i "' WC-ASL 10 9 4

U. S. NAVAL AIR ENGINEERING CENTERPH I LA DE LP K IA, PE NNS Y LVAN IA

, rn . AERONAUTICAL STRUCTURES LABORATORReport No. NAE-ASL-1094 Volume IV

Date: 31 August 1966

APPLICATION OF APPLIED LOAD RATIO STATICTEST SIMLATION TECHNIQUES TO FULL-SCALE

STRCUMES; VOLUM IV, SUPMRY OF RESULTS,CONCLUSIONS AND RECOM.NDATIOITS

by-' ; 'R. W. Gehring and 0. H. Maines

Noeoh American Aviation, TnColumbus Dv.

Columbus, OhioContract No. 1J156-45330

Distribution of thisi , I document is unlimited

MA 2 967

II.

, I

I 4D. ACC.$t21/3 iIN. ) V..1 PLATE No. 11749

DISCLAIMER NOTICE

THIS DOCUMENT IS BEST QUALITYPRACTICABLE. THE COPY FURNISHEDTO DTIC CONTAINED A SIGNIFICANTNUMBER OF PAGES WHICH DO NOTREPRODUCE LEGIBLY.

IMF W OR E -i

F1

SlTI

l ll l ll l.......................... :....

CODE:11,'~~~~~~ 011V T/kll. IIil C El

l NOTICE

Reproduction of this document In any form by otht than naval activitiesIs oot authorized except by special approval a the Secretary of the Navyor the CIf of Naval Operations al cppropriate.

The following Espionage notice can be disregarded unless-this documentis plainly marked CONFIDENTIAL or SECRET.

This document contains information affecting ?he national defense of ahe,

United States within the meaning of the Espionage Laws, Title 18, U.S.C.,Sections 793 and 794. The tconsmissfon or the revelation f ts contents inany manner to an unauthorized person is. prohibited by law.

ii

FCKEWORD

This study was performed under Naval Air Engineering Center ContractN156-45330 and is primarily an analytical study and experimental datacorrelation. The analysis and data correlation portion of the pro-gram was performed by the Structural Development Group, .Vehicle Tech-nology Engineering, North American Aviation, Inc., Cblumbus Division.The experimental data was obtained from static tests performed onaircraft horizontal stabilizer assemblies at room and elevated tempera-ture by the Aeronautical Structures Laboratory, Naval Air EngineeringCenter, Philadelphia, Pennsylvania, with Mrs. Elise Alter as ProjectEngineer. This volume represents that portion of the final reportdescribing the application of the methods of Volue I, the materialproperties data of Volume II, and the analytical-experimental datacorrelation using the data obtained from the experimental programdescribed in Volume III. Conclusions, recommendations, and problemareas related to the realistic use of applied load ratio simulationare also included in this volume. This program was begun in July1964 and the final draft of Volume IV was completed in January 1966.

r,

L

The objectives of Volume IV are to correlate analytical predictionswith experimental results, and to present a general summary of results,also related studies, special analysis and experimental problemsencountered, and detailed conclusions and recommendations; thereforethis Volume IV is a summary report that includes significant data fromVolumes I through III. The main objective of this study was to demon-strate the applicability of applied load ratio simulation and thedegree of accuracy to be expected when simulating the elevated tempera-ture static test in a room temperature environment.

The results of Part 1 of Volume IV support the use of applied loadratios for room temperature simalation of elevated temperature statictests within the limitations of the analytical methods. The accuracyto be expected for bending tests is shown to be within +i0% based uponthe failing load comparison, but depends, to a large extent, on theaccuracy of the basic material properties.

Part 2 of this study investigaes the feasibility and applicability ofapplied load ratios to the strength verification of a multi-plate,multi-fastener splice joint at elevated temperatures by room tempera-ture testing. The analytical methods used apply to axially-loadedsplices, and were successfully applied to the study of the horizontalstabilizer root end splice joint which was heated botn symmetricallyand unsymmetrically and loaded primarily by shear loads normal to thestabilizer surface and a bending moment in the splice joint area. Theaccuracy demonstrated for splice joint static test simulation is shownto be within t16% based upon the failing load comparison and dependsupon the accuracy of fastener-plate load-deformation data.

FL

ii L

VQLMZI

TABLE OF CONTENTS

FOREWORD i

SUMMARY

SYMOLS xi

REFERENCES xv

1.0 INWAODUCTION 1

2.0 DISCUSSION OF ANALYTICAL PROGRAM 2

2.1 Strength and Deformation Analysis Procedures 22.1.1 Cross-Section Analysis, Part 1 32.1.2 Inboard Splice Joint Analysis, Part 2 6

3.0 TEMPERATURE SURVEY PROGRAM 20

3.1 Part 1 - Cross-Section Temperature Survey 203.2 Part 2 - Inboard Splice Joint Temperature

Survey 243.3 Equivalent Input Temperatures for Unsym-

metrically Heated !nboarC Splice Joints 29

4.0 ANALYTICAL-EXPERIlNTAL COMPARISON SJDIES 32

4.1 Load-Deformation Curves, Part 1 324.2 Available Ultimate Strength and Critical

Station Selection, Part 1 414.3 Analytical-Experimental Deflection Com-

parisons, Part 1 484.4 Deflection Rate Study, Part 1 544.5 Material Properties investigation 604.6 Analytical Results for Plate Element and

Fastener Load Distributions at Ultimate Load,Part 2 62

4.7 Thermal Moment Restraint Study, Part 2 714.7.1 Experimental Evaluation of Thermal

Moent Restraint 754.7.2 Thermal Moment Evaluation and

Analysis Input 78

iii

5.0 APPLIED LOAD RATIO SJMMARY AND COMPARISON 81

5,1 Part 1 - Cross-Section Summary 815.2 Part 2 - Inboard Splice Joint Summary 875.3 Approximate Applied Load Ratios Using

Simplified Procedures 89

6.0 EW(PLE PROCEDURES FOR ELEVATED IEMPERATURESIRENGTH VERIFICATION 92

6.1 Example Procedure to Illustrate Cross-SectionFailure Prediction at Elevated TemperatureFrom Room Temperature Test Results 92

6.2 Example Procedure to Illustrate Splice JointFailure Prediction at Elevated TemperatureFrom Room Temperature Test Results 94

7.0 CONCLUSIONS 967.1 Conclusions on Part 1 96

7.2 Conclusions on Part 2 96

8.0 RECOMMENDATIONS 98

APPENDIX I - PHOTOGRAPHS OF HCRIZONTAL 99STABILIZER FAILURES

iv

TABLES AND ILLUSTRATIONS

PAGE

Table 2.1 Typical Material Properties Summary, Part 1, TestNo. 6 (425oF. Unsymmetrical), Elastic Axis Sta-tion 102 5

Table 2.2 M ment in Static Main Box Skin Elements Between1 and 10

Table 2.. Moment I Static din Beam Fitting Elements Be-tween U6 and U1

Table 2.4 Typical Fastener Allowables, Main Box Skin,4250F. Symmetrical (Test No. 10) 16

Table 2.5 Typical Fastener Allowables, Steel Splice Plate,4250F. Symmetrical (Test No. 10) 17

Table 2.6 Typical Fastener Allowables, Interbeams, 4250F.Symmetrical (Test No. 10) 17

Table 2.7 -iypical Fastener Allowables, Main Beam Fitting4250F. Symmetrical (Test No. 10) 17

Table 2.8 Plate Elemsnt Properties Main Box Skin, 4250F.Symmeatrical (Test No. 10 18

Table 2.9 Plate Element Properties, Steel Splice Plate,

4250 F. Symmetrical (Test No. 10) 19

Table 2.10 Plate Element Properties interbeams, 4250F.Symmetrical (Test No. I0) 19

Table 2.11 Plate Element Properties, Main Beam Fitting,4250F. Symmetrical (Test No. 10) 19

Figurf. 3.1 Main Box Spar External Surface Temperatures,Part 1, Cross-Section Analysis (Test No. 6) 22

Figure 3.2 Upper and Lower Skin Temperatures, Part 1,2Cross-Section Analysis(Test No. 6) 22

Figure 3.3 Upper Spar Flange Temperature 23

v

Figure 3.4 Lower Spar Flonge Temperature 23

Figure 3.5 Spar Web Temperature 23

Figure 3.6 Typical Spanwise Temperature DistributicnCurve, Part 2, Inboard Splice Joint Study4250F. Unsymmetrical Condition 26

Figure 3.7 Main Beam Fitting Flange ard Splice PlateTemperatures 27

Figure 3.8 Forward and Aft Interbeam Temperatures 28

Figure 4.1 Cross-Section Geometry, Part 1 Elastic AxisStation Structural Element Locations 35

Figure 4.2 Load-Deformation Curve Comparison With StaticTest Results, Room Temperature 36

Figure 4.3 Load-Deformation Curve Comparison With StaticTest Results, 2500F. Symmetrical 37

Figure 4.4 Load-Defoemation Curve Comparison With StaticTest Results, 2500F. Unsymmetrical 38

Figure 4.5 Load-Deformation Curve Comparison With StaticTest Results, 425 0F. Symmetrical 39

Figure 4.6 Load-Deformation Curve Comparison With StaticTest Results, 4250F. Unsymmetrical 40

Figure 4.7 Allowable Bending Moment vs. Span, RoomTemperature 43

Figure 4.8 Allowable Bending Moment vs. Span, 2500F.Symmetrical 4

Figure 4.9 Allowable Bending Moment vs. Span, 2500F.,45

Figure 4.10 Allowable Bending Moment vs. Span, 4250F.Symmetrical 46

Figure 4.11 Allowable Bending Moment vs. Span, 4250F.Unsymmetrical 47

vi

PAGEFigure 4.12 Comparisons of Test and Calculated De-

flections - Part i 49

Figure 4.13 Comparisons of Test and Calcul&ted De-

flections - Part 1 50

Figure 4.14 Variation of Yield Stress With Temperature 52

Figure 4.15 Stabilizer Tip Deflection vs. Temperature 53

Figure 4.16 Analytical Deflection Rate Data 55

Figure 4.17 Experimental Deflection Rate Data 55

Figure 4.18 Variation in Inelastic Strain versus ElasticAxis Station for Various Percentages ofFailure Load (4250 F. Symmetrical Test, LampsOn at Failure) 57

Figure 4.19 Variation in Inelastic Strain versus ElasticAxis Station for Various Percentages ofFailure Load (425 0F. Symmetrical Test, LampsOff at Failure) 57

Table 4.1 Failure Station Summary, Part 1 58

Table 4.2 Tensile Coupon - Predicted Yield StressSummary 60

Figure 4.20 Inboard Splice Joint Geometry 64

Figure 4.21 Inboard Splice Joint and Main Beam Cross-Section 65

Figure 4.22 Fastener Shear Loads and Plate Element Loadsat Failure, Inboard Splice Joint, Part 2,Room Temperature (Test No. 7 and Pilot Study) 66

Figure 4.23 Fastener Shear Loads and Plate Element Loadsat Failure, Inboard Splice Joint, Part 2,250 0F. Symmetrical (Test No. 9) 67

Figure 4.24 Fastener Shear Loads and Plate Element Loadsat Failure, Inboard Sp~lice Joint, Part 2,2500F. Unsymmetrical (Test No. 1), Restrainedin-Bending For Thermal Moment, (Test No. 11) 68

vii

PAGEFigure 4.25 Fastener Shear Loads and Plate Element Loads

at Failure; Inboard Splice Joint, Part 2,4250F. Symmetrical (Test No. 10) 69

Figure 4.26 Fastener Shear Loads and Plate Element Loadsat Failure, Inboard Splice Joint, Part 2,4250F. Unsymmetrical (Test No. 12) Restrained-in-Bending For Thermal Moment 70

Figure 4.27 Sketch of Thermal Moment Restraint Condition 74

Figure 4.28 Chordwise Strain Plots of Experimental DataFor Thermal Moment Restraint Investigation,Inboard Splice Joint Study, Part 2 77

Figure 4.29 Simplified Splice Joint Cross-Section 78

Table 4.3 Plate Element Axial Loads in Static Main BeamFitting Structure Due to Restrained-in-BendingThermal Moment Effects 80

Table 5.1 Cross-Section Failure and Applied Load RatioComparison, Part 1 84

Table 5.2 Per Cent Deviatior of Analysis Results fromTest Ultimate Load, Part 1 85

Table 5.3 Bending Moment Comparison at Test FailureStation, Part 1 86

Table 5.4 Summary of Results, Horizontal Stabilizer In-board Splice Joint Study, Part 2 88

Table 5.5 Applied Load Ratios by Simplified Method,Part 1 89

Figure 5.1 Variation of App2ied Load Ratio With StationFor Simplified Method 90

Table 6.1 Test-AnaWysis Summary for Example Procedure,Part 1 93

Table 6.2 Test-Analysis Summary for Example Procedure,Part 2 94

Figure AI.1 Horizontal Stabilizer Pilot Study a; RoomTemperature; Closeup of Compression SideAfter Failure 10

viii

Figure AI.2 Horizontal Stabilizer Test 1; Room TemperatureCloseup of Compression Side After Failure 102

Figure AI.3 Horizontal Stabilizer Test 2; 2500F. Sym-metrical; Closeup of Compression SurfaceAfter Failure 101

Figure AI.4 Horizontal Stabilizer Test 4; 4250 F. Sym-metrical; Closeup of Compression SurfaceAfter Failure 101

Figure 1.1.5 Horizontal Stabilizer Test 5; 250 ° F. Unsym-metrical; Closeup of Compression Surface AfterFailure 102

Figure AI.6 Horizontal Stabilizer Test 6; 4250 F. Unsym-metrical; Closeup of Compression SurfaceAfter Failure 102

Figure AI.7 Horizontal Stabilizer Test 7; Pilot Study b;Room Temperature; Closeup Showing Cracks inMain Beam and Forward Spar 103

Figure AI.8 Horizontal Stabilizer Test 9; 2500 F. Sym-metrical; Closeup of Tension Surface AfterFastener Failures 103

Figure AI.9 Horizontal Stabilizer Test 9; 2500 F. Sym-metrical; Closeup of Failure of InternalStructure 104

Figure AI.10 Horizontal Stabilizer Test 9; 2500 F. Sym-metrical; Closeup of Tension Failure ofMain Beam Flanges 104

Figure AI.11 Horizontal Stabilizer Test 10; 4250F. Sym-metrical; Closeup of Failure of InternalStructure 105

Figure AI.12 Horizontal Stabilizer Test 1I; 2500F. Un-symmetrical; Inboard Portion of Tz:AsionSurface After Failure 105

Figure AI.13 Horizontal Stabilizer Test 11; 250°F. Un-symmetrical; Closeup of Failure Main BeamTension Flanges 106

ix

Figure AI.14 Horizontal Stabilizer Test 12; 1250F. Un-symmetrical; Closeup of Failure of MainBeam Tension Flanges 106

Ix

I,

1f

tL

I(xt

X{

= flange element area, in. 2

An = element area, in.2

An = skin element area, in.2

Cf,Cr = thermal moment distribution factors for front and rear beamaraas, respectively

(I = maximum value of Yn, inches

d = depth of beam or cross-section, inches

dmax = maximum beam depth, inches

E.A. = elastic axis

Ef = flange element modulus of elasticity, psi

En = element modulus of elasticity, psi

Es = skin element modulus of elasticity, psi

ec - incremental applied constant axial strain, in./in.

ecr = critical buckling strain, in./in.

ef = elastic strain in flange element, in./in.

es = elastic strain in skin element, in./in.

Fcy = compressive yield stress, psi

Fty = tensile yield stress, psi

Ftu = tensile ultimate stress, psi

he = equivalent beam height, inches

I = cross-section ......nt of. .ai- , LA4

K joint spring constant representing the elastic slope offastener load-deformation curve, lbs./in.

Kapx = applied moment rotational strain about x axis

xi

KT = thermal moment rotation& strain

KTx = thermal moment rotational strain about x axis

px = inelastic moment rotational stra-in about x axis

L' = bending moment at elastic axis station, in.-lbs.

M = bending moment, in.-lbs.

r = critical ultimate bending moment, in.-lbs.

M O = reference ultimate bending moment at elastic axis station75.625, in.-lbs.

MT = elastic thermal moment, in.-lbs.

P = axial load, lbs.

PC = couple load, lbs.

F y = yield load, lbs.

Rap = applied load ratio

R.T. = room temperature

Scr = critical ultimate shear load in fastener, lbs.

Tf = flange temperature change from datum,, OF.

Tft = lower flange temperature, OF.

Tfu = upper flange temperature, OF.

Tf = lower surface temperature, OF.

Ts = skin temperature change from datum, OF.

T" = upper surface temperature, OF.

Tw = web temperature, OF.

Vcr = critical ultimate applied shear load, lbs.

VE.T. = ultimate shear load at elevated temperature at outboardloading former, lbs.

xii

(VR.T.)Ref. = reference ultimate shear load at room temperature at

outboard loading former, lbs.

YH' = elastic axis station, inches

Yn = distance from neutral axis to centroid of element "n", inches

ap = applied

c = constant; couple

cr = critical

E.T. = elevated temperature

e = effective; equivalent

f = flange; front

= lower

n = element no.

o = static applied

p = inelastic

R.T. = room temperature

r = rear

s = skin

T = thermal

u = upper

w = web

y = yield

xlii

Greek Symbols

af = coefficient of thermal expansion for skin element,in./in./OF.

cis = coefficient of thermal expansion for flange element,in./in./OF.

= incremental

6 = deflection, inches

iI

xiv

(a) Gehring, R.W. and Maines, C.H., U.S. Naval Air Engineering CenterReport No. NAEC-ASL-1094, "Application of Applied Load RatioStatic Test Simulation Techniques to Full-Scale Structures; Vol-ume I, Methods of Analysis and Digital Computer Programs", of 31December 1965.

