolute difference seen in the data base betweenthe direct measurement technique and the postflightmodel, which occurred at maximum power, a speed ofMach 1.95, and an altitude of 32,000 ft. This differenceis possibly the result of inlet-seal reaction as discussedearlier. Despite this difference, the figure shows verygood agreement between the direct thrust-measurementtechnique and the two models, especially considering thedata span an enormous range of aircraft and engineoperating conditions.

Note that the direct thrust-measurement techniqueshows very close agreement with the onboard modelduring the acceleration to a maximum speed ofapproximately Mach 1.6. In fact, as a general rulethroughout the flight envelope, the onboard modeltended to agree more closely with the directthrust-measurement technique than with the postflightmodel. The average percent difference between thedirect thrust-measurement technique and the onboardmodel was 2.2 percent for all military-power points and1.2 percent for all maximum-power points.



Figure 22. Relative change in percent difference in gross thrust between direct thrust-measurement technique andpostflight model as a function of flight condition; military power.

Figure 23. Relative change in percent difference in gross thrust between direct thrust-measurement technique andpostflight model as a function of flight condition; maximum power.

Altitude,ft

Mach number

0 .2 .4 .6 .8 1.0 1.2 1.4 1.6 2.01.8

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

50,000

980325

Altitude,ft

Mach number

0 .2 .4 .6 .8 1.0 1.2 1.4 1.6 2.01.8

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

50,000

980326


Figure 24. Relative change in absolute difference in gross thrust between direct thrust-measurement technique andpostflight model as a function of flight condition; military power.

Figure 25. Relative change in absolute difference in gross thrust between direct thrust-measurement technique andpostflight model as a function of flight condition; maximum power.

Altitude,ft

Mach number

0 .2 .4 .6 .8 1.0 1.2 1.4 1.6 2.01.8

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

50,000

980327

Altitude,ft

Mach number

0 .2 .4 .6 .8 1.0 1.2 1.4 1.6 2.01.8

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

50,000

980328

Figure 26. Gross thrust calculation method comparison for a maximum-power acceleration and military-powerdeceleration at an altitude of approximately 30,000 ft.

Postflight modelOnboard modelDirect measurement

5000 lbf

Acceleration

Deceleration

Grossthrust,

lbf

Mach number.8 1.0 1.2 1.4 1.6 1.8 2.0.6

980329

The low-noise characteristics of the strain gage–basedmethod is also evident in this figure. In fact, the standarddeviation of the direct thrust-measurement techniquetends to be as good as or slightly better than theanalytical methods throughout the flight envelope atstabilized flight conditions.

Table 2 shows the bias and standard deviation of thedirect thrust-measurement values relative to the resultsfrom the two analytical methods. From the data in thetable, the vast majority of direct thrust-measurementdata evidently falls below the postflight model at bothpower settings. This skew is currently underinvestigation, although obvious potential causes includeunaccounted secondary forces and postflight model

input data bias. The latter is thought to be likely becausethe direct thrust data are more equally distributed aboutthe onboard model results, as the table shows.

As airspeed increases at a given altitude, theengine-pressure drag (eq. 2) reverses at approximatelyMach 0.6 from a suction force that actually adds to netpropulsive force to a true drag that grows exponentiallywith Mach number. At Mach 2.0 and an altitude of30,000 ft, engine-pressure drag is approximately50 percent of the maximum-power gross-thrust value;the percentage is even greater at military power. Thechanges in engine-face ram drag as a percentage of grossthrust are more gradual than engine-pressure dragthroughout the flight envelope, decreasing slightly with


Table 2. Percent bias and standard deviation of the direct thrust-measurement method relative to the two analytical methods; military andmaximum power.

Postflight model Onboard model

Power settingData points

analyzed BiasStandarddeviation Bias

Standarddeviation

Military 3822 – 4.22 3.61 – 2.24 4.14

Maximum 1420 – 3.81 2.44 +1.16 2.11

Mach number, but nonetheless averaging approximately30 percent of the gross-thrust value for military powerand 14 percent at maximum power.