(b) Gehring, R.W. and Lumm, J.A., U.S. Naval Air Engineering CenterReport No. NAEC-ASL-1094, "Application of Applied Load RatioStatic Test Siriilation Techniques to Full-Scale Structures; Vol-ume II, Material Properties Studies and Evaluation", of January1966.

(c) Alter, Elise, U.S. Naval Air Engineering Center Report No. NAEC-ASL-1094, "Application of Applied Load Ratio Static Test Simula-tion Techniques to Full-Scale Strucatres; Volume III, Experi-mental Program", of January 1966.

(d) Gehring, R.W. and Engle, R.L., North American Aviation, Inc., Re-port No. NA62H-973, "A Feasibility Study of Applied Load Ratiosto Simulate Elevated Temperature Static Tests at Room Tempera-ture", of May 1963. Prepared under Department of the Navy,Bureau of Naval Weapons Contract NOw6l-0963-d.

(e) MIL-HDBK-5, Strength of Metal Aircraft Elements, dated August1962, through Change Notice No. 5, 15 June 1965.

(f) Allen, H.F., WADC Techr-ical Report 56-227, "An Experimental StudyRelating to the Prediction of Elevated-Temperature

Structural

Behavior From the Results of Tests at Room Temperature", of June---- 1956.

(g) Gatewood, B.E., Thermal Stres2es, McGraw-Hill Book Company, Inc.,1957.

xv

1.0 INTRODUCTION

A previous analytical-experimental program performed under Department ofthe Navy, Bureau of Naval Weapons Contract NOw6l-0963-d used simple boxbeam structures loaded in bending or comprsssion to verify the feasibilityof simulating elevated temperature static tests at room temperature. Testand analysis data for room temperature, and for syn-metrically and unsym-metrically heated box beam structure under static loads indicated that theeffects of material properties degradation and thermal stresses could beaccounted for analytically by using strain analysis procedures whichallowed the construction of load-deformation curves for the critical cross-section. The comparison of the curves for room temperature with those ofany other temperature environment provided the applied load ratios foryield load ultimate load and for any other desired value of permanent setat the criical cross-section. Analy+ical-experimental comparisons,usingcritical e.ement strain gage data versus calculated strains, indicatedthat this epproach was feasible.

A verification program was then instituted to perform similar studies onfull-scale aircraft components. The basic method was to perform thestatic strength simulation test at room temperature on an airframe com-ponent designed to operate in an elevated temperature environment. Then,using reliable analytical solutions for strength and deformation analysesbased on discrete element strains and temperature dependent material pro-perties, it is possible to define applied limit, yield, and ultimate loadratios to relate the room temperature test results to arq desired tempera-ture environment. The analytical procedures must be the same for roomtemperature load-deformation definition as for elevated temperature withthe inclusion of the additional temperature terms in the elevated tempera-ture analysis. It is equally essential that the analytical proceduresallow for inelastic stress-strain relationships as well as buckling,temperature-time depender' material properties, and thermal stresses.

Previous reports have defined the basic methods and procedures used forthe analytical studies (Volume I), the definition of material propertiesand Ramberg-Osgood parameters necessary for stress-strail curve descrip-tion (Volume II), and the room-temperature elevated-temperature experi-mental program performed on the horizontal stabilizer assemblies (VolumeIII). This volumA contains specific information on the use of the analy-tical studies for experimental-analytical comparisons and includes sum-maries for plane-strain cross-section and splice joint test simulation,conclusions and recommendations.

m1

2.0 DISCUSSION OF ANALYTICAL PROGRAM

The analytical portion of this study was performed using specificstrain-oriented analysis procedures applicable to elastic and in-elastic structures, and room and elevated temperatures. Load-deformation defines a common basis whereby the room temperature andelevated temperature results can be compared, with the only analy-tical differences being the inclusion of temperature terms in theequations and the temperature-dependent material properties data in-put. In oratr to define the strength, and, therefore the appliedload ratios, the analytical studies were performed using the methodsdescribed in Volume I which allow for

(a) Inelastic stress-strain relationships(b) Buckling of individual elements(c) Thermal stresLes which may be elastic or inelastic(d) Combined thermal and applied stresses on a realistic

strain basis(e) Temperature-time dependent material properties and

recovery of properties after exposure to elevated

temperature.

The methods of Volume I were utilized to define the horizontal stabi-lizer strength at room temperature and in four different elevatedtemperature envirorments in order to establish applied load ratios forroom temperature simulation. Parts 1 and 2 depend upon two differentmethod of analysis: Part 1 is concerned with characteristics of theoutboard cross sections using the plane-strain method of analysis;Part 2 is concerned with the axially-loaded mechanically-fastenedsplice joints, using the virtual work method of analysis.

2.1 Strength and Deformation Analysis Procedures

The general methods used for the analytical investigation are des-cribed in Volurme I, Section 3.2 for the cross-section analysis ofPart. 1, and Section 3.3 for the inboard splice joint analysis ofPart 2. The following sections (2.1.1 and 2.1.2) illustrate withtypical data the input data of major importance to the digital com-puter programs outlined in block diagram form in Sections 4.1 (cross-section analysis) and 4.2 (splice joint analysis) of Volume I. Thedata presented in Sections 2.1.1 and 2.1.2 are primarily the te&,.,era--ture-time dependent material properties, typical calculations anddescriptions of the v. -ous mechaniical properties, and, in Section2.1.2, a description of the applied axial load input to the axiallyloaded splice joint based on an extei ml applied bending moment.

2

2.1.1 Crqss-Section Analysis. part

Using the procedures described in Volume I, the load-deformationcurves and the ultimate loads are calculated for each instrumentedcross-section for each of the temperature environments of the testprogram. The load-deformation curves are plotted for the criticalelement in the cross-section which is defined as that having themaximum permanent set, epsm, where the maximum value is calculatedby

es Z e + ( Kx ) + (D.K )y apsM p px c py

The resulting load-deformation curves and ultimate loads are de-scribed in Sections 4.1 and 5.1 respectively.

The predicted failure station is defined by plotting the ultimatebending moment versus span (elastic axis station) curve using thecalculated ultimate values at the four instrumented cros,-sections,and by plotting on the same set of coordinates an applied bendingmoment line which is tangent to the ultimate strength c'rve. Thepoint of tangency of the two curves defines the critical stationfor ultimate strength. Yield, limit load, or any specified strengthcan be established for the "strength curve". This procedure is ap-plied to this program and is described in detail in Section 4.2.

The cross-section analysis; procedures as defined in Volume I forultimate strength and loal-deformation prediction rely heavily onthe accurate representation and definition of the basic materialproperties in any temperature environment and any exposure time orsoaking period. Therefore, the material properties must be definedat temperature to compute the strength at temperature and at roomtemperature after exposure to temperature to compute the strengthand any permanent set remaining at room temperature after exposureto the elevated temperature environment. The individual ele'ienttemperatures are defined by curves as described in Section 3.0which utilize span-wise plots of the external thermocouples at thetime of failure of the particular horizontal stabilizer and gen-eralized plots of temperature gradient data obta.ned from thetemperature survey stabilizer (Horizontal Stabilizer Test No. 2).The equivalent soaking times at temperature and the element tem-peratu-rp, z . . ro shovn in Table 2.1 along with the yield and ultimateLarson--Miller Parameters for use with the parametric curves ofVolume Ii for material properties definition (Fty aid Ftu) at tem-perature and at room temperature after exposure to the elevatedtemperature environment.. The remaining Ramberg-Osgood parameters,modulus of elasticity (E), and stress-strain curve shape factor (in)

3

are obtained from the' temperature dependent data and curves inVolume II for the specific materials. These properties, and tablesidentical to that shown in Table 2.1, were calculated and definedfor each instrumented cross-section (Elastic Axis Stations 91, 102,113, and 124) for each test horizontal stabilizer (Horizontal Stabi-lizer Tests 1, 3, 4, 5, and 6). Table 2.1 is shown as a typicalillustrative example of material properties input data for ElasticAxis Station 102, Horizontal Stabilizer Test 6, 4250 F. Unsymmetricaltemperature environment. For Test 4, 4250F. Symmetrical, where alamps-off,lamps-on study was performed, it was necessary to definetwo separate sets of material property data for the two differentsets of temperature data.

The cross-section analysis procedures also provide the necessary datain the form of applied, thermal, and inelastic moment rotational strainvalues (Kapx, KTx, and Kpx respectively) as basic input to the de-flection analysis prccedures of Section 2.2.

4

W..

~ .

6FO 0

, 0.

-4

4

oo

4.')

0or c1 WNoc

(V 0 NN660 ~ N cv (N N

e- c

E, >,

1 -0 U 3

CL 0 0 0 S.5

2.1.2 Inboard Splice Joint Analysis. Part 2

In order to perform the analysis for splice joint strength as de-scribed in Volume I and provide input data for the computer programcertain data must be obtained,and/or defined,pertaining to the ap-plied loads and the necessary mechanical properties for both fastenersand plate elements. This section describes the numerical definitionof the applied axial loads for the splice analysis as derived fromthe applied shear loads and bending moments applied to the outboardend of the splice joint. Loads in, the statically determinate struc-ture are shown which apply to all temperature environments except asnoted for the restrained-in-bending thermal moment cases of Test No's.11 and 12 which had unsymmetrical temperature distributions. A typi-cal set of calculations is shown (room temperature environment) forfastener-plate Ramberg-Osgood parameters to describe the load-deflec-tion characteristics of individual fasteners in the analysis. Inaddition, typical sets of fastener properties and plate elementproperties are tabulated for one particular temperature environmentto illustrate the magnitude and units of typical values and the exactdata required for a complete stress analysis of the splice joint.Briefly, the four sections are dO.scribed below.

(a) Load Distribution Asamptions in Inboard Splice Joint

This section includes the assumptions made to obtain theequivalent axial loading of the splice joint for analysisby the methods described in Section 3.3 of Volume I. Thecalculations to define the axial loads are shown and theseare assumed to apply in general to all temperature environ-ments. Tables 2.2 and 2.3 are presented to show the loaddistribution in the static load path of the redundantfastener-plate structure. The cross--section geometry isshown in Figures 4.20 and 4.21.

(b) Ramberg-Osgood Parameters for Fastener-Plate lntersectionsat Room Temperature

Typical ca]culations are shown for the room temperaturecase where the Ramberg-Osgood parameters are defined foreach fastener-pla~e intersection according to fastenerdiameter, plate thickness, and plate material. The cal-culated values shown are those used in the analysis ofest No. 7 and are obtained by using the curves and ma-terial property data for fastener-plate combinations in-- Volume II.

(c) Summary Tables for Fastener Properties

Tables 2.4 through 2.7 provide the necessary mechanical

/

properties to describe fastener deformation, local bearingdeformation, and ultimate strength cut-off viAlues for atypical elevated temperature test (425 0 F. Symmetrical; TestNo. 10). Four separate tables are shown, each for a parti-cular component of the splice joint; i.e., main box skin,splice plate, interbeams, and main beam fitting.

(d) Summary Tables for Plate Element Properties

Tables 2.8 through 2.11 provide the necessary temperature-dependent material properties for all plate elements in thesplice joint using Test No. 10 (425 0 F. Symmetrical) as atypical example. These properties describe the Ramberg-Osgood parameters for plate element stress-strain curverepresentation in a similar manner to that for the axially-loaded cross-section elements of Part 1. The temperaturesshown are the plate element temperatures on the tension sideof the splice at the time of failure and were obtained fromthermocouple readings and the procedure and curves of Sw,&tion3.2. In these tables the Larson-Miller values represent theeffects of 1/2 hour exposure time for the temperatures shownsince the test temperature exposure times did not vary suffi-ciently from 1/2 hour to warrant more detailed approximationsof elevated temperature properties.

The calculations for items (a) and (b) immediately follow this sectionon pages 8 through 15. The fastener property tables of item (c) areon pages 16 and 17, and the plate element property tables of item (d)are on pagesl8 and19.

7

--_

(a) Load Distribution Assumptions In Inboard Splice Joint

The oending moment is applied to the splice joint area by concentratedloads of equal magnitude at Elastic Axis Stations 80 and 95.

Froxm the original static test at room temperature the critical ulti-mate bending moment at Elastic Axis Station 49.56 is

,rC 1.13 x 106 in.-Lb.

TNo loading formers were used for more accurate matching of designutimate moment curve slope at root and to preclude a cross-sectionSf'ailre just. outboard of the splice area.

"he critical applied shear load at each loading former, Vcr, is definedby

Vcr (95-49.56) + Vcr (80-49.56) = 1,130,000

45.44 Vcr + 30.44 Vcr = 1,130,0

75.88 Vcr = 1,130,000

Vcr = 14,900# ultimate

'he cending moment at the outboard line of fasteners (Elastic AxisStation 75.625) for Vc. -- 14,900# is.

M = 14900 (95-75.625) + 14900 (-0-75.625)= 14900 (19.375) + 14900 (4.375)= 288,700 + 65,1> "

353,850 in.-lbs. at Station 75.625

The reference ultimate loads for room temperature at Station 75.625 are

defined as Mo and Vo where

S353,S5t "n.-bs. and Vo = 29,500 lbs.

I Load' D3rlbutlon Assumo .ions

Station 75.6,5 is represented by fastener l!ne I- ( to(,b7:r 'WI 13ac disr'iution between sKin arid interoeam flaes.

For a unit moment Mo = 100,000 in.-lbs. (Refer to Volume I) theaxial load distribution is

E8

Main Box Skin Load = 39,150#, Interbeam Flange Load = 1914#

For Mo = 353,850 in.-lbs. the load distribution is,

Main Box Skin Load = 138,600#, Interbeam Flange Load = 6780#

(2) Applied axial loads of 138,600# and 6780# are constant betweenStations 75.625 and 65.76 in the statically determinate plateelements.

(3) The shear load of 29,800# is assumed to be transferred inboard as acouple load in the main boN skins only, producing variable loadsin skin elements between 1 and (la . The shear load is tentr 3sferred to the static lastener-Segments represented by (1 ,ANI)and (5-) at Elastic Axis Station 65.76. O

This simplifying assumption is made because of the relativelysmall areas of the interbeam flanges and the splice plate plusthe fact that the majority of the 29,800# shear load is in thefront and rear beams. Also it may be seen from the axial loaddistribution at Station 75.625 that the maximum error in thisassumption could be only about 4% at most.

Table 2.2 on page 10 gives the average incremental loads in themain box skin plate elements between G and Q.For the thermal moment restraint conditions considered to bepresent in Test No's. 11 and 12 the couple loads shown on page79 must be added, where applicable, to the values in Tables 2.2and 2.3 for applied load distribution in the statically deter-minate structure.

9

4) )

GLA C~ kt\ 0 0 C) C)0) ' UN t o H O W

C) N) LA*

0 C)~ 4 ' ' O 0

V

a- O ) m m) 0A 0 A-0 Z UN\ 0- 0U 0 "0

U'\ or a'\ 'cN oo "o.H .- 4 .4 C4 N 4~ t 0a

0 L

U-C\ C -. cq LAl UNC4 '.dL N a k-I N ~ \L0 0 Mt \- '

x~ 0, 0\t 0\' G\ 0t L A A 0 '0 '0

4-))

o o 0 ) 0 0

0 0 5 -4 0 0 0 '0 0t 0 ' m' a' 0 0 m) C

V U.~ ('2 N N' N -(N N E' (\C- "

ci : ) C ) 0 000 0~-' 0 0 C C 0 0 , C)

ci ~0 ~ IT 00 . ~ C ~ C (

4-)0 U) G\ LA C) H ~ ~ .0 0

CC c

o~0 L L.C)Ct-dU 0 C)r-\

0) t3 .010

00 0 81' 0C 0 ~ 0\

H A N4 U ' 0U % r.- 0 a ' H CVtA~~ H HC w H 0ol G

' 0'0 0'. 0 0 0 0 0 0Y -4 LC 00 0'4.3 F(' i' R 0 0 '0 '0 0 '0 '0 '0 0i

z m0

4)' 0

H. to o4 oC . - a N N~- ..H m~ 4-\H H U W

) '0 4-

4

.4-;a 0 0 0 0 0 0 0 lCd Z +-I %r\ 0 0 0 N '0 0 0~ 0= 0

4-D tri Z. '0l ' ' r r 0 0' C

a~~~~~~. E Dt o U\0G J

d) 0 O- H Hbi N

4-D - 0-

"-4

4, 0

-4 '0 (4 0 N - '

H- H- H H- H- 0.0

0U)

0

0V-H ILt\ t10 l-% tU 4- k L r to

4-- CC ~ H 0 c7' m0 '0 '' -4 N OEn '0 ' 0 ' ' . 0 10 ' 0 1 1 (

0 rf C4 rito, 11

II 1 0 D .0 \D tr l\ t, N V r c

(b) Ramber-Ouood Parameters For Fastener-Plate Intersectionsat Room Temperatur

The following set of calculations is provided to illustrate the ap-

proximate procedure used to obtain fastener-plate Ramberg-Osgood

parameters for the individual fastener-plate intersections withinthe splice joint. P , n, and K represent the yield load per fastener,

the shape factor for the Ramberg-Osgood representation of the

fastener load-deflection curve, and the initial elastic slope ofthe curve, respectively.

Alowable bearing yield stress reference value is 50,000 psi. for

critical stainless steel sleeve as shown in Volume II. Use Figure

5.3 of Volume II for bearing yield factor.

(1) For J" Hi-Shear Blind Bolts iq .290" alum. alloy- app1's to

fastener-plate intersectionsl, Q ,® , ()and 0

Bearing Area = .25(.29) = .0725 in.2

Py = 1.02(50,000)(.0725) = 3695 lbs./fastener, n = 2.762

K = 1.02(451,750) = 460,000, Scr = 9750# (Two fasteners)

(2) For 3/8" Hi-Shear Blind Bolts in .29 " alum alloy applies to

fa nergate intersections I , ,

Bearing Area .)75(.29) = .10875 in.2

Py = 1.175(50000)(.10875) = 6380 lbs./fastener, n = 2.762

K = 1.175(451,750) = 530,OOC, Scr = 21,300# (Two fasteners)

(3) For 7/16" Hi-Shear Blind Bolts in .2201all. ally; apples to

fastener-plate jtersections 8 ,,(20, (2) 2

Bearing Area .4375(.29) .127 in.2

Py = 1.255(50000)(.127) = 7960 lbs./fastener, n = 4.55

K = 1.255(451,750) = 566,500, Scr = 35,150#(Two fasteners)

12

(4) For J" i-Shear Blind Bolts in .090" steel sPl1e plate; ap-p so fostenrplate intersections S,(

Bearing Area = .25(.09) = .0225 in. 2

P = 0.8(50,000)(.0225) = 900 lbs./fastener, n = 2.762y

K = O.8(451,750) = 361,000, Sor : 9750#(Two fasteners)

(5) For 7/16" Hi-Shear Blind Bolts in .090" steel solice ate; ap-plies to fastener-.plate intersections 3

Bearing Area = .4375(.09) = .0394 in.2

Py = 0.87(50,000)(.0394) = 1714 lbs./fastener, n : 4.55

K = 0.87(451,750) = 392,500, Scr = 35,150# (Two fasteners)

(6) For +0 Hi-Shear Blind Bolts in 0.152", un, inter-be ;9; ap-plies to fastener-plate intersections4 , , 49

Bearing Area = .25(.152) = .038 in. 2

Py = 0.865(50,000)(.038) ; 1643 lbs./fastener, n 2.762

K = 0.865(451,750) = 390,500, Scr = 9750# (Two fasteners)

(7) For 7/1611 hi-Shear Blind Bolts in 0.125" steel man beam flange;applies to fastener-plate intersections (D,

Bearing Area = .4375(.125) .0547 in. 2

P = 0.935(50,000)(.0547) = 2555 lbs./fastener, n = 4.55-y

K 0.935(451,750) = 422,000, Sr = 35,150" (Two Fasteners)

(8) For 3/8" Hi-Shear Blind Bolts in 0.187" keel main beam flange;applies to fastener-plate intersection (5 only.

2Bearing Area = 0.375(.187) = .0701 in.

P = 1.01(50,000)(.0701) = 3540 lbs./fastener, n = 2.762y

K = 1.01(451,750) = 455,500, Scr = 21,300# (Two fasteners)

13

(9) For 3/8" Hi-Shear Blind Bolts in 0.143" steel main beam flange;

applies to fastener-plate intersection 5k only.

Bearing Area = 0.375(.143) = .0536 in. 2

Py 0.93(50,000)(.0536) = 2493 lbs./fastener, n = 2.762

K = 0.93(451,750) = 420,000, Scr = 21,300# (Two fasteners)

(10) For 3/8" Hi-Shear Blind Bolts in 0.155" steel main beam flange;

applies to fastener-plate intersection a only.

Bearing Area = 0.375(.155) = .0581 in.2

P= 0.95(50,000)(.0581) = 2760 lbs./fastener, n = 2.762

K 0.95(451,750) = 429,000, Scr 21,300# (Two fasteners)

(11) For 3/8" Hi-Shear Blind Bolts in 0.168", stesl main beam flange;

applies to fastener-plate intersection 69 only.

Bearing Area = 0.375(.168) = .063 in.2

Py = 0.975(50,000)(.062) = 3070 lbs./fastener, n = 2.762

K = 0.975(451,750) = 440,500, Scr = 21,300# (Two fasteners)

(12) For 3/8" hi-Shear Blind Bolts in 0.182" steel main beam flange;

applies o fastener-plate intersection only.

Bearirg Area = O.375(.182) = .06825 in.2

Py 1.0(50,000)(.06825) = 3415 lbs./fastener, n = 2.762

K =1.0(451,750) = 451,750, Scr = 21,300# (Two fasteners)

(13) For 3/8" Hi-Shear Blind Boli s in 0.195" steel main beam flange;applies to fastener-plate intersection S only.

Bearing Area = 0.375(0.195) = .0731 in.2

P = 1.02(50,000)(.0731) = 3725 lbs./fastener, n = 2.762y

K = 1.02(451,750) 460,000, scr = 21,300# (wo fasteners)

14+

(14) For 3/8" Hi-Shear Blind Bolts in 0.207" teel main beam flange;

applies to fastener-plate intersection F6 only.

Bearing Area = 0.375(.207) = .0777 in.2

Py = 1.04(50,000)(.0777) = 4035 lbs./fastener, n = 2.762

K = 1.04(451,750) = 470,000, Scr = 21,300# (Two fasteners)

(15) For 7/16" Hi-Shear Blind Bolts in 0.224" steel main beam flange;applies to fastener-plate intersection S only.

Bearing Area = 0.4375(.224) = .0979 In.2

Py = 1.127(50,000)(.0979) = 5520 lbs./fastener, n = 4.55

K = 1.127(451,750) = 509,000, Scr = 35,150# (Two fasteners)

(16) For 7/16" Hi-Shear Blind Bolts in 0.230"1 eel mn beam flange;applies to fastener-plate intersec ion6s , &

,and & .

Bearing Area = .23(.4375) .1007 in.2

Py = 1.14(50,000)(.1007) = 5740 lbs./fastener, n = 4.55

K = 1.14(451,750) = 515,COO, Scr = 35,150# (Two fasteners)

15

L 14

410 ~ ~ 1 0 0 .4-.4000 6S4 .- .9 .t..3.

.0.'M 0. 0000000000ooo0000o

10 A (1 4.4.4 N N N N

'-4 toWt 0" 0 ,(1 O

4,

0Mt ' ' (Jf '.0~~r 0.WN33t0 0~'u

't

-C .4. (~-4 1 -4 - 4NN N N

1 9AA

'4" QQ

do Q 0 0. C;-- OH. .4.0 0 (1 - .- .4 -4

0'~ 0000 00 0 0

LP

0. AD

04 10 10~ IF~ I~ 1 )

C) 4-) 11.1-

-4 4 -4 03 '-v4

m A,.- - - :: )Zc N VC 0 N4!717

00 00

04-'-'

I .E-4c

N4 0

ID ID oc "': - C F14C .. ev,-.( "I m M VN 4% N W

>-4 rZ :-

4). 0 V 4) 0cT C

IC)E- F,,I0 0 ii 0

4 -' ) P Q

CO I .

N v1 e I) .0)ZI

I I ~17

N

S4~

'- N > -eq00 0

o o

L

4)

0. 110 00

"I

kf

to ! "

04)C (71, -

.- .. . . . . . .

4) 4-

0. W1 le\ cy O' r, i" ,0

E-oo1 4 .

02 0

110

0 0 N' N__ _"1

4 . '-4" 1-4 -

144 M4

*F

N V

3 s? c

14 "1

0 1 0 1 4 1 4 t

N' 'N . N N N NN41

In4 0 C)0

41

0-00 to

4-0 ) 4.>

0-4 4____ r00

4. 41 jo

----- 4-D1 0 - - ----

.1 ~ ~ '3 441x4 0'

4.1 Cd C. -

r-j*r- U- 4 3 0

'43.- Q1*H - 4-44

0)00 00

-3 fl

ca 4 enE-1 'NC3 N- C~

It tt -t.. m

00

! -It'

19

*. _ERQATURE SURVE PROGRA

3.1 ?artr I - Cross-Section 2emDerature ISurvey

One horizontal stabilizer Test 2 was used for temperature and tem-perature distribution studies only to provide a means of approximatingthe temperatures on internal structurel elements of the test stabilizers.Iolume III shows the location of thermocouples on the temperature surveystabilizer and also on the test stabilizers. Note that internal thermo-

couples are provided on the temperature survey stabilizer, but not onthose stabilizers which are tested to failure by externally applied loads.The construction of the stabilizers does not provide ieady access to thespars and spar flanges; therefore it was necessary to cut access holesin the main box skin for internal thermocouple installation. Flushpatches were tolerable from a temperature distribution standpoint butstructural repairs in the main box areas would be intolerable for com-parative failure analysis of the test stabilizers.

It 3hould also be noted that the external thermocouples on the tempera- .ture survey stabilizer and the test stabilizers are located in identical1ocations to allow relationships to be established between external andinternal temperatures. These relationships could best be establishedusing the temperature survey stabilizer with both internal and externalthermocouples and performing temperature studies under similar tempera-ture environments to the steady-state operating conditions during thestatic load tests. The four operating conditions were nominally asfollows:

(a) 2501F. control temperature on upper and lower skins; sym-metrically heated (upper and lower surfaces).

(b) 2500F. control temperature on upper skin; heated uppersurface and air cooled lower surface.

(c) 4250F. control temperature on upper and lower skins; sym-metrically heated (upper and lower surfaces).

(d) 4250F. control temperature on upper skin; heated uppersurface and air cooled lower surface.

Considering that joint conductivity may be a significant item in the

overall heat conductance problem the experimental temperature datafrom ,e temperature survey specimen are considere to be more reli-able ',nan t.c),-e obtained b- analytical means.

A method was devised for using thermocouple data from the temperaturestudie2 performed by NAEC-ASL (Volume III)for realistic internal tem-perature predictions f )r all test stabilize-s by relating the upperflange, spar weo, and lbwer flange temperstures to the maximum tempera-ture ( OaX. ox s3rin) over the spar flange and to the temperature

20

i

IA

gradient through the depth of the spar. This gradient is measuredby external thermocouples in the upper and lower skins over the sparflanges at the stations indicated -n Figure 3.1. The curves inFigures 3.3, 3.4, and 3.5 were derived on this basis and apply tothe four main spars and any elastic axis station by interpolatingbetween the two curves shown. Note that the inboard and outboardstations (Elastic Axis Station 90 aad Elastic Axis Station 125) tendto produce separate groups of data (Tu/Tt s 1.0) and slightly differ-ent curves for the range of Tu/Tj . These cirves are considered tobe sufficiently accurate for determining internal structural elementtemperatures and the corresponding element material properties. Thesignificance of these curves is that they require only a knowledgeof externtd temperatures, as obtained from the static test stabilizersto define the internal structural element temperatures.

Figures 3.1 and 3.2 show a typical example (425 0 F. Unsymmetrical,Test No. 6) of spanwise temperature variations which are used in con-junction with Figures 3.3, 3.4, and 3.5 to define internal elementtemperatures.

The main box skin and spar temperatures and temperature distributionsin Figures 3.1, 3.2, 3.3, 3.4, and 3.5 are designated by the followingsymbols.

Tu = upper surface (external) temperature in OF. measured bythermocouples

T = lower surface (external) temperature in OF. measured bythermocouples

Tfu = upper spar flange temperature in OF. and applies to frontspar, rear spar, and interbeams

T fy = lower spar flange temperature in OF. and applies to frontspar, rear spar, and interbeams

Tw = average spar web temperature in OF. and applies to frontspar, rear spar, and interbeams

21

" /

4X(

' 300CD

00 '-

il --Front Spar

-10._ Forward nterbeam

Aft Interbeami -- - --' -- Rear SpLr

80 90 100 110 120 130

Elastic Axis Sta. (Y'1 ), inchesFigure 3.1 Main Box Spar External Surface Temperatures

_Part 1. Cross-Section Anals:.(Test No. 6)500

40C

ili

C

C)T

,, 3003

"o -(---- Forward Main Fox ','kin

' C '

,- -- [DElastic Axis Nain Box Skin

,,------ Al t Ma.in Box Ikin" ,Flrward Magnesium 5kin

-- < Aft -.'agnesium .;kinq 0 9() 100 110 120 130

-. lastic A xis 3ta. (Y'ji), inchesFigurwe ".p P'r and Lower ikin Temp-ratures,

Part 1 , Cros5-3ection nalvsis , est lo. 6)

22

f.10 C) I'

-- _ _

Cf'

-~f i

01

.fI.I C\ 42q

H. H

L41.H-1 H o 0 H H- 0 0 0; 8 H H 8 (3 ;

4C.

23

3.2 Inboard Splice Joint Termratureurev

Internal and external thermocouples in the splice joint area in Test8 provided data for each of the four elevated temperature environ-

ments whereby the internal temperature distributions in the main-box-skin,main-beam-fitting splice joint area could be approximated for thestatic test components. As noted in Section 3.1 for the cross-sectionsurvey, the external thermocouples on the temperature survey stabilizerand the static test stabilizers are located identically to allow reli-able temperature relationships to be established. The relationshipsbetween external temperatures and internal temperatures were estab-lished by the temperature surveys of test 8 in which temperaturestudies were made with the required temperature environments of theanticipated steady-state elevated temperature test conditions. The fourelevated temperature conditions were basically those described in Items(a), (b), (c), and (d) of Section 3.1 with a fundamental differencebeing that the heated area was generally restricted to the area boundedT by the ribs at the inboard and outboard ends of the main beam fitting,and between te front and rear beams. The difference in material between.he main beam fitting (steel) and the main box skins (aluminum alloy)together with the mechanipally-fastened joint makes a theoretical tem-perature distribution highLy questionable and data from a heated sta-bilizer is considered more reliable.

Four temperature survey tests were performed for the anticipatedstatic test temperature environments by NAEC-ASL as reported in VolumeIII. As in the cross-section temperature survey tests (Section 3.1),the internal temperature predictions for all test stabilizers were madeby relating the upper and lower main beam fitting flanges and the mainbeam web temperatures to the temperature gradient between the upper andlower main box skins. The gradient is defined by the thermocouples inthe upper and lower skins directly ovor the main beam fitting flange atthe 3tations shown in Figure 3.6. The non-dimensional procedure resultsin the general sets of curves for internal element temperature pre-diction shown in Figure 3.7 for the upper and lower main beam fittingflanges and splice plate and in Figure 3.8 for the forward and aft inter-beams. Figures 3.7 and 3.8 were defined by thermocouple data from thefour temperature survey tests and are used in conjunction with the span-wise temperature distribution curves for each test as shown in thetypical example in Figure 3.6. For the specified temperature conditionin Figure 3.6 the proper value of Tu/T2 is selected at any desired sta-tion and the values of flange and web temnprntbi-ma 2_"e d-breading on the appropriate curve of Figure 3.7 or 3.8 for the particularvalue of Tu/T1 .

24

These curves are sufficiently accurate for temperature definition,both for temperature data input to the analysis as well as thedefinition of temperature-time exposure histories for definingmaterial properties.

The definition of the terms T, T, , TfU, Tfj , and Tw in Figures3.6, 3.7, and 3.8 is identical to that shown in Section 3.1.

25

C'J~1 0V v- 1

IH-

0

I 4- O '

0c

4--

C)'

CC)

.,Jo ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . 0S)IITjdujIq XP r~ a.- u ai

0 0

i. E__ jv (d~i

iDc41f __C r/ _ _

r') 1 '*1 0~ w

0 cn a)

0 0

ciCcI 0) .O H

010

HC:4

.- c 00

H C)

__ __C-4 1

271_ _ _

- 0 -- ,

_ -4

4 4.) C

41l) Wil ,

0r-' ,'

-0 ..... o)(D o

0)0

: r __ _ o _- _ -- __-.

0 pE -

0x

0)0

4- r o

__I__ , __

OD 20)efIIct4 rz-4

0 0

0

.0 I

_ _ 1_ _ .. _ _ __28

iII

3.3 Equivalent Input Temperatures for Unsyvrnetrically Heated InboardSplice Joints

For the two unsymmetrically heated inboard splice Joints (Test No's.11 and 12, 2501F. unsimmetrical and 425°F. unsymmetrical, respectively)the temperature distributions and restraint corlitions are such thatthe thermal strains computed by cross--section bending procedures andthose computed by axial procedures only,using the actual plate elementtemperatures, are not the same. Therefore, for axial load equivalenceto provide the same thermal strains as the unrestrained-in-bendingcondition, it is necessary to compute equivalent main beam flange ormain box skin temperatures to use in the analysis to replace the actualtemperature when calculating the thermal strains. In this case theskin temperature was taken from the elevated temperatlre surveys anda fictitious flange temperature for equivalent thermal strains vascalculated.

Expressions for the axial strains e s and ef, which are known fromcalculations by the cross-section thermal strain analysis, are

es =-casT + eT 7 33les =C S S+ e axial thermal strain (3.3,1)

ef =-cfTf + eT equations

where e : EsAs asTs -t EfAf fTf (3.3.2)

T = EsAs + EfAf

quating eT values and solving for flange temperature, Tf ,~1

a f (Oss + e s - e ) (333)

where Tf and Ts represent a temperature change from datum for theflange and skin elemei.ts, respectively, and es and ef are the tnermalstrains obtained from the cross-section analysis for a T step. Thetemperature change, T., in Equation (3.3.3) is the actual temperaturechange for the main box skin element. Geometric and material propertydata for ,"quation (3.3.2) is shown below.

r"

Flange and Skin Data (Inboard Splic Joint)

Af = 0.772 in.2 As = 3.58 in 2

f 28.7 x 106 (Test No. 11) Es = 9.2 x"10 6 (Test No. 11) jj-- 28.5 x 106 (Test No. 12) - 9.1 x 106 (Test No. 12)

(es-6 o -612Cf 7 x 1 = 13 x10

sActual Tf = 59°F (Test No. 11) Actual T = 51IF (Test No. ii)

-M 1II0F (Test No. 12) - 98'F (Test No. 12)

The calculated axial equivalent flange temperature for the twounzymmet-ically heated splice joints (Test ,o's. 11 and 12) are as -

follows: ( 1

Tf (13 x 10)(51) - .000197 - .000175 - 41.6OF. (Test No.11)7 x 10- 6 L

(13 x i0-6) (98) - .000122 - .000366 _

£ - x1 0-" 69.z1 .

0 F. (Test No.1l2) I)

where. e s -.000197 in./in. 1Test No. 11, 250'F Unsymm. V

e,. .000175 in./in

and .and , es =-000422 in/in.Test - 2No. 12, 425F. Unsym.

f .00036 6-in./in.

are'the thermal strain values calculated by the cross-section proce-d u.re. The actual temperature changes and the fictitious flangetemperature chn.nges are shown above only for the tension (critical)ide of the beam which is also the cooler side.