These results qualitatively imply that gross-thrustvalues as calculated by the direct thrust-measurementtechnique in equation 1 would be sensitive to significanterror in the ram-drag and pressure terms, particularly athigh speeds. Care should be taken in accounting forthese forces.

Concluding Remarks

Gross thrust was successfully computed using a straingage–based direct thrust-measurement techniqueapplied to the NASA F-15 Advanced ControlTechnology for Integrated Vehicles (ACTIVE) aircraftpowered by two F100-PW-229 engines. Results wereobtained at quasi-steady-state conditions throughout theflight envelope of the vehicle using data collected during13 missions spanning an eight-month period. More than5200 time cuts of data were processed to produce theresults in this report, representing a Mach range from 0.0to 2.0 and altitudes from near sea level to higher than45,000 ft. Both military and maximum augmentedpower settings were studied.

The strain-gage calibration process, described in thispaper, was developed in support of the ACTIVEprogram. This process has established the feasibility ofmeasuring engine main-mount loads and extracting theaxial thrust at the mount. The methodology has evolvedand benefited from four distinct ground tests: shearfixture loadings, on-aircraft hangar testing (engines notinstalled), thermal tests, and combined systems testing(engines installed and operating). Feeding results fromeach step forward to understand and minimize eachsource of error to the greatest extent possible has beenfundamental. The approach for instrumenting andcalibrating the pins presents certain challenges but offerssome distinct advantages and options: the pins areamenable to either on- or off-aircraft load calibration,and the pins easily fit in an oven for thermal testing. Inaddition, installing the pins in another aircraft or teststand is an option unique to this approach.

Another important advantage is the direct interceptionof the engine-force reactions. No concern exists for thegages responding to strain induced by other forces.Among the potential problems anticipated, none hasproven to be significantly detrimental to thethrust-measurement process. However, the influence ofthe engine-mount clamping flexibility and resultinggapping was not anticipated and has aggravated certain

problems that would have otherwise been minor ornonexistent. Revised calibration equations are warrantedto minimize the cross-axis errors and improve theprimary force calculations.

Gross thrust computed using the direct measurementtechnique differed from results obtained using apostflight aerothermodynamic engine model by anaverage of 4.2 percent at military power and 3.8 percentat maximum power, all data considered. When comparedagainst an onboard gross-thrust model within the nozzlecontroller, the differences averaged 2.2 percent atmilitary power and 1.2 percent at maximum power.Outside of a local peak occurring at military power atapproximately Mach 1.0, the percent difference tendedto decrease as speed increased. Percent differences werehighest in the upper left-hand corner of the flightenvelope where gross-thrust values are small and theeffect of absolute differences on percent is amplified.

Engine inlet–plane ram drag, engine inlet–planepressure drag, and ambient-pressure force on the nozzlewere the only secondary forces included in the processof transforming the engine-mount strain-gagemeasurements into gross thrust. These three forces werecomputed in a straightforward manner using informationreadily available on the digital aircraft and enginecontroller data buses. In fact, the direct use of the enginecontroller output eliminated the need for costlyadditional flight test instrumentation.

The engine pressure term rises exponentially withspeed and can exceed 50 percent of the gross-thrustvalue at Mach 2.0. The engine-face ram-drag termaccounted for an average of 30 percent of military-powergross thrust and 14 percent at maximum power. Theselarge values for the percentages indicate their importantrole in the accurate calculation of gross thrust using thedirect thrust-measurement technique, and care must betaken in accounting for these forces.

Other secondary forces were assumed to equal 0 lbfbecause they could not be easily quantified and theireffects were believed to be minimal for this application.These secondary forces included connector forces andinlet-seal reactions. Nozzle drag, a nonzero secondaryforce, was not included in the analysis, primarilybecause of the large computational burden required tocalculate it. The use of a large analytical model wasthought to be inconsistent with the objective here of asimplified approach to thrust calculation; and, in percentterms, the exclusion of nozzle drag from the calculationprocess had only a minor impact on the results.