For- complete compatibility and equilibrium a constant strain on theskin and flange is necessary for equivalence to the axial load case.Thi? constant strain is obtained by substituting the equivalent skinand flange temperatures in Equation (3.3.1) and solving for values ofe s and ef . The difference between this strain and the e- and efvaiues for the cross-section thermal strain analysis is snali andrepresents an additional applied compression strain on the skin and

30

L

I

flange elements in this case. This incremental strain is designatedas e and the values zre as follows.

C

e c - .000047 in./in (Test No. 11)

e c . .000103 in./in (Test No. 12)

This small strain represents an additional small compression loadin the flange and skin on the tension side as calculated below.

P = ec(EsAs + EfAf) .000047 (55x10 6) -2585 lbs (Test No. 11)

- .000103 (54.6xlO6) -5620 lbs (Test No. 12)

L. These loads are relatively small and of no major consequence in thestrength evaluation of the spli'e joint; therefore, they were notadded to the applied load input for the axial splice joint analysis.

The element input flange temperatures of 41.6°1F. for Test Eo. 11 and69,40 F. for Test No. 12 are specifically associated with the skintemperatures 51'F. and 98'F., respectively. Since the actual skintemperatures varied somevhat along the span of the splice joint areathe analysis input values of the fictitious flange temperatures werealso assumed to vary and were corrected for each element along tne spanby the gradient established by the above temperatures.

31

4.0 ANALYTICAL-EXPER IENTAL COMPARISON STUDIES

4.1 Load Deformation Curves, Part 1

The load-deformation curves, Figures 4.2 through 4.6, use the bendingmoment at the particular elastic axis station as ordinate and thecritical element strain as abscissa, where element (4) is the criticallement at all cross-sections in all temperature envronments. Element4)is the center main box skin area between the forward and aft inter-

Teams and was defined as the critical element (y the definition shownin Section 3.5.1 of Volume 1 where the element with the maximum perman-ent set after removal of all load and temperature is designated as thecritical element. ) Output data from the analysis showed element @ tohave the highest calculated permanent set strain. This was also in-dicated by the test results where the highest strain gage readings werejust forward and aft of element (?) . The mode of failure also indi-cated that skin element g4 was Ygnerally critical because of the com-pression instability fail e of the stabilizers. The single exceptionwas Horizontal Staoilizer Pilot Study a which failed in tension of thelower main box skin. The tension failure may be explained by the nearlynqually critical nature of the tension and compression covers, highbuckling (crippling) allowaoles of the compression covers, and possiblysome slight variations in the cover plate manufacture and/or materialproperties. Figure 4.'. shows the cross-section and discrete elementrepre sentatiun.

Each figure includes four aeparate groups of curves; one for each ofthe instrumented cross-sections at Elastic Axis Stations 91, 102, 113,and 124. In general, each group contains three curves; two associatedwith the analysis and one with the test data. A solid line representsthe analytical load-deformation curve calculated for the temperature-load sequence +T, +P (apply temperature, then apply load) and is basedon elevated temperature material properties. The dashed line representsthe permanent set curve where the inelastic effects are associated withthe temperature-load sequence +T, +P, -P, -T (apply temperature, applyload, remove load, then remove temperature) and room temperaturerecovery material properties after exposure to the test temperatareenvironment. The experimental data is represented by specific datapoints and a solid curve drawn through these points to indicate thatthe test--analysis comparison is between the data points and the solidanalytical curve. rne early separation of the analytical load-deforma-tion and permanent. set curves indicates the onset of initial bucklingin the forward and aft box magnesium skins which have extfemely lowcr.tical buckling strains (ecr). The data points shown represent thestrain gage results at, each e)abt~i axis station using the "tempera-ture on" strain ,eadings as the reference zero. Therefore, the straingage curves shown represent an experimental applied load-deformation

curve which is comparable to the solid analytical curve. The setteroverall reliability of the room temperature test may be noted .Ln thisgroup of curves. It should also be noted that agreement between teststrain data and plane-strain analysis results is better at the threeinboard stations which are farthest away from the local effects of theoutboard loading former. The lower strain gage readings at a givenbending moment at Elastic Axis Station 124 appear to reflect the effectsof shear lag typlcal of a shallow beam loaded by a single concentratedload. The extension of the experimental strain data into the inelasticregion at Elastic Axis Stations 102 and 113 indicates this spanwisesegment to be critical with failure expected between the two stations,probably closer to Elastic Axis Station 102 due to the larger inelasticstrains shown at this station.

Two groups of load-deformation curves are shown for Horizontal Stabi-lizer No. 4 (4250F. symmetrical temperature environment) since theheat lamps were shut off just prior to the end of the load applicationportion of the test. Te decreases in structural temperature associ-ated with the time between lamp shat-off and failure of the stabilizerapparently warranted material property recovery considerations in theanalysis. However, when the lower temperature and higher propertieswere considered, the analysis tended to over-predict the failure by awider margin than expected when compared with the results of the othertests. In order to investigate the material properties effect theanalysis was also performed using material properties associated withthe higher temperatures prior to lamp shut off. The question thenarises as to whether the material property changes from the higherto t'ie lower temperature actually had time to take plats and stabilizebefore failure oceur'red. Subsequent tensile coupon data from Hori-zontal Stabilizer No. 4 shows that the actual yield stress valueswere approximately .4% less than the values used for the "lamps off"analysis using recovered material properties at the temperature attime of failure. Refer to Section 4.5 for the tensile coupon dataevaluation. This may indicate the presence of a lag in material pro-perties vs. soaking time which may vary considerably in highly trans-ient temperature environments. In tKis case the prediction of overallstrength at any particular time wi.thin the transient temperatureenvironment may be quite difficult since material properties data maynot be sufficient or defined well enough in their present form.Further study in this area appears to be warranted where structuralcomponents subjected to severe transient heating could be loaded tofailure fairly rapidly after exposure to temperature environments wherer .. .an .recovary would occur rapdly. The mLattI&I pr'.perbies for the

-. highly transient temperature environment must be defined specificallyat the time of failure by using tensile coupons from the test componentsand reference coupons from similar plate material. This procedure 'ormaterial properties verification was investigated and the results areshown in Table 3.15 of Volume II and in Section 4.5 of this volume.

L 33

?I

Good agreement is shown by the comparison of strain gage data pointswith the analytical curve, particularly in the region of failure

j1 between Elastic Axis Stations 102 and 113 for all temperature environ-ments. In most cases the elastic slope shown by the experimental dataindicates slightly greater stLffress than the analytical data. Al-though no specific reason for this difference is known, there areseveral items which contribute to this trend. The geometry (skin -

thickness, spar flange and web thickness, tolerances, effects of local,)aas) and material properties (primarily the modulus of elasticity)may vary. The strain gages are not located exactly on the centroidof analysis element 0 . The element idealization for the analysisprocedures may tend to slightly underestimate the moment of inertiaof the cross-section. Finally, the loading former and load line set-up (canted inboard for normal surface loads at yield load) producesslightly smaller bending moments at each elastic axis station thanthose used for plotting the experimental data. This effect was pro-duced by the large deflections and relatively small radius of curva-ture of the stabilizer at yield load where the components of thenormal surface load gave somewhat reduced bending moments at eachcross-section.

The strain data plotted in Figure 4.6 for the 4250F. unsymmetricaltemperature environment (Horizontal Stabilizer Test 6) are parti-cularly noteworthy since strain data was obtained at relatively 7'

high strains (.010 in./in.) very close to ultimate load.

I3

(t

*4-0 9I

\00

E~ ) - l 0t2.0 L-El4

0C V

IA :3 0 H04

4)0 N

4-4 H4 0'0o 'o 0 O, 4H r.1r

10 J\E-4

0 -0

m co

ul -:

Q9 E-®O®©

L. 35

Element (D is the critical element at all elastic axisstations. Refer to Figure 4.1.

-- Calculated Load-Deformation----- Calculated Permanent Set

A Gage No. 2A, E.A. Sta. 91----- - ge <,.- -(6k Smi- , .A &t-a- -- --

SGage N (. 1]A(101 Sia.lar),E.A Sta, 113

WA S ) mil ),E A.Sti.124 -E.A. Sta. 91

6 kn al 313

Test

O0 5 - . _

- E.A. Sta. 102

Teit Sta. 113

1/ Analysisj

S--Te t

---A---y--I- -EA Sta. 124

00 0.002 0.00 (,OOA 0.008 0.010 0.012 0.014

Element Strain, in,!in

F'iare 4.2. Load-Deformation Curve Comparison With StaticTest Results, Room Temperature

36

Element is the critical element at all elastic axisstations. Refer to Figure 4.1.

Calculated Load-Deformation- ---- Calculated Permanent 3et

A Gage No. 2A. E.A. Sta. 91

---QZeN 7__6.1tA10

6- E . ti) a. 13 C

Ef.A. Sa. 24

5 Ana y sisj Te s

4 E{A. Sa.j 3

- - E Sa 1

EA Sa.24

flgare 0.~ ;.004 0.06 000 0.010 0.012 0.014

Fiere .3Load-Deformation Curve Conparison With StaticV Test Results, R50 0 F. Symmetrical

37

Element (4f) is the critical element at all elastic axisstations. Refer to Figure 4.1.

-I Calculated Load-Deformation----- Calculated Permanent Set

A Gage No. 2A, E.A. Sta. 91o Gage No. 7A (6A Similar), E.A. Sta. 102

7Gge o. IlA (1 A Similar .F.A. Sta. 11_3

tl 2, --" I _ : /'/ !;d/ 7 - .... - -...

TA (I.A Si , ,.1A7 S a. 4i

-- 'iI I I . , . - .

- IA

4 .-- 4 .... E L- 1 --

Element~~ S~an i.i

Tes5t

2 -47

i ~ ~ ~ ~- An~lysis EAt 2

0 0.002 COI 0.006 6.008 0.010 0.012 0.01L

Element Strain, in./in.,

r,- re 4.4 Load-Deformation Curve Comparison WithStatic .'est Reiults , 2500i. Unsymetrica!

-40

cc~~~~*, (v . ( :4,

HA A

0 -S.

-4- 4 -)

0~.~4~cp -- 7-- N-'r- -4 (-I -t m ar4 -

4-)4-'

4J 4- .4)

4 4-'

4,O C) - -'t)Cl

El- .o (0 011. - - -

w-4i4- -4I (zV- r.

00 H0 4- 443 -r -, E-4-

0 " .- * -. 4.H~' .~ 0 0

4- 4- - - __A

r:0 r I -K ;7 -a)~ .,1 00 *>- 4 A I-S -1

I

I- 0

~OT x VqT-.Ul -juaWow JuTpulo

it 39)

Elemont ( is the critical element at all elastic axisstations. Refer to Figure 4.1.

--- Calculated Load-Deformation----- Permanent Set

A Gage No. 2A, E.A. Sta. 91O Gage No. 7A (6A Similar), E.A. Sta. 102) Gage No. 11A (10A Similar), E.A. Sta. 113 Cale. Ul• Gage No. 14A (15A Similar), E.A. Sta. 124

alyst s --

" l E.A. Sta. 91 i

- ~ 7 U timate

2U

---- - -- P -

00 0.002 0.004 0.006 0.008 0.010 0.012 0.016

Element Strain. in./in.

Figure 4.6. Load-Deformation Curve Comparison With Static TestResults. 4250F. Unsymetrical I

40B

4.2 Available Ultimate Strength and Critica.Station Selection. Part I

Since the actual calculations are performed for the four specificinstrumented cross-sections, the critical cross-section at whichfailure could be expected cannot be determined directly. However,since the structure is continuous and no sudden or major changes incross-section occur within the test region, it is possible to plota curve of available ultimate strength versus elastic axis stationfor each temperature environment. The expected failure station de-fined by analysis is determined by the intersection or point oftangency of the applied moment curve with the available strengthcarve. This intersection or tangent point also indicates the valueof the critical bending moment at which the failure can be expected.The applied bending moment distribution is represented by a straightline with L' = 0 at Elastic Axis Station 135. This line drawn tan-gent to the available strength curve represents the ultimate ap-plied bending moment distribution for a single concentrated load atElastic Axis Station 135 (outboard loading former location). Thesecurves are shown in Figures 4.7 through 4.11 for each of the fivetest temperature environments. The test ultimate loads and failurestation are also indicated on these curves by individual data pointsfor each horizontal stabilizer. Figures 4.7 through 4.11 are pre-sented as another means of demonstrating the feasibility of theapplied load ratio method and comparing analytical and test data.If a heated test structure had severe spanwise transient temperaturedistributions then thermal stress and material property degradationeffects would vary considerably along the span. In this case it isentirely possible that the critical station for the simulated roomtemperature test may be quite different from that for the elevatedtemperature test. For the simu10i6ed static test this may mean aredistribution of external applied shear loads and a modified shapefor the test applied shear and bending moment curves. The addi-tional complexity and work involved to investigate critical stationvariation between elevated temperature aid room temperature isrelatively small since the required data would be available fromthe cross-section calculations. Essentially all that is requiredare additional curves similar to Figures 4.7 through 4.11 with allow-able yield and ultimate bending moment, and yield and ultimateapplied bending moments plotted versus span.

Although the curves in Figures 4.7 through 4.11 show only the ullt!-mate load comparison the same procedure can be followed for yieldload, limit load, or any selected level of the applied load whichrepresents a design criterion. Yield curves would be of particularinterest when the materials and/or structural geometry are suchthat there is a distinct difference between the yield and ultimate

4.1

load. However, in maxr realistic cases the airframe component willfail by compression instability where yield and ultimate are verynearly the same value. The horizontal stabili2er failures were ofthis type where yield data based on 0.002 in./in. offset strain ofthe load-deformation curves of Section 4.1 would not produce sig-nificantly lower curves than those shown for ultimate in Figures4.7 through 4.11. Past experience on crippling failures and thastudy of various semi-empirical approaches to crippling analysisas well as the previous box beam studies of Reference (d) tends toconfirm the yield load criteria for compression-loaded structures.For this reason, no yield curves are shown and the yield-criticaland ultimate-critical stations are assumed to be the same.

The variation in the location of the contact point of the tangentline and the calculated available ultimate strength curve shows agenerally critical span to exist between Elastic Axis Stations102 and 113 for this particular applied moment distribution.

The analytically predicted ultimate loads (bending moments) andfailure stations shown in Tables 5.1, 5.2, and 5.3 were obtainedfrom the curves of Figures 4.7 through 4.11 where the predictedfailure station is defined by the point of tangency of the appliedmoment curve with the ultimate allowable moment curve.

4

J£

0 ( Pilot Study Horizontal Stabilizer Failure

(Horiz. Stab. Pilot Study a)

A Horizontal Stabilizer Test 1

8 Calculated Available Ultimate Strength

Applied Bending Moment

/lA. E.A. EA EAS91 102 113 124Sta. Sta. Sta.a.

0

I° 4 "o ent) A.-

SSpultimat43 ent

i appl Led

2-----------------------

80 90 100 110 120 130Elastic Axs Sta. ( inches

Figure 4.7U~iaeBn!gMmn Ys Soan. Room Temnerature

i'43£

t i

o9 Failure of Horizontal Stabilizer Test 3

El Calculated Available Ultimate Strength

- Applied Bending Moment

OA. E.A. E.A. E.A.112 113 / 124

Sta.S Sta. Sta.. Sta.iC

8grir

6

I 4 , . .. -- mqmentl2

oI

w " " _appl Led,,2" mome it

S80 90 100 110 120 130 140

!i Elastic Axis Sta. (YA) inches L

i Figure 4.8Ultimate Bending Moment Y s. Sn, 250lF. SymmetricalI

4., - -- - - - - - 1S..j4

UI

D e Horizontal Stabilizer Test 5(Test Failure, E.A.Sta. 102)

19 Calculated Available Ultimate Strength

Applied Bending Moment

I E.A. VBA. E.A. 4 A91 102 113 124

Sta. Sta. Stut. Sta.

0

" 6 -t'H

If - - -4-

I ultimate4 -moment

appliad

1 1

80 90 100 )10 120 130 140Elastic Axis Sta., (Y'H), inches

Figure 4.9 ,

Iltimste Bendine Moment vs. San.250°F. Unsymmetrical

!45I

£L

(i) Failure of Horizontal Stabilizer Test 4

X Calculated points using material properties fortemperature distribution assuming heat lampswere on at time of failure. No temperature re-duction and no recovery of properties.

ID Calculated Available Ultimate Strength(Recovery of material properties for lowertemperatures as heat lamps were off prior to Lfailure)lamps-on B M n

- ___- lamps-off Bending Moments

(E.A.) E.A.) E0.A. EA91 102 113 124Sta. Sta. Sta. Sta.

IICJ

0q

o moenr-t

Fu ultimate

. -~mom nt-

S80 90 100 1 I0 120 13p102

Elastic Axis Sta. (Y'H), inches

Uliat enig Figure 4.10

UltimateBendng Moment vs. Span,4250F. Symmetrical

1

U!

II!I

G Failure of Horizontal Stabilizer Test 6

El Calculated Available Ultimate Strength

-------- Applied Bending Moment

E.A. E.A. E.A. EA91 102 113 124

Sta. Sta. CSt)a. Sta,

10 6 ! !I ........ ..

2-- 1

w :Z____ p] -ied

80 90 100 110 120 1

Elastic Axis Sta., (YA ) inches~Figure 4.iiUlt _BeXding Moment v s. Span, 425°F. Unsvetric

C 47

4.3 Analytical-Exoerimental Deflection Comparisons, Part 1

Deflection studies were performed in order to show comparison betweenanalytical and experimental deflection results. Using the proceduresshown in Section 3.4 of Volume I, the deflected shape of the horizon- Ltal stabilizer at failure, under the various thermal environments, wasobtained. The analytical spanwise deflection plots were then comparedto thp corrected experimental spanwise deflection. The correctionsto the experimental deflection data were obtained by plotting load-deflection curves for each gage and then correcting each plot to assurea zero intersection. Once each gage had been corrected to zero, themagnitude of the correction was added to the value of the deflectionof that gage for all values of the applied load.

A deflection rate study was performed using both the corrected experi-mental data and the analytical deflection data. In the case of theanalytical data, the actual test failure loads were used, rather thanthe predicted analytical failure loads, to denote the maximum appliedloads. This was done in order -o show a direct comparison betweenthe analytical and the experimental results. Using various percentagesof the test failure load to establish AP values, the values of 66, thechange in deflection corresponding to the load increment chosen, werecomputed. The deflection rate plots were then used to show a pre-diction of the failure station. A table has been included in Section4.4 showing the predicted failure staUion due to the analytical de-flection rate data, the predicted failure station due to the experi-

mental deflection rate data, and the actual test failure station.

The plots presented in this section show a similarity between the ex-perimental and analytical deflection data for the Part 1, or cross-section, portion of the study. No meaningful deflection data wasobtained experimentally for the Part 2, or joint test, portion ofthe study. Since the Part 2 portion of the study was concentrated onthe stabilizer root end joint, no spanwise test data was obtained out-side the local area of the joint. Deflection gages used in the Part2 study were located at the leading and trailing edges of the stabi-lizer. Thus, the gage readings were a function of the twist orwarping of the cross-section. If the gages had been placed on thestabilizer main spars, then an indication of the rigid body rotationof the stabilizer would have been obtained. However, some chordwise,or anticlastic deflections were present which could not be accountedfor within the limitations of the analytical procedures.

Figures 4.12 and 4.13 show spanwise plots of calculated deflectiondata and corrected experimental deflection data for the five tests.Load-deflection curves for each gage were plotted to obtain the nec-essary zero corrections for the experimentally oLAained deflectiondata.

30-30

'Room Te rper ture

0 + Tesi

, 20 - - f

0

80 90 100 110 120 130 140 150

Elastic Axis Station, inches

30 - - - - - - - - - - - - - -

2500 Un metrici el

C + Test

20- -- --x Calc lated

0

4-)

0 .

80 90 100 110 120 130 140 150

Elastic Axis Station, inches

U 30

u 42'10 SYmetrical

+ Testo20- - -

x 'Calc ated

0

890 9 00 110 120 130 140 15

Elastic Axis Station, inches

Figare 4.12.Comparison of Test and Calculated Daflections, Part I

49

304250 Unsmmet ical

+ Teso 200

x Cal ulated

o i0+ '

S10~~~ 1-

- - - - - - - - - - - - - -I

80 90 100 110 120 130 140 150

Elastic Axis Station, Inches

2500 Symm 3tri al 4-

+ Tesl +{20 .

x Calc ulat d

- i0 I...lifli1L

80 90 100 110 120 130 140 150

Elastic Axis Station, Inches

Figure 4.13.Comparison of Tast and Calculated Deflections,_Part 1

50

IT

A study of the experimental and analytical deflection data revealsthat the shape of the analytical curves agrees very well with theshape of the experimental deflection curves. In all cases the mea-sured deflection was greater than the calculated deflection. Ingeneral, it can be seen that the analytical deflection curves vary

from the experimental deflection curves as a function of some angleof rotation. If the root end of the stabilizer is rotated slightlythe two curves will match almost exactly. The necessary angles ofrotation vary from 1.70 to 3.70 . These small angles would be diffi-cult to detect during the progress of a test unless dial gage in-formation were available for root end fitting deflection. The in-board end of the stabilizer, with its attach fitting and hon,, isincapable of transferring high reaction loads from the stabilizerwithout deformation. In the analytical procedure, of course, afixed end condition is faithfully reproduced; and the root endslope is zero.

An investigation was undertaken to attempt to compare the rcsults ofL the deflection studies with those of Allen, Reference (f). In con-clusion 2 of his box beam study, Allen states that the deflectionat. ultimate load in a built-up aluminum alloy aircraft structurewhich fails by buckling is substantially independent of temperatureI; up to at least 4000 and decreases scmewhat above this point. Hethen presents a formula by which the deflection of a structure atany temperature may be determined if the room temperature deflectionis known. From Reference (f) the relationship is

'PT

[Jl K T_ Ut (4.3.1)U (E)RT

where K = 1 below 4000 F. and K = 0.9 above 500°F.Since tAhis formula applies to elastic conditions, it produces good

rresults in the elastic range. The assumption here has obviously beenI; that buckling occurs at stresses below the yield stress. The use of2024-T4 material in the Reference (f) study makes the above formulaless general. In the present study, the 7079-T6 stabilizer skins failedin buckling, iat the deflection at failure was somewhat more at 4250 F.than Allen's formula predicted. At 250 0 F. the deflebtion at failure, asstated by Allen, was substantially the same as the deflection at

51

failure in the room temperature test. If we examine the variation inyield stress with temperature for 2024-T4 and 7079-T6 the reason for Uthe inconsistency becomes apparent. A comparative plot of this varia-tion is given in Figure 4.14. L

1.00

0 k 20 24-T4

0 .6o

~7079..T6---

400

02 __ _ __ _ _

100 200 300 400 500 600 700 800Temperature, F.

Figure 4.14 Vnriation of Yield Stress With Temperature

It can be seen from Figure /+.114 that the yield strength for 7079-T6begins to drop off considerably at 300F. "whereas the 2024-T4 retainsa high percentage of the room temperature value out to 4000F. Thus,in the case of the 7079-T6 material,, the constant, K in Equation(4.3.1) was set equal to 0.9 for temperatures above 400°F. This iscompatible with Allen's work except that the temperatures have beenlowered by 1000F.

Since the load-deformation relationship for the stabilizer cross-section indicates inelastic action at failure, it was decided tomodify Allen's formula into the following form where the deflectionsare considered to be a function of both yield stress and modulus ofelasticity.

=R K ult (43.2)ult (Fy)T./ET 1

(F,)RT/-vRT

52

I

where K = I below 300 0F. and K = 0.9 above 4OO0 F. This form of theequation accounts for the variation in the yield strain with tempera-ture.

Figure 4.15 shows curves of stabilizer tip deflection versus tempera-ture. Included for comparison are the test deflections, the de-flections given by Allen's Formula, and the deflections given by themodified Allen Formula. It can be seen that a for.mula of this typeshows great promise as a means of predicting 'eflections empirically.No definite conclusions can be made at this tims, however, due to thelimited amount of available experimental data However, the compari-sons illustrate deflection procedures to be an effective means ofequivalence, or simulation since the deflections are a function ofthe spanwise strain variation.

1 30

S2Test D flection,"'20 - Ne Ajen's For mla (Eq. .3.1 ,• 4 m MC dia! &_A1 n r 1e1 A-E,I... _

Qlo-V 0 -2 ----T

100 200 300 400 500Temperature, OF.

Figure 4.15 Stabilizer Tip Deflection vs. TemperatureI

St

53

4.4 Deflection Rate Study, Part 1L

A deflection-rate study was performed using the experimental and ana-lytical deflection data. Comparison plots are included for severalcases. A table has also been included to show a comparison betweenthe actual failure station and the failure station predicted by vari-ous techniques.

Figures 4.16 and 4.17 show the analytical and experimental deflectionrate curves, respectively, for the 4250F. unsymmetrical test of Part1. Similar curves were prepared for all of the tests. This parti-cular test was chosen as an example to demonstrate the technique.

The analytical curves were drawn using the test failure load as themaxinnim load rather than the predicted analytical load. This was 1done to equalize the two sets of deflection rate data to the sameload. The load deformation procedures, as given in Section 3.2,Volume I, for the analytical failure load prediction provided in-elastic strains corresponding to the various input loads. These (inelastic strains, added to the input elastic strains, give the de-flections for the stabilizer under any given loading condition.From this data, the A6 and AP values used for Figure 4.16 were ob-tained.

The experimental values of A6 and AP were obtained from the datataken during the tests. The loads were used directly, while thedeflection data was corrected. The data correction was performedby plotting the variation of each gage versus the load and off-setting the plotted curve to the zero origin. St'dy of Figures 4.16and 4.17 shows that the plots are very similar. The failure sta-tion, as predicted by these plots, is essentially the same, andvaries by two inches from the actual failure station.

Plots similar to those shown in Figures 4.16 and 4.17 were preparedfor each of the five tests. The results of these plots are pre-sented in Table 4.1. Also included in Table 4.1 are the actualtest failure stations for each test and the failure stations aspredicted by the analytical load-deformation procedures of Section3.2, Volume I. In the case of the room temparaturo toot and the2500F. unsymmetrical test, no deflection rate solution was obtained.Due to relatively small magnitude of inelastic behavior at failureduring these two tests, there was no appreciable change in de-flection rate throughout the test, and thus no break point in theset of curves to indicate failure location. It is interesting tonote that when the experimental deflection rate curves were un-able to predict a failure station, the analytical deflection ratecurves also failed to predict a failure station. Similarly, whenone method did predict a failure station, the other did also.

54

I,.004 4250 Unsymmetrical

Test

. 0 0 3 t 1 F o r m r\ ' 7 --

Lad 7012 lbsAP I 623.3 ibs .

in./lbs. .002 Test

001 -Base Lim e

S0 90 i00 110 120 130 140 150

Elastic Axis Station, Y, .iches

Figure 4.16.An0,ytica Deflec tion Rate Data

4250 Uns mmetrical Test

.003 - 779. lbs z

.03 OUi bT )ormea 7012 lbsLoAP I 623:1 lbs./.

fir../Ibs. .002 - - -

U ~ TeasFail ie-

80 90 100 110 3.20 130 140 150

Elastic Axis Station, YA, inches

Figure 4.17. Experimental Deflection Rate Data

55

In the case of the 4250F. symmetrical test, an analytical deflectionrate study was run for the case of lamps on at failure and for lampsoff at failure. The main difference between these two analyses isthat of material properties input. Early lamp shutdown during thistest raised a question as to whether the material properties at fail-ure had recovered to temperatures indicated by thermocouples. A dis-cussion of Test 4 is given in Section 4.1 of this report.

The inelastic strains were computed for every ten percent of the testultimate load. These inelastic strains are shown in Figures 4.18 and

i 4.19. It has been shown that the change in slope of the deflection

rate curve is caused by an increase in the rate of incremental in-elastic strain. It is noteworthy that the analysis for the lampsoff at failure, which shows no incremental increase in inelasticstrain with increasing load, produced no deflection rate results.The analysis for lamps on at failure, however, produced deflectionrate rezlts which agree quite well with the actual experimental re-sults. The inelastic strain variation with increasing percentages ofapplied load for this case is shown in Figure 4.18. The incrementalchange in strain is seen to increase abruptly at full ultimate load.These deflection-rate results for this test thus support the resultsobtained in the analytical load deformation procedure of Section 4.1.We have, then, frther evidence to support the proemise that the ma-terial properties did not recover during the short time between lampshut-off and stabilizer failure in the 4250F. symmetrical test. Thesignificance in the curves of Figures 4.18 and 4.19 is that thegreater inelastic effects and the greater incremental change from 90%to 100% shown in Figure 4.18 will increase the deflection rate morethan the effects shown in Figure 4.19.

The analytical deflection-rate studies tended to divide the tests in-to two distinct types in regard to the incremental variation of in-elastic strain with increasing load. In the case of the 4250F. sym-metrical test 4, the 4250F. unsymmetrical test 6, and the 2500F.symmetrical test 3, the plots of inelastic strain versus increasing L!load were similar to Figure 4.18. The room temperature test 1 andthe 2500F. unsymmetrical test 5, however, produced plots of inelasticstrain versus increasing load which were similar to Figure 4.19.

A -*--LY,01 ofth deflection rate results s presented in Table 4.1.

The predicted and test failure stations are listed for the experi-mental and analytical deflection rates. For comparison purposes, thefailure stations predicted by the load deformation method are alsopresented.

The failure stations predicted by the deflection rate method are seento be quite similar to thuse predicted by the load-deformation pro-cedures. This is to be somewhat expected, since both methods rely on

56 ['

/.25F. &metricaI F -0 0 100 of failu e

-.0008,

0 -- 0006 - -4 -.ooo4 - -4.,

I -.0006 -- " - -o

0070 909 0 i 2 3 4

4-80I -0004 -7- -----

'-50

0

70 80 90 100 '110 120 130 140Elastic Axis Station

Figure 4.18, Variation in Inelastic Strain versusElastic Axis Station For Various Percentagesof Failure Load

(425UF. Symmetrical Test, Lamps On at Failure)

U425cF. Sqrmetrica]-.0006 lams off at lailure

.- IC% of failre

\o o ad

o-.0004 -. 5%61'

0 80%

a, -.02 -- -0 -

------------ --- -O-0 0 0 0i0 i040

70 80 90 100 110 120 130 140Elastic Axi s Station

Figure 4.19, Variation in Inelastic Strain versusElastic Axis Station For Various Percentages of iailure Load

(425UF. Symmetrical Test, Lamps Off at Failure)

I57

V 0.

H. aH

4 0

d)P 0). r4-

4.3,U) 0 fl

V) 0 ~ J0 CU) -H *rj

E! 0 jD 0 4J'

0 0 () , '

w4- rl-r

.4.'~U t'0r- (I ~ ~n0l 0r 0d 00

a U,0 N1

M 0 90 -c U) 1

-H z , 1co ~ ~ E-0

4,) C% Hn

.0~ C/ V*

C).

W ~ ' cd -e .HS 4-

0 Yd) 0 )r

-4 4-4) L+) C)4.) t4) vi

0L

the inelastic strains to find the plastic hinge location on the struc-ture. The inelastic strains used in the deflection rate study weretaken from the load-deformation output.

A comparison between the predicted failure stations and the actualtest failure stations in Table 4.1 shows that a high degree of pre-dictability exists. The maximum difference between the predictedfailure station and the actual failure station is 5.5 inches, whilethe average difference is 3.2 inches. These distances are smallenough to make the predicted failure stations lie within the failurezone of the actual test failures.

The correlation between analytical predictions and test data obtainedin this study substantiates the use of analytical failure station pre-diction as an important tool in the area of structura. testing. Theload-deformation predictions are particularly useful since they arecomputed prior to the actual testing. In this manner, the failureregion can be pre-determined and instrumented accordingly. Anotheradvantage of the load-deformation prediction method is the fact thatit accurately predicts critical areas regardless of the presence orabsence of inelastic effects. The deflection rate method may be tooinsensitive unless large inelastic strains are present, and in gener-al, requires accurate deflection measurement. Both methods work wellat elevated temperatures on aluminum alloy when buckling and increasedductility produce larger inelastic effects.

An approach to the use of these methods might be to perform the load-deformation calculations prior to test and to plot the results. An-alytical deflection rate calculations may then be made to verify theresults. Then, during the actual test, the experimental deflectionrate data can be pl otted on the same graph as the load-deformationdata. In this manner, any deviation from the predicted behavior ofthe structure will become immediately apparent. Indications of earlyfailure would then appear in time to allow corrective measures to betaken.

59

4.5 Material Prooeties Investigation

The standard practice of testing tensile coupons from each test com-ponent after completion of the static test was followed in this study.Although this procedure is specified to compare the actual materialproperties with those used in the analysis, it is considered evenmore important for tests performed at elevated temperature than atroom temperature. The greater variations associated with the ele-vated temperature environment and complex temperature-time historiesmake the prediction of important material properties, such as theyield stress, somewhat more difficult and questionable. In this studyit was mandatory to investigate the actual versus predicted propertiesto evaluate the validity of the overall horizontal stabilizer resultssince the analytical procedures depend highly upon the yield stressparameter for accurate prediction of the load-deformation curves,permanent set, and ultimate load.

The experimental coupon investigation for defining material ;-opertiesat the time of failure of each horizontal stabilizer is described inSection 3.1.3 of Volume II and the data from the individual stabilizersis summarized in Table 3.15 of Volume II. The coupons evaluated wereobtained from Elastic Axis Station 127 which, according to the tempera-ture-span plots similar to Figure 3.2, had temperature-time exposurehistories as specified in Table 4.2 which summarizes actual horizontalstabilizer yield stress parameters and those predicted by the para- [metric curves for the tensile coupon elastic axis station and coupontest temperature.

z. TeAL Coupon Prior TquLv. t L-rion- " Tenile Per Cent IT ,ie- I Tensileib. Temp. Llstic :xpoure fime at -llr Coupo R.T. Yield Iu. -on

Con- Axis Ten., Exposure Paranoter Test Yield j r , , *.

dition 3ta.,in. IF. Temp., hrs. (Yield) Temp., Stress Fty, at Stress,OF. Coupon Ft ,TableTest Temp. 3.5,Vol.II

R.T. 139 R.T. - - R.T. - 73510 73,23

25O"F. 127 208 0.587 9130 250 0.885 65103 64664

A 425 F.4 4250. 127 33 1.1,00 11570 383 0.605 4452J 42768

U.1 12 21.' 0.553 I91&) 259 0.885 65100 64493

S Un.y. ...'2 7 J3, 0.822 11680 450 0.1,49 330C0 32323

Table 42. Tensile Coupon - Predicted Yield Stress Summary* Main box skin temperature measured at Elastic Axis Sta. 125

Main box skin element time associated with prior exposure temperature• Each value is the average of two tests

60

The data summary in Table 4.2 illustrates, by comparing the last twocolumns, that the predicted values for a given temperature-time ex-posure history are verified by tensile coupon results having the samehistory. This indicates that the material properties used in the an-alytical procedures were sufficiently accurate to allow a directcomparison of overall horizontal stabilizer results such as ultimateload and load-deformation behavior.

The maximum difference between predicted and tensile coupon test pro-perties is for Horizontal Stabilizer No. 4 (Test No. 4), 4250F. Sym-metrical, and shows actual properties to be about 4% less than thepredicted values. This fact coupled with the material propertyvalues used in the analysis for lamps on and lamps off tends to indi-cate that the overall stabilizer results for the lamps on conditionmight be the more reliable comparison. The analysis material proper-ties data for Elastic Axis Station 124 (nearest to coupon area ofElastic Axis Station Station 127) indicated approximately 45000 psiused for Fty for lamps on and approximately 50000 psi used for lampsoff.