Sensor Installation and Calibration Process Recommendations

• Application of the methodology developed hereshould be evaluated on a case-by-case basis for aparticular aircraft configuration; not allconfigurations would be amenable to the technique.

• A good portion of the calibration accuracy isattributed to the gage selection and high-precisioninstallation relative to the load axes.

• Maintaining precise angular alignment of the pinrelative to the mount and the load axes will alwaysbe critical.

• For future applications, an assessment of the valueor necessity of conducting both a fixture loadingtest and a “cold loads–type” test would beadvisable. If a high-fidelity calibration effort isperformed with either test, performing both tests isprobably not necessary. Either test requires customhardware and test setup. If an off-aircraft fixtureloading is more desirable, the setup and loadingshould replicate the real aircraft as much as feasible.

• The single most beneficial improvement for anyfuture application on a similar configuration wouldbe to increase the stiffness of the main-mountclamping structure. The benefits would be reducedengine deflections and simplified calibration of themount adapter pins.

Considerations for the Analysis Process

• Simplifying assumptions regarding secondary forceterms must be given careful thought prior tocomputing gross and net thrust using the directthrust-measurement technique. Ignoring these forceterms may be acceptable on one type of airframeinstallation but not on another. In the absence of ananalytical thrust model with which to provide areference check, using carefully selected sensors toflag unaccounted secondary load paths (should theydevelop) may be prudent. For instance, alongitudinal engine displacement transducer can beused to ensure that high-speed inlet-seal bottomingis not occurring on a buried-engine installation.

• The use of output data from the digital enginecontroller, if one is available, should be capitalizedupon. Provided that high-rate dynamic-thrustmeasurement or extreme accuracy are notrequirements, the use of this readily available

information can save costs by eliminatingunnecessary specialized flight test instrumentation.

• When the flight test program commences, anddepending on the program duration, periodicin-flight tare readings of the direct thrust-measurement technique against a reference modelmay be prudent to flag a need for strain-gage sensorrecalibration. A simplified engine performancespecification model would serve well in thiscapacity and need not be as sophisticated as thepostflight model described in this report.

Proposed Additional Analyses

Several additional analyses are proposed to furtherdocument the utility of the direct thrust-measurementtechnique:

• Collect additional data in those parts of the flightenvelope where data are scarce.

• Improve the accuracy of the strain-gage calibrationequations during all-axis maneuvering and thrustvectoring.

• Analyze the sensitivity to errors in input parametersand in the calculation and rejection of secondaryforces.

• Fully quantify the accuracy below military power,particularly for cruise power settings.

• Investigate reducing bias error by using anappropriate correlation parameter readily availableon the engine data bus.

• Increase the accuracy of the benchmark analyticalmodels by maximizing the accuracy of their inputparameters.

• Study the benefit of including a simple table lookupfor nozzle drag in order to reduce high-speed thrusterror.

• Determine if high-speed inlet-seal bottoming is anissue on the F-15 airplane by analyzing longitudinalengine displacement at the engine face.

• Study the accuracy stability of the engine-mountstrain gages over long periods of time.

• Document the capability and limitations forhigh-response and dynamic-thrust calculation.

• Fully quantify the standard deviation and errorbands against the analytical methods.


References

1Society of Automotive Engineers, Incorporated,Committee E-33, “In-Flight Thrust Determination,”SAE-AIR-1703, Nov. 1985.

2Ray, Ronald J., Evaluating the Dynamic Response ofIn-Flight Thrust Calculation Techniques During ThrottleTransients, NASA TM-4591, 1994.

3Groth, Harold W., Nick E. Samanich, and Philip Z.Blumenthal, Inflight Thrust Measuring System forUnderwing Nacelles Installed on a Modified F-106Aircraft, NASA TM-X-2356, 1971.

4Fogg, Alan, “Direct Measurement of In-Flight Thrustfor Aircraft Engines,” Proceedings of the ThirteenthAnnual Symposium, Society of Flight Test Engineers,Sept. 1982, pp. 33–41.

5Smolka, James W., et al, “F-15 ACTIVE FlightResearch Program,” Fortieth Symposium Proceedings,Society of Experimental Test Pilots, Sept. 1996,pp. 112–145.