The differences are not too great, however, and the amount of data istoo limited to allow any firm conclusions to be drawn. However, thisinvestigation and the overall results shown in Tables 4.2, 5.11 5.2,and 5.3 for Horizontal Stabilizer No. 4 illustrate the significanteffect of the material properties evaluation (particularly yieldstress) on overall test results and the applied load ratios for roomtemperature simulation.

In Table 4.2 the values shown in the next to the last column aretypical main box skin element properties as defined by the exposuretemperature and time and the yield stress parametric curve shown inFigure 3.8 of Volume II. The last column in Table 4.2 denotes theactual test values obtained from horizontal stabilizer coupons.

61

4.6 Analytical Rsults for Plate Element and Fastener LoadDistributions at Ultimate LoadL Part 2

The diagrams shown in Figures 4.22 through 4.26 show the distributionof plate element axial loads and fastener shear loads on the tension(critical) side of the horizontal stabilizer for each temperatureenvironment at the calculated ultimate load. The load distributionsshown include the main box skin, the steel splice plate between theinterbeams and main beam fitting, the two outboard interbeams, andthe steel main beam fitting flange. The figures include diagrams forthe following test and analysis temperature environments and momentrestraint conditions. Based on the results of the thermal moment re-straint study of Section 4.7 the diagrams in Figures 4.24 and 4.26for Test No's. 11 and 12, respectively, represent the results forthe assumption of full elastic thermal bending moment restraint.

(a) Pilot Study b, Room Temperature, ultimate applied former

load = 18000 lbs.

(b) Test No. 9, 2500F. Symmetrical, ultimate applied formerload = 17250 lbs.

(c) Test No. 11, 2500F. Unsymmetrical, ultimate applied formerload = 16770 lbs., assumed full thermal bending moment re-straint.

(d) Test No. 10, 4250F. Symmetrical, ultimate applied formerload = 12420 lbs.

(e) Test No. 12, 4250F. Unsymmetrical, ultimate applied former

load = 13030 lbs., assumed full thermal bending moment re-straint.

The calculated ultimate loads shown above were obtained using themethods of Section 3.3 of Reference (a) where the failure criteriawere specified in Equation (3.3.8.1) on pages 31 and 32 of Reference(a). For eqch applied load investigated analytically the failurecriteria of Equation (3.3.8.1) of Reference (a) were applied afterconvergence of the calculations Vor the redundant fastener shear loads.Failure of a fastener segment, plate element, or fastener-plate in-........ (berig) is noted depending upon the failure criteria

'ests for fastener shear, plate element strain, or fastener-plate

deformation. The loads at which the splice joint failed to pass any

one of the three specified tests were obtained from an IBM 7090 com-puter analysis programmed according to the block diagram solution ofSection 4.2 of Reference (a).

62

1L

3

The splice joints were analyzed by the inelastic method, and resultswere plotted on Figures 4.22 through 4.26. The inboard end fastener(or row) had higher loads than the outboard ones, and the plateareas surrounding them were deformed inelastically. Failure did notoccur at this time. Instead, loads were first redistributed, andsuccessive outboard fasteners picked up more of the redistributedload until several fasteners failed simultaneously. In Figure AI.8,several fasteners near the inboard end have failed, demonstratingthat a single inboard fastener did not initiate the failure.

The splice joints were not analyzed herein by the elastiz method,but Reftirence (g) shows that the inboard end fastener (or row) doesnot fail immediately after becoming critical.

F For splice joint geometry and identification refer to Figures 4.20ij and 4.21.

L.

L63

>I. -a.or

40

2.15

/ 2

4j /~

let-

-V I -o

0

640

-I a

#4't I o

Z *O

i ft, - -

.1 '00

FL LK_ 0

4ci

1 A

4: 5

Sta St.Sa

1' -. ~,-~_ -48% -S

1199 M"2.&

U T 1' . Fastener She p Lo d 3 r- i

, 3

Sc l Scnl "A '

Eleen Load Faeril ure. Loads- an Plate-- ! -- Scale "A"

C 0iFigure 4+.22 Fastener Shear Loads and Plate

Element Loads t Failure, Inboard Solice-joint,

Part 2, Room Temperature (Test No,

and Pilot Studyl

66

I!

I:

''5.621 6 5 '73 ' 5R : ' i ;49.56

I I

t ' -~ - ~~ "116 - -wt -l.d

16t 45 ZIoot~4

14_- _. '..1

L 6Steel Splice P1 I

~Main Box kln

x: 12 300 _____ Scale -A"

A 8 200

L IL

SI ntder-Se- ,-

-- ,00 I

I r ! if i .-- i i I al 1 A'1 il ,

Figue 423 FsteerSear La an la

Element Loads at Falure, Inboard Splice OoinitPart 2. 25O0F. Svm (Test No, 9)

67

W Kw M -t

241-14.11, . *t -14%4 -¢ -l' tf * sss%

OUT31D Fastener 3he-r I ads-_i i -- i16 400 I-------- ,• I -- teel 3;le PI te,

Scale '3' J" M-in P-ox Skin [1o 12 3/ Scale VI

0 7 W :,"-

•~ intr-Bue

r* n,! d

Figure ~~V. i.4 F s e e h a o d n Beat ,n

C.i Inoard i

,~~~ ~ ~ ~ ~ .Reinne-nBedn Fo.hra.. omet r

0 0

0 Ii late 1-- nt Lmt

Figure 4.24 Fastener ShearLoads and PlateElement Loadsat Failure.noadSle Joint.

Restrained-in-Bending, For Thermal Moment [6

L

IE

St ., .

.-- - -- WG4 *IW I'., oi~l~

',, I ,Wr *PW 1T" " -," ' '-<,,., ._ -,-_. -- 2...' ....

OUTBID Fa tener $eA Lods

I t ale "' , t n

1- 300

i0 4 100__ __

L. Thter.Demw~

O-I -

Plate Element Loada

Figure 4.25 Fastener Shear Loads and PlateElement Loads at Failure. InbX&rdSD)~ice Joint,

Part 2. 425°F. Svmm, (TestNo. 10)

69

-TitoU -- In" -'24b4.

16 -.

Steel '. c i ,'r J

r"-.A '. . cle '. -1-._ --

I- j r'' JI

J L / - _Jl' ln ox :.

t: " ~cale V ."

0 -_0__ _ " -

Figure 4.26 Fastener Shear Loads and Plate

Element Loads at Failure, In~board Splice Joint,Part 2, 4251F. Unsymm, (Test No. 12)

Restr-ained-i n-Bnnd ns7 ,oi, Thermal Moment

70

I

4.7 Thermal Moment Restraint Study1 Part 2

Chordwise plots of experimental thermal strain data are shown inFigure 4.28 for two separate stations along the span of the mainbeam fitting. From Figure 4.20 these stations are approximatelyElastic Axis Stations 67.6, and 55.6, Page 64. The data plottedwas obtained from ASL magnetic tapes for tests 11 and 12, the un-symmetrical environment tests and is for the tension (cooler)side at the beginning of load application. This data, then, re-

-presents essentially the thermal strains for the stabilized tern-*v perature environment.

The experimental data tends to indicate a relatively large positivestrain in the main box skin at the outboard end of the main beamfitting showing that some degree of moment fixity does exist sincethe unrestrained-in-bending case should produce a negative strain

r in the main box skin for the temperature distribution present ac-cording to the thermal strain calculations. Further evidence whichtends to support the assumption of moment restraint is the plottedvalue of calculated thermal strain for the restrained-in-bendingcase which is positive (tension) and of the same order of magnitudeas the peak skin strain. These facts plus the shape of the chord-wise strain distribution curve indicate a general internalself-balancing moment condition in the stabilizer root area similar

1 to that shown in the sketch in Figure 4.Z7.

The generalized condition shown in Figure 4.27 can logically be pre-sumed to exist for several reasons. First, steep chordwise tempera-ture gradients are pres,. in the extreme root area and are producedprimarily by locally heating the main box area bounded by the frontand rear beams and the inboard and outboard root ribs. The frontand rear box beam areas were essentially protected from direct heat-ing lamp radiation by curtains and became heated only by conduction,thus maintaining a generally lower temperature level in these areas.

The presence of these relatively steep temperature gradients, notonly through the depth (thickness) of the stabilizer, but chordwiseas well, would severely affect restraint conditions when coupled withthe relative torsion and bending stiffnesses of the front and rearbox beam sections, the inboard and outbcard root ribs. and the heavy

fl main box structure outboard of the outboard root rib. A contributingfactor to the degree of restraint present could also be the twoclosely spaced loading formers and their attendant clamping actionon the structure just outboard of the outboard root rib.

Actually the moment restraint effects do not remain constant along

the span as noted by the comparison between the chordwise plots for

71

two different stations along the main beam fitting (Stations 67.6,and 55.6). Generally, the experimental strain data indicatesa gradually diminishing moment restraint from the outboard end ofthe fitting to the extreme root end. This is indicated by the re-duction in positive (+), or tension, strain values at the most inboardstation (Station 55.6).

the Initial splice joint analysis considered only those equivalentthermal strains for the axially-loaded splice which were defined bya thermal strain analysis of the full cross-section assuming thecross-section to be unrestrained-in-bending. In view of the experi-mental evidence which indicated the presence of some degree of thermalmoment restraint the 2500F. unsymmetrical and 4250 F. unsymmetricaltemperature environments (ASL tests 11 and 12) were also consideredanalytically assuming full thermal moment restraint along the lengthof the main beam fitting. The additional thermal moment is appliedas a couple load and produces additional tension in the lower surfacecover and main beam flange. The values of these thermal moments andcouple loads are shown in Table 4.3 and the net corrected staticallydeterminate load distributions are also shown in Table 4.3 for thetwo unsymmetrical temperature environments. £As a consequence of this investigation the summary tables for Part 2(Table 5.4) show two values for analytically predicted ulti-mate load and ultimate applied load ratios, as well as location offailure and mode of fe'lure. The values shown, then, represent thetwo extremes for thermal woment restraint (unrestrained and fully re-strained elastically). The good agreement between test and analysisresults for room temperature and the symmetrically heated temperatureenvironments indicates the specific criteria and parameters relatedto joint description, material properties, and local joint deformationare reasonably well defined for problem description. Therefore, it isexpected that the unsymmetrically heated cases should produce similartest-analysis comparison results since identical means of definingdeformation criteria and material parameters are employed. A signi-ficant trend is noted in the test-analysis comparison for the un-restrained-in-bending cases where the difference in the test andanalysis failing loads appears to be a function of the temperaturegradient with the largest difference occurring in the 4250F. unsym- -

metrical case.

The final results of Table 5.4, the indication of some degree ofthermal moment restraint and the diminishing restraint along the mainbeam fitting indicate that a rational assumption might be made forthermal moment restraint conditions in lieu of "exact" analyticalresults. For the type of structure typified by the horizontal stabi-lizer inboard splice joint, the localized heating, and significanttemperature gradients, the thermal moment restraint could be presumed

72

I

to be approximately one-half without producing significant errors inthe applied load ratios for room temperature simulation.

In Table 5.4 the "Failing Load Percent Dviation" column would thenshow about +9.2% for 250 F. Unsymmetrical and +5.1% for 425 0F. Un-

L symmetrical and the "Calculated Failing Load" column would show17,260 lbs. and 15,365 lbs., respectively.

Section 4.22 defines briefly the procedure used to compute the thermalmoment and shows the additional couple loads used as applied load in-put to approximate the full bending moment restraint condition.

The chordwise temperature distribution produced is peculiar to thisseries of elevated temperature tests, and was produced because theleading and trailing edge sections were shielded against direct radi-ation from the lamps while the main box area in the splice joint washeated.

7

[

L73

*KT is the thermal moment rotational restrain. strain and representsan elastic thermal moment Of MT in the case shown.L

Main BeamLFitting

BEAMAREA

N'N BEAM LX AREA

~~~ T,~N

IneBeam

LFront. Beb K T

FigureI .2 . 01c f hr a o en etan o dtoBELM \: tAREe to FiIe46frSai etSri aeRsls

I

4.7.1 Experimental Evaluation of Thermal Moment Restraint

The sketch in Figure4.27 is a generalization of the thermal momenteffects which can be presumed to be present at the outboard end ofthe steel main beam fitting. Since the main beam area only washeated and the front and rear beam areas were essentially protectedfrom direct lamp radiation the maximum temperatures and tempera-ture gradients through the depth of the cross-section tend to behigher in the main beam area. The steeper gradients in the main beamarea produce a greater tendency to bow or bend but this tendency is

- resisted by the front and rear bean areas which tend to remainstraight and unbowed since severe temperatures and temperature grad-ients are not present in these areas. This tendency plus inherenttorsional stiffness in the rib-skin combination at the outboard endof the main beam fitting could produce, essentially, a self-equi-librating restrained thermal moment condition similar to that shownin Figure 4.27. The coefficients of I" shown in Figure 4.27 on KTfor the front and rear beam areas are merely shown to representthe fact that the moment at any station is self-equilibrating and arenot intended to indicate that exactly half of the thermal moment isrestrained by either the front beam or rear beam area. If KT re-presents an elastically applied moment, MT, produced by the tempera-ture distribution the general equilibrium equation is

CfMT + CrMT = MT = (Cf + Cr)MT (4.7.1.1)

where Cf and Cr are merely unknown moment distribution factors forthe front beam and rear beam, respectively, and MT is the elasti-cally applied thermal moment in inch-lbs. The factors Cf and Crcannot be determined directly except by more rigorous indetermin-ate analysis techniques and more extensive temperature distributiondata over the leading edge box and trailing edge box areas. Further-more, any inherent torsional stiffness or axial load components inthe inboard rib area could produce similar boundary conditions atthe extreme root end. The extreme condition at both ends of themain beam fitting would define complete thermal bending moment re-straint for the main beam area where a constant moment would existalong the length of the main beam fitting. The maximum value ofthis bending .... endt in the momnt region would be equivalent tothe full value of KT (as elastic a plied strain), or MT (as an

-elastically applied bending moment) , as defined by the cross-section

thermal stress calculations. The values of the thermal moment, MT,are shown in Section 4. along with the associate I couple loads in

L the covers produced by considering MT as an additional elastic ap-plied moment. For calculation purposes for the restrained-in-bending case the couple loads, Pc, associated with the thermal

L75

[

It . . . .

moments, MT, are added, as appropriate, to the values in Tables2.2 and 2.3 as applied loads in the statically determinate struc-ture.

The thermal moment restraint system shown in Figure 4.27 is generallydemonstrated by the chordwise strain distribution near the outooardend of the main beam fitting. The chordwise strain distribution atthe stabilized temperature distribution just prior to the start ofloading is shown in Figure 428 for a cross-section near the out-board end of the main beam fitting for 2500F. Unsymmetrical (TestNo. 11) and 4500F. Unsymmetrical (Test No. 12) and for the most Iinboard cross-section (Elastic Axis Station 55.6) for Test No. 12.

The strain gages used for plotting purposes for the tension sideat the outboard station (Elastic Axis Station 67.6) were thoseidentified in Volume III as 23b, 24b, 25b, 26b, Z7b, 28b, and 29b.The strain gages plotted for the inboard station (Elastic AxisStation 55.6) are 6b, 7b, 8b, 9b, lOb, and lib.

For 4250F. Unsymmetrical (Test No. 12), calculations show the un-restrained-in-bending thermal straii, on the center main box skin Velement to be about -420 micro-inches arv the restrained-in-bendingthermal strain to be approximately +1260 micro-inches. The tem-perature moment strain, KT, alone is approximately +1680 micro-inches. For direct comparison of experimental and analyticalstrains the calculated maximum restrained value of +1260 micro-inches compares favorably with strain gage 26b which showed atension strain of +1790 micro-inches. This gage is located on theelastic axis near the outboard end of the main beam fitting wherethe outboard rib stiffness and the front and rear beam moment re-straint is the largest. The change in sign of the strain coupledwith the cross-section analysis comparison for the restrained andunrestrained cases and a favorable comparison with experimentalstrain values appears to indicate that some degree of thermal mo-ment restraint does exist. Therefore, the effects of this re- Fstraint must be allowed for in the equivalence of the full cross- Lsection to the axial.y-loaded splice joint analysis.

L

76

------ - T

Figure 4.28, Chordwise Strain Plots of Experimental DataFor Thermal Moment Restraint Investigation,

Inboard Splice Joint Study. Part 2

2800 d

Out board tation"t- .. - " E. . Sta 67.6)____ 250OF 3tn . NT t~ )i

Out oard StationL25F! (E L. p No- 12)r -, C-Inb # ard S ation

/ - (E.. Sta 55.6U I425 0 F Unsym i. (Test No. 12,

. I I• 2000 , .

0

I I

I

-- I .ulateS1 (5oF, nsymm ,Rest ained

r. 1200 -

-4I NSION

800

ICculaed%- -f15 - yin2.m -

<1 Rest ained

400- -

-1 0

Dist e Fro Elas tic klis, irches-400 1 77CCRESO

77

4.7.2 Thermal Moment Evaluation and Analysis Inpilt

In order to define equivalent axial plate elemeitt thermal strains(and input temperatures for splice plate analysis) and the thermalmoments for evaluation of the restrained vs. unrestrained momentcondition of the unsymmetrically heated splice joints,the cross-section analysis was performed on a simplified typical cross-section as shown in the following diagram.

Geometry andTemperature Symmetrical Main Box Skin

About L

Main Beam Fitting Web 2 Main BeamManBaitn _e _______ ____Fitting Flange

4 Main BeamFitting Flange

M Main Box Skin

Figure 4.29 Simplified Solice Joint Cross-Section

The actual temperatures obtained from the 250 0F. Unsymmetrical and rthe 4500F. Unsymmetrical tests were used as data input to the cross-section analysis for a +T (temperature applied) step only to obtainthe thermal strains, thermal stresses, and the thermal moment asoutput data. From this analysis the thermal moments for the twounsymmetrically heated conditions were found to be as follows.

2500F. Unsymmetrical (Test No. ii)

Thermal moment rotation, KT = .000705

From IBM data,ET r.2 j EnNn Yn, =, X.- -,, I 6 ) = .... x ,6

1~T-,2 ~-n y c - (, - - AC.-/

The thermal moment is then,

KTEI .000705(218 x 106) 102,500 in.-lbs.MT " c - 1.5

78

4500F. Unsymmetrical (Test No. 12)

Thermal moment rotation, KT = .00164

From IBM data, EI = 218 x 106 , c = 1.5 inches

The thermal moment is then,

KT EI .00164(218 x 106) = 23800iC 1.5

For an evaluation of the effects on the predicted ultimate strengthproduced by the presence of thermal bending moment restraints, theextreme restraint case can be considered in the analysis where thethermal moment is consid6red as an eastically applied constant mo-ment essentially over the length of th steel maij beam fittingbetween fastener-plate intersections 5 and 69. With theaverage couple height taken as 2.5 inches, additional couple loadsproduce an increase in tension stresses in the already-critical ten-sion side of the horizontal stabilizer splice. These values areshown below for the two temperature environments under consideration.

Couple Loads

i~02. 500Pc = 02.5 = 41000 lbs. (2500F. Unsy.mmetrical, Test No. 11)

Pc = 238,300 95300 lbs. (4500F. Unsymmetrical, Test No. 12)2.5

The above couple loads were added to the applied load input forultimate load prediction for the splice jointand the comparativeresults for the restrained versus unrestrained case are shown inTable 5.4 . The actual degree of restraint present in each tempera-ture case cannot be readily ascertainedbut some qualitative resultsand observations from strain gage data are described in Section 4.7.1.The couple loads shown above actually represent an approximate averagecouple load for the length of the splice jointi The actual valuesused in the analysis are tabulated on the following page.

L79

1 2 3 4 5 6 7 8

Test No. 11 Test No. 12

250°F. Unsymmetrical 425°F Unsymmotrical

Average -Element Couple Thermal Couple Couple Thermal Couple Average

No. Height, Moment LoadPc' Load in Moment, Load,Pc, Coupleinches M,in.-lbs. lbs. Element,Pc, MT,in.- lbs. l.oad in

Ref. lbs. lbs. ( Element,Page 78 _ _age 79) Pc, lbs.

56 2.38 102,500 43000 238,300 100,20042500 98950 T

57 2.44 42000 97,70041900 97500

58 2.45 41800 97,30051650 96900

59 2.47 41500 96,500 h41400 96350

60 2.48 41300 96,20091250 5, 96000

61 2.49 41200 95,80041100 95600

62 2.50 41000O 95,4004 0850 95000 C63 2.52 40700 94,6004 0350 93900

64 2.56 4000 93,200 [39700

9245065 2.60 39400 39200 91,70039200 i 91200 lr

66 2.63 39000 90,700 938700 90000

67 2.67 38400 89,300 U38050 88500

68 2.72 102,500 37700 238,300 87,700IiL

Table 4.3. Plate Element Axial Loads in Static Main BeamFitting Structure Due to Restrained-in BendingThermal Moment Effects. L

UL"

80

!

L 5.0 APPLIED LOAD RATIO q'MMART RD COLTAMSON

5.1 Part 1-Cross-Section Sumary

Table 5.1 summarizes the experimental and analytical results for thesix horizontal stabilizer components and the five temperature environ-ments. The data presented in the table provides a comparison betweenthe experimental and analytical results for the outboard former loadat failure, the station at which the failure occurred, the primarymode of failure, and the ultimate applied load ratio for each test.In this table the failure load only is considered for comparison pur-poses without regard for any difference in failure station. The loadshown in the table is the outboard former load since the bendingmoment at the critical cross-section is a function of that particularapplied load. A comparison of the failure locations shows generallygood agreement between the experimental and analytical results. Thesingle exception was Horizontal Stabilizer No. 1 which failed atElastic Axis Station 122 which is a significant departure from thenormal failure area between Elastic Axis Stations 102 and 105. How-ever, the failing load and the mode of failure show favorable agree-ment. The analytical procedures predict the room temperature failureto be compression instability at Elastic Axis Station 107.5 but thetangent bending moment line in Figure 4.7 passes through the testfailure data point at Station 122. This indicates that apparentlythe two stations could have been very nearly equally critical andthat a failure at Elastic Axis Station 107.5 was probably imminent.The proximity of the actual failure station to the outboard formerindicates the po3sible presence of severe shear lag effects whichviolates tne plane-strain cross-section assumptions in the analy-sis and may have contributed to the outboard station failure.

ii The room temperature outboard former loads, both experimental andanalytical, are the reference values for the applied load ratiocomputations. The experimental reference value used was the aver-age failing load of Horizontal Stabilizer tests 1 and Pilot Study a(14,900 lbs.). For each test the applied load ratios are definedfor ultimate load by

~VE.T.R(5.1.1)Rap: = VR. T. Ref.

L where Rap = applied load ratio

VE.T = outboard former elevated temperature ultimate loadL (analytical or experimental)

81

(VR.T.) outboard former room temperature reference ulti-Ref. mate load (analytical or experimental)

The comparison of the applied load ratios for each test is shown in 1Table 5.1 and, in general, shows good agreement between the pre-dicted ratio and those obtained experimentally.

For the 4250 F. (Symmetrical) test (Test No. 4) two values are givenin the table for analytical results. The purpose of the two sets,of values is to illustrate the importance of temperature-time de-pendent material property recovery data on the overall analyticalresults. During this particular test the heating lamps were turnedoff prior to the failure as a safety precaution. The element tem-peratures measured on the surface of the stabilizer indicated thata drop of approximately 70 0F. occurred on the skin between the timeof lamp shut-down and failure. The two failure loads listed inthe table represent the predicted strength based on the skin elementtemperatures measured at failure and also on the temperatures justprior to heating lamp shut-down. it is interesting to note thatthe overall test results tend to indicate that the material pro- -perties appear to reflect the temperature at the time of lamp shut-down rather than the lower temperature at failure. Subsequenttests of coupons taken from each of the test stabilizers did indi-cate the incomplete recovery of properties to the new low6r tem- rperature. This may be due to the fact that the stabilization ofcertain material properties, such as the yield stress, simply didnot have time to take place prior to the failure of the stabilizer.The results of the tensile coupon-stabilizer analysis data inputare summarized ia Table 4.2.

Table 5.2 is essentially a continuation of Table 5.1 and shows theper cent deviation of the predicted failing load from the testfailing load for each horizontal stabilizer. As in Table 5.1the outboard former load is used for comparison without consideringpredicted-failure- station ,te st-failure- station comparison.

To ensure a fair comparison of test and analytical results thesummary in Table 5.3 is presented to compare test and analytical Lbending moments at the test failure station. The per cent deviation

results from Tnhlk 5.2 combined with those from Table 5.3 provide amore complete basis for comparison of analytical and experimentalresults. The test ultimate bending moments shown are those presentat the actual failure station, and the predicted ultimate bendingmoments shown are those obtained from Figures 4.7 through 4.11 inSection 4.2 on the allowable bending moment curve at the testfailure station. As in the previous tables, two values are shownfor Test No. 4 to account for the two temperature conditions as Vpreviously discussed. L

From the summary tables it is noteworthy that the two room tempera-ture test failures produced in two supposedly identical horizontalstabilizer assemblies differed from each other by 17 inches inlocation and 16 percent in bending moment. This deviation inactual test results under identical environmental conditionsexceeds the test-analysis deviation for every test temperatureenvironment. Therefore, it can be presumed that the analytical-experimental deviations are probably within acceptable tolerancelimits for all of the tests including the two room temperaturetests.

83L.

v-I 0 IH000Lr\0 I- cQ

0)ccC 0' 0H

4H t0 0\ 0r 0

P

.4-,

a)*- 4.4 -- 'Co

gs 0 C: .4

4- H OH OH um~ C.)I O-H

091

44 00

to to w EZ l O C1 10 00 009: 0 0 F: C)S3 4)

L) H -42 00

0-s- 4-- 0 V(S0 0 0- .-ri

-P 0 r-( 140

H ~ 44'-o0, C\j -LO t \

43cx 0~ t 00j

0-.-

*0H4 0 0 r\ 0S4) (n- "0) CM C'93 0 0 HH c 4

00l Hy- H0m c l

0__ __ _ :2 0 p

rx pS. 114CDU

43 0)'0U

H -1:4 'P. t. \0 *

la to E- 00iUN C -

CM tr tl- -i 4

0-®x 100

Horiz. Test Test Outboard Former Load Per Cent DeviationStab. No. Temp. at Failure, lbs. From TestNo. Environ. Test Analysis

1 1 R.T. 14717 15230 + 3.40

2 2 - Te perature 3urvey Specimen

3 3 250 0F. 13256 13430 + 1.40Symm.

4f 14 4250F. 7716 f8921l * +13.50>Synm. t8360 -+ 8.oo

5 5 2500?. 13553 13800 + 1.80Unsym.

6 6 4250F. 7791 8o45 + 3.20U.Isymm.

PilotStudy R.T. 15100 15230 + 0.90

Table 5.2. Per Cent Deviation of Analysis Results from Test Ultimate Load,

This value denotes the analytical results obtained when the materialproperties are considered to recover to the higher values appropri-ate to the lower temperatures due to lamp shut off.

** This value denotes the results when the material properties were con-sidered to be those present at the higher temperature just prior tolamp shut-off.

85

T

"0 (D

ioriz. Test Test Ultimate Calculated Ultimate Per CentStab. Temp. Test Bending Moment, Bending Moment, Deviation 7No. Environ. (E.A. Sta.) inch-lbs. inch-lbs.

1 R.T. 122 192,000 230,000 o16.5

II2 -Temperattre Survey Specii en

3 250 0F. Symm. 103 427,000 436,000 + 2,1

4 4250F. SYmm. o 0 238,000 258,ooO0 7:$[ -78, ooo]* +14.3J*

5 250 0F. Unsym. 102 447,000 460,000 + 2.8

6 425 0F. Unsym. 103 248,OOO 258,000 + 3.8

Pilot -

Study R.T. 105 455,000 457,000 + 0.5a

Table 5.3. Bending Moment Comparison at Test Failure Station,k art I

This value denotes the analytical results obtained when the materialproperties are considered to recover to the higher values appropriateto the lower temperatures due to lamp shlit-off.

** This value denotes the results when the material properties were con-sidered to be those present at the higher temperature just prior tolamp shut-off. t

86

I!

5.2 Part 2 - Inboard Splice Joint Summary

Table 5.4 presents a complete summary of the experimental and ana-lytical results of the inboard splice joint portion of the study.Ultimate test loads are shown for each loading former, since, intheory, they should be equal. In addition a brief description ofthe test failure is shown for comparison with the failure mode pre-dicted by the analysis procedures. Using the room temperature resultsas a reference,the applied load ratios for ultimate load are calcu-lated and shown for each temperature environment for both analyticaland experimental ultimate loads. The per cent deviation of the pre-dicted ultimate load from the test ultimate load is also shown for

each temperature environment and shows generally acceptable agree-ment considering the large wimber of variables present in amechanically-fastened splice joint at elevated temperature.

Note that two calculated ultimate load values are shown for each ofthe unsymmetrically heated splice joints (Test No's. 11 and 12).The initial calculated results indicated by a double asterisk (**)were obtained by assuming no bending moment restraint for thermalstrains. The study performed in Sction 4.7 indicates that somedegree of thermal moment restraint probably exists in the rootregion due to temperature distribution and mixed materials. There-fore, a second set of calculations was performed assuming that fullbending moment restraint was present throughout the entire lengthof the joint. These results are noted by a triple asterisk (***)and show generally better agreement with the test values. From thestudy of Section 4.7 and the Test No. 12 results,it is apparent thatthe degree of thermal moment restraint lies somewhere between norestraint and full restraint. For example, the strain gage resultsfrom Test No. 12 tend to indicate approximately full bending momentrestraint at the outboard end of the splice and diminishing inboardto no restraint at the inboard end of the main box skin. In addi-tion the test failing load of approximately 14,300 lbs. (average)lies between the calculated values of 13,030 lbs. and 17,700 lbs.indicating that either extreme assumption was actually incorrect.Since the joint analysis procedures used were for axially-loadedsplices only, the thermal moment for the restrained case was takenas the elastic thermal moment and added to the applied loads as anadditional applied bending moment for purposes of this investigation.This applied bending moment was added to the ,_p "led loads as acouple load for axial load analysis where the value of this coupleload is shown in Section 4.7.2.

L

*. v.0 1A .0

0 o 0 0 0

1-44

0 4 0

0 N4

ci 00 to a *

24. 4. + .4 4+ I 4 .g

00

..3

0I W4 0. 0

4~.04U v II r r.4 to U.4 0 .0 0 .. (0 4 14 1 000 0

'44 '. ?,A r0 0 OC4' co '4'3 ON U t

A.UC' 0 U 0 U 040 '1 0.0 00 0-0 4' 0 0V40 Y zq .0 ~0 T, 0 0 r 0 0 *1. 1* 0 0

0 0 3 . 1 . .nl ~ . 0 4 40 %4 . U O . 4.0 .4 C .0 4 . 0 0 00 46.0 9 0 4 0 X C .4.4 .. ~.0 .j .4 9 5 0 .4. 4 -4

.4" Ni .4 N. C.' .4 N - 4 40 * [

0

03 0 0~ ~ 113 w. r '. . 0~00IS0 .3 fb 0 .4.0 D4

C.0 . C3 0 0 1. .

c14 0 4-3

~~~~~ N~~' * . 0a

.0004 to2.g .0 0 r0 r.

4. n 0 0, .80 ' 0 E, 4f.4 S.91 0C cC 0 ) U l fl: .0 c .,S .4 4 Q0. .00 40.4 0 0 L

o0 .-j .4 .; . 0I16j ) is0- 0 a4 0 4;94~ c .9 vA ~4l 1-00 40 +) 00.40 G4 C0 C IV0 lad

.0 C.o3 03 C___ iI0 o

4.0 -4 014 4 0o

0.4 0 . 0 C.' 4

-4- 040 c. 0,-C C6

1. 04

0~ aI1*

CV 88\ L

5.3 Arnproximate Avnlied Load Ratios Using Simplified Procedures

A simplified procedure previoasly demonstrated in Reference (d) wasemployed in order to establish the applicability of such techniquesin failure load prediction. The failure of a structure under testloading at elevated temperature is largely a function of the yieldstrength. The success at failure prediction in this program hascome as a result of careful determination of the yield stress of theelements of the stabilizer cross-section. The simplified procedureemployed here involves use of a psuedo yield moment obtained byconsidering each element in the cross-section at yield stress. Thecontribution of a single element in the cross-section to the valueof the yield moment is equal to the element yield stress times theelement area times the distance from the centroid of the element tothe centroid of the cross-section at temperature. This momentjequal toZ FyAn yn, is computed at room and elevated temperature. Then, by

Ydividing the room temperature "yield moment" value intoany of the elevated temperature "yield moment" values, an approximateapplied load ratio is obtained.

TestApplied Load Ratios Ratios

Test (TableCondition Sta. 91 Sta. 102 Sta. 113 Sta. 124 5.1

RoomTemperature 1.000 1.000 1.000 1.000 1.000

2500F. Un-Symmetrical 0.932 0.918 0.919 0.938 0.909

2500F.Symmetrical 0.905 0.887 0.884 0.917 0.888

425°F. Un-

Symmetrical 0.694 0.629 0.631 0.727 0.522

C 4250F. VSymmetrical-(ljemps off0) ,710 0.569 0.574 694 .517

4250 F.Symmetrical 0.700 0.525 0.516 1(lamps onn ___0 05 01 068 _

Table 5.5 Aplied Load Ratios bySimplified Method, Part 1

89

250 . USym.

0.807 at ta.lC

0.9

250. 0 'F ____m 0.383 __taL7.

F0

0

140.70_

.. 0.60 __ __ _

0.50 o, (am s Off)

4250 *FU Z0503 at Sta'. 107.5

- - ____ _____-4250 F. 33ym. _ _ _ _ _ _

.4 T ~Tst Applied Ioad R tios

+ +C itica Appled LodRato on t.lt~

90 100 110 120 130Elastic Axis Station, inches

Figure 5.1 Variation of Applied LoadRatio Wiith Station For Simplifis- MethodL

90

I

Table 5.5 shows the approximate applied load ratios computed at eachstation for the five tests. In the case of the 4250F. symmetricaltest, the values were computed for both lamps off at failure and forlamps on at failure. The difference in the values here is uauseddirectly by the difference in element yield stress for the two con-ditions.

Figure 5.1 shows plots of the approximate applied load ratios versuselastic axis station for each test. This plot provided not only awinimum applied load ratio for each test but also a failure station.Comparison with Table 5.1 , Section 5.1 , shows that the failurestation prediction by the approximate applied load ratio method isin very close agreement with the failure station prediction by theload-deformation procedures of Section 2.1.1. The applied loadratios, in turn, agree very well with the test applied load ratios.The only test which does not have good agreement is the 4250F.unsymmetrical test. The difference here is due to the fact that the

VI] approximate method does not consider thermal stresses. Thus, themethod will tend to over estimate the failure load for a test in thepresence of high therma gradients, and thus is unconservative.In the case of the 2500 F. unsymmetrical test, the thermal stressesare not high enough for the thermal gradient to have anr effect. Itwould appear from the results of this study that the simplified methodis quite satisfactory for failure prediction under symmetrical heating.For the ratios given in Table 5.5, only six elements out of the thirtyelements in the cross-section were used to compute the applied loadratios. These were the main box skin elements whose collapse causedfailure of the cross-section. It is not necessary in this approxi-mate method tc include all of the elements in the cross-section solong as the principal elements are considered.