6Doane, P., R. Bursey, and G. Schkolnik, “F-15ACTIVE: A Flexible Propulsion Integration Testbed,”AIAA-94-3360, June 1994.

7McDonnell Aircraft Company, “STOL/ManeuverTechnology Demonstrator: Volume 4, AxisymmetricFlight Test Report,” WL-TR-91-3083, Sept. 1991.

8United Technologies Pratt & Whitney,“F100-PW-229 Status Simulation With P/Y BBNVectoring Capability: User’s Manual for CCD1430–00.0,” FR-20248-26, Dec. 1994.

9Burcham, Frank W., Jr., An Investigation of TwoVariations of the Gas Generator Method to Calculate theThrust of the Afterburning Turbofan Engines Installed inan F-111A Airplane, NASA-TN-D-6297, 1971.

10Conners, Timothy R., Measurement Effects on theCalculation of In-Flight Thrust for an F404 TurbofanEngine, NASA TM-4140, 1989.

11Society of Automotive Engineers, Incorporated,Committee E-33, “Uncertainty of In-Flight ThrustDetermination,” SAE-AIR-1678, Aug. 1985.


REPORT DOCUMENTATION PAGE Form ApprovedOMB No. 0704-0188

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering andmaintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington,VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

4. TITLE AND SUBTITLE 5. FUNDING NUMBERS

6. AUTHOR(S)

8. PERFORMING ORGANIZATION REPORT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING AGENCY REPORT NUMBER

11. SUPPLEMENTARY NOTES

12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

13. ABSTRACT (Maximum 200 words)

14. SUBJECT TERMS 15. NUMBER OF PAGES

16. PRICE CODE

17. SECURITY CLASSIFICATION OF REPORT

18. SECURITY CLASSIFICATION OF THIS PAGE

19. SECURITY CLASSIFICATION OF ABSTRACT

20. LIMITATION OF ABSTRACT

NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)Prescribed by ANSI Std. Z39-18298-102


WU 529-20-04-00-33-00-MGT

Timothy R. Conners and Robert L. Sims

NASA Dryden Flight Research CenterP.O. Box 273Edwards, California 93523-0273

H-2266

National Aeronautics and Space AdministrationWashington, DC 20546-0001 NASA/TM-1998-206560

Direct thrust measurement using strain gages offers advantages over analytically-based thrust calculationmethods. For flight test applications, the direct measurement method typically uses a simpler sensorarrangement and minimal data processing compared to analytical techniques, which normally require costlyengine modeling and multisensor arrangements throughout the engine. Conversely, direct thrust measurementhas historically produced less than desirable accuracy because of difficulty in mounting and calibrating thestrain gages and the inability to account for secondary forces that influence the thrust reading at the enginemounts. Consequently, the strain-gage technique has normally been used for simple engine arrangements andprimarily in the subsonic speed range. This paper presents the results of a strain gage–based direct thrust-measurement technique developed by the NASA Dryden Flight Research Center and successfully applied to thefull flight envelope of an F-15 aircraft powered by two F100-PW-229 turbofan engines. Measurements havebeen obtained at quasi-steady-state operating conditions at maximum nonaugmented and maximum augmentedpower throughout the altitude range of the vehicle and to a maximum speed of Mach 2.0, and are comparedagainst results from two analytically-based thrust calculation methods. The strain-gage installation andcalibration processes are also described.

Accuracy, Analytical models, Calibration, Direct thrust measurement, F-15,F100-PW-229 engine, Strain gage, Supersonic, Thrust calculation

A03

37

Unclassified Unclassified Unclassified Unlimited

July 1998 Technical Memorandum

Presented at the 34th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Cleveland, Ohio, July 13–15,1998.

Unclassified—UnlimitedSubject Category 05

Documents

Full Flight Envelope Direct Thrust Measurement on a ... · FULL FLIGHT ENVELOPE DIRECT THRUST MEASUREMENT ON A SUPERSONIC AIRCRAFT Timothy R. Conners * and Robert L. Sims † NASA