L991

£2

6.0 EXAMPLE PROCEDURES FOR ELEVATED TEMPERATURE STRENGTHVERIFICATIOII

6.1 Example Procedure To Illustrate Cross-Section FailurePrediction at Elevated Temperature From Room TemperatureTest Results

The following procedure illustrates the method whereby the combineduse of analysis and room temperature test results could be used topredict elevated temperature strength for horizontal stabilizerbending investigated during Part 1 of this study.

1. Room temperature static test is performed and failure occurredat a reference load of 14,908 lbs. (average of HorizontalStabilizer No. 1 and No. 7).

2. Applied load ratios are obtained analytically using the cal-

culated room temperature failing load as the reference valuefor defining the theoretical applied load ratios outlined below.(See Table 5.1).(a) Temperature environment; 2500F. Symmetrical, Applied Load

Ratio, Rap = 0.883

(b) Temperature environment; 2500F. Unsymmetrical, AppliedLoad Ratio, Rap = 0.907

(c) Temperature environment; 4250F. SymmetricalApplied Load Ratio, Rap = 0.585 (Lamps off at failure)Applied Load Ratio, Rap = 0.550 (Lamps on at failure)

rd) Temperature environment; 4250 F. Unsymmetrical, Applied

Load Ratio, Rap = 0.528

3. The predicted failing loads for bending are shcwn in Table 6.1by applying the tnecretica] ratios of Item 2 to the test Lfailing load of Item 1. The table below also shows the actualtest failing loads for the elevated temperature conditions aswell as the per cent deviation of the predicted failing loadfrom the actual test load.

92

lb1 2 4

@a-0

j x 100

ASL Test Per CentTest Temperature Predicted Failing Load Actual Test DeviationNo. Condition From R.T. Test, lbs. Failing Load,lbs. from

L- Actual Test

- 3 2500F.Symm. 13180 13256 - 0.6

5 250 0F.Unsymm. 13530 13553 - 0.2

4 1425 0F.Symm. 8740 7716 +ll.7*8205 + 6.

L6_ 250F.Unsymm. 7880 7791 + 1.2

- * Lamps off at failure 1 Different Material-* Lamps assumed on at failure J Used.

Table 6.1. Test-Analysis Summry For ExampleProcedure, Part 1

The results shown in the above table show that the elevated temperaturestrength test results can be accurately predicted based upon calculatedapplied load ratios applied to room temperature test results. An actualsimulation procedure would use MIL-IIDBK-5 properties which would produce,generally, slightly conservative results. This fact seems to indicatethat the procedure could be accepted as a simulation procedure with somedegree of confidence.

L93

6.2 Example Procedure To Illustrate Splice Joint Failure Pre-diction at Elevated Temperature From Room Temperature Test Results

The following procedure illustrates the method whereby the combined useof analysis and room temperature test results could be used to predictelevated temperature strength for the horizontal stabilizer splice jointinvestigated chring this study.

1. Room temperature static test is performed and failure occurred ata reference load of 17,300 lbs.

2. Applied load ratios are obtained analytically using the calculatedroom temperature failing load as the reference value for definingthe theoretical applied load ratios outlined below.

(a) Temperature environment; 2500F. Symmetrical, Applied load -"

ratio, R = 0.958L(b) Temperature environment; 250 0F. Unsymmetrical, Applied load

ratio, R = 0.986 (R = 0.932 for thermal moment restraint)*(c) Temperature environment; 4250 F. Symmetrical, Applied load .

ratio, R = 0.690(d) Temperature environment; 4250F. Unsymmetrical, Applied load

ratio, R = 0.983 (R = 0.723 for thermal moment restraint)* L3. The predicted failing loads for the splice joint are calculated below

by applying the theoretical ratios of Item 2 to the test failing load fof Item 1. The table below also shows the test failing loads for the Lelevated temperature conditions as well as the per cent deviation ofthe predicted failing load from the actual test load.

I e Page 95

ASL Test Predicted Failing Per CentTest Temperature Load From R. T. Actual Test I eviation

No. Condition Test, lbs. Failing Load,lbs. From Actual Test

9 2500 F.Symm. 16575 15890 + 4.2

11 250 0 F.Uns5nm. 17070 15649 I + 8.316120* + 3.0*

10 4250F.Symm. 11940 12393 -3.5

12 1 425 F.Unsymm.1 1252 -12.6*1 172o_ _15.8

Table 6.2. rest-Analysis Jummary For ExampleProcedure, Part 2

The results of the Table 6.2 indicate that the predicted elevatedtemperature strength based upon calculated applied load ratios androom temperature test results may result in slightly unconservativepredictions except for Test No. 10 which was fairly realistic andslightly conservative. Possibly the use of MIL-HDBK-5 materialproperties and minimum fastener properties from a greater number oftests would ensure slightly conservative results in most cases.

See Table 6.2 on Page 94. 0

95

F.95

7.0 CONCLUSIONS

7.1 Conclusions on Part 1

In general, the results of this program indicate that the applied tiload ratio-room temperature static test simulation of elevated tem-perature design conditions can be applied to components or particularareas of components where continuous structum plane-strain cross-sections exist. The specific loading conditions satisfactorily veri-fied by this study are axial load, bending, and combined axial loadand bending, in uniform and nonuniform elevated temperature environ-ments. The further investigation of an approximate simplified ap-Lproach to the determination of applied load ratios has indicated thata relatively good estimate of elevated temperature strength can beobtained simply by using the 3levated temperature material propertiesand computing a "psuedo-yield bending moment" for the cross-section(bending critical).

The deflections of structures subjected to benaing in elevated tem-perature environments can be predicted analytically once the roomtemperature deflections are known, particularly maximum, or tip,deflections at failure. Some evidence supporting this conclusion Lwas obtained through the use of a simplified formula equating theratio of the elevated temperature test deflection and the room tem-perature test deflection to the ratio of test failure loads multiplied [by the ratio of the yield stress values of the major structural ele-ments at the test temperatures. The analytical load-deformation pro-cedures of Section 3.2, Volume I, coupled with the deflection analysisprocedures of Section 3.4, Volume I, can be used to predict thedeflected shape of a structure at failure. The prediction is limitedto plane-strain type structures such as are typically found in aircraft.The deflection rate method of failure station prediction is also shown Lto be of considerable value for elevated temperature tests as well asroom temperature tests.

Predictions are accurate to within +10% of the experimental values.

7.2 Conclusions on Part 2

Part 2 of this study demonstrated the application of applied loadratio-room temperature test simulation to multi-plate, multi-fastener, axially-loaded splice join s In uniform and nonuntiformtemperature environments. Using redundant force-type analyticalmethods which allow for temperature and inelastic effects, the ulti-inate applied load ratios were defined within realistic material pro-

perty limits and temperature ranges. Therefore, the elevated tempera-ture splice joint strengths were verified by analytically defined ap-plied load ratios and a room temperature static test to failure.

96

L2

II

In order for the applied load ratios to be calculated with sufficientaccuracy (and/or conservatism), MIL-HDBK-5 material properties shouldbe selected for the design elevated temperature environment to supple-ment those obtained experimentally by tests of fastener-plate com-

ij binations at room and elevated temperatures. The predictions areaccurate to within +16% of the experimental values.

L

FLII

121

II

8.0 RECOMNENDATIONS

Based upon the results of the work reported herein, there are limitationsinvolved with the use of the applied-load-ratio method. It is now ap-plicable to air'raft structures which fail in bending and/or axial load, Lunder steady-state elevated temperat re environments. It is thereforerecom ended that further investigations be performed so that all re-strictions could be removed.

The present investigation should be extended to the applicability of theapplied load-ratio method to:

(1) Full-scale structures, under steady-state temperature environ-ments, where failures are caused by torsion, combined torsionand bending, and by cut-outs in the structure. The methodsshould be verified by performing structural tests on suitablecomponents, such as stabilizers (for torsion, combined torsionand bending) and a typica2. fuselage (for structures with cut-outs).

(2) Laboratory-type structures, in other temperature environments,which are subjected to simple loadings. The "lamps-off" con-dition could be one of several additional temperature environ-ments. "

The applicability of the applied-load-ratio method to mechanically-fastened splice joints was established in Part 2 of the present in-vestigation. It is recommended that other types of fasteners be usedunder steady-state elevated temperatures and the same loading conditionsas reported herein; the splice joints could be welded or bonded.

98

I

APPENDIX I

PHOTOGRAPHS OF HORIZONTAL STABILIZER FAILURES

The photographs provided in this section illustrate the modes offailure which were encountered during the experimental program.Representative photographs of each failed stabilizer were selectedto clarify the failure descriptions in the summary tables ofSections 5.1 and 5.2. Figures AI.l through AI.6 show the com-pression sides of the outboard span plane-strain cross-sectionsafter failure. Particularly noteworthy in this series of photo-graphs is the similar mode of failure for the temperature rangefrom room temperature to approximately 4250 F. for this aluminumalloy multi-spar configuration. Generally the failures in theoutboard span regions were a compression instability failure ofthe aluminum alloy main box skin and spar flanges and in the samegeneral area. The single exception to this was the primary ten-sion failure of the pilot study stabilizer (Pilot study a) al-thouj'h Figure AI.I shows only the compression surface of thisstabilizer.

Figures AI.7 through AI.14 were selected to illustrate both themain beam fitting tension failures and the holo elongations in thealuminum alloy main box skins which occurred during the inboardI splice joint test program. As in the outboard regions of the sta-bilizer the failures are seen to be quite similar in mode andlocation for the temperature range from room temperature to ap-proximately 425°F. All major fractures and hole elongations shownin the photographs are on the tension surface.

99

FL[i 99

L

Figure AI.I. Horizontal Stabilizer Pilot Study a L(Pilot Study); Room Temperature; Closeup

of Compression Side After Failure F:

Figure AI.2. Horizontal Stabilizer Test 1;Room Temperature; Closeup of Compression

Side After Failure

0c

Io

[L

Figure AI.3. Horizontal Stabilizer Test 3;2500F. Symmetrical; Closeup of

Compression Surface After Failure

Figure AI . Hrzna Stabilizer Test 4

420.Symmnetrical; Closeup ofCompression Surface After Failure

101

Figure AI.5. Horizontal Stabilizer Test 5;2500 F. Unsymmetrical; Closeup ofCompression Surface After Failure

Figure AI.6. Horizontal Stabilizer Test 6; L4~25 0F. Unsymmetrical; Closeup ofCompression &arface,- After Failure

102L

' i

Figure AI.7. Horizontal Stabilizer Pilot Study bPilot Study Room Temperature; Closeup

Showing Cracks in Main Beam andForward Spar

'11

Figure AI.8. Horizontal Stabilizer Test 9;2500F. Symmetrical; Closeup of Tension

Surface After Fastener Failures

103

A' W-.Lk 3A,

prof"

Figare AI.lC Horzntal S~tabilizer Test 9;250 0F. Symmetrical; Closeup of Fair

Fiuof Maienal Flangtues

Cr4,.1"-1t~tL

k5 0 F Unymtrcl nbar ?or o

o25F Syenioncrae Afserpo Failure

105

I'L

++~o /o + 1

" [

Figure AI.13. Horizontal Stabilizer Test 112500F. Unsymmetrical; Closeup of Failure L

of Main Beam Tension Flanges

IL l x, k 1, 14 Ipa

Figure MJ.4 Horizontal Stabilizer Test 12;4250 F. Unsymmetrical; Closeup of Failure

of Main Beam Tension Flanges

lO

UNCLASS IFIEDSecurity Classification

DOC,'UMENT CONTROL DATA- R&D(Security claslificatlon of tit. body of abstsect and Indexing annotation muel be angerd when the overall report to claefied)

I ORIGINATIN G ACTIVITY (Corporate author) 2s. REPORT SECURITY C LASDIFICATIONNorth American Aviation, Inc., Columbus Division Unclassified

4300 East Fifth Avenue 2b GROUP

Columbus, Ohio 43216 None

3. REPORT TITLE

APPLICATION OF APPLIED LOAD RATIO STATIC TEST SIMULATION TECHNIQUES TOFULL-SCALE STRUCTURES; VOLUME IV, SUMMARY OF RESULTS, CONCLUSIONS ANDRECOMMENDATIONS

4. DESCRIPTIVE NOTES (Type of report and Inclusive date.)

Engineering Report - Volume IVS. AUTHOR(S) (Last none., toat name, Initial)

Gehring, R. W. and Maines, C. H.

6. REPORT DATE 74. TOTAL NO. OF PAGES 7b, NO. or REPS

August 1966 106 I 7S.. CONTRACT OR GRANT NO. 9A. ORIGINATOR'S REPORT NUMmtR(S)

N156-45330 NAEC-ASL-1094 Volume IVb. PROJECT NO.

P. A. 1-23-83C. 9b. OHER E[PORr NO(S) (Any othernummere that may ba ale IeS

Noned.

10. AVA ILAIILITY/LIMITATION NOTICES

Distribution of this document is unlimited.

I. SUPPLEMENTARY NOTES 12. SPONSORING MIUTARY ACTIVITY

Naval Air Engineering CenterNone Aeronautical Structures Laboratory

Philadelphia, Pennsylvania 19112

13. ANSTRACT

The overall objective cf the study is to investigate the applicabilityof the applied load ratio method to full scale aircraft structural components.Volume IV is a report on the correlation of theoretical and experimentalresults. The conclusion is that the applied load ratio method gives accutatepredictions for structures loaded in elevated temperature environments tofail in bending and shear. Failure predictions apply either to a mechanicallyfastened splice joint or to areas where plane-strain cross-sections exist.

FORM USFE

D D I1JAN064 14 3UNCLASSIFIEDSecurity Classffication_ .__. _

UNCLASSIFIEDSecurity Classification_

14. LINK A LINK 5 LINKcRot-KEY WORDS -ROL WT O La WT-

Block diagrans, structural analysis 4

Deflection of structuresDeformation of structuresDigital computers

Elevated temperature effectsJoint, splice-joints

Nonuniform temperature environment effectsSimulation studies

Static testsStrength analysis

Stress analysisStabilizers/Control surfaces/Structural analysis

Uniform temperature environment effects

INSTRUCTIONS1. ORIGINATING ACTIVITY* Enter the name and address imposed by security classification, using standard statementsof the contractor, subcontractor, grantee, Department of De- such as:tense activity or other organization (corporate author) Issuing (1) "Qualified requesters may obtain copies of thisthe report. report from DDC."2. REPORT SECUTITY CLASSIFICATION: Enter the ove, (2) "Foreign announcement and dissemination of thisall security classification of the report. Indicate whether"Restricted Data" Is Included. Marking In to be in accord- report by DDC is not authorized.ance with appnuiriate security regulations. (3) "U. S. Government ngencles may obtain copies of

this report directly from DDC. Other qualified DDC2b. GROUP: Automatic downgradingt is specified in DoD Di-usrshlreettrog

rective 5200. 10 and Armed Forces Industrial Manual. Enter users shall request through

the group nunber. Also, when applicable, show that optional .markings havb been used for Group 3 and Group 4 as author- (4) "U. S. military agencies may obtain copies of thisized. report directly from DDC. Other qualified users

3. REPORT TITLE. Enter the complete report title in all shall request throughcapital letters. Titles in all cases should be unclassified.If a meaningful title cannot be selected without classifica-tion, show title classification in all capitals in parenthesis (5) "All distribution of this report is controlled. Qual-immediately following the title. ified DDC users shall request through4. DESCRIPTIVE NOTES: If appropriate, enter the type of ,"report, e.g., interim, progress, summary, annual, or final. If the report has been furnished to the Office of TechnicalGive the inclusive dates when a specific reporting period is Services, Department of Commerce, for sale to the public, ndi-covered, cate this fact and enter the price, if known.5. AUTHOR(S): Enter the name(s) of author(s) as shown on IL SUPPLEMENTARY NOTES: Use for additional explans.or in the report. Enter last name, 'irat name, middle initial, tory notes.If military, show rank and beanch o.1 service. The name ofthe principal -uthor is an absolute minimum requirement. 12. SPONSORING MILITARY ACTIVITY: Enter the name of

the departmental project office or laboratory sponsoring (pay6. REPORT DATE Enter the date of the report as day, Ing for) the research and development. Include address.month, year, or month, year. If more than one dte appearson the report, use date of publication. 13. AP, STRACT: Enter an abstract giving a brief and factual -on. tOTA NUMBER, OFe PAGES: Te tota geo summary of the document indicative of the report, even though7s. TOTAL NUMBER OF PAGES- The total page count it may also appear elsewhere in the body of the technical re-should follow normal pagination procedures, Le., enter the port. If additional space is required, a continuation sheet ahallnumber of pages containing information. be attached.7b. NUMBER OF REFERENCES: Enter the total number of It is highly desirable that the abstract of classified reportsreferences cited in the report, be unclassified. Each paragraph of the abstract shall end with8&. CONTRACT OR GRANT NUMBER. If appropriate, enter an ivdication of the military security classification of the in-the applicable number of the contract or grant under which formation in the paragraph, represented 2s (Ts). (S), (C), or (U).the report was written. There is ro limitation on the hongth of the abstract. How-Sb, 8c, & 8d. PROJECT NUMBER. Enter the appropriate ever, the suggested length is from 150 to 225 words.military department Identification. such as project number,subproject number, system numbers, task number, etc. 14. KEY WORDS: Key words are technically meaningful terms

or short phrases that characterize a report and may be used as9a. ORIGINATOR'S REPORT NUMBER(S): Enter the offl- Index entries for cataloging the report Key words must becial report number by which the document will be identified selected so that no security clastification is required. Identi-and controlled by the originating activity. This number must fiers, such as equipment model designation, trade name, militarybe unique to this report, project code name, geographic location, may be used as key9b. OTHER REPORT NUMBER(S)- If the report has been words but will be followed by an indication of technical con-assigned any other report numbers (either by the originator text The assignment of links, rales, and weights Is optional.or by the sponsor), also enter this number(s).10. AVAILADILITY/LIMITATION NOTICES: Enter any lim-itations on further dissemination of the report, other than those

DD I 1473(BACK) UNCLASSIFIED